1200 files changed, 151238 insertions, 33183 deletions

diff --git a/drivers/atm/ambassador.c b/drivers/atm/ambassador.c
index 77a7480d..62a7607 100644
--- a/drivers/atm/ambassador.c
+++ b/drivers/atm/ambassador.c
@@ -1403,7 +1403,7 @@ static void amb_free_rx_skb (struct atm_vcc * atm_vcc, struct sk_buff * skb) {
   rx.host_address = cpu_to_be32 (virt_to_bus (skb->data));
   
   skb->data = skb->head;
-  skb->tail = skb->head;
+  skb_reset_tail_pointer(skb);
   skb->len = 0;
   
   if (!rx_give (dev, &rx, pool)) {
diff --git a/drivers/base/Makefile b/drivers/base/Makefile
index 4e22ce3..48029aa 100644
--- a/drivers/base/Makefile
+++ b/drivers/base/Makefile
@@ -10,7 +10,7 @@ obj-$(CONFIG_CMA) += dma-contiguous.o
 obj-y			+= power/
 obj-$(CONFIG_HAS_DMA)	+= dma-mapping.o
 obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o
-obj-$(CONFIG_DMA_SHARED_BUFFER) += dma-buf.o
+obj-$(CONFIG_DMA_SHARED_BUFFER) += dma-buf.o reservation.o
 obj-$(CONFIG_ISA)	+= isa.o
 obj-$(CONFIG_FW_LOADER)	+= firmware_class.o
 obj-$(CONFIG_NUMA)	+= node.o
diff --git a/drivers/base/reservation.c b/drivers/base/reservation.c
new file mode 100644
index 0000000..a73fbf3
--- /dev/null
+++ b/drivers/base/reservation.c
@@ -0,0 +1,39 @@
+/*
+ * Copyright (C) 2012-2013 Canonical Ltd
+ *
+ * Based on bo.c which bears the following copyright notice,
+ * but is dual licensed:
+ *
+ * Copyright (c) 2006-2009 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+/*
+ * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
+ */
+
+#include <linux/reservation.h>
+#include <linux/export.h>
+
+DEFINE_WW_CLASS(reservation_ww_class);
+EXPORT_SYMBOL(reservation_ww_class);
diff --git a/drivers/bcma/Kconfig b/drivers/bcma/Kconfig
index 8b4221c..380a200 100644
--- a/drivers/bcma/Kconfig
+++ b/drivers/bcma/Kconfig
@@ -26,6 +26,7 @@ config BCMA_HOST_PCI_POSSIBLE
 config BCMA_HOST_PCI
 	bool "Support for BCMA on PCI-host bus"
 	depends on BCMA_HOST_PCI_POSSIBLE
+	default y
 
 config BCMA_DRIVER_PCI_HOSTMODE
 	bool "Driver for PCI core working in hostmode"
diff --git a/drivers/bcma/bcma_private.h b/drivers/bcma/bcma_private.h
index 79595a0..0215f9a 100644
--- a/drivers/bcma/bcma_private.h
+++ b/drivers/bcma/bcma_private.h
@@ -22,6 +22,8 @@
 struct bcma_bus;
 
 /* main.c */
+bool bcma_wait_value(struct bcma_device *core, u16 reg, u32 mask, u32 value,
+		     int timeout);
 int bcma_bus_register(struct bcma_bus *bus);
 void bcma_bus_unregister(struct bcma_bus *bus);
 int __init bcma_bus_early_register(struct bcma_bus *bus,
diff --git a/drivers/bcma/core.c b/drivers/bcma/core.c
index 17b26ce..37a5ffe 100644
--- a/drivers/bcma/core.c
+++ b/drivers/bcma/core.c
@@ -9,6 +9,25 @@
 #include <linux/export.h>
 #include <linux/bcma/bcma.h>
 
+static bool bcma_core_wait_value(struct bcma_device *core, u16 reg, u32 mask,
+				 u32 value, int timeout)
+{
+	unsigned long deadline = jiffies + timeout;
+	u32 val;
+
+	do {
+		val = bcma_aread32(core, reg);
+		if ((val & mask) == value)
+			return true;
+		cpu_relax();
+		udelay(10);
+	} while (!time_after_eq(jiffies, deadline));
+
+	bcma_warn(core->bus, "Timeout waiting for register 0x%04X!\n", reg);
+
+	return false;
+}
+
 bool bcma_core_is_enabled(struct bcma_device *core)
 {
 	if ((bcma_aread32(core, BCMA_IOCTL) & (BCMA_IOCTL_CLK | BCMA_IOCTL_FGC))
@@ -25,13 +44,15 @@ void bcma_core_disable(struct bcma_device *core, u32 flags)
 	if (bcma_aread32(core, BCMA_RESET_CTL) & BCMA_RESET_CTL_RESET)
 		return;
 
-	bcma_awrite32(core, BCMA_IOCTL, flags);
-	bcma_aread32(core, BCMA_IOCTL);
-	udelay(10);
+	bcma_core_wait_value(core, BCMA_RESET_ST, ~0, 0, 300);
 
 	bcma_awrite32(core, BCMA_RESET_CTL, BCMA_RESET_CTL_RESET);
 	bcma_aread32(core, BCMA_RESET_CTL);
 	udelay(1);
+
+	bcma_awrite32(core, BCMA_IOCTL, flags);
+	bcma_aread32(core, BCMA_IOCTL);
+	udelay(10);
 }
 EXPORT_SYMBOL_GPL(bcma_core_disable);
 
@@ -43,6 +64,7 @@ int bcma_core_enable(struct bcma_device *core, u32 flags)
 	bcma_aread32(core, BCMA_IOCTL);
 
 	bcma_awrite32(core, BCMA_RESET_CTL, 0);
+	bcma_aread32(core, BCMA_RESET_CTL);
 	udelay(1);
 
 	bcma_awrite32(core, BCMA_IOCTL, (BCMA_IOCTL_CLK | flags));
diff --git a/drivers/bcma/driver_chipcommon.c b/drivers/bcma/driver_chipcommon.c
index 036c674..b068f98 100644
--- a/drivers/bcma/driver_chipcommon.c
+++ b/drivers/bcma/driver_chipcommon.c
@@ -140,8 +140,15 @@ void bcma_core_chipcommon_init(struct bcma_drv_cc *cc)
 	bcma_core_chipcommon_early_init(cc);
 
 	if (cc->core->id.rev >= 20) {
-		bcma_cc_write32(cc, BCMA_CC_GPIOPULLUP, 0);
-		bcma_cc_write32(cc, BCMA_CC_GPIOPULLDOWN, 0);
+		u32 pullup = 0, pulldown = 0;
+
+		if (cc->core->bus->chipinfo.id == BCMA_CHIP_ID_BCM43142) {
+			pullup = 0x402e0;
+			pulldown = 0x20500;
+		}
+
+		bcma_cc_write32(cc, BCMA_CC_GPIOPULLUP, pullup);
+		bcma_cc_write32(cc, BCMA_CC_GPIOPULLDOWN, pulldown);
 	}
 
 	if (cc->capabilities & BCMA_CC_CAP_PMU)
diff --git a/drivers/bcma/driver_chipcommon_pmu.c b/drivers/bcma/driver_chipcommon_pmu.c
index edca73a..5081a8c 100644
--- a/drivers/bcma/driver_chipcommon_pmu.c
+++ b/drivers/bcma/driver_chipcommon_pmu.c
@@ -56,6 +56,109 @@ void bcma_chipco_regctl_maskset(struct bcma_drv_cc *cc, u32 offset, u32 mask,
 }
 EXPORT_SYMBOL_GPL(bcma_chipco_regctl_maskset);
 
+static u32 bcma_pmu_xtalfreq(struct bcma_drv_cc *cc)
+{
+	u32 ilp_ctl, alp_hz;
+
+	if (!(bcma_cc_read32(cc, BCMA_CC_PMU_STAT) &
+	      BCMA_CC_PMU_STAT_EXT_LPO_AVAIL))
+		return 0;
+
+	bcma_cc_write32(cc, BCMA_CC_PMU_XTAL_FREQ,
+			BIT(BCMA_CC_PMU_XTAL_FREQ_MEASURE_SHIFT));
+	usleep_range(1000, 2000);
+
+	ilp_ctl = bcma_cc_read32(cc, BCMA_CC_PMU_XTAL_FREQ);
+	ilp_ctl &= BCMA_CC_PMU_XTAL_FREQ_ILPCTL_MASK;
+
+	bcma_cc_write32(cc, BCMA_CC_PMU_XTAL_FREQ, 0);
+
+	alp_hz = ilp_ctl * 32768 / 4;
+	return (alp_hz + 50000) / 100000 * 100;
+}
+
+static void bcma_pmu2_pll_init0(struct bcma_drv_cc *cc, u32 xtalfreq)
+{
+	struct bcma_bus *bus = cc->core->bus;
+	u32 freq_tgt_target = 0, freq_tgt_current;
+	u32 pll0, mask;
+
+	switch (bus->chipinfo.id) {
+	case BCMA_CHIP_ID_BCM43142:
+		/* pmu2_xtaltab0_adfll_485 */
+		switch (xtalfreq) {
+		case 12000:
+			freq_tgt_target = 0x50D52;
+			break;
+		case 20000:
+			freq_tgt_target = 0x307FE;
+			break;
+		case 26000:
+			freq_tgt_target = 0x254EA;
+			break;
+		case 37400:
+			freq_tgt_target = 0x19EF8;
+			break;
+		case 52000:
+			freq_tgt_target = 0x12A75;
+			break;
+		}
+		break;
+	}
+
+	if (!freq_tgt_target) {
+		bcma_err(bus, "Unknown TGT frequency for xtalfreq %d\n",
+			 xtalfreq);
+		return;
+	}
+
+	pll0 = bcma_chipco_pll_read(cc, BCMA_CC_PMU15_PLL_PLLCTL0);
+	freq_tgt_current = (pll0 & BCMA_CC_PMU15_PLL_PC0_FREQTGT_MASK) >>
+		BCMA_CC_PMU15_PLL_PC0_FREQTGT_SHIFT;
+
+	if (freq_tgt_current == freq_tgt_target) {
+		bcma_debug(bus, "Target TGT frequency already set\n");
+		return;
+	}
+
+	/* Turn off PLL */
+	switch (bus->chipinfo.id) {
+	case BCMA_CHIP_ID_BCM43142:
+		mask = (u32)~(BCMA_RES_4314_HT_AVAIL |
+			      BCMA_RES_4314_MACPHY_CLK_AVAIL);
+
+		bcma_cc_mask32(cc, BCMA_CC_PMU_MINRES_MSK, mask);
+		bcma_cc_mask32(cc, BCMA_CC_PMU_MAXRES_MSK, mask);
+		bcma_wait_value(cc->core, BCMA_CLKCTLST,
+				BCMA_CLKCTLST_HAVEHT, 0, 20000);
+		break;
+	}
+
+	pll0 &= ~BCMA_CC_PMU15_PLL_PC0_FREQTGT_MASK;
+	pll0 |= freq_tgt_target << BCMA_CC_PMU15_PLL_PC0_FREQTGT_SHIFT;
+	bcma_chipco_pll_write(cc, BCMA_CC_PMU15_PLL_PLLCTL0, pll0);
+
+	/* Flush */
+	if (cc->pmu.rev >= 2)
+		bcma_cc_set32(cc, BCMA_CC_PMU_CTL, BCMA_CC_PMU_CTL_PLL_UPD);
+
+	/* TODO: Do we need to update OTP? */
+}
+
+static void bcma_pmu_pll_init(struct bcma_drv_cc *cc)
+{
+	struct bcma_bus *bus = cc->core->bus;
+	u32 xtalfreq = bcma_pmu_xtalfreq(cc);
+
+	switch (bus->chipinfo.id) {
+	case BCMA_CHIP_ID_BCM43142:
+		if (xtalfreq == 0)
+			xtalfreq = 20000;
+		bcma_pmu2_pll_init0(cc, xtalfreq);
+		break;
+	}
+}
+
 static void bcma_pmu_resources_init(struct bcma_drv_cc *cc)
 {
 	struct bcma_bus *bus = cc->core->bus;
@@ -66,6 +169,25 @@ static void bcma_pmu_resources_init(struct bcma_drv_cc *cc)
 		min_msk = 0x200D;
 		max_msk = 0xFFFF;
 		break;
+	case BCMA_CHIP_ID_BCM43142:
+		min_msk = BCMA_RES_4314_LPLDO_PU |
+			  BCMA_RES_4314_PMU_SLEEP_DIS |
+			  BCMA_RES_4314_PMU_BG_PU |
+			  BCMA_RES_4314_CBUCK_LPOM_PU |
+			  BCMA_RES_4314_CBUCK_PFM_PU |
+			  BCMA_RES_4314_CLDO_PU |
+			  BCMA_RES_4314_LPLDO2_LVM |
+			  BCMA_RES_4314_WL_PMU_PU |
+			  BCMA_RES_4314_LDO3P3_PU |
+			  BCMA_RES_4314_OTP_PU |
+			  BCMA_RES_4314_WL_PWRSW_PU |
+			  BCMA_RES_4314_LQ_AVAIL |
+			  BCMA_RES_4314_LOGIC_RET |
+			  BCMA_RES_4314_MEM_SLEEP |
+			  BCMA_RES_4314_MACPHY_RET |
+			  BCMA_RES_4314_WL_CORE_READY;
+		max_msk = 0x3FFFFFFF;
+		break;
 	default:
 		bcma_debug(bus, "PMU resource config unknown or not needed for device 0x%04X\n",
 			   bus->chipinfo.id);
@@ -165,6 +287,7 @@ void bcma_pmu_init(struct bcma_drv_cc *cc)
 		bcma_cc_set32(cc, BCMA_CC_PMU_CTL,
 			     BCMA_CC_PMU_CTL_NOILPONW);
 
+	bcma_pmu_pll_init(cc);
 	bcma_pmu_resources_init(cc);
 	bcma_pmu_workarounds(cc);
 }
diff --git a/drivers/bcma/driver_chipcommon_sflash.c b/drivers/bcma/driver_chipcommon_sflash.c
index e6ed4fe..4d07cce 100644
--- a/drivers/bcma/driver_chipcommon_sflash.c
+++ b/drivers/bcma/driver_chipcommon_sflash.c
@@ -30,7 +30,7 @@ struct bcma_sflash_tbl_e {
 	u16 numblocks;
 };
 
-static struct bcma_sflash_tbl_e bcma_sflash_st_tbl[] = {
+static const struct bcma_sflash_tbl_e bcma_sflash_st_tbl[] = {
 	{ "M25P20", 0x11, 0x10000, 4, },
 	{ "M25P40", 0x12, 0x10000, 8, },
 
@@ -41,7 +41,7 @@ static struct bcma_sflash_tbl_e bcma_sflash_st_tbl[] = {
 	{ 0 },
 };
 
-static struct bcma_sflash_tbl_e bcma_sflash_sst_tbl[] = {
+static const struct bcma_sflash_tbl_e bcma_sflash_sst_tbl[] = {
 	{ "SST25WF512", 1, 0x1000, 16, },
 	{ "SST25VF512", 0x48, 0x1000, 16, },
 	{ "SST25WF010", 2, 0x1000, 32, },
@@ -59,7 +59,7 @@ static struct bcma_sflash_tbl_e bcma_sflash_sst_tbl[] = {
 	{ 0 },
 };
 
-static struct bcma_sflash_tbl_e bcma_sflash_at_tbl[] = {
+static const struct bcma_sflash_tbl_e bcma_sflash_at_tbl[] = {
 	{ "AT45DB011", 0xc, 256, 512, },
 	{ "AT45DB021", 0x14, 256, 1024, },
 	{ "AT45DB041", 0x1c, 256, 2048, },
@@ -89,7 +89,7 @@ int bcma_sflash_init(struct bcma_drv_cc *cc)
 {
 	struct bcma_bus *bus = cc->core->bus;
 	struct bcma_sflash *sflash = &cc->sflash;
-	struct bcma_sflash_tbl_e *e;
+	const struct bcma_sflash_tbl_e *e;
 	u32 id, id2;
 
 	switch (cc->capabilities & BCMA_CC_CAP_FLASHT) {
diff --git a/drivers/bcma/host_pci.c b/drivers/bcma/host_pci.c
index fbf2759..a355e63 100644
--- a/drivers/bcma/host_pci.c
+++ b/drivers/bcma/host_pci.c
@@ -275,6 +275,7 @@ static DEFINE_PCI_DEVICE_TABLE(bcma_pci_bridge_tbl) = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_BROADCOM, 0x4357) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_BROADCOM, 0x4358) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_BROADCOM, 0x4359) },
+	{ PCI_DEVICE(PCI_VENDOR_ID_BROADCOM, 0x4365) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_BROADCOM, 0x4727) },
 	{ 0, },
 };
diff --git a/drivers/bcma/main.c b/drivers/bcma/main.c
index f72f52b..0067422 100644
--- a/drivers/bcma/main.c
+++ b/drivers/bcma/main.c
@@ -93,6 +93,25 @@ struct bcma_device *bcma_find_core_unit(struct bcma_bus *bus, u16 coreid,
 	return NULL;
 }
 
+bool bcma_wait_value(struct bcma_device *core, u16 reg, u32 mask, u32 value,
+		     int timeout)
+{
+	unsigned long deadline = jiffies + timeout;
+	u32 val;
+
+	do {
+		val = bcma_read32(core, reg);
+		if ((val & mask) == value)
+			return true;
+		cpu_relax();
+		udelay(10);
+	} while (!time_after_eq(jiffies, deadline));
+
+	bcma_warn(core->bus, "Timeout waiting for register 0x%04X!\n", reg);
+
+	return false;
+}
+
 static void bcma_release_core_dev(struct device *dev)
 {
 	struct bcma_device *core = container_of(dev, struct bcma_device, dev);
diff --git a/drivers/bcma/sprom.c b/drivers/bcma/sprom.c
index 8934298..72bf454 100644
--- a/drivers/bcma/sprom.c
+++ b/drivers/bcma/sprom.c
@@ -72,12 +72,12 @@ fail:
  * R/W ops.
  **************************************************/
 
-static void bcma_sprom_read(struct bcma_bus *bus, u16 offset, u16 *sprom)
+static void bcma_sprom_read(struct bcma_bus *bus, u16 offset, u16 *sprom,
+			    size_t words)
 {
 	int i;
-	for (i = 0; i < SSB_SPROMSIZE_WORDS_R4; i++)
-		sprom[i] = bcma_read16(bus->drv_cc.core,
-				       offset + (i * 2));
+	for (i = 0; i < words; i++)
+		sprom[i] = bcma_read16(bus->drv_cc.core, offset + (i * 2));
 }
 
 /**************************************************
@@ -124,29 +124,29 @@ static inline u8 bcma_crc8(u8 crc, u8 data)
 	return t[crc ^ data];
 }
 
-static u8 bcma_sprom_crc(const u16 *sprom)
+static u8 bcma_sprom_crc(const u16 *sprom, size_t words)
 {
 	int word;
 	u8 crc = 0xFF;
 
-	for (word = 0; word < SSB_SPROMSIZE_WORDS_R4 - 1; word++) {
+	for (word = 0; word < words - 1; word++) {
 		crc = bcma_crc8(crc, sprom[word] & 0x00FF);
 		crc = bcma_crc8(crc, (sprom[word] & 0xFF00) >> 8);
 	}
-	crc = bcma_crc8(crc, sprom[SSB_SPROMSIZE_WORDS_R4 - 1] & 0x00FF);
+	crc = bcma_crc8(crc, sprom[words - 1] & 0x00FF);
 	crc ^= 0xFF;
 
 	return crc;
 }
 
-static int bcma_sprom_check_crc(const u16 *sprom)
+static int bcma_sprom_check_crc(const u16 *sprom, size_t words)
 {
 	u8 crc;
 	u8 expected_crc;
 	u16 tmp;
 
-	crc = bcma_sprom_crc(sprom);
-	tmp = sprom[SSB_SPROMSIZE_WORDS_R4 - 1] & SSB_SPROM_REVISION_CRC;
+	crc = bcma_sprom_crc(sprom, words);
+	tmp = sprom[words - 1] & SSB_SPROM_REVISION_CRC;
 	expected_crc = tmp >> SSB_SPROM_REVISION_CRC_SHIFT;
 	if (crc != expected_crc)
 		return -EPROTO;
@@ -154,21 +154,25 @@ static int bcma_sprom_check_crc(const u16 *sprom)
 	return 0;
 }
 
-static int bcma_sprom_valid(const u16 *sprom)
+static int bcma_sprom_valid(struct bcma_bus *bus, const u16 *sprom,
+			    size_t words)
 {
 	u16 revision;
 	int err;
 
-	err = bcma_sprom_check_crc(sprom);
+	err = bcma_sprom_check_crc(sprom, words);
 	if (err)
 		return err;
 
-	revision = sprom[SSB_SPROMSIZE_WORDS_R4 - 1] & SSB_SPROM_REVISION_REV;
-	if (revision != 8 && revision != 9) {
+	revision = sprom[words - 1] & SSB_SPROM_REVISION_REV;
+	if (revision != 8 && revision != 9 && revision != 10) {
 		pr_err("Unsupported SPROM revision: %d\n", revision);
 		return -ENOENT;
 	}
 
+	bus->sprom.revision = revision;
+	bcma_debug(bus, "Found SPROM revision %d\n", revision);
+
 	return 0;
 }
 
@@ -208,9 +212,6 @@ static void bcma_sprom_extract_r8(struct bcma_bus *bus, const u16 *sprom)
 	BUILD_BUG_ON(ARRAY_SIZE(pwr_info_offset) !=
 			ARRAY_SIZE(bus->sprom.core_pwr_info));
 
-	bus->sprom.revision = sprom[SSB_SPROMSIZE_WORDS_R4 - 1] &
-		SSB_SPROM_REVISION_REV;
-
 	for (i = 0; i < 3; i++) {
 		v = sprom[SPOFF(SSB_SPROM8_IL0MAC) + i];
 		*(((__be16 *)bus->sprom.il0mac) + i) = cpu_to_be16(v);
@@ -502,7 +503,7 @@ static bool bcma_sprom_onchip_available(struct bcma_bus *bus)
 	case BCMA_CHIP_ID_BCM4331:
 		present = chip_status & BCMA_CC_CHIPST_4331_OTP_PRESENT;
 		break;
-
+	case BCMA_CHIP_ID_BCM43142:
 	case BCMA_CHIP_ID_BCM43224:
 	case BCMA_CHIP_ID_BCM43225:
 		/* for these chips OTP is always available */
@@ -550,7 +551,9 @@ int bcma_sprom_get(struct bcma_bus *bus)
 {
 	u16 offset = BCMA_CC_SPROM;
 	u16 *sprom;
-	int err = 0;
+	size_t sprom_sizes[] = { SSB_SPROMSIZE_WORDS_R4,
+				 SSB_SPROMSIZE_WORDS_R10, };
+	int i, err = 0;
 
 	if (!bus->drv_cc.core)
 		return -EOPNOTSUPP;
@@ -579,32 +582,37 @@ int bcma_sprom_get(struct bcma_bus *bus)
 		}
 	}
 
-	sprom = kcalloc(SSB_SPROMSIZE_WORDS_R4, sizeof(u16),
-			GFP_KERNEL);
-	if (!sprom)
-		return -ENOMEM;
-
 	if (bus->chipinfo.id == BCMA_CHIP_ID_BCM4331 ||
 	    bus->chipinfo.id == BCMA_CHIP_ID_BCM43431)
 		bcma_chipco_bcm4331_ext_pa_lines_ctl(&bus->drv_cc, false);
 
 	bcma_debug(bus, "SPROM offset 0x%x\n", offset);
-	bcma_sprom_read(bus, offset, sprom);
+	for (i = 0; i < ARRAY_SIZE(sprom_sizes); i++) {
+		size_t words = sprom_sizes[i];
+
+		sprom = kcalloc(words, sizeof(u16), GFP_KERNEL);
+		if (!sprom)
+			return -ENOMEM;
+
+		bcma_sprom_read(bus, offset, sprom, words);
+		err = bcma_sprom_valid(bus, sprom, words);
+		if (!err)
+			break;
+
+		kfree(sprom);
+	}
 
 	if (bus->chipinfo.id == BCMA_CHIP_ID_BCM4331 ||
 	    bus->chipinfo.id == BCMA_CHIP_ID_BCM43431)
 		bcma_chipco_bcm4331_ext_pa_lines_ctl(&bus->drv_cc, true);
 
-	err = bcma_sprom_valid(sprom);
 	if (err) {
-		bcma_warn(bus, "invalid sprom read from the PCIe card, try to use fallback sprom\n");
+		bcma_warn(bus, "Invalid SPROM read from the PCIe card, trying to use fallback SPROM\n");
 		err = bcma_fill_sprom_with_fallback(bus, &bus->sprom);
-		goto out;
+	} else {
+		bcma_sprom_extract_r8(bus, sprom);
+		kfree(sprom);
 	}
 
-	bcma_sprom_extract_r8(bus, sprom);
-
-out:
-	kfree(sprom);
 	return err;
 }
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index aff789d..4ad2ad9 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -372,7 +372,7 @@ enum rbd_dev_flags {
 	RBD_DEV_FLAG_REMOVING,	/* this mapping is being removed */
 };
 
-static DEFINE_MUTEX(ctl_mutex);	  /* Serialize open/close/setup/teardown */
+static DEFINE_MUTEX(client_mutex);	/* Serialize client creation */
 
 static LIST_HEAD(rbd_dev_list);    /* devices */
 static DEFINE_SPINLOCK(rbd_dev_list_lock);
@@ -489,10 +489,8 @@ static int rbd_open(struct block_device *bdev, fmode_t mode)
 	if (removing)
 		return -ENOENT;
 
-	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
 	(void) get_device(&rbd_dev->dev);
 	set_device_ro(bdev, rbd_dev->mapping.read_only);
-	mutex_unlock(&ctl_mutex);
 
 	return 0;
 }
@@ -507,9 +505,7 @@ static void rbd_release(struct gendisk *disk, fmode_t mode)
 	spin_unlock_irq(&rbd_dev->lock);
 	rbd_assert(open_count_before > 0);
 
-	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
 	put_device(&rbd_dev->dev);
-	mutex_unlock(&ctl_mutex);
 }
 
 static const struct block_device_operations rbd_bd_ops = {
@@ -520,7 +516,7 @@ static const struct block_device_operations rbd_bd_ops = {
 
 /*
  * Initialize an rbd client instance.  Success or not, this function
- * consumes ceph_opts.
+ * consumes ceph_opts.  Caller holds client_mutex.
  */
 static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts)
 {
@@ -535,30 +531,25 @@ static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts)
 	kref_init(&rbdc->kref);
 	INIT_LIST_HEAD(&rbdc->node);
 
-	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
-
 	rbdc->client = ceph_create_client(ceph_opts, rbdc, 0, 0);
 	if (IS_ERR(rbdc->client))
-		goto out_mutex;
+		goto out_rbdc;
 	ceph_opts = NULL; /* Now rbdc->client is responsible for ceph_opts */
 
 	ret = ceph_open_session(rbdc->client);
 	if (ret < 0)
-		goto out_err;
+		goto out_client;
 
 	spin_lock(&rbd_client_list_lock);
 	list_add_tail(&rbdc->node, &rbd_client_list);
 	spin_unlock(&rbd_client_list_lock);
 
-	mutex_unlock(&ctl_mutex);
 	dout("%s: rbdc %p\n", __func__, rbdc);
 
 	return rbdc;
-
-out_err:
+out_client:
 	ceph_destroy_client(rbdc->client);
-out_mutex:
-	mutex_unlock(&ctl_mutex);
+out_rbdc:
 	kfree(rbdc);
 out_opt:
 	if (ceph_opts)
@@ -682,11 +673,13 @@ static struct rbd_client *rbd_get_client(struct ceph_options *ceph_opts)
 {
 	struct rbd_client *rbdc;
 
+	mutex_lock_nested(&client_mutex, SINGLE_DEPTH_NESTING);
 	rbdc = rbd_client_find(ceph_opts);
 	if (rbdc)	/* using an existing client */
 		ceph_destroy_options(ceph_opts);
 	else
 		rbdc = rbd_client_create(ceph_opts);
+	mutex_unlock(&client_mutex);
 
 	return rbdc;
 }
@@ -840,7 +833,6 @@ static int rbd_header_from_disk(struct rbd_device *rbd_dev,
 
 	/* We won't fail any more, fill in the header */
 
-	down_write(&rbd_dev->header_rwsem);
 	if (first_time) {
 		header->object_prefix = object_prefix;
 		header->obj_order = ondisk->options.order;
@@ -869,8 +861,6 @@ static int rbd_header_from_disk(struct rbd_device *rbd_dev,
 		if (rbd_dev->mapping.size != header->image_size)
 			rbd_dev->mapping.size = header->image_size;
 
-	up_write(&rbd_dev->header_rwsem);
-
 	return 0;
 out_2big:
 	ret = -EIO;
@@ -1126,6 +1116,7 @@ static void zero_bio_chain(struct bio *chain, int start_ofs)
 				buf = bvec_kmap_irq(bv, &flags);
 				memset(buf + remainder, 0,
 				       bv->bv_len - remainder);
+				flush_dcache_page(bv->bv_page);
 				bvec_kunmap_irq(buf, &flags);
 			}
 			pos += bv->bv_len;
@@ -1153,11 +1144,12 @@ static void zero_pages(struct page **pages, u64 offset, u64 end)
 		unsigned long flags;
 		void *kaddr;
 
-		page_offset = (size_t)(offset & ~PAGE_MASK);
-		length = min(PAGE_SIZE - page_offset, (size_t)(end - offset));
+		page_offset = offset & ~PAGE_MASK;
+		length = min_t(size_t, PAGE_SIZE - page_offset, end - offset);
 		local_irq_save(flags);
 		kaddr = kmap_atomic(*page);
 		memset(kaddr + page_offset, 0, length);
+		flush_dcache_page(*page);
 		kunmap_atomic(kaddr);
 		local_irq_restore(flags);
 
@@ -2171,9 +2163,9 @@ static int rbd_img_request_fill(struct rbd_img_request *img_request,
 	struct rbd_obj_request *obj_request = NULL;
 	struct rbd_obj_request *next_obj_request;
 	bool write_request = img_request_write_test(img_request);
-	struct bio *bio_list;
+	struct bio *bio_list = 0;
 	unsigned int bio_offset = 0;
-	struct page **pages;
+	struct page **pages = 0;
 	u64 img_offset;
 	u64 resid;
 	u16 opcode;
@@ -2535,6 +2527,7 @@ static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request)
 	 */
 	orig_request = obj_request->obj_request;
 	obj_request->obj_request = NULL;
+	rbd_obj_request_put(orig_request);
 	rbd_assert(orig_request);
 	rbd_assert(orig_request->img_request);
 
@@ -2555,7 +2548,6 @@ static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request)
 	if (!rbd_dev->parent_overlap) {
 		struct ceph_osd_client *osdc;
 
-		rbd_obj_request_put(orig_request);
 		osdc = &rbd_dev->rbd_client->client->osdc;
 		result = rbd_obj_request_submit(osdc, orig_request);
 		if (!result)
@@ -2585,7 +2577,6 @@ static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request)
 out:
 	if (orig_request->result)
 		rbd_obj_request_complete(orig_request);
-	rbd_obj_request_put(orig_request);
 }
 
 static int rbd_img_obj_exists_submit(struct rbd_obj_request *obj_request)
@@ -2859,7 +2850,7 @@ static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
 		(unsigned int)opcode);
 	ret = rbd_dev_refresh(rbd_dev);
 	if (ret)
-		rbd_warn(rbd_dev, ": header refresh error (%d)\n", ret);
+		rbd_warn(rbd_dev, "header refresh error (%d)\n", ret);
 
 	rbd_obj_notify_ack(rbd_dev, notify_id);
 }
@@ -3339,8 +3330,8 @@ static int rbd_dev_refresh(struct rbd_device *rbd_dev)
 	int ret;
 
 	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+	down_write(&rbd_dev->header_rwsem);
 	mapping_size = rbd_dev->mapping.size;
-	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
 	if (rbd_dev->image_format == 1)
 		ret = rbd_dev_v1_header_info(rbd_dev);
 	else
@@ -3349,7 +3340,8 @@ static int rbd_dev_refresh(struct rbd_device *rbd_dev)
 	/* If it's a mapped snapshot, validate its EXISTS flag */
 
 	rbd_exists_validate(rbd_dev);
-	mutex_unlock(&ctl_mutex);
+	up_write(&rbd_dev->header_rwsem);
+
 	if (mapping_size != rbd_dev->mapping.size) {
 		sector_t size;
 
@@ -3813,6 +3805,7 @@ static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev)
 	void *end;
 	u64 pool_id;
 	char *image_id;
+	u64 snap_id;
 	u64 overlap;
 	int ret;
 
@@ -3872,24 +3865,56 @@ static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev)
 			(unsigned long long)pool_id, U32_MAX);
 		goto out_err;
 	}
-	parent_spec->pool_id = pool_id;
 
 	image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
 	if (IS_ERR(image_id)) {
 		ret = PTR_ERR(image_id);
 		goto out_err;
 	}
-	parent_spec->image_id = image_id;
-	ceph_decode_64_safe(&p, end, parent_spec->snap_id, out_err);
+	ceph_decode_64_safe(&p, end, snap_id, out_err);
 	ceph_decode_64_safe(&p, end, overlap, out_err);
 
-	if (overlap) {
-		rbd_spec_put(rbd_dev->parent_spec);
+	/*
+	 * The parent won't change (except when the clone is
+	 * flattened, already handled that).  So we only need to
+	 * record the parent spec we have not already done so.
+	 */
+	if (!rbd_dev->parent_spec) {
+		parent_spec->pool_id = pool_id;
+		parent_spec->image_id = image_id;
+		parent_spec->snap_id = snap_id;
 		rbd_dev->parent_spec = parent_spec;
 		parent_spec = NULL;	/* rbd_dev now owns this */
-		rbd_dev->parent_overlap = overlap;
-	} else {
-		rbd_warn(rbd_dev, "ignoring parent of clone with overlap 0\n");
+	}
+
+	/*
+	 * We always update the parent overlap.  If it's zero we
+	 * treat it specially.
+	 */
+	rbd_dev->parent_overlap = overlap;
+	smp_mb();
+	if (!overlap) {
+
+		/* A null parent_spec indicates it's the initial probe */
+
+		if (parent_spec) {
+			/*
+			 * The overlap has become zero, so the clone
+			 * must have been resized down to 0 at some
+			 * point.  Treat this the same as a flatten.
+			 */
+			rbd_dev_parent_put(rbd_dev);
+			pr_info("%s: clone image now standalone\n",
+				rbd_dev->disk->disk_name);
+		} else {
+			/*
+			 * For the initial probe, if we find the
+			 * overlap is zero we just pretend there was
+			 * no parent image.
+			 */
+			rbd_warn(rbd_dev, "ignoring parent of "
+						"clone with overlap 0\n");
+		}
 	}
 out:
 	ret = 0;
@@ -4245,16 +4270,14 @@ static int rbd_dev_v2_header_info(struct rbd_device *rbd_dev)
 	bool first_time = rbd_dev->header.object_prefix == NULL;
 	int ret;
 
-	down_write(&rbd_dev->header_rwsem);
-
 	ret = rbd_dev_v2_image_size(rbd_dev);
 	if (ret)
-		goto out;
+		return ret;
 
 	if (first_time) {
 		ret = rbd_dev_v2_header_onetime(rbd_dev);
 		if (ret)
-			goto out;
+			return ret;
 	}
 
 	/*
@@ -4269,7 +4292,7 @@ static int rbd_dev_v2_header_info(struct rbd_device *rbd_dev)
 
 		ret = rbd_dev_v2_parent_info(rbd_dev);
 		if (ret)
-			goto out;
+			return ret;
 
 		/*
 		 * Print a warning if this is the initial probe and
@@ -4290,8 +4313,6 @@ static int rbd_dev_v2_header_info(struct rbd_device *rbd_dev)
 
 	ret = rbd_dev_v2_snap_context(rbd_dev);
 	dout("rbd_dev_v2_snap_context returned %d\n", ret);
-out:
-	up_write(&rbd_dev->header_rwsem);
 
 	return ret;
 }
@@ -4301,8 +4322,6 @@ static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
 	struct device *dev;
 	int ret;
 
-	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
-
 	dev = &rbd_dev->dev;
 	dev->bus = &rbd_bus_type;
 	dev->type = &rbd_device_type;
@@ -4311,8 +4330,6 @@ static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
 	dev_set_name(dev, "%d", rbd_dev->dev_id);
 	ret = device_register(dev);
 
-	mutex_unlock(&ctl_mutex);
-
 	return ret;
 }
 
@@ -5059,23 +5076,6 @@ err_out_module:
 	return (ssize_t)rc;
 }
 
-static struct rbd_device *__rbd_get_dev(unsigned long dev_id)
-{
-	struct list_head *tmp;
-	struct rbd_device *rbd_dev;
-
-	spin_lock(&rbd_dev_list_lock);
-	list_for_each(tmp, &rbd_dev_list) {
-		rbd_dev = list_entry(tmp, struct rbd_device, node);
-		if (rbd_dev->dev_id == dev_id) {
-			spin_unlock(&rbd_dev_list_lock);
-			return rbd_dev;
-		}
-	}
-	spin_unlock(&rbd_dev_list_lock);
-	return NULL;
-}
-
 static void rbd_dev_device_release(struct device *dev)
 {
 	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
@@ -5120,8 +5120,10 @@ static ssize_t rbd_remove(struct bus_type *bus,
 			  size_t count)
 {
 	struct rbd_device *rbd_dev = NULL;
-	int target_id;
+	struct list_head *tmp;
+	int dev_id;
 	unsigned long ul;
+	bool already = false;
 	int ret;
 
 	ret = strict_strtoul(buf, 10, &ul);
@@ -5129,37 +5131,40 @@ static ssize_t rbd_remove(struct bus_type *bus,
 		return ret;
 
 	/* convert to int; abort if we lost anything in the conversion */
-	target_id = (int) ul;
-	if (target_id != ul)
+	dev_id = (int)ul;
+	if (dev_id != ul)
 		return -EINVAL;
 
-	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
-
-	rbd_dev = __rbd_get_dev(target_id);
-	if (!rbd_dev) {
-		ret = -ENOENT;
-		goto done;
+	ret = -ENOENT;
+	spin_lock(&rbd_dev_list_lock);
+	list_for_each(tmp, &rbd_dev_list) {
+		rbd_dev = list_entry(tmp, struct rbd_device, node);
+		if (rbd_dev->dev_id == dev_id) {
+			ret = 0;
+			break;
+		}
+	}
+	if (!ret) {
+		spin_lock_irq(&rbd_dev->lock);
+		if (rbd_dev->open_count)
+			ret = -EBUSY;
+		else
+			already = test_and_set_bit(RBD_DEV_FLAG_REMOVING,
+							&rbd_dev->flags);
+		spin_unlock_irq(&rbd_dev->lock);
 	}
+	spin_unlock(&rbd_dev_list_lock);
+	if (ret < 0 || already)
+		return ret;
 
-	spin_lock_irq(&rbd_dev->lock);
-	if (rbd_dev->open_count)
-		ret = -EBUSY;
-	else
-		set_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags);
-	spin_unlock_irq(&rbd_dev->lock);
-	if (ret < 0)
-		goto done;
 	rbd_bus_del_dev(rbd_dev);
 	ret = rbd_dev_header_watch_sync(rbd_dev, false);
 	if (ret)
 		rbd_warn(rbd_dev, "failed to cancel watch event (%d)\n", ret);
 	rbd_dev_image_release(rbd_dev);
 	module_put(THIS_MODULE);
-	ret = count;
-done:
-	mutex_unlock(&ctl_mutex);
 
-	return ret;
+	return count;
 }
 
 /*
@@ -5267,6 +5272,7 @@ static void __exit rbd_exit(void)
 module_init(rbd_init);
 module_exit(rbd_exit);
 
+MODULE_AUTHOR("Alex Elder <elder@inktank.com>");
 MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
 MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
 MODULE_DESCRIPTION("rados block device");
diff --git a/drivers/bluetooth/btmrvl_sdio.c b/drivers/bluetooth/btmrvl_sdio.c
index 13693b7..75c2626 100644
--- a/drivers/bluetooth/btmrvl_sdio.c
+++ b/drivers/bluetooth/btmrvl_sdio.c
@@ -554,6 +554,7 @@ static int btmrvl_sdio_card_to_host(struct btmrvl_private *priv)
 	skb = bt_skb_alloc(num_blocks * blksz + BTSDIO_DMA_ALIGN, GFP_ATOMIC);
 	if (skb == NULL) {
 		BT_ERR("No free skb");
+		ret = -ENOMEM;
 		goto exit;
 	}
 
diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
index 7a7e5f8..de4cf4d 100644
--- a/drivers/bluetooth/btusb.c
+++ b/drivers/bluetooth/btusb.c
@@ -57,6 +57,9 @@ static struct usb_device_id btusb_table[] = {
 	/* Apple-specific (Broadcom) devices */
 	{ USB_VENDOR_AND_INTERFACE_INFO(0x05ac, 0xff, 0x01, 0x01) },
 
+	/* MediaTek MT76x0E */
+	{ USB_DEVICE(0x0e8d, 0x763f) },
+
 	/* Broadcom SoftSailing reporting vendor specific */
 	{ USB_DEVICE(0x0a5c, 0x21e1) },
 
@@ -1616,6 +1619,7 @@ static struct usb_driver btusb_driver = {
 #ifdef CONFIG_PM
 	.suspend	= btusb_suspend,
 	.resume		= btusb_resume,
+	.reset_resume	= btusb_resume,
 #endif
 	.id_table	= btusb_table,
 	.supports_autosuspend = 1,
diff --git a/drivers/char/agp/ati-agp.c b/drivers/char/agp/ati-agp.c
index 0628d7b..03c1dc1 100644
--- a/drivers/char/agp/ati-agp.c
+++ b/drivers/char/agp/ati-agp.c
@@ -236,14 +236,14 @@ static int ati_configure(void)
 static int agp_ati_suspend(struct pci_dev *dev, pm_message_t state)
 {
 	pci_save_state(dev);
-	pci_set_power_state(dev, 3);
+	pci_set_power_state(dev, PCI_D3hot);
 
 	return 0;
 }
 
 static int agp_ati_resume(struct pci_dev *dev)
 {
-	pci_set_power_state(dev, 0);
+	pci_set_power_state(dev, PCI_D0);
 	pci_restore_state(dev);
 
 	return ati_configure();
diff --git a/drivers/char/agp/frontend.c b/drivers/char/agp/frontend.c
index 2e04433..1b19239 100644
--- a/drivers/char/agp/frontend.c
+++ b/drivers/char/agp/frontend.c
@@ -603,7 +603,8 @@ static int agp_mmap(struct file *file, struct vm_area_struct *vma)
 			vma->vm_ops = kerninfo.vm_ops;
 		} else if (io_remap_pfn_range(vma, vma->vm_start,
 				(kerninfo.aper_base + offset) >> PAGE_SHIFT,
-					    size, vma->vm_page_prot)) {
+				size,
+				pgprot_writecombine(vma->vm_page_prot))) {
 			goto out_again;
 		}
 		mutex_unlock(&(agp_fe.agp_mutex));
@@ -618,8 +619,9 @@ static int agp_mmap(struct file *file, struct vm_area_struct *vma)
 		if (kerninfo.vm_ops) {
 			vma->vm_ops = kerninfo.vm_ops;
 		} else if (io_remap_pfn_range(vma, vma->vm_start,
-					    kerninfo.aper_base >> PAGE_SHIFT,
-					    size, vma->vm_page_prot)) {
+				kerninfo.aper_base >> PAGE_SHIFT,
+				size,
+				pgprot_writecombine(vma->vm_page_prot))) {
 			goto out_again;
 		}
 		mutex_unlock(&(agp_fe.agp_mutex));
diff --git a/drivers/char/agp/nvidia-agp.c b/drivers/char/agp/nvidia-agp.c
index 62be3ec..be42a23 100644
--- a/drivers/char/agp/nvidia-agp.c
+++ b/drivers/char/agp/nvidia-agp.c
@@ -399,8 +399,8 @@ static void agp_nvidia_remove(struct pci_dev *pdev)
 #ifdef CONFIG_PM
 static int agp_nvidia_suspend(struct pci_dev *pdev, pm_message_t state)
 {
-	pci_save_state (pdev);
-	pci_set_power_state (pdev, 3);
+	pci_save_state(pdev);
+	pci_set_power_state(pdev, PCI_D3hot);
 
 	return 0;
 }
@@ -408,7 +408,7 @@ static int agp_nvidia_suspend(struct pci_dev *pdev, pm_message_t state)
 static int agp_nvidia_resume(struct pci_dev *pdev)
 {
 	/* set power state 0 and restore PCI space */
-	pci_set_power_state (pdev, 0);
+	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 
 	/* reconfigure AGP hardware again */
diff --git a/drivers/char/mwave/tp3780i.c b/drivers/char/mwave/tp3780i.c
index c689697..04e6d6a 100644
--- a/drivers/char/mwave/tp3780i.c
+++ b/drivers/char/mwave/tp3780i.c
@@ -479,6 +479,7 @@ int tp3780I_QueryAbilities(THINKPAD_BD_DATA * pBDData, MW_ABILITIES * pAbilities
 	PRINTK_2(TRACE_TP3780I,
 		"tp3780i::tp3780I_QueryAbilities entry pBDData %p\n", pBDData);
 
+	memset(pAbilities, 0, sizeof(*pAbilities));
 	/* fill out standard constant fields */
 	pAbilities->instr_per_sec = pBDData->rDspSettings.uIps;
 	pAbilities->data_size = pBDData->rDspSettings.uDStoreSize;
diff --git a/drivers/crypto/talitos.c b/drivers/crypto/talitos.c
index 5b2b5e6..661dc3e 100644
--- a/drivers/crypto/talitos.c
+++ b/drivers/crypto/talitos.c
@@ -1112,64 +1112,6 @@ static int sg_count(struct scatterlist *sg_list, int nbytes, bool *chained)
 	return sg_nents;
 }
 
-/**
- * sg_copy_end_to_buffer - Copy end data from SG list to a linear buffer
- * @sgl:		 The SG list
- * @nents:		 Number of SG entries
- * @buf:		 Where to copy to
- * @buflen:		 The number of bytes to copy
- * @skip:		 The number of bytes to skip before copying.
- *                       Note: skip + buflen should equal SG total size.
- *
- * Returns the number of copied bytes.
- *
- **/
-static size_t sg_copy_end_to_buffer(struct scatterlist *sgl, unsigned int nents,
-				    void *buf, size_t buflen, unsigned int skip)
-{
-	unsigned int offset = 0;
-	unsigned int boffset = 0;
-	struct sg_mapping_iter miter;
-	unsigned long flags;
-	unsigned int sg_flags = SG_MITER_ATOMIC;
-	size_t total_buffer = buflen + skip;
-
-	sg_flags |= SG_MITER_FROM_SG;
-
-	sg_miter_start(&miter, sgl, nents, sg_flags);
-
-	local_irq_save(flags);
-
-	while (sg_miter_next(&miter) && offset < total_buffer) {
-		unsigned int len;
-		unsigned int ignore;
-
-		if ((offset + miter.length) > skip) {
-			if (offset < skip) {
-				/* Copy part of this segment */
-				ignore = skip - offset;
-				len = miter.length - ignore;
-				if (boffset + len > buflen)
-					len = buflen - boffset;
-				memcpy(buf + boffset, miter.addr + ignore, len);
-			} else {
-				/* Copy all of this segment (up to buflen) */
-				len = miter.length;
-				if (boffset + len > buflen)
-					len = buflen - boffset;
-				memcpy(buf + boffset, miter.addr, len);
-			}
-			boffset += len;
-		}
-		offset += miter.length;
-	}
-
-	sg_miter_stop(&miter);
-
-	local_irq_restore(flags);
-	return boffset;
-}
-
 /*
  * allocate and map the extended descriptor
  */
@@ -1800,7 +1742,7 @@ static int ahash_process_req(struct ahash_request *areq, unsigned int nbytes)
 
 	if (to_hash_later) {
 		int nents = sg_count(areq->src, nbytes, &chained);
-		sg_copy_end_to_buffer(areq->src, nents,
+		sg_pcopy_to_buffer(areq->src, nents,
 				      req_ctx->bufnext,
 				      to_hash_later,
 				      nbytes - to_hash_later);
diff --git a/drivers/dma/iop-adma.c b/drivers/dma/iop-adma.c
index c9cc08c..cc727ec 100644
--- a/drivers/dma/iop-adma.c
+++ b/drivers/dma/iop-adma.c
@@ -1017,7 +1017,7 @@ iop_adma_xor_val_self_test(struct iop_adma_device *device)
 	struct page *xor_srcs[IOP_ADMA_NUM_SRC_TEST];
 	struct page *zero_sum_srcs[IOP_ADMA_NUM_SRC_TEST + 1];
 	dma_addr_t dma_srcs[IOP_ADMA_NUM_SRC_TEST + 1];
-	dma_addr_t dma_addr, dest_dma;
+	dma_addr_t dest_dma;
 	struct dma_async_tx_descriptor *tx;
 	struct dma_chan *dma_chan;
 	dma_cookie_t cookie;
@@ -1516,7 +1516,7 @@ static int iop_adma_probe(struct platform_device *pdev)
 			goto err_free_iop_chan;
 	}
 
-	dev_info(&pdev->dev, "Intel(R) IOP: ( %s%s%s%s%s%s%s)\n",
+	dev_info(&pdev->dev, "Intel(R) IOP: ( %s%s%s%s%s%s)\n",
 		 dma_has_cap(DMA_PQ, dma_dev->cap_mask) ? "pq " : "",
 		 dma_has_cap(DMA_PQ_VAL, dma_dev->cap_mask) ? "pq_val " : "",
 		 dma_has_cap(DMA_XOR, dma_dev->cap_mask) ? "xor " : "",
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index b16c50e..a7c54c8 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -139,6 +139,7 @@ config DRM_I915
 	select BACKLIGHT_CLASS_DEVICE if ACPI
 	select VIDEO_OUTPUT_CONTROL if ACPI
 	select INPUT if ACPI
+	select THERMAL if ACPI
 	select ACPI_VIDEO if ACPI
 	select ACPI_BUTTON if ACPI
 	help
@@ -213,6 +214,8 @@ source "drivers/gpu/drm/mgag200/Kconfig"
 
 source "drivers/gpu/drm/cirrus/Kconfig"
 
+source "drivers/gpu/drm/rcar-du/Kconfig"
+
 source "drivers/gpu/drm/shmobile/Kconfig"
 
 source "drivers/gpu/drm/omapdrm/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 1c9f243..801bcaf 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -12,7 +12,8 @@ drm-y       :=	drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \
 		drm_platform.o drm_sysfs.o drm_hashtab.o drm_mm.o \
 		drm_crtc.o drm_modes.o drm_edid.o \
 		drm_info.o drm_debugfs.o drm_encoder_slave.o \
-		drm_trace_points.o drm_global.o drm_prime.o
+		drm_trace_points.o drm_global.o drm_prime.o \
+		drm_rect.o
 
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
 drm-$(CONFIG_DRM_GEM_CMA_HELPER) += drm_gem_cma_helper.o
@@ -48,6 +49,7 @@ obj-$(CONFIG_DRM_EXYNOS) +=exynos/
 obj-$(CONFIG_DRM_GMA500) += gma500/
 obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
+obj-$(CONFIG_DRM_RCAR_DU) += rcar-du/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
 obj-$(CONFIG_DRM_OMAP)	+= omapdrm/
 obj-$(CONFIG_DRM_TILCDC)	+= tilcdc/
diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index 02e52d5..622d4ae 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -348,8 +348,24 @@ int ast_gem_create(struct drm_device *dev,
 int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr);
 int ast_bo_unpin(struct ast_bo *bo);
 
-int ast_bo_reserve(struct ast_bo *bo, bool no_wait);
-void ast_bo_unreserve(struct ast_bo *bo);
+static inline int ast_bo_reserve(struct ast_bo *bo, bool no_wait)
+{
+	int ret;
+
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	if (ret) {
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p\n", bo);
+		return ret;
+	}
+	return 0;
+}
+
+static inline void ast_bo_unreserve(struct ast_bo *bo)
+{
+	ttm_bo_unreserve(&bo->bo);
+}
+
 void ast_ttm_placement(struct ast_bo *bo, int domain);
 int ast_bo_push_sysram(struct ast_bo *bo);
 int ast_mmap(struct file *filp, struct vm_area_struct *vma);
diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index fbc0823..7b33e14 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -51,7 +51,7 @@ static void ast_dirty_update(struct ast_fbdev *afbdev,
 	struct ast_bo *bo;
 	int src_offset, dst_offset;
 	int bpp = (afbdev->afb.base.bits_per_pixel + 7)/8;
-	int ret;
+	int ret = -EBUSY;
 	bool unmap = false;
 	bool store_for_later = false;
 	int x2, y2;
@@ -65,7 +65,8 @@ static void ast_dirty_update(struct ast_fbdev *afbdev,
 	 * then the BO is being moved and we should
 	 * store up the damage until later.
 	 */
-	ret = ast_bo_reserve(bo, true);
+	if (!in_interrupt())
+		ret = ast_bo_reserve(bo, true);
 	if (ret) {
 		if (ret != -EBUSY)
 			return;
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index 09da339..98d6708 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -271,26 +271,19 @@ int ast_mm_init(struct ast_private *ast)
 		return ret;
 	}
 
-	ast->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 0),
-				    pci_resource_len(dev->pdev, 0),
-				    DRM_MTRR_WC);
+	ast->fb_mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 0),
+					pci_resource_len(dev->pdev, 0));
 
 	return 0;
 }
 
 void ast_mm_fini(struct ast_private *ast)
 {
-	struct drm_device *dev = ast->dev;
 	ttm_bo_device_release(&ast->ttm.bdev);
 
 	ast_ttm_global_release(ast);
 
-	if (ast->fb_mtrr >= 0) {
-		drm_mtrr_del(ast->fb_mtrr,
-			     pci_resource_start(dev->pdev, 0),
-			     pci_resource_len(dev->pdev, 0), DRM_MTRR_WC);
-		ast->fb_mtrr = -1;
-	}
+	arch_phys_wc_del(ast->fb_mtrr);
 }
 
 void ast_ttm_placement(struct ast_bo *bo, int domain)
@@ -310,24 +303,6 @@ void ast_ttm_placement(struct ast_bo *bo, int domain)
 	bo->placement.num_busy_placement = c;
 }
 
-int ast_bo_reserve(struct ast_bo *bo, bool no_wait)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
-	if (ret) {
-		if (ret != -ERESTARTSYS && ret != -EBUSY)
-			DRM_ERROR("reserve failed %p\n", bo);
-		return ret;
-	}
-	return 0;
-}
-
-void ast_bo_unreserve(struct ast_bo *bo)
-{
-	ttm_bo_unreserve(&bo->bo);
-}
-
 int ast_bo_create(struct drm_device *dev, int size, int align,
 		  uint32_t flags, struct ast_bo **pastbo)
 {
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.h b/drivers/gpu/drm/cirrus/cirrus_drv.h
index 7ca0595..bae5560 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.h
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.h
@@ -240,8 +240,25 @@ void cirrus_ttm_placement(struct cirrus_bo *bo, int domain);
 int cirrus_bo_create(struct drm_device *dev, int size, int align,
 		     uint32_t flags, struct cirrus_bo **pcirrusbo);
 int cirrus_mmap(struct file *filp, struct vm_area_struct *vma);
-int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait);
-void cirrus_bo_unreserve(struct cirrus_bo *bo);
+
+static inline int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait)
+{
+	int ret;
+
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	if (ret) {
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p\n", bo);
+		return ret;
+	}
+	return 0;
+}
+
+static inline void cirrus_bo_unreserve(struct cirrus_bo *bo)
+{
+	ttm_bo_unreserve(&bo->bo);
+}
+
 int cirrus_bo_push_sysram(struct cirrus_bo *bo);
 int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr);
 #endif				/* __CIRRUS_DRV_H__ */
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index 3541b56..b27e956 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -25,7 +25,7 @@ static void cirrus_dirty_update(struct cirrus_fbdev *afbdev,
 	struct cirrus_bo *bo;
 	int src_offset, dst_offset;
 	int bpp = (afbdev->gfb.base.bits_per_pixel + 7)/8;
-	int ret;
+	int ret = -EBUSY;
 	bool unmap = false;
 	bool store_for_later = false;
 	int x2, y2;
@@ -39,7 +39,8 @@ static void cirrus_dirty_update(struct cirrus_fbdev *afbdev,
 	 * then the BO is being moved and we should
 	 * store up the damage until later.
 	 */
-	ret = cirrus_bo_reserve(bo, true);
+	if (!in_interrupt())
+		ret = cirrus_bo_reserve(bo, true);
 	if (ret) {
 		if (ret != -EBUSY)
 			return;
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 2ed8cfc..0047012 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -271,9 +271,8 @@ int cirrus_mm_init(struct cirrus_device *cirrus)
 		return ret;
 	}
 
-	cirrus->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 0),
-				    pci_resource_len(dev->pdev, 0),
-				    DRM_MTRR_WC);
+	cirrus->fb_mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 0),
+					   pci_resource_len(dev->pdev, 0));
 
 	cirrus->mm_inited = true;
 	return 0;
@@ -281,8 +280,6 @@ int cirrus_mm_init(struct cirrus_device *cirrus)
 
 void cirrus_mm_fini(struct cirrus_device *cirrus)
 {
-	struct drm_device *dev = cirrus->dev;
-
 	if (!cirrus->mm_inited)
 		return;
 
@@ -290,12 +287,8 @@ void cirrus_mm_fini(struct cirrus_device *cirrus)
 
 	cirrus_ttm_global_release(cirrus);
 
-	if (cirrus->fb_mtrr >= 0) {
-		drm_mtrr_del(cirrus->fb_mtrr,
-			     pci_resource_start(dev->pdev, 0),
-			     pci_resource_len(dev->pdev, 0), DRM_MTRR_WC);
-		cirrus->fb_mtrr = -1;
-	}
+	arch_phys_wc_del(cirrus->fb_mtrr);
+	cirrus->fb_mtrr = 0;
 }
 
 void cirrus_ttm_placement(struct cirrus_bo *bo, int domain)
@@ -315,24 +308,6 @@ void cirrus_ttm_placement(struct cirrus_bo *bo, int domain)
 	bo->placement.num_busy_placement = c;
 }
 
-int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
-	if (ret) {
-		if (ret != -ERESTARTSYS && ret != -EBUSY)
-			DRM_ERROR("reserve failed %p\n", bo);
-		return ret;
-	}
-	return 0;
-}
-
-void cirrus_bo_unreserve(struct cirrus_bo *bo)
-{
-	ttm_bo_unreserve(&bo->bo);
-}
-
 int cirrus_bo_create(struct drm_device *dev, int size, int align,
 		  uint32_t flags, struct cirrus_bo **pcirrusbo)
 {
diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c
index 0128147..5a4dbb4 100644
--- a/drivers/gpu/drm/drm_bufs.c
+++ b/drivers/gpu/drm/drm_bufs.c
@@ -210,12 +210,16 @@ static int drm_addmap_core(struct drm_device * dev, resource_size_t offset,
 		if (drm_core_has_MTRR(dev)) {
 			if (map->type == _DRM_FRAME_BUFFER ||
 			    (map->flags & _DRM_WRITE_COMBINING)) {
-				map->mtrr = mtrr_add(map->offset, map->size,
-						     MTRR_TYPE_WRCOMB, 1);
+				map->mtrr =
+					arch_phys_wc_add(map->offset, map->size);
 			}
 		}
 		if (map->type == _DRM_REGISTERS) {
-			map->handle = ioremap(map->offset, map->size);
+			if (map->flags & _DRM_WRITE_COMBINING)
+				map->handle = ioremap_wc(map->offset,
+							 map->size);
+			else
+				map->handle = ioremap(map->offset, map->size);
 			if (!map->handle) {
 				kfree(map);
 				return -ENOMEM;
@@ -410,6 +414,15 @@ int drm_addmap_ioctl(struct drm_device *dev, void *data,
 
 	/* avoid a warning on 64-bit, this casting isn't very nice, but the API is set so too late */
 	map->handle = (void *)(unsigned long)maplist->user_token;
+
+	/*
+	 * It appears that there are no users of this value whatsoever --
+	 * drmAddMap just discards it.  Let's not encourage its use.
+	 * (Keeping drm_addmap_core's returned mtrr value would be wrong --
+	 *  it's not a real mtrr index anymore.)
+	 */
+	map->mtrr = -1;
+
 	return 0;
 }
 
@@ -451,11 +464,8 @@ int drm_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
 		iounmap(map->handle);
 		/* FALLTHROUGH */
 	case _DRM_FRAME_BUFFER:
-		if (drm_core_has_MTRR(dev) && map->mtrr >= 0) {
-			int retcode;
-			retcode = mtrr_del(map->mtrr, map->offset, map->size);
-			DRM_DEBUG("mtrr_del=%d\n", retcode);
-		}
+		if (drm_core_has_MTRR(dev))
+			arch_phys_wc_del(map->mtrr);
 		break;
 	case _DRM_SHM:
 		vfree(map->handle);
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index e7e9242..fc83bb9 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -29,6 +29,7 @@
  *      Dave Airlie <airlied@linux.ie>
  *      Jesse Barnes <jesse.barnes@intel.com>
  */
+#include <linux/ctype.h>
 #include <linux/list.h>
 #include <linux/slab.h>
 #include <linux/export.h>
@@ -91,7 +92,7 @@ EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
 
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
-	char *fnname(int val)					\
+	const char *fnname(int val)				\
 	{							\
 		int i;						\
 		for (i = 0; i < ARRAY_SIZE(list); i++) {	\
@@ -104,7 +105,7 @@ EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
 /*
  * Global properties
  */
-static struct drm_prop_enum_list drm_dpms_enum_list[] =
+static const struct drm_prop_enum_list drm_dpms_enum_list[] =
 {	{ DRM_MODE_DPMS_ON, "On" },
 	{ DRM_MODE_DPMS_STANDBY, "Standby" },
 	{ DRM_MODE_DPMS_SUSPEND, "Suspend" },
@@ -116,7 +117,7 @@ DRM_ENUM_NAME_FN(drm_get_dpms_name, drm_dpms_enum_list)
 /*
  * Optional properties
  */
-static struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
+static const struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
 {
 	{ DRM_MODE_SCALE_NONE, "None" },
 	{ DRM_MODE_SCALE_FULLSCREEN, "Full" },
@@ -124,7 +125,7 @@ static struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
 	{ DRM_MODE_SCALE_ASPECT, "Full aspect" },
 };
 
-static struct drm_prop_enum_list drm_dithering_mode_enum_list[] =
+static const struct drm_prop_enum_list drm_dithering_mode_enum_list[] =
 {
 	{ DRM_MODE_DITHERING_OFF, "Off" },
 	{ DRM_MODE_DITHERING_ON, "On" },
@@ -134,7 +135,7 @@ static struct drm_prop_enum_list drm_dithering_mode_enum_list[] =
 /*
  * Non-global properties, but "required" for certain connectors.
  */
-static struct drm_prop_enum_list drm_dvi_i_select_enum_list[] =
+static const struct drm_prop_enum_list drm_dvi_i_select_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Automatic, "Automatic" }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_DVID,      "DVI-D"     }, /* DVI-I  */
@@ -143,7 +144,7 @@ static struct drm_prop_enum_list drm_dvi_i_select_enum_list[] =
 
 DRM_ENUM_NAME_FN(drm_get_dvi_i_select_name, drm_dvi_i_select_enum_list)
 
-static struct drm_prop_enum_list drm_dvi_i_subconnector_enum_list[] =
+static const struct drm_prop_enum_list drm_dvi_i_subconnector_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Unknown,   "Unknown"   }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_DVID,      "DVI-D"     }, /* DVI-I  */
@@ -153,7 +154,7 @@ static struct drm_prop_enum_list drm_dvi_i_subconnector_enum_list[] =
 DRM_ENUM_NAME_FN(drm_get_dvi_i_subconnector_name,
 		 drm_dvi_i_subconnector_enum_list)
 
-static struct drm_prop_enum_list drm_tv_select_enum_list[] =
+static const struct drm_prop_enum_list drm_tv_select_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Automatic, "Automatic" }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_Composite, "Composite" }, /* TV-out */
@@ -164,7 +165,7 @@ static struct drm_prop_enum_list drm_tv_select_enum_list[] =
 
 DRM_ENUM_NAME_FN(drm_get_tv_select_name, drm_tv_select_enum_list)
 
-static struct drm_prop_enum_list drm_tv_subconnector_enum_list[] =
+static const struct drm_prop_enum_list drm_tv_subconnector_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Unknown,   "Unknown"   }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_Composite, "Composite" }, /* TV-out */
@@ -176,7 +177,7 @@ static struct drm_prop_enum_list drm_tv_subconnector_enum_list[] =
 DRM_ENUM_NAME_FN(drm_get_tv_subconnector_name,
 		 drm_tv_subconnector_enum_list)
 
-static struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
+static const struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
 	{ DRM_MODE_DIRTY_OFF,      "Off"      },
 	{ DRM_MODE_DIRTY_ON,       "On"       },
 	{ DRM_MODE_DIRTY_ANNOTATE, "Annotate" },
@@ -184,7 +185,7 @@ static struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
 
 struct drm_conn_prop_enum_list {
 	int type;
-	char *name;
+	const char *name;
 	int count;
 };
 
@@ -210,7 +211,7 @@ static struct drm_conn_prop_enum_list drm_connector_enum_list[] =
 	{ DRM_MODE_CONNECTOR_VIRTUAL, "Virtual", 0},
 };
 
-static struct drm_prop_enum_list drm_encoder_enum_list[] =
+static const struct drm_prop_enum_list drm_encoder_enum_list[] =
 {	{ DRM_MODE_ENCODER_NONE, "None" },
 	{ DRM_MODE_ENCODER_DAC, "DAC" },
 	{ DRM_MODE_ENCODER_TMDS, "TMDS" },
@@ -219,7 +220,7 @@ static struct drm_prop_enum_list drm_encoder_enum_list[] =
 	{ DRM_MODE_ENCODER_VIRTUAL, "Virtual" },
 };
 
-char *drm_get_encoder_name(struct drm_encoder *encoder)
+const char *drm_get_encoder_name(const struct drm_encoder *encoder)
 {
 	static char buf[32];
 
@@ -230,7 +231,7 @@ char *drm_get_encoder_name(struct drm_encoder *encoder)
 }
 EXPORT_SYMBOL(drm_get_encoder_name);
 
-char *drm_get_connector_name(struct drm_connector *connector)
+const char *drm_get_connector_name(const struct drm_connector *connector)
 {
 	static char buf[32];
 
@@ -241,7 +242,7 @@ char *drm_get_connector_name(struct drm_connector *connector)
 }
 EXPORT_SYMBOL(drm_get_connector_name);
 
-char *drm_get_connector_status_name(enum drm_connector_status status)
+const char *drm_get_connector_status_name(enum drm_connector_status status)
 {
 	if (status == connector_status_connected)
 		return "connected";
@@ -252,6 +253,28 @@ char *drm_get_connector_status_name(enum drm_connector_status status)
 }
 EXPORT_SYMBOL(drm_get_connector_status_name);
 
+static char printable_char(int c)
+{
+	return isascii(c) && isprint(c) ? c : '?';
+}
+
+const char *drm_get_format_name(uint32_t format)
+{
+	static char buf[32];
+
+	snprintf(buf, sizeof(buf),
+		 "%c%c%c%c %s-endian (0x%08x)",
+		 printable_char(format & 0xff),
+		 printable_char((format >> 8) & 0xff),
+		 printable_char((format >> 16) & 0xff),
+		 printable_char((format >> 24) & 0x7f),
+		 format & DRM_FORMAT_BIG_ENDIAN ? "big" : "little",
+		 format);
+
+	return buf;
+}
+EXPORT_SYMBOL(drm_get_format_name);
+
 /**
  * drm_mode_object_get - allocate a new modeset identifier
  * @dev: DRM device
@@ -569,16 +592,8 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 		}
 
 		list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
-			if (plane->fb == fb) {
-				/* should turn off the crtc */
-				ret = plane->funcs->disable_plane(plane);
-				if (ret)
-					DRM_ERROR("failed to disable plane with busy fb\n");
-				/* disconnect the plane from the fb and crtc: */
-				__drm_framebuffer_unreference(plane->fb);
-				plane->fb = NULL;
-				plane->crtc = NULL;
-			}
+			if (plane->fb == fb)
+				drm_plane_force_disable(plane);
 		}
 		drm_modeset_unlock_all(dev);
 	}
@@ -593,7 +608,7 @@ EXPORT_SYMBOL(drm_framebuffer_remove);
  * @crtc: CRTC object to init
  * @funcs: callbacks for the new CRTC
  *
- * Inits a new object created as base part of an driver crtc object.
+ * Inits a new object created as base part of a driver crtc object.
  *
  * RETURNS:
  * Zero on success, error code on failure.
@@ -628,11 +643,12 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 EXPORT_SYMBOL(drm_crtc_init);
 
 /**
- * drm_crtc_cleanup - Cleans up the core crtc usage.
+ * drm_crtc_cleanup - Clean up the core crtc usage
  * @crtc: CRTC to cleanup
  *
- * Cleanup @crtc. Removes from drm modesetting space
- * does NOT free object, caller does that.
+ * This function cleans up @crtc and removes it from the DRM mode setting
+ * core. Note that the function does *not* free the crtc structure itself,
+ * this is the responsibility of the caller.
  */
 void drm_crtc_cleanup(struct drm_crtc *crtc)
 {
@@ -657,7 +673,7 @@ EXPORT_SYMBOL(drm_crtc_cleanup);
 void drm_mode_probed_add(struct drm_connector *connector,
 			 struct drm_display_mode *mode)
 {
-	list_add(&mode->head, &connector->probed_modes);
+	list_add_tail(&mode->head, &connector->probed_modes);
 }
 EXPORT_SYMBOL(drm_mode_probed_add);
 
@@ -803,6 +819,21 @@ void drm_encoder_cleanup(struct drm_encoder *encoder)
 }
 EXPORT_SYMBOL(drm_encoder_cleanup);
 
+/**
+ * drm_plane_init - Initialise a new plane object
+ * @dev: DRM device
+ * @plane: plane object to init
+ * @possible_crtcs: bitmask of possible CRTCs
+ * @funcs: callbacks for the new plane
+ * @formats: array of supported formats (%DRM_FORMAT_*)
+ * @format_count: number of elements in @formats
+ * @priv: plane is private (hidden from userspace)?
+ *
+ * Inits a new object created as base part of a driver plane object.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure.
+ */
 int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 		   unsigned long possible_crtcs,
 		   const struct drm_plane_funcs *funcs,
@@ -851,6 +882,14 @@ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 }
 EXPORT_SYMBOL(drm_plane_init);
 
+/**
+ * drm_plane_cleanup - Clean up the core plane usage
+ * @plane: plane to cleanup
+ *
+ * This function cleans up @plane and removes it from the DRM mode setting
+ * core. Note that the function does *not* free the plane structure itself,
+ * this is the responsibility of the caller.
+ */
 void drm_plane_cleanup(struct drm_plane *plane)
 {
 	struct drm_device *dev = plane->dev;
@@ -868,6 +907,32 @@ void drm_plane_cleanup(struct drm_plane *plane)
 EXPORT_SYMBOL(drm_plane_cleanup);
 
 /**
+ * drm_plane_force_disable - Forcibly disable a plane
+ * @plane: plane to disable
+ *
+ * Forces the plane to be disabled.
+ *
+ * Used when the plane's current framebuffer is destroyed,
+ * and when restoring fbdev mode.
+ */
+void drm_plane_force_disable(struct drm_plane *plane)
+{
+	int ret;
+
+	if (!plane->fb)
+		return;
+
+	ret = plane->funcs->disable_plane(plane);
+	if (ret)
+		DRM_ERROR("failed to disable plane with busy fb\n");
+	/* disconnect the plane from the fb and crtc: */
+	__drm_framebuffer_unreference(plane->fb);
+	plane->fb = NULL;
+	plane->crtc = NULL;
+}
+EXPORT_SYMBOL(drm_plane_force_disable);
+
+/**
  * drm_mode_create - create a new display mode
  * @dev: DRM device
  *
@@ -1740,7 +1805,7 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 
 	plane_resp->plane_id = plane->base.id;
 	plane_resp->possible_crtcs = plane->possible_crtcs;
-	plane_resp->gamma_size = plane->gamma_size;
+	plane_resp->gamma_size = 0;
 
 	/*
 	 * This ioctl is called twice, once to determine how much space is
@@ -1834,7 +1899,8 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 		if (fb->pixel_format == plane->format_types[i])
 			break;
 	if (i == plane->format_count) {
-		DRM_DEBUG_KMS("Invalid pixel format 0x%08x\n", fb->pixel_format);
+		DRM_DEBUG_KMS("Invalid pixel format %s\n",
+			      drm_get_format_name(fb->pixel_format));
 		ret = -EINVAL;
 		goto out;
 	}
@@ -1906,18 +1972,31 @@ out:
 int drm_mode_set_config_internal(struct drm_mode_set *set)
 {
 	struct drm_crtc *crtc = set->crtc;
-	struct drm_framebuffer *fb, *old_fb;
+	struct drm_framebuffer *fb;
+	struct drm_crtc *tmp;
 	int ret;
 
-	old_fb = crtc->fb;
+	/*
+	 * NOTE: ->set_config can also disable other crtcs (if we steal all
+	 * connectors from it), hence we need to refcount the fbs across all
+	 * crtcs. Atomic modeset will have saner semantics ...
+	 */
+	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head)
+		tmp->old_fb = tmp->fb;
+
 	fb = set->fb;
 
 	ret = crtc->funcs->set_config(set);
 	if (ret == 0) {
-		if (old_fb)
-			drm_framebuffer_unreference(old_fb);
-		if (fb)
-			drm_framebuffer_reference(fb);
+		/* crtc->fb must be updated by ->set_config, enforces this. */
+		WARN_ON(fb != crtc->fb);
+	}
+
+	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head) {
+		if (tmp->fb)
+			drm_framebuffer_reference(tmp->fb);
+		if (tmp->old_fb)
+			drm_framebuffer_unreference(tmp->old_fb);
 	}
 
 	return ret;
@@ -2099,10 +2178,10 @@ out:
 	return ret;
 }
 
-int drm_mode_cursor_ioctl(struct drm_device *dev,
-			void *data, struct drm_file *file_priv)
+static int drm_mode_cursor_common(struct drm_device *dev,
+				  struct drm_mode_cursor2 *req,
+				  struct drm_file *file_priv)
 {
-	struct drm_mode_cursor *req = data;
 	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	int ret = 0;
@@ -2122,13 +2201,17 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 
 	mutex_lock(&crtc->mutex);
 	if (req->flags & DRM_MODE_CURSOR_BO) {
-		if (!crtc->funcs->cursor_set) {
+		if (!crtc->funcs->cursor_set && !crtc->funcs->cursor_set2) {
 			ret = -ENXIO;
 			goto out;
 		}
 		/* Turns off the cursor if handle is 0 */
-		ret = crtc->funcs->cursor_set(crtc, file_priv, req->handle,
-					      req->width, req->height);
+		if (crtc->funcs->cursor_set2)
+			ret = crtc->funcs->cursor_set2(crtc, file_priv, req->handle,
+						      req->width, req->height, req->hot_x, req->hot_y);
+		else
+			ret = crtc->funcs->cursor_set(crtc, file_priv, req->handle,
+						      req->width, req->height);
 	}
 
 	if (req->flags & DRM_MODE_CURSOR_MOVE) {
@@ -2143,6 +2226,25 @@ out:
 	mutex_unlock(&crtc->mutex);
 
 	return ret;
+
+}
+int drm_mode_cursor_ioctl(struct drm_device *dev,
+			void *data, struct drm_file *file_priv)
+{
+	struct drm_mode_cursor *req = data;
+	struct drm_mode_cursor2 new_req;
+
+	memcpy(&new_req, req, sizeof(struct drm_mode_cursor));
+	new_req.hot_x = new_req.hot_y = 0;
+
+	return drm_mode_cursor_common(dev, &new_req, file_priv);
+}
+
+int drm_mode_cursor2_ioctl(struct drm_device *dev,
+			   void *data, struct drm_file *file_priv)
+{
+	struct drm_mode_cursor2 *req = data;
+	return drm_mode_cursor_common(dev, req, file_priv);
 }
 
 /* Original addfb only supported RGB formats, so figure out which one */
@@ -2312,7 +2414,8 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r)
 
 	ret = format_check(r);
 	if (ret) {
-		DRM_DEBUG_KMS("bad framebuffer format 0x%08x\n", r->pixel_format);
+		DRM_DEBUG_KMS("bad framebuffer format %s\n",
+			      drm_get_format_name(r->pixel_format));
 		return ret;
 	}
 
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index ed1334e..738a429 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -189,13 +189,14 @@ prune:
 	if (list_empty(&connector->modes))
 		return 0;
 
+	list_for_each_entry(mode, &connector->modes, head)
+		mode->vrefresh = drm_mode_vrefresh(mode);
+
 	drm_mode_sort(&connector->modes);
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] probed modes :\n", connector->base.id,
 			drm_get_connector_name(connector));
 	list_for_each_entry(mode, &connector->modes, head) {
-		mode->vrefresh = drm_mode_vrefresh(mode);
-
 		drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
 		drm_mode_debug_printmodeline(mode);
 	}
@@ -564,14 +565,13 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 
 	DRM_DEBUG_KMS("\n");
 
-	if (!set)
-		return -EINVAL;
-
-	if (!set->crtc)
-		return -EINVAL;
+	BUG_ON(!set);
+	BUG_ON(!set->crtc);
+	BUG_ON(!set->crtc->helper_private);
 
-	if (!set->crtc->helper_private)
-		return -EINVAL;
+	/* Enforce sane interface api - has been abused by the fb helper. */
+	BUG_ON(!set->mode && set->fb);
+	BUG_ON(set->fb && set->num_connectors == 0);
 
 	crtc_funcs = set->crtc->helper_private;
 
@@ -645,11 +645,6 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 			mode_changed = true;
 		} else if (set->fb == NULL) {
 			mode_changed = true;
-		} else if (set->fb->depth != set->crtc->fb->depth) {
-			mode_changed = true;
-		} else if (set->fb->bits_per_pixel !=
-			   set->crtc->fb->bits_per_pixel) {
-			mode_changed = true;
 		} else if (set->fb->pixel_format !=
 			   set->crtc->fb->pixel_format) {
 			mode_changed = true;
@@ -759,12 +754,6 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 				ret = -EINVAL;
 				goto fail;
 			}
-			DRM_DEBUG_KMS("Setting connector DPMS state to on\n");
-			for (i = 0; i < set->num_connectors; i++) {
-				DRM_DEBUG_KMS("\t[CONNECTOR:%d:%s] set DPMS on\n", set->connectors[i]->base.id,
-					      drm_get_connector_name(set->connectors[i]));
-				set->connectors[i]->funcs->dpms(set->connectors[i], DRM_MODE_DPMS_ON);
-			}
 		}
 		drm_helper_disable_unused_functions(dev);
 	} else if (fb_changed) {
@@ -782,6 +771,22 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 		}
 	}
 
+	/*
+	 * crtc set_config helpers implicit set the crtc and all connected
+	 * encoders to DPMS on for a full mode set. But for just an fb update it
+	 * doesn't do that. To not confuse userspace, do an explicit DPMS_ON
+	 * unconditionally. This will also ensure driver internal dpms state is
+	 * consistent again.
+	 */
+	if (set->crtc->enabled) {
+		DRM_DEBUG_KMS("Setting connector DPMS state to on\n");
+		for (i = 0; i < set->num_connectors; i++) {
+			DRM_DEBUG_KMS("\t[CONNECTOR:%d:%s] set DPMS on\n", set->connectors[i]->base.id,
+				      drm_get_connector_name(set->connectors[i]));
+			set->connectors[i]->funcs->dpms(set->connectors[i], DRM_MODE_DPMS_ON);
+		}
+	}
+
 	kfree(save_connectors);
 	kfree(save_encoders);
 	kfree(save_crtcs);
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 9cc247f..99fcd7c 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -166,6 +166,7 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 };
 
 #define DRM_CORE_IOCTL_COUNT	ARRAY_SIZE( drm_ioctls )
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 9e62bbe..95d6f4b 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -968,6 +968,9 @@ bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 	u8 csum = 0;
 	struct edid *edid = (struct edid *)raw_edid;
 
+	if (WARN_ON(!raw_edid))
+		return false;
+
 	if (edid_fixup > 8 || edid_fixup < 0)
 		edid_fixup = 6;
 
@@ -1010,15 +1013,15 @@ bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 		break;
 	}
 
-	return 1;
+	return true;
 
 bad:
-	if (raw_edid && print_bad_edid) {
+	if (print_bad_edid) {
 		printk(KERN_ERR "Raw EDID:\n");
 		print_hex_dump(KERN_ERR, " \t", DUMP_PREFIX_NONE, 16, 1,
 			       raw_edid, EDID_LENGTH, false);
 	}
-	return 0;
+	return false;
 }
 EXPORT_SYMBOL(drm_edid_block_valid);
 
@@ -1706,11 +1709,11 @@ static struct drm_display_mode *drm_mode_detailed(struct drm_device *dev,
 		return NULL;
 
 	if (pt->misc & DRM_EDID_PT_STEREO) {
-		printk(KERN_WARNING "stereo mode not supported\n");
+		DRM_DEBUG_KMS("stereo mode not supported\n");
 		return NULL;
 	}
 	if (!(pt->misc & DRM_EDID_PT_SEPARATE_SYNC)) {
-		printk(KERN_WARNING "composite sync not supported\n");
+		DRM_DEBUG_KMS("composite sync not supported\n");
 	}
 
 	/* it is incorrect if hsync/vsync width is zero */
@@ -2321,6 +2324,31 @@ u8 *drm_find_cea_extension(struct edid *edid)
 }
 EXPORT_SYMBOL(drm_find_cea_extension);
 
+/*
+ * Calculate the alternate clock for the CEA mode
+ * (60Hz vs. 59.94Hz etc.)
+ */
+static unsigned int
+cea_mode_alternate_clock(const struct drm_display_mode *cea_mode)
+{
+	unsigned int clock = cea_mode->clock;
+
+	if (cea_mode->vrefresh % 6 != 0)
+		return clock;
+
+	/*
+	 * edid_cea_modes contains the 59.94Hz
+	 * variant for 240 and 480 line modes,
+	 * and the 60Hz variant otherwise.
+	 */
+	if (cea_mode->vdisplay == 240 || cea_mode->vdisplay == 480)
+		clock = clock * 1001 / 1000;
+	else
+		clock = DIV_ROUND_UP(clock * 1000, 1001);
+
+	return clock;
+}
+
 /**
  * drm_match_cea_mode - look for a CEA mode matching given mode
  * @to_match: display mode
@@ -2339,21 +2367,9 @@ u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 		const struct drm_display_mode *cea_mode = &edid_cea_modes[mode];
 		unsigned int clock1, clock2;
 
-		clock1 = clock2 = cea_mode->clock;
-
 		/* Check both 60Hz and 59.94Hz */
-		if (cea_mode->vrefresh % 6 == 0) {
-			/*
-			 * edid_cea_modes contains the 59.94Hz
-			 * variant for 240 and 480 line modes,
-			 * and the 60Hz variant otherwise.
-			 */
-			if (cea_mode->vdisplay == 240 ||
-			    cea_mode->vdisplay == 480)
-				clock1 = clock1 * 1001 / 1000;
-			else
-				clock2 = DIV_ROUND_UP(clock2 * 1000, 1001);
-		}
+		clock1 = cea_mode->clock;
+		clock2 = cea_mode_alternate_clock(cea_mode);
 
 		if ((KHZ2PICOS(to_match->clock) == KHZ2PICOS(clock1) ||
 		     KHZ2PICOS(to_match->clock) == KHZ2PICOS(clock2)) &&
@@ -2364,6 +2380,66 @@ u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 }
 EXPORT_SYMBOL(drm_match_cea_mode);
 
+static int
+add_alternate_cea_modes(struct drm_connector *connector, struct edid *edid)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_display_mode *mode, *tmp;
+	LIST_HEAD(list);
+	int modes = 0;
+
+	/* Don't add CEA modes if the CEA extension block is missing */
+	if (!drm_find_cea_extension(edid))
+		return 0;
+
+	/*
+	 * Go through all probed modes and create a new mode
+	 * with the alternate clock for certain CEA modes.
+	 */
+	list_for_each_entry(mode, &connector->probed_modes, head) {
+		const struct drm_display_mode *cea_mode;
+		struct drm_display_mode *newmode;
+		u8 cea_mode_idx = drm_match_cea_mode(mode) - 1;
+		unsigned int clock1, clock2;
+
+		if (cea_mode_idx >= ARRAY_SIZE(edid_cea_modes))
+			continue;
+
+		cea_mode = &edid_cea_modes[cea_mode_idx];
+
+		clock1 = cea_mode->clock;
+		clock2 = cea_mode_alternate_clock(cea_mode);
+
+		if (clock1 == clock2)
+			continue;
+
+		if (mode->clock != clock1 && mode->clock != clock2)
+			continue;
+
+		newmode = drm_mode_duplicate(dev, cea_mode);
+		if (!newmode)
+			continue;
+
+		/*
+		 * The current mode could be either variant. Make
+		 * sure to pick the "other" clock for the new mode.
+		 */
+		if (mode->clock != clock1)
+			newmode->clock = clock1;
+		else
+			newmode->clock = clock2;
+
+		list_add_tail(&newmode->head, &list);
+	}
+
+	list_for_each_entry_safe(mode, tmp, &list, head) {
+		list_del(&mode->head);
+		drm_mode_probed_add(connector, mode);
+		modes++;
+	}
+
+	return modes;
+}
 
 static int
 do_cea_modes (struct drm_connector *connector, u8 *db, u8 len)
@@ -2946,6 +3022,7 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)
 	if (edid->features & DRM_EDID_FEATURE_DEFAULT_GTF)
 		num_modes += add_inferred_modes(connector, edid);
 	num_modes += add_cea_modes(connector, edid);
+	num_modes += add_alternate_cea_modes(connector, edid);
 
 	if (quirks & (EDID_QUIRK_PREFER_LARGE_60 | EDID_QUIRK_PREFER_LARGE_75))
 		edid_fixup_preferred(connector, quirks);
diff --git a/drivers/gpu/drm/drm_edid_load.c b/drivers/gpu/drm/drm_edid_load.c
index fa445dd..a4f5ce1 100644
--- a/drivers/gpu/drm/drm_edid_load.c
+++ b/drivers/gpu/drm/drm_edid_load.c
@@ -186,12 +186,11 @@ static u8 *edid_load(struct drm_connector *connector, char *name,
 		goto relfw_out;
 	}
 
-	edid = kmalloc(fwsize, GFP_KERNEL);
+	edid = kmemdup(fwdata, fwsize, GFP_KERNEL);
 	if (edid == NULL) {
 		err = -ENOMEM;
 		goto relfw_out;
 	}
-	memcpy(edid, fwdata, fwsize);
 
 	if (!drm_edid_block_valid(edid, 0, print_bad_edid)) {
 		connector->bad_edid_counter++;
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index b78cbe7..3d13ca6e2 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -168,6 +168,9 @@ static void drm_fb_helper_save_lut_atomic(struct drm_crtc *crtc, struct drm_fb_h
 	uint16_t *r_base, *g_base, *b_base;
 	int i;
 
+	if (helper->funcs->gamma_get == NULL)
+		return;
+
 	r_base = crtc->gamma_store;
 	g_base = r_base + crtc->gamma_size;
 	b_base = g_base + crtc->gamma_size;
@@ -284,13 +287,27 @@ EXPORT_SYMBOL(drm_fb_helper_debug_leave);
  */
 bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 {
+	struct drm_device *dev = fb_helper->dev;
+	struct drm_plane *plane;
 	bool error = false;
-	int i, ret;
+	int i;
+
+	drm_warn_on_modeset_not_all_locked(dev);
 
-	drm_warn_on_modeset_not_all_locked(fb_helper->dev);
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+		drm_plane_force_disable(plane);
 
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_mode_set *mode_set = &fb_helper->crtc_info[i].mode_set;
+		struct drm_crtc *crtc = mode_set->crtc;
+		int ret;
+
+		if (crtc->funcs->cursor_set) {
+			ret = crtc->funcs->cursor_set(crtc, NULL, 0, 0, 0);
+			if (ret)
+				error = true;
+		}
+
 		ret = drm_mode_set_config_internal(mode_set);
 		if (ret)
 			error = true;
@@ -583,6 +600,14 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 		return 0;
 	}
 
+	/*
+	 * The driver really shouldn't advertise pseudo/directcolor
+	 * visuals if it can't deal with the palette.
+	 */
+	if (WARN_ON(!fb_helper->funcs->gamma_set ||
+		    !fb_helper->funcs->gamma_get))
+		return -EINVAL;
+
 	pindex = regno;
 
 	if (fb->bits_per_pixel == 16) {
@@ -626,12 +651,19 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 {
 	struct drm_fb_helper *fb_helper = info->par;
+	struct drm_device *dev = fb_helper->dev;
 	struct drm_crtc_helper_funcs *crtc_funcs;
 	u16 *red, *green, *blue, *transp;
 	struct drm_crtc *crtc;
 	int i, j, rc = 0;
 	int start;
 
+	drm_modeset_lock_all(dev);
+	if (!drm_fb_helper_is_bound(fb_helper)) {
+		drm_modeset_unlock_all(dev);
+		return -EBUSY;
+	}
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 		crtc_funcs = crtc->helper_private;
@@ -654,10 +686,13 @@ int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 
 			rc = setcolreg(crtc, hred, hgreen, hblue, start++, info);
 			if (rc)
-				return rc;
+				goto out;
 		}
-		crtc_funcs->load_lut(crtc);
+		if (crtc_funcs->load_lut)
+			crtc_funcs->load_lut(crtc);
 	}
+ out:
+	drm_modeset_unlock_all(dev);
 	return rc;
 }
 EXPORT_SYMBOL(drm_fb_helper_setcmap);
diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c
index 429e07d..3a24385 100644
--- a/drivers/gpu/drm/drm_fops.c
+++ b/drivers/gpu/drm/drm_fops.c
@@ -271,6 +271,11 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 	priv->uid = current_euid();
 	priv->pid = get_pid(task_pid(current));
 	priv->minor = idr_find(&drm_minors_idr, minor_id);
+	if (!priv->minor) {
+		ret = -ENODEV;
+		goto out_put_pid;
+	}
+
 	priv->ioctl_count = 0;
 	/* for compatibility root is always authenticated */
 	priv->authenticated = capable(CAP_SYS_ADMIN);
@@ -292,7 +297,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 	if (dev->driver->open) {
 		ret = dev->driver->open(dev, priv);
 		if (ret < 0)
-			goto out_free;
+			goto out_prime_destroy;
 	}
 
 
@@ -304,7 +309,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 		if (!priv->minor->master) {
 			mutex_unlock(&dev->struct_mutex);
 			ret = -ENOMEM;
-			goto out_free;
+			goto out_close;
 		}
 
 		priv->is_master = 1;
@@ -322,7 +327,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 				drm_master_put(&priv->minor->master);
 				drm_master_put(&priv->master);
 				mutex_unlock(&dev->struct_mutex);
-				goto out_free;
+				goto out_close;
 			}
 		}
 		mutex_lock(&dev->struct_mutex);
@@ -333,7 +338,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 				drm_master_put(&priv->minor->master);
 				drm_master_put(&priv->master);
 				mutex_unlock(&dev->struct_mutex);
-				goto out_free;
+				goto out_close;
 			}
 		}
 		mutex_unlock(&dev->struct_mutex);
@@ -367,7 +372,17 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 #endif
 
 	return 0;
-      out_free:
+
+out_close:
+	if (dev->driver->postclose)
+		dev->driver->postclose(dev, priv);
+out_prime_destroy:
+	if (drm_core_check_feature(dev, DRIVER_PRIME))
+		drm_prime_destroy_file_private(&priv->prime);
+	if (dev->driver->driver_features & DRIVER_GEM)
+		drm_gem_release(dev, priv);
+out_put_pid:
+	put_pid(priv->pid);
 	kfree(priv);
 	filp->private_data = NULL;
 	return ret;
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index cf919e3..603f256 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -108,12 +108,8 @@ drm_gem_init(struct drm_device *dev)
 		return -ENOMEM;
 	}
 
-	if (drm_mm_init(&mm->offset_manager, DRM_FILE_PAGE_OFFSET_START,
-			DRM_FILE_PAGE_OFFSET_SIZE)) {
-		drm_ht_remove(&mm->offset_hash);
-		kfree(mm);
-		return -ENOMEM;
-	}
+	drm_mm_init(&mm->offset_manager, DRM_FILE_PAGE_OFFSET_START,
+		    DRM_FILE_PAGE_OFFSET_SIZE);
 
 	return 0;
 }
@@ -453,25 +449,21 @@ drm_gem_flink_ioctl(struct drm_device *dev, void *data,
 	spin_lock(&dev->object_name_lock);
 	if (!obj->name) {
 		ret = idr_alloc(&dev->object_name_idr, obj, 1, 0, GFP_NOWAIT);
-		obj->name = ret;
-		args->name = (uint64_t) obj->name;
-		spin_unlock(&dev->object_name_lock);
-		idr_preload_end();
-
 		if (ret < 0)
 			goto err;
-		ret = 0;
+
+		obj->name = ret;
 
 		/* Allocate a reference for the name table.  */
 		drm_gem_object_reference(obj);
-	} else {
-		args->name = (uint64_t) obj->name;
-		spin_unlock(&dev->object_name_lock);
-		idr_preload_end();
-		ret = 0;
 	}
 
+	args->name = (uint64_t) obj->name;
+	ret = 0;
+
 err:
+	spin_unlock(&dev->object_name_lock);
+	idr_preload_end();
 	drm_gem_object_unreference_unlocked(obj);
 	return ret;
 }
@@ -644,6 +636,59 @@ void drm_gem_vm_close(struct vm_area_struct *vma)
 }
 EXPORT_SYMBOL(drm_gem_vm_close);
 
+/**
+ * drm_gem_mmap_obj - memory map a GEM object
+ * @obj: the GEM object to map
+ * @obj_size: the object size to be mapped, in bytes
+ * @vma: VMA for the area to be mapped
+ *
+ * Set up the VMA to prepare mapping of the GEM object using the gem_vm_ops
+ * provided by the driver. Depending on their requirements, drivers can either
+ * provide a fault handler in their gem_vm_ops (in which case any accesses to
+ * the object will be trapped, to perform migration, GTT binding, surface
+ * register allocation, or performance monitoring), or mmap the buffer memory
+ * synchronously after calling drm_gem_mmap_obj.
+ *
+ * This function is mainly intended to implement the DMABUF mmap operation, when
+ * the GEM object is not looked up based on its fake offset. To implement the
+ * DRM mmap operation, drivers should use the drm_gem_mmap() function.
+ *
+ * NOTE: This function has to be protected with dev->struct_mutex
+ *
+ * Return 0 or success or -EINVAL if the object size is smaller than the VMA
+ * size, or if no gem_vm_ops are provided.
+ */
+int drm_gem_mmap_obj(struct drm_gem_object *obj, unsigned long obj_size,
+		     struct vm_area_struct *vma)
+{
+	struct drm_device *dev = obj->dev;
+
+	lockdep_assert_held(&dev->struct_mutex);
+
+	/* Check for valid size. */
+	if (obj_size < vma->vm_end - vma->vm_start)
+		return -EINVAL;
+
+	if (!dev->driver->gem_vm_ops)
+		return -EINVAL;
+
+	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
+	vma->vm_ops = dev->driver->gem_vm_ops;
+	vma->vm_private_data = obj;
+	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+
+	/* Take a ref for this mapping of the object, so that the fault
+	 * handler can dereference the mmap offset's pointer to the object.
+	 * This reference is cleaned up by the corresponding vm_close
+	 * (which should happen whether the vma was created by this call, or
+	 * by a vm_open due to mremap or partial unmap or whatever).
+	 */
+	drm_gem_object_reference(obj);
+
+	drm_vm_open_locked(dev, vma);
+	return 0;
+}
+EXPORT_SYMBOL(drm_gem_mmap_obj);
 
 /**
  * drm_gem_mmap - memory map routine for GEM objects
@@ -653,11 +698,9 @@ EXPORT_SYMBOL(drm_gem_vm_close);
  * If a driver supports GEM object mapping, mmap calls on the DRM file
  * descriptor will end up here.
  *
- * If we find the object based on the offset passed in (vma->vm_pgoff will
+ * Look up the GEM object based on the offset passed in (vma->vm_pgoff will
  * contain the fake offset we created when the GTT map ioctl was called on
- * the object), we set up the driver fault handler so that any accesses
- * to the object can be trapped, to perform migration, GTT binding, surface
- * register allocation, or performance monitoring.
+ * the object) and map it with a call to drm_gem_mmap_obj().
  */
 int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 {
@@ -665,7 +708,6 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct drm_device *dev = priv->minor->dev;
 	struct drm_gem_mm *mm = dev->mm_private;
 	struct drm_local_map *map = NULL;
-	struct drm_gem_object *obj;
 	struct drm_hash_item *hash;
 	int ret = 0;
 
@@ -686,32 +728,7 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 		goto out_unlock;
 	}
 
-	/* Check for valid size. */
-	if (map->size < vma->vm_end - vma->vm_start) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	obj = map->handle;
-	if (!obj->dev->driver->gem_vm_ops) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_ops = obj->dev->driver->gem_vm_ops;
-	vma->vm_private_data = map->handle;
-	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
-
-	/* Take a ref for this mapping of the object, so that the fault
-	 * handler can dereference the mmap offset's pointer to the object.
-	 * This reference is cleaned up by the corresponding vm_close
-	 * (which should happen whether the vma was created by this call, or
-	 * by a vm_open due to mremap or partial unmap or whatever).
-	 */
-	drm_gem_object_reference(obj);
-
-	drm_vm_open_locked(dev, vma);
+	ret = drm_gem_mmap_obj(map->handle, map->size, vma);
 
 out_unlock:
 	mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index 0a7e011..ece72a8 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -21,6 +21,7 @@
 #include <linux/slab.h>
 #include <linux/mutex.h>
 #include <linux/export.h>
+#include <linux/dma-buf.h>
 #include <linux/dma-mapping.h>
 
 #include <drm/drmP.h>
@@ -32,11 +33,44 @@ static unsigned int get_gem_mmap_offset(struct drm_gem_object *obj)
 	return (unsigned int)obj->map_list.hash.key << PAGE_SHIFT;
 }
 
-static void drm_gem_cma_buf_destroy(struct drm_device *drm,
-		struct drm_gem_cma_object *cma_obj)
+/*
+ * __drm_gem_cma_create - Create a GEM CMA object without allocating memory
+ * @drm: The drm device
+ * @size: The GEM object size
+ *
+ * This function creates and initializes a GEM CMA object of the given size, but
+ * doesn't allocate any memory to back the object.
+ *
+ * Return a struct drm_gem_cma_object* on success or ERR_PTR values on failure.
+ */
+static struct drm_gem_cma_object *
+__drm_gem_cma_create(struct drm_device *drm, unsigned int size)
 {
-	dma_free_writecombine(drm->dev, cma_obj->base.size, cma_obj->vaddr,
-			cma_obj->paddr);
+	struct drm_gem_cma_object *cma_obj;
+	struct drm_gem_object *gem_obj;
+	int ret;
+
+	cma_obj = kzalloc(sizeof(*cma_obj), GFP_KERNEL);
+	if (!cma_obj)
+		return ERR_PTR(-ENOMEM);
+
+	gem_obj = &cma_obj->base;
+
+	ret = drm_gem_object_init(drm, gem_obj, size);
+	if (ret)
+		goto error;
+
+	ret = drm_gem_create_mmap_offset(gem_obj);
+	if (ret) {
+		drm_gem_object_release(gem_obj);
+		goto error;
+	}
+
+	return cma_obj;
+
+error:
+	kfree(cma_obj);
+	return ERR_PTR(ret);
 }
 
 /*
@@ -49,44 +83,42 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm,
 		unsigned int size)
 {
 	struct drm_gem_cma_object *cma_obj;
-	struct drm_gem_object *gem_obj;
+	struct sg_table *sgt = NULL;
 	int ret;
 
 	size = round_up(size, PAGE_SIZE);
 
-	cma_obj = kzalloc(sizeof(*cma_obj), GFP_KERNEL);
-	if (!cma_obj)
-		return ERR_PTR(-ENOMEM);
+	cma_obj = __drm_gem_cma_create(drm, size);
+	if (IS_ERR(cma_obj))
+		return cma_obj;
 
 	cma_obj->vaddr = dma_alloc_writecombine(drm->dev, size,
 			&cma_obj->paddr, GFP_KERNEL | __GFP_NOWARN);
 	if (!cma_obj->vaddr) {
-		dev_err(drm->dev, "failed to allocate buffer with size %d\n", size);
+		dev_err(drm->dev, "failed to allocate buffer with size %d\n",
+			size);
 		ret = -ENOMEM;
-		goto err_dma_alloc;
+		goto error;
 	}
 
-	gem_obj = &cma_obj->base;
+	sgt = kzalloc(sizeof(*cma_obj->sgt), GFP_KERNEL);
+	if (sgt == NULL) {
+		ret = -ENOMEM;
+		goto error;
+	}
 
-	ret = drm_gem_object_init(drm, gem_obj, size);
-	if (ret)
-		goto err_obj_init;
+	ret = dma_get_sgtable(drm->dev, sgt, cma_obj->vaddr,
+			      cma_obj->paddr, size);
+	if (ret < 0)
+		goto error;
 
-	ret = drm_gem_create_mmap_offset(gem_obj);
-	if (ret)
-		goto err_create_mmap_offset;
+	cma_obj->sgt = sgt;
 
 	return cma_obj;
 
-err_create_mmap_offset:
-	drm_gem_object_release(gem_obj);
-
-err_obj_init:
-	drm_gem_cma_buf_destroy(drm, cma_obj);
-
-err_dma_alloc:
-	kfree(cma_obj);
-
+error:
+	kfree(sgt);
+	drm_gem_cma_free_object(&cma_obj->base);
 	return ERR_PTR(ret);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_create);
@@ -143,11 +175,20 @@ void drm_gem_cma_free_object(struct drm_gem_object *gem_obj)
 	if (gem_obj->map_list.map)
 		drm_gem_free_mmap_offset(gem_obj);
 
-	drm_gem_object_release(gem_obj);
-
 	cma_obj = to_drm_gem_cma_obj(gem_obj);
 
-	drm_gem_cma_buf_destroy(gem_obj->dev, cma_obj);
+	if (cma_obj->vaddr) {
+		dma_free_writecombine(gem_obj->dev->dev, cma_obj->base.size,
+				      cma_obj->vaddr, cma_obj->paddr);
+		if (cma_obj->sgt) {
+			sg_free_table(cma_obj->sgt);
+			kfree(cma_obj->sgt);
+		}
+	} else if (gem_obj->import_attach) {
+		drm_prime_gem_destroy(gem_obj, cma_obj->sgt);
+	}
+
+	drm_gem_object_release(gem_obj);
 
 	kfree(cma_obj);
 }
@@ -174,10 +215,7 @@ int drm_gem_cma_dumb_create(struct drm_file *file_priv,
 
 	cma_obj = drm_gem_cma_create_with_handle(file_priv, dev,
 			args->size, &args->handle);
-	if (IS_ERR(cma_obj))
-		return PTR_ERR(cma_obj);
-
-	return 0;
+	return PTR_RET(cma_obj);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create);
 
@@ -215,13 +253,26 @@ const struct vm_operations_struct drm_gem_cma_vm_ops = {
 };
 EXPORT_SYMBOL_GPL(drm_gem_cma_vm_ops);
 
+static int drm_gem_cma_mmap_obj(struct drm_gem_cma_object *cma_obj,
+				struct vm_area_struct *vma)
+{
+	int ret;
+
+	ret = remap_pfn_range(vma, vma->vm_start, cma_obj->paddr >> PAGE_SHIFT,
+			vma->vm_end - vma->vm_start, vma->vm_page_prot);
+	if (ret)
+		drm_gem_vm_close(vma);
+
+	return ret;
+}
+
 /*
  * drm_gem_cma_mmap - (struct file_operation)->mmap callback function
  */
 int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
 {
-	struct drm_gem_object *gem_obj;
 	struct drm_gem_cma_object *cma_obj;
+	struct drm_gem_object *gem_obj;
 	int ret;
 
 	ret = drm_gem_mmap(filp, vma);
@@ -231,12 +282,7 @@ int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
 	gem_obj = vma->vm_private_data;
 	cma_obj = to_drm_gem_cma_obj(gem_obj);
 
-	ret = remap_pfn_range(vma, vma->vm_start, cma_obj->paddr >> PAGE_SHIFT,
-			vma->vm_end - vma->vm_start, vma->vm_page_prot);
-	if (ret)
-		drm_gem_vm_close(vma);
-
-	return ret;
+	return drm_gem_cma_mmap_obj(cma_obj, vma);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
 
@@ -270,3 +316,82 @@ void drm_gem_cma_describe(struct drm_gem_cma_object *cma_obj, struct seq_file *m
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_describe);
 #endif
+
+/* low-level interface prime helpers */
+struct sg_table *drm_gem_cma_prime_get_sg_table(struct drm_gem_object *obj)
+{
+	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
+	struct sg_table *sgt;
+	int ret;
+
+	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
+	if (!sgt)
+		return NULL;
+
+	ret = dma_get_sgtable(obj->dev->dev, sgt, cma_obj->vaddr,
+			      cma_obj->paddr, obj->size);
+	if (ret < 0)
+		goto out;
+
+	return sgt;
+
+out:
+	kfree(sgt);
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_get_sg_table);
+
+struct drm_gem_object *
+drm_gem_cma_prime_import_sg_table(struct drm_device *dev, size_t size,
+				  struct sg_table *sgt)
+{
+	struct drm_gem_cma_object *cma_obj;
+
+	if (sgt->nents != 1)
+		return ERR_PTR(-EINVAL);
+
+	/* Create a CMA GEM buffer. */
+	cma_obj = __drm_gem_cma_create(dev, size);
+	if (IS_ERR(cma_obj))
+		return ERR_PTR(PTR_ERR(cma_obj));
+
+	cma_obj->paddr = sg_dma_address(sgt->sgl);
+	cma_obj->sgt = sgt;
+
+	DRM_DEBUG_PRIME("dma_addr = 0x%x, size = %zu\n", cma_obj->paddr, size);
+
+	return &cma_obj->base;
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_import_sg_table);
+
+int drm_gem_cma_prime_mmap(struct drm_gem_object *obj,
+			   struct vm_area_struct *vma)
+{
+	struct drm_gem_cma_object *cma_obj;
+	struct drm_device *dev = obj->dev;
+	int ret;
+
+	mutex_lock(&dev->struct_mutex);
+	ret = drm_gem_mmap_obj(obj, obj->size, vma);
+	mutex_unlock(&dev->struct_mutex);
+	if (ret < 0)
+		return ret;
+
+	cma_obj = to_drm_gem_cma_obj(obj);
+	return drm_gem_cma_mmap_obj(cma_obj, vma);
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_mmap);
+
+void *drm_gem_cma_prime_vmap(struct drm_gem_object *obj)
+{
+	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
+
+	return cma_obj->vaddr;
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_vmap);
+
+void drm_gem_cma_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
+{
+	/* Nothing to do */
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_vunmap);
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index e77bd8b..ffd7a7b 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -38,6 +38,9 @@
 
 #include <linux/pci.h>
 #include <linux/export.h>
+#ifdef CONFIG_X86
+#include <asm/mtrr.h>
+#endif
 
 /**
  * Get the bus id.
@@ -181,7 +184,17 @@ int drm_getmap(struct drm_device *dev, void *data,
 	map->type = r_list->map->type;
 	map->flags = r_list->map->flags;
 	map->handle = (void *)(unsigned long) r_list->user_token;
-	map->mtrr = r_list->map->mtrr;
+
+#ifdef CONFIG_X86
+	/*
+	 * There appears to be exactly one user of the mtrr index: dritest.
+	 * It's easy enough to keep it working on non-PAT systems.
+	 */
+	map->mtrr = phys_wc_to_mtrr_index(r_list->map->mtrr);
+#else
+	map->mtrr = -1;
+#endif
+
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index 07cf99c..543b9b3 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -669,7 +669,7 @@ int drm_mm_clean(struct drm_mm * mm)
 }
 EXPORT_SYMBOL(drm_mm_clean);
 
-int drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
+void drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
 {
 	INIT_LIST_HEAD(&mm->hole_stack);
 	INIT_LIST_HEAD(&mm->unused_nodes);
@@ -690,8 +690,6 @@ int drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
 	list_add_tail(&mm->head_node.hole_stack, &mm->hole_stack);
 
 	mm->color_adjust = NULL;
-
-	return 0;
 }
 EXPORT_SYMBOL(drm_mm_init);
 
@@ -699,8 +697,8 @@ void drm_mm_takedown(struct drm_mm * mm)
 {
 	struct drm_mm_node *entry, *next;
 
-	if (!list_empty(&mm->head_node.node_list)) {
-		DRM_ERROR("Memory manager not clean. Delaying takedown\n");
+	if (WARN(!list_empty(&mm->head_node.node_list),
+		 "Memory manager not clean. Delaying takedown\n")) {
 		return;
 	}
 
@@ -716,36 +714,37 @@ void drm_mm_takedown(struct drm_mm * mm)
 }
 EXPORT_SYMBOL(drm_mm_takedown);
 
-void drm_mm_debug_table(struct drm_mm *mm, const char *prefix)
+static unsigned long drm_mm_debug_hole(struct drm_mm_node *entry,
+				       const char *prefix)
 {
-	struct drm_mm_node *entry;
-	unsigned long total_used = 0, total_free = 0, total = 0;
 	unsigned long hole_start, hole_end, hole_size;
 
-	hole_start = drm_mm_hole_node_start(&mm->head_node);
-	hole_end = drm_mm_hole_node_end(&mm->head_node);
-	hole_size = hole_end - hole_start;
-	if (hole_size)
+	if (entry->hole_follows) {
+		hole_start = drm_mm_hole_node_start(entry);
+		hole_end = drm_mm_hole_node_end(entry);
+		hole_size = hole_end - hole_start;
 		printk(KERN_DEBUG "%s 0x%08lx-0x%08lx: %8lu: free\n",
 			prefix, hole_start, hole_end,
 			hole_size);
-	total_free += hole_size;
+		return hole_size;
+	}
+
+	return 0;
+}
+
+void drm_mm_debug_table(struct drm_mm *mm, const char *prefix)
+{
+	struct drm_mm_node *entry;
+	unsigned long total_used = 0, total_free = 0, total = 0;
+
+	total_free += drm_mm_debug_hole(&mm->head_node, prefix);
 
 	drm_mm_for_each_node(entry, mm) {
 		printk(KERN_DEBUG "%s 0x%08lx-0x%08lx: %8lu: used\n",
 			prefix, entry->start, entry->start + entry->size,
 			entry->size);
 		total_used += entry->size;
-
-		if (entry->hole_follows) {
-			hole_start = drm_mm_hole_node_start(entry);
-			hole_end = drm_mm_hole_node_end(entry);
-			hole_size = hole_end - hole_start;
-			printk(KERN_DEBUG "%s 0x%08lx-0x%08lx: %8lu: free\n",
-				prefix, hole_start, hole_end,
-				hole_size);
-			total_free += hole_size;
-		}
+		total_free += drm_mm_debug_hole(entry, prefix);
 	}
 	total = total_free + total_used;
 
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index a371ff8..a6729bf 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -535,6 +535,8 @@ int drm_display_mode_from_videomode(const struct videomode *vm,
 		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
 	if (vm->flags & DISPLAY_FLAGS_DOUBLESCAN)
 		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	if (vm->flags & DISPLAY_FLAGS_DOUBLECLK)
+		dmode->flags |= DRM_MODE_FLAG_DBLCLK;
 	drm_mode_set_name(dmode);
 
 	return 0;
@@ -787,16 +789,17 @@ EXPORT_SYMBOL(drm_mode_set_crtcinfo);
  * LOCKING:
  * None.
  *
- * Copy an existing mode into another mode, preserving the object id
- * of the destination mode.
+ * Copy an existing mode into another mode, preserving the object id and
+ * list head of the destination mode.
  */
 void drm_mode_copy(struct drm_display_mode *dst, const struct drm_display_mode *src)
 {
 	int id = dst->base.id;
+	struct list_head head = dst->head;
 
 	*dst = *src;
 	dst->base.id = id;
-	INIT_LIST_HEAD(&dst->head);
+	dst->head = head;
 }
 EXPORT_SYMBOL(drm_mode_copy);
 
@@ -1017,6 +1020,11 @@ static int drm_mode_compare(void *priv, struct list_head *lh_a, struct list_head
 	diff = b->hdisplay * b->vdisplay - a->hdisplay * a->vdisplay;
 	if (diff)
 		return diff;
+
+	diff = b->vrefresh - a->vrefresh;
+	if (diff)
+		return diff;
+
 	diff = b->clock - a->clock;
 	return diff;
 }
diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 14194b6..80c0b2b 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -278,10 +278,10 @@ static int drm_pci_agp_init(struct drm_device *dev)
 		}
 		if (drm_core_has_MTRR(dev)) {
 			if (dev->agp)
-				dev->agp->agp_mtrr =
-					mtrr_add(dev->agp->agp_info.aper_base,
-						 dev->agp->agp_info.aper_size *
-						 1024 * 1024, MTRR_TYPE_WRCOMB, 1);
+				dev->agp->agp_mtrr = arch_phys_wc_add(
+					dev->agp->agp_info.aper_base,
+					dev->agp->agp_info.aper_size *
+					1024 * 1024);
 		}
 	}
 	return 0;
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 5b7b911..85e450e 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -62,20 +62,125 @@ struct drm_prime_member {
 	struct dma_buf *dma_buf;
 	uint32_t handle;
 };
-static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle);
+
+struct drm_prime_attachment {
+	struct sg_table *sgt;
+	enum dma_data_direction dir;
+};
+
+static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle)
+{
+	struct drm_prime_member *member;
+
+	member = kmalloc(sizeof(*member), GFP_KERNEL);
+	if (!member)
+		return -ENOMEM;
+
+	get_dma_buf(dma_buf);
+	member->dma_buf = dma_buf;
+	member->handle = handle;
+	list_add(&member->entry, &prime_fpriv->head);
+	return 0;
+}
+
+static int drm_gem_map_attach(struct dma_buf *dma_buf,
+			      struct device *target_dev,
+			      struct dma_buf_attachment *attach)
+{
+	struct drm_prime_attachment *prime_attach;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+
+	prime_attach = kzalloc(sizeof(*prime_attach), GFP_KERNEL);
+	if (!prime_attach)
+		return -ENOMEM;
+
+	prime_attach->dir = DMA_NONE;
+	attach->priv = prime_attach;
+
+	if (!dev->driver->gem_prime_pin)
+		return 0;
+
+	return dev->driver->gem_prime_pin(obj);
+}
+
+static void drm_gem_map_detach(struct dma_buf *dma_buf,
+			       struct dma_buf_attachment *attach)
+{
+	struct drm_prime_attachment *prime_attach = attach->priv;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+	struct sg_table *sgt;
+
+	if (dev->driver->gem_prime_unpin)
+		dev->driver->gem_prime_unpin(obj);
+
+	if (!prime_attach)
+		return;
+
+	sgt = prime_attach->sgt;
+	if (sgt) {
+		if (prime_attach->dir != DMA_NONE)
+			dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents,
+					prime_attach->dir);
+		sg_free_table(sgt);
+	}
+
+	kfree(sgt);
+	kfree(prime_attach);
+	attach->priv = NULL;
+}
+
+static void drm_prime_remove_buf_handle_locked(
+		struct drm_prime_file_private *prime_fpriv,
+		struct dma_buf *dma_buf)
+{
+	struct drm_prime_member *member, *safe;
+
+	list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) {
+		if (member->dma_buf == dma_buf) {
+			dma_buf_put(dma_buf);
+			list_del(&member->entry);
+			kfree(member);
+		}
+	}
+}
 
 static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		enum dma_data_direction dir)
 {
+	struct drm_prime_attachment *prime_attach = attach->priv;
 	struct drm_gem_object *obj = attach->dmabuf->priv;
 	struct sg_table *sgt;
 
+	if (WARN_ON(dir == DMA_NONE || !prime_attach))
+		return ERR_PTR(-EINVAL);
+
+	/* return the cached mapping when possible */
+	if (prime_attach->dir == dir)
+		return prime_attach->sgt;
+
+	/*
+	 * two mappings with different directions for the same attachment are
+	 * not allowed
+	 */
+	if (WARN_ON(prime_attach->dir != DMA_NONE))
+		return ERR_PTR(-EBUSY);
+
 	mutex_lock(&obj->dev->struct_mutex);
 
 	sgt = obj->dev->driver->gem_prime_get_sg_table(obj);
 
-	if (!IS_ERR_OR_NULL(sgt))
-		dma_map_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+	if (!IS_ERR(sgt)) {
+		if (!dma_map_sg(attach->dev, sgt->sgl, sgt->nents, dir)) {
+			sg_free_table(sgt);
+			kfree(sgt);
+			sgt = ERR_PTR(-ENOMEM);
+		} else {
+			prime_attach->sgt = sgt;
+			prime_attach->dir = dir;
+		}
+	}
 
 	mutex_unlock(&obj->dev->struct_mutex);
 	return sgt;
@@ -84,9 +189,7 @@ static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach,
 static void drm_gem_unmap_dma_buf(struct dma_buf_attachment *attach,
 		struct sg_table *sgt, enum dma_data_direction dir)
 {
-	dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
-	sg_free_table(sgt);
-	kfree(sgt);
+	/* nothing to be done here */
 }
 
 static void drm_gem_dmabuf_release(struct dma_buf *dma_buf)
@@ -142,10 +245,18 @@ static void drm_gem_dmabuf_kunmap(struct dma_buf *dma_buf,
 static int drm_gem_dmabuf_mmap(struct dma_buf *dma_buf,
 		struct vm_area_struct *vma)
 {
-	return -EINVAL;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+
+	if (!dev->driver->gem_prime_mmap)
+		return -ENOSYS;
+
+	return dev->driver->gem_prime_mmap(obj, vma);
 }
 
 static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
+	.attach = drm_gem_map_attach,
+	.detach = drm_gem_map_detach,
 	.map_dma_buf = drm_gem_map_dma_buf,
 	.unmap_dma_buf = drm_gem_unmap_dma_buf,
 	.release = drm_gem_dmabuf_release,
@@ -185,11 +296,6 @@ static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
 struct dma_buf *drm_gem_prime_export(struct drm_device *dev,
 				     struct drm_gem_object *obj, int flags)
 {
-	if (dev->driver->gem_prime_pin) {
-		int ret = dev->driver->gem_prime_pin(obj);
-		if (ret)
-			return ERR_PTR(ret);
-	}
 	return dma_buf_export(obj, &drm_gem_prime_dmabuf_ops, obj->size, flags);
 }
 EXPORT_SYMBOL(drm_gem_prime_export);
@@ -235,15 +341,34 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 	ret = drm_prime_add_buf_handle(&file_priv->prime,
 				       obj->export_dma_buf, handle);
 	if (ret)
-		goto out;
+		goto fail_put_dmabuf;
+
+	ret = dma_buf_fd(buf, flags);
+	if (ret < 0)
+		goto fail_rm_handle;
 
-	*prime_fd = dma_buf_fd(buf, flags);
+	*prime_fd = ret;
 	mutex_unlock(&file_priv->prime.lock);
 	return 0;
 
 out_have_obj:
 	get_dma_buf(dmabuf);
-	*prime_fd = dma_buf_fd(dmabuf, flags);
+	ret = dma_buf_fd(dmabuf, flags);
+	if (ret < 0) {
+		dma_buf_put(dmabuf);
+	} else {
+		*prime_fd = ret;
+		ret = 0;
+	}
+
+	goto out;
+
+fail_rm_handle:
+	drm_prime_remove_buf_handle_locked(&file_priv->prime, buf);
+fail_put_dmabuf:
+	/* clear NOT to be checked when releasing dma_buf */
+	obj->export_dma_buf = NULL;
+	dma_buf_put(buf);
 out:
 	drm_gem_object_unreference_unlocked(obj);
 	mutex_unlock(&file_priv->prime.lock);
@@ -276,7 +401,7 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
 
 	attach = dma_buf_attach(dma_buf, dev->dev);
 	if (IS_ERR(attach))
-		return ERR_PTR(PTR_ERR(attach));
+		return ERR_CAST(attach);
 
 	get_dma_buf(dma_buf);
 
@@ -412,8 +537,10 @@ struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages)
 	int ret;
 
 	sg = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
-	if (!sg)
+	if (!sg) {
+		ret = -ENOMEM;
 		goto out;
+	}
 
 	ret = sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
 				nr_pages << PAGE_SHIFT, GFP_KERNEL);
@@ -423,7 +550,7 @@ struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages)
 	return sg;
 out:
 	kfree(sg);
-	return NULL;
+	return ERR_PTR(ret);
 }
 EXPORT_SYMBOL(drm_prime_pages_to_sg);
 
@@ -492,21 +619,6 @@ void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv)
 }
 EXPORT_SYMBOL(drm_prime_destroy_file_private);
 
-static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle)
-{
-	struct drm_prime_member *member;
-
-	member = kmalloc(sizeof(*member), GFP_KERNEL);
-	if (!member)
-		return -ENOMEM;
-
-	get_dma_buf(dma_buf);
-	member->dma_buf = dma_buf;
-	member->handle = handle;
-	list_add(&member->entry, &prime_fpriv->head);
-	return 0;
-}
-
 int drm_prime_lookup_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t *handle)
 {
 	struct drm_prime_member *member;
@@ -523,16 +635,8 @@ EXPORT_SYMBOL(drm_prime_lookup_buf_handle);
 
 void drm_prime_remove_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf)
 {
-	struct drm_prime_member *member, *safe;
-
 	mutex_lock(&prime_fpriv->lock);
-	list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) {
-		if (member->dma_buf == dma_buf) {
-			dma_buf_put(dma_buf);
-			list_del(&member->entry);
-			kfree(member);
-		}
-	}
+	drm_prime_remove_buf_handle_locked(prime_fpriv, dma_buf);
 	mutex_unlock(&prime_fpriv->lock);
 }
 EXPORT_SYMBOL(drm_prime_remove_buf_handle);
diff --git a/drivers/gpu/drm/drm_rect.c b/drivers/gpu/drm/drm_rect.c
new file mode 100644
index 0000000..7047ca0
--- /dev/null
+++ b/drivers/gpu/drm/drm_rect.c
@@ -0,0 +1,295 @@
+/*
+ * Copyright (C) 2011-2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/kernel.h>
+#include <drm/drmP.h>
+#include <drm/drm_rect.h>
+
+/**
+ * drm_rect_intersect - intersect two rectangles
+ * @r1: first rectangle
+ * @r2: second rectangle
+ *
+ * Calculate the intersection of rectangles @r1 and @r2.
+ * @r1 will be overwritten with the intersection.
+ *
+ * RETURNS:
+ * %true if rectangle @r1 is still visible after the operation,
+ * %false otherwise.
+ */
+bool drm_rect_intersect(struct drm_rect *r1, const struct drm_rect *r2)
+{
+	r1->x1 = max(r1->x1, r2->x1);
+	r1->y1 = max(r1->y1, r2->y1);
+	r1->x2 = min(r1->x2, r2->x2);
+	r1->y2 = min(r1->y2, r2->y2);
+
+	return drm_rect_visible(r1);
+}
+EXPORT_SYMBOL(drm_rect_intersect);
+
+/**
+ * drm_rect_clip_scaled - perform a scaled clip operation
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @clip: clip rectangle
+ * @hscale: horizontal scaling factor
+ * @vscale: vertical scaling factor
+ *
+ * Clip rectangle @dst by rectangle @clip. Clip rectangle @src by the
+ * same amounts multiplied by @hscale and @vscale.
+ *
+ * RETURNS:
+ * %true if rectangle @dst is still visible after being clipped,
+ * %false otherwise
+ */
+bool drm_rect_clip_scaled(struct drm_rect *src, struct drm_rect *dst,
+			  const struct drm_rect *clip,
+			  int hscale, int vscale)
+{
+	int diff;
+
+	diff = clip->x1 - dst->x1;
+	if (diff > 0) {
+		int64_t tmp = src->x1 + (int64_t) diff * hscale;
+		src->x1 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+	diff = clip->y1 - dst->y1;
+	if (diff > 0) {
+		int64_t tmp = src->y1 + (int64_t) diff * vscale;
+		src->y1 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+	diff = dst->x2 - clip->x2;
+	if (diff > 0) {
+		int64_t tmp = src->x2 - (int64_t) diff * hscale;
+		src->x2 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+	diff = dst->y2 - clip->y2;
+	if (diff > 0) {
+		int64_t tmp = src->y2 - (int64_t) diff * vscale;
+		src->y2 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+
+	return drm_rect_intersect(dst, clip);
+}
+EXPORT_SYMBOL(drm_rect_clip_scaled);
+
+static int drm_calc_scale(int src, int dst)
+{
+	int scale = 0;
+
+	if (src < 0 || dst < 0)
+		return -EINVAL;
+
+	if (dst == 0)
+		return 0;
+
+	scale = src / dst;
+
+	return scale;
+}
+
+/**
+ * drm_rect_calc_hscale - calculate the horizontal scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_hscale: minimum allowed horizontal scaling factor
+ * @max_hscale: maximum allowed horizontal scaling factor
+ *
+ * Calculate the horizontal scaling factor as
+ * (@src width) / (@dst width).
+ *
+ * RETURNS:
+ * The horizontal scaling factor, or errno of out of limits.
+ */
+int drm_rect_calc_hscale(const struct drm_rect *src,
+			 const struct drm_rect *dst,
+			 int min_hscale, int max_hscale)
+{
+	int src_w = drm_rect_width(src);
+	int dst_w = drm_rect_width(dst);
+	int hscale = drm_calc_scale(src_w, dst_w);
+
+	if (hscale < 0 || dst_w == 0)
+		return hscale;
+
+	if (hscale < min_hscale || hscale > max_hscale)
+		return -ERANGE;
+
+	return hscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_hscale);
+
+/**
+ * drm_rect_calc_vscale - calculate the vertical scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_vscale: minimum allowed vertical scaling factor
+ * @max_vscale: maximum allowed vertical scaling factor
+ *
+ * Calculate the vertical scaling factor as
+ * (@src height) / (@dst height).
+ *
+ * RETURNS:
+ * The vertical scaling factor, or errno of out of limits.
+ */
+int drm_rect_calc_vscale(const struct drm_rect *src,
+			 const struct drm_rect *dst,
+			 int min_vscale, int max_vscale)
+{
+	int src_h = drm_rect_height(src);
+	int dst_h = drm_rect_height(dst);
+	int vscale = drm_calc_scale(src_h, dst_h);
+
+	if (vscale < 0 || dst_h == 0)
+		return vscale;
+
+	if (vscale < min_vscale || vscale > max_vscale)
+		return -ERANGE;
+
+	return vscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_vscale);
+
+/**
+ * drm_calc_hscale_relaxed - calculate the horizontal scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_hscale: minimum allowed horizontal scaling factor
+ * @max_hscale: maximum allowed horizontal scaling factor
+ *
+ * Calculate the horizontal scaling factor as
+ * (@src width) / (@dst width).
+ *
+ * If the calculated scaling factor is below @min_vscale,
+ * decrease the height of rectangle @dst to compensate.
+ *
+ * If the calculated scaling factor is above @max_vscale,
+ * decrease the height of rectangle @src to compensate.
+ *
+ * RETURNS:
+ * The horizontal scaling factor.
+ */
+int drm_rect_calc_hscale_relaxed(struct drm_rect *src,
+				 struct drm_rect *dst,
+				 int min_hscale, int max_hscale)
+{
+	int src_w = drm_rect_width(src);
+	int dst_w = drm_rect_width(dst);
+	int hscale = drm_calc_scale(src_w, dst_w);
+
+	if (hscale < 0 || dst_w == 0)
+		return hscale;
+
+	if (hscale < min_hscale) {
+		int max_dst_w = src_w / min_hscale;
+
+		drm_rect_adjust_size(dst, max_dst_w - dst_w, 0);
+
+		return min_hscale;
+	}
+
+	if (hscale > max_hscale) {
+		int max_src_w = dst_w * max_hscale;
+
+		drm_rect_adjust_size(src, max_src_w - src_w, 0);
+
+		return max_hscale;
+	}
+
+	return hscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_hscale_relaxed);
+
+/**
+ * drm_rect_calc_vscale_relaxed - calculate the vertical scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_vscale: minimum allowed vertical scaling factor
+ * @max_vscale: maximum allowed vertical scaling factor
+ *
+ * Calculate the vertical scaling factor as
+ * (@src height) / (@dst height).
+ *
+ * If the calculated scaling factor is below @min_vscale,
+ * decrease the height of rectangle @dst to compensate.
+ *
+ * If the calculated scaling factor is above @max_vscale,
+ * decrease the height of rectangle @src to compensate.
+ *
+ * RETURNS:
+ * The vertical scaling factor.
+ */
+int drm_rect_calc_vscale_relaxed(struct drm_rect *src,
+				 struct drm_rect *dst,
+				 int min_vscale, int max_vscale)
+{
+	int src_h = drm_rect_height(src);
+	int dst_h = drm_rect_height(dst);
+	int vscale = drm_calc_scale(src_h, dst_h);
+
+	if (vscale < 0 || dst_h == 0)
+		return vscale;
+
+	if (vscale < min_vscale) {
+		int max_dst_h = src_h / min_vscale;
+
+		drm_rect_adjust_size(dst, 0, max_dst_h - dst_h);
+
+		return min_vscale;
+	}
+
+	if (vscale > max_vscale) {
+		int max_src_h = dst_h * max_vscale;
+
+		drm_rect_adjust_size(src, 0, max_src_h - src_h);
+
+		return max_vscale;
+	}
+
+	return vscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_vscale_relaxed);
+
+/**
+ * drm_rect_debug_print - print the rectangle information
+ * @r: rectangle to print
+ * @fixed_point: rectangle is in 16.16 fixed point format
+ */
+void drm_rect_debug_print(const struct drm_rect *r, bool fixed_point)
+{
+	int w = drm_rect_width(r);
+	int h = drm_rect_height(r);
+
+	if (fixed_point)
+		DRM_DEBUG_KMS("%d.%06ux%d.%06u%+d.%06u%+d.%06u\n",
+			      w >> 16, ((w & 0xffff) * 15625) >> 10,
+			      h >> 16, ((h & 0xffff) * 15625) >> 10,
+			      r->x1 >> 16, ((r->x1 & 0xffff) * 15625) >> 10,
+			      r->y1 >> 16, ((r->y1 & 0xffff) * 15625) >> 10);
+	else
+		DRM_DEBUG_KMS("%dx%d%+d%+d\n", w, h, r->x1, r->y1);
+}
+EXPORT_SYMBOL(drm_rect_debug_print);
diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c
index 16f3ec5..327ca19 100644
--- a/drivers/gpu/drm/drm_stub.c
+++ b/drivers/gpu/drm/drm_stub.c
@@ -203,7 +203,7 @@ EXPORT_SYMBOL(drm_master_put);
 int drm_setmaster_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv)
 {
-	int ret;
+	int ret = 0;
 
 	if (file_priv->is_master)
 		return 0;
@@ -229,7 +229,7 @@ int drm_setmaster_ioctl(struct drm_device *dev, void *data,
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	return 0;
+	return ret;
 }
 
 int drm_dropmaster_ioctl(struct drm_device *dev, void *data,
@@ -451,14 +451,8 @@ void drm_put_dev(struct drm_device *dev)
 
 	drm_lastclose(dev);
 
-	if (drm_core_has_MTRR(dev) && drm_core_has_AGP(dev) &&
-	    dev->agp && dev->agp->agp_mtrr >= 0) {
-		int retval;
-		retval = mtrr_del(dev->agp->agp_mtrr,
-				  dev->agp->agp_info.aper_base,
-				  dev->agp->agp_info.aper_size * 1024 * 1024);
-		DRM_DEBUG("mtrr_del=%d\n", retval);
-	}
+	if (drm_core_has_MTRR(dev) && drm_core_has_AGP(dev) && dev->agp)
+		arch_phys_wc_del(dev->agp->agp_mtrr);
 
 	if (dev->driver->unload)
 		dev->driver->unload(dev);
diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index 0229665..2290b3b 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -30,14 +30,14 @@ static struct device_type drm_sysfs_device_minor = {
 };
 
 /**
- * drm_class_suspend - DRM class suspend hook
+ * __drm_class_suspend - internal DRM class suspend routine
  * @dev: Linux device to suspend
  * @state: power state to enter
  *
  * Just figures out what the actual struct drm_device associated with
  * @dev is and calls its suspend hook, if present.
  */
-static int drm_class_suspend(struct device *dev, pm_message_t state)
+static int __drm_class_suspend(struct device *dev, pm_message_t state)
 {
 	if (dev->type == &drm_sysfs_device_minor) {
 		struct drm_minor *drm_minor = to_drm_minor(dev);
@@ -52,6 +52,26 @@ static int drm_class_suspend(struct device *dev, pm_message_t state)
 }
 
 /**
+ * drm_class_suspend - internal DRM class suspend hook. Simply calls
+ * __drm_class_suspend() with the correct pm state.
+ * @dev: Linux device to suspend
+ */
+static int drm_class_suspend(struct device *dev)
+{
+	return __drm_class_suspend(dev, PMSG_SUSPEND);
+}
+
+/**
+ * drm_class_freeze - internal DRM class freeze hook. Simply calls
+ * __drm_class_suspend() with the correct pm state.
+ * @dev: Linux device to freeze
+ */
+static int drm_class_freeze(struct device *dev)
+{
+	return __drm_class_suspend(dev, PMSG_FREEZE);
+}
+
+/**
  * drm_class_resume - DRM class resume hook
  * @dev: Linux device to resume
  *
@@ -72,6 +92,12 @@ static int drm_class_resume(struct device *dev)
 	return 0;
 }
 
+static const struct dev_pm_ops drm_class_dev_pm_ops = {
+	.suspend	= drm_class_suspend,
+	.resume		= drm_class_resume,
+	.freeze		= drm_class_freeze,
+};
+
 static char *drm_devnode(struct device *dev, umode_t *mode)
 {
 	return kasprintf(GFP_KERNEL, "dri/%s", dev_name(dev));
@@ -106,8 +132,7 @@ struct class *drm_sysfs_create(struct module *owner, char *name)
 		goto err_out;
 	}
 
-	class->suspend = drm_class_suspend;
-	class->resume = drm_class_resume;
+	class->pm = &drm_class_dev_pm_ops;
 
 	err = class_create_file(class, &class_attr_version.attr);
 	if (err)
diff --git a/drivers/gpu/drm/drm_trace.h b/drivers/gpu/drm/drm_trace.h
index 03ea964..27cc95f 100644
--- a/drivers/gpu/drm/drm_trace.h
+++ b/drivers/gpu/drm/drm_trace.h
@@ -21,7 +21,7 @@ TRACE_EVENT(drm_vblank_event,
 		    __entry->crtc = crtc;
 		    __entry->seq = seq;
 		    ),
-	    TP_printk("crtc=%d, seq=%d", __entry->crtc, __entry->seq)
+	    TP_printk("crtc=%d, seq=%u", __entry->crtc, __entry->seq)
 );
 
 TRACE_EVENT(drm_vblank_event_queued,
@@ -37,7 +37,7 @@ TRACE_EVENT(drm_vblank_event_queued,
 		    __entry->crtc = crtc;
 		    __entry->seq = seq;
 		    ),
-	    TP_printk("pid=%d, crtc=%d, seq=%d", __entry->pid, __entry->crtc, \
+	    TP_printk("pid=%d, crtc=%d, seq=%u", __entry->pid, __entry->crtc, \
 		      __entry->seq)
 );
 
@@ -54,7 +54,7 @@ TRACE_EVENT(drm_vblank_event_delivered,
 		    __entry->crtc = crtc;
 		    __entry->seq = seq;
 		    ),
-	    TP_printk("pid=%d, crtc=%d, seq=%d", __entry->pid, __entry->crtc, \
+	    TP_printk("pid=%d, crtc=%d, seq=%u", __entry->pid, __entry->crtc, \
 		      __entry->seq)
 );
 
diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c
index 67969e2..feb2003 100644
--- a/drivers/gpu/drm/drm_vm.c
+++ b/drivers/gpu/drm/drm_vm.c
@@ -43,18 +43,19 @@
 static void drm_vm_open(struct vm_area_struct *vma);
 static void drm_vm_close(struct vm_area_struct *vma);
 
-static pgprot_t drm_io_prot(uint32_t map_type, struct vm_area_struct *vma)
+static pgprot_t drm_io_prot(struct drm_local_map *map,
+			    struct vm_area_struct *vma)
 {
 	pgprot_t tmp = vm_get_page_prot(vma->vm_flags);
 
 #if defined(__i386__) || defined(__x86_64__)
-	if (boot_cpu_data.x86 > 3 && map_type != _DRM_AGP) {
-		pgprot_val(tmp) |= _PAGE_PCD;
-		pgprot_val(tmp) &= ~_PAGE_PWT;
-	}
+	if (map->type == _DRM_REGISTERS && !(map->flags & _DRM_WRITE_COMBINING))
+		tmp = pgprot_noncached(tmp);
+	else
+		tmp = pgprot_writecombine(tmp);
 #elif defined(__powerpc__)
 	pgprot_val(tmp) |= _PAGE_NO_CACHE;
-	if (map_type == _DRM_REGISTERS)
+	if (map->type == _DRM_REGISTERS)
 		pgprot_val(tmp) |= _PAGE_GUARDED;
 #elif defined(__ia64__)
 	if (efi_range_is_wc(vma->vm_start, vma->vm_end -
@@ -250,13 +251,8 @@ static void drm_vm_shm_close(struct vm_area_struct *vma)
 			switch (map->type) {
 			case _DRM_REGISTERS:
 			case _DRM_FRAME_BUFFER:
-				if (drm_core_has_MTRR(dev) && map->mtrr >= 0) {
-					int retcode;
-					retcode = mtrr_del(map->mtrr,
-							   map->offset,
-							   map->size);
-					DRM_DEBUG("mtrr_del = %d\n", retcode);
-				}
+				if (drm_core_has_MTRR(dev))
+					arch_phys_wc_del(map->mtrr);
 				iounmap(map->handle);
 				break;
 			case _DRM_SHM:
@@ -617,7 +613,7 @@ int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 	case _DRM_FRAME_BUFFER:
 	case _DRM_REGISTERS:
 		offset = drm_core_get_reg_ofs(dev);
-		vma->vm_page_prot = drm_io_prot(map->type, vma);
+		vma->vm_page_prot = drm_io_prot(map, vma);
 		if (io_remap_pfn_range(vma, vma->vm_start,
 				       (map->offset + offset) >> PAGE_SHIFT,
 				       vma->vm_end - vma->vm_start,
diff --git a/drivers/gpu/drm/exynos/exynos_ddc.c b/drivers/gpu/drm/exynos/exynos_ddc.c
index 4e9b5ba..95c75ed 100644
--- a/drivers/gpu/drm/exynos/exynos_ddc.c
+++ b/drivers/gpu/drm/exynos/exynos_ddc.c
@@ -53,6 +53,8 @@ static struct of_device_id hdmiddc_match_types[] = {
 	{
 		.compatible = "samsung,exynos5-hdmiddc",
 	}, {
+		.compatible = "samsung,exynos4210-hdmiddc",
+	}, {
 		/* end node */
 	}
 };
diff --git a/drivers/gpu/drm/exynos/exynos_drm_buf.c b/drivers/gpu/drm/exynos/exynos_drm_buf.c
index 57affae..b8ac06d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_buf.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_buf.c
@@ -24,8 +24,6 @@ static int lowlevel_buffer_allocate(struct drm_device *dev,
 	enum dma_attr attr;
 	unsigned int nr_pages;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (buf->dma_addr) {
 		DRM_DEBUG_KMS("already allocated.\n");
 		return 0;
@@ -59,8 +57,7 @@ static int lowlevel_buffer_allocate(struct drm_device *dev,
 		dma_addr_t start_addr;
 		unsigned int i = 0;
 
-		buf->pages = kzalloc(sizeof(struct page) * nr_pages,
-					GFP_KERNEL);
+		buf->pages = drm_calloc_large(nr_pages, sizeof(struct page *));
 		if (!buf->pages) {
 			DRM_ERROR("failed to allocate pages.\n");
 			return -ENOMEM;
@@ -71,8 +68,8 @@ static int lowlevel_buffer_allocate(struct drm_device *dev,
 					&buf->dma_attrs);
 		if (!buf->kvaddr) {
 			DRM_ERROR("failed to allocate buffer.\n");
-			kfree(buf->pages);
-			return -ENOMEM;
+			ret = -ENOMEM;
+			goto err_free;
 		}
 
 		start_addr = buf->dma_addr;
@@ -109,9 +106,9 @@ err_free_attrs:
 	dma_free_attrs(dev->dev, buf->size, buf->pages,
 			(dma_addr_t)buf->dma_addr, &buf->dma_attrs);
 	buf->dma_addr = (dma_addr_t)NULL;
-
+err_free:
 	if (!is_drm_iommu_supported(dev))
-		kfree(buf->pages);
+		drm_free_large(buf->pages);
 
 	return ret;
 }
@@ -119,8 +116,6 @@ err_free_attrs:
 static void lowlevel_buffer_deallocate(struct drm_device *dev,
 		unsigned int flags, struct exynos_drm_gem_buf *buf)
 {
-	DRM_DEBUG_KMS("%s.\n", __FILE__);
-
 	if (!buf->dma_addr) {
 		DRM_DEBUG_KMS("dma_addr is invalid.\n");
 		return;
@@ -138,7 +133,7 @@ static void lowlevel_buffer_deallocate(struct drm_device *dev,
 	if (!is_drm_iommu_supported(dev)) {
 		dma_free_attrs(dev->dev, buf->size, buf->kvaddr,
 				(dma_addr_t)buf->dma_addr, &buf->dma_attrs);
-		kfree(buf->pages);
+		drm_free_large(buf->pages);
 	} else
 		dma_free_attrs(dev->dev, buf->size, buf->pages,
 				(dma_addr_t)buf->dma_addr, &buf->dma_attrs);
@@ -151,7 +146,6 @@ struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev,
 {
 	struct exynos_drm_gem_buf *buffer;
 
-	DRM_DEBUG_KMS("%s.\n", __FILE__);
 	DRM_DEBUG_KMS("desired size = 0x%x\n", size);
 
 	buffer = kzalloc(sizeof(*buffer), GFP_KERNEL);
@@ -167,8 +161,6 @@ struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev,
 void exynos_drm_fini_buf(struct drm_device *dev,
 				struct exynos_drm_gem_buf *buffer)
 {
-	DRM_DEBUG_KMS("%s.\n", __FILE__);
-
 	if (!buffer) {
 		DRM_DEBUG_KMS("buffer is null.\n");
 		return;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_connector.c b/drivers/gpu/drm/exynos/exynos_drm_connector.c
index 8bcc13a..02a8bc5 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_connector.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_connector.c
@@ -34,7 +34,6 @@ convert_to_display_mode(struct drm_display_mode *mode,
 			struct exynos_drm_panel_info *panel)
 {
 	struct fb_videomode *timing = &panel->timing;
-	DRM_DEBUG_KMS("%s\n", __FILE__);
 
 	mode->clock = timing->pixclock / 1000;
 	mode->vrefresh = timing->refresh;
@@ -58,37 +57,6 @@ convert_to_display_mode(struct drm_display_mode *mode,
 		mode->flags |= DRM_MODE_FLAG_DBLSCAN;
 }
 
-/* convert drm_display_mode to exynos_video_timings */
-static inline void
-convert_to_video_timing(struct fb_videomode *timing,
-			struct drm_display_mode *mode)
-{
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	memset(timing, 0, sizeof(*timing));
-
-	timing->pixclock = mode->clock * 1000;
-	timing->refresh = drm_mode_vrefresh(mode);
-
-	timing->xres = mode->hdisplay;
-	timing->right_margin = mode->hsync_start - mode->hdisplay;
-	timing->hsync_len = mode->hsync_end - mode->hsync_start;
-	timing->left_margin = mode->htotal - mode->hsync_end;
-
-	timing->yres = mode->vdisplay;
-	timing->lower_margin = mode->vsync_start - mode->vdisplay;
-	timing->vsync_len = mode->vsync_end - mode->vsync_start;
-	timing->upper_margin = mode->vtotal - mode->vsync_end;
-
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-		timing->vmode = FB_VMODE_INTERLACED;
-	else
-		timing->vmode = FB_VMODE_NONINTERLACED;
-
-	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
-		timing->vmode |= FB_VMODE_DOUBLE;
-}
-
 static int exynos_drm_connector_get_modes(struct drm_connector *connector)
 {
 	struct exynos_drm_connector *exynos_connector =
@@ -99,8 +67,6 @@ static int exynos_drm_connector_get_modes(struct drm_connector *connector)
 	unsigned int count = 0;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!display_ops) {
 		DRM_DEBUG_KMS("display_ops is null.\n");
 		return 0;
@@ -168,15 +134,12 @@ static int exynos_drm_connector_mode_valid(struct drm_connector *connector,
 					to_exynos_connector(connector);
 	struct exynos_drm_manager *manager = exynos_connector->manager;
 	struct exynos_drm_display_ops *display_ops = manager->display_ops;
-	struct fb_videomode timing;
 	int ret = MODE_BAD;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	convert_to_video_timing(&timing, mode);
-
-	if (display_ops && display_ops->check_timing)
-		if (!display_ops->check_timing(manager->dev, (void *)&timing))
+	if (display_ops && display_ops->check_mode)
+		if (!display_ops->check_mode(manager->dev, mode))
 			ret = MODE_OK;
 
 	return ret;
@@ -190,8 +153,6 @@ struct drm_encoder *exynos_drm_best_encoder(struct drm_connector *connector)
 	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	obj = drm_mode_object_find(dev, exynos_connector->encoder_id,
 				   DRM_MODE_OBJECT_ENCODER);
 	if (!obj) {
@@ -234,8 +195,6 @@ void exynos_drm_display_power(struct drm_connector *connector, int mode)
 static void exynos_drm_connector_dpms(struct drm_connector *connector,
 					int mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * in case that drm_crtc_helper_set_mode() is called,
 	 * encoder/crtc->funcs->dpms() will be just returned
@@ -282,8 +241,6 @@ exynos_drm_connector_detect(struct drm_connector *connector, bool force)
 					manager->display_ops;
 	enum drm_connector_status status = connector_status_disconnected;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (display_ops && display_ops->is_connected) {
 		if (display_ops->is_connected(manager->dev))
 			status = connector_status_connected;
@@ -299,8 +256,6 @@ static void exynos_drm_connector_destroy(struct drm_connector *connector)
 	struct exynos_drm_connector *exynos_connector =
 		to_exynos_connector(connector);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
 	kfree(exynos_connector);
@@ -322,8 +277,6 @@ struct drm_connector *exynos_drm_connector_create(struct drm_device *dev,
 	int type;
 	int err;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_connector = kzalloc(sizeof(*exynos_connector), GFP_KERNEL);
 	if (!exynos_connector) {
 		DRM_ERROR("failed to allocate connector\n");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_core.c b/drivers/gpu/drm/exynos/exynos_drm_core.c
index 4667c9f..1bef6dc 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_core.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_core.c
@@ -27,8 +27,6 @@ static int exynos_drm_create_enc_conn(struct drm_device *dev,
 	struct drm_connector *connector;
 	int ret;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	subdrv->manager->dev = subdrv->dev;
 
 	/* create and initialize a encoder for this sub driver. */
@@ -102,8 +100,6 @@ static int exynos_drm_subdrv_probe(struct drm_device *dev,
 static void exynos_drm_subdrv_remove(struct drm_device *dev,
 				      struct exynos_drm_subdrv *subdrv)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (subdrv->remove)
 		subdrv->remove(dev, subdrv->dev);
 }
@@ -114,8 +110,6 @@ int exynos_drm_device_register(struct drm_device *dev)
 	unsigned int fine_cnt = 0;
 	int err;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!dev)
 		return -EINVAL;
 
@@ -158,8 +152,6 @@ int exynos_drm_device_unregister(struct drm_device *dev)
 {
 	struct exynos_drm_subdrv *subdrv;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!dev) {
 		WARN(1, "Unexpected drm device unregister!\n");
 		return -EINVAL;
@@ -176,8 +168,6 @@ EXPORT_SYMBOL_GPL(exynos_drm_device_unregister);
 
 int exynos_drm_subdrv_register(struct exynos_drm_subdrv *subdrv)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!subdrv)
 		return -EINVAL;
 
@@ -189,8 +179,6 @@ EXPORT_SYMBOL_GPL(exynos_drm_subdrv_register);
 
 int exynos_drm_subdrv_unregister(struct exynos_drm_subdrv *subdrv)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!subdrv)
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.c b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
index c200e4d..9a35d17 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
@@ -76,8 +76,6 @@ static void exynos_drm_crtc_dpms(struct drm_crtc *crtc, int mode)
 
 static void exynos_drm_crtc_prepare(struct drm_crtc *crtc)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* drm framework doesn't check NULL. */
 }
 
@@ -85,8 +83,6 @@ static void exynos_drm_crtc_commit(struct drm_crtc *crtc)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
 	exynos_plane_commit(exynos_crtc->plane);
 	exynos_plane_dpms(exynos_crtc->plane, DRM_MODE_DPMS_ON);
@@ -97,8 +93,6 @@ exynos_drm_crtc_mode_fixup(struct drm_crtc *crtc,
 			    const struct drm_display_mode *mode,
 			    struct drm_display_mode *adjusted_mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* drm framework doesn't check NULL */
 	return true;
 }
@@ -115,8 +109,6 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	int pipe = exynos_crtc->pipe;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * copy the mode data adjusted by mode_fixup() into crtc->mode
 	 * so that hardware can be seet to proper mode.
@@ -139,7 +131,7 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	return 0;
 }
 
-static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+static int exynos_drm_crtc_mode_set_commit(struct drm_crtc *crtc, int x, int y,
 					  struct drm_framebuffer *old_fb)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
@@ -148,8 +140,6 @@ static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	unsigned int crtc_h;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* when framebuffer changing is requested, crtc's dpms should be on */
 	if (exynos_crtc->dpms > DRM_MODE_DPMS_ON) {
 		DRM_ERROR("failed framebuffer changing request.\n");
@@ -169,18 +159,16 @@ static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	return 0;
 }
 
-static void exynos_drm_crtc_load_lut(struct drm_crtc *crtc)
+static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+					  struct drm_framebuffer *old_fb)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-	/* drm framework doesn't check NULL */
+	return exynos_drm_crtc_mode_set_commit(crtc, x, y, old_fb);
 }
 
 static void exynos_drm_crtc_disable(struct drm_crtc *crtc)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_plane_dpms(exynos_crtc->plane, DRM_MODE_DPMS_OFF);
 	exynos_drm_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
 }
@@ -192,7 +180,6 @@ static struct drm_crtc_helper_funcs exynos_crtc_helper_funcs = {
 	.mode_fixup	= exynos_drm_crtc_mode_fixup,
 	.mode_set	= exynos_drm_crtc_mode_set,
 	.mode_set_base	= exynos_drm_crtc_mode_set_base,
-	.load_lut	= exynos_drm_crtc_load_lut,
 	.disable	= exynos_drm_crtc_disable,
 };
 
@@ -206,8 +193,6 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 	struct drm_framebuffer *old_fb = crtc->fb;
 	int ret = -EINVAL;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* when the page flip is requested, crtc's dpms should be on */
 	if (exynos_crtc->dpms > DRM_MODE_DPMS_ON) {
 		DRM_ERROR("failed page flip request.\n");
@@ -237,7 +222,7 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 		spin_unlock_irq(&dev->event_lock);
 
 		crtc->fb = fb;
-		ret = exynos_drm_crtc_mode_set_base(crtc, crtc->x, crtc->y,
+		ret = exynos_drm_crtc_mode_set_commit(crtc, crtc->x, crtc->y,
 						    NULL);
 		if (ret) {
 			crtc->fb = old_fb;
@@ -260,8 +245,6 @@ static void exynos_drm_crtc_destroy(struct drm_crtc *crtc)
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 	struct exynos_drm_private *private = crtc->dev->dev_private;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	private->crtc[exynos_crtc->pipe] = NULL;
 
 	drm_crtc_cleanup(crtc);
@@ -276,8 +259,6 @@ static int exynos_drm_crtc_set_property(struct drm_crtc *crtc,
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (property == dev_priv->crtc_mode_property) {
 		enum exynos_crtc_mode mode = val;
 
@@ -322,8 +303,6 @@ static void exynos_drm_crtc_attach_mode_property(struct drm_crtc *crtc)
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop = dev_priv->crtc_mode_property;
 	if (!prop) {
 		prop = drm_property_create_enum(dev, 0, "mode", mode_names,
@@ -343,8 +322,6 @@ int exynos_drm_crtc_create(struct drm_device *dev, unsigned int nr)
 	struct exynos_drm_private *private = dev->dev_private;
 	struct drm_crtc *crtc;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_crtc = kzalloc(sizeof(*exynos_crtc), GFP_KERNEL);
 	if (!exynos_crtc) {
 		DRM_ERROR("failed to allocate exynos crtc\n");
@@ -379,8 +356,6 @@ int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int crtc)
 	struct exynos_drm_crtc *exynos_crtc =
 		to_exynos_crtc(private->crtc[crtc]);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (exynos_crtc->dpms != DRM_MODE_DPMS_ON)
 		return -EPERM;
 
@@ -396,8 +371,6 @@ void exynos_drm_crtc_disable_vblank(struct drm_device *dev, int crtc)
 	struct exynos_drm_crtc *exynos_crtc =
 		to_exynos_crtc(private->crtc[crtc]);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (exynos_crtc->dpms != DRM_MODE_DPMS_ON)
 		return;
 
@@ -413,8 +386,6 @@ void exynos_drm_crtc_finish_pageflip(struct drm_device *dev, int crtc)
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(drm_crtc);
 	unsigned long flags;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	spin_lock_irqsave(&dev->event_lock, flags);
 
 	list_for_each_entry_safe(e, t, &dev_priv->pageflip_event_list,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c b/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
index ff7f2a8..a0f997e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
@@ -71,8 +71,6 @@ static struct sg_table *
 	unsigned int i;
 	int nents, ret;
 
-	DRM_DEBUG_PRIME("%s\n", __FILE__);
-
 	/* just return current sgt if already requested. */
 	if (exynos_attach->dir == dir && exynos_attach->is_mapped)
 		return &exynos_attach->sgt;
@@ -133,8 +131,6 @@ static void exynos_dmabuf_release(struct dma_buf *dmabuf)
 {
 	struct exynos_drm_gem_obj *exynos_gem_obj = dmabuf->priv;
 
-	DRM_DEBUG_PRIME("%s\n", __FILE__);
-
 	/*
 	 * exynos_dmabuf_release() call means that file object's
 	 * f_count is 0 and it calls drm_gem_object_handle_unreference()
@@ -219,8 +215,6 @@ struct drm_gem_object *exynos_dmabuf_prime_import(struct drm_device *drm_dev,
 	struct exynos_drm_gem_buf *buffer;
 	int ret;
 
-	DRM_DEBUG_PRIME("%s\n", __FILE__);
-
 	/* is this one of own objects? */
 	if (dma_buf->ops == &exynos_dmabuf_ops) {
 		struct drm_gem_object *obj;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index ba6d995..ca2729a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -46,8 +46,6 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 	int ret;
 	int nr;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	private = kzalloc(sizeof(struct exynos_drm_private), GFP_KERNEL);
 	if (!private) {
 		DRM_ERROR("failed to allocate private\n");
@@ -140,8 +138,6 @@ err_crtc:
 
 static int exynos_drm_unload(struct drm_device *dev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	exynos_drm_fbdev_fini(dev);
 	exynos_drm_device_unregister(dev);
 	drm_vblank_cleanup(dev);
@@ -159,8 +155,7 @@ static int exynos_drm_unload(struct drm_device *dev)
 static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv;
-
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
+	int ret;
 
 	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
 	if (!file_priv)
@@ -168,7 +163,13 @@ static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 
 	file->driver_priv = file_priv;
 
-	return exynos_drm_subdrv_open(dev, file);
+	ret = exynos_drm_subdrv_open(dev, file);
+	if (ret) {
+		kfree(file_priv);
+		file->driver_priv = NULL;
+	}
+
+	return ret;
 }
 
 static void exynos_drm_preclose(struct drm_device *dev,
@@ -178,8 +179,6 @@ static void exynos_drm_preclose(struct drm_device *dev,
 	struct drm_pending_vblank_event *e, *t;
 	unsigned long flags;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	/* release events of current file */
 	spin_lock_irqsave(&dev->event_lock, flags);
 	list_for_each_entry_safe(e, t, &private->pageflip_event_list,
@@ -196,8 +195,6 @@ static void exynos_drm_preclose(struct drm_device *dev,
 
 static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!file->driver_priv)
 		return;
 
@@ -207,8 +204,6 @@ static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 
 static void exynos_drm_lastclose(struct drm_device *dev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	exynos_drm_fbdev_restore_mode(dev);
 }
 
@@ -292,8 +287,6 @@ static struct drm_driver exynos_drm_driver = {
 
 static int exynos_drm_platform_probe(struct platform_device *pdev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
 	exynos_drm_driver.num_ioctls = DRM_ARRAY_SIZE(exynos_ioctls);
 
@@ -302,8 +295,6 @@ static int exynos_drm_platform_probe(struct platform_device *pdev)
 
 static int exynos_drm_platform_remove(struct platform_device *pdev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	drm_platform_exit(&exynos_drm_driver, pdev);
 
 	return 0;
@@ -322,8 +313,6 @@ static int __init exynos_drm_init(void)
 {
 	int ret;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 #ifdef CONFIG_DRM_EXYNOS_FIMD
 	ret = platform_driver_register(&fimd_driver);
 	if (ret < 0)
@@ -455,8 +444,6 @@ out_fimd:
 
 static void __exit exynos_drm_exit(void)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	platform_device_unregister(exynos_drm_pdev);
 
 	platform_driver_unregister(&exynos_drm_platform_driver);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index 680a7c1..eaa1966 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -142,7 +142,7 @@ struct exynos_drm_overlay {
  * @is_connected: check for that display is connected or not.
  * @get_edid: get edid modes from display driver.
  * @get_panel: get panel object from display driver.
- * @check_timing: check if timing is valid or not.
+ * @check_mode: check if mode is valid or not.
  * @power_on: display device on or off.
  */
 struct exynos_drm_display_ops {
@@ -151,7 +151,7 @@ struct exynos_drm_display_ops {
 	struct edid *(*get_edid)(struct device *dev,
 			struct drm_connector *connector);
 	void *(*get_panel)(struct device *dev);
-	int (*check_timing)(struct device *dev, void *timing);
+	int (*check_mode)(struct device *dev, struct drm_display_mode *mode);
 	int (*power_on)(struct device *dev, int mode);
 };
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.c b/drivers/gpu/drm/exynos/exynos_drm_encoder.c
index c63721f..a99a033 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_encoder.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.c
@@ -61,7 +61,7 @@ static void exynos_drm_encoder_dpms(struct drm_encoder *encoder, int mode)
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 	struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder);
 
-	DRM_DEBUG_KMS("%s, encoder dpms: %d\n", __FILE__, mode);
+	DRM_DEBUG_KMS("encoder dpms: %d\n", mode);
 
 	if (exynos_encoder->dpms == mode) {
 		DRM_DEBUG_KMS("desired dpms mode is same as previous one.\n");
@@ -104,8 +104,6 @@ exynos_drm_encoder_mode_fixup(struct drm_encoder *encoder,
 	struct exynos_drm_manager *manager = exynos_drm_get_manager(encoder);
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		if (connector->encoder == encoder)
 			if (manager_ops && manager_ops->mode_fixup)
@@ -155,8 +153,6 @@ static void exynos_drm_encoder_mode_set(struct drm_encoder *encoder,
 	struct exynos_drm_manager *manager;
 	struct exynos_drm_manager_ops *manager_ops;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		if (connector->encoder == encoder) {
 			struct exynos_drm_encoder *exynos_encoder;
@@ -189,8 +185,6 @@ static void exynos_drm_encoder_mode_set(struct drm_encoder *encoder,
 
 static void exynos_drm_encoder_prepare(struct drm_encoder *encoder)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* drm framework doesn't check NULL. */
 }
 
@@ -200,8 +194,6 @@ static void exynos_drm_encoder_commit(struct drm_encoder *encoder)
 	struct exynos_drm_manager *manager = exynos_encoder->manager;
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (manager_ops && manager_ops->commit)
 		manager_ops->commit(manager->dev);
 
@@ -274,8 +266,6 @@ static void exynos_drm_encoder_destroy(struct drm_encoder *encoder)
 	struct exynos_drm_encoder *exynos_encoder =
 		to_exynos_encoder(encoder);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_encoder->manager->pipe = -1;
 
 	drm_encoder_cleanup(encoder);
@@ -315,8 +305,6 @@ void exynos_drm_encoder_setup(struct drm_device *dev)
 {
 	struct drm_encoder *encoder;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head)
 		encoder->possible_clones = exynos_drm_encoder_clones(encoder);
 }
@@ -329,8 +317,6 @@ exynos_drm_encoder_create(struct drm_device *dev,
 	struct drm_encoder *encoder;
 	struct exynos_drm_encoder *exynos_encoder;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!manager || !possible_crtcs)
 		return NULL;
 
@@ -427,8 +413,6 @@ void exynos_drm_encoder_crtc_dpms(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 	int mode = *(int *)data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (manager_ops && manager_ops->dpms)
 		manager_ops->dpms(manager->dev, mode);
 
@@ -449,8 +433,6 @@ void exynos_drm_encoder_crtc_pipe(struct drm_encoder *encoder, void *data)
 		to_exynos_encoder(encoder)->manager;
 	int pipe = *(int *)data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * when crtc is detached from encoder, this pipe is used
 	 * to select manager operation
@@ -465,8 +447,6 @@ void exynos_drm_encoder_plane_mode_set(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	struct exynos_drm_overlay *overlay = data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (overlay_ops && overlay_ops->mode_set)
 		overlay_ops->mode_set(manager->dev, overlay);
 }
@@ -478,8 +458,6 @@ void exynos_drm_encoder_plane_commit(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	int zpos = DEFAULT_ZPOS;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (data)
 		zpos = *(int *)data;
 
@@ -494,8 +472,6 @@ void exynos_drm_encoder_plane_enable(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	int zpos = DEFAULT_ZPOS;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (data)
 		zpos = *(int *)data;
 
@@ -510,8 +486,6 @@ void exynos_drm_encoder_plane_disable(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	int zpos = DEFAULT_ZPOS;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (data)
 		zpos = *(int *)data;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 0e04f4e..c2d149f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -70,8 +70,6 @@ static void exynos_drm_fb_destroy(struct drm_framebuffer *fb)
 	struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
 	unsigned int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* make sure that overlay data are updated before relesing fb. */
 	exynos_drm_encoder_complete_scanout(fb);
 
@@ -97,8 +95,6 @@ static int exynos_drm_fb_create_handle(struct drm_framebuffer *fb,
 {
 	struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* This fb should have only one gem object. */
 	if (WARN_ON(exynos_fb->buf_cnt != 1))
 		return -EINVAL;
@@ -112,8 +108,6 @@ static int exynos_drm_fb_dirty(struct drm_framebuffer *fb,
 				unsigned color, struct drm_clip_rect *clips,
 				unsigned num_clips)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO */
 
 	return 0;
@@ -225,8 +219,6 @@ exynos_user_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 	struct exynos_drm_fb *exynos_fb;
 	int i, ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL);
 	if (!exynos_fb) {
 		DRM_ERROR("failed to allocate exynos drm framebuffer\n");
@@ -293,8 +285,6 @@ struct exynos_drm_gem_buf *exynos_drm_fb_buffer(struct drm_framebuffer *fb,
 	struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
 	struct exynos_drm_gem_buf *buffer;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (index >= MAX_FB_BUFFER)
 		return NULL;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 8f007aa..8e60bd6 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -43,8 +43,6 @@ static int exynos_drm_fb_mmap(struct fb_info *info,
 	unsigned long vm_size;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
 
 	vm_size = vma->vm_end - vma->vm_start;
@@ -84,8 +82,6 @@ static int exynos_drm_fbdev_update(struct drm_fb_helper *helper,
 	unsigned int size = fb->width * fb->height * (fb->bits_per_pixel >> 3);
 	unsigned long offset;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth);
 	drm_fb_helper_fill_var(fbi, helper, fb->width, fb->height);
 
@@ -148,8 +144,6 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper,
 	unsigned long size;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	DRM_DEBUG_KMS("surface width(%d), height(%d) and bpp(%d\n",
 			sizes->surface_width, sizes->surface_height,
 			sizes->surface_bpp);
@@ -238,8 +232,6 @@ int exynos_drm_fbdev_init(struct drm_device *dev)
 	unsigned int num_crtc;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!dev->mode_config.num_crtc || !dev->mode_config.num_connector)
 		return 0;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
index 4a1616a..61b094f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
@@ -175,8 +175,6 @@ static void fimc_sw_reset(struct fimc_context *ctx)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* stop dma operation */
 	cfg = fimc_read(EXYNOS_CISTATUS);
 	if (EXYNOS_CISTATUS_GET_ENVID_STATUS(cfg)) {
@@ -210,8 +208,6 @@ static void fimc_sw_reset(struct fimc_context *ctx)
 
 static int fimc_set_camblk_fimd0_wb(struct fimc_context *ctx)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return regmap_update_bits(ctx->sysreg, SYSREG_CAMERA_BLK,
 				  SYSREG_FIMD0WB_DEST_MASK,
 				  ctx->id << SYSREG_FIMD0WB_DEST_SHIFT);
@@ -221,7 +217,7 @@ static void fimc_set_type_ctrl(struct fimc_context *ctx, enum fimc_wb wb)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:wb[%d]\n", __func__, wb);
+	DRM_DEBUG_KMS("wb[%d]\n", wb);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	cfg &= ~(EXYNOS_CIGCTRL_TESTPATTERN_MASK |
@@ -257,10 +253,10 @@ static void fimc_set_polarity(struct fimc_context *ctx,
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:inv_pclk[%d]inv_vsync[%d]\n",
-		__func__, pol->inv_pclk, pol->inv_vsync);
-	DRM_DEBUG_KMS("%s:inv_href[%d]inv_hsync[%d]\n",
-		__func__, pol->inv_href, pol->inv_hsync);
+	DRM_DEBUG_KMS("inv_pclk[%d]inv_vsync[%d]\n",
+		pol->inv_pclk, pol->inv_vsync);
+	DRM_DEBUG_KMS("inv_href[%d]inv_hsync[%d]\n",
+		pol->inv_href, pol->inv_hsync);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	cfg &= ~(EXYNOS_CIGCTRL_INVPOLPCLK | EXYNOS_CIGCTRL_INVPOLVSYNC |
@@ -282,7 +278,7 @@ static void fimc_handle_jpeg(struct fimc_context *ctx, bool enable)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	if (enable)
@@ -298,7 +294,7 @@ static void fimc_handle_irq(struct fimc_context *ctx, bool enable,
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]overflow[%d]level[%d]\n", __func__,
+	DRM_DEBUG_KMS("enable[%d]overflow[%d]level[%d]\n",
 			enable, overflow, level);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
@@ -319,8 +315,6 @@ static void fimc_clear_irq(struct fimc_context *ctx)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	cfg |= EXYNOS_CIGCTRL_IRQ_CLR;
 	fimc_write(cfg, EXYNOS_CIGCTRL);
@@ -335,7 +329,7 @@ static bool fimc_check_ovf(struct fimc_context *ctx)
 	flag = EXYNOS_CISTATUS_OVFIY | EXYNOS_CISTATUS_OVFICB |
 		EXYNOS_CISTATUS_OVFICR;
 
-	DRM_DEBUG_KMS("%s:flag[0x%x]\n", __func__, flag);
+	DRM_DEBUG_KMS("flag[0x%x]\n", flag);
 
 	if (status & flag) {
 		cfg = fimc_read(EXYNOS_CIWDOFST);
@@ -364,7 +358,7 @@ static bool fimc_check_frame_end(struct fimc_context *ctx)
 
 	cfg = fimc_read(EXYNOS_CISTATUS);
 
-	DRM_DEBUG_KMS("%s:cfg[0x%x]\n", __func__, cfg);
+	DRM_DEBUG_KMS("cfg[0x%x]\n", cfg);
 
 	if (!(cfg & EXYNOS_CISTATUS_FRAMEEND))
 		return false;
@@ -380,15 +374,13 @@ static int fimc_get_buf_id(struct fimc_context *ctx)
 	u32 cfg;
 	int frame_cnt, buf_id;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	cfg = fimc_read(EXYNOS_CISTATUS2);
 	frame_cnt = EXYNOS_CISTATUS2_GET_FRAMECOUNT_BEFORE(cfg);
 
 	if (frame_cnt == 0)
 		frame_cnt = EXYNOS_CISTATUS2_GET_FRAMECOUNT_PRESENT(cfg);
 
-	DRM_DEBUG_KMS("%s:present[%d]before[%d]\n", __func__,
+	DRM_DEBUG_KMS("present[%d]before[%d]\n",
 		EXYNOS_CISTATUS2_GET_FRAMECOUNT_PRESENT(cfg),
 		EXYNOS_CISTATUS2_GET_FRAMECOUNT_BEFORE(cfg));
 
@@ -398,7 +390,7 @@ static int fimc_get_buf_id(struct fimc_context *ctx)
 	}
 
 	buf_id = frame_cnt - 1;
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__, buf_id);
+	DRM_DEBUG_KMS("buf_id[%d]\n", buf_id);
 
 	return buf_id;
 }
@@ -407,7 +399,7 @@ static void fimc_handle_lastend(struct fimc_context *ctx, bool enable)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	cfg = fimc_read(EXYNOS_CIOCTRL);
 	if (enable)
@@ -424,7 +416,7 @@ static int fimc_src_set_fmt_order(struct fimc_context *ctx, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	/* RGB */
 	cfg = fimc_read(EXYNOS_CISCCTRL);
@@ -497,7 +489,7 @@ static int fimc_src_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = fimc_read(EXYNOS_MSCTRL);
 	cfg &= ~EXYNOS_MSCTRL_INFORMAT_RGB;
@@ -557,8 +549,7 @@ static int fimc_src_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg1, cfg2;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg1 = fimc_read(EXYNOS_MSCTRL);
 	cfg1 &= ~(EXYNOS_MSCTRL_FLIP_X_MIRROR |
@@ -621,10 +612,9 @@ static int fimc_set_window(struct fimc_context *ctx,
 	v1 = pos->y;
 	v2 = sz->vsize - pos->h - pos->y;
 
-	DRM_DEBUG_KMS("%s:x[%d]y[%d]w[%d]h[%d]hsize[%d]vsize[%d]\n",
-	__func__, pos->x, pos->y, pos->w, pos->h, sz->hsize, sz->vsize);
-	DRM_DEBUG_KMS("%s:h1[%d]h2[%d]v1[%d]v2[%d]\n", __func__,
-		h1, h2, v1, v2);
+	DRM_DEBUG_KMS("x[%d]y[%d]w[%d]h[%d]hsize[%d]vsize[%d]\n",
+		pos->x, pos->y, pos->w, pos->h, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("h1[%d]h2[%d]v1[%d]v2[%d]\n", h1, h2, v1, v2);
 
 	/*
 	 * set window offset 1, 2 size
@@ -653,8 +643,8 @@ static int fimc_src_set_size(struct device *dev, int swap,
 	struct drm_exynos_sz img_sz = *sz;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]hsize[%d]vsize[%d]\n",
-		__func__, swap, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("swap[%d]hsize[%d]vsize[%d]\n",
+		swap, sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = (EXYNOS_ORGISIZE_HORIZONTAL(img_sz.hsize) |
@@ -662,8 +652,7 @@ static int fimc_src_set_size(struct device *dev, int swap,
 
 	fimc_write(cfg, EXYNOS_ORGISIZE);
 
-	DRM_DEBUG_KMS("%s:x[%d]y[%d]w[%d]h[%d]\n", __func__,
-		pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("x[%d]y[%d]w[%d]h[%d]\n", pos->x, pos->y, pos->w, pos->h);
 
 	if (swap) {
 		img_pos.w = pos->h;
@@ -720,7 +709,7 @@ static int fimc_src_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > FIMC_MAX_SRC) {
@@ -772,7 +761,7 @@ static int fimc_dst_set_fmt_order(struct fimc_context *ctx, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	/* RGB */
 	cfg = fimc_read(EXYNOS_CISCCTRL);
@@ -851,7 +840,7 @@ static int fimc_dst_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = fimc_read(EXYNOS_CIEXTEN);
 
@@ -919,8 +908,7 @@ static int fimc_dst_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg = fimc_read(EXYNOS_CITRGFMT);
 	cfg &= ~EXYNOS_CITRGFMT_FLIP_MASK;
@@ -970,7 +958,7 @@ static int fimc_dst_set_transf(struct device *dev,
 
 static int fimc_get_ratio_shift(u32 src, u32 dst, u32 *ratio, u32 *shift)
 {
-	DRM_DEBUG_KMS("%s:src[%d]dst[%d]\n", __func__, src, dst);
+	DRM_DEBUG_KMS("src[%d]dst[%d]\n", src, dst);
 
 	if (src >= dst * 64) {
 		DRM_ERROR("failed to make ratio and shift.\n");
@@ -1039,20 +1027,20 @@ static int fimc_set_prescaler(struct fimc_context *ctx, struct fimc_scaler *sc,
 
 	pre_dst_width = src_w / pre_hratio;
 	pre_dst_height = src_h / pre_vratio;
-	DRM_DEBUG_KMS("%s:pre_dst_width[%d]pre_dst_height[%d]\n", __func__,
+	DRM_DEBUG_KMS("pre_dst_width[%d]pre_dst_height[%d]\n",
 		pre_dst_width, pre_dst_height);
-	DRM_DEBUG_KMS("%s:pre_hratio[%d]hfactor[%d]pre_vratio[%d]vfactor[%d]\n",
-		__func__, pre_hratio, hfactor, pre_vratio, vfactor);
+	DRM_DEBUG_KMS("pre_hratio[%d]hfactor[%d]pre_vratio[%d]vfactor[%d]\n",
+		pre_hratio, hfactor, pre_vratio, vfactor);
 
 	sc->hratio = (src_w << 14) / (dst_w << hfactor);
 	sc->vratio = (src_h << 14) / (dst_h << vfactor);
 	sc->up_h = (dst_w >= src_w) ? true : false;
 	sc->up_v = (dst_h >= src_h) ? true : false;
-	DRM_DEBUG_KMS("%s:hratio[%d]vratio[%d]up_h[%d]up_v[%d]\n",
-	__func__, sc->hratio, sc->vratio, sc->up_h, sc->up_v);
+	DRM_DEBUG_KMS("hratio[%d]vratio[%d]up_h[%d]up_v[%d]\n",
+		sc->hratio, sc->vratio, sc->up_h, sc->up_v);
 
 	shfactor = FIMC_SHFACTOR - (hfactor + vfactor);
-	DRM_DEBUG_KMS("%s:shfactor[%d]\n", __func__, shfactor);
+	DRM_DEBUG_KMS("shfactor[%d]\n", shfactor);
 
 	cfg = (EXYNOS_CISCPRERATIO_SHFACTOR(shfactor) |
 		EXYNOS_CISCPRERATIO_PREHORRATIO(pre_hratio) |
@@ -1070,10 +1058,10 @@ static void fimc_set_scaler(struct fimc_context *ctx, struct fimc_scaler *sc)
 {
 	u32 cfg, cfg_ext;
 
-	DRM_DEBUG_KMS("%s:range[%d]bypass[%d]up_h[%d]up_v[%d]\n",
-		__func__, sc->range, sc->bypass, sc->up_h, sc->up_v);
-	DRM_DEBUG_KMS("%s:hratio[%d]vratio[%d]\n",
-		__func__, sc->hratio, sc->vratio);
+	DRM_DEBUG_KMS("range[%d]bypass[%d]up_h[%d]up_v[%d]\n",
+		sc->range, sc->bypass, sc->up_h, sc->up_v);
+	DRM_DEBUG_KMS("hratio[%d]vratio[%d]\n",
+		sc->hratio, sc->vratio);
 
 	cfg = fimc_read(EXYNOS_CISCCTRL);
 	cfg &= ~(EXYNOS_CISCCTRL_SCALERBYPASS |
@@ -1113,8 +1101,8 @@ static int fimc_dst_set_size(struct device *dev, int swap,
 	struct drm_exynos_sz img_sz = *sz;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]hsize[%d]vsize[%d]\n",
-		__func__, swap, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("swap[%d]hsize[%d]vsize[%d]\n",
+		swap, sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = (EXYNOS_ORGOSIZE_HORIZONTAL(img_sz.hsize) |
@@ -1122,8 +1110,7 @@ static int fimc_dst_set_size(struct device *dev, int swap,
 
 	fimc_write(cfg, EXYNOS_ORGOSIZE);
 
-	DRM_DEBUG_KMS("%s:x[%d]y[%d]w[%d]h[%d]\n",
-		__func__, pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("x[%d]y[%d]w[%d]h[%d]\n", pos->x, pos->y, pos->w, pos->h);
 
 	/* CSC ITU */
 	cfg = fimc_read(EXYNOS_CIGCTRL);
@@ -1180,7 +1167,7 @@ static int fimc_dst_get_buf_seq(struct fimc_context *ctx)
 		if (cfg & (mask << i))
 			buf_num++;
 
-	DRM_DEBUG_KMS("%s:buf_num[%d]\n", __func__, buf_num);
+	DRM_DEBUG_KMS("buf_num[%d]\n", buf_num);
 
 	return buf_num;
 }
@@ -1194,8 +1181,7 @@ static int fimc_dst_set_buf_seq(struct fimc_context *ctx, u32 buf_id,
 	u32 mask = 0x00000001 << buf_id;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]buf_type[%d]\n", __func__,
-		buf_id, buf_type);
+	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	mutex_lock(&ctx->lock);
 
@@ -1252,7 +1238,7 @@ static int fimc_dst_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > FIMC_MAX_DST) {
@@ -1302,7 +1288,7 @@ static struct exynos_drm_ipp_ops fimc_dst_ops = {
 
 static int fimc_clk_ctrl(struct fimc_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	if (enable) {
 		clk_prepare_enable(ctx->clocks[FIMC_CLK_GATE]);
@@ -1326,7 +1312,7 @@ static irqreturn_t fimc_irq_handler(int irq, void *dev_id)
 		c_node->event_work;
 	int buf_id;
 
-	DRM_DEBUG_KMS("%s:fimc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("fimc id[%d]\n", ctx->id);
 
 	fimc_clear_irq(ctx);
 	if (fimc_check_ovf(ctx))
@@ -1339,7 +1325,7 @@ static irqreturn_t fimc_irq_handler(int irq, void *dev_id)
 	if (buf_id < 0)
 		return IRQ_HANDLED;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__, buf_id);
+	DRM_DEBUG_KMS("buf_id[%d]\n", buf_id);
 
 	if (fimc_dst_set_buf_seq(ctx, buf_id, IPP_BUF_DEQUEUE) < 0) {
 		DRM_ERROR("failed to dequeue.\n");
@@ -1357,8 +1343,6 @@ static int fimc_init_prop_list(struct exynos_drm_ippdrv *ippdrv)
 {
 	struct drm_exynos_ipp_prop_list *prop_list;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop_list = devm_kzalloc(ippdrv->dev, sizeof(*prop_list), GFP_KERNEL);
 	if (!prop_list) {
 		DRM_ERROR("failed to alloc property list.\n");
@@ -1402,7 +1386,7 @@ static inline bool fimc_check_drm_flip(enum drm_exynos_flip flip)
 	case EXYNOS_DRM_FLIP_BOTH:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return false;
 	}
 }
@@ -1419,8 +1403,6 @@ static int fimc_ippdrv_check_property(struct device *dev,
 	bool swap;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	for_each_ipp_ops(i) {
 		if ((i == EXYNOS_DRM_OPS_SRC) &&
 			(property->cmd == IPP_CMD_WB))
@@ -1526,8 +1508,6 @@ static void fimc_clear_addr(struct fimc_context *ctx)
 {
 	int i;
 
-	DRM_DEBUG_KMS("%s:\n", __func__);
-
 	for (i = 0; i < FIMC_MAX_SRC; i++) {
 		fimc_write(0, EXYNOS_CIIYSA(i));
 		fimc_write(0, EXYNOS_CIICBSA(i));
@@ -1545,8 +1525,6 @@ static int fimc_ippdrv_reset(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* reset h/w block */
 	fimc_sw_reset(ctx);
 
@@ -1570,7 +1548,7 @@ static int fimc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	int ret, i;
 	u32 cfg0, cfg1;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	if (!c_node) {
 		DRM_ERROR("failed to get c_node.\n");
@@ -1679,7 +1657,7 @@ static void fimc_ippdrv_stop(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	struct drm_exynos_ipp_set_wb set_wb = {0, 0};
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	switch (cmd) {
 	case IPP_CMD_M2M:
@@ -1869,8 +1847,7 @@ static int fimc_probe(struct platform_device *pdev)
 		goto err_put_clk;
 	}
 
-	DRM_DEBUG_KMS("%s:id[%d]ippdrv[0x%x]\n", __func__, ctx->id,
-		(int)ippdrv);
+	DRM_DEBUG_KMS("id[%d]ippdrv[0x%x]\n", ctx->id, (int)ippdrv);
 
 	mutex_init(&ctx->lock);
 	platform_set_drvdata(pdev, ctx);
@@ -1917,7 +1894,7 @@ static int fimc_suspend(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (pm_runtime_suspended(dev))
 		return 0;
@@ -1929,7 +1906,7 @@ static int fimc_resume(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (!pm_runtime_suspended(dev))
 		return fimc_clk_ctrl(ctx, true);
@@ -1943,7 +1920,7 @@ static int fimc_runtime_suspend(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  fimc_clk_ctrl(ctx, false);
 }
@@ -1952,7 +1929,7 @@ static int fimc_runtime_resume(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  fimc_clk_ctrl(ctx, true);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 97c61db..3e106be 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -63,14 +63,24 @@
 
 struct fimd_driver_data {
 	unsigned int timing_base;
+
+	unsigned int has_shadowcon:1;
+	unsigned int has_clksel:1;
+};
+
+static struct fimd_driver_data s3c64xx_fimd_driver_data = {
+	.timing_base = 0x0,
+	.has_clksel = 1,
 };
 
 static struct fimd_driver_data exynos4_fimd_driver_data = {
 	.timing_base = 0x0,
+	.has_shadowcon = 1,
 };
 
 static struct fimd_driver_data exynos5_fimd_driver_data = {
 	.timing_base = 0x20000,
+	.has_shadowcon = 1,
 };
 
 struct fimd_win_data {
@@ -107,10 +117,13 @@ struct fimd_context {
 	atomic_t			wait_vsync_event;
 
 	struct exynos_drm_panel_info *panel;
+	struct fimd_driver_data *driver_data;
 };
 
 #ifdef CONFIG_OF
 static const struct of_device_id fimd_driver_dt_match[] = {
+	{ .compatible = "samsung,s3c6400-fimd",
+	  .data = &s3c64xx_fimd_driver_data },
 	{ .compatible = "samsung,exynos4210-fimd",
 	  .data = &exynos4_fimd_driver_data },
 	{ .compatible = "samsung,exynos5250-fimd",
@@ -137,8 +150,6 @@ static inline struct fimd_driver_data *drm_fimd_get_driver_data(
 
 static bool fimd_display_is_connected(struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return true;
@@ -148,15 +159,11 @@ static void *fimd_get_panel(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return ctx->panel;
 }
 
-static int fimd_check_timing(struct device *dev, void *timing)
+static int fimd_check_mode(struct device *dev, struct drm_display_mode *mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return 0;
@@ -164,8 +171,6 @@ static int fimd_check_timing(struct device *dev, void *timing)
 
 static int fimd_display_power_on(struct device *dev, int mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO */
 
 	return 0;
@@ -175,7 +180,7 @@ static struct exynos_drm_display_ops fimd_display_ops = {
 	.type = EXYNOS_DISPLAY_TYPE_LCD,
 	.is_connected = fimd_display_is_connected,
 	.get_panel = fimd_get_panel,
-	.check_timing = fimd_check_timing,
+	.check_mode = fimd_check_mode,
 	.power_on = fimd_display_power_on,
 };
 
@@ -183,7 +188,7 @@ static void fimd_dpms(struct device *subdrv_dev, int mode)
 {
 	struct fimd_context *ctx = get_fimd_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode);
+	DRM_DEBUG_KMS("%d\n", mode);
 
 	mutex_lock(&ctx->lock);
 
@@ -221,8 +226,6 @@ static void fimd_apply(struct device *subdrv_dev)
 	struct fimd_win_data *win_data;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	for (i = 0; i < WINDOWS_NR; i++) {
 		win_data = &ctx->win_data[i];
 		if (win_data->enabled && (ovl_ops && ovl_ops->commit))
@@ -239,15 +242,12 @@ static void fimd_commit(struct device *dev)
 	struct exynos_drm_panel_info *panel = ctx->panel;
 	struct fb_videomode *timing = &panel->timing;
 	struct fimd_driver_data *driver_data;
-	struct platform_device *pdev = to_platform_device(dev);
 	u32 val;
 
-	driver_data = drm_fimd_get_driver_data(pdev);
+	driver_data = ctx->driver_data;
 	if (ctx->suspended)
 		return;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* setup polarity values from machine code. */
 	writel(ctx->vidcon1, ctx->regs + driver_data->timing_base + VIDCON1);
 
@@ -274,6 +274,11 @@ static void fimd_commit(struct device *dev)
 	val = ctx->vidcon0;
 	val &= ~(VIDCON0_CLKVAL_F_MASK | VIDCON0_CLKDIR);
 
+	if (ctx->driver_data->has_clksel) {
+		val &= ~VIDCON0_CLKSEL_MASK;
+		val |= VIDCON0_CLKSEL_LCD;
+	}
+
 	if (ctx->clkdiv > 1)
 		val |= VIDCON0_CLKVAL_F(ctx->clkdiv - 1) | VIDCON0_CLKDIR;
 	else
@@ -292,8 +297,6 @@ static int fimd_enable_vblank(struct device *dev)
 	struct fimd_context *ctx = get_fimd_context(dev);
 	u32 val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return -EPERM;
 
@@ -319,8 +322,6 @@ static void fimd_disable_vblank(struct device *dev)
 	struct fimd_context *ctx = get_fimd_context(dev);
 	u32 val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
@@ -370,8 +371,6 @@ static void fimd_win_mode_set(struct device *dev,
 	int win;
 	unsigned long offset;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!overlay) {
 		dev_err(dev, "overlay is NULL\n");
 		return;
@@ -381,7 +380,7 @@ static void fimd_win_mode_set(struct device *dev,
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	offset = overlay->fb_x * (overlay->bpp >> 3);
@@ -418,8 +417,6 @@ static void fimd_win_set_pixfmt(struct device *dev, unsigned int win)
 	struct fimd_win_data *win_data = &ctx->win_data[win];
 	unsigned long val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	val = WINCONx_ENWIN;
 
 	switch (win_data->bpp) {
@@ -478,8 +475,6 @@ static void fimd_win_set_colkey(struct device *dev, unsigned int win)
 	struct fimd_context *ctx = get_fimd_context(dev);
 	unsigned int keycon0 = 0, keycon1 = 0;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	keycon0 = ~(WxKEYCON0_KEYBL_EN | WxKEYCON0_KEYEN_F |
 			WxKEYCON0_DIRCON) | WxKEYCON0_COMPKEY(0);
 
@@ -489,6 +484,33 @@ static void fimd_win_set_colkey(struct device *dev, unsigned int win)
 	writel(keycon1, ctx->regs + WKEYCON1_BASE(win));
 }
 
+/**
+ * shadow_protect_win() - disable updating values from shadow registers at vsync
+ *
+ * @win: window to protect registers for
+ * @protect: 1 to protect (disable updates)
+ */
+static void fimd_shadow_protect_win(struct fimd_context *ctx,
+							int win, bool protect)
+{
+	u32 reg, bits, val;
+
+	if (ctx->driver_data->has_shadowcon) {
+		reg = SHADOWCON;
+		bits = SHADOWCON_WINx_PROTECT(win);
+	} else {
+		reg = PRTCON;
+		bits = PRTCON_PROTECT;
+	}
+
+	val = readl(ctx->regs + reg);
+	if (protect)
+		val |= bits;
+	else
+		val &= ~bits;
+	writel(val, ctx->regs + reg);
+}
+
 static void fimd_win_commit(struct device *dev, int zpos)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
@@ -498,21 +520,19 @@ static void fimd_win_commit(struct device *dev, int zpos)
 	unsigned int last_x;
 	unsigned int last_y;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
 
 	/*
-	 * SHADOWCON register is used for enabling timing.
+	 * SHADOWCON/PRTCON register is used for enabling timing.
 	 *
 	 * for example, once only width value of a register is set,
 	 * if the dma is started then fimd hardware could malfunction so
@@ -522,9 +542,7 @@ static void fimd_win_commit(struct device *dev, int zpos)
 	 */
 
 	/* protect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val |= SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, true);
 
 	/* buffer start address */
 	val = (unsigned long)win_data->dma_addr;
@@ -602,10 +620,13 @@ static void fimd_win_commit(struct device *dev, int zpos)
 	writel(val, ctx->regs + WINCON(win));
 
 	/* Enable DMA channel and unprotect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val |= SHADOWCON_CHx_ENABLE(win);
-	val &= ~SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, false);
+
+	if (ctx->driver_data->has_shadowcon) {
+		val = readl(ctx->regs + SHADOWCON);
+		val |= SHADOWCON_CHx_ENABLE(win);
+		writel(val, ctx->regs + SHADOWCON);
+	}
 
 	win_data->enabled = true;
 }
@@ -617,12 +638,10 @@ static void fimd_win_disable(struct device *dev, int zpos)
 	int win = zpos;
 	u32 val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
@@ -634,9 +653,7 @@ static void fimd_win_disable(struct device *dev, int zpos)
 	}
 
 	/* protect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val |= SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, true);
 
 	/* wincon */
 	val = readl(ctx->regs + WINCON(win));
@@ -644,10 +661,13 @@ static void fimd_win_disable(struct device *dev, int zpos)
 	writel(val, ctx->regs + WINCON(win));
 
 	/* unprotect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val &= ~SHADOWCON_CHx_ENABLE(win);
-	val &= ~SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	if (ctx->driver_data->has_shadowcon) {
+		val = readl(ctx->regs + SHADOWCON);
+		val &= ~SHADOWCON_CHx_ENABLE(win);
+		writel(val, ctx->regs + SHADOWCON);
+	}
+
+	fimd_shadow_protect_win(ctx, win, false);
 
 	win_data->enabled = false;
 }
@@ -697,8 +717,6 @@ out:
 
 static int fimd_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * enable drm irq mode.
 	 * - with irq_enabled = 1, we can use the vblank feature.
@@ -725,8 +743,6 @@ static int fimd_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 
 static void fimd_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* detach this sub driver from iommu mapping if supported. */
 	if (is_drm_iommu_supported(drm_dev))
 		drm_iommu_detach_device(drm_dev, dev);
@@ -741,8 +757,6 @@ static int fimd_calc_clkdiv(struct fimd_context *ctx,
 	u32 best_framerate = 0;
 	u32 framerate;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	retrace = timing->left_margin + timing->hsync_len +
 				timing->right_margin + timing->xres;
 	retrace *= timing->upper_margin + timing->vsync_len +
@@ -777,10 +791,6 @@ static int fimd_calc_clkdiv(struct fimd_context *ctx,
 
 static void fimd_clear_win(struct fimd_context *ctx, int win)
 {
-	u32 val;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	writel(0, ctx->regs + WINCON(win));
 	writel(0, ctx->regs + VIDOSD_A(win));
 	writel(0, ctx->regs + VIDOSD_B(win));
@@ -789,15 +799,11 @@ static void fimd_clear_win(struct fimd_context *ctx, int win)
 	if (win == 1 || win == 2)
 		writel(0, ctx->regs + VIDOSD_D(win));
 
-	val = readl(ctx->regs + SHADOWCON);
-	val &= ~SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, false);
 }
 
 static int fimd_clock(struct fimd_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (enable) {
 		int ret;
 
@@ -883,8 +889,6 @@ static int fimd_probe(struct platform_device *pdev)
 	int win;
 	int ret = -EINVAL;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (dev->of_node) {
 		pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
 		if (!pdata) {
@@ -949,6 +953,7 @@ static int fimd_probe(struct platform_device *pdev)
 		return ret;
 	}
 
+	ctx->driver_data = drm_fimd_get_driver_data(pdev);
 	ctx->vidcon0 = pdata->vidcon0;
 	ctx->vidcon1 = pdata->vidcon1;
 	ctx->default_win = pdata->default_win;
@@ -989,8 +994,6 @@ static int fimd_remove(struct platform_device *pdev)
 	struct device *dev = &pdev->dev;
 	struct fimd_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
 	if (ctx->suspended)
@@ -1055,8 +1058,6 @@ static int fimd_runtime_suspend(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return fimd_activate(ctx, false);
 }
 
@@ -1064,14 +1065,15 @@ static int fimd_runtime_resume(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return fimd_activate(ctx, true);
 }
 #endif
 
 static struct platform_device_id fimd_driver_ids[] = {
 	{
+		.name		= "s3c64xx-fb",
+		.driver_data	= (unsigned long)&s3c64xx_fimd_driver_data,
+	}, {
 		.name		= "exynos4-fb",
 		.driver_data	= (unsigned long)&exynos4_fimd_driver_data,
 	}, {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index af75434..42a5a54 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -388,12 +388,9 @@ out:
 
 	sg_free_table(g2d_userptr->sgt);
 	kfree(g2d_userptr->sgt);
-	g2d_userptr->sgt = NULL;
 
-	kfree(g2d_userptr->pages);
-	g2d_userptr->pages = NULL;
+	drm_free_large(g2d_userptr->pages);
 	kfree(g2d_userptr);
-	g2d_userptr = NULL;
 }
 
 static dma_addr_t *g2d_userptr_get_dma_addr(struct drm_device *drm_dev,
@@ -463,11 +460,11 @@ static dma_addr_t *g2d_userptr_get_dma_addr(struct drm_device *drm_dev,
 	npages = (end - start) >> PAGE_SHIFT;
 	g2d_userptr->npages = npages;
 
-	pages = kzalloc(npages * sizeof(struct page *), GFP_KERNEL);
+	pages = drm_calloc_large(npages, sizeof(struct page *));
 	if (!pages) {
 		DRM_ERROR("failed to allocate pages.\n");
-		kfree(g2d_userptr);
-		return ERR_PTR(-ENOMEM);
+		ret = -ENOMEM;
+		goto err_free;
 	}
 
 	vma = find_vma(current->mm, userptr);
@@ -543,7 +540,6 @@ err_sg_free_table:
 
 err_free_sgt:
 	kfree(sgt);
-	sgt = NULL;
 
 err_free_userptr:
 	exynos_gem_put_pages_to_userptr(g2d_userptr->pages,
@@ -554,10 +550,10 @@ err_put_vma:
 	exynos_gem_put_vma(g2d_userptr->vma);
 
 err_free_pages:
-	kfree(pages);
+	drm_free_large(pages);
+
+err_free:
 	kfree(g2d_userptr);
-	pages = NULL;
-	g2d_userptr = NULL;
 
 	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index cf4543f..24c22a8 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -132,8 +132,6 @@ void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj)
 	struct drm_gem_object *obj;
 	struct exynos_drm_gem_buf *buf;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	obj = &exynos_gem_obj->base;
 	buf = exynos_gem_obj->buffer;
 
@@ -227,7 +225,6 @@ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
 	}
 
 	size = roundup_gem_size(size, flags);
-	DRM_DEBUG_KMS("%s\n", __FILE__);
 
 	ret = check_gem_flags(flags);
 	if (ret)
@@ -249,13 +246,14 @@ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
 	exynos_gem_obj->flags = flags;
 
 	ret = exynos_drm_alloc_buf(dev, buf, flags);
-	if (ret < 0) {
-		drm_gem_object_release(&exynos_gem_obj->base);
-		goto err_fini_buf;
-	}
+	if (ret < 0)
+		goto err_gem_fini;
 
 	return exynos_gem_obj;
 
+err_gem_fini:
+	drm_gem_object_release(&exynos_gem_obj->base);
+	kfree(exynos_gem_obj);
 err_fini_buf:
 	exynos_drm_fini_buf(dev, buf);
 	return ERR_PTR(ret);
@@ -268,8 +266,6 @@ int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data,
 	struct exynos_drm_gem_obj *exynos_gem_obj;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size);
 	if (IS_ERR(exynos_gem_obj))
 		return PTR_ERR(exynos_gem_obj);
@@ -331,8 +327,6 @@ int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
 {
 	struct drm_exynos_gem_map_off *args = data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	DRM_DEBUG_KMS("handle = 0x%x, offset = 0x%lx\n",
 			args->handle, (unsigned long)args->offset);
 
@@ -371,8 +365,6 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp,
 	unsigned long vm_size;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
 	vma->vm_private_data = obj;
 	vma->vm_ops = drm_dev->driver->gem_vm_ops;
@@ -429,9 +421,7 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
 {
 	struct drm_exynos_gem_mmap *args = data;
 	struct drm_gem_object *obj;
-	unsigned int addr;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
+	unsigned long addr;
 
 	if (!(dev->driver->driver_features & DRIVER_GEM)) {
 		DRM_ERROR("does not support GEM.\n");
@@ -473,14 +463,14 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
 
 	drm_gem_object_unreference(obj);
 
-	if (IS_ERR((void *)addr)) {
+	if (IS_ERR_VALUE(addr)) {
 		/* check filp->f_op, filp->private_data are restored */
 		if (file_priv->filp->f_op == &exynos_drm_gem_fops) {
 			file_priv->filp->f_op = fops_get(dev->driver->fops);
 			file_priv->filp->private_data = file_priv;
 		}
 		mutex_unlock(&dev->struct_mutex);
-		return PTR_ERR((void *)addr);
+		return (int)addr;
 	}
 
 	mutex_unlock(&dev->struct_mutex);
@@ -643,8 +633,6 @@ void exynos_gem_unmap_sgt_from_dma(struct drm_device *drm_dev,
 
 int exynos_drm_gem_init_object(struct drm_gem_object *obj)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return 0;
 }
 
@@ -653,8 +641,6 @@ void exynos_drm_gem_free_object(struct drm_gem_object *obj)
 	struct exynos_drm_gem_obj *exynos_gem_obj;
 	struct exynos_drm_gem_buf *buf;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_gem_obj = to_exynos_gem_obj(obj);
 	buf = exynos_gem_obj->buffer;
 
@@ -671,8 +657,6 @@ int exynos_drm_gem_dumb_create(struct drm_file *file_priv,
 	struct exynos_drm_gem_obj *exynos_gem_obj;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * alocate memory to be used for framebuffer.
 	 * - this callback would be called by user application
@@ -704,8 +688,6 @@ int exynos_drm_gem_dumb_map_offset(struct drm_file *file_priv,
 	struct drm_gem_object *obj;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	mutex_lock(&dev->struct_mutex);
 
 	/*
@@ -743,8 +725,6 @@ int exynos_drm_gem_dumb_destroy(struct drm_file *file_priv,
 {
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * obj->refcount and obj->handle_count are decreased and
 	 * if both them are 0 then exynos_drm_gem_free_object()
@@ -788,8 +768,6 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct drm_gem_object *obj;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* set vm_area_struct. */
 	ret = drm_gem_mmap(filp, vma);
 	if (ret < 0) {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
index 762f40d..472e3b2 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
@@ -400,8 +400,6 @@ static int gsc_sw_reset(struct gsc_context *ctx)
 	u32 cfg;
 	int count = GSC_RESET_TIMEOUT;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* s/w reset */
 	cfg = (GSC_SW_RESET_SRESET);
 	gsc_write(cfg, GSC_SW_RESET);
@@ -441,8 +439,6 @@ static void gsc_set_gscblk_fimd_wb(struct gsc_context *ctx, bool enable)
 {
 	u32 gscblk_cfg;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	gscblk_cfg = readl(SYSREG_GSCBLK_CFG1);
 
 	if (enable)
@@ -460,7 +456,7 @@ static void gsc_handle_irq(struct gsc_context *ctx, bool enable,
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]overflow[%d]level[%d]\n", __func__,
+	DRM_DEBUG_KMS("enable[%d]overflow[%d]level[%d]\n",
 			enable, overflow, done);
 
 	cfg = gsc_read(GSC_IRQ);
@@ -491,7 +487,7 @@ static int gsc_src_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~(GSC_IN_RGB_TYPE_MASK | GSC_IN_YUV422_1P_ORDER_MASK |
@@ -567,8 +563,7 @@ static int gsc_src_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~GSC_IN_ROT_MASK;
@@ -616,8 +611,8 @@ static int gsc_src_set_size(struct device *dev, int swap,
 	struct gsc_scaler *sc = &ctx->sc;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
-		__func__, swap, pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
+		swap, pos->x, pos->y, pos->w, pos->h);
 
 	if (swap) {
 		img_pos.w = pos->h;
@@ -634,8 +629,7 @@ static int gsc_src_set_size(struct device *dev, int swap,
 		GSC_CROPPED_HEIGHT(img_pos.h));
 	gsc_write(cfg, GSC_CROPPED_SIZE);
 
-	DRM_DEBUG_KMS("%s:hsize[%d]vsize[%d]\n",
-		__func__, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("hsize[%d]vsize[%d]\n", sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = gsc_read(GSC_SRCIMG_SIZE);
@@ -650,8 +644,7 @@ static int gsc_src_set_size(struct device *dev, int swap,
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~GSC_IN_RGB_TYPE_MASK;
 
-	DRM_DEBUG_KMS("%s:width[%d]range[%d]\n",
-		__func__, pos->w, sc->range);
+	DRM_DEBUG_KMS("width[%d]range[%d]\n", pos->w, sc->range);
 
 	if (pos->w >= GSC_WIDTH_ITU_709)
 		if (sc->range)
@@ -677,8 +670,7 @@ static int gsc_src_set_buf_seq(struct gsc_context *ctx, u32 buf_id,
 	u32 cfg;
 	u32 mask = 0x00000001 << buf_id;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]buf_type[%d]\n", __func__,
-		buf_id, buf_type);
+	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	/* mask register set */
 	cfg = gsc_read(GSC_IN_BASE_ADDR_Y_MASK);
@@ -721,7 +713,7 @@ static int gsc_src_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > GSC_MAX_SRC) {
@@ -765,7 +757,7 @@ static int gsc_dst_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = gsc_read(GSC_OUT_CON);
 	cfg &= ~(GSC_OUT_RGB_TYPE_MASK | GSC_OUT_YUV422_1P_ORDER_MASK |
@@ -838,8 +830,7 @@ static int gsc_dst_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~GSC_IN_ROT_MASK;
@@ -881,7 +872,7 @@ static int gsc_dst_set_transf(struct device *dev,
 
 static int gsc_get_ratio_shift(u32 src, u32 dst, u32 *ratio)
 {
-	DRM_DEBUG_KMS("%s:src[%d]dst[%d]\n", __func__, src, dst);
+	DRM_DEBUG_KMS("src[%d]dst[%d]\n", src, dst);
 
 	if (src >= dst * 8) {
 		DRM_ERROR("failed to make ratio and shift.\n");
@@ -944,20 +935,19 @@ static int gsc_set_prescaler(struct gsc_context *ctx, struct gsc_scaler *sc,
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:pre_hratio[%d]pre_vratio[%d]\n",
-		__func__, sc->pre_hratio, sc->pre_vratio);
+	DRM_DEBUG_KMS("pre_hratio[%d]pre_vratio[%d]\n",
+		sc->pre_hratio, sc->pre_vratio);
 
 	sc->main_hratio = (src_w << 16) / dst_w;
 	sc->main_vratio = (src_h << 16) / dst_h;
 
-	DRM_DEBUG_KMS("%s:main_hratio[%ld]main_vratio[%ld]\n",
-		__func__, sc->main_hratio, sc->main_vratio);
+	DRM_DEBUG_KMS("main_hratio[%ld]main_vratio[%ld]\n",
+		sc->main_hratio, sc->main_vratio);
 
 	gsc_get_prescaler_shfactor(sc->pre_hratio, sc->pre_vratio,
 		&sc->pre_shfactor);
 
-	DRM_DEBUG_KMS("%s:pre_shfactor[%d]\n", __func__,
-		sc->pre_shfactor);
+	DRM_DEBUG_KMS("pre_shfactor[%d]\n", sc->pre_shfactor);
 
 	cfg = (GSC_PRESC_SHFACTOR(sc->pre_shfactor) |
 		GSC_PRESC_H_RATIO(sc->pre_hratio) |
@@ -1023,8 +1013,8 @@ static void gsc_set_scaler(struct gsc_context *ctx, struct gsc_scaler *sc)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:main_hratio[%ld]main_vratio[%ld]\n",
-		__func__, sc->main_hratio, sc->main_vratio);
+	DRM_DEBUG_KMS("main_hratio[%ld]main_vratio[%ld]\n",
+		sc->main_hratio, sc->main_vratio);
 
 	gsc_set_h_coef(ctx, sc->main_hratio);
 	cfg = GSC_MAIN_H_RATIO_VALUE(sc->main_hratio);
@@ -1043,8 +1033,8 @@ static int gsc_dst_set_size(struct device *dev, int swap,
 	struct gsc_scaler *sc = &ctx->sc;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
-		__func__, swap, pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
+		swap, pos->x, pos->y, pos->w, pos->h);
 
 	if (swap) {
 		img_pos.w = pos->h;
@@ -1060,8 +1050,7 @@ static int gsc_dst_set_size(struct device *dev, int swap,
 	cfg = (GSC_SCALED_WIDTH(img_pos.w) | GSC_SCALED_HEIGHT(img_pos.h));
 	gsc_write(cfg, GSC_SCALED_SIZE);
 
-	DRM_DEBUG_KMS("%s:hsize[%d]vsize[%d]\n",
-		__func__, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("hsize[%d]vsize[%d]\n", sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = gsc_read(GSC_DSTIMG_SIZE);
@@ -1074,8 +1063,7 @@ static int gsc_dst_set_size(struct device *dev, int swap,
 	cfg = gsc_read(GSC_OUT_CON);
 	cfg &= ~GSC_OUT_RGB_TYPE_MASK;
 
-	DRM_DEBUG_KMS("%s:width[%d]range[%d]\n",
-		__func__, pos->w, sc->range);
+	DRM_DEBUG_KMS("width[%d]range[%d]\n", pos->w, sc->range);
 
 	if (pos->w >= GSC_WIDTH_ITU_709)
 		if (sc->range)
@@ -1104,7 +1092,7 @@ static int gsc_dst_get_buf_seq(struct gsc_context *ctx)
 		if (cfg & (mask << i))
 			buf_num--;
 
-	DRM_DEBUG_KMS("%s:buf_num[%d]\n", __func__, buf_num);
+	DRM_DEBUG_KMS("buf_num[%d]\n", buf_num);
 
 	return buf_num;
 }
@@ -1118,8 +1106,7 @@ static int gsc_dst_set_buf_seq(struct gsc_context *ctx, u32 buf_id,
 	u32 mask = 0x00000001 << buf_id;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]buf_type[%d]\n", __func__,
-		buf_id, buf_type);
+	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	mutex_lock(&ctx->lock);
 
@@ -1177,7 +1164,7 @@ static int gsc_dst_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > GSC_MAX_DST) {
@@ -1217,7 +1204,7 @@ static struct exynos_drm_ipp_ops gsc_dst_ops = {
 
 static int gsc_clk_ctrl(struct gsc_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	if (enable) {
 		clk_enable(ctx->gsc_clk);
@@ -1236,7 +1223,7 @@ static int gsc_get_src_buf_index(struct gsc_context *ctx)
 	u32 buf_id = GSC_MAX_SRC;
 	int ret;
 
-	DRM_DEBUG_KMS("%s:gsc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("gsc id[%d]\n", ctx->id);
 
 	cfg = gsc_read(GSC_IN_BASE_ADDR_Y_MASK);
 	curr_index = GSC_IN_CURR_GET_INDEX(cfg);
@@ -1259,7 +1246,7 @@ static int gsc_get_src_buf_index(struct gsc_context *ctx)
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:cfg[0x%x]curr_index[%d]buf_id[%d]\n", __func__, cfg,
+	DRM_DEBUG_KMS("cfg[0x%x]curr_index[%d]buf_id[%d]\n", cfg,
 		curr_index, buf_id);
 
 	return buf_id;
@@ -1271,7 +1258,7 @@ static int gsc_get_dst_buf_index(struct gsc_context *ctx)
 	u32 buf_id = GSC_MAX_DST;
 	int ret;
 
-	DRM_DEBUG_KMS("%s:gsc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("gsc id[%d]\n", ctx->id);
 
 	cfg = gsc_read(GSC_OUT_BASE_ADDR_Y_MASK);
 	curr_index = GSC_OUT_CURR_GET_INDEX(cfg);
@@ -1294,7 +1281,7 @@ static int gsc_get_dst_buf_index(struct gsc_context *ctx)
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:cfg[0x%x]curr_index[%d]buf_id[%d]\n", __func__, cfg,
+	DRM_DEBUG_KMS("cfg[0x%x]curr_index[%d]buf_id[%d]\n", cfg,
 		curr_index, buf_id);
 
 	return buf_id;
@@ -1310,7 +1297,7 @@ static irqreturn_t gsc_irq_handler(int irq, void *dev_id)
 	u32 status;
 	int buf_id[EXYNOS_DRM_OPS_MAX];
 
-	DRM_DEBUG_KMS("%s:gsc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("gsc id[%d]\n", ctx->id);
 
 	status = gsc_read(GSC_IRQ);
 	if (status & GSC_IRQ_STATUS_OR_IRQ) {
@@ -1331,7 +1318,7 @@ static irqreturn_t gsc_irq_handler(int irq, void *dev_id)
 		if (buf_id[EXYNOS_DRM_OPS_DST] < 0)
 			return IRQ_HANDLED;
 
-		DRM_DEBUG_KMS("%s:buf_id_src[%d]buf_id_dst[%d]\n", __func__,
+		DRM_DEBUG_KMS("buf_id_src[%d]buf_id_dst[%d]\n",
 			buf_id[EXYNOS_DRM_OPS_SRC], buf_id[EXYNOS_DRM_OPS_DST]);
 
 		event_work->ippdrv = ippdrv;
@@ -1350,8 +1337,6 @@ static int gsc_init_prop_list(struct exynos_drm_ippdrv *ippdrv)
 {
 	struct drm_exynos_ipp_prop_list *prop_list;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop_list = devm_kzalloc(ippdrv->dev, sizeof(*prop_list), GFP_KERNEL);
 	if (!prop_list) {
 		DRM_ERROR("failed to alloc property list.\n");
@@ -1394,7 +1379,7 @@ static inline bool gsc_check_drm_flip(enum drm_exynos_flip flip)
 	case EXYNOS_DRM_FLIP_BOTH:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return false;
 	}
 }
@@ -1411,8 +1396,6 @@ static int gsc_ippdrv_check_property(struct device *dev,
 	bool swap;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	for_each_ipp_ops(i) {
 		if ((i == EXYNOS_DRM_OPS_SRC) &&
 			(property->cmd == IPP_CMD_WB))
@@ -1521,8 +1504,6 @@ static int gsc_ippdrv_reset(struct device *dev)
 	struct gsc_scaler *sc = &ctx->sc;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* reset h/w block */
 	ret = gsc_sw_reset(ctx);
 	if (ret < 0) {
@@ -1549,7 +1530,7 @@ static int gsc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	u32 cfg;
 	int ret, i;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	if (!c_node) {
 		DRM_ERROR("failed to get c_node.\n");
@@ -1643,7 +1624,7 @@ static void gsc_ippdrv_stop(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	struct drm_exynos_ipp_set_wb set_wb = {0, 0};
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	switch (cmd) {
 	case IPP_CMD_M2M:
@@ -1728,8 +1709,7 @@ static int gsc_probe(struct platform_device *pdev)
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:id[%d]ippdrv[0x%x]\n", __func__, ctx->id,
-		(int)ippdrv);
+	DRM_DEBUG_KMS("id[%d]ippdrv[0x%x]\n", ctx->id, (int)ippdrv);
 
 	mutex_init(&ctx->lock);
 	platform_set_drvdata(pdev, ctx);
@@ -1772,7 +1752,7 @@ static int gsc_suspend(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (pm_runtime_suspended(dev))
 		return 0;
@@ -1784,7 +1764,7 @@ static int gsc_resume(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (!pm_runtime_suspended(dev))
 		return gsc_clk_ctrl(ctx, true);
@@ -1798,7 +1778,7 @@ static int gsc_runtime_suspend(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  gsc_clk_ctrl(ctx, false);
 }
@@ -1807,7 +1787,7 @@ static int gsc_runtime_resume(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __FILE__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  gsc_clk_ctrl(ctx, true);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
index 437fb94..aaa550d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
@@ -88,16 +88,12 @@ void exynos_mixer_drv_attach(struct exynos_drm_hdmi_context *ctx)
 
 void exynos_hdmi_ops_register(struct exynos_hdmi_ops *ops)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ops)
 		hdmi_ops = ops;
 }
 
 void exynos_mixer_ops_register(struct exynos_mixer_ops *ops)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ops)
 		mixer_ops = ops;
 }
@@ -106,8 +102,6 @@ static bool drm_hdmi_is_connected(struct device *dev)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->is_connected)
 		return hdmi_ops->is_connected(ctx->hdmi_ctx->ctx);
 
@@ -119,34 +113,31 @@ static struct edid *drm_hdmi_get_edid(struct device *dev,
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->get_edid)
 		return hdmi_ops->get_edid(ctx->hdmi_ctx->ctx, connector);
 
 	return NULL;
 }
 
-static int drm_hdmi_check_timing(struct device *dev, void *timing)
+static int drm_hdmi_check_mode(struct device *dev,
+		struct drm_display_mode *mode)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	* Both, mixer and hdmi should be able to handle the requested mode.
 	* If any of the two fails, return mode as BAD.
 	*/
 
-	if (mixer_ops && mixer_ops->check_timing)
-		ret = mixer_ops->check_timing(ctx->mixer_ctx->ctx, timing);
+	if (mixer_ops && mixer_ops->check_mode)
+		ret = mixer_ops->check_mode(ctx->mixer_ctx->ctx, mode);
 
 	if (ret)
 		return ret;
 
-	if (hdmi_ops && hdmi_ops->check_timing)
-		return hdmi_ops->check_timing(ctx->hdmi_ctx->ctx, timing);
+	if (hdmi_ops && hdmi_ops->check_mode)
+		return hdmi_ops->check_mode(ctx->hdmi_ctx->ctx, mode);
 
 	return 0;
 }
@@ -155,8 +146,6 @@ static int drm_hdmi_power_on(struct device *dev, int mode)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->power_on)
 		return hdmi_ops->power_on(ctx->hdmi_ctx->ctx, mode);
 
@@ -167,7 +156,7 @@ static struct exynos_drm_display_ops drm_hdmi_display_ops = {
 	.type = EXYNOS_DISPLAY_TYPE_HDMI,
 	.is_connected = drm_hdmi_is_connected,
 	.get_edid = drm_hdmi_get_edid,
-	.check_timing = drm_hdmi_check_timing,
+	.check_mode = drm_hdmi_check_mode,
 	.power_on = drm_hdmi_power_on,
 };
 
@@ -177,8 +166,6 @@ static int drm_hdmi_enable_vblank(struct device *subdrv_dev)
 	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
 	struct exynos_drm_manager *manager = subdrv->manager;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->enable_vblank)
 		return mixer_ops->enable_vblank(ctx->mixer_ctx->ctx,
 						manager->pipe);
@@ -190,8 +177,6 @@ static void drm_hdmi_disable_vblank(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->disable_vblank)
 		return mixer_ops->disable_vblank(ctx->mixer_ctx->ctx);
 }
@@ -200,8 +185,6 @@ static void drm_hdmi_wait_for_vblank(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->wait_for_vblank)
 		mixer_ops->wait_for_vblank(ctx->mixer_ctx->ctx);
 }
@@ -214,11 +197,9 @@ static void drm_hdmi_mode_fixup(struct device *subdrv_dev,
 	struct drm_display_mode *m;
 	int mode_ok;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	drm_mode_set_crtcinfo(adjusted_mode, 0);
 
-	mode_ok = drm_hdmi_check_timing(subdrv_dev, adjusted_mode);
+	mode_ok = drm_hdmi_check_mode(subdrv_dev, adjusted_mode);
 
 	/* just return if user desired mode exists. */
 	if (mode_ok == 0)
@@ -229,7 +210,7 @@ static void drm_hdmi_mode_fixup(struct device *subdrv_dev,
 	 * to adjusted_mode.
 	 */
 	list_for_each_entry(m, &connector->modes, head) {
-		mode_ok = drm_hdmi_check_timing(subdrv_dev, m);
+		mode_ok = drm_hdmi_check_mode(subdrv_dev, m);
 
 		if (mode_ok == 0) {
 			struct drm_mode_object base;
@@ -256,8 +237,6 @@ static void drm_hdmi_mode_set(struct device *subdrv_dev, void *mode)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->mode_set)
 		hdmi_ops->mode_set(ctx->hdmi_ctx->ctx, mode);
 }
@@ -267,8 +246,6 @@ static void drm_hdmi_get_max_resol(struct device *subdrv_dev,
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->get_max_resol)
 		hdmi_ops->get_max_resol(ctx->hdmi_ctx->ctx, width, height);
 }
@@ -277,8 +254,6 @@ static void drm_hdmi_commit(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->commit)
 		hdmi_ops->commit(ctx->hdmi_ctx->ctx);
 }
@@ -287,8 +262,6 @@ static void drm_hdmi_dpms(struct device *subdrv_dev, int mode)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->dpms)
 		mixer_ops->dpms(ctx->mixer_ctx->ctx, mode);
 
@@ -301,8 +274,6 @@ static void drm_hdmi_apply(struct device *subdrv_dev)
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	for (i = 0; i < MIXER_WIN_NR; i++) {
 		if (!ctx->enabled[i])
 			continue;
@@ -331,8 +302,6 @@ static void drm_mixer_mode_set(struct device *subdrv_dev,
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->win_mode_set)
 		mixer_ops->win_mode_set(ctx->mixer_ctx->ctx, overlay);
 }
@@ -342,9 +311,7 @@ static void drm_mixer_commit(struct device *subdrv_dev, int zpos)
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 	int win = (zpos == DEFAULT_ZPOS) ? MIXER_DEFAULT_WIN : zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (win < 0 || win > MIXER_WIN_NR) {
+	if (win < 0 || win >= MIXER_WIN_NR) {
 		DRM_ERROR("mixer window[%d] is wrong\n", win);
 		return;
 	}
@@ -360,9 +327,7 @@ static void drm_mixer_disable(struct device *subdrv_dev, int zpos)
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 	int win = (zpos == DEFAULT_ZPOS) ? MIXER_DEFAULT_WIN : zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (win < 0 || win > MIXER_WIN_NR) {
+	if (win < 0 || win >= MIXER_WIN_NR) {
 		DRM_ERROR("mixer window[%d] is wrong\n", win);
 		return;
 	}
@@ -392,8 +357,6 @@ static int hdmi_subdrv_probe(struct drm_device *drm_dev,
 	struct exynos_drm_subdrv *subdrv = to_subdrv(dev);
 	struct drm_hdmi_context *ctx;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!hdmi_ctx) {
 		DRM_ERROR("hdmi context not initialized.\n");
 		return -EFAULT;
@@ -440,8 +403,6 @@ static int exynos_drm_hdmi_probe(struct platform_device *pdev)
 	struct exynos_drm_subdrv *subdrv;
 	struct drm_hdmi_context *ctx;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx) {
 		DRM_LOG_KMS("failed to alloc common hdmi context.\n");
@@ -466,8 +427,6 @@ static int exynos_drm_hdmi_remove(struct platform_device *pdev)
 {
 	struct drm_hdmi_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
 	return 0;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
index 6b70944..724cab1 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
@@ -32,11 +32,11 @@ struct exynos_hdmi_ops {
 	bool (*is_connected)(void *ctx);
 	struct edid *(*get_edid)(void *ctx,
 			struct drm_connector *connector);
-	int (*check_timing)(void *ctx, struct fb_videomode *timing);
+	int (*check_mode)(void *ctx, struct drm_display_mode *mode);
 	int (*power_on)(void *ctx, int mode);
 
 	/* manager */
-	void (*mode_set)(void *ctx, void *mode);
+	void (*mode_set)(void *ctx, struct drm_display_mode *mode);
 	void (*get_max_resol)(void *ctx, unsigned int *width,
 				unsigned int *height);
 	void (*commit)(void *ctx);
@@ -57,7 +57,7 @@ struct exynos_mixer_ops {
 	void (*win_disable)(void *ctx, int zpos);
 
 	/* display */
-	int (*check_timing)(void *ctx, struct fb_videomode *timing);
+	int (*check_mode)(void *ctx, struct drm_display_mode *mode);
 };
 
 void exynos_hdmi_drv_attach(struct exynos_drm_hdmi_context *ctx);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.c b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
index be1e884..b1ef8e7 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
@@ -131,8 +131,6 @@ void exynos_platform_device_ipp_unregister(void)
 
 int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ippdrv)
 		return -EINVAL;
 
@@ -145,8 +143,6 @@ int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 
 int exynos_drm_ippdrv_unregister(struct exynos_drm_ippdrv *ippdrv)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ippdrv)
 		return -EINVAL;
 
@@ -162,8 +158,6 @@ static int ipp_create_id(struct idr *id_idr, struct mutex *lock, void *obj,
 {
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* do the allocation under our mutexlock */
 	mutex_lock(lock);
 	ret = idr_alloc(id_idr, obj, 1, 0, GFP_KERNEL);
@@ -179,7 +173,7 @@ static void *ipp_find_obj(struct idr *id_idr, struct mutex *lock, u32 id)
 {
 	void *obj;
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, id);
+	DRM_DEBUG_KMS("id[%d]\n", id);
 
 	mutex_lock(lock);
 
@@ -216,7 +210,7 @@ static struct exynos_drm_ippdrv *ipp_find_driver(struct ipp_context *ctx,
 	struct exynos_drm_ippdrv *ippdrv;
 	u32 ipp_id = property->ipp_id;
 
-	DRM_DEBUG_KMS("%s:ipp_id[%d]\n", __func__, ipp_id);
+	DRM_DEBUG_KMS("ipp_id[%d]\n", ipp_id);
 
 	if (ipp_id) {
 		/* find ipp driver using idr */
@@ -257,14 +251,13 @@ static struct exynos_drm_ippdrv *ipp_find_driver(struct ipp_context *ctx,
 		 */
 		list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 			if (ipp_check_dedicated(ippdrv, property->cmd)) {
-				DRM_DEBUG_KMS("%s:used device.\n", __func__);
+				DRM_DEBUG_KMS("used device.\n");
 				continue;
 			}
 
 			if (ippdrv->check_property &&
 			    ippdrv->check_property(ippdrv->dev, property)) {
-				DRM_DEBUG_KMS("%s:not support property.\n",
-					__func__);
+				DRM_DEBUG_KMS("not support property.\n");
 				continue;
 			}
 
@@ -283,10 +276,10 @@ static struct exynos_drm_ippdrv *ipp_find_drv_by_handle(u32 prop_id)
 	struct drm_exynos_ipp_cmd_node *c_node;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", prop_id);
 
 	if (list_empty(&exynos_drm_ippdrv_list)) {
-		DRM_DEBUG_KMS("%s:ippdrv_list is empty.\n", __func__);
+		DRM_DEBUG_KMS("ippdrv_list is empty.\n");
 		return ERR_PTR(-ENODEV);
 	}
 
@@ -296,8 +289,7 @@ static struct exynos_drm_ippdrv *ipp_find_drv_by_handle(u32 prop_id)
 	 * e.g PAUSE state, queue buf, command contro.
 	 */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
-		DRM_DEBUG_KMS("%s:count[%d]ippdrv[0x%x]\n", __func__,
-			count++, (int)ippdrv);
+		DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]\n", count++, (int)ippdrv);
 
 		if (!list_empty(&ippdrv->cmd_list)) {
 			list_for_each_entry(c_node, &ippdrv->cmd_list, list)
@@ -320,8 +312,6 @@ int exynos_drm_ipp_get_property(struct drm_device *drm_dev, void *data,
 	struct exynos_drm_ippdrv *ippdrv;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ctx) {
 		DRM_ERROR("invalid context.\n");
 		return -EINVAL;
@@ -332,7 +322,7 @@ int exynos_drm_ipp_get_property(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:ipp_id[%d]\n", __func__, prop_list->ipp_id);
+	DRM_DEBUG_KMS("ipp_id[%d]\n", prop_list->ipp_id);
 
 	if (!prop_list->ipp_id) {
 		list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list)
@@ -371,11 +361,11 @@ static void ipp_print_property(struct drm_exynos_ipp_property *property,
 	struct drm_exynos_pos *pos = &config->pos;
 	struct drm_exynos_sz *sz = &config->sz;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]ops[%s]fmt[0x%x]\n",
-		__func__, property->prop_id, idx ? "dst" : "src", config->fmt);
+	DRM_DEBUG_KMS("prop_id[%d]ops[%s]fmt[0x%x]\n",
+		property->prop_id, idx ? "dst" : "src", config->fmt);
 
-	DRM_DEBUG_KMS("%s:pos[%d %d %d %d]sz[%d %d]f[%d]r[%d]\n",
-		__func__, pos->x, pos->y, pos->w, pos->h,
+	DRM_DEBUG_KMS("pos[%d %d %d %d]sz[%d %d]f[%d]r[%d]\n",
+		pos->x, pos->y, pos->w, pos->h,
 		sz->hsize, sz->vsize, config->flip, config->degree);
 }
 
@@ -385,7 +375,7 @@ static int ipp_find_and_set_property(struct drm_exynos_ipp_property *property)
 	struct drm_exynos_ipp_cmd_node *c_node;
 	u32 prop_id = property->prop_id;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", prop_id);
 
 	ippdrv = ipp_find_drv_by_handle(prop_id);
 	if (IS_ERR(ippdrv)) {
@@ -401,8 +391,8 @@ static int ipp_find_and_set_property(struct drm_exynos_ipp_property *property)
 	list_for_each_entry(c_node, &ippdrv->cmd_list, list) {
 		if ((c_node->property.prop_id == prop_id) &&
 		    (c_node->state == IPP_STATE_STOP)) {
-			DRM_DEBUG_KMS("%s:found cmd[%d]ippdrv[0x%x]\n",
-				__func__, property->cmd, (int)ippdrv);
+			DRM_DEBUG_KMS("found cmd[%d]ippdrv[0x%x]\n",
+				property->cmd, (int)ippdrv);
 
 			c_node->property = *property;
 			return 0;
@@ -418,8 +408,6 @@ static struct drm_exynos_ipp_cmd_work *ipp_create_cmd_work(void)
 {
 	struct drm_exynos_ipp_cmd_work *cmd_work;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	cmd_work = kzalloc(sizeof(*cmd_work), GFP_KERNEL);
 	if (!cmd_work) {
 		DRM_ERROR("failed to alloc cmd_work.\n");
@@ -435,8 +423,6 @@ static struct drm_exynos_ipp_event_work *ipp_create_event_work(void)
 {
 	struct drm_exynos_ipp_event_work *event_work;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	event_work = kzalloc(sizeof(*event_work), GFP_KERNEL);
 	if (!event_work) {
 		DRM_ERROR("failed to alloc event_work.\n");
@@ -460,8 +446,6 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_cmd_node *c_node;
 	int ret, i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ctx) {
 		DRM_ERROR("invalid context.\n");
 		return -EINVAL;
@@ -486,7 +470,7 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 	 * instead of allocation.
 	 */
 	if (property->prop_id) {
-		DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+		DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 		return ipp_find_and_set_property(property);
 	}
 
@@ -512,8 +496,8 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		goto err_clear;
 	}
 
-	DRM_DEBUG_KMS("%s:created prop_id[%d]cmd[%d]ippdrv[0x%x]\n",
-		__func__, property->prop_id, property->cmd, (int)ippdrv);
+	DRM_DEBUG_KMS("created prop_id[%d]cmd[%d]ippdrv[0x%x]\n",
+		property->prop_id, property->cmd, (int)ippdrv);
 
 	/* stored property information and ippdrv in private data */
 	c_node->priv = priv;
@@ -569,8 +553,6 @@ err_clear:
 
 static void ipp_clean_cmd_node(struct drm_exynos_ipp_cmd_node *c_node)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* delete list */
 	list_del(&c_node->list);
 
@@ -593,8 +575,6 @@ static int ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
 	struct list_head *head;
 	int ret, i, count[EXYNOS_DRM_OPS_MAX] = { 0, };
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	mutex_lock(&c_node->mem_lock);
 
 	for_each_ipp_ops(i) {
@@ -602,20 +582,19 @@ static int ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
 		head = &c_node->mem_list[i];
 
 		if (list_empty(head)) {
-			DRM_DEBUG_KMS("%s:%s memory empty.\n", __func__,
-				i ? "dst" : "src");
+			DRM_DEBUG_KMS("%s memory empty.\n", i ? "dst" : "src");
 			continue;
 		}
 
 		/* find memory node entry */
 		list_for_each_entry(m_node, head, list) {
-			DRM_DEBUG_KMS("%s:%s,count[%d]m_node[0x%x]\n", __func__,
+			DRM_DEBUG_KMS("%s,count[%d]m_node[0x%x]\n",
 				i ? "dst" : "src", count[i], (int)m_node);
 			count[i]++;
 		}
 	}
 
-	DRM_DEBUG_KMS("%s:min[%d]max[%d]\n", __func__,
+	DRM_DEBUG_KMS("min[%d]max[%d]\n",
 		min(count[EXYNOS_DRM_OPS_SRC], count[EXYNOS_DRM_OPS_DST]),
 		max(count[EXYNOS_DRM_OPS_SRC], count[EXYNOS_DRM_OPS_DST]));
 
@@ -644,15 +623,14 @@ static struct drm_exynos_ipp_mem_node
 	struct list_head *head;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__, qbuf->buf_id);
+	DRM_DEBUG_KMS("buf_id[%d]\n", qbuf->buf_id);
 
 	/* source/destination memory list */
 	head = &c_node->mem_list[qbuf->ops_id];
 
 	/* find memory node from memory list */
 	list_for_each_entry(m_node, head, list) {
-		DRM_DEBUG_KMS("%s:count[%d]m_node[0x%x]\n",
-			__func__, count++, (int)m_node);
+		DRM_DEBUG_KMS("count[%d]m_node[0x%x]\n", count++, (int)m_node);
 
 		/* compare buffer id */
 		if (m_node->buf_id == qbuf->buf_id)
@@ -669,7 +647,7 @@ static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
 	struct exynos_drm_ipp_ops *ops = NULL;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s:node[0x%x]\n", __func__, (int)m_node);
+	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
 
 	if (!m_node) {
 		DRM_ERROR("invalid queue node.\n");
@@ -678,7 +656,7 @@ static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
 
 	mutex_lock(&c_node->mem_lock);
 
-	DRM_DEBUG_KMS("%s:ops_id[%d]\n", __func__, m_node->ops_id);
+	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
 
 	/* get operations callback */
 	ops = ippdrv->ops[m_node->ops_id];
@@ -714,8 +692,6 @@ static struct drm_exynos_ipp_mem_node
 	void *addr;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	mutex_lock(&c_node->mem_lock);
 
 	m_node = kzalloc(sizeof(*m_node), GFP_KERNEL);
@@ -732,14 +708,11 @@ static struct drm_exynos_ipp_mem_node
 	m_node->prop_id = qbuf->prop_id;
 	m_node->buf_id = qbuf->buf_id;
 
-	DRM_DEBUG_KMS("%s:m_node[0x%x]ops_id[%d]\n", __func__,
-		(int)m_node, qbuf->ops_id);
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]\n", __func__,
-		qbuf->prop_id, m_node->buf_id);
+	DRM_DEBUG_KMS("m_node[0x%x]ops_id[%d]\n", (int)m_node, qbuf->ops_id);
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]\n", qbuf->prop_id, m_node->buf_id);
 
 	for_each_ipp_planar(i) {
-		DRM_DEBUG_KMS("%s:i[%d]handle[0x%x]\n", __func__,
-			i, qbuf->handle[i]);
+		DRM_DEBUG_KMS("i[%d]handle[0x%x]\n", i, qbuf->handle[i]);
 
 		/* get dma address by handle */
 		if (qbuf->handle[i]) {
@@ -752,9 +725,8 @@ static struct drm_exynos_ipp_mem_node
 
 			buf_info.handles[i] = qbuf->handle[i];
 			buf_info.base[i] = *(dma_addr_t *) addr;
-			DRM_DEBUG_KMS("%s:i[%d]base[0x%x]hd[0x%x]\n",
-				__func__, i, buf_info.base[i],
-				(int)buf_info.handles[i]);
+			DRM_DEBUG_KMS("i[%d]base[0x%x]hd[0x%x]\n",
+				i, buf_info.base[i], (int)buf_info.handles[i]);
 		}
 	}
 
@@ -778,7 +750,7 @@ static int ipp_put_mem_node(struct drm_device *drm_dev,
 {
 	int i;
 
-	DRM_DEBUG_KMS("%s:node[0x%x]\n", __func__, (int)m_node);
+	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
 
 	if (!m_node) {
 		DRM_ERROR("invalid dequeue node.\n");
@@ -792,7 +764,7 @@ static int ipp_put_mem_node(struct drm_device *drm_dev,
 
 	mutex_lock(&c_node->mem_lock);
 
-	DRM_DEBUG_KMS("%s:ops_id[%d]\n", __func__, m_node->ops_id);
+	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
 
 	/* put gem buffer */
 	for_each_ipp_planar(i) {
@@ -824,8 +796,7 @@ static int ipp_get_event(struct drm_device *drm_dev,
 	struct drm_exynos_ipp_send_event *e;
 	unsigned long flags;
 
-	DRM_DEBUG_KMS("%s:ops_id[%d]buf_id[%d]\n", __func__,
-		qbuf->ops_id, qbuf->buf_id);
+	DRM_DEBUG_KMS("ops_id[%d]buf_id[%d]\n", qbuf->ops_id, qbuf->buf_id);
 
 	e = kzalloc(sizeof(*e), GFP_KERNEL);
 
@@ -857,16 +828,13 @@ static void ipp_put_event(struct drm_exynos_ipp_cmd_node *c_node,
 	struct drm_exynos_ipp_send_event *e, *te;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (list_empty(&c_node->event_list)) {
-		DRM_DEBUG_KMS("%s:event_list is empty.\n", __func__);
+		DRM_DEBUG_KMS("event_list is empty.\n");
 		return;
 	}
 
 	list_for_each_entry_safe(e, te, &c_node->event_list, base.link) {
-		DRM_DEBUG_KMS("%s:count[%d]e[0x%x]\n",
-			__func__, count++, (int)e);
+		DRM_DEBUG_KMS("count[%d]e[0x%x]\n", count++, (int)e);
 
 		/*
 		 * quf == NULL condition means all event deletion.
@@ -912,8 +880,6 @@ static int ipp_queue_buf_with_run(struct device *dev,
 	struct exynos_drm_ipp_ops *ops;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	ippdrv = ipp_find_drv_by_handle(qbuf->prop_id);
 	if (IS_ERR(ippdrv)) {
 		DRM_ERROR("failed to get ipp driver.\n");
@@ -929,12 +895,12 @@ static int ipp_queue_buf_with_run(struct device *dev,
 	property = &c_node->property;
 
 	if (c_node->state != IPP_STATE_START) {
-		DRM_DEBUG_KMS("%s:bypass for invalid state.\n" , __func__);
+		DRM_DEBUG_KMS("bypass for invalid state.\n");
 		return 0;
 	}
 
 	if (!ipp_check_mem_list(c_node)) {
-		DRM_DEBUG_KMS("%s:empty memory.\n", __func__);
+		DRM_DEBUG_KMS("empty memory.\n");
 		return 0;
 	}
 
@@ -964,8 +930,6 @@ static void ipp_clean_queue_buf(struct drm_device *drm_dev,
 {
 	struct drm_exynos_ipp_mem_node *m_node, *tm_node;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!list_empty(&c_node->mem_list[qbuf->ops_id])) {
 		/* delete list */
 		list_for_each_entry_safe(m_node, tm_node,
@@ -989,8 +953,6 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_mem_node *m_node;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!qbuf) {
 		DRM_ERROR("invalid buf parameter.\n");
 		return -EINVAL;
@@ -1001,8 +963,8 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]ops_id[%s]buf_id[%d]buf_type[%d]\n",
-		__func__, qbuf->prop_id, qbuf->ops_id ? "dst" : "src",
+	DRM_DEBUG_KMS("prop_id[%d]ops_id[%s]buf_id[%d]buf_type[%d]\n",
+		qbuf->prop_id, qbuf->ops_id ? "dst" : "src",
 		qbuf->buf_id, qbuf->buf_type);
 
 	/* find command node */
@@ -1075,8 +1037,6 @@ err_clean_node:
 static bool exynos_drm_ipp_check_valid(struct device *dev,
 		enum drm_exynos_ipp_ctrl ctrl, enum drm_exynos_ipp_state state)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (ctrl != IPP_CTRL_PLAY) {
 		if (pm_runtime_suspended(dev)) {
 			DRM_ERROR("pm:runtime_suspended.\n");
@@ -1104,7 +1064,6 @@ static bool exynos_drm_ipp_check_valid(struct device *dev,
 	default:
 		DRM_ERROR("invalid state.\n");
 		goto err_status;
-		break;
 	}
 
 	return true;
@@ -1126,8 +1085,6 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_cmd_work *cmd_work;
 	struct drm_exynos_ipp_cmd_node *c_node;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ctx) {
 		DRM_ERROR("invalid context.\n");
 		return -EINVAL;
@@ -1138,7 +1095,7 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:ctrl[%d]prop_id[%d]\n", __func__,
+	DRM_DEBUG_KMS("ctrl[%d]prop_id[%d]\n",
 		cmd_ctrl->ctrl, cmd_ctrl->prop_id);
 
 	ippdrv = ipp_find_drv_by_handle(cmd_ctrl->prop_id);
@@ -1213,7 +1170,7 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:done ctrl[%d]prop_id[%d]\n", __func__,
+	DRM_DEBUG_KMS("done ctrl[%d]prop_id[%d]\n",
 		cmd_ctrl->ctrl, cmd_ctrl->prop_id);
 
 	return 0;
@@ -1249,7 +1206,7 @@ static int ipp_set_property(struct exynos_drm_ippdrv *ippdrv,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
 	/* reset h/w block */
 	if (ippdrv->reset &&
@@ -1310,13 +1267,13 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 	struct list_head *head;
 	int ret, i;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
 	/* store command info in ippdrv */
 	ippdrv->c_node = c_node;
 
 	if (!ipp_check_mem_list(c_node)) {
-		DRM_DEBUG_KMS("%s:empty memory.\n", __func__);
+		DRM_DEBUG_KMS("empty memory.\n");
 		return -ENOMEM;
 	}
 
@@ -1343,8 +1300,7 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 				return ret;
 			}
 
-			DRM_DEBUG_KMS("%s:m_node[0x%x]\n",
-				__func__, (int)m_node);
+			DRM_DEBUG_KMS("m_node[0x%x]\n", (int)m_node);
 
 			ret = ipp_set_mem_node(ippdrv, c_node, m_node);
 			if (ret) {
@@ -1382,7 +1338,7 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, property->cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", property->cmd);
 
 	/* start operations */
 	if (ippdrv->start) {
@@ -1405,7 +1361,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 	struct list_head *head;
 	int ret = 0, i;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
 	/* put event */
 	ipp_put_event(c_node, NULL);
@@ -1418,8 +1374,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 			head = &c_node->mem_list[i];
 
 			if (list_empty(head)) {
-				DRM_DEBUG_KMS("%s:mem_list is empty.\n",
-					__func__);
+				DRM_DEBUG_KMS("mem_list is empty.\n");
 				break;
 			}
 
@@ -1439,7 +1394,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 		head = &c_node->mem_list[EXYNOS_DRM_OPS_DST];
 
 		if (list_empty(head)) {
-			DRM_DEBUG_KMS("%s:mem_list is empty.\n", __func__);
+			DRM_DEBUG_KMS("mem_list is empty.\n");
 			break;
 		}
 
@@ -1456,7 +1411,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 		head = &c_node->mem_list[EXYNOS_DRM_OPS_SRC];
 
 		if (list_empty(head)) {
-			DRM_DEBUG_KMS("%s:mem_list is empty.\n", __func__);
+			DRM_DEBUG_KMS("mem_list is empty.\n");
 			break;
 		}
 
@@ -1491,8 +1446,6 @@ void ipp_sched_cmd(struct work_struct *work)
 	struct drm_exynos_ipp_property *property;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	ippdrv = cmd_work->ippdrv;
 	if (!ippdrv) {
 		DRM_ERROR("invalid ippdrv list.\n");
@@ -1550,7 +1503,7 @@ void ipp_sched_cmd(struct work_struct *work)
 		break;
 	}
 
-	DRM_DEBUG_KMS("%s:ctrl[%d] done.\n", __func__, cmd_work->ctrl);
+	DRM_DEBUG_KMS("ctrl[%d] done.\n", cmd_work->ctrl);
 
 err_unlock:
 	mutex_unlock(&c_node->cmd_lock);
@@ -1571,8 +1524,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	int ret, i;
 
 	for_each_ipp_ops(i)
-		DRM_DEBUG_KMS("%s:%s buf_id[%d]\n", __func__,
-			i ? "dst" : "src", buf_id[i]);
+		DRM_DEBUG_KMS("%s buf_id[%d]\n", i ? "dst" : "src", buf_id[i]);
 
 	if (!drm_dev) {
 		DRM_ERROR("failed to get drm_dev.\n");
@@ -1585,12 +1537,12 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	}
 
 	if (list_empty(&c_node->event_list)) {
-		DRM_DEBUG_KMS("%s:event list is empty.\n", __func__);
+		DRM_DEBUG_KMS("event list is empty.\n");
 		return 0;
 	}
 
 	if (!ipp_check_mem_list(c_node)) {
-		DRM_DEBUG_KMS("%s:empty memory.\n", __func__);
+		DRM_DEBUG_KMS("empty memory.\n");
 		return 0;
 	}
 
@@ -1609,7 +1561,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 			}
 
 			tbuf_id[i] = m_node->buf_id;
-			DRM_DEBUG_KMS("%s:%s buf_id[%d]\n", __func__,
+			DRM_DEBUG_KMS("%s buf_id[%d]\n",
 				i ? "dst" : "src", tbuf_id[i]);
 
 			ret = ipp_put_mem_node(drm_dev, c_node, m_node);
@@ -1677,8 +1629,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	}
 
 	do_gettimeofday(&now);
-	DRM_DEBUG_KMS("%s:tv_sec[%ld]tv_usec[%ld]\n"
-		, __func__, now.tv_sec, now.tv_usec);
+	DRM_DEBUG_KMS("tv_sec[%ld]tv_usec[%ld]\n", now.tv_sec, now.tv_usec);
 	e->event.tv_sec = now.tv_sec;
 	e->event.tv_usec = now.tv_usec;
 	e->event.prop_id = property->prop_id;
@@ -1692,7 +1643,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	wake_up_interruptible(&e->base.file_priv->event_wait);
 	spin_unlock_irqrestore(&drm_dev->event_lock, flags);
 
-	DRM_DEBUG_KMS("%s:done cmd[%d]prop_id[%d]buf_id[%d]\n", __func__,
+	DRM_DEBUG_KMS("done cmd[%d]prop_id[%d]buf_id[%d]\n",
 		property->cmd, property->prop_id, tbuf_id[EXYNOS_DRM_OPS_DST]);
 
 	return 0;
@@ -1711,8 +1662,7 @@ void ipp_sched_event(struct work_struct *work)
 		return;
 	}
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__,
-		event_work->buf_id[EXYNOS_DRM_OPS_DST]);
+	DRM_DEBUG_KMS("buf_id[%d]\n", event_work->buf_id[EXYNOS_DRM_OPS_DST]);
 
 	ippdrv = event_work->ippdrv;
 	if (!ippdrv) {
@@ -1733,8 +1683,8 @@ void ipp_sched_event(struct work_struct *work)
 	 * or going out operations.
 	 */
 	if (c_node->state != IPP_STATE_START) {
-		DRM_DEBUG_KMS("%s:bypass state[%d]prop_id[%d]\n",
-			__func__, c_node->state, c_node->property.prop_id);
+		DRM_DEBUG_KMS("bypass state[%d]prop_id[%d]\n",
+			c_node->state, c_node->property.prop_id);
 		goto err_completion;
 	}
 
@@ -1759,8 +1709,6 @@ static int ipp_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 	struct exynos_drm_ippdrv *ippdrv;
 	int ret, count = 0;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* get ipp driver entry */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 		ippdrv->drm_dev = drm_dev;
@@ -1772,7 +1720,7 @@ static int ipp_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 			goto err_idr;
 		}
 
-		DRM_DEBUG_KMS("%s:count[%d]ippdrv[0x%x]ipp_id[%d]\n", __func__,
+		DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]ipp_id[%d]\n",
 			count++, (int)ippdrv, ippdrv->ipp_id);
 
 		if (ippdrv->ipp_id == 0) {
@@ -1816,8 +1764,6 @@ static void ipp_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
 	struct exynos_drm_ippdrv *ippdrv;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* get ipp driver entry */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 		if (is_drm_iommu_supported(drm_dev))
@@ -1834,8 +1780,6 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
 	struct exynos_drm_ipp_private *priv;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv) {
 		DRM_ERROR("failed to allocate priv.\n");
@@ -1846,7 +1790,7 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 
 	INIT_LIST_HEAD(&priv->event_list);
 
-	DRM_DEBUG_KMS("%s:done priv[0x%x]\n", __func__, (int)priv);
+	DRM_DEBUG_KMS("done priv[0x%x]\n", (int)priv);
 
 	return 0;
 }
@@ -1860,10 +1804,10 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 	struct drm_exynos_ipp_cmd_node *c_node, *tc_node;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s:for priv[0x%x]\n", __func__, (int)priv);
+	DRM_DEBUG_KMS("for priv[0x%x]\n", (int)priv);
 
 	if (list_empty(&exynos_drm_ippdrv_list)) {
-		DRM_DEBUG_KMS("%s:ippdrv_list is empty.\n", __func__);
+		DRM_DEBUG_KMS("ippdrv_list is empty.\n");
 		goto err_clear;
 	}
 
@@ -1873,8 +1817,8 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 
 		list_for_each_entry_safe(c_node, tc_node,
 			&ippdrv->cmd_list, list) {
-			DRM_DEBUG_KMS("%s:count[%d]ippdrv[0x%x]\n",
-				__func__, count++, (int)ippdrv);
+			DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]\n",
+				count++, (int)ippdrv);
 
 			if (c_node->priv == priv) {
 				/*
@@ -1913,8 +1857,6 @@ static int ipp_probe(struct platform_device *pdev)
 	if (!ctx)
 		return -ENOMEM;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	mutex_init(&ctx->ipp_lock);
 	mutex_init(&ctx->prop_lock);
 
@@ -1978,8 +1920,6 @@ static int ipp_remove(struct platform_device *pdev)
 {
 	struct ipp_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* unregister sub driver */
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
@@ -1999,7 +1939,7 @@ static int ipp_remove(struct platform_device *pdev)
 
 static int ipp_power_ctrl(struct ipp_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	return 0;
 }
@@ -2009,8 +1949,6 @@ static int ipp_suspend(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (pm_runtime_suspended(dev))
 		return 0;
 
@@ -2021,8 +1959,6 @@ static int ipp_resume(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!pm_runtime_suspended(dev))
 		return ipp_power_ctrl(ctx, true);
 
@@ -2035,8 +1971,6 @@ static int ipp_runtime_suspend(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return ipp_power_ctrl(ctx, false);
 }
 
@@ -2044,8 +1978,6 @@ static int ipp_runtime_resume(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return ipp_power_ctrl(ctx, true);
 }
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c
index 83efc66..6ee55e6 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c
@@ -81,8 +81,6 @@ int exynos_plane_mode_set(struct drm_plane *plane, struct drm_crtc *crtc,
 	int nr;
 	int i;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	nr = exynos_drm_fb_get_buf_cnt(fb);
 	for (i = 0; i < nr; i++) {
 		struct exynos_drm_gem_buf *buffer = exynos_drm_fb_buffer(fb, i);
@@ -159,8 +157,6 @@ void exynos_plane_dpms(struct drm_plane *plane, int mode)
 	struct exynos_plane *exynos_plane = to_exynos_plane(plane);
 	struct exynos_drm_overlay *overlay = &exynos_plane->overlay;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (mode == DRM_MODE_DPMS_ON) {
 		if (exynos_plane->enabled)
 			return;
@@ -189,8 +185,6 @@ exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 {
 	int ret;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	ret = exynos_plane_mode_set(plane, crtc, fb, crtc_x, crtc_y,
 			crtc_w, crtc_h, src_x >> 16, src_y >> 16,
 			src_w >> 16, src_h >> 16);
@@ -207,8 +201,6 @@ exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 static int exynos_disable_plane(struct drm_plane *plane)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	exynos_plane_dpms(plane, DRM_MODE_DPMS_OFF);
 
 	return 0;
@@ -218,8 +210,6 @@ static void exynos_plane_destroy(struct drm_plane *plane)
 {
 	struct exynos_plane *exynos_plane = to_exynos_plane(plane);
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	exynos_disable_plane(plane);
 	drm_plane_cleanup(plane);
 	kfree(exynos_plane);
@@ -233,8 +223,6 @@ static int exynos_plane_set_property(struct drm_plane *plane,
 	struct exynos_plane *exynos_plane = to_exynos_plane(plane);
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (property == dev_priv->plane_zpos_property) {
 		exynos_plane->overlay.zpos = val;
 		return 0;
@@ -256,8 +244,6 @@ static void exynos_plane_attach_zpos_property(struct drm_plane *plane)
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	prop = dev_priv->plane_zpos_property;
 	if (!prop) {
 		prop = drm_property_create_range(dev, 0, "zpos", 0,
@@ -277,8 +263,6 @@ struct drm_plane *exynos_plane_init(struct drm_device *dev,
 	struct exynos_plane *exynos_plane;
 	int err;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	exynos_plane = kzalloc(sizeof(struct exynos_plane), GFP_KERNEL);
 	if (!exynos_plane) {
 		DRM_ERROR("failed to allocate plane\n");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
index 9b6c709..427640a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
@@ -244,7 +244,7 @@ static int rotator_src_set_size(struct device *dev, int swap,
 	/* Get format */
 	fmt = rotator_reg_get_fmt(rot);
 	if (!rotator_check_reg_fmt(fmt)) {
-		DRM_ERROR("%s:invalid format.\n", __func__);
+		DRM_ERROR("invalid format.\n");
 		return -EINVAL;
 	}
 
@@ -287,7 +287,7 @@ static int rotator_src_set_addr(struct device *dev,
 		/* Get format */
 		fmt = rotator_reg_get_fmt(rot);
 		if (!rotator_check_reg_fmt(fmt)) {
-			DRM_ERROR("%s:invalid format.\n", __func__);
+			DRM_ERROR("invalid format.\n");
 			return -EINVAL;
 		}
 
@@ -381,7 +381,7 @@ static int rotator_dst_set_size(struct device *dev, int swap,
 	/* Get format */
 	fmt = rotator_reg_get_fmt(rot);
 	if (!rotator_check_reg_fmt(fmt)) {
-		DRM_ERROR("%s:invalid format.\n", __func__);
+		DRM_ERROR("invalid format.\n");
 		return -EINVAL;
 	}
 
@@ -422,7 +422,7 @@ static int rotator_dst_set_addr(struct device *dev,
 		/* Get format */
 		fmt = rotator_reg_get_fmt(rot);
 		if (!rotator_check_reg_fmt(fmt)) {
-			DRM_ERROR("%s:invalid format.\n", __func__);
+			DRM_ERROR("invalid format.\n");
 			return -EINVAL;
 		}
 
@@ -471,8 +471,6 @@ static int rotator_init_prop_list(struct exynos_drm_ippdrv *ippdrv)
 {
 	struct drm_exynos_ipp_prop_list *prop_list;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop_list = devm_kzalloc(ippdrv->dev, sizeof(*prop_list), GFP_KERNEL);
 	if (!prop_list) {
 		DRM_ERROR("failed to alloc property list.\n");
@@ -502,7 +500,7 @@ static inline bool rotator_check_drm_fmt(u32 fmt)
 	case DRM_FORMAT_NV12:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:not support format\n", __func__);
+		DRM_DEBUG_KMS("not support format\n");
 		return false;
 	}
 }
@@ -516,7 +514,7 @@ static inline bool rotator_check_drm_flip(enum drm_exynos_flip flip)
 	case EXYNOS_DRM_FLIP_BOTH:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return false;
 	}
 }
@@ -536,19 +534,18 @@ static int rotator_ippdrv_check_property(struct device *dev,
 
 	/* Check format configuration */
 	if (src_config->fmt != dst_config->fmt) {
-		DRM_DEBUG_KMS("%s:not support csc feature\n", __func__);
+		DRM_DEBUG_KMS("not support csc feature\n");
 		return -EINVAL;
 	}
 
 	if (!rotator_check_drm_fmt(dst_config->fmt)) {
-		DRM_DEBUG_KMS("%s:invalid format\n", __func__);
+		DRM_DEBUG_KMS("invalid format\n");
 		return -EINVAL;
 	}
 
 	/* Check transform configuration */
 	if (src_config->degree != EXYNOS_DRM_DEGREE_0) {
-		DRM_DEBUG_KMS("%s:not support source-side rotation\n",
-			__func__);
+		DRM_DEBUG_KMS("not support source-side rotation\n");
 		return -EINVAL;
 	}
 
@@ -561,51 +558,47 @@ static int rotator_ippdrv_check_property(struct device *dev,
 		/* No problem */
 		break;
 	default:
-		DRM_DEBUG_KMS("%s:invalid degree\n", __func__);
+		DRM_DEBUG_KMS("invalid degree\n");
 		return -EINVAL;
 	}
 
 	if (src_config->flip != EXYNOS_DRM_FLIP_NONE) {
-		DRM_DEBUG_KMS("%s:not support source-side flip\n", __func__);
+		DRM_DEBUG_KMS("not support source-side flip\n");
 		return -EINVAL;
 	}
 
 	if (!rotator_check_drm_flip(dst_config->flip)) {
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return -EINVAL;
 	}
 
 	/* Check size configuration */
 	if ((src_pos->x + src_pos->w > src_sz->hsize) ||
 		(src_pos->y + src_pos->h > src_sz->vsize)) {
-		DRM_DEBUG_KMS("%s:out of source buffer bound\n", __func__);
+		DRM_DEBUG_KMS("out of source buffer bound\n");
 		return -EINVAL;
 	}
 
 	if (swap) {
 		if ((dst_pos->x + dst_pos->h > dst_sz->vsize) ||
 			(dst_pos->y + dst_pos->w > dst_sz->hsize)) {
-			DRM_DEBUG_KMS("%s:out of destination buffer bound\n",
-				__func__);
+			DRM_DEBUG_KMS("out of destination buffer bound\n");
 			return -EINVAL;
 		}
 
 		if ((src_pos->w != dst_pos->h) || (src_pos->h != dst_pos->w)) {
-			DRM_DEBUG_KMS("%s:not support scale feature\n",
-				__func__);
+			DRM_DEBUG_KMS("not support scale feature\n");
 			return -EINVAL;
 		}
 	} else {
 		if ((dst_pos->x + dst_pos->w > dst_sz->hsize) ||
 			(dst_pos->y + dst_pos->h > dst_sz->vsize)) {
-			DRM_DEBUG_KMS("%s:out of destination buffer bound\n",
-				__func__);
+			DRM_DEBUG_KMS("out of destination buffer bound\n");
 			return -EINVAL;
 		}
 
 		if ((src_pos->w != dst_pos->w) || (src_pos->h != dst_pos->h)) {
-			DRM_DEBUG_KMS("%s:not support scale feature\n",
-				__func__);
+			DRM_DEBUG_KMS("not support scale feature\n");
 			return -EINVAL;
 		}
 	}
@@ -693,7 +686,7 @@ static int rotator_probe(struct platform_device *pdev)
 		goto err_ippdrv_register;
 	}
 
-	DRM_DEBUG_KMS("%s:ippdrv[0x%x]\n", __func__, (int)ippdrv);
+	DRM_DEBUG_KMS("ippdrv[0x%x]\n", (int)ippdrv);
 
 	platform_set_drvdata(pdev, rot);
 
@@ -752,8 +745,6 @@ static struct platform_device_id rotator_driver_ids[] = {
 
 static int rotator_clk_crtl(struct rot_context *rot, bool enable)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (enable) {
 		clk_enable(rot->clock);
 		rot->suspended = false;
@@ -771,8 +762,6 @@ static int rotator_suspend(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (pm_runtime_suspended(dev))
 		return 0;
 
@@ -783,8 +772,6 @@ static int rotator_resume(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!pm_runtime_suspended(dev))
 		return rotator_clk_crtl(rot, true);
 
@@ -797,8 +784,6 @@ static int rotator_runtime_suspend(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return  rotator_clk_crtl(rot, false);
 }
 
@@ -806,8 +791,6 @@ static int rotator_runtime_resume(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return  rotator_clk_crtl(rot, true);
 }
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
index 24376c1..41cc74d 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_vidi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
@@ -89,8 +89,6 @@ static bool vidi_display_is_connected(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * connection request would come from user side
 	 * to do hotplug through specific ioctl.
@@ -105,8 +103,6 @@ static struct edid *vidi_get_edid(struct device *dev,
 	struct edid *edid;
 	int edid_len;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * the edid data comes from user side and it would be set
 	 * to ctx->raw_edid through specific ioctl.
@@ -128,17 +124,13 @@ static struct edid *vidi_get_edid(struct device *dev,
 
 static void *vidi_get_panel(struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return NULL;
 }
 
-static int vidi_check_timing(struct device *dev, void *timing)
+static int vidi_check_mode(struct device *dev, struct drm_display_mode *mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return 0;
@@ -146,8 +138,6 @@ static int vidi_check_timing(struct device *dev, void *timing)
 
 static int vidi_display_power_on(struct device *dev, int mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO */
 
 	return 0;
@@ -158,7 +148,7 @@ static struct exynos_drm_display_ops vidi_display_ops = {
 	.is_connected = vidi_display_is_connected,
 	.get_edid = vidi_get_edid,
 	.get_panel = vidi_get_panel,
-	.check_timing = vidi_check_timing,
+	.check_mode = vidi_check_mode,
 	.power_on = vidi_display_power_on,
 };
 
@@ -166,7 +156,7 @@ static void vidi_dpms(struct device *subdrv_dev, int mode)
 {
 	struct vidi_context *ctx = get_vidi_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode);
+	DRM_DEBUG_KMS("%d\n", mode);
 
 	mutex_lock(&ctx->lock);
 
@@ -196,8 +186,6 @@ static void vidi_apply(struct device *subdrv_dev)
 	struct vidi_win_data *win_data;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	for (i = 0; i < WINDOWS_NR; i++) {
 		win_data = &ctx->win_data[i];
 		if (win_data->enabled && (ovl_ops && ovl_ops->commit))
@@ -212,8 +200,6 @@ static void vidi_commit(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 }
@@ -222,8 +208,6 @@ static int vidi_enable_vblank(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return -EPERM;
 
@@ -246,8 +230,6 @@ static void vidi_disable_vblank(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
@@ -271,8 +253,6 @@ static void vidi_win_mode_set(struct device *dev,
 	int win;
 	unsigned long offset;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!overlay) {
 		dev_err(dev, "overlay is NULL\n");
 		return;
@@ -282,7 +262,7 @@ static void vidi_win_mode_set(struct device *dev,
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	offset = overlay->fb_x * (overlay->bpp >> 3);
@@ -324,15 +304,13 @@ static void vidi_win_commit(struct device *dev, int zpos)
 	struct vidi_win_data *win_data;
 	int win = zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
@@ -351,12 +329,10 @@ static void vidi_win_disable(struct device *dev, int zpos)
 	struct vidi_win_data *win_data;
 	int win = zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
@@ -407,8 +383,6 @@ static void vidi_fake_vblank_handler(struct work_struct *work)
 
 static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * enable drm irq mode.
 	 * - with irq_enabled = 1, we can use the vblank feature.
@@ -431,8 +405,6 @@ static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 
 static void vidi_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 }
 
@@ -441,11 +413,6 @@ static int vidi_power_on(struct vidi_context *ctx, bool enable)
 	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
 	struct device *dev = subdrv->dev;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (enable != false && enable != true)
-		return -EINVAL;
-
 	if (enable) {
 		ctx->suspended = false;
 
@@ -483,8 +450,6 @@ static int vidi_store_connection(struct device *dev,
 	struct vidi_context *ctx = get_vidi_context(dev);
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	ret = kstrtoint(buf, 0, &ctx->connected);
 	if (ret)
 		return ret;
@@ -522,8 +487,6 @@ int vidi_connection_ioctl(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_vidi_connection *vidi = data;
 	int edid_len;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!vidi) {
 		DRM_DEBUG_KMS("user data for vidi is null.\n");
 		return -EINVAL;
@@ -592,8 +555,6 @@ static int vidi_probe(struct platform_device *pdev)
 	struct exynos_drm_subdrv *subdrv;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
 		return -ENOMEM;
@@ -625,8 +586,6 @@ static int vidi_remove(struct platform_device *pdev)
 {
 	struct vidi_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
 	if (ctx->raw_edid != (struct edid *)fake_edid_info) {
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index fd1426d..62ef597 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -83,6 +83,7 @@ struct hdmi_resources {
 	struct clk			*sclk_pixel;
 	struct clk			*sclk_hdmiphy;
 	struct clk			*hdmiphy;
+	struct clk			*mout_hdmi;
 	struct regulator_bulk_data	*regul_bulk;
 	int				regul_count;
 };
@@ -689,8 +690,6 @@ static void hdmi_reg_infoframe(struct hdmi_context *hdata,
 	u32 mod;
 	u32 vic;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mod = hdmi_reg_read(hdata, HDMI_MODE_SEL);
 	if (hdata->dvi_mode) {
 		hdmi_reg_writeb(hdata, HDMI_VSI_CON,
@@ -755,8 +754,6 @@ static struct edid *hdmi_get_edid(void *ctx, struct drm_connector *connector)
 	struct edid *raw_edid;
 	struct hdmi_context *hdata = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (!hdata->ddc_port)
 		return ERR_PTR(-ENODEV);
 
@@ -777,8 +774,6 @@ static int hdmi_find_phy_conf(struct hdmi_context *hdata, u32 pixel_clock)
 	const struct hdmiphy_config *confs;
 	int count, i;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (hdata->type == HDMI_TYPE13) {
 		confs = hdmiphy_v13_configs;
 		count = ARRAY_SIZE(hdmiphy_v13_configs);
@@ -796,18 +791,17 @@ static int hdmi_find_phy_conf(struct hdmi_context *hdata, u32 pixel_clock)
 	return -EINVAL;
 }
 
-static int hdmi_check_timing(void *ctx, struct fb_videomode *timing)
+static int hdmi_check_mode(void *ctx, struct drm_display_mode *mode)
 {
 	struct hdmi_context *hdata = ctx;
 	int ret;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%d clock=%d\n",
+		mode->hdisplay, mode->vdisplay, mode->vrefresh,
+		(mode->flags & DRM_MODE_FLAG_INTERLACE) ? true :
+		false, mode->clock * 1000);
 
-	DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", timing->xres,
-			timing->yres, timing->refresh,
-			timing->vmode);
-
-	ret = hdmi_find_phy_conf(hdata, timing->pixclock);
+	ret = hdmi_find_phy_conf(hdata, mode->clock * 1000);
 	if (ret < 0)
 		return ret;
 	return 0;
@@ -1042,7 +1036,7 @@ static void hdmi_conf_init(struct hdmi_context *hdata)
 	}
 }
 
-static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
+static void hdmi_v13_mode_apply(struct hdmi_context *hdata)
 {
 	const struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v13_conf.tg;
 	const struct hdmi_v13_core_regs *core =
@@ -1118,9 +1112,9 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 		hdmi_regs_dump(hdata, "timing apply");
 	}
 
-	clk_disable(hdata->res.sclk_hdmi);
-	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy);
-	clk_enable(hdata->res.sclk_hdmi);
+	clk_disable_unprepare(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.mout_hdmi, hdata->res.sclk_hdmiphy);
+	clk_prepare_enable(hdata->res.sclk_hdmi);
 
 	/* enable HDMI and timing generator */
 	hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN);
@@ -1131,7 +1125,7 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN);
 }
 
-static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
+static void hdmi_v14_mode_apply(struct hdmi_context *hdata)
 {
 	const struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v14_conf.tg;
 	const struct hdmi_v14_core_regs *core =
@@ -1285,9 +1279,9 @@ static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 		hdmi_regs_dump(hdata, "timing apply");
 	}
 
-	clk_disable(hdata->res.sclk_hdmi);
-	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy);
-	clk_enable(hdata->res.sclk_hdmi);
+	clk_disable_unprepare(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.mout_hdmi, hdata->res.sclk_hdmiphy);
+	clk_prepare_enable(hdata->res.sclk_hdmi);
 
 	/* enable HDMI and timing generator */
 	hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN);
@@ -1298,12 +1292,12 @@ static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN);
 }
 
-static void hdmi_timing_apply(struct hdmi_context *hdata)
+static void hdmi_mode_apply(struct hdmi_context *hdata)
 {
 	if (hdata->type == HDMI_TYPE13)
-		hdmi_v13_timing_apply(hdata);
+		hdmi_v13_mode_apply(hdata);
 	else
-		hdmi_v14_timing_apply(hdata);
+		hdmi_v14_mode_apply(hdata);
 }
 
 static void hdmiphy_conf_reset(struct hdmi_context *hdata)
@@ -1311,9 +1305,9 @@ static void hdmiphy_conf_reset(struct hdmi_context *hdata)
 	u8 buffer[2];
 	u32 reg;
 
-	clk_disable(hdata->res.sclk_hdmi);
-	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_pixel);
-	clk_enable(hdata->res.sclk_hdmi);
+	clk_disable_unprepare(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.mout_hdmi, hdata->res.sclk_pixel);
+	clk_prepare_enable(hdata->res.sclk_hdmi);
 
 	/* operation mode */
 	buffer[0] = 0x1f;
@@ -1336,8 +1330,6 @@ static void hdmiphy_conf_reset(struct hdmi_context *hdata)
 
 static void hdmiphy_poweron(struct hdmi_context *hdata)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (hdata->type == HDMI_TYPE14)
 		hdmi_reg_writemask(hdata, HDMI_PHY_CON_0, 0,
 			HDMI_PHY_POWER_OFF_EN);
@@ -1345,8 +1337,6 @@ static void hdmiphy_poweron(struct hdmi_context *hdata)
 
 static void hdmiphy_poweroff(struct hdmi_context *hdata)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (hdata->type == HDMI_TYPE14)
 		hdmi_reg_writemask(hdata, HDMI_PHY_CON_0, ~0,
 			HDMI_PHY_POWER_OFF_EN);
@@ -1410,8 +1400,6 @@ static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 
 static void hdmi_conf_apply(struct hdmi_context *hdata)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	hdmiphy_conf_reset(hdata);
 	hdmiphy_conf_apply(hdata);
 
@@ -1423,7 +1411,7 @@ static void hdmi_conf_apply(struct hdmi_context *hdata)
 	hdmi_audio_init(hdata);
 
 	/* setting core registers */
-	hdmi_timing_apply(hdata);
+	hdmi_mode_apply(hdata);
 	hdmi_audio_control(hdata, true);
 
 	hdmi_regs_dump(hdata, "start");
@@ -1569,8 +1557,7 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 			(m->vsync_start - m->vdisplay) / 2);
 		hdmi_set_reg(core->v2_blank, 2, m->vtotal / 2);
 		hdmi_set_reg(core->v1_blank, 2, (m->vtotal - m->vdisplay) / 2);
-		hdmi_set_reg(core->v_blank_f0, 2, (m->vtotal +
-			((m->vsync_end - m->vsync_start) * 4) + 5) / 2);
+		hdmi_set_reg(core->v_blank_f0, 2, m->vtotal - m->vdisplay / 2);
 		hdmi_set_reg(core->v_blank_f1, 2, m->vtotal);
 		hdmi_set_reg(core->v_sync_line_aft_2, 2, (m->vtotal / 2) + 7);
 		hdmi_set_reg(core->v_sync_line_aft_1, 2, (m->vtotal / 2) + 2);
@@ -1580,7 +1567,10 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 			(m->htotal / 2) + (m->hsync_start - m->hdisplay));
 		hdmi_set_reg(tg->vact_st, 2, (m->vtotal - m->vdisplay) / 2);
 		hdmi_set_reg(tg->vact_sz, 2, m->vdisplay / 2);
-		hdmi_set_reg(tg->vact_st2, 2, 0x249);/* Reset value + 1*/
+		hdmi_set_reg(tg->vact_st2, 2, m->vtotal - m->vdisplay / 2);
+		hdmi_set_reg(tg->vsync2, 2, (m->vtotal / 2) + 1);
+		hdmi_set_reg(tg->vsync_bot_hdmi, 2, (m->vtotal / 2) + 1);
+		hdmi_set_reg(tg->field_bot_hdmi, 2, (m->vtotal / 2) + 1);
 		hdmi_set_reg(tg->vact_st3, 2, 0x0);
 		hdmi_set_reg(tg->vact_st4, 2, 0x0);
 	} else {
@@ -1602,6 +1592,9 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 		hdmi_set_reg(tg->vact_st2, 2, 0x248); /* Reset value */
 		hdmi_set_reg(tg->vact_st3, 2, 0x47b); /* Reset value */
 		hdmi_set_reg(tg->vact_st4, 2, 0x6ae); /* Reset value */
+		hdmi_set_reg(tg->vsync2, 2, 0x233); /* Reset value */
+		hdmi_set_reg(tg->vsync_bot_hdmi, 2, 0x233); /* Reset value */
+		hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
 	}
 
 	/* Following values & calculations are same irrespective of mode type */
@@ -1633,22 +1626,19 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 	hdmi_set_reg(tg->hact_sz, 2, m->hdisplay);
 	hdmi_set_reg(tg->v_fsz, 2, m->vtotal);
 	hdmi_set_reg(tg->vsync, 2, 0x1);
-	hdmi_set_reg(tg->vsync2, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->field_chg, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->vsync_top_hdmi, 2, 0x1); /* Reset value */
-	hdmi_set_reg(tg->vsync_bot_hdmi, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->field_top_hdmi, 2, 0x1); /* Reset value */
-	hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->tg_3d, 1, 0x0);
 }
 
-static void hdmi_mode_set(void *ctx, void *mode)
+static void hdmi_mode_set(void *ctx, struct drm_display_mode *mode)
 {
 	struct hdmi_context *hdata = ctx;
 	struct drm_display_mode *m = mode;
 
-	DRM_DEBUG_KMS("[%s]: xres=%d, yres=%d, refresh=%d, intl=%s\n",
-		__func__, m->hdisplay, m->vdisplay,
+	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%s\n",
+		m->hdisplay, m->vdisplay,
 		m->vrefresh, (m->flags & DRM_MODE_FLAG_INTERLACE) ?
 		"INTERLACED" : "PROGERESSIVE");
 
@@ -1661,8 +1651,6 @@ static void hdmi_mode_set(void *ctx, void *mode)
 static void hdmi_get_max_resol(void *ctx, unsigned int *width,
 					unsigned int *height)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	*width = MAX_WIDTH;
 	*height = MAX_HEIGHT;
 }
@@ -1671,8 +1659,6 @@ static void hdmi_commit(void *ctx)
 {
 	struct hdmi_context *hdata = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&hdata->hdmi_mutex);
 	if (!hdata->powered) {
 		mutex_unlock(&hdata->hdmi_mutex);
@@ -1687,8 +1673,6 @@ static void hdmi_poweron(struct hdmi_context *hdata)
 {
 	struct hdmi_resources *res = &hdata->res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&hdata->hdmi_mutex);
 	if (hdata->powered) {
 		mutex_unlock(&hdata->hdmi_mutex);
@@ -1699,10 +1683,12 @@ static void hdmi_poweron(struct hdmi_context *hdata)
 
 	mutex_unlock(&hdata->hdmi_mutex);
 
-	regulator_bulk_enable(res->regul_count, res->regul_bulk);
-	clk_enable(res->hdmiphy);
-	clk_enable(res->hdmi);
-	clk_enable(res->sclk_hdmi);
+	if (regulator_bulk_enable(res->regul_count, res->regul_bulk))
+		DRM_DEBUG_KMS("failed to enable regulator bulk\n");
+
+	clk_prepare_enable(res->hdmiphy);
+	clk_prepare_enable(res->hdmi);
+	clk_prepare_enable(res->sclk_hdmi);
 
 	hdmiphy_poweron(hdata);
 }
@@ -1711,8 +1697,6 @@ static void hdmi_poweroff(struct hdmi_context *hdata)
 {
 	struct hdmi_resources *res = &hdata->res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&hdata->hdmi_mutex);
 	if (!hdata->powered)
 		goto out;
@@ -1725,9 +1709,9 @@ static void hdmi_poweroff(struct hdmi_context *hdata)
 	hdmiphy_conf_reset(hdata);
 	hdmiphy_poweroff(hdata);
 
-	clk_disable(res->sclk_hdmi);
-	clk_disable(res->hdmi);
-	clk_disable(res->hdmiphy);
+	clk_disable_unprepare(res->sclk_hdmi);
+	clk_disable_unprepare(res->hdmi);
+	clk_disable_unprepare(res->hdmiphy);
 	regulator_bulk_disable(res->regul_count, res->regul_bulk);
 
 	mutex_lock(&hdata->hdmi_mutex);
@@ -1742,7 +1726,7 @@ static void hdmi_dpms(void *ctx, int mode)
 {
 	struct hdmi_context *hdata = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s mode %d\n", __LINE__, __func__, mode);
+	DRM_DEBUG_KMS("mode %d\n", mode);
 
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
@@ -1765,7 +1749,7 @@ static struct exynos_hdmi_ops hdmi_ops = {
 	/* display */
 	.is_connected	= hdmi_is_connected,
 	.get_edid	= hdmi_get_edid,
-	.check_timing	= hdmi_check_timing,
+	.check_mode	= hdmi_check_mode,
 
 	/* manager */
 	.mode_set	= hdmi_mode_set,
@@ -1831,8 +1815,13 @@ static int hdmi_resources_init(struct hdmi_context *hdata)
 		DRM_ERROR("failed to get clock 'hdmiphy'\n");
 		goto fail;
 	}
+	res->mout_hdmi = devm_clk_get(dev, "mout_hdmi");
+	if (IS_ERR(res->mout_hdmi)) {
+		DRM_ERROR("failed to get clock 'mout_hdmi'\n");
+		goto fail;
+	}
 
-	clk_set_parent(res->sclk_hdmi, res->sclk_pixel);
+	clk_set_parent(res->mout_hdmi, res->sclk_pixel);
 
 	res->regul_bulk = devm_kzalloc(dev, ARRAY_SIZE(supply) *
 		sizeof(res->regul_bulk[0]), GFP_KERNEL);
@@ -1877,7 +1866,6 @@ static struct s5p_hdmi_platform_data *drm_hdmi_dt_parse_pdata
 {
 	struct device_node *np = dev->of_node;
 	struct s5p_hdmi_platform_data *pd;
-	enum of_gpio_flags flags;
 	u32 value;
 
 	pd = devm_kzalloc(dev, sizeof(*pd), GFP_KERNEL);
@@ -1891,7 +1879,7 @@ static struct s5p_hdmi_platform_data *drm_hdmi_dt_parse_pdata
 		goto err_data;
 	}
 
-	pd->hpd_gpio = of_get_named_gpio_flags(np, "hpd-gpio", 0, &flags);
+	pd->hpd_gpio = of_get_named_gpio(np, "hpd-gpio", 0);
 
 	return pd;
 
@@ -1930,6 +1918,9 @@ static struct of_device_id hdmi_match_types[] = {
 		.compatible = "samsung,exynos5-hdmi",
 		.data	= (void	*)HDMI_TYPE14,
 	}, {
+		.compatible = "samsung,exynos4212-hdmi",
+		.data	= (void	*)HDMI_TYPE14,
+	}, {
 		/* end node */
 	}
 };
@@ -1944,8 +1935,6 @@ static int hdmi_probe(struct platform_device *pdev)
 	struct resource *res;
 	int ret;
 
-	DRM_DEBUG_KMS("[%d]\n", __LINE__);
-
 	if (dev->of_node) {
 		pdata = drm_hdmi_dt_parse_pdata(dev);
 		if (IS_ERR(pdata)) {
@@ -2071,8 +2060,6 @@ static int hdmi_remove(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	pm_runtime_disable(dev);
 
 	/* hdmiphy i2c driver */
@@ -2089,8 +2076,6 @@ static int hdmi_suspend(struct device *dev)
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	disable_irq(hdata->irq);
 
 	hdata->hpd = false;
@@ -2098,7 +2083,7 @@ static int hdmi_suspend(struct device *dev)
 		drm_helper_hpd_irq_event(ctx->drm_dev);
 
 	if (pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already suspended\n", __func__);
+		DRM_DEBUG_KMS("Already suspended\n");
 		return 0;
 	}
 
@@ -2112,14 +2097,12 @@ static int hdmi_resume(struct device *dev)
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	hdata->hpd = gpio_get_value(hdata->hpd_gpio);
 
 	enable_irq(hdata->irq);
 
 	if (!pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already resumed\n", __func__);
+		DRM_DEBUG_KMS("Already resumed\n");
 		return 0;
 	}
 
@@ -2134,7 +2117,6 @@ static int hdmi_runtime_suspend(struct device *dev)
 {
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
 	hdmi_poweroff(hdata);
 
@@ -2145,7 +2127,6 @@ static int hdmi_runtime_resume(struct device *dev)
 {
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
 	hdmi_poweron(hdata);
 
diff --git a/drivers/gpu/drm/exynos/exynos_hdmiphy.c b/drivers/gpu/drm/exynos/exynos_hdmiphy.c
index ea49d13..ef04255 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmiphy.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmiphy.c
@@ -51,6 +51,10 @@ static struct of_device_id hdmiphy_match_types[] = {
 	{
 		.compatible = "samsung,exynos5-hdmiphy",
 	}, {
+		.compatible = "samsung,exynos4210-hdmiphy",
+	}, {
+		.compatible = "samsung,exynos4212-hdmiphy",
+	}, {
 		/* end node */
 	}
 };
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index 7c197d38..42ffb71 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -78,6 +78,7 @@ struct mixer_resources {
 enum mixer_version_id {
 	MXR_VER_0_0_0_16,
 	MXR_VER_16_0_33_0,
+	MXR_VER_128_0_0_184,
 };
 
 struct mixer_context {
@@ -283,17 +284,19 @@ static void mixer_cfg_scan(struct mixer_context *ctx, unsigned int height)
 	val = (ctx->interlace ? MXR_CFG_SCAN_INTERLACE :
 				MXR_CFG_SCAN_PROGRASSIVE);
 
-	/* choosing between porper HD and SD mode */
-	if (height <= 480)
-		val |= MXR_CFG_SCAN_NTSC | MXR_CFG_SCAN_SD;
-	else if (height <= 576)
-		val |= MXR_CFG_SCAN_PAL | MXR_CFG_SCAN_SD;
-	else if (height <= 720)
-		val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
-	else if (height <= 1080)
-		val |= MXR_CFG_SCAN_HD_1080 | MXR_CFG_SCAN_HD;
-	else
-		val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
+	if (ctx->mxr_ver != MXR_VER_128_0_0_184) {
+		/* choosing between proper HD and SD mode */
+		if (height <= 480)
+			val |= MXR_CFG_SCAN_NTSC | MXR_CFG_SCAN_SD;
+		else if (height <= 576)
+			val |= MXR_CFG_SCAN_PAL | MXR_CFG_SCAN_SD;
+		else if (height <= 720)
+			val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
+		else if (height <= 1080)
+			val |= MXR_CFG_SCAN_HD_1080 | MXR_CFG_SCAN_HD;
+		else
+			val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
+	}
 
 	mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_SCAN_MASK);
 }
@@ -376,7 +379,7 @@ static void vp_video_buffer(struct mixer_context *ctx, int win)
 	unsigned long flags;
 	struct hdmi_win_data *win_data;
 	unsigned int x_ratio, y_ratio;
-	unsigned int buf_num;
+	unsigned int buf_num = 1;
 	dma_addr_t luma_addr[2], chroma_addr[2];
 	bool tiled_mode = false;
 	bool crcb_mode = false;
@@ -557,6 +560,14 @@ static void mixer_graph_buffer(struct mixer_context *ctx, int win)
 	/* setup geometry */
 	mixer_reg_write(res, MXR_GRAPHIC_SPAN(win), win_data->fb_width);
 
+	/* setup display size */
+	if (ctx->mxr_ver == MXR_VER_128_0_0_184 &&
+		win == MIXER_DEFAULT_WIN) {
+		val  = MXR_MXR_RES_HEIGHT(win_data->fb_height);
+		val |= MXR_MXR_RES_WIDTH(win_data->fb_width);
+		mixer_reg_write(res, MXR_RESOLUTION, val);
+	}
+
 	val  = MXR_GRP_WH_WIDTH(win_data->crtc_width);
 	val |= MXR_GRP_WH_HEIGHT(win_data->crtc_height);
 	val |= MXR_GRP_WH_H_SCALE(x_ratio);
@@ -581,7 +592,8 @@ static void mixer_graph_buffer(struct mixer_context *ctx, int win)
 	mixer_cfg_layer(ctx, win, true);
 
 	/* layer update mandatory for mixer 16.0.33.0 */
-	if (ctx->mxr_ver == MXR_VER_16_0_33_0)
+	if (ctx->mxr_ver == MXR_VER_16_0_33_0 ||
+		ctx->mxr_ver == MXR_VER_128_0_0_184)
 		mixer_layer_update(ctx);
 
 	mixer_run(ctx);
@@ -696,8 +708,6 @@ static int mixer_enable_vblank(void *ctx, int pipe)
 	struct mixer_context *mixer_ctx = ctx;
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mixer_ctx->pipe = pipe;
 
 	/* enable vsync interrupt */
@@ -712,8 +722,6 @@ static void mixer_disable_vblank(void *ctx)
 	struct mixer_context *mixer_ctx = ctx;
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	/* disable vsync interrupt */
 	mixer_reg_writemask(res, MXR_INT_EN, 0, MXR_INT_EN_VSYNC);
 }
@@ -725,8 +733,6 @@ static void mixer_win_mode_set(void *ctx,
 	struct hdmi_win_data *win_data;
 	int win;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (!overlay) {
 		DRM_ERROR("overlay is NULL\n");
 		return;
@@ -742,7 +748,7 @@ static void mixer_win_mode_set(void *ctx,
 	if (win == DEFAULT_ZPOS)
 		win = MIXER_DEFAULT_WIN;
 
-	if (win < 0 || win > MIXER_WIN_NR) {
+	if (win < 0 || win >= MIXER_WIN_NR) {
 		DRM_ERROR("mixer window[%d] is wrong\n", win);
 		return;
 	}
@@ -776,7 +782,7 @@ static void mixer_win_commit(void *ctx, int win)
 {
 	struct mixer_context *mixer_ctx = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s, win: %d\n", __LINE__, __func__, win);
+	DRM_DEBUG_KMS("win: %d\n", win);
 
 	mutex_lock(&mixer_ctx->mixer_mutex);
 	if (!mixer_ctx->powered) {
@@ -799,7 +805,7 @@ static void mixer_win_disable(void *ctx, int win)
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 	unsigned long flags;
 
-	DRM_DEBUG_KMS("[%d] %s, win: %d\n", __LINE__, __func__, win);
+	DRM_DEBUG_KMS("win: %d\n", win);
 
 	mutex_lock(&mixer_ctx->mixer_mutex);
 	if (!mixer_ctx->powered) {
@@ -820,17 +826,21 @@ static void mixer_win_disable(void *ctx, int win)
 	mixer_ctx->win_data[win].enabled = false;
 }
 
-static int mixer_check_timing(void *ctx, struct fb_videomode *timing)
+static int mixer_check_mode(void *ctx, struct drm_display_mode *mode)
 {
+	struct mixer_context *mixer_ctx = ctx;
 	u32 w, h;
 
-	w = timing->xres;
-	h = timing->yres;
+	w = mode->hdisplay;
+	h = mode->vdisplay;
 
-	DRM_DEBUG_KMS("%s : xres=%d, yres=%d, refresh=%d, intl=%d\n",
-		__func__, timing->xres, timing->yres,
-		timing->refresh, (timing->vmode &
-		FB_VMODE_INTERLACED) ? true : false);
+	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%d\n",
+		mode->hdisplay, mode->vdisplay, mode->vrefresh,
+		(mode->flags & DRM_MODE_FLAG_INTERLACE) ? 1 : 0);
+
+	if (mixer_ctx->mxr_ver == MXR_VER_0_0_0_16 ||
+		mixer_ctx->mxr_ver == MXR_VER_128_0_0_184)
+		return 0;
 
 	if ((w >= 464 && w <= 720 && h >= 261 && h <= 576) ||
 		(w >= 1024 && w <= 1280 && h >= 576 && h <= 720) ||
@@ -891,8 +901,6 @@ static void mixer_poweron(struct mixer_context *ctx)
 {
 	struct mixer_resources *res = &ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&ctx->mixer_mutex);
 	if (ctx->powered) {
 		mutex_unlock(&ctx->mixer_mutex);
@@ -901,10 +909,10 @@ static void mixer_poweron(struct mixer_context *ctx)
 	ctx->powered = true;
 	mutex_unlock(&ctx->mixer_mutex);
 
-	clk_enable(res->mixer);
+	clk_prepare_enable(res->mixer);
 	if (ctx->vp_enabled) {
-		clk_enable(res->vp);
-		clk_enable(res->sclk_mixer);
+		clk_prepare_enable(res->vp);
+		clk_prepare_enable(res->sclk_mixer);
 	}
 
 	mixer_reg_write(res, MXR_INT_EN, ctx->int_en);
@@ -917,8 +925,6 @@ static void mixer_poweroff(struct mixer_context *ctx)
 {
 	struct mixer_resources *res = &ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&ctx->mixer_mutex);
 	if (!ctx->powered)
 		goto out;
@@ -928,10 +934,10 @@ static void mixer_poweroff(struct mixer_context *ctx)
 
 	ctx->int_en = mixer_reg_read(res, MXR_INT_EN);
 
-	clk_disable(res->mixer);
+	clk_disable_unprepare(res->mixer);
 	if (ctx->vp_enabled) {
-		clk_disable(res->vp);
-		clk_disable(res->sclk_mixer);
+		clk_disable_unprepare(res->vp);
+		clk_disable_unprepare(res->sclk_mixer);
 	}
 
 	mutex_lock(&ctx->mixer_mutex);
@@ -945,8 +951,6 @@ static void mixer_dpms(void *ctx, int mode)
 {
 	struct mixer_context *mixer_ctx = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
 		if (pm_runtime_suspended(mixer_ctx->dev))
@@ -978,7 +982,7 @@ static struct exynos_mixer_ops mixer_ops = {
 	.win_disable		= mixer_win_disable,
 
 	/* display */
-	.check_timing		= mixer_check_timing,
+	.check_mode		= mixer_check_mode,
 };
 
 static irqreturn_t mixer_irq_handler(int irq, void *arg)
@@ -1128,12 +1132,17 @@ static int vp_resources_init(struct exynos_drm_hdmi_context *ctx,
 	return 0;
 }
 
-static struct mixer_drv_data exynos5_mxr_drv_data = {
+static struct mixer_drv_data exynos5420_mxr_drv_data = {
+	.version = MXR_VER_128_0_0_184,
+	.is_vp_enabled = 0,
+};
+
+static struct mixer_drv_data exynos5250_mxr_drv_data = {
 	.version = MXR_VER_16_0_33_0,
 	.is_vp_enabled = 0,
 };
 
-static struct mixer_drv_data exynos4_mxr_drv_data = {
+static struct mixer_drv_data exynos4210_mxr_drv_data = {
 	.version = MXR_VER_0_0_0_16,
 	.is_vp_enabled = 1,
 };
@@ -1141,10 +1150,10 @@ static struct mixer_drv_data exynos4_mxr_drv_data = {
 static struct platform_device_id mixer_driver_types[] = {
 	{
 		.name		= "s5p-mixer",
-		.driver_data	= (unsigned long)&exynos4_mxr_drv_data,
+		.driver_data	= (unsigned long)&exynos4210_mxr_drv_data,
 	}, {
 		.name		= "exynos5-mixer",
-		.driver_data	= (unsigned long)&exynos5_mxr_drv_data,
+		.driver_data	= (unsigned long)&exynos5250_mxr_drv_data,
 	}, {
 		/* end node */
 	}
@@ -1153,7 +1162,13 @@ static struct platform_device_id mixer_driver_types[] = {
 static struct of_device_id mixer_match_types[] = {
 	{
 		.compatible = "samsung,exynos5-mixer",
-		.data	= &exynos5_mxr_drv_data,
+		.data	= &exynos5250_mxr_drv_data,
+	}, {
+		.compatible = "samsung,exynos5250-mixer",
+		.data	= &exynos5250_mxr_drv_data,
+	}, {
+		.compatible = "samsung,exynos5420-mixer",
+		.data	= &exynos5420_mxr_drv_data,
 	}, {
 		/* end node */
 	}
@@ -1186,8 +1201,7 @@ static int mixer_probe(struct platform_device *pdev)
 
 	if (dev->of_node) {
 		const struct of_device_id *match;
-		match = of_match_node(of_match_ptr(mixer_match_types),
-							  dev->of_node);
+		match = of_match_node(mixer_match_types, dev->of_node);
 		drv = (struct mixer_drv_data *)match->data;
 	} else {
 		drv = (struct mixer_drv_data *)
@@ -1251,10 +1265,8 @@ static int mixer_suspend(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already suspended\n", __func__);
+		DRM_DEBUG_KMS("Already suspended\n");
 		return 0;
 	}
 
@@ -1268,10 +1280,8 @@ static int mixer_resume(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (!pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already resumed\n", __func__);
+		DRM_DEBUG_KMS("Already resumed\n");
 		return 0;
 	}
 
@@ -1287,8 +1297,6 @@ static int mixer_runtime_suspend(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mixer_poweroff(ctx);
 
 	return 0;
@@ -1299,8 +1307,6 @@ static int mixer_runtime_resume(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mixer_poweron(ctx);
 
 	return 0;
diff --git a/drivers/gpu/drm/exynos/regs-mixer.h b/drivers/gpu/drm/exynos/regs-mixer.h
index 5d8dbc0..4537026 100644
--- a/drivers/gpu/drm/exynos/regs-mixer.h
+++ b/drivers/gpu/drm/exynos/regs-mixer.h
@@ -44,6 +44,9 @@
 #define MXR_CM_COEFF_Y			0x0080
 #define MXR_CM_COEFF_CB			0x0084
 #define MXR_CM_COEFF_CR			0x0088
+#define MXR_MO				0x0304
+#define MXR_RESOLUTION			0x0310
+
 #define MXR_GRAPHIC0_BASE_S		0x2024
 #define MXR_GRAPHIC1_BASE_S		0x2044
 
@@ -119,6 +122,10 @@
 #define MXR_GRP_WH_WIDTH(x)		MXR_MASK_VAL(x, 26, 16)
 #define MXR_GRP_WH_HEIGHT(x)		MXR_MASK_VAL(x, 10, 0)
 
+/* bits for MXR_RESOLUTION */
+#define MXR_MXR_RES_HEIGHT(x)		MXR_MASK_VAL(x, 26, 16)
+#define MXR_MXR_RES_WIDTH(x)		MXR_MASK_VAL(x, 10, 0)
+
 /* bits for MXR_GRAPHICn_SXY */
 #define MXR_GRP_SXY_SX(x)		MXR_MASK_VAL(x, 26, 16)
 #define MXR_GRP_SXY_SY(x)		MXR_MASK_VAL(x, 10, 0)
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 91f3ac6..40034ec 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -36,6 +36,7 @@ i915-y := i915_drv.o i915_dma.o i915_irq.o \
 	  intel_overlay.o \
 	  intel_sprite.o \
 	  intel_opregion.o \
+	  intel_sideband.o \
 	  dvo_ch7xxx.o \
 	  dvo_ch7017.o \
 	  dvo_ivch.o \
diff --git a/drivers/gpu/drm/i915/dvo_ch7xxx.c b/drivers/gpu/drm/i915/dvo_ch7xxx.c
index 3edd981..757e0fa 100644
--- a/drivers/gpu/drm/i915/dvo_ch7xxx.c
+++ b/drivers/gpu/drm/i915/dvo_ch7xxx.c
@@ -32,12 +32,14 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define CH7xxx_REG_DID		0x4b
 
 #define CH7011_VID		0x83 /* 7010 as well */
+#define CH7010B_VID		0x05
 #define CH7009A_VID		0x84
 #define CH7009B_VID		0x85
 #define CH7301_VID		0x95
 
 #define CH7xxx_VID		0x84
 #define CH7xxx_DID		0x17
+#define CH7010_DID		0x16
 
 #define CH7xxx_NUM_REGS		0x4c
 
@@ -87,11 +89,20 @@ static struct ch7xxx_id_struct {
 	char *name;
 } ch7xxx_ids[] = {
 	{ CH7011_VID, "CH7011" },
+	{ CH7010B_VID, "CH7010B" },
 	{ CH7009A_VID, "CH7009A" },
 	{ CH7009B_VID, "CH7009B" },
 	{ CH7301_VID, "CH7301" },
 };
 
+static struct ch7xxx_did_struct {
+	uint8_t did;
+	char *name;
+} ch7xxx_dids[] = {
+	{ CH7xxx_DID, "CH7XXX" },
+	{ CH7010_DID, "CH7010B" },
+};
+
 struct ch7xxx_priv {
 	bool quiet;
 };
@@ -108,6 +119,18 @@ static char *ch7xxx_get_id(uint8_t vid)
 	return NULL;
 }
 
+static char *ch7xxx_get_did(uint8_t did)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ch7xxx_dids); i++) {
+		if (ch7xxx_dids[i].did == did)
+			return ch7xxx_dids[i].name;
+	}
+
+	return NULL;
+}
+
 /** Reads an 8 bit register */
 static bool ch7xxx_readb(struct intel_dvo_device *dvo, int addr, uint8_t *ch)
 {
@@ -179,7 +202,7 @@ static bool ch7xxx_init(struct intel_dvo_device *dvo,
 	/* this will detect the CH7xxx chip on the specified i2c bus */
 	struct ch7xxx_priv *ch7xxx;
 	uint8_t vendor, device;
-	char *name;
+	char *name, *devid;
 
 	ch7xxx = kzalloc(sizeof(struct ch7xxx_priv), GFP_KERNEL);
 	if (ch7xxx == NULL)
@@ -204,7 +227,8 @@ static bool ch7xxx_init(struct intel_dvo_device *dvo,
 	if (!ch7xxx_readb(dvo, CH7xxx_REG_DID, &device))
 		goto out;
 
-	if (device != CH7xxx_DID) {
+	devid = ch7xxx_get_did(device);
+	if (!devid) {
 		DRM_DEBUG_KMS("ch7xxx not detected; got 0x%02x from %s "
 				"slave %d.\n",
 			  vendor, adapter->name, dvo->slave_addr);
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index e913d32..47d6c74 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -61,11 +61,11 @@ static int i915_capabilities(struct seq_file *m, void *data)
 
 	seq_printf(m, "gen: %d\n", info->gen);
 	seq_printf(m, "pch: %d\n", INTEL_PCH_TYPE(dev));
-#define DEV_INFO_FLAG(x) seq_printf(m, #x ": %s\n", yesno(info->x))
-#define DEV_INFO_SEP ;
-	DEV_INFO_FLAGS;
-#undef DEV_INFO_FLAG
-#undef DEV_INFO_SEP
+#define PRINT_FLAG(x)  seq_printf(m, #x ": %s\n", yesno(info->x))
+#define SEP_SEMICOLON ;
+	DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_SEMICOLON);
+#undef PRINT_FLAG
+#undef SEP_SEMICOLON
 
 	return 0;
 }
@@ -196,6 +196,32 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data)
 	} \
 } while (0)
 
+struct file_stats {
+	int count;
+	size_t total, active, inactive, unbound;
+};
+
+static int per_file_stats(int id, void *ptr, void *data)
+{
+	struct drm_i915_gem_object *obj = ptr;
+	struct file_stats *stats = data;
+
+	stats->count++;
+	stats->total += obj->base.size;
+
+	if (obj->gtt_space) {
+		if (!list_empty(&obj->ring_list))
+			stats->active += obj->base.size;
+		else
+			stats->inactive += obj->base.size;
+	} else {
+		if (!list_empty(&obj->global_list))
+			stats->unbound += obj->base.size;
+	}
+
+	return 0;
+}
+
 static int i915_gem_object_info(struct seq_file *m, void* data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -204,6 +230,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 	u32 count, mappable_count, purgeable_count;
 	size_t size, mappable_size, purgeable_size;
 	struct drm_i915_gem_object *obj;
+	struct drm_file *file;
 	int ret;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -215,7 +242,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   dev_priv->mm.object_memory);
 
 	size = count = mappable_size = mappable_count = 0;
-	count_objects(&dev_priv->mm.bound_list, gtt_list);
+	count_objects(&dev_priv->mm.bound_list, global_list);
 	seq_printf(m, "%u [%u] objects, %zu [%zu] bytes in gtt\n",
 		   count, mappable_count, size, mappable_size);
 
@@ -230,7 +257,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   count, mappable_count, size, mappable_size);
 
 	size = count = purgeable_size = purgeable_count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.unbound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list) {
 		size += obj->base.size, ++count;
 		if (obj->madv == I915_MADV_DONTNEED)
 			purgeable_size += obj->base.size, ++purgeable_count;
@@ -238,7 +265,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 	seq_printf(m, "%u unbound objects, %zu bytes\n", count, size);
 
 	size = count = mappable_size = mappable_count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		if (obj->fault_mappable) {
 			size += obj->gtt_space->size;
 			++count;
@@ -263,6 +290,21 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   dev_priv->gtt.total,
 		   dev_priv->gtt.mappable_end - dev_priv->gtt.start);
 
+	seq_printf(m, "\n");
+	list_for_each_entry_reverse(file, &dev->filelist, lhead) {
+		struct file_stats stats;
+
+		memset(&stats, 0, sizeof(stats));
+		idr_for_each(&file->object_idr, per_file_stats, &stats);
+		seq_printf(m, "%s: %u objects, %zu bytes (%zu active, %zu inactive, %zu unbound)\n",
+			   get_pid_task(file->pid, PIDTYPE_PID)->comm,
+			   stats.count,
+			   stats.total,
+			   stats.active,
+			   stats.inactive,
+			   stats.unbound);
+	}
+
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
@@ -283,7 +325,7 @@ static int i915_gem_gtt_info(struct seq_file *m, void* data)
 		return ret;
 
 	total_obj_size = total_gtt_size = count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		if (list == PINNED_LIST && obj->pin_count == 0)
 			continue;
 
@@ -570,6 +612,7 @@ static const char *ring_str(int ring)
 	case RCS: return "render";
 	case VCS: return "bsd";
 	case BCS: return "blt";
+	case VECS: return "vebox";
 	default: return "";
 	}
 }
@@ -604,73 +647,187 @@ static const char *purgeable_flag(int purgeable)
 	return purgeable ? " purgeable" : "";
 }
 
-static void print_error_buffers(struct seq_file *m,
+static bool __i915_error_ok(struct drm_i915_error_state_buf *e)
+{
+
+	if (!e->err && WARN(e->bytes > (e->size - 1), "overflow")) {
+		e->err = -ENOSPC;
+		return false;
+	}
+
+	if (e->bytes == e->size - 1 || e->err)
+		return false;
+
+	return true;
+}
+
+static bool __i915_error_seek(struct drm_i915_error_state_buf *e,
+			      unsigned len)
+{
+	if (e->pos + len <= e->start) {
+		e->pos += len;
+		return false;
+	}
+
+	/* First vsnprintf needs to fit in its entirety for memmove */
+	if (len >= e->size) {
+		e->err = -EIO;
+		return false;
+	}
+
+	return true;
+}
+
+static void __i915_error_advance(struct drm_i915_error_state_buf *e,
+				 unsigned len)
+{
+	/* If this is first printf in this window, adjust it so that
+	 * start position matches start of the buffer
+	 */
+
+	if (e->pos < e->start) {
+		const size_t off = e->start - e->pos;
+
+		/* Should not happen but be paranoid */
+		if (off > len || e->bytes) {
+			e->err = -EIO;
+			return;
+		}
+
+		memmove(e->buf, e->buf + off, len - off);
+		e->bytes = len - off;
+		e->pos = e->start;
+		return;
+	}
+
+	e->bytes += len;
+	e->pos += len;
+}
+
+static void i915_error_vprintf(struct drm_i915_error_state_buf *e,
+			       const char *f, va_list args)
+{
+	unsigned len;
+
+	if (!__i915_error_ok(e))
+		return;
+
+	/* Seek the first printf which is hits start position */
+	if (e->pos < e->start) {
+		len = vsnprintf(NULL, 0, f, args);
+		if (!__i915_error_seek(e, len))
+			return;
+	}
+
+	len = vsnprintf(e->buf + e->bytes, e->size - e->bytes, f, args);
+	if (len >= e->size - e->bytes)
+		len = e->size - e->bytes - 1;
+
+	__i915_error_advance(e, len);
+}
+
+static void i915_error_puts(struct drm_i915_error_state_buf *e,
+			    const char *str)
+{
+	unsigned len;
+
+	if (!__i915_error_ok(e))
+		return;
+
+	len = strlen(str);
+
+	/* Seek the first printf which is hits start position */
+	if (e->pos < e->start) {
+		if (!__i915_error_seek(e, len))
+			return;
+	}
+
+	if (len >= e->size - e->bytes)
+		len = e->size - e->bytes - 1;
+	memcpy(e->buf + e->bytes, str, len);
+
+	__i915_error_advance(e, len);
+}
+
+void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
+{
+	va_list args;
+
+	va_start(args, f);
+	i915_error_vprintf(e, f, args);
+	va_end(args);
+}
+
+#define err_printf(e, ...) i915_error_printf(e, __VA_ARGS__)
+#define err_puts(e, s) i915_error_puts(e, s)
+
+static void print_error_buffers(struct drm_i915_error_state_buf *m,
 				const char *name,
 				struct drm_i915_error_buffer *err,
 				int count)
 {
-	seq_printf(m, "%s [%d]:\n", name, count);
+	err_printf(m, "%s [%d]:\n", name, count);
 
 	while (count--) {
-		seq_printf(m, "  %08x %8u %02x %02x %x %x%s%s%s%s%s%s%s",
+		err_printf(m, "  %08x %8u %02x %02x %x %x",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
 			   err->write_domain,
-			   err->rseqno, err->wseqno,
-			   pin_flag(err->pinned),
-			   tiling_flag(err->tiling),
-			   dirty_flag(err->dirty),
-			   purgeable_flag(err->purgeable),
-			   err->ring != -1 ? " " : "",
-			   ring_str(err->ring),
-			   cache_level_str(err->cache_level));
+			   err->rseqno, err->wseqno);
+		err_puts(m, pin_flag(err->pinned));
+		err_puts(m, tiling_flag(err->tiling));
+		err_puts(m, dirty_flag(err->dirty));
+		err_puts(m, purgeable_flag(err->purgeable));
+		err_puts(m, err->ring != -1 ? " " : "");
+		err_puts(m, ring_str(err->ring));
+		err_puts(m, cache_level_str(err->cache_level));
 
 		if (err->name)
-			seq_printf(m, " (name: %d)", err->name);
+			err_printf(m, " (name: %d)", err->name);
 		if (err->fence_reg != I915_FENCE_REG_NONE)
-			seq_printf(m, " (fence: %d)", err->fence_reg);
+			err_printf(m, " (fence: %d)", err->fence_reg);
 
-		seq_printf(m, "\n");
+		err_puts(m, "\n");
 		err++;
 	}
 }
 
-static void i915_ring_error_state(struct seq_file *m,
+static void i915_ring_error_state(struct drm_i915_error_state_buf *m,
 				  struct drm_device *dev,
 				  struct drm_i915_error_state *error,
 				  unsigned ring)
 {
 	BUG_ON(ring >= I915_NUM_RINGS); /* shut up confused gcc */
-	seq_printf(m, "%s command stream:\n", ring_str(ring));
-	seq_printf(m, "  HEAD: 0x%08x\n", error->head[ring]);
-	seq_printf(m, "  TAIL: 0x%08x\n", error->tail[ring]);
-	seq_printf(m, "  CTL: 0x%08x\n", error->ctl[ring]);
-	seq_printf(m, "  ACTHD: 0x%08x\n", error->acthd[ring]);
-	seq_printf(m, "  IPEIR: 0x%08x\n", error->ipeir[ring]);
-	seq_printf(m, "  IPEHR: 0x%08x\n", error->ipehr[ring]);
-	seq_printf(m, "  INSTDONE: 0x%08x\n", error->instdone[ring]);
+	err_printf(m, "%s command stream:\n", ring_str(ring));
+	err_printf(m, "  HEAD: 0x%08x\n", error->head[ring]);
+	err_printf(m, "  TAIL: 0x%08x\n", error->tail[ring]);
+	err_printf(m, "  CTL: 0x%08x\n", error->ctl[ring]);
+	err_printf(m, "  ACTHD: 0x%08x\n", error->acthd[ring]);
+	err_printf(m, "  IPEIR: 0x%08x\n", error->ipeir[ring]);
+	err_printf(m, "  IPEHR: 0x%08x\n", error->ipehr[ring]);
+	err_printf(m, "  INSTDONE: 0x%08x\n", error->instdone[ring]);
 	if (ring == RCS && INTEL_INFO(dev)->gen >= 4)
-		seq_printf(m, "  BBADDR: 0x%08llx\n", error->bbaddr);
+		err_printf(m, "  BBADDR: 0x%08llx\n", error->bbaddr);
 
 	if (INTEL_INFO(dev)->gen >= 4)
-		seq_printf(m, "  INSTPS: 0x%08x\n", error->instps[ring]);
-	seq_printf(m, "  INSTPM: 0x%08x\n", error->instpm[ring]);
-	seq_printf(m, "  FADDR: 0x%08x\n", error->faddr[ring]);
+		err_printf(m, "  INSTPS: 0x%08x\n", error->instps[ring]);
+	err_printf(m, "  INSTPM: 0x%08x\n", error->instpm[ring]);
+	err_printf(m, "  FADDR: 0x%08x\n", error->faddr[ring]);
 	if (INTEL_INFO(dev)->gen >= 6) {
-		seq_printf(m, "  RC PSMI: 0x%08x\n", error->rc_psmi[ring]);
-		seq_printf(m, "  FAULT_REG: 0x%08x\n", error->fault_reg[ring]);
-		seq_printf(m, "  SYNC_0: 0x%08x [last synced 0x%08x]\n",
+		err_printf(m, "  RC PSMI: 0x%08x\n", error->rc_psmi[ring]);
+		err_printf(m, "  FAULT_REG: 0x%08x\n", error->fault_reg[ring]);
+		err_printf(m, "  SYNC_0: 0x%08x [last synced 0x%08x]\n",
 			   error->semaphore_mboxes[ring][0],
 			   error->semaphore_seqno[ring][0]);
-		seq_printf(m, "  SYNC_1: 0x%08x [last synced 0x%08x]\n",
+		err_printf(m, "  SYNC_1: 0x%08x [last synced 0x%08x]\n",
 			   error->semaphore_mboxes[ring][1],
 			   error->semaphore_seqno[ring][1]);
 	}
-	seq_printf(m, "  seqno: 0x%08x\n", error->seqno[ring]);
-	seq_printf(m, "  waiting: %s\n", yesno(error->waiting[ring]));
-	seq_printf(m, "  ring->head: 0x%08x\n", error->cpu_ring_head[ring]);
-	seq_printf(m, "  ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]);
+	err_printf(m, "  seqno: 0x%08x\n", error->seqno[ring]);
+	err_printf(m, "  waiting: %s\n", yesno(error->waiting[ring]));
+	err_printf(m, "  ring->head: 0x%08x\n", error->cpu_ring_head[ring]);
+	err_printf(m, "  ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]);
 }
 
 struct i915_error_state_file_priv {
@@ -678,9 +835,11 @@ struct i915_error_state_file_priv {
 	struct drm_i915_error_state *error;
 };
 
-static int i915_error_state(struct seq_file *m, void *unused)
+
+static int i915_error_state(struct i915_error_state_file_priv *error_priv,
+			    struct drm_i915_error_state_buf *m)
+
 {
-	struct i915_error_state_file_priv *error_priv = m->private;
 	struct drm_device *dev = error_priv->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error = error_priv->error;
@@ -688,34 +847,35 @@ static int i915_error_state(struct seq_file *m, void *unused)
 	int i, j, page, offset, elt;
 
 	if (!error) {
-		seq_printf(m, "no error state collected\n");
+		err_printf(m, "no error state collected\n");
 		return 0;
 	}
 
-	seq_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
+	err_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
 		   error->time.tv_usec);
-	seq_printf(m, "Kernel: " UTS_RELEASE "\n");
-	seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
-	seq_printf(m, "EIR: 0x%08x\n", error->eir);
-	seq_printf(m, "IER: 0x%08x\n", error->ier);
-	seq_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
-	seq_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
-	seq_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
-	seq_printf(m, "CCID: 0x%08x\n", error->ccid);
+	err_printf(m, "Kernel: " UTS_RELEASE "\n");
+	err_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
+	err_printf(m, "EIR: 0x%08x\n", error->eir);
+	err_printf(m, "IER: 0x%08x\n", error->ier);
+	err_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
+	err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
+	err_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
+	err_printf(m, "CCID: 0x%08x\n", error->ccid);
 
 	for (i = 0; i < dev_priv->num_fence_regs; i++)
-		seq_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
+		err_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
 
 	for (i = 0; i < ARRAY_SIZE(error->extra_instdone); i++)
-		seq_printf(m, "  INSTDONE_%d: 0x%08x\n", i, error->extra_instdone[i]);
+		err_printf(m, "  INSTDONE_%d: 0x%08x\n", i,
+			   error->extra_instdone[i]);
 
 	if (INTEL_INFO(dev)->gen >= 6) {
-		seq_printf(m, "ERROR: 0x%08x\n", error->error);
-		seq_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
+		err_printf(m, "ERROR: 0x%08x\n", error->error);
+		err_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
 	}
 
 	if (INTEL_INFO(dev)->gen == 7)
-		seq_printf(m, "ERR_INT: 0x%08x\n", error->err_int);
+		err_printf(m, "ERR_INT: 0x%08x\n", error->err_int);
 
 	for_each_ring(ring, dev_priv, i)
 		i915_ring_error_state(m, dev, error, i);
@@ -734,24 +894,25 @@ static int i915_error_state(struct seq_file *m, void *unused)
 		struct drm_i915_error_object *obj;
 
 		if ((obj = error->ring[i].batchbuffer)) {
-			seq_printf(m, "%s --- gtt_offset = 0x%08x\n",
+			err_printf(m, "%s --- gtt_offset = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
 			offset = 0;
 			for (page = 0; page < obj->page_count; page++) {
 				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
-					seq_printf(m, "%08x :  %08x\n", offset, obj->pages[page][elt]);
+					err_printf(m, "%08x :  %08x\n", offset,
+						   obj->pages[page][elt]);
 					offset += 4;
 				}
 			}
 		}
 
 		if (error->ring[i].num_requests) {
-			seq_printf(m, "%s --- %d requests\n",
+			err_printf(m, "%s --- %d requests\n",
 				   dev_priv->ring[i].name,
 				   error->ring[i].num_requests);
 			for (j = 0; j < error->ring[i].num_requests; j++) {
-				seq_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
+				err_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
 					   error->ring[i].requests[j].seqno,
 					   error->ring[i].requests[j].jiffies,
 					   error->ring[i].requests[j].tail);
@@ -759,13 +920,13 @@ static int i915_error_state(struct seq_file *m, void *unused)
 		}
 
 		if ((obj = error->ring[i].ringbuffer)) {
-			seq_printf(m, "%s --- ringbuffer = 0x%08x\n",
+			err_printf(m, "%s --- ringbuffer = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
 			offset = 0;
 			for (page = 0; page < obj->page_count; page++) {
 				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
-					seq_printf(m, "%08x :  %08x\n",
+					err_printf(m, "%08x :  %08x\n",
 						   offset,
 						   obj->pages[page][elt]);
 					offset += 4;
@@ -775,12 +936,12 @@ static int i915_error_state(struct seq_file *m, void *unused)
 
 		obj = error->ring[i].ctx;
 		if (obj) {
-			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+			err_printf(m, "%s --- HW Context = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
 			offset = 0;
 			for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
-				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+				err_printf(m, "[%04x] %08x %08x %08x %08x\n",
 					   offset,
 					   obj->pages[0][elt],
 					   obj->pages[0][elt+1],
@@ -806,8 +967,7 @@ i915_error_state_write(struct file *filp,
 		       size_t cnt,
 		       loff_t *ppos)
 {
-	struct seq_file *m = filp->private_data;
-	struct i915_error_state_file_priv *error_priv = m->private;
+	struct i915_error_state_file_priv *error_priv = filp->private_data;
 	struct drm_device *dev = error_priv->dev;
 	int ret;
 
@@ -842,25 +1002,81 @@ static int i915_error_state_open(struct inode *inode, struct file *file)
 		kref_get(&error_priv->error->ref);
 	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 
-	return single_open(file, i915_error_state, error_priv);
+	file->private_data = error_priv;
+
+	return 0;
 }
 
 static int i915_error_state_release(struct inode *inode, struct file *file)
 {
-	struct seq_file *m = file->private_data;
-	struct i915_error_state_file_priv *error_priv = m->private;
+	struct i915_error_state_file_priv *error_priv = file->private_data;
 
 	if (error_priv->error)
 		kref_put(&error_priv->error->ref, i915_error_state_free);
 	kfree(error_priv);
 
-	return single_release(inode, file);
+	return 0;
+}
+
+static ssize_t i915_error_state_read(struct file *file, char __user *userbuf,
+				     size_t count, loff_t *pos)
+{
+	struct i915_error_state_file_priv *error_priv = file->private_data;
+	struct drm_i915_error_state_buf error_str;
+	loff_t tmp_pos = 0;
+	ssize_t ret_count = 0;
+	int ret = 0;
+
+	memset(&error_str, 0, sizeof(error_str));
+
+	/* We need to have enough room to store any i915_error_state printf
+	 * so that we can move it to start position.
+	 */
+	error_str.size = count + 1 > PAGE_SIZE ? count + 1 : PAGE_SIZE;
+	error_str.buf = kmalloc(error_str.size,
+				GFP_TEMPORARY | __GFP_NORETRY | __GFP_NOWARN);
+
+	if (error_str.buf == NULL) {
+		error_str.size = PAGE_SIZE;
+		error_str.buf = kmalloc(error_str.size, GFP_TEMPORARY);
+	}
+
+	if (error_str.buf == NULL) {
+		error_str.size = 128;
+		error_str.buf = kmalloc(error_str.size, GFP_TEMPORARY);
+	}
+
+	if (error_str.buf == NULL)
+		return -ENOMEM;
+
+	error_str.start = *pos;
+
+	ret = i915_error_state(error_priv, &error_str);
+	if (ret)
+		goto out;
+
+	if (error_str.bytes == 0 && error_str.err) {
+		ret = error_str.err;
+		goto out;
+	}
+
+	ret_count = simple_read_from_buffer(userbuf, count, &tmp_pos,
+					    error_str.buf,
+					    error_str.bytes);
+
+	if (ret_count < 0)
+		ret = ret_count;
+	else
+		*pos = error_str.start + ret_count;
+out:
+	kfree(error_str.buf);
+	return ret ?: ret_count;
 }
 
 static const struct file_operations i915_error_state_fops = {
 	.owner = THIS_MODULE,
 	.open = i915_error_state_open,
-	.read = seq_read,
+	.read = i915_error_state_read,
 	.write = i915_error_state_write,
 	.llseek = default_llseek,
 	.release = i915_error_state_release,
@@ -941,7 +1157,7 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 			   MEMSTAT_VID_SHIFT);
 		seq_printf(m, "Current P-state: %d\n",
 			   (rgvstat & MEMSTAT_PSTATE_MASK) >> MEMSTAT_PSTATE_SHIFT);
-	} else if (IS_GEN6(dev) || IS_GEN7(dev)) {
+	} else if ((IS_GEN6(dev) || IS_GEN7(dev)) && !IS_VALLEYVIEW(dev)) {
 		u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 		u32 rp_state_limits = I915_READ(GEN6_RP_STATE_LIMITS);
 		u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
@@ -1009,6 +1225,26 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 
 		seq_printf(m, "Max overclocked frequency: %dMHz\n",
 			   dev_priv->rps.hw_max * GT_FREQUENCY_MULTIPLIER);
+	} else if (IS_VALLEYVIEW(dev)) {
+		u32 freq_sts, val;
+
+		mutex_lock(&dev_priv->rps.hw_lock);
+		freq_sts = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+		seq_printf(m, "PUNIT_REG_GPU_FREQ_STS: 0x%08x\n", freq_sts);
+		seq_printf(m, "DDR freq: %d MHz\n", dev_priv->mem_freq);
+
+		val = vlv_punit_read(dev_priv, PUNIT_FUSE_BUS1);
+		seq_printf(m, "max GPU freq: %d MHz\n",
+			   vlv_gpu_freq(dev_priv->mem_freq, val));
+
+		val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_LFM);
+		seq_printf(m, "min GPU freq: %d MHz\n",
+			   vlv_gpu_freq(dev_priv->mem_freq, val));
+
+		seq_printf(m, "current GPU freq: %d MHz\n",
+			   vlv_gpu_freq(dev_priv->mem_freq,
+					(freq_sts >> 8) & 0xff));
+		mutex_unlock(&dev_priv->rps.hw_lock);
 	} else {
 		seq_printf(m, "no P-state info available\n");
 	}
@@ -1290,6 +1526,25 @@ static int i915_fbc_status(struct seq_file *m, void *unused)
 	return 0;
 }
 
+static int i915_ips_status(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!HAS_IPS(dev)) {
+		seq_puts(m, "not supported\n");
+		return 0;
+	}
+
+	if (I915_READ(IPS_CTL) & IPS_ENABLE)
+		seq_puts(m, "enabled\n");
+	else
+		seq_puts(m, "disabled\n");
+
+	return 0;
+}
+
 static int i915_sr_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -1642,27 +1897,27 @@ static int i915_dpio_info(struct seq_file *m, void *data)
 	seq_printf(m, "DPIO_CTL: 0x%08x\n", I915_READ(DPIO_CTL));
 
 	seq_printf(m, "DPIO_DIV_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_DIV_A));
+		   vlv_dpio_read(dev_priv, _DPIO_DIV_A));
 	seq_printf(m, "DPIO_DIV_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_DIV_B));
+		   vlv_dpio_read(dev_priv, _DPIO_DIV_B));
 
 	seq_printf(m, "DPIO_REFSFR_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_REFSFR_A));
+		   vlv_dpio_read(dev_priv, _DPIO_REFSFR_A));
 	seq_printf(m, "DPIO_REFSFR_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_REFSFR_B));
+		   vlv_dpio_read(dev_priv, _DPIO_REFSFR_B));
 
 	seq_printf(m, "DPIO_CORE_CLK_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_CORE_CLK_A));
+		   vlv_dpio_read(dev_priv, _DPIO_CORE_CLK_A));
 	seq_printf(m, "DPIO_CORE_CLK_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_CORE_CLK_B));
+		   vlv_dpio_read(dev_priv, _DPIO_CORE_CLK_B));
 
-	seq_printf(m, "DPIO_LFP_COEFF_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_LFP_COEFF_A));
-	seq_printf(m, "DPIO_LFP_COEFF_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_LFP_COEFF_B));
+	seq_printf(m, "DPIO_LPF_COEFF_A: 0x%08x\n",
+		   vlv_dpio_read(dev_priv, _DPIO_LPF_COEFF_A));
+	seq_printf(m, "DPIO_LPF_COEFF_B: 0x%08x\n",
+		   vlv_dpio_read(dev_priv, _DPIO_LPF_COEFF_B));
 
 	seq_printf(m, "DPIO_FASTCLK_DISABLE: 0x%08x\n",
-		   intel_dpio_read(dev_priv, DPIO_FASTCLK_DISABLE));
+		   vlv_dpio_read(dev_priv, DPIO_FASTCLK_DISABLE));
 
 	mutex_unlock(&dev_priv->dpio_lock);
 
@@ -1780,7 +2035,8 @@ i915_drop_caches_set(void *data, u64 val)
 	}
 
 	if (val & DROP_UNBOUND) {
-		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list, gtt_list)
+		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list,
+					 global_list)
 			if (obj->pages_pin_count == 0) {
 				ret = i915_gem_object_put_pages(obj);
 				if (ret)
@@ -1812,7 +2068,11 @@ i915_max_freq_get(void *data, u64 *val)
 	if (ret)
 		return ret;
 
-	*val = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev))
+		*val = vlv_gpu_freq(dev_priv->mem_freq,
+				    dev_priv->rps.max_delay);
+	else
+		*val = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -1837,9 +2097,16 @@ i915_max_freq_set(void *data, u64 val)
 	/*
 	 * Turbo will still be enabled, but won't go above the set value.
 	 */
-	do_div(val, GT_FREQUENCY_MULTIPLIER);
-	dev_priv->rps.max_delay = val;
-	gen6_set_rps(dev, val);
+	if (IS_VALLEYVIEW(dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+		dev_priv->rps.max_delay = val;
+		gen6_set_rps(dev, val);
+	} else {
+		do_div(val, GT_FREQUENCY_MULTIPLIER);
+		dev_priv->rps.max_delay = val;
+		gen6_set_rps(dev, val);
+	}
+
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -1863,7 +2130,11 @@ i915_min_freq_get(void *data, u64 *val)
 	if (ret)
 		return ret;
 
-	*val = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev))
+		*val = vlv_gpu_freq(dev_priv->mem_freq,
+				    dev_priv->rps.min_delay);
+	else
+		*val = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -1888,9 +2159,15 @@ i915_min_freq_set(void *data, u64 val)
 	/*
 	 * Turbo will still be enabled, but won't go below the set value.
 	 */
-	do_div(val, GT_FREQUENCY_MULTIPLIER);
-	dev_priv->rps.min_delay = val;
-	gen6_set_rps(dev, val);
+	if (IS_VALLEYVIEW(dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+		dev_priv->rps.min_delay = val;
+		valleyview_set_rps(dev, val);
+	} else {
+		do_div(val, GT_FREQUENCY_MULTIPLIER);
+		dev_priv->rps.min_delay = val;
+		gen6_set_rps(dev, val);
+	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -2057,6 +2334,7 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_gem_hws", i915_hws_info, 0, (void *)RCS},
 	{"i915_gem_hws_blt", i915_hws_info, 0, (void *)BCS},
 	{"i915_gem_hws_bsd", i915_hws_info, 0, (void *)VCS},
+	{"i915_gem_hws_vebox", i915_hws_info, 0, (void *)VECS},
 	{"i915_rstdby_delays", i915_rstdby_delays, 0},
 	{"i915_cur_delayinfo", i915_cur_delayinfo, 0},
 	{"i915_delayfreq_table", i915_delayfreq_table, 0},
@@ -2066,6 +2344,7 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_ring_freq_table", i915_ring_freq_table, 0},
 	{"i915_gfxec", i915_gfxec, 0},
 	{"i915_fbc_status", i915_fbc_status, 0},
+	{"i915_ips_status", i915_ips_status, 0},
 	{"i915_sr_status", i915_sr_status, 0},
 	{"i915_opregion", i915_opregion, 0},
 	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 3b315ba..adb319b 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -42,7 +42,6 @@
 #include <linux/vga_switcheroo.h>
 #include <linux/slab.h>
 #include <acpi/video.h>
-#include <asm/pat.h>
 
 #define LP_RING(d) (&((struct drm_i915_private *)(d))->ring[RCS])
 
@@ -956,6 +955,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_BLT:
 		value = intel_ring_initialized(&dev_priv->ring[BCS]);
 		break;
+	case I915_PARAM_HAS_VEBOX:
+		value = intel_ring_initialized(&dev_priv->ring[VECS]);
+		break;
 	case I915_PARAM_HAS_RELAXED_FENCING:
 		value = 1;
 		break;
@@ -999,8 +1001,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = 1;
 		break;
 	default:
-		DRM_DEBUG_DRIVER("Unknown parameter %d\n",
-				 param->param);
+		DRM_DEBUG("Unknown parameter %d\n", param->param);
 		return -EINVAL;
 	}
 
@@ -1359,8 +1360,10 @@ static int i915_load_modeset_init(struct drm_device *dev)
 cleanup_gem:
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_cleanup_ringbuffer(dev);
+	i915_gem_context_fini(dev);
 	mutex_unlock(&dev->struct_mutex);
 	i915_gem_cleanup_aliasing_ppgtt(dev);
+	drm_mm_takedown(&dev_priv->mm.gtt_space);
 cleanup_irq:
 	drm_irq_uninstall(dev);
 cleanup_gem_stolen:
@@ -1397,29 +1400,6 @@ void i915_master_destroy(struct drm_device *dev, struct drm_master *master)
 	master->driver_priv = NULL;
 }
 
-static void
-i915_mtrr_setup(struct drm_i915_private *dev_priv, unsigned long base,
-		unsigned long size)
-{
-	dev_priv->mm.gtt_mtrr = -1;
-
-#if defined(CONFIG_X86_PAT)
-	if (cpu_has_pat)
-		return;
-#endif
-
-	/* Set up a WC MTRR for non-PAT systems.  This is more common than
-	 * one would think, because the kernel disables PAT on first
-	 * generation Core chips because WC PAT gets overridden by a UC
-	 * MTRR if present.  Even if a UC MTRR isn't present.
-	 */
-	dev_priv->mm.gtt_mtrr = mtrr_add(base, size, MTRR_TYPE_WRCOMB, 1);
-	if (dev_priv->mm.gtt_mtrr < 0) {
-		DRM_INFO("MTRR allocation failed.  Graphics "
-			 "performance may suffer.\n");
-	}
-}
-
 static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 {
 	struct apertures_struct *ap;
@@ -1431,7 +1411,7 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 		return;
 
 	ap->ranges[0].base = dev_priv->gtt.mappable_base;
-	ap->ranges[0].size = dev_priv->gtt.mappable_end - dev_priv->gtt.start;
+	ap->ranges[0].size = dev_priv->gtt.mappable_end;
 
 	primary =
 		pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
@@ -1445,15 +1425,19 @@ static void i915_dump_device_info(struct drm_i915_private *dev_priv)
 {
 	const struct intel_device_info *info = dev_priv->info;
 
-#define DEV_INFO_FLAG(name) info->name ? #name "," : ""
-#define DEV_INFO_SEP ,
+#define PRINT_S(name) "%s"
+#define SEP_EMPTY
+#define PRINT_FLAG(name) info->name ? #name "," : ""
+#define SEP_COMMA ,
 	DRM_DEBUG_DRIVER("i915 device info: gen=%i, pciid=0x%04x flags="
-			 "%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s",
+			 DEV_INFO_FOR_EACH_FLAG(PRINT_S, SEP_EMPTY),
 			 info->gen,
 			 dev_priv->dev->pdev->device,
-			 DEV_INFO_FLAGS);
-#undef DEV_INFO_FLAG
-#undef DEV_INFO_SEP
+			 DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_COMMA));
+#undef PRINT_S
+#undef SEP_EMPTY
+#undef PRINT_FLAG
+#undef SEP_COMMA
 }
 
 /**
@@ -1468,7 +1452,7 @@ static void intel_early_sanitize_regs(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev))
+	if (HAS_FPGA_DBG_UNCLAIMED(dev))
 		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
 }
 
@@ -1574,8 +1558,8 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out_rmmap;
 	}
 
-	i915_mtrr_setup(dev_priv, dev_priv->gtt.mappable_base,
-			aperture_size);
+	dev_priv->mm.gtt_mtrr = arch_phys_wc_add(dev_priv->gtt.mappable_base,
+						 aperture_size);
 
 	/* The i915 workqueue is primarily used for batched retirement of
 	 * requests (and thus managing bo) once the task has been completed
@@ -1629,6 +1613,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	spin_lock_init(&dev_priv->irq_lock);
 	spin_lock_init(&dev_priv->gpu_error.lock);
 	spin_lock_init(&dev_priv->rps.lock);
+	spin_lock_init(&dev_priv->backlight.lock);
 	mutex_init(&dev_priv->dpio_lock);
 
 	mutex_init(&dev_priv->rps.hw_lock);
@@ -1647,6 +1632,9 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	/* Start out suspended */
 	dev_priv->mm.suspended = 1;
 
+	if (HAS_POWER_WELL(dev))
+		i915_init_power_well(dev);
+
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		ret = i915_load_modeset_init(dev);
 		if (ret < 0) {
@@ -1679,12 +1667,7 @@ out_gem_unload:
 	intel_teardown_mchbar(dev);
 	destroy_workqueue(dev_priv->wq);
 out_mtrrfree:
-	if (dev_priv->mm.gtt_mtrr >= 0) {
-		mtrr_del(dev_priv->mm.gtt_mtrr,
-			 dev_priv->gtt.mappable_base,
-			 aperture_size);
-		dev_priv->mm.gtt_mtrr = -1;
-	}
+	arch_phys_wc_del(dev_priv->mm.gtt_mtrr);
 	io_mapping_free(dev_priv->gtt.mappable);
 	dev_priv->gtt.gtt_remove(dev);
 out_rmmap:
@@ -1703,6 +1686,9 @@ int i915_driver_unload(struct drm_device *dev)
 
 	intel_gpu_ips_teardown();
 
+	if (HAS_POWER_WELL(dev))
+		i915_remove_power_well(dev);
+
 	i915_teardown_sysfs(dev);
 
 	if (dev_priv->mm.inactive_shrinker.shrink)
@@ -1719,12 +1705,7 @@ int i915_driver_unload(struct drm_device *dev)
 	cancel_delayed_work_sync(&dev_priv->mm.retire_work);
 
 	io_mapping_free(dev_priv->gtt.mappable);
-	if (dev_priv->mm.gtt_mtrr >= 0) {
-		mtrr_del(dev_priv->mm.gtt_mtrr,
-			 dev_priv->gtt.mappable_base,
-			 dev_priv->gtt.mappable_end);
-		dev_priv->mm.gtt_mtrr = -1;
-	}
+	arch_phys_wc_del(dev_priv->mm.gtt_mtrr);
 
 	acpi_video_unregister();
 
@@ -1737,10 +1718,10 @@ int i915_driver_unload(struct drm_device *dev)
 		 * free the memory space allocated for the child device
 		 * config parsed from VBT
 		 */
-		if (dev_priv->child_dev && dev_priv->child_dev_num) {
-			kfree(dev_priv->child_dev);
-			dev_priv->child_dev = NULL;
-			dev_priv->child_dev_num = 0;
+		if (dev_priv->vbt.child_dev && dev_priv->vbt.child_dev_num) {
+			kfree(dev_priv->vbt.child_dev);
+			dev_priv->vbt.child_dev = NULL;
+			dev_priv->vbt.child_dev_num = 0;
 		}
 
 		vga_switcheroo_unregister_client(dev->pdev);
@@ -1773,6 +1754,7 @@ int i915_driver_unload(struct drm_device *dev)
 			i915_free_hws(dev);
 	}
 
+	drm_mm_takedown(&dev_priv->mm.gtt_space);
 	if (dev_priv->regs != NULL)
 		pci_iounmap(dev->pdev, dev_priv->regs);
 
@@ -1782,6 +1764,8 @@ int i915_driver_unload(struct drm_device *dev)
 	destroy_workqueue(dev_priv->wq);
 	pm_qos_remove_request(&dev_priv->pm_qos);
 
+	dev_priv->gtt.gtt_remove(dev);
+
 	if (dev_priv->slab)
 		kmem_cache_destroy(dev_priv->slab);
 
@@ -1796,7 +1780,7 @@ int i915_driver_open(struct drm_device *dev, struct drm_file *file)
 	struct drm_i915_file_private *file_priv;
 
 	DRM_DEBUG_DRIVER("\n");
-	file_priv = kmalloc(sizeof(*file_priv), GFP_KERNEL);
+	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
 	if (!file_priv)
 		return -ENOMEM;
 
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index a2e4953..062cbda 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -128,6 +128,10 @@ module_param_named(disable_power_well, i915_disable_power_well, int, 0600);
 MODULE_PARM_DESC(disable_power_well,
 		 "Disable the power well when possible (default: false)");
 
+int i915_enable_ips __read_mostly = 1;
+module_param_named(enable_ips, i915_enable_ips, int, 0600);
+MODULE_PARM_DESC(enable_ips, "Enable IPS (default: true)");
+
 static struct drm_driver driver;
 extern int intel_agp_enabled;
 
@@ -280,6 +284,7 @@ static const struct intel_device_info intel_ivybridge_m_info = {
 	GEN7_FEATURES,
 	.is_ivybridge = 1,
 	.is_mobile = 1,
+	.has_fbc = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_q_info = {
@@ -308,12 +313,19 @@ static const struct intel_device_info intel_valleyview_d_info = {
 static const struct intel_device_info intel_haswell_d_info = {
 	GEN7_FEATURES,
 	.is_haswell = 1,
+	.has_ddi = 1,
+	.has_fpga_dbg = 1,
+	.has_vebox_ring = 1,
 };
 
 static const struct intel_device_info intel_haswell_m_info = {
 	GEN7_FEATURES,
 	.is_haswell = 1,
 	.is_mobile = 1,
+	.has_ddi = 1,
+	.has_fpga_dbg = 1,
+	.has_fbc = 1,
+	.has_vebox_ring = 1,
 };
 
 static const struct pci_device_id pciidlist[] = {		/* aka */
@@ -445,7 +457,6 @@ void intel_detect_pch(struct drm_device *dev)
 	 */
 	if (INTEL_INFO(dev)->num_pipes == 0) {
 		dev_priv->pch_type = PCH_NOP;
-		dev_priv->num_pch_pll = 0;
 		return;
 	}
 
@@ -454,9 +465,15 @@ void intel_detect_pch(struct drm_device *dev)
 	 * make graphics device passthrough work easy for VMM, that only
 	 * need to expose ISA bridge to let driver know the real hardware
 	 * underneath. This is a requirement from virtualization team.
+	 *
+	 * In some virtualized environments (e.g. XEN), there is irrelevant
+	 * ISA bridge in the system. To work reliably, we should scan trhough
+	 * all the ISA bridge devices and check for the first match, instead
+	 * of only checking the first one.
 	 */
 	pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, NULL);
-	if (pch) {
+	while (pch) {
+		struct pci_dev *curr = pch;
 		if (pch->vendor == PCI_VENDOR_ID_INTEL) {
 			unsigned short id;
 			id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
@@ -464,37 +481,39 @@ void intel_detect_pch(struct drm_device *dev)
 
 			if (id == INTEL_PCH_IBX_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_IBX;
-				dev_priv->num_pch_pll = 2;
 				DRM_DEBUG_KMS("Found Ibex Peak PCH\n");
 				WARN_ON(!IS_GEN5(dev));
 			} else if (id == INTEL_PCH_CPT_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_CPT;
-				dev_priv->num_pch_pll = 2;
 				DRM_DEBUG_KMS("Found CougarPoint PCH\n");
 				WARN_ON(!(IS_GEN6(dev) || IS_IVYBRIDGE(dev)));
 			} else if (id == INTEL_PCH_PPT_DEVICE_ID_TYPE) {
 				/* PantherPoint is CPT compatible */
 				dev_priv->pch_type = PCH_CPT;
-				dev_priv->num_pch_pll = 2;
 				DRM_DEBUG_KMS("Found PatherPoint PCH\n");
 				WARN_ON(!(IS_GEN6(dev) || IS_IVYBRIDGE(dev)));
 			} else if (id == INTEL_PCH_LPT_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_LPT;
-				dev_priv->num_pch_pll = 0;
 				DRM_DEBUG_KMS("Found LynxPoint PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
 				WARN_ON(IS_ULT(dev));
 			} else if (id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_LPT;
-				dev_priv->num_pch_pll = 0;
 				DRM_DEBUG_KMS("Found LynxPoint LP PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
 				WARN_ON(!IS_ULT(dev));
+			} else {
+				goto check_next;
 			}
-			BUG_ON(dev_priv->num_pch_pll > I915_NUM_PLLS);
+			pci_dev_put(pch);
+			break;
 		}
-		pci_dev_put(pch);
+check_next:
+		pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, curr);
+		pci_dev_put(curr);
 	}
+	if (!pch)
+		DRM_DEBUG_KMS("No PCH found?\n");
 }
 
 bool i915_semaphore_is_enabled(struct drm_device *dev)
@@ -549,6 +568,8 @@ static int i915_drm_freeze(struct drm_device *dev)
 		 */
 		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 			dev_priv->display.crtc_disable(crtc);
+
+		intel_modeset_suspend_hw(dev);
 	}
 
 	i915_save_state(dev);
@@ -556,7 +577,7 @@ static int i915_drm_freeze(struct drm_device *dev)
 	intel_opregion_fini(dev);
 
 	console_lock();
-	intel_fbdev_set_suspend(dev, 1);
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED);
 	console_unlock();
 
 	return 0;
@@ -600,7 +621,7 @@ void intel_console_resume(struct work_struct *work)
 	struct drm_device *dev = dev_priv->dev;
 
 	console_lock();
-	intel_fbdev_set_suspend(dev, 0);
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
 	console_unlock();
 }
 
@@ -669,7 +690,7 @@ static int __i915_drm_thaw(struct drm_device *dev)
 	 * path of resume if possible.
 	 */
 	if (console_trylock()) {
-		intel_fbdev_set_suspend(dev, 0);
+		intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
 		console_unlock();
 	} else {
 		schedule_work(&dev_priv->console_resume_work);
@@ -855,37 +876,14 @@ static int gen6_do_reset(struct drm_device *dev)
 
 int intel_gpu_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret = -ENODEV;
-
 	switch (INTEL_INFO(dev)->gen) {
 	case 7:
-	case 6:
-		ret = gen6_do_reset(dev);
-		break;
-	case 5:
-		ret = ironlake_do_reset(dev);
-		break;
-	case 4:
-		ret = i965_do_reset(dev);
-		break;
-	case 2:
-		ret = i8xx_do_reset(dev);
-		break;
-	}
-
-	/* Also reset the gpu hangman. */
-	if (dev_priv->gpu_error.stop_rings) {
-		DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
-		dev_priv->gpu_error.stop_rings = 0;
-		if (ret == -ENODEV) {
-			DRM_ERROR("Reset not implemented, but ignoring "
-				  "error for simulated gpu hangs\n");
-			ret = 0;
-		}
+	case 6: return gen6_do_reset(dev);
+	case 5: return ironlake_do_reset(dev);
+	case 4: return i965_do_reset(dev);
+	case 2: return i8xx_do_reset(dev);
+	default: return -ENODEV;
 	}
-
-	return ret;
 }
 
 /**
@@ -906,6 +904,7 @@ int intel_gpu_reset(struct drm_device *dev)
 int i915_reset(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	bool simulated;
 	int ret;
 
 	if (!i915_try_reset)
@@ -915,13 +914,26 @@ int i915_reset(struct drm_device *dev)
 
 	i915_gem_reset(dev);
 
-	ret = -ENODEV;
-	if (get_seconds() - dev_priv->gpu_error.last_reset < 5)
+	simulated = dev_priv->gpu_error.stop_rings != 0;
+
+	if (!simulated && get_seconds() - dev_priv->gpu_error.last_reset < 5) {
 		DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
-	else
+		ret = -ENODEV;
+	} else {
 		ret = intel_gpu_reset(dev);
 
-	dev_priv->gpu_error.last_reset = get_seconds();
+		/* Also reset the gpu hangman. */
+		if (simulated) {
+			DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
+			dev_priv->gpu_error.stop_rings = 0;
+			if (ret == -ENODEV) {
+				DRM_ERROR("Reset not implemented, but ignoring "
+					  "error for simulated gpu hangs\n");
+				ret = 0;
+			}
+		} else
+			dev_priv->gpu_error.last_reset = get_seconds();
+	}
 	if (ret) {
 		DRM_ERROR("Failed to reset chip.\n");
 		mutex_unlock(&dev->struct_mutex);
@@ -984,12 +996,6 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	struct intel_device_info *intel_info =
 		(struct intel_device_info *) ent->driver_data;
 
-	if (intel_info->is_valleyview)
-		if(!i915_preliminary_hw_support) {
-			DRM_ERROR("Preliminary hardware support disabled\n");
-			return -ENODEV;
-		}
-
 	/* Only bind to function 0 of the device. Early generations
 	 * used function 1 as a placeholder for multi-head. This causes
 	 * us confusion instead, especially on the systems where both
@@ -1218,16 +1224,16 @@ MODULE_LICENSE("GPL and additional rights");
 static void
 ilk_dummy_write(struct drm_i915_private *dev_priv)
 {
-	/* WaIssueDummyWriteToWakeupFromRC6: Issue a dummy write to wake up the
-	 * chip from rc6 before touching it for real. MI_MODE is masked, hence
-	 * harmless to write 0 into. */
+	/* WaIssueDummyWriteToWakeupFromRC6:ilk Issue a dummy write to wake up
+	 * the chip from rc6 before touching it for real. MI_MODE is masked,
+	 * hence harmless to write 0 into. */
 	I915_WRITE_NOTRACE(MI_MODE, 0);
 }
 
 static void
 hsw_unclaimed_reg_clear(struct drm_i915_private *dev_priv, u32 reg)
 {
-	if (IS_HASWELL(dev_priv->dev) &&
+	if (HAS_FPGA_DBG_UNCLAIMED(dev_priv->dev) &&
 	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
 		DRM_ERROR("Unknown unclaimed register before writing to %x\n",
 			  reg);
@@ -1238,7 +1244,7 @@ hsw_unclaimed_reg_clear(struct drm_i915_private *dev_priv, u32 reg)
 static void
 hsw_unclaimed_reg_check(struct drm_i915_private *dev_priv, u32 reg)
 {
-	if (IS_HASWELL(dev_priv->dev) &&
+	if (HAS_FPGA_DBG_UNCLAIMED(dev_priv->dev) &&
 	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
 		DRM_ERROR("Unclaimed write to %x\n", reg);
 		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 9669a0b..a416645 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -76,6 +76,8 @@ enum plane {
 };
 #define plane_name(p) ((p) + 'A')
 
+#define sprite_name(p, s) ((p) * dev_priv->num_plane + (s) + 'A')
+
 enum port {
 	PORT_A = 0,
 	PORT_B,
@@ -86,6 +88,24 @@ enum port {
 };
 #define port_name(p) ((p) + 'A')
 
+enum intel_display_power_domain {
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_C,
+	POWER_DOMAIN_PIPE_A_PANEL_FITTER,
+	POWER_DOMAIN_PIPE_B_PANEL_FITTER,
+	POWER_DOMAIN_PIPE_C_PANEL_FITTER,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_TRANSCODER_C,
+	POWER_DOMAIN_TRANSCODER_EDP = POWER_DOMAIN_TRANSCODER_A + 0xF,
+};
+
+#define POWER_DOMAIN_PIPE(pipe) ((pipe) + POWER_DOMAIN_PIPE_A)
+#define POWER_DOMAIN_PIPE_PANEL_FITTER(pipe) \
+		((pipe) + POWER_DOMAIN_PIPE_A_PANEL_FITTER)
+#define POWER_DOMAIN_TRANSCODER(tran) ((tran) + POWER_DOMAIN_TRANSCODER_A)
+
 enum hpd_pin {
 	HPD_NONE = 0,
 	HPD_PORT_A = HPD_NONE, /* PORT_A is internal */
@@ -112,15 +132,38 @@ enum hpd_pin {
 	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
 		if ((intel_encoder)->base.crtc == (__crtc))
 
-struct intel_pch_pll {
+struct drm_i915_private;
+
+enum intel_dpll_id {
+	DPLL_ID_PRIVATE = -1, /* non-shared dpll in use */
+	/* real shared dpll ids must be >= 0 */
+	DPLL_ID_PCH_PLL_A,
+	DPLL_ID_PCH_PLL_B,
+};
+#define I915_NUM_PLLS 2
+
+struct intel_dpll_hw_state {
+	uint32_t dpll;
+	uint32_t fp0;
+	uint32_t fp1;
+};
+
+struct intel_shared_dpll {
 	int refcount; /* count of number of CRTCs sharing this PLL */
 	int active; /* count of number of active CRTCs (i.e. DPMS on) */
 	bool on; /* is the PLL actually active? Disabled during modeset */
-	int pll_reg;
-	int fp0_reg;
-	int fp1_reg;
+	const char *name;
+	/* should match the index in the dev_priv->shared_dplls array */
+	enum intel_dpll_id id;
+	struct intel_dpll_hw_state hw_state;
+	void (*enable)(struct drm_i915_private *dev_priv,
+		       struct intel_shared_dpll *pll);
+	void (*disable)(struct drm_i915_private *dev_priv,
+			struct intel_shared_dpll *pll);
+	bool (*get_hw_state)(struct drm_i915_private *dev_priv,
+			     struct intel_shared_dpll *pll,
+			     struct intel_dpll_hw_state *hw_state);
 };
-#define I915_NUM_PLLS 2
 
 /* Used by dp and fdi links */
 struct intel_link_m_n {
@@ -175,7 +218,6 @@ struct opregion_header;
 struct opregion_acpi;
 struct opregion_swsci;
 struct opregion_asle;
-struct drm_i915_private;
 
 struct intel_opregion {
 	struct opregion_header __iomem *header;
@@ -286,6 +328,8 @@ struct drm_i915_error_state {
 
 struct intel_crtc_config;
 struct intel_crtc;
+struct intel_limit;
+struct dpll;
 
 struct drm_i915_display_funcs {
 	bool (*fbc_enabled)(struct drm_device *dev);
@@ -293,11 +337,28 @@ struct drm_i915_display_funcs {
 	void (*disable_fbc)(struct drm_device *dev);
 	int (*get_display_clock_speed)(struct drm_device *dev);
 	int (*get_fifo_size)(struct drm_device *dev, int plane);
+	/**
+	 * find_dpll() - Find the best values for the PLL
+	 * @limit: limits for the PLL
+	 * @crtc: current CRTC
+	 * @target: target frequency in kHz
+	 * @refclk: reference clock frequency in kHz
+	 * @match_clock: if provided, @best_clock P divider must
+	 *               match the P divider from @match_clock
+	 *               used for LVDS downclocking
+	 * @best_clock: best PLL values found
+	 *
+	 * Returns true on success, false on failure.
+	 */
+	bool (*find_dpll)(const struct intel_limit *limit,
+			  struct drm_crtc *crtc,
+			  int target, int refclk,
+			  struct dpll *match_clock,
+			  struct dpll *best_clock);
 	void (*update_wm)(struct drm_device *dev);
 	void (*update_sprite_wm)(struct drm_device *dev, int pipe,
-				 uint32_t sprite_width, int pixel_size);
-	void (*update_linetime_wm)(struct drm_device *dev, int pipe,
-				 struct drm_display_mode *mode);
+				 uint32_t sprite_width, int pixel_size,
+				 bool enable);
 	void (*modeset_global_resources)(struct drm_device *dev);
 	/* Returns the active state of the crtc, and if the crtc is active,
 	 * fills out the pipe-config with the hw state. */
@@ -331,68 +392,56 @@ struct drm_i915_gt_funcs {
 	void (*force_wake_put)(struct drm_i915_private *dev_priv);
 };
 
-#define DEV_INFO_FLAGS \
-	DEV_INFO_FLAG(is_mobile) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_i85x) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_i915g) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_i945gm) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_g33) DEV_INFO_SEP \
-	DEV_INFO_FLAG(need_gfx_hws) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_g4x) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_pineview) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_broadwater) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_crestline) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_ivybridge) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_valleyview) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_haswell) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_force_wake) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_fbc) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_pipe_cxsr) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_hotplug) DEV_INFO_SEP \
-	DEV_INFO_FLAG(cursor_needs_physical) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_overlay) DEV_INFO_SEP \
-	DEV_INFO_FLAG(overlay_needs_physical) DEV_INFO_SEP \
-	DEV_INFO_FLAG(supports_tv) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_bsd_ring) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_blt_ring) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_llc)
+#define DEV_INFO_FOR_EACH_FLAG(func, sep) \
+	func(is_mobile) sep \
+	func(is_i85x) sep \
+	func(is_i915g) sep \
+	func(is_i945gm) sep \
+	func(is_g33) sep \
+	func(need_gfx_hws) sep \
+	func(is_g4x) sep \
+	func(is_pineview) sep \
+	func(is_broadwater) sep \
+	func(is_crestline) sep \
+	func(is_ivybridge) sep \
+	func(is_valleyview) sep \
+	func(is_haswell) sep \
+	func(has_force_wake) sep \
+	func(has_fbc) sep \
+	func(has_pipe_cxsr) sep \
+	func(has_hotplug) sep \
+	func(cursor_needs_physical) sep \
+	func(has_overlay) sep \
+	func(overlay_needs_physical) sep \
+	func(supports_tv) sep \
+	func(has_bsd_ring) sep \
+	func(has_blt_ring) sep \
+	func(has_vebox_ring) sep \
+	func(has_llc) sep \
+	func(has_ddi) sep \
+	func(has_fpga_dbg)
+
+#define DEFINE_FLAG(name) u8 name:1
+#define SEP_SEMICOLON ;
 
 struct intel_device_info {
 	u32 display_mmio_offset;
 	u8 num_pipes:3;
 	u8 gen;
-	u8 is_mobile:1;
-	u8 is_i85x:1;
-	u8 is_i915g:1;
-	u8 is_i945gm:1;
-	u8 is_g33:1;
-	u8 need_gfx_hws:1;
-	u8 is_g4x:1;
-	u8 is_pineview:1;
-	u8 is_broadwater:1;
-	u8 is_crestline:1;
-	u8 is_ivybridge:1;
-	u8 is_valleyview:1;
-	u8 has_force_wake:1;
-	u8 is_haswell:1;
-	u8 has_fbc:1;
-	u8 has_pipe_cxsr:1;
-	u8 has_hotplug:1;
-	u8 cursor_needs_physical:1;
-	u8 has_overlay:1;
-	u8 overlay_needs_physical:1;
-	u8 supports_tv:1;
-	u8 has_bsd_ring:1;
-	u8 has_blt_ring:1;
-	u8 has_llc:1;
+	DEV_INFO_FOR_EACH_FLAG(DEFINE_FLAG, SEP_SEMICOLON);
 };
 
+#undef DEFINE_FLAG
+#undef SEP_SEMICOLON
+
 enum i915_cache_level {
 	I915_CACHE_NONE = 0,
 	I915_CACHE_LLC,
 	I915_CACHE_LLC_MLC, /* gen6+, in docs at least! */
 };
 
+typedef uint32_t gen6_gtt_pte_t;
+
 /* The Graphics Translation Table is the way in which GEN hardware translates a
  * Graphics Virtual Address into a Physical Address. In addition to the normal
  * collateral associated with any va->pa translations GEN hardware also has a
@@ -428,6 +477,9 @@ struct i915_gtt {
 				   struct sg_table *st,
 				   unsigned int pg_start,
 				   enum i915_cache_level cache_level);
+	gen6_gtt_pte_t (*pte_encode)(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level);
 };
 #define gtt_total_entries(gtt) ((gtt).total >> PAGE_SHIFT)
 
@@ -449,19 +501,31 @@ struct i915_hw_ppgtt {
 			       struct sg_table *st,
 			       unsigned int pg_start,
 			       enum i915_cache_level cache_level);
+	gen6_gtt_pte_t (*pte_encode)(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level);
 	int (*enable)(struct drm_device *dev);
 	void (*cleanup)(struct i915_hw_ppgtt *ppgtt);
 };
 
+struct i915_ctx_hang_stats {
+	/* This context had batch pending when hang was declared */
+	unsigned batch_pending;
+
+	/* This context had batch active when hang was declared */
+	unsigned batch_active;
+};
 
 /* This must match up with the value previously used for execbuf2.rsvd1. */
 #define DEFAULT_CONTEXT_ID 0
 struct i915_hw_context {
+	struct kref ref;
 	int id;
 	bool is_initialized;
 	struct drm_i915_file_private *file_priv;
 	struct intel_ring_buffer *ring;
 	struct drm_i915_gem_object *obj;
+	struct i915_ctx_hang_stats hang_stats;
 };
 
 enum no_fbc_reason {
@@ -658,6 +722,7 @@ struct i915_suspend_saved_registers {
 
 struct intel_gen6_power_mgmt {
 	struct work_struct work;
+	struct delayed_work vlv_work;
 	u32 pm_iir;
 	/* lock - irqsave spinlock that protectects the work_struct and
 	 * pm_iir. */
@@ -668,6 +733,7 @@ struct intel_gen6_power_mgmt {
 	u8 cur_delay;
 	u8 min_delay;
 	u8 max_delay;
+	u8 rpe_delay;
 	u8 hw_max;
 
 	struct delayed_work delayed_resume_work;
@@ -704,6 +770,15 @@ struct intel_ilk_power_mgmt {
 	struct drm_i915_gem_object *renderctx;
 };
 
+/* Power well structure for haswell */
+struct i915_power_well {
+	struct drm_device *device;
+	spinlock_t lock;
+	/* power well enable/disable usage count */
+	int count;
+	int i915_request;
+};
+
 struct i915_dri1_state {
 	unsigned allow_batchbuffer : 1;
 	u32 __iomem *gfx_hws_cpu_addr;
@@ -812,14 +887,20 @@ struct i915_gem_mm {
 	u32 object_count;
 };
 
+struct drm_i915_error_state_buf {
+	unsigned bytes;
+	unsigned size;
+	int err;
+	u8 *buf;
+	loff_t start;
+	loff_t pos;
+};
+
 struct i915_gpu_error {
 	/* For hangcheck timer */
 #define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
 #define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
 	struct timer_list hangcheck_timer;
-	int hangcheck_count;
-	uint32_t last_acthd[I915_NUM_RINGS];
-	uint32_t prev_instdone[I915_NUM_INSTDONE_REG];
 
 	/* For reset and error_state handling. */
 	spinlock_t lock;
@@ -875,6 +956,37 @@ enum modeset_restore {
 	MODESET_SUSPENDED,
 };
 
+struct intel_vbt_data {
+	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
+	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
+
+	/* Feature bits */
+	unsigned int int_tv_support:1;
+	unsigned int lvds_dither:1;
+	unsigned int lvds_vbt:1;
+	unsigned int int_crt_support:1;
+	unsigned int lvds_use_ssc:1;
+	unsigned int display_clock_mode:1;
+	unsigned int fdi_rx_polarity_inverted:1;
+	int lvds_ssc_freq;
+	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
+
+	/* eDP */
+	int edp_rate;
+	int edp_lanes;
+	int edp_preemphasis;
+	int edp_vswing;
+	bool edp_initialized;
+	bool edp_support;
+	int edp_bpp;
+	struct edp_power_seq edp_pps;
+
+	int crt_ddc_pin;
+
+	int child_dev_num;
+	struct child_device_config *child_dev;
+};
+
 typedef struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -941,9 +1053,9 @@ typedef struct drm_i915_private {
 			HPD_MARK_DISABLED = 2
 		} hpd_mark;
 	} hpd_stats[HPD_NUM_PINS];
+	u32 hpd_event_bits;
 	struct timer_list hotplug_reenable_timer;
 
-	int num_pch_pll;
 	int num_plane;
 
 	unsigned long cfb_size;
@@ -953,6 +1065,7 @@ typedef struct drm_i915_private {
 	struct intel_fbc_work *fbc_work;
 
 	struct intel_opregion opregion;
+	struct intel_vbt_data vbt;
 
 	/* overlay */
 	struct intel_overlay *overlay;
@@ -962,37 +1075,15 @@ typedef struct drm_i915_private {
 	struct {
 		int level;
 		bool enabled;
+		spinlock_t lock; /* bl registers and the above bl fields */
 		struct backlight_device *device;
 	} backlight;
 
 	/* LVDS info */
 	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
 	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
-
-	/* Feature bits from the VBIOS */
-	unsigned int int_tv_support:1;
-	unsigned int lvds_dither:1;
-	unsigned int lvds_vbt:1;
-	unsigned int int_crt_support:1;
-	unsigned int lvds_use_ssc:1;
-	unsigned int display_clock_mode:1;
-	unsigned int fdi_rx_polarity_inverted:1;
-	int lvds_ssc_freq;
-	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
-	struct {
-		int rate;
-		int lanes;
-		int preemphasis;
-		int vswing;
-
-		bool initialized;
-		bool support;
-		int bpp;
-		struct edp_power_seq pps;
-	} edp;
 	bool no_aux_handshake;
 
-	int crt_ddc_pin;
 	struct drm_i915_fence_reg fence_regs[I915_MAX_NUM_FENCES]; /* assume 965 */
 	int fence_reg_start; /* 4 if userland hasn't ioctl'd us yet */
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
@@ -1020,16 +1111,13 @@ typedef struct drm_i915_private {
 	/* Kernel Modesetting */
 
 	struct sdvo_device_mapping sdvo_mappings[2];
-	/* indicate whether the LVDS_BORDER should be enabled or not */
-	unsigned int lvds_border_bits;
-	/* Panel fitter placement and size for Ironlake+ */
-	u32 pch_pf_pos, pch_pf_size;
 
 	struct drm_crtc *plane_to_crtc_mapping[3];
 	struct drm_crtc *pipe_to_crtc_mapping[3];
 	wait_queue_head_t pending_flip_queue;
 
-	struct intel_pch_pll pch_plls[I915_NUM_PLLS];
+	int num_shared_dpll;
+	struct intel_shared_dpll shared_dplls[I915_NUM_PLLS];
 	struct intel_ddi_plls ddi_plls;
 
 	/* Reclocking support */
@@ -1038,8 +1126,6 @@ typedef struct drm_i915_private {
 	/* indicates the reduced downclock for LVDS*/
 	int lvds_downclock;
 	u16 orig_clock;
-	int child_dev_num;
-	struct child_device_config *child_dev;
 
 	bool mchbar_need_disable;
 
@@ -1052,6 +1138,9 @@ typedef struct drm_i915_private {
 	 * mchdev_lock in intel_pm.c */
 	struct intel_ilk_power_mgmt ips;
 
+	/* Haswell power well */
+	struct i915_power_well power_well;
+
 	enum no_fbc_reason no_fbc_reason;
 
 	struct drm_mm_node *compressed_fb;
@@ -1059,6 +1148,8 @@ typedef struct drm_i915_private {
 
 	struct i915_gpu_error gpu_error;
 
+	struct drm_i915_gem_object *vlv_pctx;
+
 	/* list of fbdev register on this device */
 	struct intel_fbdev *fbdev;
 
@@ -1124,7 +1215,7 @@ struct drm_i915_gem_object {
 	struct drm_mm_node *gtt_space;
 	/** Stolen memory for this object, instead of being backed by shmem. */
 	struct drm_mm_node *stolen;
-	struct list_head gtt_list;
+	struct list_head global_list;
 
 	/** This object's place on the active/inactive lists */
 	struct list_head ring_list;
@@ -1271,9 +1362,18 @@ struct drm_i915_gem_request {
 	/** GEM sequence number associated with this request. */
 	uint32_t seqno;
 
-	/** Postion in the ringbuffer of the end of the request */
+	/** Position in the ringbuffer of the start of the request */
+	u32 head;
+
+	/** Position in the ringbuffer of the end of the request */
 	u32 tail;
 
+	/** Context related to this request */
+	struct i915_hw_context *ctx;
+
+	/** Batch buffer related to this request if any */
+	struct drm_i915_gem_object *batch_obj;
+
 	/** Time at which this request was emitted, in jiffies. */
 	unsigned long emitted_jiffies;
 
@@ -1291,6 +1391,8 @@ struct drm_i915_file_private {
 		struct list_head request_list;
 	} mm;
 	struct idr context_idr;
+
+	struct i915_ctx_hang_stats hang_stats;
 };
 
 #define INTEL_INFO(dev)	(((struct drm_i915_private *) (dev)->dev_private)->info)
@@ -1341,6 +1443,7 @@ struct drm_i915_file_private {
 
 #define HAS_BSD(dev)            (INTEL_INFO(dev)->has_bsd_ring)
 #define HAS_BLT(dev)            (INTEL_INFO(dev)->has_blt_ring)
+#define HAS_VEBOX(dev)          (INTEL_INFO(dev)->has_vebox_ring)
 #define HAS_LLC(dev)            (INTEL_INFO(dev)->has_llc)
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
 
@@ -1371,10 +1474,13 @@ struct drm_i915_file_private {
 #define HAS_PIPE_CXSR(dev) (INTEL_INFO(dev)->has_pipe_cxsr)
 #define I915_HAS_FBC(dev) (INTEL_INFO(dev)->has_fbc)
 
+#define HAS_IPS(dev)		(IS_ULT(dev))
+
 #define HAS_PIPE_CONTROL(dev) (INTEL_INFO(dev)->gen >= 5)
 
-#define HAS_DDI(dev)		(IS_HASWELL(dev))
+#define HAS_DDI(dev)		(INTEL_INFO(dev)->has_ddi)
 #define HAS_POWER_WELL(dev)	(IS_HASWELL(dev))
+#define HAS_FPGA_DBG_UNCLAIMED(dev)	(INTEL_INFO(dev)->has_fpga_dbg)
 
 #define INTEL_PCH_DEVICE_ID_MASK		0xff00
 #define INTEL_PCH_IBX_DEVICE_ID_TYPE		0x3b00
@@ -1435,6 +1541,7 @@ extern bool i915_enable_hangcheck __read_mostly;
 extern int i915_enable_ppgtt __read_mostly;
 extern unsigned int i915_preliminary_hw_support __read_mostly;
 extern int i915_disable_power_well __read_mostly;
+extern int i915_enable_ips __read_mostly;
 
 extern int i915_suspend(struct drm_device *dev, pm_message_t state);
 extern int i915_resume(struct drm_device *dev);
@@ -1486,8 +1593,6 @@ i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
 void
 i915_disable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
 
-void intel_enable_asle(struct drm_device *dev);
-
 #ifdef CONFIG_DEBUG_FS
 extern void i915_destroy_error_state(struct drm_device *dev);
 #else
@@ -1626,6 +1731,7 @@ i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
 {
 	if (obj->fence_reg != I915_FENCE_REG_NONE) {
 		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+		WARN_ON(dev_priv->fence_regs[obj->fence_reg].pin_count <= 0);
 		dev_priv->fence_regs[obj->fence_reg].pin_count--;
 	}
 }
@@ -1658,9 +1764,12 @@ void i915_gem_init_swizzling(struct drm_device *dev);
 void i915_gem_cleanup_ringbuffer(struct drm_device *dev);
 int __must_check i915_gpu_idle(struct drm_device *dev);
 int __must_check i915_gem_idle(struct drm_device *dev);
-int i915_add_request(struct intel_ring_buffer *ring,
-		     struct drm_file *file,
-		     u32 *seqno);
+int __i915_add_request(struct intel_ring_buffer *ring,
+		       struct drm_file *file,
+		       struct drm_i915_gem_object *batch_obj,
+		       u32 *seqno);
+#define i915_add_request(ring, seqno) \
+	__i915_add_request(ring, NULL, NULL, seqno)
 int __must_check i915_wait_seqno(struct intel_ring_buffer *ring,
 				 uint32_t seqno);
 int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
@@ -1705,6 +1814,21 @@ void i915_gem_context_fini(struct drm_device *dev);
 void i915_gem_context_close(struct drm_device *dev, struct drm_file *file);
 int i915_switch_context(struct intel_ring_buffer *ring,
 			struct drm_file *file, int to_id);
+void i915_gem_context_free(struct kref *ctx_ref);
+static inline void i915_gem_context_reference(struct i915_hw_context *ctx)
+{
+	kref_get(&ctx->ref);
+}
+
+static inline void i915_gem_context_unreference(struct i915_hw_context *ctx)
+{
+	kref_put(&ctx->ref, i915_gem_context_free);
+}
+
+struct i915_ctx_hang_stats * __must_check
+i915_gem_context_get_hang_stats(struct intel_ring_buffer *ring,
+				struct drm_file *file,
+				u32 id);
 int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
 				  struct drm_file *file);
 int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
@@ -1786,6 +1910,8 @@ void i915_gem_dump_object(struct drm_i915_gem_object *obj, int len,
 /* i915_debugfs.c */
 int i915_debugfs_init(struct drm_minor *minor);
 void i915_debugfs_cleanup(struct drm_minor *minor);
+__printf(2, 3)
+void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...);
 
 /* i915_suspend.c */
 extern int i915_save_state(struct drm_device *dev);
@@ -1802,7 +1928,7 @@ void i915_teardown_sysfs(struct drm_device *dev_priv);
 /* intel_i2c.c */
 extern int intel_setup_gmbus(struct drm_device *dev);
 extern void intel_teardown_gmbus(struct drm_device *dev);
-extern inline bool intel_gmbus_is_port_valid(unsigned port)
+static inline bool intel_gmbus_is_port_valid(unsigned port)
 {
 	return (port >= GMBUS_PORT_SSC && port <= GMBUS_PORT_DPD);
 }
@@ -1811,7 +1937,7 @@ extern struct i2c_adapter *intel_gmbus_get_adapter(
 		struct drm_i915_private *dev_priv, unsigned port);
 extern void intel_gmbus_set_speed(struct i2c_adapter *adapter, int speed);
 extern void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit);
-extern inline bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter)
+static inline bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter)
 {
 	return container_of(adapter, struct intel_gmbus, adapter)->force_bit;
 }
@@ -1823,14 +1949,10 @@ extern int intel_opregion_setup(struct drm_device *dev);
 extern void intel_opregion_init(struct drm_device *dev);
 extern void intel_opregion_fini(struct drm_device *dev);
 extern void intel_opregion_asle_intr(struct drm_device *dev);
-extern void intel_opregion_gse_intr(struct drm_device *dev);
-extern void intel_opregion_enable_asle(struct drm_device *dev);
 #else
 static inline void intel_opregion_init(struct drm_device *dev) { return; }
 static inline void intel_opregion_fini(struct drm_device *dev) { return; }
 static inline void intel_opregion_asle_intr(struct drm_device *dev) { return; }
-static inline void intel_opregion_gse_intr(struct drm_device *dev) { return; }
-static inline void intel_opregion_enable_asle(struct drm_device *dev) { return; }
 #endif
 
 /* intel_acpi.c */
@@ -1844,6 +1966,7 @@ static inline void intel_unregister_dsm_handler(void) { return; }
 
 /* modesetting */
 extern void intel_modeset_init_hw(struct drm_device *dev);
+extern void intel_modeset_suspend_hw(struct drm_device *dev);
 extern void intel_modeset_init(struct drm_device *dev);
 extern void intel_modeset_gem_init(struct drm_device *dev);
 extern void intel_modeset_cleanup(struct drm_device *dev);
@@ -1856,6 +1979,9 @@ extern void intel_disable_fbc(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
 extern void intel_init_pch_refclk(struct drm_device *dev);
 extern void gen6_set_rps(struct drm_device *dev, u8 val);
+extern void valleyview_set_rps(struct drm_device *dev, u8 val);
+extern int valleyview_rps_max_freq(struct drm_i915_private *dev_priv);
+extern int valleyview_rps_min_freq(struct drm_i915_private *dev_priv);
 extern void intel_detect_pch(struct drm_device *dev);
 extern int intel_trans_dp_port_sel(struct drm_crtc *crtc);
 extern int intel_enable_rc6(const struct drm_device *dev);
@@ -1867,10 +1993,11 @@ int i915_reg_read_ioctl(struct drm_device *dev, void *data,
 /* overlay */
 #ifdef CONFIG_DEBUG_FS
 extern struct intel_overlay_error_state *intel_overlay_capture_error_state(struct drm_device *dev);
-extern void intel_overlay_print_error_state(struct seq_file *m, struct intel_overlay_error_state *error);
+extern void intel_overlay_print_error_state(struct drm_i915_error_state_buf *e,
+					    struct intel_overlay_error_state *error);
 
 extern struct intel_display_error_state *intel_display_capture_error_state(struct drm_device *dev);
-extern void intel_display_print_error_state(struct seq_file *m,
+extern void intel_display_print_error_state(struct drm_i915_error_state_buf *e,
 					    struct drm_device *dev,
 					    struct intel_display_error_state *error);
 #endif
@@ -1885,8 +2012,20 @@ int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
 
 int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val);
 int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val);
-int valleyview_punit_read(struct drm_i915_private *dev_priv, u8 addr, u32 *val);
-int valleyview_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val);
+
+/* intel_sideband.c */
+u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr);
+void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val);
+u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr);
+u32 vlv_dpio_read(struct drm_i915_private *dev_priv, int reg);
+void vlv_dpio_write(struct drm_i915_private *dev_priv, int reg, u32 val);
+u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
+		   enum intel_sbi_destination destination);
+void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
+		     enum intel_sbi_destination destination);
+
+int vlv_gpu_freq(int ddr_freq, int val);
+int vlv_freq_opcode(int ddr_freq, int val);
 
 #define __i915_read(x, y) \
 	u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 9e35daf..4200c32 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -176,7 +176,7 @@ i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 
 	pinned = 0;
 	mutex_lock(&dev->struct_mutex);
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list)
 		if (obj->pin_count)
 			pinned += obj->gtt_space->size;
 	mutex_unlock(&dev->struct_mutex);
@@ -956,7 +956,7 @@ i915_gem_check_olr(struct intel_ring_buffer *ring, u32 seqno)
 
 	ret = 0;
 	if (seqno == ring->outstanding_lazy_request)
-		ret = i915_add_request(ring, NULL, NULL);
+		ret = i915_add_request(ring, NULL);
 
 	return ret;
 }
@@ -1087,6 +1087,25 @@ i915_wait_seqno(struct intel_ring_buffer *ring, uint32_t seqno)
 			    interruptible, NULL);
 }
 
+static int
+i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj,
+				     struct intel_ring_buffer *ring)
+{
+	i915_gem_retire_requests_ring(ring);
+
+	/* Manually manage the write flush as we may have not yet
+	 * retired the buffer.
+	 *
+	 * Note that the last_write_seqno is always the earlier of
+	 * the two (read/write) seqno, so if we haved successfully waited,
+	 * we know we have passed the last write.
+	 */
+	obj->last_write_seqno = 0;
+	obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
+
+	return 0;
+}
+
 /**
  * Ensures that all rendering to the object has completed and the object is
  * safe to unbind from the GTT or access from the CPU.
@@ -1107,18 +1126,7 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj,
 	if (ret)
 		return ret;
 
-	i915_gem_retire_requests_ring(ring);
-
-	/* Manually manage the write flush as we may have not yet
-	 * retired the buffer.
-	 */
-	if (obj->last_write_seqno &&
-	    i915_seqno_passed(seqno, obj->last_write_seqno)) {
-		obj->last_write_seqno = 0;
-		obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
-	}
-
-	return 0;
+	return i915_gem_object_wait_rendering__tail(obj, ring);
 }
 
 /* A nonblocking variant of the above wait. This is a highly dangerous routine
@@ -1154,19 +1162,10 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 	mutex_unlock(&dev->struct_mutex);
 	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL);
 	mutex_lock(&dev->struct_mutex);
+	if (ret)
+		return ret;
 
-	i915_gem_retire_requests_ring(ring);
-
-	/* Manually manage the write flush as we may have not yet
-	 * retired the buffer.
-	 */
-	if (obj->last_write_seqno &&
-	    i915_seqno_passed(seqno, obj->last_write_seqno)) {
-		obj->last_write_seqno = 0;
-		obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
-	}
-
-	return ret;
+	return i915_gem_object_wait_rendering__tail(obj, ring);
 }
 
 /**
@@ -1676,7 +1675,7 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
 	/* ->put_pages might need to allocate memory for the bit17 swizzle
 	 * array, hence protect them from being reaped by removing them from gtt
 	 * lists early. */
-	list_del(&obj->gtt_list);
+	list_del(&obj->global_list);
 
 	ops->put_pages(obj);
 	obj->pages = NULL;
@@ -1696,7 +1695,7 @@ __i915_gem_shrink(struct drm_i915_private *dev_priv, long target,
 
 	list_for_each_entry_safe(obj, next,
 				 &dev_priv->mm.unbound_list,
-				 gtt_list) {
+				 global_list) {
 		if ((i915_gem_object_is_purgeable(obj) || !purgeable_only) &&
 		    i915_gem_object_put_pages(obj) == 0) {
 			count += obj->base.size >> PAGE_SHIFT;
@@ -1733,7 +1732,8 @@ i915_gem_shrink_all(struct drm_i915_private *dev_priv)
 
 	i915_gem_evict_everything(dev_priv->dev);
 
-	list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list, gtt_list)
+	list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list,
+				 global_list)
 		i915_gem_object_put_pages(obj);
 }
 
@@ -1867,7 +1867,7 @@ i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
 	if (ret)
 		return ret;
 
-	list_add_tail(&obj->gtt_list, &dev_priv->mm.unbound_list);
+	list_add_tail(&obj->global_list, &dev_priv->mm.unbound_list);
 	return 0;
 }
 
@@ -2005,17 +2005,18 @@ i915_gem_get_seqno(struct drm_device *dev, u32 *seqno)
 	return 0;
 }
 
-int
-i915_add_request(struct intel_ring_buffer *ring,
-		 struct drm_file *file,
-		 u32 *out_seqno)
+int __i915_add_request(struct intel_ring_buffer *ring,
+		       struct drm_file *file,
+		       struct drm_i915_gem_object *obj,
+		       u32 *out_seqno)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_request *request;
-	u32 request_ring_position;
+	u32 request_ring_position, request_start;
 	int was_empty;
 	int ret;
 
+	request_start = intel_ring_get_tail(ring);
 	/*
 	 * Emit any outstanding flushes - execbuf can fail to emit the flush
 	 * after having emitted the batchbuffer command. Hence we need to fix
@@ -2047,7 +2048,21 @@ i915_add_request(struct intel_ring_buffer *ring,
 
 	request->seqno = intel_ring_get_seqno(ring);
 	request->ring = ring;
+	request->head = request_start;
 	request->tail = request_ring_position;
+	request->ctx = ring->last_context;
+	request->batch_obj = obj;
+
+	/* Whilst this request exists, batch_obj will be on the
+	 * active_list, and so will hold the active reference. Only when this
+	 * request is retired will the the batch_obj be moved onto the
+	 * inactive_list and lose its active reference. Hence we do not need
+	 * to explicitly hold another reference here.
+	 */
+
+	if (request->ctx)
+		i915_gem_context_reference(request->ctx);
+
 	request->emitted_jiffies = jiffies;
 	was_empty = list_empty(&ring->request_list);
 	list_add_tail(&request->list, &ring->request_list);
@@ -2100,9 +2115,114 @@ i915_gem_request_remove_from_client(struct drm_i915_gem_request *request)
 	spin_unlock(&file_priv->mm.lock);
 }
 
+static bool i915_head_inside_object(u32 acthd, struct drm_i915_gem_object *obj)
+{
+	if (acthd >= obj->gtt_offset &&
+	    acthd < obj->gtt_offset + obj->base.size)
+		return true;
+
+	return false;
+}
+
+static bool i915_head_inside_request(const u32 acthd_unmasked,
+				     const u32 request_start,
+				     const u32 request_end)
+{
+	const u32 acthd = acthd_unmasked & HEAD_ADDR;
+
+	if (request_start < request_end) {
+		if (acthd >= request_start && acthd < request_end)
+			return true;
+	} else if (request_start > request_end) {
+		if (acthd >= request_start || acthd < request_end)
+			return true;
+	}
+
+	return false;
+}
+
+static bool i915_request_guilty(struct drm_i915_gem_request *request,
+				const u32 acthd, bool *inside)
+{
+	/* There is a possibility that unmasked head address
+	 * pointing inside the ring, matches the batch_obj address range.
+	 * However this is extremely unlikely.
+	 */
+
+	if (request->batch_obj) {
+		if (i915_head_inside_object(acthd, request->batch_obj)) {
+			*inside = true;
+			return true;
+		}
+	}
+
+	if (i915_head_inside_request(acthd, request->head, request->tail)) {
+		*inside = false;
+		return true;
+	}
+
+	return false;
+}
+
+static void i915_set_reset_status(struct intel_ring_buffer *ring,
+				  struct drm_i915_gem_request *request,
+				  u32 acthd)
+{
+	struct i915_ctx_hang_stats *hs = NULL;
+	bool inside, guilty;
+
+	/* Innocent until proven guilty */
+	guilty = false;
+
+	if (ring->hangcheck.action != wait &&
+	    i915_request_guilty(request, acthd, &inside)) {
+		DRM_ERROR("%s hung %s bo (0x%x ctx %d) at 0x%x\n",
+			  ring->name,
+			  inside ? "inside" : "flushing",
+			  request->batch_obj ?
+			  request->batch_obj->gtt_offset : 0,
+			  request->ctx ? request->ctx->id : 0,
+			  acthd);
+
+		guilty = true;
+	}
+
+	/* If contexts are disabled or this is the default context, use
+	 * file_priv->reset_state
+	 */
+	if (request->ctx && request->ctx->id != DEFAULT_CONTEXT_ID)
+		hs = &request->ctx->hang_stats;
+	else if (request->file_priv)
+		hs = &request->file_priv->hang_stats;
+
+	if (hs) {
+		if (guilty)
+			hs->batch_active++;
+		else
+			hs->batch_pending++;
+	}
+}
+
+static void i915_gem_free_request(struct drm_i915_gem_request *request)
+{
+	list_del(&request->list);
+	i915_gem_request_remove_from_client(request);
+
+	if (request->ctx)
+		i915_gem_context_unreference(request->ctx);
+
+	kfree(request);
+}
+
 static void i915_gem_reset_ring_lists(struct drm_i915_private *dev_priv,
 				      struct intel_ring_buffer *ring)
 {
+	u32 completed_seqno;
+	u32 acthd;
+
+	acthd = intel_ring_get_active_head(ring);
+	completed_seqno = ring->get_seqno(ring, false);
+
 	while (!list_empty(&ring->request_list)) {
 		struct drm_i915_gem_request *request;
 
@@ -2110,9 +2230,10 @@ static void i915_gem_reset_ring_lists(struct drm_i915_private *dev_priv,
 					   struct drm_i915_gem_request,
 					   list);
 
-		list_del(&request->list);
-		i915_gem_request_remove_from_client(request);
-		kfree(request);
+		if (request->seqno > completed_seqno)
+			i915_set_reset_status(ring, request, acthd);
+
+		i915_gem_free_request(request);
 	}
 
 	while (!list_empty(&ring->active_list)) {
@@ -2193,9 +2314,7 @@ i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 		 */
 		ring->last_retired_head = request->tail;
 
-		list_del(&request->list);
-		i915_gem_request_remove_from_client(request);
-		kfree(request);
+		i915_gem_free_request(request);
 	}
 
 	/* Move any buffers on the active list that are no longer referenced
@@ -2262,7 +2381,7 @@ i915_gem_retire_work_handler(struct work_struct *work)
 	idle = true;
 	for_each_ring(ring, dev_priv, i) {
 		if (ring->gpu_caches_dirty)
-			i915_add_request(ring, NULL, NULL);
+			i915_add_request(ring, NULL);
 
 		idle &= list_empty(&ring->request_list);
 	}
@@ -2494,9 +2613,10 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 		obj->has_aliasing_ppgtt_mapping = 0;
 	}
 	i915_gem_gtt_finish_object(obj);
+	i915_gem_object_unpin_pages(obj);
 
 	list_del(&obj->mm_list);
-	list_move_tail(&obj->gtt_list, &dev_priv->mm.unbound_list);
+	list_move_tail(&obj->global_list, &dev_priv->mm.unbound_list);
 	/* Avoid an unnecessary call to unbind on rebind. */
 	obj->map_and_fenceable = true;
 
@@ -2676,18 +2796,33 @@ static inline int fence_number(struct drm_i915_private *dev_priv,
 	return fence - dev_priv->fence_regs;
 }
 
+struct write_fence {
+	struct drm_device *dev;
+	struct drm_i915_gem_object *obj;
+	int fence;
+};
+
 static void i915_gem_write_fence__ipi(void *data)
 {
+	struct write_fence *args = data;
+
+	/* Required for SNB+ with LLC */
 	wbinvd();
+
+	/* Required for VLV */
+	i915_gem_write_fence(args->dev, args->fence, args->obj);
 }
 
 static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 					 struct drm_i915_fence_reg *fence,
 					 bool enable)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int fence_reg = fence_number(dev_priv, fence);
+	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	struct write_fence args = {
+		.dev = obj->base.dev,
+		.fence = fence_number(dev_priv, fence),
+		.obj = enable ? obj : NULL,
+	};
 
 	/* In order to fully serialize access to the fenced region and
 	 * the update to the fence register we need to take extreme
@@ -2698,13 +2833,19 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 	 * SNB+ we need to take a step further and emit an explicit wbinvd()
 	 * on each processor in order to manually flush all memory
 	 * transactions before updating the fence register.
+	 *
+	 * However, Valleyview complicates matter. There the wbinvd is
+	 * insufficient and unlike SNB/IVB requires the serialising
+	 * register write. (Note that that register write by itself is
+	 * conversely not sufficient for SNB+.) To compromise, we do both.
 	 */
-	if (HAS_LLC(obj->base.dev))
-		on_each_cpu(i915_gem_write_fence__ipi, NULL, 1);
-	i915_gem_write_fence(dev, fence_reg, enable ? obj : NULL);
+	if (INTEL_INFO(args.dev)->gen >= 6)
+		on_each_cpu(i915_gem_write_fence__ipi, &args, 1);
+	else
+		i915_gem_write_fence(args.dev, args.fence, args.obj);
 
 	if (enable) {
-		obj->fence_reg = fence_reg;
+		obj->fence_reg = args.fence;
 		fence->obj = obj;
 		list_move_tail(&fence->lru_list, &dev_priv->mm.fence_list);
 	} else {
@@ -2883,7 +3024,7 @@ static void i915_gem_verify_gtt(struct drm_device *dev)
 	struct drm_i915_gem_object *obj;
 	int err = 0;
 
-	list_for_each_entry(obj, &dev_priv->mm.gtt_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.gtt_list, global_list) {
 		if (obj->gtt_space == NULL) {
 			printk(KERN_ERR "object found on GTT list with no space reserved\n");
 			err++;
@@ -2930,6 +3071,8 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	struct drm_mm_node *node;
 	u32 size, fence_size, fence_alignment, unfenced_alignment;
 	bool mappable, fenceable;
+	size_t gtt_max = map_and_fenceable ?
+		dev_priv->gtt.mappable_end : dev_priv->gtt.total;
 	int ret;
 
 	fence_size = i915_gem_get_gtt_size(dev,
@@ -2956,9 +3099,11 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	/* If the object is bigger than the entire aperture, reject it early
 	 * before evicting everything in a vain attempt to find space.
 	 */
-	if (obj->base.size >
-	    (map_and_fenceable ? dev_priv->gtt.mappable_end : dev_priv->gtt.total)) {
-		DRM_ERROR("Attempting to bind an object larger than the aperture\n");
+	if (obj->base.size > gtt_max) {
+		DRM_ERROR("Attempting to bind an object larger than the aperture: object=%zd > %s aperture=%zu\n",
+			  obj->base.size,
+			  map_and_fenceable ? "mappable" : "total",
+			  gtt_max);
 		return -E2BIG;
 	}
 
@@ -2974,14 +3119,10 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		return -ENOMEM;
 	}
 
- search_free:
-	if (map_and_fenceable)
-		ret = drm_mm_insert_node_in_range_generic(&dev_priv->mm.gtt_space, node,
-							  size, alignment, obj->cache_level,
-							  0, dev_priv->gtt.mappable_end);
-	else
-		ret = drm_mm_insert_node_generic(&dev_priv->mm.gtt_space, node,
-						 size, alignment, obj->cache_level);
+search_free:
+	ret = drm_mm_insert_node_in_range_generic(&dev_priv->mm.gtt_space, node,
+						  size, alignment,
+						  obj->cache_level, 0, gtt_max);
 	if (ret) {
 		ret = i915_gem_evict_something(dev, size, alignment,
 					       obj->cache_level,
@@ -3007,7 +3148,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		return ret;
 	}
 
-	list_move_tail(&obj->gtt_list, &dev_priv->mm.bound_list);
+	list_move_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&obj->mm_list, &dev_priv->mm.inactive_list);
 
 	obj->gtt_space = node;
@@ -3022,7 +3163,6 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 
 	obj->map_and_fenceable = mappable && fenceable;
 
-	i915_gem_object_unpin_pages(obj);
 	trace_i915_gem_object_bind(obj, map_and_fenceable);
 	i915_gem_verify_gtt(dev);
 	return 0;
@@ -3722,7 +3862,7 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
 			  const struct drm_i915_gem_object_ops *ops)
 {
 	INIT_LIST_HEAD(&obj->mm_list);
-	INIT_LIST_HEAD(&obj->gtt_list);
+	INIT_LIST_HEAD(&obj->global_list);
 	INIT_LIST_HEAD(&obj->ring_list);
 	INIT_LIST_HEAD(&obj->exec_list);
 
@@ -3822,7 +3962,13 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 		dev_priv->mm.interruptible = was_interruptible;
 	}
 
-	obj->pages_pin_count = 0;
+	/* Stolen objects don't hold a ref, but do hold pin count. Fix that up
+	 * before progressing. */
+	if (obj->stolen)
+		i915_gem_object_unpin_pages(obj);
+
+	if (WARN_ON(obj->pages_pin_count))
+		obj->pages_pin_count = 0;
 	i915_gem_object_put_pages(obj);
 	i915_gem_object_free_mmap_offset(obj);
 	i915_gem_object_release_stolen(obj);
@@ -3973,12 +4119,21 @@ static int i915_gem_init_rings(struct drm_device *dev)
 			goto cleanup_bsd_ring;
 	}
 
+	if (HAS_VEBOX(dev)) {
+		ret = intel_init_vebox_ring_buffer(dev);
+		if (ret)
+			goto cleanup_blt_ring;
+	}
+
+
 	ret = i915_gem_set_seqno(dev, ((u32)~0 - 0x1000));
 	if (ret)
-		goto cleanup_blt_ring;
+		goto cleanup_vebox_ring;
 
 	return 0;
 
+cleanup_vebox_ring:
+	intel_cleanup_ring_buffer(&dev_priv->ring[VECS]);
 cleanup_blt_ring:
 	intel_cleanup_ring_buffer(&dev_priv->ring[BCS]);
 cleanup_bsd_ring:
@@ -4453,10 +4608,10 @@ i915_gem_inactive_shrink(struct shrinker *shrinker, struct shrink_control *sc)
 	}
 
 	cnt = 0;
-	list_for_each_entry(obj, &dev_priv->mm.unbound_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list)
 		if (obj->pages_pin_count == 0)
 			cnt += obj->base.size >> PAGE_SHIFT;
-	list_for_each_entry(obj, &dev_priv->mm.inactive_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.inactive_list, global_list)
 		if (obj->pin_count == 0 && obj->pages_pin_count == 0)
 			cnt += obj->base.size >> PAGE_SHIFT;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index a1e8ecb..51b7a21 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -113,7 +113,7 @@ static int get_context_size(struct drm_device *dev)
 	case 7:
 		reg = I915_READ(GEN7_CXT_SIZE);
 		if (IS_HASWELL(dev))
-			ret = HSW_CXT_TOTAL_SIZE(reg) * 64;
+			ret = HSW_CXT_TOTAL_SIZE;
 		else
 			ret = GEN7_CXT_TOTAL_SIZE(reg) * 64;
 		break;
@@ -124,10 +124,10 @@ static int get_context_size(struct drm_device *dev)
 	return ret;
 }
 
-static void do_destroy(struct i915_hw_context *ctx)
+void i915_gem_context_free(struct kref *ctx_ref)
 {
-	if (ctx->file_priv)
-		idr_remove(&ctx->file_priv->context_idr, ctx->id);
+	struct i915_hw_context *ctx = container_of(ctx_ref,
+						   typeof(*ctx), ref);
 
 	drm_gem_object_unreference(&ctx->obj->base);
 	kfree(ctx);
@@ -145,6 +145,7 @@ create_hw_context(struct drm_device *dev,
 	if (ctx == NULL)
 		return ERR_PTR(-ENOMEM);
 
+	kref_init(&ctx->ref);
 	ctx->obj = i915_gem_alloc_object(dev, dev_priv->hw_context_size);
 	if (ctx->obj == NULL) {
 		kfree(ctx);
@@ -155,7 +156,8 @@ create_hw_context(struct drm_device *dev,
 	if (INTEL_INFO(dev)->gen >= 7) {
 		ret = i915_gem_object_set_cache_level(ctx->obj,
 						      I915_CACHE_LLC_MLC);
-		if (ret)
+		/* Failure shouldn't ever happen this early */
+		if (WARN_ON(ret))
 			goto err_out;
 	}
 
@@ -169,18 +171,18 @@ create_hw_context(struct drm_device *dev,
 	if (file_priv == NULL)
 		return ctx;
 
-	ctx->file_priv = file_priv;
-
 	ret = idr_alloc(&file_priv->context_idr, ctx, DEFAULT_CONTEXT_ID + 1, 0,
 			GFP_KERNEL);
 	if (ret < 0)
 		goto err_out;
+
+	ctx->file_priv = file_priv;
 	ctx->id = ret;
 
 	return ctx;
 
 err_out:
-	do_destroy(ctx);
+	i915_gem_context_unreference(ctx);
 	return ERR_PTR(ret);
 }
 
@@ -213,12 +215,16 @@ static int create_default_context(struct drm_i915_private *dev_priv)
 	 */
 	dev_priv->ring[RCS].default_context = ctx;
 	ret = i915_gem_object_pin(ctx->obj, CONTEXT_ALIGN, false, false);
-	if (ret)
+	if (ret) {
+		DRM_DEBUG_DRIVER("Couldn't pin %d\n", ret);
 		goto err_destroy;
+	}
 
 	ret = do_switch(ctx);
-	if (ret)
+	if (ret) {
+		DRM_DEBUG_DRIVER("Switch failed %d\n", ret);
 		goto err_unpin;
+	}
 
 	DRM_DEBUG_DRIVER("Default HW context loaded\n");
 	return 0;
@@ -226,7 +232,7 @@ static int create_default_context(struct drm_i915_private *dev_priv)
 err_unpin:
 	i915_gem_object_unpin(ctx->obj);
 err_destroy:
-	do_destroy(ctx);
+	i915_gem_context_unreference(ctx);
 	return ret;
 }
 
@@ -236,6 +242,7 @@ void i915_gem_context_init(struct drm_device *dev)
 
 	if (!HAS_HW_CONTEXTS(dev)) {
 		dev_priv->hw_contexts_disabled = true;
+		DRM_DEBUG_DRIVER("Disabling HW Contexts; old hardware\n");
 		return;
 	}
 
@@ -248,11 +255,13 @@ void i915_gem_context_init(struct drm_device *dev)
 
 	if (dev_priv->hw_context_size > (1<<20)) {
 		dev_priv->hw_contexts_disabled = true;
+		DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size\n");
 		return;
 	}
 
 	if (create_default_context(dev_priv)) {
 		dev_priv->hw_contexts_disabled = true;
+		DRM_DEBUG_DRIVER("Disabling HW Contexts; create failed\n");
 		return;
 	}
 
@@ -262,6 +271,7 @@ void i915_gem_context_init(struct drm_device *dev)
 void i915_gem_context_fini(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_hw_context *dctx = dev_priv->ring[RCS].default_context;
 
 	if (dev_priv->hw_contexts_disabled)
 		return;
@@ -271,9 +281,16 @@ void i915_gem_context_fini(struct drm_device *dev)
 	 * other code, leading to spurious errors. */
 	intel_gpu_reset(dev);
 
-	i915_gem_object_unpin(dev_priv->ring[RCS].default_context->obj);
+	i915_gem_object_unpin(dctx->obj);
 
-	do_destroy(dev_priv->ring[RCS].default_context);
+	/* When default context is created and switched to, base object refcount
+	 * will be 2 (+1 from object creation and +1 from do_switch()).
+	 * i915_gem_context_fini() will be called after gpu_idle() has switched
+	 * to default context. So we need to unreference the base object once
+	 * to offset the do_switch part, so that i915_gem_context_unreference()
+	 * can then free the base object correctly. */
+	drm_gem_object_unreference(&dctx->obj->base);
+	i915_gem_context_unreference(dctx);
 }
 
 static int context_idr_cleanup(int id, void *p, void *data)
@@ -282,11 +299,38 @@ static int context_idr_cleanup(int id, void *p, void *data)
 
 	BUG_ON(id == DEFAULT_CONTEXT_ID);
 
-	do_destroy(ctx);
-
+	i915_gem_context_unreference(ctx);
 	return 0;
 }
 
+struct i915_ctx_hang_stats *
+i915_gem_context_get_hang_stats(struct intel_ring_buffer *ring,
+				struct drm_file *file,
+				u32 id)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct i915_hw_context *to;
+
+	if (dev_priv->hw_contexts_disabled)
+		return ERR_PTR(-ENOENT);
+
+	if (ring->id != RCS)
+		return ERR_PTR(-EINVAL);
+
+	if (file == NULL)
+		return ERR_PTR(-EINVAL);
+
+	if (id == DEFAULT_CONTEXT_ID)
+		return &file_priv->hang_stats;
+
+	to = i915_gem_context_get(file->driver_priv, id);
+	if (to == NULL)
+		return ERR_PTR(-ENOENT);
+
+	return &to->hang_stats;
+}
+
 void i915_gem_context_close(struct drm_device *dev, struct drm_file *file)
 {
 	struct drm_i915_file_private *file_priv = file->driver_priv;
@@ -325,6 +369,7 @@ mi_set_context(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
+	/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw */
 	if (IS_GEN7(ring->dev))
 		intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_DISABLE);
 	else
@@ -353,13 +398,13 @@ mi_set_context(struct intel_ring_buffer *ring,
 static int do_switch(struct i915_hw_context *to)
 {
 	struct intel_ring_buffer *ring = to->ring;
-	struct drm_i915_gem_object *from_obj = ring->last_context_obj;
+	struct i915_hw_context *from = ring->last_context;
 	u32 hw_flags = 0;
 	int ret;
 
-	BUG_ON(from_obj != NULL && from_obj->pin_count == 0);
+	BUG_ON(from != NULL && from->obj != NULL && from->obj->pin_count == 0);
 
-	if (from_obj == to->obj)
+	if (from == to)
 		return 0;
 
 	ret = i915_gem_object_pin(to->obj, CONTEXT_ALIGN, false, false);
@@ -382,7 +427,7 @@ static int do_switch(struct i915_hw_context *to)
 
 	if (!to->is_initialized || is_default_context(to))
 		hw_flags |= MI_RESTORE_INHIBIT;
-	else if (WARN_ON_ONCE(from_obj == to->obj)) /* not yet expected */
+	else if (WARN_ON_ONCE(from == to)) /* not yet expected */
 		hw_flags |= MI_FORCE_RESTORE;
 
 	ret = mi_set_context(ring, to, hw_flags);
@@ -397,9 +442,9 @@ static int do_switch(struct i915_hw_context *to)
 	 * is a bit suboptimal because the retiring can occur simply after the
 	 * MI_SET_CONTEXT instead of when the next seqno has completed.
 	 */
-	if (from_obj != NULL) {
-		from_obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
-		i915_gem_object_move_to_active(from_obj, ring);
+	if (from != NULL) {
+		from->obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
+		i915_gem_object_move_to_active(from->obj, ring);
 		/* As long as MI_SET_CONTEXT is serializing, ie. it flushes the
 		 * whole damn pipeline, we don't need to explicitly mark the
 		 * object dirty. The only exception is that the context must be
@@ -407,15 +452,26 @@ static int do_switch(struct i915_hw_context *to)
 		 * able to defer doing this until we know the object would be
 		 * swapped, but there is no way to do that yet.
 		 */
-		from_obj->dirty = 1;
-		BUG_ON(from_obj->ring != ring);
-		i915_gem_object_unpin(from_obj);
+		from->obj->dirty = 1;
+		BUG_ON(from->obj->ring != ring);
+
+		ret = i915_add_request(ring, NULL);
+		if (ret) {
+			/* Too late, we've already scheduled a context switch.
+			 * Try to undo the change so that the hw state is
+			 * consistent with out tracking. In case of emergency,
+			 * scream.
+			 */
+			WARN_ON(mi_set_context(ring, from, MI_RESTORE_INHIBIT));
+			return ret;
+		}
 
-		drm_gem_object_unreference(&from_obj->base);
+		i915_gem_object_unpin(from->obj);
+		i915_gem_context_unreference(from);
 	}
 
-	drm_gem_object_reference(&to->obj->base);
-	ring->last_context_obj = to->obj;
+	i915_gem_context_reference(to);
+	ring->last_context = to;
 	to->is_initialized = true;
 
 	return 0;
@@ -444,6 +500,8 @@ int i915_switch_context(struct intel_ring_buffer *ring,
 	if (dev_priv->hw_contexts_disabled)
 		return 0;
 
+	WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
+
 	if (ring != &dev_priv->ring[RCS])
 		return 0;
 
@@ -512,8 +570,8 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 		return -ENOENT;
 	}
 
-	do_destroy(ctx);
-
+	idr_remove(&ctx->file_priv->context_idr, ctx->id);
+	i915_gem_context_unreference(ctx);
 	mutex_unlock(&dev->struct_mutex);
 
 	DRM_DEBUG_DRIVER("HW context %d destroyed\n", args->ctx_id);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 117ce38..87a3227 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -786,7 +786,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *objects,
 			obj->dirty = 1;
 			obj->last_write_seqno = intel_ring_get_seqno(ring);
 			if (obj->pin_count) /* check for potential scanout */
-				intel_mark_fb_busy(obj);
+				intel_mark_fb_busy(obj, ring);
 		}
 
 		trace_i915_gem_object_change_domain(obj, old_read, old_write);
@@ -796,13 +796,14 @@ i915_gem_execbuffer_move_to_active(struct list_head *objects,
 static void
 i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 				    struct drm_file *file,
-				    struct intel_ring_buffer *ring)
+				    struct intel_ring_buffer *ring,
+				    struct drm_i915_gem_object *obj)
 {
 	/* Unconditionally force add_request to emit a full flush. */
 	ring->gpu_caches_dirty = true;
 
 	/* Add a breadcrumb for the completion of the batch buffer */
-	(void)i915_add_request(ring, file, NULL);
+	(void)__i915_add_request(ring, file, obj, NULL);
 }
 
 static int
@@ -885,6 +886,15 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 			return -EPERM;
 		}
 		break;
+	case I915_EXEC_VEBOX:
+		ring = &dev_priv->ring[VECS];
+		if (ctx_id != 0) {
+			DRM_DEBUG("Ring %s doesn't support contexts\n",
+				  ring->name);
+			return -EPERM;
+		}
+		break;
+
 	default:
 		DRM_DEBUG("execbuf with unknown ring: %d\n",
 			  (int)(args->flags & I915_EXEC_RING_MASK));
@@ -1074,7 +1084,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	trace_i915_gem_ring_dispatch(ring, intel_ring_get_seqno(ring), flags);
 
 	i915_gem_execbuffer_move_to_active(&eb->objects, ring);
-	i915_gem_execbuffer_retire_commands(dev, file, ring);
+	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
 
 err:
 	eb_destroy(eb);
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index bdb0d77..5101ab6 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -28,8 +28,6 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-typedef uint32_t gen6_gtt_pte_t;
-
 /* PPGTT stuff */
 #define GEN6_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0xff0))
 
@@ -44,29 +42,22 @@ typedef uint32_t gen6_gtt_pte_t;
 #define GEN6_PTE_CACHE_LLC_MLC		(3 << 1)
 #define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
 
-static inline gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
-					     dma_addr_t addr,
-					     enum i915_cache_level level)
+static gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
+				      dma_addr_t addr,
+				      enum i915_cache_level level)
 {
 	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
 
 	switch (level) {
 	case I915_CACHE_LLC_MLC:
-		/* Haswell doesn't set L3 this way */
-		if (IS_HASWELL(dev))
-			pte |= GEN6_PTE_CACHE_LLC;
-		else
-			pte |= GEN6_PTE_CACHE_LLC_MLC;
+		pte |= GEN6_PTE_CACHE_LLC_MLC;
 		break;
 	case I915_CACHE_LLC:
 		pte |= GEN6_PTE_CACHE_LLC;
 		break;
 	case I915_CACHE_NONE:
-		if (IS_HASWELL(dev))
-			pte |= HSW_PTE_UNCACHED;
-		else
-			pte |= GEN6_PTE_UNCACHED;
+		pte |= GEN6_PTE_UNCACHED;
 		break;
 	default:
 		BUG();
@@ -75,16 +66,48 @@ static inline gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
 	return pte;
 }
 
-static int gen6_ppgtt_enable(struct drm_device *dev)
+#define BYT_PTE_WRITEABLE		(1 << 1)
+#define BYT_PTE_SNOOPED_BY_CPU_CACHES	(1 << 2)
+
+static gen6_gtt_pte_t byt_pte_encode(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint32_t pd_offset;
-	struct intel_ring_buffer *ring;
-	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
+	pte |= GEN6_PTE_ADDR_ENCODE(addr);
+
+	/* Mark the page as writeable.  Other platforms don't have a
+	 * setting for read-only/writable, so this matches that behavior.
+	 */
+	pte |= BYT_PTE_WRITEABLE;
+
+	if (level != I915_CACHE_NONE)
+		pte |= BYT_PTE_SNOOPED_BY_CPU_CACHES;
+
+	return pte;
+}
+
+static gen6_gtt_pte_t hsw_pte_encode(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level)
+{
+	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
+	pte |= GEN6_PTE_ADDR_ENCODE(addr);
+
+	if (level != I915_CACHE_NONE)
+		pte |= GEN6_PTE_CACHE_LLC;
+
+	return pte;
+}
+
+static void gen6_write_pdes(struct i915_hw_ppgtt *ppgtt)
+{
+	struct drm_i915_private *dev_priv = ppgtt->dev->dev_private;
 	gen6_gtt_pte_t __iomem *pd_addr;
 	uint32_t pd_entry;
 	int i;
 
+	WARN_ON(ppgtt->pd_offset & 0x3f);
 	pd_addr = (gen6_gtt_pte_t __iomem*)dev_priv->gtt.gsm +
 		ppgtt->pd_offset / sizeof(gen6_gtt_pte_t);
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
@@ -97,6 +120,19 @@ static int gen6_ppgtt_enable(struct drm_device *dev)
 		writel(pd_entry, pd_addr + i);
 	}
 	readl(pd_addr);
+}
+
+static int gen6_ppgtt_enable(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	uint32_t pd_offset;
+	struct intel_ring_buffer *ring;
+	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	int i;
+
+	BUG_ON(ppgtt->pd_offset & 0x3f);
+
+	gen6_write_pdes(ppgtt);
 
 	pd_offset = ppgtt->pd_offset;
 	pd_offset /= 64; /* in cachelines, */
@@ -154,9 +190,9 @@ static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
 	unsigned last_pte, i;
 
-	scratch_pte = gen6_pte_encode(ppgtt->dev,
-				      ppgtt->scratch_page_dma_addr,
-				      I915_CACHE_LLC);
+	scratch_pte = ppgtt->pte_encode(ppgtt->dev,
+					ppgtt->scratch_page_dma_addr,
+					I915_CACHE_LLC);
 
 	while (num_entries) {
 		last_pte = first_pte + num_entries;
@@ -191,8 +227,8 @@ static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
 		dma_addr_t page_addr;
 
 		page_addr = sg_page_iter_dma_address(&sg_iter);
-		pt_vaddr[act_pte] = gen6_pte_encode(ppgtt->dev, page_addr,
-						    cache_level);
+		pt_vaddr[act_pte] = ppgtt->pte_encode(ppgtt->dev, page_addr,
+						      cache_level);
 		if (++act_pte == I915_PPGTT_PT_ENTRIES) {
 			kunmap_atomic(pt_vaddr);
 			act_pt++;
@@ -233,8 +269,15 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	/* ppgtt PDEs reside in the global gtt pagetable, which has 512*1024
 	 * entries. For aliasing ppgtt support we just steal them at the end for
 	 * now. */
-       first_pd_entry_in_global_pt = gtt_total_entries(dev_priv->gtt);
+	first_pd_entry_in_global_pt = gtt_total_entries(dev_priv->gtt);
 
+	if (IS_HASWELL(dev)) {
+		ppgtt->pte_encode = hsw_pte_encode;
+	} else if (IS_VALLEYVIEW(dev)) {
+		ppgtt->pte_encode = byt_pte_encode;
+	} else {
+		ppgtt->pte_encode = gen6_pte_encode;
+	}
 	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
 	ppgtt->enable = gen6_ppgtt_enable;
 	ppgtt->clear_range = gen6_ppgtt_clear_range;
@@ -396,7 +439,7 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 	dev_priv->gtt.gtt_clear_range(dev, dev_priv->gtt.start / PAGE_SIZE,
 				      dev_priv->gtt.total / PAGE_SIZE);
 
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		i915_gem_clflush_object(obj);
 		i915_gem_gtt_bind_object(obj, obj->cache_level);
 	}
@@ -437,7 +480,8 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 
 	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) {
 		addr = sg_page_iter_dma_address(&sg_iter);
-		iowrite32(gen6_pte_encode(dev, addr, level), &gtt_entries[i]);
+		iowrite32(dev_priv->gtt.pte_encode(dev, addr, level),
+			  &gtt_entries[i]);
 		i++;
 	}
 
@@ -449,7 +493,7 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 	 */
 	if (i != 0)
 		WARN_ON(readl(&gtt_entries[i-1])
-			!= gen6_pte_encode(dev, addr, level));
+			!= dev_priv->gtt.pte_encode(dev, addr, level));
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
@@ -474,8 +518,9 @@ static void gen6_ggtt_clear_range(struct drm_device *dev,
 		 first_entry, num_entries, max_entries))
 		num_entries = max_entries;
 
-	scratch_pte = gen6_pte_encode(dev, dev_priv->gtt.scratch_page_dma,
-				      I915_CACHE_LLC);
+	scratch_pte = dev_priv->gtt.pte_encode(dev,
+					       dev_priv->gtt.scratch_page_dma,
+					       I915_CACHE_LLC);
 	for (i = 0; i < num_entries; i++)
 		iowrite32(scratch_pte, &gtt_base[i]);
 	readl(gtt_base);
@@ -586,7 +631,7 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 		dev_priv->mm.gtt_space.color_adjust = i915_gtt_color_adjust;
 
 	/* Mark any preallocated objects as occupied */
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		DRM_DEBUG_KMS("reserving preallocated space: %x + %zx\n",
 			      obj->gtt_offset, obj->base.size);
 
@@ -809,6 +854,13 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	} else {
 		dev_priv->gtt.gtt_probe = gen6_gmch_probe;
 		dev_priv->gtt.gtt_remove = gen6_gmch_remove;
+		if (IS_HASWELL(dev)) {
+			dev_priv->gtt.pte_encode = hsw_pte_encode;
+		} else if (IS_VALLEYVIEW(dev)) {
+			dev_priv->gtt.pte_encode = byt_pte_encode;
+		} else {
+			dev_priv->gtt.pte_encode = gen6_pte_encode;
+		}
 	}
 
 	ret = dev_priv->gtt.gtt_probe(dev, &dev_priv->gtt.total,
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 130d1db..982d473 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -62,7 +62,10 @@ static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 	 * its value of TOLUD.
 	 */
 	base = 0;
-	if (INTEL_INFO(dev)->gen >= 6) {
+	if (IS_VALLEYVIEW(dev)) {
+		pci_read_config_dword(dev->pdev, 0x5c, &base);
+		base &= ~((1<<20) - 1);
+	} else if (INTEL_INFO(dev)->gen >= 6) {
 		/* Read Base Data of Stolen Memory Register (BDSM) directly.
 		 * Note that there is also a MCHBAR miror at 0x1080c0 or
 		 * we could use device 2:0x5c instead.
@@ -136,6 +139,7 @@ static int i915_setup_compression(struct drm_device *dev, int size)
 err_fb:
 	drm_mm_put_block(compressed_fb);
 err:
+	pr_info_once("drm: not enough stolen space for compressed buffer (need %d more bytes), disabling. Hint: you may be able to increase stolen memory size in the BIOS to avoid this.\n", size);
 	return -ENOSPC;
 }
 
@@ -143,7 +147,7 @@ int i915_gem_stolen_setup_compression(struct drm_device *dev, int size)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (dev_priv->mm.stolen_base == 0)
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return -ENODEV;
 
 	if (size < dev_priv->cfb_size)
@@ -175,6 +179,9 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
+		return;
+
 	i915_gem_stolen_cleanup_compression(dev);
 	drm_mm_takedown(&dev_priv->mm.stolen);
 }
@@ -182,6 +189,7 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 int i915_gem_init_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	int bios_reserved = 0;
 
 	dev_priv->mm.stolen_base = i915_stolen_to_physical(dev);
 	if (dev_priv->mm.stolen_base == 0)
@@ -190,8 +198,12 @@ int i915_gem_init_stolen(struct drm_device *dev)
 	DRM_DEBUG_KMS("found %zd bytes of stolen memory at %08lx\n",
 		      dev_priv->gtt.stolen_size, dev_priv->mm.stolen_base);
 
+	if (IS_VALLEYVIEW(dev))
+		bios_reserved = 1024*1024; /* top 1M on VLV/BYT */
+
 	/* Basic memrange allocator for stolen space */
-	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->gtt.stolen_size);
+	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->gtt.stolen_size -
+		    bios_reserved);
 
 	return 0;
 }
@@ -270,7 +282,7 @@ _i915_gem_object_create_stolen(struct drm_device *dev,
 		goto cleanup;
 
 	obj->has_dma_mapping = true;
-	obj->pages_pin_count = 1;
+	i915_gem_object_pin_pages(obj);
 	obj->stolen = stolen;
 
 	obj->base.write_domain = I915_GEM_DOMAIN_GTT;
@@ -291,7 +303,7 @@ i915_gem_object_create_stolen(struct drm_device *dev, u32 size)
 	struct drm_i915_gem_object *obj;
 	struct drm_mm_node *stolen;
 
-	if (dev_priv->mm.stolen_base == 0)
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return NULL;
 
 	DRM_DEBUG_KMS("creating stolen object: size=%x\n", size);
@@ -322,7 +334,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 	struct drm_i915_gem_object *obj;
 	struct drm_mm_node *stolen;
 
-	if (dev_priv->mm.stolen_base == 0)
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return NULL;
 
 	DRM_DEBUG_KMS("creating preallocated stolen object: stolen_offset=%x, gtt_offset=%x, size=%x\n",
@@ -330,7 +342,6 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 
 	/* KISS and expect everything to be page-aligned */
 	BUG_ON(stolen_offset & 4095);
-	BUG_ON(gtt_offset & 4095);
 	BUG_ON(size & 4095);
 
 	if (WARN_ON(size == 0))
@@ -351,6 +362,10 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 		return NULL;
 	}
 
+	/* Some objects just need physical mem from stolen space */
+	if (gtt_offset == -1)
+		return obj;
+
 	/* To simplify the initialisation sequence between KMS and GTT,
 	 * we allow construction of the stolen object prior to
 	 * setting up the GTT space. The actual reservation will occur
@@ -371,7 +386,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 	obj->gtt_offset = gtt_offset;
 	obj->has_global_gtt_mapping = 1;
 
-	list_add_tail(&obj->gtt_list, &dev_priv->mm.bound_list);
+	list_add_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&obj->mm_list, &dev_priv->mm.inactive_list);
 
 	return obj;
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 0aa2ef0..3d92a7c 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -70,15 +70,6 @@ static const u32 hpd_status_gen4[] = {
 	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
 };
 
-static const u32 hpd_status_i965[] = {
-	 [HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
-	 [HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I965,
-	 [HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_I965,
-	 [HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
-	 [HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
-	 [HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
-};
-
 static const u32 hpd_status_i915[] = { /* i915 and valleyview are the same */
 	[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
 	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I915,
@@ -88,13 +79,12 @@ static const u32 hpd_status_i915[] = { /* i915 and valleyview are the same */
 	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
 };
 
-static void ibx_hpd_irq_setup(struct drm_device *dev);
-static void i915_hpd_irq_setup(struct drm_device *dev);
-
 /* For display hotplug interrupt */
 static void
 ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 {
+	assert_spin_locked(&dev_priv->irq_lock);
+
 	if ((dev_priv->irq_mask & mask) != 0) {
 		dev_priv->irq_mask &= ~mask;
 		I915_WRITE(DEIMR, dev_priv->irq_mask);
@@ -105,6 +95,8 @@ ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 static void
 ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 {
+	assert_spin_locked(&dev_priv->irq_lock);
+
 	if ((dev_priv->irq_mask & mask) != mask) {
 		dev_priv->irq_mask |= mask;
 		I915_WRITE(DEIMR, dev_priv->irq_mask);
@@ -112,6 +104,215 @@ ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 	}
 }
 
+static bool ivb_can_enable_err_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	enum pipe pipe;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	for_each_pipe(pipe) {
+		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+
+		if (crtc->cpu_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static bool cpt_can_enable_serr_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+	struct intel_crtc *crtc;
+
+	for_each_pipe(pipe) {
+		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+
+		if (crtc->pch_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static void ironlake_set_fifo_underrun_reporting(struct drm_device *dev,
+						 enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t bit = (pipe == PIPE_A) ? DE_PIPEA_FIFO_UNDERRUN :
+					  DE_PIPEB_FIFO_UNDERRUN;
+
+	if (enable)
+		ironlake_enable_display_irq(dev_priv, bit);
+	else
+		ironlake_disable_display_irq(dev_priv, bit);
+}
+
+static void ivybridge_set_fifo_underrun_reporting(struct drm_device *dev,
+						  bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (enable) {
+		if (!ivb_can_enable_err_int(dev))
+			return;
+
+		I915_WRITE(GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN_A |
+					 ERR_INT_FIFO_UNDERRUN_B |
+					 ERR_INT_FIFO_UNDERRUN_C);
+
+		ironlake_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	} else {
+		ironlake_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	}
+}
+
+static void ibx_set_fifo_underrun_reporting(struct intel_crtc *crtc,
+					    bool enable)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t bit = (crtc->pipe == PIPE_A) ? SDE_TRANSA_FIFO_UNDER :
+						SDE_TRANSB_FIFO_UNDER;
+
+	if (enable)
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) & ~bit);
+	else
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) | bit);
+
+	POSTING_READ(SDEIMR);
+}
+
+static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum transcoder pch_transcoder,
+					    bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (enable) {
+		if (!cpt_can_enable_serr_int(dev))
+			return;
+
+		I915_WRITE(SERR_INT, SERR_INT_TRANS_A_FIFO_UNDERRUN |
+				     SERR_INT_TRANS_B_FIFO_UNDERRUN |
+				     SERR_INT_TRANS_C_FIFO_UNDERRUN);
+
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) & ~SDE_ERROR_CPT);
+	} else {
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) | SDE_ERROR_CPT);
+	}
+
+	POSTING_READ(SDEIMR);
+}
+
+/**
+ * intel_set_cpu_fifo_underrun_reporting - enable/disable FIFO underrun messages
+ * @dev: drm device
+ * @pipe: pipe
+ * @enable: true if we want to report FIFO underrun errors, false otherwise
+ *
+ * This function makes us disable or enable CPU fifo underruns for a specific
+ * pipe. Notice that on some Gens (e.g. IVB, HSW), disabling FIFO underrun
+ * reporting for one pipe may also disable all the other CPU error interruts for
+ * the other pipes, due to the fact that there's just one interrupt mask/enable
+ * bit for all the pipes.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+					   enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	unsigned long flags;
+	bool ret;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+
+	ret = !intel_crtc->cpu_fifo_underrun_disabled;
+
+	if (enable == ret)
+		goto done;
+
+	intel_crtc->cpu_fifo_underrun_disabled = !enable;
+
+	if (IS_GEN5(dev) || IS_GEN6(dev))
+		ironlake_set_fifo_underrun_reporting(dev, pipe, enable);
+	else if (IS_GEN7(dev))
+		ivybridge_set_fifo_underrun_reporting(dev, enable);
+
+done:
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	return ret;
+}
+
+/**
+ * intel_set_pch_fifo_underrun_reporting - enable/disable FIFO underrun messages
+ * @dev: drm device
+ * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
+ * @enable: true if we want to report FIFO underrun errors, false otherwise
+ *
+ * This function makes us disable or enable PCH fifo underruns for a specific
+ * PCH transcoder. Notice that on some PCHs (e.g. CPT/PPT), disabling FIFO
+ * underrun reporting for one transcoder may also disable all the other PCH
+ * error interruts for the other transcoders, due to the fact that there's just
+ * one interrupt mask/enable bit for all the transcoders.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
+					   enum transcoder pch_transcoder,
+					   bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe p;
+	struct drm_crtc *crtc;
+	struct intel_crtc *intel_crtc;
+	unsigned long flags;
+	bool ret;
+
+	if (HAS_PCH_LPT(dev)) {
+		crtc = NULL;
+		for_each_pipe(p) {
+			struct drm_crtc *c = dev_priv->pipe_to_crtc_mapping[p];
+			if (intel_pipe_has_type(c, INTEL_OUTPUT_ANALOG)) {
+				crtc = c;
+				break;
+			}
+		}
+		if (!crtc) {
+			DRM_ERROR("PCH FIFO underrun, but no CRTC using the PCH found\n");
+			return false;
+		}
+	} else {
+		crtc = dev_priv->pipe_to_crtc_mapping[pch_transcoder];
+	}
+	intel_crtc = to_intel_crtc(crtc);
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+
+	ret = !intel_crtc->pch_fifo_underrun_disabled;
+
+	if (enable == ret)
+		goto done;
+
+	intel_crtc->pch_fifo_underrun_disabled = !enable;
+
+	if (HAS_PCH_IBX(dev))
+		ibx_set_fifo_underrun_reporting(intel_crtc, enable);
+	else
+		cpt_set_fifo_underrun_reporting(dev, pch_transcoder, enable);
+
+done:
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	return ret;
+}
+
+
 void
 i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask)
 {
@@ -142,28 +343,21 @@ i915_disable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask)
 }
 
 /**
- * intel_enable_asle - enable ASLE interrupt for OpRegion
+ * i915_enable_asle_pipestat - enable ASLE pipestat for OpRegion
  */
-void intel_enable_asle(struct drm_device *dev)
+static void i915_enable_asle_pipestat(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	unsigned long irqflags;
 
-	/* FIXME: opregion/asle for VLV */
-	if (IS_VALLEYVIEW(dev))
+	if (!dev_priv->opregion.asle || !IS_MOBILE(dev))
 		return;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 
-	if (HAS_PCH_SPLIT(dev))
-		ironlake_enable_display_irq(dev_priv, DE_GSE);
-	else {
-		i915_enable_pipestat(dev_priv, 1,
-				     PIPE_LEGACY_BLC_EVENT_ENABLE);
-		if (INTEL_INFO(dev)->gen >= 4)
-			i915_enable_pipestat(dev_priv, 0,
-					     PIPE_LEGACY_BLC_EVENT_ENABLE);
-	}
+	i915_enable_pipestat(dev_priv, 1, PIPE_LEGACY_BLC_EVENT_ENABLE);
+	if (INTEL_INFO(dev)->gen >= 4)
+		i915_enable_pipestat(dev_priv, 0, PIPE_LEGACY_BLC_EVENT_ENABLE);
 
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 }
@@ -181,10 +375,16 @@ static int
 i915_pipe_enabled(struct drm_device *dev, int pipe)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
 
-	return I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_ENABLE;
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+		/* Locking is horribly broken here, but whatever. */
+		struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+		return intel_crtc->active;
+	} else {
+		return I915_READ(PIPECONF(pipe)) & PIPECONF_ENABLE;
+	}
 }
 
 /* Called from drm generic code, passed a 'crtc', which
@@ -334,6 +534,21 @@ static int i915_get_vblank_timestamp(struct drm_device *dev, int pipe,
 						     crtc);
 }
 
+static int intel_hpd_irq_event(struct drm_device *dev, struct drm_connector *connector)
+{
+	enum drm_connector_status old_status;
+
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+	old_status = connector->status;
+
+	connector->status = connector->funcs->detect(connector, false);
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] status updated from %d to %d\n",
+		      connector->base.id,
+		      drm_get_connector_name(connector),
+		      old_status, connector->status);
+	return (old_status != connector->status);
+}
+
 /*
  * Handle hotplug events outside the interrupt handler proper.
  */
@@ -350,6 +565,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	struct drm_connector *connector;
 	unsigned long irqflags;
 	bool hpd_disabled = false;
+	bool changed = false;
+	u32 hpd_event_bits;
 
 	/* HPD irq before everything is fully set up. */
 	if (!dev_priv->enable_hotplug_processing)
@@ -359,6 +576,9 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	DRM_DEBUG_KMS("running encoder hotplug functions\n");
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+
+	hpd_event_bits = dev_priv->hpd_event_bits;
+	dev_priv->hpd_event_bits = 0;
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
 		intel_encoder = intel_connector->encoder;
@@ -373,6 +593,10 @@ static void i915_hotplug_work_func(struct work_struct *work)
 				| DRM_CONNECTOR_POLL_DISCONNECT;
 			hpd_disabled = true;
 		}
+		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
+			DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug event.\n",
+				      drm_get_connector_name(connector), intel_encoder->hpd_pin);
+		}
 	}
 	 /* if there were no outputs to poll, poll was disabled,
 	  * therefore make sure it's enabled when disabling HPD on
@@ -385,14 +609,20 @@ static void i915_hotplug_work_func(struct work_struct *work)
 
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 
-	list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
-		if (intel_encoder->hot_plug)
-			intel_encoder->hot_plug(intel_encoder);
-
+	list_for_each_entry(connector, &mode_config->connector_list, head) {
+		intel_connector = to_intel_connector(connector);
+		intel_encoder = intel_connector->encoder;
+		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
+			if (intel_encoder->hot_plug)
+				intel_encoder->hot_plug(intel_encoder);
+			if (intel_hpd_irq_event(dev, connector))
+				changed = true;
+		}
+	}
 	mutex_unlock(&mode_config->mutex);
 
-	/* Just fire off a uevent and let userspace tell us what to do */
-	drm_helper_hpd_irq_event(dev);
+	if (changed)
+		drm_kms_helper_hotplug_event(dev);
 }
 
 static void ironlake_handle_rps_change(struct drm_device *dev)
@@ -447,7 +677,6 @@ static void notify_ring(struct drm_device *dev,
 
 	wake_up_all(&ring->irq_queue);
 	if (i915_enable_hangcheck) {
-		dev_priv->gpu_error.hangcheck_count = 0;
 		mod_timer(&dev_priv->gpu_error.hangcheck_timer,
 			  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
 	}
@@ -464,25 +693,48 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	pm_iir = dev_priv->rps.pm_iir;
 	dev_priv->rps.pm_iir = 0;
 	pm_imr = I915_READ(GEN6_PMIMR);
-	I915_WRITE(GEN6_PMIMR, 0);
+	/* Make sure not to corrupt PMIMR state used by ringbuffer code */
+	I915_WRITE(GEN6_PMIMR, pm_imr & ~GEN6_PM_RPS_EVENTS);
 	spin_unlock_irq(&dev_priv->rps.lock);
 
-	if ((pm_iir & GEN6_PM_DEFERRED_EVENTS) == 0)
+	if ((pm_iir & GEN6_PM_RPS_EVENTS) == 0)
 		return;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	if (pm_iir & GEN6_PM_RP_UP_THRESHOLD)
+	if (pm_iir & GEN6_PM_RP_UP_THRESHOLD) {
 		new_delay = dev_priv->rps.cur_delay + 1;
-	else
+
+		/*
+		 * For better performance, jump directly
+		 * to RPe if we're below it.
+		 */
+		if (IS_VALLEYVIEW(dev_priv->dev) &&
+		    dev_priv->rps.cur_delay < dev_priv->rps.rpe_delay)
+			new_delay = dev_priv->rps.rpe_delay;
+	} else
 		new_delay = dev_priv->rps.cur_delay - 1;
 
 	/* sysfs frequency interfaces may have snuck in while servicing the
 	 * interrupt
 	 */
-	if (!(new_delay > dev_priv->rps.max_delay ||
-	      new_delay < dev_priv->rps.min_delay)) {
-		gen6_set_rps(dev_priv->dev, new_delay);
+	if (new_delay >= dev_priv->rps.min_delay &&
+	    new_delay <= dev_priv->rps.max_delay) {
+		if (IS_VALLEYVIEW(dev_priv->dev))
+			valleyview_set_rps(dev_priv->dev, new_delay);
+		else
+			gen6_set_rps(dev_priv->dev, new_delay);
+	}
+
+	if (IS_VALLEYVIEW(dev_priv->dev)) {
+		/*
+		 * On VLV, when we enter RC6 we may not be at the minimum
+		 * voltage level, so arm a timer to check.  It should only
+		 * fire when there's activity or once after we've entered
+		 * RC6, and then won't be re-armed until the next RPS interrupt.
+		 */
+		mod_delayed_work(dev_priv->wq, &dev_priv->rps.vlv_work,
+				 msecs_to_jiffies(100));
 	}
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
@@ -529,7 +781,7 @@ static void ivybridge_parity_work(struct work_struct *work)
 	I915_WRITE(GEN7_MISCCPCTL, misccpctl);
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	dev_priv->gt_irq_mask &= ~GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
+	dev_priv->gt_irq_mask &= ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -561,7 +813,7 @@ static void ivybridge_handle_parity_error(struct drm_device *dev)
 		return;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	dev_priv->gt_irq_mask |= GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
+	dev_priv->gt_irq_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -573,25 +825,26 @@ static void snb_gt_irq_handler(struct drm_device *dev,
 			       u32 gt_iir)
 {
 
-	if (gt_iir & (GEN6_RENDER_USER_INTERRUPT |
-		      GEN6_RENDER_PIPE_CONTROL_NOTIFY_INTERRUPT))
+	if (gt_iir &
+	    (GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
 		notify_ring(dev, &dev_priv->ring[RCS]);
-	if (gt_iir & GEN6_BSD_USER_INTERRUPT)
+	if (gt_iir & GT_BSD_USER_INTERRUPT)
 		notify_ring(dev, &dev_priv->ring[VCS]);
-	if (gt_iir & GEN6_BLITTER_USER_INTERRUPT)
+	if (gt_iir & GT_BLT_USER_INTERRUPT)
 		notify_ring(dev, &dev_priv->ring[BCS]);
 
-	if (gt_iir & (GT_GEN6_BLT_CS_ERROR_INTERRUPT |
-		      GT_GEN6_BSD_CS_ERROR_INTERRUPT |
-		      GT_RENDER_CS_ERROR_INTERRUPT)) {
+	if (gt_iir & (GT_BLT_CS_ERROR_INTERRUPT |
+		      GT_BSD_CS_ERROR_INTERRUPT |
+		      GT_RENDER_CS_MASTER_ERROR_INTERRUPT)) {
 		DRM_ERROR("GT error interrupt 0x%08x\n", gt_iir);
 		i915_handle_error(dev, false);
 	}
 
-	if (gt_iir & GT_GEN7_L3_PARITY_ERROR_INTERRUPT)
+	if (gt_iir & GT_RENDER_L3_PARITY_ERROR_INTERRUPT)
 		ivybridge_handle_parity_error(dev);
 }
 
+/* Legacy way of handling PM interrupts */
 static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 				u32 pm_iir)
 {
@@ -619,23 +872,25 @@ static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 #define HPD_STORM_DETECT_PERIOD 1000
 #define HPD_STORM_THRESHOLD 5
 
-static inline bool hotplug_irq_storm_detect(struct drm_device *dev,
-					    u32 hotplug_trigger,
-					    const u32 *hpd)
+static inline void intel_hpd_irq_handler(struct drm_device *dev,
+					 u32 hotplug_trigger,
+					 const u32 *hpd)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	unsigned long irqflags;
 	int i;
-	bool ret = false;
+	bool storm_detected = false;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	if (!hotplug_trigger)
+		return;
 
+	spin_lock(&dev_priv->irq_lock);
 	for (i = 1; i < HPD_NUM_PINS; i++) {
 
 		if (!(hpd[i] & hotplug_trigger) ||
 		    dev_priv->hpd_stats[i].hpd_mark != HPD_ENABLED)
 			continue;
 
+		dev_priv->hpd_event_bits |= (1 << i);
 		if (!time_in_range(jiffies, dev_priv->hpd_stats[i].hpd_last_jiffies,
 				   dev_priv->hpd_stats[i].hpd_last_jiffies
 				   + msecs_to_jiffies(HPD_STORM_DETECT_PERIOD))) {
@@ -643,16 +898,20 @@ static inline bool hotplug_irq_storm_detect(struct drm_device *dev,
 			dev_priv->hpd_stats[i].hpd_cnt = 0;
 		} else if (dev_priv->hpd_stats[i].hpd_cnt > HPD_STORM_THRESHOLD) {
 			dev_priv->hpd_stats[i].hpd_mark = HPD_MARK_DISABLED;
+			dev_priv->hpd_event_bits &= ~(1 << i);
 			DRM_DEBUG_KMS("HPD interrupt storm detected on PIN %d\n", i);
-			ret = true;
+			storm_detected = true;
 		} else {
 			dev_priv->hpd_stats[i].hpd_cnt++;
 		}
 	}
 
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	if (storm_detected)
+		dev_priv->display.hpd_irq_setup(dev);
+	spin_unlock(&dev_priv->irq_lock);
 
-	return ret;
+	queue_work(dev_priv->wq,
+		   &dev_priv->hotplug_work);
 }
 
 static void gmbus_irq_handler(struct drm_device *dev)
@@ -669,6 +928,38 @@ static void dp_aux_irq_handler(struct drm_device *dev)
 	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
+/* Unlike gen6_queue_rps_work() from which this function is originally derived,
+ * we must be able to deal with other PM interrupts. This is complicated because
+ * of the way in which we use the masks to defer the RPS work (which for
+ * posterity is necessary because of forcewake).
+ */
+static void hsw_pm_irq_handler(struct drm_i915_private *dev_priv,
+			       u32 pm_iir)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->rps.lock, flags);
+	dev_priv->rps.pm_iir |= pm_iir & GEN6_PM_RPS_EVENTS;
+	if (dev_priv->rps.pm_iir) {
+		I915_WRITE(GEN6_PMIMR, dev_priv->rps.pm_iir);
+		/* never want to mask useful interrupts. (also posting read) */
+		WARN_ON(I915_READ_NOTRACE(GEN6_PMIMR) & ~GEN6_PM_RPS_EVENTS);
+		/* TODO: if queue_work is slow, move it out of the spinlock */
+		queue_work(dev_priv->wq, &dev_priv->rps.work);
+	}
+	spin_unlock_irqrestore(&dev_priv->rps.lock, flags);
+
+	if (pm_iir & ~GEN6_PM_RPS_EVENTS) {
+		if (pm_iir & PM_VEBOX_USER_INTERRUPT)
+			notify_ring(dev_priv->dev, &dev_priv->ring[VECS]);
+
+		if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT) {
+			DRM_ERROR("VEBOX CS error interrupt 0x%08x\n", pm_iir);
+			i915_handle_error(dev_priv->dev, false);
+		}
+	}
+}
+
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = (struct drm_device *) arg;
@@ -727,12 +1018,9 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 					 hotplug_status);
-			if (hotplug_trigger) {
-				if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_status_i915))
-					i915_hpd_irq_setup(dev);
-				queue_work(dev_priv->wq,
-					   &dev_priv->hotplug_work);
-			}
+
+			intel_hpd_irq_handler(dev, hotplug_trigger, hpd_status_i915);
+
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
@@ -740,7 +1028,7 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 		if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
 			gmbus_irq_handler(dev);
 
-		if (pm_iir & GEN6_PM_DEFERRED_EVENTS)
+		if (pm_iir & GEN6_PM_RPS_EVENTS)
 			gen6_queue_rps_work(dev_priv, pm_iir);
 
 		I915_WRITE(GTIIR, gt_iir);
@@ -758,15 +1046,14 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 	int pipe;
 	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK;
 
-	if (hotplug_trigger) {
-		if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_ibx))
-			ibx_hpd_irq_setup(dev);
-		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
-	}
-	if (pch_iir & SDE_AUDIO_POWER_MASK)
+	intel_hpd_irq_handler(dev, hotplug_trigger, hpd_ibx);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK) {
+		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK) >>
+			       SDE_AUDIO_POWER_SHIFT);
 		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
-				 (pch_iir & SDE_AUDIO_POWER_MASK) >>
-				 SDE_AUDIO_POWER_SHIFT);
+				 port_name(port));
+	}
 
 	if (pch_iir & SDE_AUX_MASK)
 		dp_aux_irq_handler(dev);
@@ -795,10 +1082,64 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 	if (pch_iir & (SDE_TRANSB_CRC_ERR | SDE_TRANSA_CRC_ERR))
 		DRM_DEBUG_DRIVER("PCH transcoder CRC error interrupt\n");
 
-	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
-		DRM_DEBUG_DRIVER("PCH transcoder B underrun interrupt\n");
 	if (pch_iir & SDE_TRANSA_FIFO_UNDER)
-		DRM_DEBUG_DRIVER("PCH transcoder A underrun interrupt\n");
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder A FIFO underrun\n");
+
+	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_B,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder B FIFO underrun\n");
+}
+
+static void ivb_err_int_handler(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 err_int = I915_READ(GEN7_ERR_INT);
+
+	if (err_int & ERR_INT_POISON)
+		DRM_ERROR("Poison interrupt\n");
+
+	if (err_int & ERR_INT_FIFO_UNDERRUN_A)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_A, false))
+			DRM_DEBUG_DRIVER("Pipe A FIFO underrun\n");
+
+	if (err_int & ERR_INT_FIFO_UNDERRUN_B)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_B, false))
+			DRM_DEBUG_DRIVER("Pipe B FIFO underrun\n");
+
+	if (err_int & ERR_INT_FIFO_UNDERRUN_C)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_C, false))
+			DRM_DEBUG_DRIVER("Pipe C FIFO underrun\n");
+
+	I915_WRITE(GEN7_ERR_INT, err_int);
+}
+
+static void cpt_serr_int_handler(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 serr_int = I915_READ(SERR_INT);
+
+	if (serr_int & SERR_INT_POISON)
+		DRM_ERROR("PCH poison interrupt\n");
+
+	if (serr_int & SERR_INT_TRANS_A_FIFO_UNDERRUN)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder A FIFO underrun\n");
+
+	if (serr_int & SERR_INT_TRANS_B_FIFO_UNDERRUN)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_B,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder B FIFO underrun\n");
+
+	if (serr_int & SERR_INT_TRANS_C_FIFO_UNDERRUN)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_C,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder C FIFO underrun\n");
+
+	I915_WRITE(SERR_INT, serr_int);
 }
 
 static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
@@ -807,15 +1148,14 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 	int pipe;
 	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT;
 
-	if (hotplug_trigger) {
-		if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_cpt))
-			ibx_hpd_irq_setup(dev);
-		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
+	intel_hpd_irq_handler(dev, hotplug_trigger, hpd_cpt);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT) {
+		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
+			       SDE_AUDIO_POWER_SHIFT_CPT);
+		DRM_DEBUG_DRIVER("PCH audio power change on port %c\n",
+				 port_name(port));
 	}
-	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT)
-		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
-				 (pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
-				 SDE_AUDIO_POWER_SHIFT_CPT);
 
 	if (pch_iir & SDE_AUX_MASK_CPT)
 		dp_aux_irq_handler(dev);
@@ -834,6 +1174,9 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 			DRM_DEBUG_DRIVER("  pipe %c FDI IIR: 0x%08x\n",
 					 pipe_name(pipe),
 					 I915_READ(FDI_RX_IIR(pipe)));
+
+	if (pch_iir & SDE_ERROR_CPT)
+		cpt_serr_int_handler(dev);
 }
 
 static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
@@ -846,6 +1189,14 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	atomic_inc(&dev_priv->irq_received);
 
+	/* We get interrupts on unclaimed registers, so check for this before we
+	 * do any I915_{READ,WRITE}. */
+	if (IS_HASWELL(dev) &&
+	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
+		DRM_ERROR("Unclaimed register before interrupt\n");
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+	}
+
 	/* disable master interrupt before clearing iir  */
 	de_ier = I915_READ(DEIER);
 	I915_WRITE(DEIER, de_ier & ~DE_MASTER_IRQ_CONTROL);
@@ -861,6 +1212,15 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 		POSTING_READ(SDEIER);
 	}
 
+	/* On Haswell, also mask ERR_INT because we don't want to risk
+	 * generating "unclaimed register" interrupts from inside the interrupt
+	 * handler. */
+	if (IS_HASWELL(dev)) {
+		spin_lock(&dev_priv->irq_lock);
+		ironlake_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+		spin_unlock(&dev_priv->irq_lock);
+	}
+
 	gt_iir = I915_READ(GTIIR);
 	if (gt_iir) {
 		snb_gt_irq_handler(dev, dev_priv, gt_iir);
@@ -870,11 +1230,14 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	de_iir = I915_READ(DEIIR);
 	if (de_iir) {
+		if (de_iir & DE_ERR_INT_IVB)
+			ivb_err_int_handler(dev);
+
 		if (de_iir & DE_AUX_CHANNEL_A_IVB)
 			dp_aux_irq_handler(dev);
 
 		if (de_iir & DE_GSE_IVB)
-			intel_opregion_gse_intr(dev);
+			intel_opregion_asle_intr(dev);
 
 		for (i = 0; i < 3; i++) {
 			if (de_iir & (DE_PIPEA_VBLANK_IVB << (5 * i)))
@@ -901,12 +1264,21 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	pm_iir = I915_READ(GEN6_PMIIR);
 	if (pm_iir) {
-		if (pm_iir & GEN6_PM_DEFERRED_EVENTS)
+		if (IS_HASWELL(dev))
+			hsw_pm_irq_handler(dev_priv, pm_iir);
+		else if (pm_iir & GEN6_PM_RPS_EVENTS)
 			gen6_queue_rps_work(dev_priv, pm_iir);
 		I915_WRITE(GEN6_PMIIR, pm_iir);
 		ret = IRQ_HANDLED;
 	}
 
+	if (IS_HASWELL(dev)) {
+		spin_lock(&dev_priv->irq_lock);
+		if (ivb_can_enable_err_int(dev))
+			ironlake_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
+		spin_unlock(&dev_priv->irq_lock);
+	}
+
 	I915_WRITE(DEIER, de_ier);
 	POSTING_READ(DEIER);
 	if (!HAS_PCH_NOP(dev)) {
@@ -921,9 +1293,10 @@ static void ilk_gt_irq_handler(struct drm_device *dev,
 			       struct drm_i915_private *dev_priv,
 			       u32 gt_iir)
 {
-	if (gt_iir & (GT_USER_INTERRUPT | GT_PIPE_NOTIFY))
+	if (gt_iir &
+	    (GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
 		notify_ring(dev, &dev_priv->ring[RCS]);
-	if (gt_iir & GT_BSD_USER_INTERRUPT)
+	if (gt_iir & ILK_BSD_USER_INTERRUPT)
 		notify_ring(dev, &dev_priv->ring[VCS]);
 }
 
@@ -968,7 +1341,7 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 		dp_aux_irq_handler(dev);
 
 	if (de_iir & DE_GSE)
-		intel_opregion_gse_intr(dev);
+		intel_opregion_asle_intr(dev);
 
 	if (de_iir & DE_PIPEA_VBLANK)
 		drm_handle_vblank(dev, 0);
@@ -976,6 +1349,17 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	if (de_iir & DE_PIPEB_VBLANK)
 		drm_handle_vblank(dev, 1);
 
+	if (de_iir & DE_POISON)
+		DRM_ERROR("Poison interrupt\n");
+
+	if (de_iir & DE_PIPEA_FIFO_UNDERRUN)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_A, false))
+			DRM_DEBUG_DRIVER("Pipe A FIFO underrun\n");
+
+	if (de_iir & DE_PIPEB_FIFO_UNDERRUN)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_B, false))
+			DRM_DEBUG_DRIVER("Pipe B FIFO underrun\n");
+
 	if (de_iir & DE_PLANEA_FLIP_DONE) {
 		intel_prepare_page_flip(dev, 0);
 		intel_finish_page_flip_plane(dev, 0);
@@ -1002,7 +1386,7 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	if (IS_GEN5(dev) &&  de_iir & DE_PCU_EVENT)
 		ironlake_handle_rps_change(dev);
 
-	if (IS_GEN6(dev) && pm_iir & GEN6_PM_DEFERRED_EVENTS)
+	if (IS_GEN6(dev) && pm_iir & GEN6_PM_RPS_EVENTS)
 		gen6_queue_rps_work(dev_priv, pm_iir);
 
 	I915_WRITE(GTIIR, gt_iir);
@@ -1222,11 +1606,13 @@ i915_error_state_free(struct kref *error_ref)
 	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
 		i915_error_object_free(error->ring[i].batchbuffer);
 		i915_error_object_free(error->ring[i].ringbuffer);
+		i915_error_object_free(error->ring[i].ctx);
 		kfree(error->ring[i].requests);
 	}
 
 	kfree(error->active_bo);
 	kfree(error->overlay);
+	kfree(error->display);
 	kfree(error);
 }
 static void capture_bo(struct drm_i915_error_buffer *err,
@@ -1273,7 +1659,7 @@ static u32 capture_pinned_bo(struct drm_i915_error_buffer *err,
 	struct drm_i915_gem_object *obj;
 	int i = 0;
 
-	list_for_each_entry(obj, head, gtt_list) {
+	list_for_each_entry(obj, head, global_list) {
 		if (obj->pin_count == 0)
 			continue;
 
@@ -1415,7 +1801,7 @@ static void i915_gem_record_active_context(struct intel_ring_buffer *ring,
 	if (ring->id != RCS || !error->ccid)
 		return;
 
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		if ((error->ccid & PAGE_MASK) == obj->gtt_offset) {
 			ering->ctx = i915_error_object_create_sized(dev_priv,
 								    obj, 1);
@@ -1552,7 +1938,7 @@ static void i915_capture_error_state(struct drm_device *dev)
 	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list)
 		i++;
 	error->active_bo_count = i;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list)
 		if (obj->pin_count)
 			i++;
 	error->pinned_bo_count = i - error->active_bo_count;
@@ -1932,38 +2318,28 @@ ring_last_seqno(struct intel_ring_buffer *ring)
 			  struct drm_i915_gem_request, list)->seqno;
 }
 
-static bool i915_hangcheck_ring_idle(struct intel_ring_buffer *ring, bool *err)
+static bool
+ring_idle(struct intel_ring_buffer *ring, u32 seqno)
 {
-	if (list_empty(&ring->request_list) ||
-	    i915_seqno_passed(ring->get_seqno(ring, false),
-			      ring_last_seqno(ring))) {
-		/* Issue a wake-up to catch stuck h/w. */
-		if (waitqueue_active(&ring->irq_queue)) {
-			DRM_ERROR("Hangcheck timer elapsed... %s idle\n",
-				  ring->name);
-			wake_up_all(&ring->irq_queue);
-			*err = true;
-		}
-		return true;
-	}
-	return false;
+	return (list_empty(&ring->request_list) ||
+		i915_seqno_passed(seqno, ring_last_seqno(ring)));
 }
 
-static bool semaphore_passed(struct intel_ring_buffer *ring)
+static struct intel_ring_buffer *
+semaphore_waits_for(struct intel_ring_buffer *ring, u32 *seqno)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-	u32 acthd = intel_ring_get_active_head(ring) & HEAD_ADDR;
-	struct intel_ring_buffer *signaller;
-	u32 cmd, ipehr, acthd_min;
+	u32 cmd, ipehr, acthd, acthd_min;
 
 	ipehr = I915_READ(RING_IPEHR(ring->mmio_base));
 	if ((ipehr & ~(0x3 << 16)) !=
 	    (MI_SEMAPHORE_MBOX | MI_SEMAPHORE_COMPARE | MI_SEMAPHORE_REGISTER))
-		return false;
+		return NULL;
 
 	/* ACTHD is likely pointing to the dword after the actual command,
 	 * so scan backwards until we find the MBOX.
 	 */
+	acthd = intel_ring_get_active_head(ring) & HEAD_ADDR;
 	acthd_min = max((int)acthd - 3 * 4, 0);
 	do {
 		cmd = ioread32(ring->virtual_start + acthd);
@@ -1972,124 +2348,216 @@ static bool semaphore_passed(struct intel_ring_buffer *ring)
 
 		acthd -= 4;
 		if (acthd < acthd_min)
-			return false;
+			return NULL;
 	} while (1);
 
-	signaller = &dev_priv->ring[(ring->id + (((ipehr >> 17) & 1) + 1)) % 3];
-	return i915_seqno_passed(signaller->get_seqno(signaller, false),
-				 ioread32(ring->virtual_start+acthd+4)+1);
+	*seqno = ioread32(ring->virtual_start+acthd+4)+1;
+	return &dev_priv->ring[(ring->id + (((ipehr >> 17) & 1) + 1)) % 3];
 }
 
-static bool kick_ring(struct intel_ring_buffer *ring)
+static int semaphore_passed(struct intel_ring_buffer *ring)
 {
-	struct drm_device *dev = ring->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 tmp = I915_READ_CTL(ring);
-	if (tmp & RING_WAIT) {
-		DRM_ERROR("Kicking stuck wait on %s\n",
-			  ring->name);
-		I915_WRITE_CTL(ring, tmp);
-		return true;
-	}
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct intel_ring_buffer *signaller;
+	u32 seqno, ctl;
 
-	if (INTEL_INFO(dev)->gen >= 6 &&
-	    tmp & RING_WAIT_SEMAPHORE &&
-	    semaphore_passed(ring)) {
-		DRM_ERROR("Kicking stuck semaphore on %s\n",
-			  ring->name);
-		I915_WRITE_CTL(ring, tmp);
-		return true;
-	}
-	return false;
+	ring->hangcheck.deadlock = true;
+
+	signaller = semaphore_waits_for(ring, &seqno);
+	if (signaller == NULL || signaller->hangcheck.deadlock)
+		return -1;
+
+	/* cursory check for an unkickable deadlock */
+	ctl = I915_READ_CTL(signaller);
+	if (ctl & RING_WAIT_SEMAPHORE && semaphore_passed(signaller) < 0)
+		return -1;
+
+	return i915_seqno_passed(signaller->get_seqno(signaller, false), seqno);
 }
 
-static bool i915_hangcheck_hung(struct drm_device *dev)
+static void semaphore_clear_deadlocks(struct drm_i915_private *dev_priv)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-
-	if (dev_priv->gpu_error.hangcheck_count++ > 1) {
-		bool hung = true;
+	struct intel_ring_buffer *ring;
+	int i;
 
-		DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
-		i915_handle_error(dev, true);
+	for_each_ring(ring, dev_priv, i)
+		ring->hangcheck.deadlock = false;
+}
 
-		if (!IS_GEN2(dev)) {
-			struct intel_ring_buffer *ring;
-			int i;
+static enum intel_ring_hangcheck_action
+ring_stuck(struct intel_ring_buffer *ring, u32 acthd)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 tmp;
 
-			/* Is the chip hanging on a WAIT_FOR_EVENT?
-			 * If so we can simply poke the RB_WAIT bit
-			 * and break the hang. This should work on
-			 * all but the second generation chipsets.
-			 */
-			for_each_ring(ring, dev_priv, i)
-				hung &= !kick_ring(ring);
-		}
+	if (ring->hangcheck.acthd != acthd)
+		return active;
 
+	if (IS_GEN2(dev))
 		return hung;
+
+	/* Is the chip hanging on a WAIT_FOR_EVENT?
+	 * If so we can simply poke the RB_WAIT bit
+	 * and break the hang. This should work on
+	 * all but the second generation chipsets.
+	 */
+	tmp = I915_READ_CTL(ring);
+	if (tmp & RING_WAIT) {
+		DRM_ERROR("Kicking stuck wait on %s\n",
+			  ring->name);
+		I915_WRITE_CTL(ring, tmp);
+		return kick;
 	}
 
-	return false;
+	if (INTEL_INFO(dev)->gen >= 6 && tmp & RING_WAIT_SEMAPHORE) {
+		switch (semaphore_passed(ring)) {
+		default:
+			return hung;
+		case 1:
+			DRM_ERROR("Kicking stuck semaphore on %s\n",
+				  ring->name);
+			I915_WRITE_CTL(ring, tmp);
+			return kick;
+		case 0:
+			return wait;
+		}
+	}
+
+	return hung;
 }
 
 /**
  * This is called when the chip hasn't reported back with completed
- * batchbuffers in a long time. The first time this is called we simply record
- * ACTHD. If ACTHD hasn't changed by the time the hangcheck timer elapses
- * again, we assume the chip is wedged and try to fix it.
+ * batchbuffers in a long time. We keep track per ring seqno progress and
+ * if there are no progress, hangcheck score for that ring is increased.
+ * Further, acthd is inspected to see if the ring is stuck. On stuck case
+ * we kick the ring. If we see no progress on three subsequent calls
+ * we assume chip is wedged and try to fix it by resetting the chip.
  */
 void i915_hangcheck_elapsed(unsigned long data)
 {
 	struct drm_device *dev = (struct drm_device *)data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint32_t acthd[I915_NUM_RINGS], instdone[I915_NUM_INSTDONE_REG];
 	struct intel_ring_buffer *ring;
-	bool err = false, idle;
 	int i;
+	int busy_count = 0, rings_hung = 0;
+	bool stuck[I915_NUM_RINGS] = { 0 };
+#define BUSY 1
+#define KICK 5
+#define HUNG 20
+#define FIRE 30
 
 	if (!i915_enable_hangcheck)
 		return;
 
-	memset(acthd, 0, sizeof(acthd));
-	idle = true;
 	for_each_ring(ring, dev_priv, i) {
-	    idle &= i915_hangcheck_ring_idle(ring, &err);
-	    acthd[i] = intel_ring_get_active_head(ring);
-	}
+		u32 seqno, acthd;
+		bool busy = true;
+
+		semaphore_clear_deadlocks(dev_priv);
+
+		seqno = ring->get_seqno(ring, false);
+		acthd = intel_ring_get_active_head(ring);
+
+		if (ring->hangcheck.seqno == seqno) {
+			if (ring_idle(ring, seqno)) {
+				if (waitqueue_active(&ring->irq_queue)) {
+					/* Issue a wake-up to catch stuck h/w. */
+					DRM_ERROR("Hangcheck timer elapsed... %s idle\n",
+						  ring->name);
+					wake_up_all(&ring->irq_queue);
+					ring->hangcheck.score += HUNG;
+				} else
+					busy = false;
+			} else {
+				int score;
+
+				/* We always increment the hangcheck score
+				 * if the ring is busy and still processing
+				 * the same request, so that no single request
+				 * can run indefinitely (such as a chain of
+				 * batches). The only time we do not increment
+				 * the hangcheck score on this ring, if this
+				 * ring is in a legitimate wait for another
+				 * ring. In that case the waiting ring is a
+				 * victim and we want to be sure we catch the
+				 * right culprit. Then every time we do kick
+				 * the ring, add a small increment to the
+				 * score so that we can catch a batch that is
+				 * being repeatedly kicked and so responsible
+				 * for stalling the machine.
+				 */
+				ring->hangcheck.action = ring_stuck(ring,
+								    acthd);
+
+				switch (ring->hangcheck.action) {
+				case wait:
+					score = 0;
+					break;
+				case active:
+					score = BUSY;
+					break;
+				case kick:
+					score = KICK;
+					break;
+				case hung:
+					score = HUNG;
+					stuck[i] = true;
+					break;
+				}
+				ring->hangcheck.score += score;
+			}
+		} else {
+			/* Gradually reduce the count so that we catch DoS
+			 * attempts across multiple batches.
+			 */
+			if (ring->hangcheck.score > 0)
+				ring->hangcheck.score--;
+		}
 
-	/* If all work is done then ACTHD clearly hasn't advanced. */
-	if (idle) {
-		if (err) {
-			if (i915_hangcheck_hung(dev))
-				return;
+		ring->hangcheck.seqno = seqno;
+		ring->hangcheck.acthd = acthd;
+		busy_count += busy;
+	}
 
-			goto repeat;
+	for_each_ring(ring, dev_priv, i) {
+		if (ring->hangcheck.score > FIRE) {
+			DRM_ERROR("%s on %s\n",
+				  stuck[i] ? "stuck" : "no progress",
+				  ring->name);
+			rings_hung++;
 		}
-
-		dev_priv->gpu_error.hangcheck_count = 0;
-		return;
 	}
 
-	i915_get_extra_instdone(dev, instdone);
-	if (memcmp(dev_priv->gpu_error.last_acthd, acthd,
-		   sizeof(acthd)) == 0 &&
-	    memcmp(dev_priv->gpu_error.prev_instdone, instdone,
-		   sizeof(instdone)) == 0) {
-		if (i915_hangcheck_hung(dev))
-			return;
-	} else {
-		dev_priv->gpu_error.hangcheck_count = 0;
+	if (rings_hung)
+		return i915_handle_error(dev, true);
 
-		memcpy(dev_priv->gpu_error.last_acthd, acthd,
-		       sizeof(acthd));
-		memcpy(dev_priv->gpu_error.prev_instdone, instdone,
-		       sizeof(instdone));
-	}
+	if (busy_count)
+		/* Reset timer case chip hangs without another request
+		 * being added */
+		mod_timer(&dev_priv->gpu_error.hangcheck_timer,
+			  round_jiffies_up(jiffies +
+					   DRM_I915_HANGCHECK_JIFFIES));
+}
+
+static void ibx_irq_preinstall(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (HAS_PCH_NOP(dev))
+		return;
 
-repeat:
-	/* Reset timer case chip hangs without another request being added */
-	mod_timer(&dev_priv->gpu_error.hangcheck_timer,
-		  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
+	/* south display irq */
+	I915_WRITE(SDEIMR, 0xffffffff);
+	/*
+	 * SDEIER is also touched by the interrupt handler to work around missed
+	 * PCH interrupts. Hence we can't update it after the interrupt handler
+	 * is enabled - instead we unconditionally enable all PCH interrupt
+	 * sources here, but then only unmask them as needed with SDEIMR.
+	 */
+	I915_WRITE(SDEIER, 0xffffffff);
+	POSTING_READ(SDEIER);
 }
 
 /* drm_dma.h hooks
@@ -2113,19 +2581,34 @@ static void ironlake_irq_preinstall(struct drm_device *dev)
 	I915_WRITE(GTIER, 0x0);
 	POSTING_READ(GTIER);
 
-	if (HAS_PCH_NOP(dev))
-		return;
+	ibx_irq_preinstall(dev);
+}
 
-	/* south display irq */
-	I915_WRITE(SDEIMR, 0xffffffff);
-	/*
-	 * SDEIER is also touched by the interrupt handler to work around missed
-	 * PCH interrupts. Hence we can't update it after the interrupt handler
-	 * is enabled - instead we unconditionally enable all PCH interrupt
-	 * sources here, but then only unmask them as needed with SDEIMR.
-	 */
-	I915_WRITE(SDEIER, 0xffffffff);
-	POSTING_READ(SDEIER);
+static void ivybridge_irq_preinstall(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+
+	atomic_set(&dev_priv->irq_received, 0);
+
+	I915_WRITE(HWSTAM, 0xeffe);
+
+	/* XXX hotplug from PCH */
+
+	I915_WRITE(DEIMR, 0xffffffff);
+	I915_WRITE(DEIER, 0x0);
+	POSTING_READ(DEIER);
+
+	/* and GT */
+	I915_WRITE(GTIMR, 0xffffffff);
+	I915_WRITE(GTIER, 0x0);
+	POSTING_READ(GTIER);
+
+	/* Power management */
+	I915_WRITE(GEN6_PMIMR, 0xffffffff);
+	I915_WRITE(GEN6_PMIER, 0x0);
+	POSTING_READ(GEN6_PMIER);
+
+	ibx_irq_preinstall(dev);
 }
 
 static void valleyview_irq_preinstall(struct drm_device *dev)
@@ -2201,33 +2684,41 @@ static void ibx_irq_postinstall(struct drm_device *dev)
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	u32 mask;
 
-	if (HAS_PCH_IBX(dev))
-		mask = SDE_GMBUS | SDE_AUX_MASK;
-	else
-		mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT;
-
 	if (HAS_PCH_NOP(dev))
 		return;
 
+	if (HAS_PCH_IBX(dev)) {
+		mask = SDE_GMBUS | SDE_AUX_MASK | SDE_TRANSB_FIFO_UNDER |
+		       SDE_TRANSA_FIFO_UNDER | SDE_POISON;
+	} else {
+		mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT | SDE_ERROR_CPT;
+
+		I915_WRITE(SERR_INT, I915_READ(SERR_INT));
+	}
+
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
 	I915_WRITE(SDEIMR, ~mask);
 }
 
 static int ironlake_irq_postinstall(struct drm_device *dev)
 {
+	unsigned long irqflags;
+
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	/* enable kind of interrupts always enabled */
 	u32 display_mask = DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT |
 			   DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE |
-			   DE_AUX_CHANNEL_A;
-	u32 render_irqs;
+			   DE_AUX_CHANNEL_A | DE_PIPEB_FIFO_UNDERRUN |
+			   DE_PIPEA_FIFO_UNDERRUN | DE_POISON;
+	u32 gt_irqs;
 
 	dev_priv->irq_mask = ~display_mask;
 
 	/* should always can generate irq */
 	I915_WRITE(DEIIR, I915_READ(DEIIR));
 	I915_WRITE(DEIMR, dev_priv->irq_mask);
-	I915_WRITE(DEIER, display_mask | DE_PIPEA_VBLANK | DE_PIPEB_VBLANK);
+	I915_WRITE(DEIER, display_mask |
+			  DE_PIPEA_VBLANK | DE_PIPEB_VBLANK | DE_PCU_EVENT);
 	POSTING_READ(DEIER);
 
 	dev_priv->gt_irq_mask = ~0;
@@ -2235,26 +2726,28 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 
+	gt_irqs = GT_RENDER_USER_INTERRUPT;
+
 	if (IS_GEN6(dev))
-		render_irqs =
-			GT_USER_INTERRUPT |
-			GEN6_BSD_USER_INTERRUPT |
-			GEN6_BLITTER_USER_INTERRUPT;
+		gt_irqs |= GT_BLT_USER_INTERRUPT | GT_BSD_USER_INTERRUPT;
 	else
-		render_irqs =
-			GT_USER_INTERRUPT |
-			GT_PIPE_NOTIFY |
-			GT_BSD_USER_INTERRUPT;
-	I915_WRITE(GTIER, render_irqs);
+		gt_irqs |= GT_RENDER_PIPECTL_NOTIFY_INTERRUPT |
+			   ILK_BSD_USER_INTERRUPT;
+
+	I915_WRITE(GTIER, gt_irqs);
 	POSTING_READ(GTIER);
 
 	ibx_irq_postinstall(dev);
 
 	if (IS_IRONLAKE_M(dev)) {
-		/* Clear & enable PCU event interrupts */
-		I915_WRITE(DEIIR, DE_PCU_EVENT);
-		I915_WRITE(DEIER, I915_READ(DEIER) | DE_PCU_EVENT);
+		/* Enable PCU event interrupts
+		 *
+		 * spinlocking not required here for correctness since interrupt
+		 * setup is guaranteed to run in single-threaded context. But we
+		 * need it to make the assert_spin_locked happy. */
+		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 		ironlake_enable_display_irq(dev_priv, DE_PCU_EVENT);
+		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 	}
 
 	return 0;
@@ -2269,12 +2762,15 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		DE_PLANEC_FLIP_DONE_IVB |
 		DE_PLANEB_FLIP_DONE_IVB |
 		DE_PLANEA_FLIP_DONE_IVB |
-		DE_AUX_CHANNEL_A_IVB;
-	u32 render_irqs;
+		DE_AUX_CHANNEL_A_IVB |
+		DE_ERR_INT_IVB;
+	u32 pm_irqs = GEN6_PM_RPS_EVENTS;
+	u32 gt_irqs;
 
 	dev_priv->irq_mask = ~display_mask;
 
 	/* should always can generate irq */
+	I915_WRITE(GEN7_ERR_INT, I915_READ(GEN7_ERR_INT));
 	I915_WRITE(DEIIR, I915_READ(DEIIR));
 	I915_WRITE(DEIMR, dev_priv->irq_mask);
 	I915_WRITE(DEIER,
@@ -2284,16 +2780,32 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		   DE_PIPEA_VBLANK_IVB);
 	POSTING_READ(DEIER);
 
-	dev_priv->gt_irq_mask = ~GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
+	dev_priv->gt_irq_mask = ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 
-	render_irqs = GT_USER_INTERRUPT | GEN6_BSD_USER_INTERRUPT |
-		GEN6_BLITTER_USER_INTERRUPT | GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
-	I915_WRITE(GTIER, render_irqs);
+	gt_irqs = GT_RENDER_USER_INTERRUPT | GT_BSD_USER_INTERRUPT |
+		  GT_BLT_USER_INTERRUPT | GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
+	I915_WRITE(GTIER, gt_irqs);
 	POSTING_READ(GTIER);
 
+	I915_WRITE(GEN6_PMIIR, I915_READ(GEN6_PMIIR));
+	if (HAS_VEBOX(dev))
+		pm_irqs |= PM_VEBOX_USER_INTERRUPT |
+			PM_VEBOX_CS_ERROR_INTERRUPT;
+
+	/* Our enable/disable rps functions may touch these registers so
+	 * make sure to set a known state for only the non-RPS bits.
+	 * The RMW is extra paranoia since this should be called after being set
+	 * to a known state in preinstall.
+	 * */
+	I915_WRITE(GEN6_PMIMR,
+		   (I915_READ(GEN6_PMIMR) | ~GEN6_PM_RPS_EVENTS) & ~pm_irqs);
+	I915_WRITE(GEN6_PMIER,
+		   (I915_READ(GEN6_PMIER) & GEN6_PM_RPS_EVENTS) | pm_irqs);
+	POSTING_READ(GEN6_PMIER);
+
 	ibx_irq_postinstall(dev);
 
 	return 0;
@@ -2302,10 +2814,9 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 static int valleyview_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 gt_irqs;
 	u32 enable_mask;
 	u32 pipestat_enable = PLANE_FLIP_DONE_INT_EN_VLV;
-	u32 render_irqs;
-	u16 msid;
 
 	enable_mask = I915_DISPLAY_PORT_INTERRUPT;
 	enable_mask |= I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
@@ -2321,13 +2832,6 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 		I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT |
 		I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT;
 
-	/* Hack for broken MSIs on VLV */
-	pci_write_config_dword(dev_priv->dev->pdev, 0x94, 0xfee00000);
-	pci_read_config_word(dev->pdev, 0x98, &msid);
-	msid &= 0xff; /* mask out delivery bits */
-	msid |= (1<<14);
-	pci_write_config_word(dev_priv->dev->pdev, 0x98, msid);
-
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	POSTING_READ(PORT_HOTPLUG_EN);
 
@@ -2348,9 +2852,9 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 
-	render_irqs = GT_USER_INTERRUPT | GEN6_BSD_USER_INTERRUPT |
-		GEN6_BLITTER_USER_INTERRUPT;
-	I915_WRITE(GTIER, render_irqs);
+	gt_irqs = GT_RENDER_USER_INTERRUPT | GT_BSD_USER_INTERRUPT |
+		GT_BLT_USER_INTERRUPT;
+	I915_WRITE(GTIER, gt_irqs);
 	POSTING_READ(GTIER);
 
 	/* ack & enable invalid PTE error interrupts */
@@ -2402,6 +2906,8 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(DEIMR, 0xffffffff);
 	I915_WRITE(DEIER, 0x0);
 	I915_WRITE(DEIIR, I915_READ(DEIIR));
+	if (IS_GEN7(dev))
+		I915_WRITE(GEN7_ERR_INT, I915_READ(GEN7_ERR_INT));
 
 	I915_WRITE(GTIMR, 0xffffffff);
 	I915_WRITE(GTIER, 0x0);
@@ -2413,6 +2919,8 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(SDEIMR, 0xffffffff);
 	I915_WRITE(SDEIER, 0x0);
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
+	if (HAS_PCH_CPT(dev) || HAS_PCH_LPT(dev))
+		I915_WRITE(SERR_INT, I915_READ(SERR_INT));
 }
 
 static void i8xx_irq_preinstall(struct drm_device * dev)
@@ -2626,7 +3134,7 @@ static int i915_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(IER, enable_mask);
 	POSTING_READ(IER);
 
-	intel_opregion_enable_asle(dev);
+	i915_enable_asle_pipestat(dev);
 
 	return 0;
 }
@@ -2715,12 +3223,9 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 				  hotplug_status);
-			if (hotplug_trigger) {
-				if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_status_i915))
-					i915_hpd_irq_setup(dev);
-				queue_work(dev_priv->wq,
-					   &dev_priv->hotplug_work);
-			}
+
+			intel_hpd_irq_handler(dev, hotplug_trigger, hpd_status_i915);
+
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			POSTING_READ(PORT_HOTPLUG_STAT);
 		}
@@ -2860,7 +3365,7 @@ static int i965_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	POSTING_READ(PORT_HOTPLUG_EN);
 
-	intel_opregion_enable_asle(dev);
+	i915_enable_asle_pipestat(dev);
 
 	return 0;
 }
@@ -2872,6 +3377,8 @@ static void i915_hpd_irq_setup(struct drm_device *dev)
 	struct intel_encoder *intel_encoder;
 	u32 hotplug_en;
 
+	assert_spin_locked(&dev_priv->irq_lock);
+
 	if (I915_HAS_HOTPLUG(dev)) {
 		hotplug_en = I915_READ(PORT_HOTPLUG_EN);
 		hotplug_en &= ~HOTPLUG_INT_EN_MASK;
@@ -2952,17 +3459,14 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 			u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
 			u32 hotplug_trigger = hotplug_status & (IS_G4X(dev) ?
 								  HOTPLUG_INT_STATUS_G4X :
-								  HOTPLUG_INT_STATUS_I965);
+								  HOTPLUG_INT_STATUS_I915);
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 				  hotplug_status);
-			if (hotplug_trigger) {
-				if (hotplug_irq_storm_detect(dev, hotplug_trigger,
-							    IS_G4X(dev) ? hpd_status_gen4 : hpd_status_i965))
-					i915_hpd_irq_setup(dev);
-				queue_work(dev_priv->wq,
-					   &dev_priv->hotplug_work);
-			}
+
+			intel_hpd_irq_handler(dev, hotplug_trigger,
+					      IS_G4X(dev) ? hpd_status_gen4 : hpd_status_i915);
+
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
@@ -3113,9 +3617,9 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->disable_vblank = valleyview_disable_vblank;
 		dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
 	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
-		/* Share pre & uninstall handlers with ILK/SNB */
+		/* Share uninstall handlers with ILK/SNB */
 		dev->driver->irq_handler = ivybridge_irq_handler;
-		dev->driver->irq_preinstall = ironlake_irq_preinstall;
+		dev->driver->irq_preinstall = ivybridge_irq_preinstall;
 		dev->driver->irq_postinstall = ivybridge_irq_postinstall;
 		dev->driver->irq_uninstall = ironlake_irq_uninstall;
 		dev->driver->enable_vblank = ivybridge_enable_vblank;
@@ -3158,6 +3662,7 @@ void intel_hpd_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
+	unsigned long irqflags;
 	int i;
 
 	for (i = 1; i < HPD_NUM_PINS; i++) {
@@ -3170,6 +3675,11 @@ void intel_hpd_init(struct drm_device *dev)
 		if (!connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE)
 			connector->polled = DRM_CONNECTOR_POLL_HPD;
 	}
+
+	/* Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked checks happy. */
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 }
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 2d6b62e..f2326fc 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -147,15 +147,9 @@
 #define   VGA_MSR_MEM_EN (1<<1)
 #define   VGA_MSR_CGA_MODE (1<<0)
 
-/*
- * SR01 is the only VGA register touched on non-UMS setups.
- * VLV doesn't do UMS, so the sequencer index/data registers
- * are the only VGA registers which need to include
- * display_mmio_offset.
- */
-#define VGA_SR_INDEX (dev_priv->info->display_mmio_offset + 0x3c4)
+#define VGA_SR_INDEX 0x3c4
 #define SR01			1
-#define VGA_SR_DATA (dev_priv->info->display_mmio_offset + 0x3c5)
+#define VGA_SR_DATA 0x3c5
 
 #define VGA_AR_INDEX 0x3c0
 #define   VGA_AR_VID_EN (1<<5)
@@ -265,13 +259,19 @@
 #define  MI_SEMAPHORE_UPDATE	    (1<<21)
 #define  MI_SEMAPHORE_COMPARE	    (1<<20)
 #define  MI_SEMAPHORE_REGISTER	    (1<<18)
-#define  MI_SEMAPHORE_SYNC_RV	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_RB	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_VR	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_VB	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_BR	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_BV	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_INVALID  (1<<0)
+#define  MI_SEMAPHORE_SYNC_VR	    (0<<16) /* RCS  wait for VCS  (RVSYNC) */
+#define  MI_SEMAPHORE_SYNC_VER	    (1<<16) /* RCS  wait for VECS (RVESYNC) */
+#define  MI_SEMAPHORE_SYNC_BR	    (2<<16) /* RCS  wait for BCS  (RBSYNC) */
+#define  MI_SEMAPHORE_SYNC_BV	    (0<<16) /* VCS  wait for BCS  (VBSYNC) */
+#define  MI_SEMAPHORE_SYNC_VEV	    (1<<16) /* VCS  wait for VECS (VVESYNC) */
+#define  MI_SEMAPHORE_SYNC_RV	    (2<<16) /* VCS  wait for RCS  (VRSYNC) */
+#define  MI_SEMAPHORE_SYNC_RB	    (0<<16) /* BCS  wait for RCS  (BRSYNC) */
+#define  MI_SEMAPHORE_SYNC_VEB	    (1<<16) /* BCS  wait for VECS (BVESYNC) */
+#define  MI_SEMAPHORE_SYNC_VB	    (2<<16) /* BCS  wait for VCS  (BVSYNC) */
+#define  MI_SEMAPHORE_SYNC_BVE	    (0<<16) /* VECS wait for BCS  (VEBSYNC) */
+#define  MI_SEMAPHORE_SYNC_VVE	    (1<<16) /* VECS wait for VCS  (VEVSYNC) */
+#define  MI_SEMAPHORE_SYNC_RVE	    (2<<16) /* VECS wait for RCS  (VERSYNC) */
+#define  MI_SEMAPHORE_SYNC_INVALID  (3<<16)
 /*
  * 3D instructions used by the kernel
  */
@@ -342,33 +342,74 @@
 #define  DEBUG_RESET_DISPLAY		(1<<9)
 
 /*
- * DPIO - a special bus for various display related registers to hide behind:
- *  0x800c: m1, m2, n, p1, p2, k dividers
- *  0x8014: REF and SFR select
- *  0x8014: N divider, VCO select
- *  0x801c/3c: core clock bits
- *  0x8048/68: low pass filter coefficients
- *  0x8100: fast clock controls
+ * IOSF sideband
+ */
+#define VLV_IOSF_DOORBELL_REQ			(VLV_DISPLAY_BASE + 0x2100)
+#define   IOSF_DEVFN_SHIFT			24
+#define   IOSF_OPCODE_SHIFT			16
+#define   IOSF_PORT_SHIFT			8
+#define   IOSF_BYTE_ENABLES_SHIFT		4
+#define   IOSF_BAR_SHIFT			1
+#define   IOSF_SB_BUSY				(1<<0)
+#define   IOSF_PORT_PUNIT			0x4
+#define   IOSF_PORT_NC				0x11
+#define   IOSF_PORT_DPIO			0x12
+#define VLV_IOSF_DATA				(VLV_DISPLAY_BASE + 0x2104)
+#define VLV_IOSF_ADDR				(VLV_DISPLAY_BASE + 0x2108)
+
+#define PUNIT_OPCODE_REG_READ			6
+#define PUNIT_OPCODE_REG_WRITE			7
+
+#define PUNIT_REG_GPU_LFM			0xd3
+#define PUNIT_REG_GPU_FREQ_REQ			0xd4
+#define PUNIT_REG_GPU_FREQ_STS			0xd8
+#define PUNIT_REG_MEDIA_TURBO_FREQ_REQ		0xdc
+
+#define PUNIT_FUSE_BUS2				0xf6 /* bits 47:40 */
+#define PUNIT_FUSE_BUS1				0xf5 /* bits 55:48 */
+
+#define IOSF_NC_FB_GFX_FREQ_FUSE		0x1c
+#define   FB_GFX_MAX_FREQ_FUSE_SHIFT		3
+#define   FB_GFX_MAX_FREQ_FUSE_MASK		0x000007f8
+#define   FB_GFX_FGUARANTEED_FREQ_FUSE_SHIFT	11
+#define   FB_GFX_FGUARANTEED_FREQ_FUSE_MASK	0x0007f800
+#define IOSF_NC_FB_GFX_FMAX_FUSE_HI		0x34
+#define   FB_FMAX_VMIN_FREQ_HI_MASK		0x00000007
+#define IOSF_NC_FB_GFX_FMAX_FUSE_LO		0x30
+#define   FB_FMAX_VMIN_FREQ_LO_SHIFT		27
+#define   FB_FMAX_VMIN_FREQ_LO_MASK		0xf8000000
+
+/*
+ * DPIO - a special bus for various display related registers to hide behind
  *
  * DPIO is VLV only.
+ *
+ * Note: digital port B is DDI0, digital pot C is DDI1
  */
-#define DPIO_PKT			(VLV_DISPLAY_BASE + 0x2100)
-#define  DPIO_RID			(0<<24)
-#define  DPIO_OP_WRITE			(1<<16)
-#define  DPIO_OP_READ			(0<<16)
-#define  DPIO_PORTID			(0x12<<8)
-#define  DPIO_BYTE			(0xf<<4)
-#define  DPIO_BUSY			(1<<0) /* status only */
-#define DPIO_DATA			(VLV_DISPLAY_BASE + 0x2104)
-#define DPIO_REG			(VLV_DISPLAY_BASE + 0x2108)
+#define DPIO_DEVFN			0
+#define DPIO_OPCODE_REG_WRITE		1
+#define DPIO_OPCODE_REG_READ		0
+
 #define DPIO_CTL			(VLV_DISPLAY_BASE + 0x2110)
 #define  DPIO_MODSEL1			(1<<3) /* if ref clk b == 27 */
 #define  DPIO_MODSEL0			(1<<2) /* if ref clk a == 27 */
 #define  DPIO_SFR_BYPASS		(1<<1)
 #define  DPIO_RESET			(1<<0)
 
+#define _DPIO_TX3_SWING_CTL4_A		0x690
+#define _DPIO_TX3_SWING_CTL4_B		0x2a90
+#define DPIO_TX3_SWING_CTL4(pipe) _PIPE(pipe, _DPIO_TX_SWING_CTL4_A, \
+					_DPIO_TX3_SWING_CTL4_B)
+
+/*
+ * Per pipe/PLL DPIO regs
+ */
 #define _DPIO_DIV_A			0x800c
 #define   DPIO_POST_DIV_SHIFT		(28) /* 3 bits */
+#define   DPIO_POST_DIV_DAC		0
+#define   DPIO_POST_DIV_HDMIDP		1 /* DAC 225-400M rate */
+#define   DPIO_POST_DIV_LVDS1		2
+#define   DPIO_POST_DIV_LVDS2		3
 #define   DPIO_K_SHIFT			(24) /* 4 bits */
 #define   DPIO_P1_SHIFT			(21) /* 3 bits */
 #define   DPIO_P2_SHIFT			(16) /* 5 bits */
@@ -394,14 +435,111 @@
 #define _DPIO_CORE_CLK_B		0x803c
 #define DPIO_CORE_CLK(pipe) _PIPE(pipe, _DPIO_CORE_CLK_A, _DPIO_CORE_CLK_B)
 
-#define _DPIO_LFP_COEFF_A		0x8048
-#define _DPIO_LFP_COEFF_B		0x8068
-#define DPIO_LFP_COEFF(pipe) _PIPE(pipe, _DPIO_LFP_COEFF_A, _DPIO_LFP_COEFF_B)
+#define _DPIO_IREF_CTL_A		0x8040
+#define _DPIO_IREF_CTL_B		0x8060
+#define DPIO_IREF_CTL(pipe) _PIPE(pipe, _DPIO_IREF_CTL_A, _DPIO_IREF_CTL_B)
+
+#define DPIO_IREF_BCAST			0xc044
+#define _DPIO_IREF_A			0x8044
+#define _DPIO_IREF_B			0x8064
+#define DPIO_IREF(pipe) _PIPE(pipe, _DPIO_IREF_A, _DPIO_IREF_B)
+
+#define _DPIO_PLL_CML_A			0x804c
+#define _DPIO_PLL_CML_B			0x806c
+#define DPIO_PLL_CML(pipe) _PIPE(pipe, _DPIO_PLL_CML_A, _DPIO_PLL_CML_B)
+
+#define _DPIO_LPF_COEFF_A		0x8048
+#define _DPIO_LPF_COEFF_B		0x8068
+#define DPIO_LPF_COEFF(pipe) _PIPE(pipe, _DPIO_LPF_COEFF_A, _DPIO_LPF_COEFF_B)
+
+#define DPIO_CALIBRATION		0x80ac
 
 #define DPIO_FASTCLK_DISABLE		0x8100
 
-#define DPIO_DATA_CHANNEL1		0x8220
-#define DPIO_DATA_CHANNEL2		0x8420
+/*
+ * Per DDI channel DPIO regs
+ */
+
+#define _DPIO_PCS_TX_0			0x8200
+#define _DPIO_PCS_TX_1			0x8400
+#define   DPIO_PCS_TX_LANE2_RESET	(1<<16)
+#define   DPIO_PCS_TX_LANE1_RESET	(1<<7)
+#define DPIO_PCS_TX(port) _PORT(port, _DPIO_PCS_TX_0, _DPIO_PCS_TX_1)
+
+#define _DPIO_PCS_CLK_0			0x8204
+#define _DPIO_PCS_CLK_1			0x8404
+#define   DPIO_PCS_CLK_CRI_RXEB_EIOS_EN	(1<<22)
+#define   DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN (1<<21)
+#define   DPIO_PCS_CLK_DATAWIDTH_SHIFT	(6)
+#define   DPIO_PCS_CLK_SOFT_RESET	(1<<5)
+#define DPIO_PCS_CLK(port) _PORT(port, _DPIO_PCS_CLK_0, _DPIO_PCS_CLK_1)
+
+#define _DPIO_PCS_CTL_OVR1_A		0x8224
+#define _DPIO_PCS_CTL_OVR1_B		0x8424
+#define DPIO_PCS_CTL_OVER1(port) _PORT(port, _DPIO_PCS_CTL_OVR1_A, \
+				       _DPIO_PCS_CTL_OVR1_B)
+
+#define _DPIO_PCS_STAGGER0_A		0x822c
+#define _DPIO_PCS_STAGGER0_B		0x842c
+#define DPIO_PCS_STAGGER0(port) _PORT(port, _DPIO_PCS_STAGGER0_A, \
+				      _DPIO_PCS_STAGGER0_B)
+
+#define _DPIO_PCS_STAGGER1_A		0x8230
+#define _DPIO_PCS_STAGGER1_B		0x8430
+#define DPIO_PCS_STAGGER1(port) _PORT(port, _DPIO_PCS_STAGGER1_A, \
+				      _DPIO_PCS_STAGGER1_B)
+
+#define _DPIO_PCS_CLOCKBUF0_A		0x8238
+#define _DPIO_PCS_CLOCKBUF0_B		0x8438
+#define DPIO_PCS_CLOCKBUF0(port) _PORT(port, _DPIO_PCS_CLOCKBUF0_A, \
+				       _DPIO_PCS_CLOCKBUF0_B)
+
+#define _DPIO_PCS_CLOCKBUF8_A		0x825c
+#define _DPIO_PCS_CLOCKBUF8_B		0x845c
+#define DPIO_PCS_CLOCKBUF8(port) _PORT(port, _DPIO_PCS_CLOCKBUF8_A, \
+				       _DPIO_PCS_CLOCKBUF8_B)
+
+#define _DPIO_TX_SWING_CTL2_A		0x8288
+#define _DPIO_TX_SWING_CTL2_B		0x8488
+#define DPIO_TX_SWING_CTL2(port) _PORT(port, _DPIO_TX_SWING_CTL2_A, \
+				       _DPIO_TX_SWING_CTL2_B)
+
+#define _DPIO_TX_SWING_CTL3_A		0x828c
+#define _DPIO_TX_SWING_CTL3_B		0x848c
+#define DPIO_TX_SWING_CTL3(port) _PORT(port, _DPIO_TX_SWING_CTL3_A, \
+				       _DPIO_TX_SWING_CTL3_B)
+
+#define _DPIO_TX_SWING_CTL4_A		0x8290
+#define _DPIO_TX_SWING_CTL4_B		0x8490
+#define DPIO_TX_SWING_CTL4(port) _PORT(port, _DPIO_TX_SWING_CTL4_A, \
+				       _DPIO_TX_SWING_CTL4_B)
+
+#define _DPIO_TX_OCALINIT_0		0x8294
+#define _DPIO_TX_OCALINIT_1		0x8494
+#define   DPIO_TX_OCALINIT_EN		(1<<31)
+#define DPIO_TX_OCALINIT(port) _PORT(port, _DPIO_TX_OCALINIT_0, \
+				     _DPIO_TX_OCALINIT_1)
+
+#define _DPIO_TX_CTL_0			0x82ac
+#define _DPIO_TX_CTL_1			0x84ac
+#define DPIO_TX_CTL(port) _PORT(port, _DPIO_TX_CTL_0, _DPIO_TX_CTL_1)
+
+#define _DPIO_TX_LANE_0			0x82b8
+#define _DPIO_TX_LANE_1			0x84b8
+#define DPIO_TX_LANE(port) _PORT(port, _DPIO_TX_LANE_0, _DPIO_TX_LANE_1)
+
+#define _DPIO_DATA_CHANNEL1		0x8220
+#define _DPIO_DATA_CHANNEL2		0x8420
+#define DPIO_DATA_CHANNEL(port) _PORT(port, _DPIO_DATA_CHANNEL1, _DPIO_DATA_CHANNEL2)
+
+#define _DPIO_PORT0_PCS0		0x0220
+#define _DPIO_PORT0_PCS1		0x0420
+#define _DPIO_PORT1_PCS2		0x2620
+#define _DPIO_PORT1_PCS3		0x2820
+#define DPIO_DATA_LANE_A(port) _PORT(port, _DPIO_PORT0_PCS0, _DPIO_PORT1_PCS2)
+#define DPIO_DATA_LANE_B(port) _PORT(port, _DPIO_PORT0_PCS1, _DPIO_PORT1_PCS3)
+#define DPIO_DATA_CHANNEL1              0x8220
+#define DPIO_DATA_CHANNEL2              0x8420
 
 /*
  * Fence registers
@@ -443,6 +581,7 @@
 #define RENDER_RING_BASE	0x02000
 #define BSD_RING_BASE		0x04000
 #define GEN6_BSD_RING_BASE	0x12000
+#define VEBOX_RING_BASE		0x1a000
 #define BLT_RING_BASE		0x22000
 #define RING_TAIL(base)		((base)+0x30)
 #define RING_HEAD(base)		((base)+0x34)
@@ -450,12 +589,20 @@
 #define RING_CTL(base)		((base)+0x3c)
 #define RING_SYNC_0(base)	((base)+0x40)
 #define RING_SYNC_1(base)	((base)+0x44)
-#define GEN6_RVSYNC (RING_SYNC_0(RENDER_RING_BASE))
-#define GEN6_RBSYNC (RING_SYNC_1(RENDER_RING_BASE))
-#define GEN6_VRSYNC (RING_SYNC_1(GEN6_BSD_RING_BASE))
-#define GEN6_VBSYNC (RING_SYNC_0(GEN6_BSD_RING_BASE))
-#define GEN6_BRSYNC (RING_SYNC_0(BLT_RING_BASE))
-#define GEN6_BVSYNC (RING_SYNC_1(BLT_RING_BASE))
+#define RING_SYNC_2(base)	((base)+0x48)
+#define GEN6_RVSYNC	(RING_SYNC_0(RENDER_RING_BASE))
+#define GEN6_RBSYNC	(RING_SYNC_1(RENDER_RING_BASE))
+#define GEN6_RVESYNC	(RING_SYNC_2(RENDER_RING_BASE))
+#define GEN6_VBSYNC	(RING_SYNC_0(GEN6_BSD_RING_BASE))
+#define GEN6_VRSYNC	(RING_SYNC_1(GEN6_BSD_RING_BASE))
+#define GEN6_VVESYNC	(RING_SYNC_2(GEN6_BSD_RING_BASE))
+#define GEN6_BRSYNC	(RING_SYNC_0(BLT_RING_BASE))
+#define GEN6_BVSYNC	(RING_SYNC_1(BLT_RING_BASE))
+#define GEN6_BVESYNC	(RING_SYNC_2(BLT_RING_BASE))
+#define GEN6_VEBSYNC	(RING_SYNC_0(VEBOX_RING_BASE))
+#define GEN6_VERSYNC	(RING_SYNC_1(VEBOX_RING_BASE))
+#define GEN6_VEVSYNC	(RING_SYNC_2(VEBOX_RING_BASE))
+#define GEN6_NOSYNC 0
 #define RING_MAX_IDLE(base)	((base)+0x54)
 #define RING_HWS_PGA(base)	((base)+0x80)
 #define RING_HWS_PGA_GEN6(base)	((base)+0x2080)
@@ -467,6 +614,7 @@
 #define DONE_REG		0x40b0
 #define BSD_HWS_PGA_GEN7	(0x04180)
 #define BLT_HWS_PGA_GEN7	(0x04280)
+#define VEBOX_HWS_PGA_GEN7	(0x04380)
 #define RING_ACTHD(base)	((base)+0x74)
 #define RING_NOPID(base)	((base)+0x94)
 #define RING_IMR(base)		((base)+0xa8)
@@ -527,7 +675,11 @@
 
 #define ERROR_GEN6	0x040a0
 #define GEN7_ERR_INT	0x44040
-#define   ERR_INT_MMIO_UNCLAIMED (1<<13)
+#define   ERR_INT_POISON		(1<<31)
+#define   ERR_INT_MMIO_UNCLAIMED	(1<<13)
+#define   ERR_INT_FIFO_UNDERRUN_C	(1<<6)
+#define   ERR_INT_FIFO_UNDERRUN_B	(1<<3)
+#define   ERR_INT_FIFO_UNDERRUN_A	(1<<0)
 
 #define FPGA_DBG		0x42300
 #define   FPGA_DBG_RM_NOCLAIM	(1<<31)
@@ -583,24 +735,7 @@
 #define VLV_IIR		(VLV_DISPLAY_BASE + 0x20a4)
 #define VLV_IMR		(VLV_DISPLAY_BASE + 0x20a8)
 #define VLV_ISR		(VLV_DISPLAY_BASE + 0x20ac)
-#define   I915_PIPE_CONTROL_NOTIFY_INTERRUPT		(1<<18)
-#define   I915_DISPLAY_PORT_INTERRUPT			(1<<17)
-#define   I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT	(1<<15)
-#define   I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT	(1<<14) /* p-state */
-#define   I915_HWB_OOM_INTERRUPT			(1<<13)
-#define   I915_SYNC_STATUS_INTERRUPT			(1<<12)
-#define   I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT	(1<<11)
-#define   I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT	(1<<10)
-#define   I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT	(1<<9)
-#define   I915_DISPLAY_PLANE_C_FLIP_PENDING_INTERRUPT	(1<<8)
-#define   I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT		(1<<7)
-#define   I915_DISPLAY_PIPE_A_EVENT_INTERRUPT		(1<<6)
-#define   I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT		(1<<5)
-#define   I915_DISPLAY_PIPE_B_EVENT_INTERRUPT		(1<<4)
-#define   I915_DEBUG_INTERRUPT				(1<<2)
-#define   I915_USER_INTERRUPT				(1<<1)
-#define   I915_ASLE_INTERRUPT				(1<<0)
-#define   I915_BSD_USER_INTERRUPT                      (1<<25)
+#define VLV_PCBR	(VLV_DISPLAY_BASE + 0x2120)
 #define   DISPLAY_PLANE_FLIP_PENDING(plane) (1<<(11-(plane))) /* A and B only */
 #define EIR		0x020b0
 #define EMR		0x020b4
@@ -712,28 +847,6 @@
 #define CACHE_MODE_1		0x7004 /* IVB+ */
 #define   PIXEL_SUBSPAN_COLLECT_OPT_DISABLE (1<<6)
 
-/* GEN6 interrupt control
- * Note that the per-ring interrupt bits do alias with the global interrupt bits
- * in GTIMR. */
-#define GEN6_RENDER_HWSTAM	0x2098
-#define GEN6_RENDER_IMR		0x20a8
-#define   GEN6_RENDER_CONTEXT_SWITCH_INTERRUPT		(1 << 8)
-#define   GEN6_RENDER_PPGTT_PAGE_FAULT			(1 << 7)
-#define   GEN6_RENDER_TIMEOUT_COUNTER_EXPIRED		(1 << 6)
-#define   GEN6_RENDER_L3_PARITY_ERROR			(1 << 5)
-#define   GEN6_RENDER_PIPE_CONTROL_NOTIFY_INTERRUPT	(1 << 4)
-#define   GEN6_RENDER_COMMAND_PARSER_MASTER_ERROR	(1 << 3)
-#define   GEN6_RENDER_SYNC_STATUS			(1 << 2)
-#define   GEN6_RENDER_DEBUG_INTERRUPT			(1 << 1)
-#define   GEN6_RENDER_USER_INTERRUPT			(1 << 0)
-
-#define GEN6_BLITTER_HWSTAM	0x22098
-#define GEN6_BLITTER_IMR	0x220a8
-#define   GEN6_BLITTER_MI_FLUSH_DW_NOTIFY_INTERRUPT	(1 << 26)
-#define   GEN6_BLITTER_COMMAND_PARSER_MASTER_ERROR	(1 << 25)
-#define   GEN6_BLITTER_SYNC_STATUS			(1 << 24)
-#define   GEN6_BLITTER_USER_INTERRUPT			(1 << 22)
-
 #define GEN6_BLITTER_ECOSKPD	0x221d0
 #define   GEN6_BLITTER_LOCK_SHIFT			16
 #define   GEN6_BLITTER_FBC_NOTIFY			(1<<3)
@@ -744,9 +857,52 @@
 #define   GEN6_BSD_SLEEP_INDICATOR	(1 << 3)
 #define   GEN6_BSD_GO_INDICATOR		(1 << 4)
 
-#define GEN6_BSD_HWSTAM			0x12098
-#define GEN6_BSD_IMR			0x120a8
-#define   GEN6_BSD_USER_INTERRUPT	(1 << 12)
+/* On modern GEN architectures interrupt control consists of two sets
+ * of registers. The first set pertains to the ring generating the
+ * interrupt. The second control is for the functional block generating the
+ * interrupt. These are PM, GT, DE, etc.
+ *
+ * Luckily *knocks on wood* all the ring interrupt bits match up with the
+ * GT interrupt bits, so we don't need to duplicate the defines.
+ *
+ * These defines should cover us well from SNB->HSW with minor exceptions
+ * it can also work on ILK.
+ */
+#define GT_BLT_FLUSHDW_NOTIFY_INTERRUPT		(1 << 26)
+#define GT_BLT_CS_ERROR_INTERRUPT		(1 << 25)
+#define GT_BLT_USER_INTERRUPT			(1 << 22)
+#define GT_BSD_CS_ERROR_INTERRUPT		(1 << 15)
+#define GT_BSD_USER_INTERRUPT			(1 << 12)
+#define GT_RENDER_L3_PARITY_ERROR_INTERRUPT	(1 <<  5) /* !snb */
+#define GT_RENDER_PIPECTL_NOTIFY_INTERRUPT	(1 <<  4)
+#define GT_RENDER_CS_MASTER_ERROR_INTERRUPT	(1 <<  3)
+#define GT_RENDER_SYNC_STATUS_INTERRUPT		(1 <<  2)
+#define GT_RENDER_DEBUG_INTERRUPT		(1 <<  1)
+#define GT_RENDER_USER_INTERRUPT		(1 <<  0)
+
+#define PM_VEBOX_CS_ERROR_INTERRUPT		(1 << 12) /* hsw+ */
+#define PM_VEBOX_USER_INTERRUPT			(1 << 10) /* hsw+ */
+
+/* These are all the "old" interrupts */
+#define ILK_BSD_USER_INTERRUPT				(1<<5)
+#define I915_PIPE_CONTROL_NOTIFY_INTERRUPT		(1<<18)
+#define I915_DISPLAY_PORT_INTERRUPT			(1<<17)
+#define I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT	(1<<15)
+#define I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT	(1<<14) /* p-state */
+#define I915_HWB_OOM_INTERRUPT				(1<<13)
+#define I915_SYNC_STATUS_INTERRUPT			(1<<12)
+#define I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT	(1<<11)
+#define I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT	(1<<10)
+#define I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT	(1<<9)
+#define I915_DISPLAY_PLANE_C_FLIP_PENDING_INTERRUPT	(1<<8)
+#define I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT		(1<<7)
+#define I915_DISPLAY_PIPE_A_EVENT_INTERRUPT		(1<<6)
+#define I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT		(1<<5)
+#define I915_DISPLAY_PIPE_B_EVENT_INTERRUPT		(1<<4)
+#define I915_DEBUG_INTERRUPT				(1<<2)
+#define I915_USER_INTERRUPT				(1<<1)
+#define I915_ASLE_INTERRUPT				(1<<0)
+#define I915_BSD_USER_INTERRUPT				(1 << 25)
 
 #define GEN6_BSD_RNCID			0x12198
 
@@ -807,7 +963,9 @@
 #define   DPFC_CTL_EN		(1<<31)
 #define   DPFC_CTL_PLANEA	(0<<30)
 #define   DPFC_CTL_PLANEB	(1<<30)
+#define   IVB_DPFC_CTL_PLANE_SHIFT	(29)
 #define   DPFC_CTL_FENCE_EN	(1<<29)
+#define   IVB_DPFC_CTL_FENCE_EN	(1<<28)
 #define   DPFC_CTL_PERSISTENT_MODE	(1<<25)
 #define   DPFC_SR_EN		(1<<10)
 #define   DPFC_CTL_LIMIT_1X	(0<<6)
@@ -840,6 +998,7 @@
 #define ILK_DPFC_CHICKEN	0x43224
 #define ILK_FBC_RT_BASE		0x2128
 #define   ILK_FBC_RT_VALID	(1<<0)
+#define   SNB_FBC_FRONT_BUFFER	(1<<1)
 
 #define ILK_DISPLAY_CHICKEN1	0x42000
 #define   ILK_FBCQ_DIS		(1<<22)
@@ -855,6 +1014,25 @@
 #define   SNB_CPU_FENCE_ENABLE	(1<<29)
 #define DPFC_CPU_FENCE_OFFSET	0x100104
 
+/* Framebuffer compression for Ivybridge */
+#define IVB_FBC_RT_BASE			0x7020
+
+#define IPS_CTL		0x43408
+#define   IPS_ENABLE	(1 << 31)
+
+#define MSG_FBC_REND_STATE	0x50380
+#define   FBC_REND_NUKE		(1<<2)
+#define   FBC_REND_CACHE_CLEAN	(1<<1)
+
+#define _HSW_PIPE_SLICE_CHICKEN_1_A	0x420B0
+#define _HSW_PIPE_SLICE_CHICKEN_1_B	0x420B4
+#define   HSW_BYPASS_FBC_QUEUE		(1<<22)
+#define HSW_PIPE_SLICE_CHICKEN_1(pipe) _PIPE(pipe, + \
+					     _HSW_PIPE_SLICE_CHICKEN_1_A, + \
+					     _HSW_PIPE_SLICE_CHICKEN_1_B)
+
+#define HSW_CLKGATE_DISABLE_PART_1	0x46500
+#define   HSW_DPFC_GATING_DISABLE	(1<<23)
 
 /*
  * GPIO regs
@@ -963,7 +1141,10 @@
 #define   DPLL_FPA01_P1_POST_DIV_MASK	0x00ff0000 /* i915 */
 #define   DPLL_FPA01_P1_POST_DIV_MASK_PINEVIEW	0x00ff8000 /* Pineview */
 #define   DPLL_LOCK_VLV			(1<<15)
+#define   DPLL_INTEGRATED_CRI_CLK_VLV	(1<<14)
 #define   DPLL_INTEGRATED_CLOCK_VLV	(1<<13)
+#define   DPLL_PORTC_READY_MASK		(0xf << 4)
+#define   DPLL_PORTB_READY_MASK		(0xf)
 
 #define   DPLL_FPA01_P1_POST_DIV_MASK_I830	0x001f0000
 /*
@@ -1073,7 +1254,7 @@
 #define  DSTATE_PLL_D3_OFF			(1<<3)
 #define  DSTATE_GFX_CLOCK_GATING		(1<<1)
 #define  DSTATE_DOT_CLOCK_GATING		(1<<0)
-#define DSPCLK_GATE_D		0x6200
+#define DSPCLK_GATE_D	(dev_priv->info->display_mmio_offset + 0x6200)
 # define DPUNIT_B_CLOCK_GATE_DISABLE		(1 << 30) /* 965 */
 # define VSUNIT_CLOCK_GATE_DISABLE		(1 << 29) /* 965 */
 # define VRHUNIT_CLOCK_GATE_DISABLE		(1 << 28) /* 965 */
@@ -1186,6 +1367,8 @@
 #define FW_BLC_SELF_VLV		(VLV_DISPLAY_BASE + 0x6500)
 #define  FW_CSPWRDWNEN		(1<<15)
 
+#define MI_ARB_VLV		(VLV_DISPLAY_BASE + 0x6504)
+
 /*
  * Palette regs
  */
@@ -1535,14 +1718,13 @@
 					 GEN7_CXT_EXTENDED_SIZE(ctx_reg) + \
 					 GEN7_CXT_GT1_SIZE(ctx_reg) + \
 					 GEN7_CXT_VFSTATE_SIZE(ctx_reg))
-#define HSW_CXT_POWER_SIZE(ctx_reg)	((ctx_reg >> 26) & 0x3f)
-#define HSW_CXT_RING_SIZE(ctx_reg)	((ctx_reg >> 23) & 0x7)
-#define HSW_CXT_RENDER_SIZE(ctx_reg)	((ctx_reg >> 15) & 0xff)
-#define HSW_CXT_TOTAL_SIZE(ctx_reg)	(HSW_CXT_POWER_SIZE(ctx_reg) + \
-					 HSW_CXT_RING_SIZE(ctx_reg) + \
-					 HSW_CXT_RENDER_SIZE(ctx_reg) + \
-					 GEN7_CXT_VFSTATE_SIZE(ctx_reg))
-
+/* Haswell does have the CXT_SIZE register however it does not appear to be
+ * valid. Now, docs explain in dwords what is in the context object. The full
+ * size is 70720 bytes, however, the power context and execlist context will
+ * never be saved (power context is stored elsewhere, and execlists don't work
+ * on HSW) - so the final size is 66944 bytes, which rounds to 17 pages.
+ */
+#define HSW_CXT_TOTAL_SIZE		(17 * PAGE_SIZE)
 
 /*
  * Overlay regs
@@ -1691,6 +1873,12 @@
 /* SDVO is different across gen3/4 */
 #define   SDVOC_HOTPLUG_INT_STATUS_G4X		(1 << 3)
 #define   SDVOB_HOTPLUG_INT_STATUS_G4X		(1 << 2)
+/*
+ * Bspec seems to be seriously misleaded about the SDVO hpd bits on i965g/gm,
+ * since reality corrobates that they're the same as on gen3. But keep these
+ * bits here (and the comment!) to help any other lost wanderers back onto the
+ * right tracks.
+ */
 #define   SDVOC_HOTPLUG_INT_STATUS_I965		(3 << 4)
 #define   SDVOB_HOTPLUG_INT_STATUS_I965		(3 << 2)
 #define   SDVOC_HOTPLUG_INT_STATUS_I915		(1 << 7)
@@ -1702,13 +1890,6 @@
 						 PORTC_HOTPLUG_INT_STATUS | \
 						 PORTD_HOTPLUG_INT_STATUS)
 
-#define HOTPLUG_INT_STATUS_I965			(CRT_HOTPLUG_INT_STATUS | \
-						 SDVOB_HOTPLUG_INT_STATUS_I965 | \
-						 SDVOC_HOTPLUG_INT_STATUS_I965 | \
-						 PORTB_HOTPLUG_INT_STATUS | \
-						 PORTC_HOTPLUG_INT_STATUS | \
-						 PORTD_HOTPLUG_INT_STATUS)
-
 #define HOTPLUG_INT_STATUS_I915			(CRT_HOTPLUG_INT_STATUS | \
 						 SDVOB_HOTPLUG_INT_STATUS_I915 | \
 						 SDVOC_HOTPLUG_INT_STATUS_I915 | \
@@ -1967,6 +2148,10 @@
 #define   BLM_PIPE_A			(0 << 29)
 #define   BLM_PIPE_B			(1 << 29)
 #define   BLM_PIPE_C			(2 << 29) /* ivb + */
+#define   BLM_TRANSCODER_A		BLM_PIPE_A /* hsw */
+#define   BLM_TRANSCODER_B		BLM_PIPE_B
+#define   BLM_TRANSCODER_C		BLM_PIPE_C
+#define   BLM_TRANSCODER_EDP		(3 << 29)
 #define   BLM_PIPE(pipe)		((pipe) << 29)
 #define   BLM_POLARITY_I965		(1 << 28) /* gen4 only */
 #define   BLM_PHASE_IN_INTERUPT_STATUS	(1 << 26)
@@ -2540,9 +2725,7 @@
 #define   DP_PRE_EMPHASIS_SHIFT		22
 
 /* How many wires to use. I guess 3 was too hard */
-#define   DP_PORT_WIDTH_1		(0 << 19)
-#define   DP_PORT_WIDTH_2		(1 << 19)
-#define   DP_PORT_WIDTH_4		(3 << 19)
+#define   DP_PORT_WIDTH(width)		(((width) - 1) << 19)
 #define   DP_PORT_WIDTH_MASK		(7 << 19)
 
 /* Mystic DPCD version 1.1 special mode */
@@ -2646,18 +2829,20 @@
  * which is after the LUTs, so we want the bytes for our color format.
  * For our current usage, this is always 3, one byte for R, G and B.
  */
-#define _PIPEA_GMCH_DATA_M			0x70050
-#define _PIPEB_GMCH_DATA_M			0x71050
+#define _PIPEA_DATA_M_G4X	0x70050
+#define _PIPEB_DATA_M_G4X	0x71050
 
 /* Transfer unit size for display port - 1, default is 0x3f (for TU size 64) */
 #define  TU_SIZE(x)             (((x)-1) << 25) /* default size 64 */
+#define  TU_SIZE_SHIFT		25
 #define  TU_SIZE_MASK           (0x3f << 25)
 
 #define  DATA_LINK_M_N_MASK	(0xffffff)
 #define  DATA_LINK_N_MAX	(0x800000)
 
-#define _PIPEA_GMCH_DATA_N			0x70054
-#define _PIPEB_GMCH_DATA_N			0x71054
+#define _PIPEA_DATA_N_G4X	0x70054
+#define _PIPEB_DATA_N_G4X	0x71054
+#define   PIPE_GMCH_DATA_N_MASK			(0xffffff)
 
 /*
  * Computing Link M and N values for the Display Port link
@@ -2670,16 +2855,18 @@
  * Attributes and VB-ID.
  */
 
-#define _PIPEA_DP_LINK_M				0x70060
-#define _PIPEB_DP_LINK_M				0x71060
+#define _PIPEA_LINK_M_G4X	0x70060
+#define _PIPEB_LINK_M_G4X	0x71060
+#define   PIPEA_DP_LINK_M_MASK			(0xffffff)
 
-#define _PIPEA_DP_LINK_N				0x70064
-#define _PIPEB_DP_LINK_N				0x71064
+#define _PIPEA_LINK_N_G4X	0x70064
+#define _PIPEB_LINK_N_G4X	0x71064
+#define   PIPEA_DP_LINK_N_MASK			(0xffffff)
 
-#define PIPE_GMCH_DATA_M(pipe) _PIPE(pipe, _PIPEA_GMCH_DATA_M, _PIPEB_GMCH_DATA_M)
-#define PIPE_GMCH_DATA_N(pipe) _PIPE(pipe, _PIPEA_GMCH_DATA_N, _PIPEB_GMCH_DATA_N)
-#define PIPE_DP_LINK_M(pipe) _PIPE(pipe, _PIPEA_DP_LINK_M, _PIPEB_DP_LINK_M)
-#define PIPE_DP_LINK_N(pipe) _PIPE(pipe, _PIPEA_DP_LINK_N, _PIPEB_DP_LINK_N)
+#define PIPE_DATA_M_G4X(pipe) _PIPE(pipe, _PIPEA_DATA_M_G4X, _PIPEB_DATA_M_G4X)
+#define PIPE_DATA_N_G4X(pipe) _PIPE(pipe, _PIPEA_DATA_N_G4X, _PIPEB_DATA_N_G4X)
+#define PIPE_LINK_M_G4X(pipe) _PIPE(pipe, _PIPEA_LINK_M_G4X, _PIPEB_LINK_M_G4X)
+#define PIPE_LINK_N_G4X(pipe) _PIPE(pipe, _PIPEA_LINK_N_G4X, _PIPEB_LINK_N_G4X)
 
 /* Display & cursor control */
 
@@ -2715,6 +2902,7 @@
 #define   PIPECONF_INTERLACED_ILK		(3 << 21)
 #define   PIPECONF_INTERLACED_DBL_ILK		(4 << 21) /* ilk/snb only */
 #define   PIPECONF_PFIT_PF_INTERLACED_DBL_ILK	(5 << 21) /* ilk/snb only */
+#define   PIPECONF_INTERLACE_MODE_MASK		(7 << 21)
 #define   PIPECONF_CXSR_DOWNCLOCK	(1<<16)
 #define   PIPECONF_COLOR_RANGE_SELECT	(1 << 13)
 #define   PIPECONF_BPC_MASK	(0x7 << 5)
@@ -2915,6 +3103,10 @@
 #define WM3S_LP_IVB		0x45128
 #define  WM1S_LP_EN		(1<<31)
 
+#define HSW_WM_LP_VAL(lat, fbc, pri, cur) \
+	(WM3_LP_EN | ((lat) << WM1_LP_LATENCY_SHIFT) | \
+	 ((fbc) << WM1_LP_FBC_SHIFT) | ((pri) << WM1_LP_SR_SHIFT) | (cur))
+
 /* Memory latency timer register */
 #define MLTR_ILK		0x11222
 #define  MLTR_WM1_SHIFT		0
@@ -3294,7 +3486,7 @@
 #define SPRGAMC(pipe) _PIPE(pipe, _SPRA_GAMC, _SPRB_GAMC)
 #define SPRSURFLIVE(pipe) _PIPE(pipe, _SPRA_SURFLIVE, _SPRB_SURFLIVE)
 
-#define _SPACNTR		0x72180
+#define _SPACNTR		(VLV_DISPLAY_BASE + 0x72180)
 #define   SP_ENABLE			(1<<31)
 #define   SP_GEAMMA_ENABLE		(1<<30)
 #define   SP_PIXFORMAT_MASK		(0xf<<26)
@@ -3313,30 +3505,30 @@
 #define   SP_YUV_ORDER_YVYU		(2<<16)
 #define   SP_YUV_ORDER_VYUY		(3<<16)
 #define   SP_TILED			(1<<10)
-#define _SPALINOFF		0x72184
-#define _SPASTRIDE		0x72188
-#define _SPAPOS			0x7218c
-#define _SPASIZE		0x72190
-#define _SPAKEYMINVAL		0x72194
-#define _SPAKEYMSK		0x72198
-#define _SPASURF		0x7219c
-#define _SPAKEYMAXVAL		0x721a0
-#define _SPATILEOFF		0x721a4
-#define _SPACONSTALPHA		0x721a8
-#define _SPAGAMC		0x721f4
-
-#define _SPBCNTR		0x72280
-#define _SPBLINOFF		0x72284
-#define _SPBSTRIDE		0x72288
-#define _SPBPOS			0x7228c
-#define _SPBSIZE		0x72290
-#define _SPBKEYMINVAL		0x72294
-#define _SPBKEYMSK		0x72298
-#define _SPBSURF		0x7229c
-#define _SPBKEYMAXVAL		0x722a0
-#define _SPBTILEOFF		0x722a4
-#define _SPBCONSTALPHA		0x722a8
-#define _SPBGAMC		0x722f4
+#define _SPALINOFF		(VLV_DISPLAY_BASE + 0x72184)
+#define _SPASTRIDE		(VLV_DISPLAY_BASE + 0x72188)
+#define _SPAPOS			(VLV_DISPLAY_BASE + 0x7218c)
+#define _SPASIZE		(VLV_DISPLAY_BASE + 0x72190)
+#define _SPAKEYMINVAL		(VLV_DISPLAY_BASE + 0x72194)
+#define _SPAKEYMSK		(VLV_DISPLAY_BASE + 0x72198)
+#define _SPASURF		(VLV_DISPLAY_BASE + 0x7219c)
+#define _SPAKEYMAXVAL		(VLV_DISPLAY_BASE + 0x721a0)
+#define _SPATILEOFF		(VLV_DISPLAY_BASE + 0x721a4)
+#define _SPACONSTALPHA		(VLV_DISPLAY_BASE + 0x721a8)
+#define _SPAGAMC		(VLV_DISPLAY_BASE + 0x721f4)
+
+#define _SPBCNTR		(VLV_DISPLAY_BASE + 0x72280)
+#define _SPBLINOFF		(VLV_DISPLAY_BASE + 0x72284)
+#define _SPBSTRIDE		(VLV_DISPLAY_BASE + 0x72288)
+#define _SPBPOS			(VLV_DISPLAY_BASE + 0x7228c)
+#define _SPBSIZE		(VLV_DISPLAY_BASE + 0x72290)
+#define _SPBKEYMINVAL		(VLV_DISPLAY_BASE + 0x72294)
+#define _SPBKEYMSK		(VLV_DISPLAY_BASE + 0x72298)
+#define _SPBSURF		(VLV_DISPLAY_BASE + 0x7229c)
+#define _SPBKEYMAXVAL		(VLV_DISPLAY_BASE + 0x722a0)
+#define _SPBTILEOFF		(VLV_DISPLAY_BASE + 0x722a4)
+#define _SPBCONSTALPHA		(VLV_DISPLAY_BASE + 0x722a8)
+#define _SPBGAMC		(VLV_DISPLAY_BASE + 0x722f4)
 
 #define SPCNTR(pipe, plane) _PIPE(pipe * 2 + plane, _SPACNTR, _SPBCNTR)
 #define SPLINOFF(pipe, plane) _PIPE(pipe * 2 + plane, _SPALINOFF, _SPBLINOFF)
@@ -3474,6 +3666,15 @@
 #define _LGC_PALETTE_B           0x4a800
 #define LGC_PALETTE(pipe) _PIPE(pipe, _LGC_PALETTE_A, _LGC_PALETTE_B)
 
+#define _GAMMA_MODE_A		0x4a480
+#define _GAMMA_MODE_B		0x4ac80
+#define GAMMA_MODE(pipe) _PIPE(pipe, _GAMMA_MODE_A, _GAMMA_MODE_B)
+#define GAMMA_MODE_MODE_MASK	(3 << 0)
+#define GAMMA_MODE_MODE_8BIT	(0 << 0)
+#define GAMMA_MODE_MODE_10BIT	(1 << 0)
+#define GAMMA_MODE_MODE_12BIT	(2 << 0)
+#define GAMMA_MODE_MODE_SPLIT	(3 << 0)
+
 /* interrupts */
 #define DE_MASTER_IRQ_CONTROL   (1 << 31)
 #define DE_SPRITEB_FLIP_DONE    (1 << 29)
@@ -3502,7 +3703,7 @@
 #define DE_PIPEA_FIFO_UNDERRUN  (1 << 0)
 
 /* More Ivybridge lolz */
-#define DE_ERR_DEBUG_IVB		(1<<30)
+#define DE_ERR_INT_IVB			(1<<30)
 #define DE_GSE_IVB			(1<<29)
 #define DE_PCH_EVENT_IVB		(1<<28)
 #define DE_DP_A_HOTPLUG_IVB		(1<<27)
@@ -3525,21 +3726,6 @@
 #define DEIIR   0x44008
 #define DEIER   0x4400c
 
-/* GT interrupt.
- * Note that for gen6+ the ring-specific interrupt bits do alias with the
- * corresponding bits in the per-ring interrupt control registers. */
-#define GT_GEN6_BLT_FLUSHDW_NOTIFY_INTERRUPT	(1 << 26)
-#define GT_GEN6_BLT_CS_ERROR_INTERRUPT		(1 << 25)
-#define GT_GEN6_BLT_USER_INTERRUPT		(1 << 22)
-#define GT_GEN6_BSD_CS_ERROR_INTERRUPT		(1 << 15)
-#define GT_GEN6_BSD_USER_INTERRUPT		(1 << 12)
-#define GT_BSD_USER_INTERRUPT			(1 << 5) /* ilk only */
-#define GT_GEN7_L3_PARITY_ERROR_INTERRUPT	(1 << 5)
-#define GT_PIPE_NOTIFY				(1 << 4)
-#define GT_RENDER_CS_ERROR_INTERRUPT		(1 << 3)
-#define GT_SYNC_STATUS				(1 << 2)
-#define GT_USER_INTERRUPT			(1 << 0)
-
 #define GTISR   0x44010
 #define GTIMR   0x44014
 #define GTIIR   0x44018
@@ -3569,6 +3755,9 @@
 # define CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE	(1 << 5)
 # define CHICKEN3_DGMG_DONE_FIX_DISABLE		(1 << 2)
 
+#define CHICKEN_PAR1_1		0x42080
+#define  FORCE_ARB_IDLE_PLANES	(1 << 14)
+
 #define DISP_ARB_CTL	0x45000
 #define  DISP_TILE_SURFACE_SWIZZLING	(1<<13)
 #define  DISP_FBC_WM_DIS		(1<<15)
@@ -3661,6 +3850,7 @@
 				 SDE_PORTC_HOTPLUG_CPT |	\
 				 SDE_PORTB_HOTPLUG_CPT)
 #define SDE_GMBUS_CPT		(1 << 17)
+#define SDE_ERROR_CPT		(1 << 16)
 #define SDE_AUDIO_CP_REQ_C_CPT	(1 << 10)
 #define SDE_AUDIO_CP_CHG_C_CPT	(1 << 9)
 #define SDE_FDI_RXC_CPT		(1 << 8)
@@ -3685,6 +3875,12 @@
 #define SDEIIR  0xc4008
 #define SDEIER  0xc400c
 
+#define SERR_INT			0xc4040
+#define  SERR_INT_POISON		(1<<31)
+#define  SERR_INT_TRANS_C_FIFO_UNDERRUN	(1<<6)
+#define  SERR_INT_TRANS_B_FIFO_UNDERRUN	(1<<3)
+#define  SERR_INT_TRANS_A_FIFO_UNDERRUN	(1<<0)
+
 /* digital port hotplug */
 #define PCH_PORT_HOTPLUG        0xc4030		/* SHOTPLUG_CTL */
 #define PORTD_HOTPLUG_ENABLE            (1 << 20)
@@ -3734,15 +3930,15 @@
 
 #define _PCH_DPLL_A              0xc6014
 #define _PCH_DPLL_B              0xc6018
-#define _PCH_DPLL(pll) (pll == 0 ? _PCH_DPLL_A : _PCH_DPLL_B)
+#define PCH_DPLL(pll) (pll == 0 ? _PCH_DPLL_A : _PCH_DPLL_B)
 
 #define _PCH_FPA0                0xc6040
 #define  FP_CB_TUNE		(0x3<<22)
 #define _PCH_FPA1                0xc6044
 #define _PCH_FPB0                0xc6048
 #define _PCH_FPB1                0xc604c
-#define _PCH_FP0(pll) (pll == 0 ? _PCH_FPA0 : _PCH_FPB0)
-#define _PCH_FP1(pll) (pll == 0 ? _PCH_FPA1 : _PCH_FPB1)
+#define PCH_FP0(pll) (pll == 0 ? _PCH_FPA0 : _PCH_FPB0)
+#define PCH_FP1(pll) (pll == 0 ? _PCH_FPA1 : _PCH_FPB1)
 
 #define PCH_DPLL_TEST           0xc606c
 
@@ -3782,46 +3978,40 @@
 #define PCH_SSC4_AUX_PARMS      0xc6214
 
 #define PCH_DPLL_SEL		0xc7000
-#define  TRANSA_DPLL_ENABLE	(1<<3)
-#define	 TRANSA_DPLLB_SEL	(1<<0)
-#define	 TRANSA_DPLLA_SEL	0
-#define  TRANSB_DPLL_ENABLE	(1<<7)
-#define	 TRANSB_DPLLB_SEL	(1<<4)
-#define	 TRANSB_DPLLA_SEL	(0)
-#define  TRANSC_DPLL_ENABLE	(1<<11)
-#define	 TRANSC_DPLLB_SEL	(1<<8)
-#define	 TRANSC_DPLLA_SEL	(0)
+#define	 TRANS_DPLLB_SEL(pipe)		(1 << (pipe * 4))
+#define	 TRANS_DPLLA_SEL(pipe)		0
+#define  TRANS_DPLL_ENABLE(pipe)	(1 << (pipe * 4 + 3))
 
 /* transcoder */
 
-#define _TRANS_HTOTAL_A          0xe0000
-#define  TRANS_HTOTAL_SHIFT     16
-#define  TRANS_HACTIVE_SHIFT    0
-#define _TRANS_HBLANK_A          0xe0004
-#define  TRANS_HBLANK_END_SHIFT 16
-#define  TRANS_HBLANK_START_SHIFT 0
-#define _TRANS_HSYNC_A           0xe0008
-#define  TRANS_HSYNC_END_SHIFT  16
-#define  TRANS_HSYNC_START_SHIFT 0
-#define _TRANS_VTOTAL_A          0xe000c
-#define  TRANS_VTOTAL_SHIFT     16
-#define  TRANS_VACTIVE_SHIFT    0
-#define _TRANS_VBLANK_A          0xe0010
-#define  TRANS_VBLANK_END_SHIFT 16
-#define  TRANS_VBLANK_START_SHIFT 0
-#define _TRANS_VSYNC_A           0xe0014
-#define  TRANS_VSYNC_END_SHIFT  16
-#define  TRANS_VSYNC_START_SHIFT 0
-#define _TRANS_VSYNCSHIFT_A	0xe0028
-
-#define _TRANSA_DATA_M1          0xe0030
-#define _TRANSA_DATA_N1          0xe0034
-#define _TRANSA_DATA_M2          0xe0038
-#define _TRANSA_DATA_N2          0xe003c
-#define _TRANSA_DP_LINK_M1       0xe0040
-#define _TRANSA_DP_LINK_N1       0xe0044
-#define _TRANSA_DP_LINK_M2       0xe0048
-#define _TRANSA_DP_LINK_N2       0xe004c
+#define _PCH_TRANS_HTOTAL_A		0xe0000
+#define  TRANS_HTOTAL_SHIFT		16
+#define  TRANS_HACTIVE_SHIFT		0
+#define _PCH_TRANS_HBLANK_A		0xe0004
+#define  TRANS_HBLANK_END_SHIFT		16
+#define  TRANS_HBLANK_START_SHIFT	0
+#define _PCH_TRANS_HSYNC_A		0xe0008
+#define  TRANS_HSYNC_END_SHIFT		16
+#define  TRANS_HSYNC_START_SHIFT	0
+#define _PCH_TRANS_VTOTAL_A		0xe000c
+#define  TRANS_VTOTAL_SHIFT		16
+#define  TRANS_VACTIVE_SHIFT		0
+#define _PCH_TRANS_VBLANK_A		0xe0010
+#define  TRANS_VBLANK_END_SHIFT		16
+#define  TRANS_VBLANK_START_SHIFT	0
+#define _PCH_TRANS_VSYNC_A		0xe0014
+#define  TRANS_VSYNC_END_SHIFT	 	16
+#define  TRANS_VSYNC_START_SHIFT	0
+#define _PCH_TRANS_VSYNCSHIFT_A		0xe0028
+
+#define _PCH_TRANSA_DATA_M1	0xe0030
+#define _PCH_TRANSA_DATA_N1	0xe0034
+#define _PCH_TRANSA_DATA_M2	0xe0038
+#define _PCH_TRANSA_DATA_N2	0xe003c
+#define _PCH_TRANSA_LINK_M1	0xe0040
+#define _PCH_TRANSA_LINK_N1	0xe0044
+#define _PCH_TRANSA_LINK_M2	0xe0048
+#define _PCH_TRANSA_LINK_N2	0xe004c
 
 /* Per-transcoder DIP controls */
 
@@ -3890,44 +4080,45 @@
 #define HSW_TVIDEO_DIP_VSC_DATA(trans) \
 	 _TRANSCODER(trans, HSW_VIDEO_DIP_VSC_DATA_A, HSW_VIDEO_DIP_VSC_DATA_B)
 
-#define _TRANS_HTOTAL_B          0xe1000
-#define _TRANS_HBLANK_B          0xe1004
-#define _TRANS_HSYNC_B           0xe1008
-#define _TRANS_VTOTAL_B          0xe100c
-#define _TRANS_VBLANK_B          0xe1010
-#define _TRANS_VSYNC_B           0xe1014
-#define _TRANS_VSYNCSHIFT_B	 0xe1028
-
-#define TRANS_HTOTAL(pipe) _PIPE(pipe, _TRANS_HTOTAL_A, _TRANS_HTOTAL_B)
-#define TRANS_HBLANK(pipe) _PIPE(pipe, _TRANS_HBLANK_A, _TRANS_HBLANK_B)
-#define TRANS_HSYNC(pipe) _PIPE(pipe, _TRANS_HSYNC_A, _TRANS_HSYNC_B)
-#define TRANS_VTOTAL(pipe) _PIPE(pipe, _TRANS_VTOTAL_A, _TRANS_VTOTAL_B)
-#define TRANS_VBLANK(pipe) _PIPE(pipe, _TRANS_VBLANK_A, _TRANS_VBLANK_B)
-#define TRANS_VSYNC(pipe) _PIPE(pipe, _TRANS_VSYNC_A, _TRANS_VSYNC_B)
-#define TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _TRANS_VSYNCSHIFT_A, \
-				     _TRANS_VSYNCSHIFT_B)
-
-#define _TRANSB_DATA_M1          0xe1030
-#define _TRANSB_DATA_N1          0xe1034
-#define _TRANSB_DATA_M2          0xe1038
-#define _TRANSB_DATA_N2          0xe103c
-#define _TRANSB_DP_LINK_M1       0xe1040
-#define _TRANSB_DP_LINK_N1       0xe1044
-#define _TRANSB_DP_LINK_M2       0xe1048
-#define _TRANSB_DP_LINK_N2       0xe104c
-
-#define TRANSDATA_M1(pipe) _PIPE(pipe, _TRANSA_DATA_M1, _TRANSB_DATA_M1)
-#define TRANSDATA_N1(pipe) _PIPE(pipe, _TRANSA_DATA_N1, _TRANSB_DATA_N1)
-#define TRANSDATA_M2(pipe) _PIPE(pipe, _TRANSA_DATA_M2, _TRANSB_DATA_M2)
-#define TRANSDATA_N2(pipe) _PIPE(pipe, _TRANSA_DATA_N2, _TRANSB_DATA_N2)
-#define TRANSDPLINK_M1(pipe) _PIPE(pipe, _TRANSA_DP_LINK_M1, _TRANSB_DP_LINK_M1)
-#define TRANSDPLINK_N1(pipe) _PIPE(pipe, _TRANSA_DP_LINK_N1, _TRANSB_DP_LINK_N1)
-#define TRANSDPLINK_M2(pipe) _PIPE(pipe, _TRANSA_DP_LINK_M2, _TRANSB_DP_LINK_M2)
-#define TRANSDPLINK_N2(pipe) _PIPE(pipe, _TRANSA_DP_LINK_N2, _TRANSB_DP_LINK_N2)
-
-#define _TRANSACONF              0xf0008
-#define _TRANSBCONF              0xf1008
-#define TRANSCONF(plane) _PIPE(plane, _TRANSACONF, _TRANSBCONF)
+#define _PCH_TRANS_HTOTAL_B          0xe1000
+#define _PCH_TRANS_HBLANK_B          0xe1004
+#define _PCH_TRANS_HSYNC_B           0xe1008
+#define _PCH_TRANS_VTOTAL_B          0xe100c
+#define _PCH_TRANS_VBLANK_B          0xe1010
+#define _PCH_TRANS_VSYNC_B           0xe1014
+#define _PCH_TRANS_VSYNCSHIFT_B	 0xe1028
+
+#define PCH_TRANS_HTOTAL(pipe) _PIPE(pipe, _PCH_TRANS_HTOTAL_A, _PCH_TRANS_HTOTAL_B)
+#define PCH_TRANS_HBLANK(pipe) _PIPE(pipe, _PCH_TRANS_HBLANK_A, _PCH_TRANS_HBLANK_B)
+#define PCH_TRANS_HSYNC(pipe) _PIPE(pipe, _PCH_TRANS_HSYNC_A, _PCH_TRANS_HSYNC_B)
+#define PCH_TRANS_VTOTAL(pipe) _PIPE(pipe, _PCH_TRANS_VTOTAL_A, _PCH_TRANS_VTOTAL_B)
+#define PCH_TRANS_VBLANK(pipe) _PIPE(pipe, _PCH_TRANS_VBLANK_A, _PCH_TRANS_VBLANK_B)
+#define PCH_TRANS_VSYNC(pipe) _PIPE(pipe, _PCH_TRANS_VSYNC_A, _PCH_TRANS_VSYNC_B)
+#define PCH_TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _PCH_TRANS_VSYNCSHIFT_A, \
+					 _PCH_TRANS_VSYNCSHIFT_B)
+
+#define _PCH_TRANSB_DATA_M1	0xe1030
+#define _PCH_TRANSB_DATA_N1	0xe1034
+#define _PCH_TRANSB_DATA_M2	0xe1038
+#define _PCH_TRANSB_DATA_N2	0xe103c
+#define _PCH_TRANSB_LINK_M1	0xe1040
+#define _PCH_TRANSB_LINK_N1	0xe1044
+#define _PCH_TRANSB_LINK_M2	0xe1048
+#define _PCH_TRANSB_LINK_N2	0xe104c
+
+#define PCH_TRANS_DATA_M1(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_M1, _PCH_TRANSB_DATA_M1)
+#define PCH_TRANS_DATA_N1(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_N1, _PCH_TRANSB_DATA_N1)
+#define PCH_TRANS_DATA_M2(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_M2, _PCH_TRANSB_DATA_M2)
+#define PCH_TRANS_DATA_N2(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_N2, _PCH_TRANSB_DATA_N2)
+#define PCH_TRANS_LINK_M1(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_M1, _PCH_TRANSB_LINK_M1)
+#define PCH_TRANS_LINK_N1(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_N1, _PCH_TRANSB_LINK_N1)
+#define PCH_TRANS_LINK_M2(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_M2, _PCH_TRANSB_LINK_M2)
+#define PCH_TRANS_LINK_N2(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_N2, _PCH_TRANSB_LINK_N2)
+
+#define _PCH_TRANSACONF              0xf0008
+#define _PCH_TRANSBCONF              0xf1008
+#define PCH_TRANSCONF(pipe) _PIPE(pipe, _PCH_TRANSACONF, _PCH_TRANSBCONF)
+#define LPT_TRANSCONF		_PCH_TRANSACONF /* lpt has only one transcoder */
 #define  TRANS_DISABLE          (0<<31)
 #define  TRANS_ENABLE           (1<<31)
 #define  TRANS_STATE_MASK       (1<<30)
@@ -4011,10 +4202,9 @@
 #define  FDI_LINK_TRAIN_600MV_3_5DB_SNB_B	(0x39<<22)
 #define  FDI_LINK_TRAIN_800MV_0DB_SNB_B		(0x38<<22)
 #define  FDI_LINK_TRAIN_VOL_EMP_MASK		(0x3f<<22)
-#define  FDI_DP_PORT_WIDTH_X1           (0<<19)
-#define  FDI_DP_PORT_WIDTH_X2           (1<<19)
-#define  FDI_DP_PORT_WIDTH_X3           (2<<19)
-#define  FDI_DP_PORT_WIDTH_X4           (3<<19)
+#define  FDI_DP_PORT_WIDTH_SHIFT		19
+#define  FDI_DP_PORT_WIDTH_MASK			(7 << FDI_DP_PORT_WIDTH_SHIFT)
+#define  FDI_DP_PORT_WIDTH(width)           (((width) - 1) << FDI_DP_PORT_WIDTH_SHIFT)
 #define  FDI_TX_ENHANCE_FRAME_ENABLE    (1<<18)
 /* Ironlake: hardwired to 1 */
 #define  FDI_TX_PLL_ENABLE              (1<<14)
@@ -4039,7 +4229,6 @@
 /* train, dp width same as FDI_TX */
 #define  FDI_FS_ERRC_ENABLE		(1<<27)
 #define  FDI_FE_ERRC_ENABLE		(1<<26)
-#define  FDI_DP_PORT_WIDTH_X8           (7<<19)
 #define  FDI_RX_POLARITY_REVERSED_LPT	(1<<16)
 #define  FDI_8BPC                       (0<<16)
 #define  FDI_10BPC                      (1<<16)
@@ -4061,9 +4250,6 @@
 #define  FDI_LINK_TRAIN_PATTERN_IDLE_CPT	(2<<8)
 #define  FDI_LINK_TRAIN_NORMAL_CPT		(3<<8)
 #define  FDI_LINK_TRAIN_PATTERN_MASK_CPT	(3<<8)
-/* LPT */
-#define  FDI_PORT_WIDTH_2X_LPT			(1<<19)
-#define  FDI_PORT_WIDTH_1X_LPT			(0<<19)
 
 #define _FDI_RXA_MISC			0xf0010
 #define _FDI_RXB_MISC			0xf1010
@@ -4309,6 +4495,7 @@
 #define   GEN6_RC_CTL_RC6_ENABLE		(1<<18)
 #define   GEN6_RC_CTL_RC1e_ENABLE		(1<<20)
 #define   GEN6_RC_CTL_RC7_ENABLE		(1<<22)
+#define   GEN7_RC_CTL_TO_MODE			(1<<28)
 #define   GEN6_RC_CTL_EI_MODE(x)		((x)<<27)
 #define   GEN6_RC_CTL_HW_ENABLE			(1<<31)
 #define GEN6_RP_DOWN_TIMEOUT			0xA010
@@ -4370,7 +4557,7 @@
 #define  GEN6_PM_RP_DOWN_THRESHOLD		(1<<4)
 #define  GEN6_PM_RP_UP_EI_EXPIRED		(1<<2)
 #define  GEN6_PM_RP_DOWN_EI_EXPIRED		(1<<1)
-#define  GEN6_PM_DEFERRED_EVENTS		(GEN6_PM_RP_UP_THRESHOLD | \
+#define  GEN6_PM_RPS_EVENTS			(GEN6_PM_RP_UP_THRESHOLD | \
 						 GEN6_PM_RP_DOWN_THRESHOLD | \
 						 GEN6_PM_RP_DOWN_TIMEOUT)
 
@@ -4392,20 +4579,6 @@
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
 #define   GEN6_PCODE_FREQ_RING_RATIO_SHIFT	16
 
-#define VLV_IOSF_DOORBELL_REQ			0x182100
-#define   IOSF_DEVFN_SHIFT			24
-#define   IOSF_OPCODE_SHIFT			16
-#define   IOSF_PORT_SHIFT			8
-#define   IOSF_BYTE_ENABLES_SHIFT		4
-#define   IOSF_BAR_SHIFT			1
-#define   IOSF_SB_BUSY				(1<<0)
-#define   IOSF_PORT_PUNIT			0x4
-#define VLV_IOSF_DATA				0x182104
-#define VLV_IOSF_ADDR				0x182108
-
-#define PUNIT_OPCODE_REG_READ			6
-#define PUNIT_OPCODE_REG_WRITE			7
-
 #define GEN6_GT_CORE_STATUS		0x138060
 #define   GEN6_CORE_CPD_STATE_MASK	(7<<4)
 #define   GEN6_RCn_MASK			7
@@ -4602,9 +4775,6 @@
 #define  TRANS_DDI_EDP_INPUT_B_ONOFF	(5<<12)
 #define  TRANS_DDI_EDP_INPUT_C_ONOFF	(6<<12)
 #define  TRANS_DDI_BFI_ENABLE		(1<<4)
-#define  TRANS_DDI_PORT_WIDTH_X1	(0<<1)
-#define  TRANS_DDI_PORT_WIDTH_X2	(1<<1)
-#define  TRANS_DDI_PORT_WIDTH_X4	(3<<1)
 
 /* DisplayPort Transport Control */
 #define DP_TP_CTL_A			0x64040
@@ -4648,9 +4818,7 @@
 #define  DDI_BUF_PORT_REVERSAL			(1<<16)
 #define  DDI_BUF_IS_IDLE			(1<<7)
 #define  DDI_A_4_LANES				(1<<4)
-#define  DDI_PORT_WIDTH_X1			(0<<1)
-#define  DDI_PORT_WIDTH_X2			(1<<1)
-#define  DDI_PORT_WIDTH_X4			(3<<1)
+#define  DDI_PORT_WIDTH(width)			(((width) - 1) << 1)
 #define  DDI_INIT_DISPLAY_DETECTED		(1<<0)
 
 /* DDI Buffer Translations */
@@ -4774,6 +4942,9 @@
 #define  SFUSE_STRAP_DDIC_DETECTED	(1<<1)
 #define  SFUSE_STRAP_DDID_DETECTED	(1<<0)
 
+#define WM_MISC				0x45260
+#define  WM_MISC_DATA_PARTITION_5_6	(1 << 0)
+
 #define WM_DBG				0x45280
 #define  WM_DBG_DISALLOW_MULTIPLE_LP	(1<<0)
 #define  WM_DBG_DISALLOW_MAXFIFO	(1<<1)
@@ -4787,6 +4958,9 @@
 #define _PIPE_A_CSC_COEFF_RV_GV	0x49020
 #define _PIPE_A_CSC_COEFF_BV	0x49024
 #define _PIPE_A_CSC_MODE	0x49028
+#define   CSC_BLACK_SCREEN_OFFSET	(1 << 2)
+#define   CSC_POSITION_BEFORE_GAMMA	(1 << 1)
+#define   CSC_MODE_YUV_TO_RGB		(1 << 0)
 #define _PIPE_A_CSC_PREOFF_HI	0x49030
 #define _PIPE_A_CSC_PREOFF_ME	0x49034
 #define _PIPE_A_CSC_PREOFF_LO	0x49038
@@ -4808,10 +4982,6 @@
 #define _PIPE_B_CSC_POSTOFF_ME	0x49144
 #define _PIPE_B_CSC_POSTOFF_LO	0x49148
 
-#define CSC_BLACK_SCREEN_OFFSET (1 << 2)
-#define CSC_POSITION_BEFORE_GAMMA (1 << 1)
-#define CSC_MODE_YUV_TO_RGB (1 << 0)
-
 #define PIPE_CSC_COEFF_RY_GY(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RY_GY, _PIPE_B_CSC_COEFF_RY_GY)
 #define PIPE_CSC_COEFF_BY(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_BY, _PIPE_B_CSC_COEFF_BY)
 #define PIPE_CSC_COEFF_RU_GU(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RU_GU, _PIPE_B_CSC_COEFF_RU_GU)
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 369b3d8..70db618 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -192,6 +192,7 @@ static void i915_restore_vga(struct drm_device *dev)
 static void i915_save_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
 
 	/* Display arbitration control */
 	if (INTEL_INFO(dev)->gen <= 4)
@@ -202,6 +203,8 @@ static void i915_save_display(struct drm_device *dev)
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		i915_save_display_reg(dev);
 
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+
 	/* LVDS state */
 	if (HAS_PCH_SPLIT(dev)) {
 		dev_priv->regfile.savePP_CONTROL = I915_READ(PCH_PP_CONTROL);
@@ -222,6 +225,8 @@ static void i915_save_display(struct drm_device *dev)
 			dev_priv->regfile.saveLVDS = I915_READ(LVDS);
 	}
 
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev))
 		dev_priv->regfile.savePFIT_CONTROL = I915_READ(PFIT_CONTROL);
 
@@ -257,6 +262,7 @@ static void i915_restore_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 mask = 0xffffffff;
+	unsigned long flags;
 
 	/* Display arbitration */
 	if (INTEL_INFO(dev)->gen <= 4)
@@ -265,6 +271,8 @@ static void i915_restore_display(struct drm_device *dev)
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		i915_restore_display_reg(dev);
 
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+
 	/* LVDS state */
 	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
 		I915_WRITE(BLC_PWM_CTL2, dev_priv->regfile.saveBLC_PWM_CTL2);
@@ -304,6 +312,8 @@ static void i915_restore_display(struct drm_device *dev)
 		I915_WRITE(PP_CONTROL, dev_priv->regfile.savePP_CONTROL);
 	}
 
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	/* only restore FBC info on the platform that supports FBC*/
 	intel_disable_fbc(dev);
 	if (I915_HAS_FBC(dev)) {
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index d5e1890..6875b56 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -212,7 +212,13 @@ static ssize_t gt_cur_freq_mhz_show(struct device *kdev,
 	int ret;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = dev_priv->rps.cur_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev_priv->dev)) {
+		u32 freq;
+		freq = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+		ret = vlv_gpu_freq(dev_priv->mem_freq, (freq >> 8) & 0xff);
+	} else {
+		ret = dev_priv->rps.cur_delay * GT_FREQUENCY_MULTIPLIER;
+	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
@@ -226,7 +232,10 @@ static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute
 	int ret;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev_priv->dev))
+		ret = vlv_gpu_freq(dev_priv->mem_freq, dev_priv->rps.max_delay);
+	else
+		ret = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
@@ -246,16 +255,25 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 	if (ret)
 		return ret;
 
-	val /= GT_FREQUENCY_MULTIPLIER;
-
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	hw_max = dev_priv->rps.hw_max;
-	non_oc_max = (rp_state_cap & 0xff);
-	hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	if (IS_VALLEYVIEW(dev_priv->dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+
+		hw_max = valleyview_rps_max_freq(dev_priv);
+		hw_min = valleyview_rps_min_freq(dev_priv);
+		non_oc_max = hw_max;
+	} else {
+		val /= GT_FREQUENCY_MULTIPLIER;
+
+		rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
+		hw_max = dev_priv->rps.hw_max;
+		non_oc_max = (rp_state_cap & 0xff);
+		hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	}
 
-	if (val < hw_min || val > hw_max || val < dev_priv->rps.min_delay) {
+	if (val < hw_min || val > hw_max ||
+	    val < dev_priv->rps.min_delay) {
 		mutex_unlock(&dev_priv->rps.hw_lock);
 		return -EINVAL;
 	}
@@ -264,8 +282,12 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 		DRM_DEBUG("User requested overclocking to %d\n",
 			  val * GT_FREQUENCY_MULTIPLIER);
 
-	if (dev_priv->rps.cur_delay > val)
-		gen6_set_rps(dev_priv->dev, val);
+	if (dev_priv->rps.cur_delay > val) {
+		if (IS_VALLEYVIEW(dev_priv->dev))
+			valleyview_set_rps(dev_priv->dev, val);
+		else
+			gen6_set_rps(dev_priv->dev, val);
+	}
 
 	dev_priv->rps.max_delay = val;
 
@@ -282,7 +304,10 @@ static ssize_t gt_min_freq_mhz_show(struct device *kdev, struct device_attribute
 	int ret;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev_priv->dev))
+		ret = vlv_gpu_freq(dev_priv->mem_freq, dev_priv->rps.min_delay);
+	else
+		ret = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
@@ -302,21 +327,32 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev,
 	if (ret)
 		return ret;
 
-	val /= GT_FREQUENCY_MULTIPLIER;
-
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	hw_max = dev_priv->rps.hw_max;
-	hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	if (IS_VALLEYVIEW(dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+
+		hw_max = valleyview_rps_max_freq(dev_priv);
+		hw_min = valleyview_rps_min_freq(dev_priv);
+	} else {
+		val /= GT_FREQUENCY_MULTIPLIER;
+
+		rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
+		hw_max = dev_priv->rps.hw_max;
+		hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	}
 
 	if (val < hw_min || val > hw_max || val > dev_priv->rps.max_delay) {
 		mutex_unlock(&dev_priv->rps.hw_lock);
 		return -EINVAL;
 	}
 
-	if (dev_priv->rps.cur_delay < val)
-		gen6_set_rps(dev_priv->dev, val);
+	if (dev_priv->rps.cur_delay < val) {
+		if (IS_VALLEYVIEW(dev))
+			valleyview_set_rps(dev, val);
+		else
+			gen6_set_rps(dev_priv->dev, val);
+	}
 
 	dev_priv->rps.min_delay = val;
 
diff --git a/drivers/gpu/drm/i915/i915_ums.c b/drivers/gpu/drm/i915/i915_ums.c
index 985a097..967da47 100644
--- a/drivers/gpu/drm/i915/i915_ums.c
+++ b/drivers/gpu/drm/i915/i915_ums.c
@@ -41,7 +41,7 @@ static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe)
 		return false;
 
 	if (HAS_PCH_SPLIT(dev))
-		dpll_reg = _PCH_DPLL(pipe);
+		dpll_reg = PCH_DPLL(pipe);
 	else
 		dpll_reg = (pipe == PIPE_A) ? _DPLL_A : _DPLL_B;
 
@@ -148,13 +148,13 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.savePFA_WIN_SZ = I915_READ(_PFA_WIN_SZ);
 		dev_priv->regfile.savePFA_WIN_POS = I915_READ(_PFA_WIN_POS);
 
-		dev_priv->regfile.saveTRANSACONF = I915_READ(_TRANSACONF);
-		dev_priv->regfile.saveTRANS_HTOTAL_A = I915_READ(_TRANS_HTOTAL_A);
-		dev_priv->regfile.saveTRANS_HBLANK_A = I915_READ(_TRANS_HBLANK_A);
-		dev_priv->regfile.saveTRANS_HSYNC_A = I915_READ(_TRANS_HSYNC_A);
-		dev_priv->regfile.saveTRANS_VTOTAL_A = I915_READ(_TRANS_VTOTAL_A);
-		dev_priv->regfile.saveTRANS_VBLANK_A = I915_READ(_TRANS_VBLANK_A);
-		dev_priv->regfile.saveTRANS_VSYNC_A = I915_READ(_TRANS_VSYNC_A);
+		dev_priv->regfile.saveTRANSACONF = I915_READ(_PCH_TRANSACONF);
+		dev_priv->regfile.saveTRANS_HTOTAL_A = I915_READ(_PCH_TRANS_HTOTAL_A);
+		dev_priv->regfile.saveTRANS_HBLANK_A = I915_READ(_PCH_TRANS_HBLANK_A);
+		dev_priv->regfile.saveTRANS_HSYNC_A = I915_READ(_PCH_TRANS_HSYNC_A);
+		dev_priv->regfile.saveTRANS_VTOTAL_A = I915_READ(_PCH_TRANS_VTOTAL_A);
+		dev_priv->regfile.saveTRANS_VBLANK_A = I915_READ(_PCH_TRANS_VBLANK_A);
+		dev_priv->regfile.saveTRANS_VSYNC_A = I915_READ(_PCH_TRANS_VSYNC_A);
 	}
 
 	dev_priv->regfile.saveDSPACNTR = I915_READ(_DSPACNTR);
@@ -205,13 +205,13 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.savePFB_WIN_SZ = I915_READ(_PFB_WIN_SZ);
 		dev_priv->regfile.savePFB_WIN_POS = I915_READ(_PFB_WIN_POS);
 
-		dev_priv->regfile.saveTRANSBCONF = I915_READ(_TRANSBCONF);
-		dev_priv->regfile.saveTRANS_HTOTAL_B = I915_READ(_TRANS_HTOTAL_B);
-		dev_priv->regfile.saveTRANS_HBLANK_B = I915_READ(_TRANS_HBLANK_B);
-		dev_priv->regfile.saveTRANS_HSYNC_B = I915_READ(_TRANS_HSYNC_B);
-		dev_priv->regfile.saveTRANS_VTOTAL_B = I915_READ(_TRANS_VTOTAL_B);
-		dev_priv->regfile.saveTRANS_VBLANK_B = I915_READ(_TRANS_VBLANK_B);
-		dev_priv->regfile.saveTRANS_VSYNC_B = I915_READ(_TRANS_VSYNC_B);
+		dev_priv->regfile.saveTRANSBCONF = I915_READ(_PCH_TRANSBCONF);
+		dev_priv->regfile.saveTRANS_HTOTAL_B = I915_READ(_PCH_TRANS_HTOTAL_B);
+		dev_priv->regfile.saveTRANS_HBLANK_B = I915_READ(_PCH_TRANS_HBLANK_B);
+		dev_priv->regfile.saveTRANS_HSYNC_B = I915_READ(_PCH_TRANS_HSYNC_B);
+		dev_priv->regfile.saveTRANS_VTOTAL_B = I915_READ(_PCH_TRANS_VTOTAL_B);
+		dev_priv->regfile.saveTRANS_VBLANK_B = I915_READ(_PCH_TRANS_VBLANK_B);
+		dev_priv->regfile.saveTRANS_VSYNC_B = I915_READ(_PCH_TRANS_VSYNC_B);
 	}
 
 	dev_priv->regfile.saveDSPBCNTR = I915_READ(_DSPBCNTR);
@@ -259,14 +259,14 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.saveDP_B = I915_READ(DP_B);
 		dev_priv->regfile.saveDP_C = I915_READ(DP_C);
 		dev_priv->regfile.saveDP_D = I915_READ(DP_D);
-		dev_priv->regfile.savePIPEA_GMCH_DATA_M = I915_READ(_PIPEA_GMCH_DATA_M);
-		dev_priv->regfile.savePIPEB_GMCH_DATA_M = I915_READ(_PIPEB_GMCH_DATA_M);
-		dev_priv->regfile.savePIPEA_GMCH_DATA_N = I915_READ(_PIPEA_GMCH_DATA_N);
-		dev_priv->regfile.savePIPEB_GMCH_DATA_N = I915_READ(_PIPEB_GMCH_DATA_N);
-		dev_priv->regfile.savePIPEA_DP_LINK_M = I915_READ(_PIPEA_DP_LINK_M);
-		dev_priv->regfile.savePIPEB_DP_LINK_M = I915_READ(_PIPEB_DP_LINK_M);
-		dev_priv->regfile.savePIPEA_DP_LINK_N = I915_READ(_PIPEA_DP_LINK_N);
-		dev_priv->regfile.savePIPEB_DP_LINK_N = I915_READ(_PIPEB_DP_LINK_N);
+		dev_priv->regfile.savePIPEA_GMCH_DATA_M = I915_READ(_PIPEA_DATA_M_G4X);
+		dev_priv->regfile.savePIPEB_GMCH_DATA_M = I915_READ(_PIPEB_DATA_M_G4X);
+		dev_priv->regfile.savePIPEA_GMCH_DATA_N = I915_READ(_PIPEA_DATA_N_G4X);
+		dev_priv->regfile.savePIPEB_GMCH_DATA_N = I915_READ(_PIPEB_DATA_N_G4X);
+		dev_priv->regfile.savePIPEA_DP_LINK_M = I915_READ(_PIPEA_LINK_M_G4X);
+		dev_priv->regfile.savePIPEB_DP_LINK_M = I915_READ(_PIPEB_LINK_M_G4X);
+		dev_priv->regfile.savePIPEA_DP_LINK_N = I915_READ(_PIPEA_LINK_N_G4X);
+		dev_priv->regfile.savePIPEB_DP_LINK_N = I915_READ(_PIPEB_LINK_N_G4X);
 	}
 	/* FIXME: regfile.save TV & SDVO state */
 
@@ -282,14 +282,14 @@ void i915_restore_display_reg(struct drm_device *dev)
 
 	/* Display port ratios (must be done before clock is set) */
 	if (SUPPORTS_INTEGRATED_DP(dev)) {
-		I915_WRITE(_PIPEA_GMCH_DATA_M, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
-		I915_WRITE(_PIPEB_GMCH_DATA_M, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
-		I915_WRITE(_PIPEA_GMCH_DATA_N, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
-		I915_WRITE(_PIPEB_GMCH_DATA_N, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
-		I915_WRITE(_PIPEA_DP_LINK_M, dev_priv->regfile.savePIPEA_DP_LINK_M);
-		I915_WRITE(_PIPEB_DP_LINK_M, dev_priv->regfile.savePIPEB_DP_LINK_M);
-		I915_WRITE(_PIPEA_DP_LINK_N, dev_priv->regfile.savePIPEA_DP_LINK_N);
-		I915_WRITE(_PIPEB_DP_LINK_N, dev_priv->regfile.savePIPEB_DP_LINK_N);
+		I915_WRITE(_PIPEA_DATA_M_G4X, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
+		I915_WRITE(_PIPEB_DATA_M_G4X, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
+		I915_WRITE(_PIPEA_DATA_N_G4X, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
+		I915_WRITE(_PIPEB_DATA_N_G4X, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
+		I915_WRITE(_PIPEA_LINK_M_G4X, dev_priv->regfile.savePIPEA_DP_LINK_M);
+		I915_WRITE(_PIPEB_LINK_M_G4X, dev_priv->regfile.savePIPEB_DP_LINK_M);
+		I915_WRITE(_PIPEA_LINK_N_G4X, dev_priv->regfile.savePIPEA_DP_LINK_N);
+		I915_WRITE(_PIPEB_LINK_N_G4X, dev_priv->regfile.savePIPEB_DP_LINK_N);
 	}
 
 	/* Fences */
@@ -379,13 +379,13 @@ void i915_restore_display_reg(struct drm_device *dev)
 		I915_WRITE(_PFA_WIN_SZ, dev_priv->regfile.savePFA_WIN_SZ);
 		I915_WRITE(_PFA_WIN_POS, dev_priv->regfile.savePFA_WIN_POS);
 
-		I915_WRITE(_TRANSACONF, dev_priv->regfile.saveTRANSACONF);
-		I915_WRITE(_TRANS_HTOTAL_A, dev_priv->regfile.saveTRANS_HTOTAL_A);
-		I915_WRITE(_TRANS_HBLANK_A, dev_priv->regfile.saveTRANS_HBLANK_A);
-		I915_WRITE(_TRANS_HSYNC_A, dev_priv->regfile.saveTRANS_HSYNC_A);
-		I915_WRITE(_TRANS_VTOTAL_A, dev_priv->regfile.saveTRANS_VTOTAL_A);
-		I915_WRITE(_TRANS_VBLANK_A, dev_priv->regfile.saveTRANS_VBLANK_A);
-		I915_WRITE(_TRANS_VSYNC_A, dev_priv->regfile.saveTRANS_VSYNC_A);
+		I915_WRITE(_PCH_TRANSACONF, dev_priv->regfile.saveTRANSACONF);
+		I915_WRITE(_PCH_TRANS_HTOTAL_A, dev_priv->regfile.saveTRANS_HTOTAL_A);
+		I915_WRITE(_PCH_TRANS_HBLANK_A, dev_priv->regfile.saveTRANS_HBLANK_A);
+		I915_WRITE(_PCH_TRANS_HSYNC_A, dev_priv->regfile.saveTRANS_HSYNC_A);
+		I915_WRITE(_PCH_TRANS_VTOTAL_A, dev_priv->regfile.saveTRANS_VTOTAL_A);
+		I915_WRITE(_PCH_TRANS_VBLANK_A, dev_priv->regfile.saveTRANS_VBLANK_A);
+		I915_WRITE(_PCH_TRANS_VSYNC_A, dev_priv->regfile.saveTRANS_VSYNC_A);
 	}
 
 	/* Restore plane info */
@@ -448,13 +448,13 @@ void i915_restore_display_reg(struct drm_device *dev)
 		I915_WRITE(_PFB_WIN_SZ, dev_priv->regfile.savePFB_WIN_SZ);
 		I915_WRITE(_PFB_WIN_POS, dev_priv->regfile.savePFB_WIN_POS);
 
-		I915_WRITE(_TRANSBCONF, dev_priv->regfile.saveTRANSBCONF);
-		I915_WRITE(_TRANS_HTOTAL_B, dev_priv->regfile.saveTRANS_HTOTAL_B);
-		I915_WRITE(_TRANS_HBLANK_B, dev_priv->regfile.saveTRANS_HBLANK_B);
-		I915_WRITE(_TRANS_HSYNC_B, dev_priv->regfile.saveTRANS_HSYNC_B);
-		I915_WRITE(_TRANS_VTOTAL_B, dev_priv->regfile.saveTRANS_VTOTAL_B);
-		I915_WRITE(_TRANS_VBLANK_B, dev_priv->regfile.saveTRANS_VBLANK_B);
-		I915_WRITE(_TRANS_VSYNC_B, dev_priv->regfile.saveTRANS_VSYNC_B);
+		I915_WRITE(_PCH_TRANSBCONF, dev_priv->regfile.saveTRANSBCONF);
+		I915_WRITE(_PCH_TRANS_HTOTAL_B, dev_priv->regfile.saveTRANS_HTOTAL_B);
+		I915_WRITE(_PCH_TRANS_HBLANK_B, dev_priv->regfile.saveTRANS_HBLANK_B);
+		I915_WRITE(_PCH_TRANS_HSYNC_B, dev_priv->regfile.saveTRANS_HSYNC_B);
+		I915_WRITE(_PCH_TRANS_VTOTAL_B, dev_priv->regfile.saveTRANS_VTOTAL_B);
+		I915_WRITE(_PCH_TRANS_VBLANK_B, dev_priv->regfile.saveTRANS_VBLANK_B);
+		I915_WRITE(_PCH_TRANS_VSYNC_B, dev_priv->regfile.saveTRANS_VSYNC_B);
 	}
 
 	/* Restore plane info */
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index 95070b2..53f2bed 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -212,7 +212,7 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 	if (!lvds_options)
 		return;
 
-	dev_priv->lvds_dither = lvds_options->pixel_dither;
+	dev_priv->vbt.lvds_dither = lvds_options->pixel_dither;
 	if (lvds_options->panel_type == 0xff)
 		return;
 
@@ -226,7 +226,7 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 	if (!lvds_lfp_data_ptrs)
 		return;
 
-	dev_priv->lvds_vbt = 1;
+	dev_priv->vbt.lvds_vbt = 1;
 
 	panel_dvo_timing = get_lvds_dvo_timing(lvds_lfp_data,
 					       lvds_lfp_data_ptrs,
@@ -238,7 +238,7 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 
 	fill_detail_timing_data(panel_fixed_mode, panel_dvo_timing);
 
-	dev_priv->lfp_lvds_vbt_mode = panel_fixed_mode;
+	dev_priv->vbt.lfp_lvds_vbt_mode = panel_fixed_mode;
 
 	DRM_DEBUG_KMS("Found panel mode in BIOS VBT tables:\n");
 	drm_mode_debug_printmodeline(panel_fixed_mode);
@@ -274,9 +274,9 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 		/* check the resolution, just to be sure */
 		if (fp_timing->x_res == panel_fixed_mode->hdisplay &&
 		    fp_timing->y_res == panel_fixed_mode->vdisplay) {
-			dev_priv->bios_lvds_val = fp_timing->lvds_reg_val;
+			dev_priv->vbt.bios_lvds_val = fp_timing->lvds_reg_val;
 			DRM_DEBUG_KMS("VBT initial LVDS value %x\n",
-				      dev_priv->bios_lvds_val);
+				      dev_priv->vbt.bios_lvds_val);
 		}
 	}
 }
@@ -316,7 +316,7 @@ parse_sdvo_panel_data(struct drm_i915_private *dev_priv,
 
 	fill_detail_timing_data(panel_fixed_mode, dvo_timing + index);
 
-	dev_priv->sdvo_lvds_vbt_mode = panel_fixed_mode;
+	dev_priv->vbt.sdvo_lvds_vbt_mode = panel_fixed_mode;
 
 	DRM_DEBUG_KMS("Found SDVO panel mode in BIOS VBT tables:\n");
 	drm_mode_debug_printmodeline(panel_fixed_mode);
@@ -345,20 +345,20 @@ parse_general_features(struct drm_i915_private *dev_priv,
 
 	general = find_section(bdb, BDB_GENERAL_FEATURES);
 	if (general) {
-		dev_priv->int_tv_support = general->int_tv_support;
-		dev_priv->int_crt_support = general->int_crt_support;
-		dev_priv->lvds_use_ssc = general->enable_ssc;
-		dev_priv->lvds_ssc_freq =
+		dev_priv->vbt.int_tv_support = general->int_tv_support;
+		dev_priv->vbt.int_crt_support = general->int_crt_support;
+		dev_priv->vbt.lvds_use_ssc = general->enable_ssc;
+		dev_priv->vbt.lvds_ssc_freq =
 			intel_bios_ssc_frequency(dev, general->ssc_freq);
-		dev_priv->display_clock_mode = general->display_clock_mode;
-		dev_priv->fdi_rx_polarity_inverted = general->fdi_rx_polarity_inverted;
+		dev_priv->vbt.display_clock_mode = general->display_clock_mode;
+		dev_priv->vbt.fdi_rx_polarity_inverted = general->fdi_rx_polarity_inverted;
 		DRM_DEBUG_KMS("BDB_GENERAL_FEATURES int_tv_support %d int_crt_support %d lvds_use_ssc %d lvds_ssc_freq %d display_clock_mode %d fdi_rx_polarity_inverted %d\n",
-			      dev_priv->int_tv_support,
-			      dev_priv->int_crt_support,
-			      dev_priv->lvds_use_ssc,
-			      dev_priv->lvds_ssc_freq,
-			      dev_priv->display_clock_mode,
-			      dev_priv->fdi_rx_polarity_inverted);
+			      dev_priv->vbt.int_tv_support,
+			      dev_priv->vbt.int_crt_support,
+			      dev_priv->vbt.lvds_use_ssc,
+			      dev_priv->vbt.lvds_ssc_freq,
+			      dev_priv->vbt.display_clock_mode,
+			      dev_priv->vbt.fdi_rx_polarity_inverted);
 	}
 }
 
@@ -375,7 +375,7 @@ parse_general_definitions(struct drm_i915_private *dev_priv,
 			int bus_pin = general->crt_ddc_gmbus_pin;
 			DRM_DEBUG_KMS("crt_ddc_bus_pin: %d\n", bus_pin);
 			if (intel_gmbus_is_port_valid(bus_pin))
-				dev_priv->crt_ddc_pin = bus_pin;
+				dev_priv->vbt.crt_ddc_pin = bus_pin;
 		} else {
 			DRM_DEBUG_KMS("BDB_GD too small (%d). Invalid.\n",
 				      block_size);
@@ -486,7 +486,7 @@ parse_driver_features(struct drm_i915_private *dev_priv,
 
 	if (SUPPORTS_EDP(dev) &&
 	    driver->lvds_config == BDB_DRIVER_FEATURE_EDP)
-		dev_priv->edp.support = 1;
+		dev_priv->vbt.edp_support = 1;
 
 	if (driver->dual_frequency)
 		dev_priv->render_reclock_avail = true;
@@ -501,20 +501,20 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	edp = find_section(bdb, BDB_EDP);
 	if (!edp) {
-		if (SUPPORTS_EDP(dev_priv->dev) && dev_priv->edp.support)
+		if (SUPPORTS_EDP(dev_priv->dev) && dev_priv->vbt.edp_support)
 			DRM_DEBUG_KMS("No eDP BDB found but eDP panel supported.\n");
 		return;
 	}
 
 	switch ((edp->color_depth >> (panel_type * 2)) & 3) {
 	case EDP_18BPP:
-		dev_priv->edp.bpp = 18;
+		dev_priv->vbt.edp_bpp = 18;
 		break;
 	case EDP_24BPP:
-		dev_priv->edp.bpp = 24;
+		dev_priv->vbt.edp_bpp = 24;
 		break;
 	case EDP_30BPP:
-		dev_priv->edp.bpp = 30;
+		dev_priv->vbt.edp_bpp = 30;
 		break;
 	}
 
@@ -522,48 +522,48 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 	edp_pps = &edp->power_seqs[panel_type];
 	edp_link_params = &edp->link_params[panel_type];
 
-	dev_priv->edp.pps = *edp_pps;
+	dev_priv->vbt.edp_pps = *edp_pps;
 
-	dev_priv->edp.rate = edp_link_params->rate ? DP_LINK_BW_2_7 :
+	dev_priv->vbt.edp_rate = edp_link_params->rate ? DP_LINK_BW_2_7 :
 		DP_LINK_BW_1_62;
 	switch (edp_link_params->lanes) {
 	case 0:
-		dev_priv->edp.lanes = 1;
+		dev_priv->vbt.edp_lanes = 1;
 		break;
 	case 1:
-		dev_priv->edp.lanes = 2;
+		dev_priv->vbt.edp_lanes = 2;
 		break;
 	case 3:
 	default:
-		dev_priv->edp.lanes = 4;
+		dev_priv->vbt.edp_lanes = 4;
 		break;
 	}
 	switch (edp_link_params->preemphasis) {
 	case 0:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
 		break;
 	case 1:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
 		break;
 	case 2:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
 		break;
 	case 3:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
 		break;
 	}
 	switch (edp_link_params->vswing) {
 	case 0:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_400;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_400;
 		break;
 	case 1:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_600;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_600;
 		break;
 	case 2:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_800;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_800;
 		break;
 	case 3:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_1200;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_1200;
 		break;
 	}
 }
@@ -611,13 +611,13 @@ parse_device_mapping(struct drm_i915_private *dev_priv,
 		DRM_DEBUG_KMS("no child dev is parsed from VBT\n");
 		return;
 	}
-	dev_priv->child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL);
-	if (!dev_priv->child_dev) {
+	dev_priv->vbt.child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL);
+	if (!dev_priv->vbt.child_dev) {
 		DRM_DEBUG_KMS("No memory space for child device\n");
 		return;
 	}
 
-	dev_priv->child_dev_num = count;
+	dev_priv->vbt.child_dev_num = count;
 	count = 0;
 	for (i = 0; i < child_device_num; i++) {
 		p_child = &(p_defs->devices[i]);
@@ -625,7 +625,7 @@ parse_device_mapping(struct drm_i915_private *dev_priv,
 			/* skip the device block if device type is invalid */
 			continue;
 		}
-		child_dev_ptr = dev_priv->child_dev + count;
+		child_dev_ptr = dev_priv->vbt.child_dev + count;
 		count++;
 		memcpy((void *)child_dev_ptr, (void *)p_child,
 					sizeof(*p_child));
@@ -638,23 +638,23 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 
-	dev_priv->crt_ddc_pin = GMBUS_PORT_VGADDC;
+	dev_priv->vbt.crt_ddc_pin = GMBUS_PORT_VGADDC;
 
 	/* LFP panel data */
-	dev_priv->lvds_dither = 1;
-	dev_priv->lvds_vbt = 0;
+	dev_priv->vbt.lvds_dither = 1;
+	dev_priv->vbt.lvds_vbt = 0;
 
 	/* SDVO panel data */
-	dev_priv->sdvo_lvds_vbt_mode = NULL;
+	dev_priv->vbt.sdvo_lvds_vbt_mode = NULL;
 
 	/* general features */
-	dev_priv->int_tv_support = 1;
-	dev_priv->int_crt_support = 1;
+	dev_priv->vbt.int_tv_support = 1;
+	dev_priv->vbt.int_crt_support = 1;
 
 	/* Default to using SSC */
-	dev_priv->lvds_use_ssc = 1;
-	dev_priv->lvds_ssc_freq = intel_bios_ssc_frequency(dev, 1);
-	DRM_DEBUG_KMS("Set default to SSC at %dMHz\n", dev_priv->lvds_ssc_freq);
+	dev_priv->vbt.lvds_use_ssc = 1;
+	dev_priv->vbt.lvds_ssc_freq = intel_bios_ssc_frequency(dev, 1);
+	DRM_DEBUG_KMS("Set default to SSC at %dMHz\n", dev_priv->vbt.lvds_ssc_freq);
 }
 
 static int __init intel_no_opregion_vbt_callback(const struct dmi_system_id *id)
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 58b4a53..3acec8c 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -84,6 +84,28 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_crt_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_crt *crt = intel_encoder_to_crt(encoder);
+	u32 tmp, flags = 0;
+
+	tmp = I915_READ(crt->adpa_reg);
+
+	if (tmp & ADPA_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (tmp & ADPA_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 /* Note: The caller is required to filter out dpms modes not supported by the
  * platform. */
 static void intel_crt_set_dpms(struct intel_encoder *encoder, int mode)
@@ -127,7 +149,7 @@ static void intel_enable_crt(struct intel_encoder *encoder)
 	intel_crt_set_dpms(encoder, crt->connector->base.dpms);
 }
 
-
+/* Special dpms function to support cloning between dvo/sdvo/crt. */
 static void intel_crt_dpms(struct drm_connector *connector, int mode)
 {
 	struct drm_device *dev = connector->dev;
@@ -158,6 +180,8 @@ static void intel_crt_dpms(struct drm_connector *connector, int mode)
 	else
 		encoder->connectors_active = true;
 
+	/* We call connector dpms manually below in case pipe dpms doesn't
+	 * change due to cloning. */
 	if (mode < old_dpms) {
 		/* From off to on, enable the pipe first. */
 		intel_crtc_update_dpms(crtc);
@@ -207,6 +231,10 @@ static bool intel_crt_compute_config(struct intel_encoder *encoder,
 	if (HAS_PCH_SPLIT(dev))
 		pipe_config->has_pch_encoder = true;
 
+	/* LPT FDI RX only supports 8bpc. */
+	if (HAS_PCH_LPT(dev))
+		pipe_config->pipe_bpp = 24;
+
 	return true;
 }
 
@@ -431,7 +459,7 @@ static bool intel_crt_detect_ddc(struct drm_connector *connector)
 
 	BUG_ON(crt->base.type != INTEL_OUTPUT_ANALOG);
 
-	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->crt_ddc_pin);
+	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->vbt.crt_ddc_pin);
 	edid = intel_crt_get_edid(connector, i2c);
 
 	if (edid) {
@@ -637,7 +665,7 @@ static int intel_crt_get_modes(struct drm_connector *connector)
 	int ret;
 	struct i2c_adapter *i2c;
 
-	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->crt_ddc_pin);
+	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->vbt.crt_ddc_pin);
 	ret = intel_crt_ddc_get_modes(connector, i2c);
 	if (ret || !IS_G4X(dev))
 		return ret;
@@ -774,6 +802,7 @@ void intel_crt_init(struct drm_device *dev)
 	crt->base.compute_config = intel_crt_compute_config;
 	crt->base.disable = intel_disable_crt;
 	crt->base.enable = intel_enable_crt;
+	crt->base.get_config = intel_crt_get_config;
 	if (I915_HAS_HOTPLUG(dev))
 		crt->base.hpd_pin = HPD_CRT;
 	if (HAS_DDI(dev))
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index fb961bb..324211a 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -174,6 +174,8 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	 * mode set "sequence for CRT port" document:
 	 * - TP1 to TP2 time with the default value
 	 * - FDI delay to 90h
+	 *
+	 * WaFDIAutoLinkSetTimingOverrride:hsw
 	 */
 	I915_WRITE(_FDI_RXA_MISC, FDI_RX_PWRDN_LANE1_VAL(2) |
 				  FDI_RX_PWRDN_LANE0_VAL(2) |
@@ -181,7 +183,8 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 
 	/* Enable the PCH Receiver FDI PLL */
 	rx_ctl_val = dev_priv->fdi_rx_config | FDI_RX_ENHANCE_FRAME_ENABLE |
-		     FDI_RX_PLL_ENABLE | ((intel_crtc->fdi_lanes - 1) << 19);
+		     FDI_RX_PLL_ENABLE |
+		     FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	I915_WRITE(_FDI_RXA_CTL, rx_ctl_val);
 	POSTING_READ(_FDI_RXA_CTL);
 	udelay(220);
@@ -209,7 +212,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 		 * port reversal bit */
 		I915_WRITE(DDI_BUF_CTL(PORT_E),
 			   DDI_BUF_CTL_ENABLE |
-			   ((intel_crtc->fdi_lanes - 1) << 1) |
+			   ((intel_crtc->config.fdi_lanes - 1) << 1) |
 			   hsw_ddi_buf_ctl_values[i / 2]);
 		POSTING_READ(DDI_BUF_CTL(PORT_E));
 
@@ -278,392 +281,6 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	DRM_ERROR("FDI link training failed!\n");
 }
 
-/* WRPLL clock dividers */
-struct wrpll_tmds_clock {
-	u32 clock;
-	u16 p;		/* Post divider */
-	u16 n2;		/* Feedback divider */
-	u16 r2;		/* Reference divider */
-};
-
-/* Table of matching values for WRPLL clocks programming for each frequency.
- * The code assumes this table is sorted. */
-static const struct wrpll_tmds_clock wrpll_tmds_clock_table[] = {
-	{19750,	38,	25,	18},
-	{20000,	48,	32,	18},
-	{21000,	36,	21,	15},
-	{21912,	42,	29,	17},
-	{22000,	36,	22,	15},
-	{23000,	36,	23,	15},
-	{23500,	40,	40,	23},
-	{23750,	26,	16,	14},
-	{24000,	36,	24,	15},
-	{25000,	36,	25,	15},
-	{25175,	26,	40,	33},
-	{25200,	30,	21,	15},
-	{26000,	36,	26,	15},
-	{27000,	30,	21,	14},
-	{27027,	18,	100,	111},
-	{27500,	30,	29,	19},
-	{28000,	34,	30,	17},
-	{28320,	26,	30,	22},
-	{28322,	32,	42,	25},
-	{28750,	24,	23,	18},
-	{29000,	30,	29,	18},
-	{29750,	32,	30,	17},
-	{30000,	30,	25,	15},
-	{30750,	30,	41,	24},
-	{31000,	30,	31,	18},
-	{31500,	30,	28,	16},
-	{32000,	30,	32,	18},
-	{32500,	28,	32,	19},
-	{33000,	24,	22,	15},
-	{34000,	28,	30,	17},
-	{35000,	26,	32,	19},
-	{35500,	24,	30,	19},
-	{36000,	26,	26,	15},
-	{36750,	26,	46,	26},
-	{37000,	24,	23,	14},
-	{37762,	22,	40,	26},
-	{37800,	20,	21,	15},
-	{38000,	24,	27,	16},
-	{38250,	24,	34,	20},
-	{39000,	24,	26,	15},
-	{40000,	24,	32,	18},
-	{40500,	20,	21,	14},
-	{40541,	22,	147,	89},
-	{40750,	18,	19,	14},
-	{41000,	16,	17,	14},
-	{41500,	22,	44,	26},
-	{41540,	22,	44,	26},
-	{42000,	18,	21,	15},
-	{42500,	22,	45,	26},
-	{43000,	20,	43,	27},
-	{43163,	20,	24,	15},
-	{44000,	18,	22,	15},
-	{44900,	20,	108,	65},
-	{45000,	20,	25,	15},
-	{45250,	20,	52,	31},
-	{46000,	18,	23,	15},
-	{46750,	20,	45,	26},
-	{47000,	20,	40,	23},
-	{48000,	18,	24,	15},
-	{49000,	18,	49,	30},
-	{49500,	16,	22,	15},
-	{50000,	18,	25,	15},
-	{50500,	18,	32,	19},
-	{51000,	18,	34,	20},
-	{52000,	18,	26,	15},
-	{52406,	14,	34,	25},
-	{53000,	16,	22,	14},
-	{54000,	16,	24,	15},
-	{54054,	16,	173,	108},
-	{54500,	14,	24,	17},
-	{55000,	12,	22,	18},
-	{56000,	14,	45,	31},
-	{56250,	16,	25,	15},
-	{56750,	14,	25,	17},
-	{57000,	16,	27,	16},
-	{58000,	16,	43,	25},
-	{58250,	16,	38,	22},
-	{58750,	16,	40,	23},
-	{59000,	14,	26,	17},
-	{59341,	14,	40,	26},
-	{59400,	16,	44,	25},
-	{60000,	16,	32,	18},
-	{60500,	12,	39,	29},
-	{61000,	14,	49,	31},
-	{62000,	14,	37,	23},
-	{62250,	14,	42,	26},
-	{63000,	12,	21,	15},
-	{63500,	14,	28,	17},
-	{64000,	12,	27,	19},
-	{65000,	14,	32,	19},
-	{65250,	12,	29,	20},
-	{65500,	12,	32,	22},
-	{66000,	12,	22,	15},
-	{66667,	14,	38,	22},
-	{66750,	10,	21,	17},
-	{67000,	14,	33,	19},
-	{67750,	14,	58,	33},
-	{68000,	14,	30,	17},
-	{68179,	14,	46,	26},
-	{68250,	14,	46,	26},
-	{69000,	12,	23,	15},
-	{70000,	12,	28,	18},
-	{71000,	12,	30,	19},
-	{72000,	12,	24,	15},
-	{73000,	10,	23,	17},
-	{74000,	12,	23,	14},
-	{74176,	8,	100,	91},
-	{74250,	10,	22,	16},
-	{74481,	12,	43,	26},
-	{74500,	10,	29,	21},
-	{75000,	12,	25,	15},
-	{75250,	10,	39,	28},
-	{76000,	12,	27,	16},
-	{77000,	12,	53,	31},
-	{78000,	12,	26,	15},
-	{78750,	12,	28,	16},
-	{79000,	10,	38,	26},
-	{79500,	10,	28,	19},
-	{80000,	12,	32,	18},
-	{81000,	10,	21,	14},
-	{81081,	6,	100,	111},
-	{81624,	8,	29,	24},
-	{82000,	8,	17,	14},
-	{83000,	10,	40,	26},
-	{83950,	10,	28,	18},
-	{84000,	10,	28,	18},
-	{84750,	6,	16,	17},
-	{85000,	6,	17,	18},
-	{85250,	10,	30,	19},
-	{85750,	10,	27,	17},
-	{86000,	10,	43,	27},
-	{87000,	10,	29,	18},
-	{88000,	10,	44,	27},
-	{88500,	10,	41,	25},
-	{89000,	10,	28,	17},
-	{89012,	6,	90,	91},
-	{89100,	10,	33,	20},
-	{90000,	10,	25,	15},
-	{91000,	10,	32,	19},
-	{92000,	10,	46,	27},
-	{93000,	10,	31,	18},
-	{94000,	10,	40,	23},
-	{94500,	10,	28,	16},
-	{95000,	10,	44,	25},
-	{95654,	10,	39,	22},
-	{95750,	10,	39,	22},
-	{96000,	10,	32,	18},
-	{97000,	8,	23,	16},
-	{97750,	8,	42,	29},
-	{98000,	8,	45,	31},
-	{99000,	8,	22,	15},
-	{99750,	8,	34,	23},
-	{100000,	6,	20,	18},
-	{100500,	6,	19,	17},
-	{101000,	6,	37,	33},
-	{101250,	8,	21,	14},
-	{102000,	6,	17,	15},
-	{102250,	6,	25,	22},
-	{103000,	8,	29,	19},
-	{104000,	8,	37,	24},
-	{105000,	8,	28,	18},
-	{106000,	8,	22,	14},
-	{107000,	8,	46,	29},
-	{107214,	8,	27,	17},
-	{108000,	8,	24,	15},
-	{108108,	8,	173,	108},
-	{109000,	6,	23,	19},
-	{110000,	6,	22,	18},
-	{110013,	6,	22,	18},
-	{110250,	8,	49,	30},
-	{110500,	8,	36,	22},
-	{111000,	8,	23,	14},
-	{111264,	8,	150,	91},
-	{111375,	8,	33,	20},
-	{112000,	8,	63,	38},
-	{112500,	8,	25,	15},
-	{113100,	8,	57,	34},
-	{113309,	8,	42,	25},
-	{114000,	8,	27,	16},
-	{115000,	6,	23,	18},
-	{116000,	8,	43,	25},
-	{117000,	8,	26,	15},
-	{117500,	8,	40,	23},
-	{118000,	6,	38,	29},
-	{119000,	8,	30,	17},
-	{119500,	8,	46,	26},
-	{119651,	8,	39,	22},
-	{120000,	8,	32,	18},
-	{121000,	6,	39,	29},
-	{121250,	6,	31,	23},
-	{121750,	6,	23,	17},
-	{122000,	6,	42,	31},
-	{122614,	6,	30,	22},
-	{123000,	6,	41,	30},
-	{123379,	6,	37,	27},
-	{124000,	6,	51,	37},
-	{125000,	6,	25,	18},
-	{125250,	4,	13,	14},
-	{125750,	4,	27,	29},
-	{126000,	6,	21,	15},
-	{127000,	6,	24,	17},
-	{127250,	6,	41,	29},
-	{128000,	6,	27,	19},
-	{129000,	6,	43,	30},
-	{129859,	4,	25,	26},
-	{130000,	6,	26,	18},
-	{130250,	6,	42,	29},
-	{131000,	6,	32,	22},
-	{131500,	6,	38,	26},
-	{131850,	6,	41,	28},
-	{132000,	6,	22,	15},
-	{132750,	6,	28,	19},
-	{133000,	6,	34,	23},
-	{133330,	6,	37,	25},
-	{134000,	6,	61,	41},
-	{135000,	6,	21,	14},
-	{135250,	6,	167,	111},
-	{136000,	6,	62,	41},
-	{137000,	6,	35,	23},
-	{138000,	6,	23,	15},
-	{138500,	6,	40,	26},
-	{138750,	6,	37,	24},
-	{139000,	6,	34,	22},
-	{139050,	6,	34,	22},
-	{139054,	6,	34,	22},
-	{140000,	6,	28,	18},
-	{141000,	6,	36,	23},
-	{141500,	6,	22,	14},
-	{142000,	6,	30,	19},
-	{143000,	6,	27,	17},
-	{143472,	4,	17,	16},
-	{144000,	6,	24,	15},
-	{145000,	6,	29,	18},
-	{146000,	6,	47,	29},
-	{146250,	6,	26,	16},
-	{147000,	6,	49,	30},
-	{147891,	6,	23,	14},
-	{148000,	6,	23,	14},
-	{148250,	6,	28,	17},
-	{148352,	4,	100,	91},
-	{148500,	6,	33,	20},
-	{149000,	6,	48,	29},
-	{150000,	6,	25,	15},
-	{151000,	4,	19,	17},
-	{152000,	6,	27,	16},
-	{152280,	6,	44,	26},
-	{153000,	6,	34,	20},
-	{154000,	6,	53,	31},
-	{155000,	6,	31,	18},
-	{155250,	6,	50,	29},
-	{155750,	6,	45,	26},
-	{156000,	6,	26,	15},
-	{157000,	6,	61,	35},
-	{157500,	6,	28,	16},
-	{158000,	6,	65,	37},
-	{158250,	6,	44,	25},
-	{159000,	6,	53,	30},
-	{159500,	6,	39,	22},
-	{160000,	6,	32,	18},
-	{161000,	4,	31,	26},
-	{162000,	4,	18,	15},
-	{162162,	4,	131,	109},
-	{162500,	4,	53,	44},
-	{163000,	4,	29,	24},
-	{164000,	4,	17,	14},
-	{165000,	4,	22,	18},
-	{166000,	4,	32,	26},
-	{167000,	4,	26,	21},
-	{168000,	4,	46,	37},
-	{169000,	4,	104,	83},
-	{169128,	4,	64,	51},
-	{169500,	4,	39,	31},
-	{170000,	4,	34,	27},
-	{171000,	4,	19,	15},
-	{172000,	4,	51,	40},
-	{172750,	4,	32,	25},
-	{172800,	4,	32,	25},
-	{173000,	4,	41,	32},
-	{174000,	4,	49,	38},
-	{174787,	4,	22,	17},
-	{175000,	4,	35,	27},
-	{176000,	4,	30,	23},
-	{177000,	4,	38,	29},
-	{178000,	4,	29,	22},
-	{178500,	4,	37,	28},
-	{179000,	4,	53,	40},
-	{179500,	4,	73,	55},
-	{180000,	4,	20,	15},
-	{181000,	4,	55,	41},
-	{182000,	4,	31,	23},
-	{183000,	4,	42,	31},
-	{184000,	4,	30,	22},
-	{184750,	4,	26,	19},
-	{185000,	4,	37,	27},
-	{186000,	4,	51,	37},
-	{187000,	4,	36,	26},
-	{188000,	4,	32,	23},
-	{189000,	4,	21,	15},
-	{190000,	4,	38,	27},
-	{190960,	4,	41,	29},
-	{191000,	4,	41,	29},
-	{192000,	4,	27,	19},
-	{192250,	4,	37,	26},
-	{193000,	4,	20,	14},
-	{193250,	4,	53,	37},
-	{194000,	4,	23,	16},
-	{194208,	4,	23,	16},
-	{195000,	4,	26,	18},
-	{196000,	4,	45,	31},
-	{197000,	4,	35,	24},
-	{197750,	4,	41,	28},
-	{198000,	4,	22,	15},
-	{198500,	4,	25,	17},
-	{199000,	4,	28,	19},
-	{200000,	4,	37,	25},
-	{201000,	4,	61,	41},
-	{202000,	4,	112,	75},
-	{202500,	4,	21,	14},
-	{203000,	4,	146,	97},
-	{204000,	4,	62,	41},
-	{204750,	4,	44,	29},
-	{205000,	4,	38,	25},
-	{206000,	4,	29,	19},
-	{207000,	4,	23,	15},
-	{207500,	4,	40,	26},
-	{208000,	4,	37,	24},
-	{208900,	4,	48,	31},
-	{209000,	4,	48,	31},
-	{209250,	4,	31,	20},
-	{210000,	4,	28,	18},
-	{211000,	4,	25,	16},
-	{212000,	4,	22,	14},
-	{213000,	4,	30,	19},
-	{213750,	4,	38,	24},
-	{214000,	4,	46,	29},
-	{214750,	4,	35,	22},
-	{215000,	4,	43,	27},
-	{216000,	4,	24,	15},
-	{217000,	4,	37,	23},
-	{218000,	4,	42,	26},
-	{218250,	4,	42,	26},
-	{218750,	4,	34,	21},
-	{219000,	4,	47,	29},
-	{220000,	4,	44,	27},
-	{220640,	4,	49,	30},
-	{220750,	4,	36,	22},
-	{221000,	4,	36,	22},
-	{222000,	4,	23,	14},
-	{222525,	4,	28,	17},
-	{222750,	4,	33,	20},
-	{227000,	4,	37,	22},
-	{230250,	4,	29,	17},
-	{233500,	4,	38,	22},
-	{235000,	4,	40,	23},
-	{238000,	4,	30,	17},
-	{241500,	2,	17,	19},
-	{245250,	2,	20,	22},
-	{247750,	2,	22,	24},
-	{253250,	2,	15,	16},
-	{256250,	2,	18,	19},
-	{262500,	2,	31,	32},
-	{267250,	2,	66,	67},
-	{268500,	2,	94,	95},
-	{270000,	2,	14,	14},
-	{272500,	2,	77,	76},
-	{273750,	2,	57,	56},
-	{280750,	2,	24,	23},
-	{281250,	2,	23,	22},
-	{286000,	2,	17,	16},
-	{291750,	2,	26,	24},
-	{296703,	2,	56,	51},
-	{297000,	2,	22,	20},
-	{298000,	2,	21,	19},
-};
-
 static void intel_ddi_mode_set(struct drm_encoder *encoder,
 			       struct drm_display_mode *mode,
 			       struct drm_display_mode *adjusted_mode)
@@ -675,7 +292,7 @@ static void intel_ddi_mode_set(struct drm_encoder *encoder,
 	int pipe = intel_crtc->pipe;
 	int type = intel_encoder->type;
 
-	DRM_DEBUG_KMS("Preparing DDI mode for Haswell on port %c, pipe %c\n",
+	DRM_DEBUG_KMS("Preparing DDI mode on port %c, pipe %c\n",
 		      port_name(port), pipe_name(pipe));
 
 	intel_crtc->eld_vld = false;
@@ -686,22 +303,7 @@ static void intel_ddi_mode_set(struct drm_encoder *encoder,
 
 		intel_dp->DP = intel_dig_port->port_reversal |
 			       DDI_BUF_CTL_ENABLE | DDI_BUF_EMP_400MV_0DB_HSW;
-		switch (intel_dp->lane_count) {
-		case 1:
-			intel_dp->DP |= DDI_PORT_WIDTH_X1;
-			break;
-		case 2:
-			intel_dp->DP |= DDI_PORT_WIDTH_X2;
-			break;
-		case 4:
-			intel_dp->DP |= DDI_PORT_WIDTH_X4;
-			break;
-		default:
-			intel_dp->DP |= DDI_PORT_WIDTH_X4;
-			WARN(1, "Unexpected DP lane count %d\n",
-			     intel_dp->lane_count);
-			break;
-		}
+		intel_dp->DP |= DDI_PORT_WIDTH(intel_dp->lane_count);
 
 		if (intel_dp->has_audio) {
 			DRM_DEBUG_DRIVER("DP audio on pipe %c on DDI\n",
@@ -748,8 +350,8 @@ intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
 	}
 
 	if (num_encoders != 1)
-		WARN(1, "%d encoders on crtc for pipe %d\n", num_encoders,
-		     intel_crtc->pipe);
+		WARN(1, "%d encoders on crtc for pipe %c\n", num_encoders,
+		     pipe_name(intel_crtc->pipe));
 
 	BUG_ON(ret == NULL);
 	return ret;
@@ -802,30 +404,227 @@ void intel_ddi_put_crtc_pll(struct drm_crtc *crtc)
 	intel_crtc->ddi_pll_sel = PORT_CLK_SEL_NONE;
 }
 
-static void intel_ddi_calculate_wrpll(int clock, int *p, int *n2, int *r2)
+#define LC_FREQ 2700
+#define LC_FREQ_2K (LC_FREQ * 2000)
+
+#define P_MIN 2
+#define P_MAX 64
+#define P_INC 2
+
+/* Constraints for PLL good behavior */
+#define REF_MIN 48
+#define REF_MAX 400
+#define VCO_MIN 2400
+#define VCO_MAX 4800
+
+#define ABS_DIFF(a, b) ((a > b) ? (a - b) : (b - a))
+
+struct wrpll_rnp {
+	unsigned p, n2, r2;
+};
+
+static unsigned wrpll_get_budget_for_freq(int clock)
+{
+	unsigned budget;
+
+	switch (clock) {
+	case 25175000:
+	case 25200000:
+	case 27000000:
+	case 27027000:
+	case 37762500:
+	case 37800000:
+	case 40500000:
+	case 40541000:
+	case 54000000:
+	case 54054000:
+	case 59341000:
+	case 59400000:
+	case 72000000:
+	case 74176000:
+	case 74250000:
+	case 81000000:
+	case 81081000:
+	case 89012000:
+	case 89100000:
+	case 108000000:
+	case 108108000:
+	case 111264000:
+	case 111375000:
+	case 148352000:
+	case 148500000:
+	case 162000000:
+	case 162162000:
+	case 222525000:
+	case 222750000:
+	case 296703000:
+	case 297000000:
+		budget = 0;
+		break;
+	case 233500000:
+	case 245250000:
+	case 247750000:
+	case 253250000:
+	case 298000000:
+		budget = 1500;
+		break;
+	case 169128000:
+	case 169500000:
+	case 179500000:
+	case 202000000:
+		budget = 2000;
+		break;
+	case 256250000:
+	case 262500000:
+	case 270000000:
+	case 272500000:
+	case 273750000:
+	case 280750000:
+	case 281250000:
+	case 286000000:
+	case 291750000:
+		budget = 4000;
+		break;
+	case 267250000:
+	case 268500000:
+		budget = 5000;
+		break;
+	default:
+		budget = 1000;
+		break;
+	}
+
+	return budget;
+}
+
+static void wrpll_update_rnp(uint64_t freq2k, unsigned budget,
+			     unsigned r2, unsigned n2, unsigned p,
+			     struct wrpll_rnp *best)
 {
-	u32 i;
+	uint64_t a, b, c, d, diff, diff_best;
 
-	for (i = 0; i < ARRAY_SIZE(wrpll_tmds_clock_table); i++)
-		if (clock <= wrpll_tmds_clock_table[i].clock)
-			break;
+	/* No best (r,n,p) yet */
+	if (best->p == 0) {
+		best->p = p;
+		best->n2 = n2;
+		best->r2 = r2;
+		return;
+	}
+
+	/*
+	 * Output clock is (LC_FREQ_2K / 2000) * N / (P * R), which compares to
+	 * freq2k.
+	 *
+	 * delta = 1e6 *
+	 *	   abs(freq2k - (LC_FREQ_2K * n2/(p * r2))) /
+	 *	   freq2k;
+	 *
+	 * and we would like delta <= budget.
+	 *
+	 * If the discrepancy is above the PPM-based budget, always prefer to
+	 * improve upon the previous solution.  However, if you're within the
+	 * budget, try to maximize Ref * VCO, that is N / (P * R^2).
+	 */
+	a = freq2k * budget * p * r2;
+	b = freq2k * budget * best->p * best->r2;
+	diff = ABS_DIFF((freq2k * p * r2), (LC_FREQ_2K * n2));
+	diff_best = ABS_DIFF((freq2k * best->p * best->r2),
+			     (LC_FREQ_2K * best->n2));
+	c = 1000000 * diff;
+	d = 1000000 * diff_best;
+
+	if (a < c && b < d) {
+		/* If both are above the budget, pick the closer */
+		if (best->p * best->r2 * diff < p * r2 * diff_best) {
+			best->p = p;
+			best->n2 = n2;
+			best->r2 = r2;
+		}
+	} else if (a >= c && b < d) {
+		/* If A is below the threshold but B is above it?  Update. */
+		best->p = p;
+		best->n2 = n2;
+		best->r2 = r2;
+	} else if (a >= c && b >= d) {
+		/* Both are below the limit, so pick the higher n2/(r2*r2) */
+		if (n2 * best->r2 * best->r2 > best->n2 * r2 * r2) {
+			best->p = p;
+			best->n2 = n2;
+			best->r2 = r2;
+		}
+	}
+	/* Otherwise a < c && b >= d, do nothing */
+}
+
+static void
+intel_ddi_calculate_wrpll(int clock /* in Hz */,
+			  unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
+{
+	uint64_t freq2k;
+	unsigned p, n2, r2;
+	struct wrpll_rnp best = { 0, 0, 0 };
+	unsigned budget;
 
-	if (i == ARRAY_SIZE(wrpll_tmds_clock_table))
-		i--;
+	freq2k = clock / 100;
 
-	*p = wrpll_tmds_clock_table[i].p;
-	*n2 = wrpll_tmds_clock_table[i].n2;
-	*r2 = wrpll_tmds_clock_table[i].r2;
+	budget = wrpll_get_budget_for_freq(clock);
 
-	if (wrpll_tmds_clock_table[i].clock != clock)
-		DRM_INFO("WRPLL: using settings for %dKHz on %dKHz mode\n",
-			 wrpll_tmds_clock_table[i].clock, clock);
+	/* Special case handling for 540 pixel clock: bypass WR PLL entirely
+	 * and directly pass the LC PLL to it. */
+	if (freq2k == 5400000) {
+		*n2_out = 2;
+		*p_out = 1;
+		*r2_out = 2;
+		return;
+	}
+
+	/*
+	 * Ref = LC_FREQ / R, where Ref is the actual reference input seen by
+	 * the WR PLL.
+	 *
+	 * We want R so that REF_MIN <= Ref <= REF_MAX.
+	 * Injecting R2 = 2 * R gives:
+	 *   REF_MAX * r2 > LC_FREQ * 2 and
+	 *   REF_MIN * r2 < LC_FREQ * 2
+	 *
+	 * Which means the desired boundaries for r2 are:
+	 *  LC_FREQ * 2 / REF_MAX < r2 < LC_FREQ * 2 / REF_MIN
+	 *
+	 */
+	for (r2 = LC_FREQ * 2 / REF_MAX + 1;
+	     r2 <= LC_FREQ * 2 / REF_MIN;
+	     r2++) {
+
+		/*
+		 * VCO = N * Ref, that is: VCO = N * LC_FREQ / R
+		 *
+		 * Once again we want VCO_MIN <= VCO <= VCO_MAX.
+		 * Injecting R2 = 2 * R and N2 = 2 * N, we get:
+		 *   VCO_MAX * r2 > n2 * LC_FREQ and
+		 *   VCO_MIN * r2 < n2 * LC_FREQ)
+		 *
+		 * Which means the desired boundaries for n2 are:
+		 * VCO_MIN * r2 / LC_FREQ < n2 < VCO_MAX * r2 / LC_FREQ
+		 */
+		for (n2 = VCO_MIN * r2 / LC_FREQ + 1;
+		     n2 <= VCO_MAX * r2 / LC_FREQ;
+		     n2++) {
 
-	DRM_DEBUG_KMS("WRPLL: %dKHz refresh rate with p=%d, n2=%d r2=%d\n",
-		      clock, *p, *n2, *r2);
+			for (p = P_MIN; p <= P_MAX; p += P_INC)
+				wrpll_update_rnp(freq2k, budget,
+						 r2, n2, p, &best);
+		}
+	}
+
+	*n2_out = best.n2;
+	*p_out = best.p;
+	*r2_out = best.r2;
+
+	DRM_DEBUG_KMS("WRPLL: %dHz refresh rate with p=%d, n2=%d r2=%d\n",
+		      clock, *p_out, *n2_out, *r2_out);
 }
 
-bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
+bool intel_ddi_pll_mode_set(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
@@ -835,6 +634,7 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 	int type = intel_encoder->type;
 	enum pipe pipe = intel_crtc->pipe;
 	uint32_t reg, val;
+	int clock = intel_crtc->config.port_clock;
 
 	/* TODO: reuse PLLs when possible (compare values) */
 
@@ -863,7 +663,7 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 		return true;
 
 	} else if (type == INTEL_OUTPUT_HDMI) {
-		int p, n2, r2;
+		unsigned p, n2, r2;
 
 		if (plls->wrpll1_refcount == 0) {
 			DRM_DEBUG_KMS("Using WRPLL 1 on pipe %c\n",
@@ -885,7 +685,7 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 		WARN(I915_READ(reg) & WRPLL_PLL_ENABLE,
 		     "WRPLL already enabled\n");
 
-		intel_ddi_calculate_wrpll(clock, &p, &n2, &r2);
+		intel_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
 
 		val = WRPLL_PLL_ENABLE | WRPLL_PLL_SELECT_LCPLL_2700 |
 		      WRPLL_DIVIDER_REFERENCE(r2) | WRPLL_DIVIDER_FEEDBACK(n2) |
@@ -995,7 +795,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 			/* Can only use the always-on power well for eDP when
 			 * not using the panel fitter, and when not using motion
 			  * blur mitigation (which we don't support). */
-			if (dev_priv->pch_pf_size)
+			if (intel_crtc->config.pch_pfit.size)
 				temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
 			else
 				temp |= TRANS_DDI_EDP_INPUT_A_ON;
@@ -1022,7 +822,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 
 	} else if (type == INTEL_OUTPUT_ANALOG) {
 		temp |= TRANS_DDI_MODE_SELECT_FDI;
-		temp |= (intel_crtc->fdi_lanes - 1) << 1;
+		temp |= (intel_crtc->config.fdi_lanes - 1) << 1;
 
 	} else if (type == INTEL_OUTPUT_DISPLAYPORT ||
 		   type == INTEL_OUTPUT_EDP) {
@@ -1030,25 +830,10 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 
 		temp |= TRANS_DDI_MODE_SELECT_DP_SST;
 
-		switch (intel_dp->lane_count) {
-		case 1:
-			temp |= TRANS_DDI_PORT_WIDTH_X1;
-			break;
-		case 2:
-			temp |= TRANS_DDI_PORT_WIDTH_X2;
-			break;
-		case 4:
-			temp |= TRANS_DDI_PORT_WIDTH_X4;
-			break;
-		default:
-			temp |= TRANS_DDI_PORT_WIDTH_X4;
-			WARN(1, "Unsupported lane count %d\n",
-			     intel_dp->lane_count);
-		}
-
+		temp |= DDI_PORT_WIDTH(intel_dp->lane_count);
 	} else {
-		WARN(1, "Invalid encoder type %d for pipe %d\n",
-		     intel_encoder->type, pipe);
+		WARN(1, "Invalid encoder type %d for pipe %c\n",
+		     intel_encoder->type, pipe_name(pipe));
 	}
 
 	I915_WRITE(TRANS_DDI_FUNC_CTL(cpu_transcoder), temp);
@@ -1148,7 +933,7 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 		}
 	}
 
-	DRM_DEBUG_KMS("No pipe for ddi port %i found\n", port);
+	DRM_DEBUG_KMS("No pipe for ddi port %c found\n", port_name(port));
 
 	return false;
 }
@@ -1334,7 +1119,7 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 		ironlake_edp_backlight_on(intel_dp);
 	}
 
-	if (intel_crtc->eld_vld) {
+	if (intel_crtc->eld_vld && type != INTEL_OUTPUT_EDP) {
 		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
 		tmp |= ((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
 		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
@@ -1352,9 +1137,12 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
-	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
-	tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
-	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+	if (intel_crtc->eld_vld && type != INTEL_OUTPUT_EDP) {
+		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+		tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) <<
+			 (pipe * 4));
+		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+	}
 
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
@@ -1366,14 +1154,14 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
 {
 	if (I915_READ(HSW_FUSE_STRAP) & HSW_CDCLK_LIMIT)
-		return 450;
+		return 450000;
 	else if ((I915_READ(LCPLL_CTL) & LCPLL_CLK_FREQ_MASK) ==
 		 LCPLL_CLK_FREQ_450)
-		return 450;
+		return 450000;
 	else if (IS_ULT(dev_priv->dev))
-		return 338;
+		return 337500;
 	else
-		return 540;
+		return 540000;
 }
 
 void intel_ddi_pll_init(struct drm_device *dev)
@@ -1386,7 +1174,7 @@ void intel_ddi_pll_init(struct drm_device *dev)
 	 * Don't even try to turn it on.
 	 */
 
-	DRM_DEBUG_KMS("CDCLK running at %dMHz\n",
+	DRM_DEBUG_KMS("CDCLK running at %dKHz\n",
 		      intel_ddi_get_cdclk_freq(dev_priv));
 
 	if (val & LCPLL_CD_SOURCE_FCLK)
@@ -1472,6 +1260,27 @@ static void intel_ddi_hot_plug(struct intel_encoder *intel_encoder)
 		intel_dp_check_link_status(intel_dp);
 }
 
+static void intel_ddi_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
+	u32 temp, flags = 0;
+
+	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
+	if (temp & TRANS_DDI_PHSYNC)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (temp & TRANS_DDI_PVSYNC)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_ddi_destroy(struct drm_encoder *encoder)
 {
 	/* HDMI has nothing special to destroy, so we can go with this. */
@@ -1482,9 +1291,13 @@ static bool intel_ddi_compute_config(struct intel_encoder *encoder,
 				     struct intel_crtc_config *pipe_config)
 {
 	int type = encoder->type;
+	int port = intel_ddi_get_encoder_port(encoder);
 
 	WARN(type == INTEL_OUTPUT_UNKNOWN, "compute_config() on unknown output!\n");
 
+	if (port == PORT_A)
+		pipe_config->cpu_transcoder = TRANSCODER_EDP;
+
 	if (type == INTEL_OUTPUT_HDMI)
 		return intel_hdmi_compute_config(encoder, pipe_config);
 	else
@@ -1518,16 +1331,6 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 		return;
 	}
 
-	if (port != PORT_A) {
-		hdmi_connector = kzalloc(sizeof(struct intel_connector),
-					 GFP_KERNEL);
-		if (!hdmi_connector) {
-			kfree(dp_connector);
-			kfree(intel_dig_port);
-			return;
-		}
-	}
-
 	intel_encoder = &intel_dig_port->base;
 	encoder = &intel_encoder->base;
 
@@ -1541,12 +1344,11 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	intel_encoder->disable = intel_disable_ddi;
 	intel_encoder->post_disable = intel_ddi_post_disable;
 	intel_encoder->get_hw_state = intel_ddi_get_hw_state;
+	intel_encoder->get_config = intel_ddi_get_config;
 
 	intel_dig_port->port = port;
 	intel_dig_port->port_reversal = I915_READ(DDI_BUF_CTL(port)) &
 					DDI_BUF_PORT_REVERSAL;
-	if (hdmi_connector)
-		intel_dig_port->hdmi.hdmi_reg = DDI_BUF_CTL(port);
 	intel_dig_port->dp.output_reg = DDI_BUF_CTL(port);
 
 	intel_encoder->type = INTEL_OUTPUT_UNKNOWN;
@@ -1554,7 +1356,21 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	intel_encoder->cloneable = false;
 	intel_encoder->hot_plug = intel_ddi_hot_plug;
 
-	if (hdmi_connector)
+	if (!intel_dp_init_connector(intel_dig_port, dp_connector)) {
+		drm_encoder_cleanup(encoder);
+		kfree(intel_dig_port);
+		kfree(dp_connector);
+		return;
+	}
+
+	if (intel_encoder->type != INTEL_OUTPUT_EDP) {
+		hdmi_connector = kzalloc(sizeof(struct intel_connector),
+					 GFP_KERNEL);
+		if (!hdmi_connector) {
+			return;
+		}
+
+		intel_dig_port->hdmi.hdmi_reg = DDI_BUF_CTL(port);
 		intel_hdmi_init_connector(intel_dig_port, hdmi_connector);
-	intel_dp_init_connector(intel_dig_port, dp_connector);
+	}
 }
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 56746dc..85f3eb7 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -46,18 +46,6 @@ static void intel_increase_pllclock(struct drm_crtc *crtc);
 static void intel_crtc_update_cursor(struct drm_crtc *crtc, bool on);
 
 typedef struct {
-	/* given values */
-	int n;
-	int m1, m2;
-	int p1, p2;
-	/* derived values */
-	int	dot;
-	int	vco;
-	int	m;
-	int	p;
-} intel_clock_t;
-
-typedef struct {
 	int	min, max;
 } intel_range_t;
 
@@ -71,24 +59,6 @@ typedef struct intel_limit intel_limit_t;
 struct intel_limit {
 	intel_range_t   dot, vco, n, m, m1, m2, p, p1;
 	intel_p2_t	    p2;
-	/**
-	 * find_pll() - Find the best values for the PLL
-	 * @limit: limits for the PLL
-	 * @crtc: current CRTC
-	 * @target: target frequency in kHz
-	 * @refclk: reference clock frequency in kHz
-	 * @match_clock: if provided, @best_clock P divider must
-	 *               match the P divider from @match_clock
-	 *               used for LVDS downclocking
-	 * @best_clock: best PLL values found
-	 *
-	 * Returns true on success, false on failure.
-	 */
-	bool (*find_pll)(const intel_limit_t *limit,
-			 struct drm_crtc *crtc,
-			 int target, int refclk,
-			 intel_clock_t *match_clock,
-			 intel_clock_t *best_clock);
 };
 
 /* FDI */
@@ -104,29 +74,6 @@ intel_pch_rawclk(struct drm_device *dev)
 	return I915_READ(PCH_RAWCLK_FREQ) & RAWCLK_FREQ_MASK;
 }
 
-static bool
-intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-		    int target, int refclk, intel_clock_t *match_clock,
-		    intel_clock_t *best_clock);
-static bool
-intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock);
-
-static bool
-intel_find_pll_g4x_dp(const intel_limit_t *, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *match_clock,
-		      intel_clock_t *best_clock);
-static bool
-intel_find_pll_ironlake_dp(const intel_limit_t *, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *match_clock,
-			   intel_clock_t *best_clock);
-
-static bool
-intel_vlv_find_best_pll(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock);
-
 static inline u32 /* units of 100MHz */
 intel_fdi_link_freq(struct drm_device *dev)
 {
@@ -148,7 +95,6 @@ static const intel_limit_t intel_limits_i8xx_dvo = {
 	.p1 = { .min = 2, .max = 33 },
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 4, .p2_fast = 2 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i8xx_lvds = {
@@ -162,7 +108,6 @@ static const intel_limit_t intel_limits_i8xx_lvds = {
 	.p1 = { .min = 1, .max = 6 },
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 14, .p2_fast = 7 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i9xx_sdvo = {
@@ -176,7 +121,6 @@ static const intel_limit_t intel_limits_i9xx_sdvo = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i9xx_lvds = {
@@ -190,7 +134,6 @@ static const intel_limit_t intel_limits_i9xx_lvds = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
 		.p2_slow = 14, .p2_fast = 7 },
-	.find_pll = intel_find_best_PLL,
 };
 
 
@@ -207,7 +150,6 @@ static const intel_limit_t intel_limits_g4x_sdvo = {
 		.p2_slow = 10,
 		.p2_fast = 10
 	},
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_g4x_hdmi = {
@@ -221,7 +163,6 @@ static const intel_limit_t intel_limits_g4x_hdmi = {
 	.p1 = { .min = 1, .max = 8},
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_g4x_single_channel_lvds = {
@@ -236,7 +177,6 @@ static const intel_limit_t intel_limits_g4x_single_channel_lvds = {
 	.p2 = { .dot_limit = 0,
 		.p2_slow = 14, .p2_fast = 14
 	},
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_g4x_dual_channel_lvds = {
@@ -251,21 +191,6 @@ static const intel_limit_t intel_limits_g4x_dual_channel_lvds = {
 	.p2 = { .dot_limit = 0,
 		.p2_slow = 7, .p2_fast = 7
 	},
-	.find_pll = intel_g4x_find_best_PLL,
-};
-
-static const intel_limit_t intel_limits_g4x_display_port = {
-	.dot = { .min = 161670, .max = 227000 },
-	.vco = { .min = 1750000, .max = 3500000},
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 97, .max = 108 },
-	.m1 = { .min = 0x10, .max = 0x12 },
-	.m2 = { .min = 0x05, .max = 0x06 },
-	.p = { .min = 10, .max = 20 },
-	.p1 = { .min = 1, .max = 2},
-	.p2 = { .dot_limit = 0,
-		.p2_slow = 10, .p2_fast = 10 },
-	.find_pll = intel_find_pll_g4x_dp,
 };
 
 static const intel_limit_t intel_limits_pineview_sdvo = {
@@ -281,7 +206,6 @@ static const intel_limit_t intel_limits_pineview_sdvo = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_pineview_lvds = {
@@ -295,7 +219,6 @@ static const intel_limit_t intel_limits_pineview_lvds = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
 		.p2_slow = 14, .p2_fast = 14 },
-	.find_pll = intel_find_best_PLL,
 };
 
 /* Ironlake / Sandybridge
@@ -314,7 +237,6 @@ static const intel_limit_t intel_limits_ironlake_dac = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_single_lvds = {
@@ -328,7 +250,6 @@ static const intel_limit_t intel_limits_ironlake_single_lvds = {
 	.p1 = { .min = 2, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 14, .p2_fast = 14 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_dual_lvds = {
@@ -342,7 +263,6 @@ static const intel_limit_t intel_limits_ironlake_dual_lvds = {
 	.p1 = { .min = 2, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 7, .p2_fast = 7 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 /* LVDS 100mhz refclk limits. */
@@ -357,7 +277,6 @@ static const intel_limit_t intel_limits_ironlake_single_lvds_100m = {
 	.p1 = { .min = 2, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 14, .p2_fast = 14 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_dual_lvds_100m = {
@@ -371,21 +290,6 @@ static const intel_limit_t intel_limits_ironlake_dual_lvds_100m = {
 	.p1 = { .min = 2, .max = 6 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 7, .p2_fast = 7 },
-	.find_pll = intel_g4x_find_best_PLL,
-};
-
-static const intel_limit_t intel_limits_ironlake_display_port = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000},
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 81, .max = 90 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 10, .max = 20 },
-	.p1 = { .min = 1, .max = 2},
-	.p2 = { .dot_limit = 0,
-		.p2_slow = 10, .p2_fast = 10 },
-	.find_pll = intel_find_pll_ironlake_dp,
 };
 
 static const intel_limit_t intel_limits_vlv_dac = {
@@ -396,15 +300,14 @@ static const intel_limit_t intel_limits_vlv_dac = {
 	.m1 = { .min = 2, .max = 3 },
 	.m2 = { .min = 11, .max = 156 },
 	.p = { .min = 10, .max = 30 },
-	.p1 = { .min = 2, .max = 3 },
+	.p1 = { .min = 1, .max = 3 },
 	.p2 = { .dot_limit = 270000,
 		.p2_slow = 2, .p2_fast = 20 },
-	.find_pll = intel_vlv_find_best_pll,
 };
 
 static const intel_limit_t intel_limits_vlv_hdmi = {
-	.dot = { .min = 20000, .max = 165000 },
-	.vco = { .min = 4000000, .max = 5994000},
+	.dot = { .min = 25000, .max = 270000 },
+	.vco = { .min = 4000000, .max = 6000000 },
 	.n = { .min = 1, .max = 7 },
 	.m = { .min = 60, .max = 300 }, /* guess */
 	.m1 = { .min = 2, .max = 3 },
@@ -413,7 +316,6 @@ static const intel_limit_t intel_limits_vlv_hdmi = {
 	.p1 = { .min = 2, .max = 3 },
 	.p2 = { .dot_limit = 270000,
 		.p2_slow = 2, .p2_fast = 20 },
-	.find_pll = intel_vlv_find_best_pll,
 };
 
 static const intel_limit_t intel_limits_vlv_dp = {
@@ -424,61 +326,11 @@ static const intel_limit_t intel_limits_vlv_dp = {
 	.m1 = { .min = 2, .max = 3 },
 	.m2 = { .min = 11, .max = 156 },
 	.p = { .min = 10, .max = 30 },
-	.p1 = { .min = 2, .max = 3 },
+	.p1 = { .min = 1, .max = 3 },
 	.p2 = { .dot_limit = 270000,
 		.p2_slow = 2, .p2_fast = 20 },
-	.find_pll = intel_vlv_find_best_pll,
 };
 
-u32 intel_dpio_read(struct drm_i915_private *dev_priv, int reg)
-{
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
-		DRM_ERROR("DPIO idle wait timed out\n");
-		return 0;
-	}
-
-	I915_WRITE(DPIO_REG, reg);
-	I915_WRITE(DPIO_PKT, DPIO_RID | DPIO_OP_READ | DPIO_PORTID |
-		   DPIO_BYTE);
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
-		DRM_ERROR("DPIO read wait timed out\n");
-		return 0;
-	}
-
-	return I915_READ(DPIO_DATA);
-}
-
-static void intel_dpio_write(struct drm_i915_private *dev_priv, int reg,
-			     u32 val)
-{
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
-		DRM_ERROR("DPIO idle wait timed out\n");
-		return;
-	}
-
-	I915_WRITE(DPIO_DATA, val);
-	I915_WRITE(DPIO_REG, reg);
-	I915_WRITE(DPIO_PKT, DPIO_RID | DPIO_OP_WRITE | DPIO_PORTID |
-		   DPIO_BYTE);
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100))
-		DRM_ERROR("DPIO write wait timed out\n");
-}
-
-static void vlv_init_dpio(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	/* Reset the DPIO config */
-	I915_WRITE(DPIO_CTL, 0);
-	POSTING_READ(DPIO_CTL);
-	I915_WRITE(DPIO_CTL, 1);
-	POSTING_READ(DPIO_CTL);
-}
-
 static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 						int refclk)
 {
@@ -497,10 +349,7 @@ static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 			else
 				limit = &intel_limits_ironlake_single_lvds;
 		}
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
-		   intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))
-		limit = &intel_limits_ironlake_display_port;
-	else
+	} else
 		limit = &intel_limits_ironlake_dac;
 
 	return limit;
@@ -521,8 +370,6 @@ static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
 		limit = &intel_limits_g4x_hdmi;
 	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO)) {
 		limit = &intel_limits_g4x_sdvo;
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
-		limit = &intel_limits_g4x_display_port;
 	} else /* The option is for other outputs */
 		limit = &intel_limits_i9xx_sdvo;
 
@@ -573,13 +420,14 @@ static void pineview_clock(int refclk, intel_clock_t *clock)
 	clock->dot = clock->vco / clock->p;
 }
 
-static void intel_clock(struct drm_device *dev, int refclk, intel_clock_t *clock)
+static uint32_t i9xx_dpll_compute_m(struct dpll *dpll)
 {
-	if (IS_PINEVIEW(dev)) {
-		pineview_clock(refclk, clock);
-		return;
-	}
-	clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2);
+	return 5 * (dpll->m1 + 2) + (dpll->m2 + 2);
+}
+
+static void i9xx_clock(int refclk, intel_clock_t *clock)
+{
+	clock->m = i9xx_dpll_compute_m(clock);
 	clock->p = clock->p1 * clock->p2;
 	clock->vco = refclk * clock->m / (clock->n + 2);
 	clock->dot = clock->vco / clock->p;
@@ -636,10 +484,9 @@ static bool intel_PLL_is_valid(struct drm_device *dev,
 }
 
 static bool
-intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
+i9xx_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 		    int target, int refclk, intel_clock_t *match_clock,
 		    intel_clock_t *best_clock)
-
 {
 	struct drm_device *dev = crtc->dev;
 	intel_clock_t clock;
@@ -668,8 +515,7 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 	     clock.m1++) {
 		for (clock.m2 = limit->m2.min;
 		     clock.m2 <= limit->m2.max; clock.m2++) {
-			/* m1 is always 0 in Pineview */
-			if (clock.m2 >= clock.m1 && !IS_PINEVIEW(dev))
+			if (clock.m2 >= clock.m1)
 				break;
 			for (clock.n = limit->n.min;
 			     clock.n <= limit->n.max; clock.n++) {
@@ -677,7 +523,66 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 					clock.p1 <= limit->p1.max; clock.p1++) {
 					int this_err;
 
-					intel_clock(dev, refclk, &clock);
+					i9xx_clock(refclk, &clock);
+					if (!intel_PLL_is_valid(dev, limit,
+								&clock))
+						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err) {
+						*best_clock = clock;
+						err = this_err;
+					}
+				}
+			}
+		}
+	}
+
+	return (err != target);
+}
+
+static bool
+pnv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+		   int target, int refclk, intel_clock_t *match_clock,
+		   intel_clock_t *best_clock)
+{
+	struct drm_device *dev = crtc->dev;
+	intel_clock_t clock;
+	int err = target;
+
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+		/*
+		 * For LVDS just rely on its current settings for dual-channel.
+		 * We haven't figured out how to reliably set up different
+		 * single/dual channel state, if we even can.
+		 */
+		if (intel_is_dual_link_lvds(dev))
+			clock.p2 = limit->p2.p2_fast;
+		else
+			clock.p2 = limit->p2.p2_slow;
+	} else {
+		if (target < limit->p2.dot_limit)
+			clock.p2 = limit->p2.p2_slow;
+		else
+			clock.p2 = limit->p2.p2_fast;
+	}
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
+	     clock.m1++) {
+		for (clock.m2 = limit->m2.min;
+		     clock.m2 <= limit->m2.max; clock.m2++) {
+			for (clock.n = limit->n.min;
+			     clock.n <= limit->n.max; clock.n++) {
+				for (clock.p1 = limit->p1.min;
+					clock.p1 <= limit->p1.max; clock.p1++) {
+					int this_err;
+
+					pineview_clock(refclk, &clock);
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
@@ -699,9 +604,9 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock)
+g4x_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+		   int target, int refclk, intel_clock_t *match_clock,
+		   intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
 	intel_clock_t clock;
@@ -712,12 +617,6 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 	found = false;
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
-		int lvds_reg;
-
-		if (HAS_PCH_SPLIT(dev))
-			lvds_reg = PCH_LVDS;
-		else
-			lvds_reg = LVDS;
 		if (intel_is_dual_link_lvds(dev))
 			clock.p2 = limit->p2.p2_fast;
 		else
@@ -742,13 +641,10 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 				     clock.p1 >= limit->p1.min; clock.p1--) {
 					int this_err;
 
-					intel_clock(dev, refclk, &clock);
+					i9xx_clock(refclk, &clock);
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
-					if (match_clock &&
-					    clock.p != match_clock->p)
-						continue;
 
 					this_err = abs(clock.dot - target);
 					if (this_err < err_most) {
@@ -765,62 +661,9 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *match_clock,
-			   intel_clock_t *best_clock)
-{
-	struct drm_device *dev = crtc->dev;
-	intel_clock_t clock;
-
-	if (target < 200000) {
-		clock.n = 1;
-		clock.p1 = 2;
-		clock.p2 = 10;
-		clock.m1 = 12;
-		clock.m2 = 9;
-	} else {
-		clock.n = 2;
-		clock.p1 = 1;
-		clock.p2 = 10;
-		clock.m1 = 14;
-		clock.m2 = 8;
-	}
-	intel_clock(dev, refclk, &clock);
-	memcpy(best_clock, &clock, sizeof(intel_clock_t));
-	return true;
-}
-
-/* DisplayPort has only two frequencies, 162MHz and 270MHz */
-static bool
-intel_find_pll_g4x_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *match_clock,
-		      intel_clock_t *best_clock)
-{
-	intel_clock_t clock;
-	if (target < 200000) {
-		clock.p1 = 2;
-		clock.p2 = 10;
-		clock.n = 2;
-		clock.m1 = 23;
-		clock.m2 = 8;
-	} else {
-		clock.p1 = 1;
-		clock.p2 = 10;
-		clock.n = 1;
-		clock.m1 = 14;
-		clock.m2 = 2;
-	}
-	clock.m = 5 * (clock.m1 + 2) + (clock.m2 + 2);
-	clock.p = (clock.p1 * clock.p2);
-	clock.dot = 96000 * clock.m / (clock.n + 2) / clock.p;
-	clock.vco = 0;
-	memcpy(best_clock, &clock, sizeof(intel_clock_t));
-	return true;
-}
-static bool
-intel_vlv_find_best_pll(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock)
+vlv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+		   int target, int refclk, intel_clock_t *match_clock,
+		   intel_clock_t *best_clock)
 {
 	u32 p1, p2, m1, m2, vco, bestn, bestm1, bestm2, bestp1, bestp2;
 	u32 m, n, fastclk;
@@ -1066,14 +909,24 @@ static void assert_pll(struct drm_i915_private *dev_priv,
 #define assert_pll_enabled(d, p) assert_pll(d, p, true)
 #define assert_pll_disabled(d, p) assert_pll(d, p, false)
 
+static struct intel_shared_dpll *
+intel_crtc_to_shared_dpll(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+
+	if (crtc->config.shared_dpll < 0)
+		return NULL;
+
+	return &dev_priv->shared_dplls[crtc->config.shared_dpll];
+}
+
 /* For ILK+ */
-static void assert_pch_pll(struct drm_i915_private *dev_priv,
-			   struct intel_pch_pll *pll,
-			   struct intel_crtc *crtc,
-			   bool state)
+static void assert_shared_dpll(struct drm_i915_private *dev_priv,
+			       struct intel_shared_dpll *pll,
+			       bool state)
 {
-	u32 val;
 	bool cur_state;
+	struct intel_dpll_hw_state hw_state;
 
 	if (HAS_PCH_LPT(dev_priv->dev)) {
 		DRM_DEBUG_DRIVER("LPT detected: skipping PCH PLL test\n");
@@ -1081,36 +934,16 @@ static void assert_pch_pll(struct drm_i915_private *dev_priv,
 	}
 
 	if (WARN (!pll,
-		  "asserting PCH PLL %s with no PLL\n", state_string(state)))
+		  "asserting DPLL %s with no DPLL\n", state_string(state)))
 		return;
 
-	val = I915_READ(pll->pll_reg);
-	cur_state = !!(val & DPLL_VCO_ENABLE);
+	cur_state = pll->get_hw_state(dev_priv, pll, &hw_state);
 	WARN(cur_state != state,
-	     "PCH PLL state for reg %x assertion failure (expected %s, current %s), val=%08x\n",
-	     pll->pll_reg, state_string(state), state_string(cur_state), val);
-
-	/* Make sure the selected PLL is correctly attached to the transcoder */
-	if (crtc && HAS_PCH_CPT(dev_priv->dev)) {
-		u32 pch_dpll;
-
-		pch_dpll = I915_READ(PCH_DPLL_SEL);
-		cur_state = pll->pll_reg == _PCH_DPLL_B;
-		if (!WARN(((pch_dpll >> (4 * crtc->pipe)) & 1) != cur_state,
-			  "PLL[%d] not attached to this transcoder %d: %08x\n",
-			  cur_state, crtc->pipe, pch_dpll)) {
-			cur_state = !!(val >> (4*crtc->pipe + 3));
-			WARN(cur_state != state,
-			     "PLL[%d] not %s on this transcoder %d: %08x\n",
-			     pll->pll_reg == _PCH_DPLL_B,
-			     state_string(state),
-			     crtc->pipe,
-			     val);
-		}
-	}
+	     "%s assertion failure (expected %s, current %s)\n",
+	     pll->name, state_string(state), state_string(cur_state));
 }
-#define assert_pch_pll_enabled(d, p, c) assert_pch_pll(d, p, c, true)
-#define assert_pch_pll_disabled(d, p, c) assert_pch_pll(d, p, c, false)
+#define assert_shared_dpll_enabled(d, p) assert_shared_dpll(d, p, true)
+#define assert_shared_dpll_disabled(d, p) assert_shared_dpll(d, p, false)
 
 static void assert_fdi_tx(struct drm_i915_private *dev_priv,
 			  enum pipe pipe, bool state)
@@ -1227,8 +1060,8 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
 		state = true;
 
-	if (!intel_using_power_well(dev_priv->dev) &&
-	    cpu_transcoder != TRANSCODER_EDP) {
+	if (!intel_display_power_enabled(dev_priv->dev,
+				POWER_DOMAIN_TRANSCODER(cpu_transcoder))) {
 		cur_state = false;
 	} else {
 		reg = PIPECONF(cpu_transcoder);
@@ -1262,12 +1095,13 @@ static void assert_plane(struct drm_i915_private *dev_priv,
 static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 				   enum pipe pipe)
 {
+	struct drm_device *dev = dev_priv->dev;
 	int reg, i;
 	u32 val;
 	int cur_pipe;
 
-	/* Planes are fixed to pipes on ILK+ */
-	if (HAS_PCH_SPLIT(dev_priv->dev) || IS_VALLEYVIEW(dev_priv->dev)) {
+	/* Primary planes are fixed to pipes on gen4+ */
+	if (INTEL_INFO(dev)->gen >= 4) {
 		reg = DSPCNTR(pipe);
 		val = I915_READ(reg);
 		WARN((val & DISPLAY_PLANE_ENABLE),
@@ -1277,7 +1111,7 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	}
 
 	/* Need to check both planes against the pipe */
-	for (i = 0; i < 2; i++) {
+	for (i = 0; i < INTEL_INFO(dev)->num_pipes; i++) {
 		reg = DSPCNTR(i);
 		val = I915_READ(reg);
 		cur_pipe = (val & DISPPLANE_SEL_PIPE_MASK) >>
@@ -1291,19 +1125,30 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
+	struct drm_device *dev = dev_priv->dev;
 	int reg, i;
 	u32 val;
 
-	if (!IS_VALLEYVIEW(dev_priv->dev))
-		return;
-
-	/* Need to check both planes against the pipe */
-	for (i = 0; i < dev_priv->num_plane; i++) {
-		reg = SPCNTR(pipe, i);
+	if (IS_VALLEYVIEW(dev)) {
+		for (i = 0; i < dev_priv->num_plane; i++) {
+			reg = SPCNTR(pipe, i);
+			val = I915_READ(reg);
+			WARN((val & SP_ENABLE),
+			     "sprite %c assertion failure, should be off on pipe %c but is still active\n",
+			     sprite_name(pipe, i), pipe_name(pipe));
+		}
+	} else if (INTEL_INFO(dev)->gen >= 7) {
+		reg = SPRCTL(pipe);
 		val = I915_READ(reg);
-		WARN((val & SP_ENABLE),
-		     "sprite %d assertion failure, should be off on pipe %c but is still active\n",
-		     pipe * 2 + i, pipe_name(pipe));
+		WARN((val & SPRITE_ENABLE),
+		     "sprite %c assertion failure, should be off on pipe %c but is still active\n",
+		     plane_name(pipe), pipe_name(pipe));
+	} else if (INTEL_INFO(dev)->gen >= 5) {
+		reg = DVSCNTR(pipe);
+		val = I915_READ(reg);
+		WARN((val & DVS_ENABLE),
+		     "sprite %c assertion failure, should be off on pipe %c but is still active\n",
+		     plane_name(pipe), pipe_name(pipe));
 	}
 }
 
@@ -1323,14 +1168,14 @@ static void assert_pch_refclk_enabled(struct drm_i915_private *dev_priv)
 	WARN(!enabled, "PCH refclk assertion failure, should be active but is disabled\n");
 }
 
-static void assert_transcoder_disabled(struct drm_i915_private *dev_priv,
-				       enum pipe pipe)
+static void assert_pch_transcoder_disabled(struct drm_i915_private *dev_priv,
+					   enum pipe pipe)
 {
 	int reg;
 	u32 val;
 	bool enabled;
 
-	reg = TRANSCONF(pipe);
+	reg = PCH_TRANSCONF(pipe);
 	val = I915_READ(reg);
 	enabled = !!(val & TRANS_ENABLE);
 	WARN(enabled,
@@ -1474,6 +1319,8 @@ static void intel_enable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	int reg;
 	u32 val;
 
+	assert_pipe_disabled(dev_priv, pipe);
+
 	/* No really, not for ILK+ */
 	BUG_ON(!IS_VALLEYVIEW(dev_priv->dev) && dev_priv->info->gen >= 5);
 
@@ -1525,156 +1372,86 @@ static void intel_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	POSTING_READ(reg);
 }
 
-/* SBI access */
-static void
-intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
-		enum intel_sbi_destination destination)
+void vlv_wait_port_ready(struct drm_i915_private *dev_priv, int port)
 {
-	u32 tmp;
+	u32 port_mask;
 
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to become ready\n");
-		return;
-	}
-
-	I915_WRITE(SBI_ADDR, (reg << 16));
-	I915_WRITE(SBI_DATA, value);
-
-	if (destination == SBI_ICLK)
-		tmp = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRWR;
+	if (!port)
+		port_mask = DPLL_PORTB_READY_MASK;
 	else
-		tmp = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IOWR;
-	I915_WRITE(SBI_CTL_STAT, SBI_BUSY | tmp);
+		port_mask = DPLL_PORTC_READY_MASK;
 
-	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
-		return;
-	}
-}
-
-static u32
-intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
-	       enum intel_sbi_destination destination)
-{
-	u32 value = 0;
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to become ready\n");
-		return 0;
-	}
-
-	I915_WRITE(SBI_ADDR, (reg << 16));
-
-	if (destination == SBI_ICLK)
-		value = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRRD;
-	else
-		value = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IORD;
-	I915_WRITE(SBI_CTL_STAT, value | SBI_BUSY);
-
-	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
-		return 0;
-	}
-
-	return I915_READ(SBI_DATA);
+	if (wait_for((I915_READ(DPLL(0)) & port_mask) == 0, 1000))
+		WARN(1, "timed out waiting for port %c ready: 0x%08x\n",
+		     'B' + port, I915_READ(DPLL(0)));
 }
 
 /**
- * ironlake_enable_pch_pll - enable PCH PLL
+ * ironlake_enable_shared_dpll - enable PCH PLL
  * @dev_priv: i915 private structure
  * @pipe: pipe PLL to enable
  *
  * The PCH PLL needs to be enabled before the PCH transcoder, since it
  * drives the transcoder clock.
  */
-static void ironlake_enable_pch_pll(struct intel_crtc *intel_crtc)
+static void ironlake_enable_shared_dpll(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	struct intel_pch_pll *pll;
-	int reg;
-	u32 val;
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
 	/* PCH PLLs only available on ILK, SNB and IVB */
 	BUG_ON(dev_priv->info->gen < 5);
-	pll = intel_crtc->pch_pll;
-	if (pll == NULL)
+	if (WARN_ON(pll == NULL))
 		return;
 
 	if (WARN_ON(pll->refcount == 0))
 		return;
 
-	DRM_DEBUG_KMS("enable PCH PLL %x (active %d, on? %d)for crtc %d\n",
-		      pll->pll_reg, pll->active, pll->on,
-		      intel_crtc->base.base.id);
-
-	/* PCH refclock must be enabled first */
-	assert_pch_refclk_enabled(dev_priv);
+	DRM_DEBUG_KMS("enable %s (active %d, on? %d)for crtc %d\n",
+		      pll->name, pll->active, pll->on,
+		      crtc->base.base.id);
 
-	if (pll->active++ && pll->on) {
-		assert_pch_pll_enabled(dev_priv, pll, NULL);
+	if (pll->active++) {
+		WARN_ON(!pll->on);
+		assert_shared_dpll_enabled(dev_priv, pll);
 		return;
 	}
+	WARN_ON(pll->on);
 
-	DRM_DEBUG_KMS("enabling PCH PLL %x\n", pll->pll_reg);
-
-	reg = pll->pll_reg;
-	val = I915_READ(reg);
-	val |= DPLL_VCO_ENABLE;
-	I915_WRITE(reg, val);
-	POSTING_READ(reg);
-	udelay(200);
-
+	DRM_DEBUG_KMS("enabling %s\n", pll->name);
+	pll->enable(dev_priv, pll);
 	pll->on = true;
 }
 
-static void intel_disable_pch_pll(struct intel_crtc *intel_crtc)
+static void intel_disable_shared_dpll(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	struct intel_pch_pll *pll = intel_crtc->pch_pll;
-	int reg;
-	u32 val;
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
-	if (pll == NULL)
+	if (WARN_ON(pll == NULL))
 	       return;
 
 	if (WARN_ON(pll->refcount == 0))
 		return;
 
-	DRM_DEBUG_KMS("disable PCH PLL %x (active %d, on? %d) for crtc %d\n",
-		      pll->pll_reg, pll->active, pll->on,
-		      intel_crtc->base.base.id);
+	DRM_DEBUG_KMS("disable %s (active %d, on? %d) for crtc %d\n",
+		      pll->name, pll->active, pll->on,
+		      crtc->base.base.id);
 
 	if (WARN_ON(pll->active == 0)) {
-		assert_pch_pll_disabled(dev_priv, pll, NULL);
+		assert_shared_dpll_disabled(dev_priv, pll);
 		return;
 	}
 
-	if (--pll->active) {
-		assert_pch_pll_enabled(dev_priv, pll, NULL);
+	assert_shared_dpll_enabled(dev_priv, pll);
+	WARN_ON(!pll->on);
+	if (--pll->active)
 		return;
-	}
-
-	DRM_DEBUG_KMS("disabling PCH PLL %x\n", pll->pll_reg);
-
-	/* Make sure transcoder isn't still depending on us */
-	assert_transcoder_disabled(dev_priv, intel_crtc->pipe);
-
-	reg = pll->pll_reg;
-	val = I915_READ(reg);
-	val &= ~DPLL_VCO_ENABLE;
-	I915_WRITE(reg, val);
-	POSTING_READ(reg);
-	udelay(200);
 
+	DRM_DEBUG_KMS("disabling %s\n", pll->name);
+	pll->disable(dev_priv, pll);
 	pll->on = false;
 }
 
@@ -1683,15 +1460,15 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t reg, val, pipeconf_val;
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
 
 	/* Make sure PCH DPLL is enabled */
-	assert_pch_pll_enabled(dev_priv,
-			       to_intel_crtc(crtc)->pch_pll,
-			       to_intel_crtc(crtc));
+	assert_shared_dpll_enabled(dev_priv,
+				   intel_crtc_to_shared_dpll(intel_crtc));
 
 	/* FDI must be feeding us bits for PCH ports */
 	assert_fdi_tx_enabled(dev_priv, pipe);
@@ -1706,7 +1483,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 		I915_WRITE(reg, val);
 	}
 
-	reg = TRANSCONF(pipe);
+	reg = PCH_TRANSCONF(pipe);
 	val = I915_READ(reg);
 	pipeconf_val = I915_READ(PIPECONF(pipe));
 
@@ -1731,7 +1508,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 
 	I915_WRITE(reg, val | TRANS_ENABLE);
 	if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
-		DRM_ERROR("failed to enable transcoder %d\n", pipe);
+		DRM_ERROR("failed to enable transcoder %c\n", pipe_name(pipe));
 }
 
 static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
@@ -1760,8 +1537,8 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	else
 		val |= TRANS_PROGRESSIVE;
 
-	I915_WRITE(TRANSCONF(TRANSCODER_A), val);
-	if (wait_for(I915_READ(_TRANSACONF) & TRANS_STATE_ENABLE, 100))
+	I915_WRITE(LPT_TRANSCONF, val);
+	if (wait_for(I915_READ(LPT_TRANSCONF) & TRANS_STATE_ENABLE, 100))
 		DRM_ERROR("Failed to enable PCH transcoder\n");
 }
 
@@ -1778,13 +1555,13 @@ static void ironlake_disable_pch_transcoder(struct drm_i915_private *dev_priv,
 	/* Ports must be off as well */
 	assert_pch_ports_disabled(dev_priv, pipe);
 
-	reg = TRANSCONF(pipe);
+	reg = PCH_TRANSCONF(pipe);
 	val = I915_READ(reg);
 	val &= ~TRANS_ENABLE;
 	I915_WRITE(reg, val);
 	/* wait for PCH transcoder off, transcoder state */
 	if (wait_for((I915_READ(reg) & TRANS_STATE_ENABLE) == 0, 50))
-		DRM_ERROR("failed to disable transcoder %d\n", pipe);
+		DRM_ERROR("failed to disable transcoder %c\n", pipe_name(pipe));
 
 	if (!HAS_PCH_IBX(dev)) {
 		/* Workaround: Clear the timing override chicken bit again. */
@@ -1799,11 +1576,11 @@ static void lpt_disable_pch_transcoder(struct drm_i915_private *dev_priv)
 {
 	u32 val;
 
-	val = I915_READ(_TRANSACONF);
+	val = I915_READ(LPT_TRANSCONF);
 	val &= ~TRANS_ENABLE;
-	I915_WRITE(_TRANSACONF, val);
+	I915_WRITE(LPT_TRANSCONF, val);
 	/* wait for PCH transcoder off, transcoder state */
-	if (wait_for((I915_READ(_TRANSACONF) & TRANS_STATE_ENABLE) == 0, 50))
+	if (wait_for((I915_READ(LPT_TRANSCONF) & TRANS_STATE_ENABLE) == 0, 50))
 		DRM_ERROR("Failed to disable PCH transcoder\n");
 
 	/* Workaround: clear timing override bit. */
@@ -1835,6 +1612,9 @@ static void intel_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 	int reg;
 	u32 val;
 
+	assert_planes_disabled(dev_priv, pipe);
+	assert_sprites_disabled(dev_priv, pipe);
+
 	if (HAS_PCH_LPT(dev_priv->dev))
 		pch_transcoder = TRANSCODER_A;
 	else
@@ -2096,7 +1876,7 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	case 1:
 		break;
 	default:
-		DRM_ERROR("Can't update plane %d in SAREA\n", plane);
+		DRM_ERROR("Can't update plane %c in SAREA\n", plane_name(plane));
 		return -EINVAL;
 	}
 
@@ -2145,6 +1925,9 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			dspcntr &= ~DISPPLANE_TILED;
 	}
 
+	if (IS_G4X(dev))
+		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
+
 	I915_WRITE(reg, dspcntr);
 
 	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
@@ -2193,7 +1976,7 @@ static int ironlake_update_plane(struct drm_crtc *crtc,
 	case 2:
 		break;
 	default:
-		DRM_ERROR("Can't update plane %d in SAREA\n", plane);
+		DRM_ERROR("Can't update plane %c in SAREA\n", plane_name(plane));
 		return -EINVAL;
 	}
 
@@ -2384,9 +2167,9 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	}
 
 	if (intel_crtc->plane > INTEL_INFO(dev)->num_pipes) {
-		DRM_ERROR("no plane for crtc: plane %d, num_pipes %d\n",
-				intel_crtc->plane,
-				INTEL_INFO(dev)->num_pipes);
+		DRM_ERROR("no plane for crtc: plane %c, num_pipes %d\n",
+			  plane_name(intel_crtc->plane),
+			  INTEL_INFO(dev)->num_pipes);
 		return -EINVAL;
 	}
 
@@ -2414,7 +2197,8 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	crtc->y = y;
 
 	if (old_fb) {
-		intel_wait_for_vblank(dev, intel_crtc->pipe);
+		if (intel_crtc->active && old_fb != fb)
+			intel_wait_for_vblank(dev, intel_crtc->pipe);
 		intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj);
 	}
 
@@ -2467,6 +2251,11 @@ static void intel_fdi_normal_train(struct drm_crtc *crtc)
 			   FDI_FE_ERRC_ENABLE);
 }
 
+static bool pipe_has_enabled_pch(struct intel_crtc *intel_crtc)
+{
+	return intel_crtc->base.enabled && intel_crtc->config.has_pch_encoder;
+}
+
 static void ivb_modeset_global_resources(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -2476,10 +2265,13 @@ static void ivb_modeset_global_resources(struct drm_device *dev)
 		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_C]);
 	uint32_t temp;
 
-	/* When everything is off disable fdi C so that we could enable fdi B
-	 * with all lanes. XXX: This misses the case where a pipe is not using
-	 * any pch resources and so doesn't need any fdi lanes. */
-	if (!pipe_B_crtc->base.enabled && !pipe_C_crtc->base.enabled) {
+	/*
+	 * When everything is off disable fdi C so that we could enable fdi B
+	 * with all lanes. Note that we don't care about enabled pipes without
+	 * an enabled pch encoder.
+	 */
+	if (!pipe_has_enabled_pch(pipe_B_crtc) &&
+	    !pipe_has_enabled_pch(pipe_C_crtc)) {
 		WARN_ON(I915_READ(FDI_RX_CTL(PIPE_B)) & FDI_RX_ENABLE);
 		WARN_ON(I915_READ(FDI_RX_CTL(PIPE_C)) & FDI_RX_ENABLE);
 
@@ -2517,8 +2309,8 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 	/* enable CPU FDI TX and PCH FDI RX */
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~(7 << 19);
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp &= ~FDI_LINK_TRAIN_NONE;
 	temp |= FDI_LINK_TRAIN_PATTERN_1;
 	I915_WRITE(reg, temp | FDI_TX_ENABLE);
@@ -2615,8 +2407,8 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 	/* enable CPU FDI TX and PCH FDI RX */
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~(7 << 19);
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp &= ~FDI_LINK_TRAIN_NONE;
 	temp |= FDI_LINK_TRAIN_PATTERN_1;
 	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2750,8 +2542,8 @@ static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
 	/* enable CPU FDI TX and PCH FDI RX */
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~(7 << 19);
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp &= ~(FDI_LINK_TRAIN_AUTO | FDI_LINK_TRAIN_NONE_IVB);
 	temp |= FDI_LINK_TRAIN_PATTERN_1_IVB;
 	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2852,8 +2644,8 @@ static void ironlake_fdi_pll_enable(struct intel_crtc *intel_crtc)
 	/* enable PCH FDI RX PLL, wait warmup plus DMI latency */
 	reg = FDI_RX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~((0x7 << 19) | (0x7 << 16));
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~(FDI_DP_PORT_WIDTH_MASK | (0x7 << 16));
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp |= (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
 	I915_WRITE(reg, temp | FDI_RX_PLL_ENABLE);
 
@@ -3085,6 +2877,30 @@ static void lpt_program_iclkip(struct drm_crtc *crtc)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static void ironlake_pch_transcoder_set_timings(struct intel_crtc *crtc,
+						enum pipe pch_transcoder)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+
+	I915_WRITE(PCH_TRANS_HTOTAL(pch_transcoder),
+		   I915_READ(HTOTAL(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_HBLANK(pch_transcoder),
+		   I915_READ(HBLANK(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_HSYNC(pch_transcoder),
+		   I915_READ(HSYNC(cpu_transcoder)));
+
+	I915_WRITE(PCH_TRANS_VTOTAL(pch_transcoder),
+		   I915_READ(VTOTAL(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_VBLANK(pch_transcoder),
+		   I915_READ(VBLANK(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_VSYNC(pch_transcoder),
+		   I915_READ(VSYNC(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_VSYNCSHIFT(pch_transcoder),
+		   I915_READ(VSYNCSHIFT(cpu_transcoder)));
+}
+
 /*
  * Enable PCH resources required for PCH ports:
  *   - PCH PLLs
@@ -3101,7 +2917,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	int pipe = intel_crtc->pipe;
 	u32 reg, temp;
 
-	assert_transcoder_disabled(dev_priv, pipe);
+	assert_pch_transcoder_disabled(dev_priv, pipe);
 
 	/* Write the TU size bits before fdi link training, so that error
 	 * detection works. */
@@ -3115,31 +2931,18 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	 * transcoder, and we actually should do this to not upset any PCH
 	 * transcoder that already use the clock when we share it.
 	 *
-	 * Note that enable_pch_pll tries to do the right thing, but get_pch_pll
-	 * unconditionally resets the pll - we need that to have the right LVDS
-	 * enable sequence. */
-	ironlake_enable_pch_pll(intel_crtc);
+	 * Note that enable_shared_dpll tries to do the right thing, but
+	 * get_shared_dpll unconditionally resets the pll - we need that to have
+	 * the right LVDS enable sequence. */
+	ironlake_enable_shared_dpll(intel_crtc);
 
 	if (HAS_PCH_CPT(dev)) {
 		u32 sel;
 
 		temp = I915_READ(PCH_DPLL_SEL);
-		switch (pipe) {
-		default:
-		case 0:
-			temp |= TRANSA_DPLL_ENABLE;
-			sel = TRANSA_DPLLB_SEL;
-			break;
-		case 1:
-			temp |= TRANSB_DPLL_ENABLE;
-			sel = TRANSB_DPLLB_SEL;
-			break;
-		case 2:
-			temp |= TRANSC_DPLL_ENABLE;
-			sel = TRANSC_DPLLB_SEL;
-			break;
-		}
-		if (intel_crtc->pch_pll->pll_reg == _PCH_DPLL_B)
+		temp |= TRANS_DPLL_ENABLE(pipe);
+		sel = TRANS_DPLLB_SEL(pipe);
+		if (intel_crtc->config.shared_dpll == DPLL_ID_PCH_PLL_B)
 			temp |= sel;
 		else
 			temp &= ~sel;
@@ -3148,14 +2951,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 
 	/* set transcoder timing, panel must allow it */
 	assert_panel_unlocked(dev_priv, pipe);
-	I915_WRITE(TRANS_HTOTAL(pipe), I915_READ(HTOTAL(pipe)));
-	I915_WRITE(TRANS_HBLANK(pipe), I915_READ(HBLANK(pipe)));
-	I915_WRITE(TRANS_HSYNC(pipe),  I915_READ(HSYNC(pipe)));
-
-	I915_WRITE(TRANS_VTOTAL(pipe), I915_READ(VTOTAL(pipe)));
-	I915_WRITE(TRANS_VBLANK(pipe), I915_READ(VBLANK(pipe)));
-	I915_WRITE(TRANS_VSYNC(pipe),  I915_READ(VSYNC(pipe)));
-	I915_WRITE(TRANS_VSYNCSHIFT(pipe),  I915_READ(VSYNCSHIFT(pipe)));
+	ironlake_pch_transcoder_set_timings(intel_crtc, pipe);
 
 	intel_fdi_normal_train(crtc);
 
@@ -3205,86 +3001,82 @@ static void lpt_pch_enable(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
-	assert_transcoder_disabled(dev_priv, TRANSCODER_A);
+	assert_pch_transcoder_disabled(dev_priv, TRANSCODER_A);
 
 	lpt_program_iclkip(crtc);
 
 	/* Set transcoder timing. */
-	I915_WRITE(_TRANS_HTOTAL_A, I915_READ(HTOTAL(cpu_transcoder)));
-	I915_WRITE(_TRANS_HBLANK_A, I915_READ(HBLANK(cpu_transcoder)));
-	I915_WRITE(_TRANS_HSYNC_A,  I915_READ(HSYNC(cpu_transcoder)));
-
-	I915_WRITE(_TRANS_VTOTAL_A, I915_READ(VTOTAL(cpu_transcoder)));
-	I915_WRITE(_TRANS_VBLANK_A, I915_READ(VBLANK(cpu_transcoder)));
-	I915_WRITE(_TRANS_VSYNC_A,  I915_READ(VSYNC(cpu_transcoder)));
-	I915_WRITE(_TRANS_VSYNCSHIFT_A, I915_READ(VSYNCSHIFT(cpu_transcoder)));
+	ironlake_pch_transcoder_set_timings(intel_crtc, PIPE_A);
 
 	lpt_enable_pch_transcoder(dev_priv, cpu_transcoder);
 }
 
-static void intel_put_pch_pll(struct intel_crtc *intel_crtc)
+static void intel_put_shared_dpll(struct intel_crtc *crtc)
 {
-	struct intel_pch_pll *pll = intel_crtc->pch_pll;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
 	if (pll == NULL)
 		return;
 
 	if (pll->refcount == 0) {
-		WARN(1, "bad PCH PLL refcount\n");
+		WARN(1, "bad %s refcount\n", pll->name);
 		return;
 	}
 
-	--pll->refcount;
-	intel_crtc->pch_pll = NULL;
+	if (--pll->refcount == 0) {
+		WARN_ON(pll->on);
+		WARN_ON(pll->active);
+	}
+
+	crtc->config.shared_dpll = DPLL_ID_PRIVATE;
 }
 
-static struct intel_pch_pll *intel_get_pch_pll(struct intel_crtc *intel_crtc, u32 dpll, u32 fp)
+static struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc, u32 dpll, u32 fp)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	struct intel_pch_pll *pll;
-	int i;
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
+	enum intel_dpll_id i;
 
-	pll = intel_crtc->pch_pll;
 	if (pll) {
-		DRM_DEBUG_KMS("CRTC:%d reusing existing PCH PLL %x\n",
-			      intel_crtc->base.base.id, pll->pll_reg);
-		goto prepare;
+		DRM_DEBUG_KMS("CRTC:%d dropping existing %s\n",
+			      crtc->base.base.id, pll->name);
+		intel_put_shared_dpll(crtc);
 	}
 
 	if (HAS_PCH_IBX(dev_priv->dev)) {
 		/* Ironlake PCH has a fixed PLL->PCH pipe mapping. */
-		i = intel_crtc->pipe;
-		pll = &dev_priv->pch_plls[i];
+		i = crtc->pipe;
+		pll = &dev_priv->shared_dplls[i];
 
-		DRM_DEBUG_KMS("CRTC:%d using pre-allocated PCH PLL %x\n",
-			      intel_crtc->base.base.id, pll->pll_reg);
+		DRM_DEBUG_KMS("CRTC:%d using pre-allocated %s\n",
+			      crtc->base.base.id, pll->name);
 
 		goto found;
 	}
 
-	for (i = 0; i < dev_priv->num_pch_pll; i++) {
-		pll = &dev_priv->pch_plls[i];
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
 
 		/* Only want to check enabled timings first */
 		if (pll->refcount == 0)
 			continue;
 
-		if (dpll == (I915_READ(pll->pll_reg) & 0x7fffffff) &&
-		    fp == I915_READ(pll->fp0_reg)) {
-			DRM_DEBUG_KMS("CRTC:%d sharing existing PCH PLL %x (refcount %d, ative %d)\n",
-				      intel_crtc->base.base.id,
-				      pll->pll_reg, pll->refcount, pll->active);
+		if (dpll == (I915_READ(PCH_DPLL(pll->id)) & 0x7fffffff) &&
+		    fp == I915_READ(PCH_FP0(pll->id))) {
+			DRM_DEBUG_KMS("CRTC:%d sharing existing %s (refcount %d, ative %d)\n",
+				      crtc->base.base.id,
+				      pll->name, pll->refcount, pll->active);
 
 			goto found;
 		}
 	}
 
 	/* Ok no matching timings, maybe there's a free one? */
-	for (i = 0; i < dev_priv->num_pch_pll; i++) {
-		pll = &dev_priv->pch_plls[i];
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
 		if (pll->refcount == 0) {
-			DRM_DEBUG_KMS("CRTC:%d allocated PCH PLL %x\n",
-				      intel_crtc->base.base.id, pll->pll_reg);
+			DRM_DEBUG_KMS("CRTC:%d allocated %s\n",
+				      crtc->base.base.id, pll->name);
 			goto found;
 		}
 	}
@@ -3292,24 +3084,32 @@ static struct intel_pch_pll *intel_get_pch_pll(struct intel_crtc *intel_crtc, u3
 	return NULL;
 
 found:
-	intel_crtc->pch_pll = pll;
-	pll->refcount++;
-	DRM_DEBUG_DRIVER("using pll %d for pipe %d\n", i, intel_crtc->pipe);
-prepare: /* separate function? */
-	DRM_DEBUG_DRIVER("switching PLL %x off\n", pll->pll_reg);
+	crtc->config.shared_dpll = i;
+	DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->name,
+			 pipe_name(crtc->pipe));
 
-	/* Wait for the clocks to stabilize before rewriting the regs */
-	I915_WRITE(pll->pll_reg, dpll & ~DPLL_VCO_ENABLE);
-	POSTING_READ(pll->pll_reg);
-	udelay(150);
+	if (pll->active == 0) {
+		memcpy(&pll->hw_state, &crtc->config.dpll_hw_state,
+		       sizeof(pll->hw_state));
+
+		DRM_DEBUG_DRIVER("setting up %s\n", pll->name);
+		WARN_ON(pll->on);
+		assert_shared_dpll_disabled(dev_priv, pll);
+
+		/* Wait for the clocks to stabilize before rewriting the regs */
+		I915_WRITE(PCH_DPLL(pll->id), dpll & ~DPLL_VCO_ENABLE);
+		POSTING_READ(PCH_DPLL(pll->id));
+		udelay(150);
+
+		I915_WRITE(PCH_FP0(pll->id), fp);
+		I915_WRITE(PCH_DPLL(pll->id), dpll & ~DPLL_VCO_ENABLE);
+	}
+	pll->refcount++;
 
-	I915_WRITE(pll->fp0_reg, fp);
-	I915_WRITE(pll->pll_reg, dpll & ~DPLL_VCO_ENABLE);
-	pll->on = false;
 	return pll;
 }
 
-void intel_cpt_verify_modeset(struct drm_device *dev, int pipe)
+static void cpt_verify_modeset(struct drm_device *dev, int pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int dslreg = PIPEDSL(pipe);
@@ -3319,10 +3119,53 @@ void intel_cpt_verify_modeset(struct drm_device *dev, int pipe)
 	udelay(500);
 	if (wait_for(I915_READ(dslreg) != temp, 5)) {
 		if (wait_for(I915_READ(dslreg) != temp, 5))
-			DRM_ERROR("mode set failed: pipe %d stuck\n", pipe);
+			DRM_ERROR("mode set failed: pipe %c stuck\n", pipe_name(pipe));
+	}
+}
+
+static void ironlake_pfit_enable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	if (crtc->config.pch_pfit.size) {
+		/* Force use of hard-coded filter coefficients
+		 * as some pre-programmed values are broken,
+		 * e.g. x201.
+		 */
+		if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
+			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3 |
+						 PF_PIPE_SEL_IVB(pipe));
+		else
+			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3);
+		I915_WRITE(PF_WIN_POS(pipe), crtc->config.pch_pfit.pos);
+		I915_WRITE(PF_WIN_SZ(pipe), crtc->config.pch_pfit.size);
 	}
 }
 
+static void intel_enable_planes(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	struct intel_plane *intel_plane;
+
+	list_for_each_entry(intel_plane, &dev->mode_config.plane_list, base.head)
+		if (intel_plane->pipe == pipe)
+			intel_plane_restore(&intel_plane->base);
+}
+
+static void intel_disable_planes(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	struct intel_plane *intel_plane;
+
+	list_for_each_entry(intel_plane, &dev->mode_config.plane_list, base.head)
+		if (intel_plane->pipe == pipe)
+			intel_plane_disable(&intel_plane->base);
+}
+
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3339,6 +3182,10 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 		return;
 
 	intel_crtc->active = true;
+
+	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+	intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
+
 	intel_update_watermarks(dev);
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
@@ -3362,22 +3209,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 		if (encoder->pre_enable)
 			encoder->pre_enable(encoder);
 
-	/* Enable panel fitting for LVDS */
-	if (dev_priv->pch_pf_size &&
-	    (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) ||
-	     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
-		/* Force use of hard-coded filter coefficients
-		 * as some pre-programmed values are broken,
-		 * e.g. x201.
-		 */
-		if (IS_IVYBRIDGE(dev))
-			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3 |
-						 PF_PIPE_SEL_IVB(pipe));
-		else
-			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3);
-		I915_WRITE(PF_WIN_POS(pipe), dev_priv->pch_pf_pos);
-		I915_WRITE(PF_WIN_SZ(pipe), dev_priv->pch_pf_size);
-	}
+	ironlake_pfit_enable(intel_crtc);
 
 	/*
 	 * On ILK+ LUT must be loaded before the pipe is running but with
@@ -3388,6 +3220,8 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	intel_enable_pipe(dev_priv, pipe,
 			  intel_crtc->config.has_pch_encoder);
 	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	intel_crtc_update_cursor(crtc, true);
 
 	if (intel_crtc->config.has_pch_encoder)
 		ironlake_pch_enable(crtc);
@@ -3396,13 +3230,11 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	intel_update_fbc(dev);
 	mutex_unlock(&dev->struct_mutex);
 
-	intel_crtc_update_cursor(crtc, true);
-
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
 
 	if (HAS_PCH_CPT(dev))
-		intel_cpt_verify_modeset(dev, intel_crtc->pipe);
+		cpt_verify_modeset(dev, intel_crtc->pipe);
 
 	/*
 	 * There seems to be a race in PCH platform hw (at least on some
@@ -3415,6 +3247,42 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	intel_wait_for_vblank(dev, intel_crtc->pipe);
 }
 
+/* IPS only exists on ULT machines and is tied to pipe A. */
+static bool hsw_crtc_supports_ips(struct intel_crtc *crtc)
+{
+	return HAS_IPS(crtc->base.dev) && crtc->pipe == PIPE_A;
+}
+
+static void hsw_enable_ips(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+
+	if (!crtc->config.ips_enabled)
+		return;
+
+	/* We can only enable IPS after we enable a plane and wait for a vblank.
+	 * We guarantee that the plane is enabled by calling intel_enable_ips
+	 * only after intel_enable_plane. And intel_enable_plane already waits
+	 * for a vblank, so all we need to do here is to enable the IPS bit. */
+	assert_plane_enabled(dev_priv, crtc->plane);
+	I915_WRITE(IPS_CTL, IPS_ENABLE);
+}
+
+static void hsw_disable_ips(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!crtc->config.ips_enabled)
+		return;
+
+	assert_plane_enabled(dev_priv, crtc->plane);
+	I915_WRITE(IPS_CTL, 0);
+
+	/* We need to wait for a vblank before we can disable the plane. */
+	intel_wait_for_vblank(dev, crtc->pipe);
+}
+
 static void haswell_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3430,6 +3298,11 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 		return;
 
 	intel_crtc->active = true;
+
+	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+	if (intel_crtc->config.has_pch_encoder)
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
+
 	intel_update_watermarks(dev);
 
 	if (intel_crtc->config.has_pch_encoder)
@@ -3441,18 +3314,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 
 	intel_ddi_enable_pipe_clock(intel_crtc);
 
-	/* Enable panel fitting for eDP */
-	if (dev_priv->pch_pf_size &&
-	    intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP)) {
-		/* Force use of hard-coded filter coefficients
-		 * as some pre-programmed values are broken,
-		 * e.g. x201.
-		 */
-		I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3 |
-					 PF_PIPE_SEL_IVB(pipe));
-		I915_WRITE(PF_WIN_POS(pipe), dev_priv->pch_pf_pos);
-		I915_WRITE(PF_WIN_SZ(pipe), dev_priv->pch_pf_size);
-	}
+	ironlake_pfit_enable(intel_crtc);
 
 	/*
 	 * On ILK+ LUT must be loaded before the pipe is running but with
@@ -3466,6 +3328,10 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_enable_pipe(dev_priv, pipe,
 			  intel_crtc->config.has_pch_encoder);
 	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	intel_crtc_update_cursor(crtc, true);
+
+	hsw_enable_ips(intel_crtc);
 
 	if (intel_crtc->config.has_pch_encoder)
 		lpt_pch_enable(crtc);
@@ -3474,8 +3340,6 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_update_fbc(dev);
 	mutex_unlock(&dev->struct_mutex);
 
-	intel_crtc_update_cursor(crtc, true);
-
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
 
@@ -3490,6 +3354,21 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_wait_for_vblank(dev, intel_crtc->pipe);
 }
 
+static void ironlake_pfit_disable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	/* To avoid upsetting the power well on haswell only disable the pfit if
+	 * it's in use. The hw state code will make sure we get this right. */
+	if (crtc->config.pch_pfit.size) {
+		I915_WRITE(PF_CTL(pipe), 0);
+		I915_WRITE(PF_WIN_POS(pipe), 0);
+		I915_WRITE(PF_WIN_SZ(pipe), 0);
+	}
+}
+
 static void ironlake_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3509,58 +3388,51 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 
 	intel_crtc_wait_for_pending_flips(crtc);
 	drm_vblank_off(dev, pipe);
-	intel_crtc_update_cursor(crtc, false);
-
-	intel_disable_plane(dev_priv, plane, pipe);
 
 	if (dev_priv->cfb_plane == plane)
 		intel_disable_fbc(dev);
 
+	intel_crtc_update_cursor(crtc, false);
+	intel_disable_planes(crtc);
+	intel_disable_plane(dev_priv, plane, pipe);
+
+	if (intel_crtc->config.has_pch_encoder)
+		intel_set_pch_fifo_underrun_reporting(dev, pipe, false);
+
 	intel_disable_pipe(dev_priv, pipe);
 
-	/* Disable PF */
-	I915_WRITE(PF_CTL(pipe), 0);
-	I915_WRITE(PF_WIN_SZ(pipe), 0);
+	ironlake_pfit_disable(intel_crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->post_disable)
 			encoder->post_disable(encoder);
 
-	ironlake_fdi_disable(crtc);
+	if (intel_crtc->config.has_pch_encoder) {
+		ironlake_fdi_disable(crtc);
 
-	ironlake_disable_pch_transcoder(dev_priv, pipe);
+		ironlake_disable_pch_transcoder(dev_priv, pipe);
+		intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
 
-	if (HAS_PCH_CPT(dev)) {
-		/* disable TRANS_DP_CTL */
-		reg = TRANS_DP_CTL(pipe);
-		temp = I915_READ(reg);
-		temp &= ~(TRANS_DP_OUTPUT_ENABLE | TRANS_DP_PORT_SEL_MASK);
-		temp |= TRANS_DP_PORT_SEL_NONE;
-		I915_WRITE(reg, temp);
-
-		/* disable DPLL_SEL */
-		temp = I915_READ(PCH_DPLL_SEL);
-		switch (pipe) {
-		case 0:
-			temp &= ~(TRANSA_DPLL_ENABLE | TRANSA_DPLLB_SEL);
-			break;
-		case 1:
-			temp &= ~(TRANSB_DPLL_ENABLE | TRANSB_DPLLB_SEL);
-			break;
-		case 2:
-			/* C shares PLL A or B */
-			temp &= ~(TRANSC_DPLL_ENABLE | TRANSC_DPLLB_SEL);
-			break;
-		default:
-			BUG(); /* wtf */
+		if (HAS_PCH_CPT(dev)) {
+			/* disable TRANS_DP_CTL */
+			reg = TRANS_DP_CTL(pipe);
+			temp = I915_READ(reg);
+			temp &= ~(TRANS_DP_OUTPUT_ENABLE |
+				  TRANS_DP_PORT_SEL_MASK);
+			temp |= TRANS_DP_PORT_SEL_NONE;
+			I915_WRITE(reg, temp);
+
+			/* disable DPLL_SEL */
+			temp = I915_READ(PCH_DPLL_SEL);
+			temp &= ~(TRANS_DPLL_ENABLE(pipe) | TRANS_DPLLB_SEL(pipe));
+			I915_WRITE(PCH_DPLL_SEL, temp);
 		}
-		I915_WRITE(PCH_DPLL_SEL, temp);
-	}
 
-	/* disable PCH DPLL */
-	intel_disable_pch_pll(intel_crtc);
+		/* disable PCH DPLL */
+		intel_disable_shared_dpll(intel_crtc);
 
-	ironlake_fdi_pll_disable(intel_crtc);
+		ironlake_fdi_pll_disable(intel_crtc);
+	}
 
 	intel_crtc->active = false;
 	intel_update_watermarks(dev);
@@ -3588,24 +3460,24 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	intel_crtc_wait_for_pending_flips(crtc);
 	drm_vblank_off(dev, pipe);
-	intel_crtc_update_cursor(crtc, false);
-
-	intel_disable_plane(dev_priv, plane, pipe);
 
+	/* FBC must be disabled before disabling the plane on HSW. */
 	if (dev_priv->cfb_plane == plane)
 		intel_disable_fbc(dev);
 
+	hsw_disable_ips(intel_crtc);
+
+	intel_crtc_update_cursor(crtc, false);
+	intel_disable_planes(crtc);
+	intel_disable_plane(dev_priv, plane, pipe);
+
+	if (intel_crtc->config.has_pch_encoder)
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, false);
 	intel_disable_pipe(dev_priv, pipe);
 
 	intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
 
-	/* XXX: Once we have proper panel fitter state tracking implemented with
-	 * hardware state read/check support we should switch to only disable
-	 * the panel fitter when we know it's used. */
-	if (intel_using_power_well(dev)) {
-		I915_WRITE(PF_CTL(pipe), 0);
-		I915_WRITE(PF_WIN_SZ(pipe), 0);
-	}
+	ironlake_pfit_disable(intel_crtc);
 
 	intel_ddi_disable_pipe_clock(intel_crtc);
 
@@ -3615,6 +3487,7 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	if (intel_crtc->config.has_pch_encoder) {
 		lpt_disable_pch_transcoder(dev_priv);
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
 		intel_ddi_fdi_disable(crtc);
 	}
 
@@ -3629,17 +3502,11 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 static void ironlake_crtc_off(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	intel_put_pch_pll(intel_crtc);
+	intel_put_shared_dpll(intel_crtc);
 }
 
 static void haswell_crtc_off(struct drm_crtc *crtc)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	/* Stop saying we're using TRANSCODER_EDP because some other CRTC might
-	 * start using it. */
-	intel_crtc->config.cpu_transcoder = (enum transcoder) intel_crtc->pipe;
-
 	intel_ddi_put_crtc_pll(crtc);
 }
 
@@ -3685,6 +3552,77 @@ g4x_fixup_plane(struct drm_i915_private *dev_priv, enum pipe pipe)
 	}
 }
 
+static void i9xx_pfit_enable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc_config *pipe_config = &crtc->config;
+
+	if (!crtc->config.gmch_pfit.control)
+		return;
+
+	/*
+	 * The panel fitter should only be adjusted whilst the pipe is disabled,
+	 * according to register description and PRM.
+	 */
+	WARN_ON(I915_READ(PFIT_CONTROL) & PFIT_ENABLE);
+	assert_pipe_disabled(dev_priv, crtc->pipe);
+
+	I915_WRITE(PFIT_PGM_RATIOS, pipe_config->gmch_pfit.pgm_ratios);
+	I915_WRITE(PFIT_CONTROL, pipe_config->gmch_pfit.control);
+
+	/* Border color in case we don't scale up to the full screen. Black by
+	 * default, change to something else for debugging. */
+	I915_WRITE(BCLRPAT(crtc->pipe), 0);
+}
+
+static void valleyview_crtc_enable(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_encoder *encoder;
+	int pipe = intel_crtc->pipe;
+	int plane = intel_crtc->plane;
+
+	WARN_ON(!crtc->enabled);
+
+	if (intel_crtc->active)
+		return;
+
+	intel_crtc->active = true;
+	intel_update_watermarks(dev);
+
+	mutex_lock(&dev_priv->dpio_lock);
+
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
+	intel_enable_pll(dev_priv, pipe);
+
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_enable)
+			encoder->pre_enable(encoder);
+
+	/* VLV wants encoder enabling _before_ the pipe is up. */
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		encoder->enable(encoder);
+
+	i9xx_pfit_enable(intel_crtc);
+
+	intel_crtc_load_lut(crtc);
+
+	intel_enable_pipe(dev_priv, pipe, false);
+	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	intel_crtc_update_cursor(crtc, true);
+
+	intel_update_fbc(dev);
+
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
 static void i9xx_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3708,17 +3646,22 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 		if (encoder->pre_enable)
 			encoder->pre_enable(encoder);
 
+	i9xx_pfit_enable(intel_crtc);
+
+	intel_crtc_load_lut(crtc);
+
 	intel_enable_pipe(dev_priv, pipe, false);
 	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	/* The fixup needs to happen before cursor is enabled */
 	if (IS_G4X(dev))
 		g4x_fixup_plane(dev_priv, pipe);
-
-	intel_crtc_load_lut(crtc);
-	intel_update_fbc(dev);
+	intel_crtc_update_cursor(crtc, true);
 
 	/* Give the overlay scaler a chance to enable if it's on this pipe */
 	intel_crtc_dpms_overlay(intel_crtc, true);
-	intel_crtc_update_cursor(crtc, true);
+
+	intel_update_fbc(dev);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
@@ -3728,20 +3671,15 @@ static void i9xx_pfit_disable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum pipe pipe;
-	uint32_t pctl = I915_READ(PFIT_CONTROL);
 
-	assert_pipe_disabled(dev_priv, crtc->pipe);
+	if (!crtc->config.gmch_pfit.control)
+		return;
 
-	if (INTEL_INFO(dev)->gen >= 4)
-		pipe = (pctl & PFIT_PIPE_MASK) >> PFIT_PIPE_SHIFT;
-	else
-		pipe = PIPE_B;
+	assert_pipe_disabled(dev_priv, crtc->pipe);
 
-	if (pipe == crtc->pipe) {
-		DRM_DEBUG_DRIVER("disabling pfit, current: 0x%08x\n", pctl);
-		I915_WRITE(PFIT_CONTROL, 0);
-	}
+	DRM_DEBUG_DRIVER("disabling pfit, current: 0x%08x\n",
+			 I915_READ(PFIT_CONTROL));
+	I915_WRITE(PFIT_CONTROL, 0);
 }
 
 static void i9xx_crtc_disable(struct drm_crtc *crtc)
@@ -3762,17 +3700,23 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	/* Give the overlay scaler a chance to disable if it's on this pipe */
 	intel_crtc_wait_for_pending_flips(crtc);
 	drm_vblank_off(dev, pipe);
-	intel_crtc_dpms_overlay(intel_crtc, false);
-	intel_crtc_update_cursor(crtc, false);
 
 	if (dev_priv->cfb_plane == plane)
 		intel_disable_fbc(dev);
 
+	intel_crtc_dpms_overlay(intel_crtc, false);
+	intel_crtc_update_cursor(crtc, false);
+	intel_disable_planes(crtc);
 	intel_disable_plane(dev_priv, plane, pipe);
+
 	intel_disable_pipe(dev_priv, pipe);
 
 	i9xx_pfit_disable(intel_crtc);
 
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->post_disable)
+			encoder->post_disable(encoder);
+
 	intel_disable_pll(dev_priv, pipe);
 
 	intel_crtc->active = false;
@@ -3845,8 +3789,8 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	/* crtc should still be enabled when we disable it. */
 	WARN_ON(!crtc->enabled);
 
-	intel_crtc->eld_vld = false;
 	dev_priv->display.crtc_disable(crtc);
+	intel_crtc->eld_vld = false;
 	intel_crtc_update_sarea(crtc, false);
 	dev_priv->display.off(crtc);
 
@@ -3977,17 +3921,131 @@ bool intel_connector_get_hw_state(struct intel_connector *connector)
 	return encoder->get_hw_state(encoder, &pipe);
 }
 
-static bool intel_crtc_compute_config(struct drm_crtc *crtc,
-				      struct intel_crtc_config *pipe_config)
+static bool ironlake_check_fdi_lanes(struct drm_device *dev, enum pipe pipe,
+				     struct intel_crtc_config *pipe_config)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *pipe_B_crtc =
+		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_B]);
+
+	DRM_DEBUG_KMS("checking fdi config on pipe %c, lanes %i\n",
+		      pipe_name(pipe), pipe_config->fdi_lanes);
+	if (pipe_config->fdi_lanes > 4) {
+		DRM_DEBUG_KMS("invalid fdi lane config on pipe %c: %i lanes\n",
+			      pipe_name(pipe), pipe_config->fdi_lanes);
+		return false;
+	}
+
+	if (IS_HASWELL(dev)) {
+		if (pipe_config->fdi_lanes > 2) {
+			DRM_DEBUG_KMS("only 2 lanes on haswell, required: %i lanes\n",
+				      pipe_config->fdi_lanes);
+			return false;
+		} else {
+			return true;
+		}
+	}
+
+	if (INTEL_INFO(dev)->num_pipes == 2)
+		return true;
+
+	/* Ivybridge 3 pipe is really complicated */
+	switch (pipe) {
+	case PIPE_A:
+		return true;
+	case PIPE_B:
+		if (dev_priv->pipe_to_crtc_mapping[PIPE_C]->enabled &&
+		    pipe_config->fdi_lanes > 2) {
+			DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %c: %i lanes\n",
+				      pipe_name(pipe), pipe_config->fdi_lanes);
+			return false;
+		}
+		return true;
+	case PIPE_C:
+		if (!pipe_has_enabled_pch(pipe_B_crtc) ||
+		    pipe_B_crtc->config.fdi_lanes <= 2) {
+			if (pipe_config->fdi_lanes > 2) {
+				DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %c: %i lanes\n",
+					      pipe_name(pipe), pipe_config->fdi_lanes);
+				return false;
+			}
+		} else {
+			DRM_DEBUG_KMS("fdi link B uses too many lanes to enable link C\n");
+			return false;
+		}
+		return true;
+	default:
+		BUG();
+	}
+}
+
+#define RETRY 1
+static int ironlake_fdi_compute_config(struct intel_crtc *intel_crtc,
+				       struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	int lane, link_bw, fdi_dotclock;
+	bool setup_ok, needs_recompute = false;
+
+retry:
+	/* FDI is a binary signal running at ~2.7GHz, encoding
+	 * each output octet as 10 bits. The actual frequency
+	 * is stored as a divider into a 100MHz clock, and the
+	 * mode pixel clock is stored in units of 1KHz.
+	 * Hence the bw of each lane in terms of the mode signal
+	 * is:
+	 */
+	link_bw = intel_fdi_link_freq(dev) * MHz(100)/KHz(1)/10;
+
+	fdi_dotclock = adjusted_mode->clock;
+	fdi_dotclock /= pipe_config->pixel_multiplier;
+
+	lane = ironlake_get_lanes_required(fdi_dotclock, link_bw,
+					   pipe_config->pipe_bpp);
+
+	pipe_config->fdi_lanes = lane;
+
+	intel_link_compute_m_n(pipe_config->pipe_bpp, lane, fdi_dotclock,
+			       link_bw, &pipe_config->fdi_m_n);
+
+	setup_ok = ironlake_check_fdi_lanes(intel_crtc->base.dev,
+					    intel_crtc->pipe, pipe_config);
+	if (!setup_ok && pipe_config->pipe_bpp > 6*3) {
+		pipe_config->pipe_bpp -= 2*3;
+		DRM_DEBUG_KMS("fdi link bw constraint, reducing pipe bpp to %i\n",
+			      pipe_config->pipe_bpp);
+		needs_recompute = true;
+		pipe_config->bw_constrained = true;
+
+		goto retry;
+	}
+
+	if (needs_recompute)
+		return RETRY;
+
+	return setup_ok ? 0 : -EINVAL;
+}
+
+static void hsw_compute_ips_config(struct intel_crtc *crtc,
+				   struct intel_crtc_config *pipe_config)
+{
+	pipe_config->ips_enabled = i915_enable_ips &&
+				   hsw_crtc_supports_ips(crtc) &&
+				   pipe_config->pipe_bpp == 24;
+}
+
+static int intel_crtc_compute_config(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		/* FDI link clock is fixed at 2.7G */
 		if (pipe_config->requested_mode.clock * 3
 		    > IRONLAKE_FDI_FREQ * 4)
-			return false;
+			return -EINVAL;
 	}
 
 	/* All interlaced capable intel hw wants timings in frames. Note though
@@ -3996,12 +4054,12 @@ static bool intel_crtc_compute_config(struct drm_crtc *crtc,
 	if (!pipe_config->timings_set)
 		drm_mode_set_crtcinfo(adjusted_mode, 0);
 
-	/* WaPruneModeWithIncorrectHsyncOffset: Cantiga+ cannot handle modes
-	 * with a hsync front porch of 0.
+	/* Cantiga+ cannot handle modes with a hsync front porch of 0.
+	 * WaPruneModeWithIncorrectHsyncOffset:ctg,elk,ilk,snb,ivb,vlv,hsw.
 	 */
 	if ((INTEL_INFO(dev)->gen > 4 || IS_G4X(dev)) &&
 		adjusted_mode->hsync_start == adjusted_mode->hdisplay)
-		return false;
+		return -EINVAL;
 
 	if ((IS_G4X(dev) || IS_VALLEYVIEW(dev)) && pipe_config->pipe_bpp > 10*3) {
 		pipe_config->pipe_bpp = 10*3; /* 12bpc is gen5+ */
@@ -4011,7 +4069,18 @@ static bool intel_crtc_compute_config(struct drm_crtc *crtc,
 		pipe_config->pipe_bpp = 8*3;
 	}
 
-	return true;
+	if (HAS_IPS(dev))
+		hsw_compute_ips_config(crtc, pipe_config);
+
+	/* XXX: PCH clock sharing is done in ->mode_set, so make sure the old
+	 * clock survives for now. */
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		pipe_config->shared_dpll = crtc->config.shared_dpll;
+
+	if (pipe_config->has_pch_encoder)
+		return ironlake_fdi_compute_config(crtc, pipe_config);
+
+	return 0;
 }
 
 static int valleyview_get_display_clock_speed(struct drm_device *dev)
@@ -4120,7 +4189,7 @@ static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
 {
 	if (i915_panel_use_ssc >= 0)
 		return i915_panel_use_ssc != 0;
-	return dev_priv->lvds_use_ssc
+	return dev_priv->vbt.lvds_use_ssc
 		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
 }
 
@@ -4156,7 +4225,7 @@ static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
 		refclk = vlv_get_refclk(crtc);
 	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
 	    intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
-		refclk = dev_priv->lvds_ssc_freq * 1000;
+		refclk = dev_priv->vbt.lvds_ssc_freq * 1000;
 		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
 			      refclk / 1000);
 	} else if (!IS_GEN2(dev)) {
@@ -4168,28 +4237,14 @@ static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
 	return refclk;
 }
 
-static void i9xx_adjust_sdvo_tv_clock(struct intel_crtc *crtc)
+static uint32_t pnv_dpll_compute_fp(struct dpll *dpll)
 {
-	unsigned dotclock = crtc->config.adjusted_mode.clock;
-	struct dpll *clock = &crtc->config.dpll;
-
-	/* SDVO TV has fixed PLL values depend on its clock range,
-	   this mirrors vbios setting. */
-	if (dotclock >= 100000 && dotclock < 140500) {
-		clock->p1 = 2;
-		clock->p2 = 10;
-		clock->n = 3;
-		clock->m1 = 16;
-		clock->m2 = 8;
-	} else if (dotclock >= 140500 && dotclock <= 200000) {
-		clock->p1 = 1;
-		clock->p2 = 10;
-		clock->n = 6;
-		clock->m1 = 12;
-		clock->m2 = 8;
-	}
+	return (1 << dpll->n) << 16 | dpll->m2;
+}
 
-	crtc->config.clock_set = true;
+static uint32_t i9xx_dpll_compute_fp(struct dpll *dpll)
+{
+	return dpll->n << 16 | dpll->m1 << 8 | dpll->m2;
 }
 
 static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
@@ -4199,18 +4254,15 @@ static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe = crtc->pipe;
 	u32 fp, fp2 = 0;
-	struct dpll *clock = &crtc->config.dpll;
 
 	if (IS_PINEVIEW(dev)) {
-		fp = (1 << clock->n) << 16 | clock->m1 << 8 | clock->m2;
+		fp = pnv_dpll_compute_fp(&crtc->config.dpll);
 		if (reduced_clock)
-			fp2 = (1 << reduced_clock->n) << 16 |
-				reduced_clock->m1 << 8 | reduced_clock->m2;
+			fp2 = pnv_dpll_compute_fp(reduced_clock);
 	} else {
-		fp = clock->n << 16 | clock->m1 << 8 | clock->m2;
+		fp = i9xx_dpll_compute_fp(&crtc->config.dpll);
 		if (reduced_clock)
-			fp2 = reduced_clock->n << 16 | reduced_clock->m1 << 8 |
-				reduced_clock->m2;
+			fp2 = i9xx_dpll_compute_fp(reduced_clock);
 	}
 
 	I915_WRITE(FP0(pipe), fp);
@@ -4225,6 +4277,68 @@ static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 	}
 }
 
+static void vlv_pllb_recal_opamp(struct drm_i915_private *dev_priv)
+{
+	u32 reg_val;
+
+	/*
+	 * PLLB opamp always calibrates to max value of 0x3f, force enable it
+	 * and set it to a reasonable value instead.
+	 */
+	reg_val = vlv_dpio_read(dev_priv, DPIO_IREF(1));
+	reg_val &= 0xffffff00;
+	reg_val |= 0x00000030;
+	vlv_dpio_write(dev_priv, DPIO_IREF(1), reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, DPIO_CALIBRATION);
+	reg_val &= 0x8cffffff;
+	reg_val = 0x8c000000;
+	vlv_dpio_write(dev_priv, DPIO_CALIBRATION, reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, DPIO_IREF(1));
+	reg_val &= 0xffffff00;
+	vlv_dpio_write(dev_priv, DPIO_IREF(1), reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, DPIO_CALIBRATION);
+	reg_val &= 0x00ffffff;
+	reg_val |= 0xb0000000;
+	vlv_dpio_write(dev_priv, DPIO_CALIBRATION, reg_val);
+}
+
+static void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
+					 struct intel_link_m_n *m_n)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	I915_WRITE(PCH_TRANS_DATA_M1(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
+	I915_WRITE(PCH_TRANS_DATA_N1(pipe), m_n->gmch_n);
+	I915_WRITE(PCH_TRANS_LINK_M1(pipe), m_n->link_m);
+	I915_WRITE(PCH_TRANS_LINK_N1(pipe), m_n->link_n);
+}
+
+static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
+					 struct intel_link_m_n *m_n)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+	enum transcoder transcoder = crtc->config.cpu_transcoder;
+
+	if (INTEL_INFO(dev)->gen >= 5) {
+		I915_WRITE(PIPE_DATA_M1(transcoder), TU_SIZE(m_n->tu) | m_n->gmch_m);
+		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
+		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
+		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
+	} else {
+		I915_WRITE(PIPE_DATA_M_G4X(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
+		I915_WRITE(PIPE_DATA_N_G4X(pipe), m_n->gmch_n);
+		I915_WRITE(PIPE_LINK_M_G4X(pipe), m_n->link_m);
+		I915_WRITE(PIPE_LINK_N_G4X(pipe), m_n->link_n);
+	}
+}
+
 static void intel_dp_set_m_n(struct intel_crtc *crtc)
 {
 	if (crtc->config.has_pch_encoder)
@@ -4237,24 +4351,16 @@ static void vlv_update_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_encoder *encoder;
 	int pipe = crtc->pipe;
-	u32 dpll, mdiv, pdiv;
+	u32 dpll, mdiv;
 	u32 bestn, bestm1, bestm2, bestp1, bestp2;
-	bool is_sdvo;
-	u32 temp;
+	bool is_hdmi;
+	u32 coreclk, reg_val, dpll_md;
 
 	mutex_lock(&dev_priv->dpio_lock);
 
-	is_sdvo = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_SDVO) ||
-		intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
-
-	dpll = DPLL_VGA_MODE_DIS;
-	dpll |= DPLL_EXT_BUFFER_ENABLE_VLV;
-	dpll |= DPLL_REFA_CLK_ENABLE_VLV;
-	dpll |= DPLL_INTEGRATED_CLOCK_VLV;
-
-	I915_WRITE(DPLL(pipe), dpll);
-	POSTING_READ(DPLL(pipe));
+	is_hdmi = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
 
 	bestn = crtc->config.dpll.n;
 	bestm1 = crtc->config.dpll.m1;
@@ -4262,72 +4368,104 @@ static void vlv_update_pll(struct intel_crtc *crtc)
 	bestp1 = crtc->config.dpll.p1;
 	bestp2 = crtc->config.dpll.p2;
 
-	/*
-	 * In Valleyview PLL and program lane counter registers are exposed
-	 * through DPIO interface
-	 */
+	/* See eDP HDMI DPIO driver vbios notes doc */
+
+	/* PLL B needs special handling */
+	if (pipe)
+		vlv_pllb_recal_opamp(dev_priv);
+
+	/* Set up Tx target for periodic Rcomp update */
+	vlv_dpio_write(dev_priv, DPIO_IREF_BCAST, 0x0100000f);
+
+	/* Disable target IRef on PLL */
+	reg_val = vlv_dpio_read(dev_priv, DPIO_IREF_CTL(pipe));
+	reg_val &= 0x00ffffff;
+	vlv_dpio_write(dev_priv, DPIO_IREF_CTL(pipe), reg_val);
+
+	/* Disable fast lock */
+	vlv_dpio_write(dev_priv, DPIO_FASTCLK_DISABLE, 0x610);
+
+	/* Set idtafcrecal before PLL is enabled */
 	mdiv = ((bestm1 << DPIO_M1DIV_SHIFT) | (bestm2 & DPIO_M2DIV_MASK));
 	mdiv |= ((bestp1 << DPIO_P1_SHIFT) | (bestp2 << DPIO_P2_SHIFT));
 	mdiv |= ((bestn << DPIO_N_SHIFT));
-	mdiv |= (1 << DPIO_POST_DIV_SHIFT);
 	mdiv |= (1 << DPIO_K_SHIFT);
+
+	/*
+	 * Post divider depends on pixel clock rate, DAC vs digital (and LVDS,
+	 * but we don't support that).
+	 * Note: don't use the DAC post divider as it seems unstable.
+	 */
+	mdiv |= (DPIO_POST_DIV_HDMIDP << DPIO_POST_DIV_SHIFT);
+	vlv_dpio_write(dev_priv, DPIO_DIV(pipe), mdiv);
+
 	mdiv |= DPIO_ENABLE_CALIBRATION;
-	intel_dpio_write(dev_priv, DPIO_DIV(pipe), mdiv);
+	vlv_dpio_write(dev_priv, DPIO_DIV(pipe), mdiv);
+
+	/* Set HBR and RBR LPF coefficients */
+	if (crtc->config.port_clock == 162000 ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_ANALOG) ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI))
+		vlv_dpio_write(dev_priv, DPIO_LPF_COEFF(pipe),
+				 0x005f0021);
+	else
+		vlv_dpio_write(dev_priv, DPIO_LPF_COEFF(pipe),
+				 0x00d0000f);
+
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP) ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT)) {
+		/* Use SSC source */
+		if (!pipe)
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df40000);
+		else
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df70000);
+	} else { /* HDMI or VGA */
+		/* Use bend source */
+		if (!pipe)
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df70000);
+		else
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df40000);
+	}
 
-	intel_dpio_write(dev_priv, DPIO_CORE_CLK(pipe), 0x01000000);
+	coreclk = vlv_dpio_read(dev_priv, DPIO_CORE_CLK(pipe));
+	coreclk = (coreclk & 0x0000ff00) | 0x01c00000;
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT) ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP))
+		coreclk |= 0x01000000;
+	vlv_dpio_write(dev_priv, DPIO_CORE_CLK(pipe), coreclk);
 
-	pdiv = (1 << DPIO_REFSEL_OVERRIDE) | (5 << DPIO_PLL_MODESEL_SHIFT) |
-		(3 << DPIO_BIAS_CURRENT_CTL_SHIFT) | (1<<20) |
-		(7 << DPIO_PLL_REFCLK_SEL_SHIFT) | (8 << DPIO_DRIVER_CTL_SHIFT) |
-		(5 << DPIO_CLK_BIAS_CTL_SHIFT);
-	intel_dpio_write(dev_priv, DPIO_REFSFR(pipe), pdiv);
+	vlv_dpio_write(dev_priv, DPIO_PLL_CML(pipe), 0x87871000);
 
-	intel_dpio_write(dev_priv, DPIO_LFP_COEFF(pipe), 0x005f003b);
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
+	/* Enable DPIO clock input */
+	dpll = DPLL_EXT_BUFFER_ENABLE_VLV | DPLL_REFA_CLK_ENABLE_VLV |
+		DPLL_VGA_MODE_DIS | DPLL_INTEGRATED_CLOCK_VLV;
+	if (pipe)
+		dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
 
 	dpll |= DPLL_VCO_ENABLE;
 	I915_WRITE(DPLL(pipe), dpll);
 	POSTING_READ(DPLL(pipe));
+	udelay(150);
+
 	if (wait_for(((I915_READ(DPLL(pipe)) & DPLL_LOCK_VLV) == DPLL_LOCK_VLV), 1))
 		DRM_ERROR("DPLL %d failed to lock\n", pipe);
 
-	intel_dpio_write(dev_priv, DPIO_FASTCLK_DISABLE, 0x620);
+	dpll_md = (crtc->config.pixel_multiplier - 1)
+		<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+	I915_WRITE(DPLL_MD(pipe), dpll_md);
+	POSTING_READ(DPLL_MD(pipe));
 
 	if (crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(crtc);
 
-	I915_WRITE(DPLL(pipe), dpll);
-
-	/* Wait for the clocks to stabilize. */
-	POSTING_READ(DPLL(pipe));
-	udelay(150);
-
-	temp = 0;
-	if (is_sdvo) {
-		temp = 0;
-		if (crtc->config.pixel_multiplier > 1) {
-			temp = (crtc->config.pixel_multiplier - 1)
-				<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-		}
-	}
-	I915_WRITE(DPLL_MD(pipe), temp);
-	POSTING_READ(DPLL_MD(pipe));
-
-	/* Now program lane control registers */
-	if(intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT)
-	   || intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI)) {
-		temp = 0x1000C4;
-		if(pipe == 1)
-			temp |= (1 << 21);
-		intel_dpio_write(dev_priv, DPIO_DATA_CHANNEL1, temp);
-	}
-
-	if(intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP)) {
-		temp = 0x1000C4;
-		if(pipe == 1)
-			temp |= (1 << 21);
-		intel_dpio_write(dev_priv, DPIO_DATA_CHANNEL2, temp);
-	}
-
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
@@ -4355,14 +4493,14 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
 
-	if (is_sdvo) {
-		if ((crtc->config.pixel_multiplier > 1) &&
-		    (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))) {
-			dpll |= (crtc->config.pixel_multiplier - 1)
-				<< SDVO_MULTIPLIER_SHIFT_HIRES;
-		}
-		dpll |= DPLL_DVO_HIGH_SPEED;
+	if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) {
+		dpll |= (crtc->config.pixel_multiplier - 1)
+			<< SDVO_MULTIPLIER_SHIFT_HIRES;
 	}
+
+	if (is_sdvo)
+		dpll |= DPLL_DVO_HIGH_SPEED;
+
 	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT))
 		dpll |= DPLL_DVO_HIGH_SPEED;
 
@@ -4391,12 +4529,8 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	if (INTEL_INFO(dev)->gen >= 4)
 		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
 
-	if (is_sdvo && intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_TVOUT))
+	if (crtc->config.sdvo_tv_clock)
 		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_TVOUT))
-		/* XXX: just matching BIOS for now */
-		/*	dpll |= PLL_REF_INPUT_TVCLKINBC; */
-		dpll |= 3;
 	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
@@ -4422,15 +4556,9 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	udelay(150);
 
 	if (INTEL_INFO(dev)->gen >= 4) {
-		u32 temp = 0;
-		if (is_sdvo) {
-			temp = 0;
-			if (crtc->config.pixel_multiplier > 1) {
-				temp = (crtc->config.pixel_multiplier - 1)
-					<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-			}
-		}
-		I915_WRITE(DPLL_MD(pipe), temp);
+		u32 dpll_md = (crtc->config.pixel_multiplier - 1)
+			<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+		I915_WRITE(DPLL_MD(pipe), dpll_md);
 	} else {
 		/* The pixel multiplier can only be updated once the
 		 * DPLL is enabled and the clocks are stable.
@@ -4442,7 +4570,6 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 }
 
 static void i8xx_update_pll(struct intel_crtc *crtc,
-			    struct drm_display_mode *adjusted_mode,
 			    intel_clock_t *reduced_clock,
 			    int num_connectors)
 {
@@ -4497,20 +4624,26 @@ static void i8xx_update_pll(struct intel_crtc *crtc,
 	I915_WRITE(DPLL(pipe), dpll);
 }
 
-static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
-				   struct drm_display_mode *mode,
-				   struct drm_display_mode *adjusted_mode)
+static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
-	uint32_t vsyncshift;
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
+	uint32_t vsyncshift, crtc_vtotal, crtc_vblank_end;
+
+	/* We need to be careful not to changed the adjusted mode, for otherwise
+	 * the hw state checker will get angry at the mismatch. */
+	crtc_vtotal = adjusted_mode->crtc_vtotal;
+	crtc_vblank_end = adjusted_mode->crtc_vblank_end;
 
 	if (!IS_GEN2(dev) && adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
 		/* the chip adds 2 halflines automatically */
-		adjusted_mode->crtc_vtotal -= 1;
-		adjusted_mode->crtc_vblank_end -= 1;
+		crtc_vtotal -= 1;
+		crtc_vblank_end -= 1;
 		vsyncshift = adjusted_mode->crtc_hsync_start
 			     - adjusted_mode->crtc_htotal / 2;
 	} else {
@@ -4532,10 +4665,10 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
 
 	I915_WRITE(VTOTAL(cpu_transcoder),
 		   (adjusted_mode->crtc_vdisplay - 1) |
-		   ((adjusted_mode->crtc_vtotal - 1) << 16));
+		   ((crtc_vtotal - 1) << 16));
 	I915_WRITE(VBLANK(cpu_transcoder),
 		   (adjusted_mode->crtc_vblank_start - 1) |
-		   ((adjusted_mode->crtc_vblank_end - 1) << 16));
+		   ((crtc_vblank_end - 1) << 16));
 	I915_WRITE(VSYNC(cpu_transcoder),
 		   (adjusted_mode->crtc_vsync_start - 1) |
 		   ((adjusted_mode->crtc_vsync_end - 1) << 16));
@@ -4555,13 +4688,52 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
 		   ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1));
 }
 
+static void intel_get_pipe_timings(struct intel_crtc *crtc,
+				   struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
+	uint32_t tmp;
+
+	tmp = I915_READ(HTOTAL(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_hdisplay = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_htotal = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(HBLANK(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_hblank_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_hblank_end = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(HSYNC(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_hsync_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_hsync_end = ((tmp >> 16) & 0xffff) + 1;
+
+	tmp = I915_READ(VTOTAL(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_vdisplay = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_vtotal = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(VBLANK(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_vblank_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_vblank_end = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(VSYNC(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_vsync_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_vsync_end = ((tmp >> 16) & 0xffff) + 1;
+
+	if (I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_INTERLACE_MASK) {
+		pipe_config->adjusted_mode.flags |= DRM_MODE_FLAG_INTERLACE;
+		pipe_config->adjusted_mode.crtc_vtotal += 1;
+		pipe_config->adjusted_mode.crtc_vblank_end += 1;
+	}
+
+	tmp = I915_READ(PIPESRC(crtc->pipe));
+	pipe_config->requested_mode.vdisplay = (tmp & 0xffff) + 1;
+	pipe_config->requested_mode.hdisplay = ((tmp >> 16) & 0xffff) + 1;
+}
+
 static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t pipeconf;
 
-	pipeconf = I915_READ(PIPECONF(intel_crtc->pipe));
+	pipeconf = 0;
 
 	if (intel_crtc->pipe == 0 && INTEL_INFO(dev)->gen < 4) {
 		/* Enable pixel doubling when the dot clock is > 90% of the (display)
@@ -4573,26 +4745,28 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 		if (intel_crtc->config.requested_mode.clock >
 		    dev_priv->display.get_display_clock_speed(dev) * 9 / 10)
 			pipeconf |= PIPECONF_DOUBLE_WIDE;
-		else
-			pipeconf &= ~PIPECONF_DOUBLE_WIDE;
 	}
 
-	/* default to 8bpc */
-	pipeconf &= ~(PIPECONF_BPC_MASK | PIPECONF_DITHER_EN);
-	if (intel_crtc->config.has_dp_encoder) {
-		if (intel_crtc->config.dither) {
-			pipeconf |= PIPECONF_6BPC |
-				    PIPECONF_DITHER_EN |
+	/* only g4x and later have fancy bpc/dither controls */
+	if (IS_G4X(dev) || IS_VALLEYVIEW(dev)) {
+		/* Bspec claims that we can't use dithering for 30bpp pipes. */
+		if (intel_crtc->config.dither && intel_crtc->config.pipe_bpp != 30)
+			pipeconf |= PIPECONF_DITHER_EN |
 				    PIPECONF_DITHER_TYPE_SP;
-		}
-	}
 
-	if (IS_VALLEYVIEW(dev) && intel_pipe_has_type(&intel_crtc->base,
-						      INTEL_OUTPUT_EDP)) {
-		if (intel_crtc->config.dither) {
-			pipeconf |= PIPECONF_6BPC |
-					PIPECONF_ENABLE |
-					I965_PIPECONF_ACTIVE;
+		switch (intel_crtc->config.pipe_bpp) {
+		case 18:
+			pipeconf |= PIPECONF_6BPC;
+			break;
+		case 24:
+			pipeconf |= PIPECONF_8BPC;
+			break;
+		case 30:
+			pipeconf |= PIPECONF_10BPC;
+			break;
+		default:
+			/* Case prevented by intel_choose_pipe_bpp_dither. */
+			BUG();
 		}
 	}
 
@@ -4602,23 +4776,17 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
 		} else {
 			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
-			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
 		}
 	}
 
-	pipeconf &= ~PIPECONF_INTERLACE_MASK;
 	if (!IS_GEN2(dev) &&
 	    intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 	else
 		pipeconf |= PIPECONF_PROGRESSIVE;
 
-	if (IS_VALLEYVIEW(dev)) {
-		if (intel_crtc->config.limited_color_range)
-			pipeconf |= PIPECONF_COLOR_RANGE_SELECT;
-		else
-			pipeconf &= ~PIPECONF_COLOR_RANGE_SELECT;
-	}
+	if (IS_VALLEYVIEW(dev) && intel_crtc->config.limited_color_range)
+		pipeconf |= PIPECONF_COLOR_RANGE_SELECT;
 
 	I915_WRITE(PIPECONF(intel_crtc->pipe), pipeconf);
 	POSTING_READ(PIPECONF(intel_crtc->pipe));
@@ -4631,16 +4799,14 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
 	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int refclk, num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
 	u32 dspcntr;
-	bool ok, has_reduced_clock = false, is_sdvo = false;
-	bool is_lvds = false, is_tv = false;
+	bool ok, has_reduced_clock = false;
+	bool is_lvds = false;
 	struct intel_encoder *encoder;
 	const intel_limit_t *limit;
 	int ret;
@@ -4650,15 +4816,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_SDVO:
-		case INTEL_OUTPUT_HDMI:
-			is_sdvo = true;
-			if (encoder->needs_tv_clock)
-				is_tv = true;
-			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
-			break;
 		}
 
 		num_connectors++;
@@ -4672,9 +4829,10 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
-			     &clock);
-	if (!ok) {
+	ok = dev_priv->display.find_dpll(limit, crtc,
+					 intel_crtc->config.port_clock,
+					 refclk, NULL, &clock);
+	if (!ok && !intel_crtc->config.clock_set) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
 	}
@@ -4689,10 +4847,10 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		 * by using the FP0/FP1. In such case we will disable the LVDS
 		 * downclock feature.
 		*/
-		has_reduced_clock = limit->find_pll(limit, crtc,
+		has_reduced_clock =
+			dev_priv->display.find_dpll(limit, crtc,
 						    dev_priv->lvds_downclock,
-						    refclk,
-						    &clock,
+						    refclk, &clock,
 						    &reduced_clock);
 	}
 	/* Compat-code for transition, will disappear. */
@@ -4704,11 +4862,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		intel_crtc->config.dpll.p2 = clock.p2;
 	}
 
-	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(intel_crtc);
-
 	if (IS_GEN2(dev))
-		i8xx_update_pll(intel_crtc, adjusted_mode,
+		i8xx_update_pll(intel_crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
 				num_connectors);
 	else if (IS_VALLEYVIEW(dev))
@@ -4716,7 +4871,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	else
 		i9xx_update_pll(intel_crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
-				num_connectors);
+                                num_connectors);
 
 	/* Set up the display plane register */
 	dspcntr = DISPPLANE_GAMMA_ENABLE;
@@ -4728,10 +4883,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 			dspcntr |= DISPPLANE_SEL_PIPE_B;
 	}
 
-	DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B');
-	drm_mode_debug_printmodeline(mode);
-
-	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
+	intel_set_pipe_timings(intel_crtc);
 
 	/* pipesrc and dspsize control the size that is scaled from,
 	 * which should always be the user's requested size.
@@ -4743,10 +4895,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 
 	i9xx_set_pipeconf(intel_crtc);
 
-	intel_enable_pipe(dev_priv, pipe, false);
-
-	intel_wait_for_vblank(dev, pipe);
-
 	I915_WRITE(DSPCNTR(plane), dspcntr);
 	POSTING_READ(DSPCNTR(plane));
 
@@ -4757,6 +4905,36 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	return ret;
 }
 
+static void i9xx_get_pfit_config(struct intel_crtc *crtc,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PFIT_CONTROL);
+
+	if (INTEL_INFO(dev)->gen < 4) {
+		if (crtc->pipe != PIPE_B)
+			return;
+
+		/* gen2/3 store dither state in pfit control, needs to match */
+		pipe_config->gmch_pfit.control = tmp & PANEL_8TO6_DITHER_ENABLE;
+	} else {
+		if ((tmp & PFIT_PIPE_MASK) != (crtc->pipe << PFIT_PIPE_SHIFT))
+			return;
+	}
+
+	if (!(tmp & PFIT_ENABLE))
+		return;
+
+	pipe_config->gmch_pfit.control = I915_READ(PFIT_CONTROL);
+	pipe_config->gmch_pfit.pgm_ratios = I915_READ(PFIT_PGM_RATIOS);
+	if (INTEL_INFO(dev)->gen < 5)
+		pipe_config->gmch_pfit.lvds_border_bits =
+			I915_READ(LVDS) & LVDS_BORDER_ENABLE;
+}
+
 static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 				 struct intel_crtc_config *pipe_config)
 {
@@ -4764,10 +4942,34 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
+	pipe_config->cpu_transcoder = crtc->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
 	tmp = I915_READ(PIPECONF(crtc->pipe));
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
+	intel_get_pipe_timings(crtc, pipe_config);
+
+	i9xx_get_pfit_config(crtc, pipe_config);
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		tmp = I915_READ(DPLL_MD(crtc->pipe));
+		pipe_config->pixel_multiplier =
+			((tmp & DPLL_MD_UDI_MULTIPLIER_MASK)
+			 >> DPLL_MD_UDI_MULTIPLIER_SHIFT) + 1;
+	} else if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) {
+		tmp = I915_READ(DPLL(crtc->pipe));
+		pipe_config->pixel_multiplier =
+			((tmp & SDVO_MULTIPLIER_MASK)
+			 >> SDVO_MULTIPLIER_SHIFT_HIRES) + 1;
+	} else {
+		/* Note that on i915G/GM the pixel multiplier is in the sdvo
+		 * port and will be fixed up in the encoder->get_config
+		 * function. */
+		pipe_config->pixel_multiplier = 1;
+	}
+
 	return true;
 }
 
@@ -4779,7 +4981,6 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 	u32 val, final;
 	bool has_lvds = false;
 	bool has_cpu_edp = false;
-	bool has_pch_edp = false;
 	bool has_panel = false;
 	bool has_ck505 = false;
 	bool can_ssc = false;
@@ -4794,25 +4995,22 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 			break;
 		case INTEL_OUTPUT_EDP:
 			has_panel = true;
-			if (intel_encoder_is_pch_edp(&encoder->base))
-				has_pch_edp = true;
-			else
+			if (enc_to_dig_port(&encoder->base)->port == PORT_A)
 				has_cpu_edp = true;
 			break;
 		}
 	}
 
 	if (HAS_PCH_IBX(dev)) {
-		has_ck505 = dev_priv->display_clock_mode;
+		has_ck505 = dev_priv->vbt.display_clock_mode;
 		can_ssc = has_ck505;
 	} else {
 		has_ck505 = false;
 		can_ssc = true;
 	}
 
-	DRM_DEBUG_KMS("has_panel %d has_lvds %d has_pch_edp %d has_cpu_edp %d has_ck505 %d\n",
-		      has_panel, has_lvds, has_pch_edp, has_cpu_edp,
-		      has_ck505);
+	DRM_DEBUG_KMS("has_panel %d has_lvds %d has_ck505 %d\n",
+		      has_panel, has_lvds, has_ck505);
 
 	/* Ironlake: try to setup display ref clock before DPLL
 	 * enabling. This is only under driver's control after
@@ -5102,7 +5300,6 @@ static int ironlake_get_refclk(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *encoder;
-	struct intel_encoder *edp_encoder = NULL;
 	int num_connectors = 0;
 	bool is_lvds = false;
 
@@ -5111,34 +5308,28 @@ static int ironlake_get_refclk(struct drm_crtc *crtc)
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_EDP:
-			edp_encoder = encoder;
-			break;
 		}
 		num_connectors++;
 	}
 
 	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
 		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
-			      dev_priv->lvds_ssc_freq);
-		return dev_priv->lvds_ssc_freq * 1000;
+			      dev_priv->vbt.lvds_ssc_freq);
+		return dev_priv->vbt.lvds_ssc_freq * 1000;
 	}
 
 	return 120000;
 }
 
-static void ironlake_set_pipeconf(struct drm_crtc *crtc,
-				  struct drm_display_mode *adjusted_mode,
-				  bool dither)
+static void ironlake_set_pipeconf(struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	uint32_t val;
 
-	val = I915_READ(PIPECONF(pipe));
+	val = 0;
 
-	val &= ~PIPECONF_BPC_MASK;
 	switch (intel_crtc->config.pipe_bpp) {
 	case 18:
 		val |= PIPECONF_6BPC;
@@ -5157,20 +5348,16 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 		BUG();
 	}
 
-	val &= ~(PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_MASK);
-	if (dither)
+	if (intel_crtc->config.dither)
 		val |= (PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_SP);
 
-	val &= ~PIPECONF_INTERLACE_MASK;
-	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 		val |= PIPECONF_INTERLACED_ILK;
 	else
 		val |= PIPECONF_PROGRESSIVE;
 
 	if (intel_crtc->config.limited_color_range)
 		val |= PIPECONF_COLOR_RANGE_SELECT;
-	else
-		val &= ~PIPECONF_COLOR_RANGE_SELECT;
 
 	I915_WRITE(PIPECONF(pipe), val);
 	POSTING_READ(PIPECONF(pipe));
@@ -5240,33 +5427,31 @@ static void intel_set_pipe_csc(struct drm_crtc *crtc)
 	}
 }
 
-static void haswell_set_pipeconf(struct drm_crtc *crtc,
-				 struct drm_display_mode *adjusted_mode,
-				 bool dither)
+static void haswell_set_pipeconf(struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	uint32_t val;
 
-	val = I915_READ(PIPECONF(cpu_transcoder));
+	val = 0;
 
-	val &= ~(PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_MASK);
-	if (dither)
+	if (intel_crtc->config.dither)
 		val |= (PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_SP);
 
-	val &= ~PIPECONF_INTERLACE_MASK_HSW;
-	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 		val |= PIPECONF_INTERLACED_ILK;
 	else
 		val |= PIPECONF_PROGRESSIVE;
 
 	I915_WRITE(PIPECONF(cpu_transcoder), val);
 	POSTING_READ(PIPECONF(cpu_transcoder));
+
+	I915_WRITE(GAMMA_MODE(intel_crtc->pipe), GAMMA_MODE_MODE_8BIT);
+	POSTING_READ(GAMMA_MODE(intel_crtc->pipe));
 }
 
 static bool ironlake_compute_clocks(struct drm_crtc *crtc,
-				    struct drm_display_mode *adjusted_mode,
 				    intel_clock_t *clock,
 				    bool *has_reduced_clock,
 				    intel_clock_t *reduced_clock)
@@ -5276,22 +5461,13 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 	struct intel_encoder *intel_encoder;
 	int refclk;
 	const intel_limit_t *limit;
-	bool ret, is_sdvo = false, is_tv = false, is_lvds = false;
+	bool ret, is_lvds = false;
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
 		switch (intel_encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_SDVO:
-		case INTEL_OUTPUT_HDMI:
-			is_sdvo = true;
-			if (intel_encoder->needs_tv_clock)
-				is_tv = true;
-			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
-			break;
 		}
 	}
 
@@ -5303,8 +5479,9 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ret = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
-			      clock);
+	ret = dev_priv->display.find_dpll(limit, crtc,
+					  to_intel_crtc(crtc)->config.port_clock,
+					  refclk, NULL, clock);
 	if (!ret)
 		return false;
 
@@ -5315,16 +5492,13 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 		 * by using the FP0/FP1. In such case we will disable the LVDS
 		 * downclock feature.
 		*/
-		*has_reduced_clock = limit->find_pll(limit, crtc,
-						     dev_priv->lvds_downclock,
-						     refclk,
-						     clock,
-						     reduced_clock);
+		*has_reduced_clock =
+			dev_priv->display.find_dpll(limit, crtc,
+						    dev_priv->lvds_downclock,
+						    refclk, clock,
+						    reduced_clock);
 	}
 
-	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(to_intel_crtc(crtc));
-
 	return true;
 }
 
@@ -5346,65 +5520,25 @@ static void cpt_enable_fdi_bc_bifurcation(struct drm_device *dev)
 	POSTING_READ(SOUTH_CHICKEN1);
 }
 
-static bool ironlake_check_fdi_lanes(struct intel_crtc *intel_crtc)
+static void ivybridge_update_fdi_bc_bifurcation(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *pipe_B_crtc =
-		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_B]);
-
-	DRM_DEBUG_KMS("checking fdi config on pipe %i, lanes %i\n",
-		      intel_crtc->pipe, intel_crtc->fdi_lanes);
-	if (intel_crtc->fdi_lanes > 4) {
-		DRM_DEBUG_KMS("invalid fdi lane config on pipe %i: %i lanes\n",
-			      intel_crtc->pipe, intel_crtc->fdi_lanes);
-		/* Clamp lanes to avoid programming the hw with bogus values. */
-		intel_crtc->fdi_lanes = 4;
-
-		return false;
-	}
-
-	if (INTEL_INFO(dev)->num_pipes == 2)
-		return true;
 
 	switch (intel_crtc->pipe) {
 	case PIPE_A:
-		return true;
+		break;
 	case PIPE_B:
-		if (dev_priv->pipe_to_crtc_mapping[PIPE_C]->enabled &&
-		    intel_crtc->fdi_lanes > 2) {
-			DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %i: %i lanes\n",
-				      intel_crtc->pipe, intel_crtc->fdi_lanes);
-			/* Clamp lanes to avoid programming the hw with bogus values. */
-			intel_crtc->fdi_lanes = 2;
-
-			return false;
-		}
-
-		if (intel_crtc->fdi_lanes > 2)
+		if (intel_crtc->config.fdi_lanes > 2)
 			WARN_ON(I915_READ(SOUTH_CHICKEN1) & FDI_BC_BIFURCATION_SELECT);
 		else
 			cpt_enable_fdi_bc_bifurcation(dev);
 
-		return true;
+		break;
 	case PIPE_C:
-		if (!pipe_B_crtc->base.enabled || pipe_B_crtc->fdi_lanes <= 2) {
-			if (intel_crtc->fdi_lanes > 2) {
-				DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %i: %i lanes\n",
-					      intel_crtc->pipe, intel_crtc->fdi_lanes);
-				/* Clamp lanes to avoid programming the hw with bogus values. */
-				intel_crtc->fdi_lanes = 2;
-
-				return false;
-			}
-		} else {
-			DRM_DEBUG_KMS("fdi link B uses too many lanes to enable link C\n");
-			return false;
-		}
-
 		cpt_enable_fdi_bc_bifurcation(dev);
 
-		return true;
+		break;
 	default:
 		BUG();
 	}
@@ -5421,78 +5555,13 @@ int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp)
 	return bps / (link_bw * 8) + 1;
 }
 
-void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
-				  struct intel_link_m_n *m_n)
+static bool ironlake_needs_fb_cb_tune(struct dpll *dpll, int factor)
 {
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe = crtc->pipe;
-
-	I915_WRITE(TRANSDATA_M1(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
-	I915_WRITE(TRANSDATA_N1(pipe), m_n->gmch_n);
-	I915_WRITE(TRANSDPLINK_M1(pipe), m_n->link_m);
-	I915_WRITE(TRANSDPLINK_N1(pipe), m_n->link_n);
-}
-
-void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-				  struct intel_link_m_n *m_n)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe = crtc->pipe;
-	enum transcoder transcoder = crtc->config.cpu_transcoder;
-
-	if (INTEL_INFO(dev)->gen >= 5) {
-		I915_WRITE(PIPE_DATA_M1(transcoder), TU_SIZE(m_n->tu) | m_n->gmch_m);
-		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
-		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
-		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
-	} else {
-		I915_WRITE(PIPE_GMCH_DATA_M(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
-		I915_WRITE(PIPE_GMCH_DATA_N(pipe), m_n->gmch_n);
-		I915_WRITE(PIPE_DP_LINK_M(pipe), m_n->link_m);
-		I915_WRITE(PIPE_DP_LINK_N(pipe), m_n->link_n);
-	}
-}
-
-static void ironlake_fdi_set_m_n(struct drm_crtc *crtc)
-{
-	struct drm_device *dev = crtc->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
-	struct intel_link_m_n m_n = {0};
-	int target_clock, lane, link_bw;
-
-	/* FDI is a binary signal running at ~2.7GHz, encoding
-	 * each output octet as 10 bits. The actual frequency
-	 * is stored as a divider into a 100MHz clock, and the
-	 * mode pixel clock is stored in units of 1KHz.
-	 * Hence the bw of each lane in terms of the mode signal
-	 * is:
-	 */
-	link_bw = intel_fdi_link_freq(dev) * MHz(100)/KHz(1)/10;
-
-	if (intel_crtc->config.pixel_target_clock)
-		target_clock = intel_crtc->config.pixel_target_clock;
-	else
-		target_clock = adjusted_mode->clock;
-
-	lane = ironlake_get_lanes_required(target_clock, link_bw,
-					   intel_crtc->config.pipe_bpp);
-
-	intel_crtc->fdi_lanes = lane;
-
-	if (intel_crtc->config.pixel_multiplier > 1)
-		link_bw *= intel_crtc->config.pixel_multiplier;
-	intel_link_compute_m_n(intel_crtc->config.pipe_bpp, lane, target_clock,
-			       link_bw, &m_n);
-
-	intel_cpu_transcoder_set_m_n(intel_crtc, &m_n);
+	return i9xx_dpll_compute_m(dpll) < factor * dpll->n;
 }
 
 static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
-				      intel_clock_t *clock, u32 *fp,
+				      u32 *fp,
 				      intel_clock_t *reduced_clock, u32 *fp2)
 {
 	struct drm_crtc *crtc = &intel_crtc->base;
@@ -5501,7 +5570,7 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	struct intel_encoder *intel_encoder;
 	uint32_t dpll;
 	int factor, num_connectors = 0;
-	bool is_lvds = false, is_sdvo = false, is_tv = false;
+	bool is_lvds = false, is_sdvo = false;
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
 		switch (intel_encoder->type) {
@@ -5511,11 +5580,6 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		case INTEL_OUTPUT_SDVO:
 		case INTEL_OUTPUT_HDMI:
 			is_sdvo = true;
-			if (intel_encoder->needs_tv_clock)
-				is_tv = true;
-			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
 			break;
 		}
 
@@ -5526,13 +5590,13 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	factor = 21;
 	if (is_lvds) {
 		if ((intel_panel_use_ssc(dev_priv) &&
-		     dev_priv->lvds_ssc_freq == 100) ||
+		     dev_priv->vbt.lvds_ssc_freq == 100) ||
 		    (HAS_PCH_IBX(dev) && intel_is_dual_link_lvds(dev)))
 			factor = 25;
-	} else if (is_sdvo && is_tv)
+	} else if (intel_crtc->config.sdvo_tv_clock)
 		factor = 20;
 
-	if (clock->m < factor * clock->n)
+	if (ironlake_needs_fb_cb_tune(&intel_crtc->config.dpll, factor))
 		*fp |= FP_CB_TUNE;
 
 	if (fp2 && (reduced_clock->m < factor * reduced_clock->n))
@@ -5544,23 +5608,21 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		dpll |= DPLLB_MODE_LVDS;
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
-	if (is_sdvo) {
-		if (intel_crtc->config.pixel_multiplier > 1) {
-			dpll |= (intel_crtc->config.pixel_multiplier - 1)
-				<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
-		}
+
+	dpll |= (intel_crtc->config.pixel_multiplier - 1)
+		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
+
+	if (is_sdvo)
 		dpll |= DPLL_DVO_HIGH_SPEED;
-	}
-	if (intel_crtc->config.has_dp_encoder &&
-	    intel_crtc->config.has_pch_encoder)
+	if (intel_crtc->config.has_dp_encoder)
 		dpll |= DPLL_DVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
-	dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	dpll |= (1 << (intel_crtc->config.dpll.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
 	/* also FPA1 */
-	dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+	dpll |= (1 << (intel_crtc->config.dpll.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
 
-	switch (clock->p2) {
+	switch (intel_crtc->config.dpll.p2) {
 	case 5:
 		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
 		break;
@@ -5575,18 +5637,12 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		break;
 	}
 
-	if (is_sdvo && is_tv)
-		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (is_tv)
-		/* XXX: just matching BIOS for now */
-		/*	dpll |= PLL_REF_INPUT_TVCLKINBC; */
-		dpll |= 3;
-	else if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2)
+	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
 		dpll |= PLL_REF_INPUT_DREFCLK;
 
-	return dpll;
+	return dpll | DPLL_VCO_ENABLE;
 }
 
 static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
@@ -5596,19 +5652,16 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
-	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
-	u32 dpll, fp = 0, fp2 = 0;
+	u32 dpll = 0, fp = 0, fp2 = 0;
 	bool ok, has_reduced_clock = false;
 	bool is_lvds = false;
 	struct intel_encoder *encoder;
+	struct intel_shared_dpll *pll;
 	int ret;
-	bool dither, fdi_config_ok;
 
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
 		switch (encoder->type) {
@@ -5623,11 +5676,9 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	WARN(!(HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)),
 	     "Unexpected PCH type %d\n", INTEL_PCH_TYPE(dev));
 
-	intel_crtc->config.cpu_transcoder = pipe;
-
-	ok = ironlake_compute_clocks(crtc, adjusted_mode, &clock,
+	ok = ironlake_compute_clocks(crtc, &clock,
 				     &has_reduced_clock, &reduced_clock);
-	if (!ok) {
+	if (!ok && !intel_crtc->config.clock_set) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
 	}
@@ -5643,34 +5694,31 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	/* Ensure that the cursor is valid for the new mode before changing... */
 	intel_crtc_update_cursor(crtc, true);
 
-	/* determine panel color depth */
-	dither = intel_crtc->config.dither;
-	if (is_lvds && dev_priv->lvds_dither)
-		dither = true;
-
-	fp = clock.n << 16 | clock.m1 << 8 | clock.m2;
-	if (has_reduced_clock)
-		fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 |
-			reduced_clock.m2;
-
-	dpll = ironlake_compute_dpll(intel_crtc, &clock, &fp, &reduced_clock,
-				     has_reduced_clock ? &fp2 : NULL);
-
-	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
-	drm_mode_debug_printmodeline(mode);
-
 	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
 	if (intel_crtc->config.has_pch_encoder) {
-		struct intel_pch_pll *pll;
+		fp = i9xx_dpll_compute_fp(&intel_crtc->config.dpll);
+		if (has_reduced_clock)
+			fp2 = i9xx_dpll_compute_fp(&reduced_clock);
+
+		dpll = ironlake_compute_dpll(intel_crtc,
+					     &fp, &reduced_clock,
+					     has_reduced_clock ? &fp2 : NULL);
+
+		intel_crtc->config.dpll_hw_state.dpll = dpll;
+		intel_crtc->config.dpll_hw_state.fp0 = fp;
+		if (has_reduced_clock)
+			intel_crtc->config.dpll_hw_state.fp1 = fp2;
+		else
+			intel_crtc->config.dpll_hw_state.fp1 = fp;
 
-		pll = intel_get_pch_pll(intel_crtc, dpll, fp);
+		pll = intel_get_shared_dpll(intel_crtc, dpll, fp);
 		if (pll == NULL) {
-			DRM_DEBUG_DRIVER("failed to find PLL for pipe %d\n",
-					 pipe);
+			DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n",
+					 pipe_name(pipe));
 			return -EINVAL;
 		}
 	} else
-		intel_put_pch_pll(intel_crtc);
+		intel_put_shared_dpll(intel_crtc);
 
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
@@ -5679,11 +5727,18 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		if (encoder->pre_pll_enable)
 			encoder->pre_pll_enable(encoder);
 
-	if (intel_crtc->pch_pll) {
-		I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
+	if (is_lvds && has_reduced_clock && i915_powersave)
+		intel_crtc->lowfreq_avail = true;
+	else
+		intel_crtc->lowfreq_avail = false;
+
+	if (intel_crtc->config.has_pch_encoder) {
+		pll = intel_crtc_to_shared_dpll(intel_crtc);
+
+		I915_WRITE(PCH_DPLL(pll->id), dpll);
 
 		/* Wait for the clocks to stabilize. */
-		POSTING_READ(intel_crtc->pch_pll->pll_reg);
+		POSTING_READ(PCH_DPLL(pll->id));
 		udelay(150);
 
 		/* The pixel multiplier can only be updated once the
@@ -5691,32 +5746,25 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		 *
 		 * So write it again.
 		 */
-		I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
-	}
+		I915_WRITE(PCH_DPLL(pll->id), dpll);
 
-	intel_crtc->lowfreq_avail = false;
-	if (intel_crtc->pch_pll) {
-		if (is_lvds && has_reduced_clock && i915_powersave) {
-			I915_WRITE(intel_crtc->pch_pll->fp1_reg, fp2);
-			intel_crtc->lowfreq_avail = true;
-		} else {
-			I915_WRITE(intel_crtc->pch_pll->fp1_reg, fp);
-		}
+		if (has_reduced_clock)
+			I915_WRITE(PCH_FP1(pll->id), fp2);
+		else
+			I915_WRITE(PCH_FP1(pll->id), fp);
 	}
 
-	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
+	intel_set_pipe_timings(intel_crtc);
 
-	/* Note, this also computes intel_crtc->fdi_lanes which is used below in
-	 * ironlake_check_fdi_lanes. */
-	intel_crtc->fdi_lanes = 0;
-	if (intel_crtc->config.has_pch_encoder)
-		ironlake_fdi_set_m_n(crtc);
-
-	fdi_config_ok = ironlake_check_fdi_lanes(intel_crtc);
+	if (intel_crtc->config.has_pch_encoder) {
+		intel_cpu_transcoder_set_m_n(intel_crtc,
+					     &intel_crtc->config.fdi_m_n);
+	}
 
-	ironlake_set_pipeconf(crtc, adjusted_mode, dither);
+	if (IS_IVYBRIDGE(dev))
+		ivybridge_update_fdi_bc_bifurcation(intel_crtc);
 
-	intel_wait_for_vblank(dev, pipe);
+	ironlake_set_pipeconf(crtc);
 
 	/* Set up the display plane register */
 	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE);
@@ -5726,9 +5774,46 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	intel_update_watermarks(dev);
 
-	intel_update_linetime_watermarks(dev, pipe, adjusted_mode);
+	return ret;
+}
 
-	return fdi_config_ok ? ret : -EINVAL;
+static void ironlake_get_fdi_m_n_config(struct intel_crtc *crtc,
+					struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder transcoder = pipe_config->cpu_transcoder;
+
+	pipe_config->fdi_m_n.link_m = I915_READ(PIPE_LINK_M1(transcoder));
+	pipe_config->fdi_m_n.link_n = I915_READ(PIPE_LINK_N1(transcoder));
+	pipe_config->fdi_m_n.gmch_m = I915_READ(PIPE_DATA_M1(transcoder))
+					& ~TU_SIZE_MASK;
+	pipe_config->fdi_m_n.gmch_n = I915_READ(PIPE_DATA_N1(transcoder));
+	pipe_config->fdi_m_n.tu = ((I915_READ(PIPE_DATA_M1(transcoder))
+				   & TU_SIZE_MASK) >> TU_SIZE_SHIFT) + 1;
+}
+
+static void ironlake_get_pfit_config(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PF_CTL(crtc->pipe));
+
+	if (tmp & PF_ENABLE) {
+		pipe_config->pch_pfit.pos = I915_READ(PF_WIN_POS(crtc->pipe));
+		pipe_config->pch_pfit.size = I915_READ(PF_WIN_SZ(crtc->pipe));
+
+		/* We currently do not free assignements of panel fitters on
+		 * ivb/hsw (since we don't use the higher upscaling modes which
+		 * differentiates them) so just WARN about this case for now. */
+		if (IS_GEN7(dev)) {
+			WARN_ON((tmp & PF_PIPE_SEL_MASK_IVB) !=
+				PF_PIPE_SEL_IVB(crtc->pipe));
+		}
+	}
 }
 
 static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
@@ -5738,42 +5823,67 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
+	pipe_config->cpu_transcoder = crtc->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
 	tmp = I915_READ(PIPECONF(crtc->pipe));
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
-	if (I915_READ(TRANSCONF(crtc->pipe)) & TRANS_ENABLE)
+	if (I915_READ(PCH_TRANSCONF(crtc->pipe)) & TRANS_ENABLE) {
+		struct intel_shared_dpll *pll;
+
 		pipe_config->has_pch_encoder = true;
 
+		tmp = I915_READ(FDI_RX_CTL(crtc->pipe));
+		pipe_config->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+					  FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+		ironlake_get_fdi_m_n_config(crtc, pipe_config);
+
+		/* XXX: Can't properly read out the pch dpll pixel multiplier
+		 * since we don't have state tracking for pch clocks yet. */
+		pipe_config->pixel_multiplier = 1;
+
+		if (HAS_PCH_IBX(dev_priv->dev)) {
+			pipe_config->shared_dpll = crtc->pipe;
+		} else {
+			tmp = I915_READ(PCH_DPLL_SEL);
+			if (tmp & TRANS_DPLLB_SEL(crtc->pipe))
+				pipe_config->shared_dpll = DPLL_ID_PCH_PLL_B;
+			else
+				pipe_config->shared_dpll = DPLL_ID_PCH_PLL_A;
+		}
+
+		pll = &dev_priv->shared_dplls[pipe_config->shared_dpll];
+
+		WARN_ON(!pll->get_hw_state(dev_priv, pll,
+					   &pipe_config->dpll_hw_state));
+	} else {
+		pipe_config->pixel_multiplier = 1;
+	}
+
+	intel_get_pipe_timings(crtc, pipe_config);
+
+	ironlake_get_pfit_config(crtc, pipe_config);
+
 	return true;
 }
 
 static void haswell_modeset_global_resources(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool enable = false;
 	struct intel_crtc *crtc;
-	struct intel_encoder *encoder;
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head) {
-		if (crtc->pipe != PIPE_A && crtc->base.enabled)
-			enable = true;
-		/* XXX: Should check for edp transcoder here, but thanks to init
-		 * sequence that's not yet available. Just in case desktop eDP
-		 * on PORT D is possible on haswell, too. */
-	}
+		if (!crtc->base.enabled)
+			continue;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
-		if (encoder->type != INTEL_OUTPUT_EDP &&
-		    encoder->connectors_active)
+		if (crtc->pipe != PIPE_A || crtc->config.pch_pfit.size ||
+		    crtc->config.cpu_transcoder != TRANSCODER_EDP)
 			enable = true;
 	}
 
-	/* Even the eDP panel fitter is outside the always-on well. */
-	if (dev_priv->pch_pf_size)
-		enable = true;
-
 	intel_set_power_well(dev, enable);
 }
 
@@ -5784,68 +5894,28 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
-	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
-	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
-	int num_connectors = 0;
-	bool is_cpu_edp = false;
-	struct intel_encoder *encoder;
 	int ret;
-	bool dither;
-
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
-		switch (encoder->type) {
-		case INTEL_OUTPUT_EDP:
-			if (!intel_encoder_is_pch_edp(&encoder->base))
-				is_cpu_edp = true;
-			break;
-		}
-
-		num_connectors++;
-	}
-
-	if (is_cpu_edp)
-		intel_crtc->config.cpu_transcoder = TRANSCODER_EDP;
-	else
-		intel_crtc->config.cpu_transcoder = pipe;
-
-	/* We are not sure yet this won't happen. */
-	WARN(!HAS_PCH_LPT(dev), "Unexpected PCH type %d\n",
-	     INTEL_PCH_TYPE(dev));
 
-	WARN(num_connectors != 1, "%d connectors attached to pipe %c\n",
-	     num_connectors, pipe_name(pipe));
-
-	WARN_ON(I915_READ(PIPECONF(intel_crtc->config.cpu_transcoder)) &
-		(PIPECONF_ENABLE | I965_PIPECONF_ACTIVE));
-
-	WARN_ON(I915_READ(DSPCNTR(plane)) & DISPLAY_PLANE_ENABLE);
-
-	if (!intel_ddi_pll_mode_set(crtc, adjusted_mode->clock))
+	if (!intel_ddi_pll_mode_set(crtc))
 		return -EINVAL;
 
 	/* Ensure that the cursor is valid for the new mode before changing... */
 	intel_crtc_update_cursor(crtc, true);
 
-	/* determine panel color depth */
-	dither = intel_crtc->config.dither;
-
-	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
-	drm_mode_debug_printmodeline(mode);
-
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_crtc->lowfreq_avail = false;
 
-	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
+	intel_set_pipe_timings(intel_crtc);
 
-	if (intel_crtc->config.has_pch_encoder)
-		ironlake_fdi_set_m_n(crtc);
+	if (intel_crtc->config.has_pch_encoder) {
+		intel_cpu_transcoder_set_m_n(intel_crtc,
+					     &intel_crtc->config.fdi_m_n);
+	}
 
-	haswell_set_pipeconf(crtc, adjusted_mode, dither);
+	haswell_set_pipeconf(crtc);
 
 	intel_set_pipe_csc(crtc);
 
@@ -5857,8 +5927,6 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 
 	intel_update_watermarks(dev);
 
-	intel_update_linetime_watermarks(dev, pipe, adjusted_mode);
-
 	return ret;
 }
 
@@ -5867,22 +5935,69 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain pfit_domain;
 	uint32_t tmp;
 
-	tmp = I915_READ(PIPECONF(crtc->config.cpu_transcoder));
+	pipe_config->cpu_transcoder = crtc->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
+	tmp = I915_READ(TRANS_DDI_FUNC_CTL(TRANSCODER_EDP));
+	if (tmp & TRANS_DDI_FUNC_ENABLE) {
+		enum pipe trans_edp_pipe;
+		switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
+		default:
+			WARN(1, "unknown pipe linked to edp transcoder\n");
+		case TRANS_DDI_EDP_INPUT_A_ONOFF:
+		case TRANS_DDI_EDP_INPUT_A_ON:
+			trans_edp_pipe = PIPE_A;
+			break;
+		case TRANS_DDI_EDP_INPUT_B_ONOFF:
+			trans_edp_pipe = PIPE_B;
+			break;
+		case TRANS_DDI_EDP_INPUT_C_ONOFF:
+			trans_edp_pipe = PIPE_C;
+			break;
+		}
+
+		if (trans_edp_pipe == crtc->pipe)
+			pipe_config->cpu_transcoder = TRANSCODER_EDP;
+	}
+
+	if (!intel_display_power_enabled(dev,
+			POWER_DOMAIN_TRANSCODER(pipe_config->cpu_transcoder)))
+		return false;
+
+	tmp = I915_READ(PIPECONF(pipe_config->cpu_transcoder));
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
 	/*
-	 * aswell has only FDI/PCH transcoder A. It is which is connected to
+	 * Haswell has only FDI/PCH transcoder A. It is which is connected to
 	 * DDI E. So just check whether this pipe is wired to DDI E and whether
 	 * the PCH transcoder is on.
 	 */
-	tmp = I915_READ(TRANS_DDI_FUNC_CTL(crtc->pipe));
+	tmp = I915_READ(TRANS_DDI_FUNC_CTL(pipe_config->cpu_transcoder));
 	if ((tmp & TRANS_DDI_PORT_MASK) == TRANS_DDI_SELECT_PORT(PORT_E) &&
-	    I915_READ(TRANSCONF(PIPE_A)) & TRANS_ENABLE)
+	    I915_READ(LPT_TRANSCONF) & TRANS_ENABLE) {
 		pipe_config->has_pch_encoder = true;
 
+		tmp = I915_READ(FDI_RX_CTL(PIPE_A));
+		pipe_config->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+					  FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+		ironlake_get_fdi_m_n_config(crtc, pipe_config);
+	}
+
+	intel_get_pipe_timings(crtc, pipe_config);
+
+	pfit_domain = POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe);
+	if (intel_display_power_enabled(dev, pfit_domain))
+		ironlake_get_pfit_config(crtc, pipe_config);
+
+	pipe_config->ips_enabled = hsw_crtc_supports_ips(crtc) &&
+				   (I915_READ(IPS_CTL) & IPS_ENABLE);
+
+	pipe_config->pixel_multiplier = 1;
 
 	return true;
 }
@@ -6120,7 +6235,7 @@ static void ironlake_write_eld(struct drm_connector *connector,
 		eldv |= IBX_ELD_VALIDB << 4;
 		eldv |= IBX_ELD_VALIDB << 8;
 	} else {
-		DRM_DEBUG_DRIVER("ELD on port %c\n", 'A' + i);
+		DRM_DEBUG_DRIVER("ELD on port %c\n", port_name(i));
 		eldv = IBX_ELD_VALIDB << ((i - 1) * 4);
 	}
 
@@ -6188,16 +6303,31 @@ void intel_crtc_load_lut(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int palreg = PALETTE(intel_crtc->pipe);
+	enum pipe pipe = intel_crtc->pipe;
+	int palreg = PALETTE(pipe);
 	int i;
+	bool reenable_ips = false;
 
 	/* The clocks have to be on to load the palette. */
 	if (!crtc->enabled || !intel_crtc->active)
 		return;
 
+	if (!HAS_PCH_SPLIT(dev_priv->dev))
+		assert_pll_enabled(dev_priv, pipe);
+
 	/* use legacy palette for Ironlake */
 	if (HAS_PCH_SPLIT(dev))
-		palreg = LGC_PALETTE(intel_crtc->pipe);
+		palreg = LGC_PALETTE(pipe);
+
+	/* Workaround : Do not read or write the pipe palette/gamma data while
+	 * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled.
+	 */
+	if (intel_crtc->config.ips_enabled &&
+	    ((I915_READ(GAMMA_MODE(pipe)) & GAMMA_MODE_MODE_MASK) ==
+	     GAMMA_MODE_MODE_SPLIT)) {
+		hsw_disable_ips(intel_crtc);
+		reenable_ips = true;
+	}
 
 	for (i = 0; i < 256; i++) {
 		I915_WRITE(palreg + 4 * i,
@@ -6205,6 +6335,9 @@ void intel_crtc_load_lut(struct drm_crtc *crtc)
 			   (intel_crtc->lut_g[i] << 8) |
 			   intel_crtc->lut_b[i]);
 	}
+
+	if (reenable_ips)
+		hsw_enable_ips(intel_crtc);
 }
 
 static void i845_update_cursor(struct drm_crtc *crtc, u32 base)
@@ -6451,7 +6584,7 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 	intel_crtc->cursor_width = width;
 	intel_crtc->cursor_height = height;
 
-	intel_crtc_update_cursor(crtc, true);
+	intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
 
 	return 0;
 fail_unpin:
@@ -6470,7 +6603,7 @@ static int intel_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	intel_crtc->cursor_x = x;
 	intel_crtc->cursor_y = y;
 
-	intel_crtc_update_cursor(crtc, true);
+	intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
 
 	return 0;
 }
@@ -6791,8 +6924,10 @@ static int intel_crtc_clock_get(struct drm_device *dev, struct drm_crtc *crtc)
 			return 0;
 		}
 
-		/* XXX: Handle the 100Mhz refclk */
-		intel_clock(dev, 96000, &clock);
+		if (IS_PINEVIEW(dev))
+			pineview_clock(96000, &clock);
+		else
+			i9xx_clock(96000, &clock);
 	} else {
 		bool is_lvds = (pipe == 1) && (I915_READ(LVDS) & LVDS_PORT_EN);
 
@@ -6804,9 +6939,9 @@ static int intel_crtc_clock_get(struct drm_device *dev, struct drm_crtc *crtc)
 			if ((dpll & PLL_REF_INPUT_MASK) ==
 			    PLLB_REF_INPUT_SPREADSPECTRUMIN) {
 				/* XXX: might not be 66MHz */
-				intel_clock(dev, 66000, &clock);
+				i9xx_clock(66000, &clock);
 			} else
-				intel_clock(dev, 48000, &clock);
+				i9xx_clock(48000, &clock);
 		} else {
 			if (dpll & PLL_P1_DIVIDE_BY_TWO)
 				clock.p1 = 2;
@@ -6819,7 +6954,7 @@ static int intel_crtc_clock_get(struct drm_device *dev, struct drm_crtc *crtc)
 			else
 				clock.p2 = 2;
 
-			intel_clock(dev, 48000, &clock);
+			i9xx_clock(48000, &clock);
 		}
 	}
 
@@ -6950,7 +7085,8 @@ void intel_mark_idle(struct drm_device *dev)
 	}
 }
 
-void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
+void intel_mark_fb_busy(struct drm_i915_gem_object *obj,
+			struct intel_ring_buffer *ring)
 {
 	struct drm_device *dev = obj->base.dev;
 	struct drm_crtc *crtc;
@@ -6962,8 +7098,12 @@ void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
 		if (!crtc->fb)
 			continue;
 
-		if (to_intel_framebuffer(crtc->fb)->obj == obj)
-			intel_increase_pllclock(crtc);
+		if (to_intel_framebuffer(crtc->fb)->obj != obj)
+			continue;
+
+		intel_increase_pllclock(crtc);
+		if (ring && intel_fbc_enabled(dev))
+			ring->fbc_dirty = true;
 	}
 }
 
@@ -6984,6 +7124,8 @@ static void intel_crtc_destroy(struct drm_crtc *crtc)
 		kfree(work);
 	}
 
+	intel_crtc_cursor_set(crtc, NULL, 0, 0, 0);
+
 	drm_crtc_cleanup(crtc);
 
 	kfree(intel_crtc);
@@ -7411,7 +7553,7 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 		goto cleanup_pending;
 
 	intel_disable_fbc(dev);
-	intel_mark_fb_busy(obj);
+	intel_mark_fb_busy(obj, NULL);
 	mutex_unlock(&dev->struct_mutex);
 
 	trace_i915_flip_request(intel_crtc->plane, obj);
@@ -7442,28 +7584,6 @@ static struct drm_crtc_helper_funcs intel_helper_funcs = {
 	.load_lut = intel_crtc_load_lut,
 };
 
-bool intel_encoder_check_is_cloned(struct intel_encoder *encoder)
-{
-	struct intel_encoder *other_encoder;
-	struct drm_crtc *crtc = &encoder->new_crtc->base;
-
-	if (WARN_ON(!crtc))
-		return false;
-
-	list_for_each_entry(other_encoder,
-			    &crtc->dev->mode_config.encoder_list,
-			    base.head) {
-
-		if (&other_encoder->new_crtc->base != crtc ||
-		    encoder == other_encoder)
-			continue;
-		else
-			return true;
-	}
-
-	return false;
-}
-
 static bool intel_encoder_crtc_ok(struct drm_encoder *encoder,
 				  struct drm_crtc *crtc)
 {
@@ -7531,13 +7651,39 @@ static void intel_modeset_commit_output_state(struct drm_device *dev)
 	}
 }
 
+static void
+connected_sink_compute_bpp(struct intel_connector * connector,
+			   struct intel_crtc_config *pipe_config)
+{
+	int bpp = pipe_config->pipe_bpp;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] checking for sink bpp constrains\n",
+		connector->base.base.id,
+		drm_get_connector_name(&connector->base));
+
+	/* Don't use an invalid EDID bpc value */
+	if (connector->base.display_info.bpc &&
+	    connector->base.display_info.bpc * 3 < bpp) {
+		DRM_DEBUG_KMS("clamping display bpp (was %d) to EDID reported max of %d\n",
+			      bpp, connector->base.display_info.bpc*3);
+		pipe_config->pipe_bpp = connector->base.display_info.bpc*3;
+	}
+
+	/* Clamp bpp to 8 on screens without EDID 1.4 */
+	if (connector->base.display_info.bpc == 0 && bpp > 24) {
+		DRM_DEBUG_KMS("clamping display bpp (was %d) to default limit of 24\n",
+			      bpp);
+		pipe_config->pipe_bpp = 24;
+	}
+}
+
 static int
-pipe_config_set_bpp(struct drm_crtc *crtc,
-		    struct drm_framebuffer *fb,
-		    struct intel_crtc_config *pipe_config)
+compute_baseline_pipe_bpp(struct intel_crtc *crtc,
+			  struct drm_framebuffer *fb,
+			  struct intel_crtc_config *pipe_config)
 {
-	struct drm_device *dev = crtc->dev;
-	struct drm_connector *connector;
+	struct drm_device *dev = crtc->base.dev;
+	struct intel_connector *connector;
 	int bpp;
 
 	switch (fb->pixel_format) {
@@ -7580,22 +7726,66 @@ pipe_config_set_bpp(struct drm_crtc *crtc,
 
 	/* Clamp display bpp to EDID value */
 	list_for_each_entry(connector, &dev->mode_config.connector_list,
-			    head) {
-		if (connector->encoder && connector->encoder->crtc != crtc)
+			    base.head) {
+		if (!connector->new_encoder ||
+		    connector->new_encoder->new_crtc != crtc)
 			continue;
 
-		/* Don't use an invalid EDID bpc value */
-		if (connector->display_info.bpc &&
-		    connector->display_info.bpc * 3 < bpp) {
-			DRM_DEBUG_KMS("clamping display bpp (was %d) to EDID reported max of %d\n",
-				      bpp, connector->display_info.bpc*3);
-			pipe_config->pipe_bpp = connector->display_info.bpc*3;
-		}
+		connected_sink_compute_bpp(connector, pipe_config);
 	}
 
 	return bpp;
 }
 
+static void intel_dump_pipe_config(struct intel_crtc *crtc,
+				   struct intel_crtc_config *pipe_config,
+				   const char *context)
+{
+	DRM_DEBUG_KMS("[CRTC:%d]%s config for pipe %c\n", crtc->base.base.id,
+		      context, pipe_name(crtc->pipe));
+
+	DRM_DEBUG_KMS("cpu_transcoder: %c\n", transcoder_name(pipe_config->cpu_transcoder));
+	DRM_DEBUG_KMS("pipe bpp: %i, dithering: %i\n",
+		      pipe_config->pipe_bpp, pipe_config->dither);
+	DRM_DEBUG_KMS("fdi/pch: %i, lanes: %i, gmch_m: %u, gmch_n: %u, link_m: %u, link_n: %u, tu: %u\n",
+		      pipe_config->has_pch_encoder,
+		      pipe_config->fdi_lanes,
+		      pipe_config->fdi_m_n.gmch_m, pipe_config->fdi_m_n.gmch_n,
+		      pipe_config->fdi_m_n.link_m, pipe_config->fdi_m_n.link_n,
+		      pipe_config->fdi_m_n.tu);
+	DRM_DEBUG_KMS("requested mode:\n");
+	drm_mode_debug_printmodeline(&pipe_config->requested_mode);
+	DRM_DEBUG_KMS("adjusted mode:\n");
+	drm_mode_debug_printmodeline(&pipe_config->adjusted_mode);
+	DRM_DEBUG_KMS("gmch pfit: control: 0x%08x, ratios: 0x%08x, lvds border: 0x%08x\n",
+		      pipe_config->gmch_pfit.control,
+		      pipe_config->gmch_pfit.pgm_ratios,
+		      pipe_config->gmch_pfit.lvds_border_bits);
+	DRM_DEBUG_KMS("pch pfit: pos: 0x%08x, size: 0x%08x\n",
+		      pipe_config->pch_pfit.pos,
+		      pipe_config->pch_pfit.size);
+	DRM_DEBUG_KMS("ips: %i\n", pipe_config->ips_enabled);
+}
+
+static bool check_encoder_cloning(struct drm_crtc *crtc)
+{
+	int num_encoders = 0;
+	bool uncloneable_encoders = false;
+	struct intel_encoder *encoder;
+
+	list_for_each_entry(encoder, &crtc->dev->mode_config.encoder_list,
+			    base.head) {
+		if (&encoder->new_crtc->base != crtc)
+			continue;
+
+		num_encoders++;
+		if (!encoder->cloneable)
+			uncloneable_encoders = true;
+	}
+
+	return !(num_encoders > 1 && uncloneable_encoders);
+}
+
 static struct intel_crtc_config *
 intel_modeset_pipe_config(struct drm_crtc *crtc,
 			  struct drm_framebuffer *fb,
@@ -7605,7 +7795,13 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 	struct drm_encoder_helper_funcs *encoder_funcs;
 	struct intel_encoder *encoder;
 	struct intel_crtc_config *pipe_config;
-	int plane_bpp;
+	int plane_bpp, ret = -EINVAL;
+	bool retry = true;
+
+	if (!check_encoder_cloning(crtc)) {
+		DRM_DEBUG_KMS("rejecting invalid cloning configuration\n");
+		return ERR_PTR(-EINVAL);
+	}
 
 	pipe_config = kzalloc(sizeof(*pipe_config), GFP_KERNEL);
 	if (!pipe_config)
@@ -7613,11 +7809,23 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 
 	drm_mode_copy(&pipe_config->adjusted_mode, mode);
 	drm_mode_copy(&pipe_config->requested_mode, mode);
-
-	plane_bpp = pipe_config_set_bpp(crtc, fb, pipe_config);
+	pipe_config->cpu_transcoder = to_intel_crtc(crtc)->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
+	/* Compute a starting value for pipe_config->pipe_bpp taking the source
+	 * plane pixel format and any sink constraints into account. Returns the
+	 * source plane bpp so that dithering can be selected on mismatches
+	 * after encoders and crtc also have had their say. */
+	plane_bpp = compute_baseline_pipe_bpp(to_intel_crtc(crtc),
+					      fb, pipe_config);
 	if (plane_bpp < 0)
 		goto fail;
 
+encoder_retry:
+	/* Ensure the port clock defaults are reset when retrying. */
+	pipe_config->port_clock = 0;
+	pipe_config->pixel_multiplier = 1;
+
 	/* Pass our mode to the connectors and the CRTC to give them a chance to
 	 * adjust it according to limitations or connector properties, and also
 	 * a chance to reject the mode entirely.
@@ -7646,11 +7854,27 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 		}
 	}
 
-	if (!(intel_crtc_compute_config(crtc, pipe_config))) {
+	/* Set default port clock if not overwritten by the encoder. Needs to be
+	 * done afterwards in case the encoder adjusts the mode. */
+	if (!pipe_config->port_clock)
+		pipe_config->port_clock = pipe_config->adjusted_mode.clock;
+
+	ret = intel_crtc_compute_config(to_intel_crtc(crtc), pipe_config);
+	if (ret < 0) {
 		DRM_DEBUG_KMS("CRTC fixup failed\n");
 		goto fail;
 	}
-	DRM_DEBUG_KMS("[CRTC:%d]\n", crtc->base.id);
+
+	if (ret == RETRY) {
+		if (WARN(!retry, "loop in pipe configuration computation\n")) {
+			ret = -EINVAL;
+			goto fail;
+		}
+
+		DRM_DEBUG_KMS("CRTC bw constrained, retrying\n");
+		retry = false;
+		goto encoder_retry;
+	}
 
 	pipe_config->dither = pipe_config->pipe_bpp != plane_bpp;
 	DRM_DEBUG_KMS("plane bpp: %i, pipe bpp: %i, dithering: %i\n",
@@ -7659,7 +7883,7 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 	return pipe_config;
 fail:
 	kfree(pipe_config);
-	return ERR_PTR(-EINVAL);
+	return ERR_PTR(ret);
 }
 
 /* Computes which crtcs are affected and sets the relevant bits in the mask. For
@@ -7755,6 +7979,9 @@ intel_modeset_affected_pipes(struct drm_crtc *crtc, unsigned *modeset_pipes,
 	 */
 	*modeset_pipes &= 1 << intel_crtc->pipe;
 	*prepare_pipes &= 1 << intel_crtc->pipe;
+
+	DRM_DEBUG_KMS("set mode pipe masks: modeset: %x, prepare: %x, disable: %x\n",
+		      *modeset_pipes, *prepare_pipes, *disable_pipes);
 }
 
 static bool intel_crtc_in_use(struct drm_crtc *crtc)
@@ -7821,31 +8048,114 @@ intel_modeset_update_state(struct drm_device *dev, unsigned prepare_pipes)
 	list_for_each_entry((intel_crtc), \
 			    &(dev)->mode_config.crtc_list, \
 			    base.head) \
-		if (mask & (1 <<(intel_crtc)->pipe)) \
+		if (mask & (1 <<(intel_crtc)->pipe))
 
 static bool
-intel_pipe_config_compare(struct intel_crtc_config *current_config,
+intel_pipe_config_compare(struct drm_device *dev,
+			  struct intel_crtc_config *current_config,
 			  struct intel_crtc_config *pipe_config)
 {
-	if (current_config->has_pch_encoder != pipe_config->has_pch_encoder) {
-		DRM_ERROR("mismatch in has_pch_encoder "
-			  "(expected %i, found %i)\n",
-			  current_config->has_pch_encoder,
-			  pipe_config->has_pch_encoder);
-		return false;
-	}
+#define PIPE_CONF_CHECK_X(name)	\
+	if (current_config->name != pipe_config->name) { \
+		DRM_ERROR("mismatch in " #name " " \
+			  "(expected 0x%08x, found 0x%08x)\n", \
+			  current_config->name, \
+			  pipe_config->name); \
+		return false; \
+	}
+
+#define PIPE_CONF_CHECK_I(name)	\
+	if (current_config->name != pipe_config->name) { \
+		DRM_ERROR("mismatch in " #name " " \
+			  "(expected %i, found %i)\n", \
+			  current_config->name, \
+			  pipe_config->name); \
+		return false; \
+	}
+
+#define PIPE_CONF_CHECK_FLAGS(name, mask)	\
+	if ((current_config->name ^ pipe_config->name) & (mask)) { \
+		DRM_ERROR("mismatch in " #name " " \
+			  "(expected %i, found %i)\n", \
+			  current_config->name & (mask), \
+			  pipe_config->name & (mask)); \
+		return false; \
+	}
+
+#define PIPE_CONF_QUIRK(quirk)	\
+	((current_config->quirks | pipe_config->quirks) & (quirk))
+
+	PIPE_CONF_CHECK_I(cpu_transcoder);
+
+	PIPE_CONF_CHECK_I(has_pch_encoder);
+	PIPE_CONF_CHECK_I(fdi_lanes);
+	PIPE_CONF_CHECK_I(fdi_m_n.gmch_m);
+	PIPE_CONF_CHECK_I(fdi_m_n.gmch_n);
+	PIPE_CONF_CHECK_I(fdi_m_n.link_m);
+	PIPE_CONF_CHECK_I(fdi_m_n.link_n);
+	PIPE_CONF_CHECK_I(fdi_m_n.tu);
+
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hdisplay);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_htotal);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hblank_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hblank_end);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hsync_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hsync_end);
+
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vdisplay);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vtotal);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vblank_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vblank_end);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vsync_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vsync_end);
+
+	if (!HAS_PCH_SPLIT(dev))
+		PIPE_CONF_CHECK_I(pixel_multiplier);
+
+	PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+			      DRM_MODE_FLAG_INTERLACE);
+
+	if (!PIPE_CONF_QUIRK(PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS)) {
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_PHSYNC);
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_NHSYNC);
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_PVSYNC);
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_NVSYNC);
+	}
+
+	PIPE_CONF_CHECK_I(requested_mode.hdisplay);
+	PIPE_CONF_CHECK_I(requested_mode.vdisplay);
+
+	PIPE_CONF_CHECK_I(gmch_pfit.control);
+	/* pfit ratios are autocomputed by the hw on gen4+ */
+	if (INTEL_INFO(dev)->gen < 4)
+		PIPE_CONF_CHECK_I(gmch_pfit.pgm_ratios);
+	PIPE_CONF_CHECK_I(gmch_pfit.lvds_border_bits);
+	PIPE_CONF_CHECK_I(pch_pfit.pos);
+	PIPE_CONF_CHECK_I(pch_pfit.size);
+
+	PIPE_CONF_CHECK_I(ips_enabled);
+
+	PIPE_CONF_CHECK_I(shared_dpll);
+	PIPE_CONF_CHECK_X(dpll_hw_state.dpll);
+	PIPE_CONF_CHECK_X(dpll_hw_state.fp0);
+	PIPE_CONF_CHECK_X(dpll_hw_state.fp1);
+
+#undef PIPE_CONF_CHECK_X
+#undef PIPE_CONF_CHECK_I
+#undef PIPE_CONF_CHECK_FLAGS
+#undef PIPE_CONF_QUIRK
 
 	return true;
 }
 
-void
-intel_modeset_check_state(struct drm_device *dev)
+static void
+check_connector_state(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct intel_crtc *crtc;
-	struct intel_encoder *encoder;
 	struct intel_connector *connector;
-	struct intel_crtc_config pipe_config;
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list,
 			    base.head) {
@@ -7856,6 +8166,13 @@ intel_modeset_check_state(struct drm_device *dev)
 		WARN(&connector->new_encoder->base != connector->base.encoder,
 		     "connector's staged encoder doesn't match current encoder\n");
 	}
+}
+
+static void
+check_encoder_state(struct drm_device *dev)
+{
+	struct intel_encoder *encoder;
+	struct intel_connector *connector;
 
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
 			    base.head) {
@@ -7907,12 +8224,23 @@ intel_modeset_check_state(struct drm_device *dev)
 		     tracked_pipe, pipe);
 
 	}
+}
+
+static void
+check_crtc_state(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	struct intel_encoder *encoder;
+	struct intel_crtc_config pipe_config;
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list,
 			    base.head) {
 		bool enabled = false;
 		bool active = false;
 
+		memset(&pipe_config, 0, sizeof(pipe_config));
+
 		DRM_DEBUG_KMS("[CRTC:%d]\n",
 			      crtc->base.base.id);
 
@@ -7927,6 +8255,7 @@ intel_modeset_check_state(struct drm_device *dev)
 			if (encoder->connectors_active)
 				active = true;
 		}
+
 		WARN(active != crtc->active,
 		     "crtc's computed active state doesn't match tracked active state "
 		     "(expected %i, found %i)\n", active, crtc->active);
@@ -7934,7 +8263,6 @@ intel_modeset_check_state(struct drm_device *dev)
 		     "crtc's computed enabled state doesn't match tracked enabled state "
 		     "(expected %i, found %i)\n", enabled, crtc->base.enabled);
 
-		memset(&pipe_config, 0, sizeof(pipe_config));
 		active = dev_priv->display.get_pipe_config(crtc,
 							   &pipe_config);
 
@@ -7942,16 +8270,86 @@ intel_modeset_check_state(struct drm_device *dev)
 		if (crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
 			active = crtc->active;
 
+		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+				    base.head) {
+			if (encoder->base.crtc != &crtc->base)
+				continue;
+			if (encoder->get_config)
+				encoder->get_config(encoder, &pipe_config);
+		}
+
 		WARN(crtc->active != active,
 		     "crtc active state doesn't match with hw state "
 		     "(expected %i, found %i)\n", crtc->active, active);
 
-		WARN(active &&
-		     !intel_pipe_config_compare(&crtc->config, &pipe_config),
-		     "pipe state doesn't match!\n");
+		if (active &&
+		    !intel_pipe_config_compare(dev, &crtc->config, &pipe_config)) {
+			WARN(1, "pipe state doesn't match!\n");
+			intel_dump_pipe_config(crtc, &pipe_config,
+					       "[hw state]");
+			intel_dump_pipe_config(crtc, &crtc->config,
+					       "[sw state]");
+		}
 	}
 }
 
+static void
+check_shared_dpll_state(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	struct intel_dpll_hw_state dpll_hw_state;
+	int i;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
+		int enabled_crtcs = 0, active_crtcs = 0;
+		bool active;
+
+		memset(&dpll_hw_state, 0, sizeof(dpll_hw_state));
+
+		DRM_DEBUG_KMS("%s\n", pll->name);
+
+		active = pll->get_hw_state(dev_priv, pll, &dpll_hw_state);
+
+		WARN(pll->active > pll->refcount,
+		     "more active pll users than references: %i vs %i\n",
+		     pll->active, pll->refcount);
+		WARN(pll->active && !pll->on,
+		     "pll in active use but not on in sw tracking\n");
+		WARN(pll->on != active,
+		     "pll on state mismatch (expected %i, found %i)\n",
+		     pll->on, active);
+
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list,
+				    base.head) {
+			if (crtc->base.enabled && intel_crtc_to_shared_dpll(crtc) == pll)
+				enabled_crtcs++;
+			if (crtc->active && intel_crtc_to_shared_dpll(crtc) == pll)
+				active_crtcs++;
+		}
+		WARN(pll->active != active_crtcs,
+		     "pll active crtcs mismatch (expected %i, found %i)\n",
+		     pll->active, active_crtcs);
+		WARN(pll->refcount != enabled_crtcs,
+		     "pll enabled crtcs mismatch (expected %i, found %i)\n",
+		     pll->refcount, enabled_crtcs);
+
+		WARN(pll->on && memcmp(&pll->hw_state, &dpll_hw_state,
+				       sizeof(dpll_hw_state)),
+		     "pll hw state mismatch\n");
+	}
+}
+
+void
+intel_modeset_check_state(struct drm_device *dev)
+{
+	check_connector_state(dev);
+	check_encoder_state(dev);
+	check_crtc_state(dev);
+	check_shared_dpll_state(dev);
+}
+
 static int __intel_set_mode(struct drm_crtc *crtc,
 			    struct drm_display_mode *mode,
 			    int x, int y, struct drm_framebuffer *fb)
@@ -7988,11 +8386,10 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 
 			goto out;
 		}
+		intel_dump_pipe_config(to_intel_crtc(crtc), pipe_config,
+				       "[modeset]");
 	}
 
-	DRM_DEBUG_KMS("set mode pipe masks: modeset: %x, prepare: %x, disable: %x\n",
-		      modeset_pipes, prepare_pipes, disable_pipes);
-
 	for_each_intel_crtc_masked(dev, disable_pipes, intel_crtc)
 		intel_crtc_disable(&intel_crtc->base);
 
@@ -8005,12 +8402,10 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 	 * to set it here already despite that we pass it down the callchain.
 	 */
 	if (modeset_pipes) {
-		enum transcoder tmp = to_intel_crtc(crtc)->config.cpu_transcoder;
 		crtc->mode = *mode;
 		/* mode_set/enable/disable functions rely on a correct pipe
 		 * config. */
 		to_intel_crtc(crtc)->config = *pipe_config;
-		to_intel_crtc(crtc)->config.cpu_transcoder = tmp;
 	}
 
 	/* Only after disabling all output pipelines that will be changed can we
@@ -8349,12 +8744,6 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 		goto fail;
 
 	if (config->mode_changed) {
-		if (set->mode) {
-			DRM_DEBUG_KMS("attempting to set mode from"
-					" userspace\n");
-			drm_mode_debug_printmodeline(set->mode);
-		}
-
 		ret = intel_set_mode(set->crtc, set->mode,
 				     set->x, set->y, set->fb);
 	} else if (config->fb_changed) {
@@ -8365,8 +8754,8 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 	}
 
 	if (ret) {
-		DRM_ERROR("failed to set mode on [CRTC:%d], err = %d\n",
-			  set->crtc->base.id, ret);
+		DRM_DEBUG_KMS("failed to set mode on [CRTC:%d], err = %d\n",
+			      set->crtc->base.id, ret);
 fail:
 		intel_set_config_restore_state(dev, config);
 
@@ -8397,23 +8786,93 @@ static void intel_cpu_pll_init(struct drm_device *dev)
 		intel_ddi_pll_init(dev);
 }
 
-static void intel_pch_pll_init(struct drm_device *dev)
+static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv,
+				      struct intel_shared_dpll *pll,
+				      struct intel_dpll_hw_state *hw_state)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	int i;
+	uint32_t val;
 
-	if (dev_priv->num_pch_pll == 0) {
-		DRM_DEBUG_KMS("No PCH PLLs on this hardware, skipping initialisation\n");
-		return;
+	val = I915_READ(PCH_DPLL(pll->id));
+	hw_state->dpll = val;
+	hw_state->fp0 = I915_READ(PCH_FP0(pll->id));
+	hw_state->fp1 = I915_READ(PCH_FP1(pll->id));
+
+	return val & DPLL_VCO_ENABLE;
+}
+
+static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	uint32_t reg, val;
+
+	/* PCH refclock must be enabled first */
+	assert_pch_refclk_enabled(dev_priv);
+
+	reg = PCH_DPLL(pll->id);
+	val = I915_READ(reg);
+	val |= DPLL_VCO_ENABLE;
+	I915_WRITE(reg, val);
+	POSTING_READ(reg);
+	udelay(200);
+}
+
+static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_crtc *crtc;
+	uint32_t reg, val;
+
+	/* Make sure no transcoder isn't still depending on us. */
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head) {
+		if (intel_crtc_to_shared_dpll(crtc) == pll)
+			assert_pch_transcoder_disabled(dev_priv, crtc->pipe);
 	}
 
-	for (i = 0; i < dev_priv->num_pch_pll; i++) {
-		dev_priv->pch_plls[i].pll_reg = _PCH_DPLL(i);
-		dev_priv->pch_plls[i].fp0_reg = _PCH_FP0(i);
-		dev_priv->pch_plls[i].fp1_reg = _PCH_FP1(i);
+	reg = PCH_DPLL(pll->id);
+	val = I915_READ(reg);
+	val &= ~DPLL_VCO_ENABLE;
+	I915_WRITE(reg, val);
+	POSTING_READ(reg);
+	udelay(200);
+}
+
+static char *ibx_pch_dpll_names[] = {
+	"PCH DPLL A",
+	"PCH DPLL B",
+};
+
+static void ibx_pch_dpll_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int i;
+
+	dev_priv->num_shared_dpll = 2;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		dev_priv->shared_dplls[i].id = i;
+		dev_priv->shared_dplls[i].name = ibx_pch_dpll_names[i];
+		dev_priv->shared_dplls[i].enable = ibx_pch_dpll_enable;
+		dev_priv->shared_dplls[i].disable = ibx_pch_dpll_disable;
+		dev_priv->shared_dplls[i].get_hw_state =
+			ibx_pch_dpll_get_hw_state;
 	}
 }
 
+static void intel_shared_dpll_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		ibx_pch_dpll_init(dev);
+	else
+		dev_priv->num_shared_dpll = 0;
+
+	BUG_ON(dev_priv->num_shared_dpll > I915_NUM_PLLS);
+	DRM_DEBUG_KMS("%i shared PLLs initialized\n",
+		      dev_priv->num_shared_dpll);
+}
+
 static void intel_crtc_init(struct drm_device *dev, int pipe)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
@@ -8436,7 +8895,6 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 	/* Swap pipes & planes for FBC on pre-965 */
 	intel_crtc->pipe = pipe;
 	intel_crtc->plane = pipe;
-	intel_crtc->config.cpu_transcoder = pipe;
 	if (IS_MOBILE(dev) && IS_GEN3(dev)) {
 		DRM_DEBUG_KMS("swapping pipes & planes for FBC\n");
 		intel_crtc->plane = !pipe;
@@ -8519,13 +8977,8 @@ static void intel_setup_outputs(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *encoder;
 	bool dpd_is_edp = false;
-	bool has_lvds;
 
-	has_lvds = intel_lvds_init(dev);
-	if (!has_lvds && !HAS_PCH_SPLIT(dev)) {
-		/* disable the panel fitter on everything but LVDS */
-		I915_WRITE(PFIT_CONTROL, 0);
-	}
+	intel_lvds_init(dev);
 
 	if (!IS_ULT(dev))
 		intel_crt_init(dev);
@@ -8598,10 +9051,8 @@ static void intel_setup_outputs(struct drm_device *dev)
 				intel_hdmi_init(dev, GEN4_HDMIB, PORT_B);
 			}
 
-			if (!found && SUPPORTS_INTEGRATED_DP(dev)) {
-				DRM_DEBUG_KMS("probing DP_B\n");
+			if (!found && SUPPORTS_INTEGRATED_DP(dev))
 				intel_dp_init(dev, DP_B, PORT_B);
-			}
 		}
 
 		/* Before G4X SDVOC doesn't have its own detect register */
@@ -8617,17 +9068,13 @@ static void intel_setup_outputs(struct drm_device *dev)
 				DRM_DEBUG_KMS("probing HDMI on SDVOC\n");
 				intel_hdmi_init(dev, GEN4_HDMIC, PORT_C);
 			}
-			if (SUPPORTS_INTEGRATED_DP(dev)) {
-				DRM_DEBUG_KMS("probing DP_C\n");
+			if (SUPPORTS_INTEGRATED_DP(dev))
 				intel_dp_init(dev, DP_C, PORT_C);
-			}
 		}
 
 		if (SUPPORTS_INTEGRATED_DP(dev) &&
-		    (I915_READ(DP_D) & DP_DETECTED)) {
-			DRM_DEBUG_KMS("probing DP_D\n");
+		    (I915_READ(DP_D) & DP_DETECTED))
 			intel_dp_init(dev, DP_D, PORT_D);
-		}
 	} else if (IS_GEN2(dev))
 		intel_dvo_init(dev);
 
@@ -8675,6 +9122,7 @@ int intel_framebuffer_init(struct drm_device *dev,
 			   struct drm_mode_fb_cmd2 *mode_cmd,
 			   struct drm_i915_gem_object *obj)
 {
+	int pitch_limit;
 	int ret;
 
 	if (obj->tiling_mode == I915_TILING_Y) {
@@ -8688,10 +9136,26 @@ int intel_framebuffer_init(struct drm_device *dev,
 		return -EINVAL;
 	}
 
-	/* FIXME <= Gen4 stride limits are bit unclear */
-	if (mode_cmd->pitches[0] > 32768) {
-		DRM_DEBUG("pitch (%d) must be at less than 32768\n",
-			  mode_cmd->pitches[0]);
+	if (INTEL_INFO(dev)->gen >= 5 && !IS_VALLEYVIEW(dev)) {
+		pitch_limit = 32*1024;
+	} else if (INTEL_INFO(dev)->gen >= 4) {
+		if (obj->tiling_mode)
+			pitch_limit = 16*1024;
+		else
+			pitch_limit = 32*1024;
+	} else if (INTEL_INFO(dev)->gen >= 3) {
+		if (obj->tiling_mode)
+			pitch_limit = 8*1024;
+		else
+			pitch_limit = 16*1024;
+	} else
+		/* XXX DSPC is limited to 4k tiled */
+		pitch_limit = 8*1024;
+
+	if (mode_cmd->pitches[0] > pitch_limit) {
+		DRM_DEBUG("%s pitch (%d) must be at less than %d\n",
+			  obj->tiling_mode ? "tiled" : "linear",
+			  mode_cmd->pitches[0], pitch_limit);
 		return -EINVAL;
 	}
 
@@ -8712,7 +9176,8 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_XRGB1555:
 	case DRM_FORMAT_ARGB1555:
 		if (INTEL_INFO(dev)->gen > 3) {
-			DRM_DEBUG("invalid format: 0x%08x\n", mode_cmd->pixel_format);
+			DRM_DEBUG("unsupported pixel format: %s\n",
+				  drm_get_format_name(mode_cmd->pixel_format));
 			return -EINVAL;
 		}
 		break;
@@ -8723,7 +9188,8 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_XBGR2101010:
 	case DRM_FORMAT_ABGR2101010:
 		if (INTEL_INFO(dev)->gen < 4) {
-			DRM_DEBUG("invalid format: 0x%08x\n", mode_cmd->pixel_format);
+			DRM_DEBUG("unsupported pixel format: %s\n",
+				  drm_get_format_name(mode_cmd->pixel_format));
 			return -EINVAL;
 		}
 		break;
@@ -8732,12 +9198,14 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_YVYU:
 	case DRM_FORMAT_VYUY:
 		if (INTEL_INFO(dev)->gen < 5) {
-			DRM_DEBUG("invalid format: 0x%08x\n", mode_cmd->pixel_format);
+			DRM_DEBUG("unsupported pixel format: %s\n",
+				  drm_get_format_name(mode_cmd->pixel_format));
 			return -EINVAL;
 		}
 		break;
 	default:
-		DRM_DEBUG("unsupported pixel format 0x%08x\n", mode_cmd->pixel_format);
+		DRM_DEBUG("unsupported pixel format: %s\n",
+			  drm_get_format_name(mode_cmd->pixel_format));
 		return -EINVAL;
 	}
 
@@ -8782,6 +9250,15 @@ static void intel_init_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (HAS_PCH_SPLIT(dev) || IS_G4X(dev))
+		dev_priv->display.find_dpll = g4x_find_best_dpll;
+	else if (IS_VALLEYVIEW(dev))
+		dev_priv->display.find_dpll = vlv_find_best_dpll;
+	else if (IS_PINEVIEW(dev))
+		dev_priv->display.find_dpll = pnv_find_best_dpll;
+	else
+		dev_priv->display.find_dpll = i9xx_find_best_dpll;
+
 	if (HAS_DDI(dev)) {
 		dev_priv->display.get_pipe_config = haswell_get_pipe_config;
 		dev_priv->display.crtc_mode_set = haswell_crtc_mode_set;
@@ -8796,6 +9273,13 @@ static void intel_init_display(struct drm_device *dev)
 		dev_priv->display.crtc_disable = ironlake_crtc_disable;
 		dev_priv->display.off = ironlake_crtc_off;
 		dev_priv->display.update_plane = ironlake_update_plane;
+	} else if (IS_VALLEYVIEW(dev)) {
+		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
+		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
+		dev_priv->display.crtc_enable = valleyview_crtc_enable;
+		dev_priv->display.crtc_disable = i9xx_crtc_disable;
+		dev_priv->display.off = i9xx_crtc_off;
+		dev_priv->display.update_plane = i9xx_update_plane;
 	} else {
 		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
 		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
@@ -9037,6 +9521,11 @@ void intel_modeset_init_hw(struct drm_device *dev)
 	mutex_unlock(&dev->struct_mutex);
 }
 
+void intel_modeset_suspend_hw(struct drm_device *dev)
+{
+	intel_suspend_hw(dev);
+}
+
 void intel_modeset_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -9082,13 +9571,13 @@ void intel_modeset_init(struct drm_device *dev)
 		for (j = 0; j < dev_priv->num_plane; j++) {
 			ret = intel_plane_init(dev, i, j);
 			if (ret)
-				DRM_DEBUG_KMS("pipe %d plane %d init failed: %d\n",
-					      i, j, ret);
+				DRM_DEBUG_KMS("pipe %c sprite %c init failed: %d\n",
+					      pipe_name(i), sprite_name(i, j), ret);
 		}
 	}
 
 	intel_cpu_pll_init(dev);
-	intel_pch_pll_init(dev);
+	intel_shared_dpll_init(dev);
 
 	/* Just disable it once at startup */
 	i915_disable_vga(dev);
@@ -9289,57 +9778,18 @@ void i915_redisable_vga(struct drm_device *dev)
 	}
 }
 
-/* Scan out the current hw modeset state, sanitizes it and maps it into the drm
- * and i915 state tracking structures. */
-void intel_modeset_setup_hw_state(struct drm_device *dev,
-				  bool force_restore)
+static void intel_modeset_readout_hw_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
-	u32 tmp;
-	struct drm_plane *plane;
 	struct intel_crtc *crtc;
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
+	int i;
 
-	if (HAS_DDI(dev)) {
-		tmp = I915_READ(TRANS_DDI_FUNC_CTL(TRANSCODER_EDP));
-
-		if (tmp & TRANS_DDI_FUNC_ENABLE) {
-			switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
-			case TRANS_DDI_EDP_INPUT_A_ON:
-			case TRANS_DDI_EDP_INPUT_A_ONOFF:
-				pipe = PIPE_A;
-				break;
-			case TRANS_DDI_EDP_INPUT_B_ONOFF:
-				pipe = PIPE_B;
-				break;
-			case TRANS_DDI_EDP_INPUT_C_ONOFF:
-				pipe = PIPE_C;
-				break;
-			default:
-				/* A bogus value has been programmed, disable
-				 * the transcoder */
-				WARN(1, "Bogus eDP source %08x\n", tmp);
-				intel_ddi_disable_transcoder_func(dev_priv,
-						TRANSCODER_EDP);
-				goto setup_pipes;
-			}
-
-			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-			crtc->config.cpu_transcoder = TRANSCODER_EDP;
-
-			DRM_DEBUG_KMS("Pipe %c using transcoder EDP\n",
-				      pipe_name(pipe));
-		}
-	}
-
-setup_pipes:
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list,
 			    base.head) {
-		enum transcoder tmp = crtc->config.cpu_transcoder;
 		memset(&crtc->config, 0, sizeof(crtc->config));
-		crtc->config.cpu_transcoder = tmp;
 
 		crtc->active = dev_priv->display.get_pipe_config(crtc,
 								 &crtc->config);
@@ -9351,16 +9801,35 @@ setup_pipes:
 			      crtc->active ? "enabled" : "disabled");
 	}
 
+	/* FIXME: Smash this into the new shared dpll infrastructure. */
 	if (HAS_DDI(dev))
 		intel_ddi_setup_hw_pll_state(dev);
 
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
+
+		pll->on = pll->get_hw_state(dev_priv, pll, &pll->hw_state);
+		pll->active = 0;
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list,
+				    base.head) {
+			if (crtc->active && intel_crtc_to_shared_dpll(crtc) == pll)
+				pll->active++;
+		}
+		pll->refcount = pll->active;
+
+		DRM_DEBUG_KMS("%s hw state readout: refcount %i\n",
+			      pll->name, pll->refcount);
+	}
+
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
 			    base.head) {
 		pipe = 0;
 
 		if (encoder->get_hw_state(encoder, &pipe)) {
-			encoder->base.crtc =
-				dev_priv->pipe_to_crtc_mapping[pipe];
+			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+			encoder->base.crtc = &crtc->base;
+			if (encoder->get_config)
+				encoder->get_config(encoder, &crtc->config);
 		} else {
 			encoder->base.crtc = NULL;
 		}
@@ -9388,6 +9857,20 @@ setup_pipes:
 			      drm_get_connector_name(&connector->base),
 			      connector->base.encoder ? "enabled" : "disabled");
 	}
+}
+
+/* Scan out the current hw modeset state, sanitizes it and maps it into the drm
+ * and i915 state tracking structures. */
+void intel_modeset_setup_hw_state(struct drm_device *dev,
+				  bool force_restore)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+	struct drm_plane *plane;
+	struct intel_crtc *crtc;
+	struct intel_encoder *encoder;
+
+	intel_modeset_readout_hw_state(dev);
 
 	/* HW state is read out, now we need to sanitize this mess. */
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
@@ -9398,6 +9881,7 @@ setup_pipes:
 	for_each_pipe(pipe) {
 		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 		intel_sanitize_crtc(crtc);
+		intel_dump_pipe_config(crtc, &crtc->config, "[setup_hw_state]");
 	}
 
 	if (force_restore) {
@@ -9440,12 +9924,23 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	struct drm_crtc *crtc;
 	struct intel_crtc *intel_crtc;
 
+	/*
+	 * Interrupts and polling as the first thing to avoid creating havoc.
+	 * Too much stuff here (turning of rps, connectors, ...) would
+	 * experience fancy races otherwise.
+	 */
+	drm_irq_uninstall(dev);
+	cancel_work_sync(&dev_priv->hotplug_work);
+	/*
+	 * Due to the hpd irq storm handling the hotplug work can re-arm the
+	 * poll handlers. Hence disable polling after hpd handling is shut down.
+	 */
 	drm_kms_helper_poll_fini(dev);
+
 	mutex_lock(&dev->struct_mutex);
 
 	intel_unregister_dsm_handler();
 
-
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		/* Skip inactive CRTCs */
 		if (!crtc->fb)
@@ -9461,17 +9956,8 @@ void intel_modeset_cleanup(struct drm_device *dev)
 
 	ironlake_teardown_rc6(dev);
 
-	if (IS_VALLEYVIEW(dev))
-		vlv_init_dpio(dev);
-
 	mutex_unlock(&dev->struct_mutex);
 
-	/* Disable the irq before mode object teardown, for the irq might
-	 * enqueue unpin/hotplug work. */
-	drm_irq_uninstall(dev);
-	cancel_work_sync(&dev_priv->hotplug_work);
-	cancel_work_sync(&dev_priv->rps.work);
-
 	/* flush any delayed tasks or pending work */
 	flush_scheduled_work();
 
@@ -9520,6 +10006,9 @@ int intel_modeset_vga_set_state(struct drm_device *dev, bool state)
 #include <linux/seq_file.h>
 
 struct intel_display_error_state {
+
+	u32 power_well_driver;
+
 	struct intel_cursor_error_state {
 		u32 control;
 		u32 position;
@@ -9528,6 +10017,7 @@ struct intel_display_error_state {
 	} cursor[I915_MAX_PIPES];
 
 	struct intel_pipe_error_state {
+		enum transcoder cpu_transcoder;
 		u32 conf;
 		u32 source;
 
@@ -9562,8 +10052,12 @@ intel_display_capture_error_state(struct drm_device *dev)
 	if (error == NULL)
 		return NULL;
 
+	if (HAS_POWER_WELL(dev))
+		error->power_well_driver = I915_READ(HSW_PWR_WELL_DRIVER);
+
 	for_each_pipe(i) {
 		cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv, i);
+		error->pipe[i].cpu_transcoder = cpu_transcoder;
 
 		if (INTEL_INFO(dev)->gen <= 6 || IS_VALLEYVIEW(dev)) {
 			error->cursor[i].control = I915_READ(CURCNTR(i));
@@ -9598,46 +10092,60 @@ intel_display_capture_error_state(struct drm_device *dev)
 		error->pipe[i].vsync = I915_READ(VSYNC(cpu_transcoder));
 	}
 
+	/* In the code above we read the registers without checking if the power
+	 * well was on, so here we have to clear the FPGA_DBG_RM_NOCLAIM bit to
+	 * prevent the next I915_WRITE from detecting it and printing an error
+	 * message. */
+	if (HAS_POWER_WELL(dev))
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+
 	return error;
 }
 
+#define err_printf(e, ...) i915_error_printf(e, __VA_ARGS__)
+
 void
-intel_display_print_error_state(struct seq_file *m,
+intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 				struct drm_device *dev,
 				struct intel_display_error_state *error)
 {
 	int i;
 
-	seq_printf(m, "Num Pipes: %d\n", INTEL_INFO(dev)->num_pipes);
+	err_printf(m, "Num Pipes: %d\n", INTEL_INFO(dev)->num_pipes);
+	if (HAS_POWER_WELL(dev))
+		err_printf(m, "PWR_WELL_CTL2: %08x\n",
+			   error->power_well_driver);
 	for_each_pipe(i) {
-		seq_printf(m, "Pipe [%d]:\n", i);
-		seq_printf(m, "  CONF: %08x\n", error->pipe[i].conf);
-		seq_printf(m, "  SRC: %08x\n", error->pipe[i].source);
-		seq_printf(m, "  HTOTAL: %08x\n", error->pipe[i].htotal);
-		seq_printf(m, "  HBLANK: %08x\n", error->pipe[i].hblank);
-		seq_printf(m, "  HSYNC: %08x\n", error->pipe[i].hsync);
-		seq_printf(m, "  VTOTAL: %08x\n", error->pipe[i].vtotal);
-		seq_printf(m, "  VBLANK: %08x\n", error->pipe[i].vblank);
-		seq_printf(m, "  VSYNC: %08x\n", error->pipe[i].vsync);
-
-		seq_printf(m, "Plane [%d]:\n", i);
-		seq_printf(m, "  CNTR: %08x\n", error->plane[i].control);
-		seq_printf(m, "  STRIDE: %08x\n", error->plane[i].stride);
+		err_printf(m, "Pipe [%d]:\n", i);
+		err_printf(m, "  CPU transcoder: %c\n",
+			   transcoder_name(error->pipe[i].cpu_transcoder));
+		err_printf(m, "  CONF: %08x\n", error->pipe[i].conf);
+		err_printf(m, "  SRC: %08x\n", error->pipe[i].source);
+		err_printf(m, "  HTOTAL: %08x\n", error->pipe[i].htotal);
+		err_printf(m, "  HBLANK: %08x\n", error->pipe[i].hblank);
+		err_printf(m, "  HSYNC: %08x\n", error->pipe[i].hsync);
+		err_printf(m, "  VTOTAL: %08x\n", error->pipe[i].vtotal);
+		err_printf(m, "  VBLANK: %08x\n", error->pipe[i].vblank);
+		err_printf(m, "  VSYNC: %08x\n", error->pipe[i].vsync);
+
+		err_printf(m, "Plane [%d]:\n", i);
+		err_printf(m, "  CNTR: %08x\n", error->plane[i].control);
+		err_printf(m, "  STRIDE: %08x\n", error->plane[i].stride);
 		if (INTEL_INFO(dev)->gen <= 3) {
-			seq_printf(m, "  SIZE: %08x\n", error->plane[i].size);
-			seq_printf(m, "  POS: %08x\n", error->plane[i].pos);
+			err_printf(m, "  SIZE: %08x\n", error->plane[i].size);
+			err_printf(m, "  POS: %08x\n", error->plane[i].pos);
 		}
 		if (INTEL_INFO(dev)->gen <= 7 && !IS_HASWELL(dev))
-			seq_printf(m, "  ADDR: %08x\n", error->plane[i].addr);
+			err_printf(m, "  ADDR: %08x\n", error->plane[i].addr);
 		if (INTEL_INFO(dev)->gen >= 4) {
-			seq_printf(m, "  SURF: %08x\n", error->plane[i].surface);
-			seq_printf(m, "  TILEOFF: %08x\n", error->plane[i].tile_offset);
+			err_printf(m, "  SURF: %08x\n", error->plane[i].surface);
+			err_printf(m, "  TILEOFF: %08x\n", error->plane[i].tile_offset);
 		}
 
-		seq_printf(m, "Cursor [%d]:\n", i);
-		seq_printf(m, "  CNTR: %08x\n", error->cursor[i].control);
-		seq_printf(m, "  POS: %08x\n", error->cursor[i].position);
-		seq_printf(m, "  BASE: %08x\n", error->cursor[i].base);
+		err_printf(m, "Cursor [%d]:\n", i);
+		err_printf(m, "  CNTR: %08x\n", error->cursor[i].control);
+		err_printf(m, "  POS: %08x\n", error->cursor[i].position);
+		err_printf(m, "  BASE: %08x\n", error->cursor[i].base);
 	}
 }
 #endif
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 70789b1..b739712 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -52,30 +52,6 @@ static bool is_edp(struct intel_dp *intel_dp)
 	return intel_dig_port->base.type == INTEL_OUTPUT_EDP;
 }
 
-/**
- * is_pch_edp - is the port on the PCH and attached to an eDP panel?
- * @intel_dp: DP struct
- *
- * Returns true if the given DP struct corresponds to a PCH DP port attached
- * to an eDP panel, false otherwise.  Helpful for determining whether we
- * may need FDI resources for a given DP output or not.
- */
-static bool is_pch_edp(struct intel_dp *intel_dp)
-{
-	return intel_dp->is_pch_edp;
-}
-
-/**
- * is_cpu_edp - is the port on the CPU and attached to an eDP panel?
- * @intel_dp: DP struct
- *
- * Returns true if the given DP struct corresponds to a CPU eDP port.
- */
-static bool is_cpu_edp(struct intel_dp *intel_dp)
-{
-	return is_edp(intel_dp) && !is_pch_edp(intel_dp);
-}
-
 static struct drm_device *intel_dp_to_dev(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -88,25 +64,6 @@ static struct intel_dp *intel_attached_dp(struct drm_connector *connector)
 	return enc_to_intel_dp(&intel_attached_encoder(connector)->base);
 }
 
-/**
- * intel_encoder_is_pch_edp - is the given encoder a PCH attached eDP?
- * @encoder: DRM encoder
- *
- * Return true if @encoder corresponds to a PCH attached eDP panel.  Needed
- * by intel_display.c.
- */
-bool intel_encoder_is_pch_edp(struct drm_encoder *encoder)
-{
-	struct intel_dp *intel_dp;
-
-	if (!encoder)
-		return false;
-
-	intel_dp = enc_to_intel_dp(encoder);
-
-	return is_pch_edp(intel_dp);
-}
-
 static void intel_dp_link_down(struct intel_dp *intel_dp);
 
 static int
@@ -344,11 +301,12 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	 * Note that PCH attached eDP panels should use a 125MHz input
 	 * clock divider.
 	 */
-	if (is_cpu_edp(intel_dp)) {
+	if (IS_VALLEYVIEW(dev)) {
+		aux_clock_divider = 100;
+	} else if (intel_dig_port->port == PORT_A) {
 		if (HAS_DDI(dev))
-			aux_clock_divider = intel_ddi_get_cdclk_freq(dev_priv) >> 1;
-		else if (IS_VALLEYVIEW(dev))
-			aux_clock_divider = 100;
+			aux_clock_divider = DIV_ROUND_CLOSEST(
+				intel_ddi_get_cdclk_freq(dev_priv), 2000);
 		else if (IS_GEN6(dev) || IS_GEN7(dev))
 			aux_clock_divider = 200; /* SNB & IVB eDP input clock at 400Mhz */
 		else
@@ -660,6 +618,49 @@ intel_dp_i2c_init(struct intel_dp *intel_dp,
 	return ret;
 }
 
+static void
+intel_dp_set_clock(struct intel_encoder *encoder,
+		   struct intel_crtc_config *pipe_config, int link_bw)
+{
+	struct drm_device *dev = encoder->base.dev;
+
+	if (IS_G4X(dev)) {
+		if (link_bw == DP_LINK_BW_1_62) {
+			pipe_config->dpll.p1 = 2;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.n = 2;
+			pipe_config->dpll.m1 = 23;
+			pipe_config->dpll.m2 = 8;
+		} else {
+			pipe_config->dpll.p1 = 1;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.n = 1;
+			pipe_config->dpll.m1 = 14;
+			pipe_config->dpll.m2 = 2;
+		}
+		pipe_config->clock_set = true;
+	} else if (IS_HASWELL(dev)) {
+		/* Haswell has special-purpose DP DDI clocks. */
+	} else if (HAS_PCH_SPLIT(dev)) {
+		if (link_bw == DP_LINK_BW_1_62) {
+			pipe_config->dpll.n = 1;
+			pipe_config->dpll.p1 = 2;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.m1 = 12;
+			pipe_config->dpll.m2 = 9;
+		} else {
+			pipe_config->dpll.n = 2;
+			pipe_config->dpll.p1 = 1;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.m1 = 14;
+			pipe_config->dpll.m2 = 8;
+		}
+		pipe_config->clock_set = true;
+	} else if (IS_VALLEYVIEW(dev)) {
+		/* FIXME: Need to figure out optimized DP clocks for vlv. */
+	}
+}
+
 bool
 intel_dp_compute_config(struct intel_encoder *encoder,
 			struct intel_crtc_config *pipe_config)
@@ -667,17 +668,18 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
-	struct drm_display_mode *mode = &pipe_config->requested_mode;
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct intel_crtc *intel_crtc = encoder->new_crtc;
 	struct intel_connector *intel_connector = intel_dp->attached_connector;
 	int lane_count, clock;
 	int max_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
 	int max_clock = intel_dp_max_link_bw(intel_dp) == DP_LINK_BW_2_7 ? 1 : 0;
 	int bpp, mode_rate;
 	static int bws[2] = { DP_LINK_BW_1_62, DP_LINK_BW_2_7 };
-	int target_clock, link_avail, link_clock;
+	int link_avail, link_clock;
 
-	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev) && !is_cpu_edp(intel_dp))
+	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev) && port != PORT_A)
 		pipe_config->has_pch_encoder = true;
 
 	pipe_config->has_dp_encoder = true;
@@ -685,12 +687,13 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
 		intel_fixed_panel_mode(intel_connector->panel.fixed_mode,
 				       adjusted_mode);
-		intel_pch_panel_fitting(dev,
-					intel_connector->panel.fitting_mode,
-					mode, adjusted_mode);
+		if (!HAS_PCH_SPLIT(dev))
+			intel_gmch_panel_fitting(intel_crtc, pipe_config,
+						 intel_connector->panel.fitting_mode);
+		else
+			intel_pch_panel_fitting(intel_crtc, pipe_config,
+						intel_connector->panel.fitting_mode);
 	}
-	/* We need to take the panel's fixed mode into account. */
-	target_clock = adjusted_mode->clock;
 
 	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
 		return false;
@@ -701,12 +704,12 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 
 	/* Walk through all bpp values. Luckily they're all nicely spaced with 2
 	 * bpc in between. */
-	bpp = min_t(int, 8*3, pipe_config->pipe_bpp);
-	if (is_edp(intel_dp) && dev_priv->edp.bpp)
-		bpp = min_t(int, bpp, dev_priv->edp.bpp);
+	bpp = pipe_config->pipe_bpp;
+	if (is_edp(intel_dp) && dev_priv->vbt.edp_bpp)
+		bpp = min_t(int, bpp, dev_priv->vbt.edp_bpp);
 
 	for (; bpp >= 6*3; bpp -= 2*3) {
-		mode_rate = intel_dp_link_required(target_clock, bpp);
+		mode_rate = intel_dp_link_required(adjusted_mode->clock, bpp);
 
 		for (clock = 0; clock <= max_clock; clock++) {
 			for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) {
@@ -741,20 +744,21 @@ found:
 
 	intel_dp->link_bw = bws[clock];
 	intel_dp->lane_count = lane_count;
-	adjusted_mode->clock = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
 	pipe_config->pipe_bpp = bpp;
-	pipe_config->pixel_target_clock = target_clock;
+	pipe_config->port_clock = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
 
 	DRM_DEBUG_KMS("DP link bw %02x lane count %d clock %d bpp %d\n",
 		      intel_dp->link_bw, intel_dp->lane_count,
-		      adjusted_mode->clock, bpp);
+		      pipe_config->port_clock, bpp);
 	DRM_DEBUG_KMS("DP link bw required %i available %i\n",
 		      mode_rate, link_avail);
 
 	intel_link_compute_m_n(bpp, lane_count,
-			       target_clock, adjusted_mode->clock,
+			       adjusted_mode->clock, pipe_config->port_clock,
 			       &pipe_config->dp_m_n);
 
+	intel_dp_set_clock(encoder, pipe_config, intel_dp->link_bw);
+
 	return true;
 }
 
@@ -773,24 +777,28 @@ void intel_dp_init_link_config(struct intel_dp *intel_dp)
 	}
 }
 
-static void ironlake_set_pll_edp(struct drm_crtc *crtc, int clock)
+static void ironlake_set_pll_cpu_edp(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = crtc->dev;
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc);
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpa_ctl;
 
-	DRM_DEBUG_KMS("eDP PLL enable for clock %d\n", clock);
+	DRM_DEBUG_KMS("eDP PLL enable for clock %d\n", crtc->config.port_clock);
 	dpa_ctl = I915_READ(DP_A);
 	dpa_ctl &= ~DP_PLL_FREQ_MASK;
 
-	if (clock < 200000) {
+	if (crtc->config.port_clock == 162000) {
 		/* For a long time we've carried around a ILK-DevA w/a for the
 		 * 160MHz clock. If we're really unlucky, it's still required.
 		 */
 		DRM_DEBUG_KMS("160MHz cpu eDP clock, might need ilk devA w/a\n");
 		dpa_ctl |= DP_PLL_FREQ_160MHZ;
+		intel_dp->DP |= DP_PLL_FREQ_160MHZ;
 	} else {
 		dpa_ctl |= DP_PLL_FREQ_270MHZ;
+		intel_dp->DP |= DP_PLL_FREQ_270MHZ;
 	}
 
 	I915_WRITE(DP_A, dpa_ctl);
@@ -806,8 +814,8 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-	struct drm_crtc *crtc = encoder->crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->crtc);
 
 	/*
 	 * There are four kinds of DP registers:
@@ -833,21 +841,11 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 
 	/* Handle DP bits in common between all three register formats */
 	intel_dp->DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	intel_dp->DP |= DP_PORT_WIDTH(intel_dp->lane_count);
 
-	switch (intel_dp->lane_count) {
-	case 1:
-		intel_dp->DP |= DP_PORT_WIDTH_1;
-		break;
-	case 2:
-		intel_dp->DP |= DP_PORT_WIDTH_2;
-		break;
-	case 4:
-		intel_dp->DP |= DP_PORT_WIDTH_4;
-		break;
-	}
 	if (intel_dp->has_audio) {
 		DRM_DEBUG_DRIVER("Enabling DP audio on pipe %c\n",
-				 pipe_name(intel_crtc->pipe));
+				 pipe_name(crtc->pipe));
 		intel_dp->DP |= DP_AUDIO_OUTPUT_ENABLE;
 		intel_write_eld(encoder, adjusted_mode);
 	}
@@ -856,7 +854,7 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 
 	/* Split out the IBX/CPU vs CPT settings */
 
-	if (is_cpu_edp(intel_dp) && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
+	if (port == PORT_A && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
 		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
 			intel_dp->DP |= DP_SYNC_HS_HIGH;
 		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
@@ -866,14 +864,8 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		if (intel_dp->link_configuration[1] & DP_LANE_COUNT_ENHANCED_FRAME_EN)
 			intel_dp->DP |= DP_ENHANCED_FRAMING;
 
-		intel_dp->DP |= intel_crtc->pipe << 29;
-
-		/* don't miss out required setting for eDP */
-		if (adjusted_mode->clock < 200000)
-			intel_dp->DP |= DP_PLL_FREQ_160MHZ;
-		else
-			intel_dp->DP |= DP_PLL_FREQ_270MHZ;
-	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
+		intel_dp->DP |= crtc->pipe << 29;
+	} else if (!HAS_PCH_CPT(dev) || port == PORT_A) {
 		if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev))
 			intel_dp->DP |= intel_dp->color_range;
 
@@ -886,22 +878,14 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		if (intel_dp->link_configuration[1] & DP_LANE_COUNT_ENHANCED_FRAME_EN)
 			intel_dp->DP |= DP_ENHANCED_FRAMING;
 
-		if (intel_crtc->pipe == 1)
+		if (crtc->pipe == 1)
 			intel_dp->DP |= DP_PIPEB_SELECT;
-
-		if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
-			/* don't miss out required setting for eDP */
-			if (adjusted_mode->clock < 200000)
-				intel_dp->DP |= DP_PLL_FREQ_160MHZ;
-			else
-				intel_dp->DP |= DP_PLL_FREQ_270MHZ;
-		}
 	} else {
 		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
 	}
 
-	if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev))
-		ironlake_set_pll_edp(crtc, adjusted_mode->clock);
+	if (port == PORT_A && !IS_VALLEYVIEW(dev))
+		ironlake_set_pll_cpu_edp(intel_dp);
 }
 
 #define IDLE_ON_MASK		(PP_ON | 0 	  | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
@@ -1290,6 +1274,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 				  enum pipe *pipe)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 tmp = I915_READ(intel_dp->output_reg);
@@ -1297,9 +1282,9 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	if (!(tmp & DP_PORT_EN))
 		return false;
 
-	if (is_cpu_edp(intel_dp) && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
+	if (port == PORT_A && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
 		*pipe = PORT_TO_PIPE_CPT(tmp);
-	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
+	} else if (!HAS_PCH_CPT(dev) || port == PORT_A) {
 		*pipe = PORT_TO_PIPE(tmp);
 	} else {
 		u32 trans_sel;
@@ -1335,9 +1320,48 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_dp_get_config(struct intel_encoder *encoder,
+				struct intel_crtc_config *pipe_config)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	u32 tmp, flags = 0;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+
+	if ((port == PORT_A) || !HAS_PCH_CPT(dev)) {
+		tmp = I915_READ(intel_dp->output_reg);
+		if (tmp & DP_SYNC_HS_HIGH)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (tmp & DP_SYNC_VS_HIGH)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	} else {
+		tmp = I915_READ(TRANS_DP_CTL(crtc->pipe));
+		if (tmp & TRANS_DP_HSYNC_ACTIVE_HIGH)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (tmp & TRANS_DP_VSYNC_ACTIVE_HIGH)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	}
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct drm_device *dev = encoder->base.dev;
 
 	/* Make sure the panel is off before trying to change the mode. But also
 	 * ensure that we have vdd while we switch off the panel. */
@@ -1347,16 +1371,17 @@ static void intel_disable_dp(struct intel_encoder *encoder)
 	ironlake_edp_panel_off(intel_dp);
 
 	/* cpu edp my only be disable _after_ the cpu pipe/plane is disabled. */
-	if (!is_cpu_edp(intel_dp))
+	if (!(port == PORT_A || IS_VALLEYVIEW(dev)))
 		intel_dp_link_down(intel_dp);
 }
 
 static void intel_post_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
 
-	if (is_cpu_edp(intel_dp)) {
+	if (port == PORT_A || IS_VALLEYVIEW(dev)) {
 		intel_dp_link_down(intel_dp);
 		if (!IS_VALLEYVIEW(dev))
 			ironlake_edp_pll_off(intel_dp);
@@ -1381,15 +1406,73 @@ static void intel_enable_dp(struct intel_encoder *encoder)
 	intel_dp_complete_link_train(intel_dp);
 	intel_dp_stop_link_train(intel_dp);
 	ironlake_edp_backlight_on(intel_dp);
+
+	if (IS_VALLEYVIEW(dev)) {
+		struct intel_digital_port *dport =
+			enc_to_dig_port(&encoder->base);
+		int channel = vlv_dport_to_channel(dport);
+
+		vlv_wait_port_ready(dev_priv, channel);
+	}
 }
 
 static void intel_pre_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev))
+	if (dport->port == PORT_A && !IS_VALLEYVIEW(dev))
 		ironlake_edp_pll_on(intel_dp);
+
+	if (IS_VALLEYVIEW(dev)) {
+		struct intel_crtc *intel_crtc =
+			to_intel_crtc(encoder->base.crtc);
+		int port = vlv_dport_to_channel(dport);
+		int pipe = intel_crtc->pipe;
+		u32 val;
+
+		val = vlv_dpio_read(dev_priv, DPIO_DATA_LANE_A(port));
+		val = 0;
+		if (pipe)
+			val |= (1<<21);
+		else
+			val &= ~(1<<21);
+		val |= 0x001000c4;
+		vlv_dpio_write(dev_priv, DPIO_DATA_CHANNEL(port), val);
+
+		vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF0(port),
+				 0x00760018);
+		vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF8(port),
+				 0x00400888);
+	}
+}
+
+static void intel_dp_pre_pll_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int port = vlv_dport_to_channel(dport);
+
+	if (!IS_VALLEYVIEW(dev))
+		return;
+
+	/* Program Tx lane resets to default */
+	vlv_dpio_write(dev_priv, DPIO_PCS_TX(port),
+			 DPIO_PCS_TX_LANE2_RESET |
+			 DPIO_PCS_TX_LANE1_RESET);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLK(port),
+			 DPIO_PCS_CLK_CRI_RXEB_EIOS_EN |
+			 DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN |
+			 (1<<DPIO_PCS_CLK_DATAWIDTH_SHIFT) |
+				 DPIO_PCS_CLK_SOFT_RESET);
+
+	/* Fix up inter-pair skew failure */
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER1(port), 0x00750f00);
+	vlv_dpio_write(dev_priv, DPIO_TX_CTL(port), 0x00001500);
+	vlv_dpio_write(dev_priv, DPIO_TX_LANE(port), 0x40400000);
 }
 
 /*
@@ -1451,10 +1534,13 @@ static uint8_t
 intel_dp_voltage_max(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (IS_GEN7(dev) && is_cpu_edp(intel_dp))
+	if (IS_VALLEYVIEW(dev))
+		return DP_TRAIN_VOLTAGE_SWING_1200;
+	else if (IS_GEN7(dev) && port == PORT_A)
 		return DP_TRAIN_VOLTAGE_SWING_800;
-	else if (HAS_PCH_CPT(dev) && !is_cpu_edp(intel_dp))
+	else if (HAS_PCH_CPT(dev) && port != PORT_A)
 		return DP_TRAIN_VOLTAGE_SWING_1200;
 	else
 		return DP_TRAIN_VOLTAGE_SWING_800;
@@ -1464,6 +1550,7 @@ static uint8_t
 intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 
 	if (HAS_DDI(dev)) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
@@ -1477,7 +1564,19 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 		default:
 			return DP_TRAIN_PRE_EMPHASIS_0;
 		}
-	} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_VALLEYVIEW(dev)) {
+		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			return DP_TRAIN_PRE_EMPHASIS_9_5;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			return DP_TRAIN_PRE_EMPHASIS_6;
+		case DP_TRAIN_VOLTAGE_SWING_800:
+			return DP_TRAIN_PRE_EMPHASIS_3_5;
+		case DP_TRAIN_VOLTAGE_SWING_1200:
+		default:
+			return DP_TRAIN_PRE_EMPHASIS_0;
+		}
+	} else if (IS_GEN7(dev) && port == PORT_A) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
 		case DP_TRAIN_VOLTAGE_SWING_400:
 			return DP_TRAIN_PRE_EMPHASIS_6;
@@ -1502,6 +1601,101 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 	}
 }
 
+static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
+	unsigned long demph_reg_value, preemph_reg_value,
+		uniqtranscale_reg_value;
+	uint8_t train_set = intel_dp->train_set[0];
+	int port = vlv_dport_to_channel(dport);
+
+	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
+	case DP_TRAIN_PRE_EMPHASIS_0:
+		preemph_reg_value = 0x0004000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x2B405555;
+			uniqtranscale_reg_value = 0x552AB83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x5548B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_800:
+			demph_reg_value = 0x2B245555;
+			uniqtranscale_reg_value = 0x5560B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_1200:
+			demph_reg_value = 0x2B405555;
+			uniqtranscale_reg_value = 0x5598DA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPHASIS_3_5:
+		preemph_reg_value = 0x0002000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x5552B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			demph_reg_value = 0x2B404848;
+			uniqtranscale_reg_value = 0x5580B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_800:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPHASIS_6:
+		preemph_reg_value = 0x0000000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x2B305555;
+			uniqtranscale_reg_value = 0x5570B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			demph_reg_value = 0x2B2B4040;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPHASIS_9_5:
+		preemph_reg_value = 0x0006000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x1B405555;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	default:
+		return 0;
+	}
+
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port), 0x00000000);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL4(port), demph_reg_value);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL2(port),
+			 uniqtranscale_reg_value);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL3(port), 0x0C782040);
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER0(port), 0x00030000);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CTL_OVER1(port), preemph_reg_value);
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port), 0x80000000);
+
+	return 0;
+}
+
 static void
 intel_get_adjust_train(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE])
 {
@@ -1669,6 +1863,7 @@ static void
 intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	enum port port = intel_dig_port->port;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	uint32_t signal_levels, mask;
 	uint8_t train_set = intel_dp->train_set[0];
@@ -1676,10 +1871,13 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 	if (HAS_DDI(dev)) {
 		signal_levels = intel_hsw_signal_levels(train_set);
 		mask = DDI_BUF_EMP_MASK;
-	} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_VALLEYVIEW(dev)) {
+		signal_levels = intel_vlv_signal_levels(intel_dp);
+		mask = 0;
+	} else if (IS_GEN7(dev) && port == PORT_A) {
 		signal_levels = intel_gen7_edp_signal_levels(train_set);
 		mask = EDP_LINK_TRAIN_VOL_EMP_MASK_IVB;
-	} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
+	} else if (IS_GEN6(dev) && port == PORT_A) {
 		signal_levels = intel_gen6_edp_signal_levels(train_set);
 		mask = EDP_LINK_TRAIN_VOL_EMP_MASK_SNB;
 	} else {
@@ -1729,8 +1927,7 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 		}
 		I915_WRITE(DP_TP_CTL(port), temp);
 
-	} else if (HAS_PCH_CPT(dev) &&
-		   (IS_GEN7(dev) || !is_cpu_edp(intel_dp))) {
+	} else if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
 		dp_reg_value &= ~DP_LINK_TRAIN_MASK_CPT;
 
 		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
@@ -1981,6 +2178,7 @@ static void
 intel_dp_link_down(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	enum port port = intel_dig_port->port;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc =
@@ -2010,7 +2208,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 
 	DRM_DEBUG_KMS("\n");
 
-	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp))) {
+	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
 		DP &= ~DP_LINK_TRAIN_MASK_CPT;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE_CPT);
 	} else {
@@ -2301,11 +2499,10 @@ intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 			return NULL;
 
 		size = (intel_connector->edid->extensions + 1) * EDID_LENGTH;
-		edid = kmalloc(size, GFP_KERNEL);
+		edid = kmemdup(intel_connector->edid, size, GFP_KERNEL);
 		if (!edid)
 			return NULL;
 
-		memcpy(edid, intel_connector->edid, size);
 		return edid;
 	}
 
@@ -2499,15 +2696,16 @@ done:
 }
 
 static void
-intel_dp_destroy(struct drm_connector *connector)
+intel_dp_connector_destroy(struct drm_connector *connector)
 {
-	struct intel_dp *intel_dp = intel_attached_dp(connector);
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 
 	if (!IS_ERR_OR_NULL(intel_connector->edid))
 		kfree(intel_connector->edid);
 
-	if (is_edp(intel_dp))
+	/* Can't call is_edp() since the encoder may have been destroyed
+	 * already. */
+	if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
 		intel_panel_fini(&intel_connector->panel);
 
 	drm_sysfs_connector_remove(connector);
@@ -2541,7 +2739,7 @@ static const struct drm_connector_funcs intel_dp_connector_funcs = {
 	.detect = intel_dp_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_dp_set_property,
-	.destroy = intel_dp_destroy,
+	.destroy = intel_dp_connector_destroy,
 };
 
 static const struct drm_connector_helper_funcs intel_dp_connector_helper_funcs = {
@@ -2588,11 +2786,11 @@ bool intel_dpd_is_edp(struct drm_device *dev)
 	struct child_device_config *p_child;
 	int i;
 
-	if (!dev_priv->child_dev_num)
+	if (!dev_priv->vbt.child_dev_num)
 		return false;
 
-	for (i = 0; i < dev_priv->child_dev_num; i++) {
-		p_child = dev_priv->child_dev + i;
+	for (i = 0; i < dev_priv->vbt.child_dev_num; i++) {
+		p_child = dev_priv->vbt.child_dev + i;
 
 		if (p_child->dvo_port == PORT_IDPD &&
 		    p_child->device_type == DEVICE_TYPE_eDP)
@@ -2670,7 +2868,7 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	DRM_DEBUG_KMS("cur t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
 		      cur.t1_t3, cur.t8, cur.t9, cur.t10, cur.t11_t12);
 
-	vbt = dev_priv->edp.pps;
+	vbt = dev_priv->vbt.edp_pps;
 
 	/* Upper limits from eDP 1.3 spec. Note that we use the clunky units of
 	 * our hw here, which are all in 100usec. */
@@ -2738,9 +2936,6 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 		pp_div_reg = PIPEA_PP_DIVISOR;
 	}
 
-	if (IS_VALLEYVIEW(dev))
-		port_sel = I915_READ(pp_on_reg) & 0xc0000000;
-
 	/* And finally store the new values in the power sequencer. */
 	pp_on = (seq->t1_t3 << PANEL_POWER_UP_DELAY_SHIFT) |
 		(seq->t8 << PANEL_LIGHT_ON_DELAY_SHIFT);
@@ -2754,8 +2949,10 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 
 	/* Haswell doesn't have any port selection bits for the panel
 	 * power sequencer any more. */
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		if (is_cpu_edp(intel_dp))
+	if (IS_VALLEYVIEW(dev)) {
+		port_sel = I915_READ(pp_on_reg) & 0xc0000000;
+	} else if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
+		if (dp_to_dig_port(intel_dp)->port == PORT_A)
 			port_sel = PANEL_POWER_PORT_DP_A;
 		else
 			port_sel = PANEL_POWER_PORT_DP_D;
@@ -2773,7 +2970,85 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 		      I915_READ(pp_div_reg));
 }
 
-void
+static bool intel_edp_init_connector(struct intel_dp *intel_dp,
+				     struct intel_connector *intel_connector)
+{
+	struct drm_connector *connector = &intel_connector->base;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_display_mode *fixed_mode = NULL;
+	struct edp_power_seq power_seq = { 0 };
+	bool has_dpcd;
+	struct drm_display_mode *scan;
+	struct edid *edid;
+
+	if (!is_edp(intel_dp))
+		return true;
+
+	intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+
+	/* Cache DPCD and EDID for edp. */
+	ironlake_edp_panel_vdd_on(intel_dp);
+	has_dpcd = intel_dp_get_dpcd(intel_dp);
+	ironlake_edp_panel_vdd_off(intel_dp, false);
+
+	if (has_dpcd) {
+		if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
+			dev_priv->no_aux_handshake =
+				intel_dp->dpcd[DP_MAX_DOWNSPREAD] &
+				DP_NO_AUX_HANDSHAKE_LINK_TRAINING;
+	} else {
+		/* if this fails, presume the device is a ghost */
+		DRM_INFO("failed to retrieve link info, disabling eDP\n");
+		return false;
+	}
+
+	/* We now know it's not a ghost, init power sequence regs. */
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
+						      &power_seq);
+
+	ironlake_edp_panel_vdd_on(intel_dp);
+	edid = drm_get_edid(connector, &intel_dp->adapter);
+	if (edid) {
+		if (drm_add_edid_modes(connector, edid)) {
+			drm_mode_connector_update_edid_property(connector,
+								edid);
+			drm_edid_to_eld(connector, edid);
+		} else {
+			kfree(edid);
+			edid = ERR_PTR(-EINVAL);
+		}
+	} else {
+		edid = ERR_PTR(-ENOENT);
+	}
+	intel_connector->edid = edid;
+
+	/* prefer fixed mode from EDID if available */
+	list_for_each_entry(scan, &connector->probed_modes, head) {
+		if ((scan->type & DRM_MODE_TYPE_PREFERRED)) {
+			fixed_mode = drm_mode_duplicate(dev, scan);
+			break;
+		}
+	}
+
+	/* fallback to VBT if available for eDP */
+	if (!fixed_mode && dev_priv->vbt.lfp_lvds_vbt_mode) {
+		fixed_mode = drm_mode_duplicate(dev,
+					dev_priv->vbt.lfp_lvds_vbt_mode);
+		if (fixed_mode)
+			fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
+	}
+
+	ironlake_edp_panel_vdd_off(intel_dp, false);
+
+	intel_panel_init(&intel_connector->panel, fixed_mode);
+	intel_panel_setup_backlight(connector);
+
+	return true;
+}
+
+bool
 intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 			struct intel_connector *intel_connector)
 {
@@ -2782,38 +3057,47 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_display_mode *fixed_mode = NULL;
-	struct edp_power_seq power_seq = { 0 };
 	enum port port = intel_dig_port->port;
 	const char *name = NULL;
-	int type;
+	int type, error;
 
 	/* Preserve the current hw state. */
 	intel_dp->DP = I915_READ(intel_dp->output_reg);
 	intel_dp->attached_connector = intel_connector;
 
-	if (HAS_PCH_SPLIT(dev) && port == PORT_D)
-		if (intel_dpd_is_edp(dev))
-			intel_dp->is_pch_edp = true;
-
+	type = DRM_MODE_CONNECTOR_DisplayPort;
 	/*
 	 * FIXME : We need to initialize built-in panels before external panels.
 	 * For X0, DP_C is fixed as eDP. Revisit this as part of VLV eDP cleanup
 	 */
-	if (IS_VALLEYVIEW(dev) && port == PORT_C) {
-		type = DRM_MODE_CONNECTOR_eDP;
-		intel_encoder->type = INTEL_OUTPUT_EDP;
-	} else if (port == PORT_A || is_pch_edp(intel_dp)) {
+	switch (port) {
+	case PORT_A:
 		type = DRM_MODE_CONNECTOR_eDP;
-		intel_encoder->type = INTEL_OUTPUT_EDP;
-	} else {
-		/* The intel_encoder->type value may be INTEL_OUTPUT_UNKNOWN for
-		 * DDI or INTEL_OUTPUT_DISPLAYPORT for the older gens, so don't
-		 * rewrite it.
-		 */
-		type = DRM_MODE_CONNECTOR_DisplayPort;
+		break;
+	case PORT_C:
+		if (IS_VALLEYVIEW(dev))
+			type = DRM_MODE_CONNECTOR_eDP;
+		break;
+	case PORT_D:
+		if (HAS_PCH_SPLIT(dev) && intel_dpd_is_edp(dev))
+			type = DRM_MODE_CONNECTOR_eDP;
+		break;
+	default:	/* silence GCC warning */
+		break;
 	}
 
+	/*
+	 * For eDP we always set the encoder type to INTEL_OUTPUT_EDP, but
+	 * for DP the encoder type can be set by the caller to
+	 * INTEL_OUTPUT_UNKNOWN for DDI, so don't rewrite it.
+	 */
+	if (type == DRM_MODE_CONNECTOR_eDP)
+		intel_encoder->type = INTEL_OUTPUT_EDP;
+
+	DRM_DEBUG_KMS("Adding %s connector on port %c\n",
+			type == DRM_MODE_CONNECTOR_eDP ? "eDP" : "DP",
+			port_name(port));
+
 	drm_connector_init(dev, connector, &intel_dp_connector_funcs, type);
 	drm_connector_helper_add(connector, &intel_dp_connector_helper_funcs);
 
@@ -2873,74 +3157,21 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		BUG();
 	}
 
-	if (is_edp(intel_dp))
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
-
-	intel_dp_i2c_init(intel_dp, intel_connector, name);
-
-	/* Cache DPCD and EDID for edp. */
-	if (is_edp(intel_dp)) {
-		bool ret;
-		struct drm_display_mode *scan;
-		struct edid *edid;
-
-		ironlake_edp_panel_vdd_on(intel_dp);
-		ret = intel_dp_get_dpcd(intel_dp);
-		ironlake_edp_panel_vdd_off(intel_dp, false);
+	error = intel_dp_i2c_init(intel_dp, intel_connector, name);
+	WARN(error, "intel_dp_i2c_init failed with error %d for port %c\n",
+	     error, port_name(port));
 
-		if (ret) {
-			if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
-				dev_priv->no_aux_handshake =
-					intel_dp->dpcd[DP_MAX_DOWNSPREAD] &
-					DP_NO_AUX_HANDSHAKE_LINK_TRAINING;
-		} else {
-			/* if this fails, presume the device is a ghost */
-			DRM_INFO("failed to retrieve link info, disabling eDP\n");
-			intel_dp_encoder_destroy(&intel_encoder->base);
-			intel_dp_destroy(connector);
-			return;
-		}
-
-		/* We now know it's not a ghost, init power sequence regs. */
-		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
-							      &power_seq);
-
-		ironlake_edp_panel_vdd_on(intel_dp);
-		edid = drm_get_edid(connector, &intel_dp->adapter);
-		if (edid) {
-			if (drm_add_edid_modes(connector, edid)) {
-				drm_mode_connector_update_edid_property(connector, edid);
-				drm_edid_to_eld(connector, edid);
-			} else {
-				kfree(edid);
-				edid = ERR_PTR(-EINVAL);
-			}
-		} else {
-			edid = ERR_PTR(-ENOENT);
+	if (!intel_edp_init_connector(intel_dp, intel_connector)) {
+		i2c_del_adapter(&intel_dp->adapter);
+		if (is_edp(intel_dp)) {
+			cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
+			mutex_lock(&dev->mode_config.mutex);
+			ironlake_panel_vdd_off_sync(intel_dp);
+			mutex_unlock(&dev->mode_config.mutex);
 		}
-		intel_connector->edid = edid;
-
-		/* prefer fixed mode from EDID if available */
-		list_for_each_entry(scan, &connector->probed_modes, head) {
-			if ((scan->type & DRM_MODE_TYPE_PREFERRED)) {
-				fixed_mode = drm_mode_duplicate(dev, scan);
-				break;
-			}
-		}
-
-		/* fallback to VBT if available for eDP */
-		if (!fixed_mode && dev_priv->lfp_lvds_vbt_mode) {
-			fixed_mode = drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode);
-			if (fixed_mode)
-				fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
-		}
-
-		ironlake_edp_panel_vdd_off(intel_dp, false);
-	}
-
-	if (is_edp(intel_dp)) {
-		intel_panel_init(&intel_connector->panel, fixed_mode);
-		intel_panel_setup_backlight(connector);
+		drm_sysfs_connector_remove(connector);
+		drm_connector_cleanup(connector);
+		return false;
 	}
 
 	intel_dp_add_properties(intel_dp, connector);
@@ -2953,6 +3184,8 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		u32 temp = I915_READ(PEG_BAND_GAP_DATA);
 		I915_WRITE(PEG_BAND_GAP_DATA, (temp & ~0xf) | 0xd);
 	}
+
+	return true;
 }
 
 void
@@ -2986,6 +3219,9 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	intel_encoder->disable = intel_disable_dp;
 	intel_encoder->post_disable = intel_post_disable_dp;
 	intel_encoder->get_hw_state = intel_dp_get_hw_state;
+	intel_encoder->get_config = intel_dp_get_config;
+	if (IS_VALLEYVIEW(dev))
+		intel_encoder->pre_pll_enable = intel_dp_pre_pll_enable;
 
 	intel_dig_port->port = port;
 	intel_dig_port->dp.output_reg = output_reg;
@@ -2995,5 +3231,9 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	intel_encoder->cloneable = false;
 	intel_encoder->hot_plug = intel_dp_hot_plug;
 
-	intel_dp_init_connector(intel_dig_port, intel_connector);
+	if (!intel_dp_init_connector(intel_dig_port, intel_connector)) {
+		drm_encoder_cleanup(encoder);
+		kfree(intel_dig_port);
+		kfree(intel_connector);
+	}
 }
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 624a9e6..c8c9b6f 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -120,7 +120,6 @@ struct intel_encoder {
 	struct intel_crtc *new_crtc;
 
 	int type;
-	bool needs_tv_clock;
 	/*
 	 * Intel hw has only one MUX where encoders could be clone, hence a
 	 * simple flag is enough to compute the possible_clones mask.
@@ -140,6 +139,12 @@ struct intel_encoder {
 	 * the encoder is active. If the encoder is enabled it also set the pipe
 	 * it is connected to in the pipe parameter. */
 	bool (*get_hw_state)(struct intel_encoder *, enum pipe *pipe);
+	/* Reconstructs the equivalent mode flags for the current hardware
+	 * state. This must be called _after_ display->get_pipe_config has
+	 * pre-filled the pipe config. Note that intel_encoder->base.crtc must
+	 * be set correctly before calling this function. */
+	void (*get_config)(struct intel_encoder *,
+			   struct intel_crtc_config *pipe_config);
 	int crtc_mask;
 	enum hpd_pin hpd_pin;
 };
@@ -177,7 +182,30 @@ struct intel_connector {
 	u8 polled;
 };
 
+typedef struct dpll {
+	/* given values */
+	int n;
+	int m1, m2;
+	int p1, p2;
+	/* derived values */
+	int	dot;
+	int	vco;
+	int	m;
+	int	p;
+} intel_clock_t;
+
 struct intel_crtc_config {
+	/**
+	 * quirks - bitfield with hw state readout quirks
+	 *
+	 * For various reasons the hw state readout code might not be able to
+	 * completely faithfully read out the current state. These cases are
+	 * tracked with quirk flags so that fastboot and state checker can act
+	 * accordingly.
+	 */
+#define PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS (1<<0) /* unreliable sync mode.flags */
+	unsigned long quirks;
+
 	struct drm_display_mode requested_mode;
 	struct drm_display_mode adjusted_mode;
 	/* This flag must be set by the encoder's compute_config callback if it
@@ -201,29 +229,67 @@ struct intel_crtc_config {
 	/* DP has a bunch of special case unfortunately, so mark the pipe
 	 * accordingly. */
 	bool has_dp_encoder;
+
+	/*
+	 * Enable dithering, used when the selected pipe bpp doesn't match the
+	 * plane bpp.
+	 */
 	bool dither;
 
 	/* Controls for the clock computation, to override various stages. */
 	bool clock_set;
 
+	/* SDVO TV has a bunch of special case. To make multifunction encoders
+	 * work correctly, we need to track this at runtime.*/
+	bool sdvo_tv_clock;
+
+	/*
+	 * crtc bandwidth limit, don't increase pipe bpp or clock if not really
+	 * required. This is set in the 2nd loop of calling encoder's
+	 * ->compute_config if the first pick doesn't work out.
+	 */
+	bool bw_constrained;
+
 	/* Settings for the intel dpll used on pretty much everything but
 	 * haswell. */
-	struct dpll {
-		unsigned n;
-		unsigned m1, m2;
-		unsigned p1, p2;
-	} dpll;
+	struct dpll dpll;
+
+	/* Selected dpll when shared or DPLL_ID_PRIVATE. */
+	enum intel_dpll_id shared_dpll;
+
+	/* Actual register state of the dpll, for shared dpll cross-checking. */
+	struct intel_dpll_hw_state dpll_hw_state;
 
 	int pipe_bpp;
 	struct intel_link_m_n dp_m_n;
-	/**
-	 * This is currently used by DP and HDMI encoders since those can have a
-	 * target pixel clock != the port link clock (which is currently stored
-	 * in adjusted_mode->clock).
+
+	/*
+	 * Frequence the dpll for the port should run at. Differs from the
+	 * adjusted dotclock e.g. for DP or 12bpc hdmi mode.
 	 */
-	int pixel_target_clock;
+	int port_clock;
+
 	/* Used by SDVO (and if we ever fix it, HDMI). */
 	unsigned pixel_multiplier;
+
+	/* Panel fitter controls for gen2-gen4 + VLV */
+	struct {
+		u32 control;
+		u32 pgm_ratios;
+		u32 lvds_border_bits;
+	} gmch_pfit;
+
+	/* Panel fitter placement and size for Ironlake+ */
+	struct {
+		u32 pos;
+		u32 size;
+	} pch_pfit;
+
+	/* FDI configuration, only valid if has_pch_encoder is set. */
+	int fdi_lanes;
+	struct intel_link_m_n fdi_m_n;
+
+	bool ips_enabled;
 };
 
 struct intel_crtc {
@@ -242,7 +308,6 @@ struct intel_crtc {
 	bool lowfreq_avail;
 	struct intel_overlay *overlay;
 	struct intel_unpin_work *unpin_work;
-	int fdi_lanes;
 
 	atomic_t unpin_work_count;
 
@@ -259,12 +324,14 @@ struct intel_crtc {
 
 	struct intel_crtc_config config;
 
-	/* We can share PLLs across outputs if the timings match */
-	struct intel_pch_pll *pch_pll;
 	uint32_t ddi_pll_sel;
 
 	/* reset counter value when the last flip was submitted */
 	unsigned int reset_counter;
+
+	/* Access to these should be protected by dev_priv->irq_lock. */
+	bool cpu_fifo_underrun_disabled;
+	bool pch_fifo_underrun_disabled;
 };
 
 struct intel_plane {
@@ -279,6 +346,18 @@ struct intel_plane {
 	unsigned int crtc_w, crtc_h;
 	uint32_t src_x, src_y;
 	uint32_t src_w, src_h;
+
+	/* Since we need to change the watermarks before/after
+	 * enabling/disabling the planes, we need to store the parameters here
+	 * as the other pieces of the struct may not reflect the values we want
+	 * for the watermark calculations. Currently only Haswell uses this.
+	 */
+	struct {
+		bool enable;
+		uint8_t bytes_per_pixel;
+		uint32_t horiz_pixels;
+	} wm;
+
 	void (*update_plane)(struct drm_plane *plane,
 			     struct drm_framebuffer *fb,
 			     struct drm_i915_gem_object *obj,
@@ -411,7 +490,6 @@ struct intel_dp {
 	uint8_t downstream_ports[DP_MAX_DOWNSTREAM_PORTS];
 	struct i2c_adapter adapter;
 	struct i2c_algo_dp_aux_data algo;
-	bool is_pch_edp;
 	uint8_t train_set[4];
 	int panel_power_up_delay;
 	int panel_power_down_delay;
@@ -431,6 +509,19 @@ struct intel_digital_port {
 	struct intel_hdmi hdmi;
 };
 
+static inline int
+vlv_dport_to_channel(struct intel_digital_port *dport)
+{
+	switch (dport->port) {
+	case PORT_B:
+		return 0;
+	case PORT_C:
+		return 1;
+	default:
+		BUG();
+	}
+}
+
 static inline struct drm_crtc *
 intel_get_crtc_for_pipe(struct drm_device *dev, int pipe)
 {
@@ -474,6 +565,7 @@ int intel_ddc_get_modes(struct drm_connector *c, struct i2c_adapter *adapter);
 extern void intel_attach_force_audio_property(struct drm_connector *connector);
 extern void intel_attach_broadcast_rgb_property(struct drm_connector *connector);
 
+extern bool intel_pipe_has_type(struct drm_crtc *crtc, int type);
 extern void intel_crt_init(struct drm_device *dev);
 extern void intel_hdmi_init(struct drm_device *dev,
 			    int hdmi_reg, enum port port);
@@ -488,13 +580,14 @@ extern bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg,
 extern void intel_dvo_init(struct drm_device *dev);
 extern void intel_tv_init(struct drm_device *dev);
 extern void intel_mark_busy(struct drm_device *dev);
-extern void intel_mark_fb_busy(struct drm_i915_gem_object *obj);
+extern void intel_mark_fb_busy(struct drm_i915_gem_object *obj,
+			       struct intel_ring_buffer *ring);
 extern void intel_mark_idle(struct drm_device *dev);
-extern bool intel_lvds_init(struct drm_device *dev);
+extern void intel_lvds_init(struct drm_device *dev);
 extern bool intel_is_dual_link_lvds(struct drm_device *dev);
 extern void intel_dp_init(struct drm_device *dev, int output_reg,
 			  enum port port);
-extern void intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
+extern bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 				    struct intel_connector *intel_connector);
 extern void intel_dp_init_link_config(struct intel_dp *intel_dp);
 extern void intel_dp_start_link_train(struct intel_dp *intel_dp);
@@ -512,7 +605,6 @@ extern void ironlake_edp_panel_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_off(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
-extern bool intel_encoder_is_pch_edp(struct drm_encoder *encoder);
 extern int intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane);
 extern void intel_flush_display_plane(struct drm_i915_private *dev_priv,
 				      enum plane plane);
@@ -524,12 +616,14 @@ extern void intel_panel_fini(struct intel_panel *panel);
 
 extern void intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
 				   struct drm_display_mode *adjusted_mode);
-extern void intel_pch_panel_fitting(struct drm_device *dev,
-				    int fitting_mode,
-				    const struct drm_display_mode *mode,
-				    struct drm_display_mode *adjusted_mode);
-extern u32 intel_panel_get_max_backlight(struct drm_device *dev);
-extern void intel_panel_set_backlight(struct drm_device *dev, u32 level);
+extern void intel_pch_panel_fitting(struct intel_crtc *crtc,
+				    struct intel_crtc_config *pipe_config,
+				    int fitting_mode);
+extern void intel_gmch_panel_fitting(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config,
+				     int fitting_mode);
+extern void intel_panel_set_backlight(struct drm_device *dev,
+				      u32 level, u32 max);
 extern int intel_panel_setup_backlight(struct drm_connector *connector);
 extern void intel_panel_enable_backlight(struct drm_device *dev,
 					 enum pipe pipe);
@@ -553,11 +647,11 @@ extern void intel_crtc_load_lut(struct drm_crtc *crtc);
 extern void intel_crtc_update_dpms(struct drm_crtc *crtc);
 extern void intel_encoder_destroy(struct drm_encoder *encoder);
 extern void intel_encoder_dpms(struct intel_encoder *encoder, int mode);
-extern bool intel_encoder_check_is_cloned(struct intel_encoder *encoder);
 extern void intel_connector_dpms(struct drm_connector *, int mode);
 extern bool intel_connector_get_hw_state(struct intel_connector *connector);
 extern void intel_modeset_check_state(struct drm_device *dev);
 extern void intel_plane_restore(struct drm_plane *plane);
+extern void intel_plane_disable(struct drm_plane *plane);
 
 
 static inline struct intel_encoder *intel_attached_encoder(struct drm_connector *connector)
@@ -565,19 +659,17 @@ static inline struct intel_encoder *intel_attached_encoder(struct drm_connector
 	return to_intel_connector(connector)->encoder;
 }
 
-static inline struct intel_dp *enc_to_intel_dp(struct drm_encoder *encoder)
-{
-	struct intel_digital_port *intel_dig_port =
-		container_of(encoder, struct intel_digital_port, base.base);
-	return &intel_dig_port->dp;
-}
-
 static inline struct intel_digital_port *
 enc_to_dig_port(struct drm_encoder *encoder)
 {
 	return container_of(encoder, struct intel_digital_port, base.base);
 }
 
+static inline struct intel_dp *enc_to_intel_dp(struct drm_encoder *encoder)
+{
+	return &enc_to_dig_port(encoder)->dp;
+}
+
 static inline struct intel_digital_port *
 dp_to_dig_port(struct intel_dp *intel_dp)
 {
@@ -607,6 +699,7 @@ intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 extern void intel_wait_for_vblank(struct drm_device *dev, int pipe);
 extern void intel_wait_for_pipe_off(struct drm_device *dev, int pipe);
 extern int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp);
+extern void vlv_wait_port_ready(struct drm_i915_private *dev_priv, int port);
 
 struct intel_load_detect_pipe {
 	struct drm_framebuffer *release_fb;
@@ -660,13 +753,9 @@ extern void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 #define assert_pipe_disabled(d, p) assert_pipe(d, p, false)
 
 extern void intel_init_clock_gating(struct drm_device *dev);
+extern void intel_suspend_hw(struct drm_device *dev);
 extern void intel_write_eld(struct drm_encoder *encoder,
 			    struct drm_display_mode *mode);
-extern void intel_cpt_verify_modeset(struct drm_device *dev, int pipe);
-extern void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n);
-extern void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n);
 extern void intel_prepare_ddi(struct drm_device *dev);
 extern void hsw_fdi_link_train(struct drm_crtc *crtc);
 extern void intel_ddi_init(struct drm_device *dev, enum port port);
@@ -675,9 +764,7 @@ extern void intel_ddi_init(struct drm_device *dev, enum port port);
 extern void intel_update_watermarks(struct drm_device *dev);
 extern void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
 					   uint32_t sprite_width,
-					   int pixel_size);
-extern void intel_update_linetime_watermarks(struct drm_device *dev, int pipe,
-			 struct drm_display_mode *mode);
+					   int pixel_size, bool enable);
 
 extern unsigned long intel_gen4_compute_page_offset(int *x, int *y,
 						    unsigned int tiling_mode,
@@ -689,8 +776,6 @@ extern int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 extern int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 				     struct drm_file *file_priv);
 
-extern u32 intel_dpio_read(struct drm_i915_private *dev_priv, int reg);
-
 /* Power-related functions, located in intel_pm.c */
 extern void intel_init_pm(struct drm_device *dev);
 /* FBC */
@@ -701,7 +786,12 @@ extern void intel_update_fbc(struct drm_device *dev);
 extern void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
 extern void intel_gpu_ips_teardown(void);
 
-extern bool intel_using_power_well(struct drm_device *dev);
+/* Power well */
+extern int i915_init_power_well(struct drm_device *dev);
+extern void i915_remove_power_well(struct drm_device *dev);
+
+extern bool intel_display_power_enabled(struct drm_device *dev,
+					enum intel_display_power_domain domain);
 extern void intel_init_power_well(struct drm_device *dev);
 extern void intel_set_power_well(struct drm_device *dev, bool enable);
 extern void intel_enable_gt_powersave(struct drm_device *dev);
@@ -719,7 +809,7 @@ extern void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 extern void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc);
 extern void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc);
 extern void intel_ddi_setup_hw_pll_state(struct drm_device *dev);
-extern bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock);
+extern bool intel_ddi_pll_mode_set(struct drm_crtc *crtc);
 extern void intel_ddi_put_crtc_pll(struct drm_crtc *crtc);
 extern void intel_ddi_set_pipe_settings(struct drm_crtc *crtc);
 extern void intel_ddi_prepare_link_retrain(struct drm_encoder *encoder);
@@ -728,5 +818,11 @@ intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector);
 extern void intel_ddi_fdi_disable(struct drm_crtc *crtc);
 
 extern void intel_display_handle_reset(struct drm_device *dev);
+extern bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+						  enum pipe pipe,
+						  bool enable);
+extern bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
+						 enum transcoder pch_transcoder,
+						 bool enable);
 
 #endif /* __INTEL_DRV_H__ */
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index cc70b16..eb2020e 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -54,6 +54,13 @@ static const struct intel_dvo_device intel_dvo_devices[] = {
 		.dev_ops = &ch7xxx_ops,
 	},
 	{
+		.type = INTEL_DVO_CHIP_TMDS,
+		.name = "ch7xxx",
+		.dvo_reg = DVOC,
+		.slave_addr = 0x75, /* For some ch7010 */
+		.dev_ops = &ch7xxx_ops,
+	},
+	{
 		.type = INTEL_DVO_CHIP_LVDS,
 		.name = "ivch",
 		.dvo_reg = DVOA,
@@ -129,6 +136,26 @@ static bool intel_dvo_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_dvo_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_dvo *intel_dvo = enc_to_intel_dvo(&encoder->base);
+	u32 tmp, flags = 0;
+
+	tmp = I915_READ(intel_dvo->dev.dvo_reg);
+	if (tmp & DVO_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (tmp & DVO_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_disable_dvo(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
@@ -153,6 +180,7 @@ static void intel_enable_dvo(struct intel_encoder *encoder)
 	intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
 }
 
+/* Special dpms function to support cloning between dvo/sdvo/crt. */
 static void intel_dvo_dpms(struct drm_connector *connector, int mode)
 {
 	struct intel_dvo *intel_dvo = intel_attached_dvo(connector);
@@ -174,6 +202,8 @@ static void intel_dvo_dpms(struct drm_connector *connector, int mode)
 		return;
 	}
 
+	/* We call connector dpms manually below in case pipe dpms doesn't
+	 * change due to cloning. */
 	if (mode == DRM_MODE_DPMS_ON) {
 		intel_dvo->base.connectors_active = true;
 
@@ -440,6 +470,7 @@ void intel_dvo_init(struct drm_device *dev)
 	intel_encoder->disable = intel_disable_dvo;
 	intel_encoder->enable = intel_enable_dvo;
 	intel_encoder->get_hw_state = intel_dvo_get_hw_state;
+	intel_encoder->get_config = intel_dvo_get_config;
 	intel_connector->get_hw_state = intel_dvo_connector_get_hw_state;
 
 	/* Now, try to find a controller */
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 6b7c3ca..dff669e 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -60,8 +60,9 @@ static struct fb_ops intelfb_ops = {
 static int intelfb_create(struct drm_fb_helper *helper,
 			  struct drm_fb_helper_surface_size *sizes)
 {
-	struct intel_fbdev *ifbdev = (struct intel_fbdev *)helper;
-	struct drm_device *dev = ifbdev->helper.dev;
+	struct intel_fbdev *ifbdev =
+		container_of(helper, struct intel_fbdev, helper);
+	struct drm_device *dev = helper->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct fb_info *info;
 	struct drm_framebuffer *fb;
@@ -108,7 +109,7 @@ static int intelfb_create(struct drm_fb_helper *helper,
 		goto out_unpin;
 	}
 
-	info->par = ifbdev;
+	info->par = helper;
 
 	ret = intel_framebuffer_init(dev, &ifbdev->ifb, &mode_cmd, obj);
 	if (ret)
@@ -217,7 +218,7 @@ static void intel_fbdev_destroy(struct drm_device *dev,
 int intel_fbdev_init(struct drm_device *dev)
 {
 	struct intel_fbdev *ifbdev;
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
 	ifbdev = kzalloc(sizeof(struct intel_fbdev), GFP_KERNEL);
@@ -242,7 +243,7 @@ int intel_fbdev_init(struct drm_device *dev)
 
 void intel_fbdev_initial_config(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	drm_fb_helper_initial_config(&dev_priv->fbdev->helper, 32);
@@ -250,7 +251,7 @@ void intel_fbdev_initial_config(struct drm_device *dev)
 
 void intel_fbdev_fini(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	if (!dev_priv->fbdev)
 		return;
 
@@ -261,7 +262,7 @@ void intel_fbdev_fini(struct drm_device *dev)
 
 void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 	struct fb_info *info;
 
@@ -274,7 +275,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 	 * been restored from swap. If the object is stolen however, it will be
 	 * full of whatever garbage was left in there.
 	 */
-	if (!state && ifbdev->ifb.obj->stolen)
+	if (state == FBINFO_STATE_RUNNING && ifbdev->ifb.obj->stolen)
 		memset_io(info->screen_base, 0, info->screen_size);
 
 	fb_set_suspend(info, state);
@@ -284,16 +285,14 @@ MODULE_LICENSE("GPL and additional rights");
 
 void intel_fb_output_poll_changed(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	drm_fb_helper_hotplug_event(&dev_priv->fbdev->helper);
 }
 
 void intel_fb_restore_mode(struct drm_device *dev)
 {
 	int ret;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_plane *plane;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	if (INTEL_INFO(dev)->num_pipes == 0)
 		return;
@@ -304,10 +303,5 @@ void intel_fb_restore_mode(struct drm_device *dev)
 	if (ret)
 		DRM_DEBUG("failed to restore crtc mode\n");
 
-	/* Be sure to shut off any planes that may be active */
-	list_for_each_entry(plane, &config->plane_list, head)
-		if (plane->enabled)
-			plane->funcs->disable_plane(plane);
-
 	drm_modeset_unlock_all(dev);
 }
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index a905793..98df2a0 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -602,7 +602,7 @@ static void intel_hdmi_mode_set(struct drm_encoder *encoder,
 	u32 hdmi_val;
 
 	hdmi_val = SDVO_ENCODING_HDMI;
-	if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev))
+	if (!HAS_PCH_SPLIT(dev))
 		hdmi_val |= intel_hdmi->color_range;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
 		hdmi_val |= SDVO_VSYNC_ACTIVE_HIGH;
@@ -658,6 +658,28 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_hdmi_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_config *pipe_config)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	u32 tmp, flags = 0;
+
+	tmp = I915_READ(intel_hdmi->hdmi_reg);
+
+	if (tmp & SDVO_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (tmp & SDVO_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_enable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
@@ -697,6 +719,14 @@ static void intel_enable_hdmi(struct intel_encoder *encoder)
 		I915_WRITE(intel_hdmi->hdmi_reg, temp);
 		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
+
+	if (IS_VALLEYVIEW(dev)) {
+		struct intel_digital_port *dport =
+			enc_to_dig_port(&encoder->base);
+		int channel = vlv_dport_to_channel(dport);
+
+		vlv_wait_port_ready(dev_priv, channel);
+	}
 }
 
 static void intel_disable_hdmi(struct intel_encoder *encoder)
@@ -775,6 +805,8 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	int clock_12bpc = pipe_config->requested_mode.clock * 3 / 2;
+	int desired_bpp;
 
 	if (intel_hdmi->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
@@ -794,14 +826,29 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 	/*
 	 * HDMI is either 12 or 8, so if the display lets 10bpc sneak
 	 * through, clamp it down. Note that g4x/vlv don't support 12bpc hdmi
-	 * outputs.
+	 * outputs. We also need to check that the higher clock still fits
+	 * within limits.
 	 */
-	if (pipe_config->pipe_bpp > 8*3 && HAS_PCH_SPLIT(dev)) {
-		DRM_DEBUG_KMS("forcing bpc to 12 for HDMI\n");
-		pipe_config->pipe_bpp = 12*3;
+	if (pipe_config->pipe_bpp > 8*3 && clock_12bpc <= 225000
+	    && HAS_PCH_SPLIT(dev)) {
+		DRM_DEBUG_KMS("picking bpc to 12 for HDMI output\n");
+		desired_bpp = 12*3;
+
+		/* Need to adjust the port link by 1.5x for 12bpc. */
+		pipe_config->port_clock = clock_12bpc;
 	} else {
-		DRM_DEBUG_KMS("forcing bpc to 8 for HDMI\n");
-		pipe_config->pipe_bpp = 8*3;
+		DRM_DEBUG_KMS("picking bpc to 8 for HDMI output\n");
+		desired_bpp = 8*3;
+	}
+
+	if (!pipe_config->bw_constrained) {
+		DRM_DEBUG_KMS("forcing pipe bpc to %i for HDMI\n", desired_bpp);
+		pipe_config->pipe_bpp = desired_bpp;
+	}
+
+	if (adjusted_mode->clock > 225000) {
+		DRM_DEBUG_KMS("too high HDMI clock, rejecting mode\n");
+		return false;
 	}
 
 	return true;
@@ -955,6 +1002,97 @@ done:
 	return 0;
 }
 
+static void intel_hdmi_pre_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc =
+		to_intel_crtc(encoder->base.crtc);
+	int port = vlv_dport_to_channel(dport);
+	int pipe = intel_crtc->pipe;
+	u32 val;
+
+	if (!IS_VALLEYVIEW(dev))
+		return;
+
+	/* Enable clock channels for this port */
+	val = vlv_dpio_read(dev_priv, DPIO_DATA_LANE_A(port));
+	val = 0;
+	if (pipe)
+		val |= (1<<21);
+	else
+		val &= ~(1<<21);
+	val |= 0x001000c4;
+	vlv_dpio_write(dev_priv, DPIO_DATA_CHANNEL(port), val);
+
+	/* HDMI 1.0V-2dB */
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port), 0);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL4(port),
+			 0x2b245f5f);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL2(port),
+			 0x5578b83a);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL3(port),
+			 0x0c782040);
+	vlv_dpio_write(dev_priv, DPIO_TX3_SWING_CTL4(port),
+			 0x2b247878);
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER0(port), 0x00030000);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CTL_OVER1(port),
+			 0x00002000);
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port),
+			 DPIO_TX_OCALINIT_EN);
+
+	/* Program lane clock */
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF0(port),
+			 0x00760018);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF8(port),
+			 0x00400888);
+}
+
+static void intel_hdmi_pre_pll_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int port = vlv_dport_to_channel(dport);
+
+	if (!IS_VALLEYVIEW(dev))
+		return;
+
+	/* Program Tx lane resets to default */
+	vlv_dpio_write(dev_priv, DPIO_PCS_TX(port),
+			 DPIO_PCS_TX_LANE2_RESET |
+			 DPIO_PCS_TX_LANE1_RESET);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLK(port),
+			 DPIO_PCS_CLK_CRI_RXEB_EIOS_EN |
+			 DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN |
+			 (1<<DPIO_PCS_CLK_DATAWIDTH_SHIFT) |
+			 DPIO_PCS_CLK_SOFT_RESET);
+
+	/* Fix up inter-pair skew failure */
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER1(port), 0x00750f00);
+	vlv_dpio_write(dev_priv, DPIO_TX_CTL(port), 0x00001500);
+	vlv_dpio_write(dev_priv, DPIO_TX_LANE(port), 0x40400000);
+
+	vlv_dpio_write(dev_priv, DPIO_PCS_CTL_OVER1(port),
+			 0x00002000);
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port),
+			 DPIO_TX_OCALINIT_EN);
+}
+
+static void intel_hdmi_post_disable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	int port = vlv_dport_to_channel(dport);
+
+	/* Reset lanes to avoid HDMI flicker (VLV w/a) */
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_dpio_write(dev_priv, DPIO_PCS_TX(port), 0x00000000);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLK(port), 0x00e00060);
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
 static void intel_hdmi_destroy(struct drm_connector *connector)
 {
 	drm_sysfs_connector_remove(connector);
@@ -1094,6 +1232,12 @@ void intel_hdmi_init(struct drm_device *dev, int hdmi_reg, enum port port)
 	intel_encoder->enable = intel_enable_hdmi;
 	intel_encoder->disable = intel_disable_hdmi;
 	intel_encoder->get_hw_state = intel_hdmi_get_hw_state;
+	intel_encoder->get_config = intel_hdmi_get_config;
+	if (IS_VALLEYVIEW(dev)) {
+		intel_encoder->pre_enable = intel_hdmi_pre_enable;
+		intel_encoder->pre_pll_enable = intel_hdmi_pre_pll_enable;
+		intel_encoder->post_disable = intel_hdmi_post_disable;
+	}
 
 	intel_encoder->type = INTEL_OUTPUT_HDMI;
 	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 817f936..021e8da 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -49,8 +49,6 @@ struct intel_lvds_connector {
 struct intel_lvds_encoder {
 	struct intel_encoder base;
 
-	u32 pfit_control;
-	u32 pfit_pgm_ratios;
 	bool is_dual_link;
 	u32 reg;
 
@@ -88,6 +86,31 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_lvds_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 lvds_reg, tmp, flags = 0;
+
+	if (HAS_PCH_SPLIT(dev))
+		lvds_reg = PCH_LVDS;
+	else
+		lvds_reg = LVDS;
+
+	tmp = I915_READ(lvds_reg);
+	if (tmp & LVDS_HSYNC_POLARITY)
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	if (tmp & LVDS_VSYNC_POLARITY)
+		flags |= DRM_MODE_FLAG_NVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_PVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 /* The LVDS pin pair needs to be on before the DPLLs are enabled.
  * This is an exception to the general rule that mode_set doesn't turn
  * things on.
@@ -118,7 +141,8 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	}
 
 	/* set the corresponsding LVDS_BORDER bit */
-	temp |= dev_priv->lvds_border_bits;
+	temp &= ~LVDS_BORDER_ENABLE;
+	temp |= intel_crtc->config.gmch_pfit.lvds_border_bits;
 	/* Set the B0-B3 data pairs corresponding to whether we're going to
 	 * set the DPLLs for dual-channel mode or not.
 	 */
@@ -136,7 +160,10 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	 * special lvds dither control bit on pch-split platforms, dithering is
 	 * only controlled through the PIPECONF reg. */
 	if (INTEL_INFO(dev)->gen == 4) {
-		if (dev_priv->lvds_dither)
+		/* Bspec wording suggests that LVDS port dithering only exists
+		 * for 18bpp panels. */
+		if (intel_crtc->config.dither &&
+		    intel_crtc->config.pipe_bpp == 18)
 			temp |= LVDS_ENABLE_DITHER;
 		else
 			temp &= ~LVDS_ENABLE_DITHER;
@@ -150,29 +177,6 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	I915_WRITE(lvds_encoder->reg, temp);
 }
 
-static void intel_pre_enable_lvds(struct intel_encoder *encoder)
-{
-	struct drm_device *dev = encoder->base.dev;
-	struct intel_lvds_encoder *enc = to_lvds_encoder(&encoder->base);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (HAS_PCH_SPLIT(dev) || !enc->pfit_control)
-		return;
-
-	/*
-	 * Enable automatic panel scaling so that non-native modes
-	 * fill the screen.  The panel fitter should only be
-	 * adjusted whilst the pipe is disabled, according to
-	 * register description and PRM.
-	 */
-	DRM_DEBUG_KMS("applying panel-fitter: %x, %x\n",
-		      enc->pfit_control,
-		      enc->pfit_pgm_ratios);
-
-	I915_WRITE(PFIT_PGM_RATIOS, enc->pfit_pgm_ratios);
-	I915_WRITE(PFIT_CONTROL, enc->pfit_control);
-}
-
 /**
  * Sets the power state for the panel.
  */
@@ -241,62 +245,6 @@ static int intel_lvds_mode_valid(struct drm_connector *connector,
 	return MODE_OK;
 }
 
-static void
-centre_horizontally(struct drm_display_mode *mode,
-		    int width)
-{
-	u32 border, sync_pos, blank_width, sync_width;
-
-	/* keep the hsync and hblank widths constant */
-	sync_width = mode->crtc_hsync_end - mode->crtc_hsync_start;
-	blank_width = mode->crtc_hblank_end - mode->crtc_hblank_start;
-	sync_pos = (blank_width - sync_width + 1) / 2;
-
-	border = (mode->hdisplay - width + 1) / 2;
-	border += border & 1; /* make the border even */
-
-	mode->crtc_hdisplay = width;
-	mode->crtc_hblank_start = width + border;
-	mode->crtc_hblank_end = mode->crtc_hblank_start + blank_width;
-
-	mode->crtc_hsync_start = mode->crtc_hblank_start + sync_pos;
-	mode->crtc_hsync_end = mode->crtc_hsync_start + sync_width;
-}
-
-static void
-centre_vertically(struct drm_display_mode *mode,
-		  int height)
-{
-	u32 border, sync_pos, blank_width, sync_width;
-
-	/* keep the vsync and vblank widths constant */
-	sync_width = mode->crtc_vsync_end - mode->crtc_vsync_start;
-	blank_width = mode->crtc_vblank_end - mode->crtc_vblank_start;
-	sync_pos = (blank_width - sync_width + 1) / 2;
-
-	border = (mode->vdisplay - height + 1) / 2;
-
-	mode->crtc_vdisplay = height;
-	mode->crtc_vblank_start = height + border;
-	mode->crtc_vblank_end = mode->crtc_vblank_start + blank_width;
-
-	mode->crtc_vsync_start = mode->crtc_vblank_start + sync_pos;
-	mode->crtc_vsync_end = mode->crtc_vsync_start + sync_width;
-}
-
-static inline u32 panel_fitter_scaling(u32 source, u32 target)
-{
-	/*
-	 * Floating point operation is not supported. So the FACTOR
-	 * is defined, which can avoid the floating point computation
-	 * when calculating the panel ratio.
-	 */
-#define ACCURACY 12
-#define FACTOR (1 << ACCURACY)
-	u32 ratio = source * FACTOR / target;
-	return (FACTOR * ratio + FACTOR/2) / FACTOR;
-}
-
 static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 				      struct intel_crtc_config *pipe_config)
 {
@@ -307,11 +255,8 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 	struct intel_connector *intel_connector =
 		&lvds_encoder->attached_connector->base;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
-	struct drm_display_mode *mode = &pipe_config->requested_mode;
 	struct intel_crtc *intel_crtc = lvds_encoder->base.new_crtc;
-	u32 pfit_control = 0, pfit_pgm_ratios = 0, border = 0;
 	unsigned int lvds_bpp;
-	int pipe;
 
 	/* Should never happen!! */
 	if (INTEL_INFO(dev)->gen < 4 && intel_crtc->pipe == 0) {
@@ -319,20 +264,18 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 		return false;
 	}
 
-	if (intel_encoder_check_is_cloned(&lvds_encoder->base))
-		return false;
-
 	if ((I915_READ(lvds_encoder->reg) & LVDS_A3_POWER_MASK) ==
 	    LVDS_A3_POWER_UP)
 		lvds_bpp = 8*3;
 	else
 		lvds_bpp = 6*3;
 
-	if (lvds_bpp != pipe_config->pipe_bpp) {
+	if (lvds_bpp != pipe_config->pipe_bpp && !pipe_config->bw_constrained) {
 		DRM_DEBUG_KMS("forcing display bpp (was %d) to LVDS (%d)\n",
 			      pipe_config->pipe_bpp, lvds_bpp);
 		pipe_config->pipe_bpp = lvds_bpp;
 	}
+
 	/*
 	 * We have timings from the BIOS for the panel, put them in
 	 * to the adjusted mode.  The CRTC will be set up for this mode,
@@ -345,139 +288,17 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 	if (HAS_PCH_SPLIT(dev)) {
 		pipe_config->has_pch_encoder = true;
 
-		intel_pch_panel_fitting(dev,
-					intel_connector->panel.fitting_mode,
-					mode, adjusted_mode);
+		intel_pch_panel_fitting(intel_crtc, pipe_config,
+					intel_connector->panel.fitting_mode);
 		return true;
+	} else {
+		intel_gmch_panel_fitting(intel_crtc, pipe_config,
+					 intel_connector->panel.fitting_mode);
 	}
 
-	/* Native modes don't need fitting */
-	if (adjusted_mode->hdisplay == mode->hdisplay &&
-	    adjusted_mode->vdisplay == mode->vdisplay)
-		goto out;
-
-	/* 965+ wants fuzzy fitting */
-	if (INTEL_INFO(dev)->gen >= 4)
-		pfit_control |= ((intel_crtc->pipe << PFIT_PIPE_SHIFT) |
-				 PFIT_FILTER_FUZZY);
-
-	/*
-	 * Enable automatic panel scaling for non-native modes so that they fill
-	 * the screen.  Should be enabled before the pipe is enabled, according
-	 * to register description and PRM.
-	 * Change the value here to see the borders for debugging
-	 */
-	for_each_pipe(pipe)
-		I915_WRITE(BCLRPAT(pipe), 0);
-
 	drm_mode_set_crtcinfo(adjusted_mode, 0);
 	pipe_config->timings_set = true;
 
-	switch (intel_connector->panel.fitting_mode) {
-	case DRM_MODE_SCALE_CENTER:
-		/*
-		 * For centered modes, we have to calculate border widths &
-		 * heights and modify the values programmed into the CRTC.
-		 */
-		centre_horizontally(adjusted_mode, mode->hdisplay);
-		centre_vertically(adjusted_mode, mode->vdisplay);
-		border = LVDS_BORDER_ENABLE;
-		break;
-
-	case DRM_MODE_SCALE_ASPECT:
-		/* Scale but preserve the aspect ratio */
-		if (INTEL_INFO(dev)->gen >= 4) {
-			u32 scaled_width = adjusted_mode->hdisplay * mode->vdisplay;
-			u32 scaled_height = mode->hdisplay * adjusted_mode->vdisplay;
-
-			/* 965+ is easy, it does everything in hw */
-			if (scaled_width > scaled_height)
-				pfit_control |= PFIT_ENABLE | PFIT_SCALING_PILLAR;
-			else if (scaled_width < scaled_height)
-				pfit_control |= PFIT_ENABLE | PFIT_SCALING_LETTER;
-			else if (adjusted_mode->hdisplay != mode->hdisplay)
-				pfit_control |= PFIT_ENABLE | PFIT_SCALING_AUTO;
-		} else {
-			u32 scaled_width = adjusted_mode->hdisplay * mode->vdisplay;
-			u32 scaled_height = mode->hdisplay * adjusted_mode->vdisplay;
-			/*
-			 * For earlier chips we have to calculate the scaling
-			 * ratio by hand and program it into the
-			 * PFIT_PGM_RATIO register
-			 */
-			if (scaled_width > scaled_height) { /* pillar */
-				centre_horizontally(adjusted_mode, scaled_height / mode->vdisplay);
-
-				border = LVDS_BORDER_ENABLE;
-				if (mode->vdisplay != adjusted_mode->vdisplay) {
-					u32 bits = panel_fitter_scaling(mode->vdisplay, adjusted_mode->vdisplay);
-					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
-							    bits << PFIT_VERT_SCALE_SHIFT);
-					pfit_control |= (PFIT_ENABLE |
-							 VERT_INTERP_BILINEAR |
-							 HORIZ_INTERP_BILINEAR);
-				}
-			} else if (scaled_width < scaled_height) { /* letter */
-				centre_vertically(adjusted_mode, scaled_width / mode->hdisplay);
-
-				border = LVDS_BORDER_ENABLE;
-				if (mode->hdisplay != adjusted_mode->hdisplay) {
-					u32 bits = panel_fitter_scaling(mode->hdisplay, adjusted_mode->hdisplay);
-					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
-							    bits << PFIT_VERT_SCALE_SHIFT);
-					pfit_control |= (PFIT_ENABLE |
-							 VERT_INTERP_BILINEAR |
-							 HORIZ_INTERP_BILINEAR);
-				}
-			} else
-				/* Aspects match, Let hw scale both directions */
-				pfit_control |= (PFIT_ENABLE |
-						 VERT_AUTO_SCALE | HORIZ_AUTO_SCALE |
-						 VERT_INTERP_BILINEAR |
-						 HORIZ_INTERP_BILINEAR);
-		}
-		break;
-
-	case DRM_MODE_SCALE_FULLSCREEN:
-		/*
-		 * Full scaling, even if it changes the aspect ratio.
-		 * Fortunately this is all done for us in hw.
-		 */
-		if (mode->vdisplay != adjusted_mode->vdisplay ||
-		    mode->hdisplay != adjusted_mode->hdisplay) {
-			pfit_control |= PFIT_ENABLE;
-			if (INTEL_INFO(dev)->gen >= 4)
-				pfit_control |= PFIT_SCALING_AUTO;
-			else
-				pfit_control |= (VERT_AUTO_SCALE |
-						 VERT_INTERP_BILINEAR |
-						 HORIZ_AUTO_SCALE |
-						 HORIZ_INTERP_BILINEAR);
-		}
-		break;
-
-	default:
-		break;
-	}
-
-out:
-	/* If not enabling scaling, be consistent and always use 0. */
-	if ((pfit_control & PFIT_ENABLE) == 0) {
-		pfit_control = 0;
-		pfit_pgm_ratios = 0;
-	}
-
-	/* Make sure pre-965 set dither correctly */
-	if (INTEL_INFO(dev)->gen < 4 && dev_priv->lvds_dither)
-		pfit_control |= PANEL_8TO6_DITHER_ENABLE;
-
-	if (pfit_control != lvds_encoder->pfit_control ||
-	    pfit_pgm_ratios != lvds_encoder->pfit_pgm_ratios) {
-		lvds_encoder->pfit_control = pfit_control;
-		lvds_encoder->pfit_pgm_ratios = pfit_pgm_ratios;
-	}
-	dev_priv->lvds_border_bits = border;
-
 	/*
 	 * XXX: It would be nice to support lower refresh rates on the
 	 * panels to reduce power consumption, and perhaps match the
@@ -953,11 +774,11 @@ static bool lvds_is_present_in_vbt(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
-	if (!dev_priv->child_dev_num)
+	if (!dev_priv->vbt.child_dev_num)
 		return true;
 
-	for (i = 0; i < dev_priv->child_dev_num; i++) {
-		struct child_device_config *child = dev_priv->child_dev + i;
+	for (i = 0; i < dev_priv->vbt.child_dev_num; i++) {
+		struct child_device_config *child = dev_priv->vbt.child_dev + i;
 
 		/* If the device type is not LFP, continue.
 		 * We have to check both the new identifiers as well as the
@@ -1045,7 +866,7 @@ static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder)
 	 */
 	val = I915_READ(lvds_encoder->reg);
 	if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
-		val = dev_priv->bios_lvds_val;
+		val = dev_priv->vbt.bios_lvds_val;
 
 	return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP;
 }
@@ -1072,7 +893,7 @@ static bool intel_lvds_supported(struct drm_device *dev)
  * Create the connector, register the LVDS DDC bus, and try to figure out what
  * modes we can display on the LVDS panel (if present).
  */
-bool intel_lvds_init(struct drm_device *dev)
+void intel_lvds_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds_encoder *lvds_encoder;
@@ -1090,43 +911,39 @@ bool intel_lvds_init(struct drm_device *dev)
 	u8 pin;
 
 	if (!intel_lvds_supported(dev))
-		return false;
+		return;
 
 	/* Skip init on machines we know falsely report LVDS */
 	if (dmi_check_system(intel_no_lvds))
-		return false;
+		return;
 
 	pin = GMBUS_PORT_PANEL;
 	if (!lvds_is_present_in_vbt(dev, &pin)) {
 		DRM_DEBUG_KMS("LVDS is not present in VBT\n");
-		return false;
+		return;
 	}
 
 	if (HAS_PCH_SPLIT(dev)) {
 		if ((I915_READ(PCH_LVDS) & LVDS_DETECTED) == 0)
-			return false;
-		if (dev_priv->edp.support) {
+			return;
+		if (dev_priv->vbt.edp_support) {
 			DRM_DEBUG_KMS("disable LVDS for eDP support\n");
-			return false;
+			return;
 		}
 	}
 
 	lvds_encoder = kzalloc(sizeof(struct intel_lvds_encoder), GFP_KERNEL);
 	if (!lvds_encoder)
-		return false;
+		return;
 
 	lvds_connector = kzalloc(sizeof(struct intel_lvds_connector), GFP_KERNEL);
 	if (!lvds_connector) {
 		kfree(lvds_encoder);
-		return false;
+		return;
 	}
 
 	lvds_encoder->attached_connector = lvds_connector;
 
-	if (!HAS_PCH_SPLIT(dev)) {
-		lvds_encoder->pfit_control = I915_READ(PFIT_CONTROL);
-	}
-
 	intel_encoder = &lvds_encoder->base;
 	encoder = &intel_encoder->base;
 	intel_connector = &lvds_connector->base;
@@ -1138,11 +955,11 @@ bool intel_lvds_init(struct drm_device *dev)
 			 DRM_MODE_ENCODER_LVDS);
 
 	intel_encoder->enable = intel_enable_lvds;
-	intel_encoder->pre_enable = intel_pre_enable_lvds;
 	intel_encoder->pre_pll_enable = intel_pre_pll_enable_lvds;
 	intel_encoder->compute_config = intel_lvds_compute_config;
 	intel_encoder->disable = intel_disable_lvds;
 	intel_encoder->get_hw_state = intel_lvds_get_hw_state;
+	intel_encoder->get_config = intel_lvds_get_config;
 	intel_connector->get_hw_state = intel_connector_get_hw_state;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
@@ -1228,11 +1045,11 @@ bool intel_lvds_init(struct drm_device *dev)
 	}
 
 	/* Failed to get EDID, what about VBT? */
-	if (dev_priv->lfp_lvds_vbt_mode) {
+	if (dev_priv->vbt.lfp_lvds_vbt_mode) {
 		DRM_DEBUG_KMS("using mode from VBT: ");
-		drm_mode_debug_printmodeline(dev_priv->lfp_lvds_vbt_mode);
+		drm_mode_debug_printmodeline(dev_priv->vbt.lfp_lvds_vbt_mode);
 
-		fixed_mode = drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode);
+		fixed_mode = drm_mode_duplicate(dev, dev_priv->vbt.lfp_lvds_vbt_mode);
 		if (fixed_mode) {
 			fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
 			goto out;
@@ -1293,7 +1110,7 @@ out:
 	intel_panel_init(&intel_connector->panel, fixed_mode);
 	intel_panel_setup_backlight(connector);
 
-	return true;
+	return;
 
 failed:
 	DRM_DEBUG_KMS("No LVDS modes found, disabling.\n");
@@ -1303,5 +1120,5 @@ failed:
 		drm_mode_destroy(dev, fixed_mode);
 	kfree(lvds_encoder);
 	kfree(lvds_connector);
-	return false;
+	return;
 }
diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index a8117e6..cfb8fb6 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -110,6 +110,10 @@ struct opregion_asle {
 	u8 rsvd[102];
 } __attribute__((packed));
 
+/* Driver readiness indicator */
+#define ASLE_ARDY_READY		(1 << 0)
+#define ASLE_ARDY_NOT_READY	(0 << 0)
+
 /* ASLE irq request bits */
 #define ASLE_SET_ALS_ILLUM     (1 << 0)
 #define ASLE_SET_BACKLIGHT     (1 << 1)
@@ -123,6 +127,12 @@ struct opregion_asle {
 #define ASLE_PFIT_FAILED	(1<<14)
 #define ASLE_PWM_FREQ_FAILED	(1<<16)
 
+/* Technology enabled indicator */
+#define ASLE_TCHE_ALS_EN	(1 << 0)
+#define ASLE_TCHE_BLC_EN	(1 << 1)
+#define ASLE_TCHE_PFIT_EN	(1 << 2)
+#define ASLE_TCHE_PFMB_EN	(1 << 3)
+
 /* ASLE backlight brightness to set */
 #define ASLE_BCLP_VALID                (1<<31)
 #define ASLE_BCLP_MSK          (~(1<<31))
@@ -152,7 +162,6 @@ static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct opregion_asle __iomem *asle = dev_priv->opregion.asle;
-	u32 max;
 
 	DRM_DEBUG_DRIVER("bclp = 0x%08x\n", bclp);
 
@@ -163,8 +172,7 @@ static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 	if (bclp > 255)
 		return ASLE_BACKLIGHT_FAILED;
 
-	max = intel_panel_get_max_backlight(dev);
-	intel_panel_set_backlight(dev, bclp * max / 255);
+	intel_panel_set_backlight(dev, bclp, 255);
 	iowrite32((bclp*0x64)/0xff | ASLE_CBLV_VALID, &asle->cblv);
 
 	return 0;
@@ -174,29 +182,22 @@ static u32 asle_set_als_illum(struct drm_device *dev, u32 alsi)
 {
 	/* alsi is the current ALS reading in lux. 0 indicates below sensor
 	   range, 0xffff indicates above sensor range. 1-0xfffe are valid */
-	return 0;
+	DRM_DEBUG_DRIVER("Illum is not supported\n");
+	return ASLE_ALS_ILLUM_FAILED;
 }
 
 static u32 asle_set_pwm_freq(struct drm_device *dev, u32 pfmb)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	if (pfmb & ASLE_PFMB_PWM_VALID) {
-		u32 blc_pwm_ctl = I915_READ(BLC_PWM_CTL);
-		u32 pwm = pfmb & ASLE_PFMB_PWM_MASK;
-		blc_pwm_ctl &= BACKLIGHT_DUTY_CYCLE_MASK;
-		pwm = pwm >> 9;
-		/* FIXME - what do we do with the PWM? */
-	}
-	return 0;
+	DRM_DEBUG_DRIVER("PWM freq is not supported\n");
+	return ASLE_PWM_FREQ_FAILED;
 }
 
 static u32 asle_set_pfit(struct drm_device *dev, u32 pfit)
 {
 	/* Panel fitting is currently controlled by the X code, so this is a
 	   noop until modesetting support works fully */
-	if (!(pfit & ASLE_PFIT_VALID))
-		return ASLE_PFIT_FAILED;
-	return 0;
+	DRM_DEBUG_DRIVER("Pfit is not supported\n");
+	return ASLE_PFIT_FAILED;
 }
 
 void intel_opregion_asle_intr(struct drm_device *dev)
@@ -231,64 +232,6 @@ void intel_opregion_asle_intr(struct drm_device *dev)
 	iowrite32(asle_stat, &asle->aslc);
 }
 
-void intel_opregion_gse_intr(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct opregion_asle __iomem *asle = dev_priv->opregion.asle;
-	u32 asle_stat = 0;
-	u32 asle_req;
-
-	if (!asle)
-		return;
-
-	asle_req = ioread32(&asle->aslc) & ASLE_REQ_MSK;
-
-	if (!asle_req) {
-		DRM_DEBUG_DRIVER("non asle set request??\n");
-		return;
-	}
-
-	if (asle_req & ASLE_SET_ALS_ILLUM) {
-		DRM_DEBUG_DRIVER("Illum is not supported\n");
-		asle_stat |= ASLE_ALS_ILLUM_FAILED;
-	}
-
-	if (asle_req & ASLE_SET_BACKLIGHT)
-		asle_stat |= asle_set_backlight(dev, ioread32(&asle->bclp));
-
-	if (asle_req & ASLE_SET_PFIT) {
-		DRM_DEBUG_DRIVER("Pfit is not supported\n");
-		asle_stat |= ASLE_PFIT_FAILED;
-	}
-
-	if (asle_req & ASLE_SET_PWM_FREQ) {
-		DRM_DEBUG_DRIVER("PWM freq is not supported\n");
-		asle_stat |= ASLE_PWM_FREQ_FAILED;
-	}
-
-	iowrite32(asle_stat, &asle->aslc);
-}
-#define ASLE_ALS_EN    (1<<0)
-#define ASLE_BLC_EN    (1<<1)
-#define ASLE_PFIT_EN   (1<<2)
-#define ASLE_PFMB_EN   (1<<3)
-
-void intel_opregion_enable_asle(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct opregion_asle __iomem *asle = dev_priv->opregion.asle;
-
-	if (asle) {
-		if (IS_MOBILE(dev))
-			intel_enable_asle(dev);
-
-		iowrite32(ASLE_ALS_EN | ASLE_BLC_EN | ASLE_PFIT_EN |
-			  ASLE_PFMB_EN,
-			  &asle->tche);
-		iowrite32(1, &asle->ardy);
-	}
-}
-
 #define ACPI_EV_DISPLAY_SWITCH (1<<0)
 #define ACPI_EV_LID            (1<<1)
 #define ACPI_EV_DOCK           (1<<2)
@@ -368,8 +311,8 @@ static void intel_didl_outputs(struct drm_device *dev)
 
 	list_for_each_entry(acpi_cdev, &acpi_video_bus->children, node) {
 		if (i >= 8) {
-			dev_printk(KERN_ERR, &dev->pdev->dev,
-				    "More than 8 outputs detected\n");
+			dev_dbg(&dev->pdev->dev,
+				"More than 8 outputs detected via ACPI\n");
 			return;
 		}
 		status =
@@ -395,8 +338,8 @@ blind_set:
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		int output_type = ACPI_OTHER_OUTPUT;
 		if (i >= 8) {
-			dev_printk(KERN_ERR, &dev->pdev->dev,
-				    "More than 8 outputs detected\n");
+			dev_dbg(&dev->pdev->dev,
+				"More than 8 outputs in connector list\n");
 			return;
 		}
 		switch (connector->connector_type) {
@@ -472,8 +415,10 @@ void intel_opregion_init(struct drm_device *dev)
 		register_acpi_notifier(&intel_opregion_notifier);
 	}
 
-	if (opregion->asle)
-		intel_opregion_enable_asle(dev);
+	if (opregion->asle) {
+		iowrite32(ASLE_TCHE_BLC_EN, &opregion->asle->tche);
+		iowrite32(ASLE_ARDY_READY, &opregion->asle->ardy);
+	}
 }
 
 void intel_opregion_fini(struct drm_device *dev)
@@ -484,6 +429,9 @@ void intel_opregion_fini(struct drm_device *dev)
 	if (!opregion->header)
 		return;
 
+	if (opregion->asle)
+		iowrite32(ASLE_ARDY_NOT_READY, &opregion->asle->ardy);
+
 	if (opregion->acpi) {
 		iowrite32(0, &opregion->acpi->drdy);
 
@@ -546,6 +494,8 @@ int intel_opregion_setup(struct drm_device *dev)
 	if (mboxes & MBOX_ASLE) {
 		DRM_DEBUG_DRIVER("ASLE supported\n");
 		opregion->asle = base + OPREGION_ASLE_OFFSET;
+
+		iowrite32(ASLE_ARDY_NOT_READY, &opregion->asle->ardy);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index 67a2501..a369881 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -217,7 +217,7 @@ static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
 	int ret;
 
 	BUG_ON(overlay->last_flip_req);
-	ret = i915_add_request(ring, NULL, &overlay->last_flip_req);
+	ret = i915_add_request(ring, &overlay->last_flip_req);
 	if (ret)
 		return ret;
 
@@ -286,7 +286,7 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 	intel_ring_emit(ring, flip_addr);
 	intel_ring_advance(ring);
 
-	return i915_add_request(ring, NULL, &overlay->last_flip_req);
+	return i915_add_request(ring, &overlay->last_flip_req);
 }
 
 static void intel_overlay_release_old_vid_tail(struct intel_overlay *overlay)
@@ -1485,14 +1485,15 @@ err:
 }
 
 void
-intel_overlay_print_error_state(struct seq_file *m, struct intel_overlay_error_state *error)
+intel_overlay_print_error_state(struct drm_i915_error_state_buf *m,
+				struct intel_overlay_error_state *error)
 {
-	seq_printf(m, "Overlay, status: 0x%08x, interrupt: 0x%08x\n",
-		   error->dovsta, error->isr);
-	seq_printf(m, "  Register file at 0x%08lx:\n",
-		   error->base);
+	i915_error_printf(m, "Overlay, status: 0x%08x, interrupt: 0x%08x\n",
+			  error->dovsta, error->isr);
+	i915_error_printf(m, "  Register file at 0x%08lx:\n",
+			  error->base);
 
-#define P(x) seq_printf(m, "    " #x ":	0x%08x\n", error->regs.x)
+#define P(x) i915_error_printf(m, "    " #x ":	0x%08x\n", error->regs.x)
 	P(OBUF_0Y);
 	P(OBUF_1Y);
 	P(OBUF_0U);
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index eb5e6e9..80bea1d 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -54,14 +54,16 @@ intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
 
 /* adjusted_mode has been preset to be the panel's fixed mode */
 void
-intel_pch_panel_fitting(struct drm_device *dev,
-			int fitting_mode,
-			const struct drm_display_mode *mode,
-			struct drm_display_mode *adjusted_mode)
+intel_pch_panel_fitting(struct intel_crtc *intel_crtc,
+			struct intel_crtc_config *pipe_config,
+			int fitting_mode)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_display_mode *mode, *adjusted_mode;
 	int x, y, width, height;
 
+	mode = &pipe_config->requested_mode;
+	adjusted_mode = &pipe_config->adjusted_mode;
+
 	x = y = width = height = 0;
 
 	/* Native modes don't need fitting */
@@ -104,17 +106,209 @@ intel_pch_panel_fitting(struct drm_device *dev,
 		}
 		break;
 
-	default:
 	case DRM_MODE_SCALE_FULLSCREEN:
 		x = y = 0;
 		width = adjusted_mode->hdisplay;
 		height = adjusted_mode->vdisplay;
 		break;
+
+	default:
+		WARN(1, "bad panel fit mode: %d\n", fitting_mode);
+		return;
 	}
 
 done:
-	dev_priv->pch_pf_pos = (x << 16) | y;
-	dev_priv->pch_pf_size = (width << 16) | height;
+	pipe_config->pch_pfit.pos = (x << 16) | y;
+	pipe_config->pch_pfit.size = (width << 16) | height;
+}
+
+static void
+centre_horizontally(struct drm_display_mode *mode,
+		    int width)
+{
+	u32 border, sync_pos, blank_width, sync_width;
+
+	/* keep the hsync and hblank widths constant */
+	sync_width = mode->crtc_hsync_end - mode->crtc_hsync_start;
+	blank_width = mode->crtc_hblank_end - mode->crtc_hblank_start;
+	sync_pos = (blank_width - sync_width + 1) / 2;
+
+	border = (mode->hdisplay - width + 1) / 2;
+	border += border & 1; /* make the border even */
+
+	mode->crtc_hdisplay = width;
+	mode->crtc_hblank_start = width + border;
+	mode->crtc_hblank_end = mode->crtc_hblank_start + blank_width;
+
+	mode->crtc_hsync_start = mode->crtc_hblank_start + sync_pos;
+	mode->crtc_hsync_end = mode->crtc_hsync_start + sync_width;
+}
+
+static void
+centre_vertically(struct drm_display_mode *mode,
+		  int height)
+{
+	u32 border, sync_pos, blank_width, sync_width;
+
+	/* keep the vsync and vblank widths constant */
+	sync_width = mode->crtc_vsync_end - mode->crtc_vsync_start;
+	blank_width = mode->crtc_vblank_end - mode->crtc_vblank_start;
+	sync_pos = (blank_width - sync_width + 1) / 2;
+
+	border = (mode->vdisplay - height + 1) / 2;
+
+	mode->crtc_vdisplay = height;
+	mode->crtc_vblank_start = height + border;
+	mode->crtc_vblank_end = mode->crtc_vblank_start + blank_width;
+
+	mode->crtc_vsync_start = mode->crtc_vblank_start + sync_pos;
+	mode->crtc_vsync_end = mode->crtc_vsync_start + sync_width;
+}
+
+static inline u32 panel_fitter_scaling(u32 source, u32 target)
+{
+	/*
+	 * Floating point operation is not supported. So the FACTOR
+	 * is defined, which can avoid the floating point computation
+	 * when calculating the panel ratio.
+	 */
+#define ACCURACY 12
+#define FACTOR (1 << ACCURACY)
+	u32 ratio = source * FACTOR / target;
+	return (FACTOR * ratio + FACTOR/2) / FACTOR;
+}
+
+void intel_gmch_panel_fitting(struct intel_crtc *intel_crtc,
+			      struct intel_crtc_config *pipe_config,
+			      int fitting_mode)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	u32 pfit_control = 0, pfit_pgm_ratios = 0, border = 0;
+	struct drm_display_mode *mode, *adjusted_mode;
+
+	mode = &pipe_config->requested_mode;
+	adjusted_mode = &pipe_config->adjusted_mode;
+
+	/* Native modes don't need fitting */
+	if (adjusted_mode->hdisplay == mode->hdisplay &&
+	    adjusted_mode->vdisplay == mode->vdisplay)
+		goto out;
+
+	switch (fitting_mode) {
+	case DRM_MODE_SCALE_CENTER:
+		/*
+		 * For centered modes, we have to calculate border widths &
+		 * heights and modify the values programmed into the CRTC.
+		 */
+		centre_horizontally(adjusted_mode, mode->hdisplay);
+		centre_vertically(adjusted_mode, mode->vdisplay);
+		border = LVDS_BORDER_ENABLE;
+		break;
+	case DRM_MODE_SCALE_ASPECT:
+		/* Scale but preserve the aspect ratio */
+		if (INTEL_INFO(dev)->gen >= 4) {
+			u32 scaled_width = adjusted_mode->hdisplay *
+				mode->vdisplay;
+			u32 scaled_height = mode->hdisplay *
+				adjusted_mode->vdisplay;
+
+			/* 965+ is easy, it does everything in hw */
+			if (scaled_width > scaled_height)
+				pfit_control |= PFIT_ENABLE |
+					PFIT_SCALING_PILLAR;
+			else if (scaled_width < scaled_height)
+				pfit_control |= PFIT_ENABLE |
+					PFIT_SCALING_LETTER;
+			else if (adjusted_mode->hdisplay != mode->hdisplay)
+				pfit_control |= PFIT_ENABLE | PFIT_SCALING_AUTO;
+		} else {
+			u32 scaled_width = adjusted_mode->hdisplay *
+				mode->vdisplay;
+			u32 scaled_height = mode->hdisplay *
+				adjusted_mode->vdisplay;
+			/*
+			 * For earlier chips we have to calculate the scaling
+			 * ratio by hand and program it into the
+			 * PFIT_PGM_RATIO register
+			 */
+			if (scaled_width > scaled_height) { /* pillar */
+				centre_horizontally(adjusted_mode,
+						    scaled_height /
+						    mode->vdisplay);
+
+				border = LVDS_BORDER_ENABLE;
+				if (mode->vdisplay != adjusted_mode->vdisplay) {
+					u32 bits = panel_fitter_scaling(mode->vdisplay, adjusted_mode->vdisplay);
+					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
+							    bits << PFIT_VERT_SCALE_SHIFT);
+					pfit_control |= (PFIT_ENABLE |
+							 VERT_INTERP_BILINEAR |
+							 HORIZ_INTERP_BILINEAR);
+				}
+			} else if (scaled_width < scaled_height) { /* letter */
+				centre_vertically(adjusted_mode,
+						  scaled_width /
+						  mode->hdisplay);
+
+				border = LVDS_BORDER_ENABLE;
+				if (mode->hdisplay != adjusted_mode->hdisplay) {
+					u32 bits = panel_fitter_scaling(mode->hdisplay, adjusted_mode->hdisplay);
+					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
+							    bits << PFIT_VERT_SCALE_SHIFT);
+					pfit_control |= (PFIT_ENABLE |
+							 VERT_INTERP_BILINEAR |
+							 HORIZ_INTERP_BILINEAR);
+				}
+			} else {
+				/* Aspects match, Let hw scale both directions */
+				pfit_control |= (PFIT_ENABLE |
+						 VERT_AUTO_SCALE | HORIZ_AUTO_SCALE |
+						 VERT_INTERP_BILINEAR |
+						 HORIZ_INTERP_BILINEAR);
+			}
+		}
+		break;
+	case DRM_MODE_SCALE_FULLSCREEN:
+		/*
+		 * Full scaling, even if it changes the aspect ratio.
+		 * Fortunately this is all done for us in hw.
+		 */
+		if (mode->vdisplay != adjusted_mode->vdisplay ||
+		    mode->hdisplay != adjusted_mode->hdisplay) {
+			pfit_control |= PFIT_ENABLE;
+			if (INTEL_INFO(dev)->gen >= 4)
+				pfit_control |= PFIT_SCALING_AUTO;
+			else
+				pfit_control |= (VERT_AUTO_SCALE |
+						 VERT_INTERP_BILINEAR |
+						 HORIZ_AUTO_SCALE |
+						 HORIZ_INTERP_BILINEAR);
+		}
+		break;
+	default:
+		WARN(1, "bad panel fit mode: %d\n", fitting_mode);
+		return;
+	}
+
+	/* 965+ wants fuzzy fitting */
+	/* FIXME: handle multiple panels by failing gracefully */
+	if (INTEL_INFO(dev)->gen >= 4)
+		pfit_control |= ((intel_crtc->pipe << PFIT_PIPE_SHIFT) |
+				 PFIT_FILTER_FUZZY);
+
+out:
+	if ((pfit_control & PFIT_ENABLE) == 0) {
+		pfit_control = 0;
+		pfit_pgm_ratios = 0;
+	}
+
+	/* Make sure pre-965 set dither correctly for 18bpp panels. */
+	if (INTEL_INFO(dev)->gen < 4 && pipe_config->pipe_bpp == 18)
+		pfit_control |= PANEL_8TO6_DITHER_ENABLE;
+
+	pipe_config->gmch_pfit.control = pfit_control;
+	pipe_config->gmch_pfit.pgm_ratios = pfit_pgm_ratios;
+	pipe_config->gmch_pfit.lvds_border_bits = border;
 }
 
 static int is_backlight_combination_mode(struct drm_device *dev)
@@ -130,11 +324,16 @@ static int is_backlight_combination_mode(struct drm_device *dev)
 	return 0;
 }
 
+/* XXX: query mode clock or hardware clock and program max PWM appropriately
+ * when it's 0.
+ */
 static u32 i915_read_blc_pwm_ctl(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val;
 
+	WARN_ON_SMP(!spin_is_locked(&dev_priv->backlight.lock));
+
 	/* Restore the CTL value if it lost, e.g. GPU reset */
 
 	if (HAS_PCH_SPLIT(dev_priv->dev)) {
@@ -164,7 +363,7 @@ static u32 i915_read_blc_pwm_ctl(struct drm_device *dev)
 	return val;
 }
 
-static u32 _intel_panel_get_max_backlight(struct drm_device *dev)
+static u32 intel_panel_get_max_backlight(struct drm_device *dev)
 {
 	u32 max;
 
@@ -182,23 +381,8 @@ static u32 _intel_panel_get_max_backlight(struct drm_device *dev)
 			max *= 0xff;
 	}
 
-	return max;
-}
-
-u32 intel_panel_get_max_backlight(struct drm_device *dev)
-{
-	u32 max;
-
-	max = _intel_panel_get_max_backlight(dev);
-	if (max == 0) {
-		/* XXX add code here to query mode clock or hardware clock
-		 * and program max PWM appropriately.
-		 */
-		pr_warn_once("fixme: max PWM is zero\n");
-		return 1;
-	}
-
 	DRM_DEBUG_DRIVER("max backlight PWM = %d\n", max);
+
 	return max;
 }
 
@@ -217,8 +401,11 @@ static u32 intel_panel_compute_brightness(struct drm_device *dev, u32 val)
 		return val;
 
 	if (i915_panel_invert_brightness > 0 ||
-	    dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS)
-		return intel_panel_get_max_backlight(dev) - val;
+	    dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS) {
+		u32 max = intel_panel_get_max_backlight(dev);
+		if (max)
+			return max - val;
+	}
 
 	return val;
 }
@@ -227,6 +414,9 @@ static u32 intel_panel_get_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	if (HAS_PCH_SPLIT(dev)) {
 		val = I915_READ(BLC_PWM_CPU_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
@@ -244,6 +434,9 @@ static u32 intel_panel_get_backlight(struct drm_device *dev)
 	}
 
 	val = intel_panel_compute_brightness(dev, val);
+
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	DRM_DEBUG_DRIVER("get backlight PWM = %d\n", val);
 	return val;
 }
@@ -270,6 +463,10 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, u32 level
 		u32 max = intel_panel_get_max_backlight(dev);
 		u8 lbpc;
 
+		/* we're screwed, but keep behaviour backwards compatible */
+		if (!max)
+			max = 1;
+
 		lbpc = level * 0xfe / max + 1;
 		level /= lbpc;
 		pci_write_config_byte(dev->pdev, PCI_LBPC, lbpc);
@@ -282,9 +479,23 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, u32 level
 	I915_WRITE(BLC_PWM_CTL, tmp | level);
 }
 
-void intel_panel_set_backlight(struct drm_device *dev, u32 level)
+/* set backlight brightness to level in range [0..max] */
+void intel_panel_set_backlight(struct drm_device *dev, u32 level, u32 max)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 freq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+
+	freq = intel_panel_get_max_backlight(dev);
+	if (!freq) {
+		/* we are screwed, bail out */
+		goto out;
+	}
+
+	/* scale to hardware */
+	level = level * freq / max;
 
 	dev_priv->backlight.level = level;
 	if (dev_priv->backlight.device)
@@ -292,11 +503,16 @@ void intel_panel_set_backlight(struct drm_device *dev, u32 level)
 
 	if (dev_priv->backlight.enabled)
 		intel_panel_actually_set_backlight(dev, level);
+out:
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
 }
 
 void intel_panel_disable_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	dev_priv->backlight.enabled = false;
 	intel_panel_actually_set_backlight(dev, 0);
@@ -314,12 +530,19 @@ void intel_panel_disable_backlight(struct drm_device *dev)
 			I915_WRITE(BLC_PWM_PCH_CTL1, tmp);
 		}
 	}
+
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
 }
 
 void intel_panel_enable_backlight(struct drm_device *dev,
 				  enum pipe pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder =
+		intel_pipe_to_cpu_transcoder(dev_priv, pipe);
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	if (dev_priv->backlight.level == 0) {
 		dev_priv->backlight.level = intel_panel_get_max_backlight(dev);
@@ -347,7 +570,10 @@ void intel_panel_enable_backlight(struct drm_device *dev,
 		else
 			tmp &= ~BLM_PIPE_SELECT;
 
-		tmp |= BLM_PIPE(pipe);
+		if (cpu_transcoder == TRANSCODER_EDP)
+			tmp |= BLM_TRANSCODER_EDP;
+		else
+			tmp |= BLM_PIPE(cpu_transcoder);
 		tmp &= ~BLM_PWM_ENABLE;
 
 		I915_WRITE(reg, tmp);
@@ -369,6 +595,8 @@ set_level:
 	 */
 	dev_priv->backlight.enabled = true;
 	intel_panel_actually_set_backlight(dev, dev_priv->backlight.level);
+
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
 }
 
 static void intel_panel_init_backlight(struct drm_device *dev)
@@ -405,7 +633,8 @@ intel_panel_detect(struct drm_device *dev)
 static int intel_panel_update_status(struct backlight_device *bd)
 {
 	struct drm_device *dev = bl_get_data(bd);
-	intel_panel_set_backlight(dev, bd->props.brightness);
+	intel_panel_set_backlight(dev, bd->props.brightness,
+				  bd->props.max_brightness);
 	return 0;
 }
 
@@ -425,6 +654,7 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct backlight_properties props;
+	unsigned long flags;
 
 	intel_panel_init_backlight(dev);
 
@@ -434,7 +664,11 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 	memset(&props, 0, sizeof(props));
 	props.type = BACKLIGHT_RAW;
 	props.brightness = dev_priv->backlight.level;
-	props.max_brightness = _intel_panel_get_max_backlight(dev);
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+	props.max_brightness = intel_panel_get_max_backlight(dev);
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	if (props.max_brightness == 0) {
 		DRM_DEBUG_DRIVER("Failed to get maximum backlight value\n");
 		return -ENODEV;
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index aa01128..ccbdd83 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -113,8 +113,8 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	fbc_ctl |= obj->fence_reg;
 	I915_WRITE(FBC_CONTROL, fbc_ctl);
 
-	DRM_DEBUG_KMS("enabled FBC, pitch %d, yoff %d, plane %d, ",
-		      cfb_pitch, crtc->y, intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled FBC, pitch %d, yoff %d, plane %c, ",
+		      cfb_pitch, crtc->y, plane_name(intel_crtc->plane));
 }
 
 static bool i8xx_fbc_enabled(struct drm_device *dev)
@@ -148,7 +148,7 @@ static void g4x_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	/* enable it... */
 	I915_WRITE(DPFC_CONTROL, I915_READ(DPFC_CONTROL) | DPFC_CTL_EN);
 
-	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
 }
 
 static void g4x_disable_fbc(struct drm_device *dev)
@@ -228,7 +228,7 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 		sandybridge_blit_fbc_update(dev);
 	}
 
-	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
 }
 
 static void ironlake_disable_fbc(struct drm_device *dev)
@@ -242,6 +242,18 @@ static void ironlake_disable_fbc(struct drm_device *dev)
 		dpfc_ctl &= ~DPFC_CTL_EN;
 		I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl);
 
+		if (IS_IVYBRIDGE(dev))
+			/* WaFbcDisableDpfcClockGating:ivb */
+			I915_WRITE(ILK_DSPCLK_GATE_D,
+				   I915_READ(ILK_DSPCLK_GATE_D) &
+				   ~ILK_DPFCUNIT_CLOCK_GATE_DISABLE);
+
+		if (IS_HASWELL(dev))
+			/* WaFbcDisableDpfcClockGating:hsw */
+			I915_WRITE(HSW_CLKGATE_DISABLE_PART_1,
+				   I915_READ(HSW_CLKGATE_DISABLE_PART_1) &
+				   ~HSW_DPFC_GATING_DISABLE);
+
 		DRM_DEBUG_KMS("disabled FBC\n");
 	}
 }
@@ -253,6 +265,47 @@ static bool ironlake_fbc_enabled(struct drm_device *dev)
 	return I915_READ(ILK_DPFC_CONTROL) & DPFC_CTL_EN;
 }
 
+static void gen7_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_framebuffer *fb = crtc->fb;
+	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	I915_WRITE(IVB_FBC_RT_BASE, obj->gtt_offset);
+
+	I915_WRITE(ILK_DPFC_CONTROL, DPFC_CTL_EN | DPFC_CTL_LIMIT_1X |
+		   IVB_DPFC_CTL_FENCE_EN |
+		   intel_crtc->plane << IVB_DPFC_CTL_PLANE_SHIFT);
+
+	if (IS_IVYBRIDGE(dev)) {
+		/* WaFbcAsynchFlipDisableFbcQueue:ivb */
+		I915_WRITE(ILK_DISPLAY_CHICKEN1, ILK_FBCQ_DIS);
+		/* WaFbcDisableDpfcClockGating:ivb */
+		I915_WRITE(ILK_DSPCLK_GATE_D,
+			   I915_READ(ILK_DSPCLK_GATE_D) |
+			   ILK_DPFCUNIT_CLOCK_GATE_DISABLE);
+	} else {
+		/* WaFbcAsynchFlipDisableFbcQueue:hsw */
+		I915_WRITE(HSW_PIPE_SLICE_CHICKEN_1(intel_crtc->pipe),
+			   HSW_BYPASS_FBC_QUEUE);
+		/* WaFbcDisableDpfcClockGating:hsw */
+		I915_WRITE(HSW_CLKGATE_DISABLE_PART_1,
+			   I915_READ(HSW_CLKGATE_DISABLE_PART_1) |
+			   HSW_DPFC_GATING_DISABLE);
+	}
+
+	I915_WRITE(SNB_DPFC_CTL_SA,
+		   SNB_CPU_FENCE_ENABLE | obj->fence_reg);
+	I915_WRITE(DPFC_CPU_FENCE_OFFSET, crtc->y);
+
+	sandybridge_blit_fbc_update(dev);
+
+	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
+}
+
 bool intel_fbc_enabled(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -378,7 +431,7 @@ void intel_disable_fbc(struct drm_device *dev)
  *   - no pixel mulitply/line duplication
  *   - no alpha buffer discard
  *   - no dual wide
- *   - framebuffer <= 2048 in width, 1536 in height
+ *   - framebuffer <= max_hdisplay in width, max_vdisplay in height
  *
  * We can't assume that any compression will take place (worst case),
  * so the compressed buffer has to be the same size as the uncompressed
@@ -396,6 +449,7 @@ void intel_update_fbc(struct drm_device *dev)
 	struct intel_framebuffer *intel_fb;
 	struct drm_i915_gem_object *obj;
 	int enable_fbc;
+	unsigned int max_hdisplay, max_vdisplay;
 
 	if (!i915_powersave)
 		return;
@@ -439,7 +493,7 @@ void intel_update_fbc(struct drm_device *dev)
 	if (enable_fbc < 0) {
 		DRM_DEBUG_KMS("fbc set to per-chip default\n");
 		enable_fbc = 1;
-		if (INTEL_INFO(dev)->gen <= 6)
+		if (INTEL_INFO(dev)->gen <= 7 && !IS_HASWELL(dev))
 			enable_fbc = 0;
 	}
 	if (!enable_fbc) {
@@ -454,13 +508,22 @@ void intel_update_fbc(struct drm_device *dev)
 		dev_priv->no_fbc_reason = FBC_UNSUPPORTED_MODE;
 		goto out_disable;
 	}
-	if ((crtc->mode.hdisplay > 2048) ||
-	    (crtc->mode.vdisplay > 1536)) {
+
+	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
+		max_hdisplay = 4096;
+		max_vdisplay = 2048;
+	} else {
+		max_hdisplay = 2048;
+		max_vdisplay = 1536;
+	}
+	if ((crtc->mode.hdisplay > max_hdisplay) ||
+	    (crtc->mode.vdisplay > max_vdisplay)) {
 		DRM_DEBUG_KMS("mode too large for compression, disabling\n");
 		dev_priv->no_fbc_reason = FBC_MODE_TOO_LARGE;
 		goto out_disable;
 	}
-	if ((IS_I915GM(dev) || IS_I945GM(dev)) && intel_crtc->plane != 0) {
+	if ((IS_I915GM(dev) || IS_I945GM(dev) || IS_HASWELL(dev)) &&
+	    intel_crtc->plane != 0) {
 		DRM_DEBUG_KMS("plane not 0, disabling compression\n");
 		dev_priv->no_fbc_reason = FBC_BAD_PLANE;
 		goto out_disable;
@@ -481,8 +544,6 @@ void intel_update_fbc(struct drm_device *dev)
 		goto out_disable;
 
 	if (i915_gem_stolen_setup_compression(dev, intel_fb->obj->base.size)) {
-		DRM_INFO("not enough stolen space for compressed buffer (need %zd bytes), disabling\n", intel_fb->obj->base.size);
-		DRM_INFO("hint: you may be able to increase stolen memory size in the BIOS to avoid this\n");
 		DRM_DEBUG_KMS("framebuffer too large, disabling compression\n");
 		dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL;
 		goto out_disable;
@@ -1633,6 +1694,10 @@ static bool ironlake_check_srwm(struct drm_device *dev, int level,
 		I915_WRITE(DISP_ARB_CTL,
 			   I915_READ(DISP_ARB_CTL) | DISP_FBC_WM_DIS);
 		return false;
+	} else if (INTEL_INFO(dev)->gen >= 6) {
+		/* enable FBC WM (except on ILK, where it must remain off) */
+		I915_WRITE(DISP_ARB_CTL,
+			   I915_READ(DISP_ARB_CTL) & ~DISP_FBC_WM_DIS);
 	}
 
 	if (display_wm > display->max_wm) {
@@ -2016,31 +2081,558 @@ static void ivybridge_update_wm(struct drm_device *dev)
 		   cursor_wm);
 }
 
-static void
-haswell_update_linetime_wm(struct drm_device *dev, int pipe,
-				 struct drm_display_mode *mode)
+static uint32_t hsw_wm_get_pixel_rate(struct drm_device *dev,
+				      struct drm_crtc *crtc)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	uint32_t pixel_rate, pfit_size;
+
+	pixel_rate = intel_crtc->config.adjusted_mode.clock;
+
+	/* We only use IF-ID interlacing. If we ever use PF-ID we'll need to
+	 * adjust the pixel_rate here. */
+
+	pfit_size = intel_crtc->config.pch_pfit.size;
+	if (pfit_size) {
+		uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
+
+		pipe_w = intel_crtc->config.requested_mode.hdisplay;
+		pipe_h = intel_crtc->config.requested_mode.vdisplay;
+		pfit_w = (pfit_size >> 16) & 0xFFFF;
+		pfit_h = pfit_size & 0xFFFF;
+		if (pipe_w < pfit_w)
+			pipe_w = pfit_w;
+		if (pipe_h < pfit_h)
+			pipe_h = pfit_h;
+
+		pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h,
+				     pfit_w * pfit_h);
+	}
+
+	return pixel_rate;
+}
+
+static uint32_t hsw_wm_method1(uint32_t pixel_rate, uint8_t bytes_per_pixel,
+			       uint32_t latency)
+{
+	uint64_t ret;
+
+	ret = (uint64_t) pixel_rate * bytes_per_pixel * latency;
+	ret = DIV_ROUND_UP_ULL(ret, 64 * 10000) + 2;
+
+	return ret;
+}
+
+static uint32_t hsw_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal,
+			       uint32_t horiz_pixels, uint8_t bytes_per_pixel,
+			       uint32_t latency)
+{
+	uint32_t ret;
+
+	ret = (latency * pixel_rate) / (pipe_htotal * 10000);
+	ret = (ret + 1) * horiz_pixels * bytes_per_pixel;
+	ret = DIV_ROUND_UP(ret, 64) + 2;
+	return ret;
+}
+
+static uint32_t hsw_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels,
+			   uint8_t bytes_per_pixel)
+{
+	return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2;
+}
+
+struct hsw_pipe_wm_parameters {
+	bool active;
+	bool sprite_enabled;
+	uint8_t pri_bytes_per_pixel;
+	uint8_t spr_bytes_per_pixel;
+	uint8_t cur_bytes_per_pixel;
+	uint32_t pri_horiz_pixels;
+	uint32_t spr_horiz_pixels;
+	uint32_t cur_horiz_pixels;
+	uint32_t pipe_htotal;
+	uint32_t pixel_rate;
+};
+
+struct hsw_wm_maximums {
+	uint16_t pri;
+	uint16_t spr;
+	uint16_t cur;
+	uint16_t fbc;
+};
+
+struct hsw_lp_wm_result {
+	bool enable;
+	bool fbc_enable;
+	uint32_t pri_val;
+	uint32_t spr_val;
+	uint32_t cur_val;
+	uint32_t fbc_val;
+};
+
+struct hsw_wm_values {
+	uint32_t wm_pipe[3];
+	uint32_t wm_lp[3];
+	uint32_t wm_lp_spr[3];
+	uint32_t wm_linetime[3];
+	bool enable_fbc_wm;
+};
+
+enum hsw_data_buf_partitioning {
+	HSW_DATA_BUF_PART_1_2,
+	HSW_DATA_BUF_PART_5_6,
+};
+
+/* For both WM_PIPE and WM_LP. */
+static uint32_t hsw_compute_pri_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t mem_value,
+				   bool is_lp)
+{
+	uint32_t method1, method2;
+
+	/* TODO: for now, assume the primary plane is always enabled. */
+	if (!params->active)
+		return 0;
+
+	method1 = hsw_wm_method1(params->pixel_rate,
+				 params->pri_bytes_per_pixel,
+				 mem_value);
+
+	if (!is_lp)
+		return method1;
+
+	method2 = hsw_wm_method2(params->pixel_rate,
+				 params->pipe_htotal,
+				 params->pri_horiz_pixels,
+				 params->pri_bytes_per_pixel,
+				 mem_value);
+
+	return min(method1, method2);
+}
+
+/* For both WM_PIPE and WM_LP. */
+static uint32_t hsw_compute_spr_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t mem_value)
+{
+	uint32_t method1, method2;
+
+	if (!params->active || !params->sprite_enabled)
+		return 0;
+
+	method1 = hsw_wm_method1(params->pixel_rate,
+				 params->spr_bytes_per_pixel,
+				 mem_value);
+	method2 = hsw_wm_method2(params->pixel_rate,
+				 params->pipe_htotal,
+				 params->spr_horiz_pixels,
+				 params->spr_bytes_per_pixel,
+				 mem_value);
+	return min(method1, method2);
+}
+
+/* For both WM_PIPE and WM_LP. */
+static uint32_t hsw_compute_cur_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t mem_value)
+{
+	if (!params->active)
+		return 0;
+
+	return hsw_wm_method2(params->pixel_rate,
+			      params->pipe_htotal,
+			      params->cur_horiz_pixels,
+			      params->cur_bytes_per_pixel,
+			      mem_value);
+}
+
+/* Only for WM_LP. */
+static uint32_t hsw_compute_fbc_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t pri_val,
+				   uint32_t mem_value)
+{
+	if (!params->active)
+		return 0;
+
+	return hsw_wm_fbc(pri_val,
+			  params->pri_horiz_pixels,
+			  params->pri_bytes_per_pixel);
+}
+
+static bool hsw_compute_lp_wm(uint32_t mem_value, struct hsw_wm_maximums *max,
+			      struct hsw_pipe_wm_parameters *params,
+			      struct hsw_lp_wm_result *result)
+{
+	enum pipe pipe;
+	uint32_t pri_val[3], spr_val[3], cur_val[3], fbc_val[3];
+
+	for (pipe = PIPE_A; pipe <= PIPE_C; pipe++) {
+		struct hsw_pipe_wm_parameters *p = &params[pipe];
+
+		pri_val[pipe] = hsw_compute_pri_wm(p, mem_value, true);
+		spr_val[pipe] = hsw_compute_spr_wm(p, mem_value);
+		cur_val[pipe] = hsw_compute_cur_wm(p, mem_value);
+		fbc_val[pipe] = hsw_compute_fbc_wm(p, pri_val[pipe], mem_value);
+	}
+
+	result->pri_val = max3(pri_val[0], pri_val[1], pri_val[2]);
+	result->spr_val = max3(spr_val[0], spr_val[1], spr_val[2]);
+	result->cur_val = max3(cur_val[0], cur_val[1], cur_val[2]);
+	result->fbc_val = max3(fbc_val[0], fbc_val[1], fbc_val[2]);
+
+	if (result->fbc_val > max->fbc) {
+		result->fbc_enable = false;
+		result->fbc_val = 0;
+	} else {
+		result->fbc_enable = true;
+	}
+
+	result->enable = result->pri_val <= max->pri &&
+			 result->spr_val <= max->spr &&
+			 result->cur_val <= max->cur;
+	return result->enable;
+}
+
+static uint32_t hsw_compute_wm_pipe(struct drm_i915_private *dev_priv,
+				    uint32_t mem_value, enum pipe pipe,
+				    struct hsw_pipe_wm_parameters *params)
+{
+	uint32_t pri_val, cur_val, spr_val;
+
+	pri_val = hsw_compute_pri_wm(params, mem_value, false);
+	spr_val = hsw_compute_spr_wm(params, mem_value);
+	cur_val = hsw_compute_cur_wm(params, mem_value);
+
+	WARN(pri_val > 127,
+	     "Primary WM error, mode not supported for pipe %c\n",
+	     pipe_name(pipe));
+	WARN(spr_val > 127,
+	     "Sprite WM error, mode not supported for pipe %c\n",
+	     pipe_name(pipe));
+	WARN(cur_val > 63,
+	     "Cursor WM error, mode not supported for pipe %c\n",
+	     pipe_name(pipe));
+
+	return (pri_val << WM0_PIPE_PLANE_SHIFT) |
+	       (spr_val << WM0_PIPE_SPRITE_SHIFT) |
+	       cur_val;
+}
+
+static uint32_t
+hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 temp;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *mode = &intel_crtc->config.adjusted_mode;
+	u32 linetime, ips_linetime;
 
-	temp = I915_READ(PIPE_WM_LINETIME(pipe));
-	temp &= ~PIPE_WM_LINETIME_MASK;
+	if (!intel_crtc_active(crtc))
+		return 0;
 
 	/* The WM are computed with base on how long it takes to fill a single
 	 * row at the given clock rate, multiplied by 8.
 	 * */
-	temp |= PIPE_WM_LINETIME_TIME(
-		((mode->crtc_hdisplay * 1000) / mode->clock) * 8);
+	linetime = DIV_ROUND_CLOSEST(mode->htotal * 1000 * 8, mode->clock);
+	ips_linetime = DIV_ROUND_CLOSEST(mode->htotal * 1000 * 8,
+					 intel_ddi_get_cdclk_freq(dev_priv));
 
-	/* IPS watermarks are only used by pipe A, and are ignored by
-	 * pipes B and C.  They are calculated similarly to the common
-	 * linetime values, except that we are using CD clock frequency
-	 * in MHz instead of pixel rate for the division.
-	 *
-	 * This is a placeholder for the IPS watermark calculation code.
-	 */
+	return PIPE_WM_LINETIME_IPS_LINETIME(ips_linetime) |
+	       PIPE_WM_LINETIME_TIME(linetime);
+}
+
+static void hsw_compute_wm_parameters(struct drm_device *dev,
+				      struct hsw_pipe_wm_parameters *params,
+				      uint32_t *wm,
+				      struct hsw_wm_maximums *lp_max_1_2,
+				      struct hsw_wm_maximums *lp_max_5_6)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	struct drm_plane *plane;
+	uint64_t sskpd = I915_READ64(MCH_SSKPD);
+	enum pipe pipe;
+	int pipes_active = 0, sprites_enabled = 0;
+
+	if ((sskpd >> 56) & 0xFF)
+		wm[0] = (sskpd >> 56) & 0xFF;
+	else
+		wm[0] = sskpd & 0xF;
+	wm[1] = ((sskpd >> 4) & 0xFF) * 5;
+	wm[2] = ((sskpd >> 12) & 0xFF) * 5;
+	wm[3] = ((sskpd >> 20) & 0x1FF) * 5;
+	wm[4] = ((sskpd >> 32) & 0x1FF) * 5;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+		struct hsw_pipe_wm_parameters *p;
+
+		pipe = intel_crtc->pipe;
+		p = &params[pipe];
+
+		p->active = intel_crtc_active(crtc);
+		if (!p->active)
+			continue;
+
+		pipes_active++;
+
+		p->pipe_htotal = intel_crtc->config.adjusted_mode.htotal;
+		p->pixel_rate = hsw_wm_get_pixel_rate(dev, crtc);
+		p->pri_bytes_per_pixel = crtc->fb->bits_per_pixel / 8;
+		p->cur_bytes_per_pixel = 4;
+		p->pri_horiz_pixels =
+			intel_crtc->config.requested_mode.hdisplay;
+		p->cur_horiz_pixels = 64;
+	}
+
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct intel_plane *intel_plane = to_intel_plane(plane);
+		struct hsw_pipe_wm_parameters *p;
+
+		pipe = intel_plane->pipe;
+		p = &params[pipe];
+
+		p->sprite_enabled = intel_plane->wm.enable;
+		p->spr_bytes_per_pixel = intel_plane->wm.bytes_per_pixel;
+		p->spr_horiz_pixels = intel_plane->wm.horiz_pixels;
+
+		if (p->sprite_enabled)
+			sprites_enabled++;
+	}
 
-	I915_WRITE(PIPE_WM_LINETIME(pipe), temp);
+	if (pipes_active > 1) {
+		lp_max_1_2->pri = lp_max_5_6->pri = sprites_enabled ? 128 : 256;
+		lp_max_1_2->spr = lp_max_5_6->spr = 128;
+		lp_max_1_2->cur = lp_max_5_6->cur = 64;
+	} else {
+		lp_max_1_2->pri = sprites_enabled ? 384 : 768;
+		lp_max_5_6->pri = sprites_enabled ? 128 : 768;
+		lp_max_1_2->spr = 384;
+		lp_max_5_6->spr = 640;
+		lp_max_1_2->cur = lp_max_5_6->cur = 255;
+	}
+	lp_max_1_2->fbc = lp_max_5_6->fbc = 15;
+}
+
+static void hsw_compute_wm_results(struct drm_device *dev,
+				   struct hsw_pipe_wm_parameters *params,
+				   uint32_t *wm,
+				   struct hsw_wm_maximums *lp_maximums,
+				   struct hsw_wm_values *results)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	struct hsw_lp_wm_result lp_results[4] = {};
+	enum pipe pipe;
+	int level, max_level, wm_lp;
+
+	for (level = 1; level <= 4; level++)
+		if (!hsw_compute_lp_wm(wm[level], lp_maximums, params,
+				       &lp_results[level - 1]))
+			break;
+	max_level = level - 1;
+
+	/* The spec says it is preferred to disable FBC WMs instead of disabling
+	 * a WM level. */
+	results->enable_fbc_wm = true;
+	for (level = 1; level <= max_level; level++) {
+		if (!lp_results[level - 1].fbc_enable) {
+			results->enable_fbc_wm = false;
+			break;
+		}
+	}
+
+	memset(results, 0, sizeof(*results));
+	for (wm_lp = 1; wm_lp <= 3; wm_lp++) {
+		const struct hsw_lp_wm_result *r;
+
+		level = (max_level == 4 && wm_lp > 1) ? wm_lp + 1 : wm_lp;
+		if (level > max_level)
+			break;
+
+		r = &lp_results[level - 1];
+		results->wm_lp[wm_lp - 1] = HSW_WM_LP_VAL(level * 2,
+							  r->fbc_val,
+							  r->pri_val,
+							  r->cur_val);
+		results->wm_lp_spr[wm_lp - 1] = r->spr_val;
+	}
+
+	for_each_pipe(pipe)
+		results->wm_pipe[pipe] = hsw_compute_wm_pipe(dev_priv, wm[0],
+							     pipe,
+							     &params[pipe]);
+
+	for_each_pipe(pipe) {
+		crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+		results->wm_linetime[pipe] = hsw_compute_linetime_wm(dev, crtc);
+	}
+}
+
+/* Find the result with the highest level enabled. Check for enable_fbc_wm in
+ * case both are at the same level. Prefer r1 in case they're the same. */
+struct hsw_wm_values *hsw_find_best_result(struct hsw_wm_values *r1,
+					   struct hsw_wm_values *r2)
+{
+	int i, val_r1 = 0, val_r2 = 0;
+
+	for (i = 0; i < 3; i++) {
+		if (r1->wm_lp[i] & WM3_LP_EN)
+			val_r1 = r1->wm_lp[i] & WM1_LP_LATENCY_MASK;
+		if (r2->wm_lp[i] & WM3_LP_EN)
+			val_r2 = r2->wm_lp[i] & WM1_LP_LATENCY_MASK;
+	}
+
+	if (val_r1 == val_r2) {
+		if (r2->enable_fbc_wm && !r1->enable_fbc_wm)
+			return r2;
+		else
+			return r1;
+	} else if (val_r1 > val_r2) {
+		return r1;
+	} else {
+		return r2;
+	}
+}
+
+/*
+ * The spec says we shouldn't write when we don't need, because every write
+ * causes WMs to be re-evaluated, expending some power.
+ */
+static void hsw_write_wm_values(struct drm_i915_private *dev_priv,
+				struct hsw_wm_values *results,
+				enum hsw_data_buf_partitioning partitioning)
+{
+	struct hsw_wm_values previous;
+	uint32_t val;
+	enum hsw_data_buf_partitioning prev_partitioning;
+	bool prev_enable_fbc_wm;
+
+	previous.wm_pipe[0] = I915_READ(WM0_PIPEA_ILK);
+	previous.wm_pipe[1] = I915_READ(WM0_PIPEB_ILK);
+	previous.wm_pipe[2] = I915_READ(WM0_PIPEC_IVB);
+	previous.wm_lp[0] = I915_READ(WM1_LP_ILK);
+	previous.wm_lp[1] = I915_READ(WM2_LP_ILK);
+	previous.wm_lp[2] = I915_READ(WM3_LP_ILK);
+	previous.wm_lp_spr[0] = I915_READ(WM1S_LP_ILK);
+	previous.wm_lp_spr[1] = I915_READ(WM2S_LP_IVB);
+	previous.wm_lp_spr[2] = I915_READ(WM3S_LP_IVB);
+	previous.wm_linetime[0] = I915_READ(PIPE_WM_LINETIME(PIPE_A));
+	previous.wm_linetime[1] = I915_READ(PIPE_WM_LINETIME(PIPE_B));
+	previous.wm_linetime[2] = I915_READ(PIPE_WM_LINETIME(PIPE_C));
+
+	prev_partitioning = (I915_READ(WM_MISC) & WM_MISC_DATA_PARTITION_5_6) ?
+			    HSW_DATA_BUF_PART_5_6 : HSW_DATA_BUF_PART_1_2;
+
+	prev_enable_fbc_wm = !(I915_READ(DISP_ARB_CTL) & DISP_FBC_WM_DIS);
+
+	if (memcmp(results->wm_pipe, previous.wm_pipe,
+		   sizeof(results->wm_pipe)) == 0 &&
+	    memcmp(results->wm_lp, previous.wm_lp,
+		   sizeof(results->wm_lp)) == 0 &&
+	    memcmp(results->wm_lp_spr, previous.wm_lp_spr,
+		   sizeof(results->wm_lp_spr)) == 0 &&
+	    memcmp(results->wm_linetime, previous.wm_linetime,
+		   sizeof(results->wm_linetime)) == 0 &&
+	    partitioning == prev_partitioning &&
+	    results->enable_fbc_wm == prev_enable_fbc_wm)
+		return;
+
+	if (previous.wm_lp[2] != 0)
+		I915_WRITE(WM3_LP_ILK, 0);
+	if (previous.wm_lp[1] != 0)
+		I915_WRITE(WM2_LP_ILK, 0);
+	if (previous.wm_lp[0] != 0)
+		I915_WRITE(WM1_LP_ILK, 0);
+
+	if (previous.wm_pipe[0] != results->wm_pipe[0])
+		I915_WRITE(WM0_PIPEA_ILK, results->wm_pipe[0]);
+	if (previous.wm_pipe[1] != results->wm_pipe[1])
+		I915_WRITE(WM0_PIPEB_ILK, results->wm_pipe[1]);
+	if (previous.wm_pipe[2] != results->wm_pipe[2])
+		I915_WRITE(WM0_PIPEC_IVB, results->wm_pipe[2]);
+
+	if (previous.wm_linetime[0] != results->wm_linetime[0])
+		I915_WRITE(PIPE_WM_LINETIME(PIPE_A), results->wm_linetime[0]);
+	if (previous.wm_linetime[1] != results->wm_linetime[1])
+		I915_WRITE(PIPE_WM_LINETIME(PIPE_B), results->wm_linetime[1]);
+	if (previous.wm_linetime[2] != results->wm_linetime[2])
+		I915_WRITE(PIPE_WM_LINETIME(PIPE_C), results->wm_linetime[2]);
+
+	if (prev_partitioning != partitioning) {
+		val = I915_READ(WM_MISC);
+		if (partitioning == HSW_DATA_BUF_PART_1_2)
+			val &= ~WM_MISC_DATA_PARTITION_5_6;
+		else
+			val |= WM_MISC_DATA_PARTITION_5_6;
+		I915_WRITE(WM_MISC, val);
+	}
+
+	if (prev_enable_fbc_wm != results->enable_fbc_wm) {
+		val = I915_READ(DISP_ARB_CTL);
+		if (results->enable_fbc_wm)
+			val &= ~DISP_FBC_WM_DIS;
+		else
+			val |= DISP_FBC_WM_DIS;
+		I915_WRITE(DISP_ARB_CTL, val);
+	}
+
+	if (previous.wm_lp_spr[0] != results->wm_lp_spr[0])
+		I915_WRITE(WM1S_LP_ILK, results->wm_lp_spr[0]);
+	if (previous.wm_lp_spr[1] != results->wm_lp_spr[1])
+		I915_WRITE(WM2S_LP_IVB, results->wm_lp_spr[1]);
+	if (previous.wm_lp_spr[2] != results->wm_lp_spr[2])
+		I915_WRITE(WM3S_LP_IVB, results->wm_lp_spr[2]);
+
+	if (results->wm_lp[0] != 0)
+		I915_WRITE(WM1_LP_ILK, results->wm_lp[0]);
+	if (results->wm_lp[1] != 0)
+		I915_WRITE(WM2_LP_ILK, results->wm_lp[1]);
+	if (results->wm_lp[2] != 0)
+		I915_WRITE(WM3_LP_ILK, results->wm_lp[2]);
+}
+
+static void haswell_update_wm(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct hsw_wm_maximums lp_max_1_2, lp_max_5_6;
+	struct hsw_pipe_wm_parameters params[3];
+	struct hsw_wm_values results_1_2, results_5_6, *best_results;
+	uint32_t wm[5];
+	enum hsw_data_buf_partitioning partitioning;
+
+	hsw_compute_wm_parameters(dev, params, wm, &lp_max_1_2, &lp_max_5_6);
+
+	hsw_compute_wm_results(dev, params, wm, &lp_max_1_2, &results_1_2);
+	if (lp_max_1_2.pri != lp_max_5_6.pri) {
+		hsw_compute_wm_results(dev, params, wm, &lp_max_5_6,
+				       &results_5_6);
+		best_results = hsw_find_best_result(&results_1_2, &results_5_6);
+	} else {
+		best_results = &results_1_2;
+	}
+
+	partitioning = (best_results == &results_1_2) ?
+		       HSW_DATA_BUF_PART_1_2 : HSW_DATA_BUF_PART_5_6;
+
+	hsw_write_wm_values(dev_priv, best_results, partitioning);
+}
+
+static void haswell_update_sprite_wm(struct drm_device *dev, int pipe,
+				     uint32_t sprite_width, int pixel_size,
+				     bool enable)
+{
+	struct drm_plane *plane;
+
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct intel_plane *intel_plane = to_intel_plane(plane);
+
+		if (intel_plane->pipe == pipe) {
+			intel_plane->wm.enable = enable;
+			intel_plane->wm.horiz_pixels = sprite_width + 1;
+			intel_plane->wm.bytes_per_pixel = pixel_size;
+			break;
+		}
+	}
+
+	haswell_update_wm(dev);
 }
 
 static bool
@@ -2120,7 +2712,8 @@ sandybridge_compute_sprite_srwm(struct drm_device *dev, int plane,
 }
 
 static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
-					 uint32_t sprite_width, int pixel_size)
+					 uint32_t sprite_width, int pixel_size,
+					 bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
@@ -2128,6 +2721,9 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 	int sprite_wm, reg;
 	int ret;
 
+	if (!enable)
+		return;
+
 	switch (pipe) {
 	case 0:
 		reg = WM0_PIPEA_ILK;
@@ -2146,15 +2742,15 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					    &sandybridge_display_wm_info,
 					    latency, &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite wm for pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite wm for pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 
 	val = I915_READ(reg);
 	val &= ~WM0_PIPE_SPRITE_MASK;
 	I915_WRITE(reg, val | (sprite_wm << WM0_PIPE_SPRITE_SHIFT));
-	DRM_DEBUG_KMS("sprite watermarks For pipe %d - %d\n", pipe, sprite_wm);
+	DRM_DEBUG_KMS("sprite watermarks For pipe %c - %d\n", pipe_name(pipe), sprite_wm);
 
 
 	ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width,
@@ -2163,8 +2759,8 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					      SNB_READ_WM1_LATENCY() * 500,
 					      &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp1 wm on pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite lp1 wm on pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 	I915_WRITE(WM1S_LP_ILK, sprite_wm);
@@ -2179,8 +2775,8 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					      SNB_READ_WM2_LATENCY() * 500,
 					      &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp2 wm on pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite lp2 wm on pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 	I915_WRITE(WM2S_LP_IVB, sprite_wm);
@@ -2191,8 +2787,8 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					      SNB_READ_WM3_LATENCY() * 500,
 					      &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp3 wm on pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite lp3 wm on pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 	I915_WRITE(WM3S_LP_IVB, sprite_wm);
@@ -2238,23 +2834,15 @@ void intel_update_watermarks(struct drm_device *dev)
 		dev_priv->display.update_wm(dev);
 }
 
-void intel_update_linetime_watermarks(struct drm_device *dev,
-		int pipe, struct drm_display_mode *mode)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (dev_priv->display.update_linetime_wm)
-		dev_priv->display.update_linetime_wm(dev, pipe, mode);
-}
-
 void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
-				    uint32_t sprite_width, int pixel_size)
+				    uint32_t sprite_width, int pixel_size,
+				    bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	if (dev_priv->display.update_sprite_wm)
 		dev_priv->display.update_sprite_wm(dev, pipe, sprite_width,
-						   pixel_size);
+						   pixel_size, enable);
 }
 
 static struct drm_i915_gem_object *
@@ -2481,6 +3069,67 @@ void gen6_set_rps(struct drm_device *dev, u8 val)
 	trace_intel_gpu_freq_change(val * 50);
 }
 
+/*
+ * Wait until the previous freq change has completed,
+ * or the timeout elapsed, and then update our notion
+ * of the current GPU frequency.
+ */
+static void vlv_update_rps_cur_delay(struct drm_i915_private *dev_priv)
+{
+	unsigned long timeout = jiffies + msecs_to_jiffies(10);
+	u32 pval;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	do {
+		pval = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+		if (time_after(jiffies, timeout)) {
+			DRM_DEBUG_DRIVER("timed out waiting for Punit\n");
+			break;
+		}
+		udelay(10);
+	} while (pval & 1);
+
+	pval >>= 8;
+
+	if (pval != dev_priv->rps.cur_delay)
+		DRM_DEBUG_DRIVER("Punit overrode GPU freq: %d MHz (%u) requested, but got %d Mhz (%u)\n",
+				 vlv_gpu_freq(dev_priv->mem_freq, dev_priv->rps.cur_delay),
+				 dev_priv->rps.cur_delay,
+				 vlv_gpu_freq(dev_priv->mem_freq, pval), pval);
+
+	dev_priv->rps.cur_delay = pval;
+}
+
+void valleyview_set_rps(struct drm_device *dev, u8 val)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	gen6_rps_limits(dev_priv, &val);
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+	WARN_ON(val > dev_priv->rps.max_delay);
+	WARN_ON(val < dev_priv->rps.min_delay);
+
+	vlv_update_rps_cur_delay(dev_priv);
+
+	DRM_DEBUG_DRIVER("GPU freq request from %d MHz (%u) to %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.cur_delay),
+			 dev_priv->rps.cur_delay,
+			 vlv_gpu_freq(dev_priv->mem_freq, val), val);
+
+	if (val == dev_priv->rps.cur_delay)
+		return;
+
+	vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
+
+	dev_priv->rps.cur_delay = val;
+
+	trace_intel_gpu_freq_change(vlv_gpu_freq(dev_priv->mem_freq, val));
+}
+
+
 static void gen6_disable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -2488,6 +3137,25 @@ static void gen6_disable_rps(struct drm_device *dev)
 	I915_WRITE(GEN6_RC_CONTROL, 0);
 	I915_WRITE(GEN6_RPNSWREQ, 1 << 31);
 	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
+	I915_WRITE(GEN6_PMIER, I915_READ(GEN6_PMIER) & ~GEN6_PM_RPS_EVENTS);
+	/* Complete PM interrupt masking here doesn't race with the rps work
+	 * item again unmasking PM interrupts because that is using a different
+	 * register (PMIMR) to mask PM interrupts. The only risk is in leaving
+	 * stale bits in PMIIR and PMIMR which gen6_enable_rps will clean up. */
+
+	spin_lock_irq(&dev_priv->rps.lock);
+	dev_priv->rps.pm_iir = 0;
+	spin_unlock_irq(&dev_priv->rps.lock);
+
+	I915_WRITE(GEN6_PMIIR, GEN6_PM_RPS_EVENTS);
+}
+
+static void valleyview_disable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE(GEN6_RC_CONTROL, 0);
+	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
 	I915_WRITE(GEN6_PMIER, 0);
 	/* Complete PM interrupt masking here doesn't race with the rps work
 	 * item again unmasking PM interrupts because that is using a different
@@ -2499,6 +3167,11 @@ static void gen6_disable_rps(struct drm_device *dev)
 	spin_unlock_irq(&dev_priv->rps.lock);
 
 	I915_WRITE(GEN6_PMIIR, I915_READ(GEN6_PMIIR));
+
+	if (dev_priv->vlv_pctx) {
+		drm_gem_object_unreference(&dev_priv->vlv_pctx->base);
+		dev_priv->vlv_pctx = NULL;
+	}
 }
 
 int intel_enable_rc6(const struct drm_device *dev)
@@ -2655,12 +3328,15 @@ static void gen6_enable_rps(struct drm_device *dev)
 	gen6_set_rps(dev_priv->dev, (gt_perf_status & 0xff00) >> 8);
 
 	/* requires MSI enabled */
-	I915_WRITE(GEN6_PMIER, GEN6_PM_DEFERRED_EVENTS);
+	I915_WRITE(GEN6_PMIER, I915_READ(GEN6_PMIER) | GEN6_PM_RPS_EVENTS);
 	spin_lock_irq(&dev_priv->rps.lock);
-	WARN_ON(dev_priv->rps.pm_iir != 0);
-	I915_WRITE(GEN6_PMIMR, 0);
+	/* FIXME: Our interrupt enabling sequence is bonghits.
+	 * dev_priv->rps.pm_iir really should be 0 here. */
+	dev_priv->rps.pm_iir = 0;
+	I915_WRITE(GEN6_PMIMR, I915_READ(GEN6_PMIMR) & ~GEN6_PM_RPS_EVENTS);
+	I915_WRITE(GEN6_PMIIR, GEN6_PM_RPS_EVENTS);
 	spin_unlock_irq(&dev_priv->rps.lock);
-	/* enable all PM interrupts */
+	/* unmask all PM interrupts */
 	I915_WRITE(GEN6_PMINTRMSK, 0);
 
 	rc6vids = 0;
@@ -2742,6 +3418,207 @@ static void gen6_update_ring_freq(struct drm_device *dev)
 	}
 }
 
+int valleyview_rps_max_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rp0;
+
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FREQ_FUSE);
+
+	rp0 = (val & FB_GFX_MAX_FREQ_FUSE_MASK) >> FB_GFX_MAX_FREQ_FUSE_SHIFT;
+	/* Clamp to max */
+	rp0 = min_t(u32, rp0, 0xea);
+
+	return rp0;
+}
+
+static int valleyview_rps_rpe_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rpe;
+
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FMAX_FUSE_LO);
+	rpe = (val & FB_FMAX_VMIN_FREQ_LO_MASK) >> FB_FMAX_VMIN_FREQ_LO_SHIFT;
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FMAX_FUSE_HI);
+	rpe |= (val & FB_FMAX_VMIN_FREQ_HI_MASK) << 5;
+
+	return rpe;
+}
+
+int valleyview_rps_min_freq(struct drm_i915_private *dev_priv)
+{
+	return vlv_punit_read(dev_priv, PUNIT_REG_GPU_LFM) & 0xff;
+}
+
+static void vlv_rps_timer_work(struct work_struct *work)
+{
+	drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t,
+						    rps.vlv_work.work);
+
+	/*
+	 * Timer fired, we must be idle.  Drop to min voltage state.
+	 * Note: we use RPe here since it should match the
+	 * Vmin we were shooting for.  That should give us better
+	 * perf when we come back out of RC6 than if we used the
+	 * min freq available.
+	 */
+	mutex_lock(&dev_priv->rps.hw_lock);
+	if (dev_priv->rps.cur_delay > dev_priv->rps.rpe_delay)
+		valleyview_set_rps(dev_priv->dev, dev_priv->rps.rpe_delay);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
+static void valleyview_setup_pctx(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *pctx;
+	unsigned long pctx_paddr;
+	u32 pcbr;
+	int pctx_size = 24*1024;
+
+	pcbr = I915_READ(VLV_PCBR);
+	if (pcbr) {
+		/* BIOS set it up already, grab the pre-alloc'd space */
+		int pcbr_offset;
+
+		pcbr_offset = (pcbr & (~4095)) - dev_priv->mm.stolen_base;
+		pctx = i915_gem_object_create_stolen_for_preallocated(dev_priv->dev,
+								      pcbr_offset,
+								      -1,
+								      pctx_size);
+		goto out;
+	}
+
+	/*
+	 * From the Gunit register HAS:
+	 * The Gfx driver is expected to program this register and ensure
+	 * proper allocation within Gfx stolen memory.  For example, this
+	 * register should be programmed such than the PCBR range does not
+	 * overlap with other ranges, such as the frame buffer, protected
+	 * memory, or any other relevant ranges.
+	 */
+	pctx = i915_gem_object_create_stolen(dev, pctx_size);
+	if (!pctx) {
+		DRM_DEBUG("not enough stolen space for PCTX, disabling\n");
+		return;
+	}
+
+	pctx_paddr = dev_priv->mm.stolen_base + pctx->stolen->start;
+	I915_WRITE(VLV_PCBR, pctx_paddr);
+
+out:
+	dev_priv->vlv_pctx = pctx;
+}
+
+static void valleyview_enable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *ring;
+	u32 gtfifodbg, val;
+	int i;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	if ((gtfifodbg = I915_READ(GTFIFODBG))) {
+		DRM_ERROR("GT fifo had a previous error %x\n", gtfifodbg);
+		I915_WRITE(GTFIFODBG, gtfifodbg);
+	}
+
+	valleyview_setup_pctx(dev);
+
+	gen6_gt_force_wake_get(dev_priv);
+
+	I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400);
+	I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000);
+	I915_WRITE(GEN6_RP_UP_EI, 66000);
+	I915_WRITE(GEN6_RP_DOWN_EI, 350000);
+
+	I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	I915_WRITE(GEN6_RP_CONTROL,
+		   GEN6_RP_MEDIA_TURBO |
+		   GEN6_RP_MEDIA_HW_NORMAL_MODE |
+		   GEN6_RP_MEDIA_IS_GFX |
+		   GEN6_RP_ENABLE |
+		   GEN6_RP_UP_BUSY_AVG |
+		   GEN6_RP_DOWN_IDLE_CONT);
+
+	I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 0x00280000);
+	I915_WRITE(GEN6_RC_EVALUATION_INTERVAL, 125000);
+	I915_WRITE(GEN6_RC_IDLE_HYSTERSIS, 25);
+
+	for_each_ring(ring, dev_priv, i)
+		I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10);
+
+	I915_WRITE(GEN6_RC6_THRESHOLD, 0xc350);
+
+	/* allows RC6 residency counter to work */
+	I915_WRITE(0x138104, _MASKED_BIT_ENABLE(0x3));
+	I915_WRITE(GEN6_RC_CONTROL,
+		   GEN7_RC_CTL_TO_MODE);
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+	switch ((val >> 6) & 3) {
+	case 0:
+	case 1:
+		dev_priv->mem_freq = 800;
+		break;
+	case 2:
+		dev_priv->mem_freq = 1066;
+		break;
+	case 3:
+		dev_priv->mem_freq = 1333;
+		break;
+	}
+	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
+
+	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & 0x10 ? "yes" : "no");
+	DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val);
+
+	dev_priv->rps.cur_delay = (val >> 8) & 0xff;
+	DRM_DEBUG_DRIVER("current GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.cur_delay),
+			 dev_priv->rps.cur_delay);
+
+	dev_priv->rps.max_delay = valleyview_rps_max_freq(dev_priv);
+	dev_priv->rps.hw_max = dev_priv->rps.max_delay;
+	DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.max_delay),
+			 dev_priv->rps.max_delay);
+
+	dev_priv->rps.rpe_delay = valleyview_rps_rpe_freq(dev_priv);
+	DRM_DEBUG_DRIVER("RPe GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.rpe_delay),
+			 dev_priv->rps.rpe_delay);
+
+	dev_priv->rps.min_delay = valleyview_rps_min_freq(dev_priv);
+	DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.min_delay),
+			 dev_priv->rps.min_delay);
+
+	DRM_DEBUG_DRIVER("setting GPU freq to %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.rpe_delay),
+			 dev_priv->rps.rpe_delay);
+
+	INIT_DELAYED_WORK(&dev_priv->rps.vlv_work, vlv_rps_timer_work);
+
+	valleyview_set_rps(dev_priv->dev, dev_priv->rps.rpe_delay);
+
+	/* requires MSI enabled */
+	I915_WRITE(GEN6_PMIER, GEN6_PM_RPS_EVENTS);
+	spin_lock_irq(&dev_priv->rps.lock);
+	WARN_ON(dev_priv->rps.pm_iir != 0);
+	I915_WRITE(GEN6_PMIMR, 0);
+	spin_unlock_irq(&dev_priv->rps.lock);
+	/* enable all PM interrupts */
+	I915_WRITE(GEN6_PMINTRMSK, 0);
+
+	gen6_gt_force_wake_put(dev_priv);
+}
+
 void ironlake_teardown_rc6(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3465,13 +4342,22 @@ void intel_disable_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	/* Interrupts should be disabled already to avoid re-arming. */
+	WARN_ON(dev->irq_enabled);
+
 	if (IS_IRONLAKE_M(dev)) {
 		ironlake_disable_drps(dev);
 		ironlake_disable_rc6(dev);
-	} else if (INTEL_INFO(dev)->gen >= 6 && !IS_VALLEYVIEW(dev)) {
+	} else if (INTEL_INFO(dev)->gen >= 6) {
 		cancel_delayed_work_sync(&dev_priv->rps.delayed_resume_work);
+		cancel_work_sync(&dev_priv->rps.work);
+		if (IS_VALLEYVIEW(dev))
+			cancel_delayed_work_sync(&dev_priv->rps.vlv_work);
 		mutex_lock(&dev_priv->rps.hw_lock);
-		gen6_disable_rps(dev);
+		if (IS_VALLEYVIEW(dev))
+			valleyview_disable_rps(dev);
+		else
+			gen6_disable_rps(dev);
 		mutex_unlock(&dev_priv->rps.hw_lock);
 	}
 }
@@ -3484,8 +4370,13 @@ static void intel_gen6_powersave_work(struct work_struct *work)
 	struct drm_device *dev = dev_priv->dev;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	gen6_enable_rps(dev);
-	gen6_update_ring_freq(dev);
+
+	if (IS_VALLEYVIEW(dev)) {
+		valleyview_enable_rps(dev);
+	} else {
+		gen6_enable_rps(dev);
+		gen6_update_ring_freq(dev);
+	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
@@ -3497,7 +4388,7 @@ void intel_enable_gt_powersave(struct drm_device *dev)
 		ironlake_enable_drps(dev);
 		ironlake_enable_rc6(dev);
 		intel_init_emon(dev);
-	} else if ((IS_GEN6(dev) || IS_GEN7(dev)) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_GEN6(dev) || IS_GEN7(dev)) {
 		/*
 		 * PCU communication is slow and this doesn't need to be
 		 * done at any specific time, so do this out of our fast path
@@ -3520,6 +4411,19 @@ static void ibx_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(SOUTH_DSPCLK_GATE_D, PCH_DPLSUNIT_CLOCK_GATE_DISABLE);
 }
 
+static void g4x_disable_trickle_feed(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe;
+
+	for_each_pipe(pipe) {
+		I915_WRITE(DSPCNTR(pipe),
+			   I915_READ(DSPCNTR(pipe)) |
+			   DISPPLANE_TRICKLE_FEED_DISABLE);
+		intel_flush_display_plane(dev_priv, pipe);
+	}
+}
+
 static void ironlake_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3579,10 +4483,12 @@ static void ironlake_init_clock_gating(struct drm_device *dev)
 		   _3D_CHICKEN2_WM_READ_PIPELINED << 16 |
 		   _3D_CHICKEN2_WM_READ_PIPELINED);
 
-	/* WaDisableRenderCachePipelinedFlush */
+	/* WaDisableRenderCachePipelinedFlush:ilk */
 	I915_WRITE(CACHE_MODE_0,
 		   _MASKED_BIT_ENABLE(CM0_PIPELINED_RENDER_FLUSH_DISABLE));
 
+	g4x_disable_trickle_feed(dev);
+
 	ibx_init_clock_gating(dev);
 }
 
@@ -3607,7 +4513,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 		val = I915_READ(TRANS_CHICKEN2(pipe));
 		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
 		val &= ~TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
-		if (dev_priv->fdi_rx_polarity_inverted)
+		if (dev_priv->vbt.fdi_rx_polarity_inverted)
 			val |= TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
 		val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
 		val &= ~TRANS_CHICKEN2_DISABLE_DEEP_COLOR_COUNTER;
@@ -3637,7 +4543,6 @@ static void gen6_check_mch_setup(struct drm_device *dev)
 static void gen6_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 	uint32_t dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE;
 
 	I915_WRITE(ILK_DSPCLK_GATE_D, dspclk_gate);
@@ -3646,11 +4551,11 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 		   I915_READ(ILK_DISPLAY_CHICKEN2) |
 		   ILK_ELPIN_409_SELECT);
 
-	/* WaDisableHiZPlanesWhenMSAAEnabled */
+	/* WaDisableHiZPlanesWhenMSAAEnabled:snb */
 	I915_WRITE(_3D_CHICKEN,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB));
 
-	/* WaSetupGtModeTdRowDispatch */
+	/* WaSetupGtModeTdRowDispatch:snb */
 	if (IS_SNB_GT1(dev))
 		I915_WRITE(GEN6_GT_MODE,
 			   _MASKED_BIT_ENABLE(GEN6_TD_FOUR_ROW_DISPATCH_DISABLE));
@@ -3677,8 +4582,8 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 	 * According to the spec, bit 11 (RCCUNIT) must also be set,
 	 * but we didn't debug actual testcases to find it out.
 	 *
-	 * Also apply WaDisableVDSUnitClockGating and
-	 * WaDisableRCPBUnitClockGating.
+	 * Also apply WaDisableVDSUnitClockGating:snb and
+	 * WaDisableRCPBUnitClockGating:snb.
 	 */
 	I915_WRITE(GEN6_UCGCTL2,
 		   GEN7_VDSUNIT_CLOCK_GATE_DISABLE |
@@ -3709,16 +4614,11 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 		   ILK_DPARBUNIT_CLOCK_GATE_ENABLE  |
 		   ILK_DPFDUNIT_CLOCK_GATE_ENABLE);
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:snb */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	g4x_disable_trickle_feed(dev);
 
 	/* The default value should be 0x200 according to docs, but the two
 	 * platforms I checked have a 0 for this. (Maybe BIOS overrides?) */
@@ -3739,7 +4639,6 @@ static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv)
 	reg |= GEN7_FF_VS_SCHED_HW;
 	reg |= GEN7_FF_DS_SCHED_HW;
 
-	/* WaVSRefCountFullforceMissDisable */
 	if (IS_HASWELL(dev_priv->dev))
 		reg &= ~GEN7_FF_VS_REF_CNT_FFME;
 
@@ -3758,65 +4657,72 @@ static void lpt_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(SOUTH_DSPCLK_GATE_D,
 			   I915_READ(SOUTH_DSPCLK_GATE_D) |
 			   PCH_LP_PARTITION_LEVEL_DISABLE);
+
+	/* WADPOClockGatingDisable:hsw */
+	I915_WRITE(_TRANSA_CHICKEN1,
+		   I915_READ(_TRANSA_CHICKEN1) |
+		   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
+}
+
+static void lpt_suspend_hw(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->pch_id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE) {
+		uint32_t val = I915_READ(SOUTH_DSPCLK_GATE_D);
+
+		val &= ~PCH_LP_PARTITION_LEVEL_DISABLE;
+		I915_WRITE(SOUTH_DSPCLK_GATE_D, val);
+	}
 }
 
 static void haswell_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 
 	I915_WRITE(WM3_LP_ILK, 0);
 	I915_WRITE(WM2_LP_ILK, 0);
 	I915_WRITE(WM1_LP_ILK, 0);
 
 	/* According to the spec, bit 13 (RCZUNIT) must be set on IVB.
-	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 * This implements the WaDisableRCZUnitClockGating:hsw workaround.
 	 */
 	I915_WRITE(GEN6_UCGCTL2, GEN6_RCZUNIT_CLOCK_GATE_DISABLE);
 
-	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:hsw workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
 		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
 
-	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	/* WaApplyL3ControlAndL3ChickenMode:hsw */
 	I915_WRITE(GEN7_L3CNTLREG1,
 			GEN7_WA_FOR_GEN7_L3_CONTROL);
 	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER,
 			GEN7_WA_L3_CHICKEN_MODE);
 
-	/* This is required by WaCatErrorRejectionIssue */
+	/* This is required by WaCatErrorRejectionIssue:hsw */
 	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
 			I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
 			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	g4x_disable_trickle_feed(dev);
 
+	/* WaVSRefCountFullforceMissDisable:hsw */
 	gen7_setup_fixed_func_scheduler(dev_priv);
 
-	/* WaDisable4x2SubspanOptimization */
+	/* WaDisable4x2SubspanOptimization:hsw */
 	I915_WRITE(CACHE_MODE_1,
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:hsw */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
-	/* WaSwitchSolVfFArbitrationPriority */
+	/* WaSwitchSolVfFArbitrationPriority:hsw */
 	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
 
-	/* XXX: This is a workaround for early silicon revisions and should be
-	 * removed later.
-	 */
-	I915_WRITE(WM_DBG,
-			I915_READ(WM_DBG) |
-			WM_DBG_DISALLOW_MULTIPLE_LP |
-			WM_DBG_DISALLOW_SPRITE |
-			WM_DBG_DISALLOW_MAXFIFO);
+	/* WaRsPkgCStateDisplayPMReq:hsw */
+	I915_WRITE(CHICKEN_PAR1_1,
+		   I915_READ(CHICKEN_PAR1_1) | FORCE_ARB_IDLE_PLANES);
 
 	lpt_init_clock_gating(dev);
 }
@@ -3824,7 +4730,6 @@ static void haswell_init_clock_gating(struct drm_device *dev)
 static void ivybridge_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 	uint32_t snpcr;
 
 	I915_WRITE(WM3_LP_ILK, 0);
@@ -3833,16 +4738,16 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 
 	I915_WRITE(ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE);
 
-	/* WaDisableEarlyCull */
+	/* WaDisableEarlyCull:ivb */
 	I915_WRITE(_3D_CHICKEN3,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_SF_DISABLE_OBJEND_CULL));
 
-	/* WaDisableBackToBackFlipFix */
+	/* WaDisableBackToBackFlipFix:ivb */
 	I915_WRITE(IVB_CHICKEN3,
 		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
 		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
 
-	/* WaDisablePSDDualDispatchEnable */
+	/* WaDisablePSDDualDispatchEnable:ivb */
 	if (IS_IVB_GT1(dev))
 		I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
 			   _MASKED_BIT_ENABLE(GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
@@ -3850,11 +4755,11 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(GEN7_HALF_SLICE_CHICKEN1_GT2,
 			   _MASKED_BIT_ENABLE(GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
 
-	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:ivb workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
 		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
 
-	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	/* WaApplyL3ControlAndL3ChickenMode:ivb */
 	I915_WRITE(GEN7_L3CNTLREG1,
 			GEN7_WA_FOR_GEN7_L3_CONTROL);
 	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER,
@@ -3867,7 +4772,7 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 
 
-	/* WaForceL3Serialization */
+	/* WaForceL3Serialization:ivb */
 	I915_WRITE(GEN7_L3SQCREG4, I915_READ(GEN7_L3SQCREG4) &
 		   ~L3SQ_URB_READ_CAM_MATCH_DISABLE);
 
@@ -3882,31 +4787,27 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 	 * but we didn't debug actual testcases to find it out.
 	 *
 	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
-	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 * This implements the WaDisableRCZUnitClockGating:ivb workaround.
 	 */
 	I915_WRITE(GEN6_UCGCTL2,
 		   GEN6_RCZUNIT_CLOCK_GATE_DISABLE |
 		   GEN6_RCCUNIT_CLOCK_GATE_DISABLE);
 
-	/* This is required by WaCatErrorRejectionIssue */
+	/* This is required by WaCatErrorRejectionIssue:ivb */
 	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
 			I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
 			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	g4x_disable_trickle_feed(dev);
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:ivb */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
+	/* WaVSRefCountFullforceMissDisable:ivb */
 	gen7_setup_fixed_func_scheduler(dev_priv);
 
-	/* WaDisable4x2SubspanOptimization */
+	/* WaDisable4x2SubspanOptimization:ivb */
 	I915_WRITE(CACHE_MODE_1,
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
@@ -3924,54 +4825,45 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 static void valleyview_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
-
-	I915_WRITE(WM3_LP_ILK, 0);
-	I915_WRITE(WM2_LP_ILK, 0);
-	I915_WRITE(WM1_LP_ILK, 0);
 
-	I915_WRITE(ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE);
+	I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE);
 
-	/* WaDisableEarlyCull */
+	/* WaDisableEarlyCull:vlv */
 	I915_WRITE(_3D_CHICKEN3,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_SF_DISABLE_OBJEND_CULL));
 
-	/* WaDisableBackToBackFlipFix */
+	/* WaDisableBackToBackFlipFix:vlv */
 	I915_WRITE(IVB_CHICKEN3,
 		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
 		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
 
-	/* WaDisablePSDDualDispatchEnable */
+	/* WaDisablePSDDualDispatchEnable:vlv */
 	I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
 		   _MASKED_BIT_ENABLE(GEN7_MAX_PS_THREAD_DEP |
 				      GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
 
-	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:vlv workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
 		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
 
-	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	/* WaApplyL3ControlAndL3ChickenMode:vlv */
 	I915_WRITE(GEN7_L3CNTLREG1, I915_READ(GEN7_L3CNTLREG1) | GEN7_L3AGDIS);
 	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER, GEN7_WA_L3_CHICKEN_MODE);
 
-	/* WaForceL3Serialization */
+	/* WaForceL3Serialization:vlv */
 	I915_WRITE(GEN7_L3SQCREG4, I915_READ(GEN7_L3SQCREG4) &
 		   ~L3SQ_URB_READ_CAM_MATCH_DISABLE);
 
-	/* WaDisableDopClockGating */
+	/* WaDisableDopClockGating:vlv */
 	I915_WRITE(GEN7_ROW_CHICKEN2,
 		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 
-	/* WaForceL3Serialization */
-	I915_WRITE(GEN7_L3SQCREG4, I915_READ(GEN7_L3SQCREG4) &
-		   ~L3SQ_URB_READ_CAM_MATCH_DISABLE);
-
-	/* This is required by WaCatErrorRejectionIssue */
+	/* This is required by WaCatErrorRejectionIssue:vlv */
 	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
 		   I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
 		   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:vlv */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
@@ -3987,10 +4879,10 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 	 * but we didn't debug actual testcases to find it out.
 	 *
 	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
-	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 * This implements the WaDisableRCZUnitClockGating:vlv workaround.
 	 *
-	 * Also apply WaDisableVDSUnitClockGating and
-	 * WaDisableRCPBUnitClockGating.
+	 * Also apply WaDisableVDSUnitClockGating:vlv and
+	 * WaDisableRCPBUnitClockGating:vlv.
 	 */
 	I915_WRITE(GEN6_UCGCTL2,
 		   GEN7_VDSUNIT_CLOCK_GATE_DISABLE |
@@ -4001,18 +4893,13 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 
 	I915_WRITE(GEN7_UCGCTL4, GEN7_L3BANK2X_CLOCK_GATE_DISABLE);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	I915_WRITE(MI_ARB_VLV, MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE);
 
 	I915_WRITE(CACHE_MODE_1,
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
 	/*
-	 * WaDisableVLVClockGating_VBIIssue
+	 * WaDisableVLVClockGating_VBIIssue:vlv
 	 * Disable clock gating on th GCFG unit to prevent a delay
 	 * in the reporting of vblank events.
 	 */
@@ -4048,6 +4935,8 @@ static void g4x_init_clock_gating(struct drm_device *dev)
 	/* WaDisableRenderCachePipelinedFlush */
 	I915_WRITE(CACHE_MODE_0,
 		   _MASKED_BIT_ENABLE(CM0_PIPELINED_RENDER_FLUSH_DISABLE));
+
+	g4x_disable_trickle_feed(dev);
 }
 
 static void crestline_init_clock_gating(struct drm_device *dev)
@@ -4059,6 +4948,8 @@ static void crestline_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(DSPCLK_GATE_D, 0);
 	I915_WRITE(RAMCLK_GATE_D, 0);
 	I915_WRITE16(DEUC, 0);
+	I915_WRITE(MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void broadwater_init_clock_gating(struct drm_device *dev)
@@ -4071,6 +4962,8 @@ static void broadwater_init_clock_gating(struct drm_device *dev)
 		   I965_ISC_CLOCK_GATE_DISABLE |
 		   I965_FBC_CLOCK_GATE_DISABLE);
 	I915_WRITE(RENCLK_GATE_D2, 0);
+	I915_WRITE(MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void gen3_init_clock_gating(struct drm_device *dev)
@@ -4110,34 +5003,50 @@ void intel_init_clock_gating(struct drm_device *dev)
 	dev_priv->display.init_clock_gating(dev);
 }
 
+void intel_suspend_hw(struct drm_device *dev)
+{
+	if (HAS_PCH_LPT(dev))
+		lpt_suspend_hw(dev);
+}
+
 /**
  * We should only use the power well if we explicitly asked the hardware to
  * enable it, so check if it's enabled and also check if we've requested it to
  * be enabled.
  */
-bool intel_using_power_well(struct drm_device *dev)
+bool intel_display_power_enabled(struct drm_device *dev,
+				 enum intel_display_power_domain domain)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev))
+	if (!HAS_POWER_WELL(dev))
+		return true;
+
+	switch (domain) {
+	case POWER_DOMAIN_PIPE_A:
+	case POWER_DOMAIN_TRANSCODER_EDP:
+		return true;
+	case POWER_DOMAIN_PIPE_B:
+	case POWER_DOMAIN_PIPE_C:
+	case POWER_DOMAIN_PIPE_A_PANEL_FITTER:
+	case POWER_DOMAIN_PIPE_B_PANEL_FITTER:
+	case POWER_DOMAIN_PIPE_C_PANEL_FITTER:
+	case POWER_DOMAIN_TRANSCODER_A:
+	case POWER_DOMAIN_TRANSCODER_B:
+	case POWER_DOMAIN_TRANSCODER_C:
 		return I915_READ(HSW_PWR_WELL_DRIVER) ==
 		       (HSW_PWR_WELL_ENABLE | HSW_PWR_WELL_STATE);
-	else
-		return true;
+	default:
+		BUG();
+	}
 }
 
-void intel_set_power_well(struct drm_device *dev, bool enable)
+static void __intel_set_power_well(struct drm_device *dev, bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool is_enabled, enable_requested;
 	uint32_t tmp;
 
-	if (!HAS_POWER_WELL(dev))
-		return;
-
-	if (!i915_disable_power_well && !enable)
-		return;
-
 	tmp = I915_READ(HSW_PWR_WELL_DRIVER);
 	is_enabled = tmp & HSW_PWR_WELL_STATE;
 	enable_requested = tmp & HSW_PWR_WELL_ENABLE;
@@ -4160,6 +5069,79 @@ void intel_set_power_well(struct drm_device *dev, bool enable)
 	}
 }
 
+static struct i915_power_well *hsw_pwr;
+
+/* Display audio driver power well request */
+void i915_request_power_well(void)
+{
+	if (WARN_ON(!hsw_pwr))
+		return;
+
+	spin_lock_irq(&hsw_pwr->lock);
+	if (!hsw_pwr->count++ &&
+			!hsw_pwr->i915_request)
+		__intel_set_power_well(hsw_pwr->device, true);
+	spin_unlock_irq(&hsw_pwr->lock);
+}
+EXPORT_SYMBOL_GPL(i915_request_power_well);
+
+/* Display audio driver power well release */
+void i915_release_power_well(void)
+{
+	if (WARN_ON(!hsw_pwr))
+		return;
+
+	spin_lock_irq(&hsw_pwr->lock);
+	WARN_ON(!hsw_pwr->count);
+	if (!--hsw_pwr->count &&
+		       !hsw_pwr->i915_request)
+		__intel_set_power_well(hsw_pwr->device, false);
+	spin_unlock_irq(&hsw_pwr->lock);
+}
+EXPORT_SYMBOL_GPL(i915_release_power_well);
+
+int i915_init_power_well(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	hsw_pwr = &dev_priv->power_well;
+
+	hsw_pwr->device = dev;
+	spin_lock_init(&hsw_pwr->lock);
+	hsw_pwr->count = 0;
+
+	return 0;
+}
+
+void i915_remove_power_well(struct drm_device *dev)
+{
+	hsw_pwr = NULL;
+}
+
+void intel_set_power_well(struct drm_device *dev, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_power_well *power_well = &dev_priv->power_well;
+
+	if (!HAS_POWER_WELL(dev))
+		return;
+
+	if (!i915_disable_power_well && !enable)
+		return;
+
+	spin_lock_irq(&power_well->lock);
+	power_well->i915_request = enable;
+
+	/* only reject "disable" power well request */
+	if (power_well->count && !enable) {
+		spin_unlock_irq(&power_well->lock);
+		return;
+	}
+
+	__intel_set_power_well(dev, enable);
+	spin_unlock_irq(&power_well->lock);
+}
+
 /*
  * Starting with Haswell, we have a "Power Down Well" that can be turned off
  * when not needed anymore. We have 4 registers that can request the power well
@@ -4190,7 +5172,12 @@ void intel_init_pm(struct drm_device *dev)
 	if (I915_HAS_FBC(dev)) {
 		if (HAS_PCH_SPLIT(dev)) {
 			dev_priv->display.fbc_enabled = ironlake_fbc_enabled;
-			dev_priv->display.enable_fbc = ironlake_enable_fbc;
+			if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
+				dev_priv->display.enable_fbc =
+					gen7_enable_fbc;
+			else
+				dev_priv->display.enable_fbc =
+					ironlake_enable_fbc;
 			dev_priv->display.disable_fbc = ironlake_disable_fbc;
 		} else if (IS_GM45(dev)) {
 			dev_priv->display.fbc_enabled = g4x_fbc_enabled;
@@ -4242,10 +5229,10 @@ void intel_init_pm(struct drm_device *dev)
 			}
 			dev_priv->display.init_clock_gating = ivybridge_init_clock_gating;
 		} else if (IS_HASWELL(dev)) {
-			if (SNB_READ_WM0_LATENCY()) {
-				dev_priv->display.update_wm = sandybridge_update_wm;
-				dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm;
-				dev_priv->display.update_linetime_wm = haswell_update_linetime_wm;
+			if (I915_READ64(MCH_SSKPD)) {
+				dev_priv->display.update_wm = haswell_update_wm;
+				dev_priv->display.update_sprite_wm =
+					haswell_update_sprite_wm;
 			} else {
 				DRM_DEBUG_KMS("Failed to read display plane latency. "
 					      "Disable CxSR\n");
@@ -4340,6 +5327,7 @@ static void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
+	/* WaRsForcewakeWaitTC0:snb */
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
@@ -4371,6 +5359,7 @@ static void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv)
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
+	/* WaRsForcewakeWaitTC0:ivb,hsw */
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
@@ -4474,6 +5463,7 @@ static void vlv_force_wake_get(struct drm_i915_private *dev_priv)
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for media to ack forcewake request.\n");
 
+	/* WaRsForcewakeWaitTC0:vlv */
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
@@ -4568,55 +5558,58 @@ int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val)
 	return 0;
 }
 
-static int vlv_punit_rw(struct drm_i915_private *dev_priv, u8 opcode,
-			u8 addr, u32 *val)
+int vlv_gpu_freq(int ddr_freq, int val)
 {
-	u32 cmd, devfn, port, be, bar;
-
-	bar = 0;
-	be = 0xf;
-	port = IOSF_PORT_PUNIT;
-	devfn = PCI_DEVFN(2, 0);
+	int mult, base;
 
-	cmd = (devfn << IOSF_DEVFN_SHIFT) | (opcode << IOSF_OPCODE_SHIFT) |
-		(port << IOSF_PORT_SHIFT) | (be << IOSF_BYTE_ENABLES_SHIFT) |
-		(bar << IOSF_BAR_SHIFT);
-
-	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
-
-	if (I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) {
-		DRM_DEBUG_DRIVER("warning: pcode (%s) mailbox access failed\n",
-				 opcode == PUNIT_OPCODE_REG_READ ?
-				 "read" : "write");
-		return -EAGAIN;
+	switch (ddr_freq) {
+	case 800:
+		mult = 20;
+		base = 120;
+		break;
+	case 1066:
+		mult = 22;
+		base = 133;
+		break;
+	case 1333:
+		mult = 21;
+		base = 125;
+		break;
+	default:
+		return -1;
 	}
 
-	I915_WRITE(VLV_IOSF_ADDR, addr);
-	if (opcode == PUNIT_OPCODE_REG_WRITE)
-		I915_WRITE(VLV_IOSF_DATA, *val);
-	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
+	return ((val - 0xbd) * mult) + base;
+}
 
-	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0,
-		     500)) {
-		DRM_ERROR("timeout waiting for pcode %s (%d) to finish\n",
-			  opcode == PUNIT_OPCODE_REG_READ ? "read" : "write",
-			  addr);
-		return -ETIMEDOUT;
+int vlv_freq_opcode(int ddr_freq, int val)
+{
+	int mult, base;
+
+	switch (ddr_freq) {
+	case 800:
+		mult = 20;
+		base = 120;
+		break;
+	case 1066:
+		mult = 22;
+		base = 133;
+		break;
+	case 1333:
+		mult = 21;
+		base = 125;
+		break;
+	default:
+		return -1;
 	}
 
-	if (opcode == PUNIT_OPCODE_REG_READ)
-		*val = I915_READ(VLV_IOSF_DATA);
-	I915_WRITE(VLV_IOSF_DATA, 0);
+	val /= mult;
+	val -= base / mult;
+	val += 0xbd;
 
-	return 0;
-}
+	if (val > 0xea)
+		val = 0xea;
 
-int valleyview_punit_read(struct drm_i915_private *dev_priv, u8 addr, u32 *val)
-{
-	return vlv_punit_rw(dev_priv, PUNIT_OPCODE_REG_READ, addr, val);
+	return val;
 }
 
-int valleyview_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val)
-{
-	return vlv_punit_rw(dev_priv, PUNIT_OPCODE_REG_WRITE, addr, &val);
-}
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 1d5d613..e51ab55 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -280,6 +280,27 @@ gen7_render_ring_cs_stall_wa(struct intel_ring_buffer *ring)
 	return 0;
 }
 
+static int gen7_ring_fbc_flush(struct intel_ring_buffer *ring, u32 value)
+{
+	int ret;
+
+	if (!ring->fbc_dirty)
+		return 0;
+
+	ret = intel_ring_begin(ring, 4);
+	if (ret)
+		return ret;
+	intel_ring_emit(ring, MI_NOOP);
+	/* WaFbcNukeOn3DBlt:ivb/hsw */
+	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
+	intel_ring_emit(ring, MSG_FBC_REND_STATE);
+	intel_ring_emit(ring, value);
+	intel_ring_advance(ring);
+
+	ring->fbc_dirty = false;
+	return 0;
+}
+
 static int
 gen7_render_ring_flush(struct intel_ring_buffer *ring,
 		       u32 invalidate_domains, u32 flush_domains)
@@ -336,6 +357,9 @@ gen7_render_ring_flush(struct intel_ring_buffer *ring,
 	intel_ring_emit(ring, 0);
 	intel_ring_advance(ring);
 
+	if (flush_domains)
+		return gen7_ring_fbc_flush(ring, FBC_REND_NUKE);
+
 	return 0;
 }
 
@@ -429,6 +453,8 @@ static int init_ring_common(struct intel_ring_buffer *ring)
 		ring->last_retired_head = -1;
 	}
 
+	memset(&ring->hangcheck, 0, sizeof(ring->hangcheck));
+
 out:
 	if (HAS_FORCE_WAKE(dev))
 		gen6_gt_force_wake_put(dev_priv);
@@ -464,9 +490,11 @@ init_pipe_control(struct intel_ring_buffer *ring)
 		goto err_unref;
 
 	pc->gtt_offset = obj->gtt_offset;
-	pc->cpu_page =  kmap(sg_page(obj->pages->sgl));
-	if (pc->cpu_page == NULL)
+	pc->cpu_page = kmap(sg_page(obj->pages->sgl));
+	if (pc->cpu_page == NULL) {
+		ret = -ENOMEM;
 		goto err_unpin;
+	}
 
 	DRM_DEBUG_DRIVER("%s pipe control offset: 0x%08x\n",
 			 ring->name, pc->gtt_offset);
@@ -515,6 +543,8 @@ static int init_render_ring(struct intel_ring_buffer *ring)
 	/* We need to disable the AsyncFlip performance optimisations in order
 	 * to use MI_WAIT_FOR_EVENT within the CS. It should already be
 	 * programmed to '1' on all products.
+	 *
+	 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv
 	 */
 	if (INTEL_INFO(dev)->gen >= 6)
 		I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE));
@@ -556,7 +586,7 @@ static int init_render_ring(struct intel_ring_buffer *ring)
 		I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));
 
 	if (HAS_L3_GPU_CACHE(dev))
-		I915_WRITE_IMR(ring, ~GEN6_RENDER_L3_PARITY_ERROR);
+		I915_WRITE_IMR(ring, ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT);
 
 	return ret;
 }
@@ -578,9 +608,16 @@ static void
 update_mboxes(struct intel_ring_buffer *ring,
 	      u32 mmio_offset)
 {
+/* NB: In order to be able to do semaphore MBOX updates for varying number
+ * of rings, it's easiest if we round up each individual update to a
+ * multiple of 2 (since ring updates must always be a multiple of 2)
+ * even though the actual update only requires 3 dwords.
+ */
+#define MBOX_UPDATE_DWORDS 4
 	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
 	intel_ring_emit(ring, mmio_offset);
 	intel_ring_emit(ring, ring->outstanding_lazy_request);
+	intel_ring_emit(ring, MI_NOOP);
 }
 
 /**
@@ -595,19 +632,24 @@ update_mboxes(struct intel_ring_buffer *ring,
 static int
 gen6_add_request(struct intel_ring_buffer *ring)
 {
-	u32 mbox1_reg;
-	u32 mbox2_reg;
-	int ret;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *useless;
+	int i, ret;
 
-	ret = intel_ring_begin(ring, 10);
+	ret = intel_ring_begin(ring, ((I915_NUM_RINGS-1) *
+				      MBOX_UPDATE_DWORDS) +
+				      4);
 	if (ret)
 		return ret;
+#undef MBOX_UPDATE_DWORDS
 
-	mbox1_reg = ring->signal_mbox[0];
-	mbox2_reg = ring->signal_mbox[1];
+	for_each_ring(useless, dev_priv, i) {
+		u32 mbox_reg = ring->signal_mbox[i];
+		if (mbox_reg != GEN6_NOSYNC)
+			update_mboxes(ring, mbox_reg);
+	}
 
-	update_mboxes(ring, mbox1_reg);
-	update_mboxes(ring, mbox2_reg);
 	intel_ring_emit(ring, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(ring, I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
 	intel_ring_emit(ring, ring->outstanding_lazy_request);
@@ -779,7 +821,7 @@ gen5_ring_get_irq(struct intel_ring_buffer *ring)
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		dev_priv->gt_irq_mask &= ~ring->irq_enable_mask;
 		I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 		POSTING_READ(GTIMR);
@@ -797,7 +839,7 @@ gen5_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		dev_priv->gt_irq_mask |= ring->irq_enable_mask;
 		I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 		POSTING_READ(GTIMR);
@@ -816,7 +858,7 @@ i9xx_ring_get_irq(struct intel_ring_buffer *ring)
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		dev_priv->irq_mask &= ~ring->irq_enable_mask;
 		I915_WRITE(IMR, dev_priv->irq_mask);
 		POSTING_READ(IMR);
@@ -834,7 +876,7 @@ i9xx_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		dev_priv->irq_mask |= ring->irq_enable_mask;
 		I915_WRITE(IMR, dev_priv->irq_mask);
 		POSTING_READ(IMR);
@@ -853,7 +895,7 @@ i8xx_ring_get_irq(struct intel_ring_buffer *ring)
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		dev_priv->irq_mask &= ~ring->irq_enable_mask;
 		I915_WRITE16(IMR, dev_priv->irq_mask);
 		POSTING_READ16(IMR);
@@ -871,7 +913,7 @@ i8xx_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		dev_priv->irq_mask |= ring->irq_enable_mask;
 		I915_WRITE16(IMR, dev_priv->irq_mask);
 		POSTING_READ16(IMR);
@@ -899,6 +941,9 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring)
 		case VCS:
 			mmio = BSD_HWS_PGA_GEN7;
 			break;
+		case VECS:
+			mmio = VEBOX_HWS_PGA_GEN7;
+			break;
 		}
 	} else if (IS_GEN6(ring->dev)) {
 		mmio = RING_HWS_PGA_GEN6(ring->mmio_base);
@@ -961,10 +1006,11 @@ gen6_ring_get_irq(struct intel_ring_buffer *ring)
 	gen6_gt_force_wake_get(dev_priv);
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		if (HAS_L3_GPU_CACHE(dev) && ring->id == RCS)
-			I915_WRITE_IMR(ring, ~(ring->irq_enable_mask |
-						GEN6_RENDER_L3_PARITY_ERROR));
+			I915_WRITE_IMR(ring,
+				       ~(ring->irq_enable_mask |
+					 GT_RENDER_L3_PARITY_ERROR_INTERRUPT));
 		else
 			I915_WRITE_IMR(ring, ~ring->irq_enable_mask);
 		dev_priv->gt_irq_mask &= ~ring->irq_enable_mask;
@@ -984,9 +1030,10 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		if (HAS_L3_GPU_CACHE(dev) && ring->id == RCS)
-			I915_WRITE_IMR(ring, ~GEN6_RENDER_L3_PARITY_ERROR);
+			I915_WRITE_IMR(ring,
+				       ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT);
 		else
 			I915_WRITE_IMR(ring, ~0);
 		dev_priv->gt_irq_mask |= ring->irq_enable_mask;
@@ -998,6 +1045,48 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring)
 	gen6_gt_force_wake_put(dev_priv);
 }
 
+static bool
+hsw_vebox_get_irq(struct intel_ring_buffer *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	if (!dev->irq_enabled)
+		return false;
+
+	spin_lock_irqsave(&dev_priv->rps.lock, flags);
+	if (ring->irq_refcount.pm++ == 0) {
+		u32 pm_imr = I915_READ(GEN6_PMIMR);
+		I915_WRITE_IMR(ring, ~ring->irq_enable_mask);
+		I915_WRITE(GEN6_PMIMR, pm_imr & ~ring->irq_enable_mask);
+		POSTING_READ(GEN6_PMIMR);
+	}
+	spin_unlock_irqrestore(&dev_priv->rps.lock, flags);
+
+	return true;
+}
+
+static void
+hsw_vebox_put_irq(struct intel_ring_buffer *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	if (!dev->irq_enabled)
+		return;
+
+	spin_lock_irqsave(&dev_priv->rps.lock, flags);
+	if (--ring->irq_refcount.pm == 0) {
+		u32 pm_imr = I915_READ(GEN6_PMIMR);
+		I915_WRITE_IMR(ring, ~0);
+		I915_WRITE(GEN6_PMIMR, pm_imr | ring->irq_enable_mask);
+		POSTING_READ(GEN6_PMIMR);
+	}
+	spin_unlock_irqrestore(&dev_priv->rps.lock, flags);
+}
+
 static int
 i965_dispatch_execbuffer(struct intel_ring_buffer *ring,
 			 u32 offset, u32 length,
@@ -1423,7 +1512,7 @@ int intel_ring_idle(struct intel_ring_buffer *ring)
 
 	/* We need to add any requests required to flush the objects and ring */
 	if (ring->outstanding_lazy_request) {
-		ret = i915_add_request(ring, NULL, NULL);
+		ret = i915_add_request(ring, NULL);
 		if (ret)
 			return ret;
 	}
@@ -1500,6 +1589,7 @@ void intel_ring_init_seqno(struct intel_ring_buffer *ring, u32 seqno)
 	}
 
 	ring->set_seqno(ring, seqno);
+	ring->hangcheck.seqno = seqno;
 }
 
 void intel_ring_advance(struct intel_ring_buffer *ring)
@@ -1546,8 +1636,8 @@ static void gen6_bsd_ring_write_tail(struct intel_ring_buffer *ring,
 		   _MASKED_BIT_DISABLE(GEN6_BSD_SLEEP_MSG_DISABLE));
 }
 
-static int gen6_ring_flush(struct intel_ring_buffer *ring,
-			   u32 invalidate, u32 flush)
+static int gen6_bsd_ring_flush(struct intel_ring_buffer *ring,
+			       u32 invalidate, u32 flush)
 {
 	uint32_t cmd;
 	int ret;
@@ -1618,9 +1708,10 @@ gen6_ring_dispatch_execbuffer(struct intel_ring_buffer *ring,
 
 /* Blitter support (SandyBridge+) */
 
-static int blt_ring_flush(struct intel_ring_buffer *ring,
-			  u32 invalidate, u32 flush)
+static int gen6_ring_flush(struct intel_ring_buffer *ring,
+			   u32 invalidate, u32 flush)
 {
+	struct drm_device *dev = ring->dev;
 	uint32_t cmd;
 	int ret;
 
@@ -1643,6 +1734,10 @@ static int blt_ring_flush(struct intel_ring_buffer *ring,
 	intel_ring_emit(ring, 0);
 	intel_ring_emit(ring, MI_NOOP);
 	intel_ring_advance(ring);
+
+	if (IS_GEN7(dev) && flush)
+		return gen7_ring_fbc_flush(ring, FBC_REND_CACHE_CLEAN);
+
 	return 0;
 }
 
@@ -1662,15 +1757,18 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 			ring->flush = gen6_render_ring_flush;
 		ring->irq_get = gen6_ring_get_irq;
 		ring->irq_put = gen6_ring_put_irq;
-		ring->irq_enable_mask = GT_USER_INTERRUPT;
+		ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
 		ring->get_seqno = gen6_ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
 		ring->sync_to = gen6_ring_sync;
-		ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_RV;
-		ring->semaphore_register[2] = MI_SEMAPHORE_SYNC_RB;
-		ring->signal_mbox[0] = GEN6_VRSYNC;
-		ring->signal_mbox[1] = GEN6_BRSYNC;
+		ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_INVALID;
+		ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_RV;
+		ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_RB;
+		ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_RVE;
+		ring->signal_mbox[RCS] = GEN6_NOSYNC;
+		ring->signal_mbox[VCS] = GEN6_VRSYNC;
+		ring->signal_mbox[BCS] = GEN6_BRSYNC;
+		ring->signal_mbox[VECS] = GEN6_VERSYNC;
 	} else if (IS_GEN5(dev)) {
 		ring->add_request = pc_render_add_request;
 		ring->flush = gen4_render_ring_flush;
@@ -1678,7 +1776,8 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 		ring->set_seqno = pc_render_set_seqno;
 		ring->irq_get = gen5_ring_get_irq;
 		ring->irq_put = gen5_ring_put_irq;
-		ring->irq_enable_mask = GT_USER_INTERRUPT | GT_PIPE_NOTIFY;
+		ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT |
+					GT_RENDER_PIPECTL_NOTIFY_INTERRUPT;
 	} else {
 		ring->add_request = i9xx_add_request;
 		if (INTEL_INFO(dev)->gen < 4)
@@ -1816,20 +1915,23 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 		/* gen6 bsd needs a special wa for tail updates */
 		if (IS_GEN6(dev))
 			ring->write_tail = gen6_bsd_ring_write_tail;
-		ring->flush = gen6_ring_flush;
+		ring->flush = gen6_bsd_ring_flush;
 		ring->add_request = gen6_add_request;
 		ring->get_seqno = gen6_ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
-		ring->irq_enable_mask = GEN6_BSD_USER_INTERRUPT;
+		ring->irq_enable_mask = GT_BSD_USER_INTERRUPT;
 		ring->irq_get = gen6_ring_get_irq;
 		ring->irq_put = gen6_ring_put_irq;
 		ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
 		ring->sync_to = gen6_ring_sync;
-		ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_VR;
-		ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore_register[2] = MI_SEMAPHORE_SYNC_VB;
-		ring->signal_mbox[0] = GEN6_RVSYNC;
-		ring->signal_mbox[1] = GEN6_BVSYNC;
+		ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_VR;
+		ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_INVALID;
+		ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_VB;
+		ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_VVE;
+		ring->signal_mbox[RCS] = GEN6_RVSYNC;
+		ring->signal_mbox[VCS] = GEN6_NOSYNC;
+		ring->signal_mbox[BCS] = GEN6_BVSYNC;
+		ring->signal_mbox[VECS] = GEN6_VEVSYNC;
 	} else {
 		ring->mmio_base = BSD_RING_BASE;
 		ring->flush = bsd_ring_flush;
@@ -1837,7 +1939,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 		ring->get_seqno = ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
 		if (IS_GEN5(dev)) {
-			ring->irq_enable_mask = GT_BSD_USER_INTERRUPT;
+			ring->irq_enable_mask = ILK_BSD_USER_INTERRUPT;
 			ring->irq_get = gen5_ring_get_irq;
 			ring->irq_put = gen5_ring_put_irq;
 		} else {
@@ -1862,20 +1964,56 @@ int intel_init_blt_ring_buffer(struct drm_device *dev)
 
 	ring->mmio_base = BLT_RING_BASE;
 	ring->write_tail = ring_write_tail;
-	ring->flush = blt_ring_flush;
+	ring->flush = gen6_ring_flush;
 	ring->add_request = gen6_add_request;
 	ring->get_seqno = gen6_ring_get_seqno;
 	ring->set_seqno = ring_set_seqno;
-	ring->irq_enable_mask = GEN6_BLITTER_USER_INTERRUPT;
+	ring->irq_enable_mask = GT_BLT_USER_INTERRUPT;
 	ring->irq_get = gen6_ring_get_irq;
 	ring->irq_put = gen6_ring_put_irq;
 	ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
 	ring->sync_to = gen6_ring_sync;
-	ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_BR;
-	ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_BV;
-	ring->semaphore_register[2] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->signal_mbox[0] = GEN6_RBSYNC;
-	ring->signal_mbox[1] = GEN6_VBSYNC;
+	ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_BR;
+	ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_BV;
+	ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_INVALID;
+	ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_BVE;
+	ring->signal_mbox[RCS] = GEN6_RBSYNC;
+	ring->signal_mbox[VCS] = GEN6_VBSYNC;
+	ring->signal_mbox[BCS] = GEN6_NOSYNC;
+	ring->signal_mbox[VECS] = GEN6_VEBSYNC;
+	ring->init = init_ring_common;
+
+	return intel_init_ring_buffer(dev, ring);
+}
+
+int intel_init_vebox_ring_buffer(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *ring = &dev_priv->ring[VECS];
+
+	ring->name = "video enhancement ring";
+	ring->id = VECS;
+
+	ring->mmio_base = VEBOX_RING_BASE;
+	ring->write_tail = ring_write_tail;
+	ring->flush = gen6_ring_flush;
+	ring->add_request = gen6_add_request;
+	ring->get_seqno = gen6_ring_get_seqno;
+	ring->set_seqno = ring_set_seqno;
+	ring->irq_enable_mask = PM_VEBOX_USER_INTERRUPT |
+		PM_VEBOX_CS_ERROR_INTERRUPT;
+	ring->irq_get = hsw_vebox_get_irq;
+	ring->irq_put = hsw_vebox_put_irq;
+	ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
+	ring->sync_to = gen6_ring_sync;
+	ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_VER;
+	ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_VEV;
+	ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_VEB;
+	ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_INVALID;
+	ring->signal_mbox[RCS] = GEN6_RVESYNC;
+	ring->signal_mbox[VCS] = GEN6_VVESYNC;
+	ring->signal_mbox[BCS] = GEN6_BVESYNC;
+	ring->signal_mbox[VECS] = GEN6_NOSYNC;
 	ring->init = init_ring_common;
 
 	return intel_init_ring_buffer(dev, ring);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index d66208c..799f04c 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -37,14 +37,25 @@ struct  intel_hw_status_page {
 #define I915_READ_SYNC_0(ring) I915_READ(RING_SYNC_0((ring)->mmio_base))
 #define I915_READ_SYNC_1(ring) I915_READ(RING_SYNC_1((ring)->mmio_base))
 
+enum intel_ring_hangcheck_action { wait, active, kick, hung };
+
+struct intel_ring_hangcheck {
+	bool deadlock;
+	u32 seqno;
+	u32 acthd;
+	int score;
+	enum intel_ring_hangcheck_action action;
+};
+
 struct  intel_ring_buffer {
 	const char	*name;
 	enum intel_ring_id {
 		RCS = 0x0,
 		VCS,
 		BCS,
+		VECS,
 	} id;
-#define I915_NUM_RINGS 3
+#define I915_NUM_RINGS 4
 	u32		mmio_base;
 	void		__iomem *virtual_start;
 	struct		drm_device *dev;
@@ -67,7 +78,10 @@ struct  intel_ring_buffer {
 	 */
 	u32		last_retired_head;
 
-	u32		irq_refcount;		/* protected by dev_priv->irq_lock */
+	struct {
+		u32	gt; /*  protected by dev_priv->irq_lock */
+		u32	pm; /*  protected by dev_priv->rps.lock (sucks) */
+	} irq_refcount;
 	u32		irq_enable_mask;	/* bitmask to enable ring interrupt */
 	u32		trace_irq_seqno;
 	u32		sync_seqno[I915_NUM_RINGS-1];
@@ -102,8 +116,11 @@ struct  intel_ring_buffer {
 				   struct intel_ring_buffer *to,
 				   u32 seqno);
 
-	u32		semaphore_register[3]; /*our mbox written by others */
-	u32		signal_mbox[2]; /* mboxes this ring signals to */
+	/* our mbox written by others */
+	u32		semaphore_register[I915_NUM_RINGS];
+	/* mboxes this ring signals to */
+	u32		signal_mbox[I915_NUM_RINGS];
+
 	/**
 	 * List of objects currently involved in rendering from the
 	 * ringbuffer.
@@ -127,6 +144,7 @@ struct  intel_ring_buffer {
 	 */
 	u32 outstanding_lazy_request;
 	bool gpu_caches_dirty;
+	bool fbc_dirty;
 
 	wait_queue_head_t irq_queue;
 
@@ -135,7 +153,9 @@ struct  intel_ring_buffer {
 	 */
 	bool itlb_before_ctx_switch;
 	struct i915_hw_context *default_context;
-	struct drm_i915_gem_object *last_context_obj;
+	struct i915_hw_context *last_context;
+
+	struct intel_ring_hangcheck hangcheck;
 
 	void *private;
 };
@@ -224,6 +244,7 @@ int intel_ring_invalidate_all_caches(struct intel_ring_buffer *ring);
 int intel_init_render_ring_buffer(struct drm_device *dev);
 int intel_init_bsd_ring_buffer(struct drm_device *dev);
 int intel_init_blt_ring_buffer(struct drm_device *dev);
+int intel_init_vebox_ring_buffer(struct drm_device *dev);
 
 u32 intel_ring_get_active_head(struct intel_ring_buffer *ring);
 void intel_ring_setup_status_page(struct intel_ring_buffer *ring);
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index d4ea6c2..2628d56 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -80,7 +80,7 @@ struct intel_sdvo {
 
 	/*
 	 * Capabilities of the SDVO device returned by
-	 * i830_sdvo_get_capabilities()
+	 * intel_sdvo_get_capabilities()
 	 */
 	struct intel_sdvo_caps caps;
 
@@ -712,6 +712,13 @@ static bool intel_sdvo_set_timing(struct intel_sdvo *intel_sdvo, u8 cmd,
 		intel_sdvo_set_value(intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2));
 }
 
+static bool intel_sdvo_get_timing(struct intel_sdvo *intel_sdvo, u8 cmd,
+				  struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_get_value(intel_sdvo, cmd, &dtd->part1, sizeof(dtd->part1)) &&
+		intel_sdvo_get_value(intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2));
+}
+
 static bool intel_sdvo_set_input_timing(struct intel_sdvo *intel_sdvo,
 					 struct intel_sdvo_dtd *dtd)
 {
@@ -726,6 +733,13 @@ static bool intel_sdvo_set_output_timing(struct intel_sdvo *intel_sdvo,
 				     SDVO_CMD_SET_OUTPUT_TIMINGS_PART1, dtd);
 }
 
+static bool intel_sdvo_get_input_timing(struct intel_sdvo *intel_sdvo,
+					struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_get_timing(intel_sdvo,
+				     SDVO_CMD_GET_INPUT_TIMINGS_PART1, dtd);
+}
+
 static bool
 intel_sdvo_create_preferred_input_timing(struct intel_sdvo *intel_sdvo,
 					 uint16_t clock,
@@ -1041,6 +1055,32 @@ intel_sdvo_get_preferred_input_mode(struct intel_sdvo *intel_sdvo,
 	return true;
 }
 
+static void i9xx_adjust_sdvo_tv_clock(struct intel_crtc_config *pipe_config)
+{
+	unsigned dotclock = pipe_config->adjusted_mode.clock;
+	struct dpll *clock = &pipe_config->dpll;
+
+	/* SDVO TV has fixed PLL values depend on its clock range,
+	   this mirrors vbios setting. */
+	if (dotclock >= 100000 && dotclock < 140500) {
+		clock->p1 = 2;
+		clock->p2 = 10;
+		clock->n = 3;
+		clock->m1 = 16;
+		clock->m2 = 8;
+	} else if (dotclock >= 140500 && dotclock <= 200000) {
+		clock->p1 = 1;
+		clock->p2 = 10;
+		clock->n = 6;
+		clock->m1 = 12;
+		clock->m2 = 8;
+	} else {
+		WARN(1, "SDVO TV clock out of range: %i\n", dotclock);
+	}
+
+	pipe_config->clock_set = true;
+}
+
 static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 				      struct intel_crtc_config *pipe_config)
 {
@@ -1066,6 +1106,7 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 		(void) intel_sdvo_get_preferred_input_mode(intel_sdvo,
 							   mode,
 							   adjusted_mode);
+		pipe_config->sdvo_tv_clock = true;
 	} else if (intel_sdvo->is_lvds) {
 		if (!intel_sdvo_set_output_timings_from_mode(intel_sdvo,
 							     intel_sdvo->sdvo_lvds_fixed_mode))
@@ -1097,6 +1138,10 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 	if (intel_sdvo->color_range)
 		pipe_config->limited_color_range = true;
 
+	/* Clock computation needs to happen after pixel multiplier. */
+	if (intel_sdvo->is_tv)
+		i9xx_adjust_sdvo_tv_clock(pipe_config);
+
 	return true;
 }
 
@@ -1174,6 +1219,7 @@ static void intel_sdvo_mode_set(struct intel_encoder *intel_encoder)
 
 	switch (intel_crtc->config.pixel_multiplier) {
 	default:
+		WARN(1, "unknown pixel mutlipler specified\n");
 	case 1: rate = SDVO_CLOCK_RATE_MULT_1X; break;
 	case 2: rate = SDVO_CLOCK_RATE_MULT_2X; break;
 	case 4: rate = SDVO_CLOCK_RATE_MULT_4X; break;
@@ -1231,7 +1277,7 @@ static bool intel_sdvo_connector_get_hw_state(struct intel_connector *connector)
 	struct intel_sdvo_connector *intel_sdvo_connector =
 		to_intel_sdvo_connector(&connector->base);
 	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(&connector->base);
-	u16 active_outputs;
+	u16 active_outputs = 0;
 
 	intel_sdvo_get_active_outputs(intel_sdvo, &active_outputs);
 
@@ -1247,7 +1293,7 @@ static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
-	u16 active_outputs;
+	u16 active_outputs = 0;
 	u32 tmp;
 
 	tmp = I915_READ(intel_sdvo->sdvo_reg);
@@ -1264,6 +1310,74 @@ static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_sdvo_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
+	struct intel_sdvo_dtd dtd;
+	int encoder_pixel_multiplier = 0;
+	u32 flags = 0, sdvox;
+	u8 val;
+	bool ret;
+
+	ret = intel_sdvo_get_input_timing(intel_sdvo, &dtd);
+	if (!ret) {
+		/* Some sdvo encoders are not spec compliant and don't
+		 * implement the mandatory get_timings function. */
+		DRM_DEBUG_DRIVER("failed to retrieve SDVO DTD\n");
+		pipe_config->quirks |= PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS;
+	} else {
+		if (dtd.part2.dtd_flags & DTD_FLAG_HSYNC_POSITIVE)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (dtd.part2.dtd_flags & DTD_FLAG_VSYNC_POSITIVE)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	}
+
+	pipe_config->adjusted_mode.flags |= flags;
+
+	/*
+	 * pixel multiplier readout is tricky: Only on i915g/gm it is stored in
+	 * the sdvo port register, on all other platforms it is part of the dpll
+	 * state. Since the general pipe state readout happens before the
+	 * encoder->get_config we so already have a valid pixel multplier on all
+	 * other platfroms.
+	 */
+	if (IS_I915G(dev) || IS_I915GM(dev)) {
+		sdvox = I915_READ(intel_sdvo->sdvo_reg);
+		pipe_config->pixel_multiplier =
+			((sdvox & SDVO_PORT_MULTIPLY_MASK)
+			 >> SDVO_PORT_MULTIPLY_SHIFT) + 1;
+	}
+
+	/* Cross check the port pixel multiplier with the sdvo encoder state. */
+	intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_CLOCK_RATE_MULT, &val, 1);
+	switch (val) {
+	case SDVO_CLOCK_RATE_MULT_1X:
+		encoder_pixel_multiplier = 1;
+		break;
+	case SDVO_CLOCK_RATE_MULT_2X:
+		encoder_pixel_multiplier = 2;
+		break;
+	case SDVO_CLOCK_RATE_MULT_4X:
+		encoder_pixel_multiplier = 4;
+		break;
+	}
+
+	if(HAS_PCH_SPLIT(dev))
+		return; /* no pixel multiplier readout support yet */
+
+	WARN(encoder_pixel_multiplier != pipe_config->pixel_multiplier,
+	     "SDVO pixel multiplier mismatch, port: %i, encoder: %i\n",
+	     pipe_config->pixel_multiplier, encoder_pixel_multiplier);
+}
+
 static void intel_disable_sdvo(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
@@ -1344,6 +1458,7 @@ static void intel_enable_sdvo(struct intel_encoder *encoder)
 	intel_sdvo_set_active_outputs(intel_sdvo, intel_sdvo->attached_output);
 }
 
+/* Special dpms function to support cloning between dvo/sdvo/crt. */
 static void intel_sdvo_dpms(struct drm_connector *connector, int mode)
 {
 	struct drm_crtc *crtc;
@@ -1365,6 +1480,8 @@ static void intel_sdvo_dpms(struct drm_connector *connector, int mode)
 		return;
 	}
 
+	/* We set active outputs manually below in case pipe dpms doesn't change
+	 * due to cloning. */
 	if (mode != DRM_MODE_DPMS_ON) {
 		intel_sdvo_set_active_outputs(intel_sdvo, 0);
 		if (0)
@@ -1495,7 +1612,7 @@ intel_sdvo_get_analog_edid(struct drm_connector *connector)
 
 	return drm_get_edid(connector,
 			    intel_gmbus_get_adapter(dev_priv,
-						    dev_priv->crt_ddc_pin));
+						    dev_priv->vbt.crt_ddc_pin));
 }
 
 static enum drm_connector_status
@@ -1625,12 +1742,9 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 	if (ret == connector_status_connected) {
 		intel_sdvo->is_tv = false;
 		intel_sdvo->is_lvds = false;
-		intel_sdvo->base.needs_tv_clock = false;
 
-		if (response & SDVO_TV_MASK) {
+		if (response & SDVO_TV_MASK)
 			intel_sdvo->is_tv = true;
-			intel_sdvo->base.needs_tv_clock = true;
-		}
 		if (response & SDVO_LVDS_MASK)
 			intel_sdvo->is_lvds = intel_sdvo->sdvo_lvds_fixed_mode != NULL;
 	}
@@ -1772,21 +1886,12 @@ static void intel_sdvo_get_lvds_modes(struct drm_connector *connector)
 	struct drm_display_mode *newmode;
 
 	/*
-	 * Attempt to get the mode list from DDC.
-	 * Assume that the preferred modes are
-	 * arranged in priority order.
-	 */
-	intel_ddc_get_modes(connector, &intel_sdvo->ddc);
-
-	/*
 	 * Fetch modes from VBT. For SDVO prefer the VBT mode since some
-	 * SDVO->LVDS transcoders can't cope with the EDID mode. Since
-	 * drm_mode_probed_add adds the mode at the head of the list we add it
-	 * last.
+	 * SDVO->LVDS transcoders can't cope with the EDID mode.
 	 */
-	if (dev_priv->sdvo_lvds_vbt_mode != NULL) {
+	if (dev_priv->vbt.sdvo_lvds_vbt_mode != NULL) {
 		newmode = drm_mode_duplicate(connector->dev,
-					     dev_priv->sdvo_lvds_vbt_mode);
+					     dev_priv->vbt.sdvo_lvds_vbt_mode);
 		if (newmode != NULL) {
 			/* Guarantee the mode is preferred */
 			newmode->type = (DRM_MODE_TYPE_PREFERRED |
@@ -1795,6 +1900,13 @@ static void intel_sdvo_get_lvds_modes(struct drm_connector *connector)
 		}
 	}
 
+	/*
+	 * Attempt to get the mode list from DDC.
+	 * Assume that the preferred modes are
+	 * arranged in priority order.
+	 */
+	intel_ddc_get_modes(connector, &intel_sdvo->ddc);
+
 	list_for_each_entry(newmode, &connector->probed_modes, head) {
 		if (newmode->type & DRM_MODE_TYPE_PREFERRED) {
 			intel_sdvo->sdvo_lvds_fixed_mode =
@@ -2329,7 +2441,6 @@ intel_sdvo_tv_init(struct intel_sdvo *intel_sdvo, int type)
 	intel_sdvo_connector->output_flag = type;
 
 	intel_sdvo->is_tv = true;
-	intel_sdvo->base.needs_tv_clock = true;
 
 	intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo);
 
@@ -2417,7 +2528,6 @@ static bool
 intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags)
 {
 	intel_sdvo->is_tv = false;
-	intel_sdvo->base.needs_tv_clock = false;
 	intel_sdvo->is_lvds = false;
 
 	/* SDVO requires XXX1 function may not exist unless it has XXX0 function.*/
@@ -2751,7 +2861,6 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *intel_encoder;
 	struct intel_sdvo *intel_sdvo;
-	u32 hotplug_mask;
 	int i;
 	intel_sdvo = kzalloc(sizeof(struct intel_sdvo), GFP_KERNEL);
 	if (!intel_sdvo)
@@ -2780,23 +2889,12 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 		}
 	}
 
-	hotplug_mask = 0;
-	if (IS_G4X(dev)) {
-		hotplug_mask = intel_sdvo->is_sdvob ?
-			SDVOB_HOTPLUG_INT_STATUS_G4X : SDVOC_HOTPLUG_INT_STATUS_G4X;
-	} else if (IS_GEN4(dev)) {
-		hotplug_mask = intel_sdvo->is_sdvob ?
-			SDVOB_HOTPLUG_INT_STATUS_I965 : SDVOC_HOTPLUG_INT_STATUS_I965;
-	} else {
-		hotplug_mask = intel_sdvo->is_sdvob ?
-			SDVOB_HOTPLUG_INT_STATUS_I915 : SDVOC_HOTPLUG_INT_STATUS_I915;
-	}
-
 	intel_encoder->compute_config = intel_sdvo_compute_config;
 	intel_encoder->disable = intel_disable_sdvo;
 	intel_encoder->mode_set = intel_sdvo_mode_set;
 	intel_encoder->enable = intel_enable_sdvo;
 	intel_encoder->get_hw_state = intel_sdvo_get_hw_state;
+	intel_encoder->get_config = intel_sdvo_get_config;
 
 	/* In default case sdvo lvds is false */
 	if (!intel_sdvo_get_capabilities(intel_sdvo, &intel_sdvo->caps))
diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c
new file mode 100644
index 0000000..9a0e6c5
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_sideband.c
@@ -0,0 +1,177 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_drv.h"
+
+/* IOSF sideband */
+static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn,
+			   u32 port, u32 opcode, u32 addr, u32 *val)
+{
+	u32 cmd, be = 0xf, bar = 0;
+	bool is_read = (opcode == PUNIT_OPCODE_REG_READ ||
+			opcode == DPIO_OPCODE_REG_READ);
+
+	cmd = (devfn << IOSF_DEVFN_SHIFT) | (opcode << IOSF_OPCODE_SHIFT) |
+		(port << IOSF_PORT_SHIFT) | (be << IOSF_BYTE_ENABLES_SHIFT) |
+		(bar << IOSF_BAR_SHIFT);
+
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+
+	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0, 5)) {
+		DRM_DEBUG_DRIVER("IOSF sideband idle wait (%s) timed out\n",
+				 is_read ? "read" : "write");
+		return -EAGAIN;
+	}
+
+	I915_WRITE(VLV_IOSF_ADDR, addr);
+	if (!is_read)
+		I915_WRITE(VLV_IOSF_DATA, *val);
+	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
+
+	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0, 5)) {
+		DRM_DEBUG_DRIVER("IOSF sideband finish wait (%s) timed out\n",
+				 is_read ? "read" : "write");
+		return -ETIMEDOUT;
+	}
+
+	if (is_read)
+		*val = I915_READ(VLV_IOSF_DATA);
+	I915_WRITE(VLV_IOSF_DATA, 0);
+
+	return 0;
+}
+
+u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr)
+{
+	u32 val = 0;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_sideband_rw(dev_priv, PCI_DEVFN(2, 0), IOSF_PORT_PUNIT,
+			PUNIT_OPCODE_REG_READ, addr, &val);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	return val;
+}
+
+void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val)
+{
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_sideband_rw(dev_priv, PCI_DEVFN(2, 0), IOSF_PORT_PUNIT,
+			PUNIT_OPCODE_REG_WRITE, addr, &val);
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
+u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr)
+{
+	u32 val = 0;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_sideband_rw(dev_priv, PCI_DEVFN(2, 0), IOSF_PORT_NC,
+			PUNIT_OPCODE_REG_READ, addr, &val);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	return val;
+}
+
+u32 vlv_dpio_read(struct drm_i915_private *dev_priv, int reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(dev_priv, DPIO_DEVFN, IOSF_PORT_DPIO,
+			DPIO_OPCODE_REG_READ, reg, &val);
+
+	return val;
+}
+
+void vlv_dpio_write(struct drm_i915_private *dev_priv, int reg, u32 val)
+{
+	vlv_sideband_rw(dev_priv, DPIO_DEVFN, IOSF_PORT_DPIO,
+			DPIO_OPCODE_REG_WRITE, reg, &val);
+}
+
+/* SBI access */
+u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
+		   enum intel_sbi_destination destination)
+{
+	u32 value = 0;
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to become ready\n");
+		return 0;
+	}
+
+	I915_WRITE(SBI_ADDR, (reg << 16));
+
+	if (destination == SBI_ICLK)
+		value = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRRD;
+	else
+		value = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IORD;
+	I915_WRITE(SBI_CTL_STAT, value | SBI_BUSY);
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
+		return 0;
+	}
+
+	return I915_READ(SBI_DATA);
+}
+
+void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
+		     enum intel_sbi_destination destination)
+{
+	u32 tmp;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to become ready\n");
+		return;
+	}
+
+	I915_WRITE(SBI_ADDR, (reg << 16));
+	I915_WRITE(SBI_DATA, value);
+
+	if (destination == SBI_ICLK)
+		tmp = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRWR;
+	else
+		tmp = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IOWR;
+	I915_WRITE(SBI_CTL_STAT, SBI_BUSY | tmp);
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
+		return;
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index c7d25c5..1fa5612 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -32,6 +32,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_rect.h>
 #include "intel_drv.h"
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
@@ -113,7 +114,7 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_framebuffer *fb,
 	crtc_w--;
 	crtc_h--;
 
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size, true);
 
 	I915_WRITE(SPSTRIDE(pipe, plane), fb->pitches[0]);
 	I915_WRITE(SPPOS(pipe, plane), (crtc_y << 16) | crtc_x);
@@ -267,7 +268,7 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	crtc_w--;
 	crtc_h--;
 
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size, true);
 
 	/*
 	 * IVB workaround: must disable low power watermarks for at least
@@ -334,6 +335,8 @@ ivb_disable_plane(struct drm_plane *plane)
 
 	dev_priv->sprite_scaling_enabled &= ~(1 << pipe);
 
+	intel_update_sprite_watermarks(dev, pipe, 0, 0, false);
+
 	/* potentially re-enable LP watermarks */
 	if (scaling_was_enabled && !dev_priv->sprite_scaling_enabled)
 		intel_update_watermarks(dev);
@@ -452,7 +455,7 @@ ilk_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	crtc_w--;
 	crtc_h--;
 
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size, true);
 
 	dvsscale = 0;
 	if (IS_GEN5(dev) || crtc_w != src_w || crtc_h != src_h)
@@ -583,6 +586,20 @@ ilk_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key)
 		key->flags = I915_SET_COLORKEY_NONE;
 }
 
+static bool
+format_is_yuv(uint32_t format)
+{
+	switch (format) {
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_YVYU:
+		return true;
+	default:
+		return false;
+	}
+}
+
 static int
 intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		   struct drm_framebuffer *fb, int crtc_x, int crtc_y,
@@ -600,9 +617,29 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
 	int ret = 0;
-	int x = src_x >> 16, y = src_y >> 16;
-	int primary_w = crtc->mode.hdisplay, primary_h = crtc->mode.vdisplay;
 	bool disable_primary = false;
+	bool visible;
+	int hscale, vscale;
+	int max_scale, min_scale;
+	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+	struct drm_rect src = {
+		/* sample coordinates in 16.16 fixed point */
+		.x1 = src_x,
+		.x2 = src_x + src_w,
+		.y1 = src_y,
+		.y2 = src_y + src_h,
+	};
+	struct drm_rect dst = {
+		/* integer pixels */
+		.x1 = crtc_x,
+		.x2 = crtc_x + crtc_w,
+		.y1 = crtc_y,
+		.y2 = crtc_y + crtc_h,
+	};
+	const struct drm_rect clip = {
+		.x2 = crtc->mode.hdisplay,
+		.y2 = crtc->mode.vdisplay,
+	};
 
 	intel_fb = to_intel_framebuffer(fb);
 	obj = intel_fb->obj;
@@ -618,19 +655,23 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	intel_plane->src_w = src_w;
 	intel_plane->src_h = src_h;
 
-	src_w = src_w >> 16;
-	src_h = src_h >> 16;
-
 	/* Pipe must be running... */
-	if (!(I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_ENABLE))
+	if (!(I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_ENABLE)) {
+		DRM_DEBUG_KMS("Pipe disabled\n");
 		return -EINVAL;
+	}
 
-	if (crtc_x >= primary_w || crtc_y >= primary_h)
+	/* Don't modify another pipe's plane */
+	if (intel_plane->pipe != intel_crtc->pipe) {
+		DRM_DEBUG_KMS("Wrong plane <-> crtc mapping\n");
 		return -EINVAL;
+	}
 
-	/* Don't modify another pipe's plane */
-	if (intel_plane->pipe != intel_crtc->pipe)
+	/* FIXME check all gen limits */
+	if (fb->width < 3 || fb->height < 3 || fb->pitches[0] > 16384) {
+		DRM_DEBUG_KMS("Unsuitable framebuffer for plane\n");
 		return -EINVAL;
+	}
 
 	/* Sprite planes can be linear or x-tiled surfaces */
 	switch (obj->tiling_mode) {
@@ -638,55 +679,123 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		case I915_TILING_X:
 			break;
 		default:
+			DRM_DEBUG_KMS("Unsupported tiling mode\n");
 			return -EINVAL;
 	}
 
 	/*
-	 * Clamp the width & height into the visible area.  Note we don't
-	 * try to scale the source if part of the visible region is offscreen.
-	 * The caller must handle that by adjusting source offset and size.
+	 * FIXME the following code does a bunch of fuzzy adjustments to the
+	 * coordinates and sizes. We probably need some way to decide whether
+	 * more strict checking should be done instead.
 	 */
-	if ((crtc_x < 0) && ((crtc_x + crtc_w) > 0)) {
-		crtc_w += crtc_x;
-		crtc_x = 0;
-	}
-	if ((crtc_x + crtc_w) <= 0) /* Nothing to display */
-		goto out;
-	if ((crtc_x + crtc_w) > primary_w)
-		crtc_w = primary_w - crtc_x;
+	max_scale = intel_plane->max_downscale << 16;
+	min_scale = intel_plane->can_scale ? 1 : (1 << 16);
+
+	hscale = drm_rect_calc_hscale_relaxed(&src, &dst, min_scale, max_scale);
+	BUG_ON(hscale < 0);
+
+	vscale = drm_rect_calc_vscale_relaxed(&src, &dst, min_scale, max_scale);
+	BUG_ON(vscale < 0);
+
+	visible = drm_rect_clip_scaled(&src, &dst, &clip, hscale, vscale);
+
+	crtc_x = dst.x1;
+	crtc_y = dst.y1;
+	crtc_w = drm_rect_width(&dst);
+	crtc_h = drm_rect_height(&dst);
+
+	if (visible) {
+		/* check again in case clipping clamped the results */
+		hscale = drm_rect_calc_hscale(&src, &dst, min_scale, max_scale);
+		if (hscale < 0) {
+			DRM_DEBUG_KMS("Horizontal scaling factor out of limits\n");
+			drm_rect_debug_print(&src, true);
+			drm_rect_debug_print(&dst, false);
+
+			return hscale;
+		}
 
-	if ((crtc_y < 0) && ((crtc_y + crtc_h) > 0)) {
-		crtc_h += crtc_y;
-		crtc_y = 0;
+		vscale = drm_rect_calc_vscale(&src, &dst, min_scale, max_scale);
+		if (vscale < 0) {
+			DRM_DEBUG_KMS("Vertical scaling factor out of limits\n");
+			drm_rect_debug_print(&src, true);
+			drm_rect_debug_print(&dst, false);
+
+			return vscale;
+		}
+
+		/* Make the source viewport size an exact multiple of the scaling factors. */
+		drm_rect_adjust_size(&src,
+				     drm_rect_width(&dst) * hscale - drm_rect_width(&src),
+				     drm_rect_height(&dst) * vscale - drm_rect_height(&src));
+
+		/* sanity check to make sure the src viewport wasn't enlarged */
+		WARN_ON(src.x1 < (int) src_x ||
+			src.y1 < (int) src_y ||
+			src.x2 > (int) (src_x + src_w) ||
+			src.y2 > (int) (src_y + src_h));
+
+		/*
+		 * Hardware doesn't handle subpixel coordinates.
+		 * Adjust to (macro)pixel boundary, but be careful not to
+		 * increase the source viewport size, because that could
+		 * push the downscaling factor out of bounds.
+		 */
+		src_x = src.x1 >> 16;
+		src_w = drm_rect_width(&src) >> 16;
+		src_y = src.y1 >> 16;
+		src_h = drm_rect_height(&src) >> 16;
+
+		if (format_is_yuv(fb->pixel_format)) {
+			src_x &= ~1;
+			src_w &= ~1;
+
+			/*
+			 * Must keep src and dst the
+			 * same if we can't scale.
+			 */
+			if (!intel_plane->can_scale)
+				crtc_w &= ~1;
+
+			if (crtc_w == 0)
+				visible = false;
+		}
 	}
-	if ((crtc_y + crtc_h) <= 0) /* Nothing to display */
-		goto out;
-	if (crtc_y + crtc_h > primary_h)
-		crtc_h = primary_h - crtc_y;
 
-	if (!crtc_w || !crtc_h) /* Again, nothing to display */
-		goto out;
+	/* Check size restrictions when scaling */
+	if (visible && (src_w != crtc_w || src_h != crtc_h)) {
+		unsigned int width_bytes;
 
-	/*
-	 * We may not have a scaler, eg. HSW does not have it any more
-	 */
-	if (!intel_plane->can_scale && (crtc_w != src_w || crtc_h != src_h))
-		return -EINVAL;
+		WARN_ON(!intel_plane->can_scale);
 
-	/*
-	 * We can take a larger source and scale it down, but
-	 * only so much...  16x is the max on SNB.
-	 */
-	if (((src_w * src_h) / (crtc_w * crtc_h)) > intel_plane->max_downscale)
-		return -EINVAL;
+		/* FIXME interlacing min height is 6 */
+
+		if (crtc_w < 3 || crtc_h < 3)
+			visible = false;
+
+		if (src_w < 3 || src_h < 3)
+			visible = false;
+
+		width_bytes = ((src_x * pixel_size) & 63) + src_w * pixel_size;
+
+		if (src_w > 2048 || src_h > 2048 ||
+		    width_bytes > 4096 || fb->pitches[0] > 4096) {
+			DRM_DEBUG_KMS("Source dimensions exceed hardware limits\n");
+			return -EINVAL;
+		}
+	}
+
+	dst.x1 = crtc_x;
+	dst.x2 = crtc_x + crtc_w;
+	dst.y1 = crtc_y;
+	dst.y2 = crtc_y + crtc_h;
 
 	/*
 	 * If the sprite is completely covering the primary plane,
 	 * we can disable the primary and save power.
 	 */
-	if ((crtc_x == 0) && (crtc_y == 0) &&
-	    (crtc_w == primary_w) && (crtc_h == primary_h))
-		disable_primary = true;
+	disable_primary = drm_rect_equals(&dst, &clip);
+	WARN_ON(disable_primary && !visible);
 
 	mutex_lock(&dev->struct_mutex);
 
@@ -708,8 +817,12 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (!disable_primary)
 		intel_enable_primary(crtc);
 
-	intel_plane->update_plane(plane, fb, obj, crtc_x, crtc_y,
-				  crtc_w, crtc_h, x, y, src_w, src_h);
+	if (visible)
+		intel_plane->update_plane(plane, fb, obj,
+					  crtc_x, crtc_y, crtc_w, crtc_h,
+					  src_x, src_y, src_w, src_h);
+	else
+		intel_plane->disable_plane(plane);
 
 	if (disable_primary)
 		intel_disable_primary(crtc);
@@ -732,7 +845,6 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 out_unlock:
 	mutex_unlock(&dev->struct_mutex);
-out:
 	return ret;
 }
 
@@ -845,6 +957,14 @@ void intel_plane_restore(struct drm_plane *plane)
 			   intel_plane->src_w, intel_plane->src_h);
 }
 
+void intel_plane_disable(struct drm_plane *plane)
+{
+	if (!plane->crtc || !plane->fb)
+		return;
+
+	intel_disable_plane(plane);
+}
+
 static const struct drm_plane_funcs intel_plane_funcs = {
 	.update_plane = intel_update_plane,
 	.disable_plane = intel_disable_plane,
@@ -918,13 +1038,15 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 		break;
 
 	case 7:
-		if (IS_HASWELL(dev) || IS_VALLEYVIEW(dev))
-			intel_plane->can_scale = false;
-		else
+		if (IS_IVYBRIDGE(dev)) {
 			intel_plane->can_scale = true;
+			intel_plane->max_downscale = 2;
+		} else {
+			intel_plane->can_scale = false;
+			intel_plane->max_downscale = 1;
+		}
 
 		if (IS_VALLEYVIEW(dev)) {
-			intel_plane->max_downscale = 1;
 			intel_plane->update_plane = vlv_update_plane;
 			intel_plane->disable_plane = vlv_disable_plane;
 			intel_plane->update_colorkey = vlv_update_colorkey;
@@ -933,7 +1055,6 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 			plane_formats = vlv_plane_formats;
 			num_plane_formats = ARRAY_SIZE(vlv_plane_formats);
 		} else {
-			intel_plane->max_downscale = 2;
 			intel_plane->update_plane = ivb_update_plane;
 			intel_plane->disable_plane = ivb_disable_plane;
 			intel_plane->update_colorkey = ivb_update_colorkey;
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index b945bc5..39debd8 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -914,9 +914,6 @@ intel_tv_compute_config(struct intel_encoder *encoder,
 	if (!tv_mode)
 		return false;
 
-	if (intel_encoder_check_is_cloned(&intel_tv->base))
-		return false;
-
 	pipe_config->adjusted_mode.clock = tv_mode->clock;
 	DRM_DEBUG_KMS("forcing bpc to 8 for TV\n");
 	pipe_config->pipe_bpp = 8*3;
@@ -1521,12 +1518,12 @@ static int tv_is_present_in_vbt(struct drm_device *dev)
 	struct child_device_config *p_child;
 	int i, ret;
 
-	if (!dev_priv->child_dev_num)
+	if (!dev_priv->vbt.child_dev_num)
 		return 1;
 
 	ret = 0;
-	for (i = 0; i < dev_priv->child_dev_num; i++) {
-		p_child = dev_priv->child_dev + i;
+	for (i = 0; i < dev_priv->vbt.child_dev_num; i++) {
+		p_child = dev_priv->vbt.child_dev + i;
 		/*
 		 * If the device type is not TV, continue.
 		 */
@@ -1564,7 +1561,7 @@ intel_tv_init(struct drm_device *dev)
 		return;
 	}
 	/* Even if we have an encoder we may not have a connector */
-	if (!dev_priv->int_tv_support)
+	if (!dev_priv->vbt.int_tv_support)
 		return;
 
 	/*
diff --git a/drivers/gpu/drm/mgag200/Makefile b/drivers/gpu/drm/mgag200/Makefile
index 7db592e..a9a0300 100644
--- a/drivers/gpu/drm/mgag200/Makefile
+++ b/drivers/gpu/drm/mgag200/Makefile
@@ -1,5 +1,5 @@
 ccflags-y := -Iinclude/drm
-mgag200-y   := mgag200_main.o mgag200_mode.o \
+mgag200-y   := mgag200_main.o mgag200_mode.o mgag200_cursor.o \
 	mgag200_drv.o mgag200_fb.o mgag200_i2c.o mgag200_ttm.o
 
 obj-$(CONFIG_DRM_MGAG200) += mgag200.o
diff --git a/drivers/gpu/drm/mgag200/mgag200_cursor.c b/drivers/gpu/drm/mgag200/mgag200_cursor.c
new file mode 100644
index 0000000..801731a
--- /dev/null
+++ b/drivers/gpu/drm/mgag200/mgag200_cursor.c
@@ -0,0 +1,275 @@
+/*
+ * Copyright 2013 Matrox Graphics
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License version 2. See the file COPYING in the main
+ * directory of this archive for more details.
+ *
+ * Author: Christopher Harvey <charvey@matrox.com>
+ */
+
+#include <drm/drmP.h>
+#include "mgag200_drv.h"
+
+static bool warn_transparent = true;
+static bool warn_palette = true;
+
+/*
+  Hide the cursor off screen. We can't disable the cursor hardware because it
+  takes too long to re-activate and causes momentary corruption
+*/
+static void mga_hide_cursor(struct mga_device *mdev)
+{
+	WREG8(MGA_CURPOSXL, 0);
+	WREG8(MGA_CURPOSXH, 0);
+	mgag200_bo_unpin(mdev->cursor.pixels_1);
+	mgag200_bo_unpin(mdev->cursor.pixels_2);
+}
+
+int mga_crtc_cursor_set(struct drm_crtc *crtc,
+			struct drm_file *file_priv,
+			uint32_t handle,
+			uint32_t width,
+			uint32_t height)
+{
+	struct drm_device *dev = (struct drm_device *)file_priv->minor->dev;
+	struct mga_device *mdev = (struct mga_device *)dev->dev_private;
+	struct mgag200_bo *pixels_1 = mdev->cursor.pixels_1;
+	struct mgag200_bo *pixels_2 = mdev->cursor.pixels_2;
+	struct mgag200_bo *pixels_current = mdev->cursor.pixels_current;
+	struct mgag200_bo *pixels_prev = mdev->cursor.pixels_prev;
+	struct drm_gem_object *obj;
+	struct mgag200_bo *bo = NULL;
+	int ret = 0;
+	unsigned int i, row, col;
+	uint32_t colour_set[16];
+	uint32_t *next_space = &colour_set[0];
+	uint32_t *palette_iter;
+	uint32_t this_colour;
+	bool found = false;
+	int colour_count = 0;
+	u64 gpu_addr;
+	u8 reg_index;
+	u8 this_row[48];
+
+	if (!pixels_1 || !pixels_2) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		return -ENOTSUPP; /* Didn't allocate space for cursors */
+	}
+
+	if ((width != 64 || height != 64) && handle) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		return -EINVAL;
+	}
+
+	BUG_ON(pixels_1 != pixels_current && pixels_1 != pixels_prev);
+	BUG_ON(pixels_2 != pixels_current && pixels_2 != pixels_prev);
+	BUG_ON(pixels_current == pixels_prev);
+
+	ret = mgag200_bo_reserve(pixels_1, true);
+	if (ret) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		return ret;
+	}
+	ret = mgag200_bo_reserve(pixels_2, true);
+	if (ret) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		mgag200_bo_unreserve(pixels_1);
+		return ret;
+	}
+
+	if (!handle) {
+		mga_hide_cursor(mdev);
+		ret = 0;
+		goto out1;
+	}
+
+	/* Move cursor buffers into VRAM if they aren't already */
+	if (!pixels_1->pin_count) {
+		ret = mgag200_bo_pin(pixels_1, TTM_PL_FLAG_VRAM,
+				     &mdev->cursor.pixels_1_gpu_addr);
+		if (ret)
+			goto out1;
+	}
+	if (!pixels_2->pin_count) {
+		ret = mgag200_bo_pin(pixels_2, TTM_PL_FLAG_VRAM,
+				     &mdev->cursor.pixels_2_gpu_addr);
+		if (ret) {
+			mgag200_bo_unpin(pixels_1);
+			goto out1;
+		}
+	}
+
+	mutex_lock(&dev->struct_mutex);
+	obj = drm_gem_object_lookup(dev, file_priv, handle);
+	if (!obj) {
+		mutex_unlock(&dev->struct_mutex);
+		ret = -ENOENT;
+		goto out1;
+	}
+	drm_gem_object_unreference(obj);
+	mutex_unlock(&dev->struct_mutex);
+
+	bo = gem_to_mga_bo(obj);
+	ret = mgag200_bo_reserve(bo, true);
+	if (ret) {
+		dev_err(&dev->pdev->dev, "failed to reserve user bo\n");
+		goto out1;
+	}
+	if (!bo->kmap.virtual) {
+		ret = ttm_bo_kmap(&bo->bo, 0, bo->bo.num_pages, &bo->kmap);
+		if (ret) {
+			dev_err(&dev->pdev->dev, "failed to kmap user buffer updates\n");
+			goto out2;
+		}
+	}
+
+	memset(&colour_set[0], 0, sizeof(uint32_t)*16);
+	/* width*height*4 = 16384 */
+	for (i = 0; i < 16384; i += 4) {
+		this_colour = ioread32(bo->kmap.virtual + i);
+		/* No transparency */
+		if (this_colour>>24 != 0xff &&
+			this_colour>>24 != 0x0) {
+			if (warn_transparent) {
+				dev_info(&dev->pdev->dev, "Video card doesn't support cursors with partial transparency.\n");
+				dev_info(&dev->pdev->dev, "Not enabling hardware cursor.\n");
+				warn_transparent = false; /* Only tell the user once. */
+			}
+			ret = -EINVAL;
+			goto out3;
+		}
+		/* Don't need to store transparent pixels as colours */
+		if (this_colour>>24 == 0x0)
+			continue;
+		found = false;
+		for (palette_iter = &colour_set[0]; palette_iter != next_space; palette_iter++) {
+			if (*palette_iter == this_colour) {
+				found = true;
+				break;
+			}
+		}
+		if (found)
+			continue;
+		/* We only support 4bit paletted cursors */
+		if (colour_count >= 16) {
+			if (warn_palette) {
+				dev_info(&dev->pdev->dev, "Video card only supports cursors with up to 16 colours.\n");
+				dev_info(&dev->pdev->dev, "Not enabling hardware cursor.\n");
+				warn_palette = false; /* Only tell the user once. */
+			}
+			ret = -EINVAL;
+			goto out3;
+		}
+		*next_space = this_colour;
+		next_space++;
+		colour_count++;
+	}
+
+	/* Program colours from cursor icon into palette */
+	for (i = 0; i < colour_count; i++) {
+		if (i <= 2)
+			reg_index = 0x8 + i*0x4;
+		else
+			reg_index = 0x60 + i*0x3;
+		WREG_DAC(reg_index, colour_set[i] & 0xff);
+		WREG_DAC(reg_index+1, colour_set[i]>>8 & 0xff);
+		WREG_DAC(reg_index+2, colour_set[i]>>16 & 0xff);
+		BUG_ON((colour_set[i]>>24 & 0xff) != 0xff);
+	}
+
+	/* Map up-coming buffer to write colour indices */
+	if (!pixels_prev->kmap.virtual) {
+		ret = ttm_bo_kmap(&pixels_prev->bo, 0,
+				  pixels_prev->bo.num_pages,
+				  &pixels_prev->kmap);
+		if (ret) {
+			dev_err(&dev->pdev->dev, "failed to kmap cursor updates\n");
+			goto out3;
+		}
+	}
+
+	/* now write colour indices into hardware cursor buffer */
+	for (row = 0; row < 64; row++) {
+		memset(&this_row[0], 0, 48);
+		for (col = 0; col < 64; col++) {
+			this_colour = ioread32(bo->kmap.virtual + 4*(col + 64*row));
+			/* write transparent pixels */
+			if (this_colour>>24 == 0x0) {
+				this_row[47 - col/8] |= 0x80>>(col%8);
+				continue;
+			}
+
+			/* write colour index here */
+			for (i = 0; i < colour_count; i++) {
+				if (colour_set[i] == this_colour) {
+					if (col % 2)
+						this_row[col/2] |= i<<4;
+					else
+						this_row[col/2] |= i;
+					break;
+				}
+			}
+		}
+		memcpy_toio(pixels_prev->kmap.virtual + row*48, &this_row[0], 48);
+	}
+
+	/* Program gpu address of cursor buffer */
+	if (pixels_prev == pixels_1)
+		gpu_addr = mdev->cursor.pixels_1_gpu_addr;
+	else
+		gpu_addr = mdev->cursor.pixels_2_gpu_addr;
+	WREG_DAC(MGA1064_CURSOR_BASE_ADR_LOW, (u8)((gpu_addr>>10) & 0xff));
+	WREG_DAC(MGA1064_CURSOR_BASE_ADR_HI, (u8)((gpu_addr>>18) & 0x3f));
+
+	/* Adjust cursor control register to turn on the cursor */
+	WREG_DAC(MGA1064_CURSOR_CTL, 4); /* 16-colour palletized cursor mode */
+
+	/* Now swap internal buffer pointers */
+	if (mdev->cursor.pixels_1 == mdev->cursor.pixels_prev) {
+		mdev->cursor.pixels_prev = mdev->cursor.pixels_2;
+		mdev->cursor.pixels_current = mdev->cursor.pixels_1;
+	} else if (mdev->cursor.pixels_1 == mdev->cursor.pixels_current) {
+		mdev->cursor.pixels_prev = mdev->cursor.pixels_1;
+		mdev->cursor.pixels_current = mdev->cursor.pixels_2;
+	} else {
+		BUG();
+	}
+	ret = 0;
+
+	ttm_bo_kunmap(&pixels_prev->kmap);
+ out3:
+	ttm_bo_kunmap(&bo->kmap);
+ out2:
+	mgag200_bo_unreserve(bo);
+ out1:
+	if (ret)
+		mga_hide_cursor(mdev);
+	mgag200_bo_unreserve(pixels_1);
+	mgag200_bo_unreserve(pixels_2);
+	return ret;
+}
+
+int mga_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
+{
+	struct mga_device *mdev = (struct mga_device *)crtc->dev->dev_private;
+	/* Our origin is at (64,64) */
+	x += 64;
+	y += 64;
+
+	BUG_ON(x <= 0);
+	BUG_ON(y <= 0);
+	BUG_ON(x & ~0xffff);
+	BUG_ON(y & ~0xffff);
+
+	WREG8(MGA_CURPOSXL, x & 0xff);
+	WREG8(MGA_CURPOSXH, (x>>8) & 0xff);
+
+	WREG8(MGA_CURPOSYL, y & 0xff);
+	WREG8(MGA_CURPOSYH, (y>>8) & 0xff);
+	return 0;
+}
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h
index bf29b2f..12e2499 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.h
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
@@ -149,6 +149,21 @@ struct mga_connector {
 	struct mga_i2c_chan *i2c;
 };
 
+struct mga_cursor {
+	/*
+	   We have to have 2 buffers for the cursor to avoid occasional
+	   corruption while switching cursor icons.
+	   If either of these is NULL, then don't do hardware cursors, and
+	   fall back to software.
+	*/
+	struct mgag200_bo *pixels_1;
+	struct mgag200_bo *pixels_2;
+	u64 pixels_1_gpu_addr, pixels_2_gpu_addr;
+	/* The currently displayed icon, this points to one of pixels_1, or pixels_2 */
+	struct mgag200_bo *pixels_current;
+	/* The previously displayed icon */
+	struct mgag200_bo *pixels_prev;
+};
 
 struct mga_mc {
 	resource_size_t			vram_size;
@@ -181,6 +196,7 @@ struct mga_device {
 	struct mga_mode_info		mode_info;
 
 	struct mga_fbdev *mfbdev;
+	struct mga_cursor cursor;
 
 	bool				suspended;
 	int				num_crtc;
@@ -198,7 +214,8 @@ struct mga_device {
 		struct ttm_bo_device bdev;
 	} ttm;
 
-	u32 reg_1e24; /* SE model number */
+	/* SE model number stored in reg 0x1e24 */
+	u32 unique_rev_id;
 };
 
 
@@ -263,8 +280,24 @@ void mgag200_i2c_destroy(struct mga_i2c_chan *i2c);
 #define DRM_FILE_PAGE_OFFSET (0x100000000ULL >> PAGE_SHIFT)
 void mgag200_ttm_placement(struct mgag200_bo *bo, int domain);
 
-int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait);
-void mgag200_bo_unreserve(struct mgag200_bo *bo);
+static inline int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait)
+{
+	int ret;
+
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	if (ret) {
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p\n", bo);
+		return ret;
+	}
+	return 0;
+}
+
+static inline void mgag200_bo_unreserve(struct mgag200_bo *bo)
+{
+	ttm_bo_unreserve(&bo->bo);
+}
+
 int mgag200_bo_create(struct drm_device *dev, int size, int align,
 		      uint32_t flags, struct mgag200_bo **pastbo);
 int mgag200_mm_init(struct mga_device *mdev);
@@ -273,4 +306,9 @@ int mgag200_mmap(struct file *filp, struct vm_area_struct *vma);
 int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr);
 int mgag200_bo_unpin(struct mgag200_bo *bo);
 int mgag200_bo_push_sysram(struct mgag200_bo *bo);
+			   /* mgag200_cursor.c */
+int mga_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
+						uint32_t handle, uint32_t width, uint32_t height);
+int mga_crtc_cursor_move(struct drm_crtc *crtc, int x, int y);
+
 #endif				/* __MGAG200_DRV_H__ */
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 5da824c..964f58c 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -27,7 +27,7 @@ static void mga_dirty_update(struct mga_fbdev *mfbdev,
 	struct mgag200_bo *bo;
 	int src_offset, dst_offset;
 	int bpp = (mfbdev->mfb.base.bits_per_pixel + 7)/8;
-	int ret;
+	int ret = -EBUSY;
 	bool unmap = false;
 	bool store_for_later = false;
 	int x2, y2;
@@ -41,7 +41,8 @@ static void mga_dirty_update(struct mga_fbdev *mfbdev,
 	 * then the BO is being moved and we should
 	 * store up the damage until later.
 	 */
-	ret = mgag200_bo_reserve(bo, true);
+	if (!in_interrupt())
+		ret = mgag200_bo_reserve(bo, true);
 	if (ret) {
 		if (ret != -EBUSY)
 			return;
diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c
index 9905923..9fa5685 100644
--- a/drivers/gpu/drm/mgag200/mgag200_main.c
+++ b/drivers/gpu/drm/mgag200/mgag200_main.c
@@ -176,7 +176,7 @@ static int mgag200_device_init(struct drm_device *dev,
 
 	/* stash G200 SE model number for later use */
 	if (IS_G200_SE(mdev))
-		mdev->reg_1e24 = RREG32(0x1e24);
+		mdev->unique_rev_id = RREG32(0x1e24);
 
 	ret = mga_vram_init(mdev);
 	if (ret)
@@ -209,7 +209,7 @@ int mgag200_driver_load(struct drm_device *dev, unsigned long flags)
 	r = mgag200_device_init(dev, flags);
 	if (r) {
 		dev_err(&dev->pdev->dev, "Fatal error during GPU init: %d\n", r);
-		goto out;
+		return r;
 	}
 	r = mgag200_mm_init(mdev);
 	if (r)
@@ -221,8 +221,27 @@ int mgag200_driver_load(struct drm_device *dev, unsigned long flags)
 	dev->mode_config.prefer_shadow = 1;
 
 	r = mgag200_modeset_init(mdev);
-	if (r)
+	if (r) {
 		dev_err(&dev->pdev->dev, "Fatal error during modeset init: %d\n", r);
+		goto out;
+	}
+
+	/* Make small buffers to store a hardware cursor (double buffered icon updates) */
+	mgag200_bo_create(dev, roundup(48*64, PAGE_SIZE), 0, 0,
+					  &mdev->cursor.pixels_1);
+	mgag200_bo_create(dev, roundup(48*64, PAGE_SIZE), 0, 0,
+					  &mdev->cursor.pixels_2);
+	if (!mdev->cursor.pixels_2 || !mdev->cursor.pixels_1)
+		goto cursor_nospace;
+	mdev->cursor.pixels_current = mdev->cursor.pixels_1;
+	mdev->cursor.pixels_prev = mdev->cursor.pixels_2;
+	goto cursor_done;
+ cursor_nospace:
+	mdev->cursor.pixels_1 = NULL;
+	mdev->cursor.pixels_2 = NULL;
+	dev_warn(&dev->pdev->dev, "Could not allocate space for cursors. Not doing hardware cursors.\n");
+ cursor_done:
+
 out:
 	if (r)
 		mgag200_driver_unload(dev);
diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c
index ee66bad..251784a 100644
--- a/drivers/gpu/drm/mgag200/mgag200_mode.c
+++ b/drivers/gpu/drm/mgag200/mgag200_mode.c
@@ -1008,7 +1008,7 @@ static int mga_crtc_mode_set(struct drm_crtc *crtc,
 
 
 	if (IS_G200_SE(mdev)) {
-		if (mdev->reg_1e24 >= 0x02) {
+		if (mdev->unique_rev_id >= 0x02) {
 			u8 hi_pri_lvl;
 			u32 bpp;
 			u32 mb;
@@ -1038,7 +1038,7 @@ static int mga_crtc_mode_set(struct drm_crtc *crtc,
 			WREG8(MGAREG_CRTCEXT_DATA, hi_pri_lvl);
 		} else {
 			WREG8(MGAREG_CRTCEXT_INDEX, 0x06);
-			if (mdev->reg_1e24 >= 0x01)
+			if (mdev->unique_rev_id >= 0x01)
 				WREG8(MGAREG_CRTCEXT_DATA, 0x03);
 			else
 				WREG8(MGAREG_CRTCEXT_DATA, 0x04);
@@ -1253,6 +1253,8 @@ static void mga_crtc_destroy(struct drm_crtc *crtc)
 
 /* These provide the minimum set of functions required to handle a CRTC */
 static const struct drm_crtc_funcs mga_crtc_funcs = {
+	.cursor_set = mga_crtc_cursor_set,
+	.cursor_move = mga_crtc_cursor_move,
 	.gamma_set = mga_crtc_gamma_set,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = mga_crtc_destroy,
@@ -1410,6 +1412,32 @@ static int mga_vga_get_modes(struct drm_connector *connector)
 	return ret;
 }
 
+static uint32_t mga_vga_calculate_mode_bandwidth(struct drm_display_mode *mode,
+							int bits_per_pixel)
+{
+	uint32_t total_area, divisor;
+	int64_t active_area, pixels_per_second, bandwidth;
+	uint64_t bytes_per_pixel = (bits_per_pixel + 7) / 8;
+
+	divisor = 1024;
+
+	if (!mode->htotal || !mode->vtotal || !mode->clock)
+		return 0;
+
+	active_area = mode->hdisplay * mode->vdisplay;
+	total_area = mode->htotal * mode->vtotal;
+
+	pixels_per_second = active_area * mode->clock * 1000;
+	do_div(pixels_per_second, total_area);
+
+	bandwidth = pixels_per_second * bytes_per_pixel * 100;
+	do_div(bandwidth, divisor);
+
+	return (uint32_t)(bandwidth);
+}
+
+#define MODE_BANDWIDTH	MODE_BAD
+
 static int mga_vga_mode_valid(struct drm_connector *connector,
 				 struct drm_display_mode *mode)
 {
@@ -1421,7 +1449,45 @@ static int mga_vga_mode_valid(struct drm_connector *connector,
 	int bpp = 32;
 	int i = 0;
 
-	/* FIXME: Add bandwidth and g200se limitations */
+	if (IS_G200_SE(mdev)) {
+		if (mdev->unique_rev_id == 0x01) {
+			if (mode->hdisplay > 1600)
+				return MODE_VIRTUAL_X;
+			if (mode->vdisplay > 1200)
+				return MODE_VIRTUAL_Y;
+			if (mga_vga_calculate_mode_bandwidth(mode, bpp)
+				> (24400 * 1024))
+				return MODE_BANDWIDTH;
+		} else if (mdev->unique_rev_id >= 0x02) {
+			if (mode->hdisplay > 1920)
+				return MODE_VIRTUAL_X;
+			if (mode->vdisplay > 1200)
+				return MODE_VIRTUAL_Y;
+			if (mga_vga_calculate_mode_bandwidth(mode, bpp)
+				> (30100 * 1024))
+				return MODE_BANDWIDTH;
+		}
+	} else if (mdev->type == G200_WB) {
+		if (mode->hdisplay > 1280)
+			return MODE_VIRTUAL_X;
+		if (mode->vdisplay > 1024)
+			return MODE_VIRTUAL_Y;
+		if (mga_vga_calculate_mode_bandwidth(mode,
+			bpp > (31877 * 1024)))
+			return MODE_BANDWIDTH;
+	} else if (mdev->type == G200_EV &&
+		(mga_vga_calculate_mode_bandwidth(mode, bpp)
+			> (32700 * 1024))) {
+		return MODE_BANDWIDTH;
+	} else if (mode->type == G200_EH &&
+		(mga_vga_calculate_mode_bandwidth(mode, bpp)
+			> (37500 * 1024))) {
+		return MODE_BANDWIDTH;
+	} else if (mode->type == G200_ER &&
+		(mga_vga_calculate_mode_bandwidth(mode,
+			bpp) > (55000 * 1024))) {
+		return MODE_BANDWIDTH;
+	}
 
 	if (mode->crtc_hdisplay > 2048 || mode->crtc_hsync_start > 4096 ||
 	    mode->crtc_hsync_end > 4096 || mode->crtc_htotal > 4096 ||
diff --git a/drivers/gpu/drm/mgag200/mgag200_reg.h b/drivers/gpu/drm/mgag200/mgag200_reg.h
index fb24d86..3ae442a6 100644
--- a/drivers/gpu/drm/mgag200/mgag200_reg.h
+++ b/drivers/gpu/drm/mgag200/mgag200_reg.h
@@ -235,7 +235,11 @@
 #define MGAREG_CRTCEXT_INDEX	0x1fde
 #define MGAREG_CRTCEXT_DATA	0x1fdf
 
-
+/* Cursor X and Y position */
+#define MGA_CURPOSXL 0x3c0c
+#define MGA_CURPOSXH 0x3c0d
+#define MGA_CURPOSYL 0x3c0e
+#define MGA_CURPOSYH 0x3c0f
 
 /* MGA bits for registers PCI_OPTION_REG */
 #define MGA1064_OPT_SYS_CLK_PCI   		( 0x00 << 0 )
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 401c989..3acb2b0 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -270,26 +270,20 @@ int mgag200_mm_init(struct mga_device *mdev)
 		return ret;
 	}
 
-	mdev->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 0),
-				    pci_resource_len(dev->pdev, 0),
-				    DRM_MTRR_WC);
+	mdev->fb_mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 0),
+					 pci_resource_len(dev->pdev, 0));
 
 	return 0;
 }
 
 void mgag200_mm_fini(struct mga_device *mdev)
 {
-	struct drm_device *dev = mdev->dev;
 	ttm_bo_device_release(&mdev->ttm.bdev);
 
 	mgag200_ttm_global_release(mdev);
 
-	if (mdev->fb_mtrr >= 0) {
-		drm_mtrr_del(mdev->fb_mtrr,
-			     pci_resource_start(dev->pdev, 0),
-			     pci_resource_len(dev->pdev, 0), DRM_MTRR_WC);
-		mdev->fb_mtrr = -1;
-	}
+	arch_phys_wc_del(mdev->fb_mtrr);
+	mdev->fb_mtrr = 0;
 }
 
 void mgag200_ttm_placement(struct mgag200_bo *bo, int domain)
@@ -309,24 +303,6 @@ void mgag200_ttm_placement(struct mgag200_bo *bo, int domain)
 	bo->placement.num_busy_placement = c;
 }
 
-int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
-	if (ret) {
-		if (ret != -ERESTARTSYS && ret != -EBUSY)
-			DRM_ERROR("reserve failed %p %d\n", bo, ret);
-		return ret;
-	}
-	return 0;
-}
-
-void mgag200_bo_unreserve(struct mgag200_bo *bo)
-{
-	ttm_bo_unreserve(&bo->bo);
-}
-
 int mgag200_bo_create(struct drm_device *dev, int size, int align,
 		  uint32_t flags, struct mgag200_bo **pmgabo)
 {
diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index a7ff6d5..ff80f12 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -15,6 +15,13 @@ config DRM_NOUVEAU
 	select ACPI_WMI if ACPI && X86
 	select MXM_WMI if ACPI && X86
 	select POWER_SUPPLY
+	# Similar to i915, we need to select ACPI_VIDEO and it's dependencies
+	select BACKLIGHT_LCD_SUPPORT if ACPI && X86
+	select BACKLIGHT_CLASS_DEVICE if ACPI && X86
+	select VIDEO_OUTPUT_CONTROL if ACPI && X86
+	select INPUT if ACPI && X86
+	select THERMAL if ACPI && X86
+	select ACPI_VIDEO if ACPI && X86
 	help
 	  Choose this option for open-source nVidia support.
 
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 998e8b4..d939a1d 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -12,7 +12,6 @@ nouveau-y += core/core/engctx.o
 nouveau-y += core/core/engine.o
 nouveau-y += core/core/enum.o
 nouveau-y += core/core/event.o
-nouveau-y += core/core/falcon.o
 nouveau-y += core/core/gpuobj.o
 nouveau-y += core/core/handle.o
 nouveau-y += core/core/mm.o
@@ -60,6 +59,8 @@ nouveau-y += core/subdev/devinit/nv10.o
 nouveau-y += core/subdev/devinit/nv1a.o
 nouveau-y += core/subdev/devinit/nv20.o
 nouveau-y += core/subdev/devinit/nv50.o
+nouveau-y += core/subdev/devinit/nva3.o
+nouveau-y += core/subdev/devinit/nvc0.o
 nouveau-y += core/subdev/fb/base.o
 nouveau-y += core/subdev/fb/nv04.o
 nouveau-y += core/subdev/fb/nv10.o
@@ -78,6 +79,17 @@ nouveau-y += core/subdev/fb/nv49.o
 nouveau-y += core/subdev/fb/nv4e.o
 nouveau-y += core/subdev/fb/nv50.o
 nouveau-y += core/subdev/fb/nvc0.o
+nouveau-y += core/subdev/fb/ramnv04.o
+nouveau-y += core/subdev/fb/ramnv10.o
+nouveau-y += core/subdev/fb/ramnv1a.o
+nouveau-y += core/subdev/fb/ramnv20.o
+nouveau-y += core/subdev/fb/ramnv40.o
+nouveau-y += core/subdev/fb/ramnv41.o
+nouveau-y += core/subdev/fb/ramnv44.o
+nouveau-y += core/subdev/fb/ramnv49.o
+nouveau-y += core/subdev/fb/ramnv4e.o
+nouveau-y += core/subdev/fb/ramnv50.o
+nouveau-y += core/subdev/fb/ramnvc0.o
 nouveau-y += core/subdev/gpio/base.o
 nouveau-y += core/subdev/gpio/nv10.o
 nouveau-y += core/subdev/gpio/nv50.o
@@ -129,12 +141,15 @@ nouveau-y += core/subdev/vm/nv44.o
 nouveau-y += core/subdev/vm/nv50.o
 nouveau-y += core/subdev/vm/nvc0.o
 
+nouveau-y += core/engine/falcon.o
+nouveau-y += core/engine/xtensa.o
 nouveau-y += core/engine/dmaobj/base.o
 nouveau-y += core/engine/dmaobj/nv04.o
 nouveau-y += core/engine/dmaobj/nv50.o
 nouveau-y += core/engine/dmaobj/nvc0.o
 nouveau-y += core/engine/dmaobj/nvd0.o
 nouveau-y += core/engine/bsp/nv84.o
+nouveau-y += core/engine/bsp/nv98.o
 nouveau-y += core/engine/bsp/nvc0.o
 nouveau-y += core/engine/bsp/nve0.o
 nouveau-y += core/engine/copy/nva3.o
@@ -185,7 +200,13 @@ nouveau-y += core/engine/fifo/nve0.o
 nouveau-y += core/engine/graph/ctxnv40.o
 nouveau-y += core/engine/graph/ctxnv50.o
 nouveau-y += core/engine/graph/ctxnvc0.o
-nouveau-y += core/engine/graph/ctxnve0.o
+nouveau-y += core/engine/graph/ctxnvc1.o
+nouveau-y += core/engine/graph/ctxnvc3.o
+nouveau-y += core/engine/graph/ctxnvc8.o
+nouveau-y += core/engine/graph/ctxnvd7.o
+nouveau-y += core/engine/graph/ctxnvd9.o
+nouveau-y += core/engine/graph/ctxnve4.o
+nouveau-y += core/engine/graph/ctxnvf0.o
 nouveau-y += core/engine/graph/nv04.o
 nouveau-y += core/engine/graph/nv10.o
 nouveau-y += core/engine/graph/nv20.o
@@ -197,7 +218,13 @@ nouveau-y += core/engine/graph/nv35.o
 nouveau-y += core/engine/graph/nv40.o
 nouveau-y += core/engine/graph/nv50.o
 nouveau-y += core/engine/graph/nvc0.o
-nouveau-y += core/engine/graph/nve0.o
+nouveau-y += core/engine/graph/nvc1.o
+nouveau-y += core/engine/graph/nvc3.o
+nouveau-y += core/engine/graph/nvc8.o
+nouveau-y += core/engine/graph/nvd7.o
+nouveau-y += core/engine/graph/nvd9.o
+nouveau-y += core/engine/graph/nve4.o
+nouveau-y += core/engine/graph/nvf0.o
 nouveau-y += core/engine/mpeg/nv31.o
 nouveau-y += core/engine/mpeg/nv40.o
 nouveau-y += core/engine/mpeg/nv50.o
@@ -209,6 +236,7 @@ nouveau-y += core/engine/software/nv10.o
 nouveau-y += core/engine/software/nv50.o
 nouveau-y += core/engine/software/nvc0.o
 nouveau-y += core/engine/vp/nv84.o
+nouveau-y += core/engine/vp/nv98.o
 nouveau-y += core/engine/vp/nvc0.o
 nouveau-y += core/engine/vp/nve0.o
 
diff --git a/drivers/gpu/drm/nouveau/core/core/mm.c b/drivers/gpu/drm/nouveau/core/core/mm.c
index 0261a11..d829172 100644
--- a/drivers/gpu/drm/nouveau/core/core/mm.c
+++ b/drivers/gpu/drm/nouveau/core/core/mm.c
@@ -208,7 +208,6 @@ nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
 	struct nouveau_mm_node *node;
 
 	if (block) {
-		mutex_init(&mm->mutex);
 		INIT_LIST_HEAD(&mm->nodes);
 		INIT_LIST_HEAD(&mm->free);
 		mm->block_size = block;
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c
index 1d9f614..1e8e75c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c
@@ -19,24 +19,19 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs, Ilia Mirkin
  */
 
-#include <core/engctx.h>
-#include <core/class.h>
-
+#include <engine/xtensa.h>
 #include <engine/bsp.h>
 
-struct nv84_bsp_priv {
-	struct nouveau_engine base;
-};
-
 /*******************************************************************************
  * BSP object classes
  ******************************************************************************/
 
 static struct nouveau_oclass
 nv84_bsp_sclass[] = {
+	{ 0x74b0, &nouveau_object_ofuncs },
 	{},
 };
 
@@ -48,7 +43,7 @@ static struct nouveau_oclass
 nv84_bsp_cclass = {
 	.handle = NV_ENGCTX(BSP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = _nouveau_engctx_ctor,
+		.ctor = _nouveau_xtensa_engctx_ctor,
 		.dtor = _nouveau_engctx_dtor,
 		.init = _nouveau_engctx_init,
 		.fini = _nouveau_engctx_fini,
@@ -66,10 +61,10 @@ nv84_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	      struct nouveau_oclass *oclass, void *data, u32 size,
 	      struct nouveau_object **pobject)
 {
-	struct nv84_bsp_priv *priv;
+	struct nouveau_xtensa *priv;
 	int ret;
 
-	ret = nouveau_engine_create(parent, engine, oclass, true,
+	ret = nouveau_xtensa_create(parent, engine, oclass, 0x103000, true,
 				    "PBSP", "bsp", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -78,6 +73,8 @@ nv84_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->unit = 0x04008000;
 	nv_engine(priv)->cclass = &nv84_bsp_cclass;
 	nv_engine(priv)->sclass = nv84_bsp_sclass;
+	priv->fifo_val = 0x1111;
+	priv->unkd28 = 0x90044;
 	return 0;
 }
 
@@ -86,8 +83,10 @@ nv84_bsp_oclass = {
 	.handle = NV_ENGINE(BSP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv84_bsp_ctor,
-		.dtor = _nouveau_engine_dtor,
-		.init = _nouveau_engine_init,
-		.fini = _nouveau_engine_fini,
+		.dtor = _nouveau_xtensa_dtor,
+		.init = _nouveau_xtensa_init,
+		.fini = _nouveau_xtensa_fini,
+		.rd32 = _nouveau_xtensa_rd32,
+		.wr32 = _nouveau_xtensa_wr32,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nv98.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nv98.c
new file mode 100644
index 0000000..8bf92b0
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nv98.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/engctx.h>
+#include <core/class.h>
+
+#include <engine/bsp.h>
+
+struct nv98_bsp_priv {
+	struct nouveau_engine base;
+};
+
+/*******************************************************************************
+ * BSP object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_bsp_sclass[] = {
+	{},
+};
+
+/*******************************************************************************
+ * BSP context
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_bsp_cclass = {
+	.handle = NV_ENGCTX(BSP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_engctx_ctor,
+		.dtor = _nouveau_engctx_dtor,
+		.init = _nouveau_engctx_init,
+		.fini = _nouveau_engctx_fini,
+		.rd32 = _nouveau_engctx_rd32,
+		.wr32 = _nouveau_engctx_wr32,
+	},
+};
+
+/*******************************************************************************
+ * BSP engine/subdev functions
+ ******************************************************************************/
+
+static int
+nv98_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv98_bsp_priv *priv;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true,
+				    "PBSP", "bsp", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->unit = 0x04008000;
+	nv_engine(priv)->cclass = &nv98_bsp_cclass;
+	nv_engine(priv)->sclass = nv98_bsp_sclass;
+	return 0;
+}
+
+struct nouveau_oclass
+nv98_bsp_oclass = {
+	.handle = NV_ENGINE(BSP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv98_bsp_ctor,
+		.dtor = _nouveau_engine_dtor,
+		.init = _nouveau_engine_init,
+		.fini = _nouveau_engine_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c
index 0a5aa6b..262c9f5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c
@@ -22,8 +22,7 @@
  * Authors: Maarten Lankhorst
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/bsp.h>
 
 struct nvc0_bsp_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c
index d4f23bb..c46882c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c
@@ -22,8 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/bsp.h>
 
 struct nve0_bsp_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h
index c92520f..241b272 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h
@@ -1,4 +1,4 @@
-static u32 nva3_pcopy_data[] = {
+uint32_t nva3_pcopy_data[] = {
 /* 0x0000: ctx_object */
 	0x00000000,
 /* 0x0004: ctx_dma */
@@ -183,7 +183,7 @@ static u32 nva3_pcopy_data[] = {
 	0x00000800,
 };
 
-static u32 nva3_pcopy_code[] = {
+uint32_t nva3_pcopy_code[] = {
 /* 0x0000: main */
 	0x04fe04bd,
 	0x3517f000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h
index 0d98c6c..98cc421 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h
@@ -1,4 +1,4 @@
-static u32 nvc0_pcopy_data[] = {
+uint32_t nvc0_pcopy_data[] = {
 /* 0x0000: ctx_object */
 	0x00000000,
 /* 0x0004: ctx_query_address_high */
@@ -171,7 +171,7 @@ static u32 nvc0_pcopy_data[] = {
 	0x00000800,
 };
 
-static u32 nvc0_pcopy_code[] = {
+uint32_t nvc0_pcopy_code[] = {
 /* 0x0000: main */
 	0x04fe04bd,
 	0x3517f000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
index d6dc2a6..f315277 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
@@ -22,16 +22,17 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/client.h>
-#include <core/falcon.h>
-#include <core/class.h>
-#include <core/enum.h>
+#include <engine/falcon.h>
+#include <engine/fifo.h>
+#include <engine/copy.h>
 
 #include <subdev/fb.h>
 #include <subdev/vm.h>
 
-#include <engine/fifo.h>
-#include <engine/copy.h>
+#include <core/client.h>
+#include <core/class.h>
+#include <core/enum.h>
+
 
 #include "fuc/nva3.fuc.h"
 
@@ -117,13 +118,6 @@ nva3_copy_intr(struct nouveau_subdev *subdev)
 }
 
 static int
-nva3_copy_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x0d);
-	return 0;
-}
-
-static int
 nva3_copy_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
@@ -142,7 +136,6 @@ nva3_copy_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nva3_copy_intr;
 	nv_engine(priv)->cclass = &nva3_copy_cclass;
 	nv_engine(priv)->sclass = nva3_copy_sclass;
-	nv_engine(priv)->tlb_flush = nva3_copy_tlb_flush;
 	nv_falcon(priv)->code.data = nva3_pcopy_code;
 	nv_falcon(priv)->code.size = sizeof(nva3_pcopy_code);
 	nv_falcon(priv)->data.data = nva3_pcopy_data;
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
index b3ed273..993df09 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
@@ -22,13 +22,15 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/falcon.h>
-#include <core/class.h>
-#include <core/enum.h>
-
+#include <engine/falcon.h>
 #include <engine/fifo.h>
 #include <engine/copy.h>
 
+#include <core/class.h>
+#include <core/enum.h>
+#include <core/class.h>
+#include <core/enum.h>
+
 #include "fuc/nvc0.fuc.h"
 
 struct nvc0_copy_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c b/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
index dbbe9e8..30f1ef1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
@@ -67,6 +67,19 @@ nve0_copy_cclass = {
  * PCOPY engine/subdev functions
  ******************************************************************************/
 
+static void
+nve0_copy_intr(struct nouveau_subdev *subdev)
+{
+	const int ce = nv_subidx(nv_object(subdev)) - NVDEV_ENGINE_COPY0;
+	struct nve0_copy_priv *priv = (void *)subdev;
+	u32 stat = nv_rd32(priv, 0x104908 + (ce * 0x1000));
+
+	if (stat) {
+		nv_warn(priv, "unhandled intr 0x%08x\n", stat);
+		nv_wr32(priv, 0x104908 + (ce * 0x1000), stat);
+	}
+}
+
 static int
 nve0_copy0_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
@@ -85,6 +98,7 @@ nve0_copy0_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00000040;
+	nv_subdev(priv)->intr = nve0_copy_intr;
 	nv_engine(priv)->cclass = &nve0_copy_cclass;
 	nv_engine(priv)->sclass = nve0_copy_sclass;
 	return 0;
@@ -108,6 +122,28 @@ nve0_copy1_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00000080;
+	nv_subdev(priv)->intr = nve0_copy_intr;
+	nv_engine(priv)->cclass = &nve0_copy_cclass;
+	nv_engine(priv)->sclass = nve0_copy_sclass;
+	return 0;
+}
+
+static int
+nve0_copy2_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nve0_copy_priv *priv;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true,
+				    "PCE2", "copy2", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->unit = 0x00200000;
+	nv_subdev(priv)->intr = nve0_copy_intr;
 	nv_engine(priv)->cclass = &nve0_copy_cclass;
 	nv_engine(priv)->sclass = nve0_copy_sclass;
 	return 0;
@@ -134,3 +170,14 @@ nve0_copy1_oclass = {
 		.fini = _nouveau_engine_fini,
 	},
 };
+
+struct nouveau_oclass
+nve0_copy2_oclass = {
+	.handle = NV_ENGINE(COPY2, 0xe0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nve0_copy2_ctor,
+		.dtor = _nouveau_engine_dtor,
+		.init = _nouveau_engine_init,
+		.fini = _nouveau_engine_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h b/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h
index 09962e4..38676c7 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h
@@ -1,4 +1,4 @@
-static uint32_t nv98_pcrypt_data[] = {
+uint32_t nv98_pcrypt_data[] = {
 /* 0x0000: ctx_dma */
 /* 0x0000: ctx_dma_query */
 	0x00000000,
@@ -150,7 +150,7 @@ static uint32_t nv98_pcrypt_data[] = {
 	0x00000000,
 };
 
-static uint32_t nv98_pcrypt_code[] = {
+uint32_t nv98_pcrypt_code[] = {
 	0x17f004bd,
 	0x0010fe35,
 	0xf10004fe,
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
index 5bc021f..2551daf 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
@@ -141,13 +141,6 @@ nv84_crypt_intr(struct nouveau_subdev *subdev)
 }
 
 static int
-nv84_crypt_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x0a);
-	return 0;
-}
-
-static int
 nv84_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
@@ -165,7 +158,6 @@ nv84_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv84_crypt_intr;
 	nv_engine(priv)->cclass = &nv84_crypt_cclass;
 	nv_engine(priv)->sclass = nv84_crypt_sclass;
-	nv_engine(priv)->tlb_flush = nv84_crypt_tlb_flush;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
index 8bf8955..c708237 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
@@ -27,11 +27,11 @@
 #include <core/enum.h>
 #include <core/class.h>
 #include <core/engctx.h>
-#include <core/falcon.h>
 
 #include <subdev/timer.h>
 #include <subdev/fb.h>
 
+#include <engine/falcon.h>
 #include <engine/fifo.h>
 #include <engine/crypt.h>
 
@@ -119,13 +119,6 @@ nv98_crypt_intr(struct nouveau_subdev *subdev)
 }
 
 static int
-nv98_crypt_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x0a);
-	return 0;
-}
-
-static int
 nv98_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
@@ -143,7 +136,6 @@ nv98_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv98_crypt_intr;
 	nv_engine(priv)->cclass = &nv98_crypt_cclass;
 	nv_engine(priv)->sclass = nv98_crypt_sclass;
-	nv_engine(priv)->tlb_flush = nv98_crypt_tlb_flush;
 	nv_falcon(priv)->code.data = nv98_pcrypt_code;
 	nv_falcon(priv)->code.size = sizeof(nv98_pcrypt_code);
 	nv_falcon(priv)->data.data = nv98_pcrypt_data;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
index 5e8c3de..ffc18b8 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
@@ -227,9 +227,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
 		device->oclass[NVDEV_ENGINE_CRYPT  ] = &nv98_crypt_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nv94_disp_oclass;
 		break;
@@ -279,9 +279,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
 		device->oclass[NVDEV_ENGINE_CRYPT  ] = &nv98_crypt_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nv94_disp_oclass;
 		break;
@@ -305,9 +305,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
 		device->oclass[NVDEV_ENGINE_CRYPT  ] = &nv98_crypt_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nv94_disp_oclass;
 		break;
@@ -319,7 +319,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -332,8 +332,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
 		device->oclass[NVDEV_ENGINE_MPEG   ] = &nv84_mpeg_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
@@ -346,7 +346,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -358,8 +358,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
@@ -372,7 +372,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -384,8 +384,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
@@ -398,7 +398,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -410,8 +410,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
index a36e64e..418f51f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
@@ -62,7 +62,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -75,7 +75,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc0_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -91,7 +91,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -104,7 +104,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -120,7 +120,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -133,7 +133,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -148,7 +148,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -161,7 +161,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -177,7 +177,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -190,7 +190,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -206,7 +206,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -219,7 +219,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc1_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -234,7 +234,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -247,7 +247,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc8_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -263,7 +263,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -276,7 +276,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvd9_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -291,7 +291,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -304,7 +304,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvd7_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
index a354e40..7aca187 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
@@ -62,7 +62,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -75,10 +75,11 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nve0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -91,7 +92,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -104,10 +105,11 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nve0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -120,7 +122,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -133,10 +135,11 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nve0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -149,7 +152,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -160,16 +163,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
-#if 0
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
-#endif
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvf0_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nvf0_disp_oclass;
-#if 0
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
 		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
+#if 0
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
index f065fc2..db8c6fd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
@@ -55,6 +55,10 @@ nva3_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
 	nv_wr32(priv, 0x61c510 + soff, 0x00000000);
 	nv_mask(priv, 0x61c500 + soff, 0x00000001, 0x00000001);
 
+	nv_mask(priv, 0x61c5d0 + soff, 0x00070001, 0x00010001); /* SPARE, HW_CTS */
+	nv_mask(priv, 0x61c568 + soff, 0x00010101, 0x00000000); /* ACR_CTRL, ?? */
+	nv_mask(priv, 0x61c578 + soff, 0x80000000, 0x80000000); /* ACR_0441_ENABLE */
+
 	/* ??? */
 	nv_mask(priv, 0x61733c, 0x00100000, 0x00100000); /* RESETF */
 	nv_mask(priv, 0x61733c, 0x10000000, 0x10000000); /* LOOKUP_EN */
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 6a38402..7ffe2f3 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -34,9 +34,9 @@
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
 #include <subdev/bios/pll.h>
+#include <subdev/devinit.h>
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/clock.h>
 
 #include "nv50.h"
 
@@ -987,10 +987,10 @@ nv50_disp_intr_unk20_0(struct nv50_disp_priv *priv, int head)
 static void
 nv50_disp_intr_unk20_1(struct nv50_disp_priv *priv, int head)
 {
-	struct nouveau_clock *clk = nouveau_clock(priv);
+	struct nouveau_devinit *devinit = nouveau_devinit(priv);
 	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
 	if (pclk)
-		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
+		devinit->pll_set(devinit, PLL_VPLL0 + head, pclk);
 }
 
 static void
@@ -1107,6 +1107,7 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
 	u32 hval, hreg = 0x614200 + (head * 0x800);
 	u32 oval, oreg;
+	u32 mask;
 	u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
 	if (conf != ~0) {
 		if (outp.location == 0 && outp.type == DCB_OUTPUT_DP) {
@@ -1133,6 +1134,7 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 			oreg = 0x614280 + (ffs(outp.or) - 1) * 0x800;
 			oval = 0x00000000;
 			hval = 0x00000000;
+			mask = 0xffffffff;
 		} else
 		if (!outp.location) {
 			if (outp.type == DCB_OUTPUT_DP)
@@ -1140,14 +1142,16 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 			oreg = 0x614300 + (ffs(outp.or) - 1) * 0x800;
 			oval = (conf & 0x0100) ? 0x00000101 : 0x00000000;
 			hval = 0x00000000;
+			mask = 0x00000707;
 		} else {
 			oreg = 0x614380 + (ffs(outp.or) - 1) * 0x800;
 			oval = 0x00000001;
 			hval = 0x00000001;
+			mask = 0x00000707;
 		}
 
 		nv_mask(priv, hreg, 0x0000000f, hval);
-		nv_mask(priv, oreg, 0x00000707, oval);
+		nv_mask(priv, oreg, mask, oval);
 	}
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 019eacd..52dd7a1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -29,15 +29,14 @@
 
 #include <engine/disp.h>
 
-#include <subdev/timer.h>
-#include <subdev/fb.h>
-#include <subdev/clock.h>
-
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
 #include <subdev/bios/pll.h>
+#include <subdev/devinit.h>
+#include <subdev/fb.h>
+#include <subdev/timer.h>
 
 #include "nv50.h"
 
@@ -738,10 +737,10 @@ nvd0_disp_intr_unk2_0(struct nv50_disp_priv *priv, int head)
 static void
 nvd0_disp_intr_unk2_1(struct nv50_disp_priv *priv, int head)
 {
-	struct nouveau_clock *clk = nouveau_clock(priv);
+	struct nouveau_devinit *devinit = nouveau_devinit(priv);
 	u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
 	if (pclk)
-		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
+		devinit->pll_set(devinit, PLL_VPLL0 + head, pclk);
 	nv_wr32(priv, 0x612200 + (head * 0x800), 0x00000000);
 }
 
@@ -959,6 +958,9 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
+	if (nv_rd32(parent, 0x022500) & 0x00000001)
+		return -ENODEV;
+
 	ret = nouveau_disp_create(parent, engine, oclass, heads,
 				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 20725b3..fb1fe6a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -54,6 +54,9 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
+	if (nv_rd32(parent, 0x022500) & 0x00000001)
+		return -ENODEV;
+
 	ret = nouveau_disp_create(parent, engine, oclass, heads,
 				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
index a488c36..42aa6b9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
@@ -54,6 +54,9 @@ nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
+	if (nv_rd32(parent, 0x022500) & 0x00000001)
+		return -ENODEV;
+
 	ret = nouveau_disp_create(parent, engine, oclass, heads,
 				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
diff --git a/drivers/gpu/drm/nouveau/core/core/falcon.c b/drivers/gpu/drm/nouveau/core/engine/falcon.c
index e05c157..3c7a31f 100644
--- a/drivers/gpu/drm/nouveau/core/core/falcon.c
+++ b/drivers/gpu/drm/nouveau/core/engine/falcon.c
@@ -20,8 +20,7 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <subdev/timer.h>
 
 u32
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
index 2b1f917..5c7433d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
@@ -320,7 +320,7 @@ nv40_fifo_init(struct nouveau_object *object)
 		break;
 	default:
 		nv_wr32(priv, 0x002230, 0x00000000);
-		nv_wr32(priv, 0x002220, ((pfb->ram.size - 512 * 1024 +
+		nv_wr32(priv, 0x002220, ((pfb->ram->size - 512 * 1024 +
 					 priv->ramfc->addr) >> 16) |
 					0x00030000);
 		break;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
index 35b94bd..7f53196 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
@@ -56,7 +56,9 @@ nv84_fifo_context_attach(struct nouveau_object *parent,
 	switch (nv_engidx(object->engine)) {
 	case NVDEV_ENGINE_SW   : return 0;
 	case NVDEV_ENGINE_GR   : addr = 0x0020; break;
+	case NVDEV_ENGINE_VP   : addr = 0x0040; break;
 	case NVDEV_ENGINE_MPEG : addr = 0x0060; break;
+	case NVDEV_ENGINE_BSP  : addr = 0x0080; break;
 	case NVDEV_ENGINE_CRYPT: addr = 0x00a0; break;
 	case NVDEV_ENGINE_COPY0: addr = 0x00c0; break;
 	default:
@@ -89,7 +91,9 @@ nv84_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 	switch (nv_engidx(object->engine)) {
 	case NVDEV_ENGINE_SW   : return 0;
 	case NVDEV_ENGINE_GR   : engn = 0; addr = 0x0020; break;
+	case NVDEV_ENGINE_VP   : engn = 3; addr = 0x0040; break;
 	case NVDEV_ENGINE_MPEG : engn = 1; addr = 0x0060; break;
+	case NVDEV_ENGINE_BSP  : engn = 5; addr = 0x0080; break;
 	case NVDEV_ENGINE_CRYPT: engn = 4; addr = 0x00a0; break;
 	case NVDEV_ENGINE_COPY0: engn = 2; addr = 0x00c0; break;
 	default:
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index 56192a7..09644fa 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -44,7 +44,8 @@ static const struct {
 	u64 subdev;
 	u64 mask;
 } fifo_engine[] = {
-	_(NVDEV_ENGINE_GR      , (1ULL << NVDEV_ENGINE_SW)),
+	_(NVDEV_ENGINE_GR      , (1ULL << NVDEV_ENGINE_SW) |
+				 (1ULL << NVDEV_ENGINE_COPY2)),
 	_(NVDEV_ENGINE_VP      , 0),
 	_(NVDEV_ENGINE_PPP     , 0),
 	_(NVDEV_ENGINE_BSP     , 0),
@@ -96,18 +97,6 @@ nve0_fifo_playlist_update(struct nve0_fifo_priv *priv, u32 engine)
 
 	mutex_lock(&nv_subdev(priv)->mutex);
 	cur = engn->playlist[engn->cur_playlist];
-	if (unlikely(cur == NULL)) {
-		int ret = nouveau_gpuobj_new(nv_object(priv), NULL,
-					     0x8000, 0x1000, 0, &cur);
-		if (ret) {
-			mutex_unlock(&nv_subdev(priv)->mutex);
-			nv_error(priv, "playlist alloc failed\n");
-			return;
-		}
-
-		engn->playlist[engn->cur_playlist] = cur;
-	}
-
 	engn->cur_playlist = !engn->cur_playlist;
 
 	for (i = 0, p = 0; i < priv->base.max; i++) {
@@ -138,10 +127,12 @@ nve0_fifo_context_attach(struct nouveau_object *parent,
 	int ret;
 
 	switch (nv_engidx(object->engine)) {
-	case NVDEV_ENGINE_SW   : return 0;
-	case NVDEV_ENGINE_GR   :
+	case NVDEV_ENGINE_SW   :
 	case NVDEV_ENGINE_COPY0:
-	case NVDEV_ENGINE_COPY1: addr = 0x0210; break;
+	case NVDEV_ENGINE_COPY1:
+	case NVDEV_ENGINE_COPY2:
+		return 0;
+	case NVDEV_ENGINE_GR   : addr = 0x0210; break;
 	case NVDEV_ENGINE_BSP  : addr = 0x0270; break;
 	case NVDEV_ENGINE_VP   : addr = 0x0250; break;
 	case NVDEV_ENGINE_PPP  : addr = 0x0260; break;
@@ -176,9 +167,10 @@ nve0_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 
 	switch (nv_engidx(object->engine)) {
 	case NVDEV_ENGINE_SW   : return 0;
-	case NVDEV_ENGINE_GR   :
 	case NVDEV_ENGINE_COPY0:
-	case NVDEV_ENGINE_COPY1: addr = 0x0210; break;
+	case NVDEV_ENGINE_COPY1:
+	case NVDEV_ENGINE_COPY2: addr = 0x0000; break;
+	case NVDEV_ENGINE_GR   : addr = 0x0210; break;
 	case NVDEV_ENGINE_BSP  : addr = 0x0270; break;
 	case NVDEV_ENGINE_VP   : addr = 0x0250; break;
 	case NVDEV_ENGINE_PPP  : addr = 0x0260; break;
@@ -194,9 +186,12 @@ nve0_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 			return -EBUSY;
 	}
 
-	nv_wo32(base, addr + 0x00, 0x00000000);
-	nv_wo32(base, addr + 0x04, 0x00000000);
-	bar->flush(bar);
+	if (addr) {
+		nv_wo32(base, addr + 0x00, 0x00000000);
+		nv_wo32(base, addr + 0x04, 0x00000000);
+		bar->flush(bar);
+	}
+
 	return 0;
 }
 
@@ -226,8 +221,10 @@ nve0_fifo_chan_ctor(struct nouveau_object *parent,
 		}
 	}
 
-	if (i == FIFO_ENGINE_NR)
+	if (i == FIFO_ENGINE_NR) {
+		nv_error(priv, "unsupported engines 0x%08x\n", args->engine);
 		return -ENODEV;
+	}
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 1,
 					  priv->user.bar.offset, 0x200,
@@ -592,13 +589,25 @@ nve0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_object **pobject)
 {
 	struct nve0_fifo_priv *priv;
-	int ret;
+	int ret, i;
 
 	ret = nouveau_fifo_create(parent, engine, oclass, 0, 4095, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
+	for (i = 0; i < FIFO_ENGINE_NR; i++) {
+		ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x8000, 0x1000,
+					 0, &priv->engine[i].playlist[0]);
+		if (ret)
+			return ret;
+
+		ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x8000, 0x1000,
+					 0, &priv->engine[i].playlist[1]);
+		if (ret)
+			return ret;
+	}
+
 	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 4096 * 0x200, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->user.mem);
 	if (ret)
@@ -629,7 +638,7 @@ nve0_fifo_dtor(struct nouveau_object *object)
 	nouveau_gpuobj_unmap(&priv->user.bar);
 	nouveau_gpuobj_ref(NULL, &priv->user.mem);
 
-	for (i = 0; i < ARRAY_SIZE(priv->engine); i++) {
+	for (i = 0; i < FIFO_ENGINE_NR; i++) {
 		nouveau_gpuobj_ref(NULL, &priv->engine[i].playlist[1]);
 		nouveau_gpuobj_ref(NULL, &priv->engine[i].playlist[0]);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
index 4cc6269..64dca26 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
@@ -24,3015 +24,1220 @@
 
 #include "nvc0.h"
 
-void
-nv_icmd(struct nvc0_graph_priv *priv, u32 icmd, u32 data)
-{
-	nv_wr32(priv, 0x400204, data);
-	nv_wr32(priv, 0x400200, icmd);
-	while (nv_rd32(priv, 0x400700) & 2) {}
-}
+struct nvc0_graph_init
+nvc0_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_9097[] = {
+	{ 0x000800,   8, 0x40, 0x00000000 },
+	{ 0x000804,   8, 0x40, 0x00000000 },
+	{ 0x000808,   8, 0x40, 0x00000400 },
+	{ 0x00080c,   8, 0x40, 0x00000300 },
+	{ 0x000810,   1, 0x04, 0x000000cf },
+	{ 0x000850,   7, 0x40, 0x00000000 },
+	{ 0x000814,   8, 0x40, 0x00000040 },
+	{ 0x000818,   8, 0x40, 0x00000001 },
+	{ 0x00081c,   8, 0x40, 0x00000000 },
+	{ 0x000820,   8, 0x40, 0x00000000 },
+	{ 0x002700,   8, 0x20, 0x00000000 },
+	{ 0x002704,   8, 0x20, 0x00000000 },
+	{ 0x002708,   8, 0x20, 0x00000000 },
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x002710,   8, 0x20, 0x00014000 },
+	{ 0x002714,   8, 0x20, 0x00000040 },
+	{ 0x001c00,  16, 0x10, 0x00000000 },
+	{ 0x001c04,  16, 0x10, 0x00000000 },
+	{ 0x001c08,  16, 0x10, 0x00000000 },
+	{ 0x001c0c,  16, 0x10, 0x00000000 },
+	{ 0x001d00,  16, 0x10, 0x00000000 },
+	{ 0x001d04,  16, 0x10, 0x00000000 },
+	{ 0x001d08,  16, 0x10, 0x00000000 },
+	{ 0x001d0c,  16, 0x10, 0x00000000 },
+	{ 0x001f00,  16, 0x08, 0x00000000 },
+	{ 0x001f04,  16, 0x08, 0x00000000 },
+	{ 0x001f80,  16, 0x08, 0x00000000 },
+	{ 0x001f84,  16, 0x08, 0x00000000 },
+	{ 0x002200,   5, 0x10, 0x00000022 },
+	{ 0x002000,   1, 0x04, 0x00000000 },
+	{ 0x002040,   1, 0x04, 0x00000011 },
+	{ 0x002080,   1, 0x04, 0x00000020 },
+	{ 0x0020c0,   1, 0x04, 0x00000030 },
+	{ 0x002100,   1, 0x04, 0x00000040 },
+	{ 0x002140,   1, 0x04, 0x00000051 },
+	{ 0x00200c,   6, 0x40, 0x00000001 },
+	{ 0x002010,   1, 0x04, 0x00000000 },
+	{ 0x002050,   1, 0x04, 0x00000000 },
+	{ 0x002090,   1, 0x04, 0x00000001 },
+	{ 0x0020d0,   1, 0x04, 0x00000002 },
+	{ 0x002110,   1, 0x04, 0x00000003 },
+	{ 0x002150,   1, 0x04, 0x00000004 },
+	{ 0x000380,   4, 0x20, 0x00000000 },
+	{ 0x000384,   4, 0x20, 0x00000000 },
+	{ 0x000388,   4, 0x20, 0x00000000 },
+	{ 0x00038c,   4, 0x20, 0x00000000 },
+	{ 0x000700,   4, 0x10, 0x00000000 },
+	{ 0x000704,   4, 0x10, 0x00000000 },
+	{ 0x000708,   4, 0x10, 0x00000000 },
+	{ 0x002800, 128, 0x04, 0x00000000 },
+	{ 0x000a00,  16, 0x20, 0x00000000 },
+	{ 0x000a04,  16, 0x20, 0x00000000 },
+	{ 0x000a08,  16, 0x20, 0x00000000 },
+	{ 0x000a0c,  16, 0x20, 0x00000000 },
+	{ 0x000a10,  16, 0x20, 0x00000000 },
+	{ 0x000a14,  16, 0x20, 0x00000000 },
+	{ 0x000c00,  16, 0x10, 0x00000000 },
+	{ 0x000c04,  16, 0x10, 0x00000000 },
+	{ 0x000c08,  16, 0x10, 0x00000000 },
+	{ 0x000c0c,  16, 0x10, 0x3f800000 },
+	{ 0x000d00,   8, 0x08, 0xffff0000 },
+	{ 0x000d04,   8, 0x08, 0xffff0000 },
+	{ 0x000e00,  16, 0x10, 0x00000000 },
+	{ 0x000e04,  16, 0x10, 0xffff0000 },
+	{ 0x000e08,  16, 0x10, 0xffff0000 },
+	{ 0x000d40,   4, 0x08, 0x00000000 },
+	{ 0x000d44,   4, 0x08, 0x00000000 },
+	{ 0x001e00,   8, 0x20, 0x00000001 },
+	{ 0x001e04,   8, 0x20, 0x00000001 },
+	{ 0x001e08,   8, 0x20, 0x00000002 },
+	{ 0x001e0c,   8, 0x20, 0x00000001 },
+	{ 0x001e10,   8, 0x20, 0x00000001 },
+	{ 0x001e14,   8, 0x20, 0x00000002 },
+	{ 0x001e18,   8, 0x20, 0x00000001 },
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x001514,   1, 0x04, 0x00000000 },
+	{ 0x000d68,   1, 0x04, 0x0000ffff },
+	{ 0x00121c,   1, 0x04, 0x0fac6881 },
+	{ 0x000fac,   1, 0x04, 0x00000001 },
+	{ 0x001538,   1, 0x04, 0x00000001 },
+	{ 0x000fe0,   2, 0x04, 0x00000000 },
+	{ 0x000fe8,   1, 0x04, 0x00000014 },
+	{ 0x000fec,   1, 0x04, 0x00000040 },
+	{ 0x000ff0,   1, 0x04, 0x00000000 },
+	{ 0x00179c,   1, 0x04, 0x00000000 },
+	{ 0x001228,   1, 0x04, 0x00000400 },
+	{ 0x00122c,   1, 0x04, 0x00000300 },
+	{ 0x001230,   1, 0x04, 0x00010001 },
+	{ 0x0007f8,   1, 0x04, 0x00000000 },
+	{ 0x0015b4,   1, 0x04, 0x00000001 },
+	{ 0x0015cc,   1, 0x04, 0x00000000 },
+	{ 0x001534,   1, 0x04, 0x00000000 },
+	{ 0x000fb0,   1, 0x04, 0x00000000 },
+	{ 0x0015d0,   1, 0x04, 0x00000000 },
+	{ 0x00153c,   1, 0x04, 0x00000000 },
+	{ 0x0016b4,   1, 0x04, 0x00000003 },
+	{ 0x000fbc,   4, 0x04, 0x0000ffff },
+	{ 0x000df8,   2, 0x04, 0x00000000 },
+	{ 0x001948,   1, 0x04, 0x00000000 },
+	{ 0x001970,   1, 0x04, 0x00000001 },
+	{ 0x00161c,   1, 0x04, 0x000009f0 },
+	{ 0x000dcc,   1, 0x04, 0x00000010 },
+	{ 0x00163c,   1, 0x04, 0x00000000 },
+	{ 0x0015e4,   1, 0x04, 0x00000000 },
+	{ 0x001160,  32, 0x04, 0x25e00040 },
+	{ 0x001880,  32, 0x04, 0x00000000 },
+	{ 0x000f84,   2, 0x04, 0x00000000 },
+	{ 0x0017c8,   2, 0x04, 0x00000000 },
+	{ 0x0017d0,   1, 0x04, 0x000000ff },
+	{ 0x0017d4,   1, 0x04, 0xffffffff },
+	{ 0x0017d8,   1, 0x04, 0x00000002 },
+	{ 0x0017dc,   1, 0x04, 0x00000000 },
+	{ 0x0015f4,   2, 0x04, 0x00000000 },
+	{ 0x001434,   2, 0x04, 0x00000000 },
+	{ 0x000d74,   1, 0x04, 0x00000000 },
+	{ 0x000dec,   1, 0x04, 0x00000001 },
+	{ 0x0013a4,   1, 0x04, 0x00000000 },
+	{ 0x001318,   1, 0x04, 0x00000001 },
+	{ 0x001644,   1, 0x04, 0x00000000 },
+	{ 0x000748,   1, 0x04, 0x00000000 },
+	{ 0x000de8,   1, 0x04, 0x00000000 },
+	{ 0x001648,   1, 0x04, 0x00000000 },
+	{ 0x0012a4,   1, 0x04, 0x00000000 },
+	{ 0x001120,   4, 0x04, 0x00000000 },
+	{ 0x001118,   1, 0x04, 0x00000000 },
+	{ 0x00164c,   1, 0x04, 0x00000000 },
+	{ 0x001658,   1, 0x04, 0x00000000 },
+	{ 0x001910,   1, 0x04, 0x00000290 },
+	{ 0x001518,   1, 0x04, 0x00000000 },
+	{ 0x00165c,   1, 0x04, 0x00000001 },
+	{ 0x001520,   1, 0x04, 0x00000000 },
+	{ 0x001604,   1, 0x04, 0x00000000 },
+	{ 0x001570,   1, 0x04, 0x00000000 },
+	{ 0x0013b0,   2, 0x04, 0x3f800000 },
+	{ 0x00020c,   1, 0x04, 0x00000000 },
+	{ 0x001670,   1, 0x04, 0x30201000 },
+	{ 0x001674,   1, 0x04, 0x70605040 },
+	{ 0x001678,   1, 0x04, 0xb8a89888 },
+	{ 0x00167c,   1, 0x04, 0xf8e8d8c8 },
+	{ 0x00166c,   1, 0x04, 0x00000000 },
+	{ 0x001680,   1, 0x04, 0x00ffff00 },
+	{ 0x0012d0,   1, 0x04, 0x00000003 },
+	{ 0x0012d4,   1, 0x04, 0x00000002 },
+	{ 0x001684,   2, 0x04, 0x00000000 },
+	{ 0x000dac,   2, 0x04, 0x00001b02 },
+	{ 0x000db4,   1, 0x04, 0x00000000 },
+	{ 0x00168c,   1, 0x04, 0x00000000 },
+	{ 0x0015bc,   1, 0x04, 0x00000000 },
+	{ 0x00156c,   1, 0x04, 0x00000000 },
+	{ 0x00187c,   1, 0x04, 0x00000000 },
+	{ 0x001110,   1, 0x04, 0x00000001 },
+	{ 0x000dc0,   3, 0x04, 0x00000000 },
+	{ 0x001234,   1, 0x04, 0x00000000 },
+	{ 0x001690,   1, 0x04, 0x00000000 },
+	{ 0x0012ac,   1, 0x04, 0x00000001 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x001000,   1, 0x04, 0x00000010 },
+	{ 0x0010fc,   1, 0x04, 0x00000000 },
+	{ 0x001290,   1, 0x04, 0x00000000 },
+	{ 0x000218,   1, 0x04, 0x00000010 },
+	{ 0x0012d8,   1, 0x04, 0x00000000 },
+	{ 0x0012dc,   1, 0x04, 0x00000010 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x00155c,   2, 0x04, 0x00000000 },
+	{ 0x001564,   1, 0x04, 0x00001fff },
+	{ 0x001574,   2, 0x04, 0x00000000 },
+	{ 0x00157c,   1, 0x04, 0x003fffff },
+	{ 0x001354,   1, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{ 0x001610,   1, 0x04, 0x00000012 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x00162c,   1, 0x04, 0x00000003 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000324,   6, 0x04, 0x3f800000 },
+	{ 0x000750,   1, 0x04, 0x00000000 },
+	{ 0x000760,   1, 0x04, 0x39291909 },
+	{ 0x000764,   1, 0x04, 0x79695949 },
+	{ 0x000768,   1, 0x04, 0xb9a99989 },
+	{ 0x00076c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x000770,   1, 0x04, 0x30201000 },
+	{ 0x000774,   1, 0x04, 0x70605040 },
+	{ 0x000778,   1, 0x04, 0x00009080 },
+	{ 0x000780,   1, 0x04, 0x39291909 },
+	{ 0x000784,   1, 0x04, 0x79695949 },
+	{ 0x000788,   1, 0x04, 0xb9a99989 },
+	{ 0x00078c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x0007d0,   1, 0x04, 0x30201000 },
+	{ 0x0007d4,   1, 0x04, 0x70605040 },
+	{ 0x0007d8,   1, 0x04, 0x00009080 },
+	{ 0x00037c,   1, 0x04, 0x00000001 },
+	{ 0x000740,   2, 0x04, 0x00000000 },
+	{ 0x002600,   1, 0x04, 0x00000000 },
+	{ 0x001918,   1, 0x04, 0x00000000 },
+	{ 0x00191c,   1, 0x04, 0x00000900 },
+	{ 0x001920,   1, 0x04, 0x00000405 },
+	{ 0x001308,   1, 0x04, 0x00000001 },
+	{ 0x001924,   1, 0x04, 0x00000000 },
+	{ 0x0013ac,   1, 0x04, 0x00000000 },
+	{ 0x00192c,   1, 0x04, 0x00000001 },
+	{ 0x00193c,   1, 0x04, 0x00002c1c },
+	{ 0x000d7c,   1, 0x04, 0x00000000 },
+	{ 0x000f8c,   1, 0x04, 0x00000000 },
+	{ 0x0002c0,   1, 0x04, 0x00000001 },
+	{ 0x001510,   1, 0x04, 0x00000000 },
+	{ 0x001940,   1, 0x04, 0x00000000 },
+	{ 0x000ff4,   2, 0x04, 0x00000000 },
+	{ 0x00194c,   2, 0x04, 0x00000000 },
+	{ 0x001968,   1, 0x04, 0x00000000 },
+	{ 0x001590,   1, 0x04, 0x0000003f },
+	{ 0x0007e8,   4, 0x04, 0x00000000 },
+	{ 0x00196c,   1, 0x04, 0x00000011 },
+	{ 0x00197c,   1, 0x04, 0x00000000 },
+	{ 0x000fcc,   2, 0x04, 0x00000000 },
+	{ 0x0002d8,   1, 0x04, 0x00000040 },
+	{ 0x001980,   1, 0x04, 0x00000080 },
+	{ 0x001504,   1, 0x04, 0x00000080 },
+	{ 0x001984,   1, 0x04, 0x00000000 },
+	{ 0x000300,   1, 0x04, 0x00000001 },
+	{ 0x0013a8,   1, 0x04, 0x00000000 },
+	{ 0x0012ec,   1, 0x04, 0x00000000 },
+	{ 0x001310,   1, 0x04, 0x00000000 },
+	{ 0x001314,   1, 0x04, 0x00000001 },
+	{ 0x001380,   1, 0x04, 0x00000000 },
+	{ 0x001384,   4, 0x04, 0x00000001 },
+	{ 0x001394,   1, 0x04, 0x00000000 },
+	{ 0x00139c,   1, 0x04, 0x00000000 },
+	{ 0x001398,   1, 0x04, 0x00000000 },
+	{ 0x001594,   1, 0x04, 0x00000000 },
+	{ 0x001598,   4, 0x04, 0x00000001 },
+	{ 0x000f54,   3, 0x04, 0x00000000 },
+	{ 0x0019bc,   1, 0x04, 0x00000000 },
+	{ 0x000f9c,   2, 0x04, 0x00000000 },
+	{ 0x0012cc,   1, 0x04, 0x00000000 },
+	{ 0x0012e8,   1, 0x04, 0x00000000 },
+	{ 0x00130c,   1, 0x04, 0x00000001 },
+	{ 0x001360,   8, 0x04, 0x00000000 },
+	{ 0x00133c,   2, 0x04, 0x00000001 },
+	{ 0x001344,   1, 0x04, 0x00000002 },
+	{ 0x001348,   2, 0x04, 0x00000001 },
+	{ 0x001350,   1, 0x04, 0x00000002 },
+	{ 0x001358,   1, 0x04, 0x00000001 },
+	{ 0x0012e4,   1, 0x04, 0x00000000 },
+	{ 0x00131c,   1, 0x04, 0x00000000 },
+	{ 0x001320,   3, 0x04, 0x00000000 },
+	{ 0x0019c0,   1, 0x04, 0x00000000 },
+	{ 0x001140,   1, 0x04, 0x00000000 },
+	{ 0x0019c4,   1, 0x04, 0x00000000 },
+	{ 0x0019c8,   1, 0x04, 0x00001500 },
+	{ 0x00135c,   1, 0x04, 0x00000000 },
+	{ 0x000f90,   1, 0x04, 0x00000000 },
+	{ 0x0019e0,   8, 0x04, 0x00000001 },
+	{ 0x0019cc,   1, 0x04, 0x00000001 },
+	{ 0x0015b8,   1, 0x04, 0x00000000 },
+	{ 0x001a00,   1, 0x04, 0x00001111 },
+	{ 0x001a04,   7, 0x04, 0x00000000 },
+	{ 0x000d6c,   2, 0x04, 0xffff0000 },
+	{ 0x0010f8,   1, 0x04, 0x00001010 },
+	{ 0x000d80,   5, 0x04, 0x00000000 },
+	{ 0x000da0,   1, 0x04, 0x00000000 },
+	{ 0x001508,   1, 0x04, 0x80000000 },
+	{ 0x00150c,   1, 0x04, 0x40000000 },
+	{ 0x001668,   1, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000008 },
+	{ 0x000d9c,   1, 0x04, 0x00000001 },
+	{ 0x0007dc,   1, 0x04, 0x00000000 },
+	{ 0x00074c,   1, 0x04, 0x00000055 },
+	{ 0x001420,   1, 0x04, 0x00000003 },
+	{ 0x0017bc,   2, 0x04, 0x00000000 },
+	{ 0x0017c4,   1, 0x04, 0x00000001 },
+	{ 0x001008,   1, 0x04, 0x00000008 },
+	{ 0x00100c,   1, 0x04, 0x00000040 },
+	{ 0x001010,   1, 0x04, 0x0000012c },
+	{ 0x000d60,   1, 0x04, 0x00000040 },
+	{ 0x00075c,   1, 0x04, 0x00000003 },
+	{ 0x001018,   1, 0x04, 0x00000020 },
+	{ 0x00101c,   1, 0x04, 0x00000001 },
+	{ 0x001020,   1, 0x04, 0x00000020 },
+	{ 0x001024,   1, 0x04, 0x00000001 },
+	{ 0x001444,   3, 0x04, 0x00000000 },
+	{ 0x000360,   1, 0x04, 0x20164010 },
+	{ 0x000364,   1, 0x04, 0x00000020 },
+	{ 0x000368,   1, 0x04, 0x00000000 },
+	{ 0x000de4,   1, 0x04, 0x00000000 },
+	{ 0x000204,   1, 0x04, 0x00000006 },
+	{ 0x000208,   1, 0x04, 0x00000000 },
+	{ 0x0002cc,   1, 0x04, 0x003fffff },
+	{ 0x0002d0,   1, 0x04, 0x00000c48 },
+	{ 0x001220,   1, 0x04, 0x00000005 },
+	{ 0x000fdc,   1, 0x04, 0x00000000 },
+	{ 0x000f98,   1, 0x04, 0x00300008 },
+	{ 0x001284,   1, 0x04, 0x04000080 },
+	{ 0x001450,   1, 0x04, 0x00300008 },
+	{ 0x001454,   1, 0x04, 0x04000080 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_902d[] = {
+	{ 0x000200,   1, 0x04, 0x000000cf },
+	{ 0x000204,   1, 0x04, 0x00000001 },
+	{ 0x000208,   1, 0x04, 0x00000020 },
+	{ 0x00020c,   1, 0x04, 0x00000001 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000214,   1, 0x04, 0x00000080 },
+	{ 0x000218,   2, 0x04, 0x00000100 },
+	{ 0x000220,   2, 0x04, 0x00000000 },
+	{ 0x000230,   1, 0x04, 0x000000cf },
+	{ 0x000234,   1, 0x04, 0x00000001 },
+	{ 0x000238,   1, 0x04, 0x00000020 },
+	{ 0x00023c,   1, 0x04, 0x00000001 },
+	{ 0x000244,   1, 0x04, 0x00000080 },
+	{ 0x000248,   2, 0x04, 0x00000100 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_9039[] = {
+	{ 0x00030c,   3, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000238,   2, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_90c0[] = {
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x000758,   1, 0x04, 0x00000100 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x000204,   3, 0x04, 0x00000000 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{ 0x00024c,   1, 0x04, 0x00000000 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_base[] = {
+	{ 0x400204,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk40xx[] = {
+	{ 0x404004,  10, 0x04, 0x00000000 },
+	{ 0x404044,   1, 0x04, 0x00000000 },
+	{ 0x404094,   1, 0x04, 0x00000000 },
+	{ 0x404098,  12, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf0000087 },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x404168,   1, 0x04, 0x00000000 },
+	{ 0x404174,   1, 0x04, 0x00000000 },
+	{ 0x404178,   2, 0x04, 0x00000000 },
+	{ 0x404200,   8, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk44xx[] = {
+	{ 0x404404,  14, 0x04, 0x00000000 },
+	{ 0x404460,   2, 0x04, 0x00000000 },
+	{ 0x404468,   1, 0x04, 0x00ffffff },
+	{ 0x40446c,   1, 0x04, 0x00000000 },
+	{ 0x404480,   1, 0x04, 0x00000001 },
+	{ 0x404498,   1, 0x04, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk46xx[] = {
+	{ 0x404604,   1, 0x04, 0x00000015 },
+	{ 0x404608,   1, 0x04, 0x00000000 },
+	{ 0x40460c,   1, 0x04, 0x00002e00 },
+	{ 0x404610,   1, 0x04, 0x00000100 },
+	{ 0x404618,   8, 0x04, 0x00000000 },
+	{ 0x404638,   1, 0x04, 0x00000004 },
+	{ 0x40463c,   8, 0x04, 0x00000000 },
+	{ 0x40465c,   1, 0x04, 0x007f0100 },
+	{ 0x404660,   7, 0x04, 0x00000000 },
+	{ 0x40467c,   1, 0x04, 0x00000002 },
+	{ 0x404680,   8, 0x04, 0x00000000 },
+	{ 0x4046a0,   1, 0x04, 0x007f0080 },
+	{ 0x4046a4,  18, 0x04, 0x00000000 },
+	{ 0x4046f0,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk47xx[] = {
+	{ 0x404700,  13, 0x04, 0x00000000 },
+	{ 0x404734,   1, 0x04, 0x00000100 },
+	{ 0x404738,   8, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x078000bf },
+	{ 0x405830,   1, 0x04, 0x02180000 },
+	{ 0x405834,   2, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk60xx[] = {
+	{ 0x406020,   1, 0x04, 0x000103c1 },
+	{ 0x406028,   4, 0x04, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk78xx[] = {
+	{ 0x407804,   1, 0x04, 0x00000023 },
+	{ 0x40780c,   1, 0x04, 0x0a418820 },
+	{ 0x407810,   1, 0x04, 0x062080e6 },
+	{ 0x407814,   1, 0x04, 0x020398a4 },
+	{ 0x407818,   1, 0x04, 0x0e629062 },
+	{ 0x40781c,   1, 0x04, 0x0a418820 },
+	{ 0x407820,   1, 0x04, 0x000000e6 },
+	{ 0x4078bc,   1, 0x04, 0x00000103 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk80xx[] = {
+	{ 0x408000,   2, 0x04, 0x00000000 },
+	{ 0x408008,   1, 0x04, 0x00000018 },
+	{ 0x40800c,   2, 0x04, 0x00000000 },
+	{ 0x408014,   1, 0x04, 0x00000069 },
+	{ 0x408018,   1, 0x04, 0xe100e100 },
+	{ 0x408064,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x0003e00d },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x02000001 },
+	{ 0x408908,   1, 0x04, 0x00c80929 },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x00200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   1, 0x04, 0x00000000 },
+	{ 0x41870c,   1, 0x04, 0x07c80000 },
+	{ 0x418710,   1, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x0006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x00000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x00100000 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000000 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_gpc_1[] = {
+	{ 0x418a00,   3, 0x04, 0x00000000 },
+	{ 0x418a0c,   1, 0x04, 0x00010000 },
+	{ 0x418a10,   3, 0x04, 0x00000000 },
+	{ 0x418a20,   3, 0x04, 0x00000000 },
+	{ 0x418a2c,   1, 0x04, 0x00010000 },
+	{ 0x418a30,   3, 0x04, 0x00000000 },
+	{ 0x418a40,   3, 0x04, 0x00000000 },
+	{ 0x418a4c,   1, 0x04, 0x00010000 },
+	{ 0x418a50,   3, 0x04, 0x00000000 },
+	{ 0x418a60,   3, 0x04, 0x00000000 },
+	{ 0x418a6c,   1, 0x04, 0x00010000 },
+	{ 0x418a70,   3, 0x04, 0x00000000 },
+	{ 0x418a80,   3, 0x04, 0x00000000 },
+	{ 0x418a8c,   1, 0x04, 0x00010000 },
+	{ 0x418a90,   3, 0x04, 0x00000000 },
+	{ 0x418aa0,   3, 0x04, 0x00000000 },
+	{ 0x418aac,   1, 0x04, 0x00010000 },
+	{ 0x418ab0,   3, 0x04, 0x00000000 },
+	{ 0x418ac0,   3, 0x04, 0x00000000 },
+	{ 0x418acc,   1, 0x04, 0x00010000 },
+	{ 0x418ad0,   3, 0x04, 0x00000000 },
+	{ 0x418ae0,   3, 0x04, 0x00000000 },
+	{ 0x418aec,   1, 0x04, 0x00010000 },
+	{ 0x418af0,   3, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x0000012a },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00000001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00060048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x02180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419f50,   2, 0x04, 0x00000000 },
+	{}
+};
 
-int
-nvc0_grctx_init(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+void
+nvc0_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 {
-	struct nouveau_bar *bar = nouveau_bar(priv);
-	struct nouveau_gpuobj *chan;
-	u32 size = (0x80000 + priv->size + 4095) & ~4095;
-	int ret, i;
-
-	/* allocate memory to for a "channel", which we'll use to generate
-	 * the default context values
-	 */
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, size, 0x1000,
-				 NVOBJ_FLAG_ZERO_ALLOC, &info->chan);
-	chan = info->chan;
-	if (ret) {
-		nv_error(priv, "failed to allocate channel memory, %d\n", ret);
-		return ret;
-	}
-
-	/* PGD pointer */
-	nv_wo32(chan, 0x0200, lower_32_bits(chan->addr + 0x1000));
-	nv_wo32(chan, 0x0204, upper_32_bits(chan->addr + 0x1000));
-	nv_wo32(chan, 0x0208, 0xffffffff);
-	nv_wo32(chan, 0x020c, 0x000000ff);
-
-	/* PGT[0] pointer */
-	nv_wo32(chan, 0x1000, 0x00000000);
-	nv_wo32(chan, 0x1004, 0x00000001 | (chan->addr + 0x2000) >> 8);
-
-	/* identity-map the whole "channel" into its own vm */
-	for (i = 0; i < size / 4096; i++) {
-		u64 addr = ((chan->addr + (i * 4096)) >> 8) | 1;
-		nv_wo32(chan, 0x2000 + (i * 8), lower_32_bits(addr));
-		nv_wo32(chan, 0x2004 + (i * 8), upper_32_bits(addr));
-	}
-
-	/* context pointer (virt) */
-	nv_wo32(chan, 0x0210, 0x00080004);
-	nv_wo32(chan, 0x0214, 0x00000000);
+	int gpc, tpc;
+	u32 offset;
 
-	bar->flush(bar);
-
-	nv_wr32(priv, 0x100cb8, (chan->addr + 0x1000) >> 8);
-	nv_wr32(priv, 0x100cbc, 0x80000001);
-	nv_wait(priv, 0x100c80, 0x00008000, 0x00008000);
-
-	/* setup default state for mmio list construction */
-	info->data = priv->mmio_data;
-	info->mmio = priv->mmio_list;
-	info->addr = 0x2000 + (i * 8);
-	info->priv = priv;
-	info->buffer_nr = 0;
+	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
 
-	if (priv->firmware) {
-		nv_wr32(priv, 0x409840, 0x00000030);
-		nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
-		nv_wr32(priv, 0x409504, 0x00000003);
-		if (!nv_wait(priv, 0x409800, 0x00000010, 0x00000010))
-			nv_error(priv, "load_ctx timeout\n");
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000018,  0, 0);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000018,  0, 0);
 
-		nv_wo32(chan, 0x8001c, 1);
-		nv_wo32(chan, 0x80020, 0);
-		nv_wo32(chan, 0x80028, 0);
-		nv_wo32(chan, 0x8002c, 0);
-		bar->flush(bar);
-		return 0;
-	}
+	mmio_list(0x405830, 0x02180000, 0, 0);
 
-	/* HUB_FUC(SET_CHAN) */
-	nv_wr32(priv, 0x409840, 0x80000000);
-	nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
-	nv_wr32(priv, 0x409504, 0x00000001);
-	if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000)) {
-		nv_error(priv, "HUB_SET_CHAN timeout\n");
-		nvc0_graph_ctxctl_debug(priv);
-		nouveau_gpuobj_ref(NULL, &info->chan);
-		return -EBUSY;
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			u32 addr = TPC_UNIT(gpc, tpc, 0x0520);
+			mmio_list(addr, 0x02180000 | offset, 0, 0);
+			offset += 0x0324;
+		}
 	}
-
-	return 0;
 }
 
 void
-nvc0_grctx_data(struct nvc0_grctx *info, u32 size, u32 align, u32 access)
+nvc0_grctx_generate_unkn(struct nvc0_graph_priv *priv)
 {
-	info->buffer[info->buffer_nr]  = info->addr;
-	info->buffer[info->buffer_nr] +=  (align - 1);
-	info->buffer[info->buffer_nr] &= ~(align - 1);
-	info->addr = info->buffer[info->buffer_nr++] + size;
-
-	info->data->size = size;
-	info->data->align = align;
-	info->data->access = access;
-	info->data++;
 }
 
 void
-nvc0_grctx_mmio(struct nvc0_grctx *info, u32 addr, u32 data, u32 shift, u32 buf)
+nvc0_grctx_generate_tpcid(struct nvc0_graph_priv *priv)
 {
-	struct nvc0_graph_priv *priv = info->priv;
-
-	info->mmio->addr = addr;
-	info->mmio->data = data;
-	info->mmio->shift = shift;
-	info->mmio->buffer = buf;
-	info->mmio++;
+	int gpc, tpc, id;
 
-	if (shift)
-		data |= info->buffer[buf] >> shift;
-	nv_wr32(priv, addr, data);
-}
-
-int
-nvc0_grctx_fini(struct nvc0_grctx *info)
-{
-	struct nvc0_graph_priv *priv = info->priv;
-	int i;
-
-	/* trigger a context unload by unsetting the "next channel valid" bit
-	 * and faking a context switch interrupt
-	 */
-	nv_mask(priv, 0x409b04, 0x80000000, 0x00000000);
-	nv_wr32(priv, 0x409000, 0x00000100);
-	if (!nv_wait(priv, 0x409b00, 0x80000000, 0x00000000)) {
-		nv_error(priv, "grctx template channel unload timeout\n");
-		return -EBUSY;
-	}
-
-	priv->data = kmalloc(priv->size, GFP_KERNEL);
-	if (priv->data) {
-		for (i = 0; i < priv->size; i += 4)
-			priv->data[i / 4] = nv_ro32(info->chan, 0x80000 + i);
-	}
-
-	nouveau_gpuobj_ref(NULL, &info->chan);
-	return priv->data ? 0 : -ENOMEM;
-}
+	for (tpc = 0, id = 0; tpc < 4; tpc++) {
+		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+			if (tpc < priv->tpc_nr[gpc]) {
+				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x698), id);
+				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x4e8), id);
+				nv_wr32(priv, GPC_UNIT(gpc, 0x0c10 + tpc * 4), id);
+				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x088), id);
+				id++;
+			}
 
-static void
-nvc0_grctx_generate_9097(struct nvc0_graph_priv *priv)
-{
-	u32 fermi = nvc0_graph_class(priv);
-	u32 mthd;
-
-	nv_mthd(priv, 0x9097, 0x0800, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0840, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0880, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0900, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0940, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0980, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0804, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0844, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0884, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0904, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0944, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0984, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0808, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0848, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0888, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x08c8, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0908, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0948, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0988, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x09c8, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x080c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x084c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x088c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x08cc, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x090c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x094c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x098c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x09cc, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x0810, 0x000000cf);
-	nv_mthd(priv, 0x9097, 0x0850, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0890, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0910, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0950, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0990, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0814, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0854, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0894, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x08d4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0914, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0954, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0994, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x09d4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0818, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0858, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0898, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x08d8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0918, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0958, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0998, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x09d8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x081c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x085c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x089c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x091c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x095c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x099c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0820, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0860, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0920, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0960, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2700, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2720, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2740, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2760, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2780, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2704, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2724, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2744, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2764, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2784, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2708, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2728, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2748, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2768, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2788, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x270c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x272c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x274c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x276c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x278c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2710, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2730, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2750, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2770, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2790, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x27b0, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x27d0, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x27f0, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2714, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2734, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2754, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2774, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2794, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x27b4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x27d4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x27f4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x1c00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ca0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ce0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cf0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ca4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ce4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cf4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ca8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cb8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ce8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cf8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c3c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cbc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ccc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cdc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cfc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1da0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1db0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1de0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1df0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1da4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1db4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1de4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1df4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1da8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1db8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1de8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1df8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d3c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dbc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ddc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dfc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f3c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fa0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fa8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fb8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fe0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fe8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ff0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ff8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fa4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fbc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fdc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fe4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ff4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ffc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2200, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2210, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2220, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2230, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2240, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2000, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2040, 0x00000011);
-	nv_mthd(priv, 0x9097, 0x2080, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x20c0, 0x00000030);
-	nv_mthd(priv, 0x9097, 0x2100, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2140, 0x00000051);
-	nv_mthd(priv, 0x9097, 0x200c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x204c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x208c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x20cc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x210c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x214c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x2010, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2050, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2090, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x20d0, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x2110, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x2150, 0x00000004);
-	nv_mthd(priv, 0x9097, 0x0380, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0384, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0388, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x038c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0700, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0710, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0720, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0730, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0704, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0714, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0724, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0734, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0708, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0718, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0728, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0738, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2800, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2804, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2808, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x280c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2810, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2814, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2818, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x281c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2820, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2824, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2828, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x282c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2830, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2834, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2838, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x283c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2840, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2844, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2848, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x284c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2850, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2854, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2858, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x285c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2860, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2864, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2868, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x286c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2870, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2874, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2878, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x287c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2880, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2884, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2888, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x288c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2890, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2894, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2898, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x289c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28b0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28b4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28d4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2900, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2904, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2908, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x290c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2910, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2914, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2918, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x291c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2920, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2924, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2928, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x292c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2930, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2934, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2938, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x293c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2940, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2944, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2948, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x294c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2950, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2954, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2958, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x295c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2960, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2964, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2968, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x296c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2970, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2974, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2978, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x297c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2980, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2984, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2988, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x298c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2990, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2994, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2998, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x299c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29b0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29b4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29d4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aa0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ac0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ae0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ba0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0be0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aa4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ac4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ae4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ba4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0be4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aa8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ac8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ae8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ba8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0be8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0acc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ab0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ad0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0af0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bf0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ab4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ad4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0af4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bf4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ca0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ce0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cf0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ca4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ce4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cf4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ca8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cb8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ce8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cf8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c0c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c1c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c2c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c3c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c4c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c5c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c6c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c7c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c8c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c9c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cac, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cbc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0ccc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cdc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cec, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cfc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0d00, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d08, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d10, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d18, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d20, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d28, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d30, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d38, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d04, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d0c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d14, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d1c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d24, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d2c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d34, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d3c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ea0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0eb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ec0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ed0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ee0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ef0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e04, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e14, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e24, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e34, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e44, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e54, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e64, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e74, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e84, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e94, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ea4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0eb4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ec4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ed4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ee4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ef4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e08, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e18, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e28, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e38, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e48, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e58, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e68, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e78, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e88, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e98, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ea8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0eb8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ec8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ed8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ee8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ef8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1e00, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e20, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e40, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e60, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e80, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ea0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ec0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ee0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e04, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e24, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e44, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e64, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e84, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ea4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ec4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ee4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e08, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e28, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e48, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e68, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e88, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ea8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ec8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ee8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e0c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e2c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e4c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e6c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e8c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eac, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ecc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eec, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e10, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e30, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e50, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e70, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e90, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eb0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ed0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ef0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e14, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e34, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e54, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e74, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e94, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1eb4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ed4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ef4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e18, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e38, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e58, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e78, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e98, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eb8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ed8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ef8, 0x00000001);
-	if (fermi == 0x9097) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(priv, 0x9097, mthd, 0x00000000);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0c08), priv->tpc_nr[gpc]);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0c8c), priv->tpc_nr[gpc]);
+		}
 	}
-	nv_mthd(priv, 0x9097, 0x030c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1944, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1514, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d68, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x121c, 0x0fac6881);
-	nv_mthd(priv, 0x9097, 0x0fac, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1538, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0fe0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fe4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fe8, 0x00000014);
-	nv_mthd(priv, 0x9097, 0x0fec, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0ff0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x179c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1228, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x122c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x1230, 0x00010001);
-	nv_mthd(priv, 0x9097, 0x07f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15b4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1534, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x153c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x16b4, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x0fbc, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0fc0, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0fc4, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0fc8, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0df8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dfc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1948, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1970, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x161c, 0x000009f0);
-	nv_mthd(priv, 0x9097, 0x0dcc, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x163c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1160, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1164, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1168, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x116c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1170, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1174, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1178, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x117c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1180, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1184, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1188, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x118c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1190, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1194, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1198, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x119c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11a0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11a4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11a8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11ac, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11b0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11b4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11b8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11bc, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11c0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11c4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11c8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11cc, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11d0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11d4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11d8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11dc, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1880, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1884, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1888, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x188c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1890, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1894, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1898, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x189c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18b0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18b4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18d4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17d0, 0x000000ff);
-	nv_mthd(priv, 0x9097, 0x17d4, 0xffffffff);
-	nv_mthd(priv, 0x9097, 0x17d8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x17dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1434, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1438, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dec, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x13a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1318, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1644, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0748, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0de8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1648, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1120, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1124, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1128, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x112c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1118, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x164c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1658, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1910, 0x00000290);
-	nv_mthd(priv, 0x9097, 0x1518, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x165c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1520, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1604, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1570, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x13b0, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x13b4, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x020c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1670, 0x30201000);
-	nv_mthd(priv, 0x9097, 0x1674, 0x70605040);
-	nv_mthd(priv, 0x9097, 0x1678, 0xb8a89888);
-	nv_mthd(priv, 0x9097, 0x167c, 0xf8e8d8c8);
-	nv_mthd(priv, 0x9097, 0x166c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1680, 0x00ffff00);
-	nv_mthd(priv, 0x9097, 0x12d0, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x12d4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1684, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1688, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dac, 0x00001b02);
-	nv_mthd(priv, 0x9097, 0x0db0, 0x00001b02);
-	nv_mthd(priv, 0x9097, 0x0db4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x168c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x156c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x187c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1110, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0dc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1234, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1690, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12ac, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x02c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0790, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0794, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0798, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x079c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x077c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1000, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x10fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1290, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0218, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x12d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12dc, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x0d94, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x155c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1560, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1564, 0x00001fff);
-	nv_mthd(priv, 0x9097, 0x1574, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1578, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x157c, 0x003fffff);
-	nv_mthd(priv, 0x9097, 0x1354, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1664, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1610, 0x00000012);
-	nv_mthd(priv, 0x9097, 0x1608, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x160c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x162c, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x0210, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0320, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0324, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0328, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x032c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0330, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0334, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0338, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0750, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0760, 0x39291909);
-	nv_mthd(priv, 0x9097, 0x0764, 0x79695949);
-	nv_mthd(priv, 0x9097, 0x0768, 0xb9a99989);
-	nv_mthd(priv, 0x9097, 0x076c, 0xf9e9d9c9);
-	nv_mthd(priv, 0x9097, 0x0770, 0x30201000);
-	nv_mthd(priv, 0x9097, 0x0774, 0x70605040);
-	nv_mthd(priv, 0x9097, 0x0778, 0x00009080);
-	nv_mthd(priv, 0x9097, 0x0780, 0x39291909);
-	nv_mthd(priv, 0x9097, 0x0784, 0x79695949);
-	nv_mthd(priv, 0x9097, 0x0788, 0xb9a99989);
-	nv_mthd(priv, 0x9097, 0x078c, 0xf9e9d9c9);
-	nv_mthd(priv, 0x9097, 0x07d0, 0x30201000);
-	nv_mthd(priv, 0x9097, 0x07d4, 0x70605040);
-	nv_mthd(priv, 0x9097, 0x07d8, 0x00009080);
-	nv_mthd(priv, 0x9097, 0x037c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0740, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0744, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2600, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1918, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x191c, 0x00000900);
-	nv_mthd(priv, 0x9097, 0x1920, 0x00000405);
-	nv_mthd(priv, 0x9097, 0x1308, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1924, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x13ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x192c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x193c, 0x00002c1c);
-	nv_mthd(priv, 0x9097, 0x0d7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x02c0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1510, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1940, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ff4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ff8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x194c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1950, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1968, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1590, 0x0000003f);
-	nv_mthd(priv, 0x9097, 0x07e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x196c, 0x00000011);
-	nv_mthd(priv, 0x9097, 0x197c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x02d8, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x1980, 0x00000080);
-	nv_mthd(priv, 0x9097, 0x1504, 0x00000080);
-	nv_mthd(priv, 0x9097, 0x1984, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0300, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x13a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1310, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1314, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1380, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1384, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1388, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x138c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1390, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1394, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x139c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1398, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1594, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1598, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x159c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15a0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15a4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0f54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fa0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x130c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1360, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1364, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1368, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x136c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1370, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1374, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1378, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x137c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x133c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1340, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1344, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1348, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x134c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1350, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1358, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x12e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x131c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1320, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1324, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1328, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1140, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19c8, 0x00001500);
-	nv_mthd(priv, 0x9097, 0x135c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19e0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19e4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19e8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19ec, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19f0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19f4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19f8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19fc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19cc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a00, 0x00001111);
-	nv_mthd(priv, 0x9097, 0x1a04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d6c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d70, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x10f8, 0x00001010);
-	nv_mthd(priv, 0x9097, 0x0d80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0da0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1508, 0x80000000);
-	nv_mthd(priv, 0x9097, 0x150c, 0x40000000);
-	nv_mthd(priv, 0x9097, 0x1668, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0318, 0x00000008);
-	nv_mthd(priv, 0x9097, 0x031c, 0x00000008);
-	nv_mthd(priv, 0x9097, 0x0d9c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x07dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x074c, 0x00000055);
-	nv_mthd(priv, 0x9097, 0x1420, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x17bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17c4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1008, 0x00000008);
-	nv_mthd(priv, 0x9097, 0x100c, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x1010, 0x0000012c);
-	nv_mthd(priv, 0x9097, 0x0d60, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x075c, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x1018, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x101c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1020, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x1024, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1444, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1448, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x144c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0360, 0x20164010);
-	nv_mthd(priv, 0x9097, 0x0364, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x0368, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0de4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0204, 0x00000006);
-	nv_mthd(priv, 0x9097, 0x0208, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x02cc, 0x003fffff);
-	nv_mthd(priv, 0x9097, 0x02d0, 0x00000c48);
-	nv_mthd(priv, 0x9097, 0x1220, 0x00000005);
-	nv_mthd(priv, 0x9097, 0x0fdc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f98, 0x00300008);
-	nv_mthd(priv, 0x9097, 0x1284, 0x04000080);
-	nv_mthd(priv, 0x9097, 0x1450, 0x00300008);
-	nv_mthd(priv, 0x9097, 0x1454, 0x04000080);
-	nv_mthd(priv, 0x9097, 0x0214, 0x00000000);
-	/* in trace, right after 0x90c0, not here */
-	nv_mthd(priv, 0x9097, 0x3410, 0x80002006);
 }
 
-static void
-nvc0_grctx_generate_9197(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r406028(struct nvc0_graph_priv *priv)
 {
-	u32 fermi = nvc0_graph_class(priv);
-	u32 mthd;
-
-	if (fermi == 0x9197) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(priv, 0x9197, mthd, 0x00000000);
+	u32 tmp[GPC_MAX / 8] = {}, i = 0;
+	for (i = 0; i < priv->gpc_nr; i++)
+		tmp[i / 8] |= priv->tpc_nr[i] << ((i % 8) * 4);
+	for (i = 0; i < 4; i++) {
+		nv_wr32(priv, 0x406028 + (i * 4), tmp[i]);
+		nv_wr32(priv, 0x405870 + (i * 4), tmp[i]);
 	}
-	nv_mthd(priv, 0x9197, 0x02e4, 0x0000b001);
 }
 
-static void
-nvc0_grctx_generate_9297(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r4060a8(struct nvc0_graph_priv *priv)
 {
-	u32 fermi = nvc0_graph_class(priv);
-	u32 mthd;
-
-	if (fermi == 0x9297) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(priv, 0x9297, mthd, 0x00000000);
+	u8  tpcnr[GPC_MAX], data[TPC_MAX];
+	int gpc, tpc, i;
+
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	memset(data, 0x1f, sizeof(data));
+
+	gpc = -1;
+	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpcnr[gpc]--;
+		data[tpc] = gpc;
 	}
-	nv_mthd(priv, 0x9297, 0x036c, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x0370, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x07a4, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x07a8, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x0374, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x0378, 0x00000020);
-}
 
-static void
-nvc0_grctx_generate_902d(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0x902d, 0x0200, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0204, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0208, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x020c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0210, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0214, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0218, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x021c, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x0220, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0224, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0230, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0234, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0238, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x023c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0244, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0248, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x024c, 0x00000100);
-}
-
-static void
-nvc0_grctx_generate_9039(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0x9039, 0x030c, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0310, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0314, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0320, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0238, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x023c, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0318, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x031c, 0x00000000);
+	for (i = 0; i < 4; i++)
+		nv_wr32(priv, 0x4060a8 + (i * 4), ((u32 *)data)[i]);
 }
 
-static void
-nvc0_grctx_generate_90c0(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r418bb8(struct nvc0_graph_priv *priv)
 {
-	int i;
-
-	for (i = 0; nv_device(priv)->chipset >= 0xd0 && i < 4; i++) {
-		nv_mthd(priv, 0x90c0, 0x2700 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2720 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2704 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2724 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2708 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2728 + (i * 0x40), 0x00000000);
+	u32 data[6] = {}, data2[2] = {};
+	u8  tpcnr[GPC_MAX];
+	u8  shift, ntpcv;
+	int gpc, tpc, i;
+
+	/* calculate first set of magics */
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+
+	gpc = -1;
+	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpcnr[gpc]--;
+
+		data[tpc / 6] |= gpc << ((tpc % 6) * 5);
 	}
-	nv_mthd(priv, 0x90c0, 0x270c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x272c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x274c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x276c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x278c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x27ac, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x27cc, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x27ec, 0x00000000);
-	for (i = 0; nv_device(priv)->chipset >= 0xd0 && i < 4; i++) {
-		nv_mthd(priv, 0x90c0, 0x2710 + (i * 0x40), 0x00014000);
-		nv_mthd(priv, 0x90c0, 0x2730 + (i * 0x40), 0x00014000);
-		nv_mthd(priv, 0x90c0, 0x2714 + (i * 0x40), 0x00000040);
-		nv_mthd(priv, 0x90c0, 0x2734 + (i * 0x40), 0x00000040);
-	}
-	nv_mthd(priv, 0x90c0, 0x030c, 0x00000001);
-	nv_mthd(priv, 0x90c0, 0x1944, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0758, 0x00000100);
-	nv_mthd(priv, 0x90c0, 0x02c4, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0790, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0794, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0798, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x079c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x07a0, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x077c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0204, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0208, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x020c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0214, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x024c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0d94, 0x00000001);
-	nv_mthd(priv, 0x90c0, 0x1608, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x160c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x1664, 0x00000000);
-}
 
-static void
-nvc0_grctx_generate_dispatch(struct nvc0_graph_priv *priv)
-{
-	int i;
+	for (; tpc < 32; tpc++)
+		data[tpc / 6] |= 7 << ((tpc % 6) * 5);
 
-	nv_wr32(priv, 0x404004, 0x00000000);
-	nv_wr32(priv, 0x404008, 0x00000000);
-	nv_wr32(priv, 0x40400c, 0x00000000);
-	nv_wr32(priv, 0x404010, 0x00000000);
-	nv_wr32(priv, 0x404014, 0x00000000);
-	nv_wr32(priv, 0x404018, 0x00000000);
-	nv_wr32(priv, 0x40401c, 0x00000000);
-	nv_wr32(priv, 0x404020, 0x00000000);
-	nv_wr32(priv, 0x404024, 0x00000000);
-	nv_wr32(priv, 0x404028, 0x00000000);
-	nv_wr32(priv, 0x40402c, 0x00000000);
-	nv_wr32(priv, 0x404044, 0x00000000);
-	nv_wr32(priv, 0x404094, 0x00000000);
-	nv_wr32(priv, 0x404098, 0x00000000);
-	nv_wr32(priv, 0x40409c, 0x00000000);
-	nv_wr32(priv, 0x4040a0, 0x00000000);
-	nv_wr32(priv, 0x4040a4, 0x00000000);
-	nv_wr32(priv, 0x4040a8, 0x00000000);
-	nv_wr32(priv, 0x4040ac, 0x00000000);
-	nv_wr32(priv, 0x4040b0, 0x00000000);
-	nv_wr32(priv, 0x4040b4, 0x00000000);
-	nv_wr32(priv, 0x4040b8, 0x00000000);
-	nv_wr32(priv, 0x4040bc, 0x00000000);
-	nv_wr32(priv, 0x4040c0, 0x00000000);
-	nv_wr32(priv, 0x4040c4, 0x00000000);
-	nv_wr32(priv, 0x4040c8, 0xf0000087);
-	nv_wr32(priv, 0x4040d4, 0x00000000);
-	nv_wr32(priv, 0x4040d8, 0x00000000);
-	nv_wr32(priv, 0x4040dc, 0x00000000);
-	nv_wr32(priv, 0x4040e0, 0x00000000);
-	nv_wr32(priv, 0x4040e4, 0x00000000);
-	nv_wr32(priv, 0x4040e8, 0x00001000);
-	nv_wr32(priv, 0x4040f8, 0x00000000);
-	nv_wr32(priv, 0x404130, 0x00000000);
-	nv_wr32(priv, 0x404134, 0x00000000);
-	nv_wr32(priv, 0x404138, 0x20000040);
-	nv_wr32(priv, 0x404150, 0x0000002e);
-	nv_wr32(priv, 0x404154, 0x00000400);
-	nv_wr32(priv, 0x404158, 0x00000200);
-	nv_wr32(priv, 0x404164, 0x00000055);
-	nv_wr32(priv, 0x404168, 0x00000000);
-	nv_wr32(priv, 0x404174, 0x00000000);
-	nv_wr32(priv, 0x404178, 0x00000000);
-	nv_wr32(priv, 0x40417c, 0x00000000);
-	for (i = 0; i < 8; i++)
-		nv_wr32(priv, 0x404200 + (i * 4), 0x00000000); /* subc */
-}
-
-static void
-nvc0_grctx_generate_macro(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404404, 0x00000000);
-	nv_wr32(priv, 0x404408, 0x00000000);
-	nv_wr32(priv, 0x40440c, 0x00000000);
-	nv_wr32(priv, 0x404410, 0x00000000);
-	nv_wr32(priv, 0x404414, 0x00000000);
-	nv_wr32(priv, 0x404418, 0x00000000);
-	nv_wr32(priv, 0x40441c, 0x00000000);
-	nv_wr32(priv, 0x404420, 0x00000000);
-	nv_wr32(priv, 0x404424, 0x00000000);
-	nv_wr32(priv, 0x404428, 0x00000000);
-	nv_wr32(priv, 0x40442c, 0x00000000);
-	nv_wr32(priv, 0x404430, 0x00000000);
-	nv_wr32(priv, 0x404434, 0x00000000);
-	nv_wr32(priv, 0x404438, 0x00000000);
-	nv_wr32(priv, 0x404460, 0x00000000);
-	nv_wr32(priv, 0x404464, 0x00000000);
-	nv_wr32(priv, 0x404468, 0x00ffffff);
-	nv_wr32(priv, 0x40446c, 0x00000000);
-	nv_wr32(priv, 0x404480, 0x00000001);
-	nv_wr32(priv, 0x404498, 0x00000001);
-}
-
-static void
-nvc0_grctx_generate_m2mf(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404604, 0x00000015);
-	nv_wr32(priv, 0x404608, 0x00000000);
-	nv_wr32(priv, 0x40460c, 0x00002e00);
-	nv_wr32(priv, 0x404610, 0x00000100);
-	nv_wr32(priv, 0x404618, 0x00000000);
-	nv_wr32(priv, 0x40461c, 0x00000000);
-	nv_wr32(priv, 0x404620, 0x00000000);
-	nv_wr32(priv, 0x404624, 0x00000000);
-	nv_wr32(priv, 0x404628, 0x00000000);
-	nv_wr32(priv, 0x40462c, 0x00000000);
-	nv_wr32(priv, 0x404630, 0x00000000);
-	nv_wr32(priv, 0x404634, 0x00000000);
-	nv_wr32(priv, 0x404638, 0x00000004);
-	nv_wr32(priv, 0x40463c, 0x00000000);
-	nv_wr32(priv, 0x404640, 0x00000000);
-	nv_wr32(priv, 0x404644, 0x00000000);
-	nv_wr32(priv, 0x404648, 0x00000000);
-	nv_wr32(priv, 0x40464c, 0x00000000);
-	nv_wr32(priv, 0x404650, 0x00000000);
-	nv_wr32(priv, 0x404654, 0x00000000);
-	nv_wr32(priv, 0x404658, 0x00000000);
-	nv_wr32(priv, 0x40465c, 0x007f0100);
-	nv_wr32(priv, 0x404660, 0x00000000);
-	nv_wr32(priv, 0x404664, 0x00000000);
-	nv_wr32(priv, 0x404668, 0x00000000);
-	nv_wr32(priv, 0x40466c, 0x00000000);
-	nv_wr32(priv, 0x404670, 0x00000000);
-	nv_wr32(priv, 0x404674, 0x00000000);
-	nv_wr32(priv, 0x404678, 0x00000000);
-	nv_wr32(priv, 0x40467c, 0x00000002);
-	nv_wr32(priv, 0x404680, 0x00000000);
-	nv_wr32(priv, 0x404684, 0x00000000);
-	nv_wr32(priv, 0x404688, 0x00000000);
-	nv_wr32(priv, 0x40468c, 0x00000000);
-	nv_wr32(priv, 0x404690, 0x00000000);
-	nv_wr32(priv, 0x404694, 0x00000000);
-	nv_wr32(priv, 0x404698, 0x00000000);
-	nv_wr32(priv, 0x40469c, 0x00000000);
-	nv_wr32(priv, 0x4046a0, 0x007f0080);
-	nv_wr32(priv, 0x4046a4, 0x00000000);
-	nv_wr32(priv, 0x4046a8, 0x00000000);
-	nv_wr32(priv, 0x4046ac, 0x00000000);
-	nv_wr32(priv, 0x4046b0, 0x00000000);
-	nv_wr32(priv, 0x4046b4, 0x00000000);
-	nv_wr32(priv, 0x4046b8, 0x00000000);
-	nv_wr32(priv, 0x4046bc, 0x00000000);
-	nv_wr32(priv, 0x4046c0, 0x00000000);
-	nv_wr32(priv, 0x4046c4, 0x00000000);
-	nv_wr32(priv, 0x4046c8, 0x00000000);
-	nv_wr32(priv, 0x4046cc, 0x00000000);
-	nv_wr32(priv, 0x4046d0, 0x00000000);
-	nv_wr32(priv, 0x4046d4, 0x00000000);
-	nv_wr32(priv, 0x4046d8, 0x00000000);
-	nv_wr32(priv, 0x4046dc, 0x00000000);
-	nv_wr32(priv, 0x4046e0, 0x00000000);
-	nv_wr32(priv, 0x4046e4, 0x00000000);
-	nv_wr32(priv, 0x4046e8, 0x00000000);
-	nv_wr32(priv, 0x4046f0, 0x00000000);
-	nv_wr32(priv, 0x4046f4, 0x00000000);
-}
+	/* and the second... */
+	shift = 0;
+	ntpcv = priv->tpc_total;
+	while (!(ntpcv & (1 << 4))) {
+		ntpcv <<= 1;
+		shift++;
+	}
 
-static void
-nvc0_grctx_generate_unk47xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404700, 0x00000000);
-	nv_wr32(priv, 0x404704, 0x00000000);
-	nv_wr32(priv, 0x404708, 0x00000000);
-	nv_wr32(priv, 0x40470c, 0x00000000);
-	nv_wr32(priv, 0x404710, 0x00000000);
-	nv_wr32(priv, 0x404714, 0x00000000);
-	nv_wr32(priv, 0x404718, 0x00000000);
-	nv_wr32(priv, 0x40471c, 0x00000000);
-	nv_wr32(priv, 0x404720, 0x00000000);
-	nv_wr32(priv, 0x404724, 0x00000000);
-	nv_wr32(priv, 0x404728, 0x00000000);
-	nv_wr32(priv, 0x40472c, 0x00000000);
-	nv_wr32(priv, 0x404730, 0x00000000);
-	nv_wr32(priv, 0x404734, 0x00000100);
-	nv_wr32(priv, 0x404738, 0x00000000);
-	nv_wr32(priv, 0x40473c, 0x00000000);
-	nv_wr32(priv, 0x404740, 0x00000000);
-	nv_wr32(priv, 0x404744, 0x00000000);
-	nv_wr32(priv, 0x404748, 0x00000000);
-	nv_wr32(priv, 0x40474c, 0x00000000);
-	nv_wr32(priv, 0x404750, 0x00000000);
-	nv_wr32(priv, 0x404754, 0x00000000);
-}
+	data2[0]  = (ntpcv << 16);
+	data2[0] |= (shift << 21);
+	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
+	for (i = 1; i < 7; i++)
+		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
 
-static void
-nvc0_grctx_generate_shaders(struct nvc0_graph_priv *priv)
-{
+	/* GPC_BROADCAST */
+	nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
 
-	if (nv_device(priv)->chipset >= 0xd0) {
-		nv_wr32(priv, 0x405800, 0x0f8000bf);
-		nv_wr32(priv, 0x405830, 0x02180218);
-		nv_wr32(priv, 0x405834, 0x08000000);
-	} else
-	if (nv_device(priv)->chipset == 0xc1) {
-		nv_wr32(priv, 0x405800, 0x0f8000bf);
-		nv_wr32(priv, 0x405830, 0x02180218);
-		nv_wr32(priv, 0x405834, 0x00000000);
-	} else {
-		nv_wr32(priv, 0x405800, 0x078000bf);
-		nv_wr32(priv, 0x405830, 0x02180000);
-		nv_wr32(priv, 0x405834, 0x00000000);
-	}
-	nv_wr32(priv, 0x405838, 0x00000000);
-	nv_wr32(priv, 0x405854, 0x00000000);
-	nv_wr32(priv, 0x405870, 0x00000001);
-	nv_wr32(priv, 0x405874, 0x00000001);
-	nv_wr32(priv, 0x405878, 0x00000001);
-	nv_wr32(priv, 0x40587c, 0x00000001);
-	nv_wr32(priv, 0x405a00, 0x00000000);
-	nv_wr32(priv, 0x405a04, 0x00000000);
-	nv_wr32(priv, 0x405a18, 0x00000000);
+	/* GPC_BROADCAST.TP_BROADCAST */
+	nv_wr32(priv, 0x419bd0, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr | data2[0]);
+	nv_wr32(priv, 0x419be4, data2[1]);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x419b00 + (i * 4), data[i]);
+
+	/* UNK78xx */
+	nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x40780c + (i * 4), data[i]);
 }
 
-static void
-nvc0_grctx_generate_unk60xx(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r406800(struct nvc0_graph_priv *priv)
 {
-	nv_wr32(priv, 0x406020, 0x000103c1);
-	nv_wr32(priv, 0x406028, 0x00000001);
-	nv_wr32(priv, 0x40602c, 0x00000001);
-	nv_wr32(priv, 0x406030, 0x00000001);
-	nv_wr32(priv, 0x406034, 0x00000001);
-}
+	u64 tpc_mask = 0, tpc_set = 0;
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc;
+	int i, a, b;
+
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++)
+		tpc_mask |= ((1ULL << priv->tpc_nr[gpc]) - 1) << (gpc * 8);
+
+	for (i = 0, gpc = -1, b = -1; i < 32; i++) {
+		a = (i * (priv->tpc_total - 1)) / 32;
+		if (a != b) {
+			b = a;
+			do {
+				gpc = (gpc + 1) % priv->gpc_nr;
+			} while (!tpcnr[gpc]);
+			tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
 
-static void
-nvc0_grctx_generate_unk64xx(struct nvc0_graph_priv *priv)
-{
+			tpc_set |= 1 << ((gpc * 8) + tpc);
+		}
 
-	nv_wr32(priv, 0x4064a8, 0x00000000);
-	nv_wr32(priv, 0x4064ac, 0x00003fff);
-	nv_wr32(priv, 0x4064b4, 0x00000000);
-	nv_wr32(priv, 0x4064b8, 0x00000000);
-	if (nv_device(priv)->chipset >= 0xd0)
-		nv_wr32(priv, 0x4064bc, 0x00000000);
-	if (nv_device(priv)->chipset == 0xc1 ||
-	    nv_device(priv)->chipset >= 0xd0) {
-		nv_wr32(priv, 0x4064c0, 0x80140078);
-		nv_wr32(priv, 0x4064c4, 0x0086ffff);
+		nv_wr32(priv, 0x406800 + (i * 0x20), lower_32_bits(tpc_set));
+		nv_wr32(priv, 0x406c00 + (i * 0x20), lower_32_bits(tpc_set ^ tpc_mask));
+		if (priv->gpc_nr > 4) {
+			nv_wr32(priv, 0x406804 + (i * 0x20), upper_32_bits(tpc_set));
+			nv_wr32(priv, 0x406c04 + (i * 0x20), upper_32_bits(tpc_set ^ tpc_mask));
+		}
 	}
 }
 
-static void
-nvc0_grctx_generate_tpbus(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 {
-	nv_wr32(priv, 0x407804, 0x00000023);
-	nv_wr32(priv, 0x40780c, 0x0a418820);
-	nv_wr32(priv, 0x407810, 0x062080e6);
-	nv_wr32(priv, 0x407814, 0x020398a4);
-	nv_wr32(priv, 0x407818, 0x0e629062);
-	nv_wr32(priv, 0x40781c, 0x0a418820);
-	nv_wr32(priv, 0x407820, 0x000000e6);
-	nv_wr32(priv, 0x4078bc, 0x00000103);
-}
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	int i;
 
-static void
-nvc0_grctx_generate_ccache(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x408000, 0x00000000);
-	nv_wr32(priv, 0x408004, 0x00000000);
-	nv_wr32(priv, 0x408008, 0x00000018);
-	nv_wr32(priv, 0x40800c, 0x00000000);
-	nv_wr32(priv, 0x408010, 0x00000000);
-	nv_wr32(priv, 0x408014, 0x00000069);
-	nv_wr32(priv, 0x408018, 0xe100e100);
-	nv_wr32(priv, 0x408064, 0x00000000);
-}
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
 
-static void
-nvc0_grctx_generate_rop(struct nvc0_graph_priv *priv)
-{
-	int chipset = nv_device(priv)->chipset;
-
-	/* ROPC_BROADCAST */
-	nv_wr32(priv, 0x408800, 0x02802a3c);
-	nv_wr32(priv, 0x408804, 0x00000040);
-	if (chipset >= 0xd0) {
-		nv_wr32(priv, 0x408808, 0x1043e005);
-		nv_wr32(priv, 0x408900, 0x3080b801);
-		nv_wr32(priv, 0x408904, 0x1043e005);
-		nv_wr32(priv, 0x408908, 0x00c8102f);
-	} else
-	if (chipset == 0xc1) {
-		nv_wr32(priv, 0x408808, 0x1003e005);
-		nv_wr32(priv, 0x408900, 0x3080b801);
-		nv_wr32(priv, 0x408904, 0x62000001);
-		nv_wr32(priv, 0x408908, 0x00c80929);
-	} else {
-		nv_wr32(priv, 0x408808, 0x0003e00d);
-		nv_wr32(priv, 0x408900, 0x3080b801);
-		nv_wr32(priv, 0x408904, 0x02000001);
-		nv_wr32(priv, 0x408908, 0x00c80929);
-	}
-	nv_wr32(priv, 0x40890c, 0x00000000);
-	nv_wr32(priv, 0x408980, 0x0000011d);
-}
+	for (i = 0; oclass->hub[i]; i++)
+		nvc0_graph_mmio(priv, oclass->hub[i]);
+	for (i = 0; oclass->gpc[i]; i++)
+		nvc0_graph_mmio(priv, oclass->gpc[i]);
 
-static void
-nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
-{
-	int chipset = nv_device(priv)->chipset;
-	int i;
+	nv_wr32(priv, 0x404154, 0x00000000);
 
-	/* GPC_BROADCAST */
-	nv_wr32(priv, 0x418380, 0x00000016);
-	nv_wr32(priv, 0x418400, 0x38004e00);
-	nv_wr32(priv, 0x418404, 0x71e0ffff);
-	nv_wr32(priv, 0x418408, 0x00000000);
-	nv_wr32(priv, 0x41840c, 0x00001008);
-	nv_wr32(priv, 0x418410, 0x0fff0fff);
-	nv_wr32(priv, 0x418414, chipset < 0xd0 ? 0x00200fff : 0x02200fff);
-	nv_wr32(priv, 0x418450, 0x00000000);
-	nv_wr32(priv, 0x418454, 0x00000000);
-	nv_wr32(priv, 0x418458, 0x00000000);
-	nv_wr32(priv, 0x41845c, 0x00000000);
-	nv_wr32(priv, 0x418460, 0x00000000);
-	nv_wr32(priv, 0x418464, 0x00000000);
-	nv_wr32(priv, 0x418468, 0x00000001);
-	nv_wr32(priv, 0x41846c, 0x00000000);
-	nv_wr32(priv, 0x418470, 0x00000000);
-	nv_wr32(priv, 0x418600, 0x0000001f);
-	nv_wr32(priv, 0x418684, 0x0000000f);
-	nv_wr32(priv, 0x418700, 0x00000002);
-	nv_wr32(priv, 0x418704, 0x00000080);
-	nv_wr32(priv, 0x418708, 0x00000000);
-	nv_wr32(priv, 0x41870c, chipset < 0xd0 ? 0x07c80000 : 0x00000000);
-	nv_wr32(priv, 0x418710, 0x00000000);
-	nv_wr32(priv, 0x418800, chipset < 0xd0 ? 0x0006860a : 0x7006860a);
-	nv_wr32(priv, 0x418808, 0x00000000);
-	nv_wr32(priv, 0x41880c, 0x00000000);
-	nv_wr32(priv, 0x418810, 0x00000000);
-	nv_wr32(priv, 0x418828, 0x00008442);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x418830, 0x10000001);
-	else
-		nv_wr32(priv, 0x418830, 0x00000001);
-	nv_wr32(priv, 0x4188d8, 0x00000008);
-	nv_wr32(priv, 0x4188e0, 0x01000000);
-	nv_wr32(priv, 0x4188e8, 0x00000000);
-	nv_wr32(priv, 0x4188ec, 0x00000000);
-	nv_wr32(priv, 0x4188f0, 0x00000000);
-	nv_wr32(priv, 0x4188f4, 0x00000000);
-	nv_wr32(priv, 0x4188f8, 0x00000000);
-	if (chipset >= 0xd0)
-		nv_wr32(priv, 0x4188fc, 0x20100008);
-	else if (chipset == 0xc1)
-		nv_wr32(priv, 0x4188fc, 0x00100018);
-	else
-		nv_wr32(priv, 0x4188fc, 0x00100000);
-	nv_wr32(priv, 0x41891c, 0x00ff00ff);
-	nv_wr32(priv, 0x418924, 0x00000000);
-	nv_wr32(priv, 0x418928, 0x00ffff00);
-	nv_wr32(priv, 0x41892c, 0x0000ff00);
-	for (i = 0; i < 8; i++) {
-		nv_wr32(priv, 0x418a00 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a04 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a08 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a0c + (i * 0x20), 0x00010000);
-		nv_wr32(priv, 0x418a10 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a14 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a18 + (i * 0x20), 0x00000000);
-	}
-	nv_wr32(priv, 0x418b00, chipset < 0xd0 ? 0x00000000 : 0x00000006);
-	nv_wr32(priv, 0x418b08, 0x0a418820);
-	nv_wr32(priv, 0x418b0c, 0x062080e6);
-	nv_wr32(priv, 0x418b10, 0x020398a4);
-	nv_wr32(priv, 0x418b14, 0x0e629062);
-	nv_wr32(priv, 0x418b18, 0x0a418820);
-	nv_wr32(priv, 0x418b1c, 0x000000e6);
-	nv_wr32(priv, 0x418bb8, 0x00000103);
-	nv_wr32(priv, 0x418c08, 0x00000001);
-	nv_wr32(priv, 0x418c10, 0x00000000);
-	nv_wr32(priv, 0x418c14, 0x00000000);
-	nv_wr32(priv, 0x418c18, 0x00000000);
-	nv_wr32(priv, 0x418c1c, 0x00000000);
-	nv_wr32(priv, 0x418c20, 0x00000000);
-	nv_wr32(priv, 0x418c24, 0x00000000);
-	nv_wr32(priv, 0x418c28, 0x00000000);
-	nv_wr32(priv, 0x418c2c, 0x00000000);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x418c6c, 0x00000001);
-	nv_wr32(priv, 0x418c80, 0x20200004);
-	nv_wr32(priv, 0x418c8c, 0x00000001);
-	nv_wr32(priv, 0x419000, 0x00000780);
-	nv_wr32(priv, 0x419004, 0x00000000);
-	nv_wr32(priv, 0x419008, 0x00000000);
-	nv_wr32(priv, 0x419014, 0x00000004);
-}
+	oclass->mods(priv, info);
+	oclass->unkn(priv);
 
-static void
-nvc0_grctx_generate_tp(struct nvc0_graph_priv *priv)
-{
-	int chipset = nv_device(priv)->chipset;
+	nvc0_grctx_generate_tpcid(priv);
+	nvc0_grctx_generate_r406028(priv);
+	nvc0_grctx_generate_r4060a8(priv);
+	nvc0_grctx_generate_r418bb8(priv);
+	nvc0_grctx_generate_r406800(priv);
 
-	/* GPC_BROADCAST.TP_BROADCAST */
-	nv_wr32(priv, 0x419818, 0x00000000);
-	nv_wr32(priv, 0x41983c, 0x00038bc7);
-	nv_wr32(priv, 0x419848, 0x00000000);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419864, 0x00000129);
-	else
-		nv_wr32(priv, 0x419864, 0x0000012a);
-	nv_wr32(priv, 0x419888, 0x00000000);
-	nv_wr32(priv, 0x419a00, 0x000001f0);
-	nv_wr32(priv, 0x419a04, 0x00000001);
-	nv_wr32(priv, 0x419a08, 0x00000023);
-	nv_wr32(priv, 0x419a0c, 0x00020000);
-	nv_wr32(priv, 0x419a10, 0x00000000);
-	nv_wr32(priv, 0x419a14, 0x00000200);
-	nv_wr32(priv, 0x419a1c, 0x00000000);
-	nv_wr32(priv, 0x419a20, 0x00000800);
-	if (chipset >= 0xd0)
-		nv_wr32(priv, 0x00419ac4, 0x0017f440);
-	else if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(priv, 0x00419ac4, 0x0007f440);
-	nv_wr32(priv, 0x419b00, 0x0a418820);
-	nv_wr32(priv, 0x419b04, 0x062080e6);
-	nv_wr32(priv, 0x419b08, 0x020398a4);
-	nv_wr32(priv, 0x419b0c, 0x0e629062);
-	nv_wr32(priv, 0x419b10, 0x0a418820);
-	nv_wr32(priv, 0x419b14, 0x000000e6);
-	nv_wr32(priv, 0x419bd0, 0x00900103);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419be0, 0x00400001);
-	else
-		nv_wr32(priv, 0x419be0, 0x00000001);
-	nv_wr32(priv, 0x419be4, 0x00000000);
-	nv_wr32(priv, 0x419c00, chipset < 0xd0 ? 0x00000002 : 0x0000000a);
-	nv_wr32(priv, 0x419c04, 0x00000006);
-	nv_wr32(priv, 0x419c08, 0x00000002);
-	nv_wr32(priv, 0x419c20, 0x00000000);
-	if (nv_device(priv)->chipset >= 0xd0) {
-		nv_wr32(priv, 0x419c24, 0x00084210);
-		nv_wr32(priv, 0x419c28, 0x3cf3cf3c);
-		nv_wr32(priv, 0x419cb0, 0x00020048);
-	} else
-	if (chipset == 0xce || chipset == 0xcf) {
-		nv_wr32(priv, 0x419cb0, 0x00020048);
-	} else {
-		nv_wr32(priv, 0x419cb0, 0x00060048);
-	}
-	nv_wr32(priv, 0x419ce8, 0x00000000);
-	nv_wr32(priv, 0x419cf4, 0x00000183);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419d20, 0x12180000);
-	else
-		nv_wr32(priv, 0x419d20, 0x02180000);
-	nv_wr32(priv, 0x419d24, 0x00001fff);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419d44, 0x02180218);
-	nv_wr32(priv, 0x419e04, 0x00000000);
-	nv_wr32(priv, 0x419e08, 0x00000000);
-	nv_wr32(priv, 0x419e0c, 0x00000000);
-	nv_wr32(priv, 0x419e10, 0x00000002);
-	nv_wr32(priv, 0x419e44, 0x001beff2);
-	nv_wr32(priv, 0x419e48, 0x00000000);
-	nv_wr32(priv, 0x419e4c, 0x0000000f);
-	nv_wr32(priv, 0x419e50, 0x00000000);
-	nv_wr32(priv, 0x419e54, 0x00000000);
-	nv_wr32(priv, 0x419e58, 0x00000000);
-	nv_wr32(priv, 0x419e5c, 0x00000000);
-	nv_wr32(priv, 0x419e60, 0x00000000);
-	nv_wr32(priv, 0x419e64, 0x00000000);
-	nv_wr32(priv, 0x419e68, 0x00000000);
-	nv_wr32(priv, 0x419e6c, 0x00000000);
-	nv_wr32(priv, 0x419e70, 0x00000000);
-	nv_wr32(priv, 0x419e74, 0x00000000);
-	nv_wr32(priv, 0x419e78, 0x00000000);
-	nv_wr32(priv, 0x419e7c, 0x00000000);
-	nv_wr32(priv, 0x419e80, 0x00000000);
-	nv_wr32(priv, 0x419e84, 0x00000000);
-	nv_wr32(priv, 0x419e88, 0x00000000);
-	nv_wr32(priv, 0x419e8c, 0x00000000);
-	nv_wr32(priv, 0x419e90, 0x00000000);
-	nv_wr32(priv, 0x419e98, 0x00000000);
-	if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(priv, 0x419ee0, 0x00011110);
-	nv_wr32(priv, 0x419f50, 0x00000000);
-	nv_wr32(priv, 0x419f54, 0x00000000);
-	if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(priv, 0x419f58, 0x00000000);
+	nvc0_graph_icmd(priv, oclass->icmd);
+	nv_wr32(priv, 0x404154, 0x00000400);
+	nvc0_graph_mthd(priv, oclass->mthd);
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
 }
 
 int
 nvc0_grctx_generate(struct nvc0_graph_priv *priv)
 {
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	struct nouveau_bar *bar = nouveau_bar(priv);
+	struct nouveau_gpuobj *chan;
 	struct nvc0_grctx info;
-	int ret, i, gpc, tpc, id;
-	u32 fermi = nvc0_graph_class(priv);
-	u32 r000260, tmp;
+	int ret, i;
 
-	ret = nvc0_grctx_init(priv, &info);
-	if (ret)
+	/* allocate memory to for a "channel", which we'll use to generate
+	 * the default context values
+	 */
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x80000 + priv->size,
+				 0x1000, NVOBJ_FLAG_ZERO_ALLOC, &chan);
+	if (ret) {
+		nv_error(priv, "failed to allocate channel memory, %d\n", ret);
 		return ret;
-
-	r000260 = nv_rd32(priv, 0x000260);
-	nv_wr32(priv, 0x000260, r000260 & ~1);
-	nv_wr32(priv, 0x400208, 0x00000000);
-
-	nvc0_grctx_generate_dispatch(priv);
-	nvc0_grctx_generate_macro(priv);
-	nvc0_grctx_generate_m2mf(priv);
-	nvc0_grctx_generate_unk47xx(priv);
-	nvc0_grctx_generate_shaders(priv);
-	nvc0_grctx_generate_unk60xx(priv);
-	nvc0_grctx_generate_unk64xx(priv);
-	nvc0_grctx_generate_tpbus(priv);
-	nvc0_grctx_generate_ccache(priv);
-	nvc0_grctx_generate_rop(priv);
-	nvc0_grctx_generate_gpc(priv);
-	nvc0_grctx_generate_tp(priv);
-
-	nv_wr32(priv, 0x404154, 0x00000000);
-
-	/* generate per-context mmio list data */
-	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000018,  0, 0);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000018,  0, 0);
-	if (nv_device(priv)->chipset != 0xc1) {
-		tmp = 0x02180000;
-		mmio_list(0x405830, tmp, 0, 0);
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-				u32 reg = TPC_UNIT(gpc, tpc, 0x0520);
-				mmio_list(reg, tmp, 0, 0);
-				tmp += 0x0324;
-			}
-		}
-	} else {
-		tmp = 0x02180000;
-		mmio_list(0x405830, 0x00000218 | tmp, 0, 0);
-		mmio_list(0x4064c4, 0x0086ffff, 0, 0);
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-				u32 reg = TPC_UNIT(gpc, tpc, 0x0520);
-				mmio_list(reg, 0x10000000 | tmp, 0, 0);
-				tmp += 0x0324;
-			}
-			for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-				u32 reg = TPC_UNIT(gpc, tpc, 0x0544);
-				mmio_list(reg, tmp, 0, 0);
-				tmp += 0x0324;
-			}
-		}
 	}
 
-	for (tpc = 0, id = 0; tpc < 4; tpc++) {
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			if (tpc < priv->tpc_nr[gpc]) {
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x698), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x4e8), id);
-				nv_wr32(priv, GPC_UNIT(gpc, 0x0c10 + tpc * 4), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x088), id);
-				id++;
-			}
-
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c08), priv->tpc_nr[gpc]);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c8c), priv->tpc_nr[gpc]);
-		}
-	}
-
-	tmp = 0;
-	for (i = 0; i < priv->gpc_nr; i++)
-		tmp |= priv->tpc_nr[i] << (i * 4);
-	nv_wr32(priv, 0x406028, tmp);
-	nv_wr32(priv, 0x405870, tmp);
-
-	nv_wr32(priv, 0x40602c, 0x00000000);
-	nv_wr32(priv, 0x405874, 0x00000000);
-	nv_wr32(priv, 0x406030, 0x00000000);
-	nv_wr32(priv, 0x405878, 0x00000000);
-	nv_wr32(priv, 0x406034, 0x00000000);
-	nv_wr32(priv, 0x40587c, 0x00000000);
-
-	if (1) {
-		u8 tpcnr[GPC_MAX], data[TPC_MAX];
-
-		memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-		memset(data, 0x1f, sizeof(data));
+	/* PGD pointer */
+	nv_wo32(chan, 0x0200, lower_32_bits(chan->addr + 0x1000));
+	nv_wo32(chan, 0x0204, upper_32_bits(chan->addr + 0x1000));
+	nv_wo32(chan, 0x0208, 0xffffffff);
+	nv_wo32(chan, 0x020c, 0x000000ff);
 
-		gpc = -1;
-		for (tpc = 0; tpc < priv->tpc_total; tpc++) {
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpcnr[gpc]);
-			tpcnr[gpc]--;
-			data[tpc] = gpc;
-		}
+	/* PGT[0] pointer */
+	nv_wo32(chan, 0x1000, 0x00000000);
+	nv_wo32(chan, 0x1004, 0x00000001 | (chan->addr + 0x2000) >> 8);
 
-		for (i = 0; i < 4; i++)
-			nv_wr32(priv, 0x4060a8 + (i * 4), ((u32 *)data)[i]);
+	/* identity-map the whole "channel" into its own vm */
+	for (i = 0; i < chan->size / 4096; i++) {
+		u64 addr = ((chan->addr + (i * 4096)) >> 8) | 1;
+		nv_wo32(chan, 0x2000 + (i * 8), lower_32_bits(addr));
+		nv_wo32(chan, 0x2004 + (i * 8), upper_32_bits(addr));
 	}
 
-	if (1) {
-		u32 data[6] = {}, data2[2] = {};
-		u8 tpcnr[GPC_MAX];
-		u8 shift, ntpcv;
-
-		/* calculate first set of magics */
-		memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	/* context pointer (virt) */
+	nv_wo32(chan, 0x0210, 0x00080004);
+	nv_wo32(chan, 0x0214, 0x00000000);
 
-		gpc = -1;
-		for (tpc = 0; tpc < priv->tpc_total; tpc++) {
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpcnr[gpc]);
-			tpcnr[gpc]--;
+	bar->flush(bar);
 
-			data[tpc / 6] |= gpc << ((tpc % 6) * 5);
-		}
+	nv_wr32(priv, 0x100cb8, (chan->addr + 0x1000) >> 8);
+	nv_wr32(priv, 0x100cbc, 0x80000001);
+	nv_wait(priv, 0x100c80, 0x00008000, 0x00008000);
 
-		for (; tpc < 32; tpc++)
-			data[tpc / 6] |= 7 << ((tpc % 6) * 5);
+	/* setup default state for mmio list construction */
+	info.priv = priv;
+	info.data = priv->mmio_data;
+	info.mmio = priv->mmio_list;
+	info.addr = 0x2000 + (i * 8);
+	info.buffer_nr = 0;
 
-		/* and the second... */
-		shift = 0;
-		ntpcv = priv->tpc_total;
-		while (!(ntpcv & (1 << 4))) {
-			ntpcv <<= 1;
-			shift++;
-		}
+	/* make channel current */
+	if (priv->firmware) {
+		nv_wr32(priv, 0x409840, 0x00000030);
+		nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
+		nv_wr32(priv, 0x409504, 0x00000003);
+		if (!nv_wait(priv, 0x409800, 0x00000010, 0x00000010))
+			nv_error(priv, "load_ctx timeout\n");
 
-		data2[0]  = (ntpcv << 16);
-		data2[0] |= (shift << 21);
-		data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
-		for (i = 1; i < 7; i++)
-			data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
-
-		/* GPC_BROADCAST */
-		nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) |
-					priv->magic_not_rop_nr);
-		for (i = 0; i < 6; i++)
-			nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
-
-		/* GPC_BROADCAST.TP_BROADCAST */
-		nv_wr32(priv, 0x419bd0, (priv->tpc_total << 8) |
-				       priv->magic_not_rop_nr |
-				       data2[0]);
-		nv_wr32(priv, 0x419be4, data2[1]);
-		for (i = 0; i < 6; i++)
-			nv_wr32(priv, 0x419b00 + (i * 4), data[i]);
-
-		/* UNK78xx */
-		nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) |
-					priv->magic_not_rop_nr);
-		for (i = 0; i < 6; i++)
-			nv_wr32(priv, 0x40780c + (i * 4), data[i]);
+		nv_wo32(chan, 0x8001c, 1);
+		nv_wo32(chan, 0x80020, 0);
+		nv_wo32(chan, 0x80028, 0);
+		nv_wo32(chan, 0x8002c, 0);
+		bar->flush(bar);
+	} else {
+		nv_wr32(priv, 0x409840, 0x80000000);
+		nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
+		nv_wr32(priv, 0x409504, 0x00000001);
+		if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000))
+			nv_error(priv, "HUB_SET_CHAN timeout\n");
 	}
 
-	if (1) {
-		u32 tpc_mask = 0, tpc_set = 0;
-		u8  tpcnr[GPC_MAX], a, b;
-
-		memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++)
-			tpc_mask |= ((1 << priv->tpc_nr[gpc]) - 1) << (gpc * 8);
-
-		for (i = 0, gpc = -1, b = -1; i < 32; i++) {
-			a = (i * (priv->tpc_total - 1)) / 32;
-			if (a != b) {
-				b = a;
-				do {
-					gpc = (gpc + 1) % priv->gpc_nr;
-				} while (!tpcnr[gpc]);
-				tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-				tpc_set |= 1 << ((gpc * 8) + tpc);
-			}
+	oclass->main(priv, &info);
 
-			nv_wr32(priv, 0x406800 + (i * 0x20), tpc_set);
-			nv_wr32(priv, 0x406c00 + (i * 0x20), tpc_set ^ tpc_mask);
-		}
+	/* trigger a context unload by unsetting the "next channel valid" bit
+	 * and faking a context switch interrupt
+	 */
+	nv_mask(priv, 0x409b04, 0x80000000, 0x00000000);
+	nv_wr32(priv, 0x409000, 0x00000100);
+	if (!nv_wait(priv, 0x409b00, 0x80000000, 0x00000000)) {
+		nv_error(priv, "grctx template channel unload timeout\n");
+		ret = -EBUSY;
+		goto done;
 	}
 
-	nv_wr32(priv, 0x400208, 0x80000000);
-
-	nv_icmd(priv, 0x00001000, 0x00000004);
-	nv_icmd(priv, 0x000000a9, 0x0000ffff);
-	nv_icmd(priv, 0x00000038, 0x0fac6881);
-	nv_icmd(priv, 0x0000003d, 0x00000001);
-	nv_icmd(priv, 0x000000e8, 0x00000400);
-	nv_icmd(priv, 0x000000e9, 0x00000400);
-	nv_icmd(priv, 0x000000ea, 0x00000400);
-	nv_icmd(priv, 0x000000eb, 0x00000400);
-	nv_icmd(priv, 0x000000ec, 0x00000400);
-	nv_icmd(priv, 0x000000ed, 0x00000400);
-	nv_icmd(priv, 0x000000ee, 0x00000400);
-	nv_icmd(priv, 0x000000ef, 0x00000400);
-	nv_icmd(priv, 0x00000078, 0x00000300);
-	nv_icmd(priv, 0x00000079, 0x00000300);
-	nv_icmd(priv, 0x0000007a, 0x00000300);
-	nv_icmd(priv, 0x0000007b, 0x00000300);
-	nv_icmd(priv, 0x0000007c, 0x00000300);
-	nv_icmd(priv, 0x0000007d, 0x00000300);
-	nv_icmd(priv, 0x0000007e, 0x00000300);
-	nv_icmd(priv, 0x0000007f, 0x00000300);
-	nv_icmd(priv, 0x00000050, 0x00000011);
-	nv_icmd(priv, 0x00000058, 0x00000008);
-	nv_icmd(priv, 0x00000059, 0x00000008);
-	nv_icmd(priv, 0x0000005a, 0x00000008);
-	nv_icmd(priv, 0x0000005b, 0x00000008);
-	nv_icmd(priv, 0x0000005c, 0x00000008);
-	nv_icmd(priv, 0x0000005d, 0x00000008);
-	nv_icmd(priv, 0x0000005e, 0x00000008);
-	nv_icmd(priv, 0x0000005f, 0x00000008);
-	nv_icmd(priv, 0x00000208, 0x00000001);
-	nv_icmd(priv, 0x00000209, 0x00000001);
-	nv_icmd(priv, 0x0000020a, 0x00000001);
-	nv_icmd(priv, 0x0000020b, 0x00000001);
-	nv_icmd(priv, 0x0000020c, 0x00000001);
-	nv_icmd(priv, 0x0000020d, 0x00000001);
-	nv_icmd(priv, 0x0000020e, 0x00000001);
-	nv_icmd(priv, 0x0000020f, 0x00000001);
-	nv_icmd(priv, 0x00000081, 0x00000001);
-	nv_icmd(priv, 0x00000085, 0x00000004);
-	nv_icmd(priv, 0x00000088, 0x00000400);
-	nv_icmd(priv, 0x00000090, 0x00000300);
-	nv_icmd(priv, 0x00000098, 0x00001001);
-	nv_icmd(priv, 0x000000e3, 0x00000001);
-	nv_icmd(priv, 0x000000da, 0x00000001);
-	nv_icmd(priv, 0x000000f8, 0x00000003);
-	nv_icmd(priv, 0x000000fa, 0x00000001);
-	nv_icmd(priv, 0x0000009f, 0x0000ffff);
-	nv_icmd(priv, 0x000000a0, 0x0000ffff);
-	nv_icmd(priv, 0x000000a1, 0x0000ffff);
-	nv_icmd(priv, 0x000000a2, 0x0000ffff);
-	nv_icmd(priv, 0x000000b1, 0x00000001);
-	nv_icmd(priv, 0x000000b2, 0x00000000);
-	nv_icmd(priv, 0x000000b3, 0x00000000);
-	nv_icmd(priv, 0x000000b4, 0x00000000);
-	nv_icmd(priv, 0x000000b5, 0x00000000);
-	nv_icmd(priv, 0x000000b6, 0x00000000);
-	nv_icmd(priv, 0x000000b7, 0x00000000);
-	nv_icmd(priv, 0x000000b8, 0x00000000);
-	nv_icmd(priv, 0x000000b9, 0x00000000);
-	nv_icmd(priv, 0x000000ba, 0x00000000);
-	nv_icmd(priv, 0x000000bb, 0x00000000);
-	nv_icmd(priv, 0x000000bc, 0x00000000);
-	nv_icmd(priv, 0x000000bd, 0x00000000);
-	nv_icmd(priv, 0x000000be, 0x00000000);
-	nv_icmd(priv, 0x000000bf, 0x00000000);
-	nv_icmd(priv, 0x000000c0, 0x00000000);
-	nv_icmd(priv, 0x000000c1, 0x00000000);
-	nv_icmd(priv, 0x000000c2, 0x00000000);
-	nv_icmd(priv, 0x000000c3, 0x00000000);
-	nv_icmd(priv, 0x000000c4, 0x00000000);
-	nv_icmd(priv, 0x000000c5, 0x00000000);
-	nv_icmd(priv, 0x000000c6, 0x00000000);
-	nv_icmd(priv, 0x000000c7, 0x00000000);
-	nv_icmd(priv, 0x000000c8, 0x00000000);
-	nv_icmd(priv, 0x000000c9, 0x00000000);
-	nv_icmd(priv, 0x000000ca, 0x00000000);
-	nv_icmd(priv, 0x000000cb, 0x00000000);
-	nv_icmd(priv, 0x000000cc, 0x00000000);
-	nv_icmd(priv, 0x000000cd, 0x00000000);
-	nv_icmd(priv, 0x000000ce, 0x00000000);
-	nv_icmd(priv, 0x000000cf, 0x00000000);
-	nv_icmd(priv, 0x000000d0, 0x00000000);
-	nv_icmd(priv, 0x000000d1, 0x00000000);
-	nv_icmd(priv, 0x000000d2, 0x00000000);
-	nv_icmd(priv, 0x000000d3, 0x00000000);
-	nv_icmd(priv, 0x000000d4, 0x00000000);
-	nv_icmd(priv, 0x000000d5, 0x00000000);
-	nv_icmd(priv, 0x000000d6, 0x00000000);
-	nv_icmd(priv, 0x000000d7, 0x00000000);
-	nv_icmd(priv, 0x000000d8, 0x00000000);
-	nv_icmd(priv, 0x000000d9, 0x00000000);
-	nv_icmd(priv, 0x00000210, 0x00000040);
-	nv_icmd(priv, 0x00000211, 0x00000040);
-	nv_icmd(priv, 0x00000212, 0x00000040);
-	nv_icmd(priv, 0x00000213, 0x00000040);
-	nv_icmd(priv, 0x00000214, 0x00000040);
-	nv_icmd(priv, 0x00000215, 0x00000040);
-	nv_icmd(priv, 0x00000216, 0x00000040);
-	nv_icmd(priv, 0x00000217, 0x00000040);
-	if (nv_device(priv)->chipset >= 0xd0) {
-		for (i = 0x0400; i <= 0x0417; i++)
-			nv_icmd(priv, i, 0x00000040);
-	}
-	nv_icmd(priv, 0x00000218, 0x0000c080);
-	nv_icmd(priv, 0x00000219, 0x0000c080);
-	nv_icmd(priv, 0x0000021a, 0x0000c080);
-	nv_icmd(priv, 0x0000021b, 0x0000c080);
-	nv_icmd(priv, 0x0000021c, 0x0000c080);
-	nv_icmd(priv, 0x0000021d, 0x0000c080);
-	nv_icmd(priv, 0x0000021e, 0x0000c080);
-	nv_icmd(priv, 0x0000021f, 0x0000c080);
-	if (nv_device(priv)->chipset >= 0xd0) {
-		for (i = 0x0440; i <= 0x0457; i++)
-			nv_icmd(priv, i, 0x0000c080);
+	priv->data = kmalloc(priv->size, GFP_KERNEL);
+	if (priv->data) {
+		for (i = 0; i < priv->size; i += 4)
+			priv->data[i / 4] = nv_ro32(chan, 0x80000 + i);
+		ret = 0;
+	} else {
+		ret = -ENOMEM;
 	}
-	nv_icmd(priv, 0x000000ad, 0x0000013e);
-	nv_icmd(priv, 0x000000e1, 0x00000010);
-	nv_icmd(priv, 0x00000290, 0x00000000);
-	nv_icmd(priv, 0x00000291, 0x00000000);
-	nv_icmd(priv, 0x00000292, 0x00000000);
-	nv_icmd(priv, 0x00000293, 0x00000000);
-	nv_icmd(priv, 0x00000294, 0x00000000);
-	nv_icmd(priv, 0x00000295, 0x00000000);
-	nv_icmd(priv, 0x00000296, 0x00000000);
-	nv_icmd(priv, 0x00000297, 0x00000000);
-	nv_icmd(priv, 0x00000298, 0x00000000);
-	nv_icmd(priv, 0x00000299, 0x00000000);
-	nv_icmd(priv, 0x0000029a, 0x00000000);
-	nv_icmd(priv, 0x0000029b, 0x00000000);
-	nv_icmd(priv, 0x0000029c, 0x00000000);
-	nv_icmd(priv, 0x0000029d, 0x00000000);
-	nv_icmd(priv, 0x0000029e, 0x00000000);
-	nv_icmd(priv, 0x0000029f, 0x00000000);
-	nv_icmd(priv, 0x000003b0, 0x00000000);
-	nv_icmd(priv, 0x000003b1, 0x00000000);
-	nv_icmd(priv, 0x000003b2, 0x00000000);
-	nv_icmd(priv, 0x000003b3, 0x00000000);
-	nv_icmd(priv, 0x000003b4, 0x00000000);
-	nv_icmd(priv, 0x000003b5, 0x00000000);
-	nv_icmd(priv, 0x000003b6, 0x00000000);
-	nv_icmd(priv, 0x000003b7, 0x00000000);
-	nv_icmd(priv, 0x000003b8, 0x00000000);
-	nv_icmd(priv, 0x000003b9, 0x00000000);
-	nv_icmd(priv, 0x000003ba, 0x00000000);
-	nv_icmd(priv, 0x000003bb, 0x00000000);
-	nv_icmd(priv, 0x000003bc, 0x00000000);
-	nv_icmd(priv, 0x000003bd, 0x00000000);
-	nv_icmd(priv, 0x000003be, 0x00000000);
-	nv_icmd(priv, 0x000003bf, 0x00000000);
-	nv_icmd(priv, 0x000002a0, 0x00000000);
-	nv_icmd(priv, 0x000002a1, 0x00000000);
-	nv_icmd(priv, 0x000002a2, 0x00000000);
-	nv_icmd(priv, 0x000002a3, 0x00000000);
-	nv_icmd(priv, 0x000002a4, 0x00000000);
-	nv_icmd(priv, 0x000002a5, 0x00000000);
-	nv_icmd(priv, 0x000002a6, 0x00000000);
-	nv_icmd(priv, 0x000002a7, 0x00000000);
-	nv_icmd(priv, 0x000002a8, 0x00000000);
-	nv_icmd(priv, 0x000002a9, 0x00000000);
-	nv_icmd(priv, 0x000002aa, 0x00000000);
-	nv_icmd(priv, 0x000002ab, 0x00000000);
-	nv_icmd(priv, 0x000002ac, 0x00000000);
-	nv_icmd(priv, 0x000002ad, 0x00000000);
-	nv_icmd(priv, 0x000002ae, 0x00000000);
-	nv_icmd(priv, 0x000002af, 0x00000000);
-	nv_icmd(priv, 0x00000420, 0x00000000);
-	nv_icmd(priv, 0x00000421, 0x00000000);
-	nv_icmd(priv, 0x00000422, 0x00000000);
-	nv_icmd(priv, 0x00000423, 0x00000000);
-	nv_icmd(priv, 0x00000424, 0x00000000);
-	nv_icmd(priv, 0x00000425, 0x00000000);
-	nv_icmd(priv, 0x00000426, 0x00000000);
-	nv_icmd(priv, 0x00000427, 0x00000000);
-	nv_icmd(priv, 0x00000428, 0x00000000);
-	nv_icmd(priv, 0x00000429, 0x00000000);
-	nv_icmd(priv, 0x0000042a, 0x00000000);
-	nv_icmd(priv, 0x0000042b, 0x00000000);
-	nv_icmd(priv, 0x0000042c, 0x00000000);
-	nv_icmd(priv, 0x0000042d, 0x00000000);
-	nv_icmd(priv, 0x0000042e, 0x00000000);
-	nv_icmd(priv, 0x0000042f, 0x00000000);
-	nv_icmd(priv, 0x000002b0, 0x00000000);
-	nv_icmd(priv, 0x000002b1, 0x00000000);
-	nv_icmd(priv, 0x000002b2, 0x00000000);
-	nv_icmd(priv, 0x000002b3, 0x00000000);
-	nv_icmd(priv, 0x000002b4, 0x00000000);
-	nv_icmd(priv, 0x000002b5, 0x00000000);
-	nv_icmd(priv, 0x000002b6, 0x00000000);
-	nv_icmd(priv, 0x000002b7, 0x00000000);
-	nv_icmd(priv, 0x000002b8, 0x00000000);
-	nv_icmd(priv, 0x000002b9, 0x00000000);
-	nv_icmd(priv, 0x000002ba, 0x00000000);
-	nv_icmd(priv, 0x000002bb, 0x00000000);
-	nv_icmd(priv, 0x000002bc, 0x00000000);
-	nv_icmd(priv, 0x000002bd, 0x00000000);
-	nv_icmd(priv, 0x000002be, 0x00000000);
-	nv_icmd(priv, 0x000002bf, 0x00000000);
-	nv_icmd(priv, 0x00000430, 0x00000000);
-	nv_icmd(priv, 0x00000431, 0x00000000);
-	nv_icmd(priv, 0x00000432, 0x00000000);
-	nv_icmd(priv, 0x00000433, 0x00000000);
-	nv_icmd(priv, 0x00000434, 0x00000000);
-	nv_icmd(priv, 0x00000435, 0x00000000);
-	nv_icmd(priv, 0x00000436, 0x00000000);
-	nv_icmd(priv, 0x00000437, 0x00000000);
-	nv_icmd(priv, 0x00000438, 0x00000000);
-	nv_icmd(priv, 0x00000439, 0x00000000);
-	nv_icmd(priv, 0x0000043a, 0x00000000);
-	nv_icmd(priv, 0x0000043b, 0x00000000);
-	nv_icmd(priv, 0x0000043c, 0x00000000);
-	nv_icmd(priv, 0x0000043d, 0x00000000);
-	nv_icmd(priv, 0x0000043e, 0x00000000);
-	nv_icmd(priv, 0x0000043f, 0x00000000);
-	nv_icmd(priv, 0x000002c0, 0x00000000);
-	nv_icmd(priv, 0x000002c1, 0x00000000);
-	nv_icmd(priv, 0x000002c2, 0x00000000);
-	nv_icmd(priv, 0x000002c3, 0x00000000);
-	nv_icmd(priv, 0x000002c4, 0x00000000);
-	nv_icmd(priv, 0x000002c5, 0x00000000);
-	nv_icmd(priv, 0x000002c6, 0x00000000);
-	nv_icmd(priv, 0x000002c7, 0x00000000);
-	nv_icmd(priv, 0x000002c8, 0x00000000);
-	nv_icmd(priv, 0x000002c9, 0x00000000);
-	nv_icmd(priv, 0x000002ca, 0x00000000);
-	nv_icmd(priv, 0x000002cb, 0x00000000);
-	nv_icmd(priv, 0x000002cc, 0x00000000);
-	nv_icmd(priv, 0x000002cd, 0x00000000);
-	nv_icmd(priv, 0x000002ce, 0x00000000);
-	nv_icmd(priv, 0x000002cf, 0x00000000);
-	nv_icmd(priv, 0x000004d0, 0x00000000);
-	nv_icmd(priv, 0x000004d1, 0x00000000);
-	nv_icmd(priv, 0x000004d2, 0x00000000);
-	nv_icmd(priv, 0x000004d3, 0x00000000);
-	nv_icmd(priv, 0x000004d4, 0x00000000);
-	nv_icmd(priv, 0x000004d5, 0x00000000);
-	nv_icmd(priv, 0x000004d6, 0x00000000);
-	nv_icmd(priv, 0x000004d7, 0x00000000);
-	nv_icmd(priv, 0x000004d8, 0x00000000);
-	nv_icmd(priv, 0x000004d9, 0x00000000);
-	nv_icmd(priv, 0x000004da, 0x00000000);
-	nv_icmd(priv, 0x000004db, 0x00000000);
-	nv_icmd(priv, 0x000004dc, 0x00000000);
-	nv_icmd(priv, 0x000004dd, 0x00000000);
-	nv_icmd(priv, 0x000004de, 0x00000000);
-	nv_icmd(priv, 0x000004df, 0x00000000);
-	nv_icmd(priv, 0x00000720, 0x00000000);
-	nv_icmd(priv, 0x00000721, 0x00000000);
-	nv_icmd(priv, 0x00000722, 0x00000000);
-	nv_icmd(priv, 0x00000723, 0x00000000);
-	nv_icmd(priv, 0x00000724, 0x00000000);
-	nv_icmd(priv, 0x00000725, 0x00000000);
-	nv_icmd(priv, 0x00000726, 0x00000000);
-	nv_icmd(priv, 0x00000727, 0x00000000);
-	nv_icmd(priv, 0x00000728, 0x00000000);
-	nv_icmd(priv, 0x00000729, 0x00000000);
-	nv_icmd(priv, 0x0000072a, 0x00000000);
-	nv_icmd(priv, 0x0000072b, 0x00000000);
-	nv_icmd(priv, 0x0000072c, 0x00000000);
-	nv_icmd(priv, 0x0000072d, 0x00000000);
-	nv_icmd(priv, 0x0000072e, 0x00000000);
-	nv_icmd(priv, 0x0000072f, 0x00000000);
-	nv_icmd(priv, 0x000008c0, 0x00000000);
-	nv_icmd(priv, 0x000008c1, 0x00000000);
-	nv_icmd(priv, 0x000008c2, 0x00000000);
-	nv_icmd(priv, 0x000008c3, 0x00000000);
-	nv_icmd(priv, 0x000008c4, 0x00000000);
-	nv_icmd(priv, 0x000008c5, 0x00000000);
-	nv_icmd(priv, 0x000008c6, 0x00000000);
-	nv_icmd(priv, 0x000008c7, 0x00000000);
-	nv_icmd(priv, 0x000008c8, 0x00000000);
-	nv_icmd(priv, 0x000008c9, 0x00000000);
-	nv_icmd(priv, 0x000008ca, 0x00000000);
-	nv_icmd(priv, 0x000008cb, 0x00000000);
-	nv_icmd(priv, 0x000008cc, 0x00000000);
-	nv_icmd(priv, 0x000008cd, 0x00000000);
-	nv_icmd(priv, 0x000008ce, 0x00000000);
-	nv_icmd(priv, 0x000008cf, 0x00000000);
-	nv_icmd(priv, 0x00000890, 0x00000000);
-	nv_icmd(priv, 0x00000891, 0x00000000);
-	nv_icmd(priv, 0x00000892, 0x00000000);
-	nv_icmd(priv, 0x00000893, 0x00000000);
-	nv_icmd(priv, 0x00000894, 0x00000000);
-	nv_icmd(priv, 0x00000895, 0x00000000);
-	nv_icmd(priv, 0x00000896, 0x00000000);
-	nv_icmd(priv, 0x00000897, 0x00000000);
-	nv_icmd(priv, 0x00000898, 0x00000000);
-	nv_icmd(priv, 0x00000899, 0x00000000);
-	nv_icmd(priv, 0x0000089a, 0x00000000);
-	nv_icmd(priv, 0x0000089b, 0x00000000);
-	nv_icmd(priv, 0x0000089c, 0x00000000);
-	nv_icmd(priv, 0x0000089d, 0x00000000);
-	nv_icmd(priv, 0x0000089e, 0x00000000);
-	nv_icmd(priv, 0x0000089f, 0x00000000);
-	nv_icmd(priv, 0x000008e0, 0x00000000);
-	nv_icmd(priv, 0x000008e1, 0x00000000);
-	nv_icmd(priv, 0x000008e2, 0x00000000);
-	nv_icmd(priv, 0x000008e3, 0x00000000);
-	nv_icmd(priv, 0x000008e4, 0x00000000);
-	nv_icmd(priv, 0x000008e5, 0x00000000);
-	nv_icmd(priv, 0x000008e6, 0x00000000);
-	nv_icmd(priv, 0x000008e7, 0x00000000);
-	nv_icmd(priv, 0x000008e8, 0x00000000);
-	nv_icmd(priv, 0x000008e9, 0x00000000);
-	nv_icmd(priv, 0x000008ea, 0x00000000);
-	nv_icmd(priv, 0x000008eb, 0x00000000);
-	nv_icmd(priv, 0x000008ec, 0x00000000);
-	nv_icmd(priv, 0x000008ed, 0x00000000);
-	nv_icmd(priv, 0x000008ee, 0x00000000);
-	nv_icmd(priv, 0x000008ef, 0x00000000);
-	nv_icmd(priv, 0x000008a0, 0x00000000);
-	nv_icmd(priv, 0x000008a1, 0x00000000);
-	nv_icmd(priv, 0x000008a2, 0x00000000);
-	nv_icmd(priv, 0x000008a3, 0x00000000);
-	nv_icmd(priv, 0x000008a4, 0x00000000);
-	nv_icmd(priv, 0x000008a5, 0x00000000);
-	nv_icmd(priv, 0x000008a6, 0x00000000);
-	nv_icmd(priv, 0x000008a7, 0x00000000);
-	nv_icmd(priv, 0x000008a8, 0x00000000);
-	nv_icmd(priv, 0x000008a9, 0x00000000);
-	nv_icmd(priv, 0x000008aa, 0x00000000);
-	nv_icmd(priv, 0x000008ab, 0x00000000);
-	nv_icmd(priv, 0x000008ac, 0x00000000);
-	nv_icmd(priv, 0x000008ad, 0x00000000);
-	nv_icmd(priv, 0x000008ae, 0x00000000);
-	nv_icmd(priv, 0x000008af, 0x00000000);
-	nv_icmd(priv, 0x000008f0, 0x00000000);
-	nv_icmd(priv, 0x000008f1, 0x00000000);
-	nv_icmd(priv, 0x000008f2, 0x00000000);
-	nv_icmd(priv, 0x000008f3, 0x00000000);
-	nv_icmd(priv, 0x000008f4, 0x00000000);
-	nv_icmd(priv, 0x000008f5, 0x00000000);
-	nv_icmd(priv, 0x000008f6, 0x00000000);
-	nv_icmd(priv, 0x000008f7, 0x00000000);
-	nv_icmd(priv, 0x000008f8, 0x00000000);
-	nv_icmd(priv, 0x000008f9, 0x00000000);
-	nv_icmd(priv, 0x000008fa, 0x00000000);
-	nv_icmd(priv, 0x000008fb, 0x00000000);
-	nv_icmd(priv, 0x000008fc, 0x00000000);
-	nv_icmd(priv, 0x000008fd, 0x00000000);
-	nv_icmd(priv, 0x000008fe, 0x00000000);
-	nv_icmd(priv, 0x000008ff, 0x00000000);
-	nv_icmd(priv, 0x0000094c, 0x000000ff);
-	nv_icmd(priv, 0x0000094d, 0xffffffff);
-	nv_icmd(priv, 0x0000094e, 0x00000002);
-	nv_icmd(priv, 0x000002ec, 0x00000001);
-	nv_icmd(priv, 0x00000303, 0x00000001);
-	nv_icmd(priv, 0x000002e6, 0x00000001);
-	nv_icmd(priv, 0x00000466, 0x00000052);
-	nv_icmd(priv, 0x00000301, 0x3f800000);
-	nv_icmd(priv, 0x00000304, 0x30201000);
-	nv_icmd(priv, 0x00000305, 0x70605040);
-	nv_icmd(priv, 0x00000306, 0xb8a89888);
-	nv_icmd(priv, 0x00000307, 0xf8e8d8c8);
-	nv_icmd(priv, 0x0000030a, 0x00ffff00);
-	nv_icmd(priv, 0x0000030b, 0x0000001a);
-	nv_icmd(priv, 0x0000030c, 0x00000001);
-	nv_icmd(priv, 0x00000318, 0x00000001);
-	nv_icmd(priv, 0x00000340, 0x00000000);
-	nv_icmd(priv, 0x00000375, 0x00000001);
-	nv_icmd(priv, 0x00000351, 0x00000100);
-	nv_icmd(priv, 0x0000037d, 0x00000006);
-	nv_icmd(priv, 0x000003a0, 0x00000002);
-	nv_icmd(priv, 0x000003aa, 0x00000001);
-	nv_icmd(priv, 0x000003a9, 0x00000001);
-	nv_icmd(priv, 0x00000380, 0x00000001);
-	nv_icmd(priv, 0x00000360, 0x00000040);
-	nv_icmd(priv, 0x00000366, 0x00000000);
-	nv_icmd(priv, 0x00000367, 0x00000000);
-	nv_icmd(priv, 0x00000368, 0x00001fff);
-	nv_icmd(priv, 0x00000370, 0x00000000);
-	nv_icmd(priv, 0x00000371, 0x00000000);
-	nv_icmd(priv, 0x00000372, 0x003fffff);
-	nv_icmd(priv, 0x0000037a, 0x00000012);
-	nv_icmd(priv, 0x000005e0, 0x00000022);
-	nv_icmd(priv, 0x000005e1, 0x00000022);
-	nv_icmd(priv, 0x000005e2, 0x00000022);
-	nv_icmd(priv, 0x000005e3, 0x00000022);
-	nv_icmd(priv, 0x000005e4, 0x00000022);
-	nv_icmd(priv, 0x00000619, 0x00000003);
-	nv_icmd(priv, 0x00000811, 0x00000003);
-	nv_icmd(priv, 0x00000812, 0x00000004);
-	nv_icmd(priv, 0x00000813, 0x00000006);
-	nv_icmd(priv, 0x00000814, 0x00000008);
-	nv_icmd(priv, 0x00000815, 0x0000000b);
-	nv_icmd(priv, 0x00000800, 0x00000001);
-	nv_icmd(priv, 0x00000801, 0x00000001);
-	nv_icmd(priv, 0x00000802, 0x00000001);
-	nv_icmd(priv, 0x00000803, 0x00000001);
-	nv_icmd(priv, 0x00000804, 0x00000001);
-	nv_icmd(priv, 0x00000805, 0x00000001);
-	nv_icmd(priv, 0x00000632, 0x00000001);
-	nv_icmd(priv, 0x00000633, 0x00000002);
-	nv_icmd(priv, 0x00000634, 0x00000003);
-	nv_icmd(priv, 0x00000635, 0x00000004);
-	nv_icmd(priv, 0x00000654, 0x3f800000);
-	nv_icmd(priv, 0x00000657, 0x3f800000);
-	nv_icmd(priv, 0x00000655, 0x3f800000);
-	nv_icmd(priv, 0x00000656, 0x3f800000);
-	nv_icmd(priv, 0x000006cd, 0x3f800000);
-	nv_icmd(priv, 0x000007f5, 0x3f800000);
-	nv_icmd(priv, 0x000007dc, 0x39291909);
-	nv_icmd(priv, 0x000007dd, 0x79695949);
-	nv_icmd(priv, 0x000007de, 0xb9a99989);
-	nv_icmd(priv, 0x000007df, 0xf9e9d9c9);
-	nv_icmd(priv, 0x000007e8, 0x00003210);
-	nv_icmd(priv, 0x000007e9, 0x00007654);
-	nv_icmd(priv, 0x000007ea, 0x00000098);
-	nv_icmd(priv, 0x000007ec, 0x39291909);
-	nv_icmd(priv, 0x000007ed, 0x79695949);
-	nv_icmd(priv, 0x000007ee, 0xb9a99989);
-	nv_icmd(priv, 0x000007ef, 0xf9e9d9c9);
-	nv_icmd(priv, 0x000007f0, 0x00003210);
-	nv_icmd(priv, 0x000007f1, 0x00007654);
-	nv_icmd(priv, 0x000007f2, 0x00000098);
-	nv_icmd(priv, 0x000005a5, 0x00000001);
-	nv_icmd(priv, 0x00000980, 0x00000000);
-	nv_icmd(priv, 0x00000981, 0x00000000);
-	nv_icmd(priv, 0x00000982, 0x00000000);
-	nv_icmd(priv, 0x00000983, 0x00000000);
-	nv_icmd(priv, 0x00000984, 0x00000000);
-	nv_icmd(priv, 0x00000985, 0x00000000);
-	nv_icmd(priv, 0x00000986, 0x00000000);
-	nv_icmd(priv, 0x00000987, 0x00000000);
-	nv_icmd(priv, 0x00000988, 0x00000000);
-	nv_icmd(priv, 0x00000989, 0x00000000);
-	nv_icmd(priv, 0x0000098a, 0x00000000);
-	nv_icmd(priv, 0x0000098b, 0x00000000);
-	nv_icmd(priv, 0x0000098c, 0x00000000);
-	nv_icmd(priv, 0x0000098d, 0x00000000);
-	nv_icmd(priv, 0x0000098e, 0x00000000);
-	nv_icmd(priv, 0x0000098f, 0x00000000);
-	nv_icmd(priv, 0x00000990, 0x00000000);
-	nv_icmd(priv, 0x00000991, 0x00000000);
-	nv_icmd(priv, 0x00000992, 0x00000000);
-	nv_icmd(priv, 0x00000993, 0x00000000);
-	nv_icmd(priv, 0x00000994, 0x00000000);
-	nv_icmd(priv, 0x00000995, 0x00000000);
-	nv_icmd(priv, 0x00000996, 0x00000000);
-	nv_icmd(priv, 0x00000997, 0x00000000);
-	nv_icmd(priv, 0x00000998, 0x00000000);
-	nv_icmd(priv, 0x00000999, 0x00000000);
-	nv_icmd(priv, 0x0000099a, 0x00000000);
-	nv_icmd(priv, 0x0000099b, 0x00000000);
-	nv_icmd(priv, 0x0000099c, 0x00000000);
-	nv_icmd(priv, 0x0000099d, 0x00000000);
-	nv_icmd(priv, 0x0000099e, 0x00000000);
-	nv_icmd(priv, 0x0000099f, 0x00000000);
-	nv_icmd(priv, 0x000009a0, 0x00000000);
-	nv_icmd(priv, 0x000009a1, 0x00000000);
-	nv_icmd(priv, 0x000009a2, 0x00000000);
-	nv_icmd(priv, 0x000009a3, 0x00000000);
-	nv_icmd(priv, 0x000009a4, 0x00000000);
-	nv_icmd(priv, 0x000009a5, 0x00000000);
-	nv_icmd(priv, 0x000009a6, 0x00000000);
-	nv_icmd(priv, 0x000009a7, 0x00000000);
-	nv_icmd(priv, 0x000009a8, 0x00000000);
-	nv_icmd(priv, 0x000009a9, 0x00000000);
-	nv_icmd(priv, 0x000009aa, 0x00000000);
-	nv_icmd(priv, 0x000009ab, 0x00000000);
-	nv_icmd(priv, 0x000009ac, 0x00000000);
-	nv_icmd(priv, 0x000009ad, 0x00000000);
-	nv_icmd(priv, 0x000009ae, 0x00000000);
-	nv_icmd(priv, 0x000009af, 0x00000000);
-	nv_icmd(priv, 0x000009b0, 0x00000000);
-	nv_icmd(priv, 0x000009b1, 0x00000000);
-	nv_icmd(priv, 0x000009b2, 0x00000000);
-	nv_icmd(priv, 0x000009b3, 0x00000000);
-	nv_icmd(priv, 0x000009b4, 0x00000000);
-	nv_icmd(priv, 0x000009b5, 0x00000000);
-	nv_icmd(priv, 0x000009b6, 0x00000000);
-	nv_icmd(priv, 0x000009b7, 0x00000000);
-	nv_icmd(priv, 0x000009b8, 0x00000000);
-	nv_icmd(priv, 0x000009b9, 0x00000000);
-	nv_icmd(priv, 0x000009ba, 0x00000000);
-	nv_icmd(priv, 0x000009bb, 0x00000000);
-	nv_icmd(priv, 0x000009bc, 0x00000000);
-	nv_icmd(priv, 0x000009bd, 0x00000000);
-	nv_icmd(priv, 0x000009be, 0x00000000);
-	nv_icmd(priv, 0x000009bf, 0x00000000);
-	nv_icmd(priv, 0x000009c0, 0x00000000);
-	nv_icmd(priv, 0x000009c1, 0x00000000);
-	nv_icmd(priv, 0x000009c2, 0x00000000);
-	nv_icmd(priv, 0x000009c3, 0x00000000);
-	nv_icmd(priv, 0x000009c4, 0x00000000);
-	nv_icmd(priv, 0x000009c5, 0x00000000);
-	nv_icmd(priv, 0x000009c6, 0x00000000);
-	nv_icmd(priv, 0x000009c7, 0x00000000);
-	nv_icmd(priv, 0x000009c8, 0x00000000);
-	nv_icmd(priv, 0x000009c9, 0x00000000);
-	nv_icmd(priv, 0x000009ca, 0x00000000);
-	nv_icmd(priv, 0x000009cb, 0x00000000);
-	nv_icmd(priv, 0x000009cc, 0x00000000);
-	nv_icmd(priv, 0x000009cd, 0x00000000);
-	nv_icmd(priv, 0x000009ce, 0x00000000);
-	nv_icmd(priv, 0x000009cf, 0x00000000);
-	nv_icmd(priv, 0x000009d0, 0x00000000);
-	nv_icmd(priv, 0x000009d1, 0x00000000);
-	nv_icmd(priv, 0x000009d2, 0x00000000);
-	nv_icmd(priv, 0x000009d3, 0x00000000);
-	nv_icmd(priv, 0x000009d4, 0x00000000);
-	nv_icmd(priv, 0x000009d5, 0x00000000);
-	nv_icmd(priv, 0x000009d6, 0x00000000);
-	nv_icmd(priv, 0x000009d7, 0x00000000);
-	nv_icmd(priv, 0x000009d8, 0x00000000);
-	nv_icmd(priv, 0x000009d9, 0x00000000);
-	nv_icmd(priv, 0x000009da, 0x00000000);
-	nv_icmd(priv, 0x000009db, 0x00000000);
-	nv_icmd(priv, 0x000009dc, 0x00000000);
-	nv_icmd(priv, 0x000009dd, 0x00000000);
-	nv_icmd(priv, 0x000009de, 0x00000000);
-	nv_icmd(priv, 0x000009df, 0x00000000);
-	nv_icmd(priv, 0x000009e0, 0x00000000);
-	nv_icmd(priv, 0x000009e1, 0x00000000);
-	nv_icmd(priv, 0x000009e2, 0x00000000);
-	nv_icmd(priv, 0x000009e3, 0x00000000);
-	nv_icmd(priv, 0x000009e4, 0x00000000);
-	nv_icmd(priv, 0x000009e5, 0x00000000);
-	nv_icmd(priv, 0x000009e6, 0x00000000);
-	nv_icmd(priv, 0x000009e7, 0x00000000);
-	nv_icmd(priv, 0x000009e8, 0x00000000);
-	nv_icmd(priv, 0x000009e9, 0x00000000);
-	nv_icmd(priv, 0x000009ea, 0x00000000);
-	nv_icmd(priv, 0x000009eb, 0x00000000);
-	nv_icmd(priv, 0x000009ec, 0x00000000);
-	nv_icmd(priv, 0x000009ed, 0x00000000);
-	nv_icmd(priv, 0x000009ee, 0x00000000);
-	nv_icmd(priv, 0x000009ef, 0x00000000);
-	nv_icmd(priv, 0x000009f0, 0x00000000);
-	nv_icmd(priv, 0x000009f1, 0x00000000);
-	nv_icmd(priv, 0x000009f2, 0x00000000);
-	nv_icmd(priv, 0x000009f3, 0x00000000);
-	nv_icmd(priv, 0x000009f4, 0x00000000);
-	nv_icmd(priv, 0x000009f5, 0x00000000);
-	nv_icmd(priv, 0x000009f6, 0x00000000);
-	nv_icmd(priv, 0x000009f7, 0x00000000);
-	nv_icmd(priv, 0x000009f8, 0x00000000);
-	nv_icmd(priv, 0x000009f9, 0x00000000);
-	nv_icmd(priv, 0x000009fa, 0x00000000);
-	nv_icmd(priv, 0x000009fb, 0x00000000);
-	nv_icmd(priv, 0x000009fc, 0x00000000);
-	nv_icmd(priv, 0x000009fd, 0x00000000);
-	nv_icmd(priv, 0x000009fe, 0x00000000);
-	nv_icmd(priv, 0x000009ff, 0x00000000);
-	nv_icmd(priv, 0x00000468, 0x00000004);
-	nv_icmd(priv, 0x0000046c, 0x00000001);
-	nv_icmd(priv, 0x00000470, 0x00000000);
-	nv_icmd(priv, 0x00000471, 0x00000000);
-	nv_icmd(priv, 0x00000472, 0x00000000);
-	nv_icmd(priv, 0x00000473, 0x00000000);
-	nv_icmd(priv, 0x00000474, 0x00000000);
-	nv_icmd(priv, 0x00000475, 0x00000000);
-	nv_icmd(priv, 0x00000476, 0x00000000);
-	nv_icmd(priv, 0x00000477, 0x00000000);
-	nv_icmd(priv, 0x00000478, 0x00000000);
-	nv_icmd(priv, 0x00000479, 0x00000000);
-	nv_icmd(priv, 0x0000047a, 0x00000000);
-	nv_icmd(priv, 0x0000047b, 0x00000000);
-	nv_icmd(priv, 0x0000047c, 0x00000000);
-	nv_icmd(priv, 0x0000047d, 0x00000000);
-	nv_icmd(priv, 0x0000047e, 0x00000000);
-	nv_icmd(priv, 0x0000047f, 0x00000000);
-	nv_icmd(priv, 0x00000480, 0x00000000);
-	nv_icmd(priv, 0x00000481, 0x00000000);
-	nv_icmd(priv, 0x00000482, 0x00000000);
-	nv_icmd(priv, 0x00000483, 0x00000000);
-	nv_icmd(priv, 0x00000484, 0x00000000);
-	nv_icmd(priv, 0x00000485, 0x00000000);
-	nv_icmd(priv, 0x00000486, 0x00000000);
-	nv_icmd(priv, 0x00000487, 0x00000000);
-	nv_icmd(priv, 0x00000488, 0x00000000);
-	nv_icmd(priv, 0x00000489, 0x00000000);
-	nv_icmd(priv, 0x0000048a, 0x00000000);
-	nv_icmd(priv, 0x0000048b, 0x00000000);
-	nv_icmd(priv, 0x0000048c, 0x00000000);
-	nv_icmd(priv, 0x0000048d, 0x00000000);
-	nv_icmd(priv, 0x0000048e, 0x00000000);
-	nv_icmd(priv, 0x0000048f, 0x00000000);
-	nv_icmd(priv, 0x00000490, 0x00000000);
-	nv_icmd(priv, 0x00000491, 0x00000000);
-	nv_icmd(priv, 0x00000492, 0x00000000);
-	nv_icmd(priv, 0x00000493, 0x00000000);
-	nv_icmd(priv, 0x00000494, 0x00000000);
-	nv_icmd(priv, 0x00000495, 0x00000000);
-	nv_icmd(priv, 0x00000496, 0x00000000);
-	nv_icmd(priv, 0x00000497, 0x00000000);
-	nv_icmd(priv, 0x00000498, 0x00000000);
-	nv_icmd(priv, 0x00000499, 0x00000000);
-	nv_icmd(priv, 0x0000049a, 0x00000000);
-	nv_icmd(priv, 0x0000049b, 0x00000000);
-	nv_icmd(priv, 0x0000049c, 0x00000000);
-	nv_icmd(priv, 0x0000049d, 0x00000000);
-	nv_icmd(priv, 0x0000049e, 0x00000000);
-	nv_icmd(priv, 0x0000049f, 0x00000000);
-	nv_icmd(priv, 0x000004a0, 0x00000000);
-	nv_icmd(priv, 0x000004a1, 0x00000000);
-	nv_icmd(priv, 0x000004a2, 0x00000000);
-	nv_icmd(priv, 0x000004a3, 0x00000000);
-	nv_icmd(priv, 0x000004a4, 0x00000000);
-	nv_icmd(priv, 0x000004a5, 0x00000000);
-	nv_icmd(priv, 0x000004a6, 0x00000000);
-	nv_icmd(priv, 0x000004a7, 0x00000000);
-	nv_icmd(priv, 0x000004a8, 0x00000000);
-	nv_icmd(priv, 0x000004a9, 0x00000000);
-	nv_icmd(priv, 0x000004aa, 0x00000000);
-	nv_icmd(priv, 0x000004ab, 0x00000000);
-	nv_icmd(priv, 0x000004ac, 0x00000000);
-	nv_icmd(priv, 0x000004ad, 0x00000000);
-	nv_icmd(priv, 0x000004ae, 0x00000000);
-	nv_icmd(priv, 0x000004af, 0x00000000);
-	nv_icmd(priv, 0x000004b0, 0x00000000);
-	nv_icmd(priv, 0x000004b1, 0x00000000);
-	nv_icmd(priv, 0x000004b2, 0x00000000);
-	nv_icmd(priv, 0x000004b3, 0x00000000);
-	nv_icmd(priv, 0x000004b4, 0x00000000);
-	nv_icmd(priv, 0x000004b5, 0x00000000);
-	nv_icmd(priv, 0x000004b6, 0x00000000);
-	nv_icmd(priv, 0x000004b7, 0x00000000);
-	nv_icmd(priv, 0x000004b8, 0x00000000);
-	nv_icmd(priv, 0x000004b9, 0x00000000);
-	nv_icmd(priv, 0x000004ba, 0x00000000);
-	nv_icmd(priv, 0x000004bb, 0x00000000);
-	nv_icmd(priv, 0x000004bc, 0x00000000);
-	nv_icmd(priv, 0x000004bd, 0x00000000);
-	nv_icmd(priv, 0x000004be, 0x00000000);
-	nv_icmd(priv, 0x000004bf, 0x00000000);
-	nv_icmd(priv, 0x000004c0, 0x00000000);
-	nv_icmd(priv, 0x000004c1, 0x00000000);
-	nv_icmd(priv, 0x000004c2, 0x00000000);
-	nv_icmd(priv, 0x000004c3, 0x00000000);
-	nv_icmd(priv, 0x000004c4, 0x00000000);
-	nv_icmd(priv, 0x000004c5, 0x00000000);
-	nv_icmd(priv, 0x000004c6, 0x00000000);
-	nv_icmd(priv, 0x000004c7, 0x00000000);
-	nv_icmd(priv, 0x000004c8, 0x00000000);
-	nv_icmd(priv, 0x000004c9, 0x00000000);
-	nv_icmd(priv, 0x000004ca, 0x00000000);
-	nv_icmd(priv, 0x000004cb, 0x00000000);
-	nv_icmd(priv, 0x000004cc, 0x00000000);
-	nv_icmd(priv, 0x000004cd, 0x00000000);
-	nv_icmd(priv, 0x000004ce, 0x00000000);
-	nv_icmd(priv, 0x000004cf, 0x00000000);
-	nv_icmd(priv, 0x00000510, 0x3f800000);
-	nv_icmd(priv, 0x00000511, 0x3f800000);
-	nv_icmd(priv, 0x00000512, 0x3f800000);
-	nv_icmd(priv, 0x00000513, 0x3f800000);
-	nv_icmd(priv, 0x00000514, 0x3f800000);
-	nv_icmd(priv, 0x00000515, 0x3f800000);
-	nv_icmd(priv, 0x00000516, 0x3f800000);
-	nv_icmd(priv, 0x00000517, 0x3f800000);
-	nv_icmd(priv, 0x00000518, 0x3f800000);
-	nv_icmd(priv, 0x00000519, 0x3f800000);
-	nv_icmd(priv, 0x0000051a, 0x3f800000);
-	nv_icmd(priv, 0x0000051b, 0x3f800000);
-	nv_icmd(priv, 0x0000051c, 0x3f800000);
-	nv_icmd(priv, 0x0000051d, 0x3f800000);
-	nv_icmd(priv, 0x0000051e, 0x3f800000);
-	nv_icmd(priv, 0x0000051f, 0x3f800000);
-	nv_icmd(priv, 0x00000520, 0x000002b6);
-	nv_icmd(priv, 0x00000529, 0x00000001);
-	nv_icmd(priv, 0x00000530, 0xffff0000);
-	nv_icmd(priv, 0x00000531, 0xffff0000);
-	nv_icmd(priv, 0x00000532, 0xffff0000);
-	nv_icmd(priv, 0x00000533, 0xffff0000);
-	nv_icmd(priv, 0x00000534, 0xffff0000);
-	nv_icmd(priv, 0x00000535, 0xffff0000);
-	nv_icmd(priv, 0x00000536, 0xffff0000);
-	nv_icmd(priv, 0x00000537, 0xffff0000);
-	nv_icmd(priv, 0x00000538, 0xffff0000);
-	nv_icmd(priv, 0x00000539, 0xffff0000);
-	nv_icmd(priv, 0x0000053a, 0xffff0000);
-	nv_icmd(priv, 0x0000053b, 0xffff0000);
-	nv_icmd(priv, 0x0000053c, 0xffff0000);
-	nv_icmd(priv, 0x0000053d, 0xffff0000);
-	nv_icmd(priv, 0x0000053e, 0xffff0000);
-	nv_icmd(priv, 0x0000053f, 0xffff0000);
-	nv_icmd(priv, 0x00000585, 0x0000003f);
-	nv_icmd(priv, 0x00000576, 0x00000003);
-	if (nv_device(priv)->chipset == 0xc1 ||
-	    nv_device(priv)->chipset >= 0xd0)
-		nv_icmd(priv, 0x0000057b, 0x00000059);
-	nv_icmd(priv, 0x00000586, 0x00000040);
-	nv_icmd(priv, 0x00000582, 0x00000080);
-	nv_icmd(priv, 0x00000583, 0x00000080);
-	nv_icmd(priv, 0x000005c2, 0x00000001);
-	nv_icmd(priv, 0x00000638, 0x00000001);
-	nv_icmd(priv, 0x00000639, 0x00000001);
-	nv_icmd(priv, 0x0000063a, 0x00000002);
-	nv_icmd(priv, 0x0000063b, 0x00000001);
-	nv_icmd(priv, 0x0000063c, 0x00000001);
-	nv_icmd(priv, 0x0000063d, 0x00000002);
-	nv_icmd(priv, 0x0000063e, 0x00000001);
-	nv_icmd(priv, 0x000008b8, 0x00000001);
-	nv_icmd(priv, 0x000008b9, 0x00000001);
-	nv_icmd(priv, 0x000008ba, 0x00000001);
-	nv_icmd(priv, 0x000008bb, 0x00000001);
-	nv_icmd(priv, 0x000008bc, 0x00000001);
-	nv_icmd(priv, 0x000008bd, 0x00000001);
-	nv_icmd(priv, 0x000008be, 0x00000001);
-	nv_icmd(priv, 0x000008bf, 0x00000001);
-	nv_icmd(priv, 0x00000900, 0x00000001);
-	nv_icmd(priv, 0x00000901, 0x00000001);
-	nv_icmd(priv, 0x00000902, 0x00000001);
-	nv_icmd(priv, 0x00000903, 0x00000001);
-	nv_icmd(priv, 0x00000904, 0x00000001);
-	nv_icmd(priv, 0x00000905, 0x00000001);
-	nv_icmd(priv, 0x00000906, 0x00000001);
-	nv_icmd(priv, 0x00000907, 0x00000001);
-	nv_icmd(priv, 0x00000908, 0x00000002);
-	nv_icmd(priv, 0x00000909, 0x00000002);
-	nv_icmd(priv, 0x0000090a, 0x00000002);
-	nv_icmd(priv, 0x0000090b, 0x00000002);
-	nv_icmd(priv, 0x0000090c, 0x00000002);
-	nv_icmd(priv, 0x0000090d, 0x00000002);
-	nv_icmd(priv, 0x0000090e, 0x00000002);
-	nv_icmd(priv, 0x0000090f, 0x00000002);
-	nv_icmd(priv, 0x00000910, 0x00000001);
-	nv_icmd(priv, 0x00000911, 0x00000001);
-	nv_icmd(priv, 0x00000912, 0x00000001);
-	nv_icmd(priv, 0x00000913, 0x00000001);
-	nv_icmd(priv, 0x00000914, 0x00000001);
-	nv_icmd(priv, 0x00000915, 0x00000001);
-	nv_icmd(priv, 0x00000916, 0x00000001);
-	nv_icmd(priv, 0x00000917, 0x00000001);
-	nv_icmd(priv, 0x00000918, 0x00000001);
-	nv_icmd(priv, 0x00000919, 0x00000001);
-	nv_icmd(priv, 0x0000091a, 0x00000001);
-	nv_icmd(priv, 0x0000091b, 0x00000001);
-	nv_icmd(priv, 0x0000091c, 0x00000001);
-	nv_icmd(priv, 0x0000091d, 0x00000001);
-	nv_icmd(priv, 0x0000091e, 0x00000001);
-	nv_icmd(priv, 0x0000091f, 0x00000001);
-	nv_icmd(priv, 0x00000920, 0x00000002);
-	nv_icmd(priv, 0x00000921, 0x00000002);
-	nv_icmd(priv, 0x00000922, 0x00000002);
-	nv_icmd(priv, 0x00000923, 0x00000002);
-	nv_icmd(priv, 0x00000924, 0x00000002);
-	nv_icmd(priv, 0x00000925, 0x00000002);
-	nv_icmd(priv, 0x00000926, 0x00000002);
-	nv_icmd(priv, 0x00000927, 0x00000002);
-	nv_icmd(priv, 0x00000928, 0x00000001);
-	nv_icmd(priv, 0x00000929, 0x00000001);
-	nv_icmd(priv, 0x0000092a, 0x00000001);
-	nv_icmd(priv, 0x0000092b, 0x00000001);
-	nv_icmd(priv, 0x0000092c, 0x00000001);
-	nv_icmd(priv, 0x0000092d, 0x00000001);
-	nv_icmd(priv, 0x0000092e, 0x00000001);
-	nv_icmd(priv, 0x0000092f, 0x00000001);
-	nv_icmd(priv, 0x00000648, 0x00000001);
-	nv_icmd(priv, 0x00000649, 0x00000001);
-	nv_icmd(priv, 0x0000064a, 0x00000001);
-	nv_icmd(priv, 0x0000064b, 0x00000001);
-	nv_icmd(priv, 0x0000064c, 0x00000001);
-	nv_icmd(priv, 0x0000064d, 0x00000001);
-	nv_icmd(priv, 0x0000064e, 0x00000001);
-	nv_icmd(priv, 0x0000064f, 0x00000001);
-	nv_icmd(priv, 0x00000650, 0x00000001);
-	nv_icmd(priv, 0x00000658, 0x0000000f);
-	nv_icmd(priv, 0x000007ff, 0x0000000a);
-	nv_icmd(priv, 0x0000066a, 0x40000000);
-	nv_icmd(priv, 0x0000066b, 0x10000000);
-	nv_icmd(priv, 0x0000066c, 0xffff0000);
-	nv_icmd(priv, 0x0000066d, 0xffff0000);
-	nv_icmd(priv, 0x000007af, 0x00000008);
-	nv_icmd(priv, 0x000007b0, 0x00000008);
-	nv_icmd(priv, 0x000007f6, 0x00000001);
-	nv_icmd(priv, 0x000006b2, 0x00000055);
-	nv_icmd(priv, 0x000007ad, 0x00000003);
-	nv_icmd(priv, 0x00000937, 0x00000001);
-	nv_icmd(priv, 0x00000971, 0x00000008);
-	nv_icmd(priv, 0x00000972, 0x00000040);
-	nv_icmd(priv, 0x00000973, 0x0000012c);
-	nv_icmd(priv, 0x0000097c, 0x00000040);
-	nv_icmd(priv, 0x00000979, 0x00000003);
-	nv_icmd(priv, 0x00000975, 0x00000020);
-	nv_icmd(priv, 0x00000976, 0x00000001);
-	nv_icmd(priv, 0x00000977, 0x00000020);
-	nv_icmd(priv, 0x00000978, 0x00000001);
-	nv_icmd(priv, 0x00000957, 0x00000003);
-	nv_icmd(priv, 0x0000095e, 0x20164010);
-	nv_icmd(priv, 0x0000095f, 0x00000020);
-	if (nv_device(priv)->chipset >= 0xd0)
-		nv_icmd(priv, 0x0000097d, 0x00000020);
-	nv_icmd(priv, 0x00000683, 0x00000006);
-	nv_icmd(priv, 0x00000685, 0x003fffff);
-	nv_icmd(priv, 0x00000687, 0x00000c48);
-	nv_icmd(priv, 0x000006a0, 0x00000005);
-	nv_icmd(priv, 0x00000840, 0x00300008);
-	nv_icmd(priv, 0x00000841, 0x04000080);
-	nv_icmd(priv, 0x00000842, 0x00300008);
-	nv_icmd(priv, 0x00000843, 0x04000080);
-	nv_icmd(priv, 0x00000818, 0x00000000);
-	nv_icmd(priv, 0x00000819, 0x00000000);
-	nv_icmd(priv, 0x0000081a, 0x00000000);
-	nv_icmd(priv, 0x0000081b, 0x00000000);
-	nv_icmd(priv, 0x0000081c, 0x00000000);
-	nv_icmd(priv, 0x0000081d, 0x00000000);
-	nv_icmd(priv, 0x0000081e, 0x00000000);
-	nv_icmd(priv, 0x0000081f, 0x00000000);
-	nv_icmd(priv, 0x00000848, 0x00000000);
-	nv_icmd(priv, 0x00000849, 0x00000000);
-	nv_icmd(priv, 0x0000084a, 0x00000000);
-	nv_icmd(priv, 0x0000084b, 0x00000000);
-	nv_icmd(priv, 0x0000084c, 0x00000000);
-	nv_icmd(priv, 0x0000084d, 0x00000000);
-	nv_icmd(priv, 0x0000084e, 0x00000000);
-	nv_icmd(priv, 0x0000084f, 0x00000000);
-	nv_icmd(priv, 0x00000850, 0x00000000);
-	nv_icmd(priv, 0x00000851, 0x00000000);
-	nv_icmd(priv, 0x00000852, 0x00000000);
-	nv_icmd(priv, 0x00000853, 0x00000000);
-	nv_icmd(priv, 0x00000854, 0x00000000);
-	nv_icmd(priv, 0x00000855, 0x00000000);
-	nv_icmd(priv, 0x00000856, 0x00000000);
-	nv_icmd(priv, 0x00000857, 0x00000000);
-	nv_icmd(priv, 0x00000738, 0x00000000);
-	nv_icmd(priv, 0x000006aa, 0x00000001);
-	nv_icmd(priv, 0x000006ab, 0x00000002);
-	nv_icmd(priv, 0x000006ac, 0x00000080);
-	nv_icmd(priv, 0x000006ad, 0x00000100);
-	nv_icmd(priv, 0x000006ae, 0x00000100);
-	nv_icmd(priv, 0x000006b1, 0x00000011);
-	nv_icmd(priv, 0x000006bb, 0x000000cf);
-	nv_icmd(priv, 0x000006ce, 0x2a712488);
-	nv_icmd(priv, 0x00000739, 0x4085c000);
-	nv_icmd(priv, 0x0000073a, 0x00000080);
-	nv_icmd(priv, 0x00000786, 0x80000100);
-	nv_icmd(priv, 0x0000073c, 0x00010100);
-	nv_icmd(priv, 0x0000073d, 0x02800000);
-	nv_icmd(priv, 0x00000787, 0x000000cf);
-	nv_icmd(priv, 0x0000078c, 0x00000008);
-	nv_icmd(priv, 0x00000792, 0x00000001);
-	nv_icmd(priv, 0x00000794, 0x00000001);
-	nv_icmd(priv, 0x00000795, 0x00000001);
-	nv_icmd(priv, 0x00000796, 0x00000001);
-	nv_icmd(priv, 0x00000797, 0x000000cf);
-	nv_icmd(priv, 0x00000836, 0x00000001);
-	nv_icmd(priv, 0x0000079a, 0x00000002);
-	nv_icmd(priv, 0x00000833, 0x04444480);
-	nv_icmd(priv, 0x000007a1, 0x00000001);
-	nv_icmd(priv, 0x000007a3, 0x00000001);
-	nv_icmd(priv, 0x000007a4, 0x00000001);
-	nv_icmd(priv, 0x000007a5, 0x00000001);
-	nv_icmd(priv, 0x00000831, 0x00000004);
-	nv_icmd(priv, 0x0000080c, 0x00000002);
-	nv_icmd(priv, 0x0000080d, 0x00000100);
-	nv_icmd(priv, 0x0000080e, 0x00000100);
-	nv_icmd(priv, 0x0000080f, 0x00000001);
-	nv_icmd(priv, 0x00000823, 0x00000002);
-	nv_icmd(priv, 0x00000824, 0x00000100);
-	nv_icmd(priv, 0x00000825, 0x00000100);
-	nv_icmd(priv, 0x00000826, 0x00000001);
-	nv_icmd(priv, 0x0000095d, 0x00000001);
-	nv_icmd(priv, 0x0000082b, 0x00000004);
-	nv_icmd(priv, 0x00000942, 0x00010001);
-	nv_icmd(priv, 0x00000943, 0x00000001);
-	nv_icmd(priv, 0x00000944, 0x00000022);
-	nv_icmd(priv, 0x000007c5, 0x00010001);
-	nv_icmd(priv, 0x00000834, 0x00000001);
-	nv_icmd(priv, 0x000007c7, 0x00000001);
-	nv_icmd(priv, 0x0000c1b0, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b1, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b2, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b3, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b4, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b5, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b6, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b7, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b8, 0x0fac6881);
-	nv_icmd(priv, 0x0000c1b9, 0x00fac688);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_icmd(priv, 0x00001000, 0x00000002);
-	nv_icmd(priv, 0x000006aa, 0x00000001);
-	nv_icmd(priv, 0x000006ad, 0x00000100);
-	nv_icmd(priv, 0x000006ae, 0x00000100);
-	nv_icmd(priv, 0x000006b1, 0x00000011);
-	nv_icmd(priv, 0x0000078c, 0x00000008);
-	nv_icmd(priv, 0x00000792, 0x00000001);
-	nv_icmd(priv, 0x00000794, 0x00000001);
-	nv_icmd(priv, 0x00000795, 0x00000001);
-	nv_icmd(priv, 0x00000796, 0x00000001);
-	nv_icmd(priv, 0x00000797, 0x000000cf);
-	nv_icmd(priv, 0x0000079a, 0x00000002);
-	nv_icmd(priv, 0x00000833, 0x04444480);
-	nv_icmd(priv, 0x000007a1, 0x00000001);
-	nv_icmd(priv, 0x000007a3, 0x00000001);
-	nv_icmd(priv, 0x000007a4, 0x00000001);
-	nv_icmd(priv, 0x000007a5, 0x00000001);
-	nv_icmd(priv, 0x00000831, 0x00000004);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_icmd(priv, 0x00001000, 0x00000014);
-	nv_icmd(priv, 0x00000351, 0x00000100);
-	nv_icmd(priv, 0x00000957, 0x00000003);
-	nv_icmd(priv, 0x0000095d, 0x00000001);
-	nv_icmd(priv, 0x0000082b, 0x00000004);
-	nv_icmd(priv, 0x00000942, 0x00010001);
-	nv_icmd(priv, 0x00000943, 0x00000001);
-	nv_icmd(priv, 0x000007c5, 0x00010001);
-	nv_icmd(priv, 0x00000834, 0x00000001);
-	nv_icmd(priv, 0x000007c7, 0x00000001);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_icmd(priv, 0x00001000, 0x00000001);
-	nv_icmd(priv, 0x0000080c, 0x00000002);
-	nv_icmd(priv, 0x0000080d, 0x00000100);
-	nv_icmd(priv, 0x0000080e, 0x00000100);
-	nv_icmd(priv, 0x0000080f, 0x00000001);
-	nv_icmd(priv, 0x00000823, 0x00000002);
-	nv_icmd(priv, 0x00000824, 0x00000100);
-	nv_icmd(priv, 0x00000825, 0x00000100);
-	nv_icmd(priv, 0x00000826, 0x00000001);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_wr32(priv, 0x400208, 0x00000000);
-	nv_wr32(priv, 0x404154, 0x00000400);
-
-	nvc0_grctx_generate_9097(priv);
-	if (fermi >= 0x9197)
-		nvc0_grctx_generate_9197(priv);
-	if (fermi >= 0x9297)
-		nvc0_grctx_generate_9297(priv);
-	nvc0_grctx_generate_902d(priv);
-	nvc0_grctx_generate_9039(priv);
-	nvc0_grctx_generate_90c0(priv);
 
-	nv_wr32(priv, 0x000260, r000260);
-
-	return nvc0_grctx_fini(&info);
+done:
+	nouveau_gpuobj_ref(NULL, &chan);
+	return ret;
 }
+
+struct nvc0_graph_init *
+nvc0_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvc0_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvc0_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvc0_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvc0_grctx_init_rop,
+	NULL
+};
+
+static struct nvc0_graph_init *
+nvc0_grctx_init_gpc[] = {
+	nvc0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc0_grctx_init_tpc,
+	NULL
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_mthd_magic[] = {
+	{ 0x3410, 1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_mthd
+nvc0_grctx_init_mthd[] = {
+	{ 0x9097, nvc0_grctx_init_9097, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvc0_grctx_init_90c0, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvc0_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc0_grctx_generate_mods,
+	.unkn = nvc0_grctx_generate_unkn,
+	.hub  = nvc0_grctx_init_hub,
+	.gpc  = nvc0_grctx_init_gpc,
+	.icmd = nvc0_grctx_init_icmd,
+	.mthd = nvc0_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
new file mode 100644
index 0000000..e5be3ee
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
@@ -0,0 +1,823 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvc1_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc1_grctx_init_9097[] = {
+	{ 0x000800,   8, 0x40, 0x00000000 },
+	{ 0x000804,   8, 0x40, 0x00000000 },
+	{ 0x000808,   8, 0x40, 0x00000400 },
+	{ 0x00080c,   8, 0x40, 0x00000300 },
+	{ 0x000810,   1, 0x04, 0x000000cf },
+	{ 0x000850,   7, 0x40, 0x00000000 },
+	{ 0x000814,   8, 0x40, 0x00000040 },
+	{ 0x000818,   8, 0x40, 0x00000001 },
+	{ 0x00081c,   8, 0x40, 0x00000000 },
+	{ 0x000820,   8, 0x40, 0x00000000 },
+	{ 0x002700,   8, 0x20, 0x00000000 },
+	{ 0x002704,   8, 0x20, 0x00000000 },
+	{ 0x002708,   8, 0x20, 0x00000000 },
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x002710,   8, 0x20, 0x00014000 },
+	{ 0x002714,   8, 0x20, 0x00000040 },
+	{ 0x001c00,  16, 0x10, 0x00000000 },
+	{ 0x001c04,  16, 0x10, 0x00000000 },
+	{ 0x001c08,  16, 0x10, 0x00000000 },
+	{ 0x001c0c,  16, 0x10, 0x00000000 },
+	{ 0x001d00,  16, 0x10, 0x00000000 },
+	{ 0x001d04,  16, 0x10, 0x00000000 },
+	{ 0x001d08,  16, 0x10, 0x00000000 },
+	{ 0x001d0c,  16, 0x10, 0x00000000 },
+	{ 0x001f00,  16, 0x08, 0x00000000 },
+	{ 0x001f04,  16, 0x08, 0x00000000 },
+	{ 0x001f80,  16, 0x08, 0x00000000 },
+	{ 0x001f84,  16, 0x08, 0x00000000 },
+	{ 0x002200,   5, 0x10, 0x00000022 },
+	{ 0x002000,   1, 0x04, 0x00000000 },
+	{ 0x002040,   1, 0x04, 0x00000011 },
+	{ 0x002080,   1, 0x04, 0x00000020 },
+	{ 0x0020c0,   1, 0x04, 0x00000030 },
+	{ 0x002100,   1, 0x04, 0x00000040 },
+	{ 0x002140,   1, 0x04, 0x00000051 },
+	{ 0x00200c,   6, 0x40, 0x00000001 },
+	{ 0x002010,   1, 0x04, 0x00000000 },
+	{ 0x002050,   1, 0x04, 0x00000000 },
+	{ 0x002090,   1, 0x04, 0x00000001 },
+	{ 0x0020d0,   1, 0x04, 0x00000002 },
+	{ 0x002110,   1, 0x04, 0x00000003 },
+	{ 0x002150,   1, 0x04, 0x00000004 },
+	{ 0x000380,   4, 0x20, 0x00000000 },
+	{ 0x000384,   4, 0x20, 0x00000000 },
+	{ 0x000388,   4, 0x20, 0x00000000 },
+	{ 0x00038c,   4, 0x20, 0x00000000 },
+	{ 0x000700,   4, 0x10, 0x00000000 },
+	{ 0x000704,   4, 0x10, 0x00000000 },
+	{ 0x000708,   4, 0x10, 0x00000000 },
+	{ 0x002800, 128, 0x04, 0x00000000 },
+	{ 0x000a00,  16, 0x20, 0x00000000 },
+	{ 0x000a04,  16, 0x20, 0x00000000 },
+	{ 0x000a08,  16, 0x20, 0x00000000 },
+	{ 0x000a0c,  16, 0x20, 0x00000000 },
+	{ 0x000a10,  16, 0x20, 0x00000000 },
+	{ 0x000a14,  16, 0x20, 0x00000000 },
+	{ 0x000c00,  16, 0x10, 0x00000000 },
+	{ 0x000c04,  16, 0x10, 0x00000000 },
+	{ 0x000c08,  16, 0x10, 0x00000000 },
+	{ 0x000c0c,  16, 0x10, 0x3f800000 },
+	{ 0x000d00,   8, 0x08, 0xffff0000 },
+	{ 0x000d04,   8, 0x08, 0xffff0000 },
+	{ 0x000e00,  16, 0x10, 0x00000000 },
+	{ 0x000e04,  16, 0x10, 0xffff0000 },
+	{ 0x000e08,  16, 0x10, 0xffff0000 },
+	{ 0x000d40,   4, 0x08, 0x00000000 },
+	{ 0x000d44,   4, 0x08, 0x00000000 },
+	{ 0x001e00,   8, 0x20, 0x00000001 },
+	{ 0x001e04,   8, 0x20, 0x00000001 },
+	{ 0x001e08,   8, 0x20, 0x00000002 },
+	{ 0x001e0c,   8, 0x20, 0x00000001 },
+	{ 0x001e10,   8, 0x20, 0x00000001 },
+	{ 0x001e14,   8, 0x20, 0x00000002 },
+	{ 0x001e18,   8, 0x20, 0x00000001 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x001514,   1, 0x04, 0x00000000 },
+	{ 0x000d68,   1, 0x04, 0x0000ffff },
+	{ 0x00121c,   1, 0x04, 0x0fac6881 },
+	{ 0x000fac,   1, 0x04, 0x00000001 },
+	{ 0x001538,   1, 0x04, 0x00000001 },
+	{ 0x000fe0,   2, 0x04, 0x00000000 },
+	{ 0x000fe8,   1, 0x04, 0x00000014 },
+	{ 0x000fec,   1, 0x04, 0x00000040 },
+	{ 0x000ff0,   1, 0x04, 0x00000000 },
+	{ 0x00179c,   1, 0x04, 0x00000000 },
+	{ 0x001228,   1, 0x04, 0x00000400 },
+	{ 0x00122c,   1, 0x04, 0x00000300 },
+	{ 0x001230,   1, 0x04, 0x00010001 },
+	{ 0x0007f8,   1, 0x04, 0x00000000 },
+	{ 0x0015b4,   1, 0x04, 0x00000001 },
+	{ 0x0015cc,   1, 0x04, 0x00000000 },
+	{ 0x001534,   1, 0x04, 0x00000000 },
+	{ 0x000fb0,   1, 0x04, 0x00000000 },
+	{ 0x0015d0,   1, 0x04, 0x00000000 },
+	{ 0x00153c,   1, 0x04, 0x00000000 },
+	{ 0x0016b4,   1, 0x04, 0x00000003 },
+	{ 0x000fbc,   4, 0x04, 0x0000ffff },
+	{ 0x000df8,   2, 0x04, 0x00000000 },
+	{ 0x001948,   1, 0x04, 0x00000000 },
+	{ 0x001970,   1, 0x04, 0x00000001 },
+	{ 0x00161c,   1, 0x04, 0x000009f0 },
+	{ 0x000dcc,   1, 0x04, 0x00000010 },
+	{ 0x00163c,   1, 0x04, 0x00000000 },
+	{ 0x0015e4,   1, 0x04, 0x00000000 },
+	{ 0x001160,  32, 0x04, 0x25e00040 },
+	{ 0x001880,  32, 0x04, 0x00000000 },
+	{ 0x000f84,   2, 0x04, 0x00000000 },
+	{ 0x0017c8,   2, 0x04, 0x00000000 },
+	{ 0x0017d0,   1, 0x04, 0x000000ff },
+	{ 0x0017d4,   1, 0x04, 0xffffffff },
+	{ 0x0017d8,   1, 0x04, 0x00000002 },
+	{ 0x0017dc,   1, 0x04, 0x00000000 },
+	{ 0x0015f4,   2, 0x04, 0x00000000 },
+	{ 0x001434,   2, 0x04, 0x00000000 },
+	{ 0x000d74,   1, 0x04, 0x00000000 },
+	{ 0x000dec,   1, 0x04, 0x00000001 },
+	{ 0x0013a4,   1, 0x04, 0x00000000 },
+	{ 0x001318,   1, 0x04, 0x00000001 },
+	{ 0x001644,   1, 0x04, 0x00000000 },
+	{ 0x000748,   1, 0x04, 0x00000000 },
+	{ 0x000de8,   1, 0x04, 0x00000000 },
+	{ 0x001648,   1, 0x04, 0x00000000 },
+	{ 0x0012a4,   1, 0x04, 0x00000000 },
+	{ 0x001120,   4, 0x04, 0x00000000 },
+	{ 0x001118,   1, 0x04, 0x00000000 },
+	{ 0x00164c,   1, 0x04, 0x00000000 },
+	{ 0x001658,   1, 0x04, 0x00000000 },
+	{ 0x001910,   1, 0x04, 0x00000290 },
+	{ 0x001518,   1, 0x04, 0x00000000 },
+	{ 0x00165c,   1, 0x04, 0x00000001 },
+	{ 0x001520,   1, 0x04, 0x00000000 },
+	{ 0x001604,   1, 0x04, 0x00000000 },
+	{ 0x001570,   1, 0x04, 0x00000000 },
+	{ 0x0013b0,   2, 0x04, 0x3f800000 },
+	{ 0x00020c,   1, 0x04, 0x00000000 },
+	{ 0x001670,   1, 0x04, 0x30201000 },
+	{ 0x001674,   1, 0x04, 0x70605040 },
+	{ 0x001678,   1, 0x04, 0xb8a89888 },
+	{ 0x00167c,   1, 0x04, 0xf8e8d8c8 },
+	{ 0x00166c,   1, 0x04, 0x00000000 },
+	{ 0x001680,   1, 0x04, 0x00ffff00 },
+	{ 0x0012d0,   1, 0x04, 0x00000003 },
+	{ 0x0012d4,   1, 0x04, 0x00000002 },
+	{ 0x001684,   2, 0x04, 0x00000000 },
+	{ 0x000dac,   2, 0x04, 0x00001b02 },
+	{ 0x000db4,   1, 0x04, 0x00000000 },
+	{ 0x00168c,   1, 0x04, 0x00000000 },
+	{ 0x0015bc,   1, 0x04, 0x00000000 },
+	{ 0x00156c,   1, 0x04, 0x00000000 },
+	{ 0x00187c,   1, 0x04, 0x00000000 },
+	{ 0x001110,   1, 0x04, 0x00000001 },
+	{ 0x000dc0,   3, 0x04, 0x00000000 },
+	{ 0x001234,   1, 0x04, 0x00000000 },
+	{ 0x001690,   1, 0x04, 0x00000000 },
+	{ 0x0012ac,   1, 0x04, 0x00000001 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x001000,   1, 0x04, 0x00000010 },
+	{ 0x0010fc,   1, 0x04, 0x00000000 },
+	{ 0x001290,   1, 0x04, 0x00000000 },
+	{ 0x000218,   1, 0x04, 0x00000010 },
+	{ 0x0012d8,   1, 0x04, 0x00000000 },
+	{ 0x0012dc,   1, 0x04, 0x00000010 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x00155c,   2, 0x04, 0x00000000 },
+	{ 0x001564,   1, 0x04, 0x00001fff },
+	{ 0x001574,   2, 0x04, 0x00000000 },
+	{ 0x00157c,   1, 0x04, 0x003fffff },
+	{ 0x001354,   1, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{ 0x001610,   1, 0x04, 0x00000012 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x00162c,   1, 0x04, 0x00000003 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000324,   6, 0x04, 0x3f800000 },
+	{ 0x000750,   1, 0x04, 0x00000000 },
+	{ 0x000760,   1, 0x04, 0x39291909 },
+	{ 0x000764,   1, 0x04, 0x79695949 },
+	{ 0x000768,   1, 0x04, 0xb9a99989 },
+	{ 0x00076c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x000770,   1, 0x04, 0x30201000 },
+	{ 0x000774,   1, 0x04, 0x70605040 },
+	{ 0x000778,   1, 0x04, 0x00009080 },
+	{ 0x000780,   1, 0x04, 0x39291909 },
+	{ 0x000784,   1, 0x04, 0x79695949 },
+	{ 0x000788,   1, 0x04, 0xb9a99989 },
+	{ 0x00078c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x0007d0,   1, 0x04, 0x30201000 },
+	{ 0x0007d4,   1, 0x04, 0x70605040 },
+	{ 0x0007d8,   1, 0x04, 0x00009080 },
+	{ 0x00037c,   1, 0x04, 0x00000001 },
+	{ 0x000740,   2, 0x04, 0x00000000 },
+	{ 0x002600,   1, 0x04, 0x00000000 },
+	{ 0x001918,   1, 0x04, 0x00000000 },
+	{ 0x00191c,   1, 0x04, 0x00000900 },
+	{ 0x001920,   1, 0x04, 0x00000405 },
+	{ 0x001308,   1, 0x04, 0x00000001 },
+	{ 0x001924,   1, 0x04, 0x00000000 },
+	{ 0x0013ac,   1, 0x04, 0x00000000 },
+	{ 0x00192c,   1, 0x04, 0x00000001 },
+	{ 0x00193c,   1, 0x04, 0x00002c1c },
+	{ 0x000d7c,   1, 0x04, 0x00000000 },
+	{ 0x000f8c,   1, 0x04, 0x00000000 },
+	{ 0x0002c0,   1, 0x04, 0x00000001 },
+	{ 0x001510,   1, 0x04, 0x00000000 },
+	{ 0x001940,   1, 0x04, 0x00000000 },
+	{ 0x000ff4,   2, 0x04, 0x00000000 },
+	{ 0x00194c,   2, 0x04, 0x00000000 },
+	{ 0x001968,   1, 0x04, 0x00000000 },
+	{ 0x001590,   1, 0x04, 0x0000003f },
+	{ 0x0007e8,   4, 0x04, 0x00000000 },
+	{ 0x00196c,   1, 0x04, 0x00000011 },
+	{ 0x00197c,   1, 0x04, 0x00000000 },
+	{ 0x000fcc,   2, 0x04, 0x00000000 },
+	{ 0x0002d8,   1, 0x04, 0x00000040 },
+	{ 0x001980,   1, 0x04, 0x00000080 },
+	{ 0x001504,   1, 0x04, 0x00000080 },
+	{ 0x001984,   1, 0x04, 0x00000000 },
+	{ 0x000300,   1, 0x04, 0x00000001 },
+	{ 0x0013a8,   1, 0x04, 0x00000000 },
+	{ 0x0012ec,   1, 0x04, 0x00000000 },
+	{ 0x001310,   1, 0x04, 0x00000000 },
+	{ 0x001314,   1, 0x04, 0x00000001 },
+	{ 0x001380,   1, 0x04, 0x00000000 },
+	{ 0x001384,   4, 0x04, 0x00000001 },
+	{ 0x001394,   1, 0x04, 0x00000000 },
+	{ 0x00139c,   1, 0x04, 0x00000000 },
+	{ 0x001398,   1, 0x04, 0x00000000 },
+	{ 0x001594,   1, 0x04, 0x00000000 },
+	{ 0x001598,   4, 0x04, 0x00000001 },
+	{ 0x000f54,   3, 0x04, 0x00000000 },
+	{ 0x0019bc,   1, 0x04, 0x00000000 },
+	{ 0x000f9c,   2, 0x04, 0x00000000 },
+	{ 0x0012cc,   1, 0x04, 0x00000000 },
+	{ 0x0012e8,   1, 0x04, 0x00000000 },
+	{ 0x00130c,   1, 0x04, 0x00000001 },
+	{ 0x001360,   8, 0x04, 0x00000000 },
+	{ 0x00133c,   2, 0x04, 0x00000001 },
+	{ 0x001344,   1, 0x04, 0x00000002 },
+	{ 0x001348,   2, 0x04, 0x00000001 },
+	{ 0x001350,   1, 0x04, 0x00000002 },
+	{ 0x001358,   1, 0x04, 0x00000001 },
+	{ 0x0012e4,   1, 0x04, 0x00000000 },
+	{ 0x00131c,   1, 0x04, 0x00000000 },
+	{ 0x001320,   3, 0x04, 0x00000000 },
+	{ 0x0019c0,   1, 0x04, 0x00000000 },
+	{ 0x001140,   1, 0x04, 0x00000000 },
+	{ 0x0019c4,   1, 0x04, 0x00000000 },
+	{ 0x0019c8,   1, 0x04, 0x00001500 },
+	{ 0x00135c,   1, 0x04, 0x00000000 },
+	{ 0x000f90,   1, 0x04, 0x00000000 },
+	{ 0x0019e0,   8, 0x04, 0x00000001 },
+	{ 0x0019cc,   1, 0x04, 0x00000001 },
+	{ 0x0015b8,   1, 0x04, 0x00000000 },
+	{ 0x001a00,   1, 0x04, 0x00001111 },
+	{ 0x001a04,   7, 0x04, 0x00000000 },
+	{ 0x000d6c,   2, 0x04, 0xffff0000 },
+	{ 0x0010f8,   1, 0x04, 0x00001010 },
+	{ 0x000d80,   5, 0x04, 0x00000000 },
+	{ 0x000da0,   1, 0x04, 0x00000000 },
+	{ 0x001508,   1, 0x04, 0x80000000 },
+	{ 0x00150c,   1, 0x04, 0x40000000 },
+	{ 0x001668,   1, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000008 },
+	{ 0x000d9c,   1, 0x04, 0x00000001 },
+	{ 0x0007dc,   1, 0x04, 0x00000000 },
+	{ 0x00074c,   1, 0x04, 0x00000055 },
+	{ 0x001420,   1, 0x04, 0x00000003 },
+	{ 0x0017bc,   2, 0x04, 0x00000000 },
+	{ 0x0017c4,   1, 0x04, 0x00000001 },
+	{ 0x001008,   1, 0x04, 0x00000008 },
+	{ 0x00100c,   1, 0x04, 0x00000040 },
+	{ 0x001010,   1, 0x04, 0x0000012c },
+	{ 0x000d60,   1, 0x04, 0x00000040 },
+	{ 0x00075c,   1, 0x04, 0x00000003 },
+	{ 0x001018,   1, 0x04, 0x00000020 },
+	{ 0x00101c,   1, 0x04, 0x00000001 },
+	{ 0x001020,   1, 0x04, 0x00000020 },
+	{ 0x001024,   1, 0x04, 0x00000001 },
+	{ 0x001444,   3, 0x04, 0x00000000 },
+	{ 0x000360,   1, 0x04, 0x20164010 },
+	{ 0x000364,   1, 0x04, 0x00000020 },
+	{ 0x000368,   1, 0x04, 0x00000000 },
+	{ 0x000de4,   1, 0x04, 0x00000000 },
+	{ 0x000204,   1, 0x04, 0x00000006 },
+	{ 0x000208,   1, 0x04, 0x00000000 },
+	{ 0x0002cc,   1, 0x04, 0x003fffff },
+	{ 0x0002d0,   1, 0x04, 0x00000c48 },
+	{ 0x001220,   1, 0x04, 0x00000005 },
+	{ 0x000fdc,   1, 0x04, 0x00000000 },
+	{ 0x000f98,   1, 0x04, 0x00300008 },
+	{ 0x001284,   1, 0x04, 0x04000080 },
+	{ 0x001450,   1, 0x04, 0x00300008 },
+	{ 0x001454,   1, 0x04, 0x04000080 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_9197[] = {
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x0002e4,   1, 0x04, 0x0000b001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180218 },
+	{ 0x405834,   2, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1003e005 },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x62000001 },
+	{ 0x408908,   1, 0x04, 0x00c80929 },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x00200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   1, 0x04, 0x00000000 },
+	{ 0x41870c,   1, 0x04, 0x07c80000 },
+	{ 0x418710,   1, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x0006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x00100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418a00,   3, 0x04, 0x00000000 },
+	{ 0x418a0c,   1, 0x04, 0x00010000 },
+	{ 0x418a10,   3, 0x04, 0x00000000 },
+	{ 0x418a20,   3, 0x04, 0x00000000 },
+	{ 0x418a2c,   1, 0x04, 0x00010000 },
+	{ 0x418a30,   3, 0x04, 0x00000000 },
+	{ 0x418a40,   3, 0x04, 0x00000000 },
+	{ 0x418a4c,   1, 0x04, 0x00010000 },
+	{ 0x418a50,   3, 0x04, 0x00000000 },
+	{ 0x418a60,   3, 0x04, 0x00000000 },
+	{ 0x418a6c,   1, 0x04, 0x00010000 },
+	{ 0x418a70,   3, 0x04, 0x00000000 },
+	{ 0x418a80,   3, 0x04, 0x00000000 },
+	{ 0x418a8c,   1, 0x04, 0x00010000 },
+	{ 0x418a90,   3, 0x04, 0x00000000 },
+	{ 0x418aa0,   3, 0x04, 0x00000000 },
+	{ 0x418aac,   1, 0x04, 0x00010000 },
+	{ 0x418ab0,   3, 0x04, 0x00000000 },
+	{ 0x418ac0,   3, 0x04, 0x00000000 },
+	{ 0x418acc,   1, 0x04, 0x00010000 },
+	{ 0x418ad0,   3, 0x04, 0x00000000 },
+	{ 0x418ae0,   3, 0x04, 0x00000000 },
+	{ 0x418aec,   1, 0x04, 0x00010000 },
+	{ 0x418af0,   3, 0x04, 0x00000000 },
+	{ 0x418b00,   1, 0x04, 0x00000000 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0007f440 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00400001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x12180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419d44,   1, 0x04, 0x02180218 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00011110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+};
+
+void
+nvc1_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	int gpc, tpc;
+	u32 offset;
+
+	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000018,  0, 0);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000018,  0, 0);
+
+	mmio_list(0x405830, 0x02180218, 0, 0);
+	mmio_list(0x4064c4, 0x0086ffff, 0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			u32 addr = TPC_UNIT(gpc, tpc, 0x0520);
+			mmio_list(addr, 0x12180000 | offset, 0, 0);
+			offset += 0x0324;
+		}
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			u32 addr = TPC_UNIT(gpc, tpc, 0x0544);
+			mmio_list(addr, 0x02180000 | offset, 0, 0);
+			offset += 0x0324;
+		}
+	}
+}
+
+void
+nvc1_grctx_generate_unkn(struct nvc0_graph_priv *priv)
+{
+	nv_mask(priv, 0x418c6c, 0x00000001, 0x00000001);
+	nv_mask(priv, 0x41980c, 0x00000010, 0x00000010);
+	nv_mask(priv, 0x419814, 0x00000004, 0x00000004);
+	nv_mask(priv, 0x4064c0, 0x80000000, 0x80000000);
+	nv_mask(priv, 0x405800, 0x08000000, 0x08000000);
+	nv_mask(priv, 0x419c00, 0x00000008, 0x00000008);
+}
+
+static struct nvc0_graph_init *
+nvc1_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvc0_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvc1_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvc0_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvc1_grctx_init_rop,
+	NULL
+};
+
+struct nvc0_graph_init *
+nvc1_grctx_init_gpc[] = {
+	nvc1_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc1_grctx_init_tpc,
+	NULL
+};
+
+static struct nvc0_graph_mthd
+nvc1_grctx_init_mthd[] = {
+	{ 0x9097, nvc1_grctx_init_9097, },
+	{ 0x9197, nvc1_grctx_init_9197, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvc0_grctx_init_90c0, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvc1_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc1),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc1_grctx_generate_mods,
+	.unkn = nvc1_grctx_generate_unkn,
+	.hub  = nvc1_grctx_init_hub,
+	.gpc  = nvc1_grctx_init_gpc,
+	.icmd = nvc1_grctx_init_icmd,
+	.mthd = nvc1_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc3.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc3.c
new file mode 100644
index 0000000..8f237b3
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc3.c
@@ -0,0 +1,99 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvc3_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x0000012a },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0007f440 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00000001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x02180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00011110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init *
+nvc3_grctx_init_gpc[] = {
+	nvc0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc3_grctx_init_tpc,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc3_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc3),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc0_grctx_generate_mods,
+	.unkn = nvc0_grctx_generate_unkn,
+	.hub  = nvc0_grctx_init_hub,
+	.gpc  = nvc3_grctx_init_gpc,
+	.icmd = nvc0_grctx_init_icmd,
+	.mthd = nvc0_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
new file mode 100644
index 0000000..d0d4ce3
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
@@ -0,0 +1,370 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvc8_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x00097d,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc8_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x0000012a },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00000001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00060048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x02180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419f50,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc8_grctx_init_9197[] = {
+	{ 0x0002e4,   1, 0x04, 0x0000b001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc8_grctx_init_9297[] = {
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x00036c,   2, 0x04, 0x00000000 },
+	{ 0x0007a4,   2, 0x04, 0x00000000 },
+	{ 0x000374,   1, 0x04, 0x00000000 },
+	{ 0x000378,   1, 0x04, 0x00000020 },
+	{}
+};
+
+static struct nvc0_graph_mthd
+nvc8_grctx_init_mthd[] = {
+	{ 0x9097, nvc1_grctx_init_9097, },
+	{ 0x9197, nvc8_grctx_init_9197, },
+	{ 0x9297, nvc8_grctx_init_9297, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvc0_grctx_init_90c0, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvc8_grctx_init_gpc[] = {
+	nvc0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc8_grctx_init_tpc,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc8_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc8),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc0_grctx_generate_mods,
+	.unkn = nvc0_grctx_generate_unkn,
+	.hub  = nvc0_grctx_init_hub,
+	.gpc  = nvc8_grctx_init_gpc,
+	.icmd = nvc8_grctx_init_icmd,
+	.mthd = nvc8_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
new file mode 100644
index 0000000..438e784
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
@@ -0,0 +1,290 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+struct nvc0_graph_init
+nvd7_grctx_init_unk40xx[] = {
+	{ 0x404004,  10, 0x04, 0x00000000 },
+	{ 0x404044,   1, 0x04, 0x00000000 },
+	{ 0x404094,   1, 0x04, 0x00000000 },
+	{ 0x404098,  12, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf0000087 },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x404168,   1, 0x04, 0x00000000 },
+	{ 0x404178,   2, 0x04, 0x00000000 },
+	{ 0x404200,   8, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180324 },
+	{ 0x405834,   1, 0x04, 0x08000000 },
+	{ 0x405838,   1, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   3, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x801a0078 },
+	{ 0x4064c4,   1, 0x04, 0x00c9ffff },
+	{ 0x4064d0,   8, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_tpc[] = {
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00008000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0017f440 },
+	{ 0x419c00,   1, 0x04, 0x0000000a },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3efbefbe },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00010110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_unk[] = {
+	{ 0x41be24,   1, 0x04, 0x00000002 },
+	{ 0x41bec0,   1, 0x04, 0x12180000 },
+	{ 0x41bec4,   1, 0x04, 0x00003fff },
+	{ 0x41bee4,   1, 0x04, 0x03240218 },
+	{ 0x41bf00,   1, 0x04, 0x0a418820 },
+	{ 0x41bf04,   1, 0x04, 0x062080e6 },
+	{ 0x41bf08,   1, 0x04, 0x020398a4 },
+	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
+	{ 0x41bf10,   1, 0x04, 0x0a418820 },
+	{ 0x41bf14,   1, 0x04, 0x000000e6 },
+	{ 0x41bfd0,   1, 0x04, 0x00900103 },
+	{ 0x41bfe0,   1, 0x04, 0x00400001 },
+	{ 0x41bfe4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static void
+nvd7_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	u32 magic[GPC_MAX][2];
+	u32 offset;
+	int gpc;
+
+	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000018,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000018,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+
+	mmio_list(0x405830, 0x02180324,  0, 0);
+	mmio_list(0x4064c4, 0x00c9ffff,  0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
+		u16 magic1 = 0x0324 * priv->tpc_nr[gpc];
+		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
+		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
+		offset += 0x0324 * priv->tpc_nr[gpc];
+	}
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
+		offset += 0x07ff * priv->tpc_nr[gpc];
+	}
+	mmio_list(0x17e91c, 0x03060609, 0, 0); /* different from kepler */
+}
+
+void
+nvd7_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	int i;
+
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+
+	for (i = 0; oclass->hub[i]; i++)
+		nvc0_graph_mmio(priv, oclass->hub[i]);
+	for (i = 0; oclass->gpc[i]; i++)
+		nvc0_graph_mmio(priv, oclass->gpc[i]);
+
+	nv_wr32(priv, 0x404154, 0x00000000);
+
+	oclass->mods(priv, info);
+	oclass->unkn(priv);
+
+	nvc0_grctx_generate_tpcid(priv);
+	nvc0_grctx_generate_r406028(priv);
+	nvc0_grctx_generate_r4060a8(priv);
+	nve4_grctx_generate_r418bb8(priv);
+	nvc0_grctx_generate_r406800(priv);
+
+	for (i = 0; i < 8; i++)
+		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
+
+	nvc0_graph_icmd(priv, oclass->icmd);
+	nv_wr32(priv, 0x404154, 0x00000400);
+	nvc0_graph_mthd(priv, oclass->mthd);
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+}
+
+
+static struct nvc0_graph_init *
+nvd7_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvd7_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvd7_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvd7_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvd9_grctx_init_rop,
+};
+
+struct nvc0_graph_init *
+nvd7_grctx_init_gpc[] = {
+	nvd7_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvd7_grctx_init_tpc,
+	nvd7_grctx_init_unk,
+	NULL
+};
+
+struct nouveau_oclass *
+nvd7_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xd7),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvd7_grctx_generate_main,
+	.mods = nvd7_grctx_generate_mods,
+	.unkn = nve4_grctx_generate_unkn,
+	.hub  = nvd7_grctx_init_hub,
+	.gpc  = nvd7_grctx_init_gpc,
+	.icmd = nvd9_grctx_init_icmd,
+	.mthd = nvd9_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
new file mode 100644
index 0000000..818a475
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
@@ -0,0 +1,515 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+struct nvc0_graph_init
+nvd9_grctx_init_90c0[] = {
+	{ 0x002700,   4, 0x40, 0x00000000 },
+	{ 0x002720,   4, 0x40, 0x00000000 },
+	{ 0x002704,   4, 0x40, 0x00000000 },
+	{ 0x002724,   4, 0x40, 0x00000000 },
+	{ 0x002708,   4, 0x40, 0x00000000 },
+	{ 0x002728,   4, 0x40, 0x00000000 },
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x002710,   4, 0x40, 0x00014000 },
+	{ 0x002730,   4, 0x40, 0x00014000 },
+	{ 0x002714,   4, 0x40, 0x00000040 },
+	{ 0x002734,   4, 0x40, 0x00000040 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x000758,   1, 0x04, 0x00000100 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x000204,   3, 0x04, 0x00000000 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{ 0x00024c,   1, 0x04, 0x00000000 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000400,  24, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x000440,  24, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x00097d,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_unk40xx[] = {
+	{ 0x404004,  11, 0x04, 0x00000000 },
+	{ 0x404044,   1, 0x04, 0x00000000 },
+	{ 0x404094,   1, 0x04, 0x00000000 },
+	{ 0x404098,  12, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf0000087 },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x404168,   1, 0x04, 0x00000000 },
+	{ 0x404178,   2, 0x04, 0x00000000 },
+	{ 0x404200,   8, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180218 },
+	{ 0x405834,   1, 0x04, 0x08000000 },
+	{ 0x405838,   1, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   3, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x80140078 },
+	{ 0x4064c4,   1, 0x04, 0x0086ffff },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1043e005 },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x1043e005 },
+	{ 0x408908,   1, 0x04, 0x00c8102f },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100008 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0017f440 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00400001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x0000000a },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3cf3cf3c },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x12180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419d44,   1, 0x04, 0x02180218 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00010110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvd9_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvd9_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvd9_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvd9_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvd9_grctx_init_rop,
+};
+
+struct nvc0_graph_init *
+nvd9_grctx_init_gpc[] = {
+	nvd9_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvd9_grctx_init_tpc,
+	NULL
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_mthd_magic[] = {
+	{ 0x3410, 1, 0x04, 0x80002006 },
+	{}
+};
+
+struct nvc0_graph_mthd
+nvd9_grctx_init_mthd[] = {
+	{ 0x9097, nvc1_grctx_init_9097, },
+	{ 0x9197, nvc8_grctx_init_9197, },
+	{ 0x9297, nvc8_grctx_init_9297, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvd9_grctx_init_90c0, },
+	{ 0x902d, nvd9_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvd9_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xd9),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc1_grctx_generate_mods,
+	.unkn = nvc1_grctx_generate_unkn,
+	.hub  = nvd9_grctx_init_hub,
+	.gpc  = nvd9_grctx_init_gpc,
+	.icmd = nvd9_grctx_init_icmd,
+	.mthd = nvd9_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
deleted file mode 100644
index ae27dae..0000000
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
+++ /dev/null
@@ -1,2793 +0,0 @@
-/*
- * Copyright 2010 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "nvc0.h"
-
-static void
-nve0_grctx_generate_icmd(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x400208, 0x80000000);
-	nv_icmd(priv, 0x001000, 0x00000004);
-	nv_icmd(priv, 0x000039, 0x00000000);
-	nv_icmd(priv, 0x00003a, 0x00000000);
-	nv_icmd(priv, 0x00003b, 0x00000000);
-	nv_icmd(priv, 0x0000a9, 0x0000ffff);
-	nv_icmd(priv, 0x000038, 0x0fac6881);
-	nv_icmd(priv, 0x00003d, 0x00000001);
-	nv_icmd(priv, 0x0000e8, 0x00000400);
-	nv_icmd(priv, 0x0000e9, 0x00000400);
-	nv_icmd(priv, 0x0000ea, 0x00000400);
-	nv_icmd(priv, 0x0000eb, 0x00000400);
-	nv_icmd(priv, 0x0000ec, 0x00000400);
-	nv_icmd(priv, 0x0000ed, 0x00000400);
-	nv_icmd(priv, 0x0000ee, 0x00000400);
-	nv_icmd(priv, 0x0000ef, 0x00000400);
-	nv_icmd(priv, 0x000078, 0x00000300);
-	nv_icmd(priv, 0x000079, 0x00000300);
-	nv_icmd(priv, 0x00007a, 0x00000300);
-	nv_icmd(priv, 0x00007b, 0x00000300);
-	nv_icmd(priv, 0x00007c, 0x00000300);
-	nv_icmd(priv, 0x00007d, 0x00000300);
-	nv_icmd(priv, 0x00007e, 0x00000300);
-	nv_icmd(priv, 0x00007f, 0x00000300);
-	nv_icmd(priv, 0x000050, 0x00000011);
-	nv_icmd(priv, 0x000058, 0x00000008);
-	nv_icmd(priv, 0x000059, 0x00000008);
-	nv_icmd(priv, 0x00005a, 0x00000008);
-	nv_icmd(priv, 0x00005b, 0x00000008);
-	nv_icmd(priv, 0x00005c, 0x00000008);
-	nv_icmd(priv, 0x00005d, 0x00000008);
-	nv_icmd(priv, 0x00005e, 0x00000008);
-	nv_icmd(priv, 0x00005f, 0x00000008);
-	nv_icmd(priv, 0x000208, 0x00000001);
-	nv_icmd(priv, 0x000209, 0x00000001);
-	nv_icmd(priv, 0x00020a, 0x00000001);
-	nv_icmd(priv, 0x00020b, 0x00000001);
-	nv_icmd(priv, 0x00020c, 0x00000001);
-	nv_icmd(priv, 0x00020d, 0x00000001);
-	nv_icmd(priv, 0x00020e, 0x00000001);
-	nv_icmd(priv, 0x00020f, 0x00000001);
-	nv_icmd(priv, 0x000081, 0x00000001);
-	nv_icmd(priv, 0x000085, 0x00000004);
-	nv_icmd(priv, 0x000088, 0x00000400);
-	nv_icmd(priv, 0x000090, 0x00000300);
-	nv_icmd(priv, 0x000098, 0x00001001);
-	nv_icmd(priv, 0x0000e3, 0x00000001);
-	nv_icmd(priv, 0x0000da, 0x00000001);
-	nv_icmd(priv, 0x0000f8, 0x00000003);
-	nv_icmd(priv, 0x0000fa, 0x00000001);
-	nv_icmd(priv, 0x00009f, 0x0000ffff);
-	nv_icmd(priv, 0x0000a0, 0x0000ffff);
-	nv_icmd(priv, 0x0000a1, 0x0000ffff);
-	nv_icmd(priv, 0x0000a2, 0x0000ffff);
-	nv_icmd(priv, 0x0000b1, 0x00000001);
-	nv_icmd(priv, 0x0000ad, 0x0000013e);
-	nv_icmd(priv, 0x0000e1, 0x00000010);
-	nv_icmd(priv, 0x000290, 0x00000000);
-	nv_icmd(priv, 0x000291, 0x00000000);
-	nv_icmd(priv, 0x000292, 0x00000000);
-	nv_icmd(priv, 0x000293, 0x00000000);
-	nv_icmd(priv, 0x000294, 0x00000000);
-	nv_icmd(priv, 0x000295, 0x00000000);
-	nv_icmd(priv, 0x000296, 0x00000000);
-	nv_icmd(priv, 0x000297, 0x00000000);
-	nv_icmd(priv, 0x000298, 0x00000000);
-	nv_icmd(priv, 0x000299, 0x00000000);
-	nv_icmd(priv, 0x00029a, 0x00000000);
-	nv_icmd(priv, 0x00029b, 0x00000000);
-	nv_icmd(priv, 0x00029c, 0x00000000);
-	nv_icmd(priv, 0x00029d, 0x00000000);
-	nv_icmd(priv, 0x00029e, 0x00000000);
-	nv_icmd(priv, 0x00029f, 0x00000000);
-	nv_icmd(priv, 0x0003b0, 0x00000000);
-	nv_icmd(priv, 0x0003b1, 0x00000000);
-	nv_icmd(priv, 0x0003b2, 0x00000000);
-	nv_icmd(priv, 0x0003b3, 0x00000000);
-	nv_icmd(priv, 0x0003b4, 0x00000000);
-	nv_icmd(priv, 0x0003b5, 0x00000000);
-	nv_icmd(priv, 0x0003b6, 0x00000000);
-	nv_icmd(priv, 0x0003b7, 0x00000000);
-	nv_icmd(priv, 0x0003b8, 0x00000000);
-	nv_icmd(priv, 0x0003b9, 0x00000000);
-	nv_icmd(priv, 0x0003ba, 0x00000000);
-	nv_icmd(priv, 0x0003bb, 0x00000000);
-	nv_icmd(priv, 0x0003bc, 0x00000000);
-	nv_icmd(priv, 0x0003bd, 0x00000000);
-	nv_icmd(priv, 0x0003be, 0x00000000);
-	nv_icmd(priv, 0x0003bf, 0x00000000);
-	nv_icmd(priv, 0x0002a0, 0x00000000);
-	nv_icmd(priv, 0x0002a1, 0x00000000);
-	nv_icmd(priv, 0x0002a2, 0x00000000);
-	nv_icmd(priv, 0x0002a3, 0x00000000);
-	nv_icmd(priv, 0x0002a4, 0x00000000);
-	nv_icmd(priv, 0x0002a5, 0x00000000);
-	nv_icmd(priv, 0x0002a6, 0x00000000);
-	nv_icmd(priv, 0x0002a7, 0x00000000);
-	nv_icmd(priv, 0x0002a8, 0x00000000);
-	nv_icmd(priv, 0x0002a9, 0x00000000);
-	nv_icmd(priv, 0x0002aa, 0x00000000);
-	nv_icmd(priv, 0x0002ab, 0x00000000);
-	nv_icmd(priv, 0x0002ac, 0x00000000);
-	nv_icmd(priv, 0x0002ad, 0x00000000);
-	nv_icmd(priv, 0x0002ae, 0x00000000);
-	nv_icmd(priv, 0x0002af, 0x00000000);
-	nv_icmd(priv, 0x000420, 0x00000000);
-	nv_icmd(priv, 0x000421, 0x00000000);
-	nv_icmd(priv, 0x000422, 0x00000000);
-	nv_icmd(priv, 0x000423, 0x00000000);
-	nv_icmd(priv, 0x000424, 0x00000000);
-	nv_icmd(priv, 0x000425, 0x00000000);
-	nv_icmd(priv, 0x000426, 0x00000000);
-	nv_icmd(priv, 0x000427, 0x00000000);
-	nv_icmd(priv, 0x000428, 0x00000000);
-	nv_icmd(priv, 0x000429, 0x00000000);
-	nv_icmd(priv, 0x00042a, 0x00000000);
-	nv_icmd(priv, 0x00042b, 0x00000000);
-	nv_icmd(priv, 0x00042c, 0x00000000);
-	nv_icmd(priv, 0x00042d, 0x00000000);
-	nv_icmd(priv, 0x00042e, 0x00000000);
-	nv_icmd(priv, 0x00042f, 0x00000000);
-	nv_icmd(priv, 0x0002b0, 0x00000000);
-	nv_icmd(priv, 0x0002b1, 0x00000000);
-	nv_icmd(priv, 0x0002b2, 0x00000000);
-	nv_icmd(priv, 0x0002b3, 0x00000000);
-	nv_icmd(priv, 0x0002b4, 0x00000000);
-	nv_icmd(priv, 0x0002b5, 0x00000000);
-	nv_icmd(priv, 0x0002b6, 0x00000000);
-	nv_icmd(priv, 0x0002b7, 0x00000000);
-	nv_icmd(priv, 0x0002b8, 0x00000000);
-	nv_icmd(priv, 0x0002b9, 0x00000000);
-	nv_icmd(priv, 0x0002ba, 0x00000000);
-	nv_icmd(priv, 0x0002bb, 0x00000000);
-	nv_icmd(priv, 0x0002bc, 0x00000000);
-	nv_icmd(priv, 0x0002bd, 0x00000000);
-	nv_icmd(priv, 0x0002be, 0x00000000);
-	nv_icmd(priv, 0x0002bf, 0x00000000);
-	nv_icmd(priv, 0x000430, 0x00000000);
-	nv_icmd(priv, 0x000431, 0x00000000);
-	nv_icmd(priv, 0x000432, 0x00000000);
-	nv_icmd(priv, 0x000433, 0x00000000);
-	nv_icmd(priv, 0x000434, 0x00000000);
-	nv_icmd(priv, 0x000435, 0x00000000);
-	nv_icmd(priv, 0x000436, 0x00000000);
-	nv_icmd(priv, 0x000437, 0x00000000);
-	nv_icmd(priv, 0x000438, 0x00000000);
-	nv_icmd(priv, 0x000439, 0x00000000);
-	nv_icmd(priv, 0x00043a, 0x00000000);
-	nv_icmd(priv, 0x00043b, 0x00000000);
-	nv_icmd(priv, 0x00043c, 0x00000000);
-	nv_icmd(priv, 0x00043d, 0x00000000);
-	nv_icmd(priv, 0x00043e, 0x00000000);
-	nv_icmd(priv, 0x00043f, 0x00000000);
-	nv_icmd(priv, 0x0002c0, 0x00000000);
-	nv_icmd(priv, 0x0002c1, 0x00000000);
-	nv_icmd(priv, 0x0002c2, 0x00000000);
-	nv_icmd(priv, 0x0002c3, 0x00000000);
-	nv_icmd(priv, 0x0002c4, 0x00000000);
-	nv_icmd(priv, 0x0002c5, 0x00000000);
-	nv_icmd(priv, 0x0002c6, 0x00000000);
-	nv_icmd(priv, 0x0002c7, 0x00000000);
-	nv_icmd(priv, 0x0002c8, 0x00000000);
-	nv_icmd(priv, 0x0002c9, 0x00000000);
-	nv_icmd(priv, 0x0002ca, 0x00000000);
-	nv_icmd(priv, 0x0002cb, 0x00000000);
-	nv_icmd(priv, 0x0002cc, 0x00000000);
-	nv_icmd(priv, 0x0002cd, 0x00000000);
-	nv_icmd(priv, 0x0002ce, 0x00000000);
-	nv_icmd(priv, 0x0002cf, 0x00000000);
-	nv_icmd(priv, 0x0004d0, 0x00000000);
-	nv_icmd(priv, 0x0004d1, 0x00000000);
-	nv_icmd(priv, 0x0004d2, 0x00000000);
-	nv_icmd(priv, 0x0004d3, 0x00000000);
-	nv_icmd(priv, 0x0004d4, 0x00000000);
-	nv_icmd(priv, 0x0004d5, 0x00000000);
-	nv_icmd(priv, 0x0004d6, 0x00000000);
-	nv_icmd(priv, 0x0004d7, 0x00000000);
-	nv_icmd(priv, 0x0004d8, 0x00000000);
-	nv_icmd(priv, 0x0004d9, 0x00000000);
-	nv_icmd(priv, 0x0004da, 0x00000000);
-	nv_icmd(priv, 0x0004db, 0x00000000);
-	nv_icmd(priv, 0x0004dc, 0x00000000);
-	nv_icmd(priv, 0x0004dd, 0x00000000);
-	nv_icmd(priv, 0x0004de, 0x00000000);
-	nv_icmd(priv, 0x0004df, 0x00000000);
-	nv_icmd(priv, 0x000720, 0x00000000);
-	nv_icmd(priv, 0x000721, 0x00000000);
-	nv_icmd(priv, 0x000722, 0x00000000);
-	nv_icmd(priv, 0x000723, 0x00000000);
-	nv_icmd(priv, 0x000724, 0x00000000);
-	nv_icmd(priv, 0x000725, 0x00000000);
-	nv_icmd(priv, 0x000726, 0x00000000);
-	nv_icmd(priv, 0x000727, 0x00000000);
-	nv_icmd(priv, 0x000728, 0x00000000);
-	nv_icmd(priv, 0x000729, 0x00000000);
-	nv_icmd(priv, 0x00072a, 0x00000000);
-	nv_icmd(priv, 0x00072b, 0x00000000);
-	nv_icmd(priv, 0x00072c, 0x00000000);
-	nv_icmd(priv, 0x00072d, 0x00000000);
-	nv_icmd(priv, 0x00072e, 0x00000000);
-	nv_icmd(priv, 0x00072f, 0x00000000);
-	nv_icmd(priv, 0x0008c0, 0x00000000);
-	nv_icmd(priv, 0x0008c1, 0x00000000);
-	nv_icmd(priv, 0x0008c2, 0x00000000);
-	nv_icmd(priv, 0x0008c3, 0x00000000);
-	nv_icmd(priv, 0x0008c4, 0x00000000);
-	nv_icmd(priv, 0x0008c5, 0x00000000);
-	nv_icmd(priv, 0x0008c6, 0x00000000);
-	nv_icmd(priv, 0x0008c7, 0x00000000);
-	nv_icmd(priv, 0x0008c8, 0x00000000);
-	nv_icmd(priv, 0x0008c9, 0x00000000);
-	nv_icmd(priv, 0x0008ca, 0x00000000);
-	nv_icmd(priv, 0x0008cb, 0x00000000);
-	nv_icmd(priv, 0x0008cc, 0x00000000);
-	nv_icmd(priv, 0x0008cd, 0x00000000);
-	nv_icmd(priv, 0x0008ce, 0x00000000);
-	nv_icmd(priv, 0x0008cf, 0x00000000);
-	nv_icmd(priv, 0x000890, 0x00000000);
-	nv_icmd(priv, 0x000891, 0x00000000);
-	nv_icmd(priv, 0x000892, 0x00000000);
-	nv_icmd(priv, 0x000893, 0x00000000);
-	nv_icmd(priv, 0x000894, 0x00000000);
-	nv_icmd(priv, 0x000895, 0x00000000);
-	nv_icmd(priv, 0x000896, 0x00000000);
-	nv_icmd(priv, 0x000897, 0x00000000);
-	nv_icmd(priv, 0x000898, 0x00000000);
-	nv_icmd(priv, 0x000899, 0x00000000);
-	nv_icmd(priv, 0x00089a, 0x00000000);
-	nv_icmd(priv, 0x00089b, 0x00000000);
-	nv_icmd(priv, 0x00089c, 0x00000000);
-	nv_icmd(priv, 0x00089d, 0x00000000);
-	nv_icmd(priv, 0x00089e, 0x00000000);
-	nv_icmd(priv, 0x00089f, 0x00000000);
-	nv_icmd(priv, 0x0008e0, 0x00000000);
-	nv_icmd(priv, 0x0008e1, 0x00000000);
-	nv_icmd(priv, 0x0008e2, 0x00000000);
-	nv_icmd(priv, 0x0008e3, 0x00000000);
-	nv_icmd(priv, 0x0008e4, 0x00000000);
-	nv_icmd(priv, 0x0008e5, 0x00000000);
-	nv_icmd(priv, 0x0008e6, 0x00000000);
-	nv_icmd(priv, 0x0008e7, 0x00000000);
-	nv_icmd(priv, 0x0008e8, 0x00000000);
-	nv_icmd(priv, 0x0008e9, 0x00000000);
-	nv_icmd(priv, 0x0008ea, 0x00000000);
-	nv_icmd(priv, 0x0008eb, 0x00000000);
-	nv_icmd(priv, 0x0008ec, 0x00000000);
-	nv_icmd(priv, 0x0008ed, 0x00000000);
-	nv_icmd(priv, 0x0008ee, 0x00000000);
-	nv_icmd(priv, 0x0008ef, 0x00000000);
-	nv_icmd(priv, 0x0008a0, 0x00000000);
-	nv_icmd(priv, 0x0008a1, 0x00000000);
-	nv_icmd(priv, 0x0008a2, 0x00000000);
-	nv_icmd(priv, 0x0008a3, 0x00000000);
-	nv_icmd(priv, 0x0008a4, 0x00000000);
-	nv_icmd(priv, 0x0008a5, 0x00000000);
-	nv_icmd(priv, 0x0008a6, 0x00000000);
-	nv_icmd(priv, 0x0008a7, 0x00000000);
-	nv_icmd(priv, 0x0008a8, 0x00000000);
-	nv_icmd(priv, 0x0008a9, 0x00000000);
-	nv_icmd(priv, 0x0008aa, 0x00000000);
-	nv_icmd(priv, 0x0008ab, 0x00000000);
-	nv_icmd(priv, 0x0008ac, 0x00000000);
-	nv_icmd(priv, 0x0008ad, 0x00000000);
-	nv_icmd(priv, 0x0008ae, 0x00000000);
-	nv_icmd(priv, 0x0008af, 0x00000000);
-	nv_icmd(priv, 0x0008f0, 0x00000000);
-	nv_icmd(priv, 0x0008f1, 0x00000000);
-	nv_icmd(priv, 0x0008f2, 0x00000000);
-	nv_icmd(priv, 0x0008f3, 0x00000000);
-	nv_icmd(priv, 0x0008f4, 0x00000000);
-	nv_icmd(priv, 0x0008f5, 0x00000000);
-	nv_icmd(priv, 0x0008f6, 0x00000000);
-	nv_icmd(priv, 0x0008f7, 0x00000000);
-	nv_icmd(priv, 0x0008f8, 0x00000000);
-	nv_icmd(priv, 0x0008f9, 0x00000000);
-	nv_icmd(priv, 0x0008fa, 0x00000000);
-	nv_icmd(priv, 0x0008fb, 0x00000000);
-	nv_icmd(priv, 0x0008fc, 0x00000000);
-	nv_icmd(priv, 0x0008fd, 0x00000000);
-	nv_icmd(priv, 0x0008fe, 0x00000000);
-	nv_icmd(priv, 0x0008ff, 0x00000000);
-	nv_icmd(priv, 0x00094c, 0x000000ff);
-	nv_icmd(priv, 0x00094d, 0xffffffff);
-	nv_icmd(priv, 0x00094e, 0x00000002);
-	nv_icmd(priv, 0x0002ec, 0x00000001);
-	nv_icmd(priv, 0x000303, 0x00000001);
-	nv_icmd(priv, 0x0002e6, 0x00000001);
-	nv_icmd(priv, 0x000466, 0x00000052);
-	nv_icmd(priv, 0x000301, 0x3f800000);
-	nv_icmd(priv, 0x000304, 0x30201000);
-	nv_icmd(priv, 0x000305, 0x70605040);
-	nv_icmd(priv, 0x000306, 0xb8a89888);
-	nv_icmd(priv, 0x000307, 0xf8e8d8c8);
-	nv_icmd(priv, 0x00030a, 0x00ffff00);
-	nv_icmd(priv, 0x00030b, 0x0000001a);
-	nv_icmd(priv, 0x00030c, 0x00000001);
-	nv_icmd(priv, 0x000318, 0x00000001);
-	nv_icmd(priv, 0x000340, 0x00000000);
-	nv_icmd(priv, 0x000375, 0x00000001);
-	nv_icmd(priv, 0x00037d, 0x00000006);
-	nv_icmd(priv, 0x0003a0, 0x00000002);
-	nv_icmd(priv, 0x0003aa, 0x00000001);
-	nv_icmd(priv, 0x0003a9, 0x00000001);
-	nv_icmd(priv, 0x000380, 0x00000001);
-	nv_icmd(priv, 0x000383, 0x00000011);
-	nv_icmd(priv, 0x000360, 0x00000040);
-	nv_icmd(priv, 0x000366, 0x00000000);
-	nv_icmd(priv, 0x000367, 0x00000000);
-	nv_icmd(priv, 0x000368, 0x00000fff);
-	nv_icmd(priv, 0x000370, 0x00000000);
-	nv_icmd(priv, 0x000371, 0x00000000);
-	nv_icmd(priv, 0x000372, 0x000fffff);
-	nv_icmd(priv, 0x00037a, 0x00000012);
-	nv_icmd(priv, 0x000619, 0x00000003);
-	nv_icmd(priv, 0x000811, 0x00000003);
-	nv_icmd(priv, 0x000812, 0x00000004);
-	nv_icmd(priv, 0x000813, 0x00000006);
-	nv_icmd(priv, 0x000814, 0x00000008);
-	nv_icmd(priv, 0x000815, 0x0000000b);
-	nv_icmd(priv, 0x000800, 0x00000001);
-	nv_icmd(priv, 0x000801, 0x00000001);
-	nv_icmd(priv, 0x000802, 0x00000001);
-	nv_icmd(priv, 0x000803, 0x00000001);
-	nv_icmd(priv, 0x000804, 0x00000001);
-	nv_icmd(priv, 0x000805, 0x00000001);
-	nv_icmd(priv, 0x000632, 0x00000001);
-	nv_icmd(priv, 0x000633, 0x00000002);
-	nv_icmd(priv, 0x000634, 0x00000003);
-	nv_icmd(priv, 0x000635, 0x00000004);
-	nv_icmd(priv, 0x000654, 0x3f800000);
-	nv_icmd(priv, 0x000657, 0x3f800000);
-	nv_icmd(priv, 0x000655, 0x3f800000);
-	nv_icmd(priv, 0x000656, 0x3f800000);
-	nv_icmd(priv, 0x0006cd, 0x3f800000);
-	nv_icmd(priv, 0x0007f5, 0x3f800000);
-	nv_icmd(priv, 0x0007dc, 0x39291909);
-	nv_icmd(priv, 0x0007dd, 0x79695949);
-	nv_icmd(priv, 0x0007de, 0xb9a99989);
-	nv_icmd(priv, 0x0007df, 0xf9e9d9c9);
-	nv_icmd(priv, 0x0007e8, 0x00003210);
-	nv_icmd(priv, 0x0007e9, 0x00007654);
-	nv_icmd(priv, 0x0007ea, 0x00000098);
-	nv_icmd(priv, 0x0007ec, 0x39291909);
-	nv_icmd(priv, 0x0007ed, 0x79695949);
-	nv_icmd(priv, 0x0007ee, 0xb9a99989);
-	nv_icmd(priv, 0x0007ef, 0xf9e9d9c9);
-	nv_icmd(priv, 0x0007f0, 0x00003210);
-	nv_icmd(priv, 0x0007f1, 0x00007654);
-	nv_icmd(priv, 0x0007f2, 0x00000098);
-	nv_icmd(priv, 0x0005a5, 0x00000001);
-	nv_icmd(priv, 0x000980, 0x00000000);
-	nv_icmd(priv, 0x000981, 0x00000000);
-	nv_icmd(priv, 0x000982, 0x00000000);
-	nv_icmd(priv, 0x000983, 0x00000000);
-	nv_icmd(priv, 0x000984, 0x00000000);
-	nv_icmd(priv, 0x000985, 0x00000000);
-	nv_icmd(priv, 0x000986, 0x00000000);
-	nv_icmd(priv, 0x000987, 0x00000000);
-	nv_icmd(priv, 0x000988, 0x00000000);
-	nv_icmd(priv, 0x000989, 0x00000000);
-	nv_icmd(priv, 0x00098a, 0x00000000);
-	nv_icmd(priv, 0x00098b, 0x00000000);
-	nv_icmd(priv, 0x00098c, 0x00000000);
-	nv_icmd(priv, 0x00098d, 0x00000000);
-	nv_icmd(priv, 0x00098e, 0x00000000);
-	nv_icmd(priv, 0x00098f, 0x00000000);
-	nv_icmd(priv, 0x000990, 0x00000000);
-	nv_icmd(priv, 0x000991, 0x00000000);
-	nv_icmd(priv, 0x000992, 0x00000000);
-	nv_icmd(priv, 0x000993, 0x00000000);
-	nv_icmd(priv, 0x000994, 0x00000000);
-	nv_icmd(priv, 0x000995, 0x00000000);
-	nv_icmd(priv, 0x000996, 0x00000000);
-	nv_icmd(priv, 0x000997, 0x00000000);
-	nv_icmd(priv, 0x000998, 0x00000000);
-	nv_icmd(priv, 0x000999, 0x00000000);
-	nv_icmd(priv, 0x00099a, 0x00000000);
-	nv_icmd(priv, 0x00099b, 0x00000000);
-	nv_icmd(priv, 0x00099c, 0x00000000);
-	nv_icmd(priv, 0x00099d, 0x00000000);
-	nv_icmd(priv, 0x00099e, 0x00000000);
-	nv_icmd(priv, 0x00099f, 0x00000000);
-	nv_icmd(priv, 0x0009a0, 0x00000000);
-	nv_icmd(priv, 0x0009a1, 0x00000000);
-	nv_icmd(priv, 0x0009a2, 0x00000000);
-	nv_icmd(priv, 0x0009a3, 0x00000000);
-	nv_icmd(priv, 0x0009a4, 0x00000000);
-	nv_icmd(priv, 0x0009a5, 0x00000000);
-	nv_icmd(priv, 0x0009a6, 0x00000000);
-	nv_icmd(priv, 0x0009a7, 0x00000000);
-	nv_icmd(priv, 0x0009a8, 0x00000000);
-	nv_icmd(priv, 0x0009a9, 0x00000000);
-	nv_icmd(priv, 0x0009aa, 0x00000000);
-	nv_icmd(priv, 0x0009ab, 0x00000000);
-	nv_icmd(priv, 0x0009ac, 0x00000000);
-	nv_icmd(priv, 0x0009ad, 0x00000000);
-	nv_icmd(priv, 0x0009ae, 0x00000000);
-	nv_icmd(priv, 0x0009af, 0x00000000);
-	nv_icmd(priv, 0x0009b0, 0x00000000);
-	nv_icmd(priv, 0x0009b1, 0x00000000);
-	nv_icmd(priv, 0x0009b2, 0x00000000);
-	nv_icmd(priv, 0x0009b3, 0x00000000);
-	nv_icmd(priv, 0x0009b4, 0x00000000);
-	nv_icmd(priv, 0x0009b5, 0x00000000);
-	nv_icmd(priv, 0x0009b6, 0x00000000);
-	nv_icmd(priv, 0x0009b7, 0x00000000);
-	nv_icmd(priv, 0x0009b8, 0x00000000);
-	nv_icmd(priv, 0x0009b9, 0x00000000);
-	nv_icmd(priv, 0x0009ba, 0x00000000);
-	nv_icmd(priv, 0x0009bb, 0x00000000);
-	nv_icmd(priv, 0x0009bc, 0x00000000);
-	nv_icmd(priv, 0x0009bd, 0x00000000);
-	nv_icmd(priv, 0x0009be, 0x00000000);
-	nv_icmd(priv, 0x0009bf, 0x00000000);
-	nv_icmd(priv, 0x0009c0, 0x00000000);
-	nv_icmd(priv, 0x0009c1, 0x00000000);
-	nv_icmd(priv, 0x0009c2, 0x00000000);
-	nv_icmd(priv, 0x0009c3, 0x00000000);
-	nv_icmd(priv, 0x0009c4, 0x00000000);
-	nv_icmd(priv, 0x0009c5, 0x00000000);
-	nv_icmd(priv, 0x0009c6, 0x00000000);
-	nv_icmd(priv, 0x0009c7, 0x00000000);
-	nv_icmd(priv, 0x0009c8, 0x00000000);
-	nv_icmd(priv, 0x0009c9, 0x00000000);
-	nv_icmd(priv, 0x0009ca, 0x00000000);
-	nv_icmd(priv, 0x0009cb, 0x00000000);
-	nv_icmd(priv, 0x0009cc, 0x00000000);
-	nv_icmd(priv, 0x0009cd, 0x00000000);
-	nv_icmd(priv, 0x0009ce, 0x00000000);
-	nv_icmd(priv, 0x0009cf, 0x00000000);
-	nv_icmd(priv, 0x0009d0, 0x00000000);
-	nv_icmd(priv, 0x0009d1, 0x00000000);
-	nv_icmd(priv, 0x0009d2, 0x00000000);
-	nv_icmd(priv, 0x0009d3, 0x00000000);
-	nv_icmd(priv, 0x0009d4, 0x00000000);
-	nv_icmd(priv, 0x0009d5, 0x00000000);
-	nv_icmd(priv, 0x0009d6, 0x00000000);
-	nv_icmd(priv, 0x0009d7, 0x00000000);
-	nv_icmd(priv, 0x0009d8, 0x00000000);
-	nv_icmd(priv, 0x0009d9, 0x00000000);
-	nv_icmd(priv, 0x0009da, 0x00000000);
-	nv_icmd(priv, 0x0009db, 0x00000000);
-	nv_icmd(priv, 0x0009dc, 0x00000000);
-	nv_icmd(priv, 0x0009dd, 0x00000000);
-	nv_icmd(priv, 0x0009de, 0x00000000);
-	nv_icmd(priv, 0x0009df, 0x00000000);
-	nv_icmd(priv, 0x0009e0, 0x00000000);
-	nv_icmd(priv, 0x0009e1, 0x00000000);
-	nv_icmd(priv, 0x0009e2, 0x00000000);
-	nv_icmd(priv, 0x0009e3, 0x00000000);
-	nv_icmd(priv, 0x0009e4, 0x00000000);
-	nv_icmd(priv, 0x0009e5, 0x00000000);
-	nv_icmd(priv, 0x0009e6, 0x00000000);
-	nv_icmd(priv, 0x0009e7, 0x00000000);
-	nv_icmd(priv, 0x0009e8, 0x00000000);
-	nv_icmd(priv, 0x0009e9, 0x00000000);
-	nv_icmd(priv, 0x0009ea, 0x00000000);
-	nv_icmd(priv, 0x0009eb, 0x00000000);
-	nv_icmd(priv, 0x0009ec, 0x00000000);
-	nv_icmd(priv, 0x0009ed, 0x00000000);
-	nv_icmd(priv, 0x0009ee, 0x00000000);
-	nv_icmd(priv, 0x0009ef, 0x00000000);
-	nv_icmd(priv, 0x0009f0, 0x00000000);
-	nv_icmd(priv, 0x0009f1, 0x00000000);
-	nv_icmd(priv, 0x0009f2, 0x00000000);
-	nv_icmd(priv, 0x0009f3, 0x00000000);
-	nv_icmd(priv, 0x0009f4, 0x00000000);
-	nv_icmd(priv, 0x0009f5, 0x00000000);
-	nv_icmd(priv, 0x0009f6, 0x00000000);
-	nv_icmd(priv, 0x0009f7, 0x00000000);
-	nv_icmd(priv, 0x0009f8, 0x00000000);
-	nv_icmd(priv, 0x0009f9, 0x00000000);
-	nv_icmd(priv, 0x0009fa, 0x00000000);
-	nv_icmd(priv, 0x0009fb, 0x00000000);
-	nv_icmd(priv, 0x0009fc, 0x00000000);
-	nv_icmd(priv, 0x0009fd, 0x00000000);
-	nv_icmd(priv, 0x0009fe, 0x00000000);
-	nv_icmd(priv, 0x0009ff, 0x00000000);
-	nv_icmd(priv, 0x000468, 0x00000004);
-	nv_icmd(priv, 0x00046c, 0x00000001);
-	nv_icmd(priv, 0x000470, 0x00000000);
-	nv_icmd(priv, 0x000471, 0x00000000);
-	nv_icmd(priv, 0x000472, 0x00000000);
-	nv_icmd(priv, 0x000473, 0x00000000);
-	nv_icmd(priv, 0x000474, 0x00000000);
-	nv_icmd(priv, 0x000475, 0x00000000);
-	nv_icmd(priv, 0x000476, 0x00000000);
-	nv_icmd(priv, 0x000477, 0x00000000);
-	nv_icmd(priv, 0x000478, 0x00000000);
-	nv_icmd(priv, 0x000479, 0x00000000);
-	nv_icmd(priv, 0x00047a, 0x00000000);
-	nv_icmd(priv, 0x00047b, 0x00000000);
-	nv_icmd(priv, 0x00047c, 0x00000000);
-	nv_icmd(priv, 0x00047d, 0x00000000);
-	nv_icmd(priv, 0x00047e, 0x00000000);
-	nv_icmd(priv, 0x00047f, 0x00000000);
-	nv_icmd(priv, 0x000480, 0x00000000);
-	nv_icmd(priv, 0x000481, 0x00000000);
-	nv_icmd(priv, 0x000482, 0x00000000);
-	nv_icmd(priv, 0x000483, 0x00000000);
-	nv_icmd(priv, 0x000484, 0x00000000);
-	nv_icmd(priv, 0x000485, 0x00000000);
-	nv_icmd(priv, 0x000486, 0x00000000);
-	nv_icmd(priv, 0x000487, 0x00000000);
-	nv_icmd(priv, 0x000488, 0x00000000);
-	nv_icmd(priv, 0x000489, 0x00000000);
-	nv_icmd(priv, 0x00048a, 0x00000000);
-	nv_icmd(priv, 0x00048b, 0x00000000);
-	nv_icmd(priv, 0x00048c, 0x00000000);
-	nv_icmd(priv, 0x00048d, 0x00000000);
-	nv_icmd(priv, 0x00048e, 0x00000000);
-	nv_icmd(priv, 0x00048f, 0x00000000);
-	nv_icmd(priv, 0x000490, 0x00000000);
-	nv_icmd(priv, 0x000491, 0x00000000);
-	nv_icmd(priv, 0x000492, 0x00000000);
-	nv_icmd(priv, 0x000493, 0x00000000);
-	nv_icmd(priv, 0x000494, 0x00000000);
-	nv_icmd(priv, 0x000495, 0x00000000);
-	nv_icmd(priv, 0x000496, 0x00000000);
-	nv_icmd(priv, 0x000497, 0x00000000);
-	nv_icmd(priv, 0x000498, 0x00000000);
-	nv_icmd(priv, 0x000499, 0x00000000);
-	nv_icmd(priv, 0x00049a, 0x00000000);
-	nv_icmd(priv, 0x00049b, 0x00000000);
-	nv_icmd(priv, 0x00049c, 0x00000000);
-	nv_icmd(priv, 0x00049d, 0x00000000);
-	nv_icmd(priv, 0x00049e, 0x00000000);
-	nv_icmd(priv, 0x00049f, 0x00000000);
-	nv_icmd(priv, 0x0004a0, 0x00000000);
-	nv_icmd(priv, 0x0004a1, 0x00000000);
-	nv_icmd(priv, 0x0004a2, 0x00000000);
-	nv_icmd(priv, 0x0004a3, 0x00000000);
-	nv_icmd(priv, 0x0004a4, 0x00000000);
-	nv_icmd(priv, 0x0004a5, 0x00000000);
-	nv_icmd(priv, 0x0004a6, 0x00000000);
-	nv_icmd(priv, 0x0004a7, 0x00000000);
-	nv_icmd(priv, 0x0004a8, 0x00000000);
-	nv_icmd(priv, 0x0004a9, 0x00000000);
-	nv_icmd(priv, 0x0004aa, 0x00000000);
-	nv_icmd(priv, 0x0004ab, 0x00000000);
-	nv_icmd(priv, 0x0004ac, 0x00000000);
-	nv_icmd(priv, 0x0004ad, 0x00000000);
-	nv_icmd(priv, 0x0004ae, 0x00000000);
-	nv_icmd(priv, 0x0004af, 0x00000000);
-	nv_icmd(priv, 0x0004b0, 0x00000000);
-	nv_icmd(priv, 0x0004b1, 0x00000000);
-	nv_icmd(priv, 0x0004b2, 0x00000000);
-	nv_icmd(priv, 0x0004b3, 0x00000000);
-	nv_icmd(priv, 0x0004b4, 0x00000000);
-	nv_icmd(priv, 0x0004b5, 0x00000000);
-	nv_icmd(priv, 0x0004b6, 0x00000000);
-	nv_icmd(priv, 0x0004b7, 0x00000000);
-	nv_icmd(priv, 0x0004b8, 0x00000000);
-	nv_icmd(priv, 0x0004b9, 0x00000000);
-	nv_icmd(priv, 0x0004ba, 0x00000000);
-	nv_icmd(priv, 0x0004bb, 0x00000000);
-	nv_icmd(priv, 0x0004bc, 0x00000000);
-	nv_icmd(priv, 0x0004bd, 0x00000000);
-	nv_icmd(priv, 0x0004be, 0x00000000);
-	nv_icmd(priv, 0x0004bf, 0x00000000);
-	nv_icmd(priv, 0x0004c0, 0x00000000);
-	nv_icmd(priv, 0x0004c1, 0x00000000);
-	nv_icmd(priv, 0x0004c2, 0x00000000);
-	nv_icmd(priv, 0x0004c3, 0x00000000);
-	nv_icmd(priv, 0x0004c4, 0x00000000);
-	nv_icmd(priv, 0x0004c5, 0x00000000);
-	nv_icmd(priv, 0x0004c6, 0x00000000);
-	nv_icmd(priv, 0x0004c7, 0x00000000);
-	nv_icmd(priv, 0x0004c8, 0x00000000);
-	nv_icmd(priv, 0x0004c9, 0x00000000);
-	nv_icmd(priv, 0x0004ca, 0x00000000);
-	nv_icmd(priv, 0x0004cb, 0x00000000);
-	nv_icmd(priv, 0x0004cc, 0x00000000);
-	nv_icmd(priv, 0x0004cd, 0x00000000);
-	nv_icmd(priv, 0x0004ce, 0x00000000);
-	nv_icmd(priv, 0x0004cf, 0x00000000);
-	nv_icmd(priv, 0x000510, 0x3f800000);
-	nv_icmd(priv, 0x000511, 0x3f800000);
-	nv_icmd(priv, 0x000512, 0x3f800000);
-	nv_icmd(priv, 0x000513, 0x3f800000);
-	nv_icmd(priv, 0x000514, 0x3f800000);
-	nv_icmd(priv, 0x000515, 0x3f800000);
-	nv_icmd(priv, 0x000516, 0x3f800000);
-	nv_icmd(priv, 0x000517, 0x3f800000);
-	nv_icmd(priv, 0x000518, 0x3f800000);
-	nv_icmd(priv, 0x000519, 0x3f800000);
-	nv_icmd(priv, 0x00051a, 0x3f800000);
-	nv_icmd(priv, 0x00051b, 0x3f800000);
-	nv_icmd(priv, 0x00051c, 0x3f800000);
-	nv_icmd(priv, 0x00051d, 0x3f800000);
-	nv_icmd(priv, 0x00051e, 0x3f800000);
-	nv_icmd(priv, 0x00051f, 0x3f800000);
-	nv_icmd(priv, 0x000520, 0x000002b6);
-	nv_icmd(priv, 0x000529, 0x00000001);
-	nv_icmd(priv, 0x000530, 0xffff0000);
-	nv_icmd(priv, 0x000531, 0xffff0000);
-	nv_icmd(priv, 0x000532, 0xffff0000);
-	nv_icmd(priv, 0x000533, 0xffff0000);
-	nv_icmd(priv, 0x000534, 0xffff0000);
-	nv_icmd(priv, 0x000535, 0xffff0000);
-	nv_icmd(priv, 0x000536, 0xffff0000);
-	nv_icmd(priv, 0x000537, 0xffff0000);
-	nv_icmd(priv, 0x000538, 0xffff0000);
-	nv_icmd(priv, 0x000539, 0xffff0000);
-	nv_icmd(priv, 0x00053a, 0xffff0000);
-	nv_icmd(priv, 0x00053b, 0xffff0000);
-	nv_icmd(priv, 0x00053c, 0xffff0000);
-	nv_icmd(priv, 0x00053d, 0xffff0000);
-	nv_icmd(priv, 0x00053e, 0xffff0000);
-	nv_icmd(priv, 0x00053f, 0xffff0000);
-	nv_icmd(priv, 0x000585, 0x0000003f);
-	nv_icmd(priv, 0x000576, 0x00000003);
-	nv_icmd(priv, 0x00057b, 0x00000059);
-	nv_icmd(priv, 0x000586, 0x00000040);
-	nv_icmd(priv, 0x000582, 0x00000080);
-	nv_icmd(priv, 0x000583, 0x00000080);
-	nv_icmd(priv, 0x0005c2, 0x00000001);
-	nv_icmd(priv, 0x000638, 0x00000001);
-	nv_icmd(priv, 0x000639, 0x00000001);
-	nv_icmd(priv, 0x00063a, 0x00000002);
-	nv_icmd(priv, 0x00063b, 0x00000001);
-	nv_icmd(priv, 0x00063c, 0x00000001);
-	nv_icmd(priv, 0x00063d, 0x00000002);
-	nv_icmd(priv, 0x00063e, 0x00000001);
-	nv_icmd(priv, 0x0008b8, 0x00000001);
-	nv_icmd(priv, 0x0008b9, 0x00000001);
-	nv_icmd(priv, 0x0008ba, 0x00000001);
-	nv_icmd(priv, 0x0008bb, 0x00000001);
-	nv_icmd(priv, 0x0008bc, 0x00000001);
-	nv_icmd(priv, 0x0008bd, 0x00000001);
-	nv_icmd(priv, 0x0008be, 0x00000001);
-	nv_icmd(priv, 0x0008bf, 0x00000001);
-	nv_icmd(priv, 0x000900, 0x00000001);
-	nv_icmd(priv, 0x000901, 0x00000001);
-	nv_icmd(priv, 0x000902, 0x00000001);
-	nv_icmd(priv, 0x000903, 0x00000001);
-	nv_icmd(priv, 0x000904, 0x00000001);
-	nv_icmd(priv, 0x000905, 0x00000001);
-	nv_icmd(priv, 0x000906, 0x00000001);
-	nv_icmd(priv, 0x000907, 0x00000001);
-	nv_icmd(priv, 0x000908, 0x00000002);
-	nv_icmd(priv, 0x000909, 0x00000002);
-	nv_icmd(priv, 0x00090a, 0x00000002);
-	nv_icmd(priv, 0x00090b, 0x00000002);
-	nv_icmd(priv, 0x00090c, 0x00000002);
-	nv_icmd(priv, 0x00090d, 0x00000002);
-	nv_icmd(priv, 0x00090e, 0x00000002);
-	nv_icmd(priv, 0x00090f, 0x00000002);
-	nv_icmd(priv, 0x000910, 0x00000001);
-	nv_icmd(priv, 0x000911, 0x00000001);
-	nv_icmd(priv, 0x000912, 0x00000001);
-	nv_icmd(priv, 0x000913, 0x00000001);
-	nv_icmd(priv, 0x000914, 0x00000001);
-	nv_icmd(priv, 0x000915, 0x00000001);
-	nv_icmd(priv, 0x000916, 0x00000001);
-	nv_icmd(priv, 0x000917, 0x00000001);
-	nv_icmd(priv, 0x000918, 0x00000001);
-	nv_icmd(priv, 0x000919, 0x00000001);
-	nv_icmd(priv, 0x00091a, 0x00000001);
-	nv_icmd(priv, 0x00091b, 0x00000001);
-	nv_icmd(priv, 0x00091c, 0x00000001);
-	nv_icmd(priv, 0x00091d, 0x00000001);
-	nv_icmd(priv, 0x00091e, 0x00000001);
-	nv_icmd(priv, 0x00091f, 0x00000001);
-	nv_icmd(priv, 0x000920, 0x00000002);
-	nv_icmd(priv, 0x000921, 0x00000002);
-	nv_icmd(priv, 0x000922, 0x00000002);
-	nv_icmd(priv, 0x000923, 0x00000002);
-	nv_icmd(priv, 0x000924, 0x00000002);
-	nv_icmd(priv, 0x000925, 0x00000002);
-	nv_icmd(priv, 0x000926, 0x00000002);
-	nv_icmd(priv, 0x000927, 0x00000002);
-	nv_icmd(priv, 0x000928, 0x00000001);
-	nv_icmd(priv, 0x000929, 0x00000001);
-	nv_icmd(priv, 0x00092a, 0x00000001);
-	nv_icmd(priv, 0x00092b, 0x00000001);
-	nv_icmd(priv, 0x00092c, 0x00000001);
-	nv_icmd(priv, 0x00092d, 0x00000001);
-	nv_icmd(priv, 0x00092e, 0x00000001);
-	nv_icmd(priv, 0x00092f, 0x00000001);
-	nv_icmd(priv, 0x000648, 0x00000001);
-	nv_icmd(priv, 0x000649, 0x00000001);
-	nv_icmd(priv, 0x00064a, 0x00000001);
-	nv_icmd(priv, 0x00064b, 0x00000001);
-	nv_icmd(priv, 0x00064c, 0x00000001);
-	nv_icmd(priv, 0x00064d, 0x00000001);
-	nv_icmd(priv, 0x00064e, 0x00000001);
-	nv_icmd(priv, 0x00064f, 0x00000001);
-	nv_icmd(priv, 0x000650, 0x00000001);
-	nv_icmd(priv, 0x000658, 0x0000000f);
-	nv_icmd(priv, 0x0007ff, 0x0000000a);
-	nv_icmd(priv, 0x00066a, 0x40000000);
-	nv_icmd(priv, 0x00066b, 0x10000000);
-	nv_icmd(priv, 0x00066c, 0xffff0000);
-	nv_icmd(priv, 0x00066d, 0xffff0000);
-	nv_icmd(priv, 0x0007af, 0x00000008);
-	nv_icmd(priv, 0x0007b0, 0x00000008);
-	nv_icmd(priv, 0x0007f6, 0x00000001);
-	nv_icmd(priv, 0x0006b2, 0x00000055);
-	nv_icmd(priv, 0x0007ad, 0x00000003);
-	nv_icmd(priv, 0x000937, 0x00000001);
-	nv_icmd(priv, 0x000971, 0x00000008);
-	nv_icmd(priv, 0x000972, 0x00000040);
-	nv_icmd(priv, 0x000973, 0x0000012c);
-	nv_icmd(priv, 0x00097c, 0x00000040);
-	nv_icmd(priv, 0x000979, 0x00000003);
-	nv_icmd(priv, 0x000975, 0x00000020);
-	nv_icmd(priv, 0x000976, 0x00000001);
-	nv_icmd(priv, 0x000977, 0x00000020);
-	nv_icmd(priv, 0x000978, 0x00000001);
-	nv_icmd(priv, 0x000957, 0x00000003);
-	nv_icmd(priv, 0x00095e, 0x20164010);
-	nv_icmd(priv, 0x00095f, 0x00000020);
-	nv_icmd(priv, 0x00097d, 0x00000020);
-	nv_icmd(priv, 0x000683, 0x00000006);
-	nv_icmd(priv, 0x000685, 0x003fffff);
-	nv_icmd(priv, 0x000687, 0x003fffff);
-	nv_icmd(priv, 0x0006a0, 0x00000005);
-	nv_icmd(priv, 0x000840, 0x00400008);
-	nv_icmd(priv, 0x000841, 0x08000080);
-	nv_icmd(priv, 0x000842, 0x00400008);
-	nv_icmd(priv, 0x000843, 0x08000080);
-	nv_icmd(priv, 0x000818, 0x00000000);
-	nv_icmd(priv, 0x000819, 0x00000000);
-	nv_icmd(priv, 0x00081a, 0x00000000);
-	nv_icmd(priv, 0x00081b, 0x00000000);
-	nv_icmd(priv, 0x00081c, 0x00000000);
-	nv_icmd(priv, 0x00081d, 0x00000000);
-	nv_icmd(priv, 0x00081e, 0x00000000);
-	nv_icmd(priv, 0x00081f, 0x00000000);
-	nv_icmd(priv, 0x000848, 0x00000000);
-	nv_icmd(priv, 0x000849, 0x00000000);
-	nv_icmd(priv, 0x00084a, 0x00000000);
-	nv_icmd(priv, 0x00084b, 0x00000000);
-	nv_icmd(priv, 0x00084c, 0x00000000);
-	nv_icmd(priv, 0x00084d, 0x00000000);
-	nv_icmd(priv, 0x00084e, 0x00000000);
-	nv_icmd(priv, 0x00084f, 0x00000000);
-	nv_icmd(priv, 0x000850, 0x00000000);
-	nv_icmd(priv, 0x000851, 0x00000000);
-	nv_icmd(priv, 0x000852, 0x00000000);
-	nv_icmd(priv, 0x000853, 0x00000000);
-	nv_icmd(priv, 0x000854, 0x00000000);
-	nv_icmd(priv, 0x000855, 0x00000000);
-	nv_icmd(priv, 0x000856, 0x00000000);
-	nv_icmd(priv, 0x000857, 0x00000000);
-	nv_icmd(priv, 0x000738, 0x00000000);
-	nv_icmd(priv, 0x0006aa, 0x00000001);
-	nv_icmd(priv, 0x0006ab, 0x00000002);
-	nv_icmd(priv, 0x0006ac, 0x00000080);
-	nv_icmd(priv, 0x0006ad, 0x00000100);
-	nv_icmd(priv, 0x0006ae, 0x00000100);
-	nv_icmd(priv, 0x0006b1, 0x00000011);
-	nv_icmd(priv, 0x0006bb, 0x000000cf);
-	nv_icmd(priv, 0x0006ce, 0x2a712488);
-	nv_icmd(priv, 0x000739, 0x4085c000);
-	nv_icmd(priv, 0x00073a, 0x00000080);
-	nv_icmd(priv, 0x000786, 0x80000100);
-	nv_icmd(priv, 0x00073c, 0x00010100);
-	nv_icmd(priv, 0x00073d, 0x02800000);
-	nv_icmd(priv, 0x000787, 0x000000cf);
-	nv_icmd(priv, 0x00078c, 0x00000008);
-	nv_icmd(priv, 0x000792, 0x00000001);
-	nv_icmd(priv, 0x000794, 0x00000001);
-	nv_icmd(priv, 0x000795, 0x00000001);
-	nv_icmd(priv, 0x000796, 0x00000001);
-	nv_icmd(priv, 0x000797, 0x000000cf);
-	nv_icmd(priv, 0x000836, 0x00000001);
-	nv_icmd(priv, 0x00079a, 0x00000002);
-	nv_icmd(priv, 0x000833, 0x04444480);
-	nv_icmd(priv, 0x0007a1, 0x00000001);
-	nv_icmd(priv, 0x0007a3, 0x00000001);
-	nv_icmd(priv, 0x0007a4, 0x00000001);
-	nv_icmd(priv, 0x0007a5, 0x00000001);
-	nv_icmd(priv, 0x000831, 0x00000004);
-	nv_icmd(priv, 0x000b07, 0x00000002);
-	nv_icmd(priv, 0x000b08, 0x00000100);
-	nv_icmd(priv, 0x000b09, 0x00000100);
-	nv_icmd(priv, 0x000b0a, 0x00000001);
-	nv_icmd(priv, 0x000a04, 0x000000ff);
-	nv_icmd(priv, 0x000a0b, 0x00000040);
-	nv_icmd(priv, 0x00097f, 0x00000100);
-	nv_icmd(priv, 0x000a02, 0x00000001);
-	nv_icmd(priv, 0x000809, 0x00000007);
-	nv_icmd(priv, 0x00c221, 0x00000040);
-	nv_icmd(priv, 0x00c1b0, 0x0000000f);
-	nv_icmd(priv, 0x00c1b1, 0x0000000f);
-	nv_icmd(priv, 0x00c1b2, 0x0000000f);
-	nv_icmd(priv, 0x00c1b3, 0x0000000f);
-	nv_icmd(priv, 0x00c1b4, 0x0000000f);
-	nv_icmd(priv, 0x00c1b5, 0x0000000f);
-	nv_icmd(priv, 0x00c1b6, 0x0000000f);
-	nv_icmd(priv, 0x00c1b7, 0x0000000f);
-	nv_icmd(priv, 0x00c1b8, 0x0fac6881);
-	nv_icmd(priv, 0x00c1b9, 0x00fac688);
-	nv_icmd(priv, 0x00c401, 0x00000001);
-	nv_icmd(priv, 0x00c402, 0x00010001);
-	nv_icmd(priv, 0x00c403, 0x00000001);
-	nv_icmd(priv, 0x00c404, 0x00000001);
-	nv_icmd(priv, 0x00c40e, 0x00000020);
-	nv_icmd(priv, 0x00c500, 0x00000003);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_icmd(priv, 0x001000, 0x00000002);
-	nv_icmd(priv, 0x0006aa, 0x00000001);
-	nv_icmd(priv, 0x0006ad, 0x00000100);
-	nv_icmd(priv, 0x0006ae, 0x00000100);
-	nv_icmd(priv, 0x0006b1, 0x00000011);
-	nv_icmd(priv, 0x00078c, 0x00000008);
-	nv_icmd(priv, 0x000792, 0x00000001);
-	nv_icmd(priv, 0x000794, 0x00000001);
-	nv_icmd(priv, 0x000795, 0x00000001);
-	nv_icmd(priv, 0x000796, 0x00000001);
-	nv_icmd(priv, 0x000797, 0x000000cf);
-	nv_icmd(priv, 0x00079a, 0x00000002);
-	nv_icmd(priv, 0x000833, 0x04444480);
-	nv_icmd(priv, 0x0007a1, 0x00000001);
-	nv_icmd(priv, 0x0007a3, 0x00000001);
-	nv_icmd(priv, 0x0007a4, 0x00000001);
-	nv_icmd(priv, 0x0007a5, 0x00000001);
-	nv_icmd(priv, 0x000831, 0x00000004);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_icmd(priv, 0x001000, 0x00000008);
-	nv_icmd(priv, 0x000039, 0x00000000);
-	nv_icmd(priv, 0x00003a, 0x00000000);
-	nv_icmd(priv, 0x00003b, 0x00000000);
-	nv_icmd(priv, 0x000380, 0x00000001);
-	nv_icmd(priv, 0x000366, 0x00000000);
-	nv_icmd(priv, 0x000367, 0x00000000);
-	nv_icmd(priv, 0x000368, 0x00000fff);
-	nv_icmd(priv, 0x000370, 0x00000000);
-	nv_icmd(priv, 0x000371, 0x00000000);
-	nv_icmd(priv, 0x000372, 0x000fffff);
-	nv_icmd(priv, 0x000813, 0x00000006);
-	nv_icmd(priv, 0x000814, 0x00000008);
-	nv_icmd(priv, 0x000957, 0x00000003);
-	nv_icmd(priv, 0x000818, 0x00000000);
-	nv_icmd(priv, 0x000819, 0x00000000);
-	nv_icmd(priv, 0x00081a, 0x00000000);
-	nv_icmd(priv, 0x00081b, 0x00000000);
-	nv_icmd(priv, 0x00081c, 0x00000000);
-	nv_icmd(priv, 0x00081d, 0x00000000);
-	nv_icmd(priv, 0x00081e, 0x00000000);
-	nv_icmd(priv, 0x00081f, 0x00000000);
-	nv_icmd(priv, 0x000848, 0x00000000);
-	nv_icmd(priv, 0x000849, 0x00000000);
-	nv_icmd(priv, 0x00084a, 0x00000000);
-	nv_icmd(priv, 0x00084b, 0x00000000);
-	nv_icmd(priv, 0x00084c, 0x00000000);
-	nv_icmd(priv, 0x00084d, 0x00000000);
-	nv_icmd(priv, 0x00084e, 0x00000000);
-	nv_icmd(priv, 0x00084f, 0x00000000);
-	nv_icmd(priv, 0x000850, 0x00000000);
-	nv_icmd(priv, 0x000851, 0x00000000);
-	nv_icmd(priv, 0x000852, 0x00000000);
-	nv_icmd(priv, 0x000853, 0x00000000);
-	nv_icmd(priv, 0x000854, 0x00000000);
-	nv_icmd(priv, 0x000855, 0x00000000);
-	nv_icmd(priv, 0x000856, 0x00000000);
-	nv_icmd(priv, 0x000857, 0x00000000);
-	nv_icmd(priv, 0x000738, 0x00000000);
-	nv_icmd(priv, 0x000b07, 0x00000002);
-	nv_icmd(priv, 0x000b08, 0x00000100);
-	nv_icmd(priv, 0x000b09, 0x00000100);
-	nv_icmd(priv, 0x000b0a, 0x00000001);
-	nv_icmd(priv, 0x000a04, 0x000000ff);
-	nv_icmd(priv, 0x00097f, 0x00000100);
-	nv_icmd(priv, 0x000a02, 0x00000001);
-	nv_icmd(priv, 0x000809, 0x00000007);
-	nv_icmd(priv, 0x00c221, 0x00000040);
-	nv_icmd(priv, 0x00c401, 0x00000001);
-	nv_icmd(priv, 0x00c402, 0x00010001);
-	nv_icmd(priv, 0x00c403, 0x00000001);
-	nv_icmd(priv, 0x00c404, 0x00000001);
-	nv_icmd(priv, 0x00c40e, 0x00000020);
-	nv_icmd(priv, 0x00c500, 0x00000003);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_icmd(priv, 0x001000, 0x00000001);
-	nv_icmd(priv, 0x000b07, 0x00000002);
-	nv_icmd(priv, 0x000b08, 0x00000100);
-	nv_icmd(priv, 0x000b09, 0x00000100);
-	nv_icmd(priv, 0x000b0a, 0x00000001);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_wr32(priv, 0x400208, 0x00000000);
-}
-
-static void
-nve0_grctx_generate_a097(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0xa097, 0x0800, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0840, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0880, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0900, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0940, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0980, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0804, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0844, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0884, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0904, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0944, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0984, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0808, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0848, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0888, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x08c8, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0908, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0948, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0988, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x09c8, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x080c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x084c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x088c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x08cc, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x090c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x094c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x098c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x09cc, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x0810, 0x000000cf);
-	nv_mthd(priv, 0xa097, 0x0850, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0890, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0910, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0950, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0990, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0814, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0854, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0894, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x08d4, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0914, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0954, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0994, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x09d4, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0818, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0858, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0898, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x08d8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0918, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0958, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0998, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x09d8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x081c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x085c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x089c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x091c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x095c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x099c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0820, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0860, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0920, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0960, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ca0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ce0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cf0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ca4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ce4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cf4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ca8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cb8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ce8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cf8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c3c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cbc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ccc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cdc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cfc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1da0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1db0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1de0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1df0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1da4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1db4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1de4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1df4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1da8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1db8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1de8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1df8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d3c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dbc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ddc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dfc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f3c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fa0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fa8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fb8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fe0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fe8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ff0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ff8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fa4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fbc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fdc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fe4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ff4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ffc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2000, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2040, 0x00000011);
-	nv_mthd(priv, 0xa097, 0x2080, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x20c0, 0x00000030);
-	nv_mthd(priv, 0xa097, 0x2100, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x2140, 0x00000051);
-	nv_mthd(priv, 0xa097, 0x200c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x204c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x208c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x20cc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x210c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x214c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x2010, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2050, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2090, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x20d0, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x2110, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x2150, 0x00000004);
-	nv_mthd(priv, 0xa097, 0x0380, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0384, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0388, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x038c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0700, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0710, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0720, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0730, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0704, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0714, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0724, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0734, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0708, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0718, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0728, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0738, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2800, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2804, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2808, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x280c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2810, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2814, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2818, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x281c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2820, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2824, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2828, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x282c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2830, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2834, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2838, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x283c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2840, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2844, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2848, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x284c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2850, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2854, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2858, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x285c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2860, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2864, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2868, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x286c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2870, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2874, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2878, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x287c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2880, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2884, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2888, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x288c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2890, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2894, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2898, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x289c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2900, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2904, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2908, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x290c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2910, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2914, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2918, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x291c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2920, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2924, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2928, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x292c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2930, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2934, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2938, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x293c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2940, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2944, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2948, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x294c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2950, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2954, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2958, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x295c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2960, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2964, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2968, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x296c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2970, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2974, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2978, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x297c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2980, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2984, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2988, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x298c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2990, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2994, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2998, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x299c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aa0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ac0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ae0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ba0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0be0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aa4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ac4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ae4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ba4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0be4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aa8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ac8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ae8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ba8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0be8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0acc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ab0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ad0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0af0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bf0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ab4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ad4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0af4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bf4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ca0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ce0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cf0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ca4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ce4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cf4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ca8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cb8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ce8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cf8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c0c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c1c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c2c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c3c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c4c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c5c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c6c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c7c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c8c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c9c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cac, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cbc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0ccc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cdc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cec, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cfc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0d00, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d08, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d10, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d18, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d20, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d28, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d30, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d38, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d04, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d0c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d14, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d1c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d24, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d2c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d34, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d3c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ea0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0eb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ec0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ed0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ee0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ef0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e04, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e14, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e24, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e34, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e44, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e54, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e64, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e74, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e84, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e94, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ea4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0eb4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ec4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ed4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ee4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ef4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e08, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e18, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e28, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e38, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e48, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e58, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e68, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e78, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e88, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e98, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ea8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0eb8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ec8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ed8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ee8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ef8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1e00, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e20, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e40, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e60, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e80, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ea0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ec0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ee0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e04, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e24, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e44, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e64, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e84, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ea4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ec4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ee4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e08, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e28, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e48, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e68, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e88, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ea8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ec8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ee8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e0c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e2c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e4c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e6c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e8c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eac, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ecc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eec, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e10, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e30, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e50, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e70, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e90, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eb0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ed0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ef0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e14, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e34, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e54, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e74, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e94, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1eb4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ed4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ef4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e18, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e38, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e58, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e78, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e98, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eb8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ed8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ef8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x3400, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3404, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3408, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x340c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3410, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3414, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3418, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x341c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3420, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3424, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3428, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x342c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3430, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3434, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3438, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x343c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3440, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3444, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3448, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x344c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3450, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3454, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3458, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x345c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3460, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3464, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3468, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x346c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3470, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3474, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3478, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x347c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3480, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3484, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3488, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x348c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3490, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3494, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3498, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x349c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3500, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3504, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3508, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x350c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3510, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3514, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3518, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x351c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3520, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3524, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3528, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x352c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3530, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3534, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3538, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x353c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3540, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3544, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3548, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x354c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3550, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3554, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3558, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x355c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3560, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3564, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3568, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x356c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3570, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3574, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3578, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x357c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3580, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3584, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3588, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x358c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3590, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3594, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3598, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x359c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x030c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1944, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1514, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d68, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x121c, 0x0fac6881);
-	nv_mthd(priv, 0xa097, 0x0fac, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1538, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0fe0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fe4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fe8, 0x00000014);
-	nv_mthd(priv, 0xa097, 0x0fec, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0ff0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x179c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1228, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x122c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x1230, 0x00010001);
-	nv_mthd(priv, 0xa097, 0x07f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15b4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1534, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x153c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x16b4, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x0fbc, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0fc0, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0fc4, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0fc8, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0df8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dfc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1948, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1970, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x161c, 0x000009f0);
-	nv_mthd(priv, 0xa097, 0x0dcc, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x163c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1160, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1164, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1168, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x116c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1170, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1174, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1178, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x117c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1180, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1184, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1188, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x118c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1190, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1194, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1198, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x119c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11a0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11a4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11a8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11ac, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11b0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11b4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11b8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11bc, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11c0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11c4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11c8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11cc, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11d0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11d4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11d8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11dc, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1880, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1884, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1888, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x188c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1890, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1894, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1898, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x189c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17d0, 0x000000ff);
-	nv_mthd(priv, 0xa097, 0x17d4, 0xffffffff);
-	nv_mthd(priv, 0xa097, 0x17d8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x17dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1434, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1438, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dec, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x13a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1318, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1644, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0748, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0de8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1648, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1120, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1124, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1128, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x112c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1118, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x164c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1658, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1910, 0x00000290);
-	nv_mthd(priv, 0xa097, 0x1518, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x165c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1520, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1604, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1570, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x13b0, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x13b4, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x020c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1670, 0x30201000);
-	nv_mthd(priv, 0xa097, 0x1674, 0x70605040);
-	nv_mthd(priv, 0xa097, 0x1678, 0xb8a89888);
-	nv_mthd(priv, 0xa097, 0x167c, 0xf8e8d8c8);
-	nv_mthd(priv, 0xa097, 0x166c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1680, 0x00ffff00);
-	nv_mthd(priv, 0xa097, 0x12d0, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x12d4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1684, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1688, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dac, 0x00001b02);
-	nv_mthd(priv, 0xa097, 0x0db0, 0x00001b02);
-	nv_mthd(priv, 0xa097, 0x0db4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x168c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x156c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x187c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1110, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0dc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1234, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1690, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12ac, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0790, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0794, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0798, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x079c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x077c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1000, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x10fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1290, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0218, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x12d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12dc, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x0d94, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x155c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1560, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1564, 0x00000fff);
-	nv_mthd(priv, 0xa097, 0x1574, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1578, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x157c, 0x000fffff);
-	nv_mthd(priv, 0xa097, 0x1354, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1610, 0x00000012);
-	nv_mthd(priv, 0xa097, 0x1608, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x160c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x260c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x162c, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x0210, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0320, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0324, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0328, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x032c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0330, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0334, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0338, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0750, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0760, 0x39291909);
-	nv_mthd(priv, 0xa097, 0x0764, 0x79695949);
-	nv_mthd(priv, 0xa097, 0x0768, 0xb9a99989);
-	nv_mthd(priv, 0xa097, 0x076c, 0xf9e9d9c9);
-	nv_mthd(priv, 0xa097, 0x0770, 0x30201000);
-	nv_mthd(priv, 0xa097, 0x0774, 0x70605040);
-	nv_mthd(priv, 0xa097, 0x0778, 0x00009080);
-	nv_mthd(priv, 0xa097, 0x0780, 0x39291909);
-	nv_mthd(priv, 0xa097, 0x0784, 0x79695949);
-	nv_mthd(priv, 0xa097, 0x0788, 0xb9a99989);
-	nv_mthd(priv, 0xa097, 0x078c, 0xf9e9d9c9);
-	nv_mthd(priv, 0xa097, 0x07d0, 0x30201000);
-	nv_mthd(priv, 0xa097, 0x07d4, 0x70605040);
-	nv_mthd(priv, 0xa097, 0x07d8, 0x00009080);
-	nv_mthd(priv, 0xa097, 0x037c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0740, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0744, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2600, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1918, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x191c, 0x00000900);
-	nv_mthd(priv, 0xa097, 0x1920, 0x00000405);
-	nv_mthd(priv, 0xa097, 0x1308, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1924, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x13ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x192c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x193c, 0x00002c1c);
-	nv_mthd(priv, 0xa097, 0x0d7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x02c0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1510, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1940, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ff4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ff8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x194c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1950, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1968, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1590, 0x0000003f);
-	nv_mthd(priv, 0xa097, 0x07e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x196c, 0x00000011);
-	nv_mthd(priv, 0xa097, 0x02e4, 0x0000b001);
-	nv_mthd(priv, 0xa097, 0x036c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0370, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x197c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x02d8, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x1980, 0x00000080);
-	nv_mthd(priv, 0xa097, 0x1504, 0x00000080);
-	nv_mthd(priv, 0xa097, 0x1984, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0300, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x13a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1310, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1314, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1380, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1384, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1388, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x138c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1390, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1394, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x139c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1398, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1594, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1598, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x159c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15a0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15a4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0f54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fa0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x130c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1360, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1364, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1368, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x136c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1370, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1374, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1378, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x137c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x133c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1340, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1344, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1348, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x134c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1350, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1358, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x12e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x131c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1320, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1324, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1328, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1140, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19c8, 0x00001500);
-	nv_mthd(priv, 0xa097, 0x135c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19e0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19e4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19e8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19ec, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19f0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19f4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19f8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19fc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19cc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a00, 0x00001111);
-	nv_mthd(priv, 0xa097, 0x1a04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d6c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d70, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x10f8, 0x00001010);
-	nv_mthd(priv, 0xa097, 0x0d80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0da0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1508, 0x80000000);
-	nv_mthd(priv, 0xa097, 0x150c, 0x40000000);
-	nv_mthd(priv, 0xa097, 0x1668, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0318, 0x00000008);
-	nv_mthd(priv, 0xa097, 0x031c, 0x00000008);
-	nv_mthd(priv, 0xa097, 0x0d9c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0374, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0378, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x07dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x074c, 0x00000055);
-	nv_mthd(priv, 0xa097, 0x1420, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x17bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17c4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1008, 0x00000008);
-	nv_mthd(priv, 0xa097, 0x100c, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x1010, 0x0000012c);
-	nv_mthd(priv, 0xa097, 0x0d60, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x075c, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x1018, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x101c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1020, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x1024, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1444, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1448, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x144c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0360, 0x20164010);
-	nv_mthd(priv, 0xa097, 0x0364, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x0368, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0de4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0204, 0x00000006);
-	nv_mthd(priv, 0xa097, 0x0208, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x02cc, 0x003fffff);
-	nv_mthd(priv, 0xa097, 0x02d0, 0x003fffff);
-	nv_mthd(priv, 0xa097, 0x1220, 0x00000005);
-	nv_mthd(priv, 0xa097, 0x0fdc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f98, 0x00400008);
-	nv_mthd(priv, 0xa097, 0x1284, 0x08000080);
-	nv_mthd(priv, 0xa097, 0x1450, 0x00400008);
-	nv_mthd(priv, 0xa097, 0x1454, 0x08000080);
-	nv_mthd(priv, 0xa097, 0x0214, 0x00000000);
-}
-
-static void
-nve0_grctx_generate_902d(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0x902d, 0x0200, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0204, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0208, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x020c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0210, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0214, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0218, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x021c, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x0220, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0224, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0230, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0234, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0238, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x023c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0244, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0248, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x024c, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x3410, 0x00000000);
-}
-
-static void
-nve0_graph_generate_unk40xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404010, 0x0);
-	nv_wr32(priv, 0x404014, 0x0);
-	nv_wr32(priv, 0x404018, 0x0);
-	nv_wr32(priv, 0x40401c, 0x0);
-	nv_wr32(priv, 0x404020, 0x0);
-	nv_wr32(priv, 0x404024, 0xe000);
-	nv_wr32(priv, 0x404028, 0x0);
-	nv_wr32(priv, 0x4040a8, 0x0);
-	nv_wr32(priv, 0x4040ac, 0x0);
-	nv_wr32(priv, 0x4040b0, 0x0);
-	nv_wr32(priv, 0x4040b4, 0x0);
-	nv_wr32(priv, 0x4040b8, 0x0);
-	nv_wr32(priv, 0x4040bc, 0x0);
-	nv_wr32(priv, 0x4040c0, 0x0);
-	nv_wr32(priv, 0x4040c4, 0x0);
-	nv_wr32(priv, 0x4040c8, 0xf800008f);
-	nv_wr32(priv, 0x4040d0, 0x0);
-	nv_wr32(priv, 0x4040d4, 0x0);
-	nv_wr32(priv, 0x4040d8, 0x0);
-	nv_wr32(priv, 0x4040dc, 0x0);
-	nv_wr32(priv, 0x4040e0, 0x0);
-	nv_wr32(priv, 0x4040e4, 0x0);
-	nv_wr32(priv, 0x4040e8, 0x1000);
-	nv_wr32(priv, 0x4040f8, 0x0);
-	nv_wr32(priv, 0x404130, 0x0);
-	nv_wr32(priv, 0x404134, 0x0);
-	nv_wr32(priv, 0x404138, 0x20000040);
-	nv_wr32(priv, 0x404150, 0x2e);
-	nv_wr32(priv, 0x404154, 0x400);
-	nv_wr32(priv, 0x404158, 0x200);
-	nv_wr32(priv, 0x404164, 0x55);
-	nv_wr32(priv, 0x4041a0, 0x0);
-	nv_wr32(priv, 0x4041a4, 0x0);
-	nv_wr32(priv, 0x4041a8, 0x0);
-	nv_wr32(priv, 0x4041ac, 0x0);
-	nv_wr32(priv, 0x404200, 0x0);
-	nv_wr32(priv, 0x404204, 0x0);
-	nv_wr32(priv, 0x404208, 0x0);
-	nv_wr32(priv, 0x40420c, 0x0);
-}
-
-static void
-nve0_graph_generate_unk44xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404404, 0x0);
-	nv_wr32(priv, 0x404408, 0x0);
-	nv_wr32(priv, 0x40440c, 0x0);
-	nv_wr32(priv, 0x404410, 0x0);
-	nv_wr32(priv, 0x404414, 0x0);
-	nv_wr32(priv, 0x404418, 0x0);
-	nv_wr32(priv, 0x40441c, 0x0);
-	nv_wr32(priv, 0x404420, 0x0);
-	nv_wr32(priv, 0x404424, 0x0);
-	nv_wr32(priv, 0x404428, 0x0);
-	nv_wr32(priv, 0x40442c, 0x0);
-	nv_wr32(priv, 0x404430, 0x0);
-	nv_wr32(priv, 0x404434, 0x0);
-	nv_wr32(priv, 0x404438, 0x0);
-	nv_wr32(priv, 0x404460, 0x0);
-	nv_wr32(priv, 0x404464, 0x0);
-	nv_wr32(priv, 0x404468, 0xffffff);
-	nv_wr32(priv, 0x40446c, 0x0);
-	nv_wr32(priv, 0x404480, 0x1);
-	nv_wr32(priv, 0x404498, 0x1);
-}
-
-static void
-nve0_graph_generate_unk46xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404604, 0x14);
-	nv_wr32(priv, 0x404608, 0x0);
-	nv_wr32(priv, 0x40460c, 0x3fff);
-	nv_wr32(priv, 0x404610, 0x100);
-	nv_wr32(priv, 0x404618, 0x0);
-	nv_wr32(priv, 0x40461c, 0x0);
-	nv_wr32(priv, 0x404620, 0x0);
-	nv_wr32(priv, 0x404624, 0x0);
-	nv_wr32(priv, 0x40462c, 0x0);
-	nv_wr32(priv, 0x404630, 0x0);
-	nv_wr32(priv, 0x404640, 0x0);
-	nv_wr32(priv, 0x404654, 0x0);
-	nv_wr32(priv, 0x404660, 0x0);
-	nv_wr32(priv, 0x404678, 0x0);
-	nv_wr32(priv, 0x40467c, 0x2);
-	nv_wr32(priv, 0x404680, 0x0);
-	nv_wr32(priv, 0x404684, 0x0);
-	nv_wr32(priv, 0x404688, 0x0);
-	nv_wr32(priv, 0x40468c, 0x0);
-	nv_wr32(priv, 0x404690, 0x0);
-	nv_wr32(priv, 0x404694, 0x0);
-	nv_wr32(priv, 0x404698, 0x0);
-	nv_wr32(priv, 0x40469c, 0x0);
-	nv_wr32(priv, 0x4046a0, 0x7f0080);
-	nv_wr32(priv, 0x4046a4, 0x0);
-	nv_wr32(priv, 0x4046a8, 0x0);
-	nv_wr32(priv, 0x4046ac, 0x0);
-	nv_wr32(priv, 0x4046b0, 0x0);
-	nv_wr32(priv, 0x4046b4, 0x0);
-	nv_wr32(priv, 0x4046b8, 0x0);
-	nv_wr32(priv, 0x4046bc, 0x0);
-	nv_wr32(priv, 0x4046c0, 0x0);
-	nv_wr32(priv, 0x4046c8, 0x0);
-	nv_wr32(priv, 0x4046cc, 0x0);
-	nv_wr32(priv, 0x4046d0, 0x0);
-}
-
-static void
-nve0_graph_generate_unk47xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404700, 0x0);
-	nv_wr32(priv, 0x404704, 0x0);
-	nv_wr32(priv, 0x404708, 0x0);
-	nv_wr32(priv, 0x404718, 0x0);
-	nv_wr32(priv, 0x40471c, 0x0);
-	nv_wr32(priv, 0x404720, 0x0);
-	nv_wr32(priv, 0x404724, 0x0);
-	nv_wr32(priv, 0x404728, 0x0);
-	nv_wr32(priv, 0x40472c, 0x0);
-	nv_wr32(priv, 0x404730, 0x0);
-	nv_wr32(priv, 0x404734, 0x100);
-	nv_wr32(priv, 0x404738, 0x0);
-	nv_wr32(priv, 0x40473c, 0x0);
-	nv_wr32(priv, 0x404744, 0x0);
-	nv_wr32(priv, 0x404748, 0x0);
-	nv_wr32(priv, 0x404754, 0x0);
-}
-
-static void
-nve0_graph_generate_unk58xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x405800, 0xf8000bf);
-	nv_wr32(priv, 0x405830, 0x2180648);
-	nv_wr32(priv, 0x405834, 0x8000000);
-	nv_wr32(priv, 0x405838, 0x0);
-	nv_wr32(priv, 0x405854, 0x0);
-	nv_wr32(priv, 0x405870, 0x1);
-	nv_wr32(priv, 0x405874, 0x1);
-	nv_wr32(priv, 0x405878, 0x1);
-	nv_wr32(priv, 0x40587c, 0x1);
-	nv_wr32(priv, 0x405a00, 0x0);
-	nv_wr32(priv, 0x405a04, 0x0);
-	nv_wr32(priv, 0x405a18, 0x0);
-	nv_wr32(priv, 0x405b00, 0x0);
-	nv_wr32(priv, 0x405b10, 0x1000);
-}
-
-static void
-nve0_graph_generate_unk60xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x406020, 0x4103c1);
-	nv_wr32(priv, 0x406028, 0x1);
-	nv_wr32(priv, 0x40602c, 0x1);
-	nv_wr32(priv, 0x406030, 0x1);
-	nv_wr32(priv, 0x406034, 0x1);
-}
-
-static void
-nve0_graph_generate_unk64xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x4064a8, 0x0);
-	nv_wr32(priv, 0x4064ac, 0x3fff);
-	nv_wr32(priv, 0x4064b4, 0x0);
-	nv_wr32(priv, 0x4064b8, 0x0);
-	nv_wr32(priv, 0x4064c0, 0x801a00f0);
-	nv_wr32(priv, 0x4064c4, 0x192ffff);
-	nv_wr32(priv, 0x4064c8, 0x1800600);
-	nv_wr32(priv, 0x4064cc, 0x0);
-	nv_wr32(priv, 0x4064d0, 0x0);
-	nv_wr32(priv, 0x4064d4, 0x0);
-	nv_wr32(priv, 0x4064d8, 0x0);
-	nv_wr32(priv, 0x4064dc, 0x0);
-	nv_wr32(priv, 0x4064e0, 0x0);
-	nv_wr32(priv, 0x4064e4, 0x0);
-	nv_wr32(priv, 0x4064e8, 0x0);
-	nv_wr32(priv, 0x4064ec, 0x0);
-	nv_wr32(priv, 0x4064fc, 0x22a);
-}
-
-static void
-nve0_graph_generate_unk70xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x407040, 0x0);
-}
-
-static void
-nve0_graph_generate_unk78xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x407804, 0x23);
-	nv_wr32(priv, 0x40780c, 0xa418820);
-	nv_wr32(priv, 0x407810, 0x62080e6);
-	nv_wr32(priv, 0x407814, 0x20398a4);
-	nv_wr32(priv, 0x407818, 0xe629062);
-	nv_wr32(priv, 0x40781c, 0xa418820);
-	nv_wr32(priv, 0x407820, 0xe6);
-	nv_wr32(priv, 0x4078bc, 0x103);
-}
-
-static void
-nve0_graph_generate_unk80xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x408000, 0x0);
-	nv_wr32(priv, 0x408004, 0x0);
-	nv_wr32(priv, 0x408008, 0x30);
-	nv_wr32(priv, 0x40800c, 0x0);
-	nv_wr32(priv, 0x408010, 0x0);
-	nv_wr32(priv, 0x408014, 0x69);
-	nv_wr32(priv, 0x408018, 0xe100e100);
-	nv_wr32(priv, 0x408064, 0x0);
-}
-
-static void
-nve0_graph_generate_unk88xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x408800, 0x2802a3c);
-	nv_wr32(priv, 0x408804, 0x40);
-	nv_wr32(priv, 0x408808, 0x1043e005);
-	nv_wr32(priv, 0x408840, 0xb);
-	nv_wr32(priv, 0x408900, 0x3080b801);
-	nv_wr32(priv, 0x408904, 0x62000001);
-	nv_wr32(priv, 0x408908, 0xc8102f);
-	nv_wr32(priv, 0x408980, 0x11d);
-}
-
-static void
-nve0_graph_generate_gpc(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x418380, 0x16);
-	nv_wr32(priv, 0x418400, 0x38004e00);
-	nv_wr32(priv, 0x418404, 0x71e0ffff);
-	nv_wr32(priv, 0x41840c, 0x1008);
-	nv_wr32(priv, 0x418410, 0xfff0fff);
-	nv_wr32(priv, 0x418414, 0x2200fff);
-	nv_wr32(priv, 0x418450, 0x0);
-	nv_wr32(priv, 0x418454, 0x0);
-	nv_wr32(priv, 0x418458, 0x0);
-	nv_wr32(priv, 0x41845c, 0x0);
-	nv_wr32(priv, 0x418460, 0x0);
-	nv_wr32(priv, 0x418464, 0x0);
-	nv_wr32(priv, 0x418468, 0x1);
-	nv_wr32(priv, 0x41846c, 0x0);
-	nv_wr32(priv, 0x418470, 0x0);
-	nv_wr32(priv, 0x418600, 0x1f);
-	nv_wr32(priv, 0x418684, 0xf);
-	nv_wr32(priv, 0x418700, 0x2);
-	nv_wr32(priv, 0x418704, 0x80);
-	nv_wr32(priv, 0x418708, 0x0);
-	nv_wr32(priv, 0x41870c, 0x0);
-	nv_wr32(priv, 0x418710, 0x0);
-	nv_wr32(priv, 0x418800, 0x7006860a);
-	nv_wr32(priv, 0x418808, 0x0);
-	nv_wr32(priv, 0x41880c, 0x0);
-	nv_wr32(priv, 0x418810, 0x0);
-	nv_wr32(priv, 0x418828, 0x44);
-	nv_wr32(priv, 0x418830, 0x10000001);
-	nv_wr32(priv, 0x4188d8, 0x8);
-	nv_wr32(priv, 0x4188e0, 0x1000000);
-	nv_wr32(priv, 0x4188e8, 0x0);
-	nv_wr32(priv, 0x4188ec, 0x0);
-	nv_wr32(priv, 0x4188f0, 0x0);
-	nv_wr32(priv, 0x4188f4, 0x0);
-	nv_wr32(priv, 0x4188f8, 0x0);
-	nv_wr32(priv, 0x4188fc, 0x20100018);
-	nv_wr32(priv, 0x41891c, 0xff00ff);
-	nv_wr32(priv, 0x418924, 0x0);
-	nv_wr32(priv, 0x418928, 0xffff00);
-	nv_wr32(priv, 0x41892c, 0xff00);
-	nv_wr32(priv, 0x418a00, 0x0);
-	nv_wr32(priv, 0x418a04, 0x0);
-	nv_wr32(priv, 0x418a08, 0x0);
-	nv_wr32(priv, 0x418a0c, 0x10000);
-	nv_wr32(priv, 0x418a10, 0x0);
-	nv_wr32(priv, 0x418a14, 0x0);
-	nv_wr32(priv, 0x418a18, 0x0);
-	nv_wr32(priv, 0x418a20, 0x0);
-	nv_wr32(priv, 0x418a24, 0x0);
-	nv_wr32(priv, 0x418a28, 0x0);
-	nv_wr32(priv, 0x418a2c, 0x10000);
-	nv_wr32(priv, 0x418a30, 0x0);
-	nv_wr32(priv, 0x418a34, 0x0);
-	nv_wr32(priv, 0x418a38, 0x0);
-	nv_wr32(priv, 0x418a40, 0x0);
-	nv_wr32(priv, 0x418a44, 0x0);
-	nv_wr32(priv, 0x418a48, 0x0);
-	nv_wr32(priv, 0x418a4c, 0x10000);
-	nv_wr32(priv, 0x418a50, 0x0);
-	nv_wr32(priv, 0x418a54, 0x0);
-	nv_wr32(priv, 0x418a58, 0x0);
-	nv_wr32(priv, 0x418a60, 0x0);
-	nv_wr32(priv, 0x418a64, 0x0);
-	nv_wr32(priv, 0x418a68, 0x0);
-	nv_wr32(priv, 0x418a6c, 0x10000);
-	nv_wr32(priv, 0x418a70, 0x0);
-	nv_wr32(priv, 0x418a74, 0x0);
-	nv_wr32(priv, 0x418a78, 0x0);
-	nv_wr32(priv, 0x418a80, 0x0);
-	nv_wr32(priv, 0x418a84, 0x0);
-	nv_wr32(priv, 0x418a88, 0x0);
-	nv_wr32(priv, 0x418a8c, 0x10000);
-	nv_wr32(priv, 0x418a90, 0x0);
-	nv_wr32(priv, 0x418a94, 0x0);
-	nv_wr32(priv, 0x418a98, 0x0);
-	nv_wr32(priv, 0x418aa0, 0x0);
-	nv_wr32(priv, 0x418aa4, 0x0);
-	nv_wr32(priv, 0x418aa8, 0x0);
-	nv_wr32(priv, 0x418aac, 0x10000);
-	nv_wr32(priv, 0x418ab0, 0x0);
-	nv_wr32(priv, 0x418ab4, 0x0);
-	nv_wr32(priv, 0x418ab8, 0x0);
-	nv_wr32(priv, 0x418ac0, 0x0);
-	nv_wr32(priv, 0x418ac4, 0x0);
-	nv_wr32(priv, 0x418ac8, 0x0);
-	nv_wr32(priv, 0x418acc, 0x10000);
-	nv_wr32(priv, 0x418ad0, 0x0);
-	nv_wr32(priv, 0x418ad4, 0x0);
-	nv_wr32(priv, 0x418ad8, 0x0);
-	nv_wr32(priv, 0x418ae0, 0x0);
-	nv_wr32(priv, 0x418ae4, 0x0);
-	nv_wr32(priv, 0x418ae8, 0x0);
-	nv_wr32(priv, 0x418aec, 0x10000);
-	nv_wr32(priv, 0x418af0, 0x0);
-	nv_wr32(priv, 0x418af4, 0x0);
-	nv_wr32(priv, 0x418af8, 0x0);
-	nv_wr32(priv, 0x418b00, 0x6);
-	nv_wr32(priv, 0x418b08, 0xa418820);
-	nv_wr32(priv, 0x418b0c, 0x62080e6);
-	nv_wr32(priv, 0x418b10, 0x20398a4);
-	nv_wr32(priv, 0x418b14, 0xe629062);
-	nv_wr32(priv, 0x418b18, 0xa418820);
-	nv_wr32(priv, 0x418b1c, 0xe6);
-	nv_wr32(priv, 0x418bb8, 0x103);
-	nv_wr32(priv, 0x418c08, 0x1);
-	nv_wr32(priv, 0x418c10, 0x0);
-	nv_wr32(priv, 0x418c14, 0x0);
-	nv_wr32(priv, 0x418c18, 0x0);
-	nv_wr32(priv, 0x418c1c, 0x0);
-	nv_wr32(priv, 0x418c20, 0x0);
-	nv_wr32(priv, 0x418c24, 0x0);
-	nv_wr32(priv, 0x418c28, 0x0);
-	nv_wr32(priv, 0x418c2c, 0x0);
-	nv_wr32(priv, 0x418c40, 0xffffffff);
-	nv_wr32(priv, 0x418c6c, 0x1);
-	nv_wr32(priv, 0x418c80, 0x20200004);
-	nv_wr32(priv, 0x418c8c, 0x1);
-	nv_wr32(priv, 0x419000, 0x780);
-	nv_wr32(priv, 0x419004, 0x0);
-	nv_wr32(priv, 0x419008, 0x0);
-	nv_wr32(priv, 0x419014, 0x4);
-}
-
-static void
-nve0_graph_generate_tpc(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x419848, 0x0);
-	nv_wr32(priv, 0x419864, 0x129);
-	nv_wr32(priv, 0x419888, 0x0);
-	nv_wr32(priv, 0x419a00, 0xf0);
-	nv_wr32(priv, 0x419a04, 0x1);
-	nv_wr32(priv, 0x419a08, 0x21);
-	nv_wr32(priv, 0x419a0c, 0x20000);
-	nv_wr32(priv, 0x419a10, 0x0);
-	nv_wr32(priv, 0x419a14, 0x200);
-	nv_wr32(priv, 0x419a1c, 0xc000);
-	nv_wr32(priv, 0x419a20, 0x800);
-	nv_wr32(priv, 0x419a30, 0x1);
-	nv_wr32(priv, 0x419ac4, 0x37f440);
-	nv_wr32(priv, 0x419c00, 0xa);
-	nv_wr32(priv, 0x419c04, 0x80000006);
-	nv_wr32(priv, 0x419c08, 0x2);
-	nv_wr32(priv, 0x419c20, 0x0);
-	nv_wr32(priv, 0x419c24, 0x84210);
-	nv_wr32(priv, 0x419c28, 0x3efbefbe);
-	nv_wr32(priv, 0x419ce8, 0x0);
-	nv_wr32(priv, 0x419cf4, 0x3203);
-	nv_wr32(priv, 0x419e04, 0x0);
-	nv_wr32(priv, 0x419e08, 0x0);
-	nv_wr32(priv, 0x419e0c, 0x0);
-	nv_wr32(priv, 0x419e10, 0x402);
-	nv_wr32(priv, 0x419e44, 0x13eff2);
-	nv_wr32(priv, 0x419e48, 0x0);
-	nv_wr32(priv, 0x419e4c, 0x7f);
-	nv_wr32(priv, 0x419e50, 0x0);
-	nv_wr32(priv, 0x419e54, 0x0);
-	nv_wr32(priv, 0x419e58, 0x0);
-	nv_wr32(priv, 0x419e5c, 0x0);
-	nv_wr32(priv, 0x419e60, 0x0);
-	nv_wr32(priv, 0x419e64, 0x0);
-	nv_wr32(priv, 0x419e68, 0x0);
-	nv_wr32(priv, 0x419e6c, 0x0);
-	nv_wr32(priv, 0x419e70, 0x0);
-	nv_wr32(priv, 0x419e74, 0x0);
-	nv_wr32(priv, 0x419e78, 0x0);
-	nv_wr32(priv, 0x419e7c, 0x0);
-	nv_wr32(priv, 0x419e80, 0x0);
-	nv_wr32(priv, 0x419e84, 0x0);
-	nv_wr32(priv, 0x419e88, 0x0);
-	nv_wr32(priv, 0x419e8c, 0x0);
-	nv_wr32(priv, 0x419e90, 0x0);
-	nv_wr32(priv, 0x419e94, 0x0);
-	nv_wr32(priv, 0x419e98, 0x0);
-	nv_wr32(priv, 0x419eac, 0x1fcf);
-	nv_wr32(priv, 0x419eb0, 0xd3f);
-	nv_wr32(priv, 0x419ec8, 0x1304f);
-	nv_wr32(priv, 0x419f30, 0x0);
-	nv_wr32(priv, 0x419f34, 0x0);
-	nv_wr32(priv, 0x419f38, 0x0);
-	nv_wr32(priv, 0x419f3c, 0x0);
-	nv_wr32(priv, 0x419f40, 0x0);
-	nv_wr32(priv, 0x419f44, 0x0);
-	nv_wr32(priv, 0x419f48, 0x0);
-	nv_wr32(priv, 0x419f4c, 0x0);
-	nv_wr32(priv, 0x419f58, 0x0);
-	nv_wr32(priv, 0x419f78, 0xb);
-}
-
-static void
-nve0_graph_generate_tpcunk(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x41be24, 0x6);
-	nv_wr32(priv, 0x41bec0, 0x12180000);
-	nv_wr32(priv, 0x41bec4, 0x37f7f);
-	nv_wr32(priv, 0x41bee4, 0x6480430);
-	nv_wr32(priv, 0x41bf00, 0xa418820);
-	nv_wr32(priv, 0x41bf04, 0x62080e6);
-	nv_wr32(priv, 0x41bf08, 0x20398a4);
-	nv_wr32(priv, 0x41bf0c, 0xe629062);
-	nv_wr32(priv, 0x41bf10, 0xa418820);
-	nv_wr32(priv, 0x41bf14, 0xe6);
-	nv_wr32(priv, 0x41bfd0, 0x900103);
-	nv_wr32(priv, 0x41bfe0, 0x400001);
-	nv_wr32(priv, 0x41bfe4, 0x0);
-}
-
-int
-nve0_grctx_generate(struct nvc0_graph_priv *priv)
-{
-	struct nvc0_grctx info;
-	int ret, i, gpc, tpc, id;
-	u32 data[6] = {}, data2[2] = {}, tmp;
-	u32 tpc_set = 0, tpc_mask = 0;
-	u32 magic[GPC_MAX][2], offset;
-	u8 tpcnr[GPC_MAX], a, b;
-	u8 shift, ntpcv;
-
-	ret = nvc0_grctx_init(priv, &info);
-	if (ret)
-		return ret;
-
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
-	nv_wr32(priv, 0x400204, 0x00000000);
-	nv_wr32(priv, 0x400208, 0x00000000);
-
-	nve0_graph_generate_unk40xx(priv);
-	nve0_graph_generate_unk44xx(priv);
-	nve0_graph_generate_unk46xx(priv);
-	nve0_graph_generate_unk47xx(priv);
-	nve0_graph_generate_unk58xx(priv);
-	nve0_graph_generate_unk60xx(priv);
-	nve0_graph_generate_unk64xx(priv);
-	nve0_graph_generate_unk70xx(priv);
-	nve0_graph_generate_unk78xx(priv);
-	nve0_graph_generate_unk80xx(priv);
-	nve0_graph_generate_unk88xx(priv);
-	nve0_graph_generate_gpc(priv);
-	nve0_graph_generate_tpc(priv);
-	nve0_graph_generate_tpcunk(priv);
-
-	nv_wr32(priv, 0x404154, 0x0);
-
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x4064cc, 0x80000000,  0, 0);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000030,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000030,  0, 0);
-	mmio_list(0x4064c8, 0x01800600,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x405830, 0x02180648,  0, 0);
-	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
-		u16 magic1 = 0x0648 * priv->tpc_nr[gpc];
-		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
-		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
-		offset += 0x0324 * priv->tpc_nr[gpc];
-	}
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
-		offset += 0x07ff * priv->tpc_nr[gpc];
-	}
-	mmio_list(0x17e91c, 0x06060609, 0, 0);
-	mmio_list(0x17e920, 0x00090a05, 0, 0);
-
-	nv_wr32(priv, 0x418c6c, 0x1);
-	nv_wr32(priv, 0x41980c, 0x10);
-	nv_wr32(priv, 0x41be08, 0x4);
-	nv_wr32(priv, 0x4064c0, 0x801a00f0);
-	nv_wr32(priv, 0x405800, 0xf8000bf);
-	nv_wr32(priv, 0x419c00, 0xa);
-
-	for (tpc = 0, id = 0; tpc < 4; tpc++) {
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			if (tpc < priv->tpc_nr[gpc]) {
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0698), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x04e8), id);
-				nv_wr32(priv, GPC_UNIT(gpc, 0x0c10 + tpc * 4), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0088), id++);
-			}
-
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c08), priv->tpc_nr[gpc]);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c8c), priv->tpc_nr[gpc]);
-		}
-	}
-
-	tmp = 0;
-	for (i = 0; i < priv->gpc_nr; i++)
-		tmp |= priv->tpc_nr[i] << (i * 4);
-	nv_wr32(priv, 0x406028, tmp);
-	nv_wr32(priv, 0x405870, tmp);
-
-	nv_wr32(priv, 0x40602c, 0x0);
-	nv_wr32(priv, 0x405874, 0x0);
-	nv_wr32(priv, 0x406030, 0x0);
-	nv_wr32(priv, 0x405878, 0x0);
-	nv_wr32(priv, 0x406034, 0x0);
-	nv_wr32(priv, 0x40587c, 0x0);
-
-	/* calculate first set of magics */
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-
-	gpc = -1;
-	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
-		do {
-			gpc = (gpc + 1) % priv->gpc_nr;
-		} while (!tpcnr[gpc]);
-		tpcnr[gpc]--;
-
-		data[tpc / 6] |= gpc << ((tpc % 6) * 5);
-	}
-
-	for (; tpc < 32; tpc++)
-		data[tpc / 6] |= 7 << ((tpc % 6) * 5);
-
-	/* and the second... */
-	shift = 0;
-	ntpcv = priv->tpc_total;
-	while (!(ntpcv & (1 << 4))) {
-		ntpcv <<= 1;
-		shift++;
-	}
-
-	data2[0]  = ntpcv << 16;
-	data2[0] |= shift << 21;
-	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
-	data2[0] |= priv->tpc_total << 8;
-	data2[0] |= priv->magic_not_rop_nr;
-	for (i = 1; i < 7; i++)
-		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
-
-	/* and write it all the various parts of PGRAPH */
-	nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) | priv->magic_not_rop_nr);
-	for (i = 0; i < 6; i++)
-		nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
-
-	nv_wr32(priv, 0x41bfd0, data2[0]);
-	nv_wr32(priv, 0x41bfe4, data2[1]);
-	for (i = 0; i < 6; i++)
-		nv_wr32(priv, 0x41bf00 + (i * 4), data[i]);
-
-	nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) | priv->magic_not_rop_nr);
-	for (i = 0; i < 6; i++)
-		nv_wr32(priv, 0x40780c + (i * 4), data[i]);
-
-
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++)
-		tpc_mask |= ((1 << priv->tpc_nr[gpc]) - 1) << (gpc * 8);
-
-	for (i = 0, gpc = -1, b = -1; i < 32; i++) {
-		a = (i * (priv->tpc_total - 1)) / 32;
-		if (a != b) {
-			b = a;
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpcnr[gpc]);
-			tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-			tpc_set |= 1 << ((gpc * 8) + tpc);
-		}
-
-		nv_wr32(priv, 0x406800 + (i * 0x20), tpc_set);
-		nv_wr32(priv, 0x406c00 + (i * 0x20), tpc_set ^ tpc_mask);
-	}
-
-	for (i = 0; i < 8; i++)
-		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
-
-	nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr);
-	if (priv->gpc_nr == 1) {
-		nv_mask(priv, 0x408850, 0x0000000f, priv->tpc_nr[0]);
-		nv_mask(priv, 0x408958, 0x0000000f, priv->tpc_nr[0]);
-	} else {
-		nv_mask(priv, 0x408850, 0x0000000f, priv->gpc_nr);
-		nv_mask(priv, 0x408958, 0x0000000f, priv->gpc_nr);
-	}
-	nv_mask(priv, 0x419f78, 0x00000001, 0x00000000);
-
-	nve0_grctx_generate_icmd(priv);
-	nve0_grctx_generate_a097(priv);
-	nve0_grctx_generate_902d(priv);
-
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
-	nv_wr32(priv, 0x418800, 0x7026860a); //XXX
-	nv_wr32(priv, 0x41be10, 0x00bb8bc7); //XXX
-	return nvc0_grctx_fini(&info);
-}
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
new file mode 100644
index 0000000..e2de73e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
@@ -0,0 +1,1018 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+struct nvc0_graph_init
+nve4_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x000039,   3, 0x01, 0x00000000 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000383,   1, 0x01, 0x00000011 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00000fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x000fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x00097d,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x003fffff },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00400008 },
+	{ 0x000841,   1, 0x01, 0x08000080 },
+	{ 0x000842,   1, 0x01, 0x00400008 },
+	{ 0x000843,   1, 0x01, 0x08000080 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x000b07,   1, 0x01, 0x00000002 },
+	{ 0x000b08,   2, 0x01, 0x00000100 },
+	{ 0x000b0a,   1, 0x01, 0x00000001 },
+	{ 0x000a04,   1, 0x01, 0x000000ff },
+	{ 0x000a0b,   1, 0x01, 0x00000040 },
+	{ 0x00097f,   1, 0x01, 0x00000100 },
+	{ 0x000a02,   1, 0x01, 0x00000001 },
+	{ 0x000809,   1, 0x01, 0x00000007 },
+	{ 0x00c221,   1, 0x01, 0x00000040 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x00c401,   1, 0x01, 0x00000001 },
+	{ 0x00c402,   1, 0x01, 0x00010001 },
+	{ 0x00c403,   2, 0x01, 0x00000001 },
+	{ 0x00c40e,   1, 0x01, 0x00000020 },
+	{ 0x00c500,   1, 0x01, 0x00000003 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000008 },
+	{ 0x000039,   3, 0x01, 0x00000000 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00000fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x000fffff },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x000b07,   1, 0x01, 0x00000002 },
+	{ 0x000b08,   2, 0x01, 0x00000100 },
+	{ 0x000b0a,   1, 0x01, 0x00000001 },
+	{ 0x000a04,   1, 0x01, 0x000000ff },
+	{ 0x00097f,   1, 0x01, 0x00000100 },
+	{ 0x000a02,   1, 0x01, 0x00000001 },
+	{ 0x000809,   1, 0x01, 0x00000007 },
+	{ 0x00c221,   1, 0x01, 0x00000040 },
+	{ 0x00c401,   1, 0x01, 0x00000001 },
+	{ 0x00c402,   1, 0x01, 0x00010001 },
+	{ 0x00c403,   2, 0x01, 0x00000001 },
+	{ 0x00c40e,   1, 0x01, 0x00000020 },
+	{ 0x00c500,   1, 0x01, 0x00000003 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x000b07,   1, 0x01, 0x00000002 },
+	{ 0x000b08,   2, 0x01, 0x00000100 },
+	{ 0x000b0a,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_a097[] = {
+	{ 0x000800,   8, 0x40, 0x00000000 },
+	{ 0x000804,   8, 0x40, 0x00000000 },
+	{ 0x000808,   8, 0x40, 0x00000400 },
+	{ 0x00080c,   8, 0x40, 0x00000300 },
+	{ 0x000810,   1, 0x04, 0x000000cf },
+	{ 0x000850,   7, 0x40, 0x00000000 },
+	{ 0x000814,   8, 0x40, 0x00000040 },
+	{ 0x000818,   8, 0x40, 0x00000001 },
+	{ 0x00081c,   8, 0x40, 0x00000000 },
+	{ 0x000820,   8, 0x40, 0x00000000 },
+	{ 0x001c00,  16, 0x10, 0x00000000 },
+	{ 0x001c04,  16, 0x10, 0x00000000 },
+	{ 0x001c08,  16, 0x10, 0x00000000 },
+	{ 0x001c0c,  16, 0x10, 0x00000000 },
+	{ 0x001d00,  16, 0x10, 0x00000000 },
+	{ 0x001d04,  16, 0x10, 0x00000000 },
+	{ 0x001d08,  16, 0x10, 0x00000000 },
+	{ 0x001d0c,  16, 0x10, 0x00000000 },
+	{ 0x001f00,  16, 0x08, 0x00000000 },
+	{ 0x001f04,  16, 0x08, 0x00000000 },
+	{ 0x001f80,  16, 0x08, 0x00000000 },
+	{ 0x001f84,  16, 0x08, 0x00000000 },
+	{ 0x002000,   1, 0x04, 0x00000000 },
+	{ 0x002040,   1, 0x04, 0x00000011 },
+	{ 0x002080,   1, 0x04, 0x00000020 },
+	{ 0x0020c0,   1, 0x04, 0x00000030 },
+	{ 0x002100,   1, 0x04, 0x00000040 },
+	{ 0x002140,   1, 0x04, 0x00000051 },
+	{ 0x00200c,   6, 0x40, 0x00000001 },
+	{ 0x002010,   1, 0x04, 0x00000000 },
+	{ 0x002050,   1, 0x04, 0x00000000 },
+	{ 0x002090,   1, 0x04, 0x00000001 },
+	{ 0x0020d0,   1, 0x04, 0x00000002 },
+	{ 0x002110,   1, 0x04, 0x00000003 },
+	{ 0x002150,   1, 0x04, 0x00000004 },
+	{ 0x000380,   4, 0x20, 0x00000000 },
+	{ 0x000384,   4, 0x20, 0x00000000 },
+	{ 0x000388,   4, 0x20, 0x00000000 },
+	{ 0x00038c,   4, 0x20, 0x00000000 },
+	{ 0x000700,   4, 0x10, 0x00000000 },
+	{ 0x000704,   4, 0x10, 0x00000000 },
+	{ 0x000708,   4, 0x10, 0x00000000 },
+	{ 0x002800, 128, 0x04, 0x00000000 },
+	{ 0x000a00,  16, 0x20, 0x00000000 },
+	{ 0x000a04,  16, 0x20, 0x00000000 },
+	{ 0x000a08,  16, 0x20, 0x00000000 },
+	{ 0x000a0c,  16, 0x20, 0x00000000 },
+	{ 0x000a10,  16, 0x20, 0x00000000 },
+	{ 0x000a14,  16, 0x20, 0x00000000 },
+	{ 0x000c00,  16, 0x10, 0x00000000 },
+	{ 0x000c04,  16, 0x10, 0x00000000 },
+	{ 0x000c08,  16, 0x10, 0x00000000 },
+	{ 0x000c0c,  16, 0x10, 0x3f800000 },
+	{ 0x000d00,   8, 0x08, 0xffff0000 },
+	{ 0x000d04,   8, 0x08, 0xffff0000 },
+	{ 0x000e00,  16, 0x10, 0x00000000 },
+	{ 0x000e04,  16, 0x10, 0xffff0000 },
+	{ 0x000e08,  16, 0x10, 0xffff0000 },
+	{ 0x000d40,   4, 0x08, 0x00000000 },
+	{ 0x000d44,   4, 0x08, 0x00000000 },
+	{ 0x001e00,   8, 0x20, 0x00000001 },
+	{ 0x001e04,   8, 0x20, 0x00000001 },
+	{ 0x001e08,   8, 0x20, 0x00000002 },
+	{ 0x001e0c,   8, 0x20, 0x00000001 },
+	{ 0x001e10,   8, 0x20, 0x00000001 },
+	{ 0x001e14,   8, 0x20, 0x00000002 },
+	{ 0x001e18,   8, 0x20, 0x00000001 },
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x001514,   1, 0x04, 0x00000000 },
+	{ 0x000d68,   1, 0x04, 0x0000ffff },
+	{ 0x00121c,   1, 0x04, 0x0fac6881 },
+	{ 0x000fac,   1, 0x04, 0x00000001 },
+	{ 0x001538,   1, 0x04, 0x00000001 },
+	{ 0x000fe0,   2, 0x04, 0x00000000 },
+	{ 0x000fe8,   1, 0x04, 0x00000014 },
+	{ 0x000fec,   1, 0x04, 0x00000040 },
+	{ 0x000ff0,   1, 0x04, 0x00000000 },
+	{ 0x00179c,   1, 0x04, 0x00000000 },
+	{ 0x001228,   1, 0x04, 0x00000400 },
+	{ 0x00122c,   1, 0x04, 0x00000300 },
+	{ 0x001230,   1, 0x04, 0x00010001 },
+	{ 0x0007f8,   1, 0x04, 0x00000000 },
+	{ 0x0015b4,   1, 0x04, 0x00000001 },
+	{ 0x0015cc,   1, 0x04, 0x00000000 },
+	{ 0x001534,   1, 0x04, 0x00000000 },
+	{ 0x000fb0,   1, 0x04, 0x00000000 },
+	{ 0x0015d0,   1, 0x04, 0x00000000 },
+	{ 0x00153c,   1, 0x04, 0x00000000 },
+	{ 0x0016b4,   1, 0x04, 0x00000003 },
+	{ 0x000fbc,   4, 0x04, 0x0000ffff },
+	{ 0x000df8,   2, 0x04, 0x00000000 },
+	{ 0x001948,   1, 0x04, 0x00000000 },
+	{ 0x001970,   1, 0x04, 0x00000001 },
+	{ 0x00161c,   1, 0x04, 0x000009f0 },
+	{ 0x000dcc,   1, 0x04, 0x00000010 },
+	{ 0x00163c,   1, 0x04, 0x00000000 },
+	{ 0x0015e4,   1, 0x04, 0x00000000 },
+	{ 0x001160,  32, 0x04, 0x25e00040 },
+	{ 0x001880,  32, 0x04, 0x00000000 },
+	{ 0x000f84,   2, 0x04, 0x00000000 },
+	{ 0x0017c8,   2, 0x04, 0x00000000 },
+	{ 0x0017d0,   1, 0x04, 0x000000ff },
+	{ 0x0017d4,   1, 0x04, 0xffffffff },
+	{ 0x0017d8,   1, 0x04, 0x00000002 },
+	{ 0x0017dc,   1, 0x04, 0x00000000 },
+	{ 0x0015f4,   2, 0x04, 0x00000000 },
+	{ 0x001434,   2, 0x04, 0x00000000 },
+	{ 0x000d74,   1, 0x04, 0x00000000 },
+	{ 0x000dec,   1, 0x04, 0x00000001 },
+	{ 0x0013a4,   1, 0x04, 0x00000000 },
+	{ 0x001318,   1, 0x04, 0x00000001 },
+	{ 0x001644,   1, 0x04, 0x00000000 },
+	{ 0x000748,   1, 0x04, 0x00000000 },
+	{ 0x000de8,   1, 0x04, 0x00000000 },
+	{ 0x001648,   1, 0x04, 0x00000000 },
+	{ 0x0012a4,   1, 0x04, 0x00000000 },
+	{ 0x001120,   4, 0x04, 0x00000000 },
+	{ 0x001118,   1, 0x04, 0x00000000 },
+	{ 0x00164c,   1, 0x04, 0x00000000 },
+	{ 0x001658,   1, 0x04, 0x00000000 },
+	{ 0x001910,   1, 0x04, 0x00000290 },
+	{ 0x001518,   1, 0x04, 0x00000000 },
+	{ 0x00165c,   1, 0x04, 0x00000001 },
+	{ 0x001520,   1, 0x04, 0x00000000 },
+	{ 0x001604,   1, 0x04, 0x00000000 },
+	{ 0x001570,   1, 0x04, 0x00000000 },
+	{ 0x0013b0,   2, 0x04, 0x3f800000 },
+	{ 0x00020c,   1, 0x04, 0x00000000 },
+	{ 0x001670,   1, 0x04, 0x30201000 },
+	{ 0x001674,   1, 0x04, 0x70605040 },
+	{ 0x001678,   1, 0x04, 0xb8a89888 },
+	{ 0x00167c,   1, 0x04, 0xf8e8d8c8 },
+	{ 0x00166c,   1, 0x04, 0x00000000 },
+	{ 0x001680,   1, 0x04, 0x00ffff00 },
+	{ 0x0012d0,   1, 0x04, 0x00000003 },
+	{ 0x0012d4,   1, 0x04, 0x00000002 },
+	{ 0x001684,   2, 0x04, 0x00000000 },
+	{ 0x000dac,   2, 0x04, 0x00001b02 },
+	{ 0x000db4,   1, 0x04, 0x00000000 },
+	{ 0x00168c,   1, 0x04, 0x00000000 },
+	{ 0x0015bc,   1, 0x04, 0x00000000 },
+	{ 0x00156c,   1, 0x04, 0x00000000 },
+	{ 0x00187c,   1, 0x04, 0x00000000 },
+	{ 0x001110,   1, 0x04, 0x00000001 },
+	{ 0x000dc0,   3, 0x04, 0x00000000 },
+	{ 0x001234,   1, 0x04, 0x00000000 },
+	{ 0x001690,   1, 0x04, 0x00000000 },
+	{ 0x0012ac,   1, 0x04, 0x00000001 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x001000,   1, 0x04, 0x00000010 },
+	{ 0x0010fc,   1, 0x04, 0x00000000 },
+	{ 0x001290,   1, 0x04, 0x00000000 },
+	{ 0x000218,   1, 0x04, 0x00000010 },
+	{ 0x0012d8,   1, 0x04, 0x00000000 },
+	{ 0x0012dc,   1, 0x04, 0x00000010 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x00155c,   2, 0x04, 0x00000000 },
+	{ 0x001564,   1, 0x04, 0x00000fff },
+	{ 0x001574,   2, 0x04, 0x00000000 },
+	{ 0x00157c,   1, 0x04, 0x000fffff },
+	{ 0x001354,   1, 0x04, 0x00000000 },
+	{ 0x001610,   1, 0x04, 0x00000012 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x00260c,   1, 0x04, 0x00000000 },
+	{ 0x0007ac,   1, 0x04, 0x00000000 },
+	{ 0x00162c,   1, 0x04, 0x00000003 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000324,   6, 0x04, 0x3f800000 },
+	{ 0x000750,   1, 0x04, 0x00000000 },
+	{ 0x000760,   1, 0x04, 0x39291909 },
+	{ 0x000764,   1, 0x04, 0x79695949 },
+	{ 0x000768,   1, 0x04, 0xb9a99989 },
+	{ 0x00076c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x000770,   1, 0x04, 0x30201000 },
+	{ 0x000774,   1, 0x04, 0x70605040 },
+	{ 0x000778,   1, 0x04, 0x00009080 },
+	{ 0x000780,   1, 0x04, 0x39291909 },
+	{ 0x000784,   1, 0x04, 0x79695949 },
+	{ 0x000788,   1, 0x04, 0xb9a99989 },
+	{ 0x00078c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x0007d0,   1, 0x04, 0x30201000 },
+	{ 0x0007d4,   1, 0x04, 0x70605040 },
+	{ 0x0007d8,   1, 0x04, 0x00009080 },
+	{ 0x00037c,   1, 0x04, 0x00000001 },
+	{ 0x000740,   2, 0x04, 0x00000000 },
+	{ 0x002600,   1, 0x04, 0x00000000 },
+	{ 0x001918,   1, 0x04, 0x00000000 },
+	{ 0x00191c,   1, 0x04, 0x00000900 },
+	{ 0x001920,   1, 0x04, 0x00000405 },
+	{ 0x001308,   1, 0x04, 0x00000001 },
+	{ 0x001924,   1, 0x04, 0x00000000 },
+	{ 0x0013ac,   1, 0x04, 0x00000000 },
+	{ 0x00192c,   1, 0x04, 0x00000001 },
+	{ 0x00193c,   1, 0x04, 0x00002c1c },
+	{ 0x000d7c,   1, 0x04, 0x00000000 },
+	{ 0x000f8c,   1, 0x04, 0x00000000 },
+	{ 0x0002c0,   1, 0x04, 0x00000001 },
+	{ 0x001510,   1, 0x04, 0x00000000 },
+	{ 0x001940,   1, 0x04, 0x00000000 },
+	{ 0x000ff4,   2, 0x04, 0x00000000 },
+	{ 0x00194c,   2, 0x04, 0x00000000 },
+	{ 0x001968,   1, 0x04, 0x00000000 },
+	{ 0x001590,   1, 0x04, 0x0000003f },
+	{ 0x0007e8,   4, 0x04, 0x00000000 },
+	{ 0x00196c,   1, 0x04, 0x00000011 },
+	{ 0x0002e4,   1, 0x04, 0x0000b001 },
+	{ 0x00036c,   2, 0x04, 0x00000000 },
+	{ 0x00197c,   1, 0x04, 0x00000000 },
+	{ 0x000fcc,   2, 0x04, 0x00000000 },
+	{ 0x0002d8,   1, 0x04, 0x00000040 },
+	{ 0x001980,   1, 0x04, 0x00000080 },
+	{ 0x001504,   1, 0x04, 0x00000080 },
+	{ 0x001984,   1, 0x04, 0x00000000 },
+	{ 0x000300,   1, 0x04, 0x00000001 },
+	{ 0x0013a8,   1, 0x04, 0x00000000 },
+	{ 0x0012ec,   1, 0x04, 0x00000000 },
+	{ 0x001310,   1, 0x04, 0x00000000 },
+	{ 0x001314,   1, 0x04, 0x00000001 },
+	{ 0x001380,   1, 0x04, 0x00000000 },
+	{ 0x001384,   4, 0x04, 0x00000001 },
+	{ 0x001394,   1, 0x04, 0x00000000 },
+	{ 0x00139c,   1, 0x04, 0x00000000 },
+	{ 0x001398,   1, 0x04, 0x00000000 },
+	{ 0x001594,   1, 0x04, 0x00000000 },
+	{ 0x001598,   4, 0x04, 0x00000001 },
+	{ 0x000f54,   3, 0x04, 0x00000000 },
+	{ 0x0019bc,   1, 0x04, 0x00000000 },
+	{ 0x000f9c,   2, 0x04, 0x00000000 },
+	{ 0x0012cc,   1, 0x04, 0x00000000 },
+	{ 0x0012e8,   1, 0x04, 0x00000000 },
+	{ 0x00130c,   1, 0x04, 0x00000001 },
+	{ 0x001360,   8, 0x04, 0x00000000 },
+	{ 0x00133c,   2, 0x04, 0x00000001 },
+	{ 0x001344,   1, 0x04, 0x00000002 },
+	{ 0x001348,   2, 0x04, 0x00000001 },
+	{ 0x001350,   1, 0x04, 0x00000002 },
+	{ 0x001358,   1, 0x04, 0x00000001 },
+	{ 0x0012e4,   1, 0x04, 0x00000000 },
+	{ 0x00131c,   1, 0x04, 0x00000000 },
+	{ 0x001320,   3, 0x04, 0x00000000 },
+	{ 0x0019c0,   1, 0x04, 0x00000000 },
+	{ 0x001140,   1, 0x04, 0x00000000 },
+	{ 0x0019c4,   1, 0x04, 0x00000000 },
+	{ 0x0019c8,   1, 0x04, 0x00001500 },
+	{ 0x00135c,   1, 0x04, 0x00000000 },
+	{ 0x000f90,   1, 0x04, 0x00000000 },
+	{ 0x0019e0,   8, 0x04, 0x00000001 },
+	{ 0x0019cc,   1, 0x04, 0x00000001 },
+	{ 0x0015b8,   1, 0x04, 0x00000000 },
+	{ 0x001a00,   1, 0x04, 0x00001111 },
+	{ 0x001a04,   7, 0x04, 0x00000000 },
+	{ 0x000d6c,   2, 0x04, 0xffff0000 },
+	{ 0x0010f8,   1, 0x04, 0x00001010 },
+	{ 0x000d80,   5, 0x04, 0x00000000 },
+	{ 0x000da0,   1, 0x04, 0x00000000 },
+	{ 0x0007a4,   2, 0x04, 0x00000000 },
+	{ 0x001508,   1, 0x04, 0x80000000 },
+	{ 0x00150c,   1, 0x04, 0x40000000 },
+	{ 0x001668,   1, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000008 },
+	{ 0x000d9c,   1, 0x04, 0x00000001 },
+	{ 0x000374,   1, 0x04, 0x00000000 },
+	{ 0x000378,   1, 0x04, 0x00000020 },
+	{ 0x0007dc,   1, 0x04, 0x00000000 },
+	{ 0x00074c,   1, 0x04, 0x00000055 },
+	{ 0x001420,   1, 0x04, 0x00000003 },
+	{ 0x0017bc,   2, 0x04, 0x00000000 },
+	{ 0x0017c4,   1, 0x04, 0x00000001 },
+	{ 0x001008,   1, 0x04, 0x00000008 },
+	{ 0x00100c,   1, 0x04, 0x00000040 },
+	{ 0x001010,   1, 0x04, 0x0000012c },
+	{ 0x000d60,   1, 0x04, 0x00000040 },
+	{ 0x00075c,   1, 0x04, 0x00000003 },
+	{ 0x001018,   1, 0x04, 0x00000020 },
+	{ 0x00101c,   1, 0x04, 0x00000001 },
+	{ 0x001020,   1, 0x04, 0x00000020 },
+	{ 0x001024,   1, 0x04, 0x00000001 },
+	{ 0x001444,   3, 0x04, 0x00000000 },
+	{ 0x000360,   1, 0x04, 0x20164010 },
+	{ 0x000364,   1, 0x04, 0x00000020 },
+	{ 0x000368,   1, 0x04, 0x00000000 },
+	{ 0x000de4,   1, 0x04, 0x00000000 },
+	{ 0x000204,   1, 0x04, 0x00000006 },
+	{ 0x000208,   1, 0x04, 0x00000000 },
+	{ 0x0002cc,   2, 0x04, 0x003fffff },
+	{ 0x001220,   1, 0x04, 0x00000005 },
+	{ 0x000fdc,   1, 0x04, 0x00000000 },
+	{ 0x000f98,   1, 0x04, 0x00400008 },
+	{ 0x001284,   1, 0x04, 0x08000080 },
+	{ 0x001450,   1, 0x04, 0x00400008 },
+	{ 0x001454,   1, 0x04, 0x08000080 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk40xx[] = {
+	{ 0x404010,   5, 0x04, 0x00000000 },
+	{ 0x404024,   1, 0x04, 0x0000e000 },
+	{ 0x404028,   1, 0x04, 0x00000000 },
+	{ 0x4040a8,   1, 0x04, 0x00000000 },
+	{ 0x4040ac,   7, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf800008f },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x4041a0,   4, 0x04, 0x00000000 },
+	{ 0x404200,   4, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk46xx[] = {
+	{ 0x404604,   1, 0x04, 0x00000014 },
+	{ 0x404608,   1, 0x04, 0x00000000 },
+	{ 0x40460c,   1, 0x04, 0x00003fff },
+	{ 0x404610,   1, 0x04, 0x00000100 },
+	{ 0x404618,   4, 0x04, 0x00000000 },
+	{ 0x40462c,   2, 0x04, 0x00000000 },
+	{ 0x404640,   1, 0x04, 0x00000000 },
+	{ 0x404654,   1, 0x04, 0x00000000 },
+	{ 0x404660,   1, 0x04, 0x00000000 },
+	{ 0x404678,   1, 0x04, 0x00000000 },
+	{ 0x40467c,   1, 0x04, 0x00000002 },
+	{ 0x404680,   8, 0x04, 0x00000000 },
+	{ 0x4046a0,   1, 0x04, 0x007f0080 },
+	{ 0x4046a4,   8, 0x04, 0x00000000 },
+	{ 0x4046c8,   3, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk47xx[] = {
+	{ 0x404700,   3, 0x04, 0x00000000 },
+	{ 0x404718,   7, 0x04, 0x00000000 },
+	{ 0x404734,   1, 0x04, 0x00000100 },
+	{ 0x404738,   2, 0x04, 0x00000000 },
+	{ 0x404744,   2, 0x04, 0x00000000 },
+	{ 0x404754,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180648 },
+	{ 0x405834,   1, 0x04, 0x08000000 },
+	{ 0x405838,   1, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk5bxx[] = {
+	{ 0x405b00,   1, 0x04, 0x00000000 },
+	{ 0x405b10,   1, 0x04, 0x00001000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk60xx[] = {
+	{ 0x406020,   1, 0x04, 0x004103c1 },
+	{ 0x406028,   4, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   2, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x801a00f0 },
+	{ 0x4064c4,   1, 0x04, 0x0192ffff },
+	{ 0x4064c8,   1, 0x04, 0x01800600 },
+	{ 0x4064cc,   9, 0x04, 0x00000000 },
+	{ 0x4064fc,   1, 0x04, 0x0000022a },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk70xx[] = {
+	{ 0x407040,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk80xx[] = {
+	{ 0x408000,   2, 0x04, 0x00000000 },
+	{ 0x408008,   1, 0x04, 0x00000030 },
+	{ 0x40800c,   2, 0x04, 0x00000000 },
+	{ 0x408014,   1, 0x04, 0x00000069 },
+	{ 0x408018,   1, 0x04, 0xe100e100 },
+	{ 0x408064,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1043e005 },
+	{ 0x408840,   1, 0x04, 0x0000000b },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x62000001 },
+	{ 0x408908,   1, 0x04, 0x00c8102f },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00000044 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c40,   1, 0x04, 0xffffffff },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_tpc[] = {
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000000f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000021 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x0000c000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419a30,   1, 0x04, 0x00000001 },
+	{ 0x419ac4,   1, 0x04, 0x0037f440 },
+	{ 0x419c00,   1, 0x04, 0x0000000a },
+	{ 0x419c04,   1, 0x04, 0x80000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3efbefbe },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00003203 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000402 },
+	{ 0x419e44,   1, 0x04, 0x0013eff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000007f },
+	{ 0x419e50,  19, 0x04, 0x00000000 },
+	{ 0x419eac,   1, 0x04, 0x00001f8f },
+	{ 0x419eb0,   1, 0x04, 0x00000d3f },
+	{ 0x419ec8,   1, 0x04, 0x0001304f },
+	{ 0x419f30,   8, 0x04, 0x00000000 },
+	{ 0x419f58,   1, 0x04, 0x00000000 },
+	{ 0x419f70,   1, 0x04, 0x00000000 },
+	{ 0x419f78,   1, 0x04, 0x0000000b },
+	{ 0x419f7c,   1, 0x04, 0x0000027a },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk[] = {
+	{ 0x41be24,   1, 0x04, 0x00000006 },
+	{ 0x41bec0,   1, 0x04, 0x12180000 },
+	{ 0x41bec4,   1, 0x04, 0x00037f7f },
+	{ 0x41bee4,   1, 0x04, 0x06480430 },
+	{ 0x41bf00,   1, 0x04, 0x0a418820 },
+	{ 0x41bf04,   1, 0x04, 0x062080e6 },
+	{ 0x41bf08,   1, 0x04, 0x020398a4 },
+	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
+	{ 0x41bf10,   1, 0x04, 0x0a418820 },
+	{ 0x41bf14,   1, 0x04, 0x000000e6 },
+	{ 0x41bfd0,   1, 0x04, 0x00900103 },
+	{ 0x41bfe0,   1, 0x04, 0x00400001 },
+	{ 0x41bfe4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static void
+nve4_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	u32 magic[GPC_MAX][2];
+	u32 offset;
+	int gpc;
+
+	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x4064cc, 0x80000000,  0, 0);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000030,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000030,  0, 0);
+	mmio_list(0x4064c8, 0x01800600,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+
+	mmio_list(0x405830, 0x02180648,  0, 0);
+	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
+		u16 magic1 = 0x0648 * priv->tpc_nr[gpc];
+		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
+		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
+		offset += 0x0324 * priv->tpc_nr[gpc];
+	}
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
+		offset += 0x07ff * priv->tpc_nr[gpc];
+	}
+
+	mmio_list(0x17e91c, 0x06060609, 0, 0);
+	mmio_list(0x17e920, 0x00090a05, 0, 0);
+}
+
+void
+nve4_grctx_generate_unkn(struct nvc0_graph_priv *priv)
+{
+	nv_mask(priv, 0x418c6c, 0x00000001, 0x00000001);
+	nv_mask(priv, 0x41980c, 0x00000010, 0x00000010);
+	nv_mask(priv, 0x41be08, 0x00000004, 0x00000004);
+	nv_mask(priv, 0x4064c0, 0x80000000, 0x80000000);
+	nv_mask(priv, 0x405800, 0x08000000, 0x08000000);
+	nv_mask(priv, 0x419c00, 0x00000008, 0x00000008);
+}
+
+void
+nve4_grctx_generate_r418bb8(struct nvc0_graph_priv *priv)
+{
+	u32 data[6] = {}, data2[2] = {};
+	u8  tpcnr[GPC_MAX];
+	u8  shift, ntpcv;
+	int gpc, tpc, i;
+
+	/* calculate first set of magics */
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+
+	gpc = -1;
+	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpcnr[gpc]--;
+
+		data[tpc / 6] |= gpc << ((tpc % 6) * 5);
+	}
+
+	for (; tpc < 32; tpc++)
+		data[tpc / 6] |= 7 << ((tpc % 6) * 5);
+
+	/* and the second... */
+	shift = 0;
+	ntpcv = priv->tpc_total;
+	while (!(ntpcv & (1 << 4))) {
+		ntpcv <<= 1;
+		shift++;
+	}
+
+	data2[0]  = (ntpcv << 16);
+	data2[0] |= (shift << 21);
+	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
+	for (i = 1; i < 7; i++)
+		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
+
+	/* GPC_BROADCAST */
+	nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
+
+	/* GPC_BROADCAST.TP_BROADCAST */
+	nv_wr32(priv, 0x41bfd0, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr | data2[0]);
+	nv_wr32(priv, 0x41bfe4, data2[1]);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x41bf00 + (i * 4), data[i]);
+
+	/* UNK78xx */
+	nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x40780c + (i * 4), data[i]);
+}
+
+void
+nve4_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	int i;
+
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+
+	for (i = 0; oclass->hub[i]; i++)
+		nvc0_graph_mmio(priv, oclass->hub[i]);
+	for (i = 0; oclass->gpc[i]; i++)
+		nvc0_graph_mmio(priv, oclass->gpc[i]);
+
+	nv_wr32(priv, 0x404154, 0x00000000);
+
+	oclass->mods(priv, info);
+	oclass->unkn(priv);
+
+	nvc0_grctx_generate_tpcid(priv);
+	nvc0_grctx_generate_r406028(priv);
+	nve4_grctx_generate_r418bb8(priv);
+	nvc0_grctx_generate_r406800(priv);
+
+	for (i = 0; i < 8; i++)
+		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
+
+	nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr);
+	if (priv->gpc_nr == 1) {
+		nv_mask(priv, 0x408850, 0x0000000f, priv->tpc_nr[0]);
+		nv_mask(priv, 0x408958, 0x0000000f, priv->tpc_nr[0]);
+	} else {
+		nv_mask(priv, 0x408850, 0x0000000f, priv->gpc_nr);
+		nv_mask(priv, 0x408958, 0x0000000f, priv->gpc_nr);
+	}
+	nv_mask(priv, 0x419f78, 0x00000001, 0x00000000);
+
+	nvc0_graph_icmd(priv, oclass->icmd);
+	nv_wr32(priv, 0x404154, 0x00000400);
+	nvc0_graph_mthd(priv, oclass->mthd);
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+
+	nv_mask(priv, 0x418800, 0x00200000, 0x00200000);
+	nv_mask(priv, 0x41be10, 0x00800000, 0x00800000);
+}
+
+static struct nvc0_graph_init *
+nve4_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nve4_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nve4_grctx_init_unk46xx,
+	nve4_grctx_init_unk47xx,
+	nve4_grctx_init_unk58xx,
+	nve4_grctx_init_unk5bxx,
+	nve4_grctx_init_unk60xx,
+	nve4_grctx_init_unk64xx,
+	nve4_grctx_init_unk70xx,
+	nvc0_grctx_init_unk78xx,
+	nve4_grctx_init_unk80xx,
+	nve4_grctx_init_rop,
+	NULL
+};
+
+struct nvc0_graph_init *
+nve4_grctx_init_gpc[] = {
+	nve4_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nve4_grctx_init_tpc,
+	nve4_grctx_init_unk,
+	NULL
+};
+
+static struct nvc0_graph_mthd
+nve4_grctx_init_mthd[] = {
+	{ 0xa097, nve4_grctx_init_a097, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nve4_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xe4),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nve4_grctx_generate_main,
+	.mods = nve4_grctx_generate_mods,
+	.unkn = nve4_grctx_generate_unkn,
+	.hub  = nve4_grctx_init_hub,
+	.gpc  = nve4_grctx_init_gpc,
+	.icmd = nve4_grctx_init_icmd,
+	.mthd = nve4_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
new file mode 100644
index 0000000..dcb2ebb
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
@@ -0,0 +1,328 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk40xx[] = {
+	{ 0x404004,   8, 0x04, 0x00000000 },
+	{ 0x404024,   1, 0x04, 0x0000e000 },
+	{ 0x404028,   8, 0x04, 0x00000000 },
+	{ 0x4040a8,   8, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf800008f },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404100,  10, 0x04, 0x00000000 },
+	{ 0x404130,   2, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x40417c,   2, 0x04, 0x00000000 },
+	{ 0x4041a0,   4, 0x04, 0x00000000 },
+	{ 0x404200,   1, 0x04, 0x0000a197 },
+	{ 0x404204,   1, 0x04, 0x0000a1c0 },
+	{ 0x404208,   1, 0x04, 0x0000a140 },
+	{ 0x40420c,   1, 0x04, 0x0000902d },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk44xx[] = {
+	{ 0x404404,  12, 0x04, 0x00000000 },
+	{ 0x404438,   1, 0x04, 0x00000000 },
+	{ 0x404460,   2, 0x04, 0x00000000 },
+	{ 0x404468,   1, 0x04, 0x00ffffff },
+	{ 0x40446c,   1, 0x04, 0x00000000 },
+	{ 0x404480,   1, 0x04, 0x00000001 },
+	{ 0x404498,   1, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk5bxx[] = {
+	{ 0x405b00,   1, 0x04, 0x00000000 },
+	{ 0x405b10,   1, 0x04, 0x00001000 },
+	{ 0x405b20,   1, 0x04, 0x04000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk60xx[] = {
+	{ 0x406020,   1, 0x04, 0x034103c1 },
+	{ 0x406028,   4, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b0,   3, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x802000f0 },
+	{ 0x4064c4,   1, 0x04, 0x0192ffff },
+	{ 0x4064c8,   1, 0x04, 0x018007c0 },
+	{ 0x4064cc,   9, 0x04, 0x00000000 },
+	{ 0x4064fc,   1, 0x04, 0x0000022a },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk88xx[] = {
+	{ 0x408800,   1, 0x04, 0x12802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1003e005 },
+	{ 0x408840,   1, 0x04, 0x0000000b },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x62000001 },
+	{ 0x408908,   1, 0x04, 0x00c8102f },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   1, 0x04, 0x00000000 },
+	{ 0x41880c,   1, 0x04, 0x00000030 },
+	{ 0x418810,   1, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00000044 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c40,   1, 0x04, 0xffffffff },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x418d24,   1, 0x04, 0x00000000 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_tpc[] = {
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000000f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000021 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x0000c000 },
+	{ 0x419a20,   1, 0x04, 0x00020800 },
+	{ 0x419a30,   1, 0x04, 0x00000001 },
+	{ 0x419ac4,   1, 0x04, 0x0037f440 },
+	{ 0x419c00,   1, 0x04, 0x0000001a },
+	{ 0x419c04,   1, 0x04, 0x80000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3efbefbe },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000203 },
+	{ 0x419e04,   1, 0x04, 0x00000000 },
+	{ 0x419e08,   1, 0x04, 0x0000001d },
+	{ 0x419e0c,   1, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00001c02 },
+	{ 0x419e44,   1, 0x04, 0x0013eff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000007f },
+	{ 0x419e50,   2, 0x04, 0x00000000 },
+	{ 0x419e58,   1, 0x04, 0x00000001 },
+	{ 0x419e5c,   3, 0x04, 0x00000000 },
+	{ 0x419e68,   1, 0x04, 0x00000002 },
+	{ 0x419e6c,  12, 0x04, 0x00000000 },
+	{ 0x419eac,   1, 0x04, 0x00001fcf },
+	{ 0x419eb0,   1, 0x04, 0x0db00da0 },
+	{ 0x419eb8,   1, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0001304f },
+	{ 0x419f30,   4, 0x04, 0x00000000 },
+	{ 0x419f40,   1, 0x04, 0x00000018 },
+	{ 0x419f44,   3, 0x04, 0x00000000 },
+	{ 0x419f58,   1, 0x04, 0x00000000 },
+	{ 0x419f70,   1, 0x04, 0x00007300 },
+	{ 0x419f78,   1, 0x04, 0x000000eb },
+	{ 0x419f7c,   1, 0x04, 0x00000404 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk[] = {
+	{ 0x41be24,   1, 0x04, 0x00000006 },
+	{ 0x41bec0,   1, 0x04, 0x10000000 },
+	{ 0x41bec4,   1, 0x04, 0x00037f7f },
+	{ 0x41bee4,   1, 0x04, 0x00000000 },
+	{ 0x41bf00,   1, 0x04, 0x0a418820 },
+	{ 0x41bf04,   1, 0x04, 0x062080e6 },
+	{ 0x41bf08,   1, 0x04, 0x020398a4 },
+	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
+	{ 0x41bf10,   1, 0x04, 0x0a418820 },
+	{ 0x41bf14,   1, 0x04, 0x000000e6 },
+	{ 0x41bfd0,   1, 0x04, 0x00900103 },
+	{ 0x41bfe0,   1, 0x04, 0x00400001 },
+	{ 0x41bfe4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static void
+nvf0_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	u32 magic[GPC_MAX][4];
+	u32 offset;
+	int gpc;
+
+	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x4064cc, 0x80000000,  0, 0);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000030,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000030,  0, 0);
+	mmio_list(0x4064c8, 0x01800600,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+
+	mmio_list(0x405830, 0x02180648,  0, 0);
+	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		u16 magic0 = 0x0218 * (priv->tpc_nr[gpc] - 1);
+		u16 magic1 = 0x0648 * (priv->tpc_nr[gpc] - 1);
+		u16 magic2 = 0x0218;
+		u16 magic3 = 0x0648;
+		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
+		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
+		offset += 0x0324 * (priv->tpc_nr[gpc] - 1);;
+		magic[gpc][2]  = 0x10000000 | (magic2 << 16) | offset;
+		magic[gpc][3]  = 0x00000000 | (magic3 << 16);
+		offset += 0x0324;
+	}
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
+		offset += 0x07ff * (priv->tpc_nr[gpc] - 1);
+		mmio_list(GPC_UNIT(gpc, 0x32c0), magic[gpc][2], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x32e4), magic[gpc][3] | offset, 0, 0);
+		offset += 0x07ff;
+	}
+
+	mmio_list(0x17e91c, 0x06060609, 0, 0);
+	mmio_list(0x17e920, 0x00090a05, 0, 0);
+}
+
+static struct nvc0_graph_init *
+nvf0_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvf0_grctx_init_unk40xx,
+	nvf0_grctx_init_unk44xx,
+	nve4_grctx_init_unk46xx,
+	nve4_grctx_init_unk47xx,
+	nve4_grctx_init_unk58xx,
+	nvf0_grctx_init_unk5bxx,
+	nvf0_grctx_init_unk60xx,
+	nvf0_grctx_init_unk64xx,
+	nve4_grctx_init_unk80xx,
+	nvf0_grctx_init_unk88xx,
+	nvd9_grctx_init_rop,
+	NULL
+};
+
+struct nvc0_graph_init *
+nvf0_grctx_init_gpc[] = {
+	nvf0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvf0_grctx_init_tpc,
+	nvf0_grctx_init_unk,
+	NULL
+};
+
+static struct nvc0_graph_mthd
+nvf0_grctx_init_mthd[] = {
+	{ 0xa197, nvc1_grctx_init_9097, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvf0_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xf0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nve4_grctx_generate_main,
+	.mods = nvf0_grctx_generate_mods,
+	.unkn = nve4_grctx_generate_unkn,
+	.hub  = nvf0_grctx_init_hub,
+	.gpc  = nvf0_grctx_init_gpc,
+	.icmd = nvc0_grctx_init_icmd,
+	.mthd = nvf0_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/com.fuc
index e6b2288..5d24b6d 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/com.fuc
@@ -23,42 +23,7 @@
  * Authors: Ben Skeggs
  */
 
-define(`mmctx_data', `.b32 eval((($2 - 1) << 26) | $1)')
-define(`queue_init', `.skip eval((2 * 4) + ((8 * 4) * 2))')
-
-ifdef(`include_code', `
-// Error codes
-define(`E_BAD_COMMAND', 0x01)
-define(`E_CMD_OVERFLOW', 0x02)
-
-// Util macros to help with debugging ucode hangs etc
-define(`T_WAIT', 0)
-define(`T_MMCTX', 1)
-define(`T_STRWAIT', 2)
-define(`T_STRINIT', 3)
-define(`T_AUTO', 4)
-define(`T_CHAN', 5)
-define(`T_LOAD', 6)
-define(`T_SAVE', 7)
-define(`T_LCHAN', 8)
-define(`T_LCTXH', 9)
-
-define(`trace_set', `
-	mov $r8 0x83c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-define(`trace_clr', `
-	mov $r8 0x85c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
+#ifdef INCLUDE_CODE
 // queue_put - add request to queue
 //
 // In : $r13 queue pointer
@@ -178,27 +143,37 @@ watchdog_clear:
 	iowr I[$r8 + 0x000] $r0
 	ret
 
-// wait_done{z,o} - wait on FUC_DONE bit to become clear/set
+// wait_donez - wait on FUC_DONE bit to become clear
+//
+// In : $r10 bit to wait on
+//
+wait_donez:
+	trace_set(T_WAIT);
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_VAL(6), 0, $r10)
+	wait_donez_ne:
+		nv_iord($r8, NV_PGRAPH_FECS_SIGNAL, 0)
+		xbit $r8 $r8 $r10
+		bra ne #wait_donez_ne
+	trace_clr(T_WAIT)
+	ret
+
+// wait_doneo - wait on FUC_DONE bit to become set
 //
 // In : $r10 bit to wait on
 //
-define(`wait_done', `
-$1:
+wait_doneo:
 	trace_set(T_WAIT);
 	mov $r8 0x818
 	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r10	// CC_SCRATCH[6] = wait bit
-	wait_done_$1:
+	iowr I[$r8 + 0x000] $r10
+	wait_doneo_e:
 		mov $r8 0x400
 		shl b32 $r8 6
-		iord $r8 I[$r8 + 0x000]	// DONE
+		iord $r8 I[$r8 + 0x000]
 		xbit $r8 $r8 $r10
-		bra $2 #wait_done_$1
+		bra e #wait_doneo_e
 	trace_clr(T_WAIT)
 	ret
-')
-wait_done(wait_donez, ne)
-wait_done(wait_doneo, e)
 
 // mmctx_size - determine size of a mmio list transfer
 //
@@ -397,4 +372,4 @@ strand_ctx_init:
 	sub b32 $r15 $r14 $r15
 	trace_clr(T_STRINIT)
 	ret
-')
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpc.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpc.fuc
new file mode 100644
index 0000000..5547c1b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpc.fuc
@@ -0,0 +1,404 @@
+/* fuc microcode for nvc0 PGRAPH/GPC
+ *
+ * Copyright 2011 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+/* TODO
+ * - bracket certain functions with scratch writes, useful for debugging
+ * - watchdog timer around ctx operations
+ */
+
+#ifdef INCLUDE_DATA
+gpc_mmio_list_head:	.b32 #mmio_list_base
+gpc_mmio_list_tail:
+tpc_mmio_list_head:	.b32 #mmio_list_base
+tpc_mmio_list_tail:
+unk_mmio_list_head:	.b32 #mmio_list_base
+unk_mmio_list_tail:	.b32 #mmio_list_base
+
+gpc_id:			.b32 0
+
+tpc_count:		.b32 0
+tpc_mask:		.b32 0
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+unk_count:		.b32 0
+unk_mask:		.b32 0
+#endif
+
+cmd_queue:		queue_init
+
+mmio_list_base:
+#endif
+
+#ifdef INCLUDE_CODE
+// reports an exception to the host
+//
+// In: $r15 error code (see nvc0.fuc)
+//
+error:
+	push $r14
+	mov $r14 -0x67ec 	// 0x9814
+	sethi $r14 0x400000
+	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
+	add b32 $r14 0x41c
+	mov $r15 1
+	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
+	pop $r14
+	ret
+
+// GPC fuc initialisation, executed by triggering ucode start, will
+// fall through to main loop after completion.
+//
+// Input:
+//   CC_SCRATCH[1]: context base
+//
+// Output:
+//   CC_SCRATCH[0]:
+//	     31:31: set to signal completion
+//   CC_SCRATCH[1]:
+//	      31:0: GPC context size
+//
+init:
+	clear b32 $r0
+	mov $sp $r0
+
+	// enable fifo access
+	mov $r1 0x1200
+	mov $r2 2
+	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
+
+	// setup i0 handler, and route all interrupts to it
+	mov $r1 #ih
+	mov $iv0 $r1
+	mov $r1 0x400
+	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
+
+	// enable fifo interrupt
+	mov $r2 4
+	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
+
+	// enable interrupts
+	bset $flags ie0
+
+	// figure out which GPC we are, and how many TPCs we have
+	mov $r1 0x608
+	shl b32 $r1 6
+	iord $r2 I[$r1 + 0x000]		// UNITS
+	mov $r3 1
+	and $r2 0x1f
+	shl b32 $r3 $r2
+	sub b32 $r3 1
+	st b32 D[$r0 + #tpc_count] $r2
+	st b32 D[$r0 + #tpc_mask] $r3
+	add b32 $r1 0x400
+	iord $r2 I[$r1 + 0x000]		// MYINDEX
+	st b32 D[$r0 + #gpc_id] $r2
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+	// figure out which, and how many, UNKs are actually present
+	mov $r14 0x0c30
+	sethi $r14 0x500000
+	clear b32 $r2
+	clear b32 $r3
+	clear b32 $r4
+	init_unk_loop:
+		call #nv_rd32
+		cmp b32 $r15 0
+		bra z #init_unk_next
+			mov $r15 1
+			shl b32 $r15 $r2
+			or $r4 $r15
+			add b32 $r3 1
+		init_unk_next:
+		add b32 $r2 1
+		add b32 $r14 4
+		cmp b32 $r2 NV_PGRAPH_GPCX_UNK__SIZE
+		bra ne #init_unk_loop
+	init_unk_done:
+	st b32 D[$r0 + #unk_count] $r3
+	st b32 D[$r0 + #unk_mask] $r4
+#endif
+
+	// initialise context base, and size tracking
+	nv_iord($r2, NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(1), 0)
+	clear b32 $r3		// track GPC context size here
+
+	// set mmctx base addresses now so we don't have to do it later,
+	// they don't currently ever change
+	mov $r4 0x700
+	shl b32 $r4 6
+	shr b32 $r5 $r2 8
+	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
+	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
+
+	// calculate GPC mmio context size
+	ld b32 $r14 D[$r0 + #gpc_mmio_list_head]
+	ld b32 $r15 D[$r0 + #gpc_mmio_list_tail]
+	call #mmctx_size
+	add b32 $r2 $r15
+	add b32 $r3 $r15
+
+	// calculate per-TPC mmio context size
+	ld b32 $r14 D[$r0 + #tpc_mmio_list_head]
+	ld b32 $r15 D[$r0 + #tpc_mmio_list_tail]
+	call #mmctx_size
+	ld b32 $r14 D[$r0 + #tpc_count]
+	mulu $r14 $r15
+	add b32 $r2 $r14
+	add b32 $r3 $r14
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+	// calculate per-UNK mmio context size
+	ld b32 $r14 D[$r0 + #unk_mmio_list_head]
+	ld b32 $r15 D[$r0 + #unk_mmio_list_tail]
+	call #mmctx_size
+	ld b32 $r14 D[$r0 + #unk_count]
+	mulu $r14 $r15
+	add b32 $r2 $r14
+	add b32 $r3 $r14
+#endif
+
+	// round up base/size to 256 byte boundary (for strand SWBASE)
+	add b32 $r4 0x1300
+	shr b32 $r3 2
+	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
+	shr b32 $r2 8
+	shr b32 $r3 6
+	add b32 $r2 1
+	add b32 $r3 1
+	shl b32 $r2 8
+	shl b32 $r3 8
+
+	// calculate size of strand context data
+	mov b32 $r15 $r2
+	call #strand_ctx_init
+	add b32 $r3 $r15
+
+	// save context size, and tell HUB we're done
+	nv_iowr(NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(1), 0, $r3)
+	clear b32 $r2
+	bset $r2 31
+	nv_iowr(NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_SET(0), 0, $r2)
+
+// Main program loop, very simple, sleeps until woken up by the interrupt
+// handler, pulls a command from the queue and executes its handler
+//
+main:
+	bset $flags $p0
+	sleep $p0
+	mov $r13 #cmd_queue
+	call #queue_get
+	bra $p1 #main
+
+	// 0x0000-0x0003 are all context transfers
+	cmpu b32 $r14 0x04
+	bra nc #main_not_ctx_xfer
+		// fetch $flags and mask off $p1/$p2
+		mov $r1 $flags
+		mov $r2 0x0006
+		not b32 $r2
+		and $r1 $r2
+		// set $p1/$p2 according to transfer type
+		shl b32 $r14 1
+		or $r1 $r14
+		mov $flags $r1
+		// transfer context data
+		call #ctx_xfer
+		bra #main
+
+	main_not_ctx_xfer:
+	shl b32 $r15 $r14 16
+	or $r15 E_BAD_COMMAND
+	call #error
+	bra #main
+
+// interrupt handler
+ih:
+	push $r8
+	mov $r8 $flags
+	push $r8
+	push $r9
+	push $r10
+	push $r11
+	push $r13
+	push $r14
+	push $r15
+	clear b32 $r0
+
+	// incoming fifo command?
+	iord $r10 I[$r0 + 0x200]	// INTR
+	and $r11 $r10 0x00000004
+	bra e #ih_no_fifo
+		// queue incoming fifo command for later processing
+		mov $r11 0x1900
+		mov $r13 #cmd_queue
+		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
+		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
+		call #queue_put
+		add b32 $r11 0x400
+		mov $r14 1
+		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
+
+	// ack, and wake up main()
+	ih_no_fifo:
+	iowr I[$r0 + 0x100] $r10	// INTR_ACK
+
+	pop $r15
+	pop $r14
+	pop $r13
+	pop $r11
+	pop $r10
+	pop $r9
+	pop $r8
+	mov $flags $r8
+	pop $r8
+	bclr $flags $p0
+	iret
+
+// Set this GPC's bit in HUB_BAR, used to signal completion of various
+// activities to the HUB fuc
+//
+hub_barrier_done:
+	mov $r15 1
+	ld b32 $r14 D[$r0 + #gpc_id]
+	shl b32 $r15 $r14
+	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
+	sethi $r14 0x400000
+	call #nv_wr32
+	ret
+
+// Disables various things, waits a bit, and re-enables them..
+//
+// Not sure how exactly this helps, perhaps "ENABLE" is not such a
+// good description for the bits we turn off?  Anyways, without this,
+// funny things happen.
+//
+ctx_redswitch:
+	mov $r14 0x614
+	shl b32 $r14 6
+	mov $r15 0x020
+	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
+	mov $r15 8
+	ctx_redswitch_delay:
+		sub b32 $r15 1
+		bra ne #ctx_redswitch_delay
+	mov $r15 0xa20
+	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
+	ret
+
+// Transfer GPC context data between GPU and storage area
+//
+// In: $r15 context base address
+//     $p1 clear on save, set on load
+//     $p2 set if opposite direction done/will be done, so:
+//		on save it means: "a load will follow this save"
+//		on load it means: "a save preceeded this load"
+//
+ctx_xfer:
+	// set context base address
+	mov $r1 0xa04
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r15// MEM_BASE
+	bra not $p1 #ctx_xfer_not_load
+		call #ctx_redswitch
+	ctx_xfer_not_load:
+
+	// strands
+	mov $r1 0x4afc
+	sethi $r1 0x20000
+	mov $r2 0xc
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
+	call #strand_wait
+	mov $r2 0x47fc
+	sethi $r2 0x20000
+	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
+	xbit $r2 $flags $p1
+	add b32 $r2 3
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
+
+	// mmio context
+	xbit $r10 $flags $p1	// direction
+	or $r10 2		// first
+	mov $r11 0x0000
+	sethi $r11 0x500000
+	ld b32 $r12 D[$r0 + #gpc_id]
+	shl b32 $r12 15
+	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
+	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
+	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
+	mov $r14 0		// not multi
+	call #mmctx_xfer
+
+	// per-TPC mmio context
+	xbit $r10 $flags $p1	// direction
+#if !NV_PGRAPH_GPCX_UNK__SIZE
+	or $r10 4		// last
+#endif
+	mov $r11 0x4000
+	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
+	ld b32 $r12 D[$r0 + #gpc_id]
+	shl b32 $r12 15
+	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
+	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
+	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
+	ld b32 $r15 D[$r0 + #tpc_mask]
+	mov $r14 0x800		// stride = 0x800
+	call #mmctx_xfer
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+	// per-UNK mmio context
+	xbit $r10 $flags $p1	// direction
+	or $r10 4		// last
+	mov $r11 0x3000
+	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_UNK0
+	ld b32 $r12 D[$r0 + #gpc_id]
+	shl b32 $r12 15
+	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_UNK0
+	ld b32 $r12 D[$r0 + #unk_mmio_list_head]
+	ld b32 $r13 D[$r0 + #unk_mmio_list_tail]
+	ld b32 $r15 D[$r0 + #unk_mask]
+	mov $r14 0x200		// stride = 0x200
+	call #mmctx_xfer
+#endif
+
+	// wait for strands to finish
+	call #strand_wait
+
+	// if load, or a save without a load following, do some
+	// unknown stuff that's done after finishing a block of
+	// strand commands
+	bra $p1 #ctx_xfer_post
+	bra not $p2 #ctx_xfer_done
+	ctx_xfer_post:
+		mov $r1 0x4afc
+		sethi $r1 0x20000
+		mov $r2 0xd
+		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
+		call #strand_wait
+
+	// mark completion in HUB's barrier
+	ctx_xfer_done:
+	call #hub_barrier_done
+	ret
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
index f7055af..5ae06a2 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nvc0 PGRAPH/GPC
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,525 +19,24 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 gpcnvc0.fuc | envyas -a -w -m fuc -V fuc3 -o gpcnvc0.fuc.h
- */
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000000
 
-/* TODO
- * - bracket certain functions with scratch writes, useful for debugging
- * - watchdog timer around ctx operations
- */
+#define CHIPSET GF100
+#include "macros.fuc"
 
 .section #nvc0_grgpc_data
-include(`nvc0.fuc')
-gpc_id:			.b32 0
-gpc_mmio_list_head:	.b32 0
-gpc_mmio_list_tail:	.b32 0
-
-tpc_count:		.b32 0
-tpc_mask:		.b32 0
-tpc_mmio_list_head:	.b32 0
-tpc_mmio_list_tail:	.b32 0
-
-cmd_queue:		queue_init
-
-// chipset descriptions
-chipsets:
-.b8  0xc0 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc0_tpc_mmio_tail
-.b8  0xc1 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc1_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc1_tpc_mmio_tail
-.b8  0xc3 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xc4 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xc8 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc0_tpc_mmio_tail
-.b8  0xce 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xcf 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvcf_tpc_mmio_tail
-.b8  0xd9 0 0 0
-.b16 #nvd9_gpc_mmio_head
-.b16 #nvd9_gpc_mmio_tail
-.b16 #nvd9_tpc_mmio_head
-.b16 #nvd9_tpc_mmio_tail
-.b8  0xd7 0 0 0
-.b16 #nvd9_gpc_mmio_head
-.b16 #nvd9_gpc_mmio_tail
-.b16 #nvd9_tpc_mmio_head
-.b16 #nvd9_tpc_mmio_tail
-.b8  0 0 0 0
-
-// GPC mmio lists
-nvc0_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 6)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-nvc0_gpc_mmio_tail:
-mmctx_data(0x000c6c, 1);
-nvc1_gpc_mmio_tail:
-
-nvd9_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 2)
-mmctx_data(0x00040c, 3)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c6c, 1)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-nvd9_gpc_mmio_tail:
-
-// TPC mmio lists
-nvc0_tpc_mmio_head:
-mmctx_data(0x000018, 1)
-mmctx_data(0x00003c, 1)
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x000300, 6)
-mmctx_data(0x0003d0, 1)
-mmctx_data(0x0003e0, 2)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 1)
-mmctx_data(0x0004b0, 1)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000520, 2)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 20)
-mmctx_data(0x000698, 1)
-mmctx_data(0x000750, 2)
-nvc0_tpc_mmio_tail:
-mmctx_data(0x000758, 1)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x0006e0, 1)
-nvcf_tpc_mmio_tail:
-mmctx_data(0x0004bc, 1)
-nvc3_tpc_mmio_tail:
-mmctx_data(0x000544, 1)
-nvc1_tpc_mmio_tail:
-
-nvd9_tpc_mmio_head:
-mmctx_data(0x000018, 1)
-mmctx_data(0x00003c, 1)
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x000300, 6)
-mmctx_data(0x0003d0, 1)
-mmctx_data(0x0003e0, 2)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 3)
-mmctx_data(0x0004b0, 1)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000520, 2)
-mmctx_data(0x000544, 1)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 20)
-mmctx_data(0x000698, 1)
-mmctx_data(0x0006e0, 1)
-mmctx_data(0x000750, 3)
-nvd9_tpc_mmio_tail:
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
 
 .section #nvc0_grgpc_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nvc0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nvc0.fuc)
-//
-error:
-	push $r14
-	mov $r14 -0x67ec 	// 0x9814
-	sethi $r14 0x400000
-	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
-	add b32 $r14 0x41c
-	mov $r15 1
-	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
-	pop $r14
-	ret
-
-// GPC fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//   CC_SCRATCH[1]: context base
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: GPC context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
-
-	// enable fifo interrupt
-	mov $r2 4
-	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
-
-	// enable interrupts
-	bset $flags ie0
-
-	// figure out which GPC we are, and how many TPCs we have
-	mov $r1 0x608
-	shl b32 $r1 6
-	iord $r2 I[$r1 + 0x000]		// UNITS
-	mov $r3 1
-	and $r2 0x1f
-	shl b32 $r3 $r2
-	sub b32 $r3 1
-	st b32 D[$r0 + #tpc_count] $r2
-	st b32 D[$r0 + #tpc_mask] $r3
-	add b32 $r1 0x400
-	iord $r2 I[$r1 + 0x000]		// MYINDEX
-	st b32 D[$r0 + #gpc_id] $r2
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r1 #chipsets - 12
-	init_find_chipset:
-		add b32 $r1 12
-		ld b32 $r3 D[$r1 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// initialise context base, and size tracking
-	init_context:
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x100]	// CC_SCRATCH[1], initial base
-	clear b32 $r3		// track GPC context size here
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't currently ever change
-	mov $r4 0x700
-	shl b32 $r4 6
-	shr b32 $r5 $r2 8
-	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
-	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
-
-	// calculate GPC mmio context size, store the chipset-specific
-	// mmio list pointers somewhere we can get at them later without
-	// re-parsing the chipset list
-	clear b32 $r14
-	clear b32 $r15
-	ld b16 $r14 D[$r1 + 4]
-	ld b16 $r15 D[$r1 + 6]
-	st b16 D[$r0 + #gpc_mmio_list_head] $r14
-	st b16 D[$r0 + #gpc_mmio_list_tail] $r15
-	call #mmctx_size
-	add b32 $r2 $r15
-	add b32 $r3 $r15
-
-	// calculate per-TPC mmio context size, store the list pointers
-	ld b16 $r14 D[$r1 + 8]
-	ld b16 $r15 D[$r1 + 10]
-	st b16 D[$r0 + #tpc_mmio_list_head] $r14
-	st b16 D[$r0 + #tpc_mmio_list_tail] $r15
-	call #mmctx_size
-	ld b32 $r14 D[$r0 + #tpc_count]
-	mulu $r14 $r15
-	add b32 $r2 $r14
-	add b32 $r3 $r14
-
-	// round up base/size to 256 byte boundary (for strand SWBASE)
-	add b32 $r4 0x1300
-	shr b32 $r3 2
-	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
-	shr b32 $r2 8
-	shr b32 $r3 6
-	add b32 $r2 1
-	add b32 $r3 1
-	shl b32 $r2 8
-	shl b32 $r3 8
-
-	// calculate size of strand context data
-	mov b32 $r15 $r2
-	call #strand_ctx_init
-	add b32 $r3 $r15
-
-	// save context size, and tell HUB we're done
-	mov $r1 0x800
-	shl b32 $r1 6
-	iowr I[$r1 + 0x100] $r3		// CC_SCRATCH[1]  = context size
-	add b32 $r1 0x800
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// 0x0000-0x0003 are all context transfers
-	cmpu b32 $r14 0x04
-	bra nc #main_not_ctx_xfer
-		// fetch $flags and mask off $p1/$p2
-		mov $r1 $flags
-		mov $r2 0x0006
-		not b32 $r2
-		and $r1 $r2
-		// set $p1/$p2 according to transfer type
-		shl b32 $r14 1
-		or $r1 $r14
-		mov $flags $r1
-		// transfer context data
-		call #ctx_xfer
-		bra #main
-
-	main_not_ctx_xfer:
-	shl b32 $r15 $r14 16
-	or $r15 E_BAD_COMMAND
-	call #error
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// ack, and wake up main()
-	ih_no_fifo:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Set this GPC's bit in HUB_BAR, used to signal completion of various
-// activities to the HUB fuc
-//
-hub_barrier_done:
-	mov $r15 1
-	ld b32 $r14 D[$r0 + #gpc_id]
-	shl b32 $r15 $r14
-	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
-	sethi $r14 0x400000
-	call #nv_wr32
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x020
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0xa20
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
-	ret
-
-// Transfer GPC context data between GPU and storage area
-//
-// In: $r15 context base address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// set context base address
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r15// MEM_BASE
-	bra not $p1 #ctx_xfer_not_load
-		call #ctx_redswitch
-	ctx_xfer_not_load:
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 2		// first
-	mov $r11 0x0000
-	sethi $r11 0x500000
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
-	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// per-TPC mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 4		// last
-	mov $r11 0x4000
-	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
-	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
-	ld b32 $r15 D[$r0 + #tpc_mask]
-	mov $r14 0x800		// stride = 0x800
-	call #mmctx_xfer
-
-	// wait for strands to finish
-	call #strand_wait
-
-	// if load, or a save without a load following, do some
-	// unknown stuff that's done after finishing a block of
-	// strand commands
-	bra $p1 #ctx_xfer_post
-	bra not $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r1 0x4afc
-		sethi $r1 0x20000
-		mov $r2 0xd
-		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
-		call #strand_wait
-
-	// mark completion in HUB's barrier
-	ctx_xfer_done:
-	call #hub_barrier_done
-	ret
-
+#include "com.fuc"
+#include "gpc.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h
index 96050dd..f2b0dea 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h
@@ -1,17 +1,19 @@
 uint32_t nvc0_grgpc_data[] = {
-/* 0x0000: gpc_id */
-	0x00000000,
-/* 0x0004: gpc_mmio_list_head */
-	0x00000000,
-/* 0x0008: gpc_mmio_list_tail */
-	0x00000000,
-/* 0x000c: tpc_count */
-	0x00000000,
-/* 0x0010: tpc_mask */
+/* 0x0000: gpc_mmio_list_head */
+	0x00000064,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x00000064,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x00000064,
+/* 0x000c: unk_mmio_list_tail */
+	0x00000064,
+/* 0x0010: gpc_id */
 	0x00000000,
-/* 0x0014: tpc_mmio_list_head */
+/* 0x0014: tpc_count */
 	0x00000000,
-/* 0x0018: tpc_mmio_list_tail */
+/* 0x0018: tpc_mask */
 	0x00000000,
 /* 0x001c: cmd_queue */
 	0x00000000,
@@ -32,153 +34,17 @@ uint32_t nvc0_grgpc_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0064: chipsets */
-	0x000000c0,
-	0x012800c8,
-	0x01e40194,
-	0x000000c1,
-	0x012c00c8,
-	0x01f80194,
-	0x000000c3,
-	0x012800c8,
-	0x01f40194,
-	0x000000c4,
-	0x012800c8,
-	0x01f40194,
-	0x000000c8,
-	0x012800c8,
-	0x01e40194,
-	0x000000ce,
-	0x012800c8,
-	0x01f40194,
-	0x000000cf,
-	0x012800c8,
-	0x01f00194,
-	0x000000d9,
-	0x0194012c,
-	0x025401f8,
-	0x00000000,
-/* 0x00c8: nvc0_gpc_mmio_head */
-	0x00000380,
-	0x14000400,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-/* 0x0128: nvc0_gpc_mmio_tail */
-	0x00000c6c,
-/* 0x012c: nvc1_gpc_mmio_tail */
-/* 0x012c: nvd9_gpc_mmio_head */
-	0x00000380,
-	0x04000400,
-	0x0800040c,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c6c,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-/* 0x0194: nvd9_gpc_mmio_tail */
-/* 0x0194: nvc0_tpc_mmio_head */
-	0x00000018,
-	0x0000003c,
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x14000300,
-	0x000003d0,
-	0x040003e0,
-	0x08000400,
-	0x00000420,
-	0x000004b0,
-	0x000004e8,
-	0x000004f4,
-	0x04000520,
-	0x0c000604,
-	0x4c000644,
-	0x00000698,
-	0x04000750,
-/* 0x01e4: nvc0_tpc_mmio_tail */
-	0x00000758,
-	0x000002c4,
-	0x000006e0,
-/* 0x01f0: nvcf_tpc_mmio_tail */
-	0x000004bc,
-/* 0x01f4: nvc3_tpc_mmio_tail */
-	0x00000544,
-/* 0x01f8: nvc1_tpc_mmio_tail */
-/* 0x01f8: nvd9_tpc_mmio_head */
-	0x00000018,
-	0x0000003c,
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x000002c4,
-	0x14000300,
-	0x000003d0,
-	0x040003e0,
-	0x08000400,
-	0x08000420,
-	0x000004b0,
-	0x000004e8,
-	0x000004f4,
-	0x04000520,
-	0x00000544,
-	0x0c000604,
-	0x4c000644,
-	0x00000698,
-	0x000006e0,
-	0x08000750,
 };
 
 uint32_t nvc0_grgpc_code[] = {
-	0x03060ef5,
+	0x03180ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -210,7 +76,7 @@ uint32_t nvc0_grgpc_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -232,63 +98,66 @@ uint32_t nvc0_grgpc_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -296,8 +165,8 @@ uint32_t nvc0_grgpc_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -306,76 +175,77 @@ uint32_t nvc0_grgpc_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -384,86 +254,74 @@ uint32_t nvc0_grgpc_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe3f09814,
-	0x8d21f440,
-	0x041ce0b7,
-	0xf401f7f0,
-	0xe0fc8d21,
-/* 0x0306: init */
-	0x04bd00f8,
-	0xf10004fe,
-	0xf0120017,
-	0x12d00227,
-	0x3e17f100,
-	0x0010fe04,
-	0x040017f1,
-	0xf0c010d0,
-	0x12d00427,
-	0x1031f400,
-	0x060817f1,
-	0xcf0614b6,
-	0x37f00012,
-	0x1f24f001,
-	0xb60432bb,
-	0x02800132,
-	0x04038003,
-	0x040010b7,
-	0x800012cf,
-	0x27f10002,
-	0x24b60800,
-	0x0022cf06,
-/* 0x035f: init_find_chipset */
-	0xb65817f0,
-	0x13980c10,
-	0x0432b800,
-	0xb00b0bf4,
-	0x1bf40034,
-/* 0x0373: init_context */
-	0xf100f8f1,
-	0xb6080027,
-	0x22cf0624,
-	0xf134bd40,
-	0xb6070047,
-	0x25950644,
-	0x0045d008,
-	0xbd4045d0,
-	0x58f4bde4,
-	0x1f58021e,
-	0x020e4003,
-	0xf5040f40,
-	0xbb013d21,
-	0x3fbb002f,
-	0x041e5800,
-	0x40051f58,
-	0x0f400a0e,
-	0x3d21f50c,
-	0x030e9801,
-	0xbb00effd,
-	0x3ebb002e,
-	0x0040b700,
-	0x0235b613,
-	0xb60043d0,
-	0x35b60825,
-	0x0120b606,
-	0xb60130b6,
-	0x34b60824,
-	0x022fb908,
-	0x026321f5,
-	0xf1003fbb,
-	0xb6080017,
-	0x13d00614,
-	0x0010b740,
-	0xf024bd08,
-	0x12d01f29,
-/* 0x0401: main */
-	0x0031f400,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe042617,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x40b7003e,
+	0x35b61300,
+	0x0043d002,
+	0xb60825b6,
+	0x20b60635,
+	0x0130b601,
+	0xb60824b6,
+	0x2fb90834,
+	0x7121f502,
+	0x003fbb02,
+	0x010007f1,
+	0xd00203f0,
+	0x04bd0003,
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+/* 0x03e9: main */
+	0x0031f404,
 	0xf00028f4,
 	0x21f41cd7,
 	0xf401f439,
@@ -474,94 +332,100 @@ uint32_t nvc0_grgpc_code[] = {
 	0x01e4b604,
 	0xfe051efd,
 	0x21f50018,
-	0x0ef404c3,
-/* 0x0431: main_not_ctx_xfer */
+	0x0ef404ad,
+/* 0x0419: main_not_ctx_xfer */
 	0x10ef94d3,
 	0xf501f5f0,
-	0xf402ec21,
-/* 0x043e: ih */
+	0xf402fe21,
+/* 0x0426: ih */
 	0x80f9c60e,
 	0xf90188fe,
 	0xf990f980,
 	0xf9b0f9a0,
 	0xf9e0f9d0,
-	0x800acff0,
-	0xf404abc4,
-	0xb7f11d0b,
-	0xd7f01900,
-	0x40becf1c,
-	0xf400bfcf,
-	0xb0b70421,
-	0xe7f00400,
-	0x00bed001,
-/* 0x0474: ih_no_fifo */
-	0xfc400ad0,
-	0xfce0fcf0,
-	0xfcb0fcd0,
-	0xfc90fca0,
-	0x0088fe80,
-	0x32f480fc,
-/* 0x048f: hub_barrier_done */
-	0xf001f800,
-	0x0e9801f7,
-	0x04febb00,
-	0x9418e7f1,
-	0xf440e3f0,
-	0x00f88d21,
-/* 0x04a4: ctx_redswitch */
-	0x0614e7f1,
-	0xf006e4b6,
-	0xefd020f7,
-	0x08f7f000,
-/* 0x04b4: ctx_redswitch_delay */
-	0xf401f2b6,
-	0xf7f1fd1b,
-	0xefd00a20,
-/* 0x04c3: ctx_xfer */
-	0xf100f800,
-	0xb60a0417,
-	0x1fd00614,
-	0x0711f400,
-	0x04a421f5,
-/* 0x04d4: ctx_xfer_not_load */
-	0x4afc17f1,
-	0xf00213f0,
-	0x12d00c27,
-	0x0721f500,
-	0xfc27f102,
-	0x0223f047,
-	0xf00020d0,
-	0x20b6012c,
-	0x0012d003,
+	0xcf04bdf0,
+	0xabc4800a,
+	0x1d0bf404,
+	0x1900b7f1,
+	0xcf1cd7f0,
+	0xbfcf40be,
+	0x0421f400,
+	0x0400b0b7,
+	0xd001e7f0,
+/* 0x045e: ih_no_fifo */
+	0x0ad000be,
+	0xfcf0fc40,
+	0xfcd0fce0,
+	0xfca0fcb0,
+	0xfe80fc90,
+	0x80fc0088,
+	0xf80032f4,
+/* 0x0479: hub_barrier_done */
+	0x01f7f001,
+	0xbb040e98,
+	0xe7f104fe,
+	0xe3f09418,
+	0x8d21f440,
+/* 0x048e: ctx_redswitch */
+	0xe7f100f8,
+	0xe4b60614,
+	0x20f7f006,
+	0xf000efd0,
+/* 0x049e: ctx_redswitch_delay */
+	0xf2b608f7,
+	0xfd1bf401,
+	0x0a20f7f1,
+	0xf800efd0,
+/* 0x04ad: ctx_xfer */
+	0x0417f100,
+	0x0614b60a,
+	0xf4001fd0,
+	0x21f50711,
+/* 0x04be: ctx_xfer_not_load */
+	0x17f1048e,
+	0x13f04afc,
+	0x0c27f002,
+	0xf50012d0,
+	0xf1021521,
+	0xf047fc27,
+	0x20d00223,
+	0x012cf000,
+	0xd00320b6,
+	0xacf00012,
+	0x02a5f001,
+	0xf000b7f0,
+	0x0c9850b3,
+	0x0fc4b604,
+	0x9800bcbb,
+	0x0d98000c,
+	0x00e7f001,
+	0x016621f5,
 	0xf001acf0,
-	0xb7f002a5,
-	0x50b3f000,
-	0xb6000c98,
-	0xbcbb0fc4,
-	0x010c9800,
-	0xf0020d98,
-	0x21f500e7,
-	0xacf0015c,
-	0x04a5f001,
-	0x4000b7f1,
-	0x9850b3f0,
-	0xc4b6000c,
-	0x00bcbb0f,
-	0x98050c98,
-	0x0f98060d,
-	0x00e7f104,
-	0x5c21f508,
-	0x0721f501,
-	0x0601f402,
-/* 0x054b: ctx_xfer_post */
-	0xf11412f4,
-	0xf04afc17,
-	0x27f00213,
-	0x0012d00d,
-	0x020721f5,
-/* 0x055c: ctx_xfer_done */
-	0x048f21f5,
-	0x000000f8,
+	0xb7f104a5,
+	0xb3f04000,
+	0x040c9850,
+	0xbb0fc4b6,
+	0x0c9800bc,
+	0x020d9801,
+	0xf1060f98,
+	0xf50800e7,
+	0xf5016621,
+	0xf4021521,
+	0x12f40601,
+/* 0x0535: ctx_xfer_post */
+	0xfc17f114,
+	0x0213f04a,
+	0xd00d27f0,
+	0x21f50012,
+/* 0x0546: ctx_xfer_done */
+	0x21f50215,
+	0x00f80479,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc
new file mode 100644
index 0000000..c2f754e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000001
+
+#define CHIPSET GF117
+#include "macros.fuc"
+
+.section #nvd7_grgpc_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
+
+.section #nvd7_grgpc_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "gpc.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc.h
new file mode 100644
index 0000000..dd346c2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc.h
@@ -0,0 +1,475 @@
+uint32_t nvd7_grgpc_data[] = {
+/* 0x0000: gpc_mmio_list_head */
+	0x0000006c,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x0000006c,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x0000006c,
+/* 0x000c: unk_mmio_list_tail */
+	0x0000006c,
+/* 0x0010: gpc_id */
+	0x00000000,
+/* 0x0014: tpc_count */
+	0x00000000,
+/* 0x0018: tpc_mask */
+	0x00000000,
+/* 0x001c: unk_count */
+	0x00000000,
+/* 0x0020: unk_mask */
+	0x00000000,
+/* 0x0024: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
+
+uint32_t nvd7_grgpc_code[] = {
+	0x03180ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe047017,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x0c30e7f1,
+	0xbd50e3f0,
+	0xbd34bd24,
+/* 0x0371: init_unk_loop */
+	0x6821f444,
+	0xf400f6b0,
+	0xf7f00f0b,
+	0x04f2bb01,
+	0xb6054ffd,
+/* 0x0386: init_unk_next */
+	0x20b60130,
+	0x04e0b601,
+	0xf40126b0,
+/* 0x0392: init_unk_done */
+	0x0380e21b,
+	0x08048007,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x0e98003e,
+	0x030f9802,
+	0x014721f5,
+	0xfd070e98,
+	0x2ebb00ef,
+	0x003ebb00,
+	0x130040b7,
+	0xd00235b6,
+	0x25b60043,
+	0x0635b608,
+	0xb60120b6,
+	0x24b60130,
+	0x0834b608,
+	0xf5022fb9,
+	0xbb027121,
+	0x07f1003f,
+	0x03f00100,
+	0x0003d002,
+	0x24bd04bd,
+	0xf11f29f0,
+	0xf0080007,
+	0x02d00203,
+/* 0x0433: main */
+	0xf404bd00,
+	0x28f40031,
+	0x24d7f000,
+	0xf43921f4,
+	0xe4b0f401,
+	0x1e18f404,
+	0xf00181fe,
+	0x20bd0627,
+	0xb60412fd,
+	0x1efd01e4,
+	0x0018fe05,
+	0x04f721f5,
+/* 0x0463: main_not_ctx_xfer */
+	0x94d30ef4,
+	0xf5f010ef,
+	0xfe21f501,
+	0xc60ef402,
+/* 0x0470: ih */
+	0x88fe80f9,
+	0xf980f901,
+	0xf9a0f990,
+	0xf9d0f9b0,
+	0xbdf0f9e0,
+	0x800acf04,
+	0xf404abc4,
+	0xb7f11d0b,
+	0xd7f01900,
+	0x40becf24,
+	0xf400bfcf,
+	0xb0b70421,
+	0xe7f00400,
+	0x00bed001,
+/* 0x04a8: ih_no_fifo */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x04c3: hub_barrier_done */
+	0xf001f800,
+	0x0e9801f7,
+	0x04febb04,
+	0x9418e7f1,
+	0xf440e3f0,
+	0x00f88d21,
+/* 0x04d8: ctx_redswitch */
+	0x0614e7f1,
+	0xf006e4b6,
+	0xefd020f7,
+	0x08f7f000,
+/* 0x04e8: ctx_redswitch_delay */
+	0xf401f2b6,
+	0xf7f1fd1b,
+	0xefd00a20,
+/* 0x04f7: ctx_xfer */
+	0xf100f800,
+	0xb60a0417,
+	0x1fd00614,
+	0x0711f400,
+	0x04d821f5,
+/* 0x0508: ctx_xfer_not_load */
+	0x4afc17f1,
+	0xf00213f0,
+	0x12d00c27,
+	0x1521f500,
+	0xfc27f102,
+	0x0223f047,
+	0xf00020d0,
+	0x20b6012c,
+	0x0012d003,
+	0xf001acf0,
+	0xb7f002a5,
+	0x50b3f000,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x000c9800,
+	0xf0010d98,
+	0x21f500e7,
+	0xacf00166,
+	0x00b7f101,
+	0x50b3f040,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x010c9800,
+	0x98020d98,
+	0xe7f1060f,
+	0x21f50800,
+	0xacf00166,
+	0x04a5f001,
+	0x3000b7f1,
+	0x9850b3f0,
+	0xc4b6040c,
+	0x00bcbb0f,
+	0x98020c98,
+	0x0f98030d,
+	0x00e7f108,
+	0x6621f502,
+	0x1521f501,
+	0x0601f402,
+/* 0x05a3: ctx_xfer_post */
+	0xf11412f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00d,
+	0x021521f5,
+/* 0x05b4: ctx_xfer_done */
+	0x04c321f5,
+	0x000000f8,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc
index 62ab231..6b906cd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nve0 PGRAPH/GPC
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,437 +19,24 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 nve0_grgpc.fuc | envyas -a -w -m fuc -V nva3 -o nve0_grgpc.fuc.h
- */
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000001
 
-/* TODO
- * - bracket certain functions with scratch writes, useful for debugging
- * - watchdog timer around ctx operations
- */
+#define CHIPSET GK100
+#include "macros.fuc"
 
 .section #nve0_grgpc_data
-include(`nve0.fuc')
-gpc_id:			.b32 0
-gpc_mmio_list_head:	.b32 0
-gpc_mmio_list_tail:	.b32 0
-
-tpc_count:		.b32 0
-tpc_mask:		.b32 0
-tpc_mmio_list_head:	.b32 0
-tpc_mmio_list_tail:	.b32 0
-
-cmd_queue:		queue_init
-
-// chipset descriptions
-chipsets:
-.b8  0xe4 0 0 0
-.b16 #nve4_gpc_mmio_head
-.b16 #nve4_gpc_mmio_tail
-.b16 #nve4_tpc_mmio_head
-.b16 #nve4_tpc_mmio_tail
-.b8  0xe7 0 0 0
-.b16 #nve4_gpc_mmio_head
-.b16 #nve4_gpc_mmio_tail
-.b16 #nve4_tpc_mmio_head
-.b16 #nve4_tpc_mmio_tail
-.b8  0xe6 0 0 0
-.b16 #nve4_gpc_mmio_head
-.b16 #nve4_gpc_mmio_tail
-.b16 #nve4_tpc_mmio_head
-.b16 #nve4_tpc_mmio_tail
-.b8  0 0 0 0
-
-// GPC mmio lists
-nve4_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 2)
-mmctx_data(0x00040c, 3)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c40, 1)
-mmctx_data(0x000c6c, 1)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-mmctx_data(0x003024, 1)
-mmctx_data(0x0030c0, 2)
-mmctx_data(0x0030e4, 1)
-mmctx_data(0x003100, 6)
-mmctx_data(0x0031d0, 1)
-mmctx_data(0x0031e0, 2)
-nve4_gpc_mmio_tail:
-
-// TPC mmio lists
-nve4_tpc_mmio_head:
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x000230, 1)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 3)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 22)
-mmctx_data(0x0006ac, 2)
-mmctx_data(0x0006c8, 1)
-mmctx_data(0x000730, 8)
-mmctx_data(0x000758, 1)
-mmctx_data(0x000778, 1)
-nve4_tpc_mmio_tail:
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
 
 .section #nve0_grgpc_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nve0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nve0.fuc)
-//
-error:
-	push $r14
-	mov $r14 -0x67ec 	// 0x9814
-	sethi $r14 0x400000
-	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
-	add b32 $r14 0x41c
-	mov $r15 1
-	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
-	pop $r14
-	ret
-
-// GPC fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//   CC_SCRATCH[1]: context base
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: GPC context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
-
-	// enable fifo interrupt
-	mov $r2 4
-	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
-
-	// enable interrupts
-	bset $flags ie0
-
-	// figure out which GPC we are, and how many TPCs we have
-	mov $r1 0x608
-	shl b32 $r1 6
-	iord $r2 I[$r1 + 0x000]		// UNITS
-	mov $r3 1
-	and $r2 0x1f
-	shl b32 $r3 $r2
-	sub b32 $r3 1
-	st b32 D[$r0 + #tpc_count] $r2
-	st b32 D[$r0 + #tpc_mask] $r3
-	add b32 $r1 0x400
-	iord $r2 I[$r1 + 0x000]		// MYINDEX
-	st b32 D[$r0 + #gpc_id] $r2
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r1 #chipsets - 12
-	init_find_chipset:
-		add b32 $r1 12
-		ld b32 $r3 D[$r1 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// initialise context base, and size tracking
-	init_context:
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x100]	// CC_SCRATCH[1], initial base
-	clear b32 $r3		// track GPC context size here
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't currently ever change
-	mov $r4 0x700
-	shl b32 $r4 6
-	shr b32 $r5 $r2 8
-	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
-	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
-
-	// calculate GPC mmio context size, store the chipset-specific
-	// mmio list pointers somewhere we can get at them later without
-	// re-parsing the chipset list
-	clear b32 $r14
-	clear b32 $r15
-	ld b16 $r14 D[$r1 + 4]
-	ld b16 $r15 D[$r1 + 6]
-	st b16 D[$r0 + #gpc_mmio_list_head] $r14
-	st b16 D[$r0 + #gpc_mmio_list_tail] $r15
-	call #mmctx_size
-	add b32 $r2 $r15
-	add b32 $r3 $r15
-
-	// calculate per-TPC mmio context size, store the list pointers
-	ld b16 $r14 D[$r1 + 8]
-	ld b16 $r15 D[$r1 + 10]
-	st b16 D[$r0 + #tpc_mmio_list_head] $r14
-	st b16 D[$r0 + #tpc_mmio_list_tail] $r15
-	call #mmctx_size
-	ld b32 $r14 D[$r0 + #tpc_count]
-	mulu $r14 $r15
-	add b32 $r2 $r14
-	add b32 $r3 $r14
-
-	// round up base/size to 256 byte boundary (for strand SWBASE)
-	add b32 $r4 0x1300
-	shr b32 $r3 2
-	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
-	shr b32 $r2 8
-	shr b32 $r3 6
-	add b32 $r2 1
-	add b32 $r3 1
-	shl b32 $r2 8
-	shl b32 $r3 8
-
-	// calculate size of strand context data
-	mov b32 $r15 $r2
-	call #strand_ctx_init
-	add b32 $r3 $r15
-
-	// save context size, and tell HUB we're done
-	mov $r1 0x800
-	shl b32 $r1 6
-	iowr I[$r1 + 0x100] $r3		// CC_SCRATCH[1]  = context size
-	add b32 $r1 0x800
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// 0x0000-0x0003 are all context transfers
-	cmpu b32 $r14 0x04
-	bra nc #main_not_ctx_xfer
-		// fetch $flags and mask off $p1/$p2
-		mov $r1 $flags
-		mov $r2 0x0006
-		not b32 $r2
-		and $r1 $r2
-		// set $p1/$p2 according to transfer type
-		shl b32 $r14 1
-		or $r1 $r14
-		mov $flags $r1
-		// transfer context data
-		call #ctx_xfer
-		bra #main
-
-	main_not_ctx_xfer:
-	shl b32 $r15 $r14 16
-	or $r15 E_BAD_COMMAND
-	call #error
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// ack, and wake up main()
-	ih_no_fifo:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Set this GPC's bit in HUB_BAR, used to signal completion of various
-// activities to the HUB fuc
-//
-hub_barrier_done:
-	mov $r15 1
-	ld b32 $r14 D[$r0 + #gpc_id]
-	shl b32 $r15 $r14
-	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
-	sethi $r14 0x400000
-	call #nv_wr32
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x020
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0xa20
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
-	ret
-
-// Transfer GPC context data between GPU and storage area
-//
-// In: $r15 context base address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// set context base address
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r15// MEM_BASE
-	bra not $p1 #ctx_xfer_not_load
-		call #ctx_redswitch
-	ctx_xfer_not_load:
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 2		// first
-	mov $r11 0x0000
-	sethi $r11 0x500000
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
-	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// per-TPC mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 4		// last
-	mov $r11 0x4000
-	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
-	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
-	ld b32 $r15 D[$r0 + #tpc_mask]
-	mov $r14 0x800		// stride = 0x800
-	call #mmctx_xfer
-
-	// wait for strands to finish
-	call #strand_wait
-
-	// if load, or a save without a load following, do some
-	// unknown stuff that's done after finishing a block of
-	// strand commands
-	bra $p1 #ctx_xfer_post
-	bra not $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r1 0x4afc
-		sethi $r1 0x20000
-		mov $r2 0xd
-		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
-		call #strand_wait
-
-	// mark completion in HUB's barrier
-	ctx_xfer_done:
-	call #hub_barrier_done
-	ret
-
+#include "com.fuc"
+#include "gpc.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h
index 09ee470..7ff5ef6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h
@@ -1,19 +1,27 @@
 uint32_t nve0_grgpc_data[] = {
-/* 0x0000: gpc_id */
+/* 0x0000: gpc_mmio_list_head */
+	0x0000006c,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x0000006c,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x0000006c,
+/* 0x000c: unk_mmio_list_tail */
+	0x0000006c,
+/* 0x0010: gpc_id */
 	0x00000000,
-/* 0x0004: gpc_mmio_list_head */
+/* 0x0014: tpc_count */
 	0x00000000,
-/* 0x0008: gpc_mmio_list_tail */
+/* 0x0018: tpc_mask */
 	0x00000000,
-/* 0x000c: tpc_count */
+/* 0x001c: unk_count */
 	0x00000000,
-/* 0x0010: tpc_mask */
+/* 0x0020: unk_mask */
 	0x00000000,
-/* 0x0014: tpc_mmio_list_head */
+/* 0x0024: cmd_queue */
 	0x00000000,
-/* 0x0018: tpc_mmio_list_tail */
 	0x00000000,
-/* 0x001c: cmd_queue */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -30,84 +38,17 @@ uint32_t nve0_grgpc_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-/* 0x0064: chipsets */
-	0x000000e4,
-	0x0110008c,
-	0x01580110,
-	0x000000e7,
-	0x0110008c,
-	0x01580110,
-	0x000000e6,
-	0x0110008c,
-	0x01580110,
-	0x00000000,
-/* 0x008c: nve4_gpc_mmio_head */
-	0x00000380,
-	0x04000400,
-	0x0800040c,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c40,
-	0x00000c6c,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-	0x00003024,
-	0x040030c0,
-	0x000030e4,
-	0x14003100,
-	0x000031d0,
-	0x040031e0,
-/* 0x0110: nve4_gpc_mmio_tail */
-/* 0x0110: nve4_tpc_mmio_head */
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x00000230,
-	0x000002c4,
-	0x08000400,
-	0x08000420,
-	0x000004e8,
-	0x000004f4,
-	0x0c000604,
-	0x54000644,
-	0x040006ac,
-	0x000006c8,
-	0x1c000730,
-	0x00000758,
-	0x00000778,
 };
 
 uint32_t nve0_grgpc_code[] = {
-	0x03060ef5,
+	0x03180ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -139,7 +80,7 @@ uint32_t nve0_grgpc_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -161,63 +102,66 @@ uint32_t nve0_grgpc_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -225,8 +169,8 @@ uint32_t nve0_grgpc_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -235,76 +179,77 @@ uint32_t nve0_grgpc_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -313,150 +258,160 @@ uint32_t nve0_grgpc_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe3f09814,
-	0x8d21f440,
-	0x041ce0b7,
-	0xf401f7f0,
-	0xe0fc8d21,
-/* 0x0306: init */
-	0x04bd00f8,
-	0xf10004fe,
-	0xf0120017,
-	0x12d00227,
-	0x3e17f100,
-	0x0010fe04,
-	0x040017f1,
-	0xf0c010d0,
-	0x12d00427,
-	0x1031f400,
-	0x060817f1,
-	0xcf0614b6,
-	0x37f00012,
-	0x1f24f001,
-	0xb60432bb,
-	0x02800132,
-	0x04038003,
-	0x040010b7,
-	0x800012cf,
-	0x27f10002,
-	0x24b60800,
-	0x0022cf06,
-/* 0x035f: init_find_chipset */
-	0xb65817f0,
-	0x13980c10,
-	0x0432b800,
-	0xb00b0bf4,
-	0x1bf40034,
-/* 0x0373: init_context */
-	0xf100f8f1,
-	0xb6080027,
-	0x22cf0624,
-	0xf134bd40,
-	0xb6070047,
-	0x25950644,
-	0x0045d008,
-	0xbd4045d0,
-	0x58f4bde4,
-	0x1f58021e,
-	0x020e4003,
-	0xf5040f40,
-	0xbb013d21,
-	0x3fbb002f,
-	0x041e5800,
-	0x40051f58,
-	0x0f400a0e,
-	0x3d21f50c,
-	0x030e9801,
-	0xbb00effd,
-	0x3ebb002e,
-	0x0040b700,
-	0x0235b613,
-	0xb60043d0,
-	0x35b60825,
-	0x0120b606,
-	0xb60130b6,
-	0x34b60824,
-	0x022fb908,
-	0x026321f5,
-	0xf1003fbb,
-	0xb6080017,
-	0x13d00614,
-	0x0010b740,
-	0xf024bd08,
-	0x12d01f29,
-/* 0x0401: main */
-	0x0031f400,
-	0xf00028f4,
-	0x21f41cd7,
-	0xf401f439,
-	0xf404e4b0,
-	0x81fe1e18,
-	0x0627f001,
-	0x12fd20bd,
-	0x01e4b604,
-	0xfe051efd,
-	0x21f50018,
-	0x0ef404c3,
-/* 0x0431: main_not_ctx_xfer */
-	0x10ef94d3,
-	0xf501f5f0,
-	0xf402ec21,
-/* 0x043e: ih */
-	0x80f9c60e,
-	0xf90188fe,
-	0xf990f980,
-	0xf9b0f9a0,
-	0xf9e0f9d0,
-	0x800acff0,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe047017,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x0c30e7f1,
+	0xbd50e3f0,
+	0xbd34bd24,
+/* 0x0371: init_unk_loop */
+	0x6821f444,
+	0xf400f6b0,
+	0xf7f00f0b,
+	0x04f2bb01,
+	0xb6054ffd,
+/* 0x0386: init_unk_next */
+	0x20b60130,
+	0x04e0b601,
+	0xf40126b0,
+/* 0x0392: init_unk_done */
+	0x0380e21b,
+	0x08048007,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x0e98003e,
+	0x030f9802,
+	0x014721f5,
+	0xfd070e98,
+	0x2ebb00ef,
+	0x003ebb00,
+	0x130040b7,
+	0xd00235b6,
+	0x25b60043,
+	0x0635b608,
+	0xb60120b6,
+	0x24b60130,
+	0x0834b608,
+	0xf5022fb9,
+	0xbb027121,
+	0x07f1003f,
+	0x03f00100,
+	0x0003d002,
+	0x24bd04bd,
+	0xf11f29f0,
+	0xf0080007,
+	0x02d00203,
+/* 0x0433: main */
+	0xf404bd00,
+	0x28f40031,
+	0x24d7f000,
+	0xf43921f4,
+	0xe4b0f401,
+	0x1e18f404,
+	0xf00181fe,
+	0x20bd0627,
+	0xb60412fd,
+	0x1efd01e4,
+	0x0018fe05,
+	0x04f721f5,
+/* 0x0463: main_not_ctx_xfer */
+	0x94d30ef4,
+	0xf5f010ef,
+	0xfe21f501,
+	0xc60ef402,
+/* 0x0470: ih */
+	0x88fe80f9,
+	0xf980f901,
+	0xf9a0f990,
+	0xf9d0f9b0,
+	0xbdf0f9e0,
+	0x800acf04,
 	0xf404abc4,
 	0xb7f11d0b,
 	0xd7f01900,
-	0x40becf1c,
+	0x40becf24,
 	0xf400bfcf,
 	0xb0b70421,
 	0xe7f00400,
 	0x00bed001,
-/* 0x0474: ih_no_fifo */
+/* 0x04a8: ih_no_fifo */
 	0xfc400ad0,
 	0xfce0fcf0,
 	0xfcb0fcd0,
 	0xfc90fca0,
 	0x0088fe80,
 	0x32f480fc,
-/* 0x048f: hub_barrier_done */
+/* 0x04c3: hub_barrier_done */
 	0xf001f800,
 	0x0e9801f7,
-	0x04febb00,
+	0x04febb04,
 	0x9418e7f1,
 	0xf440e3f0,
 	0x00f88d21,
-/* 0x04a4: ctx_redswitch */
+/* 0x04d8: ctx_redswitch */
 	0x0614e7f1,
 	0xf006e4b6,
 	0xefd020f7,
 	0x08f7f000,
-/* 0x04b4: ctx_redswitch_delay */
+/* 0x04e8: ctx_redswitch_delay */
 	0xf401f2b6,
 	0xf7f1fd1b,
 	0xefd00a20,
-/* 0x04c3: ctx_xfer */
+/* 0x04f7: ctx_xfer */
 	0xf100f800,
 	0xb60a0417,
 	0x1fd00614,
 	0x0711f400,
-	0x04a421f5,
-/* 0x04d4: ctx_xfer_not_load */
+	0x04d821f5,
+/* 0x0508: ctx_xfer_not_load */
 	0x4afc17f1,
 	0xf00213f0,
 	0x12d00c27,
-	0x0721f500,
+	0x1521f500,
 	0xfc27f102,
 	0x0223f047,
 	0xf00020d0,
@@ -465,31 +420,40 @@ uint32_t nve0_grgpc_code[] = {
 	0xf001acf0,
 	0xb7f002a5,
 	0x50b3f000,
-	0xb6000c98,
+	0xb6040c98,
 	0xbcbb0fc4,
-	0x010c9800,
-	0xf0020d98,
+	0x000c9800,
+	0xf0010d98,
 	0x21f500e7,
-	0xacf0015c,
+	0xacf00166,
+	0x00b7f101,
+	0x50b3f040,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x010c9800,
+	0x98020d98,
+	0xe7f1060f,
+	0x21f50800,
+	0xacf00166,
 	0x04a5f001,
-	0x4000b7f1,
+	0x3000b7f1,
 	0x9850b3f0,
-	0xc4b6000c,
+	0xc4b6040c,
 	0x00bcbb0f,
-	0x98050c98,
-	0x0f98060d,
-	0x00e7f104,
-	0x5c21f508,
-	0x0721f501,
+	0x98020c98,
+	0x0f98030d,
+	0x00e7f108,
+	0x6621f502,
+	0x1521f501,
 	0x0601f402,
-/* 0x054b: ctx_xfer_post */
+/* 0x05a3: ctx_xfer_post */
 	0xf11412f4,
 	0xf04afc17,
 	0x27f00213,
 	0x0012d00d,
-	0x020721f5,
-/* 0x055c: ctx_xfer_done */
-	0x048f21f5,
+	0x021521f5,
+/* 0x05b4: ctx_xfer_done */
+	0x04c321f5,
 	0x000000f8,
 	0x00000000,
 	0x00000000,
@@ -508,26 +472,4 @@ uint32_t nve0_grgpc_code[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc
new file mode 100644
index 0000000..90bbe52
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000002
+
+#define CHIPSET GK110
+#include "macros.fuc"
+
+.section #nvf0_grgpc_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
+
+.section #nvf0_grgpc_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "gpc.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc.h
new file mode 100644
index 0000000..f870507
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc.h
@@ -0,0 +1,475 @@
+uint32_t nvf0_grgpc_data[] = {
+/* 0x0000: gpc_mmio_list_head */
+	0x0000006c,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x0000006c,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x0000006c,
+/* 0x000c: unk_mmio_list_tail */
+	0x0000006c,
+/* 0x0010: gpc_id */
+	0x00000000,
+/* 0x0014: tpc_count */
+	0x00000000,
+/* 0x0018: tpc_mask */
+	0x00000000,
+/* 0x001c: unk_count */
+	0x00000000,
+/* 0x0020: unk_mask */
+	0x00000000,
+/* 0x0024: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
+
+uint32_t nvf0_grgpc_code[] = {
+	0x03180ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf0370007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f03700,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe047017,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x0c30e7f1,
+	0xbd50e3f0,
+	0xbd34bd24,
+/* 0x0371: init_unk_loop */
+	0x6821f444,
+	0xf400f6b0,
+	0xf7f00f0b,
+	0x04f2bb01,
+	0xb6054ffd,
+/* 0x0386: init_unk_next */
+	0x20b60130,
+	0x04e0b601,
+	0xf40226b0,
+/* 0x0392: init_unk_done */
+	0x0380e21b,
+	0x08048007,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x0e98003e,
+	0x030f9802,
+	0x014721f5,
+	0xfd070e98,
+	0x2ebb00ef,
+	0x003ebb00,
+	0x130040b7,
+	0xd00235b6,
+	0x25b60043,
+	0x0635b608,
+	0xb60120b6,
+	0x24b60130,
+	0x0834b608,
+	0xf5022fb9,
+	0xbb027121,
+	0x07f1003f,
+	0x03f00100,
+	0x0003d002,
+	0x24bd04bd,
+	0xf11f29f0,
+	0xf0300007,
+	0x02d00203,
+/* 0x0433: main */
+	0xf404bd00,
+	0x28f40031,
+	0x24d7f000,
+	0xf43921f4,
+	0xe4b0f401,
+	0x1e18f404,
+	0xf00181fe,
+	0x20bd0627,
+	0xb60412fd,
+	0x1efd01e4,
+	0x0018fe05,
+	0x04f721f5,
+/* 0x0463: main_not_ctx_xfer */
+	0x94d30ef4,
+	0xf5f010ef,
+	0xfe21f501,
+	0xc60ef402,
+/* 0x0470: ih */
+	0x88fe80f9,
+	0xf980f901,
+	0xf9a0f990,
+	0xf9d0f9b0,
+	0xbdf0f9e0,
+	0x800acf04,
+	0xf404abc4,
+	0xb7f11d0b,
+	0xd7f01900,
+	0x40becf24,
+	0xf400bfcf,
+	0xb0b70421,
+	0xe7f00400,
+	0x00bed001,
+/* 0x04a8: ih_no_fifo */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x04c3: hub_barrier_done */
+	0xf001f800,
+	0x0e9801f7,
+	0x04febb04,
+	0x9418e7f1,
+	0xf440e3f0,
+	0x00f88d21,
+/* 0x04d8: ctx_redswitch */
+	0x0614e7f1,
+	0xf006e4b6,
+	0xefd020f7,
+	0x08f7f000,
+/* 0x04e8: ctx_redswitch_delay */
+	0xf401f2b6,
+	0xf7f1fd1b,
+	0xefd00a20,
+/* 0x04f7: ctx_xfer */
+	0xf100f800,
+	0xb60a0417,
+	0x1fd00614,
+	0x0711f400,
+	0x04d821f5,
+/* 0x0508: ctx_xfer_not_load */
+	0x4afc17f1,
+	0xf00213f0,
+	0x12d00c27,
+	0x1521f500,
+	0xfc27f102,
+	0x0223f047,
+	0xf00020d0,
+	0x20b6012c,
+	0x0012d003,
+	0xf001acf0,
+	0xb7f002a5,
+	0x50b3f000,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x000c9800,
+	0xf0010d98,
+	0x21f500e7,
+	0xacf00166,
+	0x00b7f101,
+	0x50b3f040,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x010c9800,
+	0x98020d98,
+	0xe7f1060f,
+	0x21f50800,
+	0xacf00166,
+	0x04a5f001,
+	0x3000b7f1,
+	0x9850b3f0,
+	0xc4b6040c,
+	0x00bcbb0f,
+	0x98020c98,
+	0x0f98030d,
+	0x00e7f108,
+	0x6621f502,
+	0x1521f501,
+	0x0601f402,
+/* 0x05a3: ctx_xfer_post */
+	0xf11412f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00d,
+	0x021521f5,
+/* 0x05b4: ctx_xfer_done */
+	0x04c321f5,
+	0x000000f8,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hub.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hub.fuc
new file mode 100644
index 0000000..b82d2ae
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hub.fuc
@@ -0,0 +1,724 @@
+/* fuc microcode for nvc0 PGRAPH/HUB
+ *
+ * Copyright 2011 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#ifdef INCLUDE_DATA
+hub_mmio_list_head:	.b32 #hub_mmio_list_base
+hub_mmio_list_tail:	.b32 #hub_mmio_list_next
+
+gpc_count:		.b32 0
+rop_count:		.b32 0
+cmd_queue:		queue_init
+
+ctx_current:		.b32 0
+
+.align 256
+chan_data:
+chan_mmio_count:	.b32 0
+chan_mmio_address:	.b32 0
+
+.align 256
+xfer_data: 		.skip 256
+
+hub_mmio_list_base:
+.b32 0x0417e91c // 0x17e91c, 2
+hub_mmio_list_next:
+#endif
+
+#ifdef INCLUDE_CODE
+// reports an exception to the host
+//
+// In: $r15 error code (see nvc0.fuc)
+//
+error:
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_VAL(5), 0, $r15)
+	mov $r15 1
+	nv_iowr(NV_PGRAPH_FECS_INTR_UP_SET, 0, $r15)
+	ret
+
+// HUB fuc initialisation, executed by triggering ucode start, will
+// fall through to main loop after completion.
+//
+// Output:
+//   CC_SCRATCH[0]:
+//	     31:31: set to signal completion
+//   CC_SCRATCH[1]:
+//	      31:0: total PGRAPH context size
+//
+init:
+	clear b32 $r0
+	mov $sp $r0
+	mov $xdbase $r0
+
+	// enable fifo access
+	mov $r1 0x1200
+	mov $r2 2
+	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
+
+	// setup i0 handler, and route all interrupts to it
+	mov $r1 #ih
+	mov $iv0 $r1
+	mov $r1 0x400
+	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
+
+	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
+	mov $r3 0x404
+	shl b32 $r3 6
+	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
+	iowr I[$r3 + 0x000] $r2
+
+	// not sure what these are, route them because NVIDIA does, and
+	// the IRQ handler will signal the host if we ever get one.. we
+	// may find out if/why we need to handle these if so..
+	//
+	mov $r2 0x2004
+	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
+	mov $r2 0x200b
+	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
+	mov $r2 0x200c
+	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
+
+	// enable all INTR_UP interrupts
+	mov $r2 0xc24
+	shl b32 $r2 6
+	not b32 $r3 $r0
+	iowr I[$r2] $r3
+
+	// enable fifo, ctxsw, 9, 10, 15 interrupts
+	mov $r2 -0x78fc		// 0x8704
+	sethi $r2 0
+	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
+
+	// fifo level triggered, rest edge
+	sub b32 $r1 0x100
+	mov $r2 4
+	iowr I[$r1] $r2
+
+	// enable interrupts
+	bset $flags ie0
+
+	// fetch enabled GPC/ROP counts
+	mov $r14 -0x69fc	// 0x409604
+	sethi $r14 0x400000
+	call #nv_rd32
+	extr $r1 $r15 16:20
+	st b32 D[$r0 + #rop_count] $r1
+	and $r15 0x1f
+	st b32 D[$r0 + #gpc_count] $r15
+
+	// set BAR_REQMASK to GPC mask
+	mov $r1 1
+	shl b32 $r1 $r15
+	sub b32 $r1 1
+	mov $r2 0x40c
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r1
+	iowr I[$r2 + 0x100] $r1
+
+	// context size calculation, reserve first 256 bytes for use by fuc
+	mov $r1 256
+
+	// calculate size of mmio context data
+	ld b32 $r14 D[$r0 + #hub_mmio_list_head]
+	ld b32 $r15 D[$r0 + #hub_mmio_list_tail]
+	call #mmctx_size
+
+	// set mmctx base addresses now so we don't have to do it later,
+	// they don't (currently) ever change
+	mov $r3 0x700
+	shl b32 $r3 6
+	shr b32 $r4 $r1 8
+	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
+	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
+	add b32 $r3 0x1300
+	add b32 $r1 $r15
+	shr b32 $r15 2
+	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
+
+	// strands, base offset needs to be aligned to 256 bytes
+	shr b32 $r1 8
+	add b32 $r1 1
+	shl b32 $r1 8
+	mov b32 $r15 $r1
+	call #strand_ctx_init
+	add b32 $r1 $r15
+
+	// initialise each GPC in sequence by passing in the offset of its
+	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
+	// has previously been uploaded by the host) running.
+	//
+	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
+	// when it has completed, and return the size of its context data
+	// in GPCn_CC_SCRATCH[1]
+	//
+	ld b32 $r3 D[$r0 + #gpc_count]
+	mov $r4 0x2000
+	sethi $r4 0x500000
+	init_gpc:
+		// setup, and start GPC ucode running
+		add b32 $r14 $r4 0x804
+		mov b32 $r15 $r1
+		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
+		add b32 $r14 $r4 0x10c
+		clear b32 $r15
+		call #nv_wr32
+		add b32 $r14 $r4 0x104
+		call #nv_wr32			// ENTRY
+		add b32 $r14 $r4 0x100
+		mov $r15 2			// CTRL_START_TRIGGER
+		call #nv_wr32			// CTRL
+
+		// wait for it to complete, and adjust context size
+		add b32 $r14 $r4 0x800
+		init_gpc_wait:
+			call #nv_rd32
+			xbit $r15 $r15 31
+			bra e #init_gpc_wait
+		add b32 $r14 $r4 0x804
+		call #nv_rd32
+		add b32 $r1 $r15
+
+		// next!
+		add b32 $r4 0x8000
+		sub b32 $r3 1
+		bra ne #init_gpc
+
+	// save context size, and tell host we're ready
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_VAL(1), 0, $r1)
+	clear b32 $r1
+	bset $r1 31
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_SET(0), 0, $r1)
+
+// Main program loop, very simple, sleeps until woken up by the interrupt
+// handler, pulls a command from the queue and executes its handler
+//
+main:
+	// sleep until we have something to do
+	bset $flags $p0
+	sleep $p0
+	mov $r13 #cmd_queue
+	call #queue_get
+	bra $p1 #main
+
+	// context switch, requested by GPU?
+	cmpu b32 $r14 0x4001
+	bra ne #main_not_ctx_switch
+		trace_set(T_AUTO)
+		mov $r1 0xb00
+		shl b32 $r1 6
+		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
+		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
+
+		xbit $r3 $r1 31
+		bra e #chsw_no_prev
+			xbit $r3 $r2 31
+			bra e #chsw_prev_no_next
+				push $r2
+				mov b32 $r2 $r1
+				trace_set(T_SAVE)
+				bclr $flags $p1
+				bset $flags $p2
+				call #ctx_xfer
+				trace_clr(T_SAVE);
+				pop $r2
+				trace_set(T_LOAD);
+				bset $flags $p1
+				call #ctx_xfer
+				trace_clr(T_LOAD);
+				bra #chsw_done
+			chsw_prev_no_next:
+				push $r2
+				mov b32 $r2 $r1
+				bclr $flags $p1
+				bclr $flags $p2
+				call #ctx_xfer
+				pop $r2
+				mov $r1 0xb00
+				shl b32 $r1 6
+				iowr I[$r1] $r2
+				bra #chsw_done
+		chsw_no_prev:
+			xbit $r3 $r2 31
+			bra e #chsw_done
+				bset $flags $p1
+				bclr $flags $p2
+				call #ctx_xfer
+
+		// ack the context switch request
+		chsw_done:
+		mov $r1 0xb0c
+		shl b32 $r1 6
+		mov $r2 1
+		iowr I[$r1 + 0x000] $r2		// 0x409b0c
+		trace_clr(T_AUTO)
+		bra #main
+
+	// request to set current channel? (*not* a context switch)
+	main_not_ctx_switch:
+	cmpu b32 $r14 0x0001
+	bra ne #main_not_ctx_chan
+		mov b32 $r2 $r15
+		call #ctx_chan
+		bra #main_done
+
+	// request to store current channel context?
+	main_not_ctx_chan:
+	cmpu b32 $r14 0x0002
+	bra ne #main_not_ctx_save
+		trace_set(T_SAVE)
+		bclr $flags $p1
+		bclr $flags $p2
+		call #ctx_xfer
+		trace_clr(T_SAVE)
+		bra #main_done
+
+	main_not_ctx_save:
+		shl b32 $r15 $r14 16
+		or $r15 E_BAD_COMMAND
+		call #error
+		bra #main
+
+	main_done:
+	clear b32 $r2
+	bset $r2 31
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_SET(0), 0, $r2)
+	bra #main
+
+// interrupt handler
+ih:
+	push $r8
+	mov $r8 $flags
+	push $r8
+	push $r9
+	push $r10
+	push $r11
+	push $r13
+	push $r14
+	push $r15
+	clear b32 $r0
+
+	// incoming fifo command?
+	iord $r10 I[$r0 + 0x200]	// INTR
+	and $r11 $r10 0x00000004
+	bra e #ih_no_fifo
+		// queue incoming fifo command for later processing
+		mov $r11 0x1900
+		mov $r13 #cmd_queue
+		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
+		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
+		call #queue_put
+		add b32 $r11 0x400
+		mov $r14 1
+		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
+
+	// context switch request?
+	ih_no_fifo:
+	and $r11 $r10 0x00000100
+	bra e #ih_no_ctxsw
+		// enqueue a context switch for later processing
+		mov $r13 #cmd_queue
+		mov $r14 0x4001
+		call #queue_put
+
+	// anything we didn't handle, bring it to the host's attention
+	ih_no_ctxsw:
+	mov $r11 0x104
+	not b32 $r11
+	and $r11 $r10 $r11
+	bra e #ih_no_other
+		mov $r10 0xc1c
+		shl b32 $r10 6
+		iowr I[$r10] $r11	// INTR_UP_SET
+
+	// ack, and wake up main()
+	ih_no_other:
+	iowr I[$r0 + 0x100] $r10	// INTR_ACK
+
+	pop $r15
+	pop $r14
+	pop $r13
+	pop $r11
+	pop $r10
+	pop $r9
+	pop $r8
+	mov $flags $r8
+	pop $r8
+	bclr $flags $p0
+	iret
+
+#if CHIPSET < GK100
+// Not real sure, but, MEM_CMD 7 will hang forever if this isn't done
+ctx_4160s:
+	mov $r14 0x4160
+	sethi $r14 0x400000
+	mov $r15 1
+	call #nv_wr32
+	ctx_4160s_wait:
+		call #nv_rd32
+		xbit $r15 $r15 4
+		bra e #ctx_4160s_wait
+	ret
+
+// Without clearing again at end of xfer, some things cause PGRAPH
+// to hang with STATUS=0x00000007 until it's cleared.. fbcon can
+// still function with it set however...
+ctx_4160c:
+	mov $r14 0x4160
+	sethi $r14 0x400000
+	clear b32 $r15
+	call #nv_wr32
+	ret
+#endif
+
+// Again, not real sure
+//
+// In: $r15 value to set 0x404170 to
+//
+ctx_4170s:
+	mov $r14 0x4170
+	sethi $r14 0x400000
+	or $r15 0x10
+	call #nv_wr32
+	ret
+
+// Waits for a ctx_4170s() call to complete
+//
+ctx_4170w:
+	mov $r14 0x4170
+	sethi $r14 0x400000
+	call #nv_rd32
+	and $r15 0x10
+	bra ne #ctx_4170w
+	ret
+
+// Disables various things, waits a bit, and re-enables them..
+//
+// Not sure how exactly this helps, perhaps "ENABLE" is not such a
+// good description for the bits we turn off?  Anyways, without this,
+// funny things happen.
+//
+ctx_redswitch:
+	mov $r14 0x614
+	shl b32 $r14 6
+	mov $r15 0x270
+	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
+	mov $r15 8
+	ctx_redswitch_delay:
+		sub b32 $r15 1
+		bra ne #ctx_redswitch_delay
+	mov $r15 0x770
+	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
+	ret
+
+// Not a clue what this is for, except that unless the value is 0x10, the
+// strand context is saved (and presumably restored) incorrectly..
+//
+// In: $r15 value to set to (0x00/0x10 are used)
+//
+ctx_86c:
+	mov $r14 0x86c
+	shl b32 $r14 6
+	iowr I[$r14] $r15	// HUB(0x86c) = val
+	mov $r14 -0x75ec
+	sethi $r14 0x400000
+	call #nv_wr32		// ROP(0xa14) = val
+	mov $r14 -0x5794
+	sethi $r14 0x410000
+	call #nv_wr32		// GPC(0x86c) = val
+	ret
+
+// ctx_load - load's a channel's ctxctl data, and selects its vm
+//
+// In: $r2 channel address
+//
+ctx_load:
+	trace_set(T_CHAN)
+
+	// switch to channel, somewhat magic in parts..
+	mov $r10 12		// DONE_UNK12
+	call #wait_donez
+	mov $r1 0xa24
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r0	// 0x409a24
+	mov $r3 0xb00
+	shl b32 $r3 6
+	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
+	mov $r1 0xa0c
+	shl b32 $r1 6
+	mov $r4 7
+	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
+	iowr I[$r1 + 0x100] $r4	// MEM_CMD
+	ctx_chan_wait_0:
+		iord $r4 I[$r1 + 0x100]
+		and $r4 0x1f
+		bra ne #ctx_chan_wait_0
+	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
+
+	// load channel header, fetch PGRAPH context pointer
+	mov $xtargets $r0
+	bclr $r2 31
+	shl b32 $r2 4
+	add b32 $r2 2
+
+	trace_set(T_LCHAN)
+	mov $r1 0xa04
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r2		// MEM_BASE
+	mov $r1 0xa20
+	shl b32 $r1 6
+	mov $r2 0x0002
+	sethi $r2 0x80000000
+	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
+	mov $r1 0x10			// chan + 0x0210
+	mov $r2 #xfer_data
+	sethi $r2 0x00020000		// 16 bytes
+	xdld $r1 $r2
+	xdwait
+	trace_clr(T_LCHAN)
+
+	// update current context
+	ld b32 $r1 D[$r0 + #xfer_data + 4]
+	shl b32 $r1 24
+	ld b32 $r2 D[$r0 + #xfer_data + 0]
+	shr b32 $r2 8
+	or $r1 $r2
+	st b32 D[$r0 + #ctx_current] $r1
+
+	// set transfer base to start of context, and fetch context header
+	trace_set(T_LCTXH)
+	mov $r2 0xa04
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r1		// MEM_BASE
+	mov $r2 1
+	mov $r1 0xa20
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
+	mov $r1 #chan_data
+	sethi $r1 0x00060000		// 256 bytes
+	xdld $r0 $r1
+	xdwait
+	trace_clr(T_LCTXH)
+
+	trace_clr(T_CHAN)
+	ret
+
+// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
+//            the active channel for ctxctl, but not actually transfer
+//            any context data.  intended for use only during initial
+//            context construction.
+//
+// In: $r2 channel address
+//
+ctx_chan:
+#if CHIPSET < GK100
+	call #ctx_4160s
+#endif
+	call #ctx_load
+	mov $r10 12			// DONE_UNK12
+	call #wait_donez
+	mov $r1 0xa10
+	shl b32 $r1 6
+	mov $r2 5
+	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
+	ctx_chan_wait:
+		iord $r2 I[$r1 + 0x000]
+		or $r2 $r2
+		bra ne #ctx_chan_wait
+#if CHIPSET < GK100
+	call #ctx_4160c
+#endif
+	ret
+
+// Execute per-context state overrides list
+//
+// Only executed on the first load of a channel.  Might want to look into
+// removing this and having the host directly modify the channel's context
+// to change this state...  The nouveau DRM already builds this list as
+// it's definitely needed for NVIDIA's, so we may as well use it for now
+//
+// Input: $r1 mmio list length
+//
+ctx_mmio_exec:
+	// set transfer base to be the mmio list
+	ld b32 $r3 D[$r0 + #chan_mmio_address]
+	mov $r2 0xa04
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r3		// MEM_BASE
+
+	clear b32 $r3
+	ctx_mmio_loop:
+		// fetch next 256 bytes of mmio list if necessary
+		and $r4 $r3 0xff
+		bra ne #ctx_mmio_pull
+			mov $r5 #xfer_data
+			sethi $r5 0x00060000	// 256 bytes
+			xdld $r3 $r5
+			xdwait
+
+		// execute a single list entry
+		ctx_mmio_pull:
+		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
+		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
+		call #nv_wr32
+
+		// next!
+		add b32 $r3 8
+		sub b32 $r1 1
+		bra ne #ctx_mmio_loop
+
+	// set transfer base back to the current context
+	ctx_mmio_done:
+	ld b32 $r3 D[$r0 + #ctx_current]
+	iowr I[$r2 + 0x000] $r3		// MEM_BASE
+
+	// disable the mmio list now, we don't need/want to execute it again
+	st b32 D[$r0 + #chan_mmio_count] $r0
+	mov $r1 #chan_data
+	sethi $r1 0x00060000		// 256 bytes
+	xdst $r0 $r1
+	xdwait
+	ret
+
+// Transfer HUB context data between GPU and storage area
+//
+// In: $r2 channel address
+//     $p1 clear on save, set on load
+//     $p2 set if opposite direction done/will be done, so:
+//		on save it means: "a load will follow this save"
+//		on load it means: "a save preceeded this load"
+//
+ctx_xfer:
+	// according to mwk, some kind of wait for idle
+	mov $r15 0xc00
+	shl b32 $r15 6
+	mov $r14 4
+	iowr I[$r15 + 0x200] $r14
+	ctx_xfer_idle:
+		iord $r14 I[$r15 + 0x000]
+		and $r14 0x2000
+		bra ne #ctx_xfer_idle
+
+	bra not $p1 #ctx_xfer_pre
+	bra $p2 #ctx_xfer_pre_load
+	ctx_xfer_pre:
+		mov $r15 0x10
+		call #ctx_86c
+#if CHIPSET < GK100
+		call #ctx_4160s
+#endif
+		bra not $p1 #ctx_xfer_exec
+
+	ctx_xfer_pre_load:
+		mov $r15 2
+		call #ctx_4170s
+		call #ctx_4170w
+		call #ctx_redswitch
+		clear b32 $r15
+		call #ctx_4170s
+		call #ctx_load
+
+	// fetch context pointer, and initiate xfer on all GPCs
+	ctx_xfer_exec:
+	ld b32 $r1 D[$r0 + #ctx_current]
+	mov $r2 0x414
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
+	mov $r14 -0x5b00
+	sethi $r14 0x410000
+	mov b32 $r15 $r1
+	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
+	add b32 $r14 4
+	xbit $r15 $flags $p1
+	xbit $r2 $flags $p2
+	shl b32 $r2 1
+	or $r15 $r2
+	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
+
+	// strands
+	mov $r1 0x4afc
+	sethi $r1 0x20000
+	mov $r2 0xc
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
+	call #strand_wait
+	mov $r2 0x47fc
+	sethi $r2 0x20000
+	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
+	xbit $r2 $flags $p1
+	add b32 $r2 3
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
+
+	// mmio context
+	xbit $r10 $flags $p1	// direction
+	or $r10 6		// first, last
+	mov $r11 0		// base = 0
+	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
+	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
+	mov $r14 0		// not multi
+	call #mmctx_xfer
+
+	// wait for GPCs to all complete
+	mov $r10 8		// DONE_BAR
+	call #wait_doneo
+
+	// wait for strand xfer to complete
+	call #strand_wait
+
+	// post-op
+	bra $p1 #ctx_xfer_post
+		mov $r10 12		// DONE_UNK12
+		call #wait_donez
+		mov $r1 0xa10
+		shl b32 $r1 6
+		mov $r2 5
+		iowr I[$r1] $r2		// MEM_CMD
+		ctx_xfer_post_save_wait:
+			iord $r2 I[$r1]
+			or $r2 $r2
+			bra ne #ctx_xfer_post_save_wait
+
+	bra $p2 #ctx_xfer_done
+	ctx_xfer_post:
+		mov $r15 2
+		call #ctx_4170s
+		clear b32 $r15
+		call #ctx_86c
+		call #strand_post
+		call #ctx_4170w
+		clear b32 $r15
+		call #ctx_4170s
+
+		bra not $p1 #ctx_xfer_no_post_mmio
+		ld b32 $r1 D[$r0 + #chan_mmio_count]
+		or $r1 $r1
+		bra e #ctx_xfer_no_post_mmio
+			call #ctx_mmio_exec
+
+		ctx_xfer_no_post_mmio:
+#if CHIPSET < GK100
+		call #ctx_4160c
+#endif
+
+	ctx_xfer_done:
+	ret
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
index 7fbdebb..3ff52ba 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nvc0 PGRAPH/HUB
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,850 +19,22 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 hubnvc0.fuc | envyas -a -w -m fuc -V fuc3 -o hubnvc0.fuc.h
- */
+#define CHIPSET GF100
+#include "macros.fuc"
 
 .section #nvc0_grhub_data
-include(`nvc0.fuc')
-gpc_count:		.b32 0
-rop_count:		.b32 0
-cmd_queue:		queue_init
-hub_mmio_list_head:	.b32 0
-hub_mmio_list_tail:	.b32 0
-
-ctx_current:		.b32 0
-
-chipsets:
-.b8  0xc0 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc1 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc1_hub_mmio_tail
-.b8  0xc3 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc4 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc8 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xce 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xcf 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xd9 0 0 0
-.b16 #nvd9_hub_mmio_head
-.b16 #nvd9_hub_mmio_tail
-.b8  0xd7 0 0 0
-.b16 #nvd9_hub_mmio_head
-.b16 #nvd9_hub_mmio_tail
-.b8  0 0 0 0
-
-nvc0_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404004, 11)
-mmctx_data(0x404044, 1)
-mmctx_data(0x404094, 14)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 2)
-mmctx_data(0x404174, 3)
-mmctx_data(0x404200, 8)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 32)
-mmctx_data(0x404698, 21)
-mmctx_data(0x4046f0, 2)
-mmctx_data(0x404700, 22)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 2)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408900, 4)
-mmctx_data(0x408980, 1)
-nvc0_hub_mmio_tail:
-mmctx_data(0x4064c0, 2)
-nvc1_hub_mmio_tail:
-
-nvd9_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404004, 10)
-mmctx_data(0x404044, 1)
-mmctx_data(0x404094, 14)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 2)
-mmctx_data(0x404178, 2)
-mmctx_data(0x404200, 8)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 32)
-mmctx_data(0x404698, 21)
-mmctx_data(0x4046f0, 2)
-mmctx_data(0x404700, 22)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 5)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408900, 4)
-mmctx_data(0x408980, 1)
-nvd9_hub_mmio_tail:
-
-.align 256
-chan_data:
-chan_mmio_count:	.b32 0
-chan_mmio_address:	.b32 0
-
-.align 256
-xfer_data: 		.b32 0
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
 
 .section #nvc0_grhub_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nvc0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nvc0.fuc)
-//
-error:
-	push $r14
-	mov $r14 0x814
-	shl b32 $r14 6
-	iowr I[$r14 + 0x000] $r15	// CC_SCRATCH[5] = error code
-	mov $r14 0xc1c
-	shl b32 $r14 6
-	mov $r15 1
-	iowr I[$r14 + 0x000] $r15	// INTR_UP_SET
-	pop $r14
-	ret
-
-// HUB fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: total PGRAPH context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-	mov $xdbase $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
-
-	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
-	mov $r3 0x404
-	shl b32 $r3 6
-	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
-	iowr I[$r3 + 0x000] $r2
-
-	// not sure what these are, route them because NVIDIA does, and
-	// the IRQ handler will signal the host if we ever get one.. we
-	// may find out if/why we need to handle these if so..
-	//
-	mov $r2 0x2004
-	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
-	mov $r2 0x200b
-	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
-	mov $r2 0x200c
-	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
-
-	// enable all INTR_UP interrupts
-	mov $r2 0xc24
-	shl b32 $r2 6
-	not b32 $r3 $r0
-	iowr I[$r2] $r3
-
-	// enable fifo, ctxsw, 9, 10, 15 interrupts
-	mov $r2 -0x78fc		// 0x8704
-	sethi $r2 0
-	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
-
-	// fifo level triggered, rest edge
-	sub b32 $r1 0x100
-	mov $r2 4
-	iowr I[$r1] $r2
-
-	// enable interrupts
-	bset $flags ie0
-
-	// fetch enabled GPC/ROP counts
-	mov $r14 -0x69fc	// 0x409604
-	sethi $r14 0x400000
-	call #nv_rd32
-	extr $r1 $r15 16:20
-	st b32 D[$r0 + #rop_count] $r1
-	and $r15 0x1f
-	st b32 D[$r0 + #gpc_count] $r15
-
-	// set BAR_REQMASK to GPC mask
-	mov $r1 1
-	shl b32 $r1 $r15
-	sub b32 $r1 1
-	mov $r2 0x40c
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1
-	iowr I[$r2 + 0x100] $r1
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r15 #chipsets - 8
-	init_find_chipset:
-		add b32 $r15 8
-		ld b32 $r3 D[$r15 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// context size calculation, reserve first 256 bytes for use by fuc
-	init_context:
-	mov $r1 256
-
-	// calculate size of mmio context data
-	ld b16 $r14 D[$r15 + 4]
-	ld b16 $r15 D[$r15 + 6]
-	sethi $r14 0
-	st b32 D[$r0 + #hub_mmio_list_head] $r14
-	st b32 D[$r0 + #hub_mmio_list_tail] $r15
-	call #mmctx_size
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't (currently) ever change
-	mov $r3 0x700
-	shl b32 $r3 6
-	shr b32 $r4 $r1 8
-	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
-	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
-	add b32 $r3 0x1300
-	add b32 $r1 $r15
-	shr b32 $r15 2
-	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
-
-	// strands, base offset needs to be aligned to 256 bytes
-	shr b32 $r1 8
-	add b32 $r1 1
-	shl b32 $r1 8
-	mov b32 $r15 $r1
-	call #strand_ctx_init
-	add b32 $r1 $r15
-
-	// initialise each GPC in sequence by passing in the offset of its
-	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
-	// has previously been uploaded by the host) running.
-	//
-	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
-	// when it has completed, and return the size of its context data
-	// in GPCn_CC_SCRATCH[1]
-	//
-	ld b32 $r3 D[$r0 + #gpc_count]
-	mov $r4 0x2000
-	sethi $r4 0x500000
-	init_gpc:
-		// setup, and start GPC ucode running
-		add b32 $r14 $r4 0x804
-		mov b32 $r15 $r1
-		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
-		add b32 $r14 $r4 0x800
-		mov b32 $r15 $r2
-		call #nv_wr32			// CC_SCRATCH[0] = chipset
-		add b32 $r14 $r4 0x10c
-		clear b32 $r15
-		call #nv_wr32
-		add b32 $r14 $r4 0x104
-		call #nv_wr32			// ENTRY
-		add b32 $r14 $r4 0x100
-		mov $r15 2			// CTRL_START_TRIGGER
-		call #nv_wr32			// CTRL
-
-		// wait for it to complete, and adjust context size
-		add b32 $r14 $r4 0x800
-		init_gpc_wait:
-			call #nv_rd32
-			xbit $r15 $r15 31
-			bra e #init_gpc_wait
-		add b32 $r14 $r4 0x804
-		call #nv_rd32
-		add b32 $r1 $r15
-
-		// next!
-		add b32 $r4 0x8000
-		sub b32 $r3 1
-		bra ne #init_gpc
-
-	// save context size, and tell host we're ready
-	mov $r2 0x800
-	shl b32 $r2 6
-	iowr I[$r2 + 0x100] $r1		// CC_SCRATCH[1]  = context size
-	add b32 $r2 0x800
-	clear b32 $r1
-	bset $r1 31
-	iowr I[$r2 + 0x000] $r1		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	// sleep until we have something to do
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// context switch, requested by GPU?
-	cmpu b32 $r14 0x4001
-	bra ne #main_not_ctx_switch
-		trace_set(T_AUTO)
-		mov $r1 0xb00
-		shl b32 $r1 6
-		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
-		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
-
-		xbit $r3 $r1 31
-		bra e #chsw_no_prev
-			xbit $r3 $r2 31
-			bra e #chsw_prev_no_next
-				push $r2
-				mov b32 $r2 $r1
-				trace_set(T_SAVE)
-				bclr $flags $p1
-				bset $flags $p2
-				call #ctx_xfer
-				trace_clr(T_SAVE);
-				pop $r2
-				trace_set(T_LOAD);
-				bset $flags $p1
-				call #ctx_xfer
-				trace_clr(T_LOAD);
-				bra #chsw_done
-			chsw_prev_no_next:
-				push $r2
-				mov b32 $r2 $r1
-				bclr $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-				pop $r2
-				mov $r1 0xb00
-				shl b32 $r1 6
-				iowr I[$r1] $r2
-				bra #chsw_done
-		chsw_no_prev:
-			xbit $r3 $r2 31
-			bra e #chsw_done
-				bset $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-
-		// ack the context switch request
-		chsw_done:
-		mov $r1 0xb0c
-		shl b32 $r1 6
-		mov $r2 1
-		iowr I[$r1 + 0x000] $r2		// 0x409b0c
-		trace_clr(T_AUTO)
-		bra #main
-
-	// request to set current channel? (*not* a context switch)
-	main_not_ctx_switch:
-	cmpu b32 $r14 0x0001
-	bra ne #main_not_ctx_chan
-		mov b32 $r2 $r15
-		call #ctx_chan
-		bra #main_done
-
-	// request to store current channel context?
-	main_not_ctx_chan:
-	cmpu b32 $r14 0x0002
-	bra ne #main_not_ctx_save
-		trace_set(T_SAVE)
-		bclr $flags $p1
-		bclr $flags $p2
-		call #ctx_xfer
-		trace_clr(T_SAVE)
-		bra #main_done
-
-	main_not_ctx_save:
-		shl b32 $r15 $r14 16
-		or $r15 E_BAD_COMMAND
-		call #error
-		bra #main
-
-	main_done:
-	mov $r1 0x820
-	shl b32 $r1 6
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// context switch request?
-	ih_no_fifo:
-	and $r11 $r10 0x00000100
-	bra e #ih_no_ctxsw
-		// enqueue a context switch for later processing
-		mov $r13 #cmd_queue
-		mov $r14 0x4001
-		call #queue_put
-
-	// anything we didn't handle, bring it to the host's attention
-	ih_no_ctxsw:
-	mov $r11 0x104
-	not b32 $r11
-	and $r11 $r10 $r11
-	bra e #ih_no_other
-		mov $r10 0xc1c
-		shl b32 $r10 6
-		iowr I[$r10] $r11	// INTR_UP_SET
-
-	// ack, and wake up main()
-	ih_no_other:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Not real sure, but, MEM_CMD 7 will hang forever if this isn't done
-ctx_4160s:
-	mov $r14 0x4160
-	sethi $r14 0x400000
-	mov $r15 1
-	call #nv_wr32
-	ctx_4160s_wait:
-		call #nv_rd32
-		xbit $r15 $r15 4
-		bra e #ctx_4160s_wait
-	ret
-
-// Without clearing again at end of xfer, some things cause PGRAPH
-// to hang with STATUS=0x00000007 until it's cleared.. fbcon can
-// still function with it set however...
-ctx_4160c:
-	mov $r14 0x4160
-	sethi $r14 0x400000
-	clear b32 $r15
-	call #nv_wr32
-	ret
-
-// Again, not real sure
-//
-// In: $r15 value to set 0x404170 to
-//
-ctx_4170s:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	or $r15 0x10
-	call #nv_wr32
-	ret
-
-// Waits for a ctx_4170s() call to complete
-//
-ctx_4170w:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	call #nv_rd32
-	and $r15 0x10
-	bra ne #ctx_4170w
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x270
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0x770
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
-	ret
-
-// Not a clue what this is for, except that unless the value is 0x10, the
-// strand context is saved (and presumably restored) incorrectly..
-//
-// In: $r15 value to set to (0x00/0x10 are used)
-//
-ctx_86c:
-	mov $r14 0x86c
-	shl b32 $r14 6
-	iowr I[$r14] $r15	// HUB(0x86c) = val
-	mov $r14 -0x75ec
-	sethi $r14 0x400000
-	call #nv_wr32		// ROP(0xa14) = val
-	mov $r14 -0x5794
-	sethi $r14 0x410000
-	call #nv_wr32		// GPC(0x86c) = val
-	ret
-
-// ctx_load - load's a channel's ctxctl data, and selects its vm
-//
-// In: $r2 channel address
-//
-ctx_load:
-	trace_set(T_CHAN)
-
-	// switch to channel, somewhat magic in parts..
-	mov $r10 12		// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa24
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r0	// 0x409a24
-	mov $r3 0xb00
-	shl b32 $r3 6
-	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
-	mov $r1 0xa0c
-	shl b32 $r1 6
-	mov $r4 7
-	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
-	iowr I[$r1 + 0x100] $r4	// MEM_CMD
-	ctx_chan_wait_0:
-		iord $r4 I[$r1 + 0x100]
-		and $r4 0x1f
-		bra ne #ctx_chan_wait_0
-	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
-
-	// load channel header, fetch PGRAPH context pointer
-	mov $xtargets $r0
-	bclr $r2 31
-	shl b32 $r2 4
-	add b32 $r2 2
-
-	trace_set(T_LCHAN)
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_BASE
-	mov $r1 0xa20
-	shl b32 $r1 6
-	mov $r2 0x0002
-	sethi $r2 0x80000000
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
-	mov $r1 0x10			// chan + 0x0210
-	mov $r2 #xfer_data
-	sethi $r2 0x00020000		// 16 bytes
-	xdld $r1 $r2
-	xdwait
-	trace_clr(T_LCHAN)
-
-	// update current context
-	ld b32 $r1 D[$r0 + #xfer_data + 4]
-	shl b32 $r1 24
-	ld b32 $r2 D[$r0 + #xfer_data + 0]
-	shr b32 $r2 8
-	or $r1 $r2
-	st b32 D[$r0 + #ctx_current] $r1
-
-	// set transfer base to start of context, and fetch context header
-	trace_set(T_LCTXH)
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1		// MEM_BASE
-	mov $r2 1
-	mov $r1 0xa20
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdld $r0 $r1
-	xdwait
-	trace_clr(T_LCTXH)
-
-	trace_clr(T_CHAN)
-	ret
-
-// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
-//            the active channel for ctxctl, but not actually transfer
-//            any context data.  intended for use only during initial
-//            context construction.
-//
-// In: $r2 channel address
-//
-ctx_chan:
-	call #ctx_4160s
-	call #ctx_load
-	mov $r10 12			// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa10
-	shl b32 $r1 6
-	mov $r2 5
-	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
-	ctx_chan_wait:
-		iord $r2 I[$r1 + 0x000]
-		or $r2 $r2
-		bra ne #ctx_chan_wait
-	call #ctx_4160c
-	ret
-
-// Execute per-context state overrides list
-//
-// Only executed on the first load of a channel.  Might want to look into
-// removing this and having the host directly modify the channel's context
-// to change this state...  The nouveau DRM already builds this list as
-// it's definitely needed for NVIDIA's, so we may as well use it for now
-//
-// Input: $r1 mmio list length
-//
-ctx_mmio_exec:
-	// set transfer base to be the mmio list
-	ld b32 $r3 D[$r0 + #chan_mmio_address]
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	clear b32 $r3
-	ctx_mmio_loop:
-		// fetch next 256 bytes of mmio list if necessary
-		and $r4 $r3 0xff
-		bra ne #ctx_mmio_pull
-			mov $r5 #xfer_data
-			sethi $r5 0x00060000	// 256 bytes
-			xdld $r3 $r5
-			xdwait
-
-		// execute a single list entry
-		ctx_mmio_pull:
-		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
-		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
-		call #nv_wr32
-
-		// next!
-		add b32 $r3 8
-		sub b32 $r1 1
-		bra ne #ctx_mmio_loop
-
-	// set transfer base back to the current context
-	ctx_mmio_done:
-	ld b32 $r3 D[$r0 + #ctx_current]
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	// disable the mmio list now, we don't need/want to execute it again
-	st b32 D[$r0 + #chan_mmio_count] $r0
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdst $r0 $r1
-	xdwait
-	ret
-
-// Transfer HUB context data between GPU and storage area
-//
-// In: $r2 channel address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// according to mwk, some kind of wait for idle
-	mov $r15 0xc00
-	shl b32 $r15 6
-	mov $r14 4
-	iowr I[$r15 + 0x200] $r14
-	ctx_xfer_idle:
-		iord $r14 I[$r15 + 0x000]
-		and $r14 0x2000
-		bra ne #ctx_xfer_idle
-
-	bra not $p1 #ctx_xfer_pre
-	bra $p2 #ctx_xfer_pre_load
-	ctx_xfer_pre:
-		mov $r15 0x10
-		call #ctx_86c
-		call #ctx_4160s
-		bra not $p1 #ctx_xfer_exec
-
-	ctx_xfer_pre_load:
-		mov $r15 2
-		call #ctx_4170s
-		call #ctx_4170w
-		call #ctx_redswitch
-		clear b32 $r15
-		call #ctx_4170s
-		call #ctx_load
-
-	// fetch context pointer, and initiate xfer on all GPCs
-	ctx_xfer_exec:
-	ld b32 $r1 D[$r0 + #ctx_current]
-	mov $r2 0x414
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
-	mov $r14 -0x5b00
-	sethi $r14 0x410000
-	mov b32 $r15 $r1
-	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
-	add b32 $r14 4
-	xbit $r15 $flags $p1
-	xbit $r2 $flags $p2
-	shl b32 $r2 1
-	or $r15 $r2
-	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 6		// first, last
-	mov $r11 0		// base = 0
-	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
-	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// wait for GPCs to all complete
-	mov $r10 8		// DONE_BAR
-	call #wait_doneo
-
-	// wait for strand xfer to complete
-	call #strand_wait
-
-	// post-op
-	bra $p1 #ctx_xfer_post
-		mov $r10 12		// DONE_UNK12
-		call #wait_donez
-		mov $r1 0xa10
-		shl b32 $r1 6
-		mov $r2 5
-		iowr I[$r1] $r2		// MEM_CMD
-		ctx_xfer_post_save_wait:
-			iord $r2 I[$r1]
-			or $r2 $r2
-			bra ne #ctx_xfer_post_save_wait
-
-	bra $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r15 2
-		call #ctx_4170s
-		clear b32 $r15
-		call #ctx_86c
-		call #strand_post
-		call #ctx_4170w
-		clear b32 $r15
-		call #ctx_4170s
-
-		bra not $p1 #ctx_xfer_no_post_mmio
-		ld b32 $r1 D[$r0 + #chan_mmio_count]
-		or $r1 $r1
-		bra e #ctx_xfer_no_post_mmio
-			call #ctx_mmio_exec
-
-		ctx_xfer_no_post_mmio:
-		call #ctx_4160c
-
-	ctx_xfer_done:
-	ret
-
+#include "com.fuc"
+#include "hub.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h
index bb03d2a..b59f694 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h
@@ -1,9 +1,90 @@
 uint32_t nvc0_grhub_data[] = {
-/* 0x0000: gpc_count */
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
+	0x00000000,
+/* 0x000c: rop_count */
+	0x00000000,
+/* 0x0010: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0058: ctx_current */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
-/* 0x0004: rop_count */
 	0x00000000,
-/* 0x0008: cmd_queue */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -22,114 +103,9 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0050: hub_mmio_list_head */
 	0x00000000,
-/* 0x0054: hub_mmio_list_tail */
 	0x00000000,
-/* 0x0058: ctx_current */
 	0x00000000,
-/* 0x005c: chipsets */
-	0x000000c0,
-	0x013c00a0,
-	0x000000c1,
-	0x014000a0,
-	0x000000c3,
-	0x013c00a0,
-	0x000000c4,
-	0x013c00a0,
-	0x000000c8,
-	0x013c00a0,
-	0x000000ce,
-	0x013c00a0,
-	0x000000cf,
-	0x013c00a0,
-	0x000000d9,
-	0x01dc0140,
-	0x00000000,
-/* 0x00a0: nvc0_hub_mmio_head */
-	0x0417e91c,
-	0x04400204,
-	0x28404004,
-	0x00404044,
-	0x34404094,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x04404164,
-	0x08404174,
-	0x1c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x7c404618,
-	0x50404698,
-	0x044046f0,
-	0x54404700,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x044064b4,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x0c408900,
-	0x00408980,
-/* 0x013c: nvc0_hub_mmio_tail */
-	0x044064c0,
-/* 0x0140: nvc1_hub_mmio_tail */
-/* 0x0140: nvd9_hub_mmio_head */
-	0x0417e91c,
-	0x04400204,
-	0x24404004,
-	0x00404044,
-	0x34404094,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x04404164,
-	0x04404178,
-	0x1c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x7c404618,
-	0x50404698,
-	0x044046f0,
-	0x54404700,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x104064b4,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x0c408900,
-	0x00408980,
-/* 0x01dc: nvd9_hub_mmio_tail */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -139,10 +115,7 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0200: chan_data */
-/* 0x0200: chan_mmio_count */
 	0x00000000,
-/* 0x0204: chan_mmio_address */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -163,6 +136,7 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0200: xfer_data */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -206,19 +180,40 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0300: xfer_data */
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
 };
 
 uint32_t nvc0_grhub_code[] = {
-	0x03090ef5,
+	0x031b0ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -250,7 +245,7 @@ uint32_t nvc0_grhub_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -272,63 +267,66 @@ uint32_t nvc0_grhub_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -336,8 +334,8 @@ uint32_t nvc0_grhub_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -346,76 +344,77 @@ uint32_t nvc0_grhub_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -424,73 +423,61 @@ uint32_t nvc0_grhub_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe4b60814,
-	0x00efd006,
-	0x0c1ce7f1,
-	0xf006e4b6,
-	0xefd001f7,
-	0xf8e0fc00,
-/* 0x0309: init */
-	0xfe04bd00,
-	0x07fe0004,
-	0x0017f100,
-	0x0227f012,
-	0xf10012d0,
-	0xfe05b917,
-	0x17f10010,
-	0x10d00400,
-	0x0437f1c0,
-	0x0634b604,
-	0x200327f1,
-	0xf10032d0,
-	0xd0200427,
-	0x27f10132,
-	0x32d0200b,
-	0x0c27f102,
-	0x0732d020,
-	0x0c2427f1,
-	0xb90624b6,
-	0x23d00003,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
 	0x0427f100,
-	0x0023f087,
-	0xb70012d0,
-	0xf0010012,
-	0x12d00427,
-	0x1031f400,
-	0x9604e7f1,
-	0xf440e3f0,
-	0xf1c76821,
-	0x01018090,
-	0x801ff4f0,
-	0x17f0000f,
-	0x041fbb01,
-	0xf10112b6,
-	0xb6040c27,
-	0x21d00624,
-	0x4021d000,
-	0x080027f1,
-	0xcf0624b6,
-	0xf7f00022,
-/* 0x03a9: init_find_chipset */
-	0x08f0b654,
-	0xb800f398,
-	0x0bf40432,
-	0x0034b00b,
-	0xf8f11bf4,
-/* 0x03bd: init_context */
-	0x0017f100,
-	0x02fe5801,
-	0xf003ff58,
-	0x0e8000e3,
-	0x150f8014,
-	0x013d21f5,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
 	0x070037f1,
 	0x950634b6,
 	0x34d00814,
@@ -501,208 +488,213 @@ uint32_t nvc0_grhub_code[] = {
 	0x0815b600,
 	0xb60110b6,
 	0x1fb90814,
-	0x6321f502,
+	0x7121f502,
 	0x001fbb02,
-	0xf1000398,
+	0xf1020398,
 	0xf0200047,
-/* 0x040e: init_gpc */
+/* 0x03f6: init_gpc */
 	0x4ea05043,
 	0x1fb90804,
 	0x8d21f402,
-	0x08004ea0,
-	0xf4022fb9,
-	0x4ea08d21,
-	0xf4bd010c,
-	0xa08d21f4,
-	0xf401044e,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
 	0x4ea08d21,
-	0xf7f00100,
-	0x8d21f402,
-	0x08004ea0,
-/* 0x0440: init_gpc_wait */
-	0xc86821f4,
-	0x0bf41fff,
-	0x044ea0fa,
-	0x6821f408,
-	0xb7001fbb,
-	0xb6800040,
-	0x1bf40132,
-	0x0027f1b4,
-	0x0624b608,
-	0xb74021d0,
-	0xbd080020,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
 	0x1f19f014,
-/* 0x0473: main */
-	0xf40021d0,
-	0x28f40031,
-	0x08d7f000,
-	0xf43921f4,
-	0xe4b1f401,
-	0x1bf54001,
-	0x87f100d1,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00499,
-	0x0017f100,
-	0x0614b60b,
-	0xcf4012cf,
-	0x13c80011,
-	0x7e0bf41f,
+	0x080007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
 	0xf41f23c8,
-	0x20f95a0b,
-	0xf10212b9,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00799f0,
-	0x32f40089,
-	0x0231f401,
-	0x082921f5,
-	0x085c87f1,
-	0xbd0684b6,
+	0x20f9620b,
+	0xbd0212b9,
 	0x0799f094,
-	0xfc0089d0,
-	0x3c87f120,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d006,
-	0xf50131f4,
-	0xf1082921,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00699f0,
-	0x0ef40089,
-/* 0x0509: chsw_prev_no_next */
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f00f00,
+	0x0009d002,
+	0x31f404bd,
+	0x2f21f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
 	0xb920f931,
 	0x32f40212,
 	0x0232f401,
-	0x082921f5,
+	0x082f21f5,
 	0x17f120fc,
 	0x14b60b00,
 	0x0012d006,
-/* 0x0527: chsw_no_prev */
+/* 0x0517: chsw_no_prev */
 	0xc8130ef4,
 	0x0bf41f23,
 	0x0131f40d,
 	0xf50232f4,
-/* 0x0537: chsw_done */
-	0xf1082921,
+/* 0x0527: chsw_done */
+	0xf1082f21,
 	0xb60b0c17,
 	0x27f00614,
 	0x0012d001,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0499f094,
-	0xf50089d0,
-/* 0x0557: main_not_ctx_switch */
-	0xb0ff200e,
-	0x1bf401e4,
-	0x02f2b90d,
-	0x07b521f5,
-/* 0x0567: main_not_ctx_chan */
-	0xb0420ef4,
-	0x1bf402e4,
-	0x3c87f12e,
-	0x0684b608,
 	0x99f094bd,
-	0x0089d007,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef407bb,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
 	0xf40132f4,
 	0x21f50232,
-	0x87f10829,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00799,
-	0x110ef400,
-/* 0x0598: main_not_ctx_save */
-	0xf010ef94,
-	0x21f501f5,
-	0x0ef502ec,
-/* 0x05a6: main_done */
-	0x17f1fed1,
-	0x14b60820,
-	0xf024bd06,
-	0x12d01f29,
-	0xbe0ef500,
-/* 0x05b9: ih */
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
 	0xfe80f9fe,
 	0x80f90188,
 	0xa0f990f9,
 	0xd0f9b0f9,
 	0xf0f9e0f9,
-	0xc4800acf,
-	0x0bf404ab,
-	0x00b7f11d,
-	0x08d7f019,
-	0xcf40becf,
-	0x21f400bf,
-	0x00b0b704,
-	0x01e7f004,
-/* 0x05ef: ih_no_fifo */
-	0xe400bed0,
-	0xf40100ab,
-	0xd7f00d0b,
-	0x01e7f108,
-	0x0421f440,
-/* 0x0600: ih_no_ctxsw */
-	0x0104b7f1,
-	0xabffb0bd,
-	0x0d0bf4b4,
-	0x0c1ca7f1,
-	0xd006a4b6,
-/* 0x0616: ih_no_other */
-	0x0ad000ab,
-	0xfcf0fc40,
-	0xfcd0fce0,
-	0xfca0fcb0,
-	0xfe80fc90,
-	0x80fc0088,
-	0xf80032f4,
-/* 0x0631: ctx_4160s */
-	0x60e7f101,
-	0x40e3f041,
-	0xf401f7f0,
-/* 0x063e: ctx_4160s_wait */
-	0x21f48d21,
-	0x04ffc868,
-	0xf8fa0bf4,
-/* 0x0649: ctx_4160c */
-	0x60e7f100,
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4160s */
+	0xf101f800,
+	0xf04160e7,
+	0xf7f040e3,
+	0x8d21f401,
+/* 0x0638: ctx_4160s_wait */
+	0xc86821f4,
+	0x0bf404ff,
+/* 0x0643: ctx_4160c */
+	0xf100f8fa,
+	0xf04160e7,
+	0xf4bd40e3,
+	0xf88d21f4,
+/* 0x0651: ctx_4170s */
+	0x70e7f100,
 	0x40e3f041,
-	0x21f4f4bd,
-/* 0x0657: ctx_4170s */
-	0xf100f88d,
-	0xf04170e7,
-	0xf5f040e3,
-	0x8d21f410,
-/* 0x0666: ctx_4170w */
-	0xe7f100f8,
-	0xe3f04170,
-	0x6821f440,
-	0xf410f4f0,
-	0x00f8f31b,
-/* 0x0678: ctx_redswitch */
-	0x0614e7f1,
-	0xf106e4b6,
-	0xd00270f7,
-	0xf7f000ef,
-/* 0x0689: ctx_redswitch_delay */
-	0x01f2b608,
-	0xf1fd1bf4,
-	0xd00770f7,
-	0x00f800ef,
-/* 0x0698: ctx_86c */
-	0x086ce7f1,
-	0xd006e4b6,
-	0xe7f100ef,
-	0xe3f08a14,
-	0x8d21f440,
-	0xa86ce7f1,
-	0xf441e3f0,
+	0xf410f5f0,
 	0x00f88d21,
-/* 0x06b8: ctx_load */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0599f094,
-	0xf00089d0,
+/* 0x0660: ctx_4170w */
+	0x4170e7f1,
+	0xf440e3f0,
+	0xf4f06821,
+	0xf31bf410,
+/* 0x0672: ctx_redswitch */
+	0xe7f100f8,
+	0xe4b60614,
+	0x70f7f106,
+	0x00efd002,
+/* 0x0683: ctx_redswitch_delay */
+	0xb608f7f0,
+	0x1bf401f2,
+	0x70f7f1fd,
+	0x00efd007,
+/* 0x0692: ctx_86c */
+	0xe7f100f8,
+	0xe4b6086c,
+	0x00efd006,
+	0x8a14e7f1,
+	0xf440e3f0,
+	0xe7f18d21,
+	0xe3f0a86c,
+	0x8d21f441,
+/* 0x06b2: ctx_load */
+	0x94bd00f8,
+	0xf10599f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf004bd00,
 	0x21f40ca7,
 	0x2417f1c9,
 	0x0614b60a,
@@ -713,168 +705,169 @@ uint32_t nvc0_grhub_code[] = {
 	0x0614b60a,
 	0xd00747f0,
 	0x14d00012,
-/* 0x06f1: ctx_chan_wait_0 */
+/* 0x06ed: ctx_chan_wait_0 */
 	0x4014cf40,
 	0xf41f44f0,
 	0x32d0fa1b,
 	0x000bfe00,
 	0xb61f2af0,
 	0x20b60424,
-	0x3c87f102,
-	0x0684b608,
+	0xf094bd02,
+	0x07f10899,
+	0x03f00f00,
+	0x0009d002,
+	0x17f104bd,
+	0x14b60a04,
+	0x0012d006,
+	0x0a2017f1,
+	0xf00614b6,
+	0x23f10227,
+	0x12d08000,
+	0x1017f000,
+	0x020027f1,
+	0xfa0223f0,
+	0x03f80512,
 	0x99f094bd,
-	0x0089d008,
-	0x0a0417f1,
-	0xd00614b6,
-	0x17f10012,
-	0x14b60a20,
-	0x0227f006,
-	0x800023f1,
-	0xf00012d0,
-	0x27f11017,
-	0x23f00300,
-	0x0512fa02,
-	0x87f103f8,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00899,
-	0xc1019800,
+	0x0007f108,
+	0x0203f017,
+	0xbd0009d0,
+	0x81019804,
 	0x981814b6,
-	0x25b6c002,
+	0x25b68002,
 	0x0512fd08,
-	0xf1160180,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00999f0,
-	0x27f10089,
-	0x24b60a04,
-	0x0021d006,
-	0xf10127f0,
-	0xb60a2017,
-	0x12d00614,
-	0x0017f100,
-	0x0613f002,
-	0xf80501fa,
-	0x5c87f103,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d009,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0599f094,
-	0xf80089d0,
-/* 0x07b5: ctx_chan */
-	0x3121f500,
-	0xb821f506,
-	0x0ca7f006,
-	0xf1c921f4,
-	0xb60a1017,
-	0x27f00614,
-	0x0012d005,
-/* 0x07d0: ctx_chan_wait */
-	0xfd0012cf,
-	0x1bf40522,
-	0x4921f5fa,
-/* 0x07df: ctx_mmio_exec */
-	0x9800f806,
-	0x27f18103,
-	0x24b60a04,
-	0x0023d006,
-/* 0x07ee: ctx_mmio_loop */
-	0x34c434bd,
-	0x0f1bf4ff,
-	0x030057f1,
-	0xfa0653f0,
-	0x03f80535,
-/* 0x0800: ctx_mmio_pull */
-	0x98c04e98,
-	0x21f4c14f,
-	0x0830b68d,
-	0xf40112b6,
-/* 0x0812: ctx_mmio_done */
-	0x0398df1b,
-	0x0023d016,
-	0xf1800080,
-	0xf0020017,
+	0xbd160180,
+	0x0999f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0427f1,
+	0xd00624b6,
+	0x27f00021,
+	0x2017f101,
+	0x0614b60a,
+	0xf10012d0,
+	0xf0010017,
 	0x01fa0613,
-	0xf803f806,
-/* 0x0829: ctx_xfer */
-	0x00f7f100,
-	0x06f4b60c,
-	0xd004e7f0,
-/* 0x0836: ctx_xfer_idle */
-	0xfecf80fe,
-	0x00e4f100,
-	0xf91bf420,
-	0xf40611f4,
-/* 0x0846: ctx_xfer_pre */
-	0xf7f01102,
-	0x9821f510,
-	0x3121f506,
-	0x1c11f406,
-/* 0x0854: ctx_xfer_pre_load */
-	0xf502f7f0,
-	0xf5065721,
-	0xf5066621,
-	0xbd067821,
-	0x5721f5f4,
-	0xb821f506,
-/* 0x086d: ctx_xfer_exec */
-	0x16019806,
-	0x041427f1,
+	0xbd03f805,
+	0x0999f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x99f094bd,
+	0x0007f105,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x07bb: ctx_chan */
+	0xf500f804,
+	0xf5062b21,
+	0xf006b221,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x07d6: ctx_chan_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf5fa1bf4,
+	0xf8064321,
+/* 0x07e5: ctx_mmio_exec */
+	0x41039800,
+	0x0a0427f1,
 	0xd00624b6,
-	0xe7f10020,
-	0xe3f0a500,
-	0x021fb941,
+	0x34bd0023,
+/* 0x07f4: ctx_mmio_loop */
+	0xf4ff34c4,
+	0x57f10f1b,
+	0x53f00200,
+	0x0535fa06,
+/* 0x0806: ctx_mmio_pull */
+	0x4e9803f8,
+	0x814f9880,
 	0xb68d21f4,
-	0xfcf004e0,
-	0x022cf001,
-	0xfd0124b6,
-	0x21f405f2,
-	0xfc17f18d,
-	0x0213f04a,
-	0xd00c27f0,
-	0x21f50012,
-	0x27f10207,
-	0x23f047fc,
-	0x0020d002,
-	0xb6012cf0,
-	0x12d00320,
-	0x01acf000,
-	0xf006a5f0,
-	0x0c9800b7,
-	0x150d9814,
-	0xf500e7f0,
-	0xf0015c21,
-	0x21f508a7,
-	0x21f50103,
-	0x01f40207,
-	0x0ca7f022,
-	0xf1c921f4,
-	0xb60a1017,
-	0x27f00614,
-	0x0012d005,
-/* 0x08f4: ctx_xfer_post_save_wait */
-	0xfd0012cf,
-	0x1bf40522,
-	0x3202f4fa,
-/* 0x0900: ctx_xfer_post */
-	0xf502f7f0,
-	0xbd065721,
-	0x9821f5f4,
-	0x2621f506,
-	0x6621f502,
+	0x12b60830,
+	0xdf1bf401,
+/* 0x0818: ctx_mmio_done */
+	0xd0160398,
+	0x00800023,
+	0x0017f140,
+	0x0613f001,
+	0xf80601fa,
+/* 0x082f: ctx_xfer */
+	0xf100f803,
+	0xb60c00f7,
+	0xe7f006f4,
+	0x80fed004,
+/* 0x083c: ctx_xfer_idle */
+	0xf100fecf,
+	0xf42000e4,
+	0x11f4f91b,
+	0x1102f406,
+/* 0x084c: ctx_xfer_pre */
+	0xf510f7f0,
+	0xf5069221,
+	0xf4062b21,
+/* 0x085a: ctx_xfer_pre_load */
+	0xf7f01c11,
+	0x5121f502,
+	0x6021f506,
+	0x7221f506,
 	0xf5f4bd06,
-	0xf4065721,
-	0x01981011,
-	0x0511fd80,
-	0xf5070bf4,
-/* 0x092b: ctx_xfer_no_post_mmio */
-	0xf507df21,
-/* 0x092f: ctx_xfer_done */
-	0xf8064921,
-	0x00000000,
-	0x00000000,
+	0xf5065121,
+/* 0x0873: ctx_xfer_exec */
+	0x9806b221,
+	0x27f11601,
+	0x24b60414,
+	0x0020d006,
+	0xa500e7f1,
+	0xb941e3f0,
+	0x21f4021f,
+	0x04e0b68d,
+	0xf001fcf0,
+	0x24b6022c,
+	0x05f2fd01,
+	0xf18d21f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00c,
+	0x021521f5,
+	0x47fc27f1,
+	0xd00223f0,
+	0x2cf00020,
+	0x0320b601,
+	0xf00012d0,
+	0xa5f001ac,
+	0x00b7f006,
+	0x98000c98,
+	0xe7f0010d,
+	0x6621f500,
+	0x08a7f001,
+	0x010921f5,
+	0x021521f5,
+	0xf02201f4,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x08fa: ctx_xfer_post_save_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf4fa1bf4,
+/* 0x0906: ctx_xfer_post */
+	0xf7f03202,
+	0x5121f502,
+	0xf5f4bd06,
+	0xf5069221,
+	0xf5023421,
+	0xbd066021,
+	0x5121f5f4,
+	0x1011f406,
+	0xfd400198,
+	0x0bf40511,
+	0xe521f507,
+/* 0x0931: ctx_xfer_no_post_mmio */
+	0x4321f507,
+/* 0x0935: ctx_xfer_done */
+	0x0000f806,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc
new file mode 100644
index 0000000..afbe03a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define CHIPSET GF117
+#include "macros.fuc"
+
+.section #nvd7_grhub_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
+
+.section #nvd7_grhub_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "hub.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc.h
new file mode 100644
index 0000000..a1b9f76
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc.h
@@ -0,0 +1,921 @@
+uint32_t nvd7_grhub_data[] = {
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
+	0x00000000,
+/* 0x000c: rop_count */
+	0x00000000,
+/* 0x0010: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0058: ctx_current */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0200: xfer_data */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
+};
+
+uint32_t nvd7_grhub_code[] = {
+	0x031b0ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
+	0x0427f100,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
+	0x070037f1,
+	0x950634b6,
+	0x34d00814,
+	0x4034d000,
+	0x130030b7,
+	0xb6001fbb,
+	0x3fd002f5,
+	0x0815b600,
+	0xb60110b6,
+	0x1fb90814,
+	0x7121f502,
+	0x001fbb02,
+	0xf1020398,
+	0xf0200047,
+/* 0x03f6: init_gpc */
+	0x4ea05043,
+	0x1fb90804,
+	0x8d21f402,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
+	0x4ea08d21,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
+	0x1f19f014,
+	0x080007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
+	0xf41f23c8,
+	0x20f9620b,
+	0xbd0212b9,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f00f00,
+	0x0009d002,
+	0x31f404bd,
+	0x2f21f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
+	0xb920f931,
+	0x32f40212,
+	0x0232f401,
+	0x082f21f5,
+	0x17f120fc,
+	0x14b60b00,
+	0x0012d006,
+/* 0x0517: chsw_no_prev */
+	0xc8130ef4,
+	0x0bf41f23,
+	0x0131f40d,
+	0xf50232f4,
+/* 0x0527: chsw_done */
+	0xf1082f21,
+	0xb60b0c17,
+	0x27f00614,
+	0x0012d001,
+	0x99f094bd,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef407bb,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50232,
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
+	0xfe80f9fe,
+	0x80f90188,
+	0xa0f990f9,
+	0xd0f9b0f9,
+	0xf0f9e0f9,
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4160s */
+	0xf101f800,
+	0xf04160e7,
+	0xf7f040e3,
+	0x8d21f401,
+/* 0x0638: ctx_4160s_wait */
+	0xc86821f4,
+	0x0bf404ff,
+/* 0x0643: ctx_4160c */
+	0xf100f8fa,
+	0xf04160e7,
+	0xf4bd40e3,
+	0xf88d21f4,
+/* 0x0651: ctx_4170s */
+	0x70e7f100,
+	0x40e3f041,
+	0xf410f5f0,
+	0x00f88d21,
+/* 0x0660: ctx_4170w */
+	0x4170e7f1,
+	0xf440e3f0,
+	0xf4f06821,
+	0xf31bf410,
+/* 0x0672: ctx_redswitch */
+	0xe7f100f8,
+	0xe4b60614,
+	0x70f7f106,
+	0x00efd002,
+/* 0x0683: ctx_redswitch_delay */
+	0xb608f7f0,
+	0x1bf401f2,
+	0x70f7f1fd,
+	0x00efd007,
+/* 0x0692: ctx_86c */
+	0xe7f100f8,
+	0xe4b6086c,
+	0x00efd006,
+	0x8a14e7f1,
+	0xf440e3f0,
+	0xe7f18d21,
+	0xe3f0a86c,
+	0x8d21f441,
+/* 0x06b2: ctx_load */
+	0x94bd00f8,
+	0xf10599f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf004bd00,
+	0x21f40ca7,
+	0x2417f1c9,
+	0x0614b60a,
+	0xf10010d0,
+	0xb60b0037,
+	0x32d00634,
+	0x0c17f140,
+	0x0614b60a,
+	0xd00747f0,
+	0x14d00012,
+/* 0x06ed: ctx_chan_wait_0 */
+	0x4014cf40,
+	0xf41f44f0,
+	0x32d0fa1b,
+	0x000bfe00,
+	0xb61f2af0,
+	0x20b60424,
+	0xf094bd02,
+	0x07f10899,
+	0x03f00f00,
+	0x0009d002,
+	0x17f104bd,
+	0x14b60a04,
+	0x0012d006,
+	0x0a2017f1,
+	0xf00614b6,
+	0x23f10227,
+	0x12d08000,
+	0x1017f000,
+	0x020027f1,
+	0xfa0223f0,
+	0x03f80512,
+	0x99f094bd,
+	0x0007f108,
+	0x0203f017,
+	0xbd0009d0,
+	0x81019804,
+	0x981814b6,
+	0x25b68002,
+	0x0512fd08,
+	0xbd160180,
+	0x0999f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0427f1,
+	0xd00624b6,
+	0x27f00021,
+	0x2017f101,
+	0x0614b60a,
+	0xf10012d0,
+	0xf0010017,
+	0x01fa0613,
+	0xbd03f805,
+	0x0999f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x99f094bd,
+	0x0007f105,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x07bb: ctx_chan */
+	0xf500f804,
+	0xf5062b21,
+	0xf006b221,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x07d6: ctx_chan_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf5fa1bf4,
+	0xf8064321,
+/* 0x07e5: ctx_mmio_exec */
+	0x41039800,
+	0x0a0427f1,
+	0xd00624b6,
+	0x34bd0023,
+/* 0x07f4: ctx_mmio_loop */
+	0xf4ff34c4,
+	0x57f10f1b,
+	0x53f00200,
+	0x0535fa06,
+/* 0x0806: ctx_mmio_pull */
+	0x4e9803f8,
+	0x814f9880,
+	0xb68d21f4,
+	0x12b60830,
+	0xdf1bf401,
+/* 0x0818: ctx_mmio_done */
+	0xd0160398,
+	0x00800023,
+	0x0017f140,
+	0x0613f001,
+	0xf80601fa,
+/* 0x082f: ctx_xfer */
+	0xf100f803,
+	0xb60c00f7,
+	0xe7f006f4,
+	0x80fed004,
+/* 0x083c: ctx_xfer_idle */
+	0xf100fecf,
+	0xf42000e4,
+	0x11f4f91b,
+	0x1102f406,
+/* 0x084c: ctx_xfer_pre */
+	0xf510f7f0,
+	0xf5069221,
+	0xf4062b21,
+/* 0x085a: ctx_xfer_pre_load */
+	0xf7f01c11,
+	0x5121f502,
+	0x6021f506,
+	0x7221f506,
+	0xf5f4bd06,
+	0xf5065121,
+/* 0x0873: ctx_xfer_exec */
+	0x9806b221,
+	0x27f11601,
+	0x24b60414,
+	0x0020d006,
+	0xa500e7f1,
+	0xb941e3f0,
+	0x21f4021f,
+	0x04e0b68d,
+	0xf001fcf0,
+	0x24b6022c,
+	0x05f2fd01,
+	0xf18d21f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00c,
+	0x021521f5,
+	0x47fc27f1,
+	0xd00223f0,
+	0x2cf00020,
+	0x0320b601,
+	0xf00012d0,
+	0xa5f001ac,
+	0x00b7f006,
+	0x98000c98,
+	0xe7f0010d,
+	0x6621f500,
+	0x08a7f001,
+	0x010921f5,
+	0x021521f5,
+	0xf02201f4,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x08fa: ctx_xfer_post_save_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf4fa1bf4,
+/* 0x0906: ctx_xfer_post */
+	0xf7f03202,
+	0x5121f502,
+	0xf5f4bd06,
+	0xf5069221,
+	0xf5023421,
+	0xbd066021,
+	0x5121f5f4,
+	0x1011f406,
+	0xfd400198,
+	0x0bf40511,
+	0xe521f507,
+/* 0x0931: ctx_xfer_no_post_mmio */
+	0x4321f507,
+/* 0x0935: ctx_xfer_done */
+	0x0000f806,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc
index 7fe9d7c..d4840f1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nve0 PGRAPH/HUB
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,774 +19,22 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 nve0_grhub.fuc | envyas -a -w -m fuc -V nva3 -o nve0_grhub.fuc.h
- */
+#define CHIPSET GK100
+#include "macros.fuc"
 
 .section #nve0_grhub_data
-include(`nve0.fuc')
-gpc_count:		.b32 0
-rop_count:		.b32 0
-cmd_queue:		queue_init
-hub_mmio_list_head:	.b32 0
-hub_mmio_list_tail:	.b32 0
-
-ctx_current:		.b32 0
-
-chipsets:
-.b8  0xe4 0 0 0
-.b16 #nve4_hub_mmio_head
-.b16 #nve4_hub_mmio_tail
-.b8  0xe7 0 0 0
-.b16 #nve4_hub_mmio_head
-.b16 #nve4_hub_mmio_tail
-.b8  0xe6 0 0 0
-.b16 #nve4_hub_mmio_head
-.b16 #nve4_hub_mmio_tail
-.b8  0 0 0 0
-
-nve4_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404010, 7)
-mmctx_data(0x4040a8, 9)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 1)
-mmctx_data(0x4041a0, 4)
-mmctx_data(0x404200, 4)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 4)
-mmctx_data(0x40462c, 2)
-mmctx_data(0x404640, 1)
-mmctx_data(0x404654, 1)
-mmctx_data(0x404660, 1)
-mmctx_data(0x404678, 19)
-mmctx_data(0x4046c8, 3)
-mmctx_data(0x404700, 3)
-mmctx_data(0x404718, 10)
-mmctx_data(0x404744, 2)
-mmctx_data(0x404754, 1)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x405b00, 1)
-mmctx_data(0x405b10, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 2)
-mmctx_data(0x4064c0, 12)
-mmctx_data(0x4064fc, 1)
-mmctx_data(0x407040, 1)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408840, 1)
-mmctx_data(0x408900, 3)
-mmctx_data(0x408980, 1)
-nve4_hub_mmio_tail:
-
-.align 256
-chan_data:
-chan_mmio_count:	.b32 0
-chan_mmio_address:	.b32 0
-
-.align 256
-xfer_data: 		.b32 0
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
 
 .section #nve0_grhub_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nve0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nve0.fuc)
-//
-error:
-	push $r14
-	mov $r14 0x814
-	shl b32 $r14 6
-	iowr I[$r14 + 0x000] $r15	// CC_SCRATCH[5] = error code
-	mov $r14 0xc1c
-	shl b32 $r14 6
-	mov $r15 1
-	iowr I[$r14 + 0x000] $r15	// INTR_UP_SET
-	pop $r14
-	ret
-
-// HUB fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: total PGRAPH context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-	mov $xdbase $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
-
-	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
-	mov $r3 0x404
-	shl b32 $r3 6
-	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
-	iowr I[$r3 + 0x000] $r2
-
-	// not sure what these are, route them because NVIDIA does, and
-	// the IRQ handler will signal the host if we ever get one.. we
-	// may find out if/why we need to handle these if so..
-	//
-	mov $r2 0x2004
-	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
-	mov $r2 0x200b
-	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
-	mov $r2 0x200c
-	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
-
-	// enable all INTR_UP interrupts
-	mov $r2 0xc24
-	shl b32 $r2 6
-	not b32 $r3 $r0
-	iowr I[$r2] $r3
-
-	// enable fifo, ctxsw, 9, 10, 15 interrupts
-	mov $r2 -0x78fc		// 0x8704
-	sethi $r2 0
-	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
-
-	// fifo level triggered, rest edge
-	sub b32 $r1 0x100
-	mov $r2 4
-	iowr I[$r1] $r2
-
-	// enable interrupts
-	bset $flags ie0
-
-	// fetch enabled GPC/ROP counts
-	mov $r14 -0x69fc	// 0x409604
-	sethi $r14 0x400000
-	call #nv_rd32
-	extr $r1 $r15 16:20
-	st b32 D[$r0 + #rop_count] $r1
-	and $r15 0x1f
-	st b32 D[$r0 + #gpc_count] $r15
-
-	// set BAR_REQMASK to GPC mask
-	mov $r1 1
-	shl b32 $r1 $r15
-	sub b32 $r1 1
-	mov $r2 0x40c
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1
-	iowr I[$r2 + 0x100] $r1
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r15 #chipsets - 8
-	init_find_chipset:
-		add b32 $r15 8
-		ld b32 $r3 D[$r15 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// context size calculation, reserve first 256 bytes for use by fuc
-	init_context:
-	mov $r1 256
-
-	// calculate size of mmio context data
-	ld b16 $r14 D[$r15 + 4]
-	ld b16 $r15 D[$r15 + 6]
-	sethi $r14 0
-	st b32 D[$r0 + #hub_mmio_list_head] $r14
-	st b32 D[$r0 + #hub_mmio_list_tail] $r15
-	call #mmctx_size
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't (currently) ever change
-	mov $r3 0x700
-	shl b32 $r3 6
-	shr b32 $r4 $r1 8
-	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
-	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
-	add b32 $r3 0x1300
-	add b32 $r1 $r15
-	shr b32 $r15 2
-	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
-
-	// strands, base offset needs to be aligned to 256 bytes
-	shr b32 $r1 8
-	add b32 $r1 1
-	shl b32 $r1 8
-	mov b32 $r15 $r1
-	call #strand_ctx_init
-	add b32 $r1 $r15
-
-	// initialise each GPC in sequence by passing in the offset of its
-	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
-	// has previously been uploaded by the host) running.
-	//
-	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
-	// when it has completed, and return the size of its context data
-	// in GPCn_CC_SCRATCH[1]
-	//
-	ld b32 $r3 D[$r0 + #gpc_count]
-	mov $r4 0x2000
-	sethi $r4 0x500000
-	init_gpc:
-		// setup, and start GPC ucode running
-		add b32 $r14 $r4 0x804
-		mov b32 $r15 $r1
-		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
-		add b32 $r14 $r4 0x800
-		mov b32 $r15 $r2
-		call #nv_wr32			// CC_SCRATCH[0] = chipset
-		add b32 $r14 $r4 0x10c
-		clear b32 $r15
-		call #nv_wr32
-		add b32 $r14 $r4 0x104
-		call #nv_wr32			// ENTRY
-		add b32 $r14 $r4 0x100
-		mov $r15 2			// CTRL_START_TRIGGER
-		call #nv_wr32			// CTRL
-
-		// wait for it to complete, and adjust context size
-		add b32 $r14 $r4 0x800
-		init_gpc_wait:
-			call #nv_rd32
-			xbit $r15 $r15 31
-			bra e #init_gpc_wait
-		add b32 $r14 $r4 0x804
-		call #nv_rd32
-		add b32 $r1 $r15
-
-		// next!
-		add b32 $r4 0x8000
-		sub b32 $r3 1
-		bra ne #init_gpc
-
-	// save context size, and tell host we're ready
-	mov $r2 0x800
-	shl b32 $r2 6
-	iowr I[$r2 + 0x100] $r1		// CC_SCRATCH[1]  = context size
-	add b32 $r2 0x800
-	clear b32 $r1
-	bset $r1 31
-	iowr I[$r2 + 0x000] $r1		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	// sleep until we have something to do
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// context switch, requested by GPU?
-	cmpu b32 $r14 0x4001
-	bra ne #main_not_ctx_switch
-		trace_set(T_AUTO)
-		mov $r1 0xb00
-		shl b32 $r1 6
-		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
-		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
-
-		xbit $r3 $r1 31
-		bra e #chsw_no_prev
-			xbit $r3 $r2 31
-			bra e #chsw_prev_no_next
-				push $r2
-				mov b32 $r2 $r1
-				trace_set(T_SAVE)
-				bclr $flags $p1
-				bset $flags $p2
-				call #ctx_xfer
-				trace_clr(T_SAVE);
-				pop $r2
-				trace_set(T_LOAD);
-				bset $flags $p1
-				call #ctx_xfer
-				trace_clr(T_LOAD);
-				bra #chsw_done
-			chsw_prev_no_next:
-				push $r2
-				mov b32 $r2 $r1
-				bclr $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-				pop $r2
-				mov $r1 0xb00
-				shl b32 $r1 6
-				iowr I[$r1] $r2
-				bra #chsw_done
-		chsw_no_prev:
-			xbit $r3 $r2 31
-			bra e #chsw_done
-				bset $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-
-		// ack the context switch request
-		chsw_done:
-		mov $r1 0xb0c
-		shl b32 $r1 6
-		mov $r2 1
-		iowr I[$r1 + 0x000] $r2		// 0x409b0c
-		trace_clr(T_AUTO)
-		bra #main
-
-	// request to set current channel? (*not* a context switch)
-	main_not_ctx_switch:
-	cmpu b32 $r14 0x0001
-	bra ne #main_not_ctx_chan
-		mov b32 $r2 $r15
-		call #ctx_chan
-		bra #main_done
-
-	// request to store current channel context?
-	main_not_ctx_chan:
-	cmpu b32 $r14 0x0002
-	bra ne #main_not_ctx_save
-		trace_set(T_SAVE)
-		bclr $flags $p1
-		bclr $flags $p2
-		call #ctx_xfer
-		trace_clr(T_SAVE)
-		bra #main_done
-
-	main_not_ctx_save:
-		shl b32 $r15 $r14 16
-		or $r15 E_BAD_COMMAND
-		call #error
-		bra #main
-
-	main_done:
-	mov $r1 0x820
-	shl b32 $r1 6
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// context switch request?
-	ih_no_fifo:
-	and $r11 $r10 0x00000100
-	bra e #ih_no_ctxsw
-		// enqueue a context switch for later processing
-		mov $r13 #cmd_queue
-		mov $r14 0x4001
-		call #queue_put
-
-	// anything we didn't handle, bring it to the host's attention
-	ih_no_ctxsw:
-	mov $r11 0x104
-	not b32 $r11
-	and $r11 $r10 $r11
-	bra e #ih_no_other
-		mov $r10 0xc1c
-		shl b32 $r10 6
-		iowr I[$r10] $r11	// INTR_UP_SET
-
-	// ack, and wake up main()
-	ih_no_other:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Again, not real sure
-//
-// In: $r15 value to set 0x404170 to
-//
-ctx_4170s:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	or $r15 0x10
-	call #nv_wr32
-	ret
-
-// Waits for a ctx_4170s() call to complete
-//
-ctx_4170w:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	call #nv_rd32
-	and $r15 0x10
-	bra ne #ctx_4170w
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x270
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0x770
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
-	ret
-
-// Not a clue what this is for, except that unless the value is 0x10, the
-// strand context is saved (and presumably restored) incorrectly..
-//
-// In: $r15 value to set to (0x00/0x10 are used)
-//
-ctx_86c:
-	mov $r14 0x86c
-	shl b32 $r14 6
-	iowr I[$r14] $r15	// HUB(0x86c) = val
-	mov $r14 -0x75ec
-	sethi $r14 0x400000
-	call #nv_wr32		// ROP(0xa14) = val
-	mov $r14 -0x5794
-	sethi $r14 0x410000
-	call #nv_wr32		// GPC(0x86c) = val
-	ret
-
-// ctx_load - load's a channel's ctxctl data, and selects its vm
-//
-// In: $r2 channel address
-//
-ctx_load:
-	trace_set(T_CHAN)
-
-	// switch to channel, somewhat magic in parts..
-	mov $r10 12		// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa24
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r0	// 0x409a24
-	mov $r3 0xb00
-	shl b32 $r3 6
-	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
-	mov $r1 0xa0c
-	shl b32 $r1 6
-	mov $r4 7
-	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
-	iowr I[$r1 + 0x100] $r4	// MEM_CMD
-	ctx_chan_wait_0:
-		iord $r4 I[$r1 + 0x100]
-		and $r4 0x1f
-		bra ne #ctx_chan_wait_0
-	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
-
-	// load channel header, fetch PGRAPH context pointer
-	mov $xtargets $r0
-	bclr $r2 31
-	shl b32 $r2 4
-	add b32 $r2 2
-
-	trace_set(T_LCHAN)
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_BASE
-	mov $r1 0xa20
-	shl b32 $r1 6
-	mov $r2 0x0002
-	sethi $r2 0x80000000
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
-	mov $r1 0x10			// chan + 0x0210
-	mov $r2 #xfer_data
-	sethi $r2 0x00020000		// 16 bytes
-	xdld $r1 $r2
-	xdwait
-	trace_clr(T_LCHAN)
-
-	// update current context
-	ld b32 $r1 D[$r0 + #xfer_data + 4]
-	shl b32 $r1 24
-	ld b32 $r2 D[$r0 + #xfer_data + 0]
-	shr b32 $r2 8
-	or $r1 $r2
-	st b32 D[$r0 + #ctx_current] $r1
-
-	// set transfer base to start of context, and fetch context header
-	trace_set(T_LCTXH)
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1		// MEM_BASE
-	mov $r2 1
-	mov $r1 0xa20
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdld $r0 $r1
-	xdwait
-	trace_clr(T_LCTXH)
-
-	trace_clr(T_CHAN)
-	ret
-
-// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
-//            the active channel for ctxctl, but not actually transfer
-//            any context data.  intended for use only during initial
-//            context construction.
-//
-// In: $r2 channel address
-//
-ctx_chan:
-	call #ctx_load
-	mov $r10 12			// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa10
-	shl b32 $r1 6
-	mov $r2 5
-	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
-	ctx_chan_wait:
-		iord $r2 I[$r1 + 0x000]
-		or $r2 $r2
-		bra ne #ctx_chan_wait
-	ret
-
-// Execute per-context state overrides list
-//
-// Only executed on the first load of a channel.  Might want to look into
-// removing this and having the host directly modify the channel's context
-// to change this state...  The nouveau DRM already builds this list as
-// it's definitely needed for NVIDIA's, so we may as well use it for now
-//
-// Input: $r1 mmio list length
-//
-ctx_mmio_exec:
-	// set transfer base to be the mmio list
-	ld b32 $r3 D[$r0 + #chan_mmio_address]
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	clear b32 $r3
-	ctx_mmio_loop:
-		// fetch next 256 bytes of mmio list if necessary
-		and $r4 $r3 0xff
-		bra ne #ctx_mmio_pull
-			mov $r5 #xfer_data
-			sethi $r5 0x00060000	// 256 bytes
-			xdld $r3 $r5
-			xdwait
-
-		// execute a single list entry
-		ctx_mmio_pull:
-		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
-		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
-		call #nv_wr32
-
-		// next!
-		add b32 $r3 8
-		sub b32 $r1 1
-		bra ne #ctx_mmio_loop
-
-	// set transfer base back to the current context
-	ctx_mmio_done:
-	ld b32 $r3 D[$r0 + #ctx_current]
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	// disable the mmio list now, we don't need/want to execute it again
-	st b32 D[$r0 + #chan_mmio_count] $r0
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdst $r0 $r1
-	xdwait
-	ret
-
-// Transfer HUB context data between GPU and storage area
-//
-// In: $r2 channel address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// according to mwk, some kind of wait for idle
-	mov $r15 0xc00
-	shl b32 $r15 6
-	mov $r14 4
-	iowr I[$r15 + 0x200] $r14
-	ctx_xfer_idle:
-		iord $r14 I[$r15 + 0x000]
-		and $r14 0x2000
-		bra ne #ctx_xfer_idle
-
-	bra not $p1 #ctx_xfer_pre
-	bra $p2 #ctx_xfer_pre_load
-	ctx_xfer_pre:
-		mov $r15 0x10
-		call #ctx_86c
-		bra not $p1 #ctx_xfer_exec
-
-	ctx_xfer_pre_load:
-		mov $r15 2
-		call #ctx_4170s
-		call #ctx_4170w
-		call #ctx_redswitch
-		clear b32 $r15
-		call #ctx_4170s
-		call #ctx_load
-
-	// fetch context pointer, and initiate xfer on all GPCs
-	ctx_xfer_exec:
-	ld b32 $r1 D[$r0 + #ctx_current]
-	mov $r2 0x414
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
-	mov $r14 -0x5b00
-	sethi $r14 0x410000
-	mov b32 $r15 $r1
-	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
-	add b32 $r14 4
-	xbit $r15 $flags $p1
-	xbit $r2 $flags $p2
-	shl b32 $r2 1
-	or $r15 $r2
-	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 6		// first, last
-	mov $r11 0		// base = 0
-	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
-	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// wait for GPCs to all complete
-	mov $r10 8		// DONE_BAR
-	call #wait_doneo
-
-	// wait for strand xfer to complete
-	call #strand_wait
-
-	// post-op
-	bra $p1 #ctx_xfer_post
-		mov $r10 12		// DONE_UNK12
-		call #wait_donez
-		mov $r1 0xa10
-		shl b32 $r1 6
-		mov $r2 5
-		iowr I[$r1] $r2		// MEM_CMD
-		ctx_xfer_post_save_wait:
-			iord $r2 I[$r1]
-			or $r2 $r2
-			bra ne #ctx_xfer_post_save_wait
-
-	bra $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r15 2
-		call #ctx_4170s
-		clear b32 $r15
-		call #ctx_86c
-		call #strand_post
-		call #ctx_4170w
-		clear b32 $r15
-		call #ctx_4170s
-
-		bra not $p1 #ctx_xfer_no_post_mmio
-		ld b32 $r1 D[$r0 + #chan_mmio_count]
-		or $r1 $r1
-		bra e #ctx_xfer_no_post_mmio
-			call #ctx_mmio_exec
-
-		ctx_xfer_no_post_mmio:
-
-	ctx_xfer_done:
-	ret
-
+#include "com.fuc"
+#include "hub.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h
index e3421af..eb7bc0e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h
@@ -1,9 +1,13 @@
 uint32_t nve0_grhub_data[] = {
-/* 0x0000: gpc_count */
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
 	0x00000000,
-/* 0x0004: rop_count */
+/* 0x000c: rop_count */
 	0x00000000,
-/* 0x0008: cmd_queue */
+/* 0x0010: cmd_queue */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -22,73 +26,11 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0050: hub_mmio_list_head */
+/* 0x0058: ctx_current */
 	0x00000000,
-/* 0x0054: hub_mmio_list_tail */
 	0x00000000,
-/* 0x0058: ctx_current */
 	0x00000000,
-/* 0x005c: chipsets */
-	0x000000e4,
-	0x01440078,
-	0x000000e7,
-	0x01440078,
-	0x000000e6,
-	0x01440078,
 	0x00000000,
-/* 0x0078: nve4_hub_mmio_head */
-	0x0417e91c,
-	0x04400204,
-	0x18404010,
-	0x204040a8,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x00404164,
-	0x0c4041a0,
-	0x0c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x0c404618,
-	0x0440462c,
-	0x00404640,
-	0x00404654,
-	0x00404660,
-	0x48404678,
-	0x084046c8,
-	0x08404700,
-	0x24404718,
-	0x04404744,
-	0x00404754,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00405b00,
-	0x00405b10,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x044064b4,
-	0x2c4064c0,
-	0x004064fc,
-	0x00407040,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x00408840,
-	0x08408900,
-	0x00408980,
-/* 0x0144: nve4_hub_mmio_tail */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -127,6 +69,47 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -136,10 +119,7 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0200: chan_data */
-/* 0x0200: chan_mmio_count */
 	0x00000000,
-/* 0x0204: chan_mmio_address */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -156,6 +136,7 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0200: xfer_data */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -203,19 +184,36 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0300: xfer_data */
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
 };
 
 uint32_t nve0_grhub_code[] = {
-	0x03090ef5,
+	0x031b0ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -247,7 +245,7 @@ uint32_t nve0_grhub_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -269,63 +267,66 @@ uint32_t nve0_grhub_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -333,8 +334,8 @@ uint32_t nve0_grhub_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -343,76 +344,77 @@ uint32_t nve0_grhub_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -421,73 +423,61 @@ uint32_t nve0_grhub_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe4b60814,
-	0x00efd006,
-	0x0c1ce7f1,
-	0xf006e4b6,
-	0xefd001f7,
-	0xf8e0fc00,
-/* 0x0309: init */
-	0xfe04bd00,
-	0x07fe0004,
-	0x0017f100,
-	0x0227f012,
-	0xf10012d0,
-	0xfe05b917,
-	0x17f10010,
-	0x10d00400,
-	0x0437f1c0,
-	0x0634b604,
-	0x200327f1,
-	0xf10032d0,
-	0xd0200427,
-	0x27f10132,
-	0x32d0200b,
-	0x0c27f102,
-	0x0732d020,
-	0x0c2427f1,
-	0xb90624b6,
-	0x23d00003,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
 	0x0427f100,
-	0x0023f087,
-	0xb70012d0,
-	0xf0010012,
-	0x12d00427,
-	0x1031f400,
-	0x9604e7f1,
-	0xf440e3f0,
-	0xf1c76821,
-	0x01018090,
-	0x801ff4f0,
-	0x17f0000f,
-	0x041fbb01,
-	0xf10112b6,
-	0xb6040c27,
-	0x21d00624,
-	0x4021d000,
-	0x080027f1,
-	0xcf0624b6,
-	0xf7f00022,
-/* 0x03a9: init_find_chipset */
-	0x08f0b654,
-	0xb800f398,
-	0x0bf40432,
-	0x0034b00b,
-	0xf8f11bf4,
-/* 0x03bd: init_context */
-	0x0017f100,
-	0x02fe5801,
-	0xf003ff58,
-	0x0e8000e3,
-	0x150f8014,
-	0x013d21f5,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
 	0x070037f1,
 	0x950634b6,
 	0x34d00814,
@@ -498,196 +488,201 @@ uint32_t nve0_grhub_code[] = {
 	0x0815b600,
 	0xb60110b6,
 	0x1fb90814,
-	0x6321f502,
+	0x7121f502,
 	0x001fbb02,
-	0xf1000398,
+	0xf1020398,
 	0xf0200047,
-/* 0x040e: init_gpc */
+/* 0x03f6: init_gpc */
 	0x4ea05043,
 	0x1fb90804,
 	0x8d21f402,
-	0x08004ea0,
-	0xf4022fb9,
-	0x4ea08d21,
-	0xf4bd010c,
-	0xa08d21f4,
-	0xf401044e,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
 	0x4ea08d21,
-	0xf7f00100,
-	0x8d21f402,
-	0x08004ea0,
-/* 0x0440: init_gpc_wait */
-	0xc86821f4,
-	0x0bf41fff,
-	0x044ea0fa,
-	0x6821f408,
-	0xb7001fbb,
-	0xb6800040,
-	0x1bf40132,
-	0x0027f1b4,
-	0x0624b608,
-	0xb74021d0,
-	0xbd080020,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
 	0x1f19f014,
-/* 0x0473: main */
-	0xf40021d0,
-	0x28f40031,
-	0x08d7f000,
-	0xf43921f4,
-	0xe4b1f401,
-	0x1bf54001,
-	0x87f100d1,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00499,
-	0x0017f100,
-	0x0614b60b,
-	0xcf4012cf,
-	0x13c80011,
-	0x7e0bf41f,
+	0x080007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
 	0xf41f23c8,
-	0x20f95a0b,
-	0xf10212b9,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00799f0,
-	0x32f40089,
-	0x0231f401,
-	0x07fb21f5,
-	0x085c87f1,
-	0xbd0684b6,
+	0x20f9620b,
+	0xbd0212b9,
 	0x0799f094,
-	0xfc0089d0,
-	0x3c87f120,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d006,
-	0xf50131f4,
-	0xf107fb21,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00699f0,
-	0x0ef40089,
-/* 0x0509: chsw_prev_no_next */
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f00f00,
+	0x0009d002,
+	0x31f404bd,
+	0x0121f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
 	0xb920f931,
 	0x32f40212,
 	0x0232f401,
-	0x07fb21f5,
+	0x080121f5,
 	0x17f120fc,
 	0x14b60b00,
 	0x0012d006,
-/* 0x0527: chsw_no_prev */
+/* 0x0517: chsw_no_prev */
 	0xc8130ef4,
 	0x0bf41f23,
 	0x0131f40d,
 	0xf50232f4,
-/* 0x0537: chsw_done */
-	0xf107fb21,
+/* 0x0527: chsw_done */
+	0xf1080121,
 	0xb60b0c17,
 	0x27f00614,
 	0x0012d001,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0499f094,
-	0xf50089d0,
-/* 0x0557: main_not_ctx_switch */
-	0xb0ff200e,
-	0x1bf401e4,
-	0x02f2b90d,
-	0x078f21f5,
-/* 0x0567: main_not_ctx_chan */
-	0xb0420ef4,
-	0x1bf402e4,
-	0x3c87f12e,
-	0x0684b608,
 	0x99f094bd,
-	0x0089d007,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef40795,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
 	0xf40132f4,
 	0x21f50232,
-	0x87f107fb,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00799,
-	0x110ef400,
-/* 0x0598: main_not_ctx_save */
-	0xf010ef94,
-	0x21f501f5,
-	0x0ef502ec,
-/* 0x05a6: main_done */
-	0x17f1fed1,
-	0x14b60820,
-	0xf024bd06,
-	0x12d01f29,
-	0xbe0ef500,
-/* 0x05b9: ih */
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
 	0xfe80f9fe,
 	0x80f90188,
 	0xa0f990f9,
 	0xd0f9b0f9,
 	0xf0f9e0f9,
-	0xc4800acf,
-	0x0bf404ab,
-	0x00b7f11d,
-	0x08d7f019,
-	0xcf40becf,
-	0x21f400bf,
-	0x00b0b704,
-	0x01e7f004,
-/* 0x05ef: ih_no_fifo */
-	0xe400bed0,
-	0xf40100ab,
-	0xd7f00d0b,
-	0x01e7f108,
-	0x0421f440,
-/* 0x0600: ih_no_ctxsw */
-	0x0104b7f1,
-	0xabffb0bd,
-	0x0d0bf4b4,
-	0x0c1ca7f1,
-	0xd006a4b6,
-/* 0x0616: ih_no_other */
-	0x0ad000ab,
-	0xfcf0fc40,
-	0xfcd0fce0,
-	0xfca0fcb0,
-	0xfe80fc90,
-	0x80fc0088,
-	0xf80032f4,
-/* 0x0631: ctx_4170s */
-	0x70e7f101,
-	0x40e3f041,
-	0xf410f5f0,
-	0x00f88d21,
-/* 0x0640: ctx_4170w */
-	0x4170e7f1,
-	0xf440e3f0,
-	0xf4f06821,
-	0xf31bf410,
-/* 0x0652: ctx_redswitch */
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4170s */
+	0xf101f800,
+	0xf04170e7,
+	0xf5f040e3,
+	0x8d21f410,
+/* 0x063a: ctx_4170w */
 	0xe7f100f8,
-	0xe4b60614,
-	0x70f7f106,
-	0x00efd002,
-/* 0x0663: ctx_redswitch_delay */
-	0xb608f7f0,
-	0x1bf401f2,
-	0x70f7f1fd,
-	0x00efd007,
-/* 0x0672: ctx_86c */
-	0xe7f100f8,
-	0xe4b6086c,
-	0x00efd006,
-	0x8a14e7f1,
-	0xf440e3f0,
-	0xe7f18d21,
-	0xe3f0a86c,
-	0x8d21f441,
-/* 0x0692: ctx_load */
-	0x87f100f8,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00599,
-	0x0ca7f000,
+	0xe3f04170,
+	0x6821f440,
+	0xf410f4f0,
+	0x00f8f31b,
+/* 0x064c: ctx_redswitch */
+	0x0614e7f1,
+	0xf106e4b6,
+	0xd00270f7,
+	0xf7f000ef,
+/* 0x065d: ctx_redswitch_delay */
+	0x01f2b608,
+	0xf1fd1bf4,
+	0xd00770f7,
+	0x00f800ef,
+/* 0x066c: ctx_86c */
+	0x086ce7f1,
+	0xd006e4b6,
+	0xe7f100ef,
+	0xe3f08a14,
+	0x8d21f440,
+	0xa86ce7f1,
+	0xf441e3f0,
+	0x00f88d21,
+/* 0x068c: ctx_load */
+	0x99f094bd,
+	0x0007f105,
+	0x0203f00f,
+	0xbd0009d0,
+	0x0ca7f004,
 	0xf1c921f4,
 	0xb60a2417,
 	0x10d00614,
@@ -697,162 +692,227 @@ uint32_t nve0_grhub_code[] = {
 	0xb60a0c17,
 	0x47f00614,
 	0x0012d007,
-/* 0x06cb: ctx_chan_wait_0 */
+/* 0x06c7: ctx_chan_wait_0 */
 	0xcf4014d0,
 	0x44f04014,
 	0xfa1bf41f,
 	0xfe0032d0,
 	0x2af0000b,
 	0x0424b61f,
-	0xf10220b6,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00899f0,
-	0x17f10089,
-	0x14b60a04,
-	0x0012d006,
-	0x0a2017f1,
-	0xf00614b6,
-	0x23f10227,
-	0x12d08000,
-	0x1017f000,
-	0x030027f1,
-	0xfa0223f0,
-	0x03f80512,
-	0x085c87f1,
-	0xbd0684b6,
+	0xbd0220b6,
 	0x0899f094,
-	0x980089d0,
-	0x14b6c101,
-	0xc0029818,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0417f1,
+	0xd00614b6,
+	0x17f10012,
+	0x14b60a20,
+	0x0227f006,
+	0x800023f1,
+	0xf00012d0,
+	0x27f11017,
+	0x23f00200,
+	0x0512fa02,
+	0x94bd03f8,
+	0xf10899f0,
+	0xf0170007,
+	0x09d00203,
+	0x9804bd00,
+	0x14b68101,
+	0x80029818,
 	0xfd0825b6,
 	0x01800512,
-	0x3c87f116,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d009,
-	0x0a0427f1,
-	0xd00624b6,
-	0x27f00021,
-	0x2017f101,
-	0x0614b60a,
-	0xf10012d0,
-	0xf0020017,
+	0xf094bd16,
+	0x07f10999,
+	0x03f00f00,
+	0x0009d002,
+	0x27f104bd,
+	0x24b60a04,
+	0x0021d006,
+	0xf10127f0,
+	0xb60a2017,
+	0x12d00614,
+	0x0017f100,
+	0x0613f001,
+	0xf80501fa,
+	0xf094bd03,
+	0x07f10999,
+	0x03f01700,
+	0x0009d002,
+	0x94bd04bd,
+	0xf10599f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0795: ctx_chan */
+	0x8c21f500,
+	0x0ca7f006,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x07ac: ctx_chan_wait */
+	0xfd0012cf,
+	0x1bf40522,
+/* 0x07b7: ctx_mmio_exec */
+	0x9800f8fa,
+	0x27f14103,
+	0x24b60a04,
+	0x0023d006,
+/* 0x07c6: ctx_mmio_loop */
+	0x34c434bd,
+	0x0f1bf4ff,
+	0x020057f1,
+	0xfa0653f0,
+	0x03f80535,
+/* 0x07d8: ctx_mmio_pull */
+	0x98804e98,
+	0x21f4814f,
+	0x0830b68d,
+	0xf40112b6,
+/* 0x07ea: ctx_mmio_done */
+	0x0398df1b,
+	0x0023d016,
+	0xf1400080,
+	0xf0010017,
 	0x01fa0613,
-	0xf103f805,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00999f0,
-	0x87f10089,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00599,
-/* 0x078f: ctx_chan */
-	0xf500f800,
-	0xf0069221,
-	0x21f40ca7,
-	0x1017f1c9,
-	0x0614b60a,
-	0xd00527f0,
-/* 0x07a6: ctx_chan_wait */
-	0x12cf0012,
-	0x0522fd00,
-	0xf8fa1bf4,
-/* 0x07b1: ctx_mmio_exec */
-	0x81039800,
-	0x0a0427f1,
+	0xf803f806,
+/* 0x0801: ctx_xfer */
+	0x00f7f100,
+	0x06f4b60c,
+	0xd004e7f0,
+/* 0x080e: ctx_xfer_idle */
+	0xfecf80fe,
+	0x00e4f100,
+	0xf91bf420,
+	0xf40611f4,
+/* 0x081e: ctx_xfer_pre */
+	0xf7f00d02,
+	0x6c21f510,
+	0x1c11f406,
+/* 0x0828: ctx_xfer_pre_load */
+	0xf502f7f0,
+	0xf5062b21,
+	0xf5063a21,
+	0xbd064c21,
+	0x2b21f5f4,
+	0x8c21f506,
+/* 0x0841: ctx_xfer_exec */
+	0x16019806,
+	0x041427f1,
 	0xd00624b6,
-	0x34bd0023,
-/* 0x07c0: ctx_mmio_loop */
-	0xf4ff34c4,
-	0x57f10f1b,
-	0x53f00300,
-	0x0535fa06,
-/* 0x07d2: ctx_mmio_pull */
-	0x4e9803f8,
-	0xc14f98c0,
+	0xe7f10020,
+	0xe3f0a500,
+	0x021fb941,
 	0xb68d21f4,
-	0x12b60830,
-	0xdf1bf401,
-/* 0x07e4: ctx_mmio_done */
-	0xd0160398,
-	0x00800023,
-	0x0017f180,
-	0x0613f002,
-	0xf80601fa,
-/* 0x07fb: ctx_xfer */
-	0xf100f803,
-	0xb60c00f7,
-	0xe7f006f4,
-	0x80fed004,
-/* 0x0808: ctx_xfer_idle */
-	0xf100fecf,
-	0xf42000e4,
-	0x11f4f91b,
-	0x0d02f406,
-/* 0x0818: ctx_xfer_pre */
-	0xf510f7f0,
-	0xf4067221,
-/* 0x0822: ctx_xfer_pre_load */
-	0xf7f01c11,
-	0x3121f502,
-	0x4021f506,
-	0x5221f506,
-	0xf5f4bd06,
-	0xf5063121,
-/* 0x083b: ctx_xfer_exec */
-	0x98069221,
-	0x27f11601,
-	0x24b60414,
-	0x0020d006,
-	0xa500e7f1,
-	0xb941e3f0,
-	0x21f4021f,
-	0x04e0b68d,
-	0xf001fcf0,
-	0x24b6022c,
-	0x05f2fd01,
-	0xf18d21f4,
-	0xf04afc17,
-	0x27f00213,
-	0x0012d00c,
-	0x020721f5,
-	0x47fc27f1,
-	0xd00223f0,
-	0x2cf00020,
-	0x0320b601,
-	0xf00012d0,
-	0xa5f001ac,
-	0x00b7f006,
-	0x98140c98,
-	0xe7f0150d,
-	0x5c21f500,
-	0x08a7f001,
-	0x010321f5,
-	0x020721f5,
-	0xf02201f4,
-	0x21f40ca7,
-	0x1017f1c9,
-	0x0614b60a,
-	0xd00527f0,
-/* 0x08c2: ctx_xfer_post_save_wait */
-	0x12cf0012,
-	0x0522fd00,
-	0xf4fa1bf4,
-/* 0x08ce: ctx_xfer_post */
-	0xf7f02e02,
-	0x3121f502,
+	0xfcf004e0,
+	0x022cf001,
+	0xfd0124b6,
+	0x21f405f2,
+	0xfc17f18d,
+	0x0213f04a,
+	0xd00c27f0,
+	0x21f50012,
+	0x27f10215,
+	0x23f047fc,
+	0x0020d002,
+	0xb6012cf0,
+	0x12d00320,
+	0x01acf000,
+	0xf006a5f0,
+	0x0c9800b7,
+	0x010d9800,
+	0xf500e7f0,
+	0xf0016621,
+	0x21f508a7,
+	0x21f50109,
+	0x01f40215,
+	0x0ca7f022,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x08c8: ctx_xfer_post_save_wait */
+	0xfd0012cf,
+	0x1bf40522,
+	0x2e02f4fa,
+/* 0x08d4: ctx_xfer_post */
+	0xf502f7f0,
+	0xbd062b21,
+	0x6c21f5f4,
+	0x3421f506,
+	0x3a21f502,
 	0xf5f4bd06,
-	0xf5067221,
-	0xf5022621,
-	0xbd064021,
-	0x3121f5f4,
-	0x1011f406,
-	0xfd800198,
-	0x0bf40511,
-	0xb121f507,
-/* 0x08f9: ctx_xfer_no_post_mmio */
-/* 0x08f9: ctx_xfer_done */
-	0x0000f807,
+	0xf4062b21,
+	0x01981011,
+	0x0511fd40,
+	0xf5070bf4,
+/* 0x08ff: ctx_xfer_no_post_mmio */
+/* 0x08ff: ctx_xfer_done */
+	0xf807b721,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc
new file mode 100644
index 0000000..ec42ed2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define CHIPSET GK110
+#include "macros.fuc"
+
+.section #nvf0_grhub_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
+
+.section #nvf0_grhub_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "hub.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc.h
new file mode 100644
index 0000000..438506d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc.h
@@ -0,0 +1,918 @@
+uint32_t nvf0_grhub_data[] = {
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
+	0x00000000,
+/* 0x000c: rop_count */
+	0x00000000,
+/* 0x0010: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0058: ctx_current */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0200: xfer_data */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
+};
+
+uint32_t nvf0_grhub_code[] = {
+	0x031b0ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf0370007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f03700,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
+	0x0427f100,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
+	0x070037f1,
+	0x950634b6,
+	0x34d00814,
+	0x4034d000,
+	0x130030b7,
+	0xb6001fbb,
+	0x3fd002f5,
+	0x0815b600,
+	0xb60110b6,
+	0x1fb90814,
+	0x7121f502,
+	0x001fbb02,
+	0xf1020398,
+	0xf0200047,
+/* 0x03f6: init_gpc */
+	0x4ea05043,
+	0x1fb90804,
+	0x8d21f402,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
+	0x4ea08d21,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
+	0x1f19f014,
+	0x300007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
+	0xf41f23c8,
+	0x20f9620b,
+	0xbd0212b9,
+	0x0799f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f03700,
+	0x0009d002,
+	0x31f404bd,
+	0x0121f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
+	0xb920f931,
+	0x32f40212,
+	0x0232f401,
+	0x080121f5,
+	0x17f120fc,
+	0x14b60b00,
+	0x0012d006,
+/* 0x0517: chsw_no_prev */
+	0xc8130ef4,
+	0x0bf41f23,
+	0x0131f40d,
+	0xf50232f4,
+/* 0x0527: chsw_done */
+	0xf1080121,
+	0xb60b0c17,
+	0x27f00614,
+	0x0012d001,
+	0x99f094bd,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef40795,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50232,
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f030,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
+	0xfe80f9fe,
+	0x80f90188,
+	0xa0f990f9,
+	0xd0f9b0f9,
+	0xf0f9e0f9,
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4170s */
+	0xf101f800,
+	0xf04170e7,
+	0xf5f040e3,
+	0x8d21f410,
+/* 0x063a: ctx_4170w */
+	0xe7f100f8,
+	0xe3f04170,
+	0x6821f440,
+	0xf410f4f0,
+	0x00f8f31b,
+/* 0x064c: ctx_redswitch */
+	0x0614e7f1,
+	0xf106e4b6,
+	0xd00270f7,
+	0xf7f000ef,
+/* 0x065d: ctx_redswitch_delay */
+	0x01f2b608,
+	0xf1fd1bf4,
+	0xd00770f7,
+	0x00f800ef,
+/* 0x066c: ctx_86c */
+	0x086ce7f1,
+	0xd006e4b6,
+	0xe7f100ef,
+	0xe3f08a14,
+	0x8d21f440,
+	0xa86ce7f1,
+	0xf441e3f0,
+	0x00f88d21,
+/* 0x068c: ctx_load */
+	0x99f094bd,
+	0x0007f105,
+	0x0203f037,
+	0xbd0009d0,
+	0x0ca7f004,
+	0xf1c921f4,
+	0xb60a2417,
+	0x10d00614,
+	0x0037f100,
+	0x0634b60b,
+	0xf14032d0,
+	0xb60a0c17,
+	0x47f00614,
+	0x0012d007,
+/* 0x06c7: ctx_chan_wait_0 */
+	0xcf4014d0,
+	0x44f04014,
+	0xfa1bf41f,
+	0xfe0032d0,
+	0x2af0000b,
+	0x0424b61f,
+	0xbd0220b6,
+	0x0899f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0417f1,
+	0xd00614b6,
+	0x17f10012,
+	0x14b60a20,
+	0x0227f006,
+	0x800023f1,
+	0xf00012d0,
+	0x27f11017,
+	0x23f00200,
+	0x0512fa02,
+	0x94bd03f8,
+	0xf10899f0,
+	0xf0170007,
+	0x09d00203,
+	0x9804bd00,
+	0x14b68101,
+	0x80029818,
+	0xfd0825b6,
+	0x01800512,
+	0xf094bd16,
+	0x07f10999,
+	0x03f03700,
+	0x0009d002,
+	0x27f104bd,
+	0x24b60a04,
+	0x0021d006,
+	0xf10127f0,
+	0xb60a2017,
+	0x12d00614,
+	0x0017f100,
+	0x0613f001,
+	0xf80501fa,
+	0xf094bd03,
+	0x07f10999,
+	0x03f01700,
+	0x0009d002,
+	0x94bd04bd,
+	0xf10599f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0795: ctx_chan */
+	0x8c21f500,
+	0x0ca7f006,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x07ac: ctx_chan_wait */
+	0xfd0012cf,
+	0x1bf40522,
+/* 0x07b7: ctx_mmio_exec */
+	0x9800f8fa,
+	0x27f14103,
+	0x24b60a04,
+	0x0023d006,
+/* 0x07c6: ctx_mmio_loop */
+	0x34c434bd,
+	0x0f1bf4ff,
+	0x020057f1,
+	0xfa0653f0,
+	0x03f80535,
+/* 0x07d8: ctx_mmio_pull */
+	0x98804e98,
+	0x21f4814f,
+	0x0830b68d,
+	0xf40112b6,
+/* 0x07ea: ctx_mmio_done */
+	0x0398df1b,
+	0x0023d016,
+	0xf1400080,
+	0xf0010017,
+	0x01fa0613,
+	0xf803f806,
+/* 0x0801: ctx_xfer */
+	0x00f7f100,
+	0x06f4b60c,
+	0xd004e7f0,
+/* 0x080e: ctx_xfer_idle */
+	0xfecf80fe,
+	0x00e4f100,
+	0xf91bf420,
+	0xf40611f4,
+/* 0x081e: ctx_xfer_pre */
+	0xf7f00d02,
+	0x6c21f510,
+	0x1c11f406,
+/* 0x0828: ctx_xfer_pre_load */
+	0xf502f7f0,
+	0xf5062b21,
+	0xf5063a21,
+	0xbd064c21,
+	0x2b21f5f4,
+	0x8c21f506,
+/* 0x0841: ctx_xfer_exec */
+	0x16019806,
+	0x041427f1,
+	0xd00624b6,
+	0xe7f10020,
+	0xe3f0a500,
+	0x021fb941,
+	0xb68d21f4,
+	0xfcf004e0,
+	0x022cf001,
+	0xfd0124b6,
+	0x21f405f2,
+	0xfc17f18d,
+	0x0213f04a,
+	0xd00c27f0,
+	0x21f50012,
+	0x27f10215,
+	0x23f047fc,
+	0x0020d002,
+	0xb6012cf0,
+	0x12d00320,
+	0x01acf000,
+	0xf006a5f0,
+	0x0c9800b7,
+	0x010d9800,
+	0xf500e7f0,
+	0xf0016621,
+	0x21f508a7,
+	0x21f50109,
+	0x01f40215,
+	0x0ca7f022,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x08c8: ctx_xfer_post_save_wait */
+	0xfd0012cf,
+	0x1bf40522,
+	0x2e02f4fa,
+/* 0x08d4: ctx_xfer_post */
+	0xf502f7f0,
+	0xbd062b21,
+	0x6c21f5f4,
+	0x3421f506,
+	0x3a21f502,
+	0xf5f4bd06,
+	0xf4062b21,
+	0x01981011,
+	0x0511fd40,
+	0xf5070bf4,
+/* 0x08ff: ctx_xfer_no_post_mmio */
+/* 0x08ff: ctx_xfer_done */
+	0xf807b721,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/macros.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/macros.fuc
new file mode 100644
index 0000000..33a5a82
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/macros.fuc
@@ -0,0 +1,89 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "os.h"
+
+#define GF100 0xc0
+#define GF117 0xd7
+#define GK100 0xe0
+#define GK110 0xf0
+
+#define NV_PGRAPH_FECS_SIGNAL                                          0x409400
+#if CHIPSET < GK110
+#define NV_PGRAPH_FECS_CC_SCRATCH_VAL(n)                    ((n) * 4 + 0x409800)
+#define NV_PGRAPH_FECS_CC_SCRATCH_SET(n)                    ((n) * 4 + 0x409820)
+#define NV_PGRAPH_FECS_CC_SCRATCH_CLR(n)                    ((n) * 4 + 0x409840)
+#else
+#define NV_PGRAPH_FECS_CC_SCRATCH_VAL(n)                    ((n) * 4 + 0x409800)
+#define NV_PGRAPH_FECS_CC_SCRATCH_CLR(n)                    ((n) * 4 + 0x409840)
+#define NV_PGRAPH_FECS_CC_SCRATCH_SET(n)                    ((n) * 4 + 0x4098c0)
+#endif
+#define NV_PGRAPH_FECS_INTR_UP_SET                                     0x409c1c
+
+#if CHIPSET < GK110
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(n)              ((n) * 4 + 0x41a800)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_SET(n)              ((n) * 4 + 0x41a820)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_CLR(n)              ((n) * 4 + 0x41a840)
+#else
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(n)              ((n) * 4 + 0x41a800)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_CLR(n)              ((n) * 4 + 0x41a840)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_SET(n)              ((n) * 4 + 0x41a8c0)
+#endif
+
+#define mmctx_data(r,c) .b32 (((c - 1) << 26) | r)
+#define queue_init      .skip 72 // (2 * 4) + ((8 * 4) * 2)
+
+#define T_WAIT    0
+#define T_MMCTX   1
+#define T_STRWAIT 2
+#define T_STRINIT 3
+#define T_AUTO    4
+#define T_CHAN    5
+#define T_LOAD    6
+#define T_SAVE    7
+#define T_LCHAN   8
+#define T_LCTXH   9
+
+#define nv_mkmm(rv,r) /*
+*/	movw rv  ((r) & 0x0000fffc) /*
+*/	sethi rv ((r) & 0x00ff0000)
+#define nv_mkio(rv,r,i) /*
+*/	nv_mkmm(rv, (((r) & 0xffc) << 6) | ((i) << 2))
+
+#define nv_iord(rv,r,i) /*
+*/	nv_mkio(rv,r,i) /*
+*/	iord rv I[rv]
+#define nv_iowr(r,i,rv) /*
+*/	nv_mkio($r0,r,i) /*
+*/	iowr I[$r0] rv /*
+*/	clear b32 $r0
+
+#define trace_set(bit) /*
+*/	clear b32 $r9 /*
+*/	bset $r9 bit /*
+*/	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_SET(7), 0, $r9)
+#define trace_clr(bit) /*
+*/	clear b32 $r9 /*
+*/	bset $r9 bit /*
+*/	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_CLR(7), 0, $r9)
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nve0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nve0.fuc
deleted file mode 100644
index f16a5d5..0000000
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nve0.fuc
+++ /dev/null
@@ -1,400 +0,0 @@
-/* fuc microcode util functions for nve0 PGRAPH
- *
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-define(`mmctx_data', `.b32 eval((($2 - 1) << 26) | $1)')
-define(`queue_init', `.skip eval((2 * 4) + ((8 * 4) * 2))')
-
-ifdef(`include_code', `
-// Error codes
-define(`E_BAD_COMMAND', 0x01)
-define(`E_CMD_OVERFLOW', 0x02)
-
-// Util macros to help with debugging ucode hangs etc
-define(`T_WAIT', 0)
-define(`T_MMCTX', 1)
-define(`T_STRWAIT', 2)
-define(`T_STRINIT', 3)
-define(`T_AUTO', 4)
-define(`T_CHAN', 5)
-define(`T_LOAD', 6)
-define(`T_SAVE', 7)
-define(`T_LCHAN', 8)
-define(`T_LCTXH', 9)
-
-define(`trace_set', `
-	mov $r8 0x83c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-define(`trace_clr', `
-	mov $r8 0x85c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-// queue_put - add request to queue
-//
-// In : $r13 queue pointer
-//	$r14 command
-//	$r15 data
-//
-queue_put:
-	// make sure we have space..
-	ld b32 $r8 D[$r13 + 0x0]	// GET
-	ld b32 $r9 D[$r13 + 0x4]	// PUT
-	xor $r8 8
-	cmpu b32 $r8 $r9
-	bra ne #queue_put_next
-		mov $r15 E_CMD_OVERFLOW
-		call #error
-		ret
-
-	// store cmd/data on queue
-	queue_put_next:
-	and $r8 $r9 7
-	shl b32 $r8 3
-	add b32 $r8 $r13
-	add b32 $r8 8
-	st b32 D[$r8 + 0x0] $r14
-	st b32 D[$r8 + 0x4] $r15
-
-	// update PUT
-	add b32 $r9 1
-	and $r9 0xf
-	st b32 D[$r13 + 0x4] $r9
-	ret
-
-// queue_get - fetch request from queue
-//
-// In : $r13 queue pointer
-//
-// Out:	$p1  clear on success (data available)
-//	$r14 command
-// 	$r15 data
-//
-queue_get:
-	bset $flags $p1
-	ld b32 $r8 D[$r13 + 0x0]	// GET
-	ld b32 $r9 D[$r13 + 0x4]	// PUT
-	cmpu b32 $r8 $r9
-	bra e #queue_get_done
-		// fetch first cmd/data pair
-		and $r9 $r8 7
-		shl b32 $r9 3
-		add b32 $r9 $r13
-		add b32 $r9 8
-		ld b32 $r14 D[$r9 + 0x0]
-		ld b32 $r15 D[$r9 + 0x4]
-
-		// update GET
-		add b32 $r8 1
-		and $r8 0xf
-		st b32 D[$r13 + 0x0] $r8
-		bclr $flags $p1
-queue_get_done:
-	ret
-
-// nv_rd32 - read 32-bit value from nv register
-//
-// In : $r14 register
-// Out: $r15 value
-//
-nv_rd32:
-	mov $r11 0x728
-	shl b32 $r11 6
-	mov b32 $r12 $r14
-	bset $r12 31			// MMIO_CTRL_PENDING
-	iowr I[$r11 + 0x000] $r12	// MMIO_CTRL
-	nv_rd32_wait:
-		iord $r12 I[$r11 + 0x000]
-		xbit $r12 $r12 31
-		bra ne #nv_rd32_wait
-	mov $r10 6			// DONE_MMIO_RD
-	call #wait_doneo
-	iord $r15 I[$r11 + 0x100]	// MMIO_RDVAL
-	ret
-
-// nv_wr32 - write 32-bit value to nv register
-//
-// In : $r14 register
-//      $r15 value
-//
-nv_wr32:
-	mov $r11 0x728
-	shl b32 $r11 6
-	iowr I[$r11 + 0x200] $r15	// MMIO_WRVAL
-	mov b32 $r12 $r14
-	bset $r12 31			// MMIO_CTRL_PENDING
-	bset $r12 30			// MMIO_CTRL_WRITE
-	iowr I[$r11 + 0x000] $r12	// MMIO_CTRL
-	nv_wr32_wait:
-		iord $r12 I[$r11 + 0x000]
-		xbit $r12 $r12 31
-		bra ne #nv_wr32_wait
-	ret
-
-// (re)set watchdog timer
-//
-// In : $r15 timeout
-//
-watchdog_reset:
-	mov $r8 0x430
-	shl b32 $r8 6
-	bset $r15 31
-	iowr I[$r8 + 0x000] $r15
-	ret
-
-// clear watchdog timer
-watchdog_clear:
-	mov $r8 0x430
-	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r0
-	ret
-
-// wait_done{z,o} - wait on FUC_DONE bit to become clear/set
-//
-// In : $r10 bit to wait on
-//
-define(`wait_done', `
-$1:
-	trace_set(T_WAIT);
-	mov $r8 0x818
-	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r10	// CC_SCRATCH[6] = wait bit
-	wait_done_$1:
-		mov $r8 0x400
-		shl b32 $r8 6
-		iord $r8 I[$r8 + 0x000]	// DONE
-		xbit $r8 $r8 $r10
-		bra $2 #wait_done_$1
-	trace_clr(T_WAIT)
-	ret
-')
-wait_done(wait_donez, ne)
-wait_done(wait_doneo, e)
-
-// mmctx_size - determine size of a mmio list transfer
-//
-// In : $r14 mmio list head
-//      $r15 mmio list tail
-// Out: $r15 transfer size (in bytes)
-//
-mmctx_size:
-	clear b32 $r9
-	nv_mmctx_size_loop:
-		ld b32 $r8 D[$r14]
-		shr b32 $r8 26
-		add b32 $r8 1
-		shl b32 $r8 2
-		add b32 $r9 $r8
-		add b32 $r14 4
-		cmpu b32 $r14 $r15
-		bra ne #nv_mmctx_size_loop
-	mov b32 $r15 $r9
-	ret
-
-// mmctx_xfer - execute a list of mmio transfers
-//
-// In : $r10 flags
-//		bit 0: direction (0 = save, 1 = load)
-//		bit 1: set if first transfer
-//		bit 2: set if last transfer
-//	$r11 base
-//	$r12 mmio list head
-//	$r13 mmio list tail
-//	$r14 multi_stride
-//	$r15 multi_mask
-//
-mmctx_xfer:
-	trace_set(T_MMCTX)
-	mov $r8 0x710
-	shl b32 $r8 6
-	clear b32 $r9
-	or $r11 $r11
-	bra e #mmctx_base_disabled
-		iowr I[$r8 + 0x000] $r11	// MMCTX_BASE
-		bset $r9 0			// BASE_EN
-	mmctx_base_disabled:
-	or $r14 $r14
-	bra e #mmctx_multi_disabled
-		iowr I[$r8 + 0x200] $r14 	// MMCTX_MULTI_STRIDE
-		iowr I[$r8 + 0x300] $r15 	// MMCTX_MULTI_MASK
-		bset $r9 1			// MULTI_EN
-	mmctx_multi_disabled:
-	add b32 $r8 0x100
-
-	xbit $r11 $r10 0
-	shl b32 $r11 16			// DIR
-	bset $r11 12			// QLIMIT = 0x10
-	xbit $r14 $r10 1
-	shl b32 $r14 17
-	or $r11 $r14			// START_TRIGGER
-	iowr I[$r8 + 0x000] $r11	// MMCTX_CTRL
-
-	// loop over the mmio list, and send requests to the hw
-	mmctx_exec_loop:
-		// wait for space in mmctx queue
-		mmctx_wait_free:
-			iord $r14 I[$r8 + 0x000] // MMCTX_CTRL
-			and $r14 0x1f
-			bra e #mmctx_wait_free
-
-		// queue up an entry
-		ld b32 $r14 D[$r12]
-		or $r14 $r9
-		iowr I[$r8 + 0x300] $r14
-		add b32 $r12 4
-		cmpu b32 $r12 $r13
-		bra ne #mmctx_exec_loop
-
-	xbit $r11 $r10 2
-	bra ne #mmctx_stop
-		// wait for queue to empty
-		mmctx_fini_wait:
-			iord $r11 I[$r8 + 0x000]	// MMCTX_CTRL
-			and $r11 0x1f
-			cmpu b32 $r11 0x10
-			bra ne #mmctx_fini_wait
-		mov $r10 2				// DONE_MMCTX
-		call #wait_donez
-		bra #mmctx_done
-	mmctx_stop:
-		xbit $r11 $r10 0
-		shl b32 $r11 16			// DIR
-		bset $r11 12			// QLIMIT = 0x10
-		bset $r11 18			// STOP_TRIGGER
-		iowr I[$r8 + 0x000] $r11	// MMCTX_CTRL
-		mmctx_stop_wait:
-			// wait for STOP_TRIGGER to clear
-			iord $r11 I[$r8 + 0x000] // MMCTX_CTRL
-			xbit $r11 $r11 18
-			bra ne #mmctx_stop_wait
-	mmctx_done:
-	trace_clr(T_MMCTX)
-	ret
-
-// Wait for DONE_STRAND
-//
-strand_wait:
-	push $r10
-	mov $r10 2
-	call #wait_donez
-	pop $r10
-	ret
-
-// unknown - call before issuing strand commands
-//
-strand_pre:
-	mov $r8 0x4afc
-	sethi $r8 0x20000
-	mov $r9 0xc
-	iowr I[$r8] $r9
-	call #strand_wait
-	ret
-
-// unknown - call after issuing strand commands
-//
-strand_post:
-	mov $r8 0x4afc
-	sethi $r8 0x20000
-	mov $r9 0xd
-	iowr I[$r8] $r9
-	call #strand_wait
-	ret
-
-// Selects strand set?!
-//
-// In: $r14 id
-//
-strand_set:
-	mov $r10 0x4ffc
-	sethi $r10 0x20000
-	sub b32 $r11 $r10 0x500
-	mov $r12 0xf
-	iowr I[$r10 + 0x000] $r12		// 0x93c = 0xf
-	mov $r12 0xb
-	iowr I[$r11 + 0x000] $r12		// 0x928 = 0xb
-	call #strand_wait
-	iowr I[$r10 + 0x000] $r14		// 0x93c = <id>
-	mov $r12 0xa
-	iowr I[$r11 + 0x000] $r12		// 0x928 = 0xa
-	call #strand_wait
-	ret
-
-// Initialise strand context data
-//
-// In : $r15 context base
-// Out: $r15 context size (in bytes)
-//
-// Strandset(?) 3 hardcoded currently
-//
-strand_ctx_init:
-	trace_set(T_STRINIT)
-	call #strand_pre
-	mov $r14 3
-	call #strand_set
-	mov $r10 0x46fc
-	sethi $r10 0x20000
-	add b32 $r11 $r10 0x400
-	iowr I[$r10 + 0x100] $r0	// STRAND_FIRST_GENE = 0
-	mov $r12 1
-	iowr I[$r11 + 0x000] $r12	// STRAND_CMD = LATCH_FIRST_GENE
-	call #strand_wait
-	sub b32 $r12 $r0 1
-	iowr I[$r10 + 0x000] $r12	// STRAND_GENE_CNT = 0xffffffff
-	mov $r12 2
-	iowr I[$r11 + 0x000] $r12	// STRAND_CMD = LATCH_GENE_CNT
-	call #strand_wait
-	call #strand_post
-
-	// read the size of each strand, poke the context offset of
-	// each into STRAND_{SAVE,LOAD}_SWBASE now, no need to worry
-	// about it later then.
-	mov $r8 0x880
-	shl b32 $r8 6
-	iord $r9 I[$r8 + 0x000]		// STRANDS
-	add b32 $r8 0x2200
-	shr b32 $r14 $r15 8
-	ctx_init_strand_loop:
-		iowr I[$r8 + 0x000] $r14	// STRAND_SAVE_SWBASE
-		iowr I[$r8 + 0x100] $r14	// STRAND_LOAD_SWBASE
-		iord $r10 I[$r8 + 0x200]	// STRAND_SIZE
-		shr b32 $r10 6
-		add b32 $r10 1
-		add b32 $r14 $r10
-		add b32 $r8 4
-		sub b32 $r9 1
-		bra ne #ctx_init_strand_loop
-
-	shl b32 $r14 8
-	sub b32 $r15 $r14 $r15
-	trace_clr(T_STRINIT)
-	ret
-')
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/os.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/os.h
new file mode 100644
index 0000000..fd1d380
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/os.h
@@ -0,0 +1,7 @@
+#ifndef __NVKM_GRAPH_OS_H__
+#define __NVKM_GRAPH_OS_H__
+
+#define E_BAD_COMMAND  0x00000001
+#define E_CMD_OVERFLOW 0x00000002
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index 1ac3611..03de517 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -186,13 +186,6 @@ nv50_graph_cclass = {
  * PGRAPH engine/subdev functions
  ******************************************************************************/
 
-static int
-nv50_graph_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x00);
-	return 0;
-}
-
 static const struct nouveau_bitfield nv50_pgraph_status[] = {
 	{ 0x00000001, "BUSY" }, /* set when any bit is set */
 	{ 0x00000002, "DISPATCH" },
@@ -302,8 +295,10 @@ nv84_graph_tlb_flush(struct nouveau_engine *engine)
 				nv_rd32(priv, 0x400388));
 	}
 
-	nv50_vm_flush_engine(&engine->base, 0x00);
 
+	nv_wr32(priv, 0x100c80, 0x00000001);
+	if (!nv_wait(priv, 0x100c80, 0x00000001, 0x00000000))
+		nv_error(priv, "vm flush timeout\n");
 	nv_mask(priv, 0x400500, 0x00000001, 0x00000001);
 	spin_unlock_irqrestore(&priv->lock, flags);
 	return timeout ? -EBUSY : 0;
@@ -857,10 +852,9 @@ nv50_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	};
 
-	if (nv_device(priv)->chipset == 0x50 ||
-	    nv_device(priv)->chipset == 0xac)
-		nv_engine(priv)->tlb_flush = nv50_graph_tlb_flush;
-	else
+	/* unfortunate hw bug workaround... */
+	if (nv_device(priv)->chipset != 0x50 &&
+	    nv_device(priv)->chipset != 0xac)
 		nv_engine(priv)->tlb_flush = nv84_graph_tlb_flush;
 
 	spin_lock_init(&priv->lock);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index f9b9d82..3f4f35c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -23,14 +23,12 @@
  */
 
 #include "nvc0.h"
-#include "fuc/hubnvc0.fuc.h"
-#include "fuc/gpcnvc0.fuc.h"
 
 /*******************************************************************************
  * Graphics object classes
  ******************************************************************************/
 
-static struct nouveau_oclass
+struct nouveau_oclass
 nvc0_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0x9039, &nouveau_object_ofuncs },
@@ -39,40 +37,6 @@ nvc0_graph_sclass[] = {
 	{}
 };
 
-static struct nouveau_oclass
-nvc1_graph_sclass[] = {
-	{ 0x902d, &nouveau_object_ofuncs },
-	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
-	{ 0x9197, &nouveau_object_ofuncs },
-	{}
-};
-
-static struct nouveau_oclass
-nvc8_graph_sclass[] = {
-	{ 0x902d, &nouveau_object_ofuncs },
-	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
-	{ 0x9197, &nouveau_object_ofuncs },
-	{ 0x9297, &nouveau_object_ofuncs },
-	{}
-};
-
-u64
-nvc0_graph_units(struct nouveau_graph *graph)
-{
-	struct nvc0_graph_priv *priv = (void *)graph;
-	u64 cfg;
-
-	cfg  = (u32)priv->gpc_nr;
-	cfg |= (u32)priv->tpc_total << 8;
-	cfg |= (u64)priv->rop_nr << 32;
-
-	return cfg;
-}
-
 /*******************************************************************************
  * PGRAPH context
  ******************************************************************************/
@@ -181,60 +145,308 @@ nvc0_graph_context_dtor(struct nouveau_object *object)
 	nouveau_graph_context_destroy(&chan->base);
 }
 
-static struct nouveau_oclass
-nvc0_graph_cclass = {
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvc0_graph_context_ctor,
-		.dtor = nvc0_graph_context_dtor,
-		.init = _nouveau_graph_context_init,
-		.fini = _nouveau_graph_context_fini,
-		.rd32 = _nouveau_graph_context_rd32,
-		.wr32 = _nouveau_graph_context_wr32,
-	},
-};
-
 /*******************************************************************************
  * PGRAPH engine/subdev functions
  ******************************************************************************/
 
-static void
-nvc0_graph_ctxctl_debug_unit(struct nvc0_graph_priv *priv, u32 base)
+struct nvc0_graph_init
+nvc0_graph_init_regs[] = {
+	{ 0x400080,   1, 0x04, 0x003083c2 },
+	{ 0x400088,   1, 0x04, 0x00006fe7 },
+	{ 0x40008c,   1, 0x04, 0x00000000 },
+	{ 0x400090,   1, 0x04, 0x00000030 },
+	{ 0x40013c,   1, 0x04, 0x013901f7 },
+	{ 0x400140,   1, 0x04, 0x00000100 },
+	{ 0x400144,   1, 0x04, 0x00000000 },
+	{ 0x400148,   1, 0x04, 0x00000110 },
+	{ 0x400138,   1, 0x04, 0x00000000 },
+	{ 0x400130,   2, 0x04, 0x00000000 },
+	{ 0x400124,   1, 0x04, 0x00000002 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk40xx[] = {
+	{ 0x40415c,   1, 0x04, 0x00000000 },
+	{ 0x404170,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk44xx[] = {
+	{ 0x404488,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk78xx[] = {
+	{ 0x407808,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk60xx[] = {
+	{ 0x406024,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk80xx[] = {
+	{ 0x40803c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_gpc[] = {
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x80000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   1, 0x04, 0x80000000 },
+	{ 0x4188cc,   1, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000050 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc0_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   3, 0x04, 0x00000000 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x06060618 },
+	{ 0x419ed0,   1, 0x04, 0x0eff0e38 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk88xx[] = {
+	{ 0x40880c,   1, 0x04, 0x00000000 },
+	{ 0x408910,   9, 0x04, 0x00000000 },
+	{ 0x408950,   1, 0x04, 0x00000000 },
+	{ 0x408954,   1, 0x04, 0x0000ffff },
+	{ 0x408984,   1, 0x04, 0x00000000 },
+	{ 0x408988,   1, 0x04, 0x08040201 },
+	{ 0x40898c,   1, 0x04, 0x80402010 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_tpc_0[] = {
+	{ 0x50405c,   1, 0x04, 0x00000001 },
+	{}
+};
+
+void
+nvc0_graph_mmio(struct nvc0_graph_priv *priv, struct nvc0_graph_init *init)
 {
-	nv_error(priv, "%06x - done 0x%08x\n", base,
-		 nv_rd32(priv, base + 0x400));
-	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
-		 nv_rd32(priv, base + 0x800), nv_rd32(priv, base + 0x804),
-		 nv_rd32(priv, base + 0x808), nv_rd32(priv, base + 0x80c));
-	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
-		 nv_rd32(priv, base + 0x810), nv_rd32(priv, base + 0x814),
-		 nv_rd32(priv, base + 0x818), nv_rd32(priv, base + 0x81c));
+	for (; init && init->count; init++) {
+		u32 addr = init->addr, i;
+		for (i = 0; i < init->count; i++) {
+			nv_wr32(priv, addr, init->data);
+			addr += init->pitch;
+		}
+	}
 }
 
 void
-nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *priv)
+nvc0_graph_icmd(struct nvc0_graph_priv *priv, struct nvc0_graph_init *init)
 {
-	u32 gpcnr = nv_rd32(priv, 0x409604) & 0xffff;
-	u32 gpc;
+	u32 addr, data;
+	int i, j;
+
+	nv_wr32(priv, 0x400208, 0x80000000);
+	for (i = 0; init->count; init++, i++) {
+		if (!i || data != init->data) {
+			nv_wr32(priv, 0x400204, init->data);
+			data = init->data;
+		}
 
-	nvc0_graph_ctxctl_debug_unit(priv, 0x409000);
-	for (gpc = 0; gpc < gpcnr; gpc++)
-		nvc0_graph_ctxctl_debug_unit(priv, 0x502000 + (gpc * 0x8000));
+		addr = init->addr;
+		for (j = 0; j < init->count; j++) {
+			nv_wr32(priv, 0x400200, addr);
+			addr += init->pitch;
+			while (nv_rd32(priv, 0x400700) & 0x00000002) {}
+		}
+	}
+	nv_wr32(priv, 0x400208, 0x00000000);
+}
+
+void
+nvc0_graph_mthd(struct nvc0_graph_priv *priv, struct nvc0_graph_mthd *mthds)
+{
+	struct nvc0_graph_mthd *mthd;
+	struct nvc0_graph_init *init;
+	int i = 0, j;
+	u32 data;
+
+	while ((mthd = &mthds[i++]) && (init = mthd->init)) {
+		u32  addr = 0x80000000 | mthd->oclass;
+		for (data = 0; init->count; init++) {
+			if (data != init->data) {
+				nv_wr32(priv, 0x40448c, init->data);
+				data = init->data;
+			}
+
+			addr = (addr & 0x8000ffff) | (init->addr << 14);
+			for (j = 0; j < init->count; j++) {
+				nv_wr32(priv, 0x404488, addr);
+				addr += init->pitch << 14;
+			}
+		}
+	}
+}
+
+u64
+nvc0_graph_units(struct nouveau_graph *graph)
+{
+	struct nvc0_graph_priv *priv = (void *)graph;
+	u64 cfg;
+
+	cfg  = (u32)priv->gpc_nr;
+	cfg |= (u32)priv->tpc_total << 8;
+	cfg |= (u64)priv->rop_nr << 32;
+
+	return cfg;
 }
 
+static const struct nouveau_enum nve0_sked_error[] = {
+	{ 7, "CONSTANT_BUFFER_SIZE" },
+	{ 9, "LOCAL_MEMORY_SIZE_POS" },
+	{ 10, "LOCAL_MEMORY_SIZE_NEG" },
+	{ 11, "WARP_CSTACK_SIZE" },
+	{ 12, "TOTAL_TEMP_SIZE" },
+	{ 13, "REGISTER_COUNT" },
+	{ 18, "TOTAL_THREADS" },
+	{ 20, "PROGRAM_OFFSET" },
+	{ 21, "SHARED_MEMORY_SIZE" },
+	{ 25, "SHARED_CONFIG_TOO_SMALL" },
+	{ 26, "TOTAL_REGISTER_COUNT" },
+	{}
+};
+
+static const struct nouveau_enum nvc0_gpc_rop_error[] = {
+	{ 1, "RT_PITCH_OVERRUN" },
+	{ 4, "RT_WIDTH_OVERRUN" },
+	{ 5, "RT_HEIGHT_OVERRUN" },
+	{ 7, "ZETA_STORAGE_TYPE_MISMATCH" },
+	{ 8, "RT_STORAGE_TYPE_MISMATCH" },
+	{ 10, "RT_LINEAR_MISMATCH" },
+	{}
+};
+
 static void
-nvc0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
+nvc0_graph_trap_gpc_rop(struct nvc0_graph_priv *priv, int gpc)
 {
-	u32 ustat = nv_rd32(priv, 0x409c18);
+	u32 trap[4];
+	int i;
 
-	if (ustat & 0x00000001)
-		nv_error(priv, "CTXCTRL ucode error\n");
-	if (ustat & 0x00080000)
-		nv_error(priv, "CTXCTRL watchdog timeout\n");
-	if (ustat & ~0x00080001)
-		nv_error(priv, "CTXCTRL 0x%08x\n", ustat);
+	trap[0] = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
+	trap[1] = nv_rd32(priv, GPC_UNIT(gpc, 0x0434));
+	trap[2] = nv_rd32(priv, GPC_UNIT(gpc, 0x0438));
+	trap[3] = nv_rd32(priv, GPC_UNIT(gpc, 0x043c));
+
+	nv_error(priv, "GPC%d/PROP trap:", gpc);
+	for (i = 0; i <= 29; ++i) {
+		if (!(trap[0] & (1 << i)))
+			continue;
+		pr_cont(" ");
+		nouveau_enum_print(nvc0_gpc_rop_error, i);
+	}
+	pr_cont("\n");
 
-	nvc0_graph_ctxctl_debug(priv);
-	nv_wr32(priv, 0x409c20, ustat);
+	nv_error(priv, "x = %u, y = %u, format = %x, storage type = %x\n",
+		 trap[1] & 0xffff, trap[1] >> 16, (trap[2] >> 8) & 0x3f,
+		 trap[3] & 0xff);
+	nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+}
+
+static const struct nouveau_enum nvc0_mp_warp_error[] = {
+	{ 0x00, "NO_ERROR" },
+	{ 0x01, "STACK_MISMATCH" },
+	{ 0x05, "MISALIGNED_PC" },
+	{ 0x08, "MISALIGNED_GPR" },
+	{ 0x09, "INVALID_OPCODE" },
+	{ 0x0d, "GPR_OUT_OF_BOUNDS" },
+	{ 0x0e, "MEM_OUT_OF_BOUNDS" },
+	{ 0x0f, "UNALIGNED_MEM_ACCESS" },
+	{ 0x11, "INVALID_PARAM" },
+	{}
+};
+
+static const struct nouveau_bitfield nvc0_mp_global_error[] = {
+	{ 0x00000004, "MULTIPLE_WARP_ERRORS" },
+	{ 0x00000008, "OUT_OF_STACK_SPACE" },
+	{}
+};
+
+static void
+nvc0_graph_trap_mp(struct nvc0_graph_priv *priv, int gpc, int tpc)
+{
+	u32 werr = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x648));
+	u32 gerr = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x650));
+
+	nv_error(priv, "GPC%i/TPC%i/MP trap:", gpc, tpc);
+	nouveau_bitfield_print(nvc0_mp_global_error, gerr);
+	if (werr) {
+		pr_cont(" ");
+		nouveau_enum_print(nvc0_mp_warp_error, werr & 0xffff);
+	}
+	pr_cont("\n");
+
+	nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x648), 0x00000000);
+	nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x650), gerr);
 }
 
 static void
@@ -246,18 +458,11 @@ nvc0_graph_trap_tpc(struct nvc0_graph_priv *priv, int gpc, int tpc)
 		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x0224));
 		nv_error(priv, "GPC%d/TPC%d/TEX: 0x%08x\n", gpc, tpc, trap);
 		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0224), 0xc0000000);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000001);
 		stat &= ~0x00000001;
 	}
 
 	if (stat & 0x00000002) {
-		u32 trap0 = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x0644));
-		u32 trap1 = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x064c));
-		nv_error(priv, "GPC%d/TPC%d/MP: 0x%08x 0x%08x\n",
-			       gpc, tpc, trap0, trap1);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0644), 0x001ffffe);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x064c), 0x0000000f);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000002);
+		nvc0_graph_trap_mp(priv, gpc, tpc);
 		stat &= ~0x00000002;
 	}
 
@@ -265,7 +470,6 @@ nvc0_graph_trap_tpc(struct nvc0_graph_priv *priv, int gpc, int tpc)
 		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x0084));
 		nv_error(priv, "GPC%d/TPC%d/POLY: 0x%08x\n", gpc, tpc, trap);
 		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0084), 0xc0000000);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000004);
 		stat &= ~0x00000004;
 	}
 
@@ -273,13 +477,11 @@ nvc0_graph_trap_tpc(struct nvc0_graph_priv *priv, int gpc, int tpc)
 		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x048c));
 		nv_error(priv, "GPC%d/TPC%d/L1C: 0x%08x\n", gpc, tpc, trap);
 		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x048c), 0xc0000000);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000008);
 		stat &= ~0x00000008;
 	}
 
 	if (stat) {
 		nv_error(priv, "GPC%d/TPC%d/0x%08x: unknown\n", gpc, tpc, stat);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), stat);
 	}
 }
 
@@ -290,10 +492,7 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 	int tpc;
 
 	if (stat & 0x00000001) {
-		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
-		nv_error(priv, "GPC%d/PROP: 0x%08x\n", gpc, trap);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000001);
+		nvc0_graph_trap_gpc_rop(priv, gpc);
 		stat &= ~0x00000001;
 	}
 
@@ -301,7 +500,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0900));
 		nv_error(priv, "GPC%d/ZCULL: 0x%08x\n", gpc, trap);
 		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000002);
 		stat &= ~0x00000002;
 	}
 
@@ -309,7 +507,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x1028));
 		nv_error(priv, "GPC%d/CCACHE: 0x%08x\n", gpc, trap);
 		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000004);
 		stat &= ~0x00000004;
 	}
 
@@ -317,7 +514,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0824));
 		nv_error(priv, "GPC%d/ESETUP: 0x%08x\n", gpc, trap);
 		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000008);
 		stat &= ~0x00000009;
 	}
 
@@ -332,7 +528,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 
 	if (stat) {
 		nv_error(priv, "GPC%d/0x%08x: unknown\n", gpc, stat);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), stat);
 	}
 }
 
@@ -340,7 +535,7 @@ static void
 nvc0_graph_trap_intr(struct nvc0_graph_priv *priv)
 {
 	u32 trap = nv_rd32(priv, 0x400108);
-	int rop, gpc;
+	int rop, gpc, i;
 
 	if (trap & 0x00000001) {
 		u32 stat = nv_rd32(priv, 0x404000);
@@ -390,6 +585,24 @@ nvc0_graph_trap_intr(struct nvc0_graph_priv *priv)
 		trap &= ~0x00000080;
 	}
 
+	if (trap & 0x00000100) {
+		u32 stat = nv_rd32(priv, 0x407020);
+
+		nv_error(priv, "SKED:");
+		for (i = 0; i <= 29; ++i) {
+			if (!(stat & (1 << i)))
+				continue;
+			pr_cont(" ");
+			nouveau_enum_print(nve0_sked_error, i);
+		}
+		pr_cont("\n");
+
+		if (stat & 0x3fffffff)
+			nv_wr32(priv, 0x407020, 0x40000000);
+		nv_wr32(priv, 0x400108, 0x00000100);
+		trap &= ~0x00000100;
+	}
+
 	if (trap & 0x01000000) {
 		u32 stat = nv_rd32(priv, 0x400118);
 		for (gpc = 0; stat && gpc < priv->gpc_nr; gpc++) {
@@ -424,6 +637,46 @@ nvc0_graph_trap_intr(struct nvc0_graph_priv *priv)
 }
 
 static void
+nvc0_graph_ctxctl_debug_unit(struct nvc0_graph_priv *priv, u32 base)
+{
+	nv_error(priv, "%06x - done 0x%08x\n", base,
+		 nv_rd32(priv, base + 0x400));
+	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
+		 nv_rd32(priv, base + 0x800), nv_rd32(priv, base + 0x804),
+		 nv_rd32(priv, base + 0x808), nv_rd32(priv, base + 0x80c));
+	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
+		 nv_rd32(priv, base + 0x810), nv_rd32(priv, base + 0x814),
+		 nv_rd32(priv, base + 0x818), nv_rd32(priv, base + 0x81c));
+}
+
+void
+nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *priv)
+{
+	u32 gpcnr = nv_rd32(priv, 0x409604) & 0xffff;
+	u32 gpc;
+
+	nvc0_graph_ctxctl_debug_unit(priv, 0x409000);
+	for (gpc = 0; gpc < gpcnr; gpc++)
+		nvc0_graph_ctxctl_debug_unit(priv, 0x502000 + (gpc * 0x8000));
+}
+
+static void
+nvc0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
+{
+	u32 ustat = nv_rd32(priv, 0x409c18);
+
+	if (ustat & 0x00000001)
+		nv_error(priv, "CTXCTL ucode error\n");
+	if (ustat & 0x00080000)
+		nv_error(priv, "CTXCTL watchdog timeout\n");
+	if (ustat & ~0x00080001)
+		nv_error(priv, "CTXCTL 0x%08x\n", ustat);
+
+	nvc0_graph_ctxctl_debug(priv);
+	nv_wr32(priv, 0x409c20, ustat);
+}
+
+static void
 nvc0_graph_intr(struct nouveau_subdev *subdev)
 {
 	struct nouveau_fifo *pfifo = nouveau_fifo(subdev);
@@ -499,290 +752,6 @@ nvc0_graph_intr(struct nouveau_subdev *subdev)
 	nouveau_engctx_put(engctx);
 }
 
-int
-nvc0_graph_ctor_fw(struct nvc0_graph_priv *priv, const char *fwname,
-		   struct nvc0_graph_fuc *fuc)
-{
-	struct nouveau_device *device = nv_device(priv);
-	const struct firmware *fw;
-	char f[32];
-	int ret;
-
-	snprintf(f, sizeof(f), "nouveau/nv%02x_%s", device->chipset, fwname);
-	ret = request_firmware(&fw, f, &device->pdev->dev);
-	if (ret) {
-		snprintf(f, sizeof(f), "nouveau/%s", fwname);
-		ret = request_firmware(&fw, f, &device->pdev->dev);
-		if (ret) {
-			nv_error(priv, "failed to load %s\n", fwname);
-			return ret;
-		}
-	}
-
-	fuc->size = fw->size;
-	fuc->data = kmemdup(fw->data, fuc->size, GFP_KERNEL);
-	release_firmware(fw);
-	return (fuc->data != NULL) ? 0 : -ENOMEM;
-}
-
-static int
-nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		struct nouveau_oclass *oclass, void *data, u32 size,
-		struct nouveau_object **pobject)
-{
-	struct nouveau_device *device = nv_device(parent);
-	struct nvc0_graph_priv *priv;
-	bool enable = device->chipset != 0xd7;
-	int ret, i;
-
-	ret = nouveau_graph_create(parent, engine, oclass, enable, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	nv_subdev(priv)->unit = 0x18001000;
-	nv_subdev(priv)->intr = nvc0_graph_intr;
-	nv_engine(priv)->cclass = &nvc0_graph_cclass;
-
-	priv->base.units = nvc0_graph_units;
-
-	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
-		nv_info(priv, "using external firmware\n");
-		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
-		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
-			return -EINVAL;
-		priv->firmware = true;
-	}
-
-	switch (nvc0_graph_class(priv)) {
-	case 0x9097:
-		nv_engine(priv)->sclass = nvc0_graph_sclass;
-		break;
-	case 0x9197:
-		nv_engine(priv)->sclass = nvc1_graph_sclass;
-		break;
-	case 0x9297:
-		nv_engine(priv)->sclass = nvc8_graph_sclass;
-		break;
-	}
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b4);
-	if (ret)
-		return ret;
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b8);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < 0x1000; i += 4) {
-		nv_wo32(priv->unk4188b4, i, 0x00000010);
-		nv_wo32(priv->unk4188b8, i, 0x00000010);
-	}
-
-	priv->rop_nr = (nv_rd32(priv, 0x409604) & 0x001f0000) >> 16;
-	priv->gpc_nr =  nv_rd32(priv, 0x409604) & 0x0000001f;
-	for (i = 0; i < priv->gpc_nr; i++) {
-		priv->tpc_nr[i]  = nv_rd32(priv, GPC_UNIT(i, 0x2608));
-		priv->tpc_total += priv->tpc_nr[i];
-	}
-
-	/*XXX: these need figuring out... though it might not even matter */
-	switch (nv_device(priv)->chipset) {
-	case 0xc0:
-		if (priv->tpc_total == 11) { /* 465, 3/4/4/0, 4 */
-			priv->magic_not_rop_nr = 0x07;
-		} else
-		if (priv->tpc_total == 14) { /* 470, 3/3/4/4, 5 */
-			priv->magic_not_rop_nr = 0x05;
-		} else
-		if (priv->tpc_total == 15) { /* 480, 3/4/4/4, 6 */
-			priv->magic_not_rop_nr = 0x06;
-		}
-		break;
-	case 0xc3: /* 450, 4/0/0/0, 2 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xc4: /* 460, 3/4/0/0, 4 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	case 0xc1: /* 2/0/0/0, 1 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	case 0xc8: /* 4/4/3/4, 5 */
-		priv->magic_not_rop_nr = 0x06;
-		break;
-	case 0xce: /* 4/4/0/0, 4 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xcf: /* 4/0/0/0, 3 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xd9: /* 1/0/0/0, 1 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	}
-
-	return 0;
-}
-
-static void
-nvc0_graph_dtor_fw(struct nvc0_graph_fuc *fuc)
-{
-	kfree(fuc->data);
-	fuc->data = NULL;
-}
-
-void
-nvc0_graph_dtor(struct nouveau_object *object)
-{
-	struct nvc0_graph_priv *priv = (void *)object;
-
-	kfree(priv->data);
-
-	nvc0_graph_dtor_fw(&priv->fuc409c);
-	nvc0_graph_dtor_fw(&priv->fuc409d);
-	nvc0_graph_dtor_fw(&priv->fuc41ac);
-	nvc0_graph_dtor_fw(&priv->fuc41ad);
-
-	nouveau_gpuobj_ref(NULL, &priv->unk4188b8);
-	nouveau_gpuobj_ref(NULL, &priv->unk4188b4);
-
-	nouveau_graph_destroy(&priv->base);
-}
-
-static void
-nvc0_graph_init_obj418880(struct nvc0_graph_priv *priv)
-{
-	int i;
-
-	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
-	for (i = 0; i < 4; i++)
-		nv_wr32(priv, GPC_BCAST(0x0888) + (i * 4), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
-	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
-}
-
-static void
-nvc0_graph_init_regs(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x400080, 0x003083c2);
-	nv_wr32(priv, 0x400088, 0x00006fe7);
-	nv_wr32(priv, 0x40008c, 0x00000000);
-	nv_wr32(priv, 0x400090, 0x00000030);
-	nv_wr32(priv, 0x40013c, 0x013901f7);
-	nv_wr32(priv, 0x400140, 0x00000100);
-	nv_wr32(priv, 0x400144, 0x00000000);
-	nv_wr32(priv, 0x400148, 0x00000110);
-	nv_wr32(priv, 0x400138, 0x00000000);
-	nv_wr32(priv, 0x400130, 0x00000000);
-	nv_wr32(priv, 0x400134, 0x00000000);
-	nv_wr32(priv, 0x400124, 0x00000002);
-}
-
-static void
-nvc0_graph_init_gpc_0(struct nvc0_graph_priv *priv)
-{
-	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
-	u32 data[TPC_MAX / 8];
-	u8  tpcnr[GPC_MAX];
-	int i, gpc, tpc;
-
-	nv_wr32(priv, TPC_UNIT(0, 0, 0x5c), 1); /* affects TFB offset queries */
-
-	/*
-	 *      TP      ROP UNKVAL(magic_not_rop_nr)
-	 * 450: 4/0/0/0 2        3
-	 * 460: 3/4/0/0 4        1
-	 * 465: 3/4/4/0 4        7
-	 * 470: 3/3/4/4 5        5
-	 * 480: 3/4/4/4 6        6
-	 */
-
-	memset(data, 0x00, sizeof(data));
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
-		do {
-			gpc = (gpc + 1) % priv->gpc_nr;
-		} while (!tpcnr[gpc]);
-		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-		data[i / 8] |= tpc << ((i % 8) * 4);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
-	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
-	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
-	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0914), priv->magic_not_rop_nr << 8 |
-						  priv->tpc_nr[gpc]);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 | priv->tpc_total);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x1bd4), magicgpc918);
-	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
-}
-
-static void
-nvc0_graph_init_units(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x409c24, 0x000f0000);
-	nv_wr32(priv, 0x404000, 0xc0000000); /* DISPATCH */
-	nv_wr32(priv, 0x404600, 0xc0000000); /* M2MF */
-	nv_wr32(priv, 0x408030, 0xc0000000);
-	nv_wr32(priv, 0x40601c, 0xc0000000);
-	nv_wr32(priv, 0x404490, 0xc0000000); /* MACRO */
-	nv_wr32(priv, 0x406018, 0xc0000000);
-	nv_wr32(priv, 0x405840, 0xc0000000);
-	nv_wr32(priv, 0x405844, 0x00ffffff);
-	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
-	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
-}
-
-static void
-nvc0_graph_init_gpc_1(struct nvc0_graph_priv *priv)
-{
-	int gpc, tpc;
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
-		}
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
-	}
-}
-
-static void
-nvc0_graph_init_rop(struct nvc0_graph_priv *priv)
-{
-	int rop;
-
-	for (rop = 0; rop < priv->rop_nr; rop++) {
-		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
-		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
-	}
-}
-
 void
 nvc0_graph_init_fw(struct nvc0_graph_priv *priv, u32 fuc_base,
 		   struct nvc0_graph_fuc *code, struct nvc0_graph_fuc *data)
@@ -801,9 +770,46 @@ nvc0_graph_init_fw(struct nvc0_graph_priv *priv, u32 fuc_base,
 	}
 }
 
-static int
+static void
+nvc0_graph_init_csdata(struct nvc0_graph_priv *priv,
+		       struct nvc0_graph_init *init,
+		       u32 falcon, u32 starstar, u32 base)
+{
+	u32 addr = init->addr;
+	u32 next = addr;
+	u32 star, temp;
+
+	nv_wr32(priv, falcon + 0x01c0, 0x02000000 + starstar);
+	star = nv_rd32(priv, falcon + 0x01c4);
+	temp = nv_rd32(priv, falcon + 0x01c4);
+	if (temp > star)
+		star = temp;
+	nv_wr32(priv, falcon + 0x01c0, 0x01000000 + star);
+
+	do {
+		if (init->addr != next) {
+			while (addr < next) {
+				u32 nr = min((int)(next - addr) / 4, 32);
+				nv_wr32(priv, falcon + 0x01c4,
+					((nr - 1) << 26) | (addr - base));
+				addr += nr * 4;
+				star += 4;
+			}
+			addr = next = init->addr;
+		}
+		next += init->count * 4;
+	} while ((init++)->count);
+
+	nv_wr32(priv, falcon + 0x01c0, 0x01000004 + starstar);
+	nv_wr32(priv, falcon + 0x01c4, star);
+}
+
+int
 nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 {
+	struct nvc0_graph_oclass *oclass = (void *)nv_object(priv)->oclass;
+	struct nvc0_grctx_oclass *cclass = (void *)nv_engine(priv)->cclass;
+	struct nvc0_graph_init *init;
 	u32 r000260;
 	int i;
 
@@ -854,6 +860,38 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 			return -EBUSY;
 		}
 
+		if (nv_device(priv)->chipset >= 0xe0) {
+			nv_wr32(priv, 0x409800, 0x00000000);
+			nv_wr32(priv, 0x409500, 0x00000001);
+			nv_wr32(priv, 0x409504, 0x00000030);
+			if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
+				nv_error(priv, "fuc09 req 0x30 timeout\n");
+				return -EBUSY;
+			}
+
+			nv_wr32(priv, 0x409810, 0xb00095c8);
+			nv_wr32(priv, 0x409800, 0x00000000);
+			nv_wr32(priv, 0x409500, 0x00000001);
+			nv_wr32(priv, 0x409504, 0x00000031);
+			if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
+				nv_error(priv, "fuc09 req 0x31 timeout\n");
+				return -EBUSY;
+			}
+
+			nv_wr32(priv, 0x409810, 0x00080420);
+			nv_wr32(priv, 0x409800, 0x00000000);
+			nv_wr32(priv, 0x409500, 0x00000001);
+			nv_wr32(priv, 0x409504, 0x00000032);
+			if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
+				nv_error(priv, "fuc09 req 0x32 timeout\n");
+				return -EBUSY;
+			}
+
+			nv_wr32(priv, 0x409614, 0x00000070);
+			nv_wr32(priv, 0x409614, 0x00000770);
+			nv_wr32(priv, 0x40802c, 0x00000001);
+		}
+
 		if (priv->data == NULL) {
 			int ret = nvc0_grctx_generate(priv);
 			if (ret) {
@@ -868,31 +906,41 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 	/* load HUB microcode */
 	r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
 	nv_wr32(priv, 0x4091c0, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grhub_data) / 4; i++)
-		nv_wr32(priv, 0x4091c4, nvc0_grhub_data[i]);
+	for (i = 0; i < oclass->fecs.ucode->data.size / 4; i++)
+		nv_wr32(priv, 0x4091c4, oclass->fecs.ucode->data.data[i]);
 
 	nv_wr32(priv, 0x409180, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grhub_code) / 4; i++) {
+	for (i = 0; i < oclass->fecs.ucode->code.size / 4; i++) {
 		if ((i & 0x3f) == 0)
 			nv_wr32(priv, 0x409188, i >> 6);
-		nv_wr32(priv, 0x409184, nvc0_grhub_code[i]);
+		nv_wr32(priv, 0x409184, oclass->fecs.ucode->code.data[i]);
+	}
+
+	for (i = 0; (init = cclass->hub[i]); i++) {
+		nvc0_graph_init_csdata(priv, init, 0x409000, 0x000, 0x000000);
 	}
 
 	/* load GPC microcode */
 	nv_wr32(priv, 0x41a1c0, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grgpc_data) / 4; i++)
-		nv_wr32(priv, 0x41a1c4, nvc0_grgpc_data[i]);
+	for (i = 0; i < oclass->gpccs.ucode->data.size / 4; i++)
+		nv_wr32(priv, 0x41a1c4, oclass->gpccs.ucode->data.data[i]);
 
 	nv_wr32(priv, 0x41a180, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grgpc_code) / 4; i++) {
+	for (i = 0; i < oclass->gpccs.ucode->code.size / 4; i++) {
 		if ((i & 0x3f) == 0)
 			nv_wr32(priv, 0x41a188, i >> 6);
-		nv_wr32(priv, 0x41a184, nvc0_grgpc_code[i]);
+		nv_wr32(priv, 0x41a184, oclass->gpccs.ucode->code.data[i]);
 	}
 	nv_wr32(priv, 0x000260, r000260);
 
+	if ((init = cclass->gpc[0]))
+		nvc0_graph_init_csdata(priv, init, 0x41a000, 0x000, 0x418000);
+	if ((init = cclass->gpc[2]))
+		nvc0_graph_init_csdata(priv, init, 0x41a000, 0x004, 0x419800);
+	if ((init = cclass->gpc[3]))
+		nvc0_graph_init_csdata(priv, init, 0x41a000, 0x008, 0x41be00);
+
 	/* start HUB ucode running, it'll init the GPCs */
-	nv_wr32(priv, 0x409800, nv_device(priv)->chipset);
 	nv_wr32(priv, 0x40910c, 0x00000000);
 	nv_wr32(priv, 0x409100, 0x00000002);
 	if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000)) {
@@ -913,29 +961,104 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 	return 0;
 }
 
-static int
+int
 nvc0_graph_init(struct nouveau_object *object)
 {
+	struct nvc0_graph_oclass *oclass = (void *)object->oclass;
 	struct nvc0_graph_priv *priv = (void *)object;
-	int ret;
+	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
+	u32 data[TPC_MAX / 8] = {};
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc, rop;
+	int ret, i;
 
 	ret = nouveau_graph_init(&priv->base);
 	if (ret)
 		return ret;
 
-	nvc0_graph_init_obj418880(priv);
-	nvc0_graph_init_regs(priv);
-	/*nvc0_graph_init_unitplemented_magics(priv);*/
-	nvc0_graph_init_gpc_0(priv);
-	/*nvc0_graph_init_unitplemented_c242(priv);*/
+	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0888), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x088c), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0890), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0894), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
+	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
+
+	for (i = 0; oclass->mmio[i]; i++)
+		nvc0_graph_mmio(priv, oclass->mmio[i]);
+
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
+
+		data[i / 8] |= tpc << ((i % 8) * 4);
+	}
+
+	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
+	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
+	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
+	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0914),
+			priv->magic_not_rop_nr << 8 | priv->tpc_nr[gpc]);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 |
+			priv->tpc_total);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
+	}
+
+	if (nv_device(priv)->chipset != 0xd7)
+		nv_wr32(priv, GPC_BCAST(0x1bd4), magicgpc918);
+	else
+		nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918);
+
+	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
 
 	nv_wr32(priv, 0x400500, 0x00010001);
+
 	nv_wr32(priv, 0x400100, 0xffffffff);
 	nv_wr32(priv, 0x40013c, 0xffffffff);
 
-	nvc0_graph_init_units(priv);
-	nvc0_graph_init_gpc_1(priv);
-	nvc0_graph_init_rop(priv);
+	nv_wr32(priv, 0x409c24, 0x000f0000);
+	nv_wr32(priv, 0x404000, 0xc0000000);
+	nv_wr32(priv, 0x404600, 0xc0000000);
+	nv_wr32(priv, 0x408030, 0xc0000000);
+	nv_wr32(priv, 0x40601c, 0xc0000000);
+	nv_wr32(priv, 0x404490, 0xc0000000);
+	nv_wr32(priv, 0x406018, 0xc0000000);
+	nv_wr32(priv, 0x405840, 0xc0000000);
+	nv_wr32(priv, 0x405844, 0x00ffffff);
+	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
+	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
+		}
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
+	}
+
+	for (rop = 0; rop < priv->rop_nr; rop++) {
+		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
+		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
+	}
 
 	nv_wr32(priv, 0x400108, 0xffffffff);
 	nv_wr32(priv, 0x400138, 0xffffffff);
@@ -943,22 +1066,205 @@ nvc0_graph_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x400130, 0xffffffff);
 	nv_wr32(priv, 0x40011c, 0xffffffff);
 	nv_wr32(priv, 0x400134, 0xffffffff);
+
 	nv_wr32(priv, 0x400054, 0x34ce3464);
+	return nvc0_graph_init_ctxctl(priv);
+}
 
-	ret = nvc0_graph_init_ctxctl(priv);
+static void
+nvc0_graph_dtor_fw(struct nvc0_graph_fuc *fuc)
+{
+	kfree(fuc->data);
+	fuc->data = NULL;
+}
+
+int
+nvc0_graph_ctor_fw(struct nvc0_graph_priv *priv, const char *fwname,
+		   struct nvc0_graph_fuc *fuc)
+{
+	struct nouveau_device *device = nv_device(priv);
+	const struct firmware *fw;
+	char f[32];
+	int ret;
+
+	snprintf(f, sizeof(f), "nouveau/nv%02x_%s", device->chipset, fwname);
+	ret = request_firmware(&fw, f, &device->pdev->dev);
+	if (ret) {
+		snprintf(f, sizeof(f), "nouveau/%s", fwname);
+		ret = request_firmware(&fw, f, &device->pdev->dev);
+		if (ret) {
+			nv_error(priv, "failed to load %s\n", fwname);
+			return ret;
+		}
+	}
+
+	fuc->size = fw->size;
+	fuc->data = kmemdup(fw->data, fuc->size, GFP_KERNEL);
+	release_firmware(fw);
+	return (fuc->data != NULL) ? 0 : -ENOMEM;
+}
+
+void
+nvc0_graph_dtor(struct nouveau_object *object)
+{
+	struct nvc0_graph_priv *priv = (void *)object;
+
+	kfree(priv->data);
+
+	nvc0_graph_dtor_fw(&priv->fuc409c);
+	nvc0_graph_dtor_fw(&priv->fuc409d);
+	nvc0_graph_dtor_fw(&priv->fuc41ac);
+	nvc0_graph_dtor_fw(&priv->fuc41ad);
+
+	nouveau_gpuobj_ref(NULL, &priv->unk4188b8);
+	nouveau_gpuobj_ref(NULL, &priv->unk4188b4);
+
+	nouveau_graph_destroy(&priv->base);
+}
+
+int
+nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *bclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nvc0_graph_oclass *oclass = (void *)bclass;
+	struct nouveau_device *device = nv_device(parent);
+	struct nvc0_graph_priv *priv;
+	int ret, i;
+
+	ret = nouveau_graph_create(parent, engine, bclass,
+				   (oclass->fecs.ucode != NULL), &priv);
+	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
+	nv_subdev(priv)->unit = 0x18001000;
+	nv_subdev(priv)->intr = nvc0_graph_intr;
+
+	priv->base.units = nvc0_graph_units;
+
+	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
+		nv_info(priv, "using external firmware\n");
+		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
+		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
+		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
+		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
+			return -EINVAL;
+		priv->firmware = true;
+	}
+
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b4);
+	if (ret)
+		return ret;
+
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b8);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < 0x1000; i += 4) {
+		nv_wo32(priv->unk4188b4, i, 0x00000010);
+		nv_wo32(priv->unk4188b8, i, 0x00000010);
+	}
+
+	priv->rop_nr = (nv_rd32(priv, 0x409604) & 0x001f0000) >> 16;
+	priv->gpc_nr =  nv_rd32(priv, 0x409604) & 0x0000001f;
+	for (i = 0; i < priv->gpc_nr; i++) {
+		priv->tpc_nr[i]  = nv_rd32(priv, GPC_UNIT(i, 0x2608));
+		priv->tpc_total += priv->tpc_nr[i];
+	}
+
+	/*XXX: these need figuring out... though it might not even matter */
+	switch (nv_device(priv)->chipset) {
+	case 0xc0:
+		if (priv->tpc_total == 11) { /* 465, 3/4/4/0, 4 */
+			priv->magic_not_rop_nr = 0x07;
+		} else
+		if (priv->tpc_total == 14) { /* 470, 3/3/4/4, 5 */
+			priv->magic_not_rop_nr = 0x05;
+		} else
+		if (priv->tpc_total == 15) { /* 480, 3/4/4/4, 6 */
+			priv->magic_not_rop_nr = 0x06;
+		}
+		break;
+	case 0xc3: /* 450, 4/0/0/0, 2 */
+		priv->magic_not_rop_nr = 0x03;
+		break;
+	case 0xc4: /* 460, 3/4/0/0, 4 */
+		priv->magic_not_rop_nr = 0x01;
+		break;
+	case 0xc1: /* 2/0/0/0, 1 */
+		priv->magic_not_rop_nr = 0x01;
+		break;
+	case 0xc8: /* 4/4/3/4, 5 */
+		priv->magic_not_rop_nr = 0x06;
+		break;
+	case 0xce: /* 4/4/0/0, 4 */
+		priv->magic_not_rop_nr = 0x03;
+		break;
+	case 0xcf: /* 4/0/0/0, 3 */
+		priv->magic_not_rop_nr = 0x03;
+		break;
+	case 0xd7:
+	case 0xd9: /* 1/0/0/0, 1 */
+		priv->magic_not_rop_nr = 0x01;
+		break;
+	}
+
+	nv_engine(priv)->cclass = *oclass->cclass;
+	nv_engine(priv)->sclass =  oclass->sclass;
 	return 0;
 }
 
-struct nouveau_oclass
-nvc0_graph_oclass = {
-	.handle = NV_ENGINE(GR, 0xc0),
-	.ofuncs = &(struct nouveau_ofuncs) {
+struct nvc0_graph_init *
+nvc0_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc0_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc0_graph_init_gpc,
+	nvc0_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+#include "fuc/hubnvc0.fuc.h"
+
+struct nvc0_graph_ucode
+nvc0_graph_fecs_ucode = {
+	.code.data = nvc0_grhub_code,
+	.code.size = sizeof(nvc0_grhub_code),
+	.data.data = nvc0_grhub_data,
+	.data.size = sizeof(nvc0_grhub_data),
+};
+
+#include "fuc/gpcnvc0.fuc.h"
+
+struct nvc0_graph_ucode
+nvc0_graph_gpccs_ucode = {
+	.code.data = nvc0_grgpc_code,
+	.code.size = sizeof(nvc0_grgpc_code),
+	.data.data = nvc0_grgpc_data,
+	.data.size = sizeof(nvc0_grgpc_data),
+};
+
+struct nouveau_oclass *
+nvc0_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nvc0_graph_ctor,
 		.dtor = nvc0_graph_dtor,
 		.init = nvc0_graph_init,
 		.fini = _nouveau_graph_fini,
 	},
-};
+	.cclass = &nvc0_grctx_oclass,
+	.sclass =  nvc0_graph_sclass,
+	.mmio = nvc0_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
index c870dad..ea17a80 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
@@ -38,8 +38,8 @@
 #include <engine/fifo.h>
 #include <engine/graph.h>
 
-#define GPC_MAX 4
-#define TPC_MAX 32
+#define GPC_MAX 32
+#define TPC_MAX (GPC_MAX * 8)
 
 #define ROP_BCAST(r)      (0x408800 + (r))
 #define ROP_UNIT(u, r)    (0x410000 + (u) * 0x400 + (r))
@@ -102,74 +102,187 @@ struct nvc0_graph_chan {
 	} data[4];
 };
 
-static inline u32
-nvc0_graph_class(void *obj)
-{
-	struct nouveau_device *device = nv_device(obj);
-
-	switch (device->chipset) {
-	case 0xc0:
-	case 0xc3:
-	case 0xc4:
-	case 0xce: /* guess, mmio trace shows only 0x9097 state */
-	case 0xcf: /* guess, mmio trace shows only 0x9097 state */
-		return 0x9097;
-	case 0xc1:
-		return 0x9197;
-	case 0xc8:
-	case 0xd9:
-	case 0xd7:
-		return 0x9297;
-	case 0xe4:
-	case 0xe7:
-	case 0xe6:
-		return 0xa097;
-	default:
-		return 0;
-	}
-}
-
-void nv_icmd(struct nvc0_graph_priv *priv, u32 icmd, u32 data);
-
-static inline void
-nv_mthd(struct nvc0_graph_priv *priv, u32 class, u32 mthd, u32 data)
-{
-	nv_wr32(priv, 0x40448c, data);
-	nv_wr32(priv, 0x404488, 0x80000000 | (mthd << 14) | class);
-}
+int  nvc0_grctx_generate(struct nvc0_graph_priv *);
+
+int  nvc0_graph_context_ctor(struct nouveau_object *, struct nouveau_object *,
+			     struct nouveau_oclass *, void *, u32,
+			     struct nouveau_object **);
+void nvc0_graph_context_dtor(struct nouveau_object *);
+
+void nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *);
+
+u64  nvc0_graph_units(struct nouveau_graph *);
+int  nvc0_graph_ctor(struct nouveau_object *, struct nouveau_object *,
+		     struct nouveau_oclass *, void *data, u32 size,
+		     struct nouveau_object **);
+void nvc0_graph_dtor(struct nouveau_object *);
+int  nvc0_graph_init(struct nouveau_object *);
+int  nve4_graph_init(struct nouveau_object *);
+
+extern struct nouveau_oclass nvc0_graph_sclass[];
+
+extern struct nouveau_oclass nvc8_graph_sclass[];
+
+struct nvc0_graph_init {
+	u32 addr;
+	u8  count;
+	u8  pitch;
+	u32 data;
+};
+
+struct nvc0_graph_mthd {
+	u16 oclass;
+	struct nvc0_graph_init *init;
+};
 
 struct nvc0_grctx {
 	struct nvc0_graph_priv *priv;
 	struct nvc0_graph_data *data;
 	struct nvc0_graph_mmio *mmio;
-	struct nouveau_gpuobj *chan;
 	int buffer_nr;
 	u64 buffer[4];
 	u64 addr;
 };
 
+struct nvc0_grctx_oclass {
+	struct nouveau_oclass base;
+	/* main context generation function */
+	void  (*main)(struct nvc0_graph_priv *, struct nvc0_grctx *);
+	/* context-specific modify-on-first-load list generation function */
+	void  (*mods)(struct nvc0_graph_priv *, struct nvc0_grctx *);
+	void  (*unkn)(struct nvc0_graph_priv *);
+	/* mmio context data */
+	struct nvc0_graph_init **hub;
+	struct nvc0_graph_init **gpc;
+	/* indirect context data, generated with icmds/mthds */
+	struct nvc0_graph_init *icmd;
+	struct nvc0_graph_mthd *mthd;
+};
+
+struct nvc0_graph_ucode {
+	struct nvc0_graph_fuc code;
+	struct nvc0_graph_fuc data;
+};
+
+extern struct nvc0_graph_ucode nvc0_graph_fecs_ucode;
+extern struct nvc0_graph_ucode nvc0_graph_gpccs_ucode;
+
+struct nvc0_graph_oclass {
+	struct nouveau_oclass base;
+	struct nouveau_oclass **cclass;
+	struct nouveau_oclass *sclass;
+	struct nvc0_graph_init **mmio;
+	struct {
+		struct nvc0_graph_ucode *ucode;
+	} fecs;
+	struct {
+		struct nvc0_graph_ucode *ucode;
+	} gpccs;
+};
+
+void nvc0_graph_mmio(struct nvc0_graph_priv *, struct nvc0_graph_init *);
+void nvc0_graph_icmd(struct nvc0_graph_priv *, struct nvc0_graph_init *);
+void nvc0_graph_mthd(struct nvc0_graph_priv *, struct nvc0_graph_mthd *);
+int  nvc0_graph_init_ctxctl(struct nvc0_graph_priv *);
+
+extern struct nvc0_graph_init nvc0_graph_init_regs[];
+extern struct nvc0_graph_init nvc0_graph_init_unk40xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk44xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk78xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk60xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk58xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk80xx[];
+extern struct nvc0_graph_init nvc0_graph_init_gpc[];
+extern struct nvc0_graph_init nvc0_graph_init_unk88xx[];
+extern struct nvc0_graph_init nvc0_graph_tpc_0[];
+
+extern struct nvc0_graph_init nvc3_graph_init_unk58xx[];
+
+extern struct nvc0_graph_init nvd9_graph_init_unk58xx[];
+extern struct nvc0_graph_init nvd9_graph_init_unk64xx[];
+
+extern struct nvc0_graph_init nve4_graph_init_regs[];
+extern struct nvc0_graph_init nve4_graph_init_unk[];
+extern struct nvc0_graph_init nve4_graph_init_unk88xx[];
+
 int  nvc0_grctx_generate(struct nvc0_graph_priv *);
-int  nvc0_grctx_init(struct nvc0_graph_priv *, struct nvc0_grctx *);
-void nvc0_grctx_data(struct nvc0_grctx *, u32, u32, u32);
-void nvc0_grctx_mmio(struct nvc0_grctx *, u32, u32, u32, u32);
-int  nvc0_grctx_fini(struct nvc0_grctx *);
+void nvc0_grctx_generate_main(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc0_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc0_grctx_generate_unkn(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_tpcid(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r406028(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r4060a8(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r418bb8(struct nvc0_graph_priv *);
+void nve4_grctx_generate_r418bb8(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r406800(struct nvc0_graph_priv *);
 
-int  nve0_grctx_generate(struct nvc0_graph_priv *);
+extern struct nouveau_oclass *nvc0_grctx_oclass;
+extern struct nvc0_graph_init *nvc0_grctx_init_hub[];
+extern struct nvc0_graph_init nvc0_grctx_init_base[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk40xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk44xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk46xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk47xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk60xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk64xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk78xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk80xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_gpc_0[];
+extern struct nvc0_graph_init nvc0_grctx_init_gpc_1[];
+extern struct nvc0_graph_init nvc0_grctx_init_tpc[];
+extern struct nvc0_graph_init nvc0_grctx_init_icmd[];
+extern struct nvc0_graph_init nvd9_grctx_init_icmd[]; //
 
-#define mmio_data(s,a,p) nvc0_grctx_data(&info, (s), (a), (p))
-#define mmio_list(r,d,s,b) nvc0_grctx_mmio(&info, (r), (d), (s), (b))
+extern struct nvc0_graph_mthd nvc0_grctx_init_mthd[];
+extern struct nvc0_graph_init nvc0_grctx_init_902d[];
+extern struct nvc0_graph_init nvc0_grctx_init_9039[];
+extern struct nvc0_graph_init nvc0_grctx_init_90c0[];
+extern struct nvc0_graph_init nvc0_grctx_init_mthd_magic[];
 
-void nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *);
-int  nvc0_graph_ctor_fw(struct nvc0_graph_priv *, const char *,
-			struct nvc0_graph_fuc *);
-void nvc0_graph_dtor(struct nouveau_object *);
-void nvc0_graph_init_fw(struct nvc0_graph_priv *, u32 base,
-			struct nvc0_graph_fuc *, struct nvc0_graph_fuc *);
-int  nvc0_graph_context_ctor(struct nouveau_object *, struct nouveau_object *,
-			     struct nouveau_oclass *, void *, u32,
-			     struct nouveau_object **);
-void nvc0_graph_context_dtor(struct nouveau_object *);
+void nvc1_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc1_grctx_generate_unkn(struct nvc0_graph_priv *);
+extern struct nouveau_oclass *nvc1_grctx_oclass;
+extern struct nvc0_graph_init nvc1_grctx_init_9097[];
+
+extern struct nouveau_oclass *nvc3_grctx_oclass;
+
+extern struct nouveau_oclass *nvc8_grctx_oclass;
+extern struct nvc0_graph_init nvc8_grctx_init_9197[];
+extern struct nvc0_graph_init nvc8_grctx_init_9297[];
+
+extern struct nouveau_oclass *nvd7_grctx_oclass;
+
+extern struct nouveau_oclass *nvd9_grctx_oclass;
+extern struct nvc0_graph_init nvd9_grctx_init_rop[];
+extern struct nvc0_graph_mthd nvd9_grctx_init_mthd[];
+
+void nve4_grctx_generate_main(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nve4_grctx_generate_unkn(struct nvc0_graph_priv *);
+extern struct nouveau_oclass *nve4_grctx_oclass;
+extern struct nvc0_graph_init nve4_grctx_init_unk46xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk47xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk58xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk80xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk90xx[];
+
+extern struct nouveau_oclass *nvf0_grctx_oclass;
+
+#define mmio_data(s,a,p) do {                                                  \
+	info->buffer[info->buffer_nr] = round_up(info->addr, (a));             \
+	info->addr = info->buffer[info->buffer_nr++] + (s);                    \
+	info->data->size = (s);                                                \
+	info->data->align = (a);                                               \
+	info->data->access = (p);                                              \
+	info->data++;                                                          \
+} while(0)
 
-u64 nvc0_graph_units(struct nouveau_graph *);
+#define mmio_list(r,d,s,b) do {                                                \
+	info->mmio->addr = (r);                                                \
+	info->mmio->data = (d);                                                \
+	info->mmio->shift = (s);                                               \
+	info->mmio->buffer = (b);                                              \
+	info->mmio++;                                                          \
+	nv_wr32(priv, (r), (d) | ((s) ? (info->buffer[(b)] >> (s)) : 0));      \
+} while(0)
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
new file mode 100644
index 0000000..bc4a469
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
@@ -0,0 +1,144 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nvc1_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0x9039, &nouveau_object_ofuncs },
+	{ 0x9097, &nouveau_object_ofuncs },
+	{ 0x90c0, &nouveau_object_ofuncs },
+	{ 0x9197, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static struct nvc0_graph_init
+nvc1_graph_init_gpc[] = {
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000003 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc1_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   2, 0x04, 0x00000000 },
+	{ 0x419814,   1, 0x04, 0x00000004 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419880,   1, 0x04, 0x00000002 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init *
+nvc1_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc3_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc1_graph_init_gpc,
+	nvc1_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc1_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc1),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvc1_grctx_oclass,
+	.sclass = nvc1_graph_sclass,
+	.mmio = nvc1_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc3.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc3.c
new file mode 100644
index 0000000..d44b3b3
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc3.c
@@ -0,0 +1,110 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nvc0_graph_init
+nvc3_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x00002834 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc3_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   3, 0x04, 0x00000000 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419880,   1, 0x04, 0x00000002 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvc3_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc3_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc0_graph_init_gpc,
+	nvc3_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc3_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc3),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvc3_grctx_oclass,
+	.sclass = nvc0_graph_sclass,
+	.mmio = nvc3_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
new file mode 100644
index 0000000..02845e5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
@@ -0,0 +1,141 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+struct nouveau_oclass
+nvc8_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0x9039, &nouveau_object_ofuncs },
+	{ 0x9097, &nouveau_object_ofuncs },
+	{ 0x90c0, &nouveau_object_ofuncs },
+	{ 0x9197, &nouveau_object_ofuncs },
+	{ 0x9297, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static struct nvc0_graph_init
+nvc8_graph_init_gpc[] = {
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x80000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000050 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc8_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   3, 0x04, 0x00000000 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100f02 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x06060618 },
+	{ 0x419ed0,   1, 0x04, 0x0eff0e38 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvc8_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc0_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc8_graph_init_gpc,
+	nvc8_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc8_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc8),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvc8_grctx_oclass,
+	.sclass = nvc8_graph_sclass,
+	.mmio = nvc8_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
new file mode 100644
index 0000000..5052d7a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
@@ -0,0 +1,167 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+#include "fuc/hubnvd7.fuc.h"
+
+struct nvc0_graph_ucode
+nvd7_graph_fecs_ucode = {
+	.code.data = nvd7_grhub_code,
+	.code.size = sizeof(nvd7_grhub_code),
+	.data.data = nvd7_grhub_data,
+	.data.size = sizeof(nvd7_grhub_data),
+};
+
+#include "fuc/gpcnvd7.fuc.h"
+
+struct nvc0_graph_ucode
+nvd7_graph_gpccs_ucode = {
+	.code.data = nvd7_grgpc_code,
+	.code.size = sizeof(nvd7_grgpc_code),
+	.data.data = nvd7_grgpc_data,
+	.data.size = sizeof(nvd7_grgpc_data),
+};
+
+static struct nvc0_graph_init
+nvd7_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   2, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000003 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   1, 0x04, 0x00000000 },
+	{ 0x418e20,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc8 },
+	{ 0x419850,   2, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x02001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_graph_init_tpc_0[] = {
+	{ 0x40402c,   1, 0x04, 0x00000000 },
+	{ 0x4040f0,   1, 0x04, 0x00000000 },
+	{ 0x404174,   1, 0x04, 0x00000000 },
+	{ 0x503018,   1, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvd7_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nvd9_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvd7_graph_init_gpc,
+	nvd7_graph_init_tpc,
+	nve4_graph_init_unk,
+	nvc0_graph_init_unk88xx,
+	nvd7_graph_init_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvd7_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xd7),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvd7_grctx_oclass,
+	.sclass = nvc8_graph_sclass,
+	.mmio = nvd7_graph_init_mmio,
+	.fecs.ucode = &nvd7_graph_fecs_ucode,
+	.gpccs.ucode = &nvd7_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvd9.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvd9.c
new file mode 100644
index 0000000..652098e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvd9.c
@@ -0,0 +1,165 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nvc0_graph_init
+nvd9_graph_init_unk64xx[] = {
+	{ 0x4064f0,   3, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x00002834 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{ 0x405928,   1, 0x04, 0x00000000 },
+	{ 0x40592c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418d2c,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   2, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000003 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   1, 0x04, 0x00000000 },
+	{ 0x418e20,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419810,   1, 0x04, 0x00000000 },
+	{ 0x419814,   1, 0x04, 0x00000004 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x0000a918 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419880,   1, 0x04, 0x00000002 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419bf8,   1, 0x04, 0x00000000 },
+	{ 0x419bfc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419d48,   1, 0x04, 0x00000000 },
+	{ 0x419d4c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x02001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvd9_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nvd9_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvd9_graph_init_gpc,
+	nvd9_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvd9_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xd9),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvd9_grctx_oclass,
+	.sclass = nvc8_graph_sclass,
+	.mmio = nvd9_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
deleted file mode 100644
index 678c16f..0000000
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
+++ /dev/null
@@ -1,807 +0,0 @@
-/*
- * Copyright 2012 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "nvc0.h"
-#include "fuc/hubnve0.fuc.h"
-#include "fuc/gpcnve0.fuc.h"
-
-/*******************************************************************************
- * Graphics object classes
- ******************************************************************************/
-
-static struct nouveau_oclass
-nve0_graph_sclass[] = {
-	{ 0x902d, &nouveau_object_ofuncs },
-	{ 0xa040, &nouveau_object_ofuncs },
-	{ 0xa097, &nouveau_object_ofuncs },
-	{ 0xa0c0, &nouveau_object_ofuncs },
-	{ 0xa0b5, &nouveau_object_ofuncs },
-	{}
-};
-
-/*******************************************************************************
- * PGRAPH context
- ******************************************************************************/
-
-static struct nouveau_oclass
-nve0_graph_cclass = {
-	.handle = NV_ENGCTX(GR, 0xe0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvc0_graph_context_ctor,
-		.dtor = nvc0_graph_context_dtor,
-		.init = _nouveau_graph_context_init,
-		.fini = _nouveau_graph_context_fini,
-		.rd32 = _nouveau_graph_context_rd32,
-		.wr32 = _nouveau_graph_context_wr32,
-	},
-};
-
-/*******************************************************************************
- * PGRAPH engine/subdev functions
- ******************************************************************************/
-
-static void
-nve0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
-{
-	u32 ustat = nv_rd32(priv, 0x409c18);
-
-	if (ustat & 0x00000001)
-		nv_error(priv, "CTXCTRL ucode error\n");
-	if (ustat & 0x00080000)
-		nv_error(priv, "CTXCTRL watchdog timeout\n");
-	if (ustat & ~0x00080001)
-		nv_error(priv, "CTXCTRL 0x%08x\n", ustat);
-
-	nvc0_graph_ctxctl_debug(priv);
-	nv_wr32(priv, 0x409c20, ustat);
-}
-
-static const struct nouveau_enum nve0_mp_warp_error[] = {
-	{ 0x00, "NO_ERROR" },
-	{ 0x01, "STACK_MISMATCH" },
-	{ 0x05, "MISALIGNED_PC" },
-	{ 0x08, "MISALIGNED_GPR" },
-	{ 0x09, "INVALID_OPCODE" },
-	{ 0x0d, "GPR_OUT_OF_BOUNDS" },
-	{ 0x0e, "MEM_OUT_OF_BOUNDS" },
-	{ 0x0f, "UNALIGNED_MEM_ACCESS" },
-	{ 0x11, "INVALID_PARAM" },
-	{}
-};
-
-static const struct nouveau_enum nve0_mp_global_error[] = {
-	{ 2, "MULTIPLE_WARP_ERRORS" },
-	{ 3, "OUT_OF_STACK_SPACE" },
-	{}
-};
-
-static const struct nouveau_enum nve0_gpc_rop_error[] = {
-	{ 1, "RT_PITCH_OVERRUN" },
-	{ 4, "RT_WIDTH_OVERRUN" },
-	{ 5, "RT_HEIGHT_OVERRUN" },
-	{ 7, "ZETA_STORAGE_TYPE_MISMATCH" },
-	{ 8, "RT_STORAGE_TYPE_MISMATCH" },
-	{ 10, "RT_LINEAR_MISMATCH" },
-	{}
-};
-
-static const struct nouveau_enum nve0_sked_error[] = {
-	{ 7, "CONSTANT_BUFFER_SIZE" },
-	{ 9, "LOCAL_MEMORY_SIZE_POS" },
-	{ 10, "LOCAL_MEMORY_SIZE_NEG" },
-	{ 11, "WARP_CSTACK_SIZE" },
-	{ 12, "TOTAL_TEMP_SIZE" },
-	{ 13, "REGISTER_COUNT" },
-	{ 18, "TOTAL_THREADS" },
-	{ 20, "PROGRAM_OFFSET" },
-	{ 21, "SHARED_MEMORY_SIZE" },
-	{ 25, "SHARED_CONFIG_TOO_SMALL" },
-	{ 26, "TOTAL_REGISTER_COUNT" },
-	{}
-};
-
-static void
-nve0_graph_mp_trap(struct nvc0_graph_priv *priv, int gpc, int tp)
-{
-	int i;
-	u32 werr = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x648));
-	u32 gerr = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x650));
-
-	nv_error(priv, "GPC%i/TP%i/MP trap:", gpc, tp);
-
-	for (i = 0; i <= 31; ++i) {
-		if (!(gerr & (1 << i)))
-			continue;
-		pr_cont(" ");
-		nouveau_enum_print(nve0_mp_global_error, i);
-	}
-	if (werr) {
-		pr_cont(" ");
-		nouveau_enum_print(nve0_mp_warp_error, werr & 0xffff);
-	}
-	pr_cont("\n");
-
-	/* disable MP trap to avoid spam */
-	nv_mask(priv, TPC_UNIT(gpc, tp, 0x50c), 0x2, 0x0);
-
-	/* TODO: figure out how to resume after an MP trap */
-}
-
-static void
-nve0_graph_tp_trap(struct nvc0_graph_priv *priv, int gpc, int tp)
-{
-	u32 stat = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x508));
-
-	if (stat & 0x1) {
-		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x224));
-		nv_error(priv, "GPC%i/TP%i/TEX trap: %08x\n",
-			 gpc, tp, trap);
-
-		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x224), 0xc0000000);
-		stat &= ~0x1;
-	}
-
-	if (stat & 0x2) {
-		nve0_graph_mp_trap(priv, gpc, tp);
-		stat &= ~0x2;
-	}
-
-	if (stat & 0x4) {
-		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x084));
-		nv_error(priv, "GPC%i/TP%i/POLY trap: %08x\n",
-			 gpc, tp, trap);
-
-		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x084), 0xc0000000);
-		stat &= ~0x4;
-	}
-
-	if (stat & 0x8) {
-		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x48c));
-		nv_error(priv, "GPC%i/TP%i/L1C trap: %08x\n",
-			 gpc, tp, trap);
-
-		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x48c), 0xc0000000);
-		stat &= ~0x8;
-	}
-
-	if (stat) {
-		nv_error(priv, "GPC%i/TP%i: unknown stat %08x\n",
-			 gpc, tp, stat);
-	}
-}
-
-static void
-nve0_graph_gpc_trap(struct nvc0_graph_priv *priv)
-{
-	const u32 mask = nv_rd32(priv, 0x400118);
-	int gpc;
-
-	for (gpc = 0; gpc < 4; ++gpc) {
-		u32 stat;
-		int tp;
-
-		if (!(mask & (1 << gpc)))
-			continue;
-		stat = nv_rd32(priv, GPC_UNIT(gpc, 0x2c90));
-
-		if (stat & 0x0001) {
-			u32 trap[4];
-			int i;
-
-			trap[0] = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
-			trap[1] = nv_rd32(priv, GPC_UNIT(gpc, 0x0434));
-			trap[2] = nv_rd32(priv, GPC_UNIT(gpc, 0x0438));
-			trap[3] = nv_rd32(priv, GPC_UNIT(gpc, 0x043c));
-
-			nv_error(priv, "GPC%i/PROP trap:", gpc);
-			for (i = 0; i <= 29; ++i) {
-				if (!(trap[0] & (1 << i)))
-					continue;
-				pr_cont(" ");
-				nouveau_enum_print(nve0_gpc_rop_error, i);
-			}
-			pr_cont("\n");
-
-			nv_error(priv, "x = %u, y = %u, "
-				 "format = %x, storage type = %x\n",
-				 trap[1] & 0xffff,
-				 trap[1] >> 16,
-				 (trap[2] >> 8) & 0x3f,
-				 trap[3] & 0xff);
-
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-			stat &= ~0x0001;
-		}
-
-		if (stat & 0x0002) {
-			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0900));
-			nv_error(priv, "GPC%i/ZCULL trap: %08x\n", gpc,
-				 trap);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-			stat &= ~0x0002;
-		}
-
-		if (stat & 0x0004) {
-			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x1028));
-			nv_error(priv, "GPC%i/CCACHE trap: %08x\n", gpc,
-				 trap);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-			stat &= ~0x0004;
-		}
-
-		if (stat & 0x0008) {
-			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0824));
-			nv_error(priv, "GPC%i/ESETUP trap %08x\n", gpc,
-				 trap);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-			stat &= ~0x0008;
-		}
-
-		for (tp = 0; tp < 8; ++tp) {
-			if (stat & (1 << (16 + tp)))
-				nve0_graph_tp_trap(priv, gpc, tp);
-		}
-		stat &= ~0xff0000;
-
-		if (stat) {
-			nv_error(priv, "GPC%i: unknown stat %08x\n",
-				 gpc, stat);
-		}
-	}
-}
-
-
-static void
-nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst,
-		struct nouveau_object *engctx)
-{
-	u32 trap = nv_rd32(priv, 0x400108);
-	int i;
-	int rop;
-
-	if (trap & 0x00000001) {
-		u32 stat = nv_rd32(priv, 0x404000);
-		nv_error(priv, "DISPATCH ch %d [0x%010llx %s] 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), stat);
-		nv_wr32(priv, 0x404000, 0xc0000000);
-		nv_wr32(priv, 0x400108, 0x00000001);
-		trap &= ~0x00000001;
-	}
-
-	if (trap & 0x00000010) {
-		u32 stat = nv_rd32(priv, 0x405840);
-		nv_error(priv, "SHADER ch %d [0x%010llx %s] 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), stat);
-		nv_wr32(priv, 0x405840, 0xc0000000);
-		nv_wr32(priv, 0x400108, 0x00000010);
-		trap &= ~0x00000010;
-	}
-
-	if (trap & 0x00000100) {
-		u32 stat = nv_rd32(priv, 0x407020);
-		nv_error(priv, "SKED ch %d [0x%010llx %s]:",
-			 chid, inst, nouveau_client_name(engctx));
-
-		for (i = 0; i <= 29; ++i) {
-			if (!(stat & (1 << i)))
-				continue;
-			pr_cont(" ");
-			nouveau_enum_print(nve0_sked_error, i);
-		}
-		pr_cont("\n");
-
-		if (stat & 0x3fffffff)
-			nv_wr32(priv, 0x407020, 0x40000000);
-		nv_wr32(priv, 0x400108, 0x00000100);
-		trap &= ~0x00000100;
-	}
-
-	if (trap & 0x01000000) {
-		nv_error(priv, "GPC ch %d [0x%010llx %s]:\n",
-			 chid, inst, nouveau_client_name(engctx));
-		nve0_graph_gpc_trap(priv);
-		trap &= ~0x01000000;
-	}
-
-	if (trap & 0x02000000) {
-		for (rop = 0; rop < priv->rop_nr; rop++) {
-			u32 statz = nv_rd32(priv, ROP_UNIT(rop, 0x070));
-			u32 statc = nv_rd32(priv, ROP_UNIT(rop, 0x144));
-			nv_error(priv,
-				 "ROP%d ch %d [0x%010llx %s] 0x%08x 0x%08x\n",
-				 rop, chid, inst, nouveau_client_name(engctx),
-				 statz, statc);
-			nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
-			nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
-		}
-		nv_wr32(priv, 0x400108, 0x02000000);
-		trap &= ~0x02000000;
-	}
-
-	if (trap) {
-		nv_error(priv, "TRAP ch %d [0x%010llx %s] 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), trap);
-		nv_wr32(priv, 0x400108, trap);
-	}
-}
-
-static void
-nve0_graph_intr(struct nouveau_subdev *subdev)
-{
-	struct nouveau_fifo *pfifo = nouveau_fifo(subdev);
-	struct nouveau_engine *engine = nv_engine(subdev);
-	struct nouveau_object *engctx;
-	struct nouveau_handle *handle;
-	struct nvc0_graph_priv *priv = (void *)subdev;
-	u64 inst = nv_rd32(priv, 0x409b00) & 0x0fffffff;
-	u32 stat = nv_rd32(priv, 0x400100);
-	u32 addr = nv_rd32(priv, 0x400704);
-	u32 mthd = (addr & 0x00003ffc);
-	u32 subc = (addr & 0x00070000) >> 16;
-	u32 data = nv_rd32(priv, 0x400708);
-	u32 code = nv_rd32(priv, 0x400110);
-	u32 class = nv_rd32(priv, 0x404200 + (subc * 4));
-	int chid;
-
-	engctx = nouveau_engctx_get(engine, inst);
-	chid   = pfifo->chid(pfifo, engctx);
-
-	if (stat & 0x00000010) {
-		handle = nouveau_handle_get_class(engctx, class);
-		if (!handle || nv_call(handle->object, mthd, data)) {
-			nv_error(priv,
-				 "ILLEGAL_MTHD ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
-				 chid, inst, nouveau_client_name(engctx), subc,
-				 class, mthd, data);
-		}
-		nouveau_handle_put(handle);
-		nv_wr32(priv, 0x400100, 0x00000010);
-		stat &= ~0x00000010;
-	}
-
-	if (stat & 0x00000020) {
-		nv_error(priv,
-			 "ILLEGAL_CLASS ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), subc, class,
-			 mthd, data);
-		nv_wr32(priv, 0x400100, 0x00000020);
-		stat &= ~0x00000020;
-	}
-
-	if (stat & 0x00100000) {
-		nv_error(priv, "DATA_ERROR [");
-		nouveau_enum_print(nv50_data_error_names, code);
-		pr_cont("] ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
-			chid, inst, nouveau_client_name(engctx), subc, class,
-			mthd, data);
-		nv_wr32(priv, 0x400100, 0x00100000);
-		stat &= ~0x00100000;
-	}
-
-	if (stat & 0x00200000) {
-		nve0_graph_trap_isr(priv, chid, inst, engctx);
-		nv_wr32(priv, 0x400100, 0x00200000);
-		stat &= ~0x00200000;
-	}
-
-	if (stat & 0x00080000) {
-		nve0_graph_ctxctl_isr(priv);
-		nv_wr32(priv, 0x400100, 0x00080000);
-		stat &= ~0x00080000;
-	}
-
-	if (stat) {
-		nv_error(priv, "unknown stat 0x%08x\n", stat);
-		nv_wr32(priv, 0x400100, stat);
-	}
-
-	nv_wr32(priv, 0x400500, 0x00010001);
-	nouveau_engctx_put(engctx);
-}
-
-static int
-nve0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	       struct nouveau_oclass *oclass, void *data, u32 size,
-	       struct nouveau_object **pobject)
-{
-	struct nouveau_device *device = nv_device(parent);
-	struct nvc0_graph_priv *priv;
-	int ret, i;
-
-	ret = nouveau_graph_create(parent, engine, oclass, true, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	nv_subdev(priv)->unit = 0x18001000;
-	nv_subdev(priv)->intr = nve0_graph_intr;
-	nv_engine(priv)->cclass = &nve0_graph_cclass;
-	nv_engine(priv)->sclass = nve0_graph_sclass;
-
-	priv->base.units = nvc0_graph_units;
-
-	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
-		nv_info(priv, "using external firmware\n");
-		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
-		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
-			return -EINVAL;
-		priv->firmware = true;
-	}
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b4);
-	if (ret)
-		return ret;
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b8);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < 0x1000; i += 4) {
-		nv_wo32(priv->unk4188b4, i, 0x00000010);
-		nv_wo32(priv->unk4188b8, i, 0x00000010);
-	}
-
-	priv->gpc_nr =  nv_rd32(priv, 0x409604) & 0x0000001f;
-	priv->rop_nr = (nv_rd32(priv, 0x409604) & 0x001f0000) >> 16;
-	for (i = 0; i < priv->gpc_nr; i++) {
-		priv->tpc_nr[i] = nv_rd32(priv, GPC_UNIT(i, 0x2608));
-		priv->tpc_total += priv->tpc_nr[i];
-	}
-
-	switch (nv_device(priv)->chipset) {
-	case 0xe4:
-		if (priv->tpc_total == 8)
-			priv->magic_not_rop_nr = 3;
-		else
-		if (priv->tpc_total == 7)
-			priv->magic_not_rop_nr = 1;
-		break;
-	case 0xe7:
-	case 0xe6:
-		priv->magic_not_rop_nr = 1;
-		break;
-	default:
-		break;
-	}
-
-	return 0;
-}
-
-static void
-nve0_graph_init_obj418880(struct nvc0_graph_priv *priv)
-{
-	int i;
-
-	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
-	for (i = 0; i < 4; i++)
-		nv_wr32(priv, GPC_BCAST(0x0888) + (i * 4), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
-	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
-}
-
-static void
-nve0_graph_init_regs(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x400080, 0x003083c2);
-	nv_wr32(priv, 0x400088, 0x0001ffe7);
-	nv_wr32(priv, 0x40008c, 0x00000000);
-	nv_wr32(priv, 0x400090, 0x00000030);
-	nv_wr32(priv, 0x40013c, 0x003901f7);
-	nv_wr32(priv, 0x400140, 0x00000100);
-	nv_wr32(priv, 0x400144, 0x00000000);
-	nv_wr32(priv, 0x400148, 0x00000110);
-	nv_wr32(priv, 0x400138, 0x00000000);
-	nv_wr32(priv, 0x400130, 0x00000000);
-	nv_wr32(priv, 0x400134, 0x00000000);
-	nv_wr32(priv, 0x400124, 0x00000002);
-}
-
-static void
-nve0_graph_init_units(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x409ffc, 0x00000000);
-	nv_wr32(priv, 0x409c14, 0x00003e3e);
-	nv_wr32(priv, 0x409c24, 0x000f0000);
-
-	nv_wr32(priv, 0x404000, 0xc0000000);
-	nv_wr32(priv, 0x404600, 0xc0000000);
-	nv_wr32(priv, 0x408030, 0xc0000000);
-	nv_wr32(priv, 0x404490, 0xc0000000);
-	nv_wr32(priv, 0x406018, 0xc0000000);
-	nv_wr32(priv, 0x407020, 0xc0000000);
-	nv_wr32(priv, 0x405840, 0xc0000000);
-	nv_wr32(priv, 0x405844, 0x00ffffff);
-
-	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
-	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
-
-}
-
-static void
-nve0_graph_init_gpc_0(struct nvc0_graph_priv *priv)
-{
-	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
-	u32 data[TPC_MAX / 8];
-	u8  tpcnr[GPC_MAX];
-	int i, gpc, tpc;
-
-	nv_wr32(priv, GPC_UNIT(0, 0x3018), 0x00000001);
-
-	memset(data, 0x00, sizeof(data));
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
-		do {
-			gpc = (gpc + 1) % priv->gpc_nr;
-		} while (!tpcnr[gpc]);
-		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-		data[i / 8] |= tpc << ((i % 8) * 4);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
-	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
-	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
-	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0914), priv->magic_not_rop_nr << 8 |
-						  priv->tpc_nr[gpc]);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 | priv->tpc_total);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918);
-	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
-}
-
-static void
-nve0_graph_init_gpc_1(struct nvc0_graph_priv *priv)
-{
-	int gpc, tpc;
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x3038), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
-		}
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
-	}
-}
-
-static void
-nve0_graph_init_rop(struct nvc0_graph_priv *priv)
-{
-	int rop;
-
-	for (rop = 0; rop < priv->rop_nr; rop++) {
-		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
-		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
-	}
-}
-
-static int
-nve0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
-{
-	u32 r000260;
-	int i;
-
-	if (priv->firmware) {
-		/* load fuc microcode */
-		r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
-		nvc0_graph_init_fw(priv, 0x409000, &priv->fuc409c, &priv->fuc409d);
-		nvc0_graph_init_fw(priv, 0x41a000, &priv->fuc41ac, &priv->fuc41ad);
-		nv_wr32(priv, 0x000260, r000260);
-
-		/* start both of them running */
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x41a10c, 0x00000000);
-		nv_wr32(priv, 0x40910c, 0x00000000);
-		nv_wr32(priv, 0x41a100, 0x00000002);
-		nv_wr32(priv, 0x409100, 0x00000002);
-		if (!nv_wait(priv, 0x409800, 0x00000001, 0x00000001))
-			nv_error(priv, "0x409800 wait failed\n");
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x7fffffff);
-		nv_wr32(priv, 0x409504, 0x00000021);
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x00000000);
-		nv_wr32(priv, 0x409504, 0x00000010);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x10 timeout\n");
-			return -EBUSY;
-		}
-		priv->size = nv_rd32(priv, 0x409800);
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x00000000);
-		nv_wr32(priv, 0x409504, 0x00000016);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x16 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x00000000);
-		nv_wr32(priv, 0x409504, 0x00000025);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x25 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409800, 0x00000000);
-		nv_wr32(priv, 0x409500, 0x00000001);
-		nv_wr32(priv, 0x409504, 0x00000030);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x30 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409810, 0xb00095c8);
-		nv_wr32(priv, 0x409800, 0x00000000);
-		nv_wr32(priv, 0x409500, 0x00000001);
-		nv_wr32(priv, 0x409504, 0x00000031);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x31 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409810, 0x00080420);
-		nv_wr32(priv, 0x409800, 0x00000000);
-		nv_wr32(priv, 0x409500, 0x00000001);
-		nv_wr32(priv, 0x409504, 0x00000032);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x32 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409614, 0x00000070);
-		nv_wr32(priv, 0x409614, 0x00000770);
-		nv_wr32(priv, 0x40802c, 0x00000001);
-
-		if (priv->data == NULL) {
-			int ret = nve0_grctx_generate(priv);
-			if (ret) {
-				nv_error(priv, "failed to construct context\n");
-				return ret;
-			}
-		}
-
-		return 0;
-	}
-
-	/* load HUB microcode */
-	r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
-	nv_wr32(priv, 0x4091c0, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grhub_data) / 4; i++)
-		nv_wr32(priv, 0x4091c4, nve0_grhub_data[i]);
-
-	nv_wr32(priv, 0x409180, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grhub_code) / 4; i++) {
-		if ((i & 0x3f) == 0)
-			nv_wr32(priv, 0x409188, i >> 6);
-		nv_wr32(priv, 0x409184, nve0_grhub_code[i]);
-	}
-
-	/* load GPC microcode */
-	nv_wr32(priv, 0x41a1c0, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grgpc_data) / 4; i++)
-		nv_wr32(priv, 0x41a1c4, nve0_grgpc_data[i]);
-
-	nv_wr32(priv, 0x41a180, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grgpc_code) / 4; i++) {
-		if ((i & 0x3f) == 0)
-			nv_wr32(priv, 0x41a188, i >> 6);
-		nv_wr32(priv, 0x41a184, nve0_grgpc_code[i]);
-	}
-	nv_wr32(priv, 0x000260, r000260);
-
-	/* start HUB ucode running, it'll init the GPCs */
-	nv_wr32(priv, 0x409800, nv_device(priv)->chipset);
-	nv_wr32(priv, 0x40910c, 0x00000000);
-	nv_wr32(priv, 0x409100, 0x00000002);
-	if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000)) {
-		nv_error(priv, "HUB_INIT timed out\n");
-		nvc0_graph_ctxctl_debug(priv);
-		return -EBUSY;
-	}
-
-	priv->size = nv_rd32(priv, 0x409804);
-	if (priv->data == NULL) {
-		int ret = nve0_grctx_generate(priv);
-		if (ret) {
-			nv_error(priv, "failed to construct context\n");
-			return ret;
-		}
-	}
-
-	return 0;
-}
-
-static int
-nve0_graph_init(struct nouveau_object *object)
-{
-	struct nvc0_graph_priv *priv = (void *)object;
-	int ret;
-
-	ret = nouveau_graph_init(&priv->base);
-	if (ret)
-		return ret;
-
-	nve0_graph_init_obj418880(priv);
-	nve0_graph_init_regs(priv);
-	nve0_graph_init_gpc_0(priv);
-
-	nv_wr32(priv, 0x400500, 0x00010001);
-	nv_wr32(priv, 0x400100, 0xffffffff);
-	nv_wr32(priv, 0x40013c, 0xffffffff);
-
-	nve0_graph_init_units(priv);
-	nve0_graph_init_gpc_1(priv);
-	nve0_graph_init_rop(priv);
-
-	nv_wr32(priv, 0x400108, 0xffffffff);
-	nv_wr32(priv, 0x400138, 0xffffffff);
-	nv_wr32(priv, 0x400118, 0xffffffff);
-	nv_wr32(priv, 0x400130, 0xffffffff);
-	nv_wr32(priv, 0x40011c, 0xffffffff);
-	nv_wr32(priv, 0x400134, 0xffffffff);
-	nv_wr32(priv, 0x400054, 0x34ce3464);
-
-	ret = nve0_graph_init_ctxctl(priv);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-struct nouveau_oclass
-nve0_graph_oclass = {
-	.handle = NV_ENGINE(GR, 0xe0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nve0_graph_ctor,
-		.dtor = nvc0_graph_dtor,
-		.init = nve0_graph_init,
-		.fini = _nouveau_graph_fini,
-	},
-};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
new file mode 100644
index 0000000..05ec09c
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
@@ -0,0 +1,354 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nve4_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0xa040, &nouveau_object_ofuncs },
+	{ 0xa097, &nouveau_object_ofuncs },
+	{ 0xa0c0, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nvc0_graph_init
+nve4_graph_init_regs[] = {
+	{ 0x400080,   1, 0x04, 0x003083c2 },
+	{ 0x400088,   1, 0x04, 0x0001ffe7 },
+	{ 0x40008c,   1, 0x04, 0x00000000 },
+	{ 0x400090,   1, 0x04, 0x00000030 },
+	{ 0x40013c,   1, 0x04, 0x003901f7 },
+	{ 0x400140,   1, 0x04, 0x00000100 },
+	{ 0x400144,   1, 0x04, 0x00000000 },
+	{ 0x400148,   1, 0x04, 0x00000110 },
+	{ 0x400138,   1, 0x04, 0x00000000 },
+	{ 0x400130,   2, 0x04, 0x00000000 },
+	{ 0x400124,   1, 0x04, 0x00000002 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x0000ff34 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{ 0x405928,   1, 0x04, 0x00000000 },
+	{ 0x40592c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_unk70xx[] = {
+	{ 0x407010,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_graph_init_unk5bxx[] = {
+	{ 0x405b50,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418d2c,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   2, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000060 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   1, 0x04, 0x00000000 },
+	{ 0x418e20,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_tpc[] = {
+	{ 0x419d0c,   1, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x419850,   1, 0x04, 0x00000004 },
+	{ 0x419854,   2, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x01000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00b08bea },
+	{ 0x419c84,   1, 0x04, 0x00010384 },
+	{ 0x419cbc,   1, 0x04, 0x28137646 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c80,   1, 0x04, 0x00020232 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ee4,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00000000 },
+	{ 0x419eb4,   1, 0x04, 0x00000000 },
+	{ 0x419eb8,   3, 0x04, 0x00000000 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f74,   1, 0x04, 0x00000555 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_graph_init_unk[] = {
+	{ 0x41be04,   1, 0x04, 0x00000000 },
+	{ 0x41be08,   1, 0x04, 0x00000004 },
+	{ 0x41be0c,   1, 0x04, 0x00000000 },
+	{ 0x41be10,   1, 0x04, 0x003b8bc7 },
+	{ 0x41be14,   2, 0x04, 0x00000000 },
+	{ 0x41bfd4,   1, 0x04, 0x00800000 },
+	{ 0x41bfdc,   1, 0x04, 0x00000000 },
+	{ 0x41bff8,   1, 0x04, 0x00000000 },
+	{ 0x41bffc,   1, 0x04, 0x00000000 },
+	{ 0x41becc,   1, 0x04, 0x00000000 },
+	{ 0x41bee8,   1, 0x04, 0x00000000 },
+	{ 0x41beec,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_graph_init_unk88xx[] = {
+	{ 0x40880c,   1, 0x04, 0x00000000 },
+	{ 0x408850,   1, 0x04, 0x00000004 },
+	{ 0x408910,   9, 0x04, 0x00000000 },
+	{ 0x408950,   1, 0x04, 0x00000000 },
+	{ 0x408954,   1, 0x04, 0x0000ffff },
+	{ 0x408958,   1, 0x04, 0x00000034 },
+	{ 0x408984,   1, 0x04, 0x00000000 },
+	{ 0x408988,   1, 0x04, 0x08040201 },
+	{ 0x40898c,   1, 0x04, 0x80402010 },
+	{}
+};
+
+int
+nve4_graph_init(struct nouveau_object *object)
+{
+	struct nvc0_graph_oclass *oclass = (void *)object->oclass;
+	struct nvc0_graph_priv *priv = (void *)object;
+	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
+	u32 data[TPC_MAX / 8] = {};
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc, rop;
+	int ret, i;
+
+	ret = nouveau_graph_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0888), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x088c), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0890), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0894), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
+	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
+
+	for (i = 0; oclass->mmio[i]; i++)
+		nvc0_graph_mmio(priv, oclass->mmio[i]);
+
+	nv_wr32(priv, GPC_UNIT(0, 0x3018), 0x00000001);
+
+	memset(data, 0x00, sizeof(data));
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
+
+		data[i / 8] |= tpc << ((i % 8) * 4);
+	}
+
+	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
+	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
+	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
+	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0914),
+			priv->magic_not_rop_nr << 8 | priv->tpc_nr[gpc]);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 |
+			priv->tpc_total);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
+	}
+
+	nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918);
+	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
+
+	nv_wr32(priv, 0x400500, 0x00010001);
+
+	nv_wr32(priv, 0x400100, 0xffffffff);
+	nv_wr32(priv, 0x40013c, 0xffffffff);
+
+	nv_wr32(priv, 0x409ffc, 0x00000000);
+	nv_wr32(priv, 0x409c14, 0x00003e3e);
+	nv_wr32(priv, 0x409c24, 0x000f0001);
+	nv_wr32(priv, 0x404000, 0xc0000000);
+	nv_wr32(priv, 0x404600, 0xc0000000);
+	nv_wr32(priv, 0x408030, 0xc0000000);
+	nv_wr32(priv, 0x404490, 0xc0000000);
+	nv_wr32(priv, 0x406018, 0xc0000000);
+	nv_wr32(priv, 0x407020, 0x40000000);
+	nv_wr32(priv, 0x405840, 0xc0000000);
+	nv_wr32(priv, 0x405844, 0x00ffffff);
+	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
+	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x3038), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
+		}
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
+	}
+
+	for (rop = 0; rop < priv->rop_nr; rop++) {
+		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
+		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
+	}
+
+	nv_wr32(priv, 0x400108, 0xffffffff);
+	nv_wr32(priv, 0x400138, 0xffffffff);
+	nv_wr32(priv, 0x400118, 0xffffffff);
+	nv_wr32(priv, 0x400130, 0xffffffff);
+	nv_wr32(priv, 0x40011c, 0xffffffff);
+	nv_wr32(priv, 0x400134, 0xffffffff);
+
+	nv_wr32(priv, 0x400054, 0x34ce3464);
+	return nvc0_graph_init_ctxctl(priv);
+}
+
+static struct nvc0_graph_init *
+nve4_graph_init_mmio[] = {
+	nve4_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nve4_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nve4_graph_init_unk70xx,
+	nve4_graph_init_unk5bxx,
+	nve4_graph_init_gpc,
+	nve4_graph_init_tpc,
+	nve4_graph_init_unk,
+	nve4_graph_init_unk88xx,
+	NULL
+};
+
+#include "fuc/hubnve0.fuc.h"
+
+static struct nvc0_graph_ucode
+nve4_graph_fecs_ucode = {
+	.code.data = nve0_grhub_code,
+	.code.size = sizeof(nve0_grhub_code),
+	.data.data = nve0_grhub_data,
+	.data.size = sizeof(nve0_grhub_data),
+};
+
+#include "fuc/gpcnve0.fuc.h"
+
+static struct nvc0_graph_ucode
+nve4_graph_gpccs_ucode = {
+	.code.data = nve0_grgpc_code,
+	.code.size = sizeof(nve0_grgpc_code),
+	.data.data = nve0_grgpc_data,
+	.data.size = sizeof(nve0_grgpc_data),
+};
+
+struct nouveau_oclass *
+nve4_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xe4),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nve4_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nve4_grctx_oclass,
+	.sclass = nve4_graph_sclass,
+	.mmio = nve4_graph_init_mmio,
+	.fecs.ucode = &nve4_graph_fecs_ucode,
+	.gpccs.ucode = &nve4_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
new file mode 100644
index 0000000..2f0ac78
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
@@ -0,0 +1,248 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nvf0_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0xa140, &nouveau_object_ofuncs },
+	{ 0xa197, &nouveau_object_ofuncs },
+	{ 0xa1c0, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk40xx[] = {
+	{ 0x40415c,   1, 0x04, 0x00000000 },
+	{ 0x404170,   1, 0x04, 0x00000000 },
+	{ 0x4041b4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x0000ff00 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{ 0x405928,   1, 0x04, 0x00000000 },
+	{ 0x40592c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk70xx[] = {
+	{ 0x407010,   1, 0x04, 0x00000000 },
+	{ 0x407040,   1, 0x04, 0x80440424 },
+	{ 0x407048,   1, 0x04, 0x0000000a },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk5bxx[] = {
+	{ 0x405b44,   1, 0x04, 0x00000000 },
+	{ 0x405b50,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418d2c,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000400 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   1, 0x04, 0x00000000 },
+	{ 0x418f24,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000000 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   2, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_tpc[] = {
+	{ 0x419d0c,   1, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419aec,   1, 0x04, 0x00000000 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x419aa8,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x419850,   1, 0x04, 0x00000004 },
+	{ 0x419854,   2, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x01000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00b08bea },
+	{ 0x419c84,   1, 0x04, 0x00010384 },
+	{ 0x419cbc,   1, 0x04, 0x281b3646 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c80,   1, 0x04, 0x00020230 },
+	{ 0x419ccc,   2, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000080 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ee4,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00000000 },
+	{ 0x419eb4,   1, 0x04, 0x00000000 },
+	{ 0x419ebc,   2, 0x04, 0x00000000 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419ed0,   1, 0x04, 0x00003234 },
+	{ 0x419f74,   1, 0x04, 0x00015555 },
+	{ 0x419f80,   4, 0x04, 0x00000000 },
+	{}
+};
+
+static int
+nvf0_graph_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nvc0_graph_priv *priv = (void *)object;
+	static const struct {
+		u32 addr;
+		u32 data;
+	} magic[] = {
+		{ 0x020520, 0xfffffffc },
+		{ 0x020524, 0xfffffffe },
+		{ 0x020524, 0xfffffffc },
+		{ 0x020524, 0xfffffff8 },
+		{ 0x020524, 0xffffffe0 },
+		{ 0x020530, 0xfffffffe },
+		{ 0x02052c, 0xfffffffa },
+		{ 0x02052c, 0xfffffff0 },
+		{ 0x02052c, 0xffffffc0 },
+		{ 0x02052c, 0xffffff00 },
+		{ 0x02052c, 0xfffffc00 },
+		{ 0x02052c, 0xfffcfc00 },
+		{ 0x02052c, 0xfff0fc00 },
+		{ 0x02052c, 0xff80fc00 },
+		{ 0x020528, 0xfffffffe },
+		{ 0x020528, 0xfffffffc },
+	};
+	int i;
+
+	nv_mask(priv, 0x000200, 0x08001000, 0x00000000);
+	nv_mask(priv, 0x0206b4, 0x00000000, 0x00000000);
+	for (i = 0; i < ARRAY_SIZE(magic); i++) {
+		nv_wr32(priv, magic[i].addr, magic[i].data);
+		nv_wait(priv, magic[i].addr, 0x80000000, 0x00000000);
+	}
+
+	return nouveau_graph_fini(&priv->base, suspend);
+}
+
+static struct nvc0_graph_init *
+nvf0_graph_init_mmio[] = {
+	nve4_graph_init_regs,
+	nvf0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nvf0_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvf0_graph_init_unk70xx,
+	nvf0_graph_init_unk5bxx,
+	nvf0_graph_init_gpc,
+	nvf0_graph_init_tpc,
+	nve4_graph_init_unk,
+	nve4_graph_init_unk88xx,
+	NULL
+};
+
+#include "fuc/hubnvf0.fuc.h"
+
+static struct nvc0_graph_ucode
+nvf0_graph_fecs_ucode = {
+	.code.data = nvf0_grhub_code,
+	.code.size = sizeof(nvf0_grhub_code),
+	.data.data = nvf0_grhub_data,
+	.data.size = sizeof(nvf0_grhub_data),
+};
+
+#include "fuc/gpcnvf0.fuc.h"
+
+static struct nvc0_graph_ucode
+nvf0_graph_gpccs_ucode = {
+	.code.data = nvf0_grgpc_code,
+	.code.size = sizeof(nvf0_grgpc_code),
+	.data.data = nvf0_grgpc_data,
+	.data.size = sizeof(nvf0_grgpc_data),
+};
+
+struct nouveau_oclass *
+nvf0_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xf0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nve4_graph_init,
+		.fini = nvf0_graph_fini,
+	},
+	.cclass = &nvf0_grctx_oclass,
+	.sclass =  nvf0_graph_sclass,
+	.mmio = nvf0_graph_init_mmio,
+	.fecs.ucode = 0 ? &nvf0_graph_fecs_ucode : NULL,
+	.gpccs.ucode = &nvf0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
index bc7d12b..37a2bd9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
@@ -125,13 +125,6 @@ nv50_mpeg_cclass = {
  * PMPEG engine/subdev functions
  ******************************************************************************/
 
-int
-nv50_mpeg_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x08);
-	return 0;
-}
-
 void
 nv50_mpeg_intr(struct nouveau_subdev *subdev)
 {
@@ -191,7 +184,6 @@ nv50_mpeg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv50_vpe_intr;
 	nv_engine(priv)->cclass = &nv50_mpeg_cclass;
 	nv_engine(priv)->sclass = nv50_mpeg_sclass;
-	nv_engine(priv)->tlb_flush = nv50_mpeg_tlb_flush;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
index 8f805b4..96f5aa9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
@@ -88,7 +88,6 @@ nv84_mpeg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv50_mpeg_intr;
 	nv_engine(priv)->cclass = &nv84_mpeg_cclass;
 	nv_engine(priv)->sclass = nv84_mpeg_sclass;
-	nv_engine(priv)->tlb_flush = nv50_mpeg_tlb_flush;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c
index ebf0d86..98072c1 100644
--- a/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c
@@ -22,8 +22,7 @@
  * Authors: Maarten Lankhorst
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/ppp.h>
 
 struct nvc0_ppp_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c
index 261cd96..fd6272b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c
@@ -19,24 +19,19 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs, Ilia Mirkin
  */
 
-#include <core/engctx.h>
-#include <core/class.h>
-
+#include <engine/xtensa.h>
 #include <engine/vp.h>
 
-struct nv84_vp_priv {
-	struct nouveau_engine base;
-};
-
 /*******************************************************************************
  * VP object classes
  ******************************************************************************/
 
 static struct nouveau_oclass
 nv84_vp_sclass[] = {
+	{ 0x7476, &nouveau_object_ofuncs },
 	{},
 };
 
@@ -48,7 +43,7 @@ static struct nouveau_oclass
 nv84_vp_cclass = {
 	.handle = NV_ENGCTX(VP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = _nouveau_engctx_ctor,
+		.ctor = _nouveau_xtensa_engctx_ctor,
 		.dtor = _nouveau_engctx_dtor,
 		.init = _nouveau_engctx_init,
 		.fini = _nouveau_engctx_fini,
@@ -66,10 +61,10 @@ nv84_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
 {
-	struct nv84_vp_priv *priv;
+	struct nouveau_xtensa *priv;
 	int ret;
 
-	ret = nouveau_engine_create(parent, engine, oclass, true,
+	ret = nouveau_xtensa_create(parent, engine, oclass, 0xf000, true,
 				    "PVP", "vp", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -78,6 +73,8 @@ nv84_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->unit = 0x01020000;
 	nv_engine(priv)->cclass = &nv84_vp_cclass;
 	nv_engine(priv)->sclass = nv84_vp_sclass;
+	priv->fifo_val = 0x111;
+	priv->unkd28 = 0x9c544;
 	return 0;
 }
 
@@ -86,8 +83,10 @@ nv84_vp_oclass = {
 	.handle = NV_ENGINE(VP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv84_vp_ctor,
-		.dtor = _nouveau_engine_dtor,
-		.init = _nouveau_engine_init,
-		.fini = _nouveau_engine_fini,
+		.dtor = _nouveau_xtensa_dtor,
+		.init = _nouveau_xtensa_init,
+		.fini = _nouveau_xtensa_fini,
+		.rd32 = _nouveau_xtensa_rd32,
+		.wr32 = _nouveau_xtensa_wr32,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nv98.c b/drivers/gpu/drm/nouveau/core/engine/vp/nv98.c
new file mode 100644
index 0000000..8a8236b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nv98.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/engctx.h>
+#include <core/class.h>
+
+#include <engine/vp.h>
+
+struct nv98_vp_priv {
+	struct nouveau_engine base;
+};
+
+/*******************************************************************************
+ * VP object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_vp_sclass[] = {
+	{},
+};
+
+/*******************************************************************************
+ * PVP context
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_vp_cclass = {
+	.handle = NV_ENGCTX(VP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_engctx_ctor,
+		.dtor = _nouveau_engctx_dtor,
+		.init = _nouveau_engctx_init,
+		.fini = _nouveau_engctx_fini,
+		.rd32 = _nouveau_engctx_rd32,
+		.wr32 = _nouveau_engctx_wr32,
+	},
+};
+
+/*******************************************************************************
+ * PVP engine/subdev functions
+ ******************************************************************************/
+
+static int
+nv98_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	     struct nouveau_oclass *oclass, void *data, u32 size,
+	     struct nouveau_object **pobject)
+{
+	struct nv98_vp_priv *priv;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true,
+				    "PVP", "vp", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->unit = 0x01020000;
+	nv_engine(priv)->cclass = &nv98_vp_cclass;
+	nv_engine(priv)->sclass = nv98_vp_sclass;
+	return 0;
+}
+
+struct nouveau_oclass
+nv98_vp_oclass = {
+	.handle = NV_ENGINE(VP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv98_vp_ctor,
+		.dtor = _nouveau_engine_dtor,
+		.init = _nouveau_engine_init,
+		.fini = _nouveau_engine_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c
index f761949..1879229 100644
--- a/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c
@@ -22,8 +22,7 @@
  * Authors: Maarten Lankhorst
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/vp.h>
 
 struct nvc0_vp_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c
index 2384ce5..d28ecbf 100644
--- a/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c
@@ -22,8 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/vp.h>
 
 struct nve0_vp_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/xtensa.c b/drivers/gpu/drm/nouveau/core/engine/xtensa.c
new file mode 100644
index 0000000..0639bc5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/xtensa.c
@@ -0,0 +1,170 @@
+/*
+ * Copyright 2013 Ilia Mirkin
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <engine/xtensa.h>
+
+u32
+_nouveau_xtensa_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct nouveau_xtensa *xtensa = (void *)object;
+	return nv_rd32(xtensa, xtensa->addr + addr);
+}
+
+void
+_nouveau_xtensa_wr32(struct nouveau_object *object, u64 addr, u32 data)
+{
+	struct nouveau_xtensa *xtensa = (void *)object;
+	nv_wr32(xtensa, xtensa->addr + addr, data);
+}
+
+int
+_nouveau_xtensa_engctx_ctor(struct nouveau_object *parent,
+			    struct nouveau_object *engine,
+			    struct nouveau_oclass *oclass, void *data, u32 size,
+			    struct nouveau_object **pobject)
+{
+	struct nouveau_engctx *engctx;
+	int ret;
+
+	ret = nouveau_engctx_create(parent, engine, oclass, NULL,
+				    0x10000, 0x1000,
+				    NVOBJ_FLAG_ZERO_ALLOC, &engctx);
+	*pobject = nv_object(engctx);
+	return ret;
+}
+
+void
+_nouveau_xtensa_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_xtensa *xtensa = (void *)subdev;
+	u32 unk104 = nv_ro32(xtensa, 0xd04);
+	u32 intr = nv_ro32(xtensa, 0xc20);
+	u32 chan = nv_ro32(xtensa, 0xc28);
+	u32 unk10c = nv_ro32(xtensa, 0xd0c);
+
+	if (intr & 0x10)
+		nv_warn(xtensa, "Watchdog interrupt, engine hung.\n");
+	nv_wo32(xtensa, 0xc20, intr);
+	intr = nv_ro32(xtensa, 0xc20);
+	if (unk104 == 0x10001 && unk10c == 0x200 && chan && !intr) {
+		nv_debug(xtensa, "Enabling FIFO_CTRL\n");
+		nv_mask(xtensa, xtensa->addr + 0xd94, 0, xtensa->fifo_val);
+	}
+}
+
+int
+nouveau_xtensa_create_(struct nouveau_object *parent,
+		       struct nouveau_object *engine,
+		       struct nouveau_oclass *oclass, u32 addr, bool enable,
+		       const char *iname, const char *fname,
+		       int length, void **pobject)
+{
+	struct nouveau_xtensa *xtensa;
+	int ret;
+
+	ret = nouveau_engine_create_(parent, engine, oclass, enable, iname,
+				     fname, length, pobject);
+	xtensa = *pobject;
+	if (ret)
+		return ret;
+
+	nv_subdev(xtensa)->intr = _nouveau_xtensa_intr;
+
+	xtensa->addr = addr;
+
+	return 0;
+}
+
+int
+_nouveau_xtensa_init(struct nouveau_object *object)
+{
+	struct nouveau_device *device = nv_device(object);
+	struct nouveau_xtensa *xtensa = (void *)object;
+	const struct firmware *fw;
+	char name[32];
+	int i, ret;
+	u32 tmp;
+
+	ret = nouveau_engine_init(&xtensa->base);
+	if (ret)
+		return ret;
+
+	if (!xtensa->gpu_fw) {
+		snprintf(name, sizeof(name), "nouveau/nv84_xuc%03x",
+			 xtensa->addr >> 12);
+
+		ret = request_firmware(&fw, name, &device->pdev->dev);
+		if (ret) {
+			nv_warn(xtensa, "unable to load firmware %s\n", name);
+			return ret;
+		}
+
+		ret = nouveau_gpuobj_new(object, NULL, fw->size, 0x1000, 0,
+					 &xtensa->gpu_fw);
+		if (ret) {
+			release_firmware(fw);
+			return ret;
+		}
+
+		nv_debug(xtensa, "Loading firmware to address: 0x%llx\n",
+			 xtensa->gpu_fw->addr);
+
+		for (i = 0; i < fw->size / 4; i++)
+			nv_wo32(xtensa->gpu_fw, i * 4, *((u32 *)fw->data + i));
+		release_firmware(fw);
+	}
+
+	nv_wo32(xtensa, 0xd10, 0x1fffffff); /* ?? */
+	nv_wo32(xtensa, 0xd08, 0x0fffffff); /* ?? */
+
+	nv_wo32(xtensa, 0xd28, xtensa->unkd28); /* ?? */
+	nv_wo32(xtensa, 0xc20, 0x3f); /* INTR */
+	nv_wo32(xtensa, 0xd84, 0x3f); /* INTR_EN */
+
+	nv_wo32(xtensa, 0xcc0, xtensa->gpu_fw->addr >> 8); /* XT_REGION_BASE */
+	nv_wo32(xtensa, 0xcc4, 0x1c); /* XT_REGION_SETUP */
+	nv_wo32(xtensa, 0xcc8, xtensa->gpu_fw->size >> 8); /* XT_REGION_LIMIT */
+
+	tmp = nv_rd32(xtensa, 0x0);
+	nv_wo32(xtensa, 0xde0, tmp); /* SCRATCH_H2X */
+
+	nv_wo32(xtensa, 0xce8, 0xf); /* XT_REGION_SETUP */
+
+	nv_wo32(xtensa, 0xc20, 0x3f); /* INTR */
+	nv_wo32(xtensa, 0xd84, 0x3f); /* INTR_EN */
+
+	return 0;
+}
+
+int
+_nouveau_xtensa_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_xtensa *xtensa = (void *)object;
+
+	nv_wo32(xtensa, 0xd84, 0); /* INTR_EN */
+	nv_wo32(xtensa, 0xd94, 0); /* FIFO_CTRL */
+
+	if (!suspend)
+		nouveau_gpuobj_ref(NULL, &xtensa->gpu_fw);
+
+	return nouveau_engine_fini(&xtensa->base, suspend);
+}
diff --git a/drivers/gpu/drm/nouveau/core/include/core/device.h b/drivers/gpu/drm/nouveau/core/include/core/device.h
index 05840f3..99b6600 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/device.h
@@ -17,8 +17,7 @@ enum nv_subdev_type {
 	NVDEV_SUBDEV_DEVINIT,
 	NVDEV_SUBDEV_GPIO,
 	NVDEV_SUBDEV_I2C,
-	NVDEV_SUBDEV_CLOCK,
-	NVDEV_SUBDEV_DEVINIT_LAST = NVDEV_SUBDEV_CLOCK,
+	NVDEV_SUBDEV_DEVINIT_LAST = NVDEV_SUBDEV_I2C,
 
 	/* This grouping of subdevs are initialised right after they've
 	 * been created, and are allowed to assume any subdevs in the
@@ -35,6 +34,7 @@ enum nv_subdev_type {
 	NVDEV_SUBDEV_VM,
 	NVDEV_SUBDEV_BAR,
 	NVDEV_SUBDEV_VOLT,
+	NVDEV_SUBDEV_CLOCK,
 	NVDEV_SUBDEV_THERM,
 
 	NVDEV_ENGINE_DMAOBJ,
@@ -49,6 +49,7 @@ enum nv_subdev_type {
 	NVDEV_ENGINE_PPP,
 	NVDEV_ENGINE_COPY0,
 	NVDEV_ENGINE_COPY1,
+	NVDEV_ENGINE_COPY2,
 	NVDEV_ENGINE_UNK1C1,
 	NVDEV_ENGINE_VENC,
 	NVDEV_ENGINE_DISP,
diff --git a/drivers/gpu/drm/nouveau/core/include/core/mm.h b/drivers/gpu/drm/nouveau/core/include/core/mm.h
index 2514e81..2bf7d0e 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/mm.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/mm.h
@@ -15,8 +15,6 @@ struct nouveau_mm {
 	struct list_head nodes;
 	struct list_head free;
 
-	struct mutex mutex;
-
 	u32 block_size;
 	int heap_nodes;
 };
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/bsp.h b/drivers/gpu/drm/nouveau/core/include/engine/bsp.h
index 13ccdf5..67662e2 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/bsp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/bsp.h
@@ -2,6 +2,7 @@
 #define __NOUVEAU_BSP_H__
 
 extern struct nouveau_oclass nv84_bsp_oclass;
+extern struct nouveau_oclass nv98_bsp_oclass;
 extern struct nouveau_oclass nvc0_bsp_oclass;
 extern struct nouveau_oclass nve0_bsp_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/copy.h b/drivers/gpu/drm/nouveau/core/include/engine/copy.h
index 8cad2cf..316a28a 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/copy.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/copy.h
@@ -8,5 +8,6 @@ extern struct nouveau_oclass nvc0_copy0_oclass;
 extern struct nouveau_oclass nvc0_copy1_oclass;
 extern struct nouveau_oclass nve0_copy0_oclass;
 extern struct nouveau_oclass nve0_copy1_oclass;
+extern struct nouveau_oclass nve0_copy2_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/falcon.h b/drivers/gpu/drm/nouveau/core/include/engine/falcon.h
index 1edec38..1edec38 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/falcon.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/falcon.h
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/graph.h b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
index 5d39243..8e1b523 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/graph.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
@@ -61,8 +61,14 @@ extern struct nouveau_oclass nv34_graph_oclass;
 extern struct nouveau_oclass nv35_graph_oclass;
 extern struct nouveau_oclass nv40_graph_oclass;
 extern struct nouveau_oclass nv50_graph_oclass;
-extern struct nouveau_oclass nvc0_graph_oclass;
-extern struct nouveau_oclass nve0_graph_oclass;
+extern struct nouveau_oclass *nvc0_graph_oclass;
+extern struct nouveau_oclass *nvc1_graph_oclass;
+extern struct nouveau_oclass *nvc3_graph_oclass;
+extern struct nouveau_oclass *nvc8_graph_oclass;
+extern struct nouveau_oclass *nvd7_graph_oclass;
+extern struct nouveau_oclass *nvd9_graph_oclass;
+extern struct nouveau_oclass *nve4_graph_oclass;
+extern struct nouveau_oclass *nvf0_graph_oclass;
 
 extern const struct nouveau_bitfield nv04_graph_nsource[];
 extern struct nouveau_ofuncs nv04_graph_ofuncs;
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h b/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h
index bbf0d4a..1d1a89a 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h
@@ -54,7 +54,6 @@ extern struct nouveau_ofuncs nv50_mpeg_ofuncs;
 int  nv50_mpeg_context_ctor(struct nouveau_object *, struct nouveau_object *,
 			    struct nouveau_oclass *, void *, u32,
 			    struct nouveau_object **);
-int  nv50_mpeg_tlb_flush(struct nouveau_engine *);
 void nv50_mpeg_intr(struct nouveau_subdev *);
 int  nv50_mpeg_init(struct nouveau_object *);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/vp.h b/drivers/gpu/drm/nouveau/core/include/engine/vp.h
index d7b287b..39baebe 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/vp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/vp.h
@@ -2,6 +2,7 @@
 #define __NOUVEAU_VP_H__
 
 extern struct nouveau_oclass nv84_vp_oclass;
+extern struct nouveau_oclass nv98_vp_oclass;
 extern struct nouveau_oclass nvc0_vp_oclass;
 extern struct nouveau_oclass nve0_vp_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/xtensa.h b/drivers/gpu/drm/nouveau/core/include/engine/xtensa.h
new file mode 100644
index 0000000..306100f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/engine/xtensa.h
@@ -0,0 +1,38 @@
+#ifndef __NOUVEAU_XTENSA_H__
+#define __NOUVEAU_XTENSA_H__
+
+#include <core/engine.h>
+#include <core/engctx.h>
+#include <core/gpuobj.h>
+
+struct nouveau_xtensa {
+	struct nouveau_engine base;
+
+	u32 addr;
+	struct nouveau_gpuobj *gpu_fw;
+	u32 fifo_val;
+	u32 unkd28;
+};
+
+#define nouveau_xtensa_create(p,e,c,b,d,i,f,r)				\
+	nouveau_xtensa_create_((p), (e), (c), (b), (d), (i), (f),	\
+			       sizeof(**r),(void **)r)
+
+int _nouveau_xtensa_engctx_ctor(struct nouveau_object *,
+				struct nouveau_object *,
+				struct nouveau_oclass *, void *, u32,
+				struct nouveau_object **);
+
+void _nouveau_xtensa_intr(struct nouveau_subdev *);
+int nouveau_xtensa_create_(struct nouveau_object *,
+			   struct nouveau_object *,
+			   struct nouveau_oclass *, u32, bool,
+			   const char *, const char *,
+			   int, void **);
+#define _nouveau_xtensa_dtor _nouveau_engine_dtor
+int _nouveau_xtensa_init(struct nouveau_object *);
+int _nouveau_xtensa_fini(struct nouveau_object *, bool);
+u32  _nouveau_xtensa_rd32(struct nouveau_object *, u64);
+void _nouveau_xtensa_wr32(struct nouveau_object *, u64, u32);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
index 41b7a6a..89ee289 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
@@ -10,8 +10,6 @@ struct nvbios_pll;
 struct nouveau_clock {
 	struct nouveau_subdev base;
 
-	int (*pll_set)(struct nouveau_clock *, u32 type, u32 freq);
-
 	/*XXX: die, these are here *only* to support the completely
 	 *     bat-shit insane what-was-nouveau_hw.c code
 	 */
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h b/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
index 29e4cc1..685c9b1 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
@@ -8,6 +8,8 @@ struct nouveau_devinit {
 	struct nouveau_subdev base;
 	bool post;
 	void (*meminit)(struct nouveau_devinit *);
+	int  (*pll_set)(struct nouveau_devinit *, u32 type, u32 freq);
+
 };
 
 static inline struct nouveau_devinit *
@@ -20,11 +22,20 @@ nouveau_devinit(void *obj)
 	nouveau_devinit_create_((p), (e), (o), sizeof(**d), (void **)d)
 #define nouveau_devinit_destroy(p)                                             \
 	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_devinit_init(p) ({                                             \
+	struct nouveau_devinit *d = (p);                                       \
+	_nouveau_devinit_init(nv_object(d));                                   \
+})
+#define nouveau_devinit_fini(p,s) ({                                           \
+	struct nouveau_devinit *d = (p);                                       \
+	_nouveau_devinit_fini(nv_object(d), (s));                              \
+})
 
 int nouveau_devinit_create_(struct nouveau_object *, struct nouveau_object *,
 			    struct nouveau_oclass *, int, void **);
-int nouveau_devinit_init(struct nouveau_devinit *);
-int nouveau_devinit_fini(struct nouveau_devinit *, bool suspend);
+#define _nouveau_devinit_dtor _nouveau_subdev_dtor
+int _nouveau_devinit_init(struct nouveau_object *);
+int _nouveau_devinit_fini(struct nouveau_object *, bool suspend);
 
 extern struct nouveau_oclass nv04_devinit_oclass;
 extern struct nouveau_oclass nv05_devinit_oclass;
@@ -32,9 +43,7 @@ extern struct nouveau_oclass nv10_devinit_oclass;
 extern struct nouveau_oclass nv1a_devinit_oclass;
 extern struct nouveau_oclass nv20_devinit_oclass;
 extern struct nouveau_oclass nv50_devinit_oclass;
-
-void nv04_devinit_dtor(struct nouveau_object *);
-int  nv04_devinit_init(struct nouveau_object *);
-int  nv04_devinit_fini(struct nouveau_object *, bool);
+extern struct nouveau_oclass nva3_devinit_oclass;
+extern struct nouveau_oclass nvc0_devinit_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
index da470e6..2e74050 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
@@ -53,31 +53,7 @@ struct nouveau_fb {
 
 	bool (*memtype_valid)(struct nouveau_fb *, u32 memtype);
 
-	struct {
-		enum {
-			NV_MEM_TYPE_UNKNOWN = 0,
-			NV_MEM_TYPE_STOLEN,
-			NV_MEM_TYPE_SGRAM,
-			NV_MEM_TYPE_SDRAM,
-			NV_MEM_TYPE_DDR1,
-			NV_MEM_TYPE_DDR2,
-			NV_MEM_TYPE_DDR3,
-			NV_MEM_TYPE_GDDR2,
-			NV_MEM_TYPE_GDDR3,
-			NV_MEM_TYPE_GDDR4,
-			NV_MEM_TYPE_GDDR5
-		} type;
-		u64 stolen;
-		u64 size;
-
-		int ranks;
-		int parts;
-
-		int  (*init)(struct nouveau_fb *);
-		int  (*get)(struct nouveau_fb *, u64 size, u32 align,
-			    u32 size_nc, u32 type, struct nouveau_mem **);
-		void (*put)(struct nouveau_fb *, struct nouveau_mem **);
-	} ram;
+	struct nouveau_ram *ram;
 
 	struct nouveau_mm vram;
 	struct nouveau_mm tags;
@@ -102,18 +78,6 @@ nouveau_fb(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_FB];
 }
 
-#define nouveau_fb_create(p,e,c,d)                                             \
-	nouveau_subdev_create((p), (e), (c), 0, "PFB", "fb", (d))
-int  nouveau_fb_preinit(struct nouveau_fb *);
-void nouveau_fb_destroy(struct nouveau_fb *);
-int  nouveau_fb_init(struct nouveau_fb *);
-#define nouveau_fb_fini(p,s)                                                   \
-	nouveau_subdev_fini(&(p)->base, (s))
-
-void _nouveau_fb_dtor(struct nouveau_object *);
-int  _nouveau_fb_init(struct nouveau_object *);
-#define _nouveau_fb_fini _nouveau_subdev_fini
-
 extern struct nouveau_oclass nv04_fb_oclass;
 extern struct nouveau_oclass nv10_fb_oclass;
 extern struct nouveau_oclass nv1a_fb_oclass;
@@ -132,40 +96,31 @@ extern struct nouveau_oclass nv4e_fb_oclass;
 extern struct nouveau_oclass nv50_fb_oclass;
 extern struct nouveau_oclass nvc0_fb_oclass;
 
-struct nouveau_bios;
-int  nouveau_fb_bios_memtype(struct nouveau_bios *);
-
-bool nv04_fb_memtype_valid(struct nouveau_fb *, u32 memtype);
-
-void nv10_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
-void nv10_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
-void nv10_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
-
-int  nv20_fb_vram_init(struct nouveau_fb *);
-void nv20_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
-void nv20_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
-void nv20_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
-
-int  nv30_fb_init(struct nouveau_object *);
-void nv30_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
-
-void nv40_fb_tile_comp(struct nouveau_fb *, int i, u32 size, u32 flags,
-		       struct nouveau_fb_tile *);
-
-int  nv41_fb_vram_init(struct nouveau_fb *);
-int  nv41_fb_init(struct nouveau_object *);
-void nv41_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
-
-int  nv44_fb_vram_init(struct nouveau_fb *);
-int  nv44_fb_init(struct nouveau_object *);
-void nv44_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+struct nouveau_ram {
+	struct nouveau_object base;
+	enum {
+		NV_MEM_TYPE_UNKNOWN = 0,
+		NV_MEM_TYPE_STOLEN,
+		NV_MEM_TYPE_SGRAM,
+		NV_MEM_TYPE_SDRAM,
+		NV_MEM_TYPE_DDR1,
+		NV_MEM_TYPE_DDR2,
+		NV_MEM_TYPE_DDR3,
+		NV_MEM_TYPE_GDDR2,
+		NV_MEM_TYPE_GDDR3,
+		NV_MEM_TYPE_GDDR4,
+		NV_MEM_TYPE_GDDR5
+	} type;
+	u64 stolen;
+	u64 size;
+	u32 tags;
 
-void nv46_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+	int ranks;
+	int parts;
 
-void nv50_fb_vram_del(struct nouveau_fb *, struct nouveau_mem **);
+	int  (*get)(struct nouveau_fb *, u64 size, u32 align,
+		    u32 size_nc, u32 type, struct nouveau_mem **);
+	void (*put)(struct nouveau_fb *, struct nouveau_mem **);
+};
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/vm.h b/drivers/gpu/drm/nouveau/core/include/subdev/vm.h
index 9d595ef..f2e87b1 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/vm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/vm.h
@@ -58,7 +58,7 @@ struct nouveau_vm {
 	int refcount;
 
 	struct list_head pgd_list;
-	atomic_t engref[64]; //NVDEV_SUBDEV_NR];
+	atomic_t engref[NVDEV_SUBDEV_NR];
 
 	struct nouveau_vm_pgt *pgt;
 	u32 fpde;
@@ -117,9 +117,6 @@ int  nv04_vm_create(struct nouveau_vmmgr *, u64, u64, u64,
 		    struct nouveau_vm **);
 void nv04_vmmgr_dtor(struct nouveau_object *);
 
-void nv50_vm_flush_engine(struct nouveau_subdev *, int engine);
-void nvc0_vm_flush_engine(struct nouveau_subdev *, u64 addr, int type);
-
 /* nouveau_vm.c */
 int  nouveau_vm_create(struct nouveau_vmmgr *, u64 offset, u64 length,
 		       u64 mm_offset, u32 block, struct nouveau_vm **);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
index 649f1ce..160d27f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
@@ -53,7 +53,6 @@ nv50_bar_kmap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nv50_vm_flush_engine(nv_subdev(bar), 6);
 	return 0;
 }
 
@@ -69,7 +68,6 @@ nv50_bar_umap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nv50_vm_flush_engine(nv_subdev(bar), 6);
 	return 0;
 }
 
@@ -77,7 +75,6 @@ static void
 nv50_bar_unmap(struct nouveau_bar *bar, struct nouveau_vma *vma)
 {
 	nouveau_vm_unmap(vma);
-	nv50_vm_flush_engine(nv_subdev(bar), 6);
 	nouveau_vm_put(vma);
 }
 
@@ -147,6 +144,8 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_gpuobj_new(nv_object(priv), heap,
 				 ((limit-- - start) >> 12) * 8, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &vm->pgt[0].obj[0]);
@@ -179,6 +178,8 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_vm_ref(vm, &priv->bar1_vm, priv->pgd);
 	nouveau_vm_ref(NULL, &vm, NULL);
 	if (ret)
@@ -237,7 +238,11 @@ nv50_bar_init(struct nouveau_object *object)
 
 	nv_mask(priv, 0x000200, 0x00000100, 0x00000000);
 	nv_mask(priv, 0x000200, 0x00000100, 0x00000100);
-	nv50_vm_flush_engine(nv_subdev(priv), 6);
+	nv_wr32(priv, 0x100c80, 0x00060001);
+	if (!nv_wait(priv, 0x100c80, 0x00000001, 0x00000000)) {
+		nv_error(priv, "vm flush timeout\n");
+		return -EBUSY;
+	}
 
 	nv_wr32(priv, 0x001704, 0x00000000 | priv->mem->addr >> 12);
 	nv_wr32(priv, 0x001704, 0x40000000 | priv->mem->addr >> 12);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
index f8a44956..b2ec741 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
@@ -51,7 +51,6 @@ nvc0_bar_kmap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nvc0_vm_flush_engine(nv_subdev(bar), priv->bar[0].pgd->addr, 5);
 	return 0;
 }
 
@@ -68,18 +67,13 @@ nvc0_bar_umap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nvc0_vm_flush_engine(nv_subdev(bar), priv->bar[1].pgd->addr, 5);
 	return 0;
 }
 
 static void
 nvc0_bar_unmap(struct nouveau_bar *bar, struct nouveau_vma *vma)
 {
-	struct nvc0_bar_priv *priv = (void *)bar;
-	int i = !(vma->vm == priv->bar[0].vm);
-
 	nouveau_vm_unmap(vma);
-	nvc0_vm_flush_engine(nv_subdev(bar), priv->bar[i].pgd->addr, 5);
 	nouveau_vm_put(vma);
 }
 
@@ -116,6 +110,8 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 				 (pci_resource_len(pdev, 3) >> 12) * 8,
 				 0x1000, NVOBJ_FLAG_ZERO_ALLOC,
@@ -150,6 +146,8 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_vm_ref(vm, &priv->bar[1].vm, priv->bar[1].pgd);
 	nouveau_vm_ref(NULL, &vm, NULL);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
index 0e2c1a4..aa0fbbe 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
@@ -85,11 +85,15 @@ static void
 nouveau_bios_shadow_pramin(struct nouveau_bios *bios)
 {
 	struct nouveau_device *device = nv_device(bios);
+	u64 addr = 0;
 	u32 bar0 = 0;
 	int i;
 
 	if (device->card_type >= NV_50) {
-		u64 addr = (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8;
+		if (  device->card_type < NV_C0 ||
+		    !(nv_rd32(bios, 0x022500) & 0x00000001))
+			addr = (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8;
+
 		if (!addr) {
 			addr  = (u64)nv_rd32(bios, 0x001700) << 16;
 			addr += 0xf0000;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index c434d39..0687e64 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -10,7 +10,6 @@
 #include <subdev/bios/gpio.h>
 #include <subdev/bios/init.h>
 #include <subdev/devinit.h>
-#include <subdev/clock.h>
 #include <subdev/i2c.h>
 #include <subdev/vga.h>
 #include <subdev/gpio.h>
@@ -300,9 +299,9 @@ init_wrauxr(struct nvbios_init *init, u32 addr, u8 data)
 static void
 init_prog_pll(struct nvbios_init *init, u32 id, u32 freq)
 {
-	struct nouveau_clock *clk = nouveau_clock(init->bios);
-	if (clk && clk->pll_set && init_exec(init)) {
-		int ret = clk->pll_set(clk, id, freq);
+	struct nouveau_devinit *devinit = nouveau_devinit(init->bios);
+	if (devinit->pll_set && init_exec(init)) {
+		int ret = devinit->pll_set(devinit, id, freq);
 		if (ret)
 			warn("failed to prog pll 0x%08x to %dkHz\n", id, freq);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
index b7fd115..a142775 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
@@ -22,9 +22,10 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/clock.h>
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
+#include <subdev/clock.h>
+#include <subdev/devinit/priv.h>
 
 #include "pll.h"
 
@@ -32,272 +33,12 @@ struct nv04_clock_priv {
 	struct nouveau_clock base;
 };
 
-static int
-powerctrl_1_shift(int chip_version, int reg)
-{
-	int shift = -4;
-
-	if (chip_version < 0x17 || chip_version == 0x1a || chip_version == 0x20)
-		return shift;
-
-	switch (reg) {
-	case 0x680520:
-		shift += 4;
-	case 0x680508:
-		shift += 4;
-	case 0x680504:
-		shift += 4;
-	case 0x680500:
-		shift += 4;
-	}
-
-	/*
-	 * the shift for vpll regs is only used for nv3x chips with a single
-	 * stage pll
-	 */
-	if (shift > 4 && (chip_version < 0x32 || chip_version == 0x35 ||
-			  chip_version == 0x36 || chip_version >= 0x40))
-		shift = -4;
-
-	return shift;
-}
-
-static void
-setPLL_single(struct nv04_clock_priv *priv, u32 reg,
-	      struct nouveau_pll_vals *pv)
-{
-	int chip_version = nouveau_bios(priv)->version.chip;
-	uint32_t oldpll = nv_rd32(priv, reg);
-	int oldN = (oldpll >> 8) & 0xff, oldM = oldpll & 0xff;
-	uint32_t pll = (oldpll & 0xfff80000) | pv->log2P << 16 | pv->NM1;
-	uint32_t saved_powerctrl_1 = 0;
-	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg);
-
-	if (oldpll == pll)
-		return;	/* already set */
-
-	if (shift_powerctrl_1 >= 0) {
-		saved_powerctrl_1 = nv_rd32(priv, 0x001584);
-		nv_wr32(priv, 0x001584,
-			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
-			1 << shift_powerctrl_1);
-	}
-
-	if (oldM && pv->M1 && (oldN / oldM < pv->N1 / pv->M1))
-		/* upclock -- write new post divider first */
-		nv_wr32(priv, reg, pv->log2P << 16 | (oldpll & 0xffff));
-	else
-		/* downclock -- write new NM first */
-		nv_wr32(priv, reg, (oldpll & 0xffff0000) | pv->NM1);
-
-	if (chip_version < 0x17 && chip_version != 0x11)
-		/* wait a bit on older chips */
-		msleep(64);
-	nv_rd32(priv, reg);
-
-	/* then write the other half as well */
-	nv_wr32(priv, reg, pll);
-
-	if (shift_powerctrl_1 >= 0)
-		nv_wr32(priv, 0x001584, saved_powerctrl_1);
-}
-
-static uint32_t
-new_ramdac580(uint32_t reg1, bool ss, uint32_t ramdac580)
-{
-	bool head_a = (reg1 == 0x680508);
-
-	if (ss)	/* single stage pll mode */
-		ramdac580 |= head_a ? 0x00000100 : 0x10000000;
-	else
-		ramdac580 &= head_a ? 0xfffffeff : 0xefffffff;
-
-	return ramdac580;
-}
-
-static void
-setPLL_double_highregs(struct nv04_clock_priv *priv, u32 reg1,
-		       struct nouveau_pll_vals *pv)
-{
-	int chip_version = nouveau_bios(priv)->version.chip;
-	bool nv3035 = chip_version == 0x30 || chip_version == 0x35;
-	uint32_t reg2 = reg1 + ((reg1 == 0x680520) ? 0x5c : 0x70);
-	uint32_t oldpll1 = nv_rd32(priv, reg1);
-	uint32_t oldpll2 = !nv3035 ? nv_rd32(priv, reg2) : 0;
-	uint32_t pll1 = (oldpll1 & 0xfff80000) | pv->log2P << 16 | pv->NM1;
-	uint32_t pll2 = (oldpll2 & 0x7fff0000) | 1 << 31 | pv->NM2;
-	uint32_t oldramdac580 = 0, ramdac580 = 0;
-	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;	/* nv41+ only */
-	uint32_t saved_powerctrl_1 = 0, savedc040 = 0;
-	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg1);
-
-	/* model specific additions to generic pll1 and pll2 set up above */
-	if (nv3035) {
-		pll1 = (pll1 & 0xfcc7ffff) | (pv->N2 & 0x18) << 21 |
-		       (pv->N2 & 0x7) << 19 | 8 << 4 | (pv->M2 & 7) << 4;
-		pll2 = 0;
-	}
-	if (chip_version > 0x40 && reg1 >= 0x680508) { /* !nv40 */
-		oldramdac580 = nv_rd32(priv, 0x680580);
-		ramdac580 = new_ramdac580(reg1, single_stage, oldramdac580);
-		if (oldramdac580 != ramdac580)
-			oldpll1 = ~0;	/* force mismatch */
-		if (single_stage)
-			/* magic value used by nvidia in single stage mode */
-			pll2 |= 0x011f;
-	}
-	if (chip_version > 0x70)
-		/* magic bits set by the blob (but not the bios) on g71-73 */
-		pll1 = (pll1 & 0x7fffffff) | (single_stage ? 0x4 : 0xc) << 28;
-
-	if (oldpll1 == pll1 && oldpll2 == pll2)
-		return;	/* already set */
-
-	if (shift_powerctrl_1 >= 0) {
-		saved_powerctrl_1 = nv_rd32(priv, 0x001584);
-		nv_wr32(priv, 0x001584,
-			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
-			1 << shift_powerctrl_1);
-	}
-
-	if (chip_version >= 0x40) {
-		int shift_c040 = 14;
-
-		switch (reg1) {
-		case 0x680504:
-			shift_c040 += 2;
-		case 0x680500:
-			shift_c040 += 2;
-		case 0x680520:
-			shift_c040 += 2;
-		case 0x680508:
-			shift_c040 += 2;
-		}
-
-		savedc040 = nv_rd32(priv, 0xc040);
-		if (shift_c040 != 14)
-			nv_wr32(priv, 0xc040, savedc040 & ~(3 << shift_c040));
-	}
-
-	if (oldramdac580 != ramdac580)
-		nv_wr32(priv, 0x680580, ramdac580);
-
-	if (!nv3035)
-		nv_wr32(priv, reg2, pll2);
-	nv_wr32(priv, reg1, pll1);
-
-	if (shift_powerctrl_1 >= 0)
-		nv_wr32(priv, 0x001584, saved_powerctrl_1);
-	if (chip_version >= 0x40)
-		nv_wr32(priv, 0xc040, savedc040);
-}
-
-static void
-setPLL_double_lowregs(struct nv04_clock_priv *priv, u32 NMNMreg,
-		      struct nouveau_pll_vals *pv)
-{
-	/* When setting PLLs, there is a merry game of disabling and enabling
-	 * various bits of hardware during the process. This function is a
-	 * synthesis of six nv4x traces, nearly each card doing a subtly
-	 * different thing. With luck all the necessary bits for each card are
-	 * combined herein. Without luck it deviates from each card's formula
-	 * so as to not work on any :)
-	 */
-
-	uint32_t Preg = NMNMreg - 4;
-	bool mpll = Preg == 0x4020;
-	uint32_t oldPval = nv_rd32(priv, Preg);
-	uint32_t NMNM = pv->NM2 << 16 | pv->NM1;
-	uint32_t Pval = (oldPval & (mpll ? ~(0x77 << 16) : ~(7 << 16))) |
-			0xc << 28 | pv->log2P << 16;
-	uint32_t saved4600 = 0;
-	/* some cards have different maskc040s */
-	uint32_t maskc040 = ~(3 << 14), savedc040;
-	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;
-
-	if (nv_rd32(priv, NMNMreg) == NMNM && (oldPval & 0xc0070000) == Pval)
-		return;
-
-	if (Preg == 0x4000)
-		maskc040 = ~0x333;
-	if (Preg == 0x4058)
-		maskc040 = ~(0xc << 24);
-
-	if (mpll) {
-		struct nvbios_pll info;
-		uint8_t Pval2;
-
-		if (nvbios_pll_parse(nouveau_bios(priv), Preg, &info))
-			return;
-
-		Pval2 = pv->log2P + info.bias_p;
-		if (Pval2 > info.max_p)
-			Pval2 = info.max_p;
-		Pval |= 1 << 28 | Pval2 << 20;
-
-		saved4600 = nv_rd32(priv, 0x4600);
-		nv_wr32(priv, 0x4600, saved4600 | 8 << 28);
-	}
-	if (single_stage)
-		Pval |= mpll ? 1 << 12 : 1 << 8;
-
-	nv_wr32(priv, Preg, oldPval | 1 << 28);
-	nv_wr32(priv, Preg, Pval & ~(4 << 28));
-	if (mpll) {
-		Pval |= 8 << 20;
-		nv_wr32(priv, 0x4020, Pval & ~(0xc << 28));
-		nv_wr32(priv, 0x4038, Pval & ~(0xc << 28));
-	}
-
-	savedc040 = nv_rd32(priv, 0xc040);
-	nv_wr32(priv, 0xc040, savedc040 & maskc040);
-
-	nv_wr32(priv, NMNMreg, NMNM);
-	if (NMNMreg == 0x4024)
-		nv_wr32(priv, 0x403c, NMNM);
-
-	nv_wr32(priv, Preg, Pval);
-	if (mpll) {
-		Pval &= ~(8 << 20);
-		nv_wr32(priv, 0x4020, Pval);
-		nv_wr32(priv, 0x4038, Pval);
-		nv_wr32(priv, 0x4600, saved4600);
-	}
-
-	nv_wr32(priv, 0xc040, savedc040);
-
-	if (mpll) {
-		nv_wr32(priv, 0x4020, Pval & ~(1 << 28));
-		nv_wr32(priv, 0x4038, Pval & ~(1 << 28));
-	}
-}
-
-int
-nv04_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nv04_clock_priv *priv = (void *)clk;
-	struct nouveau_pll_vals pv;
-	struct nvbios_pll info;
-	int ret;
-
-	ret = nvbios_pll_parse(nouveau_bios(priv), type > 0x405c ?
-			       type : type - 4, &info);
-	if (ret)
-		return ret;
-
-	ret = clk->pll_calc(clk, &info, freq, &pv);
-	if (!ret)
-		return ret;
-
-	return clk->pll_prog(clk, type, &pv);
-}
-
 int
 nv04_clock_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		    int clk, struct nouveau_pll_vals *pv)
 {
 	int N1, M1, N2, M2, P;
-	int ret = nv04_pll_calc(clock, info, clk, &N1, &M1, &N2, &M2, &P);
+	int ret = nv04_pll_calc(nv_subdev(clock), info, clk, &N1, &M1, &N2, &M2, &P);
 	if (ret) {
 		pv->refclk = info->refclk;
 		pv->N1 = N1;
@@ -313,17 +54,17 @@ int
 nv04_clock_pll_prog(struct nouveau_clock *clk, u32 reg1,
 		    struct nouveau_pll_vals *pv)
 {
-	struct nv04_clock_priv *priv = (void *)clk;
+	struct nouveau_devinit *devinit = nouveau_devinit(clk);
 	int cv = nouveau_bios(clk)->version.chip;
 
 	if (cv == 0x30 || cv == 0x31 || cv == 0x35 || cv == 0x36 ||
 	    cv >= 0x40) {
 		if (reg1 > 0x405c)
-			setPLL_double_highregs(priv, reg1, pv);
+			setPLL_double_highregs(devinit, reg1, pv);
 		else
-			setPLL_double_lowregs(priv, reg1, pv);
+			setPLL_double_lowregs(devinit, reg1, pv);
 	} else
-		setPLL_single(priv, reg1, pv);
+		setPLL_single(devinit, reg1, pv);
 
 	return 0;
 }
@@ -341,7 +82,6 @@ nv04_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nv04_clock_pll_set;
 	priv->base.pll_calc = nv04_clock_pll_calc;
 	priv->base.pll_prog = nv04_clock_pll_prog;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
index a4b2b7e..0db5dbf 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
@@ -41,7 +41,6 @@ nv40_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nv04_clock_pll_set;
 	priv->base.pll_calc = nv04_clock_pll_calc;
 	priv->base.pll_prog = nv04_clock_pll_prog;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
index f4147f6..d09d3e7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
@@ -33,50 +33,6 @@ struct nv50_clock_priv {
 };
 
 static int
-nv50_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nv50_clock_priv *priv = (void *)clk;
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_pll info;
-	int N1, M1, N2, M2, P;
-	int ret;
-
-	ret = nvbios_pll_parse(bios, type, &info);
-	if (ret) {
-		nv_error(clk, "failed to retrieve pll data, %d\n", ret);
-		return ret;
-	}
-
-	ret = nv04_pll_calc(clk, &info, freq, &N1, &M1, &N2, &M2, &P);
-	if (!ret) {
-		nv_error(clk, "failed pll calculation\n");
-		return ret;
-	}
-
-	switch (info.type) {
-	case PLL_VPLL0:
-	case PLL_VPLL1:
-		nv_wr32(priv, info.reg + 0, 0x10000611);
-		nv_mask(priv, info.reg + 4, 0x00ff00ff, (M1 << 16) | N1);
-		nv_mask(priv, info.reg + 8, 0x7fff00ff, (P  << 28) |
-							(M2 << 16) | N2);
-		break;
-	case PLL_MEMORY:
-		nv_mask(priv, info.reg + 0, 0x01ff0000, (P << 22) |
-						        (info.bias_p << 19) |
-							(P << 16));
-		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
-		break;
-	default:
-		nv_mask(priv, info.reg + 0, 0x00070000, (P << 16));
-		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
-		break;
-	}
-
-	return 0;
-}
-
-static int
 nv50_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
 		struct nouveau_object **pobject)
@@ -89,7 +45,6 @@ nv50_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nv50_clock_pll_set;
 	priv->base.pll_calc = nv04_clock_pll_calc;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
index 9068c98..f074cd2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
@@ -32,47 +32,13 @@ struct nva3_clock_priv {
 	struct nouveau_clock base;
 };
 
-static int
-nva3_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nva3_clock_priv *priv = (void *)clk;
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_pll info;
-	int N, fN, M, P;
-	int ret;
-
-	ret = nvbios_pll_parse(bios, type, &info);
-	if (ret)
-		return ret;
-
-	ret = nva3_pll_calc(clk, &info, freq, &N, &fN, &M, &P);
-	if (ret < 0)
-		return ret;
-
-	switch (info.type) {
-	case PLL_VPLL0:
-	case PLL_VPLL1:
-		nv_wr32(priv, info.reg + 0, 0x50000610);
-		nv_mask(priv, info.reg + 4, 0x003fffff,
-					    (P << 16) | (M << 8) | N);
-		nv_wr32(priv, info.reg + 8, fN);
-		break;
-	default:
-		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
-		ret = -EINVAL;
-		break;
-	}
-
-	return ret;
-}
-
 int
 nva3_clock_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		    int clk, struct nouveau_pll_vals *pv)
 {
 	int ret, N, M, P;
 
-	ret = nva3_pll_calc(clock, info, clk, &N, NULL, &M, &P);
+	ret = nva3_pll_calc(nv_subdev(clock), info, clk, &N, NULL, &M, &P);
 
 	if (ret > 0) {
 		pv->refclk = info->refclk;
@@ -97,7 +63,6 @@ nva3_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nva3_clock_pll_set;
 	priv->base.pll_calc = nva3_clock_pll_calc;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
index 7c96262..439d81c2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
@@ -33,41 +33,6 @@ struct nvc0_clock_priv {
 };
 
 static int
-nvc0_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nvc0_clock_priv *priv = (void *)clk;
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_pll info;
-	int N, fN, M, P;
-	int ret;
-
-	ret = nvbios_pll_parse(bios, type, &info);
-	if (ret)
-		return ret;
-
-	ret = nva3_pll_calc(clk, &info, freq, &N, &fN, &M, &P);
-	if (ret < 0)
-		return ret;
-
-	switch (info.type) {
-	case PLL_VPLL0:
-	case PLL_VPLL1:
-	case PLL_VPLL2:
-	case PLL_VPLL3:
-		nv_mask(priv, info.reg + 0x0c, 0x00000000, 0x00000100);
-		nv_wr32(priv, info.reg + 0x04, (P << 16) | (N << 8) | M);
-		nv_wr32(priv, info.reg + 0x10, fN << 16);
-		break;
-	default:
-		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
-		ret = -EINVAL;
-		break;
-	}
-
-	return ret;
-}
-
-static int
 nvc0_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
 		struct nouveau_object **pobject)
@@ -80,7 +45,6 @@ nvc0_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nvc0_clock_pll_set;
 	priv->base.pll_calc = nva3_clock_pll_calc;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h b/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h
index ef2c007..445b14c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h
@@ -1,9 +1,9 @@
 #ifndef __NOUVEAU_PLL_H__
 #define __NOUVEAU_PLL_H__
 
-int nv04_pll_calc(struct nouveau_clock *, struct nvbios_pll *, u32 freq,
+int nv04_pll_calc(struct nouveau_subdev *, struct nvbios_pll *, u32 freq,
 		  int *N1, int *M1, int *N2, int *M2, int *P);
-int nva3_pll_calc(struct nouveau_clock *, struct nvbios_pll *, u32 freq,
+int nva3_pll_calc(struct nouveau_subdev *, struct nvbios_pll *, u32 freq,
 		  int *N, int *fN, int *M, int *P);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c
index a2ab6d0..cf1ed0d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c
@@ -21,14 +21,13 @@
  * SOFTWARE.
  */
 
-#include <subdev/clock.h>
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
 
 #include "pll.h"
 
 static int
-getMNP_single(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
+getMNP_single(struct nouveau_subdev *subdev, struct nvbios_pll *info, int clk,
 	      int *pN, int *pM, int *pP)
 {
 	/* Find M, N and P for a single stage PLL
@@ -39,7 +38,7 @@ getMNP_single(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 	 * "clk" parameter in kHz
 	 * returns calculated clock
 	 */
-	int cv = nouveau_bios(clock)->version.chip;
+	int cv = nouveau_bios(subdev)->version.chip;
 	int minvco = info->vco1.min_freq, maxvco = info->vco1.max_freq;
 	int minM = info->vco1.min_m, maxM = info->vco1.max_m;
 	int minN = info->vco1.min_n, maxN = info->vco1.max_n;
@@ -124,7 +123,7 @@ getMNP_single(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 }
 
 static int
-getMNP_double(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
+getMNP_double(struct nouveau_subdev *subdev, struct nvbios_pll *info, int clk,
 	      int *pN1, int *pM1, int *pN2, int *pM2, int *pP)
 {
 	/* Find M, N and P for a two stage PLL
@@ -135,7 +134,7 @@ getMNP_double(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 	 * "clk" parameter in kHz
 	 * returns calculated clock
 	 */
-	int chip_version = nouveau_bios(clock)->version.chip;
+	int chip_version = nouveau_bios(subdev)->version.chip;
 	int minvco1 = info->vco1.min_freq, maxvco1 = info->vco1.max_freq;
 	int minvco2 = info->vco2.min_freq, maxvco2 = info->vco2.max_freq;
 	int minU1 = info->vco1.min_inputfreq, minU2 = info->vco2.min_inputfreq;
@@ -223,20 +222,20 @@ getMNP_double(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 }
 
 int
-nv04_pll_calc(struct nouveau_clock *clk, struct nvbios_pll *info, u32 freq,
+nv04_pll_calc(struct nouveau_subdev *subdev, struct nvbios_pll *info, u32 freq,
 	      int *N1, int *M1, int *N2, int *M2, int *P)
 {
 	int ret;
 
 	if (!info->vco2.max_freq) {
-		ret = getMNP_single(clk, info, freq, N1, M1, P);
+		ret = getMNP_single(subdev, info, freq, N1, M1, P);
 		*N2 = 1;
 		*M2 = 1;
 	} else {
-		ret = getMNP_double(clk, info, freq, N1, M1, N2, M2, P);
+		ret = getMNP_double(subdev, info, freq, N1, M1, N2, M2, P);
 	}
 
 	if (!ret)
-		nv_error(clk, "unable to compute acceptable pll values\n");
+		nv_error(subdev, "unable to compute acceptable pll values\n");
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c
index eed5c16..2fe1f71 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c
@@ -29,7 +29,7 @@
 #include "pll.h"
 
 int
-nva3_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
+nva3_pll_calc(struct nouveau_subdev *subdev, struct nvbios_pll *info,
 	      u32 freq, int *pN, int *pfN, int *pM, int *P)
 {
 	u32 best_err = ~0, err;
@@ -50,8 +50,15 @@ nva3_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		u32 tmp = freq * *P * M;
 		N  = tmp / info->refclk;
 		fN = tmp % info->refclk;
-		if (!pfN && fN >= info->refclk / 2)
-			N++;
+
+		if (!pfN) {
+			if (fN >= info->refclk / 2)
+				N++;
+		} else {
+			if (fN <  info->refclk / 2)
+				N--;
+			fN = tmp - (N * info->refclk);
+		}
 
 		if (N < info->vco1.min_n)
 			continue;
@@ -66,13 +73,14 @@ nva3_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		}
 
 		if (pfN) {
-			*pfN = (((fN << 13) / info->refclk) - 4096) & 0xffff;
+			*pfN = ((fN << 13) + info->refclk / 2) / info->refclk;
+			*pfN = (*pfN - 4096) & 0xffff;
 			return freq;
 		}
 	}
 
 	if (unlikely(best_err == ~0)) {
-		nv_error(clock, "unable to find matching pll values\n");
+		nv_error(subdev, "unable to find matching pll values\n");
 		return -EINVAL;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
index 5a07a39..79c81d3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
@@ -29,18 +29,10 @@
 #include <subdev/bios/init.h>
 
 int
-nouveau_devinit_init(struct nouveau_devinit *devinit)
+_nouveau_devinit_fini(struct nouveau_object *object, bool suspend)
 {
-	int ret = nouveau_subdev_init(&devinit->base);
-	if (ret)
-		return ret;
+	struct nouveau_devinit *devinit = (void *)object;
 
-	return nvbios_init(&devinit->base, devinit->post);
-}
-
-int
-nouveau_devinit_fini(struct nouveau_devinit *devinit, bool suspend)
-{
 	/* force full reinit on resume */
 	if (suspend)
 		devinit->post = true;
@@ -49,6 +41,17 @@ nouveau_devinit_fini(struct nouveau_devinit *devinit, bool suspend)
 }
 
 int
+_nouveau_devinit_init(struct nouveau_object *object)
+{
+	struct nouveau_devinit *devinit = (void *)object;
+	int ret = nouveau_subdev_init(&devinit->base);
+	if (ret)
+		return ret;
+
+	return nvbios_init(&devinit->base, devinit->post);
+}
+
+int
 nouveau_devinit_create_(struct nouveau_object *parent,
 			struct nouveau_object *engine,
 			struct nouveau_oclass *oclass,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
index 7a72d93..b22357d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
@@ -24,10 +24,10 @@
  *
  */
 
-#include <subdev/devinit.h>
 #include <subdev/vga.h>
 
 #include "fbmem.h"
+#include "priv.h"
 
 struct nv04_devinit_priv {
 	struct nouveau_devinit base;
@@ -111,33 +111,298 @@ nv04_devinit_meminit(struct nouveau_devinit *devinit)
 }
 
 static int
-nv04_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		  struct nouveau_oclass *oclass, void *data, u32 size,
-		  struct nouveau_object **pobject)
+powerctrl_1_shift(int chip_version, int reg)
 {
-	struct nv04_devinit_priv *priv;
+	int shift = -4;
+
+	if (chip_version < 0x17 || chip_version == 0x1a || chip_version == 0x20)
+		return shift;
+
+	switch (reg) {
+	case 0x680520:
+		shift += 4;
+	case 0x680508:
+		shift += 4;
+	case 0x680504:
+		shift += 4;
+	case 0x680500:
+		shift += 4;
+	}
+
+	/*
+	 * the shift for vpll regs is only used for nv3x chips with a single
+	 * stage pll
+	 */
+	if (shift > 4 && (chip_version < 0x32 || chip_version == 0x35 ||
+			  chip_version == 0x36 || chip_version >= 0x40))
+		shift = -4;
+
+	return shift;
+}
+
+void
+setPLL_single(struct nouveau_devinit *devinit, u32 reg,
+	      struct nouveau_pll_vals *pv)
+{
+	int chip_version = nouveau_bios(devinit)->version.chip;
+	uint32_t oldpll = nv_rd32(devinit, reg);
+	int oldN = (oldpll >> 8) & 0xff, oldM = oldpll & 0xff;
+	uint32_t pll = (oldpll & 0xfff80000) | pv->log2P << 16 | pv->NM1;
+	uint32_t saved_powerctrl_1 = 0;
+	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg);
+
+	if (oldpll == pll)
+		return;	/* already set */
+
+	if (shift_powerctrl_1 >= 0) {
+		saved_powerctrl_1 = nv_rd32(devinit, 0x001584);
+		nv_wr32(devinit, 0x001584,
+			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
+			1 << shift_powerctrl_1);
+	}
+
+	if (oldM && pv->M1 && (oldN / oldM < pv->N1 / pv->M1))
+		/* upclock -- write new post divider first */
+		nv_wr32(devinit, reg, pv->log2P << 16 | (oldpll & 0xffff));
+	else
+		/* downclock -- write new NM first */
+		nv_wr32(devinit, reg, (oldpll & 0xffff0000) | pv->NM1);
+
+	if (chip_version < 0x17 && chip_version != 0x11)
+		/* wait a bit on older chips */
+		msleep(64);
+	nv_rd32(devinit, reg);
+
+	/* then write the other half as well */
+	nv_wr32(devinit, reg, pll);
+
+	if (shift_powerctrl_1 >= 0)
+		nv_wr32(devinit, 0x001584, saved_powerctrl_1);
+}
+
+static uint32_t
+new_ramdac580(uint32_t reg1, bool ss, uint32_t ramdac580)
+{
+	bool head_a = (reg1 == 0x680508);
+
+	if (ss)	/* single stage pll mode */
+		ramdac580 |= head_a ? 0x00000100 : 0x10000000;
+	else
+		ramdac580 &= head_a ? 0xfffffeff : 0xefffffff;
+
+	return ramdac580;
+}
+
+void
+setPLL_double_highregs(struct nouveau_devinit *devinit, u32 reg1,
+		       struct nouveau_pll_vals *pv)
+{
+	int chip_version = nouveau_bios(devinit)->version.chip;
+	bool nv3035 = chip_version == 0x30 || chip_version == 0x35;
+	uint32_t reg2 = reg1 + ((reg1 == 0x680520) ? 0x5c : 0x70);
+	uint32_t oldpll1 = nv_rd32(devinit, reg1);
+	uint32_t oldpll2 = !nv3035 ? nv_rd32(devinit, reg2) : 0;
+	uint32_t pll1 = (oldpll1 & 0xfff80000) | pv->log2P << 16 | pv->NM1;
+	uint32_t pll2 = (oldpll2 & 0x7fff0000) | 1 << 31 | pv->NM2;
+	uint32_t oldramdac580 = 0, ramdac580 = 0;
+	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;	/* nv41+ only */
+	uint32_t saved_powerctrl_1 = 0, savedc040 = 0;
+	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg1);
+
+	/* model specific additions to generic pll1 and pll2 set up above */
+	if (nv3035) {
+		pll1 = (pll1 & 0xfcc7ffff) | (pv->N2 & 0x18) << 21 |
+		       (pv->N2 & 0x7) << 19 | 8 << 4 | (pv->M2 & 7) << 4;
+		pll2 = 0;
+	}
+	if (chip_version > 0x40 && reg1 >= 0x680508) { /* !nv40 */
+		oldramdac580 = nv_rd32(devinit, 0x680580);
+		ramdac580 = new_ramdac580(reg1, single_stage, oldramdac580);
+		if (oldramdac580 != ramdac580)
+			oldpll1 = ~0;	/* force mismatch */
+		if (single_stage)
+			/* magic value used by nvidia in single stage mode */
+			pll2 |= 0x011f;
+	}
+	if (chip_version > 0x70)
+		/* magic bits set by the blob (but not the bios) on g71-73 */
+		pll1 = (pll1 & 0x7fffffff) | (single_stage ? 0x4 : 0xc) << 28;
+
+	if (oldpll1 == pll1 && oldpll2 == pll2)
+		return;	/* already set */
+
+	if (shift_powerctrl_1 >= 0) {
+		saved_powerctrl_1 = nv_rd32(devinit, 0x001584);
+		nv_wr32(devinit, 0x001584,
+			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
+			1 << shift_powerctrl_1);
+	}
+
+	if (chip_version >= 0x40) {
+		int shift_c040 = 14;
+
+		switch (reg1) {
+		case 0x680504:
+			shift_c040 += 2;
+		case 0x680500:
+			shift_c040 += 2;
+		case 0x680520:
+			shift_c040 += 2;
+		case 0x680508:
+			shift_c040 += 2;
+		}
+
+		savedc040 = nv_rd32(devinit, 0xc040);
+		if (shift_c040 != 14)
+			nv_wr32(devinit, 0xc040, savedc040 & ~(3 << shift_c040));
+	}
+
+	if (oldramdac580 != ramdac580)
+		nv_wr32(devinit, 0x680580, ramdac580);
+
+	if (!nv3035)
+		nv_wr32(devinit, reg2, pll2);
+	nv_wr32(devinit, reg1, pll1);
+
+	if (shift_powerctrl_1 >= 0)
+		nv_wr32(devinit, 0x001584, saved_powerctrl_1);
+	if (chip_version >= 0x40)
+		nv_wr32(devinit, 0xc040, savedc040);
+}
+
+void
+setPLL_double_lowregs(struct nouveau_devinit *devinit, u32 NMNMreg,
+		      struct nouveau_pll_vals *pv)
+{
+	/* When setting PLLs, there is a merry game of disabling and enabling
+	 * various bits of hardware during the process. This function is a
+	 * synthesis of six nv4x traces, nearly each card doing a subtly
+	 * different thing. With luck all the necessary bits for each card are
+	 * combined herein. Without luck it deviates from each card's formula
+	 * so as to not work on any :)
+	 */
+
+	uint32_t Preg = NMNMreg - 4;
+	bool mpll = Preg == 0x4020;
+	uint32_t oldPval = nv_rd32(devinit, Preg);
+	uint32_t NMNM = pv->NM2 << 16 | pv->NM1;
+	uint32_t Pval = (oldPval & (mpll ? ~(0x77 << 16) : ~(7 << 16))) |
+			0xc << 28 | pv->log2P << 16;
+	uint32_t saved4600 = 0;
+	/* some cards have different maskc040s */
+	uint32_t maskc040 = ~(3 << 14), savedc040;
+	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;
+
+	if (nv_rd32(devinit, NMNMreg) == NMNM && (oldPval & 0xc0070000) == Pval)
+		return;
+
+	if (Preg == 0x4000)
+		maskc040 = ~0x333;
+	if (Preg == 0x4058)
+		maskc040 = ~(0xc << 24);
+
+	if (mpll) {
+		struct nvbios_pll info;
+		uint8_t Pval2;
+
+		if (nvbios_pll_parse(nouveau_bios(devinit), Preg, &info))
+			return;
+
+		Pval2 = pv->log2P + info.bias_p;
+		if (Pval2 > info.max_p)
+			Pval2 = info.max_p;
+		Pval |= 1 << 28 | Pval2 << 20;
+
+		saved4600 = nv_rd32(devinit, 0x4600);
+		nv_wr32(devinit, 0x4600, saved4600 | 8 << 28);
+	}
+	if (single_stage)
+		Pval |= mpll ? 1 << 12 : 1 << 8;
+
+	nv_wr32(devinit, Preg, oldPval | 1 << 28);
+	nv_wr32(devinit, Preg, Pval & ~(4 << 28));
+	if (mpll) {
+		Pval |= 8 << 20;
+		nv_wr32(devinit, 0x4020, Pval & ~(0xc << 28));
+		nv_wr32(devinit, 0x4038, Pval & ~(0xc << 28));
+	}
+
+	savedc040 = nv_rd32(devinit, 0xc040);
+	nv_wr32(devinit, 0xc040, savedc040 & maskc040);
+
+	nv_wr32(devinit, NMNMreg, NMNM);
+	if (NMNMreg == 0x4024)
+		nv_wr32(devinit, 0x403c, NMNM);
+
+	nv_wr32(devinit, Preg, Pval);
+	if (mpll) {
+		Pval &= ~(8 << 20);
+		nv_wr32(devinit, 0x4020, Pval);
+		nv_wr32(devinit, 0x4038, Pval);
+		nv_wr32(devinit, 0x4600, saved4600);
+	}
+
+	nv_wr32(devinit, 0xc040, savedc040);
+
+	if (mpll) {
+		nv_wr32(devinit, 0x4020, Pval & ~(1 << 28));
+		nv_wr32(devinit, 0x4038, Pval & ~(1 << 28));
+	}
+}
+
+int
+nv04_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
+{
+	struct nouveau_bios *bios = nouveau_bios(devinit);
+	struct nouveau_pll_vals pv;
+	struct nvbios_pll info;
+	int cv = bios->version.chip;
+	int N1, M1, N2, M2, P;
 	int ret;
 
-	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
+	ret = nvbios_pll_parse(bios, type > 0x405c ? type : type - 4, &info);
 	if (ret)
 		return ret;
 
-	priv->base.meminit = nv04_devinit_meminit;
-	priv->owner = -1;
+	ret = nv04_pll_calc(nv_subdev(devinit), &info, freq,
+			   &N1, &M1, &N2, &M2, &P);
+	if (!ret)
+		return -EINVAL;
+
+	pv.refclk = info.refclk;
+	pv.N1 = N1;
+	pv.M1 = M1;
+	pv.N2 = N2;
+	pv.M2 = M2;
+	pv.log2P = P;
+
+	if (cv == 0x30 || cv == 0x31 || cv == 0x35 || cv == 0x36 ||
+	    cv >= 0x40) {
+		if (type > 0x405c)
+			setPLL_double_highregs(devinit, type, &pv);
+		else
+			setPLL_double_lowregs(devinit, type, &pv);
+	} else
+		setPLL_single(devinit, type, &pv);
+
 	return 0;
 }
 
-void
-nv04_devinit_dtor(struct nouveau_object *object)
+int
+nv04_devinit_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv04_devinit_priv *priv = (void *)object;
 
-	/* restore vga owner saved at first init, and lock crtc regs  */
-	nv_wrvgaowner(priv, priv->owner);
-	nv_lockvgac(priv, true);
+	/* make i2c busses accessible */
+	nv_mask(priv, 0x000200, 0x00000001, 0x00000001);
 
-	nouveau_devinit_destroy(&priv->base);
+	/* unlock extended vga crtc regs, and unslave crtcs */
+	nv_lockvgac(priv, false);
+	if (priv->owner < 0)
+		priv->owner = nv_rdvgaowner(priv);
+	nv_wrvgaowner(priv, 0);
+
+	return nouveau_devinit_fini(&priv->base, suspend);
 }
 
 int
@@ -160,21 +425,35 @@ nv04_devinit_init(struct nouveau_object *object)
 	return nouveau_devinit_init(&priv->base);
 }
 
-int
-nv04_devinit_fini(struct nouveau_object *object, bool suspend)
+void
+nv04_devinit_dtor(struct nouveau_object *object)
 {
 	struct nv04_devinit_priv *priv = (void *)object;
 
-	/* make i2c busses accessible */
-	nv_mask(priv, 0x000200, 0x00000001, 0x00000001);
+	/* restore vga owner saved at first init, and lock crtc regs  */
+	nv_wrvgaowner(priv, priv->owner);
+	nv_lockvgac(priv, true);
 
-	/* unlock extended vga crtc regs, and unslave crtcs */
-	nv_lockvgac(priv, false);
-	if (priv->owner < 0)
-		priv->owner = nv_rdvgaowner(priv);
-	nv_wrvgaowner(priv, 0);
+	nouveau_devinit_destroy(&priv->base);
+}
 
-	return nouveau_devinit_fini(&priv->base, suspend);
+static int
+nv04_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nv04_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.meminit = nv04_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
+	priv->owner = -1;
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
index 191447d..b1912a8a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
@@ -24,12 +24,12 @@
  *
  */
 
-#include <subdev/devinit.h>
 #include <subdev/bios.h>
 #include <subdev/bios/bmp.h>
 #include <subdev/vga.h>
 
 #include "fbmem.h"
+#include "priv.h"
 
 struct nv05_devinit_priv {
 	struct nouveau_devinit base;
@@ -144,6 +144,7 @@ nv05_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	priv->base.meminit = nv05_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
index eb76ffa..463b08f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
@@ -24,10 +24,10 @@
  *
  */
 
-#include <subdev/devinit.h>
 #include <subdev/vga.h>
 
 #include "fbmem.h"
+#include "priv.h"
 
 struct nv10_devinit_priv {
 	struct nouveau_devinit base;
@@ -109,6 +109,7 @@ nv10_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	priv->base.meminit = nv10_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
index 5b2ba63..e9743cd 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
@@ -22,8 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/devinit.h>
-#include <subdev/vga.h>
+#include "priv.h"
 
 struct nv1a_devinit_priv {
 	struct nouveau_devinit base;
@@ -43,6 +42,7 @@ nv1a_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
index eb32e99..6cc6080 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
@@ -24,9 +24,7 @@
  *
  */
 
-#include <subdev/devinit.h>
-#include <subdev/vga.h>
-
+#include "priv.h"
 #include "fbmem.h"
 
 struct nv20_devinit_priv {
@@ -81,6 +79,7 @@ nv20_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	priv->base.meminit = nv20_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
index 4a85778..6df7224 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012 Red Hat Inc.
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -26,37 +26,55 @@
 #include <subdev/bios/dcb.h>
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
-#include <subdev/devinit.h>
 #include <subdev/vga.h>
 
-struct nv50_devinit_priv {
-	struct nouveau_devinit base;
-};
+#include "priv.h"
 
 static int
-nv50_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		  struct nouveau_oclass *oclass, void *data, u32 size,
-		  struct nouveau_object **pobject)
+nv50_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
 {
-	struct nv50_devinit_priv *priv;
+	struct nv50_devinit_priv *priv = (void *)devinit;
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pll info;
+	int N1, M1, N2, M2, P;
 	int ret;
 
-	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
+	ret = nvbios_pll_parse(bios, type, &info);
+	if (ret) {
+		nv_error(devinit, "failed to retrieve pll data, %d\n", ret);
 		return ret;
+	}
 
-	return 0;
-}
+	ret = nv04_pll_calc(nv_subdev(devinit), &info, freq, &N1, &M1, &N2, &M2, &P);
+	if (!ret) {
+		nv_error(devinit, "failed pll calculation\n");
+		return ret;
+	}
 
-static void
-nv50_devinit_dtor(struct nouveau_object *object)
-{
-	struct nv50_devinit_priv *priv = (void *)object;
-	nouveau_devinit_destroy(&priv->base);
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+		nv_wr32(priv, info.reg + 0, 0x10000611);
+		nv_mask(priv, info.reg + 4, 0x00ff00ff, (M1 << 16) | N1);
+		nv_mask(priv, info.reg + 8, 0x7fff00ff, (P  << 28) |
+							(M2 << 16) | N2);
+		break;
+	case PLL_MEMORY:
+		nv_mask(priv, info.reg + 0, 0x01ff0000, (P << 22) |
+						        (info.bias_p << 19) |
+							(P << 16));
+		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
+		break;
+	default:
+		nv_mask(priv, info.reg + 0, 0x00070000, (P << 16));
+		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
+		break;
+	}
+
+	return 0;
 }
 
-static int
+int
 nv50_devinit_init(struct nouveau_object *object)
 {
 	struct nouveau_bios *bios = nouveau_bios(object);
@@ -103,10 +121,20 @@ nv50_devinit_init(struct nouveau_object *object)
 }
 
 static int
-nv50_devinit_fini(struct nouveau_object *object, bool suspend)
+nv50_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
 {
-	struct nv50_devinit_priv *priv = (void *)object;
-	return nouveau_devinit_fini(&priv->base, suspend);
+	struct nv50_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.pll_set = nv50_devinit_pll_set;
+	return 0;
 }
 
 struct nouveau_oclass
@@ -114,8 +142,8 @@ nv50_devinit_oclass = {
 	.handle = NV_SUBDEV(DEVINIT, 0x50),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv50_devinit_ctor,
-		.dtor = nv50_devinit_dtor,
+		.dtor = _nouveau_devinit_dtor,
 		.init = nv50_devinit_init,
-		.fini = nv50_devinit_fini,
+		.fini = _nouveau_devinit_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
new file mode 100644
index 0000000..76a68b2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
@@ -0,0 +1,87 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nva3_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
+{
+	struct nva3_devinit_priv *priv = (void *)devinit;
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pll info;
+	int N, fN, M, P;
+	int ret;
+
+	ret = nvbios_pll_parse(bios, type, &info);
+	if (ret)
+		return ret;
+
+	ret = nva3_pll_calc(nv_subdev(devinit), &info, freq, &N, &fN, &M, &P);
+	if (ret < 0)
+		return ret;
+
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+		nv_wr32(priv, info.reg + 0, 0x50000610);
+		nv_mask(priv, info.reg + 4, 0x003fffff,
+					    (P << 16) | (M << 8) | N);
+		nv_wr32(priv, info.reg + 8, fN);
+		break;
+	default:
+		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int
+nva3_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nv50_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.pll_set = nva3_devinit_pll_set;
+	return 0;
+}
+
+struct nouveau_oclass
+nva3_devinit_oclass = {
+	.handle = NV_SUBDEV(DEVINIT, 0xa3),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nva3_devinit_ctor,
+		.dtor = _nouveau_devinit_dtor,
+		.init = nv50_devinit_init,
+		.fini = _nouveau_devinit_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
new file mode 100644
index 0000000..19e265b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
@@ -0,0 +1,90 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nvc0_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
+{
+	struct nvc0_devinit_priv *priv = (void *)devinit;
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pll info;
+	int N, fN, M, P;
+	int ret;
+
+	ret = nvbios_pll_parse(bios, type, &info);
+	if (ret)
+		return ret;
+
+	ret = nva3_pll_calc(nv_subdev(devinit), &info, freq, &N, &fN, &M, &P);
+	if (ret < 0)
+		return ret;
+
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+	case PLL_VPLL2:
+	case PLL_VPLL3:
+		nv_mask(priv, info.reg + 0x0c, 0x00000000, 0x00000100);
+		nv_wr32(priv, info.reg + 0x04, (P << 16) | (N << 8) | M);
+		nv_wr32(priv, info.reg + 0x10, fN << 16);
+		break;
+	default:
+		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int
+nvc0_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nv50_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.pll_set = nvc0_devinit_pll_set;
+	if (nv_rd32(priv, 0x022500) & 0x00000001)
+		priv->base.post = true;
+	return 0;
+}
+
+struct nouveau_oclass
+nvc0_devinit_oclass = {
+	.handle = NV_SUBDEV(DEVINIT, 0xc0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_devinit_ctor,
+		.dtor = _nouveau_devinit_dtor,
+		.init = nv50_devinit_init,
+		.fini = _nouveau_devinit_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h b/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
new file mode 100644
index 0000000..7d622e2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
@@ -0,0 +1,25 @@
+#ifndef __NVKM_DEVINIT_PRIV_H__
+#define __NVKM_DEVINIT_PRIV_H__
+
+#include <subdev/bios.h>
+#include <subdev/bios/pll.h>
+#include <subdev/clock/pll.h>
+#include <subdev/devinit.h>
+
+void nv04_devinit_dtor(struct nouveau_object *);
+int  nv04_devinit_init(struct nouveau_object *);
+int  nv04_devinit_fini(struct nouveau_object *, bool);
+int  nv04_devinit_pll_set(struct nouveau_devinit *, u32, u32);
+
+void setPLL_single(struct nouveau_devinit *, u32, struct nouveau_pll_vals *);
+void setPLL_double_highregs(struct nouveau_devinit *, u32, struct nouveau_pll_vals *);
+void setPLL_double_lowregs(struct nouveau_devinit *, u32, struct nouveau_pll_vals *);
+
+
+struct nv50_devinit_priv {
+	struct nouveau_devinit base;
+};
+
+int  nv50_devinit_init(struct nouveau_object *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
index d62045f..821cd75 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
@@ -57,7 +57,57 @@ nouveau_fb_bios_memtype(struct nouveau_bios *bios)
 }
 
 int
-nouveau_fb_preinit(struct nouveau_fb *pfb)
+_nouveau_fb_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_fb *pfb = (void *)object;
+	int ret;
+
+	ret = nv_ofuncs(pfb->ram)->fini(nv_object(pfb->ram), suspend);
+	if (ret && suspend)
+		return ret;
+
+	return nouveau_subdev_fini(&pfb->base, suspend);
+}
+
+int
+_nouveau_fb_init(struct nouveau_object *object)
+{
+	struct nouveau_fb *pfb = (void *)object;
+	int ret, i;
+
+	ret = nouveau_subdev_init(&pfb->base);
+	if (ret)
+		return ret;
+
+	ret = nv_ofuncs(pfb->ram)->init(nv_object(pfb->ram));
+	if (ret)
+		return ret;
+
+	for (i = 0; i < pfb->tile.regions; i++)
+		pfb->tile.prog(pfb, i, &pfb->tile.region[i]);
+
+	return 0;
+}
+
+void
+_nouveau_fb_dtor(struct nouveau_object *object)
+{
+	struct nouveau_fb *pfb = (void *)object;
+	int i;
+
+	for (i = 0; i < pfb->tile.regions; i++)
+		pfb->tile.fini(pfb, i, &pfb->tile.region[i]);
+	nouveau_mm_fini(&pfb->tags);
+	nouveau_mm_fini(&pfb->vram);
+
+	nouveau_object_ref(NULL, (struct nouveau_object **)&pfb->ram);
+	nouveau_subdev_destroy(&pfb->base);
+}
+
+int
+nouveau_fb_create_(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, struct nouveau_oclass *ramcls,
+		   int length, void **pobject)
 {
 	static const char *name[] = {
 		[NV_MEM_TYPE_UNKNOWN] = "unknown",
@@ -72,69 +122,42 @@ nouveau_fb_preinit(struct nouveau_fb *pfb)
 		[NV_MEM_TYPE_GDDR4  ] = "GDDR4",
 		[NV_MEM_TYPE_GDDR5  ] = "GDDR5",
 	};
-	int ret, tags;
+	struct nouveau_object *ram;
+	struct nouveau_fb *pfb;
+	int ret;
 
-	tags = pfb->ram.init(pfb);
-	if (tags < 0 || !pfb->ram.size) {
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "PFB", "fb",
+				     length, pobject);
+	pfb = *pobject;
+	if (ret)
+		return ret;
+
+	ret = nouveau_object_ctor(nv_object(pfb), nv_object(pfb),
+				  ramcls, NULL, 0, &ram);
+	if (ret) {
 		nv_fatal(pfb, "error detecting memory configuration!!\n");
-		return (tags < 0) ? tags : -ERANGE;
+		return ret;
 	}
 
+	atomic_dec(&ram->parent->refcount);
+	atomic_dec(&ram->engine->refcount);
+	pfb->ram = (void *)ram;
+
 	if (!nouveau_mm_initialised(&pfb->vram)) {
-		ret = nouveau_mm_init(&pfb->vram, 0, pfb->ram.size >> 12, 1);
+		ret = nouveau_mm_init(&pfb->vram, 0, pfb->ram->size >> 12, 1);
 		if (ret)
 			return ret;
 	}
 
 	if (!nouveau_mm_initialised(&pfb->tags)) {
-		ret = nouveau_mm_init(&pfb->tags, 0, tags ? ++tags : 0, 1);
+		ret = nouveau_mm_init(&pfb->tags, 0, pfb->ram->tags ?
+				     ++pfb->ram->tags : 0, 1);
 		if (ret)
 			return ret;
 	}
 
-	nv_info(pfb, "RAM type: %s\n", name[pfb->ram.type]);
-	nv_info(pfb, "RAM size: %d MiB\n", (int)(pfb->ram.size >> 20));
-	nv_info(pfb, "   ZCOMP: %d tags\n", tags);
+	nv_info(pfb, "RAM type: %s\n", name[pfb->ram->type]);
+	nv_info(pfb, "RAM size: %d MiB\n", (int)(pfb->ram->size >> 20));
+	nv_info(pfb, "   ZCOMP: %d tags\n", pfb->ram->tags);
 	return 0;
 }
-
-void
-nouveau_fb_destroy(struct nouveau_fb *pfb)
-{
-	int i;
-
-	for (i = 0; i < pfb->tile.regions; i++)
-		pfb->tile.fini(pfb, i, &pfb->tile.region[i]);
-	nouveau_mm_fini(&pfb->tags);
-	nouveau_mm_fini(&pfb->vram);
-
-	nouveau_subdev_destroy(&pfb->base);
-}
-
-void
-_nouveau_fb_dtor(struct nouveau_object *object)
-{
-	struct nouveau_fb *pfb = (void *)object;
-	nouveau_fb_destroy(pfb);
-}
-int
-nouveau_fb_init(struct nouveau_fb *pfb)
-{
-	int ret, i;
-
-	ret = nouveau_subdev_init(&pfb->base);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < pfb->tile.regions; i++)
-		pfb->tile.prog(pfb, i, &pfb->tile.region[i]);
-
-	return 0;
-}
-
-int
-_nouveau_fb_init(struct nouveau_object *object)
-{
-	struct nouveau_fb *pfb = (void *)object;
-	return nouveau_fb_init(pfb);
-}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c
index 6e369f8..1f103c7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c
@@ -22,24 +22,8 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
-#define NV04_PFB_BOOT_0						0x00100000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
-#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
-#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
-#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
 #define NV04_PFB_CFG0						0x00100200
 
 struct nv04_fb_priv {
@@ -56,37 +40,6 @@ nv04_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 }
 
 static int
-nv04_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 boot0 = nv_rd32(pfb, NV04_PFB_BOOT_0);
-	if (boot0 & 0x00000100) {
-		pfb->ram.size  = ((boot0 >> 12) & 0xf) * 2 + 2;
-		pfb->ram.size *= 1024 * 1024;
-	} else {
-		switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
-			pfb->ram.size = 32 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
-			pfb->ram.size = 16 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
-			pfb->ram.size = 8 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
-			pfb->ram.size = 4 * 1024 * 1024;
-			break;
-		}
-	}
-
-	if ((boot0 & 0x00000038) <= 0x10)
-		pfb->ram.type = NV_MEM_TYPE_SGRAM;
-	else
-		pfb->ram.type = NV_MEM_TYPE_SDRAM;
-	return 0;
-}
-
-static int
 nv04_fb_init(struct nouveau_object *object)
 {
 	struct nv04_fb_priv *priv = (void *)object;
@@ -112,14 +65,13 @@ nv04_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv04_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv04_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv04_fb_vram_init;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c
index edbbe26..be069b5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c
@@ -24,25 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv10_fb_priv {
 	struct nouveau_fb base;
 };
 
-static int
-nv10_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 cfg0 = nv_rd32(pfb, 0x100200);
-	if (cfg0 & 0x00000001)
-		pfb->ram.type = NV_MEM_TYPE_DDR1;
-	else
-		pfb->ram.type = NV_MEM_TYPE_SDRAM;
-
-	pfb->ram.size = nv_rd32(pfb, 0x10020c) & 0xff000000;
-	return 0;
-}
-
 void
 nv10_fb_tile_init(struct nouveau_fb *pfb, int i, u32 addr, u32 size, u32 pitch,
 		  u32 flags, struct nouveau_fb_tile *tile)
@@ -78,18 +65,17 @@ nv10_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv10_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv10_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv10_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv10_fb_tile_init;
 	priv->base.tile.fini = nv10_fb_tile_fini;
 	priv->base.tile.prog = nv10_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c
index 4836684..57a2af0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c
@@ -24,38 +24,13 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv1a_fb_priv {
 	struct nouveau_fb base;
 };
 
 static int
-nv1a_fb_vram_init(struct nouveau_fb *pfb)
-{
-	struct pci_dev *bridge;
-	u32 mem, mib;
-
-	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
-	if (!bridge) {
-		nv_fatal(pfb, "no bridge device\n");
-		return -ENODEV;
-	}
-
-	if (nv_device(pfb)->chipset == 0x1a) {
-		pci_read_config_dword(bridge, 0x7c, &mem);
-		mib = ((mem >> 6) & 31) + 1;
-	} else {
-		pci_read_config_dword(bridge, 0x84, &mem);
-		mib = ((mem >> 4) & 127) + 1;
-	}
-
-	pfb->ram.type = NV_MEM_TYPE_STOLEN;
-	pfb->ram.size = mib * 1024 * 1024;
-	return 0;
-}
-
-static int
 nv1a_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -63,18 +38,17 @@ nv1a_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv1a_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv1a_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv1a_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv10_fb_tile_init;
 	priv->base.tile.fini = nv10_fb_tile_fini;
 	priv->base.tile.prog = nv10_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
index 5d14612..b18c4e6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
@@ -24,29 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv20_fb_priv {
 	struct nouveau_fb base;
 };
 
-int
-nv20_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pbus1218 = nv_rd32(pfb, 0x001218);
-
-	switch (pbus1218 & 0x00000300) {
-	case 0x00000000: pfb->ram.type = NV_MEM_TYPE_SDRAM; break;
-	case 0x00000100: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000200: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000300: pfb->ram.type = NV_MEM_TYPE_GDDR2; break;
-	}
-	pfb->ram.size  = (nv_rd32(pfb, 0x10020c) & 0xff000000);
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-
-	return nv_rd32(pfb, 0x100320);
-}
-
 void
 nv20_fb_tile_init(struct nouveau_fb *pfb, int i, u32 addr, u32 size, u32 pitch,
 		  u32 flags, struct nouveau_fb_tile *tile)
@@ -65,7 +48,7 @@ nv20_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00000000; /* Z16 */
 		else              tile->zcomp = 0x04000000; /* Z24S8 */
@@ -105,19 +88,18 @@ nv20_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv20_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv20_fb_tile_init;
 	priv->base.tile.comp = nv20_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
index 0042ace..32ccabf 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv25_fb_priv {
 	struct nouveau_fb base;
@@ -35,7 +35,7 @@ nv25_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00100000; /* Z16 */
 		else              tile->zcomp = 0x00200000; /* Z24S8 */
@@ -54,19 +54,18 @@ nv25_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv25_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv20_fb_tile_init;
 	priv->base.tile.comp = nv25_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
index a7ba0d0..bef756d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv30_fb_priv {
 	struct nouveau_fb base;
@@ -54,7 +54,7 @@ nv30_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x01000000; /* Z16 */
 		else           tile->zcomp |= 0x02000000; /* Z24S8 */
@@ -132,19 +132,18 @@ nv30_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv30_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv30_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
index 092f6f4..097d8e3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv35_fb_priv {
 	struct nouveau_fb base;
@@ -35,7 +35,7 @@ nv35_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x04000000; /* Z16 */
 		else           tile->zcomp |= 0x08000000; /* Z24S8 */
@@ -55,19 +55,18 @@ nv35_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv35_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv35_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
index 797ab3b..9d6d9df 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv36_fb_priv {
 	struct nouveau_fb base;
@@ -35,7 +35,7 @@ nv36_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x10000000; /* Z16 */
 		else           tile->zcomp |= 0x20000000; /* Z24S8 */
@@ -55,19 +55,18 @@ nv36_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv36_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv36_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
index 65e131b..33b4393 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
@@ -24,34 +24,18 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv40_fb_priv {
 	struct nouveau_fb base;
 };
 
-static int
-nv40_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pbus1218 = nv_rd32(pfb, 0x001218);
-	switch (pbus1218 & 0x00000300) {
-	case 0x00000000: pfb->ram.type = NV_MEM_TYPE_SDRAM; break;
-	case 0x00000100: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000200: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000300: pfb->ram.type = NV_MEM_TYPE_DDR2; break;
-	}
-
-	pfb->ram.size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-	return nv_rd32(pfb, 0x100320);
-}
-
 void
 nv40_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x80);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x100);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x100);
 	if ( (flags & 2) &&
 	    !nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		tile->zcomp  = 0x28000000; /* Z24S8_SPLIT_GRAD */
@@ -85,19 +69,18 @@ nv40_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv40_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv40_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv40_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c
index e9e5a08..02cd837 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c
@@ -24,28 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv41_fb_priv {
 	struct nouveau_fb base;
 };
 
-int
-nv41_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pfb474 = nv_rd32(pfb, 0x100474);
-	if (pfb474 & 0x00000004)
-		pfb->ram.type = NV_MEM_TYPE_GDDR3;
-	if (pfb474 & 0x00000002)
-		pfb->ram.type = NV_MEM_TYPE_DDR2;
-	if (pfb474 & 0x00000001)
-		pfb->ram.type = NV_MEM_TYPE_DDR1;
-
-	pfb->ram.size =   nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-	return nv_rd32(pfb, 0x100320);
-}
-
 void
 nv41_fb_tile_prog(struct nouveau_fb *pfb, int i, struct nouveau_fb_tile *tile)
 {
@@ -78,19 +62,18 @@ nv41_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv41_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv41_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv41_fb_vram_init;
 	priv->base.tile.regions = 12;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv41_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c
index ae89b50..c5246c2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c
@@ -24,27 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv44_fb_priv {
 	struct nouveau_fb base;
 };
 
-int
-nv44_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pfb474 = nv_rd32(pfb, 0x100474);
-	if (pfb474 & 0x00000004)
-		pfb->ram.type = NV_MEM_TYPE_GDDR3;
-	if (pfb474 & 0x00000002)
-		pfb->ram.type = NV_MEM_TYPE_DDR2;
-	if (pfb474 & 0x00000001)
-		pfb->ram.type = NV_MEM_TYPE_DDR1;
-
-	pfb->ram.size = nv_rd32(pfb, 0x10020c) & 0xff000000;
-	return 0;
-}
-
 static void
 nv44_fb_tile_init(struct nouveau_fb *pfb, int i, u32 addr, u32 size, u32 pitch,
 		  u32 flags, struct nouveau_fb_tile *tile)
@@ -87,18 +72,17 @@ nv44_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv44_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv44_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv44_fb_vram_init;
 	priv->base.tile.regions = 12;
 	priv->base.tile.init = nv44_fb_tile_init;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv44_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c
index 589b93e..e2b5790 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv46_fb_priv {
 	struct nouveau_fb base;
@@ -52,18 +52,17 @@ nv46_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv46_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv44_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv44_fb_vram_init;
 	priv->base.tile.regions = 15;
 	priv->base.tile.init = nv46_fb_tile_init;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv44_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c
index 818bba3..fe6a227 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv47_fb_priv {
 	struct nouveau_fb base;
@@ -38,19 +38,18 @@ nv47_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv47_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv41_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv41_fb_vram_init;
 	priv->base.tile.regions = 15;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv41_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c
index 84a31af..5eca99b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c
@@ -24,30 +24,13 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv49_fb_priv {
 	struct nouveau_fb base;
 };
 
 static int
-nv49_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pfb914 = nv_rd32(pfb, 0x100914);
-
-	switch (pfb914 & 0x00000003) {
-	case 0x00000000: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000001: pfb->ram.type = NV_MEM_TYPE_DDR2; break;
-	case 0x00000002: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000003: break;
-	}
-
-	pfb->ram.size =   nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-	return nv_rd32(pfb, 0x100320);
-}
-
-static int
 nv49_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -55,20 +38,18 @@ nv49_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv49_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv49_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv49_fb_vram_init;
 	priv->base.tile.regions = 15;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv41_fb_tile_prog;
-
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c
index 797fd55..1190b78 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c
@@ -24,21 +24,13 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv4e_fb_priv {
 	struct nouveau_fb base;
 };
 
 static int
-nv4e_fb_vram_init(struct nouveau_fb *pfb)
-{
-	pfb->ram.size = nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.type = NV_MEM_TYPE_STOLEN;
-	return 0;
-}
-
-static int
 nv4e_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -46,18 +38,17 @@ nv4e_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv4e_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv4e_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv4e_fb_vram_init;
 	priv->base.tile.regions = 12;
 	priv->base.tile.init = nv46_fb_tile_init;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv44_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 0772ec9..da614ec 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -27,7 +27,7 @@
 #include <core/engctx.h>
 #include <core/object.h>
 
-#include <subdev/fb.h>
+#include "priv.h"
 #include <subdev/bios.h>
 
 struct nv50_fb_priv {
@@ -36,7 +36,8 @@ struct nv50_fb_priv {
 	dma_addr_t r100c08;
 };
 
-static int types[0x80] = {
+int
+nv50_fb_memtype[0x80] = {
 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
 	1, 1, 1, 1, 0, 0, 0, 0, 2, 2, 2, 2, 0, 0, 0, 0,
 	1, 1, 1, 1, 1, 1, 1, 0, 2, 2, 2, 2, 2, 2, 2, 0,
@@ -50,192 +51,7 @@ static int types[0x80] = {
 static bool
 nv50_fb_memtype_valid(struct nouveau_fb *pfb, u32 memtype)
 {
-	return types[(memtype & 0xff00) >> 8] != 0;
-}
-
-static u32
-nv50_fb_vram_rblock(struct nouveau_fb *pfb)
-{
-	int i, parts, colbits, rowbitsa, rowbitsb, banks;
-	u64 rowsize, predicted;
-	u32 r0, r4, rt, ru, rblock_size;
-
-	r0 = nv_rd32(pfb, 0x100200);
-	r4 = nv_rd32(pfb, 0x100204);
-	rt = nv_rd32(pfb, 0x100250);
-	ru = nv_rd32(pfb, 0x001540);
-	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
-
-	for (i = 0, parts = 0; i < 8; i++) {
-		if (ru & (0x00010000 << i))
-			parts++;
-	}
-
-	colbits  =  (r4 & 0x0000f000) >> 12;
-	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
-	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
-	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
-
-	rowsize = parts * banks * (1 << colbits) * 8;
-	predicted = rowsize << rowbitsa;
-	if (r0 & 0x00000004)
-		predicted += rowsize << rowbitsb;
-
-	if (predicted != pfb->ram.size) {
-		nv_warn(pfb, "memory controller reports %d MiB VRAM\n",
-			(u32)(pfb->ram.size >> 20));
-	}
-
-	rblock_size = rowsize;
-	if (rt & 1)
-		rblock_size *= 3;
-
-	nv_debug(pfb, "rblock %d bytes\n", rblock_size);
-	return rblock_size;
-}
-
-static int
-nv50_fb_vram_init(struct nouveau_fb *pfb)
-{
-	struct nouveau_device *device = nv_device(pfb);
-	struct nouveau_bios *bios = nouveau_bios(device);
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 size, tags = 0;
-	int ret;
-
-	pfb->ram.size = nv_rd32(pfb, 0x10020c);
-	pfb->ram.size = (pfb->ram.size & 0xffffff00) |
-		       ((pfb->ram.size & 0x000000ff) << 32);
-
-	size = (pfb->ram.size >> 12) - rsvd_head - rsvd_tail;
-	switch (device->chipset) {
-	case 0xaa:
-	case 0xac:
-	case 0xaf: /* IGPs, no reordering, no real VRAM */
-		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size, 1);
-		if (ret)
-			return ret;
-
-		pfb->ram.type   = NV_MEM_TYPE_STOLEN;
-		pfb->ram.stolen = (u64)nv_rd32(pfb, 0x100e10) << 12;
-		break;
-	default:
-		switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
-		case 0: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-		case 1:
-			if (nouveau_fb_bios_memtype(bios) == NV_MEM_TYPE_DDR3)
-				pfb->ram.type = NV_MEM_TYPE_DDR3;
-			else
-				pfb->ram.type = NV_MEM_TYPE_DDR2;
-			break;
-		case 2: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-		case 3: pfb->ram.type = NV_MEM_TYPE_GDDR4; break;
-		case 4: pfb->ram.type = NV_MEM_TYPE_GDDR5; break;
-		default:
-			break;
-		}
-
-		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size,
-				      nv50_fb_vram_rblock(pfb) >> 12);
-		if (ret)
-			return ret;
-
-		pfb->ram.ranks = (nv_rd32(pfb, 0x100200) & 0x4) ? 2 : 1;
-		tags = nv_rd32(pfb, 0x100320);
-		break;
-	}
-
-	return tags;
-}
-
-static int
-nv50_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
-		 u32 memtype, struct nouveau_mem **pmem)
-{
-	struct nv50_fb_priv *priv = (void *)pfb;
-	struct nouveau_mm *heap = &priv->base.vram;
-	struct nouveau_mm *tags = &priv->base.tags;
-	struct nouveau_mm_node *r;
-	struct nouveau_mem *mem;
-	int comp = (memtype & 0x300) >> 8;
-	int type = (memtype & 0x07f);
-	int back = (memtype & 0x800);
-	int min, max, ret;
-
-	max = (size >> 12);
-	min = ncmin ? (ncmin >> 12) : max;
-	align >>= 12;
-
-	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
-	if (!mem)
-		return -ENOMEM;
-
-	mutex_lock(&pfb->base.mutex);
-	if (comp) {
-		if (align == 16) {
-			int n = (max >> 4) * comp;
-
-			ret = nouveau_mm_head(tags, 1, n, n, 1, &mem->tag);
-			if (ret)
-				mem->tag = NULL;
-		}
-
-		if (unlikely(!mem->tag))
-			comp = 0;
-	}
-
-	INIT_LIST_HEAD(&mem->regions);
-	mem->memtype = (comp << 7) | type;
-	mem->size = max;
-
-	type = types[type];
-	do {
-		if (back)
-			ret = nouveau_mm_tail(heap, type, max, min, align, &r);
-		else
-			ret = nouveau_mm_head(heap, type, max, min, align, &r);
-		if (ret) {
-			mutex_unlock(&pfb->base.mutex);
-			pfb->ram.put(pfb, &mem);
-			return ret;
-		}
-
-		list_add_tail(&r->rl_entry, &mem->regions);
-		max -= r->length;
-	} while (max);
-	mutex_unlock(&pfb->base.mutex);
-
-	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
-	mem->offset = (u64)r->offset << 12;
-	*pmem = mem;
-	return 0;
-}
-
-void
-nv50_fb_vram_del(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
-{
-	struct nv50_fb_priv *priv = (void *)pfb;
-	struct nouveau_mm_node *this;
-	struct nouveau_mem *mem;
-
-	mem = *pmem;
-	*pmem = NULL;
-	if (unlikely(mem == NULL))
-		return;
-
-	mutex_lock(&pfb->base.mutex);
-	while (!list_empty(&mem->regions)) {
-		this = list_first_entry(&mem->regions, typeof(*this), rl_entry);
-
-		list_del(&this->rl_entry);
-		nouveau_mm_free(&priv->base.vram, &this);
-	}
-
-	nouveau_mm_free(&priv->base.tags, &mem->tag);
-	mutex_unlock(&pfb->base.mutex);
-
-	kfree(mem);
+	return nv50_fb_memtype[(memtype & 0xff00) >> 8] != 0;
 }
 
 static const struct nouveau_enum vm_dispatch_subclients[] = {
@@ -432,7 +248,7 @@ nv50_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv50_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -449,11 +265,8 @@ nv50_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	}
 
 	priv->base.memtype_valid = nv50_fb_memtype_valid;
-	priv->base.ram.init = nv50_fb_vram_init;
-	priv->base.ram.get = nv50_fb_vram_new;
-	priv->base.ram.put = nv50_fb_vram_del;
 	nv_subdev(priv)->intr = nv50_fb_intr;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 static void
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
index 86ad592..f35d76f 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
@@ -22,9 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/fb.h>
-#include <subdev/ltcg.h>
-#include <subdev/bios.h>
+#include "priv.h"
 
 struct nvc0_fb_priv {
 	struct nouveau_fb base;
@@ -34,7 +32,6 @@ struct nvc0_fb_priv {
 
 extern const u8 nvc0_pte_storage_type_map[256];
 
-
 static bool
 nvc0_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 {
@@ -43,137 +40,6 @@ nvc0_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 }
 
 static int
-nvc0_fb_vram_init(struct nouveau_fb *pfb)
-{
-	struct nouveau_bios *bios = nouveau_bios(pfb);
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 parts = nv_rd32(pfb, 0x022438);
-	u32 pmask = nv_rd32(pfb, 0x022554);
-	u32 bsize = nv_rd32(pfb, 0x10f20c);
-	u32 offset, length;
-	bool uniform = true;
-	int ret, part;
-
-	nv_debug(pfb, "0x100800: 0x%08x\n", nv_rd32(pfb, 0x100800));
-	nv_debug(pfb, "parts 0x%08x mask 0x%08x\n", parts, pmask);
-
-	pfb->ram.type = nouveau_fb_bios_memtype(bios);
-	pfb->ram.ranks = (nv_rd32(pfb, 0x10f200) & 0x00000004) ? 2 : 1;
-
-	/* read amount of vram attached to each memory controller */
-	for (part = 0; part < parts; part++) {
-		if (!(pmask & (1 << part))) {
-			u32 psize = nv_rd32(pfb, 0x11020c + (part * 0x1000));
-			if (psize != bsize) {
-				if (psize < bsize)
-					bsize = psize;
-				uniform = false;
-			}
-
-			nv_debug(pfb, "%d: mem_amount 0x%08x\n", part, psize);
-			pfb->ram.size += (u64)psize << 20;
-		}
-	}
-
-	/* if all controllers have the same amount attached, there's no holes */
-	if (uniform) {
-		offset = rsvd_head;
-		length = (pfb->ram.size >> 12) - rsvd_head - rsvd_tail;
-		return nouveau_mm_init(&pfb->vram, offset, length, 1);
-	}
-
-	/* otherwise, address lowest common amount from 0GiB */
-	ret = nouveau_mm_init(&pfb->vram, rsvd_head, (bsize << 8) * parts, 1);
-	if (ret)
-		return ret;
-
-	/* and the rest starting from (8GiB + common_size) */
-	offset = (0x0200000000ULL >> 12) + (bsize << 8);
-	length = (pfb->ram.size >> 12) - (bsize << 8) - rsvd_tail;
-
-	ret = nouveau_mm_init(&pfb->vram, offset, length, 0);
-	if (ret) {
-		nouveau_mm_fini(&pfb->vram);
-		return ret;
-	}
-
-	return 0;
-}
-
-static int
-nvc0_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
-		 u32 memtype, struct nouveau_mem **pmem)
-{
-	struct nouveau_mm *mm = &pfb->vram;
-	struct nouveau_mm_node *r;
-	struct nouveau_mem *mem;
-	int type = (memtype & 0x0ff);
-	int back = (memtype & 0x800);
-	int ret;
-	const bool comp = nvc0_pte_storage_type_map[type] != type;
-
-	size  >>= 12;
-	align >>= 12;
-	ncmin >>= 12;
-	if (!ncmin)
-		ncmin = size;
-
-	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
-	if (!mem)
-		return -ENOMEM;
-
-	INIT_LIST_HEAD(&mem->regions);
-	mem->size = size;
-
-	mutex_lock(&pfb->base.mutex);
-	if (comp) {
-		struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb->base.base.parent);
-
-		/* compression only works with lpages */
-		if (align == (1 << (17 - 12))) {
-			int n = size >> 5;
-			ltcg->tags_alloc(ltcg, n, &mem->tag);
-		}
-		if (unlikely(!mem->tag))
-			type = nvc0_pte_storage_type_map[type];
-	}
-	mem->memtype = type;
-
-	do {
-		if (back)
-			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
-		else
-			ret = nouveau_mm_head(mm, 1, size, ncmin, align, &r);
-		if (ret) {
-			mutex_unlock(&pfb->base.mutex);
-			pfb->ram.put(pfb, &mem);
-			return ret;
-		}
-
-		list_add_tail(&r->rl_entry, &mem->regions);
-		size -= r->length;
-	} while (size);
-	mutex_unlock(&pfb->base.mutex);
-
-	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
-	mem->offset = (u64)r->offset << 12;
-	*pmem = mem;
-	return 0;
-}
-
-static void
-nvc0_fb_vram_del(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
-{
-	struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb->base.base.parent);
-
-	if ((*pmem)->tag)
-		ltcg->tags_free(ltcg, &(*pmem)->tag);
-
-	nv50_fb_vram_del(pfb, pmem);
-}
-
-static int
 nvc0_fb_init(struct nouveau_object *object)
 {
 	struct nvc0_fb_priv *priv = (void *)object;
@@ -212,15 +78,12 @@ nvc0_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nvc0_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nvc0_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nvc0_fb_memtype_valid;
-	priv->base.ram.init = nvc0_fb_vram_init;
-	priv->base.ram.get = nvc0_fb_vram_new;
-	priv->base.ram.put = nvc0_fb_vram_del;
 
 	priv->r100c10_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
 	if (priv->r100c10_page) {
@@ -231,7 +94,7 @@ nvc0_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 			return -EFAULT;
 	}
 
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
new file mode 100644
index 0000000..6c974dd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
@@ -0,0 +1,87 @@
+#ifndef __NVKM_FB_PRIV_H__
+#define __NVKM_FB_PRIV_H__
+
+#include <subdev/fb.h>
+
+#define nouveau_ram_create(p,e,o,d)                                            \
+	nouveau_object_create_((p), (e), (o), 0, sizeof(**d), (void **)d)
+#define nouveau_ram_destroy(p)                                                 \
+	nouveau_object_destroy(&(p)->base)
+#define nouveau_ram_init(p)                                                    \
+	nouveau_object_init(&(p)->base)
+#define nouveau_ram_fini(p,s)                                                  \
+	nouveau_object_fini(&(p)->base, (s))
+
+#define _nouveau_ram_dtor nouveau_object_destroy
+#define _nouveau_ram_init nouveau_object_init
+#define _nouveau_ram_fini nouveau_object_fini
+
+extern struct nouveau_oclass nv04_ram_oclass;
+extern struct nouveau_oclass nv10_ram_oclass;
+extern struct nouveau_oclass nv1a_ram_oclass;
+extern struct nouveau_oclass nv20_ram_oclass;
+extern struct nouveau_oclass nv40_ram_oclass;
+extern struct nouveau_oclass nv41_ram_oclass;
+extern struct nouveau_oclass nv44_ram_oclass;
+extern struct nouveau_oclass nv49_ram_oclass;
+extern struct nouveau_oclass nv4e_ram_oclass;
+extern struct nouveau_oclass nv50_ram_oclass;
+extern struct nouveau_oclass nvc0_ram_oclass;
+
+#define nouveau_fb_create(p,e,c,r,d)                                           \
+	nouveau_fb_create_((p), (e), (c), (r), sizeof(**d), (void **)d)
+#define nouveau_fb_destroy(p) ({                                               \
+	struct nouveau_fb *pfb = (p);                                          \
+	_nouveau_fb_dtor(nv_object(pfb));                                      \
+})
+#define nouveau_fb_init(p) ({                                                  \
+	struct nouveau_fb *pfb = (p);                                          \
+	_nouveau_fb_init(nv_object(pfb));                                      \
+})
+#define nouveau_fb_fini(p,s) ({                                                \
+	struct nouveau_fb *pfb = (p);                                          \
+	_nouveau_fb_fini(nv_object(pfb), (s));                                 \
+})
+
+int nouveau_fb_create_(struct nouveau_object *, struct nouveau_object *,
+		       struct nouveau_oclass *, struct nouveau_oclass *,
+		       int length, void **pobject);
+void _nouveau_fb_dtor(struct nouveau_object *);
+int  _nouveau_fb_init(struct nouveau_object *);
+int  _nouveau_fb_fini(struct nouveau_object *, bool);
+
+struct nouveau_bios;
+int  nouveau_fb_bios_memtype(struct nouveau_bios *);
+
+bool nv04_fb_memtype_valid(struct nouveau_fb *, u32 memtype);
+
+void nv10_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+void nv10_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
+void nv10_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+void nv20_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+void nv20_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
+void nv20_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+int  nv30_fb_init(struct nouveau_object *);
+void nv30_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+
+void nv40_fb_tile_comp(struct nouveau_fb *, int i, u32 size, u32 flags,
+		       struct nouveau_fb_tile *);
+
+int  nv41_fb_init(struct nouveau_object *);
+void nv41_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+int  nv44_fb_init(struct nouveau_object *);
+void nv44_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+void nv46_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+
+void nv50_ram_put(struct nouveau_fb *, struct nouveau_mem **);
+extern int nv50_fb_memtype[0x80];
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
new file mode 100644
index 0000000..e781080
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#define NV04_PFB_BOOT_0						0x00100000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
+#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
+#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
+#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+
+#include "priv.h"
+
+static int
+nv04_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 boot0 = nv_rd32(pfb, NV04_PFB_BOOT_0);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (boot0 & 0x00000100) {
+		ram->size  = ((boot0 >> 12) & 0xf) * 2 + 2;
+		ram->size *= 1024 * 1024;
+	} else {
+		switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
+			ram->size = 32 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
+			ram->size = 16 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
+			ram->size = 8 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
+			ram->size = 4 * 1024 * 1024;
+			break;
+		}
+	}
+
+	if ((boot0 & 0x00000038) <= 0x10)
+		ram->type = NV_MEM_TYPE_SGRAM;
+	else
+		ram->type = NV_MEM_TYPE_SDRAM;
+	return 0;
+}
+
+struct nouveau_oclass
+nv04_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv04_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv10.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv10.c
new file mode 100644
index 0000000..8311f37
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv10.c
@@ -0,0 +1,61 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv10_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 cfg0 = nv_rd32(pfb, 0x100200);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (cfg0 & 0x00000001)
+		ram->type = NV_MEM_TYPE_DDR1;
+	else
+		ram->type = NV_MEM_TYPE_SDRAM;
+
+	ram->size = nv_rd32(pfb, 0x10020c) & 0xff000000;
+	return 0;
+}
+
+
+struct nouveau_oclass
+nv10_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv10_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv1a.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv1a.c
new file mode 100644
index 0000000..d0caddf
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv1a.c
@@ -0,0 +1,71 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv1a_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	struct pci_dev *bridge;
+	u32 mem, mib;
+	int ret;
+
+	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
+	if (!bridge) {
+		nv_fatal(pfb, "no bridge device\n");
+		return -ENODEV;
+	}
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (nv_device(pfb)->chipset == 0x1a) {
+		pci_read_config_dword(bridge, 0x7c, &mem);
+		mib = ((mem >> 6) & 31) + 1;
+	} else {
+		pci_read_config_dword(bridge, 0x84, &mem);
+		mib = ((mem >> 4) & 127) + 1;
+	}
+
+	ram->type = NV_MEM_TYPE_STOLEN;
+	ram->size = mib * 1024 * 1024;
+	return 0;
+}
+
+struct nouveau_oclass
+nv1a_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv1a_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv20.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv20.c
new file mode 100644
index 0000000..fdc11bb
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv20.c
@@ -0,0 +1,63 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv20_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pbus1218 = nv_rd32(pfb, 0x001218);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	switch (pbus1218 & 0x00000300) {
+	case 0x00000000: ram->type = NV_MEM_TYPE_SDRAM; break;
+	case 0x00000100: ram->type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000200: ram->type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000300: ram->type = NV_MEM_TYPE_GDDR2; break;
+	}
+	ram->size  = (nv_rd32(pfb, 0x10020c) & 0xff000000);
+	ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	ram->tags  = nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+struct nouveau_oclass
+nv20_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv20_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv40.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv40.c
new file mode 100644
index 0000000..ee49ac4
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv40.c
@@ -0,0 +1,65 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv40_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pbus1218 = nv_rd32(pfb, 0x001218);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	switch (pbus1218 & 0x00000300) {
+	case 0x00000000: ram->type = NV_MEM_TYPE_SDRAM; break;
+	case 0x00000100: ram->type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000200: ram->type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000300: ram->type = NV_MEM_TYPE_DDR2; break;
+	}
+
+	ram->size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
+	ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	ram->tags  =  nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+
+struct nouveau_oclass
+nv40_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv40_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv41.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv41.c
new file mode 100644
index 0000000..1dab7e1
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv41.c
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv41_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pfb474 = nv_rd32(pfb, 0x100474);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (pfb474 & 0x00000004)
+		ram->type = NV_MEM_TYPE_GDDR3;
+	if (pfb474 & 0x00000002)
+		ram->type = NV_MEM_TYPE_DDR2;
+	if (pfb474 & 0x00000001)
+		ram->type = NV_MEM_TYPE_DDR1;
+
+	ram->size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
+	ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	ram->tags  =  nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+struct nouveau_oclass
+nv41_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv41_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv44.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv44.c
new file mode 100644
index 0000000..25fff84
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv44.c
@@ -0,0 +1,62 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv44_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pfb474 = nv_rd32(pfb, 0x100474);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (pfb474 & 0x00000004)
+		ram->type = NV_MEM_TYPE_GDDR3;
+	if (pfb474 & 0x00000002)
+		ram->type = NV_MEM_TYPE_DDR2;
+	if (pfb474 & 0x00000001)
+		ram->type = NV_MEM_TYPE_DDR1;
+
+	ram->size = nv_rd32(pfb, 0x10020c) & 0xff000000;
+	return 0;
+}
+
+struct nouveau_oclass
+nv44_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv44_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv49.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv49.c
new file mode 100644
index 0000000..19e3a9a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv49.c
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv49_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pfb914 = nv_rd32(pfb, 0x100914);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	switch (pfb914 & 0x00000003) {
+	case 0x00000000: pfb->ram->type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000001: pfb->ram->type = NV_MEM_TYPE_DDR2; break;
+	case 0x00000002: pfb->ram->type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000003: break;
+	}
+
+	pfb->ram->size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
+	pfb->ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	pfb->ram->tags  =  nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+struct nouveau_oclass
+nv49_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv49_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv4e.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv4e.c
new file mode 100644
index 0000000..7192aa6
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv4e.c
@@ -0,0 +1,55 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv4e_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	pfb->ram->size = nv_rd32(pfb, 0x10020c) & 0xff000000;
+	pfb->ram->type = NV_MEM_TYPE_STOLEN;
+	return 0;
+}
+
+struct nouveau_oclass
+nv4e_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv4e_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
new file mode 100644
index 0000000..af5aa7e
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
@@ -0,0 +1,232 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <core/mm.h>
+#include "priv.h"
+
+void
+nv50_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
+{
+	struct nouveau_mm_node *this;
+	struct nouveau_mem *mem;
+
+	mem = *pmem;
+	*pmem = NULL;
+	if (unlikely(mem == NULL))
+		return;
+
+	mutex_lock(&pfb->base.mutex);
+	while (!list_empty(&mem->regions)) {
+		this = list_first_entry(&mem->regions, typeof(*this), rl_entry);
+
+		list_del(&this->rl_entry);
+		nouveau_mm_free(&pfb->vram, &this);
+	}
+
+	nouveau_mm_free(&pfb->tags, &mem->tag);
+	mutex_unlock(&pfb->base.mutex);
+
+	kfree(mem);
+}
+
+static int
+nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
+	     u32 memtype, struct nouveau_mem **pmem)
+{
+	struct nouveau_mm *heap = &pfb->vram;
+	struct nouveau_mm *tags = &pfb->tags;
+	struct nouveau_mm_node *r;
+	struct nouveau_mem *mem;
+	int comp = (memtype & 0x300) >> 8;
+	int type = (memtype & 0x07f);
+	int back = (memtype & 0x800);
+	int min, max, ret;
+
+	max = (size >> 12);
+	min = ncmin ? (ncmin >> 12) : max;
+	align >>= 12;
+
+	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+	if (!mem)
+		return -ENOMEM;
+
+	mutex_lock(&pfb->base.mutex);
+	if (comp) {
+		if (align == 16) {
+			int n = (max >> 4) * comp;
+
+			ret = nouveau_mm_head(tags, 1, n, n, 1, &mem->tag);
+			if (ret)
+				mem->tag = NULL;
+		}
+
+		if (unlikely(!mem->tag))
+			comp = 0;
+	}
+
+	INIT_LIST_HEAD(&mem->regions);
+	mem->memtype = (comp << 7) | type;
+	mem->size = max;
+
+	type = nv50_fb_memtype[type];
+	do {
+		if (back)
+			ret = nouveau_mm_tail(heap, type, max, min, align, &r);
+		else
+			ret = nouveau_mm_head(heap, type, max, min, align, &r);
+		if (ret) {
+			mutex_unlock(&pfb->base.mutex);
+			pfb->ram->put(pfb, &mem);
+			return ret;
+		}
+
+		list_add_tail(&r->rl_entry, &mem->regions);
+		max -= r->length;
+	} while (max);
+	mutex_unlock(&pfb->base.mutex);
+
+	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
+	mem->offset = (u64)r->offset << 12;
+	*pmem = mem;
+	return 0;
+}
+
+static u32
+nv50_fb_vram_rblock(struct nouveau_fb *pfb, struct nouveau_ram *ram)
+{
+	int i, parts, colbits, rowbitsa, rowbitsb, banks;
+	u64 rowsize, predicted;
+	u32 r0, r4, rt, ru, rblock_size;
+
+	r0 = nv_rd32(pfb, 0x100200);
+	r4 = nv_rd32(pfb, 0x100204);
+	rt = nv_rd32(pfb, 0x100250);
+	ru = nv_rd32(pfb, 0x001540);
+	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
+
+	for (i = 0, parts = 0; i < 8; i++) {
+		if (ru & (0x00010000 << i))
+			parts++;
+	}
+
+	colbits  =  (r4 & 0x0000f000) >> 12;
+	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
+	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
+	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
+
+	rowsize = parts * banks * (1 << colbits) * 8;
+	predicted = rowsize << rowbitsa;
+	if (r0 & 0x00000004)
+		predicted += rowsize << rowbitsb;
+
+	if (predicted != ram->size) {
+		nv_warn(pfb, "memory controller reports %d MiB VRAM\n",
+			(u32)(ram->size >> 20));
+	}
+
+	rblock_size = rowsize;
+	if (rt & 1)
+		rblock_size *= 3;
+
+	nv_debug(pfb, "rblock %d bytes\n", rblock_size);
+	return rblock_size;
+}
+
+static int
+nv50_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 datasize,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_device *device = nv_device(pfb);
+	struct nouveau_bios *bios = nouveau_bios(device);
+	struct nouveau_ram *ram;
+	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
+	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
+	u32 size;
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	ram->size = nv_rd32(pfb, 0x10020c);
+	ram->size = (ram->size & 0xffffff00) |
+		       ((ram->size & 0x000000ff) << 32);
+
+	size = (ram->size >> 12) - rsvd_head - rsvd_tail;
+	switch (device->chipset) {
+	case 0xaa:
+	case 0xac:
+	case 0xaf: /* IGPs, no reordering, no real VRAM */
+		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size, 1);
+		if (ret)
+			return ret;
+
+		ram->type   = NV_MEM_TYPE_STOLEN;
+		ram->stolen = (u64)nv_rd32(pfb, 0x100e10) << 12;
+		break;
+	default:
+		switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
+		case 0: ram->type = NV_MEM_TYPE_DDR1; break;
+		case 1:
+			if (nouveau_fb_bios_memtype(bios) == NV_MEM_TYPE_DDR3)
+				ram->type = NV_MEM_TYPE_DDR3;
+			else
+				ram->type = NV_MEM_TYPE_DDR2;
+			break;
+		case 2: ram->type = NV_MEM_TYPE_GDDR3; break;
+		case 3: ram->type = NV_MEM_TYPE_GDDR4; break;
+		case 4: ram->type = NV_MEM_TYPE_GDDR5; break;
+		default:
+			break;
+		}
+
+		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size,
+				      nv50_fb_vram_rblock(pfb, ram) >> 12);
+		if (ret)
+			return ret;
+
+		ram->ranks = (nv_rd32(pfb, 0x100200) & 0x4) ? 2 : 1;
+		ram->tags  =  nv_rd32(pfb, 0x100320);
+		break;
+	}
+
+	ram->get = nv50_ram_get;
+	ram->put = nv50_ram_put;
+	return 0;
+}
+
+struct nouveau_oclass
+nv50_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv50_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
new file mode 100644
index 0000000..9c3634a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
@@ -0,0 +1,186 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/ltcg.h>
+
+#include "priv.h"
+
+extern const u8 nvc0_pte_storage_type_map[256];
+
+void
+nvc0_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
+{
+	struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb);
+
+	if ((*pmem)->tag)
+		ltcg->tags_free(ltcg, &(*pmem)->tag);
+
+	nv50_ram_put(pfb, pmem);
+}
+
+int
+nvc0_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
+	     u32 memtype, struct nouveau_mem **pmem)
+{
+	struct nouveau_mm *mm = &pfb->vram;
+	struct nouveau_mm_node *r;
+	struct nouveau_mem *mem;
+	int type = (memtype & 0x0ff);
+	int back = (memtype & 0x800);
+	const bool comp = nvc0_pte_storage_type_map[type] != type;
+	int ret;
+
+	size  >>= 12;
+	align >>= 12;
+	ncmin >>= 12;
+	if (!ncmin)
+		ncmin = size;
+
+	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+	if (!mem)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&mem->regions);
+	mem->size = size;
+
+	mutex_lock(&pfb->base.mutex);
+	if (comp) {
+		struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb);
+
+		/* compression only works with lpages */
+		if (align == (1 << (17 - 12))) {
+			int n = size >> 5;
+			ltcg->tags_alloc(ltcg, n, &mem->tag);
+		}
+
+		if (unlikely(!mem->tag))
+			type = nvc0_pte_storage_type_map[type];
+	}
+	mem->memtype = type;
+
+	do {
+		if (back)
+			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
+		else
+			ret = nouveau_mm_head(mm, 1, size, ncmin, align, &r);
+		if (ret) {
+			mutex_unlock(&pfb->base.mutex);
+			pfb->ram->put(pfb, &mem);
+			return ret;
+		}
+
+		list_add_tail(&r->rl_entry, &mem->regions);
+		size -= r->length;
+	} while (size);
+	mutex_unlock(&pfb->base.mutex);
+
+	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
+	mem->offset = (u64)r->offset << 12;
+	*pmem = mem;
+	return 0;
+}
+
+static int
+nvc0_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nouveau_ram *ram;
+	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
+	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
+	u32 parts = nv_rd32(pfb, 0x022438);
+	u32 pmask = nv_rd32(pfb, 0x022554);
+	u32 bsize = nv_rd32(pfb, 0x10f20c);
+	u32 offset, length;
+	bool uniform = true;
+	int ret, part;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	nv_debug(pfb, "0x100800: 0x%08x\n", nv_rd32(pfb, 0x100800));
+	nv_debug(pfb, "parts 0x%08x mask 0x%08x\n", parts, pmask);
+
+	ram->type = nouveau_fb_bios_memtype(bios);
+	ram->ranks = (nv_rd32(pfb, 0x10f200) & 0x00000004) ? 2 : 1;
+
+	/* read amount of vram attached to each memory controller */
+	for (part = 0; part < parts; part++) {
+		if (!(pmask & (1 << part))) {
+			u32 psize = nv_rd32(pfb, 0x11020c + (part * 0x1000));
+			if (psize != bsize) {
+				if (psize < bsize)
+					bsize = psize;
+				uniform = false;
+			}
+
+			nv_debug(pfb, "%d: mem_amount 0x%08x\n", part, psize);
+			ram->size += (u64)psize << 20;
+		}
+	}
+
+	/* if all controllers have the same amount attached, there's no holes */
+	if (uniform) {
+		offset = rsvd_head;
+		length = (ram->size >> 12) - rsvd_head - rsvd_tail;
+		ret = nouveau_mm_init(&pfb->vram, offset, length, 1);
+	} else {
+		/* otherwise, address lowest common amount from 0GiB */
+		ret = nouveau_mm_init(&pfb->vram, rsvd_head,
+				      (bsize << 8) * parts, 1);
+		if (ret)
+			return ret;
+
+		/* and the rest starting from (8GiB + common_size) */
+		offset = (0x0200000000ULL >> 12) + (bsize << 8);
+		length = (ram->size >> 12) - (bsize << 8) - rsvd_tail;
+
+		ret = nouveau_mm_init(&pfb->vram, offset, length, 0);
+		if (ret)
+			nouveau_mm_fini(&pfb->vram);
+	}
+
+	if (ret)
+		return ret;
+
+	ram->get = nvc0_ram_get;
+	ram->put = nvc0_ram_put;
+	return 0;
+}
+
+struct nouveau_oclass
+nvc0_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c
index cfc7e31..97bc5df 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c
@@ -56,7 +56,7 @@ nv50_instobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = pfb->ram.get(pfb, size, align, 0, 0x800, &node->mem);
+	ret = pfb->ram->get(pfb, size, align, 0, 0x800, &node->mem);
 	if (ret)
 		return ret;
 
@@ -71,7 +71,7 @@ nv50_instobj_dtor(struct nouveau_object *object)
 {
 	struct nv50_instobj_priv *node = (void *)object;
 	struct nouveau_fb *pfb = nouveau_fb(object);
-	pfb->ram.put(pfb, &node->mem);
+	pfb->ram->put(pfb, &node->mem);
 	nouveau_instobj_destroy(&node->base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
index fb794e9..bcca883 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
@@ -122,7 +122,7 @@ nvc0_ltcg_init_tag_ram(struct nouveau_fb *pfb, struct nvc0_ltcg_priv *priv)
 		nv_wr32(priv, 0x17e000, priv->part_nr);
 
 	/* tags for 1/4 of VRAM should be enough (8192/4 per GiB of VRAM) */
-	priv->num_tags = (pfb->ram.size >> 17) / 4;
+	priv->num_tags = (pfb->ram->size >> 17) / 4;
 	if (priv->num_tags > (1 << 17))
 		priv->num_tags = 1 << 17; /* we have 17 bits in PTE */
 	priv->num_tags = (priv->num_tags + 63) & ~63; /* round up to 64 */
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
index d796924..0cb322a 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
@@ -35,6 +35,7 @@ nv50_mc_intr[] = {
 	{ 0x00001000, NVDEV_ENGINE_GR },
 	{ 0x00004000, NVDEV_ENGINE_CRYPT },	/* NV84- */
 	{ 0x00008000, NVDEV_ENGINE_BSP },	/* NV84- */
+	{ 0x00020000, NVDEV_ENGINE_VP },	/* NV84- */
 	{ 0x00100000, NVDEV_SUBDEV_TIMER },
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },
 	{ 0x04000000, NVDEV_ENGINE_DISP },
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
index 737bd4b..c5da3ba 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
@@ -33,6 +33,7 @@ nvc0_mc_intr[] = {
 	{ 0x00000001, NVDEV_ENGINE_PPP },
 	{ 0x00000020, NVDEV_ENGINE_COPY0 },
 	{ 0x00000040, NVDEV_ENGINE_COPY1 },
+	{ 0x00000080, NVDEV_ENGINE_COPY2 },
 	{ 0x00000100, NVDEV_ENGINE_FIFO },
 	{ 0x00001000, NVDEV_ENGINE_GR },
 	{ 0x00008000, NVDEV_ENGINE_BSP },
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
index 77c67fc..67fcb6c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
@@ -236,9 +236,9 @@ nouveau_vm_unmap_pgt(struct nouveau_vm *vm, int big, u32 fpde, u32 lpde)
 			vmm->map_pgt(vpgd->obj, pde, vpgt->obj);
 		}
 
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&nv_subdev(vmm)->mutex);
 		nouveau_gpuobj_ref(NULL, &pgt);
-		mutex_lock(&vm->mm.mutex);
+		mutex_lock(&nv_subdev(vmm)->mutex);
 	}
 }
 
@@ -256,18 +256,18 @@ nouveau_vm_map_pgt(struct nouveau_vm *vm, u32 pde, u32 type)
 	pgt_size  = (1 << (vmm->pgt_bits + 12)) >> type;
 	pgt_size *= 8;
 
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 	ret = nouveau_gpuobj_new(nv_object(vm->vmm), NULL, pgt_size, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &pgt);
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	if (unlikely(ret))
 		return ret;
 
 	/* someone beat us to filling the PDE while we didn't have the lock */
 	if (unlikely(vpgt->refcount[big]++)) {
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&nv_subdev(vmm)->mutex);
 		nouveau_gpuobj_ref(NULL, &pgt);
-		mutex_lock(&vm->mm.mutex);
+		mutex_lock(&nv_subdev(vmm)->mutex);
 		return 0;
 	}
 
@@ -289,11 +289,11 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 	u32 fpde, lpde, pde;
 	int ret;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	ret = nouveau_mm_head(&vm->mm, page_shift, msize, msize, align,
 			     &vma->node);
 	if (unlikely(ret != 0)) {
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&nv_subdev(vmm)->mutex);
 		return ret;
 	}
 
@@ -314,13 +314,14 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 			if (pde != fpde)
 				nouveau_vm_unmap_pgt(vm, big, fpde, pde - 1);
 			nouveau_mm_free(&vm->mm, &vma->node);
-			mutex_unlock(&vm->mm.mutex);
+			mutex_unlock(&nv_subdev(vmm)->mutex);
 			return ret;
 		}
 	}
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 
-	vma->vm     = vm;
+	vma->vm = NULL;
+	nouveau_vm_ref(vm, &vma->vm, NULL);
 	vma->offset = (u64)vma->node->offset << 12;
 	vma->access = access;
 	return 0;
@@ -338,10 +339,12 @@ nouveau_vm_put(struct nouveau_vma *vma)
 	fpde = (vma->node->offset >> vmm->pgt_bits);
 	lpde = (vma->node->offset + vma->node->length - 1) >> vmm->pgt_bits;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	nouveau_vm_unmap_pgt(vm, vma->node->type != vmm->spg_shift, fpde, lpde);
 	nouveau_mm_free(&vm->mm, &vma->node);
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
+
+	nouveau_vm_ref(NULL, &vma->vm, NULL);
 }
 
 int
@@ -362,7 +365,7 @@ nouveau_vm_create(struct nouveau_vmmgr *vmm, u64 offset, u64 length,
 	vm->fpde = offset >> (vmm->pgt_bits + 12);
 	vm->lpde = (offset + length - 1) >> (vmm->pgt_bits + 12);
 
-	vm->pgt  = kcalloc(vm->lpde - vm->fpde + 1, sizeof(*vm->pgt), GFP_KERNEL);
+	vm->pgt  = vzalloc((vm->lpde - vm->fpde + 1) * sizeof(*vm->pgt));
 	if (!vm->pgt) {
 		kfree(vm);
 		return -ENOMEM;
@@ -371,7 +374,7 @@ nouveau_vm_create(struct nouveau_vmmgr *vmm, u64 offset, u64 length,
 	ret = nouveau_mm_init(&vm->mm, mm_offset >> 12, mm_length >> 12,
 			      block >> 12);
 	if (ret) {
-		kfree(vm->pgt);
+		vfree(vm->pgt);
 		kfree(vm);
 		return ret;
 	}
@@ -405,24 +408,25 @@ nouveau_vm_link(struct nouveau_vm *vm, struct nouveau_gpuobj *pgd)
 
 	nouveau_gpuobj_ref(pgd, &vpgd->obj);
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	for (i = vm->fpde; i <= vm->lpde; i++)
 		vmm->map_pgt(pgd, i, vm->pgt[i - vm->fpde].obj);
 	list_add(&vpgd->head, &vm->pgd_list);
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 	return 0;
 }
 
 static void
 nouveau_vm_unlink(struct nouveau_vm *vm, struct nouveau_gpuobj *mpgd)
 {
+	struct nouveau_vmmgr *vmm = vm->vmm;
 	struct nouveau_vm_pgd *vpgd, *tmp;
 	struct nouveau_gpuobj *pgd = NULL;
 
 	if (!mpgd)
 		return;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	list_for_each_entry_safe(vpgd, tmp, &vm->pgd_list, head) {
 		if (vpgd->obj == mpgd) {
 			pgd = vpgd->obj;
@@ -431,7 +435,7 @@ nouveau_vm_unlink(struct nouveau_vm *vm, struct nouveau_gpuobj *mpgd)
 			break;
 		}
 	}
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 
 	nouveau_gpuobj_ref(NULL, &pgd);
 }
@@ -446,7 +450,7 @@ nouveau_vm_del(struct nouveau_vm *vm)
 	}
 
 	nouveau_mm_fini(&vm->mm);
-	kfree(vm->pgt);
+	vfree(vm->pgt);
 	kfree(vm);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c
index e067f81..07dd1fe 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c
@@ -27,11 +27,11 @@
 
 #include <subdev/timer.h>
 #include <subdev/fb.h>
+#include <subdev/bar.h>
 #include <subdev/vm.h>
 
 struct nv50_vmmgr_priv {
 	struct nouveau_vmmgr base;
-	spinlock_t lock;
 };
 
 static void
@@ -86,8 +86,8 @@ nv50_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 
 	/* IGPs don't have real VRAM, re-target to stolen system memory */
 	target = 0;
-	if (nouveau_fb(vma->vm->vmm)->ram.stolen) {
-		phys += nouveau_fb(vma->vm->vmm)->ram.stolen;
+	if (nouveau_fb(vma->vm->vmm)->ram->stolen) {
+		phys += nouveau_fb(vma->vm->vmm)->ram->stolen;
 		target = 3;
 	}
 
@@ -151,29 +151,42 @@ nv50_vm_unmap(struct nouveau_gpuobj *pgt, u32 pte, u32 cnt)
 static void
 nv50_vm_flush(struct nouveau_vm *vm)
 {
+	struct nv50_vmmgr_priv *priv = (void *)vm->vmm;
+	struct nouveau_bar *bar = nouveau_bar(priv);
 	struct nouveau_engine *engine;
-	int i;
+	int i, vme;
+
+	bar->flush(bar);
 
+	mutex_lock(&nv_subdev(priv)->mutex);
 	for (i = 0; i < NVDEV_SUBDEV_NR; i++) {
-		if (atomic_read(&vm->engref[i])) {
-			engine = nouveau_engine(vm->vmm, i);
-			if (engine && engine->tlb_flush)
-				engine->tlb_flush(engine);
+		if (!atomic_read(&vm->engref[i]))
+			continue;
+
+		/* unfortunate hw bug workaround... */
+		engine = nouveau_engine(priv, i);
+		if (engine && engine->tlb_flush) {
+			engine->tlb_flush(engine);
+			continue;
 		}
-	}
-}
 
-void
-nv50_vm_flush_engine(struct nouveau_subdev *subdev, int engine)
-{
-	struct nv50_vmmgr_priv *priv = (void *)nouveau_vmmgr(subdev);
-	unsigned long flags;
-
-	spin_lock_irqsave(&priv->lock, flags);
-	nv_wr32(subdev, 0x100c80, (engine << 16) | 1);
-	if (!nv_wait(subdev, 0x100c80, 0x00000001, 0x00000000))
-		nv_error(subdev, "vm flush timeout: engine %d\n", engine);
-	spin_unlock_irqrestore(&priv->lock, flags);
+		switch (i) {
+		case NVDEV_ENGINE_GR   : vme = 0x00; break;
+		case NVDEV_ENGINE_VP   : vme = 0x01; break;
+		case NVDEV_SUBDEV_BAR  : vme = 0x06; break;
+		case NVDEV_ENGINE_MPEG : vme = 0x08; break;
+		case NVDEV_ENGINE_BSP  : vme = 0x09; break;
+		case NVDEV_ENGINE_CRYPT: vme = 0x0a; break;
+		case NVDEV_ENGINE_COPY0: vme = 0x0d; break;
+		default:
+			continue;
+		}
+
+		nv_wr32(priv, 0x100c80, (vme << 16) | 1);
+		if (!nv_wait(priv, 0x100c80, 0x00000001, 0x00000000))
+			nv_error(priv, "vm flush timeout: engine %d\n", vme);
+	}
+	mutex_unlock(&nv_subdev(priv)->mutex);
 }
 
 static int
@@ -211,7 +224,6 @@ nv50_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.map_sg = nv50_vm_map_sg;
 	priv->base.unmap = nv50_vm_unmap;
 	priv->base.flush = nv50_vm_flush;
-	spin_lock_init(&priv->lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
index 4c3b0a2..668cf96 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
@@ -29,10 +29,10 @@
 #include <subdev/fb.h>
 #include <subdev/vm.h>
 #include <subdev/ltcg.h>
+#include <subdev/bar.h>
 
 struct nvc0_vmmgr_priv {
 	struct nouveau_vmmgr base;
-	spinlock_t lock;
 };
 
 
@@ -160,40 +160,40 @@ nvc0_vm_unmap(struct nouveau_gpuobj *pgt, u32 pte, u32 cnt)
 	}
 }
 
-void
-nvc0_vm_flush_engine(struct nouveau_subdev *subdev, u64 addr, int type)
-{
-	struct nvc0_vmmgr_priv *priv = (void *)nouveau_vmmgr(subdev);
-	unsigned long flags;
-
-	/* looks like maybe a "free flush slots" counter, the
-	 * faster you write to 0x100cbc to more it decreases
-	 */
-	spin_lock_irqsave(&priv->lock, flags);
-	if (!nv_wait_ne(subdev, 0x100c80, 0x00ff0000, 0x00000000)) {
-		nv_error(subdev, "vm timeout 0: 0x%08x %d\n",
-			 nv_rd32(subdev, 0x100c80), type);
-	}
-
-	nv_wr32(subdev, 0x100cb8, addr >> 8);
-	nv_wr32(subdev, 0x100cbc, 0x80000000 | type);
-
-	/* wait for flush to be queued? */
-	if (!nv_wait(subdev, 0x100c80, 0x00008000, 0x00008000)) {
-		nv_error(subdev, "vm timeout 1: 0x%08x %d\n",
-			 nv_rd32(subdev, 0x100c80), type);
-	}
-	spin_unlock_irqrestore(&priv->lock, flags);
-}
-
 static void
 nvc0_vm_flush(struct nouveau_vm *vm)
 {
+	struct nvc0_vmmgr_priv *priv = (void *)vm->vmm;
+	struct nouveau_bar *bar = nouveau_bar(priv);
 	struct nouveau_vm_pgd *vpgd;
+	u32 type;
+
+	bar->flush(bar);
+
+	type = 0x00000001; /* PAGE_ALL */
+	if (atomic_read(&vm->engref[NVDEV_SUBDEV_BAR]))
+		type |= 0x00000004; /* HUB_ONLY */
 
+	mutex_lock(&nv_subdev(priv)->mutex);
 	list_for_each_entry(vpgd, &vm->pgd_list, head) {
-		nvc0_vm_flush_engine(nv_subdev(vm->vmm), vpgd->obj->addr, 1);
+		/* looks like maybe a "free flush slots" counter, the
+		 * faster you write to 0x100cbc to more it decreases
+		 */
+		if (!nv_wait_ne(priv, 0x100c80, 0x00ff0000, 0x00000000)) {
+			nv_error(priv, "vm timeout 0: 0x%08x %d\n",
+				 nv_rd32(priv, 0x100c80), type);
+		}
+
+		nv_wr32(priv, 0x100cb8, vpgd->obj->addr >> 8);
+		nv_wr32(priv, 0x100cbc, 0x80000000 | type);
+
+		/* wait for flush to be queued? */
+		if (!nv_wait(priv, 0x100c80, 0x00008000, 0x00008000)) {
+			nv_error(priv, "vm timeout 1: 0x%08x %d\n",
+				 nv_rd32(priv, 0x100c80), type);
+		}
 	}
+	mutex_unlock(&nv_subdev(priv)->mutex);
 }
 
 static int
@@ -227,7 +227,6 @@ nvc0_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.map_sg = nvc0_vm_map_sg;
 	priv->base.unmap = nvc0_vm_unmap;
 	priv->base.flush = nvc0_vm_flush;
-	spin_lock_init(&priv->lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index 1c4c6c9..8f467e7 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -129,6 +129,7 @@ nouveau_abi16_chan_fini(struct nouveau_abi16 *abi16,
 
 	if (chan->ntfy) {
 		nouveau_bo_vma_del(chan->ntfy, &chan->ntfy_vma);
+		nouveau_bo_unpin(chan->ntfy);
 		drm_gem_object_unreference_unlocked(chan->ntfy->gem);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 6aa2137..3e72876 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -1878,9 +1878,6 @@ parse_dcb_table(struct drm_device *dev, struct nvbios *bios)
 	if (dcb->version < 0x21)
 		merge_like_dcb_entries(dev, dcb);
 
-	if (!dcb->entries)
-		return -ENXIO;
-
 	/* dump connector table entries to log, if any exist */
 	idx = -1;
 	while ((conn = olddcb_conn(dev, ++idx))) {
@@ -2054,19 +2051,14 @@ nouveau_bios_posted(struct drm_device *dev)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	unsigned htotal;
 
-	if (nv_device(drm->device)->card_type >= NV_50) {
-		if (NVReadVgaCrtc(dev, 0, 0x00) == 0 &&
-		    NVReadVgaCrtc(dev, 0, 0x1a) == 0)
-			return false;
+	if (nv_device(drm->device)->card_type >= NV_50)
 		return true;
-	}
 
 	htotal  = NVReadVgaCrtc(dev, 0, 0x06);
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x07) & 0x01) << 8;
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x07) & 0x20) << 4;
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x25) & 0x01) << 10;
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x41) & 0x01) << 11;
-
 	return (htotal != 0);
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 7ff1071..4b1afb1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -255,7 +255,7 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct nouveau_fb *pfb = nouveau_fb(drm->device);
-	u32 vram_pages = pfb->ram.size >> PAGE_SHIFT;
+	u32 vram_pages = pfb->ram->size >> PAGE_SHIFT;
 
 	if (nv_device(drm->device)->card_type == NV_10 &&
 	    nvbo->tile_mode && (type & TTM_PL_FLAG_VRAM) &&
@@ -968,7 +968,7 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 		     bool no_wait_gpu, struct ttm_mem_reg *new_mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
-	struct nouveau_channel *chan = chan = drm->channel;
+	struct nouveau_channel *chan = chan = drm->ttm.chan;
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	int ret;
@@ -1052,6 +1052,7 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 			}
 
 			drm->ttm.move = mthd->exec;
+			drm->ttm.chan = chan;
 			name = mthd->name;
 			break;
 		}
@@ -1550,13 +1551,8 @@ void
 nouveau_bo_vma_del(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
 {
 	if (vma->node) {
-		if (nvbo->bo.mem.mem_type != TTM_PL_SYSTEM) {
-			spin_lock(&nvbo->bo.bdev->fence_lock);
-			ttm_bo_wait(&nvbo->bo, false, false, false);
-			spin_unlock(&nvbo->bo.bdev->fence_lock);
+		if (nvbo->bo.mem.mem_type != TTM_PL_SYSTEM)
 			nouveau_vm_unmap(vma);
-		}
-
 		nouveau_vm_put(vma);
 		list_del(&vma->head);
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index eaa80a2..e84f4c3 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -147,7 +147,7 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nouveau_cli *cli,
 		args.limit = client->vm->vmm->limit - 1;
 	} else
 	if (chan->push.buffer->bo.mem.mem_type == TTM_PL_VRAM) {
-		u64 limit = pfb->ram.size - imem->reserved - 1;
+		u64 limit = pfb->ram->size - imem->reserved - 1;
 		if (device->card_type == NV_04) {
 			/* nv04 vram pushbuf hack, retarget to its location in
 			 * the framebuffer bar rather than direct vram access..
@@ -282,7 +282,7 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 		} else {
 			args.flags = NV_DMA_TARGET_VRAM | NV_DMA_ACCESS_RDWR;
 			args.start = 0;
-			args.limit = pfb->ram.size - imem->reserved - 1;
+			args.limit = pfb->ram->size - imem->reserved - 1;
 		}
 
 		ret = nouveau_object_new(nv_object(client), chan->handle, vram,
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index f17dc2a..708b2d1 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -26,6 +26,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/ttm/ttm_execbuf_util.h>
 
 #include "nouveau_fbcon.h"
 #include "dispnv04/hw.h"
@@ -332,10 +333,15 @@ nouveau_display_create(struct drm_device *dev)
 
 	if (nouveau_modeset == 1 ||
 	    (nouveau_modeset < 0 && pclass == PCI_CLASS_DISPLAY_VGA)) {
-		if (nv_device(drm->device)->card_type < NV_50)
-			ret = nv04_display_create(dev);
-		else
-			ret = nv50_display_create(dev);
+		if (drm->vbios.dcb.entries) {
+			if (nv_device(drm->device)->card_type < NV_50)
+				ret = nv04_display_create(dev);
+			else
+				ret = nv50_display_create(dev);
+		} else {
+			ret = 0;
+		}
+
 		if (ret)
 			goto disp_create_err;
 
@@ -457,51 +463,6 @@ nouveau_display_resume(struct drm_device *dev)
 }
 
 static int
-nouveau_page_flip_reserve(struct nouveau_bo *old_bo,
-			  struct nouveau_bo *new_bo)
-{
-	int ret;
-
-	ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM);
-	if (ret)
-		return ret;
-
-	ret = ttm_bo_reserve(&new_bo->bo, false, false, false, 0);
-	if (ret)
-		goto fail;
-
-	if (likely(old_bo != new_bo)) {
-		ret = ttm_bo_reserve(&old_bo->bo, false, false, false, 0);
-		if (ret)
-			goto fail_unreserve;
-	}
-
-	return 0;
-
-fail_unreserve:
-	ttm_bo_unreserve(&new_bo->bo);
-fail:
-	nouveau_bo_unpin(new_bo);
-	return ret;
-}
-
-static void
-nouveau_page_flip_unreserve(struct nouveau_bo *old_bo,
-			    struct nouveau_bo *new_bo,
-			    struct nouveau_fence *fence)
-{
-	nouveau_bo_fence(new_bo, fence);
-	ttm_bo_unreserve(&new_bo->bo);
-
-	if (likely(old_bo != new_bo)) {
-		nouveau_bo_fence(old_bo, fence);
-		ttm_bo_unreserve(&old_bo->bo);
-	}
-
-	nouveau_bo_unpin(old_bo);
-}
-
-static int
 nouveau_page_flip_emit(struct nouveau_channel *chan,
 		       struct nouveau_bo *old_bo,
 		       struct nouveau_bo *new_bo,
@@ -563,6 +524,9 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	struct nouveau_page_flip_state *s;
 	struct nouveau_channel *chan = NULL;
 	struct nouveau_fence *fence;
+	struct list_head res;
+	struct ttm_validate_buffer res_val[2];
+	struct ww_acquire_ctx ticket;
 	int ret;
 
 	if (!drm->channel)
@@ -572,25 +536,43 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	if (!s)
 		return -ENOMEM;
 
-	/* Don't let the buffers go away while we flip */
-	ret = nouveau_page_flip_reserve(old_bo, new_bo);
-	if (ret)
-		goto fail_free;
-
-	/* Initialize a page flip struct */
-	*s = (struct nouveau_page_flip_state)
-		{ { }, event, nouveau_crtc(crtc)->index,
-		  fb->bits_per_pixel, fb->pitches[0], crtc->x, crtc->y,
-		  new_bo->bo.offset };
-
 	/* Choose the channel the flip will be handled in */
+	spin_lock(&old_bo->bo.bdev->fence_lock);
 	fence = new_bo->bo.sync_obj;
 	if (fence)
 		chan = fence->channel;
 	if (!chan)
 		chan = drm->channel;
+	spin_unlock(&old_bo->bo.bdev->fence_lock);
+
 	mutex_lock(&chan->cli->mutex);
 
+	if (new_bo != old_bo) {
+		ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM);
+		if (likely(!ret)) {
+			res_val[0].bo = &old_bo->bo;
+			res_val[1].bo = &new_bo->bo;
+			INIT_LIST_HEAD(&res);
+			list_add_tail(&res_val[0].head, &res);
+			list_add_tail(&res_val[1].head, &res);
+			ret = ttm_eu_reserve_buffers(&ticket, &res);
+			if (ret)
+				nouveau_bo_unpin(new_bo);
+		}
+	} else
+		ret = ttm_bo_reserve(&new_bo->bo, false, false, false, 0);
+
+	if (ret) {
+		mutex_unlock(&chan->cli->mutex);
+		goto fail_free;
+	}
+
+	/* Initialize a page flip struct */
+	*s = (struct nouveau_page_flip_state)
+		{ { }, event, nouveau_crtc(crtc)->index,
+		  fb->bits_per_pixel, fb->pitches[0], crtc->x, crtc->y,
+		  new_bo->bo.offset };
+
 	/* Emit a page flip */
 	if (nv_device(drm->device)->card_type >= NV_50) {
 		ret = nv50_display_flip_next(crtc, fb, chan, 0);
@@ -608,12 +590,22 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	/* Update the crtc struct and cleanup */
 	crtc->fb = fb;
 
-	nouveau_page_flip_unreserve(old_bo, new_bo, fence);
+	if (old_bo != new_bo) {
+		ttm_eu_fence_buffer_objects(&ticket, &res, fence);
+		nouveau_bo_unpin(old_bo);
+	} else {
+		nouveau_bo_fence(new_bo, fence);
+		ttm_bo_unreserve(&new_bo->bo);
+	}
 	nouveau_fence_unref(&fence);
 	return 0;
 
 fail_unreserve:
-	nouveau_page_flip_unreserve(old_bo, new_bo, NULL);
+	if (old_bo != new_bo) {
+		ttm_eu_backoff_reservation(&ticket, &res);
+		nouveau_bo_unpin(new_bo);
+	} else
+		ttm_bo_unreserve(&new_bo->bo);
 fail_free:
 	kfree(s);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 383f4e6..218a4b5 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -702,6 +702,7 @@ driver = {
 	.gem_prime_export = drm_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = nouveau_gem_prime_pin,
+	.gem_prime_unpin = nouveau_gem_prime_unpin,
 	.gem_prime_get_sg_table = nouveau_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = nouveau_gem_prime_import_sg_table,
 	.gem_prime_vmap = nouveau_gem_prime_vmap,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.h b/drivers/gpu/drm/nouveau/nouveau_drm.h
index f2b30f8..41ff7e0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.h
@@ -96,6 +96,7 @@ struct nouveau_drm {
 		int (*move)(struct nouveau_channel *,
 			    struct ttm_buffer_object *,
 			    struct ttm_mem_reg *, struct ttm_mem_reg *);
+		struct nouveau_channel *chan;
 		int mtrr;
 	} ttm;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index b035317..9352010 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -289,16 +289,13 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM);
 	if (ret) {
 		NV_ERROR(drm, "failed to pin fb: %d\n", ret);
-		nouveau_bo_ref(NULL, &nvbo);
-		goto out;
+		goto out_unref;
 	}
 
 	ret = nouveau_bo_map(nvbo);
 	if (ret) {
 		NV_ERROR(drm, "failed to map fb: %d\n", ret);
-		nouveau_bo_unpin(nvbo);
-		nouveau_bo_ref(NULL, &nvbo);
-		goto out;
+		goto out_unpin;
 	}
 
 	chan = nouveau_nofbaccel ? NULL : drm->channel;
@@ -316,13 +313,14 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	info = framebuffer_alloc(0, &pdev->dev);
 	if (!info) {
 		ret = -ENOMEM;
-		goto out_unref;
+		goto out_unlock;
 	}
 
 	ret = fb_alloc_cmap(&info->cmap, 256, 0);
 	if (ret) {
 		ret = -ENOMEM;
-		goto out_unref;
+		framebuffer_release(info);
+		goto out_unlock;
 	}
 
 	info->par = fbcon;
@@ -337,7 +335,7 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	fbcon->helper.fbdev = info;
 
 	strcpy(info->fix.id, "nouveaufb");
-	if (nouveau_nofbaccel)
+	if (!chan)
 		info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_DISABLED;
 	else
 		info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_COPYAREA |
@@ -383,8 +381,14 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	vga_switcheroo_client_fb_set(dev->pdev, info);
 	return 0;
 
-out_unref:
+out_unlock:
 	mutex_unlock(&dev->struct_mutex);
+	if (chan)
+		nouveau_bo_vma_del(nvbo, &fbcon->nouveau_fb.vma);
+out_unpin:
+	nouveau_bo_unpin(nvbo);
+out_unref:
+	nouveau_bo_ref(NULL, &nvbo);
 out:
 	return ret;
 }
@@ -413,6 +417,7 @@ nouveau_fbcon_destroy(struct drm_device *dev, struct nouveau_fbdev *fbcon)
 	if (nouveau_fb->nvbo) {
 		nouveau_bo_unmap(nouveau_fb->nvbo);
 		nouveau_bo_vma_del(nouveau_fb->nvbo, &nouveau_fb->vma);
+		nouveau_bo_unpin(nouveau_fb->nvbo);
 		drm_gem_object_unreference_unlocked(nouveau_fb->nvbo->gem);
 		nouveau_fb->nvbo = NULL;
 	}
@@ -467,10 +472,10 @@ nouveau_fbcon_init(struct drm_device *dev)
 
 	drm_fb_helper_single_add_all_connectors(&fbcon->helper);
 
-	if (pfb->ram.size <= 32 * 1024 * 1024)
+	if (pfb->ram->size <= 32 * 1024 * 1024)
 		preferred_bpp = 8;
 	else
-	if (pfb->ram.size <= 64 * 1024 * 1024)
+	if (pfb->ram->size <= 64 * 1024 * 1024)
 		preferred_bpp = 16;
 	else
 		preferred_bpp = 32;
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 6c94683..1680d91 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -35,15 +35,34 @@
 
 #include <engine/fifo.h>
 
+struct fence_work {
+	struct work_struct base;
+	struct list_head head;
+	void (*func)(void *);
+	void *data;
+};
+
+static void
+nouveau_fence_signal(struct nouveau_fence *fence)
+{
+	struct fence_work *work, *temp;
+
+	list_for_each_entry_safe(work, temp, &fence->work, head) {
+		schedule_work(&work->base);
+		list_del(&work->head);
+	}
+
+	fence->channel = NULL;
+	list_del(&fence->head);
+}
+
 void
 nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 {
 	struct nouveau_fence *fence, *fnext;
 	spin_lock(&fctx->lock);
 	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		fence->channel = NULL;
-		list_del(&fence->head);
-		nouveau_fence_unref(&fence);
+		nouveau_fence_signal(fence);
 	}
 	spin_unlock(&fctx->lock);
 }
@@ -57,6 +76,50 @@ nouveau_fence_context_new(struct nouveau_fence_chan *fctx)
 }
 
 static void
+nouveau_fence_work_handler(struct work_struct *kwork)
+{
+	struct fence_work *work = container_of(kwork, typeof(*work), base);
+	work->func(work->data);
+	kfree(work);
+}
+
+void
+nouveau_fence_work(struct nouveau_fence *fence,
+		   void (*func)(void *), void *data)
+{
+	struct nouveau_channel *chan = fence->channel;
+	struct nouveau_fence_chan *fctx;
+	struct fence_work *work = NULL;
+
+	if (nouveau_fence_done(fence)) {
+		func(data);
+		return;
+	}
+
+	fctx = chan->fence;
+	work = kmalloc(sizeof(*work), GFP_KERNEL);
+	if (!work) {
+		WARN_ON(nouveau_fence_wait(fence, false, false));
+		func(data);
+		return;
+	}
+
+	spin_lock(&fctx->lock);
+	if (!fence->channel) {
+		spin_unlock(&fctx->lock);
+		kfree(work);
+		func(data);
+		return;
+	}
+
+	INIT_WORK(&work->base, nouveau_fence_work_handler);
+	work->func = func;
+	work->data = data;
+	list_add(&work->head, &fence->work);
+	spin_unlock(&fctx->lock);
+}
+
+static void
 nouveau_fence_update(struct nouveau_channel *chan)
 {
 	struct nouveau_fence_chan *fctx = chan->fence;
@@ -67,8 +130,7 @@ nouveau_fence_update(struct nouveau_channel *chan)
 		if (fctx->read(chan) < fence->sequence)
 			break;
 
-		fence->channel = NULL;
-		list_del(&fence->head);
+		nouveau_fence_signal(fence);
 		nouveau_fence_unref(&fence);
 	}
 	spin_unlock(&fctx->lock);
@@ -265,6 +327,7 @@ nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 	if (!fence)
 		return -ENOMEM;
 
+	INIT_LIST_HEAD(&fence->work);
 	fence->sysmem = sysmem;
 	kref_init(&fence->kref);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index c899434..c57bb61 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -5,6 +5,7 @@ struct nouveau_drm;
 
 struct nouveau_fence {
 	struct list_head head;
+	struct list_head work;
 	struct kref kref;
 
 	bool sysmem;
@@ -22,6 +23,7 @@ void nouveau_fence_unref(struct nouveau_fence **);
 
 int  nouveau_fence_emit(struct nouveau_fence *, struct nouveau_channel *);
 bool nouveau_fence_done(struct nouveau_fence *);
+void nouveau_fence_work(struct nouveau_fence *, void (*)(void *), void *);
 int  nouveau_fence_wait(struct nouveau_fence *, bool lazy, bool intr);
 int  nouveau_fence_sync(struct nouveau_fence *, struct nouveau_channel *);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index b4b4d0c..e72d09c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -50,7 +50,8 @@ nouveau_gem_object_del(struct drm_gem_object *gem)
 		return;
 	nvbo->gem = NULL;
 
-	if (unlikely(nvbo->pin_refcnt)) {
+	/* Lockdep hates you for doing reserve with gem object lock held */
+	if (WARN_ON_ONCE(nvbo->pin_refcnt)) {
 		nvbo->pin_refcnt = 1;
 		nouveau_bo_unpin(nvbo);
 	}
@@ -101,6 +102,41 @@ out:
 	return ret;
 }
 
+static void
+nouveau_gem_object_delete(void *data)
+{
+	struct nouveau_vma *vma = data;
+	nouveau_vm_unmap(vma);
+	nouveau_vm_put(vma);
+	kfree(vma);
+}
+
+static void
+nouveau_gem_object_unmap(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
+{
+	const bool mapped = nvbo->bo.mem.mem_type != TTM_PL_SYSTEM;
+	struct nouveau_fence *fence = NULL;
+
+	list_del(&vma->head);
+
+	if (mapped) {
+		spin_lock(&nvbo->bo.bdev->fence_lock);
+		if (nvbo->bo.sync_obj)
+			fence = nouveau_fence_ref(nvbo->bo.sync_obj);
+		spin_unlock(&nvbo->bo.bdev->fence_lock);
+	}
+
+	if (fence) {
+		nouveau_fence_work(fence, nouveau_gem_object_delete, vma);
+	} else {
+		if (mapped)
+			nouveau_vm_unmap(vma);
+		nouveau_vm_put(vma);
+		kfree(vma);
+	}
+	nouveau_fence_unref(&fence);
+}
+
 void
 nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 {
@@ -118,10 +154,8 @@ nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 
 	vma = nouveau_bo_vma_find(nvbo, cli->base.vm);
 	if (vma) {
-		if (--vma->refcount == 0) {
-			nouveau_bo_vma_del(nvbo, vma);
-			kfree(vma);
-		}
+		if (--vma->refcount == 0)
+			nouveau_gem_object_unmap(nvbo, vma);
 	}
 	ttm_bo_unreserve(&nvbo->bo);
 }
@@ -276,10 +310,12 @@ struct validate_op {
 	struct list_head vram_list;
 	struct list_head gart_list;
 	struct list_head both_list;
+	struct ww_acquire_ctx ticket;
 };
 
 static void
-validate_fini_list(struct list_head *list, struct nouveau_fence *fence)
+validate_fini_list(struct list_head *list, struct nouveau_fence *fence,
+		   struct ww_acquire_ctx *ticket)
 {
 	struct list_head *entry, *tmp;
 	struct nouveau_bo *nvbo;
@@ -296,17 +332,24 @@ validate_fini_list(struct list_head *list, struct nouveau_fence *fence)
 
 		list_del(&nvbo->entry);
 		nvbo->reserved_by = NULL;
-		ttm_bo_unreserve(&nvbo->bo);
+		ttm_bo_unreserve_ticket(&nvbo->bo, ticket);
 		drm_gem_object_unreference_unlocked(nvbo->gem);
 	}
 }
 
 static void
-validate_fini(struct validate_op *op, struct nouveau_fence* fence)
+validate_fini_no_ticket(struct validate_op *op, struct nouveau_fence *fence)
 {
-	validate_fini_list(&op->vram_list, fence);
-	validate_fini_list(&op->gart_list, fence);
-	validate_fini_list(&op->both_list, fence);
+	validate_fini_list(&op->vram_list, fence, &op->ticket);
+	validate_fini_list(&op->gart_list, fence, &op->ticket);
+	validate_fini_list(&op->both_list, fence, &op->ticket);
+}
+
+static void
+validate_fini(struct validate_op *op, struct nouveau_fence *fence)
+{
+	validate_fini_no_ticket(op, fence);
+	ww_acquire_fini(&op->ticket);
 }
 
 static int
@@ -316,13 +359,11 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 {
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct drm_device *dev = chan->drm->dev;
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	uint32_t sequence;
 	int trycnt = 0;
 	int ret, i;
 	struct nouveau_bo *res_bo = NULL;
 
-	sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
+	ww_acquire_init(&op->ticket, &reservation_ww_class);
 retry:
 	if (++trycnt > 100000) {
 		NV_ERROR(cli, "%s failed and gave up.\n", __func__);
@@ -337,6 +378,7 @@ retry:
 		gem = drm_gem_object_lookup(dev, file_priv, b->handle);
 		if (!gem) {
 			NV_ERROR(cli, "Unknown handle 0x%08x\n", b->handle);
+			ww_acquire_done(&op->ticket);
 			validate_fini(op, NULL);
 			return -ENOENT;
 		}
@@ -351,21 +393,23 @@ retry:
 			NV_ERROR(cli, "multiple instances of buffer %d on "
 				      "validation list\n", b->handle);
 			drm_gem_object_unreference_unlocked(gem);
+			ww_acquire_done(&op->ticket);
 			validate_fini(op, NULL);
 			return -EINVAL;
 		}
 
-		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, sequence);
+		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, &op->ticket);
 		if (ret) {
-			validate_fini(op, NULL);
-			if (unlikely(ret == -EAGAIN)) {
-				sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
+			validate_fini_no_ticket(op, NULL);
+			if (unlikely(ret == -EDEADLK)) {
 				ret = ttm_bo_reserve_slowpath(&nvbo->bo, true,
-							      sequence);
+							      &op->ticket);
 				if (!ret)
 					res_bo = nvbo;
 			}
 			if (unlikely(ret)) {
+				ww_acquire_done(&op->ticket);
+				ww_acquire_fini(&op->ticket);
 				drm_gem_object_unreference_unlocked(gem);
 				if (ret != -ERESTARTSYS)
 					NV_ERROR(cli, "fail reserve\n");
@@ -389,6 +433,7 @@ retry:
 			NV_ERROR(cli, "invalid valid domains: 0x%08x\n",
 				 b->valid_domains);
 			list_add_tail(&nvbo->entry, &op->both_list);
+			ww_acquire_done(&op->ticket);
 			validate_fini(op, NULL);
 			return -EINVAL;
 		}
@@ -396,6 +441,7 @@ retry:
 			goto retry;
 	}
 
+	ww_acquire_done(&op->ticket);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.h b/drivers/gpu/drm/nouveau/nouveau_gem.h
index 8d7a3f0..502e429 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.h
@@ -36,6 +36,7 @@ extern int nouveau_gem_ioctl_info(struct drm_device *, void *,
 				  struct drm_file *);
 
 extern int nouveau_gem_prime_pin(struct drm_gem_object *);
+extern void nouveau_gem_prime_unpin(struct drm_gem_object *);
 extern struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *);
 extern struct drm_gem_object *nouveau_gem_prime_import_sg_table(
 	struct drm_device *, size_t size, struct sg_table *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index 7e0ff10a..4f6a572 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -125,7 +125,7 @@ nv50_mem_timing_calc(struct drm_device *dev, u32 freq,
 		t->reg[7] = 0x4000202 | (e->tCL - 1) << 16;
 
 		/* XXX: P.version == 1 only has DDR2 and GDDR3? */
-		if (pfb->ram.type == NV_MEM_TYPE_DDR2) {
+		if (pfb->ram->type == NV_MEM_TYPE_DDR2) {
 			t->reg[5] |= (e->tCL + 3) << 8;
 			t->reg[6] |= (t->tCWL - 2) << 8;
 			t->reg[8] |= (e->tCL - 4);
@@ -428,7 +428,7 @@ nouveau_mem_timing_calc(struct drm_device *dev, u32 freq,
 		break;
 	}
 
-	switch (pfb->ram.type * !ret) {
+	switch (pfb->ram->type * !ret) {
 	case NV_MEM_TYPE_GDDR3:
 		ret = nouveau_mem_gddr3_mr(dev, freq, e, len, boot, t);
 		break;
@@ -455,7 +455,7 @@ nouveau_mem_timing_calc(struct drm_device *dev, u32 freq,
 		else
 			dll_off = !!(ramcfg[2] & 0x40);
 
-		switch (pfb->ram.type) {
+		switch (pfb->ram->type) {
 		case NV_MEM_TYPE_GDDR3:
 			t->mr[1] &= ~0x00000040;
 			t->mr[1] |=  0x00000040 * dll_off;
@@ -522,7 +522,7 @@ nouveau_mem_timing_read(struct drm_device *dev, struct nouveau_pm_memtiming *t)
 	t->odt = 0;
 	t->drive_strength = 0;
 
-	switch (pfb->ram.type) {
+	switch (pfb->ram->type) {
 	case NV_MEM_TYPE_DDR3:
 		t->odt |= (t->mr[1] & 0x200) >> 7;
 	case NV_MEM_TYPE_DDR2:
@@ -551,7 +551,7 @@ nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
 	u32 mr[3] = { info->mr[0], info->mr[1], info->mr[2] };
 	u32 mr1_dlloff;
 
-	switch (pfb->ram.type) {
+	switch (pfb->ram->type) {
 	case NV_MEM_TYPE_DDR2:
 		tDLLK = 2000;
 		mr1_dlloff = 0x00000001;
@@ -572,7 +572,7 @@ nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
 	}
 
 	/* fetch current MRs */
-	switch (pfb->ram.type) {
+	switch (pfb->ram->type) {
 	case NV_MEM_TYPE_GDDR3:
 	case NV_MEM_TYPE_DDR3:
 		mr[2] = exec->mrg(exec, 2);
@@ -639,7 +639,7 @@ nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
 		exec->mrs (exec, 0, info->mr[0] | 0x00000000);
 		exec->wait(exec, tMRD);
 		exec->wait(exec, tDLLK);
-		if (pfb->ram.type == NV_MEM_TYPE_GDDR3)
+		if (pfb->ram->type == NV_MEM_TYPE_GDDR3)
 			exec->precharge(exec);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_prime.c b/drivers/gpu/drm/nouveau/nouveau_prime.c
index f53e108..e90468d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_prime.c
+++ b/drivers/gpu/drm/nouveau/nouveau_prime.c
@@ -84,7 +84,7 @@ struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
 int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 {
 	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
-	int ret = 0;
+	int ret;
 
 	/* pin buffer into GTT */
 	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_TT);
@@ -93,3 +93,10 @@ int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 
 	return 0;
 }
+
+void nouveau_gem_prime_unpin(struct drm_gem_object *obj)
+{
+	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
+
+	nouveau_bo_unpin(nvbo);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index f19a15a..19e3757 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -69,7 +69,7 @@ nouveau_vram_manager_del(struct ttm_mem_type_manager *man,
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
 	struct nouveau_fb *pfb = nouveau_fb(drm->device);
 	nouveau_mem_node_cleanup(mem->mm_node);
-	pfb->ram.put(pfb, (struct nouveau_mem **)&mem->mm_node);
+	pfb->ram->put(pfb, (struct nouveau_mem **)&mem->mm_node);
 }
 
 static int
@@ -88,7 +88,7 @@ nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 	if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG)
 		size_nc = 1 << nvbo->page_shift;
 
-	ret = pfb->ram.get(pfb, mem->num_pages << PAGE_SHIFT,
+	ret = pfb->ram->get(pfb, mem->num_pages << PAGE_SHIFT,
 			   mem->page_alignment << PAGE_SHIFT, size_nc,
 			   (nvbo->tile_flags >> 8) & 0x3ff, &node);
 	if (ret) {
@@ -111,7 +111,7 @@ nouveau_vram_manager_debug(struct ttm_mem_type_manager *man, const char *prefix)
 	struct nouveau_mm_node *r;
 	u32 total = 0, free = 0;
 
-	mutex_lock(&mm->mutex);
+	mutex_lock(&nv_subdev(pfb)->mutex);
 	list_for_each_entry(r, &mm->nodes, nl_entry) {
 		printk(KERN_DEBUG "%s %d: 0x%010llx 0x%010llx\n",
 		       prefix, r->type, ((u64)r->offset << 12),
@@ -121,7 +121,7 @@ nouveau_vram_manager_debug(struct ttm_mem_type_manager *man, const char *prefix)
 		if (!r->type)
 			free += r->length;
 	}
-	mutex_unlock(&mm->mutex);
+	mutex_unlock(&nv_subdev(pfb)->mutex);
 
 	printk(KERN_DEBUG "%s  total: 0x%010llx free: 0x%010llx\n",
 	       prefix, (u64)total << 12, (u64)free << 12);
@@ -168,9 +168,6 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct nouveau_mem *node;
 
-	if (unlikely((mem->num_pages << PAGE_SHIFT) >= 512 * 1024 * 1024))
-		return -ENOMEM;
-
 	node = kzalloc(sizeof(*node), GFP_KERNEL);
 	if (!node)
 		return -ENOMEM;
@@ -386,7 +383,7 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 	}
 
 	/* VRAM init */
-	drm->gem.vram_available  = nouveau_fb(drm->device)->ram.size;
+	drm->gem.vram_available  = nouveau_fb(drm->device)->ram->size;
 	drm->gem.vram_available -= nouveau_instmem(drm->device)->reserved;
 
 	ret = ttm_bo_init_mm(&drm->ttm.bdev, TTM_PL_VRAM,
@@ -396,15 +393,12 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 		return ret;
 	}
 
-	drm->ttm.mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 1),
-				     pci_resource_len(dev->pdev, 1),
-				     DRM_MTRR_WC);
+	drm->ttm.mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 1),
+					 pci_resource_len(dev->pdev, 1));
 
 	/* GART init */
 	if (drm->agp.stat != ENABLED) {
 		drm->gem.gart_available = nouveau_vmmgr(drm->device)->limit;
-		if (drm->gem.gart_available > 512 * 1024 * 1024)
-			drm->gem.gart_available = 512 * 1024 * 1024;
 	} else {
 		drm->gem.gart_available = drm->agp.size;
 	}
@@ -433,10 +427,6 @@ nouveau_ttm_fini(struct nouveau_drm *drm)
 
 	nouveau_ttm_global_release(drm);
 
-	if (drm->ttm.mtrr >= 0) {
-		drm_mtrr_del(drm->ttm.mtrr,
-			     pci_resource_start(drm->dev->pdev, 1),
-			     pci_resource_len(drm->dev->pdev, 1), DRM_MTRR_WC);
-		drm->ttm.mtrr = -1;
-	}
+	arch_phys_wc_del(drm->ttm.mtrr);
+	drm->ttm.mtrr = 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index dd5e01f..54dc635 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -159,7 +159,7 @@ nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NV50_DMA_CONF0_ENABLE |
 					         NV50_DMA_CONF0_PART_256,
 				     }, sizeof(struct nv_dma_class), &object);
@@ -172,7 +172,7 @@ nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NV50_DMA_CONF0_ENABLE | 0x70 |
 					         NV50_DMA_CONF0_PART_256,
 				 }, sizeof(struct nv_dma_class), &object);
@@ -185,7 +185,7 @@ nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NV50_DMA_CONF0_ENABLE | 0x7a |
 					         NV50_DMA_CONF0_PART_256,
 				 }, sizeof(struct nv_dma_class), &object);
@@ -204,7 +204,7 @@ nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVC0_DMA_CONF0_ENABLE,
 				     }, sizeof(struct nv_dma_class), &object);
 	if (ret)
@@ -216,7 +216,7 @@ nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVC0_DMA_CONF0_ENABLE | 0xfe,
 				 }, sizeof(struct nv_dma_class), &object);
 	if (ret)
@@ -228,7 +228,7 @@ nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVC0_DMA_CONF0_ENABLE | 0xfe,
 				 }, sizeof(struct nv_dma_class), &object);
 	return ret;
@@ -246,7 +246,7 @@ nvd0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVD0_DMA_CONF0_ENABLE |
 						 NVD0_DMA_CONF0_PAGE_LP,
 				     }, sizeof(struct nv_dma_class), &object);
@@ -259,7 +259,7 @@ nvd0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVD0_DMA_CONF0_ENABLE | 0xfe |
 						 NVD0_DMA_CONF0_PAGE_LP,
 				 }, sizeof(struct nv_dma_class), &object);
@@ -316,7 +316,7 @@ nv50_dmac_create(struct nouveau_object *core, u32 bclass, u8 head,
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 				 }, sizeof(struct nv_dma_class), &object);
 	if (ret)
 		return ret;
diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c
index 69620e3..4efc33f 100644
--- a/drivers/gpu/drm/nouveau/nv50_pm.c
+++ b/drivers/gpu/drm/nouveau/nv50_pm.c
@@ -493,12 +493,12 @@ mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
 	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
 
 	if (mr <= 1) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			hwsq_wr32(hwsq, 0x1002c8 + ((mr - 0) * 4), data);
 		hwsq_wr32(hwsq, 0x1002c0 + ((mr - 0) * 4), data);
 	} else
 	if (mr <= 3) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			hwsq_wr32(hwsq, 0x1002e8 + ((mr - 2) * 4), data);
 		hwsq_wr32(hwsq, 0x1002e0 + ((mr - 2) * 4), data);
 	}
diff --git a/drivers/gpu/drm/nouveau/nva3_pm.c b/drivers/gpu/drm/nouveau/nva3_pm.c
index 863f010..0d0ed59 100644
--- a/drivers/gpu/drm/nouveau/nva3_pm.c
+++ b/drivers/gpu/drm/nouveau/nva3_pm.c
@@ -389,12 +389,12 @@ mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
 	struct nouveau_device *device = nouveau_dev(exec->dev);
 	struct nouveau_fb *pfb = nouveau_fb(device);
 	if (mr <= 1) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			nv_wr32(device, 0x1002c8 + ((mr - 0) * 4), data);
 		nv_wr32(device, 0x1002c0 + ((mr - 0) * 4), data);
 	} else
 	if (mr <= 3) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			nv_wr32(device, 0x1002e8 + ((mr - 2) * 4), data);
 		nv_wr32(device, 0x1002e0 + ((mr - 2) * 4), data);
 	}
diff --git a/drivers/gpu/drm/nouveau/nvc0_pm.c b/drivers/gpu/drm/nouveau/nvc0_pm.c
index 0d34eb5..3b7041c 100644
--- a/drivers/gpu/drm/nouveau/nvc0_pm.c
+++ b/drivers/gpu/drm/nouveau/nvc0_pm.c
@@ -477,7 +477,7 @@ mclk_mrg(struct nouveau_mem_exec_func *exec, int mr)
 {
 	struct nouveau_device *device = nouveau_dev(exec->dev);
 	struct nouveau_fb *pfb = nouveau_fb(device);
-	if (pfb->ram.type != NV_MEM_TYPE_GDDR5) {
+	if (pfb->ram->type != NV_MEM_TYPE_GDDR5) {
 		if (mr <= 1)
 			return nv_rd32(device, 0x10f300 + ((mr - 0) * 4));
 		return nv_rd32(device, 0x10f320 + ((mr - 2) * 4));
@@ -496,15 +496,15 @@ mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
 {
 	struct nouveau_device *device = nouveau_dev(exec->dev);
 	struct nouveau_fb *pfb = nouveau_fb(device);
-	if (pfb->ram.type != NV_MEM_TYPE_GDDR5) {
+	if (pfb->ram->type != NV_MEM_TYPE_GDDR5) {
 		if (mr <= 1) {
 			nv_wr32(device, 0x10f300 + ((mr - 0) * 4), data);
-			if (pfb->ram.ranks > 1)
+			if (pfb->ram->ranks > 1)
 				nv_wr32(device, 0x10f308 + ((mr - 0) * 4), data);
 		} else
 		if (mr <= 3) {
 			nv_wr32(device, 0x10f320 + ((mr - 2) * 4), data);
-			if (pfb->ram.ranks > 1)
+			if (pfb->ram->ranks > 1)
 				nv_wr32(device, 0x10f328 + ((mr - 2) * 4), data);
 		}
 	} else {
diff --git a/drivers/gpu/drm/omapdrm/Kconfig b/drivers/gpu/drm/omapdrm/Kconfig
index 09f65dc..20c41e7 100644
--- a/drivers/gpu/drm/omapdrm/Kconfig
+++ b/drivers/gpu/drm/omapdrm/Kconfig
@@ -1,7 +1,7 @@
 
 config DRM_OMAP
 	tristate "OMAP DRM"
-	depends on DRM && !CONFIG_FB_OMAP2
+	depends on DRM
 	depends on ARCH_OMAP2PLUS || ARCH_MULTIPLATFORM
 	depends on OMAP2_DSS
 	select DRM_KMS_HELPER
diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c
index 79b200a..11a5263 100644
--- a/drivers/gpu/drm/omapdrm/omap_crtc.c
+++ b/drivers/gpu/drm/omapdrm/omap_crtc.c
@@ -40,7 +40,7 @@ struct omap_crtc {
 	 * mgr->id.)  Eventually this will be replaced w/ something
 	 * more common-panel-framework-y
 	 */
-	struct omap_overlay_manager mgr;
+	struct omap_overlay_manager *mgr;
 
 	struct omap_video_timings timings;
 	bool enabled;
@@ -90,7 +90,32 @@ uint32_t pipe2vbl(struct drm_crtc *crtc)
  * job of sequencing the setup of the video pipe in the proper order
  */
 
+/* ovl-mgr-id -> crtc */
+static struct omap_crtc *omap_crtcs[8];
+
 /* we can probably ignore these until we support command-mode panels: */
+static int omap_crtc_connect(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	if (mgr->output)
+		return -EINVAL;
+
+	if ((mgr->supported_outputs & dst->id) == 0)
+		return -EINVAL;
+
+	dst->manager = mgr;
+	mgr->output = dst;
+
+	return 0;
+}
+
+static void omap_crtc_disconnect(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	mgr->output->manager = NULL;
+	mgr->output = NULL;
+}
+
 static void omap_crtc_start_update(struct omap_overlay_manager *mgr)
 {
 }
@@ -107,7 +132,7 @@ static void omap_crtc_disable(struct omap_overlay_manager *mgr)
 static void omap_crtc_set_timings(struct omap_overlay_manager *mgr,
 		const struct omap_video_timings *timings)
 {
-	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
+	struct omap_crtc *omap_crtc = omap_crtcs[mgr->id];
 	DBG("%s", omap_crtc->name);
 	omap_crtc->timings = *timings;
 	omap_crtc->full_update = true;
@@ -116,7 +141,7 @@ static void omap_crtc_set_timings(struct omap_overlay_manager *mgr,
 static void omap_crtc_set_lcd_config(struct omap_overlay_manager *mgr,
 		const struct dss_lcd_mgr_config *config)
 {
-	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
+	struct omap_crtc *omap_crtc = omap_crtcs[mgr->id];
 	DBG("%s", omap_crtc->name);
 	dispc_mgr_set_lcd_config(omap_crtc->channel, config);
 }
@@ -135,6 +160,8 @@ static void omap_crtc_unregister_framedone_handler(
 }
 
 static const struct dss_mgr_ops mgr_ops = {
+		.connect = omap_crtc_connect,
+		.disconnect = omap_crtc_disconnect,
 		.start_update = omap_crtc_start_update,
 		.enable = omap_crtc_enable,
 		.disable = omap_crtc_disable,
@@ -253,10 +280,6 @@ static int omap_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 			NULL, NULL);
 }
 
-static void omap_crtc_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static void vblank_cb(void *arg)
 {
 	struct drm_crtc *crtc = arg;
@@ -366,7 +389,6 @@ static const struct drm_crtc_helper_funcs omap_crtc_helper_funcs = {
 	.prepare = omap_crtc_prepare,
 	.commit = omap_crtc_commit,
 	.mode_set_base = omap_crtc_mode_set_base,
-	.load_lut = omap_crtc_load_lut,
 };
 
 const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc)
@@ -569,7 +591,7 @@ static void omap_crtc_pre_apply(struct omap_drm_apply *apply)
 	} else {
 		if (encoder) {
 			omap_encoder_set_enabled(encoder, false);
-			omap_encoder_update(encoder, &omap_crtc->mgr,
+			omap_encoder_update(encoder, omap_crtc->mgr,
 					&omap_crtc->timings);
 			omap_encoder_set_enabled(encoder, true);
 			omap_crtc->full_update = false;
@@ -595,6 +617,11 @@ static const char *channel_names[] = {
 		[OMAP_DSS_CHANNEL_LCD2] = "lcd2",
 };
 
+void omap_crtc_pre_init(void)
+{
+	dss_install_mgr_ops(&mgr_ops);
+}
+
 /* initialize crtc */
 struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 		struct drm_plane *plane, enum omap_channel channel, int id)
@@ -635,9 +662,7 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 	omap_irq_register(dev, &omap_crtc->error_irq);
 
 	/* temporary: */
-	omap_crtc->mgr.id = channel;
-
-	dss_install_mgr_ops(&mgr_ops);
+	omap_crtc->mgr = omap_dss_get_overlay_manager(channel);
 
 	/* TODO: fix hard-coded setup.. add properties! */
 	info = &omap_crtc->info;
@@ -651,6 +676,8 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 
 	omap_plane_install_properties(omap_crtc->plane, &crtc->base);
 
+	omap_crtcs[channel] = omap_crtc;
+
 	return crtc;
 
 fail:
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c
index 826586f..a3004f1 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.c
+++ b/drivers/gpu/drm/omapdrm/omap_drv.c
@@ -65,10 +65,8 @@ static int get_connector_type(struct omap_dss_device *dssdev)
 	switch (dssdev->type) {
 	case OMAP_DISPLAY_TYPE_HDMI:
 		return DRM_MODE_CONNECTOR_HDMIA;
-	case OMAP_DISPLAY_TYPE_DPI:
-		if (!strcmp(dssdev->name, "dvi"))
-			return DRM_MODE_CONNECTOR_DVID;
-		/* fallthrough */
+	case OMAP_DISPLAY_TYPE_DVI:
+		return DRM_MODE_CONNECTOR_DVID;
 	default:
 		return DRM_MODE_CONNECTOR_Unknown;
 	}
@@ -97,6 +95,9 @@ static int omap_modeset_init(struct drm_device *dev)
 	int num_mgrs = dss_feat_get_num_mgrs();
 	int num_crtcs;
 	int i, id = 0;
+	int r;
+
+	omap_crtc_pre_init();
 
 	drm_mode_config_init(dev);
 
@@ -116,6 +117,7 @@ static int omap_modeset_init(struct drm_device *dev)
 		struct drm_connector *connector;
 		struct drm_encoder *encoder;
 		enum omap_channel channel;
+		struct omap_overlay_manager *mgr;
 
 		if (!dssdev->driver) {
 			dev_warn(dev->dev, "%s has no driver.. skipping it\n",
@@ -131,6 +133,13 @@ static int omap_modeset_init(struct drm_device *dev)
 			continue;
 		}
 
+		r = dssdev->driver->connect(dssdev);
+		if (r) {
+			dev_err(dev->dev, "could not connect display: %s\n",
+					dssdev->name);
+			continue;
+		}
+
 		encoder = omap_encoder_init(dev, dssdev);
 
 		if (!encoder) {
@@ -172,8 +181,9 @@ static int omap_modeset_init(struct drm_device *dev)
 		 * other possible channels to which the encoder can connect are
 		 * not considered.
 		 */
-		channel = dssdev->output->dispc_channel;
 
+		mgr = omapdss_find_mgr_from_display(dssdev);
+		channel = mgr->id;
 		/*
 		 * if this channel hasn't already been taken by a previously
 		 * allocated crtc, we create a new crtc for it
@@ -247,6 +257,9 @@ static int omap_modeset_init(struct drm_device *dev)
 		struct drm_encoder *encoder = priv->encoders[i];
 		struct omap_dss_device *dssdev =
 					omap_encoder_get_dssdev(encoder);
+		struct omap_dss_device *output;
+
+		output = omapdss_find_output_from_display(dssdev);
 
 		/* figure out which crtc's we can connect the encoder to: */
 		encoder->possible_crtcs = 0;
@@ -259,9 +272,11 @@ static int omap_modeset_init(struct drm_device *dev)
 			supported_outputs =
 				dss_feat_get_supported_outputs(crtc_channel);
 
-			if (supported_outputs & dssdev->output->id)
+			if (supported_outputs & output->id)
 				encoder->possible_crtcs |= (1 << id);
 		}
+
+		omap_dss_put_device(output);
 	}
 
 	DBG("registered %d planes, %d crtcs, %d encoders and %d connectors\n",
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h
index 215a20d..14f17da 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.h
+++ b/drivers/gpu/drm/omapdrm/omap_drv.h
@@ -157,6 +157,7 @@ const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc);
 enum omap_channel omap_crtc_channel(struct drm_crtc *crtc);
 int omap_crtc_apply(struct drm_crtc *crtc,
 		struct omap_drm_apply *apply);
+void omap_crtc_pre_init(void);
 struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 		struct drm_plane *plane, enum omap_channel channel, int id);
 
diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
index b11ce60..002988d 100644
--- a/drivers/gpu/drm/omapdrm/omap_fbdev.c
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -281,21 +281,7 @@ fail:
 	return ret;
 }
 
-static void omap_crtc_fb_gamma_set(struct drm_crtc *crtc,
-		u16 red, u16 green, u16 blue, int regno)
-{
-	DBG("fbdev: set gamma");
-}
-
-static void omap_crtc_fb_gamma_get(struct drm_crtc *crtc,
-		u16 *red, u16 *green, u16 *blue, int regno)
-{
-	DBG("fbdev: get gamma");
-}
-
 static struct drm_fb_helper_funcs omap_fb_helper_funcs = {
-	.gamma_set = omap_crtc_fb_gamma_set,
-	.gamma_get = omap_crtc_fb_gamma_get,
 	.fb_probe = omap_fbdev_create,
 };
 
diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
index be7cd97..4fcca8d 100644
--- a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
+++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
@@ -136,44 +136,21 @@ static void omap_gem_dmabuf_kunmap(struct dma_buf *buffer,
 	kunmap(pages[page_num]);
 }
 
-/*
- * TODO maybe we can split up drm_gem_mmap to avoid duplicating
- * some here.. or at least have a drm_dmabuf_mmap helper.
- */
 static int omap_gem_dmabuf_mmap(struct dma_buf *buffer,
 		struct vm_area_struct *vma)
 {
 	struct drm_gem_object *obj = buffer->priv;
+	struct drm_device *dev = obj->dev;
 	int ret = 0;
 
 	if (WARN_ON(!obj->filp))
 		return -EINVAL;
 
-	/* Check for valid size. */
-	if (omap_gem_mmap_size(obj) < vma->vm_end - vma->vm_start) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	if (!obj->dev->driver->gem_vm_ops) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_ops = obj->dev->driver->gem_vm_ops;
-	vma->vm_private_data = obj;
-	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
-
-	/* Take a ref for this mapping of the object, so that the fault
-	 * handler can dereference the mmap offset's pointer to the object.
-	 * This reference is cleaned up by the corresponding vm_close
-	 * (which should happen whether the vma was created by this call, or
-	 * by a vm_open due to mremap or partial unmap or whatever).
-	 */
-	vma->vm_ops->open(vma);
-
-out_unlock:
+	mutex_lock(&dev->struct_mutex);
+	ret = drm_gem_mmap_obj(obj, omap_gem_mmap_size(obj), vma);
+	mutex_unlock(&dev->struct_mutex);
+	if (ret < 0)
+		return ret;
 
 	return omap_gem_mmap_obj(obj, vma);
 }
diff --git a/drivers/gpu/drm/qxl/qxl_cmd.c b/drivers/gpu/drm/qxl/qxl_cmd.c
index f867714..93c2f2c 100644
--- a/drivers/gpu/drm/qxl/qxl_cmd.c
+++ b/drivers/gpu/drm/qxl/qxl_cmd.c
@@ -49,6 +49,11 @@ void qxl_ring_free(struct qxl_ring *ring)
 	kfree(ring);
 }
 
+void qxl_ring_init_hdr(struct qxl_ring *ring)
+{
+	ring->ring->header.notify_on_prod = ring->n_elements;
+}
+
 struct qxl_ring *
 qxl_ring_create(struct qxl_ring_header *header,
 		int element_size,
@@ -69,7 +74,7 @@ qxl_ring_create(struct qxl_ring_header *header,
 	ring->prod_notify = prod_notify;
 	ring->push_event = push_event;
 	if (set_prod_notify)
-		header->notify_on_prod = ring->n_elements;
+		qxl_ring_init_hdr(ring);
 	spin_lock_init(&ring->lock);
 	return ring;
 }
@@ -87,7 +92,7 @@ static int qxl_check_header(struct qxl_ring *ring)
 	return ret;
 }
 
-static int qxl_check_idle(struct qxl_ring *ring)
+int qxl_check_idle(struct qxl_ring *ring)
 {
 	int ret;
 	struct qxl_ring_header *header = &(ring->ring->header);
@@ -375,8 +380,8 @@ void qxl_io_destroy_primary(struct qxl_device *qdev)
 	wait_for_io_cmd(qdev, 0, QXL_IO_DESTROY_PRIMARY_ASYNC);
 }
 
-void qxl_io_create_primary(struct qxl_device *qdev, unsigned width,
-			   unsigned height, unsigned offset, struct qxl_bo *bo)
+void qxl_io_create_primary(struct qxl_device *qdev,
+			   unsigned offset, struct qxl_bo *bo)
 {
 	struct qxl_surface_create *create;
 
@@ -384,8 +389,8 @@ void qxl_io_create_primary(struct qxl_device *qdev, unsigned width,
 		 qdev->ram_header);
 	create = &qdev->ram_header->create_surface;
 	create->format = bo->surf.format;
-	create->width = width;
-	create->height = height;
+	create->width = bo->surf.width;
+	create->height = bo->surf.height;
 	create->stride = bo->surf.stride;
 	create->mem = qxl_bo_physical_address(qdev, bo, offset);
 
diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
index 823d29e..f76f5dd 100644
--- a/drivers/gpu/drm/qxl/qxl_display.c
+++ b/drivers/gpu/drm/qxl/qxl_display.c
@@ -30,53 +30,9 @@
 #include "qxl_object.h"
 #include "drm_crtc_helper.h"
 
-static void qxl_crtc_set_to_mode(struct qxl_device *qdev,
-				 struct drm_connector *connector,
-				 struct qxl_head *head)
+static bool qxl_head_enabled(struct qxl_head *head)
 {
-	struct drm_device *dev = connector->dev;
-	struct drm_display_mode *mode, *t;
-	int width = head->width;
-	int height = head->height;
-
-	if (width < 320 || height < 240) {
-		qxl_io_log(qdev, "%s: bad head: %dx%d", width, height);
-		width = 1024;
-		height = 768;
-	}
-	if (width * height * 4 > 16*1024*1024) {
-		width = 1024;
-		height = 768;
-	}
-	/* TODO: go over regular modes and removed preferred? */
-	list_for_each_entry_safe(mode, t, &connector->probed_modes, head)
-		drm_mode_remove(connector, mode);
-	mode = drm_cvt_mode(dev, width, height, 60, false, false, false);
-	mode->type |= DRM_MODE_TYPE_PREFERRED;
-	mode->status = MODE_OK;
-	drm_mode_probed_add(connector, mode);
-	qxl_io_log(qdev, "%s: %d x %d\n", __func__, width, height);
-}
-
-void qxl_crtc_set_from_monitors_config(struct qxl_device *qdev)
-{
-	struct drm_connector *connector;
-	int i;
-	struct drm_device *dev = qdev->ddev;
-
-	i = 0;
-	qxl_io_log(qdev, "%s: %d, %d\n", __func__,
-		   dev->mode_config.num_connector,
-		   qdev->monitors_config->count);
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (i > qdev->monitors_config->count) {
-			/* crtc will be reported as disabled */
-			continue;
-		}
-		qxl_crtc_set_to_mode(qdev, connector,
-				     &qdev->monitors_config->heads[i]);
-		++i;
-	}
+	return head->width && head->height;
 }
 
 void qxl_alloc_client_monitors_config(struct qxl_device *qdev, unsigned count)
@@ -106,7 +62,6 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 	int num_monitors;
 	uint32_t crc;
 
-	BUG_ON(!qdev->monitors_config);
 	num_monitors = qdev->rom->client_monitors_config.count;
 	crc = crc32(0, (const uint8_t *)&qdev->rom->client_monitors_config,
 		  sizeof(qdev->rom->client_monitors_config));
@@ -117,8 +72,8 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 		return 1;
 	}
 	if (num_monitors > qdev->monitors_config->max_allowed) {
-		DRM_INFO("client monitors list will be truncated: %d < %d\n",
-			 qdev->monitors_config->max_allowed, num_monitors);
+		DRM_DEBUG_KMS("client monitors list will be truncated: %d < %d\n",
+			      qdev->monitors_config->max_allowed, num_monitors);
 		num_monitors = qdev->monitors_config->max_allowed;
 	} else {
 		num_monitors = qdev->rom->client_monitors_config.count;
@@ -132,18 +87,15 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 			&qdev->rom->client_monitors_config.heads[i];
 		struct qxl_head *client_head =
 			&qdev->client_monitors_config->heads[i];
-		struct qxl_head *head = &qdev->monitors_config->heads[i];
-		client_head->x = head->x = c_rect->left;
-		client_head->y = head->y = c_rect->top;
-		client_head->width = head->width =
-						c_rect->right - c_rect->left;
-		client_head->height = head->height =
-						c_rect->bottom - c_rect->top;
-		client_head->surface_id = head->surface_id = 0;
-		client_head->id = head->id = i;
-		client_head->flags = head->flags = 0;
-		QXL_DEBUG(qdev, "read %dx%d+%d+%d\n", head->width, head->height,
-			  head->x, head->y);
+		client_head->x = c_rect->left;
+		client_head->y = c_rect->top;
+		client_head->width = c_rect->right - c_rect->left;
+		client_head->height = c_rect->bottom - c_rect->top;
+		client_head->surface_id = 0;
+		client_head->id = i;
+		client_head->flags = 0;
+		DRM_DEBUG_KMS("read %dx%d+%d+%d\n", client_head->width, client_head->height,
+			  client_head->x, client_head->y);
 	}
 	return 0;
 }
@@ -155,10 +107,7 @@ void qxl_display_read_client_monitors_config(struct qxl_device *qdev)
 		qxl_io_log(qdev, "failed crc check for client_monitors_config,"
 				 " retrying\n");
 	}
-	qxl_crtc_set_from_monitors_config(qdev);
-	/* fire off a uevent and let userspace tell us what to do */
-	qxl_io_log(qdev, "calling drm_sysfs_hotplug_event\n");
-	drm_sysfs_hotplug_event(qdev->ddev);
+	drm_helper_hpd_irq_event(qdev->ddev);
 }
 
 static int qxl_add_monitors_config_modes(struct drm_connector *connector)
@@ -170,9 +119,9 @@ static int qxl_add_monitors_config_modes(struct drm_connector *connector)
 	struct drm_display_mode *mode = NULL;
 	struct qxl_head *head;
 
-	if (!qdev->monitors_config)
+	if (!qdev->client_monitors_config)
 		return 0;
-	head = &qdev->monitors_config->heads[h];
+	head = &qdev->client_monitors_config->heads[h];
 
 	mode = drm_cvt_mode(dev, head->width, head->height, 60, false, false,
 			    false);
@@ -222,12 +171,6 @@ static int qxl_add_common_modes(struct drm_connector *connector)
 	return i - 1;
 }
 
-static void qxl_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
-			       u16 *blue, uint32_t start, uint32_t size)
-{
-	/* TODO */
-}
-
 static void qxl_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct qxl_crtc *qxl_crtc = to_qxl_crtc(crtc);
@@ -255,11 +198,11 @@ qxl_hide_cursor(struct qxl_device *qdev)
 	qxl_release_unreserve(qdev, release);
 }
 
-static int qxl_crtc_cursor_set(struct drm_crtc *crtc,
-			       struct drm_file *file_priv,
-			       uint32_t handle,
-			       uint32_t width,
-			       uint32_t height)
+static int qxl_crtc_cursor_set2(struct drm_crtc *crtc,
+				struct drm_file *file_priv,
+				uint32_t handle,
+				uint32_t width,
+				uint32_t height, int32_t hot_x, int32_t hot_y)
 {
 	struct drm_device *dev = crtc->dev;
 	struct qxl_device *qdev = dev->dev_private;
@@ -315,8 +258,8 @@ static int qxl_crtc_cursor_set(struct drm_crtc *crtc,
 	cursor->header.type = SPICE_CURSOR_TYPE_ALPHA;
 	cursor->header.width = 64;
 	cursor->header.height = 64;
-	cursor->header.hot_spot_x = 0;
-	cursor->header.hot_spot_y = 0;
+	cursor->header.hot_spot_x = hot_x;
+	cursor->header.hot_spot_y = hot_y;
 	cursor->data_size = size;
 	cursor->chunk.next_chunk = 0;
 	cursor->chunk.prev_chunk = 0;
@@ -397,9 +340,8 @@ static int qxl_crtc_cursor_move(struct drm_crtc *crtc,
 
 
 static const struct drm_crtc_funcs qxl_crtc_funcs = {
-	.cursor_set = qxl_crtc_cursor_set,
+	.cursor_set2 = qxl_crtc_cursor_set2,
 	.cursor_move = qxl_crtc_cursor_move,
-	.gamma_set = qxl_crtc_gamma_set,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = qxl_crtc_destroy,
 };
@@ -506,7 +448,7 @@ qxl_send_monitors_config(struct qxl_device *qdev)
 	for (i = 0 ; i < qdev->monitors_config->count ; ++i) {
 		struct qxl_head *head = &qdev->monitors_config->heads[i];
 
-		if (head->y > 8192 || head->y < head->x ||
+		if (head->y > 8192 || head->x > 8192 ||
 		    head->width > 8192 || head->height > 8192) {
 			DRM_ERROR("head %d wrong: %dx%d+%d+%d\n",
 				  i, head->width, head->height,
@@ -517,16 +459,19 @@ qxl_send_monitors_config(struct qxl_device *qdev)
 	qxl_io_monitors_config(qdev);
 }
 
-static void qxl_monitors_config_set_single(struct qxl_device *qdev,
-					   unsigned x, unsigned y,
-					   unsigned width, unsigned height)
+static void qxl_monitors_config_set(struct qxl_device *qdev,
+				    int index,
+				    unsigned x, unsigned y,
+				    unsigned width, unsigned height,
+				    unsigned surf_id)
 {
-	DRM_DEBUG("%dx%d+%d+%d\n", width, height, x, y);
-	qdev->monitors_config->count = 1;
-	qdev->monitors_config->heads[0].x = x;
-	qdev->monitors_config->heads[0].y = y;
-	qdev->monitors_config->heads[0].width = width;
-	qdev->monitors_config->heads[0].height = height;
+	DRM_DEBUG_KMS("%d:%dx%d+%d+%d\n", index, width, height, x, y);
+	qdev->monitors_config->heads[index].x = x;
+	qdev->monitors_config->heads[index].y = y;
+	qdev->monitors_config->heads[index].width = width;
+	qdev->monitors_config->heads[index].height = height;
+	qdev->monitors_config->heads[index].surface_id = surf_id;
+
 }
 
 static int qxl_crtc_mode_set(struct drm_crtc *crtc,
@@ -540,10 +485,11 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	struct qxl_mode *m = (void *)mode->private;
 	struct qxl_framebuffer *qfb;
 	struct qxl_bo *bo, *old_bo = NULL;
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
 	uint32_t width, height, base_offset;
 	bool recreate_primary = false;
 	int ret;
-
+	int surf_id;
 	if (!crtc->fb) {
 		DRM_DEBUG_KMS("No FB bound\n");
 		return 0;
@@ -567,7 +513,8 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 		  adjusted_mode->hdisplay,
 		  adjusted_mode->vdisplay);
 
-	recreate_primary = true;
+	if (qcrtc->index == 0)
+		recreate_primary = true;
 
 	width = mode->hdisplay;
 	height = mode->vdisplay;
@@ -588,8 +535,11 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 			   "recreate primary: %dx%d (was %dx%d,%d,%d)\n",
 			   width, height, bo->surf.width,
 			   bo->surf.height, bo->surf.stride, bo->surf.format);
-		qxl_io_create_primary(qdev, width, height, base_offset, bo);
+		qxl_io_create_primary(qdev, base_offset, bo);
 		bo->is_primary = true;
+		surf_id = 0;
+	} else {
+		surf_id = bo->surface_id;
 	}
 
 	if (old_bo && old_bo != bo) {
@@ -599,11 +549,9 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 		qxl_bo_unreserve(old_bo);
 	}
 
-	if (qdev->monitors_config->count == 0) {
-		qxl_monitors_config_set_single(qdev, x, y,
-					       mode->hdisplay,
-					       mode->vdisplay);
-	}
+	qxl_monitors_config_set(qdev, qcrtc->index, x, y,
+				mode->hdisplay,
+				mode->vdisplay, surf_id);
 	return 0;
 }
 
@@ -619,21 +567,36 @@ static void qxl_crtc_commit(struct drm_crtc *crtc)
 	DRM_DEBUG("\n");
 }
 
-static void qxl_crtc_load_lut(struct drm_crtc *crtc)
+static void qxl_crtc_disable(struct drm_crtc *crtc)
 {
-	DRM_DEBUG("\n");
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	if (crtc->fb) {
+		struct qxl_framebuffer *qfb = to_qxl_framebuffer(crtc->fb);
+		struct qxl_bo *bo = gem_to_qxl_bo(qfb->obj);
+		int ret;
+		ret = qxl_bo_reserve(bo, false);
+		qxl_bo_unpin(bo);
+		qxl_bo_unreserve(bo);
+		crtc->fb = NULL;
+	}
+
+	qxl_monitors_config_set(qdev, qcrtc->index, 0, 0, 0, 0, 0);
+
+	qxl_send_monitors_config(qdev);
 }
 
 static const struct drm_crtc_helper_funcs qxl_crtc_helper_funcs = {
 	.dpms = qxl_crtc_dpms,
+	.disable = qxl_crtc_disable,
 	.mode_fixup = qxl_crtc_mode_fixup,
 	.mode_set = qxl_crtc_mode_set,
 	.prepare = qxl_crtc_prepare,
 	.commit = qxl_crtc_commit,
-	.load_lut = qxl_crtc_load_lut,
 };
 
-static int qdev_crtc_init(struct drm_device *dev, int num_crtc)
+static int qdev_crtc_init(struct drm_device *dev, int crtc_id)
 {
 	struct qxl_crtc *qxl_crtc;
 
@@ -642,7 +605,7 @@ static int qdev_crtc_init(struct drm_device *dev, int num_crtc)
 		return -ENOMEM;
 
 	drm_crtc_init(dev, &qxl_crtc->base, &qxl_crtc_funcs);
-
+	qxl_crtc->index = crtc_id;
 	drm_mode_crtc_set_gamma_size(&qxl_crtc->base, 256);
 	drm_crtc_helper_add(&qxl_crtc->base, &qxl_crtc_helper_funcs);
 	return 0;
@@ -670,18 +633,13 @@ static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
 		struct drm_encoder *encoder)
 {
 	int i;
+	struct qxl_output *output = drm_encoder_to_qxl_output(encoder);
 	struct qxl_head *head;
 	struct drm_display_mode *mode;
 
 	BUG_ON(!encoder);
 	/* TODO: ugly, do better */
-	for (i = 0 ; (encoder->possible_crtcs != (1 << i)) && i < 32; ++i)
-		;
-	if (encoder->possible_crtcs != (1 << i)) {
-		DRM_ERROR("encoder has wrong possible_crtcs: %x\n",
-			  encoder->possible_crtcs);
-		return;
-	}
+	i = output->index;
 	if (!qdev->monitors_config ||
 	    qdev->monitors_config->max_allowed <= i) {
 		DRM_ERROR(
@@ -699,7 +657,6 @@ static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
 		DRM_DEBUG("missing for multiple monitors: no head holes\n");
 	head = &qdev->monitors_config->heads[i];
 	head->id = i;
-	head->surface_id = 0;
 	if (encoder->crtc->enabled) {
 		mode = &encoder->crtc->mode;
 		head->width = mode->hdisplay;
@@ -714,8 +671,8 @@ static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
 		head->x = 0;
 		head->y = 0;
 	}
-	DRM_DEBUG("setting head %d to +%d+%d %dx%d\n",
-		  i, head->x, head->y, head->width, head->height);
+	DRM_DEBUG_KMS("setting head %d to +%d+%d %dx%d out of %d\n",
+		      i, head->x, head->y, head->width, head->height, qdev->monitors_config->count);
 	head->flags = 0;
 	/* TODO - somewhere else to call this for multiple monitors
 	 * (config_commit?) */
@@ -810,8 +767,9 @@ static enum drm_connector_status qxl_conn_detect(
 
 	/* The first monitor is always connected */
 	connected = (output->index == 0) ||
-		    (qdev->monitors_config &&
-		     qdev->monitors_config->count > output->index);
+		    (qdev->client_monitors_config &&
+		     qdev->client_monitors_config->count > output->index &&
+		     qxl_head_enabled(&qdev->client_monitors_config->heads[output->index]));
 
 	DRM_DEBUG("\n");
 	return connected ? connector_status_connected
@@ -875,6 +833,8 @@ static int qdev_output_init(struct drm_device *dev, int num_output)
 	drm_encoder_init(dev, &qxl_output->enc, &qxl_enc_funcs,
 			 DRM_MODE_ENCODER_VIRTUAL);
 
+	/* we get HPD via client monitors config */
+	connector->polled = DRM_CONNECTOR_POLL_HPD;
 	encoder->possible_crtcs = 1 << num_output;
 	drm_mode_connector_attach_encoder(&qxl_output->base,
 					  &qxl_output->enc);
@@ -914,16 +874,14 @@ static const struct drm_mode_config_funcs qxl_mode_funcs = {
 	.fb_create = qxl_user_framebuffer_create,
 };
 
-int qxl_modeset_init(struct qxl_device *qdev)
+int qxl_create_monitors_object(struct qxl_device *qdev)
 {
-	int i;
 	int ret;
 	struct drm_gem_object *gobj;
-	int max_allowed = QXL_NUM_OUTPUTS;
+	int max_allowed = qxl_num_crtc;
 	int monitors_config_size = sizeof(struct qxl_monitors_config) +
-				   max_allowed * sizeof(struct qxl_head);
+		max_allowed * sizeof(struct qxl_head);
 
-	drm_mode_config_init(qdev->ddev);
 	ret = qxl_gem_object_create(qdev, monitors_config_size, 0,
 				    QXL_GEM_DOMAIN_VRAM,
 				    false, false, NULL, &gobj);
@@ -932,13 +890,59 @@ int qxl_modeset_init(struct qxl_device *qdev)
 		return -ENOMEM;
 	}
 	qdev->monitors_config_bo = gem_to_qxl_bo(gobj);
+
+	ret = qxl_bo_reserve(qdev->monitors_config_bo, false);
+	if (ret)
+		return ret;
+
+	ret = qxl_bo_pin(qdev->monitors_config_bo, QXL_GEM_DOMAIN_VRAM, NULL);
+	if (ret) {
+		qxl_bo_unreserve(qdev->monitors_config_bo);
+		return ret;
+	}
+
+	qxl_bo_unreserve(qdev->monitors_config_bo);
+
 	qxl_bo_kmap(qdev->monitors_config_bo, NULL);
+
 	qdev->monitors_config = qdev->monitors_config_bo->kptr;
 	qdev->ram_header->monitors_config =
 		qxl_bo_physical_address(qdev, qdev->monitors_config_bo, 0);
 
 	memset(qdev->monitors_config, 0, monitors_config_size);
 	qdev->monitors_config->max_allowed = max_allowed;
+	return 0;
+}
+
+int qxl_destroy_monitors_object(struct qxl_device *qdev)
+{
+	int ret;
+
+	qdev->monitors_config = NULL;
+	qdev->ram_header->monitors_config = 0;
+
+	qxl_bo_kunmap(qdev->monitors_config_bo);
+	ret = qxl_bo_reserve(qdev->monitors_config_bo, false);
+	if (ret)
+		return ret;
+
+	qxl_bo_unpin(qdev->monitors_config_bo);
+	qxl_bo_unreserve(qdev->monitors_config_bo);
+
+	qxl_bo_unref(&qdev->monitors_config_bo);
+	return 0;
+}
+
+int qxl_modeset_init(struct qxl_device *qdev)
+{
+	int i;
+	int ret;
+
+	drm_mode_config_init(qdev->ddev);
+
+	ret = qxl_create_monitors_object(qdev);
+	if (ret)
+		return ret;
 
 	qdev->ddev->mode_config.funcs = (void *)&qxl_mode_funcs;
 
@@ -949,7 +953,7 @@ int qxl_modeset_init(struct qxl_device *qdev)
 	qdev->ddev->mode_config.max_height = 8192;
 
 	qdev->ddev->mode_config.fb_base = qdev->vram_base;
-	for (i = 0 ; i < QXL_NUM_OUTPUTS; ++i) {
+	for (i = 0 ; i < qxl_num_crtc; ++i) {
 		qdev_crtc_init(qdev->ddev, i);
 		qdev_output_init(qdev->ddev, i);
 	}
@@ -966,6 +970,8 @@ int qxl_modeset_init(struct qxl_device *qdev)
 void qxl_modeset_fini(struct qxl_device *qdev)
 {
 	qxl_fbdev_fini(qdev);
+
+	qxl_destroy_monitors_object(qdev);
 	if (qdev->mode_info.mode_config_initialized) {
 		drm_mode_config_cleanup(qdev->ddev);
 		qdev->mode_info.mode_config_initialized = false;
diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c
index aa291d8..df0b577 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.c
+++ b/drivers/gpu/drm/qxl/qxl_drv.c
@@ -33,8 +33,9 @@
 
 #include "drmP.h"
 #include "drm/drm.h"
-
+#include "drm_crtc_helper.h"
 #include "qxl_drv.h"
+#include "qxl_object.h"
 
 extern int qxl_max_ioctls;
 static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
@@ -47,10 +48,14 @@ static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
 MODULE_DEVICE_TABLE(pci, pciidlist);
 
 static int qxl_modeset = -1;
+int qxl_num_crtc = 4;
 
 MODULE_PARM_DESC(modeset, "Disable/Enable modesetting");
 module_param_named(modeset, qxl_modeset, int, 0400);
 
+MODULE_PARM_DESC(num_heads, "Number of virtual crtcs to expose (default 4)");
+module_param_named(num_heads, qxl_num_crtc, int, 0400);
+
 static struct drm_driver qxl_driver;
 static struct pci_driver qxl_pci_driver;
 
@@ -73,13 +78,6 @@ qxl_pci_remove(struct pci_dev *pdev)
 	drm_put_dev(dev);
 }
 
-static struct pci_driver qxl_pci_driver = {
-	 .name = DRIVER_NAME,
-	 .id_table = pciidlist,
-	 .probe = qxl_pci_probe,
-	 .remove = qxl_pci_remove,
-};
-
 static const struct file_operations qxl_fops = {
 	.owner = THIS_MODULE,
 	.open = drm_open,
@@ -90,6 +88,130 @@ static const struct file_operations qxl_fops = {
 	.mmap = qxl_mmap,
 };
 
+static int qxl_drm_freeze(struct drm_device *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_crtc *crtc;
+
+	drm_kms_helper_poll_disable(dev);
+
+	console_lock();
+	qxl_fbdev_set_suspend(qdev, 1);
+	console_unlock();
+
+	/* unpin the front buffers */
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
+		if (crtc->enabled)
+			(*crtc_funcs->disable)(crtc);
+	}
+
+	qxl_destroy_monitors_object(qdev);
+	qxl_surf_evict(qdev);
+	qxl_vram_evict(qdev);
+
+	while (!qxl_check_idle(qdev->command_ring));
+	while (!qxl_check_idle(qdev->release_ring))
+		qxl_queue_garbage_collect(qdev, 1);
+
+	pci_save_state(pdev);
+
+	return 0;
+}
+
+static int qxl_drm_resume(struct drm_device *dev, bool thaw)
+{
+	struct qxl_device *qdev = dev->dev_private;
+
+	qdev->ram_header->int_mask = QXL_INTERRUPT_MASK;
+	if (!thaw) {
+		qxl_reinit_memslots(qdev);
+		qxl_ring_init_hdr(qdev->release_ring);
+	}
+
+	qxl_create_monitors_object(qdev);
+	drm_helper_resume_force_mode(dev);
+
+	console_lock();
+	qxl_fbdev_set_suspend(qdev, 0);
+	console_unlock();
+
+	drm_kms_helper_poll_enable(dev);
+	return 0;
+}
+
+static int qxl_pm_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	int error;
+
+	error = qxl_drm_freeze(drm_dev);
+	if (error)
+		return error;
+
+	pci_disable_device(pdev);
+	pci_set_power_state(pdev, PCI_D3hot);
+	return 0;
+}
+
+static int qxl_pm_resume(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+
+	pci_set_power_state(pdev, PCI_D0);
+	pci_restore_state(pdev);
+	if (pci_enable_device(pdev)) {
+		return -EIO;
+	}
+
+	return qxl_drm_resume(drm_dev, false);
+}
+
+static int qxl_pm_thaw(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+
+	return qxl_drm_resume(drm_dev, true);
+}
+
+static int qxl_pm_freeze(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+
+	return qxl_drm_freeze(drm_dev);
+}
+
+static int qxl_pm_restore(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct qxl_device *qdev = drm_dev->dev_private;
+
+	qxl_io_reset(qdev);
+	return qxl_drm_resume(drm_dev, false);
+}
+
+static const struct dev_pm_ops qxl_pm_ops = {
+	.suspend = qxl_pm_suspend,
+	.resume = qxl_pm_resume,
+	.freeze = qxl_pm_freeze,
+	.thaw = qxl_pm_thaw,
+	.poweroff = qxl_pm_freeze,
+	.restore = qxl_pm_restore,
+};
+static struct pci_driver qxl_pci_driver = {
+	 .name = DRIVER_NAME,
+	 .id_table = pciidlist,
+	 .probe = qxl_pci_probe,
+	 .remove = qxl_pci_remove,
+	 .driver.pm = &qxl_pm_ops,
+};
+
 static struct drm_driver qxl_driver = {
 	.driver_features = DRIVER_GEM | DRIVER_MODESET |
 			   DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED,
diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h
index 43d06ab..aacb791 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.h
+++ b/drivers/gpu/drm/qxl/qxl_drv.h
@@ -55,11 +55,10 @@
 #define DRIVER_MINOR 1
 #define DRIVER_PATCHLEVEL 0
 
-#define QXL_NUM_OUTPUTS 1
-
 #define QXL_DEBUGFS_MAX_COMPONENTS		32
 
 extern int qxl_log_level;
+extern int qxl_num_crtc;
 
 enum {
 	QXL_INFO_LEVEL = 1,
@@ -139,6 +138,7 @@ struct qxl_reloc_list {
 
 struct qxl_crtc {
 	struct drm_crtc base;
+	int index;
 	int cur_x;
 	int cur_y;
 };
@@ -156,7 +156,7 @@ struct qxl_framebuffer {
 
 #define to_qxl_crtc(x) container_of(x, struct qxl_crtc, base)
 #define drm_connector_to_qxl_output(x) container_of(x, struct qxl_output, base)
-#define drm_encoder_to_qxl_output(x) container_of(x, struct qxl_output, base)
+#define drm_encoder_to_qxl_output(x) container_of(x, struct qxl_output, enc)
 #define to_qxl_framebuffer(x) container_of(x, struct qxl_framebuffer, base)
 
 struct qxl_mman {
@@ -331,6 +331,10 @@ void qxl_modeset_fini(struct qxl_device *qdev);
 int qxl_bo_init(struct qxl_device *qdev);
 void qxl_bo_fini(struct qxl_device *qdev);
 
+void qxl_reinit_memslots(struct qxl_device *qdev);
+int qxl_surf_evict(struct qxl_device *qdev);
+int qxl_vram_evict(struct qxl_device *qdev);
+
 struct qxl_ring *qxl_ring_create(struct qxl_ring_header *header,
 				 int element_size,
 				 int n_elements,
@@ -338,6 +342,8 @@ struct qxl_ring *qxl_ring_create(struct qxl_ring_header *header,
 				 bool set_prod_notify,
 				 wait_queue_head_t *push_event);
 void qxl_ring_free(struct qxl_ring *ring);
+void qxl_ring_init_hdr(struct qxl_ring *ring);
+int qxl_check_idle(struct qxl_ring *ring);
 
 static inline void *
 qxl_fb_virtual_address(struct qxl_device *qdev, unsigned long physical)
@@ -365,6 +371,7 @@ void qxl_fbdev_fini(struct qxl_device *qdev);
 int qxl_get_handle_for_primary_fb(struct qxl_device *qdev,
 				  struct drm_file *file_priv,
 				  uint32_t *handle);
+void qxl_fbdev_set_suspend(struct qxl_device *qdev, int state);
 
 /* qxl_display.c */
 int
@@ -374,6 +381,8 @@ qxl_framebuffer_init(struct drm_device *dev,
 		     struct drm_gem_object *obj);
 void qxl_display_read_client_monitors_config(struct qxl_device *qdev);
 void qxl_send_monitors_config(struct qxl_device *qdev);
+int qxl_create_monitors_object(struct qxl_device *qdev);
+int qxl_destroy_monitors_object(struct qxl_device *qdev);
 
 /* used by qxl_debugfs only */
 void qxl_crtc_set_from_monitors_config(struct qxl_device *qdev);
@@ -435,7 +444,7 @@ void qxl_update_screen(struct qxl_device *qxl);
 /* qxl io operations (qxl_cmd.c) */
 
 void qxl_io_create_primary(struct qxl_device *qdev,
-			   unsigned width, unsigned height, unsigned offset,
+			   unsigned offset,
 			   struct qxl_bo *bo);
 void qxl_io_destroy_primary(struct qxl_device *qdev);
 void qxl_io_memslot_add(struct qxl_device *qdev, uint8_t id);
@@ -528,6 +537,7 @@ irqreturn_t qxl_irq_handler(DRM_IRQ_ARGS);
 
 /* qxl_fb.c */
 int qxl_fb_init(struct qxl_device *qdev);
+bool qxl_fbdev_qobj_is_fb(struct qxl_device *qdev, struct qxl_bo *qobj);
 
 int qxl_debugfs_add_files(struct qxl_device *qdev,
 			  struct drm_info_list *files,
diff --git a/drivers/gpu/drm/qxl/qxl_fb.c b/drivers/gpu/drm/qxl/qxl_fb.c
index b3c5127..76f39d8 100644
--- a/drivers/gpu/drm/qxl/qxl_fb.c
+++ b/drivers/gpu/drm/qxl/qxl_fb.c
@@ -520,10 +520,6 @@ static int qxl_fbdev_destroy(struct drm_device *dev, struct qxl_fbdev *qfbdev)
 }
 
 static struct drm_fb_helper_funcs qxl_fb_helper_funcs = {
-	/* TODO
-	.gamma_set = qxl_crtc_fb_gamma_set,
-	.gamma_get = qxl_crtc_fb_gamma_get,
-	*/
 	.fb_probe = qxl_fb_find_or_create_single,
 };
 
@@ -542,7 +538,7 @@ int qxl_fbdev_init(struct qxl_device *qdev)
 	qfbdev->helper.funcs = &qxl_fb_helper_funcs;
 
 	ret = drm_fb_helper_init(qdev->ddev, &qfbdev->helper,
-				 1 /* num_crtc - QXL supports just 1 */,
+				 qxl_num_crtc /* num_crtc - QXL supports just 1 */,
 				 QXLFB_CONN_LIMIT);
 	if (ret) {
 		kfree(qfbdev);
@@ -564,4 +560,14 @@ void qxl_fbdev_fini(struct qxl_device *qdev)
 	qdev->mode_info.qfbdev = NULL;
 }
 
+void qxl_fbdev_set_suspend(struct qxl_device *qdev, int state)
+{
+	fb_set_suspend(qdev->mode_info.qfbdev->helper.fbdev, state);
+}
 
+bool qxl_fbdev_qobj_is_fb(struct qxl_device *qdev, struct qxl_bo *qobj)
+{
+	if (qobj == gem_to_qxl_bo(qdev->mode_info.qfbdev->qfb.obj))
+		return true;
+	return false;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_ioctl.c b/drivers/gpu/drm/qxl/qxl_ioctl.c
index a30f294..27f45e4 100644
--- a/drivers/gpu/drm/qxl/qxl_ioctl.c
+++ b/drivers/gpu/drm/qxl/qxl_ioctl.c
@@ -188,6 +188,12 @@ static int qxl_execbuffer_ioctl(struct drm_device *dev, void *data,
 		/* TODO copy slow path code from i915 */
 		fb_cmd = qxl_bo_kmap_atomic_page(qdev, cmd_bo, (release->release_offset & PAGE_SIZE));
 		unwritten = __copy_from_user_inatomic_nocache(fb_cmd + sizeof(union qxl_release_info) + (release->release_offset & ~PAGE_SIZE), (void *)(unsigned long)user_cmd.command, user_cmd.command_size);
+
+		{
+			struct qxl_drawable *draw = fb_cmd;
+
+			draw->mm_time = qdev->rom->mm_clock;
+		}
 		qxl_bo_kunmap_atomic_page(qdev, cmd_bo, fb_cmd);
 		if (unwritten) {
 			DRM_ERROR("got unwritten %d\n", unwritten);
diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c
index e27ce2a..9e8da9e 100644
--- a/drivers/gpu/drm/qxl/qxl_kms.c
+++ b/drivers/gpu/drm/qxl/qxl_kms.c
@@ -26,6 +26,7 @@
 #include "qxl_drv.h"
 #include "qxl_object.h"
 
+#include <drm/drm_crtc_helper.h>
 #include <linux/io-mapping.h>
 
 int qxl_log_level;
@@ -72,21 +73,28 @@ static bool qxl_check_device(struct qxl_device *qdev)
 	return true;
 }
 
+static void setup_hw_slot(struct qxl_device *qdev, int slot_index,
+			  struct qxl_memslot *slot)
+{
+	qdev->ram_header->mem_slot.mem_start = slot->start_phys_addr;
+	qdev->ram_header->mem_slot.mem_end = slot->end_phys_addr;
+	qxl_io_memslot_add(qdev, slot_index);
+}
+
 static uint8_t setup_slot(struct qxl_device *qdev, uint8_t slot_index_offset,
 	unsigned long start_phys_addr, unsigned long end_phys_addr)
 {
 	uint64_t high_bits;
 	struct qxl_memslot *slot;
 	uint8_t slot_index;
-	struct qxl_ram_header *ram_header = qdev->ram_header;
 
 	slot_index = qdev->rom->slots_start + slot_index_offset;
 	slot = &qdev->mem_slots[slot_index];
 	slot->start_phys_addr = start_phys_addr;
 	slot->end_phys_addr = end_phys_addr;
-	ram_header->mem_slot.mem_start = slot->start_phys_addr;
-	ram_header->mem_slot.mem_end = slot->end_phys_addr;
-	qxl_io_memslot_add(qdev, slot_index);
+
+	setup_hw_slot(qdev, slot_index, slot);
+
 	slot->generation = qdev->rom->slot_generation;
 	high_bits = slot_index << qdev->slot_gen_bits;
 	high_bits |= slot->generation;
@@ -95,6 +103,12 @@ static uint8_t setup_slot(struct qxl_device *qdev, uint8_t slot_index_offset,
 	return slot_index;
 }
 
+void qxl_reinit_memslots(struct qxl_device *qdev)
+{
+	setup_hw_slot(qdev, qdev->main_mem_slot, &qdev->mem_slots[qdev->main_mem_slot]);
+	setup_hw_slot(qdev, qdev->surfaces_mem_slot, &qdev->mem_slots[qdev->surfaces_mem_slot]);
+}
+
 static void qxl_gc_work(struct work_struct *work)
 {
 	struct qxl_device *qdev = container_of(work, struct qxl_device, gc_work);
@@ -294,6 +308,8 @@ int qxl_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out;
 	}
 
+	drm_kms_helper_poll_init(qdev->ddev);
+
 	return 0;
 out:
 	kfree(qdev);
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index d9b12e7..1191fe7 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -363,3 +363,13 @@ int qxl_bo_list_add(struct qxl_reloc_list *reloc_list, struct qxl_bo *bo)
 		return ret;
 	return 0;
 }
+
+int qxl_surf_evict(struct qxl_device *qdev)
+{
+	return ttm_bo_evict_mm(&qdev->mman.bdev, TTM_PL_PRIV0);
+}
+
+int qxl_vram_evict(struct qxl_device *qdev)
+{
+	return ttm_bo_evict_mm(&qdev->mman.bdev, TTM_PL_VRAM);
+}
diff --git a/drivers/gpu/drm/qxl/qxl_object.h b/drivers/gpu/drm/qxl/qxl_object.h
index b4fd89f..ee7ad79 100644
--- a/drivers/gpu/drm/qxl/qxl_object.h
+++ b/drivers/gpu/drm/qxl/qxl_object.h
@@ -57,11 +57,6 @@ static inline unsigned long qxl_bo_size(struct qxl_bo *bo)
 	return bo->tbo.num_pages << PAGE_SHIFT;
 }
 
-static inline bool qxl_bo_is_reserved(struct qxl_bo *bo)
-{
-	return !!atomic_read(&bo->tbo.reserved);
-}
-
 static inline u64 qxl_bo_mmap_offset(struct qxl_bo *bo)
 {
 	return bo->tbo.addr_space_offset;
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index 86c5e36..c3df52c 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -76,7 +76,10 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
 	atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \
-	si_blit_shaders.o radeon_prime.o radeon_uvd.o
+	si_blit_shaders.o radeon_prime.o radeon_uvd.o cik.o cik_blit_shaders.o \
+	r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \
+	rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \
+	trinity_smc.o ni_dpm.o si_smc.o si_dpm.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o
diff --git a/drivers/gpu/drm/radeon/ObjectID.h b/drivers/gpu/drm/radeon/ObjectID.h
index ca4b038..0619269 100644
--- a/drivers/gpu/drm/radeon/ObjectID.h
+++ b/drivers/gpu/drm/radeon/ObjectID.h
@@ -69,6 +69,8 @@
 #define ENCODER_OBJECT_ID_ALMOND                  0x22
 #define ENCODER_OBJECT_ID_TRAVIS                  0x23
 #define ENCODER_OBJECT_ID_NUTMEG                  0x22
+#define ENCODER_OBJECT_ID_HDMI_ANX9805            0x26
+
 /* Kaleidoscope (KLDSCP) Class Display Hardware (internal) */
 #define ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1   0x13
 #define ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1    0x14
@@ -86,6 +88,8 @@
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY1        0x20
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY2        0x21
 #define ENCODER_OBJECT_ID_INTERNAL_VCE            0x24
+#define ENCODER_OBJECT_ID_INTERNAL_UNIPHY3        0x25
+#define ENCODER_OBJECT_ID_INTERNAL_AMCLK          0x27
 
 #define ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO    0xFF
 
@@ -364,6 +368,14 @@
                                                  GRAPH_OBJECT_ENUM_ID2 << ENUM_ID_SHIFT |\
                                                  ENCODER_OBJECT_ID_INTERNAL_UNIPHY2 << OBJECT_ID_SHIFT)
 
+#define ENCODER_INTERNAL_UNIPHY3_ENUM_ID1         ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
+                                                 ENCODER_OBJECT_ID_INTERNAL_UNIPHY3 << OBJECT_ID_SHIFT)
+
+#define ENCODER_INTERNAL_UNIPHY3_ENUM_ID2         ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID2 << ENUM_ID_SHIFT |\
+                                                 ENCODER_OBJECT_ID_INTERNAL_UNIPHY3 << OBJECT_ID_SHIFT)
+
 #define ENCODER_GENERAL_EXTERNAL_DVO_ENUM_ID1    ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
                                                   GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                   ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO << OBJECT_ID_SHIFT)
@@ -392,6 +404,10 @@
                                                   GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                   ENCODER_OBJECT_ID_INTERNAL_VCE << OBJECT_ID_SHIFT)
 
+#define ENCODER_HDMI_ANX9805_ENUM_ID1            ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
+                                                  ENCODER_OBJECT_ID_HDMI_ANX9805 << OBJECT_ID_SHIFT)
+
 /****************************************************/
 /* Connector Object ID definition - Shared with BIOS */
 /****************************************************/
@@ -461,6 +477,14 @@
                                                  GRAPH_OBJECT_ENUM_ID4 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D << OBJECT_ID_SHIFT)
 
+#define CONNECTOR_SINGLE_LINK_DVI_D_ENUM_ID5   ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID5 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D << OBJECT_ID_SHIFT)
+
+#define CONNECTOR_SINGLE_LINK_DVI_D_ENUM_ID6   ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID6 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D << OBJECT_ID_SHIFT)
+
 #define CONNECTOR_DUAL_LINK_DVI_D_ENUM_ID1     ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D << OBJECT_ID_SHIFT)
@@ -473,6 +497,10 @@
                                                  GRAPH_OBJECT_ENUM_ID3 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D << OBJECT_ID_SHIFT)
 
+#define CONNECTOR_DUAL_LINK_DVI_D_ENUM_ID4     ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID4 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D << OBJECT_ID_SHIFT)
+
 #define CONNECTOR_VGA_ENUM_ID1                 ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_VGA << OBJECT_ID_SHIFT)
@@ -541,6 +569,18 @@
                                                  GRAPH_OBJECT_ENUM_ID3 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
 
+#define CONNECTOR_HDMI_TYPE_A_ENUM_ID4         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID4 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
+
+#define CONNECTOR_HDMI_TYPE_A_ENUM_ID5         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID5 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
+
+#define CONNECTOR_HDMI_TYPE_A_ENUM_ID6         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID6 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
+
 #define CONNECTOR_HDMI_TYPE_B_ENUM_ID1         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_HDMI_TYPE_B << OBJECT_ID_SHIFT)
diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h
index 0ee5737..16b120c 100644
--- a/drivers/gpu/drm/radeon/atombios.h
+++ b/drivers/gpu/drm/radeon/atombios.h
@@ -74,6 +74,8 @@
 #define ATOM_PPLL2            1
 #define ATOM_DCPLL            2
 #define ATOM_PPLL0            2
+#define ATOM_PPLL3            3
+
 #define ATOM_EXT_PLL1         8
 #define ATOM_EXT_PLL2         9
 #define ATOM_EXT_CLOCK        10
@@ -259,7 +261,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT AdjustDisplayPll;											 //Atomic Table,  used by various SW componentes. 
   USHORT AdjustMemoryController;                 //Atomic Table,  indirectly used by various SW components,called from SetMemoryClock                
   USHORT EnableASIC_StaticPwrMgt;                //Atomic Table,  only used by Bios
-  USHORT ASIC_StaticPwrMgtStatusChange;          //Obsolete ,     only used by Bios   
+  USHORT SetUniphyInstance;                      //Atomic Table,  only used by Bios   
   USHORT DAC_LoadDetection;                      //Atomic Table,  directly used by various SW components,latest version 1.2  
   USHORT LVTMAEncoderControl;                    //Atomic Table,directly used by various SW components,latest version 1.3
   USHORT HW_Misc_Operation;                      //Atomic Table,  directly used by various SW components,latest version 1.1 
@@ -271,7 +273,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT TVEncoderControl;                       //Function Table,directly used by various SW components,latest version 1.1
   USHORT PatchMCSetting;                         //only used by BIOS
   USHORT MC_SEQ_Control;                         //only used by BIOS
-  USHORT TV1OutputControl;                       //Atomic Table,  Obsolete from Ry6xx, use DAC2 Output instead
+  USHORT Gfx_Harvesting;                         //Atomic Table,  Obsolete from Ry6xx, Now only used by BIOS for GFX harvesting
   USHORT EnableScaler;                           //Atomic Table,  used only by Bios
   USHORT BlankCRTC;                              //Atomic Table,  directly used by various SW components,latest version 1.1 
   USHORT EnableCRTC;                             //Atomic Table,  directly used by various SW components,latest version 1.1 
@@ -328,7 +330,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
 #define UNIPHYTransmitterControl			     DIG1TransmitterControl
 #define LVTMATransmitterControl				     DIG2TransmitterControl
 #define SetCRTC_DPM_State                        GetConditionalGoldenSetting
-#define SetUniphyInstance                        ASIC_StaticPwrMgtStatusChange
+#define ASIC_StaticPwrMgtStatusChange            SetUniphyInstance 
 #define HPDInterruptService                      ReadHWAssistedI2CStatus
 #define EnableVGA_Access                         GetSCLKOverMCLKRatio
 #define EnableYUV                                GetDispObjectInfo                         
@@ -338,7 +340,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
 #define TMDSAEncoderControl                      PatchMCSetting
 #define LVDSEncoderControl                       MC_SEQ_Control
 #define LCD1OutputControl                        HW_Misc_Operation
-
+#define TV1OutputControl                         Gfx_Harvesting
 
 typedef struct _ATOM_MASTER_COMMAND_TABLE
 {
@@ -478,11 +480,11 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3
 typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4
 {
 #if ATOM_BIG_ENDIAN
-  ULONG  ucPostDiv;          //return parameter: post divider which is used to program to register directly
+  ULONG  ucPostDiv:8;        //return parameter: post divider which is used to program to register directly
   ULONG  ulClock:24;         //Input= target clock, output = actual clock 
 #else
   ULONG  ulClock:24;         //Input= target clock, output = actual clock 
-  ULONG  ucPostDiv;          //return parameter: post divider which is used to program to register directly
+  ULONG  ucPostDiv:8;        //return parameter: post divider which is used to program to register directly
 #endif
 }COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4;
 
@@ -504,6 +506,32 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5
   UCHAR   ucReserved;                       
 }COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5;
 
+
+typedef struct _COMPUTE_GPU_CLOCK_INPUT_PARAMETERS_V1_6
+{
+  ATOM_COMPUTE_CLOCK_FREQ  ulClock;         //Input Parameter
+  ULONG   ulReserved[2];
+}COMPUTE_GPU_CLOCK_INPUT_PARAMETERS_V1_6;
+
+//ATOM_COMPUTE_CLOCK_FREQ.ulComputeClockFlag
+#define COMPUTE_GPUCLK_INPUT_FLAG_CLK_TYPE_MASK            0x0f
+#define COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK           0x00
+#define COMPUTE_GPUCLK_INPUT_FLAG_SCLK                     0x01
+
+typedef struct _COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6
+{
+  COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4  ulClock;         //Output Parameter: ucPostDiv=DFS divider
+  ATOM_S_MPLL_FB_DIVIDER   ulFbDiv;         //Output Parameter: PLL FB divider
+  UCHAR   ucPllRefDiv;                      //Output Parameter: PLL ref divider      
+  UCHAR   ucPllPostDiv;                     //Output Parameter: PLL post divider      
+  UCHAR   ucPllCntlFlag;                    //Output Flags: control flag
+  UCHAR   ucReserved;                       
+}COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6;
+
+//ucPllCntlFlag
+#define SPLL_CNTL_FLAG_VCO_MODE_MASK            0x03 
+
+
 // ucInputFlag
 #define ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN  1   // 1-StrobeMode, 0-PerformanceMode
 
@@ -1686,6 +1714,7 @@ typedef struct _PIXEL_CLOCK_PARAMETERS_V6
 #define PIXEL_CLOCK_V6_MISC_HDMI_30BPP              0x08
 #define PIXEL_CLOCK_V6_MISC_HDMI_48BPP              0x0c
 #define PIXEL_CLOCK_V6_MISC_REF_DIV_SRC             0x10
+#define PIXEL_CLOCK_V6_MISC_GEN_DPREFCLK            0x40
 
 typedef struct _GET_DISP_PLL_STATUS_INPUT_PARAMETERS_V2
 {
@@ -2102,6 +2131,17 @@ typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3
 }DVO_ENCODER_CONTROL_PARAMETERS_V3;
 #define DVO_ENCODER_CONTROL_PS_ALLOCATION_V3	DVO_ENCODER_CONTROL_PARAMETERS_V3
 
+typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V1_4
+{
+  USHORT usPixelClock; 
+  UCHAR  ucDVOConfig;
+  UCHAR  ucAction;														//ATOM_ENABLE/ATOM_DISABLE/ATOM_HPD_INIT
+  UCHAR  ucBitPerColor;                       //please refer to definition of PANEL_xBIT_PER_COLOR
+  UCHAR  ucReseved[3];
+}DVO_ENCODER_CONTROL_PARAMETERS_V1_4;
+#define DVO_ENCODER_CONTROL_PS_ALLOCATION_V1_4	DVO_ENCODER_CONTROL_PARAMETERS_V1_4
+
+
 //ucTableFormatRevision=1
 //ucTableContentRevision=3 structure is not changed but usMisc add bit 1 as another input for 
 // bit1=0: non-coherent mode
@@ -2165,7 +2205,7 @@ typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3
 #define SET_ASIC_VOLTAGE_MODE_SOURCE_B         0x4
 
 #define	SET_ASIC_VOLTAGE_MODE_SET_VOLTAGE      0x0
-#define	SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL      0x1	
+#define	SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL      0x1
 #define	SET_ASIC_VOLTAGE_MODE_GET_GPIOMASK     0x2
 
 typedef struct	_SET_VOLTAGE_PARAMETERS
@@ -2200,15 +2240,20 @@ typedef struct	_SET_VOLTAGE_PARAMETERS_V1_3
 //SET_VOLTAGE_PARAMETERS_V3.ucVoltageMode
 #define ATOM_SET_VOLTAGE                     0        //Set voltage Level
 #define ATOM_INIT_VOLTAGE_REGULATOR          3        //Init Regulator
-#define ATOM_SET_VOLTAGE_PHASE               4        //Set Vregulator Phase
-#define ATOM_GET_MAX_VOLTAGE                 6        //Get Max Voltage, not used in SetVoltageTable v1.3
-#define ATOM_GET_VOLTAGE_LEVEL               6        //Get Voltage level from vitual voltage ID
+#define ATOM_SET_VOLTAGE_PHASE               4        //Set Vregulator Phase, only for SVID/PVID regulator
+#define ATOM_GET_MAX_VOLTAGE                 6        //Get Max Voltage, not used from SetVoltageTable v1.3
+#define ATOM_GET_VOLTAGE_LEVEL               6        //Get Voltage level from vitual voltage ID, not used for SetVoltage v1.4
+#define ATOM_GET_LEAKAGE_ID                  8        //Get Leakage Voltage Id ( starting from SMU7x IP ), SetVoltage v1.4 
 
 // define vitual voltage id in usVoltageLevel
 #define ATOM_VIRTUAL_VOLTAGE_ID0             0xff01
 #define ATOM_VIRTUAL_VOLTAGE_ID1             0xff02
 #define ATOM_VIRTUAL_VOLTAGE_ID2             0xff03
 #define ATOM_VIRTUAL_VOLTAGE_ID3             0xff04
+#define ATOM_VIRTUAL_VOLTAGE_ID4             0xff05
+#define ATOM_VIRTUAL_VOLTAGE_ID5             0xff06
+#define ATOM_VIRTUAL_VOLTAGE_ID6             0xff07
+#define ATOM_VIRTUAL_VOLTAGE_ID7             0xff08
 
 typedef struct _SET_VOLTAGE_PS_ALLOCATION
 {
@@ -2628,7 +2673,8 @@ typedef struct _ATOM_FIRMWARE_INFO_V2_2
   ULONG                           ulFirmwareRevision;
   ULONG                           ulDefaultEngineClock;       //In 10Khz unit
   ULONG                           ulDefaultMemoryClock;       //In 10Khz unit
-  ULONG                           ulReserved[2];
+  ULONG                           ulSPLL_OutputFreq;          //In 10Khz unit  
+  ULONG                           ulGPUPLL_OutputFreq;        //In 10Khz unit
   ULONG                           ulReserved1;                //Was ulMaxEngineClockPLL_Output; //In 10Khz unit*
   ULONG                           ulReserved2;                //Was ulMaxMemoryClockPLL_Output; //In 10Khz unit*
   ULONG                           ulMaxPixelClockPLL_Output;  //In 10Khz unit
@@ -3813,6 +3859,12 @@ typedef struct _ATOM_GPIO_PIN_ASSIGNMENT
   UCHAR                    ucGPIO_ID;
 }ATOM_GPIO_PIN_ASSIGNMENT;
 
+//ucGPIO_ID pre-define id for multiple usage
+//from SMU7.x, if ucGPIO_ID=PP_AC_DC_SWITCH_GPIO_PINID in GPIO_LUTTable, AC/DC swithing feature is enable
+#define PP_AC_DC_SWITCH_GPIO_PINID          60
+//from SMU7.x, if ucGPIO_ID=VDDC_REGULATOR_VRHOT_GPIO_PINID in GPIO_LUTable, VRHot feature is enable
+#define VDDC_VRHOT_GPIO_PINID               61
+
 typedef struct _ATOM_GPIO_PIN_LUT
 {
   ATOM_COMMON_TABLE_HEADER  sHeader;
@@ -4074,17 +4126,19 @@ typedef struct _EXT_DISPLAY_PATH
 
 //usCaps
 #define  EXT_DISPLAY_PATH_CAPS__HBR2_DISABLE          0x01
+#define  EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN        0x02
 
 typedef  struct _ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO
 {
   ATOM_COMMON_TABLE_HEADER sHeader;
   UCHAR                    ucGuid [NUMBER_OF_UCHAR_FOR_GUID];     // a GUID is a 16 byte long string
   EXT_DISPLAY_PATH         sPath[MAX_NUMBER_OF_EXT_DISPLAY_PATH]; // total of fixed 7 entries.
-  UCHAR                    ucChecksum;                            // a  simple Checksum of the sum of whole structure equal to 0x0. 
+  UCHAR                    ucChecksum;                            // a simple Checksum of the sum of whole structure equal to 0x0. 
   UCHAR                    uc3DStereoPinId;                       // use for eDP panel
   UCHAR                    ucRemoteDisplayConfig;
   UCHAR                    uceDPToLVDSRxId;
-  UCHAR                    Reserved[4];                           // for potential expansion
+  UCHAR                    ucFixDPVoltageSwing;                   // usCaps[1]=1, this indicate DP_LANE_SET value
+  UCHAR                    Reserved[3];                           // for potential expansion
 }ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO;
 
 //Related definitions, all records are different but they have a commond header
@@ -4416,6 +4470,13 @@ typedef struct _ATOM_VOLTAGE_CONTROL
 #define	VOLTAGE_CONTROL_ID_CHL822x						0x08									
 #define	VOLTAGE_CONTROL_ID_VT1586M						0x09
 #define VOLTAGE_CONTROL_ID_UP1637 						0x0A
+#define	VOLTAGE_CONTROL_ID_CHL8214            0x0B
+#define	VOLTAGE_CONTROL_ID_UP1801             0x0C
+#define	VOLTAGE_CONTROL_ID_ST6788A            0x0D
+#define VOLTAGE_CONTROL_ID_CHLIR3564SVI2      0x0E
+#define VOLTAGE_CONTROL_ID_AD527x      	      0x0F
+#define VOLTAGE_CONTROL_ID_NCP81022    	      0x10
+#define VOLTAGE_CONTROL_ID_LTC2635			  0x11
 
 typedef struct  _ATOM_VOLTAGE_OBJECT
 {
@@ -4458,6 +4519,15 @@ typedef struct _ATOM_VOLTAGE_OBJECT_HEADER_V3{
 	 USHORT		usSize;													//Size of Object	
 }ATOM_VOLTAGE_OBJECT_HEADER_V3;
 
+// ATOM_VOLTAGE_OBJECT_HEADER_V3.ucVoltageMode
+#define VOLTAGE_OBJ_GPIO_LUT                 0        //VOLTAGE and GPIO Lookup table ->ATOM_GPIO_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_VR_I2C_INIT_SEQ          3        //VOLTAGE REGULATOR INIT sequece through I2C -> ATOM_I2C_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_PHASE_LUT                4        //Set Vregulator Phase lookup table ->ATOM_GPIO_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_SVID2                    7        //Indicate voltage control by SVID2 ->ATOM_SVID2_VOLTAGE_OBJECT_V3
+#define	VOLTAGE_OBJ_PWRBOOST_LEAKAGE_LUT     0x10     //Powerboost Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+#define	VOLTAGE_OBJ_HIGH_STATE_LEAKAGE_LUT   0x11     //High voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_HIGH1_STATE_LEAKAGE_LUT  0x12     //High1 voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+
 typedef struct  _VOLTAGE_LUT_ENTRY_V2
 {
 	 ULONG		ulVoltageId;									  // The Voltage ID which is used to program GPIO register
@@ -4473,7 +4543,7 @@ typedef struct  _LEAKAGE_VOLTAGE_LUT_ENTRY_V2
 
 typedef struct  _ATOM_I2C_VOLTAGE_OBJECT_V3
 {
-   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;    // voltage mode = VOLTAGE_OBJ_VR_I2C_INIT_SEQ
    UCHAR	ucVoltageRegulatorId;					  //Indicate Voltage Regulator Id
    UCHAR    ucVoltageControlI2cLine;
    UCHAR    ucVoltageControlAddress;
@@ -4484,7 +4554,7 @@ typedef struct  _ATOM_I2C_VOLTAGE_OBJECT_V3
 
 typedef struct  _ATOM_GPIO_VOLTAGE_OBJECT_V3
 {
-   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;   
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;   // voltage mode = VOLTAGE_OBJ_GPIO_LUT or VOLTAGE_OBJ_PHASE_LUT
    UCHAR    ucVoltageGpioCntlId;         // default is 0 which indicate control through CG VID mode 
    UCHAR    ucGpioEntryNum;              // indiate the entry numbers of Votlage/Gpio value Look up table
    UCHAR    ucPhaseDelay;                // phase delay in unit of micro second
@@ -4495,7 +4565,7 @@ typedef struct  _ATOM_GPIO_VOLTAGE_OBJECT_V3
 
 typedef struct  _ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
 {
-   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;    // voltage mode = 0x10/0x11/0x12
    UCHAR    ucLeakageCntlId;             // default is 0
    UCHAR    ucLeakageEntryNum;           // indicate the entry number of LeakageId/Voltage Lut table
    UCHAR    ucReserved[2];               
@@ -4503,10 +4573,26 @@ typedef struct  _ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
    LEAKAGE_VOLTAGE_LUT_ENTRY_V2 asLeakageIdLut[1];   
 }ATOM_LEAKAGE_VOLTAGE_OBJECT_V3;
 
+
+typedef struct  _ATOM_SVID2_VOLTAGE_OBJECT_V3
+{
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;    // voltage mode = VOLTAGE_OBJ_SVID2
+// 14:7 – PSI0_VID
+// 6 – PSI0_EN
+// 5 – PSI1
+// 4:2 – load line slope trim. 
+// 1:0 – offset trim, 
+   USHORT   usLoadLine_PSI;    
+// GPU GPIO pin Id to SVID2 regulator VRHot pin. possible value 0~31. 0 means GPIO0, 31 means GPIO31
+   UCHAR    ucReserved[2];
+   ULONG    ulReserved;
+}ATOM_SVID2_VOLTAGE_OBJECT_V3;
+
 typedef union _ATOM_VOLTAGE_OBJECT_V3{
   ATOM_GPIO_VOLTAGE_OBJECT_V3 asGpioVoltageObj;
   ATOM_I2C_VOLTAGE_OBJECT_V3 asI2cVoltageObj;
   ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 asLeakageObj;
+  ATOM_SVID2_VOLTAGE_OBJECT_V3 asSVID2Obj;
 }ATOM_VOLTAGE_OBJECT_V3;
 
 typedef struct  _ATOM_VOLTAGE_OBJECT_INFO_V3_1
@@ -4536,6 +4622,21 @@ typedef struct  _ATOM_ASIC_PROFILING_INFO
 	ATOM_ASIC_PROFILE_VOLTAGE			asVoltage;
 }ATOM_ASIC_PROFILING_INFO;
 
+typedef struct  _ATOM_ASIC_PROFILING_INFO_V2_1
+{
+  ATOM_COMMON_TABLE_HEADER			asHeader; 
+  UCHAR  ucLeakageBinNum;                // indicate the entry number of LeakageId/Voltage Lut table
+  USHORT usLeakageBinArrayOffset;        // offset of USHORT Leakage Bin list array ( from lower LeakageId to higher) 
+
+  UCHAR  ucElbVDDC_Num;               
+  USHORT usElbVDDC_IdArrayOffset;        // offset of USHORT virtual VDDC voltage id ( 0xff01~0xff08 )
+  USHORT usElbVDDC_LevelArrayOffset;     // offset of 2 dimension voltage level USHORT array
+
+  UCHAR  ucElbVDDCI_Num;
+  USHORT usElbVDDCI_IdArrayOffset;       // offset of USHORT virtual VDDCI voltage id ( 0xff01~0xff08 )
+  USHORT usElbVDDCI_LevelArrayOffset;    // offset of 2 dimension voltage level USHORT array
+}ATOM_ASIC_PROFILING_INFO_V2_1;
+
 typedef struct _ATOM_POWER_SOURCE_OBJECT
 {
 	UCHAR	ucPwrSrcId;													// Power source
@@ -4652,6 +4753,8 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V6
 #define SYS_INFO_LVDSMISC__888_BPC                                                   0x04
 #define SYS_INFO_LVDSMISC__OVERRIDE_EN                                               0x08
 #define SYS_INFO_LVDSMISC__BLON_ACTIVE_LOW                                           0x10
+// new since Trinity
+#define SYS_INFO_LVDSMISC__TRAVIS_LVDS_VOL_OVERRIDE_EN                               0x20
 
 // not used any more
 #define SYS_INFO_LVDSMISC__VSYNC_ACTIVE_LOW                                          0x04
@@ -4752,6 +4855,29 @@ typedef struct _ATOM_FUSION_SYSTEM_INFO_V1
   ATOM_INTEGRATED_SYSTEM_INFO_V6    sIntegratedSysInfo;   
   ULONG  ulPowerplayTable[128];  
 }ATOM_FUSION_SYSTEM_INFO_V1; 
+
+
+typedef struct _ATOM_TDP_CONFIG_BITS
+{
+#if ATOM_BIG_ENDIAN
+  ULONG   uReserved:2;
+  ULONG   uTDP_Value:14;  // Original TDP value in tens of milli watts
+  ULONG   uCTDP_Value:14; // Override value in tens of milli watts
+  ULONG   uCTDP_Enable:2; // = (uCTDP_Value > uTDP_Value? 2: (uCTDP_Value < uTDP_Value))
+#else
+  ULONG   uCTDP_Enable:2; // = (uCTDP_Value > uTDP_Value? 2: (uCTDP_Value < uTDP_Value))
+  ULONG   uCTDP_Value:14; // Override value in tens of milli watts
+  ULONG   uTDP_Value:14;  // Original TDP value in tens of milli watts
+  ULONG   uReserved:2;
+#endif
+}ATOM_TDP_CONFIG_BITS;
+
+typedef union _ATOM_TDP_CONFIG
+{
+  ATOM_TDP_CONFIG_BITS TDP_config;
+  ULONG            TDP_config_all;
+}ATOM_TDP_CONFIG;
+
 /**********************************************************************************************************************
   ATOM_FUSION_SYSTEM_INFO_V1 Description
 sIntegratedSysInfo:               refer to ATOM_INTEGRATED_SYSTEM_INFO_V6 definition.
@@ -4784,7 +4910,8 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
   UCHAR  ucMemoryType;  
   UCHAR  ucUMAChannelNumber;
   UCHAR  strVBIOSMsg[40];
-  ULONG  ulReserved[20];
+  ATOM_TDP_CONFIG  asTdpConfig;
+  ULONG  ulReserved[19];
   ATOM_AVAILABLE_SCLK_LIST   sAvail_SCLK[5];
   ULONG  ulGMCRestoreResetTime;
   ULONG  ulMinimumNClk;
@@ -4809,7 +4936,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
   USHORT GnbTdpLimit;
   USHORT usMaxLVDSPclkFreqInSingleLink;
   UCHAR  ucLvdsMisc;
-  UCHAR  ucLVDSReserved;
+  UCHAR  ucTravisLVDSVolAdjust;
   UCHAR  ucLVDSPwrOnSeqDIGONtoDE_in4Ms;
   UCHAR  ucLVDSPwrOnSeqDEtoVARY_BL_in4Ms;
   UCHAR  ucLVDSPwrOffSeqVARY_BLtoDE_in4Ms;
@@ -4817,7 +4944,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
   UCHAR  ucLVDSOffToOnDelay_in4Ms;
   UCHAR  ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms;
   UCHAR  ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms;
-  UCHAR  ucLVDSReserved1;
+  UCHAR  ucMinAllowedBL_Level;
   ULONG  ulLCDBitDepthControlVal;
   ULONG  ulNbpStateMemclkFreq[4];
   USHORT usNBP2Voltage;               
@@ -4846,6 +4973,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
 #define SYS_INFO_GPUCAPS__TMDSHDMI_COHERENT_SINGLEPLL_MODE                0x01
 #define SYS_INFO_GPUCAPS__DP_SINGLEPLL_MODE                               0x02
 #define SYS_INFO_GPUCAPS__DISABLE_AUX_MODE_DETECT                         0x08
+#define SYS_INFO_GPUCAPS__ENABEL_DFS_BYPASS                               0x10
 
 /**********************************************************************************************************************
   ATOM_INTEGRATED_SYSTEM_INFO_V1_7 Description
@@ -4945,6 +5073,9 @@ ucLVDSMisc:                       [bit0] LVDS 888bit panel mode =0: LVDS 888 pan
                                   [bit2] LVDS 888bit per color mode  =0: 666 bit per color =1:888 bit per color
                                   [bit3] LVDS parameter override enable  =0: ucLvdsMisc parameter are not used =1: ucLvdsMisc parameter should be used
                                   [bit4] Polarity of signal sent to digital BLON output pin. =0: not inverted(active high) =1: inverted ( active low )
+                                  [bit5] Travid LVDS output voltage override enable, when =1, use ucTravisLVDSVolAdjust value to overwrite Traivs register LVDS_CTRL_4
+ucTravisLVDSVolAdjust             When ucLVDSMisc[5]=1,it means platform SBIOS want to overwrite TravisLVDSVoltage. Then VBIOS will use ucTravisLVDSVolAdjust 
+                                  value to program Travis register LVDS_CTRL_4
 ucLVDSPwrOnSeqDIGONtoDE_in4Ms:    LVDS power up sequence time in unit of 4ms, time delay from DIGON signal active to data enable signal active( DE ).
                                   =0 mean use VBIOS default which is 8 ( 32ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
@@ -4964,18 +5095,241 @@ ucLVDSOffToOnDelay_in4Ms:         LVDS power down sequence time in unit of 4ms.
                                   =0 means to use VBIOS default delay which is 125 ( 500ms ).
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
 
-ucLVDSPwrOnVARY_BLtoBLON_in4Ms:   LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
+ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms:
+                                  LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
                                   =0 means to use VBIOS default delay which is 0 ( 0ms ).
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
 
-ucLVDSPwrOffBLONtoVARY_BL_in4Ms:  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
+ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms:  
+                                  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
                                   =0 means to use VBIOS default delay which is 0 ( 0ms ).
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
 
+ucMinAllowedBL_Level:             Lowest LCD backlight PWM level. This is customer platform specific parameters. By default it is 0. 
+
 ulNbpStateMemclkFreq[4]:          system memory clock frequncey in unit of 10Khz in different NB pstate. 
 
 **********************************************************************************************************************/
 
+// this IntegrateSystemInfoTable is used for Kaveri & Kabini APU
+typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_8
+{
+  ATOM_COMMON_TABLE_HEADER   sHeader;
+  ULONG  ulBootUpEngineClock;
+  ULONG  ulDentistVCOFreq;
+  ULONG  ulBootUpUMAClock;
+  ATOM_CLK_VOLT_CAPABILITY   sDISPCLK_Voltage[4];
+  ULONG  ulBootUpReqDisplayVector;
+  ULONG  ulVBIOSMisc;
+  ULONG  ulGPUCapInfo;
+  ULONG  ulDISP_CLK2Freq;
+  USHORT usRequestedPWMFreqInHz;
+  UCHAR  ucHtcTmpLmt;
+  UCHAR  ucHtcHystLmt;
+  ULONG  ulReserved2;
+  ULONG  ulSystemConfig;            
+  ULONG  ulCPUCapInfo;
+  ULONG  ulReserved3;
+  USHORT usGPUReservedSysMemSize;
+  USHORT usExtDispConnInfoOffset;
+  USHORT usPanelRefreshRateRange;     
+  UCHAR  ucMemoryType;  
+  UCHAR  ucUMAChannelNumber;
+  UCHAR  strVBIOSMsg[40];
+  ATOM_TDP_CONFIG  asTdpConfig;
+  ULONG  ulReserved[19];
+  ATOM_AVAILABLE_SCLK_LIST   sAvail_SCLK[5];
+  ULONG  ulGMCRestoreResetTime;
+  ULONG  ulReserved4;
+  ULONG  ulIdleNClk;
+  ULONG  ulDDR_DLL_PowerUpTime;
+  ULONG  ulDDR_PLL_PowerUpTime;
+  USHORT usPCIEClkSSPercentage;
+  USHORT usPCIEClkSSType;
+  USHORT usLvdsSSPercentage;
+  USHORT usLvdsSSpreadRateIn10Hz;
+  USHORT usHDMISSPercentage;
+  USHORT usHDMISSpreadRateIn10Hz;
+  USHORT usDVISSPercentage;
+  USHORT usDVISSpreadRateIn10Hz;
+  ULONG  ulGPUReservedSysMemBaseAddrLo;
+  ULONG  ulGPUReservedSysMemBaseAddrHi;
+  ULONG  ulReserved5[3];
+  USHORT usMaxLVDSPclkFreqInSingleLink;
+  UCHAR  ucLvdsMisc;
+  UCHAR  ucTravisLVDSVolAdjust;
+  UCHAR  ucLVDSPwrOnSeqDIGONtoDE_in4Ms;
+  UCHAR  ucLVDSPwrOnSeqDEtoVARY_BL_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqVARY_BLtoDE_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqDEtoDIGON_in4Ms;
+  UCHAR  ucLVDSOffToOnDelay_in4Ms;
+  UCHAR  ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms;
+  UCHAR  ucMinAllowedBL_Level;
+  ULONG  ulLCDBitDepthControlVal;
+  ULONG  ulNbpStateMemclkFreq[4];
+  ULONG  ulReserved6;               
+  ULONG  ulNbpStateNClkFreq[4];
+  USHORT usNBPStateVoltage[4];            
+  USHORT usBootUpNBVoltage;   
+  USHORT usReserved2; 
+  ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO sExtDispConnInfo;
+}ATOM_INTEGRATED_SYSTEM_INFO_V1_8;
+
+/**********************************************************************************************************************
+  ATOM_INTEGRATED_SYSTEM_INFO_V1_8 Description
+ulBootUpEngineClock:              VBIOS bootup Engine clock frequency, in 10kHz unit. if it is equal 0, then VBIOS use pre-defined bootup engine clock
+ulDentistVCOFreq:                 Dentist VCO clock in 10kHz unit. 
+ulBootUpUMAClock:                 System memory boot up clock frequency in 10Khz unit. 
+sDISPCLK_Voltage:                 Report Display clock frequency requirement on GNB voltage(up to 4 voltage levels).
+ 
+ulBootUpReqDisplayVector:         VBIOS boot up display IDs, following are supported devices in Trinity projects:
+                                  ATOM_DEVICE_CRT1_SUPPORT                  0x0001
+                                  ATOM_DEVICE_DFP1_SUPPORT                  0x0008
+                                  ATOM_DEVICE_DFP6_SUPPORT                  0x0040
+                                  ATOM_DEVICE_DFP2_SUPPORT                  0x0080
+                                  ATOM_DEVICE_DFP3_SUPPORT                  0x0200
+                                  ATOM_DEVICE_DFP4_SUPPORT                  0x0400
+                                  ATOM_DEVICE_DFP5_SUPPORT                  0x0800
+                                  ATOM_DEVICE_LCD1_SUPPORT                  0x0002
+
+ulVBIOSMisc:      	              Miscellenous flags for VBIOS requirement and interface 
+                                  bit[0]=0: INT15 callback function Get LCD EDID ( ax=4e08, bl=1b ) is not supported by SBIOS. 
+                                        =1: INT15 callback function Get LCD EDID ( ax=4e08, bl=1b ) is supported by SBIOS.
+                                  bit[1]=0: INT15 callback function Get boot display( ax=4e08, bl=01h) is not supported by SBIOS
+                                        =1: INT15 callback function Get boot display( ax=4e08, bl=01h) is supported by SBIOS
+                                  bit[2]=0: INT15 callback function Get panel Expansion ( ax=4e08, bl=02h) is not supported by SBIOS
+                                        =1: INT15 callback function Get panel Expansion ( ax=4e08, bl=02h) is supported by SBIOS
+                                  bit[3]=0: VBIOS fast boot is disable
+                                        =1: VBIOS fast boot is enable. ( VBIOS skip display device detection in every set mode if LCD panel is connect and LID is open)
+
+ulGPUCapInfo:                     bit[0~2]= Reserved
+                                  bit[3]=0: Enable AUX HW mode detection logic
+                                        =1: Disable AUX HW mode detection logic
+                                  bit[4]=0: Disable DFS bypass feature
+                                        =1: Enable DFS bypass feature
+
+usRequestedPWMFreqInHz:           When it's set to 0x0 by SBIOS: the LCD BackLight is not controlled by GPU(SW). 
+                                  Any attempt to change BL using VBIOS function or enable VariBri from PP table is not effective since ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==0;
+                                  
+                                  When it's set to a non-zero frequency, the BackLight is controlled by GPU (SW) in one of two ways below:
+                                  1. SW uses the GPU BL PWM output to control the BL, in chis case, this non-zero frequency determines what freq GPU should use;
+                                  VBIOS will set up proper PWM frequency and ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==1,as the result,
+                                  Changing BL using VBIOS function is functional in both driver and non-driver present environment; 
+                                  and enabling VariBri under the driver environment from PP table is optional.
+
+                                  2. SW uses other means to control BL (like DPCD),this non-zero frequency serves as a flag only indicating
+                                  that BL control from GPU is expected.
+                                  VBIOS will NOT set up PWM frequency but make ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==1
+                                  Changing BL using VBIOS function could be functional in both driver and non-driver present environment,but
+                                  it's per platform 
+                                  and enabling VariBri under the driver environment from PP table is optional.
+
+ucHtcTmpLmt:                      Refer to D18F3x64 bit[22:16], HtcTmpLmt. Threshold on value to enter HTC_active state.
+ucHtcHystLmt:                     Refer to D18F3x64 bit[27:24], HtcHystLmt. 
+                                  To calculate threshold off value to exit HTC_active state, which is Threshold on vlaue minus ucHtcHystLmt.
+
+ulSystemConfig:                   Bit[0]=0: PCIE Power Gating Disabled 
+                                        =1: PCIE Power Gating Enabled
+                                  Bit[1]=0: DDR-DLL shut-down feature disabled.
+                                         1: DDR-DLL shut-down feature enabled.
+                                  Bit[2]=0: DDR-PLL Power down feature disabled.
+                                         1: DDR-PLL Power down feature enabled. 
+                                  Bit[3]=0: GNB DPM is disabled
+                                        =1: GNB DPM is enabled                                
+ulCPUCapInfo:                     TBD
+
+usExtDispConnInfoOffset:          Offset to sExtDispConnInfo inside the structure
+usPanelRefreshRateRange:          Bit vector for LCD supported refresh rate range. If DRR is requestd by the platform, at least two bits need to be set
+                                  to indicate a range.
+                                  SUPPORTED_LCD_REFRESHRATE_30Hz          0x0004
+                                  SUPPORTED_LCD_REFRESHRATE_40Hz          0x0008
+                                  SUPPORTED_LCD_REFRESHRATE_50Hz          0x0010
+                                  SUPPORTED_LCD_REFRESHRATE_60Hz          0x0020
+
+ucMemoryType:                     [3:0]=1:DDR1;=2:DDR2;=3:DDR3;=5:GDDR5; [7:4] is reserved.
+ucUMAChannelNumber:      	        System memory channel numbers. 
+
+strVBIOSMsg[40]:                  VBIOS boot up customized message string 
+
+sAvail_SCLK[5]:                   Arrays to provide availabe list of SLCK and corresponding voltage, order from low to high  
+
+ulGMCRestoreResetTime:            GMC power restore and GMC reset time to calculate data reconnection latency. Unit in ns. 
+ulIdleNClk:                       NCLK speed while memory runs in self-refresh state, used to calculate self-refresh latency. Unit in 10kHz.
+ulDDR_DLL_PowerUpTime:            DDR PHY DLL power up time. Unit in ns.
+ulDDR_PLL_PowerUpTime:            DDR PHY PLL power up time. Unit in ns.
+
+usPCIEClkSSPercentage:            PCIE Clock Spread Spectrum Percentage in unit 0.01%; 100 mean 1%.
+usPCIEClkSSType:                  PCIE Clock Spread Spectrum Type. 0 for Down spread(default); 1 for Center spread.
+usLvdsSSPercentage:               LVDS panel ( not include eDP ) Spread Spectrum Percentage in unit of 0.01%, =0, use VBIOS default setting. 
+usLvdsSSpreadRateIn10Hz:          LVDS panel ( not include eDP ) Spread Spectrum frequency in unit of 10Hz, =0, use VBIOS default setting. 
+usHDMISSPercentage:               HDMI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
+usHDMISSpreadRateIn10Hz:          HDMI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+usDVISSPercentage:                DVI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
+usDVISSpreadRateIn10Hz:           DVI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+
+usGPUReservedSysMemSize:          Reserved system memory size for ACP engine in APU GNB, units in MB. 0/2/4MB based on CMOS options, current default could be 0MB. KV only, not on KB.
+ulGPUReservedSysMemBaseAddrLo:    Low 32 bits base address to the reserved system memory. 
+ulGPUReservedSysMemBaseAddrHi:    High 32 bits base address to the reserved system memory. 
+
+usMaxLVDSPclkFreqInSingleLink:    Max pixel clock LVDS panel single link, if=0 means VBIOS use default threhold, right now it is 85Mhz
+ucLVDSMisc:                       [bit0] LVDS 888bit panel mode =0: LVDS 888 panel in LDI mode, =1: LVDS 888 panel in FPDI mode
+                                  [bit1] LVDS panel lower and upper link mapping =0: lower link and upper link not swap, =1: lower link and upper link are swapped
+                                  [bit2] LVDS 888bit per color mode  =0: 666 bit per color =1:888 bit per color
+                                  [bit3] LVDS parameter override enable  =0: ucLvdsMisc parameter are not used =1: ucLvdsMisc parameter should be used
+                                  [bit4] Polarity of signal sent to digital BLON output pin. =0: not inverted(active high) =1: inverted ( active low )
+                                  [bit5] Travid LVDS output voltage override enable, when =1, use ucTravisLVDSVolAdjust value to overwrite Traivs register LVDS_CTRL_4
+ucTravisLVDSVolAdjust             When ucLVDSMisc[5]=1,it means platform SBIOS want to overwrite TravisLVDSVoltage. Then VBIOS will use ucTravisLVDSVolAdjust 
+                                  value to program Travis register LVDS_CTRL_4
+ucLVDSPwrOnSeqDIGONtoDE_in4Ms:    
+                                  LVDS power up sequence time in unit of 4ms, time delay from DIGON signal active to data enable signal active( DE ).
+                                  =0 mean use VBIOS default which is 8 ( 32ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOnDEtoVARY_BL_in4Ms:     
+                                  LVDS power up sequence time in unit of 4ms., time delay from DE( data enable ) active to Vary Brightness enable signal active( VARY_BL ).  
+                                  =0 mean use VBIOS default which is 90 ( 360ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOffVARY_BLtoDE_in4Ms:    
+                                  LVDS power down sequence time in unit of 4ms, time delay from data enable ( DE ) signal off to LCDVCC (DIGON) off. 
+                                  =0 mean use VBIOS default delay which is 8 ( 32ms ). The LVDS power down sequence is as following: BLON->VARY_BL->DE->DIGON
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOffDEtoDIGON_in4Ms:      
+                                   LVDS power down sequence time in unit of 4ms, time delay from vary brightness enable signal( VARY_BL) off to data enable ( DE ) signal off. 
+                                  =0 mean use VBIOS default which is 90 ( 360ms ). The LVDS power down sequence is as following: BLON->VARY_BL->DE->DIGON
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSOffToOnDelay_in4Ms:         
+                                  LVDS power down sequence time in unit of 4ms. Time delay from DIGON signal off to DIGON signal active. 
+                                  =0 means to use VBIOS default delay which is 125 ( 500ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms:
+                                  LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
+                                  =0 means to use VBIOS default delay which is 0 ( 0ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms:  
+                                  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
+                                  =0 means to use VBIOS default delay which is 0 ( 0ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucMinAllowedBL_Level:             Lowest LCD backlight PWM level. This is customer platform specific parameters. By default it is 0. 
+
+ulLCDBitDepthControlVal:          GPU display control encoder bit dither control setting, used to program register mmFMT_BIT_DEPTH_CONTROL
+
+ulNbpStateMemclkFreq[4]:          system memory clock frequncey in unit of 10Khz in different NB P-State(P0, P1, P2 & P3).
+ulNbpStateNClkFreq[4]:            NB P-State NClk frequency in different NB P-State
+usNBPStateVoltage[4]:             NB P-State (P0/P1 & P2/P3) voltage; NBP3 refers to lowes voltage
+usBootUpNBVoltage:                NB P-State voltage during boot up before driver loaded 
+sExtDispConnInfo:                 Display connector information table provided to VBIOS
+
+**********************************************************************************************************************/
+
+// this Table is used for Kaveri/Kabini APU
+typedef struct _ATOM_FUSION_SYSTEM_INFO_V2
+{
+  ATOM_INTEGRATED_SYSTEM_INFO_V1_8    sIntegratedSysInfo;       // refer to ATOM_INTEGRATED_SYSTEM_INFO_V1_8 definition
+  ULONG                               ulPowerplayTable[128];    // Update comments here to link new powerplay table definition structure
+}ATOM_FUSION_SYSTEM_INFO_V2; 
+
+
 /**************************************************************************/
 // This portion is only used when ext thermal chip or engine/memory clock SS chip is populated on a design
 //Memory SS Info Table
@@ -5026,22 +5380,24 @@ typedef struct _ATOM_ASIC_SS_ASSIGNMENT
 
 //Define ucClockIndication, SW uses the IDs below to search if the SS is required/enabled on a clock branch/signal type.
 //SS is not required or enabled if a match is not found.
-#define ASIC_INTERNAL_MEMORY_SS			1
-#define ASIC_INTERNAL_ENGINE_SS			2
-#define ASIC_INTERNAL_UVD_SS        3
-#define ASIC_INTERNAL_SS_ON_TMDS    4
-#define ASIC_INTERNAL_SS_ON_HDMI    5
-#define ASIC_INTERNAL_SS_ON_LVDS    6
-#define ASIC_INTERNAL_SS_ON_DP      7
-#define ASIC_INTERNAL_SS_ON_DCPLL   8
-#define ASIC_EXTERNAL_SS_ON_DP_CLOCK 9
-#define ASIC_INTERNAL_VCE_SS        10
+#define ASIC_INTERNAL_MEMORY_SS	         1
+#define ASIC_INTERNAL_ENGINE_SS	         2
+#define ASIC_INTERNAL_UVD_SS             3
+#define ASIC_INTERNAL_SS_ON_TMDS         4
+#define ASIC_INTERNAL_SS_ON_HDMI         5
+#define ASIC_INTERNAL_SS_ON_LVDS         6
+#define ASIC_INTERNAL_SS_ON_DP           7
+#define ASIC_INTERNAL_SS_ON_DCPLL        8
+#define ASIC_EXTERNAL_SS_ON_DP_CLOCK     9
+#define ASIC_INTERNAL_VCE_SS             10
+#define ASIC_INTERNAL_GPUPLL_SS          11
+
 
 typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V2
 {
 	ULONG								ulTargetClockRange;						//For mem/engine/uvd, Clock Out frequence (VCO ), in unit of 10Khz
                                                     //For TMDS/HDMI/LVDS, it is pixel clock , for DP, it is link clock ( 27000 or 16200 )
-  USHORT              usSpreadSpectrumPercentage;		//in unit of 0.01%
+  USHORT              usSpreadSpectrumPercentage;		//in unit of 0.01% or 0.001%, decided by ucSpreadSpectrumMode bit4
 	USHORT							usSpreadRateIn10Hz;						//in unit of 10Hz, modulation freq
   UCHAR               ucClockIndication;					  //Indicate which clock source needs SS
 	UCHAR								ucSpreadSpectrumMode;					//Bit0=0 Down Spread,=1 Center Spread, bit1=0: internal SS bit1=1: external SS
@@ -5079,6 +5435,11 @@ typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V3
 	UCHAR								ucReserved[2];
 }ATOM_ASIC_SS_ASSIGNMENT_V3;
 
+//ATOM_ASIC_SS_ASSIGNMENT_V3.ucSpreadSpectrumMode
+#define SS_MODE_V3_CENTRE_SPREAD_MASK             0x01
+#define SS_MODE_V3_EXTERNAL_SS_MASK               0x02
+#define SS_MODE_V3_PERCENTAGE_DIV_BY_1000_MASK    0x10
+
 typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V3
 {
   ATOM_COMMON_TABLE_HEADER	      sHeader; 
@@ -5719,6 +6080,7 @@ typedef struct _INDIRECT_IO_ACCESS
 #define INDIRECT_IO_PCIE           3
 #define INDIRECT_IO_PCIEP          4
 #define INDIRECT_IO_NBMISC         5
+#define INDIRECT_IO_SMU            5
 
 #define INDIRECT_IO_PLL_READ       INDIRECT_IO_PLL   | INDIRECT_READ
 #define INDIRECT_IO_PLL_WRITE      INDIRECT_IO_PLL   | INDIRECT_WRITE
@@ -5730,6 +6092,8 @@ typedef struct _INDIRECT_IO_ACCESS
 #define INDIRECT_IO_PCIEP_WRITE    INDIRECT_IO_PCIEP | INDIRECT_WRITE
 #define INDIRECT_IO_NBMISC_READ    INDIRECT_IO_NBMISC | INDIRECT_READ
 #define INDIRECT_IO_NBMISC_WRITE   INDIRECT_IO_NBMISC | INDIRECT_WRITE
+#define INDIRECT_IO_SMU_READ       INDIRECT_IO_SMU | INDIRECT_READ
+#define INDIRECT_IO_SMU_WRITE      INDIRECT_IO_SMU | INDIRECT_WRITE
 
 typedef struct _ATOM_OEM_INFO
 { 
@@ -5875,6 +6239,7 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE
 #define _64Mx32             0x43
 #define _128Mx8             0x51
 #define _128Mx16            0x52
+#define _128Mx32            0x53
 #define _256Mx8             0x61
 #define _256Mx16            0x62
 
@@ -5893,6 +6258,8 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE
 #define PROMOS              MOSEL
 #define KRETON              INFINEON
 #define ELIXIR              NANYA
+#define MEZZA               ELPIDA
+
 
 /////////////Support for GDDR5 MC uCode to reside in upper 64K of ROM/////////////
 
@@ -6625,6 +6992,10 @@ typedef struct _ATOM_DISP_OUT_INFO_V3
 	ASIC_TRANSMITTER_INFO_V2  asTransmitterInfo[1];     // for alligment only
 }ATOM_DISP_OUT_INFO_V3;
 
+//ucDispCaps
+#define DISPLAY_CAPS__DP_PCLK_FROM_PPLL        0x01
+#define DISPLAY_CAPS__FORCE_DISPDEV_CONNECTED  0x02
+
 typedef enum CORE_REF_CLK_SOURCE{
   CLOCK_SRC_XTALIN=0,
   CLOCK_SRC_XO_IN=1,
@@ -6829,6 +7200,17 @@ typedef struct _DIG_TRANSMITTER_INFO_HEADER_V3_1{
   USHORT usPhyPllSettingOffset;          // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy Pll Settings
 }DIG_TRANSMITTER_INFO_HEADER_V3_1;
 
+typedef struct _DIG_TRANSMITTER_INFO_HEADER_V3_2{  
+  ATOM_COMMON_TABLE_HEADER sHeader;  
+  USHORT usDPVsPreEmphSettingOffset;     // offset of PHY_ANALOG_SETTING_INFO * with DP Voltage Swing and Pre-Emphasis for each Link clock 
+  USHORT usPhyAnalogRegListOffset;       // offset of CLOCK_CONDITION_REGESTER_INFO* with None-DP mode Analog Setting's register Info 
+  USHORT usPhyAnalogSettingOffset;       // offset of CLOCK_CONDITION_SETTING_ENTRY* with None-DP mode Analog Setting for each link clock range
+  USHORT usPhyPllRegListOffset;          // offset of CLOCK_CONDITION_REGESTER_INFO* with Phy Pll register Info 
+  USHORT usPhyPllSettingOffset;          // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy Pll Settings
+  USHORT usDPSSRegListOffset;            // offset of CLOCK_CONDITION_REGESTER_INFO* with Phy SS Pll register Info 
+  USHORT usDPSSSettingOffset;            // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy SS Pll Settings
+}DIG_TRANSMITTER_INFO_HEADER_V3_2;
+
 typedef struct _CLOCK_CONDITION_REGESTER_INFO{
   USHORT usRegisterIndex;
   UCHAR  ucStartBit;
@@ -6852,12 +7234,24 @@ typedef struct _PHY_CONDITION_REG_VAL{
   ULONG  ulRegVal;
 }PHY_CONDITION_REG_VAL;
 
+typedef struct _PHY_CONDITION_REG_VAL_V2{
+  ULONG  ulCondition;
+  UCHAR  ucCondition2;
+  ULONG  ulRegVal;
+}PHY_CONDITION_REG_VAL_V2;
+
 typedef struct _PHY_CONDITION_REG_INFO{
   USHORT usRegIndex;
   USHORT usSize;
   PHY_CONDITION_REG_VAL asRegVal[1];
 }PHY_CONDITION_REG_INFO;
 
+typedef struct _PHY_CONDITION_REG_INFO_V2{
+  USHORT usRegIndex;
+  USHORT usSize;
+  PHY_CONDITION_REG_VAL_V2 asRegVal[1];
+}PHY_CONDITION_REG_INFO_V2;
+
 typedef struct _PHY_ANALOG_SETTING_INFO{
   UCHAR  ucEncodeMode;
   UCHAR  ucPhySel;
@@ -6865,6 +7259,25 @@ typedef struct _PHY_ANALOG_SETTING_INFO{
   PHY_CONDITION_REG_INFO  asAnalogSetting[1];
 }PHY_ANALOG_SETTING_INFO;
 
+typedef struct _PHY_ANALOG_SETTING_INFO_V2{
+  UCHAR  ucEncodeMode;
+  UCHAR  ucPhySel;
+  USHORT usSize;
+  PHY_CONDITION_REG_INFO_V2  asAnalogSetting[1];
+}PHY_ANALOG_SETTING_INFO_V2;
+
+typedef struct _GFX_HAVESTING_PARAMETERS {
+  UCHAR ucGfxBlkId;                        //GFX blk id to be harvested, like CU, RB or PRIM
+  UCHAR ucReserved;                        //reserved 
+  UCHAR ucActiveUnitNumPerSH;              //requested active CU/RB/PRIM number per shader array
+  UCHAR ucMaxUnitNumPerSH;                 //max CU/RB/PRIM number per shader array   
+} GFX_HAVESTING_PARAMETERS;
+
+//ucGfxBlkId
+#define GFX_HARVESTING_CU_ID               0
+#define GFX_HARVESTING_RB_ID               1
+#define GFX_HARVESTING_PRIM_ID             2
+
 /****************************************************************************/	
 //Portion VI: Definitinos for vbios MC scratch registers that driver used
 /****************************************************************************/
@@ -6875,8 +7288,17 @@ typedef struct _PHY_ANALOG_SETTING_INFO{
 #define MC_MISC0__MEMORY_TYPE__GDDR3  0x30000000
 #define MC_MISC0__MEMORY_TYPE__GDDR4  0x40000000
 #define MC_MISC0__MEMORY_TYPE__GDDR5  0x50000000
+#define MC_MISC0__MEMORY_TYPE__HBM    0x60000000
 #define MC_MISC0__MEMORY_TYPE__DDR3   0xB0000000
 
+#define ATOM_MEM_TYPE_DDR_STRING      "DDR"
+#define ATOM_MEM_TYPE_DDR2_STRING     "DDR2"
+#define ATOM_MEM_TYPE_GDDR3_STRING    "GDDR3"
+#define ATOM_MEM_TYPE_GDDR4_STRING    "GDDR4"
+#define ATOM_MEM_TYPE_GDDR5_STRING    "GDDR5"
+#define ATOM_MEM_TYPE_HBM_STRING      "HBM"
+#define ATOM_MEM_TYPE_DDR3_STRING     "DDR3"
+
 /****************************************************************************/	
 //Portion VI: Definitinos being oboselete
 /****************************************************************************/
@@ -7274,6 +7696,7 @@ typedef struct _ATOM_PPLIB_THERMALCONTROLLER
 #define ATOM_PP_THERMALCONTROLLER_NISLANDS  15
 #define ATOM_PP_THERMALCONTROLLER_SISLANDS  16
 #define ATOM_PP_THERMALCONTROLLER_LM96163   17
+#define ATOM_PP_THERMALCONTROLLER_CISLANDS  18
 
 // Thermal controller 'combo type' to use an external controller for Fan control and an internal controller for thermal.
 // We probably should reserve the bit 0x80 for this use.
@@ -7316,6 +7739,8 @@ typedef struct _ATOM_PPLIB_EXTENDEDHEADER
     // Add extra system parameters here, always adjust size to include all fields.
     USHORT  usVCETableOffset; //points to ATOM_PPLIB_VCE_Table
     USHORT  usUVDTableOffset;   //points to ATOM_PPLIB_UVD_Table
+    USHORT  usSAMUTableOffset;  //points to ATOM_PPLIB_SAMU_Table
+    USHORT  usPPMTableOffset;   //points to ATOM_PPLIB_PPM_Table
 } ATOM_PPLIB_EXTENDEDHEADER;
 
 //// ATOM_PPLIB_POWERPLAYTABLE::ulPlatformCaps
@@ -7337,7 +7762,10 @@ typedef struct _ATOM_PPLIB_EXTENDEDHEADER
 #define ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL 0x8000                   // Does the driver control VDDCI independently from VDDC.
 #define ATOM_PP_PLATFORM_CAP_REGULATOR_HOT 0x00010000               // Enable the 'regulator hot' feature.
 #define ATOM_PP_PLATFORM_CAP_BACO          0x00020000               // Does the driver supports BACO state.
-
+#define ATOM_PP_PLATFORM_CAP_NEW_CAC_VOLTAGE   0x00040000           // Does the driver supports new CAC voltage table.
+#define ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY   0x00080000     // Does the driver supports revert GPIO5 polarity.
+#define ATOM_PP_PLATFORM_CAP_OUTPUT_THERMAL2GPIO17   0x00100000     // Does the driver supports thermal2GPIO17.
+#define ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE   0x00200000   // Does the driver supports VR HOT GPIO Configurable.
 
 typedef struct _ATOM_PPLIB_POWERPLAYTABLE
 {
@@ -7398,7 +7826,7 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE4
     USHORT                     usVddcDependencyOnMCLKOffset;
     USHORT                     usMaxClockVoltageOnDCOffset;
     USHORT                     usVddcPhaseShedLimitsTableOffset;    // Points to ATOM_PPLIB_PhaseSheddingLimits_Table
-    USHORT                     usReserved;  
+    USHORT                     usMvddDependencyOnMCLKOffset;  
 } ATOM_PPLIB_POWERPLAYTABLE4, *LPATOM_PPLIB_POWERPLAYTABLE4;
 
 typedef struct _ATOM_PPLIB_POWERPLAYTABLE5
@@ -7563,6 +7991,17 @@ typedef struct _ATOM_PPLIB_SI_CLOCK_INFO
 
 } ATOM_PPLIB_SI_CLOCK_INFO;
 
+typedef struct _ATOM_PPLIB_CI_CLOCK_INFO
+{
+      USHORT usEngineClockLow;
+      UCHAR  ucEngineClockHigh;
+
+      USHORT usMemoryClockLow;
+      UCHAR  ucMemoryClockHigh;
+      
+      UCHAR  ucPCIEGen;
+      USHORT usPCIELane;
+} ATOM_PPLIB_CI_CLOCK_INFO;
 
 typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO
 
@@ -7680,8 +8119,8 @@ typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table
 
 typedef struct _ATOM_PPLIB_CAC_Leakage_Record
 {
-    USHORT usVddc;  // We use this field for the "fake" standardized VDDC for power calculations                                                  
-    ULONG  ulLeakageValue;
+    USHORT usVddc;  // We use this field for the "fake" standardized VDDC for power calculations; For CI and newer, we use this as the real VDDC value.
+    ULONG  ulLeakageValue;  // For CI and newer we use this as the "fake" standar VDDC value.
 }ATOM_PPLIB_CAC_Leakage_Record;
 
 typedef struct _ATOM_PPLIB_CAC_Leakage_Table
@@ -7796,6 +8235,42 @@ typedef struct _ATOM_PPLIB_UVD_Table
 //    ATOM_PPLIB_UVD_State_Table states;
 }ATOM_PPLIB_UVD_Table;
 
+
+typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Record
+{
+      USHORT usVoltage;
+      USHORT usSAMClockLow;
+      UCHAR  ucSAMClockHigh;
+}ATOM_PPLIB_SAMClk_Voltage_Limit_Record;
+
+typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Table{
+    UCHAR numEntries;
+    ATOM_PPLIB_SAMClk_Voltage_Limit_Record entries[1];
+}ATOM_PPLIB_SAMClk_Voltage_Limit_Table;
+
+typedef struct _ATOM_PPLIB_SAMU_Table
+{
+      UCHAR revid;
+      ATOM_PPLIB_SAMClk_Voltage_Limit_Table limits;
+}ATOM_PPLIB_SAMU_Table;
+
+#define ATOM_PPM_A_A    1
+#define ATOM_PPM_A_I    2
+typedef struct _ATOM_PPLIB_PPM_Table
+{
+      UCHAR  ucRevId;
+      UCHAR  ucPpmDesign;          //A+I or A+A
+      USHORT usCpuCoreNumber;
+      ULONG  ulPlatformTDP;
+      ULONG  ulSmallACPlatformTDP;
+      ULONG  ulPlatformTDC;
+      ULONG  ulSmallACPlatformTDC;
+      ULONG  ulApuTDP;
+      ULONG  ulDGpuTDP;  
+      ULONG  ulDGpuUlvPower;
+      ULONG  ulTjmax;
+} ATOM_PPLIB_PPM_Table;
+
 /**************************************************************************/
 
 
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index d5df8fd..b9d3b43 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -555,7 +555,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 		if (rdev->family < CHIP_RV770)
 			radeon_crtc->pll_flags |= RADEON_PLL_PREFER_MINM_OVER_MAXP;
 		/* use frac fb div on APUs */
-		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev))
+		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev))
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 		/* use frac fb div on RS780/RS880 */
 		if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
@@ -743,7 +743,7 @@ static void atombios_crtc_set_disp_eng_pll(struct radeon_device *rdev,
 			 * SetPixelClock provides the dividers
 			 */
 			args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk);
-			if (ASIC_IS_DCE61(rdev))
+			if (ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev))
 				args.v6.ucPpll = ATOM_EXT_PLL1;
 			else if (ASIC_IS_DCE6(rdev))
 				args.v6.ucPpll = ATOM_PPLL0;
@@ -1143,7 +1143,9 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	}
 
 	if (tiling_flags & RADEON_TILING_MACRO) {
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			tmp = rdev->config.cik.tile_config;
+		else if (rdev->family >= CHIP_TAHITI)
 			tmp = rdev->config.si.tile_config;
 		else if (rdev->family >= CHIP_CAYMAN)
 			tmp = rdev->config.cayman.tile_config;
@@ -1170,11 +1172,29 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 		fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw);
 		fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh);
 		fb_format |= EVERGREEN_GRPH_MACRO_TILE_ASPECT(mtaspect);
+		if (rdev->family >= CHIP_BONAIRE) {
+			/* XXX need to know more about the surface tiling mode */
+			fb_format |= CIK_GRPH_MICRO_TILE_MODE(CIK_DISPLAY_MICRO_TILING);
+		}
 	} else if (tiling_flags & RADEON_TILING_MICRO)
 		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1);
 
-	if ((rdev->family == CHIP_TAHITI) ||
-	    (rdev->family == CHIP_PITCAIRN))
+	if (rdev->family >= CHIP_BONAIRE) {
+		u32 num_pipe_configs = rdev->config.cik.max_tile_pipes;
+		u32 num_rb = rdev->config.cik.max_backends_per_se;
+		if (num_pipe_configs > 8)
+			num_pipe_configs = 8;
+		if (num_pipe_configs == 8)
+			fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P8_32x32_16x16);
+		else if (num_pipe_configs == 4) {
+			if (num_rb == 4)
+				fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P4_16x16);
+			else if (num_rb < 4)
+				fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P4_8x16);
+		} else if (num_pipe_configs == 2)
+			fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P2);
+	} else if ((rdev->family == CHIP_TAHITI) ||
+		   (rdev->family == CHIP_PITCAIRN))
 		fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P8_32x32_8x16);
 	else if (rdev->family == CHIP_VERDE)
 		fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P4_8x16);
@@ -1224,8 +1244,12 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(EVERGREEN_GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels);
 	WREG32(EVERGREEN_GRPH_ENABLE + radeon_crtc->crtc_offset, 1);
 
-	WREG32(EVERGREEN_DESKTOP_HEIGHT + radeon_crtc->crtc_offset,
-	       target_fb->height);
+	if (rdev->family >= CHIP_BONAIRE)
+		WREG32(CIK_LB_DESKTOP_HEIGHT + radeon_crtc->crtc_offset,
+		       target_fb->height);
+	else
+		WREG32(EVERGREEN_DESKTOP_HEIGHT + radeon_crtc->crtc_offset,
+		       target_fb->height);
 	x &= ~3;
 	y &= ~1;
 	WREG32(EVERGREEN_VIEWPORT_START + radeon_crtc->crtc_offset,
@@ -1597,6 +1621,12 @@ static int radeon_get_shared_nondp_ppll(struct drm_crtc *crtc)
  *
  * Asic specific PLL information
  *
+ * DCE 8.x
+ * KB/KV
+ * - PPLL1, PPLL2 are available for all UNIPHY (both DP and non-DP)
+ * CI
+ * - PPLL0, PPLL1, PPLL2 are available for all UNIPHY (both DP and non-DP) and DAC
+ *
  * DCE 6.1
  * - PPLL2 is only available to UNIPHYA (both DP and non-DP)
  * - PPLL0, PPLL1 are available for UNIPHYB/C/D/E/F (both DP and non-DP)
@@ -1623,7 +1653,47 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 	u32 pll_in_use;
 	int pll;
 
-	if (ASIC_IS_DCE61(rdev)) {
+	if (ASIC_IS_DCE8(rdev)) {
+		if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(radeon_crtc->encoder))) {
+			if (rdev->clock.dp_extclk)
+				/* skip PPLL programming if using ext clock */
+				return ATOM_PPLL_INVALID;
+			else {
+				/* use the same PPLL for all DP monitors */
+				pll = radeon_get_shared_dp_ppll(crtc);
+				if (pll != ATOM_PPLL_INVALID)
+					return pll;
+			}
+		} else {
+			/* use the same PPLL for all monitors with the same clock */
+			pll = radeon_get_shared_nondp_ppll(crtc);
+			if (pll != ATOM_PPLL_INVALID)
+				return pll;
+		}
+		/* otherwise, pick one of the plls */
+		if ((rdev->family == CHIP_KAVERI) ||
+		    (rdev->family == CHIP_KABINI)) {
+			/* KB/KV has PPLL1 and PPLL2 */
+			pll_in_use = radeon_get_pll_use_mask(crtc);
+			if (!(pll_in_use & (1 << ATOM_PPLL2)))
+				return ATOM_PPLL2;
+			if (!(pll_in_use & (1 << ATOM_PPLL1)))
+				return ATOM_PPLL1;
+			DRM_ERROR("unable to allocate a PPLL\n");
+			return ATOM_PPLL_INVALID;
+		} else {
+			/* CI has PPLL0, PPLL1, and PPLL2 */
+			pll_in_use = radeon_get_pll_use_mask(crtc);
+			if (!(pll_in_use & (1 << ATOM_PPLL2)))
+				return ATOM_PPLL2;
+			if (!(pll_in_use & (1 << ATOM_PPLL1)))
+				return ATOM_PPLL1;
+			if (!(pll_in_use & (1 << ATOM_PPLL0)))
+				return ATOM_PPLL0;
+			DRM_ERROR("unable to allocate a PPLL\n");
+			return ATOM_PPLL_INVALID;
+		}
+	} else if (ASIC_IS_DCE61(rdev)) {
 		struct radeon_encoder_atom_dig *dig =
 			radeon_encoder->enc_priv;
 
@@ -1771,6 +1841,9 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc,
 	atombios_crtc_set_base(crtc, x, y, old_fb);
 	atombios_overscan_setup(crtc, mode, adjusted_mode);
 	atombios_scaler_setup(crtc);
+	/* update the hw version fpr dpm */
+	radeon_crtc->hw_mode = *adjusted_mode;
+
 	return 0;
 }
 
@@ -1861,7 +1934,7 @@ static void atombios_crtc_disable(struct drm_crtc *crtc)
 		break;
 	case ATOM_PPLL0:
 		/* disable the ppll */
-		if (ASIC_IS_DCE61(rdev))
+		if ((rdev->family == CHIP_ARUBA) || (rdev->family == CHIP_BONAIRE))
 			atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id,
 						  0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
 		break;
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index 8406c82..092275d 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -186,6 +186,13 @@ void radeon_atom_backlight_init(struct radeon_encoder *radeon_encoder,
 	u8 backlight_level;
 	char bl_name[16];
 
+	/* Mac laptops with multiple GPUs use the gmux driver for backlight
+	 * so don't register a backlight device
+	 */
+	if ((rdev->pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE) &&
+	    (rdev->pdev->device == 0x6741))
+		return;
+
 	if (!radeon_encoder->enc_priv)
 		return;
 
@@ -296,6 +303,7 @@ static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		return true;
 	default:
 		return false;
@@ -479,11 +487,11 @@ static u8 radeon_atom_get_bpc(struct drm_encoder *encoder)
 	}
 }
 
-
 union dvo_encoder_control {
 	ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION ext_tmds;
 	DVO_ENCODER_CONTROL_PS_ALLOCATION dvo;
 	DVO_ENCODER_CONTROL_PS_ALLOCATION_V3 dvo_v3;
+	DVO_ENCODER_CONTROL_PS_ALLOCATION_V1_4 dvo_v4;
 };
 
 void
@@ -533,6 +541,13 @@ atombios_dvo_setup(struct drm_encoder *encoder, int action)
 			args.dvo_v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
 			args.dvo_v3.ucDVOConfig = 0; /* XXX */
 			break;
+		case 4:
+			/* DCE8 */
+			args.dvo_v4.ucAction = action;
+			args.dvo_v4.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+			args.dvo_v4.ucDVOConfig = 0; /* XXX */
+			args.dvo_v4.ucBitPerColor = radeon_atom_get_bpc(encoder);
+			break;
 		default:
 			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
 			break;
@@ -915,10 +930,14 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo
 				args.v4.ucLaneNum = 4;
 
 			if (ENCODER_MODE_IS_DP(args.v4.ucEncoderMode)) {
-				if (dp_clock == 270000)
-					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ;
-				else if (dp_clock == 540000)
+				if (dp_clock == 540000)
 					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_5_40GHZ;
+				else if (dp_clock == 324000)
+					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_3_24GHZ;
+				else if (dp_clock == 270000)
+					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ;
+				else
+					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_1_62GHZ;
 			}
 			args.v4.acConfig.ucDigSel = dig->dig_encoder;
 			args.v4.ucBitPerColor = radeon_atom_get_bpc(encoder);
@@ -1012,6 +1031,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl);
 		break;
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
@@ -1271,6 +1291,9 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 				else
 					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYE;
 				break;
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+				args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYG;
+				break;
 			}
 			if (is_dp)
 				args.v5.ucLaneNum = dp_lane_count;
@@ -1735,6 +1758,7 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode)
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 		radeon_atom_encoder_dpms_dig(encoder, mode);
 		break;
@@ -1872,6 +1896,7 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 				dig = radeon_encoder->enc_priv;
 				switch (dig->dig_encoder) {
@@ -1893,6 +1918,9 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
 				case 5:
 					args.v2.ucEncoderID = ASIC_INT_DIG6_ENCODER_ID;
 					break;
+				case 6:
+					args.v2.ucEncoderID = ASIC_INT_DIG7_ENCODER_ID;
+					break;
 				}
 				break;
 			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
@@ -1955,7 +1983,13 @@ atombios_apply_encoder_quirks(struct drm_encoder *encoder,
 	/* set scaler clears this on some chips */
 	if (ASIC_IS_AVIVO(rdev) &&
 	    (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)))) {
-		if (ASIC_IS_DCE4(rdev)) {
+		if (ASIC_IS_DCE8(rdev)) {
+			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+				WREG32(CIK_LB_DATA_FORMAT + radeon_crtc->crtc_offset,
+				       CIK_INTERLEAVE_EN);
+			else
+				WREG32(CIK_LB_DATA_FORMAT + radeon_crtc->crtc_offset, 0);
+		} else if (ASIC_IS_DCE4(rdev)) {
 			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
 				WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset,
 				       EVERGREEN_INTERLEAVE_EN);
@@ -2002,6 +2036,9 @@ static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder)
 			else
 				return 4;
 			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+			return 6;
+			break;
 		}
 	} else if (ASIC_IS_DCE4(rdev)) {
 		/* DCE4/5 */
@@ -2086,6 +2123,7 @@ radeon_atom_encoder_init(struct radeon_device *rdev)
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0);
 			break;
@@ -2130,6 +2168,7 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 		/* handled in dpms */
 		break;
@@ -2395,6 +2434,7 @@ static void radeon_atom_encoder_disable(struct drm_encoder *encoder)
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 		/* handled in dpms */
 		break;
@@ -2626,6 +2666,7 @@ radeon_add_atom_encoder(struct drm_device *dev,
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
 			radeon_encoder->rmx_type = RMX_FULL;
 			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c
new file mode 100644
index 0000000..0bfd55e
--- /dev/null
+++ b/drivers/gpu/drm/radeon/btc_dpm.c
@@ -0,0 +1,2751 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "btcd.h"
+#include "r600_dpm.h"
+#include "cypress_dpm.h"
+#include "btc_dpm.h"
+#include "atom.h"
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define MC_CG_SEQ_DRAMCONF_S0       0x05
+#define MC_CG_SEQ_DRAMCONF_S1       0x06
+#define MC_CG_SEQ_YCLK_SUSPEND      0x04
+#define MC_CG_SEQ_YCLK_RESUME       0x0a
+
+#define SMC_RAM_END 0x8000
+
+#ifndef BTC_MGCG_SEQUENCE
+#define BTC_MGCG_SEQUENCE  300
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps);
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+
+
+//********* BARTS **************//
+static const u32 barts_cgcg_cgls_default[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define BARTS_CGCG_CGLS_DEFAULT_LENGTH sizeof(barts_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 barts_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7912, 0x001f4180,
+	0x00000644, 0x000f3812, 0x001f4180
+};
+#define BARTS_CGCG_CGLS_DISABLE_LENGTH sizeof(barts_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 barts_cgcg_cgls_enable[] =
+{
+	/* 0x0000c124, 0x84180000, 0x00180000, */
+	0x00000644, 0x000f7892, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define BARTS_CGCG_CGLS_ENABLE_LENGTH sizeof(barts_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+static const u32 barts_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00600100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00000100, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x0000977c, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009784, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000100, 0xffffffff,
+	0x0000d0c0, 0xff000100, 0xffffffff,
+	0x0000802c, 0x40000000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0x40010000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define BARTS_MGCG_DEFAULT_LENGTH sizeof(barts_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 barts_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define BARTS_MGCG_DISABLE_LENGTH sizeof(barts_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 barts_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x81944000, 0xffffffff
+};
+#define BARTS_MGCG_ENABLE_LENGTH sizeof(barts_mgcg_enable) / (3 * sizeof(u32))
+
+//********* CAICOS **************//
+static const u32 caicos_cgcg_cgls_default[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAICOS_CGCG_CGLS_DEFAULT_LENGTH sizeof(caicos_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 caicos_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7912, 0x001f4180,
+	0x00000644, 0x000f3812, 0x001f4180
+};
+#define CAICOS_CGCG_CGLS_DISABLE_LENGTH sizeof(caicos_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 caicos_cgcg_cgls_enable[] =
+{
+	/* 0x0000c124, 0x84180000, 0x00180000, */
+	0x00000644, 0x000f7892, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define CAICOS_CGCG_CGLS_ENABLE_LENGTH sizeof(caicos_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+static const u32 caicos_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00600100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00000100, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x0000977c, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009784, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000100, 0xffffffff,
+	0x0000d0c0, 0xff000100, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAICOS_MGCG_DEFAULT_LENGTH sizeof(caicos_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 caicos_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define CAICOS_MGCG_DISABLE_LENGTH sizeof(caicos_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 caicos_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x46944040, 0xffffffff
+};
+#define CAICOS_MGCG_ENABLE_LENGTH sizeof(caicos_mgcg_enable) / (3 * sizeof(u32))
+
+//********* TURKS **************//
+static const u32 turks_cgcg_cgls_default[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define TURKS_CGCG_CGLS_DEFAULT_LENGTH  sizeof(turks_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 turks_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7912, 0x001f4180,
+	0x00000644, 0x000f3812, 0x001f4180
+};
+#define TURKS_CGCG_CGLS_DISABLE_LENGTH sizeof(turks_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 turks_cgcg_cgls_enable[] =
+{
+	/* 0x0000c124, 0x84180000, 0x00180000, */
+	0x00000644, 0x000f7892, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define TURKS_CGCG_CGLS_ENABLE_LENGTH sizeof(turks_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+// These are the sequences for turks_mgcg_shls
+static const u32 turks_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00600100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00000100, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x0000977c, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009784, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000100, 0xffffffff,
+	0x0000d0c0, 0x00000100, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define TURKS_MGCG_DEFAULT_LENGTH sizeof(turks_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 turks_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define TURKS_MGCG_DISABLE_LENGTH sizeof(turks_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 turks_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x6e944000, 0xffffffff
+};
+#define TURKS_MGCG_ENABLE_LENGTH sizeof(turks_mgcg_enable) / (3 * sizeof(u32))
+
+#endif
+
+#ifndef BTC_SYSLS_SEQUENCE
+#define BTC_SYSLS_SEQUENCE  100
+
+
+//********* BARTS **************//
+static const u32 barts_sysls_default[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define BARTS_SYSLS_DEFAULT_LENGTH sizeof(barts_sysls_default) / (3 * sizeof(u32))
+
+static const u32 barts_sysls_disable[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+#define BARTS_SYSLS_DISABLE_LENGTH sizeof(barts_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 barts_sysls_enable[] =
+{
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define BARTS_SYSLS_ENABLE_LENGTH sizeof(barts_sysls_enable) / (3 * sizeof(u32))
+
+//********* CAICOS **************//
+static const u32 caicos_sysls_default[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define CAICOS_SYSLS_DEFAULT_LENGTH sizeof(caicos_sysls_default) / (3 * sizeof(u32))
+
+static const u32 caicos_sysls_disable[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+#define CAICOS_SYSLS_DISABLE_LENGTH sizeof(caicos_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 caicos_sysls_enable[] =
+{
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff
+};
+#define CAICOS_SYSLS_ENABLE_LENGTH sizeof(caicos_sysls_enable) / (3 * sizeof(u32))
+
+//********* TURKS **************//
+static const u32 turks_sysls_default[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define TURKS_SYSLS_DEFAULT_LENGTH sizeof(turks_sysls_default) / (3 * sizeof(u32))
+
+static const u32 turks_sysls_disable[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+#define TURKS_SYSLS_DISABLE_LENGTH sizeof(turks_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 turks_sysls_enable[] =
+{
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define TURKS_SYSLS_ENABLE_LENGTH sizeof(turks_sysls_enable) / (3 * sizeof(u32))
+
+#endif
+
+u32 btc_valid_sclk[40] =
+{
+	5000,   10000,  15000,  20000,  25000,  30000,  35000,  40000,  45000,  50000,
+	55000,  60000,  65000,  70000,  75000,  80000,  85000,  90000,  95000,  100000,
+	105000, 110000, 11500,  120000, 125000, 130000, 135000, 140000, 145000, 150000,
+	155000, 160000, 165000, 170000, 175000, 180000, 185000, 190000, 195000, 200000
+};
+
+static const struct radeon_blacklist_clocks btc_blacklist_clocks[] =
+{
+        { 10000, 30000, RADEON_SCLK_UP },
+        { 15000, 30000, RADEON_SCLK_UP },
+        { 20000, 30000, RADEON_SCLK_UP },
+        { 25000, 30000, RADEON_SCLK_UP }
+};
+
+void btc_apply_voltage_dependency_rules(struct radeon_clock_voltage_dependency_table *table,
+					u32 clock, u16 max_voltage, u16 *voltage)
+{
+	u32 i;
+
+	if ((table == NULL) || (table->count == 0))
+		return;
+
+	for (i= 0; i < table->count; i++) {
+		if (clock <= table->entries[i].clk) {
+			if (*voltage < table->entries[i].v)
+				*voltage = (u16)((table->entries[i].v < max_voltage) ?
+						  table->entries[i].v : max_voltage);
+			return;
+		}
+	}
+
+	*voltage = (*voltage > max_voltage) ? *voltage : max_voltage;
+}
+
+static u32 btc_find_valid_clock(struct radeon_clock_array *clocks,
+				u32 max_clock, u32 requested_clock)
+{
+	unsigned int i;
+
+	if ((clocks == NULL) || (clocks->count == 0))
+		return (requested_clock < max_clock) ? requested_clock : max_clock;
+
+	for (i = 0; i < clocks->count; i++) {
+		if (clocks->values[i] >= requested_clock)
+			return (clocks->values[i] < max_clock) ? clocks->values[i] : max_clock;
+	}
+
+	return (clocks->values[clocks->count - 1] < max_clock) ?
+		clocks->values[clocks->count - 1] : max_clock;
+}
+
+static u32 btc_get_valid_mclk(struct radeon_device *rdev,
+			      u32 max_mclk, u32 requested_mclk)
+{
+	return btc_find_valid_clock(&rdev->pm.dpm.dyn_state.valid_mclk_values,
+				    max_mclk, requested_mclk);
+}
+
+static u32 btc_get_valid_sclk(struct radeon_device *rdev,
+			      u32 max_sclk, u32 requested_sclk)
+{
+	return btc_find_valid_clock(&rdev->pm.dpm.dyn_state.valid_sclk_values,
+				    max_sclk, requested_sclk);
+}
+
+void btc_skip_blacklist_clocks(struct radeon_device *rdev,
+			       const u32 max_sclk, const u32 max_mclk,
+			       u32 *sclk, u32 *mclk)
+{
+	int i, num_blacklist_clocks;
+
+	if ((sclk == NULL) || (mclk == NULL))
+		return;
+
+	num_blacklist_clocks = ARRAY_SIZE(btc_blacklist_clocks);
+
+	for (i = 0; i < num_blacklist_clocks; i++) {
+		if ((btc_blacklist_clocks[i].sclk == *sclk) &&
+		    (btc_blacklist_clocks[i].mclk == *mclk))
+			break;
+	}
+
+	if (i < num_blacklist_clocks) {
+		if (btc_blacklist_clocks[i].action == RADEON_SCLK_UP) {
+			*sclk = btc_get_valid_sclk(rdev, max_sclk, *sclk + 1);
+
+			if (*sclk < max_sclk)
+				btc_skip_blacklist_clocks(rdev, max_sclk, max_mclk, sclk, mclk);
+		}
+	}
+}
+
+void btc_adjust_clock_combinations(struct radeon_device *rdev,
+				   const struct radeon_clock_and_voltage_limits *max_limits,
+				   struct rv7xx_pl *pl)
+{
+
+	if ((pl->mclk == 0) || (pl->sclk == 0))
+		return;
+
+	if (pl->mclk == pl->sclk)
+		return;
+
+	if (pl->mclk > pl->sclk) {
+		if (((pl->mclk + (pl->sclk - 1)) / pl->sclk) > rdev->pm.dpm.dyn_state.mclk_sclk_ratio)
+			pl->sclk = btc_get_valid_sclk(rdev,
+						      max_limits->sclk,
+						      (pl->mclk +
+						       (rdev->pm.dpm.dyn_state.mclk_sclk_ratio - 1)) /
+						      rdev->pm.dpm.dyn_state.mclk_sclk_ratio);
+	} else {
+		if ((pl->sclk - pl->mclk) > rdev->pm.dpm.dyn_state.sclk_mclk_delta)
+			pl->mclk = btc_get_valid_mclk(rdev,
+						      max_limits->mclk,
+						      pl->sclk -
+						      rdev->pm.dpm.dyn_state.sclk_mclk_delta);
+	}
+}
+
+static u16 btc_find_voltage(struct atom_voltage_table *table, u16 voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (voltage <= table->entries[i].value)
+			return table->entries[i].value;
+	}
+
+	return table->entries[table->count - 1].value;
+}
+
+void btc_apply_voltage_delta_rules(struct radeon_device *rdev,
+				   u16 max_vddc, u16 max_vddci,
+				   u16 *vddc, u16 *vddci)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u16 new_voltage;
+
+	if ((0 == *vddc) || (0 == *vddci))
+		return;
+
+	if (*vddc > *vddci) {
+		if ((*vddc - *vddci) > rdev->pm.dpm.dyn_state.vddc_vddci_delta) {
+			new_voltage = btc_find_voltage(&eg_pi->vddci_voltage_table,
+						       (*vddc - rdev->pm.dpm.dyn_state.vddc_vddci_delta));
+			*vddci = (new_voltage < max_vddci) ? new_voltage : max_vddci;
+		}
+	} else {
+		if ((*vddci - *vddc) > rdev->pm.dpm.dyn_state.vddc_vddci_delta) {
+			new_voltage = btc_find_voltage(&eg_pi->vddc_voltage_table,
+						       (*vddci - rdev->pm.dpm.dyn_state.vddc_vddci_delta));
+			*vddc = (new_voltage < max_vddc) ? new_voltage : max_vddc;
+		}
+	}
+}
+
+static void btc_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					     bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp, bif;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if (enable) {
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+				tmp |= LC_GEN2_EN_STRAP;
+
+				tmp |= LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+				udelay(10);
+				tmp &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+			}
+		}
+	} else {
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp &= ~LC_GEN2_EN_STRAP;
+			}
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+		}
+	}
+}
+
+static void btc_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					 bool enable)
+{
+	btc_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+static int btc_disable_ulv(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->ulv.supported) {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_DisableULV) != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+	return 0;
+}
+
+static int btc_populate_ulv_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table)
+{
+	int ret = -EINVAL;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_pl *ulv_pl = eg_pi->ulv.pl;
+
+	if (ulv_pl->vddc) {
+		ret = cypress_convert_power_level_to_smc(rdev,
+							 ulv_pl,
+							 &table->ULVState.levels[0],
+							 PPSMC_DISPLAY_WATERMARK_LOW);
+		if (ret == 0) {
+			table->ULVState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+			table->ULVState.levels[0].ACIndex = 1;
+
+			table->ULVState.levels[1] = table->ULVState.levels[0];
+			table->ULVState.levels[2] = table->ULVState.levels[0];
+
+			table->ULVState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+			WREG32(CG_ULV_CONTROL, BTC_CGULVCONTROL_DFLT);
+			WREG32(CG_ULV_PARAMETER, BTC_CGULVPARAMETER_DFLT);
+		}
+	}
+
+	return ret;
+}
+
+static int btc_populate_smc_acpi_state(struct radeon_device *rdev,
+				       RV770_SMC_STATETABLE *table)
+{
+	int ret = cypress_populate_smc_acpi_state(rdev, table);
+
+	if (ret == 0) {
+		table->ACPIState.levels[0].ACIndex = 0;
+		table->ACPIState.levels[1].ACIndex = 0;
+		table->ACPIState.levels[2].ACIndex = 0;
+	}
+
+	return ret;
+}
+
+void btc_program_mgcg_hw_sequence(struct radeon_device *rdev,
+				  const u32 *sequence, u32 count)
+{
+	u32 i, length = count * 3;
+	u32 tmp;
+
+	for (i = 0; i < length; i+=3) {
+		tmp = RREG32(sequence[i]);
+		tmp &= ~sequence[i+2];
+		tmp |= sequence[i+1] & sequence[i+2];
+		WREG32(sequence[i], tmp);
+	}
+}
+
+static void btc_cg_clock_gating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (rdev->family == CHIP_BARTS) {
+		p = (const u32 *)&barts_cgcg_cgls_default;
+		count = BARTS_CGCG_CGLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_TURKS) {
+		p = (const u32 *)&turks_cgcg_cgls_default;
+		count = TURKS_CGCG_CGLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_CAICOS) {
+		p = (const u32 *)&caicos_cgcg_cgls_default;
+		count = CAICOS_CGCG_CGLS_DEFAULT_LENGTH;
+	} else
+		return;
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_cg_clock_gating_enable(struct radeon_device *rdev,
+				       bool enable)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (enable) {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_cgcg_cgls_enable;
+			count = BARTS_CGCG_CGLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_cgcg_cgls_enable;
+			count = TURKS_CGCG_CGLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_cgcg_cgls_enable;
+			count = CAICOS_CGCG_CGLS_ENABLE_LENGTH;
+		} else
+			return;
+	} else {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_cgcg_cgls_disable;
+			count = BARTS_CGCG_CGLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_cgcg_cgls_disable;
+			count = TURKS_CGCG_CGLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_cgcg_cgls_disable;
+			count = CAICOS_CGCG_CGLS_DISABLE_LENGTH;
+		} else
+			return;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_mg_clock_gating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (rdev->family == CHIP_BARTS) {
+		p = (const u32 *)&barts_mgcg_default;
+		count = BARTS_MGCG_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_TURKS) {
+		p = (const u32 *)&turks_mgcg_default;
+		count = TURKS_MGCG_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_CAICOS) {
+		p = (const u32 *)&caicos_mgcg_default;
+		count = CAICOS_MGCG_DEFAULT_LENGTH;
+	} else
+		return;
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_mg_clock_gating_enable(struct radeon_device *rdev,
+				       bool enable)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (enable) {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_mgcg_enable;
+			count = BARTS_MGCG_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_mgcg_enable;
+			count = TURKS_MGCG_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_mgcg_enable;
+			count = CAICOS_MGCG_ENABLE_LENGTH;
+		} else
+			return;
+	} else {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_mgcg_disable[0];
+			count = BARTS_MGCG_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_mgcg_disable[0];
+			count = TURKS_MGCG_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_mgcg_disable[0];
+			count = CAICOS_MGCG_DISABLE_LENGTH;
+		} else
+			return;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_ls_clock_gating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (rdev->family == CHIP_BARTS) {
+		p = (const u32 *)&barts_sysls_default;
+		count = BARTS_SYSLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_TURKS) {
+		p = (const u32 *)&turks_sysls_default;
+		count = TURKS_SYSLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_CAICOS) {
+		p = (const u32 *)&caicos_sysls_default;
+		count = CAICOS_SYSLS_DEFAULT_LENGTH;
+	} else
+		return;
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_ls_clock_gating_enable(struct radeon_device *rdev,
+				       bool enable)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (enable) {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_sysls_enable;
+			count = BARTS_SYSLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_sysls_enable;
+			count = TURKS_SYSLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_sysls_enable;
+			count = CAICOS_SYSLS_ENABLE_LENGTH;
+		} else
+			return;
+	} else {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_sysls_disable;
+			count = BARTS_SYSLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_sysls_disable;
+			count = TURKS_SYSLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_sysls_disable;
+			count = CAICOS_SYSLS_DISABLE_LENGTH;
+		} else
+			return;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+bool btc_dpm_enabled(struct radeon_device *rdev)
+{
+	if (rv770_is_smc_running(rdev))
+		return true;
+	else
+		return false;
+}
+
+static int btc_init_smc_table(struct radeon_device *rdev,
+			      struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	RV770_SMC_STATETABLE *table = &pi->smc_statetable;
+	int ret;
+
+	memset(table, 0, sizeof(RV770_SMC_STATETABLE));
+
+	cypress_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+        case THERMAL_TYPE_EVERGREEN:
+        case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+        case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+        default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	ret = cypress_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->sclk_deep_sleep)
+		WREG32_P(SCLK_PSKIP_CNTL, PSKIP_ON_ALLOW_STOP_HI(32),
+			 ~PSKIP_ON_ALLOW_STOP_HI_MASK);
+
+	ret = btc_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->ulv.supported) {
+		ret = btc_populate_ulv_state(rdev, table);
+		if (ret)
+			eg_pi->ulv.supported = false;
+	}
+
+	table->driverState = table->initialState;
+
+	return rv770_copy_bytes_to_smc(rdev,
+				       pi->state_table_start,
+				       (u8 *)table,
+				       sizeof(RV770_SMC_STATETABLE),
+				       pi->sram_end);
+}
+
+static void btc_set_at_for_uvd(struct radeon_device *rdev,
+			       struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int idx = 0;
+
+	if (r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2))
+		idx = 1;
+
+	if ((idx == 1) && !eg_pi->smu_uvd_hs) {
+		pi->rlp = 10;
+		pi->rmp = 100;
+		pi->lhp = 100;
+		pi->lmp = 10;
+	} else {
+		pi->rlp = eg_pi->ats[idx].rlp;
+		pi->rmp = eg_pi->ats[idx].rmp;
+		pi->lhp = eg_pi->ats[idx].lhp;
+		pi->lmp = eg_pi->ats[idx].lmp;
+	}
+
+}
+
+void btc_notify_uvd_to_smc(struct radeon_device *rdev,
+			   struct radeon_ps *radeon_new_state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2)) {
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_uvd_enabled, 1);
+		eg_pi->uvd_enabled = true;
+	} else {
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_uvd_enabled, 0);
+		eg_pi->uvd_enabled = false;
+	}
+}
+
+int btc_reset_to_default(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_ResetToDefaults) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void btc_stop_smc(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (((RREG32(LB_SYNC_RESET_SEL) & LB_SYNC_RESET_SEL_MASK) >> LB_SYNC_RESET_SEL_SHIFT) != 1)
+			break;
+		udelay(1);
+	}
+	udelay(100);
+
+	r7xx_stop_smc(rdev);
+}
+
+void btc_read_arb_registers(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct evergreen_arb_registers *arb_registers =
+		&eg_pi->bootup_arb_registers;
+
+	arb_registers->mc_arb_dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	arb_registers->mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+	arb_registers->mc_arb_rfsh_rate = RREG32(MC_ARB_RFSH_RATE);
+	arb_registers->mc_arb_burst_time = RREG32(MC_ARB_BURST_TIME);
+}
+
+
+static void btc_set_arb0_registers(struct radeon_device *rdev,
+				   struct evergreen_arb_registers *arb_registers)
+{
+	u32 val;
+
+	WREG32(MC_ARB_DRAM_TIMING,  arb_registers->mc_arb_dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2, arb_registers->mc_arb_dram_timing2);
+
+	val = (arb_registers->mc_arb_rfsh_rate & POWERMODE0_MASK) >>
+		POWERMODE0_SHIFT;
+	WREG32_P(MC_ARB_RFSH_RATE, POWERMODE0(val), ~POWERMODE0_MASK);
+
+	val = (arb_registers->mc_arb_burst_time & STATE0_MASK) >>
+		STATE0_SHIFT;
+	WREG32_P(MC_ARB_BURST_TIME, STATE0(val), ~STATE0_MASK);
+}
+
+static void btc_set_boot_state_timing(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->ulv.supported)
+		btc_set_arb0_registers(rdev, &eg_pi->bootup_arb_registers);
+}
+
+static bool btc_is_state_ulv_compatible(struct radeon_device *rdev,
+					struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_pl *ulv_pl = eg_pi->ulv.pl;
+
+	if (state->low.mclk != ulv_pl->mclk)
+		return false;
+
+	if (state->low.vddci != ulv_pl->vddci)
+		return false;
+
+	/* XXX check minclocks, etc. */
+
+	return true;
+}
+
+
+static int btc_set_ulv_dram_timing(struct radeon_device *rdev)
+{
+	u32 val;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_pl *ulv_pl = eg_pi->ulv.pl;
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    ulv_pl->sclk,
+					    ulv_pl->mclk);
+
+	val = rv770_calculate_memory_refresh_rate(rdev, ulv_pl->sclk);
+	WREG32_P(MC_ARB_RFSH_RATE, POWERMODE0(val), ~POWERMODE0_MASK);
+
+	val = cypress_calculate_burst_time(rdev, ulv_pl->sclk, ulv_pl->mclk);
+	WREG32_P(MC_ARB_BURST_TIME, STATE0(val), ~STATE0_MASK);
+
+	return 0;
+}
+
+static int btc_enable_ulv(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableULV) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int btc_set_power_state_conditionally_enable_ulv(struct radeon_device *rdev,
+							struct radeon_ps *radeon_new_state)
+{
+	int ret = 0;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->ulv.supported) {
+		if (btc_is_state_ulv_compatible(rdev, radeon_new_state)) {
+			// Set ARB[0] to reflect the DRAM timing needed for ULV.
+			ret = btc_set_ulv_dram_timing(rdev);
+			if (ret == 0)
+				ret = btc_enable_ulv(rdev);
+		}
+	}
+
+	return ret;
+}
+
+static bool btc_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg)
+{
+	bool result = true;
+
+	switch (in_reg) {
+	case MC_SEQ_RAS_TIMING >> 2:
+		*out_reg = MC_SEQ_RAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_CAS_TIMING >> 2:
+		*out_reg = MC_SEQ_CAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING2 >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING2_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D1_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D1_LP >> 2;
+		break;
+        case MC_PMG_CMD_EMRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS1 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+		break;
+        default:
+		result = false;
+		break;
+	}
+
+	return result;
+}
+
+static void btc_set_valid_flag(struct evergreen_mc_reg_table *table)
+{
+	u8 i, j;
+
+	for (i = 0; i < table->last; i++) {
+		for (j = 1; j < table->num_entries; j++) {
+			if (table->mc_reg_table_entry[j-1].mc_data[i] !=
+			    table->mc_reg_table_entry[j].mc_data[i]) {
+				table->valid_flag |= (1 << i);
+				break;
+			}
+		}
+	}
+}
+
+static int btc_set_mc_special_registers(struct radeon_device *rdev,
+					struct evergreen_mc_reg_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 i, j, k;
+	u32 tmp;
+
+	for (i = 0, j = table->last; i < table->last; i++) {
+		switch (table->mc_reg_address[i].s1) {
+		case MC_SEQ_MISC1 >> 2:
+			tmp = RREG32(MC_PMG_CMD_EMRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					((tmp & 0xffff0000)) |
+					((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16);
+			}
+			j++;
+
+			if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			tmp = RREG32(MC_PMG_CMD_MRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(tmp & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+				if (!pi->mem_gddr5)
+					table->mc_reg_table_entry[k].mc_data[j] |= 0x100;
+			}
+			j++;
+
+			if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		case MC_SEQ_RESERVE_M >> 2:
+			tmp = RREG32(MC_PMG_CMD_MRS1);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(tmp & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+			}
+			j++;
+
+			if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		default:
+			break;
+		}
+	}
+
+	table->last = j;
+
+	return 0;
+}
+
+static void btc_set_s0_mc_reg_index(struct evergreen_mc_reg_table *table)
+{
+	u32 i;
+	u16 address;
+
+	for (i = 0; i < table->last; i++) {
+		table->mc_reg_address[i].s0 =
+			btc_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ?
+			address : table->mc_reg_address[i].s1;
+	}
+}
+
+static int btc_copy_vbios_mc_reg_table(struct atom_mc_reg_table *table,
+				       struct evergreen_mc_reg_table *eg_table)
+{
+	u8 i, j;
+
+	if (table->last > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+		return -EINVAL;
+
+	if (table->num_entries > MAX_AC_TIMING_ENTRIES)
+		return -EINVAL;
+
+	for (i = 0; i < table->last; i++)
+		eg_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1;
+	eg_table->last = table->last;
+
+	for (i = 0; i < table->num_entries; i++) {
+		eg_table->mc_reg_table_entry[i].mclk_max =
+			table->mc_reg_table_entry[i].mclk_max;
+		for(j = 0; j < table->last; j++)
+			eg_table->mc_reg_table_entry[i].mc_data[j] =
+				table->mc_reg_table_entry[i].mc_data[j];
+	}
+	eg_table->num_entries = table->num_entries;
+
+	return 0;
+}
+
+static int btc_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	int ret;
+	struct atom_mc_reg_table *table;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct evergreen_mc_reg_table *eg_table = &eg_pi->mc_reg_table;
+	u8 module_index = rv770_get_memory_module_index(rdev);
+
+	table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL);
+	if (!table)
+		return -ENOMEM;
+
+	/* Program additional LP registers that are no longer programmed by VBIOS */
+	WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING));
+	WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2));
+	WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0));
+	WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1));
+	WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0));
+	WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1));
+	WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1));
+
+	ret = radeon_atom_init_mc_reg_table(rdev, module_index, table);
+
+	if (ret)
+		goto init_mc_done;
+
+	ret = btc_copy_vbios_mc_reg_table(table, eg_table);
+
+	if (ret)
+		goto init_mc_done;
+
+	btc_set_s0_mc_reg_index(eg_table);
+	ret = btc_set_mc_special_registers(rdev, eg_table);
+
+	if (ret)
+		goto init_mc_done;
+
+	btc_set_valid_flag(eg_table);
+
+init_mc_done:
+	kfree(table);
+
+	return ret;
+}
+
+static void btc_init_stutter_mode(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	if (pi->mclk_stutter_mode_threshold) {
+		if (pi->mem_gddr5) {
+			tmp = RREG32(MC_PMG_AUTO_CFG);
+			if ((0x200 & tmp) == 0) {
+				tmp = (tmp & 0xfffffc0b) | 0x204;
+				WREG32(MC_PMG_AUTO_CFG, tmp);
+			}
+		}
+	}
+}
+
+bool btc_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+	u32 switch_limit = pi->mem_gddr5 ? 450 : 100;
+
+	if (vblank_time < switch_limit)
+		return true;
+	else
+		return false;
+
+}
+
+static void btc_apply_state_adjust_rules(struct radeon_device *rdev,
+					 struct radeon_ps *rps)
+{
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	struct radeon_clock_and_voltage_limits *max_limits;
+	bool disable_mclk_switching;
+	u32 mclk, sclk;
+	u16 vddc, vddci;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
+	    btc_dpm_vblank_too_short(rdev))
+		disable_mclk_switching = true;
+	else
+		disable_mclk_switching = false;
+
+	if (rdev->pm.dpm.ac_power)
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
+	else
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc;
+
+	if (rdev->pm.dpm.ac_power == false) {
+		if (ps->high.mclk > max_limits->mclk)
+			ps->high.mclk = max_limits->mclk;
+		if (ps->high.sclk > max_limits->sclk)
+			ps->high.sclk = max_limits->sclk;
+		if (ps->high.vddc > max_limits->vddc)
+			ps->high.vddc = max_limits->vddc;
+		if (ps->high.vddci > max_limits->vddci)
+			ps->high.vddci = max_limits->vddci;
+
+		if (ps->medium.mclk > max_limits->mclk)
+			ps->medium.mclk = max_limits->mclk;
+		if (ps->medium.sclk > max_limits->sclk)
+			ps->medium.sclk = max_limits->sclk;
+		if (ps->medium.vddc > max_limits->vddc)
+			ps->medium.vddc = max_limits->vddc;
+		if (ps->medium.vddci > max_limits->vddci)
+			ps->medium.vddci = max_limits->vddci;
+
+		if (ps->low.mclk > max_limits->mclk)
+			ps->low.mclk = max_limits->mclk;
+		if (ps->low.sclk > max_limits->sclk)
+			ps->low.sclk = max_limits->sclk;
+		if (ps->low.vddc > max_limits->vddc)
+			ps->low.vddc = max_limits->vddc;
+		if (ps->low.vddci > max_limits->vddci)
+			ps->low.vddci = max_limits->vddci;
+	}
+
+	/* XXX validate the min clocks required for display */
+
+	if (disable_mclk_switching) {
+		sclk = ps->low.sclk;
+		mclk = ps->high.mclk;
+		vddc = ps->low.vddc;
+		vddci = ps->high.vddci;
+	} else {
+		sclk = ps->low.sclk;
+		mclk = ps->low.mclk;
+		vddc = ps->low.vddc;
+		vddci = ps->low.vddci;
+	}
+
+	/* adjusted low state */
+	ps->low.sclk = sclk;
+	ps->low.mclk = mclk;
+	ps->low.vddc = vddc;
+	ps->low.vddci = vddci;
+
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->low.sclk, &ps->low.mclk);
+
+	/* adjusted medium, high states */
+	if (ps->medium.sclk < ps->low.sclk)
+		ps->medium.sclk = ps->low.sclk;
+	if (ps->medium.vddc < ps->low.vddc)
+		ps->medium.vddc = ps->low.vddc;
+	if (ps->high.sclk < ps->medium.sclk)
+		ps->high.sclk = ps->medium.sclk;
+	if (ps->high.vddc < ps->medium.vddc)
+		ps->high.vddc = ps->medium.vddc;
+
+	if (disable_mclk_switching) {
+		mclk = ps->low.mclk;
+		if (mclk < ps->medium.mclk)
+			mclk = ps->medium.mclk;
+		if (mclk < ps->high.mclk)
+			mclk = ps->high.mclk;
+		ps->low.mclk = mclk;
+		ps->low.vddci = vddci;
+		ps->medium.mclk = mclk;
+		ps->medium.vddci = vddci;
+		ps->high.mclk = mclk;
+		ps->high.vddci = vddci;
+	} else {
+		if (ps->medium.mclk < ps->low.mclk)
+			ps->medium.mclk = ps->low.mclk;
+		if (ps->medium.vddci < ps->low.vddci)
+			ps->medium.vddci = ps->low.vddci;
+		if (ps->high.mclk < ps->medium.mclk)
+			ps->high.mclk = ps->medium.mclk;
+		if (ps->high.vddci < ps->medium.vddci)
+			ps->high.vddci = ps->medium.vddci;
+	}
+
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->medium.sclk, &ps->medium.mclk);
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->high.sclk, &ps->high.mclk);
+
+	btc_adjust_clock_combinations(rdev, max_limits, &ps->low);
+	btc_adjust_clock_combinations(rdev, max_limits, &ps->medium);
+	btc_adjust_clock_combinations(rdev, max_limits, &ps->high);
+
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+					   ps->low.sclk, max_limits->vddc, &ps->low.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+					   ps->low.mclk, max_limits->vddci, &ps->low.vddci);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+					   ps->low.mclk, max_limits->vddc, &ps->low.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+					   rdev->clock.current_dispclk, max_limits->vddc, &ps->low.vddc);
+
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+					   ps->medium.sclk, max_limits->vddc, &ps->medium.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+					   ps->medium.mclk, max_limits->vddci, &ps->medium.vddci);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+					   ps->medium.mclk, max_limits->vddc, &ps->medium.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+					   rdev->clock.current_dispclk, max_limits->vddc, &ps->medium.vddc);
+
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+					   ps->high.sclk, max_limits->vddc, &ps->high.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+					   ps->high.mclk, max_limits->vddci, &ps->high.vddci);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+					   ps->high.mclk, max_limits->vddc, &ps->high.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+					   rdev->clock.current_dispclk, max_limits->vddc, &ps->high.vddc);
+
+	btc_apply_voltage_delta_rules(rdev, max_limits->vddc, max_limits->vddci,
+				      &ps->low.vddc, &ps->low.vddci);
+	btc_apply_voltage_delta_rules(rdev, max_limits->vddc, max_limits->vddci,
+				      &ps->medium.vddc, &ps->medium.vddci);
+	btc_apply_voltage_delta_rules(rdev, max_limits->vddc, max_limits->vddci,
+				      &ps->high.vddc, &ps->high.vddci);
+
+	if ((ps->high.vddc <= rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc) &&
+	    (ps->medium.vddc <= rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc) &&
+	    (ps->low.vddc <= rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc))
+		ps->dc_compatible = true;
+	else
+		ps->dc_compatible = false;
+
+	if (ps->low.vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+		ps->low.flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+	if (ps->medium.vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+		ps->medium.flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+	if (ps->high.vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+		ps->high.flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+}
+
+static void btc_update_current_ps(struct radeon_device *rdev,
+				  struct radeon_ps *rps)
+{
+	struct rv7xx_ps *new_ps = rv770_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	eg_pi->current_rps = *rps;
+	eg_pi->current_ps = *new_ps;
+	eg_pi->current_rps.ps_priv = &eg_pi->current_ps;
+}
+
+static void btc_update_requested_ps(struct radeon_device *rdev,
+				    struct radeon_ps *rps)
+{
+	struct rv7xx_ps *new_ps = rv770_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	eg_pi->requested_rps = *rps;
+	eg_pi->requested_ps = *new_ps;
+	eg_pi->requested_rps.ps_priv = &eg_pi->requested_ps;
+}
+
+void btc_dpm_reset_asic(struct radeon_device *rdev)
+{
+	rv770_restrict_performance_levels_before_switch(rdev);
+	btc_disable_ulv(rdev);
+	btc_set_boot_state_timing(rdev);
+	rv770_set_boot_state(rdev);
+}
+
+int btc_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	btc_update_requested_ps(rdev, new_ps);
+
+	btc_apply_state_adjust_rules(rdev, &eg_pi->requested_rps);
+
+	return 0;
+}
+
+int btc_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+	struct radeon_ps *old_ps = &eg_pi->current_rps;
+	int ret;
+
+	ret = btc_disable_ulv(rdev);
+	btc_set_boot_state_timing(rdev);
+	ret = rv770_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_before_state_change(rdev, new_ps, old_ps);
+
+	rv770_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	btc_set_at_for_uvd(rdev, new_ps);
+	if (eg_pi->smu_uvd_hs)
+		btc_notify_uvd_to_smc(rdev, new_ps);
+	ret = cypress_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("cypress_upload_sw_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("cypress_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_memory_timing_parameters(rdev, new_ps);
+
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps);
+
+	ret = btc_set_power_state_conditionally_enable_ulv(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("btc_set_power_state_conditionally_enable_ulv failed\n");
+		return ret;
+	}
+
+	ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("rv770_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void btc_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+
+	btc_update_current_ps(rdev, new_ps);
+}
+
+int btc_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		btc_cg_clock_gating_default(rdev);
+
+	if (btc_dpm_enabled(rdev))
+		return -EINVAL;
+
+	if (pi->mg_clock_gating)
+		btc_mg_clock_gating_default(rdev);
+
+	if (eg_pi->ls_clock_gating)
+		btc_ls_clock_gating_default(rdev);
+
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = cypress_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+
+	if (pi->mvdd_control) {
+		ret = cypress_get_mvdd_configuration(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_get_mvdd_configuration failed\n");
+			return ret;
+		}
+	}
+
+	if (eg_pi->dynamic_ac_timing) {
+		ret = btc_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv770_enable_backbias(rdev, true);
+
+	if (pi->dynamic_ss)
+		cypress_enable_spread_spectrum(rdev, true);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	rv770_program_engine_speed_parameters(rdev);
+	cypress_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+
+	if (pi->dynamic_pcie_gen2)
+		btc_enable_dynamic_pcie_gen2(rdev, true);
+
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+	ret = cypress_get_table_locations(rdev);
+	if (ret) {
+		DRM_ERROR("cypress_get_table_locations failed\n");
+		return ret;
+	}
+	ret = btc_init_smc_table(rdev, boot_ps);
+	if (ret)
+		return ret;
+
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("cypress_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_response_times(rdev);
+	r7xx_start_smc(rdev);
+	ret = cypress_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("cypress_notify_smc_display_change failed\n");
+		return ret;
+	}
+	cypress_enable_sclk_control(rdev, true);
+
+	if (eg_pi->memory_transition)
+		cypress_enable_mclk_control(rdev, true);
+
+	cypress_start_dpm(rdev);
+
+	if (pi->gfx_clock_gating)
+		btc_cg_clock_gating_enable(rdev, true);
+
+	if (pi->mg_clock_gating)
+		btc_mg_clock_gating_enable(rdev, true);
+
+	if (eg_pi->ls_clock_gating)
+		btc_ls_clock_gating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	btc_init_stutter_mode(rdev);
+
+	btc_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+
+	return 0;
+};
+
+void btc_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (!btc_dpm_enabled(rdev))
+		return;
+
+	rv770_clear_vc(rdev);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+
+	if (pi->dynamic_pcie_gen2)
+		btc_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		btc_cg_clock_gating_enable(rdev, false);
+
+	if (pi->mg_clock_gating)
+		btc_mg_clock_gating_enable(rdev, false);
+
+	if (eg_pi->ls_clock_gating)
+		btc_ls_clock_gating_enable(rdev, false);
+
+	rv770_stop_dpm(rdev);
+	btc_reset_to_default(rdev);
+	btc_stop_smc(rdev);
+	cypress_enable_spread_spectrum(rdev, false);
+
+	btc_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+}
+
+void btc_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	rv770_get_memory_type(rdev);
+	rv740_read_clock_registers(rdev);
+	btc_read_arb_registers(rdev);
+	rv770_read_voltage_smio_registers(rdev);
+
+	if (eg_pi->pcie_performance_request)
+		cypress_advertise_gen2_capability(rdev);
+
+	rv770_get_pcie_gen2_status(rdev);
+	rv770_enable_acpi_pm(rdev);
+}
+
+int btc_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	u16 data_offset, size;
+	u8 frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	eg_pi = kzalloc(sizeof(struct evergreen_power_info), GFP_KERNEL);
+	if (eg_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = eg_pi;
+	pi = &eg_pi->rv7xx;
+
+	rv770_get_max_vddc(rdev);
+
+	eg_pi->ulv.supported = false;
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = rv7xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+	ret = r600_parse_extended_power_table(rdev);
+	if (ret)
+		return ret;
+
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries =
+		kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL);
+	if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) {
+		r600_free_extended_power_table(rdev);
+		return -ENOMEM;
+	}
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 800;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 800;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 800;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->mclk_strobe_mode_threshold = 40000;
+	pi->mclk_edc_enable_threshold = 40000;
+	eg_pi->mclk_edc_wr_enable_threshold = 40000;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[0].rlp = RV770_RLP_DFLT;
+	eg_pi->ats[0].rmp = RV770_RMP_DFLT;
+	eg_pi->ats[0].lhp = RV770_LHP_DFLT;
+	eg_pi->ats[0].lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[1].rlp = BTC_RLP_UVD_DFLT;
+	eg_pi->ats[1].rmp = BTC_RMP_UVD_DFLT;
+	eg_pi->ats[1].lhp = BTC_LHP_UVD_DFLT;
+	eg_pi->ats[1].lmp = BTC_LMP_UVD_DFLT;
+
+	eg_pi->smu_uvd_hs = true;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = CYPRESS_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+	eg_pi->ls_clock_gating = false;
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	if (rdev->flags & RADEON_IS_MOBILITY)
+		pi->dcodt = true;
+	else
+		pi->dcodt = false;
+
+	pi->ulps = true;
+
+	eg_pi->dynamic_ac_timing = true;
+	eg_pi->abm = true;
+	eg_pi->mcls = true;
+	eg_pi->light_sleep = true;
+	eg_pi->memory_transition = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	if (rdev->family == CHIP_BARTS)
+		eg_pi->dll_default_on = true;
+	else
+		eg_pi->dll_default_on = false;
+
+	eg_pi->sclk_deep_sleep = false;
+	if (ASIC_IS_LOMBOK(rdev))
+		pi->mclk_stutter_mode_threshold = 30000;
+	else
+		pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+
+	rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 4;
+	rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200;
+	rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2 = 900;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.count = ARRAY_SIZE(btc_valid_sclk);
+	rdev->pm.dpm.dyn_state.valid_sclk_values.values = btc_valid_sclk;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL;
+
+	if (rdev->family == CHIP_TURKS)
+		rdev->pm.dpm.dyn_state.sclk_mclk_delta = 15000;
+	else
+		rdev->pm.dpm.dyn_state.sclk_mclk_delta = 10000;
+
+	return 0;
+}
+
+void btc_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+	kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries);
+	r600_free_extended_power_table(rdev);
+}
+
+u32 btc_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *requested_state = rv770_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->low.sclk;
+	else
+		return requested_state->high.sclk;
+}
+
+u32 btc_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *requested_state = rv770_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->low.mclk;
+	else
+		return requested_state->high.mclk;
+}
diff --git a/drivers/gpu/drm/radeon/btc_dpm.h b/drivers/gpu/drm/radeon/btc_dpm.h
new file mode 100644
index 0000000..1a15e0e
--- /dev/null
+++ b/drivers/gpu/drm/radeon/btc_dpm.h
@@ -0,0 +1,57 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __BTC_DPM_H__
+#define __BTC_DPM_H__
+
+#define BTC_RLP_UVD_DFLT                              20
+#define BTC_RMP_UVD_DFLT                              50
+#define BTC_LHP_UVD_DFLT                              50
+#define BTC_LMP_UVD_DFLT                              20
+#define BARTS_MGCGCGTSSMCTRL_DFLT                     0x81944000
+#define TURKS_MGCGCGTSSMCTRL_DFLT                     0x6e944000
+#define CAICOS_MGCGCGTSSMCTRL_DFLT                    0x46944040
+#define BTC_CGULVPARAMETER_DFLT                       0x00040035
+#define BTC_CGULVCONTROL_DFLT                         0x00001450
+
+extern u32 btc_valid_sclk[40];
+
+void btc_read_arb_registers(struct radeon_device *rdev);
+void btc_program_mgcg_hw_sequence(struct radeon_device *rdev,
+				  const u32 *sequence, u32 count);
+void btc_skip_blacklist_clocks(struct radeon_device *rdev,
+			       const u32 max_sclk, const u32 max_mclk,
+			       u32 *sclk, u32 *mclk);
+void btc_adjust_clock_combinations(struct radeon_device *rdev,
+				   const struct radeon_clock_and_voltage_limits *max_limits,
+				   struct rv7xx_pl *pl);
+void btc_apply_voltage_dependency_rules(struct radeon_clock_voltage_dependency_table *table,
+					u32 clock, u16 max_voltage, u16 *voltage);
+void btc_apply_voltage_delta_rules(struct radeon_device *rdev,
+				   u16 max_vddc, u16 max_vddci,
+				   u16 *vddc, u16 *vddci);
+bool btc_dpm_enabled(struct radeon_device *rdev);
+int btc_reset_to_default(struct radeon_device *rdev);
+void btc_notify_uvd_to_smc(struct radeon_device *rdev,
+			   struct radeon_ps *radeon_new_state);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/btcd.h b/drivers/gpu/drm/radeon/btcd.h
new file mode 100644
index 0000000..29e32de
--- /dev/null
+++ b/drivers/gpu/drm/radeon/btcd.h
@@ -0,0 +1,181 @@
+/*
+ * Copyright 2010 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef _BTCD_H_
+#define _BTCD_H_
+
+/* pm registers */
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define	CG_BIF_REQ_AND_RSP				0x7f4
+#define		CG_CLIENT_REQ(x)			((x) << 0)
+#define		CG_CLIENT_REQ_MASK			(0xff << 0)
+#define		CG_CLIENT_REQ_SHIFT			0
+#define		CG_CLIENT_RESP(x)			((x) << 8)
+#define		CG_CLIENT_RESP_MASK			(0xff << 8)
+#define		CG_CLIENT_RESP_SHIFT			8
+#define		CLIENT_CG_REQ(x)			((x) << 16)
+#define		CLIENT_CG_REQ_MASK			(0xff << 16)
+#define		CLIENT_CG_REQ_SHIFT			16
+#define		CLIENT_CG_RESP(x)			((x) << 24)
+#define		CLIENT_CG_RESP_MASK			(0xff << 24)
+#define		CLIENT_CG_RESP_SHIFT			24
+
+#define	SCLK_PSKIP_CNTL					0x8c0
+#define		PSKIP_ON_ALLOW_STOP_HI(x)		((x) << 16)
+#define		PSKIP_ON_ALLOW_STOP_HI_MASK		(0xff << 16)
+#define		PSKIP_ON_ALLOW_STOP_HI_SHIFT		16
+
+#define	CG_ULV_CONTROL					0x8c8
+#define	CG_ULV_PARAMETER				0x8cc
+
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE0_SHIFT			0
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE1_SHIFT			8
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE2_SHIFT			16
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+#define		POWERMODE3_SHIFT			24
+
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE0_SHIFT				0
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE1_SHIFT				5
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE2_SHIFT				10
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+#define		STATE3_SHIFT				15
+
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_PMG_AUTO_CFG                                 0x28d4
+
+#define MC_SEQ_STATUS_M                                 0x29f4
+#       define PMG_PWRSTATE                             (1 << 16)
+
+#define MC_SEQ_MISC0                                    0x2a00
+#define         MC_SEQ_MISC0_GDDR5_SHIFT                28
+#define         MC_SEQ_MISC0_GDDR5_MASK                 0xf0000000
+#define         MC_SEQ_MISC0_GDDR5_VALUE                5
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
+#define MC_SEQ_MISC3                                    0x2a2c
+
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_CG                                       0x2a68
+#define		CG_SEQ_REQ(x)				((x) << 0)
+#define		CG_SEQ_REQ_MASK				(0xff << 0)
+#define		CG_SEQ_REQ_SHIFT			0
+#define		CG_SEQ_RESP(x)				((x) << 8)
+#define		CG_SEQ_RESP_MASK			(0xff << 8)
+#define		CG_SEQ_RESP_SHIFT			8
+#define		SEQ_CG_REQ(x)				((x) << 16)
+#define		SEQ_CG_REQ_MASK				(0xff << 16)
+#define		SEQ_CG_REQ_SHIFT			16
+#define		SEQ_CG_RESP(x)				((x) << 24)
+#define		SEQ_CG_RESP_MASK			(0xff << 24)
+#define		SEQ_CG_RESP_SHIFT			24
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+
+#define	LB_SYNC_RESET_SEL				0x6b28
+#define		LB_SYNC_RESET_SEL_MASK			(3 << 0)
+#define		LB_SYNC_RESET_SEL_SHIFT			0
+
+/* PCIE link stuff */
+#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
+#       define LC_GEN2_EN_STRAP                           (1 << 0)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 1)
+#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 5)
+#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 6)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
+#       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
+#       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
+#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 24)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
new file mode 100644
index 0000000..ed1d910
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -0,0 +1,6987 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#include <linux/firmware.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "radeon_asic.h"
+#include "cikd.h"
+#include "atom.h"
+#include "cik_blit_shaders.h"
+
+/* GFX */
+#define CIK_PFP_UCODE_SIZE 2144
+#define CIK_ME_UCODE_SIZE 2144
+#define CIK_CE_UCODE_SIZE 2144
+/* compute */
+#define CIK_MEC_UCODE_SIZE 4192
+/* interrupts */
+#define BONAIRE_RLC_UCODE_SIZE 2048
+#define KB_RLC_UCODE_SIZE 2560
+#define KV_RLC_UCODE_SIZE 2560
+/* gddr controller */
+#define CIK_MC_UCODE_SIZE 7866
+/* sdma */
+#define CIK_SDMA_UCODE_SIZE 1050
+#define CIK_SDMA_UCODE_VERSION 64
+
+MODULE_FIRMWARE("radeon/BONAIRE_pfp.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_me.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_ce.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_mec.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_mc.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_rlc.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_sdma.bin");
+MODULE_FIRMWARE("radeon/KAVERI_pfp.bin");
+MODULE_FIRMWARE("radeon/KAVERI_me.bin");
+MODULE_FIRMWARE("radeon/KAVERI_ce.bin");
+MODULE_FIRMWARE("radeon/KAVERI_mec.bin");
+MODULE_FIRMWARE("radeon/KAVERI_rlc.bin");
+MODULE_FIRMWARE("radeon/KAVERI_sdma.bin");
+MODULE_FIRMWARE("radeon/KABINI_pfp.bin");
+MODULE_FIRMWARE("radeon/KABINI_me.bin");
+MODULE_FIRMWARE("radeon/KABINI_ce.bin");
+MODULE_FIRMWARE("radeon/KABINI_mec.bin");
+MODULE_FIRMWARE("radeon/KABINI_rlc.bin");
+MODULE_FIRMWARE("radeon/KABINI_sdma.bin");
+
+extern int r600_ih_ring_alloc(struct radeon_device *rdev);
+extern void r600_ih_ring_fini(struct radeon_device *rdev);
+extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
+extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
+extern bool evergreen_is_display_hung(struct radeon_device *rdev);
+extern void si_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
+extern void si_rlc_fini(struct radeon_device *rdev);
+extern int si_rlc_init(struct radeon_device *rdev);
+static void cik_rlc_stop(struct radeon_device *rdev);
+
+/*
+ * Indirect registers accessor
+ */
+u32 cik_pciep_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(PCIE_INDEX, reg);
+	(void)RREG32(PCIE_INDEX);
+	r = RREG32(PCIE_DATA);
+	return r;
+}
+
+void cik_pciep_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(PCIE_INDEX, reg);
+	(void)RREG32(PCIE_INDEX);
+	WREG32(PCIE_DATA, v);
+	(void)RREG32(PCIE_DATA);
+}
+
+static const u32 bonaire_golden_spm_registers[] =
+{
+	0x30800, 0xe0ffffff, 0xe0000000
+};
+
+static const u32 bonaire_golden_common_registers[] =
+{
+	0xc770, 0xffffffff, 0x00000800,
+	0xc774, 0xffffffff, 0x00000800,
+	0xc798, 0xffffffff, 0x00007fbf,
+	0xc79c, 0xffffffff, 0x00007faf
+};
+
+static const u32 bonaire_golden_registers[] =
+{
+	0x3354, 0x00000333, 0x00000333,
+	0x3350, 0x000c0fc0, 0x00040200,
+	0x9a10, 0x00010000, 0x00058208,
+	0x3c000, 0xffff1fff, 0x00140000,
+	0x3c200, 0xfdfc0fff, 0x00000100,
+	0x3c234, 0x40000000, 0x40000200,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0x0002021c, 0x00020200,
+	0xc78, 0x00000080, 0x00000000,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0xf0311fff, 0x80300000,
+	0x98f8, 0x73773777, 0x12010001,
+	0x350c, 0x00810000, 0x408af000,
+	0x7030, 0x31000111, 0x00000011,
+	0x2f48, 0x73773777, 0x12010001,
+	0x220c, 0x00007fb6, 0x0021a1b1,
+	0x2210, 0x00007fb6, 0x002021b1,
+	0x2180, 0x00007fb6, 0x00002191,
+	0x2218, 0x00007fb6, 0x002121b1,
+	0x221c, 0x00007fb6, 0x002021b1,
+	0x21dc, 0x00007fb6, 0x00002191,
+	0x21e0, 0x00007fb6, 0x00002191,
+	0x3628, 0x0000003f, 0x0000000a,
+	0x362c, 0x0000003f, 0x0000000a,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8bf0, 0x00002001, 0x00000001,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x30a04, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x4d8, 0x00000fff, 0x00000100,
+	0x3e78, 0x00000001, 0x00000002,
+	0x9100, 0x03000000, 0x0362c688,
+	0x8c00, 0x000000ff, 0x00000001,
+	0xe40, 0x00001fff, 0x00001fff,
+	0x9060, 0x0000007f, 0x00000020,
+	0x9508, 0x00010000, 0x00010000,
+	0xac14, 0x000003ff, 0x000000f3,
+	0xac0c, 0xffffffff, 0x00001032
+};
+
+static const u32 bonaire_mgcg_cgcg_init[] =
+{
+	0xc420, 0xffffffff, 0xfffffffc,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c2a0, 0xffffffff, 0x00000100,
+	0x3c208, 0xffffffff, 0x00000100,
+	0x3c2c0, 0xffffffff, 0xc0000100,
+	0x3c2c8, 0xffffffff, 0xc0000100,
+	0x3c2c4, 0xffffffff, 0xc0000100,
+	0x55e4, 0xffffffff, 0x00600100,
+	0x3c280, 0xffffffff, 0x00000100,
+	0x3c214, 0xffffffff, 0x06000100,
+	0x3c220, 0xffffffff, 0x00000100,
+	0x3c218, 0xffffffff, 0x06000100,
+	0x3c204, 0xffffffff, 0x00000100,
+	0x3c2e0, 0xffffffff, 0x00000100,
+	0x3c224, 0xffffffff, 0x00000100,
+	0x3c200, 0xffffffff, 0x00000100,
+	0x3c230, 0xffffffff, 0x00000100,
+	0x3c234, 0xffffffff, 0x00000100,
+	0x3c250, 0xffffffff, 0x00000100,
+	0x3c254, 0xffffffff, 0x00000100,
+	0x3c258, 0xffffffff, 0x00000100,
+	0x3c25c, 0xffffffff, 0x00000100,
+	0x3c260, 0xffffffff, 0x00000100,
+	0x3c27c, 0xffffffff, 0x00000100,
+	0x3c278, 0xffffffff, 0x00000100,
+	0x3c210, 0xffffffff, 0x06000100,
+	0x3c290, 0xffffffff, 0x00000100,
+	0x3c274, 0xffffffff, 0x00000100,
+	0x3c2b4, 0xffffffff, 0x00000100,
+	0x3c2b0, 0xffffffff, 0x00000100,
+	0x3c270, 0xffffffff, 0x00000100,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c020, 0xffffffff, 0x00010000,
+	0x3c024, 0xffffffff, 0x00030002,
+	0x3c028, 0xffffffff, 0x00040007,
+	0x3c02c, 0xffffffff, 0x00060005,
+	0x3c030, 0xffffffff, 0x00090008,
+	0x3c034, 0xffffffff, 0x00010000,
+	0x3c038, 0xffffffff, 0x00030002,
+	0x3c03c, 0xffffffff, 0x00040007,
+	0x3c040, 0xffffffff, 0x00060005,
+	0x3c044, 0xffffffff, 0x00090008,
+	0x3c048, 0xffffffff, 0x00010000,
+	0x3c04c, 0xffffffff, 0x00030002,
+	0x3c050, 0xffffffff, 0x00040007,
+	0x3c054, 0xffffffff, 0x00060005,
+	0x3c058, 0xffffffff, 0x00090008,
+	0x3c05c, 0xffffffff, 0x00010000,
+	0x3c060, 0xffffffff, 0x00030002,
+	0x3c064, 0xffffffff, 0x00040007,
+	0x3c068, 0xffffffff, 0x00060005,
+	0x3c06c, 0xffffffff, 0x00090008,
+	0x3c070, 0xffffffff, 0x00010000,
+	0x3c074, 0xffffffff, 0x00030002,
+	0x3c078, 0xffffffff, 0x00040007,
+	0x3c07c, 0xffffffff, 0x00060005,
+	0x3c080, 0xffffffff, 0x00090008,
+	0x3c084, 0xffffffff, 0x00010000,
+	0x3c088, 0xffffffff, 0x00030002,
+	0x3c08c, 0xffffffff, 0x00040007,
+	0x3c090, 0xffffffff, 0x00060005,
+	0x3c094, 0xffffffff, 0x00090008,
+	0x3c098, 0xffffffff, 0x00010000,
+	0x3c09c, 0xffffffff, 0x00030002,
+	0x3c0a0, 0xffffffff, 0x00040007,
+	0x3c0a4, 0xffffffff, 0x00060005,
+	0x3c0a8, 0xffffffff, 0x00090008,
+	0x3c000, 0xffffffff, 0x96e00200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc424, 0xffffffff, 0x0020003f,
+	0x38, 0xffffffff, 0x0140001c,
+	0x3c, 0x000f0000, 0x000f0000,
+	0x220, 0xffffffff, 0xC060000C,
+	0x224, 0xc0000fff, 0x00000100,
+	0xf90, 0xffffffff, 0x00000100,
+	0xf98, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd00c, 0xff000ff0, 0x00000100,
+	0xd80c, 0xff000ff0, 0x00000100
+};
+
+static const u32 spectre_golden_spm_registers[] =
+{
+	0x30800, 0xe0ffffff, 0xe0000000
+};
+
+static const u32 spectre_golden_common_registers[] =
+{
+	0xc770, 0xffffffff, 0x00000800,
+	0xc774, 0xffffffff, 0x00000800,
+	0xc798, 0xffffffff, 0x00007fbf,
+	0xc79c, 0xffffffff, 0x00007faf
+};
+
+static const u32 spectre_golden_registers[] =
+{
+	0x3c000, 0xffff1fff, 0x96940200,
+	0x3c00c, 0xffff0001, 0xff000000,
+	0x3c200, 0xfffc0fff, 0x00000100,
+	0x6ed8, 0x00010101, 0x00010000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0xfffffffc, 0x00020200,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0xf0311fff, 0x80300000,
+	0x98f8, 0x73773777, 0x12010001,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x2f48, 0x73773777, 0x12010001,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x28350, 0x3f3f3fff, 0x00000082,
+	0x28355, 0x0000003f, 0x00000000,
+	0x3e78, 0x00000001, 0x00000002,
+	0x913c, 0xffff03df, 0x00000004,
+	0xc768, 0x00000008, 0x00000008,
+	0x8c00, 0x000008ff, 0x00000800,
+	0x9508, 0x00010000, 0x00010000,
+	0xac0c, 0xffffffff, 0x54763210,
+	0x214f8, 0x01ff01ff, 0x00000002,
+	0x21498, 0x007ff800, 0x00200000,
+	0x2015c, 0xffffffff, 0x00000f40,
+	0x30934, 0xffffffff, 0x00000001
+};
+
+static const u32 spectre_mgcg_cgcg_init[] =
+{
+	0xc420, 0xffffffff, 0xfffffffc,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c2a0, 0xffffffff, 0x00000100,
+	0x3c208, 0xffffffff, 0x00000100,
+	0x3c2c0, 0xffffffff, 0x00000100,
+	0x3c2c8, 0xffffffff, 0x00000100,
+	0x3c2c4, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00600100,
+	0x3c280, 0xffffffff, 0x00000100,
+	0x3c214, 0xffffffff, 0x06000100,
+	0x3c220, 0xffffffff, 0x00000100,
+	0x3c218, 0xffffffff, 0x06000100,
+	0x3c204, 0xffffffff, 0x00000100,
+	0x3c2e0, 0xffffffff, 0x00000100,
+	0x3c224, 0xffffffff, 0x00000100,
+	0x3c200, 0xffffffff, 0x00000100,
+	0x3c230, 0xffffffff, 0x00000100,
+	0x3c234, 0xffffffff, 0x00000100,
+	0x3c250, 0xffffffff, 0x00000100,
+	0x3c254, 0xffffffff, 0x00000100,
+	0x3c258, 0xffffffff, 0x00000100,
+	0x3c25c, 0xffffffff, 0x00000100,
+	0x3c260, 0xffffffff, 0x00000100,
+	0x3c27c, 0xffffffff, 0x00000100,
+	0x3c278, 0xffffffff, 0x00000100,
+	0x3c210, 0xffffffff, 0x06000100,
+	0x3c290, 0xffffffff, 0x00000100,
+	0x3c274, 0xffffffff, 0x00000100,
+	0x3c2b4, 0xffffffff, 0x00000100,
+	0x3c2b0, 0xffffffff, 0x00000100,
+	0x3c270, 0xffffffff, 0x00000100,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c020, 0xffffffff, 0x00010000,
+	0x3c024, 0xffffffff, 0x00030002,
+	0x3c028, 0xffffffff, 0x00040007,
+	0x3c02c, 0xffffffff, 0x00060005,
+	0x3c030, 0xffffffff, 0x00090008,
+	0x3c034, 0xffffffff, 0x00010000,
+	0x3c038, 0xffffffff, 0x00030002,
+	0x3c03c, 0xffffffff, 0x00040007,
+	0x3c040, 0xffffffff, 0x00060005,
+	0x3c044, 0xffffffff, 0x00090008,
+	0x3c048, 0xffffffff, 0x00010000,
+	0x3c04c, 0xffffffff, 0x00030002,
+	0x3c050, 0xffffffff, 0x00040007,
+	0x3c054, 0xffffffff, 0x00060005,
+	0x3c058, 0xffffffff, 0x00090008,
+	0x3c05c, 0xffffffff, 0x00010000,
+	0x3c060, 0xffffffff, 0x00030002,
+	0x3c064, 0xffffffff, 0x00040007,
+	0x3c068, 0xffffffff, 0x00060005,
+	0x3c06c, 0xffffffff, 0x00090008,
+	0x3c070, 0xffffffff, 0x00010000,
+	0x3c074, 0xffffffff, 0x00030002,
+	0x3c078, 0xffffffff, 0x00040007,
+	0x3c07c, 0xffffffff, 0x00060005,
+	0x3c080, 0xffffffff, 0x00090008,
+	0x3c084, 0xffffffff, 0x00010000,
+	0x3c088, 0xffffffff, 0x00030002,
+	0x3c08c, 0xffffffff, 0x00040007,
+	0x3c090, 0xffffffff, 0x00060005,
+	0x3c094, 0xffffffff, 0x00090008,
+	0x3c098, 0xffffffff, 0x00010000,
+	0x3c09c, 0xffffffff, 0x00030002,
+	0x3c0a0, 0xffffffff, 0x00040007,
+	0x3c0a4, 0xffffffff, 0x00060005,
+	0x3c0a8, 0xffffffff, 0x00090008,
+	0x3c0ac, 0xffffffff, 0x00010000,
+	0x3c0b0, 0xffffffff, 0x00030002,
+	0x3c0b4, 0xffffffff, 0x00040007,
+	0x3c0b8, 0xffffffff, 0x00060005,
+	0x3c0bc, 0xffffffff, 0x00090008,
+	0x3c000, 0xffffffff, 0x96e00200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc424, 0xffffffff, 0x0020003f,
+	0x38, 0xffffffff, 0x0140001c,
+	0x3c, 0x000f0000, 0x000f0000,
+	0x220, 0xffffffff, 0xC060000C,
+	0x224, 0xc0000fff, 0x00000100,
+	0xf90, 0xffffffff, 0x00000100,
+	0xf98, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd00c, 0xff000ff0, 0x00000100,
+	0xd80c, 0xff000ff0, 0x00000100
+};
+
+static const u32 kalindi_golden_spm_registers[] =
+{
+	0x30800, 0xe0ffffff, 0xe0000000
+};
+
+static const u32 kalindi_golden_common_registers[] =
+{
+	0xc770, 0xffffffff, 0x00000800,
+	0xc774, 0xffffffff, 0x00000800,
+	0xc798, 0xffffffff, 0x00007fbf,
+	0xc79c, 0xffffffff, 0x00007faf
+};
+
+static const u32 kalindi_golden_registers[] =
+{
+	0x3c000, 0xffffdfff, 0x6e944040,
+	0x55e4, 0xff607fff, 0xfc000100,
+	0x3c220, 0xff000fff, 0x00000100,
+	0x3c224, 0xff000fff, 0x00000100,
+	0x3c200, 0xfffc0fff, 0x00000100,
+	0x6ed8, 0x00010101, 0x00010000,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0xf0311fff, 0x80300000,
+	0x98f8, 0x73773777, 0x12010001,
+	0x98fc, 0xffffffff, 0x00000010,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x8030, 0x00001f0f, 0x0000100a,
+	0x2f48, 0x73773777, 0x12010001,
+	0x2408, 0x000fffff, 0x000c007f,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ffcfff,
+	0x30a04, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x4d8, 0x00000fff, 0x00000100,
+	0x3e78, 0x00000001, 0x00000002,
+	0xc768, 0x00000008, 0x00000008,
+	0x8c00, 0x000000ff, 0x00000003,
+	0x214f8, 0x01ff01ff, 0x00000002,
+	0x21498, 0x007ff800, 0x00200000,
+	0x2015c, 0xffffffff, 0x00000f40,
+	0x88c4, 0x001f3ae3, 0x00000082,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x30934, 0xffffffff, 0x00000000
+};
+
+static const u32 kalindi_mgcg_cgcg_init[] =
+{
+	0xc420, 0xffffffff, 0xfffffffc,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c2a0, 0xffffffff, 0x00000100,
+	0x3c208, 0xffffffff, 0x00000100,
+	0x3c2c0, 0xffffffff, 0x00000100,
+	0x3c2c8, 0xffffffff, 0x00000100,
+	0x3c2c4, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00600100,
+	0x3c280, 0xffffffff, 0x00000100,
+	0x3c214, 0xffffffff, 0x06000100,
+	0x3c220, 0xffffffff, 0x00000100,
+	0x3c218, 0xffffffff, 0x06000100,
+	0x3c204, 0xffffffff, 0x00000100,
+	0x3c2e0, 0xffffffff, 0x00000100,
+	0x3c224, 0xffffffff, 0x00000100,
+	0x3c200, 0xffffffff, 0x00000100,
+	0x3c230, 0xffffffff, 0x00000100,
+	0x3c234, 0xffffffff, 0x00000100,
+	0x3c250, 0xffffffff, 0x00000100,
+	0x3c254, 0xffffffff, 0x00000100,
+	0x3c258, 0xffffffff, 0x00000100,
+	0x3c25c, 0xffffffff, 0x00000100,
+	0x3c260, 0xffffffff, 0x00000100,
+	0x3c27c, 0xffffffff, 0x00000100,
+	0x3c278, 0xffffffff, 0x00000100,
+	0x3c210, 0xffffffff, 0x06000100,
+	0x3c290, 0xffffffff, 0x00000100,
+	0x3c274, 0xffffffff, 0x00000100,
+	0x3c2b4, 0xffffffff, 0x00000100,
+	0x3c2b0, 0xffffffff, 0x00000100,
+	0x3c270, 0xffffffff, 0x00000100,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c020, 0xffffffff, 0x00010000,
+	0x3c024, 0xffffffff, 0x00030002,
+	0x3c028, 0xffffffff, 0x00040007,
+	0x3c02c, 0xffffffff, 0x00060005,
+	0x3c030, 0xffffffff, 0x00090008,
+	0x3c034, 0xffffffff, 0x00010000,
+	0x3c038, 0xffffffff, 0x00030002,
+	0x3c03c, 0xffffffff, 0x00040007,
+	0x3c040, 0xffffffff, 0x00060005,
+	0x3c044, 0xffffffff, 0x00090008,
+	0x3c000, 0xffffffff, 0x96e00200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc424, 0xffffffff, 0x0020003f,
+	0x38, 0xffffffff, 0x0140001c,
+	0x3c, 0x000f0000, 0x000f0000,
+	0x220, 0xffffffff, 0xC060000C,
+	0x224, 0xc0000fff, 0x00000100,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd00c, 0xff000ff0, 0x00000100,
+	0xd80c, 0xff000ff0, 0x00000100
+};
+
+static void cik_init_golden_registers(struct radeon_device *rdev)
+{
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+		radeon_program_register_sequence(rdev,
+						 bonaire_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(bonaire_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 bonaire_golden_registers,
+						 (const u32)ARRAY_SIZE(bonaire_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 bonaire_golden_common_registers,
+						 (const u32)ARRAY_SIZE(bonaire_golden_common_registers));
+		radeon_program_register_sequence(rdev,
+						 bonaire_golden_spm_registers,
+						 (const u32)ARRAY_SIZE(bonaire_golden_spm_registers));
+		break;
+	case CHIP_KABINI:
+		radeon_program_register_sequence(rdev,
+						 kalindi_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(kalindi_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 kalindi_golden_registers,
+						 (const u32)ARRAY_SIZE(kalindi_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 kalindi_golden_common_registers,
+						 (const u32)ARRAY_SIZE(kalindi_golden_common_registers));
+		radeon_program_register_sequence(rdev,
+						 kalindi_golden_spm_registers,
+						 (const u32)ARRAY_SIZE(kalindi_golden_spm_registers));
+		break;
+	case CHIP_KAVERI:
+		radeon_program_register_sequence(rdev,
+						 spectre_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(spectre_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 spectre_golden_registers,
+						 (const u32)ARRAY_SIZE(spectre_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 spectre_golden_common_registers,
+						 (const u32)ARRAY_SIZE(spectre_golden_common_registers));
+		radeon_program_register_sequence(rdev,
+						 spectre_golden_spm_registers,
+						 (const u32)ARRAY_SIZE(spectre_golden_spm_registers));
+		break;
+	default:
+		break;
+	}
+}
+
+/**
+ * cik_get_xclk - get the xclk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Returns the reference clock used by the gfx engine
+ * (CIK).
+ */
+u32 cik_get_xclk(struct radeon_device *rdev)
+{
+        u32 reference_clock = rdev->clock.spll.reference_freq;
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (RREG32_SMC(GENERAL_PWRMGT) & GPU_COUNTER_CLK)
+			return reference_clock / 2;
+	} else {
+		if (RREG32_SMC(CG_CLKPIN_CNTL) & XTALIN_DIVIDE)
+			return reference_clock / 4;
+	}
+	return reference_clock;
+}
+
+/**
+ * cik_mm_rdoorbell - read a doorbell dword
+ *
+ * @rdev: radeon_device pointer
+ * @offset: byte offset into the aperture
+ *
+ * Returns the value in the doorbell aperture at the
+ * requested offset (CIK).
+ */
+u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset)
+{
+	if (offset < rdev->doorbell.size) {
+		return readl(((void __iomem *)rdev->doorbell.ptr) + offset);
+	} else {
+		DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", offset);
+		return 0;
+	}
+}
+
+/**
+ * cik_mm_wdoorbell - write a doorbell dword
+ *
+ * @rdev: radeon_device pointer
+ * @offset: byte offset into the aperture
+ * @v: value to write
+ *
+ * Writes @v to the doorbell aperture at the
+ * requested offset (CIK).
+ */
+void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v)
+{
+	if (offset < rdev->doorbell.size) {
+		writel(v, ((void __iomem *)rdev->doorbell.ptr) + offset);
+	} else {
+		DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", offset);
+	}
+}
+
+#define BONAIRE_IO_MC_REGS_SIZE 36
+
+static const u32 bonaire_io_mc_regs[BONAIRE_IO_MC_REGS_SIZE][2] =
+{
+	{0x00000070, 0x04400000},
+	{0x00000071, 0x80c01803},
+	{0x00000072, 0x00004004},
+	{0x00000073, 0x00000100},
+	{0x00000074, 0x00ff0000},
+	{0x00000075, 0x34000000},
+	{0x00000076, 0x08000014},
+	{0x00000077, 0x00cc08ec},
+	{0x00000078, 0x00000400},
+	{0x00000079, 0x00000000},
+	{0x0000007a, 0x04090000},
+	{0x0000007c, 0x00000000},
+	{0x0000007e, 0x4408a8e8},
+	{0x0000007f, 0x00000304},
+	{0x00000080, 0x00000000},
+	{0x00000082, 0x00000001},
+	{0x00000083, 0x00000002},
+	{0x00000084, 0xf3e4f400},
+	{0x00000085, 0x052024e3},
+	{0x00000087, 0x00000000},
+	{0x00000088, 0x01000000},
+	{0x0000008a, 0x1c0a0000},
+	{0x0000008b, 0xff010000},
+	{0x0000008d, 0xffffefff},
+	{0x0000008e, 0xfff3efff},
+	{0x0000008f, 0xfff3efbf},
+	{0x00000092, 0xf7ffffff},
+	{0x00000093, 0xffffff7f},
+	{0x00000095, 0x00101101},
+	{0x00000096, 0x00000fff},
+	{0x00000097, 0x00116fff},
+	{0x00000098, 0x60010000},
+	{0x00000099, 0x10010000},
+	{0x0000009a, 0x00006000},
+	{0x0000009b, 0x00001000},
+	{0x0000009f, 0x00b48000}
+};
+
+/**
+ * cik_srbm_select - select specific register instances
+ *
+ * @rdev: radeon_device pointer
+ * @me: selected ME (micro engine)
+ * @pipe: pipe
+ * @queue: queue
+ * @vmid: VMID
+ *
+ * Switches the currently active registers instances.  Some
+ * registers are instanced per VMID, others are instanced per
+ * me/pipe/queue combination.
+ */
+static void cik_srbm_select(struct radeon_device *rdev,
+			    u32 me, u32 pipe, u32 queue, u32 vmid)
+{
+	u32 srbm_gfx_cntl = (PIPEID(pipe & 0x3) |
+			     MEID(me & 0x3) |
+			     VMID(vmid & 0xf) |
+			     QUEUEID(queue & 0x7));
+	WREG32(SRBM_GFX_CNTL, srbm_gfx_cntl);
+}
+
+/* ucode loading */
+/**
+ * ci_mc_load_microcode - load MC ucode into the hw
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Load the GDDR MC ucode into the hw (CIK).
+ * Returns 0 on success, error on failure.
+ */
+static int ci_mc_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	u32 running, blackout = 0;
+	u32 *io_mc_regs;
+	int i, ucode_size, regs_size;
+
+	if (!rdev->mc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+	default:
+		io_mc_regs = (u32 *)&bonaire_io_mc_regs;
+		ucode_size = CIK_MC_UCODE_SIZE;
+		regs_size = BONAIRE_IO_MC_REGS_SIZE;
+		break;
+	}
+
+	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
+
+	if (running == 0) {
+		if (running) {
+			blackout = RREG32(MC_SHARED_BLACKOUT_CNTL);
+			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout | 1);
+		}
+
+		/* reset the engine and set to writable */
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000010);
+
+		/* load mc io regs */
+		for (i = 0; i < regs_size; i++) {
+			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+		}
+		/* load the MC ucode */
+		fw_data = (const __be32 *)rdev->mc_fw->data;
+		for (i = 0; i < ucode_size; i++)
+			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+
+		/* put the engine back into the active state */
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000004);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000001);
+
+		/* wait for training to complete */
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D0)
+				break;
+			udelay(1);
+		}
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D1)
+				break;
+			udelay(1);
+		}
+
+		if (running)
+			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout);
+	}
+
+	return 0;
+}
+
+/**
+ * cik_init_microcode - load ucode images from disk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Use the firmware interface to load the ucode images into
+ * the driver (not loaded into hw).
+ * Returns 0 on success, error on failure.
+ */
+static int cik_init_microcode(struct radeon_device *rdev)
+{
+	struct platform_device *pdev;
+	const char *chip_name;
+	size_t pfp_req_size, me_req_size, ce_req_size,
+		mec_req_size, rlc_req_size, mc_req_size,
+		sdma_req_size;
+	char fw_name[30];
+	int err;
+
+	DRM_DEBUG("\n");
+
+	pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0);
+	err = IS_ERR(pdev);
+	if (err) {
+		printk(KERN_ERR "radeon_cp: Failed to register firmware\n");
+		return -EINVAL;
+	}
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+		chip_name = "BONAIRE";
+		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
+		me_req_size = CIK_ME_UCODE_SIZE * 4;
+		ce_req_size = CIK_CE_UCODE_SIZE * 4;
+		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
+		rlc_req_size = BONAIRE_RLC_UCODE_SIZE * 4;
+		mc_req_size = CIK_MC_UCODE_SIZE * 4;
+		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		break;
+	case CHIP_KAVERI:
+		chip_name = "KAVERI";
+		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
+		me_req_size = CIK_ME_UCODE_SIZE * 4;
+		ce_req_size = CIK_CE_UCODE_SIZE * 4;
+		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
+		rlc_req_size = KV_RLC_UCODE_SIZE * 4;
+		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		break;
+	case CHIP_KABINI:
+		chip_name = "KABINI";
+		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
+		me_req_size = CIK_ME_UCODE_SIZE * 4;
+		ce_req_size = CIK_CE_UCODE_SIZE * 4;
+		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
+		rlc_req_size = KB_RLC_UCODE_SIZE * 4;
+		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		break;
+	default: BUG();
+	}
+
+	DRM_INFO("Loading %s Microcode\n", chip_name);
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->pfp_fw->size != pfp_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->pfp_fw->size, fw_name);
+		err = -EINVAL;
+		goto out;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->me_fw->size != me_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->me_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->ce_fw->size != ce_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->ce_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name);
+	err = request_firmware(&rdev->mec_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->mec_fw->size != mec_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->mec_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+	err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->rlc_fw->size != rlc_req_size) {
+		printk(KERN_ERR
+		       "cik_rlc: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->rlc_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name);
+	err = request_firmware(&rdev->sdma_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->sdma_fw->size != sdma_req_size) {
+		printk(KERN_ERR
+		       "cik_sdma: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->sdma_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	/* No MC ucode on APUs */
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+		err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev);
+		if (err)
+			goto out;
+		if (rdev->mc_fw->size != mc_req_size) {
+			printk(KERN_ERR
+			       "cik_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	}
+
+out:
+	platform_device_unregister(pdev);
+
+	if (err) {
+		if (err != -EINVAL)
+			printk(KERN_ERR
+			       "cik_cp: Failed to load firmware \"%s\"\n",
+			       fw_name);
+		release_firmware(rdev->pfp_fw);
+		rdev->pfp_fw = NULL;
+		release_firmware(rdev->me_fw);
+		rdev->me_fw = NULL;
+		release_firmware(rdev->ce_fw);
+		rdev->ce_fw = NULL;
+		release_firmware(rdev->rlc_fw);
+		rdev->rlc_fw = NULL;
+		release_firmware(rdev->mc_fw);
+		rdev->mc_fw = NULL;
+	}
+	return err;
+}
+
+/*
+ * Core functions
+ */
+/**
+ * cik_tiling_mode_table_init - init the hw tiling table
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Starting with SI, the tiling setup is done globally in a
+ * set of 32 tiling modes.  Rather than selecting each set of
+ * parameters per surface as on older asics, we just select
+ * which index in the tiling table we want to use, and the
+ * surface uses those parameters (CIK).
+ */
+static void cik_tiling_mode_table_init(struct radeon_device *rdev)
+{
+	const u32 num_tile_mode_states = 32;
+	const u32 num_secondary_tile_mode_states = 16;
+	u32 reg_offset, gb_tile_moden, split_equal_to_row_size;
+	u32 num_pipe_configs;
+	u32 num_rbs = rdev->config.cik.max_backends_per_se *
+		rdev->config.cik.max_shader_engines;
+
+	switch (rdev->config.cik.mem_row_size_in_kb) {
+	case 1:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_1KB;
+		break;
+	case 2:
+	default:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_2KB;
+		break;
+	case 4:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_4KB;
+		break;
+	}
+
+	num_pipe_configs = rdev->config.cik.max_tile_pipes;
+	if (num_pipe_configs > 8)
+		num_pipe_configs = 8; /* ??? */
+
+	if (num_pipe_configs == 8) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+				break;
+			case 1:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+				break;
+			case 2:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 3:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+				break;
+			case 4:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 5:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+				break;
+			case 6:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 7:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 8:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16));
+				break;
+			case 9:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+				break;
+			case 10:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 11:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 12:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 13:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+				break;
+			case 14:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 16:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 17:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 27:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+				break;
+			case 28:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 29:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 30:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+		for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 1:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 2:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 3:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 4:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 5:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			case 6:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_2_BANK));
+				break;
+			case 8:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 9:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 10:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 11:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 12:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 13:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			case 14:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_2_BANK));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else if (num_pipe_configs == 4) {
+		if (num_rbs == 4) {
+			for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+				switch (reg_offset) {
+				case 0:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+					break;
+				case 1:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+					break;
+				case 2:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 3:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+					break;
+				case 4:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 5:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+					break;
+				case 6:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 7:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 8:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16));
+					break;
+				case 9:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+					break;
+				case 10:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 11:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 12:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 13:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+					break;
+				case 14:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 16:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 17:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 27:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+					break;
+				case 28:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 29:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 30:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				default:
+					gb_tile_moden = 0;
+					break;
+				}
+				rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+				WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+			}
+		} else if (num_rbs < 4) {
+			for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+				switch (reg_offset) {
+				case 0:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+					break;
+				case 1:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+					break;
+				case 2:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 3:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+					break;
+				case 4:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 5:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+					break;
+				case 6:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 7:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 8:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16));
+					break;
+				case 9:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+					break;
+				case 10:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 11:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 12:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 13:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+					break;
+				case 14:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 16:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 17:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 27:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+					break;
+				case 28:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 29:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 30:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				default:
+					gb_tile_moden = 0;
+					break;
+				}
+				rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+				WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+			}
+		}
+		for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 1:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 2:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 3:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 4:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 5:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 6:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			case 8:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 9:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 10:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 11:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 12:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 13:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 14:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else if (num_pipe_configs == 2) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+				break;
+			case 1:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+				break;
+			case 2:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 3:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+				break;
+			case 4:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 5:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+				break;
+			case 6:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 7:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 8:
+				gb_tile_moden = ARRAY_MODE(ARRAY_LINEAR_ALIGNED);
+				break;
+			case 9:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+				break;
+			case 10:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 11:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 12:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 13:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+				break;
+			case 14:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 16:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 17:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 27:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+				break;
+			case 28:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 29:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 30:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+		for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 1:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 2:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 3:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 4:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 5:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 6:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 8:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 9:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 10:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 11:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 12:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 13:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 14:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else
+		DRM_ERROR("unknown num pipe config: 0x%x\n", num_pipe_configs);
+}
+
+/**
+ * cik_select_se_sh - select which SE, SH to address
+ *
+ * @rdev: radeon_device pointer
+ * @se_num: shader engine to address
+ * @sh_num: sh block to address
+ *
+ * Select which SE, SH combinations to address. Certain
+ * registers are instanced per SE or SH.  0xffffffff means
+ * broadcast to all SEs or SHs (CIK).
+ */
+static void cik_select_se_sh(struct radeon_device *rdev,
+			     u32 se_num, u32 sh_num)
+{
+	u32 data = INSTANCE_BROADCAST_WRITES;
+
+	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff))
+		data |= SH_BROADCAST_WRITES | SE_BROADCAST_WRITES;
+	else if (se_num == 0xffffffff)
+		data |= SE_BROADCAST_WRITES | SH_INDEX(sh_num);
+	else if (sh_num == 0xffffffff)
+		data |= SH_BROADCAST_WRITES | SE_INDEX(se_num);
+	else
+		data |= SH_INDEX(sh_num) | SE_INDEX(se_num);
+	WREG32(GRBM_GFX_INDEX, data);
+}
+
+/**
+ * cik_create_bitmask - create a bitmask
+ *
+ * @bit_width: length of the mask
+ *
+ * create a variable length bit mask (CIK).
+ * Returns the bitmask.
+ */
+static u32 cik_create_bitmask(u32 bit_width)
+{
+	u32 i, mask = 0;
+
+	for (i = 0; i < bit_width; i++) {
+		mask <<= 1;
+		mask |= 1;
+	}
+	return mask;
+}
+
+/**
+ * cik_select_se_sh - select which SE, SH to address
+ *
+ * @rdev: radeon_device pointer
+ * @max_rb_num: max RBs (render backends) for the asic
+ * @se_num: number of SEs (shader engines) for the asic
+ * @sh_per_se: number of SH blocks per SE for the asic
+ *
+ * Calculates the bitmask of disabled RBs (CIK).
+ * Returns the disabled RB bitmask.
+ */
+static u32 cik_get_rb_disabled(struct radeon_device *rdev,
+			      u32 max_rb_num, u32 se_num,
+			      u32 sh_per_se)
+{
+	u32 data, mask;
+
+	data = RREG32(CC_RB_BACKEND_DISABLE);
+	if (data & 1)
+		data &= BACKEND_DISABLE_MASK;
+	else
+		data = 0;
+	data |= RREG32(GC_USER_RB_BACKEND_DISABLE);
+
+	data >>= BACKEND_DISABLE_SHIFT;
+
+	mask = cik_create_bitmask(max_rb_num / se_num / sh_per_se);
+
+	return data & mask;
+}
+
+/**
+ * cik_setup_rb - setup the RBs on the asic
+ *
+ * @rdev: radeon_device pointer
+ * @se_num: number of SEs (shader engines) for the asic
+ * @sh_per_se: number of SH blocks per SE for the asic
+ * @max_rb_num: max RBs (render backends) for the asic
+ *
+ * Configures per-SE/SH RB registers (CIK).
+ */
+static void cik_setup_rb(struct radeon_device *rdev,
+			 u32 se_num, u32 sh_per_se,
+			 u32 max_rb_num)
+{
+	int i, j;
+	u32 data, mask;
+	u32 disabled_rbs = 0;
+	u32 enabled_rbs = 0;
+
+	for (i = 0; i < se_num; i++) {
+		for (j = 0; j < sh_per_se; j++) {
+			cik_select_se_sh(rdev, i, j);
+			data = cik_get_rb_disabled(rdev, max_rb_num, se_num, sh_per_se);
+			disabled_rbs |= data << ((i * sh_per_se + j) * CIK_RB_BITMAP_WIDTH_PER_SH);
+		}
+	}
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+
+	mask = 1;
+	for (i = 0; i < max_rb_num; i++) {
+		if (!(disabled_rbs & mask))
+			enabled_rbs |= mask;
+		mask <<= 1;
+	}
+
+	for (i = 0; i < se_num; i++) {
+		cik_select_se_sh(rdev, i, 0xffffffff);
+		data = 0;
+		for (j = 0; j < sh_per_se; j++) {
+			switch (enabled_rbs & 3) {
+			case 1:
+				data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2);
+				break;
+			case 2:
+				data |= (RASTER_CONFIG_RB_MAP_3 << (i * sh_per_se + j) * 2);
+				break;
+			case 3:
+			default:
+				data |= (RASTER_CONFIG_RB_MAP_2 << (i * sh_per_se + j) * 2);
+				break;
+			}
+			enabled_rbs >>= 2;
+		}
+		WREG32(PA_SC_RASTER_CONFIG, data);
+	}
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+}
+
+/**
+ * cik_gpu_init - setup the 3D engine
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Configures the 3D engine and tiling configuration
+ * registers so that the 3D engine is usable.
+ */
+static void cik_gpu_init(struct radeon_device *rdev)
+{
+	u32 gb_addr_config = RREG32(GB_ADDR_CONFIG);
+	u32 mc_shared_chmap, mc_arb_ramcfg;
+	u32 hdp_host_path_cntl;
+	u32 tmp;
+	int i, j;
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+		rdev->config.cik.max_shader_engines = 2;
+		rdev->config.cik.max_tile_pipes = 4;
+		rdev->config.cik.max_cu_per_sh = 7;
+		rdev->config.cik.max_sh_per_se = 1;
+		rdev->config.cik.max_backends_per_se = 2;
+		rdev->config.cik.max_texture_channel_caches = 4;
+		rdev->config.cik.max_gprs = 256;
+		rdev->config.cik.max_gs_threads = 32;
+		rdev->config.cik.max_hw_contexts = 8;
+
+		rdev->config.cik.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.cik.sc_prim_fifo_size_backend = 0x100;
+		rdev->config.cik.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	case CHIP_KAVERI:
+		/* TODO */
+		break;
+	case CHIP_KABINI:
+	default:
+		rdev->config.cik.max_shader_engines = 1;
+		rdev->config.cik.max_tile_pipes = 2;
+		rdev->config.cik.max_cu_per_sh = 2;
+		rdev->config.cik.max_sh_per_se = 1;
+		rdev->config.cik.max_backends_per_se = 1;
+		rdev->config.cik.max_texture_channel_caches = 2;
+		rdev->config.cik.max_gprs = 256;
+		rdev->config.cik.max_gs_threads = 16;
+		rdev->config.cik.max_hw_contexts = 8;
+
+		rdev->config.cik.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.cik.sc_prim_fifo_size_backend = 0x100;
+		rdev->config.cik.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	}
+
+	/* Initialize HDP */
+	for (i = 0, j = 0; i < 32; i++, j += 0x18) {
+		WREG32((0x2c14 + j), 0x00000000);
+		WREG32((0x2c18 + j), 0x00000000);
+		WREG32((0x2c1c + j), 0x00000000);
+		WREG32((0x2c20 + j), 0x00000000);
+		WREG32((0x2c24 + j), 0x00000000);
+	}
+
+	WREG32(GRBM_CNTL, GRBM_READ_TIMEOUT(0xff));
+
+	WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN);
+
+	mc_shared_chmap = RREG32(MC_SHARED_CHMAP);
+	mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG);
+
+	rdev->config.cik.num_tile_pipes = rdev->config.cik.max_tile_pipes;
+	rdev->config.cik.mem_max_burst_length_bytes = 256;
+	tmp = (mc_arb_ramcfg & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT;
+	rdev->config.cik.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024;
+	if (rdev->config.cik.mem_row_size_in_kb > 4)
+		rdev->config.cik.mem_row_size_in_kb = 4;
+	/* XXX use MC settings? */
+	rdev->config.cik.shader_engine_tile_size = 32;
+	rdev->config.cik.num_gpus = 1;
+	rdev->config.cik.multi_gpu_tile_size = 64;
+
+	/* fix up row size */
+	gb_addr_config &= ~ROW_SIZE_MASK;
+	switch (rdev->config.cik.mem_row_size_in_kb) {
+	case 1:
+	default:
+		gb_addr_config |= ROW_SIZE(0);
+		break;
+	case 2:
+		gb_addr_config |= ROW_SIZE(1);
+		break;
+	case 4:
+		gb_addr_config |= ROW_SIZE(2);
+		break;
+	}
+
+	/* setup tiling info dword.  gb_addr_config is not adequate since it does
+	 * not have bank info, so create a custom tiling dword.
+	 * bits 3:0   num_pipes
+	 * bits 7:4   num_banks
+	 * bits 11:8  group_size
+	 * bits 15:12 row_size
+	 */
+	rdev->config.cik.tile_config = 0;
+	switch (rdev->config.cik.num_tile_pipes) {
+	case 1:
+		rdev->config.cik.tile_config |= (0 << 0);
+		break;
+	case 2:
+		rdev->config.cik.tile_config |= (1 << 0);
+		break;
+	case 4:
+		rdev->config.cik.tile_config |= (2 << 0);
+		break;
+	case 8:
+	default:
+		/* XXX what about 12? */
+		rdev->config.cik.tile_config |= (3 << 0);
+		break;
+	}
+	if ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT)
+		rdev->config.cik.tile_config |= 1 << 4;
+	else
+		rdev->config.cik.tile_config |= 0 << 4;
+	rdev->config.cik.tile_config |=
+		((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8;
+	rdev->config.cik.tile_config |=
+		((gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT) << 12;
+
+	WREG32(GB_ADDR_CONFIG, gb_addr_config);
+	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMIF_ADDR_CALC, gb_addr_config);
+	WREG32(SDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET, gb_addr_config & 0x70);
+	WREG32(SDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET, gb_addr_config & 0x70);
+	WREG32(UVD_UDEC_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DB_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config);
+
+	cik_tiling_mode_table_init(rdev);
+
+	cik_setup_rb(rdev, rdev->config.cik.max_shader_engines,
+		     rdev->config.cik.max_sh_per_se,
+		     rdev->config.cik.max_backends_per_se);
+
+	/* set HW defaults for 3D engine */
+	WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60));
+
+	WREG32(SX_DEBUG_1, 0x20);
+
+	WREG32(TA_CNTL_AUX, 0x00010000);
+
+	tmp = RREG32(SPI_CONFIG_CNTL);
+	tmp |= 0x03000000;
+	WREG32(SPI_CONFIG_CNTL, tmp);
+
+	WREG32(SQ_CONFIG, 1);
+
+	WREG32(DB_DEBUG, 0);
+
+	tmp = RREG32(DB_DEBUG2) & ~0xf00fffff;
+	tmp |= 0x00000400;
+	WREG32(DB_DEBUG2, tmp);
+
+	tmp = RREG32(DB_DEBUG3) & ~0x0002021c;
+	tmp |= 0x00020200;
+	WREG32(DB_DEBUG3, tmp);
+
+	tmp = RREG32(CB_HW_CONTROL) & ~0x00010000;
+	tmp |= 0x00018208;
+	WREG32(CB_HW_CONTROL, tmp);
+
+	WREG32(SPI_CONFIG_CNTL_1, VTX_DONE_DELAY(4));
+
+	WREG32(PA_SC_FIFO_SIZE, (SC_FRONTEND_PRIM_FIFO_SIZE(rdev->config.cik.sc_prim_fifo_size_frontend) |
+				 SC_BACKEND_PRIM_FIFO_SIZE(rdev->config.cik.sc_prim_fifo_size_backend) |
+				 SC_HIZ_TILE_FIFO_SIZE(rdev->config.cik.sc_hiz_tile_fifo_size) |
+				 SC_EARLYZ_TILE_FIFO_SIZE(rdev->config.cik.sc_earlyz_tile_fifo_size)));
+
+	WREG32(VGT_NUM_INSTANCES, 1);
+
+	WREG32(CP_PERFMON_CNTL, 0);
+
+	WREG32(SQ_CONFIG, 0);
+
+	WREG32(PA_SC_FORCE_EOV_MAX_CNTS, (FORCE_EOV_MAX_CLK_CNT(4095) |
+					  FORCE_EOV_MAX_REZ_CNT(255)));
+
+	WREG32(VGT_CACHE_INVALIDATION, CACHE_INVALIDATION(VC_AND_TC) |
+	       AUTO_INVLD_EN(ES_AND_GS_AUTO));
+
+	WREG32(VGT_GS_VERTEX_REUSE, 16);
+	WREG32(PA_SC_LINE_STIPPLE_STATE, 0);
+
+	tmp = RREG32(HDP_MISC_CNTL);
+	tmp |= HDP_FLUSH_INVALIDATE_CACHE;
+	WREG32(HDP_MISC_CNTL, tmp);
+
+	hdp_host_path_cntl = RREG32(HDP_HOST_PATH_CNTL);
+	WREG32(HDP_HOST_PATH_CNTL, hdp_host_path_cntl);
+
+	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
+	WREG32(PA_SC_ENHANCE, ENABLE_PA_SC_OUT_OF_ORDER);
+
+	udelay(50);
+}
+
+/*
+ * GPU scratch registers helpers function.
+ */
+/**
+ * cik_scratch_init - setup driver info for CP scratch regs
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the number and offset of the CP scratch registers.
+ * NOTE: use of CP scratch registers is a legacy inferface and
+ * is not used by default on newer asics (r6xx+).  On newer asics,
+ * memory buffers are used for fences rather than scratch regs.
+ */
+static void cik_scratch_init(struct radeon_device *rdev)
+{
+	int i;
+
+	rdev->scratch.num_reg = 7;
+	rdev->scratch.reg_base = SCRATCH_REG0;
+	for (i = 0; i < rdev->scratch.num_reg; i++) {
+		rdev->scratch.free[i] = true;
+		rdev->scratch.reg[i] = rdev->scratch.reg_base + (i * 4);
+	}
+}
+
+/**
+ * cik_ring_test - basic gfx ring test
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Allocate a scratch register and write to it using the gfx ring (CIK).
+ * Provides a basic gfx ring test to verify that the ring is working.
+ * Used by cik_cp_gfx_resume();
+ * Returns 0 on success, error on failure.
+ */
+int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	uint32_t scratch;
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	r = radeon_scratch_get(rdev, &scratch);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to get scratch reg (%d).\n", r);
+		return r;
+	}
+	WREG32(scratch, 0xCAFEDEAD);
+	r = radeon_ring_lock(rdev, ring, 3);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", ring->idx, r);
+		radeon_scratch_free(rdev, scratch);
+		return r;
+	}
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+	radeon_ring_write(ring, ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2));
+	radeon_ring_write(ring, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev, ring);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i);
+	} else {
+		DRM_ERROR("radeon: ring %d test failed (scratch(0x%04X)=0x%08X)\n",
+			  ring->idx, scratch, tmp);
+		r = -EINVAL;
+	}
+	radeon_scratch_free(rdev, scratch);
+	return r;
+}
+
+/**
+ * cik_fence_gfx_ring_emit - emit a fence on the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ * @fence: radeon fence object
+ *
+ * Emits a fence sequnce number on the gfx ring and flushes
+ * GPU caches.
+ */
+void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
+			     struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	/* EVENT_WRITE_EOP - flush caches, send int */
+	radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
+	radeon_ring_write(ring, (EOP_TCL1_ACTION_EN |
+				 EOP_TC_ACTION_EN |
+				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 EVENT_INDEX(5)));
+	radeon_ring_write(ring, addr & 0xfffffffc);
+	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | DATA_SEL(1) | INT_SEL(2));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, 0);
+	/* HDP flush */
+	/* We should be using the new WAIT_REG_MEM special op packet here
+	 * but it causes the CP to hang
+	 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+}
+
+/**
+ * cik_fence_compute_ring_emit - emit a fence on the compute ring
+ *
+ * @rdev: radeon_device pointer
+ * @fence: radeon fence object
+ *
+ * Emits a fence sequnce number on the compute ring and flushes
+ * GPU caches.
+ */
+void cik_fence_compute_ring_emit(struct radeon_device *rdev,
+				 struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	/* RELEASE_MEM - flush caches, send int */
+	radeon_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 5));
+	radeon_ring_write(ring, (EOP_TCL1_ACTION_EN |
+				 EOP_TC_ACTION_EN |
+				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 EVENT_INDEX(5)));
+	radeon_ring_write(ring, DATA_SEL(1) | INT_SEL(2));
+	radeon_ring_write(ring, addr & 0xfffffffc);
+	radeon_ring_write(ring, upper_32_bits(addr));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, 0);
+	/* HDP flush */
+	/* We should be using the new WAIT_REG_MEM special op packet here
+	 * but it causes the CP to hang
+	 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+}
+
+void cik_semaphore_ring_emit(struct radeon_device *rdev,
+			     struct radeon_ring *ring,
+			     struct radeon_semaphore *semaphore,
+			     bool emit_wait)
+{
+	uint64_t addr = semaphore->gpu_addr;
+	unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL;
+
+	radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel);
+}
+
+/*
+ * IB stuff
+ */
+/**
+ * cik_ring_ib_execute - emit an IB (Indirect Buffer) on the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ * @ib: radeon indirect buffer object
+ *
+ * Emits an DE (drawing engine) or CE (constant engine) IB
+ * on the gfx ring.  IBs are usually generated by userspace
+ * acceleration drivers and submitted to the kernel for
+ * sheduling on the ring.  This function schedules the IB
+ * on the gfx ring for execution by the GPU.
+ */
+void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	u32 header, control = INDIRECT_BUFFER_VALID;
+
+	if (ib->is_const_ib) {
+		/* set switch buffer packet before const IB */
+		radeon_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		radeon_ring_write(ring, 0);
+
+		header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2);
+	} else {
+		u32 next_rptr;
+		if (ring->rptr_save_reg) {
+			next_rptr = ring->wptr + 3 + 4;
+			radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+			radeon_ring_write(ring, ((ring->rptr_save_reg -
+						  PACKET3_SET_UCONFIG_REG_START) >> 2));
+			radeon_ring_write(ring, next_rptr);
+		} else if (rdev->wb.enabled) {
+			next_rptr = ring->wptr + 5 + 4;
+			radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+			radeon_ring_write(ring, WRITE_DATA_DST_SEL(1));
+			radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
+			radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
+			radeon_ring_write(ring, next_rptr);
+		}
+
+		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+	}
+
+	control |= ib->length_dw |
+		(ib->vm ? (ib->vm->id << 24) : 0);
+
+	radeon_ring_write(ring, header);
+	radeon_ring_write(ring,
+#ifdef __BIG_ENDIAN
+			  (2 << 0) |
+#endif
+			  (ib->gpu_addr & 0xFFFFFFFC));
+	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF);
+	radeon_ring_write(ring, control);
+}
+
+/**
+ * cik_ib_test - basic gfx ring IB test
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Allocate an IB and execute it on the gfx ring (CIK).
+ * Provides a basic gfx ring test to verify that IBs are working.
+ * Returns 0 on success, error on failure.
+ */
+int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	struct radeon_ib ib;
+	uint32_t scratch;
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	r = radeon_scratch_get(rdev, &scratch);
+	if (r) {
+		DRM_ERROR("radeon: failed to get scratch reg (%d).\n", r);
+		return r;
+	}
+	WREG32(scratch, 0xCAFEDEAD);
+	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
+	if (r) {
+		DRM_ERROR("radeon: failed to get ib (%d).\n", r);
+		return r;
+	}
+	ib.ptr[0] = PACKET3(PACKET3_SET_UCONFIG_REG, 1);
+	ib.ptr[1] = ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2);
+	ib.ptr[2] = 0xDEADBEEF;
+	ib.length_dw = 3;
+	r = radeon_ib_schedule(rdev, &ib, NULL);
+	if (r) {
+		radeon_scratch_free(rdev, scratch);
+		radeon_ib_free(rdev, &ib);
+		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
+		return r;
+	}
+	r = radeon_fence_wait(ib.fence, false);
+	if (r) {
+		DRM_ERROR("radeon: fence wait failed (%d).\n", r);
+		return r;
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i);
+	} else {
+		DRM_ERROR("radeon: ib test failed (scratch(0x%04X)=0x%08X)\n",
+			  scratch, tmp);
+		r = -EINVAL;
+	}
+	radeon_scratch_free(rdev, scratch);
+	radeon_ib_free(rdev, &ib);
+	return r;
+}
+
+/*
+ * CP.
+ * On CIK, gfx and compute now have independant command processors.
+ *
+ * GFX
+ * Gfx consists of a single ring and can process both gfx jobs and
+ * compute jobs.  The gfx CP consists of three microengines (ME):
+ * PFP - Pre-Fetch Parser
+ * ME - Micro Engine
+ * CE - Constant Engine
+ * The PFP and ME make up what is considered the Drawing Engine (DE).
+ * The CE is an asynchronous engine used for updating buffer desciptors
+ * used by the DE so that they can be loaded into cache in parallel
+ * while the DE is processing state update packets.
+ *
+ * Compute
+ * The compute CP consists of two microengines (ME):
+ * MEC1 - Compute MicroEngine 1
+ * MEC2 - Compute MicroEngine 2
+ * Each MEC supports 4 compute pipes and each pipe supports 8 queues.
+ * The queues are exposed to userspace and are programmed directly
+ * by the compute runtime.
+ */
+/**
+ * cik_cp_gfx_enable - enable/disable the gfx CP MEs
+ *
+ * @rdev: radeon_device pointer
+ * @enable: enable or disable the MEs
+ *
+ * Halts or unhalts the gfx MEs.
+ */
+static void cik_cp_gfx_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32(CP_ME_CNTL, 0);
+	else {
+		WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT));
+		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	}
+	udelay(50);
+}
+
+/**
+ * cik_cp_gfx_load_microcode - load the gfx CP ME ucode
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Loads the gfx PFP, ME, and CE ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw)
+		return -EINVAL;
+
+	cik_cp_gfx_enable(rdev, false);
+
+	/* PFP */
+	fw_data = (const __be32 *)rdev->pfp_fw->data;
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	for (i = 0; i < CIK_PFP_UCODE_SIZE; i++)
+		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+
+	/* CE */
+	fw_data = (const __be32 *)rdev->ce_fw->data;
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	for (i = 0; i < CIK_CE_UCODE_SIZE; i++)
+		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_CE_UCODE_ADDR, 0);
+
+	/* ME */
+	fw_data = (const __be32 *)rdev->me_fw->data;
+	WREG32(CP_ME_RAM_WADDR, 0);
+	for (i = 0; i < CIK_ME_UCODE_SIZE; i++)
+		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_ME_RAM_WADDR, 0);
+
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	WREG32(CP_ME_RAM_WADDR, 0);
+	WREG32(CP_ME_RAM_RADDR, 0);
+	return 0;
+}
+
+/**
+ * cik_cp_gfx_start - start the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enables the ring and loads the clear state context and other
+ * packets required to init the ring.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_gfx_start(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	int r, i;
+
+	/* init the CP */
+	WREG32(CP_MAX_CONTEXT, rdev->config.cik.max_hw_contexts - 1);
+	WREG32(CP_ENDIAN_SWAP, 0);
+	WREG32(CP_DEVICE_ID, 1);
+
+	cik_cp_gfx_enable(rdev, true);
+
+	r = radeon_ring_lock(rdev, ring, cik_default_size + 17);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+
+	/* init the CE partitions.  CE only used for gfx on CIK */
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2));
+	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
+	radeon_ring_write(ring, 0xc000);
+	radeon_ring_write(ring, 0xc000);
+
+	/* setup clear context state */
+	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	radeon_ring_write(ring, 0x80000000);
+	radeon_ring_write(ring, 0x80000000);
+
+	for (i = 0; i < cik_default_size; i++)
+		radeon_ring_write(ring, cik_default_state[i]);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	/* set clear context state */
+	radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(ring, 0x00000316);
+	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
+
+	radeon_ring_unlock_commit(rdev, ring);
+
+	return 0;
+}
+
+/**
+ * cik_cp_gfx_fini - stop the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the gfx ring and tear down the driver ring
+ * info.
+ */
+static void cik_cp_gfx_fini(struct radeon_device *rdev)
+{
+	cik_cp_gfx_enable(rdev, false);
+	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+}
+
+/**
+ * cik_cp_gfx_resume - setup the gfx ring buffer registers
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Program the location and size of the gfx ring buffer
+ * and test it to make sure it's working.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_gfx_resume(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	u64 rb_addr;
+	int r;
+
+	WREG32(CP_SEM_WAIT_TIMER, 0x0);
+	WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
+
+	/* Set the write pointer delay */
+	WREG32(CP_RB_WPTR_DELAY, 0);
+
+	/* set the RB to use vmid 0 */
+	WREG32(CP_RB_VMID, 0);
+
+	WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF);
+
+	/* ring 0 - compute and gfx */
+	/* Set ring buffer size */
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	rb_bufsz = drm_order(ring->ring_size / 8);
+	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB0_CNTL, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB0_WPTR, ring->wptr);
+
+	/* set the wb address wether it's enabled or not */
+	WREG32(CP_RB0_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(CP_RB0_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF);
+
+	/* scratch register shadowing is no longer supported */
+	WREG32(SCRATCH_UMSK, 0);
+
+	if (!rdev->wb.enabled)
+		tmp |= RB_NO_UPDATE;
+
+	mdelay(1);
+	WREG32(CP_RB0_CNTL, tmp);
+
+	rb_addr = ring->gpu_addr >> 8;
+	WREG32(CP_RB0_BASE, rb_addr);
+	WREG32(CP_RB0_BASE_HI, upper_32_bits(rb_addr));
+
+	ring->rptr = RREG32(CP_RB0_RPTR);
+
+	/* start the ring */
+	cik_cp_gfx_start(rdev);
+	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true;
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	if (r) {
+		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+		return r;
+	}
+	return 0;
+}
+
+u32 cik_compute_ring_get_rptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring)
+{
+	u32 rptr;
+
+
+
+	if (rdev->wb.enabled) {
+		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
+	} else {
+		cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0);
+		rptr = RREG32(CP_HQD_PQ_RPTR);
+		cik_srbm_select(rdev, 0, 0, 0, 0);
+	}
+	rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return rptr;
+}
+
+u32 cik_compute_ring_get_wptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring)
+{
+	u32 wptr;
+
+	if (rdev->wb.enabled) {
+		wptr = le32_to_cpu(rdev->wb.wb[ring->wptr_offs/4]);
+	} else {
+		cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0);
+		wptr = RREG32(CP_HQD_PQ_WPTR);
+		cik_srbm_select(rdev, 0, 0, 0, 0);
+	}
+	wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return wptr;
+}
+
+void cik_compute_ring_set_wptr(struct radeon_device *rdev,
+			       struct radeon_ring *ring)
+{
+	u32 wptr = (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask;
+
+	rdev->wb.wb[ring->wptr_offs/4] = cpu_to_le32(wptr);
+	WDOORBELL32(ring->doorbell_offset, wptr);
+}
+
+/**
+ * cik_cp_compute_enable - enable/disable the compute CP MEs
+ *
+ * @rdev: radeon_device pointer
+ * @enable: enable or disable the MEs
+ *
+ * Halts or unhalts the compute MEs.
+ */
+static void cik_cp_compute_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32(CP_MEC_CNTL, 0);
+	else
+		WREG32(CP_MEC_CNTL, (MEC_ME1_HALT | MEC_ME2_HALT));
+	udelay(50);
+}
+
+/**
+ * cik_cp_compute_load_microcode - load the compute CP ME ucode
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Loads the compute MEC1&2 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->mec_fw)
+		return -EINVAL;
+
+	cik_cp_compute_enable(rdev, false);
+
+	/* MEC1 */
+	fw_data = (const __be32 *)rdev->mec_fw->data;
+	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+	for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
+		WREG32(CP_MEC_ME1_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+
+	if (rdev->family == CHIP_KAVERI) {
+		/* MEC2 */
+		fw_data = (const __be32 *)rdev->mec_fw->data;
+		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
+			WREG32(CP_MEC_ME2_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+	}
+
+	return 0;
+}
+
+/**
+ * cik_cp_compute_start - start the compute queues
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enable the compute queues.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_compute_start(struct radeon_device *rdev)
+{
+	cik_cp_compute_enable(rdev, true);
+
+	return 0;
+}
+
+/**
+ * cik_cp_compute_fini - stop the compute queues
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the compute queues and tear down the driver queue
+ * info.
+ */
+static void cik_cp_compute_fini(struct radeon_device *rdev)
+{
+	int i, idx, r;
+
+	cik_cp_compute_enable(rdev, false);
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			idx = CAYMAN_RING_TYPE_CP1_INDEX;
+		else
+			idx = CAYMAN_RING_TYPE_CP2_INDEX;
+
+		if (rdev->ring[idx].mqd_obj) {
+			r = radeon_bo_reserve(rdev->ring[idx].mqd_obj, false);
+			if (unlikely(r != 0))
+				dev_warn(rdev->dev, "(%d) reserve MQD bo failed\n", r);
+
+			radeon_bo_unpin(rdev->ring[idx].mqd_obj);
+			radeon_bo_unreserve(rdev->ring[idx].mqd_obj);
+
+			radeon_bo_unref(&rdev->ring[idx].mqd_obj);
+			rdev->ring[idx].mqd_obj = NULL;
+		}
+	}
+}
+
+static void cik_mec_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->mec.hpd_eop_obj) {
+		r = radeon_bo_reserve(rdev->mec.hpd_eop_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve HPD EOP bo failed\n", r);
+		radeon_bo_unpin(rdev->mec.hpd_eop_obj);
+		radeon_bo_unreserve(rdev->mec.hpd_eop_obj);
+
+		radeon_bo_unref(&rdev->mec.hpd_eop_obj);
+		rdev->mec.hpd_eop_obj = NULL;
+	}
+}
+
+#define MEC_HPD_SIZE 2048
+
+static int cik_mec_init(struct radeon_device *rdev)
+{
+	int r;
+	u32 *hpd;
+
+	/*
+	 * KV:    2 MEC, 4 Pipes/MEC, 8 Queues/Pipe - 64 Queues total
+	 * CI/KB: 1 MEC, 4 Pipes/MEC, 8 Queues/Pipe - 32 Queues total
+	 */
+	if (rdev->family == CHIP_KAVERI)
+		rdev->mec.num_mec = 2;
+	else
+		rdev->mec.num_mec = 1;
+	rdev->mec.num_pipe = 4;
+	rdev->mec.num_queue = rdev->mec.num_mec * rdev->mec.num_pipe * 8;
+
+	if (rdev->mec.hpd_eop_obj == NULL) {
+		r = radeon_bo_create(rdev,
+				     rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2,
+				     PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_GTT, NULL,
+				     &rdev->mec.hpd_eop_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create HDP EOP bo failed\n", r);
+			return r;
+		}
+	}
+
+	r = radeon_bo_reserve(rdev->mec.hpd_eop_obj, false);
+	if (unlikely(r != 0)) {
+		cik_mec_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->mec.hpd_eop_obj, RADEON_GEM_DOMAIN_GTT,
+			  &rdev->mec.hpd_eop_gpu_addr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) pin HDP EOP bo failed\n", r);
+		cik_mec_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->mec.hpd_eop_obj, (void **)&hpd);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map HDP EOP bo failed\n", r);
+		cik_mec_fini(rdev);
+		return r;
+	}
+
+	/* clear memory.  Not sure if this is required or not */
+	memset(hpd, 0, rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2);
+
+	radeon_bo_kunmap(rdev->mec.hpd_eop_obj);
+	radeon_bo_unreserve(rdev->mec.hpd_eop_obj);
+
+	return 0;
+}
+
+struct hqd_registers
+{
+	u32 cp_mqd_base_addr;
+	u32 cp_mqd_base_addr_hi;
+	u32 cp_hqd_active;
+	u32 cp_hqd_vmid;
+	u32 cp_hqd_persistent_state;
+	u32 cp_hqd_pipe_priority;
+	u32 cp_hqd_queue_priority;
+	u32 cp_hqd_quantum;
+	u32 cp_hqd_pq_base;
+	u32 cp_hqd_pq_base_hi;
+	u32 cp_hqd_pq_rptr;
+	u32 cp_hqd_pq_rptr_report_addr;
+	u32 cp_hqd_pq_rptr_report_addr_hi;
+	u32 cp_hqd_pq_wptr_poll_addr;
+	u32 cp_hqd_pq_wptr_poll_addr_hi;
+	u32 cp_hqd_pq_doorbell_control;
+	u32 cp_hqd_pq_wptr;
+	u32 cp_hqd_pq_control;
+	u32 cp_hqd_ib_base_addr;
+	u32 cp_hqd_ib_base_addr_hi;
+	u32 cp_hqd_ib_rptr;
+	u32 cp_hqd_ib_control;
+	u32 cp_hqd_iq_timer;
+	u32 cp_hqd_iq_rptr;
+	u32 cp_hqd_dequeue_request;
+	u32 cp_hqd_dma_offload;
+	u32 cp_hqd_sema_cmd;
+	u32 cp_hqd_msg_type;
+	u32 cp_hqd_atomic0_preop_lo;
+	u32 cp_hqd_atomic0_preop_hi;
+	u32 cp_hqd_atomic1_preop_lo;
+	u32 cp_hqd_atomic1_preop_hi;
+	u32 cp_hqd_hq_scheduler0;
+	u32 cp_hqd_hq_scheduler1;
+	u32 cp_mqd_control;
+};
+
+struct bonaire_mqd
+{
+	u32 header;
+	u32 dispatch_initiator;
+	u32 dimensions[3];
+	u32 start_idx[3];
+	u32 num_threads[3];
+	u32 pipeline_stat_enable;
+	u32 perf_counter_enable;
+	u32 pgm[2];
+	u32 tba[2];
+	u32 tma[2];
+	u32 pgm_rsrc[2];
+	u32 vmid;
+	u32 resource_limits;
+	u32 static_thread_mgmt01[2];
+	u32 tmp_ring_size;
+	u32 static_thread_mgmt23[2];
+	u32 restart[3];
+	u32 thread_trace_enable;
+	u32 reserved1;
+	u32 user_data[16];
+	u32 vgtcs_invoke_count[2];
+	struct hqd_registers queue_state;
+	u32 dequeue_cntr;
+	u32 interrupt_queue[64];
+};
+
+/**
+ * cik_cp_compute_resume - setup the compute queue registers
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Program the compute queues and test them to make sure they
+ * are working.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_compute_resume(struct radeon_device *rdev)
+{
+	int r, i, idx;
+	u32 tmp;
+	bool use_doorbell = true;
+	u64 hqd_gpu_addr;
+	u64 mqd_gpu_addr;
+	u64 eop_gpu_addr;
+	u64 wb_gpu_addr;
+	u32 *buf;
+	struct bonaire_mqd *mqd;
+
+	r = cik_cp_compute_start(rdev);
+	if (r)
+		return r;
+
+	/* fix up chicken bits */
+	tmp = RREG32(CP_CPF_DEBUG);
+	tmp |= (1 << 23);
+	WREG32(CP_CPF_DEBUG, tmp);
+
+	/* init the pipes */
+	for (i = 0; i < (rdev->mec.num_pipe * rdev->mec.num_mec); i++) {
+		int me = (i < 4) ? 1 : 2;
+		int pipe = (i < 4) ? i : (i - 4);
+
+		eop_gpu_addr = rdev->mec.hpd_eop_gpu_addr + (i * MEC_HPD_SIZE * 2);
+
+		cik_srbm_select(rdev, me, pipe, 0, 0);
+
+		/* write the EOP addr */
+		WREG32(CP_HPD_EOP_BASE_ADDR, eop_gpu_addr >> 8);
+		WREG32(CP_HPD_EOP_BASE_ADDR_HI, upper_32_bits(eop_gpu_addr) >> 8);
+
+		/* set the VMID assigned */
+		WREG32(CP_HPD_EOP_VMID, 0);
+
+		/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+		tmp = RREG32(CP_HPD_EOP_CONTROL);
+		tmp &= ~EOP_SIZE_MASK;
+		tmp |= drm_order(MEC_HPD_SIZE / 8);
+		WREG32(CP_HPD_EOP_CONTROL, tmp);
+	}
+	cik_srbm_select(rdev, 0, 0, 0, 0);
+
+	/* init the queues.  Just two for now. */
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			idx = CAYMAN_RING_TYPE_CP1_INDEX;
+		else
+			idx = CAYMAN_RING_TYPE_CP2_INDEX;
+
+		if (rdev->ring[idx].mqd_obj == NULL) {
+			r = radeon_bo_create(rdev,
+					     sizeof(struct bonaire_mqd),
+					     PAGE_SIZE, true,
+					     RADEON_GEM_DOMAIN_GTT, NULL,
+					     &rdev->ring[idx].mqd_obj);
+			if (r) {
+				dev_warn(rdev->dev, "(%d) create MQD bo failed\n", r);
+				return r;
+			}
+		}
+
+		r = radeon_bo_reserve(rdev->ring[idx].mqd_obj, false);
+		if (unlikely(r != 0)) {
+			cik_cp_compute_fini(rdev);
+			return r;
+		}
+		r = radeon_bo_pin(rdev->ring[idx].mqd_obj, RADEON_GEM_DOMAIN_GTT,
+				  &mqd_gpu_addr);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) pin MQD bo failed\n", r);
+			cik_cp_compute_fini(rdev);
+			return r;
+		}
+		r = radeon_bo_kmap(rdev->ring[idx].mqd_obj, (void **)&buf);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) map MQD bo failed\n", r);
+			cik_cp_compute_fini(rdev);
+			return r;
+		}
+
+		/* doorbell offset */
+		rdev->ring[idx].doorbell_offset =
+			(rdev->ring[idx].doorbell_page_num * PAGE_SIZE) + 0;
+
+		/* init the mqd struct */
+		memset(buf, 0, sizeof(struct bonaire_mqd));
+
+		mqd = (struct bonaire_mqd *)buf;
+		mqd->header = 0xC0310800;
+		mqd->static_thread_mgmt01[0] = 0xffffffff;
+		mqd->static_thread_mgmt01[1] = 0xffffffff;
+		mqd->static_thread_mgmt23[0] = 0xffffffff;
+		mqd->static_thread_mgmt23[1] = 0xffffffff;
+
+		cik_srbm_select(rdev, rdev->ring[idx].me,
+				rdev->ring[idx].pipe,
+				rdev->ring[idx].queue, 0);
+
+		/* disable wptr polling */
+		tmp = RREG32(CP_PQ_WPTR_POLL_CNTL);
+		tmp &= ~WPTR_POLL_EN;
+		WREG32(CP_PQ_WPTR_POLL_CNTL, tmp);
+
+		/* enable doorbell? */
+		mqd->queue_state.cp_hqd_pq_doorbell_control =
+			RREG32(CP_HQD_PQ_DOORBELL_CONTROL);
+		if (use_doorbell)
+			mqd->queue_state.cp_hqd_pq_doorbell_control |= DOORBELL_EN;
+		else
+			mqd->queue_state.cp_hqd_pq_doorbell_control &= ~DOORBELL_EN;
+		WREG32(CP_HQD_PQ_DOORBELL_CONTROL,
+		       mqd->queue_state.cp_hqd_pq_doorbell_control);
+
+		/* disable the queue if it's active */
+		mqd->queue_state.cp_hqd_dequeue_request = 0;
+		mqd->queue_state.cp_hqd_pq_rptr = 0;
+		mqd->queue_state.cp_hqd_pq_wptr= 0;
+		if (RREG32(CP_HQD_ACTIVE) & 1) {
+			WREG32(CP_HQD_DEQUEUE_REQUEST, 1);
+			for (i = 0; i < rdev->usec_timeout; i++) {
+				if (!(RREG32(CP_HQD_ACTIVE) & 1))
+					break;
+				udelay(1);
+			}
+			WREG32(CP_HQD_DEQUEUE_REQUEST, mqd->queue_state.cp_hqd_dequeue_request);
+			WREG32(CP_HQD_PQ_RPTR, mqd->queue_state.cp_hqd_pq_rptr);
+			WREG32(CP_HQD_PQ_WPTR, mqd->queue_state.cp_hqd_pq_wptr);
+		}
+
+		/* set the pointer to the MQD */
+		mqd->queue_state.cp_mqd_base_addr = mqd_gpu_addr & 0xfffffffc;
+		mqd->queue_state.cp_mqd_base_addr_hi = upper_32_bits(mqd_gpu_addr);
+		WREG32(CP_MQD_BASE_ADDR, mqd->queue_state.cp_mqd_base_addr);
+		WREG32(CP_MQD_BASE_ADDR_HI, mqd->queue_state.cp_mqd_base_addr_hi);
+		/* set MQD vmid to 0 */
+		mqd->queue_state.cp_mqd_control = RREG32(CP_MQD_CONTROL);
+		mqd->queue_state.cp_mqd_control &= ~MQD_VMID_MASK;
+		WREG32(CP_MQD_CONTROL, mqd->queue_state.cp_mqd_control);
+
+		/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+		hqd_gpu_addr = rdev->ring[idx].gpu_addr >> 8;
+		mqd->queue_state.cp_hqd_pq_base = hqd_gpu_addr;
+		mqd->queue_state.cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+		WREG32(CP_HQD_PQ_BASE, mqd->queue_state.cp_hqd_pq_base);
+		WREG32(CP_HQD_PQ_BASE_HI, mqd->queue_state.cp_hqd_pq_base_hi);
+
+		/* set up the HQD, this is similar to CP_RB0_CNTL */
+		mqd->queue_state.cp_hqd_pq_control = RREG32(CP_HQD_PQ_CONTROL);
+		mqd->queue_state.cp_hqd_pq_control &=
+			~(QUEUE_SIZE_MASK | RPTR_BLOCK_SIZE_MASK);
+
+		mqd->queue_state.cp_hqd_pq_control |=
+			drm_order(rdev->ring[idx].ring_size / 8);
+		mqd->queue_state.cp_hqd_pq_control |=
+			(drm_order(RADEON_GPU_PAGE_SIZE/8) << 8);
+#ifdef __BIG_ENDIAN
+		mqd->queue_state.cp_hqd_pq_control |= BUF_SWAP_32BIT;
+#endif
+		mqd->queue_state.cp_hqd_pq_control &=
+			~(UNORD_DISPATCH | ROQ_PQ_IB_FLIP | PQ_VOLATILE);
+		mqd->queue_state.cp_hqd_pq_control |=
+			PRIV_STATE | KMD_QUEUE; /* assuming kernel queue control */
+		WREG32(CP_HQD_PQ_CONTROL, mqd->queue_state.cp_hqd_pq_control);
+
+		/* only used if CP_PQ_WPTR_POLL_CNTL.WPTR_POLL_EN=1 */
+		if (i == 0)
+			wb_gpu_addr = rdev->wb.gpu_addr + CIK_WB_CP1_WPTR_OFFSET;
+		else
+			wb_gpu_addr = rdev->wb.gpu_addr + CIK_WB_CP2_WPTR_OFFSET;
+		mqd->queue_state.cp_hqd_pq_wptr_poll_addr = wb_gpu_addr & 0xfffffffc;
+		mqd->queue_state.cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+		WREG32(CP_HQD_PQ_WPTR_POLL_ADDR, mqd->queue_state.cp_hqd_pq_wptr_poll_addr);
+		WREG32(CP_HQD_PQ_WPTR_POLL_ADDR_HI,
+		       mqd->queue_state.cp_hqd_pq_wptr_poll_addr_hi);
+
+		/* set the wb address wether it's enabled or not */
+		if (i == 0)
+			wb_gpu_addr = rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET;
+		else
+			wb_gpu_addr = rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET;
+		mqd->queue_state.cp_hqd_pq_rptr_report_addr = wb_gpu_addr & 0xfffffffc;
+		mqd->queue_state.cp_hqd_pq_rptr_report_addr_hi =
+			upper_32_bits(wb_gpu_addr) & 0xffff;
+		WREG32(CP_HQD_PQ_RPTR_REPORT_ADDR,
+		       mqd->queue_state.cp_hqd_pq_rptr_report_addr);
+		WREG32(CP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+		       mqd->queue_state.cp_hqd_pq_rptr_report_addr_hi);
+
+		/* enable the doorbell if requested */
+		if (use_doorbell) {
+			mqd->queue_state.cp_hqd_pq_doorbell_control =
+				RREG32(CP_HQD_PQ_DOORBELL_CONTROL);
+			mqd->queue_state.cp_hqd_pq_doorbell_control &= ~DOORBELL_OFFSET_MASK;
+			mqd->queue_state.cp_hqd_pq_doorbell_control |=
+				DOORBELL_OFFSET(rdev->ring[idx].doorbell_offset / 4);
+			mqd->queue_state.cp_hqd_pq_doorbell_control |= DOORBELL_EN;
+			mqd->queue_state.cp_hqd_pq_doorbell_control &=
+				~(DOORBELL_SOURCE | DOORBELL_HIT);
+
+		} else {
+			mqd->queue_state.cp_hqd_pq_doorbell_control = 0;
+		}
+		WREG32(CP_HQD_PQ_DOORBELL_CONTROL,
+		       mqd->queue_state.cp_hqd_pq_doorbell_control);
+
+		/* read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+		rdev->ring[idx].wptr = 0;
+		mqd->queue_state.cp_hqd_pq_wptr = rdev->ring[idx].wptr;
+		WREG32(CP_HQD_PQ_WPTR, mqd->queue_state.cp_hqd_pq_wptr);
+		rdev->ring[idx].rptr = RREG32(CP_HQD_PQ_RPTR);
+		mqd->queue_state.cp_hqd_pq_rptr = rdev->ring[idx].rptr;
+
+		/* set the vmid for the queue */
+		mqd->queue_state.cp_hqd_vmid = 0;
+		WREG32(CP_HQD_VMID, mqd->queue_state.cp_hqd_vmid);
+
+		/* activate the queue */
+		mqd->queue_state.cp_hqd_active = 1;
+		WREG32(CP_HQD_ACTIVE, mqd->queue_state.cp_hqd_active);
+
+		cik_srbm_select(rdev, 0, 0, 0, 0);
+
+		radeon_bo_kunmap(rdev->ring[idx].mqd_obj);
+		radeon_bo_unreserve(rdev->ring[idx].mqd_obj);
+
+		rdev->ring[idx].ready = true;
+		r = radeon_ring_test(rdev, idx, &rdev->ring[idx]);
+		if (r)
+			rdev->ring[idx].ready = false;
+	}
+
+	return 0;
+}
+
+static void cik_cp_enable(struct radeon_device *rdev, bool enable)
+{
+	cik_cp_gfx_enable(rdev, enable);
+	cik_cp_compute_enable(rdev, enable);
+}
+
+static int cik_cp_load_microcode(struct radeon_device *rdev)
+{
+	int r;
+
+	r = cik_cp_gfx_load_microcode(rdev);
+	if (r)
+		return r;
+	r = cik_cp_compute_load_microcode(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void cik_cp_fini(struct radeon_device *rdev)
+{
+	cik_cp_gfx_fini(rdev);
+	cik_cp_compute_fini(rdev);
+}
+
+static int cik_cp_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* Reset all cp blocks */
+	WREG32(GRBM_SOFT_RESET, SOFT_RESET_CP);
+	RREG32(GRBM_SOFT_RESET);
+	mdelay(15);
+	WREG32(GRBM_SOFT_RESET, 0);
+	RREG32(GRBM_SOFT_RESET);
+
+	r = cik_cp_load_microcode(rdev);
+	if (r)
+		return r;
+
+	r = cik_cp_gfx_resume(rdev);
+	if (r)
+		return r;
+	r = cik_cp_compute_resume(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/*
+ * sDMA - System DMA
+ * Starting with CIK, the GPU has new asynchronous
+ * DMA engines.  These engines are used for compute
+ * and gfx.  There are two DMA engines (SDMA0, SDMA1)
+ * and each one supports 1 ring buffer used for gfx
+ * and 2 queues used for compute.
+ *
+ * The programming model is very similar to the CP
+ * (ring buffer, IBs, etc.), but sDMA has it's own
+ * packet format that is different from the PM4 format
+ * used by the CP. sDMA supports copying data, writing
+ * embedded data, solid fills, and a number of other
+ * things.  It also has support for tiling/detiling of
+ * buffers.
+ */
+/**
+ * cik_sdma_ring_ib_execute - Schedule an IB on the DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @ib: IB object to schedule
+ *
+ * Schedule an IB in the DMA ring (CIK).
+ */
+void cik_sdma_ring_ib_execute(struct radeon_device *rdev,
+			      struct radeon_ib *ib)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	u32 extra_bits = (ib->vm ? ib->vm->id : 0) & 0xf;
+
+	if (rdev->wb.enabled) {
+		u32 next_rptr = ring->wptr + 5;
+		while ((next_rptr & 7) != 4)
+			next_rptr++;
+		next_rptr += 4;
+		radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
+		radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
+		radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
+		radeon_ring_write(ring, 1); /* number of DWs to follow */
+		radeon_ring_write(ring, next_rptr);
+	}
+
+	/* IB packet must end on a 8 DW boundary */
+	while ((ring->wptr & 7) != 4)
+		radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_INDIRECT_BUFFER, 0, extra_bits));
+	radeon_ring_write(ring, ib->gpu_addr & 0xffffffe0); /* base must be 32 byte aligned */
+	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xffffffff);
+	radeon_ring_write(ring, ib->length_dw);
+
+}
+
+/**
+ * cik_sdma_fence_ring_emit - emit a fence on the DMA ring
+ *
+ * @rdev: radeon_device pointer
+ * @fence: radeon fence object
+ *
+ * Add a DMA fence packet to the ring to write
+ * the fence seq number and DMA trap packet to generate
+ * an interrupt if needed (CIK).
+ */
+void cik_sdma_fence_ring_emit(struct radeon_device *rdev,
+			      struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+	u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) |
+			  SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */
+	u32 ref_and_mask;
+
+	if (fence->ring == R600_RING_TYPE_DMA_INDEX)
+		ref_and_mask = SDMA0;
+	else
+		ref_and_mask = SDMA1;
+
+	/* write the fence */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_FENCE, 0, 0));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+	radeon_ring_write(ring, fence->seq);
+	/* generate an interrupt */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_TRAP, 0, 0));
+	/* flush HDP */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits));
+	radeon_ring_write(ring, GPU_HDP_FLUSH_DONE);
+	radeon_ring_write(ring, GPU_HDP_FLUSH_REQ);
+	radeon_ring_write(ring, ref_and_mask); /* REFERENCE */
+	radeon_ring_write(ring, ref_and_mask); /* MASK */
+	radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */
+}
+
+/**
+ * cik_sdma_semaphore_ring_emit - emit a semaphore on the dma ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ * @semaphore: radeon semaphore object
+ * @emit_wait: wait or signal semaphore
+ *
+ * Add a DMA semaphore packet to the ring wait on or signal
+ * other rings (CIK).
+ */
+void cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
+				  struct radeon_ring *ring,
+				  struct radeon_semaphore *semaphore,
+				  bool emit_wait)
+{
+	u64 addr = semaphore->gpu_addr;
+	u32 extra_bits = emit_wait ? 0 : SDMA_SEMAPHORE_EXTRA_S;
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SEMAPHORE, 0, extra_bits));
+	radeon_ring_write(ring, addr & 0xfffffff8);
+	radeon_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+}
+
+/**
+ * cik_sdma_gfx_stop - stop the gfx async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the gfx async dma ring buffers (CIK).
+ */
+static void cik_sdma_gfx_stop(struct radeon_device *rdev)
+{
+	u32 rb_cntl, reg_offset;
+	int i;
+
+	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			reg_offset = SDMA0_REGISTER_OFFSET;
+		else
+			reg_offset = SDMA1_REGISTER_OFFSET;
+		rb_cntl = RREG32(SDMA0_GFX_RB_CNTL + reg_offset);
+		rb_cntl &= ~SDMA_RB_ENABLE;
+		WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl);
+		WREG32(SDMA0_GFX_IB_CNTL + reg_offset, 0);
+	}
+}
+
+/**
+ * cik_sdma_rlc_stop - stop the compute async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the compute async dma queues (CIK).
+ */
+static void cik_sdma_rlc_stop(struct radeon_device *rdev)
+{
+	/* XXX todo */
+}
+
+/**
+ * cik_sdma_enable - stop the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ * @enable: enable/disable the DMA MEs.
+ *
+ * Halt or unhalt the async dma engines (CIK).
+ */
+static void cik_sdma_enable(struct radeon_device *rdev, bool enable)
+{
+	u32 me_cntl, reg_offset;
+	int i;
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			reg_offset = SDMA0_REGISTER_OFFSET;
+		else
+			reg_offset = SDMA1_REGISTER_OFFSET;
+		me_cntl = RREG32(SDMA0_ME_CNTL + reg_offset);
+		if (enable)
+			me_cntl &= ~SDMA_HALT;
+		else
+			me_cntl |= SDMA_HALT;
+		WREG32(SDMA0_ME_CNTL + reg_offset, me_cntl);
+	}
+}
+
+/**
+ * cik_sdma_gfx_resume - setup and start the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them (CIK).
+ * Returns 0 for success, error for failure.
+ */
+static int cik_sdma_gfx_resume(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	u32 rb_cntl, ib_cntl;
+	u32 rb_bufsz;
+	u32 reg_offset, wb_offset;
+	int i, r;
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0) {
+			ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
+			reg_offset = SDMA0_REGISTER_OFFSET;
+			wb_offset = R600_WB_DMA_RPTR_OFFSET;
+		} else {
+			ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
+			reg_offset = SDMA1_REGISTER_OFFSET;
+			wb_offset = CAYMAN_WB_DMA1_RPTR_OFFSET;
+		}
+
+		WREG32(SDMA0_SEM_INCOMPLETE_TIMER_CNTL + reg_offset, 0);
+		WREG32(SDMA0_SEM_WAIT_FAIL_TIMER_CNTL + reg_offset, 0);
+
+		/* Set ring buffer size in dwords */
+		rb_bufsz = drm_order(ring->ring_size / 4);
+		rb_cntl = rb_bufsz << 1;
+#ifdef __BIG_ENDIAN
+		rb_cntl |= SDMA_RB_SWAP_ENABLE | SDMA_RPTR_WRITEBACK_SWAP_ENABLE;
+#endif
+		WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl);
+
+		/* Initialize the ring buffer's read and write pointers */
+		WREG32(SDMA0_GFX_RB_RPTR + reg_offset, 0);
+		WREG32(SDMA0_GFX_RB_WPTR + reg_offset, 0);
+
+		/* set the wb address whether it's enabled or not */
+		WREG32(SDMA0_GFX_RB_RPTR_ADDR_HI + reg_offset,
+		       upper_32_bits(rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+		WREG32(SDMA0_GFX_RB_RPTR_ADDR_LO + reg_offset,
+		       ((rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC));
+
+		if (rdev->wb.enabled)
+			rb_cntl |= SDMA_RPTR_WRITEBACK_ENABLE;
+
+		WREG32(SDMA0_GFX_RB_BASE + reg_offset, ring->gpu_addr >> 8);
+		WREG32(SDMA0_GFX_RB_BASE_HI + reg_offset, ring->gpu_addr >> 40);
+
+		ring->wptr = 0;
+		WREG32(SDMA0_GFX_RB_WPTR + reg_offset, ring->wptr << 2);
+
+		ring->rptr = RREG32(SDMA0_GFX_RB_RPTR + reg_offset) >> 2;
+
+		/* enable DMA RB */
+		WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl | SDMA_RB_ENABLE);
+
+		ib_cntl = SDMA_IB_ENABLE;
+#ifdef __BIG_ENDIAN
+		ib_cntl |= SDMA_IB_SWAP_ENABLE;
+#endif
+		/* enable DMA IBs */
+		WREG32(SDMA0_GFX_IB_CNTL + reg_offset, ib_cntl);
+
+		ring->ready = true;
+
+		r = radeon_ring_test(rdev, ring->idx, ring);
+		if (r) {
+			ring->ready = false;
+			return r;
+		}
+	}
+
+	radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
+
+	return 0;
+}
+
+/**
+ * cik_sdma_rlc_resume - setup and start the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the compute DMA queues and enable them (CIK).
+ * Returns 0 for success, error for failure.
+ */
+static int cik_sdma_rlc_resume(struct radeon_device *rdev)
+{
+	/* XXX todo */
+	return 0;
+}
+
+/**
+ * cik_sdma_load_microcode - load the sDMA ME ucode
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Loads the sDMA0/1 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_sdma_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->sdma_fw)
+		return -EINVAL;
+
+	/* stop the gfx rings and rlc compute queues */
+	cik_sdma_gfx_stop(rdev);
+	cik_sdma_rlc_stop(rdev);
+
+	/* halt the MEs */
+	cik_sdma_enable(rdev, false);
+
+	/* sdma0 */
+	fw_data = (const __be32 *)rdev->sdma_fw->data;
+	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+	WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+	/* sdma1 */
+	fw_data = (const __be32 *)rdev->sdma_fw->data;
+	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+	WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+	return 0;
+}
+
+/**
+ * cik_sdma_resume - setup and start the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the DMA engines and enable them (CIK).
+ * Returns 0 for success, error for failure.
+ */
+static int cik_sdma_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* Reset dma */
+	WREG32(SRBM_SOFT_RESET, SOFT_RESET_SDMA | SOFT_RESET_SDMA1);
+	RREG32(SRBM_SOFT_RESET);
+	udelay(50);
+	WREG32(SRBM_SOFT_RESET, 0);
+	RREG32(SRBM_SOFT_RESET);
+
+	r = cik_sdma_load_microcode(rdev);
+	if (r)
+		return r;
+
+	/* unhalt the MEs */
+	cik_sdma_enable(rdev, true);
+
+	/* start the gfx rings and rlc compute queues */
+	r = cik_sdma_gfx_resume(rdev);
+	if (r)
+		return r;
+	r = cik_sdma_rlc_resume(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * cik_sdma_fini - tear down the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the async dma engines and free the rings (CIK).
+ */
+static void cik_sdma_fini(struct radeon_device *rdev)
+{
+	/* stop the gfx rings and rlc compute queues */
+	cik_sdma_gfx_stop(rdev);
+	cik_sdma_rlc_stop(rdev);
+	/* halt the MEs */
+	cik_sdma_enable(rdev, false);
+	radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]);
+	radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]);
+	/* XXX - compute dma queue tear down */
+}
+
+/**
+ * cik_copy_dma - copy pages using the DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @src_offset: src GPU address
+ * @dst_offset: dst GPU address
+ * @num_gpu_pages: number of GPU pages to xfer
+ * @fence: radeon fence object
+ *
+ * Copy GPU paging using the DMA engine (CIK).
+ * Used by the radeon ttm implementation to move pages if
+ * registered as the asic copy callback.
+ */
+int cik_copy_dma(struct radeon_device *rdev,
+		 uint64_t src_offset, uint64_t dst_offset,
+		 unsigned num_gpu_pages,
+		 struct radeon_fence **fence)
+{
+	struct radeon_semaphore *sem = NULL;
+	int ring_index = rdev->asic->copy.dma_ring_index;
+	struct radeon_ring *ring = &rdev->ring[ring_index];
+	u32 size_in_bytes, cur_size_in_bytes;
+	int i, num_loops;
+	int r = 0;
+
+	r = radeon_semaphore_create(rdev, &sem);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		return r;
+	}
+
+	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
+	num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff);
+	r = radeon_ring_lock(rdev, ring, num_loops * 7 + 14);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		radeon_semaphore_free(rdev, &sem, NULL);
+		return r;
+	}
+
+	if (radeon_fence_need_sync(*fence, ring->idx)) {
+		radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring,
+					    ring->idx);
+		radeon_fence_note_sync(*fence, ring->idx);
+	} else {
+		radeon_semaphore_free(rdev, &sem, NULL);
+	}
+
+	for (i = 0; i < num_loops; i++) {
+		cur_size_in_bytes = size_in_bytes;
+		if (cur_size_in_bytes > 0x1fffff)
+			cur_size_in_bytes = 0x1fffff;
+		size_in_bytes -= cur_size_in_bytes;
+		radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_COPY_SUB_OPCODE_LINEAR, 0));
+		radeon_ring_write(ring, cur_size_in_bytes);
+		radeon_ring_write(ring, 0); /* src/dst endian swap */
+		radeon_ring_write(ring, src_offset & 0xffffffff);
+		radeon_ring_write(ring, upper_32_bits(src_offset) & 0xffffffff);
+		radeon_ring_write(ring, dst_offset & 0xfffffffc);
+		radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xffffffff);
+		src_offset += cur_size_in_bytes;
+		dst_offset += cur_size_in_bytes;
+	}
+
+	r = radeon_fence_emit(rdev, fence, ring->idx);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+
+	radeon_ring_unlock_commit(rdev, ring);
+	radeon_semaphore_free(rdev, &sem, *fence);
+
+	return r;
+}
+
+/**
+ * cik_sdma_ring_test - simple async dma engine test
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Test the DMA engine by writing using it to write an
+ * value to memory. (CIK).
+ * Returns 0 for success, error for failure.
+ */
+int cik_sdma_ring_test(struct radeon_device *rdev,
+		       struct radeon_ring *ring)
+{
+	unsigned i;
+	int r;
+	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	u32 tmp;
+
+	if (!ptr) {
+		DRM_ERROR("invalid vram scratch pointer\n");
+		return -EINVAL;
+	}
+
+	tmp = 0xCAFEDEAD;
+	writel(tmp, ptr);
+
+	r = radeon_ring_lock(rdev, ring, 4);
+	if (r) {
+		DRM_ERROR("radeon: dma failed to lock ring %d (%d).\n", ring->idx, r);
+		return r;
+	}
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
+	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
+	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xffffffff);
+	radeon_ring_write(ring, 1); /* number of DWs to follow */
+	radeon_ring_write(ring, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev, ring);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = readl(ptr);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i);
+	} else {
+		DRM_ERROR("radeon: ring %d test failed (0x%08X)\n",
+			  ring->idx, tmp);
+		r = -EINVAL;
+	}
+	return r;
+}
+
+/**
+ * cik_sdma_ib_test - test an IB on the DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Test a simple IB in the DMA ring (CIK).
+ * Returns 0 on success, error on failure.
+ */
+int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	struct radeon_ib ib;
+	unsigned i;
+	int r;
+	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	u32 tmp = 0;
+
+	if (!ptr) {
+		DRM_ERROR("invalid vram scratch pointer\n");
+		return -EINVAL;
+	}
+
+	tmp = 0xCAFEDEAD;
+	writel(tmp, ptr);
+
+	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
+	if (r) {
+		DRM_ERROR("radeon: failed to get ib (%d).\n", r);
+		return r;
+	}
+
+	ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+	ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
+	ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xffffffff;
+	ib.ptr[3] = 1;
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.length_dw = 5;
+
+	r = radeon_ib_schedule(rdev, &ib, NULL);
+	if (r) {
+		radeon_ib_free(rdev, &ib);
+		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
+		return r;
+	}
+	r = radeon_fence_wait(ib.fence, false);
+	if (r) {
+		DRM_ERROR("radeon: fence wait failed (%d).\n", r);
+		return r;
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = readl(ptr);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i);
+	} else {
+		DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp);
+		r = -EINVAL;
+	}
+	radeon_ib_free(rdev, &ib);
+	return r;
+}
+
+
+static void cik_print_gpu_status_regs(struct radeon_device *rdev)
+{
+	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
+		RREG32(GRBM_STATUS));
+	dev_info(rdev->dev, "  GRBM_STATUS2=0x%08X\n",
+		RREG32(GRBM_STATUS2));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
+		RREG32(GRBM_STATUS_SE0));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE1=0x%08X\n",
+		RREG32(GRBM_STATUS_SE1));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE2=0x%08X\n",
+		RREG32(GRBM_STATUS_SE2));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE3=0x%08X\n",
+		RREG32(GRBM_STATUS_SE3));
+	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
+		RREG32(SRBM_STATUS));
+	dev_info(rdev->dev, "  SRBM_STATUS2=0x%08X\n",
+		RREG32(SRBM_STATUS2));
+	dev_info(rdev->dev, "  SDMA0_STATUS_REG   = 0x%08X\n",
+		RREG32(SDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET));
+	dev_info(rdev->dev, "  SDMA1_STATUS_REG   = 0x%08X\n",
+		 RREG32(SDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET));
+	dev_info(rdev->dev, "  CP_STAT = 0x%08x\n", RREG32(CP_STAT));
+	dev_info(rdev->dev, "  CP_STALLED_STAT1 = 0x%08x\n",
+		 RREG32(CP_STALLED_STAT1));
+	dev_info(rdev->dev, "  CP_STALLED_STAT2 = 0x%08x\n",
+		 RREG32(CP_STALLED_STAT2));
+	dev_info(rdev->dev, "  CP_STALLED_STAT3 = 0x%08x\n",
+		 RREG32(CP_STALLED_STAT3));
+	dev_info(rdev->dev, "  CP_CPF_BUSY_STAT = 0x%08x\n",
+		 RREG32(CP_CPF_BUSY_STAT));
+	dev_info(rdev->dev, "  CP_CPF_STALLED_STAT1 = 0x%08x\n",
+		 RREG32(CP_CPF_STALLED_STAT1));
+	dev_info(rdev->dev, "  CP_CPF_STATUS = 0x%08x\n", RREG32(CP_CPF_STATUS));
+	dev_info(rdev->dev, "  CP_CPC_BUSY_STAT = 0x%08x\n", RREG32(CP_CPC_BUSY_STAT));
+	dev_info(rdev->dev, "  CP_CPC_STALLED_STAT1 = 0x%08x\n",
+		 RREG32(CP_CPC_STALLED_STAT1));
+	dev_info(rdev->dev, "  CP_CPC_STATUS = 0x%08x\n", RREG32(CP_CPC_STATUS));
+}
+
+/**
+ * cik_gpu_check_soft_reset - check which blocks are busy
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Check which blocks are busy and return the relevant reset
+ * mask to be used by cik_gpu_soft_reset().
+ * Returns a mask of the blocks to be reset.
+ */
+static u32 cik_gpu_check_soft_reset(struct radeon_device *rdev)
+{
+	u32 reset_mask = 0;
+	u32 tmp;
+
+	/* GRBM_STATUS */
+	tmp = RREG32(GRBM_STATUS);
+	if (tmp & (PA_BUSY | SC_BUSY |
+		   BCI_BUSY | SX_BUSY |
+		   TA_BUSY | VGT_BUSY |
+		   DB_BUSY | CB_BUSY |
+		   GDS_BUSY | SPI_BUSY |
+		   IA_BUSY | IA_BUSY_NO_DMA))
+		reset_mask |= RADEON_RESET_GFX;
+
+	if (tmp & (CP_BUSY | CP_COHERENCY_BUSY))
+		reset_mask |= RADEON_RESET_CP;
+
+	/* GRBM_STATUS2 */
+	tmp = RREG32(GRBM_STATUS2);
+	if (tmp & RLC_BUSY)
+		reset_mask |= RADEON_RESET_RLC;
+
+	/* SDMA0_STATUS_REG */
+	tmp = RREG32(SDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET);
+	if (!(tmp & SDMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* SDMA1_STATUS_REG */
+	tmp = RREG32(SDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET);
+	if (!(tmp & SDMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS2 */
+	tmp = RREG32(SRBM_STATUS2);
+	if (tmp & SDMA_BUSY)
+		reset_mask |= RADEON_RESET_DMA;
+
+	if (tmp & SDMA1_BUSY)
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS */
+	tmp = RREG32(SRBM_STATUS);
+
+	if (tmp & IH_BUSY)
+		reset_mask |= RADEON_RESET_IH;
+
+	if (tmp & SEM_BUSY)
+		reset_mask |= RADEON_RESET_SEM;
+
+	if (tmp & GRBM_RQ_PENDING)
+		reset_mask |= RADEON_RESET_GRBM;
+
+	if (tmp & VMC_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
+
+	if (tmp & (MCB_BUSY | MCB_NON_DISPLAY_BUSY |
+		   MCC_BUSY | MCD_BUSY))
+		reset_mask |= RADEON_RESET_MC;
+
+	if (evergreen_is_display_hung(rdev))
+		reset_mask |= RADEON_RESET_DISPLAY;
+
+	/* Skip MC reset as it's mostly likely not hung, just busy */
+	if (reset_mask & RADEON_RESET_MC) {
+		DRM_DEBUG("MC busy: 0x%08X, clearing.\n", reset_mask);
+		reset_mask &= ~RADEON_RESET_MC;
+	}
+
+	return reset_mask;
+}
+
+/**
+ * cik_gpu_soft_reset - soft reset GPU
+ *
+ * @rdev: radeon_device pointer
+ * @reset_mask: mask of which blocks to reset
+ *
+ * Soft reset the blocks specified in @reset_mask.
+ */
+static void cik_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+{
+	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
+
+	if (reset_mask == 0)
+		return;
+
+	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
+
+	cik_print_gpu_status_regs(rdev);
+	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
+		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
+		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+
+	/* stop the rlc */
+	cik_rlc_stop(rdev);
+
+	/* Disable GFX parsing/prefetching */
+	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
+
+	/* Disable MEC parsing/prefetching */
+	WREG32(CP_MEC_CNTL, MEC_ME1_HALT | MEC_ME2_HALT);
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		/* sdma0 */
+		tmp = RREG32(SDMA0_ME_CNTL + SDMA0_REGISTER_OFFSET);
+		tmp |= SDMA_HALT;
+		WREG32(SDMA0_ME_CNTL + SDMA0_REGISTER_OFFSET, tmp);
+	}
+	if (reset_mask & RADEON_RESET_DMA1) {
+		/* sdma1 */
+		tmp = RREG32(SDMA0_ME_CNTL + SDMA1_REGISTER_OFFSET);
+		tmp |= SDMA_HALT;
+		WREG32(SDMA0_ME_CNTL + SDMA1_REGISTER_OFFSET, tmp);
+	}
+
+	evergreen_mc_stop(rdev, &save);
+	if (evergreen_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))
+		grbm_soft_reset = SOFT_RESET_CP | SOFT_RESET_GFX;
+
+	if (reset_mask & RADEON_RESET_CP) {
+		grbm_soft_reset |= SOFT_RESET_CP;
+
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+	}
+
+	if (reset_mask & RADEON_RESET_DMA)
+		srbm_soft_reset |= SOFT_RESET_SDMA;
+
+	if (reset_mask & RADEON_RESET_DMA1)
+		srbm_soft_reset |= SOFT_RESET_SDMA1;
+
+	if (reset_mask & RADEON_RESET_DISPLAY)
+		srbm_soft_reset |= SOFT_RESET_DC;
+
+	if (reset_mask & RADEON_RESET_RLC)
+		grbm_soft_reset |= SOFT_RESET_RLC;
+
+	if (reset_mask & RADEON_RESET_SEM)
+		srbm_soft_reset |= SOFT_RESET_SEM;
+
+	if (reset_mask & RADEON_RESET_IH)
+		srbm_soft_reset |= SOFT_RESET_IH;
+
+	if (reset_mask & RADEON_RESET_GRBM)
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+
+	if (reset_mask & RADEON_RESET_VMC)
+		srbm_soft_reset |= SOFT_RESET_VMC;
+
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		if (reset_mask & RADEON_RESET_MC)
+			srbm_soft_reset |= SOFT_RESET_MC;
+	}
+
+	if (grbm_soft_reset) {
+		tmp = RREG32(GRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(rdev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+	}
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+	}
+
+	/* Wait a little for things to settle down */
+	udelay(50);
+
+	evergreen_mc_resume(rdev, &save);
+	udelay(50);
+
+	cik_print_gpu_status_regs(rdev);
+}
+
+/**
+ * cik_asic_reset - soft reset GPU
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Look up which blocks are hung and attempt
+ * to reset them.
+ * Returns 0 for success.
+ */
+int cik_asic_reset(struct radeon_device *rdev)
+{
+	u32 reset_mask;
+
+	reset_mask = cik_gpu_check_soft_reset(rdev);
+
+	if (reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, true);
+
+	cik_gpu_soft_reset(rdev, reset_mask);
+
+	reset_mask = cik_gpu_check_soft_reset(rdev);
+
+	if (!reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	return 0;
+}
+
+/**
+ * cik_gfx_is_lockup - check if the 3D engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the 3D engine is locked up (CIK).
+ * Returns true if the engine is locked, false if not.
+ */
+bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = cik_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & (RADEON_RESET_GFX |
+			    RADEON_RESET_COMPUTE |
+			    RADEON_RESET_CP))) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force CP activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/**
+ * cik_sdma_is_lockup - Check if the DMA engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the async DMA engine is locked up (CIK).
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = cik_gpu_check_soft_reset(rdev);
+	u32 mask;
+
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		mask = RADEON_RESET_DMA;
+	else
+		mask = RADEON_RESET_DMA1;
+
+	if (!(reset_mask & mask)) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force ring activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/* MC */
+/**
+ * cik_mc_program - program the GPU memory controller
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set the location of vram, gart, and AGP in the GPU's
+ * physical address space (CIK).
+ */
+static void cik_mc_program(struct radeon_device *rdev)
+{
+	struct evergreen_mc_save save;
+	u32 tmp;
+	int i, j;
+
+	/* Initialize HDP */
+	for (i = 0, j = 0; i < 32; i++, j += 0x18) {
+		WREG32((0x2c14 + j), 0x00000000);
+		WREG32((0x2c18 + j), 0x00000000);
+		WREG32((0x2c1c + j), 0x00000000);
+		WREG32((0x2c20 + j), 0x00000000);
+		WREG32((0x2c24 + j), 0x00000000);
+	}
+	WREG32(HDP_REG_COHERENCY_FLUSH_CNTL, 0);
+
+	evergreen_mc_stop(rdev, &save);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	/* Lockout access through VGA aperture*/
+	WREG32(VGA_HDP_CONTROL, VGA_MEMORY_DISABLE);
+	/* Update configuration */
+	WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR,
+	       rdev->mc.vram_start >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+	       rdev->mc.vram_end >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
+	       rdev->vram_scratch.gpu_addr >> 12);
+	tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16;
+	tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF);
+	WREG32(MC_VM_FB_LOCATION, tmp);
+	/* XXX double check these! */
+	WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8));
+	WREG32(HDP_NONSURFACE_INFO, (2 << 7) | (1 << 30));
+	WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF);
+	WREG32(MC_VM_AGP_BASE, 0);
+	WREG32(MC_VM_AGP_TOP, 0x0FFFFFFF);
+	WREG32(MC_VM_AGP_BOT, 0x0FFFFFFF);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	evergreen_mc_resume(rdev, &save);
+	/* we need to own VRAM, so turn off the VGA renderer here
+	 * to stop it overwriting our objects */
+	rv515_vga_render_disable(rdev);
+}
+
+/**
+ * cik_mc_init - initialize the memory controller driver params
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Look up the amount of vram, vram width, and decide how to place
+ * vram and gart within the GPU's physical address space (CIK).
+ * Returns 0 for success.
+ */
+static int cik_mc_init(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int chansize, numchan;
+
+	/* Get VRAM informations */
+	rdev->mc.vram_is_ddr = true;
+	tmp = RREG32(MC_ARB_RAMCFG);
+	if (tmp & CHANSIZE_MASK) {
+		chansize = 64;
+	} else {
+		chansize = 32;
+	}
+	tmp = RREG32(MC_SHARED_CHMAP);
+	switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) {
+	case 0:
+	default:
+		numchan = 1;
+		break;
+	case 1:
+		numchan = 2;
+		break;
+	case 2:
+		numchan = 4;
+		break;
+	case 3:
+		numchan = 8;
+		break;
+	case 4:
+		numchan = 3;
+		break;
+	case 5:
+		numchan = 6;
+		break;
+	case 6:
+		numchan = 10;
+		break;
+	case 7:
+		numchan = 12;
+		break;
+	case 8:
+		numchan = 16;
+		break;
+	}
+	rdev->mc.vram_width = numchan * chansize;
+	/* Could aper size report 0 ? */
+	rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0);
+	rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0);
+	/* size in MB on si */
+	rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
+	rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
+	rdev->mc.visible_vram_size = rdev->mc.aper_size;
+	si_vram_gtt_location(rdev, &rdev->mc);
+	radeon_update_bandwidth_info(rdev);
+
+	return 0;
+}
+
+/*
+ * GART
+ * VMID 0 is the physical GPU addresses as used by the kernel.
+ * VMIDs 1-15 are used for userspace clients and are handled
+ * by the radeon vm/hsa code.
+ */
+/**
+ * cik_pcie_gart_tlb_flush - gart tlb flush callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Flush the TLB for the VMID 0 page table (CIK).
+ */
+void cik_pcie_gart_tlb_flush(struct radeon_device *rdev)
+{
+	/* flush hdp cache */
+	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0);
+
+	/* bits 0-15 are the VM contexts0-15 */
+	WREG32(VM_INVALIDATE_REQUEST, 0x1);
+}
+
+/**
+ * cik_pcie_gart_enable - gart enable
+ *
+ * @rdev: radeon_device pointer
+ *
+ * This sets up the TLBs, programs the page tables for VMID0,
+ * sets up the hw for VMIDs 1-15 which are allocated on
+ * demand, and sets up the global locations for the LDS, GDS,
+ * and GPUVM for FSA64 clients (CIK).
+ * Returns 0 for success, errors for failure.
+ */
+static int cik_pcie_gart_enable(struct radeon_device *rdev)
+{
+	int r, i;
+
+	if (rdev->gart.robj == NULL) {
+		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
+		return -EINVAL;
+	}
+	r = radeon_gart_table_vram_pin(rdev);
+	if (r)
+		return r;
+	radeon_gart_restore(rdev);
+	/* Setup TLB control */
+	WREG32(MC_VM_MX_L1_TLB_CNTL,
+	       (0xA << 7) |
+	       ENABLE_L1_TLB |
+	       SYSTEM_ACCESS_MODE_NOT_IN_SYS |
+	       ENABLE_ADVANCED_DRIVER_MODEL |
+	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
+	/* Setup L2 cache */
+	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE |
+	       ENABLE_L2_FRAGMENT_PROCESSING |
+	       ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
+	       ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
+	       EFFECTIVE_L2_QUEUE_SIZE(7) |
+	       CONTEXT1_IDENTITY_ACCESS_MODE(1));
+	WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE);
+	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
+	       L2_CACHE_BIGK_FRAGMENT_SIZE(6));
+	/* setup context0 */
+	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12);
+	WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12);
+	WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR, rdev->gart.table_addr >> 12);
+	WREG32(VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR,
+			(u32)(rdev->dummy_page.addr >> 12));
+	WREG32(VM_CONTEXT0_CNTL2, 0);
+	WREG32(VM_CONTEXT0_CNTL, (ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
+				  RANGE_PROTECTION_FAULT_ENABLE_DEFAULT));
+
+	WREG32(0x15D4, 0);
+	WREG32(0x15D8, 0);
+	WREG32(0x15DC, 0);
+
+	/* empty context1-15 */
+	/* FIXME start with 4G, once using 2 level pt switch to full
+	 * vm size space
+	 */
+	/* set vm size, must be a multiple of 4 */
+	WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
+	WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn);
+	for (i = 1; i < 16; i++) {
+		if (i < 8)
+			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
+			       rdev->gart.table_addr >> 12);
+		else
+			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
+			       rdev->gart.table_addr >> 12);
+	}
+
+	/* enable context1-15 */
+	WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
+	       (u32)(rdev->dummy_page.addr >> 12));
+	WREG32(VM_CONTEXT1_CNTL2, 4);
+	WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) |
+				RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+				PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT |
+				VALID_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT |
+				READ_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT |
+				WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT);
+
+	/* TC cache setup ??? */
+	WREG32(TC_CFG_L1_LOAD_POLICY0, 0);
+	WREG32(TC_CFG_L1_LOAD_POLICY1, 0);
+	WREG32(TC_CFG_L1_STORE_POLICY, 0);
+
+	WREG32(TC_CFG_L2_LOAD_POLICY0, 0);
+	WREG32(TC_CFG_L2_LOAD_POLICY1, 0);
+	WREG32(TC_CFG_L2_STORE_POLICY0, 0);
+	WREG32(TC_CFG_L2_STORE_POLICY1, 0);
+	WREG32(TC_CFG_L2_ATOMIC_POLICY, 0);
+
+	WREG32(TC_CFG_L1_VOLATILE, 0);
+	WREG32(TC_CFG_L2_VOLATILE, 0);
+
+	if (rdev->family == CHIP_KAVERI) {
+		u32 tmp = RREG32(CHUB_CONTROL);
+		tmp &= ~BYPASS_VM;
+		WREG32(CHUB_CONTROL, tmp);
+	}
+
+	/* XXX SH_MEM regs */
+	/* where to put LDS, scratch, GPUVM in FSA64 space */
+	for (i = 0; i < 16; i++) {
+		cik_srbm_select(rdev, 0, 0, 0, i);
+		/* CP and shaders */
+		WREG32(SH_MEM_CONFIG, 0);
+		WREG32(SH_MEM_APE1_BASE, 1);
+		WREG32(SH_MEM_APE1_LIMIT, 0);
+		WREG32(SH_MEM_BASES, 0);
+		/* SDMA GFX */
+		WREG32(SDMA0_GFX_VIRTUAL_ADDR + SDMA0_REGISTER_OFFSET, 0);
+		WREG32(SDMA0_GFX_APE1_CNTL + SDMA0_REGISTER_OFFSET, 0);
+		WREG32(SDMA0_GFX_VIRTUAL_ADDR + SDMA1_REGISTER_OFFSET, 0);
+		WREG32(SDMA0_GFX_APE1_CNTL + SDMA1_REGISTER_OFFSET, 0);
+		/* XXX SDMA RLC - todo */
+	}
+	cik_srbm_select(rdev, 0, 0, 0, 0);
+
+	cik_pcie_gart_tlb_flush(rdev);
+	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
+		 (unsigned)(rdev->mc.gtt_size >> 20),
+		 (unsigned long long)rdev->gart.table_addr);
+	rdev->gart.ready = true;
+	return 0;
+}
+
+/**
+ * cik_pcie_gart_disable - gart disable
+ *
+ * @rdev: radeon_device pointer
+ *
+ * This disables all VM page table (CIK).
+ */
+static void cik_pcie_gart_disable(struct radeon_device *rdev)
+{
+	/* Disable all tables */
+	WREG32(VM_CONTEXT0_CNTL, 0);
+	WREG32(VM_CONTEXT1_CNTL, 0);
+	/* Setup TLB control */
+	WREG32(MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE_NOT_IN_SYS |
+	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
+	/* Setup L2 cache */
+	WREG32(VM_L2_CNTL,
+	       ENABLE_L2_FRAGMENT_PROCESSING |
+	       ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
+	       ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
+	       EFFECTIVE_L2_QUEUE_SIZE(7) |
+	       CONTEXT1_IDENTITY_ACCESS_MODE(1));
+	WREG32(VM_L2_CNTL2, 0);
+	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
+	       L2_CACHE_BIGK_FRAGMENT_SIZE(6));
+	radeon_gart_table_vram_unpin(rdev);
+}
+
+/**
+ * cik_pcie_gart_fini - vm fini callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tears down the driver GART/VM setup (CIK).
+ */
+static void cik_pcie_gart_fini(struct radeon_device *rdev)
+{
+	cik_pcie_gart_disable(rdev);
+	radeon_gart_table_vram_free(rdev);
+	radeon_gart_fini(rdev);
+}
+
+/* vm parser */
+/**
+ * cik_ib_parse - vm ib_parse callback
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer pointer
+ *
+ * CIK uses hw IB checking so this is a nop (CIK).
+ */
+int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	return 0;
+}
+
+/*
+ * vm
+ * VMID 0 is the physical GPU addresses as used by the kernel.
+ * VMIDs 1-15 are used for userspace clients and are handled
+ * by the radeon vm/hsa code.
+ */
+/**
+ * cik_vm_init - cik vm init callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Inits cik specific vm parameters (number of VMs, base of vram for
+ * VMIDs 1-15) (CIK).
+ * Returns 0 for success.
+ */
+int cik_vm_init(struct radeon_device *rdev)
+{
+	/* number of VMs */
+	rdev->vm_manager.nvm = 16;
+	/* base offset of vram pages */
+	if (rdev->flags & RADEON_IS_IGP) {
+		u64 tmp = RREG32(MC_VM_FB_OFFSET);
+		tmp <<= 22;
+		rdev->vm_manager.vram_base_offset = tmp;
+	} else
+		rdev->vm_manager.vram_base_offset = 0;
+
+	return 0;
+}
+
+/**
+ * cik_vm_fini - cik vm fini callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down any asic specific VM setup (CIK).
+ */
+void cik_vm_fini(struct radeon_device *rdev)
+{
+}
+
+/**
+ * cik_vm_flush - cik vm flush using the CP
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Update the page table base and flush the VM TLB
+ * using the CP (CIK).
+ */
+void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+{
+	struct radeon_ring *ring = &rdev->ring[ridx];
+
+	if (vm == NULL)
+		return;
+
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	if (vm->id < 8) {
+		radeon_ring_write(ring,
+				  (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+	} else {
+		radeon_ring_write(ring,
+				  (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+	}
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+
+	/* update SH_MEM_* regs */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, VMID(vm->id));
+
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 6));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, SH_MEM_BASES >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, 0); /* SH_MEM_BASES */
+	radeon_ring_write(ring, 0); /* SH_MEM_CONFIG */
+	radeon_ring_write(ring, 1); /* SH_MEM_APE1_BASE */
+	radeon_ring_write(ring, 0); /* SH_MEM_APE1_LIMIT */
+
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, VMID(0));
+
+	/* HDP flush */
+	/* We should be using the WAIT_REG_MEM packet here like in
+	 * cik_fence_ring_emit(), but it causes the CP to hang in this
+	 * context...
+	 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+
+	/* bits 0-15 are the VM contexts0-15 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 1 << vm->id);
+
+	/* compute doesn't have PFP */
+	if (ridx == RADEON_RING_TYPE_GFX_INDEX) {
+		/* sync PFP to ME, otherwise we might get invalid PFP reads */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+}
+
+/**
+ * cik_vm_set_page - update the page tables using sDMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using CP or sDMA (CIK).
+ */
+void cik_vm_set_page(struct radeon_device *rdev,
+		     struct radeon_ib *ib,
+		     uint64_t pe,
+		     uint64_t addr, unsigned count,
+		     uint32_t incr, uint32_t flags)
+{
+	uint32_t r600_flags = cayman_vm_page_flags(rdev, flags);
+	uint64_t value;
+	unsigned ndw;
+
+	if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) {
+		/* CP */
+		while (count) {
+			ndw = 2 + count * 2;
+			if (ndw > 0x3FFE)
+				ndw = 0x3FFE;
+
+			ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw);
+			ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) |
+						    WRITE_DATA_DST_SEL(1));
+			ib->ptr[ib->length_dw++] = pe;
+			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+			for (; ndw > 2; ndw -= 2, --count, pe += 8) {
+				if (flags & RADEON_VM_PAGE_SYSTEM) {
+					value = radeon_vm_map_gart(rdev, addr);
+					value &= 0xFFFFFFFFFFFFF000ULL;
+				} else if (flags & RADEON_VM_PAGE_VALID) {
+					value = addr;
+				} else {
+					value = 0;
+				}
+				addr += incr;
+				value |= r600_flags;
+				ib->ptr[ib->length_dw++] = value;
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
+			}
+		}
+	} else {
+		/* DMA */
+		if (flags & RADEON_VM_PAGE_SYSTEM) {
+			while (count) {
+				ndw = count * 2;
+				if (ndw > 0xFFFFE)
+					ndw = 0xFFFFE;
+
+				/* for non-physically contiguous pages (system) */
+				ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+				ib->ptr[ib->length_dw++] = pe;
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+				ib->ptr[ib->length_dw++] = ndw;
+				for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+					if (flags & RADEON_VM_PAGE_SYSTEM) {
+						value = radeon_vm_map_gart(rdev, addr);
+						value &= 0xFFFFFFFFFFFFF000ULL;
+					} else if (flags & RADEON_VM_PAGE_VALID) {
+						value = addr;
+					} else {
+						value = 0;
+					}
+					addr += incr;
+					value |= r600_flags;
+					ib->ptr[ib->length_dw++] = value;
+					ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				}
+			}
+		} else {
+			while (count) {
+				ndw = count;
+				if (ndw > 0x7FFFF)
+					ndw = 0x7FFFF;
+
+				if (flags & RADEON_VM_PAGE_VALID)
+					value = addr;
+				else
+					value = 0;
+				/* for physically contiguous pages (vram) */
+				ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
+				ib->ptr[ib->length_dw++] = pe; /* dst addr */
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+				ib->ptr[ib->length_dw++] = r600_flags; /* mask */
+				ib->ptr[ib->length_dw++] = 0;
+				ib->ptr[ib->length_dw++] = value; /* value */
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				ib->ptr[ib->length_dw++] = incr; /* increment size */
+				ib->ptr[ib->length_dw++] = 0;
+				ib->ptr[ib->length_dw++] = ndw; /* number of entries */
+				pe += ndw * 8;
+				addr += ndw * incr;
+				count -= ndw;
+			}
+		}
+		while (ib->length_dw & 0x7)
+			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0);
+	}
+}
+
+/**
+ * cik_dma_vm_flush - cik vm flush using sDMA
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Update the page table base and flush the VM TLB
+ * using sDMA (CIK).
+ */
+void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+{
+	struct radeon_ring *ring = &rdev->ring[ridx];
+	u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) |
+			  SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */
+	u32 ref_and_mask;
+
+	if (vm == NULL)
+		return;
+
+	if (ridx == R600_RING_TYPE_DMA_INDEX)
+		ref_and_mask = SDMA0;
+	else
+		ref_and_mask = SDMA1;
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	if (vm->id < 8) {
+		radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+	} else {
+		radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+	}
+	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+
+	/* update SH_MEM_* regs */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, VMID(vm->id));
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_BASES >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_CONFIG >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_APE1_BASE >> 2);
+	radeon_ring_write(ring, 1);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_APE1_LIMIT >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, VMID(0));
+
+	/* flush HDP */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits));
+	radeon_ring_write(ring, GPU_HDP_FLUSH_DONE);
+	radeon_ring_write(ring, GPU_HDP_FLUSH_REQ);
+	radeon_ring_write(ring, ref_and_mask); /* REFERENCE */
+	radeon_ring_write(ring, ref_and_mask); /* MASK */
+	radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */
+
+	/* flush TLB */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+	radeon_ring_write(ring, 1 << vm->id);
+}
+
+/*
+ * RLC
+ * The RLC is a multi-purpose microengine that handles a
+ * variety of functions, the most important of which is
+ * the interrupt controller.
+ */
+/**
+ * cik_rlc_stop - stop the RLC ME
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Halt the RLC ME (MicroEngine) (CIK).
+ */
+static void cik_rlc_stop(struct radeon_device *rdev)
+{
+	int i, j, k;
+	u32 mask, tmp;
+
+	tmp = RREG32(CP_INT_CNTL_RING0);
+	tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+
+	tmp = RREG32(RLC_CGCG_CGLS_CTRL) & 0xfffffffc;
+	WREG32(RLC_CGCG_CGLS_CTRL, tmp);
+
+	WREG32(RLC_CNTL, 0);
+
+	for (i = 0; i < rdev->config.cik.max_shader_engines; i++) {
+		for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) {
+			cik_select_se_sh(rdev, i, j);
+			for (k = 0; k < rdev->usec_timeout; k++) {
+				if (RREG32(RLC_SERDES_CU_MASTER_BUSY) == 0)
+					break;
+				udelay(1);
+			}
+		}
+	}
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+
+	mask = SE_MASTER_BUSY_MASK | GC_MASTER_BUSY | TC0_MASTER_BUSY | TC1_MASTER_BUSY;
+	for (k = 0; k < rdev->usec_timeout; k++) {
+		if ((RREG32(RLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+/**
+ * cik_rlc_start - start the RLC ME
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Unhalt the RLC ME (MicroEngine) (CIK).
+ */
+static void cik_rlc_start(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(RLC_CNTL, RLC_ENABLE);
+
+	tmp = RREG32(CP_INT_CNTL_RING0);
+	tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	udelay(50);
+}
+
+/**
+ * cik_rlc_resume - setup the RLC hw
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Initialize the RLC registers, load the ucode,
+ * and start the RLC (CIK).
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_rlc_resume(struct radeon_device *rdev)
+{
+	u32 i, size;
+	u32 clear_state_info[3];
+	const __be32 *fw_data;
+
+	if (!rdev->rlc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+	default:
+		size = BONAIRE_RLC_UCODE_SIZE;
+		break;
+	case CHIP_KAVERI:
+		size = KV_RLC_UCODE_SIZE;
+		break;
+	case CHIP_KABINI:
+		size = KB_RLC_UCODE_SIZE;
+		break;
+	}
+
+	cik_rlc_stop(rdev);
+
+	WREG32(GRBM_SOFT_RESET, SOFT_RESET_RLC);
+	RREG32(GRBM_SOFT_RESET);
+	udelay(50);
+	WREG32(GRBM_SOFT_RESET, 0);
+	RREG32(GRBM_SOFT_RESET);
+	udelay(50);
+
+	WREG32(RLC_LB_CNTR_INIT, 0);
+	WREG32(RLC_LB_CNTR_MAX, 0x00008000);
+
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+	WREG32(RLC_LB_INIT_CU_MASK, 0xffffffff);
+	WREG32(RLC_LB_PARAMS, 0x00600408);
+	WREG32(RLC_LB_CNTL, 0x80000004);
+
+	WREG32(RLC_MC_CNTL, 0);
+	WREG32(RLC_UCODE_CNTL, 0);
+
+	fw_data = (const __be32 *)rdev->rlc_fw->data;
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+	for (i = 0; i < size; i++)
+		WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(RLC_GPM_UCODE_ADDR, 0);
+
+	/* XXX */
+	clear_state_info[0] = 0;//upper_32_bits(rdev->rlc.save_restore_gpu_addr);
+	clear_state_info[1] = 0;//rdev->rlc.save_restore_gpu_addr;
+	clear_state_info[2] = 0;//cik_default_size;
+	WREG32(RLC_GPM_SCRATCH_ADDR, 0x3d);
+	for (i = 0; i < 3; i++)
+		WREG32(RLC_GPM_SCRATCH_DATA, clear_state_info[i]);
+	WREG32(RLC_DRIVER_DMA_STATUS, 0);
+
+	cik_rlc_start(rdev);
+
+	return 0;
+}
+
+/*
+ * Interrupts
+ * Starting with r6xx, interrupts are handled via a ring buffer.
+ * Ring buffers are areas of GPU accessible memory that the GPU
+ * writes interrupt vectors into and the host reads vectors out of.
+ * There is a rptr (read pointer) that determines where the
+ * host is currently reading, and a wptr (write pointer)
+ * which determines where the GPU has written.  When the
+ * pointers are equal, the ring is idle.  When the GPU
+ * writes vectors to the ring buffer, it increments the
+ * wptr.  When there is an interrupt, the host then starts
+ * fetching commands and processing them until the pointers are
+ * equal again at which point it updates the rptr.
+ */
+
+/**
+ * cik_enable_interrupts - Enable the interrupt ring buffer
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enable the interrupt ring buffer (CIK).
+ */
+static void cik_enable_interrupts(struct radeon_device *rdev)
+{
+	u32 ih_cntl = RREG32(IH_CNTL);
+	u32 ih_rb_cntl = RREG32(IH_RB_CNTL);
+
+	ih_cntl |= ENABLE_INTR;
+	ih_rb_cntl |= IH_RB_ENABLE;
+	WREG32(IH_CNTL, ih_cntl);
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+	rdev->ih.enabled = true;
+}
+
+/**
+ * cik_disable_interrupts - Disable the interrupt ring buffer
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable the interrupt ring buffer (CIK).
+ */
+static void cik_disable_interrupts(struct radeon_device *rdev)
+{
+	u32 ih_rb_cntl = RREG32(IH_RB_CNTL);
+	u32 ih_cntl = RREG32(IH_CNTL);
+
+	ih_rb_cntl &= ~IH_RB_ENABLE;
+	ih_cntl &= ~ENABLE_INTR;
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+	WREG32(IH_CNTL, ih_cntl);
+	/* set rptr, wptr to 0 */
+	WREG32(IH_RB_RPTR, 0);
+	WREG32(IH_RB_WPTR, 0);
+	rdev->ih.enabled = false;
+	rdev->ih.rptr = 0;
+}
+
+/**
+ * cik_disable_interrupt_state - Disable all interrupt sources
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Clear all interrupt enable bits used by the driver (CIK).
+ */
+static void cik_disable_interrupt_state(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	/* gfx ring */
+	WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	/* sdma */
+	tmp = RREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
+	WREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET, tmp);
+	tmp = RREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
+	WREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET, tmp);
+	/* compute queues */
+	WREG32(CP_ME1_PIPE0_INT_CNTL, 0);
+	WREG32(CP_ME1_PIPE1_INT_CNTL, 0);
+	WREG32(CP_ME1_PIPE2_INT_CNTL, 0);
+	WREG32(CP_ME1_PIPE3_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE0_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE1_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE2_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE3_INT_CNTL, 0);
+	/* grbm */
+	WREG32(GRBM_INT_CNTL, 0);
+	/* vline/vblank, etc. */
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
+	if (rdev->num_crtc >= 4) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
+	}
+
+	/* dac hotplug */
+	WREG32(DAC_AUTODETECT_INT_CONTROL, 0);
+
+	/* digital hotplug */
+	tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD1_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD2_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD2_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD3_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD3_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD4_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD4_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD5_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD5_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD6_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD6_INT_CONTROL, tmp);
+
+}
+
+/**
+ * cik_irq_init - init and enable the interrupt ring
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Allocate a ring buffer for the interrupt controller,
+ * enable the RLC, disable interrupts, enable the IH
+ * ring buffer and enable it (CIK).
+ * Called at device load and reume.
+ * Returns 0 for success, errors for failure.
+ */
+static int cik_irq_init(struct radeon_device *rdev)
+{
+	int ret = 0;
+	int rb_bufsz;
+	u32 interrupt_cntl, ih_cntl, ih_rb_cntl;
+
+	/* allocate ring */
+	ret = r600_ih_ring_alloc(rdev);
+	if (ret)
+		return ret;
+
+	/* disable irqs */
+	cik_disable_interrupts(rdev);
+
+	/* init rlc */
+	ret = cik_rlc_resume(rdev);
+	if (ret) {
+		r600_ih_ring_fini(rdev);
+		return ret;
+	}
+
+	/* setup interrupt control */
+	/* XXX this should actually be a bus address, not an MC address. same on older asics */
+	WREG32(INTERRUPT_CNTL2, rdev->ih.gpu_addr >> 8);
+	interrupt_cntl = RREG32(INTERRUPT_CNTL);
+	/* IH_DUMMY_RD_OVERRIDE=0 - dummy read disabled with msi, enabled without msi
+	 * IH_DUMMY_RD_OVERRIDE=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl &= ~IH_DUMMY_RD_OVERRIDE;
+	/* IH_REQ_NONSNOOP_EN=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl &= ~IH_REQ_NONSNOOP_EN;
+	WREG32(INTERRUPT_CNTL, interrupt_cntl);
+
+	WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8);
+	rb_bufsz = drm_order(rdev->ih.ring_size / 4);
+
+	ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE |
+		      IH_WPTR_OVERFLOW_CLEAR |
+		      (rb_bufsz << 1));
+
+	if (rdev->wb.enabled)
+		ih_rb_cntl |= IH_WPTR_WRITEBACK_ENABLE;
+
+	/* set the writeback address whether it's enabled or not */
+	WREG32(IH_RB_WPTR_ADDR_LO, (rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(IH_RB_WPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFF);
+
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+
+	/* set rptr, wptr to 0 */
+	WREG32(IH_RB_RPTR, 0);
+	WREG32(IH_RB_WPTR, 0);
+
+	/* Default settings for IH_CNTL (disabled at first) */
+	ih_cntl = MC_WRREQ_CREDIT(0x10) | MC_WR_CLEAN_CNT(0x10) | MC_VMID(0);
+	/* RPTR_REARM only works if msi's are enabled */
+	if (rdev->msi_enabled)
+		ih_cntl |= RPTR_REARM;
+	WREG32(IH_CNTL, ih_cntl);
+
+	/* force the active interrupt state to all disabled */
+	cik_disable_interrupt_state(rdev);
+
+	pci_set_master(rdev->pdev);
+
+	/* enable irqs */
+	cik_enable_interrupts(rdev);
+
+	return ret;
+}
+
+/**
+ * cik_irq_set - enable/disable interrupt sources
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enable interrupt sources on the GPU (vblanks, hpd,
+ * etc.) (CIK).
+ * Returns 0 for success, errors for failure.
+ */
+int cik_irq_set(struct radeon_device *rdev)
+{
+	u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE |
+		PRIV_INSTR_INT_ENABLE | PRIV_REG_INT_ENABLE;
+	u32 cp_m1p0, cp_m1p1, cp_m1p2, cp_m1p3;
+	u32 cp_m2p0, cp_m2p1, cp_m2p2, cp_m2p3;
+	u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
+	u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6;
+	u32 grbm_int_cntl = 0;
+	u32 dma_cntl, dma_cntl1;
+
+	if (!rdev->irq.installed) {
+		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
+		return -EINVAL;
+	}
+	/* don't enable anything if the ih is disabled */
+	if (!rdev->ih.enabled) {
+		cik_disable_interrupts(rdev);
+		/* force the active interrupt state to all disabled */
+		cik_disable_interrupt_state(rdev);
+		return 0;
+	}
+
+	hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
+
+	dma_cntl = RREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
+	dma_cntl1 = RREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
+
+	cp_m1p0 = RREG32(CP_ME1_PIPE0_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m1p1 = RREG32(CP_ME1_PIPE1_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m1p2 = RREG32(CP_ME1_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m1p3 = RREG32(CP_ME1_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p0 = RREG32(CP_ME2_PIPE0_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p1 = RREG32(CP_ME2_PIPE1_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p2 = RREG32(CP_ME2_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p3 = RREG32(CP_ME2_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+
+	/* enable CP interrupts on all rings */
+	if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) {
+		DRM_DEBUG("cik_irq_set: sw int gfx\n");
+		cp_int_cntl |= TIME_STAMP_INT_ENABLE;
+	}
+	if (atomic_read(&rdev->irq.ring_int[CAYMAN_RING_TYPE_CP1_INDEX])) {
+		struct radeon_ring *ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+		DRM_DEBUG("si_irq_set: sw int cp1\n");
+		if (ring->me == 1) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m1p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m1p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp1 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else if (ring->me == 2) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m2p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m2p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp1 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else {
+			DRM_DEBUG("si_irq_set: sw int cp1 invalid me %d\n", ring->me);
+		}
+	}
+	if (atomic_read(&rdev->irq.ring_int[CAYMAN_RING_TYPE_CP2_INDEX])) {
+		struct radeon_ring *ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+		DRM_DEBUG("si_irq_set: sw int cp2\n");
+		if (ring->me == 1) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m1p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m1p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp2 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else if (ring->me == 2) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m2p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m2p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp2 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else {
+			DRM_DEBUG("si_irq_set: sw int cp2 invalid me %d\n", ring->me);
+		}
+	}
+
+	if (atomic_read(&rdev->irq.ring_int[R600_RING_TYPE_DMA_INDEX])) {
+		DRM_DEBUG("cik_irq_set: sw int dma\n");
+		dma_cntl |= TRAP_ENABLE;
+	}
+
+	if (atomic_read(&rdev->irq.ring_int[CAYMAN_RING_TYPE_DMA1_INDEX])) {
+		DRM_DEBUG("cik_irq_set: sw int dma1\n");
+		dma_cntl1 |= TRAP_ENABLE;
+	}
+
+	if (rdev->irq.crtc_vblank_int[0] ||
+	    atomic_read(&rdev->irq.pflip[0])) {
+		DRM_DEBUG("cik_irq_set: vblank 0\n");
+		crtc1 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[1] ||
+	    atomic_read(&rdev->irq.pflip[1])) {
+		DRM_DEBUG("cik_irq_set: vblank 1\n");
+		crtc2 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[2] ||
+	    atomic_read(&rdev->irq.pflip[2])) {
+		DRM_DEBUG("cik_irq_set: vblank 2\n");
+		crtc3 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[3] ||
+	    atomic_read(&rdev->irq.pflip[3])) {
+		DRM_DEBUG("cik_irq_set: vblank 3\n");
+		crtc4 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[4] ||
+	    atomic_read(&rdev->irq.pflip[4])) {
+		DRM_DEBUG("cik_irq_set: vblank 4\n");
+		crtc5 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[5] ||
+	    atomic_read(&rdev->irq.pflip[5])) {
+		DRM_DEBUG("cik_irq_set: vblank 5\n");
+		crtc6 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.hpd[0]) {
+		DRM_DEBUG("cik_irq_set: hpd 1\n");
+		hpd1 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[1]) {
+		DRM_DEBUG("cik_irq_set: hpd 2\n");
+		hpd2 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[2]) {
+		DRM_DEBUG("cik_irq_set: hpd 3\n");
+		hpd3 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[3]) {
+		DRM_DEBUG("cik_irq_set: hpd 4\n");
+		hpd4 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[4]) {
+		DRM_DEBUG("cik_irq_set: hpd 5\n");
+		hpd5 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[5]) {
+		DRM_DEBUG("cik_irq_set: hpd 6\n");
+		hpd6 |= DC_HPDx_INT_EN;
+	}
+
+	WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+
+	WREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET, dma_cntl);
+	WREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET, dma_cntl1);
+
+	WREG32(CP_ME1_PIPE0_INT_CNTL, cp_m1p0);
+	WREG32(CP_ME1_PIPE1_INT_CNTL, cp_m1p1);
+	WREG32(CP_ME1_PIPE2_INT_CNTL, cp_m1p2);
+	WREG32(CP_ME1_PIPE3_INT_CNTL, cp_m1p3);
+	WREG32(CP_ME2_PIPE0_INT_CNTL, cp_m2p0);
+	WREG32(CP_ME2_PIPE1_INT_CNTL, cp_m2p1);
+	WREG32(CP_ME2_PIPE2_INT_CNTL, cp_m2p2);
+	WREG32(CP_ME2_PIPE3_INT_CNTL, cp_m2p3);
+
+	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
+
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2);
+	if (rdev->num_crtc >= 4) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, crtc3);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, crtc4);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, crtc5);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6);
+	}
+
+	WREG32(DC_HPD1_INT_CONTROL, hpd1);
+	WREG32(DC_HPD2_INT_CONTROL, hpd2);
+	WREG32(DC_HPD3_INT_CONTROL, hpd3);
+	WREG32(DC_HPD4_INT_CONTROL, hpd4);
+	WREG32(DC_HPD5_INT_CONTROL, hpd5);
+	WREG32(DC_HPD6_INT_CONTROL, hpd6);
+
+	return 0;
+}
+
+/**
+ * cik_irq_ack - ack interrupt sources
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Ack interrupt sources on the GPU (vblanks, hpd,
+ * etc.) (CIK).  Certain interrupts sources are sw
+ * generated and do not require an explicit ack.
+ */
+static inline void cik_irq_ack(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	rdev->irq.stat_regs.cik.disp_int = RREG32(DISP_INTERRUPT_STATUS);
+	rdev->irq.stat_regs.cik.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE);
+	rdev->irq.stat_regs.cik.disp_int_cont2 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE2);
+	rdev->irq.stat_regs.cik.disp_int_cont3 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE3);
+	rdev->irq.stat_regs.cik.disp_int_cont4 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE4);
+	rdev->irq.stat_regs.cik.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5);
+	rdev->irq.stat_regs.cik.disp_int_cont6 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE6);
+
+	if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VBLANK_INTERRUPT)
+		WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VLINE_INTERRUPT)
+		WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VLINE_ACK);
+	if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VBLANK_INTERRUPT)
+		WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VLINE_INTERRUPT)
+		WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK);
+
+	if (rdev->num_crtc >= 4) {
+		if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->num_crtc >= 6) {
+		if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->irq.stat_regs.cik.disp_int & DC_HPD1_INTERRUPT) {
+		tmp = RREG32(DC_HPD1_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD1_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont & DC_HPD2_INTERRUPT) {
+		tmp = RREG32(DC_HPD2_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD2_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+		tmp = RREG32(DC_HPD3_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD3_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+		tmp = RREG32(DC_HPD4_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD4_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD5_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD6_INT_CONTROL, tmp);
+	}
+}
+
+/**
+ * cik_irq_disable - disable interrupts
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable interrupts on the hw (CIK).
+ */
+static void cik_irq_disable(struct radeon_device *rdev)
+{
+	cik_disable_interrupts(rdev);
+	/* Wait and acknowledge irq */
+	mdelay(1);
+	cik_irq_ack(rdev);
+	cik_disable_interrupt_state(rdev);
+}
+
+/**
+ * cik_irq_disable - disable interrupts for suspend
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable interrupts and stop the RLC (CIK).
+ * Used for suspend.
+ */
+static void cik_irq_suspend(struct radeon_device *rdev)
+{
+	cik_irq_disable(rdev);
+	cik_rlc_stop(rdev);
+}
+
+/**
+ * cik_irq_fini - tear down interrupt support
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable interrupts on the hw and free the IH ring
+ * buffer (CIK).
+ * Used for driver unload.
+ */
+static void cik_irq_fini(struct radeon_device *rdev)
+{
+	cik_irq_suspend(rdev);
+	r600_ih_ring_fini(rdev);
+}
+
+/**
+ * cik_get_ih_wptr - get the IH ring buffer wptr
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Get the IH ring buffer wptr from either the register
+ * or the writeback memory buffer (CIK).  Also check for
+ * ring buffer overflow and deal with it.
+ * Used by cik_irq_process().
+ * Returns the value of the wptr.
+ */
+static inline u32 cik_get_ih_wptr(struct radeon_device *rdev)
+{
+	u32 wptr, tmp;
+
+	if (rdev->wb.enabled)
+		wptr = le32_to_cpu(rdev->wb.wb[R600_WB_IH_WPTR_OFFSET/4]);
+	else
+		wptr = RREG32(IH_RB_WPTR);
+
+	if (wptr & RB_OVERFLOW) {
+		/* When a ring buffer overflow happen start parsing interrupt
+		 * from the last not overwritten vector (wptr + 16). Hopefully
+		 * this should allow us to catchup.
+		 */
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
+			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
+		tmp = RREG32(IH_RB_CNTL);
+		tmp |= IH_WPTR_OVERFLOW_CLEAR;
+		WREG32(IH_RB_CNTL, tmp);
+	}
+	return (wptr & rdev->ih.ptr_mask);
+}
+
+/*        CIK IV Ring
+ * Each IV ring entry is 128 bits:
+ * [7:0]    - interrupt source id
+ * [31:8]   - reserved
+ * [59:32]  - interrupt source data
+ * [63:60]  - reserved
+ * [71:64]  - RINGID
+ *            CP:
+ *            ME_ID [1:0], PIPE_ID[1:0], QUEUE_ID[2:0]
+ *            QUEUE_ID - for compute, which of the 8 queues owned by the dispatcher
+ *                     - for gfx, hw shader state (0=PS...5=LS, 6=CS)
+ *            ME_ID - 0 = gfx, 1 = first 4 CS pipes, 2 = second 4 CS pipes
+ *            PIPE_ID - ME0 0=3D
+ *                    - ME1&2 compute dispatcher (4 pipes each)
+ *            SDMA:
+ *            INSTANCE_ID [1:0], QUEUE_ID[1:0]
+ *            INSTANCE_ID - 0 = sdma0, 1 = sdma1
+ *            QUEUE_ID - 0 = gfx, 1 = rlc0, 2 = rlc1
+ * [79:72]  - VMID
+ * [95:80]  - PASID
+ * [127:96] - reserved
+ */
+/**
+ * cik_irq_process - interrupt handler
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Interrupt hander (CIK).  Walk the IH ring,
+ * ack interrupts and schedule work to handle
+ * interrupt events.
+ * Returns irq process return code.
+ */
+int cik_irq_process(struct radeon_device *rdev)
+{
+	struct radeon_ring *cp1_ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	struct radeon_ring *cp2_ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	u32 wptr;
+	u32 rptr;
+	u32 src_id, src_data, ring_id;
+	u8 me_id, pipe_id, queue_id;
+	u32 ring_index;
+	bool queue_hotplug = false;
+	bool queue_reset = false;
+
+	if (!rdev->ih.enabled || rdev->shutdown)
+		return IRQ_NONE;
+
+	wptr = cik_get_ih_wptr(rdev);
+
+restart_ih:
+	/* is somebody else already processing irqs? */
+	if (atomic_xchg(&rdev->ih.lock, 1))
+		return IRQ_NONE;
+
+	rptr = rdev->ih.rptr;
+	DRM_DEBUG("cik_irq_process start: rptr %d, wptr %d\n", rptr, wptr);
+
+	/* Order reading of wptr vs. reading of IH ring data */
+	rmb();
+
+	/* display interrupts */
+	cik_irq_ack(rdev);
+
+	while (rptr != wptr) {
+		/* wptr/rptr are in bytes! */
+		ring_index = rptr / 4;
+		src_id =  le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff;
+		src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff;
+		ring_id = le32_to_cpu(rdev->ih.ring[ring_index + 2]) & 0xff;
+
+		switch (src_id) {
+		case 1: /* D1 vblank/vline */
+			switch (src_data) {
+			case 0: /* D1 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[0]) {
+						drm_handle_vblank(rdev->ddev, 0);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[0]))
+						radeon_crtc_handle_flip(rdev, 0);
+					rdev->irq.stat_regs.cik.disp_int &= ~LB_D1_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D1 vblank\n");
+				}
+				break;
+			case 1: /* D1 vline */
+				if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int &= ~LB_D1_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D1 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 2: /* D2 vblank/vline */
+			switch (src_data) {
+			case 0: /* D2 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[1]) {
+						drm_handle_vblank(rdev->ddev, 1);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[1]))
+						radeon_crtc_handle_flip(rdev, 1);
+					rdev->irq.stat_regs.cik.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D2 vblank\n");
+				}
+				break;
+			case 1: /* D2 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont &= ~LB_D2_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D2 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 3: /* D3 vblank/vline */
+			switch (src_data) {
+			case 0: /* D3 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[2]) {
+						drm_handle_vblank(rdev->ddev, 2);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[2]))
+						radeon_crtc_handle_flip(rdev, 2);
+					rdev->irq.stat_regs.cik.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D3 vblank\n");
+				}
+				break;
+			case 1: /* D3 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont2 &= ~LB_D3_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D3 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 4: /* D4 vblank/vline */
+			switch (src_data) {
+			case 0: /* D4 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[3]) {
+						drm_handle_vblank(rdev->ddev, 3);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[3]))
+						radeon_crtc_handle_flip(rdev, 3);
+					rdev->irq.stat_regs.cik.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D4 vblank\n");
+				}
+				break;
+			case 1: /* D4 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont3 &= ~LB_D4_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D4 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 5: /* D5 vblank/vline */
+			switch (src_data) {
+			case 0: /* D5 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[4]) {
+						drm_handle_vblank(rdev->ddev, 4);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[4]))
+						radeon_crtc_handle_flip(rdev, 4);
+					rdev->irq.stat_regs.cik.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D5 vblank\n");
+				}
+				break;
+			case 1: /* D5 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont4 &= ~LB_D5_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D5 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 6: /* D6 vblank/vline */
+			switch (src_data) {
+			case 0: /* D6 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[5]) {
+						drm_handle_vblank(rdev->ddev, 5);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[5]))
+						radeon_crtc_handle_flip(rdev, 5);
+					rdev->irq.stat_regs.cik.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D6 vblank\n");
+				}
+				break;
+			case 1: /* D6 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont5 &= ~LB_D6_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D6 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 42: /* HPD hotplug */
+			switch (src_data) {
+			case 0:
+				if (rdev->irq.stat_regs.cik.disp_int & DC_HPD1_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int &= ~DC_HPD1_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD1\n");
+				}
+				break;
+			case 1:
+				if (rdev->irq.stat_regs.cik.disp_int_cont & DC_HPD2_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont &= ~DC_HPD2_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD2\n");
+				}
+				break;
+			case 2:
+				if (rdev->irq.stat_regs.cik.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont2 &= ~DC_HPD3_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD3\n");
+				}
+				break;
+			case 3:
+				if (rdev->irq.stat_regs.cik.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont3 &= ~DC_HPD4_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD4\n");
+				}
+				break;
+			case 4:
+				if (rdev->irq.stat_regs.cik.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont4 &= ~DC_HPD5_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD5\n");
+				}
+				break;
+			case 5:
+				if (rdev->irq.stat_regs.cik.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont5 &= ~DC_HPD6_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD6\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 146:
+		case 147:
+			dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
+			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
+				RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
+				RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+			/* reset addr and status */
+			WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
+			break;
+		case 176: /* GFX RB CP_INT */
+		case 177: /* GFX IB CP_INT */
+			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			break;
+		case 181: /* CP EOP event */
+			DRM_DEBUG("IH: CP EOP\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x60) >> 5;
+			pipe_id = (ring_id & 0x18) >> 3;
+			queue_id = (ring_id & 0x7) >> 0;
+			switch (me_id) {
+			case 0:
+				radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+				break;
+			case 1:
+			case 2:
+				if ((cp1_ring->me == me_id) & (cp1_ring->pipe == pipe_id))
+					radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
+				if ((cp2_ring->me == me_id) & (cp2_ring->pipe == pipe_id))
+					radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+				break;
+			}
+			break;
+		case 184: /* CP Privileged reg access */
+			DRM_ERROR("Illegal register access in command stream\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x60) >> 5;
+			pipe_id = (ring_id & 0x18) >> 3;
+			queue_id = (ring_id & 0x7) >> 0;
+			switch (me_id) {
+			case 0:
+				/* This results in a full GPU reset, but all we need to do is soft
+				 * reset the CP for gfx
+				 */
+				queue_reset = true;
+				break;
+			case 1:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			case 2:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			}
+			break;
+		case 185: /* CP Privileged inst */
+			DRM_ERROR("Illegal instruction in command stream\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x60) >> 5;
+			pipe_id = (ring_id & 0x18) >> 3;
+			queue_id = (ring_id & 0x7) >> 0;
+			switch (me_id) {
+			case 0:
+				/* This results in a full GPU reset, but all we need to do is soft
+				 * reset the CP for gfx
+				 */
+				queue_reset = true;
+				break;
+			case 1:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			case 2:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			}
+			break;
+		case 224: /* SDMA trap event */
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x3) >> 0;
+			queue_id = (ring_id & 0xc) >> 2;
+			DRM_DEBUG("IH: SDMA trap\n");
+			switch (me_id) {
+			case 0:
+				switch (queue_id) {
+				case 0:
+					radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
+					break;
+				case 1:
+					/* XXX compute */
+					break;
+				case 2:
+					/* XXX compute */
+					break;
+				}
+				break;
+			case 1:
+				switch (queue_id) {
+				case 0:
+					radeon_fence_process(rdev, CAYMAN_RING_TYPE_DMA1_INDEX);
+					break;
+				case 1:
+					/* XXX compute */
+					break;
+				case 2:
+					/* XXX compute */
+					break;
+				}
+				break;
+			}
+			break;
+		case 241: /* SDMA Privileged inst */
+		case 247: /* SDMA Privileged inst */
+			DRM_ERROR("Illegal instruction in SDMA command stream\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x3) >> 0;
+			queue_id = (ring_id & 0xc) >> 2;
+			switch (me_id) {
+			case 0:
+				switch (queue_id) {
+				case 0:
+					queue_reset = true;
+					break;
+				case 1:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				case 2:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				}
+				break;
+			case 1:
+				switch (queue_id) {
+				case 0:
+					queue_reset = true;
+					break;
+				case 1:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				case 2:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				}
+				break;
+			}
+			break;
+		case 233: /* GUI IDLE */
+			DRM_DEBUG("IH: GUI idle\n");
+			break;
+		default:
+			DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+			break;
+		}
+
+		/* wptr/rptr are in bytes! */
+		rptr += 16;
+		rptr &= rdev->ih.ptr_mask;
+	}
+	if (queue_hotplug)
+		schedule_work(&rdev->hotplug_work);
+	if (queue_reset)
+		schedule_work(&rdev->reset_work);
+	rdev->ih.rptr = rptr;
+	WREG32(IH_RB_RPTR, rdev->ih.rptr);
+	atomic_set(&rdev->ih.lock, 0);
+
+	/* make sure wptr hasn't changed while processing */
+	wptr = cik_get_ih_wptr(rdev);
+	if (wptr != rptr)
+		goto restart_ih;
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * startup/shutdown callbacks
+ */
+/**
+ * cik_startup - program the asic to a functional state
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Programs the asic to a functional state (CIK).
+ * Called by cik_init() and cik_resume().
+ * Returns 0 for success, error for failure.
+ */
+static int cik_startup(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	int r;
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
+		    !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw) {
+			r = cik_init_microcode(rdev);
+			if (r) {
+				DRM_ERROR("Failed to load firmware!\n");
+				return r;
+			}
+		}
+	} else {
+		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
+		    !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw ||
+		    !rdev->mc_fw) {
+			r = cik_init_microcode(rdev);
+			if (r) {
+				DRM_ERROR("Failed to load firmware!\n");
+				return r;
+			}
+		}
+
+		r = ci_mc_load_microcode(rdev);
+		if (r) {
+			DRM_ERROR("Failed to load MC firmware!\n");
+			return r;
+		}
+	}
+
+	r = r600_vram_scratch_init(rdev);
+	if (r)
+		return r;
+
+	cik_mc_program(rdev);
+	r = cik_pcie_gart_enable(rdev);
+	if (r)
+		return r;
+	cik_gpu_init(rdev);
+
+	/* allocate rlc buffers */
+	r = si_rlc_init(rdev);
+	if (r) {
+		DRM_ERROR("Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
+	/* allocate mec buffers */
+	r = cik_mec_init(rdev);
+	if (r) {
+		DRM_ERROR("Failed to init MEC BOs!\n");
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_DMA_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing DMA fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_DMA1_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing DMA fences (%d).\n", r);
+		return r;
+	}
+
+	r = cik_uvd_resume(rdev);
+	if (!r) {
+		r = radeon_fence_driver_start_ring(rdev,
+						   R600_RING_TYPE_UVD_INDEX);
+		if (r)
+			dev_err(rdev->dev, "UVD fences init error (%d).\n", r);
+	}
+	if (r)
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+
+	/* Enable IRQ */
+	if (!rdev->irq.installed) {
+		r = radeon_irq_kms_init(rdev);
+		if (r)
+			return r;
+	}
+
+	r = cik_irq_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: IH init failed (%d).\n", r);
+		radeon_irq_kms_fini(rdev);
+		return r;
+	}
+	cik_irq_set(rdev);
+
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
+			     CP_RB0_RPTR, CP_RB0_WPTR,
+			     0, 0xfffff, RADEON_CP_PACKET2);
+	if (r)
+		return r;
+
+	/* set up the compute queues */
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET,
+			     CP_HQD_PQ_RPTR, CP_HQD_PQ_WPTR,
+			     0, 0xfffff, PACKET3(PACKET3_NOP, 0x3FFF));
+	if (r)
+		return r;
+	ring->me = 1; /* first MEC */
+	ring->pipe = 0; /* first pipe */
+	ring->queue = 0; /* first queue */
+	ring->wptr_offs = CIK_WB_CP1_WPTR_OFFSET;
+
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET,
+			     CP_HQD_PQ_RPTR, CP_HQD_PQ_WPTR,
+			     0, 0xffffffff, PACKET3(PACKET3_NOP, 0x3FFF));
+	if (r)
+		return r;
+	/* dGPU only have 1 MEC */
+	ring->me = 1; /* first MEC */
+	ring->pipe = 0; /* first pipe */
+	ring->queue = 1; /* second queue */
+	ring->wptr_offs = CIK_WB_CP2_WPTR_OFFSET;
+
+	ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET,
+			     SDMA0_GFX_RB_RPTR + SDMA0_REGISTER_OFFSET,
+			     SDMA0_GFX_RB_WPTR + SDMA0_REGISTER_OFFSET,
+			     2, 0xfffffffc, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
+	if (r)
+		return r;
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET,
+			     SDMA0_GFX_RB_RPTR + SDMA1_REGISTER_OFFSET,
+			     SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET,
+			     2, 0xfffffffc, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
+	if (r)
+		return r;
+
+	r = cik_cp_resume(rdev);
+	if (r)
+		return r;
+
+	r = cik_sdma_resume(rdev);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	if (ring->ring_size) {
+		r = radeon_ring_init(rdev, ring, ring->ring_size,
+				     R600_WB_UVD_RPTR_OFFSET,
+				     UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
+				     0, 0xfffff, RADEON_CP_PACKET2);
+		if (!r)
+			r = r600_uvd_init(rdev);
+		if (r)
+			DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+	}
+
+	r = radeon_ib_pool_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_vm_manager_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "vm manager initialization failed (%d).\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+/**
+ * cik_resume - resume the asic to a functional state
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Programs the asic to a functional state (CIK).
+ * Called at resume.
+ * Returns 0 for success, error for failure.
+ */
+int cik_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* post card */
+	atom_asic_init(rdev->mode_info.atom_context);
+
+	/* init golden registers */
+	cik_init_golden_registers(rdev);
+
+	rdev->accel_working = true;
+	r = cik_startup(rdev);
+	if (r) {
+		DRM_ERROR("cik startup failed on resume\n");
+		rdev->accel_working = false;
+		return r;
+	}
+
+	return r;
+
+}
+
+/**
+ * cik_suspend - suspend the asic
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Bring the chip into a state suitable for suspend (CIK).
+ * Called at suspend.
+ * Returns 0 for success.
+ */
+int cik_suspend(struct radeon_device *rdev)
+{
+	radeon_vm_manager_fini(rdev);
+	cik_cp_enable(rdev, false);
+	cik_sdma_enable(rdev, false);
+	r600_uvd_rbc_stop(rdev);
+	radeon_uvd_suspend(rdev);
+	cik_irq_suspend(rdev);
+	radeon_wb_disable(rdev);
+	cik_pcie_gart_disable(rdev);
+	return 0;
+}
+
+/* Plan is to move initialization in that function and use
+ * helper function so that radeon_device_init pretty much
+ * do nothing more than calling asic specific function. This
+ * should also allow to remove a bunch of callback function
+ * like vram_info.
+ */
+/**
+ * cik_init - asic specific driver and hw init
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Setup asic specific driver variables and program the hw
+ * to a functional state (CIK).
+ * Called at driver startup.
+ * Returns 0 for success, errors for failure.
+ */
+int cik_init(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	int r;
+
+	/* Read BIOS */
+	if (!radeon_get_bios(rdev)) {
+		if (ASIC_IS_AVIVO(rdev))
+			return -EINVAL;
+	}
+	/* Must be an ATOMBIOS */
+	if (!rdev->is_atom_bios) {
+		dev_err(rdev->dev, "Expecting atombios for cayman GPU\n");
+		return -EINVAL;
+	}
+	r = radeon_atombios_init(rdev);
+	if (r)
+		return r;
+
+	/* Post card if necessary */
+	if (!radeon_card_posted(rdev)) {
+		if (!rdev->bios) {
+			dev_err(rdev->dev, "Card not posted and no BIOS - ignoring\n");
+			return -EINVAL;
+		}
+		DRM_INFO("GPU not posted. posting now...\n");
+		atom_asic_init(rdev->mode_info.atom_context);
+	}
+	/* init golden registers */
+	cik_init_golden_registers(rdev);
+	/* Initialize scratch registers */
+	cik_scratch_init(rdev);
+	/* Initialize surface registers */
+	radeon_surface_init(rdev);
+	/* Initialize clocks */
+	radeon_get_clock_info(rdev->ddev);
+
+	/* Fence driver */
+	r = radeon_fence_driver_init(rdev);
+	if (r)
+		return r;
+
+	/* initialize memory controller */
+	r = cik_mc_init(rdev);
+	if (r)
+		return r;
+	/* Memory manager */
+	r = radeon_bo_init(rdev);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+	r = radeon_doorbell_get(rdev, &ring->doorbell_page_num);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+	r = radeon_doorbell_get(rdev, &ring->doorbell_page_num);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 256 * 1024);
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 256 * 1024);
+
+	r = radeon_uvd_init(rdev);
+	if (!r) {
+		ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+		ring->ring_obj = NULL;
+		r600_ring_init(rdev, ring, 4096);
+	}
+
+	rdev->ih.ring_obj = NULL;
+	r600_ih_ring_init(rdev, 64 * 1024);
+
+	r = r600_pcie_gart_init(rdev);
+	if (r)
+		return r;
+
+	rdev->accel_working = true;
+	r = cik_startup(rdev);
+	if (r) {
+		dev_err(rdev->dev, "disabling GPU acceleration\n");
+		cik_cp_fini(rdev);
+		cik_sdma_fini(rdev);
+		cik_irq_fini(rdev);
+		si_rlc_fini(rdev);
+		cik_mec_fini(rdev);
+		radeon_wb_fini(rdev);
+		radeon_ib_pool_fini(rdev);
+		radeon_vm_manager_fini(rdev);
+		radeon_irq_kms_fini(rdev);
+		cik_pcie_gart_fini(rdev);
+		rdev->accel_working = false;
+	}
+
+	/* Don't start up if the MC ucode is missing.
+	 * The default clocks and voltages before the MC ucode
+	 * is loaded are not suffient for advanced operations.
+	 */
+	if (!rdev->mc_fw && !(rdev->flags & RADEON_IS_IGP)) {
+		DRM_ERROR("radeon: MC ucode required for NI+.\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * cik_fini - asic specific driver and hw fini
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down the asic specific driver variables and program the hw
+ * to an idle state (CIK).
+ * Called at driver unload.
+ */
+void cik_fini(struct radeon_device *rdev)
+{
+	cik_cp_fini(rdev);
+	cik_sdma_fini(rdev);
+	cik_irq_fini(rdev);
+	si_rlc_fini(rdev);
+	cik_mec_fini(rdev);
+	radeon_wb_fini(rdev);
+	radeon_vm_manager_fini(rdev);
+	radeon_ib_pool_fini(rdev);
+	radeon_irq_kms_fini(rdev);
+	radeon_uvd_fini(rdev);
+	cik_pcie_gart_fini(rdev);
+	r600_vram_scratch_fini(rdev);
+	radeon_gem_fini(rdev);
+	radeon_fence_driver_fini(rdev);
+	radeon_bo_fini(rdev);
+	radeon_atombios_fini(rdev);
+	kfree(rdev->bios);
+	rdev->bios = NULL;
+}
+
+/* display watermark setup */
+/**
+ * dce8_line_buffer_adjust - Set up the line buffer
+ *
+ * @rdev: radeon_device pointer
+ * @radeon_crtc: the selected display controller
+ * @mode: the current display mode on the selected display
+ * controller
+ *
+ * Setup up the line buffer allocation for
+ * the selected display controller (CIK).
+ * Returns the line buffer size in pixels.
+ */
+static u32 dce8_line_buffer_adjust(struct radeon_device *rdev,
+				   struct radeon_crtc *radeon_crtc,
+				   struct drm_display_mode *mode)
+{
+	u32 tmp;
+
+	/*
+	 * Line Buffer Setup
+	 * There are 6 line buffers, one for each display controllers.
+	 * There are 3 partitions per LB. Select the number of partitions
+	 * to enable based on the display width.  For display widths larger
+	 * than 4096, you need use to use 2 display controllers and combine
+	 * them using the stereo blender.
+	 */
+	if (radeon_crtc->base.enabled && mode) {
+		if (mode->crtc_hdisplay < 1920)
+			tmp = 1;
+		else if (mode->crtc_hdisplay < 2560)
+			tmp = 2;
+		else if (mode->crtc_hdisplay < 4096)
+			tmp = 0;
+		else {
+			DRM_DEBUG_KMS("Mode too big for LB!\n");
+			tmp = 0;
+		}
+	} else
+		tmp = 1;
+
+	WREG32(LB_MEMORY_CTRL + radeon_crtc->crtc_offset,
+	       LB_MEMORY_CONFIG(tmp) | LB_MEMORY_SIZE(0x6B0));
+
+	if (radeon_crtc->base.enabled && mode) {
+		switch (tmp) {
+		case 0:
+		default:
+			return 4096 * 2;
+		case 1:
+			return 1920 * 2;
+		case 2:
+			return 2560 * 2;
+		}
+	}
+
+	/* controller not enabled, so no lb used */
+	return 0;
+}
+
+/**
+ * cik_get_number_of_dram_channels - get the number of dram channels
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Look up the number of video ram channels (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the number of dram channels
+ */
+static u32 cik_get_number_of_dram_channels(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(MC_SHARED_CHMAP);
+
+	switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) {
+	case 0:
+	default:
+		return 1;
+	case 1:
+		return 2;
+	case 2:
+		return 4;
+	case 3:
+		return 8;
+	case 4:
+		return 3;
+	case 5:
+		return 6;
+	case 6:
+		return 10;
+	case 7:
+		return 12;
+	case 8:
+		return 16;
+	}
+}
+
+struct dce8_wm_params {
+	u32 dram_channels; /* number of dram channels */
+	u32 yclk;          /* bandwidth per dram data pin in kHz */
+	u32 sclk;          /* engine clock in kHz */
+	u32 disp_clk;      /* display clock in kHz */
+	u32 src_width;     /* viewport width */
+	u32 active_time;   /* active display time in ns */
+	u32 blank_time;    /* blank time in ns */
+	bool interlaced;    /* mode is interlaced */
+	fixed20_12 vsc;    /* vertical scale ratio */
+	u32 num_heads;     /* number of active crtcs */
+	u32 bytes_per_pixel; /* bytes per pixel display + overlay */
+	u32 lb_size;       /* line buffer allocated to pipe */
+	u32 vtaps;         /* vertical scaler taps */
+};
+
+/**
+ * dce8_dram_bandwidth - get the dram bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the raw dram bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dram bandwidth in MBytes/s
+ */
+static u32 dce8_dram_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate raw DRAM Bandwidth */
+	fixed20_12 dram_efficiency; /* 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	dram_efficiency.full = dfixed_const(7);
+	dram_efficiency.full = dfixed_div(dram_efficiency, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, dram_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_dram_bandwidth_for_display - get the dram bandwidth for display
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the dram bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dram bandwidth for display in MBytes/s
+ */
+static u32 dce8_dram_bandwidth_for_display(struct dce8_wm_params *wm)
+{
+	/* Calculate DRAM Bandwidth and the part allocated to display. */
+	fixed20_12 disp_dram_allocation; /* 0.3 to 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	disp_dram_allocation.full = dfixed_const(3); /* XXX worse case value 0.3 */
+	disp_dram_allocation.full = dfixed_div(disp_dram_allocation, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, disp_dram_allocation);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_data_return_bandwidth - get the data return bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the data return bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the data return bandwidth in MBytes/s
+ */
+static u32 dce8_data_return_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the display Data return Bandwidth */
+	fixed20_12 return_efficiency; /* 0.8 */
+	fixed20_12 sclk, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	sclk.full = dfixed_const(wm->sclk);
+	sclk.full = dfixed_div(sclk, a);
+	a.full = dfixed_const(10);
+	return_efficiency.full = dfixed_const(8);
+	return_efficiency.full = dfixed_div(return_efficiency, a);
+	a.full = dfixed_const(32);
+	bandwidth.full = dfixed_mul(a, sclk);
+	bandwidth.full = dfixed_mul(bandwidth, return_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_dmif_request_bandwidth - get the dmif bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the dmif bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dmif bandwidth in MBytes/s
+ */
+static u32 dce8_dmif_request_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the DMIF Request Bandwidth */
+	fixed20_12 disp_clk_request_efficiency; /* 0.8 */
+	fixed20_12 disp_clk, bandwidth;
+	fixed20_12 a, b;
+
+	a.full = dfixed_const(1000);
+	disp_clk.full = dfixed_const(wm->disp_clk);
+	disp_clk.full = dfixed_div(disp_clk, a);
+	a.full = dfixed_const(32);
+	b.full = dfixed_mul(a, disp_clk);
+
+	a.full = dfixed_const(10);
+	disp_clk_request_efficiency.full = dfixed_const(8);
+	disp_clk_request_efficiency.full = dfixed_div(disp_clk_request_efficiency, a);
+
+	bandwidth.full = dfixed_mul(b, disp_clk_request_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_available_bandwidth - get the min available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the min available bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the min available bandwidth in MBytes/s
+ */
+static u32 dce8_available_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the Available bandwidth. Display can use this temporarily but not in average. */
+	u32 dram_bandwidth = dce8_dram_bandwidth(wm);
+	u32 data_return_bandwidth = dce8_data_return_bandwidth(wm);
+	u32 dmif_req_bandwidth = dce8_dmif_request_bandwidth(wm);
+
+	return min(dram_bandwidth, min(data_return_bandwidth, dmif_req_bandwidth));
+}
+
+/**
+ * dce8_average_bandwidth - get the average available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the average available bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the average available bandwidth in MBytes/s
+ */
+static u32 dce8_average_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the display mode Average Bandwidth
+	 * DisplayMode should contain the source and destination dimensions,
+	 * timing, etc.
+	 */
+	fixed20_12 bpp;
+	fixed20_12 line_time;
+	fixed20_12 src_width;
+	fixed20_12 bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	line_time.full = dfixed_const(wm->active_time + wm->blank_time);
+	line_time.full = dfixed_div(line_time, a);
+	bpp.full = dfixed_const(wm->bytes_per_pixel);
+	src_width.full = dfixed_const(wm->src_width);
+	bandwidth.full = dfixed_mul(src_width, bpp);
+	bandwidth.full = dfixed_mul(bandwidth, wm->vsc);
+	bandwidth.full = dfixed_div(bandwidth, line_time);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_latency_watermark - get the latency watermark
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the latency watermark (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the latency watermark in ns
+ */
+static u32 dce8_latency_watermark(struct dce8_wm_params *wm)
+{
+	/* First calculate the latency in ns */
+	u32 mc_latency = 2000; /* 2000 ns. */
+	u32 available_bandwidth = dce8_available_bandwidth(wm);
+	u32 worst_chunk_return_time = (512 * 8 * 1000) / available_bandwidth;
+	u32 cursor_line_pair_return_time = (128 * 4 * 1000) / available_bandwidth;
+	u32 dc_latency = 40000000 / wm->disp_clk; /* dc pipe latency */
+	u32 other_heads_data_return_time = ((wm->num_heads + 1) * worst_chunk_return_time) +
+		(wm->num_heads * cursor_line_pair_return_time);
+	u32 latency = mc_latency + other_heads_data_return_time + dc_latency;
+	u32 max_src_lines_per_dst_line, lb_fill_bw, line_fill_time;
+	u32 tmp, dmif_size = 12288;
+	fixed20_12 a, b, c;
+
+	if (wm->num_heads == 0)
+		return 0;
+
+	a.full = dfixed_const(2);
+	b.full = dfixed_const(1);
+	if ((wm->vsc.full > a.full) ||
+	    ((wm->vsc.full > b.full) && (wm->vtaps >= 3)) ||
+	    (wm->vtaps >= 5) ||
+	    ((wm->vsc.full >= a.full) && wm->interlaced))
+		max_src_lines_per_dst_line = 4;
+	else
+		max_src_lines_per_dst_line = 2;
+
+	a.full = dfixed_const(available_bandwidth);
+	b.full = dfixed_const(wm->num_heads);
+	a.full = dfixed_div(a, b);
+
+	b.full = dfixed_const(mc_latency + 512);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(b, c);
+
+	c.full = dfixed_const(dmif_size);
+	b.full = dfixed_div(c, b);
+
+	tmp = min(dfixed_trunc(a), dfixed_trunc(b));
+
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(c, b);
+	c.full = dfixed_const(wm->bytes_per_pixel);
+	b.full = dfixed_mul(b, c);
+
+	lb_fill_bw = min(tmp, dfixed_trunc(b));
+
+	a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel);
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(lb_fill_bw);
+	b.full = dfixed_div(c, b);
+	a.full = dfixed_div(a, b);
+	line_fill_time = dfixed_trunc(a);
+
+	if (line_fill_time < wm->active_time)
+		return latency;
+	else
+		return latency + (line_fill_time - wm->active_time);
+
+}
+
+/**
+ * dce8_average_bandwidth_vs_dram_bandwidth_for_display - check
+ * average and available dram bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Check if the display average bandwidth fits in the display
+ * dram bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce8_average_bandwidth_vs_dram_bandwidth_for_display(struct dce8_wm_params *wm)
+{
+	if (dce8_average_bandwidth(wm) <=
+	    (dce8_dram_bandwidth_for_display(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce8_average_bandwidth_vs_available_bandwidth - check
+ * average and available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Check if the display average bandwidth fits in the display
+ * available bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce8_average_bandwidth_vs_available_bandwidth(struct dce8_wm_params *wm)
+{
+	if (dce8_average_bandwidth(wm) <=
+	    (dce8_available_bandwidth(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce8_check_latency_hiding - check latency hiding
+ *
+ * @wm: watermark calculation data
+ *
+ * Check latency hiding (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce8_check_latency_hiding(struct dce8_wm_params *wm)
+{
+	u32 lb_partitions = wm->lb_size / wm->src_width;
+	u32 line_time = wm->active_time + wm->blank_time;
+	u32 latency_tolerant_lines;
+	u32 latency_hiding;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1);
+	if (wm->vsc.full > a.full)
+		latency_tolerant_lines = 1;
+	else {
+		if (lb_partitions <= (wm->vtaps + 1))
+			latency_tolerant_lines = 1;
+		else
+			latency_tolerant_lines = 2;
+	}
+
+	latency_hiding = (latency_tolerant_lines * line_time + wm->blank_time);
+
+	if (dce8_latency_watermark(wm) <= latency_hiding)
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce8_program_watermarks - program display watermarks
+ *
+ * @rdev: radeon_device pointer
+ * @radeon_crtc: the selected display controller
+ * @lb_size: line buffer size
+ * @num_heads: number of display controllers in use
+ *
+ * Calculate and program the display watermarks for the
+ * selected display controller (CIK).
+ */
+static void dce8_program_watermarks(struct radeon_device *rdev,
+				    struct radeon_crtc *radeon_crtc,
+				    u32 lb_size, u32 num_heads)
+{
+	struct drm_display_mode *mode = &radeon_crtc->base.mode;
+	struct dce8_wm_params wm;
+	u32 pixel_period;
+	u32 line_time = 0;
+	u32 latency_watermark_a = 0, latency_watermark_b = 0;
+	u32 tmp, wm_mask;
+
+	if (radeon_crtc->base.enabled && num_heads && mode) {
+		pixel_period = 1000000 / (u32)mode->clock;
+		line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535);
+
+		wm.yclk = rdev->pm.current_mclk * 10;
+		wm.sclk = rdev->pm.current_sclk * 10;
+		wm.disp_clk = mode->clock;
+		wm.src_width = mode->crtc_hdisplay;
+		wm.active_time = mode->crtc_hdisplay * pixel_period;
+		wm.blank_time = line_time - wm.active_time;
+		wm.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm.interlaced = true;
+		wm.vsc = radeon_crtc->vsc;
+		wm.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm.vtaps = 2;
+		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm.lb_size = lb_size;
+		wm.dram_channels = cik_get_number_of_dram_channels(rdev);
+		wm.num_heads = num_heads;
+
+		/* set for high clocks */
+		latency_watermark_a = min(dce8_latency_watermark(&wm), (u32)65535);
+		/* set for low clocks */
+		/* wm.yclk = low clk; wm.sclk = low clk */
+		latency_watermark_b = min(dce8_latency_watermark(&wm), (u32)65535);
+
+		/* possibly force display priority to high */
+		/* should really do this at mode validation time... */
+		if (!dce8_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
+		    !dce8_average_bandwidth_vs_available_bandwidth(&wm) ||
+		    !dce8_check_latency_hiding(&wm) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+		}
+	}
+
+	/* select wm A */
+	wm_mask = RREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset);
+	tmp = wm_mask;
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(1);
+	WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_a) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* select wm B */
+	tmp = RREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset);
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(2);
+	WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_b) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* restore original selection */
+	WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, wm_mask);
+}
+
+/**
+ * dce8_bandwidth_update - program display watermarks
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Calculate and program the display watermarks and line
+ * buffer allocation (CIK).
+ */
+void dce8_bandwidth_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode = NULL;
+	u32 num_heads = 0, lb_size;
+	int i;
+
+	radeon_update_display_priority(rdev);
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (rdev->mode_info.crtcs[i]->base.enabled)
+			num_heads++;
+	}
+	for (i = 0; i < rdev->num_crtc; i++) {
+		mode = &rdev->mode_info.crtcs[i]->base.mode;
+		lb_size = dce8_line_buffer_adjust(rdev, rdev->mode_info.crtcs[i], mode);
+		dce8_program_watermarks(rdev, rdev->mode_info.crtcs[i], lb_size, num_heads);
+	}
+}
+
+/**
+ * cik_get_gpu_clock_counter - return GPU clock counter snapshot
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Fetches a GPU clock counter snapshot (SI).
+ * Returns the 64 bit clock counter snapshot.
+ */
+uint64_t cik_get_gpu_clock_counter(struct radeon_device *rdev)
+{
+	uint64_t clock;
+
+	mutex_lock(&rdev->gpu_clock_mutex);
+	WREG32(RLC_CAPTURE_GPU_CLOCK_COUNT, 1);
+	clock = (uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_LSB) |
+	        ((uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_MSB) << 32ULL);
+	mutex_unlock(&rdev->gpu_clock_mutex);
+	return clock;
+}
+
+static int cik_set_uvd_clock(struct radeon_device *rdev, u32 clock,
+                              u32 cntl_reg, u32 status_reg)
+{
+	int r, i;
+	struct atom_clock_dividers dividers;
+	uint32_t tmp;
+
+	r = radeon_atom_get_clock_dividers(rdev, COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK,
+					   clock, false, &dividers);
+	if (r)
+		return r;
+
+	tmp = RREG32_SMC(cntl_reg);
+	tmp &= ~(DCLK_DIR_CNTL_EN|DCLK_DIVIDER_MASK);
+	tmp |= dividers.post_divider;
+	WREG32_SMC(cntl_reg, tmp);
+
+	for (i = 0; i < 100; i++) {
+		if (RREG32_SMC(status_reg) & DCLK_STATUS)
+			break;
+		mdelay(10);
+	}
+	if (i == 100)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
+{
+	int r = 0;
+
+	r = cik_set_uvd_clock(rdev, vclk, CG_VCLK_CNTL, CG_VCLK_STATUS);
+	if (r)
+		return r;
+
+	r = cik_set_uvd_clock(rdev, dclk, CG_DCLK_CNTL, CG_DCLK_STATUS);
+	return r;
+}
+
+int cik_uvd_resume(struct radeon_device *rdev)
+{
+	uint64_t addr;
+	uint32_t size;
+	int r;
+
+	r = radeon_uvd_resume(rdev);
+	if (r)
+		return r;
+
+	/* programm the VCPU memory controller bits 0-27 */
+	addr = rdev->uvd.gpu_addr >> 3;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = RADEON_UVD_STACK_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET1, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = RADEON_UVD_HEAP_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET2, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE2, size);
+
+	/* bits 28-31 */
+	addr = (rdev->uvd.gpu_addr >> 28) & 0xF;
+	WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0));
+
+	/* bits 32-39 */
+	addr = (rdev->uvd.gpu_addr >> 32) & 0xFF;
+	WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31));
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/cik_blit_shaders.c b/drivers/gpu/drm/radeon/cik_blit_shaders.c
new file mode 100644
index 0000000..ff13118
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik_blit_shaders.c
@@ -0,0 +1,246 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *     Alex Deucher <alexander.deucher@amd.com>
+ */
+
+#include <linux/types.h>
+#include <linux/bug.h>
+#include <linux/kernel.h>
+
+const u32 cik_default_state[] =
+{
+	0xc0066900,
+	0x00000000,
+	0x00000060, /* DB_RENDER_CONTROL */
+	0x00000000, /* DB_COUNT_CONTROL */
+	0x00000000, /* DB_DEPTH_VIEW */
+	0x0000002a, /* DB_RENDER_OVERRIDE */
+	0x00000000, /* DB_RENDER_OVERRIDE2 */
+	0x00000000, /* DB_HTILE_DATA_BASE */
+
+	0xc0046900,
+	0x00000008,
+	0x00000000, /* DB_DEPTH_BOUNDS_MIN */
+	0x00000000, /* DB_DEPTH_BOUNDS_MAX */
+	0x00000000, /* DB_STENCIL_CLEAR */
+	0x00000000, /* DB_DEPTH_CLEAR */
+
+	0xc0036900,
+	0x0000000f,
+	0x00000000, /* DB_DEPTH_INFO */
+	0x00000000, /* DB_Z_INFO */
+	0x00000000, /* DB_STENCIL_INFO */
+
+	0xc0016900,
+	0x00000080,
+	0x00000000, /* PA_SC_WINDOW_OFFSET */
+
+	0xc00d6900,
+	0x00000083,
+	0x0000ffff, /* PA_SC_CLIPRECT_RULE */
+	0x00000000, /* PA_SC_CLIPRECT_0_TL */
+	0x20002000, /* PA_SC_CLIPRECT_0_BR */
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0xaaaaaaaa, /* PA_SC_EDGERULE */
+	0x00000000, /* PA_SU_HARDWARE_SCREEN_OFFSET */
+	0x0000000f, /* CB_TARGET_MASK */
+	0x0000000f, /* CB_SHADER_MASK */
+
+	0xc0226900,
+	0x00000094,
+	0x80000000, /* PA_SC_VPORT_SCISSOR_0_TL */
+	0x20002000, /* PA_SC_VPORT_SCISSOR_0_BR */
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x00000000, /* PA_SC_VPORT_ZMIN_0 */
+	0x3f800000, /* PA_SC_VPORT_ZMAX_0 */
+
+	0xc0046900,
+	0x00000100,
+	0xffffffff, /* VGT_MAX_VTX_INDX */
+	0x00000000, /* VGT_MIN_VTX_INDX */
+	0x00000000, /* VGT_INDX_OFFSET */
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_INDX */
+
+	0xc0046900,
+	0x00000105,
+	0x00000000, /* CB_BLEND_RED */
+	0x00000000, /* CB_BLEND_GREEN */
+	0x00000000, /* CB_BLEND_BLUE */
+	0x00000000, /* CB_BLEND_ALPHA */
+
+	0xc0016900,
+	0x000001e0,
+	0x00000000, /* CB_BLEND0_CONTROL */
+
+	0xc00c6900,
+	0x00000200,
+	0x00000000, /* DB_DEPTH_CONTROL */
+	0x00000000, /* DB_EQAA */
+	0x00cc0010, /* CB_COLOR_CONTROL */
+	0x00000210, /* DB_SHADER_CONTROL */
+	0x00010000, /* PA_CL_CLIP_CNTL */
+	0x00000004, /* PA_SU_SC_MODE_CNTL */
+	0x00000100, /* PA_CL_VTE_CNTL */
+	0x00000000, /* PA_CL_VS_OUT_CNTL */
+	0x00000000, /* PA_CL_NANINF_CNTL */
+	0x00000000, /* PA_SU_LINE_STIPPLE_CNTL */
+	0x00000000, /* PA_SU_LINE_STIPPLE_SCALE */
+	0x00000000, /* PA_SU_PRIM_FILTER_CNTL */
+
+	0xc0116900,
+	0x00000280,
+	0x00000000, /* PA_SU_POINT_SIZE */
+	0x00000000, /* PA_SU_POINT_MINMAX */
+	0x00000008, /* PA_SU_LINE_CNTL */
+	0x00000000, /* PA_SC_LINE_STIPPLE */
+	0x00000000, /* VGT_OUTPUT_PATH_CNTL */
+	0x00000000, /* VGT_HOS_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000, /* VGT_GS_MODE */
+
+	0xc0026900,
+	0x00000292,
+	0x00000000, /* PA_SC_MODE_CNTL_0 */
+	0x00000000, /* PA_SC_MODE_CNTL_1 */
+
+	0xc0016900,
+	0x000002a1,
+	0x00000000, /* VGT_PRIMITIVEID_EN */
+
+	0xc0016900,
+	0x000002a5,
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_EN */
+
+	0xc0026900,
+	0x000002a8,
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_0 */
+	0x00000000,
+
+	0xc0026900,
+	0x000002ad,
+	0x00000000, /* VGT_REUSE_OFF */
+	0x00000000,
+
+	0xc0016900,
+	0x000002d5,
+	0x00000000, /* VGT_SHADER_STAGES_EN */
+
+	0xc0016900,
+	0x000002dc,
+	0x0000aa00, /* DB_ALPHA_TO_MASK */
+
+	0xc0066900,
+	0x000002de,
+	0x00000000, /* PA_SU_POLY_OFFSET_DB_FMT_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0026900,
+	0x000002e5,
+	0x00000000, /* VGT_STRMOUT_CONFIG */
+	0x00000000,
+
+	0xc01b6900,
+	0x000002f5,
+	0x76543210, /* PA_SC_CENTROID_PRIORITY_0 */
+	0xfedcba98, /* PA_SC_CENTROID_PRIORITY_1 */
+	0x00000000, /* PA_SC_LINE_CNTL */
+	0x00000000, /* PA_SC_AA_CONFIG */
+	0x00000005, /* PA_SU_VTX_CNTL */
+	0x3f800000, /* PA_CL_GB_VERT_CLIP_ADJ */
+	0x3f800000, /* PA_CL_GB_VERT_DISC_ADJ */
+	0x3f800000, /* PA_CL_GB_HORZ_CLIP_ADJ */
+	0x3f800000, /* PA_CL_GB_HORZ_DISC_ADJ */
+	0x00000000, /* PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0 */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0xffffffff, /* PA_SC_AA_MASK_X0Y0_X1Y0 */
+	0xffffffff,
+
+	0xc0026900,
+	0x00000316,
+	0x0000000e, /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	0x00000010, /*  */
+};
+
+const u32 cik_default_size = ARRAY_SIZE(cik_default_state);
diff --git a/drivers/gpu/drm/radeon/cik_blit_shaders.h b/drivers/gpu/drm/radeon/cik_blit_shaders.h
new file mode 100644
index 0000000..dfe7314
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik_blit_shaders.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef CIK_BLIT_SHADERS_H
+#define CIK_BLIT_SHADERS_H
+
+extern const u32 cik_default_state[];
+
+extern const u32 cik_default_size;
+
+#endif
diff --git a/drivers/gpu/drm/radeon/cik_reg.h b/drivers/gpu/drm/radeon/cik_reg.h
new file mode 100644
index 0000000..d71e46d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik_reg.h
@@ -0,0 +1,147 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef __CIK_REG_H__
+#define __CIK_REG_H__
+
+#define CIK_DC_GPIO_HPD_MASK                      0x65b0
+#define CIK_DC_GPIO_HPD_A                         0x65b4
+#define CIK_DC_GPIO_HPD_EN                        0x65b8
+#define CIK_DC_GPIO_HPD_Y                         0x65bc
+
+#define CIK_GRPH_CONTROL                          0x6804
+#       define CIK_GRPH_DEPTH(x)                  (((x) & 0x3) << 0)
+#       define CIK_GRPH_DEPTH_8BPP                0
+#       define CIK_GRPH_DEPTH_16BPP               1
+#       define CIK_GRPH_DEPTH_32BPP               2
+#       define CIK_GRPH_NUM_BANKS(x)              (((x) & 0x3) << 2)
+#       define CIK_ADDR_SURF_2_BANK               0
+#       define CIK_ADDR_SURF_4_BANK               1
+#       define CIK_ADDR_SURF_8_BANK               2
+#       define CIK_ADDR_SURF_16_BANK              3
+#       define CIK_GRPH_Z(x)                      (((x) & 0x3) << 4)
+#       define CIK_GRPH_BANK_WIDTH(x)             (((x) & 0x3) << 6)
+#       define CIK_ADDR_SURF_BANK_WIDTH_1         0
+#       define CIK_ADDR_SURF_BANK_WIDTH_2         1
+#       define CIK_ADDR_SURF_BANK_WIDTH_4         2
+#       define CIK_ADDR_SURF_BANK_WIDTH_8         3
+#       define CIK_GRPH_FORMAT(x)                 (((x) & 0x7) << 8)
+/* 8 BPP */
+#       define CIK_GRPH_FORMAT_INDEXED            0
+/* 16 BPP */
+#       define CIK_GRPH_FORMAT_ARGB1555           0
+#       define CIK_GRPH_FORMAT_ARGB565            1
+#       define CIK_GRPH_FORMAT_ARGB4444           2
+#       define CIK_GRPH_FORMAT_AI88               3
+#       define CIK_GRPH_FORMAT_MONO16             4
+#       define CIK_GRPH_FORMAT_BGRA5551           5
+/* 32 BPP */
+#       define CIK_GRPH_FORMAT_ARGB8888           0
+#       define CIK_GRPH_FORMAT_ARGB2101010        1
+#       define CIK_GRPH_FORMAT_32BPP_DIG          2
+#       define CIK_GRPH_FORMAT_8B_ARGB2101010     3
+#       define CIK_GRPH_FORMAT_BGRA1010102        4
+#       define CIK_GRPH_FORMAT_8B_BGRA1010102     5
+#       define CIK_GRPH_FORMAT_RGB111110          6
+#       define CIK_GRPH_FORMAT_BGR101111          7
+#       define CIK_GRPH_BANK_HEIGHT(x)            (((x) & 0x3) << 11)
+#       define CIK_ADDR_SURF_BANK_HEIGHT_1        0
+#       define CIK_ADDR_SURF_BANK_HEIGHT_2        1
+#       define CIK_ADDR_SURF_BANK_HEIGHT_4        2
+#       define CIK_ADDR_SURF_BANK_HEIGHT_8        3
+#       define CIK_GRPH_TILE_SPLIT(x)             (((x) & 0x7) << 13)
+#       define CIK_ADDR_SURF_TILE_SPLIT_64B       0
+#       define CIK_ADDR_SURF_TILE_SPLIT_128B      1
+#       define CIK_ADDR_SURF_TILE_SPLIT_256B      2
+#       define CIK_ADDR_SURF_TILE_SPLIT_512B      3
+#       define CIK_ADDR_SURF_TILE_SPLIT_1KB       4
+#       define CIK_ADDR_SURF_TILE_SPLIT_2KB       5
+#       define CIK_ADDR_SURF_TILE_SPLIT_4KB       6
+#       define CIK_GRPH_MACRO_TILE_ASPECT(x)      (((x) & 0x3) << 18)
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_1  0
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_2  1
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_4  2
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_8  3
+#       define CIK_GRPH_ARRAY_MODE(x)             (((x) & 0x7) << 20)
+#       define CIK_GRPH_ARRAY_LINEAR_GENERAL      0
+#       define CIK_GRPH_ARRAY_LINEAR_ALIGNED      1
+#       define CIK_GRPH_ARRAY_1D_TILED_THIN1      2
+#       define CIK_GRPH_ARRAY_2D_TILED_THIN1      4
+#       define CIK_GRPH_PIPE_CONFIG(x)		 (((x) & 0x1f) << 24)
+#       define CIK_ADDR_SURF_P2			 0
+#       define CIK_ADDR_SURF_P4_8x16		 4
+#       define CIK_ADDR_SURF_P4_16x16		 5
+#       define CIK_ADDR_SURF_P4_16x32		 6
+#       define CIK_ADDR_SURF_P4_32x32		 7
+#       define CIK_ADDR_SURF_P8_16x16_8x16	 8
+#       define CIK_ADDR_SURF_P8_16x32_8x16	 9
+#       define CIK_ADDR_SURF_P8_32x32_8x16	 10
+#       define CIK_ADDR_SURF_P8_16x32_16x16	 11
+#       define CIK_ADDR_SURF_P8_32x32_16x16	 12
+#       define CIK_ADDR_SURF_P8_32x32_16x32	 13
+#       define CIK_ADDR_SURF_P8_32x64_32x32	 14
+#       define CIK_GRPH_MICRO_TILE_MODE(x)       (((x) & 0x7) << 29)
+#       define CIK_DISPLAY_MICRO_TILING          0
+#       define CIK_THIN_MICRO_TILING             1
+#       define CIK_DEPTH_MICRO_TILING            2
+#       define CIK_ROTATED_MICRO_TILING          4
+
+/* CUR blocks at 0x6998, 0x7598, 0x10198, 0x10d98, 0x11998, 0x12598 */
+#define CIK_CUR_CONTROL                           0x6998
+#       define CIK_CURSOR_EN                      (1 << 0)
+#       define CIK_CURSOR_MODE(x)                 (((x) & 0x3) << 8)
+#       define CIK_CURSOR_MONO                    0
+#       define CIK_CURSOR_24_1                    1
+#       define CIK_CURSOR_24_8_PRE_MULT           2
+#       define CIK_CURSOR_24_8_UNPRE_MULT         3
+#       define CIK_CURSOR_2X_MAGNIFY              (1 << 16)
+#       define CIK_CURSOR_FORCE_MC_ON             (1 << 20)
+#       define CIK_CURSOR_URGENT_CONTROL(x)       (((x) & 0x7) << 24)
+#       define CIK_CURSOR_URGENT_ALWAYS           0
+#       define CIK_CURSOR_URGENT_1_8              1
+#       define CIK_CURSOR_URGENT_1_4              2
+#       define CIK_CURSOR_URGENT_3_8              3
+#       define CIK_CURSOR_URGENT_1_2              4
+#define CIK_CUR_SURFACE_ADDRESS                   0x699c
+#       define CIK_CUR_SURFACE_ADDRESS_MASK       0xfffff000
+#define CIK_CUR_SIZE                              0x69a0
+#define CIK_CUR_SURFACE_ADDRESS_HIGH              0x69a4
+#define CIK_CUR_POSITION                          0x69a8
+#define CIK_CUR_HOT_SPOT                          0x69ac
+#define CIK_CUR_COLOR1                            0x69b0
+#define CIK_CUR_COLOR2                            0x69b4
+#define CIK_CUR_UPDATE                            0x69b8
+#       define CIK_CURSOR_UPDATE_PENDING          (1 << 0)
+#       define CIK_CURSOR_UPDATE_TAKEN            (1 << 1)
+#       define CIK_CURSOR_UPDATE_LOCK             (1 << 16)
+#       define CIK_CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24)
+
+#define CIK_ALPHA_CONTROL                         0x6af0
+#       define CIK_CURSOR_ALPHA_BLND_ENA          (1 << 1)
+
+#define CIK_LB_DATA_FORMAT                        0x6b00
+#       define CIK_INTERLEAVE_EN                  (1 << 3)
+
+#define CIK_LB_DESKTOP_HEIGHT                     0x6b0c
+
+#endif
diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h
new file mode 100644
index 0000000..63514b9
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cikd.h
@@ -0,0 +1,1297 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef CIK_H
+#define CIK_H
+
+#define BONAIRE_GB_ADDR_CONFIG_GOLDEN        0x12010001
+
+#define CIK_RB_BITMAP_WIDTH_PER_SH  2
+
+/* SMC IND registers */
+#define GENERAL_PWRMGT                                    0xC0200000
+#       define GPU_COUNTER_CLK                            (1 << 15)
+
+#define CG_CLKPIN_CNTL                                    0xC05001A0
+#       define XTALIN_DIVIDE                              (1 << 1)
+
+#define PCIE_INDEX  					0x38
+#define PCIE_DATA  					0x3C
+
+#define VGA_HDP_CONTROL  				0x328
+#define		VGA_MEMORY_DISABLE				(1 << 4)
+
+#define DMIF_ADDR_CALC  				0xC00
+
+#define	SRBM_GFX_CNTL				        0xE44
+#define		PIPEID(x)					((x) << 0)
+#define		MEID(x)						((x) << 2)
+#define		VMID(x)						((x) << 4)
+#define		QUEUEID(x)					((x) << 8)
+
+#define	SRBM_STATUS2				        0xE4C
+#define		SDMA_BUSY 				(1 << 5)
+#define		SDMA1_BUSY 				(1 << 6)
+#define	SRBM_STATUS				        0xE50
+#define		UVD_RQ_PENDING 				(1 << 1)
+#define		GRBM_RQ_PENDING 			(1 << 5)
+#define		VMC_BUSY 				(1 << 8)
+#define		MCB_BUSY 				(1 << 9)
+#define		MCB_NON_DISPLAY_BUSY 			(1 << 10)
+#define		MCC_BUSY 				(1 << 11)
+#define		MCD_BUSY 				(1 << 12)
+#define		SEM_BUSY 				(1 << 14)
+#define		IH_BUSY 				(1 << 17)
+#define		UVD_BUSY 				(1 << 19)
+
+#define	SRBM_SOFT_RESET				        0xE60
+#define		SOFT_RESET_BIF				(1 << 1)
+#define		SOFT_RESET_R0PLL			(1 << 4)
+#define		SOFT_RESET_DC				(1 << 5)
+#define		SOFT_RESET_SDMA1			(1 << 6)
+#define		SOFT_RESET_GRBM				(1 << 8)
+#define		SOFT_RESET_HDP				(1 << 9)
+#define		SOFT_RESET_IH				(1 << 10)
+#define		SOFT_RESET_MC				(1 << 11)
+#define		SOFT_RESET_ROM				(1 << 14)
+#define		SOFT_RESET_SEM				(1 << 15)
+#define		SOFT_RESET_VMC				(1 << 17)
+#define		SOFT_RESET_SDMA				(1 << 20)
+#define		SOFT_RESET_TST				(1 << 21)
+#define		SOFT_RESET_REGBB		       	(1 << 22)
+#define		SOFT_RESET_ORB				(1 << 23)
+#define		SOFT_RESET_VCE				(1 << 24)
+
+#define VM_L2_CNTL					0x1400
+#define		ENABLE_L2_CACHE					(1 << 0)
+#define		ENABLE_L2_FRAGMENT_PROCESSING			(1 << 1)
+#define		L2_CACHE_PTE_ENDIAN_SWAP_MODE(x)		((x) << 2)
+#define		L2_CACHE_PDE_ENDIAN_SWAP_MODE(x)		((x) << 4)
+#define		ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE		(1 << 9)
+#define		ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE	(1 << 10)
+#define		EFFECTIVE_L2_QUEUE_SIZE(x)			(((x) & 7) << 15)
+#define		CONTEXT1_IDENTITY_ACCESS_MODE(x)		(((x) & 3) << 19)
+#define VM_L2_CNTL2					0x1404
+#define		INVALIDATE_ALL_L1_TLBS				(1 << 0)
+#define		INVALIDATE_L2_CACHE				(1 << 1)
+#define		INVALIDATE_CACHE_MODE(x)			((x) << 26)
+#define			INVALIDATE_PTE_AND_PDE_CACHES		0
+#define			INVALIDATE_ONLY_PTE_CACHES		1
+#define			INVALIDATE_ONLY_PDE_CACHES		2
+#define VM_L2_CNTL3					0x1408
+#define		BANK_SELECT(x)					((x) << 0)
+#define		L2_CACHE_UPDATE_MODE(x)				((x) << 6)
+#define		L2_CACHE_BIGK_FRAGMENT_SIZE(x)			((x) << 15)
+#define		L2_CACHE_BIGK_ASSOCIATIVITY			(1 << 20)
+#define	VM_L2_STATUS					0x140C
+#define		L2_BUSY						(1 << 0)
+#define VM_CONTEXT0_CNTL				0x1410
+#define		ENABLE_CONTEXT					(1 << 0)
+#define		PAGE_TABLE_DEPTH(x)				(((x) & 3) << 1)
+#define		RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 3)
+#define		RANGE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 4)
+#define		DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT	(1 << 6)
+#define		DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT	(1 << 7)
+#define		PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 9)
+#define		PDE0_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 10)
+#define		VALID_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 12)
+#define		VALID_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 13)
+#define		READ_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 15)
+#define		READ_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 16)
+#define		WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 18)
+#define		WRITE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 19)
+#define VM_CONTEXT1_CNTL				0x1414
+#define VM_CONTEXT0_CNTL2				0x1430
+#define VM_CONTEXT1_CNTL2				0x1434
+#define	VM_CONTEXT8_PAGE_TABLE_BASE_ADDR		0x1438
+#define	VM_CONTEXT9_PAGE_TABLE_BASE_ADDR		0x143c
+#define	VM_CONTEXT10_PAGE_TABLE_BASE_ADDR		0x1440
+#define	VM_CONTEXT11_PAGE_TABLE_BASE_ADDR		0x1444
+#define	VM_CONTEXT12_PAGE_TABLE_BASE_ADDR		0x1448
+#define	VM_CONTEXT13_PAGE_TABLE_BASE_ADDR		0x144c
+#define	VM_CONTEXT14_PAGE_TABLE_BASE_ADDR		0x1450
+#define	VM_CONTEXT15_PAGE_TABLE_BASE_ADDR		0x1454
+
+#define VM_INVALIDATE_REQUEST				0x1478
+#define VM_INVALIDATE_RESPONSE				0x147c
+
+#define	VM_CONTEXT1_PROTECTION_FAULT_STATUS		0x14DC
+
+#define	VM_CONTEXT1_PROTECTION_FAULT_ADDR		0x14FC
+
+#define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR	0x1518
+#define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR	0x151c
+
+#define	VM_CONTEXT0_PAGE_TABLE_BASE_ADDR		0x153c
+#define	VM_CONTEXT1_PAGE_TABLE_BASE_ADDR		0x1540
+#define	VM_CONTEXT2_PAGE_TABLE_BASE_ADDR		0x1544
+#define	VM_CONTEXT3_PAGE_TABLE_BASE_ADDR		0x1548
+#define	VM_CONTEXT4_PAGE_TABLE_BASE_ADDR		0x154c
+#define	VM_CONTEXT5_PAGE_TABLE_BASE_ADDR		0x1550
+#define	VM_CONTEXT6_PAGE_TABLE_BASE_ADDR		0x1554
+#define	VM_CONTEXT7_PAGE_TABLE_BASE_ADDR		0x1558
+#define	VM_CONTEXT0_PAGE_TABLE_START_ADDR		0x155c
+#define	VM_CONTEXT1_PAGE_TABLE_START_ADDR		0x1560
+
+#define	VM_CONTEXT0_PAGE_TABLE_END_ADDR			0x157C
+#define	VM_CONTEXT1_PAGE_TABLE_END_ADDR			0x1580
+
+#define MC_SHARED_CHMAP						0x2004
+#define		NOOFCHAN_SHIFT					12
+#define		NOOFCHAN_MASK					0x0000f000
+#define MC_SHARED_CHREMAP					0x2008
+
+#define CHUB_CONTROL					0x1864
+#define		BYPASS_VM					(1 << 0)
+
+#define	MC_VM_FB_LOCATION				0x2024
+#define	MC_VM_AGP_TOP					0x2028
+#define	MC_VM_AGP_BOT					0x202C
+#define	MC_VM_AGP_BASE					0x2030
+#define	MC_VM_SYSTEM_APERTURE_LOW_ADDR			0x2034
+#define	MC_VM_SYSTEM_APERTURE_HIGH_ADDR			0x2038
+#define	MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR		0x203C
+
+#define	MC_VM_MX_L1_TLB_CNTL				0x2064
+#define		ENABLE_L1_TLB					(1 << 0)
+#define		ENABLE_L1_FRAGMENT_PROCESSING			(1 << 1)
+#define		SYSTEM_ACCESS_MODE_PA_ONLY			(0 << 3)
+#define		SYSTEM_ACCESS_MODE_USE_SYS_MAP			(1 << 3)
+#define		SYSTEM_ACCESS_MODE_IN_SYS			(2 << 3)
+#define		SYSTEM_ACCESS_MODE_NOT_IN_SYS			(3 << 3)
+#define		SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU	(0 << 5)
+#define		ENABLE_ADVANCED_DRIVER_MODEL			(1 << 6)
+#define	MC_VM_FB_OFFSET					0x2068
+
+#define MC_SHARED_BLACKOUT_CNTL           		0x20ac
+
+#define	MC_ARB_RAMCFG					0x2760
+#define		NOOFBANK_SHIFT					0
+#define		NOOFBANK_MASK					0x00000003
+#define		NOOFRANK_SHIFT					2
+#define		NOOFRANK_MASK					0x00000004
+#define		NOOFROWS_SHIFT					3
+#define		NOOFROWS_MASK					0x00000038
+#define		NOOFCOLS_SHIFT					6
+#define		NOOFCOLS_MASK					0x000000C0
+#define		CHANSIZE_SHIFT					8
+#define		CHANSIZE_MASK					0x00000100
+#define		NOOFGROUPS_SHIFT				12
+#define		NOOFGROUPS_MASK					0x00001000
+
+#define MC_SEQ_SUP_CNTL           			0x28c8
+#define		RUN_MASK      				(1 << 0)
+#define MC_SEQ_SUP_PGM           			0x28cc
+
+#define	MC_SEQ_TRAIN_WAKEUP_CNTL			0x28e8
+#define		TRAIN_DONE_D0      			(1 << 30)
+#define		TRAIN_DONE_D1      			(1 << 31)
+
+#define MC_IO_PAD_CNTL_D0           			0x29d0
+#define		MEM_FALL_OUT_CMD      			(1 << 8)
+
+#define MC_SEQ_IO_DEBUG_INDEX           		0x2a44
+#define MC_SEQ_IO_DEBUG_DATA           			0x2a48
+
+#define	HDP_HOST_PATH_CNTL				0x2C00
+#define	HDP_NONSURFACE_BASE				0x2C04
+#define	HDP_NONSURFACE_INFO				0x2C08
+#define	HDP_NONSURFACE_SIZE				0x2C0C
+
+#define HDP_ADDR_CONFIG  				0x2F48
+#define HDP_MISC_CNTL					0x2F4C
+#define 	HDP_FLUSH_INVALIDATE_CACHE			(1 << 0)
+
+#define IH_RB_CNTL                                        0x3e00
+#       define IH_RB_ENABLE                               (1 << 0)
+#       define IH_RB_SIZE(x)                              ((x) << 1) /* log2 */
+#       define IH_RB_FULL_DRAIN_ENABLE                    (1 << 6)
+#       define IH_WPTR_WRITEBACK_ENABLE                   (1 << 8)
+#       define IH_WPTR_WRITEBACK_TIMER(x)                 ((x) << 9) /* log2 */
+#       define IH_WPTR_OVERFLOW_ENABLE                    (1 << 16)
+#       define IH_WPTR_OVERFLOW_CLEAR                     (1 << 31)
+#define IH_RB_BASE                                        0x3e04
+#define IH_RB_RPTR                                        0x3e08
+#define IH_RB_WPTR                                        0x3e0c
+#       define RB_OVERFLOW                                (1 << 0)
+#       define WPTR_OFFSET_MASK                           0x3fffc
+#define IH_RB_WPTR_ADDR_HI                                0x3e10
+#define IH_RB_WPTR_ADDR_LO                                0x3e14
+#define IH_CNTL                                           0x3e18
+#       define ENABLE_INTR                                (1 << 0)
+#       define IH_MC_SWAP(x)                              ((x) << 1)
+#       define IH_MC_SWAP_NONE                            0
+#       define IH_MC_SWAP_16BIT                           1
+#       define IH_MC_SWAP_32BIT                           2
+#       define IH_MC_SWAP_64BIT                           3
+#       define RPTR_REARM                                 (1 << 4)
+#       define MC_WRREQ_CREDIT(x)                         ((x) << 15)
+#       define MC_WR_CLEAN_CNT(x)                         ((x) << 20)
+#       define MC_VMID(x)                                 ((x) << 25)
+
+#define	CONFIG_MEMSIZE					0x5428
+
+#define INTERRUPT_CNTL                                    0x5468
+#       define IH_DUMMY_RD_OVERRIDE                       (1 << 0)
+#       define IH_DUMMY_RD_EN                             (1 << 1)
+#       define IH_REQ_NONSNOOP_EN                         (1 << 3)
+#       define GEN_IH_INT_EN                              (1 << 8)
+#define INTERRUPT_CNTL2                                   0x546c
+
+#define HDP_MEM_COHERENCY_FLUSH_CNTL			0x5480
+
+#define	BIF_FB_EN						0x5490
+#define		FB_READ_EN					(1 << 0)
+#define		FB_WRITE_EN					(1 << 1)
+
+#define HDP_REG_COHERENCY_FLUSH_CNTL			0x54A0
+
+#define GPU_HDP_FLUSH_REQ				0x54DC
+#define GPU_HDP_FLUSH_DONE				0x54E0
+#define		CP0					(1 << 0)
+#define		CP1					(1 << 1)
+#define		CP2					(1 << 2)
+#define		CP3					(1 << 3)
+#define		CP4					(1 << 4)
+#define		CP5					(1 << 5)
+#define		CP6					(1 << 6)
+#define		CP7					(1 << 7)
+#define		CP8					(1 << 8)
+#define		CP9					(1 << 9)
+#define		SDMA0					(1 << 10)
+#define		SDMA1					(1 << 11)
+
+/* 0x6b04, 0x7704, 0x10304, 0x10f04, 0x11b04, 0x12704 */
+#define	LB_MEMORY_CTRL					0x6b04
+#define		LB_MEMORY_SIZE(x)			((x) << 0)
+#define		LB_MEMORY_CONFIG(x)			((x) << 20)
+
+#define	DPG_WATERMARK_MASK_CONTROL			0x6cc8
+#       define LATENCY_WATERMARK_MASK(x)		((x) << 8)
+#define	DPG_PIPE_LATENCY_CONTROL			0x6ccc
+#       define LATENCY_LOW_WATERMARK(x)			((x) << 0)
+#       define LATENCY_HIGH_WATERMARK(x)		((x) << 16)
+
+/* 0x6b24, 0x7724, 0x10324, 0x10f24, 0x11b24, 0x12724 */
+#define LB_VLINE_STATUS                                 0x6b24
+#       define VLINE_OCCURRED                           (1 << 0)
+#       define VLINE_ACK                                (1 << 4)
+#       define VLINE_STAT                               (1 << 12)
+#       define VLINE_INTERRUPT                          (1 << 16)
+#       define VLINE_INTERRUPT_TYPE                     (1 << 17)
+/* 0x6b2c, 0x772c, 0x1032c, 0x10f2c, 0x11b2c, 0x1272c */
+#define LB_VBLANK_STATUS                                0x6b2c
+#       define VBLANK_OCCURRED                          (1 << 0)
+#       define VBLANK_ACK                               (1 << 4)
+#       define VBLANK_STAT                              (1 << 12)
+#       define VBLANK_INTERRUPT                         (1 << 16)
+#       define VBLANK_INTERRUPT_TYPE                    (1 << 17)
+
+/* 0x6b20, 0x7720, 0x10320, 0x10f20, 0x11b20, 0x12720 */
+#define LB_INTERRUPT_MASK                               0x6b20
+#       define VBLANK_INTERRUPT_MASK                    (1 << 0)
+#       define VLINE_INTERRUPT_MASK                     (1 << 4)
+#       define VLINE2_INTERRUPT_MASK                    (1 << 8)
+
+#define DISP_INTERRUPT_STATUS                           0x60f4
+#       define LB_D1_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D1_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD1_INTERRUPT                        (1 << 17)
+#       define DC_HPD1_RX_INTERRUPT                     (1 << 18)
+#       define DACA_AUTODETECT_INTERRUPT                (1 << 22)
+#       define DACB_AUTODETECT_INTERRUPT                (1 << 23)
+#       define DC_I2C_SW_DONE_INTERRUPT                 (1 << 24)
+#       define DC_I2C_HW_DONE_INTERRUPT                 (1 << 25)
+#define DISP_INTERRUPT_STATUS_CONTINUE                  0x60f8
+#       define LB_D2_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D2_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD2_INTERRUPT                        (1 << 17)
+#       define DC_HPD2_RX_INTERRUPT                     (1 << 18)
+#       define DISP_TIMER_INTERRUPT                     (1 << 24)
+#define DISP_INTERRUPT_STATUS_CONTINUE2                 0x60fc
+#       define LB_D3_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D3_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD3_INTERRUPT                        (1 << 17)
+#       define DC_HPD3_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE3                 0x6100
+#       define LB_D4_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D4_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD4_INTERRUPT                        (1 << 17)
+#       define DC_HPD4_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE4                 0x614c
+#       define LB_D5_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D5_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD5_INTERRUPT                        (1 << 17)
+#       define DC_HPD5_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE5                 0x6150
+#       define LB_D6_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D6_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD6_INTERRUPT                        (1 << 17)
+#       define DC_HPD6_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE6                 0x6780
+
+#define	DAC_AUTODETECT_INT_CONTROL			0x67c8
+
+#define DC_HPD1_INT_STATUS                              0x601c
+#define DC_HPD2_INT_STATUS                              0x6028
+#define DC_HPD3_INT_STATUS                              0x6034
+#define DC_HPD4_INT_STATUS                              0x6040
+#define DC_HPD5_INT_STATUS                              0x604c
+#define DC_HPD6_INT_STATUS                              0x6058
+#       define DC_HPDx_INT_STATUS                       (1 << 0)
+#       define DC_HPDx_SENSE                            (1 << 1)
+#       define DC_HPDx_SENSE_DELAYED                    (1 << 4)
+#       define DC_HPDx_RX_INT_STATUS                    (1 << 8)
+
+#define DC_HPD1_INT_CONTROL                             0x6020
+#define DC_HPD2_INT_CONTROL                             0x602c
+#define DC_HPD3_INT_CONTROL                             0x6038
+#define DC_HPD4_INT_CONTROL                             0x6044
+#define DC_HPD5_INT_CONTROL                             0x6050
+#define DC_HPD6_INT_CONTROL                             0x605c
+#       define DC_HPDx_INT_ACK                          (1 << 0)
+#       define DC_HPDx_INT_POLARITY                     (1 << 8)
+#       define DC_HPDx_INT_EN                           (1 << 16)
+#       define DC_HPDx_RX_INT_ACK                       (1 << 20)
+#       define DC_HPDx_RX_INT_EN                        (1 << 24)
+
+#define DC_HPD1_CONTROL                                   0x6024
+#define DC_HPD2_CONTROL                                   0x6030
+#define DC_HPD3_CONTROL                                   0x603c
+#define DC_HPD4_CONTROL                                   0x6048
+#define DC_HPD5_CONTROL                                   0x6054
+#define DC_HPD6_CONTROL                                   0x6060
+#       define DC_HPDx_CONNECTION_TIMER(x)                ((x) << 0)
+#       define DC_HPDx_RX_INT_TIMER(x)                    ((x) << 16)
+#       define DC_HPDx_EN                                 (1 << 28)
+
+#define	GRBM_CNTL					0x8000
+#define		GRBM_READ_TIMEOUT(x)				((x) << 0)
+
+#define	GRBM_STATUS2					0x8008
+#define		ME0PIPE1_CMDFIFO_AVAIL_MASK			0x0000000F
+#define		ME0PIPE1_CF_RQ_PENDING				(1 << 4)
+#define		ME0PIPE1_PF_RQ_PENDING				(1 << 5)
+#define		ME1PIPE0_RQ_PENDING				(1 << 6)
+#define		ME1PIPE1_RQ_PENDING				(1 << 7)
+#define		ME1PIPE2_RQ_PENDING				(1 << 8)
+#define		ME1PIPE3_RQ_PENDING				(1 << 9)
+#define		ME2PIPE0_RQ_PENDING				(1 << 10)
+#define		ME2PIPE1_RQ_PENDING				(1 << 11)
+#define		ME2PIPE2_RQ_PENDING				(1 << 12)
+#define		ME2PIPE3_RQ_PENDING				(1 << 13)
+#define		RLC_RQ_PENDING 					(1 << 14)
+#define		RLC_BUSY 					(1 << 24)
+#define		TC_BUSY 					(1 << 25)
+#define		CPF_BUSY 					(1 << 28)
+#define		CPC_BUSY 					(1 << 29)
+#define		CPG_BUSY 					(1 << 30)
+
+#define	GRBM_STATUS					0x8010
+#define		ME0PIPE0_CMDFIFO_AVAIL_MASK			0x0000000F
+#define		SRBM_RQ_PENDING					(1 << 5)
+#define		ME0PIPE0_CF_RQ_PENDING				(1 << 7)
+#define		ME0PIPE0_PF_RQ_PENDING				(1 << 8)
+#define		GDS_DMA_RQ_PENDING				(1 << 9)
+#define		DB_CLEAN					(1 << 12)
+#define		CB_CLEAN					(1 << 13)
+#define		TA_BUSY 					(1 << 14)
+#define		GDS_BUSY 					(1 << 15)
+#define		WD_BUSY_NO_DMA 					(1 << 16)
+#define		VGT_BUSY					(1 << 17)
+#define		IA_BUSY_NO_DMA					(1 << 18)
+#define		IA_BUSY						(1 << 19)
+#define		SX_BUSY 					(1 << 20)
+#define		WD_BUSY 					(1 << 21)
+#define		SPI_BUSY					(1 << 22)
+#define		BCI_BUSY					(1 << 23)
+#define		SC_BUSY 					(1 << 24)
+#define		PA_BUSY 					(1 << 25)
+#define		DB_BUSY 					(1 << 26)
+#define		CP_COHERENCY_BUSY      				(1 << 28)
+#define		CP_BUSY 					(1 << 29)
+#define		CB_BUSY 					(1 << 30)
+#define		GUI_ACTIVE					(1 << 31)
+#define	GRBM_STATUS_SE0					0x8014
+#define	GRBM_STATUS_SE1					0x8018
+#define	GRBM_STATUS_SE2					0x8038
+#define	GRBM_STATUS_SE3					0x803C
+#define		SE_DB_CLEAN					(1 << 1)
+#define		SE_CB_CLEAN					(1 << 2)
+#define		SE_BCI_BUSY					(1 << 22)
+#define		SE_VGT_BUSY					(1 << 23)
+#define		SE_PA_BUSY					(1 << 24)
+#define		SE_TA_BUSY					(1 << 25)
+#define		SE_SX_BUSY					(1 << 26)
+#define		SE_SPI_BUSY					(1 << 27)
+#define		SE_SC_BUSY					(1 << 29)
+#define		SE_DB_BUSY					(1 << 30)
+#define		SE_CB_BUSY					(1 << 31)
+
+#define	GRBM_SOFT_RESET					0x8020
+#define		SOFT_RESET_CP					(1 << 0)  /* All CP blocks */
+#define		SOFT_RESET_RLC					(1 << 2)  /* RLC */
+#define		SOFT_RESET_GFX					(1 << 16) /* GFX */
+#define		SOFT_RESET_CPF					(1 << 17) /* CP fetcher shared by gfx and compute */
+#define		SOFT_RESET_CPC					(1 << 18) /* CP Compute (MEC1/2) */
+#define		SOFT_RESET_CPG					(1 << 19) /* CP GFX (PFP, ME, CE) */
+
+#define GRBM_INT_CNTL                                   0x8060
+#       define RDERR_INT_ENABLE                         (1 << 0)
+#       define GUI_IDLE_INT_ENABLE                      (1 << 19)
+
+#define CP_CPC_STATUS					0x8210
+#define CP_CPC_BUSY_STAT				0x8214
+#define CP_CPC_STALLED_STAT1				0x8218
+#define CP_CPF_STATUS					0x821c
+#define CP_CPF_BUSY_STAT				0x8220
+#define CP_CPF_STALLED_STAT1				0x8224
+
+#define CP_MEC_CNTL					0x8234
+#define		MEC_ME2_HALT					(1 << 28)
+#define		MEC_ME1_HALT					(1 << 30)
+
+#define CP_MEC_CNTL					0x8234
+#define		MEC_ME2_HALT					(1 << 28)
+#define		MEC_ME1_HALT					(1 << 30)
+
+#define CP_STALLED_STAT3				0x8670
+#define CP_STALLED_STAT1				0x8674
+#define CP_STALLED_STAT2				0x8678
+
+#define CP_STAT						0x8680
+
+#define CP_ME_CNTL					0x86D8
+#define		CP_CE_HALT					(1 << 24)
+#define		CP_PFP_HALT					(1 << 26)
+#define		CP_ME_HALT					(1 << 28)
+
+#define	CP_RB0_RPTR					0x8700
+#define	CP_RB_WPTR_DELAY				0x8704
+
+#define CP_MEQ_THRESHOLDS				0x8764
+#define		MEQ1_START(x)				((x) << 0)
+#define		MEQ2_START(x)				((x) << 8)
+
+#define	VGT_VTX_VECT_EJECT_REG				0x88B0
+
+#define	VGT_CACHE_INVALIDATION				0x88C4
+#define		CACHE_INVALIDATION(x)				((x) << 0)
+#define			VC_ONLY						0
+#define			TC_ONLY						1
+#define			VC_AND_TC					2
+#define		AUTO_INVLD_EN(x)				((x) << 6)
+#define			NO_AUTO						0
+#define			ES_AUTO						1
+#define			GS_AUTO						2
+#define			ES_AND_GS_AUTO					3
+
+#define	VGT_GS_VERTEX_REUSE				0x88D4
+
+#define CC_GC_SHADER_ARRAY_CONFIG			0x89bc
+#define		INACTIVE_CUS_MASK			0xFFFF0000
+#define		INACTIVE_CUS_SHIFT			16
+#define GC_USER_SHADER_ARRAY_CONFIG			0x89c0
+
+#define	PA_CL_ENHANCE					0x8A14
+#define		CLIP_VTX_REORDER_ENA				(1 << 0)
+#define		NUM_CLIP_SEQ(x)					((x) << 1)
+
+#define	PA_SC_FORCE_EOV_MAX_CNTS			0x8B24
+#define		FORCE_EOV_MAX_CLK_CNT(x)			((x) << 0)
+#define		FORCE_EOV_MAX_REZ_CNT(x)			((x) << 16)
+
+#define	PA_SC_FIFO_SIZE					0x8BCC
+#define		SC_FRONTEND_PRIM_FIFO_SIZE(x)			((x) << 0)
+#define		SC_BACKEND_PRIM_FIFO_SIZE(x)			((x) << 6)
+#define		SC_HIZ_TILE_FIFO_SIZE(x)			((x) << 15)
+#define		SC_EARLYZ_TILE_FIFO_SIZE(x)			((x) << 23)
+
+#define	PA_SC_ENHANCE					0x8BF0
+#define		ENABLE_PA_SC_OUT_OF_ORDER			(1 << 0)
+#define		DISABLE_PA_SC_GUIDANCE				(1 << 13)
+
+#define	SQ_CONFIG					0x8C00
+
+#define	SH_MEM_BASES					0x8C28
+/* if PTR32, these are the bases for scratch and lds */
+#define		PRIVATE_BASE(x)					((x) << 0) /* scratch */
+#define		SHARED_BASE(x)					((x) << 16) /* LDS */
+#define	SH_MEM_APE1_BASE				0x8C2C
+/* if PTR32, this is the base location of GPUVM */
+#define	SH_MEM_APE1_LIMIT				0x8C30
+/* if PTR32, this is the upper limit of GPUVM */
+#define	SH_MEM_CONFIG					0x8C34
+#define		PTR32						(1 << 0)
+#define		ALIGNMENT_MODE(x)				((x) << 2)
+#define			SH_MEM_ALIGNMENT_MODE_DWORD			0
+#define			SH_MEM_ALIGNMENT_MODE_DWORD_STRICT		1
+#define			SH_MEM_ALIGNMENT_MODE_STRICT			2
+#define			SH_MEM_ALIGNMENT_MODE_UNALIGNED			3
+#define		DEFAULT_MTYPE(x)				((x) << 4)
+#define		APE1_MTYPE(x)					((x) << 7)
+
+#define	SX_DEBUG_1					0x9060
+
+#define	SPI_CONFIG_CNTL					0x9100
+
+#define	SPI_CONFIG_CNTL_1				0x913C
+#define		VTX_DONE_DELAY(x)				((x) << 0)
+#define		INTERP_ONE_PRIM_PER_ROW				(1 << 4)
+
+#define	TA_CNTL_AUX					0x9508
+
+#define DB_DEBUG					0x9830
+#define DB_DEBUG2					0x9834
+#define DB_DEBUG3					0x9838
+
+#define CC_RB_BACKEND_DISABLE				0x98F4
+#define		BACKEND_DISABLE(x)     			((x) << 16)
+#define GB_ADDR_CONFIG  				0x98F8
+#define		NUM_PIPES(x)				((x) << 0)
+#define		NUM_PIPES_MASK				0x00000007
+#define		NUM_PIPES_SHIFT				0
+#define		PIPE_INTERLEAVE_SIZE(x)			((x) << 4)
+#define		PIPE_INTERLEAVE_SIZE_MASK		0x00000070
+#define		PIPE_INTERLEAVE_SIZE_SHIFT		4
+#define		NUM_SHADER_ENGINES(x)			((x) << 12)
+#define		NUM_SHADER_ENGINES_MASK			0x00003000
+#define		NUM_SHADER_ENGINES_SHIFT		12
+#define		SHADER_ENGINE_TILE_SIZE(x)     		((x) << 16)
+#define		SHADER_ENGINE_TILE_SIZE_MASK		0x00070000
+#define		SHADER_ENGINE_TILE_SIZE_SHIFT		16
+#define		ROW_SIZE(x)             		((x) << 28)
+#define		ROW_SIZE_MASK				0x30000000
+#define		ROW_SIZE_SHIFT				28
+
+#define	GB_TILE_MODE0					0x9910
+#       define ARRAY_MODE(x)					((x) << 2)
+#              define	ARRAY_LINEAR_GENERAL			0
+#              define	ARRAY_LINEAR_ALIGNED			1
+#              define	ARRAY_1D_TILED_THIN1			2
+#              define	ARRAY_2D_TILED_THIN1			4
+#              define	ARRAY_PRT_TILED_THIN1			5
+#              define	ARRAY_PRT_2D_TILED_THIN1		6
+#       define PIPE_CONFIG(x)					((x) << 6)
+#              define	ADDR_SURF_P2				0
+#              define	ADDR_SURF_P4_8x16			4
+#              define	ADDR_SURF_P4_16x16			5
+#              define	ADDR_SURF_P4_16x32			6
+#              define	ADDR_SURF_P4_32x32			7
+#              define	ADDR_SURF_P8_16x16_8x16			8
+#              define	ADDR_SURF_P8_16x32_8x16			9
+#              define	ADDR_SURF_P8_32x32_8x16			10
+#              define	ADDR_SURF_P8_16x32_16x16		11
+#              define	ADDR_SURF_P8_32x32_16x16		12
+#              define	ADDR_SURF_P8_32x32_16x32		13
+#              define	ADDR_SURF_P8_32x64_32x32		14
+#       define TILE_SPLIT(x)					((x) << 11)
+#              define	ADDR_SURF_TILE_SPLIT_64B		0
+#              define	ADDR_SURF_TILE_SPLIT_128B		1
+#              define	ADDR_SURF_TILE_SPLIT_256B		2
+#              define	ADDR_SURF_TILE_SPLIT_512B		3
+#              define	ADDR_SURF_TILE_SPLIT_1KB		4
+#              define	ADDR_SURF_TILE_SPLIT_2KB		5
+#              define	ADDR_SURF_TILE_SPLIT_4KB		6
+#       define MICRO_TILE_MODE_NEW(x)				((x) << 22)
+#              define	ADDR_SURF_DISPLAY_MICRO_TILING		0
+#              define	ADDR_SURF_THIN_MICRO_TILING		1
+#              define	ADDR_SURF_DEPTH_MICRO_TILING		2
+#              define	ADDR_SURF_ROTATED_MICRO_TILING		3
+#       define SAMPLE_SPLIT(x)					((x) << 25)
+#              define	ADDR_SURF_SAMPLE_SPLIT_1		0
+#              define	ADDR_SURF_SAMPLE_SPLIT_2		1
+#              define	ADDR_SURF_SAMPLE_SPLIT_4		2
+#              define	ADDR_SURF_SAMPLE_SPLIT_8		3
+
+#define	GB_MACROTILE_MODE0					0x9990
+#       define BANK_WIDTH(x)					((x) << 0)
+#              define	ADDR_SURF_BANK_WIDTH_1			0
+#              define	ADDR_SURF_BANK_WIDTH_2			1
+#              define	ADDR_SURF_BANK_WIDTH_4			2
+#              define	ADDR_SURF_BANK_WIDTH_8			3
+#       define BANK_HEIGHT(x)					((x) << 2)
+#              define	ADDR_SURF_BANK_HEIGHT_1			0
+#              define	ADDR_SURF_BANK_HEIGHT_2			1
+#              define	ADDR_SURF_BANK_HEIGHT_4			2
+#              define	ADDR_SURF_BANK_HEIGHT_8			3
+#       define MACRO_TILE_ASPECT(x)				((x) << 4)
+#              define	ADDR_SURF_MACRO_ASPECT_1		0
+#              define	ADDR_SURF_MACRO_ASPECT_2		1
+#              define	ADDR_SURF_MACRO_ASPECT_4		2
+#              define	ADDR_SURF_MACRO_ASPECT_8		3
+#       define NUM_BANKS(x)					((x) << 6)
+#              define	ADDR_SURF_2_BANK			0
+#              define	ADDR_SURF_4_BANK			1
+#              define	ADDR_SURF_8_BANK			2
+#              define	ADDR_SURF_16_BANK			3
+
+#define	CB_HW_CONTROL					0x9A10
+
+#define	GC_USER_RB_BACKEND_DISABLE			0x9B7C
+#define		BACKEND_DISABLE_MASK			0x00FF0000
+#define		BACKEND_DISABLE_SHIFT			16
+
+#define	TCP_CHAN_STEER_LO				0xac0c
+#define	TCP_CHAN_STEER_HI				0xac10
+
+#define	TC_CFG_L1_LOAD_POLICY0				0xAC68
+#define	TC_CFG_L1_LOAD_POLICY1				0xAC6C
+#define	TC_CFG_L1_STORE_POLICY				0xAC70
+#define	TC_CFG_L2_LOAD_POLICY0				0xAC74
+#define	TC_CFG_L2_LOAD_POLICY1				0xAC78
+#define	TC_CFG_L2_STORE_POLICY0				0xAC7C
+#define	TC_CFG_L2_STORE_POLICY1				0xAC80
+#define	TC_CFG_L2_ATOMIC_POLICY				0xAC84
+#define	TC_CFG_L1_VOLATILE				0xAC88
+#define	TC_CFG_L2_VOLATILE				0xAC8C
+
+#define	CP_RB0_BASE					0xC100
+#define	CP_RB0_CNTL					0xC104
+#define		RB_BUFSZ(x)					((x) << 0)
+#define		RB_BLKSZ(x)					((x) << 8)
+#define		BUF_SWAP_32BIT					(2 << 16)
+#define		RB_NO_UPDATE					(1 << 27)
+#define		RB_RPTR_WR_ENA					(1 << 31)
+
+#define	CP_RB0_RPTR_ADDR				0xC10C
+#define		RB_RPTR_SWAP_32BIT				(2 << 0)
+#define	CP_RB0_RPTR_ADDR_HI				0xC110
+#define	CP_RB0_WPTR					0xC114
+
+#define	CP_DEVICE_ID					0xC12C
+#define	CP_ENDIAN_SWAP					0xC140
+#define	CP_RB_VMID					0xC144
+
+#define	CP_PFP_UCODE_ADDR				0xC150
+#define	CP_PFP_UCODE_DATA				0xC154
+#define	CP_ME_RAM_RADDR					0xC158
+#define	CP_ME_RAM_WADDR					0xC15C
+#define	CP_ME_RAM_DATA					0xC160
+
+#define	CP_CE_UCODE_ADDR				0xC168
+#define	CP_CE_UCODE_DATA				0xC16C
+#define	CP_MEC_ME1_UCODE_ADDR				0xC170
+#define	CP_MEC_ME1_UCODE_DATA				0xC174
+#define	CP_MEC_ME2_UCODE_ADDR				0xC178
+#define	CP_MEC_ME2_UCODE_DATA				0xC17C
+
+#define CP_INT_CNTL_RING0                               0xC1A8
+#       define CNTX_BUSY_INT_ENABLE                     (1 << 19)
+#       define CNTX_EMPTY_INT_ENABLE                    (1 << 20)
+#       define PRIV_INSTR_INT_ENABLE                    (1 << 22)
+#       define PRIV_REG_INT_ENABLE                      (1 << 23)
+#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
+#       define CP_RINGID2_INT_ENABLE                    (1 << 29)
+#       define CP_RINGID1_INT_ENABLE                    (1 << 30)
+#       define CP_RINGID0_INT_ENABLE                    (1 << 31)
+
+#define CP_INT_STATUS_RING0                             0xC1B4
+#       define PRIV_INSTR_INT_STAT                      (1 << 22)
+#       define PRIV_REG_INT_STAT                        (1 << 23)
+#       define TIME_STAMP_INT_STAT                      (1 << 26)
+#       define CP_RINGID2_INT_STAT                      (1 << 29)
+#       define CP_RINGID1_INT_STAT                      (1 << 30)
+#       define CP_RINGID0_INT_STAT                      (1 << 31)
+
+#define CP_CPF_DEBUG                                    0xC200
+
+#define CP_PQ_WPTR_POLL_CNTL                            0xC20C
+#define		WPTR_POLL_EN      			(1 << 31)
+
+#define CP_ME1_PIPE0_INT_CNTL                           0xC214
+#define CP_ME1_PIPE1_INT_CNTL                           0xC218
+#define CP_ME1_PIPE2_INT_CNTL                           0xC21C
+#define CP_ME1_PIPE3_INT_CNTL                           0xC220
+#define CP_ME2_PIPE0_INT_CNTL                           0xC224
+#define CP_ME2_PIPE1_INT_CNTL                           0xC228
+#define CP_ME2_PIPE2_INT_CNTL                           0xC22C
+#define CP_ME2_PIPE3_INT_CNTL                           0xC230
+#       define DEQUEUE_REQUEST_INT_ENABLE               (1 << 13)
+#       define WRM_POLL_TIMEOUT_INT_ENABLE              (1 << 17)
+#       define PRIV_REG_INT_ENABLE                      (1 << 23)
+#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
+#       define GENERIC2_INT_ENABLE                      (1 << 29)
+#       define GENERIC1_INT_ENABLE                      (1 << 30)
+#       define GENERIC0_INT_ENABLE                      (1 << 31)
+#define CP_ME1_PIPE0_INT_STATUS                         0xC214
+#define CP_ME1_PIPE1_INT_STATUS                         0xC218
+#define CP_ME1_PIPE2_INT_STATUS                         0xC21C
+#define CP_ME1_PIPE3_INT_STATUS                         0xC220
+#define CP_ME2_PIPE0_INT_STATUS                         0xC224
+#define CP_ME2_PIPE1_INT_STATUS                         0xC228
+#define CP_ME2_PIPE2_INT_STATUS                         0xC22C
+#define CP_ME2_PIPE3_INT_STATUS                         0xC230
+#       define DEQUEUE_REQUEST_INT_STATUS               (1 << 13)
+#       define WRM_POLL_TIMEOUT_INT_STATUS              (1 << 17)
+#       define PRIV_REG_INT_STATUS                      (1 << 23)
+#       define TIME_STAMP_INT_STATUS                    (1 << 26)
+#       define GENERIC2_INT_STATUS                      (1 << 29)
+#       define GENERIC1_INT_STATUS                      (1 << 30)
+#       define GENERIC0_INT_STATUS                      (1 << 31)
+
+#define	CP_MAX_CONTEXT					0xC2B8
+
+#define	CP_RB0_BASE_HI					0xC2C4
+
+#define RLC_CNTL                                          0xC300
+#       define RLC_ENABLE                                 (1 << 0)
+
+#define RLC_MC_CNTL                                       0xC30C
+
+#define RLC_LB_CNTR_MAX                                   0xC348
+
+#define RLC_LB_CNTL                                       0xC364
+
+#define RLC_LB_CNTR_INIT                                  0xC36C
+
+#define RLC_SAVE_AND_RESTORE_BASE                         0xC374
+#define RLC_DRIVER_DMA_STATUS                             0xC378
+
+#define RLC_GPM_UCODE_ADDR                                0xC388
+#define RLC_GPM_UCODE_DATA                                0xC38C
+#define RLC_GPU_CLOCK_COUNT_LSB                           0xC390
+#define RLC_GPU_CLOCK_COUNT_MSB                           0xC394
+#define RLC_CAPTURE_GPU_CLOCK_COUNT                       0xC398
+#define RLC_UCODE_CNTL                                    0xC39C
+
+#define RLC_CGCG_CGLS_CTRL                                0xC424
+
+#define RLC_LB_INIT_CU_MASK                               0xC43C
+
+#define RLC_LB_PARAMS                                     0xC444
+
+#define RLC_SERDES_CU_MASTER_BUSY                         0xC484
+#define RLC_SERDES_NONCU_MASTER_BUSY                      0xC488
+#       define SE_MASTER_BUSY_MASK                        0x0000ffff
+#       define GC_MASTER_BUSY                             (1 << 16)
+#       define TC0_MASTER_BUSY                            (1 << 17)
+#       define TC1_MASTER_BUSY                            (1 << 18)
+
+#define RLC_GPM_SCRATCH_ADDR                              0xC4B0
+#define RLC_GPM_SCRATCH_DATA                              0xC4B4
+
+#define CP_HPD_EOP_BASE_ADDR                              0xC904
+#define CP_HPD_EOP_BASE_ADDR_HI                           0xC908
+#define CP_HPD_EOP_VMID                                   0xC90C
+#define CP_HPD_EOP_CONTROL                                0xC910
+#define		EOP_SIZE(x)				((x) << 0)
+#define		EOP_SIZE_MASK      			(0x3f << 0)
+#define CP_MQD_BASE_ADDR                                  0xC914
+#define CP_MQD_BASE_ADDR_HI                               0xC918
+#define CP_HQD_ACTIVE                                     0xC91C
+#define CP_HQD_VMID                                       0xC920
+
+#define CP_HQD_PQ_BASE                                    0xC934
+#define CP_HQD_PQ_BASE_HI                                 0xC938
+#define CP_HQD_PQ_RPTR                                    0xC93C
+#define CP_HQD_PQ_RPTR_REPORT_ADDR                        0xC940
+#define CP_HQD_PQ_RPTR_REPORT_ADDR_HI                     0xC944
+#define CP_HQD_PQ_WPTR_POLL_ADDR                          0xC948
+#define CP_HQD_PQ_WPTR_POLL_ADDR_HI                       0xC94C
+#define CP_HQD_PQ_DOORBELL_CONTROL                        0xC950
+#define		DOORBELL_OFFSET(x)			((x) << 2)
+#define		DOORBELL_OFFSET_MASK			(0x1fffff << 2)
+#define		DOORBELL_SOURCE      			(1 << 28)
+#define		DOORBELL_SCHD_HIT      			(1 << 29)
+#define		DOORBELL_EN      			(1 << 30)
+#define		DOORBELL_HIT      			(1 << 31)
+#define CP_HQD_PQ_WPTR                                    0xC954
+#define CP_HQD_PQ_CONTROL                                 0xC958
+#define		QUEUE_SIZE(x)				((x) << 0)
+#define		QUEUE_SIZE_MASK      			(0x3f << 0)
+#define		RPTR_BLOCK_SIZE(x)			((x) << 8)
+#define		RPTR_BLOCK_SIZE_MASK			(0x3f << 8)
+#define		PQ_VOLATILE      			(1 << 26)
+#define		NO_UPDATE_RPTR      			(1 << 27)
+#define		UNORD_DISPATCH      			(1 << 28)
+#define		ROQ_PQ_IB_FLIP      			(1 << 29)
+#define		PRIV_STATE      			(1 << 30)
+#define		KMD_QUEUE      				(1 << 31)
+
+#define CP_HQD_DEQUEUE_REQUEST                          0xC974
+
+#define CP_MQD_CONTROL                                  0xC99C
+#define		MQD_VMID(x)				((x) << 0)
+#define		MQD_VMID_MASK      			(0xf << 0)
+
+#define PA_SC_RASTER_CONFIG                             0x28350
+#       define RASTER_CONFIG_RB_MAP_0                   0
+#       define RASTER_CONFIG_RB_MAP_1                   1
+#       define RASTER_CONFIG_RB_MAP_2                   2
+#       define RASTER_CONFIG_RB_MAP_3                   3
+
+#define VGT_EVENT_INITIATOR                             0x28a90
+#       define SAMPLE_STREAMOUTSTATS1                   (1 << 0)
+#       define SAMPLE_STREAMOUTSTATS2                   (2 << 0)
+#       define SAMPLE_STREAMOUTSTATS3                   (3 << 0)
+#       define CACHE_FLUSH_TS                           (4 << 0)
+#       define CACHE_FLUSH                              (6 << 0)
+#       define CS_PARTIAL_FLUSH                         (7 << 0)
+#       define VGT_STREAMOUT_RESET                      (10 << 0)
+#       define END_OF_PIPE_INCR_DE                      (11 << 0)
+#       define END_OF_PIPE_IB_END                       (12 << 0)
+#       define RST_PIX_CNT                              (13 << 0)
+#       define VS_PARTIAL_FLUSH                         (15 << 0)
+#       define PS_PARTIAL_FLUSH                         (16 << 0)
+#       define CACHE_FLUSH_AND_INV_TS_EVENT             (20 << 0)
+#       define ZPASS_DONE                               (21 << 0)
+#       define CACHE_FLUSH_AND_INV_EVENT                (22 << 0)
+#       define PERFCOUNTER_START                        (23 << 0)
+#       define PERFCOUNTER_STOP                         (24 << 0)
+#       define PIPELINESTAT_START                       (25 << 0)
+#       define PIPELINESTAT_STOP                        (26 << 0)
+#       define PERFCOUNTER_SAMPLE                       (27 << 0)
+#       define SAMPLE_PIPELINESTAT                      (30 << 0)
+#       define SO_VGT_STREAMOUT_FLUSH                   (31 << 0)
+#       define SAMPLE_STREAMOUTSTATS                    (32 << 0)
+#       define RESET_VTX_CNT                            (33 << 0)
+#       define VGT_FLUSH                                (36 << 0)
+#       define BOTTOM_OF_PIPE_TS                        (40 << 0)
+#       define DB_CACHE_FLUSH_AND_INV                   (42 << 0)
+#       define FLUSH_AND_INV_DB_DATA_TS                 (43 << 0)
+#       define FLUSH_AND_INV_DB_META                    (44 << 0)
+#       define FLUSH_AND_INV_CB_DATA_TS                 (45 << 0)
+#       define FLUSH_AND_INV_CB_META                    (46 << 0)
+#       define CS_DONE                                  (47 << 0)
+#       define PS_DONE                                  (48 << 0)
+#       define FLUSH_AND_INV_CB_PIXEL_DATA              (49 << 0)
+#       define THREAD_TRACE_START                       (51 << 0)
+#       define THREAD_TRACE_STOP                        (52 << 0)
+#       define THREAD_TRACE_FLUSH                       (54 << 0)
+#       define THREAD_TRACE_FINISH                      (55 << 0)
+#       define PIXEL_PIPE_STAT_CONTROL                  (56 << 0)
+#       define PIXEL_PIPE_STAT_DUMP                     (57 << 0)
+#       define PIXEL_PIPE_STAT_RESET                    (58 << 0)
+
+#define	SCRATCH_REG0					0x30100
+#define	SCRATCH_REG1					0x30104
+#define	SCRATCH_REG2					0x30108
+#define	SCRATCH_REG3					0x3010C
+#define	SCRATCH_REG4					0x30110
+#define	SCRATCH_REG5					0x30114
+#define	SCRATCH_REG6					0x30118
+#define	SCRATCH_REG7					0x3011C
+
+#define	SCRATCH_UMSK					0x30140
+#define	SCRATCH_ADDR					0x30144
+
+#define	CP_SEM_WAIT_TIMER				0x301BC
+
+#define	CP_SEM_INCOMPLETE_TIMER_CNTL			0x301C8
+
+#define	CP_WAIT_REG_MEM_TIMEOUT				0x301D0
+
+#define GRBM_GFX_INDEX          			0x30800
+#define		INSTANCE_INDEX(x)			((x) << 0)
+#define		SH_INDEX(x)     			((x) << 8)
+#define		SE_INDEX(x)     			((x) << 16)
+#define		SH_BROADCAST_WRITES      		(1 << 29)
+#define		INSTANCE_BROADCAST_WRITES      		(1 << 30)
+#define		SE_BROADCAST_WRITES      		(1 << 31)
+
+#define	VGT_ESGS_RING_SIZE				0x30900
+#define	VGT_GSVS_RING_SIZE				0x30904
+#define	VGT_PRIMITIVE_TYPE				0x30908
+#define	VGT_INDEX_TYPE					0x3090C
+
+#define	VGT_NUM_INDICES					0x30930
+#define	VGT_NUM_INSTANCES				0x30934
+#define	VGT_TF_RING_SIZE				0x30938
+#define	VGT_HS_OFFCHIP_PARAM				0x3093C
+#define	VGT_TF_MEMORY_BASE				0x30940
+
+#define	PA_SU_LINE_STIPPLE_VALUE			0x30a00
+#define	PA_SC_LINE_STIPPLE_STATE			0x30a04
+
+#define	SQC_CACHES					0x30d20
+
+#define	CP_PERFMON_CNTL					0x36020
+
+#define	CGTS_TCC_DISABLE				0x3c00c
+#define	CGTS_USER_TCC_DISABLE				0x3c010
+#define		TCC_DISABLE_MASK				0xFFFF0000
+#define		TCC_DISABLE_SHIFT				16
+
+#define	CB_CGTT_SCLK_CTRL				0x3c2a0
+
+/*
+ * PM4
+ */
+#define	PACKET_TYPE0	0
+#define	PACKET_TYPE1	1
+#define	PACKET_TYPE2	2
+#define	PACKET_TYPE3	3
+
+#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
+#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
+#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
+#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
+#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+			 (((reg) >> 2) & 0xFFFF) |			\
+			 ((n) & 0x3FFF) << 16)
+#define CP_PACKET2			0x80000000
+#define		PACKET2_PAD_SHIFT		0
+#define		PACKET2_PAD_MASK		(0x3fffffff << 0)
+
+#define PACKET2(v)	(CP_PACKET2 | REG_SET(PACKET2_PAD, (v)))
+
+#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+			 (((op) & 0xFF) << 8) |				\
+			 ((n) & 0x3FFF) << 16)
+
+#define PACKET3_COMPUTE(op, n) (PACKET3(op, n) | 1 << 1)
+
+/* Packet 3 types */
+#define	PACKET3_NOP					0x10
+#define	PACKET3_SET_BASE				0x11
+#define		PACKET3_BASE_INDEX(x)                  ((x) << 0)
+#define			CE_PARTITION_BASE		3
+#define	PACKET3_CLEAR_STATE				0x12
+#define	PACKET3_INDEX_BUFFER_SIZE			0x13
+#define	PACKET3_DISPATCH_DIRECT				0x15
+#define	PACKET3_DISPATCH_INDIRECT			0x16
+#define	PACKET3_ATOMIC_GDS				0x1D
+#define	PACKET3_ATOMIC_MEM				0x1E
+#define	PACKET3_OCCLUSION_QUERY				0x1F
+#define	PACKET3_SET_PREDICATION				0x20
+#define	PACKET3_REG_RMW					0x21
+#define	PACKET3_COND_EXEC				0x22
+#define	PACKET3_PRED_EXEC				0x23
+#define	PACKET3_DRAW_INDIRECT				0x24
+#define	PACKET3_DRAW_INDEX_INDIRECT			0x25
+#define	PACKET3_INDEX_BASE				0x26
+#define	PACKET3_DRAW_INDEX_2				0x27
+#define	PACKET3_CONTEXT_CONTROL				0x28
+#define	PACKET3_INDEX_TYPE				0x2A
+#define	PACKET3_DRAW_INDIRECT_MULTI			0x2C
+#define	PACKET3_DRAW_INDEX_AUTO				0x2D
+#define	PACKET3_NUM_INSTANCES				0x2F
+#define	PACKET3_DRAW_INDEX_MULTI_AUTO			0x30
+#define	PACKET3_INDIRECT_BUFFER_CONST			0x33
+#define	PACKET3_STRMOUT_BUFFER_UPDATE			0x34
+#define	PACKET3_DRAW_INDEX_OFFSET_2			0x35
+#define	PACKET3_DRAW_PREAMBLE				0x36
+#define	PACKET3_WRITE_DATA				0x37
+#define		WRITE_DATA_DST_SEL(x)                   ((x) << 8)
+                /* 0 - register
+		 * 1 - memory (sync - via GRBM)
+		 * 2 - gl2
+		 * 3 - gds
+		 * 4 - reserved
+		 * 5 - memory (async - direct)
+		 */
+#define		WR_ONE_ADDR                             (1 << 16)
+#define		WR_CONFIRM                              (1 << 20)
+#define		WRITE_DATA_CACHE_POLICY(x)              ((x) << 25)
+                /* 0 - LRU
+		 * 1 - Stream
+		 */
+#define		WRITE_DATA_ENGINE_SEL(x)                ((x) << 30)
+                /* 0 - me
+		 * 1 - pfp
+		 * 2 - ce
+		 */
+#define	PACKET3_DRAW_INDEX_INDIRECT_MULTI		0x38
+#define	PACKET3_MEM_SEMAPHORE				0x39
+#              define PACKET3_SEM_USE_MAILBOX       (0x1 << 16)
+#              define PACKET3_SEM_SEL_SIGNAL_TYPE   (0x1 << 20) /* 0 = increment, 1 = write 1 */
+#              define PACKET3_SEM_CLIENT_CODE	    ((x) << 24) /* 0 = CP, 1 = CB, 2 = DB */
+#              define PACKET3_SEM_SEL_SIGNAL	    (0x6 << 29)
+#              define PACKET3_SEM_SEL_WAIT	    (0x7 << 29)
+#define	PACKET3_COPY_DW					0x3B
+#define	PACKET3_WAIT_REG_MEM				0x3C
+#define		WAIT_REG_MEM_FUNCTION(x)                ((x) << 0)
+                /* 0 - always
+		 * 1 - <
+		 * 2 - <=
+		 * 3 - ==
+		 * 4 - !=
+		 * 5 - >=
+		 * 6 - >
+		 */
+#define		WAIT_REG_MEM_MEM_SPACE(x)               ((x) << 4)
+                /* 0 - reg
+		 * 1 - mem
+		 */
+#define		WAIT_REG_MEM_OPERATION(x)               ((x) << 6)
+                /* 0 - wait_reg_mem
+		 * 1 - wr_wait_wr_reg
+		 */
+#define		WAIT_REG_MEM_ENGINE(x)                  ((x) << 8)
+                /* 0 - me
+		 * 1 - pfp
+		 */
+#define	PACKET3_INDIRECT_BUFFER				0x3F
+#define		INDIRECT_BUFFER_TCL2_VOLATILE           (1 << 22)
+#define		INDIRECT_BUFFER_VALID                   (1 << 23)
+#define		INDIRECT_BUFFER_CACHE_POLICY(x)         ((x) << 28)
+                /* 0 - LRU
+		 * 1 - Stream
+		 * 2 - Bypass
+		 */
+#define	PACKET3_COPY_DATA				0x40
+#define	PACKET3_PFP_SYNC_ME				0x42
+#define	PACKET3_SURFACE_SYNC				0x43
+#              define PACKET3_DEST_BASE_0_ENA      (1 << 0)
+#              define PACKET3_DEST_BASE_1_ENA      (1 << 1)
+#              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
+#              define PACKET3_CB1_DEST_BASE_ENA    (1 << 7)
+#              define PACKET3_CB2_DEST_BASE_ENA    (1 << 8)
+#              define PACKET3_CB3_DEST_BASE_ENA    (1 << 9)
+#              define PACKET3_CB4_DEST_BASE_ENA    (1 << 10)
+#              define PACKET3_CB5_DEST_BASE_ENA    (1 << 11)
+#              define PACKET3_CB6_DEST_BASE_ENA    (1 << 12)
+#              define PACKET3_CB7_DEST_BASE_ENA    (1 << 13)
+#              define PACKET3_DB_DEST_BASE_ENA     (1 << 14)
+#              define PACKET3_TCL1_VOL_ACTION_ENA  (1 << 15)
+#              define PACKET3_TC_VOL_ACTION_ENA    (1 << 16) /* L2 */
+#              define PACKET3_TC_WB_ACTION_ENA     (1 << 18) /* L2 */
+#              define PACKET3_DEST_BASE_2_ENA      (1 << 19)
+#              define PACKET3_DEST_BASE_3_ENA      (1 << 21)
+#              define PACKET3_TCL1_ACTION_ENA      (1 << 22)
+#              define PACKET3_TC_ACTION_ENA        (1 << 23) /* L2 */
+#              define PACKET3_CB_ACTION_ENA        (1 << 25)
+#              define PACKET3_DB_ACTION_ENA        (1 << 26)
+#              define PACKET3_SH_KCACHE_ACTION_ENA (1 << 27)
+#              define PACKET3_SH_KCACHE_VOL_ACTION_ENA (1 << 28)
+#              define PACKET3_SH_ICACHE_ACTION_ENA (1 << 29)
+#define	PACKET3_COND_WRITE				0x45
+#define	PACKET3_EVENT_WRITE				0x46
+#define		EVENT_TYPE(x)                           ((x) << 0)
+#define		EVENT_INDEX(x)                          ((x) << 8)
+                /* 0 - any non-TS event
+		 * 1 - ZPASS_DONE, PIXEL_PIPE_STAT_*
+		 * 2 - SAMPLE_PIPELINESTAT
+		 * 3 - SAMPLE_STREAMOUTSTAT*
+		 * 4 - *S_PARTIAL_FLUSH
+		 * 5 - EOP events
+		 * 6 - EOS events
+		 */
+#define	PACKET3_EVENT_WRITE_EOP				0x47
+#define		EOP_TCL1_VOL_ACTION_EN                  (1 << 12)
+#define		EOP_TC_VOL_ACTION_EN                    (1 << 13) /* L2 */
+#define		EOP_TC_WB_ACTION_EN                     (1 << 15) /* L2 */
+#define		EOP_TCL1_ACTION_EN                      (1 << 16)
+#define		EOP_TC_ACTION_EN                        (1 << 17) /* L2 */
+#define		EOP_CACHE_POLICY(x)                     ((x) << 25)
+                /* 0 - LRU
+		 * 1 - Stream
+		 * 2 - Bypass
+		 */
+#define		EOP_TCL2_VOLATILE                       (1 << 27)
+#define		DATA_SEL(x)                             ((x) << 29)
+                /* 0 - discard
+		 * 1 - send low 32bit data
+		 * 2 - send 64bit data
+		 * 3 - send 64bit GPU counter value
+		 * 4 - send 64bit sys counter value
+		 */
+#define		INT_SEL(x)                              ((x) << 24)
+                /* 0 - none
+		 * 1 - interrupt only (DATA_SEL = 0)
+		 * 2 - interrupt when data write is confirmed
+		 */
+#define		DST_SEL(x)                              ((x) << 16)
+                /* 0 - MC
+		 * 1 - TC/L2
+		 */
+#define	PACKET3_EVENT_WRITE_EOS				0x48
+#define	PACKET3_RELEASE_MEM				0x49
+#define	PACKET3_PREAMBLE_CNTL				0x4A
+#              define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE     (2 << 28)
+#              define PACKET3_PREAMBLE_END_CLEAR_STATE       (3 << 28)
+#define	PACKET3_DMA_DATA				0x50
+#define	PACKET3_AQUIRE_MEM				0x58
+#define	PACKET3_REWIND					0x59
+#define	PACKET3_LOAD_UCONFIG_REG			0x5E
+#define	PACKET3_LOAD_SH_REG				0x5F
+#define	PACKET3_LOAD_CONFIG_REG				0x60
+#define	PACKET3_LOAD_CONTEXT_REG			0x61
+#define	PACKET3_SET_CONFIG_REG				0x68
+#define		PACKET3_SET_CONFIG_REG_START			0x00008000
+#define		PACKET3_SET_CONFIG_REG_END			0x0000b000
+#define	PACKET3_SET_CONTEXT_REG				0x69
+#define		PACKET3_SET_CONTEXT_REG_START			0x00028000
+#define		PACKET3_SET_CONTEXT_REG_END			0x00029000
+#define	PACKET3_SET_CONTEXT_REG_INDIRECT		0x73
+#define	PACKET3_SET_SH_REG				0x76
+#define		PACKET3_SET_SH_REG_START			0x0000b000
+#define		PACKET3_SET_SH_REG_END				0x0000c000
+#define	PACKET3_SET_SH_REG_OFFSET			0x77
+#define	PACKET3_SET_QUEUE_REG				0x78
+#define	PACKET3_SET_UCONFIG_REG				0x79
+#define		PACKET3_SET_UCONFIG_REG_START			0x00030000
+#define		PACKET3_SET_UCONFIG_REG_END			0x00031000
+#define	PACKET3_SCRATCH_RAM_WRITE			0x7D
+#define	PACKET3_SCRATCH_RAM_READ			0x7E
+#define	PACKET3_LOAD_CONST_RAM				0x80
+#define	PACKET3_WRITE_CONST_RAM				0x81
+#define	PACKET3_DUMP_CONST_RAM				0x83
+#define	PACKET3_INCREMENT_CE_COUNTER			0x84
+#define	PACKET3_INCREMENT_DE_COUNTER			0x85
+#define	PACKET3_WAIT_ON_CE_COUNTER			0x86
+#define	PACKET3_WAIT_ON_DE_COUNTER_DIFF			0x88
+#define	PACKET3_SWITCH_BUFFER				0x8B
+
+/* SDMA - first instance at 0xd000, second at 0xd800 */
+#define SDMA0_REGISTER_OFFSET                             0x0 /* not a register */
+#define SDMA1_REGISTER_OFFSET                             0x800 /* not a register */
+
+#define	SDMA0_UCODE_ADDR                                  0xD000
+#define	SDMA0_UCODE_DATA                                  0xD004
+
+#define SDMA0_CNTL                                        0xD010
+#       define TRAP_ENABLE                                (1 << 0)
+#       define SEM_INCOMPLETE_INT_ENABLE                  (1 << 1)
+#       define SEM_WAIT_INT_ENABLE                        (1 << 2)
+#       define DATA_SWAP_ENABLE                           (1 << 3)
+#       define FENCE_SWAP_ENABLE                          (1 << 4)
+#       define AUTO_CTXSW_ENABLE                          (1 << 18)
+#       define CTXEMPTY_INT_ENABLE                        (1 << 28)
+
+#define SDMA0_TILING_CONFIG  				  0xD018
+
+#define SDMA0_SEM_INCOMPLETE_TIMER_CNTL                   0xD020
+#define SDMA0_SEM_WAIT_FAIL_TIMER_CNTL                    0xD024
+
+#define SDMA0_STATUS_REG                                  0xd034
+#       define SDMA_IDLE                                  (1 << 0)
+
+#define SDMA0_ME_CNTL                                     0xD048
+#       define SDMA_HALT                                  (1 << 0)
+
+#define SDMA0_GFX_RB_CNTL                                 0xD200
+#       define SDMA_RB_ENABLE                             (1 << 0)
+#       define SDMA_RB_SIZE(x)                            ((x) << 1) /* log2 */
+#       define SDMA_RB_SWAP_ENABLE                        (1 << 9) /* 8IN32 */
+#       define SDMA_RPTR_WRITEBACK_ENABLE                 (1 << 12)
+#       define SDMA_RPTR_WRITEBACK_SWAP_ENABLE            (1 << 13)  /* 8IN32 */
+#       define SDMA_RPTR_WRITEBACK_TIMER(x)               ((x) << 16) /* log2 */
+#define SDMA0_GFX_RB_BASE                                 0xD204
+#define SDMA0_GFX_RB_BASE_HI                              0xD208
+#define SDMA0_GFX_RB_RPTR                                 0xD20C
+#define SDMA0_GFX_RB_WPTR                                 0xD210
+
+#define SDMA0_GFX_RB_RPTR_ADDR_HI                         0xD220
+#define SDMA0_GFX_RB_RPTR_ADDR_LO                         0xD224
+#define SDMA0_GFX_IB_CNTL                                 0xD228
+#       define SDMA_IB_ENABLE                             (1 << 0)
+#       define SDMA_IB_SWAP_ENABLE                        (1 << 4)
+#       define SDMA_SWITCH_INSIDE_IB                      (1 << 8)
+#       define SDMA_CMD_VMID(x)                           ((x) << 16)
+
+#define SDMA0_GFX_VIRTUAL_ADDR                            0xD29C
+#define SDMA0_GFX_APE1_CNTL                               0xD2A0
+
+#define SDMA_PACKET(op, sub_op, e)	((((e) & 0xFFFF) << 16) |	\
+					 (((sub_op) & 0xFF) << 8) |	\
+					 (((op) & 0xFF) << 0))
+/* sDMA opcodes */
+#define	SDMA_OPCODE_NOP					  0
+#define	SDMA_OPCODE_COPY				  1
+#       define SDMA_COPY_SUB_OPCODE_LINEAR                0
+#       define SDMA_COPY_SUB_OPCODE_TILED                 1
+#       define SDMA_COPY_SUB_OPCODE_SOA                   3
+#       define SDMA_COPY_SUB_OPCODE_LINEAR_SUB_WINDOW     4
+#       define SDMA_COPY_SUB_OPCODE_TILED_SUB_WINDOW      5
+#       define SDMA_COPY_SUB_OPCODE_T2T_SUB_WINDOW        6
+#define	SDMA_OPCODE_WRITE				  2
+#       define SDMA_WRITE_SUB_OPCODE_LINEAR               0
+#       define SDMA_WRTIE_SUB_OPCODE_TILED                1
+#define	SDMA_OPCODE_INDIRECT_BUFFER			  4
+#define	SDMA_OPCODE_FENCE				  5
+#define	SDMA_OPCODE_TRAP				  6
+#define	SDMA_OPCODE_SEMAPHORE				  7
+#       define SDMA_SEMAPHORE_EXTRA_O                     (1 << 13)
+                /* 0 - increment
+		 * 1 - write 1
+		 */
+#       define SDMA_SEMAPHORE_EXTRA_S                     (1 << 14)
+                /* 0 - wait
+		 * 1 - signal
+		 */
+#       define SDMA_SEMAPHORE_EXTRA_M                     (1 << 15)
+                /* mailbox */
+#define	SDMA_OPCODE_POLL_REG_MEM			  8
+#       define SDMA_POLL_REG_MEM_EXTRA_OP(x)              ((x) << 10)
+                /* 0 - wait_reg_mem
+		 * 1 - wr_wait_wr_reg
+		 */
+#       define SDMA_POLL_REG_MEM_EXTRA_FUNC(x)            ((x) << 12)
+                /* 0 - always
+		 * 1 - <
+		 * 2 - <=
+		 * 3 - ==
+		 * 4 - !=
+		 * 5 - >=
+		 * 6 - >
+		 */
+#       define SDMA_POLL_REG_MEM_EXTRA_M                  (1 << 15)
+                /* 0 = register
+		 * 1 = memory
+		 */
+#define	SDMA_OPCODE_COND_EXEC				  9
+#define	SDMA_OPCODE_CONSTANT_FILL			  11
+#       define SDMA_CONSTANT_FILL_EXTRA_SIZE(x)           ((x) << 14)
+                /* 0 = byte fill
+		 * 2 = DW fill
+		 */
+#define	SDMA_OPCODE_GENERATE_PTE_PDE			  12
+#define	SDMA_OPCODE_TIMESTAMP				  13
+#       define SDMA_TIMESTAMP_SUB_OPCODE_SET_LOCAL        0
+#       define SDMA_TIMESTAMP_SUB_OPCODE_GET_LOCAL        1
+#       define SDMA_TIMESTAMP_SUB_OPCODE_GET_GLOBAL       2
+#define	SDMA_OPCODE_SRBM_WRITE				  14
+#       define SDMA_SRBM_WRITE_EXTRA_BYTE_ENABLE(x)       ((x) << 12)
+                /* byte mask */
+
+/* UVD */
+
+#define UVD_UDEC_ADDR_CONFIG		0xef4c
+#define UVD_UDEC_DB_ADDR_CONFIG		0xef50
+#define UVD_UDEC_DBW_ADDR_CONFIG	0xef54
+
+#define UVD_LMI_EXT40_ADDR		0xf498
+#define UVD_LMI_ADDR_EXT		0xf594
+#define UVD_VCPU_CACHE_OFFSET0		0xf608
+#define UVD_VCPU_CACHE_SIZE0		0xf60c
+#define UVD_VCPU_CACHE_OFFSET1		0xf610
+#define UVD_VCPU_CACHE_SIZE1		0xf614
+#define UVD_VCPU_CACHE_OFFSET2		0xf618
+#define UVD_VCPU_CACHE_SIZE2		0xf61c
+
+#define UVD_RBC_RB_RPTR			0xf690
+#define UVD_RBC_RB_WPTR			0xf694
+
+/* UVD clocks */
+
+#define CG_DCLK_CNTL			0xC050009C
+#	define DCLK_DIVIDER_MASK	0x7f
+#	define DCLK_DIR_CNTL_EN		(1 << 8)
+#define CG_DCLK_STATUS			0xC05000A0
+#	define DCLK_STATUS		(1 << 0)
+#define CG_VCLK_CNTL			0xC05000A4
+#define CG_VCLK_STATUS			0xC05000A8
+
+#endif
diff --git a/drivers/gpu/drm/radeon/clearstate_cayman.h b/drivers/gpu/drm/radeon/clearstate_cayman.h
new file mode 100644
index 0000000..c003394
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_cayman.h
@@ -0,0 +1,1081 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+static const u32 SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0x00000000, // DB_DEPTH_INFO
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_DEPTH_SIZE
+    0x00000000, // DB_DEPTH_SLICE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_15
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+    0x00000000, // SX_MISC
+    0x00000000, // SX_SURFACE_SYNC
+    0x00000000, // SX_SCATTER_EXPORT_BASE
+    0x00000000, // SX_SCATTER_EXPORT_SIZE
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0x00000000, // CP_RINGID
+    0x00000000, // CP_VMID
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_VTX_SEMANTIC_0
+    0x00000000, // SQ_VTX_SEMANTIC_1
+    0x00000000, // SQ_VTX_SEMANTIC_2
+    0x00000000, // SQ_VTX_SEMANTIC_3
+    0x00000000, // SQ_VTX_SEMANTIC_4
+    0x00000000, // SQ_VTX_SEMANTIC_5
+    0x00000000, // SQ_VTX_SEMANTIC_6
+    0x00000000, // SQ_VTX_SEMANTIC_7
+    0x00000000, // SQ_VTX_SEMANTIC_8
+    0x00000000, // SQ_VTX_SEMANTIC_9
+    0x00000000, // SQ_VTX_SEMANTIC_10
+    0x00000000, // SQ_VTX_SEMANTIC_11
+    0x00000000, // SQ_VTX_SEMANTIC_12
+    0x00000000, // SQ_VTX_SEMANTIC_13
+    0x00000000, // SQ_VTX_SEMANTIC_14
+    0x00000000, // SQ_VTX_SEMANTIC_15
+    0x00000000, // SQ_VTX_SEMANTIC_16
+    0x00000000, // SQ_VTX_SEMANTIC_17
+    0x00000000, // SQ_VTX_SEMANTIC_18
+    0x00000000, // SQ_VTX_SEMANTIC_19
+    0x00000000, // SQ_VTX_SEMANTIC_20
+    0x00000000, // SQ_VTX_SEMANTIC_21
+    0x00000000, // SQ_VTX_SEMANTIC_22
+    0x00000000, // SQ_VTX_SEMANTIC_23
+    0x00000000, // SQ_VTX_SEMANTIC_24
+    0x00000000, // SQ_VTX_SEMANTIC_25
+    0x00000000, // SQ_VTX_SEMANTIC_26
+    0x00000000, // SQ_VTX_SEMANTIC_27
+    0x00000000, // SQ_VTX_SEMANTIC_28
+    0x00000000, // SQ_VTX_SEMANTIC_29
+    0x00000000, // SQ_VTX_SEMANTIC_30
+    0x00000000, // SQ_VTX_SEMANTIC_31
+    0xffffffff, // VGT_MAX_VTX_INDX
+    0x00000000, // VGT_MIN_VTX_INDX
+    0x00000000, // VGT_INDX_OFFSET
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0x00000000, // SX_ALPHA_TEST_CONTROL
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCILREFMASK
+    0x00000000, // DB_STENCILREFMASK_BF
+    0x00000000, // SX_ALPHA_REF
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0x00000000, // SPI_VS_OUT_ID_0
+    0x00000000, // SPI_VS_OUT_ID_1
+    0x00000000, // SPI_VS_OUT_ID_2
+    0x00000000, // SPI_VS_OUT_ID_3
+    0x00000000, // SPI_VS_OUT_ID_4
+    0x00000000, // SPI_VS_OUT_ID_5
+    0x00000000, // SPI_VS_OUT_ID_6
+    0x00000000, // SPI_VS_OUT_ID_7
+    0x00000000, // SPI_VS_OUT_ID_8
+    0x00000000, // SPI_VS_OUT_ID_9
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0x00000001, // SPI_THREAD_GROUPING
+    0x00000002, // SPI_PS_IN_CONTROL_0
+    0x00000000, // SPI_PS_IN_CONTROL_1
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000000, // SPI_INPUT_Z
+    0x00000000, // SPI_FOG_CNTL
+    0x00000000, // SPI_BARYC_CNTL
+    0x00000000, // SPI_PS_IN_CONTROL_2
+    0x00000000, // SPI_COMPUTE_INPUT_CNTL
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_X
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Y
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Z
+    0x00000000, // SPI_GPR_MGMT
+    0x00000000, // SPI_LDS_MGMT
+    0x00000000, // SPI_STACK_MGMT
+    0x00000000, // SPI_WAVE_MGMT_1
+    0x00000000, // SPI_WAVE_MGMT_2
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_ADDR_BASE
+    0x00003fff, // GDS_ADDR_SIZE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_ORDERED_COUNT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_APPEND_CONSUME_UAV0
+    0x00000000, // GDS_APPEND_CONSUME_UAV1
+    0x00000000, // GDS_APPEND_CONSUME_UAV2
+    0x00000000, // GDS_APPEND_CONSUME_UAV3
+    0x00000000, // GDS_APPEND_CONSUME_UAV4
+    0x00000000, // GDS_APPEND_CONSUME_UAV5
+    0x00000000, // GDS_APPEND_CONSUME_UAV6
+    0x00000000, // GDS_APPEND_CONSUME_UAV7
+    0x00000000, // GDS_APPEND_CONSUME_UAV8
+    0x00000000, // GDS_APPEND_CONSUME_UAV9
+    0x00000000, // GDS_APPEND_CONSUME_UAV10
+    0x00000000, // GDS_APPEND_CONSUME_UAV11
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const u32 SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+    0x00000000, // VGT_DMA_BASE_HI
+    0x00000000, // VGT_DMA_BASE
+};
+static const u32 SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // DB_DEPTH_CONTROL
+    0x00000000, // DB_EQAA
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000200, // DB_SHADER_CONTROL
+    0x00000000, // PA_CL_CLIP_CNTL
+    0x00000000, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0x00000000, // SQ_LSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_HSTMP_RING_ITEMSIZE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_PS
+    0x00000000, // SQ_PGM_RESOURCES_PS
+    0x00000000, // SQ_PGM_RESOURCES_2_PS
+    0x00000000, // SQ_PGM_EXPORTS_PS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_VS
+    0x00000000, // SQ_PGM_RESOURCES_VS
+    0x00000000, // SQ_PGM_RESOURCES_2_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_GS
+    0x00000000, // SQ_PGM_RESOURCES_GS
+    0x00000000, // SQ_PGM_RESOURCES_2_GS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_ES
+    0x00000000, // SQ_PGM_RESOURCES_ES
+    0x00000000, // SQ_PGM_RESOURCES_2_ES
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_FS
+    0x00000000, // SQ_PGM_RESOURCES_FS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_HS
+    0x00000000, // SQ_PGM_RESOURCES_HS
+    0x00000000, // SQ_PGM_RESOURCES_2_HS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_LS
+    0x00000000, // SQ_PGM_RESOURCES_LS
+    0x00000000, // SQ_PGM_RESOURCES_2_LS
+};
+static const u32 SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // SQ_LDS_ALLOC
+    0x00000000, // SQ_LDS_ALLOC_PS
+    0x00000000, // SQ_VTX_SEMANTIC_CLEAR
+    0, // HOLE
+    0x00000000, // SQ_THREAD_TRACE_CTRL
+    0, // HOLE
+    0x00000000, // SQ_ESGS_RING_ITEMSIZE
+    0x00000000, // SQ_GSVS_RING_ITEMSIZE
+    0x00000000, // SQ_ESTMP_RING_ITEMSIZE
+    0x00000000, // SQ_GSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_VSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_PSTMP_RING_ITEMSIZE
+    0, // HOLE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_1
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_2
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_3
+    0x00000000, // SQ_GSVS_RING_OFFSET_1
+    0x00000000, // SQ_GSVS_RING_OFFSET_2
+    0x00000000, // SQ_GSVS_RING_OFFSET_3
+    0x00000000, // SQ_GWS_RING_OFFSET
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_15
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0x00000000, // VGT_OUTPUT_PATH_CNTL
+    0x00000000, // VGT_HOS_CNTL
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0x00000000, // VGT_HOS_REUSE_DEPTH
+    0x00000000, // VGT_GROUP_PRIM_TYPE
+    0x00000000, // VGT_GROUP_FIRST_DECR
+    0x00000000, // VGT_GROUP_DECR
+    0x00000000, // VGT_GROUP_VECT_0_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_CNTL
+    0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL
+    0x00000000, // VGT_GS_MODE
+    0, // HOLE
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0x00000100, // VGT_GS_PER_ES
+    0x00000080, // VGT_ES_PER_GS
+    0x00000002, // VGT_GS_PER_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_GS_OUT_PRIM_TYPE
+    0x00000000, // IA_ENHANCE
+};
+static const u32 SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // VGT_DMA_MAX_SIZE
+    0x00000000, // VGT_DMA_INDEX_TYPE
+    0, // HOLE
+    0x00000000, // VGT_PRIMITIVEID_EN
+    0x00000000, // VGT_DMA_NUM_INSTANCES
+};
+static const u32 SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_INSTANCE_STEP_RATE_0
+    0x00000000, // VGT_INSTANCE_STEP_RATE_1
+    0x000000ff, // IA_MULTI_VGT_PARAM
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_REUSE_OFF
+    0x00000000, // VGT_VTX_CNT_EN
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_3
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+};
+static const u32 SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0x00000000, // VGT_STRMOUT_CONFIG
+    0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
+    0x00000000, // CB_IMMED0_BASE
+    0x00000000, // CB_IMMED1_BASE
+    0x00000000, // CB_IMMED2_BASE
+    0x00000000, // CB_IMMED3_BASE
+    0x00000000, // CB_IMMED4_BASE
+    0x00000000, // CB_IMMED5_BASE
+    0x00000000, // CB_IMMED6_BASE
+    0x00000000, // CB_IMMED7_BASE
+    0x00000000, // CB_IMMED8_BASE
+    0x00000000, // CB_IMMED9_BASE
+    0x00000000, // CB_IMMED10_BASE
+    0x00000000, // CB_IMMED11_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_CENTROID_PRIORITY_0
+    0x00000000, // PA_SC_CENTROID_PRIORITY_1
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3
+    0xffffffff, // PA_SC_AA_MASK_X0Y0_X1Y0
+    0xffffffff, // PA_SC_AA_MASK_X0Y1_X1Y1
+    0x00000000, // CB_CLRCMP_CONTROL
+    0x00000000, // CB_CLRCMP_SRC
+    0x00000000, // CB_CLRCMP_DST
+    0x00000000, // CB_CLRCMP_MSK
+    0, // HOLE
+    0, // HOLE
+    0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL
+    0x00000010, // VGT_OUT_DEALLOC_CNTL
+    0x00000000, // CB_COLOR0_BASE
+    0x00000000, // CB_COLOR0_PITCH
+    0x00000000, // CB_COLOR0_SLICE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0x00000000, // CB_COLOR0_DIM
+    0x00000000, // CB_COLOR0_CMASK
+    0x00000000, // CB_COLOR0_CMASK_SLICE
+    0x00000000, // CB_COLOR0_FMASK
+    0x00000000, // CB_COLOR0_FMASK_SLICE
+    0x00000000, // CB_COLOR0_CLEAR_WORD0
+    0x00000000, // CB_COLOR0_CLEAR_WORD1
+    0x00000000, // CB_COLOR0_CLEAR_WORD2
+    0x00000000, // CB_COLOR0_CLEAR_WORD3
+    0x00000000, // CB_COLOR1_BASE
+    0x00000000, // CB_COLOR1_PITCH
+    0x00000000, // CB_COLOR1_SLICE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0x00000000, // CB_COLOR1_DIM
+    0x00000000, // CB_COLOR1_CMASK
+    0x00000000, // CB_COLOR1_CMASK_SLICE
+    0x00000000, // CB_COLOR1_FMASK
+    0x00000000, // CB_COLOR1_FMASK_SLICE
+    0x00000000, // CB_COLOR1_CLEAR_WORD0
+    0x00000000, // CB_COLOR1_CLEAR_WORD1
+    0x00000000, // CB_COLOR1_CLEAR_WORD2
+    0x00000000, // CB_COLOR1_CLEAR_WORD3
+    0x00000000, // CB_COLOR2_BASE
+    0x00000000, // CB_COLOR2_PITCH
+    0x00000000, // CB_COLOR2_SLICE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0x00000000, // CB_COLOR2_DIM
+    0x00000000, // CB_COLOR2_CMASK
+    0x00000000, // CB_COLOR2_CMASK_SLICE
+    0x00000000, // CB_COLOR2_FMASK
+    0x00000000, // CB_COLOR2_FMASK_SLICE
+    0x00000000, // CB_COLOR2_CLEAR_WORD0
+    0x00000000, // CB_COLOR2_CLEAR_WORD1
+    0x00000000, // CB_COLOR2_CLEAR_WORD2
+    0x00000000, // CB_COLOR2_CLEAR_WORD3
+    0x00000000, // CB_COLOR3_BASE
+    0x00000000, // CB_COLOR3_PITCH
+    0x00000000, // CB_COLOR3_SLICE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0x00000000, // CB_COLOR3_DIM
+    0x00000000, // CB_COLOR3_CMASK
+    0x00000000, // CB_COLOR3_CMASK_SLICE
+    0x00000000, // CB_COLOR3_FMASK
+    0x00000000, // CB_COLOR3_FMASK_SLICE
+    0x00000000, // CB_COLOR3_CLEAR_WORD0
+    0x00000000, // CB_COLOR3_CLEAR_WORD1
+    0x00000000, // CB_COLOR3_CLEAR_WORD2
+    0x00000000, // CB_COLOR3_CLEAR_WORD3
+    0x00000000, // CB_COLOR4_BASE
+    0x00000000, // CB_COLOR4_PITCH
+    0x00000000, // CB_COLOR4_SLICE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0x00000000, // CB_COLOR4_DIM
+    0x00000000, // CB_COLOR4_CMASK
+    0x00000000, // CB_COLOR4_CMASK_SLICE
+    0x00000000, // CB_COLOR4_FMASK
+    0x00000000, // CB_COLOR4_FMASK_SLICE
+    0x00000000, // CB_COLOR4_CLEAR_WORD0
+    0x00000000, // CB_COLOR4_CLEAR_WORD1
+    0x00000000, // CB_COLOR4_CLEAR_WORD2
+    0x00000000, // CB_COLOR4_CLEAR_WORD3
+    0x00000000, // CB_COLOR5_BASE
+    0x00000000, // CB_COLOR5_PITCH
+    0x00000000, // CB_COLOR5_SLICE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0x00000000, // CB_COLOR5_DIM
+    0x00000000, // CB_COLOR5_CMASK
+    0x00000000, // CB_COLOR5_CMASK_SLICE
+    0x00000000, // CB_COLOR5_FMASK
+    0x00000000, // CB_COLOR5_FMASK_SLICE
+    0x00000000, // CB_COLOR5_CLEAR_WORD0
+    0x00000000, // CB_COLOR5_CLEAR_WORD1
+    0x00000000, // CB_COLOR5_CLEAR_WORD2
+    0x00000000, // CB_COLOR5_CLEAR_WORD3
+    0x00000000, // CB_COLOR6_BASE
+    0x00000000, // CB_COLOR6_PITCH
+    0x00000000, // CB_COLOR6_SLICE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0x00000000, // CB_COLOR6_DIM
+    0x00000000, // CB_COLOR6_CMASK
+    0x00000000, // CB_COLOR6_CMASK_SLICE
+    0x00000000, // CB_COLOR6_FMASK
+    0x00000000, // CB_COLOR6_FMASK_SLICE
+    0x00000000, // CB_COLOR6_CLEAR_WORD0
+    0x00000000, // CB_COLOR6_CLEAR_WORD1
+    0x00000000, // CB_COLOR6_CLEAR_WORD2
+    0x00000000, // CB_COLOR6_CLEAR_WORD3
+    0x00000000, // CB_COLOR7_BASE
+    0x00000000, // CB_COLOR7_PITCH
+    0x00000000, // CB_COLOR7_SLICE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0x00000000, // CB_COLOR7_DIM
+    0x00000000, // CB_COLOR7_CMASK
+    0x00000000, // CB_COLOR7_CMASK_SLICE
+    0x00000000, // CB_COLOR7_FMASK
+    0x00000000, // CB_COLOR7_FMASK_SLICE
+    0x00000000, // CB_COLOR7_CLEAR_WORD0
+    0x00000000, // CB_COLOR7_CLEAR_WORD1
+    0x00000000, // CB_COLOR7_CLEAR_WORD2
+    0x00000000, // CB_COLOR7_CLEAR_WORD3
+    0x00000000, // CB_COLOR8_BASE
+    0x00000000, // CB_COLOR8_PITCH
+    0x00000000, // CB_COLOR8_SLICE
+    0x00000000, // CB_COLOR8_VIEW
+    0x00000000, // CB_COLOR8_INFO
+    0x00000000, // CB_COLOR8_ATTRIB
+    0x00000000, // CB_COLOR8_DIM
+    0x00000000, // CB_COLOR9_BASE
+    0x00000000, // CB_COLOR9_PITCH
+    0x00000000, // CB_COLOR9_SLICE
+    0x00000000, // CB_COLOR9_VIEW
+    0x00000000, // CB_COLOR9_INFO
+    0x00000000, // CB_COLOR9_ATTRIB
+    0x00000000, // CB_COLOR9_DIM
+    0x00000000, // CB_COLOR10_BASE
+    0x00000000, // CB_COLOR10_PITCH
+    0x00000000, // CB_COLOR10_SLICE
+    0x00000000, // CB_COLOR10_VIEW
+    0x00000000, // CB_COLOR10_INFO
+    0x00000000, // CB_COLOR10_ATTRIB
+    0x00000000, // CB_COLOR10_DIM
+    0x00000000, // CB_COLOR11_BASE
+    0x00000000, // CB_COLOR11_PITCH
+    0x00000000, // CB_COLOR11_SLICE
+    0x00000000, // CB_COLOR11_VIEW
+    0x00000000, // CB_COLOR11_INFO
+    0x00000000, // CB_COLOR11_ATTRIB
+    0x00000000, // CB_COLOR11_DIM
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_15
+};
+static const struct cs_extent_def SECT_CONTEXT_defs[] =
+{
+    {SECT_CONTEXT_def_1, 0x0000a000, 488 },
+    {SECT_CONTEXT_def_2, 0x0000a1f5, 6 },
+    {SECT_CONTEXT_def_3, 0x0000a200, 55 },
+    {SECT_CONTEXT_def_4, 0x0000a23a, 99 },
+    {SECT_CONTEXT_def_5, 0x0000a29e, 5 },
+    {SECT_CONTEXT_def_6, 0x0000a2a5, 56 },
+    {SECT_CONTEXT_def_7, 0x0000a2de, 290 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CLEAR_def_1[] =
+{
+    0xffffffff, // SQ_TEX_SAMPLER_CLEAR
+    0xffffffff, // SQ_TEX_RESOURCE_CLEAR
+    0xffffffff, // SQ_LOOP_BOOL_CLEAR
+};
+static const struct cs_extent_def SECT_CLEAR_defs[] =
+{
+    {SECT_CLEAR_def_1, 0x0000ffc0, 3 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CTRLCONST_def_1[] =
+{
+    0x00000000, // SQ_VTX_BASE_VTX_LOC
+    0x00000000, // SQ_VTX_START_INST_LOC
+};
+static const struct cs_extent_def SECT_CTRLCONST_defs[] =
+{
+    {SECT_CTRLCONST_def_1, 0x0000f3fc, 2 },
+    { 0, 0, 0 }
+};
+struct cs_section_def cayman_cs_data[] = {
+    { SECT_CONTEXT_defs, SECT_CONTEXT },
+    { SECT_CLEAR_defs, SECT_CLEAR },
+    { SECT_CTRLCONST_defs, SECT_CTRLCONST },
+    { 0, SECT_NONE }
+};
diff --git a/drivers/gpu/drm/radeon/clearstate_defs.h b/drivers/gpu/drm/radeon/clearstate_defs.h
new file mode 100644
index 0000000..3eda707
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_defs.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef CLEARSTATE_DEFS_H
+#define CLEARSTATE_DEFS_H
+
+enum section_id {
+    SECT_NONE,
+    SECT_CONTEXT,
+    SECT_CLEAR,
+    SECT_CTRLCONST
+};
+
+struct cs_extent_def {
+    const unsigned int *extent;
+    const unsigned int reg_index;
+    const unsigned int reg_count;
+};
+
+struct cs_section_def {
+    const struct cs_extent_def *section;
+    const enum section_id id;
+};
+
+#endif
diff --git a/drivers/gpu/drm/radeon/clearstate_evergreen.h b/drivers/gpu/drm/radeon/clearstate_evergreen.h
new file mode 100644
index 0000000..4791d85
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_evergreen.h
@@ -0,0 +1,1080 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+static const u32 SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_DEPTH_SIZE
+    0x00000000, // DB_DEPTH_SLICE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_15
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+    0x00000000, // SX_MISC
+    0x00000000, // SX_SURFACE_SYNC
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_VTX_SEMANTIC_0
+    0x00000000, // SQ_VTX_SEMANTIC_1
+    0x00000000, // SQ_VTX_SEMANTIC_2
+    0x00000000, // SQ_VTX_SEMANTIC_3
+    0x00000000, // SQ_VTX_SEMANTIC_4
+    0x00000000, // SQ_VTX_SEMANTIC_5
+    0x00000000, // SQ_VTX_SEMANTIC_6
+    0x00000000, // SQ_VTX_SEMANTIC_7
+    0x00000000, // SQ_VTX_SEMANTIC_8
+    0x00000000, // SQ_VTX_SEMANTIC_9
+    0x00000000, // SQ_VTX_SEMANTIC_10
+    0x00000000, // SQ_VTX_SEMANTIC_11
+    0x00000000, // SQ_VTX_SEMANTIC_12
+    0x00000000, // SQ_VTX_SEMANTIC_13
+    0x00000000, // SQ_VTX_SEMANTIC_14
+    0x00000000, // SQ_VTX_SEMANTIC_15
+    0x00000000, // SQ_VTX_SEMANTIC_16
+    0x00000000, // SQ_VTX_SEMANTIC_17
+    0x00000000, // SQ_VTX_SEMANTIC_18
+    0x00000000, // SQ_VTX_SEMANTIC_19
+    0x00000000, // SQ_VTX_SEMANTIC_20
+    0x00000000, // SQ_VTX_SEMANTIC_21
+    0x00000000, // SQ_VTX_SEMANTIC_22
+    0x00000000, // SQ_VTX_SEMANTIC_23
+    0x00000000, // SQ_VTX_SEMANTIC_24
+    0x00000000, // SQ_VTX_SEMANTIC_25
+    0x00000000, // SQ_VTX_SEMANTIC_26
+    0x00000000, // SQ_VTX_SEMANTIC_27
+    0x00000000, // SQ_VTX_SEMANTIC_28
+    0x00000000, // SQ_VTX_SEMANTIC_29
+    0x00000000, // SQ_VTX_SEMANTIC_30
+    0x00000000, // SQ_VTX_SEMANTIC_31
+    0xffffffff, // VGT_MAX_VTX_INDX
+    0x00000000, // VGT_MIN_VTX_INDX
+    0x00000000, // VGT_INDX_OFFSET
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0x00000000, // SX_ALPHA_TEST_CONTROL
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCILREFMASK
+    0x00000000, // DB_STENCILREFMASK_BF
+    0x00000000, // SX_ALPHA_REF
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0x00000000, // SPI_VS_OUT_ID_0
+    0x00000000, // SPI_VS_OUT_ID_1
+    0x00000000, // SPI_VS_OUT_ID_2
+    0x00000000, // SPI_VS_OUT_ID_3
+    0x00000000, // SPI_VS_OUT_ID_4
+    0x00000000, // SPI_VS_OUT_ID_5
+    0x00000000, // SPI_VS_OUT_ID_6
+    0x00000000, // SPI_VS_OUT_ID_7
+    0x00000000, // SPI_VS_OUT_ID_8
+    0x00000000, // SPI_VS_OUT_ID_9
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0x00000001, // SPI_THREAD_GROUPING
+    0x00000000, // SPI_PS_IN_CONTROL_0
+    0x00000000, // SPI_PS_IN_CONTROL_1
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000000, // SPI_INPUT_Z
+    0x00000000, // SPI_FOG_CNTL
+    0x00000000, // SPI_BARYC_CNTL
+    0x00000000, // SPI_PS_IN_CONTROL_2
+    0x00000000, // SPI_COMPUTE_INPUT_CNTL
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_X
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Y
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Z
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_ADDR_BASE
+    0x00003fff, // GDS_ADDR_SIZE
+    0x00000001, // GDS_ORDERED_WAVE_PER_SE
+    0x00000000, // GDS_APPEND_CONSUME_UAV0
+    0x00000000, // GDS_APPEND_CONSUME_UAV1
+    0x00000000, // GDS_APPEND_CONSUME_UAV2
+    0x00000000, // GDS_APPEND_CONSUME_UAV3
+    0x00000000, // GDS_APPEND_CONSUME_UAV4
+    0x00000000, // GDS_APPEND_CONSUME_UAV5
+    0x00000000, // GDS_APPEND_CONSUME_UAV6
+    0x00000000, // GDS_APPEND_CONSUME_UAV7
+    0x00000000, // GDS_APPEND_CONSUME_UAV8
+    0x00000000, // GDS_APPEND_CONSUME_UAV9
+    0x00000000, // GDS_APPEND_CONSUME_UAV10
+    0x00000000, // GDS_APPEND_CONSUME_UAV11
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const u32 SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+    0x00000000, // VGT_DMA_BASE_HI
+    0x00000000, // VGT_DMA_BASE
+};
+static const u32 SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // DB_DEPTH_CONTROL
+    0, // HOLE
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000200, // DB_SHADER_CONTROL
+    0x00000000, // PA_CL_CLIP_CNTL
+    0x00000000, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0x00000000, // SQ_LSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_HSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_DYN_GPR_RESOURCE_LIMIT_1
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_PS
+    0x00000000, // SQ_PGM_RESOURCES_PS
+    0x00000000, // SQ_PGM_RESOURCES_2_PS
+    0x00000000, // SQ_PGM_EXPORTS_PS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_VS
+    0x00000000, // SQ_PGM_RESOURCES_VS
+    0x00000000, // SQ_PGM_RESOURCES_2_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_GS
+    0x00000000, // SQ_PGM_RESOURCES_GS
+    0x00000000, // SQ_PGM_RESOURCES_2_GS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_ES
+    0x00000000, // SQ_PGM_RESOURCES_ES
+    0x00000000, // SQ_PGM_RESOURCES_2_ES
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_FS
+    0x00000000, // SQ_PGM_RESOURCES_FS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_HS
+    0x00000000, // SQ_PGM_RESOURCES_HS
+    0x00000000, // SQ_PGM_RESOURCES_2_HS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_LS
+    0x00000000, // SQ_PGM_RESOURCES_LS
+    0x00000000, // SQ_PGM_RESOURCES_2_LS
+};
+static const u32 SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // SQ_LDS_ALLOC
+    0x00000000, // SQ_LDS_ALLOC_PS
+    0x00000000, // SQ_VTX_SEMANTIC_CLEAR
+    0, // HOLE
+    0x00000000, // SQ_THREAD_TRACE_CTRL
+    0, // HOLE
+    0x00000000, // SQ_ESGS_RING_ITEMSIZE
+    0x00000000, // SQ_GSVS_RING_ITEMSIZE
+    0x00000000, // SQ_ESTMP_RING_ITEMSIZE
+    0x00000000, // SQ_GSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_VSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_PSTMP_RING_ITEMSIZE
+    0, // HOLE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_1
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_2
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_3
+    0x00000000, // SQ_GSVS_RING_OFFSET_1
+    0x00000000, // SQ_GSVS_RING_OFFSET_2
+    0x00000000, // SQ_GSVS_RING_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_15
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0x00000000, // VGT_OUTPUT_PATH_CNTL
+    0x00000000, // VGT_HOS_CNTL
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0x00000000, // VGT_HOS_REUSE_DEPTH
+    0x00000000, // VGT_GROUP_PRIM_TYPE
+    0x00000000, // VGT_GROUP_FIRST_DECR
+    0x00000000, // VGT_GROUP_DECR
+    0x00000000, // VGT_GROUP_VECT_0_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_CNTL
+    0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL
+    0x00000000, // VGT_GS_MODE
+    0, // HOLE
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0x00000000, // VGT_GS_PER_ES
+    0x00000000, // VGT_ES_PER_GS
+    0x00000000, // VGT_GS_PER_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_GS_OUT_PRIM_TYPE
+};
+static const u32 SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // VGT_DMA_MAX_SIZE
+    0x00000000, // VGT_DMA_INDEX_TYPE
+    0, // HOLE
+    0x00000000, // VGT_PRIMITIVEID_EN
+    0x00000000, // VGT_DMA_NUM_INSTANCES
+};
+static const u32 SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_INSTANCE_STEP_RATE_0
+    0x00000000, // VGT_INSTANCE_STEP_RATE_1
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_REUSE_OFF
+    0x00000000, // VGT_VTX_CNT_EN
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_3
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0x00000000, // VGT_LS_SIZE
+    0x00000000, // VGT_HS_SIZE
+    0x00000000, // VGT_LS_HS_ALLOC
+    0x00000000, // VGT_HS_PATCH_CONST
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+};
+static const u32 SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0x00000000, // VGT_STRMOUT_CONFIG
+    0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
+    0x00000000, // CB_IMMED0_BASE
+    0x00000000, // CB_IMMED1_BASE
+    0x00000000, // CB_IMMED2_BASE
+    0x00000000, // CB_IMMED3_BASE
+    0x00000000, // CB_IMMED4_BASE
+    0x00000000, // CB_IMMED5_BASE
+    0x00000000, // CB_IMMED6_BASE
+    0x00000000, // CB_IMMED7_BASE
+    0x00000000, // CB_IMMED8_BASE
+    0x00000000, // CB_IMMED9_BASE
+    0x00000000, // CB_IMMED10_BASE
+    0x00000000, // CB_IMMED11_BASE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_4
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_5
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_6
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_7
+    0xffffffff, // PA_SC_AA_MASK
+    0x00000000, // CB_CLRCMP_CONTROL
+    0x00000000, // CB_CLRCMP_SRC
+    0x00000000, // CB_CLRCMP_DST
+    0x00000000, // CB_CLRCMP_MSK
+    0, // HOLE
+    0, // HOLE
+    0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL
+    0x00000010, // VGT_OUT_DEALLOC_CNTL
+    0x00000000, // CB_COLOR0_BASE
+    0x00000000, // CB_COLOR0_PITCH
+    0x00000000, // CB_COLOR0_SLICE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0x00000000, // CB_COLOR0_DIM
+    0x00000000, // CB_COLOR0_CMASK
+    0x00000000, // CB_COLOR0_CMASK_SLICE
+    0x00000000, // CB_COLOR0_FMASK
+    0x00000000, // CB_COLOR0_FMASK_SLICE
+    0x00000000, // CB_COLOR0_CLEAR_WORD0
+    0x00000000, // CB_COLOR0_CLEAR_WORD1
+    0x00000000, // CB_COLOR0_CLEAR_WORD2
+    0x00000000, // CB_COLOR0_CLEAR_WORD3
+    0x00000000, // CB_COLOR1_BASE
+    0x00000000, // CB_COLOR1_PITCH
+    0x00000000, // CB_COLOR1_SLICE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0x00000000, // CB_COLOR1_DIM
+    0x00000000, // CB_COLOR1_CMASK
+    0x00000000, // CB_COLOR1_CMASK_SLICE
+    0x00000000, // CB_COLOR1_FMASK
+    0x00000000, // CB_COLOR1_FMASK_SLICE
+    0x00000000, // CB_COLOR1_CLEAR_WORD0
+    0x00000000, // CB_COLOR1_CLEAR_WORD1
+    0x00000000, // CB_COLOR1_CLEAR_WORD2
+    0x00000000, // CB_COLOR1_CLEAR_WORD3
+    0x00000000, // CB_COLOR2_BASE
+    0x00000000, // CB_COLOR2_PITCH
+    0x00000000, // CB_COLOR2_SLICE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0x00000000, // CB_COLOR2_DIM
+    0x00000000, // CB_COLOR2_CMASK
+    0x00000000, // CB_COLOR2_CMASK_SLICE
+    0x00000000, // CB_COLOR2_FMASK
+    0x00000000, // CB_COLOR2_FMASK_SLICE
+    0x00000000, // CB_COLOR2_CLEAR_WORD0
+    0x00000000, // CB_COLOR2_CLEAR_WORD1
+    0x00000000, // CB_COLOR2_CLEAR_WORD2
+    0x00000000, // CB_COLOR2_CLEAR_WORD3
+    0x00000000, // CB_COLOR3_BASE
+    0x00000000, // CB_COLOR3_PITCH
+    0x00000000, // CB_COLOR3_SLICE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0x00000000, // CB_COLOR3_DIM
+    0x00000000, // CB_COLOR3_CMASK
+    0x00000000, // CB_COLOR3_CMASK_SLICE
+    0x00000000, // CB_COLOR3_FMASK
+    0x00000000, // CB_COLOR3_FMASK_SLICE
+    0x00000000, // CB_COLOR3_CLEAR_WORD0
+    0x00000000, // CB_COLOR3_CLEAR_WORD1
+    0x00000000, // CB_COLOR3_CLEAR_WORD2
+    0x00000000, // CB_COLOR3_CLEAR_WORD3
+    0x00000000, // CB_COLOR4_BASE
+    0x00000000, // CB_COLOR4_PITCH
+    0x00000000, // CB_COLOR4_SLICE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0x00000000, // CB_COLOR4_DIM
+    0x00000000, // CB_COLOR4_CMASK
+    0x00000000, // CB_COLOR4_CMASK_SLICE
+    0x00000000, // CB_COLOR4_FMASK
+    0x00000000, // CB_COLOR4_FMASK_SLICE
+    0x00000000, // CB_COLOR4_CLEAR_WORD0
+    0x00000000, // CB_COLOR4_CLEAR_WORD1
+    0x00000000, // CB_COLOR4_CLEAR_WORD2
+    0x00000000, // CB_COLOR4_CLEAR_WORD3
+    0x00000000, // CB_COLOR5_BASE
+    0x00000000, // CB_COLOR5_PITCH
+    0x00000000, // CB_COLOR5_SLICE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0x00000000, // CB_COLOR5_DIM
+    0x00000000, // CB_COLOR5_CMASK
+    0x00000000, // CB_COLOR5_CMASK_SLICE
+    0x00000000, // CB_COLOR5_FMASK
+    0x00000000, // CB_COLOR5_FMASK_SLICE
+    0x00000000, // CB_COLOR5_CLEAR_WORD0
+    0x00000000, // CB_COLOR5_CLEAR_WORD1
+    0x00000000, // CB_COLOR5_CLEAR_WORD2
+    0x00000000, // CB_COLOR5_CLEAR_WORD3
+    0x00000000, // CB_COLOR6_BASE
+    0x00000000, // CB_COLOR6_PITCH
+    0x00000000, // CB_COLOR6_SLICE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0x00000000, // CB_COLOR6_DIM
+    0x00000000, // CB_COLOR6_CMASK
+    0x00000000, // CB_COLOR6_CMASK_SLICE
+    0x00000000, // CB_COLOR6_FMASK
+    0x00000000, // CB_COLOR6_FMASK_SLICE
+    0x00000000, // CB_COLOR6_CLEAR_WORD0
+    0x00000000, // CB_COLOR6_CLEAR_WORD1
+    0x00000000, // CB_COLOR6_CLEAR_WORD2
+    0x00000000, // CB_COLOR6_CLEAR_WORD3
+    0x00000000, // CB_COLOR7_BASE
+    0x00000000, // CB_COLOR7_PITCH
+    0x00000000, // CB_COLOR7_SLICE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0x00000000, // CB_COLOR7_DIM
+    0x00000000, // CB_COLOR7_CMASK
+    0x00000000, // CB_COLOR7_CMASK_SLICE
+    0x00000000, // CB_COLOR7_FMASK
+    0x00000000, // CB_COLOR7_FMASK_SLICE
+    0x00000000, // CB_COLOR7_CLEAR_WORD0
+    0x00000000, // CB_COLOR7_CLEAR_WORD1
+    0x00000000, // CB_COLOR7_CLEAR_WORD2
+    0x00000000, // CB_COLOR7_CLEAR_WORD3
+    0x00000000, // CB_COLOR8_BASE
+    0x00000000, // CB_COLOR8_PITCH
+    0x00000000, // CB_COLOR8_SLICE
+    0x00000000, // CB_COLOR8_VIEW
+    0x00000000, // CB_COLOR8_INFO
+    0x00000000, // CB_COLOR8_ATTRIB
+    0x00000000, // CB_COLOR8_DIM
+    0x00000000, // CB_COLOR9_BASE
+    0x00000000, // CB_COLOR9_PITCH
+    0x00000000, // CB_COLOR9_SLICE
+    0x00000000, // CB_COLOR9_VIEW
+    0x00000000, // CB_COLOR9_INFO
+    0x00000000, // CB_COLOR9_ATTRIB
+    0x00000000, // CB_COLOR9_DIM
+    0x00000000, // CB_COLOR10_BASE
+    0x00000000, // CB_COLOR10_PITCH
+    0x00000000, // CB_COLOR10_SLICE
+    0x00000000, // CB_COLOR10_VIEW
+    0x00000000, // CB_COLOR10_INFO
+    0x00000000, // CB_COLOR10_ATTRIB
+    0x00000000, // CB_COLOR10_DIM
+    0x00000000, // CB_COLOR11_BASE
+    0x00000000, // CB_COLOR11_PITCH
+    0x00000000, // CB_COLOR11_SLICE
+    0x00000000, // CB_COLOR11_VIEW
+    0x00000000, // CB_COLOR11_INFO
+    0x00000000, // CB_COLOR11_ATTRIB
+    0x00000000, // CB_COLOR11_DIM
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_15
+};
+static const struct cs_extent_def SECT_CONTEXT_defs[] =
+{
+    {SECT_CONTEXT_def_1, 0x0000a000, 488 },
+    {SECT_CONTEXT_def_2, 0x0000a1f5, 6 },
+    {SECT_CONTEXT_def_3, 0x0000a200, 55 },
+    {SECT_CONTEXT_def_4, 0x0000a23a, 98 },
+    {SECT_CONTEXT_def_5, 0x0000a29e, 5 },
+    {SECT_CONTEXT_def_6, 0x0000a2a5, 56 },
+    {SECT_CONTEXT_def_7, 0x0000a2de, 290 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CLEAR_def_1[] =
+{
+    0xffffffff, // SQ_TEX_SAMPLER_CLEAR
+    0xffffffff, // SQ_TEX_RESOURCE_CLEAR
+    0xffffffff, // SQ_LOOP_BOOL_CLEAR
+};
+static const struct cs_extent_def SECT_CLEAR_defs[] =
+{
+    {SECT_CLEAR_def_1, 0x0000ffc0, 3 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CTRLCONST_def_1[] =
+{
+    0x00000000, // SQ_VTX_BASE_VTX_LOC
+    0x00000000, // SQ_VTX_START_INST_LOC
+};
+static const struct cs_extent_def SECT_CTRLCONST_defs[] =
+{
+    {SECT_CTRLCONST_def_1, 0x0000f3fc, 2 },
+    { 0, 0, 0 }
+};
+struct cs_section_def evergreen_cs_data[] = {
+    { SECT_CONTEXT_defs, SECT_CONTEXT },
+    { SECT_CLEAR_defs, SECT_CLEAR },
+    { SECT_CTRLCONST_defs, SECT_CTRLCONST },
+    { 0, SECT_NONE }
+};
diff --git a/drivers/gpu/drm/radeon/clearstate_si.h b/drivers/gpu/drm/radeon/clearstate_si.h
new file mode 100644
index 0000000..b994cb2
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_si.h
@@ -0,0 +1,941 @@
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+static const u32 si_SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_DEPTH_BOUNDS_MIN
+    0x00000000, // DB_DEPTH_BOUNDS_MAX
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0x00000000, // DB_DEPTH_INFO
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_DEPTH_SIZE
+    0x00000000, // DB_DEPTH_SLICE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // TA_BC_BASE_ADDR
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // COHER_DEST_BASE_2
+    0x00000000, // COHER_DEST_BASE_3
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+};
+static const u32 si_SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0x00000000, // CP_RINGID
+    0x00000000, // CP_VMID
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0xffffffff, // VGT_MAX_VTX_INDX
+    0x00000000, // VGT_MIN_VTX_INDX
+    0x00000000, // VGT_INDX_OFFSET
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0, // HOLE
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCIL_CONTROL
+    0x00000000, // DB_STENCILREFMASK
+    0x00000000, // DB_STENCILREFMASK_BF
+    0, // HOLE
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0, // HOLE
+    0x00000000, // SPI_PS_INPUT_ENA
+    0x00000000, // SPI_PS_INPUT_ADDR
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000002, // SPI_PS_IN_CONTROL
+    0, // HOLE
+    0x00000000, // SPI_BARYC_CNTL
+    0, // HOLE
+    0x00000000, // SPI_TMPRING_SIZE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SPI_WAVE_MGMT_1
+    0x00000000, // SPI_WAVE_MGMT_2
+    0x00000000, // SPI_SHADER_POS_FORMAT
+    0x00000000, // SPI_SHADER_Z_FORMAT
+    0x00000000, // SPI_SHADER_COL_FORMAT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const u32 si_SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+    0x00000000, // VGT_DMA_BASE_HI
+    0x00000000, // VGT_DMA_BASE
+};
+static const u32 si_SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // DB_DEPTH_CONTROL
+    0x00000000, // DB_EQAA
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000000, // DB_SHADER_CONTROL
+    0x00090000, // PA_CL_CLIP_CNTL
+    0x00000004, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0x00000000, // VGT_OUTPUT_PATH_CNTL
+    0x00000000, // VGT_HOS_CNTL
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0x00000000, // VGT_HOS_REUSE_DEPTH
+    0x00000000, // VGT_GROUP_PRIM_TYPE
+    0x00000000, // VGT_GROUP_FIRST_DECR
+    0x00000000, // VGT_GROUP_DECR
+    0x00000000, // VGT_GROUP_VECT_0_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_CNTL
+    0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL
+    0x00000000, // VGT_GS_MODE
+    0, // HOLE
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0x00000100, // VGT_GS_PER_ES
+    0x00000080, // VGT_ES_PER_GS
+    0x00000002, // VGT_GS_PER_VS
+    0x00000000, // VGT_GSVS_RING_OFFSET_1
+    0x00000000, // VGT_GSVS_RING_OFFSET_2
+    0x00000000, // VGT_GSVS_RING_OFFSET_3
+    0x00000000, // VGT_GS_OUT_PRIM_TYPE
+    0x00000000, // IA_ENHANCE
+};
+static const u32 si_SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // VGT_PRIMITIVEID_EN
+};
+static const u32 si_SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_PRIMITIVEID_RESET
+};
+static const u32 si_SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_INSTANCE_STEP_RATE_0
+    0x00000000, // VGT_INSTANCE_STEP_RATE_1
+    0x000000ff, // IA_MULTI_VGT_PARAM
+    0x00000000, // VGT_ESGS_RING_ITEMSIZE
+    0x00000000, // VGT_GSVS_RING_ITEMSIZE
+    0x00000000, // VGT_REUSE_OFF
+    0x00000000, // VGT_VTX_CNT_EN
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_0
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_1
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_2
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_3
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0x00000000, // VGT_GS_VERT_ITEMSIZE
+    0x00000000, // VGT_GS_VERT_ITEMSIZE_1
+    0x00000000, // VGT_GS_VERT_ITEMSIZE_2
+    0x00000000, // VGT_GS_VERT_ITEMSIZE_3
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+    0, // HOLE
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0x00000000, // VGT_STRMOUT_CONFIG
+    0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_CENTROID_PRIORITY_0
+    0x00000000, // PA_SC_CENTROID_PRIORITY_1
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3
+    0xffffffff, // PA_SC_AA_MASK_X0Y0_X1Y0
+    0xffffffff, // PA_SC_AA_MASK_X0Y1_X1Y1
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL
+    0x00000010, // VGT_OUT_DEALLOC_CNTL
+    0x00000000, // CB_COLOR0_BASE
+    0x00000000, // CB_COLOR0_PITCH
+    0x00000000, // CB_COLOR0_SLICE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR0_CMASK
+    0x00000000, // CB_COLOR0_CMASK_SLICE
+    0x00000000, // CB_COLOR0_FMASK
+    0x00000000, // CB_COLOR0_FMASK_SLICE
+    0x00000000, // CB_COLOR0_CLEAR_WORD0
+    0x00000000, // CB_COLOR0_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR1_BASE
+    0x00000000, // CB_COLOR1_PITCH
+    0x00000000, // CB_COLOR1_SLICE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR1_CMASK
+    0x00000000, // CB_COLOR1_CMASK_SLICE
+    0x00000000, // CB_COLOR1_FMASK
+    0x00000000, // CB_COLOR1_FMASK_SLICE
+    0x00000000, // CB_COLOR1_CLEAR_WORD0
+    0x00000000, // CB_COLOR1_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR2_BASE
+    0x00000000, // CB_COLOR2_PITCH
+    0x00000000, // CB_COLOR2_SLICE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR2_CMASK
+    0x00000000, // CB_COLOR2_CMASK_SLICE
+    0x00000000, // CB_COLOR2_FMASK
+    0x00000000, // CB_COLOR2_FMASK_SLICE
+    0x00000000, // CB_COLOR2_CLEAR_WORD0
+    0x00000000, // CB_COLOR2_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR3_BASE
+    0x00000000, // CB_COLOR3_PITCH
+    0x00000000, // CB_COLOR3_SLICE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR3_CMASK
+    0x00000000, // CB_COLOR3_CMASK_SLICE
+    0x00000000, // CB_COLOR3_FMASK
+    0x00000000, // CB_COLOR3_FMASK_SLICE
+    0x00000000, // CB_COLOR3_CLEAR_WORD0
+    0x00000000, // CB_COLOR3_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR4_BASE
+    0x00000000, // CB_COLOR4_PITCH
+    0x00000000, // CB_COLOR4_SLICE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR4_CMASK
+    0x00000000, // CB_COLOR4_CMASK_SLICE
+    0x00000000, // CB_COLOR4_FMASK
+    0x00000000, // CB_COLOR4_FMASK_SLICE
+    0x00000000, // CB_COLOR4_CLEAR_WORD0
+    0x00000000, // CB_COLOR4_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR5_BASE
+    0x00000000, // CB_COLOR5_PITCH
+    0x00000000, // CB_COLOR5_SLICE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR5_CMASK
+    0x00000000, // CB_COLOR5_CMASK_SLICE
+    0x00000000, // CB_COLOR5_FMASK
+    0x00000000, // CB_COLOR5_FMASK_SLICE
+    0x00000000, // CB_COLOR5_CLEAR_WORD0
+    0x00000000, // CB_COLOR5_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR6_BASE
+    0x00000000, // CB_COLOR6_PITCH
+    0x00000000, // CB_COLOR6_SLICE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR6_CMASK
+    0x00000000, // CB_COLOR6_CMASK_SLICE
+    0x00000000, // CB_COLOR6_FMASK
+    0x00000000, // CB_COLOR6_FMASK_SLICE
+    0x00000000, // CB_COLOR6_CLEAR_WORD0
+    0x00000000, // CB_COLOR6_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR7_BASE
+    0x00000000, // CB_COLOR7_PITCH
+    0x00000000, // CB_COLOR7_SLICE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR7_CMASK
+    0x00000000, // CB_COLOR7_CMASK_SLICE
+    0x00000000, // CB_COLOR7_FMASK
+    0x00000000, // CB_COLOR7_FMASK_SLICE
+    0x00000000, // CB_COLOR7_CLEAR_WORD0
+    0x00000000, // CB_COLOR7_CLEAR_WORD1
+};
+static const struct cs_extent_def si_SECT_CONTEXT_defs[] =
+{
+    {si_SECT_CONTEXT_def_1, 0x0000a000, 212 },
+    {si_SECT_CONTEXT_def_2, 0x0000a0d8, 272 },
+    {si_SECT_CONTEXT_def_3, 0x0000a1f5, 6 },
+    {si_SECT_CONTEXT_def_4, 0x0000a200, 157 },
+    {si_SECT_CONTEXT_def_5, 0x0000a2a1, 1 },
+    {si_SECT_CONTEXT_def_6, 0x0000a2a3, 1 },
+    {si_SECT_CONTEXT_def_7, 0x0000a2a5, 233 },
+    { 0, 0, 0 }
+};
+static const struct cs_section_def si_cs_data[] = {
+    { si_SECT_CONTEXT_defs, SECT_CONTEXT },
+    { 0, SECT_NONE }
+};
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c
new file mode 100644
index 0000000..9bcdd17
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cypress_dpm.c
@@ -0,0 +1,2189 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "evergreend.h"
+#include "r600_dpm.h"
+#include "cypress_dpm.h"
+#include "atom.h"
+
+#define SMC_RAM_END 0x8000
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define MC_CG_SEQ_DRAMCONF_S0       0x05
+#define MC_CG_SEQ_DRAMCONF_S1       0x06
+#define MC_CG_SEQ_YCLK_SUSPEND      0x04
+#define MC_CG_SEQ_YCLK_RESUME       0x0a
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps);
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+
+static void cypress_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+						 bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp, bif;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if (enable) {
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+				tmp |= LC_GEN2_EN_STRAP;
+
+				tmp |= LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+				udelay(10);
+				tmp &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+			}
+		}
+	} else {
+		if (!pi->boot_in_gen2) {
+			tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+			tmp &= ~LC_GEN2_EN_STRAP;
+		}
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+	}
+}
+
+static void cypress_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					     bool enable)
+{
+	cypress_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+#if 0
+static int cypress_enter_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->gfx_clock_gating) {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+
+		RREG32(GB_ADDR_CONFIG);
+	}
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_SwitchToMinimumPower),
+		 ~HOST_SMC_MSG_MASK);
+
+	udelay(7000);
+
+	return 0;
+}
+#endif
+
+static void cypress_gfx_clock_gating_enable(struct radeon_device *rdev,
+					    bool enable)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (enable) {
+		if (eg_pi->light_sleep) {
+			WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+			WREG32_CG(CG_CGLS_TILE_0, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_1, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_2, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_3, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_4, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_5, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_6, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_7, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_8, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_9, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_10, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_11, 0xFFFFFFFF);
+
+			WREG32_P(SCLK_PWRMGT_CNTL, DYN_LIGHT_SLEEP_EN, ~DYN_LIGHT_SLEEP_EN);
+		}
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+
+		if (eg_pi->light_sleep) {
+			WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_LIGHT_SLEEP_EN);
+
+			WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+			WREG32_CG(CG_CGLS_TILE_0, 0);
+			WREG32_CG(CG_CGLS_TILE_1, 0);
+			WREG32_CG(CG_CGLS_TILE_2, 0);
+			WREG32_CG(CG_CGLS_TILE_3, 0);
+			WREG32_CG(CG_CGLS_TILE_4, 0);
+			WREG32_CG(CG_CGLS_TILE_5, 0);
+			WREG32_CG(CG_CGLS_TILE_6, 0);
+			WREG32_CG(CG_CGLS_TILE_7, 0);
+			WREG32_CG(CG_CGLS_TILE_8, 0);
+			WREG32_CG(CG_CGLS_TILE_9, 0);
+			WREG32_CG(CG_CGLS_TILE_10, 0);
+			WREG32_CG(CG_CGLS_TILE_11, 0);
+		}
+	}
+}
+
+static void cypress_mg_clock_gating_enable(struct radeon_device *rdev,
+					   bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (enable) {
+		u32 cgts_sm_ctrl_reg;
+
+		if (rdev->family == CHIP_CEDAR)
+			cgts_sm_ctrl_reg = CEDAR_MGCGCGTSSMCTRL_DFLT;
+		else if (rdev->family == CHIP_REDWOOD)
+			cgts_sm_ctrl_reg = REDWOOD_MGCGCGTSSMCTRL_DFLT;
+		else
+			cgts_sm_ctrl_reg = CYPRESS_MGCGCGTSSMCTRL_DFLT;
+
+		WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+		WREG32_CG(CG_CGTT_LOCAL_0, CYPRESS_MGCGTTLOCAL0_DFLT);
+		WREG32_CG(CG_CGTT_LOCAL_1, CYPRESS_MGCGTTLOCAL1_DFLT & 0xFFFFCFFF);
+		WREG32_CG(CG_CGTT_LOCAL_2, CYPRESS_MGCGTTLOCAL2_DFLT);
+		WREG32_CG(CG_CGTT_LOCAL_3, CYPRESS_MGCGTTLOCAL3_DFLT);
+
+		if (pi->mgcgtssm)
+			WREG32(CGTS_SM_CTRL_REG, cgts_sm_ctrl_reg);
+
+		if (eg_pi->mcls) {
+			WREG32_P(MC_CITF_MISC_RD_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_CITF_MISC_WR_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_CITF_MISC_VM_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_HUB_MISC_HUB_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_HUB_MISC_VM_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_HUB_MISC_SIP_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_XPB_CLK_GAT, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(VM_L2_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+		}
+	} else {
+		WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+		WREG32_CG(CG_CGTT_LOCAL_0, 0xFFFFFFFF);
+		WREG32_CG(CG_CGTT_LOCAL_1, 0xFFFFFFFF);
+		WREG32_CG(CG_CGTT_LOCAL_2, 0xFFFFFFFF);
+		WREG32_CG(CG_CGTT_LOCAL_3, 0xFFFFFFFF);
+
+		if (pi->mgcgtssm)
+			WREG32(CGTS_SM_CTRL_REG, 0x81f44bc0);
+	}
+}
+
+void cypress_enable_spread_spectrum(struct radeon_device *rdev,
+				    bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (pi->sclk_ss)
+			WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+
+		if (pi->mclk_ss)
+			WREG32_P(MPLL_CNTL_MODE, SS_SSEN, ~SS_SSEN);
+	} else {
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+		WREG32_P(MPLL_CNTL_MODE, 0, ~SS_SSEN);
+		WREG32_P(MPLL_CNTL_MODE, 0, ~SS_DSMODE_EN);
+	}
+}
+
+void cypress_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+void cypress_enable_sclk_control(struct radeon_device *rdev,
+				 bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+}
+
+void cypress_enable_mclk_control(struct radeon_device *rdev,
+				 bool enable)
+{
+	if (enable)
+		WREG32_P(MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+	else
+		WREG32_P(MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+int cypress_notify_smc_display_change(struct radeon_device *rdev,
+				      bool has_display)
+{
+	PPSMC_Msg msg = has_display ?
+		(PPSMC_Msg)PPSMC_MSG_HasDisplay : (PPSMC_Msg)PPSMC_MSG_NoDisplay;
+
+	if (rv770_send_msg_to_smc(rdev, msg) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+void cypress_program_response_times(struct radeon_device *rdev)
+{
+	u32 reference_clock;
+	u32 mclk_switch_limit;
+
+	reference_clock = radeon_get_xclk(rdev);
+	mclk_switch_limit = (460 * reference_clock) / 100;
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mclk_switch_lim,
+				      mclk_switch_limit);
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mvdd_chg_time, 1);
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mc_block_delay, 0xAA);
+
+	rv770_program_response_times(rdev);
+
+	if (ASIC_IS_LOMBOK(rdev))
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_is_asic_lombok, 1);
+
+}
+
+static int cypress_pcie_performance_request(struct radeon_device *rdev,
+					    u8 perf_req, bool advertise)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 tmp;
+
+	udelay(10);
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) && (tmp & LC_CURRENT_DATA_RATE))
+		return 0;
+
+#if defined(CONFIG_ACPI)
+	if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) ||
+	    (perf_req == PCIE_PERF_REQ_PECI_GEN2)) {
+		eg_pi->pcie_performance_request_registered = true;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	} else if ((perf_req == PCIE_PERF_REQ_REMOVE_REGISTRY) &&
+		   eg_pi->pcie_performance_request_registered) {
+		eg_pi->pcie_performance_request_registered = false;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	}
+#endif
+
+	return 0;
+}
+
+void cypress_advertise_gen2_capability(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+#if defined(CONFIG_ACPI)
+	radeon_acpi_pcie_notify_device_ready(rdev);
+#endif
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+		pi->pcie_gen2 = true;
+	else
+		pi->pcie_gen2 = false;
+
+	if (!pi->pcie_gen2)
+		cypress_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, true);
+
+}
+
+static enum radeon_pcie_gen cypress_get_maximum_link_speed(struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+
+	if (state->high.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		return 1;
+	return 0;
+}
+
+void cypress_notify_link_speed_change_after_state_change(struct radeon_device *rdev,
+							 struct radeon_ps *radeon_new_state,
+							 struct radeon_ps *radeon_current_state)
+{
+	enum radeon_pcie_gen pcie_link_speed_target =
+		cypress_get_maximum_link_speed(radeon_new_state);
+	enum radeon_pcie_gen pcie_link_speed_current =
+		cypress_get_maximum_link_speed(radeon_current_state);
+	u8 request;
+
+	if (pcie_link_speed_target < pcie_link_speed_current) {
+		if (pcie_link_speed_target == RADEON_PCIE_GEN1)
+			request = PCIE_PERF_REQ_PECI_GEN1;
+		else if (pcie_link_speed_target == RADEON_PCIE_GEN2)
+			request = PCIE_PERF_REQ_PECI_GEN2;
+		else
+			request = PCIE_PERF_REQ_PECI_GEN3;
+
+		cypress_pcie_performance_request(rdev, request, false);
+	}
+}
+
+void cypress_notify_link_speed_change_before_state_change(struct radeon_device *rdev,
+							  struct radeon_ps *radeon_new_state,
+							  struct radeon_ps *radeon_current_state)
+{
+	enum radeon_pcie_gen pcie_link_speed_target =
+		cypress_get_maximum_link_speed(radeon_new_state);
+	enum radeon_pcie_gen pcie_link_speed_current =
+		cypress_get_maximum_link_speed(radeon_current_state);
+	u8 request;
+
+	if (pcie_link_speed_target > pcie_link_speed_current) {
+		if (pcie_link_speed_target == RADEON_PCIE_GEN1)
+			request = PCIE_PERF_REQ_PECI_GEN1;
+		else if (pcie_link_speed_target == RADEON_PCIE_GEN2)
+			request = PCIE_PERF_REQ_PECI_GEN2;
+		else
+			request = PCIE_PERF_REQ_PECI_GEN3;
+
+		cypress_pcie_performance_request(rdev, request, false);
+	}
+}
+
+static int cypress_populate_voltage_value(struct radeon_device *rdev,
+					  struct atom_voltage_table *table,
+					  u16 value, RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (value <= table->entries[i].value) {
+			voltage->index = (u8)i;
+			voltage->value = cpu_to_be16(table->entries[i].value);
+			break;
+		}
+	}
+
+	if (i == table->count)
+		return -EINVAL;
+
+	return 0;
+}
+
+u8 cypress_get_strobe_mode_settings(struct radeon_device *rdev, u32 mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 result = 0;
+	bool strobe_mode = false;
+
+	if (pi->mem_gddr5) {
+		if (mclk <= pi->mclk_strobe_mode_threshold)
+			strobe_mode = true;
+		result = cypress_get_mclk_frequency_ratio(rdev, mclk, strobe_mode);
+
+		if (strobe_mode)
+			result |= SMC_STROBE_ENABLE;
+	}
+
+	return result;
+}
+
+u32 cypress_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf)
+{
+	u32 ref_clk = rdev->clock.mpll.reference_freq;
+	u32 vco = clkf * ref_clk;
+
+	/* 100 Mhz ref clk */
+	if (ref_clk == 10000) {
+		if (vco > 500000)
+			return 0xC6;
+		if (vco > 400000)
+			return 0x9D;
+		if (vco > 330000)
+			return 0x6C;
+		if (vco > 250000)
+			return 0x2B;
+		if (vco >  160000)
+			return 0x5B;
+		if (vco > 120000)
+			return 0x0A;
+		return 0x4B;
+	}
+
+	/* 27 Mhz ref clk */
+	if (vco > 250000)
+		return 0x8B;
+	if (vco > 200000)
+		return 0xCC;
+	if (vco > 150000)
+		return 0x9B;
+	return 0x6B;
+}
+
+static int cypress_populate_mclk_value(struct radeon_device *rdev,
+				       u32 engine_clock, u32 memory_clock,
+				       RV7XX_SMC_MCLK_VALUE *mclk,
+				       bool strobe_mode, bool dll_state_on)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl =
+		pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl =
+		pi->clk_regs.rv770.dll_cntl;
+	u32 mpll_ss1 = pi->clk_regs.rv770.mpll_ss1;
+	u32 mpll_ss2 = pi->clk_regs.rv770.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 ibias;
+	u32 dll_speed;
+	int ret;
+	u32 mc_seq_misc7;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, strobe_mode, &dividers);
+	if (ret)
+		return ret;
+
+	if (!strobe_mode) {
+		mc_seq_misc7 = RREG32(MC_SEQ_MISC7);
+
+		if(mc_seq_misc7 & 0x8000000)
+			dividers.post_div = 1;
+	}
+
+	ibias = cypress_map_clkf_to_ibias(rdev, dividers.whole_fb_div);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(dividers.ref_div);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+	mpll_ad_func_cntl |= CLKF(dividers.whole_fb_div);
+	mpll_ad_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(dividers.ref_div);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+		mpll_dq_func_cntl |= CLKF(dividers.whole_fb_div);
+		mpll_dq_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (strobe_mode)
+			mpll_dq_func_cntl &= ~PDNB;
+		else
+			mpll_dq_func_cntl |= PDNB;
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 decoded_ref = rv740_get_decoded_reference_divider(dividers.ref_div);
+			u32 clk_s = reference_clock * 5 / (decoded_ref * ss.rate);
+			u32 clk_v = ss.percentage *
+				(0x4000 * dividers.whole_fb_div + 0x800 * dividers.frac_fb_div) / (clk_s * 625);
+
+			mpll_ss1 &= ~CLKV_MASK;
+			mpll_ss1 |= CLKV(clk_v);
+
+			mpll_ss2 &= ~CLKS_MASK;
+			mpll_ss2 |= CLKS(clk_s);
+		}
+	}
+
+	dll_speed = rv740_get_dll_speed(pi->mem_gddr5,
+					memory_clock);
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(dll_speed);
+	if (dll_state_on)
+		mclk_pwrmgt_cntl |= (MRDCKA0_PDNB |
+				     MRDCKA1_PDNB |
+				     MRDCKB0_PDNB |
+				     MRDCKB1_PDNB |
+				     MRDCKC0_PDNB |
+				     MRDCKC1_PDNB |
+				     MRDCKD0_PDNB |
+				     MRDCKD1_PDNB);
+	else
+		mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+				      MRDCKA1_PDNB |
+				      MRDCKB0_PDNB |
+				      MRDCKB1_PDNB |
+				      MRDCKC0_PDNB |
+				      MRDCKC1_PDNB |
+				      MRDCKD0_PDNB |
+				      MRDCKD1_PDNB);
+
+	mclk->mclk770.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->mclk770.vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->mclk770.vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+u8 cypress_get_mclk_frequency_ratio(struct radeon_device *rdev,
+				    u32 memory_clock, bool strobe_mode)
+{
+	u8 mc_para_index;
+
+	if (rdev->family >= CHIP_BARTS) {
+		if (strobe_mode) {
+			if (memory_clock < 10000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 47500)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 10000) / 2500);
+		} else {
+			if (memory_clock < 65000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 135000)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 60000) / 5000);
+		}
+	} else {
+		if (strobe_mode) {
+			if (memory_clock < 10000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 47500)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 10000) / 2500);
+		} else {
+			if (memory_clock < 40000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 115000)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 40000) / 5000);
+		}
+	}
+	return mc_para_index;
+}
+
+static int cypress_populate_mvdd_value(struct radeon_device *rdev,
+				       u32 mclk,
+				       RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (!pi->mvdd_control) {
+		voltage->index = eg_pi->mvdd_high_index;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+		return 0;
+	}
+
+	if (mclk <= pi->mvdd_split_frequency) {
+		voltage->index = eg_pi->mvdd_low_index;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = eg_pi->mvdd_high_index;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+
+	return 0;
+}
+
+int cypress_convert_power_level_to_smc(struct radeon_device *rdev,
+				       struct rv7xx_pl *pl,
+				       RV770_SMC_HW_PERFORMANCE_LEVEL *level,
+				       u8 watermark_level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int ret;
+	bool dll_state_on;
+
+	level->gen2PCIE = pi->pcie_gen2 ?
+		((pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0) : 0;
+	level->gen2XSP  = (pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0;
+	level->backbias = (pl->flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? 1 : 0;
+	level->displayWatermark = watermark_level;
+
+	ret = rv740_populate_sclk_value(rdev, pl->sclk, &level->sclk);
+	if (ret)
+		return ret;
+
+	level->mcFlags =  0;
+	if (pi->mclk_stutter_mode_threshold &&
+	    (pl->mclk <= pi->mclk_stutter_mode_threshold) &&
+	    !eg_pi->uvd_enabled) {
+		level->mcFlags |= SMC_MC_STUTTER_EN;
+		if (eg_pi->sclk_deep_sleep)
+			level->stateFlags |= PPSMC_STATEFLAG_AUTO_PULSE_SKIP;
+		else
+			level->stateFlags &= ~PPSMC_STATEFLAG_AUTO_PULSE_SKIP;
+	}
+
+	if (pi->mem_gddr5) {
+		if (pl->mclk > pi->mclk_edc_enable_threshold)
+			level->mcFlags |= SMC_MC_EDC_RD_FLAG;
+
+		if (pl->mclk > eg_pi->mclk_edc_wr_enable_threshold)
+			level->mcFlags |= SMC_MC_EDC_WR_FLAG;
+
+		level->strobeMode = cypress_get_strobe_mode_settings(rdev, pl->mclk);
+
+		if (level->strobeMode & SMC_STROBE_ENABLE) {
+			if (cypress_get_mclk_frequency_ratio(rdev, pl->mclk, true) >=
+			    ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf))
+				dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+			else
+				dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false;
+		} else
+			dll_state_on = eg_pi->dll_default_on;
+
+		ret = cypress_populate_mclk_value(rdev,
+						  pl->sclk,
+						  pl->mclk,
+						  &level->mclk,
+						  (level->strobeMode & SMC_STROBE_ENABLE) != 0,
+						  dll_state_on);
+	} else {
+		ret = cypress_populate_mclk_value(rdev,
+						  pl->sclk,
+						  pl->mclk,
+						  &level->mclk,
+						  true,
+						  true);
+	}
+	if (ret)
+		return ret;
+
+	ret = cypress_populate_voltage_value(rdev,
+					     &eg_pi->vddc_voltage_table,
+					     pl->vddc,
+					     &level->vddc);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddci_control) {
+		ret = cypress_populate_voltage_value(rdev,
+						     &eg_pi->vddci_voltage_table,
+						     pl->vddci,
+						     &level->vddci);
+		if (ret)
+			return ret;
+	}
+
+	ret = cypress_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int cypress_convert_power_state_to_smc(struct radeon_device *rdev,
+					      struct radeon_ps *radeon_state,
+					      RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int ret;
+
+	if (!(radeon_state->caps & ATOM_PPLIB_DISALLOW_ON_DC))
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	ret = cypress_convert_power_level_to_smc(rdev,
+						 &state->low,
+						 &smc_state->levels[0],
+						 PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = cypress_convert_power_level_to_smc(rdev,
+						 &state->medium,
+						 &smc_state->levels[1],
+						 PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = cypress_convert_power_level_to_smc(rdev,
+						 &state->high,
+						 &smc_state->levels[2],
+						 PPSMC_DISPLAY_WATERMARK_HIGH);
+	if (ret)
+		return ret;
+
+	smc_state->levels[0].arbValue = MC_CG_ARB_FREQ_F1;
+	smc_state->levels[1].arbValue = MC_CG_ARB_FREQ_F2;
+	smc_state->levels[2].arbValue = MC_CG_ARB_FREQ_F3;
+
+	if (eg_pi->dynamic_ac_timing) {
+		smc_state->levels[0].ACIndex = 2;
+		smc_state->levels[1].ACIndex = 3;
+		smc_state->levels[2].ACIndex = 4;
+	} else {
+		smc_state->levels[0].ACIndex = 0;
+		smc_state->levels[1].ACIndex = 0;
+		smc_state->levels[2].ACIndex = 0;
+	}
+
+	rv770_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	return rv770_populate_smc_t(rdev, radeon_state, smc_state);
+}
+
+static void cypress_convert_mc_registers(struct evergreen_mc_reg_entry *entry,
+					 SMC_Evergreen_MCRegisterSet *data,
+					 u32 num_entries, u32 valid_flag)
+{
+	u32 i, j;
+
+	for (i = 0, j = 0; j < num_entries; j++) {
+		if (valid_flag & (1 << j)) {
+			data->value[i] = cpu_to_be32(entry->mc_data[j]);
+			i++;
+		}
+	}
+}
+
+static void cypress_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev,
+						      struct rv7xx_pl *pl,
+						      SMC_Evergreen_MCRegisterSet *mc_reg_table_data)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < eg_pi->mc_reg_table.num_entries; i++) {
+		if (pl->mclk <=
+		    eg_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max)
+			break;
+	}
+
+	if ((i == eg_pi->mc_reg_table.num_entries) && (i > 0))
+		--i;
+
+	cypress_convert_mc_registers(&eg_pi->mc_reg_table.mc_reg_table_entry[i],
+				     mc_reg_table_data,
+				     eg_pi->mc_reg_table.last,
+				     eg_pi->mc_reg_table.valid_flag);
+}
+
+static void cypress_convert_mc_reg_table_to_smc(struct radeon_device *rdev,
+						struct radeon_ps *radeon_state,
+						SMC_Evergreen_MCRegisters *mc_reg_table)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &state->low,
+						  &mc_reg_table->data[2]);
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &state->medium,
+						  &mc_reg_table->data[3]);
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &state->high,
+						  &mc_reg_table->data[4]);
+}
+
+int cypress_upload_sw_state(struct radeon_device *rdev,
+			    struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 address = pi->state_table_start +
+		offsetof(RV770_SMC_STATETABLE, driverState);
+	RV770_SMC_SWSTATE state = { 0 };
+	int ret;
+
+	ret = cypress_convert_power_state_to_smc(rdev, radeon_new_state, &state);
+	if (ret)
+		return ret;
+
+	return rv770_copy_bytes_to_smc(rdev, address, (u8 *)&state,
+				    sizeof(RV770_SMC_SWSTATE),
+				    pi->sram_end);
+}
+
+int cypress_upload_mc_reg_table(struct radeon_device *rdev,
+				struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	SMC_Evergreen_MCRegisters mc_reg_table = { 0 };
+	u16 address;
+
+	cypress_convert_mc_reg_table_to_smc(rdev, radeon_new_state, &mc_reg_table);
+
+	address = eg_pi->mc_reg_table_start +
+		(u16)offsetof(SMC_Evergreen_MCRegisters, data[2]);
+
+	return rv770_copy_bytes_to_smc(rdev, address,
+				       (u8 *)&mc_reg_table.data[2],
+				       sizeof(SMC_Evergreen_MCRegisterSet) * 3,
+				       pi->sram_end);
+}
+
+u32 cypress_calculate_burst_time(struct radeon_device *rdev,
+				 u32 engine_clock, u32 memory_clock)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 multiplier = pi->mem_gddr5 ? 1 : 2;
+	u32 result = (4 * multiplier * engine_clock) / (memory_clock / 2);
+	u32 burst_time;
+
+	if (result <= 4)
+		burst_time = 0;
+	else if (result < 8)
+		burst_time = result - 4;
+	else {
+		burst_time = result / 2 ;
+		if (burst_time > 18)
+			burst_time = 18;
+	}
+
+	return burst_time;
+}
+
+void cypress_program_memory_timing_parameters(struct radeon_device *rdev,
+					      struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_ps *new_state = rv770_get_ps(radeon_new_state);
+	u32 mc_arb_burst_time = RREG32(MC_ARB_BURST_TIME);
+
+	mc_arb_burst_time &= ~(STATE1_MASK | STATE2_MASK | STATE3_MASK);
+
+	mc_arb_burst_time |= STATE1(cypress_calculate_burst_time(rdev,
+								 new_state->low.sclk,
+								 new_state->low.mclk));
+	mc_arb_burst_time |= STATE2(cypress_calculate_burst_time(rdev,
+								 new_state->medium.sclk,
+								 new_state->medium.mclk));
+	mc_arb_burst_time |= STATE3(cypress_calculate_burst_time(rdev,
+								 new_state->high.sclk,
+								 new_state->high.mclk));
+
+	rv730_program_memory_timing_parameters(rdev, radeon_new_state);
+
+	WREG32(MC_ARB_BURST_TIME, mc_arb_burst_time);
+}
+
+static void cypress_populate_mc_reg_addresses(struct radeon_device *rdev,
+					      SMC_Evergreen_MCRegisters *mc_reg_table)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0, j = 0; j < eg_pi->mc_reg_table.last; j++) {
+		if (eg_pi->mc_reg_table.valid_flag & (1 << j)) {
+			mc_reg_table->address[i].s0 =
+				cpu_to_be16(eg_pi->mc_reg_table.mc_reg_address[j].s0);
+			mc_reg_table->address[i].s1 =
+				cpu_to_be16(eg_pi->mc_reg_table.mc_reg_address[j].s1);
+			i++;
+		}
+	}
+
+	mc_reg_table->last = (u8)i;
+}
+
+static void cypress_set_mc_reg_address_table(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i = 0;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RAS_TIMING_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RAS_TIMING >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_CAS_TIMING_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_CAS_TIMING >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC_TIMING_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC_TIMING >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC_TIMING2_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC_TIMING2 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RD_CTL_D0_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RD_CTL_D0 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RD_CTL_D1_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RD_CTL_D1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_WR_CTL_D0_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_WR_CTL_D0 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_WR_CTL_D1_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_WR_CTL_D1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_PMG_CMD_EMRS >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_PMG_CMD_MRS >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_PMG_CMD_MRS1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC1 >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RESERVE_M >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RESERVE_M >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC3 >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC3 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.last = (u8)i;
+}
+
+static void cypress_retrieve_ac_timing_for_one_entry(struct radeon_device *rdev,
+						     struct evergreen_mc_reg_entry *entry)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i;
+
+	for (i = 0; i < eg_pi->mc_reg_table.last; i++)
+		entry->mc_data[i] =
+			RREG32(eg_pi->mc_reg_table.mc_reg_address[i].s1 << 2);
+
+}
+
+static void cypress_retrieve_ac_timing_for_all_ranges(struct radeon_device *rdev,
+						      struct atom_memory_clock_range_table *range_table)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0; i < range_table->num_entries; i++) {
+		eg_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max =
+			range_table->mclk[i];
+		radeon_atom_set_ac_timing(rdev, range_table->mclk[i]);
+		cypress_retrieve_ac_timing_for_one_entry(rdev,
+							 &eg_pi->mc_reg_table.mc_reg_table_entry[i]);
+	}
+
+	eg_pi->mc_reg_table.num_entries = range_table->num_entries;
+	eg_pi->mc_reg_table.valid_flag = 0;
+
+	for (i = 0; i < eg_pi->mc_reg_table.last; i++) {
+		for (j = 1; j < range_table->num_entries; j++) {
+			if (eg_pi->mc_reg_table.mc_reg_table_entry[j-1].mc_data[i] !=
+			    eg_pi->mc_reg_table.mc_reg_table_entry[j].mc_data[i]) {
+				eg_pi->mc_reg_table.valid_flag |= (1 << i);
+				break;
+			}
+		}
+	}
+}
+
+static int cypress_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 module_index = rv770_get_memory_module_index(rdev);
+	struct atom_memory_clock_range_table range_table = { 0 };
+	int ret;
+
+	ret = radeon_atom_get_mclk_range_table(rdev,
+					       pi->mem_gddr5,
+					       module_index, &range_table);
+	if (ret)
+		return ret;
+
+	cypress_retrieve_ac_timing_for_all_ranges(rdev, &range_table);
+
+	return 0;
+}
+
+static void cypress_wait_for_mc_sequencer(struct radeon_device *rdev, u8 value)
+{
+	u32 i, j;
+	u32 channels = 2;
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK))
+		channels = 4;
+	else if (rdev->family == CHIP_CEDAR)
+		channels = 1;
+
+	for (i = 0; i < channels; i++) {
+		if ((rdev->family == CHIP_CYPRESS) ||
+		    (rdev->family == CHIP_HEMLOCK)) {
+			WREG32_P(MC_CONFIG_MCD, MC_RD_ENABLE_MCD(i), ~MC_RD_ENABLE_MCD_MASK);
+			WREG32_P(MC_CG_CONFIG_MCD, MC_RD_ENABLE_MCD(i), ~MC_RD_ENABLE_MCD_MASK);
+		} else {
+			WREG32_P(MC_CONFIG, MC_RD_ENABLE(i), ~MC_RD_ENABLE_MASK);
+			WREG32_P(MC_CG_CONFIG, MC_RD_ENABLE(i), ~MC_RD_ENABLE_MASK);
+		}
+		for (j = 0; j < rdev->usec_timeout; j++) {
+			if (((RREG32(MC_SEQ_CG) & CG_SEQ_RESP_MASK) >> CG_SEQ_RESP_SHIFT) == value)
+				break;
+			udelay(1);
+		}
+	}
+}
+
+static void cypress_force_mc_use_s1(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	u32 strobe_mode;
+	u32 mc_seq_cg;
+	int i;
+
+	if (RREG32(MC_SEQ_STATUS_M) & PMG_PWRSTATE)
+		return;
+
+	radeon_atom_set_ac_timing(rdev, boot_state->low.mclk);
+	radeon_mc_wait_for_idle(rdev);
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK)) {
+		WREG32(MC_CONFIG_MCD, 0xf);
+		WREG32(MC_CG_CONFIG_MCD, 0xf);
+	} else {
+		WREG32(MC_CONFIG, 0xf);
+		WREG32(MC_CG_CONFIG, 0xf);
+	}
+
+	for (i = 0; i < rdev->num_crtc; i++)
+		radeon_wait_for_vblank(rdev, i);
+
+	WREG32(MC_SEQ_CG, MC_CG_SEQ_YCLK_SUSPEND);
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_SUSPEND);
+
+	strobe_mode = cypress_get_strobe_mode_settings(rdev,
+						       boot_state->low.mclk);
+
+	mc_seq_cg = CG_SEQ_REQ(MC_CG_SEQ_DRAMCONF_S1);
+	mc_seq_cg |= SEQ_CG_RESP(strobe_mode);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(MC_SEQ_STATUS_M) & PMG_PWRSTATE)
+			break;
+		udelay(1);
+	}
+
+	mc_seq_cg &= ~CG_SEQ_REQ_MASK;
+	mc_seq_cg |= CG_SEQ_REQ(MC_CG_SEQ_YCLK_RESUME);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_RESUME);
+}
+
+static void cypress_copy_ac_timing_from_s1_to_s0(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 value;
+	u32 i;
+
+	for (i = 0; i < eg_pi->mc_reg_table.last; i++) {
+		value = RREG32(eg_pi->mc_reg_table.mc_reg_address[i].s1 << 2);
+		WREG32(eg_pi->mc_reg_table.mc_reg_address[i].s0 << 2, value);
+	}
+}
+
+static void cypress_force_mc_use_s0(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	u32 strobe_mode;
+	u32 mc_seq_cg;
+	int i;
+
+	cypress_copy_ac_timing_from_s1_to_s0(rdev);
+	radeon_mc_wait_for_idle(rdev);
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK)) {
+		WREG32(MC_CONFIG_MCD, 0xf);
+		WREG32(MC_CG_CONFIG_MCD, 0xf);
+	} else {
+		WREG32(MC_CONFIG, 0xf);
+		WREG32(MC_CG_CONFIG, 0xf);
+	}
+
+	for (i = 0; i < rdev->num_crtc; i++)
+		radeon_wait_for_vblank(rdev, i);
+
+	WREG32(MC_SEQ_CG, MC_CG_SEQ_YCLK_SUSPEND);
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_SUSPEND);
+
+	strobe_mode = cypress_get_strobe_mode_settings(rdev,
+						       boot_state->low.mclk);
+
+	mc_seq_cg = CG_SEQ_REQ(MC_CG_SEQ_DRAMCONF_S0);
+	mc_seq_cg |= SEQ_CG_RESP(strobe_mode);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (!(RREG32(MC_SEQ_STATUS_M) & PMG_PWRSTATE))
+			break;
+		udelay(1);
+	}
+
+	mc_seq_cg &= ~CG_SEQ_REQ_MASK;
+	mc_seq_cg |= CG_SEQ_REQ(MC_CG_SEQ_YCLK_RESUME);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_RESUME);
+}
+
+static int cypress_populate_initial_mvdd_value(struct radeon_device *rdev,
+					       RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	voltage->index = eg_pi->mvdd_high_index;
+	voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+
+	return 0;
+}
+
+int cypress_populate_smc_initial_state(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_initial_state,
+				       RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_ps *initial_state = rv770_get_ps(radeon_initial_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 a_t;
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.mclk770.vDLL_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.dll_cntl);
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss1);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk770.mclk_value =
+		cpu_to_be32(initial_state->low.mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->low.sclk);
+
+	table->initialState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+
+	table->initialState.levels[0].ACIndex = 0;
+
+	cypress_populate_voltage_value(rdev,
+				       &eg_pi->vddc_voltage_table,
+				       initial_state->low.vddc,
+				       &table->initialState.levels[0].vddc);
+
+	if (eg_pi->vddci_control)
+		cypress_populate_voltage_value(rdev,
+					       &eg_pi->vddci_voltage_table,
+					       initial_state->low.vddci,
+					       &table->initialState.levels[0].vddci);
+
+	cypress_populate_initial_mvdd_value(rdev,
+					    &table->initialState.levels[0].mvdd);
+
+	a_t = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(a_t);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+	if (initial_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		table->initialState.levels[0].gen2XSP = 1;
+	else
+		table->initialState.levels[0].gen2XSP = 0;
+
+	if (pi->mem_gddr5) {
+		table->initialState.levels[0].strobeMode =
+			cypress_get_strobe_mode_settings(rdev,
+							 initial_state->low.mclk);
+
+		if (initial_state->low.mclk > pi->mclk_edc_enable_threshold)
+			table->initialState.levels[0].mcFlags = SMC_MC_EDC_RD_FLAG | SMC_MC_EDC_WR_FLAG;
+		else
+			table->initialState.levels[0].mcFlags =  0;
+	}
+
+	table->initialState.levels[1] = table->initialState.levels[0];
+	table->initialState.levels[2] = table->initialState.levels[0];
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	return 0;
+}
+
+int cypress_populate_smc_acpi_state(struct radeon_device *rdev,
+				    RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl =
+		pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 mclk_pwrmgt_cntl =
+		pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl =
+		pi->clk_regs.rv770.dll_cntl;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		cypress_populate_voltage_value(rdev,
+					       &eg_pi->vddc_voltage_table,
+					       pi->acpi_vddc,
+					       &table->ACPIState.levels[0].vddc);
+		if (pi->pcie_gen2) {
+			if (pi->acpi_pcie_gen2)
+				table->ACPIState.levels[0].gen2PCIE = 1;
+			else
+				table->ACPIState.levels[0].gen2PCIE = 0;
+		} else
+			table->ACPIState.levels[0].gen2PCIE = 0;
+		if (pi->acpi_pcie_gen2)
+			table->ACPIState.levels[0].gen2XSP = 1;
+		else
+			table->ACPIState.levels[0].gen2XSP = 0;
+	} else {
+		cypress_populate_voltage_value(rdev,
+					       &eg_pi->vddc_voltage_table,
+					       pi->min_vddc_in_table,
+					       &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	if (eg_pi->acpi_vddci) {
+		if (eg_pi->vddci_control) {
+			cypress_populate_voltage_value(rdev,
+						       &eg_pi->vddci_voltage_table,
+						       eg_pi->acpi_vddci,
+						       &table->ACPIState.levels[0].vddci);
+		}
+	}
+
+	mpll_ad_func_cntl &= ~PDNB;
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	if (pi->mem_gddr5)
+		mpll_dq_func_cntl &= ~PDNB;
+	mpll_dq_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN | BYPASS;
+
+	mclk_pwrmgt_cntl |= (MRDCKA0_RESET |
+			     MRDCKA1_RESET |
+			     MRDCKB0_RESET |
+			     MRDCKB1_RESET |
+			     MRDCKC0_RESET |
+			     MRDCKC1_RESET |
+			     MRDCKD0_RESET |
+			     MRDCKD1_RESET);
+
+	mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+			      MRDCKA1_PDNB |
+			      MRDCKB0_PDNB |
+			      MRDCKB1_PDNB |
+			      MRDCKC0_PDNB |
+			      MRDCKC1_PDNB |
+			      MRDCKD0_PDNB |
+			      MRDCKD1_PDNB);
+
+	dll_cntl |= (MRDCKA0_BYPASS |
+		     MRDCKA1_BYPASS |
+		     MRDCKB0_BYPASS |
+		     MRDCKB1_BYPASS |
+		     MRDCKC0_BYPASS |
+		     MRDCKC1_BYPASS |
+		     MRDCKD0_BYPASS |
+		     MRDCKD1_BYPASS);
+
+	/* evergreen only */
+	if (rdev->family <= CHIP_HEMLOCK)
+		spll_func_cntl |= SPLL_RESET | SPLL_SLEEP | SPLL_BYPASS_EN;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(mpll_dq_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk770.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	cypress_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	if (eg_pi->dynamic_ac_timing)
+		table->ACPIState.levels[0].ACIndex = 1;
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	return 0;
+}
+
+static void cypress_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev,
+							  struct atom_voltage_table *voltage_table)
+{
+	unsigned int i, diff;
+
+	if (voltage_table->count <= MAX_NO_VREG_STEPS)
+		return;
+
+	diff = voltage_table->count - MAX_NO_VREG_STEPS;
+
+	for (i= 0; i < MAX_NO_VREG_STEPS; i++)
+		voltage_table->entries[i] = voltage_table->entries[i + diff];
+
+	voltage_table->count = MAX_NO_VREG_STEPS;
+}
+
+int cypress_construct_voltage_tables(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int ret;
+
+	ret = radeon_atom_get_voltage_table(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0,
+					    &eg_pi->vddc_voltage_table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddc_voltage_table.count > MAX_NO_VREG_STEPS)
+		cypress_trim_voltage_table_to_fit_state_table(rdev,
+							      &eg_pi->vddc_voltage_table);
+
+	if (eg_pi->vddci_control) {
+		ret = radeon_atom_get_voltage_table(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0,
+						    &eg_pi->vddci_voltage_table);
+		if (ret)
+			return ret;
+
+		if (eg_pi->vddci_voltage_table.count > MAX_NO_VREG_STEPS)
+			cypress_trim_voltage_table_to_fit_state_table(rdev,
+								      &eg_pi->vddci_voltage_table);
+	}
+
+	return 0;
+}
+
+static void cypress_populate_smc_voltage_table(struct radeon_device *rdev,
+					       struct atom_voltage_table *voltage_table,
+					       RV770_SMC_STATETABLE *table)
+{
+	unsigned int i;
+
+	for (i = 0; i < voltage_table->count; i++) {
+		table->highSMIO[i] = 0;
+		table->lowSMIO[i] |= cpu_to_be32(voltage_table->entries[i].smio_low);
+	}
+}
+
+int cypress_populate_smc_voltage_tables(struct radeon_device *rdev,
+					RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	unsigned char i;
+
+	if (eg_pi->vddc_voltage_table.count) {
+		cypress_populate_smc_voltage_table(rdev,
+						   &eg_pi->vddc_voltage_table,
+						   table);
+
+		table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDC] = 0;
+		table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDC] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+			if (pi->max_vddc_in_table <=
+			    eg_pi->vddc_voltage_table.entries[i].value) {
+				table->maxVDDCIndexInPPTable = i;
+				break;
+			}
+		}
+	}
+
+	if (eg_pi->vddci_voltage_table.count) {
+		cypress_populate_smc_voltage_table(rdev,
+						   &eg_pi->vddci_voltage_table,
+						   table);
+
+		table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDCI] = 0;
+		table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDCI] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+	}
+
+	return 0;
+}
+
+static u32 cypress_get_mclk_split_point(struct atom_memory_info *memory_info)
+{
+	if ((memory_info->mem_type == MEM_TYPE_GDDR3) ||
+	    (memory_info->mem_type == MEM_TYPE_DDR3))
+		return 30000;
+
+	return 0;
+}
+
+int cypress_get_mvdd_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u8 module_index;
+	struct atom_memory_info memory_info;
+	u32 tmp = RREG32(GENERAL_PWRMGT);
+
+	if (!(tmp & BACKBIAS_PAD_EN)) {
+		eg_pi->mvdd_high_index = 0;
+		eg_pi->mvdd_low_index = 1;
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	if (tmp & BACKBIAS_VALUE)
+		eg_pi->mvdd_high_index = 1;
+	else
+		eg_pi->mvdd_high_index = 0;
+
+	eg_pi->mvdd_low_index =
+		(eg_pi->mvdd_high_index == 0) ? 1 : 0;
+
+	module_index = rv770_get_memory_module_index(rdev);
+
+	if (radeon_atom_get_memory_info(rdev, module_index, &memory_info)) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	pi->mvdd_split_frequency =
+		cypress_get_mclk_split_point(&memory_info);
+
+	if (pi->mvdd_split_frequency == 0) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	return 0;
+}
+
+static int cypress_init_smc_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	RV770_SMC_STATETABLE *table = &pi->smc_statetable;
+	int ret;
+
+	memset(table, 0, sizeof(RV770_SMC_STATETABLE));
+
+	cypress_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+        case THERMAL_TYPE_EVERGREEN:
+        case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+        case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+        default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	ret = cypress_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	ret = cypress_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	return rv770_copy_bytes_to_smc(rdev,
+				       pi->state_table_start,
+				       (u8 *)table, sizeof(RV770_SMC_STATETABLE),
+				       pi->sram_end);
+}
+
+int cypress_populate_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	SMC_Evergreen_MCRegisters mc_reg_table = { 0 };
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_seq_index, 1);
+
+	cypress_populate_mc_reg_addresses(rdev, &mc_reg_table);
+
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &boot_state->low,
+						  &mc_reg_table.data[0]);
+
+	cypress_convert_mc_registers(&eg_pi->mc_reg_table.mc_reg_table_entry[0],
+				     &mc_reg_table.data[1], eg_pi->mc_reg_table.last,
+				     eg_pi->mc_reg_table.valid_flag);
+
+	cypress_convert_mc_reg_table_to_smc(rdev, radeon_boot_state, &mc_reg_table);
+
+	return rv770_copy_bytes_to_smc(rdev, eg_pi->mc_reg_table_start,
+				       (u8 *)&mc_reg_table, sizeof(SMC_Evergreen_MCRegisters),
+				       pi->sram_end);
+}
+
+int cypress_get_table_locations(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION +
+					EVERGREEN_SMC_FIRMWARE_HEADER_stateTable,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	pi->state_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION +
+					EVERGREEN_SMC_FIRMWARE_HEADER_softRegisters,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	pi->soft_regs_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION +
+					EVERGREEN_SMC_FIRMWARE_HEADER_mcRegisterTable,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	eg_pi->mc_reg_table_start = (u16)tmp;
+
+	return 0;
+}
+
+void cypress_enable_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MASK | DISP2_GAP_MASK);
+	tmp |= (DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE) |
+		DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE));
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK) |
+		DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE));
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void cypress_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp, pipe;
+	int i;
+
+	tmp = RREG32(CG_DISPLAY_GAP_CNTL) & ~(DISP1_GAP_MASK | DISP2_GAP_MASK);
+	if (rdev->pm.dpm.new_active_crtc_count > 0)
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	if (rdev->pm.dpm.new_active_crtc_count > 1)
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+
+	tmp = RREG32(DCCG_DISP_SLOW_SELECT_REG);
+	pipe = (tmp & DCCG_DISP1_SLOW_SELECT_MASK) >> DCCG_DISP1_SLOW_SELECT_SHIFT;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 0) &&
+	    (!(rdev->pm.dpm.new_active_crtcs & (1 << pipe)))) {
+		/* find the first active crtc */
+		for (i = 0; i < rdev->num_crtc; i++) {
+			if (rdev->pm.dpm.new_active_crtcs & (1 << i))
+				break;
+		}
+		if (i == rdev->num_crtc)
+			pipe = 0;
+		else
+			pipe = i;
+
+		tmp &= ~DCCG_DISP1_SLOW_SELECT_MASK;
+		tmp |= DCCG_DISP1_SLOW_SELECT(pipe);
+		WREG32(DCCG_DISP_SLOW_SELECT_REG, tmp);
+	}
+
+	cypress_notify_smc_display_change(rdev, rdev->pm.dpm.new_active_crtc_count > 0);
+}
+
+void cypress_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	rv740_read_clock_registers(rdev);
+	rv770_read_voltage_smio_registers(rdev);
+	rv770_get_max_vddc(rdev);
+	rv770_get_memory_type(rdev);
+
+	if (eg_pi->pcie_performance_request)
+		eg_pi->pcie_performance_request_registered = false;
+
+	if (eg_pi->pcie_performance_request)
+		cypress_advertise_gen2_capability(rdev);
+
+	rv770_get_pcie_gen2_status(rdev);
+
+	rv770_enable_acpi_pm(rdev);
+}
+
+int cypress_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		rv770_restore_cgcg(rdev);
+
+	if (rv770_dpm_enabled(rdev))
+		return -EINVAL;
+
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = cypress_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+
+	if (pi->mvdd_control) {
+		ret = cypress_get_mvdd_configuration(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_get_mvdd_configuration failed\n");
+			return ret;
+		}
+	}
+
+	if (eg_pi->dynamic_ac_timing) {
+		cypress_set_mc_reg_address_table(rdev);
+		cypress_force_mc_use_s0(rdev, boot_ps);
+		ret = cypress_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+		cypress_force_mc_use_s1(rdev, boot_ps);
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv770_enable_backbias(rdev, true);
+
+	if (pi->dynamic_ss)
+		cypress_enable_spread_spectrum(rdev, true);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	rv770_program_engine_speed_parameters(rdev);
+	cypress_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+
+	if (pi->dynamic_pcie_gen2)
+		cypress_enable_dynamic_pcie_gen2(rdev, true);
+
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+
+	ret = cypress_get_table_locations(rdev);
+	if (ret) {
+		DRM_ERROR("cypress_get_table_locations failed\n");
+		return ret;
+	}
+	ret = cypress_init_smc_table(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("cypress_init_smc_table failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("cypress_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_response_times(rdev);
+
+	r7xx_start_smc(rdev);
+
+	ret = cypress_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("cypress_notify_smc_display_change failed\n");
+		return ret;
+	}
+	cypress_enable_sclk_control(rdev, true);
+
+	if (eg_pi->memory_transition)
+		cypress_enable_mclk_control(rdev, true);
+
+	cypress_start_dpm(rdev);
+
+	if (pi->gfx_clock_gating)
+		cypress_gfx_clock_gating_enable(rdev, true);
+
+	if (pi->mg_clock_gating)
+		cypress_mg_clock_gating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	return 0;
+}
+
+void cypress_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!rv770_dpm_enabled(rdev))
+		return;
+
+	rv770_clear_vc(rdev);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+
+	if (pi->dynamic_pcie_gen2)
+		cypress_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		cypress_gfx_clock_gating_enable(rdev, false);
+
+	if (pi->mg_clock_gating)
+		cypress_mg_clock_gating_enable(rdev, false);
+
+	rv770_stop_dpm(rdev);
+	r7xx_stop_smc(rdev);
+
+	cypress_enable_spread_spectrum(rdev, false);
+
+	if (eg_pi->dynamic_ac_timing)
+		cypress_force_mc_use_s1(rdev, boot_ps);
+
+	rv770_reset_smio_status(rdev);
+}
+
+int cypress_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	ret = rv770_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_before_state_change(rdev, new_ps, old_ps);
+
+	rv770_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	ret = cypress_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("cypress_upload_sw_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("cypress_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_memory_timing_parameters(rdev, new_ps);
+
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps);
+
+	ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("rv770_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void cypress_dpm_reset_asic(struct radeon_device *rdev)
+{
+	rv770_restrict_performance_levels_before_switch(rdev);
+	rv770_set_boot_state(rdev);
+}
+
+void cypress_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	cypress_program_display_gap(rdev);
+}
+
+int cypress_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	uint16_t data_offset, size;
+	uint8_t frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	eg_pi = kzalloc(sizeof(struct evergreen_power_info), GFP_KERNEL);
+	if (eg_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = eg_pi;
+	pi = &eg_pi->rv7xx;
+
+	rv770_get_max_vddc(rdev);
+
+	eg_pi->ulv.supported = false;
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = rv7xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->mclk_strobe_mode_threshold = 40000;
+	pi->mclk_edc_enable_threshold = 40000;
+	eg_pi->mclk_edc_wr_enable_threshold = 40000;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = CYPRESS_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK))
+		pi->gfx_clock_gating = false;
+	else
+		pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+	eg_pi->ls_clock_gating = false;
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	if (rdev->flags & RADEON_IS_MOBILITY)
+		pi->dcodt = true;
+	else
+		pi->dcodt = false;
+
+	pi->ulps = true;
+
+	eg_pi->dynamic_ac_timing = true;
+	eg_pi->abm = true;
+	eg_pi->mcls = true;
+	eg_pi->light_sleep = true;
+	eg_pi->memory_transition = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK) ||
+	    (rdev->family == CHIP_JUNIPER))
+		eg_pi->dll_default_on = true;
+	else
+		eg_pi->dll_default_on = false;
+
+	eg_pi->sclk_deep_sleep = false;
+	pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+
+	return 0;
+}
+
+void cypress_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+bool cypress_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+	u32 switch_limit = pi->mem_gddr5 ? 450 : 300;
+
+	if (vblank_time < switch_limit)
+		return true;
+	else
+		return false;
+
+}
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.h b/drivers/gpu/drm/radeon/cypress_dpm.h
new file mode 100644
index 0000000..4c3f18c
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cypress_dpm.h
@@ -0,0 +1,160 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __CYPRESS_DPM_H__
+#define __CYPRESS_DPM_H__
+
+#include "rv770_dpm.h"
+#include "evergreen_smc.h"
+
+struct evergreen_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct evergreen_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	u16 valid_flag;
+	struct evergreen_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES];
+	SMC_Evergreen_MCRegisterAddress mc_reg_address[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct evergreen_ulv_param {
+	bool supported;
+	struct rv7xx_pl *pl;
+};
+
+struct evergreen_arb_registers {
+	u32 mc_arb_dram_timing;
+	u32 mc_arb_dram_timing2;
+	u32 mc_arb_rfsh_rate;
+	u32 mc_arb_burst_time;
+};
+
+struct at {
+	u32 rlp;
+	u32 rmp;
+	u32 lhp;
+	u32 lmp;
+};
+
+struct evergreen_power_info {
+	/* must be first! */
+	struct rv7xx_power_info rv7xx;
+	/* flags */
+	bool vddci_control;
+	bool dynamic_ac_timing;
+	bool abm;
+	bool mcls;
+	bool light_sleep;
+	bool memory_transition;
+	bool pcie_performance_request;
+	bool pcie_performance_request_registered;
+	bool sclk_deep_sleep;
+	bool dll_default_on;
+	bool ls_clock_gating;
+	bool smu_uvd_hs;
+	bool uvd_enabled;
+	/* stored values */
+	u16 acpi_vddci;
+	u8 mvdd_high_index;
+	u8 mvdd_low_index;
+	u32 mclk_edc_wr_enable_threshold;
+	struct evergreen_mc_reg_table mc_reg_table;
+	struct atom_voltage_table vddc_voltage_table;
+	struct atom_voltage_table vddci_voltage_table;
+	struct evergreen_arb_registers bootup_arb_registers;
+	struct evergreen_ulv_param ulv;
+	struct at ats[2];
+	/* smc offsets */
+	u16 mc_reg_table_start;
+	struct radeon_ps current_rps;
+	struct rv7xx_ps current_ps;
+	struct radeon_ps requested_rps;
+	struct rv7xx_ps requested_ps;
+};
+
+#define CYPRESS_HASI_DFLT                               400000
+#define CYPRESS_MGCGTTLOCAL0_DFLT                       0x00000000
+#define CYPRESS_MGCGTTLOCAL1_DFLT                       0x00000000
+#define CYPRESS_MGCGTTLOCAL2_DFLT                       0x00000000
+#define CYPRESS_MGCGTTLOCAL3_DFLT                       0x00000000
+#define CYPRESS_MGCGCGTSSMCTRL_DFLT                     0x81944bc0
+#define REDWOOD_MGCGCGTSSMCTRL_DFLT                     0x6e944040
+#define CEDAR_MGCGCGTSSMCTRL_DFLT                       0x46944040
+#define CYPRESS_VRC_DFLT                                0xC00033
+
+#define PCIE_PERF_REQ_REMOVE_REGISTRY   0
+#define PCIE_PERF_REQ_FORCE_LOWPOWER    1
+#define PCIE_PERF_REQ_PECI_GEN1         2
+#define PCIE_PERF_REQ_PECI_GEN2         3
+#define PCIE_PERF_REQ_PECI_GEN3         4
+
+int cypress_convert_power_level_to_smc(struct radeon_device *rdev,
+				       struct rv7xx_pl *pl,
+				       RV770_SMC_HW_PERFORMANCE_LEVEL *level,
+				       u8 watermark_level);
+int cypress_populate_smc_acpi_state(struct radeon_device *rdev,
+				    RV770_SMC_STATETABLE *table);
+int cypress_populate_smc_voltage_tables(struct radeon_device *rdev,
+					RV770_SMC_STATETABLE *table);
+int cypress_populate_smc_initial_state(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_initial_state,
+				       RV770_SMC_STATETABLE *table);
+u32 cypress_calculate_burst_time(struct radeon_device *rdev,
+				 u32 engine_clock, u32 memory_clock);
+void cypress_notify_link_speed_change_before_state_change(struct radeon_device *rdev,
+							  struct radeon_ps *radeon_new_state,
+							  struct radeon_ps *radeon_current_state);
+int cypress_upload_sw_state(struct radeon_device *rdev,
+			    struct radeon_ps *radeon_new_state);
+int cypress_upload_mc_reg_table(struct radeon_device *rdev,
+				struct radeon_ps *radeon_new_state);
+void cypress_program_memory_timing_parameters(struct radeon_device *rdev,
+					      struct radeon_ps *radeon_new_state);
+void cypress_notify_link_speed_change_after_state_change(struct radeon_device *rdev,
+							 struct radeon_ps *radeon_new_state,
+							 struct radeon_ps *radeon_current_state);
+int cypress_construct_voltage_tables(struct radeon_device *rdev);
+int cypress_get_mvdd_configuration(struct radeon_device *rdev);
+void cypress_enable_spread_spectrum(struct radeon_device *rdev,
+				    bool enable);
+void cypress_enable_display_gap(struct radeon_device *rdev);
+int cypress_get_table_locations(struct radeon_device *rdev);
+int cypress_populate_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_boot_state);
+void cypress_program_response_times(struct radeon_device *rdev);
+int cypress_notify_smc_display_change(struct radeon_device *rdev,
+				      bool has_display);
+void cypress_enable_sclk_control(struct radeon_device *rdev,
+				 bool enable);
+void cypress_enable_mclk_control(struct radeon_device *rdev,
+				 bool enable);
+void cypress_start_dpm(struct radeon_device *rdev);
+void cypress_advertise_gen2_capability(struct radeon_device *rdev);
+u32 cypress_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf);
+u8 cypress_get_mclk_frequency_ratio(struct radeon_device *rdev,
+				    u32 memory_clock, bool strobe_mode);
+u8 cypress_get_strobe_mode_settings(struct radeon_device *rdev, u32 mclk);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 0f89ce3..e49059d 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -33,9 +33,7 @@
 #include "avivod.h"
 #include "evergreen_reg.h"
 #include "evergreen_blit_shaders.h"
-
-#define EVERGREEN_PFP_UCODE_SIZE 1120
-#define EVERGREEN_PM4_UCODE_SIZE 1376
+#include "radeon_ucode.h"
 
 static const u32 crtc_offsets[6] =
 {
@@ -47,9 +45,98 @@ static const u32 crtc_offsets[6] =
 	EVERGREEN_CRTC5_REGISTER_OFFSET
 };
 
+#include "clearstate_evergreen.h"
+
+static u32 sumo_rlc_save_restore_register_list[] =
+{
+	0x98fc,
+	0x9830,
+	0x9834,
+	0x9838,
+	0x9870,
+	0x9874,
+	0x8a14,
+	0x8b24,
+	0x8bcc,
+	0x8b10,
+	0x8d00,
+	0x8d04,
+	0x8c00,
+	0x8c04,
+	0x8c08,
+	0x8c0c,
+	0x8d8c,
+	0x8c20,
+	0x8c24,
+	0x8c28,
+	0x8c18,
+	0x8c1c,
+	0x8cf0,
+	0x8e2c,
+	0x8e38,
+	0x8c30,
+	0x9508,
+	0x9688,
+	0x9608,
+	0x960c,
+	0x9610,
+	0x9614,
+	0x88c4,
+	0x88d4,
+	0xa008,
+	0x900c,
+	0x9100,
+	0x913c,
+	0x98f8,
+	0x98f4,
+	0x9b7c,
+	0x3f8c,
+	0x8950,
+	0x8954,
+	0x8a18,
+	0x8b28,
+	0x9144,
+	0x9148,
+	0x914c,
+	0x3f90,
+	0x3f94,
+	0x915c,
+	0x9160,
+	0x9178,
+	0x917c,
+	0x9180,
+	0x918c,
+	0x9190,
+	0x9194,
+	0x9198,
+	0x919c,
+	0x91a8,
+	0x91ac,
+	0x91b0,
+	0x91b4,
+	0x91b8,
+	0x91c4,
+	0x91c8,
+	0x91cc,
+	0x91d0,
+	0x91d4,
+	0x91e0,
+	0x91e4,
+	0x91ec,
+	0x91f0,
+	0x91f4,
+	0x9200,
+	0x9204,
+	0x929c,
+	0x9150,
+	0x802c,
+};
+static u32 sumo_rlc_save_restore_register_list_size = ARRAY_SIZE(sumo_rlc_save_restore_register_list);
+
 static void evergreen_gpu_init(struct radeon_device *rdev);
 void evergreen_fini(struct radeon_device *rdev);
 void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
+void evergreen_program_aspm(struct radeon_device *rdev);
 extern void cayman_cp_int_cntl_setup(struct radeon_device *rdev,
 				     int ring, u32 cp_int_cntl);
 
@@ -1417,8 +1504,8 @@ void evergreen_pm_misc(struct radeon_device *rdev)
 	struct radeon_voltage *voltage = &ps->clock_info[req_cm_idx].voltage;
 
 	if (voltage->type == VOLTAGE_SW) {
-		/* 0xff01 is a flag rather then an actual voltage */
-		if (voltage->voltage == 0xff01)
+		/* 0xff0x are flags rather then an actual voltage */
+		if ((voltage->voltage & 0xff00) == 0xff00)
 			return;
 		if (voltage->voltage && (voltage->voltage != rdev->pm.current_vddc)) {
 			radeon_atom_set_voltage(rdev, voltage->voltage, SET_VOLTAGE_TYPE_ASIC_VDDC);
@@ -1438,8 +1525,8 @@ void evergreen_pm_misc(struct radeon_device *rdev)
 			voltage = &rdev->pm.power_state[req_ps_idx].
 				clock_info[rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_cm_idx].voltage;
 
-		/* 0xff01 is a flag rather then an actual voltage */
-		if (voltage->vddci == 0xff01)
+		/* 0xff0x are flags rather then an actual voltage */
+		if ((voltage->vddci & 0xff00) == 0xff00)
 			return;
 		if (voltage->vddci && (voltage->vddci != rdev->pm.current_vddci)) {
 			radeon_atom_set_voltage(rdev, voltage->vddci, SET_VOLTAGE_TYPE_ASIC_VDDCI);
@@ -2036,7 +2123,8 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 					 u32 lb_size, u32 num_heads)
 {
 	struct drm_display_mode *mode = &radeon_crtc->base.mode;
-	struct evergreen_wm_params wm;
+	struct evergreen_wm_params wm_low, wm_high;
+	u32 dram_channels;
 	u32 pixel_period;
 	u32 line_time = 0;
 	u32 latency_watermark_a = 0, latency_watermark_b = 0;
@@ -2052,39 +2140,81 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 		line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535);
 		priority_a_cnt = 0;
 		priority_b_cnt = 0;
+		dram_channels = evergreen_get_number_of_dram_channels(rdev);
+
+		/* watermark for high clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_high.yclk =
+				radeon_dpm_get_mclk(rdev, false) * 10;
+			wm_high.sclk =
+				radeon_dpm_get_sclk(rdev, false) * 10;
+		} else {
+			wm_high.yclk = rdev->pm.current_mclk * 10;
+			wm_high.sclk = rdev->pm.current_sclk * 10;
+		}
 
-		wm.yclk = rdev->pm.current_mclk * 10;
-		wm.sclk = rdev->pm.current_sclk * 10;
-		wm.disp_clk = mode->clock;
-		wm.src_width = mode->crtc_hdisplay;
-		wm.active_time = mode->crtc_hdisplay * pixel_period;
-		wm.blank_time = line_time - wm.active_time;
-		wm.interlaced = false;
+		wm_high.disp_clk = mode->clock;
+		wm_high.src_width = mode->crtc_hdisplay;
+		wm_high.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_high.blank_time = line_time - wm_high.active_time;
+		wm_high.interlaced = false;
 		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-			wm.interlaced = true;
-		wm.vsc = radeon_crtc->vsc;
-		wm.vtaps = 1;
+			wm_high.interlaced = true;
+		wm_high.vsc = radeon_crtc->vsc;
+		wm_high.vtaps = 1;
 		if (radeon_crtc->rmx_type != RMX_OFF)
-			wm.vtaps = 2;
-		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
-		wm.lb_size = lb_size;
-		wm.dram_channels = evergreen_get_number_of_dram_channels(rdev);
-		wm.num_heads = num_heads;
+			wm_high.vtaps = 2;
+		wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_high.lb_size = lb_size;
+		wm_high.dram_channels = dram_channels;
+		wm_high.num_heads = num_heads;
+
+		/* watermark for low clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_low.yclk =
+				radeon_dpm_get_mclk(rdev, true) * 10;
+			wm_low.sclk =
+				radeon_dpm_get_sclk(rdev, true) * 10;
+		} else {
+			wm_low.yclk = rdev->pm.current_mclk * 10;
+			wm_low.sclk = rdev->pm.current_sclk * 10;
+		}
+
+		wm_low.disp_clk = mode->clock;
+		wm_low.src_width = mode->crtc_hdisplay;
+		wm_low.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_low.blank_time = line_time - wm_low.active_time;
+		wm_low.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_low.interlaced = true;
+		wm_low.vsc = radeon_crtc->vsc;
+		wm_low.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm_low.vtaps = 2;
+		wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_low.lb_size = lb_size;
+		wm_low.dram_channels = dram_channels;
+		wm_low.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(evergreen_latency_watermark(&wm), (u32)65535);
+		latency_watermark_a = min(evergreen_latency_watermark(&wm_high), (u32)65535);
 		/* set for low clocks */
-		/* wm.yclk = low clk; wm.sclk = low clk */
-		latency_watermark_b = min(evergreen_latency_watermark(&wm), (u32)65535);
+		latency_watermark_b = min(evergreen_latency_watermark(&wm_low), (u32)65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
-		if (!evergreen_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
-		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm) ||
-		    !evergreen_check_latency_hiding(&wm) ||
+		if (!evergreen_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) ||
+		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm_high) ||
+		    !evergreen_check_latency_hiding(&wm_high) ||
 		    (rdev->disp_priority == 2)) {
-			DRM_DEBUG_KMS("force priority to high\n");
+			DRM_DEBUG_KMS("force priority a to high\n");
 			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+		}
+		if (!evergreen_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) ||
+		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm_low) ||
+		    !evergreen_check_latency_hiding(&wm_low) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority b to high\n");
 			priority_b_cnt |= PRIORITY_ALWAYS_ON;
 		}
 
@@ -2137,6 +2267,10 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 	WREG32(PRIORITY_A_CNT + radeon_crtc->crtc_offset, priority_a_cnt);
 	WREG32(PRIORITY_B_CNT + radeon_crtc->crtc_offset, priority_b_cnt);
 
+	/* save values for DPM */
+	radeon_crtc->line_time = line_time;
+	radeon_crtc->wm_high = latency_watermark_a;
+	radeon_crtc->wm_low = latency_watermark_b;
 }
 
 /**
@@ -3120,10 +3254,8 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 		u32 efuse_straps_4;
 		u32 efuse_straps_3;
 
-		WREG32(RCU_IND_INDEX, 0x204);
-		efuse_straps_4 = RREG32(RCU_IND_DATA);
-		WREG32(RCU_IND_INDEX, 0x203);
-		efuse_straps_3 = RREG32(RCU_IND_DATA);
+		efuse_straps_4 = RREG32_RCU(0x204);
+		efuse_straps_3 = RREG32_RCU(0x203);
 		tmp = (((efuse_straps_4 & 0xf) << 4) |
 		      ((efuse_straps_3 & 0xf0000000) >> 28));
 	} else {
@@ -3727,6 +3859,262 @@ bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
+/*
+ * RLC
+ */
+#define RLC_SAVE_RESTORE_LIST_END_MARKER    0x00000000
+#define RLC_CLEAR_STATE_END_MARKER          0x00000001
+
+void sumo_rlc_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	/* save restore block */
+	if (rdev->rlc.save_restore_obj) {
+		r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r);
+		radeon_bo_unpin(rdev->rlc.save_restore_obj);
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
+		radeon_bo_unref(&rdev->rlc.save_restore_obj);
+		rdev->rlc.save_restore_obj = NULL;
+	}
+
+	/* clear state block */
+	if (rdev->rlc.clear_state_obj) {
+		r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r);
+		radeon_bo_unpin(rdev->rlc.clear_state_obj);
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+
+		radeon_bo_unref(&rdev->rlc.clear_state_obj);
+		rdev->rlc.clear_state_obj = NULL;
+	}
+}
+
+int sumo_rlc_init(struct radeon_device *rdev)
+{
+	u32 *src_ptr;
+	volatile u32 *dst_ptr;
+	u32 dws, data, i, j, k, reg_num;
+	u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index;
+	u64 reg_list_mc_addr;
+	struct cs_section_def *cs_data;
+	int r;
+
+	src_ptr = rdev->rlc.reg_list;
+	dws = rdev->rlc.reg_list_size;
+	cs_data = rdev->rlc.cs_data;
+
+	/* save restore block */
+	if (rdev->rlc.save_restore_obj == NULL) {
+		r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.save_restore_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
+			return r;
+		}
+	}
+
+	r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
+	if (unlikely(r != 0)) {
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->rlc.save_restore_gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+		dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	/* write the sr buffer */
+	dst_ptr = rdev->rlc.sr_ptr;
+	/* format:
+	 * dw0: (reg2 << 16) | reg1
+	 * dw1: reg1 save space
+	 * dw2: reg2 save space
+	 */
+	for (i = 0; i < dws; i++) {
+		data = src_ptr[i] >> 2;
+		i++;
+		if (i < dws)
+			data |= (src_ptr[i] >> 2) << 16;
+		j = (((i - 1) * 3) / 2);
+		dst_ptr[j] = data;
+	}
+	j = ((i * 3) / 2);
+	dst_ptr[j] = RLC_SAVE_RESTORE_LIST_END_MARKER;
+
+	radeon_bo_kunmap(rdev->rlc.save_restore_obj);
+	radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
+	/* clear state block */
+	reg_list_num = 0;
+	dws = 0;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_list_num++;
+			dws += cs_data[i].section[j].reg_count;
+		}
+	}
+	reg_list_blk_index = (3 * reg_list_num + 2);
+	dws += reg_list_blk_index;
+
+	if (rdev->rlc.clear_state_obj == NULL) {
+		r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
+			sumo_rlc_fini(rdev);
+			return r;
+		}
+	}
+	r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
+	if (unlikely(r != 0)) {
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->rlc.clear_state_gpu_addr);
+	if (r) {
+
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+		dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	/* set up the cs buffer */
+	dst_ptr = rdev->rlc.cs_ptr;
+	reg_list_hdr_blk_index = 0;
+	reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4);
+	data = upper_32_bits(reg_list_mc_addr);
+	dst_ptr[reg_list_hdr_blk_index] = data;
+	reg_list_hdr_blk_index++;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_num = cs_data[i].section[j].reg_count;
+			data = reg_list_mc_addr & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = 0x08000000 | (reg_num * 4);
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			for (k = 0; k < reg_num; k++) {
+				data = cs_data[i].section[j].extent[k];
+				dst_ptr[reg_list_blk_index + k] = data;
+			}
+			reg_list_mc_addr += reg_num * 4;
+			reg_list_blk_index += reg_num;
+		}
+	}
+	dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER;
+
+	radeon_bo_kunmap(rdev->rlc.clear_state_obj);
+	radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+
+	return 0;
+}
+
+static void evergreen_rlc_start(struct radeon_device *rdev)
+{
+	u32 mask = RLC_ENABLE;
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		mask |= GFX_POWER_GATING_ENABLE | GFX_POWER_GATING_SRC;
+	}
+
+	WREG32(RLC_CNTL, mask);
+}
+
+int evergreen_rlc_resume(struct radeon_device *rdev)
+{
+	u32 i;
+	const __be32 *fw_data;
+
+	if (!rdev->rlc_fw)
+		return -EINVAL;
+
+	r600_rlc_stop(rdev);
+
+	WREG32(RLC_HB_CNTL, 0);
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (rdev->family == CHIP_ARUBA) {
+			u32 always_on_bitmap =
+				3 | (3 << (16 * rdev->config.cayman.max_shader_engines));
+			/* find out the number of active simds */
+			u32 tmp = (RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffff0000) >> 16;
+			tmp |= 0xffffffff << rdev->config.cayman.max_simds_per_se;
+			tmp = hweight32(~tmp);
+			if (tmp == rdev->config.cayman.max_simds_per_se) {
+				WREG32(TN_RLC_LB_ALWAYS_ACTIVE_SIMD_MASK, always_on_bitmap);
+				WREG32(TN_RLC_LB_PARAMS, 0x00601004);
+				WREG32(TN_RLC_LB_INIT_SIMD_MASK, 0xffffffff);
+				WREG32(TN_RLC_LB_CNTR_INIT, 0x00000000);
+				WREG32(TN_RLC_LB_CNTR_MAX, 0x00002000);
+			}
+		} else {
+			WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
+			WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
+		}
+		WREG32(TN_RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+		WREG32(TN_RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	} else {
+		WREG32(RLC_HB_BASE, 0);
+		WREG32(RLC_HB_RPTR, 0);
+		WREG32(RLC_HB_WPTR, 0);
+		WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
+		WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
+	}
+	WREG32(RLC_MC_CNTL, 0);
+	WREG32(RLC_UCODE_CNTL, 0);
+
+	fw_data = (const __be32 *)rdev->rlc_fw->data;
+	if (rdev->family >= CHIP_ARUBA) {
+		for (i = 0; i < ARUBA_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	} else if (rdev->family >= CHIP_CAYMAN) {
+		for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	} else {
+		for (i = 0; i < EVERGREEN_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	}
+	WREG32(RLC_UCODE_ADDR, 0);
+
+	evergreen_rlc_start(rdev);
+
+	return 0;
+}
+
 /* Interrupts */
 
 u32 evergreen_get_vblank_counter(struct radeon_device *rdev, int crtc)
@@ -3805,6 +4193,7 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
 	u32 afmt1 = 0, afmt2 = 0, afmt3 = 0, afmt4 = 0, afmt5 = 0, afmt6 = 0;
 	u32 dma_cntl, dma_cntl1 = 0;
+	u32 thermal_int = 0;
 
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
@@ -3824,6 +4213,12 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN;
 	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
 	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	if (rdev->family == CHIP_ARUBA)
+		thermal_int = RREG32(TN_CG_THERMAL_INT_CTRL) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+	else
+		thermal_int = RREG32(CG_THERMAL_INT) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
 
 	afmt1 = RREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET) & ~AFMT_AZ_FORMAT_WTRIG_MASK;
 	afmt2 = RREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET) & ~AFMT_AZ_FORMAT_WTRIG_MASK;
@@ -3869,6 +4264,11 @@ int evergreen_irq_set(struct radeon_device *rdev)
 		}
 	}
 
+	if (rdev->irq.dpm_thermal) {
+		DRM_DEBUG("dpm thermal\n");
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+	}
+
 	if (rdev->irq.crtc_vblank_int[0] ||
 	    atomic_read(&rdev->irq.pflip[0])) {
 		DRM_DEBUG("evergreen_irq_set: vblank 0\n");
@@ -3990,6 +4390,10 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	WREG32(DC_HPD4_INT_CONTROL, hpd4);
 	WREG32(DC_HPD5_INT_CONTROL, hpd5);
 	WREG32(DC_HPD6_INT_CONTROL, hpd6);
+	if (rdev->family == CHIP_ARUBA)
+		WREG32(TN_CG_THERMAL_INT_CTRL, thermal_int);
+	else
+		WREG32(CG_THERMAL_INT, thermal_int);
 
 	WREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, afmt1);
 	WREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, afmt2);
@@ -4181,6 +4585,7 @@ int evergreen_irq_process(struct radeon_device *rdev)
 	u32 ring_index;
 	bool queue_hotplug = false;
 	bool queue_hdmi = false;
+	bool queue_thermal = false;
 
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
@@ -4502,6 +4907,16 @@ restart_ih:
 			DRM_DEBUG("IH: DMA trap\n");
 			radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
 			break;
+		case 230: /* thermal low to high */
+			DRM_DEBUG("IH: thermal low to high\n");
+			rdev->pm.dpm.thermal.high_to_low = false;
+			queue_thermal = true;
+			break;
+		case 231: /* thermal high to low */
+			DRM_DEBUG("IH: thermal high to low\n");
+			rdev->pm.dpm.thermal.high_to_low = true;
+			queue_thermal = true;
+			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
 			break;
@@ -4524,6 +4939,8 @@ restart_ih:
 		schedule_work(&rdev->hotplug_work);
 	if (queue_hdmi)
 		schedule_work(&rdev->audio_work);
+	if (queue_thermal && rdev->pm.dpm_enabled)
+		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
 	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
@@ -4680,6 +5097,8 @@ static int evergreen_startup(struct radeon_device *rdev)
 
 	/* enable pcie gen2 link */
 	evergreen_pcie_gen2_enable(rdev);
+	/* enable aspm */
+	evergreen_program_aspm(rdev);
 
 	if (ASIC_IS_DCE5(rdev)) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
@@ -4725,6 +5144,18 @@ static int evergreen_startup(struct radeon_device *rdev)
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
+	/* allocate rlc buffers */
+	if (rdev->flags & RADEON_IS_IGP) {
+		rdev->rlc.reg_list = sumo_rlc_save_restore_register_list;
+		rdev->rlc.reg_list_size = sumo_rlc_save_restore_register_list_size;
+		rdev->rlc.cs_data = evergreen_cs_data;
+		r = sumo_rlc_init(rdev);
+		if (r) {
+			DRM_ERROR("Failed to init rlc BOs!\n");
+			return r;
+		}
+	}
+
 	/* allocate wb buffer */
 	r = radeon_wb_init(rdev);
 	if (r)
@@ -4956,6 +5387,8 @@ int evergreen_init(struct radeon_device *rdev)
 		r700_cp_fini(rdev);
 		r600_dma_fini(rdev);
 		r600_irq_fini(rdev);
+		if (rdev->flags & RADEON_IS_IGP)
+			sumo_rlc_fini(rdev);
 		radeon_wb_fini(rdev);
 		radeon_ib_pool_fini(rdev);
 		radeon_irq_kms_fini(rdev);
@@ -4984,6 +5417,8 @@ void evergreen_fini(struct radeon_device *rdev)
 	r700_cp_fini(rdev);
 	r600_dma_fini(rdev);
 	r600_irq_fini(rdev);
+	if (rdev->flags & RADEON_IS_IGP)
+		sumo_rlc_fini(rdev);
 	radeon_wb_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
@@ -5061,3 +5496,150 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
 		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	}
 }
+
+void evergreen_program_aspm(struct radeon_device *rdev)
+{
+	u32 data, orig;
+	u32 pcie_lc_cntl, pcie_lc_cntl_old;
+	bool disable_l0s, disable_l1 = false, disable_plloff_in_l1 = false;
+	/* fusion_platform = true
+	 * if the system is a fusion system
+	 * (APU or DGPU in a fusion system).
+	 * todo: check if the system is a fusion platform.
+	 */
+	bool fusion_platform = false;
+
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		return;
+
+	switch (rdev->family) {
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+	case CHIP_JUNIPER:
+	case CHIP_REDWOOD:
+	case CHIP_CEDAR:
+	case CHIP_SUMO:
+	case CHIP_SUMO2:
+	case CHIP_PALM:
+	case CHIP_ARUBA:
+		disable_l0s = true;
+		break;
+	default:
+		disable_l0s = false;
+		break;
+	}
+
+	if (rdev->flags & RADEON_IS_IGP)
+		fusion_platform = true; /* XXX also dGPUs in a fusion system */
+
+	data = orig = RREG32_PIF_PHY0(PB0_PIF_PAIRING);
+	if (fusion_platform)
+		data &= ~MULTI_PIF;
+	else
+		data |= MULTI_PIF;
+	if (data != orig)
+		WREG32_PIF_PHY0(PB0_PIF_PAIRING, data);
+
+	data = orig = RREG32_PIF_PHY1(PB1_PIF_PAIRING);
+	if (fusion_platform)
+		data &= ~MULTI_PIF;
+	else
+		data |= MULTI_PIF;
+	if (data != orig)
+		WREG32_PIF_PHY1(PB1_PIF_PAIRING, data);
+
+	pcie_lc_cntl = pcie_lc_cntl_old = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	pcie_lc_cntl &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK);
+	if (!disable_l0s) {
+		if (rdev->family >= CHIP_BARTS)
+			pcie_lc_cntl |= LC_L0S_INACTIVITY(7);
+		else
+			pcie_lc_cntl |= LC_L0S_INACTIVITY(3);
+	}
+
+	if (!disable_l1) {
+		if (rdev->family >= CHIP_BARTS)
+			pcie_lc_cntl |= LC_L1_INACTIVITY(7);
+		else
+			pcie_lc_cntl |= LC_L1_INACTIVITY(8);
+
+		if (!disable_plloff_in_l1) {
+			data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (data != orig)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+			data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (data != orig)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+			data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (data != orig)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+			data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (data != orig)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+
+			if (rdev->family >= CHIP_BARTS) {
+				data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				data |= PLL_RAMP_UP_TIME_0(4);
+				if (data != orig)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+				data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				data |= PLL_RAMP_UP_TIME_1(4);
+				if (data != orig)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+				data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				data |= PLL_RAMP_UP_TIME_0(4);
+				if (data != orig)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+				data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				data |= PLL_RAMP_UP_TIME_1(4);
+				if (data != orig)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+			}
+
+			data = orig = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+			data &= ~LC_DYN_LANES_PWR_STATE_MASK;
+			data |= LC_DYN_LANES_PWR_STATE(3);
+			if (data != orig)
+				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
+
+			if (rdev->family >= CHIP_BARTS) {
+				data = orig = RREG32_PIF_PHY0(PB0_PIF_CNTL);
+				data &= ~LS2_EXIT_TIME_MASK;
+				data |= LS2_EXIT_TIME(1);
+				if (data != orig)
+					WREG32_PIF_PHY0(PB0_PIF_CNTL, data);
+
+				data = orig = RREG32_PIF_PHY1(PB1_PIF_CNTL);
+				data &= ~LS2_EXIT_TIME_MASK;
+				data |= LS2_EXIT_TIME(1);
+				if (data != orig)
+					WREG32_PIF_PHY1(PB1_PIF_CNTL, data);
+			}
+		}
+	}
+
+	/* evergreen parts only */
+	if (rdev->family < CHIP_BARTS)
+		pcie_lc_cntl |= LC_PMI_TO_L1_DIS;
+
+	if (pcie_lc_cntl != pcie_lc_cntl_old)
+		WREG32_PCIE_PORT(PCIE_LC_CNTL, pcie_lc_cntl);
+}
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index ed7c8a7..b9c6f76 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -128,14 +128,7 @@ static void evergreen_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
 	uint8_t *frame = buffer + 3;
-
-	/* Our header values (type, version, length) should be alright, Intel
-	 * is using the same. Checksum function also seems to be OK, it works
-	 * fine for audio infoframe. However calculated value is always lower
-	 * by 2 in comparison to fglrx. It breaks displaying anything in case
-	 * of TVs that strictly check the checksum. Hack it manually here to
-	 * workaround this issue. */
-	frame[0x0] += 2;
+	uint8_t *header = buffer;
 
 	WREG32(AFMT_AVI_INFO0 + offset,
 		frame[0x0] | (frame[0x1] << 8) | (frame[0x2] << 16) | (frame[0x3] << 24));
@@ -144,7 +137,7 @@ static void evergreen_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	WREG32(AFMT_AVI_INFO2 + offset,
 		frame[0x8] | (frame[0x9] << 8) | (frame[0xA] << 16) | (frame[0xB] << 24));
 	WREG32(AFMT_AVI_INFO3 + offset,
-		frame[0xC] | (frame[0xD] << 8));
+		frame[0xC] | (frame[0xD] << 8) | (header[1] << 24));
 }
 
 static void evergreen_audio_set_dto(struct drm_encoder *encoder, u32 clock)
diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h
index 881aba2..8a4e641 100644
--- a/drivers/gpu/drm/radeon/evergreen_reg.h
+++ b/drivers/gpu/drm/radeon/evergreen_reg.h
@@ -24,7 +24,16 @@
 #ifndef __EVERGREEN_REG_H__
 #define __EVERGREEN_REG_H__
 
+/* trinity */
+#define TN_SMC_IND_INDEX_0                              0x200
+#define TN_SMC_IND_DATA_0                               0x204
+
 /* evergreen */
+#define EVERGREEN_PIF_PHY0_INDEX                        0x8
+#define EVERGREEN_PIF_PHY0_DATA                         0xc
+#define EVERGREEN_PIF_PHY1_INDEX                        0x10
+#define EVERGREEN_PIF_PHY1_DATA                         0x14
+
 #define EVERGREEN_VGA_MEMORY_BASE_ADDRESS               0x310
 #define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH          0x324
 #define EVERGREEN_D3VGA_CONTROL                         0x3e0
@@ -40,6 +49,9 @@
 #define EVERGREEN_AUDIO_PLL1_DIV			0x5b4
 #define EVERGREEN_AUDIO_PLL1_UNK			0x5bc
 
+#define EVERGREEN_CG_IND_ADDR                           0x8f8
+#define EVERGREEN_CG_IND_DATA                           0x8fc
+
 #define EVERGREEN_AUDIO_ENABLE				0x5e78
 #define EVERGREEN_AUDIO_VENDOR_ID			0x5ec0
 
diff --git a/drivers/gpu/drm/radeon/evergreen_smc.h b/drivers/gpu/drm/radeon/evergreen_smc.h
new file mode 100644
index 0000000..76ada8c
--- /dev/null
+++ b/drivers/gpu/drm/radeon/evergreen_smc.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __EVERGREEN_SMC_H__
+#define __EVERGREEN_SMC_H__
+
+#include "rv770_smc.h"
+
+#pragma pack(push, 1)
+
+#define SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE 16
+
+struct SMC_Evergreen_MCRegisterAddress
+{
+    uint16_t s0;
+    uint16_t s1;
+};
+
+typedef struct SMC_Evergreen_MCRegisterAddress SMC_Evergreen_MCRegisterAddress;
+
+
+struct SMC_Evergreen_MCRegisterSet
+{
+    uint32_t value[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+};
+
+typedef struct SMC_Evergreen_MCRegisterSet SMC_Evergreen_MCRegisterSet;
+
+struct SMC_Evergreen_MCRegisters
+{
+    uint8_t                             last;
+    uint8_t                             reserved[3];
+    SMC_Evergreen_MCRegisterAddress     address[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+    SMC_Evergreen_MCRegisterSet         data[5];
+};
+
+typedef struct SMC_Evergreen_MCRegisters SMC_Evergreen_MCRegisters;
+
+#define EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION 0x100
+
+#define EVERGREEN_SMC_FIRMWARE_HEADER_softRegisters   0x0
+#define EVERGREEN_SMC_FIRMWARE_HEADER_stateTable      0xC
+#define EVERGREEN_SMC_FIRMWARE_HEADER_mcRegisterTable 0x20
+
+
+#pragma pack(pop)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 75c0563..a7baf67 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -48,6 +48,293 @@
 #define SUMO_GB_ADDR_CONFIG_GOLDEN           0x02010002
 #define SUMO2_GB_ADDR_CONFIG_GOLDEN          0x02010002
 
+/* pm registers */
+#define	SMC_MSG						0x20c
+#define		HOST_SMC_MSG(x)				((x) << 0)
+#define		HOST_SMC_MSG_MASK			(0xff << 0)
+#define		HOST_SMC_MSG_SHIFT			0
+#define		HOST_SMC_RESP(x)			((x) << 8)
+#define		HOST_SMC_RESP_MASK			(0xff << 8)
+#define		HOST_SMC_RESP_SHIFT			8
+#define		SMC_HOST_MSG(x)				((x) << 16)
+#define		SMC_HOST_MSG_MASK			(0xff << 16)
+#define		SMC_HOST_MSG_SHIFT			16
+#define		SMC_HOST_RESP(x)			((x) << 24)
+#define		SMC_HOST_RESP_MASK			(0xff << 24)
+#define		SMC_HOST_RESP_SHIFT			24
+
+#define DCCG_DISP_SLOW_SELECT_REG                       0x4fc
+#define		DCCG_DISP1_SLOW_SELECT(x)		((x) << 0)
+#define		DCCG_DISP1_SLOW_SELECT_MASK		(7 << 0)
+#define		DCCG_DISP1_SLOW_SELECT_SHIFT		0
+#define		DCCG_DISP2_SLOW_SELECT(x)		((x) << 4)
+#define		DCCG_DISP2_SLOW_SELECT_MASK		(7 << 4)
+#define		DCCG_DISP2_SLOW_SELECT_SHIFT		4
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define MPLL_CNTL_MODE                                  0x61c
+#       define SS_SSEN                                  (1 << 24)
+#       define SS_DSMODE_EN                             (1 << 25)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
+#define	MCLK_PWRMGT_CNTL				0x648
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_PDNB                             (1 << 8)
+#       define MRDCKA1_PDNB                             (1 << 9)
+#       define MRDCKB0_PDNB                             (1 << 10)
+#       define MRDCKB1_PDNB                             (1 << 11)
+#       define MRDCKC0_PDNB                             (1 << 12)
+#       define MRDCKC1_PDNB                             (1 << 13)
+#       define MRDCKD0_PDNB                             (1 << 14)
+#       define MRDCKD1_PDNB                             (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define CG_AT                                           0x6d4
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+
+#define CG_DISPLAY_GAP_CNTL                               0x714
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define	CG_BIF_REQ_AND_RSP				0x7f4
+#define		CG_CLIENT_REQ(x)			((x) << 0)
+#define		CG_CLIENT_REQ_MASK			(0xff << 0)
+#define		CG_CLIENT_REQ_SHIFT			0
+#define		CG_CLIENT_RESP(x)			((x) << 8)
+#define		CG_CLIENT_RESP_MASK			(0xff << 8)
+#define		CG_CLIENT_RESP_SHIFT			8
+#define		CLIENT_CG_REQ(x)			((x) << 16)
+#define		CLIENT_CG_REQ_MASK			(0xff << 16)
+#define		CLIENT_CG_REQ_SHIFT			16
+#define		CLIENT_CG_RESP(x)			((x) << 24)
+#define		CLIENT_CG_RESP_MASK			(0xff << 24)
+#define		CLIENT_CG_RESP_SHIFT			24
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+
+#define	MPLL_SS1					0x85c
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x860
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
+#define	CG_IND_ADDR					0x8f8
+#define	CG_IND_DATA					0x8fc
+/* CGIND regs */
+#define	CG_CGTT_LOCAL_0					0x00
+#define	CG_CGTT_LOCAL_1					0x01
+#define	CG_CGTT_LOCAL_2					0x02
+#define	CG_CGTT_LOCAL_3					0x03
+#define	CG_CGLS_TILE_0					0x20
+#define	CG_CGLS_TILE_1					0x21
+#define	CG_CGLS_TILE_2					0x22
+#define	CG_CGLS_TILE_3					0x23
+#define	CG_CGLS_TILE_4					0x24
+#define	CG_CGLS_TILE_5					0x25
+#define	CG_CGLS_TILE_6					0x26
+#define	CG_CGLS_TILE_7					0x27
+#define	CG_CGLS_TILE_8					0x28
+#define	CG_CGLS_TILE_9					0x29
+#define	CG_CGLS_TILE_10					0x2a
+#define	CG_CGLS_TILE_11					0x2b
+
+#define VM_L2_CG                                        0x15c0
+
+#define MC_CONFIG                                       0x2000
+
+#define MC_CONFIG_MCD                                   0x20a0
+#define MC_CG_CONFIG_MCD                                0x20a4
+#define		MC_RD_ENABLE_MCD(x)			((x) << 8)
+#define		MC_RD_ENABLE_MCD_MASK			(7 << 8)
+
+#define MC_HUB_MISC_HUB_CG                              0x20b8
+#define MC_HUB_MISC_VM_CG                               0x20bc
+#define MC_HUB_MISC_SIP_CG                              0x20c0
+
+#define MC_XPB_CLK_GAT                                  0x2478
+
+#define MC_CG_CONFIG                                    0x25bc
+#define		MC_RD_ENABLE(x)				((x) << 4)
+#define		MC_RD_ENABLE_MASK			(3 << 4)
+
+#define MC_CITF_MISC_RD_CG                              0x2648
+#define MC_CITF_MISC_WR_CG                              0x264c
+#define MC_CITF_MISC_VM_CG                              0x2650
+#       define MEM_LS_ENABLE                            (1 << 19)
+
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_SEQ_STATUS_M                                 0x29f4
+#       define PMG_PWRSTATE                             (1 << 16)
+
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
+#define MC_SEQ_MISC3                                    0x2a2c
+
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_CG                                       0x2a68
+#define		CG_SEQ_REQ(x)				((x) << 0)
+#define		CG_SEQ_REQ_MASK				(0xff << 0)
+#define		CG_SEQ_REQ_SHIFT			0
+#define		CG_SEQ_RESP(x)				((x) << 8)
+#define		CG_SEQ_RESP_MASK			(0xff << 8)
+#define		CG_SEQ_RESP_SHIFT			8
+#define		SEQ_CG_REQ(x)				((x) << 16)
+#define		SEQ_CG_REQ_MASK				(0xff << 16)
+#define		SEQ_CG_REQ_SHIFT			16
+#define		SEQ_CG_RESP(x)				((x) << 24)
+#define		SEQ_CG_RESP_MASK			(0xff << 24)
+#define		SEQ_CG_RESP_SHIFT			24
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+
+#define CGTS_SM_CTRL_REG                                0x9150
+
 /* Registers */
 
 #define RCU_IND_INDEX           			0x100
@@ -90,6 +377,34 @@
 #define CG_VCLK_STATUS                                  0x61c
 #define	CG_SCRATCH1					0x820
 
+#define RLC_CNTL                                        0x3f00
+#       define RLC_ENABLE                               (1 << 0)
+#       define GFX_POWER_GATING_ENABLE                  (1 << 7)
+#       define GFX_POWER_GATING_SRC                     (1 << 8)
+#       define DYN_PER_SIMD_PG_ENABLE                   (1 << 27)
+#       define LB_CNT_SPIM_ACTIVE                       (1 << 30)
+#       define LOAD_BALANCE_ENABLE                      (1 << 31)
+
+#define RLC_HB_BASE                                       0x3f10
+#define RLC_HB_CNTL                                       0x3f0c
+#define RLC_HB_RPTR                                       0x3f20
+#define RLC_HB_WPTR                                       0x3f1c
+#define RLC_HB_WPTR_LSB_ADDR                              0x3f14
+#define RLC_HB_WPTR_MSB_ADDR                              0x3f18
+#define RLC_MC_CNTL                                       0x3f44
+#define RLC_UCODE_CNTL                                    0x3f48
+#define RLC_UCODE_ADDR                                    0x3f2c
+#define RLC_UCODE_DATA                                    0x3f30
+
+/* new for TN */
+#define TN_RLC_SAVE_AND_RESTORE_BASE                      0x3f10
+#define TN_RLC_LB_CNTR_MAX                                0x3f14
+#define TN_RLC_LB_CNTR_INIT                               0x3f18
+#define TN_RLC_CLEAR_STATE_RESTORE_BASE                   0x3f20
+#define TN_RLC_LB_INIT_SIMD_MASK                          0x3fe4
+#define TN_RLC_LB_ALWAYS_ACTIVE_SIMD_MASK                 0x3fe8
+#define TN_RLC_LB_PARAMS                                  0x3fec
+
 #define GRBM_GFX_INDEX          			0x802C
 #define		INSTANCE_INDEX(x)			((x) << 0)
 #define		SE_INDEX(x)     			((x) << 16)
@@ -503,6 +818,30 @@
 #define	CG_THERMAL_CTRL					0x72c
 #define		TOFFSET_MASK			        0x00003FE0
 #define		TOFFSET_SHIFT			        5
+#define		DIG_THERM_DPM(x)			((x) << 14)
+#define		DIG_THERM_DPM_MASK			0x003FC000
+#define		DIG_THERM_DPM_SHIFT			14
+
+#define	CG_THERMAL_INT					0x734
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+
+#define	TN_CG_THERMAL_INT_CTRL				0x738
+#define		TN_DIG_THERM_INTH(x)			((x) << 0)
+#define		TN_DIG_THERM_INTH_MASK			0x000000FF
+#define		TN_DIG_THERM_INTH_SHIFT			0
+#define		TN_DIG_THERM_INTL(x)			((x) << 8)
+#define		TN_DIG_THERM_INTL_MASK			0x0000FF00
+#define		TN_DIG_THERM_INTL_SHIFT			8
+#define 	TN_THERM_INT_MASK_HIGH			(1 << 24)
+#define 	TN_THERM_INT_MASK_LOW			(1 << 25)
+
 #define	CG_MULT_THERMAL_STATUS				0x740
 #define		ASIC_T(x)			        ((x) << 16)
 #define		ASIC_T_MASK			        0x07FF0000
@@ -510,6 +849,7 @@
 #define	CG_TS0_STATUS					0x760
 #define		TS0_ADC_DOUT_MASK			0x000003FF
 #define		TS0_ADC_DOUT_SHIFT			0
+
 /* APU */
 #define	CG_THERMAL_STATUS			        0x678
 
@@ -992,7 +1332,48 @@
 #define	DMA_PACKET_CONSTANT_FILL                0xd
 #define	DMA_PACKET_NOP                          0xf
 
-/* PCIE link stuff */
+/* PIF PHY0 indirect regs */
+#define PB0_PIF_CNTL                                      0x10
+#       define LS2_EXIT_TIME(x)                           ((x) << 17)
+#       define LS2_EXIT_TIME_MASK                         (0x7 << 17)
+#       define LS2_EXIT_TIME_SHIFT                        17
+#define PB0_PIF_PAIRING                                   0x11
+#       define MULTI_PIF                                  (1 << 25)
+#define PB0_PIF_PWRDOWN_0                                 0x12
+#       define PLL_POWER_STATE_IN_TXS2_0(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_0(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_SHIFT             10
+#       define PLL_RAMP_UP_TIME_0(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_0_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_0_SHIFT                   24
+#define PB0_PIF_PWRDOWN_1                                 0x13
+#       define PLL_POWER_STATE_IN_TXS2_1(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_1(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_SHIFT             10
+#       define PLL_RAMP_UP_TIME_1(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_1_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_1_SHIFT                   24
+/* PIF PHY1 indirect regs */
+#define PB1_PIF_CNTL                                      0x10
+#define PB1_PIF_PAIRING                                   0x11
+#define PB1_PIF_PWRDOWN_0                                 0x12
+#define PB1_PIF_PWRDOWN_1                                 0x13
+/* PCIE PORT indirect regs */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
 #define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
 #define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
 #       define LC_LINK_WIDTH_SHIFT                        0
@@ -1012,6 +1393,9 @@
 #       define LC_SHORT_RECONFIG_EN                       (1 << 11)
 #       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
 #       define LC_UPCONFIGURE_DIS                         (1 << 13)
+#       define LC_DYN_LANES_PWR_STATE(x)                  ((x) << 21)
+#       define LC_DYN_LANES_PWR_STATE_MASK                (0x3 << 21)
+#       define LC_DYN_LANES_PWR_STATE_SHIFT               21
 #define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
 #       define LC_GEN2_EN_STRAP                           (1 << 0)
 #       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 1)
@@ -1020,6 +1404,9 @@
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
 #       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
 #       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
 #       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
 #       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 8458330..f30127c 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -33,6 +33,135 @@
 #include "atom.h"
 #include "ni_reg.h"
 #include "cayman_blit_shaders.h"
+#include "radeon_ucode.h"
+#include "clearstate_cayman.h"
+
+static u32 tn_rlc_save_restore_register_list[] =
+{
+	0x98fc,
+	0x98f0,
+	0x9834,
+	0x9838,
+	0x9870,
+	0x9874,
+	0x8a14,
+	0x8b24,
+	0x8bcc,
+	0x8b10,
+	0x8c30,
+	0x8d00,
+	0x8d04,
+	0x8c00,
+	0x8c04,
+	0x8c10,
+	0x8c14,
+	0x8d8c,
+	0x8cf0,
+	0x8e38,
+	0x9508,
+	0x9688,
+	0x9608,
+	0x960c,
+	0x9610,
+	0x9614,
+	0x88c4,
+	0x8978,
+	0x88d4,
+	0x900c,
+	0x9100,
+	0x913c,
+	0x90e8,
+	0x9354,
+	0xa008,
+	0x98f8,
+	0x9148,
+	0x914c,
+	0x3f94,
+	0x98f4,
+	0x9b7c,
+	0x3f8c,
+	0x8950,
+	0x8954,
+	0x8a18,
+	0x8b28,
+	0x9144,
+	0x3f90,
+	0x915c,
+	0x9160,
+	0x9178,
+	0x917c,
+	0x9180,
+	0x918c,
+	0x9190,
+	0x9194,
+	0x9198,
+	0x919c,
+	0x91a8,
+	0x91ac,
+	0x91b0,
+	0x91b4,
+	0x91b8,
+	0x91c4,
+	0x91c8,
+	0x91cc,
+	0x91d0,
+	0x91d4,
+	0x91e0,
+	0x91e4,
+	0x91ec,
+	0x91f0,
+	0x91f4,
+	0x9200,
+	0x9204,
+	0x929c,
+	0x8030,
+	0x9150,
+	0x9a60,
+	0x920c,
+	0x9210,
+	0x9228,
+	0x922c,
+	0x9244,
+	0x9248,
+	0x91e8,
+	0x9294,
+	0x9208,
+	0x9224,
+	0x9240,
+	0x9220,
+	0x923c,
+	0x9258,
+	0x9744,
+	0xa200,
+	0xa204,
+	0xa208,
+	0xa20c,
+	0x8d58,
+	0x9030,
+	0x9034,
+	0x9038,
+	0x903c,
+	0x9040,
+	0x9654,
+	0x897c,
+	0xa210,
+	0xa214,
+	0x9868,
+	0xa02c,
+	0x9664,
+	0x9698,
+	0x949c,
+	0x8e10,
+	0x8e18,
+	0x8c50,
+	0x8c58,
+	0x8c60,
+	0x8c68,
+	0x89b4,
+	0x9830,
+	0x802c,
+};
+static u32 tn_rlc_save_restore_register_list_size = ARRAY_SIZE(tn_rlc_save_restore_register_list);
 
 extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
@@ -44,36 +173,29 @@ extern void evergreen_irq_suspend(struct radeon_device *rdev);
 extern int evergreen_mc_init(struct radeon_device *rdev);
 extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
 extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
-extern void si_rlc_fini(struct radeon_device *rdev);
-extern int si_rlc_init(struct radeon_device *rdev);
-
-#define EVERGREEN_PFP_UCODE_SIZE 1120
-#define EVERGREEN_PM4_UCODE_SIZE 1376
-#define EVERGREEN_RLC_UCODE_SIZE 768
-#define BTC_MC_UCODE_SIZE 6024
-
-#define CAYMAN_PFP_UCODE_SIZE 2176
-#define CAYMAN_PM4_UCODE_SIZE 2176
-#define CAYMAN_RLC_UCODE_SIZE 1024
-#define CAYMAN_MC_UCODE_SIZE 6037
-
-#define ARUBA_RLC_UCODE_SIZE 1536
+extern void evergreen_program_aspm(struct radeon_device *rdev);
+extern void sumo_rlc_fini(struct radeon_device *rdev);
+extern int sumo_rlc_init(struct radeon_device *rdev);
 
 /* Firmware Names */
 MODULE_FIRMWARE("radeon/BARTS_pfp.bin");
 MODULE_FIRMWARE("radeon/BARTS_me.bin");
 MODULE_FIRMWARE("radeon/BARTS_mc.bin");
+MODULE_FIRMWARE("radeon/BARTS_smc.bin");
 MODULE_FIRMWARE("radeon/BTC_rlc.bin");
 MODULE_FIRMWARE("radeon/TURKS_pfp.bin");
 MODULE_FIRMWARE("radeon/TURKS_me.bin");
 MODULE_FIRMWARE("radeon/TURKS_mc.bin");
+MODULE_FIRMWARE("radeon/TURKS_smc.bin");
 MODULE_FIRMWARE("radeon/CAICOS_pfp.bin");
 MODULE_FIRMWARE("radeon/CAICOS_me.bin");
 MODULE_FIRMWARE("radeon/CAICOS_mc.bin");
+MODULE_FIRMWARE("radeon/CAICOS_smc.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_pfp.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_me.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_mc.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_rlc.bin");
+MODULE_FIRMWARE("radeon/CAYMAN_smc.bin");
 MODULE_FIRMWARE("radeon/ARUBA_pfp.bin");
 MODULE_FIRMWARE("radeon/ARUBA_me.bin");
 MODULE_FIRMWARE("radeon/ARUBA_rlc.bin");
@@ -566,6 +688,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 	const char *chip_name;
 	const char *rlc_chip_name;
 	size_t pfp_req_size, me_req_size, rlc_req_size, mc_req_size;
+	size_t smc_req_size = 0;
 	char fw_name[30];
 	int err;
 
@@ -586,6 +709,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = EVERGREEN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = EVERGREEN_RLC_UCODE_SIZE * 4;
 		mc_req_size = BTC_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(BARTS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_TURKS:
 		chip_name = "TURKS";
@@ -594,6 +718,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = EVERGREEN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = EVERGREEN_RLC_UCODE_SIZE * 4;
 		mc_req_size = BTC_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(TURKS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CAICOS:
 		chip_name = "CAICOS";
@@ -602,6 +727,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = EVERGREEN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = EVERGREEN_RLC_UCODE_SIZE * 4;
 		mc_req_size = BTC_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(CAICOS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CAYMAN:
 		chip_name = "CAYMAN";
@@ -610,6 +736,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = CAYMAN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = CAYMAN_RLC_UCODE_SIZE * 4;
 		mc_req_size = CAYMAN_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(CAYMAN_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_ARUBA:
 		chip_name = "ARUBA";
@@ -672,6 +799,20 @@ int ni_init_microcode(struct radeon_device *rdev)
 			err = -EINVAL;
 		}
 	}
+
+	if ((rdev->family >= CHIP_BARTS) && (rdev->family <= CHIP_CAYMAN)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+		err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev);
+		if (err)
+			goto out;
+		if (rdev->smc_fw->size != smc_req_size) {
+			printk(KERN_ERR
+			       "ni_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	}
+
 out:
 	platform_device_unregister(pdev);
 
@@ -692,6 +833,14 @@ out:
 	return err;
 }
 
+int tn_get_temp(struct radeon_device *rdev)
+{
+	u32 temp = RREG32_SMC(TN_CURRENT_GNB_TEMP) & 0x7ff;
+	int actual_temp = (temp / 8) - 49;
+
+	return actual_temp * 1000;
+}
+
 /*
  * Core functions
  */
@@ -1027,6 +1176,16 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
 
 	udelay(50);
+
+	/* set clockgating golden values on TN */
+	if (rdev->family == CHIP_ARUBA) {
+		tmp = RREG32_CG(CG_CGTT_LOCAL_0);
+		tmp &= ~0x00380000;
+		WREG32_CG(CG_CGTT_LOCAL_0, tmp);
+                tmp = RREG32_CG(CG_CGTT_LOCAL_1);
+		tmp &= ~0x0e000000;
+		WREG32_CG(CG_CGTT_LOCAL_1, tmp);
+	}
 }
 
 /*
@@ -1928,6 +2087,8 @@ static int cayman_startup(struct radeon_device *rdev)
 
 	/* enable pcie gen2 link */
 	evergreen_pcie_gen2_enable(rdev);
+	/* enable aspm */
+	evergreen_program_aspm(rdev);
 
 	if (rdev->flags & RADEON_IS_IGP) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) {
@@ -1972,7 +2133,10 @@ static int cayman_startup(struct radeon_device *rdev)
 
 	/* allocate rlc buffers */
 	if (rdev->flags & RADEON_IS_IGP) {
-		r = si_rlc_init(rdev);
+		rdev->rlc.reg_list = tn_rlc_save_restore_register_list;
+		rdev->rlc.reg_list_size = tn_rlc_save_restore_register_list_size;
+		rdev->rlc.cs_data = cayman_cs_data;
+		r = sumo_rlc_init(rdev);
 		if (r) {
 			DRM_ERROR("Failed to init rlc BOs!\n");
 			return r;
@@ -2229,7 +2393,7 @@ int cayman_init(struct radeon_device *rdev)
 		cayman_dma_fini(rdev);
 		r600_irq_fini(rdev);
 		if (rdev->flags & RADEON_IS_IGP)
-			si_rlc_fini(rdev);
+			sumo_rlc_fini(rdev);
 		radeon_wb_fini(rdev);
 		radeon_ib_pool_fini(rdev);
 		radeon_vm_manager_fini(rdev);
@@ -2260,7 +2424,7 @@ void cayman_fini(struct radeon_device *rdev)
 	cayman_dma_fini(rdev);
 	r600_irq_fini(rdev);
 	if (rdev->flags & RADEON_IS_IGP)
-		si_rlc_fini(rdev);
+		sumo_rlc_fini(rdev);
 	radeon_wb_fini(rdev);
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c
new file mode 100644
index 0000000..559cf24
--- /dev/null
+++ b/drivers/gpu/drm/radeon/ni_dpm.c
@@ -0,0 +1,4370 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "nid.h"
+#include "r600_dpm.h"
+#include "ni_dpm.h"
+#include "atom.h"
+#include <linux/math64.h>
+#include <linux/seq_file.h>
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define SMC_RAM_END 0xC000
+
+static const struct ni_cac_weights cac_weights_cayman_xt =
+{
+	0x15,
+	0x2,
+	0x19,
+	0x2,
+	0x8,
+	0x14,
+	0x2,
+	0x16,
+	0xE,
+	0x17,
+	0x13,
+	0x2B,
+	0x10,
+	0x7,
+	0x5,
+	0x5,
+	0x5,
+	0x2,
+	0x3,
+	0x9,
+	0x10,
+	0x10,
+	0x2B,
+	0xA,
+	0x9,
+	0x4,
+	0xD,
+	0xD,
+	0x3E,
+	0x18,
+	0x14,
+	0,
+	0x3,
+	0x3,
+	0x5,
+	0,
+	0x2,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0x1CC,
+	0,
+	0x164,
+	1,
+	1,
+	1,
+	1,
+	12,
+	12,
+	12,
+	0x12,
+	0x1F,
+	132,
+	5,
+	7,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0 },
+	true
+};
+
+static const struct ni_cac_weights cac_weights_cayman_pro =
+{
+	0x16,
+	0x4,
+	0x10,
+	0x2,
+	0xA,
+	0x16,
+	0x2,
+	0x18,
+	0x10,
+	0x1A,
+	0x16,
+	0x2D,
+	0x12,
+	0xA,
+	0x6,
+	0x6,
+	0x6,
+	0x2,
+	0x4,
+	0xB,
+	0x11,
+	0x11,
+	0x2D,
+	0xC,
+	0xC,
+	0x7,
+	0x10,
+	0x10,
+	0x3F,
+	0x1A,
+	0x16,
+	0,
+	0x7,
+	0x4,
+	0x6,
+	1,
+	0x2,
+	0x1,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0x30,
+	0,
+	0x1CF,
+	0,
+	0x166,
+	1,
+	1,
+	1,
+	1,
+	12,
+	12,
+	12,
+	0x15,
+	0x1F,
+	132,
+	6,
+	6,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0 },
+	true
+};
+
+static const struct ni_cac_weights cac_weights_cayman_le =
+{
+	0x7,
+	0xE,
+	0x1,
+	0xA,
+	0x1,
+	0x3F,
+	0x2,
+	0x18,
+	0x10,
+	0x1A,
+	0x1,
+	0x3F,
+	0x1,
+	0xE,
+	0x6,
+	0x6,
+	0x6,
+	0x2,
+	0x4,
+	0x9,
+	0x1A,
+	0x1A,
+	0x2C,
+	0xA,
+	0x11,
+	0x8,
+	0x19,
+	0x19,
+	0x1,
+	0x1,
+	0x1A,
+	0,
+	0x8,
+	0x5,
+	0x8,
+	0x1,
+	0x3,
+	0x1,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0x38,
+	0x38,
+	0x239,
+	0x3,
+	0x18A,
+	1,
+	1,
+	1,
+	1,
+	12,
+	12,
+	12,
+	0x15,
+	0x22,
+	132,
+	6,
+	6,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0 },
+	true
+};
+
+#define NISLANDS_MGCG_SEQUENCE  300
+
+static const u32 cayman_cgcg_cgls_default[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_CGCG_CGLS_DEFAULT_LENGTH sizeof(cayman_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 cayman_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7902, 0x001f4180,
+	0x00000644, 0x000f3802, 0x001f4180
+};
+#define CAYMAN_CGCG_CGLS_DISABLE_LENGTH sizeof(cayman_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 cayman_cgcg_cgls_enable[] =
+{
+	0x00000644, 0x000f7882, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define CAYMAN_CGCG_CGLS_ENABLE_LENGTH  sizeof(cayman_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+static const u32 cayman_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00003fc4, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00000100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x00008984, 0x06000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00800200, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x00009744, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009664, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000104, 0xffffffff,
+	0x0000d0c0, 0x00000100, 0xffffffff,
+	0x0000d8c0, 0x00000100, 0xffffffff,
+	0x0000802c, 0x40000000, 0xffffffff,
+	0x00003fc4, 0x40000000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009240, 0x00070000, 0xffffffff,
+	0x00009244, 0x00030002, 0xffffffff,
+	0x00009248, 0x00050004, 0xffffffff,
+	0x00009254, 0x00010006, 0xffffffff,
+	0x00009258, 0x00090008, 0xffffffff,
+	0x0000925c, 0x00070000, 0xffffffff,
+	0x00009260, 0x00030002, 0xffffffff,
+	0x00009264, 0x00050004, 0xffffffff,
+	0x00009270, 0x00010006, 0xffffffff,
+	0x00009274, 0x00090008, 0xffffffff,
+	0x00009278, 0x00070000, 0xffffffff,
+	0x0000927c, 0x00030002, 0xffffffff,
+	0x00009280, 0x00050004, 0xffffffff,
+	0x0000928c, 0x00010006, 0xffffffff,
+	0x00009290, 0x00090008, 0xffffffff,
+	0x000092a8, 0x00070000, 0xffffffff,
+	0x000092ac, 0x00030002, 0xffffffff,
+	0x000092b0, 0x00050004, 0xffffffff,
+	0x000092bc, 0x00010006, 0xffffffff,
+	0x000092c0, 0x00090008, 0xffffffff,
+	0x000092c4, 0x00070000, 0xffffffff,
+	0x000092c8, 0x00030002, 0xffffffff,
+	0x000092cc, 0x00050004, 0xffffffff,
+	0x000092d8, 0x00010006, 0xffffffff,
+	0x000092dc, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0x40010000, 0xffffffff,
+	0x00003fc4, 0x40010000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009240, 0x00070000, 0xffffffff,
+	0x00009244, 0x00030002, 0xffffffff,
+	0x00009248, 0x00050004, 0xffffffff,
+	0x00009254, 0x00010006, 0xffffffff,
+	0x00009258, 0x00090008, 0xffffffff,
+	0x0000925c, 0x00070000, 0xffffffff,
+	0x00009260, 0x00030002, 0xffffffff,
+	0x00009264, 0x00050004, 0xffffffff,
+	0x00009270, 0x00010006, 0xffffffff,
+	0x00009274, 0x00090008, 0xffffffff,
+	0x00009278, 0x00070000, 0xffffffff,
+	0x0000927c, 0x00030002, 0xffffffff,
+	0x00009280, 0x00050004, 0xffffffff,
+	0x0000928c, 0x00010006, 0xffffffff,
+	0x00009290, 0x00090008, 0xffffffff,
+	0x000092a8, 0x00070000, 0xffffffff,
+	0x000092ac, 0x00030002, 0xffffffff,
+	0x000092b0, 0x00050004, 0xffffffff,
+	0x000092bc, 0x00010006, 0xffffffff,
+	0x000092c0, 0x00090008, 0xffffffff,
+	0x000092c4, 0x00070000, 0xffffffff,
+	0x000092c8, 0x00030002, 0xffffffff,
+	0x000092cc, 0x00050004, 0xffffffff,
+	0x000092d8, 0x00010006, 0xffffffff,
+	0x000092dc, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00003fc4, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_MGCG_DEFAULT_LENGTH sizeof(cayman_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 cayman_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define CAYMAN_MGCG_DISABLE_LENGTH   sizeof(cayman_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 cayman_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00600000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x96944200, 0xffffffff
+};
+
+#define CAYMAN_MGCG_ENABLE_LENGTH   sizeof(cayman_mgcg_enable) / (3 * sizeof(u32))
+
+#define NISLANDS_SYSLS_SEQUENCE  100
+
+static const u32 cayman_sysls_default[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00008dfc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_SYSLS_DEFAULT_LENGTH sizeof(cayman_sysls_default) / (3 * sizeof(u32))
+
+static const u32 cayman_sysls_disable[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x0000d0c0, 0x00000000, 0xffffffff,
+	0x0000d8c0, 0x00000000, 0xffffffff,
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00008dfc, 0x0000007f, 0xffffffff
+};
+#define CAYMAN_SYSLS_DISABLE_LENGTH sizeof(cayman_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 cayman_sysls_enable[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x0000d8bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000903, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00008dfc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_SYSLS_ENABLE_LENGTH sizeof(cayman_sysls_enable) / (3 * sizeof(u32))
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+
+struct ni_power_info *ni_get_pi(struct radeon_device *rdev)
+{
+        struct ni_power_info *pi = rdev->pm.dpm.priv;
+
+        return pi;
+}
+
+struct ni_ps *ni_get_ps(struct radeon_ps *rps)
+{
+	struct ni_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+static void ni_calculate_leakage_for_v_and_t_formula(const struct ni_leakage_coeffients *coeff,
+						     u16 v, s32 t,
+						     u32 ileakage,
+						     u32 *leakage)
+{
+	s64 kt, kv, leakage_w, i_leakage, vddc, temperature;
+
+	i_leakage = div64_s64(drm_int2fixp(ileakage), 1000);
+	vddc = div64_s64(drm_int2fixp(v), 1000);
+	temperature = div64_s64(drm_int2fixp(t), 1000);
+
+	kt = drm_fixp_mul(div64_s64(drm_int2fixp(coeff->at), 1000),
+			  drm_fixp_exp(drm_fixp_mul(div64_s64(drm_int2fixp(coeff->bt), 1000), temperature)));
+	kv = drm_fixp_mul(div64_s64(drm_int2fixp(coeff->av), 1000),
+			  drm_fixp_exp(drm_fixp_mul(div64_s64(drm_int2fixp(coeff->bv), 1000), vddc)));
+
+	leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc);
+
+	*leakage = drm_fixp2int(leakage_w * 1000);
+}
+
+static void ni_calculate_leakage_for_v_and_t(struct radeon_device *rdev,
+					     const struct ni_leakage_coeffients *coeff,
+					     u16 v,
+					     s32 t,
+					     u32 i_leakage,
+					     u32 *leakage)
+{
+	ni_calculate_leakage_for_v_and_t_formula(coeff, v, t, i_leakage, leakage);
+}
+
+bool ni_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+	u32 switch_limit = pi->mem_gddr5 ? 450 : 300;
+
+	if (vblank_time < switch_limit)
+		return true;
+	else
+		return false;
+
+}
+
+static void ni_apply_state_adjust_rules(struct radeon_device *rdev,
+					struct radeon_ps *rps)
+{
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct radeon_clock_and_voltage_limits *max_limits;
+	bool disable_mclk_switching;
+	u32 mclk, sclk;
+	u16 vddc, vddci;
+	int i;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
+	    ni_dpm_vblank_too_short(rdev))
+		disable_mclk_switching = true;
+	else
+		disable_mclk_switching = false;
+
+	if (rdev->pm.dpm.ac_power)
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
+	else
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc;
+
+	if (rdev->pm.dpm.ac_power == false) {
+		for (i = 0; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk > max_limits->mclk)
+				ps->performance_levels[i].mclk = max_limits->mclk;
+			if (ps->performance_levels[i].sclk > max_limits->sclk)
+				ps->performance_levels[i].sclk = max_limits->sclk;
+			if (ps->performance_levels[i].vddc > max_limits->vddc)
+				ps->performance_levels[i].vddc = max_limits->vddc;
+			if (ps->performance_levels[i].vddci > max_limits->vddci)
+				ps->performance_levels[i].vddci = max_limits->vddci;
+		}
+	}
+
+	/* XXX validate the min clocks required for display */
+
+	if (disable_mclk_switching) {
+		mclk  = ps->performance_levels[ps->performance_level_count - 1].mclk;
+		sclk = ps->performance_levels[0].sclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[ps->performance_level_count - 1].vddci;
+	} else {
+		sclk = ps->performance_levels[0].sclk;
+		mclk = ps->performance_levels[0].mclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[0].vddci;
+	}
+
+	/* adjusted low state */
+	ps->performance_levels[0].sclk = sclk;
+	ps->performance_levels[0].mclk = mclk;
+	ps->performance_levels[0].vddc = vddc;
+	ps->performance_levels[0].vddci = vddci;
+
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->performance_levels[0].sclk,
+				  &ps->performance_levels[0].mclk);
+
+	for (i = 1; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk)
+			ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk;
+		if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc)
+			ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc;
+	}
+
+	if (disable_mclk_switching) {
+		mclk = ps->performance_levels[0].mclk;
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (mclk < ps->performance_levels[i].mclk)
+				mclk = ps->performance_levels[i].mclk;
+		}
+		for (i = 0; i < ps->performance_level_count; i++) {
+			ps->performance_levels[i].mclk = mclk;
+			ps->performance_levels[i].vddci = vddci;
+		}
+	} else {
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk < ps->performance_levels[i - 1].mclk)
+				ps->performance_levels[i].mclk = ps->performance_levels[i - 1].mclk;
+			if (ps->performance_levels[i].vddci < ps->performance_levels[i - 1].vddci)
+				ps->performance_levels[i].vddci = ps->performance_levels[i - 1].vddci;
+		}
+	}
+
+	for (i = 1; i < ps->performance_level_count; i++)
+		btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+					  &ps->performance_levels[i].sclk,
+					  &ps->performance_levels[i].mclk);
+
+	for (i = 0; i < ps->performance_level_count; i++)
+		btc_adjust_clock_combinations(rdev, max_limits,
+					      &ps->performance_levels[i]);
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+						   ps->performance_levels[i].sclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddci, &ps->performance_levels[i].vddci);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+						   rdev->clock.current_dispclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+	}
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_delta_rules(rdev,
+					      max_limits->vddc, max_limits->vddci,
+					      &ps->performance_levels[i].vddc,
+					      &ps->performance_levels[i].vddci);
+	}
+
+	ps->dc_compatible = true;
+	for (i = 0; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].vddc > rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc)
+			ps->dc_compatible = false;
+
+		if (ps->performance_levels[i].vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+			ps->performance_levels[i].flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+	}
+}
+
+static void ni_cg_clockgating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	ps = (const u32 *)&cayman_cgcg_cgls_default;
+	count = CAYMAN_CGCG_CGLS_DEFAULT_LENGTH;
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_gfx_clockgating_enable(struct radeon_device *rdev,
+				      bool enable)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	if (enable) {
+		ps = (const u32 *)&cayman_cgcg_cgls_enable;
+		count = CAYMAN_CGCG_CGLS_ENABLE_LENGTH;
+	} else {
+		ps = (const u32 *)&cayman_cgcg_cgls_disable;
+		count = CAYMAN_CGCG_CGLS_DISABLE_LENGTH;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_mg_clockgating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	ps = (const u32 *)&cayman_mgcg_default;
+	count = CAYMAN_MGCG_DEFAULT_LENGTH;
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_mg_clockgating_enable(struct radeon_device *rdev,
+				     bool enable)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	if (enable) {
+		ps = (const u32 *)&cayman_mgcg_enable;
+		count = CAYMAN_MGCG_ENABLE_LENGTH;
+	} else {
+		ps = (const u32 *)&cayman_mgcg_disable;
+		count = CAYMAN_MGCG_DISABLE_LENGTH;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_ls_clockgating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	ps = (const u32 *)&cayman_sysls_default;
+	count = CAYMAN_SYSLS_DEFAULT_LENGTH;
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_ls_clockgating_enable(struct radeon_device *rdev,
+				     bool enable)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	if (enable) {
+		ps = (const u32 *)&cayman_sysls_enable;
+		count = CAYMAN_SYSLS_ENABLE_LENGTH;
+	} else {
+		ps = (const u32 *)&cayman_sysls_disable;
+		count = CAYMAN_SYSLS_DISABLE_LENGTH;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+
+}
+
+static int ni_patch_single_dependency_table_based_on_leakage(struct radeon_device *rdev,
+							     struct radeon_clock_voltage_dependency_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 i;
+
+	if (table) {
+		for (i = 0; i < table->count; i++) {
+			if (0xff01 == table->entries[i].v) {
+				if (pi->max_vddc == 0)
+					return -EINVAL;
+				table->entries[i].v = pi->max_vddc;
+			}
+		}
+	}
+	return 0;
+}
+
+static int ni_patch_dependency_tables_based_on_leakage(struct radeon_device *rdev)
+{
+	int ret = 0;
+
+	ret = ni_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk);
+
+	ret = ni_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk);
+	return ret;
+}
+
+static void ni_stop_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+}
+
+#if 0
+static int ni_notify_hw_of_power_source(struct radeon_device *rdev,
+					bool ac_power)
+{
+	if (ac_power)
+		return (rv770_send_msg_to_smc(rdev, PPSMC_MSG_RunningOnAC) == PPSMC_Result_OK) ?
+			0 : -EINVAL;
+
+	return 0;
+}
+#endif
+
+static PPSMC_Result ni_send_msg_to_smc_with_parameter(struct radeon_device *rdev,
+						      PPSMC_Msg msg, u32 parameter)
+{
+	WREG32(SMC_SCRATCH0, parameter);
+	return rv770_send_msg_to_smc(rdev, msg);
+}
+
+static int ni_restrict_performance_levels_before_switch(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 1) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+int ni_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	u32 levels;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 1) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		levels = ps->performance_level_count - 1;
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) {
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+static void ni_stop_smc(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(LB_SYNC_RESET_SEL) & LB_SYNC_RESET_SEL_MASK;
+		if (tmp != 1)
+			break;
+		udelay(1);
+	}
+
+	udelay(100);
+
+	r7xx_stop_smc(rdev);
+}
+
+static int ni_process_firmware_header(struct radeon_device *rdev)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_stateTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	pi->state_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_softRegisters,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	pi->soft_regs_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	eg_pi->mc_reg_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_fanTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->fan_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->arb_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_cacTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->cac_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_spllTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->spll_table_start = (u16)tmp;
+
+
+	return ret;
+}
+
+static void ni_read_clock_registers(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	ni_pi->clock_registers.cg_spll_func_cntl = RREG32(CG_SPLL_FUNC_CNTL);
+	ni_pi->clock_registers.cg_spll_func_cntl_2 = RREG32(CG_SPLL_FUNC_CNTL_2);
+	ni_pi->clock_registers.cg_spll_func_cntl_3 = RREG32(CG_SPLL_FUNC_CNTL_3);
+	ni_pi->clock_registers.cg_spll_func_cntl_4 = RREG32(CG_SPLL_FUNC_CNTL_4);
+	ni_pi->clock_registers.cg_spll_spread_spectrum = RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	ni_pi->clock_registers.cg_spll_spread_spectrum_2 = RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+	ni_pi->clock_registers.mpll_ad_func_cntl = RREG32(MPLL_AD_FUNC_CNTL);
+	ni_pi->clock_registers.mpll_ad_func_cntl_2 = RREG32(MPLL_AD_FUNC_CNTL_2);
+	ni_pi->clock_registers.mpll_dq_func_cntl = RREG32(MPLL_DQ_FUNC_CNTL);
+	ni_pi->clock_registers.mpll_dq_func_cntl_2 = RREG32(MPLL_DQ_FUNC_CNTL_2);
+	ni_pi->clock_registers.mclk_pwrmgt_cntl = RREG32(MCLK_PWRMGT_CNTL);
+	ni_pi->clock_registers.dll_cntl = RREG32(DLL_CNTL);
+	ni_pi->clock_registers.mpll_ss1 = RREG32(MPLL_SS1);
+	ni_pi->clock_registers.mpll_ss2 = RREG32(MPLL_SS2);
+}
+
+#if 0
+static int ni_enter_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->gfx_clock_gating) {
+                WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+                WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+        }
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_SwitchToMinimumPower),
+                 ~HOST_SMC_MSG_MASK);
+
+	udelay(25000);
+
+	return 0;
+}
+#endif
+
+static void ni_program_response_times(struct radeon_device *rdev)
+{
+	u32 voltage_response_time, backbias_response_time, acpi_delay_time, vbi_time_out;
+	u32 vddc_dly, bb_dly, acpi_dly, vbi_dly, mclk_switch_limit;
+	u32 reference_clock;
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mvdd_chg_time, 1);
+
+	voltage_response_time = (u32)rdev->pm.dpm.voltage_response_time;
+	backbias_response_time = (u32)rdev->pm.dpm.backbias_response_time;
+
+	if (voltage_response_time == 0)
+		voltage_response_time = 1000;
+
+	if (backbias_response_time == 0)
+		backbias_response_time = 1000;
+
+	acpi_delay_time = 15000;
+	vbi_time_out = 100000;
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	vddc_dly = (voltage_response_time  * reference_clock) / 1600;
+	bb_dly   = (backbias_response_time * reference_clock) / 1600;
+	acpi_dly = (acpi_delay_time * reference_clock) / 1600;
+	vbi_dly  = (vbi_time_out * reference_clock) / 1600;
+
+	mclk_switch_limit = (460 * reference_clock) / 100;
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_delay_vreg,  vddc_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_delay_bbias, bb_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_delay_acpi,  acpi_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mclk_chg_timeout, vbi_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mc_block_delay, 0xAA);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mclk_switch_lim, mclk_switch_limit);
+}
+
+static void ni_populate_smc_voltage_table(struct radeon_device *rdev,
+					  struct atom_voltage_table *voltage_table,
+					  NISLANDS_SMC_STATETABLE *table)
+{
+	unsigned int i;
+
+	for (i = 0; i < voltage_table->count; i++) {
+		table->highSMIO[i] = 0;
+		table->lowSMIO[i] |= cpu_to_be32(voltage_table->entries[i].smio_low);
+	}
+}
+
+static void ni_populate_smc_voltage_tables(struct radeon_device *rdev,
+					   NISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	unsigned char i;
+
+	if (eg_pi->vddc_voltage_table.count) {
+		ni_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
+		table->voltageMaskTable.highMask[NISLANDS_SMC_VOLTAGEMASK_VDDC] = 0;
+		table->voltageMaskTable.lowMask[NISLANDS_SMC_VOLTAGEMASK_VDDC] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+			if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
+				table->maxVDDCIndexInPPTable = i;
+				break;
+			}
+		}
+	}
+
+	if (eg_pi->vddci_voltage_table.count) {
+		ni_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
+
+		table->voltageMaskTable.highMask[NISLANDS_SMC_VOLTAGEMASK_VDDCI] = 0;
+		table->voltageMaskTable.lowMask[NISLANDS_SMC_VOLTAGEMASK_VDDCI] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+	}
+}
+
+static int ni_populate_voltage_value(struct radeon_device *rdev,
+				     struct atom_voltage_table *table,
+				     u16 value,
+				     NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (value <= table->entries[i].value) {
+			voltage->index = (u8)i;
+			voltage->value = cpu_to_be16(table->entries[i].value);
+			break;
+		}
+	}
+
+	if (i >= table->count)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void ni_populate_mvdd_value(struct radeon_device *rdev,
+				   u32 mclk,
+				   NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (!pi->mvdd_control) {
+		voltage->index = eg_pi->mvdd_high_index;
+                voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+		return;
+	}
+
+	if (mclk <= pi->mvdd_split_frequency) {
+		voltage->index = eg_pi->mvdd_low_index;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = eg_pi->mvdd_high_index;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+}
+
+static int ni_get_std_voltage_value(struct radeon_device *rdev,
+				    NISLANDS_SMC_VOLTAGE_VALUE *voltage,
+				    u16 *std_voltage)
+{
+	if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries &&
+	    ((u32)voltage->index < rdev->pm.dpm.dyn_state.cac_leakage_table.count))
+		*std_voltage = rdev->pm.dpm.dyn_state.cac_leakage_table.entries[voltage->index].vddc;
+	else
+		*std_voltage = be16_to_cpu(voltage->value);
+
+	return 0;
+}
+
+static void ni_populate_std_voltage_value(struct radeon_device *rdev,
+					  u16 value, u8 index,
+					  NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	voltage->index = index;
+	voltage->value = cpu_to_be16(value);
+}
+
+static u32 ni_get_smc_power_scaling_factor(struct radeon_device *rdev)
+{
+	u32 xclk_period;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 tmp = RREG32(CG_CAC_CTRL) & TID_CNT_MASK;
+
+	xclk_period = (1000000000UL / xclk);
+	xclk_period /= 10000UL;
+
+	return tmp * xclk_period;
+}
+
+static u32 ni_scale_power_for_smc(u32 power_in_watts, u32 scaling_factor)
+{
+	return (power_in_watts * scaling_factor) << 2;
+}
+
+static u32 ni_calculate_power_boost_limit(struct radeon_device *rdev,
+					  struct radeon_ps *radeon_state,
+					  u32 near_tdp_limit)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 power_boost_limit = 0;
+	int ret;
+
+	if (ni_pi->enable_power_containment &&
+	    ni_pi->use_power_boost_limit) {
+		NISLANDS_SMC_VOLTAGE_VALUE vddc;
+		u16 std_vddc_med;
+		u16 std_vddc_high;
+		u64 tmp, n, d;
+
+		if (state->performance_level_count < 3)
+			return 0;
+
+		ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[state->performance_level_count - 2].vddc,
+						&vddc);
+		if (ret)
+			return 0;
+
+		ret = ni_get_std_voltage_value(rdev, &vddc, &std_vddc_med);
+		if (ret)
+			return 0;
+
+		ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[state->performance_level_count - 1].vddc,
+						&vddc);
+		if (ret)
+			return 0;
+
+		ret = ni_get_std_voltage_value(rdev, &vddc, &std_vddc_high);
+		if (ret)
+			return 0;
+
+		n = ((u64)near_tdp_limit * ((u64)std_vddc_med * (u64)std_vddc_med) * 90);
+		d = ((u64)std_vddc_high * (u64)std_vddc_high * 100);
+		tmp = div64_u64(n, d);
+
+		if (tmp >> 32)
+			return 0;
+		power_boost_limit = (u32)tmp;
+	}
+
+	return power_boost_limit;
+}
+
+static int ni_calculate_adjusted_tdp_limits(struct radeon_device *rdev,
+					    bool adjust_polarity,
+					    u32 tdp_adjustment,
+					    u32 *tdp_limit,
+					    u32 *near_tdp_limit)
+{
+	if (tdp_adjustment > (u32)rdev->pm.dpm.tdp_od_limit)
+		return -EINVAL;
+
+	if (adjust_polarity) {
+		*tdp_limit = ((100 + tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		*near_tdp_limit = rdev->pm.dpm.near_tdp_limit + (*tdp_limit - rdev->pm.dpm.tdp_limit);
+	} else {
+		*tdp_limit = ((100 - tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		*near_tdp_limit = rdev->pm.dpm.near_tdp_limit - (rdev->pm.dpm.tdp_limit - *tdp_limit);
+	}
+
+	return 0;
+}
+
+static int ni_populate_smc_tdp_limits(struct radeon_device *rdev,
+				      struct radeon_ps *radeon_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	if (ni_pi->enable_power_containment) {
+		NISLANDS_SMC_STATETABLE *smc_table = &ni_pi->smc_statetable;
+		u32 scaling_factor = ni_get_smc_power_scaling_factor(rdev);
+		u32 tdp_limit;
+		u32 near_tdp_limit;
+		u32 power_boost_limit;
+		int ret;
+
+		if (scaling_factor == 0)
+			return -EINVAL;
+
+		memset(smc_table, 0, sizeof(NISLANDS_SMC_STATETABLE));
+
+		ret = ni_calculate_adjusted_tdp_limits(rdev,
+						       false, /* ??? */
+						       rdev->pm.dpm.tdp_adjustment,
+						       &tdp_limit,
+						       &near_tdp_limit);
+		if (ret)
+			return ret;
+
+		power_boost_limit = ni_calculate_power_boost_limit(rdev, radeon_state,
+								   near_tdp_limit);
+
+		smc_table->dpm2Params.TDPLimit =
+			cpu_to_be32(ni_scale_power_for_smc(tdp_limit, scaling_factor));
+		smc_table->dpm2Params.NearTDPLimit =
+			cpu_to_be32(ni_scale_power_for_smc(near_tdp_limit, scaling_factor));
+		smc_table->dpm2Params.SafePowerLimit =
+			cpu_to_be32(ni_scale_power_for_smc((near_tdp_limit * NISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT) / 100,
+							   scaling_factor));
+		smc_table->dpm2Params.PowerBoostLimit =
+			cpu_to_be32(ni_scale_power_for_smc(power_boost_limit, scaling_factor));
+
+		ret = rv770_copy_bytes_to_smc(rdev,
+					      (u16)(pi->state_table_start + offsetof(NISLANDS_SMC_STATETABLE, dpm2Params) +
+						    offsetof(PP_NIslands_DPM2Parameters, TDPLimit)),
+					      (u8 *)(&smc_table->dpm2Params.TDPLimit),
+					      sizeof(u32) * 4, pi->sram_end);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+int ni_copy_and_switch_arb_sets(struct radeon_device *rdev,
+				u32 arb_freq_src, u32 arb_freq_dest)
+{
+	u32 mc_arb_dram_timing;
+	u32 mc_arb_dram_timing2;
+	u32 burst_time;
+	u32 mc_cg_config;
+
+	switch (arb_freq_src) {
+        case MC_CG_ARB_FREQ_F0:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE0_MASK) >> STATE0_SHIFT;
+		break;
+        case MC_CG_ARB_FREQ_F1:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING_1);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2_1);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE1_MASK) >> STATE1_SHIFT;
+		break;
+        case MC_CG_ARB_FREQ_F2:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING_2);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2_2);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE2_MASK) >> STATE2_SHIFT;
+		break;
+        case MC_CG_ARB_FREQ_F3:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING_3);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2_3);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE3_MASK) >> STATE3_SHIFT;
+		break;
+        default:
+		return -EINVAL;
+	}
+
+	switch (arb_freq_dest) {
+        case MC_CG_ARB_FREQ_F0:
+		WREG32(MC_ARB_DRAM_TIMING, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE0(burst_time), ~STATE0_MASK);
+		break;
+        case MC_CG_ARB_FREQ_F1:
+		WREG32(MC_ARB_DRAM_TIMING_1, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2_1, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE1(burst_time), ~STATE1_MASK);
+		break;
+        case MC_CG_ARB_FREQ_F2:
+		WREG32(MC_ARB_DRAM_TIMING_2, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2_2, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE2(burst_time), ~STATE2_MASK);
+		break;
+        case MC_CG_ARB_FREQ_F3:
+		WREG32(MC_ARB_DRAM_TIMING_3, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2_3, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE3(burst_time), ~STATE3_MASK);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	mc_cg_config = RREG32(MC_CG_CONFIG) | 0x0000000F;
+	WREG32(MC_CG_CONFIG, mc_cg_config);
+	WREG32_P(MC_ARB_CG, CG_ARB_REQ(arb_freq_dest), ~CG_ARB_REQ_MASK);
+
+	return 0;
+}
+
+static int ni_init_arb_table_index(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev, ni_pi->arb_table_start,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp &= 0x00FFFFFF;
+	tmp |= ((u32)MC_CG_ARB_FREQ_F1) << 24;
+
+	return rv770_write_smc_sram_dword(rdev, ni_pi->arb_table_start,
+					  tmp, pi->sram_end);
+}
+
+static int ni_initial_switch_from_arb_f0_to_f1(struct radeon_device *rdev)
+{
+	return ni_copy_and_switch_arb_sets(rdev, MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1);
+}
+
+static int ni_force_switch_to_arb_f0(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev, ni_pi->arb_table_start,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp = (tmp >> 24) & 0xff;
+
+	if (tmp == MC_CG_ARB_FREQ_F0)
+		return 0;
+
+	return ni_copy_and_switch_arb_sets(rdev, tmp, MC_CG_ARB_FREQ_F0);
+}
+
+static int ni_populate_memory_timing_parameters(struct radeon_device *rdev,
+						struct rv7xx_pl *pl,
+						SMC_NIslands_MCArbDramTimingRegisterSet *arb_regs)
+{
+	u32 dram_timing;
+	u32 dram_timing2;
+
+	arb_regs->mc_arb_rfsh_rate =
+		(u8)rv770_calculate_memory_refresh_rate(rdev, pl->sclk);
+
+
+	radeon_atom_set_engine_dram_timings(rdev,
+                                            pl->sclk,
+                                            pl->mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	arb_regs->mc_arb_dram_timing  = cpu_to_be32(dram_timing);
+	arb_regs->mc_arb_dram_timing2 = cpu_to_be32(dram_timing2);
+
+	return 0;
+}
+
+static int ni_do_program_memory_timing_parameters(struct radeon_device *rdev,
+						  struct radeon_ps *radeon_state,
+						  unsigned int first_arb_set)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	SMC_NIslands_MCArbDramTimingRegisterSet arb_regs = { 0 };
+	int i, ret = 0;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ret = ni_populate_memory_timing_parameters(rdev, &state->performance_levels[i], &arb_regs);
+		if (ret)
+			break;
+
+		ret = rv770_copy_bytes_to_smc(rdev,
+					      (u16)(ni_pi->arb_table_start +
+						    offsetof(SMC_NIslands_MCArbDramTimingRegisters, data) +
+						    sizeof(SMC_NIslands_MCArbDramTimingRegisterSet) * (first_arb_set + i)),
+					      (u8 *)&arb_regs,
+					      (u16)sizeof(SMC_NIslands_MCArbDramTimingRegisterSet),
+					      pi->sram_end);
+		if (ret)
+			break;
+	}
+	return ret;
+}
+
+static int ni_program_memory_timing_parameters(struct radeon_device *rdev,
+					       struct radeon_ps *radeon_new_state)
+{
+	return ni_do_program_memory_timing_parameters(rdev, radeon_new_state,
+						      NISLANDS_DRIVER_STATE_ARB_INDEX);
+}
+
+static void ni_populate_initial_mvdd_value(struct radeon_device *rdev,
+					   struct NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	voltage->index = eg_pi->mvdd_high_index;
+	voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+}
+
+static int ni_populate_smc_initial_state(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_initial_state,
+					 NISLANDS_SMC_STATETABLE *table)
+{
+	struct ni_ps *initial_state = ni_get_ps(radeon_initial_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 reg;
+	int ret;
+
+	table->initialState.levels[0].mclk.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ad_func_cntl_2);
+	table->initialState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(ni_pi->clock_registers.mpll_dq_func_cntl_2);
+	table->initialState.levels[0].mclk.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.vDLL_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.dll_cntl);
+	table->initialState.levels[0].mclk.vMPLL_SS =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ss1);
+	table->initialState.levels[0].mclk.vMPLL_SS2 =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ss2);
+	table->initialState.levels[0].mclk.mclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl_4);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_spread_spectrum_2);
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].sclk);
+	table->initialState.levels[0].arbRefreshState =
+		NISLANDS_INITIAL_STATE_ARB_INDEX;
+
+	table->initialState.levels[0].ACIndex = 0;
+
+	ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+					initial_state->performance_levels[0].vddc,
+					&table->initialState.levels[0].vddc);
+	if (!ret) {
+		u16 std_vddc;
+
+		ret = ni_get_std_voltage_value(rdev,
+					       &table->initialState.levels[0].vddc,
+					       &std_vddc);
+		if (!ret)
+			ni_populate_std_voltage_value(rdev, std_vddc,
+						      table->initialState.levels[0].vddc.index,
+						      &table->initialState.levels[0].std_vddc);
+	}
+
+	if (eg_pi->vddci_control)
+		ni_populate_voltage_value(rdev,
+					  &eg_pi->vddci_voltage_table,
+					  initial_state->performance_levels[0].vddci,
+					  &table->initialState.levels[0].vddci);
+
+	ni_populate_initial_mvdd_value(rdev, &table->initialState.levels[0].mvdd);
+
+	reg = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(reg);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+
+	if (pi->mem_gddr5) {
+		table->initialState.levels[0].strobeMode =
+			cypress_get_strobe_mode_settings(rdev,
+							 initial_state->performance_levels[0].mclk);
+
+		if (initial_state->performance_levels[0].mclk > pi->mclk_edc_enable_threshold)
+			table->initialState.levels[0].mcFlags = NISLANDS_SMC_MC_EDC_RD_FLAG | NISLANDS_SMC_MC_EDC_WR_FLAG;
+		else
+			table->initialState.levels[0].mcFlags =  0;
+	}
+
+	table->initialState.levelCount = 1;
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	table->initialState.levels[0].dpm2.MaxPS = 0;
+	table->initialState.levels[0].dpm2.NearTDPDec = 0;
+	table->initialState.levels[0].dpm2.AboveSafeInc = 0;
+	table->initialState.levels[0].dpm2.BelowSafeInc = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->initialState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->initialState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int ni_populate_smc_acpi_state(struct radeon_device *rdev,
+				      NISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 mpll_ad_func_cntl   = ni_pi->clock_registers.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = ni_pi->clock_registers.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl   = ni_pi->clock_registers.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = ni_pi->clock_registers.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl      = ni_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2    = ni_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3    = ni_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4    = ni_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 mclk_pwrmgt_cntl    = ni_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32 dll_cntl            = ni_pi->clock_registers.dll_cntl;
+	u32 reg;
+	int ret;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		ret = ni_populate_voltage_value(rdev,
+						&eg_pi->vddc_voltage_table,
+						pi->acpi_vddc, &table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = ni_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc, &std_vddc);
+			if (!ret)
+				ni_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+
+		if (pi->pcie_gen2) {
+			if (pi->acpi_pcie_gen2)
+				table->ACPIState.levels[0].gen2PCIE = 1;
+			else
+				table->ACPIState.levels[0].gen2PCIE = 0;
+		} else {
+			table->ACPIState.levels[0].gen2PCIE = 0;
+		}
+	} else {
+		ret = ni_populate_voltage_value(rdev,
+						&eg_pi->vddc_voltage_table,
+						pi->min_vddc_in_table,
+						&table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = ni_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc,
+						       &std_vddc);
+			if (!ret)
+				ni_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	if (eg_pi->acpi_vddci) {
+		if (eg_pi->vddci_control)
+			ni_populate_voltage_value(rdev,
+						  &eg_pi->vddci_voltage_table,
+						  eg_pi->acpi_vddci,
+						  &table->ACPIState.levels[0].vddci);
+	}
+
+
+	mpll_ad_func_cntl &= ~PDNB;
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+        if (pi->mem_gddr5)
+                mpll_dq_func_cntl &= ~PDNB;
+        mpll_dq_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN | BYPASS;
+
+
+	mclk_pwrmgt_cntl |= (MRDCKA0_RESET |
+			     MRDCKA1_RESET |
+			     MRDCKB0_RESET |
+			     MRDCKB1_RESET |
+			     MRDCKC0_RESET |
+			     MRDCKC1_RESET |
+			     MRDCKD0_RESET |
+			     MRDCKD1_RESET);
+
+	mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+			      MRDCKA1_PDNB |
+			      MRDCKB0_PDNB |
+			      MRDCKB1_PDNB |
+			      MRDCKC0_PDNB |
+			      MRDCKC1_PDNB |
+			      MRDCKD0_PDNB |
+			      MRDCKD1_PDNB);
+
+	dll_cntl |= (MRDCKA0_BYPASS |
+                     MRDCKA1_BYPASS |
+                     MRDCKB0_BYPASS |
+                     MRDCKB1_BYPASS |
+                     MRDCKC0_BYPASS |
+                     MRDCKC1_BYPASS |
+                     MRDCKD0_BYPASS |
+                     MRDCKD1_BYPASS);
+
+        spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	table->ACPIState.levels[0].mclk.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 = cpu_to_be32(spll_func_cntl_4);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	ni_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	if (eg_pi->dynamic_ac_timing)
+		table->ACPIState.levels[0].ACIndex = 1;
+
+	table->ACPIState.levels[0].dpm2.MaxPS = 0;
+	table->ACPIState.levels[0].dpm2.NearTDPDec = 0;
+	table->ACPIState.levels[0].dpm2.AboveSafeInc = 0;
+	table->ACPIState.levels[0].dpm2.BelowSafeInc = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int ni_init_smc_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret;
+	struct radeon_ps *radeon_boot_state = rdev->pm.dpm.boot_ps;
+	NISLANDS_SMC_STATETABLE *table = &ni_pi->smc_statetable;
+
+	memset(table, 0, sizeof(NISLANDS_SMC_STATETABLE));
+
+	ni_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+	case THERMAL_TYPE_NI:
+	case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+	case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+	default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	ret = ni_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	ret = ni_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	table->ULVState = table->initialState;
+
+	ret = ni_do_program_memory_timing_parameters(rdev, radeon_boot_state,
+						     NISLANDS_INITIAL_STATE_ARB_INDEX);
+	if (ret)
+		return ret;
+
+	return rv770_copy_bytes_to_smc(rdev, pi->state_table_start, (u8 *)table,
+				       sizeof(NISLANDS_SMC_STATETABLE), pi->sram_end);
+}
+
+static int ni_calculate_sclk_params(struct radeon_device *rdev,
+				    u32 engine_clock,
+				    NISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = ni_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = ni_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = ni_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4 = ni_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum = ni_pi->clock_registers.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = ni_pi->clock_registers.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+
+	tmp = (u64) engine_clock * reference_divider * dividers.post_div * 16834;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_PDIV_A(dividers.post_div);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = engine_clock;
+	sclk->vCG_SPLL_FUNC_CNTL = spll_func_cntl;
+	sclk->vCG_SPLL_FUNC_CNTL_2 = spll_func_cntl_2;
+	sclk->vCG_SPLL_FUNC_CNTL_3 = spll_func_cntl_3;
+	sclk->vCG_SPLL_FUNC_CNTL_4 = spll_func_cntl_4;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cg_spll_spread_spectrum;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cg_spll_spread_spectrum_2;
+
+	return 0;
+}
+
+static int ni_populate_sclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  NISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	NISLANDS_SMC_SCLK_VALUE sclk_tmp;
+	int ret;
+
+	ret = ni_calculate_sclk_params(rdev, engine_clock, &sclk_tmp);
+	if (!ret) {
+		sclk->sclk_value = cpu_to_be32(sclk_tmp.sclk_value);
+		sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL);
+		sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_2);
+		sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_3);
+		sclk->vCG_SPLL_FUNC_CNTL_4 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_4);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM_2);
+	}
+
+	return ret;
+}
+
+static int ni_init_smc_spll_table(struct radeon_device *rdev)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	SMC_NISLANDS_SPLL_DIV_TABLE *spll_table;
+	NISLANDS_SMC_SCLK_VALUE sclk_params;
+	u32 fb_div;
+	u32 p_div;
+	u32 clk_s;
+	u32 clk_v;
+	u32 sclk = 0;
+	int i, ret;
+	u32 tmp;
+
+	if (ni_pi->spll_table_start == 0)
+		return -EINVAL;
+
+	spll_table = kzalloc(sizeof(SMC_NISLANDS_SPLL_DIV_TABLE), GFP_KERNEL);
+	if (spll_table == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < 256; i++) {
+		ret = ni_calculate_sclk_params(rdev, sclk, &sclk_params);
+		if (ret)
+			break;
+
+		p_div = (sclk_params.vCG_SPLL_FUNC_CNTL & SPLL_PDIV_A_MASK) >> SPLL_PDIV_A_SHIFT;
+		fb_div = (sclk_params.vCG_SPLL_FUNC_CNTL_3 & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT;
+		clk_s = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM & CLK_S_MASK) >> CLK_S_SHIFT;
+		clk_v = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM_2 & CLK_V_MASK) >> CLK_V_SHIFT;
+
+		fb_div &= ~0x00001FFF;
+		fb_div >>= 1;
+		clk_v >>= 6;
+
+		if (p_div & ~(SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT))
+			ret = -EINVAL;
+
+		if (clk_s & ~(SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT))
+			ret = -EINVAL;
+
+		if (clk_s & ~(SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT))
+			ret = -EINVAL;
+
+		if (clk_v & ~(SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT))
+			ret = -EINVAL;
+
+		if (ret)
+			break;
+
+		tmp = ((fb_div << SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_MASK) |
+			((p_div << SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_MASK);
+		spll_table->freq[i] = cpu_to_be32(tmp);
+
+		tmp = ((clk_v << SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_MASK) |
+			((clk_s << SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK);
+		spll_table->ss[i] = cpu_to_be32(tmp);
+
+		sclk += 512;
+	}
+
+	if (!ret)
+		ret = rv770_copy_bytes_to_smc(rdev, ni_pi->spll_table_start, (u8 *)spll_table,
+					      sizeof(SMC_NISLANDS_SPLL_DIV_TABLE), pi->sram_end);
+
+	kfree(spll_table);
+
+	return ret;
+}
+
+static int ni_populate_mclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  u32 memory_clock,
+				  NISLANDS_SMC_MCLK_VALUE *mclk,
+				  bool strobe_mode,
+				  bool dll_state_on)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 mpll_ad_func_cntl = ni_pi->clock_registers.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = ni_pi->clock_registers.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl = ni_pi->clock_registers.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = ni_pi->clock_registers.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl = ni_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32 dll_cntl = ni_pi->clock_registers.dll_cntl;
+	u32 mpll_ss1 = ni_pi->clock_registers.mpll_ss1;
+	u32 mpll_ss2 = ni_pi->clock_registers.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 ibias;
+	u32 dll_speed;
+	int ret;
+	u32 mc_seq_misc7;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, strobe_mode, &dividers);
+	if (ret)
+		return ret;
+
+	if (!strobe_mode) {
+		mc_seq_misc7 = RREG32(MC_SEQ_MISC7);
+
+		if (mc_seq_misc7 & 0x8000000)
+			dividers.post_div = 1;
+	}
+
+	ibias = cypress_map_clkf_to_ibias(rdev, dividers.whole_fb_div);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(dividers.ref_div);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+	mpll_ad_func_cntl |= CLKF(dividers.whole_fb_div);
+	mpll_ad_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(dividers.ref_div);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+		mpll_dq_func_cntl |= CLKF(dividers.whole_fb_div);
+		mpll_dq_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (strobe_mode)
+			mpll_dq_func_cntl &= ~PDNB;
+		else
+			mpll_dq_func_cntl |= PDNB;
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 decoded_ref = rv740_get_decoded_reference_divider(dividers.ref_div);
+			u32 clk_s = reference_clock * 5 / (decoded_ref * ss.rate);
+			u32 clk_v = ss.percentage *
+				(0x4000 * dividers.whole_fb_div + 0x800 * dividers.frac_fb_div) / (clk_s * 625);
+
+			mpll_ss1 &= ~CLKV_MASK;
+			mpll_ss1 |= CLKV(clk_v);
+
+			mpll_ss2 &= ~CLKS_MASK;
+			mpll_ss2 |= CLKS(clk_s);
+		}
+	}
+
+	dll_speed = rv740_get_dll_speed(pi->mem_gddr5,
+					memory_clock);
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(dll_speed);
+	if (dll_state_on)
+		mclk_pwrmgt_cntl |= (MRDCKA0_PDNB |
+				     MRDCKA1_PDNB |
+				     MRDCKB0_PDNB |
+				     MRDCKB1_PDNB |
+				     MRDCKC0_PDNB |
+				     MRDCKC1_PDNB |
+				     MRDCKD0_PDNB |
+				     MRDCKD1_PDNB);
+	else
+		mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+				      MRDCKA1_PDNB |
+				      MRDCKB0_PDNB |
+				      MRDCKB1_PDNB |
+				      MRDCKC0_PDNB |
+				      MRDCKC1_PDNB |
+				      MRDCKD0_PDNB |
+				      MRDCKD1_PDNB);
+
+
+	mclk->mclk_value = cpu_to_be32(memory_clock);
+	mclk->vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+static void ni_populate_smc_sp(struct radeon_device *rdev,
+			       struct radeon_ps *radeon_state,
+			       NISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_ps *ps = ni_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < ps->performance_level_count - 1; i++)
+		smc_state->levels[i].bSP = cpu_to_be32(pi->dsp);
+
+	smc_state->levels[ps->performance_level_count - 1].bSP =
+		cpu_to_be32(pi->psp);
+}
+
+static int ni_convert_power_level_to_smc(struct radeon_device *rdev,
+					 struct rv7xx_pl *pl,
+					 NISLANDS_SMC_HW_PERFORMANCE_LEVEL *level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret;
+	bool dll_state_on;
+	u16 std_vddc;
+	u32 tmp = RREG32(DC_STUTTER_CNTL);
+
+	level->gen2PCIE = pi->pcie_gen2 ?
+		((pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0) : 0;
+
+	ret = ni_populate_sclk_value(rdev, pl->sclk, &level->sclk);
+	if (ret)
+		return ret;
+
+	level->mcFlags =  0;
+	if (pi->mclk_stutter_mode_threshold &&
+	    (pl->mclk <= pi->mclk_stutter_mode_threshold) &&
+	    !eg_pi->uvd_enabled &&
+	    (tmp & DC_STUTTER_ENABLE_A) &&
+	    (tmp & DC_STUTTER_ENABLE_B))
+		level->mcFlags |= NISLANDS_SMC_MC_STUTTER_EN;
+
+	if (pi->mem_gddr5) {
+		if (pl->mclk > pi->mclk_edc_enable_threshold)
+			level->mcFlags |= NISLANDS_SMC_MC_EDC_RD_FLAG;
+		if (pl->mclk > eg_pi->mclk_edc_wr_enable_threshold)
+			level->mcFlags |= NISLANDS_SMC_MC_EDC_WR_FLAG;
+
+		level->strobeMode = cypress_get_strobe_mode_settings(rdev, pl->mclk);
+
+		if (level->strobeMode & NISLANDS_SMC_STROBE_ENABLE) {
+			if (cypress_get_mclk_frequency_ratio(rdev, pl->mclk, true) >=
+			    ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf))
+				dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+			else
+				dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false;
+		} else {
+			dll_state_on = false;
+			if (pl->mclk > ni_pi->mclk_rtt_mode_threshold)
+				level->mcFlags |= NISLANDS_SMC_MC_RTT_ENABLE;
+		}
+
+		ret = ni_populate_mclk_value(rdev, pl->sclk, pl->mclk,
+					     &level->mclk,
+					     (level->strobeMode & NISLANDS_SMC_STROBE_ENABLE) != 0,
+					     dll_state_on);
+	} else
+		ret = ni_populate_mclk_value(rdev, pl->sclk, pl->mclk, &level->mclk, 1, 1);
+
+	if (ret)
+		return ret;
+
+	ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+					pl->vddc, &level->vddc);
+	if (ret)
+		return ret;
+
+	ret = ni_get_std_voltage_value(rdev, &level->vddc, &std_vddc);
+	if (ret)
+		return ret;
+
+	ni_populate_std_voltage_value(rdev, std_vddc,
+				      level->vddc.index, &level->std_vddc);
+
+	if (eg_pi->vddci_control) {
+		ret = ni_populate_voltage_value(rdev, &eg_pi->vddci_voltage_table,
+						pl->vddci, &level->vddci);
+		if (ret)
+			return ret;
+	}
+
+	ni_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int ni_populate_smc_t(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_state,
+			     NISLANDS_SMC_SWSTATE *smc_state)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 a_t;
+	u32 t_l, t_h;
+	u32 high_bsp;
+	int i, ret;
+
+	if (state->performance_level_count >= 9)
+		return -EINVAL;
+
+	if (state->performance_level_count < 2) {
+		a_t = CG_R(0xffff) | CG_L(0);
+		smc_state->levels[0].aT = cpu_to_be32(a_t);
+		return 0;
+	}
+
+	smc_state->levels[0].aT = cpu_to_be32(0);
+
+	for (i = 0; i <= state->performance_level_count - 2; i++) {
+		if (eg_pi->uvd_enabled)
+			ret = r600_calculate_at(
+				1000 * (i * (eg_pi->smu_uvd_hs ? 2 : 8) + 2),
+				100 * R600_AH_DFLT,
+				state->performance_levels[i + 1].sclk,
+				state->performance_levels[i].sclk,
+				&t_l,
+				&t_h);
+		else
+			ret = r600_calculate_at(
+				1000 * (i + 1),
+				100 * R600_AH_DFLT,
+				state->performance_levels[i + 1].sclk,
+				state->performance_levels[i].sclk,
+				&t_l,
+				&t_h);
+
+		if (ret) {
+			t_h = (i + 1) * 1000 - 50 * R600_AH_DFLT;
+			t_l = (i + 1) * 1000 + 50 * R600_AH_DFLT;
+		}
+
+		a_t = be32_to_cpu(smc_state->levels[i].aT) & ~CG_R_MASK;
+		a_t |= CG_R(t_l * pi->bsp / 20000);
+		smc_state->levels[i].aT = cpu_to_be32(a_t);
+
+		high_bsp = (i == state->performance_level_count - 2) ?
+			pi->pbsp : pi->bsp;
+
+		a_t = CG_R(0xffff) | CG_L(t_h * high_bsp / 20000);
+		smc_state->levels[i + 1].aT = cpu_to_be32(a_t);
+	}
+
+	return 0;
+}
+
+static int ni_populate_power_containment_values(struct radeon_device *rdev,
+						struct radeon_ps *radeon_state,
+						NISLANDS_SMC_SWSTATE *smc_state)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 prev_sclk;
+	u32 max_sclk;
+	u32 min_sclk;
+	int i, ret;
+	u32 tdp_limit;
+	u32 near_tdp_limit;
+	u32 power_boost_limit;
+	u8 max_ps_percent;
+
+	if (ni_pi->enable_power_containment == false)
+		return 0;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	ret = ni_calculate_adjusted_tdp_limits(rdev,
+					       false, /* ??? */
+					       rdev->pm.dpm.tdp_adjustment,
+					       &tdp_limit,
+					       &near_tdp_limit);
+	if (ret)
+		return ret;
+
+	power_boost_limit = ni_calculate_power_boost_limit(rdev, radeon_state, near_tdp_limit);
+
+	ret = rv770_write_smc_sram_dword(rdev,
+					 pi->state_table_start +
+					 offsetof(NISLANDS_SMC_STATETABLE, dpm2Params) +
+					 offsetof(PP_NIslands_DPM2Parameters, PowerBoostLimit),
+					 ni_scale_power_for_smc(power_boost_limit, ni_get_smc_power_scaling_factor(rdev)),
+					 pi->sram_end);
+	if (ret)
+		power_boost_limit = 0;
+
+	smc_state->levels[0].dpm2.MaxPS = 0;
+	smc_state->levels[0].dpm2.NearTDPDec = 0;
+	smc_state->levels[0].dpm2.AboveSafeInc = 0;
+	smc_state->levels[0].dpm2.BelowSafeInc = 0;
+	smc_state->levels[0].stateFlags |= power_boost_limit ? PPSMC_STATEFLAG_POWERBOOST : 0;
+
+	for (i = 1; i < state->performance_level_count; i++) {
+		prev_sclk = state->performance_levels[i-1].sclk;
+		max_sclk  = state->performance_levels[i].sclk;
+		max_ps_percent = (i != (state->performance_level_count - 1)) ?
+			NISLANDS_DPM2_MAXPS_PERCENT_M : NISLANDS_DPM2_MAXPS_PERCENT_H;
+
+		if (max_sclk < prev_sclk)
+			return -EINVAL;
+
+		if ((max_ps_percent == 0) || (prev_sclk == max_sclk) || eg_pi->uvd_enabled)
+			min_sclk = max_sclk;
+		else if (1 == i)
+			min_sclk = prev_sclk;
+		else
+			min_sclk = (prev_sclk * (u32)max_ps_percent) / 100;
+
+		if (min_sclk < state->performance_levels[0].sclk)
+			min_sclk = state->performance_levels[0].sclk;
+
+		if (min_sclk == 0)
+			return -EINVAL;
+
+		smc_state->levels[i].dpm2.MaxPS =
+			(u8)((NISLANDS_DPM2_MAX_PULSE_SKIP * (max_sclk - min_sclk)) / max_sclk);
+		smc_state->levels[i].dpm2.NearTDPDec = NISLANDS_DPM2_NEAR_TDP_DEC;
+		smc_state->levels[i].dpm2.AboveSafeInc = NISLANDS_DPM2_ABOVE_SAFE_INC;
+		smc_state->levels[i].dpm2.BelowSafeInc = NISLANDS_DPM2_BELOW_SAFE_INC;
+		smc_state->levels[i].stateFlags |=
+			((i != (state->performance_level_count - 1)) && power_boost_limit) ?
+			PPSMC_STATEFLAG_POWERBOOST : 0;
+	}
+
+	return 0;
+}
+
+static int ni_populate_sq_ramping_values(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 NISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 sq_power_throttle;
+	u32 sq_power_throttle2;
+	bool enable_sq_ramping = ni_pi->enable_sq_ramping;
+	int i;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	if (rdev->pm.dpm.sq_ramping_threshold == 0)
+		return -EINVAL;
+
+	if (NISLANDS_DPM2_SQ_RAMP_MAX_POWER > (MAX_POWER_MASK >> MAX_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_MIN_POWER > (MIN_POWER_MASK >> MIN_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA > (MAX_POWER_DELTA_MASK >> MAX_POWER_DELTA_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO <= (LTI_RATIO_MASK >> LTI_RATIO_SHIFT))
+		enable_sq_ramping = false;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		sq_power_throttle  = 0;
+		sq_power_throttle2 = 0;
+
+		if ((state->performance_levels[i].sclk >= rdev->pm.dpm.sq_ramping_threshold) &&
+		    enable_sq_ramping) {
+			sq_power_throttle |= MAX_POWER(NISLANDS_DPM2_SQ_RAMP_MAX_POWER);
+			sq_power_throttle |= MIN_POWER(NISLANDS_DPM2_SQ_RAMP_MIN_POWER);
+			sq_power_throttle2 |= MAX_POWER_DELTA(NISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA);
+			sq_power_throttle2 |= STI_SIZE(NISLANDS_DPM2_SQ_RAMP_STI_SIZE);
+			sq_power_throttle2 |= LTI_RATIO(NISLANDS_DPM2_SQ_RAMP_LTI_RATIO);
+		} else {
+			sq_power_throttle |= MAX_POWER_MASK | MIN_POWER_MASK;
+			sq_power_throttle2 |= MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+		}
+
+		smc_state->levels[i].SQPowerThrottle   = cpu_to_be32(sq_power_throttle);
+		smc_state->levels[i].SQPowerThrottle_2 = cpu_to_be32(sq_power_throttle2);
+	}
+
+	return 0;
+}
+
+static int ni_enable_power_containment(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_new_state,
+				       bool enable)
+{
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	PPSMC_Result smc_result;
+	int ret = 0;
+
+	if (ni_pi->enable_power_containment) {
+		if (enable) {
+			if (!r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2)) {
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_TDPClampingActive);
+				if (smc_result != PPSMC_Result_OK) {
+					ret = -EINVAL;
+					ni_pi->pc_enabled = false;
+				} else {
+					ni_pi->pc_enabled = true;
+				}
+			}
+		} else {
+			smc_result = rv770_send_msg_to_smc(rdev, PPSMC_TDPClampingInactive);
+			if (smc_result != PPSMC_Result_OK)
+				ret = -EINVAL;
+			ni_pi->pc_enabled = false;
+		}
+	}
+
+	return ret;
+}
+
+static int ni_convert_power_state_to_smc(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 NISLANDS_SMC_SWSTATE *smc_state)
+{
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i, ret;
+	u32 threshold = state->performance_levels[state->performance_level_count - 1].sclk * 100 / 100;
+
+	if (!(radeon_state->caps & ATOM_PPLIB_DISALLOW_ON_DC))
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	smc_state->levelCount = 0;
+
+	if (state->performance_level_count > NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE)
+		return -EINVAL;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ret = ni_convert_power_level_to_smc(rdev, &state->performance_levels[i],
+						    &smc_state->levels[i]);
+		smc_state->levels[i].arbRefreshState =
+			(u8)(NISLANDS_DRIVER_STATE_ARB_INDEX + i);
+
+		if (ret)
+			return ret;
+
+		if (ni_pi->enable_power_containment)
+			smc_state->levels[i].displayWatermark =
+				(state->performance_levels[i].sclk < threshold) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+		else
+			smc_state->levels[i].displayWatermark = (i < 2) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+
+		if (eg_pi->dynamic_ac_timing)
+			smc_state->levels[i].ACIndex = NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i;
+		else
+			smc_state->levels[i].ACIndex = 0;
+
+		smc_state->levelCount++;
+	}
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_watermark_threshold,
+				      cpu_to_be32(threshold / 512));
+
+	ni_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	ret = ni_populate_power_containment_values(rdev, radeon_state, smc_state);
+	if (ret)
+		ni_pi->enable_power_containment = false;
+
+	ret = ni_populate_sq_ramping_values(rdev, radeon_state, smc_state);
+	if (ret)
+		ni_pi->enable_sq_ramping = false;
+
+	return ni_populate_smc_t(rdev, radeon_state, smc_state);
+}
+
+static int ni_upload_sw_state(struct radeon_device *rdev,
+			      struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 address = pi->state_table_start +
+		offsetof(NISLANDS_SMC_STATETABLE, driverState);
+	u16 state_size = sizeof(NISLANDS_SMC_SWSTATE) +
+		((NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1) * sizeof(NISLANDS_SMC_HW_PERFORMANCE_LEVEL));
+	int ret;
+	NISLANDS_SMC_SWSTATE *smc_state = kzalloc(state_size, GFP_KERNEL);
+
+	if (smc_state == NULL)
+		return -ENOMEM;
+
+	ret = ni_convert_power_state_to_smc(rdev, radeon_new_state, smc_state);
+	if (ret)
+		goto done;
+
+	ret = rv770_copy_bytes_to_smc(rdev, address, (u8 *)smc_state, state_size, pi->sram_end);
+
+done:
+	kfree(smc_state);
+
+	return ret;
+}
+
+static int ni_set_mc_special_registers(struct radeon_device *rdev,
+				       struct ni_mc_reg_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 i, j, k;
+	u32 temp_reg;
+
+	for (i = 0, j = table->last; i < table->last; i++) {
+		switch (table->mc_reg_address[i].s1) {
+		case MC_SEQ_MISC1 >> 2:
+			if (j >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			temp_reg = RREG32(MC_PMG_CMD_EMRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					((temp_reg & 0xffff0000)) |
+					((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16);
+			j++;
+			if (j >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			temp_reg = RREG32(MC_PMG_CMD_MRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+			for(k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+				if (!pi->mem_gddr5)
+					table->mc_reg_table_entry[k].mc_data[j] |= 0x100;
+			}
+			j++;
+			if (j > SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		case MC_SEQ_RESERVE_M >> 2:
+			temp_reg = RREG32(MC_PMG_CMD_MRS1);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+			for (k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+			j++;
+			if (j > SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		default:
+			break;
+		}
+	}
+
+	table->last = j;
+
+	return 0;
+}
+
+static bool ni_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg)
+{
+	bool result = true;
+
+	switch (in_reg) {
+        case  MC_SEQ_RAS_TIMING >> 2:
+		*out_reg = MC_SEQ_RAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_CAS_TIMING >> 2:
+		*out_reg = MC_SEQ_CAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING2 >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING2_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D1_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D1_LP >> 2;
+		break;
+        case MC_PMG_CMD_EMRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS1 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+		break;
+        case MC_SEQ_PMG_TIMING >> 2:
+		*out_reg = MC_SEQ_PMG_TIMING_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS2 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS2_LP >> 2;
+		break;
+        default:
+		result = false;
+		break;
+	}
+
+	return result;
+}
+
+static void ni_set_valid_flag(struct ni_mc_reg_table *table)
+{
+	u8 i, j;
+
+	for (i = 0; i < table->last; i++) {
+		for (j = 1; j < table->num_entries; j++) {
+			if (table->mc_reg_table_entry[j-1].mc_data[i] != table->mc_reg_table_entry[j].mc_data[i]) {
+				table->valid_flag |= 1 << i;
+				break;
+			}
+		}
+	}
+}
+
+static void ni_set_s0_mc_reg_index(struct ni_mc_reg_table *table)
+{
+	u32 i;
+	u16 address;
+
+	for (i = 0; i < table->last; i++)
+		table->mc_reg_address[i].s0 =
+			ni_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ?
+			address : table->mc_reg_address[i].s1;
+}
+
+static int ni_copy_vbios_mc_reg_table(struct atom_mc_reg_table *table,
+				      struct ni_mc_reg_table *ni_table)
+{
+	u8 i, j;
+
+	if (table->last > SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+		return -EINVAL;
+	if (table->num_entries > MAX_AC_TIMING_ENTRIES)
+		return -EINVAL;
+
+	for (i = 0; i < table->last; i++)
+		ni_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1;
+	ni_table->last = table->last;
+
+	for (i = 0; i < table->num_entries; i++) {
+		ni_table->mc_reg_table_entry[i].mclk_max =
+			table->mc_reg_table_entry[i].mclk_max;
+		for (j = 0; j < table->last; j++)
+			ni_table->mc_reg_table_entry[i].mc_data[j] =
+				table->mc_reg_table_entry[i].mc_data[j];
+	}
+	ni_table->num_entries = table->num_entries;
+
+	return 0;
+}
+
+static int ni_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret;
+	struct atom_mc_reg_table *table;
+	struct ni_mc_reg_table *ni_table = &ni_pi->mc_reg_table;
+	u8 module_index = rv770_get_memory_module_index(rdev);
+
+        table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL);
+        if (!table)
+                return -ENOMEM;
+
+	WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING));
+	WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2));
+	WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1));
+	WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0));
+	WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1));
+	WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0));
+	WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1));
+	WREG32(MC_SEQ_PMG_TIMING_LP, RREG32(MC_SEQ_PMG_TIMING));
+	WREG32(MC_SEQ_PMG_CMD_MRS2_LP, RREG32(MC_PMG_CMD_MRS2));
+
+	ret = radeon_atom_init_mc_reg_table(rdev, module_index, table);
+
+        if (ret)
+                goto init_mc_done;
+
+	ret = ni_copy_vbios_mc_reg_table(table, ni_table);
+
+        if (ret)
+                goto init_mc_done;
+
+	ni_set_s0_mc_reg_index(ni_table);
+
+	ret = ni_set_mc_special_registers(rdev, ni_table);
+
+        if (ret)
+                goto init_mc_done;
+
+	ni_set_valid_flag(ni_table);
+
+init_mc_done:
+        kfree(table);
+
+	return ret;
+}
+
+static void ni_populate_mc_reg_addresses(struct radeon_device *rdev,
+					 SMC_NIslands_MCRegisters *mc_reg_table)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0, j = 0; j < ni_pi->mc_reg_table.last; j++) {
+		if (ni_pi->mc_reg_table.valid_flag & (1 << j)) {
+			if (i >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				break;
+			mc_reg_table->address[i].s0 =
+				cpu_to_be16(ni_pi->mc_reg_table.mc_reg_address[j].s0);
+			mc_reg_table->address[i].s1 =
+				cpu_to_be16(ni_pi->mc_reg_table.mc_reg_address[j].s1);
+			i++;
+		}
+	}
+	mc_reg_table->last = (u8)i;
+}
+
+
+static void ni_convert_mc_registers(struct ni_mc_reg_entry *entry,
+				    SMC_NIslands_MCRegisterSet *data,
+				    u32 num_entries, u32 valid_flag)
+{
+	u32 i, j;
+
+	for (i = 0, j = 0; j < num_entries; j++) {
+		if (valid_flag & (1 << j)) {
+			data->value[i] = cpu_to_be32(entry->mc_data[j]);
+			i++;
+		}
+	}
+}
+
+static void ni_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev,
+						 struct rv7xx_pl *pl,
+						 SMC_NIslands_MCRegisterSet *mc_reg_table_data)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < ni_pi->mc_reg_table.num_entries; i++) {
+		if (pl->mclk <= ni_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max)
+			break;
+	}
+
+	if ((i == ni_pi->mc_reg_table.num_entries) && (i > 0))
+		--i;
+
+	ni_convert_mc_registers(&ni_pi->mc_reg_table.mc_reg_table_entry[i],
+				mc_reg_table_data,
+				ni_pi->mc_reg_table.last,
+				ni_pi->mc_reg_table.valid_flag);
+}
+
+static void ni_convert_mc_reg_table_to_smc(struct radeon_device *rdev,
+					   struct radeon_ps *radeon_state,
+					   SMC_NIslands_MCRegisters *mc_reg_table)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ni_convert_mc_reg_table_entry_to_smc(rdev,
+						     &state->performance_levels[i],
+						     &mc_reg_table->data[NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i]);
+	}
+}
+
+static int ni_populate_mc_reg_table(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *boot_state = ni_get_ps(radeon_boot_state);
+	SMC_NIslands_MCRegisters *mc_reg_table = &ni_pi->smc_mc_reg_table;
+
+	memset(mc_reg_table, 0, sizeof(SMC_NIslands_MCRegisters));
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_seq_index, 1);
+
+	ni_populate_mc_reg_addresses(rdev, mc_reg_table);
+
+	ni_convert_mc_reg_table_entry_to_smc(rdev, &boot_state->performance_levels[0],
+					     &mc_reg_table->data[0]);
+
+	ni_convert_mc_registers(&ni_pi->mc_reg_table.mc_reg_table_entry[0],
+				&mc_reg_table->data[1],
+				ni_pi->mc_reg_table.last,
+				ni_pi->mc_reg_table.valid_flag);
+
+	ni_convert_mc_reg_table_to_smc(rdev, radeon_boot_state, mc_reg_table);
+
+	return rv770_copy_bytes_to_smc(rdev, eg_pi->mc_reg_table_start,
+				       (u8 *)mc_reg_table,
+				       sizeof(SMC_NIslands_MCRegisters),
+				       pi->sram_end);
+}
+
+static int ni_upload_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *ni_new_state = ni_get_ps(radeon_new_state);
+	SMC_NIslands_MCRegisters *mc_reg_table = &ni_pi->smc_mc_reg_table;
+	u16 address;
+
+	memset(mc_reg_table, 0, sizeof(SMC_NIslands_MCRegisters));
+
+	ni_convert_mc_reg_table_to_smc(rdev, radeon_new_state, mc_reg_table);
+
+	address = eg_pi->mc_reg_table_start +
+		(u16)offsetof(SMC_NIslands_MCRegisters, data[NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT]);
+
+	return rv770_copy_bytes_to_smc(rdev, address,
+				       (u8 *)&mc_reg_table->data[NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT],
+				       sizeof(SMC_NIslands_MCRegisterSet) * ni_new_state->performance_level_count,
+				       pi->sram_end);
+}
+
+static int ni_init_driver_calculated_leakage_table(struct radeon_device *rdev,
+						   PP_NIslands_CACTABLES *cac_tables)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 leakage = 0;
+	unsigned int i, j, table_size;
+	s32 t;
+	u32 smc_leakage, max_leakage = 0;
+	u32 scaling_factor;
+
+	table_size = eg_pi->vddc_voltage_table.count;
+
+	if (SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES < table_size)
+		table_size = SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES;
+
+	scaling_factor = ni_get_smc_power_scaling_factor(rdev);
+
+	for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++) {
+		for (j = 0; j < table_size; j++) {
+			t = (1000 * ((i + 1) * 8));
+
+			if (t < ni_pi->cac_data.leakage_minimum_temperature)
+				t = ni_pi->cac_data.leakage_minimum_temperature;
+
+			ni_calculate_leakage_for_v_and_t(rdev,
+							 &ni_pi->cac_data.leakage_coefficients,
+							 eg_pi->vddc_voltage_table.entries[j].value,
+							 t,
+							 ni_pi->cac_data.i_leakage,
+							 &leakage);
+
+			smc_leakage = ni_scale_power_for_smc(leakage, scaling_factor) / 1000;
+			if (smc_leakage > max_leakage)
+				max_leakage = smc_leakage;
+
+			cac_tables->cac_lkge_lut[i][j] = cpu_to_be32(smc_leakage);
+		}
+	}
+
+	for (j = table_size; j < SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+		for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++)
+			cac_tables->cac_lkge_lut[i][j] = cpu_to_be32(max_leakage);
+	}
+	return 0;
+}
+
+static int ni_init_simplified_leakage_table(struct radeon_device *rdev,
+					    PP_NIslands_CACTABLES *cac_tables)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_cac_leakage_table *leakage_table =
+		&rdev->pm.dpm.dyn_state.cac_leakage_table;
+	u32 i, j, table_size;
+	u32 smc_leakage, max_leakage = 0;
+	u32 scaling_factor;
+
+	if (!leakage_table)
+		return -EINVAL;
+
+	table_size = leakage_table->count;
+
+	if (eg_pi->vddc_voltage_table.count != table_size)
+		table_size = (eg_pi->vddc_voltage_table.count < leakage_table->count) ?
+			eg_pi->vddc_voltage_table.count : leakage_table->count;
+
+	if (SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES < table_size)
+		table_size = SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES;
+
+	if (table_size == 0)
+		return -EINVAL;
+
+	scaling_factor = ni_get_smc_power_scaling_factor(rdev);
+
+	for (j = 0; j < table_size; j++) {
+		smc_leakage = leakage_table->entries[j].leakage;
+
+		if (smc_leakage > max_leakage)
+			max_leakage = smc_leakage;
+
+		for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++)
+			cac_tables->cac_lkge_lut[i][j] =
+				cpu_to_be32(ni_scale_power_for_smc(smc_leakage, scaling_factor));
+	}
+
+	for (j = table_size; j < SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+		for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++)
+			cac_tables->cac_lkge_lut[i][j] =
+				cpu_to_be32(ni_scale_power_for_smc(max_leakage, scaling_factor));
+	}
+	return 0;
+}
+
+static int ni_initialize_smc_cac_tables(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	PP_NIslands_CACTABLES *cac_tables = NULL;
+	int i, ret;
+        u32 reg;
+
+	if (ni_pi->enable_cac == false)
+		return 0;
+
+	cac_tables = kzalloc(sizeof(PP_NIslands_CACTABLES), GFP_KERNEL);
+	if (!cac_tables)
+		return -ENOMEM;
+
+	reg = RREG32(CG_CAC_CTRL) & ~(TID_CNT_MASK | TID_UNIT_MASK);
+	reg |= (TID_CNT(ni_pi->cac_weights->tid_cnt) |
+		TID_UNIT(ni_pi->cac_weights->tid_unit));
+	WREG32(CG_CAC_CTRL, reg);
+
+	for (i = 0; i < NISLANDS_DCCAC_MAX_LEVELS; i++)
+		ni_pi->dc_cac_table[i] = ni_pi->cac_weights->dc_cac[i];
+
+	for (i = 0; i < SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES; i++)
+		cac_tables->cac_bif_lut[i] = ni_pi->cac_weights->pcie_cac[i];
+
+	ni_pi->cac_data.i_leakage = rdev->pm.dpm.cac_leakage;
+	ni_pi->cac_data.pwr_const = 0;
+	ni_pi->cac_data.dc_cac_value = ni_pi->dc_cac_table[NISLANDS_DCCAC_LEVEL_0];
+	ni_pi->cac_data.bif_cac_value = 0;
+	ni_pi->cac_data.mc_wr_weight = ni_pi->cac_weights->mc_write_weight;
+	ni_pi->cac_data.mc_rd_weight = ni_pi->cac_weights->mc_read_weight;
+	ni_pi->cac_data.allow_ovrflw = 0;
+	ni_pi->cac_data.l2num_win_tdp = ni_pi->lta_window_size;
+	ni_pi->cac_data.num_win_tdp = 0;
+	ni_pi->cac_data.lts_truncate_n = ni_pi->lts_truncate;
+
+	if (ni_pi->driver_calculate_cac_leakage)
+		ret = ni_init_driver_calculated_leakage_table(rdev, cac_tables);
+	else
+		ret = ni_init_simplified_leakage_table(rdev, cac_tables);
+
+	if (ret)
+		goto done_free;
+
+	cac_tables->pwr_const      = cpu_to_be32(ni_pi->cac_data.pwr_const);
+	cac_tables->dc_cacValue    = cpu_to_be32(ni_pi->cac_data.dc_cac_value);
+	cac_tables->bif_cacValue   = cpu_to_be32(ni_pi->cac_data.bif_cac_value);
+	cac_tables->AllowOvrflw    = ni_pi->cac_data.allow_ovrflw;
+	cac_tables->MCWrWeight     = ni_pi->cac_data.mc_wr_weight;
+	cac_tables->MCRdWeight     = ni_pi->cac_data.mc_rd_weight;
+	cac_tables->numWin_TDP     = ni_pi->cac_data.num_win_tdp;
+	cac_tables->l2numWin_TDP   = ni_pi->cac_data.l2num_win_tdp;
+	cac_tables->lts_truncate_n = ni_pi->cac_data.lts_truncate_n;
+
+	ret = rv770_copy_bytes_to_smc(rdev, ni_pi->cac_table_start, (u8 *)cac_tables,
+				      sizeof(PP_NIslands_CACTABLES), pi->sram_end);
+
+done_free:
+	if (ret) {
+		ni_pi->enable_cac = false;
+		ni_pi->enable_power_containment = false;
+	}
+
+	kfree(cac_tables);
+
+	return 0;
+}
+
+static int ni_initialize_hardware_cac_manager(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 reg;
+
+	if (!ni_pi->enable_cac ||
+	    !ni_pi->cac_configuration_required)
+		return 0;
+
+	if (ni_pi->cac_weights == NULL)
+		return -EINVAL;
+
+	reg = RREG32_CG(CG_CAC_REGION_1_WEIGHT_0) & ~(WEIGHT_TCP_SIG0_MASK |
+						      WEIGHT_TCP_SIG1_MASK |
+						      WEIGHT_TA_SIG_MASK);
+	reg |= (WEIGHT_TCP_SIG0(ni_pi->cac_weights->weight_tcp_sig0) |
+		WEIGHT_TCP_SIG1(ni_pi->cac_weights->weight_tcp_sig1) |
+		WEIGHT_TA_SIG(ni_pi->cac_weights->weight_ta_sig));
+	WREG32_CG(CG_CAC_REGION_1_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_1_WEIGHT_1) & ~(WEIGHT_TCC_EN0_MASK |
+						      WEIGHT_TCC_EN1_MASK |
+						      WEIGHT_TCC_EN2_MASK);
+	reg |= (WEIGHT_TCC_EN0(ni_pi->cac_weights->weight_tcc_en0) |
+		WEIGHT_TCC_EN1(ni_pi->cac_weights->weight_tcc_en1) |
+		WEIGHT_TCC_EN2(ni_pi->cac_weights->weight_tcc_en2));
+	WREG32_CG(CG_CAC_REGION_1_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_2_WEIGHT_0) & ~(WEIGHT_CB_EN0_MASK |
+						      WEIGHT_CB_EN1_MASK |
+						      WEIGHT_CB_EN2_MASK |
+						      WEIGHT_CB_EN3_MASK);
+	reg |= (WEIGHT_CB_EN0(ni_pi->cac_weights->weight_cb_en0) |
+		WEIGHT_CB_EN1(ni_pi->cac_weights->weight_cb_en1) |
+		WEIGHT_CB_EN2(ni_pi->cac_weights->weight_cb_en2) |
+		WEIGHT_CB_EN3(ni_pi->cac_weights->weight_cb_en3));
+	WREG32_CG(CG_CAC_REGION_2_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_2_WEIGHT_1) & ~(WEIGHT_DB_SIG0_MASK |
+						      WEIGHT_DB_SIG1_MASK |
+						      WEIGHT_DB_SIG2_MASK |
+						      WEIGHT_DB_SIG3_MASK);
+	reg |= (WEIGHT_DB_SIG0(ni_pi->cac_weights->weight_db_sig0) |
+		WEIGHT_DB_SIG1(ni_pi->cac_weights->weight_db_sig1) |
+		WEIGHT_DB_SIG2(ni_pi->cac_weights->weight_db_sig2) |
+		WEIGHT_DB_SIG3(ni_pi->cac_weights->weight_db_sig3));
+	WREG32_CG(CG_CAC_REGION_2_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_2_WEIGHT_2) & ~(WEIGHT_SXM_SIG0_MASK |
+						      WEIGHT_SXM_SIG1_MASK |
+						      WEIGHT_SXM_SIG2_MASK |
+						      WEIGHT_SXS_SIG0_MASK |
+						      WEIGHT_SXS_SIG1_MASK);
+	reg |= (WEIGHT_SXM_SIG0(ni_pi->cac_weights->weight_sxm_sig0) |
+		WEIGHT_SXM_SIG1(ni_pi->cac_weights->weight_sxm_sig1) |
+		WEIGHT_SXM_SIG2(ni_pi->cac_weights->weight_sxm_sig2) |
+		WEIGHT_SXS_SIG0(ni_pi->cac_weights->weight_sxs_sig0) |
+		WEIGHT_SXS_SIG1(ni_pi->cac_weights->weight_sxs_sig1));
+	WREG32_CG(CG_CAC_REGION_2_WEIGHT_2, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_3_WEIGHT_0) & ~(WEIGHT_XBR_0_MASK |
+						      WEIGHT_XBR_1_MASK |
+						      WEIGHT_XBR_2_MASK |
+						      WEIGHT_SPI_SIG0_MASK);
+	reg |= (WEIGHT_XBR_0(ni_pi->cac_weights->weight_xbr_0) |
+		WEIGHT_XBR_1(ni_pi->cac_weights->weight_xbr_1) |
+		WEIGHT_XBR_2(ni_pi->cac_weights->weight_xbr_2) |
+		WEIGHT_SPI_SIG0(ni_pi->cac_weights->weight_spi_sig0));
+	WREG32_CG(CG_CAC_REGION_3_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_3_WEIGHT_1) & ~(WEIGHT_SPI_SIG1_MASK |
+						      WEIGHT_SPI_SIG2_MASK |
+						      WEIGHT_SPI_SIG3_MASK |
+						      WEIGHT_SPI_SIG4_MASK |
+						      WEIGHT_SPI_SIG5_MASK);
+	reg |= (WEIGHT_SPI_SIG1(ni_pi->cac_weights->weight_spi_sig1) |
+		WEIGHT_SPI_SIG2(ni_pi->cac_weights->weight_spi_sig2) |
+		WEIGHT_SPI_SIG3(ni_pi->cac_weights->weight_spi_sig3) |
+		WEIGHT_SPI_SIG4(ni_pi->cac_weights->weight_spi_sig4) |
+		WEIGHT_SPI_SIG5(ni_pi->cac_weights->weight_spi_sig5));
+	WREG32_CG(CG_CAC_REGION_3_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_0) & ~(WEIGHT_LDS_SIG0_MASK |
+						      WEIGHT_LDS_SIG1_MASK |
+						      WEIGHT_SC_MASK);
+	reg |= (WEIGHT_LDS_SIG0(ni_pi->cac_weights->weight_lds_sig0) |
+		WEIGHT_LDS_SIG1(ni_pi->cac_weights->weight_lds_sig1) |
+		WEIGHT_SC(ni_pi->cac_weights->weight_sc));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_1) & ~(WEIGHT_BIF_MASK |
+						      WEIGHT_CP_MASK |
+						      WEIGHT_PA_SIG0_MASK |
+						      WEIGHT_PA_SIG1_MASK |
+						      WEIGHT_VGT_SIG0_MASK);
+	reg |= (WEIGHT_BIF(ni_pi->cac_weights->weight_bif) |
+		WEIGHT_CP(ni_pi->cac_weights->weight_cp) |
+		WEIGHT_PA_SIG0(ni_pi->cac_weights->weight_pa_sig0) |
+		WEIGHT_PA_SIG1(ni_pi->cac_weights->weight_pa_sig1) |
+		WEIGHT_VGT_SIG0(ni_pi->cac_weights->weight_vgt_sig0));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_2) & ~(WEIGHT_VGT_SIG1_MASK |
+						      WEIGHT_VGT_SIG2_MASK |
+						      WEIGHT_DC_SIG0_MASK |
+						      WEIGHT_DC_SIG1_MASK |
+						      WEIGHT_DC_SIG2_MASK);
+	reg |= (WEIGHT_VGT_SIG1(ni_pi->cac_weights->weight_vgt_sig1) |
+		WEIGHT_VGT_SIG2(ni_pi->cac_weights->weight_vgt_sig2) |
+		WEIGHT_DC_SIG0(ni_pi->cac_weights->weight_dc_sig0) |
+		WEIGHT_DC_SIG1(ni_pi->cac_weights->weight_dc_sig1) |
+		WEIGHT_DC_SIG2(ni_pi->cac_weights->weight_dc_sig2));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_2, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_3) & ~(WEIGHT_DC_SIG3_MASK |
+						      WEIGHT_UVD_SIG0_MASK |
+						      WEIGHT_UVD_SIG1_MASK |
+						      WEIGHT_SPARE0_MASK |
+						      WEIGHT_SPARE1_MASK);
+	reg |= (WEIGHT_DC_SIG3(ni_pi->cac_weights->weight_dc_sig3) |
+		WEIGHT_UVD_SIG0(ni_pi->cac_weights->weight_uvd_sig0) |
+		WEIGHT_UVD_SIG1(ni_pi->cac_weights->weight_uvd_sig1) |
+		WEIGHT_SPARE0(ni_pi->cac_weights->weight_spare0) |
+		WEIGHT_SPARE1(ni_pi->cac_weights->weight_spare1));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_3, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_5_WEIGHT_0) & ~(WEIGHT_SQ_VSP_MASK |
+						      WEIGHT_SQ_VSP0_MASK);
+	reg |= (WEIGHT_SQ_VSP(ni_pi->cac_weights->weight_sq_vsp) |
+		WEIGHT_SQ_VSP0(ni_pi->cac_weights->weight_sq_vsp0));
+	WREG32_CG(CG_CAC_REGION_5_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_5_WEIGHT_1) & ~(WEIGHT_SQ_GPR_MASK);
+	reg |= WEIGHT_SQ_GPR(ni_pi->cac_weights->weight_sq_gpr);
+	WREG32_CG(CG_CAC_REGION_5_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_OVERRIDE_4) & ~(OVR_MODE_SPARE_0_MASK |
+							OVR_VAL_SPARE_0_MASK |
+							OVR_MODE_SPARE_1_MASK |
+							OVR_VAL_SPARE_1_MASK);
+	reg |= (OVR_MODE_SPARE_0(ni_pi->cac_weights->ovr_mode_spare_0) |
+		OVR_VAL_SPARE_0(ni_pi->cac_weights->ovr_val_spare_0) |
+		OVR_MODE_SPARE_1(ni_pi->cac_weights->ovr_mode_spare_1) |
+		OVR_VAL_SPARE_1(ni_pi->cac_weights->ovr_val_spare_1));
+	WREG32_CG(CG_CAC_REGION_4_OVERRIDE_4, reg);
+
+	reg = RREG32(SQ_CAC_THRESHOLD) & ~(VSP_MASK |
+					   VSP0_MASK |
+					   GPR_MASK);
+	reg |= (VSP(ni_pi->cac_weights->vsp) |
+		VSP0(ni_pi->cac_weights->vsp0) |
+		GPR(ni_pi->cac_weights->gpr));
+	WREG32(SQ_CAC_THRESHOLD, reg);
+
+	reg = (MCDW_WR_ENABLE |
+	       MCDX_WR_ENABLE |
+	       MCDY_WR_ENABLE |
+	       MCDZ_WR_ENABLE |
+	       INDEX(0x09D4));
+	WREG32(MC_CG_CONFIG, reg);
+
+	reg = (READ_WEIGHT(ni_pi->cac_weights->mc_read_weight) |
+	       WRITE_WEIGHT(ni_pi->cac_weights->mc_write_weight) |
+	       ALLOW_OVERFLOW);
+	WREG32(MC_CG_DATAPORT, reg);
+
+	return 0;
+}
+
+static int ni_enable_smc_cac(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_new_state,
+			     bool enable)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret = 0;
+	PPSMC_Result smc_result;
+
+	if (ni_pi->enable_cac) {
+		if (enable) {
+			if (!r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2)) {
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_CollectCAC_PowerCorreln);
+
+				if (ni_pi->support_cac_long_term_average) {
+					smc_result = rv770_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgEnable);
+					if (PPSMC_Result_OK != smc_result)
+						ni_pi->support_cac_long_term_average = false;
+				}
+
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableCac);
+				if (PPSMC_Result_OK != smc_result)
+					ret = -EINVAL;
+
+				ni_pi->cac_enabled = (PPSMC_Result_OK == smc_result) ? true : false;
+			}
+		} else if (ni_pi->cac_enabled) {
+			smc_result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_DisableCac);
+
+			ni_pi->cac_enabled = false;
+
+			if (ni_pi->support_cac_long_term_average) {
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgDisable);
+				if (PPSMC_Result_OK != smc_result)
+					ni_pi->support_cac_long_term_average = false;
+			}
+		}
+	}
+
+	return ret;
+}
+
+static int ni_pcie_performance_request(struct radeon_device *rdev,
+				       u8 perf_req, bool advertise)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+#if defined(CONFIG_ACPI)
+	if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) ||
+            (perf_req == PCIE_PERF_REQ_PECI_GEN2)) {
+		if (eg_pi->pcie_performance_request_registered == false)
+			radeon_acpi_pcie_notify_device_ready(rdev);
+		eg_pi->pcie_performance_request_registered = true;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	} else if ((perf_req == PCIE_PERF_REQ_REMOVE_REGISTRY) &&
+                   eg_pi->pcie_performance_request_registered) {
+		eg_pi->pcie_performance_request_registered = false;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	}
+#endif
+	return 0;
+}
+
+static int ni_advertise_gen2_capability(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+        tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+        if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+            (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+                pi->pcie_gen2 = true;
+        else
+		pi->pcie_gen2 = false;
+
+	if (!pi->pcie_gen2)
+		ni_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, true);
+
+	return 0;
+}
+
+static void ni_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					    bool enable)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        u32 tmp, bif;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+		if (enable) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+			}
+			tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+			tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+			tmp |= LC_GEN2_EN_STRAP;
+
+			tmp |= LC_CLR_FAILED_SPD_CHANGE_CNT;
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+			udelay(10);
+			tmp &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+		} else {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp &= ~LC_GEN2_EN_STRAP;
+			}
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+		}
+	}
+}
+
+static void ni_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					bool enable)
+{
+	ni_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+                WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+void ni_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					   struct radeon_ps *new_ps,
+					   struct radeon_ps *old_ps)
+{
+	struct ni_ps *new_state = ni_get_ps(new_ps);
+	struct ni_ps *current_state = ni_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->performance_levels[new_state->performance_level_count - 1].sclk >=
+	    current_state->performance_levels[current_state->performance_level_count - 1].sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+void ni_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps,
+					  struct radeon_ps *old_ps)
+{
+	struct ni_ps *new_state = ni_get_ps(new_ps);
+	struct ni_ps *current_state = ni_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->performance_levels[new_state->performance_level_count - 1].sclk <
+	    current_state->performance_levels[current_state->performance_level_count - 1].sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+void ni_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	ni_read_clock_registers(rdev);
+	btc_read_arb_registers(rdev);
+	rv770_get_memory_type(rdev);
+	if (eg_pi->pcie_performance_request)
+		ni_advertise_gen2_capability(rdev);
+	rv770_get_pcie_gen2_status(rdev);
+	rv770_enable_acpi_pm(rdev);
+}
+
+void ni_update_current_ps(struct radeon_device *rdev,
+			  struct radeon_ps *rps)
+{
+	struct ni_ps *new_ps = ni_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	eg_pi->current_rps = *rps;
+	ni_pi->current_ps = *new_ps;
+	eg_pi->current_rps.ps_priv = &ni_pi->current_ps;
+}
+
+void ni_update_requested_ps(struct radeon_device *rdev,
+			    struct radeon_ps *rps)
+{
+	struct ni_ps *new_ps = ni_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	eg_pi->requested_rps = *rps;
+	ni_pi->requested_ps = *new_ps;
+	eg_pi->requested_rps.ps_priv = &ni_pi->requested_ps;
+}
+
+int ni_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		ni_cg_clockgating_default(rdev);
+        if (btc_dpm_enabled(rdev))
+                return -EINVAL;
+	if (pi->mg_clock_gating)
+		ni_mg_clockgating_default(rdev);
+	if (eg_pi->ls_clock_gating)
+		ni_ls_clockgating_default(rdev);
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = cypress_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = ni_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+	}
+	if (pi->dynamic_ss)
+		cypress_enable_spread_spectrum(rdev, true);
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	cypress_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+	if (pi->dynamic_pcie_gen2)
+		ni_enable_dynamic_pcie_gen2(rdev, true);
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+	ret = ni_process_firmware_header(rdev);
+	if (ret) {
+		DRM_ERROR("ni_process_firmware_header failed\n");
+		return ret;
+	}
+	ret = ni_initial_switch_from_arb_f0_to_f1(rdev);
+	if (ret) {
+		DRM_ERROR("ni_initial_switch_from_arb_f0_to_f1 failed\n");
+		return ret;
+	}
+	ret = ni_init_smc_table(rdev);
+	if (ret) {
+		DRM_ERROR("ni_init_smc_table failed\n");
+		return ret;
+	}
+	ret = ni_init_smc_spll_table(rdev);
+	if (ret) {
+		DRM_ERROR("ni_init_smc_spll_table failed\n");
+		return ret;
+	}
+	ret = ni_init_arb_table_index(rdev);
+	if (ret) {
+		DRM_ERROR("ni_init_arb_table_index failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = ni_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("ni_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = ni_initialize_smc_cac_tables(rdev);
+	if (ret) {
+		DRM_ERROR("ni_initialize_smc_cac_tables failed\n");
+		return ret;
+	}
+	ret = ni_initialize_hardware_cac_manager(rdev);
+	if (ret) {
+		DRM_ERROR("ni_initialize_hardware_cac_manager failed\n");
+		return ret;
+	}
+	ret = ni_populate_smc_tdp_limits(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("ni_populate_smc_tdp_limits failed\n");
+		return ret;
+	}
+	ni_program_response_times(rdev);
+	r7xx_start_smc(rdev);
+	ret = cypress_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("cypress_notify_smc_display_change failed\n");
+		return ret;
+	}
+	cypress_enable_sclk_control(rdev, true);
+	if (eg_pi->memory_transition)
+		cypress_enable_mclk_control(rdev, true);
+	cypress_start_dpm(rdev);
+	if (pi->gfx_clock_gating)
+		ni_gfx_clockgating_enable(rdev, true);
+	if (pi->mg_clock_gating)
+		ni_mg_clockgating_enable(rdev, true);
+	if (eg_pi->ls_clock_gating)
+		ni_ls_clockgating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, 0xff * 1000);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	ni_update_current_ps(rdev, boot_ps);
+
+	return 0;
+}
+
+void ni_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!btc_dpm_enabled(rdev))
+		return;
+	rv770_clear_vc(rdev);
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+	ni_enable_power_containment(rdev, boot_ps, false);
+	ni_enable_smc_cac(rdev, boot_ps, false);
+	cypress_enable_spread_spectrum(rdev, false);
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, false);
+	if (pi->dynamic_pcie_gen2)
+		ni_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		ni_gfx_clockgating_enable(rdev, false);
+	if (pi->mg_clock_gating)
+		ni_mg_clockgating_enable(rdev, false);
+	if (eg_pi->ls_clock_gating)
+		ni_ls_clockgating_enable(rdev, false);
+	ni_stop_dpm(rdev);
+	btc_reset_to_default(rdev);
+	ni_stop_smc(rdev);
+	ni_force_switch_to_arb_f0(rdev);
+
+	ni_update_current_ps(rdev, boot_ps);
+}
+
+static int ni_power_control_set_level(struct radeon_device *rdev)
+{
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	int ret;
+
+	ret = ni_restrict_performance_levels_before_switch(rdev);
+	if (ret)
+		return ret;
+	ret = rv770_halt_smc(rdev);
+	if (ret)
+		return ret;
+	ret = ni_populate_smc_tdp_limits(rdev, new_ps);
+	if (ret)
+		return ret;
+	ret = rv770_resume_smc(rdev);
+	if (ret)
+		return ret;
+	ret = rv770_set_sw_state(rdev);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+int ni_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	ni_update_requested_ps(rdev, new_ps);
+
+	ni_apply_state_adjust_rules(rdev, &eg_pi->requested_rps);
+
+	return 0;
+}
+
+int ni_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+	struct radeon_ps *old_ps = &eg_pi->current_rps;
+	int ret;
+
+	ret = ni_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("ni_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	ni_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = ni_enable_power_containment(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("ni_enable_power_containment failed\n");
+		return ret;
+	}
+	ret = ni_enable_smc_cac(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("ni_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	if (eg_pi->smu_uvd_hs)
+		btc_notify_uvd_to_smc(rdev, new_ps);
+	ret = ni_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("ni_upload_sw_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = ni_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("ni_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = ni_program_memory_timing_parameters(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("ni_program_memory_timing_parameters failed\n");
+		return ret;
+	}
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	ni_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+	ret = ni_enable_smc_cac(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("ni_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = ni_enable_power_containment(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("ni_enable_power_containment failed\n");
+		return ret;
+	}
+
+	/* update tdp */
+	ret = ni_power_control_set_level(rdev);
+	if (ret) {
+		DRM_ERROR("ni_power_control_set_level failed\n");
+		return ret;
+	}
+
+	ret = ni_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("ni_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void ni_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+
+	ni_update_current_ps(rdev, new_ps);
+}
+
+void ni_dpm_reset_asic(struct radeon_device *rdev)
+{
+	ni_restrict_performance_levels_before_switch(rdev);
+	rv770_set_boot_state(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void ni_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					  struct radeon_ps *rps,
+					  struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					  u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV770_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV770_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void ni_parse_pplib_clock_info(struct radeon_device *rdev,
+				      struct radeon_ps *rps, int index,
+				      union pplib_clock_info *clock_info)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *ps = ni_get_ps(rps);
+	u16 vddc;
+	struct rv7xx_pl *pl = &ps->performance_levels[index];
+
+	ps->performance_level_count = index + 1;
+
+	pl->sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow);
+	pl->sclk |= clock_info->evergreen.ucEngineClockHigh << 16;
+	pl->mclk = le16_to_cpu(clock_info->evergreen.usMemoryClockLow);
+	pl->mclk |= clock_info->evergreen.ucMemoryClockHigh << 16;
+
+	pl->vddc = le16_to_cpu(clock_info->evergreen.usVDDC);
+	pl->vddci = le16_to_cpu(clock_info->evergreen.usVDDCI);
+	pl->flags = le32_to_cpu(clock_info->evergreen.ulFlags);
+
+	/* patch up vddc if necessary */
+	if (pl->vddc == 0xff01) {
+		if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0)
+			pl->vddc = vddc;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) {
+		pi->acpi_vddc = pl->vddc;
+		eg_pi->acpi_vddci = pl->vddci;
+		if (ps->performance_levels[0].flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+			pi->acpi_pcie_gen2 = true;
+		else
+			pi->acpi_pcie_gen2 = false;
+	}
+
+	if (rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) {
+		eg_pi->ulv.supported = true;
+		eg_pi->ulv.pl = pl;
+	}
+
+	if (pi->min_vddc_in_table > pl->vddc)
+		pi->min_vddc_in_table = pl->vddc;
+
+	if (pi->max_vddc_in_table < pl->vddc)
+		pi->max_vddc_in_table = pl->vddc;
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+		pl->vddci = vddci;
+	}
+
+	if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
+	    ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci;
+	}
+}
+
+static int ni_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct ni_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			ps = kzalloc(sizeof(struct ni_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			ni_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info,
+							 power_info->pplib.ucNonClockSize);
+			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+				clock_info = (union pplib_clock_info *)
+					(mode_info->atom_context->bios + data_offset +
+					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+					 (power_state->v1.ucClockStateIndices[j] *
+					  power_info->pplib.ucClockInfoSize));
+				ni_parse_pplib_clock_info(rdev,
+							  &rdev->pm.dpm.ps[i], j,
+							  clock_info);
+			}
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int ni_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	struct ni_power_info *ni_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	u16 data_offset, size;
+	u8 frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	ni_pi = kzalloc(sizeof(struct ni_power_info), GFP_KERNEL);
+	if (ni_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = ni_pi;
+	eg_pi = &ni_pi->eg;
+	pi = &eg_pi->rv7xx;
+
+	rv770_get_max_vddc(rdev);
+
+	eg_pi->ulv.supported = false;
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = ni_parse_power_table(rdev);
+	if (ret)
+		return ret;
+	ret = r600_parse_extended_power_table(rdev);
+	if (ret)
+		return ret;
+
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries =
+		kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL);
+	if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) {
+		r600_free_extended_power_table(rdev);
+		return -ENOMEM;
+	}
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 720;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 810;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 900;
+
+	ni_patch_dependency_tables_based_on_leakage(rdev);
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[0].rlp = RV770_RLP_DFLT;
+	eg_pi->ats[0].rmp = RV770_RMP_DFLT;
+	eg_pi->ats[0].lhp = RV770_LHP_DFLT;
+	eg_pi->ats[0].lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[1].rlp = BTC_RLP_UVD_DFLT;
+	eg_pi->ats[1].rmp = BTC_RMP_UVD_DFLT;
+	eg_pi->ats[1].lhp = BTC_LHP_UVD_DFLT;
+	eg_pi->ats[1].lmp = BTC_LMP_UVD_DFLT;
+
+	eg_pi->smu_uvd_hs = true;
+
+	if (rdev->pdev->device == 0x6707) {
+		pi->mclk_strobe_mode_threshold = 55000;
+		pi->mclk_edc_enable_threshold = 55000;
+		eg_pi->mclk_edc_wr_enable_threshold = 55000;
+	} else {
+		pi->mclk_strobe_mode_threshold = 40000;
+		pi->mclk_edc_enable_threshold = 40000;
+		eg_pi->mclk_edc_wr_enable_threshold = 40000;
+	}
+	ni_pi->mclk_rtt_mode_threshold = eg_pi->mclk_edc_wr_enable_threshold;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = CYPRESS_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+	eg_pi->ls_clock_gating = false;
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	pi->dcodt = true;
+
+	pi->ulps = true;
+
+	eg_pi->dynamic_ac_timing = true;
+	eg_pi->abm = true;
+	eg_pi->mcls = true;
+	eg_pi->light_sleep = true;
+	eg_pi->memory_transition = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	eg_pi->dll_default_on = false;
+
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+
+	rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 3;
+	rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200;
+	rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2 = 900;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.count = ARRAY_SIZE(btc_valid_sclk);
+	rdev->pm.dpm.dyn_state.valid_sclk_values.values = btc_valid_sclk;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL;
+	rdev->pm.dpm.dyn_state.sclk_mclk_delta = 12500;
+
+	ni_pi->cac_data.leakage_coefficients.at = 516;
+	ni_pi->cac_data.leakage_coefficients.bt = 18;
+	ni_pi->cac_data.leakage_coefficients.av = 51;
+	ni_pi->cac_data.leakage_coefficients.bv = 2957;
+
+	switch (rdev->pdev->device) {
+	case 0x6700:
+	case 0x6701:
+	case 0x6702:
+	case 0x6703:
+	case 0x6718:
+		ni_pi->cac_weights = &cac_weights_cayman_xt;
+		break;
+	case 0x6705:
+	case 0x6719:
+	case 0x671D:
+	case 0x671C:
+	default:
+		ni_pi->cac_weights = &cac_weights_cayman_pro;
+		break;
+	case 0x6704:
+	case 0x6706:
+	case 0x6707:
+	case 0x6708:
+	case 0x6709:
+		ni_pi->cac_weights = &cac_weights_cayman_le;
+		break;
+	}
+
+	if (ni_pi->cac_weights->enable_power_containment_by_default) {
+		ni_pi->enable_power_containment = true;
+		ni_pi->enable_cac = true;
+		ni_pi->enable_sq_ramping = true;
+	} else {
+		ni_pi->enable_power_containment = false;
+		ni_pi->enable_cac = false;
+		ni_pi->enable_sq_ramping = false;
+	}
+
+	ni_pi->driver_calculate_cac_leakage = false;
+	ni_pi->cac_configuration_required = true;
+
+	if (ni_pi->cac_configuration_required) {
+		ni_pi->support_cac_long_term_average = true;
+		ni_pi->lta_window_size = ni_pi->cac_weights->l2_lta_window_size;
+		ni_pi->lts_truncate = ni_pi->cac_weights->lts_truncate;
+	} else {
+		ni_pi->support_cac_long_term_average = false;
+		ni_pi->lta_window_size = 0;
+		ni_pi->lts_truncate = 0;
+	}
+
+	ni_pi->use_power_boost_limit = true;
+
+	return 0;
+}
+
+void ni_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+	kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries);
+	r600_free_extended_power_table(rdev);
+}
+
+void ni_dpm_print_power_state(struct radeon_device *rdev,
+			      struct radeon_ps *rps)
+{
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct rv7xx_pl *pl;
+	int i;
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	for (i = 0; i < ps->performance_level_count; i++) {
+		pl = &ps->performance_levels[i];
+		if (rdev->family >= CHIP_TAHITI)
+			printk("\t\tpower level %d    sclk: %u mclk: %u vddc: %u vddci: %u pcie gen: %u\n",
+			       i, pl->sclk, pl->mclk, pl->vddc, pl->vddci, pl->pcie_gen + 1);
+		else
+			printk("\t\tpower level %d    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+			       i, pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void ni_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct rv7xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_INDEX_MASK) >>
+		CURRENT_STATE_INDEX_SHIFT;
+
+	if (current_index >= ps->performance_level_count) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->performance_levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+			   current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+	}
+}
+
+u32 ni_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *requested_state = ni_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->performance_levels[0].sclk;
+	else
+		return requested_state->performance_levels[requested_state->performance_level_count - 1].sclk;
+}
+
+u32 ni_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *requested_state = ni_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->performance_levels[0].mclk;
+	else
+		return requested_state->performance_levels[requested_state->performance_level_count - 1].mclk;
+}
+
diff --git a/drivers/gpu/drm/radeon/ni_dpm.h b/drivers/gpu/drm/radeon/ni_dpm.h
new file mode 100644
index 0000000..6bbee91
--- /dev/null
+++ b/drivers/gpu/drm/radeon/ni_dpm.h
@@ -0,0 +1,250 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __NI_DPM_H__
+#define __NI_DPM_H__
+
+#include "cypress_dpm.h"
+#include "btc_dpm.h"
+#include "nislands_smc.h"
+
+struct ni_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2;
+	u32 mpll_ss1;
+	u32 mpll_ss2;
+};
+
+struct ni_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct ni_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	u16 valid_flag;
+	struct ni_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES];
+	SMC_NIslands_MCRegisterAddress mc_reg_address[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+#define NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT 2
+
+enum ni_dc_cac_level
+{
+	NISLANDS_DCCAC_LEVEL_0 = 0,
+	NISLANDS_DCCAC_LEVEL_1,
+	NISLANDS_DCCAC_LEVEL_2,
+	NISLANDS_DCCAC_LEVEL_3,
+	NISLANDS_DCCAC_LEVEL_4,
+	NISLANDS_DCCAC_LEVEL_5,
+	NISLANDS_DCCAC_LEVEL_6,
+	NISLANDS_DCCAC_LEVEL_7,
+	NISLANDS_DCCAC_MAX_LEVELS
+};
+
+struct ni_leakage_coeffients
+{
+	u32 at;
+	u32 bt;
+	u32 av;
+	u32 bv;
+	s32 t_slope;
+	s32 t_intercept;
+	u32 t_ref;
+};
+
+struct ni_cac_data
+{
+	struct ni_leakage_coeffients leakage_coefficients;
+	u32 i_leakage;
+	s32 leakage_minimum_temperature;
+	u32 pwr_const;
+	u32 dc_cac_value;
+	u32 bif_cac_value;
+	u32 lkge_pwr;
+	u8 mc_wr_weight;
+	u8 mc_rd_weight;
+	u8 allow_ovrflw;
+	u8 num_win_tdp;
+	u8 l2num_win_tdp;
+	u8 lts_truncate_n;
+};
+
+struct ni_cac_weights
+{
+	u32 weight_tcp_sig0;
+	u32 weight_tcp_sig1;
+	u32 weight_ta_sig;
+	u32 weight_tcc_en0;
+	u32 weight_tcc_en1;
+	u32 weight_tcc_en2;
+	u32 weight_cb_en0;
+	u32 weight_cb_en1;
+	u32 weight_cb_en2;
+	u32 weight_cb_en3;
+	u32 weight_db_sig0;
+	u32 weight_db_sig1;
+	u32 weight_db_sig2;
+	u32 weight_db_sig3;
+	u32 weight_sxm_sig0;
+	u32 weight_sxm_sig1;
+	u32 weight_sxm_sig2;
+	u32 weight_sxs_sig0;
+	u32 weight_sxs_sig1;
+	u32 weight_xbr_0;
+	u32 weight_xbr_1;
+	u32 weight_xbr_2;
+	u32 weight_spi_sig0;
+	u32 weight_spi_sig1;
+	u32 weight_spi_sig2;
+	u32 weight_spi_sig3;
+	u32 weight_spi_sig4;
+	u32 weight_spi_sig5;
+	u32 weight_lds_sig0;
+	u32 weight_lds_sig1;
+	u32 weight_sc;
+	u32 weight_bif;
+	u32 weight_cp;
+	u32 weight_pa_sig0;
+	u32 weight_pa_sig1;
+	u32 weight_vgt_sig0;
+	u32 weight_vgt_sig1;
+	u32 weight_vgt_sig2;
+	u32 weight_dc_sig0;
+	u32 weight_dc_sig1;
+	u32 weight_dc_sig2;
+	u32 weight_dc_sig3;
+	u32 weight_uvd_sig0;
+	u32 weight_uvd_sig1;
+	u32 weight_spare0;
+	u32 weight_spare1;
+	u32 weight_sq_vsp;
+	u32 weight_sq_vsp0;
+	u32 weight_sq_gpr;
+	u32 ovr_mode_spare_0;
+	u32 ovr_val_spare_0;
+	u32 ovr_mode_spare_1;
+	u32 ovr_val_spare_1;
+	u32 vsp;
+	u32 vsp0;
+	u32 gpr;
+	u8 mc_read_weight;
+	u8 mc_write_weight;
+	u32 tid_cnt;
+	u32 tid_unit;
+	u32 l2_lta_window_size;
+	u32 lts_truncate;
+	u32 dc_cac[NISLANDS_DCCAC_MAX_LEVELS];
+	u32 pcie_cac[SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES];
+	bool enable_power_containment_by_default;
+};
+
+struct ni_ps {
+	u16 performance_level_count;
+	bool dc_compatible;
+	struct rv7xx_pl performance_levels[NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+};
+
+struct ni_power_info {
+	/* must be first! */
+	struct evergreen_power_info eg;
+	struct ni_clock_registers clock_registers;
+	struct ni_mc_reg_table mc_reg_table;
+	u32 mclk_rtt_mode_threshold;
+	/* flags */
+	bool use_power_boost_limit;
+	bool support_cac_long_term_average;
+	bool cac_enabled;
+	bool cac_configuration_required;
+	bool driver_calculate_cac_leakage;
+	bool pc_enabled;
+	bool enable_power_containment;
+	bool enable_cac;
+	bool enable_sq_ramping;
+	/* smc offsets */
+	u16 arb_table_start;
+	u16 fan_table_start;
+	u16 cac_table_start;
+	u16 spll_table_start;
+	/* CAC stuff */
+	struct ni_cac_data cac_data;
+	u32 dc_cac_table[NISLANDS_DCCAC_MAX_LEVELS];
+	const struct ni_cac_weights *cac_weights;
+	u8 lta_window_size;
+	u8 lts_truncate;
+	struct ni_ps current_ps;
+	struct ni_ps requested_ps;
+	/* scratch structs */
+	SMC_NIslands_MCRegisters smc_mc_reg_table;
+	NISLANDS_SMC_STATETABLE smc_statetable;
+};
+
+#define NISLANDS_INITIAL_STATE_ARB_INDEX    0
+#define NISLANDS_ACPI_STATE_ARB_INDEX       1
+#define NISLANDS_ULV_STATE_ARB_INDEX        2
+#define NISLANDS_DRIVER_STATE_ARB_INDEX     3
+
+#define NISLANDS_DPM2_MAX_PULSE_SKIP        256
+
+#define NISLANDS_DPM2_NEAR_TDP_DEC          10
+#define NISLANDS_DPM2_ABOVE_SAFE_INC        5
+#define NISLANDS_DPM2_BELOW_SAFE_INC        20
+
+#define NISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT            80
+
+#define NISLANDS_DPM2_MAXPS_PERCENT_H                   90
+#define NISLANDS_DPM2_MAXPS_PERCENT_M                   0
+
+#define NISLANDS_DPM2_SQ_RAMP_MAX_POWER                 0x3FFF
+#define NISLANDS_DPM2_SQ_RAMP_MIN_POWER                 0x12
+#define NISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA           0x15
+#define NISLANDS_DPM2_SQ_RAMP_STI_SIZE                  0x1E
+#define NISLANDS_DPM2_SQ_RAMP_LTI_RATIO                 0xF
+
+int ni_copy_and_switch_arb_sets(struct radeon_device *rdev,
+				u32 arb_freq_src, u32 arb_freq_dest);
+void ni_update_current_ps(struct radeon_device *rdev,
+			  struct radeon_ps *rps);
+void ni_update_requested_ps(struct radeon_device *rdev,
+			    struct radeon_ps *rps);
+
+void ni_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					   struct radeon_ps *new_ps,
+					   struct radeon_ps *old_ps);
+void ni_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps,
+					  struct radeon_ps *old_ps);
+
+bool ni_dpm_vblank_too_short(struct radeon_device *rdev);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index e226faf..fe24a93 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -489,6 +489,571 @@
 #       define CACHE_FLUSH_AND_INV_EVENT_TS                     (0x14 << 0)
 #       define CACHE_FLUSH_AND_INV_EVENT                        (0x16 << 0)
 
+/* TN SMU registers */
+#define	TN_CURRENT_GNB_TEMP				0x1F390
+
+/* pm registers */
+#define	SMC_MSG						0x20c
+#define		HOST_SMC_MSG(x)				((x) << 0)
+#define		HOST_SMC_MSG_MASK			(0xff << 0)
+#define		HOST_SMC_MSG_SHIFT			0
+#define		HOST_SMC_RESP(x)			((x) << 8)
+#define		HOST_SMC_RESP_MASK			(0xff << 8)
+#define		HOST_SMC_RESP_SHIFT			8
+#define		SMC_HOST_MSG(x)				((x) << 16)
+#define		SMC_HOST_MSG_MASK			(0xff << 16)
+#define		SMC_HOST_MSG_SHIFT			16
+#define		SMC_HOST_RESP(x)			((x) << 24)
+#define		SMC_HOST_RESP_MASK			(0xff << 24)
+#define		SMC_HOST_RESP_SHIFT			24
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define		SPLL_PDIV_A_SHIFT			20
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_FB_DIV_SHIFT			0
+#define		SPLL_DITHEN				(1 << 28)
+
+#define MPLL_CNTL_MODE                                  0x61c
+#       define SS_SSEN                                  (1 << 24)
+#       define SS_DSMODE_EN                             (1 << 25)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
+#define	MCLK_PWRMGT_CNTL				0x648
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_PDNB                             (1 << 8)
+#       define MRDCKA1_PDNB                             (1 << 9)
+#       define MRDCKB0_PDNB                             (1 << 10)
+#       define MRDCKB1_PDNB                             (1 << 11)
+#       define MRDCKC0_PDNB                             (1 << 12)
+#       define MRDCKC1_PDNB                             (1 << 13)
+#       define MRDCKD0_PDNB                             (1 << 14)
+#       define MRDCKD1_PDNB                             (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x66c
+#       define CURRENT_STATE_INDEX_MASK                   (0xf << 4)
+#       define CURRENT_STATE_INDEX_SHIFT                  4
+
+#define CG_AT                                           0x6d4
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+
+#define	CG_BIF_REQ_AND_RSP				0x7f4
+#define		CG_CLIENT_REQ(x)			((x) << 0)
+#define		CG_CLIENT_REQ_MASK			(0xff << 0)
+#define		CG_CLIENT_REQ_SHIFT			0
+#define		CG_CLIENT_RESP(x)			((x) << 8)
+#define		CG_CLIENT_RESP_MASK			(0xff << 8)
+#define		CG_CLIENT_RESP_SHIFT			8
+#define		CLIENT_CG_REQ(x)			((x) << 16)
+#define		CLIENT_CG_REQ_MASK			(0xff << 16)
+#define		CLIENT_CG_REQ_SHIFT			16
+#define		CLIENT_CG_RESP(x)			((x) << 24)
+#define		CLIENT_CG_RESP_MASK			(0xff << 24)
+#define		CLIENT_CG_RESP_SHIFT			24
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define		CLK_S_SHIFT				4
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+#define		CLK_V_SHIFT				0
+
+#define SMC_SCRATCH0                                    0x81c
+
+#define	CG_SPLL_FUNC_CNTL_4				0x850
+
+#define	MPLL_SS1					0x85c
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x860
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
+#define	CG_CAC_CTRL					0x88c
+#define		TID_CNT(x)				((x) << 0)
+#define		TID_CNT_MASK				(0x3fff << 0)
+#define		TID_UNIT(x)				((x) << 14)
+#define		TID_UNIT_MASK				(0xf << 14)
+
+#define	CG_IND_ADDR					0x8f8
+#define	CG_IND_DATA					0x8fc
+/* CGIND regs */
+#define	CG_CGTT_LOCAL_0					0x00
+#define	CG_CGTT_LOCAL_1					0x01
+
+#define MC_CG_CONFIG                                    0x25bc
+#define         MCDW_WR_ENABLE                          (1 << 0)
+#define         MCDX_WR_ENABLE                          (1 << 1)
+#define         MCDY_WR_ENABLE                          (1 << 2)
+#define         MCDZ_WR_ENABLE                          (1 << 3)
+#define		MC_RD_ENABLE(x)				((x) << 4)
+#define		MC_RD_ENABLE_MASK			(3 << 4)
+#define		INDEX(x)				((x) << 6)
+#define		INDEX_MASK				(0xfff << 6)
+#define		INDEX_SHIFT				6
+
+#define	MC_ARB_CAC_CNTL					0x2750
+#define         ENABLE                                  (1 << 0)
+#define		READ_WEIGHT(x)				((x) << 1)
+#define		READ_WEIGHT_MASK			(0x3f << 1)
+#define		READ_WEIGHT_SHIFT			1
+#define		WRITE_WEIGHT(x)				((x) << 7)
+#define		WRITE_WEIGHT_MASK			(0x3f << 7)
+#define		WRITE_WEIGHT_SHIFT			7
+#define         ALLOW_OVERFLOW                          (1 << 13)
+
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE0_SHIFT			0
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE1_SHIFT			8
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE2_SHIFT			16
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+#define		POWERMODE3_SHIFT			24
+
+#define MC_ARB_CG                                       0x27e8
+#define		CG_ARB_REQ(x)				((x) << 0)
+#define		CG_ARB_REQ_MASK				(0xff << 0)
+#define		CG_ARB_REQ_SHIFT			0
+#define		CG_ARB_RESP(x)				((x) << 8)
+#define		CG_ARB_RESP_MASK			(0xff << 8)
+#define		CG_ARB_RESP_SHIFT			8
+#define		ARB_CG_REQ(x)				((x) << 16)
+#define		ARB_CG_REQ_MASK				(0xff << 16)
+#define		ARB_CG_REQ_SHIFT			16
+#define		ARB_CG_RESP(x)				((x) << 24)
+#define		ARB_CG_RESP_MASK			(0xff << 24)
+#define		ARB_CG_RESP_SHIFT			24
+
+#define	MC_ARB_DRAM_TIMING_1				0x27f0
+#define	MC_ARB_DRAM_TIMING_2				0x27f4
+#define	MC_ARB_DRAM_TIMING_3				0x27f8
+#define	MC_ARB_DRAM_TIMING2_1				0x27fc
+#define	MC_ARB_DRAM_TIMING2_2				0x2800
+#define	MC_ARB_DRAM_TIMING2_3				0x2804
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE0_SHIFT				0
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE1_SHIFT				5
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE2_SHIFT				10
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+#define		STATE3_SHIFT				15
+
+#define MC_CG_DATAPORT                                  0x2884
+
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+#define MC_SEQ_PMG_TIMING                               0x28b0
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_SEQ_MISC0                                    0x2a00
+#define         MC_SEQ_MISC0_GDDR5_SHIFT                28
+#define         MC_SEQ_MISC0_GDDR5_MASK                 0xf0000000
+#define         MC_SEQ_MISC0_GDDR5_VALUE                5
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
+#define MC_SEQ_MISC3                                    0x2a2c
+
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+#define MC_SEQ_PMG_TIMING_LP                            0x2b4c
+
+#define MC_PMG_CMD_MRS2                                 0x2b5c
+#define MC_SEQ_PMG_CMD_MRS2_LP                          0x2b60
+
+#define	LB_SYNC_RESET_SEL				0x6b28
+#define		LB_SYNC_RESET_SEL_MASK			(3 << 0)
+#define		LB_SYNC_RESET_SEL_SHIFT			0
+
+#define	DC_STUTTER_CNTL					0x6b30
+#define		DC_STUTTER_ENABLE_A			(1 << 0)
+#define		DC_STUTTER_ENABLE_B			(1 << 1)
+
+#define SQ_CAC_THRESHOLD                                0x8e4c
+#define		VSP(x)					((x) << 0)
+#define		VSP_MASK				(0xff << 0)
+#define		VSP_SHIFT				0
+#define		VSP0(x)					((x) << 8)
+#define		VSP0_MASK				(0xff << 8)
+#define		VSP0_SHIFT				8
+#define		GPR(x)					((x) << 16)
+#define		GPR_MASK				(0xff << 16)
+#define		GPR_SHIFT				16
+
+#define SQ_POWER_THROTTLE                               0x8e58
+#define		MIN_POWER(x)				((x) << 0)
+#define		MIN_POWER_MASK				(0x3fff << 0)
+#define		MIN_POWER_SHIFT				0
+#define		MAX_POWER(x)				((x) << 16)
+#define		MAX_POWER_MASK				(0x3fff << 16)
+#define		MAX_POWER_SHIFT				0
+#define SQ_POWER_THROTTLE2                              0x8e5c
+#define		MAX_POWER_DELTA(x)			((x) << 0)
+#define		MAX_POWER_DELTA_MASK			(0x3fff << 0)
+#define		MAX_POWER_DELTA_SHIFT			0
+#define		STI_SIZE(x)				((x) << 16)
+#define		STI_SIZE_MASK				(0x3ff << 16)
+#define		STI_SIZE_SHIFT				16
+#define		LTI_RATIO(x)				((x) << 27)
+#define		LTI_RATIO_MASK				(0xf << 27)
+#define		LTI_RATIO_SHIFT				27
+
+/* CG indirect registers */
+#define CG_CAC_REGION_1_WEIGHT_0                        0x83
+#define		WEIGHT_TCP_SIG0(x)			((x) << 0)
+#define		WEIGHT_TCP_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_TCP_SIG0_SHIFT			0
+#define		WEIGHT_TCP_SIG1(x)			((x) << 6)
+#define		WEIGHT_TCP_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_TCP_SIG1_SHIFT			6
+#define		WEIGHT_TA_SIG(x)			((x) << 12)
+#define		WEIGHT_TA_SIG_MASK			(0x3f << 12)
+#define		WEIGHT_TA_SIG_SHIFT			12
+#define CG_CAC_REGION_1_WEIGHT_1                        0x84
+#define		WEIGHT_TCC_EN0(x)			((x) << 0)
+#define		WEIGHT_TCC_EN0_MASK			(0x3f << 0)
+#define		WEIGHT_TCC_EN0_SHIFT			0
+#define		WEIGHT_TCC_EN1(x)			((x) << 6)
+#define		WEIGHT_TCC_EN1_MASK			(0x3f << 6)
+#define		WEIGHT_TCC_EN1_SHIFT			6
+#define		WEIGHT_TCC_EN2(x)			((x) << 12)
+#define		WEIGHT_TCC_EN2_MASK			(0x3f << 12)
+#define		WEIGHT_TCC_EN2_SHIFT			12
+#define		WEIGHT_TCC_EN3(x)			((x) << 18)
+#define		WEIGHT_TCC_EN3_MASK			(0x3f << 18)
+#define		WEIGHT_TCC_EN3_SHIFT			18
+#define CG_CAC_REGION_2_WEIGHT_0                        0x85
+#define		WEIGHT_CB_EN0(x)			((x) << 0)
+#define		WEIGHT_CB_EN0_MASK			(0x3f << 0)
+#define		WEIGHT_CB_EN0_SHIFT			0
+#define		WEIGHT_CB_EN1(x)			((x) << 6)
+#define		WEIGHT_CB_EN1_MASK			(0x3f << 6)
+#define		WEIGHT_CB_EN1_SHIFT			6
+#define		WEIGHT_CB_EN2(x)			((x) << 12)
+#define		WEIGHT_CB_EN2_MASK			(0x3f << 12)
+#define		WEIGHT_CB_EN2_SHIFT			12
+#define		WEIGHT_CB_EN3(x)			((x) << 18)
+#define		WEIGHT_CB_EN3_MASK			(0x3f << 18)
+#define		WEIGHT_CB_EN3_SHIFT			18
+#define CG_CAC_REGION_2_WEIGHT_1                        0x86
+#define		WEIGHT_DB_SIG0(x)			((x) << 0)
+#define		WEIGHT_DB_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_DB_SIG0_SHIFT			0
+#define		WEIGHT_DB_SIG1(x)			((x) << 6)
+#define		WEIGHT_DB_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_DB_SIG1_SHIFT			6
+#define		WEIGHT_DB_SIG2(x)			((x) << 12)
+#define		WEIGHT_DB_SIG2_MASK			(0x3f << 12)
+#define		WEIGHT_DB_SIG2_SHIFT			12
+#define		WEIGHT_DB_SIG3(x)			((x) << 18)
+#define		WEIGHT_DB_SIG3_MASK			(0x3f << 18)
+#define		WEIGHT_DB_SIG3_SHIFT			18
+#define CG_CAC_REGION_2_WEIGHT_2                        0x87
+#define		WEIGHT_SXM_SIG0(x)			((x) << 0)
+#define		WEIGHT_SXM_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_SXM_SIG0_SHIFT			0
+#define		WEIGHT_SXM_SIG1(x)			((x) << 6)
+#define		WEIGHT_SXM_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_SXM_SIG1_SHIFT			6
+#define		WEIGHT_SXM_SIG2(x)			((x) << 12)
+#define		WEIGHT_SXM_SIG2_MASK			(0x3f << 12)
+#define		WEIGHT_SXM_SIG2_SHIFT			12
+#define		WEIGHT_SXS_SIG0(x)			((x) << 18)
+#define		WEIGHT_SXS_SIG0_MASK			(0x3f << 18)
+#define		WEIGHT_SXS_SIG0_SHIFT			18
+#define		WEIGHT_SXS_SIG1(x)			((x) << 24)
+#define		WEIGHT_SXS_SIG1_MASK			(0x3f << 24)
+#define		WEIGHT_SXS_SIG1_SHIFT			24
+#define CG_CAC_REGION_3_WEIGHT_0                        0x88
+#define		WEIGHT_XBR_0(x)				((x) << 0)
+#define		WEIGHT_XBR_0_MASK			(0x3f << 0)
+#define		WEIGHT_XBR_0_SHIFT			0
+#define		WEIGHT_XBR_1(x)				((x) << 6)
+#define		WEIGHT_XBR_1_MASK			(0x3f << 6)
+#define		WEIGHT_XBR_1_SHIFT			6
+#define		WEIGHT_XBR_2(x)				((x) << 12)
+#define		WEIGHT_XBR_2_MASK			(0x3f << 12)
+#define		WEIGHT_XBR_2_SHIFT			12
+#define		WEIGHT_SPI_SIG0(x)			((x) << 18)
+#define		WEIGHT_SPI_SIG0_MASK			(0x3f << 18)
+#define		WEIGHT_SPI_SIG0_SHIFT			18
+#define CG_CAC_REGION_3_WEIGHT_1                        0x89
+#define		WEIGHT_SPI_SIG1(x)			((x) << 0)
+#define		WEIGHT_SPI_SIG1_MASK			(0x3f << 0)
+#define		WEIGHT_SPI_SIG1_SHIFT			0
+#define		WEIGHT_SPI_SIG2(x)			((x) << 6)
+#define		WEIGHT_SPI_SIG2_MASK			(0x3f << 6)
+#define		WEIGHT_SPI_SIG2_SHIFT			6
+#define		WEIGHT_SPI_SIG3(x)			((x) << 12)
+#define		WEIGHT_SPI_SIG3_MASK			(0x3f << 12)
+#define		WEIGHT_SPI_SIG3_SHIFT			12
+#define		WEIGHT_SPI_SIG4(x)			((x) << 18)
+#define		WEIGHT_SPI_SIG4_MASK			(0x3f << 18)
+#define		WEIGHT_SPI_SIG4_SHIFT			18
+#define		WEIGHT_SPI_SIG5(x)			((x) << 24)
+#define		WEIGHT_SPI_SIG5_MASK			(0x3f << 24)
+#define		WEIGHT_SPI_SIG5_SHIFT			24
+#define CG_CAC_REGION_4_WEIGHT_0                        0x8a
+#define		WEIGHT_LDS_SIG0(x)			((x) << 0)
+#define		WEIGHT_LDS_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_LDS_SIG0_SHIFT			0
+#define		WEIGHT_LDS_SIG1(x)			((x) << 6)
+#define		WEIGHT_LDS_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_LDS_SIG1_SHIFT			6
+#define		WEIGHT_SC(x)				((x) << 24)
+#define		WEIGHT_SC_MASK				(0x3f << 24)
+#define		WEIGHT_SC_SHIFT				24
+#define CG_CAC_REGION_4_WEIGHT_1                        0x8b
+#define		WEIGHT_BIF(x)				((x) << 0)
+#define		WEIGHT_BIF_MASK				(0x3f << 0)
+#define		WEIGHT_BIF_SHIFT			0
+#define		WEIGHT_CP(x)				((x) << 6)
+#define		WEIGHT_CP_MASK				(0x3f << 6)
+#define		WEIGHT_CP_SHIFT				6
+#define		WEIGHT_PA_SIG0(x)			((x) << 12)
+#define		WEIGHT_PA_SIG0_MASK			(0x3f << 12)
+#define		WEIGHT_PA_SIG0_SHIFT			12
+#define		WEIGHT_PA_SIG1(x)			((x) << 18)
+#define		WEIGHT_PA_SIG1_MASK			(0x3f << 18)
+#define		WEIGHT_PA_SIG1_SHIFT			18
+#define		WEIGHT_VGT_SIG0(x)			((x) << 24)
+#define		WEIGHT_VGT_SIG0_MASK			(0x3f << 24)
+#define		WEIGHT_VGT_SIG0_SHIFT			24
+#define CG_CAC_REGION_4_WEIGHT_2                        0x8c
+#define		WEIGHT_VGT_SIG1(x)			((x) << 0)
+#define		WEIGHT_VGT_SIG1_MASK			(0x3f << 0)
+#define		WEIGHT_VGT_SIG1_SHIFT			0
+#define		WEIGHT_VGT_SIG2(x)			((x) << 6)
+#define		WEIGHT_VGT_SIG2_MASK			(0x3f << 6)
+#define		WEIGHT_VGT_SIG2_SHIFT			6
+#define		WEIGHT_DC_SIG0(x)			((x) << 12)
+#define		WEIGHT_DC_SIG0_MASK			(0x3f << 12)
+#define		WEIGHT_DC_SIG0_SHIFT			12
+#define		WEIGHT_DC_SIG1(x)			((x) << 18)
+#define		WEIGHT_DC_SIG1_MASK			(0x3f << 18)
+#define		WEIGHT_DC_SIG1_SHIFT			18
+#define		WEIGHT_DC_SIG2(x)			((x) << 24)
+#define		WEIGHT_DC_SIG2_MASK			(0x3f << 24)
+#define		WEIGHT_DC_SIG2_SHIFT			24
+#define CG_CAC_REGION_4_WEIGHT_3                        0x8d
+#define		WEIGHT_DC_SIG3(x)			((x) << 0)
+#define		WEIGHT_DC_SIG3_MASK			(0x3f << 0)
+#define		WEIGHT_DC_SIG3_SHIFT			0
+#define		WEIGHT_UVD_SIG0(x)			((x) << 6)
+#define		WEIGHT_UVD_SIG0_MASK			(0x3f << 6)
+#define		WEIGHT_UVD_SIG0_SHIFT			6
+#define		WEIGHT_UVD_SIG1(x)			((x) << 12)
+#define		WEIGHT_UVD_SIG1_MASK			(0x3f << 12)
+#define		WEIGHT_UVD_SIG1_SHIFT			12
+#define		WEIGHT_SPARE0(x)			((x) << 18)
+#define		WEIGHT_SPARE0_MASK			(0x3f << 18)
+#define		WEIGHT_SPARE0_SHIFT			18
+#define		WEIGHT_SPARE1(x)			((x) << 24)
+#define		WEIGHT_SPARE1_MASK			(0x3f << 24)
+#define		WEIGHT_SPARE1_SHIFT			24
+#define CG_CAC_REGION_5_WEIGHT_0                        0x8e
+#define		WEIGHT_SQ_VSP(x)			((x) << 0)
+#define		WEIGHT_SQ_VSP_MASK			(0x3fff << 0)
+#define		WEIGHT_SQ_VSP_SHIFT			0
+#define		WEIGHT_SQ_VSP0(x)			((x) << 14)
+#define		WEIGHT_SQ_VSP0_MASK			(0x3fff << 14)
+#define		WEIGHT_SQ_VSP0_SHIFT			14
+#define CG_CAC_REGION_4_OVERRIDE_4                      0xab
+#define		OVR_MODE_SPARE_0(x)			((x) << 16)
+#define		OVR_MODE_SPARE_0_MASK			(0x1 << 16)
+#define		OVR_MODE_SPARE_0_SHIFT			16
+#define		OVR_VAL_SPARE_0(x)			((x) << 17)
+#define		OVR_VAL_SPARE_0_MASK			(0x1 << 17)
+#define		OVR_VAL_SPARE_0_SHIFT			17
+#define		OVR_MODE_SPARE_1(x)			((x) << 18)
+#define		OVR_MODE_SPARE_1_MASK			(0x3f << 18)
+#define		OVR_MODE_SPARE_1_SHIFT			18
+#define		OVR_VAL_SPARE_1(x)			((x) << 19)
+#define		OVR_VAL_SPARE_1_MASK			(0x3f << 19)
+#define		OVR_VAL_SPARE_1_SHIFT			19
+#define CG_CAC_REGION_5_WEIGHT_1                        0xb7
+#define		WEIGHT_SQ_GPR(x)			((x) << 0)
+#define		WEIGHT_SQ_GPR_MASK			(0x3fff << 0)
+#define		WEIGHT_SQ_GPR_SHIFT			0
+#define		WEIGHT_SQ_LDS(x)			((x) << 14)
+#define		WEIGHT_SQ_LDS_MASK			(0x3fff << 14)
+#define		WEIGHT_SQ_LDS_SHIFT			14
+
+/* PCIE link stuff */
+#define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
+#define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
+#       define LC_LINK_WIDTH_SHIFT                        0
+#       define LC_LINK_WIDTH_MASK                         0x7
+#       define LC_LINK_WIDTH_X0                           0
+#       define LC_LINK_WIDTH_X1                           1
+#       define LC_LINK_WIDTH_X2                           2
+#       define LC_LINK_WIDTH_X4                           3
+#       define LC_LINK_WIDTH_X8                           4
+#       define LC_LINK_WIDTH_X16                          6
+#       define LC_LINK_WIDTH_RD_SHIFT                     4
+#       define LC_LINK_WIDTH_RD_MASK                      0x70
+#       define LC_RECONFIG_ARC_MISSING_ESCAPE             (1 << 7)
+#       define LC_RECONFIG_NOW                            (1 << 8)
+#       define LC_RENEGOTIATION_SUPPORT                   (1 << 9)
+#       define LC_RENEGOTIATE_EN                          (1 << 10)
+#       define LC_SHORT_RECONFIG_EN                       (1 << 11)
+#       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
+#       define LC_UPCONFIGURE_DIS                         (1 << 13)
+#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
+#       define LC_GEN2_EN_STRAP                           (1 << 0)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 1)
+#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 5)
+#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 6)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
+#       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
+#       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
+#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 24)
+#define MM_CFGREGS_CNTL                                   0x544c
+#       define MM_WR_TO_CFG_EN                            (1 << 3)
+#define LINK_CNTL2                                        0x88 /* F0 */
+#       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
+#       define SELECTABLE_DEEMPHASIS                      (1 << 6)
+
 /*
  * UVD
  */
diff --git a/drivers/gpu/drm/radeon/nislands_smc.h b/drivers/gpu/drm/radeon/nislands_smc.h
new file mode 100644
index 0000000..3cf8fc0
--- /dev/null
+++ b/drivers/gpu/drm/radeon/nislands_smc.h
@@ -0,0 +1,329 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __NISLANDS_SMC_H__
+#define __NISLANDS_SMC_H__
+
+#pragma pack(push, 1)
+
+#define NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE 16
+
+struct PP_NIslands_Dpm2PerfLevel
+{
+    uint8_t     MaxPS;
+    uint8_t     TgtAct;
+    uint8_t     MaxPS_StepInc;
+    uint8_t     MaxPS_StepDec;
+    uint8_t     PSST;
+    uint8_t     NearTDPDec;
+    uint8_t     AboveSafeInc;
+    uint8_t     BelowSafeInc;
+    uint8_t     PSDeltaLimit;
+    uint8_t     PSDeltaWin;
+    uint8_t     Reserved[6];
+};
+
+typedef struct PP_NIslands_Dpm2PerfLevel PP_NIslands_Dpm2PerfLevel;
+
+struct PP_NIslands_DPM2Parameters
+{
+    uint32_t    TDPLimit;
+    uint32_t    NearTDPLimit;
+    uint32_t    SafePowerLimit;
+    uint32_t    PowerBoostLimit;
+};
+typedef struct PP_NIslands_DPM2Parameters PP_NIslands_DPM2Parameters;
+
+struct NISLANDS_SMC_SCLK_VALUE
+{
+    uint32_t        vCG_SPLL_FUNC_CNTL;
+    uint32_t        vCG_SPLL_FUNC_CNTL_2;
+    uint32_t        vCG_SPLL_FUNC_CNTL_3;
+    uint32_t        vCG_SPLL_FUNC_CNTL_4;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM_2;
+    uint32_t        sclk_value;
+};
+
+typedef struct NISLANDS_SMC_SCLK_VALUE NISLANDS_SMC_SCLK_VALUE;
+
+struct NISLANDS_SMC_MCLK_VALUE
+{
+    uint32_t        vMPLL_FUNC_CNTL;
+    uint32_t        vMPLL_FUNC_CNTL_1;
+    uint32_t        vMPLL_FUNC_CNTL_2;
+    uint32_t        vMPLL_AD_FUNC_CNTL;
+    uint32_t        vMPLL_AD_FUNC_CNTL_2;
+    uint32_t        vMPLL_DQ_FUNC_CNTL;
+    uint32_t        vMPLL_DQ_FUNC_CNTL_2;
+    uint32_t        vMCLK_PWRMGT_CNTL;
+    uint32_t        vDLL_CNTL;
+    uint32_t        vMPLL_SS;
+    uint32_t        vMPLL_SS2;
+    uint32_t        mclk_value;
+};
+
+typedef struct NISLANDS_SMC_MCLK_VALUE NISLANDS_SMC_MCLK_VALUE;
+
+struct NISLANDS_SMC_VOLTAGE_VALUE
+{
+    uint16_t             value;
+    uint8_t              index;
+    uint8_t              padding;
+};
+
+typedef struct NISLANDS_SMC_VOLTAGE_VALUE NISLANDS_SMC_VOLTAGE_VALUE;
+
+struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL
+{
+    uint8_t                     arbValue;
+    uint8_t                     ACIndex;
+    uint8_t                     displayWatermark;
+    uint8_t                     gen2PCIE;
+    uint8_t                     reserved1;
+    uint8_t                     reserved2;
+    uint8_t                     strobeMode;
+    uint8_t                     mcFlags;
+    uint32_t                    aT;
+    uint32_t                    bSP;
+    NISLANDS_SMC_SCLK_VALUE     sclk;
+    NISLANDS_SMC_MCLK_VALUE     mclk;
+    NISLANDS_SMC_VOLTAGE_VALUE  vddc;
+    NISLANDS_SMC_VOLTAGE_VALUE  mvdd;
+    NISLANDS_SMC_VOLTAGE_VALUE  vddci;
+    NISLANDS_SMC_VOLTAGE_VALUE  std_vddc;
+    uint32_t                    powergate_en;
+    uint8_t                     hUp;
+    uint8_t                     hDown;
+    uint8_t                     stateFlags;
+    uint8_t                     arbRefreshState;
+    uint32_t                    SQPowerThrottle;
+    uint32_t                    SQPowerThrottle_2;
+    uint32_t                    reserved[2];
+    PP_NIslands_Dpm2PerfLevel   dpm2;
+};
+
+#define NISLANDS_SMC_STROBE_RATIO    0x0F
+#define NISLANDS_SMC_STROBE_ENABLE   0x10
+
+#define NISLANDS_SMC_MC_EDC_RD_FLAG  0x01
+#define NISLANDS_SMC_MC_EDC_WR_FLAG  0x02
+#define NISLANDS_SMC_MC_RTT_ENABLE   0x04
+#define NISLANDS_SMC_MC_STUTTER_EN   0x08
+
+typedef struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL NISLANDS_SMC_HW_PERFORMANCE_LEVEL;
+
+struct NISLANDS_SMC_SWSTATE
+{
+    uint8_t                             flags;
+    uint8_t                             levelCount;
+    uint8_t                             padding2;
+    uint8_t                             padding3;
+    NISLANDS_SMC_HW_PERFORMANCE_LEVEL   levels[1];
+};
+
+typedef struct NISLANDS_SMC_SWSTATE NISLANDS_SMC_SWSTATE;
+
+#define NISLANDS_SMC_VOLTAGEMASK_VDDC  0
+#define NISLANDS_SMC_VOLTAGEMASK_MVDD  1
+#define NISLANDS_SMC_VOLTAGEMASK_VDDCI 2
+#define NISLANDS_SMC_VOLTAGEMASK_MAX   4
+
+struct NISLANDS_SMC_VOLTAGEMASKTABLE
+{
+    uint8_t  highMask[NISLANDS_SMC_VOLTAGEMASK_MAX];
+    uint32_t lowMask[NISLANDS_SMC_VOLTAGEMASK_MAX];
+};
+
+typedef struct NISLANDS_SMC_VOLTAGEMASKTABLE NISLANDS_SMC_VOLTAGEMASKTABLE;
+
+#define NISLANDS_MAX_NO_VREG_STEPS 32
+
+struct NISLANDS_SMC_STATETABLE
+{
+    uint8_t                             thermalProtectType;
+    uint8_t                             systemFlags;
+    uint8_t                             maxVDDCIndexInPPTable;
+    uint8_t                             extraFlags;
+    uint8_t                             highSMIO[NISLANDS_MAX_NO_VREG_STEPS];
+    uint32_t                            lowSMIO[NISLANDS_MAX_NO_VREG_STEPS];
+    NISLANDS_SMC_VOLTAGEMASKTABLE       voltageMaskTable;
+    PP_NIslands_DPM2Parameters          dpm2Params;
+    NISLANDS_SMC_SWSTATE                initialState;
+    NISLANDS_SMC_SWSTATE                ACPIState;
+    NISLANDS_SMC_SWSTATE                ULVState;
+    NISLANDS_SMC_SWSTATE                driverState;
+    NISLANDS_SMC_HW_PERFORMANCE_LEVEL   dpmLevels[NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1];
+};
+
+typedef struct NISLANDS_SMC_STATETABLE NISLANDS_SMC_STATETABLE;
+
+#define NI_SMC_SOFT_REGISTERS_START        0x108
+
+#define NI_SMC_SOFT_REGISTER_mclk_chg_timeout        0x0
+#define NI_SMC_SOFT_REGISTER_delay_bbias             0xC
+#define NI_SMC_SOFT_REGISTER_delay_vreg              0x10
+#define NI_SMC_SOFT_REGISTER_delay_acpi              0x2C
+#define NI_SMC_SOFT_REGISTER_seq_index               0x64
+#define NI_SMC_SOFT_REGISTER_mvdd_chg_time           0x68
+#define NI_SMC_SOFT_REGISTER_mclk_switch_lim         0x78
+#define NI_SMC_SOFT_REGISTER_watermark_threshold     0x80
+#define NI_SMC_SOFT_REGISTER_mc_block_delay          0x84
+#define NI_SMC_SOFT_REGISTER_uvd_enabled             0x98
+
+#define SMC_NISLANDS_MC_TPP_CAC_NUM_OF_ENTRIES 16
+#define SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
+#define SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 16
+#define SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES 4
+
+struct SMC_NISLANDS_MC_TPP_CAC_TABLE
+{
+    uint32_t    tpp[SMC_NISLANDS_MC_TPP_CAC_NUM_OF_ENTRIES];
+    uint32_t    cacValue[SMC_NISLANDS_MC_TPP_CAC_NUM_OF_ENTRIES];
+};
+
+typedef struct SMC_NISLANDS_MC_TPP_CAC_TABLE SMC_NISLANDS_MC_TPP_CAC_TABLE;
+
+
+struct PP_NIslands_CACTABLES
+{
+    uint32_t                cac_bif_lut[SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES];
+    uint32_t                cac_lkge_lut[SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES][SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES];
+
+    uint32_t                pwr_const;
+
+    uint32_t                dc_cacValue;
+    uint32_t                bif_cacValue;
+    uint32_t                lkge_pwr;
+
+    uint8_t                 cac_width;
+    uint8_t                 window_size_p2;
+
+    uint8_t                 num_drop_lsb;
+    uint8_t                 padding_0;
+
+    uint32_t                last_power;
+
+    uint8_t                 AllowOvrflw;
+    uint8_t                 MCWrWeight;
+    uint8_t                 MCRdWeight;
+    uint8_t                 padding_1[9];
+
+    uint8_t                 enableWinAvg;
+    uint8_t                 numWin_TDP;
+    uint8_t                 l2numWin_TDP;
+    uint8_t                 WinIndex;
+
+    uint32_t                dynPwr_TDP[4];
+    uint32_t                lkgePwr_TDP[4];
+    uint32_t                power_TDP[4];
+    uint32_t                avg_dynPwr_TDP;
+    uint32_t                avg_lkgePwr_TDP;
+    uint32_t                avg_power_TDP;
+    uint32_t                lts_power_TDP;
+    uint8_t                 lts_truncate_n;
+    uint8_t                 padding_2[7];
+};
+
+typedef struct PP_NIslands_CACTABLES PP_NIslands_CACTABLES;
+
+#define SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE 32
+#define SMC_NISLANDS_MC_REGISTER_ARRAY_SET_COUNT 20
+
+struct SMC_NIslands_MCRegisterAddress
+{
+    uint16_t s0;
+    uint16_t s1;
+};
+
+typedef struct SMC_NIslands_MCRegisterAddress SMC_NIslands_MCRegisterAddress;
+
+
+struct SMC_NIslands_MCRegisterSet
+{
+    uint32_t value[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+typedef struct SMC_NIslands_MCRegisterSet SMC_NIslands_MCRegisterSet;
+
+struct SMC_NIslands_MCRegisters
+{
+    uint8_t                             last;
+    uint8_t                             reserved[3];
+    SMC_NIslands_MCRegisterAddress      address[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+    SMC_NIslands_MCRegisterSet          data[SMC_NISLANDS_MC_REGISTER_ARRAY_SET_COUNT];
+};
+
+typedef struct SMC_NIslands_MCRegisters SMC_NIslands_MCRegisters;
+
+struct SMC_NIslands_MCArbDramTimingRegisterSet
+{
+    uint32_t mc_arb_dram_timing;
+    uint32_t mc_arb_dram_timing2;
+    uint8_t  mc_arb_rfsh_rate;
+    uint8_t  padding[3];
+};
+
+typedef struct SMC_NIslands_MCArbDramTimingRegisterSet SMC_NIslands_MCArbDramTimingRegisterSet;
+
+struct SMC_NIslands_MCArbDramTimingRegisters
+{
+    uint8_t                                     arb_current;
+    uint8_t                                     reserved[3];
+    SMC_NIslands_MCArbDramTimingRegisterSet     data[20];
+};
+
+typedef struct SMC_NIslands_MCArbDramTimingRegisters SMC_NIslands_MCArbDramTimingRegisters;
+
+struct SMC_NISLANDS_SPLL_DIV_TABLE
+{
+    uint32_t    freq[256];
+    uint32_t    ss[256];
+};
+
+#define SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_MASK  0x01ffffff
+#define SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT 0
+#define SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_MASK   0xfe000000
+#define SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT  25
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_MASK   0x000fffff
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT  0
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK   0xfff00000
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT  20
+
+typedef struct SMC_NISLANDS_SPLL_DIV_TABLE SMC_NISLANDS_SPLL_DIV_TABLE;
+
+#define NISLANDS_SMC_FIRMWARE_HEADER_LOCATION 0x100
+
+#define NISLANDS_SMC_FIRMWARE_HEADER_version                   0x0
+#define NISLANDS_SMC_FIRMWARE_HEADER_flags                     0x4
+#define NISLANDS_SMC_FIRMWARE_HEADER_softRegisters             0x8
+#define NISLANDS_SMC_FIRMWARE_HEADER_stateTable                0xC
+#define NISLANDS_SMC_FIRMWARE_HEADER_fanTable                  0x10
+#define NISLANDS_SMC_FIRMWARE_HEADER_cacTable                  0x14
+#define NISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable           0x20
+#define NISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable 0x2C
+#define NISLANDS_SMC_FIRMWARE_HEADER_spllTable                 0x30
+
+#pragma pack(pop)
+
+#endif
+
diff --git a/drivers/gpu/drm/radeon/ppsmc.h b/drivers/gpu/drm/radeon/ppsmc.h
new file mode 100644
index 0000000..b5564a3
--- /dev/null
+++ b/drivers/gpu/drm/radeon/ppsmc.h
@@ -0,0 +1,116 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef PP_SMC_H
+#define PP_SMC_H
+
+#pragma pack(push, 1)
+
+#define PPSMC_SWSTATE_FLAG_DC                           0x01
+#define PPSMC_SWSTATE_FLAG_UVD                          0x02
+#define PPSMC_SWSTATE_FLAG_VCE                          0x04
+#define PPSMC_SWSTATE_FLAG_PCIE_X1                      0x08
+
+#define PPSMC_THERMAL_PROTECT_TYPE_INTERNAL             0x00
+#define PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL             0x01
+#define PPSMC_THERMAL_PROTECT_TYPE_NONE                 0xff
+
+#define PPSMC_SYSTEMFLAG_GPIO_DC                        0x01
+#define PPSMC_SYSTEMFLAG_STEPVDDC                       0x02
+#define PPSMC_SYSTEMFLAG_GDDR5                          0x04
+#define PPSMC_SYSTEMFLAG_DISABLE_BABYSTEP               0x08
+#define PPSMC_SYSTEMFLAG_REGULATOR_HOT                  0x10
+#define PPSMC_SYSTEMFLAG_REGULATOR_HOT_ANALOG           0x20
+#define PPSMC_SYSTEMFLAG_REGULATOR_HOT_PROG_GPIO        0x40
+
+#define PPSMC_EXTRAFLAGS_AC2DC_ACTION_MASK              0x07
+#define PPSMC_EXTRAFLAGS_AC2DC_DONT_WAIT_FOR_VBLANK     0x08
+#define PPSMC_EXTRAFLAGS_AC2DC_ACTION_GOTODPMLOWSTATE   0x00
+#define PPSMC_EXTRAFLAGS_AC2DC_ACTION_GOTOINITIALSTATE  0x01
+#define PPSMC_EXTRAFLAGS_AC2DC_GPIO5_POLARITY_HIGH      0x02
+
+#define PPSMC_DISPLAY_WATERMARK_LOW                     0
+#define PPSMC_DISPLAY_WATERMARK_HIGH                    1
+
+#define PPSMC_STATEFLAG_AUTO_PULSE_SKIP    0x01
+#define PPSMC_STATEFLAG_POWERBOOST         0x02
+#define PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE 0x20
+#define PPSMC_STATEFLAG_DEEPSLEEP_BYPASS   0x40
+
+#define PPSMC_Result_OK             ((uint8_t)0x01)
+#define PPSMC_Result_Failed         ((uint8_t)0xFF)
+
+typedef uint8_t PPSMC_Result;
+
+#define PPSMC_MSG_Halt                      ((uint8_t)0x10)
+#define PPSMC_MSG_Resume                    ((uint8_t)0x11)
+#define PPSMC_MSG_ZeroLevelsDisabled        ((uint8_t)0x13)
+#define PPSMC_MSG_OneLevelsDisabled         ((uint8_t)0x14)
+#define PPSMC_MSG_TwoLevelsDisabled         ((uint8_t)0x15)
+#define PPSMC_MSG_EnableThermalInterrupt    ((uint8_t)0x16)
+#define PPSMC_MSG_RunningOnAC               ((uint8_t)0x17)
+#define PPSMC_MSG_SwitchToSwState           ((uint8_t)0x20)
+#define PPSMC_MSG_SwitchToInitialState      ((uint8_t)0x40)
+#define PPSMC_MSG_NoForcedLevel             ((uint8_t)0x41)
+#define PPSMC_MSG_ForceHigh                 ((uint8_t)0x42)
+#define PPSMC_MSG_ForceMediumOrHigh         ((uint8_t)0x43)
+#define PPSMC_MSG_SwitchToMinimumPower      ((uint8_t)0x51)
+#define PPSMC_MSG_ResumeFromMinimumPower    ((uint8_t)0x52)
+#define PPSMC_MSG_EnableCac                 ((uint8_t)0x53)
+#define PPSMC_MSG_DisableCac                ((uint8_t)0x54)
+#define PPSMC_TDPClampingActive             ((uint8_t)0x59)
+#define PPSMC_TDPClampingInactive           ((uint8_t)0x5A)
+#define PPSMC_MSG_NoDisplay                 ((uint8_t)0x5D)
+#define PPSMC_MSG_HasDisplay                ((uint8_t)0x5E)
+#define PPSMC_MSG_UVDPowerOFF               ((uint8_t)0x60)
+#define PPSMC_MSG_UVDPowerON                ((uint8_t)0x61)
+#define PPSMC_MSG_EnableULV                 ((uint8_t)0x62)
+#define PPSMC_MSG_DisableULV                ((uint8_t)0x63)
+#define PPSMC_MSG_EnterULV                  ((uint8_t)0x64)
+#define PPSMC_MSG_ExitULV                   ((uint8_t)0x65)
+#define PPSMC_CACLongTermAvgEnable          ((uint8_t)0x6E)
+#define PPSMC_CACLongTermAvgDisable         ((uint8_t)0x6F)
+#define PPSMC_MSG_CollectCAC_PowerCorreln   ((uint8_t)0x7A)
+#define PPSMC_FlushDataCache                ((uint8_t)0x80)
+#define PPSMC_MSG_SetEnabledLevels          ((uint8_t)0x82)
+#define PPSMC_MSG_SetForcedLevels           ((uint8_t)0x83)
+#define PPSMC_MSG_ResetToDefaults           ((uint8_t)0x84)
+#define PPSMC_MSG_EnableDTE                 ((uint8_t)0x87)
+#define PPSMC_MSG_DisableDTE                ((uint8_t)0x88)
+#define PPSMC_MSG_ThrottleOVRDSCLKDS        ((uint8_t)0x96)
+#define PPSMC_MSG_CancelThrottleOVRDSCLKDS  ((uint8_t)0x97)
+
+/* TN */
+#define PPSMC_MSG_DPM_Config                ((uint32_t) 0x102)
+#define PPSMC_MSG_DPM_ForceState            ((uint32_t) 0x104)
+#define PPSMC_MSG_PG_SIMD_Config            ((uint32_t) 0x108)
+#define PPSMC_MSG_DPM_N_LevelsDisabled      ((uint32_t) 0x112)
+#define PPSMC_MSG_DCE_RemoveVoltageAdjustment   ((uint32_t) 0x11d)
+#define PPSMC_MSG_DCE_AllowVoltageAdjustment    ((uint32_t) 0x11e)
+#define PPSMC_MSG_UVD_DPM_Config            ((uint32_t) 0x124)
+
+
+typedef uint16_t PPSMC_Msg;
+
+#pragma pack(pop)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index d0314ec..c9affef 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -3077,6 +3077,10 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 			flags |= RADEON_SURF_TILE_COLOR_BOTH;
 		if (tiling_flags & RADEON_TILING_MACRO)
 			flags |= RADEON_SURF_TILE_COLOR_MACRO;
+		/* setting pitch to 0 disables tiling */
+		if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO))
+				== 0)
+			pitch = 0;
 	} else if (rdev->family <= CHIP_RV280) {
 		if (tiling_flags & (RADEON_TILING_MACRO))
 			flags |= R200_SURF_TILE_COLOR_MACRO;
@@ -3094,13 +3098,6 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 	if (tiling_flags & RADEON_TILING_SWAP_32BIT)
 		flags |= RADEON_SURF_AP0_SWP_32BPP | RADEON_SURF_AP1_SWP_32BPP;
 
-	/* when we aren't tiling the pitch seems to needs to be furtherdivided down. - tested on power5 + rn50 server */
-	if (tiling_flags & (RADEON_TILING_SWAP_16BIT | RADEON_TILING_SWAP_32BIT)) {
-		if (!(tiling_flags & (RADEON_TILING_MACRO | RADEON_TILING_MICRO)))
-			if (ASIC_IS_RN50(rdev))
-				pitch /= 16;
-	}
-
 	/* r100/r200 divide by 16 */
 	if (rdev->family < CHIP_R300)
 		flags |= pitch / 16;
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 6948eb8..2d3655f 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -38,18 +38,7 @@
 #include "r600d.h"
 #include "atom.h"
 #include "avivod.h"
-
-#define PFP_UCODE_SIZE 576
-#define PM4_UCODE_SIZE 1792
-#define RLC_UCODE_SIZE 768
-#define R700_PFP_UCODE_SIZE 848
-#define R700_PM4_UCODE_SIZE 1360
-#define R700_RLC_UCODE_SIZE 1024
-#define EVERGREEN_PFP_UCODE_SIZE 1120
-#define EVERGREEN_PM4_UCODE_SIZE 1376
-#define EVERGREEN_RLC_UCODE_SIZE 768
-#define CAYMAN_RLC_UCODE_SIZE 1024
-#define ARUBA_RLC_UCODE_SIZE 1536
+#include "radeon_ucode.h"
 
 /* Firmware Names */
 MODULE_FIRMWARE("radeon/R600_pfp.bin");
@@ -68,24 +57,32 @@ MODULE_FIRMWARE("radeon/RS780_pfp.bin");
 MODULE_FIRMWARE("radeon/RS780_me.bin");
 MODULE_FIRMWARE("radeon/RV770_pfp.bin");
 MODULE_FIRMWARE("radeon/RV770_me.bin");
+MODULE_FIRMWARE("radeon/RV770_smc.bin");
 MODULE_FIRMWARE("radeon/RV730_pfp.bin");
 MODULE_FIRMWARE("radeon/RV730_me.bin");
+MODULE_FIRMWARE("radeon/RV730_smc.bin");
+MODULE_FIRMWARE("radeon/RV740_smc.bin");
 MODULE_FIRMWARE("radeon/RV710_pfp.bin");
 MODULE_FIRMWARE("radeon/RV710_me.bin");
+MODULE_FIRMWARE("radeon/RV710_smc.bin");
 MODULE_FIRMWARE("radeon/R600_rlc.bin");
 MODULE_FIRMWARE("radeon/R700_rlc.bin");
 MODULE_FIRMWARE("radeon/CEDAR_pfp.bin");
 MODULE_FIRMWARE("radeon/CEDAR_me.bin");
 MODULE_FIRMWARE("radeon/CEDAR_rlc.bin");
+MODULE_FIRMWARE("radeon/CEDAR_smc.bin");
 MODULE_FIRMWARE("radeon/REDWOOD_pfp.bin");
 MODULE_FIRMWARE("radeon/REDWOOD_me.bin");
 MODULE_FIRMWARE("radeon/REDWOOD_rlc.bin");
+MODULE_FIRMWARE("radeon/REDWOOD_smc.bin");
 MODULE_FIRMWARE("radeon/JUNIPER_pfp.bin");
 MODULE_FIRMWARE("radeon/JUNIPER_me.bin");
 MODULE_FIRMWARE("radeon/JUNIPER_rlc.bin");
+MODULE_FIRMWARE("radeon/JUNIPER_smc.bin");
 MODULE_FIRMWARE("radeon/CYPRESS_pfp.bin");
 MODULE_FIRMWARE("radeon/CYPRESS_me.bin");
 MODULE_FIRMWARE("radeon/CYPRESS_rlc.bin");
+MODULE_FIRMWARE("radeon/CYPRESS_smc.bin");
 MODULE_FIRMWARE("radeon/PALM_pfp.bin");
 MODULE_FIRMWARE("radeon/PALM_me.bin");
 MODULE_FIRMWARE("radeon/SUMO_rlc.bin");
@@ -108,6 +105,7 @@ static void r600_gpu_init(struct radeon_device *rdev);
 void r600_fini(struct radeon_device *rdev);
 void r600_irq_disable(struct radeon_device *rdev);
 static void r600_pcie_gen2_enable(struct radeon_device *rdev);
+extern int evergreen_rlc_resume(struct radeon_device *rdev);
 
 /**
  * r600_get_xclk - get the xclk
@@ -2149,7 +2147,8 @@ int r600_init_microcode(struct radeon_device *rdev)
 	struct platform_device *pdev;
 	const char *chip_name;
 	const char *rlc_chip_name;
-	size_t pfp_req_size, me_req_size, rlc_req_size;
+	const char *smc_chip_name = "RV770";
+	size_t pfp_req_size, me_req_size, rlc_req_size, smc_req_size = 0;
 	char fw_name[30];
 	int err;
 
@@ -2195,32 +2194,51 @@ int r600_init_microcode(struct radeon_device *rdev)
 	case CHIP_RV770:
 		chip_name = "RV770";
 		rlc_chip_name = "R700";
+		smc_chip_name = "RV770";
+		smc_req_size = ALIGN(RV770_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_RV730:
-	case CHIP_RV740:
 		chip_name = "RV730";
 		rlc_chip_name = "R700";
+		smc_chip_name = "RV730";
+		smc_req_size = ALIGN(RV730_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_RV710:
 		chip_name = "RV710";
 		rlc_chip_name = "R700";
+		smc_chip_name = "RV710";
+		smc_req_size = ALIGN(RV710_SMC_UCODE_SIZE, 4);
+		break;
+	case CHIP_RV740:
+		chip_name = "RV730";
+		rlc_chip_name = "R700";
+		smc_chip_name = "RV740";
+		smc_req_size = ALIGN(RV740_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CEDAR:
 		chip_name = "CEDAR";
 		rlc_chip_name = "CEDAR";
+		smc_chip_name = "CEDAR";
+		smc_req_size = ALIGN(CEDAR_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_REDWOOD:
 		chip_name = "REDWOOD";
 		rlc_chip_name = "REDWOOD";
+		smc_chip_name = "REDWOOD";
+		smc_req_size = ALIGN(REDWOOD_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_JUNIPER:
 		chip_name = "JUNIPER";
 		rlc_chip_name = "JUNIPER";
+		smc_chip_name = "JUNIPER";
+		smc_req_size = ALIGN(JUNIPER_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CYPRESS:
 	case CHIP_HEMLOCK:
 		chip_name = "CYPRESS";
 		rlc_chip_name = "CYPRESS";
+		smc_chip_name = "CYPRESS";
+		smc_req_size = ALIGN(CYPRESS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_PALM:
 		chip_name = "PALM";
@@ -2246,9 +2264,9 @@ int r600_init_microcode(struct radeon_device *rdev)
 		me_req_size = R700_PM4_UCODE_SIZE * 4;
 		rlc_req_size = R700_RLC_UCODE_SIZE * 4;
 	} else {
-		pfp_req_size = PFP_UCODE_SIZE * 4;
-		me_req_size = PM4_UCODE_SIZE * 12;
-		rlc_req_size = RLC_UCODE_SIZE * 4;
+		pfp_req_size = R600_PFP_UCODE_SIZE * 4;
+		me_req_size = R600_PM4_UCODE_SIZE * 12;
+		rlc_req_size = R600_RLC_UCODE_SIZE * 4;
 	}
 
 	DRM_INFO("Loading %s Microcode\n", chip_name);
@@ -2287,6 +2305,19 @@ int r600_init_microcode(struct radeon_device *rdev)
 		err = -EINVAL;
 	}
 
+	if ((rdev->family >= CHIP_RV770) && (rdev->family <= CHIP_HEMLOCK)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", smc_chip_name);
+		err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev);
+		if (err)
+			goto out;
+		if (rdev->smc_fw->size != smc_req_size) {
+			printk(KERN_ERR
+			       "smc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->smc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	}
+
 out:
 	platform_device_unregister(pdev);
 
@@ -2301,6 +2332,8 @@ out:
 		rdev->me_fw = NULL;
 		release_firmware(rdev->rlc_fw);
 		rdev->rlc_fw = NULL;
+		release_firmware(rdev->smc_fw);
+		rdev->smc_fw = NULL;
 	}
 	return err;
 }
@@ -2331,13 +2364,13 @@ static int r600_cp_load_microcode(struct radeon_device *rdev)
 
 	fw_data = (const __be32 *)rdev->me_fw->data;
 	WREG32(CP_ME_RAM_WADDR, 0);
-	for (i = 0; i < PM4_UCODE_SIZE * 3; i++)
+	for (i = 0; i < R600_PM4_UCODE_SIZE * 3; i++)
 		WREG32(CP_ME_RAM_DATA,
 		       be32_to_cpup(fw_data++));
 
 	fw_data = (const __be32 *)rdev->pfp_fw->data;
 	WREG32(CP_PFP_UCODE_ADDR, 0);
-	for (i = 0; i < PFP_UCODE_SIZE; i++)
+	for (i = 0; i < R600_PFP_UCODE_SIZE; i++)
 		WREG32(CP_PFP_UCODE_DATA,
 		       be32_to_cpup(fw_data++));
 
@@ -3789,7 +3822,7 @@ static void r600_rlc_start(struct radeon_device *rdev)
 	WREG32(RLC_CNTL, RLC_ENABLE);
 }
 
-static int r600_rlc_init(struct radeon_device *rdev)
+static int r600_rlc_resume(struct radeon_device *rdev)
 {
 	u32 i;
 	const __be32 *fw_data;
@@ -3801,45 +3834,22 @@ static int r600_rlc_init(struct radeon_device *rdev)
 
 	WREG32(RLC_HB_CNTL, 0);
 
-	if (rdev->family == CHIP_ARUBA) {
-		WREG32(TN_RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
-		WREG32(TN_RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
-	}
-	if (rdev->family <= CHIP_CAYMAN) {
-		WREG32(RLC_HB_BASE, 0);
-		WREG32(RLC_HB_RPTR, 0);
-		WREG32(RLC_HB_WPTR, 0);
-	}
-	if (rdev->family <= CHIP_CAICOS) {
-		WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
-		WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
-	}
+	WREG32(RLC_HB_BASE, 0);
+	WREG32(RLC_HB_RPTR, 0);
+	WREG32(RLC_HB_WPTR, 0);
+	WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
+	WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
 
 	fw_data = (const __be32 *)rdev->rlc_fw->data;
-	if (rdev->family >= CHIP_ARUBA) {
-		for (i = 0; i < ARUBA_RLC_UCODE_SIZE; i++) {
-			WREG32(RLC_UCODE_ADDR, i);
-			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
-		}
-	} else if (rdev->family >= CHIP_CAYMAN) {
-		for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) {
-			WREG32(RLC_UCODE_ADDR, i);
-			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
-		}
-	} else if (rdev->family >= CHIP_CEDAR) {
-		for (i = 0; i < EVERGREEN_RLC_UCODE_SIZE; i++) {
-			WREG32(RLC_UCODE_ADDR, i);
-			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
-		}
-	} else if (rdev->family >= CHIP_RV770) {
+	if (rdev->family >= CHIP_RV770) {
 		for (i = 0; i < R700_RLC_UCODE_SIZE; i++) {
 			WREG32(RLC_UCODE_ADDR, i);
 			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
 		}
 	} else {
-		for (i = 0; i < RLC_UCODE_SIZE; i++) {
+		for (i = 0; i < R600_RLC_UCODE_SIZE; i++) {
 			WREG32(RLC_UCODE_ADDR, i);
 			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
 		}
@@ -3947,7 +3957,10 @@ int r600_irq_init(struct radeon_device *rdev)
 	r600_disable_interrupts(rdev);
 
 	/* init rlc */
-	ret = r600_rlc_init(rdev);
+	if (rdev->family >= CHIP_CEDAR)
+		ret = evergreen_rlc_resume(rdev);
+	else
+		ret = r600_rlc_resume(rdev);
 	if (ret) {
 		r600_ih_ring_fini(rdev);
 		return ret;
@@ -4028,6 +4041,7 @@ int r600_irq_set(struct radeon_device *rdev)
 	u32 hdmi0, hdmi1;
 	u32 d1grph = 0, d2grph = 0;
 	u32 dma_cntl;
+	u32 thermal_int = 0;
 
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
@@ -4062,8 +4076,21 @@ int r600_irq_set(struct radeon_device *rdev)
 		hdmi0 = RREG32(HDMI0_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK;
 		hdmi1 = RREG32(HDMI1_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK;
 	}
+
 	dma_cntl = RREG32(DMA_CNTL) & ~TRAP_ENABLE;
 
+	if ((rdev->family > CHIP_R600) && (rdev->family < CHIP_RV770)) {
+		thermal_int = RREG32(CG_THERMAL_INT) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+	} else if (rdev->family >= CHIP_RV770) {
+		thermal_int = RREG32(RV770_CG_THERMAL_INT) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+	}
+	if (rdev->irq.dpm_thermal) {
+		DRM_DEBUG("dpm thermal\n");
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+	}
+
 	if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) {
 		DRM_DEBUG("r600_irq_set: sw int\n");
 		cp_int_cntl |= RB_INT_ENABLE;
@@ -4145,6 +4172,11 @@ int r600_irq_set(struct radeon_device *rdev)
 		WREG32(HDMI0_AUDIO_PACKET_CONTROL, hdmi0);
 		WREG32(HDMI1_AUDIO_PACKET_CONTROL, hdmi1);
 	}
+	if ((rdev->family > CHIP_R600) && (rdev->family < CHIP_RV770)) {
+		WREG32(CG_THERMAL_INT, thermal_int);
+	} else if (rdev->family >= CHIP_RV770) {
+		WREG32(RV770_CG_THERMAL_INT, thermal_int);
+	}
 
 	return 0;
 }
@@ -4336,6 +4368,7 @@ int r600_irq_process(struct radeon_device *rdev)
 	u32 ring_index;
 	bool queue_hotplug = false;
 	bool queue_hdmi = false;
+	bool queue_thermal = false;
 
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
@@ -4503,6 +4536,16 @@ restart_ih:
 			DRM_DEBUG("IH: DMA trap\n");
 			radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
 			break;
+		case 230: /* thermal low to high */
+			DRM_DEBUG("IH: thermal low to high\n");
+			rdev->pm.dpm.thermal.high_to_low = false;
+			queue_thermal = true;
+			break;
+		case 231: /* thermal high to low */
+			DRM_DEBUG("IH: thermal high to low\n");
+			rdev->pm.dpm.thermal.high_to_low = true;
+			queue_thermal = true;
+			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
 			break;
@@ -4519,6 +4562,8 @@ restart_ih:
 		schedule_work(&rdev->hotplug_work);
 	if (queue_hdmi)
 		schedule_work(&rdev->audio_work);
+	if (queue_thermal && rdev->pm.dpm_enabled)
+		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
 	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c
new file mode 100644
index 0000000..b88f54b
--- /dev/null
+++ b/drivers/gpu/drm/radeon/r600_dpm.c
@@ -0,0 +1,1048 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "r600d.h"
+#include "r600_dpm.h"
+#include "atom.h"
+
+const u32 r600_utc[R600_PM_NUMBER_OF_TC] =
+{
+	R600_UTC_DFLT_00,
+	R600_UTC_DFLT_01,
+	R600_UTC_DFLT_02,
+	R600_UTC_DFLT_03,
+	R600_UTC_DFLT_04,
+	R600_UTC_DFLT_05,
+	R600_UTC_DFLT_06,
+	R600_UTC_DFLT_07,
+	R600_UTC_DFLT_08,
+	R600_UTC_DFLT_09,
+	R600_UTC_DFLT_10,
+	R600_UTC_DFLT_11,
+	R600_UTC_DFLT_12,
+	R600_UTC_DFLT_13,
+	R600_UTC_DFLT_14,
+};
+
+const u32 r600_dtc[R600_PM_NUMBER_OF_TC] =
+{
+	R600_DTC_DFLT_00,
+	R600_DTC_DFLT_01,
+	R600_DTC_DFLT_02,
+	R600_DTC_DFLT_03,
+	R600_DTC_DFLT_04,
+	R600_DTC_DFLT_05,
+	R600_DTC_DFLT_06,
+	R600_DTC_DFLT_07,
+	R600_DTC_DFLT_08,
+	R600_DTC_DFLT_09,
+	R600_DTC_DFLT_10,
+	R600_DTC_DFLT_11,
+	R600_DTC_DFLT_12,
+	R600_DTC_DFLT_13,
+	R600_DTC_DFLT_14,
+};
+
+void r600_dpm_print_class_info(u32 class, u32 class2)
+{
+	printk("\tui class: ");
+	switch (class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) {
+	case ATOM_PPLIB_CLASSIFICATION_UI_NONE:
+	default:
+		printk("none\n");
+		break;
+	case ATOM_PPLIB_CLASSIFICATION_UI_BATTERY:
+		printk("battery\n");
+		break;
+	case ATOM_PPLIB_CLASSIFICATION_UI_BALANCED:
+		printk("balanced\n");
+		break;
+	case ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE:
+		printk("performance\n");
+		break;
+	}
+	printk("\tinternal class: ");
+	if (((class & ~ATOM_PPLIB_CLASSIFICATION_UI_MASK) == 0) &&
+	    (class2 == 0))
+		printk("none");
+	else {
+		if (class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+			printk("boot ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+			printk("thermal ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_LIMITEDPOWERSOURCE)
+			printk("limited_pwr ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_REST)
+			printk("rest ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_FORCED)
+			printk("forced ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_3DPERFORMANCE)
+			printk("3d_perf ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_OVERDRIVETEMPLATE)
+			printk("ovrdrv ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+			printk("uvd ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_3DLOW)
+			printk("3d_low ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_ACPI)
+			printk("acpi ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_HD2STATE)
+			printk("uvd_hd2 ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_HDSTATE)
+			printk("uvd_hd ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_SDSTATE)
+			printk("uvd_sd ");
+		if (class2 & ATOM_PPLIB_CLASSIFICATION2_LIMITEDPOWERSOURCE_2)
+			printk("limited_pwr2 ");
+		if (class2 & ATOM_PPLIB_CLASSIFICATION2_ULV)
+			printk("ulv ");
+		if (class2 & ATOM_PPLIB_CLASSIFICATION2_MVC)
+			printk("uvd_mvc ");
+	}
+	printk("\n");
+}
+
+void r600_dpm_print_cap_info(u32 caps)
+{
+	printk("\tcaps: ");
+	if (caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY)
+		printk("single_disp ");
+	if (caps & ATOM_PPLIB_SUPPORTS_VIDEO_PLAYBACK)
+		printk("video ");
+	if (caps & ATOM_PPLIB_DISALLOW_ON_DC)
+		printk("no_dc ");
+	printk("\n");
+}
+
+void r600_dpm_print_ps_status(struct radeon_device *rdev,
+			      struct radeon_ps *rps)
+{
+	printk("\tstatus: ");
+	if (rps == rdev->pm.dpm.current_ps)
+		printk("c ");
+	if (rps == rdev->pm.dpm.requested_ps)
+		printk("r ");
+	if (rps == rdev->pm.dpm.boot_ps)
+		printk("b ");
+	printk("\n");
+}
+
+u32 r600_dpm_get_vblank_time(struct radeon_device *rdev)
+{
+	struct drm_device *dev = rdev->ddev;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+	u32 line_time_us, vblank_lines;
+	u32 vblank_time_us = 0xffffffff; /* if the displays are off, vblank time is max */
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		radeon_crtc = to_radeon_crtc(crtc);
+		if (crtc->enabled && radeon_crtc->enabled && radeon_crtc->hw_mode.clock) {
+			line_time_us = (radeon_crtc->hw_mode.crtc_htotal * 1000) /
+				radeon_crtc->hw_mode.clock;
+			vblank_lines = radeon_crtc->hw_mode.crtc_vblank_end -
+				radeon_crtc->hw_mode.crtc_vdisplay +
+				(radeon_crtc->v_border * 2);
+			vblank_time_us = vblank_lines * line_time_us;
+			break;
+		}
+	}
+
+	return vblank_time_us;
+}
+
+void r600_calculate_u_and_p(u32 i, u32 r_c, u32 p_b,
+			    u32 *p, u32 *u)
+{
+	u32 b_c = 0;
+	u32 i_c;
+	u32 tmp;
+
+	i_c = (i * r_c) / 100;
+	tmp = i_c >> p_b;
+
+	while (tmp) {
+		b_c++;
+		tmp >>= 1;
+	}
+
+	*u = (b_c + 1) / 2;
+	*p = i_c / (1 << (2 * (*u)));
+}
+
+int r600_calculate_at(u32 t, u32 h, u32 fh, u32 fl, u32 *tl, u32 *th)
+{
+	u32 k, a, ah, al;
+	u32 t1;
+
+	if ((fl == 0) || (fh == 0) || (fl > fh))
+		return -EINVAL;
+
+	k = (100 * fh) / fl;
+	t1 = (t * (k - 100));
+	a = (1000 * (100 * h + t1)) / (10000 + (t1 / 100));
+	a = (a + 5) / 10;
+	ah = ((a * t) + 5000) / 10000;
+	al = a - ah;
+
+	*th = t - ah;
+	*tl = t + al;
+
+	return 0;
+}
+
+void r600_gfx_clockgating_enable(struct radeon_device *rdev, bool enable)
+{
+	int i;
+
+	if (enable) {
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+
+		WREG32(CG_RLC_REQ_AND_RSP, 0x2);
+
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (((RREG32(CG_RLC_REQ_AND_RSP) & CG_RLC_RSP_TYPE_MASK) >> CG_RLC_RSP_TYPE_SHIFT) == 1)
+				break;
+			udelay(1);
+		}
+
+		WREG32(CG_RLC_REQ_AND_RSP, 0x0);
+
+		WREG32(GRBM_PWR_CNTL, 0x1);
+		RREG32(GRBM_PWR_CNTL);
+	}
+}
+
+void r600_dynamicpm_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+}
+
+void r600_enable_thermal_protection(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	else
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+}
+
+void r600_enable_acpi_pm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+void r600_enable_dynamic_pcie_gen2(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+bool r600_dynamicpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32(GENERAL_PWRMGT) & GLOBAL_PWRMGT_EN)
+		return true;
+	else
+		return false;
+}
+
+void r600_enable_sclk_control(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~SCLK_PWRMGT_OFF);
+	else
+		WREG32_P(GENERAL_PWRMGT, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+}
+
+void r600_enable_mclk_control(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+	else
+		WREG32_P(MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+void r600_enable_spll_bypass(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_SPLL_FUNC_CNTL, SPLL_BYPASS_EN, ~SPLL_BYPASS_EN);
+	else
+		WREG32_P(CG_SPLL_FUNC_CNTL, 0, ~SPLL_BYPASS_EN);
+}
+
+void r600_wait_for_spll_change(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(CG_SPLL_FUNC_CNTL) & SPLL_CHG_STATUS)
+			break;
+		udelay(1);
+	}
+}
+
+void r600_set_bsp(struct radeon_device *rdev, u32 u, u32 p)
+{
+	WREG32(CG_BSP, BSP(p) | BSU(u));
+}
+
+void r600_set_at(struct radeon_device *rdev,
+		 u32 l_to_m, u32 m_to_h,
+		 u32 h_to_m, u32 m_to_l)
+{
+	WREG32(CG_RT, FLS(l_to_m) | FMS(m_to_h));
+	WREG32(CG_LT, FHS(h_to_m) | FMS(m_to_l));
+}
+
+void r600_set_tc(struct radeon_device *rdev,
+		 u32 index, u32 u_t, u32 d_t)
+{
+	WREG32(CG_FFCT_0 + (index * 4), UTC_0(u_t) | DTC_0(d_t));
+}
+
+void r600_select_td(struct radeon_device *rdev,
+		    enum r600_td td)
+{
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+void r600_set_vrc(struct radeon_device *rdev, u32 vrv)
+{
+	WREG32(CG_FTV, vrv);
+}
+
+void r600_set_tpu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_TPC, TPU(u), ~TPU_MASK);
+}
+
+void r600_set_tpc(struct radeon_device *rdev, u32 c)
+{
+	WREG32_P(CG_TPC, TPCC(c), ~TPCC_MASK);
+}
+
+void r600_set_sstu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_SSP, CG_SSTU(u), ~CG_SSTU_MASK);
+}
+
+void r600_set_sst(struct radeon_device *rdev, u32 t)
+{
+	WREG32_P(CG_SSP, CG_SST(t), ~CG_SST_MASK);
+}
+
+void r600_set_git(struct radeon_device *rdev, u32 t)
+{
+	WREG32_P(CG_GIT, CG_GICST(t), ~CG_GICST_MASK);
+}
+
+void r600_set_fctu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_FC_T, FC_TU(u), ~FC_TU_MASK);
+}
+
+void r600_set_fct(struct radeon_device *rdev, u32 t)
+{
+	WREG32_P(CG_FC_T, FC_T(t), ~FC_T_MASK);
+}
+
+void r600_set_ctxcgtt3d_rphc(struct radeon_device *rdev, u32 p)
+{
+	WREG32_P(CG_CTX_CGTT3D_R, PHC(p), ~PHC_MASK);
+}
+
+void r600_set_ctxcgtt3d_rsdc(struct radeon_device *rdev, u32 s)
+{
+	WREG32_P(CG_CTX_CGTT3D_R, SDC(s), ~SDC_MASK);
+}
+
+void r600_set_vddc3d_oorsu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_VDDC3D_OOR, SU(u), ~SU_MASK);
+}
+
+void r600_set_vddc3d_oorphc(struct radeon_device *rdev, u32 p)
+{
+	WREG32_P(CG_VDDC3D_OOR, PHC(p), ~PHC_MASK);
+}
+
+void r600_set_vddc3d_oorsdc(struct radeon_device *rdev, u32 s)
+{
+	WREG32_P(CG_VDDC3D_OOR, SDC(s), ~SDC_MASK);
+}
+
+void r600_set_mpll_lock_time(struct radeon_device *rdev, u32 lock_time)
+{
+	WREG32_P(MPLL_TIME, MPLL_LOCK_TIME(lock_time), ~MPLL_LOCK_TIME_MASK);
+}
+
+void r600_set_mpll_reset_time(struct radeon_device *rdev, u32 reset_time)
+{
+	WREG32_P(MPLL_TIME, MPLL_RESET_TIME(reset_time), ~MPLL_RESET_TIME_MASK);
+}
+
+void r600_engine_clock_entry_enable(struct radeon_device *rdev,
+				    u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 STEP_0_SPLL_ENTRY_VALID, ~STEP_0_SPLL_ENTRY_VALID);
+	else
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 0, ~STEP_0_SPLL_ENTRY_VALID);
+}
+
+void r600_engine_clock_entry_enable_pulse_skipping(struct radeon_device *rdev,
+						   u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 STEP_0_SPLL_STEP_ENABLE, ~STEP_0_SPLL_STEP_ENABLE);
+	else
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 0, ~STEP_0_SPLL_STEP_ENABLE);
+}
+
+void r600_engine_clock_entry_enable_post_divider(struct radeon_device *rdev,
+						 u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 STEP_0_POST_DIV_EN, ~STEP_0_POST_DIV_EN);
+	else
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 0, ~STEP_0_POST_DIV_EN);
+}
+
+void r600_engine_clock_entry_set_post_divider(struct radeon_device *rdev,
+					      u32 index, u32 divider)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_POST_DIV(divider), ~STEP_0_SPLL_POST_DIV_MASK);
+}
+
+void r600_engine_clock_entry_set_reference_divider(struct radeon_device *rdev,
+						   u32 index, u32 divider)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_REF_DIV(divider), ~STEP_0_SPLL_REF_DIV_MASK);
+}
+
+void r600_engine_clock_entry_set_feedback_divider(struct radeon_device *rdev,
+						  u32 index, u32 divider)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_FB_DIV(divider), ~STEP_0_SPLL_FB_DIV_MASK);
+}
+
+void r600_engine_clock_entry_set_step_time(struct radeon_device *rdev,
+					   u32 index, u32 step_time)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_STEP_TIME(step_time), ~STEP_0_SPLL_STEP_TIME_MASK);
+}
+
+void r600_vid_rt_set_ssu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(VID_RT, SSTU(u), ~SSTU_MASK);
+}
+
+void r600_vid_rt_set_vru(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(VID_RT, VID_CRTU(u), ~VID_CRTU_MASK);
+}
+
+void r600_vid_rt_set_vrt(struct radeon_device *rdev, u32 rt)
+{
+	WREG32_P(VID_RT, VID_CRT(rt), ~VID_CRT_MASK);
+}
+
+void r600_voltage_control_enable_pins(struct radeon_device *rdev,
+				      u64 mask)
+{
+	WREG32(LOWER_GPIO_ENABLE, mask & 0xffffffff);
+	WREG32(UPPER_GPIO_ENABLE, upper_32_bits(mask));
+}
+
+
+void r600_voltage_control_program_voltages(struct radeon_device *rdev,
+					   enum r600_power_level index, u64 pins)
+{
+	u32 tmp, mask;
+	u32 ix = 3 - (3 & index);
+
+	WREG32(CTXSW_VID_LOWER_GPIO_CNTL + (ix * 4), pins & 0xffffffff);
+
+	mask = 7 << (3 * ix);
+	tmp = RREG32(VID_UPPER_GPIO_CNTL);
+	tmp = (tmp & ~mask) | ((pins >> (32 - (3 * ix))) & mask);
+	WREG32(VID_UPPER_GPIO_CNTL, tmp);
+}
+
+void r600_voltage_control_deactivate_static_control(struct radeon_device *rdev,
+						    u64 mask)
+{
+	u32 gpio;
+
+	gpio = RREG32(GPIOPAD_MASK);
+	gpio &= ~mask;
+	WREG32(GPIOPAD_MASK, gpio);
+
+	gpio = RREG32(GPIOPAD_EN);
+	gpio &= ~mask;
+	WREG32(GPIOPAD_EN, gpio);
+
+	gpio = RREG32(GPIOPAD_A);
+	gpio &= ~mask;
+	WREG32(GPIOPAD_A, gpio);
+}
+
+void r600_power_level_enable(struct radeon_device *rdev,
+			     enum r600_power_level index, bool enable)
+{
+	u32 ix = 3 - (3 & index);
+
+	if (enable)
+		WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), CTXSW_FREQ_STATE_ENABLE,
+			 ~CTXSW_FREQ_STATE_ENABLE);
+	else
+		WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), 0,
+			 ~CTXSW_FREQ_STATE_ENABLE);
+}
+
+void r600_power_level_set_voltage_index(struct radeon_device *rdev,
+					enum r600_power_level index, u32 voltage_index)
+{
+	u32 ix = 3 - (3 & index);
+
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4),
+		 CTXSW_FREQ_VIDS_CFG_INDEX(voltage_index), ~CTXSW_FREQ_VIDS_CFG_INDEX_MASK);
+}
+
+void r600_power_level_set_mem_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 mem_clock_index)
+{
+	u32 ix = 3 - (3 & index);
+
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4),
+		 CTXSW_FREQ_MCLK_CFG_INDEX(mem_clock_index), ~CTXSW_FREQ_MCLK_CFG_INDEX_MASK);
+}
+
+void r600_power_level_set_eng_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 eng_clock_index)
+{
+	u32 ix = 3 - (3 & index);
+
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4),
+		 CTXSW_FREQ_SCLK_CFG_INDEX(eng_clock_index), ~CTXSW_FREQ_SCLK_CFG_INDEX_MASK);
+}
+
+void r600_power_level_set_watermark_id(struct radeon_device *rdev,
+				       enum r600_power_level index,
+				       enum r600_display_watermark watermark_id)
+{
+	u32 ix = 3 - (3 & index);
+	u32 tmp = 0;
+
+	if (watermark_id == R600_DISPLAY_WATERMARK_HIGH)
+		tmp = CTXSW_FREQ_DISPLAY_WATERMARK;
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), tmp, ~CTXSW_FREQ_DISPLAY_WATERMARK);
+}
+
+void r600_power_level_set_pcie_gen2(struct radeon_device *rdev,
+				    enum r600_power_level index, bool compatible)
+{
+	u32 ix = 3 - (3 & index);
+	u32 tmp = 0;
+
+	if (compatible)
+		tmp = CTXSW_FREQ_GEN2PCIE_VOLT;
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), tmp, ~CTXSW_FREQ_GEN2PCIE_VOLT);
+}
+
+enum r600_power_level r600_power_level_get_current_index(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK;
+	tmp >>= CURRENT_PROFILE_INDEX_SHIFT;
+	return tmp;
+}
+
+enum r600_power_level r600_power_level_get_target_index(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & TARGET_PROFILE_INDEX_MASK;
+	tmp >>= TARGET_PROFILE_INDEX_SHIFT;
+	return tmp;
+}
+
+void r600_power_level_set_enter_index(struct radeon_device *rdev,
+				      enum r600_power_level index)
+{
+	WREG32_P(TARGET_AND_CURRENT_PROFILE_INDEX, DYN_PWR_ENTER_INDEX(index),
+		 ~DYN_PWR_ENTER_INDEX_MASK);
+}
+
+void r600_wait_for_power_level_unequal(struct radeon_device *rdev,
+				       enum r600_power_level index)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_target_index(rdev) != index)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_current_index(rdev) != index)
+			break;
+		udelay(1);
+	}
+}
+
+void r600_wait_for_power_level(struct radeon_device *rdev,
+			       enum r600_power_level index)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_target_index(rdev) == index)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_current_index(rdev) == index)
+			break;
+		udelay(1);
+	}
+}
+
+void r600_start_dpm(struct radeon_device *rdev)
+{
+	r600_enable_sclk_control(rdev, false);
+	r600_enable_mclk_control(rdev, false);
+
+	r600_dynamicpm_enable(rdev, true);
+
+	radeon_wait_for_vblank(rdev, 0);
+	radeon_wait_for_vblank(rdev, 1);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_sclk_control(rdev, true);
+	r600_enable_mclk_control(rdev, true);
+}
+
+void r600_stop_dpm(struct radeon_device *rdev)
+{
+	r600_dynamicpm_enable(rdev, false);
+}
+
+int r600_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	return 0;
+}
+
+void r600_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+
+}
+
+bool r600_is_uvd_state(u32 class, u32 class2)
+{
+	if (class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		return true;
+	if (class & ATOM_PPLIB_CLASSIFICATION_HD2STATE)
+		return true;
+	if (class & ATOM_PPLIB_CLASSIFICATION_HDSTATE)
+		return true;
+	if (class & ATOM_PPLIB_CLASSIFICATION_SDSTATE)
+		return true;
+	if (class2 & ATOM_PPLIB_CLASSIFICATION2_MVC)
+		return true;
+	return false;
+}
+
+int r600_set_thermal_temperature_range(struct radeon_device *rdev,
+				       int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK);
+	WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor)
+{
+	switch (sensor) {
+	case THERMAL_TYPE_RV6XX:
+	case THERMAL_TYPE_RV770:
+	case THERMAL_TYPE_EVERGREEN:
+	case THERMAL_TYPE_SUMO:
+	case THERMAL_TYPE_NI:
+	case THERMAL_TYPE_SI:
+		return true;
+	case THERMAL_TYPE_ADT7473_WITH_INTERNAL:
+	case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		return false; /* need special handling */
+	case THERMAL_TYPE_NONE:
+	case THERMAL_TYPE_EXTERNAL:
+	case THERMAL_TYPE_EXTERNAL_GPIO:
+	default:
+		return false;
+	}
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE4 pplib4;
+	struct _ATOM_PPLIB_POWERPLAYTABLE5 pplib5;
+};
+
+union fan_info {
+	struct _ATOM_PPLIB_FANTABLE fan;
+	struct _ATOM_PPLIB_FANTABLE2 fan2;
+};
+
+static int r600_parse_clk_voltage_dep_table(struct radeon_clock_voltage_dependency_table *radeon_table,
+					    ATOM_PPLIB_Clock_Voltage_Dependency_Table *atom_table)
+{
+	u32 size = atom_table->ucNumEntries *
+		sizeof(struct radeon_clock_voltage_dependency_entry);
+	int i;
+
+	radeon_table->entries = kzalloc(size, GFP_KERNEL);
+	if (!radeon_table->entries)
+		return -ENOMEM;
+
+	for (i = 0; i < atom_table->ucNumEntries; i++) {
+		radeon_table->entries[i].clk = le16_to_cpu(atom_table->entries[i].usClockLow) |
+			(atom_table->entries[i].ucClockHigh << 16);
+		radeon_table->entries[i].v = le16_to_cpu(atom_table->entries[i].usVoltage);
+	}
+	radeon_table->count = atom_table->ucNumEntries;
+
+	return 0;
+}
+
+/* sizeof(ATOM_PPLIB_EXTENDEDHEADER) */
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V2 12
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V3 14
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V4 16
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V5 18
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V6 20
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V7 22
+
+int r600_parse_extended_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	union power_info *power_info;
+	union fan_info *fan_info;
+	ATOM_PPLIB_Clock_Voltage_Dependency_Table *dep_table;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	int ret, i;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	/* fan table */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE3)) {
+		if (power_info->pplib3.usFanTableOffset) {
+			fan_info = (union fan_info *)(mode_info->atom_context->bios + data_offset +
+						      le16_to_cpu(power_info->pplib3.usFanTableOffset));
+			rdev->pm.dpm.fan.t_hyst = fan_info->fan.ucTHyst;
+			rdev->pm.dpm.fan.t_min = le16_to_cpu(fan_info->fan.usTMin);
+			rdev->pm.dpm.fan.t_med = le16_to_cpu(fan_info->fan.usTMed);
+			rdev->pm.dpm.fan.t_high = le16_to_cpu(fan_info->fan.usTHigh);
+			rdev->pm.dpm.fan.pwm_min = le16_to_cpu(fan_info->fan.usPWMMin);
+			rdev->pm.dpm.fan.pwm_med = le16_to_cpu(fan_info->fan.usPWMMed);
+			rdev->pm.dpm.fan.pwm_high = le16_to_cpu(fan_info->fan.usPWMHigh);
+			if (fan_info->fan.ucFanTableFormat >= 2)
+				rdev->pm.dpm.fan.t_max = le16_to_cpu(fan_info->fan2.usTMax);
+			else
+				rdev->pm.dpm.fan.t_max = 10900;
+			rdev->pm.dpm.fan.cycle_delay = 100000;
+			rdev->pm.dpm.fan.ucode_fan_control = true;
+		}
+	}
+
+	/* clock dependancy tables, shedding tables */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE4)) {
+		if (power_info->pplib4.usVddcDependencyOnSCLKOffset) {
+			dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddcDependencyOnSCLKOffset));
+			ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+							       dep_table);
+			if (ret)
+				return ret;
+		}
+		if (power_info->pplib4.usVddciDependencyOnMCLKOffset) {
+			dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddciDependencyOnMCLKOffset));
+			ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+							       dep_table);
+			if (ret) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				return ret;
+			}
+		}
+		if (power_info->pplib4.usVddcDependencyOnMCLKOffset) {
+			dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddcDependencyOnMCLKOffset));
+			ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+							       dep_table);
+			if (ret) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				return ret;
+			}
+		}
+		if (power_info->pplib4.usMaxClockVoltageOnDCOffset) {
+			ATOM_PPLIB_Clock_Voltage_Limit_Table *clk_v =
+				(ATOM_PPLIB_Clock_Voltage_Limit_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usMaxClockVoltageOnDCOffset));
+			if (clk_v->ucNumEntries) {
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk =
+					le16_to_cpu(clk_v->entries[0].usSclkLow) |
+					(clk_v->entries[0].ucSclkHigh << 16);
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk =
+					le16_to_cpu(clk_v->entries[0].usMclkLow) |
+					(clk_v->entries[0].ucMclkHigh << 16);
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc =
+					le16_to_cpu(clk_v->entries[0].usVddc);
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddci =
+					le16_to_cpu(clk_v->entries[0].usVddci);
+			}
+		}
+		if (power_info->pplib4.usVddcPhaseShedLimitsTableOffset) {
+			ATOM_PPLIB_PhaseSheddingLimits_Table *psl =
+				(ATOM_PPLIB_PhaseSheddingLimits_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddcPhaseShedLimitsTableOffset));
+
+			rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries =
+				kzalloc(psl->ucNumEntries *
+					sizeof(struct radeon_phase_shedding_limits_entry),
+					GFP_KERNEL);
+			if (!rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+				return -ENOMEM;
+			}
+
+			for (i = 0; i < psl->ucNumEntries; i++) {
+				rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].sclk =
+					le16_to_cpu(psl->entries[i].usSclkLow) |
+					(psl->entries[i].ucSclkHigh << 16);
+				rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].mclk =
+					le16_to_cpu(psl->entries[i].usMclkLow) |
+					(psl->entries[i].ucMclkHigh << 16);
+				rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].voltage =
+					le16_to_cpu(psl->entries[i].usVoltage);
+			}
+			rdev->pm.dpm.dyn_state.phase_shedding_limits_table.count =
+				psl->ucNumEntries;
+		}
+	}
+
+	/* cac data */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE5)) {
+		rdev->pm.dpm.tdp_limit = le32_to_cpu(power_info->pplib5.ulTDPLimit);
+		rdev->pm.dpm.near_tdp_limit = le32_to_cpu(power_info->pplib5.ulNearTDPLimit);
+		rdev->pm.dpm.near_tdp_limit_adjusted = rdev->pm.dpm.near_tdp_limit;
+		rdev->pm.dpm.tdp_od_limit = le16_to_cpu(power_info->pplib5.usTDPODLimit);
+		if (rdev->pm.dpm.tdp_od_limit)
+			rdev->pm.dpm.power_control = true;
+		else
+			rdev->pm.dpm.power_control = false;
+		rdev->pm.dpm.tdp_adjustment = 0;
+		rdev->pm.dpm.sq_ramping_threshold = le32_to_cpu(power_info->pplib5.ulSQRampingThreshold);
+		rdev->pm.dpm.cac_leakage = le32_to_cpu(power_info->pplib5.ulCACLeakage);
+		rdev->pm.dpm.load_line_slope = le16_to_cpu(power_info->pplib5.usLoadLineSlope);
+		if (power_info->pplib5.usCACLeakageTableOffset) {
+			ATOM_PPLIB_CAC_Leakage_Table *cac_table =
+				(ATOM_PPLIB_CAC_Leakage_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib5.usCACLeakageTableOffset));
+			u32 size = cac_table->ucNumEntries * sizeof(struct radeon_cac_leakage_table);
+			rdev->pm.dpm.dyn_state.cac_leakage_table.entries = kzalloc(size, GFP_KERNEL);
+			if (!rdev->pm.dpm.dyn_state.cac_leakage_table.entries) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+				return -ENOMEM;
+			}
+			for (i = 0; i < cac_table->ucNumEntries; i++) {
+				rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc =
+					le16_to_cpu(cac_table->entries[i].usVddc);
+				rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].leakage =
+					le32_to_cpu(cac_table->entries[i].ulLeakageValue);
+			}
+			rdev->pm.dpm.dyn_state.cac_leakage_table.count = cac_table->ucNumEntries;
+		}
+	}
+
+	/* ppm table */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE3)) {
+		ATOM_PPLIB_EXTENDEDHEADER *ext_hdr = (ATOM_PPLIB_EXTENDEDHEADER *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib3.usExtendendedHeaderOffset));
+		if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V5) &&
+		    ext_hdr->usPPMTableOffset) {
+			ATOM_PPLIB_PPM_Table *ppm = (ATOM_PPLIB_PPM_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(ext_hdr->usPPMTableOffset));
+			rdev->pm.dpm.dyn_state.ppm_table =
+				kzalloc(sizeof(struct radeon_ppm_table), GFP_KERNEL);
+			if (!rdev->pm.dpm.dyn_state.ppm_table) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.cac_leakage_table.entries);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.dyn_state.ppm_table->ppm_design = ppm->ucPpmDesign;
+			rdev->pm.dpm.dyn_state.ppm_table->cpu_core_number =
+				le16_to_cpu(ppm->usCpuCoreNumber);
+			rdev->pm.dpm.dyn_state.ppm_table->platform_tdp =
+				le32_to_cpu(ppm->ulPlatformTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->small_ac_platform_tdp =
+				le32_to_cpu(ppm->ulSmallACPlatformTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->platform_tdc =
+				le32_to_cpu(ppm->ulPlatformTDC);
+			rdev->pm.dpm.dyn_state.ppm_table->small_ac_platform_tdc =
+				le32_to_cpu(ppm->ulSmallACPlatformTDC);
+			rdev->pm.dpm.dyn_state.ppm_table->apu_tdp =
+				le32_to_cpu(ppm->ulApuTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->dgpu_tdp =
+				le32_to_cpu(ppm->ulDGpuTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->dgpu_ulv_power =
+				le32_to_cpu(ppm->ulDGpuUlvPower);
+			rdev->pm.dpm.dyn_state.ppm_table->tj_max =
+				le32_to_cpu(ppm->ulTjmax);
+		}
+	}
+
+	return 0;
+}
+
+void r600_free_extended_power_table(struct radeon_device *rdev)
+{
+	if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries)
+		kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+	if (rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries)
+		kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+	if (rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries)
+		kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+	if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries)
+		kfree(rdev->pm.dpm.dyn_state.cac_leakage_table.entries);
+	if (rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries)
+		kfree(rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries);
+	if (rdev->pm.dpm.dyn_state.ppm_table)
+		kfree(rdev->pm.dpm.dyn_state.ppm_table);
+}
+
+enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev,
+					       u32 sys_mask,
+					       enum radeon_pcie_gen asic_gen,
+					       enum radeon_pcie_gen default_gen)
+{
+	switch (asic_gen) {
+	case RADEON_PCIE_GEN1:
+		return RADEON_PCIE_GEN1;
+	case RADEON_PCIE_GEN2:
+		return RADEON_PCIE_GEN2;
+	case RADEON_PCIE_GEN3:
+		return RADEON_PCIE_GEN3;
+	default:
+		if ((sys_mask & DRM_PCIE_SPEED_80) && (default_gen == RADEON_PCIE_GEN3))
+			return RADEON_PCIE_GEN3;
+		else if ((sys_mask & DRM_PCIE_SPEED_50) && (default_gen == RADEON_PCIE_GEN2))
+			return RADEON_PCIE_GEN2;
+		else
+			return RADEON_PCIE_GEN1;
+	}
+	return RADEON_PCIE_GEN1;
+}
diff --git a/drivers/gpu/drm/radeon/r600_dpm.h b/drivers/gpu/drm/radeon/r600_dpm.h
new file mode 100644
index 0000000..7c822d9
--- /dev/null
+++ b/drivers/gpu/drm/radeon/r600_dpm.h
@@ -0,0 +1,227 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __R600_DPM_H__
+#define __R600_DPM_H__
+
+#define R600_ASI_DFLT                                10000
+#define R600_BSP_DFLT                                0x41EB
+#define R600_BSU_DFLT                                0x2
+#define R600_AH_DFLT                                 5
+#define R600_RLP_DFLT                                25
+#define R600_RMP_DFLT                                65
+#define R600_LHP_DFLT                                40
+#define R600_LMP_DFLT                                15
+#define R600_TD_DFLT                                 0
+#define R600_UTC_DFLT_00                             0x24
+#define R600_UTC_DFLT_01                             0x22
+#define R600_UTC_DFLT_02                             0x22
+#define R600_UTC_DFLT_03                             0x22
+#define R600_UTC_DFLT_04                             0x22
+#define R600_UTC_DFLT_05                             0x22
+#define R600_UTC_DFLT_06                             0x22
+#define R600_UTC_DFLT_07                             0x22
+#define R600_UTC_DFLT_08                             0x22
+#define R600_UTC_DFLT_09                             0x22
+#define R600_UTC_DFLT_10                             0x22
+#define R600_UTC_DFLT_11                             0x22
+#define R600_UTC_DFLT_12                             0x22
+#define R600_UTC_DFLT_13                             0x22
+#define R600_UTC_DFLT_14                             0x22
+#define R600_DTC_DFLT_00                             0x24
+#define R600_DTC_DFLT_01                             0x22
+#define R600_DTC_DFLT_02                             0x22
+#define R600_DTC_DFLT_03                             0x22
+#define R600_DTC_DFLT_04                             0x22
+#define R600_DTC_DFLT_05                             0x22
+#define R600_DTC_DFLT_06                             0x22
+#define R600_DTC_DFLT_07                             0x22
+#define R600_DTC_DFLT_08                             0x22
+#define R600_DTC_DFLT_09                             0x22
+#define R600_DTC_DFLT_10                             0x22
+#define R600_DTC_DFLT_11                             0x22
+#define R600_DTC_DFLT_12                             0x22
+#define R600_DTC_DFLT_13                             0x22
+#define R600_DTC_DFLT_14                             0x22
+#define R600_VRC_DFLT                                0x0000C003
+#define R600_VOLTAGERESPONSETIME_DFLT                1000
+#define R600_BACKBIASRESPONSETIME_DFLT               1000
+#define R600_VRU_DFLT                                0x3
+#define R600_SPLLSTEPTIME_DFLT                       0x1000
+#define R600_SPLLSTEPUNIT_DFLT                       0x3
+#define R600_TPU_DFLT                                0
+#define R600_TPC_DFLT                                0x200
+#define R600_SSTU_DFLT                               0
+#define R600_SST_DFLT                                0x00C8
+#define R600_GICST_DFLT                              0x200
+#define R600_FCT_DFLT                                0x0400
+#define R600_FCTU_DFLT                               0
+#define R600_CTXCGTT3DRPHC_DFLT                      0x20
+#define R600_CTXCGTT3DRSDC_DFLT                      0x40
+#define R600_VDDC3DOORPHC_DFLT                       0x100
+#define R600_VDDC3DOORSDC_DFLT                       0x7
+#define R600_VDDC3DOORSU_DFLT                        0
+#define R600_MPLLLOCKTIME_DFLT                       100
+#define R600_MPLLRESETTIME_DFLT                      150
+#define R600_VCOSTEPPCT_DFLT                          20
+#define R600_ENDINGVCOSTEPPCT_DFLT                    5
+#define R600_REFERENCEDIVIDER_DFLT                    4
+
+#define R600_PM_NUMBER_OF_TC 15
+#define R600_PM_NUMBER_OF_SCLKS 20
+#define R600_PM_NUMBER_OF_MCLKS 4
+#define R600_PM_NUMBER_OF_VOLTAGE_LEVELS 4
+#define R600_PM_NUMBER_OF_ACTIVITY_LEVELS 3
+
+/* XXX are these ok? */
+#define R600_TEMP_RANGE_MIN (90 * 1000)
+#define R600_TEMP_RANGE_MAX (120 * 1000)
+
+enum r600_power_level {
+	R600_POWER_LEVEL_LOW = 0,
+	R600_POWER_LEVEL_MEDIUM = 1,
+	R600_POWER_LEVEL_HIGH = 2,
+	R600_POWER_LEVEL_CTXSW = 3,
+};
+
+enum r600_td {
+	R600_TD_AUTO,
+	R600_TD_UP,
+	R600_TD_DOWN,
+};
+
+enum r600_display_watermark {
+	R600_DISPLAY_WATERMARK_LOW = 0,
+	R600_DISPLAY_WATERMARK_HIGH = 1,
+};
+
+enum r600_display_gap
+{
+    R600_PM_DISPLAY_GAP_VBLANK_OR_WM = 0,
+    R600_PM_DISPLAY_GAP_VBLANK       = 1,
+    R600_PM_DISPLAY_GAP_WATERMARK    = 2,
+    R600_PM_DISPLAY_GAP_IGNORE       = 3,
+};
+
+extern const u32 r600_utc[R600_PM_NUMBER_OF_TC];
+extern const u32 r600_dtc[R600_PM_NUMBER_OF_TC];
+
+void r600_dpm_print_class_info(u32 class, u32 class2);
+void r600_dpm_print_cap_info(u32 caps);
+void r600_dpm_print_ps_status(struct radeon_device *rdev,
+			      struct radeon_ps *rps);
+u32 r600_dpm_get_vblank_time(struct radeon_device *rdev);
+bool r600_is_uvd_state(u32 class, u32 class2);
+void r600_calculate_u_and_p(u32 i, u32 r_c, u32 p_b,
+			    u32 *p, u32 *u);
+int r600_calculate_at(u32 t, u32 h, u32 fh, u32 fl, u32 *tl, u32 *th);
+void r600_gfx_clockgating_enable(struct radeon_device *rdev, bool enable);
+void r600_dynamicpm_enable(struct radeon_device *rdev, bool enable);
+void r600_enable_thermal_protection(struct radeon_device *rdev, bool enable);
+void r600_enable_acpi_pm(struct radeon_device *rdev);
+void r600_enable_dynamic_pcie_gen2(struct radeon_device *rdev, bool enable);
+bool r600_dynamicpm_enabled(struct radeon_device *rdev);
+void r600_enable_sclk_control(struct radeon_device *rdev, bool enable);
+void r600_enable_mclk_control(struct radeon_device *rdev, bool enable);
+void r600_enable_spll_bypass(struct radeon_device *rdev, bool enable);
+void r600_wait_for_spll_change(struct radeon_device *rdev);
+void r600_set_bsp(struct radeon_device *rdev, u32 u, u32 p);
+void r600_set_at(struct radeon_device *rdev,
+		 u32 l_to_m, u32 m_to_h,
+		 u32 h_to_m, u32 m_to_l);
+void r600_set_tc(struct radeon_device *rdev, u32 index, u32 u_t, u32 d_t);
+void r600_select_td(struct radeon_device *rdev, enum r600_td td);
+void r600_set_vrc(struct radeon_device *rdev, u32 vrv);
+void r600_set_tpu(struct radeon_device *rdev, u32 u);
+void r600_set_tpc(struct radeon_device *rdev, u32 c);
+void r600_set_sstu(struct radeon_device *rdev, u32 u);
+void r600_set_sst(struct radeon_device *rdev, u32 t);
+void r600_set_git(struct radeon_device *rdev, u32 t);
+void r600_set_fctu(struct radeon_device *rdev, u32 u);
+void r600_set_fct(struct radeon_device *rdev, u32 t);
+void r600_set_ctxcgtt3d_rphc(struct radeon_device *rdev, u32 p);
+void r600_set_ctxcgtt3d_rsdc(struct radeon_device *rdev, u32 s);
+void r600_set_vddc3d_oorsu(struct radeon_device *rdev, u32 u);
+void r600_set_vddc3d_oorphc(struct radeon_device *rdev, u32 p);
+void r600_set_vddc3d_oorsdc(struct radeon_device *rdev, u32 s);
+void r600_set_mpll_lock_time(struct radeon_device *rdev, u32 lock_time);
+void r600_set_mpll_reset_time(struct radeon_device *rdev, u32 reset_time);
+void r600_engine_clock_entry_enable(struct radeon_device *rdev,
+				    u32 index, bool enable);
+void r600_engine_clock_entry_enable_pulse_skipping(struct radeon_device *rdev,
+						   u32 index, bool enable);
+void r600_engine_clock_entry_enable_post_divider(struct radeon_device *rdev,
+						 u32 index, bool enable);
+void r600_engine_clock_entry_set_post_divider(struct radeon_device *rdev,
+					      u32 index, u32 divider);
+void r600_engine_clock_entry_set_reference_divider(struct radeon_device *rdev,
+						   u32 index, u32 divider);
+void r600_engine_clock_entry_set_feedback_divider(struct radeon_device *rdev,
+						  u32 index, u32 divider);
+void r600_engine_clock_entry_set_step_time(struct radeon_device *rdev,
+					   u32 index, u32 step_time);
+void r600_vid_rt_set_ssu(struct radeon_device *rdev, u32 u);
+void r600_vid_rt_set_vru(struct radeon_device *rdev, u32 u);
+void r600_vid_rt_set_vrt(struct radeon_device *rdev, u32 rt);
+void r600_voltage_control_enable_pins(struct radeon_device *rdev,
+				      u64 mask);
+void r600_voltage_control_program_voltages(struct radeon_device *rdev,
+					   enum r600_power_level index, u64 pins);
+void r600_voltage_control_deactivate_static_control(struct radeon_device *rdev,
+						    u64 mask);
+void r600_power_level_enable(struct radeon_device *rdev,
+			     enum r600_power_level index, bool enable);
+void r600_power_level_set_voltage_index(struct radeon_device *rdev,
+					enum r600_power_level index, u32 voltage_index);
+void r600_power_level_set_mem_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 mem_clock_index);
+void r600_power_level_set_eng_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 eng_clock_index);
+void r600_power_level_set_watermark_id(struct radeon_device *rdev,
+				       enum r600_power_level index,
+				       enum r600_display_watermark watermark_id);
+void r600_power_level_set_pcie_gen2(struct radeon_device *rdev,
+				    enum r600_power_level index, bool compatible);
+enum r600_power_level r600_power_level_get_current_index(struct radeon_device *rdev);
+enum r600_power_level r600_power_level_get_target_index(struct radeon_device *rdev);
+void r600_power_level_set_enter_index(struct radeon_device *rdev,
+				      enum r600_power_level index);
+void r600_wait_for_power_level_unequal(struct radeon_device *rdev,
+				       enum r600_power_level index);
+void r600_wait_for_power_level(struct radeon_device *rdev,
+			       enum r600_power_level index);
+void r600_start_dpm(struct radeon_device *rdev);
+void r600_stop_dpm(struct radeon_device *rdev);
+
+int r600_set_thermal_temperature_range(struct radeon_device *rdev,
+				       int min_temp, int max_temp);
+bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor);
+
+int r600_parse_extended_power_table(struct radeon_device *rdev);
+void r600_free_extended_power_table(struct radeon_device *rdev);
+
+enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev,
+					       u32 sys_mask,
+					       enum radeon_pcie_gen asic_gen,
+					       enum radeon_pcie_gen default_gen);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 456750a..e73b2a7 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -133,14 +133,7 @@ static void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
 	uint8_t *frame = buffer + 3;
-
-	/* Our header values (type, version, length) should be alright, Intel
-	 * is using the same. Checksum function also seems to be OK, it works
-	 * fine for audio infoframe. However calculated value is always lower
-	 * by 2 in comparison to fglrx. It breaks displaying anything in case
-	 * of TVs that strictly check the checksum. Hack it manually here to
-	 * workaround this issue. */
-	frame[0x0] += 2;
+	uint8_t *header = buffer;
 
 	WREG32(HDMI0_AVI_INFO0 + offset,
 		frame[0x0] | (frame[0x1] << 8) | (frame[0x2] << 16) | (frame[0x3] << 24));
@@ -149,7 +142,7 @@ static void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	WREG32(HDMI0_AVI_INFO2 + offset,
 		frame[0x8] | (frame[0x9] << 8) | (frame[0xA] << 16) | (frame[0xB] << 24));
 	WREG32(HDMI0_AVI_INFO3 + offset,
-		frame[0xC] | (frame[0xD] << 8));
+		frame[0xC] | (frame[0xD] << 8) | (header[1] << 24));
 }
 
 /*
diff --git a/drivers/gpu/drm/radeon/r600_reg.h b/drivers/gpu/drm/radeon/r600_reg.h
index 909219b..3ef2026 100644
--- a/drivers/gpu/drm/radeon/r600_reg.h
+++ b/drivers/gpu/drm/radeon/r600_reg.h
@@ -31,6 +31,12 @@
 #define R600_PCIE_PORT_INDEX                0x0038
 #define R600_PCIE_PORT_DATA                 0x003c
 
+#define R600_RCU_INDEX                      0x0100
+#define R600_RCU_DATA                       0x0104
+
+#define R600_UVD_CTX_INDEX                  0xf4a0
+#define R600_UVD_CTX_DATA                   0xf4a4
+
 #define R600_MC_VM_FB_LOCATION			0x2180
 #define		R600_MC_FB_BASE_MASK			0x0000FFFF
 #define		R600_MC_FB_BASE_SHIFT			0
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 79df558..f1b3084 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -302,10 +302,25 @@
 #define	GRBM_SOFT_RESET					0x8020
 #define		SOFT_RESET_CP					(1<<0)
 
+#define	CG_THERMAL_CTRL					0x7F0
+#define		DIG_THERM_DPM(x)			((x) << 12)
+#define		DIG_THERM_DPM_MASK			0x000FF000
+#define		DIG_THERM_DPM_SHIFT			12
 #define	CG_THERMAL_STATUS				0x7F4
 #define		ASIC_T(x)			        ((x) << 0)
 #define		ASIC_T_MASK			        0x1FF
 #define		ASIC_T_SHIFT			        0
+#define	CG_THERMAL_INT					0x7F8
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+
+#define	RV770_CG_THERMAL_INT				0x734
 
 #define	HDP_HOST_PATH_CNTL				0x2C00
 #define	HDP_NONSURFACE_BASE				0x2C04
@@ -684,10 +699,6 @@
 #define RLC_UCODE_ADDR                                    0x3f2c
 #define RLC_UCODE_DATA                                    0x3f30
 
-/* new for TN */
-#define TN_RLC_SAVE_AND_RESTORE_BASE                      0x3f10
-#define TN_RLC_CLEAR_STATE_RESTORE_BASE                   0x3f20
-
 #define SRBM_SOFT_RESET                                   0xe60
 #       define SOFT_RESET_DMA                             (1 << 12)
 #       define SOFT_RESET_RLC                             (1 << 13)
@@ -1148,6 +1159,219 @@
 #       define AFMT_AZ_FORMAT_WTRIG_ACK      (1 << 29)
 #       define AFMT_AZ_AUDIO_ENABLE_CHG_ACK  (1 << 30)
 
+/* Power management */
+#define CG_SPLL_FUNC_CNTL                                 0x600
+#       define SPLL_RESET                                (1 << 0)
+#       define SPLL_SLEEP                                (1 << 1)
+#       define SPLL_REF_DIV(x)                           ((x) << 2)
+#       define SPLL_REF_DIV_MASK                         (7 << 2)
+#       define SPLL_FB_DIV(x)                            ((x) << 5)
+#       define SPLL_FB_DIV_MASK                          (0xff << 5)
+#       define SPLL_PULSEEN                              (1 << 13)
+#       define SPLL_PULSENUM(x)                          ((x) << 14)
+#       define SPLL_PULSENUM_MASK                        (3 << 14)
+#       define SPLL_SW_HILEN(x)                          ((x) << 16)
+#       define SPLL_SW_HILEN_MASK                        (0xf << 16)
+#       define SPLL_SW_LOLEN(x)                          ((x) << 20)
+#       define SPLL_SW_LOLEN_MASK                        (0xf << 20)
+#       define SPLL_DIVEN                                (1 << 24)
+#       define SPLL_BYPASS_EN                            (1 << 25)
+#       define SPLL_CHG_STATUS                           (1 << 29)
+#       define SPLL_CTLREQ                               (1 << 30)
+#       define SPLL_CTLACK                               (1 << 31)
+
+#define GENERAL_PWRMGT                                    0x618
+#       define GLOBAL_PWRMGT_EN                           (1 << 0)
+#       define STATIC_PM_EN                               (1 << 1)
+#       define MOBILE_SU                                  (1 << 2)
+#       define THERMAL_PROTECTION_DIS                     (1 << 3)
+#       define THERMAL_PROTECTION_TYPE                    (1 << 4)
+#       define ENABLE_GEN2PCIE                            (1 << 5)
+#       define SW_GPIO_INDEX(x)                           ((x) << 6)
+#       define SW_GPIO_INDEX_MASK                         (3 << 6)
+#       define LOW_VOLT_D2_ACPI                           (1 << 8)
+#       define LOW_VOLT_D3_ACPI                           (1 << 9)
+#       define VOLT_PWRMGT_EN                             (1 << 10)
+#define CG_TPC                                            0x61c
+#       define TPCC(x)                                    ((x) << 0)
+#       define TPCC_MASK                                  (0x7fffff << 0)
+#       define TPU(x)                                     ((x) << 23)
+#       define TPU_MASK                                   (0x1f << 23)
+#define SCLK_PWRMGT_CNTL                                  0x620
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_TURNOFF                               (1 << 1)
+#       define SPLL_TURNOFF                               (1 << 2)
+#       define SU_SCLK_USE_BCLK                           (1 << 3)
+#       define DYNAMIC_GFX_ISLAND_PWR_DOWN                (1 << 4)
+#       define DYNAMIC_GFX_ISLAND_PWR_LP                  (1 << 5)
+#       define CLK_TURN_ON_STAGGER                        (1 << 6)
+#       define CLK_TURN_OFF_STAGGER                       (1 << 7)
+#       define FIR_FORCE_TREND_SEL                        (1 << 8)
+#       define FIR_TREND_MODE                             (1 << 9)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 10)
+#       define VDDC3D_TURNOFF_D1                          (1 << 11)
+#       define VDDC3D_TURNOFF_D2                          (1 << 12)
+#       define VDDC3D_TURNOFF_D3                          (1 << 13)
+#       define SPLL_TURNOFF_D2                            (1 << 14)
+#       define SCLK_LOW_D1                                (1 << 15)
+#       define DYN_GFX_CLK_OFF_MC_EN                      (1 << 16)
+#define MCLK_PWRMGT_CNTL                                  0x624
+#       define MPLL_PWRMGT_OFF                            (1 << 0)
+#       define YCLK_TURNOFF                               (1 << 1)
+#       define MPLL_TURNOFF                               (1 << 2)
+#       define SU_MCLK_USE_BCLK                           (1 << 3)
+#       define DLL_READY                                  (1 << 4)
+#       define MC_BUSY                                    (1 << 5)
+#       define MC_INT_CNTL                                (1 << 7)
+#       define MRDCKA_SLEEP                               (1 << 8)
+#       define MRDCKB_SLEEP                               (1 << 9)
+#       define MRDCKC_SLEEP                               (1 << 10)
+#       define MRDCKD_SLEEP                               (1 << 11)
+#       define MRDCKE_SLEEP                               (1 << 12)
+#       define MRDCKF_SLEEP                               (1 << 13)
+#       define MRDCKG_SLEEP                               (1 << 14)
+#       define MRDCKH_SLEEP                               (1 << 15)
+#       define MRDCKA_RESET                               (1 << 16)
+#       define MRDCKB_RESET                               (1 << 17)
+#       define MRDCKC_RESET                               (1 << 18)
+#       define MRDCKD_RESET                               (1 << 19)
+#       define MRDCKE_RESET                               (1 << 20)
+#       define MRDCKF_RESET                               (1 << 21)
+#       define MRDCKG_RESET                               (1 << 22)
+#       define MRDCKH_RESET                               (1 << 23)
+#       define DLL_READY_READ                             (1 << 24)
+#       define USE_DISPLAY_GAP                            (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                  (1 << 26)
+#       define USE_DISPLAY_GAP_CTXSW                      (1 << 27)
+#       define MPLL_TURNOFF_D2                            (1 << 28)
+#       define USE_DISPLAY_URGENT_CTXSW                   (1 << 29)
+
+#define MPLL_TIME                                         0x634
+#       define MPLL_LOCK_TIME(x)                          ((x) << 0)
+#       define MPLL_LOCK_TIME_MASK                        (0xffff << 0)
+#       define MPLL_RESET_TIME(x)                         ((x) << 16)
+#       define MPLL_RESET_TIME_MASK                       (0xffff << 16)
+
+#define SCLK_FREQ_SETTING_STEP_0_PART1                    0x648
+#       define STEP_0_SPLL_POST_DIV(x)                    ((x) << 0)
+#       define STEP_0_SPLL_POST_DIV_MASK                  (0xff << 0)
+#       define STEP_0_SPLL_FB_DIV(x)                      ((x) << 8)
+#       define STEP_0_SPLL_FB_DIV_MASK                    (0xff << 8)
+#       define STEP_0_SPLL_REF_DIV(x)                     ((x) << 16)
+#       define STEP_0_SPLL_REF_DIV_MASK                   (7 << 16)
+#       define STEP_0_SPLL_STEP_TIME(x)                   ((x) << 19)
+#       define STEP_0_SPLL_STEP_TIME_MASK                 (0x1fff << 19)
+#define SCLK_FREQ_SETTING_STEP_0_PART2                    0x64c
+#       define STEP_0_PULSE_HIGH_CNT(x)                   ((x) << 0)
+#       define STEP_0_PULSE_HIGH_CNT_MASK                 (0x1ff << 0)
+#       define STEP_0_POST_DIV_EN                         (1 << 9)
+#       define STEP_0_SPLL_STEP_ENABLE                    (1 << 30)
+#       define STEP_0_SPLL_ENTRY_VALID                    (1 << 31)
+
+#define VID_RT                                            0x6f8
+#       define VID_CRT(x)                                 ((x) << 0)
+#       define VID_CRT_MASK                               (0x1fff << 0)
+#       define VID_CRTU(x)                                ((x) << 13)
+#       define VID_CRTU_MASK                              (7 << 13)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (7 << 16)
+#define CTXSW_PROFILE_INDEX                               0x6fc
+#       define CTXSW_FREQ_VIDS_CFG_INDEX(x)               ((x) << 0)
+#       define CTXSW_FREQ_VIDS_CFG_INDEX_MASK             (3 << 0)
+#       define CTXSW_FREQ_VIDS_CFG_INDEX_SHIFT            0
+#       define CTXSW_FREQ_MCLK_CFG_INDEX(x)               ((x) << 2)
+#       define CTXSW_FREQ_MCLK_CFG_INDEX_MASK             (3 << 2)
+#       define CTXSW_FREQ_MCLK_CFG_INDEX_SHIFT            2
+#       define CTXSW_FREQ_SCLK_CFG_INDEX(x)               ((x) << 4)
+#       define CTXSW_FREQ_SCLK_CFG_INDEX_MASK             (0x1f << 4)
+#       define CTXSW_FREQ_SCLK_CFG_INDEX_SHIFT            4
+#       define CTXSW_FREQ_STATE_SPLL_RESET_EN             (1 << 9)
+#       define CTXSW_FREQ_STATE_ENABLE                    (1 << 10)
+#       define CTXSW_FREQ_DISPLAY_WATERMARK               (1 << 11)
+#       define CTXSW_FREQ_GEN2PCIE_VOLT                   (1 << 12)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x70c
+#       define TARGET_PROFILE_INDEX_MASK                  (3 << 0)
+#       define TARGET_PROFILE_INDEX_SHIFT                 0
+#       define CURRENT_PROFILE_INDEX_MASK                 (3 << 2)
+#       define CURRENT_PROFILE_INDEX_SHIFT                2
+#       define DYN_PWR_ENTER_INDEX(x)                     ((x) << 4)
+#       define DYN_PWR_ENTER_INDEX_MASK                   (3 << 4)
+#       define DYN_PWR_ENTER_INDEX_SHIFT                  4
+#       define CURR_MCLK_INDEX_MASK                       (3 << 6)
+#       define CURR_MCLK_INDEX_SHIFT                      6
+#       define CURR_SCLK_INDEX_MASK                       (0x1f << 8)
+#       define CURR_SCLK_INDEX_SHIFT                      8
+#       define CURR_VID_INDEX_MASK                        (3 << 13)
+#       define CURR_VID_INDEX_SHIFT                       13
+
+#define LOWER_GPIO_ENABLE                                 0x710
+#define UPPER_GPIO_ENABLE                                 0x714
+#define CTXSW_VID_LOWER_GPIO_CNTL                         0x718
+
+#define VID_UPPER_GPIO_CNTL                               0x740
+#define CG_CTX_CGTT3D_R                                   0x744
+#       define PHC(x)                                     ((x) << 0)
+#       define PHC_MASK                                   (0x1ff << 0)
+#       define SDC(x)                                     ((x) << 9)
+#       define SDC_MASK                                   (0x3fff << 9)
+#define CG_VDDC3D_OOR                                     0x748
+#       define SU(x)                                      ((x) << 23)
+#       define SU_MASK                                    (0xf << 23)
+#define CG_FTV                                            0x74c
+#define CG_FFCT_0                                         0x750
+#       define UTC_0(x)                                   ((x) << 0)
+#       define UTC_0_MASK                                 (0x3ff << 0)
+#       define DTC_0(x)                                   ((x) << 10)
+#       define DTC_0_MASK                                 (0x3ff << 10)
+
+#define CG_BSP                                            0x78c
+#       define BSP(x)                                     ((x) << 0)
+#       define BSP_MASK                                   (0xffff << 0)
+#       define BSU(x)                                     ((x) << 16)
+#       define BSU_MASK                                   (0xf << 16)
+#define CG_RT                                             0x790
+#       define FLS(x)                                     ((x) << 0)
+#       define FLS_MASK                                   (0xffff << 0)
+#       define FMS(x)                                     ((x) << 16)
+#       define FMS_MASK                                   (0xffff << 16)
+#define CG_LT                                             0x794
+#       define FHS(x)                                     ((x) << 0)
+#       define FHS_MASK                                   (0xffff << 0)
+#define CG_GIT                                            0x798
+#       define CG_GICST(x)                                ((x) << 0)
+#       define CG_GICST_MASK                              (0xffff << 0)
+#       define CG_GIPOT(x)                                ((x) << 16)
+#       define CG_GIPOT_MASK                              (0xffff << 16)
+
+#define CG_SSP                                            0x7a8
+#       define CG_SST(x)                                  ((x) << 0)
+#       define CG_SST_MASK                                (0xffff << 0)
+#       define CG_SSTU(x)                                 ((x) << 16)
+#       define CG_SSTU_MASK                               (0xf << 16)
+
+#define CG_RLC_REQ_AND_RSP                                0x7c4
+#       define RLC_CG_REQ_TYPE_MASK                       0xf
+#       define RLC_CG_REQ_TYPE_SHIFT                      0
+#       define CG_RLC_RSP_TYPE_MASK                       0xf0
+#       define CG_RLC_RSP_TYPE_SHIFT                      4
+
+#define CG_FC_T                                           0x7cc
+#       define FC_T(x)                                    ((x) << 0)
+#       define FC_T_MASK                                  (0xffff << 0)
+#       define FC_TU(x)                                   ((x) << 16)
+#       define FC_TU_MASK                                 (0x1f << 16)
+
+#define GPIOPAD_MASK                                      0x1798
+#define GPIOPAD_A                                         0x179c
+#define GPIOPAD_EN                                        0x17a0
+
+#define GRBM_PWR_CNTL                                     0x800c
+#       define REQ_TYPE_MASK                              0xf
+#       define REQ_TYPE_SHIFT                             0
+#       define RSP_TYPE_MASK                              0xf0
+#       define RSP_TYPE_SHIFT                             4
+
 /*
  * UVD
  */
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 142ce6c..9b7025d 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -96,6 +96,7 @@ extern int radeon_pcie_gen2;
 extern int radeon_msi;
 extern int radeon_lockup_timeout;
 extern int radeon_fastfb;
+extern int radeon_dpm;
 
 /*
  * Copy from radeon_drv.h so we don't have to include both and have conflicting
@@ -150,6 +151,13 @@ extern int radeon_fastfb;
 #define RADEON_RESET_MC				(1 << 10)
 #define RADEON_RESET_DISPLAY			(1 << 11)
 
+/* max cursor sizes (in pixels) */
+#define CURSOR_WIDTH 64
+#define CURSOR_HEIGHT 64
+
+#define CIK_CURSOR_WIDTH 128
+#define CIK_CURSOR_HEIGHT 128
+
 /*
  * Errata workarounds.
  */
@@ -192,6 +200,7 @@ struct radeon_clock {
 	uint32_t default_mclk;
 	uint32_t default_sclk;
 	uint32_t default_dispclk;
+	uint32_t current_dispclk;
 	uint32_t dp_extclk;
 	uint32_t max_pixel_clock;
 };
@@ -211,13 +220,51 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 				   u32 clock,
 				   bool strobe_mode,
 				   struct atom_clock_dividers *dividers);
+int radeon_atom_get_memory_pll_dividers(struct radeon_device *rdev,
+					u32 clock,
+					bool strobe_mode,
+					struct atom_mpll_param *mpll_param);
 void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type);
+int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev,
+					  u16 voltage_level, u8 voltage_type,
+					  u32 *gpio_value, u32 *gpio_mask);
+void radeon_atom_set_engine_dram_timings(struct radeon_device *rdev,
+					 u32 eng_clock, u32 mem_clock);
+int radeon_atom_get_voltage_step(struct radeon_device *rdev,
+				 u8 voltage_type, u16 *voltage_step);
+int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage);
+int radeon_atom_get_leakage_vddc_based_on_leakage_idx(struct radeon_device *rdev,
+						      u16 *voltage,
+						      u16 leakage_idx);
+int radeon_atom_round_to_true_voltage(struct radeon_device *rdev,
+				      u8 voltage_type,
+				      u16 nominal_voltage,
+				      u16 *true_voltage);
+int radeon_atom_get_min_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *min_voltage);
+int radeon_atom_get_max_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *max_voltage);
+int radeon_atom_get_voltage_table(struct radeon_device *rdev,
+				  u8 voltage_type, u8 voltage_mode,
+				  struct atom_voltage_table *voltage_table);
+bool radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
+				 u8 voltage_type, u8 voltage_mode);
+void radeon_atom_update_memory_dll(struct radeon_device *rdev,
+				   u32 mem_clock);
+void radeon_atom_set_ac_timing(struct radeon_device *rdev,
+			       u32 mem_clock);
+int radeon_atom_init_mc_reg_table(struct radeon_device *rdev,
+				  u8 module_index,
+				  struct atom_mc_reg_table *reg_table);
+int radeon_atom_get_memory_info(struct radeon_device *rdev,
+				u8 module_index, struct atom_memory_info *mem_info);
+int radeon_atom_get_mclk_range_table(struct radeon_device *rdev,
+				     bool gddr5, u8 module_index,
+				     struct atom_memory_clock_range_table *mclk_range_table);
+int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage);
 void rs690_pm_info(struct radeon_device *rdev);
-extern int rv6xx_get_temp(struct radeon_device *rdev);
-extern int rv770_get_temp(struct radeon_device *rdev);
-extern int evergreen_get_temp(struct radeon_device *rdev);
-extern int sumo_get_temp(struct radeon_device *rdev);
-extern int si_get_temp(struct radeon_device *rdev);
 extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
 				    unsigned *bankh, unsigned *mtaspect,
 				    unsigned *tile_split);
@@ -549,6 +596,20 @@ struct radeon_scratch {
 int radeon_scratch_get(struct radeon_device *rdev, uint32_t *reg);
 void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg);
 
+/*
+ * GPU doorbell structures, functions & helpers
+ */
+struct radeon_doorbell {
+	u32			num_pages;
+	bool			free[1024];
+	/* doorbell mmio */
+	resource_size_t			base;
+	resource_size_t			size;
+	void __iomem			*ptr;
+};
+
+int radeon_doorbell_get(struct radeon_device *rdev, u32 *page);
+void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell);
 
 /*
  * IRQS.
@@ -600,10 +661,21 @@ struct evergreen_irq_stat_regs {
 	u32 afmt_status6;
 };
 
+struct cik_irq_stat_regs {
+	u32 disp_int;
+	u32 disp_int_cont;
+	u32 disp_int_cont2;
+	u32 disp_int_cont3;
+	u32 disp_int_cont4;
+	u32 disp_int_cont5;
+	u32 disp_int_cont6;
+};
+
 union radeon_irq_stat_regs {
 	struct r500_irq_stat_regs r500;
 	struct r600_irq_stat_regs r600;
 	struct evergreen_irq_stat_regs evergreen;
+	struct cik_irq_stat_regs cik;
 };
 
 #define RADEON_MAX_HPD_PINS 6
@@ -620,6 +692,7 @@ struct radeon_irq {
 	bool				hpd[RADEON_MAX_HPD_PINS];
 	bool				afmt[RADEON_MAX_AFMT_BLOCKS];
 	union radeon_irq_stat_regs	stat_regs;
+	bool				dpm_thermal;
 };
 
 int radeon_irq_kms_init(struct radeon_device *rdev);
@@ -677,6 +750,22 @@ struct radeon_ring {
 	u32			idx;
 	u64			last_semaphore_signal_addr;
 	u64			last_semaphore_wait_addr;
+	/* for CIK queues */
+	u32 me;
+	u32 pipe;
+	u32 queue;
+	struct radeon_bo	*mqd_obj;
+	u32 doorbell_page_num;
+	u32 doorbell_offset;
+	unsigned		wptr_offs;
+};
+
+struct radeon_mec {
+	struct radeon_bo	*hpd_eop_obj;
+	u64			hpd_eop_gpu_addr;
+	u32 num_pipe;
+	u32 num_mec;
+	u32 num_queue;
 };
 
 /*
@@ -778,15 +867,22 @@ struct r600_blit {
 };
 
 /*
- * SI RLC stuff
+ * RLC stuff
  */
-struct si_rlc {
+#include "clearstate_defs.h"
+
+struct radeon_rlc {
 	/* for power gating */
 	struct radeon_bo	*save_restore_obj;
 	uint64_t		save_restore_gpu_addr;
+	volatile uint32_t	*sr_ptr;
+	u32                     *reg_list;
+	u32                     reg_list_size;
 	/* for clear state */
 	struct radeon_bo	*clear_state_obj;
 	uint64_t		clear_state_gpu_addr;
+	volatile uint32_t	*cs_ptr;
+	struct cs_section_def   *cs_data;
 };
 
 int radeon_ib_get(struct radeon_device *rdev, int ring,
@@ -883,6 +979,7 @@ struct radeon_cs_parser {
 	u32			cs_flags;
 	u32			ring;
 	s32			priority;
+	struct ww_acquire_ctx	ticket;
 };
 
 extern int radeon_cs_finish_pages(struct radeon_cs_parser *p);
@@ -934,6 +1031,8 @@ struct radeon_wb {
 #define CAYMAN_WB_DMA1_RPTR_OFFSET   2304
 #define R600_WB_UVD_RPTR_OFFSET  2560
 #define R600_WB_EVENT_OFFSET     3072
+#define CIK_WB_CP1_WPTR_OFFSET     3328
+#define CIK_WB_CP2_WPTR_OFFSET     3584
 
 /**
  * struct radeon_pm - power management datas
@@ -958,6 +1057,7 @@ struct radeon_wb {
 enum radeon_pm_method {
 	PM_METHOD_PROFILE,
 	PM_METHOD_DYNPM,
+	PM_METHOD_DPM,
 };
 
 enum radeon_dynpm_state {
@@ -983,11 +1083,24 @@ enum radeon_voltage_type {
 };
 
 enum radeon_pm_state_type {
+	/* not used for dpm */
 	POWER_STATE_TYPE_DEFAULT,
 	POWER_STATE_TYPE_POWERSAVE,
+	/* user selectable states */
 	POWER_STATE_TYPE_BATTERY,
 	POWER_STATE_TYPE_BALANCED,
 	POWER_STATE_TYPE_PERFORMANCE,
+	/* internal states */
+	POWER_STATE_TYPE_INTERNAL_UVD,
+	POWER_STATE_TYPE_INTERNAL_UVD_SD,
+	POWER_STATE_TYPE_INTERNAL_UVD_HD,
+	POWER_STATE_TYPE_INTERNAL_UVD_HD2,
+	POWER_STATE_TYPE_INTERNAL_UVD_MVC,
+	POWER_STATE_TYPE_INTERNAL_BOOT,
+	POWER_STATE_TYPE_INTERNAL_THERMAL,
+	POWER_STATE_TYPE_INTERNAL_ACPI,
+	POWER_STATE_TYPE_INTERNAL_ULV,
+	POWER_STATE_TYPE_INTERNAL_3DPERF,
 };
 
 enum radeon_pm_profile_type {
@@ -1016,12 +1129,17 @@ struct radeon_pm_profile {
 
 enum radeon_int_thermal_type {
 	THERMAL_TYPE_NONE,
+	THERMAL_TYPE_EXTERNAL,
+	THERMAL_TYPE_EXTERNAL_GPIO,
 	THERMAL_TYPE_RV6XX,
 	THERMAL_TYPE_RV770,
+	THERMAL_TYPE_ADT7473_WITH_INTERNAL,
 	THERMAL_TYPE_EVERGREEN,
 	THERMAL_TYPE_SUMO,
 	THERMAL_TYPE_NI,
 	THERMAL_TYPE_SI,
+	THERMAL_TYPE_EMC2103_WITH_INTERNAL,
+	THERMAL_TYPE_CI,
 };
 
 struct radeon_voltage {
@@ -1075,6 +1193,201 @@ struct radeon_power_state {
  */
 #define RADEON_MODE_OVERCLOCK_MARGIN 500 /* 5 MHz */
 
+enum radeon_dpm_auto_throttle_src {
+	RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL,
+	RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL
+};
+
+enum radeon_dpm_event_src {
+	RADEON_DPM_EVENT_SRC_ANALOG = 0,
+	RADEON_DPM_EVENT_SRC_EXTERNAL = 1,
+	RADEON_DPM_EVENT_SRC_DIGITAL = 2,
+	RADEON_DPM_EVENT_SRC_ANALOG_OR_EXTERNAL = 3,
+	RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL = 4
+};
+
+struct radeon_ps {
+	u32 caps; /* vbios flags */
+	u32 class; /* vbios flags */
+	u32 class2; /* vbios flags */
+	/* UVD clocks */
+	u32 vclk;
+	u32 dclk;
+	/* asic priv */
+	void *ps_priv;
+};
+
+struct radeon_dpm_thermal {
+	/* thermal interrupt work */
+	struct work_struct work;
+	/* low temperature threshold */
+	int                min_temp;
+	/* high temperature threshold */
+	int                max_temp;
+	/* was interrupt low to high or high to low */
+	bool               high_to_low;
+};
+
+enum radeon_clk_action
+{
+	RADEON_SCLK_UP = 1,
+	RADEON_SCLK_DOWN
+};
+
+struct radeon_blacklist_clocks
+{
+	u32 sclk;
+	u32 mclk;
+	enum radeon_clk_action action;
+};
+
+struct radeon_clock_and_voltage_limits {
+	u32 sclk;
+	u32 mclk;
+	u32 vddc;
+	u32 vddci;
+};
+
+struct radeon_clock_array {
+	u32 count;
+	u32 *values;
+};
+
+struct radeon_clock_voltage_dependency_entry {
+	u32 clk;
+	u16 v;
+};
+
+struct radeon_clock_voltage_dependency_table {
+	u32 count;
+	struct radeon_clock_voltage_dependency_entry *entries;
+};
+
+struct radeon_cac_leakage_entry {
+	u16 vddc;
+	u32 leakage;
+};
+
+struct radeon_cac_leakage_table {
+	u32 count;
+	struct radeon_cac_leakage_entry *entries;
+};
+
+struct radeon_phase_shedding_limits_entry {
+	u16 voltage;
+	u32 sclk;
+	u32 mclk;
+};
+
+struct radeon_phase_shedding_limits_table {
+	u32 count;
+	struct radeon_phase_shedding_limits_entry *entries;
+};
+
+struct radeon_ppm_table {
+	u8 ppm_design;
+	u16 cpu_core_number;
+	u32 platform_tdp;
+	u32 small_ac_platform_tdp;
+	u32 platform_tdc;
+	u32 small_ac_platform_tdc;
+	u32 apu_tdp;
+	u32 dgpu_tdp;
+	u32 dgpu_ulv_power;
+	u32 tj_max;
+};
+
+struct radeon_dpm_dynamic_state {
+	struct radeon_clock_voltage_dependency_table vddc_dependency_on_sclk;
+	struct radeon_clock_voltage_dependency_table vddci_dependency_on_mclk;
+	struct radeon_clock_voltage_dependency_table vddc_dependency_on_mclk;
+	struct radeon_clock_voltage_dependency_table vddc_dependency_on_dispclk;
+	struct radeon_clock_array valid_sclk_values;
+	struct radeon_clock_array valid_mclk_values;
+	struct radeon_clock_and_voltage_limits max_clock_voltage_on_dc;
+	struct radeon_clock_and_voltage_limits max_clock_voltage_on_ac;
+	u32 mclk_sclk_ratio;
+	u32 sclk_mclk_delta;
+	u16 vddc_vddci_delta;
+	u16 min_vddc_for_pcie_gen2;
+	struct radeon_cac_leakage_table cac_leakage_table;
+	struct radeon_phase_shedding_limits_table phase_shedding_limits_table;
+	struct radeon_ppm_table *ppm_table;
+};
+
+struct radeon_dpm_fan {
+	u16 t_min;
+	u16 t_med;
+	u16 t_high;
+	u16 pwm_min;
+	u16 pwm_med;
+	u16 pwm_high;
+	u8 t_hyst;
+	u32 cycle_delay;
+	u16 t_max;
+	bool ucode_fan_control;
+};
+
+enum radeon_pcie_gen {
+	RADEON_PCIE_GEN1 = 0,
+	RADEON_PCIE_GEN2 = 1,
+	RADEON_PCIE_GEN3 = 2,
+	RADEON_PCIE_GEN_INVALID = 0xffff
+};
+
+enum radeon_dpm_forced_level {
+	RADEON_DPM_FORCED_LEVEL_AUTO = 0,
+	RADEON_DPM_FORCED_LEVEL_LOW = 1,
+	RADEON_DPM_FORCED_LEVEL_HIGH = 2,
+};
+
+struct radeon_dpm {
+	struct radeon_ps        *ps;
+	/* number of valid power states */
+	int                     num_ps;
+	/* current power state that is active */
+	struct radeon_ps        *current_ps;
+	/* requested power state */
+	struct radeon_ps        *requested_ps;
+	/* boot up power state */
+	struct radeon_ps        *boot_ps;
+	/* default uvd power state */
+	struct radeon_ps        *uvd_ps;
+	enum radeon_pm_state_type state;
+	enum radeon_pm_state_type user_state;
+	u32                     platform_caps;
+	u32                     voltage_response_time;
+	u32                     backbias_response_time;
+	void                    *priv;
+	u32			new_active_crtcs;
+	int			new_active_crtc_count;
+	u32			current_active_crtcs;
+	int			current_active_crtc_count;
+	struct radeon_dpm_dynamic_state dyn_state;
+	struct radeon_dpm_fan fan;
+	u32 tdp_limit;
+	u32 near_tdp_limit;
+	u32 near_tdp_limit_adjusted;
+	u32 sq_ramping_threshold;
+	u32 cac_leakage;
+	u16 tdp_od_limit;
+	u32 tdp_adjustment;
+	u16 load_line_slope;
+	bool power_control;
+	bool ac_power;
+	/* special states active */
+	bool                    thermal_active;
+	bool                    uvd_active;
+	/* thermal handling */
+	struct radeon_dpm_thermal thermal;
+	/* forced levels */
+	enum radeon_dpm_forced_level forced_level;
+};
+
+void radeon_dpm_enable_power_state(struct radeon_device *rdev,
+				    enum radeon_pm_state_type dpm_state);
+
+
 struct radeon_pm {
 	struct mutex		mutex;
 	/* write locked while reprogramming mclk */
@@ -1128,6 +1441,9 @@ struct radeon_pm {
 	/* internal thermal controller on rv6xx+ */
 	enum radeon_int_thermal_type int_thermal_type;
 	struct device	        *int_hwmon_dev;
+	/* dpm */
+	bool                    dpm_enabled;
+	struct radeon_dpm       dpm;
 };
 
 int radeon_pm_get_type_index(struct radeon_device *rdev,
@@ -1266,6 +1582,10 @@ struct radeon_asic {
 		int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp);
 		bool (*is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp);
 		void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+
+		u32 (*get_rptr)(struct radeon_device *rdev, struct radeon_ring *ring);
+		u32 (*get_wptr)(struct radeon_device *rdev, struct radeon_ring *ring);
+		void (*set_wptr)(struct radeon_device *rdev, struct radeon_ring *ring);
 	} ring[RADEON_NUM_RINGS];
 	/* irqs */
 	struct {
@@ -1325,7 +1645,7 @@ struct radeon_asic {
 		bool (*sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 		void (*set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 	} hpd;
-	/* power management */
+	/* static power management */
 	struct {
 		void (*misc)(struct radeon_device *rdev);
 		void (*prepare)(struct radeon_device *rdev);
@@ -1340,7 +1660,26 @@ struct radeon_asic {
 		void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes);
 		void (*set_clock_gating)(struct radeon_device *rdev, int enable);
 		int (*set_uvd_clocks)(struct radeon_device *rdev, u32 vclk, u32 dclk);
+		int (*get_temperature)(struct radeon_device *rdev);
 	} pm;
+	/* dynamic power management */
+	struct {
+		int (*init)(struct radeon_device *rdev);
+		void (*setup_asic)(struct radeon_device *rdev);
+		int (*enable)(struct radeon_device *rdev);
+		void (*disable)(struct radeon_device *rdev);
+		int (*pre_set_power_state)(struct radeon_device *rdev);
+		int (*set_power_state)(struct radeon_device *rdev);
+		void (*post_set_power_state)(struct radeon_device *rdev);
+		void (*display_configuration_changed)(struct radeon_device *rdev);
+		void (*fini)(struct radeon_device *rdev);
+		u32 (*get_sclk)(struct radeon_device *rdev, bool low);
+		u32 (*get_mclk)(struct radeon_device *rdev, bool low);
+		void (*print_power_state)(struct radeon_device *rdev, struct radeon_ps *ps);
+		void (*debugfs_print_current_performance_level)(struct radeon_device *rdev, struct seq_file *m);
+		int (*force_performance_level)(struct radeon_device *rdev, enum radeon_dpm_forced_level level);
+		bool (*vblank_too_short)(struct radeon_device *rdev);
+	} dpm;
 	/* pageflipping */
 	struct {
 		void (*pre_page_flip)(struct radeon_device *rdev, int crtc);
@@ -1505,6 +1844,36 @@ struct si_asic {
 	uint32_t tile_mode_array[32];
 };
 
+struct cik_asic {
+	unsigned max_shader_engines;
+	unsigned max_tile_pipes;
+	unsigned max_cu_per_sh;
+	unsigned max_sh_per_se;
+	unsigned max_backends_per_se;
+	unsigned max_texture_channel_caches;
+	unsigned max_gprs;
+	unsigned max_gs_threads;
+	unsigned max_hw_contexts;
+	unsigned sc_prim_fifo_size_frontend;
+	unsigned sc_prim_fifo_size_backend;
+	unsigned sc_hiz_tile_fifo_size;
+	unsigned sc_earlyz_tile_fifo_size;
+
+	unsigned num_tile_pipes;
+	unsigned num_backends_per_se;
+	unsigned backend_disable_mask_per_asic;
+	unsigned backend_map;
+	unsigned num_texture_channel_caches;
+	unsigned mem_max_burst_length_bytes;
+	unsigned mem_row_size_in_kb;
+	unsigned shader_engine_tile_size;
+	unsigned num_gpus;
+	unsigned multi_gpu_tile_size;
+
+	unsigned tile_config;
+	uint32_t tile_mode_array[32];
+};
+
 union radeon_asic_config {
 	struct r300_asic	r300;
 	struct r100_asic	r100;
@@ -1513,6 +1882,7 @@ union radeon_asic_config {
 	struct evergreen_asic	evergreen;
 	struct cayman_asic	cayman;
 	struct si_asic		si;
+	struct cik_asic		cik;
 };
 
 /*
@@ -1657,6 +2027,7 @@ struct radeon_device {
 	struct radeon_gart		gart;
 	struct radeon_mode_info		mode_info;
 	struct radeon_scratch		scratch;
+	struct radeon_doorbell		doorbell;
 	struct radeon_mman		mman;
 	struct radeon_fence_driver	fence_drv[RADEON_NUM_RINGS];
 	wait_queue_head_t		fence_queue;
@@ -1684,13 +2055,18 @@ struct radeon_device {
 	const struct firmware *mc_fw;	/* NI MC firmware */
 	const struct firmware *ce_fw;	/* SI CE firmware */
 	const struct firmware *uvd_fw;	/* UVD firmware */
+	const struct firmware *mec_fw;	/* CIK MEC firmware */
+	const struct firmware *sdma_fw;	/* CIK SDMA firmware */
+	const struct firmware *smc_fw;	/* SMC firmware */
 	struct r600_blit r600_blit;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
 	struct r600_ih ih; /* r6/700 interrupt ring */
-	struct si_rlc rlc;
+	struct radeon_rlc rlc;
+	struct radeon_mec mec;
 	struct work_struct hotplug_work;
 	struct work_struct audio_work;
+	struct work_struct reset_work;
 	int num_crtc; /* number of crtcs */
 	struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */
 	bool audio_enabled;
@@ -1727,6 +2103,9 @@ void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg);
 void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 
+u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset);
+void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v);
+
 /*
  * Cast helper
  */
@@ -1754,6 +2133,18 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 #define WREG32_PCIE(reg, v) rv370_pcie_wreg(rdev, (reg), (v))
 #define RREG32_PCIE_PORT(reg) rdev->pciep_rreg(rdev, (reg))
 #define WREG32_PCIE_PORT(reg, v) rdev->pciep_wreg(rdev, (reg), (v))
+#define RREG32_SMC(reg) tn_smc_rreg(rdev, (reg))
+#define WREG32_SMC(reg, v) tn_smc_wreg(rdev, (reg), (v))
+#define RREG32_RCU(reg) r600_rcu_rreg(rdev, (reg))
+#define WREG32_RCU(reg, v) r600_rcu_wreg(rdev, (reg), (v))
+#define RREG32_CG(reg) eg_cg_rreg(rdev, (reg))
+#define WREG32_CG(reg, v) eg_cg_wreg(rdev, (reg), (v))
+#define RREG32_PIF_PHY0(reg) eg_pif_phy0_rreg(rdev, (reg))
+#define WREG32_PIF_PHY0(reg, v) eg_pif_phy0_wreg(rdev, (reg), (v))
+#define RREG32_PIF_PHY1(reg) eg_pif_phy1_rreg(rdev, (reg))
+#define WREG32_PIF_PHY1(reg, v) eg_pif_phy1_wreg(rdev, (reg), (v))
+#define RREG32_UVD_CTX(reg) r600_uvd_ctx_rreg(rdev, (reg))
+#define WREG32_UVD_CTX(reg, v) r600_uvd_ctx_wreg(rdev, (reg), (v))
 #define WREG32_P(reg, val, mask)				\
 	do {							\
 		uint32_t tmp_ = RREG32(reg);			\
@@ -1774,6 +2165,9 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 #define RREG32_IO(reg) r100_io_rreg(rdev, (reg))
 #define WREG32_IO(reg, v) r100_io_wreg(rdev, (reg), (v))
 
+#define RDOORBELL32(offset) cik_mm_rdoorbell(rdev, (offset))
+#define WDOORBELL32(offset, v) cik_mm_wdoorbell(rdev, (offset), (v))
+
 /*
  * Indirect registers accessor
  */
@@ -1792,6 +2186,96 @@ static inline void rv370_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uin
 	WREG32(RADEON_PCIE_DATA, (v));
 }
 
+static inline u32 tn_smc_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(TN_SMC_IND_INDEX_0, (reg));
+	r = RREG32(TN_SMC_IND_DATA_0);
+	return r;
+}
+
+static inline void tn_smc_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(TN_SMC_IND_INDEX_0, (reg));
+	WREG32(TN_SMC_IND_DATA_0, (v));
+}
+
+static inline u32 r600_rcu_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(R600_RCU_INDEX, ((reg) & 0x1fff));
+	r = RREG32(R600_RCU_DATA);
+	return r;
+}
+
+static inline void r600_rcu_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(R600_RCU_INDEX, ((reg) & 0x1fff));
+	WREG32(R600_RCU_DATA, (v));
+}
+
+static inline u32 eg_cg_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(EVERGREEN_CG_IND_ADDR, ((reg) & 0xffff));
+	r = RREG32(EVERGREEN_CG_IND_DATA);
+	return r;
+}
+
+static inline void eg_cg_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(EVERGREEN_CG_IND_ADDR, ((reg) & 0xffff));
+	WREG32(EVERGREEN_CG_IND_DATA, (v));
+}
+
+static inline u32 eg_pif_phy0_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(EVERGREEN_PIF_PHY0_INDEX, ((reg) & 0xffff));
+	r = RREG32(EVERGREEN_PIF_PHY0_DATA);
+	return r;
+}
+
+static inline void eg_pif_phy0_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(EVERGREEN_PIF_PHY0_INDEX, ((reg) & 0xffff));
+	WREG32(EVERGREEN_PIF_PHY0_DATA, (v));
+}
+
+static inline u32 eg_pif_phy1_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(EVERGREEN_PIF_PHY1_INDEX, ((reg) & 0xffff));
+	r = RREG32(EVERGREEN_PIF_PHY1_DATA);
+	return r;
+}
+
+static inline void eg_pif_phy1_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(EVERGREEN_PIF_PHY1_INDEX, ((reg) & 0xffff));
+	WREG32(EVERGREEN_PIF_PHY1_DATA, (v));
+}
+
+static inline u32 r600_uvd_ctx_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(R600_UVD_CTX_INDEX, ((reg) & 0x1ff));
+	r = RREG32(R600_UVD_CTX_DATA);
+	return r;
+}
+
+static inline void r600_uvd_ctx_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(R600_UVD_CTX_INDEX, ((reg) & 0x1ff));
+	WREG32(R600_UVD_CTX_DATA, (v));
+}
+
 void r100_pll_errata_after_index(struct radeon_device *rdev);
 
 
@@ -1840,6 +2324,16 @@ void r100_pll_errata_after_index(struct radeon_device *rdev);
 			     (rdev->flags & RADEON_IS_IGP))
 #define ASIC_IS_DCE64(rdev) ((rdev->family == CHIP_OLAND))
 #define ASIC_IS_NODCE(rdev) ((rdev->family == CHIP_HAINAN))
+#define ASIC_IS_DCE8(rdev) ((rdev->family >= CHIP_BONAIRE))
+
+#define ASIC_IS_LOMBOK(rdev) ((rdev->ddev->pdev->device == 0x6849) || \
+			      (rdev->ddev->pdev->device == 0x6850) || \
+			      (rdev->ddev->pdev->device == 0x6858) || \
+			      (rdev->ddev->pdev->device == 0x6859) || \
+			      (rdev->ddev->pdev->device == 0x6840) || \
+			      (rdev->ddev->pdev->device == 0x6841) || \
+			      (rdev->ddev->pdev->device == 0x6842) || \
+			      (rdev->ddev->pdev->device == 0x6843))
 
 /*
  * BIOS helpers.
@@ -1892,6 +2386,9 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)].ib_parse((rdev), (ib))
 #define radeon_ring_is_lockup(rdev, r, cp) (rdev)->asic->ring[(r)].is_lockup((rdev), (cp))
 #define radeon_ring_vm_flush(rdev, r, vm) (rdev)->asic->ring[(r)].vm_flush((rdev), (r), (vm))
+#define radeon_ring_get_rptr(rdev, r) (rdev)->asic->ring[(r)->idx].get_rptr((rdev), (r))
+#define radeon_ring_get_wptr(rdev, r) (rdev)->asic->ring[(r)->idx].get_wptr((rdev), (r))
+#define radeon_ring_set_wptr(rdev, r) (rdev)->asic->ring[(r)->idx].set_wptr((rdev), (r))
 #define radeon_irq_set(rdev) (rdev)->asic->irq.set((rdev))
 #define radeon_irq_process(rdev) (rdev)->asic->irq.process((rdev))
 #define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc))
@@ -1915,6 +2412,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l))
 #define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e))
 #define radeon_set_uvd_clocks(rdev, v, d) (rdev)->asic->pm.set_uvd_clocks((rdev), (v), (d))
+#define radeon_get_temperature(rdev) (rdev)->asic->pm.get_temperature((rdev))
 #define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s)))
 #define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r)))
 #define radeon_bandwidth_update(rdev) (rdev)->asic->display.bandwidth_update((rdev))
@@ -1935,6 +2433,21 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_mc_wait_for_idle(rdev) (rdev)->asic->mc_wait_for_idle((rdev))
 #define radeon_get_xclk(rdev) (rdev)->asic->get_xclk((rdev))
 #define radeon_get_gpu_clock_counter(rdev) (rdev)->asic->get_gpu_clock_counter((rdev))
+#define radeon_dpm_init(rdev) rdev->asic->dpm.init((rdev))
+#define radeon_dpm_setup_asic(rdev) rdev->asic->dpm.setup_asic((rdev))
+#define radeon_dpm_enable(rdev) rdev->asic->dpm.enable((rdev))
+#define radeon_dpm_disable(rdev) rdev->asic->dpm.disable((rdev))
+#define radeon_dpm_pre_set_power_state(rdev) rdev->asic->dpm.pre_set_power_state((rdev))
+#define radeon_dpm_set_power_state(rdev) rdev->asic->dpm.set_power_state((rdev))
+#define radeon_dpm_post_set_power_state(rdev) rdev->asic->dpm.post_set_power_state((rdev))
+#define radeon_dpm_display_configuration_changed(rdev) rdev->asic->dpm.display_configuration_changed((rdev))
+#define radeon_dpm_fini(rdev) rdev->asic->dpm.fini((rdev))
+#define radeon_dpm_get_sclk(rdev, l) rdev->asic->dpm.get_sclk((rdev), (l))
+#define radeon_dpm_get_mclk(rdev, l) rdev->asic->dpm.get_mclk((rdev), (l))
+#define radeon_dpm_print_power_state(rdev, ps) rdev->asic->dpm.print_power_state((rdev), (ps))
+#define radeon_dpm_debugfs_print_current_performance_level(rdev, m) rdev->asic->dpm.debugfs_print_current_performance_level((rdev), (m))
+#define radeon_dpm_force_performance_level(rdev, l) rdev->asic->dpm.force_performance_level((rdev), (l))
+#define radeon_dpm_vblank_too_short(rdev) rdev->asic->dpm.vblank_too_short((rdev))
 
 /* Common functions */
 /* AGP */
@@ -2054,6 +2567,10 @@ extern int ni_mc_load_microcode(struct radeon_device *rdev);
 #if defined(CONFIG_ACPI)
 extern int radeon_acpi_init(struct radeon_device *rdev);
 extern void radeon_acpi_fini(struct radeon_device *rdev);
+extern bool radeon_acpi_is_pcie_performance_request_supported(struct radeon_device *rdev);
+extern int radeon_acpi_pcie_performance_request(struct radeon_device *rdev,
+						u8 perf_req, bool advertise);
+extern int radeon_acpi_pcie_notify_device_ready(struct radeon_device *rdev);
 #else
 static inline int radeon_acpi_init(struct radeon_device *rdev) { return 0; }
 static inline void radeon_acpi_fini(struct radeon_device *rdev) { }
diff --git a/drivers/gpu/drm/radeon/radeon_acpi.c b/drivers/gpu/drm/radeon/radeon_acpi.c
index 196d28d..10f98c7 100644
--- a/drivers/gpu/drm/radeon/radeon_acpi.c
+++ b/drivers/gpu/drm/radeon/radeon_acpi.c
@@ -78,6 +78,22 @@ struct atcs_verify_interface {
 	u32 function_bits;	/* supported functions bit vector */
 } __packed;
 
+#define ATCS_VALID_FLAGS_MASK	0x3
+
+struct atcs_pref_req_input {
+	u16 size;		/* structure size in bytes (includes size field) */
+	u16 client_id;		/* client id (bit 2-0: func num, 7-3: dev num, 15-8: bus num) */
+	u16 valid_flags_mask;	/* valid flags mask */
+	u16 flags;		/* flags */
+	u8 req_type;		/* request type */
+	u8 perf_req;		/* performance request */
+} __packed;
+
+struct atcs_pref_req_output {
+	u16 size;		/* structure size in bytes (includes size field) */
+	u8 ret_val;		/* return value */
+} __packed;
+
 /* Call the ATIF method
  */
 /**
@@ -506,6 +522,135 @@ out:
 }
 
 /**
+ * radeon_acpi_is_pcie_performance_request_supported
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Check if the ATCS pcie_perf_req and pcie_dev_rdy methods
+ * are supported (all asics).
+ * returns true if supported, false if not.
+ */
+bool radeon_acpi_is_pcie_performance_request_supported(struct radeon_device *rdev)
+{
+	struct radeon_atcs *atcs = &rdev->atcs;
+
+	if (atcs->functions.pcie_perf_req && atcs->functions.pcie_dev_rdy)
+		return true;
+
+	return false;
+}
+
+/**
+ * radeon_acpi_pcie_notify_device_ready
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Executes the PCIE_DEVICE_READY_NOTIFICATION method
+ * (all asics).
+ * returns 0 on success, error on failure.
+ */
+int radeon_acpi_pcie_notify_device_ready(struct radeon_device *rdev)
+{
+	acpi_handle handle;
+	union acpi_object *info;
+	struct radeon_atcs *atcs = &rdev->atcs;
+
+	/* Get the device handle */
+	handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev);
+	if (!handle)
+		return -EINVAL;
+
+	if (!atcs->functions.pcie_dev_rdy)
+		return -EINVAL;
+
+	info = radeon_atcs_call(handle, ATCS_FUNCTION_PCIE_DEVICE_READY_NOTIFICATION, NULL);
+	if (!info)
+		return -EIO;
+
+	kfree(info);
+
+	return 0;
+}
+
+/**
+ * radeon_acpi_pcie_performance_request
+ *
+ * @rdev: radeon_device pointer
+ * @perf_req: requested perf level (pcie gen speed)
+ * @advertise: set advertise caps flag if set
+ *
+ * Executes the PCIE_PERFORMANCE_REQUEST method to
+ * change the pcie gen speed (all asics).
+ * returns 0 on success, error on failure.
+ */
+int radeon_acpi_pcie_performance_request(struct radeon_device *rdev,
+					 u8 perf_req, bool advertise)
+{
+	acpi_handle handle;
+	union acpi_object *info;
+	struct radeon_atcs *atcs = &rdev->atcs;
+	struct atcs_pref_req_input atcs_input;
+	struct atcs_pref_req_output atcs_output;
+	struct acpi_buffer params;
+	size_t size;
+	u32 retry = 3;
+
+	/* Get the device handle */
+	handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev);
+	if (!handle)
+		return -EINVAL;
+
+	if (!atcs->functions.pcie_perf_req)
+		return -EINVAL;
+
+	atcs_input.size = sizeof(struct atcs_pref_req_input);
+	/* client id (bit 2-0: func num, 7-3: dev num, 15-8: bus num) */
+	atcs_input.client_id = rdev->pdev->devfn | (rdev->pdev->bus->number << 8);
+	atcs_input.valid_flags_mask = ATCS_VALID_FLAGS_MASK;
+	atcs_input.flags = ATCS_WAIT_FOR_COMPLETION;
+	if (advertise)
+		atcs_input.flags |= ATCS_ADVERTISE_CAPS;
+	atcs_input.req_type = ATCS_PCIE_LINK_SPEED;
+	atcs_input.perf_req = perf_req;
+
+	params.length = sizeof(struct atcs_pref_req_input);
+	params.pointer = &atcs_input;
+
+	while (retry--) {
+		info = radeon_atcs_call(handle, ATCS_FUNCTION_PCIE_PERFORMANCE_REQUEST, &params);
+		if (!info)
+			return -EIO;
+
+		memset(&atcs_output, 0, sizeof(atcs_output));
+
+		size = *(u16 *) info->buffer.pointer;
+		if (size < 3) {
+			DRM_INFO("ATCS buffer is too small: %zu\n", size);
+			kfree(info);
+			return -EINVAL;
+		}
+		size = min(sizeof(atcs_output), size);
+
+		memcpy(&atcs_output, info->buffer.pointer, size);
+
+		kfree(info);
+
+		switch (atcs_output.ret_val) {
+		case ATCS_REQUEST_REFUSED:
+		default:
+			return -EINVAL;
+		case ATCS_REQUEST_COMPLETE:
+			return 0;
+		case ATCS_REQUEST_IN_PROGRESS:
+			udelay(10);
+			break;
+		}
+	}
+
+	return 0;
+}
+
+/**
  * radeon_acpi_event - handle notify events
  *
  * @nb: notifier block
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index a2802b47..0970774 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -126,7 +126,11 @@ static void radeon_register_accessor_init(struct radeon_device *rdev)
 		rdev->mc_rreg = &rs780_mc_rreg;
 		rdev->mc_wreg = &rs780_mc_wreg;
 	}
-	if (rdev->family >= CHIP_R600) {
+
+	if (rdev->family >= CHIP_BONAIRE) {
+		rdev->pciep_rreg = &cik_pciep_rreg;
+		rdev->pciep_wreg = &cik_pciep_wreg;
+	} else if (rdev->family >= CHIP_R600) {
 		rdev->pciep_rreg = &r600_pciep_rreg;
 		rdev->pciep_wreg = &r600_pciep_wreg;
 	}
@@ -192,6 +196,9 @@ static struct radeon_asic r100_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -268,6 +275,9 @@ static struct radeon_asic r200_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -344,6 +354,9 @@ static struct radeon_asic r300_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -420,6 +433,9 @@ static struct radeon_asic r300_asic_pcie = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -496,6 +512,9 @@ static struct radeon_asic r420_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -572,6 +591,9 @@ static struct radeon_asic rs400_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -648,6 +670,9 @@ static struct radeon_asic rs600_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -726,6 +751,9 @@ static struct radeon_asic rs690_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -804,6 +832,9 @@ static struct radeon_asic rv515_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -880,6 +911,9 @@ static struct radeon_asic r520_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -957,6 +991,9 @@ static struct radeon_asic r600_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -966,6 +1003,9 @@ static struct radeon_asic r600_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1012,6 +1052,115 @@ static struct radeon_asic r600_asic = {
 		.get_pcie_lanes = &r600_get_pcie_lanes,
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
+		.get_temperature = &rv6xx_get_temp,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
+};
+
+static struct radeon_asic rv6xx_asic = {
+	.init = &r600_init,
+	.fini = &r600_fini,
+	.suspend = &r600_suspend,
+	.resume = &r600_resume,
+	.vga_set_state = &r600_vga_set_state,
+	.asic_reset = &r600_asic_reset,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.get_xclk = &r600_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
+	.gart = {
+		.tlb_flush = &r600_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &r600_ring_ib_execute,
+			.emit_fence = &r600_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &r600_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[R600_RING_TYPE_DMA_INDEX] = {
+			.ib_execute = &r600_dma_ring_ib_execute,
+			.emit_fence = &r600_dma_fence_ring_emit,
+			.emit_semaphore = &r600_dma_semaphore_ring_emit,
+			.cs_parse = &r600_dma_cs_parse,
+			.ring_test = &r600_dma_ring_test,
+			.ib_test = &r600_dma_ib_test,
+			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		}
+	},
+	.irq = {
+		.set = &r600_irq_set,
+		.process = &r600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rv515_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+		.set_backlight_level = &atombios_set_backlight_level,
+		.get_backlight_level = &atombios_get_backlight_level,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r600_copy_dma,
+		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &r600_copy_dma,
+		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r600_hpd_init,
+		.fini = &r600_hpd_fini,
+		.sense = &r600_hpd_sense,
+		.set_polarity = &r600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = &radeon_atom_get_memory_clock,
+		.set_memory_clock = &radeon_atom_set_memory_clock,
+		.get_pcie_lanes = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
+		.set_clock_gating = NULL,
+		.get_temperature = &rv6xx_get_temp,
+	},
+	.dpm = {
+		.init = &rv6xx_dpm_init,
+		.setup_asic = &rv6xx_setup_asic,
+		.enable = &rv6xx_dpm_enable,
+		.disable = &rv6xx_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &rv6xx_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &rv6xx_dpm_display_configuration_changed,
+		.fini = &rv6xx_dpm_fini,
+		.get_sclk = &rv6xx_dpm_get_sclk,
+		.get_mclk = &rv6xx_dpm_get_mclk,
+		.print_power_state = &rv6xx_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv6xx_dpm_debugfs_print_current_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1045,6 +1194,9 @@ static struct radeon_asic rs780_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1054,6 +1206,9 @@ static struct radeon_asic rs780_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1100,6 +1255,21 @@ static struct radeon_asic rs780_asic = {
 		.get_pcie_lanes = NULL,
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
+		.get_temperature = &rv6xx_get_temp,
+	},
+	.dpm = {
+		.init = &rs780_dpm_init,
+		.setup_asic = &rs780_dpm_setup_asic,
+		.enable = &rs780_dpm_enable,
+		.disable = &rs780_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &rs780_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &rs780_dpm_display_configuration_changed,
+		.fini = &rs780_dpm_fini,
+		.get_sclk = &rs780_dpm_get_sclk,
+		.get_mclk = &rs780_dpm_get_mclk,
+		.print_power_state = &rs780_dpm_print_power_state,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1133,6 +1303,9 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1142,6 +1315,9 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1151,6 +1327,9 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1198,6 +1377,24 @@ static struct radeon_asic rv770_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = &radeon_atom_set_clock_gating,
 		.set_uvd_clocks = &rv770_set_uvd_clocks,
+		.get_temperature = &rv770_get_temp,
+	},
+	.dpm = {
+		.init = &rv770_dpm_init,
+		.setup_asic = &rv770_dpm_setup_asic,
+		.enable = &rv770_dpm_enable,
+		.disable = &rv770_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &rv770_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &rv770_dpm_display_configuration_changed,
+		.fini = &rv770_dpm_fini,
+		.get_sclk = &rv770_dpm_get_sclk,
+		.get_mclk = &rv770_dpm_get_mclk,
+		.print_power_state = &rv770_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &rv770_dpm_force_performance_level,
+		.vblank_too_short = &rv770_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1231,6 +1428,9 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &evergreen_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1240,6 +1440,9 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1249,6 +1452,9 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1296,6 +1502,24 @@ static struct radeon_asic evergreen_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &evergreen_set_uvd_clocks,
+		.get_temperature = &evergreen_get_temp,
+	},
+	.dpm = {
+		.init = &cypress_dpm_init,
+		.setup_asic = &cypress_dpm_setup_asic,
+		.enable = &cypress_dpm_enable,
+		.disable = &cypress_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &cypress_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &cypress_dpm_display_configuration_changed,
+		.fini = &cypress_dpm_fini,
+		.get_sclk = &rv770_dpm_get_sclk,
+		.get_mclk = &rv770_dpm_get_mclk,
+		.print_power_state = &rv770_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &rv770_dpm_force_performance_level,
+		.vblank_too_short = &cypress_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1329,6 +1553,9 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &evergreen_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1338,6 +1565,9 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1347,6 +1577,9 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1394,6 +1627,23 @@ static struct radeon_asic sumo_asic = {
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &sumo_set_uvd_clocks,
+		.get_temperature = &sumo_get_temp,
+	},
+	.dpm = {
+		.init = &sumo_dpm_init,
+		.setup_asic = &sumo_dpm_setup_asic,
+		.enable = &sumo_dpm_enable,
+		.disable = &sumo_dpm_disable,
+		.pre_set_power_state = &sumo_dpm_pre_set_power_state,
+		.set_power_state = &sumo_dpm_set_power_state,
+		.post_set_power_state = &sumo_dpm_post_set_power_state,
+		.display_configuration_changed = &sumo_dpm_display_configuration_changed,
+		.fini = &sumo_dpm_fini,
+		.get_sclk = &sumo_dpm_get_sclk,
+		.get_mclk = &sumo_dpm_get_mclk,
+		.print_power_state = &sumo_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &sumo_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &sumo_dpm_force_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1427,6 +1677,9 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &evergreen_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1436,6 +1689,9 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1445,6 +1701,9 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1492,6 +1751,24 @@ static struct radeon_asic btc_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &evergreen_set_uvd_clocks,
+		.get_temperature = &evergreen_get_temp,
+	},
+	.dpm = {
+		.init = &btc_dpm_init,
+		.setup_asic = &btc_dpm_setup_asic,
+		.enable = &btc_dpm_enable,
+		.disable = &btc_dpm_disable,
+		.pre_set_power_state = &btc_dpm_pre_set_power_state,
+		.set_power_state = &btc_dpm_set_power_state,
+		.post_set_power_state = &btc_dpm_post_set_power_state,
+		.display_configuration_changed = &cypress_dpm_display_configuration_changed,
+		.fini = &btc_dpm_fini,
+		.get_sclk = &btc_dpm_get_sclk,
+		.get_mclk = &btc_dpm_get_mclk,
+		.print_power_state = &rv770_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &rv770_dpm_force_performance_level,
+		.vblank_too_short = &btc_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1533,6 +1810,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1544,6 +1824,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1555,6 +1838,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1566,6 +1852,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1577,6 +1866,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1586,6 +1878,9 @@ static struct radeon_asic cayman_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1633,6 +1928,24 @@ static struct radeon_asic cayman_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &evergreen_set_uvd_clocks,
+		.get_temperature = &evergreen_get_temp,
+	},
+	.dpm = {
+		.init = &ni_dpm_init,
+		.setup_asic = &ni_dpm_setup_asic,
+		.enable = &ni_dpm_enable,
+		.disable = &ni_dpm_disable,
+		.pre_set_power_state = &ni_dpm_pre_set_power_state,
+		.set_power_state = &ni_dpm_set_power_state,
+		.post_set_power_state = &ni_dpm_post_set_power_state,
+		.display_configuration_changed = &cypress_dpm_display_configuration_changed,
+		.fini = &ni_dpm_fini,
+		.get_sclk = &ni_dpm_get_sclk,
+		.get_mclk = &ni_dpm_get_mclk,
+		.print_power_state = &ni_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &ni_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &ni_dpm_force_performance_level,
+		.vblank_too_short = &ni_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1674,6 +1987,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1685,6 +2001,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1696,6 +2015,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1707,6 +2029,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1718,6 +2043,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1727,6 +2055,9 @@ static struct radeon_asic trinity_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1772,6 +2103,23 @@ static struct radeon_asic trinity_asic = {
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &sumo_set_uvd_clocks,
+		.get_temperature = &tn_get_temp,
+	},
+	.dpm = {
+		.init = &trinity_dpm_init,
+		.setup_asic = &trinity_dpm_setup_asic,
+		.enable = &trinity_dpm_enable,
+		.disable = &trinity_dpm_disable,
+		.pre_set_power_state = &trinity_dpm_pre_set_power_state,
+		.set_power_state = &trinity_dpm_set_power_state,
+		.post_set_power_state = &trinity_dpm_post_set_power_state,
+		.display_configuration_changed = &trinity_dpm_display_configuration_changed,
+		.fini = &trinity_dpm_fini,
+		.get_sclk = &trinity_dpm_get_sclk,
+		.get_mclk = &trinity_dpm_get_mclk,
+		.print_power_state = &trinity_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &trinity_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &trinity_dpm_force_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1813,6 +2161,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &si_ring_ib_execute,
@@ -1824,6 +2175,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &si_ring_ib_execute,
@@ -1835,6 +2189,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1846,6 +2203,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1857,6 +2217,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1866,6 +2229,9 @@ static struct radeon_asic si_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1911,6 +2277,334 @@ static struct radeon_asic si_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &si_set_uvd_clocks,
+		.get_temperature = &si_get_temp,
+	},
+	.dpm = {
+		.init = &si_dpm_init,
+		.setup_asic = &si_dpm_setup_asic,
+		.enable = &si_dpm_enable,
+		.disable = &si_dpm_disable,
+		.pre_set_power_state = &si_dpm_pre_set_power_state,
+		.set_power_state = &si_dpm_set_power_state,
+		.post_set_power_state = &si_dpm_post_set_power_state,
+		.display_configuration_changed = &si_dpm_display_configuration_changed,
+		.fini = &si_dpm_fini,
+		.get_sclk = &ni_dpm_get_sclk,
+		.get_mclk = &ni_dpm_get_mclk,
+		.print_power_state = &ni_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &si_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &si_dpm_force_performance_level,
+		.vblank_too_short = &ni_dpm_vblank_too_short,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
+};
+
+static struct radeon_asic ci_asic = {
+	.init = &cik_init,
+	.fini = &cik_fini,
+	.suspend = &cik_suspend,
+	.resume = &cik_resume,
+	.asic_reset = &cik_asic_reset,
+	.vga_set_state = &r600_vga_set_state,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &cik_get_xclk,
+	.get_gpu_clock_counter = &cik_get_gpu_clock_counter,
+	.gart = {
+		.tlb_flush = &cik_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.vm = {
+		.init = &cik_vm_init,
+		.fini = &cik_vm_fini,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.set_page = &cik_vm_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_gfx_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP1_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP2_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[R600_RING_TYPE_DMA_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		}
+	},
+	.irq = {
+		.set = &cik_irq_set,
+		.process = &cik_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &dce8_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = NULL,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &cik_copy_dma,
+		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &cik_copy_dma,
+		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = &radeon_atom_get_memory_clock,
+		.set_memory_clock = &radeon_atom_set_memory_clock,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+		.set_uvd_clocks = &cik_set_uvd_clocks,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
+};
+
+static struct radeon_asic kv_asic = {
+	.init = &cik_init,
+	.fini = &cik_fini,
+	.suspend = &cik_suspend,
+	.resume = &cik_resume,
+	.asic_reset = &cik_asic_reset,
+	.vga_set_state = &r600_vga_set_state,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &cik_get_xclk,
+	.get_gpu_clock_counter = &cik_get_gpu_clock_counter,
+	.gart = {
+		.tlb_flush = &cik_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.vm = {
+		.init = &cik_vm_init,
+		.fini = &cik_vm_fini,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.set_page = &cik_vm_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_gfx_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP1_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP2_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[R600_RING_TYPE_DMA_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		}
+	},
+	.irq = {
+		.set = &cik_irq_set,
+		.process = &cik_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &dce8_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = NULL,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &cik_copy_dma,
+		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &cik_copy_dma,
+		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = &radeon_atom_get_memory_clock,
+		.set_memory_clock = &radeon_atom_set_memory_clock,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+		.set_uvd_clocks = &cik_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1999,16 +2693,15 @@ int radeon_asic_init(struct radeon_device *rdev)
 		rdev->asic = &r520_asic;
 		break;
 	case CHIP_R600:
+		rdev->asic = &r600_asic;
+		break;
 	case CHIP_RV610:
 	case CHIP_RV630:
 	case CHIP_RV620:
 	case CHIP_RV635:
 	case CHIP_RV670:
-		rdev->asic = &r600_asic;
-		if (rdev->family == CHIP_R600)
-			rdev->has_uvd = false;
-		else
-			rdev->has_uvd = true;
+		rdev->asic = &rv6xx_asic;
+		rdev->has_uvd = true;
 		break;
 	case CHIP_RS780:
 	case CHIP_RS880:
@@ -2082,6 +2775,19 @@ int radeon_asic_init(struct radeon_device *rdev)
 		else
 			rdev->has_uvd = true;
 		break;
+	case CHIP_BONAIRE:
+		rdev->asic = &ci_asic;
+		rdev->num_crtc = 6;
+		break;
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+		rdev->asic = &kv_asic;
+		/* set num crtcs */
+		if (rdev->family == CHIP_KAVERI)
+			rdev->num_crtc = 4;
+		else
+			rdev->num_crtc = 2;
+		break;
 	default:
 		/* FIXME: not supported yet */
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index a72759e..45d0693 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -47,6 +47,12 @@ u8 atombios_get_backlight_level(struct radeon_encoder *radeon_encoder);
 void radeon_legacy_set_backlight_level(struct radeon_encoder *radeon_encoder, u8 level);
 u8 radeon_legacy_get_backlight_level(struct radeon_encoder *radeon_encoder);
 
+u32 radeon_ring_generic_get_rptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring);
+u32 radeon_ring_generic_get_wptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring);
+void radeon_ring_generic_set_wptr(struct radeon_device *rdev,
+				  struct radeon_ring *ring);
 
 /*
  * r100,rv100,rs100,rv200,rs200
@@ -395,6 +401,35 @@ void r600_kms_blit_copy(struct radeon_device *rdev,
 int r600_mc_wait_for_idle(struct radeon_device *rdev);
 u32 r600_get_xclk(struct radeon_device *rdev);
 uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev);
+int rv6xx_get_temp(struct radeon_device *rdev);
+int r600_dpm_pre_set_power_state(struct radeon_device *rdev);
+void r600_dpm_post_set_power_state(struct radeon_device *rdev);
+/* rv6xx dpm */
+int rv6xx_dpm_init(struct radeon_device *rdev);
+int rv6xx_dpm_enable(struct radeon_device *rdev);
+void rv6xx_dpm_disable(struct radeon_device *rdev);
+int rv6xx_dpm_set_power_state(struct radeon_device *rdev);
+void rv6xx_setup_asic(struct radeon_device *rdev);
+void rv6xx_dpm_display_configuration_changed(struct radeon_device *rdev);
+void rv6xx_dpm_fini(struct radeon_device *rdev);
+u32 rv6xx_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 rv6xx_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void rv6xx_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *ps);
+void rv6xx_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m);
+/* rs780 dpm */
+int rs780_dpm_init(struct radeon_device *rdev);
+int rs780_dpm_enable(struct radeon_device *rdev);
+void rs780_dpm_disable(struct radeon_device *rdev);
+int rs780_dpm_set_power_state(struct radeon_device *rdev);
+void rs780_dpm_setup_asic(struct radeon_device *rdev);
+void rs780_dpm_display_configuration_changed(struct radeon_device *rdev);
+void rs780_dpm_fini(struct radeon_device *rdev);
+u32 rs780_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void rs780_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *ps);
 
 /* uvd */
 int r600_uvd_init(struct radeon_device *rdev);
@@ -428,6 +463,24 @@ int rv770_copy_dma(struct radeon_device *rdev,
 u32 rv770_get_xclk(struct radeon_device *rdev);
 int rv770_uvd_resume(struct radeon_device *rdev);
 int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int rv770_get_temp(struct radeon_device *rdev);
+/* rv7xx pm */
+int rv770_dpm_init(struct radeon_device *rdev);
+int rv770_dpm_enable(struct radeon_device *rdev);
+void rv770_dpm_disable(struct radeon_device *rdev);
+int rv770_dpm_set_power_state(struct radeon_device *rdev);
+void rv770_dpm_setup_asic(struct radeon_device *rdev);
+void rv770_dpm_display_configuration_changed(struct radeon_device *rdev);
+void rv770_dpm_fini(struct radeon_device *rdev);
+u32 rv770_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 rv770_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void rv770_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *ps);
+void rv770_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m);
+int rv770_dpm_force_performance_level(struct radeon_device *rdev,
+				      enum radeon_dpm_forced_level level);
+bool rv770_dpm_vblank_too_short(struct radeon_device *rdev);
 
 /*
  * evergreen
@@ -482,6 +535,45 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		       struct radeon_fence **fence);
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable);
 void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
+int evergreen_get_temp(struct radeon_device *rdev);
+int sumo_get_temp(struct radeon_device *rdev);
+int tn_get_temp(struct radeon_device *rdev);
+int cypress_dpm_init(struct radeon_device *rdev);
+void cypress_dpm_setup_asic(struct radeon_device *rdev);
+int cypress_dpm_enable(struct radeon_device *rdev);
+void cypress_dpm_disable(struct radeon_device *rdev);
+int cypress_dpm_set_power_state(struct radeon_device *rdev);
+void cypress_dpm_display_configuration_changed(struct radeon_device *rdev);
+void cypress_dpm_fini(struct radeon_device *rdev);
+bool cypress_dpm_vblank_too_short(struct radeon_device *rdev);
+int btc_dpm_init(struct radeon_device *rdev);
+void btc_dpm_setup_asic(struct radeon_device *rdev);
+int btc_dpm_enable(struct radeon_device *rdev);
+void btc_dpm_disable(struct radeon_device *rdev);
+int btc_dpm_pre_set_power_state(struct radeon_device *rdev);
+int btc_dpm_set_power_state(struct radeon_device *rdev);
+void btc_dpm_post_set_power_state(struct radeon_device *rdev);
+void btc_dpm_fini(struct radeon_device *rdev);
+u32 btc_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 btc_dpm_get_mclk(struct radeon_device *rdev, bool low);
+bool btc_dpm_vblank_too_short(struct radeon_device *rdev);
+int sumo_dpm_init(struct radeon_device *rdev);
+int sumo_dpm_enable(struct radeon_device *rdev);
+void sumo_dpm_disable(struct radeon_device *rdev);
+int sumo_dpm_pre_set_power_state(struct radeon_device *rdev);
+int sumo_dpm_set_power_state(struct radeon_device *rdev);
+void sumo_dpm_post_set_power_state(struct radeon_device *rdev);
+void sumo_dpm_setup_asic(struct radeon_device *rdev);
+void sumo_dpm_display_configuration_changed(struct radeon_device *rdev);
+void sumo_dpm_fini(struct radeon_device *rdev);
+u32 sumo_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 sumo_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void sumo_dpm_print_power_state(struct radeon_device *rdev,
+				struct radeon_ps *ps);
+void sumo_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						      struct seq_file *m);
+int sumo_dpm_force_performance_level(struct radeon_device *rdev,
+				     enum radeon_dpm_forced_level level);
 
 /*
  * cayman
@@ -516,6 +608,41 @@ bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 
+int ni_dpm_init(struct radeon_device *rdev);
+void ni_dpm_setup_asic(struct radeon_device *rdev);
+int ni_dpm_enable(struct radeon_device *rdev);
+void ni_dpm_disable(struct radeon_device *rdev);
+int ni_dpm_pre_set_power_state(struct radeon_device *rdev);
+int ni_dpm_set_power_state(struct radeon_device *rdev);
+void ni_dpm_post_set_power_state(struct radeon_device *rdev);
+void ni_dpm_fini(struct radeon_device *rdev);
+u32 ni_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 ni_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void ni_dpm_print_power_state(struct radeon_device *rdev,
+			      struct radeon_ps *ps);
+void ni_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m);
+int ni_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level);
+bool ni_dpm_vblank_too_short(struct radeon_device *rdev);
+int trinity_dpm_init(struct radeon_device *rdev);
+int trinity_dpm_enable(struct radeon_device *rdev);
+void trinity_dpm_disable(struct radeon_device *rdev);
+int trinity_dpm_pre_set_power_state(struct radeon_device *rdev);
+int trinity_dpm_set_power_state(struct radeon_device *rdev);
+void trinity_dpm_post_set_power_state(struct radeon_device *rdev);
+void trinity_dpm_setup_asic(struct radeon_device *rdev);
+void trinity_dpm_display_configuration_changed(struct radeon_device *rdev);
+void trinity_dpm_fini(struct radeon_device *rdev);
+u32 trinity_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 trinity_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void trinity_dpm_print_power_state(struct radeon_device *rdev,
+				   struct radeon_ps *ps);
+void trinity_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+							 struct seq_file *m);
+int trinity_dpm_force_performance_level(struct radeon_device *rdev,
+					enum radeon_dpm_forced_level level);
+
 /* DCE6 - SI */
 void dce6_bandwidth_update(struct radeon_device *rdev);
 
@@ -552,5 +679,82 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 u32 si_get_xclk(struct radeon_device *rdev);
 uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
 int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int si_get_temp(struct radeon_device *rdev);
+int si_dpm_init(struct radeon_device *rdev);
+void si_dpm_setup_asic(struct radeon_device *rdev);
+int si_dpm_enable(struct radeon_device *rdev);
+void si_dpm_disable(struct radeon_device *rdev);
+int si_dpm_pre_set_power_state(struct radeon_device *rdev);
+int si_dpm_set_power_state(struct radeon_device *rdev);
+void si_dpm_post_set_power_state(struct radeon_device *rdev);
+void si_dpm_fini(struct radeon_device *rdev);
+void si_dpm_display_configuration_changed(struct radeon_device *rdev);
+void si_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m);
+int si_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level);
+
+/* DCE8 - CIK */
+void dce8_bandwidth_update(struct radeon_device *rdev);
+
+/*
+ * cik
+ */
+uint64_t cik_get_gpu_clock_counter(struct radeon_device *rdev);
+u32 cik_get_xclk(struct radeon_device *rdev);
+uint32_t cik_pciep_rreg(struct radeon_device *rdev, uint32_t reg);
+void cik_pciep_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
+int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int cik_uvd_resume(struct radeon_device *rdev);
+void cik_sdma_fence_ring_emit(struct radeon_device *rdev,
+			      struct radeon_fence *fence);
+void cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
+				  struct radeon_ring *ring,
+				  struct radeon_semaphore *semaphore,
+				  bool emit_wait);
+void cik_sdma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+int cik_copy_dma(struct radeon_device *rdev,
+		 uint64_t src_offset, uint64_t dst_offset,
+		 unsigned num_gpu_pages,
+		 struct radeon_fence **fence);
+int cik_sdma_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
+void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
+			     struct radeon_fence *fence);
+void cik_fence_compute_ring_emit(struct radeon_device *rdev,
+				 struct radeon_fence *fence);
+void cik_semaphore_ring_emit(struct radeon_device *rdev,
+			     struct radeon_ring *cp,
+			     struct radeon_semaphore *semaphore,
+			     bool emit_wait);
+void cik_pcie_gart_tlb_flush(struct radeon_device *rdev);
+int cik_init(struct radeon_device *rdev);
+void cik_fini(struct radeon_device *rdev);
+int cik_suspend(struct radeon_device *rdev);
+int cik_resume(struct radeon_device *rdev);
+bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+int cik_asic_reset(struct radeon_device *rdev);
+void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int cik_irq_set(struct radeon_device *rdev);
+int cik_irq_process(struct radeon_device *rdev);
+int cik_vm_init(struct radeon_device *rdev);
+void cik_vm_fini(struct radeon_device *rdev);
+void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void cik_vm_set_page(struct radeon_device *rdev,
+		     struct radeon_ib *ib,
+		     uint64_t pe,
+		     uint64_t addr, unsigned count,
+		     uint32_t incr, uint32_t flags);
+void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
+u32 cik_compute_ring_get_rptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring);
+u32 cik_compute_ring_get_wptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring);
+void cik_compute_ring_set_wptr(struct radeon_device *rdev,
+			       struct radeon_ring *ring);
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index dea6f63c..fbdaff5 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -56,10 +56,6 @@ extern void
 radeon_add_legacy_encoder(struct drm_device *dev, uint32_t encoder_enum,
 			  uint32_t supported_device);
 
-/* local */
-static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
-				    u16 voltage_id, u16 *voltage);
-
 union atom_supported_devices {
 	struct _ATOM_SUPPORTED_DEVICES_INFO info;
 	struct _ATOM_SUPPORTED_DEVICES_INFO_2 info_2;
@@ -1247,6 +1243,7 @@ bool radeon_atom_get_clock_info(struct drm_device *dev)
 			}
 			rdev->clock.dp_extclk =
 				le16_to_cpu(firmware_info->info_21.usUniphyDPModeExtClkFreq);
+			rdev->clock.current_dispclk = rdev->clock.default_dispclk;
 		}
 		*dcpll = *p1pll;
 
@@ -1269,6 +1266,7 @@ union igp_info {
 	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
 	struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6;
 	struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7 info_7;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_8 info_8;
 };
 
 bool radeon_atombios_sideport_present(struct radeon_device *rdev)
@@ -1438,6 +1436,22 @@ static void radeon_atombios_get_igp_ss_overrides(struct radeon_device *rdev,
 				break;
 			}
 			break;
+		case 8:
+			switch (id) {
+			case ASIC_INTERNAL_SS_ON_TMDS:
+				percentage = le16_to_cpu(igp_info->info_8.usDVISSPercentage);
+				rate = le16_to_cpu(igp_info->info_8.usDVISSpreadRateIn10Hz);
+				break;
+			case ASIC_INTERNAL_SS_ON_HDMI:
+				percentage = le16_to_cpu(igp_info->info_8.usHDMISSPercentage);
+				rate = le16_to_cpu(igp_info->info_8.usHDMISSpreadRateIn10Hz);
+				break;
+			case ASIC_INTERNAL_SS_ON_LVDS:
+				percentage = le16_to_cpu(igp_info->info_8.usLvdsSSPercentage);
+				rate = le16_to_cpu(igp_info->info_8.usLvdsSSpreadRateIn10Hz);
+				break;
+			}
+			break;
 		default:
 			DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev);
 			break;
@@ -1499,6 +1513,10 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev,
 						le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadSpectrumPercentage);
 					ss->type = ss_info->info_2.asSpreadSpectrum[i].ucSpreadSpectrumMode;
 					ss->rate = le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadRateIn10Hz);
+					if ((crev == 2) &&
+					    ((id == ASIC_INTERNAL_ENGINE_SS) ||
+					     (id == ASIC_INTERNAL_MEMORY_SS)))
+						ss->rate /= 100;
 					return true;
 				}
 			}
@@ -1513,6 +1531,9 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev,
 						le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadSpectrumPercentage);
 					ss->type = ss_info->info_3.asSpreadSpectrum[i].ucSpreadSpectrumMode;
 					ss->rate = le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadRateIn10Hz);
+					if ((id == ASIC_INTERNAL_ENGINE_SS) ||
+					    (id == ASIC_INTERNAL_MEMORY_SS))
+						ss->rate /= 100;
 					if (rdev->flags & RADEON_IS_IGP)
 						radeon_atombios_get_igp_ss_overrides(rdev, ss, id);
 					return true;
@@ -1927,6 +1948,7 @@ static const char *pp_lib_thermal_controller_names[] = {
 	"Northern Islands",
 	"Southern Islands",
 	"lm96163",
+	"Sea Islands",
 };
 
 union power_info {
@@ -1944,6 +1966,7 @@ union pplib_clock_info {
 	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
 	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
 	struct _ATOM_PPLIB_SI_CLOCK_INFO si;
+	struct _ATOM_PPLIB_CI_CLOCK_INFO ci;
 };
 
 union pplib_power_state {
@@ -2209,6 +2232,11 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 				 (controller->ucFanParameters &
 				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
 			rdev->pm.int_thermal_type = THERMAL_TYPE_SI;
+		} else if (controller->ucType == ATOM_PP_THERMALCONTROLLER_CISLANDS) {
+			DRM_INFO("Internal thermal controller %s fan control\n",
+				 (controller->ucFanParameters &
+				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_CI;
 		} else if ((controller->ucType ==
 			    ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) ||
 			   (controller->ucType ==
@@ -2241,8 +2269,8 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 	}
 }
 
-static void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
-						 u16 *vddc, u16 *vddci)
+void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
+					  u16 *vddc, u16 *vddci, u16 *mvdd)
 {
 	struct radeon_mode_info *mode_info = &rdev->mode_info;
 	int index = GetIndexIntoMasterTable(DATA, FirmwareInfo);
@@ -2252,6 +2280,7 @@ static void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
 
 	*vddc = 0;
 	*vddci = 0;
+	*mvdd = 0;
 
 	if (atom_parse_data_header(mode_info->atom_context, index, NULL,
 				   &frev, &crev, &data_offset)) {
@@ -2259,8 +2288,10 @@ static void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
 			(union firmware_info *)(mode_info->atom_context->bios +
 						data_offset);
 		*vddc = le16_to_cpu(firmware_info->info_14.usBootUpVDDCVoltage);
-		if ((frev == 2) && (crev >= 2))
+		if ((frev == 2) && (crev >= 2)) {
 			*vddci = le16_to_cpu(firmware_info->info_22.usBootUpVDDCIVoltage);
+			*mvdd = le16_to_cpu(firmware_info->info_22.usBootUpMVDDCVoltage);
+		}
 	}
 }
 
@@ -2271,9 +2302,9 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 	int j;
 	u32 misc = le32_to_cpu(non_clock_info->ulCapsAndSettings);
 	u32 misc2 = le16_to_cpu(non_clock_info->usClassification);
-	u16 vddc, vddci;
+	u16 vddc, vddci, mvdd;
 
-	radeon_atombios_get_default_voltages(rdev, &vddc, &vddci);
+	radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
 
 	rdev->pm.power_state[state_index].misc = misc;
 	rdev->pm.power_state[state_index].misc2 = misc2;
@@ -2316,7 +2347,13 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 			rdev->pm.default_vddc = rdev->pm.power_state[state_index].clock_info[0].voltage.voltage;
 			rdev->pm.default_vddci = rdev->pm.power_state[state_index].clock_info[0].voltage.vddci;
 		} else {
-			/* patch the table values with the default slck/mclk from firmware info */
+			u16 max_vddci = 0;
+
+			if (ASIC_IS_DCE4(rdev))
+				radeon_atom_get_max_voltage(rdev,
+							    SET_VOLTAGE_TYPE_ASIC_VDDCI,
+							    &max_vddci);
+			/* patch the table values with the default sclk/mclk from firmware info */
 			for (j = 0; j < mode_index; j++) {
 				rdev->pm.power_state[state_index].clock_info[j].mclk =
 					rdev->clock.default_mclk;
@@ -2325,6 +2362,9 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 				if (vddc)
 					rdev->pm.power_state[state_index].clock_info[j].voltage.voltage =
 						vddc;
+				if (max_vddci)
+					rdev->pm.power_state[state_index].clock_info[j].voltage.vddci =
+						max_vddci;
 			}
 		}
 	}
@@ -2347,6 +2387,15 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 			sclk |= clock_info->rs780.ucLowEngineClockHigh << 16;
 			rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk;
 		}
+	} else if (rdev->family >= CHIP_BONAIRE) {
+		sclk = le16_to_cpu(clock_info->ci.usEngineClockLow);
+		sclk |= clock_info->ci.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow);
+		mclk |= clock_info->ci.ucMemoryClockHigh << 16;
+		rdev->pm.power_state[state_index].clock_info[mode_index].mclk = mclk;
+		rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk;
+		rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type =
+			VOLTAGE_NONE;
 	} else if (rdev->family >= CHIP_TAHITI) {
 		sclk = le16_to_cpu(clock_info->si.usEngineClockLow);
 		sclk |= clock_info->si.ucEngineClockHigh << 16;
@@ -2392,6 +2441,10 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 	case ATOM_VIRTUAL_VOLTAGE_ID1:
 	case ATOM_VIRTUAL_VOLTAGE_ID2:
 	case ATOM_VIRTUAL_VOLTAGE_ID3:
+	case ATOM_VIRTUAL_VOLTAGE_ID4:
+	case ATOM_VIRTUAL_VOLTAGE_ID5:
+	case ATOM_VIRTUAL_VOLTAGE_ID6:
+	case ATOM_VIRTUAL_VOLTAGE_ID7:
 		if (radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC,
 					     rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage,
 					     &vddc) == 0)
@@ -2667,6 +2720,8 @@ union get_clock_dividers {
 	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3 v3;
 	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4 v4;
 	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5 v5;
+	struct _COMPUTE_GPU_CLOCK_INPUT_PARAMETERS_V1_6 v6_in;
+	struct _COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6 v6_out;
 };
 
 int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
@@ -2699,7 +2754,8 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 		break;
 	case 2:
 	case 3:
-		/* r6xx, r7xx, evergreen, ni */
+	case 5:
+		/* r6xx, r7xx, evergreen, ni, si */
 		if (rdev->family <= CHIP_RV770) {
 			args.v2.ucAction = clock_type;
 			args.v2.ulClock = cpu_to_le32(clock);	/* 10 khz */
@@ -2732,6 +2788,9 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 				dividers->vco_mode = (args.v3.ucCntlFlag &
 						      ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE) ? 1 : 0;
 			} else {
+				/* for SI we use ComputeMemoryClockParam for memory plls */
+				if (rdev->family >= CHIP_TAHITI)
+					return -EINVAL;
 				args.v5.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
 				if (strobe_mode)
 					args.v5.ucInputFlag = ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN;
@@ -2757,9 +2816,76 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 
 		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
-		dividers->post_div = args.v4.ucPostDiv;
+		dividers->post_divider = dividers->post_div = args.v4.ucPostDiv;
 		dividers->real_clock = le32_to_cpu(args.v4.ulClock);
 		break;
+	case 6:
+		/* CI */
+		/* COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, COMPUTE_GPUCLK_INPUT_FLAG_SCLK */
+		args.v6_in.ulClock.ulComputeClockFlag = clock_type;
+		args.v6_in.ulClock.ulClockFreq = cpu_to_le32(clock);	/* 10 khz */
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		dividers->whole_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDiv);
+		dividers->frac_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDivFrac);
+		dividers->ref_div = args.v6_out.ucPllRefDiv;
+		dividers->post_div = args.v6_out.ucPllPostDiv;
+		dividers->flags = args.v6_out.ucPllCntlFlag;
+		dividers->real_clock = le32_to_cpu(args.v6_out.ulClock.ulClock);
+		dividers->post_divider = args.v6_out.ulClock.ucPostDiv;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
+int radeon_atom_get_memory_pll_dividers(struct radeon_device *rdev,
+					u32 clock,
+					bool strobe_mode,
+					struct atom_mpll_param *mpll_param)
+{
+	COMPUTE_MEMORY_CLOCK_PARAM_PARAMETERS_V2_1 args;
+	int index = GetIndexIntoMasterTable(COMMAND, ComputeMemoryClockParam);
+	u8 frev, crev;
+
+	memset(&args, 0, sizeof(args));
+	memset(mpll_param, 0, sizeof(struct atom_mpll_param));
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return -EINVAL;
+
+	switch (frev) {
+	case 2:
+		switch (crev) {
+		case 1:
+			/* SI */
+			args.ulClock = cpu_to_le32(clock);	/* 10 khz */
+			args.ucInputFlag = 0;
+			if (strobe_mode)
+				args.ucInputFlag |= MPLL_INPUT_FLAG_STROBE_MODE_EN;
+
+			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+			mpll_param->clkfrac = le16_to_cpu(args.ulFbDiv.usFbDivFrac);
+			mpll_param->clkf = le16_to_cpu(args.ulFbDiv.usFbDiv);
+			mpll_param->post_div = args.ucPostDiv;
+			mpll_param->dll_speed = args.ucDllSpeed;
+			mpll_param->bwcntl = args.ucBWCntl;
+			mpll_param->vco_mode =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_VCO_MODE_MASK) ? 1 : 0;
+			mpll_param->yclk_sel =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_BYPASS_DQ_PLL) ? 1 : 0;
+			mpll_param->qdr =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_QDR_ENABLE) ? 1 : 0;
+			mpll_param->half_rate =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_AD_HALF_RATE) ? 1 : 0;
+			break;
+		default:
+			return -EINVAL;
+		}
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -2819,6 +2945,48 @@ void radeon_atom_set_memory_clock(struct radeon_device *rdev,
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
+void radeon_atom_set_engine_dram_timings(struct radeon_device *rdev,
+					 u32 eng_clock, u32 mem_clock)
+{
+	SET_ENGINE_CLOCK_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
+	u32 tmp;
+
+	memset(&args, 0, sizeof(args));
+
+	tmp = eng_clock & SET_CLOCK_FREQ_MASK;
+	tmp |= (COMPUTE_ENGINE_PLL_PARAM << 24);
+
+	args.ulTargetEngineClock = cpu_to_le32(tmp);
+	if (mem_clock)
+		args.sReserved.ulClock = cpu_to_le32(mem_clock & SET_CLOCK_FREQ_MASK);
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+void radeon_atom_update_memory_dll(struct radeon_device *rdev,
+				   u32 mem_clock)
+{
+	u32 args;
+	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
+
+	args = cpu_to_le32(mem_clock);	/* 10 khz */
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+void radeon_atom_set_ac_timing(struct radeon_device *rdev,
+			       u32 mem_clock)
+{
+	SET_MEMORY_CLOCK_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
+	u32 tmp = mem_clock | (COMPUTE_MEMORY_PLL_PARAM << 24);
+
+	args.ulTargetMemoryClock = cpu_to_le32(tmp);	/* 10 khz */
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
 union set_voltage {
 	struct _SET_VOLTAGE_PS_ALLOCATION alloc;
 	struct _SET_VOLTAGE_PARAMETERS v1;
@@ -2863,8 +3031,8 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
-static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
-				    u16 voltage_id, u16 *voltage)
+int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage)
 {
 	union set_voltage args;
 	int index = GetIndexIntoMasterTable(COMMAND, SetVoltage);
@@ -2902,6 +3070,695 @@ static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
 	return 0;
 }
 
+int radeon_atom_get_leakage_vddc_based_on_leakage_idx(struct radeon_device *rdev,
+						      u16 *voltage,
+						      u16 leakage_idx)
+{
+	return radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC, leakage_idx, voltage);
+}
+
+int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev,
+					  u16 voltage_level, u8 voltage_type,
+					  u32 *gpio_value, u32 *gpio_mask)
+{
+	union set_voltage args;
+	int index = GetIndexIntoMasterTable(COMMAND, SetVoltage);
+	u8 frev, crev;
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return -EINVAL;
+
+	switch (crev) {
+	case 1:
+		return -EINVAL;
+	case 2:
+		args.v2.ucVoltageType = voltage_type;
+		args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_GET_GPIOMASK;
+		args.v2.usVoltageLevel = cpu_to_le16(voltage_level);
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*gpio_mask = le32_to_cpu(*(u32 *)&args.v2);
+
+		args.v2.ucVoltageType = voltage_type;
+		args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL;
+		args.v2.usVoltageLevel = cpu_to_le16(voltage_level);
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*gpio_value = le32_to_cpu(*(u32 *)&args.v2);
+		break;
+	default:
+		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+union voltage_object_info {
+	struct _ATOM_VOLTAGE_OBJECT_INFO v1;
+	struct _ATOM_VOLTAGE_OBJECT_INFO_V2 v2;
+	struct _ATOM_VOLTAGE_OBJECT_INFO_V3_1 v3;
+};
+
+union voltage_object {
+	struct _ATOM_VOLTAGE_OBJECT v1;
+	struct _ATOM_VOLTAGE_OBJECT_V2 v2;
+	union _ATOM_VOLTAGE_OBJECT_V3 v3;
+};
+
+static ATOM_VOLTAGE_OBJECT *atom_lookup_voltage_object_v1(ATOM_VOLTAGE_OBJECT_INFO *v1,
+							  u8 voltage_type)
+{
+	u32 size = le16_to_cpu(v1->sHeader.usStructureSize);
+	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO, asVoltageObj[0]);
+	u8 *start = (u8 *)v1;
+
+	while (offset < size) {
+		ATOM_VOLTAGE_OBJECT *vo = (ATOM_VOLTAGE_OBJECT *)(start + offset);
+		if (vo->ucVoltageType == voltage_type)
+			return vo;
+		offset += offsetof(ATOM_VOLTAGE_OBJECT, asFormula.ucVIDAdjustEntries) +
+			vo->asFormula.ucNumOfVoltageEntries;
+	}
+	return NULL;
+}
+
+static ATOM_VOLTAGE_OBJECT_V2 *atom_lookup_voltage_object_v2(ATOM_VOLTAGE_OBJECT_INFO_V2 *v2,
+							     u8 voltage_type)
+{
+	u32 size = le16_to_cpu(v2->sHeader.usStructureSize);
+	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO_V2, asVoltageObj[0]);
+	u8 *start = (u8*)v2;
+
+	while (offset < size) {
+		ATOM_VOLTAGE_OBJECT_V2 *vo = (ATOM_VOLTAGE_OBJECT_V2 *)(start + offset);
+		if (vo->ucVoltageType == voltage_type)
+			return vo;
+		offset += offsetof(ATOM_VOLTAGE_OBJECT_V2, asFormula.asVIDAdjustEntries) +
+			(vo->asFormula.ucNumOfVoltageEntries * sizeof(VOLTAGE_LUT_ENTRY));
+	}
+	return NULL;
+}
+
+static ATOM_VOLTAGE_OBJECT_V3 *atom_lookup_voltage_object_v3(ATOM_VOLTAGE_OBJECT_INFO_V3_1 *v3,
+							     u8 voltage_type, u8 voltage_mode)
+{
+	u32 size = le16_to_cpu(v3->sHeader.usStructureSize);
+	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO_V3_1, asVoltageObj[0]);
+	u8 *start = (u8*)v3;
+
+	while (offset < size) {
+		ATOM_VOLTAGE_OBJECT_V3 *vo = (ATOM_VOLTAGE_OBJECT_V3 *)(start + offset);
+		if ((vo->asGpioVoltageObj.sHeader.ucVoltageType == voltage_type) &&
+		    (vo->asGpioVoltageObj.sHeader.ucVoltageMode == voltage_mode))
+			return vo;
+		offset += le16_to_cpu(vo->asGpioVoltageObj.sHeader.usSize);
+	}
+	return NULL;
+}
+
+bool
+radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
+			    u8 voltage_type, u8 voltage_mode)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 1:
+		case 2:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+				if (voltage_object &&
+				    (voltage_object->v1.asControl.ucVoltageControlId == VOLTAGE_CONTROLLED_BY_GPIO))
+					return true;
+				break;
+			case 2:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+				if (voltage_object &&
+				    (voltage_object->v2.asControl.ucVoltageControlId == VOLTAGE_CONTROLLED_BY_GPIO))
+					return true;
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return false;
+			}
+			break;
+		case 3:
+			switch (crev) {
+			case 1:
+				if (atom_lookup_voltage_object_v3(&voltage_info->v3,
+								  voltage_type, voltage_mode))
+					return true;
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return false;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return false;
+		}
+
+	}
+	return false;
+}
+
+int radeon_atom_get_max_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *max_voltage)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (crev) {
+		case 1:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA *formula =
+					&voltage_object->v1.asFormula;
+				if (formula->ucFlag & 1)
+					*max_voltage =
+						le16_to_cpu(formula->usVoltageBaseLevel) +
+						formula->ucNumOfVoltageEntries / 2 *
+						le16_to_cpu(formula->usVoltageStep);
+				else
+					*max_voltage =
+						le16_to_cpu(formula->usVoltageBaseLevel) +
+						(formula->ucNumOfVoltageEntries - 1) *
+						le16_to_cpu(formula->usVoltageStep);
+				return 0;
+			}
+			break;
+		case 2:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA_V2 *formula =
+					&voltage_object->v2.asFormula;
+				if (formula->ucNumOfVoltageEntries) {
+					*max_voltage =
+						le16_to_cpu(formula->asVIDAdjustEntries[
+								    formula->ucNumOfVoltageEntries - 1
+								    ].usVoltageValue);
+					return 0;
+				}
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_get_min_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *min_voltage)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (crev) {
+		case 1:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA *formula =
+					&voltage_object->v1.asFormula;
+				*min_voltage =
+					le16_to_cpu(formula->usVoltageBaseLevel);
+				return 0;
+			}
+			break;
+		case 2:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA_V2 *formula =
+					&voltage_object->v2.asFormula;
+				if (formula->ucNumOfVoltageEntries) {
+					*min_voltage =
+						le16_to_cpu(formula->asVIDAdjustEntries[
+								    0
+								    ].usVoltageValue);
+					return 0;
+				}
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_get_voltage_step(struct radeon_device *rdev,
+				 u8 voltage_type, u16 *voltage_step)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (crev) {
+		case 1:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA *formula =
+					&voltage_object->v1.asFormula;
+				if (formula->ucFlag & 1)
+					*voltage_step =
+						(le16_to_cpu(formula->usVoltageStep) + 1) / 2;
+				else
+					*voltage_step =
+						le16_to_cpu(formula->usVoltageStep);
+				return 0;
+			}
+			break;
+		case 2:
+			return -EINVAL;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_round_to_true_voltage(struct radeon_device *rdev,
+				      u8 voltage_type,
+				      u16 nominal_voltage,
+				      u16 *true_voltage)
+{
+	u16 min_voltage, max_voltage, voltage_step;
+
+	if (radeon_atom_get_max_voltage(rdev, voltage_type, &max_voltage))
+		return -EINVAL;
+	if (radeon_atom_get_min_voltage(rdev, voltage_type, &min_voltage))
+		return -EINVAL;
+	if (radeon_atom_get_voltage_step(rdev, voltage_type, &voltage_step))
+		return -EINVAL;
+
+	if (nominal_voltage <= min_voltage)
+		*true_voltage = min_voltage;
+	else if (nominal_voltage >= max_voltage)
+		*true_voltage = max_voltage;
+	else
+		*true_voltage = min_voltage +
+			((nominal_voltage - min_voltage) / voltage_step) *
+			voltage_step;
+
+	return 0;
+}
+
+int radeon_atom_get_voltage_table(struct radeon_device *rdev,
+				  u8 voltage_type, u8 voltage_mode,
+				  struct atom_voltage_table *voltage_table)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	int i, ret;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 1:
+		case 2:
+			switch (crev) {
+			case 1:
+				DRM_ERROR("old table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			case 2:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+				if (voltage_object) {
+					ATOM_VOLTAGE_FORMULA_V2 *formula =
+						&voltage_object->v2.asFormula;
+					if (formula->ucNumOfVoltageEntries > MAX_VOLTAGE_ENTRIES)
+						return -EINVAL;
+					for (i = 0; i < formula->ucNumOfVoltageEntries; i++) {
+						voltage_table->entries[i].value =
+							le16_to_cpu(formula->asVIDAdjustEntries[i].usVoltageValue);
+						ret = radeon_atom_get_voltage_gpio_settings(rdev,
+											    voltage_table->entries[i].value,
+											    voltage_type,
+											    &voltage_table->entries[i].smio_low,
+											    &voltage_table->mask_low);
+						if (ret)
+							return ret;
+					}
+					voltage_table->count = formula->ucNumOfVoltageEntries;
+					return 0;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		case 3:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v3(&voltage_info->v3,
+								      voltage_type, voltage_mode);
+				if (voltage_object) {
+					ATOM_GPIO_VOLTAGE_OBJECT_V3 *gpio =
+						&voltage_object->v3.asGpioVoltageObj;
+					if (gpio->ucGpioEntryNum > MAX_VOLTAGE_ENTRIES)
+						return -EINVAL;
+					for (i = 0; i < gpio->ucGpioEntryNum; i++) {
+						voltage_table->entries[i].value =
+							le16_to_cpu(gpio->asVolGpioLut[i].usVoltageValue);
+						voltage_table->entries[i].smio_low =
+							le32_to_cpu(gpio->asVolGpioLut[i].ulVoltageId);
+					}
+					voltage_table->mask_low = le32_to_cpu(gpio->ulGpioMaskVal);
+					voltage_table->count = gpio->ucGpioEntryNum;
+					voltage_table->phase_delay = gpio->ucPhaseDelay;
+					return 0;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+	}
+	return -EINVAL;
+}
+
+union vram_info {
+	struct _ATOM_VRAM_INFO_V3 v1_3;
+	struct _ATOM_VRAM_INFO_V4 v1_4;
+	struct _ATOM_VRAM_INFO_HEADER_V2_1 v2_1;
+};
+
+int radeon_atom_get_memory_info(struct radeon_device *rdev,
+				u8 module_index, struct atom_memory_info *mem_info)
+{
+	int index = GetIndexIntoMasterTable(DATA, VRAM_Info);
+	u8 frev, crev, i;
+	u16 data_offset, size;
+	union vram_info *vram_info;
+	u8 *p;
+
+	memset(mem_info, 0, sizeof(struct atom_memory_info));
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		vram_info = (union vram_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+		switch (frev) {
+		case 1:
+			switch (crev) {
+			case 3:
+				/* r6xx */
+				if (module_index < vram_info->v1_3.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V3 *vram_module =
+						(ATOM_VRAM_MODULE_V3 *)vram_info->v1_3.aVramInfo;
+					p = (u8 *)vram_info->v1_3.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						vram_module = (ATOM_VRAM_MODULE_V3 *)p;
+						if (le16_to_cpu(vram_module->usSize) == 0)
+							return -EINVAL;
+						p += le16_to_cpu(vram_module->usSize);
+					}
+					mem_info->mem_vendor = vram_module->asMemory.ucMemoryVenderID & 0xf;
+					mem_info->mem_type = vram_module->asMemory.ucMemoryType & 0xf0;
+				} else
+					return -EINVAL;
+				break;
+			case 4:
+				/* r7xx, evergreen */
+				if (module_index < vram_info->v1_4.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V4 *vram_module =
+						(ATOM_VRAM_MODULE_V4 *)vram_info->v1_4.aVramInfo;
+					p = (u8 *)vram_info->v1_4.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						vram_module = (ATOM_VRAM_MODULE_V4 *)p;
+						if (le16_to_cpu(vram_module->usModuleSize) == 0)
+							return -EINVAL;
+						p += le16_to_cpu(vram_module->usModuleSize);
+					}
+					mem_info->mem_vendor = vram_module->ucMemoryVenderID & 0xf;
+					mem_info->mem_type = vram_module->ucMemoryType & 0xf0;
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		case 2:
+			switch (crev) {
+			case 1:
+				/* ni */
+				if (module_index < vram_info->v2_1.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V7 *vram_module =
+						(ATOM_VRAM_MODULE_V7 *)vram_info->v2_1.aVramInfo;
+					p = (u8 *)vram_info->v2_1.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						vram_module = (ATOM_VRAM_MODULE_V7 *)p;
+						if (le16_to_cpu(vram_module->usModuleSize) == 0)
+							return -EINVAL;
+						p += le16_to_cpu(vram_module->usModuleSize);
+					}
+					mem_info->mem_vendor = vram_module->ucMemoryVenderID & 0xf;
+					mem_info->mem_type = vram_module->ucMemoryType & 0xf0;
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		}
+		return 0;
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_get_mclk_range_table(struct radeon_device *rdev,
+				     bool gddr5, u8 module_index,
+				     struct atom_memory_clock_range_table *mclk_range_table)
+{
+	int index = GetIndexIntoMasterTable(DATA, VRAM_Info);
+	u8 frev, crev, i;
+	u16 data_offset, size;
+	union vram_info *vram_info;
+	u32 mem_timing_size = gddr5 ?
+		sizeof(ATOM_MEMORY_TIMING_FORMAT_V2) : sizeof(ATOM_MEMORY_TIMING_FORMAT);
+	u8 *p;
+
+	memset(mclk_range_table, 0, sizeof(struct atom_memory_clock_range_table));
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		vram_info = (union vram_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+		switch (frev) {
+		case 1:
+			switch (crev) {
+			case 3:
+				DRM_ERROR("old table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			case 4:
+				/* r7xx, evergreen */
+				if (module_index < vram_info->v1_4.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V4 *vram_module =
+						(ATOM_VRAM_MODULE_V4 *)vram_info->v1_4.aVramInfo;
+					ATOM_MEMORY_TIMING_FORMAT *format;
+					p = (u8 *)vram_info->v1_4.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						vram_module = (ATOM_VRAM_MODULE_V4 *)p;
+						if (le16_to_cpu(vram_module->usModuleSize) == 0)
+							return -EINVAL;
+						p += le16_to_cpu(vram_module->usModuleSize);
+					}
+					mclk_range_table->num_entries = (u8)
+						((vram_module->usModuleSize - offsetof(ATOM_VRAM_MODULE_V4, asMemTiming)) /
+						 mem_timing_size);
+					p = (u8 *)vram_module->asMemTiming;
+					for (i = 0; i < mclk_range_table->num_entries; i++) {
+						format = (ATOM_MEMORY_TIMING_FORMAT *)p;
+						mclk_range_table->mclk[i] = le32_to_cpu(format->ulClkRange);
+						p += mem_timing_size;
+					}
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		case 2:
+			DRM_ERROR("new table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		}
+		return 0;
+	}
+	return -EINVAL;
+}
+
+#define MEM_ID_MASK           0xff000000
+#define MEM_ID_SHIFT          24
+#define CLOCK_RANGE_MASK      0x00ffffff
+#define CLOCK_RANGE_SHIFT     0
+#define LOW_NIBBLE_MASK       0xf
+#define DATA_EQU_PREV         0
+#define DATA_FROM_TABLE       4
+
+int radeon_atom_init_mc_reg_table(struct radeon_device *rdev,
+				  u8 module_index,
+				  struct atom_mc_reg_table *reg_table)
+{
+	int index = GetIndexIntoMasterTable(DATA, VRAM_Info);
+	u8 frev, crev, num_entries, t_mem_id, num_ranges = 0;
+	u32 i = 0, j;
+	u16 data_offset, size;
+	union vram_info *vram_info;
+
+	memset(reg_table, 0, sizeof(struct atom_mc_reg_table));
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		vram_info = (union vram_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+		switch (frev) {
+		case 1:
+			DRM_ERROR("old table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		case 2:
+			switch (crev) {
+			case 1:
+				if (module_index < vram_info->v2_1.ucNumOfVRAMModule) {
+					ATOM_INIT_REG_BLOCK *reg_block =
+						(ATOM_INIT_REG_BLOCK *)
+						((u8 *)vram_info + le16_to_cpu(vram_info->v2_1.usMemClkPatchTblOffset));
+					ATOM_MEMORY_SETTING_DATA_BLOCK *reg_data =
+						(ATOM_MEMORY_SETTING_DATA_BLOCK *)
+						((u8 *)reg_block + (2 * sizeof(u16)) +
+						 le16_to_cpu(reg_block->usRegIndexTblSize));
+					num_entries = (u8)((le16_to_cpu(reg_block->usRegIndexTblSize)) /
+							   sizeof(ATOM_INIT_REG_INDEX_FORMAT)) - 1;
+					if (num_entries > VBIOS_MC_REGISTER_ARRAY_SIZE)
+						return -EINVAL;
+					while (!(reg_block->asRegIndexBuf[i].ucPreRegDataLength & ACCESS_PLACEHOLDER) &&
+					      (i < num_entries)) {
+						reg_table->mc_reg_address[i].s1 =
+							(u16)(le16_to_cpu(reg_block->asRegIndexBuf[i].usRegIndex));
+						reg_table->mc_reg_address[i].pre_reg_data =
+							(u8)(reg_block->asRegIndexBuf[i].ucPreRegDataLength);
+						i++;
+					}
+					reg_table->last = i;
+					while ((*(u32 *)reg_data != END_OF_REG_DATA_BLOCK) &&
+					       (num_ranges < VBIOS_MAX_AC_TIMING_ENTRIES)) {
+						t_mem_id = (u8)((*(u32 *)reg_data & MEM_ID_MASK) >> MEM_ID_SHIFT);
+						if (module_index == t_mem_id) {
+							reg_table->mc_reg_table_entry[num_ranges].mclk_max =
+								(u32)((*(u32 *)reg_data & CLOCK_RANGE_MASK) >> CLOCK_RANGE_SHIFT);
+							for (i = 0, j = 1; i < reg_table->last; i++) {
+								if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_FROM_TABLE) {
+									reg_table->mc_reg_table_entry[num_ranges].mc_data[i] =
+										(u32)*((u32 *)reg_data + j);
+									j++;
+								} else if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_EQU_PREV) {
+									reg_table->mc_reg_table_entry[num_ranges].mc_data[i] =
+										reg_table->mc_reg_table_entry[num_ranges].mc_data[i - 1];
+								}
+							}
+							num_ranges++;
+						}
+						reg_data = (ATOM_MEMORY_SETTING_DATA_BLOCK *)
+							((u8 *)reg_data + le16_to_cpu(reg_block->usRegDataBlkSize));
+					}
+					if (*(u32 *)reg_data != END_OF_REG_DATA_BLOCK)
+						return -EINVAL;
+					reg_table->num_entries = num_ranges;
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		}
+		return 0;
+	}
+	return -EINVAL;
+}
+
 void radeon_atom_initialize_bios_scratch_regs(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 7e265a5..13a130f 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -106,7 +106,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 		radeon_bo_list_add_object(&p->relocs[i].lobj,
 					  &p->validated);
 	}
-	return radeon_bo_list_validate(&p->validated, p->ring);
+	return radeon_bo_list_validate(&p->ticket, &p->validated, p->ring);
 }
 
 static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
@@ -314,15 +314,17 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
  * If error is set than unvalidate buffer, otherwise just free memory
  * used by parsing context.
  **/
-static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
+static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bool backoff)
 {
 	unsigned i;
 
 	if (!error) {
-		ttm_eu_fence_buffer_objects(&parser->validated,
+		ttm_eu_fence_buffer_objects(&parser->ticket,
+					    &parser->validated,
 					    parser->ib.fence);
-	} else {
-		ttm_eu_backoff_reservation(&parser->validated);
+	} else if (backoff) {
+		ttm_eu_backoff_reservation(&parser->ticket,
+					   &parser->validated);
 	}
 
 	if (parser->relocs != NULL) {
@@ -535,7 +537,7 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	r = radeon_cs_parser_init(&parser, data);
 	if (r) {
 		DRM_ERROR("Failed to initialize parser !\n");
-		radeon_cs_parser_fini(&parser, r);
+		radeon_cs_parser_fini(&parser, r, false);
 		up_read(&rdev->exclusive_lock);
 		r = radeon_cs_handle_lockup(rdev, r);
 		return r;
@@ -544,12 +546,13 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	if (r) {
 		if (r != -ERESTARTSYS)
 			DRM_ERROR("Failed to parse relocation %d!\n", r);
-		radeon_cs_parser_fini(&parser, r);
+		radeon_cs_parser_fini(&parser, r, false);
 		up_read(&rdev->exclusive_lock);
 		r = radeon_cs_handle_lockup(rdev, r);
 		return r;
 	}
 
+	/* XXX pick SD/HD/MVC */
 	if (parser.ring == R600_RING_TYPE_UVD_INDEX)
 		radeon_uvd_note_usage(rdev);
 
@@ -562,7 +565,7 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		goto out;
 	}
 out:
-	radeon_cs_parser_fini(&parser, r);
+	radeon_cs_parser_fini(&parser, r, true);
 	up_read(&rdev->exclusive_lock);
 	r = radeon_cs_handle_lockup(rdev, r);
 	return r;
diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c
index b097d5b..9630e8d 100644
--- a/drivers/gpu/drm/radeon/radeon_cursor.c
+++ b/drivers/gpu/drm/radeon/radeon_cursor.c
@@ -27,9 +27,6 @@
 #include <drm/radeon_drm.h>
 #include "radeon.h"
 
-#define CURSOR_WIDTH 64
-#define CURSOR_HEIGHT 64
-
 static void radeon_lock_cursor(struct drm_crtc *crtc, bool lock)
 {
 	struct radeon_device *rdev = crtc->dev->dev_private;
@@ -167,7 +164,8 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc,
 		goto unpin;
 	}
 
-	if ((width > CURSOR_WIDTH) || (height > CURSOR_HEIGHT)) {
+	if ((width > radeon_crtc->max_cursor_width) ||
+	    (height > radeon_crtc->max_cursor_height)) {
 		DRM_ERROR("bad cursor width or height %d x %d\n", width, height);
 		return -EINVAL;
 	}
@@ -233,11 +231,11 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 	DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y);
 
 	if (x < 0) {
-		xorigin = min(-x, CURSOR_WIDTH - 1);
+		xorigin = min(-x, radeon_crtc->max_cursor_width - 1);
 		x = 0;
 	}
 	if (y < 0) {
-		yorigin = min(-y, CURSOR_HEIGHT - 1);
+		yorigin = min(-y, radeon_crtc->max_cursor_height - 1);
 		y = 0;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index b0dc0b6..82335e3 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -95,6 +95,9 @@ static const char radeon_family_name[][16] = {
 	"VERDE",
 	"OLAND",
 	"HAINAN",
+	"BONAIRE",
+	"KAVERI",
+	"KABINI",
 	"LAST",
 };
 
@@ -229,6 +232,94 @@ void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg)
 }
 
 /*
+ * GPU doorbell aperture helpers function.
+ */
+/**
+ * radeon_doorbell_init - Init doorbell driver information.
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Init doorbell driver information (CIK)
+ * Returns 0 on success, error on failure.
+ */
+int radeon_doorbell_init(struct radeon_device *rdev)
+{
+	int i;
+
+	/* doorbell bar mapping */
+	rdev->doorbell.base = pci_resource_start(rdev->pdev, 2);
+	rdev->doorbell.size = pci_resource_len(rdev->pdev, 2);
+
+	/* limit to 4 MB for now */
+	if (rdev->doorbell.size > (4 * 1024 * 1024))
+		rdev->doorbell.size = 4 * 1024 * 1024;
+
+	rdev->doorbell.ptr = ioremap(rdev->doorbell.base, rdev->doorbell.size);
+	if (rdev->doorbell.ptr == NULL) {
+		return -ENOMEM;
+	}
+	DRM_INFO("doorbell mmio base: 0x%08X\n", (uint32_t)rdev->doorbell.base);
+	DRM_INFO("doorbell mmio size: %u\n", (unsigned)rdev->doorbell.size);
+
+	rdev->doorbell.num_pages = rdev->doorbell.size / PAGE_SIZE;
+
+	for (i = 0; i < rdev->doorbell.num_pages; i++) {
+		rdev->doorbell.free[i] = true;
+	}
+	return 0;
+}
+
+/**
+ * radeon_doorbell_fini - Tear down doorbell driver information.
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down doorbell driver information (CIK)
+ */
+void radeon_doorbell_fini(struct radeon_device *rdev)
+{
+	iounmap(rdev->doorbell.ptr);
+	rdev->doorbell.ptr = NULL;
+}
+
+/**
+ * radeon_doorbell_get - Allocate a doorbell page
+ *
+ * @rdev: radeon_device pointer
+ * @doorbell: doorbell page number
+ *
+ * Allocate a doorbell page for use by the driver (all asics).
+ * Returns 0 on success or -EINVAL on failure.
+ */
+int radeon_doorbell_get(struct radeon_device *rdev, u32 *doorbell)
+{
+	int i;
+
+	for (i = 0; i < rdev->doorbell.num_pages; i++) {
+		if (rdev->doorbell.free[i]) {
+			rdev->doorbell.free[i] = false;
+			*doorbell = i;
+			return 0;
+		}
+	}
+	return -EINVAL;
+}
+
+/**
+ * radeon_doorbell_free - Free a doorbell page
+ *
+ * @rdev: radeon_device pointer
+ * @doorbell: doorbell page number
+ *
+ * Free a doorbell page allocated for use by the driver (all asics)
+ */
+void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell)
+{
+	if (doorbell < rdev->doorbell.num_pages)
+		rdev->doorbell.free[doorbell] = true;
+}
+
+/*
  * radeon_wb_*()
  * Writeback is the the method by which the the GPU updates special pages
  * in memory with the status of certain GPU events (fences, ring pointers,
@@ -1145,8 +1236,13 @@ int radeon_device_init(struct radeon_device *rdev,
 	/* Registers mapping */
 	/* TODO: block userspace mapping of io register */
 	spin_lock_init(&rdev->mmio_idx_lock);
-	rdev->rmmio_base = pci_resource_start(rdev->pdev, 2);
-	rdev->rmmio_size = pci_resource_len(rdev->pdev, 2);
+	if (rdev->family >= CHIP_BONAIRE) {
+		rdev->rmmio_base = pci_resource_start(rdev->pdev, 5);
+		rdev->rmmio_size = pci_resource_len(rdev->pdev, 5);
+	} else {
+		rdev->rmmio_base = pci_resource_start(rdev->pdev, 2);
+		rdev->rmmio_size = pci_resource_len(rdev->pdev, 2);
+	}
 	rdev->rmmio = ioremap(rdev->rmmio_base, rdev->rmmio_size);
 	if (rdev->rmmio == NULL) {
 		return -ENOMEM;
@@ -1154,6 +1250,10 @@ int radeon_device_init(struct radeon_device *rdev,
 	DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)rdev->rmmio_base);
 	DRM_INFO("register mmio size: %u\n", (unsigned)rdev->rmmio_size);
 
+	/* doorbell bar mapping */
+	if (rdev->family >= CHIP_BONAIRE)
+		radeon_doorbell_init(rdev);
+
 	/* io port mapping */
 	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
 		if (pci_resource_flags(rdev->pdev, i) & IORESOURCE_IO) {
@@ -1231,6 +1331,8 @@ void radeon_device_fini(struct radeon_device *rdev)
 	rdev->rio_mem = NULL;
 	iounmap(rdev->rmmio);
 	rdev->rmmio = NULL;
+	if (rdev->family >= CHIP_BONAIRE)
+		radeon_doorbell_fini(rdev);
 	radeon_debugfs_remove_files(rdev);
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index eb18bb7..c2b67b4 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -153,7 +153,13 @@ static void dce5_crtc_load_lut(struct drm_crtc *crtc)
 		NI_OUTPUT_CSC_OVL_MODE(NI_OUTPUT_CSC_BYPASS)));
 	/* XXX match this to the depth of the crtc fmt block, move to modeset? */
 	WREG32(0x6940 + radeon_crtc->crtc_offset, 0);
-
+	if (ASIC_IS_DCE8(rdev)) {
+		/* XXX this only needs to be programmed once per crtc at startup,
+		 * not sure where the best place for it is
+		 */
+		WREG32(CIK_ALPHA_CONTROL + radeon_crtc->crtc_offset,
+		       CIK_CURSOR_ALPHA_BLND_ENA);
+	}
 }
 
 static void legacy_crtc_load_lut(struct drm_crtc *crtc)
@@ -512,6 +518,14 @@ static void radeon_crtc_init(struct drm_device *dev, int index)
 	radeon_crtc->crtc_id = index;
 	rdev->mode_info.crtcs[index] = radeon_crtc;
 
+	if (rdev->family >= CHIP_BONAIRE) {
+		radeon_crtc->max_cursor_width = CIK_CURSOR_WIDTH;
+		radeon_crtc->max_cursor_height = CIK_CURSOR_HEIGHT;
+	} else {
+		radeon_crtc->max_cursor_width = CURSOR_WIDTH;
+		radeon_crtc->max_cursor_height = CURSOR_HEIGHT;
+	}
+
 #if 0
 	radeon_crtc->mode_set.crtc = &radeon_crtc->base;
 	radeon_crtc->mode_set.connectors = (struct drm_connector **)(radeon_crtc + 1);
@@ -530,7 +544,7 @@ static void radeon_crtc_init(struct drm_device *dev, int index)
 		radeon_legacy_init_crtc(dev, radeon_crtc);
 }
 
-static const char *encoder_names[37] = {
+static const char *encoder_names[38] = {
 	"NONE",
 	"INTERNAL_LVDS",
 	"INTERNAL_TMDS1",
@@ -567,7 +581,8 @@ static const char *encoder_names[37] = {
 	"INTERNAL_UNIPHY2",
 	"NUTMEG",
 	"TRAVIS",
-	"INTERNAL_VCE"
+	"INTERNAL_VCE",
+	"INTERNAL_UNIPHY3",
 };
 
 static const char *hpd_names[6] = {
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 094e7e5..e5419b3 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -74,9 +74,10 @@
  *   2.31.0 - Add fastfb support for rs690
  *   2.32.0 - new info request for rings working
  *   2.33.0 - Add SI tiling mode array query
+ *   2.34.0 - Add CIK tiling mode array query
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	33
+#define KMS_DRIVER_MINOR	34
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -127,6 +128,7 @@ struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
 							size_t size,
 							struct sg_table *sg);
 int radeon_gem_prime_pin(struct drm_gem_object *obj);
+void radeon_gem_prime_unpin(struct drm_gem_object *obj);
 void *radeon_gem_prime_vmap(struct drm_gem_object *obj);
 void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
 extern long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd,
@@ -164,6 +166,7 @@ int radeon_pcie_gen2 = -1;
 int radeon_msi = -1;
 int radeon_lockup_timeout = 10000;
 int radeon_fastfb = 0;
+int radeon_dpm = -1;
 
 MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers");
 module_param_named(no_wb, radeon_no_wb, int, 0444);
@@ -219,6 +222,9 @@ module_param_named(lockup_timeout, radeon_lockup_timeout, int, 0444);
 MODULE_PARM_DESC(fastfb, "Direct FB access for IGP chips (0 = disable, 1 = enable)");
 module_param_named(fastfb, radeon_fastfb, int, 0444);
 
+MODULE_PARM_DESC(dpm, "DPM support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(dpm, radeon_dpm, int, 0444);
+
 static struct pci_device_id pciidlist[] = {
 	radeon_PCI_IDS
 };
@@ -422,6 +428,7 @@ static struct drm_driver kms_driver = {
 	.gem_prime_export = drm_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = radeon_gem_prime_pin,
+	.gem_prime_unpin = radeon_gem_prime_unpin,
 	.gem_prime_get_sg_table = radeon_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = radeon_gem_prime_import_sg_table,
 	.gem_prime_vmap = radeon_gem_prime_vmap,
diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h
index 36e9803..3c82890 100644
--- a/drivers/gpu/drm/radeon/radeon_family.h
+++ b/drivers/gpu/drm/radeon/radeon_family.h
@@ -93,6 +93,9 @@ enum radeon_family {
 	CHIP_VERDE,
 	CHIP_OLAND,
 	CHIP_HAINAN,
+	CHIP_BONAIRE,
+	CHIP_KAVERI,
+	CHIP_KABINI,
 	CHIP_LAST,
 };
 
diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c
index 5a99d43..bcdefd1 100644
--- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
@@ -82,6 +82,23 @@ static void radeon_hotplug_work_func(struct work_struct *work)
 }
 
 /**
+ * radeon_irq_reset_work_func - execute gpu reset
+ *
+ * @work: work struct
+ *
+ * Execute scheduled gpu reset (cayman+).
+ * This function is called when the irq handler
+ * thinks we need a gpu reset.
+ */
+static void radeon_irq_reset_work_func(struct work_struct *work)
+{
+	struct radeon_device *rdev = container_of(work, struct radeon_device,
+						  reset_work);
+
+	radeon_gpu_reset(rdev);
+}
+
+/**
  * radeon_driver_irq_preinstall_kms - drm irq preinstall callback
  *
  * @dev: drm dev pointer
@@ -99,6 +116,7 @@ void radeon_driver_irq_preinstall_kms(struct drm_device *dev)
 	/* Disable *all* interrupts */
 	for (i = 0; i < RADEON_NUM_RINGS; i++)
 		atomic_set(&rdev->irq.ring_int[i], 0);
+	rdev->irq.dpm_thermal = false;
 	for (i = 0; i < RADEON_MAX_HPD_PINS; i++)
 		rdev->irq.hpd[i] = false;
 	for (i = 0; i < RADEON_MAX_CRTCS; i++) {
@@ -146,6 +164,7 @@ void radeon_driver_irq_uninstall_kms(struct drm_device *dev)
 	/* Disable *all* interrupts */
 	for (i = 0; i < RADEON_NUM_RINGS; i++)
 		atomic_set(&rdev->irq.ring_int[i], 0);
+	rdev->irq.dpm_thermal = false;
 	for (i = 0; i < RADEON_MAX_HPD_PINS; i++)
 		rdev->irq.hpd[i] = false;
 	for (i = 0; i < RADEON_MAX_CRTCS; i++) {
@@ -243,6 +262,7 @@ int radeon_irq_kms_init(struct radeon_device *rdev)
 
 	INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func);
 	INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi);
+	INIT_WORK(&rdev->reset_work, radeon_irq_reset_work_func);
 
 	spin_lock_init(&rdev->irq.lock);
 	r = drm_vblank_init(rdev->ddev, rdev->num_crtc);
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 4f2d4f4..49ff3d1 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -229,7 +229,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		*value = rdev->accel_working;
 		break;
 	case RADEON_INFO_TILING_CONFIG:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.tile_config;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.tile_config;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.tile_config;
@@ -281,7 +283,10 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			*value = rdev->clock.spll.reference_freq * 10;
 		break;
 	case RADEON_INFO_NUM_BACKENDS:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_backends_per_se *
+				rdev->config.cik.max_shader_engines;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_backends_per_se *
 				rdev->config.si.max_shader_engines;
 		else if (rdev->family >= CHIP_CAYMAN)
@@ -298,7 +303,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		}
 		break;
 	case RADEON_INFO_NUM_TILE_PIPES:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_tile_pipes;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_tile_pipes;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.max_tile_pipes;
@@ -316,7 +323,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		*value = 1;
 		break;
 	case RADEON_INFO_BACKEND_MAP:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			return -EINVAL;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.backend_map;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.backend_map;
@@ -343,7 +352,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		*value = RADEON_IB_VM_MAX_SIZE;
 		break;
 	case RADEON_INFO_MAX_PIPES:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_cu_per_sh;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_cu_per_sh;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.max_pipes_per_simd;
@@ -367,7 +378,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		value64 = radeon_get_gpu_clock_counter(rdev);
 		break;
 	case RADEON_INFO_MAX_SE:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_shader_engines;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_shader_engines;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.max_shader_engines;
@@ -377,7 +390,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			*value = 1;
 		break;
 	case RADEON_INFO_MAX_SH_PER_SE:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_sh_per_se;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_sh_per_se;
 		else
 			return -EINVAL;
@@ -407,12 +422,16 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		}
 		break;
 	case RADEON_INFO_SI_TILE_MODE_ARRAY:
-		if (rdev->family < CHIP_TAHITI) {
-			DRM_DEBUG_KMS("tile mode array is si only!\n");
+		if (rdev->family >= CHIP_BONAIRE) {
+			value = rdev->config.cik.tile_mode_array;
+			value_size = sizeof(uint32_t)*32;
+		} else if (rdev->family >= CHIP_TAHITI) {
+			value = rdev->config.si.tile_mode_array;
+			value_size = sizeof(uint32_t)*32;
+		} else {
+			DRM_DEBUG_KMS("tile mode array is si+ only!\n");
 			return -EINVAL;
 		}
-		value = rdev->config.si.tile_mode_array;
-		value_size = sizeof(uint32_t)*32;
 		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 69ad4fe..8296632 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -307,6 +307,8 @@ struct radeon_crtc {
 	uint64_t cursor_addr;
 	int cursor_width;
 	int cursor_height;
+	int max_cursor_width;
+	int max_cursor_height;
 	uint32_t legacy_display_base_addr;
 	uint32_t legacy_cursor_offset;
 	enum radeon_rmx_type rmx_type;
@@ -329,6 +331,11 @@ struct radeon_crtc {
 	u32 pll_flags;
 	struct drm_encoder *encoder;
 	struct drm_connector *connector;
+	/* for dpm */
+	u32 line_time;
+	u32 wm_low;
+	u32 wm_high;
+	struct drm_display_mode hw_mode;
 };
 
 struct radeon_encoder_primary_dac {
@@ -512,12 +519,99 @@ struct atom_clock_dividers {
 	bool enable_dithen;
 	u32 vco_mode;
 	u32 real_clock;
+	/* added for CI */
+	u32 post_divider;
+	u32 flags;
+};
+
+struct atom_mpll_param {
+	union {
+		struct {
+#ifdef __BIG_ENDIAN
+			u32 reserved : 8;
+			u32 clkfrac : 12;
+			u32 clkf : 12;
+#else
+			u32 clkf : 12;
+			u32 clkfrac : 12;
+			u32 reserved : 8;
+#endif
+		};
+		u32 fb_div;
+	};
+	u32 post_div;
+	u32 bwcntl;
+	u32 dll_speed;
+	u32 vco_mode;
+	u32 yclk_sel;
+	u32 qdr;
+	u32 half_rate;
+};
+
+#define MEM_TYPE_GDDR5  0x50
+#define MEM_TYPE_GDDR4  0x40
+#define MEM_TYPE_GDDR3  0x30
+#define MEM_TYPE_DDR2   0x20
+#define MEM_TYPE_GDDR1  0x10
+#define MEM_TYPE_DDR3   0xb0
+#define MEM_TYPE_MASK   0xf0
+
+struct atom_memory_info {
+	u8 mem_vendor;
+	u8 mem_type;
+};
+
+#define MAX_AC_TIMING_ENTRIES 16
+
+struct atom_memory_clock_range_table
+{
+	u8 num_entries;
+	u8 rsv[3];
+	u32 mclk[MAX_AC_TIMING_ENTRIES];
+};
+
+#define VBIOS_MC_REGISTER_ARRAY_SIZE 32
+#define VBIOS_MAX_AC_TIMING_ENTRIES 20
+
+struct atom_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[VBIOS_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct atom_mc_register_address {
+	u16 s1;
+	u8 pre_reg_data;
+};
+
+struct atom_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	struct atom_mc_reg_entry mc_reg_table_entry[VBIOS_MAX_AC_TIMING_ENTRIES];
+	struct atom_mc_register_address mc_reg_address[VBIOS_MC_REGISTER_ARRAY_SIZE];
+};
+
+#define MAX_VOLTAGE_ENTRIES 32
+
+struct atom_voltage_table_entry
+{
+	u16 value;
+	u32 smio_low;
+};
+
+struct atom_voltage_table
+{
+	u32 count;
+	u32 mask_low;
+	u32 phase_delay;
+	struct atom_voltage_table_entry entries[MAX_VOLTAGE_ENTRIES];
 };
 
 extern enum radeon_tv_std
 radeon_combios_get_tv_info(struct radeon_device *rdev);
 extern enum radeon_tv_std
 radeon_atombios_get_tv_info(struct radeon_device *rdev);
+extern void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
+						 u16 *vddc, u16 *vddci, u16 *mvdd);
 
 extern struct drm_connector *
 radeon_get_connector_for_encoder(struct drm_encoder *encoder);
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 1424ccd..0219d26 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -322,8 +322,8 @@ int radeon_bo_init(struct radeon_device *rdev)
 {
 	/* Add an MTRR for the VRAM */
 	if (!rdev->fastfb_working) {
-		rdev->mc.vram_mtrr = mtrr_add(rdev->mc.aper_base, rdev->mc.aper_size,
-			MTRR_TYPE_WRCOMB, 1);
+		rdev->mc.vram_mtrr = arch_phys_wc_add(rdev->mc.aper_base,
+						      rdev->mc.aper_size);
 	}
 	DRM_INFO("Detected VRAM RAM=%lluM, BAR=%lluM\n",
 		rdev->mc.mc_vram_size >> 20,
@@ -336,6 +336,7 @@ int radeon_bo_init(struct radeon_device *rdev)
 void radeon_bo_fini(struct radeon_device *rdev)
 {
 	radeon_ttm_fini(rdev);
+	arch_phys_wc_del(rdev->mc.vram_mtrr);
 }
 
 void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
@@ -348,14 +349,15 @@ void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 	}
 }
 
-int radeon_bo_list_validate(struct list_head *head, int ring)
+int radeon_bo_list_validate(struct ww_acquire_ctx *ticket,
+			    struct list_head *head, int ring)
 {
 	struct radeon_bo_list *lobj;
 	struct radeon_bo *bo;
 	u32 domain;
 	int r;
 
-	r = ttm_eu_reserve_buffers(head);
+	r = ttm_eu_reserve_buffers(ticket, head);
 	if (unlikely(r != 0)) {
 		return r;
 	}
@@ -398,7 +400,7 @@ int radeon_bo_get_surface_reg(struct radeon_bo *bo)
 	int steal;
 	int i;
 
-	BUG_ON(!radeon_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->tbo.resv->lock.base);
 
 	if (!bo->tiling_flags)
 		return 0;
@@ -524,7 +526,8 @@ void radeon_bo_get_tiling_flags(struct radeon_bo *bo,
 				uint32_t *tiling_flags,
 				uint32_t *pitch)
 {
-	BUG_ON(!radeon_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->tbo.resv->lock.base);
+
 	if (tiling_flags)
 		*tiling_flags = bo->tiling_flags;
 	if (pitch)
@@ -534,7 +537,8 @@ void radeon_bo_get_tiling_flags(struct radeon_bo *bo,
 int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved,
 				bool force_drop)
 {
-	BUG_ON(!radeon_bo_is_reserved(bo) && !force_drop);
+	if (!force_drop)
+		lockdep_assert_held(&bo->tbo.resv->lock.base);
 
 	if (!(bo->tiling_flags & RADEON_TILING_SURFACE))
 		return 0;
@@ -617,26 +621,3 @@ int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, bool no_wait)
 	ttm_bo_unreserve(&bo->tbo);
 	return r;
 }
-
-
-/**
- * radeon_bo_reserve - reserve bo
- * @bo:		bo structure
- * @no_intr:	don't return -ERESTARTSYS on pending signal
- *
- * Returns:
- * -ERESTARTSYS: A wait for the buffer to become unreserved was interrupted by
- * a signal. Release all buffer reservations and return to user-space.
- */
-int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr)
-{
-	int r;
-
-	r = ttm_bo_reserve(&bo->tbo, !no_intr, false, false, 0);
-	if (unlikely(r != 0)) {
-		if (r != -ERESTARTSYS)
-			dev_err(bo->rdev->dev, "%p reserve failed\n", bo);
-		return r;
-	}
-	return 0;
-}
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index e2cb80a..91519a5 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -52,7 +52,27 @@ static inline unsigned radeon_mem_type_to_domain(u32 mem_type)
 	return 0;
 }
 
-int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr);
+/**
+ * radeon_bo_reserve - reserve bo
+ * @bo:		bo structure
+ * @no_intr:	don't return -ERESTARTSYS on pending signal
+ *
+ * Returns:
+ * -ERESTARTSYS: A wait for the buffer to become unreserved was interrupted by
+ * a signal. Release all buffer reservations and return to user-space.
+ */
+static inline int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr)
+{
+	int r;
+
+	r = ttm_bo_reserve(&bo->tbo, !no_intr, false, false, 0);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			dev_err(bo->rdev->dev, "%p reserve failed\n", bo);
+		return r;
+	}
+	return 0;
+}
 
 static inline void radeon_bo_unreserve(struct radeon_bo *bo)
 {
@@ -78,11 +98,6 @@ static inline unsigned long radeon_bo_size(struct radeon_bo *bo)
 	return bo->tbo.num_pages << PAGE_SHIFT;
 }
 
-static inline bool radeon_bo_is_reserved(struct radeon_bo *bo)
-{
-	return ttm_bo_is_reserved(&bo->tbo);
-}
-
 static inline unsigned radeon_bo_ngpu_pages(struct radeon_bo *bo)
 {
 	return (bo->tbo.num_pages << PAGE_SHIFT) / RADEON_GPU_PAGE_SIZE;
@@ -128,7 +143,8 @@ extern int radeon_bo_init(struct radeon_device *rdev);
 extern void radeon_bo_fini(struct radeon_device *rdev);
 extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 				struct list_head *head);
-extern int radeon_bo_list_validate(struct list_head *head, int ring);
+extern int radeon_bo_list_validate(struct ww_acquire_ctx *ticket,
+				   struct list_head *head, int ring);
 extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo,
 				struct vm_area_struct *vma);
 extern int radeon_bo_set_tiling_flags(struct radeon_bo *bo,
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 788c64c..f374c46 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -388,7 +388,8 @@ static ssize_t radeon_get_pm_method(struct device *dev,
 	int pm = rdev->pm.pm_method;
 
 	return snprintf(buf, PAGE_SIZE, "%s\n",
-			(pm == PM_METHOD_DYNPM) ? "dynpm" : "profile");
+			(pm == PM_METHOD_DYNPM) ? "dynpm" :
+			(pm == PM_METHOD_PROFILE) ? "profile" : "dpm");
 }
 
 static ssize_t radeon_set_pm_method(struct device *dev,
@@ -399,6 +400,11 @@ static ssize_t radeon_set_pm_method(struct device *dev,
 	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
 	struct radeon_device *rdev = ddev->dev_private;
 
+	/* we don't support the legacy modes with dpm */
+	if (rdev->pm.pm_method == PM_METHOD_DPM) {
+		count = -EINVAL;
+		goto fail;
+	}
 
 	if (strncmp("dynpm", buf, strlen("dynpm")) == 0) {
 		mutex_lock(&rdev->pm.mutex);
@@ -423,8 +429,96 @@ fail:
 	return count;
 }
 
+static ssize_t radeon_get_dpm_state(struct device *dev,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+	enum radeon_pm_state_type pm = rdev->pm.dpm.user_state;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n",
+			(pm == POWER_STATE_TYPE_BATTERY) ? "battery" :
+			(pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance");
+}
+
+static ssize_t radeon_set_dpm_state(struct device *dev,
+				    struct device_attribute *attr,
+				    const char *buf,
+				    size_t count)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+
+	mutex_lock(&rdev->pm.mutex);
+	if (strncmp("battery", buf, strlen("battery")) == 0)
+		rdev->pm.dpm.user_state = POWER_STATE_TYPE_BATTERY;
+	else if (strncmp("balanced", buf, strlen("balanced")) == 0)
+		rdev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED;
+	else if (strncmp("performance", buf, strlen("performance")) == 0)
+		rdev->pm.dpm.user_state = POWER_STATE_TYPE_PERFORMANCE;
+	else {
+		mutex_unlock(&rdev->pm.mutex);
+		count = -EINVAL;
+		goto fail;
+	}
+	mutex_unlock(&rdev->pm.mutex);
+	radeon_pm_compute_clocks(rdev);
+fail:
+	return count;
+}
+
+static ssize_t radeon_get_dpm_forced_performance_level(struct device *dev,
+						       struct device_attribute *attr,
+						       char *buf)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+	enum radeon_dpm_forced_level level = rdev->pm.dpm.forced_level;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n",
+			(level == RADEON_DPM_FORCED_LEVEL_AUTO) ? "auto" :
+			(level == RADEON_DPM_FORCED_LEVEL_LOW) ? "low" : "high");
+}
+
+static ssize_t radeon_set_dpm_forced_performance_level(struct device *dev,
+						       struct device_attribute *attr,
+						       const char *buf,
+						       size_t count)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+	enum radeon_dpm_forced_level level;
+	int ret = 0;
+
+	mutex_lock(&rdev->pm.mutex);
+	if (strncmp("low", buf, strlen("low")) == 0) {
+		level = RADEON_DPM_FORCED_LEVEL_LOW;
+	} else if (strncmp("high", buf, strlen("high")) == 0) {
+		level = RADEON_DPM_FORCED_LEVEL_HIGH;
+	} else if (strncmp("auto", buf, strlen("auto")) == 0) {
+		level = RADEON_DPM_FORCED_LEVEL_AUTO;
+	} else {
+		mutex_unlock(&rdev->pm.mutex);
+		count = -EINVAL;
+		goto fail;
+	}
+	if (rdev->asic->dpm.force_performance_level) {
+		ret = radeon_dpm_force_performance_level(rdev, level);
+		if (ret)
+			count = -EINVAL;
+	}
+	mutex_unlock(&rdev->pm.mutex);
+fail:
+	return count;
+}
+
 static DEVICE_ATTR(power_profile, S_IRUGO | S_IWUSR, radeon_get_pm_profile, radeon_set_pm_profile);
 static DEVICE_ATTR(power_method, S_IRUGO | S_IWUSR, radeon_get_pm_method, radeon_set_pm_method);
+static DEVICE_ATTR(power_dpm_state, S_IRUGO | S_IWUSR, radeon_get_dpm_state, radeon_set_dpm_state);
+static DEVICE_ATTR(power_dpm_force_performance_level, S_IRUGO | S_IWUSR,
+		   radeon_get_dpm_forced_performance_level,
+		   radeon_set_dpm_forced_performance_level);
 
 static ssize_t radeon_hwmon_show_temp(struct device *dev,
 				      struct device_attribute *attr,
@@ -434,27 +528,10 @@ static ssize_t radeon_hwmon_show_temp(struct device *dev,
 	struct radeon_device *rdev = ddev->dev_private;
 	int temp;
 
-	switch (rdev->pm.int_thermal_type) {
-	case THERMAL_TYPE_RV6XX:
-		temp = rv6xx_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_RV770:
-		temp = rv770_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_EVERGREEN:
-	case THERMAL_TYPE_NI:
-		temp = evergreen_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_SUMO:
-		temp = sumo_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_SI:
-		temp = si_get_temp(rdev);
-		break;
-	default:
+	if (rdev->asic->pm.get_temperature)
+		temp = radeon_get_temperature(rdev);
+	else
 		temp = 0;
-		break;
-	}
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", temp);
 }
@@ -492,8 +569,7 @@ static int radeon_hwmon_init(struct radeon_device *rdev)
 	case THERMAL_TYPE_NI:
 	case THERMAL_TYPE_SUMO:
 	case THERMAL_TYPE_SI:
-		/* No support for TN yet */
-		if (rdev->family == CHIP_ARUBA)
+		if (rdev->asic->pm.get_temperature == NULL)
 			return err;
 		rdev->pm.int_hwmon_dev = hwmon_device_register(rdev->dev);
 		if (IS_ERR(rdev->pm.int_hwmon_dev)) {
@@ -526,7 +602,289 @@ static void radeon_hwmon_fini(struct radeon_device *rdev)
 	}
 }
 
-void radeon_pm_suspend(struct radeon_device *rdev)
+static void radeon_dpm_thermal_work_handler(struct work_struct *work)
+{
+	struct radeon_device *rdev =
+		container_of(work, struct radeon_device,
+			     pm.dpm.thermal.work);
+	/* switch to the thermal state */
+	enum radeon_pm_state_type dpm_state = POWER_STATE_TYPE_INTERNAL_THERMAL;
+
+	if (!rdev->pm.dpm_enabled)
+		return;
+
+	if (rdev->asic->pm.get_temperature) {
+		int temp = radeon_get_temperature(rdev);
+
+		if (temp < rdev->pm.dpm.thermal.min_temp)
+			/* switch back the user state */
+			dpm_state = rdev->pm.dpm.user_state;
+	} else {
+		if (rdev->pm.dpm.thermal.high_to_low)
+			/* switch back the user state */
+			dpm_state = rdev->pm.dpm.user_state;
+	}
+	radeon_dpm_enable_power_state(rdev, dpm_state);
+}
+
+static struct radeon_ps *radeon_dpm_pick_power_state(struct radeon_device *rdev,
+						     enum radeon_pm_state_type dpm_state)
+{
+	int i;
+	struct radeon_ps *ps;
+	u32 ui_class;
+	bool single_display = (rdev->pm.dpm.new_active_crtc_count < 2) ?
+		true : false;
+
+	/* check if the vblank period is too short to adjust the mclk */
+	if (single_display && rdev->asic->dpm.vblank_too_short) {
+		if (radeon_dpm_vblank_too_short(rdev))
+			single_display = false;
+	}
+
+	/* certain older asics have a separare 3D performance state,
+	 * so try that first if the user selected performance
+	 */
+	if (dpm_state == POWER_STATE_TYPE_PERFORMANCE)
+		dpm_state = POWER_STATE_TYPE_INTERNAL_3DPERF;
+	/* balanced states don't exist at the moment */
+	if (dpm_state == POWER_STATE_TYPE_BALANCED)
+		dpm_state = POWER_STATE_TYPE_PERFORMANCE;
+
+restart_search:
+	/* Pick the best power state based on current conditions */
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		ps = &rdev->pm.dpm.ps[i];
+		ui_class = ps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK;
+		switch (dpm_state) {
+		/* user states */
+		case POWER_STATE_TYPE_BATTERY:
+			if (ui_class == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) {
+				if (ps->caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY) {
+					if (single_display)
+						return ps;
+				} else
+					return ps;
+			}
+			break;
+		case POWER_STATE_TYPE_BALANCED:
+			if (ui_class == ATOM_PPLIB_CLASSIFICATION_UI_BALANCED) {
+				if (ps->caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY) {
+					if (single_display)
+						return ps;
+				} else
+					return ps;
+			}
+			break;
+		case POWER_STATE_TYPE_PERFORMANCE:
+			if (ui_class == ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+				if (ps->caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY) {
+					if (single_display)
+						return ps;
+				} else
+					return ps;
+			}
+			break;
+		/* internal states */
+		case POWER_STATE_TYPE_INTERNAL_UVD:
+			return rdev->pm.dpm.uvd_ps;
+		case POWER_STATE_TYPE_INTERNAL_UVD_SD:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_UVD_HD:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_HDSTATE)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_UVD_HD2:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_HD2STATE)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_UVD_MVC:
+			if (ps->class2 & ATOM_PPLIB_CLASSIFICATION2_MVC)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_BOOT:
+			return rdev->pm.dpm.boot_ps;
+		case POWER_STATE_TYPE_INTERNAL_THERMAL:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_ACPI:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_ACPI)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_ULV:
+			if (ps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_3DPERF:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_3DPERFORMANCE)
+				return ps;
+			break;
+		default:
+			break;
+		}
+	}
+	/* use a fallback state if we didn't match */
+	switch (dpm_state) {
+	case POWER_STATE_TYPE_INTERNAL_UVD_SD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD2:
+	case POWER_STATE_TYPE_INTERNAL_UVD_MVC:
+		return rdev->pm.dpm.uvd_ps;
+	case POWER_STATE_TYPE_INTERNAL_THERMAL:
+		dpm_state = POWER_STATE_TYPE_INTERNAL_ACPI;
+		goto restart_search;
+	case POWER_STATE_TYPE_INTERNAL_ACPI:
+		dpm_state = POWER_STATE_TYPE_BATTERY;
+		goto restart_search;
+	case POWER_STATE_TYPE_BATTERY:
+	case POWER_STATE_TYPE_BALANCED:
+	case POWER_STATE_TYPE_INTERNAL_3DPERF:
+		dpm_state = POWER_STATE_TYPE_PERFORMANCE;
+		goto restart_search;
+	default:
+		break;
+	}
+
+	return NULL;
+}
+
+static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev)
+{
+	int i;
+	struct radeon_ps *ps;
+	enum radeon_pm_state_type dpm_state;
+	int ret;
+
+	/* if dpm init failed */
+	if (!rdev->pm.dpm_enabled)
+		return;
+
+	if (rdev->pm.dpm.user_state != rdev->pm.dpm.state) {
+		/* add other state override checks here */
+		if ((!rdev->pm.dpm.thermal_active) &&
+		    (!rdev->pm.dpm.uvd_active))
+			rdev->pm.dpm.state = rdev->pm.dpm.user_state;
+	}
+	dpm_state = rdev->pm.dpm.state;
+
+	ps = radeon_dpm_pick_power_state(rdev, dpm_state);
+	if (ps)
+		rdev->pm.dpm.requested_ps = ps;
+	else
+		return;
+
+	/* no need to reprogram if nothing changed unless we are on BTC+ */
+	if (rdev->pm.dpm.current_ps == rdev->pm.dpm.requested_ps) {
+		if ((rdev->family < CHIP_BARTS) || (rdev->flags & RADEON_IS_IGP)) {
+			/* for pre-BTC and APUs if the num crtcs changed but state is the same,
+			 * all we need to do is update the display configuration.
+			 */
+			if (rdev->pm.dpm.new_active_crtcs != rdev->pm.dpm.current_active_crtcs) {
+				/* update display watermarks based on new power state */
+				radeon_bandwidth_update(rdev);
+				/* update displays */
+				radeon_dpm_display_configuration_changed(rdev);
+				rdev->pm.dpm.current_active_crtcs = rdev->pm.dpm.new_active_crtcs;
+				rdev->pm.dpm.current_active_crtc_count = rdev->pm.dpm.new_active_crtc_count;
+			}
+			return;
+		} else {
+			/* for BTC+ if the num crtcs hasn't changed and state is the same,
+			 * nothing to do, if the num crtcs is > 1 and state is the same,
+			 * update display configuration.
+			 */
+			if (rdev->pm.dpm.new_active_crtcs ==
+			    rdev->pm.dpm.current_active_crtcs) {
+				return;
+			} else {
+				if ((rdev->pm.dpm.current_active_crtc_count > 1) &&
+				    (rdev->pm.dpm.new_active_crtc_count > 1)) {
+					/* update display watermarks based on new power state */
+					radeon_bandwidth_update(rdev);
+					/* update displays */
+					radeon_dpm_display_configuration_changed(rdev);
+					rdev->pm.dpm.current_active_crtcs = rdev->pm.dpm.new_active_crtcs;
+					rdev->pm.dpm.current_active_crtc_count = rdev->pm.dpm.new_active_crtc_count;
+					return;
+				}
+			}
+		}
+	}
+
+	printk("switching from power state:\n");
+	radeon_dpm_print_power_state(rdev, rdev->pm.dpm.current_ps);
+	printk("switching to power state:\n");
+	radeon_dpm_print_power_state(rdev, rdev->pm.dpm.requested_ps);
+
+	mutex_lock(&rdev->ddev->struct_mutex);
+	down_write(&rdev->pm.mclk_lock);
+	mutex_lock(&rdev->ring_lock);
+
+	ret = radeon_dpm_pre_set_power_state(rdev);
+	if (ret)
+		goto done;
+
+	/* update display watermarks based on new power state */
+	radeon_bandwidth_update(rdev);
+	/* update displays */
+	radeon_dpm_display_configuration_changed(rdev);
+
+	rdev->pm.dpm.current_active_crtcs = rdev->pm.dpm.new_active_crtcs;
+	rdev->pm.dpm.current_active_crtc_count = rdev->pm.dpm.new_active_crtc_count;
+
+	/* wait for the rings to drain */
+	for (i = 0; i < RADEON_NUM_RINGS; i++) {
+		struct radeon_ring *ring = &rdev->ring[i];
+		if (ring->ready)
+			radeon_fence_wait_empty_locked(rdev, i);
+	}
+
+	/* program the new power state */
+	radeon_dpm_set_power_state(rdev);
+
+	/* update current power state */
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps;
+
+	radeon_dpm_post_set_power_state(rdev);
+
+done:
+	mutex_unlock(&rdev->ring_lock);
+	up_write(&rdev->pm.mclk_lock);
+	mutex_unlock(&rdev->ddev->struct_mutex);
+}
+
+void radeon_dpm_enable_power_state(struct radeon_device *rdev,
+				   enum radeon_pm_state_type dpm_state)
+{
+	if (!rdev->pm.dpm_enabled)
+		return;
+
+	mutex_lock(&rdev->pm.mutex);
+	switch (dpm_state) {
+	case POWER_STATE_TYPE_INTERNAL_THERMAL:
+		rdev->pm.dpm.thermal_active = true;
+		break;
+	case POWER_STATE_TYPE_INTERNAL_UVD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_SD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD2:
+	case POWER_STATE_TYPE_INTERNAL_UVD_MVC:
+		rdev->pm.dpm.uvd_active = true;
+		break;
+	default:
+		rdev->pm.dpm.thermal_active = false;
+		rdev->pm.dpm.uvd_active = false;
+		break;
+	}
+	rdev->pm.dpm.state = dpm_state;
+	mutex_unlock(&rdev->pm.mutex);
+	radeon_pm_compute_clocks(rdev);
+}
+
+static void radeon_pm_suspend_old(struct radeon_device *rdev)
 {
 	mutex_lock(&rdev->pm.mutex);
 	if (rdev->pm.pm_method == PM_METHOD_DYNPM) {
@@ -538,11 +896,30 @@ void radeon_pm_suspend(struct radeon_device *rdev)
 	cancel_delayed_work_sync(&rdev->pm.dynpm_idle_work);
 }
 
-void radeon_pm_resume(struct radeon_device *rdev)
+static void radeon_pm_suspend_dpm(struct radeon_device *rdev)
+{
+	mutex_lock(&rdev->pm.mutex);
+	/* disable dpm */
+	radeon_dpm_disable(rdev);
+	/* reset the power state */
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps;
+	rdev->pm.dpm_enabled = false;
+	mutex_unlock(&rdev->pm.mutex);
+}
+
+void radeon_pm_suspend(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_suspend_dpm(rdev);
+	else
+		radeon_pm_suspend_old(rdev);
+}
+
+static void radeon_pm_resume_old(struct radeon_device *rdev)
 {
 	/* set up the default clocks if the MC ucode is loaded */
 	if ((rdev->family >= CHIP_BARTS) &&
-	    (rdev->family <= CHIP_CAYMAN) &&
+	    (rdev->family <= CHIP_HAINAN) &&
 	    rdev->mc_fw) {
 		if (rdev->pm.default_vddc)
 			radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
@@ -573,12 +950,50 @@ void radeon_pm_resume(struct radeon_device *rdev)
 	radeon_pm_compute_clocks(rdev);
 }
 
-int radeon_pm_init(struct radeon_device *rdev)
+static void radeon_pm_resume_dpm(struct radeon_device *rdev)
+{
+	int ret;
+
+	/* asic init will reset to the boot state */
+	mutex_lock(&rdev->pm.mutex);
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps;
+	radeon_dpm_setup_asic(rdev);
+	ret = radeon_dpm_enable(rdev);
+	mutex_unlock(&rdev->pm.mutex);
+	if (ret) {
+		DRM_ERROR("radeon: dpm resume failed\n");
+		if ((rdev->family >= CHIP_BARTS) &&
+		    (rdev->family <= CHIP_HAINAN) &&
+		    rdev->mc_fw) {
+			if (rdev->pm.default_vddc)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
+							SET_VOLTAGE_TYPE_ASIC_VDDC);
+			if (rdev->pm.default_vddci)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddci,
+							SET_VOLTAGE_TYPE_ASIC_VDDCI);
+			if (rdev->pm.default_sclk)
+				radeon_set_engine_clock(rdev, rdev->pm.default_sclk);
+			if (rdev->pm.default_mclk)
+				radeon_set_memory_clock(rdev, rdev->pm.default_mclk);
+		}
+	} else {
+		rdev->pm.dpm_enabled = true;
+		radeon_pm_compute_clocks(rdev);
+	}
+}
+
+void radeon_pm_resume(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_resume_dpm(rdev);
+	else
+		radeon_pm_resume_old(rdev);
+}
+
+static int radeon_pm_init_old(struct radeon_device *rdev)
 {
 	int ret;
 
-	/* default to profile method */
-	rdev->pm.pm_method = PM_METHOD_PROFILE;
 	rdev->pm.profile = PM_PROFILE_DEFAULT;
 	rdev->pm.dynpm_state = DYNPM_STATE_DISABLED;
 	rdev->pm.dynpm_planned_action = DYNPM_ACTION_NONE;
@@ -599,7 +1014,7 @@ int radeon_pm_init(struct radeon_device *rdev)
 		radeon_pm_init_profile(rdev);
 		/* set up the default clocks if the MC ucode is loaded */
 		if ((rdev->family >= CHIP_BARTS) &&
-		    (rdev->family <= CHIP_CAYMAN) &&
+		    (rdev->family <= CHIP_HAINAN) &&
 		    rdev->mc_fw) {
 			if (rdev->pm.default_vddc)
 				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
@@ -640,7 +1055,145 @@ int radeon_pm_init(struct radeon_device *rdev)
 	return 0;
 }
 
-void radeon_pm_fini(struct radeon_device *rdev)
+static void radeon_dpm_print_power_states(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		printk("== power state %d ==\n", i);
+		radeon_dpm_print_power_state(rdev, &rdev->pm.dpm.ps[i]);
+	}
+}
+
+static int radeon_pm_init_dpm(struct radeon_device *rdev)
+{
+	int ret;
+
+	/* default to performance state */
+	rdev->pm.dpm.state = POWER_STATE_TYPE_BALANCED;
+	rdev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED;
+	rdev->pm.default_sclk = rdev->clock.default_sclk;
+	rdev->pm.default_mclk = rdev->clock.default_mclk;
+	rdev->pm.current_sclk = rdev->clock.default_sclk;
+	rdev->pm.current_mclk = rdev->clock.default_mclk;
+	rdev->pm.int_thermal_type = THERMAL_TYPE_NONE;
+
+	if (rdev->bios && rdev->is_atom_bios)
+		radeon_atombios_get_power_modes(rdev);
+	else
+		return -EINVAL;
+
+	/* set up the internal thermal sensor if applicable */
+	ret = radeon_hwmon_init(rdev);
+	if (ret)
+		return ret;
+
+	INIT_WORK(&rdev->pm.dpm.thermal.work, radeon_dpm_thermal_work_handler);
+	mutex_lock(&rdev->pm.mutex);
+	radeon_dpm_init(rdev);
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps;
+	radeon_dpm_print_power_states(rdev);
+	radeon_dpm_setup_asic(rdev);
+	ret = radeon_dpm_enable(rdev);
+	mutex_unlock(&rdev->pm.mutex);
+	if (ret) {
+		rdev->pm.dpm_enabled = false;
+		if ((rdev->family >= CHIP_BARTS) &&
+		    (rdev->family <= CHIP_HAINAN) &&
+		    rdev->mc_fw) {
+			if (rdev->pm.default_vddc)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
+							SET_VOLTAGE_TYPE_ASIC_VDDC);
+			if (rdev->pm.default_vddci)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddci,
+							SET_VOLTAGE_TYPE_ASIC_VDDCI);
+			if (rdev->pm.default_sclk)
+				radeon_set_engine_clock(rdev, rdev->pm.default_sclk);
+			if (rdev->pm.default_mclk)
+				radeon_set_memory_clock(rdev, rdev->pm.default_mclk);
+		}
+		DRM_ERROR("radeon: dpm initialization failed\n");
+		return ret;
+	}
+	rdev->pm.dpm_enabled = true;
+	radeon_pm_compute_clocks(rdev);
+
+	if (rdev->pm.num_power_states > 1) {
+		ret = device_create_file(rdev->dev, &dev_attr_power_dpm_state);
+		if (ret)
+			DRM_ERROR("failed to create device file for dpm state\n");
+		ret = device_create_file(rdev->dev, &dev_attr_power_dpm_force_performance_level);
+		if (ret)
+			DRM_ERROR("failed to create device file for dpm state\n");
+		/* XXX: these are noops for dpm but are here for backwards compat */
+		ret = device_create_file(rdev->dev, &dev_attr_power_profile);
+		if (ret)
+			DRM_ERROR("failed to create device file for power profile\n");
+		ret = device_create_file(rdev->dev, &dev_attr_power_method);
+		if (ret)
+			DRM_ERROR("failed to create device file for power method\n");
+
+		if (radeon_debugfs_pm_init(rdev)) {
+			DRM_ERROR("Failed to register debugfs file for dpm!\n");
+		}
+
+		DRM_INFO("radeon: dpm initialized\n");
+	}
+
+	return 0;
+}
+
+int radeon_pm_init(struct radeon_device *rdev)
+{
+	/* enable dpm on rv6xx+ */
+	switch (rdev->family) {
+	case CHIP_RV610:
+	case CHIP_RV630:
+	case CHIP_RV620:
+	case CHIP_RV635:
+	case CHIP_RV670:
+	case CHIP_RS780:
+	case CHIP_RS880:
+	case CHIP_RV770:
+	case CHIP_RV730:
+	case CHIP_RV710:
+	case CHIP_RV740:
+	case CHIP_CEDAR:
+	case CHIP_REDWOOD:
+	case CHIP_JUNIPER:
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+	case CHIP_PALM:
+	case CHIP_SUMO:
+	case CHIP_SUMO2:
+	case CHIP_BARTS:
+	case CHIP_TURKS:
+	case CHIP_CAICOS:
+	case CHIP_CAYMAN:
+	case CHIP_ARUBA:
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+	case CHIP_HAINAN:
+		if (radeon_dpm == 1)
+			rdev->pm.pm_method = PM_METHOD_DPM;
+		else
+			rdev->pm.pm_method = PM_METHOD_PROFILE;
+		break;
+	default:
+		/* default to profile method */
+		rdev->pm.pm_method = PM_METHOD_PROFILE;
+		break;
+	}
+
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		return radeon_pm_init_dpm(rdev);
+	else
+		return radeon_pm_init_old(rdev);
+}
+
+static void radeon_pm_fini_old(struct radeon_device *rdev)
 {
 	if (rdev->pm.num_power_states > 1) {
 		mutex_lock(&rdev->pm.mutex);
@@ -668,7 +1221,36 @@ void radeon_pm_fini(struct radeon_device *rdev)
 	radeon_hwmon_fini(rdev);
 }
 
-void radeon_pm_compute_clocks(struct radeon_device *rdev)
+static void radeon_pm_fini_dpm(struct radeon_device *rdev)
+{
+	if (rdev->pm.num_power_states > 1) {
+		mutex_lock(&rdev->pm.mutex);
+		radeon_dpm_disable(rdev);
+		mutex_unlock(&rdev->pm.mutex);
+
+		device_remove_file(rdev->dev, &dev_attr_power_dpm_state);
+		device_remove_file(rdev->dev, &dev_attr_power_dpm_force_performance_level);
+		/* XXX backwards compat */
+		device_remove_file(rdev->dev, &dev_attr_power_profile);
+		device_remove_file(rdev->dev, &dev_attr_power_method);
+	}
+	radeon_dpm_fini(rdev);
+
+	if (rdev->pm.power_state)
+		kfree(rdev->pm.power_state);
+
+	radeon_hwmon_fini(rdev);
+}
+
+void radeon_pm_fini(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_fini_dpm(rdev);
+	else
+		radeon_pm_fini_old(rdev);
+}
+
+static void radeon_pm_compute_clocks_old(struct radeon_device *rdev)
 {
 	struct drm_device *ddev = rdev->ddev;
 	struct drm_crtc *crtc;
@@ -739,6 +1321,46 @@ void radeon_pm_compute_clocks(struct radeon_device *rdev)
 	mutex_unlock(&rdev->pm.mutex);
 }
 
+static void radeon_pm_compute_clocks_dpm(struct radeon_device *rdev)
+{
+	struct drm_device *ddev = rdev->ddev;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+
+	mutex_lock(&rdev->pm.mutex);
+
+	/* update active crtc counts */
+	rdev->pm.dpm.new_active_crtcs = 0;
+	rdev->pm.dpm.new_active_crtc_count = 0;
+	list_for_each_entry(crtc,
+		&ddev->mode_config.crtc_list, head) {
+		radeon_crtc = to_radeon_crtc(crtc);
+		if (crtc->enabled) {
+			rdev->pm.dpm.new_active_crtcs |= (1 << radeon_crtc->crtc_id);
+			rdev->pm.dpm.new_active_crtc_count++;
+		}
+	}
+
+	/* update battery/ac status */
+	if (power_supply_is_system_supplied() > 0)
+		rdev->pm.dpm.ac_power = true;
+	else
+		rdev->pm.dpm.ac_power = false;
+
+	radeon_dpm_change_power_state_locked(rdev);
+
+	mutex_unlock(&rdev->pm.mutex);
+
+}
+
+void radeon_pm_compute_clocks(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_compute_clocks_dpm(rdev);
+	else
+		radeon_pm_compute_clocks_old(rdev);
+}
+
 static bool radeon_pm_in_vbl(struct radeon_device *rdev)
 {
 	int  crtc, vpos, hpos, vbl_status;
@@ -842,19 +1464,28 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	struct radeon_device *rdev = dev->dev_private;
 
-	seq_printf(m, "default engine clock: %u0 kHz\n", rdev->pm.default_sclk);
-	/* radeon_get_engine_clock is not reliable on APUs so just print the current clock */
-	if ((rdev->family >= CHIP_PALM) && (rdev->flags & RADEON_IS_IGP))
-		seq_printf(m, "current engine clock: %u0 kHz\n", rdev->pm.current_sclk);
-	else
-		seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
-	seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
-	if (rdev->asic->pm.get_memory_clock)
-		seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
-	if (rdev->pm.current_vddc)
-		seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc);
-	if (rdev->asic->pm.get_pcie_lanes)
-		seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev));
+	if (rdev->pm.dpm_enabled) {
+		mutex_lock(&rdev->pm.mutex);
+		if (rdev->asic->dpm.debugfs_print_current_performance_level)
+			radeon_dpm_debugfs_print_current_performance_level(rdev, m);
+		else
+			seq_printf(m, "Debugfs support not implemented for this asic\n");
+		mutex_unlock(&rdev->pm.mutex);
+	} else {
+		seq_printf(m, "default engine clock: %u0 kHz\n", rdev->pm.default_sclk);
+		/* radeon_get_engine_clock is not reliable on APUs so just print the current clock */
+		if ((rdev->family >= CHIP_PALM) && (rdev->flags & RADEON_IS_IGP))
+			seq_printf(m, "current engine clock: %u0 kHz\n", rdev->pm.current_sclk);
+		else
+			seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
+		seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
+		if (rdev->asic->pm.get_memory_clock)
+			seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
+		if (rdev->pm.current_vddc)
+			seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc);
+		if (rdev->asic->pm.get_pcie_lanes)
+			seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev));
+	}
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index 4940af7..65b9eab 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -88,11 +88,19 @@ int radeon_gem_prime_pin(struct drm_gem_object *obj)
 
 	/* pin buffer into GTT */
 	ret = radeon_bo_pin(bo, RADEON_GEM_DOMAIN_GTT, NULL);
-	if (ret) {
-		radeon_bo_unreserve(bo);
-		return ret;
-	}
 	radeon_bo_unreserve(bo);
+	return ret;
+}
+
+void radeon_gem_prime_unpin(struct drm_gem_object *obj)
+{
+	struct radeon_bo *bo = gem_to_radeon_bo(obj);
+	int ret = 0;
 
-	return 0;
+	ret = radeon_bo_reserve(bo, false);
+	if (unlikely(ret != 0))
+		return;
+
+	radeon_bo_unpin(bo);
+	radeon_bo_unreserve(bo);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index 7e2c2b7..62d5497 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -57,6 +57,7 @@
 #include "evergreen_reg.h"
 #include "ni_reg.h"
 #include "si_reg.h"
+#include "cik_reg.h"
 
 #define RADEON_MC_AGP_LOCATION		0x014c
 #define		RADEON_MC_AGP_START_MASK	0x0000FFFF
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index 82434018..5f1c51a 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -357,6 +357,38 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev,
 	}
 }
 
+u32 radeon_ring_generic_get_rptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring)
+{
+	u32 rptr;
+
+	if (rdev->wb.enabled && ring != &rdev->ring[R600_RING_TYPE_UVD_INDEX])
+		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
+	else
+		rptr = RREG32(ring->rptr_reg);
+	rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return rptr;
+}
+
+u32 radeon_ring_generic_get_wptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring)
+{
+	u32 wptr;
+
+	wptr = RREG32(ring->wptr_reg);
+	wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return wptr;
+}
+
+void radeon_ring_generic_set_wptr(struct radeon_device *rdev,
+				  struct radeon_ring *ring)
+{
+	WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask);
+	(void)RREG32(ring->wptr_reg);
+}
+
 /**
  * radeon_ring_free_size - update the free size
  *
@@ -367,13 +399,7 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev,
  */
 void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	u32 rptr;
-
-	if (rdev->wb.enabled && ring != &rdev->ring[R600_RING_TYPE_UVD_INDEX])
-		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
-	else
-		rptr = RREG32(ring->rptr_reg);
-	ring->rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+	ring->rptr = radeon_ring_get_rptr(rdev, ring);
 	/* This works because ring_size is a power of 2 */
 	ring->ring_free_dw = (ring->rptr + (ring->ring_size / 4));
 	ring->ring_free_dw -= ring->wptr;
@@ -465,8 +491,7 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
 		radeon_ring_write(ring, ring->nop);
 	}
 	DRM_MEMORYBARRIER();
-	WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask);
-	(void)RREG32(ring->wptr_reg);
+	radeon_ring_set_wptr(rdev, ring);
 }
 
 /**
@@ -568,7 +593,6 @@ void radeon_ring_lockup_update(struct radeon_ring *ring)
 bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	unsigned long cjiffies, elapsed;
-	uint32_t rptr;
 
 	cjiffies = jiffies;
 	if (!time_after(cjiffies, ring->last_activity)) {
@@ -576,8 +600,7 @@ bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 		radeon_ring_lockup_update(ring);
 		return false;
 	}
-	rptr = RREG32(ring->rptr_reg);
-	ring->rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+	ring->rptr = radeon_ring_get_rptr(rdev, ring);
 	if (ring->rptr != ring->last_rptr) {
 		/* CP is still working no lockup */
 		radeon_ring_lockup_update(ring);
@@ -804,9 +827,9 @@ static int radeon_debugfs_ring_info(struct seq_file *m, void *data)
 
 	radeon_ring_free_size(rdev, ring);
 	count = (ring->ring_size / 4) - ring->ring_free_dw;
-	tmp = RREG32(ring->wptr_reg) >> ring->ptr_reg_shift;
+	tmp = radeon_ring_get_wptr(rdev, ring);
 	seq_printf(m, "wptr(0x%04x): 0x%08x [%5d]\n", ring->wptr_reg, tmp, tmp);
-	tmp = RREG32(ring->rptr_reg) >> ring->ptr_reg_shift;
+	tmp = radeon_ring_get_rptr(rdev, ring);
 	seq_printf(m, "rptr(0x%04x): 0x%08x [%5d]\n", ring->rptr_reg, tmp, tmp);
 	if (ring->rptr_save_reg) {
 		seq_printf(m, "rptr next(0x%04x): 0x%08x\n", ring->rptr_save_reg,
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index bbed4af..f4d6bce 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -35,7 +35,6 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 {
 	struct radeon_bo *vram_obj = NULL;
 	struct radeon_bo **gtt_obj = NULL;
-	struct radeon_fence *fence = NULL;
 	uint64_t gtt_addr, vram_addr;
 	unsigned i, n, size;
 	int r, ring;
@@ -81,37 +80,38 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	}
 	r = radeon_bo_reserve(vram_obj, false);
 	if (unlikely(r != 0))
-		goto out_cleanup;
+		goto out_unref;
 	r = radeon_bo_pin(vram_obj, RADEON_GEM_DOMAIN_VRAM, &vram_addr);
 	if (r) {
 		DRM_ERROR("Failed to pin VRAM object\n");
-		goto out_cleanup;
+		goto out_unres;
 	}
 	for (i = 0; i < n; i++) {
 		void *gtt_map, *vram_map;
 		void **gtt_start, **gtt_end;
 		void **vram_start, **vram_end;
+		struct radeon_fence *fence = NULL;
 
 		r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
 				     RADEON_GEM_DOMAIN_GTT, NULL, gtt_obj + i);
 		if (r) {
 			DRM_ERROR("Failed to create GTT object %d\n", i);
-			goto out_cleanup;
+			goto out_lclean;
 		}
 
 		r = radeon_bo_reserve(gtt_obj[i], false);
 		if (unlikely(r != 0))
-			goto out_cleanup;
+			goto out_lclean_unref;
 		r = radeon_bo_pin(gtt_obj[i], RADEON_GEM_DOMAIN_GTT, &gtt_addr);
 		if (r) {
 			DRM_ERROR("Failed to pin GTT object %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unres;
 		}
 
 		r = radeon_bo_kmap(gtt_obj[i], &gtt_map);
 		if (r) {
 			DRM_ERROR("Failed to map GTT object %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		for (gtt_start = gtt_map, gtt_end = gtt_map + size;
@@ -127,13 +127,13 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 			r = radeon_copy_blit(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
 		if (r) {
 			DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		r = radeon_fence_wait(fence, false);
 		if (r) {
 			DRM_ERROR("Failed to wait for GTT->VRAM fence %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		radeon_fence_unref(&fence);
@@ -141,7 +141,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		r = radeon_bo_kmap(vram_obj, &vram_map);
 		if (r) {
 			DRM_ERROR("Failed to map VRAM object after copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		for (gtt_start = gtt_map, gtt_end = gtt_map + size,
@@ -160,7 +160,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 					  (vram_addr - rdev->mc.vram_start +
 					   (void*)gtt_start - gtt_map));
 				radeon_bo_kunmap(vram_obj);
-				goto out_cleanup;
+				goto out_lclean_unpin;
 			}
 			*vram_start = vram_start;
 		}
@@ -173,13 +173,13 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 			r = radeon_copy_blit(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
 		if (r) {
 			DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		r = radeon_fence_wait(fence, false);
 		if (r) {
 			DRM_ERROR("Failed to wait for VRAM->GTT fence %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		radeon_fence_unref(&fence);
@@ -187,7 +187,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		r = radeon_bo_kmap(gtt_obj[i], &gtt_map);
 		if (r) {
 			DRM_ERROR("Failed to map GTT object after copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		for (gtt_start = gtt_map, gtt_end = gtt_map + size,
@@ -206,7 +206,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 					  (gtt_addr - rdev->mc.gtt_start +
 					   (void*)vram_start - vram_map));
 				radeon_bo_kunmap(gtt_obj[i]);
-				goto out_cleanup;
+				goto out_lclean_unpin;
 			}
 		}
 
@@ -214,31 +214,32 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 
 		DRM_INFO("Tested GTT->VRAM and VRAM->GTT copy for GTT offset 0x%llx\n",
 			 gtt_addr - rdev->mc.gtt_start);
+		continue;
+
+out_lclean_unpin:
+		radeon_bo_unpin(gtt_obj[i]);
+out_lclean_unres:
+		radeon_bo_unreserve(gtt_obj[i]);
+out_lclean_unref:
+		radeon_bo_unref(&gtt_obj[i]);
+out_lclean:
+		for (--i; i >= 0; --i) {
+			radeon_bo_unpin(gtt_obj[i]);
+			radeon_bo_unreserve(gtt_obj[i]);
+			radeon_bo_unref(&gtt_obj[i]);
+		}
+		if (fence)
+			radeon_fence_unref(&fence);
+		break;
 	}
 
+	radeon_bo_unpin(vram_obj);
+out_unres:
+	radeon_bo_unreserve(vram_obj);
+out_unref:
+	radeon_bo_unref(&vram_obj);
 out_cleanup:
-	if (vram_obj) {
-		if (radeon_bo_is_reserved(vram_obj)) {
-			radeon_bo_unpin(vram_obj);
-			radeon_bo_unreserve(vram_obj);
-		}
-		radeon_bo_unref(&vram_obj);
-	}
-	if (gtt_obj) {
-		for (i = 0; i < n; i++) {
-			if (gtt_obj[i]) {
-				if (radeon_bo_is_reserved(gtt_obj[i])) {
-					radeon_bo_unpin(gtt_obj[i]);
-					radeon_bo_unreserve(gtt_obj[i]);
-				}
-				radeon_bo_unref(&gtt_obj[i]);
-			}
-		}
-		kfree(gtt_obj);
-	}
-	if (fence) {
-		radeon_fence_unref(&fence);
-	}
+	kfree(gtt_obj);
 	if (r) {
 		printk(KERN_WARNING "Error while testing BO move.\n");
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_ucode.h b/drivers/gpu/drm/radeon/radeon_ucode.h
new file mode 100644
index 0000000..d8b05f7
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_ucode.h
@@ -0,0 +1,129 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RADEON_UCODE_H__
+#define __RADEON_UCODE_H__
+
+/* CP */
+#define R600_PFP_UCODE_SIZE          576
+#define R600_PM4_UCODE_SIZE          1792
+#define R700_PFP_UCODE_SIZE          848
+#define R700_PM4_UCODE_SIZE          1360
+#define EVERGREEN_PFP_UCODE_SIZE     1120
+#define EVERGREEN_PM4_UCODE_SIZE     1376
+#define CAYMAN_PFP_UCODE_SIZE        2176
+#define CAYMAN_PM4_UCODE_SIZE        2176
+#define SI_PFP_UCODE_SIZE            2144
+#define SI_PM4_UCODE_SIZE            2144
+#define SI_CE_UCODE_SIZE             2144
+
+/* RLC */
+#define R600_RLC_UCODE_SIZE          768
+#define R700_RLC_UCODE_SIZE          1024
+#define EVERGREEN_RLC_UCODE_SIZE     768
+#define CAYMAN_RLC_UCODE_SIZE        1024
+#define ARUBA_RLC_UCODE_SIZE         1536
+#define SI_RLC_UCODE_SIZE            2048
+
+/* MC */
+#define BTC_MC_UCODE_SIZE            6024
+#define CAYMAN_MC_UCODE_SIZE         6037
+#define SI_MC_UCODE_SIZE             7769
+#define OLAND_MC_UCODE_SIZE          7863
+
+/* SMC */
+#define RV770_SMC_UCODE_START        0x0100
+#define RV770_SMC_UCODE_SIZE         0x410d
+#define RV770_SMC_INT_VECTOR_START   0xffc0
+#define RV770_SMC_INT_VECTOR_SIZE    0x0040
+
+#define RV730_SMC_UCODE_START        0x0100
+#define RV730_SMC_UCODE_SIZE         0x412c
+#define RV730_SMC_INT_VECTOR_START   0xffc0
+#define RV730_SMC_INT_VECTOR_SIZE    0x0040
+
+#define RV710_SMC_UCODE_START        0x0100
+#define RV710_SMC_UCODE_SIZE         0x3f1f
+#define RV710_SMC_INT_VECTOR_START   0xffc0
+#define RV710_SMC_INT_VECTOR_SIZE    0x0040
+
+#define RV740_SMC_UCODE_START        0x0100
+#define RV740_SMC_UCODE_SIZE         0x41c5
+#define RV740_SMC_INT_VECTOR_START   0xffc0
+#define RV740_SMC_INT_VECTOR_SIZE    0x0040
+
+#define CEDAR_SMC_UCODE_START        0x0100
+#define CEDAR_SMC_UCODE_SIZE         0x5d50
+#define CEDAR_SMC_INT_VECTOR_START   0xffc0
+#define CEDAR_SMC_INT_VECTOR_SIZE    0x0040
+
+#define REDWOOD_SMC_UCODE_START      0x0100
+#define REDWOOD_SMC_UCODE_SIZE       0x5f0a
+#define REDWOOD_SMC_INT_VECTOR_START 0xffc0
+#define REDWOOD_SMC_INT_VECTOR_SIZE  0x0040
+
+#define JUNIPER_SMC_UCODE_START      0x0100
+#define JUNIPER_SMC_UCODE_SIZE       0x5f1f
+#define JUNIPER_SMC_INT_VECTOR_START 0xffc0
+#define JUNIPER_SMC_INT_VECTOR_SIZE  0x0040
+
+#define CYPRESS_SMC_UCODE_START      0x0100
+#define CYPRESS_SMC_UCODE_SIZE       0x61f7
+#define CYPRESS_SMC_INT_VECTOR_START 0xffc0
+#define CYPRESS_SMC_INT_VECTOR_SIZE  0x0040
+
+#define BARTS_SMC_UCODE_START        0x0100
+#define BARTS_SMC_UCODE_SIZE         0x6107
+#define BARTS_SMC_INT_VECTOR_START   0xffc0
+#define BARTS_SMC_INT_VECTOR_SIZE    0x0040
+
+#define TURKS_SMC_UCODE_START        0x0100
+#define TURKS_SMC_UCODE_SIZE         0x605b
+#define TURKS_SMC_INT_VECTOR_START   0xffc0
+#define TURKS_SMC_INT_VECTOR_SIZE    0x0040
+
+#define CAICOS_SMC_UCODE_START       0x0100
+#define CAICOS_SMC_UCODE_SIZE        0x5fbd
+#define CAICOS_SMC_INT_VECTOR_START  0xffc0
+#define CAICOS_SMC_INT_VECTOR_SIZE   0x0040
+
+#define CAYMAN_SMC_UCODE_START       0x0100
+#define CAYMAN_SMC_UCODE_SIZE        0x79ec
+#define CAYMAN_SMC_INT_VECTOR_START  0xffc0
+#define CAYMAN_SMC_INT_VECTOR_SIZE   0x0040
+
+#define TAHITI_SMC_UCODE_START       0x10000
+#define TAHITI_SMC_UCODE_SIZE        0xf458
+
+#define PITCAIRN_SMC_UCODE_START     0x10000
+#define PITCAIRN_SMC_UCODE_SIZE      0xe9f4
+
+#define VERDE_SMC_UCODE_START        0x10000
+#define VERDE_SMC_UCODE_SIZE         0xebe4
+
+#define OLAND_SMC_UCODE_START        0x10000
+#define OLAND_SMC_UCODE_SIZE         0xe7b4
+
+#define HAINAN_SMC_UCODE_START       0x10000
+#define HAINAN_SMC_UCODE_SIZE        0xe67C
+
+#endif
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index cad735d..41efcec 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -44,11 +44,13 @@
 #define FIRMWARE_CYPRESS	"radeon/CYPRESS_uvd.bin"
 #define FIRMWARE_SUMO		"radeon/SUMO_uvd.bin"
 #define FIRMWARE_TAHITI		"radeon/TAHITI_uvd.bin"
+#define FIRMWARE_BONAIRE	"radeon/BONAIRE_uvd.bin"
 
 MODULE_FIRMWARE(FIRMWARE_RV710);
 MODULE_FIRMWARE(FIRMWARE_CYPRESS);
 MODULE_FIRMWARE(FIRMWARE_SUMO);
 MODULE_FIRMWARE(FIRMWARE_TAHITI);
+MODULE_FIRMWARE(FIRMWARE_BONAIRE);
 
 static void radeon_uvd_idle_work_handler(struct work_struct *work);
 
@@ -100,6 +102,12 @@ int radeon_uvd_init(struct radeon_device *rdev)
 		fw_name = FIRMWARE_TAHITI;
 		break;
 
+	case CHIP_BONAIRE:
+	case CHIP_KABINI:
+	case CHIP_KAVERI:
+		fw_name = FIRMWARE_BONAIRE;
+		break;
+
 	default:
 		return -EINVAL;
 	}
@@ -542,6 +550,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 			       struct radeon_fence **fence)
 {
 	struct ttm_validate_buffer tv;
+	struct ww_acquire_ctx ticket;
 	struct list_head head;
 	struct radeon_ib ib;
 	uint64_t addr;
@@ -553,7 +562,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	INIT_LIST_HEAD(&head);
 	list_add(&tv.head, &head);
 
-	r = ttm_eu_reserve_buffers(&head);
+	r = ttm_eu_reserve_buffers(&ticket, &head);
 	if (r)
 		return r;
 
@@ -561,16 +570,12 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	radeon_uvd_force_into_uvd_segment(bo);
 
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
-	if (r) {
-		ttm_eu_backoff_reservation(&head);
-		return r;
-	}
+	if (r) 
+		goto err;
 
 	r = radeon_ib_get(rdev, ring, &ib, NULL, 16);
-	if (r) {
-		ttm_eu_backoff_reservation(&head);
-		return r;
-	}
+	if (r)
+		goto err;
 
 	addr = radeon_bo_gpu_offset(bo);
 	ib.ptr[0] = PACKET0(UVD_GPCOM_VCPU_DATA0, 0);
@@ -584,11 +589,9 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	ib.length_dw = 16;
 
 	r = radeon_ib_schedule(rdev, &ib, NULL);
-	if (r) {
-		ttm_eu_backoff_reservation(&head);
-		return r;
-	}
-	ttm_eu_fence_buffer_objects(&head, ib.fence);
+	if (r)
+		goto err;
+	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
 
 	if (fence)
 		*fence = radeon_fence_ref(ib.fence);
@@ -596,6 +599,10 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	radeon_ib_free(rdev, &ib);
 	radeon_bo_unref(&bo);
 	return 0;
+
+err:
+	ttm_eu_backoff_reservation(&ticket, &head);
+	return r;
 }
 
 /* multiple fence commands without any stream commands in between can
@@ -691,11 +698,19 @@ static void radeon_uvd_idle_work_handler(struct work_struct *work)
 	struct radeon_device *rdev =
 		container_of(work, struct radeon_device, uvd.idle_work.work);
 
-	if (radeon_fence_count_emitted(rdev, R600_RING_TYPE_UVD_INDEX) == 0)
-		radeon_set_uvd_clocks(rdev, 0, 0);
-	else
+	if (radeon_fence_count_emitted(rdev, R600_RING_TYPE_UVD_INDEX) == 0) {
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			mutex_lock(&rdev->pm.mutex);
+			rdev->pm.dpm.uvd_active = false;
+			mutex_unlock(&rdev->pm.mutex);
+			radeon_pm_compute_clocks(rdev);
+		} else {
+			radeon_set_uvd_clocks(rdev, 0, 0);
+		}
+	} else {
 		schedule_delayed_work(&rdev->uvd.idle_work,
 				      msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS));
+	}
 }
 
 void radeon_uvd_note_usage(struct radeon_device *rdev)
@@ -703,8 +718,14 @@ void radeon_uvd_note_usage(struct radeon_device *rdev)
 	bool set_clocks = !cancel_delayed_work_sync(&rdev->uvd.idle_work);
 	set_clocks &= schedule_delayed_work(&rdev->uvd.idle_work,
 					    msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS));
-	if (set_clocks)
-		radeon_set_uvd_clocks(rdev, 53300, 40000);
+	if (set_clocks) {
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			/* XXX pick SD/HD/MVC */
+			radeon_dpm_enable_power_state(rdev, POWER_STATE_TYPE_INTERNAL_UVD);
+		} else {
+			radeon_set_uvd_clocks(rdev, 53300, 40000);
+		}
+	}
 }
 
 static unsigned radeon_uvd_calc_upll_post_div(unsigned vco_freq,
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index 55880d5..d8ddfb3 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -248,13 +248,16 @@ struct rs690_watermark {
 };
 
 static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
-				  struct radeon_crtc *crtc,
-				  struct rs690_watermark *wm)
+					 struct radeon_crtc *crtc,
+					 struct rs690_watermark *wm,
+					 bool low)
 {
 	struct drm_display_mode *mode = &crtc->base.mode;
 	fixed20_12 a, b, c;
 	fixed20_12 pclk, request_fifo_depth, tolerable_latency, estimated_width;
 	fixed20_12 consumption_time, line_time, chunk_time, read_delay_latency;
+	fixed20_12 sclk, core_bandwidth, max_bandwidth;
+	u32 selected_sclk;
 
 	if (!crtc->base.enabled) {
 		/* FIXME: wouldn't it better to set priority mark to maximum */
@@ -262,6 +265,21 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
 		return;
 	}
 
+	if (((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880)) &&
+	    (rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
+		selected_sclk = radeon_dpm_get_sclk(rdev, low);
+	else
+		selected_sclk = rdev->pm.current_sclk;
+
+	/* sclk in Mhz */
+	a.full = dfixed_const(100);
+	sclk.full = dfixed_const(selected_sclk);
+	sclk.full = dfixed_div(sclk, a);
+
+	/* core_bandwidth = sclk(Mhz) * 16 */
+	a.full = dfixed_const(16);
+	core_bandwidth.full = dfixed_div(rdev->pm.sclk, a);
+
 	if (crtc->vsc.full > dfixed_const(2))
 		wm->num_line_pair.full = dfixed_const(2);
 	else
@@ -322,36 +340,36 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
 	wm->active_time.full = dfixed_div(wm->active_time, a);
 
 	/* Maximun bandwidth is the minimun bandwidth of all component */
-	rdev->pm.max_bandwidth = rdev->pm.core_bandwidth;
+	max_bandwidth = core_bandwidth;
 	if (rdev->mc.igp_sideport_enabled) {
-		if (rdev->pm.max_bandwidth.full > rdev->pm.sideport_bandwidth.full &&
+		if (max_bandwidth.full > rdev->pm.sideport_bandwidth.full &&
 			rdev->pm.sideport_bandwidth.full)
-			rdev->pm.max_bandwidth = rdev->pm.sideport_bandwidth;
+			max_bandwidth = rdev->pm.sideport_bandwidth;
 		read_delay_latency.full = dfixed_const(370 * 800 * 1000);
 		read_delay_latency.full = dfixed_div(read_delay_latency,
 			rdev->pm.igp_sideport_mclk);
 	} else {
-		if (rdev->pm.max_bandwidth.full > rdev->pm.k8_bandwidth.full &&
+		if (max_bandwidth.full > rdev->pm.k8_bandwidth.full &&
 			rdev->pm.k8_bandwidth.full)
-			rdev->pm.max_bandwidth = rdev->pm.k8_bandwidth;
-		if (rdev->pm.max_bandwidth.full > rdev->pm.ht_bandwidth.full &&
+			max_bandwidth = rdev->pm.k8_bandwidth;
+		if (max_bandwidth.full > rdev->pm.ht_bandwidth.full &&
 			rdev->pm.ht_bandwidth.full)
-			rdev->pm.max_bandwidth = rdev->pm.ht_bandwidth;
+			max_bandwidth = rdev->pm.ht_bandwidth;
 		read_delay_latency.full = dfixed_const(5000);
 	}
 
 	/* sclk = system clocks(ns) = 1000 / max_bandwidth / 16 */
 	a.full = dfixed_const(16);
-	rdev->pm.sclk.full = dfixed_mul(rdev->pm.max_bandwidth, a);
+	sclk.full = dfixed_mul(max_bandwidth, a);
 	a.full = dfixed_const(1000);
-	rdev->pm.sclk.full = dfixed_div(a, rdev->pm.sclk);
+	sclk.full = dfixed_div(a, sclk);
 	/* Determine chunk time
 	 * ChunkTime = the time it takes the DCP to send one chunk of data
 	 * to the LB which consists of pipeline delay and inter chunk gap
 	 * sclk = system clock(ns)
 	 */
 	a.full = dfixed_const(256 * 13);
-	chunk_time.full = dfixed_mul(rdev->pm.sclk, a);
+	chunk_time.full = dfixed_mul(sclk, a);
 	a.full = dfixed_const(10);
 	chunk_time.full = dfixed_div(chunk_time, a);
 
@@ -415,175 +433,200 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
 	}
 }
 
-void rs690_bandwidth_update(struct radeon_device *rdev)
+static void rs690_compute_mode_priority(struct radeon_device *rdev,
+					struct rs690_watermark *wm0,
+					struct rs690_watermark *wm1,
+					struct drm_display_mode *mode0,
+					struct drm_display_mode *mode1,
+					u32 *d1mode_priority_a_cnt,
+					u32 *d2mode_priority_a_cnt)
 {
-	struct drm_display_mode *mode0 = NULL;
-	struct drm_display_mode *mode1 = NULL;
-	struct rs690_watermark wm0;
-	struct rs690_watermark wm1;
-	u32 tmp;
-	u32 d1mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
-	u32 d2mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
 	fixed20_12 priority_mark02, priority_mark12, fill_rate;
 	fixed20_12 a, b;
 
-	radeon_update_display_priority(rdev);
-
-	if (rdev->mode_info.crtcs[0]->base.enabled)
-		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
-	if (rdev->mode_info.crtcs[1]->base.enabled)
-		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
-	/*
-	 * Set display0/1 priority up in the memory controller for
-	 * modes if the user specifies HIGH for displaypriority
-	 * option.
-	 */
-	if ((rdev->disp_priority == 2) &&
-	    ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))) {
-		tmp = RREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER);
-		tmp &= C_000104_MC_DISP0R_INIT_LAT;
-		tmp &= C_000104_MC_DISP1R_INIT_LAT;
-		if (mode0)
-			tmp |= S_000104_MC_DISP0R_INIT_LAT(1);
-		if (mode1)
-			tmp |= S_000104_MC_DISP1R_INIT_LAT(1);
-		WREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER, tmp);
-	}
-	rs690_line_buffer_adjust(rdev, mode0, mode1);
-
-	if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))
-		WREG32(R_006C9C_DCP_CONTROL, 0);
-	if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
-		WREG32(R_006C9C_DCP_CONTROL, 2);
-
-	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0);
-	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1);
-
-	tmp = (wm0.lb_request_fifo_depth - 1);
-	tmp |= (wm1.lb_request_fifo_depth - 1) << 16;
-	WREG32(R_006D58_LB_MAX_REQ_OUTSTANDING, tmp);
+	*d1mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
+	*d2mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
 
 	if (mode0 && mode1) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_mul(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_mul(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			b.full = dfixed_mul(wm1.dbpp, wm1.num_line_pair);
+			a.full = wm0->num_line_pair.full;
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			b.full = dfixed_mul(wm1->dbpp, wm1->num_line_pair);
 		else
-			b.full = wm1.num_line_pair.full;
+			b.full = wm1->num_line_pair.full;
 		a.full += b.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2) {
-			d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
-			d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
 		}
 	} else if (mode0) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_mul(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_mul(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = wm0->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
 		if (rdev->disp_priority == 2)
-			d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
 	} else if (mode1) {
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			a.full = dfixed_mul(wm1.dbpp, wm1.num_line_pair);
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			a.full = dfixed_mul(wm1->dbpp, wm1->num_line_pair);
 		else
-			a.full = wm1.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm1.sclk, a);
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = wm1->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm1->sclk, a);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2)
-			d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
 	}
+}
+
+void rs690_bandwidth_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode0 = NULL;
+	struct drm_display_mode *mode1 = NULL;
+	struct rs690_watermark wm0_high, wm0_low;
+	struct rs690_watermark wm1_high, wm1_low;
+	u32 tmp;
+	u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt;
+	u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt;
+
+	radeon_update_display_priority(rdev);
+
+	if (rdev->mode_info.crtcs[0]->base.enabled)
+		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
+	if (rdev->mode_info.crtcs[1]->base.enabled)
+		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
+	/*
+	 * Set display0/1 priority up in the memory controller for
+	 * modes if the user specifies HIGH for displaypriority
+	 * option.
+	 */
+	if ((rdev->disp_priority == 2) &&
+	    ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))) {
+		tmp = RREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER);
+		tmp &= C_000104_MC_DISP0R_INIT_LAT;
+		tmp &= C_000104_MC_DISP1R_INIT_LAT;
+		if (mode0)
+			tmp |= S_000104_MC_DISP0R_INIT_LAT(1);
+		if (mode1)
+			tmp |= S_000104_MC_DISP1R_INIT_LAT(1);
+		WREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER, tmp);
+	}
+	rs690_line_buffer_adjust(rdev, mode0, mode1);
+
+	if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))
+		WREG32(R_006C9C_DCP_CONTROL, 0);
+	if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+		WREG32(R_006C9C_DCP_CONTROL, 2);
+
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_high, false);
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_high, false);
+
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_low, true);
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_low, true);
+
+	tmp = (wm0_high.lb_request_fifo_depth - 1);
+	tmp |= (wm1_high.lb_request_fifo_depth - 1) << 16;
+	WREG32(R_006D58_LB_MAX_REQ_OUTSTANDING, tmp);
+
+	rs690_compute_mode_priority(rdev,
+				    &wm0_high, &wm1_high,
+				    mode0, mode1,
+				    &d1mode_priority_a_cnt, &d2mode_priority_a_cnt);
+	rs690_compute_mode_priority(rdev,
+				    &wm0_low, &wm1_low,
+				    mode0, mode1,
+				    &d1mode_priority_b_cnt, &d2mode_priority_b_cnt);
 
 	WREG32(R_006548_D1MODE_PRIORITY_A_CNT, d1mode_priority_a_cnt);
-	WREG32(R_00654C_D1MODE_PRIORITY_B_CNT, d1mode_priority_a_cnt);
+	WREG32(R_00654C_D1MODE_PRIORITY_B_CNT, d1mode_priority_b_cnt);
 	WREG32(R_006D48_D2MODE_PRIORITY_A_CNT, d2mode_priority_a_cnt);
-	WREG32(R_006D4C_D2MODE_PRIORITY_B_CNT, d2mode_priority_a_cnt);
+	WREG32(R_006D4C_D2MODE_PRIORITY_B_CNT, d2mode_priority_b_cnt);
 }
 
 uint32_t rs690_mc_rreg(struct radeon_device *rdev, uint32_t reg)
diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c b/drivers/gpu/drm/radeon/rs780_dpm.c
new file mode 100644
index 0000000..bef832a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rs780_dpm.c
@@ -0,0 +1,963 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rs780d.h"
+#include "r600_dpm.h"
+#include "rs780_dpm.h"
+#include "atom.h"
+
+static struct igp_ps *rs780_get_ps(struct radeon_ps *rps)
+{
+	struct igp_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+static struct igp_power_info *rs780_get_pi(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void rs780_get_pm_mode_parameters(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct radeon_mode_info *minfo = &rdev->mode_info;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+	int i;
+
+	/* defaults */
+	pi->crtc_id = 0;
+	pi->refresh_rate = 60;
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		crtc = (struct drm_crtc *)minfo->crtcs[i];
+		if (crtc && crtc->enabled) {
+			radeon_crtc = to_radeon_crtc(crtc);
+			pi->crtc_id = radeon_crtc->crtc_id;
+			if (crtc->mode.htotal && crtc->mode.vtotal)
+				pi->refresh_rate =
+					(crtc->mode.clock * 1000) /
+					(crtc->mode.htotal * crtc->mode.vtotal);
+			break;
+		}
+	}
+}
+
+static void rs780_voltage_scaling_enable(struct radeon_device *rdev, bool enable);
+
+static int rs780_initialize_dpm_power_state(struct radeon_device *rdev,
+					    struct radeon_ps *boot_ps)
+{
+	struct atom_clock_dividers dividers;
+	struct igp_ps *default_state = rs780_get_ps(boot_ps);
+	int i, ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     default_state->sclk_low, false, &dividers);
+	if (ret)
+		return ret;
+
+	r600_engine_clock_entry_set_reference_divider(rdev, 0, dividers.ref_div);
+	r600_engine_clock_entry_set_feedback_divider(rdev, 0, dividers.fb_div);
+	r600_engine_clock_entry_set_post_divider(rdev, 0, dividers.post_div);
+
+	if (dividers.enable_post_div)
+		r600_engine_clock_entry_enable_post_divider(rdev, 0, true);
+	else
+		r600_engine_clock_entry_enable_post_divider(rdev, 0, false);
+
+	r600_engine_clock_entry_set_step_time(rdev, 0, R600_SST_DFLT);
+	r600_engine_clock_entry_enable_pulse_skipping(rdev, 0, false);
+
+	r600_engine_clock_entry_enable(rdev, 0, true);
+	for (i = 1; i < R600_PM_NUMBER_OF_SCLKS; i++)
+		r600_engine_clock_entry_enable(rdev, i, false);
+
+	r600_enable_mclk_control(rdev, false);
+	r600_voltage_control_enable_pins(rdev, 0);
+
+	return 0;
+}
+
+static int rs780_initialize_dpm_parameters(struct radeon_device *rdev,
+					   struct radeon_ps *boot_ps)
+{
+	int ret = 0;
+	int i;
+
+	r600_set_bsp(rdev, R600_BSU_DFLT, R600_BSP_DFLT);
+
+	r600_set_at(rdev, 0, 0, 0, 0);
+
+	r600_set_git(rdev, R600_GICST_DFLT);
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		r600_set_tc(rdev, i, 0, 0);
+
+	r600_select_td(rdev, R600_TD_DFLT);
+	r600_set_vrc(rdev, 0);
+
+	r600_set_tpu(rdev, R600_TPU_DFLT);
+	r600_set_tpc(rdev, R600_TPC_DFLT);
+
+	r600_set_sstu(rdev, R600_SSTU_DFLT);
+	r600_set_sst(rdev, R600_SST_DFLT);
+
+	r600_set_fctu(rdev, R600_FCTU_DFLT);
+	r600_set_fct(rdev, R600_FCT_DFLT);
+
+	r600_set_vddc3d_oorsu(rdev, R600_VDDC3DOORSU_DFLT);
+	r600_set_vddc3d_oorphc(rdev, R600_VDDC3DOORPHC_DFLT);
+	r600_set_vddc3d_oorsdc(rdev, R600_VDDC3DOORSDC_DFLT);
+	r600_set_ctxcgtt3d_rphc(rdev, R600_CTXCGTT3DRPHC_DFLT);
+	r600_set_ctxcgtt3d_rsdc(rdev, R600_CTXCGTT3DRSDC_DFLT);
+
+	r600_vid_rt_set_vru(rdev, R600_VRU_DFLT);
+	r600_vid_rt_set_vrt(rdev, R600_VOLTAGERESPONSETIME_DFLT);
+	r600_vid_rt_set_ssu(rdev, R600_SPLLSTEPUNIT_DFLT);
+
+	ret = rs780_initialize_dpm_power_state(rdev, boot_ps);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_LOW,     0);
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_MEDIUM,  0);
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_HIGH,    0);
+
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_LOW,    0);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_MEDIUM, 0);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_HIGH,   0);
+
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_LOW,    0);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_MEDIUM, 0);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_HIGH,   0);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_LOW,    R600_DISPLAY_WATERMARK_HIGH);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_MEDIUM, R600_DISPLAY_WATERMARK_HIGH);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_HIGH,   R600_DISPLAY_WATERMARK_HIGH);
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_CTXSW, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+
+	r600_power_level_set_enter_index(rdev, R600_POWER_LEVEL_LOW);
+
+	r600_set_vrc(rdev, RS780_CGFTV_DFLT);
+
+	return ret;
+}
+
+static void rs780_start_dpm(struct radeon_device *rdev)
+{
+	r600_enable_sclk_control(rdev, false);
+	r600_enable_mclk_control(rdev, false);
+
+	r600_dynamicpm_enable(rdev, true);
+
+	radeon_wait_for_vblank(rdev, 0);
+	radeon_wait_for_vblank(rdev, 1);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_sclk_control(rdev, true);
+}
+
+
+static void rs780_preset_ranges_slow_clk_fbdiv_en(struct radeon_device *rdev)
+{
+	WREG32_P(FVTHROT_SLOW_CLK_FEEDBACK_DIV_REG1, RANGE_SLOW_CLK_FEEDBACK_DIV_EN,
+		 ~RANGE_SLOW_CLK_FEEDBACK_DIV_EN);
+
+	WREG32_P(FVTHROT_SLOW_CLK_FEEDBACK_DIV_REG1,
+		 RANGE0_SLOW_CLK_FEEDBACK_DIV(RS780_SLOWCLKFEEDBACKDIV_DFLT),
+		 ~RANGE0_SLOW_CLK_FEEDBACK_DIV_MASK);
+}
+
+static void rs780_preset_starting_fbdiv(struct radeon_device *rdev)
+{
+	u32 fbdiv = (RREG32(CG_SPLL_FUNC_CNTL) & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT;
+
+	WREG32_P(FVTHROT_FBDIV_REG1, STARTING_FEEDBACK_DIV(fbdiv),
+		 ~STARTING_FEEDBACK_DIV_MASK);
+
+	WREG32_P(FVTHROT_FBDIV_REG2, FORCED_FEEDBACK_DIV(fbdiv),
+		 ~FORCED_FEEDBACK_DIV_MASK);
+
+	WREG32_P(FVTHROT_FBDIV_REG1, FORCE_FEEDBACK_DIV, ~FORCE_FEEDBACK_DIV);
+}
+
+static void rs780_voltage_scaling_init(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct drm_device *dev = rdev->ddev;
+	u32 fv_throt_pwm_fb_div_range[3];
+	u32 fv_throt_pwm_range[4];
+
+	if (dev->pdev->device == 0x9614) {
+		fv_throt_pwm_fb_div_range[0] = RS780D_FVTHROTPWMFBDIVRANGEREG0_DFLT;
+		fv_throt_pwm_fb_div_range[1] = RS780D_FVTHROTPWMFBDIVRANGEREG1_DFLT;
+		fv_throt_pwm_fb_div_range[2] = RS780D_FVTHROTPWMFBDIVRANGEREG2_DFLT;
+	} else if ((dev->pdev->device == 0x9714) ||
+		   (dev->pdev->device == 0x9715)) {
+		fv_throt_pwm_fb_div_range[0] = RS880D_FVTHROTPWMFBDIVRANGEREG0_DFLT;
+		fv_throt_pwm_fb_div_range[1] = RS880D_FVTHROTPWMFBDIVRANGEREG1_DFLT;
+		fv_throt_pwm_fb_div_range[2] = RS880D_FVTHROTPWMFBDIVRANGEREG2_DFLT;
+	} else {
+		fv_throt_pwm_fb_div_range[0] = RS780_FVTHROTPWMFBDIVRANGEREG0_DFLT;
+		fv_throt_pwm_fb_div_range[1] = RS780_FVTHROTPWMFBDIVRANGEREG1_DFLT;
+		fv_throt_pwm_fb_div_range[2] = RS780_FVTHROTPWMFBDIVRANGEREG2_DFLT;
+	}
+
+	if (pi->pwm_voltage_control) {
+		fv_throt_pwm_range[0] = pi->min_voltage;
+		fv_throt_pwm_range[1] = pi->min_voltage;
+		fv_throt_pwm_range[2] = pi->max_voltage;
+		fv_throt_pwm_range[3] = pi->max_voltage;
+	} else {
+		fv_throt_pwm_range[0] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE3_GPIO_DFLT : RS780_FVTHROTPWMRANGE0_GPIO_DFLT;
+		fv_throt_pwm_range[1] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE2_GPIO_DFLT : RS780_FVTHROTPWMRANGE1_GPIO_DFLT;
+		fv_throt_pwm_range[2] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE1_GPIO_DFLT : RS780_FVTHROTPWMRANGE2_GPIO_DFLT;
+		fv_throt_pwm_range[3] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE0_GPIO_DFLT : RS780_FVTHROTPWMRANGE3_GPIO_DFLT;
+	}
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 STARTING_PWM_HIGHTIME(pi->max_voltage),
+		 ~STARTING_PWM_HIGHTIME_MASK);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 NUMBER_OF_CYCLES_IN_PERIOD(pi->num_of_cycles_in_period),
+		 ~NUMBER_OF_CYCLES_IN_PERIOD_MASK);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0, FORCE_STARTING_PWM_HIGHTIME,
+		 ~FORCE_STARTING_PWM_HIGHTIME);
+
+	if (pi->invert_pwm_required)
+		WREG32_P(FVTHROT_PWM_CTRL_REG0, INVERT_PWM_WAVEFORM, ~INVERT_PWM_WAVEFORM);
+	else
+		WREG32_P(FVTHROT_PWM_CTRL_REG0, 0, ~INVERT_PWM_WAVEFORM);
+
+	rs780_voltage_scaling_enable(rdev, true);
+
+	WREG32(FVTHROT_PWM_CTRL_REG1,
+	       (MIN_PWM_HIGHTIME(pi->min_voltage) |
+		MAX_PWM_HIGHTIME(pi->max_voltage)));
+
+	WREG32(FVTHROT_PWM_US_REG0, RS780_FVTHROTPWMUSREG0_DFLT);
+	WREG32(FVTHROT_PWM_US_REG1, RS780_FVTHROTPWMUSREG1_DFLT);
+	WREG32(FVTHROT_PWM_DS_REG0, RS780_FVTHROTPWMDSREG0_DFLT);
+	WREG32(FVTHROT_PWM_DS_REG1, RS780_FVTHROTPWMDSREG1_DFLT);
+
+	WREG32_P(FVTHROT_PWM_FEEDBACK_DIV_REG1,
+		 RANGE0_PWM_FEEDBACK_DIV(fv_throt_pwm_fb_div_range[0]),
+		 ~RANGE0_PWM_FEEDBACK_DIV_MASK);
+
+	WREG32(FVTHROT_PWM_FEEDBACK_DIV_REG2,
+	       (RANGE1_PWM_FEEDBACK_DIV(fv_throt_pwm_fb_div_range[1]) |
+		RANGE2_PWM_FEEDBACK_DIV(fv_throt_pwm_fb_div_range[2])));
+
+	WREG32(FVTHROT_PWM_FEEDBACK_DIV_REG3,
+	       (RANGE0_PWM(fv_throt_pwm_range[1]) |
+		RANGE1_PWM(fv_throt_pwm_range[2])));
+	WREG32(FVTHROT_PWM_FEEDBACK_DIV_REG4,
+	       (RANGE2_PWM(fv_throt_pwm_range[1]) |
+		RANGE3_PWM(fv_throt_pwm_range[2])));
+}
+
+static void rs780_clk_scaling_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(FVTHROT_CNTRL_REG, ENABLE_FV_THROT | ENABLE_FV_UPDATE,
+			 ~(ENABLE_FV_THROT | ENABLE_FV_UPDATE));
+	else
+		WREG32_P(FVTHROT_CNTRL_REG, 0,
+			 ~(ENABLE_FV_THROT | ENABLE_FV_UPDATE));
+}
+
+static void rs780_voltage_scaling_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(FVTHROT_CNTRL_REG, ENABLE_FV_THROT_IO, ~ENABLE_FV_THROT_IO);
+	else
+		WREG32_P(FVTHROT_CNTRL_REG, 0, ~ENABLE_FV_THROT_IO);
+}
+
+static void rs780_set_engine_clock_wfc(struct radeon_device *rdev)
+{
+	WREG32(FVTHROT_UTC0, RS780_FVTHROTUTC0_DFLT);
+	WREG32(FVTHROT_UTC1, RS780_FVTHROTUTC1_DFLT);
+	WREG32(FVTHROT_UTC2, RS780_FVTHROTUTC2_DFLT);
+	WREG32(FVTHROT_UTC3, RS780_FVTHROTUTC3_DFLT);
+	WREG32(FVTHROT_UTC4, RS780_FVTHROTUTC4_DFLT);
+
+	WREG32(FVTHROT_DTC0, RS780_FVTHROTDTC0_DFLT);
+	WREG32(FVTHROT_DTC1, RS780_FVTHROTDTC1_DFLT);
+	WREG32(FVTHROT_DTC2, RS780_FVTHROTDTC2_DFLT);
+	WREG32(FVTHROT_DTC3, RS780_FVTHROTDTC3_DFLT);
+	WREG32(FVTHROT_DTC4, RS780_FVTHROTDTC4_DFLT);
+}
+
+static void rs780_set_engine_clock_sc(struct radeon_device *rdev)
+{
+	WREG32_P(FVTHROT_FBDIV_REG2,
+		 FB_DIV_TIMER_VAL(RS780_FBDIVTIMERVAL_DFLT),
+		 ~FB_DIV_TIMER_VAL_MASK);
+
+	WREG32_P(FVTHROT_CNTRL_REG,
+		 REFRESH_RATE_DIVISOR(0) | MINIMUM_CIP(0xf),
+		 ~(REFRESH_RATE_DIVISOR_MASK | MINIMUM_CIP_MASK));
+}
+
+static void rs780_set_engine_clock_tdc(struct radeon_device *rdev)
+{
+	WREG32_P(FVTHROT_CNTRL_REG, 0, ~(FORCE_TREND_SEL | TREND_SEL_MODE));
+}
+
+static void rs780_set_engine_clock_ssc(struct radeon_device *rdev)
+{
+	WREG32(FVTHROT_FB_US_REG0, RS780_FVTHROTFBUSREG0_DFLT);
+	WREG32(FVTHROT_FB_US_REG1, RS780_FVTHROTFBUSREG1_DFLT);
+	WREG32(FVTHROT_FB_DS_REG0, RS780_FVTHROTFBDSREG0_DFLT);
+	WREG32(FVTHROT_FB_DS_REG1, RS780_FVTHROTFBDSREG1_DFLT);
+
+	WREG32_P(FVTHROT_FBDIV_REG1, MAX_FEEDBACK_STEP(1), ~MAX_FEEDBACK_STEP_MASK);
+}
+
+static void rs780_program_at(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	WREG32(FVTHROT_TARGET_REG, 30000000 / pi->refresh_rate);
+	WREG32(FVTHROT_CB1, 1000000 * 5 / pi->refresh_rate);
+	WREG32(FVTHROT_CB2, 1000000 * 10 / pi->refresh_rate);
+	WREG32(FVTHROT_CB3, 1000000 * 30 / pi->refresh_rate);
+	WREG32(FVTHROT_CB4, 1000000 * 50 / pi->refresh_rate);
+}
+
+static void rs780_disable_vbios_powersaving(struct radeon_device *rdev)
+{
+	WREG32_P(CG_INTGFX_MISC, 0, ~0xFFF00000);
+}
+
+static void rs780_force_voltage_to_high(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct igp_ps *current_state = rs780_get_ps(rdev->pm.dpm.current_ps);
+
+	if ((current_state->max_voltage == RS780_VDDC_LEVEL_HIGH) &&
+	    (current_state->min_voltage == RS780_VDDC_LEVEL_HIGH))
+		return;
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL);
+
+	udelay(1);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 STARTING_PWM_HIGHTIME(pi->max_voltage),
+		 ~STARTING_PWM_HIGHTIME_MASK);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 FORCE_STARTING_PWM_HIGHTIME, ~FORCE_STARTING_PWM_HIGHTIME);
+
+	WREG32_P(FVTHROT_PWM_FEEDBACK_DIV_REG1, 0,
+		~RANGE_PWM_FEEDBACK_DIV_EN);
+
+	udelay(1);
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL);
+}
+
+static int rs780_set_engine_clock_scaling(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps,
+					  struct radeon_ps *old_ps)
+{
+	struct atom_clock_dividers min_dividers, max_dividers, current_max_dividers;
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *old_state = rs780_get_ps(old_ps);
+	int ret;
+
+	if ((new_state->sclk_high == old_state->sclk_high) &&
+	    (new_state->sclk_low == old_state->sclk_low))
+		return 0;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     new_state->sclk_low, false, &min_dividers);
+	if (ret)
+		return ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     new_state->sclk_high, false, &max_dividers);
+	if (ret)
+		return ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     old_state->sclk_high, false, &current_max_dividers);
+	if (ret)
+		return ret;
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL);
+
+	WREG32_P(FVTHROT_FBDIV_REG2, FORCED_FEEDBACK_DIV(max_dividers.fb_div),
+		 ~FORCED_FEEDBACK_DIV_MASK);
+	WREG32_P(FVTHROT_FBDIV_REG1, STARTING_FEEDBACK_DIV(max_dividers.fb_div),
+		 ~STARTING_FEEDBACK_DIV_MASK);
+	WREG32_P(FVTHROT_FBDIV_REG1, FORCE_FEEDBACK_DIV, ~FORCE_FEEDBACK_DIV);
+
+	udelay(100);
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL);
+
+	if (max_dividers.fb_div > min_dividers.fb_div) {
+		WREG32_P(FVTHROT_FBDIV_REG0,
+			 MIN_FEEDBACK_DIV(min_dividers.fb_div) |
+			 MAX_FEEDBACK_DIV(max_dividers.fb_div),
+			 ~(MIN_FEEDBACK_DIV_MASK | MAX_FEEDBACK_DIV_MASK));
+
+		WREG32_P(FVTHROT_FBDIV_REG1, 0, ~FORCE_FEEDBACK_DIV);
+	}
+
+	return 0;
+}
+
+static void rs780_set_engine_clock_spc(struct radeon_device *rdev,
+				       struct radeon_ps *new_ps,
+				       struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *old_state = rs780_get_ps(old_ps);
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	if ((new_state->sclk_high == old_state->sclk_high) &&
+	    (new_state->sclk_low == old_state->sclk_low))
+		return;
+
+	if (pi->crtc_id == 0)
+		WREG32_P(CG_INTGFX_MISC, 0, ~FVTHROT_VBLANK_SEL);
+	else
+		WREG32_P(CG_INTGFX_MISC, FVTHROT_VBLANK_SEL, ~FVTHROT_VBLANK_SEL);
+
+}
+
+static void rs780_activate_engine_clk_scaling(struct radeon_device *rdev,
+					      struct radeon_ps *new_ps,
+					      struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *old_state = rs780_get_ps(old_ps);
+
+	if ((new_state->sclk_high == old_state->sclk_high) &&
+	    (new_state->sclk_low == old_state->sclk_low))
+		return;
+
+	rs780_clk_scaling_enable(rdev, true);
+}
+
+static u32 rs780_get_voltage_for_vddc_level(struct radeon_device *rdev,
+					    enum rs780_vddc_level vddc)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	if (vddc == RS780_VDDC_LEVEL_HIGH)
+		return pi->max_voltage;
+	else if (vddc == RS780_VDDC_LEVEL_LOW)
+		return pi->min_voltage;
+	else
+		return pi->max_voltage;
+}
+
+static void rs780_enable_voltage_scaling(struct radeon_device *rdev,
+					 struct radeon_ps *new_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	enum rs780_vddc_level vddc_high, vddc_low;
+
+	udelay(100);
+
+	if ((new_state->max_voltage == RS780_VDDC_LEVEL_HIGH) &&
+	    (new_state->min_voltage == RS780_VDDC_LEVEL_HIGH))
+		return;
+
+	vddc_high = rs780_get_voltage_for_vddc_level(rdev,
+						     new_state->max_voltage);
+	vddc_low = rs780_get_voltage_for_vddc_level(rdev,
+						    new_state->min_voltage);
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL);
+
+	udelay(1);
+	if (vddc_high > vddc_low) {
+		WREG32_P(FVTHROT_PWM_FEEDBACK_DIV_REG1,
+			 RANGE_PWM_FEEDBACK_DIV_EN, ~RANGE_PWM_FEEDBACK_DIV_EN);
+
+		WREG32_P(FVTHROT_PWM_CTRL_REG0, 0, ~FORCE_STARTING_PWM_HIGHTIME);
+	} else if (vddc_high == vddc_low) {
+		if (pi->max_voltage != vddc_high) {
+			WREG32_P(FVTHROT_PWM_CTRL_REG0,
+				 STARTING_PWM_HIGHTIME(vddc_high),
+				 ~STARTING_PWM_HIGHTIME_MASK);
+
+			WREG32_P(FVTHROT_PWM_CTRL_REG0,
+				 FORCE_STARTING_PWM_HIGHTIME,
+				 ~FORCE_STARTING_PWM_HIGHTIME);
+		}
+	}
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL);
+}
+
+static void rs780_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						     struct radeon_ps *new_ps,
+						     struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *current_state = rs780_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->sclk_high >= current_state->sclk_high)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+static void rs780_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						    struct radeon_ps *new_ps,
+						    struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *current_state = rs780_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->sclk_high < current_state->sclk_high)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+int rs780_dpm_enable(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	rs780_get_pm_mode_parameters(rdev);
+	rs780_disable_vbios_powersaving(rdev);
+
+	if (r600_dynamicpm_enabled(rdev))
+		return -EINVAL;
+	ret = rs780_initialize_dpm_parameters(rdev, boot_ps);
+	if (ret)
+		return ret;
+	rs780_start_dpm(rdev);
+
+	rs780_preset_ranges_slow_clk_fbdiv_en(rdev);
+	rs780_preset_starting_fbdiv(rdev);
+	if (pi->voltage_control)
+		rs780_voltage_scaling_init(rdev);
+	rs780_clk_scaling_enable(rdev, true);
+	rs780_set_engine_clock_sc(rdev);
+	rs780_set_engine_clock_wfc(rdev);
+	rs780_program_at(rdev);
+	rs780_set_engine_clock_tdc(rdev);
+	rs780_set_engine_clock_ssc(rdev);
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, true);
+
+	if (rdev->irq.installed && (rdev->pm.int_thermal_type == THERMAL_TYPE_RV6XX)) {
+		ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	return 0;
+}
+
+void rs780_dpm_disable(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	r600_dynamicpm_enable(rdev, false);
+
+	rs780_clk_scaling_enable(rdev, false);
+	rs780_voltage_scaling_enable(rdev, false);
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, false);
+
+	if (rdev->irq.installed &&
+	    (rdev->pm.int_thermal_type == THERMAL_TYPE_RV6XX)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+}
+
+int rs780_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	rs780_get_pm_mode_parameters(rdev);
+
+	rs780_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+
+	if (pi->voltage_control) {
+		rs780_force_voltage_to_high(rdev);
+		mdelay(5);
+	}
+
+	ret = rs780_set_engine_clock_scaling(rdev, new_ps, old_ps);
+	if (ret)
+		return ret;
+	rs780_set_engine_clock_spc(rdev, new_ps, old_ps);
+
+	rs780_activate_engine_clk_scaling(rdev, new_ps, old_ps);
+
+	if (pi->voltage_control)
+		rs780_enable_voltage_scaling(rdev, new_ps);
+
+	rs780_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	return 0;
+}
+
+void rs780_dpm_setup_asic(struct radeon_device *rdev)
+{
+
+}
+
+void rs780_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	rs780_get_pm_mode_parameters(rdev);
+	rs780_program_at(rdev);
+}
+
+union igp_info {
+	struct _ATOM_INTEGRATED_SYSTEM_INFO info;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
+};
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void rs780_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					     struct radeon_ps *rps,
+					     struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					     u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RS780_DEFAULT_VCLK_FREQ;
+		rps->dclk = RS780_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void rs780_parse_pplib_clock_info(struct radeon_device *rdev,
+					 struct radeon_ps *rps,
+					 union pplib_clock_info *clock_info)
+{
+	struct igp_ps *ps = rs780_get_ps(rps);
+	u32 sclk;
+
+	sclk = le16_to_cpu(clock_info->rs780.usLowEngineClockLow);
+	sclk |= clock_info->rs780.ucLowEngineClockHigh << 16;
+	ps->sclk_low = sclk;
+	sclk = le16_to_cpu(clock_info->rs780.usHighEngineClockLow);
+	sclk |= clock_info->rs780.ucHighEngineClockHigh << 16;
+	ps->sclk_high = sclk;
+	switch (le16_to_cpu(clock_info->rs780.usVDDC)) {
+	case ATOM_PPLIB_RS780_VOLTAGE_NONE:
+	default:
+		ps->min_voltage = RS780_VDDC_LEVEL_UNKNOWN;
+		ps->max_voltage = RS780_VDDC_LEVEL_UNKNOWN;
+		break;
+	case ATOM_PPLIB_RS780_VOLTAGE_LOW:
+		ps->min_voltage = RS780_VDDC_LEVEL_LOW;
+		ps->max_voltage = RS780_VDDC_LEVEL_LOW;
+		break;
+	case ATOM_PPLIB_RS780_VOLTAGE_HIGH:
+		ps->min_voltage = RS780_VDDC_LEVEL_HIGH;
+		ps->max_voltage = RS780_VDDC_LEVEL_HIGH;
+		break;
+	case ATOM_PPLIB_RS780_VOLTAGE_VARIABLE:
+		ps->min_voltage = RS780_VDDC_LEVEL_LOW;
+		ps->max_voltage = RS780_VDDC_LEVEL_HIGH;
+		break;
+	}
+	ps->flags = le32_to_cpu(clock_info->rs780.ulFlags);
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		ps->sclk_low = rdev->clock.default_sclk;
+		ps->sclk_high = rdev->clock.default_sclk;
+		ps->min_voltage = RS780_VDDC_LEVEL_HIGH;
+		ps->max_voltage = RS780_VDDC_LEVEL_HIGH;
+	}
+}
+
+static int rs780_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct igp_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			clock_info = (union pplib_clock_info *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+				 (power_state->v1.ucClockStateIndices[0] *
+				  power_info->pplib.ucClockInfoSize));
+			ps = kzalloc(sizeof(struct igp_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			rs780_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info,
+							 power_info->pplib.ucNonClockSize);
+			rs780_parse_pplib_clock_info(rdev,
+						     &rdev->pm.dpm.ps[i],
+						     clock_info);
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int rs780_dpm_init(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi;
+	int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo);
+	union igp_info *info;
+	u16 data_offset;
+	u8 frev, crev;
+	int ret;
+
+	pi = kzalloc(sizeof(struct igp_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	ret = rs780_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	pi->voltage_control = false;
+	pi->gfx_clock_gating = true;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		info = (union igp_info *)(rdev->mode_info.atom_context->bios + data_offset);
+
+		/* Get various system informations from bios */
+		switch (crev) {
+		case 1:
+			pi->num_of_cycles_in_period =
+				info->info.ucNumberOfCyclesInPeriod;
+			pi->num_of_cycles_in_period |=
+				info->info.ucNumberOfCyclesInPeriodHi << 8;
+			pi->invert_pwm_required =
+				(pi->num_of_cycles_in_period & 0x8000) ? true : false;
+			pi->boot_voltage = info->info.ucStartingPWM_HighTime;
+			pi->max_voltage = info->info.ucMaxNBVoltage;
+			pi->max_voltage |= info->info.ucMaxNBVoltageHigh << 8;
+			pi->min_voltage = info->info.ucMinNBVoltage;
+			pi->min_voltage |= info->info.ucMinNBVoltageHigh << 8;
+			pi->inter_voltage_low =
+				le16_to_cpu(info->info.usInterNBVoltageLow);
+			pi->inter_voltage_high =
+				le16_to_cpu(info->info.usInterNBVoltageHigh);
+			pi->voltage_control = true;
+			pi->bootup_uma_clk = info->info.usK8MemoryClock * 100;
+			break;
+		case 2:
+			pi->num_of_cycles_in_period =
+				le16_to_cpu(info->info_2.usNumberOfCyclesInPeriod);
+			pi->invert_pwm_required =
+				(pi->num_of_cycles_in_period & 0x8000) ? true : false;
+			pi->boot_voltage =
+				le16_to_cpu(info->info_2.usBootUpNBVoltage);
+			pi->max_voltage =
+				le16_to_cpu(info->info_2.usMaxNBVoltage);
+			pi->min_voltage =
+				le16_to_cpu(info->info_2.usMinNBVoltage);
+			pi->system_config =
+				le32_to_cpu(info->info_2.ulSystemConfig);
+			pi->pwm_voltage_control =
+				(pi->system_config & 0x4) ? true : false;
+			pi->voltage_control = true;
+			pi->bootup_uma_clk = le32_to_cpu(info->info_2.ulBootUpUMAClock);
+			break;
+		default:
+			DRM_ERROR("No integrated system info for your GPU\n");
+			return -EINVAL;
+		}
+		if (pi->min_voltage > pi->max_voltage)
+			pi->voltage_control = false;
+		if (pi->pwm_voltage_control) {
+			if ((pi->num_of_cycles_in_period == 0) ||
+			    (pi->max_voltage == 0) ||
+			    (pi->min_voltage == 0))
+				pi->voltage_control = false;
+		} else {
+			if ((pi->num_of_cycles_in_period == 0) ||
+			    (pi->max_voltage == 0))
+				pi->voltage_control = false;
+		}
+
+		return 0;
+	}
+	radeon_dpm_fini(rdev);
+	return -EINVAL;
+}
+
+void rs780_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct igp_ps *ps = rs780_get_ps(rps);
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	printk("\t\tpower level 0    sclk: %u vddc_index: %d\n",
+	       ps->sclk_low, ps->min_voltage);
+	printk("\t\tpower level 1    sclk: %u vddc_index: %d\n",
+	       ps->sclk_high, ps->max_voltage);
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void rs780_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 rs780_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct igp_ps *requested_state = rs780_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->sclk_low;
+	else
+		return requested_state->sclk_high;
+}
+
+u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	return pi->bootup_uma_clk;
+}
diff --git a/drivers/gpu/drm/radeon/rs780_dpm.h b/drivers/gpu/drm/radeon/rs780_dpm.h
new file mode 100644
index 0000000..47a40b1
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rs780_dpm.h
@@ -0,0 +1,109 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RS780_DPM_H__
+#define __RS780_DPM_H__
+
+enum rs780_vddc_level {
+	RS780_VDDC_LEVEL_UNKNOWN = 0,
+	RS780_VDDC_LEVEL_LOW = 1,
+	RS780_VDDC_LEVEL_HIGH = 2,
+};
+
+struct igp_power_info {
+	/* flags */
+	bool invert_pwm_required;
+	bool pwm_voltage_control;
+	bool voltage_control;
+	bool gfx_clock_gating;
+	/* stored values */
+	u32 system_config;
+	u32 bootup_uma_clk;
+	u16 max_voltage;
+	u16 min_voltage;
+	u16 boot_voltage;
+	u16 inter_voltage_low;
+	u16 inter_voltage_high;
+	u16 num_of_cycles_in_period;
+	/* variable */
+	int crtc_id;
+	int refresh_rate;
+};
+
+struct igp_ps {
+	enum rs780_vddc_level min_voltage;
+	enum rs780_vddc_level max_voltage;
+	u32 sclk_low;
+	u32 sclk_high;
+	u32 flags;
+};
+
+#define RS780_CGFTV_DFLT                 0x0303000f
+#define RS780_FBDIVTIMERVAL_DFLT         0x2710
+
+#define RS780_FVTHROTUTC0_DFLT   0x04010040
+#define RS780_FVTHROTUTC1_DFLT   0x04010040
+#define RS780_FVTHROTUTC2_DFLT   0x04010040
+#define RS780_FVTHROTUTC3_DFLT   0x04010040
+#define RS780_FVTHROTUTC4_DFLT   0x04010040
+
+#define RS780_FVTHROTDTC0_DFLT 0x04010040
+#define RS780_FVTHROTDTC1_DFLT 0x04010040
+#define RS780_FVTHROTDTC2_DFLT 0x04010040
+#define RS780_FVTHROTDTC3_DFLT 0x04010040
+#define RS780_FVTHROTDTC4_DFLT 0x04010040
+
+#define RS780_FVTHROTFBUSREG0_DFLT       0x00001001
+#define RS780_FVTHROTFBUSREG1_DFLT       0x00002002
+#define RS780_FVTHROTFBDSREG0_DFLT       0x00004001
+#define RS780_FVTHROTFBDSREG1_DFLT       0x00020010
+
+#define RS780_FVTHROTPWMUSREG0_DFLT      0x00002001
+#define RS780_FVTHROTPWMUSREG1_DFLT      0x00004003
+#define RS780_FVTHROTPWMDSREG0_DFLT      0x00002001
+#define RS780_FVTHROTPWMDSREG1_DFLT      0x00004003
+
+#define RS780_FVTHROTPWMFBDIVRANGEREG0_DFLT  0x37
+#define RS780_FVTHROTPWMFBDIVRANGEREG1_DFLT  0x4b
+#define RS780_FVTHROTPWMFBDIVRANGEREG2_DFLT  0x8b
+
+#define RS780D_FVTHROTPWMFBDIVRANGEREG0_DFLT  0x8b
+#define RS780D_FVTHROTPWMFBDIVRANGEREG1_DFLT  0x8c
+#define RS780D_FVTHROTPWMFBDIVRANGEREG2_DFLT  0xb5
+
+#define RS880D_FVTHROTPWMFBDIVRANGEREG0_DFLT  0x8d
+#define RS880D_FVTHROTPWMFBDIVRANGEREG1_DFLT  0x8e
+#define RS880D_FVTHROTPWMFBDIVRANGEREG2_DFLT  0xBa
+
+#define RS780_FVTHROTPWMRANGE0_GPIO_DFLT  0x1a
+#define RS780_FVTHROTPWMRANGE1_GPIO_DFLT  0x1a
+#define RS780_FVTHROTPWMRANGE2_GPIO_DFLT  0x0
+#define RS780_FVTHROTPWMRANGE3_GPIO_DFLT  0x0
+
+#define RS780_SLOWCLKFEEDBACKDIV_DFLT 110
+
+#define RS780_CGCLKGATING_DFLT           0x0000E204
+
+#define RS780_DEFAULT_VCLK_FREQ  53300 /* 10 khz */
+#define RS780_DEFAULT_DCLK_FREQ  40000 /* 10 khz */
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rs780d.h b/drivers/gpu/drm/radeon/rs780d.h
new file mode 100644
index 0000000..b1142ed
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rs780d.h
@@ -0,0 +1,168 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RS780D_H__
+#define __RS780D_H__
+
+#define CG_SPLL_FUNC_CNTL                                 0x600
+#       define SPLL_RESET                                (1 << 0)
+#       define SPLL_SLEEP                                (1 << 1)
+#       define SPLL_REF_DIV(x)                           ((x) << 2)
+#       define SPLL_REF_DIV_MASK                         (7 << 2)
+#       define SPLL_FB_DIV(x)                            ((x) << 5)
+#       define SPLL_FB_DIV_MASK                          (0xff << 2)
+#       define SPLL_FB_DIV_SHIFT                         2
+#       define SPLL_PULSEEN                              (1 << 13)
+#       define SPLL_PULSENUM(x)                          ((x) << 14)
+#       define SPLL_PULSENUM_MASK                        (3 << 14)
+#       define SPLL_SW_HILEN(x)                          ((x) << 16)
+#       define SPLL_SW_HILEN_MASK                        (0xf << 16)
+#       define SPLL_SW_LOLEN(x)                          ((x) << 20)
+#       define SPLL_SW_LOLEN_MASK                        (0xf << 20)
+#       define SPLL_DIVEN                                (1 << 24)
+#       define SPLL_BYPASS_EN                            (1 << 25)
+#       define SPLL_CHG_STATUS                           (1 << 29)
+#       define SPLL_CTLREQ                               (1 << 30)
+#       define SPLL_CTLACK                               (1 << 31)
+
+/* RS780/RS880 PM */
+#define	FVTHROT_CNTRL_REG				0x3000
+#define		DONT_WAIT_FOR_FBDIV_WRAP		(1 << 0)
+#define		MINIMUM_CIP(x)				((x) << 1)
+#define		MINIMUM_CIP_SHIFT			1
+#define		MINIMUM_CIP_MASK			0x1fffffe
+#define		REFRESH_RATE_DIVISOR(x)			((x) << 25)
+#define		REFRESH_RATE_DIVISOR_SHIFT		25
+#define		REFRESH_RATE_DIVISOR_MASK		(0x3 << 25)
+#define		ENABLE_FV_THROT				(1 << 27)
+#define		ENABLE_FV_UPDATE			(1 << 28)
+#define		TREND_SEL_MODE				(1 << 29)
+#define		FORCE_TREND_SEL				(1 << 30)
+#define		ENABLE_FV_THROT_IO			(1 << 31)
+#define	FVTHROT_TARGET_REG				0x3004
+#define		TARGET_IDLE_COUNT(x)			((x) << 0)
+#define		TARGET_IDLE_COUNT_MASK			0xffffff
+#define		TARGET_IDLE_COUNT_SHIFT			0
+#define	FVTHROT_CB1					0x3008
+#define	FVTHROT_CB2					0x300c
+#define	FVTHROT_CB3					0x3010
+#define	FVTHROT_CB4					0x3014
+#define	FVTHROT_UTC0					0x3018
+#define	FVTHROT_UTC1					0x301c
+#define	FVTHROT_UTC2					0x3020
+#define	FVTHROT_UTC3					0x3024
+#define	FVTHROT_UTC4					0x3028
+#define	FVTHROT_DTC0					0x302c
+#define	FVTHROT_DTC1					0x3030
+#define	FVTHROT_DTC2					0x3034
+#define	FVTHROT_DTC3					0x3038
+#define	FVTHROT_DTC4					0x303c
+#define	FVTHROT_FBDIV_REG0				0x3040
+#define		MIN_FEEDBACK_DIV(x)			((x) << 0)
+#define		MIN_FEEDBACK_DIV_MASK			0xfff
+#define		MIN_FEEDBACK_DIV_SHIFT			0
+#define		MAX_FEEDBACK_DIV(x)			((x) << 12)
+#define		MAX_FEEDBACK_DIV_MASK			(0xfff << 12)
+#define		MAX_FEEDBACK_DIV_SHIFT			12
+#define	FVTHROT_FBDIV_REG1				0x3044
+#define		MAX_FEEDBACK_STEP(x)			((x) << 0)
+#define		MAX_FEEDBACK_STEP_MASK			0xfff
+#define		MAX_FEEDBACK_STEP_SHIFT			0
+#define		STARTING_FEEDBACK_DIV(x)		((x) << 12)
+#define		STARTING_FEEDBACK_DIV_MASK		(0xfff << 12)
+#define		STARTING_FEEDBACK_DIV_SHIFT		12
+#define		FORCE_FEEDBACK_DIV			(1 << 24)
+#define	FVTHROT_FBDIV_REG2				0x3048
+#define		FORCED_FEEDBACK_DIV(x)			((x) << 0)
+#define		FORCED_FEEDBACK_DIV_MASK		0xfff
+#define		FORCED_FEEDBACK_DIV_SHIFT		0
+#define		FB_DIV_TIMER_VAL(x)			((x) << 12)
+#define		FB_DIV_TIMER_VAL_MASK			(0xffff << 12)
+#define		FB_DIV_TIMER_VAL_SHIFT			12
+#define	FVTHROT_FB_US_REG0				0x304c
+#define	FVTHROT_FB_US_REG1				0x3050
+#define	FVTHROT_FB_DS_REG0				0x3054
+#define	FVTHROT_FB_DS_REG1				0x3058
+#define	FVTHROT_PWM_CTRL_REG0				0x305c
+#define		STARTING_PWM_HIGHTIME(x)		((x) << 0)
+#define		STARTING_PWM_HIGHTIME_MASK		0xfff
+#define		STARTING_PWM_HIGHTIME_SHIFT		0
+#define		NUMBER_OF_CYCLES_IN_PERIOD(x)		((x) << 12)
+#define		NUMBER_OF_CYCLES_IN_PERIOD_MASK		(0xfff << 12)
+#define		NUMBER_OF_CYCLES_IN_PERIOD_SHIFT	12
+#define		FORCE_STARTING_PWM_HIGHTIME		(1 << 24)
+#define		INVERT_PWM_WAVEFORM			(1 << 25)
+#define	FVTHROT_PWM_CTRL_REG1				0x3060
+#define		MIN_PWM_HIGHTIME(x)			((x) << 0)
+#define		MIN_PWM_HIGHTIME_MASK			0xfff
+#define		MIN_PWM_HIGHTIME_SHIFT			0
+#define		MAX_PWM_HIGHTIME(x)			((x) << 12)
+#define		MAX_PWM_HIGHTIME_MASK			(0xfff << 12)
+#define		MAX_PWM_HIGHTIME_SHIFT			12
+#define	FVTHROT_PWM_US_REG0				0x3064
+#define	FVTHROT_PWM_US_REG1				0x3068
+#define	FVTHROT_PWM_DS_REG0				0x306c
+#define	FVTHROT_PWM_DS_REG1				0x3070
+#define	FVTHROT_STATUS_REG0				0x3074
+#define		CURRENT_FEEDBACK_DIV_MASK		0xfff
+#define		CURRENT_FEEDBACK_DIV_SHIFT		0
+#define	FVTHROT_STATUS_REG1				0x3078
+#define	FVTHROT_STATUS_REG2				0x307c
+#define	CG_INTGFX_MISC					0x3080
+#define		FVTHROT_VBLANK_SEL			(1 << 9)
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG1			0x308c
+#define		RANGE0_PWM_FEEDBACK_DIV(x)		((x) << 0)
+#define		RANGE0_PWM_FEEDBACK_DIV_MASK		0xfff
+#define		RANGE0_PWM_FEEDBACK_DIV_SHIFT		0
+#define		RANGE_PWM_FEEDBACK_DIV_EN		(1 << 12)
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG2			0x3090
+#define		RANGE1_PWM_FEEDBACK_DIV(x)		((x) << 0)
+#define		RANGE1_PWM_FEEDBACK_DIV_MASK		0xfff
+#define		RANGE1_PWM_FEEDBACK_DIV_SHIFT		0
+#define		RANGE2_PWM_FEEDBACK_DIV(x)		((x) << 12)
+#define		RANGE2_PWM_FEEDBACK_DIV_MASK		(0xfff << 12)
+#define		RANGE2_PWM_FEEDBACK_DIV_SHIFT		12
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG3			0x3094
+#define		RANGE0_PWM(x)				((x) << 0)
+#define		RANGE0_PWM_MASK				0xfff
+#define		RANGE0_PWM_SHIFT			0
+#define		RANGE1_PWM(x)				((x) << 12)
+#define		RANGE1_PWM_MASK				(0xfff << 12)
+#define		RANGE1_PWM_SHIFT			12
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG4			0x3098
+#define		RANGE2_PWM(x)				((x) << 0)
+#define		RANGE2_PWM_MASK				0xfff
+#define		RANGE2_PWM_SHIFT			0
+#define		RANGE3_PWM(x)				((x) << 12)
+#define		RANGE3_PWM_MASK				(0xfff << 12)
+#define		RANGE3_PWM_SHIFT			12
+#define	FVTHROT_SLOW_CLK_FEEDBACK_DIV_REG1		0x30ac
+#define		RANGE0_SLOW_CLK_FEEDBACK_DIV(x)		((x) << 0)
+#define		RANGE0_SLOW_CLK_FEEDBACK_DIV_MASK	0xfff
+#define		RANGE0_SLOW_CLK_FEEDBACK_DIV_SHIFT	0
+#define		RANGE_SLOW_CLK_FEEDBACK_DIV_EN		(1 << 12)
+
+#define	GFX_MACRO_BYPASS_CNTL				0x30c0
+#define		SPLL_BYPASS_CNTL			(1 << 0)
+#define		UPLL_BYPASS_CNTL			(1 << 1)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 21c7d7b..8ea1573 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -937,13 +937,16 @@ struct rv515_watermark {
 };
 
 static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
-				  struct radeon_crtc *crtc,
-				  struct rv515_watermark *wm)
+					 struct radeon_crtc *crtc,
+					 struct rv515_watermark *wm,
+					 bool low)
 {
 	struct drm_display_mode *mode = &crtc->base.mode;
 	fixed20_12 a, b, c;
 	fixed20_12 pclk, request_fifo_depth, tolerable_latency, estimated_width;
 	fixed20_12 consumption_time, line_time, chunk_time, read_delay_latency;
+	fixed20_12 sclk;
+	u32 selected_sclk;
 
 	if (!crtc->base.enabled) {
 		/* FIXME: wouldn't it better to set priority mark to maximum */
@@ -951,6 +954,18 @@ static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
 		return;
 	}
 
+	/* rv6xx, rv7xx */
+	if ((rdev->family >= CHIP_RV610) &&
+	    (rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
+		selected_sclk = radeon_dpm_get_sclk(rdev, low);
+	else
+		selected_sclk = rdev->pm.current_sclk;
+
+	/* sclk in Mhz */
+	a.full = dfixed_const(100);
+	sclk.full = dfixed_const(selected_sclk);
+	sclk.full = dfixed_div(sclk, a);
+
 	if (crtc->vsc.full > dfixed_const(2))
 		wm->num_line_pair.full = dfixed_const(2);
 	else
@@ -1016,7 +1031,7 @@ static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
 	 * sclk = system clock(Mhz)
 	 */
 	a.full = dfixed_const(600 * 1000);
-	chunk_time.full = dfixed_div(a, rdev->pm.sclk);
+	chunk_time.full = dfixed_div(a, sclk);
 	read_delay_latency.full = dfixed_const(1000);
 
 	/* Determine the worst case latency
@@ -1077,152 +1092,177 @@ static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
 	}
 }
 
-void rv515_bandwidth_avivo_update(struct radeon_device *rdev)
+static void rv515_compute_mode_priority(struct radeon_device *rdev,
+					struct rv515_watermark *wm0,
+					struct rv515_watermark *wm1,
+					struct drm_display_mode *mode0,
+					struct drm_display_mode *mode1,
+					u32 *d1mode_priority_a_cnt,
+					u32 *d2mode_priority_a_cnt)
 {
-	struct drm_display_mode *mode0 = NULL;
-	struct drm_display_mode *mode1 = NULL;
-	struct rv515_watermark wm0;
-	struct rv515_watermark wm1;
-	u32 tmp;
-	u32 d1mode_priority_a_cnt = MODE_PRIORITY_OFF;
-	u32 d2mode_priority_a_cnt = MODE_PRIORITY_OFF;
 	fixed20_12 priority_mark02, priority_mark12, fill_rate;
 	fixed20_12 a, b;
 
-	if (rdev->mode_info.crtcs[0]->base.enabled)
-		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
-	if (rdev->mode_info.crtcs[1]->base.enabled)
-		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
-	rs690_line_buffer_adjust(rdev, mode0, mode1);
-
-	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0);
-	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1);
-
-	tmp = wm0.lb_request_fifo_depth;
-	tmp |= wm1.lb_request_fifo_depth << 16;
-	WREG32(LB_MAX_REQ_OUTSTANDING, tmp);
+	*d1mode_priority_a_cnt = MODE_PRIORITY_OFF;
+	*d2mode_priority_a_cnt = MODE_PRIORITY_OFF;
 
 	if (mode0 && mode1) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_div(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_div(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			b.full = dfixed_div(wm1.dbpp, wm1.num_line_pair);
+			a.full = wm0->num_line_pair.full;
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			b.full = dfixed_div(wm1->dbpp, wm1->num_line_pair);
 		else
-			b.full = wm1.num_line_pair.full;
+			b.full = wm1->num_line_pair.full;
 		a.full += b.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			priority_mark02.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			priority_mark12.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2) {
-			d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
-			d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
 		}
 	} else if (mode0) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_div(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_div(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
+			a.full = wm0->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			priority_mark02.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
 		if (rdev->disp_priority == 2)
-			d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
 	} else if (mode1) {
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			a.full = dfixed_div(wm1.dbpp, wm1.num_line_pair);
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			a.full = dfixed_div(wm1->dbpp, wm1->num_line_pair);
 		else
-			a.full = wm1.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm1.sclk, a);
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
+			a.full = wm1->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm1->sclk, a);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			priority_mark12.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2)
-			d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
 	}
+}
+
+void rv515_bandwidth_avivo_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode0 = NULL;
+	struct drm_display_mode *mode1 = NULL;
+	struct rv515_watermark wm0_high, wm0_low;
+	struct rv515_watermark wm1_high, wm1_low;
+	u32 tmp;
+	u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt;
+	u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt;
+
+	if (rdev->mode_info.crtcs[0]->base.enabled)
+		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
+	if (rdev->mode_info.crtcs[1]->base.enabled)
+		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
+	rs690_line_buffer_adjust(rdev, mode0, mode1);
+
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_high, false);
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_high, false);
+
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_low, false);
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_low, false);
+
+	tmp = wm0_high.lb_request_fifo_depth;
+	tmp |= wm1_high.lb_request_fifo_depth << 16;
+	WREG32(LB_MAX_REQ_OUTSTANDING, tmp);
+
+	rv515_compute_mode_priority(rdev,
+				    &wm0_high, &wm1_high,
+				    mode0, mode1,
+				    &d1mode_priority_a_cnt, &d2mode_priority_a_cnt);
+	rv515_compute_mode_priority(rdev,
+				    &wm0_low, &wm1_low,
+				    mode0, mode1,
+				    &d1mode_priority_b_cnt, &d2mode_priority_b_cnt);
 
 	WREG32(D1MODE_PRIORITY_A_CNT, d1mode_priority_a_cnt);
-	WREG32(D1MODE_PRIORITY_B_CNT, d1mode_priority_a_cnt);
+	WREG32(D1MODE_PRIORITY_B_CNT, d1mode_priority_b_cnt);
 	WREG32(D2MODE_PRIORITY_A_CNT, d2mode_priority_a_cnt);
-	WREG32(D2MODE_PRIORITY_B_CNT, d2mode_priority_a_cnt);
+	WREG32(D2MODE_PRIORITY_B_CNT, d2mode_priority_b_cnt);
 }
 
 void rv515_bandwidth_update(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c
new file mode 100644
index 0000000..8303de2
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c
@@ -0,0 +1,2085 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv6xxd.h"
+#include "r600_dpm.h"
+#include "rv6xx_dpm.h"
+#include "atom.h"
+#include <linux/seq_file.h>
+
+static u32 rv6xx_scale_count_given_unit(struct radeon_device *rdev,
+					u32 unscaled_count, u32 unit);
+
+static struct rv6xx_ps *rv6xx_get_ps(struct radeon_ps *rps)
+{
+	struct rv6xx_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+static struct rv6xx_power_info *rv6xx_get_pi(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void rv6xx_force_pcie_gen1(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int i;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	tmp &= LC_GEN2_EN;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	tmp |= LC_INITIATE_LINK_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (!(RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & LC_CURRENT_DATA_RATE))
+			break;
+		udelay(1);
+	}
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	tmp &= ~LC_INITIATE_LINK_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+}
+
+static void rv6xx_enable_pcie_gen2_support(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+		tmp |= LC_GEN2_EN;
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+	}
+}
+
+static void rv6xx_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					       bool enable)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+	if (enable)
+		tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+	else
+		tmp |= LC_HW_VOLTAGE_IF_CONTROL(0);
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+}
+
+static void rv6xx_enable_l0s(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L0S_INACTIVITY_MASK;
+	tmp |= LC_L0S_INACTIVITY(3);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv6xx_enable_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	tmp &= ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(4);
+	tmp &= ~LC_PMI_TO_L1_DIS;
+	tmp &= ~LC_ASPM_TO_L1_DIS;
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv6xx_enable_pll_sleep_in_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(8);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+
+	/* NOTE, this is a PCIE indirect reg, not PCIE PORT */
+	tmp = RREG32_PCIE(PCIE_P_CNTL);
+	tmp |= P_PLL_PWRDN_IN_L1L23;
+	tmp &= ~P_PLL_BUF_PDNB;
+	tmp &= ~P_PLL_PDNB;
+	tmp |= P_ALLOW_PRX_FRONTEND_SHUTOFF;
+	WREG32_PCIE(PCIE_P_CNTL, tmp);
+}
+
+static int rv6xx_convert_clock_to_stepping(struct radeon_device *rdev,
+					   u32 clock, struct rv6xx_sclk_stepping *step)
+{
+	int ret;
+	struct atom_clock_dividers dividers;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	if (dividers.enable_post_div)
+		step->post_divider = 2 + (dividers.post_div & 0xF) + (dividers.post_div >> 4);
+	else
+		step->post_divider = 1;
+
+	step->vco_frequency = clock * step->post_divider;
+
+	return 0;
+}
+
+static void rv6xx_output_stepping(struct radeon_device *rdev,
+				  u32 step_index, struct rv6xx_sclk_stepping *step)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+	u32 fb_divider;
+	u32 spll_step_count = rv6xx_scale_count_given_unit(rdev,
+							   R600_SPLLSTEPTIME_DFLT *
+							   pi->spll_ref_div,
+							   R600_SPLLSTEPUNIT_DFLT);
+
+	r600_engine_clock_entry_enable(rdev, step_index, true);
+	r600_engine_clock_entry_enable_pulse_skipping(rdev, step_index, false);
+
+	if (step->post_divider == 1)
+		r600_engine_clock_entry_enable_post_divider(rdev, step_index, false);
+	else {
+		u32 lo_len = (step->post_divider - 2) / 2;
+		u32 hi_len = step->post_divider - 2 - lo_len;
+
+		r600_engine_clock_entry_enable_post_divider(rdev, step_index, true);
+		r600_engine_clock_entry_set_post_divider(rdev, step_index, (hi_len << 4) | lo_len);
+	}
+
+	fb_divider = ((step->vco_frequency * pi->spll_ref_div) / ref_clk) >>
+		pi->fb_div_scale;
+
+	r600_engine_clock_entry_set_reference_divider(rdev, step_index,
+						      pi->spll_ref_div - 1);
+	r600_engine_clock_entry_set_feedback_divider(rdev, step_index, fb_divider);
+	r600_engine_clock_entry_set_step_time(rdev, step_index, spll_step_count);
+
+}
+
+static struct rv6xx_sclk_stepping rv6xx_next_vco_step(struct radeon_device *rdev,
+						      struct rv6xx_sclk_stepping *cur,
+						      bool increasing_vco, u32 step_size)
+{
+	struct rv6xx_sclk_stepping next;
+
+	next.post_divider = cur->post_divider;
+
+	if (increasing_vco)
+		next.vco_frequency = (cur->vco_frequency * (100 + step_size)) / 100;
+	else
+		next.vco_frequency = (cur->vco_frequency * 100 + 99 + step_size) / (100 + step_size);
+
+	return next;
+}
+
+static bool rv6xx_can_step_post_div(struct radeon_device *rdev,
+				    struct rv6xx_sclk_stepping *cur,
+                                    struct rv6xx_sclk_stepping *target)
+{
+	return (cur->post_divider > target->post_divider) &&
+		((cur->vco_frequency * target->post_divider) <=
+		 (target->vco_frequency * (cur->post_divider - 1)));
+}
+
+static struct rv6xx_sclk_stepping rv6xx_next_post_div_step(struct radeon_device *rdev,
+							   struct rv6xx_sclk_stepping *cur,
+							   struct rv6xx_sclk_stepping *target)
+{
+	struct rv6xx_sclk_stepping next = *cur;
+
+	while (rv6xx_can_step_post_div(rdev, &next, target))
+		next.post_divider--;
+
+	return next;
+}
+
+static bool rv6xx_reached_stepping_target(struct radeon_device *rdev,
+					  struct rv6xx_sclk_stepping *cur,
+					  struct rv6xx_sclk_stepping *target,
+					  bool increasing_vco)
+{
+	return (increasing_vco && (cur->vco_frequency >= target->vco_frequency)) ||
+		(!increasing_vco && (cur->vco_frequency <= target->vco_frequency));
+}
+
+static void rv6xx_generate_steps(struct radeon_device *rdev,
+				 u32 low, u32 high,
+                                 u32 start_index, u8 *end_index)
+{
+	struct rv6xx_sclk_stepping cur;
+	struct rv6xx_sclk_stepping target;
+	bool increasing_vco;
+	u32 step_index = start_index;
+
+	rv6xx_convert_clock_to_stepping(rdev, low, &cur);
+	rv6xx_convert_clock_to_stepping(rdev, high, &target);
+
+	rv6xx_output_stepping(rdev, step_index++, &cur);
+
+	increasing_vco = (target.vco_frequency >= cur.vco_frequency);
+
+	if (target.post_divider > cur.post_divider)
+		cur.post_divider = target.post_divider;
+
+	while (1) {
+		struct rv6xx_sclk_stepping next;
+
+		if (rv6xx_can_step_post_div(rdev, &cur, &target))
+			next = rv6xx_next_post_div_step(rdev, &cur, &target);
+		else
+			next = rv6xx_next_vco_step(rdev, &cur, increasing_vco, R600_VCOSTEPPCT_DFLT);
+
+		if (rv6xx_reached_stepping_target(rdev, &next, &target, increasing_vco)) {
+			struct rv6xx_sclk_stepping tiny =
+				rv6xx_next_vco_step(rdev, &target, !increasing_vco, R600_ENDINGVCOSTEPPCT_DFLT);
+			tiny.post_divider = next.post_divider;
+
+			if (!rv6xx_reached_stepping_target(rdev, &tiny, &cur, !increasing_vco))
+				rv6xx_output_stepping(rdev, step_index++, &tiny);
+
+			if ((next.post_divider != target.post_divider) &&
+			    (next.vco_frequency != target.vco_frequency)) {
+				struct rv6xx_sclk_stepping final_vco;
+
+				final_vco.vco_frequency = target.vco_frequency;
+				final_vco.post_divider = next.post_divider;
+
+				rv6xx_output_stepping(rdev, step_index++, &final_vco);
+			}
+
+			rv6xx_output_stepping(rdev, step_index++, &target);
+			break;
+		} else
+			rv6xx_output_stepping(rdev, step_index++, &next);
+
+		cur = next;
+	}
+
+	*end_index = (u8)step_index - 1;
+
+}
+
+static void rv6xx_generate_single_step(struct radeon_device *rdev,
+				       u32 clock, u32 index)
+{
+	struct rv6xx_sclk_stepping step;
+
+	rv6xx_convert_clock_to_stepping(rdev, clock, &step);
+	rv6xx_output_stepping(rdev, index, &step);
+}
+
+static void rv6xx_invalidate_intermediate_steps_range(struct radeon_device *rdev,
+						      u32 start_index, u32 end_index)
+{
+	u32 step_index;
+
+	for (step_index = start_index + 1; step_index < end_index; step_index++)
+		r600_engine_clock_entry_enable(rdev, step_index, false);
+}
+
+static void rv6xx_set_engine_spread_spectrum_clk_s(struct radeon_device *rdev,
+						   u32 index, u32 clk_s)
+{
+	WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+		 CLKS(clk_s), ~CLKS_MASK);
+}
+
+static void rv6xx_set_engine_spread_spectrum_clk_v(struct radeon_device *rdev,
+						   u32 index, u32 clk_v)
+{
+	WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+		 CLKV(clk_v), ~CLKV_MASK);
+}
+
+static void rv6xx_enable_engine_spread_spectrum(struct radeon_device *rdev,
+						u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+			 SSEN, ~SSEN);
+	else
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+			 0, ~SSEN);
+}
+
+static void rv6xx_set_memory_spread_spectrum_clk_s(struct radeon_device *rdev,
+						   u32 clk_s)
+{
+	WREG32_P(CG_MPLL_SPREAD_SPECTRUM, CLKS(clk_s), ~CLKS_MASK);
+}
+
+static void rv6xx_set_memory_spread_spectrum_clk_v(struct radeon_device *rdev,
+						   u32 clk_v)
+{
+	WREG32_P(CG_MPLL_SPREAD_SPECTRUM, CLKV(clk_v), ~CLKV_MASK);
+}
+
+static void rv6xx_enable_memory_spread_spectrum(struct radeon_device *rdev,
+						bool enable)
+{
+	if (enable)
+		WREG32_P(CG_MPLL_SPREAD_SPECTRUM, SSEN, ~SSEN);
+	else
+		WREG32_P(CG_MPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+}
+
+static void rv6xx_enable_dynamic_spread_spectrum(struct radeon_device *rdev,
+						 bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+}
+
+static void rv6xx_memory_clock_entry_enable_post_divider(struct radeon_device *rdev,
+							 u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4),
+			 LEVEL0_MPLL_DIV_EN, ~LEVEL0_MPLL_DIV_EN);
+	else
+		WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4), 0, ~LEVEL0_MPLL_DIV_EN);
+}
+
+static void rv6xx_memory_clock_entry_set_post_divider(struct radeon_device *rdev,
+						      u32 index, u32 divider)
+{
+	WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4),
+		 LEVEL0_MPLL_POST_DIV(divider), ~LEVEL0_MPLL_POST_DIV_MASK);
+}
+
+static void rv6xx_memory_clock_entry_set_feedback_divider(struct radeon_device *rdev,
+							  u32 index, u32 divider)
+{
+	WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4), LEVEL0_MPLL_FB_DIV(divider),
+		 ~LEVEL0_MPLL_FB_DIV_MASK);
+}
+
+static void rv6xx_memory_clock_entry_set_reference_divider(struct radeon_device *rdev,
+							   u32 index, u32 divider)
+{
+	WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4),
+		 LEVEL0_MPLL_REF_DIV(divider), ~LEVEL0_MPLL_REF_DIV_MASK);
+}
+
+static void rv6xx_vid_response_set_brt(struct radeon_device *rdev, u32 rt)
+{
+	WREG32_P(VID_RT, BRT(rt), ~BRT_MASK);
+}
+
+static void rv6xx_enable_engine_feedback_and_reference_sync(struct radeon_device *rdev)
+{
+	WREG32_P(SPLL_CNTL_MODE, SPLL_DIV_SYNC, ~SPLL_DIV_SYNC);
+}
+
+static u64 rv6xx_clocks_per_unit(u32 unit)
+{
+	u64 tmp = 1 << (2 * unit);
+
+	return tmp;
+}
+
+static u32 rv6xx_scale_count_given_unit(struct radeon_device *rdev,
+					u32 unscaled_count, u32 unit)
+{
+	u32 count_per_unit = (u32)rv6xx_clocks_per_unit(unit);
+
+	return (unscaled_count + count_per_unit - 1) / count_per_unit;
+}
+
+static u32 rv6xx_compute_count_for_delay(struct radeon_device *rdev,
+					 u32 delay_us, u32 unit)
+{
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+
+	return rv6xx_scale_count_given_unit(rdev, delay_us * (ref_clk / 100), unit);
+}
+
+static void rv6xx_calculate_engine_speed_stepping_parameters(struct radeon_device *rdev,
+							     struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.sclks[R600_POWER_LEVEL_LOW] =
+		state->low.sclk;
+	pi->hw.sclks[R600_POWER_LEVEL_MEDIUM] =
+		state->medium.sclk;
+	pi->hw.sclks[R600_POWER_LEVEL_HIGH] =
+		state->high.sclk;
+
+	pi->hw.low_sclk_index = R600_POWER_LEVEL_LOW;
+	pi->hw.medium_sclk_index = R600_POWER_LEVEL_MEDIUM;
+	pi->hw.high_sclk_index = R600_POWER_LEVEL_HIGH;
+}
+
+static void rv6xx_calculate_memory_clock_stepping_parameters(struct radeon_device *rdev,
+							     struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.mclks[R600_POWER_LEVEL_CTXSW] =
+		state->high.mclk;
+	pi->hw.mclks[R600_POWER_LEVEL_HIGH] =
+		state->high.mclk;
+	pi->hw.mclks[R600_POWER_LEVEL_MEDIUM] =
+		state->medium.mclk;
+	pi->hw.mclks[R600_POWER_LEVEL_LOW] =
+		state->low.mclk;
+
+	pi->hw.high_mclk_index = R600_POWER_LEVEL_HIGH;
+
+	if (state->high.mclk == state->medium.mclk)
+		pi->hw.medium_mclk_index =
+			pi->hw.high_mclk_index;
+	else
+		pi->hw.medium_mclk_index = R600_POWER_LEVEL_MEDIUM;
+
+
+	if (state->medium.mclk == state->low.mclk)
+		pi->hw.low_mclk_index =
+			pi->hw.medium_mclk_index;
+	else
+		pi->hw.low_mclk_index = R600_POWER_LEVEL_LOW;
+}
+
+static void rv6xx_calculate_voltage_stepping_parameters(struct radeon_device *rdev,
+							struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.vddc[R600_POWER_LEVEL_CTXSW] = state->high.vddc;
+	pi->hw.vddc[R600_POWER_LEVEL_HIGH] = state->high.vddc;
+	pi->hw.vddc[R600_POWER_LEVEL_MEDIUM] = state->medium.vddc;
+	pi->hw.vddc[R600_POWER_LEVEL_LOW] = state->low.vddc;
+
+	pi->hw.backbias[R600_POWER_LEVEL_CTXSW] =
+		(state->high.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+	pi->hw.backbias[R600_POWER_LEVEL_HIGH] =
+		(state->high.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+	pi->hw.backbias[R600_POWER_LEVEL_MEDIUM] =
+		(state->medium.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+	pi->hw.backbias[R600_POWER_LEVEL_LOW] =
+		(state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+
+	pi->hw.pcie_gen2[R600_POWER_LEVEL_HIGH] =
+		(state->high.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? true : false;
+	pi->hw.pcie_gen2[R600_POWER_LEVEL_MEDIUM] =
+		(state->medium.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? true : false;
+	pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW] =
+		(state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? true : false;
+
+	pi->hw.high_vddc_index = R600_POWER_LEVEL_HIGH;
+
+	if ((state->high.vddc == state->medium.vddc) &&
+	    ((state->high.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ==
+	     (state->medium.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE)))
+		pi->hw.medium_vddc_index =
+			pi->hw.high_vddc_index;
+	else
+		pi->hw.medium_vddc_index = R600_POWER_LEVEL_MEDIUM;
+
+	if ((state->medium.vddc == state->low.vddc) &&
+	    ((state->medium.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ==
+	     (state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE)))
+		pi->hw.low_vddc_index =
+			pi->hw.medium_vddc_index;
+	else
+		pi->hw.medium_vddc_index = R600_POWER_LEVEL_LOW;
+}
+
+static inline u32 rv6xx_calculate_vco_frequency(u32 ref_clock,
+						struct atom_clock_dividers *dividers,
+						u32 fb_divider_scale)
+{
+	return ref_clock * ((dividers->fb_div & ~1) << fb_divider_scale) /
+		(dividers->ref_div + 1);
+}
+
+static inline u32 rv6xx_calculate_spread_spectrum_clk_v(u32 vco_freq, u32 ref_freq,
+							u32 ss_rate, u32 ss_percent,
+							u32 fb_divider_scale)
+{
+	u32 fb_divider = vco_freq / ref_freq;
+
+	return (ss_percent * ss_rate * 4 * (fb_divider * fb_divider) /
+		(5375 * ((vco_freq * 10) / (4096 >> fb_divider_scale))));
+}
+
+static inline u32 rv6xx_calculate_spread_spectrum_clk_s(u32 ss_rate, u32 ref_freq)
+{
+	return (((ref_freq * 10) / (ss_rate * 2)) - 1) / 4;
+}
+
+static void rv6xx_program_engine_spread_spectrum(struct radeon_device *rdev,
+						 u32 clock, enum r600_power_level level)
+{
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	struct radeon_atom_ss ss;
+	u32 vco_freq, clk_v, clk_s;
+
+	rv6xx_enable_engine_spread_spectrum(rdev, level, false);
+
+	if (clock && pi->sclk_ss) {
+		if (radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, clock, false, &dividers) == 0) {
+			vco_freq = rv6xx_calculate_vco_frequency(ref_clk, &dividers,
+								 pi->fb_div_scale);
+
+			if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+							     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+				clk_v = rv6xx_calculate_spread_spectrum_clk_v(vco_freq,
+									      (ref_clk / (dividers.ref_div + 1)),
+									      ss.rate,
+									      ss.percentage,
+									      pi->fb_div_scale);
+
+				clk_s = rv6xx_calculate_spread_spectrum_clk_s(ss.rate,
+									      (ref_clk / (dividers.ref_div + 1)));
+
+				rv6xx_set_engine_spread_spectrum_clk_v(rdev, level, clk_v);
+				rv6xx_set_engine_spread_spectrum_clk_s(rdev, level, clk_s);
+				rv6xx_enable_engine_spread_spectrum(rdev, level, true);
+			}
+		}
+	}
+}
+
+static void rv6xx_program_sclk_spread_spectrum_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_engine_spread_spectrum(rdev,
+					     pi->hw.sclks[R600_POWER_LEVEL_HIGH],
+					     R600_POWER_LEVEL_HIGH);
+
+	rv6xx_program_engine_spread_spectrum(rdev,
+					     pi->hw.sclks[R600_POWER_LEVEL_MEDIUM],
+					     R600_POWER_LEVEL_MEDIUM);
+
+}
+
+static int rv6xx_program_mclk_stepping_entry(struct radeon_device *rdev,
+					     u32 entry, u32 clock)
+{
+	struct atom_clock_dividers dividers;
+
+	if (radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM, clock, false, &dividers))
+	    return -EINVAL;
+
+
+	rv6xx_memory_clock_entry_set_reference_divider(rdev, entry, dividers.ref_div);
+	rv6xx_memory_clock_entry_set_feedback_divider(rdev, entry, dividers.fb_div);
+	rv6xx_memory_clock_entry_set_post_divider(rdev, entry, dividers.post_div);
+
+	if (dividers.enable_post_div)
+		rv6xx_memory_clock_entry_enable_post_divider(rdev, entry, true);
+	else
+		rv6xx_memory_clock_entry_enable_post_divider(rdev, entry, false);
+
+	return 0;
+}
+
+static void rv6xx_program_mclk_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	int i;
+
+	for (i = 1; i < R600_PM_NUMBER_OF_MCLKS; i++) {
+		if (pi->hw.mclks[i])
+			rv6xx_program_mclk_stepping_entry(rdev, i,
+							  pi->hw.mclks[i]);
+	}
+}
+
+static void rv6xx_find_memory_clock_with_highest_vco(struct radeon_device *rdev,
+						     u32 requested_memory_clock,
+						     u32 ref_clk,
+						     struct atom_clock_dividers *dividers,
+						     u32 *vco_freq)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct atom_clock_dividers req_dividers;
+	u32 vco_freq_temp;
+
+	if (radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					   requested_memory_clock, false, &req_dividers) == 0) {
+		vco_freq_temp = rv6xx_calculate_vco_frequency(ref_clk, &req_dividers,
+							      pi->fb_div_scale);
+
+		if (vco_freq_temp > *vco_freq) {
+			*dividers = req_dividers;
+			*vco_freq = vco_freq_temp;
+		}
+	}
+}
+
+static void rv6xx_program_mclk_spread_spectrum_parameters(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 ref_clk = rdev->clock.mpll.reference_freq;
+	struct atom_clock_dividers dividers;
+	struct radeon_atom_ss ss;
+	u32 vco_freq = 0, clk_v, clk_s;
+
+	rv6xx_enable_memory_spread_spectrum(rdev, false);
+
+	if (pi->mclk_ss) {
+		rv6xx_find_memory_clock_with_highest_vco(rdev,
+							 pi->hw.mclks[pi->hw.high_mclk_index],
+							 ref_clk,
+							 &dividers,
+							 &vco_freq);
+
+		rv6xx_find_memory_clock_with_highest_vco(rdev,
+							 pi->hw.mclks[pi->hw.medium_mclk_index],
+							 ref_clk,
+							 &dividers,
+							 &vco_freq);
+
+		rv6xx_find_memory_clock_with_highest_vco(rdev,
+							 pi->hw.mclks[pi->hw.low_mclk_index],
+							 ref_clk,
+							 &dividers,
+							 &vco_freq);
+
+		if (vco_freq) {
+			if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+							     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+				clk_v = rv6xx_calculate_spread_spectrum_clk_v(vco_freq,
+									     (ref_clk / (dividers.ref_div + 1)),
+									     ss.rate,
+									     ss.percentage,
+									     pi->fb_div_scale);
+
+				clk_s = rv6xx_calculate_spread_spectrum_clk_s(ss.rate,
+									     (ref_clk / (dividers.ref_div + 1)));
+
+				rv6xx_set_memory_spread_spectrum_clk_v(rdev, clk_v);
+				rv6xx_set_memory_spread_spectrum_clk_s(rdev, clk_s);
+				rv6xx_enable_memory_spread_spectrum(rdev, true);
+			}
+		}
+	}
+}
+
+static int rv6xx_program_voltage_stepping_entry(struct radeon_device *rdev,
+						u32 entry, u16 voltage)
+{
+	u32 mask, set_pins;
+	int ret;
+
+	ret = radeon_atom_get_voltage_gpio_settings(rdev, voltage,
+						    SET_VOLTAGE_TYPE_ASIC_VDDC,
+						    &set_pins, &mask);
+	if (ret)
+		return ret;
+
+	r600_voltage_control_program_voltages(rdev, entry, set_pins);
+
+	return 0;
+}
+
+static void rv6xx_program_voltage_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	int i;
+
+	for (i = 1; i < R600_PM_NUMBER_OF_VOLTAGE_LEVELS; i++)
+		rv6xx_program_voltage_stepping_entry(rdev, i,
+						     pi->hw.vddc[i]);
+
+}
+
+static void rv6xx_program_backbias_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->hw.backbias[1])
+		WREG32_P(VID_UPPER_GPIO_CNTL, MEDIUM_BACKBIAS_VALUE, ~MEDIUM_BACKBIAS_VALUE);
+	else
+		WREG32_P(VID_UPPER_GPIO_CNTL, 0, ~MEDIUM_BACKBIAS_VALUE);
+
+	if (pi->hw.backbias[2])
+		WREG32_P(VID_UPPER_GPIO_CNTL, HIGH_BACKBIAS_VALUE, ~HIGH_BACKBIAS_VALUE);
+	else
+		WREG32_P(VID_UPPER_GPIO_CNTL, 0, ~HIGH_BACKBIAS_VALUE);
+}
+
+static void rv6xx_program_sclk_spread_spectrum_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_engine_spread_spectrum(rdev,
+					     pi->hw.sclks[R600_POWER_LEVEL_LOW],
+					     R600_POWER_LEVEL_LOW);
+}
+
+static void rv6xx_program_mclk_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->hw.mclks[0])
+		rv6xx_program_mclk_stepping_entry(rdev, 0,
+						  pi->hw.mclks[0]);
+}
+
+static void rv6xx_program_voltage_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_voltage_stepping_entry(rdev, 0,
+					     pi->hw.vddc[0]);
+
+}
+
+static void rv6xx_program_backbias_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->hw.backbias[0])
+		WREG32_P(VID_UPPER_GPIO_CNTL, LOW_BACKBIAS_VALUE, ~LOW_BACKBIAS_VALUE);
+	else
+		WREG32_P(VID_UPPER_GPIO_CNTL, 0, ~LOW_BACKBIAS_VALUE);
+}
+
+static u32 calculate_memory_refresh_rate(struct radeon_device *rdev,
+					 u32 engine_clock)
+{
+	u32 dram_rows, dram_refresh_rate;
+	u32 tmp;
+
+	tmp = (RREG32(RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT;
+	dram_rows = 1 << (tmp + 10);
+	dram_refresh_rate = 1 << ((RREG32(MC_SEQ_RESERVE_M) & 0x3) + 3);
+
+	return ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64;
+}
+
+static void rv6xx_program_memory_timing_parameters(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 sqm_ratio;
+	u32 arb_refresh_rate;
+	u32 high_clock;
+
+	if (pi->hw.sclks[R600_POWER_LEVEL_HIGH] <
+	    (pi->hw.sclks[R600_POWER_LEVEL_LOW] * 0xFF / 0x40))
+		high_clock = pi->hw.sclks[R600_POWER_LEVEL_HIGH];
+	else
+		high_clock =
+			pi->hw.sclks[R600_POWER_LEVEL_LOW] * 0xFF / 0x40;
+
+	radeon_atom_set_engine_dram_timings(rdev, high_clock, 0);
+
+	sqm_ratio = (STATE0(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_LOW]) |
+		     STATE1(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_MEDIUM]) |
+		     STATE2(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_HIGH]) |
+		     STATE3(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_HIGH]));
+	WREG32(SQM_RATIO, sqm_ratio);
+
+	arb_refresh_rate =
+		(POWERMODE0(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_LOW])) |
+		 POWERMODE1(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_MEDIUM])) |
+		 POWERMODE2(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_MEDIUM])) |
+		 POWERMODE3(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_HIGH])));
+	WREG32(ARB_RFSH_RATE, arb_refresh_rate);
+}
+
+static void rv6xx_program_mpll_timing_parameters(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_set_mpll_lock_time(rdev, R600_MPLLLOCKTIME_DFLT *
+				pi->mpll_ref_div);
+	r600_set_mpll_reset_time(rdev, R600_MPLLRESETTIME_DFLT);
+}
+
+static void rv6xx_program_bsp(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+
+	r600_calculate_u_and_p(R600_ASI_DFLT,
+			       ref_clk, 16,
+			       &pi->bsp,
+			       &pi->bsu);
+
+	r600_set_bsp(rdev, pi->bsu, pi->bsp);
+}
+
+static void rv6xx_program_at(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_set_at(rdev,
+		    (pi->hw.rp[0] * pi->bsp) / 200,
+		    (pi->hw.rp[1] * pi->bsp) / 200,
+		    (pi->hw.lp[2] * pi->bsp) / 200,
+		    (pi->hw.lp[1] * pi->bsp) / 200);
+}
+
+static void rv6xx_program_git(struct radeon_device *rdev)
+{
+	r600_set_git(rdev, R600_GICST_DFLT);
+}
+
+static void rv6xx_program_tp(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		r600_set_tc(rdev, i, r600_utc[i], r600_dtc[i]);
+
+	r600_select_td(rdev, R600_TD_DFLT);
+}
+
+static void rv6xx_program_vc(struct radeon_device *rdev)
+{
+	r600_set_vrc(rdev, R600_VRC_DFLT);
+}
+
+static void rv6xx_clear_vc(struct radeon_device *rdev)
+{
+	r600_set_vrc(rdev, 0);
+}
+
+static void rv6xx_program_tpp(struct radeon_device *rdev)
+{
+	r600_set_tpu(rdev, R600_TPU_DFLT);
+	r600_set_tpc(rdev, R600_TPC_DFLT);
+}
+
+static void rv6xx_program_sstp(struct radeon_device *rdev)
+{
+	r600_set_sstu(rdev, R600_SSTU_DFLT);
+	r600_set_sst(rdev, R600_SST_DFLT);
+}
+
+static void rv6xx_program_fcp(struct radeon_device *rdev)
+{
+	r600_set_fctu(rdev, R600_FCTU_DFLT);
+	r600_set_fct(rdev, R600_FCT_DFLT);
+}
+
+static void rv6xx_program_vddc3d_parameters(struct radeon_device *rdev)
+{
+	r600_set_vddc3d_oorsu(rdev, R600_VDDC3DOORSU_DFLT);
+	r600_set_vddc3d_oorphc(rdev, R600_VDDC3DOORPHC_DFLT);
+	r600_set_vddc3d_oorsdc(rdev, R600_VDDC3DOORSDC_DFLT);
+	r600_set_ctxcgtt3d_rphc(rdev, R600_CTXCGTT3DRPHC_DFLT);
+	r600_set_ctxcgtt3d_rsdc(rdev, R600_CTXCGTT3DRSDC_DFLT);
+}
+
+static void rv6xx_program_voltage_timing_parameters(struct radeon_device *rdev)
+{
+	u32 rt;
+
+	r600_vid_rt_set_vru(rdev, R600_VRU_DFLT);
+
+	r600_vid_rt_set_vrt(rdev,
+			    rv6xx_compute_count_for_delay(rdev,
+							  rdev->pm.dpm.voltage_response_time,
+							  R600_VRU_DFLT));
+
+	rt = rv6xx_compute_count_for_delay(rdev,
+					   rdev->pm.dpm.backbias_response_time,
+					   R600_VRU_DFLT);
+
+	rv6xx_vid_response_set_brt(rdev, (rt + 0x1F) >> 5);
+}
+
+static void rv6xx_program_engine_speed_parameters(struct radeon_device *rdev)
+{
+	r600_vid_rt_set_ssu(rdev, R600_SPLLSTEPUNIT_DFLT);
+	rv6xx_enable_engine_feedback_and_reference_sync(rdev);
+}
+
+static u64 rv6xx_get_master_voltage_mask(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u64 master_mask = 0;
+	int i;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_VOLTAGE_LEVELS; i++) {
+		u32 tmp_mask, tmp_set_pins;
+		int ret;
+
+		ret = radeon_atom_get_voltage_gpio_settings(rdev,
+							    pi->hw.vddc[i],
+							    SET_VOLTAGE_TYPE_ASIC_VDDC,
+							    &tmp_set_pins, &tmp_mask);
+
+		if (ret == 0)
+			master_mask |= tmp_mask;
+	}
+
+	return master_mask;
+}
+
+static void rv6xx_program_voltage_gpio_pins(struct radeon_device *rdev)
+{
+	r600_voltage_control_enable_pins(rdev,
+					 rv6xx_get_master_voltage_mask(rdev));
+}
+
+static void rv6xx_enable_static_voltage_control(struct radeon_device *rdev,
+						struct radeon_ps *new_ps,
+						bool enable)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+
+	if (enable)
+		radeon_atom_set_voltage(rdev,
+					new_state->low.vddc,
+					SET_VOLTAGE_TYPE_ASIC_VDDC);
+	else
+		r600_voltage_control_deactivate_static_control(rdev,
+							       rv6xx_get_master_voltage_mask(rdev));
+}
+
+static void rv6xx_enable_display_gap(struct radeon_device *rdev, bool enable)
+{
+	if (enable) {
+		u32 tmp = (DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM) |
+			   DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM) |
+			   DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+			   DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+			   VBI_TIMER_COUNT(0x3FFF) |
+			   VBI_TIMER_UNIT(7));
+		WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+
+		WREG32_P(MCLK_PWRMGT_CNTL, USE_DISPLAY_GAP, ~USE_DISPLAY_GAP);
+	} else
+		WREG32_P(MCLK_PWRMGT_CNTL, 0, ~USE_DISPLAY_GAP);
+}
+
+static void rv6xx_program_power_level_enter_state(struct radeon_device *rdev)
+{
+	r600_power_level_set_enter_index(rdev, R600_POWER_LEVEL_MEDIUM);
+}
+
+static void rv6xx_calculate_t(u32 l_f, u32 h_f, int h,
+			      int d_l, int d_r, u8 *l, u8 *r)
+{
+	int a_n, a_d, h_r, l_r;
+
+	h_r = d_l;
+	l_r = 100 - d_r;
+
+	a_n = (int)h_f * d_l + (int)l_f * (h - d_r);
+	a_d = (int)l_f * l_r + (int)h_f * h_r;
+
+	if (a_d != 0) {
+		*l = d_l - h_r * a_n / a_d;
+		*r = d_r + l_r * a_n / a_d;
+	}
+}
+
+static void rv6xx_calculate_ap(struct radeon_device *rdev,
+			       struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.lp[0] = 0;
+	pi->hw.rp[R600_PM_NUMBER_OF_ACTIVITY_LEVELS - 1]
+		= 100;
+
+	rv6xx_calculate_t(state->low.sclk,
+			  state->medium.sclk,
+			  R600_AH_DFLT,
+			  R600_LMP_DFLT,
+			  R600_RLP_DFLT,
+			  &pi->hw.lp[1],
+			  &pi->hw.rp[0]);
+
+	rv6xx_calculate_t(state->medium.sclk,
+			  state->high.sclk,
+			  R600_AH_DFLT,
+			  R600_LHP_DFLT,
+			  R600_RMP_DFLT,
+			  &pi->hw.lp[2],
+			  &pi->hw.rp[1]);
+
+}
+
+static void rv6xx_calculate_stepping_parameters(struct radeon_device *rdev,
+						struct radeon_ps *new_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+
+	rv6xx_calculate_engine_speed_stepping_parameters(rdev, new_state);
+	rv6xx_calculate_memory_clock_stepping_parameters(rdev, new_state);
+	rv6xx_calculate_voltage_stepping_parameters(rdev, new_state);
+	rv6xx_calculate_ap(rdev, new_state);
+}
+
+static void rv6xx_program_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_mclk_stepping_parameters_except_lowest_entry(rdev);
+	if (pi->voltage_control)
+		rv6xx_program_voltage_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_backbias_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_sclk_spread_spectrum_parameters_except_lowest_entry(rdev);
+	rv6xx_program_mclk_spread_spectrum_parameters(rdev);
+	rv6xx_program_memory_timing_parameters(rdev);
+}
+
+static void rv6xx_program_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_mclk_stepping_parameters_lowest_entry(rdev);
+	if (pi->voltage_control)
+		rv6xx_program_voltage_stepping_parameters_lowest_entry(rdev);
+	rv6xx_program_backbias_stepping_parameters_lowest_entry(rdev);
+	rv6xx_program_sclk_spread_spectrum_parameters_lowest_entry(rdev);
+}
+
+static void rv6xx_program_power_level_low(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_LOW,
+					   pi->hw.low_vddc_index);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_LOW,
+					     pi->hw.low_mclk_index);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_LOW,
+					     pi->hw.low_sclk_index);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_LOW,
+					  R600_DISPLAY_WATERMARK_LOW);
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_LOW,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW]);
+}
+
+static void rv6xx_program_power_level_low_to_lowest_state(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_LOW, 0);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_LOW, 0);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_LOW, 0);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_LOW,
+					  R600_DISPLAY_WATERMARK_LOW);
+
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_LOW,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW]);
+
+}
+
+static void rv6xx_program_power_level_medium(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					  pi->hw.medium_vddc_index);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					    pi->hw.medium_mclk_index);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					    pi->hw.medium_sclk_index);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_MEDIUM,
+					 R600_DISPLAY_WATERMARK_LOW);
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_MEDIUM,
+				      pi->hw.pcie_gen2[R600_POWER_LEVEL_MEDIUM]);
+}
+
+static void rv6xx_program_power_level_medium_for_transition(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_mclk_stepping_entry(rdev,
+					  R600_POWER_LEVEL_CTXSW,
+					  pi->hw.mclks[pi->hw.low_mclk_index]);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_MEDIUM, 1);
+
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					     R600_POWER_LEVEL_CTXSW);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					     pi->hw.medium_sclk_index);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_MEDIUM,
+					  R600_DISPLAY_WATERMARK_LOW);
+
+	rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_MEDIUM,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW]);
+}
+
+static void rv6xx_program_power_level_high(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_HIGH,
+					   pi->hw.high_vddc_index);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_HIGH,
+					     pi->hw.high_mclk_index);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_HIGH,
+					     pi->hw.high_sclk_index);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_HIGH,
+					  R600_DISPLAY_WATERMARK_HIGH);
+
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_HIGH,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_HIGH]);
+}
+
+static void rv6xx_enable_backbias(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_PAD_EN | BACKBIAS_DPM_CNTL,
+			 ~(BACKBIAS_PAD_EN | BACKBIAS_DPM_CNTL));
+	else
+		WREG32_P(GENERAL_PWRMGT, 0,
+			 ~(BACKBIAS_VALUE | BACKBIAS_PAD_EN | BACKBIAS_DPM_CNTL));
+}
+
+static void rv6xx_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	if (RREG32(AVIVO_D1CRTC_CONTROL) & AVIVO_CRTC_EN) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	} else if (RREG32(AVIVO_D2CRTC_CONTROL) & AVIVO_CRTC_EN) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+	} else {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	}
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void rv6xx_set_sw_voltage_to_safe(struct radeon_device *rdev,
+					 struct radeon_ps *new_ps,
+					 struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+	u16 safe_voltage;
+
+	safe_voltage = (new_state->low.vddc >= old_state->low.vddc) ?
+		new_state->low.vddc : old_state->low.vddc;
+
+	rv6xx_program_voltage_stepping_entry(rdev, R600_POWER_LEVEL_CTXSW,
+					     safe_voltage);
+
+	WREG32_P(GENERAL_PWRMGT, SW_GPIO_INDEX(R600_POWER_LEVEL_CTXSW),
+		 ~SW_GPIO_INDEX_MASK);
+}
+
+static void rv6xx_set_sw_voltage_to_low(struct radeon_device *rdev,
+					struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	rv6xx_program_voltage_stepping_entry(rdev, R600_POWER_LEVEL_CTXSW,
+					     old_state->low.vddc);
+
+	WREG32_P(GENERAL_PWRMGT, SW_GPIO_INDEX(R600_POWER_LEVEL_CTXSW),
+		~SW_GPIO_INDEX_MASK);
+}
+
+static void rv6xx_set_safe_backbias(struct radeon_device *rdev,
+				    struct radeon_ps *new_ps,
+				    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if ((new_state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) &&
+	    (old_state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE))
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_VALUE, ~BACKBIAS_VALUE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~BACKBIAS_VALUE);
+}
+
+static void rv6xx_set_safe_pcie_gen2(struct radeon_device *rdev,
+				     struct radeon_ps *new_ps,
+				     struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if ((new_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) !=
+	    (old_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2))
+		rv6xx_force_pcie_gen1(rdev);
+}
+
+static void rv6xx_enable_dynamic_voltage_control(struct radeon_device *rdev,
+						 bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN);
+}
+
+static void rv6xx_enable_dynamic_backbias_control(struct radeon_device *rdev,
+						  bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_DPM_CNTL, ~BACKBIAS_DPM_CNTL);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~BACKBIAS_DPM_CNTL);
+}
+
+static int rv6xx_step_sw_voltage(struct radeon_device *rdev,
+				 u16 initial_voltage,
+				 u16 target_voltage)
+{
+	u16 current_voltage;
+	u16 true_target_voltage;
+	u16 voltage_step;
+	int signed_voltage_step;
+
+	if ((radeon_atom_get_voltage_step(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					  &voltage_step)) ||
+	    (radeon_atom_round_to_true_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					       initial_voltage, &current_voltage)) ||
+	    (radeon_atom_round_to_true_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					       target_voltage, &true_target_voltage)))
+		return -EINVAL;
+
+	if (true_target_voltage < current_voltage)
+		signed_voltage_step = -(int)voltage_step;
+	else
+		signed_voltage_step = voltage_step;
+
+	while (current_voltage != true_target_voltage) {
+		current_voltage += signed_voltage_step;
+		rv6xx_program_voltage_stepping_entry(rdev, R600_POWER_LEVEL_CTXSW,
+						     current_voltage);
+		msleep((rdev->pm.dpm.voltage_response_time + 999) / 1000);
+	}
+
+	return 0;
+}
+
+static int rv6xx_step_voltage_if_increasing(struct radeon_device *rdev,
+					    struct radeon_ps *new_ps,
+					    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if (new_state->low.vddc > old_state->low.vddc)
+		return rv6xx_step_sw_voltage(rdev,
+					     old_state->low.vddc,
+					     new_state->low.vddc);
+
+	return 0;
+}
+
+static int rv6xx_step_voltage_if_decreasing(struct radeon_device *rdev,
+					    struct radeon_ps *new_ps,
+					    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if (new_state->low.vddc < old_state->low.vddc)
+		return rv6xx_step_sw_voltage(rdev,
+					     old_state->low.vddc,
+					     new_state->low.vddc);
+	else
+		return 0;
+}
+
+static void rv6xx_enable_high(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if ((pi->restricted_levels < 1) ||
+	    (pi->restricted_levels == 3))
+		r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, true);
+}
+
+static void rv6xx_enable_medium(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->restricted_levels < 2)
+		r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+}
+
+static void rv6xx_set_dpm_event_sources(struct radeon_device *rdev, u32 sources)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	bool want_thermal_protection;
+	enum radeon_dpm_event_src dpm_event_src;
+
+	switch (sources) {
+        case 0:
+        default:
+		want_thermal_protection = false;
+		break;
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL;
+		break;
+
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL;
+		break;
+
+        case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) |
+	      (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL;
+		break;
+	}
+
+	if (want_thermal_protection) {
+		WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK);
+		if (pi->thermal_protection)
+			WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	} else {
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+	}
+}
+
+static void rv6xx_enable_auto_throttle_source(struct radeon_device *rdev,
+					      enum radeon_dpm_auto_throttle_src source,
+					      bool enable)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (enable) {
+		if (!(pi->active_auto_throttle_sources & (1 << source))) {
+			pi->active_auto_throttle_sources |= 1 << source;
+			rv6xx_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	} else {
+		if (pi->active_auto_throttle_sources & (1 << source)) {
+			pi->active_auto_throttle_sources &= ~(1 << source);
+			rv6xx_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	}
+}
+
+
+static void rv6xx_enable_thermal_protection(struct radeon_device *rdev,
+					    bool enable)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->active_auto_throttle_sources)
+		r600_enable_thermal_protection(rdev, enable);
+}
+
+static void rv6xx_generate_transition_stepping(struct radeon_device *rdev,
+					       struct radeon_ps *new_ps,
+					       struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_generate_steps(rdev,
+			     old_state->low.sclk,
+			     new_state->low.sclk,
+			     0, &pi->hw.medium_sclk_index);
+}
+
+static void rv6xx_generate_low_step(struct radeon_device *rdev,
+				    struct radeon_ps *new_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.low_sclk_index = 0;
+	rv6xx_generate_single_step(rdev,
+				   new_state->low.sclk,
+				   0);
+}
+
+static void rv6xx_invalidate_intermediate_steps(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_invalidate_intermediate_steps_range(rdev, 0,
+						  pi->hw.medium_sclk_index);
+}
+
+static void rv6xx_generate_stepping_table(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.low_sclk_index = 0;
+
+	rv6xx_generate_steps(rdev,
+			     new_state->low.sclk,
+			     new_state->medium.sclk,
+			     0,
+			     &pi->hw.medium_sclk_index);
+	rv6xx_generate_steps(rdev,
+			     new_state->medium.sclk,
+			     new_state->high.sclk,
+			     pi->hw.medium_sclk_index,
+			     &pi->hw.high_sclk_index);
+}
+
+static void rv6xx_enable_spread_spectrum(struct radeon_device *rdev,
+					 bool enable)
+{
+	if (enable)
+		rv6xx_enable_dynamic_spread_spectrum(rdev, true);
+	else {
+		rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_LOW, false);
+		rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_MEDIUM, false);
+		rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_HIGH, false);
+		rv6xx_enable_dynamic_spread_spectrum(rdev, false);
+		rv6xx_enable_memory_spread_spectrum(rdev, false);
+	}
+}
+
+static void rv6xx_reset_lvtm_data_sync(struct radeon_device *rdev)
+{
+	if (ASIC_IS_DCE3(rdev))
+		WREG32_P(DCE3_LVTMA_DATA_SYNCHRONIZATION, LVTMA_PFREQCHG, ~LVTMA_PFREQCHG);
+	else
+		WREG32_P(LVTMA_DATA_SYNCHRONIZATION, LVTMA_PFREQCHG, ~LVTMA_PFREQCHG);
+}
+
+static void rv6xx_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					   struct radeon_ps *new_ps,
+					   bool enable)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+
+	if (enable) {
+		rv6xx_enable_bif_dynamic_pcie_gen2(rdev, true);
+		rv6xx_enable_pcie_gen2_support(rdev);
+		r600_enable_dynamic_pcie_gen2(rdev, true);
+	} else {
+		if (!(new_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2))
+			rv6xx_force_pcie_gen1(rdev);
+		rv6xx_enable_bif_dynamic_pcie_gen2(rdev, false);
+		r600_enable_dynamic_pcie_gen2(rdev, false);
+	}
+}
+
+static void rv6xx_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						     struct radeon_ps *new_ps,
+						     struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *current_state = rv6xx_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk >= current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+static void rv6xx_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						    struct radeon_ps *new_ps,
+						    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *current_state = rv6xx_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk < current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+int rv6xx_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (r600_dynamicpm_enabled(rdev))
+		return -EINVAL;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_backbias(rdev, true);
+
+	if (pi->dynamic_ss)
+		rv6xx_enable_spread_spectrum(rdev, true);
+
+	rv6xx_program_mpll_timing_parameters(rdev);
+	rv6xx_program_bsp(rdev);
+	rv6xx_program_git(rdev);
+	rv6xx_program_tp(rdev);
+	rv6xx_program_tpp(rdev);
+	rv6xx_program_sstp(rdev);
+	rv6xx_program_fcp(rdev);
+	rv6xx_program_vddc3d_parameters(rdev);
+	rv6xx_program_voltage_timing_parameters(rdev);
+	rv6xx_program_engine_speed_parameters(rdev);
+
+	rv6xx_enable_display_gap(rdev, true);
+	if (pi->display_gap == false)
+		rv6xx_enable_display_gap(rdev, false);
+
+	rv6xx_program_power_level_enter_state(rdev);
+
+	rv6xx_calculate_stepping_parameters(rdev, boot_ps);
+
+	if (pi->voltage_control)
+		rv6xx_program_voltage_gpio_pins(rdev);
+
+	rv6xx_generate_stepping_table(rdev, boot_ps);
+
+	rv6xx_program_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_stepping_parameters_lowest_entry(rdev);
+
+	rv6xx_program_power_level_low(rdev);
+	rv6xx_program_power_level_medium(rdev);
+	rv6xx_program_power_level_high(rdev);
+	rv6xx_program_vc(rdev);
+	rv6xx_program_at(rdev);
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	rv6xx_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	r600_start_dpm(rdev);
+
+	if (pi->voltage_control)
+		rv6xx_enable_static_voltage_control(rdev, boot_ps, false);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_enable_dynamic_pcie_gen2(rdev, boot_ps, true);
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, true);
+
+	return 0;
+}
+
+void rv6xx_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!r600_dynamicpm_enabled(rdev))
+		return;
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+	rv6xx_enable_display_gap(rdev, false);
+	rv6xx_clear_vc(rdev);
+	r600_set_at(rdev, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF);
+
+	if (pi->thermal_protection)
+		r600_enable_thermal_protection(rdev, false);
+
+	r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_backbias(rdev, false);
+
+	rv6xx_enable_spread_spectrum(rdev, false);
+
+	if (pi->voltage_control)
+		rv6xx_enable_static_voltage_control(rdev, boot_ps, true);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_enable_dynamic_pcie_gen2(rdev, boot_ps, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, false);
+
+	r600_stop_dpm(rdev);
+}
+
+int rv6xx_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	rv6xx_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+
+	rv6xx_clear_vc(rdev);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_set_at(rdev, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF);
+
+	if (pi->thermal_protection)
+		r600_enable_thermal_protection(rdev, false);
+
+	r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	rv6xx_generate_transition_stepping(rdev, new_ps, old_ps);
+	rv6xx_program_power_level_medium_for_transition(rdev);
+
+	if (pi->voltage_control) {
+		rv6xx_set_sw_voltage_to_safe(rdev, new_ps, old_ps);
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+			rv6xx_set_sw_voltage_to_low(rdev, old_ps);
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_set_safe_backbias(rdev, new_ps, old_ps);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_set_safe_pcie_gen2(rdev, new_ps, old_ps);
+
+	if (pi->voltage_control)
+		rv6xx_enable_dynamic_voltage_control(rdev, false);
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_dynamic_backbias_control(rdev, false);
+
+	if (pi->voltage_control) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+			rv6xx_step_voltage_if_increasing(rdev, new_ps, old_ps);
+		msleep((rdev->pm.dpm.voltage_response_time + 999) / 1000);
+	}
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, false);
+	r600_wait_for_power_level_unequal(rdev, R600_POWER_LEVEL_LOW);
+
+	rv6xx_generate_low_step(rdev, new_ps);
+	rv6xx_invalidate_intermediate_steps(rdev);
+	rv6xx_calculate_stepping_parameters(rdev, new_ps);
+	rv6xx_program_stepping_parameters_lowest_entry(rdev);
+	rv6xx_program_power_level_low_to_lowest_state(rdev);
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	if (pi->voltage_control) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC) {
+			ret = rv6xx_step_voltage_if_decreasing(rdev, new_ps, old_ps);
+			if (ret)
+				return ret;
+		}
+		rv6xx_enable_dynamic_voltage_control(rdev, true);
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_dynamic_backbias_control(rdev, true);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_enable_dynamic_pcie_gen2(rdev, new_ps, true);
+
+	rv6xx_reset_lvtm_data_sync(rdev);
+
+	rv6xx_generate_stepping_table(rdev, new_ps);
+	rv6xx_program_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_power_level_low(rdev);
+	rv6xx_program_power_level_medium(rdev);
+	rv6xx_program_power_level_high(rdev);
+	rv6xx_enable_medium(rdev);
+	rv6xx_enable_high(rdev);
+
+	if (pi->thermal_protection)
+		rv6xx_enable_thermal_protection(rdev, true);
+	rv6xx_program_vc(rdev);
+	rv6xx_program_at(rdev);
+
+	rv6xx_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	return 0;
+}
+
+void rv6xx_setup_asic(struct radeon_device *rdev)
+{
+	r600_enable_acpi_pm(rdev);
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s)
+		rv6xx_enable_l0s(rdev);
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1)
+		rv6xx_enable_l1(rdev);
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1)
+		rv6xx_enable_pll_sleep_in_l1(rdev);
+}
+
+void rv6xx_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	rv6xx_program_display_gap(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void rv6xx_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					     struct radeon_ps *rps,
+					     struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV6XX_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV6XX_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void rv6xx_parse_pplib_clock_info(struct radeon_device *rdev,
+					 struct radeon_ps *rps, int index,
+					 union pplib_clock_info *clock_info)
+{
+	struct rv6xx_ps *ps = rv6xx_get_ps(rps);
+	u32 sclk, mclk;
+	u16 vddc;
+	struct rv6xx_pl *pl;
+
+	switch (index) {
+	case 0:
+		pl = &ps->low;
+		break;
+	case 1:
+		pl = &ps->medium;
+		break;
+	case 2:
+	default:
+		pl = &ps->high;
+		break;
+	}
+
+	sclk = le16_to_cpu(clock_info->r600.usEngineClockLow);
+	sclk |= clock_info->r600.ucEngineClockHigh << 16;
+	mclk = le16_to_cpu(clock_info->r600.usMemoryClockLow);
+	mclk |= clock_info->r600.ucMemoryClockHigh << 16;
+
+	pl->mclk = mclk;
+	pl->sclk = sclk;
+	pl->vddc = le16_to_cpu(clock_info->r600.usVDDC);
+	pl->flags = le32_to_cpu(clock_info->r600.ulFlags);
+
+	/* patch up vddc if necessary */
+	if (pl->vddc == 0xff01) {
+		if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0)
+			pl->vddc = vddc;
+	}
+
+	/* fix up pcie gen2 */
+	if (pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) {
+		if ((rdev->family == CHIP_RV610) || (rdev->family == CHIP_RV630)) {
+			if (pl->vddc < 1100)
+				pl->flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+		}
+	}
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+	}
+}
+
+static int rv6xx_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct rv6xx_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			ps = kzalloc(sizeof(struct rv6xx_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			rv6xx_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info);
+			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+				clock_info = (union pplib_clock_info *)
+					(mode_info->atom_context->bios + data_offset +
+					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+					 (power_state->v1.ucClockStateIndices[j] *
+					  power_info->pplib.ucClockInfoSize));
+				rv6xx_parse_pplib_clock_info(rdev,
+							     &rdev->pm.dpm.ps[i], j,
+							     clock_info);
+			}
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int rv6xx_dpm_init(struct radeon_device *rdev)
+{
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	uint16_t data_offset, size;
+	uint8_t frev, crev;
+	struct atom_clock_dividers dividers;
+	struct rv6xx_power_info *pi;
+	int ret;
+
+	pi = kzalloc(sizeof(struct rv6xx_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	ret = rv6xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->spll_ref_div = dividers.ref_div + 1;
+	else
+		pi->spll_ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->mpll_ref_div = dividers.ref_div + 1;
+	else
+		pi->mpll_ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	if (rdev->family >= CHIP_RV670)
+		pi->fb_div_scale = 1;
+	else
+		pi->fb_div_scale = 0;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->gfx_clock_gating = true;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = false;
+	}
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	return 0;
+}
+
+void rv6xx_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct rv6xx_ps *ps = rv6xx_get_ps(rps);
+	struct rv6xx_pl *pl;
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	pl = &ps->low;
+	printk("\t\tpower level 0    sclk: %u mclk: %u vddc: %u\n",
+	       pl->sclk, pl->mclk, pl->vddc);
+	pl = &ps->medium;
+	printk("\t\tpower level 1    sclk: %u mclk: %u vddc: %u\n",
+	       pl->sclk, pl->mclk, pl->vddc);
+	pl = &ps->high;
+	printk("\t\tpower level 2    sclk: %u mclk: %u vddc: %u\n",
+	       pl->sclk, pl->mclk, pl->vddc);
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void rv6xx_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct rv6xx_ps *ps = rv6xx_get_ps(rps);
+	struct rv6xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK) >>
+		CURRENT_PROFILE_INDEX_SHIFT;
+
+	if (current_index > 2) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		if (current_index == 0)
+			pl = &ps->low;
+		else if (current_index == 1)
+			pl = &ps->medium;
+		else /* current_index == 2 */
+			pl = &ps->high;
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u\n",
+			   current_index, pl->sclk, pl->mclk, pl->vddc);
+	}
+}
+
+void rv6xx_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 rv6xx_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct rv6xx_ps *requested_state = rv6xx_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.sclk;
+	else
+		return requested_state->high.sclk;
+}
+
+u32 rv6xx_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct rv6xx_ps *requested_state = rv6xx_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.mclk;
+	else
+		return requested_state->high.mclk;
+}
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.h b/drivers/gpu/drm/radeon/rv6xx_dpm.h
new file mode 100644
index 0000000..8035d53
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.h
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#ifndef __RV6XX_DPM_H__
+#define __RV6XX_DPM_H__
+
+#include "r600_dpm.h"
+
+/* Represents a single SCLK step. */
+struct rv6xx_sclk_stepping
+{
+    u32 vco_frequency;
+    u32 post_divider;
+};
+
+struct rv6xx_pm_hw_state {
+	u32 sclks[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+	u32 mclks[R600_PM_NUMBER_OF_MCLKS];
+	u16 vddc[R600_PM_NUMBER_OF_VOLTAGE_LEVELS];
+	bool backbias[R600_PM_NUMBER_OF_VOLTAGE_LEVELS];
+	bool pcie_gen2[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+	u8 high_sclk_index;
+	u8 medium_sclk_index;
+	u8 low_sclk_index;
+	u8 high_mclk_index;
+	u8 medium_mclk_index;
+	u8 low_mclk_index;
+	u8 high_vddc_index;
+	u8 medium_vddc_index;
+	u8 low_vddc_index;
+	u8 rp[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+	u8 lp[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+};
+
+struct rv6xx_power_info {
+	/* flags */
+	bool voltage_control;
+	bool sclk_ss;
+	bool mclk_ss;
+	bool dynamic_ss;
+	bool dynamic_pcie_gen2;
+	bool thermal_protection;
+	bool display_gap;
+	bool gfx_clock_gating;
+	/* clk values */
+	u32 fb_div_scale;
+	u32 spll_ref_div;
+	u32 mpll_ref_div;
+	u32 bsu;
+	u32 bsp;
+	/* */
+	u32 active_auto_throttle_sources;
+	/* current power state */
+	u32 restricted_levels;
+	struct rv6xx_pm_hw_state hw;
+};
+
+struct rv6xx_pl {
+	u32 sclk;
+	u32 mclk;
+	u16 vddc;
+	u32 flags;
+};
+
+struct rv6xx_ps {
+	struct rv6xx_pl high;
+	struct rv6xx_pl medium;
+	struct rv6xx_pl low;
+};
+
+#define RV6XX_DEFAULT_VCLK_FREQ  40000 /* 10 khz */
+#define RV6XX_DEFAULT_DCLK_FREQ  30000 /* 10 khz */
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv6xxd.h b/drivers/gpu/drm/radeon/rv6xxd.h
new file mode 100644
index 0000000..34e86f9
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv6xxd.h
@@ -0,0 +1,246 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef RV6XXD_H
+#define RV6XXD_H
+
+/* RV6xx power management */
+#define SPLL_CNTL_MODE                                    0x60c
+#       define SPLL_DIV_SYNC                              (1 << 5)
+
+#define GENERAL_PWRMGT                                    0x618
+#       define GLOBAL_PWRMGT_EN                           (1 << 0)
+#       define STATIC_PM_EN                               (1 << 1)
+#       define MOBILE_SU                                  (1 << 2)
+#       define THERMAL_PROTECTION_DIS                     (1 << 3)
+#       define THERMAL_PROTECTION_TYPE                    (1 << 4)
+#       define ENABLE_GEN2PCIE                            (1 << 5)
+#       define SW_GPIO_INDEX(x)                           ((x) << 6)
+#       define SW_GPIO_INDEX_MASK                         (3 << 6)
+#       define LOW_VOLT_D2_ACPI                           (1 << 8)
+#       define LOW_VOLT_D3_ACPI                           (1 << 9)
+#       define VOLT_PWRMGT_EN                             (1 << 10)
+#       define BACKBIAS_PAD_EN                            (1 << 16)
+#       define BACKBIAS_VALUE                             (1 << 17)
+#       define BACKBIAS_DPM_CNTL                          (1 << 18)
+#       define DYN_SPREAD_SPECTRUM_EN                     (1 << 21)
+
+#define MCLK_PWRMGT_CNTL                                  0x624
+#       define MPLL_PWRMGT_OFF                            (1 << 0)
+#       define YCLK_TURNOFF                               (1 << 1)
+#       define MPLL_TURNOFF                               (1 << 2)
+#       define SU_MCLK_USE_BCLK                           (1 << 3)
+#       define DLL_READY                                  (1 << 4)
+#       define MC_BUSY                                    (1 << 5)
+#       define MC_INT_CNTL                                (1 << 7)
+#       define MRDCKA_SLEEP                               (1 << 8)
+#       define MRDCKB_SLEEP                               (1 << 9)
+#       define MRDCKC_SLEEP                               (1 << 10)
+#       define MRDCKD_SLEEP                               (1 << 11)
+#       define MRDCKE_SLEEP                               (1 << 12)
+#       define MRDCKF_SLEEP                               (1 << 13)
+#       define MRDCKG_SLEEP                               (1 << 14)
+#       define MRDCKH_SLEEP                               (1 << 15)
+#       define MRDCKA_RESET                               (1 << 16)
+#       define MRDCKB_RESET                               (1 << 17)
+#       define MRDCKC_RESET                               (1 << 18)
+#       define MRDCKD_RESET                               (1 << 19)
+#       define MRDCKE_RESET                               (1 << 20)
+#       define MRDCKF_RESET                               (1 << 21)
+#       define MRDCKG_RESET                               (1 << 22)
+#       define MRDCKH_RESET                               (1 << 23)
+#       define DLL_READY_READ                             (1 << 24)
+#       define USE_DISPLAY_GAP                            (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                  (1 << 26)
+#       define USE_DISPLAY_GAP_CTXSW                      (1 << 27)
+#       define MPLL_TURNOFF_D2                            (1 << 28)
+#       define USE_DISPLAY_URGENT_CTXSW                   (1 << 29)
+
+#define MPLL_FREQ_LEVEL_0                                 0x6e8
+#       define LEVEL0_MPLL_POST_DIV(x)                    ((x) << 0)
+#       define LEVEL0_MPLL_POST_DIV_MASK                  (0xff << 0)
+#       define LEVEL0_MPLL_FB_DIV(x)                      ((x) << 8)
+#       define LEVEL0_MPLL_FB_DIV_MASK                    (0xfff << 8)
+#       define LEVEL0_MPLL_REF_DIV(x)                     ((x) << 20)
+#       define LEVEL0_MPLL_REF_DIV_MASK                   (0x3f << 20)
+#       define LEVEL0_MPLL_DIV_EN                         (1 << 28)
+#       define LEVEL0_DLL_BYPASS                          (1 << 29)
+#       define LEVEL0_DLL_RESET                           (1 << 30)
+
+#define VID_RT                                            0x6f8
+#       define VID_CRT(x)                                 ((x) << 0)
+#       define VID_CRT_MASK                               (0x1fff << 0)
+#       define VID_CRTU(x)                                ((x) << 13)
+#       define VID_CRTU_MASK                              (7 << 13)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (7 << 16)
+#       define VID_SWT(x)                                 ((x) << 19)
+#       define VID_SWT_MASK                               (0x1f << 19)
+#       define BRT(x)                                     ((x) << 24)
+#       define BRT_MASK                                   (0xff << 24)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x70c
+#       define TARGET_PROFILE_INDEX_MASK                  (3 << 0)
+#       define TARGET_PROFILE_INDEX_SHIFT                 0
+#       define CURRENT_PROFILE_INDEX_MASK                 (3 << 2)
+#       define CURRENT_PROFILE_INDEX_SHIFT                2
+#       define DYN_PWR_ENTER_INDEX(x)                     ((x) << 4)
+#       define DYN_PWR_ENTER_INDEX_MASK                   (3 << 4)
+#       define DYN_PWR_ENTER_INDEX_SHIFT                  4
+#       define CURR_MCLK_INDEX_MASK                       (3 << 6)
+#       define CURR_MCLK_INDEX_SHIFT                      6
+#       define CURR_SCLK_INDEX_MASK                       (0x1f << 8)
+#       define CURR_SCLK_INDEX_SHIFT                      8
+#       define CURR_VID_INDEX_MASK                        (3 << 13)
+#       define CURR_VID_INDEX_SHIFT                       13
+
+#define VID_UPPER_GPIO_CNTL                               0x740
+#       define CTXSW_UPPER_GPIO_VALUES(x)                 ((x) << 0)
+#       define CTXSW_UPPER_GPIO_VALUES_MASK               (7 << 0)
+#       define HIGH_UPPER_GPIO_VALUES(x)                  ((x) << 3)
+#       define HIGH_UPPER_GPIO_VALUES_MASK                (7 << 3)
+#       define MEDIUM_UPPER_GPIO_VALUES(x)                ((x) << 6)
+#       define MEDIUM_UPPER_GPIO_VALUES_MASK              (7 << 6)
+#       define LOW_UPPER_GPIO_VALUES(x)                   ((x) << 9)
+#       define LOW_UPPER_GPIO_VALUES_MASK                 (7 << 9)
+#       define CTXSW_BACKBIAS_VALUE                       (1 << 12)
+#       define HIGH_BACKBIAS_VALUE                        (1 << 13)
+#       define MEDIUM_BACKBIAS_VALUE                      (1 << 14)
+#       define LOW_BACKBIAS_VALUE                         (1 << 15)
+
+#define CG_DISPLAY_GAP_CNTL                               0x7dc
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define CG_THERMAL_CTRL                                   0x7f0
+#       define DPM_EVENT_SRC(x)                           ((x) << 0)
+#       define DPM_EVENT_SRC_MASK                         (7 << 0)
+#       define THERM_INC_CLK                              (1 << 3)
+#       define TOFFSET(x)                                 ((x) << 4)
+#       define TOFFSET_MASK                               (0xff << 4)
+#       define DIG_THERM_DPM(x)                           ((x) << 12)
+#       define DIG_THERM_DPM_MASK                         (0xff << 12)
+#       define CTF_SEL(x)                                 ((x) << 20)
+#       define CTF_SEL_MASK                               (7 << 20)
+#       define CTF_PAD_POLARITY                           (1 << 23)
+#       define CTF_PAD_EN                                 (1 << 24)
+
+#define CG_SPLL_SPREAD_SPECTRUM_LOW                       0x820
+#       define SSEN                                       (1 << 0)
+#       define CLKS(x)                                    ((x) << 3)
+#       define CLKS_MASK                                  (0xff << 3)
+#       define CLKS_SHIFT                                 3
+#       define CLKV(x)                                    ((x) << 11)
+#       define CLKV_MASK                                  (0x7ff << 11)
+#       define CLKV_SHIFT                                 11
+#define CG_MPLL_SPREAD_SPECTRUM                           0x830
+
+#define CITF_CNTL					0x200c
+#       define BLACKOUT_RD                              (1 << 0)
+#       define BLACKOUT_WR                              (1 << 1)
+
+#define RAMCFG						0x2408
+#define		NOOFBANK_SHIFT					0
+#define		NOOFBANK_MASK					0x00000001
+#define		NOOFRANK_SHIFT					1
+#define		NOOFRANK_MASK					0x00000002
+#define		NOOFROWS_SHIFT					2
+#define		NOOFROWS_MASK					0x0000001C
+#define		NOOFCOLS_SHIFT					5
+#define		NOOFCOLS_MASK					0x00000060
+#define		CHANSIZE_SHIFT					7
+#define		CHANSIZE_MASK					0x00000080
+#define		BURSTLENGTH_SHIFT				8
+#define		BURSTLENGTH_MASK				0x00000100
+#define		CHANSIZE_OVERRIDE				(1 << 10)
+
+#define SQM_RATIO					0x2424
+#       define STATE0(x)                                ((x) << 0)
+#       define STATE0_MASK                              (0xff << 0)
+#       define STATE1(x)                                ((x) << 8)
+#       define STATE1_MASK                              (0xff << 8)
+#       define STATE2(x)                                ((x) << 16)
+#       define STATE2_MASK                              (0xff << 16)
+#       define STATE3(x)                                ((x) << 24)
+#       define STATE3_MASK                              (0xff << 24)
+
+#define ARB_RFSH_CNTL					0x2460
+#       define ENABLE                                   (1 << 0)
+#define ARB_RFSH_RATE					0x2464
+#       define POWERMODE0(x)                            ((x) << 0)
+#       define POWERMODE0_MASK                          (0xff << 0)
+#       define POWERMODE1(x)                            ((x) << 8)
+#       define POWERMODE1_MASK                          (0xff << 8)
+#       define POWERMODE2(x)                            ((x) << 16)
+#       define POWERMODE2_MASK                          (0xff << 16)
+#       define POWERMODE3(x)                            ((x) << 24)
+#       define POWERMODE3_MASK                          (0xff << 24)
+
+#define MC_SEQ_DRAM					0x2608
+#       define CKE_DYN                                  (1 << 12)
+
+#define MC_SEQ_CMD					0x26c4
+
+#define MC_SEQ_RESERVE_S				0x2890
+#define MC_SEQ_RESERVE_M				0x2894
+
+#define LVTMA_DATA_SYNCHRONIZATION                      0x7adc
+#       define LVTMA_PFREQCHG                           (1 << 8)
+#define DCE3_LVTMA_DATA_SYNCHRONIZATION                 0x7f98
+
+/* PCIE indirect regs */
+#define PCIE_P_CNTL                                       0x40
+#       define P_PLL_PWRDN_IN_L1L23                       (1 << 3)
+#       define P_PLL_BUF_PDNB                             (1 << 4)
+#       define P_PLL_PDNB                                 (1 << 9)
+#       define P_ALLOW_PRX_FRONTEND_SHUTOFF               (1 << 12)
+/* PCIE PORT indirect regs */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
+#define PCIE_LC_SPEED_CNTL                                0xa4
+#       define LC_GEN2_EN                                 (1 << 0)
+#       define LC_INITIATE_LINK_SPEED_CHANGE              (1 << 7)
+#       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 24)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv730_dpm.c b/drivers/gpu/drm/radeon/rv730_dpm.c
new file mode 100644
index 0000000..3f5e1cf
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv730_dpm.c
@@ -0,0 +1,508 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv730d.h"
+#include "r600_dpm.h"
+#include "rv770_dpm.h"
+#include "atom.h"
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps);
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+
+int rv730_populate_sclk_value(struct radeon_device *rdev,
+			      u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = pi->clk_regs.rv730.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = pi->clk_regs.rv730.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = pi->clk_regs.rv730.cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum = pi->clk_regs.rv730.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = pi->clk_regs.rv730.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider, post_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	if (dividers.enable_post_div)
+		post_divider = ((dividers.post_div >> 4) & 0xf) +
+			(dividers.post_div & 0xf) + 2;
+	else
+		post_divider = 1;
+
+	tmp = (u64) engine_clock * reference_divider * post_divider * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	/* set up registers */
+	if (dividers.enable_post_div)
+		spll_func_cntl |= SPLL_DIVEN;
+	else
+		spll_func_cntl &= ~SPLL_DIVEN;
+	spll_func_cntl &= ~(SPLL_HILEN_MASK | SPLL_LOLEN_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_HILEN((dividers.post_div >> 4) & 0xf);
+	spll_func_cntl |= SPLL_LOLEN(dividers.post_div & 0xf);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * post_divider;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = cpu_to_be32(engine_clock);
+	sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(cg_spll_spread_spectrum);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(cg_spll_spread_spectrum_2);
+
+	return 0;
+}
+
+int rv730_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      LPRV7XX_SMC_MCLK_VALUE mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mclk_pwrmgt_cntl = pi->clk_regs.rv730.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv730.dll_cntl;
+	u32 mpll_func_cntl = pi->clk_regs.rv730.mpll_func_cntl;
+	u32 mpll_func_cntl_2 = pi->clk_regs.rv730.mpll_func_cntl2;
+	u32 mpll_func_cntl_3 = pi->clk_regs.rv730.mpll_func_cntl3;
+	u32 mpll_ss = pi->clk_regs.rv730.mpll_ss;
+	u32 mpll_ss2 = pi->clk_regs.rv730.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 post_divider, reference_divider;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = dividers.ref_div + 1;
+
+	if (dividers.enable_post_div)
+		post_divider = ((dividers.post_div >> 4) & 0xf) +
+			(dividers.post_div & 0xf) + 2;
+	else
+		post_divider = 1;
+
+	/* setup the registers */
+	if (dividers.enable_post_div)
+		mpll_func_cntl |= MPLL_DIVEN;
+	else
+		mpll_func_cntl &= ~MPLL_DIVEN;
+
+	mpll_func_cntl &= ~(MPLL_REF_DIV_MASK | MPLL_HILEN_MASK | MPLL_LOLEN_MASK);
+	mpll_func_cntl |= MPLL_REF_DIV(dividers.ref_div);
+	mpll_func_cntl |= MPLL_HILEN((dividers.post_div >> 4) & 0xf);
+	mpll_func_cntl |= MPLL_LOLEN(dividers.post_div & 0xf);
+
+	mpll_func_cntl_3 &= ~MPLL_FB_DIV_MASK;
+	mpll_func_cntl_3 |= MPLL_FB_DIV(dividers.fb_div);
+	if (dividers.enable_dithen)
+		mpll_func_cntl_3 |= MPLL_DITHEN;
+	else
+		mpll_func_cntl_3 &= ~MPLL_DITHEN;
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * post_divider;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = ss.percentage * dividers.fb_div / (clk_s * 10000);
+
+			mpll_ss &= ~CLK_S_MASK;
+			mpll_ss |= CLK_S(clk_s);
+			mpll_ss |= SSEN;
+
+			mpll_ss2 &= ~CLK_V_MASK;
+			mpll_ss |= CLK_V(clk_v);
+		}
+	}
+
+
+	mclk->mclk730.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk730.vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->mclk730.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk730.vMPLL_FUNC_CNTL = cpu_to_be32(mpll_func_cntl);
+	mclk->mclk730.vMPLL_FUNC_CNTL2 = cpu_to_be32(mpll_func_cntl_2);
+	mclk->mclk730.vMPLL_FUNC_CNTL3 = cpu_to_be32(mpll_func_cntl_3);
+	mclk->mclk730.vMPLL_SS = cpu_to_be32(mpll_ss);
+	mclk->mclk730.vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+void rv730_read_clock_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->clk_regs.rv730.cg_spll_func_cntl =
+		RREG32(CG_SPLL_FUNC_CNTL);
+	pi->clk_regs.rv730.cg_spll_func_cntl_2 =
+		RREG32(CG_SPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv730.cg_spll_func_cntl_3 =
+		RREG32(CG_SPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv730.cg_spll_spread_spectrum =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv730.cg_spll_spread_spectrum_2 =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+
+	pi->clk_regs.rv730.mclk_pwrmgt_cntl =
+		RREG32(TCI_MCLK_PWRMGT_CNTL);
+	pi->clk_regs.rv730.dll_cntl =
+		RREG32(TCI_DLL_CNTL);
+	pi->clk_regs.rv730.mpll_func_cntl =
+		RREG32(CG_MPLL_FUNC_CNTL);
+	pi->clk_regs.rv730.mpll_func_cntl2 =
+		RREG32(CG_MPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv730.mpll_func_cntl3 =
+		RREG32(CG_MPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv730.mpll_ss =
+		RREG32(CG_TCI_MPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv730.mpll_ss2 =
+		RREG32(CG_TCI_MPLL_SPREAD_SPECTRUM_2);
+}
+
+int rv730_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mpll_func_cntl = 0;
+	u32 mpll_func_cntl_2 = 0 ;
+	u32 mpll_func_cntl_3 = 0;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 spll_func_cntl;
+	u32 spll_func_cntl_2;
+	u32 spll_func_cntl_3;
+
+	table->ACPIState = table->initialState;
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		rv770_populate_vddc_value(rdev, pi->acpi_vddc,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = pi->pcie_gen2 ?
+			pi->acpi_pcie_gen2 : 0;
+		table->ACPIState.levels[0].gen2XSP =
+			pi->acpi_pcie_gen2;
+	} else {
+		rv770_populate_vddc_value(rdev, pi->min_vddc_in_table,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	mpll_func_cntl = pi->clk_regs.rv730.mpll_func_cntl;
+	mpll_func_cntl_2 = pi->clk_regs.rv730.mpll_func_cntl2;
+	mpll_func_cntl_3 = pi->clk_regs.rv730.mpll_func_cntl3;
+
+	mpll_func_cntl |= MPLL_RESET | MPLL_BYPASS_EN;
+	mpll_func_cntl &= ~MPLL_SLEEP;
+
+	mpll_func_cntl_2 &= ~MCLK_MUX_SEL_MASK;
+	mpll_func_cntl_2 |= MCLK_MUX_SEL(1);
+
+	mclk_pwrmgt_cntl = (MRDCKA_RESET |
+			    MRDCKB_RESET |
+			    MRDCKC_RESET |
+			    MRDCKD_RESET |
+			    MRDCKE_RESET |
+			    MRDCKF_RESET |
+			    MRDCKG_RESET |
+			    MRDCKH_RESET |
+			    MRDCKA_SLEEP |
+			    MRDCKB_SLEEP |
+			    MRDCKC_SLEEP |
+			    MRDCKD_SLEEP |
+			    MRDCKE_SLEEP |
+			    MRDCKF_SLEEP |
+			    MRDCKG_SLEEP |
+			    MRDCKH_SLEEP);
+
+	dll_cntl = 0xff000000;
+
+	spll_func_cntl = pi->clk_regs.rv730.cg_spll_func_cntl;
+	spll_func_cntl_2 = pi->clk_regs.rv730.cg_spll_func_cntl_2;
+	spll_func_cntl_3 = pi->clk_regs.rv730.cg_spll_func_cntl_3;
+
+	spll_func_cntl |= SPLL_RESET | SPLL_BYPASS_EN;
+	spll_func_cntl &= ~SPLL_SLEEP;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL = cpu_to_be32(mpll_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL2 = cpu_to_be32(mpll_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL3 = cpu_to_be32(mpll_func_cntl_3);
+	table->ACPIState.levels[0].mclk.mclk730.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk730.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk730.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	rv770_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	return 0;
+}
+
+int rv730_populate_smc_initial_state(struct radeon_device *rdev,
+				     struct radeon_ps *radeon_state,
+				     RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_ps *initial_state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 a_t;
+
+	table->initialState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_func_cntl);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL2 =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_func_cntl2);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL3 =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_func_cntl3);
+	table->initialState.levels[0].mclk.mclk730.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.mclk730.vDLL_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.dll_cntl);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_SS =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_ss);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_SS2 =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk730.mclk_value =
+		cpu_to_be32(initial_state->low.mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->low.sclk);
+
+	table->initialState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+
+	table->initialState.levels[0].seqValue =
+		rv770_get_seq_value(rdev, &initial_state->low);
+
+	rv770_populate_vddc_value(rdev,
+				  initial_state->low.vddc,
+				  &table->initialState.levels[0].vddc);
+	rv770_populate_initial_mvdd_value(rdev,
+					  &table->initialState.levels[0].mvdd);
+
+	a_t = CG_R(0xffff) | CG_L(0);
+
+	table->initialState.levels[0].aT = cpu_to_be32(a_t);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+	if (initial_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		table->initialState.levels[0].gen2XSP = 1;
+	else
+		table->initialState.levels[0].gen2XSP = 0;
+
+	table->initialState.levels[1] = table->initialState.levels[0];
+	table->initialState.levels[2] = table->initialState.levels[0];
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	return 0;
+}
+
+void rv730_program_memory_timing_parameters(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	u32 arb_refresh_rate = 0;
+	u32 dram_timing = 0;
+	u32 dram_timing2 = 0;
+	u32 old_dram_timing = 0;
+	u32 old_dram_timing2 = 0;
+
+	arb_refresh_rate = RREG32(MC_ARB_RFSH_RATE) &
+		~(POWERMODE1_MASK | POWERMODE2_MASK | POWERMODE3_MASK);
+	arb_refresh_rate |=
+		(POWERMODE1(rv770_calculate_memory_refresh_rate(rdev, state->low.sclk)) |
+		 POWERMODE2(rv770_calculate_memory_refresh_rate(rdev, state->medium.sclk)) |
+		 POWERMODE3(rv770_calculate_memory_refresh_rate(rdev, state->high.sclk)));
+	WREG32(MC_ARB_RFSH_RATE, arb_refresh_rate);
+
+	/* save the boot dram timings */
+	old_dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	old_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    state->high.sclk,
+					    state->high.mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	WREG32(MC_ARB_DRAM_TIMING_3, dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2_3, dram_timing2);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    state->medium.sclk,
+					    state->medium.mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	WREG32(MC_ARB_DRAM_TIMING_2, dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2_2, dram_timing2);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    state->low.sclk,
+					    state->low.mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	WREG32(MC_ARB_DRAM_TIMING_1, dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2_1, dram_timing2);
+
+	/* restore the boot dram timings */
+	WREG32(MC_ARB_DRAM_TIMING, old_dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2, old_dram_timing2);
+
+}
+
+void rv730_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(TCI_MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+void rv730_stop_dpm(struct radeon_device *rdev)
+{
+	PPSMC_Result result;
+
+	result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_TwoLevelsDisabled);
+
+	if (result != PPSMC_Result_OK)
+		DRM_ERROR("Could not force DPM to low\n");
+
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+
+	WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(TCI_MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+void rv730_program_dcodt(struct radeon_device *rdev, bool use_dcodt)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 i = use_dcodt ? 0 : 1;
+	u32 mc4_io_pad_cntl;
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_DQ_PAD_CNTL_D0_I0);
+	mc4_io_pad_cntl &= 0xFFFFFF00;
+	mc4_io_pad_cntl |= pi->odt_value_0[i];
+	WREG32(MC4_IO_DQ_PAD_CNTL_D0_I0, mc4_io_pad_cntl);
+	WREG32(MC4_IO_DQ_PAD_CNTL_D0_I1, mc4_io_pad_cntl);
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_QS_PAD_CNTL_D0_I0);
+	mc4_io_pad_cntl &= 0xFFFFFF00;
+	mc4_io_pad_cntl |= pi->odt_value_1[i];
+	WREG32(MC4_IO_QS_PAD_CNTL_D0_I0, mc4_io_pad_cntl);
+	WREG32(MC4_IO_QS_PAD_CNTL_D0_I1, mc4_io_pad_cntl);
+}
+
+void rv730_get_odt_values(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mc4_io_pad_cntl;
+
+	pi->odt_value_0[0] = (u8)0;
+	pi->odt_value_1[0] = (u8)0x80;
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_DQ_PAD_CNTL_D0_I0);
+	pi->odt_value_0[1] = (u8)(mc4_io_pad_cntl & 0xff);
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_QS_PAD_CNTL_D0_I0);
+	pi->odt_value_1[1] = (u8)(mc4_io_pad_cntl & 0xff);
+}
diff --git a/drivers/gpu/drm/radeon/rv730d.h b/drivers/gpu/drm/radeon/rv730d.h
new file mode 100644
index 0000000..f0a7954
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv730d.h
@@ -0,0 +1,165 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef RV730_H
+#define RV730_H
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_DIVEN				(1 << 2)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_HILEN(x)				((x) << 12)
+#define		SPLL_HILEN_MASK				(0xf << 12)
+#define		SPLL_LOLEN(x)				((x) << 16)
+#define		SPLL_LOLEN_MASK				(0xf << 16)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define	CG_MPLL_FUNC_CNTL				0x624
+#define		MPLL_RESET				(1 << 0)
+#define		MPLL_SLEEP				(1 << 1)
+#define		MPLL_DIVEN				(1 << 2)
+#define		MPLL_BYPASS_EN				(1 << 3)
+#define		MPLL_REF_DIV(x)				((x) << 4)
+#define		MPLL_REF_DIV_MASK			(0x3f << 4)
+#define		MPLL_HILEN(x)				((x) << 12)
+#define		MPLL_HILEN_MASK				(0xf << 12)
+#define		MPLL_LOLEN(x)				((x) << 16)
+#define		MPLL_LOLEN_MASK				(0xf << 16)
+#define	CG_MPLL_FUNC_CNTL_2				0x628
+#define		MCLK_MUX_SEL(x)				((x) << 0)
+#define		MCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_MPLL_FUNC_CNTL_3				0x62c
+#define		MPLL_FB_DIV(x)				((x) << 0)
+#define		MPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		MPLL_DITHEN				(1 << 28)
+
+#define	CG_TCI_MPLL_SPREAD_SPECTRUM			0x634
+#define	CG_TCI_MPLL_SPREAD_SPECTRUM_2			0x638
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+
+#define	TCI_MCLK_PWRMGT_CNTL				0x648
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA_SLEEP                             (1 << 8)
+#       define MRDCKB_SLEEP                             (1 << 9)
+#       define MRDCKC_SLEEP                             (1 << 10)
+#       define MRDCKD_SLEEP                             (1 << 11)
+#       define MRDCKE_SLEEP                             (1 << 12)
+#       define MRDCKF_SLEEP                             (1 << 13)
+#       define MRDCKG_SLEEP                             (1 << 14)
+#       define MRDCKH_SLEEP                             (1 << 15)
+#       define MRDCKA_RESET                             (1 << 16)
+#       define MRDCKB_RESET                             (1 << 17)
+#       define MRDCKC_RESET                             (1 << 18)
+#       define MRDCKD_RESET                             (1 << 19)
+#       define MRDCKE_RESET                             (1 << 20)
+#       define MRDCKF_RESET                             (1 << 21)
+#       define MRDCKG_RESET                             (1 << 22)
+#       define MRDCKH_RESET                             (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	TCI_DLL_CNTL					0x64c
+
+#define	CG_PG_CNTL					0x858
+#       define PWRGATE_ENABLE                           (1 << 0)
+
+#define	CG_AT				                0x6d4
+#define		CG_R(x)					((x) << 0)
+#define		CG_R_MASK				(0xffff << 0)
+#define		CG_L(x)					((x) << 16)
+#define		CG_L_MASK				(0xffff << 16)
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+
+#define	MC_ARB_DRAM_TIMING_1				0x27f0
+#define	MC_ARB_DRAM_TIMING_2				0x27f4
+#define	MC_ARB_DRAM_TIMING_3				0x27f8
+#define	MC_ARB_DRAM_TIMING2_1				0x27fc
+#define	MC_ARB_DRAM_TIMING2_2				0x2800
+#define	MC_ARB_DRAM_TIMING2_3				0x2804
+
+#define	MC4_IO_DQ_PAD_CNTL_D0_I0			0x2978
+#define	MC4_IO_DQ_PAD_CNTL_D0_I1			0x297c
+#define	MC4_IO_QS_PAD_CNTL_D0_I0			0x2980
+#define	MC4_IO_QS_PAD_CNTL_D0_I1			0x2984
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv740_dpm.c b/drivers/gpu/drm/radeon/rv740_dpm.c
new file mode 100644
index 0000000..c4c8da5
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv740_dpm.c
@@ -0,0 +1,416 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv740d.h"
+#include "r600_dpm.h"
+#include "rv770_dpm.h"
+#include "atom.h"
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+
+u32 rv740_get_decoded_reference_divider(u32 encoded_ref)
+{
+	u32 ref = 0;
+
+	switch (encoded_ref) {
+        case 0:
+		ref = 1;
+		break;
+        case 16:
+		ref = 2;
+		break;
+        case 17:
+		ref = 3;
+		break;
+        case 18:
+		ref = 2;
+		break;
+        case 19:
+		ref = 3;
+		break;
+        case 20:
+		ref = 4;
+		break;
+        case 21:
+		ref = 5;
+		break;
+        default:
+		DRM_ERROR("Invalid encoded Reference Divider\n");
+		ref = 0;
+		break;
+	}
+
+	return ref;
+}
+
+struct dll_speed_setting {
+	u16 min;
+	u16 max;
+	u32 dll_speed;
+};
+
+static struct dll_speed_setting dll_speed_table[16] =
+{
+	{ 270, 320, 0x0f },
+	{ 240, 270, 0x0e },
+	{ 200, 240, 0x0d },
+	{ 180, 200, 0x0c },
+	{ 160, 180, 0x0b },
+	{ 140, 160, 0x0a },
+	{ 120, 140, 0x09 },
+	{ 110, 120, 0x08 },
+	{  95, 110, 0x07 },
+	{  85,  95, 0x06 },
+	{  78,  85, 0x05 },
+	{  70,  78, 0x04 },
+	{  65,  70, 0x03 },
+	{  60,  65, 0x02 },
+	{  42,  60, 0x01 },
+	{  00,  42, 0x00 }
+};
+
+u32 rv740_get_dll_speed(bool is_gddr5, u32 memory_clock)
+{
+	int i;
+	u32 factor;
+	u16 data_rate;
+
+	if (is_gddr5)
+		factor = 4;
+	else
+		factor = 2;
+
+	data_rate = (u16)(memory_clock * factor / 1000);
+
+	if (data_rate < dll_speed_table[0].max) {
+		for (i = 0; i < 16; i++) {
+			if (data_rate > dll_speed_table[i].min &&
+			    data_rate <= dll_speed_table[i].max)
+				return dll_speed_table[i].dll_speed;
+		}
+	}
+
+	DRM_DEBUG_KMS("Target MCLK greater than largest MCLK in DLL speed table\n");
+
+	return 0x0f;
+}
+
+int rv740_populate_sclk_value(struct radeon_device *rdev, u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum = pi->clk_regs.rv770.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = pi->clk_regs.rv770.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	tmp = (u64) engine_clock * reference_divider * dividers.post_div * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_PDIV_A(dividers.post_div);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = cpu_to_be32(engine_clock);
+	sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(cg_spll_spread_spectrum);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(cg_spll_spread_spectrum_2);
+
+	return 0;
+}
+
+int rv740_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      RV7XX_SMC_MCLK_VALUE *mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mpll_ad_func_cntl = pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl = pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl = pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv770.dll_cntl;
+	u32 mpll_ss1 = pi->clk_regs.rv770.mpll_ss1;
+	u32 mpll_ss2 = pi->clk_regs.rv770.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 ibias;
+	u32 dll_speed;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	ibias = rv770_map_clkf_to_ibias(rdev, dividers.whole_fb_div);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(dividers.ref_div);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+	mpll_ad_func_cntl |= CLKF(dividers.whole_fb_div);
+	mpll_ad_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(dividers.ref_div);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+		mpll_dq_func_cntl |= CLKF(dividers.whole_fb_div);
+		mpll_dq_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 decoded_ref = rv740_get_decoded_reference_divider(dividers.ref_div);
+			u32 clk_s = reference_clock * 5 / (decoded_ref * ss.rate);
+			u32 clk_v = 0x40000 * ss.percentage *
+				(dividers.whole_fb_div + (dividers.frac_fb_div / 8)) / (clk_s * 10000);
+
+			mpll_ss1 &= ~CLKV_MASK;
+			mpll_ss1 |= CLKV(clk_v);
+
+			mpll_ss2 &= ~CLKS_MASK;
+			mpll_ss2 |= CLKS(clk_s);
+		}
+	}
+
+	dll_speed = rv740_get_dll_speed(pi->mem_gddr5,
+					memory_clock);
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(dll_speed);
+
+	mclk->mclk770.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->mclk770.vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->mclk770.vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+void rv740_read_clock_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->clk_regs.rv770.cg_spll_func_cntl =
+		RREG32(CG_SPLL_FUNC_CNTL);
+	pi->clk_regs.rv770.cg_spll_func_cntl_2 =
+		RREG32(CG_SPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv770.cg_spll_func_cntl_3 =
+		RREG32(CG_SPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum_2 =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+
+	pi->clk_regs.rv770.mpll_ad_func_cntl =
+		RREG32(MPLL_AD_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_ad_func_cntl_2 =
+		RREG32(MPLL_AD_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mpll_dq_func_cntl =
+		RREG32(MPLL_DQ_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_dq_func_cntl_2 =
+		RREG32(MPLL_DQ_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mclk_pwrmgt_cntl =
+		RREG32(MCLK_PWRMGT_CNTL);
+	pi->clk_regs.rv770.dll_cntl = RREG32(DLL_CNTL);
+	pi->clk_regs.rv770.mpll_ss1 = RREG32(MPLL_SS1);
+	pi->clk_regs.rv770.mpll_ss2 = RREG32(MPLL_SS2);
+}
+
+int rv740_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mpll_ad_func_cntl = pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl = pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl = pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 mclk_pwrmgt_cntl = pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv770.dll_cntl;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		rv770_populate_vddc_value(rdev, pi->acpi_vddc,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE =
+			pi->pcie_gen2 ?
+			pi->acpi_pcie_gen2 : 0;
+		table->ACPIState.levels[0].gen2XSP =
+			pi->acpi_pcie_gen2;
+	} else {
+		rv770_populate_vddc_value(rdev, pi->min_vddc_in_table,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	mpll_dq_func_cntl_2 |= BYPASS | BIAS_GEN_PDNB | RESET_EN;
+
+	mclk_pwrmgt_cntl |= (MRDCKA0_RESET |
+			     MRDCKA1_RESET |
+			     MRDCKB0_RESET |
+			     MRDCKB1_RESET |
+			     MRDCKC0_RESET |
+			     MRDCKC1_RESET |
+			     MRDCKD0_RESET |
+			     MRDCKD1_RESET);
+
+	dll_cntl |= (MRDCKA0_BYPASS |
+		     MRDCKA1_BYPASS |
+		     MRDCKB0_BYPASS |
+		     MRDCKB1_BYPASS |
+		     MRDCKC0_BYPASS |
+		     MRDCKC1_BYPASS |
+		     MRDCKD0_BYPASS |
+		     MRDCKD1_BYPASS);
+
+	spll_func_cntl |= SPLL_RESET | SPLL_SLEEP | SPLL_BYPASS_EN;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk770.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	rv770_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	return 0;
+}
+
+void rv740_enable_mclk_spread_spectrum(struct radeon_device *rdev,
+				       bool enable)
+{
+	if (enable)
+		WREG32_P(MPLL_CNTL_MODE, SS_SSEN, ~SS_SSEN);
+	else
+		WREG32_P(MPLL_CNTL_MODE, 0, ~SS_SSEN);
+}
+
+u8 rv740_get_mclk_frequency_ratio(u32 memory_clock)
+{
+	u8 mc_para_index;
+
+	if ((memory_clock < 10000) || (memory_clock > 47500))
+		mc_para_index = 0x00;
+	else
+		mc_para_index = (u8)((memory_clock - 10000) / 2500);
+
+	return mc_para_index;
+}
diff --git a/drivers/gpu/drm/radeon/rv740d.h b/drivers/gpu/drm/radeon/rv740d.h
new file mode 100644
index 0000000..fe5ab07
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv740d.h
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef RV740_H
+#define RV740_H
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define	MPLL_CNTL_MODE					0x61c
+#define		SS_SSEN					(1 << 24)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define	MCLK_PWRMGT_CNTL				0x648
+#define		DLL_SPEED(x)				((x) << 0)
+#define		DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_SLEEP                            (1 << 8)
+#       define MRDCKA1_SLEEP                            (1 << 9)
+#       define MRDCKB0_SLEEP                            (1 << 10)
+#       define MRDCKB1_SLEEP                            (1 << 11)
+#       define MRDCKC0_SLEEP                            (1 << 12)
+#       define MRDCKC1_SLEEP                            (1 << 13)
+#       define MRDCKD0_SLEEP                            (1 << 14)
+#       define MRDCKD1_SLEEP                            (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+
+#define	MPLL_SS1					0x85c
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x860
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c
new file mode 100644
index 0000000..d914e04
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_dpm.c
@@ -0,0 +1,2521 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv770d.h"
+#include "r600_dpm.h"
+#include "rv770_dpm.h"
+#include "cypress_dpm.h"
+#include "atom.h"
+#include <linux/seq_file.h>
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define MC_CG_SEQ_DRAMCONF_S0       0x05
+#define MC_CG_SEQ_DRAMCONF_S1       0x06
+
+#define PCIE_BUS_CLK                10000
+#define TCLK                        (PCIE_BUS_CLK / 10)
+
+#define SMC_RAM_END 0xC000
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps)
+{
+	struct rv7xx_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void rv770_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					       bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if (enable) {
+		tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+		tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+		tmp |= LC_GEN2_EN_STRAP;
+	} else {
+		if (!pi->boot_in_gen2) {
+			tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+			tmp &= ~LC_GEN2_EN_STRAP;
+		}
+	}
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+
+}
+
+static void rv770_enable_l0s(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L0S_INACTIVITY_MASK;
+	tmp |= LC_L0S_INACTIVITY(3);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv770_enable_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	tmp &= ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(4);
+	tmp &= ~LC_PMI_TO_L1_DIS;
+	tmp &= ~LC_ASPM_TO_L1_DIS;
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv770_enable_pll_sleep_in_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(8);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+
+	/* NOTE, this is a PCIE indirect reg, not PCIE PORT */
+	tmp = RREG32_PCIE(PCIE_P_CNTL);
+	tmp |= P_PLL_PWRDN_IN_L1L23;
+	tmp &= ~P_PLL_BUF_PDNB;
+	tmp &= ~P_PLL_PDNB;
+	tmp |= P_ALLOW_PRX_FRONTEND_SHUTOFF;
+	WREG32_PCIE(PCIE_P_CNTL, tmp);
+}
+
+static void rv770_gfx_clock_gating_enable(struct radeon_device *rdev,
+					  bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_TILING_CONFIG);
+	}
+}
+
+static void rv770_mg_clock_gating_enable(struct radeon_device *rdev,
+					 bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		u32 mgcg_cgtt_local0;
+
+		if (rdev->family == CHIP_RV770)
+			mgcg_cgtt_local0 = RV770_MGCGTTLOCAL0_DFLT;
+		else
+			mgcg_cgtt_local0 = RV7XX_MGCGTTLOCAL0_DFLT;
+
+		WREG32(CG_CGTT_LOCAL_0, mgcg_cgtt_local0);
+		WREG32(CG_CGTT_LOCAL_1, (RV770_MGCGTTLOCAL1_DFLT & 0xFFFFCFFF));
+
+		if (pi->mgcgtssm)
+			WREG32(CGTS_SM_CTRL_REG, RV770_MGCGCGTSSMCTRL_DFLT);
+	} else {
+		WREG32(CG_CGTT_LOCAL_0, 0xFFFFFFFF);
+		WREG32(CG_CGTT_LOCAL_1, 0xFFFFCFFF);
+	}
+}
+
+void rv770_restore_cgcg(struct radeon_device *rdev)
+{
+	bool dpm_en = false, cg_en = false;
+
+	if (RREG32(GENERAL_PWRMGT) & GLOBAL_PWRMGT_EN)
+		dpm_en = true;
+	if (RREG32(SCLK_PWRMGT_CNTL) & DYN_GFX_CLK_OFF_EN)
+		cg_en = true;
+
+	if (dpm_en && !cg_en)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+}
+
+static void rv770_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+void rv770_stop_dpm(struct radeon_device *rdev)
+{
+	PPSMC_Result result;
+
+	result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_TwoLevelsDisabled);
+
+	if (result != PPSMC_Result_OK)
+		DRM_ERROR("Could not force DPM to low.\n");
+
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+
+	WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+bool rv770_dpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32(GENERAL_PWRMGT) & GLOBAL_PWRMGT_EN)
+		return true;
+	else
+		return false;
+}
+
+void rv770_enable_thermal_protection(struct radeon_device *rdev,
+				     bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	else
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+}
+
+void rv770_enable_acpi_pm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+u8 rv770_get_seq_value(struct radeon_device *rdev,
+		       struct rv7xx_pl *pl)
+{
+	return (pl->flags & ATOM_PPLIB_R600_FLAGS_LOWPOWER) ?
+		MC_CG_SEQ_DRAMCONF_S0 : MC_CG_SEQ_DRAMCONF_S1;
+}
+
+int rv770_read_smc_soft_register(struct radeon_device *rdev,
+				 u16 reg_offset, u32 *value)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	return rv770_read_smc_sram_dword(rdev,
+					 pi->soft_regs_start + reg_offset,
+					 value, pi->sram_end);
+}
+
+int rv770_write_smc_soft_register(struct radeon_device *rdev,
+				  u16 reg_offset, u32 value)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	return rv770_write_smc_sram_dword(rdev,
+					  pi->soft_regs_start + reg_offset,
+					  value, pi->sram_end);
+}
+
+int rv770_populate_smc_t(struct radeon_device *rdev,
+			 struct radeon_ps *radeon_state,
+			 RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+	int a_n;
+	int a_d;
+	u8 l[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+	u8 r[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+	u32 a_t;
+
+	l[0] = 0;
+	r[2] = 100;
+
+	a_n = (int)state->medium.sclk * pi->lmp +
+		(int)state->low.sclk * (R600_AH_DFLT - pi->rlp);
+	a_d = (int)state->low.sclk * (100 - (int)pi->rlp) +
+		(int)state->medium.sclk * pi->lmp;
+
+	l[1] = (u8)(pi->lmp - (int)pi->lmp * a_n / a_d);
+	r[0] = (u8)(pi->rlp + (100 - (int)pi->rlp) * a_n / a_d);
+
+	a_n = (int)state->high.sclk * pi->lhp + (int)state->medium.sclk *
+		(R600_AH_DFLT - pi->rmp);
+	a_d = (int)state->medium.sclk * (100 - (int)pi->rmp) +
+		(int)state->high.sclk * pi->lhp;
+
+	l[2] = (u8)(pi->lhp - (int)pi->lhp * a_n / a_d);
+	r[1] = (u8)(pi->rmp + (100 - (int)pi->rmp) * a_n / a_d);
+
+	for (i = 0; i < (RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1); i++) {
+		a_t = CG_R(r[i] * pi->bsp / 200) | CG_L(l[i] * pi->bsp / 200);
+		smc_state->levels[i].aT = cpu_to_be32(a_t);
+	}
+
+	a_t = CG_R(r[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1] * pi->pbsp / 200) |
+		CG_L(l[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1] * pi->pbsp / 200);
+
+	smc_state->levels[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1].aT =
+		cpu_to_be32(a_t);
+
+	return 0;
+}
+
+int rv770_populate_smc_sp(struct radeon_device *rdev,
+			  struct radeon_ps *radeon_state,
+			  RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < (RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1); i++)
+		smc_state->levels[i].bSP = cpu_to_be32(pi->dsp);
+
+	smc_state->levels[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1].bSP =
+		cpu_to_be32(pi->psp);
+
+	return 0;
+}
+
+static void rv770_calculate_fractional_mpll_feedback_divider(u32 memory_clock,
+							     u32 reference_clock,
+							     bool gddr5,
+							     struct atom_clock_dividers *dividers,
+							     u32 *clkf,
+							     u32 *clkfrac)
+{
+	u32 post_divider, reference_divider, feedback_divider8;
+	u32 fyclk;
+
+	if (gddr5)
+		fyclk = (memory_clock * 8) / 2;
+	else
+		fyclk = (memory_clock * 4) / 2;
+
+	post_divider = dividers->post_div;
+	reference_divider = dividers->ref_div;
+
+	feedback_divider8 =
+		(8 * fyclk * reference_divider * post_divider) / reference_clock;
+
+	*clkf = feedback_divider8 / 8;
+	*clkfrac = feedback_divider8 % 8;
+}
+
+static int rv770_encode_yclk_post_div(u32 postdiv, u32 *encoded_postdiv)
+{
+	int ret = 0;
+
+	switch (postdiv) {
+        case 1:
+		*encoded_postdiv = 0;
+		break;
+        case 2:
+		*encoded_postdiv = 1;
+		break;
+        case 4:
+		*encoded_postdiv = 2;
+		break;
+        case 8:
+		*encoded_postdiv = 3;
+		break;
+        case 16:
+		*encoded_postdiv = 4;
+		break;
+        default:
+		ret = -EINVAL;
+		break;
+	}
+
+    return ret;
+}
+
+u32 rv770_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf)
+{
+	if (clkf <= 0x10)
+		return 0x4B;
+	if (clkf <= 0x19)
+		return 0x5B;
+	if (clkf <= 0x21)
+		return 0x2B;
+	if (clkf <= 0x27)
+		return 0x6C;
+	if (clkf <= 0x31)
+		return 0x9D;
+	return 0xC6;
+}
+
+static int rv770_populate_mclk_value(struct radeon_device *rdev,
+				     u32 engine_clock, u32 memory_clock,
+				     RV7XX_SMC_MCLK_VALUE *mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 encoded_reference_dividers[] = { 0, 16, 17, 20, 21 };
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl =
+		pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv770.dll_cntl;
+	struct atom_clock_dividers dividers;
+	u32 reference_clock = rdev->clock.mpll.reference_freq;
+	u32 clkf, clkfrac;
+	u32 postdiv_yclk;
+	u32 ibias;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	if ((dividers.ref_div < 1) || (dividers.ref_div > 5))
+		return -EINVAL;
+
+	rv770_calculate_fractional_mpll_feedback_divider(memory_clock, reference_clock,
+							 pi->mem_gddr5,
+							 &dividers, &clkf, &clkfrac);
+
+	ret = rv770_encode_yclk_post_div(dividers.post_div, &postdiv_yclk);
+	if (ret)
+		return ret;
+
+	ibias = rv770_map_clkf_to_ibias(rdev, clkf);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(encoded_reference_dividers[dividers.ref_div - 1]);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(postdiv_yclk);
+	mpll_ad_func_cntl |= CLKF(clkf);
+	mpll_ad_func_cntl |= CLKFRAC(clkfrac);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		rv770_calculate_fractional_mpll_feedback_divider(memory_clock,
+								 reference_clock,
+								 pi->mem_gddr5,
+								 &dividers, &clkf, &clkfrac);
+
+		ibias = rv770_map_clkf_to_ibias(rdev, clkf);
+
+		ret = rv770_encode_yclk_post_div(dividers.post_div, &postdiv_yclk);
+		if (ret)
+			return ret;
+
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(encoded_reference_dividers[dividers.ref_div - 1]);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(postdiv_yclk);
+		mpll_dq_func_cntl |= CLKF(clkf);
+		mpll_dq_func_cntl |= CLKFRAC(clkfrac);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	mclk->mclk770.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	return 0;
+}
+
+static int rv770_populate_sclk_value(struct radeon_device *rdev,
+				     u32 engine_clock,
+				     RV770_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl =
+		pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum =
+		pi->clk_regs.rv770.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 =
+		pi->clk_regs.rv770.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider, post_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	if (dividers.enable_post_div)
+		post_divider = (0x0f & (dividers.post_div >> 4)) + (0x0f & dividers.post_div) + 2;
+	else
+		post_divider = 1;
+
+	tmp = (u64) engine_clock * reference_divider * post_divider * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	if (dividers.enable_post_div)
+		spll_func_cntl |= SPLL_DIVEN;
+	else
+		spll_func_cntl &= ~SPLL_DIVEN;
+	spll_func_cntl &= ~(SPLL_HILEN_MASK | SPLL_LOLEN_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_HILEN((dividers.post_div >> 4) & 0xf);
+	spll_func_cntl |= SPLL_LOLEN(dividers.post_div & 0xf);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * post_divider;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLKS_MASK;
+			cg_spll_spread_spectrum |= CLKS(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLKV_MASK;
+			cg_spll_spread_spectrum_2 |= CLKV(clk_v);
+		}
+	}
+
+	sclk->sclk_value = cpu_to_be32(engine_clock);
+	sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(cg_spll_spread_spectrum);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(cg_spll_spread_spectrum_2);
+
+	return 0;
+}
+
+int rv770_populate_vddc_value(struct radeon_device *rdev, u16 vddc,
+			      RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	if (!pi->voltage_control) {
+		voltage->index = 0;
+		voltage->value = 0;
+		return 0;
+	}
+
+	for (i = 0; i < pi->valid_vddc_entries; i++) {
+		if (vddc <= pi->vddc_table[i].vddc) {
+			voltage->index = pi->vddc_table[i].vddc_index;
+			voltage->value = cpu_to_be16(vddc);
+			break;
+		}
+	}
+
+	if (i == pi->valid_vddc_entries)
+		return -EINVAL;
+
+	return 0;
+}
+
+int rv770_populate_mvdd_value(struct radeon_device *rdev, u32 mclk,
+			      RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (!pi->mvdd_control) {
+		voltage->index = MVDD_HIGH_INDEX;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+		return 0;
+	}
+
+	if (mclk <= pi->mvdd_split_frequency) {
+		voltage->index = MVDD_LOW_INDEX;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = MVDD_HIGH_INDEX;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+
+	return 0;
+}
+
+static int rv770_convert_power_level_to_smc(struct radeon_device *rdev,
+					    struct rv7xx_pl *pl,
+					    RV770_SMC_HW_PERFORMANCE_LEVEL *level,
+					    u8 watermark_level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int ret;
+
+	level->gen2PCIE = pi->pcie_gen2 ?
+		((pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0) : 0;
+	level->gen2XSP  = (pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0;
+	level->backbias = (pl->flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? 1 : 0;
+	level->displayWatermark = watermark_level;
+
+	if (rdev->family == CHIP_RV740)
+		ret = rv740_populate_sclk_value(rdev, pl->sclk,
+						&level->sclk);
+	else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_sclk_value(rdev, pl->sclk,
+						&level->sclk);
+	else
+		ret = rv770_populate_sclk_value(rdev, pl->sclk,
+						&level->sclk);
+	if (ret)
+		return ret;
+
+	if (rdev->family == CHIP_RV740) {
+		if (pi->mem_gddr5) {
+			if (pl->mclk <= pi->mclk_strobe_mode_threshold)
+				level->strobeMode =
+					rv740_get_mclk_frequency_ratio(pl->mclk) | 0x10;
+			else
+				level->strobeMode = 0;
+
+			if (pl->mclk > pi->mclk_edc_enable_threshold)
+				level->mcFlags = SMC_MC_EDC_RD_FLAG | SMC_MC_EDC_WR_FLAG;
+			else
+				level->mcFlags =  0;
+		}
+		ret = rv740_populate_mclk_value(rdev, pl->sclk,
+						pl->mclk, &level->mclk);
+	} else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_mclk_value(rdev, pl->sclk,
+						pl->mclk, &level->mclk);
+	else
+		ret = rv770_populate_mclk_value(rdev, pl->sclk,
+						pl->mclk, &level->mclk);
+	if (ret)
+		return ret;
+
+	ret = rv770_populate_vddc_value(rdev, pl->vddc,
+					&level->vddc);
+	if (ret)
+		return ret;
+
+	ret = rv770_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int rv770_convert_power_state_to_smc(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state,
+					    RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	int ret;
+
+	if (!(radeon_state->caps & ATOM_PPLIB_DISALLOW_ON_DC))
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	ret = rv770_convert_power_level_to_smc(rdev,
+					       &state->low,
+					       &smc_state->levels[0],
+					       PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = rv770_convert_power_level_to_smc(rdev,
+					       &state->medium,
+					       &smc_state->levels[1],
+					       PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = rv770_convert_power_level_to_smc(rdev,
+					       &state->high,
+					       &smc_state->levels[2],
+					       PPSMC_DISPLAY_WATERMARK_HIGH);
+	if (ret)
+		return ret;
+
+	smc_state->levels[0].arbValue = MC_CG_ARB_FREQ_F1;
+	smc_state->levels[1].arbValue = MC_CG_ARB_FREQ_F2;
+	smc_state->levels[2].arbValue = MC_CG_ARB_FREQ_F3;
+
+	smc_state->levels[0].seqValue = rv770_get_seq_value(rdev,
+							    &state->low);
+	smc_state->levels[1].seqValue = rv770_get_seq_value(rdev,
+							    &state->medium);
+	smc_state->levels[2].seqValue = rv770_get_seq_value(rdev,
+							    &state->high);
+
+	rv770_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	return rv770_populate_smc_t(rdev, radeon_state, smc_state);
+
+}
+
+u32 rv770_calculate_memory_refresh_rate(struct radeon_device *rdev,
+					u32 engine_clock)
+{
+	u32 dram_rows;
+	u32 dram_refresh_rate;
+	u32 mc_arb_rfsh_rate;
+	u32 tmp;
+
+	tmp = (RREG32(MC_ARB_RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT;
+	dram_rows = 1 << (tmp + 10);
+	tmp = RREG32(MC_SEQ_MISC0) & 3;
+	dram_refresh_rate = 1 << (tmp + 3);
+	mc_arb_rfsh_rate = ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64;
+
+	return mc_arb_rfsh_rate;
+}
+
+static void rv770_program_memory_timing_parameters(struct radeon_device *rdev,
+						   struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 sqm_ratio;
+	u32 arb_refresh_rate;
+	u32 high_clock;
+
+	if (state->high.sclk < (state->low.sclk * 0xFF / 0x40))
+		high_clock = state->high.sclk;
+	else
+		high_clock = (state->low.sclk * 0xFF / 0x40);
+
+	radeon_atom_set_engine_dram_timings(rdev, high_clock,
+					    state->high.mclk);
+
+	sqm_ratio =
+		STATE0(64 * high_clock / pi->boot_sclk) |
+		STATE1(64 * high_clock / state->low.sclk) |
+		STATE2(64 * high_clock / state->medium.sclk) |
+		STATE3(64 * high_clock / state->high.sclk);
+	WREG32(MC_ARB_SQM_RATIO, sqm_ratio);
+
+	arb_refresh_rate =
+		POWERMODE0(rv770_calculate_memory_refresh_rate(rdev, pi->boot_sclk)) |
+		POWERMODE1(rv770_calculate_memory_refresh_rate(rdev, state->low.sclk)) |
+		POWERMODE2(rv770_calculate_memory_refresh_rate(rdev, state->medium.sclk)) |
+		POWERMODE3(rv770_calculate_memory_refresh_rate(rdev, state->high.sclk));
+	WREG32(MC_ARB_RFSH_RATE, arb_refresh_rate);
+}
+
+void rv770_enable_backbias(struct radeon_device *rdev,
+			   bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_PAD_EN, ~BACKBIAS_PAD_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~(BACKBIAS_VALUE | BACKBIAS_PAD_EN));
+}
+
+static void rv770_enable_spread_spectrum(struct radeon_device *rdev,
+					 bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (pi->sclk_ss)
+			WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+
+		if (pi->mclk_ss) {
+			if (rdev->family == CHIP_RV740)
+				rv740_enable_mclk_spread_spectrum(rdev, true);
+		}
+	} else {
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+
+		WREG32_P(CG_MPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+
+		if (rdev->family == CHIP_RV740)
+			rv740_enable_mclk_spread_spectrum(rdev, false);
+	}
+}
+
+static void rv770_program_mpll_timing_parameters(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if ((rdev->family == CHIP_RV770) && !pi->mem_gddr5) {
+		WREG32(MPLL_TIME,
+		       (MPLL_LOCK_TIME(R600_MPLLLOCKTIME_DFLT * pi->ref_div) |
+			MPLL_RESET_TIME(R600_MPLLRESETTIME_DFLT)));
+	}
+}
+
+void rv770_setup_bsp(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(pi->asi,
+			       xclk,
+			       16,
+			       &pi->bsp,
+			       &pi->bsu);
+
+	r600_calculate_u_and_p(pi->pasi,
+			       xclk,
+			       16,
+			       &pi->pbsp,
+			       &pi->pbsu);
+
+	pi->dsp = BSP(pi->bsp) | BSU(pi->bsu);
+	pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu);
+
+	WREG32(CG_BSP, pi->dsp);
+
+}
+
+void rv770_program_git(struct radeon_device *rdev)
+{
+	WREG32_P(CG_GIT, CG_GICST(R600_GICST_DFLT), ~CG_GICST_MASK);
+}
+
+void rv770_program_tp(struct radeon_device *rdev)
+{
+	int i;
+	enum r600_td td = R600_TD_DFLT;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		WREG32(CG_FFCT_0 + (i * 4), (UTC_0(r600_utc[i]) | DTC_0(r600_dtc[i])));
+
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+void rv770_program_tpp(struct radeon_device *rdev)
+{
+	WREG32(CG_TPC, R600_TPC_DFLT);
+}
+
+void rv770_program_sstp(struct radeon_device *rdev)
+{
+	WREG32(CG_SSP, (SSTU(R600_SSTU_DFLT) | SST(R600_SST_DFLT)));
+}
+
+void rv770_program_engine_speed_parameters(struct radeon_device *rdev)
+{
+	WREG32_P(SPLL_CNTL_MODE, SPLL_DIV_SYNC, ~SPLL_DIV_SYNC);
+}
+
+static void rv770_enable_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+		DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE));
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+void rv770_program_vc(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	WREG32(CG_FTV, pi->vrc);
+}
+
+void rv770_clear_vc(struct radeon_device *rdev)
+{
+	WREG32(CG_FTV, 0);
+}
+
+int rv770_upload_firmware(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int ret;
+
+	rv770_reset_smc(rdev);
+	rv770_stop_smc_clock(rdev);
+
+	ret = rv770_load_smc_ucode(rdev, pi->sram_end);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int rv770_populate_smc_acpi_state(struct radeon_device *rdev,
+					 RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl =
+		pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		rv770_populate_vddc_value(rdev, pi->acpi_vddc,
+					  &table->ACPIState.levels[0].vddc);
+		if (pi->pcie_gen2) {
+			if (pi->acpi_pcie_gen2)
+				table->ACPIState.levels[0].gen2PCIE = 1;
+			else
+				table->ACPIState.levels[0].gen2PCIE = 0;
+		} else
+			table->ACPIState.levels[0].gen2PCIE = 0;
+		if (pi->acpi_pcie_gen2)
+			table->ACPIState.levels[0].gen2XSP = 1;
+		else
+			table->ACPIState.levels[0].gen2XSP = 0;
+	} else {
+		rv770_populate_vddc_value(rdev, pi->min_vddc_in_table,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	mpll_dq_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	mclk_pwrmgt_cntl = (MRDCKA0_RESET |
+			    MRDCKA1_RESET |
+			    MRDCKB0_RESET |
+			    MRDCKB1_RESET |
+			    MRDCKC0_RESET |
+			    MRDCKC1_RESET |
+			    MRDCKD0_RESET |
+			    MRDCKD1_RESET);
+
+	dll_cntl = 0xff000000;
+
+	spll_func_cntl |= SPLL_RESET | SPLL_SLEEP | SPLL_BYPASS_EN;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk770.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	rv770_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	return 0;
+}
+
+int rv770_populate_initial_mvdd_value(struct radeon_device *rdev,
+				      RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if ((pi->s0_vid_lower_smio_cntl & pi->mvdd_mask_low) ==
+	     (pi->mvdd_low_smio[MVDD_LOW_INDEX] & pi->mvdd_mask_low) ) {
+		voltage->index = MVDD_LOW_INDEX;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = MVDD_HIGH_INDEX;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+
+	return 0;
+}
+
+static int rv770_populate_smc_initial_state(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state,
+					    RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_ps *initial_state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 a_t;
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.mclk770.vDLL_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.dll_cntl);
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss1);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk770.mclk_value =
+		cpu_to_be32(initial_state->low.mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->low.sclk);
+
+	table->initialState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+
+	table->initialState.levels[0].seqValue =
+		rv770_get_seq_value(rdev, &initial_state->low);
+
+	rv770_populate_vddc_value(rdev,
+				  initial_state->low.vddc,
+				  &table->initialState.levels[0].vddc);
+	rv770_populate_initial_mvdd_value(rdev,
+					  &table->initialState.levels[0].mvdd);
+
+	a_t = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(a_t);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+	if (initial_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		table->initialState.levels[0].gen2XSP = 1;
+	else
+		table->initialState.levels[0].gen2XSP = 0;
+
+	if (rdev->family == CHIP_RV740) {
+		if (pi->mem_gddr5) {
+			if (initial_state->low.mclk <= pi->mclk_strobe_mode_threshold)
+				table->initialState.levels[0].strobeMode =
+					rv740_get_mclk_frequency_ratio(initial_state->low.mclk) | 0x10;
+			else
+				table->initialState.levels[0].strobeMode = 0;
+
+			if (initial_state->low.mclk >= pi->mclk_edc_enable_threshold)
+				table->initialState.levels[0].mcFlags = SMC_MC_EDC_RD_FLAG | SMC_MC_EDC_WR_FLAG;
+			else
+				table->initialState.levels[0].mcFlags =  0;
+		}
+	}
+
+	table->initialState.levels[1] = table->initialState.levels[0];
+	table->initialState.levels[2] = table->initialState.levels[0];
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	return 0;
+}
+
+static int rv770_populate_smc_vddc_table(struct radeon_device *rdev,
+					 RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < pi->valid_vddc_entries; i++) {
+		table->highSMIO[pi->vddc_table[i].vddc_index] =
+			pi->vddc_table[i].high_smio;
+		table->lowSMIO[pi->vddc_table[i].vddc_index] =
+			cpu_to_be32(pi->vddc_table[i].low_smio);
+	}
+
+	table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDC] = 0;
+	table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDC] =
+		cpu_to_be32(pi->vddc_mask_low);
+
+	for (i = 0;
+	     ((i < pi->valid_vddc_entries) &&
+	      (pi->max_vddc_in_table >
+	       pi->vddc_table[i].vddc));
+	     i++);
+
+	table->maxVDDCIndexInPPTable =
+		pi->vddc_table[i].vddc_index;
+
+	return 0;
+}
+
+static int rv770_populate_smc_mvdd_table(struct radeon_device *rdev,
+					 RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->mvdd_control) {
+		table->lowSMIO[MVDD_HIGH_INDEX] |=
+			cpu_to_be32(pi->mvdd_low_smio[MVDD_HIGH_INDEX]);
+		table->lowSMIO[MVDD_LOW_INDEX] |=
+			cpu_to_be32(pi->mvdd_low_smio[MVDD_LOW_INDEX]);
+
+		table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_MVDD] = 0;
+		table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_MVDD] =
+			cpu_to_be32(pi->mvdd_mask_low);
+	}
+
+	return 0;
+}
+
+static int rv770_init_smc_table(struct radeon_device *rdev,
+				struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	RV770_SMC_STATETABLE *table = &pi->smc_statetable;
+	int ret;
+
+	memset(table, 0, sizeof(RV770_SMC_STATETABLE));
+
+	pi->boot_sclk = boot_state->low.sclk;
+
+	rv770_populate_smc_vddc_table(rdev, table);
+	rv770_populate_smc_mvdd_table(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+        case THERMAL_TYPE_RV770:
+        case THERMAL_TYPE_ADT7473_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+        case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+        case THERMAL_TYPE_EXTERNAL_GPIO:
+        default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC) {
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_DONT_WAIT_FOR_VBLANK_ON_ALERT)
+			table->extraFlags |= PPSMC_EXTRAFLAGS_AC2DC_DONT_WAIT_FOR_VBLANK;
+
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_GOTO_BOOT_ON_ALERT)
+			table->extraFlags |= PPSMC_EXTRAFLAGS_AC2DC_ACTION_GOTOINITIALSTATE;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	else
+		ret = rv770_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	if (rdev->family == CHIP_RV740)
+		ret = rv740_populate_smc_acpi_state(rdev, table);
+	else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_smc_acpi_state(rdev, table);
+	else
+		ret = rv770_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	return rv770_copy_bytes_to_smc(rdev,
+				       pi->state_table_start,
+				       (const u8 *)table,
+				       sizeof(RV770_SMC_STATETABLE),
+				       pi->sram_end);
+}
+
+static int rv770_construct_vddc_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 min, max, step;
+	u32 steps = 0;
+	u8 vddc_index = 0;
+	u32 i;
+
+	radeon_atom_get_min_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, &min);
+	radeon_atom_get_max_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, &max);
+	radeon_atom_get_voltage_step(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, &step);
+
+	steps = (max - min) / step + 1;
+
+	if (steps > MAX_NO_VREG_STEPS)
+		return -EINVAL;
+
+	for (i = 0; i < steps; i++) {
+		u32 gpio_pins, gpio_mask;
+
+		pi->vddc_table[i].vddc = (u16)(min + i * step);
+		radeon_atom_get_voltage_gpio_settings(rdev,
+						      pi->vddc_table[i].vddc,
+						      SET_VOLTAGE_TYPE_ASIC_VDDC,
+						      &gpio_pins, &gpio_mask);
+		pi->vddc_table[i].low_smio = gpio_pins & gpio_mask;
+		pi->vddc_table[i].high_smio = 0;
+		pi->vddc_mask_low = gpio_mask;
+		if (i > 0) {
+			if ((pi->vddc_table[i].low_smio !=
+			     pi->vddc_table[i - 1].low_smio ) ||
+			     (pi->vddc_table[i].high_smio !=
+			      pi->vddc_table[i - 1].high_smio))
+				vddc_index++;
+		}
+		pi->vddc_table[i].vddc_index = vddc_index;
+	}
+
+	pi->valid_vddc_entries = (u8)steps;
+
+	return 0;
+}
+
+static u32 rv770_get_mclk_split_point(struct atom_memory_info *memory_info)
+{
+	if (memory_info->mem_type == MEM_TYPE_GDDR3)
+		return 30000;
+
+	return 0;
+}
+
+static int rv770_get_mvdd_pin_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 gpio_pins, gpio_mask;
+
+	radeon_atom_get_voltage_gpio_settings(rdev,
+					      MVDD_HIGH_VALUE, SET_VOLTAGE_TYPE_ASIC_MVDDC,
+					      &gpio_pins, &gpio_mask);
+	pi->mvdd_mask_low = gpio_mask;
+	pi->mvdd_low_smio[MVDD_HIGH_INDEX] =
+		gpio_pins & gpio_mask;
+
+	radeon_atom_get_voltage_gpio_settings(rdev,
+					      MVDD_LOW_VALUE, SET_VOLTAGE_TYPE_ASIC_MVDDC,
+					      &gpio_pins, &gpio_mask);
+	pi->mvdd_low_smio[MVDD_LOW_INDEX] =
+		gpio_pins & gpio_mask;
+
+	return 0;
+}
+
+u8 rv770_get_memory_module_index(struct radeon_device *rdev)
+{
+	return (u8) ((RREG32(BIOS_SCRATCH_4) >> 16) & 0xff);
+}
+
+static int rv770_get_mvdd_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 memory_module_index;
+	struct atom_memory_info memory_info;
+
+	memory_module_index = rv770_get_memory_module_index(rdev);
+
+	if (radeon_atom_get_memory_info(rdev, memory_module_index, &memory_info)) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	pi->mvdd_split_frequency =
+		rv770_get_mclk_split_point(&memory_info);
+
+	if (pi->mvdd_split_frequency == 0) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	return rv770_get_mvdd_pin_configuration(rdev);
+}
+
+void rv770_enable_voltage_control(struct radeon_device *rdev,
+				  bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN);
+}
+
+static void rv770_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	if (rdev->pm.dpm.new_active_crtcs & 1) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	} else if (rdev->pm.dpm.new_active_crtcs & 2) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+	} else {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	}
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void rv770_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					   bool enable)
+{
+	rv770_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+static void r7xx_program_memory_timing_parameters(struct radeon_device *rdev,
+						  struct radeon_ps *radeon_new_state)
+{
+	if ((rdev->family == CHIP_RV730) ||
+	    (rdev->family == CHIP_RV710) ||
+	    (rdev->family == CHIP_RV740))
+		rv730_program_memory_timing_parameters(rdev, radeon_new_state);
+	else
+		rv770_program_memory_timing_parameters(rdev, radeon_new_state);
+}
+
+static int rv770_upload_sw_state(struct radeon_device *rdev,
+				 struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 address = pi->state_table_start +
+		offsetof(RV770_SMC_STATETABLE, driverState);
+	RV770_SMC_SWSTATE state = { 0 };
+	int ret;
+
+	ret = rv770_convert_power_state_to_smc(rdev, radeon_new_state, &state);
+	if (ret)
+		return ret;
+
+	return rv770_copy_bytes_to_smc(rdev, address, (const u8 *)&state,
+				       sizeof(RV770_SMC_SWSTATE),
+				       pi->sram_end);
+}
+
+int rv770_halt_smc(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_Halt) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	if (rv770_wait_for_smc_inactive(rdev) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+int rv770_resume_smc(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_Resume) != PPSMC_Result_OK)
+		return -EINVAL;
+	return 0;
+}
+
+int rv770_set_sw_state(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToSwState) != PPSMC_Result_OK)
+		return -EINVAL;
+	return 0;
+}
+
+int rv770_set_boot_state(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToInitialState) != PPSMC_Result_OK)
+		return -EINVAL;
+	return 0;
+}
+
+void rv770_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					      struct radeon_ps *new_ps,
+					      struct radeon_ps *old_ps)
+{
+	struct rv7xx_ps *new_state = rv770_get_ps(new_ps);
+	struct rv7xx_ps *current_state = rv770_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk >= current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+void rv770_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					     struct radeon_ps *new_ps,
+					     struct radeon_ps *old_ps)
+{
+	struct rv7xx_ps *new_state = rv770_get_ps(new_ps);
+	struct rv7xx_ps *current_state = rv770_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk < current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+int rv770_restrict_performance_levels_before_switch(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, (PPSMC_Msg)(PPSMC_MSG_NoForcedLevel)) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	if (rv770_send_msg_to_smc(rdev, (PPSMC_Msg)(PPSMC_MSG_TwoLevelsDisabled)) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+int rv770_dpm_force_performance_level(struct radeon_device *rdev,
+				      enum radeon_dpm_forced_level level)
+{
+	PPSMC_Msg msg;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_ZeroLevelsDisabled) != PPSMC_Result_OK)
+			return -EINVAL;
+		msg = PPSMC_MSG_ForceHigh;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+			return -EINVAL;
+		msg = (PPSMC_Msg)(PPSMC_MSG_TwoLevelsDisabled);
+	} else {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+			return -EINVAL;
+		msg = (PPSMC_Msg)(PPSMC_MSG_ZeroLevelsDisabled);
+	}
+
+	if (rv770_send_msg_to_smc(rdev, msg) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+void r7xx_start_smc(struct radeon_device *rdev)
+{
+	rv770_start_smc(rdev);
+	rv770_start_smc_clock(rdev);
+}
+
+
+void r7xx_stop_smc(struct radeon_device *rdev)
+{
+	rv770_reset_smc(rdev);
+	rv770_stop_smc_clock(rdev);
+}
+
+static void rv770_read_clock_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->clk_regs.rv770.cg_spll_func_cntl =
+		RREG32(CG_SPLL_FUNC_CNTL);
+	pi->clk_regs.rv770.cg_spll_func_cntl_2 =
+		RREG32(CG_SPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv770.cg_spll_func_cntl_3 =
+		RREG32(CG_SPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum_2 =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+	pi->clk_regs.rv770.mpll_ad_func_cntl =
+		RREG32(MPLL_AD_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_ad_func_cntl_2 =
+		RREG32(MPLL_AD_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mpll_dq_func_cntl =
+		RREG32(MPLL_DQ_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_dq_func_cntl_2 =
+		RREG32(MPLL_DQ_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mclk_pwrmgt_cntl =
+		RREG32(MCLK_PWRMGT_CNTL);
+	pi->clk_regs.rv770.dll_cntl = RREG32(DLL_CNTL);
+}
+
+static void r7xx_read_clock_registers(struct radeon_device *rdev)
+{
+	if (rdev->family == CHIP_RV740)
+		rv740_read_clock_registers(rdev);
+	else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_read_clock_registers(rdev);
+	else
+		rv770_read_clock_registers(rdev);
+}
+
+void rv770_read_voltage_smio_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->s0_vid_lower_smio_cntl =
+		RREG32(S0_VID_LOWER_SMIO_CNTL);
+}
+
+void rv770_reset_smio_status(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 sw_smio_index, vid_smio_cntl;
+
+	sw_smio_index =
+		(RREG32(GENERAL_PWRMGT) & SW_SMIO_INDEX_MASK) >> SW_SMIO_INDEX_SHIFT;
+	switch (sw_smio_index) {
+        case 3:
+		vid_smio_cntl = RREG32(S3_VID_LOWER_SMIO_CNTL);
+		break;
+        case 2:
+		vid_smio_cntl = RREG32(S2_VID_LOWER_SMIO_CNTL);
+		break;
+        case 1:
+		vid_smio_cntl = RREG32(S1_VID_LOWER_SMIO_CNTL);
+		break;
+        case 0:
+		return;
+        default:
+		vid_smio_cntl = pi->s0_vid_lower_smio_cntl;
+		break;
+	}
+
+	WREG32(S0_VID_LOWER_SMIO_CNTL, vid_smio_cntl);
+	WREG32_P(GENERAL_PWRMGT, SW_SMIO_INDEX(0), ~SW_SMIO_INDEX_MASK);
+}
+
+void rv770_get_memory_type(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	tmp = RREG32(MC_SEQ_MISC0);
+
+	if (((tmp & MC_SEQ_MISC0_GDDR5_MASK) >> MC_SEQ_MISC0_GDDR5_SHIFT) ==
+	    MC_SEQ_MISC0_GDDR5_VALUE)
+		pi->mem_gddr5 = true;
+	else
+		pi->mem_gddr5 = false;
+
+}
+
+void rv770_get_pcie_gen2_status(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+		pi->pcie_gen2 = true;
+	else
+		pi->pcie_gen2 = false;
+
+	if (pi->pcie_gen2) {
+		if (tmp & LC_CURRENT_DATA_RATE)
+			pi->boot_in_gen2 = true;
+		else
+			pi->boot_in_gen2 = false;
+	} else
+		pi->boot_in_gen2 = false;
+}
+
+#if 0
+static int rv770_enter_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->gfx_clock_gating) {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_TILING_CONFIG);
+	}
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_SwitchToMinimumPower),
+		 ~HOST_SMC_MSG_MASK);
+
+	udelay(7000);
+
+	return 0;
+}
+
+static int rv770_exit_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_ResumeFromMinimumPower),
+		 ~HOST_SMC_MSG_MASK);
+
+	udelay(7000);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (((RREG32(SMC_MSG) & HOST_SMC_RESP_MASK) >> HOST_SMC_RESP_SHIFT) == 1)
+			break;
+		udelay(1000);
+	}
+
+	if (pi->gfx_clock_gating)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+
+	return 0;
+}
+#endif
+
+static void rv770_get_mclk_odt_threshold(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 memory_module_index;
+	struct atom_memory_info memory_info;
+
+	pi->mclk_odt_threshold = 0;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710)) {
+		memory_module_index = rv770_get_memory_module_index(rdev);
+
+		if (radeon_atom_get_memory_info(rdev, memory_module_index, &memory_info))
+			return;
+
+		if (memory_info.mem_type == MEM_TYPE_DDR2 ||
+		    memory_info.mem_type == MEM_TYPE_DDR3)
+			pi->mclk_odt_threshold = 30000;
+	}
+}
+
+void rv770_get_max_vddc(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 vddc;
+
+	if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc))
+		pi->max_vddc = 0;
+	else
+		pi->max_vddc = vddc;
+}
+
+void rv770_program_response_times(struct radeon_device *rdev)
+{
+	u32 voltage_response_time, backbias_response_time;
+	u32 acpi_delay_time, vbi_time_out;
+	u32 vddc_dly, bb_dly, acpi_dly, vbi_dly;
+	u32 reference_clock;
+
+	voltage_response_time = (u32)rdev->pm.dpm.voltage_response_time;
+	backbias_response_time = (u32)rdev->pm.dpm.backbias_response_time;
+
+	if (voltage_response_time == 0)
+		voltage_response_time = 1000;
+
+	if (backbias_response_time == 0)
+		backbias_response_time = 1000;
+
+	acpi_delay_time = 15000;
+	vbi_time_out = 100000;
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	vddc_dly = (voltage_response_time  * reference_clock) / 1600;
+	bb_dly = (backbias_response_time * reference_clock) / 1600;
+	acpi_dly = (acpi_delay_time * reference_clock) / 1600;
+	vbi_dly = (vbi_time_out * reference_clock) / 1600;
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_delay_vreg, vddc_dly);
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_delay_bbias, bb_dly);
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_delay_acpi, acpi_dly);
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mclk_chg_timeout, vbi_dly);
+#if 0
+	/* XXX look up hw revision */
+	if (WEKIVA_A21)
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_baby_step_timer,
+					      0x10);
+#endif
+}
+
+static void rv770_program_dcodt_before_state_switch(struct radeon_device *rdev,
+						    struct radeon_ps *radeon_new_state,
+						    struct radeon_ps *radeon_current_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct rv7xx_ps *new_state = rv770_get_ps(radeon_new_state);
+	struct rv7xx_ps *current_state = rv770_get_ps(radeon_current_state);
+	bool current_use_dc = false;
+	bool new_use_dc = false;
+
+	if (pi->mclk_odt_threshold == 0)
+		return;
+
+	if (current_state->high.mclk <= pi->mclk_odt_threshold)
+		current_use_dc = true;
+
+	if (new_state->high.mclk <= pi->mclk_odt_threshold)
+		new_use_dc = true;
+
+	if (current_use_dc == new_use_dc)
+		return;
+
+	if (!current_use_dc && new_use_dc)
+		return;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_program_dcodt(rdev, new_use_dc);
+}
+
+static void rv770_program_dcodt_after_state_switch(struct radeon_device *rdev,
+						   struct radeon_ps *radeon_new_state,
+						   struct radeon_ps *radeon_current_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct rv7xx_ps *new_state = rv770_get_ps(radeon_new_state);
+	struct rv7xx_ps *current_state = rv770_get_ps(radeon_current_state);
+	bool current_use_dc = false;
+	bool new_use_dc = false;
+
+	if (pi->mclk_odt_threshold == 0)
+		return;
+
+	if (current_state->high.mclk <= pi->mclk_odt_threshold)
+		current_use_dc = true;
+
+	if (new_state->high.mclk <= pi->mclk_odt_threshold)
+		new_use_dc = true;
+
+	if (current_use_dc == new_use_dc)
+		return;
+
+	if (current_use_dc && !new_use_dc)
+		return;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_program_dcodt(rdev, new_use_dc);
+}
+
+static void rv770_retrieve_odt_values(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->mclk_odt_threshold == 0)
+		return;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_get_odt_values(rdev);
+}
+
+static void rv770_set_dpm_event_sources(struct radeon_device *rdev, u32 sources)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	bool want_thermal_protection;
+	enum radeon_dpm_event_src dpm_event_src;
+
+	switch (sources) {
+        case 0:
+        default:
+		want_thermal_protection = false;
+		break;
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL;
+		break;
+
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL;
+		break;
+
+        case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) |
+	      (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL;
+		break;
+	}
+
+	if (want_thermal_protection) {
+		WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK);
+		if (pi->thermal_protection)
+			WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	} else {
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+	}
+}
+
+void rv770_enable_auto_throttle_source(struct radeon_device *rdev,
+				       enum radeon_dpm_auto_throttle_src source,
+				       bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (!(pi->active_auto_throttle_sources & (1 << source))) {
+			pi->active_auto_throttle_sources |= 1 << source;
+			rv770_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	} else {
+		if (pi->active_auto_throttle_sources & (1 << source)) {
+			pi->active_auto_throttle_sources &= ~(1 << source);
+			rv770_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	}
+}
+
+int rv770_set_thermal_temperature_range(struct radeon_device *rdev,
+					int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK);
+	WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+int rv770_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		rv770_restore_cgcg(rdev);
+
+	if (rv770_dpm_enabled(rdev))
+		return -EINVAL;
+
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = rv770_construct_vddc_table(rdev);
+		if (ret) {
+			DRM_ERROR("rv770_construct_vddc_table failed\n");
+			return ret;
+		}
+	}
+
+	if (pi->dcodt)
+		rv770_retrieve_odt_values(rdev);
+
+	if (pi->mvdd_control) {
+		ret = rv770_get_mvdd_configuration(rdev);
+		if (ret) {
+			DRM_ERROR("rv770_get_mvdd_configuration failed\n");
+			return ret;
+		}
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv770_enable_backbias(rdev, true);
+
+	rv770_enable_spread_spectrum(rdev, true);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+
+	rv770_program_mpll_timing_parameters(rdev);
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	rv770_program_engine_speed_parameters(rdev);
+	rv770_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+
+	if (pi->dynamic_pcie_gen2)
+		rv770_enable_dynamic_pcie_gen2(rdev, true);
+
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+	ret = rv770_init_smc_table(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("rv770_init_smc_table failed\n");
+		return ret;
+	}
+
+	rv770_program_response_times(rdev);
+	r7xx_start_smc(rdev);
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_start_dpm(rdev);
+	else
+		rv770_start_dpm(rdev);
+
+	if (pi->gfx_clock_gating)
+		rv770_gfx_clock_gating_enable(rdev, true);
+
+	if (pi->mg_clock_gating)
+		rv770_mg_clock_gating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	return 0;
+}
+
+void rv770_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (!rv770_dpm_enabled(rdev))
+		return;
+
+	rv770_clear_vc(rdev);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+
+	rv770_enable_spread_spectrum(rdev, false);
+
+	if (pi->dynamic_pcie_gen2)
+		rv770_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		rv770_gfx_clock_gating_enable(rdev, false);
+
+	if (pi->mg_clock_gating)
+		rv770_mg_clock_gating_enable(rdev, false);
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_stop_dpm(rdev);
+	else
+		rv770_stop_dpm(rdev);
+
+	r7xx_stop_smc(rdev);
+	rv770_reset_smio_status(rdev);
+}
+
+int rv770_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	ret = rv770_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	rv770_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	ret = rv770_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("rv770_upload_sw_state failed\n");
+		return ret;
+	}
+	r7xx_program_memory_timing_parameters(rdev, new_ps);
+	if (pi->dcodt)
+		rv770_program_dcodt_before_state_switch(rdev, new_ps, old_ps);
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	if (pi->dcodt)
+		rv770_program_dcodt_after_state_switch(rdev, new_ps, old_ps);
+	rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("rv770_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void rv770_dpm_reset_asic(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	rv770_restrict_performance_levels_before_switch(rdev);
+	if (pi->dcodt)
+		rv770_program_dcodt_before_state_switch(rdev, boot_ps, boot_ps);
+	rv770_set_boot_state(rdev);
+	if (pi->dcodt)
+		rv770_program_dcodt_after_state_switch(rdev, boot_ps, boot_ps);
+}
+
+void rv770_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	r7xx_read_clock_registers(rdev);
+	rv770_read_voltage_smio_registers(rdev);
+	rv770_get_memory_type(rdev);
+	if (pi->dcodt)
+		rv770_get_mclk_odt_threshold(rdev);
+	rv770_get_pcie_gen2_status(rdev);
+
+	rv770_enable_acpi_pm(rdev);
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s)
+		rv770_enable_l0s(rdev);
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1)
+		rv770_enable_l1(rdev);
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1)
+		rv770_enable_pll_sleep_in_l1(rdev);
+}
+
+void rv770_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	rv770_program_display_gap(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void rv7xx_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					     struct radeon_ps *rps,
+					     struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					     u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV770_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV770_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void rv7xx_parse_pplib_clock_info(struct radeon_device *rdev,
+					 struct radeon_ps *rps, int index,
+					 union pplib_clock_info *clock_info)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	u32 sclk, mclk;
+	u16 vddc;
+	struct rv7xx_pl *pl;
+
+	switch (index) {
+	case 0:
+		pl = &ps->low;
+		break;
+	case 1:
+		pl = &ps->medium;
+		break;
+	case 2:
+	default:
+		pl = &ps->high;
+		break;
+	}
+
+	if (rdev->family >= CHIP_CEDAR) {
+		sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow);
+		sclk |= clock_info->evergreen.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->evergreen.usMemoryClockLow);
+		mclk |= clock_info->evergreen.ucMemoryClockHigh << 16;
+
+		pl->vddc = le16_to_cpu(clock_info->evergreen.usVDDC);
+		pl->vddci = le16_to_cpu(clock_info->evergreen.usVDDCI);
+		pl->flags = le32_to_cpu(clock_info->evergreen.ulFlags);
+	} else {
+		sclk = le16_to_cpu(clock_info->r600.usEngineClockLow);
+		sclk |= clock_info->r600.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->r600.usMemoryClockLow);
+		mclk |= clock_info->r600.ucMemoryClockHigh << 16;
+
+		pl->vddc = le16_to_cpu(clock_info->r600.usVDDC);
+		pl->flags = le32_to_cpu(clock_info->r600.ulFlags);
+	}
+
+	pl->mclk = mclk;
+	pl->sclk = sclk;
+
+	/* patch up vddc if necessary */
+	if (pl->vddc == 0xff01) {
+		if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0)
+			pl->vddc = vddc;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) {
+		pi->acpi_vddc = pl->vddc;
+		if (rdev->family >= CHIP_CEDAR)
+			eg_pi->acpi_vddci = pl->vddci;
+		if (ps->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+			pi->acpi_pcie_gen2 = true;
+		else
+			pi->acpi_pcie_gen2 = false;
+	}
+
+	if (rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) {
+		if (rdev->family >= CHIP_BARTS) {
+			eg_pi->ulv.supported = true;
+			eg_pi->ulv.pl = pl;
+		}
+	}
+
+	if (pi->min_vddc_in_table > pl->vddc)
+		pi->min_vddc_in_table = pl->vddc;
+
+	if (pi->max_vddc_in_table < pl->vddc)
+		pi->max_vddc_in_table = pl->vddc;
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+		pl->vddci = vddci;
+	}
+
+	if (rdev->family >= CHIP_BARTS) {
+		if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
+		    ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk;
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk;
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc;
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci;
+		}
+	}
+}
+
+int rv7xx_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct rv7xx_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			ps = kzalloc(sizeof(struct rv7xx_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			rv7xx_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info,
+							 power_info->pplib.ucNonClockSize);
+			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+				clock_info = (union pplib_clock_info *)
+					(mode_info->atom_context->bios + data_offset +
+					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+					 (power_state->v1.ucClockStateIndices[j] *
+					  power_info->pplib.ucClockInfoSize));
+				rv7xx_parse_pplib_clock_info(rdev,
+							     &rdev->pm.dpm.ps[i], j,
+							     clock_info);
+			}
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int rv770_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	uint16_t data_offset, size;
+	uint8_t frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	pi = kzalloc(sizeof(struct rv7xx_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	rv770_get_max_vddc(rdev);
+
+	pi->acpi_vddc = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = rv7xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->mclk_strobe_mode_threshold = 30000;
+	pi->mclk_edc_enable_threshold = 30000;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = false;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = RV770_HASI_DFLT;
+	pi->vrc = RV770_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	if (rdev->flags & RADEON_IS_MOBILITY)
+		pi->dcodt = true;
+	else
+		pi->dcodt = false;
+
+	pi->ulps = true;
+
+	pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+	pi->state_table_start = RV770_SMC_TABLE_ADDRESS;
+	pi->soft_regs_start = RV770_SMC_SOFT_REGISTERS_START;
+
+	return 0;
+}
+
+void rv770_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	struct rv7xx_pl *pl;
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	if (rdev->family >= CHIP_CEDAR) {
+		pl = &ps->low;
+		printk("\t\tpower level 0    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+		pl = &ps->medium;
+		printk("\t\tpower level 1    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+		pl = &ps->high;
+		printk("\t\tpower level 2    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+	} else {
+		pl = &ps->low;
+		printk("\t\tpower level 0    sclk: %u mclk: %u vddc: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc);
+		pl = &ps->medium;
+		printk("\t\tpower level 1    sclk: %u mclk: %u vddc: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc);
+		pl = &ps->high;
+		printk("\t\tpower level 2    sclk: %u mclk: %u vddc: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc);
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void rv770_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	struct rv7xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK) >>
+		CURRENT_PROFILE_INDEX_SHIFT;
+
+	if (current_index > 2) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		if (current_index == 0)
+			pl = &ps->low;
+		else if (current_index == 1)
+			pl = &ps->medium;
+		else /* current_index == 2 */
+			pl = &ps->high;
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		if (rdev->family >= CHIP_CEDAR) {
+			seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+				   current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+		} else {
+			seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u\n",
+				   current_index, pl->sclk, pl->mclk, pl->vddc);
+		}
+	}
+}
+
+void rv770_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 rv770_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct rv7xx_ps *requested_state = rv770_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.sclk;
+	else
+		return requested_state->high.sclk;
+}
+
+u32 rv770_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct rv7xx_ps *requested_state = rv770_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.mclk;
+	else
+		return requested_state->high.mclk;
+}
+
+bool rv770_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+
+	if (vblank_time < 300)
+		return true;
+	else
+		return false;
+
+}
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.h b/drivers/gpu/drm/radeon/rv770_dpm.h
new file mode 100644
index 0000000..96b1b2a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_dpm.h
@@ -0,0 +1,289 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RV770_DPM_H__
+#define __RV770_DPM_H__
+
+#include "rv770_smc.h"
+
+struct rv770_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 mpll_ss1;
+	u32 mpll_ss2;
+};
+
+struct rv730_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 mpll_func_cntl;
+	u32 mpll_func_cntl2;
+	u32 mpll_func_cntl3;
+	u32 mpll_ss;
+	u32 mpll_ss2;
+};
+
+union r7xx_clock_registers {
+	struct rv770_clock_registers rv770;
+	struct rv730_clock_registers rv730;
+};
+
+struct vddc_table_entry {
+	u16 vddc;
+	u8 vddc_index;
+	u8 high_smio;
+	u32 low_smio;
+};
+
+#define MAX_NO_OF_MVDD_VALUES 2
+#define MAX_NO_VREG_STEPS 32
+
+struct rv7xx_power_info {
+	/* flags */
+	bool mem_gddr5;
+	bool pcie_gen2;
+	bool dynamic_pcie_gen2;
+	bool acpi_pcie_gen2;
+	bool boot_in_gen2;
+	bool voltage_control; /* vddc */
+	bool mvdd_control;
+	bool sclk_ss;
+	bool mclk_ss;
+	bool dynamic_ss;
+	bool gfx_clock_gating;
+	bool mg_clock_gating;
+	bool mgcgtssm;
+	bool power_gating;
+	bool thermal_protection;
+	bool display_gap;
+	bool dcodt;
+	bool ulps;
+	/* registers */
+	union r7xx_clock_registers clk_regs;
+	u32 s0_vid_lower_smio_cntl;
+	/* voltage */
+	u32 vddc_mask_low;
+	u32 mvdd_mask_low;
+	u32 mvdd_split_frequency;
+	u32 mvdd_low_smio[MAX_NO_OF_MVDD_VALUES];
+	u16 max_vddc;
+	u16 max_vddc_in_table;
+	u16 min_vddc_in_table;
+	struct vddc_table_entry vddc_table[MAX_NO_VREG_STEPS];
+	u8 valid_vddc_entries;
+	/* dc odt */
+	u32 mclk_odt_threshold;
+	u8 odt_value_0[2];
+	u8 odt_value_1[2];
+	/* stored values */
+	u32 boot_sclk;
+	u16 acpi_vddc;
+	u32 ref_div;
+	u32 active_auto_throttle_sources;
+	u32 mclk_stutter_mode_threshold;
+	u32 mclk_strobe_mode_threshold;
+	u32 mclk_edc_enable_threshold;
+	u32 bsp;
+	u32 bsu;
+	u32 pbsp;
+	u32 pbsu;
+	u32 dsp;
+	u32 psp;
+	u32 asi;
+	u32 pasi;
+	u32 vrc;
+	u32 restricted_levels;
+	u32 rlp;
+	u32 rmp;
+	u32 lhp;
+	u32 lmp;
+	/* smc offsets */
+	u16 state_table_start;
+	u16 soft_regs_start;
+	u16 sram_end;
+	/* scratch structs */
+	RV770_SMC_STATETABLE smc_statetable;
+};
+
+struct rv7xx_pl {
+	u32 sclk;
+	u32 mclk;
+	u16 vddc;
+	u16 vddci; /* eg+ only */
+	u32 flags;
+	enum radeon_pcie_gen pcie_gen; /* si+ only */
+};
+
+struct rv7xx_ps {
+	struct rv7xx_pl high;
+	struct rv7xx_pl medium;
+	struct rv7xx_pl low;
+	bool dc_compatible;
+};
+
+#define RV770_RLP_DFLT                                10
+#define RV770_RMP_DFLT                                25
+#define RV770_LHP_DFLT                                25
+#define RV770_LMP_DFLT                                10
+#define RV770_VRC_DFLT                                0x003f
+#define RV770_ASI_DFLT                                1000
+#define RV770_HASI_DFLT                               200000
+#define RV770_MGCGTTLOCAL0_DFLT                       0x00100000
+#define RV7XX_MGCGTTLOCAL0_DFLT                       0
+#define RV770_MGCGTTLOCAL1_DFLT                       0xFFFF0000
+#define RV770_MGCGCGTSSMCTRL_DFLT                     0x55940000
+
+#define MVDD_LOW_INDEX  0
+#define MVDD_HIGH_INDEX 1
+
+#define MVDD_LOW_VALUE  0
+#define MVDD_HIGH_VALUE 0xffff
+
+#define RV770_DEFAULT_VCLK_FREQ  53300 /* 10 khz */
+#define RV770_DEFAULT_DCLK_FREQ  40000 /* 10 khz */
+
+/* rv730/rv710 */
+int rv730_populate_sclk_value(struct radeon_device *rdev,
+			      u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk);
+int rv730_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      LPRV7XX_SMC_MCLK_VALUE mclk);
+void rv730_read_clock_registers(struct radeon_device *rdev);
+int rv730_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table);
+int rv730_populate_smc_initial_state(struct radeon_device *rdev,
+				     struct radeon_ps *radeon_initial_state,
+				     RV770_SMC_STATETABLE *table);
+void rv730_program_memory_timing_parameters(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state);
+void rv730_power_gating_enable(struct radeon_device *rdev,
+			       bool enable);
+void rv730_start_dpm(struct radeon_device *rdev);
+void rv730_stop_dpm(struct radeon_device *rdev);
+void rv730_program_dcodt(struct radeon_device *rdev, bool use_dcodt);
+void rv730_get_odt_values(struct radeon_device *rdev);
+
+/* rv740 */
+int rv740_populate_sclk_value(struct radeon_device *rdev, u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk);
+int rv740_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      RV7XX_SMC_MCLK_VALUE *mclk);
+void rv740_read_clock_registers(struct radeon_device *rdev);
+int rv740_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table);
+void rv740_enable_mclk_spread_spectrum(struct radeon_device *rdev,
+				       bool enable);
+u8 rv740_get_mclk_frequency_ratio(u32 memory_clock);
+u32 rv740_get_dll_speed(bool is_gddr5, u32 memory_clock);
+u32 rv740_get_decoded_reference_divider(u32 encoded_ref);
+
+/* rv770 */
+u32 rv770_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf);
+int rv770_populate_vddc_value(struct radeon_device *rdev, u16 vddc,
+			      RV770_SMC_VOLTAGE_VALUE *voltage);
+int rv770_populate_mvdd_value(struct radeon_device *rdev, u32 mclk,
+			      RV770_SMC_VOLTAGE_VALUE *voltage);
+u8 rv770_get_seq_value(struct radeon_device *rdev,
+		       struct rv7xx_pl *pl);
+int rv770_populate_initial_mvdd_value(struct radeon_device *rdev,
+				      RV770_SMC_VOLTAGE_VALUE *voltage);
+u32 rv770_calculate_memory_refresh_rate(struct radeon_device *rdev,
+					u32 engine_clock);
+void rv770_program_response_times(struct radeon_device *rdev);
+int rv770_populate_smc_sp(struct radeon_device *rdev,
+			  struct radeon_ps *radeon_state,
+			  RV770_SMC_SWSTATE *smc_state);
+int rv770_populate_smc_t(struct radeon_device *rdev,
+			 struct radeon_ps *radeon_state,
+			 RV770_SMC_SWSTATE *smc_state);
+void rv770_read_voltage_smio_registers(struct radeon_device *rdev);
+void rv770_get_memory_type(struct radeon_device *rdev);
+void r7xx_start_smc(struct radeon_device *rdev);
+u8 rv770_get_memory_module_index(struct radeon_device *rdev);
+void rv770_get_max_vddc(struct radeon_device *rdev);
+void rv770_get_pcie_gen2_status(struct radeon_device *rdev);
+void rv770_enable_acpi_pm(struct radeon_device *rdev);
+void rv770_restore_cgcg(struct radeon_device *rdev);
+bool rv770_dpm_enabled(struct radeon_device *rdev);
+void rv770_enable_voltage_control(struct radeon_device *rdev,
+				  bool enable);
+void rv770_enable_backbias(struct radeon_device *rdev,
+			   bool enable);
+void rv770_enable_thermal_protection(struct radeon_device *rdev,
+				     bool enable);
+void rv770_enable_auto_throttle_source(struct radeon_device *rdev,
+				       enum radeon_dpm_auto_throttle_src source,
+				       bool enable);
+void rv770_setup_bsp(struct radeon_device *rdev);
+void rv770_program_git(struct radeon_device *rdev);
+void rv770_program_tp(struct radeon_device *rdev);
+void rv770_program_tpp(struct radeon_device *rdev);
+void rv770_program_sstp(struct radeon_device *rdev);
+void rv770_program_engine_speed_parameters(struct radeon_device *rdev);
+void rv770_program_vc(struct radeon_device *rdev);
+void rv770_clear_vc(struct radeon_device *rdev);
+int rv770_upload_firmware(struct radeon_device *rdev);
+void rv770_stop_dpm(struct radeon_device *rdev);
+void r7xx_stop_smc(struct radeon_device *rdev);
+void rv770_reset_smio_status(struct radeon_device *rdev);
+int rv770_restrict_performance_levels_before_switch(struct radeon_device *rdev);
+int rv770_dpm_force_performance_level(struct radeon_device *rdev,
+				      enum radeon_dpm_forced_level level);
+int rv770_halt_smc(struct radeon_device *rdev);
+int rv770_resume_smc(struct radeon_device *rdev);
+int rv770_set_sw_state(struct radeon_device *rdev);
+int rv770_set_boot_state(struct radeon_device *rdev);
+int rv7xx_parse_power_table(struct radeon_device *rdev);
+void rv770_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					      struct radeon_ps *new_ps,
+					      struct radeon_ps *old_ps);
+void rv770_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					     struct radeon_ps *new_ps,
+					     struct radeon_ps *old_ps);
+
+/* smc */
+int rv770_read_smc_soft_register(struct radeon_device *rdev,
+				 u16 reg_offset, u32 *value);
+int rv770_write_smc_soft_register(struct radeon_device *rdev,
+				  u16 reg_offset, u32 value);
+
+/* thermal */
+int rv770_set_thermal_temperature_range(struct radeon_device *rdev,
+					int min_temp, int max_temp);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv770_smc.c b/drivers/gpu/drm/radeon/rv770_smc.c
new file mode 100644
index 0000000..ab95da5
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_smc.c
@@ -0,0 +1,621 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include <linux/firmware.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "rv770d.h"
+#include "rv770_dpm.h"
+#include "rv770_smc.h"
+#include "atom.h"
+#include "radeon_ucode.h"
+
+#define FIRST_SMC_INT_VECT_REG 0xFFD8
+#define FIRST_INT_VECT_S19     0xFFC0
+
+static const u8 rv770_smc_int_vectors[] =
+{
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x0C, 0xD7,
+	0x08, 0x2B, 0x08, 0x10,
+	0x03, 0x51, 0x03, 0x51,
+	0x03, 0x51, 0x03, 0x51
+};
+
+static const u8 rv730_smc_int_vectors[] =
+{
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x0C, 0xBB,
+	0x08, 0x30, 0x08, 0x15,
+	0x03, 0x56, 0x03, 0x56,
+	0x03, 0x56, 0x03, 0x56
+};
+
+static const u8 rv710_smc_int_vectors[] =
+{
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x0C, 0xCB,
+	0x08, 0x1F, 0x08, 0x04,
+	0x03, 0x51, 0x03, 0x51,
+	0x03, 0x51, 0x03, 0x51
+};
+
+static const u8 rv740_smc_int_vectors[] =
+{
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x0C, 0xD7,
+	0x08, 0x2B, 0x08, 0x10,
+	0x03, 0x51, 0x03, 0x51,
+	0x03, 0x51, 0x03, 0x51
+};
+
+static const u8 cedar_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 redwood_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 juniper_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 cypress_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 barts_smc_int_vectors[] =
+{
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x12, 0xAA,
+	0x0C, 0x2F, 0x15, 0xF6,
+	0x15, 0xF6, 0x05, 0x0A,
+	0x05, 0x0A, 0x05, 0x0A
+};
+
+static const u8 turks_smc_int_vectors[] =
+{
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x12, 0xAA,
+	0x0C, 0x2F, 0x15, 0xF6,
+	0x15, 0xF6, 0x05, 0x0A,
+	0x05, 0x0A, 0x05, 0x0A
+};
+
+static const u8 caicos_smc_int_vectors[] =
+{
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x12, 0xAA,
+	0x0C, 0x2F, 0x15, 0xF6,
+	0x15, 0xF6, 0x05, 0x0A,
+	0x05, 0x0A, 0x05, 0x0A
+};
+
+static const u8 cayman_smc_int_vectors[] =
+{
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x18, 0xEA,
+	0x12, 0x20, 0x1C, 0x34,
+	0x1C, 0x34, 0x08, 0x72,
+	0x08, 0x72, 0x08, 0x72
+};
+
+int rv770_set_smc_sram_address(struct radeon_device *rdev,
+			       u16 smc_address, u16 limit)
+{
+	u32 addr;
+
+	if (smc_address & 3)
+		return -EINVAL;
+	if ((smc_address + 3) > limit)
+		return -EINVAL;
+
+	addr = smc_address;
+	addr |= SMC_SRAM_AUTO_INC_DIS;
+
+	WREG32(SMC_SRAM_ADDR, addr);
+
+	return 0;
+}
+
+int rv770_copy_bytes_to_smc(struct radeon_device *rdev,
+			    u16 smc_start_address, const u8 *src,
+			    u16 byte_count, u16 limit)
+{
+	u32 data, original_data, extra_shift;
+	u16 addr;
+	int ret;
+
+	if (smc_start_address & 3)
+		return -EINVAL;
+	if ((smc_start_address + byte_count) > limit)
+		return -EINVAL;
+
+	addr = smc_start_address;
+
+	while (byte_count >= 4) {
+		/* SMC address space is BE */
+		data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3];
+
+		ret = rv770_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_SRAM_DATA, data);
+
+		src += 4;
+		byte_count -= 4;
+		addr += 4;
+	}
+
+	/* RMW for final bytes */
+	if (byte_count > 0) {
+		data = 0;
+
+		ret = rv770_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		original_data = RREG32(SMC_SRAM_DATA);
+
+		extra_shift = 8 * (4 - byte_count);
+
+		while (byte_count > 0) {
+			/* SMC address space is BE */
+			data = (data << 8) + *src++;
+			byte_count--;
+		}
+
+		data <<= extra_shift;
+
+		data |= (original_data & ~((~0UL) << extra_shift));
+
+		ret = rv770_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_SRAM_DATA, data);
+	}
+
+	return 0;
+}
+
+static int rv770_program_interrupt_vectors(struct radeon_device *rdev,
+					   u32 smc_first_vector, const u8 *src,
+					   u32 byte_count)
+{
+	u32 tmp, i;
+
+	if (byte_count % 4)
+		return -EINVAL;
+
+	if (smc_first_vector < FIRST_SMC_INT_VECT_REG) {
+		tmp = FIRST_SMC_INT_VECT_REG - smc_first_vector;
+
+		if (tmp > byte_count)
+			return 0;
+
+		byte_count -= tmp;
+		src += tmp;
+		smc_first_vector = FIRST_SMC_INT_VECT_REG;
+	}
+
+	for (i = 0; i < byte_count; i += 4) {
+		/* SMC address space is BE */
+		tmp = (src[i] << 24) | (src[i + 1] << 16) | (src[i + 2] << 8) | src[i + 3];
+
+		WREG32(SMC_ISR_FFD8_FFDB + i, tmp);
+	}
+
+	return 0;
+}
+
+void rv770_start_smc(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, SMC_RST_N, ~SMC_RST_N);
+}
+
+void rv770_reset_smc(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, 0, ~SMC_RST_N);
+}
+
+void rv770_stop_smc_clock(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, 0, ~SMC_CLK_EN);
+}
+
+void rv770_start_smc_clock(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, SMC_CLK_EN, ~SMC_CLK_EN);
+}
+
+bool rv770_is_smc_running(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32(SMC_IO);
+
+	if ((tmp & SMC_RST_N) && (tmp & SMC_CLK_EN))
+		return true;
+	else
+		return false;
+}
+
+PPSMC_Result rv770_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg)
+{
+	u32 tmp;
+	int i;
+	PPSMC_Result result;
+
+	if (!rv770_is_smc_running(rdev))
+		return PPSMC_Result_Failed;
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(msg), ~HOST_SMC_MSG_MASK);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(SMC_MSG) & HOST_SMC_RESP_MASK;
+		tmp >>= HOST_SMC_RESP_SHIFT;
+		if (tmp != 0)
+			break;
+		udelay(1);
+	}
+
+	tmp = RREG32(SMC_MSG) & HOST_SMC_RESP_MASK;
+	tmp >>= HOST_SMC_RESP_SHIFT;
+
+	result = (PPSMC_Result)tmp;
+	return result;
+}
+
+PPSMC_Result rv770_wait_for_smc_inactive(struct radeon_device *rdev)
+{
+	int i;
+	PPSMC_Result result = PPSMC_Result_OK;
+
+	if (!rv770_is_smc_running(rdev))
+		return result;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SMC_IO) & SMC_STOP_MODE)
+			break;
+		udelay(1);
+	}
+
+	return result;
+}
+
+static void rv770_clear_smc_sram(struct radeon_device *rdev, u16 limit)
+{
+	u16 i;
+
+	for (i = 0;  i < limit; i += 4) {
+		rv770_set_smc_sram_address(rdev, i, limit);
+		WREG32(SMC_SRAM_DATA, 0);
+	}
+}
+
+int rv770_load_smc_ucode(struct radeon_device *rdev,
+			 u16 limit)
+{
+	int ret;
+	const u8 *int_vect;
+	u16 int_vect_start_address;
+	u16 int_vect_size;
+	const u8 *ucode_data;
+	u16 ucode_start_address;
+	u16 ucode_size;
+
+	if (!rdev->smc_fw)
+		return -EINVAL;
+
+	rv770_clear_smc_sram(rdev, limit);
+
+	switch (rdev->family) {
+	case CHIP_RV770:
+		ucode_start_address = RV770_SMC_UCODE_START;
+		ucode_size = RV770_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv770_smc_int_vectors;
+		int_vect_start_address = RV770_SMC_INT_VECTOR_START;
+		int_vect_size = RV770_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_RV730:
+		ucode_start_address = RV730_SMC_UCODE_START;
+		ucode_size = RV730_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv730_smc_int_vectors;
+		int_vect_start_address = RV730_SMC_INT_VECTOR_START;
+		int_vect_size = RV730_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_RV710:
+		ucode_start_address = RV710_SMC_UCODE_START;
+		ucode_size = RV710_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv710_smc_int_vectors;
+		int_vect_start_address = RV710_SMC_INT_VECTOR_START;
+		int_vect_size = RV710_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_RV740:
+		ucode_start_address = RV740_SMC_UCODE_START;
+		ucode_size = RV740_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv740_smc_int_vectors;
+		int_vect_start_address = RV740_SMC_INT_VECTOR_START;
+		int_vect_size = RV740_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CEDAR:
+		ucode_start_address = CEDAR_SMC_UCODE_START;
+		ucode_size = CEDAR_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&cedar_smc_int_vectors;
+		int_vect_start_address = CEDAR_SMC_INT_VECTOR_START;
+		int_vect_size = CEDAR_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_REDWOOD:
+		ucode_start_address = REDWOOD_SMC_UCODE_START;
+		ucode_size = REDWOOD_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&redwood_smc_int_vectors;
+		int_vect_start_address = REDWOOD_SMC_INT_VECTOR_START;
+		int_vect_size = REDWOOD_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_JUNIPER:
+		ucode_start_address = JUNIPER_SMC_UCODE_START;
+		ucode_size = JUNIPER_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&juniper_smc_int_vectors;
+		int_vect_start_address = JUNIPER_SMC_INT_VECTOR_START;
+		int_vect_size = JUNIPER_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+		ucode_start_address = CYPRESS_SMC_UCODE_START;
+		ucode_size = CYPRESS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&cypress_smc_int_vectors;
+		int_vect_start_address = CYPRESS_SMC_INT_VECTOR_START;
+		int_vect_size = CYPRESS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_BARTS:
+		ucode_start_address = BARTS_SMC_UCODE_START;
+		ucode_size = BARTS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&barts_smc_int_vectors;
+		int_vect_start_address = BARTS_SMC_INT_VECTOR_START;
+		int_vect_size = BARTS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_TURKS:
+		ucode_start_address = TURKS_SMC_UCODE_START;
+		ucode_size = TURKS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&turks_smc_int_vectors;
+		int_vect_start_address = TURKS_SMC_INT_VECTOR_START;
+		int_vect_size = TURKS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CAICOS:
+		ucode_start_address = CAICOS_SMC_UCODE_START;
+		ucode_size = CAICOS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&caicos_smc_int_vectors;
+		int_vect_start_address = CAICOS_SMC_INT_VECTOR_START;
+		int_vect_size = CAICOS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CAYMAN:
+		ucode_start_address = CAYMAN_SMC_UCODE_START;
+		ucode_size = CAYMAN_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&cayman_smc_int_vectors;
+		int_vect_start_address = CAYMAN_SMC_INT_VECTOR_START;
+		int_vect_size = CAYMAN_SMC_INT_VECTOR_SIZE;
+		break;
+	default:
+		DRM_ERROR("unknown asic in smc ucode loader\n");
+		BUG();
+	}
+
+	/* load the ucode */
+	ucode_data = (const u8 *)rdev->smc_fw->data;
+	ret = rv770_copy_bytes_to_smc(rdev, ucode_start_address,
+				      ucode_data, ucode_size, limit);
+	if (ret)
+		return ret;
+
+	/* set up the int vectors */
+	ret = rv770_program_interrupt_vectors(rdev, int_vect_start_address,
+					      int_vect, int_vect_size);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+int rv770_read_smc_sram_dword(struct radeon_device *rdev,
+			      u16 smc_address, u32 *value, u16 limit)
+{
+	int ret;
+
+	ret = rv770_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	*value = RREG32(SMC_SRAM_DATA);
+
+	return 0;
+}
+
+int rv770_write_smc_sram_dword(struct radeon_device *rdev,
+			       u16 smc_address, u32 value, u16 limit)
+{
+	int ret;
+
+	ret = rv770_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	WREG32(SMC_SRAM_DATA, value);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/rv770_smc.h b/drivers/gpu/drm/radeon/rv770_smc.h
new file mode 100644
index 0000000..f78d92a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_smc.h
@@ -0,0 +1,209 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RV770_SMC_H__
+#define __RV770_SMC_H__
+
+#include "ppsmc.h"
+
+#pragma pack(push, 1)
+
+#define RV770_SMC_TABLE_ADDRESS 0xB000
+
+#define RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE    3
+
+struct RV770_SMC_SCLK_VALUE
+{
+    uint32_t        vCG_SPLL_FUNC_CNTL;
+    uint32_t        vCG_SPLL_FUNC_CNTL_2;
+    uint32_t        vCG_SPLL_FUNC_CNTL_3;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM_2;
+    uint32_t        sclk_value;
+};
+
+typedef struct RV770_SMC_SCLK_VALUE RV770_SMC_SCLK_VALUE;
+
+struct RV770_SMC_MCLK_VALUE
+{
+    uint32_t        vMPLL_AD_FUNC_CNTL;
+    uint32_t        vMPLL_AD_FUNC_CNTL_2;
+    uint32_t        vMPLL_DQ_FUNC_CNTL;
+    uint32_t        vMPLL_DQ_FUNC_CNTL_2;
+    uint32_t        vMCLK_PWRMGT_CNTL;
+    uint32_t        vDLL_CNTL;
+    uint32_t        vMPLL_SS;
+    uint32_t        vMPLL_SS2;
+    uint32_t        mclk_value;
+};
+
+typedef struct RV770_SMC_MCLK_VALUE RV770_SMC_MCLK_VALUE;
+
+
+struct RV730_SMC_MCLK_VALUE
+{
+    uint32_t        vMCLK_PWRMGT_CNTL;
+    uint32_t        vDLL_CNTL;
+    uint32_t        vMPLL_FUNC_CNTL;
+    uint32_t        vMPLL_FUNC_CNTL2;
+    uint32_t        vMPLL_FUNC_CNTL3;
+    uint32_t        vMPLL_SS;
+    uint32_t        vMPLL_SS2;
+    uint32_t        mclk_value;
+};
+
+typedef struct RV730_SMC_MCLK_VALUE RV730_SMC_MCLK_VALUE;
+
+struct RV770_SMC_VOLTAGE_VALUE
+{
+    uint16_t             value;
+    uint8_t              index;
+    uint8_t              padding;
+};
+
+typedef struct RV770_SMC_VOLTAGE_VALUE RV770_SMC_VOLTAGE_VALUE;
+
+union RV7XX_SMC_MCLK_VALUE
+{
+    RV770_SMC_MCLK_VALUE    mclk770;
+    RV730_SMC_MCLK_VALUE    mclk730;
+};
+
+typedef union RV7XX_SMC_MCLK_VALUE RV7XX_SMC_MCLK_VALUE, *LPRV7XX_SMC_MCLK_VALUE;
+
+struct RV770_SMC_HW_PERFORMANCE_LEVEL
+{
+    uint8_t                 arbValue;
+    union{
+        uint8_t             seqValue;
+        uint8_t             ACIndex;
+    };
+    uint8_t                 displayWatermark;
+    uint8_t                 gen2PCIE;
+    uint8_t                 gen2XSP;
+    uint8_t                 backbias;
+    uint8_t                 strobeMode;
+    uint8_t                 mcFlags;
+    uint32_t                aT;
+    uint32_t                bSP;
+    RV770_SMC_SCLK_VALUE    sclk;
+    RV7XX_SMC_MCLK_VALUE    mclk;
+    RV770_SMC_VOLTAGE_VALUE vddc;
+    RV770_SMC_VOLTAGE_VALUE mvdd;
+    RV770_SMC_VOLTAGE_VALUE vddci;
+    uint8_t                 reserved1;
+    uint8_t                 reserved2;
+    uint8_t                 stateFlags;
+    uint8_t                 padding;
+};
+
+#define SMC_STROBE_RATIO    0x0F
+#define SMC_STROBE_ENABLE   0x10
+
+#define SMC_MC_EDC_RD_FLAG  0x01
+#define SMC_MC_EDC_WR_FLAG  0x02
+#define SMC_MC_RTT_ENABLE   0x04
+#define SMC_MC_STUTTER_EN   0x08
+
+typedef struct RV770_SMC_HW_PERFORMANCE_LEVEL RV770_SMC_HW_PERFORMANCE_LEVEL;
+
+struct RV770_SMC_SWSTATE
+{
+    uint8_t           flags;
+    uint8_t           padding1;
+    uint8_t           padding2;
+    uint8_t           padding3;
+    RV770_SMC_HW_PERFORMANCE_LEVEL levels[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+};
+
+typedef struct RV770_SMC_SWSTATE RV770_SMC_SWSTATE;
+
+#define RV770_SMC_VOLTAGEMASK_VDDC 0
+#define RV770_SMC_VOLTAGEMASK_MVDD 1
+#define RV770_SMC_VOLTAGEMASK_VDDCI 2
+#define RV770_SMC_VOLTAGEMASK_MAX  4
+
+struct RV770_SMC_VOLTAGEMASKTABLE
+{
+    uint8_t  highMask[RV770_SMC_VOLTAGEMASK_MAX];
+    uint32_t lowMask[RV770_SMC_VOLTAGEMASK_MAX];
+};
+
+typedef struct RV770_SMC_VOLTAGEMASKTABLE RV770_SMC_VOLTAGEMASKTABLE;
+
+#define MAX_NO_VREG_STEPS 32
+
+struct RV770_SMC_STATETABLE
+{
+    uint8_t             thermalProtectType;
+    uint8_t             systemFlags;
+    uint8_t             maxVDDCIndexInPPTable;
+    uint8_t             extraFlags;
+    uint8_t             highSMIO[MAX_NO_VREG_STEPS];
+    uint32_t            lowSMIO[MAX_NO_VREG_STEPS];
+    RV770_SMC_VOLTAGEMASKTABLE voltageMaskTable;
+    RV770_SMC_SWSTATE   initialState;
+    RV770_SMC_SWSTATE   ACPIState;
+    RV770_SMC_SWSTATE   driverState;
+    RV770_SMC_SWSTATE   ULVState;
+};
+
+typedef struct RV770_SMC_STATETABLE RV770_SMC_STATETABLE;
+
+#define PPSMC_STATEFLAG_AUTO_PULSE_SKIP 0x01
+
+#pragma pack(pop)
+
+#define RV770_SMC_SOFT_REGISTERS_START        0x104
+
+#define RV770_SMC_SOFT_REGISTER_mclk_chg_timeout        0x0
+#define RV770_SMC_SOFT_REGISTER_baby_step_timer         0x8
+#define RV770_SMC_SOFT_REGISTER_delay_bbias             0xC
+#define RV770_SMC_SOFT_REGISTER_delay_vreg              0x10
+#define RV770_SMC_SOFT_REGISTER_delay_acpi              0x2C
+#define RV770_SMC_SOFT_REGISTER_seq_index               0x64
+#define RV770_SMC_SOFT_REGISTER_mvdd_chg_time           0x68
+#define RV770_SMC_SOFT_REGISTER_mclk_switch_lim         0x78
+#define RV770_SMC_SOFT_REGISTER_mc_block_delay          0x90
+#define RV770_SMC_SOFT_REGISTER_uvd_enabled             0x9C
+#define RV770_SMC_SOFT_REGISTER_is_asic_lombok          0xA0
+
+int rv770_set_smc_sram_address(struct radeon_device *rdev,
+			       u16 smc_address, u16 limit);
+int rv770_copy_bytes_to_smc(struct radeon_device *rdev,
+			    u16 smc_start_address, const u8 *src,
+			    u16 byte_count, u16 limit);
+void rv770_start_smc(struct radeon_device *rdev);
+void rv770_reset_smc(struct radeon_device *rdev);
+void rv770_stop_smc_clock(struct radeon_device *rdev);
+void rv770_start_smc_clock(struct radeon_device *rdev);
+bool rv770_is_smc_running(struct radeon_device *rdev);
+PPSMC_Result rv770_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg);
+PPSMC_Result rv770_wait_for_smc_inactive(struct radeon_device *rdev);
+int rv770_read_smc_sram_dword(struct radeon_device *rdev,
+			      u16 smc_address, u32 *value, u16 limit);
+int rv770_write_smc_sram_dword(struct radeon_device *rdev,
+			       u16 smc_address, u32 value, u16 limit);
+int rv770_load_smc_ucode(struct radeon_device *rdev,
+			 u16 limit);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h
index 85b1626..6bef2b7 100644
--- a/drivers/gpu/drm/radeon/rv770d.h
+++ b/drivers/gpu/drm/radeon/rv770d.h
@@ -62,6 +62,246 @@
 #	define UPLL_FB_DIV(x)				((x) << 0)
 #	define UPLL_FB_DIV_MASK				0x01FFFFFF
 
+/* pm registers */
+#define	SMC_SRAM_ADDR					0x200
+#define		SMC_SRAM_AUTO_INC_DIS				(1 << 16)
+#define	SMC_SRAM_DATA					0x204
+#define	SMC_IO						0x208
+#define		SMC_RST_N					(1 << 0)
+#define		SMC_STOP_MODE					(1 << 2)
+#define		SMC_CLK_EN					(1 << 11)
+#define	SMC_MSG						0x20c
+#define		HOST_SMC_MSG(x)					((x) << 0)
+#define		HOST_SMC_MSG_MASK				(0xff << 0)
+#define		HOST_SMC_MSG_SHIFT				0
+#define		HOST_SMC_RESP(x)				((x) << 8)
+#define		HOST_SMC_RESP_MASK				(0xff << 8)
+#define		HOST_SMC_RESP_SHIFT				8
+#define		SMC_HOST_MSG(x)					((x) << 16)
+#define		SMC_HOST_MSG_MASK				(0xff << 16)
+#define		SMC_HOST_MSG_SHIFT				16
+#define		SMC_HOST_RESP(x)				((x) << 24)
+#define		SMC_HOST_RESP_MASK				(0xff << 24)
+#define		SMC_HOST_RESP_SHIFT				24
+
+#define	SMC_ISR_FFD8_FFDB				0x218
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_DIVEN				(1 << 2)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_HILEN(x)				((x) << 12)
+#define		SPLL_HILEN_MASK				(0xf << 12)
+#define		SPLL_LOLEN(x)				((x) << 16)
+#define		SPLL_LOLEN_MASK				(0xf << 16)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define	SPLL_CNTL_MODE					0x610
+#define		SPLL_DIV_SYNC				(1 << 5)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define CG_TPC                                            0x640
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#define	MCLK_PWRMGT_CNTL				0x648
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_SLEEP                            (1 << 8)
+#       define MRDCKA1_SLEEP                            (1 << 9)
+#       define MRDCKB0_SLEEP                            (1 << 10)
+#       define MRDCKB1_SLEEP                            (1 << 11)
+#       define MRDCKC0_SLEEP                            (1 << 12)
+#       define MRDCKC1_SLEEP                            (1 << 13)
+#       define MRDCKD0_SLEEP                            (1 << 14)
+#       define MRDCKD1_SLEEP                            (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define MPLL_TIME                                         0x654
+#       define MPLL_LOCK_TIME(x)			((x) << 0)
+#       define MPLL_LOCK_TIME_MASK			(0xffff << 0)
+#       define MPLL_RESET_TIME(x)			((x) << 16)
+#       define MPLL_RESET_TIME_MASK			(0xffff << 16)
+
+#define CG_CLKPIN_CNTL                                    0x660
+#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+#       define XTALIN_DIVIDE                              (1 << 9)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x66c
+#       define CURRENT_PROFILE_INDEX_MASK                 (0xf << 4)
+#       define CURRENT_PROFILE_INDEX_SHIFT                4
+
+#define S0_VID_LOWER_SMIO_CNTL                            0x678
+#define S1_VID_LOWER_SMIO_CNTL                            0x67c
+#define S2_VID_LOWER_SMIO_CNTL                            0x680
+#define S3_VID_LOWER_SMIO_CNTL                            0x684
+
+#define CG_FTV                                            0x690
+#define CG_FFCT_0                                         0x694
+#       define UTC_0(x)                                   ((x) << 0)
+#       define UTC_0_MASK                                 (0x3ff << 0)
+#       define DTC_0(x)                                   ((x) << 10)
+#       define DTC_0_MASK                                 (0x3ff << 10)
+
+#define CG_BSP                                          0x6d0
+#       define BSP(x)					((x) << 0)
+#       define BSP_MASK					(0xffff << 0)
+#       define BSU(x)					((x) << 16)
+#       define BSU_MASK					(0xf << 16)
+#define CG_AT                                           0x6d4
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+#define CG_GIT                                          0x6d8
+#       define CG_GICST(x)                              ((x) << 0)
+#       define CG_GICST_MASK                            (0xffff << 0)
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+
+#define CG_SSP                                            0x6e8
+#       define SST(x)                                     ((x) << 0)
+#       define SST_MASK                                   (0xffff << 0)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (0xf << 16)
+
+#define CG_DISPLAY_GAP_CNTL                               0x714
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLKS(x)					((x) << 4)
+#define		CLKS_MASK				(0xfff << 4)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	CG_MPLL_SPREAD_SPECTRUM				0x798
+#define CG_UPLL_SPREAD_SPECTRUM				0x79c
+#	define SSEN_MASK				0x00000001
+
+#define CG_CGTT_LOCAL_0                                   0x7d0
+#define CG_CGTT_LOCAL_1                                   0x7d4
+
+#define BIOS_SCRATCH_4                                    0x1734
+
+#define MC_SEQ_MISC0                                      0x2a00
+#define         MC_SEQ_MISC0_GDDR5_SHIFT                  28
+#define         MC_SEQ_MISC0_GDDR5_MASK                   0xf0000000
+#define         MC_SEQ_MISC0_GDDR5_VALUE                  5
+
+#define MC_ARB_SQM_RATIO                                  0x2770
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0xff << 0)
+#define		STATE1(x)				((x) << 8)
+#define		STATE1_MASK				(0xff << 8)
+#define		STATE2(x)				((x) << 16)
+#define		STATE2_MASK				(0xff << 16)
+#define		STATE3(x)				((x) << 24)
+#define		STATE3_MASK				(0xff << 24)
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+
+#define CGTS_SM_CTRL_REG                                  0x9150
+
 /* Registers */
 #define	CB_COLOR0_BASE					0x28040
 #define	CB_COLOR1_BASE					0x28044
@@ -86,8 +326,8 @@
 #define	CONFIG_MEMSIZE					0x5428
 
 #define	CP_ME_CNTL					0x86D8
-#define		CP_ME_HALT					(1<<28)
-#define		CP_PFP_HALT					(1<<26)
+#define		CP_ME_HALT					(1 << 28)
+#define		CP_PFP_HALT					(1 << 26)
 #define	CP_ME_RAM_DATA					0xC160
 #define	CP_ME_RAM_RADDR					0xC158
 #define	CP_ME_RAM_WADDR					0xC15C
@@ -157,9 +397,22 @@
 #define		GUI_ACTIVE					(1<<31)
 #define	GRBM_STATUS2					0x8014
 
-#define CG_CLKPIN_CNTL                                    0x660
-#       define MUX_TCLK_TO_XCLK                           (1 << 8)
-#       define XTALIN_DIVIDE                              (1 << 9)
+#define	CG_THERMAL_CTRL					0x72C
+#define 	DPM_EVENT_SRC(x)			((x) << 0)
+#define 	DPM_EVENT_SRC_MASK			(7 << 0)
+#define		DIG_THERM_DPM(x)			((x) << 14)
+#define		DIG_THERM_DPM_MASK			0x003FC000
+#define		DIG_THERM_DPM_SHIFT			14
+
+#define	CG_THERMAL_INT					0x734
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
 
 #define	CG_MULT_THERMAL_STATUS				0x740
 #define		ASIC_T(x)			        ((x) << 16)
@@ -662,7 +915,22 @@
 #define D1GRPH_SECONDARY_SURFACE_ADDRESS_HIGH             0x691c
 #define D2GRPH_SECONDARY_SURFACE_ADDRESS_HIGH             0x611c
 
-/* PCIE link stuff */
+/* PCIE indirect regs */
+#define PCIE_P_CNTL                                       0x40
+#       define P_PLL_PWRDN_IN_L1L23                       (1 << 3)
+#       define P_PLL_BUF_PDNB                             (1 << 4)
+#       define P_PLL_PDNB                                 (1 << 9)
+#       define P_ALLOW_PRX_FRONTEND_SHUTOFF               (1 << 12)
+/* PCIE PORT regs */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
 #define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
 #define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
 #       define LC_LINK_WIDTH_SHIFT                        0
@@ -690,6 +958,9 @@
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
 #       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
 #       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
 #       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
 #       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index a1b0da6..2349067 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -32,40 +32,43 @@
 #include "sid.h"
 #include "atom.h"
 #include "si_blit_shaders.h"
+#include "clearstate_si.h"
+#include "radeon_ucode.h"
 
-#define SI_PFP_UCODE_SIZE 2144
-#define SI_PM4_UCODE_SIZE 2144
-#define SI_CE_UCODE_SIZE 2144
-#define SI_RLC_UCODE_SIZE 2048
-#define SI_MC_UCODE_SIZE 7769
-#define OLAND_MC_UCODE_SIZE 7863
 
 MODULE_FIRMWARE("radeon/TAHITI_pfp.bin");
 MODULE_FIRMWARE("radeon/TAHITI_me.bin");
 MODULE_FIRMWARE("radeon/TAHITI_ce.bin");
 MODULE_FIRMWARE("radeon/TAHITI_mc.bin");
 MODULE_FIRMWARE("radeon/TAHITI_rlc.bin");
+MODULE_FIRMWARE("radeon/TAHITI_smc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin");
 MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
 MODULE_FIRMWARE("radeon/VERDE_me.bin");
 MODULE_FIRMWARE("radeon/VERDE_ce.bin");
 MODULE_FIRMWARE("radeon/VERDE_mc.bin");
 MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
+MODULE_FIRMWARE("radeon/VERDE_smc.bin");
 MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
 MODULE_FIRMWARE("radeon/OLAND_me.bin");
 MODULE_FIRMWARE("radeon/OLAND_ce.bin");
 MODULE_FIRMWARE("radeon/OLAND_mc.bin");
 MODULE_FIRMWARE("radeon/OLAND_rlc.bin");
+MODULE_FIRMWARE("radeon/OLAND_smc.bin");
 MODULE_FIRMWARE("radeon/HAINAN_pfp.bin");
 MODULE_FIRMWARE("radeon/HAINAN_me.bin");
 MODULE_FIRMWARE("radeon/HAINAN_ce.bin");
 MODULE_FIRMWARE("radeon/HAINAN_mc.bin");
 MODULE_FIRMWARE("radeon/HAINAN_rlc.bin");
+MODULE_FIRMWARE("radeon/HAINAN_smc.bin");
 
+static void si_pcie_gen3_enable(struct radeon_device *rdev);
+static void si_program_aspm(struct radeon_device *rdev);
 extern int r600_ih_ring_alloc(struct radeon_device *rdev);
 extern void r600_ih_ring_fini(struct radeon_device *rdev);
 extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
@@ -75,6 +78,228 @@ extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 
+static const u32 verde_rlc_save_restore_register_list[] =
+{
+	(0x8000 << 16) | (0x98f4 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x98f4 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0xe80 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0xe80 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x89bc >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x89bc >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x8c1c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x8c1c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x98f0 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xe7c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9148 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9148 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9150 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x897c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8d8c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac54 >> 2),
+	0X00000000,
+	0x3,
+	(0x9c00 << 16) | (0x98f8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9910 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9914 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9918 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x991c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9920 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9924 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9928 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x992c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9930 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9934 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9938 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x993c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9940 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9944 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9948 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x994c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9950 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9954 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9958 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x995c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9960 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9964 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9968 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x996c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9970 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9974 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9978 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x997c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9980 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9984 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9988 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x998c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c00 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c04 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c08 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9b7c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9b7c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0xe84 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0xe84 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x89c0 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x89c0 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x914c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x914c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x8c20 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x8c20 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9354 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9354 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9060 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9364 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9100 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x913c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e0 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e4 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e8 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e0 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e4 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8bcc >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8b24 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88c4 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e50 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c0c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e58 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e5c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9508 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x950c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9494 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac0c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac10 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xae00 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac08 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88d4 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88c8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88cc >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x89b0 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8b10 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8a14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9830 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9834 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9838 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9a10 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8001 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8001 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8041 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8041 << 16) | (0x9874 >> 2),
+	0x00000000,
+	0x00000000
+};
+
 static const u32 tahiti_golden_rlc_registers[] =
 {
 	0xc424, 0xffffffff, 0x00601005,
@@ -1320,6 +1545,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 	const char *chip_name;
 	const char *rlc_chip_name;
 	size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size;
+	size_t smc_req_size;
 	char fw_name[30];
 	int err;
 
@@ -1341,6 +1567,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(TAHITI_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_PITCAIRN:
 		chip_name = "PITCAIRN";
@@ -1350,6 +1577,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(PITCAIRN_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_VERDE:
 		chip_name = "VERDE";
@@ -1359,6 +1587,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(VERDE_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_OLAND:
 		chip_name = "OLAND";
@@ -1368,6 +1597,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(OLAND_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_HAINAN:
 		chip_name = "HAINAN";
@@ -1377,6 +1607,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(HAINAN_SMC_UCODE_SIZE, 4);
 		break;
 	default: BUG();
 	}
@@ -1439,6 +1670,17 @@ static int si_init_microcode(struct radeon_device *rdev)
 		err = -EINVAL;
 	}
 
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+	err = request_firmware(&rdev->smc_fw, fw_name, &pdev->dev);
+	if (err)
+		goto out;
+	if (rdev->smc_fw->size != smc_req_size) {
+		printk(KERN_ERR
+		       "si_smc: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->smc_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
 out:
 	platform_device_unregister(pdev);
 
@@ -1457,6 +1699,8 @@ out:
 		rdev->rlc_fw = NULL;
 		release_firmware(rdev->mc_fw);
 		rdev->mc_fw = NULL;
+		release_firmware(rdev->smc_fw);
+		rdev->smc_fw = NULL;
 	}
 	return err;
 }
@@ -1792,7 +2036,8 @@ static void dce6_program_watermarks(struct radeon_device *rdev,
 					 u32 lb_size, u32 num_heads)
 {
 	struct drm_display_mode *mode = &radeon_crtc->base.mode;
-	struct dce6_wm_params wm;
+	struct dce6_wm_params wm_low, wm_high;
+	u32 dram_channels;
 	u32 pixel_period;
 	u32 line_time = 0;
 	u32 latency_watermark_a = 0, latency_watermark_b = 0;
@@ -1808,38 +2053,83 @@ static void dce6_program_watermarks(struct radeon_device *rdev,
 		priority_a_cnt = 0;
 		priority_b_cnt = 0;
 
-		wm.yclk = rdev->pm.current_mclk * 10;
-		wm.sclk = rdev->pm.current_sclk * 10;
-		wm.disp_clk = mode->clock;
-		wm.src_width = mode->crtc_hdisplay;
-		wm.active_time = mode->crtc_hdisplay * pixel_period;
-		wm.blank_time = line_time - wm.active_time;
-		wm.interlaced = false;
-		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-			wm.interlaced = true;
-		wm.vsc = radeon_crtc->vsc;
-		wm.vtaps = 1;
-		if (radeon_crtc->rmx_type != RMX_OFF)
-			wm.vtaps = 2;
-		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
-		wm.lb_size = lb_size;
 		if (rdev->family == CHIP_ARUBA)
-			wm.dram_channels = evergreen_get_number_of_dram_channels(rdev);
+			dram_channels = evergreen_get_number_of_dram_channels(rdev);
 		else
-			wm.dram_channels = si_get_number_of_dram_channels(rdev);
-		wm.num_heads = num_heads;
+			dram_channels = si_get_number_of_dram_channels(rdev);
+
+		/* watermark for high clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_high.yclk =
+				radeon_dpm_get_mclk(rdev, false) * 10;
+			wm_high.sclk =
+				radeon_dpm_get_sclk(rdev, false) * 10;
+		} else {
+			wm_high.yclk = rdev->pm.current_mclk * 10;
+			wm_high.sclk = rdev->pm.current_sclk * 10;
+		}
+
+		wm_high.disp_clk = mode->clock;
+		wm_high.src_width = mode->crtc_hdisplay;
+		wm_high.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_high.blank_time = line_time - wm_high.active_time;
+		wm_high.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_high.interlaced = true;
+		wm_high.vsc = radeon_crtc->vsc;
+		wm_high.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm_high.vtaps = 2;
+		wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_high.lb_size = lb_size;
+		wm_high.dram_channels = dram_channels;
+		wm_high.num_heads = num_heads;
+
+		/* watermark for low clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_low.yclk =
+				radeon_dpm_get_mclk(rdev, true) * 10;
+			wm_low.sclk =
+				radeon_dpm_get_sclk(rdev, true) * 10;
+		} else {
+			wm_low.yclk = rdev->pm.current_mclk * 10;
+			wm_low.sclk = rdev->pm.current_sclk * 10;
+		}
+
+		wm_low.disp_clk = mode->clock;
+		wm_low.src_width = mode->crtc_hdisplay;
+		wm_low.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_low.blank_time = line_time - wm_low.active_time;
+		wm_low.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_low.interlaced = true;
+		wm_low.vsc = radeon_crtc->vsc;
+		wm_low.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm_low.vtaps = 2;
+		wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_low.lb_size = lb_size;
+		wm_low.dram_channels = dram_channels;
+		wm_low.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(dce6_latency_watermark(&wm), (u32)65535);
+		latency_watermark_a = min(dce6_latency_watermark(&wm_high), (u32)65535);
 		/* set for low clocks */
-		/* wm.yclk = low clk; wm.sclk = low clk */
-		latency_watermark_b = min(dce6_latency_watermark(&wm), (u32)65535);
+		latency_watermark_b = min(dce6_latency_watermark(&wm_low), (u32)65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
-		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
-		    !dce6_average_bandwidth_vs_available_bandwidth(&wm) ||
-		    !dce6_check_latency_hiding(&wm) ||
+		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) ||
+		    !dce6_average_bandwidth_vs_available_bandwidth(&wm_high) ||
+		    !dce6_check_latency_hiding(&wm_high) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+			priority_b_cnt |= PRIORITY_ALWAYS_ON;
+		}
+		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) ||
+		    !dce6_average_bandwidth_vs_available_bandwidth(&wm_low) ||
+		    !dce6_check_latency_hiding(&wm_low) ||
 		    (rdev->disp_priority == 2)) {
 			DRM_DEBUG_KMS("force priority to high\n");
 			priority_a_cnt |= PRIORITY_ALWAYS_ON;
@@ -1895,6 +2185,10 @@ static void dce6_program_watermarks(struct radeon_device *rdev,
 	WREG32(PRIORITY_A_CNT + radeon_crtc->crtc_offset, priority_a_cnt);
 	WREG32(PRIORITY_B_CNT + radeon_crtc->crtc_offset, priority_b_cnt);
 
+	/* save values for DPM */
+	radeon_crtc->line_time = line_time;
+	radeon_crtc->wm_high = latency_watermark_a;
+	radeon_crtc->wm_low = latency_watermark_b;
 }
 
 void dce6_bandwidth_update(struct radeon_device *rdev)
@@ -3535,8 +3829,8 @@ static void si_mc_program(struct radeon_device *rdev)
 	}
 }
 
-static void si_vram_gtt_location(struct radeon_device *rdev,
-				 struct radeon_mc *mc)
+void si_vram_gtt_location(struct radeon_device *rdev,
+			  struct radeon_mc *mc)
 {
 	if (mc->mc_vram_size > 0xFFC0000000ULL) {
 		/* leave room for at least 1024M GTT */
@@ -4282,6 +4576,450 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 }
 
 /*
+ *  Power and clock gating
+ */
+static void si_wait_for_rlc_serdes(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(RLC_SERDES_MASTER_BUSY_0) == 0)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(RLC_SERDES_MASTER_BUSY_1) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void si_enable_gui_idle_interrupt(struct radeon_device *rdev,
+					 bool enable)
+{
+	u32 tmp = RREG32(CP_INT_CNTL_RING0);
+	u32 mask;
+	int i;
+
+	if (enable)
+		tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	else
+		tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	if (!enable) {
+		/* read a gfx register */
+		tmp = RREG32(DB_DEPTH_INFO);
+
+		mask = RLC_BUSY_STATUS | GFX_POWER_STATUS | GFX_CLOCK_STATUS | GFX_LS_STATUS;
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if ((RREG32(RLC_STAT) & mask) == (GFX_CLOCK_STATUS | GFX_POWER_STATUS))
+				break;
+			udelay(1);
+		}
+	}
+}
+
+static void si_set_uvd_dcm(struct radeon_device *rdev,
+			   bool sw_mode)
+{
+	u32 tmp, tmp2;
+
+	tmp = RREG32(UVD_CGC_CTRL);
+	tmp &= ~(CLK_OD_MASK | CG_DT_MASK);
+	tmp |= DCM | CG_DT(1) | CLK_OD(4);
+
+	if (sw_mode) {
+		tmp &= ~0x7ffff800;
+		tmp2 = DYN_OR_EN | DYN_RR_EN | G_DIV_ID(7);
+	} else {
+		tmp |= 0x7ffff800;
+		tmp2 = 0;
+	}
+
+	WREG32(UVD_CGC_CTRL, tmp);
+	WREG32_UVD_CTX(UVD_CGC_CTRL2, tmp2);
+}
+
+static void si_init_uvd_internal_cg(struct radeon_device *rdev)
+{
+	bool hw_mode = true;
+
+	if (hw_mode) {
+		si_set_uvd_dcm(rdev, false);
+	} else {
+		u32 tmp = RREG32(UVD_CGC_CTRL);
+		tmp &= ~DCM;
+		WREG32(UVD_CGC_CTRL, tmp);
+	}
+}
+
+static u32 si_halt_rlc(struct radeon_device *rdev)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(RLC_CNTL);
+
+	if (data & RLC_ENABLE) {
+		data &= ~RLC_ENABLE;
+		WREG32(RLC_CNTL, data);
+
+		si_wait_for_rlc_serdes(rdev);
+	}
+
+	return orig;
+}
+
+static void si_update_rlc(struct radeon_device *rdev, u32 rlc)
+{
+	u32 tmp;
+
+	tmp = RREG32(RLC_CNTL);
+	if (tmp != rlc)
+		WREG32(RLC_CNTL, rlc);
+}
+
+static void si_enable_dma_pg(struct radeon_device *rdev, bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(DMA_PG);
+	if (enable)
+		data |= PG_CNTL_ENABLE;
+	else
+		data &= ~PG_CNTL_ENABLE;
+	if (orig != data)
+		WREG32(DMA_PG, data);
+}
+
+static void si_init_dma_pg(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(DMA_PGFSM_WRITE,  0x00002000);
+	WREG32(DMA_PGFSM_CONFIG, 0x100010ff);
+
+	for (tmp = 0; tmp < 5; tmp++)
+		WREG32(DMA_PGFSM_WRITE, 0);
+}
+
+static void si_enable_gfx_cgpg(struct radeon_device *rdev,
+			       bool enable)
+{
+	u32 tmp;
+
+	if (enable) {
+		tmp = RLC_PUD(0x10) | RLC_PDD(0x10) | RLC_TTPD(0x10) | RLC_MSD(0x10);
+		WREG32(RLC_TTOP_D, tmp);
+
+		tmp = RREG32(RLC_PG_CNTL);
+		tmp |= GFX_PG_ENABLE;
+		WREG32(RLC_PG_CNTL, tmp);
+
+		tmp = RREG32(RLC_AUTO_PG_CTRL);
+		tmp |= AUTO_PG_EN;
+		WREG32(RLC_AUTO_PG_CTRL, tmp);
+	} else {
+		tmp = RREG32(RLC_AUTO_PG_CTRL);
+		tmp &= ~AUTO_PG_EN;
+		WREG32(RLC_AUTO_PG_CTRL, tmp);
+
+		tmp = RREG32(DB_RENDER_CONTROL);
+	}
+}
+
+static void si_init_gfx_cgpg(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+
+	tmp = RREG32(RLC_PG_CNTL);
+	tmp |= GFX_PG_SRC;
+	WREG32(RLC_PG_CNTL, tmp);
+
+	WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+
+	tmp = RREG32(RLC_AUTO_PG_CTRL);
+
+	tmp &= ~GRBM_REG_SGIT_MASK;
+	tmp |= GRBM_REG_SGIT(0x700);
+	tmp &= ~PG_AFTER_GRBM_REG_ST_MASK;
+	WREG32(RLC_AUTO_PG_CTRL, tmp);
+}
+
+static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh)
+{
+	u32 mask = 0, tmp, tmp1;
+	int i;
+
+	si_select_se_sh(rdev, se, sh);
+	tmp = RREG32(CC_GC_SHADER_ARRAY_CONFIG);
+	tmp1 = RREG32(GC_USER_SHADER_ARRAY_CONFIG);
+	si_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+
+	tmp &= 0xffff0000;
+
+	tmp |= tmp1;
+	tmp >>= 16;
+
+	for (i = 0; i < rdev->config.si.max_cu_per_sh; i ++) {
+		mask <<= 1;
+		mask |= 1;
+	}
+
+	return (~tmp) & mask;
+}
+
+static void si_init_ao_cu_mask(struct radeon_device *rdev)
+{
+	u32 i, j, k, active_cu_number = 0;
+	u32 mask, counter, cu_bitmap;
+	u32 tmp = 0;
+
+	for (i = 0; i < rdev->config.si.max_shader_engines; i++) {
+		for (j = 0; j < rdev->config.si.max_sh_per_se; j++) {
+			mask = 1;
+			cu_bitmap = 0;
+			counter  = 0;
+			for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) {
+				if (si_get_cu_active_bitmap(rdev, i, j) & mask) {
+					if (counter < 2)
+						cu_bitmap |= mask;
+					counter++;
+				}
+				mask <<= 1;
+			}
+
+			active_cu_number += counter;
+			tmp |= (cu_bitmap << (i * 16 + j * 8));
+		}
+	}
+
+	WREG32(RLC_PG_AO_CU_MASK, tmp);
+
+	tmp = RREG32(RLC_MAX_PG_CU);
+	tmp &= ~MAX_PU_CU_MASK;
+	tmp |= MAX_PU_CU(active_cu_number);
+	WREG32(RLC_MAX_PG_CU, tmp);
+}
+
+static void si_enable_cgcg(struct radeon_device *rdev,
+			   bool enable)
+{
+	u32 data, orig, tmp;
+
+	orig = data = RREG32(RLC_CGCG_CGLS_CTRL);
+
+	si_enable_gui_idle_interrupt(rdev, enable);
+
+	if (enable) {
+		WREG32(RLC_GCPM_GENERAL_3, 0x00000080);
+
+		tmp = si_halt_rlc(rdev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00b000ff);
+
+		si_wait_for_rlc_serdes(rdev);
+
+		si_update_rlc(rdev, tmp);
+
+		WREG32(RLC_SERDES_WR_CTRL, 0x007000ff);
+
+		data |= CGCG_EN | CGLS_EN;
+	} else {
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+
+		data &= ~(CGCG_EN | CGLS_EN);
+	}
+
+	if (orig != data)
+		WREG32(RLC_CGCG_CGLS_CTRL, data);
+}
+
+static void si_enable_mgcg(struct radeon_device *rdev,
+			   bool enable)
+{
+	u32 data, orig, tmp = 0;
+
+	if (enable) {
+		orig = data = RREG32(CGTS_SM_CTRL_REG);
+		data = 0x96940200;
+		if (orig != data)
+			WREG32(CGTS_SM_CTRL_REG, data);
+
+		orig = data = RREG32(CP_MEM_SLP_CNTL);
+		data |= CP_MEM_LS_EN;
+		if (orig != data)
+			WREG32(CP_MEM_SLP_CNTL, data);
+
+		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data &= 0xffffffc0;
+		if (orig != data)
+			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
+
+		tmp = si_halt_rlc(rdev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00d000ff);
+
+		si_update_rlc(rdev, tmp);
+	} else {
+		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data |= 0x00000003;
+		if (orig != data)
+			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
+
+		data = RREG32(CP_MEM_SLP_CNTL);
+		if (data & CP_MEM_LS_EN) {
+			data &= ~CP_MEM_LS_EN;
+			WREG32(CP_MEM_SLP_CNTL, data);
+		}
+		orig = data = RREG32(CGTS_SM_CTRL_REG);
+		data |= LS_OVERRIDE | OVERRIDE;
+		if (orig != data)
+			WREG32(CGTS_SM_CTRL_REG, data);
+
+		tmp = si_halt_rlc(rdev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00e000ff);
+
+		si_update_rlc(rdev, tmp);
+	}
+}
+
+static void si_enable_uvd_mgcg(struct radeon_device *rdev,
+			       bool enable)
+{
+	u32 orig, data, tmp;
+
+	if (enable) {
+		tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL);
+		tmp |= 0x3fff;
+		WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp);
+
+		orig = data = RREG32(UVD_CGC_CTRL);
+		data |= DCM;
+		if (orig != data)
+			WREG32(UVD_CGC_CTRL, data);
+
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_0, 0);
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_1, 0);
+	} else {
+		tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL);
+		tmp &= ~0x3fff;
+		WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp);
+
+		orig = data = RREG32(UVD_CGC_CTRL);
+		data &= ~DCM;
+		if (orig != data)
+			WREG32(UVD_CGC_CTRL, data);
+
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_0, 0xffffffff);
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_1, 0xffffffff);
+	}
+}
+
+static const u32 mc_cg_registers[] =
+{
+	MC_HUB_MISC_HUB_CG,
+	MC_HUB_MISC_SIP_CG,
+	MC_HUB_MISC_VM_CG,
+	MC_XPB_CLK_GAT,
+	ATC_MISC_CG,
+	MC_CITF_MISC_WR_CG,
+	MC_CITF_MISC_RD_CG,
+	MC_CITF_MISC_VM_CG,
+	VM_L2_CG,
+};
+
+static void si_enable_mc_ls(struct radeon_device *rdev,
+			    bool enable)
+{
+	int i;
+	u32 orig, data;
+
+	for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
+		orig = data = RREG32(mc_cg_registers[i]);
+		if (enable)
+			data |= MC_LS_ENABLE;
+		else
+			data &= ~MC_LS_ENABLE;
+		if (data != orig)
+			WREG32(mc_cg_registers[i], data);
+	}
+}
+
+
+static void si_init_cg(struct radeon_device *rdev)
+{
+	bool has_uvd = true;
+
+	si_enable_mgcg(rdev, true);
+	si_enable_cgcg(rdev, true);
+	/* disable MC LS on Tahiti */
+	if (rdev->family == CHIP_TAHITI)
+		si_enable_mc_ls(rdev, false);
+	if (has_uvd) {
+		si_enable_uvd_mgcg(rdev, true);
+		si_init_uvd_internal_cg(rdev);
+	}
+}
+
+static void si_fini_cg(struct radeon_device *rdev)
+{
+	bool has_uvd = true;
+
+	if (has_uvd)
+		si_enable_uvd_mgcg(rdev, false);
+	si_enable_cgcg(rdev, false);
+	si_enable_mgcg(rdev, false);
+}
+
+static void si_init_pg(struct radeon_device *rdev)
+{
+	bool has_pg = false;
+
+	/* only cape verde supports PG */
+	if (rdev->family == CHIP_VERDE)
+		has_pg = true;
+
+	if (has_pg) {
+		si_init_ao_cu_mask(rdev);
+		si_init_dma_pg(rdev);
+		si_enable_dma_pg(rdev, true);
+		si_init_gfx_cgpg(rdev);
+		si_enable_gfx_cgpg(rdev, true);
+	} else {
+		WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+		WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	}
+}
+
+static void si_fini_pg(struct radeon_device *rdev)
+{
+	bool has_pg = false;
+
+	/* only cape verde supports PG */
+	if (rdev->family == CHIP_VERDE)
+		has_pg = true;
+
+	if (has_pg) {
+		si_enable_dma_pg(rdev, false);
+		si_enable_gfx_cgpg(rdev, false);
+	}
+}
+
+/*
  * RLC
  */
 void si_rlc_fini(struct radeon_device *rdev)
@@ -4313,8 +5051,15 @@ void si_rlc_fini(struct radeon_device *rdev)
 	}
 }
 
+#define RLC_CLEAR_STATE_END_MARKER          0x00000001
+
 int si_rlc_init(struct radeon_device *rdev)
 {
+	volatile u32 *dst_ptr;
+	u32 dws, data, i, j, k, reg_num;
+	u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index;
+	u64 reg_list_mc_addr;
+	const struct cs_section_def *cs_data = si_cs_data;
 	int r;
 
 	/* save restore block */
@@ -4335,18 +5080,44 @@ int si_rlc_init(struct radeon_device *rdev)
 	}
 	r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM,
 			  &rdev->rlc.save_restore_gpu_addr);
-	radeon_bo_unreserve(rdev->rlc.save_restore_obj);
 	if (r) {
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
 		dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r);
 		si_rlc_fini(rdev);
 		return r;
 	}
 
+	if (rdev->family == CHIP_VERDE) {
+		r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r);
+			si_rlc_fini(rdev);
+		return r;
+		}
+		/* write the sr buffer */
+		dst_ptr = rdev->rlc.sr_ptr;
+		for (i = 0; i < ARRAY_SIZE(verde_rlc_save_restore_register_list); i++) {
+			dst_ptr[i] = verde_rlc_save_restore_register_list[i];
+		}
+		radeon_bo_kunmap(rdev->rlc.save_restore_obj);
+	}
+	radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
 	/* clear state block */
+	reg_list_num = 0;
+	dws = 0;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_list_num++;
+			dws += cs_data[i].section[j].reg_count;
+		}
+	}
+	reg_list_blk_index = (3 * reg_list_num + 2);
+	dws += reg_list_blk_index;
+
 	if (rdev->rlc.clear_state_obj == NULL) {
-		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_VRAM, NULL,
-				     &rdev->rlc.clear_state_obj);
+		r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
 			si_rlc_fini(rdev);
@@ -4360,24 +5131,113 @@ int si_rlc_init(struct radeon_device *rdev)
 	}
 	r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM,
 			  &rdev->rlc.clear_state_gpu_addr);
-	radeon_bo_unreserve(rdev->rlc.clear_state_obj);
 	if (r) {
+
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
 		dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r);
 		si_rlc_fini(rdev);
 		return r;
 	}
+	r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r);
+		si_rlc_fini(rdev);
+		return r;
+	}
+	/* set up the cs buffer */
+	dst_ptr = rdev->rlc.cs_ptr;
+	reg_list_hdr_blk_index = 0;
+	reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4);
+	data = upper_32_bits(reg_list_mc_addr);
+	dst_ptr[reg_list_hdr_blk_index] = data;
+	reg_list_hdr_blk_index++;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_num = cs_data[i].section[j].reg_count;
+			data = reg_list_mc_addr & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = 0x08000000 | (reg_num * 4);
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			for (k = 0; k < reg_num; k++) {
+				data = cs_data[i].section[j].extent[k];
+				dst_ptr[reg_list_blk_index + k] = data;
+			}
+			reg_list_mc_addr += reg_num * 4;
+			reg_list_blk_index += reg_num;
+		}
+	}
+	dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER;
+
+	radeon_bo_kunmap(rdev->rlc.clear_state_obj);
+	radeon_bo_unreserve(rdev->rlc.clear_state_obj);
 
 	return 0;
 }
 
+static void si_rlc_reset(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(GRBM_SOFT_RESET);
+
+	tmp |= SOFT_RESET_RLC;
+	WREG32(GRBM_SOFT_RESET, tmp);
+	udelay(50);
+	tmp &= ~SOFT_RESET_RLC;
+	WREG32(GRBM_SOFT_RESET, tmp);
+	udelay(50);
+}
+
 static void si_rlc_stop(struct radeon_device *rdev)
 {
 	WREG32(RLC_CNTL, 0);
+
+	si_enable_gui_idle_interrupt(rdev, false);
+
+	si_wait_for_rlc_serdes(rdev);
 }
 
 static void si_rlc_start(struct radeon_device *rdev)
 {
 	WREG32(RLC_CNTL, RLC_ENABLE);
+
+	si_enable_gui_idle_interrupt(rdev, true);
+
+	udelay(50);
+}
+
+static bool si_lbpw_supported(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	/* Enable LBPW only for DDR3 */
+	tmp = RREG32(MC_SEQ_MISC0);
+	if ((tmp & 0xF0000000) == 0xB0000000)
+		return true;
+	return false;
+}
+
+static void si_enable_lbpw(struct radeon_device *rdev, bool enable)
+{
+	u32 tmp;
+
+	tmp = RREG32(RLC_LB_CNTL);
+	if (enable)
+		tmp |= LOAD_BALANCE_ENABLE;
+	else
+		tmp &= ~LOAD_BALANCE_ENABLE;
+	WREG32(RLC_LB_CNTL, tmp);
+
+	if (!enable) {
+		si_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+		WREG32(SPI_LB_CU_MASK, 0x00ff);
+	}
 }
 
 static int si_rlc_resume(struct radeon_device *rdev)
@@ -4390,14 +5250,18 @@ static int si_rlc_resume(struct radeon_device *rdev)
 
 	si_rlc_stop(rdev);
 
+	si_rlc_reset(rdev);
+
+	si_init_pg(rdev);
+
+	si_init_cg(rdev);
+
 	WREG32(RLC_RL_BASE, 0);
 	WREG32(RLC_RL_SIZE, 0);
 	WREG32(RLC_LB_CNTL, 0);
 	WREG32(RLC_LB_CNTR_MAX, 0xffffffff);
 	WREG32(RLC_LB_CNTR_INIT, 0);
-
-	WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
-	WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	WREG32(RLC_LB_INIT_CU_MASK, 0xffffffff);
 
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
@@ -4409,6 +5273,8 @@ static int si_rlc_resume(struct radeon_device *rdev)
 	}
 	WREG32(RLC_UCODE_ADDR, 0);
 
+	si_enable_lbpw(rdev, si_lbpw_supported(rdev));
+
 	si_rlc_start(rdev);
 
 	return 0;
@@ -4578,6 +5444,7 @@ int si_irq_set(struct radeon_device *rdev)
 	u32 grbm_int_cntl = 0;
 	u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
 	u32 dma_cntl, dma_cntl1;
+	u32 thermal_int = 0;
 
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
@@ -4603,6 +5470,9 @@ int si_irq_set(struct radeon_device *rdev)
 	dma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
 	dma_cntl1 = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
 
+	thermal_int = RREG32(CG_THERMAL_INT) &
+		~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+
 	/* enable CP interrupts on all rings */
 	if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) {
 		DRM_DEBUG("si_irq_set: sw int gfx\n");
@@ -4689,6 +5559,11 @@ int si_irq_set(struct radeon_device *rdev)
 
 	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
 
+	if (rdev->irq.dpm_thermal) {
+		DRM_DEBUG("dpm thermal\n");
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+	}
+
 	if (rdev->num_crtc >= 2) {
 		WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
 		WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2);
@@ -4724,6 +5599,8 @@ int si_irq_set(struct radeon_device *rdev)
 		WREG32(DC_HPD6_INT_CONTROL, hpd6);
 	}
 
+	WREG32(CG_THERMAL_INT, thermal_int);
+
 	return 0;
 }
 
@@ -4888,6 +5765,7 @@ int si_irq_process(struct radeon_device *rdev)
 	u32 src_id, src_data, ring_id;
 	u32 ring_index;
 	bool queue_hotplug = false;
+	bool queue_thermal = false;
 
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
@@ -5158,6 +6036,16 @@ restart_ih:
 			DRM_DEBUG("IH: DMA trap\n");
 			radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
 			break;
+		case 230: /* thermal low to high */
+			DRM_DEBUG("IH: thermal low to high\n");
+			rdev->pm.dpm.thermal.high_to_low = false;
+			queue_thermal = true;
+			break;
+		case 231: /* thermal high to low */
+			DRM_DEBUG("IH: thermal high to low\n");
+			rdev->pm.dpm.thermal.high_to_low = true;
+			queue_thermal = true;
+			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
 			break;
@@ -5176,6 +6064,8 @@ restart_ih:
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
+	if (queue_thermal && rdev->pm.dpm_enabled)
+		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
 	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
@@ -5270,6 +6160,11 @@ static int si_startup(struct radeon_device *rdev)
 	struct radeon_ring *ring;
 	int r;
 
+	/* enable pcie gen2/3 link */
+	si_pcie_gen3_enable(rdev);
+	/* enable aspm */
+	si_program_aspm(rdev);
+
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
 	    !rdev->rlc_fw || !rdev->mc_fw) {
 		r = si_init_microcode(rdev);
@@ -5609,6 +6504,8 @@ void si_fini(struct radeon_device *rdev)
 	cayman_dma_fini(rdev);
 	si_irq_fini(rdev);
 	si_rlc_fini(rdev);
+	si_fini_cg(rdev);
+	si_fini_pg(rdev);
 	radeon_wb_fini(rdev);
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
@@ -5735,3 +6632,361 @@ int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
 
 	return 0;
 }
+
+static void si_pcie_gen3_enable(struct radeon_device *rdev)
+{
+	struct pci_dev *root = rdev->pdev->bus->self;
+	int bridge_pos, gpu_pos;
+	u32 speed_cntl, mask, current_data_rate;
+	int ret, i;
+	u16 tmp16;
+
+	if (radeon_pcie_gen2 == 0)
+		return;
+
+	if (rdev->flags & RADEON_IS_IGP)
+		return;
+
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		return;
+
+	ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask);
+	if (ret != 0)
+		return;
+
+	if (!(mask & (DRM_PCIE_SPEED_50 | DRM_PCIE_SPEED_80)))
+		return;
+
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	current_data_rate = (speed_cntl & LC_CURRENT_DATA_RATE_MASK) >>
+		LC_CURRENT_DATA_RATE_SHIFT;
+	if (mask & DRM_PCIE_SPEED_80) {
+		if (current_data_rate == 2) {
+			DRM_INFO("PCIE gen 3 link speeds already enabled\n");
+			return;
+		}
+		DRM_INFO("enabling PCIE gen 3 link speeds, disable with radeon.pcie_gen2=0\n");
+	} else if (mask & DRM_PCIE_SPEED_50) {
+		if (current_data_rate == 1) {
+			DRM_INFO("PCIE gen 2 link speeds already enabled\n");
+			return;
+		}
+		DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n");
+	}
+
+	bridge_pos = pci_pcie_cap(root);
+	if (!bridge_pos)
+		return;
+
+	gpu_pos = pci_pcie_cap(rdev->pdev);
+	if (!gpu_pos)
+		return;
+
+	if (mask & DRM_PCIE_SPEED_80) {
+		/* re-try equalization if gen3 is not already enabled */
+		if (current_data_rate != 2) {
+			u16 bridge_cfg, gpu_cfg;
+			u16 bridge_cfg2, gpu_cfg2;
+			u32 max_lw, current_lw, tmp;
+
+			pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg);
+			pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg);
+
+			tmp16 = bridge_cfg | PCI_EXP_LNKCTL_HAWD;
+			pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16);
+
+			tmp16 = gpu_cfg | PCI_EXP_LNKCTL_HAWD;
+			pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16);
+
+			tmp = RREG32_PCIE(PCIE_LC_STATUS1);
+			max_lw = (tmp & LC_DETECTED_LINK_WIDTH_MASK) >> LC_DETECTED_LINK_WIDTH_SHIFT;
+			current_lw = (tmp & LC_OPERATING_LINK_WIDTH_MASK) >> LC_OPERATING_LINK_WIDTH_SHIFT;
+
+			if (current_lw < max_lw) {
+				tmp = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+				if (tmp & LC_RENEGOTIATION_SUPPORT) {
+					tmp &= ~(LC_LINK_WIDTH_MASK | LC_UPCONFIGURE_DIS);
+					tmp |= (max_lw << LC_LINK_WIDTH_SHIFT);
+					tmp |= LC_UPCONFIGURE_SUPPORT | LC_RENEGOTIATE_EN | LC_RECONFIG_NOW;
+					WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, tmp);
+				}
+			}
+
+			for (i = 0; i < 10; i++) {
+				/* check status */
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_DEVSTA, &tmp16);
+				if (tmp16 & PCI_EXP_DEVSTA_TRPND)
+					break;
+
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg);
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg);
+
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &bridge_cfg2);
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &gpu_cfg2);
+
+				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
+				tmp |= LC_SET_QUIESCE;
+				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+
+				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
+				tmp |= LC_REDO_EQ;
+				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+
+				mdelay(100);
+
+				/* linkctl */
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &tmp16);
+				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
+				tmp16 |= (bridge_cfg & PCI_EXP_LNKCTL_HAWD);
+				pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16);
+
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &tmp16);
+				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
+				tmp16 |= (gpu_cfg & PCI_EXP_LNKCTL_HAWD);
+				pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16);
+
+				/* linkctl2 */
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &tmp16);
+				tmp16 &= ~((1 << 4) | (7 << 9));
+				tmp16 |= (bridge_cfg2 & ((1 << 4) | (7 << 9)));
+				pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, tmp16);
+
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16);
+				tmp16 &= ~((1 << 4) | (7 << 9));
+				tmp16 |= (gpu_cfg2 & ((1 << 4) | (7 << 9)));
+				pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16);
+
+				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
+				tmp &= ~LC_SET_QUIESCE;
+				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+			}
+		}
+	}
+
+	/* set the link speed */
+	speed_cntl |= LC_FORCE_EN_SW_SPEED_CHANGE | LC_FORCE_DIS_HW_SPEED_CHANGE;
+	speed_cntl &= ~LC_FORCE_DIS_SW_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
+
+	pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16);
+	tmp16 &= ~0xf;
+	if (mask & DRM_PCIE_SPEED_80)
+		tmp16 |= 3; /* gen3 */
+	else if (mask & DRM_PCIE_SPEED_50)
+		tmp16 |= 2; /* gen2 */
+	else
+		tmp16 |= 1; /* gen1 */
+	pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16);
+
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	speed_cntl |= LC_INITIATE_LINK_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+		if ((speed_cntl & LC_INITIATE_LINK_SPEED_CHANGE) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void si_program_aspm(struct radeon_device *rdev)
+{
+	u32 data, orig;
+	bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false;
+	bool disable_clkreq = false;
+
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		return;
+
+	orig = data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL);
+	data &= ~LC_XMIT_N_FTS_MASK;
+	data |= LC_XMIT_N_FTS(0x24) | LC_XMIT_N_FTS_OVERRIDE_EN;
+	if (orig != data)
+		WREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL, data);
+
+	orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL3);
+	data |= LC_GO_TO_RECOVERY;
+	if (orig != data)
+		WREG32_PCIE_PORT(PCIE_LC_CNTL3, data);
+
+	orig = data = RREG32_PCIE(PCIE_P_CNTL);
+	data |= P_IGNORE_EDB_ERR;
+	if (orig != data)
+		WREG32_PCIE(PCIE_P_CNTL, data);
+
+	orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	data &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK);
+	data |= LC_PMI_TO_L1_DIS;
+	if (!disable_l0s)
+		data |= LC_L0S_INACTIVITY(7);
+
+	if (!disable_l1) {
+		data |= LC_L1_INACTIVITY(7);
+		data &= ~LC_PMI_TO_L1_DIS;
+		if (orig != data)
+			WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+
+		if (!disable_plloff_in_l1) {
+			bool clk_req_support;
+
+			orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (orig != data)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+			orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (orig != data)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+			orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (orig != data)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+			orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (orig != data)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+
+			if ((rdev->family != CHIP_OLAND) && (rdev->family != CHIP_HAINAN)) {
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_2);
+				data &= ~PLL_RAMP_UP_TIME_2_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_2, data);
+
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_3);
+				data &= ~PLL_RAMP_UP_TIME_3_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_3, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_2);
+				data &= ~PLL_RAMP_UP_TIME_2_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_2, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_3);
+				data &= ~PLL_RAMP_UP_TIME_3_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_3, data);
+			}
+			orig = data = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+			data &= ~LC_DYN_LANES_PWR_STATE_MASK;
+			data |= LC_DYN_LANES_PWR_STATE(3);
+			if (orig != data)
+				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
+
+			orig = data = RREG32_PIF_PHY0(PB0_PIF_CNTL);
+			data &= ~LS2_EXIT_TIME_MASK;
+			if ((rdev->family == CHIP_OLAND) || (rdev->family == CHIP_HAINAN))
+				data |= LS2_EXIT_TIME(5);
+			if (orig != data)
+				WREG32_PIF_PHY0(PB0_PIF_CNTL, data);
+
+			orig = data = RREG32_PIF_PHY1(PB1_PIF_CNTL);
+			data &= ~LS2_EXIT_TIME_MASK;
+			if ((rdev->family == CHIP_OLAND) || (rdev->family == CHIP_HAINAN))
+				data |= LS2_EXIT_TIME(5);
+			if (orig != data)
+				WREG32_PIF_PHY1(PB1_PIF_CNTL, data);
+
+			if (!disable_clkreq) {
+				struct pci_dev *root = rdev->pdev->bus->self;
+				u32 lnkcap;
+
+				clk_req_support = false;
+				pcie_capability_read_dword(root, PCI_EXP_LNKCAP, &lnkcap);
+				if (lnkcap & PCI_EXP_LNKCAP_CLKPM)
+					clk_req_support = true;
+			} else {
+				clk_req_support = false;
+			}
+
+			if (clk_req_support) {
+				orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL2);
+				data |= LC_ALLOW_PDWN_IN_L1 | LC_ALLOW_PDWN_IN_L23;
+				if (orig != data)
+					WREG32_PCIE_PORT(PCIE_LC_CNTL2, data);
+
+				orig = data = RREG32(THM_CLK_CNTL);
+				data &= ~(CMON_CLK_SEL_MASK | TMON_CLK_SEL_MASK);
+				data |= CMON_CLK_SEL(1) | TMON_CLK_SEL(1);
+				if (orig != data)
+					WREG32(THM_CLK_CNTL, data);
+
+				orig = data = RREG32(MISC_CLK_CNTL);
+				data &= ~(DEEP_SLEEP_CLK_SEL_MASK | ZCLK_SEL_MASK);
+				data |= DEEP_SLEEP_CLK_SEL(1) | ZCLK_SEL(1);
+				if (orig != data)
+					WREG32(MISC_CLK_CNTL, data);
+
+				orig = data = RREG32(CG_CLKPIN_CNTL);
+				data &= ~BCLK_AS_XCLK;
+				if (orig != data)
+					WREG32(CG_CLKPIN_CNTL, data);
+
+				orig = data = RREG32(CG_CLKPIN_CNTL_2);
+				data &= ~FORCE_BIF_REFCLK_EN;
+				if (orig != data)
+					WREG32(CG_CLKPIN_CNTL_2, data);
+
+				orig = data = RREG32(MPLL_BYPASSCLK_SEL);
+				data &= ~MPLL_CLKOUT_SEL_MASK;
+				data |= MPLL_CLKOUT_SEL(4);
+				if (orig != data)
+					WREG32(MPLL_BYPASSCLK_SEL, data);
+
+				orig = data = RREG32(SPLL_CNTL_MODE);
+				data &= ~SPLL_REFCLK_SEL_MASK;
+				if (orig != data)
+					WREG32(SPLL_CNTL_MODE, data);
+			}
+		}
+	} else {
+		if (orig != data)
+			WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+	}
+
+	orig = data = RREG32_PCIE(PCIE_CNTL2);
+	data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | REPLAY_MEM_LS_EN;
+	if (orig != data)
+		WREG32_PCIE(PCIE_CNTL2, data);
+
+	if (!disable_l0s) {
+		data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL);
+		if((data & LC_N_FTS_MASK) == LC_N_FTS_MASK) {
+			data = RREG32_PCIE(PCIE_LC_STATUS1);
+			if ((data & LC_REVERSE_XMIT) && (data & LC_REVERSE_RCVR)) {
+				orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+				data &= ~LC_L0S_INACTIVITY_MASK;
+				if (orig != data)
+					WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+			}
+		}
+	}
+}
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
new file mode 100644
index 0000000..73aaa2e
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -0,0 +1,6432 @@
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "sid.h"
+#include "r600_dpm.h"
+#include "si_dpm.h"
+#include "atom.h"
+#include <linux/math64.h>
+#include <linux/seq_file.h>
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define SMC_RAM_END                 0x20000
+
+#define DDR3_DRAM_ROWS              0x2000
+
+#define SCLK_MIN_DEEPSLEEP_FREQ     1350
+
+static const struct si_cac_config_reg cac_weights_tahiti[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0xc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x101, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0xc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x8fc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x95, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x34e, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x1a1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0xda, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x46, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x208, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0xe7, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x948, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x167, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x31, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x18e, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_tahiti[] =
+{
+	{ 0x143, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+
+};
+
+static const struct si_cac_config_reg cac_override_tahiti[] =
+{
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_tahiti =
+{
+	((1 << 16) | 27027),
+	6,
+	0,
+	4,
+	95,
+	{
+		0UL,
+		0UL,
+		4521550UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		40
+	},
+	595000000UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_tahiti =
+{
+	{ 1159409, 0, 0, 0, 0 },
+	{ 777, 0, 0, 0, 0 },
+	2,
+	54000,
+	127000,
+	25,
+	2,
+	10,
+	13,
+	{ 27, 31, 35, 39, 43, 47, 54, 61, 67, 74, 81, 88, 95, 0, 0, 0 },
+	{ 240888759, 221057860, 235370597, 162287531, 158510299, 131423027, 116673180, 103067515, 87941937, 76209048, 68209175, 64090048, 58301890, 0, 0, 0 },
+	{ 12024, 11189, 11451, 8411, 7939, 6666, 5681, 4905, 4241, 3720, 3354, 3122, 2890, 0, 0, 0 },
+	85,
+	false
+};
+
+static const struct si_dte_data dte_data_tahiti_le =
+{
+	{ 0x1E8480, 0x7A1200, 0x2160EC0, 0x3938700, 0 },
+	{ 0x7D, 0x7D, 0x4E4, 0xB00, 0 },
+	0x5,
+	0xAFC8,
+	0x64,
+	0x32,
+	1,
+	0,
+	0x10,
+	{ 0x78, 0x7C, 0x82, 0x88, 0x8E, 0x94, 0x9A, 0xA0, 0xA6, 0xAC, 0xB0, 0xB4, 0xB8, 0xBC, 0xC0, 0xC4 },
+	{ 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700 },
+	{ 0x2AF8, 0x2AF8, 0x29BB, 0x27F9, 0x2637, 0x2475, 0x22B3, 0x20F1, 0x1F2F, 0x1D6D, 0x1734, 0x1414, 0x10F4, 0xDD4, 0xAB4, 0x794 },
+	85,
+	true
+};
+
+static const struct si_dte_data dte_data_tahiti_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x7D0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_new_zealand =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0 },
+	{ 0x29B, 0x3E9, 0x537, 0x7D2, 0 },
+	0x5,
+	0xAFC8,
+	0x69,
+	0x32,
+	1,
+	0,
+	0x10,
+	{ 0x82, 0xA0, 0xB4, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0xDAC, 0x1388, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685 },
+	85,
+	true
+};
+
+static const struct si_dte_data dte_data_aruba_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x1000, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_malta =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+struct si_cac_config_reg cac_weights_pitcairn[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x8a, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x24d, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x19, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0xc11, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x7f3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x403, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x367, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x4c9, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x45d, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x36d, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x534, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x5da, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x880, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0x201, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x9f, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x1f, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5de, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x7b, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x13, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0xf9, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x66, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x13, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x186, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_pitcairn[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_override_pitcairn[] =
+{
+    { 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_pitcairn =
+{
+	((1 << 16) | 27027),
+	5,
+	0,
+	6,
+	100,
+	{
+		51600000UL,
+		1800000UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_pitcairn =
+{
+	{ 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0 },
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	0,
+	false
+};
+
+static const struct si_dte_data dte_data_curacao_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x1D17, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_curacao_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x1D17, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_neptune_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x3A2F, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_cac_config_reg cac_weights_chelsea_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x2BD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_chelsea_xt[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x30A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_heathrow[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x362, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_cape_verde_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x315, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_cape_verde[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x3BA, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_cape_verde[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_override_cape_verde[] =
+{
+    { 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_cape_verde =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	7,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_cape_verde =
+{
+	{ 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0 },
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	0,
+	false
+};
+
+static const struct si_dte_data dte_data_venus_xtx =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x71C, 0xAAB, 0xE39, 0x11C7, 0x0 },
+	5,
+	55000,
+	0x69,
+	0xA,
+	1,
+	0,
+	0x3,
+	{ 0x96, 0xB4, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x895440, 0x3D0900, 0x989680, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0xD6D8, 0x88B8, 0x1555, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_venus_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0xBDA, 0x11C7, 0x17B4, 0x1DA1, 0x0 },
+	5,
+	55000,
+	0x69,
+	0xA,
+	1,
+	0,
+	0x3,
+	{ 0x96, 0xB4, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x895440, 0x3D0900, 0x989680, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0xAFC8, 0x88B8, 0x238E, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_venus_pro =
+{
+	{  0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x11C7, 0x1AAB, 0x238E, 0x2C72, 0x0 },
+	5,
+	55000,
+	0x69,
+	0xA,
+	1,
+	0,
+	0x3,
+	{ 0x96, 0xB4, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x895440, 0x3D0900, 0x989680, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x88B8, 0x88B8, 0x3555, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+struct si_cac_config_reg cac_weights_oland[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x3BA, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_mars_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_mars_xt[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x60, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_oland_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x90, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_oland_xt[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x120, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_oland[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_mars_pro[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_override_oland[] =
+{
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_oland =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	7,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_powertune_data powertune_data_mars_pro =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	7,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_oland =
+{
+	{ 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0 },
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	0,
+	false
+};
+
+static const struct si_dte_data dte_data_mars_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	55000,
+	105,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0xF627, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_sun_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	55000,
+	105,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0xD555, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+
+static const struct si_cac_config_reg cac_weights_hainan[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x2d9, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x22b, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x21c, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x1dc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x24e, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x35e, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1143, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0xe17, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x441, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x28b, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0xabe, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xf11, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x907, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0xb45, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0xd1e, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xa2c, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x62, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x1f3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x42, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x709, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x3a, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x357, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x9f, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x314, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x6d, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x1b9, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_hainan =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	9,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+struct ni_power_info *ni_get_pi(struct radeon_device *rdev);
+struct ni_ps *ni_get_ps(struct radeon_ps *rps);
+
+static int si_populate_voltage_value(struct radeon_device *rdev,
+				     const struct atom_voltage_table *table,
+				     u16 value, SISLANDS_SMC_VOLTAGE_VALUE *voltage);
+static int si_get_std_voltage_value(struct radeon_device *rdev,
+				    SISLANDS_SMC_VOLTAGE_VALUE *voltage,
+				    u16 *std_voltage);
+static int si_write_smc_soft_register(struct radeon_device *rdev,
+				      u16 reg_offset, u32 value);
+static int si_convert_power_level_to_smc(struct radeon_device *rdev,
+					 struct rv7xx_pl *pl,
+					 SISLANDS_SMC_HW_PERFORMANCE_LEVEL *level);
+static int si_calculate_sclk_params(struct radeon_device *rdev,
+				    u32 engine_clock,
+				    SISLANDS_SMC_SCLK_VALUE *sclk);
+
+static struct si_power_info *si_get_pi(struct radeon_device *rdev)
+{
+        struct si_power_info *pi = rdev->pm.dpm.priv;
+
+        return pi;
+}
+
+static void si_calculate_leakage_for_v_and_t_formula(const struct ni_leakage_coeffients *coeff,
+						     u16 v, s32 t, u32 ileakage, u32 *leakage)
+{
+	s64 kt, kv, leakage_w, i_leakage, vddc;
+	s64 temperature, t_slope, t_intercept, av, bv, t_ref;
+
+	i_leakage = drm_int2fixp(ileakage / 100);
+	vddc = div64_s64(drm_int2fixp(v), 1000);
+	temperature = div64_s64(drm_int2fixp(t), 1000);
+
+	t_slope = div64_s64(drm_int2fixp(coeff->t_slope), 100000000);
+	t_intercept = div64_s64(drm_int2fixp(coeff->t_intercept), 100000000);
+	av = div64_s64(drm_int2fixp(coeff->av), 100000000);
+	bv = div64_s64(drm_int2fixp(coeff->bv), 100000000);
+	t_ref = drm_int2fixp(coeff->t_ref);
+
+	kt = drm_fixp_div(drm_fixp_exp(drm_fixp_mul(drm_fixp_mul(t_slope, vddc) + t_intercept, temperature)),
+			  drm_fixp_exp(drm_fixp_mul(drm_fixp_mul(t_slope, vddc) + t_intercept, t_ref)));
+	kv = drm_fixp_mul(av, drm_fixp_exp(drm_fixp_mul(bv, vddc)));
+
+	leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc);
+
+	*leakage = drm_fixp2int(leakage_w * 1000);
+}
+
+static void si_calculate_leakage_for_v_and_t(struct radeon_device *rdev,
+					     const struct ni_leakage_coeffients *coeff,
+					     u16 v,
+					     s32 t,
+					     u32 i_leakage,
+					     u32 *leakage)
+{
+	si_calculate_leakage_for_v_and_t_formula(coeff, v, t, i_leakage, leakage);
+}
+
+static void si_calculate_leakage_for_v_formula(const struct ni_leakage_coeffients *coeff,
+					       const u32 fixed_kt, u16 v,
+					       u32 ileakage, u32 *leakage)
+{
+	s64 kt, kv, leakage_w, i_leakage, vddc;
+
+	i_leakage = div64_s64(drm_int2fixp(ileakage), 100);
+	vddc = div64_s64(drm_int2fixp(v), 1000);
+
+	kt = div64_s64(drm_int2fixp(fixed_kt), 100000000);
+	kv = drm_fixp_mul(div64_s64(drm_int2fixp(coeff->av), 100000000),
+			  drm_fixp_exp(drm_fixp_mul(div64_s64(drm_int2fixp(coeff->bv), 100000000), vddc)));
+
+	leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc);
+
+	*leakage = drm_fixp2int(leakage_w * 1000);
+}
+
+static void si_calculate_leakage_for_v(struct radeon_device *rdev,
+				       const struct ni_leakage_coeffients *coeff,
+				       const u32 fixed_kt,
+				       u16 v,
+				       u32 i_leakage,
+				       u32 *leakage)
+{
+	si_calculate_leakage_for_v_formula(coeff, fixed_kt, v, i_leakage, leakage);
+}
+
+
+static void si_update_dte_from_pl2(struct radeon_device *rdev,
+				   struct si_dte_data *dte_data)
+{
+	u32 p_limit1 = rdev->pm.dpm.tdp_limit;
+	u32 p_limit2 = rdev->pm.dpm.near_tdp_limit;
+	u32 k = dte_data->k;
+	u32 t_max = dte_data->max_t;
+	u32 t_split[5] = { 10, 15, 20, 25, 30 };
+	u32 t_0 = dte_data->t0;
+	u32 i;
+
+	if (p_limit2 != 0 && p_limit2 <= p_limit1) {
+		dte_data->tdep_count = 3;
+
+		for (i = 0; i < k; i++) {
+			dte_data->r[i] =
+				(t_split[i] * (t_max - t_0/(u32)1000) * (1 << 14)) /
+				(p_limit2  * (u32)100);
+		}
+
+		dte_data->tdep_r[1] = dte_data->r[4] * 2;
+
+		for (i = 2; i < SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE; i++) {
+			dte_data->tdep_r[i] = dte_data->r[4];
+		}
+	} else {
+		DRM_ERROR("Invalid PL2! DTE will not be updated.\n");
+	}
+}
+
+static void si_initialize_powertune_defaults(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	bool update_dte_from_pl2 = false;
+
+	if (rdev->family == CHIP_TAHITI) {
+		si_pi->cac_weights = cac_weights_tahiti;
+		si_pi->lcac_config = lcac_tahiti;
+		si_pi->cac_override = cac_override_tahiti;
+		si_pi->powertune_data = &powertune_data_tahiti;
+		si_pi->dte_data = dte_data_tahiti;
+
+		switch (rdev->pdev->device) {
+		case 0x6798:
+			si_pi->dte_data.enable_dte_by_default = true;
+			break;
+		case 0x6799:
+			si_pi->dte_data = dte_data_new_zealand;
+			break;
+		case 0x6790:
+		case 0x6791:
+		case 0x6792:
+		case 0x679E:
+			si_pi->dte_data = dte_data_aruba_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x679B:
+			si_pi->dte_data = dte_data_malta;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x679A:
+			si_pi->dte_data = dte_data_tahiti_pro;
+			update_dte_from_pl2 = true;
+			break;
+		default:
+			if (si_pi->dte_data.enable_dte_by_default == true)
+				DRM_ERROR("DTE is not enabled!\n");
+			break;
+		}
+	} else if (rdev->family == CHIP_PITCAIRN) {
+		switch (rdev->pdev->device) {
+		case 0x6810:
+		case 0x6818:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_curacao_xt;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6819:
+		case 0x6811:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_curacao_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6800:
+		case 0x6806:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_neptune_xt;
+			update_dte_from_pl2 = true;
+			break;
+		default:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_pitcairn;
+		}
+	} else if (rdev->family == CHIP_VERDE) {
+		si_pi->lcac_config = lcac_cape_verde;
+		si_pi->cac_override = cac_override_cape_verde;
+		si_pi->powertune_data = &powertune_data_cape_verde;
+
+		switch (rdev->pdev->device) {
+		case 0x683B:
+		case 0x683F:
+		case 0x6829:
+			si_pi->cac_weights = cac_weights_cape_verde_pro;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x6825:
+		case 0x6827:
+			si_pi->cac_weights = cac_weights_heathrow;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x6824:
+		case 0x682D:
+			si_pi->cac_weights = cac_weights_chelsea_xt;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x682F:
+			si_pi->cac_weights = cac_weights_chelsea_pro;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x6820:
+			si_pi->cac_weights = cac_weights_heathrow;
+			si_pi->dte_data = dte_data_venus_xtx;
+			break;
+		case 0x6821:
+			si_pi->cac_weights = cac_weights_heathrow;
+			si_pi->dte_data = dte_data_venus_xt;
+			break;
+		case 0x6823:
+			si_pi->cac_weights = cac_weights_chelsea_pro;
+			si_pi->dte_data = dte_data_venus_pro;
+			break;
+		case 0x682B:
+			si_pi->cac_weights = cac_weights_chelsea_pro;
+			si_pi->dte_data = dte_data_venus_pro;
+			break;
+		default:
+			si_pi->cac_weights = cac_weights_cape_verde;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		}
+	} else if (rdev->family == CHIP_OLAND) {
+		switch (rdev->pdev->device) {
+		case 0x6601:
+		case 0x6621:
+		case 0x6603:
+			si_pi->cac_weights = cac_weights_mars_pro;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6600:
+		case 0x6606:
+		case 0x6620:
+			si_pi->cac_weights = cac_weights_mars_xt;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6611:
+			si_pi->cac_weights = cac_weights_oland_pro;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6610:
+			si_pi->cac_weights = cac_weights_oland_xt;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		default:
+			si_pi->cac_weights = cac_weights_oland;
+			si_pi->lcac_config = lcac_oland;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_oland;
+			si_pi->dte_data = dte_data_oland;
+			break;
+		}
+	} else if (rdev->family == CHIP_HAINAN) {
+		si_pi->cac_weights = cac_weights_hainan;
+		si_pi->lcac_config = lcac_oland;
+		si_pi->cac_override = cac_override_oland;
+		si_pi->powertune_data = &powertune_data_hainan;
+		si_pi->dte_data = dte_data_sun_xt;
+		update_dte_from_pl2 = true;
+	} else {
+		DRM_ERROR("Unknown SI asic revision, failed to initialize PowerTune!\n");
+		return;
+	}
+
+	ni_pi->enable_power_containment = false;
+	ni_pi->enable_cac = false;
+	ni_pi->enable_sq_ramping = false;
+	si_pi->enable_dte = false;
+
+	if (si_pi->powertune_data->enable_powertune_by_default) {
+		ni_pi->enable_power_containment= true;
+		ni_pi->enable_cac = true;
+		if (si_pi->dte_data.enable_dte_by_default) {
+			si_pi->enable_dte = true;
+			if (update_dte_from_pl2)
+				si_update_dte_from_pl2(rdev, &si_pi->dte_data);
+
+		}
+		ni_pi->enable_sq_ramping = true;
+	}
+
+	ni_pi->driver_calculate_cac_leakage = true;
+	ni_pi->cac_configuration_required = true;
+
+	if (ni_pi->cac_configuration_required) {
+		ni_pi->support_cac_long_term_average = true;
+		si_pi->dyn_powertune_data.l2_lta_window_size =
+			si_pi->powertune_data->l2_lta_window_size_default;
+		si_pi->dyn_powertune_data.lts_truncate =
+			si_pi->powertune_data->lts_truncate_default;
+	} else {
+		ni_pi->support_cac_long_term_average = false;
+		si_pi->dyn_powertune_data.l2_lta_window_size = 0;
+		si_pi->dyn_powertune_data.lts_truncate = 0;
+	}
+
+	si_pi->dyn_powertune_data.disable_uvd_powertune = false;
+}
+
+static u32 si_get_smc_power_scaling_factor(struct radeon_device *rdev)
+{
+	return 1;
+}
+
+static u32 si_calculate_cac_wintime(struct radeon_device *rdev)
+{
+	u32 xclk;
+	u32 wintime;
+	u32 cac_window;
+	u32 cac_window_size;
+
+	xclk = radeon_get_xclk(rdev);
+
+	if (xclk == 0)
+		return 0;
+
+	cac_window = RREG32(CG_CAC_CTRL) & CAC_WINDOW_MASK;
+	cac_window_size = ((cac_window & 0xFFFF0000) >> 16) * (cac_window & 0x0000FFFF);
+
+	wintime = (cac_window_size * 100) / xclk;
+
+	return wintime;
+}
+
+static u32 si_scale_power_for_smc(u32 power_in_watts, u32 scaling_factor)
+{
+	return power_in_watts;
+}
+
+static int si_calculate_adjusted_tdp_limits(struct radeon_device *rdev,
+					    bool adjust_polarity,
+					    u32 tdp_adjustment,
+					    u32 *tdp_limit,
+					    u32 *near_tdp_limit)
+{
+	u32 adjustment_delta, max_tdp_limit;
+
+	if (tdp_adjustment > (u32)rdev->pm.dpm.tdp_od_limit)
+		return -EINVAL;
+
+	max_tdp_limit = ((100 + 100) * rdev->pm.dpm.tdp_limit) / 100;
+
+	if (adjust_polarity) {
+		*tdp_limit = ((100 + tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		*near_tdp_limit = rdev->pm.dpm.near_tdp_limit_adjusted + (*tdp_limit - rdev->pm.dpm.tdp_limit);
+	} else {
+		*tdp_limit = ((100 - tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		adjustment_delta  = rdev->pm.dpm.tdp_limit - *tdp_limit;
+		if (adjustment_delta < rdev->pm.dpm.near_tdp_limit_adjusted)
+			*near_tdp_limit = rdev->pm.dpm.near_tdp_limit_adjusted - adjustment_delta;
+		else
+			*near_tdp_limit = 0;
+	}
+
+	if ((*tdp_limit <= 0) || (*tdp_limit > max_tdp_limit))
+		return -EINVAL;
+	if ((*near_tdp_limit <= 0) || (*near_tdp_limit > *tdp_limit))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int si_populate_smc_tdp_limits(struct radeon_device *rdev,
+				      struct radeon_ps *radeon_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (ni_pi->enable_power_containment) {
+		SISLANDS_SMC_STATETABLE *smc_table = &si_pi->smc_statetable;
+		PP_SIslands_PAPMParameters *papm_parm;
+		struct radeon_ppm_table *ppm = rdev->pm.dpm.dyn_state.ppm_table;
+		u32 scaling_factor = si_get_smc_power_scaling_factor(rdev);
+		u32 tdp_limit;
+		u32 near_tdp_limit;
+		int ret;
+
+		if (scaling_factor == 0)
+			return -EINVAL;
+
+		memset(smc_table, 0, sizeof(SISLANDS_SMC_STATETABLE));
+
+		ret = si_calculate_adjusted_tdp_limits(rdev,
+						       false, /* ??? */
+						       rdev->pm.dpm.tdp_adjustment,
+						       &tdp_limit,
+						       &near_tdp_limit);
+		if (ret)
+			return ret;
+
+		smc_table->dpm2Params.TDPLimit =
+			cpu_to_be32(si_scale_power_for_smc(tdp_limit, scaling_factor) * 1000);
+		smc_table->dpm2Params.NearTDPLimit =
+			cpu_to_be32(si_scale_power_for_smc(near_tdp_limit, scaling_factor) * 1000);
+		smc_table->dpm2Params.SafePowerLimit =
+			cpu_to_be32(si_scale_power_for_smc((near_tdp_limit * SISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT) / 100, scaling_factor) * 1000);
+
+		ret = si_copy_bytes_to_smc(rdev,
+					   (si_pi->state_table_start + offsetof(SISLANDS_SMC_STATETABLE, dpm2Params) +
+						 offsetof(PP_SIslands_DPM2Parameters, TDPLimit)),
+					   (u8 *)(&(smc_table->dpm2Params.TDPLimit)),
+					   sizeof(u32) * 3,
+					   si_pi->sram_end);
+		if (ret)
+			return ret;
+
+		if (si_pi->enable_ppm) {
+			papm_parm = &si_pi->papm_parm;
+			memset(papm_parm, 0, sizeof(PP_SIslands_PAPMParameters));
+			papm_parm->NearTDPLimitTherm = cpu_to_be32(ppm->dgpu_tdp);
+			papm_parm->dGPU_T_Limit = cpu_to_be32(ppm->tj_max);
+			papm_parm->dGPU_T_Warning = cpu_to_be32(95);
+			papm_parm->dGPU_T_Hysteresis = cpu_to_be32(5);
+			papm_parm->PlatformPowerLimit = 0xffffffff;
+			papm_parm->NearTDPLimitPAPM = 0xffffffff;
+
+			ret = si_copy_bytes_to_smc(rdev, si_pi->papm_cfg_table_start,
+						   (u8 *)papm_parm,
+						   sizeof(PP_SIslands_PAPMParameters),
+						   si_pi->sram_end);
+			if (ret)
+				return ret;
+		}
+	}
+	return 0;
+}
+
+static int si_populate_smc_tdp_limits_2(struct radeon_device *rdev,
+					struct radeon_ps *radeon_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (ni_pi->enable_power_containment) {
+		SISLANDS_SMC_STATETABLE *smc_table = &si_pi->smc_statetable;
+		u32 scaling_factor = si_get_smc_power_scaling_factor(rdev);
+		int ret;
+
+		memset(smc_table, 0, sizeof(SISLANDS_SMC_STATETABLE));
+
+		smc_table->dpm2Params.NearTDPLimit =
+			cpu_to_be32(si_scale_power_for_smc(rdev->pm.dpm.near_tdp_limit_adjusted, scaling_factor) * 1000);
+		smc_table->dpm2Params.SafePowerLimit =
+			cpu_to_be32(si_scale_power_for_smc((rdev->pm.dpm.near_tdp_limit_adjusted * SISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT) / 100, scaling_factor) * 1000);
+
+		ret = si_copy_bytes_to_smc(rdev,
+					   (si_pi->state_table_start +
+					    offsetof(SISLANDS_SMC_STATETABLE, dpm2Params) +
+					    offsetof(PP_SIslands_DPM2Parameters, NearTDPLimit)),
+					   (u8 *)(&(smc_table->dpm2Params.NearTDPLimit)),
+					   sizeof(u32) * 2,
+					   si_pi->sram_end);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static u16 si_calculate_power_efficiency_ratio(struct radeon_device *rdev,
+					       const u16 prev_std_vddc,
+					       const u16 curr_std_vddc)
+{
+	u64 margin = (u64)SISLANDS_DPM2_PWREFFICIENCYRATIO_MARGIN;
+	u64 prev_vddc = (u64)prev_std_vddc;
+	u64 curr_vddc = (u64)curr_std_vddc;
+	u64 pwr_efficiency_ratio, n, d;
+
+	if ((prev_vddc == 0) || (curr_vddc == 0))
+		return 0;
+
+	n = div64_u64((u64)1024 * curr_vddc * curr_vddc * ((u64)1000 + margin), (u64)1000);
+	d = prev_vddc * prev_vddc;
+	pwr_efficiency_ratio = div64_u64(n, d);
+
+	if (pwr_efficiency_ratio > (u64)0xFFFF)
+		return 0;
+
+	return (u16)pwr_efficiency_ratio;
+}
+
+static bool si_should_disable_uvd_powertune(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (si_pi->dyn_powertune_data.disable_uvd_powertune &&
+	    radeon_state->vclk && radeon_state->dclk)
+		return true;
+
+	return false;
+}
+
+static int si_populate_power_containment_values(struct radeon_device *rdev,
+						struct radeon_ps *radeon_state,
+						SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	SISLANDS_SMC_VOLTAGE_VALUE vddc;
+	u32 prev_sclk;
+	u32 max_sclk;
+	u32 min_sclk;
+	u16 prev_std_vddc;
+	u16 curr_std_vddc;
+	int i;
+	u16 pwr_efficiency_ratio;
+	u8 max_ps_percent;
+	bool disable_uvd_power_tune;
+	int ret;
+
+	if (ni_pi->enable_power_containment == false)
+		return 0;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	disable_uvd_power_tune = si_should_disable_uvd_powertune(rdev, radeon_state);
+
+	smc_state->levels[0].dpm2.MaxPS = 0;
+	smc_state->levels[0].dpm2.NearTDPDec = 0;
+	smc_state->levels[0].dpm2.AboveSafeInc = 0;
+	smc_state->levels[0].dpm2.BelowSafeInc = 0;
+	smc_state->levels[0].dpm2.PwrEfficiencyRatio = 0;
+
+	for (i = 1; i < state->performance_level_count; i++) {
+		prev_sclk = state->performance_levels[i-1].sclk;
+		max_sclk  = state->performance_levels[i].sclk;
+		if (i == 1)
+			max_ps_percent = SISLANDS_DPM2_MAXPS_PERCENT_M;
+		else
+			max_ps_percent = SISLANDS_DPM2_MAXPS_PERCENT_H;
+
+		if (prev_sclk > max_sclk)
+			return -EINVAL;
+
+		if ((max_ps_percent == 0) ||
+		    (prev_sclk == max_sclk) ||
+		    disable_uvd_power_tune) {
+			min_sclk = max_sclk;
+		} else if (i == 1) {
+			min_sclk = prev_sclk;
+		} else {
+			min_sclk = (prev_sclk * (u32)max_ps_percent) / 100;
+		}
+
+		if (min_sclk < state->performance_levels[0].sclk)
+			min_sclk = state->performance_levels[0].sclk;
+
+		if (min_sclk == 0)
+			return -EINVAL;
+
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[i-1].vddc, &vddc);
+		if (ret)
+			return ret;
+
+		ret = si_get_std_voltage_value(rdev, &vddc, &prev_std_vddc);
+		if (ret)
+			return ret;
+
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[i].vddc, &vddc);
+		if (ret)
+			return ret;
+
+		ret = si_get_std_voltage_value(rdev, &vddc, &curr_std_vddc);
+		if (ret)
+			return ret;
+
+		pwr_efficiency_ratio = si_calculate_power_efficiency_ratio(rdev,
+									   prev_std_vddc, curr_std_vddc);
+
+		smc_state->levels[i].dpm2.MaxPS = (u8)((SISLANDS_DPM2_MAX_PULSE_SKIP * (max_sclk - min_sclk)) / max_sclk);
+		smc_state->levels[i].dpm2.NearTDPDec = SISLANDS_DPM2_NEAR_TDP_DEC;
+		smc_state->levels[i].dpm2.AboveSafeInc = SISLANDS_DPM2_ABOVE_SAFE_INC;
+		smc_state->levels[i].dpm2.BelowSafeInc = SISLANDS_DPM2_BELOW_SAFE_INC;
+		smc_state->levels[i].dpm2.PwrEfficiencyRatio = cpu_to_be16(pwr_efficiency_ratio);
+	}
+
+	return 0;
+}
+
+static int si_populate_sq_ramping_values(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 sq_power_throttle, sq_power_throttle2;
+	bool enable_sq_ramping = ni_pi->enable_sq_ramping;
+	int i;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	if (rdev->pm.dpm.sq_ramping_threshold == 0)
+		return -EINVAL;
+
+	if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER > (MAX_POWER_MASK >> MAX_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (SISLANDS_DPM2_SQ_RAMP_MIN_POWER > (MIN_POWER_MASK >> MIN_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA > (MAX_POWER_DELTA_MASK >> MAX_POWER_DELTA_SHIFT))
+		enable_sq_ramping = false;
+
+	if (SISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO <= (LTI_RATIO_MASK >> LTI_RATIO_SHIFT))
+		enable_sq_ramping = false;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		sq_power_throttle = 0;
+		sq_power_throttle2 = 0;
+
+		if ((state->performance_levels[i].sclk >= rdev->pm.dpm.sq_ramping_threshold) &&
+		    enable_sq_ramping) {
+			sq_power_throttle |= MAX_POWER(SISLANDS_DPM2_SQ_RAMP_MAX_POWER);
+			sq_power_throttle |= MIN_POWER(SISLANDS_DPM2_SQ_RAMP_MIN_POWER);
+			sq_power_throttle2 |= MAX_POWER_DELTA(SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA);
+			sq_power_throttle2 |= STI_SIZE(SISLANDS_DPM2_SQ_RAMP_STI_SIZE);
+			sq_power_throttle2 |= LTI_RATIO(SISLANDS_DPM2_SQ_RAMP_LTI_RATIO);
+		} else {
+			sq_power_throttle |= MAX_POWER_MASK | MIN_POWER_MASK;
+			sq_power_throttle2 |= MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+		}
+
+		smc_state->levels[i].SQPowerThrottle = cpu_to_be32(sq_power_throttle);
+		smc_state->levels[i].SQPowerThrottle_2 = cpu_to_be32(sq_power_throttle2);
+	}
+
+	return 0;
+}
+
+static int si_enable_power_containment(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_new_state,
+				       bool enable)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	PPSMC_Result smc_result;
+	int ret = 0;
+
+	if (ni_pi->enable_power_containment) {
+		if (enable) {
+			if (!si_should_disable_uvd_powertune(rdev, radeon_new_state)) {
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_TDPClampingActive);
+				if (smc_result != PPSMC_Result_OK) {
+					ret = -EINVAL;
+					ni_pi->pc_enabled = false;
+				} else {
+					ni_pi->pc_enabled = true;
+				}
+			}
+		} else {
+			smc_result = si_send_msg_to_smc(rdev, PPSMC_TDPClampingInactive);
+			if (smc_result != PPSMC_Result_OK)
+				ret = -EINVAL;
+			ni_pi->pc_enabled = false;
+		}
+	}
+
+	return ret;
+}
+
+static int si_initialize_smc_dte_tables(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret = 0;
+	struct si_dte_data *dte_data = &si_pi->dte_data;
+	Smc_SIslands_DTE_Configuration *dte_tables = NULL;
+	u32 table_size;
+	u8 tdep_count;
+	u32 i;
+
+	if (dte_data == NULL)
+		si_pi->enable_dte = false;
+
+	if (si_pi->enable_dte == false)
+		return 0;
+
+	if (dte_data->k <= 0)
+		return -EINVAL;
+
+	dte_tables = kzalloc(sizeof(Smc_SIslands_DTE_Configuration), GFP_KERNEL);
+	if (dte_tables == NULL) {
+		si_pi->enable_dte = false;
+		return -ENOMEM;
+	}
+
+	table_size = dte_data->k;
+
+	if (table_size > SMC_SISLANDS_DTE_MAX_FILTER_STAGES)
+		table_size = SMC_SISLANDS_DTE_MAX_FILTER_STAGES;
+
+	tdep_count = dte_data->tdep_count;
+	if (tdep_count > SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE)
+		tdep_count = SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE;
+
+	dte_tables->K = cpu_to_be32(table_size);
+	dte_tables->T0 = cpu_to_be32(dte_data->t0);
+	dte_tables->MaxT = cpu_to_be32(dte_data->max_t);
+	dte_tables->WindowSize = dte_data->window_size;
+	dte_tables->temp_select = dte_data->temp_select;
+	dte_tables->DTE_mode = dte_data->dte_mode;
+	dte_tables->Tthreshold = cpu_to_be32(dte_data->t_threshold);
+
+	if (tdep_count > 0)
+		table_size--;
+
+	for (i = 0; i < table_size; i++) {
+		dte_tables->tau[i] = cpu_to_be32(dte_data->tau[i]);
+		dte_tables->R[i]   = cpu_to_be32(dte_data->r[i]);
+	}
+
+	dte_tables->Tdep_count = tdep_count;
+
+	for (i = 0; i < (u32)tdep_count; i++) {
+		dte_tables->T_limits[i] = dte_data->t_limits[i];
+		dte_tables->Tdep_tau[i] = cpu_to_be32(dte_data->tdep_tau[i]);
+		dte_tables->Tdep_R[i] = cpu_to_be32(dte_data->tdep_r[i]);
+	}
+
+	ret = si_copy_bytes_to_smc(rdev, si_pi->dte_table_start, (u8 *)dte_tables,
+				   sizeof(Smc_SIslands_DTE_Configuration), si_pi->sram_end);
+	kfree(dte_tables);
+
+	return ret;
+}
+
+static int si_get_cac_std_voltage_max_min(struct radeon_device *rdev,
+					  u16 *max, u16 *min)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct radeon_cac_leakage_table *table =
+		&rdev->pm.dpm.dyn_state.cac_leakage_table;
+	u32 i;
+	u32 v0_loadline;
+
+
+	if (table == NULL)
+		return -EINVAL;
+
+	*max = 0;
+	*min = 0xFFFF;
+
+	for (i = 0; i < table->count; i++) {
+		if (table->entries[i].vddc > *max)
+			*max = table->entries[i].vddc;
+		if (table->entries[i].vddc < *min)
+			*min = table->entries[i].vddc;
+	}
+
+	if (si_pi->powertune_data->lkge_lut_v0_percent > 100)
+		return -EINVAL;
+
+	v0_loadline = (*min) * (100 - si_pi->powertune_data->lkge_lut_v0_percent) / 100;
+
+	if (v0_loadline > 0xFFFFUL)
+		return -EINVAL;
+
+	*min = (u16)v0_loadline;
+
+	if ((*min > *max) || (*max == 0) || (*min == 0))
+		return -EINVAL;
+
+	return 0;
+}
+
+static u16 si_get_cac_std_voltage_step(u16 max, u16 min)
+{
+	return ((max - min) + (SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES - 1)) /
+		SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES;
+}
+
+static int si_init_dte_leakage_table(struct radeon_device *rdev,
+				     PP_SIslands_CacConfig *cac_tables,
+				     u16 vddc_max, u16 vddc_min, u16 vddc_step,
+				     u16 t0, u16 t_step)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 leakage;
+	unsigned int i, j;
+	s32 t;
+	u32 smc_leakage;
+	u32 scaling_factor;
+	u16 voltage;
+
+	scaling_factor = si_get_smc_power_scaling_factor(rdev);
+
+	for (i = 0; i < SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES ; i++) {
+		t = (1000 * (i * t_step + t0));
+
+		for (j = 0; j < SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+			voltage = vddc_max - (vddc_step * j);
+
+			si_calculate_leakage_for_v_and_t(rdev,
+							 &si_pi->powertune_data->leakage_coefficients,
+							 voltage,
+							 t,
+							 si_pi->dyn_powertune_data.cac_leakage,
+							 &leakage);
+
+			smc_leakage = si_scale_power_for_smc(leakage, scaling_factor) / 4;
+
+			if (smc_leakage > 0xFFFF)
+				smc_leakage = 0xFFFF;
+
+			cac_tables->cac_lkge_lut[i][SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES-1-j] =
+				cpu_to_be16((u16)smc_leakage);
+		}
+	}
+	return 0;
+}
+
+static int si_init_simplified_leakage_table(struct radeon_device *rdev,
+					    PP_SIslands_CacConfig *cac_tables,
+					    u16 vddc_max, u16 vddc_min, u16 vddc_step)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 leakage;
+	unsigned int i, j;
+	u32 smc_leakage;
+	u32 scaling_factor;
+	u16 voltage;
+
+	scaling_factor = si_get_smc_power_scaling_factor(rdev);
+
+	for (j = 0; j < SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+		voltage = vddc_max - (vddc_step * j);
+
+		si_calculate_leakage_for_v(rdev,
+					   &si_pi->powertune_data->leakage_coefficients,
+					   si_pi->powertune_data->fixed_kt,
+					   voltage,
+					   si_pi->dyn_powertune_data.cac_leakage,
+					   &leakage);
+
+		smc_leakage = si_scale_power_for_smc(leakage, scaling_factor) / 4;
+
+		if (smc_leakage > 0xFFFF)
+			smc_leakage = 0xFFFF;
+
+		for (i = 0; i < SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES ; i++)
+			cac_tables->cac_lkge_lut[i][SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES-1-j] =
+				cpu_to_be16((u16)smc_leakage);
+	}
+	return 0;
+}
+
+static int si_initialize_smc_cac_tables(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	PP_SIslands_CacConfig *cac_tables = NULL;
+	u16 vddc_max, vddc_min, vddc_step;
+	u16 t0, t_step;
+	u32 load_line_slope, reg;
+	int ret = 0;
+	u32 ticks_per_us = radeon_get_xclk(rdev) / 100;
+
+	if (ni_pi->enable_cac == false)
+		return 0;
+
+	cac_tables = kzalloc(sizeof(PP_SIslands_CacConfig), GFP_KERNEL);
+	if (!cac_tables)
+		return -ENOMEM;
+
+	reg = RREG32(CG_CAC_CTRL) & ~CAC_WINDOW_MASK;
+	reg |= CAC_WINDOW(si_pi->powertune_data->cac_window);
+	WREG32(CG_CAC_CTRL, reg);
+
+	si_pi->dyn_powertune_data.cac_leakage = rdev->pm.dpm.cac_leakage;
+	si_pi->dyn_powertune_data.dc_pwr_value =
+		si_pi->powertune_data->dc_cac[NISLANDS_DCCAC_LEVEL_0];
+	si_pi->dyn_powertune_data.wintime = si_calculate_cac_wintime(rdev);
+	si_pi->dyn_powertune_data.shift_n = si_pi->powertune_data->shift_n_default;
+
+	si_pi->dyn_powertune_data.leakage_minimum_temperature = 80 * 1000;
+
+	ret = si_get_cac_std_voltage_max_min(rdev, &vddc_max, &vddc_min);
+	if (ret)
+		goto done_free;
+
+	vddc_step = si_get_cac_std_voltage_step(vddc_max, vddc_min);
+	vddc_min = vddc_max - (vddc_step * (SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES - 1));
+	t_step = 4;
+	t0 = 60;
+
+	if (si_pi->enable_dte || ni_pi->driver_calculate_cac_leakage)
+		ret = si_init_dte_leakage_table(rdev, cac_tables,
+						vddc_max, vddc_min, vddc_step,
+						t0, t_step);
+	else
+		ret = si_init_simplified_leakage_table(rdev, cac_tables,
+						       vddc_max, vddc_min, vddc_step);
+	if (ret)
+		goto done_free;
+
+	load_line_slope = ((u32)rdev->pm.dpm.load_line_slope << SMC_SISLANDS_SCALE_R) / 100;
+
+	cac_tables->l2numWin_TDP = cpu_to_be32(si_pi->dyn_powertune_data.l2_lta_window_size);
+	cac_tables->lts_truncate_n = si_pi->dyn_powertune_data.lts_truncate;
+	cac_tables->SHIFT_N = si_pi->dyn_powertune_data.shift_n;
+	cac_tables->lkge_lut_V0 = cpu_to_be32((u32)vddc_min);
+	cac_tables->lkge_lut_Vstep = cpu_to_be32((u32)vddc_step);
+	cac_tables->R_LL = cpu_to_be32(load_line_slope);
+	cac_tables->WinTime = cpu_to_be32(si_pi->dyn_powertune_data.wintime);
+	cac_tables->calculation_repeats = cpu_to_be32(2);
+	cac_tables->dc_cac = cpu_to_be32(0);
+	cac_tables->log2_PG_LKG_SCALE = 12;
+	cac_tables->cac_temp = si_pi->powertune_data->operating_temp;
+	cac_tables->lkge_lut_T0 = cpu_to_be32((u32)t0);
+	cac_tables->lkge_lut_Tstep = cpu_to_be32((u32)t_step);
+
+	ret = si_copy_bytes_to_smc(rdev, si_pi->cac_table_start, (u8 *)cac_tables,
+				   sizeof(PP_SIslands_CacConfig), si_pi->sram_end);
+
+	if (ret)
+		goto done_free;
+
+	ret = si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_ticks_per_us, ticks_per_us);
+
+done_free:
+	if (ret) {
+		ni_pi->enable_cac = false;
+		ni_pi->enable_power_containment = false;
+	}
+
+	kfree(cac_tables);
+
+	return 0;
+}
+
+static int si_program_cac_config_registers(struct radeon_device *rdev,
+					   const struct si_cac_config_reg *cac_config_regs)
+{
+	const struct si_cac_config_reg *config_regs = cac_config_regs;
+	u32 data = 0, offset;
+
+	if (!config_regs)
+		return -EINVAL;
+
+	while (config_regs->offset != 0xFFFFFFFF) {
+		switch (config_regs->type) {
+		case SISLANDS_CACCONFIG_CGIND:
+			offset = SMC_CG_IND_START + config_regs->offset;
+			if (offset < SMC_CG_IND_END)
+				data = RREG32_SMC(offset);
+			break;
+		default:
+			data = RREG32(config_regs->offset << 2);
+			break;
+		}
+
+		data &= ~config_regs->mask;
+		data |= ((config_regs->value << config_regs->shift) & config_regs->mask);
+
+		switch (config_regs->type) {
+		case SISLANDS_CACCONFIG_CGIND:
+			offset = SMC_CG_IND_START + config_regs->offset;
+			if (offset < SMC_CG_IND_END)
+				WREG32_SMC(offset, data);
+			break;
+		default:
+			WREG32(config_regs->offset << 2, data);
+			break;
+		}
+		config_regs++;
+	}
+	return 0;
+}
+
+static int si_initialize_hardware_cac_manager(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+
+	if ((ni_pi->enable_cac == false) ||
+	    (ni_pi->cac_configuration_required == false))
+		return 0;
+
+	ret = si_program_cac_config_registers(rdev, si_pi->lcac_config);
+	if (ret)
+		return ret;
+	ret = si_program_cac_config_registers(rdev, si_pi->cac_override);
+	if (ret)
+		return ret;
+	ret = si_program_cac_config_registers(rdev, si_pi->cac_weights);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int si_enable_smc_cac(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_new_state,
+			     bool enable)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	PPSMC_Result smc_result;
+	int ret = 0;
+
+	if (ni_pi->enable_cac) {
+		if (enable) {
+			if (!si_should_disable_uvd_powertune(rdev, radeon_new_state)) {
+				if (ni_pi->support_cac_long_term_average) {
+					smc_result = si_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgEnable);
+					if (smc_result != PPSMC_Result_OK)
+						ni_pi->support_cac_long_term_average = false;
+				}
+
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableCac);
+				if (smc_result != PPSMC_Result_OK) {
+					ret = -EINVAL;
+					ni_pi->cac_enabled = false;
+				} else {
+					ni_pi->cac_enabled = true;
+				}
+
+				if (si_pi->enable_dte) {
+					smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableDTE);
+					if (smc_result != PPSMC_Result_OK)
+						ret = -EINVAL;
+				}
+			}
+		} else if (ni_pi->cac_enabled) {
+			if (si_pi->enable_dte)
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_DisableDTE);
+
+			smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_DisableCac);
+
+			ni_pi->cac_enabled = false;
+
+			if (ni_pi->support_cac_long_term_average)
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgDisable);
+		}
+	}
+	return ret;
+}
+
+static int si_init_smc_spll_table(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	SMC_SISLANDS_SPLL_DIV_TABLE *spll_table;
+	SISLANDS_SMC_SCLK_VALUE sclk_params;
+	u32 fb_div, p_div;
+	u32 clk_s, clk_v;
+	u32 sclk = 0;
+	int ret = 0;
+	u32 tmp;
+	int i;
+
+	if (si_pi->spll_table_start == 0)
+		return -EINVAL;
+
+	spll_table = kzalloc(sizeof(SMC_SISLANDS_SPLL_DIV_TABLE), GFP_KERNEL);
+	if (spll_table == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < 256; i++) {
+		ret = si_calculate_sclk_params(rdev, sclk, &sclk_params);
+		if (ret)
+			break;
+
+		p_div = (sclk_params.vCG_SPLL_FUNC_CNTL & SPLL_PDIV_A_MASK) >> SPLL_PDIV_A_SHIFT;
+		fb_div = (sclk_params.vCG_SPLL_FUNC_CNTL_3 & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT;
+		clk_s = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM & CLK_S_MASK) >> CLK_S_SHIFT;
+		clk_v = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM_2 & CLK_V_MASK) >> CLK_V_SHIFT;
+
+		fb_div &= ~0x00001FFF;
+		fb_div >>= 1;
+		clk_v >>= 6;
+
+		if (p_div & ~(SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT))
+			ret = -EINVAL;
+		if (fb_div & ~(SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT))
+			ret = -EINVAL;
+		if (clk_s & ~(SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT))
+			ret = -EINVAL;
+		if (clk_v & ~(SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT))
+			ret = -EINVAL;
+
+		if (ret)
+			break;
+
+		tmp = ((fb_div << SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_MASK) |
+			((p_div << SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_MASK);
+		spll_table->freq[i] = cpu_to_be32(tmp);
+
+		tmp = ((clk_v << SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_MASK) |
+			((clk_s << SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_MASK);
+		spll_table->ss[i] = cpu_to_be32(tmp);
+
+		sclk += 512;
+	}
+
+
+	if (!ret)
+		ret = si_copy_bytes_to_smc(rdev, si_pi->spll_table_start,
+					   (u8 *)spll_table, sizeof(SMC_SISLANDS_SPLL_DIV_TABLE),
+					   si_pi->sram_end);
+
+	if (ret)
+		ni_pi->enable_power_containment = false;
+
+	kfree(spll_table);
+
+	return ret;
+}
+
+static void si_apply_state_adjust_rules(struct radeon_device *rdev,
+					struct radeon_ps *rps)
+{
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct radeon_clock_and_voltage_limits *max_limits;
+	bool disable_mclk_switching;
+	u32 mclk, sclk;
+	u16 vddc, vddci;
+	int i;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
+	    ni_dpm_vblank_too_short(rdev))
+		disable_mclk_switching = true;
+	else
+		disable_mclk_switching = false;
+
+	if (rdev->pm.dpm.ac_power)
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
+	else
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc;
+
+	for (i = ps->performance_level_count - 2; i >= 0; i--) {
+		if (ps->performance_levels[i].vddc > ps->performance_levels[i+1].vddc)
+			ps->performance_levels[i].vddc = ps->performance_levels[i+1].vddc;
+	}
+	if (rdev->pm.dpm.ac_power == false) {
+		for (i = 0; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk > max_limits->mclk)
+				ps->performance_levels[i].mclk = max_limits->mclk;
+			if (ps->performance_levels[i].sclk > max_limits->sclk)
+				ps->performance_levels[i].sclk = max_limits->sclk;
+			if (ps->performance_levels[i].vddc > max_limits->vddc)
+				ps->performance_levels[i].vddc = max_limits->vddc;
+			if (ps->performance_levels[i].vddci > max_limits->vddci)
+				ps->performance_levels[i].vddci = max_limits->vddci;
+		}
+	}
+
+	/* XXX validate the min clocks required for display */
+
+	if (disable_mclk_switching) {
+		mclk  = ps->performance_levels[ps->performance_level_count - 1].mclk;
+		sclk = ps->performance_levels[0].sclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[ps->performance_level_count - 1].vddci;
+	} else {
+		sclk = ps->performance_levels[0].sclk;
+		mclk = ps->performance_levels[0].mclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[0].vddci;
+	}
+
+	/* adjusted low state */
+	ps->performance_levels[0].sclk = sclk;
+	ps->performance_levels[0].mclk = mclk;
+	ps->performance_levels[0].vddc = vddc;
+	ps->performance_levels[0].vddci = vddci;
+
+	for (i = 1; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk)
+			ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk;
+		if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc)
+			ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc;
+	}
+
+	if (disable_mclk_switching) {
+		mclk = ps->performance_levels[0].mclk;
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (mclk < ps->performance_levels[i].mclk)
+				mclk = ps->performance_levels[i].mclk;
+		}
+		for (i = 0; i < ps->performance_level_count; i++) {
+			ps->performance_levels[i].mclk = mclk;
+			ps->performance_levels[i].vddci = vddci;
+		}
+	} else {
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk < ps->performance_levels[i - 1].mclk)
+				ps->performance_levels[i].mclk = ps->performance_levels[i - 1].mclk;
+			if (ps->performance_levels[i].vddci < ps->performance_levels[i - 1].vddci)
+				ps->performance_levels[i].vddci = ps->performance_levels[i - 1].vddci;
+		}
+	}
+
+        for (i = 0; i < ps->performance_level_count; i++)
+                btc_adjust_clock_combinations(rdev, max_limits,
+                                              &ps->performance_levels[i]);
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+						   ps->performance_levels[i].sclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddci, &ps->performance_levels[i].vddci);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+						   rdev->clock.current_dispclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+	}
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_delta_rules(rdev,
+					      max_limits->vddc, max_limits->vddci,
+					      &ps->performance_levels[i].vddc,
+					      &ps->performance_levels[i].vddci);
+	}
+
+	ps->dc_compatible = true;
+	for (i = 0; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].vddc > rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc)
+			ps->dc_compatible = false;
+	}
+
+}
+
+#if 0
+static int si_read_smc_soft_register(struct radeon_device *rdev,
+				     u16 reg_offset, u32 *value)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	return si_read_smc_sram_dword(rdev,
+				      si_pi->soft_regs_start + reg_offset, value,
+				      si_pi->sram_end);
+}
+#endif
+
+static int si_write_smc_soft_register(struct radeon_device *rdev,
+				      u16 reg_offset, u32 value)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	return si_write_smc_sram_dword(rdev,
+				       si_pi->soft_regs_start + reg_offset,
+				       value, si_pi->sram_end);
+}
+
+static bool si_is_special_1gb_platform(struct radeon_device *rdev)
+{
+	bool ret = false;
+	u32 tmp, width, row, column, bank, density;
+	bool is_memory_gddr5, is_special;
+
+	tmp = RREG32(MC_SEQ_MISC0);
+	is_memory_gddr5 = (MC_SEQ_MISC0_GDDR5_VALUE == ((tmp & MC_SEQ_MISC0_GDDR5_MASK) >> MC_SEQ_MISC0_GDDR5_SHIFT));
+	is_special = (MC_SEQ_MISC0_REV_ID_VALUE == ((tmp & MC_SEQ_MISC0_REV_ID_MASK) >> MC_SEQ_MISC0_REV_ID_SHIFT))
+		& (MC_SEQ_MISC0_VEN_ID_VALUE == ((tmp & MC_SEQ_MISC0_VEN_ID_MASK) >> MC_SEQ_MISC0_VEN_ID_SHIFT));
+
+	WREG32(MC_SEQ_IO_DEBUG_INDEX, 0xb);
+	width = ((RREG32(MC_SEQ_IO_DEBUG_DATA) >> 1) & 1) ? 16 : 32;
+
+	tmp = RREG32(MC_ARB_RAMCFG);
+	row = ((tmp & NOOFROWS_MASK) >> NOOFROWS_SHIFT) + 10;
+	column = ((tmp & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT) + 8;
+	bank = ((tmp & NOOFBANK_MASK) >> NOOFBANK_SHIFT) + 2;
+
+	density = (1 << (row + column - 20 + bank)) * width;
+
+	if ((rdev->pdev->device == 0x6819) &&
+	    is_memory_gddr5 && is_special && (density == 0x400))
+		ret = true;
+
+	return ret;
+}
+
+static void si_get_leakage_vddc(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u16 vddc, count = 0;
+	int i, ret;
+
+	for (i = 0; i < SISLANDS_MAX_LEAKAGE_COUNT; i++) {
+		ret = radeon_atom_get_leakage_vddc_based_on_leakage_idx(rdev, &vddc, SISLANDS_LEAKAGE_INDEX0 + i);
+
+		if (!ret && (vddc > 0) && (vddc != (SISLANDS_LEAKAGE_INDEX0 + i))) {
+			si_pi->leakage_voltage.entries[count].voltage = vddc;
+			si_pi->leakage_voltage.entries[count].leakage_index =
+				SISLANDS_LEAKAGE_INDEX0 + i;
+			count++;
+		}
+	}
+	si_pi->leakage_voltage.count = count;
+}
+
+static int si_get_leakage_voltage_from_leakage_index(struct radeon_device *rdev,
+						     u32 index, u16 *leakage_voltage)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int i;
+
+	if (leakage_voltage == NULL)
+		return -EINVAL;
+
+	if ((index & 0xff00) != 0xff00)
+		return -EINVAL;
+
+	if ((index & 0xff) > SISLANDS_MAX_LEAKAGE_COUNT + 1)
+		return -EINVAL;
+
+	if (index < SISLANDS_LEAKAGE_INDEX0)
+		return -EINVAL;
+
+	for (i = 0; i < si_pi->leakage_voltage.count; i++) {
+		if (si_pi->leakage_voltage.entries[i].leakage_index == index) {
+			*leakage_voltage = si_pi->leakage_voltage.entries[i].voltage;
+			return 0;
+		}
+	}
+	return -EAGAIN;
+}
+
+static void si_set_dpm_event_sources(struct radeon_device *rdev, u32 sources)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	bool want_thermal_protection;
+	enum radeon_dpm_event_src dpm_event_src;
+
+	switch (sources) {
+	case 0:
+	default:
+		want_thermal_protection = false;
+                break;
+	case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL;
+		break;
+	case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL;
+		break;
+	case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) |
+	      (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL;
+		break;
+	}
+
+	if (want_thermal_protection) {
+		WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK);
+		if (pi->thermal_protection)
+			WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	} else {
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+	}
+}
+
+static void si_enable_auto_throttle_source(struct radeon_device *rdev,
+					   enum radeon_dpm_auto_throttle_src source,
+					   bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (!(pi->active_auto_throttle_sources & (1 << source))) {
+			pi->active_auto_throttle_sources |= 1 << source;
+			si_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	} else {
+		if (pi->active_auto_throttle_sources & (1 << source)) {
+			pi->active_auto_throttle_sources &= ~(1 << source);
+			si_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	}
+}
+
+static void si_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+static void si_stop_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+}
+
+static void si_enable_sclk_control(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+
+}
+
+#if 0
+static int si_notify_hardware_of_thermal_state(struct radeon_device *rdev,
+					       u32 thermal_level)
+{
+	PPSMC_Result ret;
+
+	if (thermal_level == 0) {
+		ret = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+		if (ret == PPSMC_Result_OK)
+			return 0;
+		else
+			return -EINVAL;
+	}
+	return 0;
+}
+
+static void si_notify_hardware_vpu_recovery_event(struct radeon_device *rdev)
+{
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_tdr_is_about_to_happen, true);
+}
+#endif
+
+#if 0
+static int si_notify_hw_of_powersource(struct radeon_device *rdev, bool ac_power)
+{
+	if (ac_power)
+		return (si_send_msg_to_smc(rdev, PPSMC_MSG_RunningOnAC) == PPSMC_Result_OK) ?
+			0 : -EINVAL;
+
+	return 0;
+}
+#endif
+
+static PPSMC_Result si_send_msg_to_smc_with_parameter(struct radeon_device *rdev,
+						      PPSMC_Msg msg, u32 parameter)
+{
+	WREG32(SMC_SCRATCH0, parameter);
+	return si_send_msg_to_smc(rdev, msg);
+}
+
+static int si_restrict_performance_levels_before_switch(struct radeon_device *rdev)
+{
+	if (si_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 1) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+int si_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	u32 levels;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 1) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		levels = ps->performance_level_count - 1;
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) {
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+static int si_set_boot_state(struct radeon_device *rdev)
+{
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToInitialState) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_set_sw_state(struct radeon_device *rdev)
+{
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToSwState) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_halt_smc(struct radeon_device *rdev)
+{
+	if (si_send_msg_to_smc(rdev, PPSMC_MSG_Halt) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (si_wait_for_smc_inactive(rdev) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_resume_smc(struct radeon_device *rdev)
+{
+	if (si_send_msg_to_smc(rdev, PPSMC_FlushDataCache) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_Resume) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static void si_dpm_start_smc(struct radeon_device *rdev)
+{
+	si_program_jump_on_start(rdev);
+	si_start_smc(rdev);
+	si_start_smc_clock(rdev);
+}
+
+static void si_dpm_stop_smc(struct radeon_device *rdev)
+{
+	si_reset_smc(rdev);
+	si_stop_smc_clock(rdev);
+}
+
+static int si_process_firmware_header(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_stateTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+        si_pi->state_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_softRegisters,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->soft_regs_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->mc_reg_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->arb_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_CacConfigTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->cac_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_DteConfiguration,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->dte_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_spllTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->spll_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_PAPMParameters,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->papm_cfg_table_start = tmp;
+
+	return ret;
+}
+
+static void si_read_clock_registers(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	si_pi->clock_registers.cg_spll_func_cntl = RREG32(CG_SPLL_FUNC_CNTL);
+	si_pi->clock_registers.cg_spll_func_cntl_2 = RREG32(CG_SPLL_FUNC_CNTL_2);
+	si_pi->clock_registers.cg_spll_func_cntl_3 = RREG32(CG_SPLL_FUNC_CNTL_3);
+	si_pi->clock_registers.cg_spll_func_cntl_4 = RREG32(CG_SPLL_FUNC_CNTL_4);
+	si_pi->clock_registers.cg_spll_spread_spectrum = RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	si_pi->clock_registers.cg_spll_spread_spectrum_2 = RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+	si_pi->clock_registers.dll_cntl = RREG32(DLL_CNTL);
+	si_pi->clock_registers.mclk_pwrmgt_cntl = RREG32(MCLK_PWRMGT_CNTL);
+	si_pi->clock_registers.mpll_ad_func_cntl = RREG32(MPLL_AD_FUNC_CNTL);
+	si_pi->clock_registers.mpll_dq_func_cntl = RREG32(MPLL_DQ_FUNC_CNTL);
+	si_pi->clock_registers.mpll_func_cntl = RREG32(MPLL_FUNC_CNTL);
+	si_pi->clock_registers.mpll_func_cntl_1 = RREG32(MPLL_FUNC_CNTL_1);
+	si_pi->clock_registers.mpll_func_cntl_2 = RREG32(MPLL_FUNC_CNTL_2);
+	si_pi->clock_registers.mpll_ss1 = RREG32(MPLL_SS1);
+	si_pi->clock_registers.mpll_ss2 = RREG32(MPLL_SS2);
+}
+
+static void si_enable_thermal_protection(struct radeon_device *rdev,
+					  bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	else
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+}
+
+static void si_enable_acpi_power_management(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+#if 0
+static int si_enter_ulp_state(struct radeon_device *rdev)
+{
+	WREG32(SMC_MESSAGE_0, PPSMC_MSG_SwitchToMinimumPower);
+
+	udelay(25000);
+
+	return 0;
+}
+
+static int si_exit_ulp_state(struct radeon_device *rdev)
+{
+	int i;
+
+	WREG32(SMC_MESSAGE_0, PPSMC_MSG_ResumeFromMinimumPower);
+
+	udelay(7000);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SMC_RESP_0) == 1)
+			break;
+		udelay(1000);
+	}
+
+	return 0;
+}
+#endif
+
+static int si_notify_smc_display_change(struct radeon_device *rdev,
+				     bool has_display)
+{
+	PPSMC_Msg msg = has_display ?
+		PPSMC_MSG_HasDisplay : PPSMC_MSG_NoDisplay;
+
+	return (si_send_msg_to_smc(rdev, msg) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static void si_program_response_times(struct radeon_device *rdev)
+{
+	u32 voltage_response_time, backbias_response_time, acpi_delay_time, vbi_time_out;
+	u32 vddc_dly, acpi_dly, vbi_dly;
+	u32 reference_clock;
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_mvdd_chg_time, 1);
+
+	voltage_response_time = (u32)rdev->pm.dpm.voltage_response_time;
+        backbias_response_time = (u32)rdev->pm.dpm.backbias_response_time;
+
+	if (voltage_response_time == 0)
+		voltage_response_time = 1000;
+
+	acpi_delay_time = 15000;
+	vbi_time_out = 100000;
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	vddc_dly = (voltage_response_time  * reference_clock) / 100;
+	acpi_dly = (acpi_delay_time * reference_clock) / 100;
+	vbi_dly  = (vbi_time_out * reference_clock) / 100;
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_delay_vreg,  vddc_dly);
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_delay_acpi,  acpi_dly);
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_mclk_chg_timeout, vbi_dly);
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_mc_block_delay, 0xAA);
+}
+
+static void si_program_ds_registers(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 tmp = 1; /* XXX: 0x10 on tahiti A0 */
+
+	if (eg_pi->sclk_deep_sleep) {
+		WREG32_P(MISC_CLK_CNTL, DEEP_SLEEP_CLK_SEL(tmp), ~DEEP_SLEEP_CLK_SEL_MASK);
+		WREG32_P(CG_SPLL_AUTOSCALE_CNTL, AUTOSCALE_ON_SS_CLEAR,
+			 ~AUTOSCALE_ON_SS_CLEAR);
+	}
+}
+
+static void si_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp, pipe;
+	int i;
+
+	tmp = RREG32(CG_DISPLAY_GAP_CNTL) & ~(DISP1_GAP_MASK | DISP2_GAP_MASK);
+	if (rdev->pm.dpm.new_active_crtc_count > 0)
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	if (rdev->pm.dpm.new_active_crtc_count > 1)
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+
+	tmp = RREG32(DCCG_DISP_SLOW_SELECT_REG);
+	pipe = (tmp & DCCG_DISP1_SLOW_SELECT_MASK) >> DCCG_DISP1_SLOW_SELECT_SHIFT;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 0) &&
+	    (!(rdev->pm.dpm.new_active_crtcs & (1 << pipe)))) {
+		/* find the first active crtc */
+		for (i = 0; i < rdev->num_crtc; i++) {
+			if (rdev->pm.dpm.new_active_crtcs & (1 << i))
+				break;
+		}
+		if (i == rdev->num_crtc)
+			pipe = 0;
+		else
+			pipe = i;
+
+		tmp &= ~DCCG_DISP1_SLOW_SELECT_MASK;
+		tmp |= DCCG_DISP1_SLOW_SELECT(pipe);
+		WREG32(DCCG_DISP_SLOW_SELECT_REG, tmp);
+	}
+
+	si_notify_smc_display_change(rdev, rdev->pm.dpm.new_active_crtc_count > 0);
+}
+
+static void si_enable_spread_spectrum(struct radeon_device *rdev, bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (pi->sclk_ss)
+			WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+	} else {
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+	}
+}
+
+static void si_setup_bsp(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(pi->asi,
+			       xclk,
+			       16,
+			       &pi->bsp,
+			       &pi->bsu);
+
+	r600_calculate_u_and_p(pi->pasi,
+			       xclk,
+			       16,
+			       &pi->pbsp,
+			       &pi->pbsu);
+
+
+        pi->dsp = BSP(pi->bsp) | BSU(pi->bsu);
+	pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu);
+
+	WREG32(CG_BSP, pi->dsp);
+}
+
+static void si_program_git(struct radeon_device *rdev)
+{
+	WREG32_P(CG_GIT, CG_GICST(R600_GICST_DFLT), ~CG_GICST_MASK);
+}
+
+static void si_program_tp(struct radeon_device *rdev)
+{
+	int i;
+	enum r600_td td = R600_TD_DFLT;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		WREG32(CG_FFCT_0 + (i * 4), (UTC_0(r600_utc[i]) | DTC_0(r600_dtc[i])));
+
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+static void si_program_tpp(struct radeon_device *rdev)
+{
+	WREG32(CG_TPC, R600_TPC_DFLT);
+}
+
+static void si_program_sstp(struct radeon_device *rdev)
+{
+	WREG32(CG_SSP, (SSTU(R600_SSTU_DFLT) | SST(R600_SST_DFLT)));
+}
+
+static void si_enable_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+		DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE));
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void si_program_vc(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	WREG32(CG_FTV, pi->vrc);
+}
+
+static void si_clear_vc(struct radeon_device *rdev)
+{
+	WREG32(CG_FTV, 0);
+}
+
+static u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock)
+{
+	u8 mc_para_index;
+
+	if (memory_clock < 10000)
+		mc_para_index = 0;
+	else if (memory_clock >= 80000)
+		mc_para_index = 0x0f;
+	else
+		mc_para_index = (u8)((memory_clock - 10000) / 5000 + 1);
+	return mc_para_index;
+}
+
+static u8 si_get_mclk_frequency_ratio(u32 memory_clock, bool strobe_mode)
+{
+	u8 mc_para_index;
+
+	if (strobe_mode) {
+		if (memory_clock < 12500)
+			mc_para_index = 0x00;
+		else if (memory_clock > 47500)
+			mc_para_index = 0x0f;
+		else
+			mc_para_index = (u8)((memory_clock - 10000) / 2500);
+	} else {
+		if (memory_clock < 65000)
+			mc_para_index = 0x00;
+		else if (memory_clock > 135000)
+			mc_para_index = 0x0f;
+		else
+			mc_para_index = (u8)((memory_clock - 60000) / 5000);
+	}
+	return mc_para_index;
+}
+
+static u8 si_get_strobe_mode_settings(struct radeon_device *rdev, u32 mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	bool strobe_mode = false;
+	u8 result = 0;
+
+	if (mclk <= pi->mclk_strobe_mode_threshold)
+		strobe_mode = true;
+
+	if (pi->mem_gddr5)
+		result = si_get_mclk_frequency_ratio(mclk, strobe_mode);
+	else
+		result = si_get_ddr3_mclk_frequency_ratio(mclk);
+
+	if (strobe_mode)
+		result |= SISLANDS_SMC_STROBE_ENABLE;
+
+	return result;
+}
+
+static int si_upload_firmware(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+
+	si_reset_smc(rdev);
+	si_stop_smc_clock(rdev);
+
+	ret = si_load_smc_ucode(rdev, si_pi->sram_end);
+
+	return ret;
+}
+
+static bool si_validate_phase_shedding_tables(struct radeon_device *rdev,
+					      const struct atom_voltage_table *table,
+					      const struct radeon_phase_shedding_limits_table *limits)
+{
+	u32 data, num_bits, num_levels;
+
+	if ((table == NULL) || (limits == NULL))
+		return false;
+
+	data = table->mask_low;
+
+	num_bits = hweight32(data);
+
+	if (num_bits == 0)
+		return false;
+
+	num_levels = (1 << num_bits);
+
+	if (table->count != num_levels)
+		return false;
+
+	if (limits->count != (num_levels - 1))
+		return false;
+
+	return true;
+}
+
+static void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev,
+						     struct atom_voltage_table *voltage_table)
+{
+	unsigned int i, diff;
+
+	if (voltage_table->count <= SISLANDS_MAX_NO_VREG_STEPS)
+		return;
+
+	diff = voltage_table->count - SISLANDS_MAX_NO_VREG_STEPS;
+
+	for (i= 0; i < SISLANDS_MAX_NO_VREG_STEPS; i++)
+		voltage_table->entries[i] = voltage_table->entries[i + diff];
+
+	voltage_table->count = SISLANDS_MAX_NO_VREG_STEPS;
+}
+
+static int si_construct_voltage_tables(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+
+	ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
+					    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddc_voltage_table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+		si_trim_voltage_table_to_fit_state_table(rdev, &eg_pi->vddc_voltage_table);
+
+	if (eg_pi->vddci_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDCI,
+						    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddci_voltage_table);
+		if (ret)
+			return ret;
+
+		if (eg_pi->vddci_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+			si_trim_voltage_table_to_fit_state_table(rdev, &eg_pi->vddci_voltage_table);
+	}
+
+	if (pi->mvdd_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_MVDDC,
+						    VOLTAGE_OBJ_GPIO_LUT, &si_pi->mvdd_voltage_table);
+
+		if (ret) {
+			pi->mvdd_control = false;
+			return ret;
+		}
+
+		if (si_pi->mvdd_voltage_table.count == 0) {
+			pi->mvdd_control = false;
+			return -EINVAL;
+		}
+
+		if (si_pi->mvdd_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+			si_trim_voltage_table_to_fit_state_table(rdev, &si_pi->mvdd_voltage_table);
+	}
+
+	if (si_pi->vddc_phase_shed_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
+						    VOLTAGE_OBJ_PHASE_LUT, &si_pi->vddc_phase_shed_table);
+		if (ret)
+			si_pi->vddc_phase_shed_control = false;
+
+		if ((si_pi->vddc_phase_shed_table.count == 0) ||
+		    (si_pi->vddc_phase_shed_table.count > SISLANDS_MAX_NO_VREG_STEPS))
+			si_pi->vddc_phase_shed_control = false;
+	}
+
+	return 0;
+}
+
+static void si_populate_smc_voltage_table(struct radeon_device *rdev,
+					  const struct atom_voltage_table *voltage_table,
+					  SISLANDS_SMC_STATETABLE *table)
+{
+	unsigned int i;
+
+	for (i = 0; i < voltage_table->count; i++)
+		table->lowSMIO[i] |= cpu_to_be32(voltage_table->entries[i].smio_low);
+}
+
+static int si_populate_smc_voltage_tables(struct radeon_device *rdev,
+					  SISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u8 i;
+
+	if (eg_pi->vddc_voltage_table.count) {
+		si_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
+		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+			if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
+				table->maxVDDCIndexInPPTable = i;
+				break;
+			}
+		}
+	}
+
+	if (eg_pi->vddci_voltage_table.count) {
+		si_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
+
+		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDCI] =
+			cpu_to_be32(eg_pi->vddci_voltage_table.mask_low);
+	}
+
+
+	if (si_pi->mvdd_voltage_table.count) {
+		si_populate_smc_voltage_table(rdev, &si_pi->mvdd_voltage_table, table);
+
+		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_MVDD] =
+			cpu_to_be32(si_pi->mvdd_voltage_table.mask_low);
+	}
+
+	if (si_pi->vddc_phase_shed_control) {
+		if (si_validate_phase_shedding_tables(rdev, &si_pi->vddc_phase_shed_table,
+						      &rdev->pm.dpm.dyn_state.phase_shedding_limits_table)) {
+			si_populate_smc_voltage_table(rdev, &si_pi->vddc_phase_shed_table, table);
+
+			table->phaseMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+				cpu_to_be32(si_pi->vddc_phase_shed_table.mask_low);
+
+			si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_phase_shedding_delay,
+						   (u32)si_pi->vddc_phase_shed_table.phase_delay);
+		} else {
+			si_pi->vddc_phase_shed_control = false;
+		}
+	}
+
+	return 0;
+}
+
+static int si_populate_voltage_value(struct radeon_device *rdev,
+				     const struct atom_voltage_table *table,
+				     u16 value, SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (value <= table->entries[i].value) {
+			voltage->index = (u8)i;
+			voltage->value = cpu_to_be16(table->entries[i].value);
+			break;
+		}
+	}
+
+	if (i >= table->count)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int si_populate_mvdd_value(struct radeon_device *rdev, u32 mclk,
+				  SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (pi->mvdd_control) {
+		if (mclk <= pi->mvdd_split_frequency)
+			voltage->index = 0;
+		else
+			voltage->index = (u8)(si_pi->mvdd_voltage_table.count) - 1;
+
+		voltage->value = cpu_to_be16(si_pi->mvdd_voltage_table.entries[voltage->index].value);
+	}
+	return 0;
+}
+
+static int si_get_std_voltage_value(struct radeon_device *rdev,
+				    SISLANDS_SMC_VOLTAGE_VALUE *voltage,
+				    u16 *std_voltage)
+{
+	u16 v_index;
+	bool voltage_found = false;
+	*std_voltage = be16_to_cpu(voltage->value);
+
+	if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_NEW_CAC_VOLTAGE) {
+			if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries == NULL)
+				return -EINVAL;
+
+			for (v_index = 0; (u32)v_index < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) {
+				if (be16_to_cpu(voltage->value) ==
+				    (u16)rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) {
+					voltage_found = true;
+					if ((u32)v_index < rdev->pm.dpm.dyn_state.cac_leakage_table.count)
+						*std_voltage =
+							rdev->pm.dpm.dyn_state.cac_leakage_table.entries[v_index].vddc;
+					else
+						*std_voltage =
+							rdev->pm.dpm.dyn_state.cac_leakage_table.entries[rdev->pm.dpm.dyn_state.cac_leakage_table.count-1].vddc;
+					break;
+				}
+			}
+
+			if (!voltage_found) {
+				for (v_index = 0; (u32)v_index < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) {
+					if (be16_to_cpu(voltage->value) <=
+					    (u16)rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) {
+						voltage_found = true;
+						if ((u32)v_index < rdev->pm.dpm.dyn_state.cac_leakage_table.count)
+							*std_voltage =
+								rdev->pm.dpm.dyn_state.cac_leakage_table.entries[v_index].vddc;
+						else
+							*std_voltage =
+								rdev->pm.dpm.dyn_state.cac_leakage_table.entries[rdev->pm.dpm.dyn_state.cac_leakage_table.count-1].vddc;
+						break;
+					}
+				}
+			}
+		} else {
+			if ((u32)voltage->index < rdev->pm.dpm.dyn_state.cac_leakage_table.count)
+				*std_voltage = rdev->pm.dpm.dyn_state.cac_leakage_table.entries[voltage->index].vddc;
+		}
+	}
+
+	return 0;
+}
+
+static int si_populate_std_voltage_value(struct radeon_device *rdev,
+					 u16 value, u8 index,
+					 SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	voltage->index = index;
+	voltage->value = cpu_to_be16(value);
+
+	return 0;
+}
+
+static int si_populate_phase_shedding_value(struct radeon_device *rdev,
+					    const struct radeon_phase_shedding_limits_table *limits,
+					    u16 voltage, u32 sclk, u32 mclk,
+					    SISLANDS_SMC_VOLTAGE_VALUE *smc_voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < limits->count; i++) {
+		if ((voltage <= limits->entries[i].voltage) &&
+		    (sclk <= limits->entries[i].sclk) &&
+		    (mclk <= limits->entries[i].mclk))
+			break;
+	}
+
+	smc_voltage->phase_settings = (u8)i;
+
+	return 0;
+}
+
+static int si_init_arb_table_index(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = si_read_smc_sram_dword(rdev, si_pi->arb_table_start, &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp &= 0x00FFFFFF;
+	tmp |= MC_CG_ARB_FREQ_F1 << 24;
+
+	return si_write_smc_sram_dword(rdev, si_pi->arb_table_start,  tmp, si_pi->sram_end);
+}
+
+static int si_initial_switch_from_arb_f0_to_f1(struct radeon_device *rdev)
+{
+	return ni_copy_and_switch_arb_sets(rdev, MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1);
+}
+
+static int si_reset_to_default(struct radeon_device *rdev)
+{
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_ResetToDefaults) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_force_switch_to_arb_f0(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = si_read_smc_sram_dword(rdev, si_pi->arb_table_start,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp = (tmp >> 24) & 0xff;
+
+	if (tmp == MC_CG_ARB_FREQ_F0)
+		return 0;
+
+	return ni_copy_and_switch_arb_sets(rdev, tmp, MC_CG_ARB_FREQ_F0);
+}
+
+static u32 si_calculate_memory_refresh_rate(struct radeon_device *rdev,
+					    u32 engine_clock)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 dram_rows;
+	u32 dram_refresh_rate;
+	u32 mc_arb_rfsh_rate;
+	u32 tmp = (RREG32(MC_ARB_RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT;
+
+	if (pi->mem_gddr5)
+		dram_rows = 1 << (tmp + 10);
+	else
+		dram_rows = DDR3_DRAM_ROWS;
+
+	dram_refresh_rate = 1 << ((RREG32(MC_SEQ_MISC0) & 0x3) + 3);
+	mc_arb_rfsh_rate = ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64;
+
+	return mc_arb_rfsh_rate;
+}
+
+static int si_populate_memory_timing_parameters(struct radeon_device *rdev,
+						struct rv7xx_pl *pl,
+						SMC_SIslands_MCArbDramTimingRegisterSet *arb_regs)
+{
+	u32 dram_timing;
+	u32 dram_timing2;
+	u32 burst_time;
+
+	arb_regs->mc_arb_rfsh_rate =
+		(u8)si_calculate_memory_refresh_rate(rdev, pl->sclk);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    pl->sclk,
+                                            pl->mclk);
+
+	dram_timing  = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+	burst_time = RREG32(MC_ARB_BURST_TIME) & STATE0_MASK;
+
+	arb_regs->mc_arb_dram_timing  = cpu_to_be32(dram_timing);
+	arb_regs->mc_arb_dram_timing2 = cpu_to_be32(dram_timing2);
+	arb_regs->mc_arb_burst_time = (u8)burst_time;
+
+	return 0;
+}
+
+static int si_do_program_memory_timing_parameters(struct radeon_device *rdev,
+						  struct radeon_ps *radeon_state,
+						  unsigned int first_arb_set)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	SMC_SIslands_MCArbDramTimingRegisterSet arb_regs = { 0 };
+	int i, ret = 0;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ret = si_populate_memory_timing_parameters(rdev, &state->performance_levels[i], &arb_regs);
+		if (ret)
+			break;
+		ret = si_copy_bytes_to_smc(rdev,
+					   si_pi->arb_table_start +
+					   offsetof(SMC_SIslands_MCArbDramTimingRegisters, data) +
+					   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet) * (first_arb_set + i),
+					   (u8 *)&arb_regs,
+					   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet),
+					   si_pi->sram_end);
+		if (ret)
+			break;
+        }
+
+	return ret;
+}
+
+static int si_program_memory_timing_parameters(struct radeon_device *rdev,
+					       struct radeon_ps *radeon_new_state)
+{
+	return si_do_program_memory_timing_parameters(rdev, radeon_new_state,
+						      SISLANDS_DRIVER_STATE_ARB_INDEX);
+}
+
+static int si_populate_initial_mvdd_value(struct radeon_device *rdev,
+					  struct SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (pi->mvdd_control)
+		return si_populate_voltage_value(rdev, &si_pi->mvdd_voltage_table,
+						 si_pi->mvdd_bootup_value, voltage);
+
+	return 0;
+}
+
+static int si_populate_smc_initial_state(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_initial_state,
+					 SISLANDS_SMC_STATETABLE *table)
+{
+	struct ni_ps *initial_state = ni_get_ps(radeon_initial_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 reg;
+	int ret;
+
+	table->initialState.levels[0].mclk.vDLL_CNTL =
+		cpu_to_be32(si_pi->clock_registers.dll_cntl);
+	table->initialState.levels[0].mclk.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mpll_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_FUNC_CNTL_1 =
+		cpu_to_be32(si_pi->clock_registers.mpll_func_cntl_1);
+	table->initialState.levels[0].mclk.vMPLL_FUNC_CNTL_2 =
+		cpu_to_be32(si_pi->clock_registers.mpll_func_cntl_2);
+	table->initialState.levels[0].mclk.vMPLL_SS =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss1);
+	table->initialState.levels[0].mclk.vMPLL_SS2 =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl_4);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2  =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].sclk);
+
+	table->initialState.levels[0].arbRefreshState =
+		SISLANDS_INITIAL_STATE_ARB_INDEX;
+
+	table->initialState.levels[0].ACIndex = 0;
+
+	ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+					initial_state->performance_levels[0].vddc,
+					&table->initialState.levels[0].vddc);
+
+	if (!ret) {
+		u16 std_vddc;
+
+		ret = si_get_std_voltage_value(rdev,
+					       &table->initialState.levels[0].vddc,
+					       &std_vddc);
+		if (!ret)
+			si_populate_std_voltage_value(rdev, std_vddc,
+						      table->initialState.levels[0].vddc.index,
+						      &table->initialState.levels[0].std_vddc);
+	}
+
+	if (eg_pi->vddci_control)
+		si_populate_voltage_value(rdev,
+					  &eg_pi->vddci_voltage_table,
+					  initial_state->performance_levels[0].vddci,
+					  &table->initialState.levels[0].vddci);
+
+	if (si_pi->vddc_phase_shed_control)
+		si_populate_phase_shedding_value(rdev,
+						 &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+						 initial_state->performance_levels[0].vddc,
+						 initial_state->performance_levels[0].sclk,
+						 initial_state->performance_levels[0].mclk,
+						 &table->initialState.levels[0].vddc);
+
+	si_populate_initial_mvdd_value(rdev, &table->initialState.levels[0].mvdd);
+
+	reg = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(reg);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	table->initialState.levels[0].gen2PCIE = (u8)si_pi->boot_pcie_gen;
+
+	if (pi->mem_gddr5) {
+		table->initialState.levels[0].strobeMode =
+			si_get_strobe_mode_settings(rdev,
+						    initial_state->performance_levels[0].mclk);
+
+		if (initial_state->performance_levels[0].mclk > pi->mclk_edc_enable_threshold)
+			table->initialState.levels[0].mcFlags = SISLANDS_SMC_MC_EDC_RD_FLAG | SISLANDS_SMC_MC_EDC_WR_FLAG;
+		else
+			table->initialState.levels[0].mcFlags =  0;
+	}
+
+	table->initialState.levelCount = 1;
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	table->initialState.levels[0].dpm2.MaxPS = 0;
+	table->initialState.levels[0].dpm2.NearTDPDec = 0;
+	table->initialState.levels[0].dpm2.AboveSafeInc = 0;
+	table->initialState.levels[0].dpm2.BelowSafeInc = 0;
+	table->initialState.levels[0].dpm2.PwrEfficiencyRatio = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->initialState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->initialState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int si_populate_smc_acpi_state(struct radeon_device *rdev,
+				      SISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 spll_func_cntl = si_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = si_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = si_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4 = si_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 dll_cntl = si_pi->clock_registers.dll_cntl;
+	u32 mclk_pwrmgt_cntl = si_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32 mpll_ad_func_cntl = si_pi->clock_registers.mpll_ad_func_cntl;
+	u32 mpll_dq_func_cntl = si_pi->clock_registers.mpll_dq_func_cntl;
+	u32 mpll_func_cntl = si_pi->clock_registers.mpll_func_cntl;
+	u32 mpll_func_cntl_1 = si_pi->clock_registers.mpll_func_cntl_1;
+	u32 mpll_func_cntl_2 = si_pi->clock_registers.mpll_func_cntl_2;
+	u32 reg;
+	int ret;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						pi->acpi_vddc, &table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = si_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc, &std_vddc);
+			if (!ret)
+				si_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+		table->ACPIState.levels[0].gen2PCIE = si_pi->acpi_pcie_gen;
+
+		if (si_pi->vddc_phase_shed_control) {
+			si_populate_phase_shedding_value(rdev,
+							 &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+							 pi->acpi_vddc,
+							 0,
+							 0,
+							 &table->ACPIState.levels[0].vddc);
+		}
+	} else {
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						pi->min_vddc_in_table, &table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = si_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc, &std_vddc);
+
+			if (!ret)
+				si_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+		table->ACPIState.levels[0].gen2PCIE = (u8)r600_get_pcie_gen_support(rdev,
+										    si_pi->sys_pcie_mask,
+										    si_pi->boot_pcie_gen,
+										    RADEON_PCIE_GEN1);
+
+		if (si_pi->vddc_phase_shed_control)
+			si_populate_phase_shedding_value(rdev,
+							 &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+							 pi->min_vddc_in_table,
+							 0,
+							 0,
+							 &table->ACPIState.levels[0].vddc);
+	}
+
+	if (pi->acpi_vddc) {
+		if (eg_pi->acpi_vddci)
+			si_populate_voltage_value(rdev, &eg_pi->vddci_voltage_table,
+						  eg_pi->acpi_vddci,
+						  &table->ACPIState.levels[0].vddci);
+	}
+
+	mclk_pwrmgt_cntl |= MRDCK0_RESET | MRDCK1_RESET;
+	mclk_pwrmgt_cntl &= ~(MRDCK0_PDNB | MRDCK1_PDNB);
+
+	dll_cntl &= ~(MRDCK0_BYPASS | MRDCK1_BYPASS);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.vDLL_CNTL =
+		cpu_to_be32(dll_cntl);
+	table->ACPIState.levels[0].mclk.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_FUNC_CNTL =
+		cpu_to_be32(mpll_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_FUNC_CNTL_1 =
+		cpu_to_be32(mpll_func_cntl_1);
+	table->ACPIState.levels[0].mclk.vMPLL_FUNC_CNTL_2 =
+		cpu_to_be32(mpll_func_cntl_2);
+	table->ACPIState.levels[0].mclk.vMPLL_SS =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss1);
+	table->ACPIState.levels[0].mclk.vMPLL_SS2 =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss2);
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(spll_func_cntl_3);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 =
+		cpu_to_be32(spll_func_cntl_4);
+
+	table->ACPIState.levels[0].mclk.mclk_value = 0;
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	si_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	if (eg_pi->dynamic_ac_timing)
+		table->ACPIState.levels[0].ACIndex = 0;
+
+	table->ACPIState.levels[0].dpm2.MaxPS = 0;
+	table->ACPIState.levels[0].dpm2.NearTDPDec = 0;
+	table->ACPIState.levels[0].dpm2.AboveSafeInc = 0;
+	table->ACPIState.levels[0].dpm2.BelowSafeInc = 0;
+	table->ACPIState.levels[0].dpm2.PwrEfficiencyRatio = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int si_populate_ulv_state(struct radeon_device *rdev,
+				 SISLANDS_SMC_SWSTATE *state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	u32 sclk_in_sr = 1350; /* ??? */
+	int ret;
+
+	ret = si_convert_power_level_to_smc(rdev, &ulv->pl,
+					    &state->levels[0]);
+	if (!ret) {
+		if (eg_pi->sclk_deep_sleep) {
+			if (sclk_in_sr <= SCLK_MIN_DEEPSLEEP_FREQ)
+				state->levels[0].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_BYPASS;
+			else
+				state->levels[0].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE;
+		}
+		if (ulv->one_pcie_lane_in_ulv)
+			state->flags |= PPSMC_SWSTATE_FLAG_PCIE_X1;
+		state->levels[0].arbRefreshState = (u8)(SISLANDS_ULV_STATE_ARB_INDEX);
+		state->levels[0].ACIndex = 1;
+		state->levels[0].std_vddc = state->levels[0].vddc;
+		state->levelCount = 1;
+
+		state->flags |= PPSMC_SWSTATE_FLAG_DC;
+	}
+
+	return ret;
+}
+
+static int si_program_ulv_memory_timing_parameters(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	SMC_SIslands_MCArbDramTimingRegisterSet arb_regs = { 0 };
+	int ret;
+
+	ret = si_populate_memory_timing_parameters(rdev, &ulv->pl,
+						   &arb_regs);
+	if (ret)
+		return ret;
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_ulv_volt_change_delay,
+				   ulv->volt_change_delay);
+
+	ret = si_copy_bytes_to_smc(rdev,
+				   si_pi->arb_table_start +
+				   offsetof(SMC_SIslands_MCArbDramTimingRegisters, data) +
+				   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet) * SISLANDS_ULV_STATE_ARB_INDEX,
+				   (u8 *)&arb_regs,
+				   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet),
+				   si_pi->sram_end);
+
+	return ret;
+}
+
+static void si_get_mvdd_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->mvdd_split_frequency = 30000;
+}
+
+static int si_init_smc_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct radeon_ps *radeon_boot_state = rdev->pm.dpm.boot_ps;
+	const struct si_ulv_param *ulv = &si_pi->ulv;
+	SISLANDS_SMC_STATETABLE  *table = &si_pi->smc_statetable;
+	int ret;
+	u32 lane_width;
+	u32 vr_hot_gpio;
+
+	si_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+	case THERMAL_TYPE_SI:
+	case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+	case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+	default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT) {
+		if ((rdev->pdev->device != 0x6818) && (rdev->pdev->device != 0x6819))
+			table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY)
+		table->systemFlags |= PPSMC_EXTRAFLAGS_AC2DC_GPIO5_POLARITY_HIGH;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE) {
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT_PROG_GPIO;
+		vr_hot_gpio = rdev->pm.dpm.backbias_response_time;
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_vr_hot_gpio,
+					   vr_hot_gpio);
+	}
+
+	ret = si_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	ret = si_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	ret = si_do_program_memory_timing_parameters(rdev, radeon_boot_state,
+						     SISLANDS_INITIAL_STATE_ARB_INDEX);
+	if (ret)
+		return ret;
+
+	if (ulv->supported && ulv->pl.vddc) {
+		ret = si_populate_ulv_state(rdev, &table->ULVState);
+		if (ret)
+			return ret;
+
+		ret = si_program_ulv_memory_timing_parameters(rdev);
+		if (ret)
+			return ret;
+
+		WREG32(CG_ULV_CONTROL, ulv->cg_ulv_control);
+		WREG32(CG_ULV_PARAMETER, ulv->cg_ulv_parameter);
+
+		lane_width = radeon_get_pcie_lanes(rdev);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width, lane_width);
+	} else {
+		table->ULVState = table->initialState;
+	}
+
+	return si_copy_bytes_to_smc(rdev, si_pi->state_table_start,
+				    (u8 *)table, sizeof(SISLANDS_SMC_STATETABLE),
+				    si_pi->sram_end);
+}
+
+static int si_calculate_sclk_params(struct radeon_device *rdev,
+				    u32 engine_clock,
+				    SISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = si_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = si_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = si_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4 = si_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum = si_pi->clock_registers.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = si_pi->clock_registers.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	tmp = (u64) engine_clock * reference_divider * dividers.post_div * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_PDIV_A(dividers.post_div);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+        spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+        spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+        spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = engine_clock;
+	sclk->vCG_SPLL_FUNC_CNTL = spll_func_cntl;
+	sclk->vCG_SPLL_FUNC_CNTL_2 = spll_func_cntl_2;
+	sclk->vCG_SPLL_FUNC_CNTL_3 = spll_func_cntl_3;
+	sclk->vCG_SPLL_FUNC_CNTL_4 = spll_func_cntl_4;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cg_spll_spread_spectrum;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cg_spll_spread_spectrum_2;
+
+	return 0;
+}
+
+static int si_populate_sclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  SISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	SISLANDS_SMC_SCLK_VALUE sclk_tmp;
+	int ret;
+
+	ret = si_calculate_sclk_params(rdev, engine_clock, &sclk_tmp);
+	if (!ret) {
+		sclk->sclk_value = cpu_to_be32(sclk_tmp.sclk_value);
+		sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL);
+		sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_2);
+		sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_3);
+		sclk->vCG_SPLL_FUNC_CNTL_4 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_4);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM_2);
+	}
+
+	return ret;
+}
+
+static int si_populate_mclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  u32 memory_clock,
+				  SISLANDS_SMC_MCLK_VALUE *mclk,
+				  bool strobe_mode,
+				  bool dll_state_on)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32  dll_cntl = si_pi->clock_registers.dll_cntl;
+	u32  mclk_pwrmgt_cntl = si_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32  mpll_ad_func_cntl = si_pi->clock_registers.mpll_ad_func_cntl;
+	u32  mpll_dq_func_cntl = si_pi->clock_registers.mpll_dq_func_cntl;
+	u32  mpll_func_cntl = si_pi->clock_registers.mpll_func_cntl;
+	u32  mpll_func_cntl_1 = si_pi->clock_registers.mpll_func_cntl_1;
+	u32  mpll_func_cntl_2 = si_pi->clock_registers.mpll_func_cntl_2;
+	u32  mpll_ss1 = si_pi->clock_registers.mpll_ss1;
+	u32  mpll_ss2 = si_pi->clock_registers.mpll_ss2;
+	struct atom_mpll_param mpll_param;
+	int ret;
+
+	ret = radeon_atom_get_memory_pll_dividers(rdev, memory_clock, strobe_mode, &mpll_param);
+	if (ret)
+		return ret;
+
+	mpll_func_cntl &= ~BWCTRL_MASK;
+	mpll_func_cntl |= BWCTRL(mpll_param.bwcntl);
+
+	mpll_func_cntl_1 &= ~(CLKF_MASK | CLKFRAC_MASK | VCO_MODE_MASK);
+	mpll_func_cntl_1 |= CLKF(mpll_param.clkf) |
+		CLKFRAC(mpll_param.clkfrac) | VCO_MODE(mpll_param.vco_mode);
+
+	mpll_ad_func_cntl &= ~YCLK_POST_DIV_MASK;
+	mpll_ad_func_cntl |= YCLK_POST_DIV(mpll_param.post_div);
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(YCLK_SEL_MASK | YCLK_POST_DIV_MASK);
+		mpll_dq_func_cntl |= YCLK_SEL(mpll_param.yclk_sel) |
+			YCLK_POST_DIV(mpll_param.post_div);
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 freq_nom;
+		u32 tmp;
+		u32 reference_clock = rdev->clock.mpll.reference_freq;
+
+		if (pi->mem_gddr5)
+			freq_nom = memory_clock * 4;
+		else
+			freq_nom = memory_clock * 2;
+
+		tmp = freq_nom / reference_clock;
+		tmp = tmp * tmp;
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+                                                     ASIC_INTERNAL_MEMORY_SS, freq_nom)) {
+			u32 clks = reference_clock * 5 / ss.rate;
+			u32 clkv = (u32)((((131 * ss.percentage * ss.rate) / 100) * tmp) / freq_nom);
+
+                        mpll_ss1 &= ~CLKV_MASK;
+                        mpll_ss1 |= CLKV(clkv);
+
+                        mpll_ss2 &= ~CLKS_MASK;
+                        mpll_ss2 |= CLKS(clks);
+		}
+	}
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(mpll_param.dll_speed);
+
+	if (dll_state_on)
+		mclk_pwrmgt_cntl |= MRDCK0_PDNB | MRDCK1_PDNB;
+	else
+		mclk_pwrmgt_cntl &= ~(MRDCK0_PDNB | MRDCK1_PDNB);
+
+	mclk->mclk_value = cpu_to_be32(memory_clock);
+	mclk->vMPLL_FUNC_CNTL = cpu_to_be32(mpll_func_cntl);
+	mclk->vMPLL_FUNC_CNTL_1 = cpu_to_be32(mpll_func_cntl_1);
+	mclk->vMPLL_FUNC_CNTL_2 = cpu_to_be32(mpll_func_cntl_2);
+	mclk->vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+static void si_populate_smc_sp(struct radeon_device *rdev,
+			       struct radeon_ps *radeon_state,
+			       SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_ps *ps = ni_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < ps->performance_level_count - 1; i++)
+		smc_state->levels[i].bSP = cpu_to_be32(pi->dsp);
+
+	smc_state->levels[ps->performance_level_count - 1].bSP =
+		cpu_to_be32(pi->psp);
+}
+
+static int si_convert_power_level_to_smc(struct radeon_device *rdev,
+					 struct rv7xx_pl *pl,
+					 SISLANDS_SMC_HW_PERFORMANCE_LEVEL *level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+	bool dll_state_on;
+	u16 std_vddc;
+	bool gmc_pg = false;
+
+	if (eg_pi->pcie_performance_request &&
+	    (si_pi->force_pcie_gen != RADEON_PCIE_GEN_INVALID))
+		level->gen2PCIE = (u8)si_pi->force_pcie_gen;
+	else
+		level->gen2PCIE = (u8)pl->pcie_gen;
+
+	ret = si_populate_sclk_value(rdev, pl->sclk, &level->sclk);
+	if (ret)
+		return ret;
+
+	level->mcFlags =  0;
+
+	if (pi->mclk_stutter_mode_threshold &&
+	    (pl->mclk <= pi->mclk_stutter_mode_threshold) &&
+	    !eg_pi->uvd_enabled &&
+	    (RREG32(DPG_PIPE_STUTTER_CONTROL) & STUTTER_ENABLE) &&
+	    (rdev->pm.dpm.new_active_crtc_count <= 2)) {
+		level->mcFlags |= SISLANDS_SMC_MC_STUTTER_EN;
+
+		if (gmc_pg)
+			level->mcFlags |= SISLANDS_SMC_MC_PG_EN;
+	}
+
+	if (pi->mem_gddr5) {
+		if (pl->mclk > pi->mclk_edc_enable_threshold)
+			level->mcFlags |= SISLANDS_SMC_MC_EDC_RD_FLAG;
+
+		if (pl->mclk > eg_pi->mclk_edc_wr_enable_threshold)
+			level->mcFlags |= SISLANDS_SMC_MC_EDC_WR_FLAG;
+
+		level->strobeMode = si_get_strobe_mode_settings(rdev, pl->mclk);
+
+		if (level->strobeMode & SISLANDS_SMC_STROBE_ENABLE) {
+			if (si_get_mclk_frequency_ratio(pl->mclk, true) >=
+			    ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf))
+				dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+			else
+				dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false;
+		} else {
+			dll_state_on = false;
+		}
+	} else {
+		level->strobeMode = si_get_strobe_mode_settings(rdev,
+								pl->mclk);
+
+		dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+	}
+
+	ret = si_populate_mclk_value(rdev,
+				     pl->sclk,
+				     pl->mclk,
+				     &level->mclk,
+				     (level->strobeMode & SISLANDS_SMC_STROBE_ENABLE) != 0, dll_state_on);
+	if (ret)
+		return ret;
+
+	ret = si_populate_voltage_value(rdev,
+					&eg_pi->vddc_voltage_table,
+					pl->vddc, &level->vddc);
+	if (ret)
+		return ret;
+
+
+	ret = si_get_std_voltage_value(rdev, &level->vddc, &std_vddc);
+	if (ret)
+		return ret;
+
+	ret = si_populate_std_voltage_value(rdev, std_vddc,
+					    level->vddc.index, &level->std_vddc);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddci_control) {
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddci_voltage_table,
+						pl->vddci, &level->vddci);
+		if (ret)
+			return ret;
+	}
+
+	if (si_pi->vddc_phase_shed_control) {
+		ret = si_populate_phase_shedding_value(rdev,
+						       &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+						       pl->vddc,
+						       pl->sclk,
+						       pl->mclk,
+						       &level->vddc);
+		if (ret)
+			return ret;
+	}
+
+	level->MaxPoweredUpCU = si_pi->max_cu;
+
+	ret = si_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int si_populate_smc_t(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_state,
+			     SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 a_t;
+	u32 t_l, t_h;
+	u32 high_bsp;
+	int i, ret;
+
+	if (state->performance_level_count >= 9)
+		return -EINVAL;
+
+	if (state->performance_level_count < 2) {
+		a_t = CG_R(0xffff) | CG_L(0);
+		smc_state->levels[0].aT = cpu_to_be32(a_t);
+		return 0;
+	}
+
+	smc_state->levels[0].aT = cpu_to_be32(0);
+
+	for (i = 0; i <= state->performance_level_count - 2; i++) {
+		ret = r600_calculate_at(
+			(50 / SISLANDS_MAX_HARDWARE_POWERLEVELS) * 100 * (i + 1),
+			100 * R600_AH_DFLT,
+			state->performance_levels[i + 1].sclk,
+			state->performance_levels[i].sclk,
+			&t_l,
+			&t_h);
+
+		if (ret) {
+			t_h = (i + 1) * 1000 - 50 * R600_AH_DFLT;
+			t_l = (i + 1) * 1000 + 50 * R600_AH_DFLT;
+		}
+
+		a_t = be32_to_cpu(smc_state->levels[i].aT) & ~CG_R_MASK;
+		a_t |= CG_R(t_l * pi->bsp / 20000);
+		smc_state->levels[i].aT = cpu_to_be32(a_t);
+
+		high_bsp = (i == state->performance_level_count - 2) ?
+			pi->pbsp : pi->bsp;
+		a_t = CG_R(0xffff) | CG_L(t_h * high_bsp / 20000);
+		smc_state->levels[i + 1].aT = cpu_to_be32(a_t);
+	}
+
+	return 0;
+}
+
+static int si_disable_ulv(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+
+	if (ulv->supported)
+		return (si_send_msg_to_smc(rdev, PPSMC_MSG_DisableULV) == PPSMC_Result_OK) ?
+			0 : -EINVAL;
+
+	return 0;
+}
+
+static bool si_is_state_ulv_compatible(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_state)
+{
+	const struct si_power_info *si_pi = si_get_pi(rdev);
+	const struct si_ulv_param *ulv = &si_pi->ulv;
+	const struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+
+	if (state->performance_levels[0].mclk != ulv->pl.mclk)
+		return false;
+
+	/* XXX validate against display requirements! */
+
+	for (i = 0; i < rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count; i++) {
+		if (rdev->clock.current_dispclk <=
+		    rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[i].clk) {
+			if (ulv->pl.vddc <
+			    rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[i].v)
+				return false;
+		}
+	}
+
+	if ((radeon_state->vclk != 0) || (radeon_state->dclk != 0))
+		return false;
+
+	return true;
+}
+
+static int si_set_power_state_conditionally_enable_ulv(struct radeon_device *rdev,
+						       struct radeon_ps *radeon_new_state)
+{
+	const struct si_power_info *si_pi = si_get_pi(rdev);
+	const struct si_ulv_param *ulv = &si_pi->ulv;
+
+	if (ulv->supported) {
+		if (si_is_state_ulv_compatible(rdev, radeon_new_state))
+			return (si_send_msg_to_smc(rdev, PPSMC_MSG_EnableULV) == PPSMC_Result_OK) ?
+				0 : -EINVAL;
+	}
+	return 0;
+}
+
+static int si_convert_power_state_to_smc(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i, ret;
+	u32 threshold;
+	u32 sclk_in_sr = 1350; /* ??? */
+
+	if (state->performance_level_count > SISLANDS_MAX_HARDWARE_POWERLEVELS)
+		return -EINVAL;
+
+	threshold = state->performance_levels[state->performance_level_count-1].sclk * 100 / 100;
+
+	if (radeon_state->vclk && radeon_state->dclk) {
+		eg_pi->uvd_enabled = true;
+		if (eg_pi->smu_uvd_hs)
+			smc_state->flags |= PPSMC_SWSTATE_FLAG_UVD;
+	} else {
+		eg_pi->uvd_enabled = false;
+	}
+
+	if (state->dc_compatible)
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	smc_state->levelCount = 0;
+	for (i = 0; i < state->performance_level_count; i++) {
+		if (eg_pi->sclk_deep_sleep) {
+			if ((i == 0) || si_pi->sclk_deep_sleep_above_low) {
+				if (sclk_in_sr <= SCLK_MIN_DEEPSLEEP_FREQ)
+					smc_state->levels[i].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_BYPASS;
+				else
+					smc_state->levels[i].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE;
+			}
+		}
+
+		ret = si_convert_power_level_to_smc(rdev, &state->performance_levels[i],
+						    &smc_state->levels[i]);
+		smc_state->levels[i].arbRefreshState =
+			(u8)(SISLANDS_DRIVER_STATE_ARB_INDEX + i);
+
+		if (ret)
+			return ret;
+
+		if (ni_pi->enable_power_containment)
+			smc_state->levels[i].displayWatermark =
+				(state->performance_levels[i].sclk < threshold) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+		else
+			smc_state->levels[i].displayWatermark = (i < 2) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+
+		if (eg_pi->dynamic_ac_timing)
+			smc_state->levels[i].ACIndex = SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i;
+		else
+			smc_state->levels[i].ACIndex = 0;
+
+		smc_state->levelCount++;
+	}
+
+	si_write_smc_soft_register(rdev,
+				   SI_SMC_SOFT_REGISTER_watermark_threshold,
+				   threshold / 512);
+
+	si_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	ret = si_populate_power_containment_values(rdev, radeon_state, smc_state);
+	if (ret)
+		ni_pi->enable_power_containment = false;
+
+	ret = si_populate_sq_ramping_values(rdev, radeon_state, smc_state);
+        if (ret)
+		ni_pi->enable_sq_ramping = false;
+
+	return si_populate_smc_t(rdev, radeon_state, smc_state);
+}
+
+static int si_upload_sw_state(struct radeon_device *rdev,
+			      struct radeon_ps *radeon_new_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *new_state = ni_get_ps(radeon_new_state);
+	int ret;
+	u32 address = si_pi->state_table_start +
+		offsetof(SISLANDS_SMC_STATETABLE, driverState);
+	u32 state_size = sizeof(SISLANDS_SMC_SWSTATE) +
+		((new_state->performance_level_count - 1) *
+		 sizeof(SISLANDS_SMC_HW_PERFORMANCE_LEVEL));
+	SISLANDS_SMC_SWSTATE *smc_state = &si_pi->smc_statetable.driverState;
+
+	memset(smc_state, 0, state_size);
+
+	ret = si_convert_power_state_to_smc(rdev, radeon_new_state, smc_state);
+	if (ret)
+		return ret;
+
+	ret = si_copy_bytes_to_smc(rdev, address, (u8 *)smc_state,
+				   state_size, si_pi->sram_end);
+
+	return ret;
+}
+
+static int si_upload_ulv_state(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	int ret = 0;
+
+	if (ulv->supported && ulv->pl.vddc) {
+		u32 address = si_pi->state_table_start +
+			offsetof(SISLANDS_SMC_STATETABLE, ULVState);
+		SISLANDS_SMC_SWSTATE *smc_state = &si_pi->smc_statetable.ULVState;
+		u32 state_size = sizeof(SISLANDS_SMC_SWSTATE);
+
+		memset(smc_state, 0, state_size);
+
+		ret = si_populate_ulv_state(rdev, smc_state);
+		if (!ret)
+			ret = si_copy_bytes_to_smc(rdev, address, (u8 *)smc_state,
+						   state_size, si_pi->sram_end);
+	}
+
+	return ret;
+}
+
+static int si_upload_smc_data(struct radeon_device *rdev)
+{
+	struct radeon_crtc *radeon_crtc = NULL;
+	int i;
+
+	if (rdev->pm.dpm.new_active_crtc_count == 0)
+		return 0;
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (rdev->pm.dpm.new_active_crtcs & (1 << i)) {
+			radeon_crtc = rdev->mode_info.crtcs[i];
+			break;
+		}
+	}
+
+	if (radeon_crtc == NULL)
+		return 0;
+
+	if (radeon_crtc->line_time <= 0)
+		return 0;
+
+	if (si_write_smc_soft_register(rdev,
+				       SI_SMC_SOFT_REGISTER_crtc_index,
+				       radeon_crtc->crtc_id) != PPSMC_Result_OK)
+		return 0;
+
+	if (si_write_smc_soft_register(rdev,
+				       SI_SMC_SOFT_REGISTER_mclk_change_block_cp_min,
+				       radeon_crtc->wm_high / radeon_crtc->line_time) != PPSMC_Result_OK)
+		return 0;
+
+	if (si_write_smc_soft_register(rdev,
+				       SI_SMC_SOFT_REGISTER_mclk_change_block_cp_max,
+				       radeon_crtc->wm_low / radeon_crtc->line_time) != PPSMC_Result_OK)
+		return 0;
+
+	return 0;
+}
+
+static int si_set_mc_special_registers(struct radeon_device *rdev,
+				       struct si_mc_reg_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 i, j, k;
+	u32 temp_reg;
+
+	for (i = 0, j = table->last; i < table->last; i++) {
+		if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+			return -EINVAL;
+		switch (table->mc_reg_address[i].s1 << 2) {
+		case MC_SEQ_MISC1:
+			temp_reg = RREG32(MC_PMG_CMD_EMRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					((temp_reg & 0xffff0000)) |
+					((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16);
+			j++;
+			if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			temp_reg = RREG32(MC_PMG_CMD_MRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+				if (!pi->mem_gddr5)
+					table->mc_reg_table_entry[k].mc_data[j] |= 0x100;
+			}
+			j++;
+			if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			if (!pi->mem_gddr5) {
+				table->mc_reg_address[j].s1 = MC_PMG_AUTO_CMD >> 2;
+				table->mc_reg_address[j].s0 = MC_PMG_AUTO_CMD >> 2;
+				for (k = 0; k < table->num_entries; k++)
+					table->mc_reg_table_entry[k].mc_data[j] =
+						(table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16;
+				j++;
+				if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+					return -EINVAL;
+			}
+			break;
+		case MC_SEQ_RESERVE_M:
+			temp_reg = RREG32(MC_PMG_CMD_MRS1);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+			for(k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+			j++;
+			if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		default:
+			break;
+		}
+	}
+
+	table->last = j;
+
+	return 0;
+}
+
+static bool si_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg)
+{
+	bool result = true;
+
+	switch (in_reg) {
+	case  MC_SEQ_RAS_TIMING >> 2:
+		*out_reg = MC_SEQ_RAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_CAS_TIMING >> 2:
+		*out_reg = MC_SEQ_CAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING2 >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING2_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D1_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D1_LP >> 2;
+		break;
+        case MC_PMG_CMD_EMRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS1 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+		break;
+        case MC_SEQ_PMG_TIMING >> 2:
+		*out_reg = MC_SEQ_PMG_TIMING_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS2 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS2_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_2 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_2_LP >> 2;
+		break;
+        default:
+		result = false;
+		break;
+	}
+
+	return result;
+}
+
+static void si_set_valid_flag(struct si_mc_reg_table *table)
+{
+	u8 i, j;
+
+	for (i = 0; i < table->last; i++) {
+		for (j = 1; j < table->num_entries; j++) {
+			if (table->mc_reg_table_entry[j-1].mc_data[i] != table->mc_reg_table_entry[j].mc_data[i]) {
+				table->valid_flag |= 1 << i;
+				break;
+			}
+		}
+	}
+}
+
+static void si_set_s0_mc_reg_index(struct si_mc_reg_table *table)
+{
+	u32 i;
+	u16 address;
+
+	for (i = 0; i < table->last; i++)
+		table->mc_reg_address[i].s0 = si_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ?
+			address : table->mc_reg_address[i].s1;
+
+}
+
+static int si_copy_vbios_mc_reg_table(struct atom_mc_reg_table *table,
+				      struct si_mc_reg_table *si_table)
+{
+	u8 i, j;
+
+	if (table->last > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+		return -EINVAL;
+	if (table->num_entries > MAX_AC_TIMING_ENTRIES)
+		return -EINVAL;
+
+	for (i = 0; i < table->last; i++)
+		si_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1;
+	si_table->last = table->last;
+
+	for (i = 0; i < table->num_entries; i++) {
+		si_table->mc_reg_table_entry[i].mclk_max =
+			table->mc_reg_table_entry[i].mclk_max;
+		for (j = 0; j < table->last; j++) {
+			si_table->mc_reg_table_entry[i].mc_data[j] =
+				table->mc_reg_table_entry[i].mc_data[j];
+		}
+	}
+	si_table->num_entries = table->num_entries;
+
+	return 0;
+}
+
+static int si_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct atom_mc_reg_table *table;
+	struct si_mc_reg_table *si_table = &si_pi->mc_reg_table;
+	u8 module_index = rv770_get_memory_module_index(rdev);
+	int ret;
+
+	table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL);
+	if (!table)
+		return -ENOMEM;
+
+	WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING));
+	WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2));
+	WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1));
+	WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0));
+	WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1));
+	WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0));
+	WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1));
+	WREG32(MC_SEQ_PMG_TIMING_LP, RREG32(MC_SEQ_PMG_TIMING));
+	WREG32(MC_SEQ_PMG_CMD_MRS2_LP, RREG32(MC_PMG_CMD_MRS2));
+	WREG32(MC_SEQ_WR_CTL_2_LP, RREG32(MC_SEQ_WR_CTL_2));
+
+        ret = radeon_atom_init_mc_reg_table(rdev, module_index, table);
+        if (ret)
+                goto init_mc_done;
+
+        ret = si_copy_vbios_mc_reg_table(table, si_table);
+        if (ret)
+                goto init_mc_done;
+
+	si_set_s0_mc_reg_index(si_table);
+
+	ret = si_set_mc_special_registers(rdev, si_table);
+        if (ret)
+                goto init_mc_done;
+
+	si_set_valid_flag(si_table);
+
+init_mc_done:
+	kfree(table);
+
+	return ret;
+
+}
+
+static void si_populate_mc_reg_addresses(struct radeon_device *rdev,
+					 SMC_SIslands_MCRegisters *mc_reg_table)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0, j = 0; j < si_pi->mc_reg_table.last; j++) {
+		if (si_pi->mc_reg_table.valid_flag & (1 << j)) {
+			if (i >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				break;
+			mc_reg_table->address[i].s0 =
+				cpu_to_be16(si_pi->mc_reg_table.mc_reg_address[j].s0);
+			mc_reg_table->address[i].s1 =
+				cpu_to_be16(si_pi->mc_reg_table.mc_reg_address[j].s1);
+			i++;
+		}
+	}
+	mc_reg_table->last = (u8)i;
+}
+
+static void si_convert_mc_registers(const struct si_mc_reg_entry *entry,
+				    SMC_SIslands_MCRegisterSet *data,
+				    u32 num_entries, u32 valid_flag)
+{
+	u32 i, j;
+
+	for(i = 0, j = 0; j < num_entries; j++) {
+		if (valid_flag & (1 << j)) {
+			data->value[i] = cpu_to_be32(entry->mc_data[j]);
+			i++;
+		}
+	}
+}
+
+static void si_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev,
+						 struct rv7xx_pl *pl,
+						 SMC_SIslands_MCRegisterSet *mc_reg_table_data)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < si_pi->mc_reg_table.num_entries; i++) {
+		if (pl->mclk <= si_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max)
+			break;
+	}
+
+	if ((i == si_pi->mc_reg_table.num_entries) && (i > 0))
+		--i;
+
+	si_convert_mc_registers(&si_pi->mc_reg_table.mc_reg_table_entry[i],
+				mc_reg_table_data, si_pi->mc_reg_table.last,
+				si_pi->mc_reg_table.valid_flag);
+}
+
+static void si_convert_mc_reg_table_to_smc(struct radeon_device *rdev,
+					   struct radeon_ps *radeon_state,
+					   SMC_SIslands_MCRegisters *mc_reg_table)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		si_convert_mc_reg_table_entry_to_smc(rdev,
+						     &state->performance_levels[i],
+						     &mc_reg_table->data[SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i]);
+	}
+}
+
+static int si_populate_mc_reg_table(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct ni_ps *boot_state = ni_get_ps(radeon_boot_state);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	SMC_SIslands_MCRegisters *smc_mc_reg_table = &si_pi->smc_mc_reg_table;
+
+	memset(smc_mc_reg_table, 0, sizeof(SMC_SIslands_MCRegisters));
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_seq_index, 1);
+
+	si_populate_mc_reg_addresses(rdev, smc_mc_reg_table);
+
+	si_convert_mc_reg_table_entry_to_smc(rdev, &boot_state->performance_levels[0],
+					     &smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_INITIAL_SLOT]);
+
+	si_convert_mc_registers(&si_pi->mc_reg_table.mc_reg_table_entry[0],
+				&smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_ACPI_SLOT],
+				si_pi->mc_reg_table.last,
+				si_pi->mc_reg_table.valid_flag);
+
+	if (ulv->supported && ulv->pl.vddc != 0)
+		si_convert_mc_reg_table_entry_to_smc(rdev, &ulv->pl,
+						     &smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_ULV_SLOT]);
+	else
+		si_convert_mc_registers(&si_pi->mc_reg_table.mc_reg_table_entry[0],
+					&smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_ULV_SLOT],
+					si_pi->mc_reg_table.last,
+					si_pi->mc_reg_table.valid_flag);
+
+	si_convert_mc_reg_table_to_smc(rdev, radeon_boot_state, smc_mc_reg_table);
+
+	return si_copy_bytes_to_smc(rdev, si_pi->mc_reg_table_start,
+				    (u8 *)smc_mc_reg_table,
+				    sizeof(SMC_SIslands_MCRegisters), si_pi->sram_end);
+}
+
+static int si_upload_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_new_state)
+{
+	struct ni_ps *new_state = ni_get_ps(radeon_new_state);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 address = si_pi->mc_reg_table_start +
+		offsetof(SMC_SIslands_MCRegisters,
+			 data[SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT]);
+	SMC_SIslands_MCRegisters *smc_mc_reg_table = &si_pi->smc_mc_reg_table;
+
+	memset(smc_mc_reg_table, 0, sizeof(SMC_SIslands_MCRegisters));
+
+	si_convert_mc_reg_table_to_smc(rdev, radeon_new_state, smc_mc_reg_table);
+
+
+	return si_copy_bytes_to_smc(rdev, address,
+				    (u8 *)&smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT],
+				    sizeof(SMC_SIslands_MCRegisterSet) * new_state->performance_level_count,
+				    si_pi->sram_end);
+
+}
+
+static void si_enable_voltage_control(struct radeon_device *rdev, bool enable)
+{
+        if (enable)
+                WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN);
+        else
+                WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN);
+}
+
+static enum radeon_pcie_gen si_get_maximum_link_speed(struct radeon_device *rdev,
+						      struct radeon_ps *radeon_state)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+	u16 pcie_speed, max_speed = 0;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		pcie_speed = state->performance_levels[i].pcie_gen;
+		if (max_speed < pcie_speed)
+			max_speed = pcie_speed;
+	}
+	return max_speed;
+}
+
+static u16 si_get_current_pcie_speed(struct radeon_device *rdev)
+{
+	u32 speed_cntl;
+
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & LC_CURRENT_DATA_RATE_MASK;
+	speed_cntl >>= LC_CURRENT_DATA_RATE_SHIFT;
+
+	return (u16)speed_cntl;
+}
+
+static void si_request_link_speed_change_before_state_change(struct radeon_device *rdev,
+							     struct radeon_ps *radeon_new_state,
+							     struct radeon_ps *radeon_current_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	enum radeon_pcie_gen target_link_speed = si_get_maximum_link_speed(rdev, radeon_new_state);
+	enum radeon_pcie_gen current_link_speed;
+
+	if (si_pi->force_pcie_gen == RADEON_PCIE_GEN_INVALID)
+		current_link_speed = si_get_maximum_link_speed(rdev, radeon_current_state);
+	else
+		current_link_speed = si_pi->force_pcie_gen;
+
+	si_pi->force_pcie_gen = RADEON_PCIE_GEN_INVALID;
+	si_pi->pspp_notify_required = false;
+	if (target_link_speed > current_link_speed) {
+		switch (target_link_speed) {
+#if defined(CONFIG_ACPI)
+		case RADEON_PCIE_GEN3:
+			if (radeon_acpi_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN3, false) == 0)
+				break;
+			si_pi->force_pcie_gen = RADEON_PCIE_GEN2;
+			if (current_link_speed == RADEON_PCIE_GEN2)
+				break;
+		case RADEON_PCIE_GEN2:
+			if (radeon_acpi_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, false) == 0)
+				break;
+#endif
+		default:
+			si_pi->force_pcie_gen = si_get_current_pcie_speed(rdev);
+			break;
+		}
+	} else {
+		if (target_link_speed < current_link_speed)
+			si_pi->pspp_notify_required = true;
+	}
+}
+
+static void si_notify_link_speed_change_after_state_change(struct radeon_device *rdev,
+							   struct radeon_ps *radeon_new_state,
+							   struct radeon_ps *radeon_current_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	enum radeon_pcie_gen target_link_speed = si_get_maximum_link_speed(rdev, radeon_new_state);
+	u8 request;
+
+	if (si_pi->pspp_notify_required) {
+		if (target_link_speed == RADEON_PCIE_GEN3)
+			request = PCIE_PERF_REQ_PECI_GEN3;
+		else if (target_link_speed == RADEON_PCIE_GEN2)
+			request = PCIE_PERF_REQ_PECI_GEN2;
+		else
+			request = PCIE_PERF_REQ_PECI_GEN1;
+
+		if ((request == PCIE_PERF_REQ_PECI_GEN1) &&
+		    (si_get_current_pcie_speed(rdev) > 0))
+			return;
+
+#if defined(CONFIG_ACPI)
+		radeon_acpi_pcie_performance_request(rdev, request, false);
+#endif
+	}
+}
+
+#if 0
+static int si_ds_request(struct radeon_device *rdev,
+			 bool ds_status_on, u32 count_write)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->sclk_deep_sleep) {
+		if (ds_status_on)
+			return (si_send_msg_to_smc(rdev, PPSMC_MSG_CancelThrottleOVRDSCLKDS) ==
+				PPSMC_Result_OK) ?
+				0 : -EINVAL;
+		else
+			return (si_send_msg_to_smc(rdev, PPSMC_MSG_ThrottleOVRDSCLKDS) ==
+				PPSMC_Result_OK) ? 0 : -EINVAL;
+	}
+	return 0;
+}
+#endif
+
+static void si_set_max_cu_value(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (rdev->family == CHIP_VERDE) {
+		switch (rdev->pdev->device) {
+		case 0x6820:
+		case 0x6825:
+		case 0x6821:
+		case 0x6823:
+		case 0x6827:
+			si_pi->max_cu = 10;
+			break;
+		case 0x682D:
+		case 0x6824:
+		case 0x682F:
+		case 0x6826:
+			si_pi->max_cu = 8;
+			break;
+		case 0x6828:
+		case 0x6830:
+		case 0x6831:
+		case 0x6838:
+		case 0x6839:
+		case 0x683D:
+			si_pi->max_cu = 10;
+			break;
+		case 0x683B:
+		case 0x683F:
+		case 0x6829:
+			si_pi->max_cu = 8;
+			break;
+		default:
+			si_pi->max_cu = 0;
+			break;
+		}
+	} else {
+		si_pi->max_cu = 0;
+	}
+}
+
+static int si_patch_single_dependency_table_based_on_leakage(struct radeon_device *rdev,
+							     struct radeon_clock_voltage_dependency_table *table)
+{
+	u32 i;
+	int j;
+	u16 leakage_voltage;
+
+	if (table) {
+		for (i = 0; i < table->count; i++) {
+			switch (si_get_leakage_voltage_from_leakage_index(rdev,
+									  table->entries[i].v,
+									  &leakage_voltage)) {
+			case 0:
+				table->entries[i].v = leakage_voltage;
+				break;
+			case -EAGAIN:
+				return -EINVAL;
+			case -EINVAL:
+			default:
+				break;
+			}
+		}
+
+		for (j = (table->count - 2); j >= 0; j--) {
+			table->entries[j].v = (table->entries[j].v <= table->entries[j + 1].v) ?
+				table->entries[j].v : table->entries[j + 1].v;
+		}
+	}
+	return 0;
+}
+
+static int si_patch_dependency_tables_based_on_leakage(struct radeon_device *rdev)
+{
+	int ret = 0;
+
+	ret = si_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk);
+	ret = si_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk);
+	ret = si_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk);
+	return ret;
+}
+
+static void si_set_pcie_lane_width_in_smc(struct radeon_device *rdev,
+					  struct radeon_ps *radeon_new_state,
+					  struct radeon_ps *radeon_current_state)
+{
+	u32 lane_width;
+	u32 new_lane_width =
+		(radeon_new_state->caps & ATOM_PPLIB_PCIE_LINK_WIDTH_MASK) >> ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT;
+	u32 current_lane_width =
+		(radeon_current_state->caps & ATOM_PPLIB_PCIE_LINK_WIDTH_MASK) >> ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT;
+
+	if (new_lane_width != current_lane_width) {
+		radeon_set_pcie_lanes(rdev, new_lane_width);
+		lane_width = radeon_get_pcie_lanes(rdev);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width, lane_width);
+	}
+}
+
+void si_dpm_setup_asic(struct radeon_device *rdev)
+{
+	rv770_get_memory_type(rdev);
+	si_read_clock_registers(rdev);
+	si_enable_acpi_power_management(rdev);
+}
+
+static int si_set_thermal_temperature_range(struct radeon_device *rdev,
+					int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK);
+	WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+int si_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (si_is_smc_running(rdev))
+		return -EINVAL;
+	if (pi->voltage_control)
+		si_enable_voltage_control(rdev, true);
+	if (pi->mvdd_control)
+		si_get_mvdd_configuration(rdev);
+	if (pi->voltage_control) {
+		ret = si_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("si_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = si_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+	}
+	if (pi->dynamic_ss)
+		si_enable_spread_spectrum(rdev, true);
+	if (pi->thermal_protection)
+		si_enable_thermal_protection(rdev, true);
+	si_setup_bsp(rdev);
+	si_program_git(rdev);
+	si_program_tp(rdev);
+	si_program_tpp(rdev);
+	si_program_sstp(rdev);
+	si_enable_display_gap(rdev);
+	si_program_vc(rdev);
+	ret = si_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("si_upload_firmware failed\n");
+		return ret;
+	}
+	ret = si_process_firmware_header(rdev);
+	if (ret) {
+		DRM_ERROR("si_process_firmware_header failed\n");
+		return ret;
+	}
+	ret = si_initial_switch_from_arb_f0_to_f1(rdev);
+	if (ret) {
+		DRM_ERROR("si_initial_switch_from_arb_f0_to_f1 failed\n");
+		return ret;
+	}
+	ret = si_init_smc_table(rdev);
+	if (ret) {
+		DRM_ERROR("si_init_smc_table failed\n");
+		return ret;
+	}
+	ret = si_init_smc_spll_table(rdev);
+	if (ret) {
+		DRM_ERROR("si_init_smc_spll_table failed\n");
+		return ret;
+	}
+	ret = si_init_arb_table_index(rdev);
+	if (ret) {
+		DRM_ERROR("si_init_arb_table_index failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = si_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("si_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = si_initialize_smc_cac_tables(rdev);
+	if (ret) {
+		DRM_ERROR("si_initialize_smc_cac_tables failed\n");
+		return ret;
+	}
+	ret = si_initialize_hardware_cac_manager(rdev);
+	if (ret) {
+		DRM_ERROR("si_initialize_hardware_cac_manager failed\n");
+		return ret;
+	}
+	ret = si_initialize_smc_dte_tables(rdev);
+	if (ret) {
+		DRM_ERROR("si_initialize_smc_dte_tables failed\n");
+		return ret;
+	}
+	ret = si_populate_smc_tdp_limits(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("si_populate_smc_tdp_limits failed\n");
+		return ret;
+	}
+	ret = si_populate_smc_tdp_limits_2(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("si_populate_smc_tdp_limits_2 failed\n");
+		return ret;
+	}
+	si_program_response_times(rdev);
+	si_program_ds_registers(rdev);
+	si_dpm_start_smc(rdev);
+	ret = si_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("si_notify_smc_display_change failed\n");
+		return ret;
+	}
+	si_enable_sclk_control(rdev, true);
+	si_start_dpm(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = si_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	ni_update_current_ps(rdev, boot_ps);
+
+	return 0;
+}
+
+void si_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!si_is_smc_running(rdev))
+		return;
+	si_disable_ulv(rdev);
+	si_clear_vc(rdev);
+	if (pi->thermal_protection)
+		si_enable_thermal_protection(rdev, false);
+	si_enable_power_containment(rdev, boot_ps, false);
+	si_enable_smc_cac(rdev, boot_ps, false);
+	si_enable_spread_spectrum(rdev, false);
+	si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, false);
+	si_stop_dpm(rdev);
+	si_reset_to_default(rdev);
+	si_dpm_stop_smc(rdev);
+	si_force_switch_to_arb_f0(rdev);
+
+	ni_update_current_ps(rdev, boot_ps);
+}
+
+int si_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	ni_update_requested_ps(rdev, new_ps);
+
+	si_apply_state_adjust_rules(rdev, &eg_pi->requested_rps);
+
+	return 0;
+}
+
+static int si_power_control_set_level(struct radeon_device *rdev)
+{
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	int ret;
+
+	ret = si_restrict_performance_levels_before_switch(rdev);
+	if (ret)
+		return ret;
+	ret = si_halt_smc(rdev);
+	if (ret)
+		return ret;
+	ret = si_populate_smc_tdp_limits(rdev, new_ps);
+	if (ret)
+		return ret;
+	ret = si_populate_smc_tdp_limits_2(rdev, new_ps);
+	if (ret)
+		return ret;
+	ret = si_resume_smc(rdev);
+	if (ret)
+		return ret;
+	ret = si_set_sw_state(rdev);
+	if (ret)
+		return ret;
+	return 0;
+}
+
+int si_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+	struct radeon_ps *old_ps = &eg_pi->current_rps;
+	int ret;
+
+	ret = si_disable_ulv(rdev);
+	if (ret) {
+		DRM_ERROR("si_disable_ulv failed\n");
+		return ret;
+	}
+	ret = si_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("si_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	if (eg_pi->pcie_performance_request)
+		si_request_link_speed_change_before_state_change(rdev, new_ps, old_ps);
+	ni_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = si_enable_power_containment(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("si_enable_power_containment failed\n");
+		return ret;
+	}
+	ret = si_enable_smc_cac(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("si_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = si_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("si_halt_smc failed\n");
+		return ret;
+	}
+	ret = si_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("si_upload_sw_state failed\n");
+		return ret;
+	}
+	ret = si_upload_smc_data(rdev);
+	if (ret) {
+		DRM_ERROR("si_upload_smc_data failed\n");
+		return ret;
+	}
+	ret = si_upload_ulv_state(rdev);
+	if (ret) {
+		DRM_ERROR("si_upload_ulv_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = si_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("si_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = si_program_memory_timing_parameters(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("si_program_memory_timing_parameters failed\n");
+		return ret;
+	}
+	si_set_pcie_lane_width_in_smc(rdev, new_ps, old_ps);
+
+	ret = si_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("si_resume_smc failed\n");
+		return ret;
+	}
+	ret = si_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("si_set_sw_state failed\n");
+		return ret;
+	}
+	ni_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+	if (eg_pi->pcie_performance_request)
+		si_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps);
+	ret = si_set_power_state_conditionally_enable_ulv(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("si_set_power_state_conditionally_enable_ulv failed\n");
+		return ret;
+	}
+	ret = si_enable_smc_cac(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("si_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = si_enable_power_containment(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("si_enable_power_containment failed\n");
+		return ret;
+	}
+
+	ret = si_power_control_set_level(rdev);
+	if (ret) {
+		DRM_ERROR("si_power_control_set_level failed\n");
+		return ret;
+	}
+
+#if 0
+	/* XXX */
+	ret = si_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("si_dpm_force_performance_level failed\n");
+		return ret;
+	}
+#else
+	rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO;
+#endif
+
+	return 0;
+}
+
+void si_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+
+	ni_update_current_ps(rdev, new_ps);
+}
+
+
+void si_dpm_reset_asic(struct radeon_device *rdev)
+{
+	si_restrict_performance_levels_before_switch(rdev);
+	si_disable_ulv(rdev);
+	si_set_boot_state(rdev);
+}
+
+void si_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	si_program_display_gap(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+	struct _ATOM_PPLIB_SI_CLOCK_INFO si;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void si_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					  struct radeon_ps *rps,
+					  struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					  u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV770_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV770_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void si_parse_pplib_clock_info(struct radeon_device *rdev,
+				      struct radeon_ps *rps, int index,
+				      union pplib_clock_info *clock_info)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *ps = ni_get_ps(rps);
+	u16 leakage_voltage;
+	struct rv7xx_pl *pl = &ps->performance_levels[index];
+	int ret;
+
+	ps->performance_level_count = index + 1;
+
+	pl->sclk = le16_to_cpu(clock_info->si.usEngineClockLow);
+	pl->sclk |= clock_info->si.ucEngineClockHigh << 16;
+	pl->mclk = le16_to_cpu(clock_info->si.usMemoryClockLow);
+	pl->mclk |= clock_info->si.ucMemoryClockHigh << 16;
+
+	pl->vddc = le16_to_cpu(clock_info->si.usVDDC);
+	pl->vddci = le16_to_cpu(clock_info->si.usVDDCI);
+	pl->flags = le32_to_cpu(clock_info->si.ulFlags);
+	pl->pcie_gen = r600_get_pcie_gen_support(rdev,
+						 si_pi->sys_pcie_mask,
+						 si_pi->boot_pcie_gen,
+						 clock_info->si.ucPCIEGen);
+
+	/* patch up vddc if necessary */
+	ret = si_get_leakage_voltage_from_leakage_index(rdev, pl->vddc,
+							&leakage_voltage);
+	if (ret == 0)
+		pl->vddc = leakage_voltage;
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) {
+		pi->acpi_vddc = pl->vddc;
+		eg_pi->acpi_vddci = pl->vddci;
+		si_pi->acpi_pcie_gen = pl->pcie_gen;
+	}
+
+	if ((rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) &&
+	    index == 0) {
+		/* XXX disable for A0 tahiti */
+		si_pi->ulv.supported = true;
+		si_pi->ulv.pl = *pl;
+		si_pi->ulv.one_pcie_lane_in_ulv = false;
+		si_pi->ulv.volt_change_delay = SISLANDS_ULVVOLTAGECHANGEDELAY_DFLT;
+		si_pi->ulv.cg_ulv_parameter = SISLANDS_CGULVPARAMETER_DFLT;
+		si_pi->ulv.cg_ulv_control = SISLANDS_CGULVCONTROL_DFLT;
+	}
+
+	if (pi->min_vddc_in_table > pl->vddc)
+		pi->min_vddc_in_table = pl->vddc;
+
+	if (pi->max_vddc_in_table < pl->vddc)
+		pi->max_vddc_in_table = pl->vddc;
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+		pl->vddci = vddci;
+		si_pi->mvdd_bootup_value = mvdd;
+	}
+
+	if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
+	    ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci;
+	}
+}
+
+static int si_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j, k, non_clock_array_index, clock_array_index;
+	union pplib_clock_info *clock_info;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	u8 *power_state_offset;
+	struct ni_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	state_array = (struct _StateArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
+	clock_info_array = (struct _ClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
+	non_clock_info_array = (struct _NonClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  state_array->ucNumEntries, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	power_state_offset = (u8 *)state_array->states;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+	for (i = 0; i < state_array->ucNumEntries; i++) {
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			&non_clock_info_array->nonClockInfo[non_clock_array_index];
+		if (!rdev->pm.power_state[i].clock_info)
+			return -EINVAL;
+		ps = kzalloc(sizeof(struct ni_ps), GFP_KERNEL);
+		if (ps == NULL) {
+			kfree(rdev->pm.dpm.ps);
+			return -ENOMEM;
+		}
+		rdev->pm.dpm.ps[i].ps_priv = ps;
+		si_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+					      non_clock_info,
+					      non_clock_info_array->ucEntrySize);
+		k = 0;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			if (clock_array_index >= clock_info_array->ucNumEntries)
+				continue;
+			if (k >= SISLANDS_MAX_HARDWARE_POWERLEVELS)
+				break;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
+			si_parse_pplib_clock_info(rdev,
+						  &rdev->pm.dpm.ps[i], k,
+						  clock_info);
+			k++;
+		}
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
+	}
+	rdev->pm.dpm.num_ps = state_array->ucNumEntries;
+	return 0;
+}
+
+int si_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	struct ni_power_info *ni_pi;
+	struct si_power_info *si_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	u16 data_offset, size;
+	u8 frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+	u32 mask;
+
+	si_pi = kzalloc(sizeof(struct si_power_info), GFP_KERNEL);
+	if (si_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = si_pi;
+	ni_pi = &si_pi->ni;
+	eg_pi = &ni_pi->eg;
+	pi = &eg_pi->rv7xx;
+
+	ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask);
+	if (ret)
+		si_pi->sys_pcie_mask = 0;
+	else
+		si_pi->sys_pcie_mask = mask;
+	si_pi->force_pcie_gen = RADEON_PCIE_GEN_INVALID;
+	si_pi->boot_pcie_gen = si_get_current_pcie_speed(rdev);
+
+	si_set_max_cu_value(rdev);
+
+	rv770_get_max_vddc(rdev);
+	si_get_leakage_vddc(rdev);
+	si_patch_dependency_tables_based_on_leakage(rdev);
+
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = si_parse_power_table(rdev);
+	if (ret)
+		return ret;
+	ret = r600_parse_extended_power_table(rdev);
+	if (ret)
+		return ret;
+
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries =
+		kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL);
+	if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) {
+		r600_free_extended_power_table(rdev);
+		return -ENOMEM;
+	}
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 720;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 810;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 900;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	eg_pi->smu_uvd_hs = false;
+
+	pi->mclk_strobe_mode_threshold = 40000;
+	if (si_is_special_1gb_platform(rdev))
+		pi->mclk_stutter_mode_threshold = 0;
+	else
+		pi->mclk_stutter_mode_threshold = pi->mclk_strobe_mode_threshold;
+	pi->mclk_edc_enable_threshold = 40000;
+	eg_pi->mclk_edc_wr_enable_threshold = 40000;
+
+	ni_pi->mclk_rtt_mode_threshold = eg_pi->mclk_edc_wr_enable_threshold;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_GPIO_LUT);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, VOLTAGE_OBJ_GPIO_LUT);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, VOLTAGE_OBJ_GPIO_LUT);
+
+	si_pi->vddc_phase_shed_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_PHASE_LUT);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = SISLANDS_VRC_DFLT;
+
+	pi->gfx_clock_gating = true;
+
+	eg_pi->sclk_deep_sleep = true;
+	si_pi->sclk_deep_sleep_above_low = false;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	eg_pi->dynamic_ac_timing = true;
+
+	eg_pi->light_sleep = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	si_pi->sram_end = SMC_RAM_END;
+
+	rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 4;
+	rdev->pm.dpm.dyn_state.sclk_mclk_delta = 15000;
+	rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.values = NULL;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL;
+
+	si_initialize_powertune_defaults(rdev);
+
+	return 0;
+}
+
+void si_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+	kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries);
+	r600_free_extended_power_table(rdev);
+}
+
+void si_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct rv7xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_INDEX_MASK) >>
+		CURRENT_STATE_INDEX_SHIFT;
+
+	if (current_index >= ps->performance_level_count) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->performance_levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u vddci: %u pcie gen: %u\n",
+			   current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci, pl->pcie_gen + 1);
+	}
+}
diff --git a/drivers/gpu/drm/radeon/si_dpm.h b/drivers/gpu/drm/radeon/si_dpm.h
new file mode 100644
index 0000000..4ce5032
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_dpm.h
@@ -0,0 +1,227 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SI_DPM_H__
+#define __SI_DPM_H__
+
+#include "ni_dpm.h"
+#include "sislands_smc.h"
+
+enum si_cac_config_reg_type
+{
+	SISLANDS_CACCONFIG_MMR = 0,
+	SISLANDS_CACCONFIG_CGIND,
+	SISLANDS_CACCONFIG_MAX
+};
+
+struct si_cac_config_reg
+{
+	u32 offset;
+	u32 mask;
+	u32 shift;
+	u32 value;
+	enum si_cac_config_reg_type type;
+};
+
+struct si_powertune_data
+{
+	u32 cac_window;
+	u32 l2_lta_window_size_default;
+	u8 lts_truncate_default;
+	u8 shift_n_default;
+	u8 operating_temp;
+	struct ni_leakage_coeffients leakage_coefficients;
+	u32 fixed_kt;
+	u32 lkge_lut_v0_percent;
+	u8 dc_cac[NISLANDS_DCCAC_MAX_LEVELS];
+	bool enable_powertune_by_default;
+};
+
+struct si_dyn_powertune_data
+{
+	u32 cac_leakage;
+	s32 leakage_minimum_temperature;
+	u32 wintime;
+	u32 l2_lta_window_size;
+	u8 lts_truncate;
+	u8 shift_n;
+	u8 dc_pwr_value;
+	bool disable_uvd_powertune;
+};
+
+struct si_dte_data
+{
+	u32 tau[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+	u32 r[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+	u32 k;
+	u32 t0;
+	u32 max_t;
+	u8 window_size;
+	u8 temp_select;
+	u8 dte_mode;
+	u8 tdep_count;
+	u8 t_limits[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+	u32 tdep_tau[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+	u32 tdep_r[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+	u32 t_threshold;
+	bool enable_dte_by_default;
+};
+
+struct si_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 dll_cntl;
+	u32 mclk_pwrmgt_cntl;
+	u32 mpll_ad_func_cntl;
+	u32 mpll_dq_func_cntl;
+	u32 mpll_func_cntl;
+	u32 mpll_func_cntl_1;
+	u32 mpll_func_cntl_2;
+	u32 mpll_ss1;
+	u32 mpll_ss2;
+};
+
+struct si_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct si_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	u16 valid_flag;
+	struct si_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES];
+	SMC_NIslands_MCRegisterAddress mc_reg_address[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+#define SISLANDS_MCREGISTERTABLE_INITIAL_SLOT               0
+#define SISLANDS_MCREGISTERTABLE_ACPI_SLOT                  1
+#define SISLANDS_MCREGISTERTABLE_ULV_SLOT                   2
+#define SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT     3
+
+struct si_leakage_voltage_entry
+{
+	u16 voltage;
+	u16 leakage_index;
+};
+
+#define SISLANDS_LEAKAGE_INDEX0     0xff01
+#define SISLANDS_MAX_LEAKAGE_COUNT  4
+
+struct si_leakage_voltage
+{
+	u16 count;
+	struct si_leakage_voltage_entry entries[SISLANDS_MAX_LEAKAGE_COUNT];
+};
+
+#define SISLANDS_MAX_HARDWARE_POWERLEVELS 5
+
+struct si_ulv_param {
+	bool supported;
+	u32 cg_ulv_control;
+	u32 cg_ulv_parameter;
+	u32 volt_change_delay;
+	struct rv7xx_pl pl;
+	bool one_pcie_lane_in_ulv;
+};
+
+struct si_power_info {
+	/* must be first! */
+	struct ni_power_info ni;
+	struct si_clock_registers clock_registers;
+	struct si_mc_reg_table mc_reg_table;
+	struct atom_voltage_table mvdd_voltage_table;
+	struct atom_voltage_table vddc_phase_shed_table;
+	struct si_leakage_voltage leakage_voltage;
+	u16 mvdd_bootup_value;
+	struct si_ulv_param ulv;
+	u32 max_cu;
+	/* pcie gen */
+	enum radeon_pcie_gen force_pcie_gen;
+	enum radeon_pcie_gen boot_pcie_gen;
+	enum radeon_pcie_gen acpi_pcie_gen;
+	u32 sys_pcie_mask;
+	/* flags */
+	bool enable_dte;
+	bool enable_ppm;
+	bool vddc_phase_shed_control;
+	bool pspp_notify_required;
+	bool sclk_deep_sleep_above_low;
+	/* smc offsets */
+	u32 sram_end;
+	u32 state_table_start;
+	u32 soft_regs_start;
+	u32 mc_reg_table_start;
+	u32 arb_table_start;
+	u32 cac_table_start;
+	u32 dte_table_start;
+	u32 spll_table_start;
+	u32 papm_cfg_table_start;
+	/* CAC stuff */
+	const struct si_cac_config_reg *cac_weights;
+	const struct si_cac_config_reg *lcac_config;
+	const struct si_cac_config_reg *cac_override;
+	const struct si_powertune_data *powertune_data;
+	struct si_dyn_powertune_data dyn_powertune_data;
+	/* DTE stuff */
+	struct si_dte_data dte_data;
+	/* scratch structs */
+	SMC_SIslands_MCRegisters smc_mc_reg_table;
+	SISLANDS_SMC_STATETABLE smc_statetable;
+	PP_SIslands_PAPMParameters papm_parm;
+};
+
+#define SISLANDS_INITIAL_STATE_ARB_INDEX    0
+#define SISLANDS_ACPI_STATE_ARB_INDEX       1
+#define SISLANDS_ULV_STATE_ARB_INDEX        2
+#define SISLANDS_DRIVER_STATE_ARB_INDEX     3
+
+#define SISLANDS_DPM2_MAX_PULSE_SKIP        256
+
+#define SISLANDS_DPM2_NEAR_TDP_DEC          10
+#define SISLANDS_DPM2_ABOVE_SAFE_INC        5
+#define SISLANDS_DPM2_BELOW_SAFE_INC        20
+
+#define SISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT            80
+
+#define SISLANDS_DPM2_MAXPS_PERCENT_H                   99
+#define SISLANDS_DPM2_MAXPS_PERCENT_M                   99
+
+#define SISLANDS_DPM2_SQ_RAMP_MAX_POWER                 0x3FFF
+#define SISLANDS_DPM2_SQ_RAMP_MIN_POWER                 0x12
+#define SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA           0x15
+#define SISLANDS_DPM2_SQ_RAMP_STI_SIZE                  0x1E
+#define SISLANDS_DPM2_SQ_RAMP_LTI_RATIO                 0xF
+
+#define SISLANDS_DPM2_PWREFFICIENCYRATIO_MARGIN         10
+
+#define SISLANDS_VRC_DFLT                               0xC000B3
+#define SISLANDS_ULVVOLTAGECHANGEDELAY_DFLT             1687
+#define SISLANDS_CGULVPARAMETER_DFLT                    0x00040035
+#define SISLANDS_CGULVCONTROL_DFLT                      0x1f007550
+
+
+#endif
diff --git a/drivers/gpu/drm/radeon/si_smc.c b/drivers/gpu/drm/radeon/si_smc.c
new file mode 100644
index 0000000..5f524c0
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_smc.c
@@ -0,0 +1,284 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include <linux/firmware.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "sid.h"
+#include "ppsmc.h"
+#include "radeon_ucode.h"
+
+int si_set_smc_sram_address(struct radeon_device *rdev,
+			    u32 smc_address, u32 limit)
+{
+	if (smc_address & 3)
+		return -EINVAL;
+	if ((smc_address + 3) > limit)
+		return -EINVAL;
+
+	WREG32(SMC_IND_INDEX_0, smc_address);
+	WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0);
+
+	return 0;
+}
+
+int si_copy_bytes_to_smc(struct radeon_device *rdev,
+			 u32 smc_start_address,
+			 const u8 *src, u32 byte_count, u32 limit)
+{
+	int ret;
+	u32 data, original_data, addr, extra_shift;
+
+	if (smc_start_address & 3)
+		return -EINVAL;
+	if ((smc_start_address + byte_count) > limit)
+		return -EINVAL;
+
+	addr = smc_start_address;
+
+	while (byte_count >= 4) {
+		/* SMC address space is BE */
+		data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3];
+
+		ret = si_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_IND_DATA_0, data);
+
+		src += 4;
+		byte_count -= 4;
+		addr += 4;
+	}
+
+	/* RMW for the final bytes */
+	if (byte_count > 0) {
+		data = 0;
+
+		ret = si_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		original_data = RREG32(SMC_IND_DATA_0);
+
+		extra_shift = 8 * (4 - byte_count);
+
+		while (byte_count > 0) {
+			/* SMC address space is BE */
+			data = (data << 8) + *src++;
+			byte_count--;
+		}
+
+		data <<= extra_shift;
+
+		data |= (original_data & ~((~0UL) << extra_shift));
+
+		ret = si_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_IND_DATA_0, data);
+	}
+	return 0;
+}
+
+void si_start_smc(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL);
+
+	tmp &= ~RST_REG;
+
+	WREG32_SMC(SMC_SYSCON_RESET_CNTL, tmp);
+}
+
+void si_reset_smc(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+
+	tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL);
+	tmp |= RST_REG;
+	WREG32_SMC(SMC_SYSCON_RESET_CNTL, tmp);
+}
+
+int si_program_jump_on_start(struct radeon_device *rdev)
+{
+	static u8 data[] = { 0x0E, 0x00, 0x40, 0x40 };
+
+	return si_copy_bytes_to_smc(rdev, 0x0, data, 4, sizeof(data)+1);
+}
+
+void si_stop_smc_clock(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+
+	tmp |= CK_DISABLE;
+
+	WREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0, tmp);
+}
+
+void si_start_smc_clock(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+
+	tmp &= ~CK_DISABLE;
+
+	WREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0, tmp);
+}
+
+bool si_is_smc_running(struct radeon_device *rdev)
+{
+	u32 rst = RREG32_SMC(SMC_SYSCON_RESET_CNTL);
+	u32 clk = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+
+	if (!(rst & RST_REG) && !(clk & CK_DISABLE))
+		return true;
+
+	return false;
+}
+
+PPSMC_Result si_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg)
+{
+	u32 tmp;
+	int i;
+
+	if (!si_is_smc_running(rdev))
+		return PPSMC_Result_Failed;
+
+	WREG32(SMC_MESSAGE_0, msg);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(SMC_RESP_0);
+		if (tmp != 0)
+			break;
+		udelay(1);
+	}
+	tmp = RREG32(SMC_RESP_0);
+
+	return (PPSMC_Result)tmp;
+}
+
+PPSMC_Result si_wait_for_smc_inactive(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int i;
+
+	if (!si_is_smc_running(rdev))
+		return PPSMC_Result_OK;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+		if ((tmp & CKEN) == 0)
+			break;
+		udelay(1);
+	}
+
+	return PPSMC_Result_OK;
+}
+
+int si_load_smc_ucode(struct radeon_device *rdev, u32 limit)
+{
+	u32 ucode_start_address;
+	u32 ucode_size;
+	const u8 *src;
+	u32 data;
+
+	if (!rdev->smc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		ucode_start_address = TAHITI_SMC_UCODE_START;
+		ucode_size = TAHITI_SMC_UCODE_SIZE;
+		break;
+	case CHIP_PITCAIRN:
+		ucode_start_address = PITCAIRN_SMC_UCODE_START;
+		ucode_size = PITCAIRN_SMC_UCODE_SIZE;
+		break;
+	case CHIP_VERDE:
+		ucode_start_address = VERDE_SMC_UCODE_START;
+		ucode_size = VERDE_SMC_UCODE_SIZE;
+		break;
+	case CHIP_OLAND:
+		ucode_start_address = OLAND_SMC_UCODE_START;
+		ucode_size = OLAND_SMC_UCODE_SIZE;
+		break;
+	case CHIP_HAINAN:
+		ucode_start_address = HAINAN_SMC_UCODE_START;
+		ucode_size = HAINAN_SMC_UCODE_SIZE;
+		break;
+	default:
+		DRM_ERROR("unknown asic in smc ucode loader\n");
+		BUG();
+	}
+
+	if (ucode_size & 3)
+		return -EINVAL;
+
+	src = (const u8 *)rdev->smc_fw->data;
+	WREG32(SMC_IND_INDEX_0, ucode_start_address);
+	WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0);
+	while (ucode_size >= 4) {
+		/* SMC address space is BE */
+		data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3];
+
+		WREG32(SMC_IND_DATA_0, data);
+
+		src += 4;
+		ucode_size -= 4;
+	}
+	WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0);
+
+	return 0;
+}
+
+int si_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			   u32 *value, u32 limit)
+{
+	int ret;
+
+	ret = si_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	*value = RREG32(SMC_IND_DATA_0);
+	return 0;
+}
+
+int si_write_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			    u32 value, u32 limit)
+{
+	int ret;
+
+	ret = si_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	WREG32(SMC_IND_DATA_0, value);
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 8f2d7d4..12a20eb 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -30,6 +30,94 @@
 #define VERDE_GB_ADDR_CONFIG_GOLDEN         0x12010002
 #define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02010001
 
+#define SI_MAX_SH_GPRS           256
+#define SI_MAX_TEMP_GPRS         16
+#define SI_MAX_SH_THREADS        256
+#define SI_MAX_SH_STACK_ENTRIES  4096
+#define SI_MAX_FRC_EOV_CNT       16384
+#define SI_MAX_BACKENDS          8
+#define SI_MAX_BACKENDS_MASK     0xFF
+#define SI_MAX_BACKENDS_PER_SE_MASK     0x0F
+#define SI_MAX_SIMDS             12
+#define SI_MAX_SIMDS_MASK        0x0FFF
+#define SI_MAX_SIMDS_PER_SE_MASK        0x00FF
+#define SI_MAX_PIPES             8
+#define SI_MAX_PIPES_MASK        0xFF
+#define SI_MAX_PIPES_PER_SIMD_MASK      0x3F
+#define SI_MAX_LDS_NUM           0xFFFF
+#define SI_MAX_TCC               16
+#define SI_MAX_TCC_MASK          0xFFFF
+
+/* SMC IND accessor regs */
+#define SMC_IND_INDEX_0                              0x200
+#define SMC_IND_DATA_0                               0x204
+
+#define SMC_IND_ACCESS_CNTL                          0x228
+#       define AUTO_INCREMENT_IND_0                  (1 << 0)
+#define SMC_MESSAGE_0                                0x22c
+#define SMC_RESP_0                                   0x230
+
+/* CG IND registers are accessed via SMC indirect space + SMC_CG_IND_START */
+#define SMC_CG_IND_START                    0xc0030000
+#define SMC_CG_IND_END                      0xc0040000
+
+#define	CG_CGTT_LOCAL_0				0x400
+#define	CG_CGTT_LOCAL_1				0x401
+
+/* SMC IND registers */
+#define	SMC_SYSCON_RESET_CNTL				0x80000000
+#       define RST_REG                                  (1 << 0)
+#define	SMC_SYSCON_CLOCK_CNTL_0				0x80000004
+#       define CK_DISABLE                               (1 << 0)
+#       define CKEN                                     (1 << 24)
+
+#define VGA_HDP_CONTROL  				0x328
+#define		VGA_MEMORY_DISABLE				(1 << 4)
+
+#define DCCG_DISP_SLOW_SELECT_REG                       0x4fc
+#define		DCCG_DISP1_SLOW_SELECT(x)		((x) << 0)
+#define		DCCG_DISP1_SLOW_SELECT_MASK		(7 << 0)
+#define		DCCG_DISP1_SLOW_SELECT_SHIFT		0
+#define		DCCG_DISP2_SLOW_SELECT(x)		((x) << 4)
+#define		DCCG_DISP2_SLOW_SELECT_MASK		(7 << 4)
+#define		DCCG_DISP2_SLOW_SELECT_SHIFT		4
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define		SPLL_PDIV_A_SHIFT			20
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_FB_DIV_SHIFT			0
+#define		SPLL_DITHEN				(1 << 28)
+#define	CG_SPLL_FUNC_CNTL_4				0x60c
+
+#define	SPLL_CNTL_MODE					0x618
+#	define SPLL_REFCLK_SEL(x)			((x) << 8)
+#	define SPLL_REFCLK_SEL_MASK			0xFF00
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x620
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define		CLK_S_SHIFT				4
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x624
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+#define		CLK_V_SHIFT				0
+
+#define	CG_SPLL_AUTOSCALE_CNTL				0x62c
+#       define AUTOSCALE_ON_SS_CLEAR                    (1 << 9)
+
 /* discrete uvd clocks */
 #define	CG_UPLL_FUNC_CNTL				0x634
 #	define UPLL_RESET_MASK				0x00000001
@@ -59,6 +147,45 @@
 #define	CG_UPLL_SPREAD_SPECTRUM				0x650
 #	define SSEN_MASK				0x00000001
 
+#define	MPLL_BYPASSCLK_SEL				0x65c
+#	define MPLL_CLKOUT_SEL(x)			((x) << 8)
+#	define MPLL_CLKOUT_SEL_MASK			0xFF00
+
+#define CG_CLKPIN_CNTL                                    0x660
+#       define XTALIN_DIVIDE                              (1 << 1)
+#       define BCLK_AS_XCLK                               (1 << 2)
+#define CG_CLKPIN_CNTL_2                                  0x664
+#       define FORCE_BIF_REFCLK_EN                        (1 << 3)
+#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+
+#define	THM_CLK_CNTL					0x66c
+#	define CMON_CLK_SEL(x)				((x) << 0)
+#	define CMON_CLK_SEL_MASK			0xFF
+#	define TMON_CLK_SEL(x)				((x) << 8)
+#	define TMON_CLK_SEL_MASK			0xFF00
+#define	MISC_CLK_CNTL					0x670
+#	define DEEP_SLEEP_CLK_SEL(x)			((x) << 0)
+#	define DEEP_SLEEP_CLK_SEL_MASK			0xFF
+#	define ZCLK_SEL(x)				((x) << 8)
+#	define ZCLK_SEL_MASK				0xFF00
+
+#define	CG_THERMAL_CTRL					0x700
+#define 	DPM_EVENT_SRC(x)			((x) << 0)
+#define 	DPM_EVENT_SRC_MASK			(7 << 0)
+#define		DIG_THERM_DPM(x)			((x) << 14)
+#define		DIG_THERM_DPM_MASK			0x003FC000
+#define		DIG_THERM_DPM_SHIFT			14
+
+#define	CG_THERMAL_INT					0x708
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+
 #define	CG_MULT_THERMAL_STATUS					0x714
 #define		ASIC_MAX_TEMP(x)				((x) << 0)
 #define		ASIC_MAX_TEMP_MASK				0x000001ff
@@ -67,31 +194,89 @@
 #define		CTF_TEMP_MASK					0x0003fe00
 #define		CTF_TEMP_SHIFT					9
 
-#define SI_MAX_SH_GPRS           256
-#define SI_MAX_TEMP_GPRS         16
-#define SI_MAX_SH_THREADS        256
-#define SI_MAX_SH_STACK_ENTRIES  4096
-#define SI_MAX_FRC_EOV_CNT       16384
-#define SI_MAX_BACKENDS          8
-#define SI_MAX_BACKENDS_MASK     0xFF
-#define SI_MAX_BACKENDS_PER_SE_MASK     0x0F
-#define SI_MAX_SIMDS             12
-#define SI_MAX_SIMDS_MASK        0x0FFF
-#define SI_MAX_SIMDS_PER_SE_MASK        0x00FF
-#define SI_MAX_PIPES             8
-#define SI_MAX_PIPES_MASK        0xFF
-#define SI_MAX_PIPES_PER_SIMD_MASK      0x3F
-#define SI_MAX_LDS_NUM           0xFFFF
-#define SI_MAX_TCC               16
-#define SI_MAX_TCC_MASK          0xFFFF
-
-#define VGA_HDP_CONTROL  				0x328
-#define		VGA_MEMORY_DISABLE				(1 << 4)
-
-#define CG_CLKPIN_CNTL                                    0x660
-#       define XTALIN_DIVIDE                              (1 << 1)
-#define CG_CLKPIN_CNTL_2                                  0x664
-#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+#define GENERAL_PWRMGT                                  0x780
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (1 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#define CG_TPC                                            0x784
+#define SCLK_PWRMGT_CNTL                                  0x788
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x798
+#       define CURRENT_STATE_INDEX_MASK                   (0xf << 4)
+#       define CURRENT_STATE_INDEX_SHIFT                  4
+
+#define CG_FTV                                            0x7bc
+
+#define CG_FFCT_0                                         0x7c0
+#       define UTC_0(x)                                   ((x) << 0)
+#       define UTC_0_MASK                                 (0x3ff << 0)
+#       define DTC_0(x)                                   ((x) << 10)
+#       define DTC_0_MASK                                 (0x3ff << 10)
+
+#define CG_BSP                                          0x7fc
+#       define BSP(x)					((x) << 0)
+#       define BSP_MASK					(0xffff << 0)
+#       define BSU(x)					((x) << 16)
+#       define BSU_MASK					(0xf << 16)
+#define CG_AT                                           0x800
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+
+#define CG_GIT                                          0x804
+#       define CG_GICST(x)                              ((x) << 0)
+#       define CG_GICST_MASK                            (0xffff << 0)
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+
+#define CG_SSP                                            0x80c
+#       define SST(x)                                     ((x) << 0)
+#       define SST_MASK                                   (0xffff << 0)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (0xf << 16)
+
+#define CG_DISPLAY_GAP_CNTL                               0x828
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define	CG_ULV_CONTROL					0x878
+#define	CG_ULV_PARAMETER				0x87c
+
+#define	SMC_SCRATCH0					0x884
+
+#define	CG_CAC_CTRL					0x8b8
+#	define CAC_WINDOW(x)				((x) << 0)
+#	define CAC_WINDOW_MASK				0x00ffffff
 
 #define DMIF_ADDR_CONFIG  				0xBD4
 
@@ -203,6 +388,10 @@
 #define	VM_CONTEXT0_PAGE_TABLE_END_ADDR			0x157C
 #define	VM_CONTEXT1_PAGE_TABLE_END_ADDR			0x1580
 
+#define VM_L2_CG           				0x15c0
+#define		MC_CG_ENABLE				(1 << 18)
+#define		MC_LS_ENABLE				(1 << 19)
+
 #define MC_SHARED_CHMAP						0x2004
 #define		NOOFCHAN_SHIFT					12
 #define		NOOFCHAN_MASK					0x0000f000
@@ -228,6 +417,17 @@
 
 #define MC_SHARED_BLACKOUT_CNTL           		0x20ac
 
+#define MC_HUB_MISC_HUB_CG           			0x20b8
+#define MC_HUB_MISC_VM_CG           			0x20bc
+
+#define MC_HUB_MISC_SIP_CG           			0x20c0
+
+#define MC_XPB_CLK_GAT           			0x2478
+
+#define MC_CITF_MISC_RD_CG           			0x2648
+#define MC_CITF_MISC_WR_CG           			0x264c
+#define MC_CITF_MISC_VM_CG           			0x2650
+
 #define	MC_ARB_RAMCFG					0x2760
 #define		NOOFBANK_SHIFT					0
 #define		NOOFBANK_MASK					0x00000003
@@ -243,6 +443,23 @@
 #define		NOOFGROUPS_SHIFT				12
 #define		NOOFGROUPS_MASK					0x00001000
 
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE0_SHIFT				0
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE1_SHIFT				5
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE2_SHIFT				10
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+#define		STATE3_SHIFT				15
+
 #define	MC_SEQ_TRAIN_WAKEUP_CNTL			0x2808
 #define		TRAIN_DONE_D0      			(1 << 30)
 #define		TRAIN_DONE_D1      			(1 << 31)
@@ -250,13 +467,105 @@
 #define MC_SEQ_SUP_CNTL           			0x28c8
 #define		RUN_MASK      				(1 << 0)
 #define MC_SEQ_SUP_PGM           			0x28cc
+#define MC_PMG_AUTO_CMD           			0x28d0
 
 #define MC_IO_PAD_CNTL_D0           			0x29d0
 #define		MEM_FALL_OUT_CMD      			(1 << 8)
 
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+#define MC_SEQ_PMG_TIMING                               0x28b0
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_SEQ_MISC0           				0x2a00
+#define 	MC_SEQ_MISC0_VEN_ID_SHIFT               8
+#define 	MC_SEQ_MISC0_VEN_ID_MASK                0x00000f00
+#define 	MC_SEQ_MISC0_VEN_ID_VALUE               3
+#define 	MC_SEQ_MISC0_REV_ID_SHIFT               12
+#define 	MC_SEQ_MISC0_REV_ID_MASK                0x0000f000
+#define 	MC_SEQ_MISC0_REV_ID_VALUE               1
+#define 	MC_SEQ_MISC0_GDDR5_SHIFT                28
+#define 	MC_SEQ_MISC0_GDDR5_MASK                 0xf0000000
+#define 	MC_SEQ_MISC0_GDDR5_VALUE                5
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
 #define MC_SEQ_IO_DEBUG_INDEX           		0x2a44
 #define MC_SEQ_IO_DEBUG_DATA           			0x2a48
 
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+#define MC_SEQ_PMG_TIMING_LP                            0x2b4c
+
+#define MC_SEQ_WR_CTL_2                                 0x2b54
+#define MC_SEQ_WR_CTL_2_LP                              0x2b58
+#define MC_PMG_CMD_MRS2                                 0x2b5c
+#define MC_SEQ_PMG_CMD_MRS2_LP                          0x2b60
+
+#define	MCLK_PWRMGT_CNTL				0x2ba0
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCK0_PDNB                              (1 << 8)
+#       define MRDCK1_PDNB                              (1 << 9)
+#       define MRDCK0_RESET                             (1 << 16)
+#       define MRDCK1_RESET                             (1 << 17)
+#       define DLL_READY_READ                           (1 << 24)
+#define	DLL_CNTL					0x2ba4
+#       define MRDCK0_BYPASS                            (1 << 24)
+#       define MRDCK1_BYPASS                            (1 << 25)
+
+#define	MPLL_FUNC_CNTL					0x2bb4
+#define		BWCTRL(x)				((x) << 20)
+#define		BWCTRL_MASK				(0xff << 20)
+#define	MPLL_FUNC_CNTL_1				0x2bb8
+#define		VCO_MODE(x)				((x) << 0)
+#define		VCO_MODE_MASK				(3 << 0)
+#define		CLKFRAC(x)				((x) << 4)
+#define		CLKFRAC_MASK				(0xfff << 4)
+#define		CLKF(x)					((x) << 16)
+#define		CLKF_MASK				(0xfff << 16)
+#define	MPLL_FUNC_CNTL_2				0x2bbc
+#define	MPLL_AD_FUNC_CNTL				0x2bc0
+#define		YCLK_POST_DIV(x)			((x) << 0)
+#define		YCLK_POST_DIV_MASK			(7 << 0)
+#define	MPLL_DQ_FUNC_CNTL				0x2bc4
+#define		YCLK_SEL(x)				((x) << 4)
+#define		YCLK_SEL_MASK				(1 << 4)
+
+#define	MPLL_SS1					0x2bcc
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x2bd0
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
 #define	HDP_HOST_PATH_CNTL				0x2C00
 #define	HDP_NONSURFACE_BASE				0x2C04
 #define	HDP_NONSURFACE_INFO				0x2C08
@@ -266,6 +575,8 @@
 #define HDP_MISC_CNTL					0x2F4C
 #define 	HDP_FLUSH_INVALIDATE_CACHE			(1 << 0)
 
+#define ATC_MISC_CG           				0x3350
+
 #define IH_RB_CNTL                                        0x3e00
 #       define IH_RB_ENABLE                               (1 << 0)
 #       define IH_IB_SIZE(x)                              ((x) << 1) /* log2 */
@@ -424,6 +735,9 @@
 #       define DC_HPDx_RX_INT_TIMER(x)                    ((x) << 16)
 #       define DC_HPDx_EN                                 (1 << 28)
 
+#define DPG_PIPE_STUTTER_CONTROL                          0x6cd4
+#       define STUTTER_ENABLE                             (1 << 0)
+
 /* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */
 #define CRTC_STATUS_FRAME_COUNT                         0x6e98
 
@@ -599,6 +913,24 @@
 
 #define	SQC_CACHES					0x8C08
 
+#define SQ_POWER_THROTTLE                               0x8e58
+#define		MIN_POWER(x)				((x) << 0)
+#define		MIN_POWER_MASK				(0x3fff << 0)
+#define		MIN_POWER_SHIFT				0
+#define		MAX_POWER(x)				((x) << 16)
+#define		MAX_POWER_MASK				(0x3fff << 16)
+#define		MAX_POWER_SHIFT				0
+#define SQ_POWER_THROTTLE2                              0x8e5c
+#define		MAX_POWER_DELTA(x)			((x) << 0)
+#define		MAX_POWER_DELTA_MASK			(0x3fff << 0)
+#define		MAX_POWER_DELTA_SHIFT			0
+#define		STI_SIZE(x)				((x) << 16)
+#define		STI_SIZE_MASK				(0x3ff << 16)
+#define		STI_SIZE_SHIFT				16
+#define		LTI_RATIO(x)				((x) << 27)
+#define		LTI_RATIO_MASK				(0xf << 27)
+#define		LTI_RATIO_SHIFT				27
+
 #define	SX_DEBUG_1					0x9060
 
 #define	SPI_STATIC_THREAD_MGMT_1			0x90E0
@@ -616,6 +948,11 @@
 #define	CGTS_USER_TCC_DISABLE				0x914C
 #define		TCC_DISABLE_MASK				0xFFFF0000
 #define		TCC_DISABLE_SHIFT				16
+#define	CGTS_SM_CTRL_REG				0x9150
+#define		OVERRIDE				(1 << 21)
+#define		LS_OVERRIDE				(1 << 22)
+
+#define	SPI_LB_CU_MASK					0x9354
 
 #define	TA_CNTL_AUX					0x9508
 
@@ -705,6 +1042,8 @@
 #define	CB_PERFCOUNTER3_SELECT0				0x9a38
 #define	CB_PERFCOUNTER3_SELECT1				0x9a3c
 
+#define	CB_CGTT_SCLK_CTRL				0x9a60
+
 #define	GC_USER_RB_BACKEND_DISABLE			0x9B7C
 #define		BACKEND_DISABLE_MASK			0x00FF0000
 #define		BACKEND_DISABLE_SHIFT			16
@@ -762,6 +1101,9 @@
 #       define CP_RINGID1_INT_STAT                      (1 << 30)
 #       define CP_RINGID0_INT_STAT                      (1 << 31)
 
+#define	CP_MEM_SLP_CNTL					0xC1E4
+#       define CP_MEM_LS_EN                             (1 << 0)
+
 #define	CP_DEBUG					0xC1FC
 
 #define RLC_CNTL                                          0xC300
@@ -769,6 +1111,7 @@
 #define RLC_RL_BASE                                       0xC304
 #define RLC_RL_SIZE                                       0xC308
 #define RLC_LB_CNTL                                       0xC30C
+#       define LOAD_BALANCE_ENABLE                        (1 << 0)
 #define RLC_SAVE_AND_RESTORE_BASE                         0xC310
 #define RLC_LB_CNTR_MAX                                   0xC314
 #define RLC_LB_CNTR_INIT                                  0xC318
@@ -783,6 +1126,56 @@
 #define RLC_CAPTURE_GPU_CLOCK_COUNT                       0xC340
 #define RLC_MC_CNTL                                       0xC344
 #define RLC_UCODE_CNTL                                    0xC348
+#define RLC_STAT                                          0xC34C
+#       define RLC_BUSY_STATUS                            (1 << 0)
+#       define GFX_POWER_STATUS                           (1 << 1)
+#       define GFX_CLOCK_STATUS                           (1 << 2)
+#       define GFX_LS_STATUS                              (1 << 3)
+
+#define	RLC_PG_CNTL					0xC35C
+#	define GFX_PG_ENABLE				(1 << 0)
+#	define GFX_PG_SRC				(1 << 1)
+
+#define	RLC_CGTT_MGCG_OVERRIDE				0xC400
+#define	RLC_CGCG_CGLS_CTRL				0xC404
+#	define CGCG_EN					(1 << 0)
+#	define CGLS_EN					(1 << 1)
+
+#define	RLC_TTOP_D					0xC414
+#	define RLC_PUD(x)				((x) << 0)
+#	define RLC_PUD_MASK				(0xff << 0)
+#	define RLC_PDD(x)				((x) << 8)
+#	define RLC_PDD_MASK				(0xff << 8)
+#	define RLC_TTPD(x)				((x) << 16)
+#	define RLC_TTPD_MASK				(0xff << 16)
+#	define RLC_MSD(x)				((x) << 24)
+#	define RLC_MSD_MASK				(0xff << 24)
+
+#define RLC_LB_INIT_CU_MASK                               0xC41C
+
+#define	RLC_PG_AO_CU_MASK				0xC42C
+#define	RLC_MAX_PG_CU					0xC430
+#	define MAX_PU_CU(x)				((x) << 0)
+#	define MAX_PU_CU_MASK				(0xff << 0)
+#define	RLC_AUTO_PG_CTRL				0xC434
+#	define AUTO_PG_EN				(1 << 0)
+#	define GRBM_REG_SGIT(x)				((x) << 3)
+#	define GRBM_REG_SGIT_MASK			(0xffff << 3)
+#	define PG_AFTER_GRBM_REG_ST(x)			((x) << 19)
+#	define PG_AFTER_GRBM_REG_ST_MASK		(0x1fff << 19)
+
+#define RLC_SERDES_WR_MASTER_MASK_0                       0xC454
+#define RLC_SERDES_WR_MASTER_MASK_1                       0xC458
+#define RLC_SERDES_WR_CTRL                                0xC45C
+
+#define RLC_SERDES_MASTER_BUSY_0                          0xC464
+#define RLC_SERDES_MASTER_BUSY_1                          0xC468
+
+#define RLC_GCPM_GENERAL_3                                0xC478
+
+#define	DB_RENDER_CONTROL				0x28000
+
+#define DB_DEPTH_INFO                                   0x2803c
 
 #define PA_SC_RASTER_CONFIG                             0x28350
 #       define RASTER_CONFIG_RB_MAP_0                   0
@@ -829,6 +1222,146 @@
 #       define THREAD_TRACE_FLUSH                       (54 << 0)
 #       define THREAD_TRACE_FINISH                      (55 << 0)
 
+/* PIF PHY0 registers idx/data 0x8/0xc */
+#define PB0_PIF_CNTL                                      0x10
+#       define LS2_EXIT_TIME(x)                           ((x) << 17)
+#       define LS2_EXIT_TIME_MASK                         (0x7 << 17)
+#       define LS2_EXIT_TIME_SHIFT                        17
+#define PB0_PIF_PAIRING                                   0x11
+#       define MULTI_PIF                                  (1 << 25)
+#define PB0_PIF_PWRDOWN_0                                 0x12
+#       define PLL_POWER_STATE_IN_TXS2_0(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_0(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_SHIFT             10
+#       define PLL_RAMP_UP_TIME_0(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_0_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_0_SHIFT                   24
+#define PB0_PIF_PWRDOWN_1                                 0x13
+#       define PLL_POWER_STATE_IN_TXS2_1(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_1(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_SHIFT             10
+#       define PLL_RAMP_UP_TIME_1(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_1_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_1_SHIFT                   24
+
+#define PB0_PIF_PWRDOWN_2                                 0x17
+#       define PLL_POWER_STATE_IN_TXS2_2(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_2_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_2_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_2(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_2_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_2_SHIFT             10
+#       define PLL_RAMP_UP_TIME_2(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_2_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_2_SHIFT                   24
+#define PB0_PIF_PWRDOWN_3                                 0x18
+#       define PLL_POWER_STATE_IN_TXS2_3(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_3_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_3_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_3(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_3_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_3_SHIFT             10
+#       define PLL_RAMP_UP_TIME_3(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_3_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_3_SHIFT                   24
+/* PIF PHY1 registers idx/data 0x10/0x14 */
+#define PB1_PIF_CNTL                                      0x10
+#define PB1_PIF_PAIRING                                   0x11
+#define PB1_PIF_PWRDOWN_0                                 0x12
+#define PB1_PIF_PWRDOWN_1                                 0x13
+
+#define PB1_PIF_PWRDOWN_2                                 0x17
+#define PB1_PIF_PWRDOWN_3                                 0x18
+/* PCIE registers idx/data 0x30/0x34 */
+#define PCIE_CNTL2                                        0x1c /* PCIE */
+#       define SLV_MEM_LS_EN                              (1 << 16)
+#       define MST_MEM_LS_EN                              (1 << 18)
+#       define REPLAY_MEM_LS_EN                           (1 << 19)
+#define PCIE_LC_STATUS1                                   0x28 /* PCIE */
+#       define LC_REVERSE_RCVR                            (1 << 0)
+#       define LC_REVERSE_XMIT                            (1 << 1)
+#       define LC_OPERATING_LINK_WIDTH_MASK               (0x7 << 2)
+#       define LC_OPERATING_LINK_WIDTH_SHIFT              2
+#       define LC_DETECTED_LINK_WIDTH_MASK                (0x7 << 5)
+#       define LC_DETECTED_LINK_WIDTH_SHIFT               5
+
+#define PCIE_P_CNTL                                       0x40 /* PCIE */
+#       define P_IGNORE_EDB_ERR                           (1 << 6)
+
+/* PCIE PORT registers idx/data 0x38/0x3c */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
+#define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
+#       define LC_LINK_WIDTH_SHIFT                        0
+#       define LC_LINK_WIDTH_MASK                         0x7
+#       define LC_LINK_WIDTH_X0                           0
+#       define LC_LINK_WIDTH_X1                           1
+#       define LC_LINK_WIDTH_X2                           2
+#       define LC_LINK_WIDTH_X4                           3
+#       define LC_LINK_WIDTH_X8                           4
+#       define LC_LINK_WIDTH_X16                          6
+#       define LC_LINK_WIDTH_RD_SHIFT                     4
+#       define LC_LINK_WIDTH_RD_MASK                      0x70
+#       define LC_RECONFIG_ARC_MISSING_ESCAPE             (1 << 7)
+#       define LC_RECONFIG_NOW                            (1 << 8)
+#       define LC_RENEGOTIATION_SUPPORT                   (1 << 9)
+#       define LC_RENEGOTIATE_EN                          (1 << 10)
+#       define LC_SHORT_RECONFIG_EN                       (1 << 11)
+#       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
+#       define LC_UPCONFIGURE_DIS                         (1 << 13)
+#       define LC_DYN_LANES_PWR_STATE(x)                  ((x) << 21)
+#       define LC_DYN_LANES_PWR_STATE_MASK                (0x3 << 21)
+#       define LC_DYN_LANES_PWR_STATE_SHIFT               21
+#define PCIE_LC_N_FTS_CNTL                                0xa3 /* PCIE_P */
+#       define LC_XMIT_N_FTS(x)                           ((x) << 0)
+#       define LC_XMIT_N_FTS_MASK                         (0xff << 0)
+#       define LC_XMIT_N_FTS_SHIFT                        0
+#       define LC_XMIT_N_FTS_OVERRIDE_EN                  (1 << 8)
+#       define LC_N_FTS_MASK                              (0xff << 24)
+#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
+#       define LC_GEN2_EN_STRAP                           (1 << 0)
+#       define LC_GEN3_EN_STRAP                           (1 << 1)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 2)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_MASK         (0x3 << 3)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_SHIFT        3
+#       define LC_FORCE_EN_SW_SPEED_CHANGE                (1 << 5)
+#       define LC_FORCE_DIS_SW_SPEED_CHANGE               (1 << 6)
+#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 7)
+#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 8)
+#       define LC_INITIATE_LINK_SPEED_CHANGE              (1 << 9)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 10)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     10
+#       define LC_CURRENT_DATA_RATE_MASK                  (0x3 << 13) /* 0/1/2 = gen1/2/3 */
+#       define LC_CURRENT_DATA_RATE_SHIFT                 13
+#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 16)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 18)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 19)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN3               (1 << 20)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN3                (1 << 21)
+
+#define PCIE_LC_CNTL2                                     0xb1
+#       define LC_ALLOW_PDWN_IN_L1                        (1 << 17)
+#       define LC_ALLOW_PDWN_IN_L23                       (1 << 18)
+
+#define PCIE_LC_CNTL3                                     0xb5 /* PCIE_P */
+#       define LC_GO_TO_RECOVERY                          (1 << 30)
+#define PCIE_LC_CNTL4                                     0xb6 /* PCIE_P */
+#       define LC_REDO_EQ                                 (1 << 5)
+#       define LC_SET_QUIESCE                             (1 << 13)
+
 /*
  * UVD
  */
@@ -838,6 +1371,21 @@
 #define UVD_RBC_RB_RPTR					0xF690
 #define UVD_RBC_RB_WPTR					0xF694
 
+#define	UVD_CGC_CTRL					0xF4B0
+#	define DCM					(1 << 0)
+#	define CG_DT(x)					((x) << 2)
+#	define CG_DT_MASK				(0xf << 2)
+#	define CLK_OD(x)				((x) << 6)
+#	define CLK_OD_MASK				(0x1f << 6)
+
+ /* UVD CTX indirect */
+#define	UVD_CGC_MEM_CTRL				0xC0
+#define	UVD_CGC_CTRL2					0xC1
+#	define DYN_OR_EN				(1 << 0)
+#	define DYN_RR_EN				(1 << 1)
+#	define G_DIV_ID(x)				((x) << 2)
+#	define G_DIV_ID_MASK				(0x7 << 2)
+
 /*
  * PM4
  */
@@ -1082,6 +1630,11 @@
 #       define DMA_IDLE                                   (1 << 0)
 #define DMA_TILING_CONFIG  				  0xd0b8
 
+#define	DMA_PG						0xd0d4
+#	define PG_CNTL_ENABLE				(1 << 0)
+#define	DMA_PGFSM_CONFIG				0xd0d8
+#define	DMA_PGFSM_WRITE					0xd0dc
+
 #define DMA_PACKET(cmd, b, t, s, n)	((((cmd) & 0xF) << 28) |	\
 					 (((b) & 0x1) << 26) |		\
 					 (((t) & 0x1) << 23) |		\
diff --git a/drivers/gpu/drm/radeon/sislands_smc.h b/drivers/gpu/drm/radeon/sislands_smc.h
new file mode 100644
index 0000000..5578e98
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sislands_smc.h
@@ -0,0 +1,397 @@
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef PP_SISLANDS_SMC_H
+#define PP_SISLANDS_SMC_H
+
+#include "ppsmc.h"
+
+#pragma pack(push, 1)
+
+#define SISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE 16
+
+struct PP_SIslands_Dpm2PerfLevel
+{
+    uint8_t MaxPS;
+    uint8_t TgtAct;
+    uint8_t MaxPS_StepInc;
+    uint8_t MaxPS_StepDec;
+    uint8_t PSSamplingTime;
+    uint8_t NearTDPDec;
+    uint8_t AboveSafeInc;
+    uint8_t BelowSafeInc;
+    uint8_t PSDeltaLimit;
+    uint8_t PSDeltaWin;
+    uint16_t PwrEfficiencyRatio;
+    uint8_t Reserved[4];
+};
+
+typedef struct PP_SIslands_Dpm2PerfLevel PP_SIslands_Dpm2PerfLevel;
+
+struct PP_SIslands_DPM2Status
+{
+    uint32_t    dpm2Flags;
+    uint8_t     CurrPSkip;
+    uint8_t     CurrPSkipPowerShift;
+    uint8_t     CurrPSkipTDP;
+    uint8_t     CurrPSkipOCP;
+    uint8_t     MaxSPLLIndex;
+    uint8_t     MinSPLLIndex;
+    uint8_t     CurrSPLLIndex;
+    uint8_t     InfSweepMode;
+    uint8_t     InfSweepDir;
+    uint8_t     TDPexceeded;
+    uint8_t     reserved;
+    uint8_t     SwitchDownThreshold;
+    uint32_t    SwitchDownCounter;
+    uint32_t    SysScalingFactor;
+};
+
+typedef struct PP_SIslands_DPM2Status PP_SIslands_DPM2Status;
+
+struct PP_SIslands_DPM2Parameters
+{
+    uint32_t    TDPLimit;
+    uint32_t    NearTDPLimit;
+    uint32_t    SafePowerLimit;
+    uint32_t    PowerBoostLimit;
+    uint32_t    MinLimitDelta;
+};
+typedef struct PP_SIslands_DPM2Parameters PP_SIslands_DPM2Parameters;
+
+struct PP_SIslands_PAPMStatus
+{
+    uint32_t    EstimatedDGPU_T;
+    uint32_t    EstimatedDGPU_P;
+    uint32_t    EstimatedAPU_T;
+    uint32_t    EstimatedAPU_P;
+    uint8_t     dGPU_T_Limit_Exceeded;
+    uint8_t     reserved[3];
+};
+typedef struct PP_SIslands_PAPMStatus PP_SIslands_PAPMStatus;
+
+struct PP_SIslands_PAPMParameters
+{
+    uint32_t    NearTDPLimitTherm;
+    uint32_t    NearTDPLimitPAPM;
+    uint32_t    PlatformPowerLimit;
+    uint32_t    dGPU_T_Limit;
+    uint32_t    dGPU_T_Warning;
+    uint32_t    dGPU_T_Hysteresis;
+};
+typedef struct PP_SIslands_PAPMParameters PP_SIslands_PAPMParameters;
+
+struct SISLANDS_SMC_SCLK_VALUE
+{
+    uint32_t    vCG_SPLL_FUNC_CNTL;
+    uint32_t    vCG_SPLL_FUNC_CNTL_2;
+    uint32_t    vCG_SPLL_FUNC_CNTL_3;
+    uint32_t    vCG_SPLL_FUNC_CNTL_4;
+    uint32_t    vCG_SPLL_SPREAD_SPECTRUM;
+    uint32_t    vCG_SPLL_SPREAD_SPECTRUM_2;
+    uint32_t    sclk_value;
+};
+
+typedef struct SISLANDS_SMC_SCLK_VALUE SISLANDS_SMC_SCLK_VALUE;
+
+struct SISLANDS_SMC_MCLK_VALUE
+{
+    uint32_t    vMPLL_FUNC_CNTL;
+    uint32_t    vMPLL_FUNC_CNTL_1;
+    uint32_t    vMPLL_FUNC_CNTL_2;
+    uint32_t    vMPLL_AD_FUNC_CNTL;
+    uint32_t    vMPLL_DQ_FUNC_CNTL;
+    uint32_t    vMCLK_PWRMGT_CNTL;
+    uint32_t    vDLL_CNTL;
+    uint32_t    vMPLL_SS;
+    uint32_t    vMPLL_SS2;
+    uint32_t    mclk_value;
+};
+
+typedef struct SISLANDS_SMC_MCLK_VALUE SISLANDS_SMC_MCLK_VALUE;
+
+struct SISLANDS_SMC_VOLTAGE_VALUE
+{
+    uint16_t    value;
+    uint8_t     index;
+    uint8_t     phase_settings;
+};
+
+typedef struct SISLANDS_SMC_VOLTAGE_VALUE SISLANDS_SMC_VOLTAGE_VALUE;
+
+struct SISLANDS_SMC_HW_PERFORMANCE_LEVEL
+{
+    uint8_t                     ACIndex;
+    uint8_t                     displayWatermark;
+    uint8_t                     gen2PCIE;
+    uint8_t                     UVDWatermark;
+    uint8_t                     VCEWatermark;
+    uint8_t                     strobeMode;
+    uint8_t                     mcFlags;
+    uint8_t                     padding;
+    uint32_t                    aT;
+    uint32_t                    bSP;
+    SISLANDS_SMC_SCLK_VALUE     sclk;
+    SISLANDS_SMC_MCLK_VALUE     mclk;
+    SISLANDS_SMC_VOLTAGE_VALUE  vddc;
+    SISLANDS_SMC_VOLTAGE_VALUE  mvdd;
+    SISLANDS_SMC_VOLTAGE_VALUE  vddci;
+    SISLANDS_SMC_VOLTAGE_VALUE  std_vddc;
+    uint8_t                     hysteresisUp;
+    uint8_t                     hysteresisDown;
+    uint8_t                     stateFlags;
+    uint8_t                     arbRefreshState;
+    uint32_t                    SQPowerThrottle;
+    uint32_t                    SQPowerThrottle_2;
+    uint32_t                    MaxPoweredUpCU;
+    SISLANDS_SMC_VOLTAGE_VALUE  high_temp_vddc;
+    SISLANDS_SMC_VOLTAGE_VALUE  low_temp_vddc;
+    uint32_t                    reserved[2];
+    PP_SIslands_Dpm2PerfLevel   dpm2;
+};
+
+#define SISLANDS_SMC_STROBE_RATIO    0x0F
+#define SISLANDS_SMC_STROBE_ENABLE   0x10
+
+#define SISLANDS_SMC_MC_EDC_RD_FLAG  0x01
+#define SISLANDS_SMC_MC_EDC_WR_FLAG  0x02
+#define SISLANDS_SMC_MC_RTT_ENABLE   0x04
+#define SISLANDS_SMC_MC_STUTTER_EN   0x08
+#define SISLANDS_SMC_MC_PG_EN        0x10
+
+typedef struct SISLANDS_SMC_HW_PERFORMANCE_LEVEL SISLANDS_SMC_HW_PERFORMANCE_LEVEL;
+
+struct SISLANDS_SMC_SWSTATE
+{
+    uint8_t                             flags;
+    uint8_t                             levelCount;
+    uint8_t                             padding2;
+    uint8_t                             padding3;
+    SISLANDS_SMC_HW_PERFORMANCE_LEVEL   levels[1];
+};
+
+typedef struct SISLANDS_SMC_SWSTATE SISLANDS_SMC_SWSTATE;
+
+#define SISLANDS_SMC_VOLTAGEMASK_VDDC  0
+#define SISLANDS_SMC_VOLTAGEMASK_MVDD  1
+#define SISLANDS_SMC_VOLTAGEMASK_VDDCI 2
+#define SISLANDS_SMC_VOLTAGEMASK_MAX   4
+
+struct SISLANDS_SMC_VOLTAGEMASKTABLE
+{
+    uint32_t lowMask[SISLANDS_SMC_VOLTAGEMASK_MAX];
+};
+
+typedef struct SISLANDS_SMC_VOLTAGEMASKTABLE SISLANDS_SMC_VOLTAGEMASKTABLE;
+
+#define SISLANDS_MAX_NO_VREG_STEPS 32
+
+struct SISLANDS_SMC_STATETABLE
+{
+    uint8_t                             thermalProtectType;
+    uint8_t                             systemFlags;
+    uint8_t                             maxVDDCIndexInPPTable;
+    uint8_t                             extraFlags;
+    uint32_t                            lowSMIO[SISLANDS_MAX_NO_VREG_STEPS];
+    SISLANDS_SMC_VOLTAGEMASKTABLE       voltageMaskTable;
+    SISLANDS_SMC_VOLTAGEMASKTABLE       phaseMaskTable;
+    PP_SIslands_DPM2Parameters          dpm2Params;
+    SISLANDS_SMC_SWSTATE                initialState;
+    SISLANDS_SMC_SWSTATE                ACPIState;
+    SISLANDS_SMC_SWSTATE                ULVState;
+    SISLANDS_SMC_SWSTATE                driverState;
+    SISLANDS_SMC_HW_PERFORMANCE_LEVEL   dpmLevels[SISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1];
+};
+
+typedef struct SISLANDS_SMC_STATETABLE SISLANDS_SMC_STATETABLE;
+
+#define SI_SMC_SOFT_REGISTER_mclk_chg_timeout         0x0
+#define SI_SMC_SOFT_REGISTER_delay_vreg               0xC
+#define SI_SMC_SOFT_REGISTER_delay_acpi               0x28
+#define SI_SMC_SOFT_REGISTER_seq_index                0x5C
+#define SI_SMC_SOFT_REGISTER_mvdd_chg_time            0x60
+#define SI_SMC_SOFT_REGISTER_mclk_switch_lim          0x70
+#define SI_SMC_SOFT_REGISTER_watermark_threshold      0x78
+#define SI_SMC_SOFT_REGISTER_phase_shedding_delay     0x88
+#define SI_SMC_SOFT_REGISTER_ulv_volt_change_delay    0x8C
+#define SI_SMC_SOFT_REGISTER_mc_block_delay           0x98
+#define SI_SMC_SOFT_REGISTER_ticks_per_us             0xA8
+#define SI_SMC_SOFT_REGISTER_crtc_index               0xC4
+#define SI_SMC_SOFT_REGISTER_mclk_change_block_cp_min 0xC8
+#define SI_SMC_SOFT_REGISTER_mclk_change_block_cp_max 0xCC
+#define SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width  0xF4
+#define SI_SMC_SOFT_REGISTER_tdr_is_about_to_happen   0xFC
+#define SI_SMC_SOFT_REGISTER_vr_hot_gpio              0x100
+
+#define SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
+#define SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 32
+
+#define SMC_SISLANDS_SCALE_I  7
+#define SMC_SISLANDS_SCALE_R 12
+
+struct PP_SIslands_CacConfig
+{
+    uint16_t   cac_lkge_lut[SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES][SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES];
+    uint32_t   lkge_lut_V0;
+    uint32_t   lkge_lut_Vstep;
+    uint32_t   WinTime;
+    uint32_t   R_LL;
+    uint32_t   calculation_repeats;
+    uint32_t   l2numWin_TDP;
+    uint32_t   dc_cac;
+    uint8_t    lts_truncate_n;
+    uint8_t    SHIFT_N;
+    uint8_t    log2_PG_LKG_SCALE;
+    uint8_t    cac_temp;
+    uint32_t   lkge_lut_T0;
+    uint32_t   lkge_lut_Tstep;
+};
+
+typedef struct PP_SIslands_CacConfig PP_SIslands_CacConfig;
+
+#define SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE 16
+#define SMC_SISLANDS_MC_REGISTER_ARRAY_SET_COUNT 20
+
+struct SMC_SIslands_MCRegisterAddress
+{
+    uint16_t s0;
+    uint16_t s1;
+};
+
+typedef struct SMC_SIslands_MCRegisterAddress SMC_SIslands_MCRegisterAddress;
+
+struct SMC_SIslands_MCRegisterSet
+{
+    uint32_t value[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+typedef struct SMC_SIslands_MCRegisterSet SMC_SIslands_MCRegisterSet;
+
+struct SMC_SIslands_MCRegisters
+{
+    uint8_t                             last;
+    uint8_t                             reserved[3];
+    SMC_SIslands_MCRegisterAddress      address[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+    SMC_SIslands_MCRegisterSet          data[SMC_SISLANDS_MC_REGISTER_ARRAY_SET_COUNT];
+};
+
+typedef struct SMC_SIslands_MCRegisters SMC_SIslands_MCRegisters;
+
+struct SMC_SIslands_MCArbDramTimingRegisterSet
+{
+    uint32_t mc_arb_dram_timing;
+    uint32_t mc_arb_dram_timing2;
+    uint8_t  mc_arb_rfsh_rate;
+    uint8_t  mc_arb_burst_time;
+    uint8_t  padding[2];
+};
+
+typedef struct SMC_SIslands_MCArbDramTimingRegisterSet SMC_SIslands_MCArbDramTimingRegisterSet;
+
+struct SMC_SIslands_MCArbDramTimingRegisters
+{
+    uint8_t                                     arb_current;
+    uint8_t                                     reserved[3];
+    SMC_SIslands_MCArbDramTimingRegisterSet     data[16];
+};
+
+typedef struct SMC_SIslands_MCArbDramTimingRegisters SMC_SIslands_MCArbDramTimingRegisters;
+
+struct SMC_SISLANDS_SPLL_DIV_TABLE
+{
+    uint32_t    freq[256];
+    uint32_t    ss[256];
+};
+
+#define SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_MASK  0x01ffffff
+#define SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT 0
+#define SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_MASK   0xfe000000
+#define SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT  25
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_MASK   0x000fffff
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT  0
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_MASK   0xfff00000
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT  20
+
+typedef struct SMC_SISLANDS_SPLL_DIV_TABLE SMC_SISLANDS_SPLL_DIV_TABLE;
+
+#define SMC_SISLANDS_DTE_MAX_FILTER_STAGES 5
+
+#define SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE 16
+
+struct Smc_SIslands_DTE_Configuration
+{
+    uint32_t tau[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+    uint32_t R[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+    uint32_t K;
+    uint32_t T0;
+    uint32_t MaxT;
+    uint8_t  WindowSize;
+    uint8_t  Tdep_count;
+    uint8_t  temp_select;
+    uint8_t  DTE_mode;
+    uint8_t  T_limits[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+    uint32_t Tdep_tau[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+    uint32_t Tdep_R[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+    uint32_t Tthreshold;
+};
+
+typedef struct Smc_SIslands_DTE_Configuration Smc_SIslands_DTE_Configuration;
+
+#define SMC_SISLANDS_DTE_STATUS_FLAG_DTE_ON 1
+
+#define SISLANDS_SMC_FIRMWARE_HEADER_LOCATION 0x10000
+
+#define SISLANDS_SMC_FIRMWARE_HEADER_version                   0x0
+#define SISLANDS_SMC_FIRMWARE_HEADER_flags                     0x4
+#define SISLANDS_SMC_FIRMWARE_HEADER_softRegisters             0xC
+#define SISLANDS_SMC_FIRMWARE_HEADER_stateTable                0x10
+#define SISLANDS_SMC_FIRMWARE_HEADER_fanTable                  0x14
+#define SISLANDS_SMC_FIRMWARE_HEADER_CacConfigTable            0x18
+#define SISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable           0x24
+#define SISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable 0x30
+#define SISLANDS_SMC_FIRMWARE_HEADER_spllTable                 0x38
+#define SISLANDS_SMC_FIRMWARE_HEADER_DteConfiguration          0x40
+#define SISLANDS_SMC_FIRMWARE_HEADER_PAPMParameters            0x48
+
+#pragma pack(pop)
+
+int si_set_smc_sram_address(struct radeon_device *rdev,
+			    u32 smc_address, u32 limit);
+int si_copy_bytes_to_smc(struct radeon_device *rdev,
+			 u32 smc_start_address,
+			 const u8 *src, u32 byte_count, u32 limit);
+void si_start_smc(struct radeon_device *rdev);
+void si_reset_smc(struct radeon_device *rdev);
+int si_program_jump_on_start(struct radeon_device *rdev);
+void si_stop_smc_clock(struct radeon_device *rdev);
+void si_start_smc_clock(struct radeon_device *rdev);
+bool si_is_smc_running(struct radeon_device *rdev);
+PPSMC_Result si_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg);
+PPSMC_Result si_wait_for_smc_inactive(struct radeon_device *rdev);
+int si_load_smc_ucode(struct radeon_device *rdev, u32 limit);
+int si_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			   u32 *value, u32 limit);
+int si_write_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			    u32 value, u32 limit);
+
+#endif
+
diff --git a/drivers/gpu/drm/radeon/sumo_dpm.c b/drivers/gpu/drm/radeon/sumo_dpm.c
new file mode 100644
index 0000000..11b6b99
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumo_dpm.c
@@ -0,0 +1,1876 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "sumod.h"
+#include "r600_dpm.h"
+#include "cypress_dpm.h"
+#include "sumo_dpm.h"
+#include <linux/seq_file.h>
+
+#define SUMO_MAX_DEEPSLEEP_DIVIDER_ID 5
+#define SUMO_MINIMUM_ENGINE_CLOCK 800
+#define BOOST_DPM_LEVEL 7
+
+static const u32 sumo_utc[SUMO_PM_NUMBER_OF_TC] =
+{
+	SUMO_UTC_DFLT_00,
+	SUMO_UTC_DFLT_01,
+	SUMO_UTC_DFLT_02,
+	SUMO_UTC_DFLT_03,
+	SUMO_UTC_DFLT_04,
+	SUMO_UTC_DFLT_05,
+	SUMO_UTC_DFLT_06,
+	SUMO_UTC_DFLT_07,
+	SUMO_UTC_DFLT_08,
+	SUMO_UTC_DFLT_09,
+	SUMO_UTC_DFLT_10,
+	SUMO_UTC_DFLT_11,
+	SUMO_UTC_DFLT_12,
+	SUMO_UTC_DFLT_13,
+	SUMO_UTC_DFLT_14,
+};
+
+static const u32 sumo_dtc[SUMO_PM_NUMBER_OF_TC] =
+{
+	SUMO_DTC_DFLT_00,
+	SUMO_DTC_DFLT_01,
+	SUMO_DTC_DFLT_02,
+	SUMO_DTC_DFLT_03,
+	SUMO_DTC_DFLT_04,
+	SUMO_DTC_DFLT_05,
+	SUMO_DTC_DFLT_06,
+	SUMO_DTC_DFLT_07,
+	SUMO_DTC_DFLT_08,
+	SUMO_DTC_DFLT_09,
+	SUMO_DTC_DFLT_10,
+	SUMO_DTC_DFLT_11,
+	SUMO_DTC_DFLT_12,
+	SUMO_DTC_DFLT_13,
+	SUMO_DTC_DFLT_14,
+};
+
+struct sumo_ps *sumo_get_ps(struct radeon_ps *rps)
+{
+	struct sumo_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+struct sumo_power_info *sumo_get_pi(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void sumo_gfx_clockgating_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+#define CGCG_CGTT_LOCAL0_MASK 0xE5BFFFFF
+#define CGCG_CGTT_LOCAL1_MASK 0xEFFF07FF
+
+static void sumo_mg_clockgating_enable(struct radeon_device *rdev, bool enable)
+{
+	u32 local0;
+	u32 local1;
+
+	local0 = RREG32(CG_CGTT_LOCAL_0);
+	local1 = RREG32(CG_CGTT_LOCAL_1);
+
+	if (enable) {
+		WREG32(CG_CGTT_LOCAL_0, (0 & CGCG_CGTT_LOCAL0_MASK) | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32(CG_CGTT_LOCAL_1, (0 & CGCG_CGTT_LOCAL1_MASK) | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+	} else {
+		WREG32(CG_CGTT_LOCAL_0, (0xFFFFFFFF & CGCG_CGTT_LOCAL0_MASK) | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32(CG_CGTT_LOCAL_1, (0xFFFFCFFF & CGCG_CGTT_LOCAL1_MASK) | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+	}
+}
+
+static void sumo_program_git(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(SUMO_GICST_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32_P(CG_GIT, CG_GICST(p), ~CG_GICST_MASK);
+}
+
+static void sumo_program_grsd(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 grs = 256 * 25 / 100;
+
+	r600_calculate_u_and_p(1, xclk, 14, &p, &u);
+
+	WREG32(CG_GCOOR, PHC(grs) | SDC(p) | SU(u));
+}
+
+void sumo_gfx_clockgating_initialize(struct radeon_device *rdev)
+{
+	sumo_program_git(rdev);
+	sumo_program_grsd(rdev);
+}
+
+static void sumo_gfx_powergating_initialize(struct radeon_device *rdev)
+{
+	u32 rcu_pwr_gating_cntl;
+	u32 p, u;
+	u32 p_c, p_p, d_p;
+	u32 r_t, i_t;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	if (rdev->family == CHIP_PALM) {
+		p_c = 4;
+		d_p = 10;
+		r_t = 10;
+		i_t = 4;
+		p_p = 50 + 1000/200 + 6 * 32;
+	} else {
+		p_c = 16;
+		d_p = 50;
+		r_t = 50;
+		i_t  = 50;
+		p_p = 113;
+	}
+
+	WREG32(CG_SCRATCH2, 0x01B60A17);
+
+	r600_calculate_u_and_p(SUMO_GFXPOWERGATINGT_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32_P(CG_PWR_GATING_CNTL, PGP(p) | PGU(u),
+		 ~(PGP_MASK | PGU_MASK));
+
+	r600_calculate_u_and_p(SUMO_VOLTAGEDROPT_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32_P(CG_CG_VOLTAGE_CNTL, PGP(p) | PGU(u),
+		 ~(PGP_MASK | PGU_MASK));
+
+	if (rdev->family == CHIP_PALM) {
+		WREG32_RCU(RCU_PWR_GATING_SEQ0, 0x10103210);
+		WREG32_RCU(RCU_PWR_GATING_SEQ1, 0x10101010);
+	} else {
+		WREG32_RCU(RCU_PWR_GATING_SEQ0, 0x76543210);
+		WREG32_RCU(RCU_PWR_GATING_SEQ1, 0xFEDCBA98);
+	}
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL);
+	rcu_pwr_gating_cntl &=
+		~(RSVD_MASK | PCV_MASK | PGS_MASK);
+	rcu_pwr_gating_cntl |= PCV(p_c) | PGS(1) | PWR_GATING_EN;
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl &= ~PCP_MASK;
+		rcu_pwr_gating_cntl |= PCP(0x77);
+	}
+	WREG32_RCU(RCU_PWR_GATING_CNTL, rcu_pwr_gating_cntl);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_2);
+	rcu_pwr_gating_cntl &= ~(MPPU_MASK | MPPD_MASK);
+	rcu_pwr_gating_cntl |= MPPU(p_p) | MPPD(50);
+	WREG32_RCU(RCU_PWR_GATING_CNTL_2, rcu_pwr_gating_cntl);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_3);
+	rcu_pwr_gating_cntl &= ~(DPPU_MASK | DPPD_MASK);
+	rcu_pwr_gating_cntl |= DPPU(d_p) | DPPD(50);
+	WREG32_RCU(RCU_PWR_GATING_CNTL_3, rcu_pwr_gating_cntl);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_4);
+	rcu_pwr_gating_cntl &= ~(RT_MASK | IT_MASK);
+	rcu_pwr_gating_cntl |= RT(r_t) | IT(i_t);
+	WREG32_RCU(RCU_PWR_GATING_CNTL_4, rcu_pwr_gating_cntl);
+
+	if (rdev->family == CHIP_PALM)
+		WREG32_RCU(RCU_PWR_GATING_CNTL_5, 0xA02);
+
+	sumo_smu_pg_init(rdev);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL);
+	rcu_pwr_gating_cntl &=
+		~(RSVD_MASK | PCV_MASK | PGS_MASK);
+	rcu_pwr_gating_cntl |= PCV(p_c) | PGS(4) | PWR_GATING_EN;
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl &= ~PCP_MASK;
+		rcu_pwr_gating_cntl |= PCP(0x77);
+	}
+	WREG32_RCU(RCU_PWR_GATING_CNTL, rcu_pwr_gating_cntl);
+
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_2);
+		rcu_pwr_gating_cntl &= ~(MPPU_MASK | MPPD_MASK);
+		rcu_pwr_gating_cntl |= MPPU(113) | MPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_2, rcu_pwr_gating_cntl);
+
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_3);
+		rcu_pwr_gating_cntl &= ~(DPPU_MASK | DPPD_MASK);
+		rcu_pwr_gating_cntl |= DPPU(16) | DPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_3, rcu_pwr_gating_cntl);
+	}
+
+	sumo_smu_pg_init(rdev);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL);
+	rcu_pwr_gating_cntl &=
+		~(RSVD_MASK | PCV_MASK | PGS_MASK);
+	rcu_pwr_gating_cntl |= PGS(5) | PWR_GATING_EN;
+
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl |= PCV(4);
+		rcu_pwr_gating_cntl &= ~PCP_MASK;
+		rcu_pwr_gating_cntl |= PCP(0x77);
+	} else
+		rcu_pwr_gating_cntl |= PCV(11);
+	WREG32_RCU(RCU_PWR_GATING_CNTL, rcu_pwr_gating_cntl);
+
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_2);
+		rcu_pwr_gating_cntl &= ~(MPPU_MASK | MPPD_MASK);
+		rcu_pwr_gating_cntl |= MPPU(113) | MPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_2, rcu_pwr_gating_cntl);
+
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_3);
+		rcu_pwr_gating_cntl &= ~(DPPU_MASK | DPPD_MASK);
+		rcu_pwr_gating_cntl |= DPPU(22) | DPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_3, rcu_pwr_gating_cntl);
+	}
+
+	sumo_smu_pg_init(rdev);
+}
+
+static void sumo_gfx_powergating_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_PWR_GATING_CNTL, DYN_PWR_DOWN_EN, ~DYN_PWR_DOWN_EN);
+	else {
+		WREG32_P(CG_PWR_GATING_CNTL, 0, ~DYN_PWR_DOWN_EN);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+static int sumo_enable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_initialize(rdev);
+	if (pi->enable_gfx_power_gating)
+		sumo_gfx_powergating_initialize(rdev);
+	if (pi->enable_mg_clock_gating)
+		sumo_mg_clockgating_enable(rdev, true);
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_enable(rdev, true);
+	if (pi->enable_gfx_power_gating)
+		sumo_gfx_powergating_enable(rdev, true);
+
+	return 0;
+}
+
+static void sumo_disable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_enable(rdev, false);
+	if (pi->enable_gfx_power_gating)
+		sumo_gfx_powergating_enable(rdev, false);
+	if (pi->enable_mg_clock_gating)
+		sumo_mg_clockgating_enable(rdev, false);
+}
+
+static void sumo_calculate_bsp(struct radeon_device *rdev,
+			       u32 high_clk)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 xclk = radeon_get_xclk(rdev);
+
+	pi->pasi = 65535 * 100 / high_clk;
+	pi->asi = 65535 * 100 / high_clk;
+
+	r600_calculate_u_and_p(pi->asi,
+			       xclk, 16, &pi->bsp, &pi->bsu);
+
+	r600_calculate_u_and_p(pi->pasi,
+			       xclk, 16, &pi->pbsp, &pi->pbsu);
+
+	pi->dsp = BSP(pi->bsp) | BSU(pi->bsu);
+	pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu);
+}
+
+static void sumo_init_bsp(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	WREG32(CG_BSP_0, pi->psp);
+}
+
+
+static void sumo_program_bsp(struct radeon_device *rdev,
+			     struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	u32 i;
+	u32 highest_engine_clock = ps->levels[ps->num_levels - 1].sclk;
+
+	if (ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		highest_engine_clock = pi->boost_pl.sclk;
+
+	sumo_calculate_bsp(rdev, highest_engine_clock);
+
+	for (i = 0; i < ps->num_levels - 1; i++)
+		WREG32(CG_BSP_0 + (i * 4), pi->dsp);
+
+	WREG32(CG_BSP_0 + (i * 4), pi->psp);
+
+	if (ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		WREG32(CG_BSP_0 + (BOOST_DPM_LEVEL * 4), pi->psp);
+}
+
+static void sumo_write_at(struct radeon_device *rdev,
+			  u32 index, u32 value)
+{
+	if (index == 0)
+		WREG32(CG_AT_0, value);
+	else if (index == 1)
+		WREG32(CG_AT_1, value);
+	else if (index == 2)
+		WREG32(CG_AT_2, value);
+	else if (index == 3)
+		WREG32(CG_AT_3, value);
+	else if (index == 4)
+		WREG32(CG_AT_4, value);
+	else if (index == 5)
+		WREG32(CG_AT_5, value);
+	else if (index == 6)
+		WREG32(CG_AT_6, value);
+	else if (index == 7)
+		WREG32(CG_AT_7, value);
+}
+
+static void sumo_program_at(struct radeon_device *rdev,
+			    struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	u32 asi;
+	u32 i;
+	u32 m_a;
+	u32 a_t;
+	u32 r[SUMO_MAX_HARDWARE_POWERLEVELS];
+	u32 l[SUMO_MAX_HARDWARE_POWERLEVELS];
+
+	r[0] = SUMO_R_DFLT0;
+	r[1] = SUMO_R_DFLT1;
+	r[2] = SUMO_R_DFLT2;
+	r[3] = SUMO_R_DFLT3;
+	r[4] = SUMO_R_DFLT4;
+
+	l[0] = SUMO_L_DFLT0;
+	l[1] = SUMO_L_DFLT1;
+	l[2] = SUMO_L_DFLT2;
+	l[3] = SUMO_L_DFLT3;
+	l[4] = SUMO_L_DFLT4;
+
+	for (i = 0; i < ps->num_levels; i++) {
+		asi = (i == ps->num_levels - 1) ? pi->pasi : pi->asi;
+
+		m_a = asi * ps->levels[i].sclk / 100;
+
+		a_t = CG_R(m_a * r[i] / 100) | CG_L(m_a * l[i] / 100);
+
+		sumo_write_at(rdev, i, a_t);
+	}
+
+	if (ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE) {
+		asi = pi->pasi;
+
+		m_a = asi * pi->boost_pl.sclk / 100;
+
+		a_t = CG_R(m_a * r[ps->num_levels - 1] / 100) |
+			CG_L(m_a * l[ps->num_levels - 1] / 100);
+
+		sumo_write_at(rdev, BOOST_DPM_LEVEL, a_t);
+	}
+}
+
+static void sumo_program_tp(struct radeon_device *rdev)
+{
+	int i;
+	enum r600_td td = R600_TD_DFLT;
+
+	for (i = 0; i < SUMO_PM_NUMBER_OF_TC; i++) {
+		WREG32_P(CG_FFCT_0 + (i * 4), UTC_0(sumo_utc[i]), ~UTC_0_MASK);
+		WREG32_P(CG_FFCT_0 + (i * 4), DTC_0(sumo_dtc[i]), ~DTC_0_MASK);
+	}
+
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+void sumo_program_vc(struct radeon_device *rdev, u32 vrc)
+{
+	WREG32(CG_FTV, vrc);
+}
+
+void sumo_clear_vc(struct radeon_device *rdev)
+{
+	WREG32(CG_FTV, 0);
+}
+
+void sumo_program_sstp(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(SUMO_SST_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32(CG_SSP, SSTU(u) | SST(p));
+}
+
+static void sumo_set_divider_value(struct radeon_device *rdev,
+				   u32 index, u32 divider)
+{
+	u32 reg_index = index / 4;
+	u32 field_index = index % 4;
+
+	if (field_index == 0)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_0_DIV(divider), ~SCLK_FSTATE_0_DIV_MASK);
+	else if (field_index == 1)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_1_DIV(divider), ~SCLK_FSTATE_1_DIV_MASK);
+	else if (field_index == 2)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_2_DIV(divider), ~SCLK_FSTATE_2_DIV_MASK);
+	else if (field_index == 3)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_3_DIV(divider), ~SCLK_FSTATE_3_DIV_MASK);
+}
+
+static void sumo_set_ds_dividers(struct radeon_device *rdev,
+				 u32 index, u32 divider)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_sclk_ds) {
+		u32 dpm_ctrl = RREG32(CG_SCLK_DPM_CTRL_6);
+
+		dpm_ctrl &= ~(0x7 << (index * 3));
+		dpm_ctrl |= (divider << (index * 3));
+		WREG32(CG_SCLK_DPM_CTRL_6, dpm_ctrl);
+	}
+}
+
+static void sumo_set_ss_dividers(struct radeon_device *rdev,
+				 u32 index, u32 divider)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_sclk_ds) {
+		u32 dpm_ctrl = RREG32(CG_SCLK_DPM_CTRL_11);
+
+		dpm_ctrl &= ~(0x7 << (index * 3));
+		dpm_ctrl |= (divider << (index * 3));
+		WREG32(CG_SCLK_DPM_CTRL_11, dpm_ctrl);
+	}
+}
+
+static void sumo_set_vid(struct radeon_device *rdev, u32 index, u32 vid)
+{
+	u32 voltage_cntl = RREG32(CG_DPM_VOLTAGE_CNTL);
+
+	voltage_cntl &= ~(DPM_STATE0_LEVEL_MASK << (index * 2));
+	voltage_cntl |= (vid << (DPM_STATE0_LEVEL_SHIFT + index * 2));
+	WREG32(CG_DPM_VOLTAGE_CNTL, voltage_cntl);
+}
+
+static void sumo_set_allos_gnb_slow(struct radeon_device *rdev, u32 index, u32 gnb_slow)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 temp = gnb_slow;
+	u32 cg_sclk_dpm_ctrl_3;
+
+	if (pi->driver_nbps_policy_disable)
+		temp = 1;
+
+	cg_sclk_dpm_ctrl_3 = RREG32(CG_SCLK_DPM_CTRL_3);
+	cg_sclk_dpm_ctrl_3 &= ~(GNB_SLOW_FSTATE_0_MASK << index);
+	cg_sclk_dpm_ctrl_3 |= (temp << (GNB_SLOW_FSTATE_0_SHIFT + index));
+
+	WREG32(CG_SCLK_DPM_CTRL_3, cg_sclk_dpm_ctrl_3);
+}
+
+static void sumo_program_power_level(struct radeon_device *rdev,
+				     struct sumo_pl *pl, u32 index)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	int ret;
+	struct atom_clock_dividers dividers;
+	u32 ds_en = RREG32(DEEP_SLEEP_CNTL) & ENABLE_DS;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     pl->sclk, false, &dividers);
+	if (ret)
+		return;
+
+	sumo_set_divider_value(rdev, index, dividers.post_div);
+
+	sumo_set_vid(rdev, index, pl->vddc_index);
+
+	if (pl->ss_divider_index == 0 || pl->ds_divider_index == 0) {
+		if (ds_en)
+			WREG32_P(DEEP_SLEEP_CNTL, 0, ~ENABLE_DS);
+	} else {
+		sumo_set_ss_dividers(rdev, index, pl->ss_divider_index);
+		sumo_set_ds_dividers(rdev, index, pl->ds_divider_index);
+
+		if (!ds_en)
+			WREG32_P(DEEP_SLEEP_CNTL, ENABLE_DS, ~ENABLE_DS);
+	}
+
+	sumo_set_allos_gnb_slow(rdev, index, pl->allow_gnb_slow);
+
+	if (pi->enable_boost)
+		sumo_set_tdp_limit(rdev, index, pl->sclk_dpm_tdp_limit);
+}
+
+static void sumo_power_level_enable(struct radeon_device *rdev, u32 index, bool enable)
+{
+	u32 reg_index = index / 4;
+	u32 field_index = index % 4;
+
+	if (field_index == 0)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_0_VLD : 0, ~SCLK_FSTATE_0_VLD);
+	else if (field_index == 1)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_1_VLD : 0, ~SCLK_FSTATE_1_VLD);
+	else if (field_index == 2)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_2_VLD : 0, ~SCLK_FSTATE_2_VLD);
+	else if (field_index == 3)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_3_VLD : 0, ~SCLK_FSTATE_3_VLD);
+}
+
+static bool sumo_dpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32(CG_SCLK_DPM_CTRL_3) & DPM_SCLK_ENABLE)
+		return true;
+	else
+		return false;
+}
+
+static void sumo_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, DPM_SCLK_ENABLE, ~DPM_SCLK_ENABLE);
+}
+
+static void sumo_stop_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, 0, ~DPM_SCLK_ENABLE);
+}
+
+static void sumo_set_forced_mode(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_SCLK_DPM_CTRL_3, FORCE_SCLK_STATE_EN, ~FORCE_SCLK_STATE_EN);
+	else
+		WREG32_P(CG_SCLK_DPM_CTRL_3, 0, ~FORCE_SCLK_STATE_EN);
+}
+
+static void sumo_set_forced_mode_enabled(struct radeon_device *rdev)
+{
+	int i;
+
+	sumo_set_forced_mode(rdev, true);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(CG_SCLK_STATUS) & SCLK_OVERCLK_DETECT)
+			break;
+		udelay(1);
+	}
+}
+
+static void sumo_wait_for_level_0(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_SCLK_INDEX_MASK) == 0)
+			break;
+		udelay(1);
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_INDEX_MASK) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void sumo_set_forced_mode_disabled(struct radeon_device *rdev)
+{
+	sumo_set_forced_mode(rdev, false);
+}
+
+static void sumo_enable_power_level_0(struct radeon_device *rdev)
+{
+	sumo_power_level_enable(rdev, 0, true);
+}
+
+static void sumo_patch_boost_state(struct radeon_device *rdev,
+				   struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+
+	if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE) {
+		pi->boost_pl = new_ps->levels[new_ps->num_levels - 1];
+		pi->boost_pl.sclk = pi->sys_info.boost_sclk;
+		pi->boost_pl.vddc_index = pi->sys_info.boost_vid_2bit;
+		pi->boost_pl.sclk_dpm_tdp_limit = pi->sys_info.sclk_dpm_tdp_limit_boost;
+	}
+}
+
+static void sumo_pre_notify_alt_vddnb_change(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *old_ps = sumo_get_ps(old_rps);
+	u32 nbps1_old = 0;
+	u32 nbps1_new = 0;
+
+	if (old_ps != NULL)
+		nbps1_old = (old_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE) ? 1 : 0;
+
+	nbps1_new = (new_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE) ? 1 : 0;
+
+	if (nbps1_old == 1 && nbps1_new == 0)
+		sumo_smu_notify_alt_vddnb_change(rdev, 0, 0);
+}
+
+static void sumo_post_notify_alt_vddnb_change(struct radeon_device *rdev,
+					      struct radeon_ps *new_rps,
+					      struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *old_ps = sumo_get_ps(old_rps);
+	u32 nbps1_old = 0;
+	u32 nbps1_new = 0;
+
+	if (old_ps != NULL)
+		nbps1_old = (old_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)? 1 : 0;
+
+	nbps1_new = (new_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)? 1 : 0;
+
+	if (nbps1_old == 0 && nbps1_new == 1)
+		sumo_smu_notify_alt_vddnb_change(rdev, 1, 1);
+}
+
+static void sumo_enable_boost(struct radeon_device *rdev,
+			      struct radeon_ps *rps,
+			      bool enable)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+
+	if (enable) {
+		if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+			sumo_boost_state_enable(rdev, true);
+	} else
+		sumo_boost_state_enable(rdev, false);
+}
+
+static void sumo_set_forced_level(struct radeon_device *rdev, u32 index)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, FORCE_SCLK_STATE(index), ~FORCE_SCLK_STATE_MASK);
+}
+
+static void sumo_set_forced_level_0(struct radeon_device *rdev)
+{
+	sumo_set_forced_level(rdev, 0);
+}
+
+static void sumo_program_wl(struct radeon_device *rdev,
+			    struct radeon_ps *rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+	u32 dpm_ctrl4 = RREG32(CG_SCLK_DPM_CTRL_4);
+
+	dpm_ctrl4 &= 0xFFFFFF00;
+	dpm_ctrl4 |= (1 << (new_ps->num_levels - 1));
+
+	if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		dpm_ctrl4 |= (1 << BOOST_DPM_LEVEL);
+
+	WREG32(CG_SCLK_DPM_CTRL_4, dpm_ctrl4);
+}
+
+static void sumo_program_power_levels_0_to_n(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *old_ps = sumo_get_ps(old_rps);
+	u32 i;
+	u32 n_current_state_levels = (old_ps == NULL) ? 1 : old_ps->num_levels;
+
+	for (i = 0; i < new_ps->num_levels; i++) {
+		sumo_program_power_level(rdev, &new_ps->levels[i], i);
+		sumo_power_level_enable(rdev, i, true);
+	}
+
+	for (i = new_ps->num_levels; i < n_current_state_levels; i++)
+		sumo_power_level_enable(rdev, i, false);
+
+	if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		sumo_program_power_level(rdev, &pi->boost_pl, BOOST_DPM_LEVEL);
+}
+
+static void sumo_enable_acpi_pm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+static void sumo_program_power_level_enter_state(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_5, SCLK_FSTATE_BOOTUP(0), ~SCLK_FSTATE_BOOTUP_MASK);
+}
+
+static void sumo_program_acpi_power_level(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	int ret;
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             pi->acpi_pl.sclk,
+					     false, &dividers);
+	if (ret)
+		return;
+
+	WREG32_P(CG_ACPI_CNTL, SCLK_ACPI_DIV(dividers.post_div), ~SCLK_ACPI_DIV_MASK);
+	WREG32_P(CG_ACPI_VOLTAGE_CNTL, 0, ~ACPI_VOLTAGE_EN);
+}
+
+static void sumo_program_bootup_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 dpm_ctrl4 = RREG32(CG_SCLK_DPM_CTRL_4);
+	u32 i;
+
+	sumo_program_power_level(rdev, &pi->boot_pl, 0);
+
+	dpm_ctrl4 &= 0xFFFFFF00;
+	WREG32(CG_SCLK_DPM_CTRL_4, dpm_ctrl4);
+
+	for (i = 1; i < 8; i++)
+		sumo_power_level_enable(rdev, i, false);
+}
+
+static void sumo_setup_uvd_clocks(struct radeon_device *rdev,
+				  struct radeon_ps *new_rps,
+				  struct radeon_ps *old_rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_gfx_power_gating) {
+		sumo_gfx_powergating_enable(rdev, false);
+	}
+
+	radeon_set_uvd_clocks(rdev, new_rps->vclk, new_rps->dclk);
+
+	if (pi->enable_gfx_power_gating) {
+		if (!pi->disable_gfx_power_gating_in_uvd ||
+		    !r600_is_uvd_state(new_rps->class, new_rps->class2))
+			sumo_gfx_powergating_enable(rdev, true);
+	}
+}
+
+static void sumo_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						    struct radeon_ps *new_rps,
+						    struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *current_ps = sumo_get_ps(old_rps);
+
+	if ((new_rps->vclk == old_rps->vclk) &&
+	    (new_rps->dclk == old_rps->dclk))
+		return;
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk >=
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	sumo_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+static void sumo_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						   struct radeon_ps *new_rps,
+						   struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *current_ps = sumo_get_ps(old_rps);
+
+	if ((new_rps->vclk == old_rps->vclk) &&
+	    (new_rps->dclk == old_rps->dclk))
+		return;
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk <
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	sumo_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+void sumo_take_smu_control(struct radeon_device *rdev, bool enable)
+{
+/* This bit selects who handles display phy powergating.
+ * Clear the bit to let atom handle it.
+ * Set it to let the driver handle it.
+ * For now we just let atom handle it.
+ */
+#if 0
+	u32 v = RREG32(DOUT_SCRATCH3);
+
+	if (enable)
+		v |= 0x4;
+	else
+		v &= 0xFFFFFFFB;
+
+	WREG32(DOUT_SCRATCH3, v);
+#endif
+}
+
+static void sumo_enable_sclk_ds(struct radeon_device *rdev, bool enable)
+{
+	if (enable) {
+		u32 deep_sleep_cntl = RREG32(DEEP_SLEEP_CNTL);
+		u32 deep_sleep_cntl2 = RREG32(DEEP_SLEEP_CNTL2);
+		u32 t = 1;
+
+		deep_sleep_cntl &= ~R_DIS;
+		deep_sleep_cntl &= ~HS_MASK;
+		deep_sleep_cntl |= HS(t > 4095 ? 4095 : t);
+
+		deep_sleep_cntl2 |= LB_UFP_EN;
+		deep_sleep_cntl2 &= INOUT_C_MASK;
+		deep_sleep_cntl2 |= INOUT_C(0xf);
+
+		WREG32(DEEP_SLEEP_CNTL2, deep_sleep_cntl2);
+		WREG32(DEEP_SLEEP_CNTL, deep_sleep_cntl);
+	} else
+		WREG32_P(DEEP_SLEEP_CNTL, 0, ~ENABLE_DS);
+}
+
+static void sumo_program_bootup_at(struct radeon_device *rdev)
+{
+	WREG32_P(CG_AT_0, CG_R(0xffff), ~CG_R_MASK);
+	WREG32_P(CG_AT_0, CG_L(0), ~CG_L_MASK);
+}
+
+static void sumo_reset_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, FIR_RESET, ~FIR_RESET);
+}
+
+static void sumo_start_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_RESET);
+}
+
+static void sumo_program_ttp(struct radeon_device *rdev)
+{
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 p, u;
+	u32 cg_sclk_dpm_ctrl_5 = RREG32(CG_SCLK_DPM_CTRL_5);
+
+	r600_calculate_u_and_p(1000,
+			       xclk, 16, &p, &u);
+
+	cg_sclk_dpm_ctrl_5 &= ~(TT_TP_MASK | TT_TU_MASK);
+	cg_sclk_dpm_ctrl_5 |= TT_TP(p) | TT_TU(u);
+
+	WREG32(CG_SCLK_DPM_CTRL_5, cg_sclk_dpm_ctrl_5);
+}
+
+static void sumo_program_ttt(struct radeon_device *rdev)
+{
+	u32 cg_sclk_dpm_ctrl_3 = RREG32(CG_SCLK_DPM_CTRL_3);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	cg_sclk_dpm_ctrl_3 &= ~(GNB_TT_MASK | GNB_THERMTHRO_MASK);
+	cg_sclk_dpm_ctrl_3 |= GNB_TT(pi->thermal_auto_throttling + 49);
+
+	WREG32(CG_SCLK_DPM_CTRL_3, cg_sclk_dpm_ctrl_3);
+}
+
+
+static void sumo_enable_voltage_scaling(struct radeon_device *rdev, bool enable)
+{
+	if (enable) {
+		WREG32_P(CG_DPM_VOLTAGE_CNTL, DPM_VOLTAGE_EN, ~DPM_VOLTAGE_EN);
+		WREG32_P(CG_CG_VOLTAGE_CNTL, 0, ~CG_VOLTAGE_EN);
+	} else {
+		WREG32_P(CG_CG_VOLTAGE_CNTL, CG_VOLTAGE_EN, ~CG_VOLTAGE_EN);
+		WREG32_P(CG_DPM_VOLTAGE_CNTL, 0, ~DPM_VOLTAGE_EN);
+	}
+}
+
+static void sumo_override_cnb_thermal_events(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, CNB_THERMTHRO_MASK_SCLK,
+		 ~CNB_THERMTHRO_MASK_SCLK);
+}
+
+static void sumo_program_dc_hto(struct radeon_device *rdev)
+{
+	u32 cg_sclk_dpm_ctrl_4 = RREG32(CG_SCLK_DPM_CTRL_4);
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(100000,
+			       xclk, 14, &p, &u);
+
+	cg_sclk_dpm_ctrl_4 &= ~(DC_HDC_MASK | DC_HU_MASK);
+	cg_sclk_dpm_ctrl_4 |= DC_HDC(p) | DC_HU(u);
+
+	WREG32(CG_SCLK_DPM_CTRL_4, cg_sclk_dpm_ctrl_4);
+}
+
+static void sumo_force_nbp_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+
+	if (!pi->driver_nbps_policy_disable) {
+		if (new_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)
+			WREG32_P(CG_SCLK_DPM_CTRL_3, FORCE_NB_PSTATE_1, ~FORCE_NB_PSTATE_1);
+		else
+			WREG32_P(CG_SCLK_DPM_CTRL_3, 0, ~FORCE_NB_PSTATE_1);
+	}
+}
+
+u32 sumo_get_sleep_divider_from_id(u32 id)
+{
+	return 1 << id;
+}
+
+u32 sumo_get_sleep_divider_id_from_clock(struct radeon_device *rdev,
+					 u32 sclk,
+					 u32 min_sclk_in_sr)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 i;
+	u32 temp;
+	u32 min = (min_sclk_in_sr > SUMO_MINIMUM_ENGINE_CLOCK) ?
+		min_sclk_in_sr : SUMO_MINIMUM_ENGINE_CLOCK;
+
+	if (sclk < min)
+		return 0;
+
+	if (!pi->enable_sclk_ds)
+		return 0;
+
+	for (i = SUMO_MAX_DEEPSLEEP_DIVIDER_ID;  ; i--) {
+		temp = sclk / sumo_get_sleep_divider_from_id(i);
+
+		if (temp >= min || i == 0)
+			break;
+	}
+	return i;
+}
+
+static u32 sumo_get_valid_engine_clock(struct radeon_device *rdev,
+				       u32 lower_limit)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 i;
+
+	for (i = 0; i < pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries; i++) {
+		if (pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency >= lower_limit)
+			return pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency;
+	}
+
+	return pi->sys_info.sclk_voltage_mapping_table.entries[pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries - 1].sclk_frequency;
+}
+
+static void sumo_patch_thermal_state(struct radeon_device *rdev,
+				     struct sumo_ps *ps,
+				     struct sumo_ps *current_ps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 current_vddc;
+	u32 current_sclk;
+	u32 current_index = 0;
+
+	if (current_ps) {
+		current_vddc = current_ps->levels[current_index].vddc_index;
+		current_sclk = current_ps->levels[current_index].sclk;
+	} else {
+		current_vddc = pi->boot_pl.vddc_index;
+		current_sclk = pi->boot_pl.sclk;
+	}
+
+	ps->levels[0].vddc_index = current_vddc;
+
+	if (ps->levels[0].sclk > current_sclk)
+		ps->levels[0].sclk = current_sclk;
+
+	ps->levels[0].ss_divider_index =
+		sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[0].sclk, sclk_in_sr);
+
+	ps->levels[0].ds_divider_index =
+		sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[0].sclk, SUMO_MINIMUM_ENGINE_CLOCK);
+
+	if (ps->levels[0].ds_divider_index > ps->levels[0].ss_divider_index + 1)
+		ps->levels[0].ds_divider_index = ps->levels[0].ss_divider_index + 1;
+
+	if (ps->levels[0].ss_divider_index == ps->levels[0].ds_divider_index) {
+		if (ps->levels[0].ss_divider_index > 1)
+			ps->levels[0].ss_divider_index = ps->levels[0].ss_divider_index - 1;
+	}
+
+	if (ps->levels[0].ss_divider_index == 0)
+		ps->levels[0].ds_divider_index = 0;
+
+	if (ps->levels[0].ds_divider_index == 0)
+		ps->levels[0].ss_divider_index = 0;
+}
+
+static void sumo_apply_state_adjust_rules(struct radeon_device *rdev,
+					  struct radeon_ps *new_rps,
+					  struct radeon_ps *old_rps)
+{
+	struct sumo_ps *ps = sumo_get_ps(new_rps);
+	struct sumo_ps *current_ps = sumo_get_ps(old_rps);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 min_voltage = 0; /* ??? */
+	u32 min_sclk = pi->sys_info.min_sclk; /* XXX check against disp reqs */
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 i;
+
+	if (new_rps->class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+		return sumo_patch_thermal_state(rdev, ps, current_ps);
+
+	if (pi->enable_boost) {
+		if (new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE)
+			ps->flags |= SUMO_POWERSTATE_FLAGS_BOOST_STATE;
+	}
+
+	if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) ||
+	    (new_rps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE) ||
+	    (new_rps->class & ATOM_PPLIB_CLASSIFICATION_HDSTATE))
+		ps->flags |= SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE;
+
+	for (i = 0; i < ps->num_levels; i++) {
+		if (ps->levels[i].vddc_index < min_voltage)
+			ps->levels[i].vddc_index = min_voltage;
+
+		if (ps->levels[i].sclk < min_sclk)
+			ps->levels[i].sclk =
+				sumo_get_valid_engine_clock(rdev, min_sclk);
+
+		ps->levels[i].ss_divider_index =
+			sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[i].sclk, sclk_in_sr);
+
+		ps->levels[i].ds_divider_index =
+			sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[i].sclk, SUMO_MINIMUM_ENGINE_CLOCK);
+
+		if (ps->levels[i].ds_divider_index > ps->levels[i].ss_divider_index + 1)
+			ps->levels[i].ds_divider_index = ps->levels[i].ss_divider_index + 1;
+
+		if (ps->levels[i].ss_divider_index == ps->levels[i].ds_divider_index) {
+			if (ps->levels[i].ss_divider_index > 1)
+				ps->levels[i].ss_divider_index = ps->levels[i].ss_divider_index - 1;
+		}
+
+		if (ps->levels[i].ss_divider_index == 0)
+			ps->levels[i].ds_divider_index = 0;
+
+		if (ps->levels[i].ds_divider_index == 0)
+			ps->levels[i].ss_divider_index = 0;
+
+		if (ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)
+			ps->levels[i].allow_gnb_slow = 1;
+		else if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) ||
+			 (new_rps->class2 & ATOM_PPLIB_CLASSIFICATION2_MVC))
+			ps->levels[i].allow_gnb_slow = 0;
+		else if (i == ps->num_levels - 1)
+			ps->levels[i].allow_gnb_slow = 0;
+		else
+			ps->levels[i].allow_gnb_slow = 1;
+	}
+}
+
+static void sumo_cleanup_asic(struct radeon_device *rdev)
+{
+	sumo_take_smu_control(rdev, false);
+}
+
+static int sumo_set_thermal_temperature_range(struct radeon_device *rdev,
+					      int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(49 + (high_temp / 1000)), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(49 + (low_temp / 1000)), ~DIG_THERM_INTL_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+static void sumo_update_current_ps(struct radeon_device *rdev,
+				   struct radeon_ps *rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	pi->current_rps = *rps;
+	pi->current_ps = *new_ps;
+	pi->current_rps.ps_priv = &pi->current_ps;
+}
+
+static void sumo_update_requested_ps(struct radeon_device *rdev,
+				     struct radeon_ps *rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	pi->requested_rps = *rps;
+	pi->requested_ps = *new_ps;
+	pi->requested_rps.ps_priv = &pi->requested_ps;
+}
+
+int sumo_dpm_enable(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	int ret;
+
+	if (sumo_dpm_enabled(rdev))
+		return -EINVAL;
+
+	ret = sumo_enable_clock_power_gating(rdev);
+	if (ret)
+		return ret;
+	sumo_program_bootup_state(rdev);
+	sumo_init_bsp(rdev);
+	sumo_reset_am(rdev);
+	sumo_program_tp(rdev);
+	sumo_program_bootup_at(rdev);
+	sumo_start_am(rdev);
+	if (pi->enable_auto_thermal_throttling) {
+		sumo_program_ttp(rdev);
+		sumo_program_ttt(rdev);
+	}
+	sumo_program_dc_hto(rdev);
+	sumo_program_power_level_enter_state(rdev);
+	sumo_enable_voltage_scaling(rdev, true);
+	sumo_program_sstp(rdev);
+	sumo_program_vc(rdev, SUMO_VRC_DFLT);
+	sumo_override_cnb_thermal_events(rdev);
+	sumo_start_dpm(rdev);
+	sumo_wait_for_level_0(rdev);
+	if (pi->enable_sclk_ds)
+		sumo_enable_sclk_ds(rdev, true);
+	if (pi->enable_boost)
+		sumo_enable_boost_timer(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		ret = sumo_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	sumo_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+
+	return 0;
+}
+
+void sumo_dpm_disable(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (!sumo_dpm_enabled(rdev))
+		return;
+	sumo_disable_clock_power_gating(rdev);
+	if (pi->enable_sclk_ds)
+		sumo_enable_sclk_ds(rdev, false);
+	sumo_clear_vc(rdev);
+	sumo_wait_for_level_0(rdev);
+	sumo_stop_dpm(rdev);
+	sumo_enable_voltage_scaling(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	sumo_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+}
+
+int sumo_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	sumo_update_requested_ps(rdev, new_ps);
+
+	if (pi->enable_dynamic_patch_ps)
+		sumo_apply_state_adjust_rules(rdev,
+					      &pi->requested_rps,
+					      &pi->current_rps);
+
+	return 0;
+}
+
+int sumo_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+	struct radeon_ps *old_ps = &pi->current_rps;
+
+	if (pi->enable_dpm)
+		sumo_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	if (pi->enable_boost) {
+		sumo_enable_boost(rdev, new_ps, false);
+		sumo_patch_boost_state(rdev, new_ps);
+	}
+	if (pi->enable_dpm) {
+		sumo_pre_notify_alt_vddnb_change(rdev, new_ps, old_ps);
+		sumo_enable_power_level_0(rdev);
+		sumo_set_forced_level_0(rdev);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_wait_for_level_0(rdev);
+		sumo_program_power_levels_0_to_n(rdev, new_ps, old_ps);
+		sumo_program_wl(rdev, new_ps);
+		sumo_program_bsp(rdev, new_ps);
+		sumo_program_at(rdev, new_ps);
+		sumo_force_nbp_state(rdev, new_ps);
+		sumo_set_forced_mode_disabled(rdev);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_set_forced_mode_disabled(rdev);
+		sumo_post_notify_alt_vddnb_change(rdev, new_ps, old_ps);
+	}
+	if (pi->enable_boost)
+		sumo_enable_boost(rdev, new_ps, true);
+	if (pi->enable_dpm)
+		sumo_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO;
+
+	return 0;
+}
+
+void sumo_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+
+	sumo_update_current_ps(rdev, new_ps);
+}
+
+void sumo_dpm_reset_asic(struct radeon_device *rdev)
+{
+	sumo_program_bootup_state(rdev);
+	sumo_enable_power_level_0(rdev);
+	sumo_set_forced_level_0(rdev);
+	sumo_set_forced_mode_enabled(rdev);
+	sumo_wait_for_level_0(rdev);
+	sumo_set_forced_mode_disabled(rdev);
+	sumo_set_forced_mode_enabled(rdev);
+	sumo_set_forced_mode_disabled(rdev);
+}
+
+void sumo_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	sumo_initialize_m3_arb(rdev);
+	pi->fw_version = sumo_get_running_fw_version(rdev);
+	DRM_INFO("Found smc ucode version: 0x%08x\n", pi->fw_version);
+	sumo_program_acpi_power_level(rdev);
+	sumo_enable_acpi_pm(rdev);
+	sumo_take_smu_control(rdev, true);
+}
+
+void sumo_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void sumo_patch_boot_state(struct radeon_device *rdev,
+				  struct sumo_ps *ps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	ps->num_levels = 1;
+	ps->flags = 0;
+	ps->levels[0] = pi->boot_pl;
+}
+
+static void sumo_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					    struct radeon_ps *rps,
+					    struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					    u8 table_rev)
+{
+	struct sumo_ps *ps = sumo_get_ps(rps);
+
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		rdev->pm.dpm.boot_ps = rps;
+		sumo_patch_boot_state(rdev, ps);
+	}
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void sumo_parse_pplib_clock_info(struct radeon_device *rdev,
+					struct radeon_ps *rps, int index,
+					union pplib_clock_info *clock_info)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	struct sumo_pl *pl = &ps->levels[index];
+	u32 sclk;
+
+	sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow);
+	sclk |= clock_info->sumo.ucEngineClockHigh << 16;
+	pl->sclk = sclk;
+	pl->vddc_index = clock_info->sumo.vddcIndex;
+	pl->sclk_dpm_tdp_limit = clock_info->sumo.tdpLimit;
+
+	ps->num_levels = index + 1;
+
+	if (pi->enable_sclk_ds) {
+		pl->ds_divider_index = 5;
+		pl->ss_divider_index = 4;
+	}
+}
+
+static int sumo_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j, k, non_clock_array_index, clock_array_index;
+	union pplib_clock_info *clock_info;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	u8 *power_state_offset;
+	struct sumo_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	state_array = (struct _StateArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
+	clock_info_array = (struct _ClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
+	non_clock_info_array = (struct _NonClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  state_array->ucNumEntries, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	power_state_offset = (u8 *)state_array->states;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+	for (i = 0; i < state_array->ucNumEntries; i++) {
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			&non_clock_info_array->nonClockInfo[non_clock_array_index];
+		if (!rdev->pm.power_state[i].clock_info)
+			return -EINVAL;
+		ps = kzalloc(sizeof(struct sumo_ps), GFP_KERNEL);
+		if (ps == NULL) {
+			kfree(rdev->pm.dpm.ps);
+			return -ENOMEM;
+		}
+		rdev->pm.dpm.ps[i].ps_priv = ps;
+		k = 0;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			if (k >= SUMO_MAX_HARDWARE_POWERLEVELS)
+				break;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
+			sumo_parse_pplib_clock_info(rdev,
+						    &rdev->pm.dpm.ps[i], k,
+						    clock_info);
+			k++;
+		}
+		sumo_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+						non_clock_info,
+						non_clock_info_array->ucEntrySize);
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
+	}
+	rdev->pm.dpm.num_ps = state_array->ucNumEntries;
+	return 0;
+}
+
+u32 sumo_convert_vid2_to_vid7(struct radeon_device *rdev,
+			      struct sumo_vid_mapping_table *vid_mapping_table,
+			      u32 vid_2bit)
+{
+	u32 i;
+
+	for (i = 0; i < vid_mapping_table->num_entries; i++) {
+		if (vid_mapping_table->entries[i].vid_2bit == vid_2bit)
+			return vid_mapping_table->entries[i].vid_7bit;
+	}
+
+	return vid_mapping_table->entries[vid_mapping_table->num_entries - 1].vid_7bit;
+}
+
+static u16 sumo_convert_voltage_index_to_value(struct radeon_device *rdev,
+					       u32 vid_2bit)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 vid_7bit = sumo_convert_vid2_to_vid7(rdev, &pi->sys_info.vid_mapping_table, vid_2bit);
+
+	if (vid_7bit > 0x7C)
+		return 0;
+
+	return (15500 - vid_7bit * 125 + 5) / 10;
+}
+
+static void sumo_construct_display_voltage_mapping_table(struct radeon_device *rdev,
+							 struct sumo_disp_clock_voltage_mapping_table *disp_clk_voltage_mapping_table,
+							 ATOM_CLK_VOLT_CAPABILITY *table)
+{
+	u32 i;
+
+	for (i = 0; i < SUMO_MAX_NUMBER_VOLTAGES; i++) {
+		if (table[i].ulMaximumSupportedCLK == 0)
+			break;
+
+		disp_clk_voltage_mapping_table->display_clock_frequency[i] =
+			table[i].ulMaximumSupportedCLK;
+	}
+
+	disp_clk_voltage_mapping_table->num_max_voltage_levels = i;
+
+	if (disp_clk_voltage_mapping_table->num_max_voltage_levels == 0) {
+		disp_clk_voltage_mapping_table->display_clock_frequency[0] = 80000;
+		disp_clk_voltage_mapping_table->num_max_voltage_levels = 1;
+	}
+}
+
+void sumo_construct_sclk_voltage_mapping_table(struct radeon_device *rdev,
+					       struct sumo_sclk_voltage_mapping_table *sclk_voltage_mapping_table,
+					       ATOM_AVAILABLE_SCLK_LIST *table)
+{
+	u32 i;
+	u32 n = 0;
+	u32 prev_sclk = 0;
+
+	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) {
+		if (table[i].ulSupportedSCLK > prev_sclk) {
+			sclk_voltage_mapping_table->entries[n].sclk_frequency =
+				table[i].ulSupportedSCLK;
+			sclk_voltage_mapping_table->entries[n].vid_2bit =
+				table[i].usVoltageIndex;
+			prev_sclk = table[i].ulSupportedSCLK;
+			n++;
+		}
+	}
+
+	sclk_voltage_mapping_table->num_max_dpm_entries = n;
+}
+
+void sumo_construct_vid_mapping_table(struct radeon_device *rdev,
+				      struct sumo_vid_mapping_table *vid_mapping_table,
+				      ATOM_AVAILABLE_SCLK_LIST *table)
+{
+	u32 i, j;
+
+	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) {
+		if (table[i].ulSupportedSCLK != 0) {
+			vid_mapping_table->entries[table[i].usVoltageIndex].vid_7bit =
+				table[i].usVoltageID;
+			vid_mapping_table->entries[table[i].usVoltageIndex].vid_2bit =
+				table[i].usVoltageIndex;
+		}
+	}
+
+	for (i = 0; i < SUMO_MAX_NUMBER_VOLTAGES; i++) {
+		if (vid_mapping_table->entries[i].vid_7bit == 0) {
+			for (j = i + 1; j < SUMO_MAX_NUMBER_VOLTAGES; j++) {
+				if (vid_mapping_table->entries[j].vid_7bit != 0) {
+					vid_mapping_table->entries[i] =
+						vid_mapping_table->entries[j];
+					vid_mapping_table->entries[j].vid_7bit = 0;
+					break;
+				}
+			}
+
+			if (j == SUMO_MAX_NUMBER_VOLTAGES)
+				break;
+		}
+	}
+
+	vid_mapping_table->num_entries = i;
+}
+
+union igp_info {
+	struct _ATOM_INTEGRATED_SYSTEM_INFO info;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V5 info_5;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6;
+};
+
+static int sumo_parse_sys_info_table(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo);
+	union igp_info *igp_info;
+	u8 frev, crev;
+	u16 data_offset;
+	int i;
+
+	if (atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		igp_info = (union igp_info *)(mode_info->atom_context->bios +
+					      data_offset);
+
+		if (crev != 6) {
+			DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev);
+			return -EINVAL;
+		}
+		pi->sys_info.bootup_sclk = le32_to_cpu(igp_info->info_6.ulBootUpEngineClock);
+		pi->sys_info.min_sclk = le32_to_cpu(igp_info->info_6.ulMinEngineClock);
+		pi->sys_info.bootup_uma_clk = le32_to_cpu(igp_info->info_6.ulBootUpUMAClock);
+		pi->sys_info.bootup_nb_voltage_index =
+			le16_to_cpu(igp_info->info_6.usBootUpNBVoltage);
+		if (igp_info->info_6.ucHtcTmpLmt == 0)
+			pi->sys_info.htc_tmp_lmt = 203;
+		else
+			pi->sys_info.htc_tmp_lmt = igp_info->info_6.ucHtcTmpLmt;
+		if (igp_info->info_6.ucHtcHystLmt == 0)
+			pi->sys_info.htc_hyst_lmt = 5;
+		else
+			pi->sys_info.htc_hyst_lmt = igp_info->info_6.ucHtcHystLmt;
+		if (pi->sys_info.htc_tmp_lmt <= pi->sys_info.htc_hyst_lmt) {
+			DRM_ERROR("The htcTmpLmt should be larger than htcHystLmt.\n");
+		}
+		for (i = 0; i < NUMBER_OF_M3ARB_PARAM_SETS; i++) {
+			pi->sys_info.csr_m3_arb_cntl_default[i] =
+				le32_to_cpu(igp_info->info_6.ulCSR_M3_ARB_CNTL_DEFAULT[i]);
+			pi->sys_info.csr_m3_arb_cntl_uvd[i] =
+				le32_to_cpu(igp_info->info_6.ulCSR_M3_ARB_CNTL_UVD[i]);
+			pi->sys_info.csr_m3_arb_cntl_fs3d[i] =
+				le32_to_cpu(igp_info->info_6.ulCSR_M3_ARB_CNTL_FS3D[i]);
+		}
+		pi->sys_info.sclk_dpm_boost_margin =
+			le32_to_cpu(igp_info->info_6.SclkDpmBoostMargin);
+		pi->sys_info.sclk_dpm_throttle_margin =
+			le32_to_cpu(igp_info->info_6.SclkDpmThrottleMargin);
+		pi->sys_info.sclk_dpm_tdp_limit_pg =
+			le16_to_cpu(igp_info->info_6.SclkDpmTdpLimitPG);
+		pi->sys_info.gnb_tdp_limit = le16_to_cpu(igp_info->info_6.GnbTdpLimit);
+		pi->sys_info.sclk_dpm_tdp_limit_boost =
+			le16_to_cpu(igp_info->info_6.SclkDpmTdpLimitBoost);
+		pi->sys_info.boost_sclk = le32_to_cpu(igp_info->info_6.ulBoostEngineCLock);
+		pi->sys_info.boost_vid_2bit = igp_info->info_6.ulBoostVid_2bit;
+		if (igp_info->info_6.EnableBoost)
+			pi->sys_info.enable_boost = true;
+		else
+			pi->sys_info.enable_boost = false;
+		sumo_construct_display_voltage_mapping_table(rdev,
+							     &pi->sys_info.disp_clk_voltage_mapping_table,
+							     igp_info->info_6.sDISPCLK_Voltage);
+		sumo_construct_sclk_voltage_mapping_table(rdev,
+							  &pi->sys_info.sclk_voltage_mapping_table,
+							  igp_info->info_6.sAvail_SCLK);
+		sumo_construct_vid_mapping_table(rdev, &pi->sys_info.vid_mapping_table,
+						 igp_info->info_6.sAvail_SCLK);
+
+	}
+	return 0;
+}
+
+static void sumo_construct_boot_and_acpi_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	pi->boot_pl.sclk = pi->sys_info.bootup_sclk;
+	pi->boot_pl.vddc_index = pi->sys_info.bootup_nb_voltage_index;
+	pi->boot_pl.ds_divider_index = 0;
+	pi->boot_pl.ss_divider_index = 0;
+	pi->boot_pl.allow_gnb_slow = 1;
+	pi->acpi_pl = pi->boot_pl;
+	pi->current_ps.num_levels = 1;
+	pi->current_ps.levels[0] = pi->boot_pl;
+}
+
+int sumo_dpm_init(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi;
+	u32 hw_rev = (RREG32(HW_REV) & ATI_REV_ID_MASK) >> ATI_REV_ID_SHIFT;
+	int ret;
+
+	pi = kzalloc(sizeof(struct sumo_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	pi->driver_nbps_policy_disable = false;
+	if ((rdev->family == CHIP_PALM) && (hw_rev < 3))
+		pi->disable_gfx_power_gating_in_uvd = true;
+	else
+		pi->disable_gfx_power_gating_in_uvd = false;
+	pi->enable_alt_vddnb = true;
+	pi->enable_sclk_ds = true;
+	pi->enable_dynamic_m3_arbiter = false;
+	pi->enable_dynamic_patch_ps = true;
+	pi->enable_gfx_power_gating = true;
+	pi->enable_gfx_clock_gating = true;
+	pi->enable_mg_clock_gating = true;
+	pi->enable_auto_thermal_throttling = true;
+
+	ret = sumo_parse_sys_info_table(rdev);
+	if (ret)
+		return ret;
+
+	sumo_construct_boot_and_acpi_state(rdev);
+
+	ret = sumo_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->asi = RV770_ASI_DFLT;
+	pi->thermal_auto_throttling = pi->sys_info.htc_tmp_lmt;
+	pi->enable_boost = pi->sys_info.enable_boost;
+	pi->enable_dpm = true;
+
+	return 0;
+}
+
+void sumo_dpm_print_power_state(struct radeon_device *rdev,
+				struct radeon_ps *rps)
+{
+	int i;
+	struct sumo_ps *ps = sumo_get_ps(rps);
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	for (i = 0; i < ps->num_levels; i++) {
+		struct sumo_pl *pl = &ps->levels[i];
+		printk("\t\tpower level %d    sclk: %u vddc: %u\n",
+		       i, pl->sclk,
+		       sumo_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void sumo_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						      struct seq_file *m)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	struct sumo_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_INDEX_MASK) >>
+		CURR_INDEX_SHIFT;
+
+	if (current_index == BOOST_DPM_LEVEL) {
+		pl = &pi->boost_pl;
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
+			   current_index, pl->sclk,
+			   sumo_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	} else if (current_index >= ps->num_levels) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
+			   current_index, pl->sclk,
+			   sumo_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+}
+
+void sumo_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	sumo_cleanup_asic(rdev); /* ??? */
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 sumo_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *requested_state = sumo_get_ps(&pi->requested_rps);
+
+	if (low)
+		return requested_state->levels[0].sclk;
+	else
+		return requested_state->levels[requested_state->num_levels - 1].sclk;
+}
+
+u32 sumo_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	return pi->sys_info.bootup_uma_clk;
+}
+
+int sumo_dpm_force_performance_level(struct radeon_device *rdev,
+				     enum radeon_dpm_forced_level level)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *rps = &pi->current_rps;
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	int i;
+
+	if (ps->num_levels <= 1)
+		return 0;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		sumo_power_level_enable(rdev, ps->num_levels - 1, true);
+		sumo_set_forced_level(rdev, ps->num_levels - 1);
+		sumo_set_forced_mode_enabled(rdev);
+		for (i = 0; i < ps->num_levels - 1; i++) {
+			sumo_power_level_enable(rdev, i, false);
+		}
+		sumo_set_forced_mode(rdev, false);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_set_forced_mode(rdev, false);
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		sumo_power_level_enable(rdev, 0, true);
+		sumo_set_forced_level(rdev, 0);
+		sumo_set_forced_mode_enabled(rdev);
+		for (i = 1; i < ps->num_levels; i++) {
+			sumo_power_level_enable(rdev, i, false);
+		}
+		sumo_set_forced_mode(rdev, false);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_set_forced_mode(rdev, false);
+	} else {
+		for (i = 0; i < ps->num_levels; i++) {
+			sumo_power_level_enable(rdev, i, true);
+		}
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/sumo_dpm.h b/drivers/gpu/drm/radeon/sumo_dpm.h
new file mode 100644
index 0000000..07dda29
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumo_dpm.h
@@ -0,0 +1,220 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SUMO_DPM_H__
+#define __SUMO_DPM_H__
+
+#include "atom.h"
+
+#define SUMO_MAX_HARDWARE_POWERLEVELS 5
+#define SUMO_PM_NUMBER_OF_TC 15
+
+struct sumo_pl {
+	u32 sclk;
+	u32 vddc_index;
+	u32 ds_divider_index;
+	u32 ss_divider_index;
+	u32 allow_gnb_slow;
+	u32 sclk_dpm_tdp_limit;
+};
+
+/* used for the flags field */
+#define SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE (1 << 0)
+#define SUMO_POWERSTATE_FLAGS_BOOST_STATE       (1 << 1)
+
+struct sumo_ps {
+	struct sumo_pl levels[SUMO_MAX_HARDWARE_POWERLEVELS];
+	u32 num_levels;
+	/* flags */
+	u32 flags;
+};
+
+#define NUMBER_OF_M3ARB_PARAM_SETS 10
+#define SUMO_MAX_NUMBER_VOLTAGES    4
+
+struct sumo_disp_clock_voltage_mapping_table {
+	u32 num_max_voltage_levels;
+	u32 display_clock_frequency[SUMO_MAX_NUMBER_VOLTAGES];
+};
+
+struct sumo_vid_mapping_entry {
+	u16 vid_2bit;
+	u16 vid_7bit;
+};
+
+struct sumo_vid_mapping_table {
+	u32 num_entries;
+	struct sumo_vid_mapping_entry entries[SUMO_MAX_NUMBER_VOLTAGES];
+};
+
+struct sumo_sclk_voltage_mapping_entry {
+	u32 sclk_frequency;
+	u16 vid_2bit;
+	u16 rsv;
+};
+
+struct sumo_sclk_voltage_mapping_table {
+	u32 num_max_dpm_entries;
+	struct sumo_sclk_voltage_mapping_entry entries[SUMO_MAX_HARDWARE_POWERLEVELS];
+};
+
+struct sumo_sys_info {
+	u32 bootup_sclk;
+	u32 min_sclk;
+	u32 bootup_uma_clk;
+	u16 bootup_nb_voltage_index;
+	u8 htc_tmp_lmt;
+	u8 htc_hyst_lmt;
+	struct sumo_sclk_voltage_mapping_table sclk_voltage_mapping_table;
+	struct sumo_disp_clock_voltage_mapping_table disp_clk_voltage_mapping_table;
+	struct sumo_vid_mapping_table vid_mapping_table;
+	u32 csr_m3_arb_cntl_default[NUMBER_OF_M3ARB_PARAM_SETS];
+	u32 csr_m3_arb_cntl_uvd[NUMBER_OF_M3ARB_PARAM_SETS];
+	u32 csr_m3_arb_cntl_fs3d[NUMBER_OF_M3ARB_PARAM_SETS];
+	u32 sclk_dpm_boost_margin;
+	u32 sclk_dpm_throttle_margin;
+	u32 sclk_dpm_tdp_limit_pg;
+	u32 gnb_tdp_limit;
+	u32 sclk_dpm_tdp_limit_boost;
+	u32 boost_sclk;
+	u32 boost_vid_2bit;
+	bool enable_boost;
+};
+
+struct sumo_power_info {
+	u32 asi;
+	u32 pasi;
+	u32 bsp;
+	u32 bsu;
+	u32 pbsp;
+	u32 pbsu;
+	u32 dsp;
+	u32 psp;
+	u32 thermal_auto_throttling;
+	u32 uvd_m3_arbiter;
+	u32 fw_version;
+	struct sumo_sys_info sys_info;
+	struct sumo_pl acpi_pl;
+	struct sumo_pl boot_pl;
+	struct sumo_pl boost_pl;
+	bool disable_gfx_power_gating_in_uvd;
+	bool driver_nbps_policy_disable;
+	bool enable_alt_vddnb;
+	bool enable_dynamic_m3_arbiter;
+	bool enable_gfx_clock_gating;
+	bool enable_gfx_power_gating;
+	bool enable_mg_clock_gating;
+	bool enable_sclk_ds;
+	bool enable_auto_thermal_throttling;
+	bool enable_dynamic_patch_ps;
+	bool enable_dpm;
+	bool enable_boost;
+	struct radeon_ps current_rps;
+	struct sumo_ps current_ps;
+	struct radeon_ps requested_rps;
+	struct sumo_ps requested_ps;
+};
+
+#define SUMO_UTC_DFLT_00                     0x48
+#define SUMO_UTC_DFLT_01                     0x44
+#define SUMO_UTC_DFLT_02                     0x44
+#define SUMO_UTC_DFLT_03                     0x44
+#define SUMO_UTC_DFLT_04                     0x44
+#define SUMO_UTC_DFLT_05                     0x44
+#define SUMO_UTC_DFLT_06                     0x44
+#define SUMO_UTC_DFLT_07                     0x44
+#define SUMO_UTC_DFLT_08                     0x44
+#define SUMO_UTC_DFLT_09                     0x44
+#define SUMO_UTC_DFLT_10                     0x44
+#define SUMO_UTC_DFLT_11                     0x44
+#define SUMO_UTC_DFLT_12                     0x44
+#define SUMO_UTC_DFLT_13                     0x44
+#define SUMO_UTC_DFLT_14                     0x44
+
+#define SUMO_DTC_DFLT_00                     0x48
+#define SUMO_DTC_DFLT_01                     0x44
+#define SUMO_DTC_DFLT_02                     0x44
+#define SUMO_DTC_DFLT_03                     0x44
+#define SUMO_DTC_DFLT_04                     0x44
+#define SUMO_DTC_DFLT_05                     0x44
+#define SUMO_DTC_DFLT_06                     0x44
+#define SUMO_DTC_DFLT_07                     0x44
+#define SUMO_DTC_DFLT_08                     0x44
+#define SUMO_DTC_DFLT_09                     0x44
+#define SUMO_DTC_DFLT_10                     0x44
+#define SUMO_DTC_DFLT_11                     0x44
+#define SUMO_DTC_DFLT_12                     0x44
+#define SUMO_DTC_DFLT_13                     0x44
+#define SUMO_DTC_DFLT_14                     0x44
+
+#define SUMO_AH_DFLT               5
+
+#define SUMO_R_DFLT0               70
+#define SUMO_R_DFLT1               70
+#define SUMO_R_DFLT2               70
+#define SUMO_R_DFLT3               70
+#define SUMO_R_DFLT4               100
+
+#define SUMO_L_DFLT0               0
+#define SUMO_L_DFLT1               20
+#define SUMO_L_DFLT2               20
+#define SUMO_L_DFLT3               20
+#define SUMO_L_DFLT4               20
+#define SUMO_VRC_DFLT              0x30033
+#define SUMO_MGCGTTLOCAL0_DFLT     0
+#define SUMO_MGCGTTLOCAL1_DFLT     0
+#define SUMO_GICST_DFLT            19
+#define SUMO_SST_DFLT              8
+#define SUMO_VOLTAGEDROPT_DFLT     1
+#define SUMO_GFXPOWERGATINGT_DFLT  100
+
+/* sumo_dpm.c */
+void sumo_gfx_clockgating_initialize(struct radeon_device *rdev);
+void sumo_program_vc(struct radeon_device *rdev, u32 vrc);
+void sumo_clear_vc(struct radeon_device *rdev);
+void sumo_program_sstp(struct radeon_device *rdev);
+void sumo_take_smu_control(struct radeon_device *rdev, bool enable);
+void sumo_construct_sclk_voltage_mapping_table(struct radeon_device *rdev,
+					       struct sumo_sclk_voltage_mapping_table *sclk_voltage_mapping_table,
+					       ATOM_AVAILABLE_SCLK_LIST *table);
+void sumo_construct_vid_mapping_table(struct radeon_device *rdev,
+				      struct sumo_vid_mapping_table *vid_mapping_table,
+				      ATOM_AVAILABLE_SCLK_LIST *table);
+u32 sumo_convert_vid2_to_vid7(struct radeon_device *rdev,
+			      struct sumo_vid_mapping_table *vid_mapping_table,
+			      u32 vid_2bit);
+u32 sumo_get_sleep_divider_from_id(u32 id);
+u32 sumo_get_sleep_divider_id_from_clock(struct radeon_device *rdev,
+					 u32 sclk,
+					 u32 min_sclk_in_sr);
+
+/* sumo_smc.c */
+void sumo_initialize_m3_arb(struct radeon_device *rdev);
+void sumo_smu_pg_init(struct radeon_device *rdev);
+void sumo_set_tdp_limit(struct radeon_device *rdev, u32 index, u32 tdp_limit);
+void sumo_smu_notify_alt_vddnb_change(struct radeon_device *rdev,
+				      bool powersaving, bool force_nbps1);
+void sumo_boost_state_enable(struct radeon_device *rdev, bool enable);
+void sumo_enable_boost_timer(struct radeon_device *rdev);
+u32 sumo_get_running_fw_version(struct radeon_device *rdev);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/sumo_smc.c b/drivers/gpu/drm/radeon/sumo_smc.c
new file mode 100644
index 0000000..18abba5
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumo_smc.c
@@ -0,0 +1,222 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "sumod.h"
+#include "sumo_dpm.h"
+#include "ppsmc.h"
+
+#define SUMO_SMU_SERVICE_ROUTINE_PG_INIT        1
+#define SUMO_SMU_SERVICE_ROUTINE_ALTVDDNB_NOTIFY  27
+#define SUMO_SMU_SERVICE_ROUTINE_GFX_SRV_ID_20  20
+
+struct sumo_ps *sumo_get_ps(struct radeon_ps *rps);
+struct sumo_power_info *sumo_get_pi(struct radeon_device *rdev);
+
+static void sumo_send_msg_to_smu(struct radeon_device *rdev, u32 id)
+{
+	u32 gfx_int_req;
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_STATUS) & INT_DONE)
+			break;
+		udelay(1);
+	}
+
+	gfx_int_req = SERV_INDEX(id) | INT_REQ;
+	WREG32(GFX_INT_REQ, gfx_int_req);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_REQ) & INT_REQ)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_STATUS) & INT_ACK)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_STATUS) & INT_DONE)
+			break;
+		udelay(1);
+	}
+
+	gfx_int_req &= ~INT_REQ;
+	WREG32(GFX_INT_REQ, gfx_int_req);
+}
+
+void sumo_initialize_m3_arb(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 i;
+
+	if (!pi->enable_dynamic_m3_arbiter)
+		return;
+
+	for (i = 0; i < NUMBER_OF_M3ARB_PARAM_SETS; i++)
+		WREG32_RCU(MCU_M3ARB_PARAMS + (i * 4),
+			   pi->sys_info.csr_m3_arb_cntl_default[i]);
+
+	for (; i < NUMBER_OF_M3ARB_PARAM_SETS * 2; i++)
+		WREG32_RCU(MCU_M3ARB_PARAMS + (i * 4),
+			   pi->sys_info.csr_m3_arb_cntl_uvd[i % NUMBER_OF_M3ARB_PARAM_SETS]);
+
+	for (; i < NUMBER_OF_M3ARB_PARAM_SETS * 3; i++)
+		WREG32_RCU(MCU_M3ARB_PARAMS + (i * 4),
+			   pi->sys_info.csr_m3_arb_cntl_fs3d[i % NUMBER_OF_M3ARB_PARAM_SETS]);
+}
+
+static bool sumo_is_alt_vddnb_supported(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	bool return_code = false;
+
+	if (!pi->enable_alt_vddnb)
+		return return_code;
+
+	if ((rdev->family == CHIP_SUMO) || (rdev->family == CHIP_SUMO2)) {
+		if (pi->fw_version >= 0x00010C00)
+			return_code = true;
+	}
+
+	return return_code;
+}
+
+void sumo_smu_notify_alt_vddnb_change(struct radeon_device *rdev,
+				      bool powersaving, bool force_nbps1)
+{
+	u32 param = 0;
+
+	if (!sumo_is_alt_vddnb_supported(rdev))
+		return;
+
+	if (powersaving)
+		param |= 1;
+
+	if (force_nbps1)
+		param |= 2;
+
+	WREG32_RCU(RCU_ALTVDDNB_NOTIFY, param);
+
+	sumo_send_msg_to_smu(rdev, SUMO_SMU_SERVICE_ROUTINE_ALTVDDNB_NOTIFY);
+}
+
+void sumo_smu_pg_init(struct radeon_device *rdev)
+{
+	sumo_send_msg_to_smu(rdev, SUMO_SMU_SERVICE_ROUTINE_PG_INIT);
+}
+
+static u32 sumo_power_of_4(u32 unit)
+{
+	u32 ret = 1;
+	u32 i;
+
+	for (i = 0; i < unit; i++)
+		ret *= 4;
+
+	return ret;
+}
+
+void sumo_enable_boost_timer(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 period, unit, timer_value;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	unit = (RREG32_RCU(RCU_LCLK_SCALING_CNTL) & LCLK_SCALING_TIMER_PRESCALER_MASK)
+		>> LCLK_SCALING_TIMER_PRESCALER_SHIFT;
+
+	period = 100 * (xclk / 100 / sumo_power_of_4(unit));
+
+	timer_value = (period << 16) | (unit << 4);
+
+	WREG32_RCU(RCU_GNB_PWR_REP_TIMER_CNTL, timer_value);
+	WREG32_RCU(RCU_BOOST_MARGIN, pi->sys_info.sclk_dpm_boost_margin);
+	WREG32_RCU(RCU_THROTTLE_MARGIN, pi->sys_info.sclk_dpm_throttle_margin);
+	WREG32_RCU(GNB_TDP_LIMIT, pi->sys_info.gnb_tdp_limit);
+	WREG32_RCU(RCU_SclkDpmTdpLimitPG, pi->sys_info.sclk_dpm_tdp_limit_pg);
+
+	sumo_send_msg_to_smu(rdev, SUMO_SMU_SERVICE_ROUTINE_GFX_SRV_ID_20);
+}
+
+void sumo_set_tdp_limit(struct radeon_device *rdev, u32 index, u32 tdp_limit)
+{
+	u32 regoffset = 0;
+	u32 shift = 0;
+	u32 mask = 0xFFF;
+	u32 sclk_dpm_tdp_limit;
+
+	switch (index) {
+	case 0:
+		regoffset = RCU_SclkDpmTdpLimit01;
+		shift = 16;
+		break;
+	case 1:
+		regoffset = RCU_SclkDpmTdpLimit01;
+		shift = 0;
+		break;
+	case 2:
+		regoffset = RCU_SclkDpmTdpLimit23;
+		shift = 16;
+		break;
+	case 3:
+		regoffset = RCU_SclkDpmTdpLimit23;
+		shift = 0;
+		break;
+	case 4:
+		regoffset = RCU_SclkDpmTdpLimit47;
+		shift = 16;
+		break;
+	case 7:
+		regoffset = RCU_SclkDpmTdpLimit47;
+		shift = 0;
+		break;
+	default:
+		break;
+	}
+
+	sclk_dpm_tdp_limit = RREG32_RCU(regoffset);
+	sclk_dpm_tdp_limit &= ~(mask << shift);
+	sclk_dpm_tdp_limit |= (tdp_limit << shift);
+	WREG32_RCU(regoffset, sclk_dpm_tdp_limit);
+}
+
+void sumo_boost_state_enable(struct radeon_device *rdev, bool enable)
+{
+	u32 boost_disable = RREG32_RCU(RCU_GPU_BOOST_DISABLE);
+
+	boost_disable &= 0xFFFFFFFE;
+	boost_disable |= (enable ? 0 : 1);
+	WREG32_RCU(RCU_GPU_BOOST_DISABLE, boost_disable);
+}
+
+u32 sumo_get_running_fw_version(struct radeon_device *rdev)
+{
+	return RREG32_RCU(RCU_FW_VERSION);
+}
+
diff --git a/drivers/gpu/drm/radeon/sumod.h b/drivers/gpu/drm/radeon/sumod.h
new file mode 100644
index 0000000..7c9c2d4
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumod.h
@@ -0,0 +1,372 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef _SUMOD_H_
+#define _SUMOD_H_
+
+/* pm registers */
+
+/* rcu */
+#define RCU_FW_VERSION                                  0x30c
+
+#define RCU_PWR_GATING_SEQ0                             0x408
+#define RCU_PWR_GATING_SEQ1                             0x40c
+#define RCU_PWR_GATING_CNTL                             0x410
+#       define PWR_GATING_EN                            (1 << 0)
+#       define RSVD_MASK                                (0x3 << 1)
+#       define PCV(x)                                   ((x) << 3)
+#       define PCV_MASK                                 (0x1f << 3)
+#       define PCV_SHIFT                                3
+#       define PCP(x)                                   ((x) << 8)
+#       define PCP_MASK                                 (0xf << 8)
+#       define PCP_SHIFT                                8
+#       define RPW(x)                                   ((x) << 16)
+#       define RPW_MASK                                 (0xf << 16)
+#       define RPW_SHIFT                                16
+#       define ID(x)                                    ((x) << 24)
+#       define ID_MASK                                  (0xf << 24)
+#       define ID_SHIFT                                 24
+#       define PGS(x)                                   ((x) << 28)
+#       define PGS_MASK                                 (0xf << 28)
+#       define PGS_SHIFT                                28
+
+#define RCU_ALTVDDNB_NOTIFY                             0x430
+#define RCU_LCLK_SCALING_CNTL                           0x434
+#       define LCLK_SCALING_EN                          (1 << 0)
+#       define LCLK_SCALING_TYPE                        (1 << 1)
+#       define LCLK_SCALING_TIMER_PRESCALER(x)          ((x) << 4)
+#       define LCLK_SCALING_TIMER_PRESCALER_MASK        (0xf << 4)
+#       define LCLK_SCALING_TIMER_PRESCALER_SHIFT       4
+#       define LCLK_SCALING_TIMER_PERIOD(x)             ((x) << 16)
+#       define LCLK_SCALING_TIMER_PERIOD_MASK           (0xf << 16)
+#       define LCLK_SCALING_TIMER_PERIOD_SHIFT          16
+
+#define RCU_PWR_GATING_CNTL_2                           0x4a0
+#       define MPPU(x)                                  ((x) << 0)
+#       define MPPU_MASK                                (0xffff << 0)
+#       define MPPU_SHIFT                               0
+#       define MPPD(x)                                  ((x) << 16)
+#       define MPPD_MASK                                (0xffff << 16)
+#       define MPPD_SHIFT                               16
+#define RCU_PWR_GATING_CNTL_3                           0x4a4
+#       define DPPU(x)                                  ((x) << 0)
+#       define DPPU_MASK                                (0xffff << 0)
+#       define DPPU_SHIFT                               0
+#       define DPPD(x)                                  ((x) << 16)
+#       define DPPD_MASK                                (0xffff << 16)
+#       define DPPD_SHIFT                               16
+#define RCU_PWR_GATING_CNTL_4                           0x4a8
+#       define RT(x)                                    ((x) << 0)
+#       define RT_MASK                                  (0xffff << 0)
+#       define RT_SHIFT                                 0
+#       define IT(x)                                    ((x) << 16)
+#       define IT_MASK                                  (0xffff << 16)
+#       define IT_SHIFT                                 16
+
+/* yes these two have the same address */
+#define RCU_PWR_GATING_CNTL_5                           0x504
+#define RCU_GPU_BOOST_DISABLE                           0x508
+
+#define MCU_M3ARB_INDEX                                 0x504
+#define MCU_M3ARB_PARAMS                                0x508
+
+#define RCU_GNB_PWR_REP_TIMER_CNTL                      0x50C
+
+#define RCU_SclkDpmTdpLimit01                           0x514
+#define RCU_SclkDpmTdpLimit23                           0x518
+#define RCU_SclkDpmTdpLimit47                           0x51C
+#define RCU_SclkDpmTdpLimitPG                           0x520
+
+#define GNB_TDP_LIMIT                                   0x540
+#define RCU_BOOST_MARGIN                                0x544
+#define RCU_THROTTLE_MARGIN                             0x548
+
+#define SMU_PCIE_PG_ARGS                                0x58C
+#define SMU_PCIE_PG_ARGS_2                              0x598
+#define SMU_PCIE_PG_ARGS_3                              0x59C
+
+/* mmio */
+#define RCU_STATUS                                      0x11c
+#       define GMC_PWR_GATER_BUSY                       (1 << 8)
+#       define GFX_PWR_GATER_BUSY                       (1 << 9)
+#       define UVD_PWR_GATER_BUSY                       (1 << 10)
+#       define PCIE_PWR_GATER_BUSY                      (1 << 11)
+#       define GMC_PWR_GATER_STATE                      (1 << 12)
+#       define GFX_PWR_GATER_STATE                      (1 << 13)
+#       define UVD_PWR_GATER_STATE                      (1 << 14)
+#       define PCIE_PWR_GATER_STATE                     (1 << 15)
+#       define GFX1_PWR_GATER_BUSY                      (1 << 16)
+#       define GFX2_PWR_GATER_BUSY                      (1 << 17)
+#       define GFX1_PWR_GATER_STATE                     (1 << 18)
+#       define GFX2_PWR_GATER_STATE                     (1 << 19)
+
+#define GFX_INT_REQ                                     0x120
+#       define INT_REQ                                  (1 << 0)
+#       define SERV_INDEX(x)                            ((x) << 1)
+#       define SERV_INDEX_MASK                          (0xff << 1)
+#       define SERV_INDEX_SHIFT                         1
+#define GFX_INT_STATUS                                  0x124
+#       define INT_ACK                                  (1 << 0)
+#       define INT_DONE                                 (1 << 1)
+
+#define CG_SCLK_CNTL                                    0x600
+#       define SCLK_DIVIDER(x)                          ((x) << 0)
+#       define SCLK_DIVIDER_MASK                        (0x7f << 0)
+#       define SCLK_DIVIDER_SHIFT                       0
+#define CG_SCLK_STATUS                                  0x604
+#       define SCLK_OVERCLK_DETECT                      (1 << 2)
+
+#define CG_DCLK_CNTL                                    0x610
+#       define DCLK_DIVIDER_MASK                        0x7f
+#       define DCLK_DIR_CNTL_EN                         (1 << 8)
+#define CG_DCLK_STATUS                                  0x614
+#       define DCLK_STATUS                              (1 << 0)
+#define CG_VCLK_CNTL                                    0x618
+#       define VCLK_DIVIDER_MASK                        0x7f
+#       define VCLK_DIR_CNTL_EN                         (1 << 8)
+#define CG_VCLK_STATUS                                  0x61c
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define STATIC_PM_EN                             (1 << 1)
+
+#define SCLK_PWRMGT_CNTL                                0x644
+#       define SCLK_PWRMGT_OFF                          (1 << 0)
+#       define SCLK_LOW_D1                              (1 << 1)
+#       define FIR_RESET                                (1 << 4)
+#       define FIR_FORCE_TREND_SEL                      (1 << 5)
+#       define FIR_TREND_MODE                           (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                       (1 << 7)
+#       define GFX_CLK_FORCE_ON                         (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                      (1 << 9)
+#       define GFX_CLK_FORCE_OFF                        (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                      (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                      (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                      (1 << 13)
+#       define GFX_VOLTAGE_CHANGE_EN                    (1 << 16)
+#       define GFX_VOLTAGE_CHANGE_MODE                  (1 << 17)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                0x66c
+#       define TARG_SCLK_INDEX(x)                       ((x) << 6)
+#       define TARG_SCLK_INDEX_MASK                     (0x7 << 6)
+#       define TARG_SCLK_INDEX_SHIFT                    6
+#       define CURR_SCLK_INDEX(x)                       ((x) << 9)
+#       define CURR_SCLK_INDEX_MASK                     (0x7 << 9)
+#       define CURR_SCLK_INDEX_SHIFT                    9
+#       define TARG_INDEX(x)                            ((x) << 12)
+#       define TARG_INDEX_MASK                          (0x7 << 12)
+#       define TARG_INDEX_SHIFT                         12
+#       define CURR_INDEX(x)                            ((x) << 15)
+#       define CURR_INDEX_MASK                          (0x7 << 15)
+#       define CURR_INDEX_SHIFT                         15
+
+#define CG_SCLK_DPM_CTRL                                0x684
+#       define SCLK_FSTATE_0_DIV(x)                     ((x) << 0)
+#       define SCLK_FSTATE_0_DIV_MASK                   (0x7f << 0)
+#       define SCLK_FSTATE_0_DIV_SHIFT                  0
+#       define SCLK_FSTATE_0_VLD                        (1 << 7)
+#       define SCLK_FSTATE_1_DIV(x)                     ((x) << 8)
+#       define SCLK_FSTATE_1_DIV_MASK                   (0x7f << 8)
+#       define SCLK_FSTATE_1_DIV_SHIFT                  8
+#       define SCLK_FSTATE_1_VLD                        (1 << 15)
+#       define SCLK_FSTATE_2_DIV(x)                     ((x) << 16)
+#       define SCLK_FSTATE_2_DIV_MASK                   (0x7f << 16)
+#       define SCLK_FSTATE_2_DIV_SHIFT                  16
+#       define SCLK_FSTATE_2_VLD                        (1 << 23)
+#       define SCLK_FSTATE_3_DIV(x)                     ((x) << 24)
+#       define SCLK_FSTATE_3_DIV_MASK                   (0x7f << 24)
+#       define SCLK_FSTATE_3_DIV_SHIFT                  24
+#       define SCLK_FSTATE_3_VLD                        (1 << 31)
+#define CG_SCLK_DPM_CTRL_2                              0x688
+#define CG_GCOOR                                        0x68c
+#       define PHC(x)                                   ((x) << 0)
+#       define PHC_MASK                                 (0x1f << 0)
+#       define PHC_SHIFT                                0
+#       define SDC(x)                                   ((x) << 9)
+#       define SDC_MASK                                 (0x3ff << 9)
+#       define SDC_SHIFT                                9
+#       define SU(x)                                    ((x) << 23)
+#       define SU_MASK                                  (0xf << 23)
+#       define SU_SHIFT                                 23
+#       define DIV_ID(x)                                ((x) << 28)
+#       define DIV_ID_MASK                              (0x7 << 28)
+#       define DIV_ID_SHIFT                             28
+
+#define CG_FTV                                          0x690
+#define CG_FFCT_0                                       0x694
+#       define UTC_0(x)                                 ((x) << 0)
+#       define UTC_0_MASK                               (0x3ff << 0)
+#       define UTC_0_SHIFT                              0
+#       define DTC_0(x)                                 ((x) << 10)
+#       define DTC_0_MASK                               (0x3ff << 10)
+#       define DTC_0_SHIFT                              10
+
+#define CG_GIT                                          0x6d8
+#       define CG_GICST(x)                              ((x) << 0)
+#       define CG_GICST_MASK                            (0xffff << 0)
+#       define CG_GICST_SHIFT                           0
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+#       define CG_GIPOT_SHIFT                           16
+
+#define CG_SCLK_DPM_CTRL_3                              0x6e0
+#       define FORCE_SCLK_STATE(x)                      ((x) << 0)
+#       define FORCE_SCLK_STATE_MASK                    (0x7 << 0)
+#       define FORCE_SCLK_STATE_SHIFT                   0
+#       define FORCE_SCLK_STATE_EN                      (1 << 3)
+#       define GNB_TT(x)                                ((x) << 8)
+#       define GNB_TT_MASK                              (0xff << 8)
+#       define GNB_TT_SHIFT                             8
+#       define GNB_THERMTHRO_MASK                       (1 << 16)
+#       define CNB_THERMTHRO_MASK_SCLK                  (1 << 17)
+#       define DPM_SCLK_ENABLE                          (1 << 18)
+#       define GNB_SLOW_FSTATE_0_MASK                   (1 << 23)
+#       define GNB_SLOW_FSTATE_0_SHIFT                  23
+#       define FORCE_NB_PSTATE_1                        (1 << 31)
+
+#define CG_SSP                                          0x6e8
+#       define SST(x)                                   ((x) << 0)
+#       define SST_MASK                                 (0xffff << 0)
+#       define SST_SHIFT                                0
+#       define SSTU(x)                                  ((x) << 16)
+#       define SSTU_MASK                                (0xffff << 16)
+#       define SSTU_SHIFT                               16
+
+#define CG_ACPI_CNTL                                    0x70c
+#       define SCLK_ACPI_DIV(x)                         ((x) << 0)
+#       define SCLK_ACPI_DIV_MASK                       (0x7f << 0)
+#       define SCLK_ACPI_DIV_SHIFT                      0
+
+#define CG_SCLK_DPM_CTRL_4                              0x71c
+#       define DC_HDC(x)                                ((x) << 14)
+#       define DC_HDC_MASK                              (0x3fff << 14)
+#       define DC_HDC_SHIFT                             14
+#       define DC_HU(x)                                 ((x) << 28)
+#       define DC_HU_MASK                               (0xf << 28)
+#       define DC_HU_SHIFT                              28
+#define CG_SCLK_DPM_CTRL_5                              0x720
+#       define SCLK_FSTATE_BOOTUP(x)                    ((x) << 0)
+#       define SCLK_FSTATE_BOOTUP_MASK                  (0x7 << 0)
+#       define SCLK_FSTATE_BOOTUP_SHIFT                 0
+#       define TT_TP(x)                                 ((x) << 3)
+#       define TT_TP_MASK                               (0xffff << 3)
+#       define TT_TP_SHIFT                              3
+#       define TT_TU(x)                                 ((x) << 19)
+#       define TT_TU_MASK                               (0xff << 19)
+#       define TT_TU_SHIFT                              19
+#define CG_SCLK_DPM_CTRL_6                              0x724
+#define CG_AT_0                                         0x728
+#       define CG_R(x)                                  ((x) << 0)
+#       define CG_R_MASK                                (0xffff << 0)
+#       define CG_R_SHIFT                               0
+#       define CG_L(x)                                  ((x) << 16)
+#       define CG_L_MASK                                (0xffff << 16)
+#       define CG_L_SHIFT                               16
+#define CG_AT_1                                         0x72c
+#define CG_AT_2                                         0x730
+#define	CG_THERMAL_INT					0x734
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+#define CG_AT_3                                         0x738
+#define CG_AT_4                                         0x73c
+#define CG_AT_5                                         0x740
+#define CG_AT_6                                         0x744
+#define CG_AT_7                                         0x748
+
+#define CG_BSP_0                                        0x750
+#       define BSP(x)                                   ((x) << 0)
+#       define BSP_MASK                                 (0xffff << 0)
+#       define BSP_SHIFT                                0
+#       define BSU(x)                                   ((x) << 16)
+#       define BSU_MASK                                 (0xf << 16)
+#       define BSU_SHIFT                                16
+
+#define CG_CG_VOLTAGE_CNTL                              0x770
+#       define REQ                                      (1 << 0)
+#       define LEVEL(x)                                 ((x) << 1)
+#       define LEVEL_MASK                               (0x3 << 1)
+#       define LEVEL_SHIFT                              1
+#       define CG_VOLTAGE_EN                            (1 << 3)
+#       define FORCE                                    (1 << 4)
+#       define PERIOD(x)                                ((x) << 8)
+#       define PERIOD_MASK                              (0xffff << 8)
+#       define PERIOD_SHIFT                             8
+#       define UNIT(x)                                  ((x) << 24)
+#       define UNIT_MASK                                (0xf << 24)
+#       define UNIT_SHIFT                               24
+
+#define CG_ACPI_VOLTAGE_CNTL                            0x780
+#       define ACPI_VOLTAGE_EN                          (1 << 8)
+
+#define CG_DPM_VOLTAGE_CNTL                             0x788
+#       define DPM_STATE0_LEVEL_MASK                    (0x3 << 0)
+#       define DPM_STATE0_LEVEL_SHIFT                   0
+#       define DPM_VOLTAGE_EN                           (1 << 16)
+
+#define CG_PWR_GATING_CNTL                              0x7ac
+#       define DYN_PWR_DOWN_EN                          (1 << 0)
+#       define ACPI_PWR_DOWN_EN                         (1 << 1)
+#       define GFX_CLK_OFF_PWR_DOWN_EN                  (1 << 2)
+#       define IOC_DISGPU_PWR_DOWN_EN                   (1 << 3)
+#       define FORCE_POWR_ON                            (1 << 4)
+#       define PGP(x)                                   ((x) << 8)
+#       define PGP_MASK                                 (0xffff << 8)
+#       define PGP_SHIFT                                8
+#       define PGU(x)                                   ((x) << 24)
+#       define PGU_MASK                                 (0xf << 24)
+#       define PGU_SHIFT                                24
+
+#define CG_CGTT_LOCAL_0                                 0x7d0
+#define CG_CGTT_LOCAL_1                                 0x7d4
+
+#define DEEP_SLEEP_CNTL                                 0x818
+#       define R_DIS                                    (1 << 3)
+#       define HS(x)                                    ((x) << 4)
+#       define HS_MASK                                  (0xfff << 4)
+#       define HS_SHIFT                                 4
+#       define ENABLE_DS                                (1 << 31)
+#define DEEP_SLEEP_CNTL2                                0x81c
+#       define LB_UFP_EN                                (1 << 0)
+#       define INOUT_C(x)                               ((x) << 4)
+#       define INOUT_C_MASK                             (0xff << 4)
+#       define INOUT_C_SHIFT                            4
+
+#define CG_SCRATCH2                                     0x824
+
+#define CG_SCLK_DPM_CTRL_11                             0x830
+
+#define HW_REV   					0x5564
+#       define ATI_REV_ID_MASK                          (0xf << 28)
+#       define ATI_REV_ID_SHIFT                         28
+/* 0 = A0, 1 = A1, 2 = B0, 3 = C0, etc. */
+
+#define DOUT_SCRATCH3   				0x611c
+
+#define GB_ADDR_CONFIG  				0x98f8
+
+#endif
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c
new file mode 100644
index 0000000..a1eb5f5
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinity_dpm.c
@@ -0,0 +1,1949 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "trinityd.h"
+#include "r600_dpm.h"
+#include "trinity_dpm.h"
+#include <linux/seq_file.h>
+
+#define TRINITY_MAX_DEEPSLEEP_DIVIDER_ID 5
+#define TRINITY_MINIMUM_ENGINE_CLOCK 800
+#define SCLK_MIN_DIV_INTV_SHIFT     12
+#define TRINITY_DISPCLK_BYPASS_THRESHOLD 10000
+
+#ifndef TRINITY_MGCG_SEQUENCE
+#define TRINITY_MGCG_SEQUENCE  100
+
+static const u32 trinity_mgcg_shls_default[] =
+{
+	/* Register, Value, Mask */
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00003fc4, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00000100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x00008984, 0x06000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00800200, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x00009744, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009664, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000104, 0xffffffff,
+	0x0000d0c0, 0x00000100, 0xffffffff,
+	0x0000d8c0, 0x00000100, 0xffffffff,
+	0x0000951c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff
+};
+
+static const u32 trinity_mgcg_shls_enable[] =
+{
+	/* Register, Value, Mask */
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0x000133FF,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xE00B03FC,
+	0x00009150, 0x96944200, 0xffffffff
+};
+
+static const u32 trinity_mgcg_shls_disable[] =
+{
+	/* Register, Value, Mask */
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0x000133FF,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xE00B03FC
+};
+#endif
+
+#ifndef TRINITY_SYSLS_SEQUENCE
+#define TRINITY_SYSLS_SEQUENCE  100
+
+static const u32 trinity_sysls_default[] =
+{
+	/* Register, Value, Mask */
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x0000641c, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+
+static const u32 trinity_sysls_disable[] =
+{
+	/* Register, Value, Mask */
+	0x0000d0c0, 0x00000000, 0xffffffff,
+	0x0000d8c0, 0x00000000, 0xffffffff,
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x0000641c, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+
+static const u32 trinity_sysls_enable[] =
+{
+	/* Register, Value, Mask */
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x0000d8bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000903, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x0000641c, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#endif
+
+static const u32 trinity_override_mgpg_sequences[] =
+{
+	/* Register, Value */
+	0x00000200, 0xE030032C,
+	0x00000204, 0x00000FFF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x800010FF,
+	0x00000200, 0x0001f198,
+	0x00000204, 0x0003ffff,
+	0x00000200, 0x0001f19C,
+	0x00000204, 0x3fffffff,
+	0x00000200, 0xE030032C,
+	0x00000204, 0x00000000,
+};
+
+static void trinity_program_clk_gating_hw_sequence(struct radeon_device *rdev,
+						   const u32 *seq, u32 count);
+static void trinity_override_dynamic_mg_powergating(struct radeon_device *rdev);
+static void trinity_apply_state_adjust_rules(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps);
+
+struct trinity_ps *trinity_get_ps(struct radeon_ps *rps)
+{
+	struct trinity_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void trinity_gfx_powergating_initialize(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 p, u;
+	u32 value;
+	struct atom_clock_dividers dividers;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 sssd = 1;
+	int ret;
+	u32 hw_rev = (RREG32(HW_REV) & ATI_REV_ID_MASK) >> ATI_REV_ID_SHIFT;
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             25000, false, &dividers);
+	if (ret)
+		return;
+
+	value = RREG32_SMC(GFX_POWER_GATING_CNTL);
+	value &= ~(SSSD_MASK | PDS_DIV_MASK);
+	if (sssd)
+		value |= SSSD(1);
+	value |= PDS_DIV(dividers.post_div);
+	WREG32_SMC(GFX_POWER_GATING_CNTL, value);
+
+	r600_calculate_u_and_p(500, xclk, 16, &p, &u);
+
+	WREG32(CG_PG_CTRL, SP(p) | SU(u));
+
+	WREG32_P(CG_GIPOTS, CG_GIPOT(p), ~CG_GIPOT_MASK);
+
+	/* XXX double check hw_rev */
+	if (pi->override_dynamic_mgpg && (hw_rev == 0))
+		trinity_override_dynamic_mg_powergating(rdev);
+
+}
+
+#define CGCG_CGTT_LOCAL0_MASK       0xFFFF33FF
+#define CGCG_CGTT_LOCAL1_MASK       0xFFFB0FFE
+#define CGTS_SM_CTRL_REG_DISABLE    0x00600000
+#define CGTS_SM_CTRL_REG_ENABLE     0x96944200
+
+static void trinity_mg_clockgating_enable(struct radeon_device *rdev,
+					  bool enable)
+{
+	u32 local0;
+	u32 local1;
+
+	if (enable) {
+		local0 = RREG32_CG(CG_CGTT_LOCAL_0);
+		local1 = RREG32_CG(CG_CGTT_LOCAL_1);
+
+		WREG32_CG(CG_CGTT_LOCAL_0,
+			  (0x00380000 & CGCG_CGTT_LOCAL0_MASK) | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32_CG(CG_CGTT_LOCAL_1,
+			  (0x0E000000 & CGCG_CGTT_LOCAL1_MASK) | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+
+		WREG32(CGTS_SM_CTRL_REG, CGTS_SM_CTRL_REG_ENABLE);
+	} else {
+		WREG32(CGTS_SM_CTRL_REG, CGTS_SM_CTRL_REG_DISABLE);
+
+		local0 = RREG32_CG(CG_CGTT_LOCAL_0);
+		local1 = RREG32_CG(CG_CGTT_LOCAL_1);
+
+		WREG32_CG(CG_CGTT_LOCAL_0,
+			  CGCG_CGTT_LOCAL0_MASK | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32_CG(CG_CGTT_LOCAL_1,
+			  CGCG_CGTT_LOCAL1_MASK | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+	}
+}
+
+static void trinity_mg_clockgating_initialize(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *seq = NULL;
+
+	seq = &trinity_mgcg_shls_default[0];
+	count = sizeof(trinity_mgcg_shls_default) / (3 * sizeof(u32));
+
+	trinity_program_clk_gating_hw_sequence(rdev, seq, count);
+}
+
+static void trinity_gfx_clockgating_enable(struct radeon_device *rdev,
+					   bool enable)
+{
+	if (enable) {
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+static void trinity_program_clk_gating_hw_sequence(struct radeon_device *rdev,
+						   const u32 *seq, u32 count)
+{
+	u32 i, length = count * 3;
+
+	for (i = 0; i < length; i += 3)
+		WREG32_P(seq[i], seq[i+1], ~seq[i+2]);
+}
+
+static void trinity_program_override_mgpg_sequences(struct radeon_device *rdev,
+						    const u32 *seq, u32 count)
+{
+	u32  i, length = count * 2;
+
+	for (i = 0; i < length; i += 2)
+		WREG32(seq[i], seq[i+1]);
+
+}
+
+static void trinity_override_dynamic_mg_powergating(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *seq = NULL;
+
+	seq = &trinity_override_mgpg_sequences[0];
+	count = sizeof(trinity_override_mgpg_sequences) / (2 * sizeof(u32));
+
+	trinity_program_override_mgpg_sequences(rdev, seq, count);
+}
+
+static void trinity_ls_clockgating_enable(struct radeon_device *rdev,
+					  bool enable)
+{
+	u32 count;
+	const u32 *seq = NULL;
+
+	if (enable) {
+		seq = &trinity_sysls_enable[0];
+		count = sizeof(trinity_sysls_enable) / (3 * sizeof(u32));
+	} else {
+		seq = &trinity_sysls_disable[0];
+		count = sizeof(trinity_sysls_disable) / (3 * sizeof(u32));
+	}
+
+	trinity_program_clk_gating_hw_sequence(rdev, seq, count);
+}
+
+static void trinity_gfx_powergating_enable(struct radeon_device *rdev,
+					   bool enable)
+{
+	if (enable) {
+		if (RREG32_SMC(CC_SMU_TST_EFUSE1_MISC) & RB_BACKEND_DISABLE_MASK)
+			WREG32_SMC(SMU_SCRATCH_A, (RREG32_SMC(SMU_SCRATCH_A) | 0x01));
+
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_PWR_DOWN_EN, ~DYN_PWR_DOWN_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_PWR_DOWN_EN);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+static void trinity_gfx_dynamic_mgpg_enable(struct radeon_device *rdev,
+					    bool enable)
+{
+	u32 value;
+
+	if (enable) {
+		value = RREG32_SMC(PM_I_CNTL_1);
+		value &= ~DS_PG_CNTL_MASK;
+		value |= DS_PG_CNTL(1);
+		WREG32_SMC(PM_I_CNTL_1, value);
+
+		value = RREG32_SMC(SMU_S_PG_CNTL);
+		value &= ~DS_PG_EN_MASK;
+		value |= DS_PG_EN(1);
+		WREG32_SMC(SMU_S_PG_CNTL, value);
+	} else {
+		value = RREG32_SMC(SMU_S_PG_CNTL);
+		value &= ~DS_PG_EN_MASK;
+		WREG32_SMC(SMU_S_PG_CNTL, value);
+
+		value = RREG32_SMC(PM_I_CNTL_1);
+		value &= ~DS_PG_CNTL_MASK;
+		WREG32_SMC(PM_I_CNTL_1, value);
+	}
+
+	trinity_gfx_dynamic_mgpg_config(rdev);
+
+}
+
+static void trinity_enable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_initialize(rdev);
+	if (pi->enable_mg_clock_gating)
+		trinity_mg_clockgating_initialize(rdev);
+	if (pi->enable_gfx_power_gating)
+		trinity_gfx_powergating_initialize(rdev);
+	if (pi->enable_mg_clock_gating) {
+		trinity_ls_clockgating_enable(rdev, true);
+		trinity_mg_clockgating_enable(rdev, true);
+	}
+	if (pi->enable_gfx_clock_gating)
+		trinity_gfx_clockgating_enable(rdev, true);
+	if (pi->enable_gfx_dynamic_mgpg)
+		trinity_gfx_dynamic_mgpg_enable(rdev, true);
+	if (pi->enable_gfx_power_gating)
+		trinity_gfx_powergating_enable(rdev, true);
+}
+
+static void trinity_disable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->enable_gfx_power_gating)
+		trinity_gfx_powergating_enable(rdev, false);
+	if (pi->enable_gfx_dynamic_mgpg)
+		trinity_gfx_dynamic_mgpg_enable(rdev, false);
+	if (pi->enable_gfx_clock_gating)
+		trinity_gfx_clockgating_enable(rdev, false);
+	if (pi->enable_mg_clock_gating) {
+		trinity_mg_clockgating_enable(rdev, false);
+		trinity_ls_clockgating_enable(rdev, false);
+	}
+}
+
+static void trinity_set_divider_value(struct radeon_device *rdev,
+				      u32 index, u32 sclk)
+{
+	struct atom_clock_dividers  dividers;
+	int ret;
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             sclk, false, &dividers);
+	if (ret)
+		return;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~CLK_DIVIDER_MASK;
+	value |= CLK_DIVIDER(dividers.post_div);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             sclk/2, false, &dividers);
+	if (ret)
+		return;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_PG_CNTL + ix);
+	value &= ~PD_SCLK_DIVIDER_MASK;
+	value |= PD_SCLK_DIVIDER(dividers.post_div);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_PG_CNTL + ix, value);
+}
+
+static void trinity_set_ds_dividers(struct radeon_device *rdev,
+				    u32 index, u32 divider)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~DS_DIV_MASK;
+	value |= DS_DIV(divider);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_ss_dividers(struct radeon_device *rdev,
+				    u32 index, u32 divider)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~DS_SH_DIV_MASK;
+	value |= DS_SH_DIV(divider);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_vid(struct radeon_device *rdev, u32 index, u32 vid)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 vid_7bit = sumo_convert_vid2_to_vid7(rdev, &pi->sys_info.vid_mapping_table, vid);
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~VID_MASK;
+	value |= VID(vid_7bit);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~LVRT_MASK;
+	value |= LVRT(0);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+}
+
+static void trinity_set_allos_gnb_slow(struct radeon_device *rdev,
+				       u32 index, u32 gnb_slow)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix);
+	value &= ~GNB_SLOW_MASK;
+	value |= GNB_SLOW(gnb_slow);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix, value);
+}
+
+static void trinity_set_force_nbp_state(struct radeon_device *rdev,
+					u32 index, u32 force_nbp_state)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix);
+	value &= ~FORCE_NBPS1_MASK;
+	value |= FORCE_NBPS1(force_nbp_state);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix, value);
+}
+
+static void trinity_set_display_wm(struct radeon_device *rdev,
+				   u32 index, u32 wm)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~DISPLAY_WM_MASK;
+	value |= DISPLAY_WM(wm);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_vce_wm(struct radeon_device *rdev,
+			       u32 index, u32 wm)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~VCE_WM_MASK;
+	value |= VCE_WM(wm);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_at(struct radeon_device *rdev,
+			   u32 index, u32 at)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_AT + ix);
+	value &= ~AT_MASK;
+	value |= AT(at);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_AT + ix, value);
+}
+
+static void trinity_program_power_level(struct radeon_device *rdev,
+					struct trinity_pl *pl, u32 index)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (index >= SUMO_MAX_HARDWARE_POWERLEVELS)
+		return;
+
+	trinity_set_divider_value(rdev, index, pl->sclk);
+	trinity_set_vid(rdev, index, pl->vddc_index);
+	trinity_set_ss_dividers(rdev, index, pl->ss_divider_index);
+	trinity_set_ds_dividers(rdev, index, pl->ds_divider_index);
+	trinity_set_allos_gnb_slow(rdev, index, pl->allow_gnb_slow);
+	trinity_set_force_nbp_state(rdev, index, pl->force_nbp_state);
+	trinity_set_display_wm(rdev, index, pl->display_wm);
+	trinity_set_vce_wm(rdev, index, pl->vce_wm);
+	trinity_set_at(rdev, index, pi->at[index]);
+}
+
+static void trinity_power_level_enable_disable(struct radeon_device *rdev,
+					       u32 index, bool enable)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~STATE_VALID_MASK;
+	if (enable)
+		value |= STATE_VALID(1);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+}
+
+static bool trinity_dpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32_SMC(SMU_SCLK_DPM_CNTL) & SCLK_DPM_EN(1))
+		return true;
+	else
+		return false;
+}
+
+static void trinity_start_dpm(struct radeon_device *rdev)
+{
+	u32 value = RREG32_SMC(SMU_SCLK_DPM_CNTL);
+
+	value &= ~(SCLK_DPM_EN_MASK | SCLK_DPM_BOOT_STATE_MASK | VOLTAGE_CHG_EN_MASK);
+	value |= SCLK_DPM_EN(1) | SCLK_DPM_BOOT_STATE(0) | VOLTAGE_CHG_EN(1);
+	WREG32_SMC(SMU_SCLK_DPM_CNTL, value);
+
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+	WREG32_P(CG_CG_VOLTAGE_CNTL, 0, ~EN);
+
+	trinity_dpm_config(rdev, true);
+}
+
+static void trinity_wait_for_dpm_enabled(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SCLK_PWRMGT_CNTL) & DYNAMIC_PM_EN)
+			break;
+		udelay(1);
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & TARGET_STATE_MASK) == 0)
+			break;
+		udelay(1);
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_MASK) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void trinity_stop_dpm(struct radeon_device *rdev)
+{
+	u32 sclk_dpm_cntl;
+
+	WREG32_P(CG_CG_VOLTAGE_CNTL, EN, ~EN);
+
+	sclk_dpm_cntl = RREG32_SMC(SMU_SCLK_DPM_CNTL);
+	sclk_dpm_cntl &= ~(SCLK_DPM_EN_MASK | VOLTAGE_CHG_EN_MASK);
+	WREG32_SMC(SMU_SCLK_DPM_CNTL, sclk_dpm_cntl);
+
+	trinity_dpm_config(rdev, false);
+}
+
+static void trinity_start_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~(RESET_SCLK_CNT | RESET_BUSY_CNT));
+}
+
+static void trinity_reset_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, RESET_SCLK_CNT | RESET_BUSY_CNT,
+		 ~(RESET_SCLK_CNT | RESET_BUSY_CNT));
+}
+
+static void trinity_wait_for_level_0(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_MASK) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void trinity_enable_power_level_0(struct radeon_device *rdev)
+{
+	trinity_power_level_enable_disable(rdev, 0, true);
+}
+
+static void trinity_force_level_0(struct radeon_device *rdev)
+{
+	trinity_dpm_force_state(rdev, 0);
+}
+
+static void trinity_unforce_levels(struct radeon_device *rdev)
+{
+	trinity_dpm_no_forced_level(rdev);
+}
+
+static void trinity_program_power_levels_0_to_n(struct radeon_device *rdev,
+						struct radeon_ps *new_rps,
+						struct radeon_ps *old_rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(new_rps);
+	struct trinity_ps *old_ps = trinity_get_ps(old_rps);
+	u32 i;
+	u32 n_current_state_levels = (old_ps == NULL) ? 1 : old_ps->num_levels;
+
+	for (i = 0; i < new_ps->num_levels; i++) {
+		trinity_program_power_level(rdev, &new_ps->levels[i], i);
+		trinity_power_level_enable_disable(rdev, i, true);
+	}
+
+	for (i = new_ps->num_levels; i < n_current_state_levels; i++)
+		trinity_power_level_enable_disable(rdev, i, false);
+}
+
+static void trinity_program_bootup_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i;
+
+	trinity_program_power_level(rdev, &pi->boot_pl, 0);
+	trinity_power_level_enable_disable(rdev, 0, true);
+
+	for (i = 1; i < 8; i++)
+		trinity_power_level_enable_disable(rdev, i, false);
+}
+
+static void trinity_setup_uvd_clock_table(struct radeon_device *rdev,
+					  struct radeon_ps *rps)
+{
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	u32 uvdstates = (ps->vclk_low_divider |
+			 ps->vclk_high_divider << 8 |
+			 ps->dclk_low_divider << 16 |
+			 ps->dclk_high_divider << 24);
+
+	WREG32_SMC(SMU_UVD_DPM_STATES, uvdstates);
+}
+
+static void trinity_setup_uvd_dpm_interval(struct radeon_device *rdev,
+					   u32 interval)
+{
+	u32 p, u;
+	u32 tp = RREG32_SMC(PM_TP);
+	u32 val;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(interval, xclk, 16, &p, &u);
+
+	val = (p + tp - 1) / tp;
+
+	WREG32_SMC(SMU_UVD_DPM_CNTL, val);
+}
+
+static bool trinity_uvd_clocks_zero(struct radeon_ps *rps)
+{
+	if ((rps->vclk == 0) && (rps->dclk == 0))
+		return true;
+	else
+		return false;
+}
+
+static bool trinity_uvd_clocks_equal(struct radeon_ps *rps1,
+				     struct radeon_ps *rps2)
+{
+	struct trinity_ps *ps1 = trinity_get_ps(rps1);
+	struct trinity_ps *ps2 = trinity_get_ps(rps2);
+
+	if ((rps1->vclk == rps2->vclk) &&
+	    (rps1->dclk == rps2->dclk) &&
+	    (ps1->vclk_low_divider == ps2->vclk_low_divider) &&
+	    (ps1->vclk_high_divider == ps2->vclk_high_divider) &&
+	    (ps1->dclk_low_divider == ps2->dclk_low_divider) &&
+	    (ps1->dclk_high_divider == ps2->dclk_high_divider))
+		return true;
+	else
+		return false;
+}
+
+static void trinity_setup_uvd_clocks(struct radeon_device *rdev,
+				     struct radeon_ps *new_rps,
+				     struct radeon_ps *old_rps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->enable_gfx_power_gating) {
+		trinity_gfx_powergating_enable(rdev, false);
+	}
+
+	if (pi->uvd_dpm) {
+		if (trinity_uvd_clocks_zero(new_rps) &&
+		    !trinity_uvd_clocks_zero(old_rps)) {
+			trinity_setup_uvd_dpm_interval(rdev, 0);
+		} else if (!trinity_uvd_clocks_zero(new_rps)) {
+			trinity_setup_uvd_clock_table(rdev, new_rps);
+
+			if (trinity_uvd_clocks_zero(old_rps)) {
+				u32 tmp = RREG32(CG_MISC_REG);
+				tmp &= 0xfffffffd;
+				WREG32(CG_MISC_REG, tmp);
+
+				radeon_set_uvd_clocks(rdev, new_rps->vclk, new_rps->dclk);
+
+				trinity_setup_uvd_dpm_interval(rdev, 3000);
+			}
+		}
+		trinity_uvd_dpm_config(rdev);
+	} else {
+		if (trinity_uvd_clocks_zero(new_rps) ||
+		    trinity_uvd_clocks_equal(new_rps, old_rps))
+			return;
+
+		radeon_set_uvd_clocks(rdev, new_rps->vclk, new_rps->dclk);
+	}
+
+	if (pi->enable_gfx_power_gating) {
+		trinity_gfx_powergating_enable(rdev, true);
+	}
+}
+
+static void trinity_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						       struct radeon_ps *new_rps,
+						       struct radeon_ps *old_rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(new_rps);
+	struct trinity_ps *current_ps = trinity_get_ps(new_rps);
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk >=
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	trinity_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+static void trinity_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						      struct radeon_ps *new_rps,
+						      struct radeon_ps *old_rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(new_rps);
+	struct trinity_ps *current_ps = trinity_get_ps(old_rps);
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk <
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	trinity_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+static void trinity_program_ttt(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 value = RREG32_SMC(SMU_SCLK_DPM_TTT);
+
+	value &= ~(HT_MASK | LT_MASK);
+	value |= HT((pi->thermal_auto_throttling + 49) * 8);
+	value |= LT((pi->thermal_auto_throttling + 49 - pi->sys_info.htc_hyst_lmt) * 8);
+	WREG32_SMC(SMU_SCLK_DPM_TTT, value);
+}
+
+static void trinity_enable_att(struct radeon_device *rdev)
+{
+	u32 value = RREG32_SMC(SMU_SCLK_DPM_TT_CNTL);
+
+	value &= ~SCLK_TT_EN_MASK;
+	value |= SCLK_TT_EN(1);
+	WREG32_SMC(SMU_SCLK_DPM_TT_CNTL, value);
+}
+
+static void trinity_program_sclk_dpm(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 tp = RREG32_SMC(PM_TP);
+	u32 ni;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 value;
+
+	r600_calculate_u_and_p(400, xclk, 16, &p, &u);
+
+	ni = (p + tp - 1) / tp;
+
+	value = RREG32_SMC(PM_I_CNTL_1);
+	value &= ~SCLK_DPM_MASK;
+	value |= SCLK_DPM(ni);
+	WREG32_SMC(PM_I_CNTL_1, value);
+}
+
+static int trinity_set_thermal_temperature_range(struct radeon_device *rdev,
+						 int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+        if (low_temp < min_temp)
+		low_temp = min_temp;
+        if (high_temp > max_temp)
+		high_temp = max_temp;
+        if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+                return -EINVAL;
+        }
+
+	WREG32_P(CG_THERMAL_INT_CTRL, DIG_THERM_INTH(49 + (high_temp / 1000)), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT_CTRL, DIG_THERM_INTL(49 + (low_temp / 1000)), ~DIG_THERM_INTL_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+static void trinity_update_current_ps(struct radeon_device *rdev,
+				      struct radeon_ps *rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->current_rps = *rps;
+	pi->current_ps = *new_ps;
+	pi->current_rps.ps_priv = &pi->current_ps;
+}
+
+static void trinity_update_requested_ps(struct radeon_device *rdev,
+					struct radeon_ps *rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->requested_rps = *rps;
+	pi->requested_ps = *new_ps;
+	pi->requested_rps.ps_priv = &pi->requested_ps;
+}
+
+int trinity_dpm_enable(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	int ret;
+
+	trinity_acquire_mutex(rdev);
+
+	if (trinity_dpm_enabled(rdev)) {
+		trinity_release_mutex(rdev);
+		return -EINVAL;
+	}
+
+	trinity_enable_clock_power_gating(rdev);
+	trinity_program_bootup_state(rdev);
+	sumo_program_vc(rdev, 0x00C00033);
+	trinity_start_am(rdev);
+	if (pi->enable_auto_thermal_throttling) {
+		trinity_program_ttt(rdev);
+		trinity_enable_att(rdev);
+	}
+	trinity_program_sclk_dpm(rdev);
+	trinity_start_dpm(rdev);
+	trinity_wait_for_dpm_enabled(rdev);
+	trinity_release_mutex(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		ret = trinity_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret) {
+			trinity_release_mutex(rdev);
+			return ret;
+		}
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	trinity_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+
+	return 0;
+}
+
+void trinity_dpm_disable(struct radeon_device *rdev)
+{
+	trinity_acquire_mutex(rdev);
+	if (!trinity_dpm_enabled(rdev)) {
+		trinity_release_mutex(rdev);
+		return;
+	}
+	trinity_disable_clock_power_gating(rdev);
+	sumo_clear_vc(rdev);
+	trinity_wait_for_level_0(rdev);
+	trinity_stop_dpm(rdev);
+	trinity_reset_am(rdev);
+	trinity_release_mutex(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	trinity_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+}
+
+static void trinity_get_min_sclk_divider(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->min_sclk_did =
+		(RREG32_SMC(CC_SMU_MISC_FUSES) & MinSClkDid_MASK) >> MinSClkDid_SHIFT;
+}
+
+static void trinity_setup_nbp_sim(struct radeon_device *rdev,
+				  struct radeon_ps *rps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct trinity_ps *new_ps = trinity_get_ps(rps);
+	u32 nbpsconfig;
+
+	if (pi->sys_info.nb_dpm_enable) {
+		nbpsconfig = RREG32_SMC(NB_PSTATE_CONFIG);
+		nbpsconfig &= ~(Dpm0PgNbPsLo_MASK | Dpm0PgNbPsHi_MASK | DpmXNbPsLo_MASK | DpmXNbPsHi_MASK);
+		nbpsconfig |= (Dpm0PgNbPsLo(new_ps->Dpm0PgNbPsLo) |
+			       Dpm0PgNbPsHi(new_ps->Dpm0PgNbPsHi) |
+			       DpmXNbPsLo(new_ps->DpmXNbPsLo) |
+			       DpmXNbPsHi(new_ps->DpmXNbPsHi));
+		WREG32_SMC(NB_PSTATE_CONFIG, nbpsconfig);
+	}
+}
+
+int trinity_dpm_force_performance_level(struct radeon_device *rdev,
+					enum radeon_dpm_forced_level level)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps *rps = &pi->current_rps;
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	int i, ret;
+
+	if (ps->num_levels <= 1)
+		return 0;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		/* not supported by the hw */
+		return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		ret = trinity_dpm_n_levels_disabled(rdev, ps->num_levels - 1);
+		if (ret)
+			return ret;
+	} else {
+		for (i = 0; i < ps->num_levels; i++) {
+			ret = trinity_dpm_n_levels_disabled(rdev, 0);
+			if (ret)
+				return ret;
+		}
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+int trinity_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	trinity_update_requested_ps(rdev, new_ps);
+
+	trinity_apply_state_adjust_rules(rdev,
+					 &pi->requested_rps,
+					 &pi->current_rps);
+
+	return 0;
+}
+
+int trinity_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+	struct radeon_ps *old_ps = &pi->current_rps;
+
+	trinity_acquire_mutex(rdev);
+	if (pi->enable_dpm) {
+		trinity_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+		trinity_enable_power_level_0(rdev);
+		trinity_force_level_0(rdev);
+		trinity_wait_for_level_0(rdev);
+		trinity_setup_nbp_sim(rdev, new_ps);
+		trinity_program_power_levels_0_to_n(rdev, new_ps, old_ps);
+		trinity_force_level_0(rdev);
+		trinity_unforce_levels(rdev);
+		trinity_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+		rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO;
+	}
+	trinity_release_mutex(rdev);
+
+	return 0;
+}
+
+void trinity_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+
+	trinity_update_current_ps(rdev, new_ps);
+}
+
+void trinity_dpm_setup_asic(struct radeon_device *rdev)
+{
+	trinity_acquire_mutex(rdev);
+	sumo_program_sstp(rdev);
+	sumo_take_smu_control(rdev, true);
+	trinity_get_min_sclk_divider(rdev);
+	trinity_release_mutex(rdev);
+}
+
+void trinity_dpm_reset_asic(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	trinity_acquire_mutex(rdev);
+	if (pi->enable_dpm) {
+		trinity_enable_power_level_0(rdev);
+		trinity_force_level_0(rdev);
+		trinity_wait_for_level_0(rdev);
+		trinity_program_bootup_state(rdev);
+		trinity_force_level_0(rdev);
+		trinity_unforce_levels(rdev);
+	}
+	trinity_release_mutex(rdev);
+}
+
+static u16 trinity_convert_voltage_index_to_value(struct radeon_device *rdev,
+						  u32 vid_2bit)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 vid_7bit = sumo_convert_vid2_to_vid7(rdev, &pi->sys_info.vid_mapping_table, vid_2bit);
+	u32 svi_mode = (RREG32_SMC(PM_CONFIG) & SVI_Mode) ? 1 : 0;
+	u32 step = (svi_mode == 0) ? 1250 : 625;
+	u32 delta = vid_7bit * step + 50;
+
+	if (delta > 155000)
+		return 0;
+
+	return (155000 - delta) / 100;
+}
+
+static void trinity_patch_boot_state(struct radeon_device *rdev,
+				     struct trinity_ps *ps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	ps->num_levels = 1;
+	ps->nbps_flags = 0;
+	ps->bapm_flags = 0;
+	ps->levels[0] = pi->boot_pl;
+}
+
+static u8 trinity_calculate_vce_wm(struct radeon_device *rdev, u32 sclk)
+{
+	if (sclk < 20000)
+		return 1;
+	return 0;
+}
+
+static void trinity_construct_boot_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->boot_pl.sclk = pi->sys_info.bootup_sclk;
+	pi->boot_pl.vddc_index = pi->sys_info.bootup_nb_voltage_index;
+	pi->boot_pl.ds_divider_index = 0;
+	pi->boot_pl.ss_divider_index = 0;
+	pi->boot_pl.allow_gnb_slow = 1;
+	pi->boot_pl.force_nbp_state = 0;
+	pi->boot_pl.display_wm = 0;
+	pi->boot_pl.vce_wm = 0;
+	pi->current_ps.num_levels = 1;
+	pi->current_ps.levels[0] = pi->boot_pl;
+}
+
+static u8 trinity_get_sleep_divider_id_from_clock(struct radeon_device *rdev,
+						  u32 sclk, u32 min_sclk_in_sr)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i;
+	u32 temp;
+	u32 min = (min_sclk_in_sr > TRINITY_MINIMUM_ENGINE_CLOCK) ?
+		min_sclk_in_sr : TRINITY_MINIMUM_ENGINE_CLOCK;
+
+	if (sclk < min)
+		return 0;
+
+	if (!pi->enable_sclk_ds)
+		return 0;
+
+	for (i = TRINITY_MAX_DEEPSLEEP_DIVIDER_ID;  ; i--) {
+		temp = sclk / sumo_get_sleep_divider_from_id(i);
+		if (temp >= min || i == 0)
+			break;
+	}
+
+	return (u8)i;
+}
+
+static u32 trinity_get_valid_engine_clock(struct radeon_device *rdev,
+					  u32 lower_limit)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i;
+
+	for (i = 0; i < pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries; i++) {
+		if (pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency >= lower_limit)
+			return pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency;
+	}
+
+	if (i == pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries)
+		DRM_ERROR("engine clock out of range!");
+
+	return 0;
+}
+
+static void trinity_patch_thermal_state(struct radeon_device *rdev,
+					struct trinity_ps *ps,
+					struct trinity_ps *current_ps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 current_vddc;
+	u32 current_sclk;
+	u32 current_index = 0;
+
+	if (current_ps) {
+		current_vddc = current_ps->levels[current_index].vddc_index;
+		current_sclk = current_ps->levels[current_index].sclk;
+	} else {
+		current_vddc = pi->boot_pl.vddc_index;
+		current_sclk = pi->boot_pl.sclk;
+	}
+
+	ps->levels[0].vddc_index = current_vddc;
+
+	if (ps->levels[0].sclk > current_sclk)
+		ps->levels[0].sclk = current_sclk;
+
+	ps->levels[0].ds_divider_index =
+		trinity_get_sleep_divider_id_from_clock(rdev, ps->levels[0].sclk, sclk_in_sr);
+	ps->levels[0].ss_divider_index = ps->levels[0].ds_divider_index;
+	ps->levels[0].allow_gnb_slow = 1;
+	ps->levels[0].force_nbp_state = 0;
+	ps->levels[0].display_wm = 0;
+	ps->levels[0].vce_wm =
+		trinity_calculate_vce_wm(rdev, ps->levels[0].sclk);
+}
+
+static u8 trinity_calculate_display_wm(struct radeon_device *rdev,
+				       struct trinity_ps *ps, u32 index)
+{
+	if (ps == NULL || ps->num_levels <= 1)
+		return 0;
+	else if (ps->num_levels == 2) {
+		if (index == 0)
+			return 0;
+		else
+			return 1;
+	} else {
+		if (index == 0)
+			return 0;
+		else if (ps->levels[index].sclk < 30000)
+			return 0;
+		else
+			return 1;
+	}
+}
+
+static u32 trinity_get_uvd_clock_index(struct radeon_device *rdev,
+				       struct radeon_ps *rps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < 4; i++) {
+		if ((rps->vclk == pi->sys_info.uvd_clock_table_entries[i].vclk) &&
+		    (rps->dclk == pi->sys_info.uvd_clock_table_entries[i].dclk))
+		    break;
+	}
+
+	if (i >= 4) {
+		DRM_ERROR("UVD clock index not found!\n");
+		i = 3;
+	}
+	return i;
+}
+
+static void trinity_adjust_uvd_state(struct radeon_device *rdev,
+				     struct radeon_ps *rps)
+{
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 high_index = 0;
+	u32 low_index = 0;
+
+	if (pi->uvd_dpm && r600_is_uvd_state(rps->class, rps->class2)) {
+		high_index = trinity_get_uvd_clock_index(rdev, rps);
+
+		switch(high_index) {
+		case 3:
+		case 2:
+			low_index = 1;
+			break;
+		case 1:
+		case 0:
+		default:
+			low_index = 0;
+			break;
+		}
+
+		ps->vclk_low_divider =
+			pi->sys_info.uvd_clock_table_entries[high_index].vclk_did;
+		ps->dclk_low_divider =
+			pi->sys_info.uvd_clock_table_entries[high_index].dclk_did;
+		ps->vclk_high_divider =
+			pi->sys_info.uvd_clock_table_entries[low_index].vclk_did;
+		ps->dclk_high_divider =
+			pi->sys_info.uvd_clock_table_entries[low_index].dclk_did;
+	}
+}
+
+
+
+static void trinity_apply_state_adjust_rules(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps)
+{
+	struct trinity_ps *ps = trinity_get_ps(new_rps);
+	struct trinity_ps *current_ps = trinity_get_ps(old_rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 min_voltage = 0; /* ??? */
+	u32 min_sclk = pi->sys_info.min_sclk; /* XXX check against disp reqs */
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 i;
+	bool force_high;
+	u32 num_active_displays = rdev->pm.dpm.new_active_crtc_count;
+
+	if (new_rps->class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+		return trinity_patch_thermal_state(rdev, ps, current_ps);
+
+	trinity_adjust_uvd_state(rdev, new_rps);
+
+	for (i = 0; i < ps->num_levels; i++) {
+		if (ps->levels[i].vddc_index < min_voltage)
+			ps->levels[i].vddc_index = min_voltage;
+
+		if (ps->levels[i].sclk < min_sclk)
+			ps->levels[i].sclk =
+				trinity_get_valid_engine_clock(rdev, min_sclk);
+
+		ps->levels[i].ds_divider_index =
+			sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[i].sclk, sclk_in_sr);
+
+		ps->levels[i].ss_divider_index = ps->levels[i].ds_divider_index;
+
+		ps->levels[i].allow_gnb_slow = 1;
+		ps->levels[i].force_nbp_state = 0;
+		ps->levels[i].display_wm =
+			trinity_calculate_display_wm(rdev, ps, i);
+		ps->levels[i].vce_wm =
+			trinity_calculate_vce_wm(rdev, ps->levels[0].sclk);
+	}
+
+	if ((new_rps->class & (ATOM_PPLIB_CLASSIFICATION_HDSTATE | ATOM_PPLIB_CLASSIFICATION_SDSTATE)) ||
+	    ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY))
+		ps->bapm_flags |= TRINITY_POWERSTATE_FLAGS_BAPM_DISABLE;
+
+	if (pi->sys_info.nb_dpm_enable) {
+		ps->Dpm0PgNbPsLo = 0x1;
+		ps->Dpm0PgNbPsHi = 0x0;
+		ps->DpmXNbPsLo = 0x2;
+		ps->DpmXNbPsHi = 0x1;
+
+		if ((new_rps->class & (ATOM_PPLIB_CLASSIFICATION_HDSTATE | ATOM_PPLIB_CLASSIFICATION_SDSTATE)) ||
+		    ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY)) {
+			force_high = ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_HDSTATE) ||
+				      ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE) &&
+				       (pi->sys_info.uma_channel_number == 1)));
+			force_high = (num_active_displays >= 3) || force_high;
+			ps->Dpm0PgNbPsLo = force_high ? 0x2 : 0x3;
+			ps->Dpm0PgNbPsHi = 0x1;
+			ps->DpmXNbPsLo = force_high ? 0x2 : 0x3;
+			ps->DpmXNbPsHi = 0x2;
+			ps->levels[ps->num_levels - 1].allow_gnb_slow = 0;
+		}
+	}
+}
+
+static void trinity_cleanup_asic(struct radeon_device *rdev)
+{
+	sumo_take_smu_control(rdev, false);
+}
+
+#if 0
+static void trinity_pre_display_configuration_change(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->voltage_drop_in_dce)
+		trinity_dce_enable_voltage_adjustment(rdev, false);
+}
+#endif
+
+static void trinity_add_dccac_value(struct radeon_device *rdev)
+{
+	u32 gpu_cac_avrg_cntl_window_size;
+	u32 num_active_displays = rdev->pm.dpm.new_active_crtc_count;
+	u64 disp_clk = rdev->clock.default_dispclk / 100;
+	u32 dc_cac_value;
+
+	gpu_cac_avrg_cntl_window_size =
+		(RREG32_SMC(GPU_CAC_AVRG_CNTL) & WINDOW_SIZE_MASK) >> WINDOW_SIZE_SHIFT;
+
+	dc_cac_value = (u32)((14213 * disp_clk * disp_clk * (u64)num_active_displays) >>
+			     (32 - gpu_cac_avrg_cntl_window_size));
+
+	WREG32_SMC(DC_CAC_VALUE, dc_cac_value);
+}
+
+void trinity_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->voltage_drop_in_dce)
+		trinity_dce_enable_voltage_adjustment(rdev, true);
+	trinity_add_dccac_value(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void trinity_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					       struct radeon_ps *rps,
+					       struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					       u8 table_rev)
+{
+	struct trinity_ps *ps = trinity_get_ps(rps);
+
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		rdev->pm.dpm.boot_ps = rps;
+		trinity_patch_boot_state(rdev, ps);
+	}
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void trinity_parse_pplib_clock_info(struct radeon_device *rdev,
+					   struct radeon_ps *rps, int index,
+					   union pplib_clock_info *clock_info)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	struct trinity_pl *pl = &ps->levels[index];
+	u32 sclk;
+
+	sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow);
+	sclk |= clock_info->sumo.ucEngineClockHigh << 16;
+	pl->sclk = sclk;
+	pl->vddc_index = clock_info->sumo.vddcIndex;
+
+	ps->num_levels = index + 1;
+
+	if (pi->enable_sclk_ds) {
+		pl->ds_divider_index = 5;
+		pl->ss_divider_index = 5;
+	}
+}
+
+static int trinity_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j, k, non_clock_array_index, clock_array_index;
+	union pplib_clock_info *clock_info;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	u8 *power_state_offset;
+	struct sumo_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	state_array = (struct _StateArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
+	clock_info_array = (struct _ClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
+	non_clock_info_array = (struct _NonClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  state_array->ucNumEntries, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	power_state_offset = (u8 *)state_array->states;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+	for (i = 0; i < state_array->ucNumEntries; i++) {
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			&non_clock_info_array->nonClockInfo[non_clock_array_index];
+		if (!rdev->pm.power_state[i].clock_info)
+			return -EINVAL;
+		ps = kzalloc(sizeof(struct sumo_ps), GFP_KERNEL);
+		if (ps == NULL) {
+			kfree(rdev->pm.dpm.ps);
+			return -ENOMEM;
+		}
+		rdev->pm.dpm.ps[i].ps_priv = ps;
+		k = 0;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			if (clock_array_index >= clock_info_array->ucNumEntries)
+				continue;
+			if (k >= SUMO_MAX_HARDWARE_POWERLEVELS)
+				break;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
+			trinity_parse_pplib_clock_info(rdev,
+						       &rdev->pm.dpm.ps[i], k,
+						       clock_info);
+			k++;
+		}
+		trinity_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+						   non_clock_info,
+						   non_clock_info_array->ucEntrySize);
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
+	}
+	rdev->pm.dpm.num_ps = state_array->ucNumEntries;
+	return 0;
+}
+
+union igp_info {
+	struct _ATOM_INTEGRATED_SYSTEM_INFO info;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V5 info_5;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7 info_7;
+};
+
+static u32 trinity_convert_did_to_freq(struct radeon_device *rdev, u8 did)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 divider;
+
+	if (did >= 8 && did <= 0x3f)
+		divider = did * 25;
+	else if (did > 0x3f && did <= 0x5f)
+		divider = (did - 64) * 50 + 1600;
+	else if (did > 0x5f && did <= 0x7e)
+		divider = (did - 96) * 100 + 3200;
+	else if (did == 0x7f)
+		divider = 128 * 100;
+	else
+		return 10000;
+
+	return ((pi->sys_info.dentist_vco_freq * 100) + (divider - 1)) / divider;
+}
+
+static int trinity_parse_sys_info_table(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo);
+	union igp_info *igp_info;
+	u8 frev, crev;
+	u16 data_offset;
+	int i;
+
+	if (atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		igp_info = (union igp_info *)(mode_info->atom_context->bios +
+					      data_offset);
+
+		if (crev != 7) {
+			DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev);
+			return -EINVAL;
+		}
+		pi->sys_info.bootup_sclk = le32_to_cpu(igp_info->info_7.ulBootUpEngineClock);
+		pi->sys_info.min_sclk = le32_to_cpu(igp_info->info_7.ulMinEngineClock);
+		pi->sys_info.bootup_uma_clk = le32_to_cpu(igp_info->info_7.ulBootUpUMAClock);
+		pi->sys_info.dentist_vco_freq = le32_to_cpu(igp_info->info_7.ulDentistVCOFreq);
+		pi->sys_info.bootup_nb_voltage_index =
+			le16_to_cpu(igp_info->info_7.usBootUpNBVoltage);
+		if (igp_info->info_7.ucHtcTmpLmt == 0)
+			pi->sys_info.htc_tmp_lmt = 203;
+		else
+			pi->sys_info.htc_tmp_lmt = igp_info->info_7.ucHtcTmpLmt;
+		if (igp_info->info_7.ucHtcHystLmt == 0)
+			pi->sys_info.htc_hyst_lmt = 5;
+		else
+			pi->sys_info.htc_hyst_lmt = igp_info->info_7.ucHtcHystLmt;
+		if (pi->sys_info.htc_tmp_lmt <= pi->sys_info.htc_hyst_lmt) {
+			DRM_ERROR("The htcTmpLmt should be larger than htcHystLmt.\n");
+		}
+
+		if (pi->enable_nbps_policy)
+			pi->sys_info.nb_dpm_enable = igp_info->info_7.ucNBDPMEnable;
+		else
+			pi->sys_info.nb_dpm_enable = 0;
+
+		for (i = 0; i < TRINITY_NUM_NBPSTATES; i++) {
+			pi->sys_info.nbp_mclk[i] = le32_to_cpu(igp_info->info_7.ulNbpStateMemclkFreq[i]);
+			pi->sys_info.nbp_nclk[i] = le32_to_cpu(igp_info->info_7.ulNbpStateNClkFreq[i]);
+		}
+
+		pi->sys_info.nbp_voltage_index[0] = le16_to_cpu(igp_info->info_7.usNBP0Voltage);
+		pi->sys_info.nbp_voltage_index[1] = le16_to_cpu(igp_info->info_7.usNBP1Voltage);
+		pi->sys_info.nbp_voltage_index[2] = le16_to_cpu(igp_info->info_7.usNBP2Voltage);
+		pi->sys_info.nbp_voltage_index[3] = le16_to_cpu(igp_info->info_7.usNBP3Voltage);
+
+		if (!pi->sys_info.nb_dpm_enable) {
+			for (i = 1; i < TRINITY_NUM_NBPSTATES; i++) {
+				pi->sys_info.nbp_mclk[i] = pi->sys_info.nbp_mclk[0];
+				pi->sys_info.nbp_nclk[i] = pi->sys_info.nbp_nclk[0];
+				pi->sys_info.nbp_voltage_index[i] = pi->sys_info.nbp_voltage_index[0];
+			}
+		}
+
+		pi->sys_info.uma_channel_number = igp_info->info_7.ucUMAChannelNumber;
+
+		sumo_construct_sclk_voltage_mapping_table(rdev,
+							  &pi->sys_info.sclk_voltage_mapping_table,
+							  igp_info->info_7.sAvail_SCLK);
+		sumo_construct_vid_mapping_table(rdev, &pi->sys_info.vid_mapping_table,
+						 igp_info->info_7.sAvail_SCLK);
+
+		pi->sys_info.uvd_clock_table_entries[0].vclk_did =
+			igp_info->info_7.ucDPMState0VclkFid;
+		pi->sys_info.uvd_clock_table_entries[1].vclk_did =
+			igp_info->info_7.ucDPMState1VclkFid;
+		pi->sys_info.uvd_clock_table_entries[2].vclk_did =
+			igp_info->info_7.ucDPMState2VclkFid;
+		pi->sys_info.uvd_clock_table_entries[3].vclk_did =
+			igp_info->info_7.ucDPMState3VclkFid;
+
+		pi->sys_info.uvd_clock_table_entries[0].dclk_did =
+			igp_info->info_7.ucDPMState0DclkFid;
+		pi->sys_info.uvd_clock_table_entries[1].dclk_did =
+			igp_info->info_7.ucDPMState1DclkFid;
+		pi->sys_info.uvd_clock_table_entries[2].dclk_did =
+			igp_info->info_7.ucDPMState2DclkFid;
+		pi->sys_info.uvd_clock_table_entries[3].dclk_did =
+			igp_info->info_7.ucDPMState3DclkFid;
+
+		for (i = 0; i < 4; i++) {
+			pi->sys_info.uvd_clock_table_entries[i].vclk =
+				trinity_convert_did_to_freq(rdev,
+							    pi->sys_info.uvd_clock_table_entries[i].vclk_did);
+			pi->sys_info.uvd_clock_table_entries[i].dclk =
+				trinity_convert_did_to_freq(rdev,
+							    pi->sys_info.uvd_clock_table_entries[i].dclk_did);
+		}
+
+
+
+	}
+	return 0;
+}
+
+int trinity_dpm_init(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi;
+	int ret, i;
+
+	pi = kzalloc(sizeof(struct trinity_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++)
+		pi->at[i] = TRINITY_AT_DFLT;
+
+	pi->enable_nbps_policy = true;
+	pi->enable_sclk_ds = true;
+	pi->enable_gfx_power_gating = true;
+	pi->enable_gfx_clock_gating = true;
+	pi->enable_mg_clock_gating = true;
+	pi->enable_gfx_dynamic_mgpg = true; /* ??? */
+	pi->override_dynamic_mgpg = true;
+	pi->enable_auto_thermal_throttling = true;
+	pi->voltage_drop_in_dce = false; /* need to restructure dpm/modeset interaction */
+	pi->uvd_dpm = true; /* ??? */
+
+	ret = trinity_parse_sys_info_table(rdev);
+	if (ret)
+		return ret;
+
+	trinity_construct_boot_state(rdev);
+
+	ret = trinity_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	pi->thermal_auto_throttling = pi->sys_info.htc_tmp_lmt;
+	pi->enable_dpm = true;
+
+	return 0;
+}
+
+void trinity_dpm_print_power_state(struct radeon_device *rdev,
+				   struct radeon_ps *rps)
+{
+	int i;
+	struct trinity_ps *ps = trinity_get_ps(rps);
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	for (i = 0; i < ps->num_levels; i++) {
+		struct trinity_pl *pl = &ps->levels[i];
+		printk("\t\tpower level %d    sclk: %u vddc: %u\n",
+		       i, pl->sclk,
+		       trinity_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void trinity_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+							 struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	struct trinity_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_MASK) >>
+		CURRENT_STATE_SHIFT;
+
+	if (current_index >= ps->num_levels) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
+			   current_index, pl->sclk,
+			   trinity_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+}
+
+void trinity_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	trinity_cleanup_asic(rdev); /* ??? */
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 trinity_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct trinity_ps *requested_state = trinity_get_ps(&pi->requested_rps);
+
+	if (low)
+		return requested_state->levels[0].sclk;
+	else
+		return requested_state->levels[requested_state->num_levels - 1].sclk;
+}
+
+u32 trinity_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	return pi->sys_info.bootup_uma_clk;
+}
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.h b/drivers/gpu/drm/radeon/trinity_dpm.h
new file mode 100644
index 0000000..e82df07
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinity_dpm.h
@@ -0,0 +1,132 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __TRINITY_DPM_H__
+#define __TRINITY_DPM_H__
+
+#include "sumo_dpm.h"
+
+#define TRINITY_SIZEOF_DPM_STATE_TABLE (SMU_SCLK_DPM_STATE_1_CNTL_0 - SMU_SCLK_DPM_STATE_0_CNTL_0)
+
+struct trinity_pl {
+	u32 sclk;
+	u8 vddc_index;
+	u8 ds_divider_index;
+	u8 ss_divider_index;
+	u8 allow_gnb_slow;
+	u8 force_nbp_state;
+	u8 display_wm;
+	u8 vce_wm;
+};
+
+#define TRINITY_POWERSTATE_FLAGS_NBPS_FORCEHIGH  (1 << 0)
+#define TRINITY_POWERSTATE_FLAGS_NBPS_LOCKTOHIGH (1 << 1)
+#define TRINITY_POWERSTATE_FLAGS_NBPS_LOCKTOLOW  (1 << 2)
+
+#define TRINITY_POWERSTATE_FLAGS_BAPM_DISABLE    (1 << 0)
+
+struct trinity_ps {
+	u32 num_levels;
+	struct trinity_pl levels[SUMO_MAX_HARDWARE_POWERLEVELS];
+
+	u32 nbps_flags;
+	u32 bapm_flags;
+
+	u8 Dpm0PgNbPsLo;
+	u8 Dpm0PgNbPsHi;
+	u8 DpmXNbPsLo;
+	u8 DpmXNbPsHi;
+
+	u32 vclk_low_divider;
+	u32 vclk_high_divider;
+	u32 dclk_low_divider;
+	u32 dclk_high_divider;
+};
+
+#define TRINITY_NUM_NBPSTATES   4
+
+struct trinity_uvd_clock_table_entry
+{
+	u32 vclk;
+	u32 dclk;
+	u8 vclk_did;
+	u8 dclk_did;
+	u8 rsv[2];
+};
+
+struct trinity_sys_info {
+	u32 bootup_uma_clk;
+	u32 bootup_sclk;
+	u32 min_sclk;
+	u32 dentist_vco_freq;
+	u32 nb_dpm_enable;
+	u32 nbp_mclk[TRINITY_NUM_NBPSTATES];
+	u32 nbp_nclk[TRINITY_NUM_NBPSTATES];
+	u16 nbp_voltage_index[TRINITY_NUM_NBPSTATES];
+	u16 bootup_nb_voltage_index;
+	u8 htc_tmp_lmt;
+	u8 htc_hyst_lmt;
+	struct sumo_sclk_voltage_mapping_table sclk_voltage_mapping_table;
+	struct sumo_vid_mapping_table vid_mapping_table;
+	u32 uma_channel_number;
+	struct trinity_uvd_clock_table_entry uvd_clock_table_entries[4];
+};
+
+struct trinity_power_info {
+	u32 at[SUMO_MAX_HARDWARE_POWERLEVELS];
+	u32 dpm_interval;
+	u32 thermal_auto_throttling;
+	struct trinity_sys_info sys_info;
+	struct trinity_pl boot_pl;
+	u32 min_sclk_did;
+	bool enable_nbps_policy;
+	bool voltage_drop_in_dce;
+	bool override_dynamic_mgpg;
+	bool enable_gfx_clock_gating;
+	bool enable_gfx_power_gating;
+	bool enable_mg_clock_gating;
+	bool enable_gfx_dynamic_mgpg;
+	bool enable_auto_thermal_throttling;
+	bool enable_dpm;
+	bool enable_sclk_ds;
+	bool uvd_dpm;
+	struct radeon_ps current_rps;
+	struct trinity_ps current_ps;
+	struct radeon_ps requested_rps;
+	struct trinity_ps requested_ps;
+};
+
+#define TRINITY_AT_DFLT            30
+
+/* trinity_smc.c */
+int trinity_dpm_config(struct radeon_device *rdev, bool enable);
+int trinity_uvd_dpm_config(struct radeon_device *rdev);
+int trinity_dpm_force_state(struct radeon_device *rdev, u32 n);
+int trinity_dpm_n_levels_disabled(struct radeon_device *rdev, u32 n);
+int trinity_dpm_no_forced_level(struct radeon_device *rdev);
+int trinity_dce_enable_voltage_adjustment(struct radeon_device *rdev,
+					  bool enable);
+int trinity_gfx_dynamic_mgpg_config(struct radeon_device *rdev);
+void trinity_acquire_mutex(struct radeon_device *rdev);
+void trinity_release_mutex(struct radeon_device *rdev);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/trinity_smc.c b/drivers/gpu/drm/radeon/trinity_smc.c
new file mode 100644
index 0000000..a42d89f
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinity_smc.c
@@ -0,0 +1,122 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "trinityd.h"
+#include "trinity_dpm.h"
+#include "ppsmc.h"
+
+struct trinity_ps *trinity_get_ps(struct radeon_ps *rps);
+struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev);
+
+static int trinity_notify_message_to_smu(struct radeon_device *rdev, u32 id)
+{
+	int i;
+	u32 v = 0;
+
+	WREG32(SMC_MESSAGE_0, id);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SMC_RESP_0) != 0)
+			break;
+		udelay(1);
+	}
+	v = RREG32(SMC_RESP_0);
+
+	if (v != 1) {
+		if (v == 0xFF) {
+			DRM_ERROR("SMC failed to handle the message!\n");
+			return -EINVAL;
+		} else if (v == 0xFE) {
+			DRM_ERROR("Unknown SMC message!\n");
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+int trinity_dpm_config(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_SMC(SMU_SCRATCH0, 1);
+	else
+		WREG32_SMC(SMU_SCRATCH0, 0);
+
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DPM_Config);
+}
+
+int trinity_dpm_force_state(struct radeon_device *rdev, u32 n)
+{
+	WREG32_SMC(SMU_SCRATCH0, n);
+
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DPM_ForceState);
+}
+
+int trinity_dpm_n_levels_disabled(struct radeon_device *rdev, u32 n)
+{
+	WREG32_SMC(SMU_SCRATCH0, n);
+
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DPM_N_LevelsDisabled);
+}
+
+int trinity_uvd_dpm_config(struct radeon_device *rdev)
+{
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_UVD_DPM_Config);
+}
+
+int trinity_dpm_no_forced_level(struct radeon_device *rdev)
+{
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_NoForcedLevel);
+}
+
+int trinity_dce_enable_voltage_adjustment(struct radeon_device *rdev,
+					  bool enable)
+{
+	if (enable)
+		return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DCE_AllowVoltageAdjustment);
+	else
+		return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DCE_RemoveVoltageAdjustment);
+}
+
+int trinity_gfx_dynamic_mgpg_config(struct radeon_device *rdev)
+{
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_PG_SIMD_Config);
+}
+
+void trinity_acquire_mutex(struct radeon_device *rdev)
+{
+	int i;
+
+	WREG32(SMC_INT_REQ, 1);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(SMC_INT_REQ) & 0xffff) == 1)
+			break;
+		udelay(1);
+	}
+}
+
+void trinity_release_mutex(struct radeon_device *rdev)
+{
+	WREG32(SMC_INT_REQ, 0);
+}
diff --git a/drivers/gpu/drm/radeon/trinityd.h b/drivers/gpu/drm/radeon/trinityd.h
new file mode 100644
index 0000000..fd32e27
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinityd.h
@@ -0,0 +1,228 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef _TRINITYD_H_
+#define _TRINITYD_H_
+
+/* pm registers */
+
+/* cg */
+#define CG_CGTT_LOCAL_0                                 0x0
+#define CG_CGTT_LOCAL_1                                 0x1
+
+/* smc */
+#define SMU_SCLK_DPM_STATE_0_CNTL_0                     0x1f000
+#       define STATE_VALID(x)                           ((x) << 0)
+#       define STATE_VALID_MASK                         (0xff << 0)
+#       define STATE_VALID_SHIFT                        0
+#       define CLK_DIVIDER(x)                           ((x) << 8)
+#       define CLK_DIVIDER_MASK                         (0xff << 8)
+#       define CLK_DIVIDER_SHIFT                        8
+#       define VID(x)                                   ((x) << 16)
+#       define VID_MASK                                 (0xff << 16)
+#       define VID_SHIFT                                16
+#       define LVRT(x)                                  ((x) << 24)
+#       define LVRT_MASK                                (0xff << 24)
+#       define LVRT_SHIFT                               24
+#define SMU_SCLK_DPM_STATE_0_CNTL_1                     0x1f004
+#       define DS_DIV(x)                                ((x) << 0)
+#       define DS_DIV_MASK                              (0xff << 0)
+#       define DS_DIV_SHIFT                             0
+#       define DS_SH_DIV(x)                             ((x) << 8)
+#       define DS_SH_DIV_MASK                           (0xff << 8)
+#       define DS_SH_DIV_SHIFT                          8
+#       define DISPLAY_WM(x)                            ((x) << 16)
+#       define DISPLAY_WM_MASK                          (0xff << 16)
+#       define DISPLAY_WM_SHIFT                         16
+#       define VCE_WM(x)                                ((x) << 24)
+#       define VCE_WM_MASK                              (0xff << 24)
+#       define VCE_WM_SHIFT                             24
+
+#define SMU_SCLK_DPM_STATE_0_CNTL_3                     0x1f00c
+#       define GNB_SLOW(x)                              ((x) << 0)
+#       define GNB_SLOW_MASK                            (0xff << 0)
+#       define GNB_SLOW_SHIFT                           0
+#       define FORCE_NBPS1(x)                           ((x) << 8)
+#       define FORCE_NBPS1_MASK                         (0xff << 8)
+#       define FORCE_NBPS1_SHIFT                        8
+#define SMU_SCLK_DPM_STATE_0_AT                         0x1f010
+#       define AT(x)                                    ((x) << 0)
+#       define AT_MASK                                  (0xff << 0)
+#       define AT_SHIFT                                 0
+
+#define SMU_SCLK_DPM_STATE_0_PG_CNTL                    0x1f014
+#       define PD_SCLK_DIVIDER(x)                       ((x) << 16)
+#       define PD_SCLK_DIVIDER_MASK                     (0xff << 16)
+#       define PD_SCLK_DIVIDER_SHIFT                    16
+
+#define SMU_SCLK_DPM_STATE_1_CNTL_0                     0x1f020
+
+#define SMU_SCLK_DPM_CNTL                               0x1f100
+#       define SCLK_DPM_EN(x)                           ((x) << 0)
+#       define SCLK_DPM_EN_MASK                         (0xff << 0)
+#       define SCLK_DPM_EN_SHIFT                        0
+#       define SCLK_DPM_BOOT_STATE(x)                   ((x) << 16)
+#       define SCLK_DPM_BOOT_STATE_MASK                 (0xff << 16)
+#       define SCLK_DPM_BOOT_STATE_SHIFT                16
+#       define VOLTAGE_CHG_EN(x)                        ((x) << 24)
+#       define VOLTAGE_CHG_EN_MASK                      (0xff << 24)
+#       define VOLTAGE_CHG_EN_SHIFT                     24
+
+#define SMU_SCLK_DPM_TT_CNTL                            0x1f108
+#       define SCLK_TT_EN(x)                            ((x) << 0)
+#       define SCLK_TT_EN_MASK                          (0xff << 0)
+#       define SCLK_TT_EN_SHIFT                         0
+#define SMU_SCLK_DPM_TTT                                0x1f10c
+#       define LT(x)                                    ((x) << 0)
+#       define LT_MASK                                  (0xffff << 0)
+#       define LT_SHIFT                                 0
+#       define HT(x)                                    ((x) << 16)
+#       define HT_MASK                                  (0xffff << 16)
+#       define HT_SHIFT                                 16
+
+#define SMU_UVD_DPM_STATES                              0x1f1a0
+#define SMU_UVD_DPM_CNTL                                0x1f1a4
+
+#define SMU_S_PG_CNTL                                   0x1f118
+#       define DS_PG_EN(x)                              ((x) << 16)
+#       define DS_PG_EN_MASK                            (0xff << 16)
+#       define DS_PG_EN_SHIFT                           16
+
+#define GFX_POWER_GATING_CNTL                           0x1f38c
+#       define PDS_DIV(x)                               ((x) << 0)
+#       define PDS_DIV_MASK                             (0xff << 0)
+#       define PDS_DIV_SHIFT                            0
+#       define SSSD(x)                                  ((x) << 8)
+#       define SSSD_MASK                                (0xff << 8)
+#       define SSSD_SHIFT                               8
+
+#define PM_CONFIG                                       0x1f428
+#       define SVI_Mode                                 (1 << 29)
+
+#define PM_I_CNTL_1                                     0x1f464
+#       define SCLK_DPM(x)                              ((x) << 0)
+#       define SCLK_DPM_MASK                            (0xff << 0)
+#       define SCLK_DPM_SHIFT                           0
+#       define DS_PG_CNTL(x)                            ((x) << 16)
+#       define DS_PG_CNTL_MASK                          (0xff << 16)
+#       define DS_PG_CNTL_SHIFT                         16
+#define PM_TP                                           0x1f468
+
+#define NB_PSTATE_CONFIG                                0x1f5f8
+#       define Dpm0PgNbPsLo(x)                          ((x) << 0)
+#       define Dpm0PgNbPsLo_MASK                        (3 << 0)
+#       define Dpm0PgNbPsLo_SHIFT                       0
+#       define Dpm0PgNbPsHi(x)                          ((x) << 2)
+#       define Dpm0PgNbPsHi_MASK                        (3 << 2)
+#       define Dpm0PgNbPsHi_SHIFT                       2
+#       define DpmXNbPsLo(x)                            ((x) << 4)
+#       define DpmXNbPsLo_MASK                          (3 << 4)
+#       define DpmXNbPsLo_SHIFT                         4
+#       define DpmXNbPsHi(x)                            ((x) << 6)
+#       define DpmXNbPsHi_MASK                          (3 << 6)
+#       define DpmXNbPsHi_SHIFT                         6
+
+#define DC_CAC_VALUE                                    0x1f908
+
+#define GPU_CAC_AVRG_CNTL                               0x1f920
+#       define WINDOW_SIZE(x)                           ((x) << 0)
+#       define WINDOW_SIZE_MASK                         (0xff << 0)
+#       define WINDOW_SIZE_SHIFT                        0
+
+#define CC_SMU_MISC_FUSES                               0xe0001004
+#       define MinSClkDid(x)                   ((x) << 2)
+#       define MinSClkDid_MASK                 (0x7f << 2)
+#       define MinSClkDid_SHIFT                2
+
+#define CC_SMU_TST_EFUSE1_MISC                          0xe000101c
+#       define RB_BACKEND_DISABLE(x)                    ((x) << 16)
+#       define RB_BACKEND_DISABLE_MASK                  (3 << 16)
+#       define RB_BACKEND_DISABLE_SHIFT                 16
+
+#define SMU_SCRATCH_A                                   0xe0003024
+
+#define SMU_SCRATCH0                                    0xe0003040
+
+/* mmio */
+#define SMC_INT_REQ                                     0x220
+
+#define SMC_MESSAGE_0                                   0x22c
+#define SMC_RESP_0                                      0x230
+
+#define GENERAL_PWRMGT                                  0x670
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+
+#define SCLK_PWRMGT_CNTL                                0x678
+#       define DYN_PWR_DOWN_EN                          (1 << 2)
+#       define RESET_BUSY_CNT                           (1 << 4)
+#       define RESET_SCLK_CNT                           (1 << 5)
+#       define DYN_GFX_CLK_OFF_EN                       (1 << 7)
+#       define GFX_CLK_FORCE_ON                         (1 << 8)
+#       define DYNAMIC_PM_EN                            (1 << 21)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                0x684
+#       define TARGET_STATE(x)                          ((x) << 0)
+#       define TARGET_STATE_MASK                        (0xf << 0)
+#       define TARGET_STATE_SHIFT                       0
+#       define CURRENT_STATE(x)                         ((x) << 4)
+#       define CURRENT_STATE_MASK                       (0xf << 4)
+#       define CURRENT_STATE_SHIFT                      4
+
+#define CG_GIPOTS                                       0x6d8
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+#       define CG_GIPOT_SHIFT                           16
+
+#define CG_PG_CTRL                                      0x6e0
+#       define SP(x)                                    ((x) << 0)
+#       define SP_MASK                                  (0xffff << 0)
+#       define SP_SHIFT                                 0
+#       define SU(x)                                    ((x) << 16)
+#       define SU_MASK                                  (0xffff << 16)
+#       define SU_SHIFT                                 16
+
+#define CG_MISC_REG                                     0x708
+
+#define CG_THERMAL_INT_CTRL                             0x738
+#       define DIG_THERM_INTH(x)                        ((x) << 0)
+#       define DIG_THERM_INTH_MASK                      (0xff << 0)
+#       define DIG_THERM_INTH_SHIFT                     0
+#       define DIG_THERM_INTL(x)                        ((x) << 8)
+#       define DIG_THERM_INTL_MASK                      (0xff << 8)
+#       define DIG_THERM_INTL_SHIFT                     8
+#       define THERM_INTH_MASK                          (1 << 24)
+#       define THERM_INTL_MASK                          (1 << 25)
+
+#define CG_CG_VOLTAGE_CNTL                              0x770
+#       define EN                                       (1 << 9)
+
+#define HW_REV   					0x5564
+#       define ATI_REV_ID_MASK                          (0xf << 28)
+#       define ATI_REV_ID_SHIFT                         28
+/* 0 = A0, 1 = A1, 2 = B0, 3 = C0, etc. */
+
+#define CGTS_SM_CTRL_REG                                0x9150
+
+#define GB_ADDR_CONFIG                                  0x98f8
+
+#endif
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig
new file mode 100644
index 0000000..72887df
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/Kconfig
@@ -0,0 +1,9 @@
+config DRM_RCAR_DU
+	tristate "DRM Support for R-Car Display Unit"
+	depends on DRM && ARM
+	select DRM_KMS_HELPER
+	select DRM_KMS_CMA_HELPER
+	select DRM_GEM_CMA_HELPER
+	help
+	  Choose this option if you have an R-Car chipset.
+	  If M is selected the module will be called rcar-du-drm.
diff --git a/drivers/gpu/drm/rcar-du/Makefile b/drivers/gpu/drm/rcar-du/Makefile
new file mode 100644
index 0000000..7333c00
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/Makefile
@@ -0,0 +1,8 @@
+rcar-du-drm-y := rcar_du_crtc.o \
+		 rcar_du_drv.o \
+		 rcar_du_kms.o \
+		 rcar_du_lvds.o \
+		 rcar_du_plane.o \
+		 rcar_du_vga.o
+
+obj-$(CONFIG_DRM_RCAR_DU)	+= rcar-du-drm.o
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
new file mode 100644
index 0000000..24183fb
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
@@ -0,0 +1,595 @@
+/*
+ * rcar_du_crtc.c  --  R-Car Display Unit CRTCs
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/clk.h>
+#include <linux/mutex.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_lvds.h"
+#include "rcar_du_plane.h"
+#include "rcar_du_regs.h"
+#include "rcar_du_vga.h"
+
+#define to_rcar_crtc(c)	container_of(c, struct rcar_du_crtc, crtc)
+
+static u32 rcar_du_crtc_read(struct rcar_du_crtc *rcrtc, u32 reg)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	return rcar_du_read(rcdu, rcrtc->mmio_offset + reg);
+}
+
+static void rcar_du_crtc_write(struct rcar_du_crtc *rcrtc, u32 reg, u32 data)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg, data);
+}
+
+static void rcar_du_crtc_clr(struct rcar_du_crtc *rcrtc, u32 reg, u32 clr)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg,
+		      rcar_du_read(rcdu, rcrtc->mmio_offset + reg) & ~clr);
+}
+
+static void rcar_du_crtc_set(struct rcar_du_crtc *rcrtc, u32 reg, u32 set)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg,
+		      rcar_du_read(rcdu, rcrtc->mmio_offset + reg) | set);
+}
+
+static void rcar_du_crtc_clr_set(struct rcar_du_crtc *rcrtc, u32 reg,
+				 u32 clr, u32 set)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+	u32 value = rcar_du_read(rcdu, rcrtc->mmio_offset + reg);
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg, (value & ~clr) | set);
+}
+
+static void rcar_du_crtc_set_display_timing(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	const struct drm_display_mode *mode = &crtc->mode;
+	unsigned long clk;
+	u32 value;
+	u32 div;
+
+	/* Dot clock */
+	clk = clk_get_rate(rcdu->clock);
+	div = DIV_ROUND_CLOSEST(clk, mode->clock * 1000);
+	div = clamp(div, 1U, 64U) - 1;
+
+	rcar_du_write(rcdu, rcrtc->index ? ESCR2 : ESCR,
+		      ESCR_DCLKSEL_CLKS | div);
+	rcar_du_write(rcdu, rcrtc->index ? OTAR2 : OTAR, 0);
+
+	/* Signal polarities */
+	value = ((mode->flags & DRM_MODE_FLAG_PVSYNC) ? 0 : DSMR_VSL)
+	      | ((mode->flags & DRM_MODE_FLAG_PHSYNC) ? 0 : DSMR_HSL)
+	      | DSMR_DIPM_DE;
+	rcar_du_crtc_write(rcrtc, DSMR, value);
+
+	/* Display timings */
+	rcar_du_crtc_write(rcrtc, HDSR, mode->htotal - mode->hsync_start - 19);
+	rcar_du_crtc_write(rcrtc, HDER, mode->htotal - mode->hsync_start +
+					mode->hdisplay - 19);
+	rcar_du_crtc_write(rcrtc, HSWR, mode->hsync_end -
+					mode->hsync_start - 1);
+	rcar_du_crtc_write(rcrtc, HCR,  mode->htotal - 1);
+
+	rcar_du_crtc_write(rcrtc, VDSR, mode->vtotal - mode->vsync_end - 2);
+	rcar_du_crtc_write(rcrtc, VDER, mode->vtotal - mode->vsync_end +
+					mode->vdisplay - 2);
+	rcar_du_crtc_write(rcrtc, VSPR, mode->vtotal - mode->vsync_end +
+					mode->vsync_start - 1);
+	rcar_du_crtc_write(rcrtc, VCR,  mode->vtotal - 1);
+
+	rcar_du_crtc_write(rcrtc, DESR,  mode->htotal - mode->hsync_start);
+	rcar_du_crtc_write(rcrtc, DEWR,  mode->hdisplay);
+}
+
+static void rcar_du_crtc_set_routing(struct rcar_du_crtc *rcrtc)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+	u32 dorcr = rcar_du_read(rcdu, DORCR);
+
+	dorcr &= ~(DORCR_PG2T | DORCR_DK2S | DORCR_PG2D_MASK);
+
+	/* Set the DU1 pins sources. Select CRTC 0 if explicitly requested and
+	 * CRTC 1 in all other cases to avoid cloning CRTC 0 to DU0 and DU1 by
+	 * default.
+	 */
+	if (rcrtc->outputs & (1 << 1) && rcrtc->index == 0)
+		dorcr |= DORCR_PG2D_DS1;
+	else
+		dorcr |= DORCR_PG2T | DORCR_DK2S | DORCR_PG2D_DS2;
+
+	rcar_du_write(rcdu, DORCR, dorcr);
+}
+
+static void __rcar_du_start_stop(struct rcar_du_device *rcdu, bool start)
+{
+	rcar_du_write(rcdu, DSYSR,
+		      (rcar_du_read(rcdu, DSYSR) & ~(DSYSR_DRES | DSYSR_DEN)) |
+		      (start ? DSYSR_DEN : DSYSR_DRES));
+}
+
+static void rcar_du_start_stop(struct rcar_du_device *rcdu, bool start)
+{
+	/* Many of the configuration bits are only updated when the display
+	 * reset (DRES) bit in DSYSR is set to 1, disabling *both* CRTCs. Some
+	 * of those bits could be pre-configured, but others (especially the
+	 * bits related to plane assignment to display timing controllers) need
+	 * to be modified at runtime.
+	 *
+	 * Restart the display controller if a start is requested. Sorry for the
+	 * flicker. It should be possible to move most of the "DRES-update" bits
+	 * setup to driver initialization time and minimize the number of cases
+	 * when the display controller will have to be restarted.
+	 */
+	if (start) {
+		if (rcdu->used_crtcs++ != 0)
+			__rcar_du_start_stop(rcdu, false);
+		__rcar_du_start_stop(rcdu, true);
+	} else {
+		if (--rcdu->used_crtcs == 0)
+			__rcar_du_start_stop(rcdu, false);
+	}
+}
+
+void rcar_du_crtc_route_output(struct drm_crtc *crtc, unsigned int output)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	/* Store the route from the CRTC output to the DU output. The DU will be
+	 * configured when starting the CRTC.
+	 */
+	rcrtc->outputs |= 1 << output;
+}
+
+void rcar_du_crtc_update_planes(struct drm_crtc *crtc)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+	struct rcar_du_plane *planes[RCAR_DU_NUM_HW_PLANES];
+	unsigned int num_planes = 0;
+	unsigned int prio = 0;
+	unsigned int i;
+	u32 dptsr = 0;
+	u32 dspr = 0;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		struct rcar_du_plane *plane = &rcdu->planes.planes[i];
+		unsigned int j;
+
+		if (plane->crtc != &rcrtc->crtc || !plane->enabled)
+			continue;
+
+		/* Insert the plane in the sorted planes array. */
+		for (j = num_planes++; j > 0; --j) {
+			if (planes[j-1]->zpos <= plane->zpos)
+				break;
+			planes[j] = planes[j-1];
+		}
+
+		planes[j] = plane;
+		prio += plane->format->planes * 4;
+	}
+
+	for (i = 0; i < num_planes; ++i) {
+		struct rcar_du_plane *plane = planes[i];
+		unsigned int index = plane->hwindex;
+
+		prio -= 4;
+		dspr |= (index + 1) << prio;
+		dptsr |= DPTSR_PnDK(index) |  DPTSR_PnTS(index);
+
+		if (plane->format->planes == 2) {
+			index = (index + 1) % 8;
+
+			prio -= 4;
+			dspr |= (index + 1) << prio;
+			dptsr |= DPTSR_PnDK(index) |  DPTSR_PnTS(index);
+		}
+	}
+
+	/* Select display timing and dot clock generator 2 for planes associated
+	 * with superposition controller 2.
+	 */
+	if (rcrtc->index) {
+		u32 value = rcar_du_read(rcdu, DPTSR);
+
+		/* The DPTSR register is updated when the display controller is
+		 * stopped. We thus need to restart the DU. Once again, sorry
+		 * for the flicker. One way to mitigate the issue would be to
+		 * pre-associate planes with CRTCs (either with a fixed 4/4
+		 * split, or through a module parameter). Flicker would then
+		 * occur only if we need to break the pre-association.
+		 */
+		if (value != dptsr) {
+			rcar_du_write(rcdu, DPTSR, dptsr);
+			if (rcdu->used_crtcs) {
+				__rcar_du_start_stop(rcdu, false);
+				__rcar_du_start_stop(rcdu, true);
+			}
+		}
+	}
+
+	rcar_du_write(rcdu, rcrtc->index ? DS2PR : DS1PR, dspr);
+}
+
+static void rcar_du_crtc_start(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	unsigned int i;
+
+	if (rcrtc->started)
+		return;
+
+	if (WARN_ON(rcrtc->plane->format == NULL))
+		return;
+
+	/* Set display off and background to black */
+	rcar_du_crtc_write(rcrtc, DOOR, DOOR_RGB(0, 0, 0));
+	rcar_du_crtc_write(rcrtc, BPOR, BPOR_RGB(0, 0, 0));
+
+	/* Configure display timings and output routing */
+	rcar_du_crtc_set_display_timing(rcrtc);
+	rcar_du_crtc_set_routing(rcrtc);
+
+	mutex_lock(&rcdu->planes.lock);
+	rcrtc->plane->enabled = true;
+	rcar_du_crtc_update_planes(crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	/* Setup planes. */
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		struct rcar_du_plane *plane = &rcdu->planes.planes[i];
+
+		if (plane->crtc != crtc || !plane->enabled)
+			continue;
+
+		rcar_du_plane_setup(plane);
+	}
+
+	/* Select master sync mode. This enables display operation in master
+	 * sync mode (with the HSYNC and VSYNC signals configured as outputs and
+	 * actively driven).
+	 */
+	rcar_du_crtc_clr_set(rcrtc, DSYSR, DSYSR_TVM_MASK, DSYSR_TVM_MASTER);
+
+	rcar_du_start_stop(rcdu, true);
+
+	rcrtc->started = true;
+}
+
+static void rcar_du_crtc_stop(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+
+	if (!rcrtc->started)
+		return;
+
+	mutex_lock(&rcdu->planes.lock);
+	rcrtc->plane->enabled = false;
+	rcar_du_crtc_update_planes(crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	/* Select switch sync mode. This stops display operation and configures
+	 * the HSYNC and VSYNC signals as inputs.
+	 */
+	rcar_du_crtc_clr_set(rcrtc, DSYSR, DSYSR_TVM_MASK, DSYSR_TVM_SWITCH);
+
+	rcar_du_start_stop(rcdu, false);
+
+	rcrtc->started = false;
+}
+
+void rcar_du_crtc_suspend(struct rcar_du_crtc *rcrtc)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_crtc_stop(rcrtc);
+	rcar_du_put(rcdu);
+}
+
+void rcar_du_crtc_resume(struct rcar_du_crtc *rcrtc)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	if (rcrtc->dpms != DRM_MODE_DPMS_ON)
+		return;
+
+	rcar_du_get(rcdu);
+	rcar_du_crtc_start(rcrtc);
+}
+
+static void rcar_du_crtc_update_base(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+
+	rcar_du_plane_compute_base(rcrtc->plane, crtc->fb);
+	rcar_du_plane_update_base(rcrtc->plane);
+}
+
+static void rcar_du_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	if (rcrtc->dpms == mode)
+		return;
+
+	if (mode == DRM_MODE_DPMS_ON) {
+		rcar_du_get(rcdu);
+		rcar_du_crtc_start(rcrtc);
+	} else {
+		rcar_du_crtc_stop(rcrtc);
+		rcar_du_put(rcdu);
+	}
+
+	rcrtc->dpms = mode;
+}
+
+static bool rcar_du_crtc_mode_fixup(struct drm_crtc *crtc,
+				    const struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+	/* TODO Fixup modes */
+	return true;
+}
+
+static void rcar_du_crtc_mode_prepare(struct drm_crtc *crtc)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	/* We need to access the hardware during mode set, acquire a reference
+	 * to the DU.
+	 */
+	rcar_du_get(rcdu);
+
+	/* Stop the CRTC and release the plane. Force the DPMS mode to off as a
+	 * result.
+	 */
+	rcar_du_crtc_stop(rcrtc);
+	rcar_du_plane_release(rcrtc->plane);
+
+	rcrtc->dpms = DRM_MODE_DPMS_OFF;
+}
+
+static int rcar_du_crtc_mode_set(struct drm_crtc *crtc,
+				 struct drm_display_mode *mode,
+				 struct drm_display_mode *adjusted_mode,
+				 int x, int y,
+				 struct drm_framebuffer *old_fb)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+	const struct rcar_du_format_info *format;
+	int ret;
+
+	format = rcar_du_format_info(crtc->fb->pixel_format);
+	if (format == NULL) {
+		dev_dbg(rcdu->dev, "mode_set: unsupported format %08x\n",
+			crtc->fb->pixel_format);
+		ret = -EINVAL;
+		goto error;
+	}
+
+	ret = rcar_du_plane_reserve(rcrtc->plane, format);
+	if (ret < 0)
+		goto error;
+
+	rcrtc->plane->format = format;
+	rcrtc->plane->pitch = crtc->fb->pitches[0];
+
+	rcrtc->plane->src_x = x;
+	rcrtc->plane->src_y = y;
+	rcrtc->plane->width = mode->hdisplay;
+	rcrtc->plane->height = mode->vdisplay;
+
+	rcar_du_plane_compute_base(rcrtc->plane, crtc->fb);
+
+	rcrtc->outputs = 0;
+
+	return 0;
+
+error:
+	/* There's no rollback/abort operation to clean up in case of error. We
+	 * thus need to release the reference to the DU acquired in prepare()
+	 * here.
+	 */
+	rcar_du_put(rcdu);
+	return ret;
+}
+
+static void rcar_du_crtc_mode_commit(struct drm_crtc *crtc)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	/* We're done, restart the CRTC and set the DPMS mode to on. The
+	 * reference to the DU acquired at prepare() time will thus be released
+	 * by the DPMS handler (possibly called by the disable() handler).
+	 */
+	rcar_du_crtc_start(rcrtc);
+	rcrtc->dpms = DRM_MODE_DPMS_ON;
+}
+
+static int rcar_du_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+				      struct drm_framebuffer *old_fb)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	rcrtc->plane->src_x = x;
+	rcrtc->plane->src_y = y;
+
+	rcar_du_crtc_update_base(to_rcar_crtc(crtc));
+
+	return 0;
+}
+
+static void rcar_du_crtc_disable(struct drm_crtc *crtc)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	rcar_du_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+	rcar_du_plane_release(rcrtc->plane);
+}
+
+static const struct drm_crtc_helper_funcs crtc_helper_funcs = {
+	.dpms = rcar_du_crtc_dpms,
+	.mode_fixup = rcar_du_crtc_mode_fixup,
+	.prepare = rcar_du_crtc_mode_prepare,
+	.commit = rcar_du_crtc_mode_commit,
+	.mode_set = rcar_du_crtc_mode_set,
+	.mode_set_base = rcar_du_crtc_mode_set_base,
+	.disable = rcar_du_crtc_disable,
+};
+
+void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc,
+				   struct drm_file *file)
+{
+	struct drm_pending_vblank_event *event;
+	struct drm_device *dev = rcrtc->crtc.dev;
+	unsigned long flags;
+
+	/* Destroy the pending vertical blanking event associated with the
+	 * pending page flip, if any, and disable vertical blanking interrupts.
+	 */
+	spin_lock_irqsave(&dev->event_lock, flags);
+	event = rcrtc->event;
+	if (event && event->base.file_priv == file) {
+		rcrtc->event = NULL;
+		event->base.destroy(&event->base);
+		drm_vblank_put(dev, rcrtc->index);
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
+
+static void rcar_du_crtc_finish_page_flip(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_pending_vblank_event *event;
+	struct drm_device *dev = rcrtc->crtc.dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	event = rcrtc->event;
+	rcrtc->event = NULL;
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	if (event == NULL)
+		return;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	drm_send_vblank_event(dev, rcrtc->index, event);
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	drm_vblank_put(dev, rcrtc->index);
+}
+
+static int rcar_du_crtc_page_flip(struct drm_crtc *crtc,
+				  struct drm_framebuffer *fb,
+				  struct drm_pending_vblank_event *event)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+	struct drm_device *dev = rcrtc->crtc.dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	if (rcrtc->event != NULL) {
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+		return -EBUSY;
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	crtc->fb = fb;
+	rcar_du_crtc_update_base(rcrtc);
+
+	if (event) {
+		event->pipe = rcrtc->index;
+		drm_vblank_get(dev, rcrtc->index);
+		spin_lock_irqsave(&dev->event_lock, flags);
+		rcrtc->event = event;
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	}
+
+	return 0;
+}
+
+static const struct drm_crtc_funcs crtc_funcs = {
+	.destroy = drm_crtc_cleanup,
+	.set_config = drm_crtc_helper_set_config,
+	.page_flip = rcar_du_crtc_page_flip,
+};
+
+int rcar_du_crtc_create(struct rcar_du_device *rcdu, unsigned int index)
+{
+	struct rcar_du_crtc *rcrtc = &rcdu->crtcs[index];
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	int ret;
+
+	rcrtc->mmio_offset = index ? DISP2_REG_OFFSET : 0;
+	rcrtc->index = index;
+	rcrtc->dpms = DRM_MODE_DPMS_OFF;
+	rcrtc->plane = &rcdu->planes.planes[index];
+
+	rcrtc->plane->crtc = crtc;
+
+	ret = drm_crtc_init(rcdu->ddev, crtc, &crtc_funcs);
+	if (ret < 0)
+		return ret;
+
+	drm_crtc_helper_add(crtc, &crtc_helper_funcs);
+
+	return 0;
+}
+
+void rcar_du_crtc_enable_vblank(struct rcar_du_crtc *rcrtc, bool enable)
+{
+	if (enable) {
+		rcar_du_crtc_write(rcrtc, DSRCR, DSRCR_VBCL);
+		rcar_du_crtc_set(rcrtc, DIER, DIER_VBE);
+	} else {
+		rcar_du_crtc_clr(rcrtc, DIER, DIER_VBE);
+	}
+}
+
+void rcar_du_crtc_irq(struct rcar_du_crtc *rcrtc)
+{
+	u32 status;
+
+	status = rcar_du_crtc_read(rcrtc, DSSR);
+	rcar_du_crtc_write(rcrtc, DSRCR, status & DSRCR_MASK);
+
+	if (status & DSSR_VBK) {
+		drm_handle_vblank(rcrtc->crtc.dev, rcrtc->index);
+		rcar_du_crtc_finish_page_flip(rcrtc);
+	}
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
new file mode 100644
index 0000000..2a0365b
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
@@ -0,0 +1,50 @@
+/*
+ * rcar_du_crtc.h  --  R-Car Display Unit CRTCs
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_CRTC_H__
+#define __RCAR_DU_CRTC_H__
+
+#include <linux/mutex.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+
+struct rcar_du_device;
+struct rcar_du_plane;
+
+struct rcar_du_crtc {
+	struct drm_crtc crtc;
+
+	unsigned int mmio_offset;
+	unsigned int index;
+	bool started;
+
+	struct drm_pending_vblank_event *event;
+	unsigned int outputs;
+	int dpms;
+
+	struct rcar_du_plane *plane;
+};
+
+int rcar_du_crtc_create(struct rcar_du_device *rcdu, unsigned int index);
+void rcar_du_crtc_enable_vblank(struct rcar_du_crtc *rcrtc, bool enable);
+void rcar_du_crtc_irq(struct rcar_du_crtc *rcrtc);
+void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc,
+				   struct drm_file *file);
+void rcar_du_crtc_suspend(struct rcar_du_crtc *rcrtc);
+void rcar_du_crtc_resume(struct rcar_du_crtc *rcrtc);
+
+void rcar_du_crtc_route_output(struct drm_crtc *crtc, unsigned int output);
+void rcar_du_crtc_update_planes(struct drm_crtc *crtc);
+
+#endif /* __RCAR_DU_CRTC_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.c b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
new file mode 100644
index 0000000..ff82877
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
@@ -0,0 +1,325 @@
+/*
+ * rcar_du_drv.c  --  R-Car Display Unit DRM driver
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/pm.h>
+#include <linux/slab.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_regs.h"
+
+/* -----------------------------------------------------------------------------
+ * Core device operations
+ */
+
+/*
+ * rcar_du_get - Acquire a reference to the DU
+ *
+ * Acquiring a reference enables the device clock and setup core registers. A
+ * reference must be held before accessing any hardware registers.
+ *
+ * This function must be called with the DRM mode_config lock held.
+ *
+ * Return 0 in case of success or a negative error code otherwise.
+ */
+int rcar_du_get(struct rcar_du_device *rcdu)
+{
+	int ret;
+
+	if (rcdu->use_count)
+		goto done;
+
+	/* Enable clocks before accessing the hardware. */
+	ret = clk_prepare_enable(rcdu->clock);
+	if (ret < 0)
+		return ret;
+
+	/* Enable extended features */
+	rcar_du_write(rcdu, DEFR, DEFR_CODE | DEFR_DEFE);
+	rcar_du_write(rcdu, DEFR2, DEFR2_CODE | DEFR2_DEFE2G);
+	rcar_du_write(rcdu, DEFR3, DEFR3_CODE | DEFR3_DEFE3);
+	rcar_du_write(rcdu, DEFR4, DEFR4_CODE);
+	rcar_du_write(rcdu, DEFR5, DEFR5_CODE | DEFR5_DEFE5);
+
+	/* Use DS1PR and DS2PR to configure planes priorities and connects the
+	 * superposition 0 to DU0 pins. DU1 pins will be configured dynamically.
+	 */
+	rcar_du_write(rcdu, DORCR, DORCR_PG1D_DS1 | DORCR_DPRS);
+
+done:
+	rcdu->use_count++;
+	return 0;
+}
+
+/*
+ * rcar_du_put - Release a reference to the DU
+ *
+ * Releasing the last reference disables the device clock.
+ *
+ * This function must be called with the DRM mode_config lock held.
+ */
+void rcar_du_put(struct rcar_du_device *rcdu)
+{
+	if (--rcdu->use_count)
+		return;
+
+	clk_disable_unprepare(rcdu->clock);
+}
+
+/* -----------------------------------------------------------------------------
+ * DRM operations
+ */
+
+static int rcar_du_unload(struct drm_device *dev)
+{
+	drm_kms_helper_poll_fini(dev);
+	drm_mode_config_cleanup(dev);
+	drm_vblank_cleanup(dev);
+	drm_irq_uninstall(dev);
+
+	dev->dev_private = NULL;
+
+	return 0;
+}
+
+static int rcar_du_load(struct drm_device *dev, unsigned long flags)
+{
+	struct platform_device *pdev = dev->platformdev;
+	struct rcar_du_platform_data *pdata = pdev->dev.platform_data;
+	struct rcar_du_device *rcdu;
+	struct resource *ioarea;
+	struct resource *mem;
+	int ret;
+
+	if (pdata == NULL) {
+		dev_err(dev->dev, "no platform data\n");
+		return -ENODEV;
+	}
+
+	rcdu = devm_kzalloc(&pdev->dev, sizeof(*rcdu), GFP_KERNEL);
+	if (rcdu == NULL) {
+		dev_err(dev->dev, "failed to allocate private data\n");
+		return -ENOMEM;
+	}
+
+	rcdu->dev = &pdev->dev;
+	rcdu->pdata = pdata;
+	rcdu->ddev = dev;
+	dev->dev_private = rcdu;
+
+	/* I/O resources and clocks */
+	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (mem == NULL) {
+		dev_err(&pdev->dev, "failed to get memory resource\n");
+		return -EINVAL;
+	}
+
+	ioarea = devm_request_mem_region(&pdev->dev, mem->start,
+					 resource_size(mem), pdev->name);
+	if (ioarea == NULL) {
+		dev_err(&pdev->dev, "failed to request memory region\n");
+		return -EBUSY;
+	}
+
+	rcdu->mmio = devm_ioremap_nocache(&pdev->dev, ioarea->start,
+					  resource_size(ioarea));
+	if (rcdu->mmio == NULL) {
+		dev_err(&pdev->dev, "failed to remap memory resource\n");
+		return -ENOMEM;
+	}
+
+	rcdu->clock = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(rcdu->clock)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		return -ENOENT;
+	}
+
+	/* DRM/KMS objects */
+	ret = rcar_du_modeset_init(rcdu);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to initialize DRM/KMS\n");
+		goto done;
+	}
+
+	/* IRQ and vblank handling */
+	ret = drm_vblank_init(dev, (1 << rcdu->num_crtcs) - 1);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to initialize vblank\n");
+		goto done;
+	}
+
+	ret = drm_irq_install(dev);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to install IRQ handler\n");
+		goto done;
+	}
+
+	platform_set_drvdata(pdev, rcdu);
+
+done:
+	if (ret)
+		rcar_du_unload(dev);
+
+	return ret;
+}
+
+static void rcar_du_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct rcar_du_device *rcdu = dev->dev_private;
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->crtcs); ++i)
+		rcar_du_crtc_cancel_page_flip(&rcdu->crtcs[i], file);
+}
+
+static irqreturn_t rcar_du_irq(int irq, void *arg)
+{
+	struct drm_device *dev = arg;
+	struct rcar_du_device *rcdu = dev->dev_private;
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->crtcs); ++i)
+		rcar_du_crtc_irq(&rcdu->crtcs[i]);
+
+	return IRQ_HANDLED;
+}
+
+static int rcar_du_enable_vblank(struct drm_device *dev, int crtc)
+{
+	struct rcar_du_device *rcdu = dev->dev_private;
+
+	rcar_du_crtc_enable_vblank(&rcdu->crtcs[crtc], true);
+
+	return 0;
+}
+
+static void rcar_du_disable_vblank(struct drm_device *dev, int crtc)
+{
+	struct rcar_du_device *rcdu = dev->dev_private;
+
+	rcar_du_crtc_enable_vblank(&rcdu->crtcs[crtc], false);
+}
+
+static const struct file_operations rcar_du_fops = {
+	.owner		= THIS_MODULE,
+	.open		= drm_open,
+	.release	= drm_release,
+	.unlocked_ioctl	= drm_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= drm_compat_ioctl,
+#endif
+	.poll		= drm_poll,
+	.read		= drm_read,
+	.fasync		= drm_fasync,
+	.llseek		= no_llseek,
+	.mmap		= drm_gem_cma_mmap,
+};
+
+static struct drm_driver rcar_du_driver = {
+	.driver_features	= DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET
+				| DRIVER_PRIME,
+	.load			= rcar_du_load,
+	.unload			= rcar_du_unload,
+	.preclose		= rcar_du_preclose,
+	.irq_handler		= rcar_du_irq,
+	.get_vblank_counter	= drm_vblank_count,
+	.enable_vblank		= rcar_du_enable_vblank,
+	.disable_vblank		= rcar_du_disable_vblank,
+	.gem_free_object	= drm_gem_cma_free_object,
+	.gem_vm_ops		= &drm_gem_cma_vm_ops,
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_cma_dmabuf_import,
+	.gem_prime_export	= drm_gem_cma_dmabuf_export,
+	.dumb_create		= rcar_du_dumb_create,
+	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,
+	.dumb_destroy		= drm_gem_cma_dumb_destroy,
+	.fops			= &rcar_du_fops,
+	.name			= "rcar-du",
+	.desc			= "Renesas R-Car Display Unit",
+	.date			= "20130110",
+	.major			= 1,
+	.minor			= 0,
+};
+
+/* -----------------------------------------------------------------------------
+ * Power management
+ */
+
+#if CONFIG_PM_SLEEP
+static int rcar_du_pm_suspend(struct device *dev)
+{
+	struct rcar_du_device *rcdu = dev_get_drvdata(dev);
+
+	drm_kms_helper_poll_disable(rcdu->ddev);
+	/* TODO Suspend the CRTC */
+
+	return 0;
+}
+
+static int rcar_du_pm_resume(struct device *dev)
+{
+	struct rcar_du_device *rcdu = dev_get_drvdata(dev);
+
+	/* TODO Resume the CRTC */
+
+	drm_kms_helper_poll_enable(rcdu->ddev);
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops rcar_du_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(rcar_du_pm_suspend, rcar_du_pm_resume)
+};
+
+/* -----------------------------------------------------------------------------
+ * Platform driver
+ */
+
+static int rcar_du_probe(struct platform_device *pdev)
+{
+	return drm_platform_init(&rcar_du_driver, pdev);
+}
+
+static int rcar_du_remove(struct platform_device *pdev)
+{
+	drm_platform_exit(&rcar_du_driver, pdev);
+
+	return 0;
+}
+
+static struct platform_driver rcar_du_platform_driver = {
+	.probe		= rcar_du_probe,
+	.remove		= rcar_du_remove,
+	.driver		= {
+		.owner	= THIS_MODULE,
+		.name	= "rcar-du",
+		.pm	= &rcar_du_pm_ops,
+	},
+};
+
+module_platform_driver(rcar_du_platform_driver);
+
+MODULE_AUTHOR("Laurent Pinchart <laurent.pinchart@ideasonboard.com>");
+MODULE_DESCRIPTION("Renesas R-Car Display Unit DRM Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.h b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
new file mode 100644
index 0000000..193cc59
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
@@ -0,0 +1,66 @@
+/*
+ * rcar_du_drv.h  --  R-Car Display Unit DRM driver
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_DRV_H__
+#define __RCAR_DU_DRV_H__
+
+#include <linux/kernel.h>
+#include <linux/mutex.h>
+#include <linux/platform_data/rcar-du.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_plane.h"
+
+struct clk;
+struct device;
+struct drm_device;
+
+struct rcar_du_device {
+	struct device *dev;
+	const struct rcar_du_platform_data *pdata;
+
+	void __iomem *mmio;
+	struct clk *clock;
+	unsigned int use_count;
+
+	struct drm_device *ddev;
+
+	struct rcar_du_crtc crtcs[2];
+	unsigned int used_crtcs;
+	unsigned int num_crtcs;
+
+	struct {
+		struct rcar_du_plane planes[RCAR_DU_NUM_SW_PLANES];
+		unsigned int free;
+		struct mutex lock;
+
+		struct drm_property *alpha;
+		struct drm_property *colorkey;
+		struct drm_property *zpos;
+	} planes;
+};
+
+int rcar_du_get(struct rcar_du_device *rcdu);
+void rcar_du_put(struct rcar_du_device *rcdu);
+
+static inline u32 rcar_du_read(struct rcar_du_device *rcdu, u32 reg)
+{
+	return ioread32(rcdu->mmio + reg);
+}
+
+static inline void rcar_du_write(struct rcar_du_device *rcdu, u32 reg, u32 data)
+{
+	iowrite32(data, rcdu->mmio + reg);
+}
+
+#endif /* __RCAR_DU_DRV_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
new file mode 100644
index 0000000..d30c2e2
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
@@ -0,0 +1,265 @@
+/*
+ * rcar_du_kms.c  --  R-Car Display Unit Mode Setting
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_lvds.h"
+#include "rcar_du_regs.h"
+#include "rcar_du_vga.h"
+
+/* -----------------------------------------------------------------------------
+ * Format helpers
+ */
+
+static const struct rcar_du_format_info rcar_du_format_infos[] = {
+	{
+		.fourcc = DRM_FORMAT_RGB565,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP | PnMR_DDDF_16BPP,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_ARGB1555,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_ALP | PnMR_DDDF_ARGB,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_XRGB1555,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_ALP | PnMR_DDDF_ARGB,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_XRGB8888,
+		.bpp = 32,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP | PnMR_DDDF_16BPP,
+		.edf = PnDDCR4_EDF_RGB888,
+	}, {
+		.fourcc = DRM_FORMAT_ARGB8888,
+		.bpp = 32,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_ALP | PnMR_DDDF_16BPP,
+		.edf = PnDDCR4_EDF_ARGB8888,
+	}, {
+		.fourcc = DRM_FORMAT_UYVY,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_YUYV,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_NV12,
+		.bpp = 12,
+		.planes = 2,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_NV21,
+		.bpp = 12,
+		.planes = 2,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		/* In YUV 4:2:2, only NV16 is supported (NV61 isn't) */
+		.fourcc = DRM_FORMAT_NV16,
+		.bpp = 16,
+		.planes = 2,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	},
+};
+
+const struct rcar_du_format_info *rcar_du_format_info(u32 fourcc)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(rcar_du_format_infos); ++i) {
+		if (rcar_du_format_infos[i].fourcc == fourcc)
+			return &rcar_du_format_infos[i];
+	}
+
+	return NULL;
+}
+
+/* -----------------------------------------------------------------------------
+ * Common connector and encoder functions
+ */
+
+struct drm_encoder *
+rcar_du_connector_best_encoder(struct drm_connector *connector)
+{
+	struct rcar_du_connector *rcon = to_rcar_connector(connector);
+
+	return &rcon->encoder->encoder;
+}
+
+void rcar_du_encoder_mode_prepare(struct drm_encoder *encoder)
+{
+}
+
+void rcar_du_encoder_mode_set(struct drm_encoder *encoder,
+			      struct drm_display_mode *mode,
+			      struct drm_display_mode *adjusted_mode)
+{
+	struct rcar_du_encoder *renc = to_rcar_encoder(encoder);
+
+	rcar_du_crtc_route_output(encoder->crtc, renc->output);
+}
+
+void rcar_du_encoder_mode_commit(struct drm_encoder *encoder)
+{
+}
+
+/* -----------------------------------------------------------------------------
+ * Frame buffer
+ */
+
+int rcar_du_dumb_create(struct drm_file *file, struct drm_device *dev,
+			struct drm_mode_create_dumb *args)
+{
+	unsigned int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	unsigned int align;
+
+	/* The pitch must be aligned to a 16 pixels boundary. */
+	align = 16 * args->bpp / 8;
+	args->pitch = roundup(max(args->pitch, min_pitch), align);
+
+	return drm_gem_cma_dumb_create(file, dev, args);
+}
+
+static struct drm_framebuffer *
+rcar_du_fb_create(struct drm_device *dev, struct drm_file *file_priv,
+		  struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	const struct rcar_du_format_info *format;
+	unsigned int align;
+
+	format = rcar_du_format_info(mode_cmd->pixel_format);
+	if (format == NULL) {
+		dev_dbg(dev->dev, "unsupported pixel format %08x\n",
+			mode_cmd->pixel_format);
+		return ERR_PTR(-EINVAL);
+	}
+
+	align = 16 * format->bpp / 8;
+
+	if (mode_cmd->pitches[0] & (align - 1) ||
+	    mode_cmd->pitches[0] >= 8192) {
+		dev_dbg(dev->dev, "invalid pitch value %u\n",
+			mode_cmd->pitches[0]);
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (format->planes == 2) {
+		if (mode_cmd->pitches[1] != mode_cmd->pitches[0]) {
+			dev_dbg(dev->dev,
+				"luma and chroma pitches do not match\n");
+			return ERR_PTR(-EINVAL);
+		}
+	}
+
+	return drm_fb_cma_create(dev, file_priv, mode_cmd);
+}
+
+static const struct drm_mode_config_funcs rcar_du_mode_config_funcs = {
+	.fb_create = rcar_du_fb_create,
+};
+
+int rcar_du_modeset_init(struct rcar_du_device *rcdu)
+{
+	struct drm_device *dev = rcdu->ddev;
+	struct drm_encoder *encoder;
+	unsigned int i;
+	int ret;
+
+	drm_mode_config_init(rcdu->ddev);
+
+	rcdu->ddev->mode_config.min_width = 0;
+	rcdu->ddev->mode_config.min_height = 0;
+	rcdu->ddev->mode_config.max_width = 4095;
+	rcdu->ddev->mode_config.max_height = 2047;
+	rcdu->ddev->mode_config.funcs = &rcar_du_mode_config_funcs;
+
+	ret = rcar_du_plane_init(rcdu);
+	if (ret < 0)
+		return ret;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->crtcs); ++i) {
+		ret = rcar_du_crtc_create(rcdu, i);
+		if (ret < 0)
+			return ret;
+	}
+
+	rcdu->used_crtcs = 0;
+	rcdu->num_crtcs = i;
+
+	for (i = 0; i < rcdu->pdata->num_encoders; ++i) {
+		const struct rcar_du_encoder_data *pdata =
+			&rcdu->pdata->encoders[i];
+
+		if (pdata->output >= ARRAY_SIZE(rcdu->crtcs)) {
+			dev_warn(rcdu->dev,
+				 "encoder %u references unexisting output %u, skipping\n",
+				 i, pdata->output);
+			continue;
+		}
+
+		switch (pdata->encoder) {
+		case RCAR_DU_ENCODER_VGA:
+			rcar_du_vga_init(rcdu, &pdata->u.vga, pdata->output);
+			break;
+
+		case RCAR_DU_ENCODER_LVDS:
+			rcar_du_lvds_init(rcdu, &pdata->u.lvds, pdata->output);
+			break;
+
+		default:
+			break;
+		}
+	}
+
+	/* Set the possible CRTCs and possible clones. All encoders can be
+	 * driven by the CRTC associated with the output they're connected to,
+	 * as well as by CRTC 0.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		struct rcar_du_encoder *renc = to_rcar_encoder(encoder);
+
+		encoder->possible_crtcs = (1 << 0) | (1 << renc->output);
+		encoder->possible_clones = 1 << 0;
+	}
+
+	ret = rcar_du_plane_register(rcdu);
+	if (ret < 0)
+		return ret;
+
+	drm_kms_helper_poll_init(rcdu->ddev);
+
+	drm_helper_disable_unused_functions(rcdu->ddev);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.h b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
new file mode 100644
index 0000000..dba4722
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
@@ -0,0 +1,62 @@
+/*
+ * rcar_du_kms.h  --  R-Car Display Unit Mode Setting
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_KMS_H__
+#define __RCAR_DU_KMS_H__
+
+#include <linux/types.h>
+
+#include <drm/drm_crtc.h>
+
+struct rcar_du_device;
+
+struct rcar_du_format_info {
+	u32 fourcc;
+	unsigned int bpp;
+	unsigned int planes;
+	unsigned int pnmr;
+	unsigned int edf;
+};
+
+struct rcar_du_encoder {
+	struct drm_encoder encoder;
+	unsigned int output;
+};
+
+#define to_rcar_encoder(e) \
+	container_of(e, struct rcar_du_encoder, encoder)
+
+struct rcar_du_connector {
+	struct drm_connector connector;
+	struct rcar_du_encoder *encoder;
+};
+
+#define to_rcar_connector(c) \
+	container_of(c, struct rcar_du_connector, connector)
+
+const struct rcar_du_format_info *rcar_du_format_info(u32 fourcc);
+
+struct drm_encoder *
+rcar_du_connector_best_encoder(struct drm_connector *connector);
+void rcar_du_encoder_mode_prepare(struct drm_encoder *encoder);
+void rcar_du_encoder_mode_set(struct drm_encoder *encoder,
+			      struct drm_display_mode *mode,
+			      struct drm_display_mode *adjusted_mode);
+void rcar_du_encoder_mode_commit(struct drm_encoder *encoder);
+
+int rcar_du_modeset_init(struct rcar_du_device *rcdu);
+
+int rcar_du_dumb_create(struct drm_file *file, struct drm_device *dev,
+			struct drm_mode_create_dumb *args);
+
+#endif /* __RCAR_DU_KMS_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvds.c b/drivers/gpu/drm/rcar-du/rcar_du_lvds.c
new file mode 100644
index 0000000..7aefe72
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvds.c
@@ -0,0 +1,216 @@
+/*
+ * rcar_du_lvds.c  --  R-Car Display Unit LVDS Encoder and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_lvds.h"
+
+struct rcar_du_lvds_connector {
+	struct rcar_du_connector connector;
+
+	const struct rcar_du_panel_data *panel;
+};
+
+#define to_rcar_lvds_connector(c) \
+	container_of(c, struct rcar_du_lvds_connector, connector.connector)
+
+/* -----------------------------------------------------------------------------
+ * Connector
+ */
+
+static int rcar_du_lvds_connector_get_modes(struct drm_connector *connector)
+{
+	struct rcar_du_lvds_connector *lvdscon = to_rcar_lvds_connector(connector);
+	struct drm_display_mode *mode;
+
+	mode = drm_mode_create(connector->dev);
+	if (mode == NULL)
+		return 0;
+
+	mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
+	mode->clock = lvdscon->panel->mode.clock;
+	mode->hdisplay = lvdscon->panel->mode.hdisplay;
+	mode->hsync_start = lvdscon->panel->mode.hsync_start;
+	mode->hsync_end = lvdscon->panel->mode.hsync_end;
+	mode->htotal = lvdscon->panel->mode.htotal;
+	mode->vdisplay = lvdscon->panel->mode.vdisplay;
+	mode->vsync_start = lvdscon->panel->mode.vsync_start;
+	mode->vsync_end = lvdscon->panel->mode.vsync_end;
+	mode->vtotal = lvdscon->panel->mode.vtotal;
+	mode->flags = lvdscon->panel->mode.flags;
+
+	drm_mode_set_name(mode);
+	drm_mode_probed_add(connector, mode);
+
+	return 1;
+}
+
+static int rcar_du_lvds_connector_mode_valid(struct drm_connector *connector,
+					    struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static const struct drm_connector_helper_funcs connector_helper_funcs = {
+	.get_modes = rcar_du_lvds_connector_get_modes,
+	.mode_valid = rcar_du_lvds_connector_mode_valid,
+	.best_encoder = rcar_du_connector_best_encoder,
+};
+
+static void rcar_du_lvds_connector_destroy(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+}
+
+static enum drm_connector_status
+rcar_du_lvds_connector_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static const struct drm_connector_funcs connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = rcar_du_lvds_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = rcar_du_lvds_connector_destroy,
+};
+
+static int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
+				       struct rcar_du_encoder *renc,
+				       const struct rcar_du_panel_data *panel)
+{
+	struct rcar_du_lvds_connector *lvdscon;
+	struct drm_connector *connector;
+	int ret;
+
+	lvdscon = devm_kzalloc(rcdu->dev, sizeof(*lvdscon), GFP_KERNEL);
+	if (lvdscon == NULL)
+		return -ENOMEM;
+
+	lvdscon->panel = panel;
+
+	connector = &lvdscon->connector.connector;
+	connector->display_info.width_mm = panel->width_mm;
+	connector->display_info.height_mm = panel->height_mm;
+
+	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
+				 DRM_MODE_CONNECTOR_LVDS);
+	if (ret < 0)
+		return ret;
+
+	drm_connector_helper_add(connector, &connector_helper_funcs);
+	ret = drm_sysfs_connector_add(connector);
+	if (ret < 0)
+		return ret;
+
+	drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	drm_object_property_set_value(&connector->base,
+		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
+
+	ret = drm_mode_connector_attach_encoder(connector, &renc->encoder);
+	if (ret < 0)
+		return ret;
+
+	connector->encoder = &renc->encoder;
+	lvdscon->connector.encoder = renc;
+
+	return 0;
+}
+
+/* -----------------------------------------------------------------------------
+ * Encoder
+ */
+
+static void rcar_du_lvds_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool rcar_du_lvds_encoder_mode_fixup(struct drm_encoder *encoder,
+					   const struct drm_display_mode *mode,
+					   struct drm_display_mode *adjusted_mode)
+{
+	const struct drm_display_mode *panel_mode;
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector;
+	bool found = false;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		if (connector->encoder == encoder) {
+			found = true;
+			break;
+		}
+	}
+
+	if (!found) {
+		dev_dbg(dev->dev, "mode_fixup: no connector found\n");
+		return false;
+	}
+
+	if (list_empty(&connector->modes)) {
+		dev_dbg(dev->dev, "mode_fixup: empty modes list\n");
+		return false;
+	}
+
+	panel_mode = list_first_entry(&connector->modes,
+				      struct drm_display_mode, head);
+
+	/* We're not allowed to modify the resolution. */
+	if (mode->hdisplay != panel_mode->hdisplay ||
+	    mode->vdisplay != panel_mode->vdisplay)
+		return false;
+
+	/* The flat panel mode is fixed, just copy it to the adjusted mode. */
+	drm_mode_copy(adjusted_mode, panel_mode);
+
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs encoder_helper_funcs = {
+	.dpms = rcar_du_lvds_encoder_dpms,
+	.mode_fixup = rcar_du_lvds_encoder_mode_fixup,
+	.prepare = rcar_du_encoder_mode_prepare,
+	.commit = rcar_du_encoder_mode_commit,
+	.mode_set = rcar_du_encoder_mode_set,
+};
+
+static const struct drm_encoder_funcs encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+int rcar_du_lvds_init(struct rcar_du_device *rcdu,
+		      const struct rcar_du_encoder_lvds_data *data,
+		      unsigned int output)
+{
+	struct rcar_du_encoder *renc;
+	int ret;
+
+	renc = devm_kzalloc(rcdu->dev, sizeof(*renc), GFP_KERNEL);
+	if (renc == NULL)
+		return -ENOMEM;
+
+	renc->output = output;
+
+	ret = drm_encoder_init(rcdu->ddev, &renc->encoder, &encoder_funcs,
+			       DRM_MODE_ENCODER_LVDS);
+	if (ret < 0)
+		return ret;
+
+	drm_encoder_helper_add(&renc->encoder, &encoder_helper_funcs);
+
+	return rcar_du_lvds_connector_init(rcdu, renc, &data->panel);
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvds.h b/drivers/gpu/drm/rcar-du/rcar_du_lvds.h
new file mode 100644
index 0000000..b47f832
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvds.h
@@ -0,0 +1,24 @@
+/*
+ * rcar_du_lvds.h  --  R-Car Display Unit LVDS Encoder and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_LVDS_H__
+#define __RCAR_DU_LVDS_H__
+
+struct rcar_du_device;
+struct rcar_du_encoder_lvds_data;
+
+int rcar_du_lvds_init(struct rcar_du_device *rcdu,
+		      const struct rcar_du_encoder_lvds_data *data,
+		      unsigned int output);
+
+#endif /* __RCAR_DU_LVDS_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.c b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
new file mode 100644
index 0000000..a65f81d
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
@@ -0,0 +1,507 @@
+/*
+ * rcar_du_plane.c  --  R-Car Display Unit Planes
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_plane.h"
+#include "rcar_du_regs.h"
+
+#define RCAR_DU_COLORKEY_NONE		(0 << 24)
+#define RCAR_DU_COLORKEY_SOURCE		(1 << 24)
+#define RCAR_DU_COLORKEY_MASK		(1 << 24)
+
+struct rcar_du_kms_plane {
+	struct drm_plane plane;
+	struct rcar_du_plane *hwplane;
+};
+
+static inline struct rcar_du_plane *to_rcar_plane(struct drm_plane *plane)
+{
+	return container_of(plane, struct rcar_du_kms_plane, plane)->hwplane;
+}
+
+static u32 rcar_du_plane_read(struct rcar_du_device *rcdu,
+			      unsigned int index, u32 reg)
+{
+	return rcar_du_read(rcdu, index * PLANE_OFF + reg);
+}
+
+static void rcar_du_plane_write(struct rcar_du_device *rcdu,
+				unsigned int index, u32 reg, u32 data)
+{
+	rcar_du_write(rcdu, index * PLANE_OFF + reg, data);
+}
+
+int rcar_du_plane_reserve(struct rcar_du_plane *plane,
+			  const struct rcar_du_format_info *format)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	unsigned int i;
+	int ret = -EBUSY;
+
+	mutex_lock(&rcdu->planes.lock);
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		if (!(rcdu->planes.free & (1 << i)))
+			continue;
+
+		if (format->planes == 1 ||
+		    rcdu->planes.free & (1 << ((i + 1) % 8)))
+			break;
+	}
+
+	if (i == ARRAY_SIZE(rcdu->planes.planes))
+		goto done;
+
+	rcdu->planes.free &= ~(1 << i);
+	if (format->planes == 2)
+		rcdu->planes.free &= ~(1 << ((i + 1) % 8));
+
+	plane->hwindex = i;
+
+	ret = 0;
+
+done:
+	mutex_unlock(&rcdu->planes.lock);
+	return ret;
+}
+
+void rcar_du_plane_release(struct rcar_du_plane *plane)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+
+	if (plane->hwindex == -1)
+		return;
+
+	mutex_lock(&rcdu->planes.lock);
+	rcdu->planes.free |= 1 << plane->hwindex;
+	if (plane->format->planes == 2)
+		rcdu->planes.free |= 1 << ((plane->hwindex + 1) % 8);
+	mutex_unlock(&rcdu->planes.lock);
+
+	plane->hwindex = -1;
+}
+
+void rcar_du_plane_update_base(struct rcar_du_plane *plane)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	unsigned int index = plane->hwindex;
+
+	/* According to the datasheet the Y position is expressed in raster line
+	 * units. However, 32bpp formats seem to require a doubled Y position
+	 * value. Similarly, for the second plane, NV12 and NV21 formats seem to
+	 * require a halved Y position value.
+	 */
+	rcar_du_plane_write(rcdu, index, PnSPXR, plane->src_x);
+	rcar_du_plane_write(rcdu, index, PnSPYR, plane->src_y *
+			    (plane->format->bpp == 32 ? 2 : 1));
+	rcar_du_plane_write(rcdu, index, PnDSA0R, plane->dma[0]);
+
+	if (plane->format->planes == 2) {
+		index = (index + 1) % 8;
+
+		rcar_du_plane_write(rcdu, index, PnSPXR, plane->src_x);
+		rcar_du_plane_write(rcdu, index, PnSPYR, plane->src_y *
+				    (plane->format->bpp == 16 ? 2 : 1) / 2);
+		rcar_du_plane_write(rcdu, index, PnDSA0R, plane->dma[1]);
+	}
+}
+
+void rcar_du_plane_compute_base(struct rcar_du_plane *plane,
+				struct drm_framebuffer *fb)
+{
+	struct drm_gem_cma_object *gem;
+
+	gem = drm_fb_cma_get_gem_obj(fb, 0);
+	plane->dma[0] = gem->paddr + fb->offsets[0];
+
+	if (plane->format->planes == 2) {
+		gem = drm_fb_cma_get_gem_obj(fb, 1);
+		plane->dma[1] = gem->paddr + fb->offsets[1];
+	}
+}
+
+static void rcar_du_plane_setup_mode(struct rcar_du_plane *plane,
+				     unsigned int index)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	u32 colorkey;
+	u32 pnmr;
+
+	/* The PnALPHAR register controls alpha-blending in 16bpp formats
+	 * (ARGB1555 and XRGB1555).
+	 *
+	 * For ARGB, set the alpha value to 0, and enable alpha-blending when
+	 * the A bit is 0. This maps A=0 to alpha=0 and A=1 to alpha=255.
+	 *
+	 * For XRGB, set the alpha value to the plane-wide alpha value and
+	 * enable alpha-blending regardless of the X bit value.
+	 */
+	if (plane->format->fourcc != DRM_FORMAT_XRGB1555)
+		rcar_du_plane_write(rcdu, index, PnALPHAR, PnALPHAR_ABIT_0);
+	else
+		rcar_du_plane_write(rcdu, index, PnALPHAR,
+				    PnALPHAR_ABIT_X | plane->alpha);
+
+	pnmr = PnMR_BM_MD | plane->format->pnmr;
+
+	/* Disable color keying when requested. YUV formats have the
+	 * PnMR_SPIM_TP_OFF bit set in their pnmr field, disabling color keying
+	 * automatically.
+	 */
+	if ((plane->colorkey & RCAR_DU_COLORKEY_MASK) == RCAR_DU_COLORKEY_NONE)
+		pnmr |= PnMR_SPIM_TP_OFF;
+
+	/* For packed YUV formats we need to select the U/V order. */
+	if (plane->format->fourcc == DRM_FORMAT_YUYV)
+		pnmr |= PnMR_YCDF_YUYV;
+
+	rcar_du_plane_write(rcdu, index, PnMR, pnmr);
+
+	switch (plane->format->fourcc) {
+	case DRM_FORMAT_RGB565:
+		colorkey = ((plane->colorkey & 0xf80000) >> 8)
+			 | ((plane->colorkey & 0x00fc00) >> 5)
+			 | ((plane->colorkey & 0x0000f8) >> 3);
+		rcar_du_plane_write(rcdu, index, PnTC2R, colorkey);
+		break;
+
+	case DRM_FORMAT_ARGB1555:
+	case DRM_FORMAT_XRGB1555:
+		colorkey = ((plane->colorkey & 0xf80000) >> 9)
+			 | ((plane->colorkey & 0x00f800) >> 6)
+			 | ((plane->colorkey & 0x0000f8) >> 3);
+		rcar_du_plane_write(rcdu, index, PnTC2R, colorkey);
+		break;
+
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ARGB8888:
+		rcar_du_plane_write(rcdu, index, PnTC3R,
+				    PnTC3R_CODE | (plane->colorkey & 0xffffff));
+		break;
+	}
+}
+
+static void __rcar_du_plane_setup(struct rcar_du_plane *plane,
+				  unsigned int index)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	u32 ddcr2 = PnDDCR2_CODE;
+	u32 ddcr4;
+	u32 mwr;
+
+	/* Data format
+	 *
+	 * The data format is selected by the DDDF field in PnMR and the EDF
+	 * field in DDCR4.
+	 */
+	ddcr4 = rcar_du_plane_read(rcdu, index, PnDDCR4);
+	ddcr4 &= ~PnDDCR4_EDF_MASK;
+	ddcr4 |= plane->format->edf | PnDDCR4_CODE;
+
+	rcar_du_plane_setup_mode(plane, index);
+
+	if (plane->format->planes == 2) {
+		if (plane->hwindex != index) {
+			if (plane->format->fourcc == DRM_FORMAT_NV12 ||
+			    plane->format->fourcc == DRM_FORMAT_NV21)
+				ddcr2 |= PnDDCR2_Y420;
+
+			if (plane->format->fourcc == DRM_FORMAT_NV21)
+				ddcr2 |= PnDDCR2_NV21;
+
+			ddcr2 |= PnDDCR2_DIVU;
+		} else {
+			ddcr2 |= PnDDCR2_DIVY;
+		}
+	}
+
+	rcar_du_plane_write(rcdu, index, PnDDCR2, ddcr2);
+	rcar_du_plane_write(rcdu, index, PnDDCR4, ddcr4);
+
+	/* Memory pitch (expressed in pixels) */
+	if (plane->format->planes == 2)
+		mwr = plane->pitch;
+	else
+		mwr = plane->pitch * 8 / plane->format->bpp;
+
+	rcar_du_plane_write(rcdu, index, PnMWR, mwr);
+
+	/* Destination position and size */
+	rcar_du_plane_write(rcdu, index, PnDSXR, plane->width);
+	rcar_du_plane_write(rcdu, index, PnDSYR, plane->height);
+	rcar_du_plane_write(rcdu, index, PnDPXR, plane->dst_x);
+	rcar_du_plane_write(rcdu, index, PnDPYR, plane->dst_y);
+
+	/* Wrap-around and blinking, disabled */
+	rcar_du_plane_write(rcdu, index, PnWASPR, 0);
+	rcar_du_plane_write(rcdu, index, PnWAMWR, 4095);
+	rcar_du_plane_write(rcdu, index, PnBTR, 0);
+	rcar_du_plane_write(rcdu, index, PnMLR, 0);
+}
+
+void rcar_du_plane_setup(struct rcar_du_plane *plane)
+{
+	__rcar_du_plane_setup(plane, plane->hwindex);
+	if (plane->format->planes == 2)
+		__rcar_du_plane_setup(plane, (plane->hwindex + 1) % 8);
+
+	rcar_du_plane_update_base(plane);
+}
+
+static int
+rcar_du_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
+		       struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+		       unsigned int crtc_w, unsigned int crtc_h,
+		       uint32_t src_x, uint32_t src_y,
+		       uint32_t src_w, uint32_t src_h)
+{
+	struct rcar_du_plane *rplane = to_rcar_plane(plane);
+	struct rcar_du_device *rcdu = plane->dev->dev_private;
+	const struct rcar_du_format_info *format;
+	unsigned int nplanes;
+	int ret;
+
+	format = rcar_du_format_info(fb->pixel_format);
+	if (format == NULL) {
+		dev_dbg(rcdu->dev, "%s: unsupported format %08x\n", __func__,
+			fb->pixel_format);
+		return -EINVAL;
+	}
+
+	if (src_w >> 16 != crtc_w || src_h >> 16 != crtc_h) {
+		dev_dbg(rcdu->dev, "%s: scaling not supported\n", __func__);
+		return -EINVAL;
+	}
+
+	nplanes = rplane->format ? rplane->format->planes : 0;
+
+	/* Reallocate hardware planes if the number of required planes has
+	 * changed.
+	 */
+	if (format->planes != nplanes) {
+		rcar_du_plane_release(rplane);
+		ret = rcar_du_plane_reserve(rplane, format);
+		if (ret < 0)
+			return ret;
+	}
+
+	rplane->crtc = crtc;
+	rplane->format = format;
+	rplane->pitch = fb->pitches[0];
+
+	rplane->src_x = src_x >> 16;
+	rplane->src_y = src_y >> 16;
+	rplane->dst_x = crtc_x;
+	rplane->dst_y = crtc_y;
+	rplane->width = crtc_w;
+	rplane->height = crtc_h;
+
+	rcar_du_plane_compute_base(rplane, fb);
+	rcar_du_plane_setup(rplane);
+
+	mutex_lock(&rcdu->planes.lock);
+	rplane->enabled = true;
+	rcar_du_crtc_update_planes(rplane->crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	return 0;
+}
+
+static int rcar_du_plane_disable(struct drm_plane *plane)
+{
+	struct rcar_du_device *rcdu = plane->dev->dev_private;
+	struct rcar_du_plane *rplane = to_rcar_plane(plane);
+
+	if (!rplane->enabled)
+		return 0;
+
+	mutex_lock(&rcdu->planes.lock);
+	rplane->enabled = false;
+	rcar_du_crtc_update_planes(rplane->crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	rcar_du_plane_release(rplane);
+
+	rplane->crtc = NULL;
+	rplane->format = NULL;
+
+	return 0;
+}
+
+/* Both the .set_property and the .update_plane operations are called with the
+ * mode_config lock held. There is this no need to explicitly protect access to
+ * the alpha and colorkey fields and the mode register.
+ */
+static void rcar_du_plane_set_alpha(struct rcar_du_plane *plane, u32 alpha)
+{
+	if (plane->alpha == alpha)
+		return;
+
+	plane->alpha = alpha;
+	if (!plane->enabled || plane->format->fourcc != DRM_FORMAT_XRGB1555)
+		return;
+
+	rcar_du_plane_setup_mode(plane, plane->hwindex);
+}
+
+static void rcar_du_plane_set_colorkey(struct rcar_du_plane *plane,
+				       u32 colorkey)
+{
+	if (plane->colorkey == colorkey)
+		return;
+
+	plane->colorkey = colorkey;
+	if (!plane->enabled)
+		return;
+
+	rcar_du_plane_setup_mode(plane, plane->hwindex);
+}
+
+static void rcar_du_plane_set_zpos(struct rcar_du_plane *plane,
+				   unsigned int zpos)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+
+	mutex_lock(&rcdu->planes.lock);
+	if (plane->zpos == zpos)
+		goto done;
+
+	plane->zpos = zpos;
+	if (!plane->enabled)
+		goto done;
+
+	rcar_du_crtc_update_planes(plane->crtc);
+
+done:
+	mutex_unlock(&rcdu->planes.lock);
+}
+
+static int rcar_du_plane_set_property(struct drm_plane *plane,
+				      struct drm_property *property,
+				      uint64_t value)
+{
+	struct rcar_du_device *rcdu = plane->dev->dev_private;
+	struct rcar_du_plane *rplane = to_rcar_plane(plane);
+
+	if (property == rcdu->planes.alpha)
+		rcar_du_plane_set_alpha(rplane, value);
+	else if (property == rcdu->planes.colorkey)
+		rcar_du_plane_set_colorkey(rplane, value);
+	else if (property == rcdu->planes.zpos)
+		rcar_du_plane_set_zpos(rplane, value);
+	else
+		return -EINVAL;
+
+	return 0;
+}
+
+static const struct drm_plane_funcs rcar_du_plane_funcs = {
+	.update_plane = rcar_du_plane_update,
+	.disable_plane = rcar_du_plane_disable,
+	.set_property = rcar_du_plane_set_property,
+	.destroy = drm_plane_cleanup,
+};
+
+static const uint32_t formats[] = {
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV16,
+};
+
+int rcar_du_plane_init(struct rcar_du_device *rcdu)
+{
+	unsigned int i;
+
+	mutex_init(&rcdu->planes.lock);
+	rcdu->planes.free = 0xff;
+
+	rcdu->planes.alpha =
+		drm_property_create_range(rcdu->ddev, 0, "alpha", 0, 255);
+	if (rcdu->planes.alpha == NULL)
+		return -ENOMEM;
+
+	/* The color key is expressed as an RGB888 triplet stored in a 32-bit
+	 * integer in XRGB8888 format. Bit 24 is used as a flag to disable (0)
+	 * or enable source color keying (1).
+	 */
+	rcdu->planes.colorkey =
+		drm_property_create_range(rcdu->ddev, 0, "colorkey",
+					  0, 0x01ffffff);
+	if (rcdu->planes.colorkey == NULL)
+		return -ENOMEM;
+
+	rcdu->planes.zpos =
+		drm_property_create_range(rcdu->ddev, 0, "zpos", 1, 7);
+	if (rcdu->planes.zpos == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		struct rcar_du_plane *plane = &rcdu->planes.planes[i];
+
+		plane->dev = rcdu;
+		plane->hwindex = -1;
+		plane->alpha = 255;
+		plane->colorkey = RCAR_DU_COLORKEY_NONE;
+		plane->zpos = 0;
+	}
+
+	return 0;
+}
+
+int rcar_du_plane_register(struct rcar_du_device *rcdu)
+{
+	unsigned int i;
+	int ret;
+
+	for (i = 0; i < RCAR_DU_NUM_KMS_PLANES; ++i) {
+		struct rcar_du_kms_plane *plane;
+
+		plane = devm_kzalloc(rcdu->dev, sizeof(*plane), GFP_KERNEL);
+		if (plane == NULL)
+			return -ENOMEM;
+
+		plane->hwplane = &rcdu->planes.planes[i + 2];
+		plane->hwplane->zpos = 1;
+
+		ret = drm_plane_init(rcdu->ddev, &plane->plane,
+				     (1 << rcdu->num_crtcs) - 1,
+				     &rcar_du_plane_funcs, formats,
+				     ARRAY_SIZE(formats), false);
+		if (ret < 0)
+			return ret;
+
+		drm_object_attach_property(&plane->plane.base,
+					   rcdu->planes.alpha, 255);
+		drm_object_attach_property(&plane->plane.base,
+					   rcdu->planes.colorkey,
+					   RCAR_DU_COLORKEY_NONE);
+		drm_object_attach_property(&plane->plane.base,
+					   rcdu->planes.zpos, 1);
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.h b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
new file mode 100644
index 0000000..5397dba
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
@@ -0,0 +1,67 @@
+/*
+ * rcar_du_plane.h  --  R-Car Display Unit Planes
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_PLANE_H__
+#define __RCAR_DU_PLANE_H__
+
+struct drm_crtc;
+struct drm_framebuffer;
+struct rcar_du_device;
+struct rcar_du_format_info;
+
+/* The RCAR DU has 8 hardware planes, shared between KMS planes and CRTCs. As
+ * using KMS planes requires at least one of the CRTCs being enabled, no more
+ * than 7 KMS planes can be available. We thus create 7 KMS planes and
+ * 9 software planes (one for each KMS planes and one for each CRTC).
+ */
+
+#define RCAR_DU_NUM_KMS_PLANES		7
+#define RCAR_DU_NUM_HW_PLANES		8
+#define RCAR_DU_NUM_SW_PLANES		9
+
+struct rcar_du_plane {
+	struct rcar_du_device *dev;
+	struct drm_crtc *crtc;
+
+	bool enabled;
+
+	int hwindex;		/* 0-based, -1 means unused */
+	unsigned int alpha;
+	unsigned int colorkey;
+	unsigned int zpos;
+
+	const struct rcar_du_format_info *format;
+
+	unsigned long dma[2];
+	unsigned int pitch;
+
+	unsigned int width;
+	unsigned int height;
+
+	unsigned int src_x;
+	unsigned int src_y;
+	unsigned int dst_x;
+	unsigned int dst_y;
+};
+
+int rcar_du_plane_init(struct rcar_du_device *rcdu);
+int rcar_du_plane_register(struct rcar_du_device *rcdu);
+void rcar_du_plane_setup(struct rcar_du_plane *plane);
+void rcar_du_plane_update_base(struct rcar_du_plane *plane);
+void rcar_du_plane_compute_base(struct rcar_du_plane *plane,
+				struct drm_framebuffer *fb);
+int rcar_du_plane_reserve(struct rcar_du_plane *plane,
+			  const struct rcar_du_format_info *format);
+void rcar_du_plane_release(struct rcar_du_plane *plane);
+
+#endif /* __RCAR_DU_PLANE_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_regs.h b/drivers/gpu/drm/rcar-du/rcar_du_regs.h
new file mode 100644
index 0000000..69f21f1
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_regs.h
@@ -0,0 +1,445 @@
+/*
+ * rcar_du_regs.h  --  R-Car Display Unit Registers Definitions
+ *
+ * Copyright (C) 2013 Renesas Electronics Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation.
+ */
+
+#ifndef __RCAR_DU_REGS_H__
+#define __RCAR_DU_REGS_H__
+
+#define DISP2_REG_OFFSET	 0x30000
+
+/* -----------------------------------------------------------------------------
+ * Display Control Registers
+ */
+
+#define DSYSR			0x00000	/* display 1 */
+#define D2SYSR			0x30000	/* display 2 */
+#define DSYSR_ILTS		(1 << 29)
+#define DSYSR_DSEC		(1 << 20)
+#define DSYSR_IUPD		(1 << 16)
+#define DSYSR_DRES		(1 << 9)
+#define DSYSR_DEN		(1 << 8)
+#define DSYSR_TVM_MASTER	(0 << 6)
+#define DSYSR_TVM_SWITCH	(1 << 6)
+#define DSYSR_TVM_TVSYNC	(2 << 6)
+#define DSYSR_TVM_MASK		(3 << 6)
+#define DSYSR_SCM_INT_NONE	(0 << 4)
+#define DSYSR_SCM_INT_SYNC	(2 << 4)
+#define DSYSR_SCM_INT_VIDEO	(3 << 4)
+
+#define DSMR			0x00004
+#define D2SMR			0x30004
+#define DSMR_VSPM		(1 << 28)
+#define DSMR_ODPM		(1 << 27)
+#define DSMR_DIPM_DISP		(0 << 25)
+#define DSMR_DIPM_CSYNC		(1 << 25)
+#define DSMR_DIPM_DE		(3 << 25)
+#define DSMR_DIPM_MASK		(3 << 25)
+#define DSMR_CSPM		(1 << 24)
+#define DSMR_DIL		(1 << 19)
+#define DSMR_VSL		(1 << 18)
+#define DSMR_HSL		(1 << 17)
+#define DSMR_DDIS		(1 << 16)
+#define DSMR_CDEL		(1 << 15)
+#define DSMR_CDEM_CDE		(0 << 13)
+#define DSMR_CDEM_LOW		(2 << 13)
+#define DSMR_CDEM_HIGH		(3 << 13)
+#define DSMR_CDEM_MASK		(3 << 13)
+#define DSMR_CDED		(1 << 12)
+#define DSMR_ODEV		(1 << 8)
+#define DSMR_CSY_VH_OR		(0 << 6)
+#define DSMR_CSY_333		(2 << 6)
+#define DSMR_CSY_222		(3 << 6)
+#define DSMR_CSY_MASK		(3 << 6)
+
+#define DSSR			0x00008
+#define D2SSR			0x30008
+#define DSSR_VC1FB_DSA0		(0 << 30)
+#define DSSR_VC1FB_DSA1		(1 << 30)
+#define DSSR_VC1FB_DSA2		(2 << 30)
+#define DSSR_VC1FB_INIT		(3 << 30)
+#define DSSR_VC1FB_MASK		(3 << 30)
+#define DSSR_VC0FB_DSA0		(0 << 28)
+#define DSSR_VC0FB_DSA1		(1 << 28)
+#define DSSR_VC0FB_DSA2		(2 << 28)
+#define DSSR_VC0FB_INIT		(3 << 28)
+#define DSSR_VC0FB_MASK		(3 << 28)
+#define DSSR_DFB(n)		(1 << ((n)+15))
+#define DSSR_TVR		(1 << 15)
+#define DSSR_FRM		(1 << 14)
+#define DSSR_VBK		(1 << 11)
+#define DSSR_RINT		(1 << 9)
+#define DSSR_HBK		(1 << 8)
+#define DSSR_ADC(n)		(1 << ((n)-1))
+
+#define DSRCR			0x0000c
+#define D2SRCR			0x3000c
+#define DSRCR_TVCL		(1 << 15)
+#define DSRCR_FRCL		(1 << 14)
+#define DSRCR_VBCL		(1 << 11)
+#define DSRCR_RICL		(1 << 9)
+#define DSRCR_HBCL		(1 << 8)
+#define DSRCR_ADCL(n)		(1 << ((n)-1))
+#define DSRCR_MASK		0x0000cbff
+
+#define DIER			0x00010
+#define D2IER			0x30010
+#define DIER_TVE		(1 << 15)
+#define DIER_FRE		(1 << 14)
+#define DIER_VBE		(1 << 11)
+#define DIER_RIE		(1 << 9)
+#define DIER_HBE		(1 << 8)
+#define DIER_ADCE(n)		(1 << ((n)-1))
+
+#define CPCR			0x00014
+#define CPCR_CP4CE		(1 << 19)
+#define CPCR_CP3CE		(1 << 18)
+#define CPCR_CP2CE		(1 << 17)
+#define CPCR_CP1CE		(1 << 16)
+
+#define DPPR			0x00018
+#define DPPR_DPE(n)		(1 << ((n)*4-1))
+#define DPPR_DPS(n, p)		(((p)-1) << DPPR_DPS_SHIFT(n))
+#define DPPR_DPS_SHIFT(n)	(((n)-1)*4)
+#define DPPR_BPP16		(DPPR_DPE(8) | DPPR_DPS(8, 1))	/* plane1 */
+#define DPPR_BPP32_P1		(DPPR_DPE(7) | DPPR_DPS(7, 1))
+#define DPPR_BPP32_P2		(DPPR_DPE(8) | DPPR_DPS(8, 2))
+#define DPPR_BPP32		(DPPR_BPP32_P1 | DPPR_BPP32_P2)	/* plane1 & 2 */
+
+#define DEFR			0x00020
+#define D2EFR			0x30020
+#define DEFR_CODE		(0x7773 << 16)
+#define DEFR_EXSL		(1 << 12)
+#define DEFR_EXVL		(1 << 11)
+#define DEFR_EXUP		(1 << 5)
+#define DEFR_VCUP		(1 << 4)
+#define DEFR_DEFE		(1 << 0)
+
+#define DAPCR			0x00024
+#define DAPCR_CODE		(0x7773 << 16)
+#define DAPCR_AP2E		(1 << 4)
+#define DAPCR_AP1E		(1 << 0)
+
+#define DCPCR			0x00028
+#define DCPCR_CODE		(0x7773 << 16)
+#define DCPCR_CA2B		(1 << 13)
+#define DCPCR_CD2F		(1 << 12)
+#define DCPCR_DC2E		(1 << 8)
+#define DCPCR_CAB		(1 << 5)
+#define DCPCR_CDF		(1 << 4)
+#define DCPCR_DCE		(1 << 0)
+
+#define DEFR2			0x00034
+#define D2EFR2			0x30034
+#define DEFR2_CODE		(0x7775 << 16)
+#define DEFR2_DEFE2G		(1 << 0)
+
+#define DEFR3			0x00038
+#define D2EFR3			0x30038
+#define DEFR3_CODE		(0x7776 << 16)
+#define DEFR3_EVDA		(1 << 14)
+#define DEFR3_EVDM_1		(1 << 12)
+#define DEFR3_EVDM_2		(2 << 12)
+#define DEFR3_EVDM_3		(3 << 12)
+#define DEFR3_VMSM2_EMA		(1 << 6)
+#define DEFR3_VMSM1_ENA		(1 << 4)
+#define DEFR3_DEFE3		(1 << 0)
+
+#define DEFR4			0x0003c
+#define D2EFR4			0x3003c
+#define DEFR4_CODE		(0x7777 << 16)
+#define DEFR4_LRUO		(1 << 5)
+#define DEFR4_SPCE		(1 << 4)
+
+#define DVCSR			0x000d0
+#define DVCSR_VCnFB2_DSA0(n)	(0 << ((n)*2+16))
+#define DVCSR_VCnFB2_DSA1(n)	(1 << ((n)*2+16))
+#define DVCSR_VCnFB2_DSA2(n)	(2 << ((n)*2+16))
+#define DVCSR_VCnFB2_INIT(n)	(3 << ((n)*2+16))
+#define DVCSR_VCnFB2_MASK(n)	(3 << ((n)*2+16))
+#define DVCSR_VCnFB_DSA0(n)	(0 << ((n)*2))
+#define DVCSR_VCnFB_DSA1(n)	(1 << ((n)*2))
+#define DVCSR_VCnFB_DSA2(n)	(2 << ((n)*2))
+#define DVCSR_VCnFB_INIT(n)	(3 << ((n)*2))
+#define DVCSR_VCnFB_MASK(n)	(3 << ((n)*2))
+
+#define DEFR5			0x000e0
+#define DEFR5_CODE		(0x66 << 24)
+#define DEFR5_YCRGB2_DIS	(0 << 14)
+#define DEFR5_YCRGB2_PRI1	(1 << 14)
+#define DEFR5_YCRGB2_PRI2	(2 << 14)
+#define DEFR5_YCRGB2_PRI3	(3 << 14)
+#define DEFR5_YCRGB2_MASK	(3 << 14)
+#define DEFR5_YCRGB1_DIS	(0 << 12)
+#define DEFR5_YCRGB1_PRI1	(1 << 12)
+#define DEFR5_YCRGB1_PRI2	(2 << 12)
+#define DEFR5_YCRGB1_PRI3	(3 << 12)
+#define DEFR5_YCRGB1_MASK	(3 << 12)
+#define DEFR5_DEFE5		(1 << 0)
+
+#define DDLTR			0x000e4
+#define DDLTR_CODE		(0x7766 << 16)
+#define DDLTR_DLAR2		(1 << 6)
+#define DDLTR_DLAY2		(1 << 5)
+#define DDLTR_DLAY1		(1 << 1)
+
+#define DEFR6			0x000e8
+#define DEFR6_CODE		(0x7778 << 16)
+#define DEFR6_ODPM22_D2SMR	(0 << 10)
+#define DEFR6_ODPM22_DISP	(2 << 10)
+#define DEFR6_ODPM22_CDE	(3 << 10)
+#define DEFR6_ODPM22_MASK	(3 << 10)
+#define DEFR6_ODPM12_DSMR	(0 << 8)
+#define DEFR6_ODPM12_DISP	(2 << 8)
+#define DEFR6_ODPM12_CDE	(3 << 8)
+#define DEFR6_ODPM12_MASK	(3 << 8)
+#define DEFR6_TCNE2		(1 << 6)
+#define DEFR6_MLOS1		(1 << 2)
+#define DEFR6_DEFAULT		(DEFR6_CODE | DEFR6_TCNE2)
+
+/* -----------------------------------------------------------------------------
+ * Display Timing Generation Registers
+ */
+
+#define HDSR			0x00040
+#define HDER			0x00044
+#define VDSR			0x00048
+#define VDER			0x0004c
+#define HCR			0x00050
+#define HSWR			0x00054
+#define VCR			0x00058
+#define VSPR			0x0005c
+#define EQWR			0x00060
+#define SPWR			0x00064
+#define CLAMPSR			0x00070
+#define CLAMPWR			0x00074
+#define DESR			0x00078
+#define DEWR			0x0007c
+
+/* -----------------------------------------------------------------------------
+ * Display Attribute Registers
+ */
+
+#define CP1TR			0x00080
+#define CP2TR			0x00084
+#define CP3TR			0x00088
+#define CP4TR			0x0008c
+
+#define DOOR			0x00090
+#define DOOR_RGB(r, g, b)	(((r) << 18) | ((g) << 10) | ((b) << 2))
+#define CDER			0x00094
+#define CDER_RGB(r, g, b)	(((r) << 18) | ((g) << 10) | ((b) << 2))
+#define BPOR			0x00098
+#define BPOR_RGB(r, g, b)	(((r) << 18) | ((g) << 10) | ((b) << 2))
+
+#define RINTOFSR		0x0009c
+
+#define DSHPR			0x000c8
+#define DSHPR_CODE		(0x7776 << 16)
+#define DSHPR_PRIH		(0xa << 4)
+#define DSHPR_PRIL_BPP16	(0x8 << 0)
+#define DSHPR_PRIL_BPP32	(0x9 << 0)
+
+/* -----------------------------------------------------------------------------
+ * Display Plane Registers
+ */
+
+#define PLANE_OFF		0x00100
+
+#define PnMR			0x00100 /* plane 1 */
+#define PnMR_VISL_VIN0		(0 << 26)	/* use Video Input 0 */
+#define PnMR_VISL_VIN1		(1 << 26)	/* use Video Input 1 */
+#define PnMR_VISL_VIN2		(2 << 26)	/* use Video Input 2 */
+#define PnMR_VISL_VIN3		(3 << 26)	/* use Video Input 3 */
+#define PnMR_YCDF_YUYV		(1 << 20)	/* YUYV format */
+#define PnMR_TC_R		(0 << 17)	/* Tranparent color is PnTC1R */
+#define PnMR_TC_CP		(1 << 17)	/* Tranparent color is color palette */
+#define PnMR_WAE		(1 << 16)	/* Wrap around Enable */
+#define PnMR_SPIM_TP		(0 << 12)	/* Transparent Color */
+#define PnMR_SPIM_ALP		(1 << 12)	/* Alpha Blending */
+#define PnMR_SPIM_EOR		(2 << 12)	/* EOR */
+#define PnMR_SPIM_TP_OFF	(1 << 14)	/* No Transparent Color */
+#define PnMR_CPSL_CP1		(0 << 8)	/* Color Palette selected 1 */
+#define PnMR_CPSL_CP2		(1 << 8)	/* Color Palette selected 2 */
+#define PnMR_CPSL_CP3		(2 << 8)	/* Color Palette selected 3 */
+#define PnMR_CPSL_CP4		(3 << 8)	/* Color Palette selected 4 */
+#define PnMR_DC			(1 << 7)	/* Display Area Change */
+#define PnMR_BM_MD		(0 << 4)	/* Manual Display Change Mode */
+#define PnMR_BM_AR		(1 << 4)	/* Auto Rendering Mode */
+#define PnMR_BM_AD		(2 << 4)	/* Auto Display Change Mode */
+#define PnMR_BM_VC		(3 << 4)	/* Video Capture Mode */
+#define PnMR_DDDF_8BPP		(0 << 0)	/* 8bit */
+#define PnMR_DDDF_16BPP		(1 << 0)	/* 16bit or 32bit */
+#define PnMR_DDDF_ARGB		(2 << 0)	/* ARGB */
+#define PnMR_DDDF_YC		(3 << 0)	/* YC */
+#define PnMR_DDDF_MASK		(3 << 0)
+
+#define PnMWR			0x00104
+
+#define PnALPHAR		0x00108
+#define PnALPHAR_ABIT_1		(0 << 12)
+#define PnALPHAR_ABIT_0		(1 << 12)
+#define PnALPHAR_ABIT_X		(2 << 12)
+
+#define PnDSXR			0x00110
+#define PnDSYR			0x00114
+#define PnDPXR			0x00118
+#define PnDPYR			0x0011c
+
+#define PnDSA0R			0x00120
+#define PnDSA1R			0x00124
+#define PnDSA2R			0x00128
+#define PnDSA_MASK		0xfffffff0
+
+#define PnSPXR			0x00130
+#define PnSPYR			0x00134
+#define PnWASPR			0x00138
+#define PnWAMWR			0x0013c
+
+#define PnBTR			0x00140
+
+#define PnTC1R			0x00144
+#define PnTC2R			0x00148
+#define PnTC3R			0x0014c
+#define PnTC3R_CODE		(0x66 << 24)
+
+#define PnMLR			0x00150
+
+#define PnSWAPR			0x00180
+#define PnSWAPR_DIGN		(1 << 4)
+#define PnSWAPR_SPQW		(1 << 3)
+#define PnSWAPR_SPLW		(1 << 2)
+#define PnSWAPR_SPWD		(1 << 1)
+#define PnSWAPR_SPBY		(1 << 0)
+
+#define PnDDCR			0x00184
+#define PnDDCR_CODE		(0x7775 << 16)
+#define PnDDCR_LRGB1		(1 << 11)
+#define PnDDCR_LRGB0		(1 << 10)
+
+#define PnDDCR2			0x00188
+#define PnDDCR2_CODE		(0x7776 << 16)
+#define PnDDCR2_NV21		(1 << 5)
+#define PnDDCR2_Y420		(1 << 4)
+#define PnDDCR2_DIVU		(1 << 1)
+#define PnDDCR2_DIVY		(1 << 0)
+
+#define PnDDCR4			0x00190
+#define PnDDCR4_CODE		(0x7766 << 16)
+#define PnDDCR4_SDFS_RGB	(0 << 4)
+#define PnDDCR4_SDFS_YC		(5 << 4)
+#define PnDDCR4_SDFS_MASK	(7 << 4)
+#define PnDDCR4_EDF_NONE	(0 << 0)
+#define PnDDCR4_EDF_ARGB8888	(1 << 0)
+#define PnDDCR4_EDF_RGB888	(2 << 0)
+#define PnDDCR4_EDF_RGB666	(3 << 0)
+#define PnDDCR4_EDF_MASK	(7 << 0)
+
+#define APnMR			0x0a100
+#define APnMR_WAE		(1 << 16)	/* Wrap around Enable */
+#define APnMR_DC		(1 << 7)	/* Display Area Change */
+#define APnMR_BM_MD		(0 << 4)	/* Manual Display Change Mode */
+#define APnMR_BM_AD		(2 << 4)	/* Auto Display Change Mode */
+
+#define APnMWR			0x0a104
+#define APnDSA0R		0x0a120
+#define APnDSA1R		0x0a124
+#define APnDSA2R		0x0a128
+#define APnMLR			0x0a150
+
+/* -----------------------------------------------------------------------------
+ * Display Capture Registers
+ */
+
+#define DCMWR			0x0c104
+#define DC2MWR			0x0c204
+#define DCSAR			0x0c120
+#define DC2SAR			0x0c220
+#define DCMLR			0x0c150
+#define DC2MLR			0x0c250
+
+/* -----------------------------------------------------------------------------
+ * Color Palette Registers
+ */
+
+#define CP1_000R		0x01000
+#define CP1_255R		0x013fc
+#define CP2_000R		0x02000
+#define CP2_255R		0x023fc
+#define CP3_000R		0x03000
+#define CP3_255R		0x033fc
+#define CP4_000R		0x04000
+#define CP4_255R		0x043fc
+
+/* -----------------------------------------------------------------------------
+ * External Synchronization Control Registers
+ */
+
+#define ESCR			0x10000
+#define ESCR2			0x31000
+#define ESCR_DCLKOINV		(1 << 25)
+#define ESCR_DCLKSEL_DCLKIN	(0 << 20)
+#define ESCR_DCLKSEL_CLKS	(1 << 20)
+#define ESCR_DCLKSEL_MASK	(1 << 20)
+#define ESCR_DCLKDIS		(1 << 16)
+#define ESCR_SYNCSEL_OFF	(0 << 8)
+#define ESCR_SYNCSEL_EXVSYNC	(2 << 8)
+#define ESCR_SYNCSEL_EXHSYNC	(3 << 8)
+#define ESCR_FRQSEL_MASK	(0x3f << 0)
+
+#define OTAR			0x10004
+#define OTAR2			0x31004
+
+/* -----------------------------------------------------------------------------
+ * Dual Display Output Control Registers
+ */
+
+#define DORCR			0x11000
+#define DORCR_PG2T		(1 << 30)
+#define DORCR_DK2S		(1 << 28)
+#define DORCR_PG2D_DS1		(0 << 24)
+#define DORCR_PG2D_DS2		(1 << 24)
+#define DORCR_PG2D_FIX0		(2 << 24)
+#define DORCR_PG2D_DOOR		(3 << 24)
+#define DORCR_PG2D_MASK		(3 << 24)
+#define DORCR_DR1D		(1 << 21)
+#define DORCR_PG1D_DS1		(0 << 16)
+#define DORCR_PG1D_DS2		(1 << 16)
+#define DORCR_PG1D_FIX0		(2 << 16)
+#define DORCR_PG1D_DOOR		(3 << 16)
+#define DORCR_PG1D_MASK		(3 << 16)
+#define DORCR_RGPV		(1 << 4)
+#define DORCR_DPRS		(1 << 0)
+
+#define DPTSR			0x11004
+#define DPTSR_PnDK(n)		(1 << ((n) + 16))
+#define DPTSR_PnTS(n)		(1 << (n))
+
+#define DAPTSR			0x11008
+#define DAPTSR_APnDK(n)		(1 << ((n) + 16))
+#define DAPTSR_APnTS(n)		(1 << (n))
+
+#define DS1PR			0x11020
+#define DS2PR			0x11024
+
+/* -----------------------------------------------------------------------------
+ * YC-RGB Conversion Coefficient Registers
+ */
+
+#define YNCR			0x11080
+#define YNOR			0x11084
+#define CRNOR			0x11088
+#define CBNOR			0x1108c
+#define RCRCR			0x11090
+#define GCRCR			0x11094
+#define GCBCR			0x11098
+#define BCBCR			0x1109c
+
+#endif /* __RCAR_DU_REGS_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vga.c b/drivers/gpu/drm/rcar-du/rcar_du_vga.c
new file mode 100644
index 0000000..327289e
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vga.c
@@ -0,0 +1,149 @@
+/*
+ * rcar_du_vga.c  --  R-Car Display Unit VGA DAC and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_vga.h"
+
+/* -----------------------------------------------------------------------------
+ * Connector
+ */
+
+static int rcar_du_vga_connector_get_modes(struct drm_connector *connector)
+{
+	return 0;
+}
+
+static int rcar_du_vga_connector_mode_valid(struct drm_connector *connector,
+					    struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static const struct drm_connector_helper_funcs connector_helper_funcs = {
+	.get_modes = rcar_du_vga_connector_get_modes,
+	.mode_valid = rcar_du_vga_connector_mode_valid,
+	.best_encoder = rcar_du_connector_best_encoder,
+};
+
+static void rcar_du_vga_connector_destroy(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+}
+
+static enum drm_connector_status
+rcar_du_vga_connector_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_unknown;
+}
+
+static const struct drm_connector_funcs connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = rcar_du_vga_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = rcar_du_vga_connector_destroy,
+};
+
+static int rcar_du_vga_connector_init(struct rcar_du_device *rcdu,
+				      struct rcar_du_encoder *renc)
+{
+	struct rcar_du_connector *rcon;
+	struct drm_connector *connector;
+	int ret;
+
+	rcon = devm_kzalloc(rcdu->dev, sizeof(*rcon), GFP_KERNEL);
+	if (rcon == NULL)
+		return -ENOMEM;
+
+	connector = &rcon->connector;
+	connector->display_info.width_mm = 0;
+	connector->display_info.height_mm = 0;
+
+	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
+				 DRM_MODE_CONNECTOR_VGA);
+	if (ret < 0)
+		return ret;
+
+	drm_connector_helper_add(connector, &connector_helper_funcs);
+	ret = drm_sysfs_connector_add(connector);
+	if (ret < 0)
+		return ret;
+
+	drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	drm_object_property_set_value(&connector->base,
+		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
+
+	ret = drm_mode_connector_attach_encoder(connector, &renc->encoder);
+	if (ret < 0)
+		return ret;
+
+	connector->encoder = &renc->encoder;
+	rcon->encoder = renc;
+
+	return 0;
+}
+
+/* -----------------------------------------------------------------------------
+ * Encoder
+ */
+
+static void rcar_du_vga_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool rcar_du_vga_encoder_mode_fixup(struct drm_encoder *encoder,
+					   const struct drm_display_mode *mode,
+					   struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs encoder_helper_funcs = {
+	.dpms = rcar_du_vga_encoder_dpms,
+	.mode_fixup = rcar_du_vga_encoder_mode_fixup,
+	.prepare = rcar_du_encoder_mode_prepare,
+	.commit = rcar_du_encoder_mode_commit,
+	.mode_set = rcar_du_encoder_mode_set,
+};
+
+static const struct drm_encoder_funcs encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+int rcar_du_vga_init(struct rcar_du_device *rcdu,
+		     const struct rcar_du_encoder_vga_data *data,
+		     unsigned int output)
+{
+	struct rcar_du_encoder *renc;
+	int ret;
+
+	renc = devm_kzalloc(rcdu->dev, sizeof(*renc), GFP_KERNEL);
+	if (renc == NULL)
+		return -ENOMEM;
+
+	renc->output = output;
+
+	ret = drm_encoder_init(rcdu->ddev, &renc->encoder, &encoder_funcs,
+			       DRM_MODE_ENCODER_DAC);
+	if (ret < 0)
+		return ret;
+
+	drm_encoder_helper_add(&renc->encoder, &encoder_helper_funcs);
+
+	return rcar_du_vga_connector_init(rcdu, renc);
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vga.h b/drivers/gpu/drm/rcar-du/rcar_du_vga.h
new file mode 100644
index 0000000..66b4d2d
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vga.h
@@ -0,0 +1,24 @@
+/*
+ * rcar_du_vga.h  --  R-Car Display Unit VGA DAC and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_VGA_H__
+#define __RCAR_DU_VGA_H__
+
+struct rcar_du_device;
+struct rcar_du_encoder_vga_data;
+
+int rcar_du_vga_init(struct rcar_du_device *rcdu,
+		     const struct rcar_du_encoder_vga_data *data,
+		     unsigned int output);
+
+#endif /* __RCAR_DU_VGA_H__ */
diff --git a/drivers/gpu/drm/savage/savage_bci.c b/drivers/gpu/drm/savage/savage_bci.c
index b55c1d6..bd6b2cf 100644
--- a/drivers/gpu/drm/savage/savage_bci.c
+++ b/drivers/gpu/drm/savage/savage_bci.c
@@ -570,9 +570,6 @@ int savage_driver_firstopen(struct drm_device *dev)
 	unsigned int fb_rsrc, aper_rsrc;
 	int ret = 0;
 
-	dev_priv->mtrr[0].handle = -1;
-	dev_priv->mtrr[1].handle = -1;
-	dev_priv->mtrr[2].handle = -1;
 	if (S3_SAVAGE3D_SERIES(dev_priv->chipset)) {
 		fb_rsrc = 0;
 		fb_base = pci_resource_start(dev->pdev, 0);
@@ -584,21 +581,14 @@ int savage_driver_firstopen(struct drm_device *dev)
 		if (pci_resource_len(dev->pdev, 0) == 0x08000000) {
 			/* Don't make MMIO write-cobining! We need 3
 			 * MTRRs. */
-			dev_priv->mtrr[0].base = fb_base;
-			dev_priv->mtrr[0].size = 0x01000000;
-			dev_priv->mtrr[0].handle =
-			    drm_mtrr_add(dev_priv->mtrr[0].base,
-				         dev_priv->mtrr[0].size, DRM_MTRR_WC);
-			dev_priv->mtrr[1].base = fb_base + 0x02000000;
-			dev_priv->mtrr[1].size = 0x02000000;
-			dev_priv->mtrr[1].handle =
-			    drm_mtrr_add(dev_priv->mtrr[1].base,
-					 dev_priv->mtrr[1].size, DRM_MTRR_WC);
-			dev_priv->mtrr[2].base = fb_base + 0x04000000;
-			dev_priv->mtrr[2].size = 0x04000000;
-			dev_priv->mtrr[2].handle =
-			    drm_mtrr_add(dev_priv->mtrr[2].base,
-					 dev_priv->mtrr[2].size, DRM_MTRR_WC);
+			dev_priv->mtrr_handles[0] =
+				arch_phys_wc_add(fb_base, 0x01000000);
+			dev_priv->mtrr_handles[1] =
+				arch_phys_wc_add(fb_base + 0x02000000,
+						 0x02000000);
+			dev_priv->mtrr_handles[2] =
+				arch_phys_wc_add(fb_base + 0x04000000,
+						0x04000000);
 		} else {
 			DRM_ERROR("strange pci_resource_len %08llx\n",
 				  (unsigned long long)
@@ -616,11 +606,9 @@ int savage_driver_firstopen(struct drm_device *dev)
 		if (pci_resource_len(dev->pdev, 1) == 0x08000000) {
 			/* Can use one MTRR to cover both fb and
 			 * aperture. */
-			dev_priv->mtrr[0].base = fb_base;
-			dev_priv->mtrr[0].size = 0x08000000;
-			dev_priv->mtrr[0].handle =
-			    drm_mtrr_add(dev_priv->mtrr[0].base,
-					 dev_priv->mtrr[0].size, DRM_MTRR_WC);
+			dev_priv->mtrr_handles[0] =
+				arch_phys_wc_add(fb_base,
+						 0x08000000);
 		} else {
 			DRM_ERROR("strange pci_resource_len %08llx\n",
 				  (unsigned long long)
@@ -660,11 +648,10 @@ void savage_driver_lastclose(struct drm_device *dev)
 	drm_savage_private_t *dev_priv = dev->dev_private;
 	int i;
 
-	for (i = 0; i < 3; ++i)
-		if (dev_priv->mtrr[i].handle >= 0)
-			drm_mtrr_del(dev_priv->mtrr[i].handle,
-				 dev_priv->mtrr[i].base,
-				 dev_priv->mtrr[i].size, DRM_MTRR_WC);
+	for (i = 0; i < 3; ++i) {
+		arch_phys_wc_del(dev_priv->mtrr_handles[i]);
+		dev_priv->mtrr_handles[i] = 0;
+	}
 }
 
 int savage_driver_unload(struct drm_device *dev)
diff --git a/drivers/gpu/drm/savage/savage_drv.h b/drivers/gpu/drm/savage/savage_drv.h
index df2aac6..c05082a 100644
--- a/drivers/gpu/drm/savage/savage_drv.h
+++ b/drivers/gpu/drm/savage/savage_drv.h
@@ -160,10 +160,7 @@ typedef struct drm_savage_private {
 	drm_local_map_t *cmd_dma;
 	drm_local_map_t fake_dma;
 
-	struct {
-		int handle;
-		unsigned long base, size;
-	} mtrr[3];
+	int mtrr_handles[3];
 
 	/* BCI and status-related stuff */
 	volatile uint32_t *status_ptr, *bci_ptr;
diff --git a/drivers/gpu/drm/shmobile/Kconfig b/drivers/gpu/drm/shmobile/Kconfig
index 7e7d52b..ca498d1 100644
--- a/drivers/gpu/drm/shmobile/Kconfig
+++ b/drivers/gpu/drm/shmobile/Kconfig
@@ -1,6 +1,6 @@
 config DRM_SHMOBILE
 	tristate "DRM Support for SH Mobile"
-	depends on DRM && (SUPERH || ARCH_SHMOBILE)
+	depends on DRM && (ARM || SUPERH)
 	select DRM_KMS_HELPER
 	select DRM_KMS_CMA_HELPER
 	select DRM_GEM_CMA_HELPER
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.c b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
index f6e0b53..edc1018 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
@@ -90,7 +90,7 @@ static int shmob_drm_setup_clocks(struct shmob_drm_device *sdev,
 		return -EINVAL;
 	}
 
-	clk = clk_get(sdev->dev, clkname);
+	clk = devm_clk_get(sdev->dev, clkname);
 	if (IS_ERR(clk)) {
 		dev_err(sdev->dev, "cannot get dot clock %s\n", clkname);
 		return PTR_ERR(clk);
@@ -106,21 +106,12 @@ static int shmob_drm_setup_clocks(struct shmob_drm_device *sdev,
 
 static int shmob_drm_unload(struct drm_device *dev)
 {
-	struct shmob_drm_device *sdev = dev->dev_private;
-
 	drm_kms_helper_poll_fini(dev);
 	drm_mode_config_cleanup(dev);
 	drm_vblank_cleanup(dev);
 	drm_irq_uninstall(dev);
 
-	if (sdev->clock)
-		clk_put(sdev->clock);
-
-	if (sdev->mmio)
-		iounmap(sdev->mmio);
-
 	dev->dev_private = NULL;
-	kfree(sdev);
 
 	return 0;
 }
@@ -139,7 +130,7 @@ static int shmob_drm_load(struct drm_device *dev, unsigned long flags)
 		return -EINVAL;
 	}
 
-	sdev = kzalloc(sizeof(*sdev), GFP_KERNEL);
+	sdev = devm_kzalloc(&pdev->dev, sizeof(*sdev), GFP_KERNEL);
 	if (sdev == NULL) {
 		dev_err(dev->dev, "failed to allocate private data\n");
 		return -ENOMEM;
@@ -156,29 +147,28 @@ static int shmob_drm_load(struct drm_device *dev, unsigned long flags)
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (res == NULL) {
 		dev_err(&pdev->dev, "failed to get memory resource\n");
-		ret = -EINVAL;
-		goto done;
+		return -EINVAL;
 	}
 
-	sdev->mmio = ioremap_nocache(res->start, resource_size(res));
+	sdev->mmio = devm_ioremap_nocache(&pdev->dev, res->start,
+					  resource_size(res));
 	if (sdev->mmio == NULL) {
 		dev_err(&pdev->dev, "failed to remap memory resource\n");
-		ret = -ENOMEM;
-		goto done;
+		return -ENOMEM;
 	}
 
 	ret = shmob_drm_setup_clocks(sdev, pdata->clk_source);
 	if (ret < 0)
-		goto done;
+		return ret;
 
 	ret = shmob_drm_init_interface(sdev);
 	if (ret < 0)
-		goto done;
+		return ret;
 
 	ret = shmob_drm_modeset_init(sdev);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to initialize mode setting\n");
-		goto done;
+		return ret;
 	}
 
 	for (i = 0; i < 4; ++i) {
@@ -273,7 +263,8 @@ static const struct file_operations shmob_drm_fops = {
 };
 
 static struct drm_driver shmob_drm_driver = {
-	.driver_features	= DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET,
+	.driver_features	= DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET
+				| DRIVER_PRIME,
 	.load			= shmob_drm_load,
 	.unload			= shmob_drm_unload,
 	.preclose		= shmob_drm_preclose,
@@ -283,6 +274,10 @@ static struct drm_driver shmob_drm_driver = {
 	.disable_vblank		= shmob_drm_disable_vblank,
 	.gem_free_object	= drm_gem_cma_free_object,
 	.gem_vm_ops		= &drm_gem_cma_vm_ops,
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_cma_dmabuf_import,
+	.gem_prime_export	= drm_gem_cma_dmabuf_export,
 	.dumb_create		= drm_gem_cma_dumb_create,
 	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,
 	.dumb_destroy		= drm_gem_cma_dumb_destroy,
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_kms.c b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
index c291ee3..fc0ef0c 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_kms.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
@@ -116,7 +116,7 @@ shmob_drm_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 	}
 
 	if (mode_cmd->pitches[0] & 7 || mode_cmd->pitches[0] >= 65536) {
-		dev_dbg(dev->dev, "valid pitch value %u\n",
+		dev_dbg(dev->dev, "invalid pitch value %u\n",
 			mode_cmd->pitches[0]);
 		return ERR_PTR(-EINVAL);
 	}
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_plane.c b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
index e1eb899..060ae03 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_plane.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
@@ -166,7 +166,7 @@ void shmob_drm_plane_setup(struct drm_plane *plane)
 {
 	struct shmob_drm_plane *splane = to_shmob_plane(plane);
 
-	if (plane->fb == NULL || !plane->enabled)
+	if (plane->fb == NULL)
 		return;
 
 	__shmob_drm_plane_setup(splane, plane->fb);
@@ -221,11 +221,8 @@ static int shmob_drm_plane_disable(struct drm_plane *plane)
 
 static void shmob_drm_plane_destroy(struct drm_plane *plane)
 {
-	struct shmob_drm_plane *splane = to_shmob_plane(plane);
-
 	shmob_drm_plane_disable(plane);
 	drm_plane_cleanup(plane);
-	kfree(splane);
 }
 
 static const struct drm_plane_funcs shmob_drm_plane_funcs = {
@@ -251,7 +248,7 @@ int shmob_drm_plane_create(struct shmob_drm_device *sdev, unsigned int index)
 	struct shmob_drm_plane *splane;
 	int ret;
 
-	splane = kzalloc(sizeof(*splane), GFP_KERNEL);
+	splane = devm_kzalloc(sdev->dev, sizeof(*splane), GFP_KERNEL);
 	if (splane == NULL)
 		return -ENOMEM;
 
@@ -261,8 +258,6 @@ int shmob_drm_plane_create(struct shmob_drm_device *sdev, unsigned int index)
 	ret = drm_plane_init(sdev->ddev, &splane->plane, 1,
 			     &shmob_drm_plane_funcs, formats,
 			     ARRAY_SIZE(formats), false);
-	if (ret < 0)
-		kfree(splane);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
index 5dd3c7d..7418dcd 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
@@ -42,7 +42,8 @@ struct tilcdc_crtc {
 
 static void unref_worker(struct work_struct *work)
 {
-	struct tilcdc_crtc *tilcdc_crtc = container_of(work, struct tilcdc_crtc, work);
+	struct tilcdc_crtc *tilcdc_crtc =
+		container_of(work, struct tilcdc_crtc, work);
 	struct drm_device *dev = tilcdc_crtc->base.dev;
 	struct drm_framebuffer *fb;
 
@@ -55,10 +56,12 @@ static void unref_worker(struct work_struct *work)
 static void set_scanout(struct drm_crtc *crtc, int n)
 {
 	static const uint32_t base_reg[] = {
-			LCDC_DMA_FB_BASE_ADDR_0_REG, LCDC_DMA_FB_BASE_ADDR_1_REG,
+			LCDC_DMA_FB_BASE_ADDR_0_REG,
+			LCDC_DMA_FB_BASE_ADDR_1_REG,
 	};
 	static const uint32_t ceil_reg[] = {
-			LCDC_DMA_FB_CEILING_ADDR_0_REG, LCDC_DMA_FB_CEILING_ADDR_1_REG,
+			LCDC_DMA_FB_CEILING_ADDR_0_REG,
+			LCDC_DMA_FB_CEILING_ADDR_1_REG,
 	};
 	static const uint32_t stat[] = {
 			LCDC_END_OF_FRAME0, LCDC_END_OF_FRAME1,
@@ -194,7 +197,8 @@ static void tilcdc_crtc_dpms(struct drm_crtc *crtc, int mode)
 		tilcdc_crtc->frame_done = false;
 		stop(crtc);
 
-		/* if necessary wait for framedone irq which will still come
+		/*
+		 * if necessary wait for framedone irq which will still come
 		 * before putting things to sleep..
 		 */
 		if (priv->rev == 2) {
@@ -289,17 +293,24 @@ static int tilcdc_crtc_mode_set(struct drm_crtc *crtc,
 	reg = tilcdc_read(dev, LCDC_RASTER_TIMING_2_REG) & ~0x000fff00;
 	reg |= LCDC_AC_BIAS_FREQUENCY(info->ac_bias) |
 		LCDC_AC_BIAS_TRANSITIONS_PER_INT(info->ac_bias_intrpt);
+
+	/*
+	 * subtract one from hfp, hbp, hsw because the hardware uses
+	 * a value of 0 as 1
+	 */
 	if (priv->rev == 2) {
-		reg |= (hfp & 0x300) >> 8;
-		reg |= (hbp & 0x300) >> 4;
-		reg |= (hsw & 0x3c0) << 21;
+		/* clear bits we're going to set */
+		reg &= ~0x78000033;
+		reg |= ((hfp-1) & 0x300) >> 8;
+		reg |= ((hbp-1) & 0x300) >> 4;
+		reg |= ((hsw-1) & 0x3c0) << 21;
 	}
 	tilcdc_write(dev, LCDC_RASTER_TIMING_2_REG, reg);
 
 	reg = (((mode->hdisplay >> 4) - 1) << 4) |
-		((hbp & 0xff) << 24) |
-		((hfp & 0xff) << 16) |
-		((hsw & 0x3f) << 10);
+		(((hbp-1) & 0xff) << 24) |
+		(((hfp-1) & 0xff) << 16) |
+		(((hsw-1) & 0x3f) << 10);
 	if (priv->rev == 2)
 		reg |= (((mode->hdisplay >> 4) - 1) & 0x40) >> 3;
 	tilcdc_write(dev, LCDC_RASTER_TIMING_0_REG, reg);
@@ -307,9 +318,24 @@ static int tilcdc_crtc_mode_set(struct drm_crtc *crtc,
 	reg = ((mode->vdisplay - 1) & 0x3ff) |
 		((vbp & 0xff) << 24) |
 		((vfp & 0xff) << 16) |
-		((vsw & 0x3f) << 10);
+		(((vsw-1) & 0x3f) << 10);
 	tilcdc_write(dev, LCDC_RASTER_TIMING_1_REG, reg);
 
+	/*
+	 * be sure to set Bit 10 for the V2 LCDC controller,
+	 * otherwise limited to 1024 pixels width, stopping
+	 * 1920x1080 being suppoted.
+	 */
+	if (priv->rev == 2) {
+		if ((mode->vdisplay - 1) & 0x400) {
+			tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG,
+				LCDC_LPP_B10);
+		} else {
+			tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG,
+				LCDC_LPP_B10);
+		}
+	}
+
 	/* Configure display type: */
 	reg = tilcdc_read(dev, LCDC_RASTER_CTRL_REG) &
 		~(LCDC_TFT_MODE | LCDC_MONO_8BIT_MODE | LCDC_MONOCHROME_MODE |
@@ -384,10 +410,6 @@ static int tilcdc_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	return 0;
 }
 
-static void tilcdc_crtc_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static const struct drm_crtc_funcs tilcdc_crtc_funcs = {
 		.destroy        = tilcdc_crtc_destroy,
 		.set_config     = drm_crtc_helper_set_config,
@@ -401,7 +423,6 @@ static const struct drm_crtc_helper_funcs tilcdc_crtc_helper_funcs = {
 		.commit         = tilcdc_crtc_commit,
 		.mode_set       = tilcdc_crtc_mode_set,
 		.mode_set_base  = tilcdc_crtc_mode_set_base,
-		.load_lut       = tilcdc_crtc_load_lut,
 };
 
 int tilcdc_crtc_max_width(struct drm_crtc *crtc)
@@ -422,7 +443,12 @@ int tilcdc_crtc_mode_valid(struct drm_crtc *crtc, struct drm_display_mode *mode)
 {
 	struct tilcdc_drm_private *priv = crtc->dev->dev_private;
 	unsigned int bandwidth;
+	uint32_t hbp, hfp, hsw, vbp, vfp, vsw;
 
+	/*
+	 * check to see if the width is within the range that
+	 * the LCD Controller physically supports
+	 */
 	if (mode->hdisplay > tilcdc_crtc_max_width(crtc))
 		return MODE_VIRTUAL_X;
 
@@ -433,10 +459,70 @@ int tilcdc_crtc_mode_valid(struct drm_crtc *crtc, struct drm_display_mode *mode)
 	if (mode->vdisplay > 2048)
 		return MODE_VIRTUAL_Y;
 
+	DBG("Processing mode %dx%d@%d with pixel clock %d",
+		mode->hdisplay, mode->vdisplay,
+		drm_mode_vrefresh(mode), mode->clock);
+
+	hbp = mode->htotal - mode->hsync_end;
+	hfp = mode->hsync_start - mode->hdisplay;
+	hsw = mode->hsync_end - mode->hsync_start;
+	vbp = mode->vtotal - mode->vsync_end;
+	vfp = mode->vsync_start - mode->vdisplay;
+	vsw = mode->vsync_end - mode->vsync_start;
+
+	if ((hbp-1) & ~0x3ff) {
+		DBG("Pruning mode: Horizontal Back Porch out of range");
+		return MODE_HBLANK_WIDE;
+	}
+
+	if ((hfp-1) & ~0x3ff) {
+		DBG("Pruning mode: Horizontal Front Porch out of range");
+		return MODE_HBLANK_WIDE;
+	}
+
+	if ((hsw-1) & ~0x3ff) {
+		DBG("Pruning mode: Horizontal Sync Width out of range");
+		return MODE_HSYNC_WIDE;
+	}
+
+	if (vbp & ~0xff) {
+		DBG("Pruning mode: Vertical Back Porch out of range");
+		return MODE_VBLANK_WIDE;
+	}
+
+	if (vfp & ~0xff) {
+		DBG("Pruning mode: Vertical Front Porch out of range");
+		return MODE_VBLANK_WIDE;
+	}
+
+	if ((vsw-1) & ~0x3f) {
+		DBG("Pruning mode: Vertical Sync Width out of range");
+		return MODE_VSYNC_WIDE;
+	}
+
+	/*
+	 * some devices have a maximum allowed pixel clock
+	 * configured from the DT
+	 */
+	if (mode->clock > priv->max_pixelclock) {
+		DBG("Pruning mode: pixel clock too high");
+		return MODE_CLOCK_HIGH;
+	}
+
+	/*
+	 * some devices further limit the max horizontal resolution
+	 * configured from the DT
+	 */
+	if (mode->hdisplay > priv->max_width)
+		return MODE_BAD_WIDTH;
+
 	/* filter out modes that would require too much memory bandwidth: */
-	bandwidth = mode->hdisplay * mode->vdisplay * drm_mode_vrefresh(mode);
-	if (bandwidth > priv->max_bandwidth)
+	bandwidth = mode->hdisplay * mode->vdisplay *
+		drm_mode_vrefresh(mode);
+	if (bandwidth > priv->max_bandwidth) {
+		DBG("Pruning mode: exceeds defined bandwidth limit");
 		return MODE_BAD;
+	}
 
 	return MODE_OK;
 }
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index 2b5461b..40b71da 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -26,6 +26,7 @@
 #include "drm_fb_helper.h"
 
 static LIST_HEAD(module_list);
+static bool slave_probing;
 
 void tilcdc_module_init(struct tilcdc_module *mod, const char *name,
 		const struct tilcdc_module_ops *funcs)
@@ -41,6 +42,11 @@ void tilcdc_module_cleanup(struct tilcdc_module *mod)
 	list_del(&mod->list);
 }
 
+void tilcdc_slave_probedefer(bool defered)
+{
+	slave_probing = defered;
+}
+
 static struct of_device_id tilcdc_of_match[];
 
 static struct drm_framebuffer *tilcdc_fb_create(struct drm_device *dev,
@@ -157,7 +163,9 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	struct platform_device *pdev = dev->platformdev;
 	struct device_node *node = pdev->dev.of_node;
 	struct tilcdc_drm_private *priv;
+	struct tilcdc_module *mod;
 	struct resource *res;
+	u32 bpp = 0;
 	int ret;
 
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
@@ -210,7 +218,20 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 #endif
 
 	if (of_property_read_u32(node, "max-bandwidth", &priv->max_bandwidth))
-		priv->max_bandwidth = 1280 * 1024 * 60;
+		priv->max_bandwidth = TILCDC_DEFAULT_MAX_BANDWIDTH;
+
+	DBG("Maximum Bandwidth Value %d", priv->max_bandwidth);
+
+	if (of_property_read_u32(node, "ti,max-width", &priv->max_width))
+		priv->max_width = TILCDC_DEFAULT_MAX_WIDTH;
+
+	DBG("Maximum Horizontal Pixel Width Value %dpixels", priv->max_width);
+
+	if (of_property_read_u32(node, "ti,max-pixelclock",
+					&priv->max_pixelclock))
+		priv->max_pixelclock = TILCDC_DEFAULT_MAX_PIXELCLOCK;
+
+	DBG("Maximum Pixel Clock Value %dKHz", priv->max_pixelclock);
 
 	pm_runtime_enable(dev->dev);
 
@@ -256,7 +277,15 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 
 	platform_set_drvdata(pdev, dev);
 
-	priv->fbdev = drm_fbdev_cma_init(dev, 16,
+
+	list_for_each_entry(mod, &module_list, list) {
+		DBG("%s: preferred_bpp: %d", mod->name, mod->preferred_bpp);
+		bpp = mod->preferred_bpp;
+		if (bpp > 0)
+			break;
+	}
+
+	priv->fbdev = drm_fbdev_cma_init(dev, bpp,
 			dev->mode_config.num_crtc,
 			dev->mode_config.num_connector);
 
@@ -557,6 +586,10 @@ static int tilcdc_pdev_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
+	/* defer probing if slave is in deferred probing */
+	if (slave_probing == true)
+		return -EPROBE_DEFER;
+
 	return drm_platform_init(&tilcdc_driver, pdev);
 }
 
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.h b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
index 8242b5a..0938036 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
@@ -34,6 +34,18 @@
 #include <drm/drm_gem_cma_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 
+/* Defaulting to pixel clock defined on AM335x */
+#define TILCDC_DEFAULT_MAX_PIXELCLOCK  126000
+/* Defaulting to max width as defined on AM335x */
+#define TILCDC_DEFAULT_MAX_WIDTH  2048
+/*
+ * This may need some tweaking, but want to allow at least 1280x1024@60
+ * with optimized DDR & EMIF settings tweaked 1920x1080@24 appears to
+ * be supportable
+ */
+#define TILCDC_DEFAULT_MAX_BANDWIDTH  (1280*1024*60)
+
+
 struct tilcdc_drm_private {
 	void __iomem *mmio;
 
@@ -43,6 +55,16 @@ struct tilcdc_drm_private {
 
 	/* don't attempt resolutions w/ higher W * H * Hz: */
 	uint32_t max_bandwidth;
+	/*
+	 * Pixel Clock will be restricted to some value as
+	 * defined in the device datasheet measured in KHz
+	 */
+	uint32_t max_pixelclock;
+	/*
+	 * Max allowable width is limited on a per device basis
+	 * measured in pixels
+	 */
+	uint32_t max_width;
 
 	/* register contents saved across suspend/resume: */
 	u32 saved_register[12];
@@ -89,12 +111,13 @@ struct tilcdc_module {
 	const char *name;
 	struct list_head list;
 	const struct tilcdc_module_ops *funcs;
+	unsigned int preferred_bpp;
 };
 
 void tilcdc_module_init(struct tilcdc_module *mod, const char *name,
 		const struct tilcdc_module_ops *funcs);
 void tilcdc_module_cleanup(struct tilcdc_module *mod);
-
+void tilcdc_slave_probedefer(bool defered);
 
 /* Panel config that needs to be set in the crtc, but is not coming from
  * the mode timings.  The display module is expected to call
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
index 0917665..86c6732 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_panel.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
@@ -393,6 +393,8 @@ static int panel_probe(struct platform_device *pdev)
 		goto fail;
 	}
 
+	mod->preferred_bpp = panel_mod->info->bpp;
+
 	panel_mod->backlight = of_find_backlight_by_node(node);
 	if (panel_mod->backlight)
 		dev_info(&pdev->dev, "found backlight\n");
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_regs.h b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
index 17fd1b4..1bf5e25 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_regs.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
@@ -80,6 +80,7 @@
 #define LCDC_INVERT_PIXEL_CLOCK                  BIT(22)
 #define LCDC_INVERT_HSYNC                        BIT(21)
 #define LCDC_INVERT_VSYNC                        BIT(20)
+#define LCDC_LPP_B10                             BIT(26)
 
 /* LCDC Block */
 #define LCDC_PID_REG                             0x0
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_slave.c b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
index db1d2fc..dfffaf0 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_slave.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
@@ -298,6 +298,7 @@ static int slave_probe(struct platform_device *pdev)
 	struct tilcdc_module *mod;
 	struct pinctrl *pinctrl;
 	uint32_t i2c_phandle;
+	struct i2c_adapter *slavei2c;
 	int ret = -EINVAL;
 
 	/* bail out early if no DT data: */
@@ -306,42 +307,48 @@ static int slave_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
-	slave_mod = kzalloc(sizeof(*slave_mod), GFP_KERNEL);
-	if (!slave_mod)
-		return -ENOMEM;
-
-	mod = &slave_mod->base;
-
-	tilcdc_module_init(mod, "slave", &slave_module_ops);
-
-	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
-	if (IS_ERR(pinctrl))
-		dev_warn(&pdev->dev, "pins are not configured\n");
-
+	/* Bail out early if i2c not specified */
 	if (of_property_read_u32(node, "i2c", &i2c_phandle)) {
 		dev_err(&pdev->dev, "could not get i2c bus phandle\n");
-		goto fail;
+		return ret;
 	}
 
 	i2c_node = of_find_node_by_phandle(i2c_phandle);
 	if (!i2c_node) {
 		dev_err(&pdev->dev, "could not get i2c bus node\n");
-		goto fail;
+		return ret;
 	}
 
-	slave_mod->i2c = of_find_i2c_adapter_by_node(i2c_node);
-	if (!slave_mod->i2c) {
+	/* but defer the probe if it can't be initialized it might come later */
+	slavei2c = of_find_i2c_adapter_by_node(i2c_node);
+	of_node_put(i2c_node);
+
+	if (!slavei2c) {
+		ret = -EPROBE_DEFER;
+		tilcdc_slave_probedefer(true);
 		dev_err(&pdev->dev, "could not get i2c\n");
-		goto fail;
+		return ret;
 	}
 
-	of_node_put(i2c_node);
+	slave_mod = kzalloc(sizeof(*slave_mod), GFP_KERNEL);
+	if (!slave_mod)
+		return -ENOMEM;
 
-	return 0;
+	mod = &slave_mod->base;
 
-fail:
-	slave_destroy(mod);
-	return ret;
+	mod->preferred_bpp = slave_info.bpp;
+
+	slave_mod->i2c = slavei2c;
+
+	tilcdc_module_init(mod, "slave", &slave_module_ops);
+
+	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
+	if (IS_ERR(pinctrl))
+		dev_warn(&pdev->dev, "pins are not configured\n");
+
+	tilcdc_slave_probedefer(false);
+
+	return 0;
 }
 
 static int slave_remove(struct platform_device *pdev)
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
index a36788f..925c7cd 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
@@ -354,6 +354,8 @@ static int tfp410_probe(struct platform_device *pdev)
 		goto fail;
 	}
 
+	mod->preferred_bpp = dvi_info.bpp;
+
 	i2c_node = of_find_node_by_phandle(i2c_phandle);
 	if (!i2c_node) {
 		dev_err(&pdev->dev, "could not get i2c bus node\n");
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 9b07b7d..cb9dd67 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -150,6 +150,9 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	if (bo->ttm)
 		ttm_tt_destroy(bo->ttm);
 	atomic_dec(&bo->glob->bo_count);
+	if (bo->resv == &bo->ttm_resv)
+		reservation_object_fini(&bo->ttm_resv);
+
 	if (bo->destroy)
 		bo->destroy(bo);
 	else {
@@ -158,24 +161,12 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	ttm_mem_global_free(bdev->glob->mem_glob, acc_size);
 }
 
-static int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo,
-				  bool interruptible)
-{
-	if (interruptible) {
-		return wait_event_interruptible(bo->event_queue,
-					       !ttm_bo_is_reserved(bo));
-	} else {
-		wait_event(bo->event_queue, !ttm_bo_is_reserved(bo));
-		return 0;
-	}
-}
-
 void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_mem_type_manager *man;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 
 	if (!(bo->mem.placement & TTM_PL_FLAG_NO_EVICT)) {
 
@@ -191,6 +182,7 @@ void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 		}
 	}
 }
+EXPORT_SYMBOL(ttm_bo_add_to_lru);
 
 int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
 {
@@ -213,71 +205,6 @@ int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
 	return put_count;
 }
 
-int ttm_bo_reserve_nolru(struct ttm_buffer_object *bo,
-			  bool interruptible,
-			  bool no_wait, bool use_sequence, uint32_t sequence)
-{
-	int ret;
-
-	while (unlikely(atomic_xchg(&bo->reserved, 1) != 0)) {
-		/**
-		 * Deadlock avoidance for multi-bo reserving.
-		 */
-		if (use_sequence && bo->seq_valid) {
-			/**
-			 * We've already reserved this one.
-			 */
-			if (unlikely(sequence == bo->val_seq))
-				return -EDEADLK;
-			/**
-			 * Already reserved by a thread that will not back
-			 * off for us. We need to back off.
-			 */
-			if (unlikely(sequence - bo->val_seq < (1 << 31)))
-				return -EAGAIN;
-		}
-
-		if (no_wait)
-			return -EBUSY;
-
-		ret = ttm_bo_wait_unreserved(bo, interruptible);
-
-		if (unlikely(ret))
-			return ret;
-	}
-
-	if (use_sequence) {
-		bool wake_up = false;
-		/**
-		 * Wake up waiters that may need to recheck for deadlock,
-		 * if we decreased the sequence number.
-		 */
-		if (unlikely((bo->val_seq - sequence < (1 << 31))
-			     || !bo->seq_valid))
-			wake_up = true;
-
-		/*
-		 * In the worst case with memory ordering these values can be
-		 * seen in the wrong order. However since we call wake_up_all
-		 * in that case, this will hopefully not pose a problem,
-		 * and the worst case would only cause someone to accidentally
-		 * hit -EAGAIN in ttm_bo_reserve when they see old value of
-		 * val_seq. However this would only happen if seq_valid was
-		 * written before val_seq was, and just means some slightly
-		 * increased cpu usage
-		 */
-		bo->val_seq = sequence;
-		bo->seq_valid = true;
-		if (wake_up)
-			wake_up_all(&bo->event_queue);
-	} else {
-		bo->seq_valid = false;
-	}
-
-	return 0;
-}
-EXPORT_SYMBOL(ttm_bo_reserve);
-
 static void ttm_bo_ref_bug(struct kref *list_kref)
 {
 	BUG();
@@ -290,89 +217,16 @@ void ttm_bo_list_ref_sub(struct ttm_buffer_object *bo, int count,
 		 (never_free) ? ttm_bo_ref_bug : ttm_bo_release_list);
 }
 
-int ttm_bo_reserve(struct ttm_buffer_object *bo,
-		   bool interruptible,
-		   bool no_wait, bool use_sequence, uint32_t sequence)
-{
-	struct ttm_bo_global *glob = bo->glob;
-	int put_count = 0;
-	int ret;
-
-	ret = ttm_bo_reserve_nolru(bo, interruptible, no_wait, use_sequence,
-				   sequence);
-	if (likely(ret == 0)) {
-		spin_lock(&glob->lru_lock);
-		put_count = ttm_bo_del_from_lru(bo);
-		spin_unlock(&glob->lru_lock);
-		ttm_bo_list_ref_sub(bo, put_count, true);
-	}
-
-	return ret;
-}
-
-int ttm_bo_reserve_slowpath_nolru(struct ttm_buffer_object *bo,
-				  bool interruptible, uint32_t sequence)
-{
-	bool wake_up = false;
-	int ret;
-
-	while (unlikely(atomic_xchg(&bo->reserved, 1) != 0)) {
-		WARN_ON(bo->seq_valid && sequence == bo->val_seq);
-
-		ret = ttm_bo_wait_unreserved(bo, interruptible);
-
-		if (unlikely(ret))
-			return ret;
-	}
-
-	if ((bo->val_seq - sequence < (1 << 31)) || !bo->seq_valid)
-		wake_up = true;
-
-	/**
-	 * Wake up waiters that may need to recheck for deadlock,
-	 * if we decreased the sequence number.
-	 */
-	bo->val_seq = sequence;
-	bo->seq_valid = true;
-	if (wake_up)
-		wake_up_all(&bo->event_queue);
-
-	return 0;
-}
-
-int ttm_bo_reserve_slowpath(struct ttm_buffer_object *bo,
-			    bool interruptible, uint32_t sequence)
-{
-	struct ttm_bo_global *glob = bo->glob;
-	int put_count, ret;
-
-	ret = ttm_bo_reserve_slowpath_nolru(bo, interruptible, sequence);
-	if (likely(!ret)) {
-		spin_lock(&glob->lru_lock);
-		put_count = ttm_bo_del_from_lru(bo);
-		spin_unlock(&glob->lru_lock);
-		ttm_bo_list_ref_sub(bo, put_count, true);
-	}
-	return ret;
-}
-EXPORT_SYMBOL(ttm_bo_reserve_slowpath);
-
-void ttm_bo_unreserve_locked(struct ttm_buffer_object *bo)
-{
-	ttm_bo_add_to_lru(bo);
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
-}
-
-void ttm_bo_unreserve(struct ttm_buffer_object *bo)
+void ttm_bo_del_sub_from_lru(struct ttm_buffer_object *bo)
 {
-	struct ttm_bo_global *glob = bo->glob;
+	int put_count;
 
-	spin_lock(&glob->lru_lock);
-	ttm_bo_unreserve_locked(bo);
-	spin_unlock(&glob->lru_lock);
+	spin_lock(&bo->glob->lru_lock);
+	put_count = ttm_bo_del_from_lru(bo);
+	spin_unlock(&bo->glob->lru_lock);
+	ttm_bo_list_ref_sub(bo, put_count, true);
 }
-EXPORT_SYMBOL(ttm_bo_unreserve);
+EXPORT_SYMBOL(ttm_bo_del_sub_from_lru);
 
 /*
  * Call bo->mutex locked.
@@ -544,17 +398,7 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 	}
 	ttm_bo_mem_put(bo, &bo->mem);
 
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
-
-	/*
-	 * Since the final reference to this bo may not be dropped by
-	 * the current task we have to put a memory barrier here to make
-	 * sure the changes done in this function are always visible.
-	 *
-	 * This function only needs protection against the final kref_put.
-	 */
-	smp_mb__before_atomic_dec();
+	ww_mutex_unlock (&bo->resv->lock);
 }
 
 static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
@@ -586,10 +430,8 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 		sync_obj = driver->sync_obj_ref(bo->sync_obj);
 	spin_unlock(&bdev->fence_lock);
 
-	if (!ret) {
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
-	}
+	if (!ret)
+		ww_mutex_unlock(&bo->resv->lock);
 
 	kref_get(&bo->list_kref);
 	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
@@ -639,8 +481,7 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 		sync_obj = driver->sync_obj_ref(bo->sync_obj);
 		spin_unlock(&bdev->fence_lock);
 
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
+		ww_mutex_unlock(&bo->resv->lock);
 		spin_unlock(&glob->lru_lock);
 
 		ret = driver->sync_obj_wait(sync_obj, false, interruptible);
@@ -678,8 +519,7 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 		spin_unlock(&bdev->fence_lock);
 
 	if (ret || unlikely(list_empty(&bo->ddestroy))) {
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
+		ww_mutex_unlock(&bo->resv->lock);
 		spin_unlock(&glob->lru_lock);
 		return ret;
 	}
@@ -831,7 +671,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 		goto out;
 	}
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 
 	evict_mem = bo->mem;
 	evict_mem.mm_node = NULL;
@@ -1121,7 +961,7 @@ int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	struct ttm_mem_reg mem;
 	struct ttm_bo_device *bdev = bo->bdev;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 
 	/*
 	 * FIXME: It's possible to pipeline buffer moves.
@@ -1180,7 +1020,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 {
 	int ret;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 	/* Check that range is valid */
 	if (placement->lpfn || placement->fpfn)
 		if (placement->fpfn > placement->lpfn ||
@@ -1239,6 +1079,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	int ret = 0;
 	unsigned long num_pages;
 	struct ttm_mem_global *mem_glob = bdev->glob->mem_glob;
+	bool locked;
 
 	ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false);
 	if (ret) {
@@ -1265,8 +1106,6 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	kref_init(&bo->kref);
 	kref_init(&bo->list_kref);
 	atomic_set(&bo->cpu_writers, 0);
-	atomic_set(&bo->reserved, 1);
-	init_waitqueue_head(&bo->event_queue);
 	INIT_LIST_HEAD(&bo->lru);
 	INIT_LIST_HEAD(&bo->ddestroy);
 	INIT_LIST_HEAD(&bo->swap);
@@ -1284,37 +1123,34 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	bo->mem.bus.io_reserved_count = 0;
 	bo->priv_flags = 0;
 	bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
-	bo->seq_valid = false;
 	bo->persistent_swap_storage = persistent_swap_storage;
 	bo->acc_size = acc_size;
 	bo->sg = sg;
+	bo->resv = &bo->ttm_resv;
+	reservation_object_init(bo->resv);
 	atomic_inc(&bo->glob->bo_count);
 
 	ret = ttm_bo_check_placement(bo, placement);
-	if (unlikely(ret != 0))
-		goto out_err;
 
 	/*
 	 * For ttm_bo_type_device buffers, allocate
 	 * address space from the device.
 	 */
-	if (bo->type == ttm_bo_type_device ||
-	    bo->type == ttm_bo_type_sg) {
+	if (likely(!ret) &&
+	    (bo->type == ttm_bo_type_device ||
+	     bo->type == ttm_bo_type_sg))
 		ret = ttm_bo_setup_vm(bo);
-		if (ret)
-			goto out_err;
-	}
 
-	ret = ttm_bo_validate(bo, placement, interruptible, false);
-	if (ret)
-		goto out_err;
+	locked = ww_mutex_trylock(&bo->resv->lock);
+	WARN_ON(!locked);
 
-	ttm_bo_unreserve(bo);
-	return 0;
+	if (likely(!ret))
+		ret = ttm_bo_validate(bo, placement, interruptible, false);
 
-out_err:
 	ttm_bo_unreserve(bo);
-	ttm_bo_unref(&bo);
+
+	if (unlikely(ret))
+		ttm_bo_unref(&bo);
 
 	return ret;
 }
@@ -1619,9 +1455,7 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 		goto out_no_sys;
 
 	bdev->addr_space_rb = RB_ROOT;
-	ret = drm_mm_init(&bdev->addr_space_mm, file_page_offset, 0x10000000);
-	if (unlikely(ret != 0))
-		goto out_no_addr_mm;
+	drm_mm_init(&bdev->addr_space_mm, file_page_offset, 0x10000000);
 
 	INIT_DELAYED_WORK(&bdev->wq, ttm_bo_delayed_workqueue);
 	INIT_LIST_HEAD(&bdev->ddestroy);
@@ -1635,8 +1469,6 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 	mutex_unlock(&glob->device_list_mutex);
 
 	return 0;
-out_no_addr_mm:
-	ttm_bo_clean_mm(bdev, 0);
 out_no_sys:
 	return ret;
 }
@@ -1927,8 +1759,7 @@ out:
 	 * already swapped buffer.
 	 */
 
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
+	ww_mutex_unlock(&bo->resv->lock);
 	kref_put(&bo->list_kref, ttm_bo_release_list);
 	return ret;
 }
diff --git a/drivers/gpu/drm/ttm/ttm_bo_manager.c b/drivers/gpu/drm/ttm/ttm_bo_manager.c
index 9212494..e4367f9 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_manager.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_manager.c
@@ -103,18 +103,12 @@ static int ttm_bo_man_init(struct ttm_mem_type_manager *man,
 			   unsigned long p_size)
 {
 	struct ttm_range_manager *rman;
-	int ret;
 
 	rman = kzalloc(sizeof(*rman), GFP_KERNEL);
 	if (!rman)
 		return -ENOMEM;
 
-	ret = drm_mm_init(&rman->mm, 0, p_size);
-	if (ret) {
-		kfree(rman);
-		return ret;
-	}
-
+	drm_mm_init(&rman->mm, 0, p_size);
 	spin_lock_init(&rman->lock);
 	man->priv = rman;
 	return 0;
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index af89458..319cf41 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -433,6 +433,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	struct ttm_buffer_object *fbo;
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_bo_driver *driver = bdev->driver;
+	int ret;
 
 	fbo = kmalloc(sizeof(*fbo), GFP_KERNEL);
 	if (!fbo)
@@ -445,7 +446,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	 * TODO: Explicit member copy would probably be better here.
 	 */
 
-	init_waitqueue_head(&fbo->event_queue);
 	INIT_LIST_HEAD(&fbo->ddestroy);
 	INIT_LIST_HEAD(&fbo->lru);
 	INIT_LIST_HEAD(&fbo->swap);
@@ -463,6 +463,10 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	kref_init(&fbo->kref);
 	fbo->destroy = &ttm_transfered_destroy;
 	fbo->acc_size = 0;
+	fbo->resv = &fbo->ttm_resv;
+	reservation_object_init(fbo->resv);
+	ret = ww_mutex_trylock(&fbo->resv->lock);
+	WARN_ON(!ret);
 
 	*new_obj = fbo;
 	return 0;
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index 7b90def..6c91178 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -32,7 +32,8 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 
-static void ttm_eu_backoff_reservation_locked(struct list_head *list)
+static void ttm_eu_backoff_reservation_locked(struct list_head *list,
+					      struct ww_acquire_ctx *ticket)
 {
 	struct ttm_validate_buffer *entry;
 
@@ -41,14 +42,12 @@ static void ttm_eu_backoff_reservation_locked(struct list_head *list)
 		if (!entry->reserved)
 			continue;
 
+		entry->reserved = false;
 		if (entry->removed) {
 			ttm_bo_add_to_lru(bo);
 			entry->removed = false;
-
 		}
-		entry->reserved = false;
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
+		ww_mutex_unlock(&bo->resv->lock);
 	}
 }
 
@@ -82,7 +81,8 @@ static void ttm_eu_list_ref_sub(struct list_head *list)
 	}
 }
 
-void ttm_eu_backoff_reservation(struct list_head *list)
+void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
+				struct list_head *list)
 {
 	struct ttm_validate_buffer *entry;
 	struct ttm_bo_global *glob;
@@ -93,7 +93,8 @@ void ttm_eu_backoff_reservation(struct list_head *list)
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 	spin_lock(&glob->lru_lock);
-	ttm_eu_backoff_reservation_locked(list);
+	ttm_eu_backoff_reservation_locked(list, ticket);
+	ww_acquire_fini(ticket);
 	spin_unlock(&glob->lru_lock);
 }
 EXPORT_SYMBOL(ttm_eu_backoff_reservation);
@@ -110,12 +111,12 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation);
  * buffers in different orders.
  */
 
-int ttm_eu_reserve_buffers(struct list_head *list)
+int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
+			   struct list_head *list)
 {
 	struct ttm_bo_global *glob;
 	struct ttm_validate_buffer *entry;
 	int ret;
-	uint32_t val_seq;
 
 	if (list_empty(list))
 		return 0;
@@ -129,9 +130,7 @@ int ttm_eu_reserve_buffers(struct list_head *list)
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 
-	spin_lock(&glob->lru_lock);
-	val_seq = entry->bo->bdev->val_seq++;
-
+	ww_acquire_init(ticket, &reservation_ww_class);
 retry:
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
@@ -140,49 +139,34 @@ retry:
 		if (entry->reserved)
 			continue;
 
-		ret = ttm_bo_reserve_nolru(bo, true, true, true, val_seq);
-		switch (ret) {
-		case 0:
-			break;
-		case -EBUSY:
-			ttm_eu_del_from_lru_locked(list);
-			spin_unlock(&glob->lru_lock);
-			ret = ttm_bo_reserve_nolru(bo, true, false,
-						   true, val_seq);
-			spin_lock(&glob->lru_lock);
-			if (!ret)
-				break;
-
-			if (unlikely(ret != -EAGAIN))
-				goto err;
 
-			/* fallthrough */
-		case -EAGAIN:
-			ttm_eu_backoff_reservation_locked(list);
+		ret = ttm_bo_reserve_nolru(bo, true, false, true, ticket);
 
-			/*
-			 * temporarily increase sequence number every retry,
-			 * to prevent us from seeing our old reservation
-			 * sequence when someone else reserved the buffer,
-			 * but hasn't updated the seq_valid/seqno members yet.
+		if (ret == -EDEADLK) {
+			/* uh oh, we lost out, drop every reservation and try
+			 * to only reserve this buffer, then start over if
+			 * this succeeds.
 			 */
-			val_seq = entry->bo->bdev->val_seq++;
-
+			spin_lock(&glob->lru_lock);
+			ttm_eu_backoff_reservation_locked(list, ticket);
 			spin_unlock(&glob->lru_lock);
 			ttm_eu_list_ref_sub(list);
-			ret = ttm_bo_reserve_slowpath_nolru(bo, true, val_seq);
-			if (unlikely(ret != 0))
-				return ret;
-			spin_lock(&glob->lru_lock);
+			ret = ww_mutex_lock_slow_interruptible(&bo->resv->lock,
+							       ticket);
+			if (unlikely(ret != 0)) {
+				if (ret == -EINTR)
+					ret = -ERESTARTSYS;
+				goto err_fini;
+			}
+
 			entry->reserved = true;
 			if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
 				ret = -EBUSY;
 				goto err;
 			}
 			goto retry;
-		default:
+		} else if (ret)
 			goto err;
-		}
 
 		entry->reserved = true;
 		if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
@@ -191,21 +175,27 @@ retry:
 		}
 	}
 
+	ww_acquire_done(ticket);
+	spin_lock(&glob->lru_lock);
 	ttm_eu_del_from_lru_locked(list);
 	spin_unlock(&glob->lru_lock);
 	ttm_eu_list_ref_sub(list);
-
 	return 0;
 
 err:
-	ttm_eu_backoff_reservation_locked(list);
+	spin_lock(&glob->lru_lock);
+	ttm_eu_backoff_reservation_locked(list, ticket);
 	spin_unlock(&glob->lru_lock);
 	ttm_eu_list_ref_sub(list);
+err_fini:
+	ww_acquire_done(ticket);
+	ww_acquire_fini(ticket);
 	return ret;
 }
 EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 
-void ttm_eu_fence_buffer_objects(struct list_head *list, void *sync_obj)
+void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
+				 struct list_head *list, void *sync_obj)
 {
 	struct ttm_validate_buffer *entry;
 	struct ttm_buffer_object *bo;
@@ -228,11 +218,13 @@ void ttm_eu_fence_buffer_objects(struct list_head *list, void *sync_obj)
 		bo = entry->bo;
 		entry->old_sync_obj = bo->sync_obj;
 		bo->sync_obj = driver->sync_obj_ref(sync_obj);
-		ttm_bo_unreserve_locked(bo);
+		ttm_bo_add_to_lru(bo);
+		ww_mutex_unlock(&bo->resv->lock);
 		entry->reserved = false;
 	}
 	spin_unlock(&bdev->fence_lock);
 	spin_unlock(&glob->lru_lock);
+	ww_acquire_fini(ticket);
 
 	list_for_each_entry(entry, list, head) {
 		if (entry->old_sync_obj)
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index dc0c065..97e9d61 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -393,19 +393,6 @@ static struct fb_ops udlfb_ops = {
 	.fb_release = udl_fb_release,
 };
 
-static void udl_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
-			   u16 blue, int regno)
-{
-}
-
-static void udl_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
-			     u16 *blue, int regno)
-{
-	*red = 0;
-	*green = 0;
-	*blue = 0;
-}
-
 static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb,
 				      struct drm_file *file,
 				      unsigned flags, unsigned color,
@@ -558,8 +545,6 @@ out:
 }
 
 static struct drm_fb_helper_funcs udl_fb_helper_funcs = {
-	.gamma_set = udl_crtc_fb_gamma_set,
-	.gamma_get = udl_crtc_fb_gamma_get,
 	.fb_probe = udlfb_create,
 };
 
diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c
index e96d234..2ae1eb7 100644
--- a/drivers/gpu/drm/udl/udl_modeset.c
+++ b/drivers/gpu/drm/udl/udl_modeset.c
@@ -363,10 +363,6 @@ static void udl_crtc_destroy(struct drm_crtc *crtc)
 	kfree(crtc);
 }
 
-static void udl_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static void udl_crtc_prepare(struct drm_crtc *crtc)
 {
 }
@@ -383,7 +379,6 @@ static struct drm_crtc_helper_funcs udl_helper_funcs = {
 	.prepare = udl_crtc_prepare,
 	.commit = udl_crtc_commit,
 	.disable = udl_crtc_disable,
-	.load_lut = udl_load_lut,
 };
 
 static const struct drm_crtc_funcs udl_crtc_funcs = {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
index 5fae06a..d4e54fc 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
@@ -302,7 +302,7 @@ void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin)
 	uint32_t old_mem_type = bo->mem.mem_type;
 	int ret;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 	BUG_ON(old_mem_type != TTM_PL_VRAM &&
 	       old_mem_type != VMW_PL_GMR);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 07dfd82..78e2164 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -565,8 +565,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		dev_priv->has_gmr = false;
 	}
 
-	dev_priv->mmio_mtrr = drm_mtrr_add(dev_priv->mmio_start,
-					   dev_priv->mmio_size, DRM_MTRR_WC);
+	dev_priv->mmio_mtrr = arch_phys_wc_add(dev_priv->mmio_start,
+					       dev_priv->mmio_size);
 
 	dev_priv->mmio_virt = ioremap_wc(dev_priv->mmio_start,
 					 dev_priv->mmio_size);
@@ -664,8 +664,7 @@ out_no_device:
 out_err4:
 	iounmap(dev_priv->mmio_virt);
 out_err3:
-	drm_mtrr_del(dev_priv->mmio_mtrr, dev_priv->mmio_start,
-		     dev_priv->mmio_size, DRM_MTRR_WC);
+	arch_phys_wc_del(dev_priv->mmio_mtrr);
 	if (dev_priv->has_gmr)
 		(void) ttm_bo_clean_mm(&dev_priv->bdev, VMW_PL_GMR);
 	(void)ttm_bo_clean_mm(&dev_priv->bdev, TTM_PL_VRAM);
@@ -709,8 +708,7 @@ static int vmw_driver_unload(struct drm_device *dev)
 
 	ttm_object_device_release(&dev_priv->tdev);
 	iounmap(dev_priv->mmio_virt);
-	drm_mtrr_del(dev_priv->mmio_mtrr, dev_priv->mmio_start,
-		     dev_priv->mmio_size, DRM_MTRR_WC);
+	arch_phys_wc_del(dev_priv->mmio_mtrr);
 	if (dev_priv->has_gmr)
 		(void)ttm_bo_clean_mm(&dev_priv->bdev, VMW_PL_GMR);
 	(void)ttm_bo_clean_mm(&dev_priv->bdev, TTM_PL_VRAM);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 394e647..599f646 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -1432,6 +1432,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	struct vmw_fence_obj *fence = NULL;
 	struct vmw_resource *error_resource;
 	struct list_head resource_list;
+	struct ww_acquire_ctx ticket;
 	uint32_t handle;
 	void *cmd;
 	int ret;
@@ -1488,7 +1489,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	if (unlikely(ret != 0))
 		goto out_err;
 
-	ret = ttm_eu_reserve_buffers(&sw_context->validate_nodes);
+	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes);
 	if (unlikely(ret != 0))
 		goto out_err;
 
@@ -1537,7 +1538,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 		DRM_ERROR("Fence submission error. Syncing.\n");
 
 	vmw_resource_list_unreserve(&sw_context->resource_list, false);
-	ttm_eu_fence_buffer_objects(&sw_context->validate_nodes,
+	ttm_eu_fence_buffer_objects(&ticket, &sw_context->validate_nodes,
 				    (void *) fence);
 
 	if (unlikely(dev_priv->pinned_bo != NULL &&
@@ -1570,7 +1571,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 out_err:
 	vmw_resource_relocations_free(&sw_context->res_relocations);
 	vmw_free_relocations(sw_context);
-	ttm_eu_backoff_reservation(&sw_context->validate_nodes);
+	ttm_eu_backoff_reservation(&ticket, &sw_context->validate_nodes);
 	vmw_resource_list_unreserve(&sw_context->resource_list, true);
 	vmw_clear_validations(sw_context);
 	if (unlikely(dev_priv->pinned_bo != NULL &&
@@ -1644,6 +1645,7 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	struct list_head validate_list;
 	struct ttm_validate_buffer pinned_val, query_val;
 	struct vmw_fence_obj *lfence = NULL;
+	struct ww_acquire_ctx ticket;
 
 	if (dev_priv->pinned_bo == NULL)
 		goto out_unlock;
@@ -1657,7 +1659,7 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	list_add_tail(&query_val.head, &validate_list);
 
 	do {
-		ret = ttm_eu_reserve_buffers(&validate_list);
+		ret = ttm_eu_reserve_buffers(&ticket, &validate_list);
 	} while (ret == -ERESTARTSYS);
 
 	if (unlikely(ret != 0)) {
@@ -1684,7 +1686,7 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 						  NULL);
 		fence = lfence;
 	}
-	ttm_eu_fence_buffer_objects(&validate_list, (void *) fence);
+	ttm_eu_fence_buffer_objects(&ticket, &validate_list, (void *) fence);
 	if (lfence != NULL)
 		vmw_fence_obj_unreference(&lfence);
 
@@ -1696,7 +1698,7 @@ out_unlock:
 	return;
 
 out_no_emit:
-	ttm_eu_backoff_reservation(&validate_list);
+	ttm_eu_backoff_reservation(&ticket, &validate_list);
 out_no_reserve:
 	ttm_bo_unref(&query_val.bo);
 	ttm_bo_unref(&pinned_val.bo);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 3e3c7ab..d4607b2 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -174,7 +174,6 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 			   uint32_t handle, uint32_t width, uint32_t height)
 {
 	struct vmw_private *dev_priv = vmw_priv(crtc->dev);
-	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
 	struct vmw_display_unit *du = vmw_crtc_to_du(crtc);
 	struct vmw_surface *surface = NULL;
 	struct vmw_dma_buffer *dmabuf = NULL;
@@ -197,6 +196,8 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	}
 
 	if (handle) {
+		struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+
 		ret = vmw_user_lookup_handle(dev_priv, tfile,
 					     handle, &surface, &dmabuf);
 		if (ret) {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index bc78425..7953d1f 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -958,13 +958,13 @@ void vmw_resource_unreserve(struct vmw_resource *res,
 	if (new_backup && new_backup != res->backup) {
 
 		if (res->backup) {
-			BUG_ON(!ttm_bo_is_reserved(&res->backup->base));
+			lockdep_assert_held(&res->backup->base.resv->lock.base);
 			list_del_init(&res->mob_head);
 			vmw_dmabuf_unreference(&res->backup);
 		}
 
 		res->backup = vmw_dmabuf_reference(new_backup);
-		BUG_ON(!ttm_bo_is_reserved(&new_backup->base));
+		lockdep_assert_held(&new_backup->base.resv->lock.base);
 		list_add_tail(&res->mob_head, &new_backup->res_list);
 	}
 	if (new_backup)
@@ -990,9 +990,11 @@ void vmw_resource_unreserve(struct vmw_resource *res,
  * @val_buf:        On successful return contains data about the
  *                  reserved and validated backup buffer.
  */
-int vmw_resource_check_buffer(struct vmw_resource *res,
-			      bool interruptible,
-			      struct ttm_validate_buffer *val_buf)
+static int
+vmw_resource_check_buffer(struct vmw_resource *res,
+			  struct ww_acquire_ctx *ticket,
+			  bool interruptible,
+			  struct ttm_validate_buffer *val_buf)
 {
 	struct list_head val_list;
 	bool backup_dirty = false;
@@ -1007,7 +1009,7 @@ int vmw_resource_check_buffer(struct vmw_resource *res,
 	INIT_LIST_HEAD(&val_list);
 	val_buf->bo = ttm_bo_reference(&res->backup->base);
 	list_add_tail(&val_buf->head, &val_list);
-	ret = ttm_eu_reserve_buffers(&val_list);
+	ret = ttm_eu_reserve_buffers(ticket, &val_list);
 	if (unlikely(ret != 0))
 		goto out_no_reserve;
 
@@ -1025,7 +1027,7 @@ int vmw_resource_check_buffer(struct vmw_resource *res,
 	return 0;
 
 out_no_validate:
-	ttm_eu_backoff_reservation(&val_list);
+	ttm_eu_backoff_reservation(ticket, &val_list);
 out_no_reserve:
 	ttm_bo_unref(&val_buf->bo);
 	if (backup_dirty)
@@ -1069,7 +1071,9 @@ int vmw_resource_reserve(struct vmw_resource *res, bool no_backup)
  *.
  * @val_buf:        Backup buffer information.
  */
-void vmw_resource_backoff_reservation(struct ttm_validate_buffer *val_buf)
+static void
+vmw_resource_backoff_reservation(struct ww_acquire_ctx *ticket,
+				 struct ttm_validate_buffer *val_buf)
 {
 	struct list_head val_list;
 
@@ -1078,7 +1082,7 @@ void vmw_resource_backoff_reservation(struct ttm_validate_buffer *val_buf)
 
 	INIT_LIST_HEAD(&val_list);
 	list_add_tail(&val_buf->head, &val_list);
-	ttm_eu_backoff_reservation(&val_list);
+	ttm_eu_backoff_reservation(ticket, &val_list);
 	ttm_bo_unref(&val_buf->bo);
 }
 
@@ -1092,12 +1096,13 @@ int vmw_resource_do_evict(struct vmw_resource *res)
 {
 	struct ttm_validate_buffer val_buf;
 	const struct vmw_res_func *func = res->func;
+	struct ww_acquire_ctx ticket;
 	int ret;
 
 	BUG_ON(!func->may_evict);
 
 	val_buf.bo = NULL;
-	ret = vmw_resource_check_buffer(res, true, &val_buf);
+	ret = vmw_resource_check_buffer(res, &ticket, true, &val_buf);
 	if (unlikely(ret != 0))
 		return ret;
 
@@ -1112,7 +1117,7 @@ int vmw_resource_do_evict(struct vmw_resource *res)
 	res->backup_dirty = true;
 	res->res_dirty = false;
 out_no_unbind:
-	vmw_resource_backoff_reservation(&val_buf);
+	vmw_resource_backoff_reservation(&ticket, &val_buf);
 
 	return ret;
 }
diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h
index a1607d6..790ddf1 100644
--- a/drivers/gpu/host1x/dev.h
+++ b/drivers/gpu/host1x/dev.h
@@ -73,7 +73,7 @@ struct host1x_syncpt_ops {
 	void (*restore_wait_base)(struct host1x_syncpt *syncpt);
 	void (*load_wait_base)(struct host1x_syncpt *syncpt);
 	u32 (*load)(struct host1x_syncpt *syncpt);
-	void (*cpu_incr)(struct host1x_syncpt *syncpt);
+	int (*cpu_incr)(struct host1x_syncpt *syncpt);
 	int (*patch_wait)(struct host1x_syncpt *syncpt, void *patch_addr);
 };
 
@@ -157,10 +157,10 @@ static inline u32 host1x_hw_syncpt_load(struct host1x *host,
 	return host->syncpt_op->load(sp);
 }
 
-static inline void host1x_hw_syncpt_cpu_incr(struct host1x *host,
-					     struct host1x_syncpt *sp)
+static inline int host1x_hw_syncpt_cpu_incr(struct host1x *host,
+					    struct host1x_syncpt *sp)
 {
-	host->syncpt_op->cpu_incr(sp);
+	return host->syncpt_op->cpu_incr(sp);
 }
 
 static inline int host1x_hw_syncpt_patch_wait(struct host1x *host,
diff --git a/drivers/gpu/host1x/drm/dc.c b/drivers/gpu/host1x/drm/dc.c
index 8c04943..5360e5a 100644
--- a/drivers/gpu/host1x/drm/dc.c
+++ b/drivers/gpu/host1x/drm/dc.c
@@ -79,6 +79,9 @@ static int tegra_plane_disable(struct drm_plane *plane)
 	struct tegra_plane *p = to_tegra_plane(plane);
 	unsigned long value;
 
+	if (!plane->crtc)
+		return 0;
+
 	value = WINDOW_A_SELECT << p->index;
 	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
 
@@ -140,6 +143,7 @@ static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
 static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 			     struct drm_framebuffer *fb)
 {
+	unsigned int format = tegra_dc_format(fb->pixel_format);
 	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
 	unsigned long value;
 
@@ -150,6 +154,7 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 
 	tegra_dc_writel(dc, bo->paddr + value, DC_WINBUF_START_ADDR);
 	tegra_dc_writel(dc, fb->pitches[0], DC_WIN_LINE_STRIDE);
+	tegra_dc_writel(dc, format, DC_WIN_COLOR_DEPTH);
 
 	value = GENERAL_UPDATE | WIN_A_UPDATE;
 	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
diff --git a/drivers/gpu/host1x/drm/drm.c b/drivers/gpu/host1x/drm/drm.c
index 2b561c9..e184b00 100644
--- a/drivers/gpu/host1x/drm/drm.c
+++ b/drivers/gpu/host1x/drm/drm.c
@@ -148,6 +148,7 @@ int host1x_drm_init(struct host1x_drm *host1x, struct drm_device *drm)
 				dev_err(host1x->dev,
 					"DRM setup failed for %s: %d\n",
 					dev_name(client->dev), err);
+				mutex_unlock(&host1x->clients_lock);
 				return err;
 			}
 		}
@@ -175,6 +176,7 @@ int host1x_drm_exit(struct host1x_drm *host1x)
 				dev_err(host1x->dev,
 					"DRM cleanup failed for %s: %d\n",
 					dev_name(client->dev), err);
+				mutex_unlock(&host1x->clients_lock);
 				return err;
 			}
 		}
@@ -257,6 +259,13 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 	if (err < 0)
 		return err;
 
+	/*
+	 * We don't use the drm_irq_install() helpers provided by the DRM
+	 * core, so we need to set this manually in order to allow the
+	 * DRM_IOCTL_WAIT_VBLANK to operate correctly.
+	 */
+	drm->irq_enabled = 1;
+
 	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
 	if (err < 0)
 		return err;
@@ -378,8 +387,7 @@ static int tegra_syncpt_incr(struct drm_device *drm, void *data,
 	if (!sp)
 		return -EINVAL;
 
-	host1x_syncpt_incr(sp);
-	return 0;
+	return host1x_syncpt_incr(sp);
 }
 
 static int tegra_syncpt_wait(struct drm_device *drm, void *data,
@@ -605,7 +613,7 @@ static void tegra_debugfs_cleanup(struct drm_minor *minor)
 #endif
 
 struct drm_driver tegra_drm_driver = {
-	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
+	.driver_features = DRIVER_MODESET | DRIVER_GEM,
 	.load = tegra_drm_load,
 	.unload = tegra_drm_unload,
 	.open = tegra_drm_open,
diff --git a/drivers/gpu/host1x/drm/gr2d.c b/drivers/gpu/host1x/drm/gr2d.c
index 6a45ae0..27ffcf1 100644
--- a/drivers/gpu/host1x/drm/gr2d.c
+++ b/drivers/gpu/host1x/drm/gr2d.c
@@ -84,7 +84,7 @@ static struct host1x_bo *host1x_bo_lookup(struct drm_device *drm,
 
 	gem = drm_gem_object_lookup(drm, file, handle);
 	if (!gem)
-		return 0;
+		return NULL;
 
 	mutex_lock(&drm->struct_mutex);
 	drm_gem_object_unreference(gem);
@@ -135,8 +135,10 @@ static int gr2d_submit(struct host1x_drm_context *context,
 			goto fail;
 
 		bo = host1x_bo_lookup(drm, file, cmdbuf.handle);
-		if (!bo)
+		if (!bo) {
+			err = -ENOENT;
 			goto fail;
+		}
 
 		host1x_job_add_gather(job, bo, cmdbuf.words, cmdbuf.offset);
 		num_cmdbufs--;
@@ -158,8 +160,10 @@ static int gr2d_submit(struct host1x_drm_context *context,
 		reloc->cmdbuf = cmdbuf;
 		reloc->target = target;
 
-		if (!reloc->target || !reloc->cmdbuf)
+		if (!reloc->target || !reloc->cmdbuf) {
+			err = -ENOENT;
 			goto fail;
+		}
 	}
 
 	err = copy_from_user(job->waitchk, waitchks,
@@ -281,7 +285,7 @@ static int gr2d_probe(struct platform_device *pdev)
 	if (!gr2d->channel)
 		return -ENOMEM;
 
-	*syncpts = host1x_syncpt_request(dev, 0);
+	*syncpts = host1x_syncpt_request(dev, false);
 	if (!(*syncpts)) {
 		host1x_channel_free(gr2d->channel);
 		return -ENOMEM;
diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c
index 590b69d..2ee4ad5 100644
--- a/drivers/gpu/host1x/hw/cdma_hw.c
+++ b/drivers/gpu/host1x/hw/cdma_hw.c
@@ -44,7 +44,7 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr,
 	u32 i;
 
 	for (i = 0; i < syncpt_incrs; i++)
-		host1x_syncpt_cpu_incr(cdma->timeout.syncpt);
+		host1x_syncpt_incr(cdma->timeout.syncpt);
 
 	/* after CPU incr, ensure shadow is up to date */
 	host1x_syncpt_load(cdma->timeout.syncpt);
diff --git a/drivers/gpu/host1x/hw/syncpt_hw.c b/drivers/gpu/host1x/hw/syncpt_hw.c
index 6117499..0cf6095 100644
--- a/drivers/gpu/host1x/hw/syncpt_hw.c
+++ b/drivers/gpu/host1x/hw/syncpt_hw.c
@@ -77,21 +77,19 @@ static u32 syncpt_load(struct host1x_syncpt *sp)
  * Write a cpu syncpoint increment to the hardware, without touching
  * the cache.
  */
-static void syncpt_cpu_incr(struct host1x_syncpt *sp)
+static int syncpt_cpu_incr(struct host1x_syncpt *sp)
 {
 	struct host1x *host = sp->host;
 	u32 reg_offset = sp->id / 32;
 
 	if (!host1x_syncpt_client_managed(sp) &&
-	    host1x_syncpt_idle(sp)) {
-		dev_err(host->dev, "Trying to increment syncpoint id %d beyond max\n",
-			sp->id);
-		host1x_debug_dump(sp->host);
-		return;
-	}
+	    host1x_syncpt_idle(sp))
+		return -EINVAL;
 	host1x_sync_writel(host, BIT_MASK(sp->id),
 			   HOST1X_SYNC_SYNCPT_CPU_INCR(reg_offset));
 	wmb();
+
+	return 0;
 }
 
 /* remove a wait pointed to by patch_addr */
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index f665d67..cc80766 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -228,17 +228,15 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 	void *cmdbuf_page_addr = NULL;
 
 	/* pin & patch the relocs for one gather */
-	while (i < job->num_relocs) {
+	for (i = 0; i < job->num_relocs; i++) {
 		struct host1x_reloc *reloc = &job->relocarray[i];
 		u32 reloc_addr = (job->reloc_addr_phys[i] +
 			reloc->target_offset) >> reloc->shift;
 		u32 *target;
 
 		/* skip all other gathers */
-		if (!(reloc->cmdbuf && cmdbuf == reloc->cmdbuf)) {
-			i++;
+		if (cmdbuf != reloc->cmdbuf)
 			continue;
-		}
 
 		if (last_page != reloc->cmdbuf_offset >> PAGE_SHIFT) {
 			if (cmdbuf_page_addr)
@@ -257,9 +255,6 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 
 		target = cmdbuf_page_addr + (reloc->cmdbuf_offset & ~PAGE_MASK);
 		*target = reloc_addr;
-
-		/* mark this gather as handled */
-		reloc->cmdbuf = 0;
 	}
 
 	if (cmdbuf_page_addr)
@@ -268,15 +263,15 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 	return 0;
 }
 
-static int check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
+static bool check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
 		       unsigned int offset)
 {
 	offset *= sizeof(u32);
 
 	if (reloc->cmdbuf != cmdbuf || reloc->cmdbuf_offset != offset)
-		return -EINVAL;
+		return false;
 
-	return 0;
+	return true;
 }
 
 struct host1x_firewall {
@@ -307,10 +302,10 @@ static int check_mask(struct host1x_firewall *fw)
 
 		if (mask & 1) {
 			if (fw->job->is_addr_reg(fw->dev, fw->class, reg)) {
-				bool bad_reloc = check_reloc(fw->reloc,
-							     fw->cmdbuf_id,
-							     fw->offset);
-				if (!fw->num_relocs || bad_reloc)
+				if (!fw->num_relocs)
+					return -EINVAL;
+				if (!check_reloc(fw->reloc, fw->cmdbuf_id,
+						 fw->offset))
 					return -EINVAL;
 				fw->reloc++;
 				fw->num_relocs--;
@@ -330,14 +325,14 @@ static int check_incr(struct host1x_firewall *fw)
 	u32 count = fw->count;
 	u32 reg = fw->reg;
 
-	while (fw) {
+	while (count) {
 		if (fw->words == 0)
 			return -EINVAL;
 
 		if (fw->job->is_addr_reg(fw->dev, fw->class, reg)) {
-			bool bad_reloc = check_reloc(fw->reloc, fw->cmdbuf_id,
-						     fw->offset);
-			if (!fw->num_relocs || bad_reloc)
+			if (!fw->num_relocs)
+				return -EINVAL;
+			if (!check_reloc(fw->reloc, fw->cmdbuf_id, fw->offset))
 				return -EINVAL;
 			fw->reloc++;
 			fw->num_relocs--;
@@ -361,9 +356,9 @@ static int check_nonincr(struct host1x_firewall *fw)
 			return -EINVAL;
 
 		if (is_addr_reg) {
-			bool bad_reloc = check_reloc(fw->reloc, fw->cmdbuf_id,
-						     fw->offset);
-			if (!fw->num_relocs || bad_reloc)
+			if (!fw->num_relocs)
+				return -EINVAL;
+			if (!check_reloc(fw->reloc, fw->cmdbuf_id, fw->offset))
 				return -EINVAL;
 			fw->reloc++;
 			fw->num_relocs--;
@@ -376,69 +371,58 @@ static int check_nonincr(struct host1x_firewall *fw)
 	return 0;
 }
 
-static int validate(struct host1x_job *job, struct device *dev,
-		    struct host1x_job_gather *g)
+static int validate(struct host1x_firewall *fw, struct host1x_job_gather *g)
 {
-	u32 *cmdbuf_base;
+	u32 *cmdbuf_base = (u32 *)fw->job->gather_copy_mapped +
+		(g->offset / sizeof(u32));
 	int err = 0;
-	struct host1x_firewall fw;
 
-	fw.job = job;
-	fw.dev = dev;
-	fw.reloc = job->relocarray;
-	fw.num_relocs = job->num_relocs;
-	fw.cmdbuf_id = g->bo;
-
-	fw.offset = 0;
-	fw.class = 0;
-
-	if (!job->is_addr_reg)
+	if (!fw->job->is_addr_reg)
 		return 0;
 
-	cmdbuf_base = host1x_bo_mmap(g->bo);
-	if (!cmdbuf_base)
-		return -ENOMEM;
+	fw->words = g->words;
+	fw->cmdbuf_id = g->bo;
+	fw->offset = 0;
 
-	fw.words = g->words;
-	while (fw.words && !err) {
-		u32 word = cmdbuf_base[fw.offset];
+	while (fw->words && !err) {
+		u32 word = cmdbuf_base[fw->offset];
 		u32 opcode = (word & 0xf0000000) >> 28;
 
-		fw.mask = 0;
-		fw.reg = 0;
-		fw.count = 0;
-		fw.words--;
-		fw.offset++;
+		fw->mask = 0;
+		fw->reg = 0;
+		fw->count = 0;
+		fw->words--;
+		fw->offset++;
 
 		switch (opcode) {
 		case 0:
-			fw.class = word >> 6 & 0x3ff;
-			fw.mask = word & 0x3f;
-			fw.reg = word >> 16 & 0xfff;
-			err = check_mask(&fw);
+			fw->class = word >> 6 & 0x3ff;
+			fw->mask = word & 0x3f;
+			fw->reg = word >> 16 & 0xfff;
+			err = check_mask(fw);
 			if (err)
 				goto out;
 			break;
 		case 1:
-			fw.reg = word >> 16 & 0xfff;
-			fw.count = word & 0xffff;
-			err = check_incr(&fw);
+			fw->reg = word >> 16 & 0xfff;
+			fw->count = word & 0xffff;
+			err = check_incr(fw);
 			if (err)
 				goto out;
 			break;
 
 		case 2:
-			fw.reg = word >> 16 & 0xfff;
-			fw.count = word & 0xffff;
-			err = check_nonincr(&fw);
+			fw->reg = word >> 16 & 0xfff;
+			fw->count = word & 0xffff;
+			err = check_nonincr(fw);
 			if (err)
 				goto out;
 			break;
 
 		case 3:
-			fw.mask = word & 0xffff;
-			fw.reg = word >> 16 & 0xfff;
-			err = check_mask(&fw);
+			fw->mask = word & 0xffff;
+			fw->reg = word >> 16 & 0xfff;
+			err = check_mask(fw);
 			if (err)
 				goto out;
 			break;
@@ -453,21 +437,26 @@ static int validate(struct host1x_job *job, struct device *dev,
 	}
 
 	/* No relocs should remain at this point */
-	if (fw.num_relocs)
+	if (fw->num_relocs)
 		err = -EINVAL;
 
 out:
-	host1x_bo_munmap(g->bo, cmdbuf_base);
-
 	return err;
 }
 
 static inline int copy_gathers(struct host1x_job *job, struct device *dev)
 {
+	struct host1x_firewall fw;
 	size_t size = 0;
 	size_t offset = 0;
 	int i;
 
+	fw.job = job;
+	fw.dev = dev;
+	fw.reloc = job->relocarray;
+	fw.num_relocs = job->num_relocs;
+	fw.class = 0;
+
 	for (i = 0; i < job->num_gathers; i++) {
 		struct host1x_job_gather *g = &job->gathers[i];
 		size += g->words * sizeof(u32);
@@ -488,14 +477,19 @@ static inline int copy_gathers(struct host1x_job *job, struct device *dev)
 		struct host1x_job_gather *g = &job->gathers[i];
 		void *gather;
 
+		/* Copy the gather */
 		gather = host1x_bo_mmap(g->bo);
 		memcpy(job->gather_copy_mapped + offset, gather + g->offset,
 		       g->words * sizeof(u32));
 		host1x_bo_munmap(g->bo, gather);
 
+		/* Store the location in the buffer */
 		g->base = job->gather_copy;
 		g->offset = offset;
-		g->bo = NULL;
+
+		/* Validate the job */
+		if (validate(&fw, g))
+			return -EINVAL;
 
 		offset += g->words * sizeof(u32);
 	}
@@ -540,20 +534,11 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev)
 			if (job->gathers[j].bo == g->bo)
 				job->gathers[j].handled = true;
 
-		err = 0;
-
-		if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL))
-			err = validate(job, dev, g);
-
+		err = do_relocs(job, g->bo);
 		if (err)
-			dev_err(dev, "Job invalid (err=%d)\n", err);
-
-		if (!err)
-			err = do_relocs(job, g->bo);
-
-		if (!err)
-			err = do_waitchks(job, host, g->bo);
+			break;
 
+		err = do_waitchks(job, host, g->bo);
 		if (err)
 			break;
 	}
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index 4b49345..409745b 100644
--- a/drivers/gpu/host1x/syncpt.c
+++ b/drivers/gpu/host1x/syncpt.c
@@ -32,7 +32,7 @@
 
 static struct host1x_syncpt *_host1x_syncpt_alloc(struct host1x *host,
 						  struct device *dev,
-						  int client_managed)
+						  bool client_managed)
 {
 	int i;
 	struct host1x_syncpt *sp = host->syncpt;
@@ -40,7 +40,8 @@ static struct host1x_syncpt *_host1x_syncpt_alloc(struct host1x *host,
 
 	for (i = 0; i < host->info->nb_pts && sp->name; i++, sp++)
 		;
-	if (sp->dev)
+
+	if (i >= host->info->nb_pts)
 		return NULL;
 
 	name = kasprintf(GFP_KERNEL, "%02d-%s", sp->id,
@@ -128,22 +129,11 @@ u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp)
 }
 
 /*
- * Write a cpu syncpoint increment to the hardware, without touching
- * the cache. Caller is responsible for host being powered.
- */
-void host1x_syncpt_cpu_incr(struct host1x_syncpt *sp)
-{
-	host1x_hw_syncpt_cpu_incr(sp->host, sp);
-}
-
-/*
  * Increment syncpoint value from cpu, updating cache
  */
-void host1x_syncpt_incr(struct host1x_syncpt *sp)
+int host1x_syncpt_incr(struct host1x_syncpt *sp)
 {
-	if (host1x_syncpt_client_managed(sp))
-		host1x_syncpt_incr_max(sp, 1);
-	host1x_syncpt_cpu_incr(sp);
+	return host1x_hw_syncpt_cpu_incr(sp->host, sp);
 }
 
 /*
@@ -331,7 +321,7 @@ int host1x_syncpt_init(struct host1x *host)
 	host1x_syncpt_restore(host);
 
 	/* Allocate sync point to use for clearing waits for expired fences */
-	host->nop_sp = _host1x_syncpt_alloc(host, NULL, 0);
+	host->nop_sp = _host1x_syncpt_alloc(host, NULL, false);
 	if (!host->nop_sp)
 		return -ENOMEM;
 
@@ -339,7 +329,7 @@ int host1x_syncpt_init(struct host1x *host)
 }
 
 struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
-					    int client_managed)
+					    bool client_managed)
 {
 	struct host1x *host = dev_get_drvdata(dev->parent);
 	return _host1x_syncpt_alloc(host, dev, client_managed);
@@ -353,7 +343,7 @@ void host1x_syncpt_free(struct host1x_syncpt *sp)
 	kfree(sp->name);
 	sp->dev = NULL;
 	sp->name = NULL;
-	sp->client_managed = 0;
+	sp->client_managed = false;
 }
 
 void host1x_syncpt_deinit(struct host1x *host)
diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h
index c998061..267c0b9 100644
--- a/drivers/gpu/host1x/syncpt.h
+++ b/drivers/gpu/host1x/syncpt.h
@@ -36,7 +36,7 @@ struct host1x_syncpt {
 	atomic_t max_val;
 	u32 base_val;
 	const char *name;
-	int client_managed;
+	bool client_managed;
 	struct host1x *host;
 	struct device *dev;
 
@@ -94,7 +94,7 @@ static inline bool host1x_syncpt_check_max(struct host1x_syncpt *sp, u32 real)
 }
 
 /* Return true if sync point is client managed. */
-static inline int host1x_syncpt_client_managed(struct host1x_syncpt *sp)
+static inline bool host1x_syncpt_client_managed(struct host1x_syncpt *sp)
 {
 	return sp->client_managed;
 }
@@ -115,9 +115,6 @@ static inline bool host1x_syncpt_idle(struct host1x_syncpt *sp)
 /* Return pointer to struct denoting sync point id. */
 struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, u32 id);
 
-/* Request incrementing a sync point. */
-void host1x_syncpt_cpu_incr(struct host1x_syncpt *sp);
-
 /* Load current value from hardware to the shadow register. */
 u32 host1x_syncpt_load(struct host1x_syncpt *sp);
 
@@ -133,8 +130,8 @@ void host1x_syncpt_restore(struct host1x *host);
 /* Read current wait base value into shadow register and return it. */
 u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp);
 
-/* Increment sync point and its max. */
-void host1x_syncpt_incr(struct host1x_syncpt *sp);
+/* Request incrementing a sync point. */
+int host1x_syncpt_incr(struct host1x_syncpt *sp);
 
 /* Indicate future operations by incrementing the sync point max. */
 u32 host1x_syncpt_incr_max(struct host1x_syncpt *sp, u32 incrs);
@@ -157,7 +154,7 @@ u32 host1x_syncpt_id(struct host1x_syncpt *sp);
 
 /* Allocate a sync point for a device. */
 struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
-		int client_managed);
+					    bool client_managed);
 
 /* Free a sync point. */
 void host1x_syncpt_free(struct host1x_syncpt *sp);
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 71c2c71..34fbc2f 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -3269,9 +3269,9 @@ static int cma_netdev_change(struct net_device *ndev, struct rdma_id_private *id
 }
 
 static int cma_netdev_callback(struct notifier_block *self, unsigned long event,
-			       void *ctx)
+			       void *ptr)
 {
-	struct net_device *ndev = (struct net_device *)ctx;
+	struct net_device *ndev = netdev_notifier_info_to_dev(ptr);
 	struct cma_device *cma_dev;
 	struct rdma_id_private *id_priv;
 	int ret = NOTIFY_DONE;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 23d7343..a188d31 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -1161,7 +1161,7 @@ static void netdev_removed(struct mlx4_ib_dev *dev, int port)
 static int mlx4_ib_netdev_event(struct notifier_block *this, unsigned long event,
 				void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct mlx4_ib_dev *ibdev;
 	struct net_device *oldnd;
 	struct mlx4_ib_iboe *iboe;
diff --git a/drivers/iommu/msm_iommu_dev.c b/drivers/iommu/msm_iommu_dev.c
index 9144a6b..6ba3514 100644
--- a/drivers/iommu/msm_iommu_dev.c
+++ b/drivers/iommu/msm_iommu_dev.c
@@ -291,25 +291,20 @@ static int msm_iommu_ctx_probe(struct platform_device *pdev)
 {
 	struct msm_iommu_ctx_dev *c = pdev->dev.platform_data;
 	struct msm_iommu_drvdata *drvdata;
-	struct msm_iommu_ctx_drvdata *ctx_drvdata = NULL;
+	struct msm_iommu_ctx_drvdata *ctx_drvdata;
 	int i, ret;
-	if (!c || !pdev->dev.parent) {
-		ret = -EINVAL;
-		goto fail;
-	}
 
-	drvdata = dev_get_drvdata(pdev->dev.parent);
+	if (!c || !pdev->dev.parent)
+		return -EINVAL;
 
-	if (!drvdata) {
-		ret = -ENODEV;
-		goto fail;
-	}
+	drvdata = dev_get_drvdata(pdev->dev.parent);
+	if (!drvdata)
+		return -ENODEV;
 
 	ctx_drvdata = kzalloc(sizeof(*ctx_drvdata), GFP_KERNEL);
-	if (!ctx_drvdata) {
-		ret = -ENOMEM;
-		goto fail;
-	}
+	if (!ctx_drvdata)
+		return -ENOMEM;
+
 	ctx_drvdata->num = c->num;
 	ctx_drvdata->pdev = pdev;
 
@@ -403,6 +398,7 @@ static int __init msm_iommu_driver_init(void)
 
 	ret = platform_driver_register(&msm_iommu_ctx_driver);
 	if (ret != 0) {
+		platform_driver_unregister(&msm_iommu_driver);
 		pr_err("Failed to register IOMMU context driver\n");
 		goto error;
 	}
diff --git a/drivers/isdn/i4l/isdn_net.c b/drivers/isdn/i4l/isdn_net.c
index 88d657d..8b98d53 100644
--- a/drivers/isdn/i4l/isdn_net.c
+++ b/drivers/isdn/i4l/isdn_net.c
@@ -885,7 +885,7 @@ isdn_net_log_skb(struct sk_buff *skb, isdn_net_local *lp)
 
 	addinfo[0] = '\0';
 	/* This check stolen from 2.1.72 dev_queue_xmit_nit() */
-	if (p < skb->data || skb->network_header >= skb->tail) {
+	if (p < skb->data || skb_network_header(skb) >= skb_tail_pointer(skb)) {
 		/* fall back to old isdn_net_log_packet method() */
 		char *buf = skb->data;
 
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 3835321..b45b240 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -25,6 +25,9 @@ menuconfig NETDEVICES
 # that for each of the symbols.
 if NETDEVICES
 
+config MII
+	tristate
+
 config NET_CORE
 	default y
 	bool "Network core driver support"
@@ -100,13 +103,6 @@ config NET_FC
 	  adaptor below. You also should have said Y to "SCSI support" and
 	  "SCSI generic support".
 
-config MII
-	tristate "Generic Media Independent Interface device support"
-	help
-	  Most ethernet controllers have MII transceiver either as an external
-	  or internal device.  It is safe to say Y or M here even if your
-	  ethernet card lacks MII.
-
 config IFB
 	tristate "Intermediate Functional Block support"
 	depends on NET_CLS_ACT
@@ -244,6 +240,16 @@ config VIRTIO_NET
 	  This is the virtual network driver for virtio.  It can be used with
 	  lguest or QEMU based VMMs (like KVM or Xen).  Say Y or M.
 
+config NLMON
+	tristate "Virtual netlink monitoring device"
+	---help---
+	  This option enables a monitoring net device for netlink skbs. The
+	  purpose of this is to analyze netlink messages with packet sockets.
+	  Thus applications like tcpdump will be able to see local netlink
+	  messages if they tap into the netlink device, record pcaps for further
+	  diagnostics, etc. This is mostly intended for developers or support
+	  to debug netlink issues. If unsure, say N.
+
 endif # NET_CORE
 
 config SUNGEM_PHY
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index ef3d090..3fef8a8 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -22,6 +22,7 @@ obj-$(CONFIG_TUN) += tun.o
 obj-$(CONFIG_VETH) += veth.o
 obj-$(CONFIG_VIRTIO_NET) += virtio_net.o
 obj-$(CONFIG_VXLAN) += vxlan.o
+obj-$(CONFIG_NLMON) += nlmon.o
 
 #
 # Networking Drivers
diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
index e02cc26..4ea8ed1 100644
--- a/drivers/net/bonding/bond_alb.c
+++ b/drivers/net/bonding/bond_alb.c
@@ -1056,7 +1056,7 @@ static int alb_set_slave_mac_addr(struct slave *slave, u8 addr[])
  *
  */
 
-static void alb_swap_mac_addr(struct bonding *bond, struct slave *slave1, struct slave *slave2)
+static void alb_swap_mac_addr(struct slave *slave1, struct slave *slave2)
 {
 	u8 tmp_mac_addr[ETH_ALEN];
 
@@ -1129,6 +1129,7 @@ static void alb_change_hw_addr_on_detach(struct bonding *bond, struct slave *sla
 {
 	int perm_curr_diff;
 	int perm_bond_diff;
+	struct slave *found_slave;
 
 	perm_curr_diff = !ether_addr_equal_64bits(slave->perm_hwaddr,
 						  slave->dev->dev_addr);
@@ -1136,21 +1137,12 @@ static void alb_change_hw_addr_on_detach(struct bonding *bond, struct slave *sla
 						  bond->dev->dev_addr);
 
 	if (perm_curr_diff && perm_bond_diff) {
-		struct slave *tmp_slave;
-		int i, found = 0;
-
-		bond_for_each_slave(bond, tmp_slave, i) {
-			if (ether_addr_equal_64bits(slave->perm_hwaddr,
-						    tmp_slave->dev->dev_addr)) {
-				found = 1;
-				break;
-			}
-		}
+		found_slave = bond_slave_has_mac(bond, slave->perm_hwaddr);
 
-		if (found) {
+		if (found_slave) {
 			/* locking: needs RTNL and nothing else */
-			alb_swap_mac_addr(bond, slave, tmp_slave);
-			alb_fasten_mac_swap(bond, slave, tmp_slave);
+			alb_swap_mac_addr(slave, found_slave);
+			alb_fasten_mac_swap(bond, slave, found_slave);
 		}
 	}
 }
@@ -1175,16 +1167,13 @@ static void alb_change_hw_addr_on_detach(struct bonding *bond, struct slave *sla
  * @slave.
  *
  * assumption: this function is called before @slave is attached to the
- * 	       bond slave list.
- *
- * caller must hold the bond lock for write since the mac addresses are compared
- * and may be swapped.
+ *	       bond slave list.
  */
 static int alb_handle_addr_collision_on_attach(struct bonding *bond, struct slave *slave)
 {
-	struct slave *tmp_slave1, *tmp_slave2, *free_mac_slave;
+	struct slave *tmp_slave1, *free_mac_slave = NULL;
 	struct slave *has_bond_addr = bond->curr_active_slave;
-	int i, j, found = 0;
+	int i;
 
 	if (bond->slave_cnt == 0) {
 		/* this is the first slave */
@@ -1196,15 +1185,7 @@ static int alb_handle_addr_collision_on_attach(struct bonding *bond, struct slav
 	 * slaves in the bond.
 	 */
 	if (!ether_addr_equal_64bits(slave->perm_hwaddr, bond->dev->dev_addr)) {
-		bond_for_each_slave(bond, tmp_slave1, i) {
-			if (ether_addr_equal_64bits(tmp_slave1->dev->dev_addr,
-						    slave->dev->dev_addr)) {
-				found = 1;
-				break;
-			}
-		}
-
-		if (!found)
+		if (!bond_slave_has_mac(bond, slave->dev->dev_addr))
 			return 0;
 
 		/* Try setting slave mac to bond address and fall-through
@@ -1215,19 +1196,8 @@ static int alb_handle_addr_collision_on_attach(struct bonding *bond, struct slav
 	/* The slave's address is equal to the address of the bond.
 	 * Search for a spare address in the bond for this slave.
 	 */
-	free_mac_slave = NULL;
-
 	bond_for_each_slave(bond, tmp_slave1, i) {
-		found = 0;
-		bond_for_each_slave(bond, tmp_slave2, j) {
-			if (ether_addr_equal_64bits(tmp_slave1->perm_hwaddr,
-						    tmp_slave2->dev->dev_addr)) {
-				found = 1;
-				break;
-			}
-		}
-
-		if (!found) {
+		if (!bond_slave_has_mac(bond, tmp_slave1->perm_hwaddr)) {
 			/* no slave has tmp_slave1's perm addr
 			 * as its curr addr
 			 */
@@ -1607,15 +1577,7 @@ int bond_alb_init_slave(struct bonding *bond, struct slave *slave)
 		return res;
 	}
 
-	/* caller must hold the bond lock for write since the mac addresses
-	 * are compared and may be swapped.
-	 */
-	read_lock(&bond->lock);
-
 	res = alb_handle_addr_collision_on_attach(bond, slave);
-
-	read_unlock(&bond->lock);
-
 	if (res) {
 		return res;
 	}
@@ -1698,7 +1660,6 @@ void bond_alb_handle_active_change(struct bonding *bond, struct slave *new_slave
 	__acquires(&bond->curr_slave_lock)
 {
 	struct slave *swap_slave;
-	int i;
 
 	if (bond->curr_active_slave == new_slave) {
 		return;
@@ -1720,17 +1681,8 @@ void bond_alb_handle_active_change(struct bonding *bond, struct slave *new_slave
 	/* set the new curr_active_slave to the bonds mac address
 	 * i.e. swap mac addresses of old curr_active_slave and new curr_active_slave
 	 */
-	if (!swap_slave) {
-		struct slave *tmp_slave;
-		/* find slave that is holding the bond's mac address */
-		bond_for_each_slave(bond, tmp_slave, i) {
-			if (ether_addr_equal_64bits(tmp_slave->dev->dev_addr,
-						    bond->dev->dev_addr)) {
-				swap_slave = tmp_slave;
-				break;
-			}
-		}
-	}
+	if (!swap_slave)
+		swap_slave = bond_slave_has_mac(bond, bond->dev->dev_addr);
 
 	/*
 	 * Arrange for swap_slave and new_slave to temporarily be
@@ -1750,16 +1702,12 @@ void bond_alb_handle_active_change(struct bonding *bond, struct slave *new_slave
 	/* curr_active_slave must be set before calling alb_swap_mac_addr */
 	if (swap_slave) {
 		/* swap mac address */
-		alb_swap_mac_addr(bond, swap_slave, new_slave);
-	} else {
-		/* set the new_slave to the bond mac address */
-		alb_set_slave_mac_addr(new_slave, bond->dev->dev_addr);
-	}
-
-	if (swap_slave) {
+		alb_swap_mac_addr(swap_slave, new_slave);
 		alb_fasten_mac_swap(bond, swap_slave, new_slave);
 		read_lock(&bond->lock);
 	} else {
+		/* set the new_slave to the bond mac address */
+		alb_set_slave_mac_addr(new_slave, bond->dev->dev_addr);
 		read_lock(&bond->lock);
 		alb_send_learning_packets(new_slave, bond->dev->dev_addr);
 	}
@@ -1776,9 +1724,8 @@ int bond_alb_set_mac_address(struct net_device *bond_dev, void *addr)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
 	struct sockaddr *sa = addr;
-	struct slave *slave, *swap_slave;
+	struct slave *swap_slave;
 	int res;
-	int i;
 
 	if (!is_valid_ether_addr(sa->sa_data)) {
 		return -EADDRNOTAVAIL;
@@ -1799,18 +1746,10 @@ int bond_alb_set_mac_address(struct net_device *bond_dev, void *addr)
 		return 0;
 	}
 
-	swap_slave = NULL;
-
-	bond_for_each_slave(bond, slave, i) {
-		if (ether_addr_equal_64bits(slave->dev->dev_addr,
-					    bond_dev->dev_addr)) {
-			swap_slave = slave;
-			break;
-		}
-	}
+	swap_slave = bond_slave_has_mac(bond, bond_dev->dev_addr);
 
 	if (swap_slave) {
-		alb_swap_mac_addr(bond, swap_slave, bond->curr_active_slave);
+		alb_swap_mac_addr(swap_slave, bond->curr_active_slave);
 		alb_fasten_mac_swap(bond, swap_slave, bond->curr_active_slave);
 	} else {
 		alb_set_slave_mac_addr(bond->curr_active_slave, bond_dev->dev_addr);
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index f975696..07f257d4 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -104,6 +104,7 @@ static char *xmit_hash_policy;
 static int arp_interval = BOND_LINK_ARP_INTERV;
 static char *arp_ip_target[BOND_MAX_ARP_TARGETS];
 static char *arp_validate;
+static char *arp_all_targets;
 static char *fail_over_mac;
 static int all_slaves_active = 0;
 static struct bond_params bonding_defaults;
@@ -166,6 +167,8 @@ module_param(arp_validate, charp, 0);
 MODULE_PARM_DESC(arp_validate, "validate src/dst of ARP probes; "
 			       "0 for none (default), 1 for active, "
 			       "2 for backup, 3 for all");
+module_param(arp_all_targets, charp, 0);
+MODULE_PARM_DESC(arp_all_targets, "fail on any/all arp targets timeout; 0 for any (default), 1 for all");
 module_param(fail_over_mac, charp, 0);
 MODULE_PARM_DESC(fail_over_mac, "For active-backup, do not set all slaves to "
 				"the same MAC; 0 for none (default), "
@@ -216,6 +219,12 @@ const struct bond_parm_tbl xmit_hashtype_tbl[] = {
 {	NULL,			-1},
 };
 
+const struct bond_parm_tbl arp_all_targets_tbl[] = {
+{	"any",			BOND_ARP_TARGETS_ANY},
+{	"all",			BOND_ARP_TARGETS_ALL},
+{	NULL,			-1},
+};
+
 const struct bond_parm_tbl arp_validate_tbl[] = {
 {	"none",			BOND_ARP_VALIDATE_NONE},
 {	"active",		BOND_ARP_VALIDATE_ACTIVE},
@@ -706,45 +715,6 @@ static int bond_set_allmulti(struct bonding *bond, int inc)
 	return err;
 }
 
-/*
- * Add a Multicast address to slaves
- * according to mode
- */
-static void bond_mc_add(struct bonding *bond, void *addr)
-{
-	if (USES_PRIMARY(bond->params.mode)) {
-		/* write lock already acquired */
-		if (bond->curr_active_slave)
-			dev_mc_add(bond->curr_active_slave->dev, addr);
-	} else {
-		struct slave *slave;
-		int i;
-
-		bond_for_each_slave(bond, slave, i)
-			dev_mc_add(slave->dev, addr);
-	}
-}
-
-/*
- * Remove a multicast address from slave
- * according to mode
- */
-static void bond_mc_del(struct bonding *bond, void *addr)
-{
-	if (USES_PRIMARY(bond->params.mode)) {
-		/* write lock already acquired */
-		if (bond->curr_active_slave)
-			dev_mc_del(bond->curr_active_slave->dev, addr);
-	} else {
-		struct slave *slave;
-		int i;
-		bond_for_each_slave(bond, slave, i) {
-			dev_mc_del(slave->dev, addr);
-		}
-	}
-}
-
-
 static void __bond_resend_igmp_join_requests(struct net_device *dev)
 {
 	struct in_device *in_dev;
@@ -810,17 +780,15 @@ static void bond_resend_igmp_join_requests_delayed(struct work_struct *work)
 	bond_resend_igmp_join_requests(bond);
 }
 
-/*
- * flush all members of flush->mc_list from device dev->mc_list
+/* Flush bond's hardware addresses from slave
  */
-static void bond_mc_list_flush(struct net_device *bond_dev,
+static void bond_hw_addr_flush(struct net_device *bond_dev,
 			       struct net_device *slave_dev)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
-	struct netdev_hw_addr *ha;
 
-	netdev_for_each_mc_addr(ha, bond_dev)
-		dev_mc_del(slave_dev, ha->addr);
+	dev_uc_unsync(slave_dev, bond_dev);
+	dev_mc_unsync(slave_dev, bond_dev);
 
 	if (bond->params.mode == BOND_MODE_8023AD) {
 		/* del lacpdu mc addr from mc list */
@@ -832,22 +800,14 @@ static void bond_mc_list_flush(struct net_device *bond_dev,
 
 /*--------------------------- Active slave change ---------------------------*/
 
-/*
- * Update the mc list and multicast-related flags for the new and
- * old active slaves (if any) according to the multicast mode, and
- * promiscuous flags unconditionally.
+/* Update the hardware address list and promisc/allmulti for the new and
+ * old active slaves (if any).  Modes that are !USES_PRIMARY keep all
+ * slaves up date at all times; only the USES_PRIMARY modes need to call
+ * this function to swap these settings during a failover.
  */
-static void bond_mc_swap(struct bonding *bond, struct slave *new_active,
-			 struct slave *old_active)
+static void bond_hw_addr_swap(struct bonding *bond, struct slave *new_active,
+			      struct slave *old_active)
 {
-	struct netdev_hw_addr *ha;
-
-	if (!USES_PRIMARY(bond->params.mode))
-		/* nothing to do -  mc list is already up-to-date on
-		 * all slaves
-		 */
-		return;
-
 	if (old_active) {
 		if (bond->dev->flags & IFF_PROMISC)
 			dev_set_promiscuity(old_active->dev, -1);
@@ -855,10 +815,7 @@ static void bond_mc_swap(struct bonding *bond, struct slave *new_active,
 		if (bond->dev->flags & IFF_ALLMULTI)
 			dev_set_allmulti(old_active->dev, -1);
 
-		netif_addr_lock_bh(bond->dev);
-		netdev_for_each_mc_addr(ha, bond->dev)
-			dev_mc_del(old_active->dev, ha->addr);
-		netif_addr_unlock_bh(bond->dev);
+		bond_hw_addr_flush(bond->dev, old_active->dev);
 	}
 
 	if (new_active) {
@@ -870,12 +827,29 @@ static void bond_mc_swap(struct bonding *bond, struct slave *new_active,
 			dev_set_allmulti(new_active->dev, 1);
 
 		netif_addr_lock_bh(bond->dev);
-		netdev_for_each_mc_addr(ha, bond->dev)
-			dev_mc_add(new_active->dev, ha->addr);
+		dev_uc_sync(new_active->dev, bond->dev);
+		dev_mc_sync(new_active->dev, bond->dev);
 		netif_addr_unlock_bh(bond->dev);
 	}
 }
 
+/**
+ * bond_set_dev_addr - clone slave's address to bond
+ * @bond_dev: bond net device
+ * @slave_dev: slave net device
+ *
+ * Should be called with RTNL held.
+ */
+static void bond_set_dev_addr(struct net_device *bond_dev,
+			      struct net_device *slave_dev)
+{
+	pr_debug("bond_dev=%p slave_dev=%p slave_dev->addr_len=%d\n",
+		 bond_dev, slave_dev, slave_dev->addr_len);
+	memcpy(bond_dev->dev_addr, slave_dev->dev_addr, slave_dev->addr_len);
+	bond_dev->addr_assign_type = NET_ADDR_STOLEN;
+	call_netdevice_notifiers(NETDEV_CHANGEADDR, bond_dev);
+}
+
 /*
  * bond_do_fail_over_mac
  *
@@ -898,11 +872,9 @@ static void bond_do_fail_over_mac(struct bonding *bond,
 	switch (bond->params.fail_over_mac) {
 	case BOND_FOM_ACTIVE:
 		if (new_active) {
-			memcpy(bond->dev->dev_addr,  new_active->dev->dev_addr,
-			       new_active->dev->addr_len);
 			write_unlock_bh(&bond->curr_slave_lock);
 			read_unlock(&bond->lock);
-			call_netdevice_notifiers(NETDEV_CHANGEADDR, bond->dev);
+			bond_set_dev_addr(bond->dev, new_active->dev);
 			read_lock(&bond->lock);
 			write_lock_bh(&bond->curr_slave_lock);
 		}
@@ -1090,7 +1062,7 @@ void bond_change_active_slave(struct bonding *bond, struct slave *new_active)
 	}
 
 	if (USES_PRIMARY(bond->params.mode))
-		bond_mc_swap(bond, new_active, old_active);
+		bond_hw_addr_swap(bond, new_active, old_active);
 
 	if (bond_is_lb(bond)) {
 		bond_alb_handle_active_change(bond, new_active);
@@ -1333,17 +1305,6 @@ static void bond_netpoll_cleanup(struct net_device *bond_dev)
 
 /*---------------------------------- IOCTL ----------------------------------*/
 
-static void bond_set_dev_addr(struct net_device *bond_dev,
-			      struct net_device *slave_dev)
-{
-	pr_debug("bond_dev=%p\n", bond_dev);
-	pr_debug("slave_dev=%p\n", slave_dev);
-	pr_debug("slave_dev->addr_len=%d\n", slave_dev->addr_len);
-	memcpy(bond_dev->dev_addr, slave_dev->dev_addr, slave_dev->addr_len);
-	bond_dev->addr_assign_type = NET_ADDR_SET;
-	call_netdevice_notifiers(NETDEV_CHANGEADDR, bond_dev);
-}
-
 static netdev_features_t bond_fix_features(struct net_device *dev,
 	netdev_features_t features)
 {
@@ -1425,8 +1386,6 @@ done:
 static void bond_setup_by_slave(struct net_device *bond_dev,
 				struct net_device *slave_dev)
 {
-	struct bonding *bond = netdev_priv(bond_dev);
-
 	bond_dev->header_ops	    = slave_dev->header_ops;
 
 	bond_dev->type		    = slave_dev->type;
@@ -1435,7 +1394,6 @@ static void bond_setup_by_slave(struct net_device *bond_dev,
 
 	memcpy(bond_dev->broadcast, slave_dev->broadcast,
 		slave_dev->addr_len);
-	bond->setup_by_slave = 1;
 }
 
 /* On bonding slaves other than the currently active slave, suppress
@@ -1533,10 +1491,9 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
 	struct bonding *bond = netdev_priv(bond_dev);
 	const struct net_device_ops *slave_ops = slave_dev->netdev_ops;
 	struct slave *new_slave = NULL;
-	struct netdev_hw_addr *ha;
 	struct sockaddr addr;
 	int link_reporting;
-	int res = 0;
+	int res = 0, i;
 
 	if (!bond->params.use_carrier &&
 	    slave_dev->ethtool_ops->get_link == NULL &&
@@ -1643,7 +1600,7 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
 
 	/* If this is the first slave, then we need to set the master's hardware
 	 * address to be the same as the slave's. */
-	if (bond->slave_cnt == 0 && bond->dev_addr_from_first)
+	if (!bond->slave_cnt && bond->dev->addr_assign_type == NET_ADDR_RANDOM)
 		bond_set_dev_addr(bond->dev, slave_dev);
 
 	new_slave = kzalloc(sizeof(struct slave), GFP_KERNEL);
@@ -1713,10 +1670,8 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
 			goto err_close;
 	}
 
-	/* If the mode USES_PRIMARY, then the new slave gets the
-	 * master's promisc (and mc) settings only if it becomes the
-	 * curr_active_slave, and that is taken care of later when calling
-	 * bond_change_active()
+	/* If the mode USES_PRIMARY, then the following is handled by
+	 * bond_change_active_slave().
 	 */
 	if (!USES_PRIMARY(bond->params.mode)) {
 		/* set promiscuity level to new slave */
@@ -1734,9 +1689,10 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
 		}
 
 		netif_addr_lock_bh(bond_dev);
-		/* upload master's mc_list to new slave */
-		netdev_for_each_mc_addr(ha, bond_dev)
-			dev_mc_add(slave_dev, ha->addr);
+
+		dev_mc_sync_multiple(slave_dev, bond_dev);
+		dev_uc_sync_multiple(slave_dev, bond_dev);
+
 		netif_addr_unlock_bh(bond_dev);
 	}
 
@@ -1766,6 +1722,8 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
 
 	new_slave->last_arp_rx = jiffies -
 		(msecs_to_jiffies(bond->params.arp_interval) + 1);
+	for (i = 0; i < BOND_MAX_ARP_TARGETS; i++)
+		new_slave->target_last_arp_rx[i] = new_slave->last_arp_rx;
 
 	if (bond->params.miimon && !bond->params.use_carrier) {
 		link_reporting = bond_check_dev_link(bond, slave_dev, 1);
@@ -1915,11 +1873,9 @@ err_dest_symlinks:
 	bond_destroy_slave_symlinks(bond_dev, slave_dev);
 
 err_detach:
-	if (!USES_PRIMARY(bond->params.mode)) {
-		netif_addr_lock_bh(bond_dev);
-		bond_mc_list_flush(bond_dev, slave_dev);
-		netif_addr_unlock_bh(bond_dev);
-	}
+	if (!USES_PRIMARY(bond->params.mode))
+		bond_hw_addr_flush(bond_dev, slave_dev);
+
 	bond_del_vlans_from_slave(bond, slave_dev);
 	write_lock_bh(&bond->lock);
 	bond_detach_slave(bond, new_slave);
@@ -2089,7 +2045,6 @@ static int __bond_release_one(struct net_device *bond_dev,
 	if (bond->slave_cnt == 0) {
 		bond_set_carrier(bond);
 		eth_hw_addr_random(bond_dev);
-		bond->dev_addr_from_first = true;
 
 		if (bond_vlan_used(bond)) {
 			pr_warning("%s: Warning: clearing HW address of %s while it still has VLANs.\n",
@@ -2118,9 +2073,8 @@ static int __bond_release_one(struct net_device *bond_dev,
 
 	bond_del_vlans_from_slave(bond, slave_dev);
 
-	/* If the mode USES_PRIMARY, then we should only remove its
-	 * promisc and mc settings if it was the curr_active_slave, but that was
-	 * already taken care of above when we detached the slave
+	/* If the mode USES_PRIMARY, then this cases was handled above by
+	 * bond_change_active_slave(..., NULL)
 	 */
 	if (!USES_PRIMARY(bond->params.mode)) {
 		/* unset promiscuity level from slave */
@@ -2131,10 +2085,7 @@ static int __bond_release_one(struct net_device *bond_dev,
 		if (bond_dev->flags & IFF_ALLMULTI)
 			dev_set_allmulti(slave_dev, -1);
 
-		/* flush master's mc_list from slave */
-		netif_addr_lock_bh(bond_dev);
-		bond_mc_list_flush(bond_dev, slave_dev);
-		netif_addr_unlock_bh(bond_dev);
+		bond_hw_addr_flush(bond_dev, slave_dev);
 	}
 
 	bond_upper_dev_unlink(bond_dev, slave_dev);
@@ -2672,18 +2623,19 @@ static void bond_arp_send_all(struct bonding *bond, struct slave *slave)
 static void bond_validate_arp(struct bonding *bond, struct slave *slave, __be32 sip, __be32 tip)
 {
 	int i;
-	__be32 *targets = bond->params.arp_targets;
 
-	for (i = 0; (i < BOND_MAX_ARP_TARGETS) && targets[i]; i++) {
-		pr_debug("bva: sip %pI4 tip %pI4 t[%d] %pI4 bhti(tip) %d\n",
-			 &sip, &tip, i, &targets[i],
-			 bond_has_this_ip(bond, tip));
-		if (sip == targets[i]) {
-			if (bond_has_this_ip(bond, tip))
-				slave->last_arp_rx = jiffies;
-			return;
-		}
+	if (!sip || !bond_has_this_ip(bond, tip)) {
+		pr_debug("bva: sip %pI4 tip %pI4 not found\n", &sip, &tip);
+		return;
+	}
+
+	i = bond_get_targets_ip(bond->params.arp_targets, sip);
+	if (i == -1) {
+		pr_debug("bva: sip %pI4 not found in targets\n", &sip);
+		return;
 	}
+	slave->last_arp_rx = jiffies;
+	slave->target_last_arp_rx[i] = jiffies;
 }
 
 static int bond_arp_rcv(const struct sk_buff *skb, struct bonding *bond,
@@ -2698,6 +2650,10 @@ static int bond_arp_rcv(const struct sk_buff *skb, struct bonding *bond,
 		return RX_HANDLER_ANOTHER;
 
 	read_lock(&bond->lock);
+
+	if (!slave_do_arp_validate(bond, slave))
+		goto out_unlock;
+
 	alen = arp_hdr_len(bond->dev);
 
 	pr_debug("bond_arp_rcv: bond %s skb->dev %s\n",
@@ -2737,10 +2693,17 @@ static int bond_arp_rcv(const struct sk_buff *skb, struct bonding *bond,
 	 * configuration, the ARP probe will (hopefully) travel from
 	 * the active, through one switch, the router, then the other
 	 * switch before reaching the backup.
+	 *
+	 * We 'trust' the arp requests if there is an active slave and
+	 * it received valid arp reply(s) after it became active. This
+	 * is done to avoid endless looping when we can't reach the
+	 * arp_ip_target and fool ourselves with our own arp requests.
 	 */
 	if (bond_is_active_slave(slave))
 		bond_validate_arp(bond, slave, sip, tip);
-	else
+	else if (bond->curr_active_slave &&
+		 time_after(slave_last_rx(bond, bond->curr_active_slave),
+			    bond->curr_active_slave->jiffies))
 		bond_validate_arp(bond, slave, tip, sip);
 
 out_unlock:
@@ -3225,7 +3188,7 @@ static int bond_slave_netdev_event(unsigned long event,
 
 	switch (event) {
 	case NETDEV_UNREGISTER:
-		if (bond->setup_by_slave)
+		if (bond_dev->type != ARPHRD_ETHER)
 			bond_release_and_destroy(bond_dev, slave_dev);
 		else
 			bond_release(bond_dev, slave_dev);
@@ -3289,7 +3252,7 @@ static int bond_slave_netdev_event(unsigned long event,
 static int bond_netdev_event(struct notifier_block *this,
 			     unsigned long event, void *ptr)
 {
-	struct net_device *event_dev = (struct net_device *)ptr;
+	struct net_device *event_dev = netdev_notifier_info_to_dev(ptr);
 
 	pr_debug("event_dev: %s, event: %lx\n",
 		 event_dev ? event_dev->name : "None",
@@ -3672,19 +3635,6 @@ static int bond_do_ioctl(struct net_device *bond_dev, struct ifreq *ifr, int cmd
 	return res;
 }
 
-static bool bond_addr_in_mc_list(unsigned char *addr,
-				 struct netdev_hw_addr_list *list,
-				 int addrlen)
-{
-	struct netdev_hw_addr *ha;
-
-	netdev_hw_addr_list_for_each(ha, list)
-		if (!memcmp(ha->addr, addr, addrlen))
-			return true;
-
-	return false;
-}
-
 static void bond_change_rx_flags(struct net_device *bond_dev, int change)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
@@ -3698,35 +3648,29 @@ static void bond_change_rx_flags(struct net_device *bond_dev, int change)
 				  bond_dev->flags & IFF_ALLMULTI ? 1 : -1);
 }
 
-static void bond_set_multicast_list(struct net_device *bond_dev)
+static void bond_set_rx_mode(struct net_device *bond_dev)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
-	struct netdev_hw_addr *ha;
-	bool found;
+	struct slave *slave;
+	int i;
 
 	read_lock(&bond->lock);
 
-	/* looking for addresses to add to slaves' mc list */
-	netdev_for_each_mc_addr(ha, bond_dev) {
-		found = bond_addr_in_mc_list(ha->addr, &bond->mc_list,
-					     bond_dev->addr_len);
-		if (!found)
-			bond_mc_add(bond, ha->addr);
-	}
-
-	/* looking for addresses to delete from slaves' list */
-	netdev_hw_addr_list_for_each(ha, &bond->mc_list) {
-		found = bond_addr_in_mc_list(ha->addr, &bond_dev->mc,
-					     bond_dev->addr_len);
-		if (!found)
-			bond_mc_del(bond, ha->addr);
+	if (USES_PRIMARY(bond->params.mode)) {
+		read_lock(&bond->curr_slave_lock);
+		slave = bond->curr_active_slave;
+		if (slave) {
+			dev_uc_sync(slave->dev, bond_dev);
+			dev_mc_sync(slave->dev, bond_dev);
+		}
+		read_unlock(&bond->curr_slave_lock);
+	} else {
+		bond_for_each_slave(bond, slave, i) {
+			dev_uc_sync_multiple(slave->dev, bond_dev);
+			dev_mc_sync_multiple(slave->dev, bond_dev);
+		}
 	}
 
-	/* save master's multicast list */
-	__hw_addr_flush(&bond->mc_list);
-	__hw_addr_add_multiple(&bond->mc_list, &bond_dev->mc,
-			       bond_dev->addr_len, NETDEV_HW_ADDR_T_MULTICAST);
-
 	read_unlock(&bond->lock);
 }
 
@@ -3871,11 +3815,10 @@ static int bond_set_mac_address(struct net_device *bond_dev, void *addr)
 	pr_debug("bond=%p, name=%s\n",
 		 bond, bond_dev ? bond_dev->name : "None");
 
-	/*
-	 * If fail_over_mac is set to active, do nothing and return
-	 * success.  Returning an error causes ifenslave to fail.
+	/* If fail_over_mac is enabled, do nothing and return success.
+	 * Returning an error causes ifenslave to fail.
 	 */
-	if (bond->params.fail_over_mac == BOND_FOM_ACTIVE)
+	if (bond->params.fail_over_mac)
 		return 0;
 
 	if (!is_valid_ether_addr(sa->sa_data))
@@ -4333,7 +4276,7 @@ static const struct net_device_ops bond_netdev_ops = {
 	.ndo_get_stats64	= bond_get_stats,
 	.ndo_do_ioctl		= bond_do_ioctl,
 	.ndo_change_rx_flags	= bond_change_rx_flags,
-	.ndo_set_rx_mode	= bond_set_multicast_list,
+	.ndo_set_rx_mode	= bond_set_rx_mode,
 	.ndo_change_mtu		= bond_change_mtu,
 	.ndo_set_mac_address	= bond_set_mac_address,
 	.ndo_neigh_setup	= bond_neigh_setup,
@@ -4438,8 +4381,6 @@ static void bond_uninit(struct net_device *bond_dev)
 
 	bond_debug_unregister(bond);
 
-	__hw_addr_flush(&bond->mc_list);
-
 	list_for_each_entry_safe(vlan, tmp, &bond->vlan_list, vlan_list) {
 		list_del(&vlan->vlan_list);
 		kfree(vlan);
@@ -4484,6 +4425,7 @@ int bond_parse_parm(const char *buf, const struct bond_parm_tbl *tbl)
 static int bond_check_params(struct bond_params *params)
 {
 	int arp_validate_value, fail_over_mac_value, primary_reselect_value, i;
+	int arp_all_targets_value;
 
 	/*
 	 * Convert string parameters.
@@ -4674,7 +4616,11 @@ static int bond_check_params(struct bond_params *params)
 				   arp_ip_target[i]);
 			arp_interval = 0;
 		} else {
-			arp_target[arp_ip_count++] = ip;
+			if (bond_get_targets_ip(arp_target, ip) == -1)
+				arp_target[arp_ip_count++] = ip;
+			else
+				pr_warning("Warning: duplicate address %pI4 in arp_ip_target, skipping\n",
+					   &ip);
 		}
 	}
 
@@ -4705,6 +4651,18 @@ static int bond_check_params(struct bond_params *params)
 	} else
 		arp_validate_value = 0;
 
+	arp_all_targets_value = 0;
+	if (arp_all_targets) {
+		arp_all_targets_value = bond_parse_parm(arp_all_targets,
+							arp_all_targets_tbl);
+
+		if (arp_all_targets_value == -1) {
+			pr_err("Error: invalid arp_all_targets_value \"%s\"\n",
+			       arp_all_targets);
+			arp_all_targets_value = 0;
+		}
+	}
+
 	if (miimon) {
 		pr_info("MII link monitoring set to %d ms\n", miimon);
 	} else if (arp_interval) {
@@ -4769,6 +4727,7 @@ static int bond_check_params(struct bond_params *params)
 	params->num_peer_notif = num_peer_notif;
 	params->arp_interval = arp_interval;
 	params->arp_validate = arp_validate_value;
+	params->arp_all_targets = arp_all_targets_value;
 	params->updelay = updelay;
 	params->downdelay = downdelay;
 	params->use_carrier = use_carrier;
@@ -4845,12 +4804,9 @@ static int bond_init(struct net_device *bond_dev)
 
 	/* Ensure valid dev_addr */
 	if (is_zero_ether_addr(bond_dev->dev_addr) &&
-	    bond_dev->addr_assign_type == NET_ADDR_PERM) {
+	    bond_dev->addr_assign_type == NET_ADDR_PERM)
 		eth_hw_addr_random(bond_dev);
-		bond->dev_addr_from_first = true;
-	}
 
-	__hw_addr_init(&bond->mc_list);
 	return 0;
 }
 
@@ -4923,7 +4879,7 @@ static int __net_init bond_net_init(struct net *net)
 
 	bond_create_proc_dir(bn);
 	bond_create_sysfs(bn);
-	
+
 	return 0;
 }
 
diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
index d7434e0..dc36a3d 100644
--- a/drivers/net/bonding/bond_sysfs.c
+++ b/drivers/net/bonding/bond_sysfs.c
@@ -231,8 +231,7 @@ static ssize_t bonding_show_slaves(struct device *d,
 }
 
 /*
- * Set the slaves in the current bond.  The bond interface must be
- * up for this to succeed.
+ * Set the slaves in the current bond.
  * This is supposed to be only thin wrapper for bond_enslave and bond_release.
  * All hard work should be done there.
  */
@@ -363,7 +362,6 @@ static DEVICE_ATTR(mode, S_IRUGO | S_IWUSR,
 
 /*
  * Show and set the bonding transmit hash method.
- * The bond interface must be down to change the xmit hash policy.
  */
 static ssize_t bonding_show_xmit_hash(struct device *d,
 				      struct device_attribute *attr,
@@ -383,20 +381,12 @@ static ssize_t bonding_store_xmit_hash(struct device *d,
 	int new_value, ret = count;
 	struct bonding *bond = to_bond(d);
 
-	if (bond->dev->flags & IFF_UP) {
-		pr_err("%s: Interface is up. Unable to update xmit policy.\n",
-		       bond->dev->name);
-		ret = -EPERM;
-		goto out;
-	}
-
 	new_value = bond_parse_parm(buf, xmit_hashtype_tbl);
 	if (new_value < 0)  {
 		pr_err("%s: Ignoring invalid xmit hash policy value %.*s.\n",
 		       bond->dev->name,
 		       (int)strlen(buf) - 1, buf);
 		ret = -EINVAL;
-		goto out;
 	} else {
 		bond->params.xmit_policy = new_value;
 		bond_set_mode_ops(bond, bond->params.mode);
@@ -404,7 +394,7 @@ static ssize_t bonding_store_xmit_hash(struct device *d,
 			bond->dev->name,
 			xmit_hashtype_tbl[new_value].modename, new_value);
 	}
-out:
+
 	return ret;
 }
 static DEVICE_ATTR(xmit_hash_policy, S_IRUGO | S_IWUSR,
@@ -453,6 +443,44 @@ static ssize_t bonding_store_arp_validate(struct device *d,
 
 static DEVICE_ATTR(arp_validate, S_IRUGO | S_IWUSR, bonding_show_arp_validate,
 		   bonding_store_arp_validate);
+/*
+ * Show and set arp_all_targets.
+ */
+static ssize_t bonding_show_arp_all_targets(struct device *d,
+					 struct device_attribute *attr,
+					 char *buf)
+{
+	struct bonding *bond = to_bond(d);
+	int value = bond->params.arp_all_targets;
+
+	return sprintf(buf, "%s %d\n", arp_all_targets_tbl[value].modename,
+		       value);
+}
+
+static ssize_t bonding_store_arp_all_targets(struct device *d,
+					  struct device_attribute *attr,
+					  const char *buf, size_t count)
+{
+	struct bonding *bond = to_bond(d);
+	int new_value;
+
+	new_value = bond_parse_parm(buf, arp_all_targets_tbl);
+	if (new_value < 0) {
+		pr_err("%s: Ignoring invalid arp_all_targets value %s\n",
+		       bond->dev->name, buf);
+		return -EINVAL;
+	}
+	pr_info("%s: setting arp_all_targets to %s (%d).\n",
+		bond->dev->name, arp_all_targets_tbl[new_value].modename,
+		new_value);
+
+	bond->params.arp_all_targets = new_value;
+
+	return count;
+}
+
+static DEVICE_ATTR(arp_all_targets, S_IRUGO | S_IWUSR,
+		   bonding_show_arp_all_targets, bonding_store_arp_all_targets);
 
 /*
  * Show and store fail_over_mac.  User only allowed to change the
@@ -600,10 +628,11 @@ static ssize_t bonding_store_arp_targets(struct device *d,
 					 struct device_attribute *attr,
 					 const char *buf, size_t count)
 {
-	__be32 newtarget;
-	int i = 0, done = 0, ret = count;
 	struct bonding *bond = to_bond(d);
-	__be32 *targets;
+	struct slave *slave;
+	__be32 newtarget, *targets;
+	unsigned long *targets_rx;
+	int ind, i, j, ret = -EINVAL;
 
 	targets = bond->params.arp_targets;
 	newtarget = in_aton(buf + 1);
@@ -612,57 +641,63 @@ static ssize_t bonding_store_arp_targets(struct device *d,
 		if ((newtarget == 0) || (newtarget == htonl(INADDR_BROADCAST))) {
 			pr_err("%s: invalid ARP target %pI4 specified for addition\n",
 			       bond->dev->name, &newtarget);
-			ret = -EINVAL;
 			goto out;
 		}
-		/* look for an empty slot to put the target in, and check for dupes */
-		for (i = 0; (i < BOND_MAX_ARP_TARGETS) && !done; i++) {
-			if (targets[i] == newtarget) { /* duplicate */
-				pr_err("%s: ARP target %pI4 is already present\n",
-				       bond->dev->name, &newtarget);
-				ret = -EINVAL;
-				goto out;
-			}
-			if (targets[i] == 0) {
-				pr_info("%s: adding ARP target %pI4.\n",
-					bond->dev->name, &newtarget);
-				done = 1;
-				targets[i] = newtarget;
-			}
+
+		if (bond_get_targets_ip(targets, newtarget) != -1) { /* dup */
+			pr_err("%s: ARP target %pI4 is already present\n",
+			       bond->dev->name, &newtarget);
+			goto out;
 		}
-		if (!done) {
+
+		ind = bond_get_targets_ip(targets, 0); /* first free slot */
+		if (ind == -1) {
 			pr_err("%s: ARP target table is full!\n",
 			       bond->dev->name);
-			ret = -EINVAL;
 			goto out;
 		}
 
+		pr_info("%s: adding ARP target %pI4.\n", bond->dev->name,
+			 &newtarget);
+		/* not to race with bond_arp_rcv */
+		write_lock_bh(&bond->lock);
+		bond_for_each_slave(bond, slave, i)
+			slave->target_last_arp_rx[ind] = jiffies;
+		targets[ind] = newtarget;
+		write_unlock_bh(&bond->lock);
 	} else if (buf[0] == '-')	{
 		if ((newtarget == 0) || (newtarget == htonl(INADDR_BROADCAST))) {
 			pr_err("%s: invalid ARP target %pI4 specified for removal\n",
 			       bond->dev->name, &newtarget);
-			ret = -EINVAL;
 			goto out;
 		}
 
-		for (i = 0; (i < BOND_MAX_ARP_TARGETS) && !done; i++) {
-			if (targets[i] == newtarget) {
-				int j;
-				pr_info("%s: removing ARP target %pI4.\n",
-					bond->dev->name, &newtarget);
-				for (j = i; (j < (BOND_MAX_ARP_TARGETS-1)) && targets[j+1]; j++)
-					targets[j] = targets[j+1];
-
-				targets[j] = 0;
-				done = 1;
-			}
-		}
-		if (!done) {
-			pr_info("%s: unable to remove nonexistent ARP target %pI4.\n",
+		ind = bond_get_targets_ip(targets, newtarget);
+		if (ind == -1) {
+			pr_err("%s: unable to remove nonexistent ARP target %pI4.\n",
 				bond->dev->name, &newtarget);
-			ret = -EINVAL;
 			goto out;
 		}
+
+		if (ind == 0 && !targets[1] && bond->params.arp_interval)
+			pr_warn("%s: removing last arp target with arp_interval on\n",
+				bond->dev->name);
+
+		pr_info("%s: removing ARP target %pI4.\n", bond->dev->name,
+			&newtarget);
+
+		write_lock_bh(&bond->lock);
+		bond_for_each_slave(bond, slave, i) {
+			targets_rx = slave->target_last_arp_rx;
+			j = ind;
+			for (; (j < BOND_MAX_ARP_TARGETS-1) && targets[j+1]; j++)
+				targets_rx[j] = targets_rx[j+1];
+			targets_rx[j] = 0;
+		}
+		for (i = ind; (i < BOND_MAX_ARP_TARGETS-1) && targets[i+1]; i++)
+			targets[i] = targets[i+1];
+		targets[i] = 0;
+		write_unlock_bh(&bond->lock);
 	} else {
 		pr_err("no command found in arp_ip_targets file for bond %s. Use +<addr> or -<addr>.\n",
 		       bond->dev->name);
@@ -670,6 +705,7 @@ static ssize_t bonding_store_arp_targets(struct device *d,
 		goto out;
 	}
 
+	ret = count;
 out:
 	return ret;
 }
@@ -1645,6 +1681,7 @@ static struct attribute *per_bond_attrs[] = {
 	&dev_attr_mode.attr,
 	&dev_attr_fail_over_mac.attr,
 	&dev_attr_arp_validate.attr,
+	&dev_attr_arp_all_targets.attr,
 	&dev_attr_arp_interval.attr,
 	&dev_attr_arp_ip_target.attr,
 	&dev_attr_downdelay.attr,
diff --git a/drivers/net/bonding/bonding.h b/drivers/net/bonding/bonding.h
index f989e15..42d1c65 100644
--- a/drivers/net/bonding/bonding.h
+++ b/drivers/net/bonding/bonding.h
@@ -144,6 +144,7 @@ struct bond_params {
 	u8 num_peer_notif;
 	int arp_interval;
 	int arp_validate;
+	int arp_all_targets;
 	int use_carrier;
 	int fail_over_mac;
 	int updelay;
@@ -179,6 +180,7 @@ struct slave {
 	int    delay;
 	unsigned long jiffies;
 	unsigned long last_arp_rx;
+	unsigned long target_last_arp_rx[BOND_MAX_ARP_TARGETS];
 	s8     link;    /* one of BOND_LINK_XXXX */
 	s8     new_link;
 	u8     backup:1,   /* indicates backup slave. Value corresponds with
@@ -224,14 +226,12 @@ struct bonding {
 	rwlock_t lock;
 	rwlock_t curr_slave_lock;
 	u8	 send_peer_notif;
-	s8	 setup_by_slave;
 	u8       igmp_retrans;
 #ifdef CONFIG_PROC_FS
 	struct   proc_dir_entry *proc_entry;
 	char     proc_file_name[IFNAMSIZ];
 #endif /* CONFIG_PROC_FS */
 	struct   list_head bond_list;
-	struct   netdev_hw_addr_list mc_list;
 	int      (*xmit_hash_policy)(struct sk_buff *, int);
 	u16      rr_tx_counter;
 	struct   ad_bond_info ad_info;
@@ -248,7 +248,6 @@ struct bonding {
 	/* debugging support via debugfs */
 	struct	 dentry *debug_dir;
 #endif /* CONFIG_DEBUG_FS */
-	bool	dev_addr_from_first;
 };
 
 static inline bool bond_vlan_used(struct bonding *bond)
@@ -323,6 +322,9 @@ static inline bool bond_is_active_slave(struct slave *slave)
 #define BOND_FOM_ACTIVE			1
 #define BOND_FOM_FOLLOW			2
 
+#define BOND_ARP_TARGETS_ANY		0
+#define BOND_ARP_TARGETS_ALL		1
+
 #define BOND_ARP_VALIDATE_NONE		0
 #define BOND_ARP_VALIDATE_ACTIVE	(1 << BOND_STATE_ACTIVE)
 #define BOND_ARP_VALIDATE_BACKUP	(1 << BOND_STATE_BACKUP)
@@ -335,11 +337,31 @@ static inline int slave_do_arp_validate(struct bonding *bond,
 	return bond->params.arp_validate & (1 << bond_slave_state(slave));
 }
 
+/* Get the oldest arp which we've received on this slave for bond's
+ * arp_targets.
+ */
+static inline unsigned long slave_oldest_target_arp_rx(struct bonding *bond,
+						       struct slave *slave)
+{
+	int i = 1;
+	unsigned long ret = slave->target_last_arp_rx[0];
+
+	for (; (i < BOND_MAX_ARP_TARGETS) && bond->params.arp_targets[i]; i++)
+		if (time_before(slave->target_last_arp_rx[i], ret))
+			ret = slave->target_last_arp_rx[i];
+
+	return ret;
+}
+
 static inline unsigned long slave_last_rx(struct bonding *bond,
 					struct slave *slave)
 {
-	if (slave_do_arp_validate(bond, slave))
-		return slave->last_arp_rx;
+	if (slave_do_arp_validate(bond, slave)) {
+		if (bond->params.arp_all_targets == BOND_ARP_TARGETS_ALL)
+			return slave_oldest_target_arp_rx(bond, slave);
+		else
+			return slave->last_arp_rx;
+	}
 
 	return slave->dev->last_rx;
 }
@@ -465,12 +487,29 @@ static inline struct slave *bond_slave_has_mac(struct bonding *bond,
 	return NULL;
 }
 
+/* Check if the ip is present in arp ip list, or first free slot if ip == 0
+ * Returns -1 if not found, index if found
+ */
+static inline int bond_get_targets_ip(__be32 *targets, __be32 ip)
+{
+	int i;
+
+	for (i = 0; i < BOND_MAX_ARP_TARGETS; i++)
+		if (targets[i] == ip)
+			return i;
+		else if (targets[i] == 0)
+			break;
+
+	return -1;
+}
+
 /* exported from bond_main.c */
 extern int bond_net_id;
 extern const struct bond_parm_tbl bond_lacp_tbl[];
 extern const struct bond_parm_tbl bond_mode_tbl[];
 extern const struct bond_parm_tbl xmit_hashtype_tbl[];
 extern const struct bond_parm_tbl arp_validate_tbl[];
+extern const struct bond_parm_tbl arp_all_targets_tbl[];
 extern const struct bond_parm_tbl fail_over_mac_tbl[];
 extern const struct bond_parm_tbl pri_reselect_tbl[];
 extern struct bond_parm_tbl ad_select_tbl[];
diff --git a/drivers/net/caif/caif_serial.c b/drivers/net/caif/caif_serial.c
index 77be3cb..34dea95 100644
--- a/drivers/net/caif/caif_serial.c
+++ b/drivers/net/caif/caif_serial.c
@@ -35,8 +35,9 @@ MODULE_ALIAS_LDISC(N_CAIF);
 #define OFF 0
 #define CAIF_MAX_MTU 4096
 
-/*This list is protected by the rtnl lock. */
+static DEFINE_SPINLOCK(ser_lock);
 static LIST_HEAD(ser_list);
+static LIST_HEAD(ser_release_list);
 
 static bool ser_loop;
 module_param(ser_loop, bool, S_IRUGO);
@@ -308,6 +309,28 @@ static void ldisc_tx_wakeup(struct tty_struct *tty)
 }
 
 
+static void ser_release(struct work_struct *work)
+{
+	struct list_head list;
+	struct ser_device *ser, *tmp;
+
+	spin_lock(&ser_lock);
+	list_replace_init(&ser_release_list, &list);
+	spin_unlock(&ser_lock);
+
+	if (!list_empty(&list)) {
+		rtnl_lock();
+		list_for_each_entry_safe(ser, tmp, &list, node) {
+			dev_close(ser->dev);
+			unregister_netdevice(ser->dev);
+			debugfs_deinit(ser);
+		}
+		rtnl_unlock();
+	}
+}
+
+static DECLARE_WORK(ser_release_work, ser_release);
+
 static int ldisc_open(struct tty_struct *tty)
 {
 	struct ser_device *ser;
@@ -321,6 +344,9 @@ static int ldisc_open(struct tty_struct *tty)
 	if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_TTY_CONFIG))
 		return -EPERM;
 
+	/* release devices to avoid name collision */
+	ser_release(NULL);
+
 	sprintf(name, "cf%s", tty->name);
 	dev = alloc_netdev(sizeof(*ser), name, caifdev_setup);
 	if (!dev)
@@ -341,7 +367,9 @@ static int ldisc_open(struct tty_struct *tty)
 		return -ENODEV;
 	}
 
+	spin_lock(&ser_lock);
 	list_add(&ser->node, &ser_list);
+	spin_unlock(&ser_lock);
 	rtnl_unlock();
 	netif_stop_queue(dev);
 	update_tty_status(ser);
@@ -351,19 +379,13 @@ static int ldisc_open(struct tty_struct *tty)
 static void ldisc_close(struct tty_struct *tty)
 {
 	struct ser_device *ser = tty->disc_data;
-	/* Remove may be called inside or outside of rtnl_lock */
-	int islocked = rtnl_is_locked();
 
-	if (!islocked)
-		rtnl_lock();
-	/* device is freed automagically by net-sysfs */
-	dev_close(ser->dev);
-	unregister_netdevice(ser->dev);
-	list_del(&ser->node);
-	debugfs_deinit(ser);
 	tty_kref_put(ser->tty);
-	if (!islocked)
-		rtnl_unlock();
+
+	spin_lock(&ser_lock);
+	list_move(&ser->node, &ser_release_list);
+	spin_unlock(&ser_lock);
+	schedule_work(&ser_release_work);
 }
 
 /* The line discipline structure. */
@@ -438,16 +460,11 @@ static int __init caif_ser_init(void)
 
 static void __exit caif_ser_exit(void)
 {
-	struct ser_device *ser = NULL;
-	struct list_head *node;
-	struct list_head *_tmp;
-
-	list_for_each_safe(node, _tmp, &ser_list) {
-		ser = list_entry(node, struct ser_device, node);
-		dev_close(ser->dev);
-		unregister_netdevice(ser->dev);
-		list_del(node);
-	}
+	spin_lock(&ser_lock);
+	list_splice(&ser_list, &ser_release_list);
+	spin_unlock(&ser_lock);
+	ser_release(NULL);
+	cancel_work_sync(&ser_release_work);
 	tty_unregister_ldisc(N_CAIF);
 	debugfs_remove_recursive(debugfsdir);
 }
diff --git a/drivers/net/can/Kconfig b/drivers/net/can/Kconfig
index e456b70..3c06947 100644
--- a/drivers/net/can/Kconfig
+++ b/drivers/net/can/Kconfig
@@ -102,12 +102,9 @@ config CAN_JANZ_ICAN3
 	  This driver can also be built as a module. If so, the module will be
 	  called janz-ican3.ko.
 
-config HAVE_CAN_FLEXCAN
-	bool
-
 config CAN_FLEXCAN
 	tristate "Support for Freescale FLEXCAN based chips"
-	depends on HAVE_CAN_FLEXCAN
+	depends on ARM || PPC
 	---help---
 	  Say Y here if you want to support for Freescale FlexCAN.
 
diff --git a/drivers/net/can/at91_can.c b/drivers/net/can/at91_can.c
index db52f441..dbbe97a 100644
--- a/drivers/net/can/at91_can.c
+++ b/drivers/net/can/at91_can.c
@@ -1220,7 +1220,7 @@ static ssize_t at91_sysfs_set_mb0_id(struct device *dev,
 		goto out;
 	}
 
-	err = strict_strtoul(buf, 0, &can_id);
+	err = kstrtoul(buf, 0, &can_id);
 	if (err) {
 		ret = err;
 		goto out;
@@ -1264,8 +1264,6 @@ static const struct of_device_id at91_can_dt_ids[] = {
 	}
 };
 MODULE_DEVICE_TABLE(of, at91_can_dt_ids);
-#else
-#define at91_can_dt_ids NULL
 #endif
 
 static const struct at91_devtype_data *at91_can_get_driver_data(struct platform_device *pdev)
@@ -1393,8 +1391,6 @@ static int at91_can_remove(struct platform_device *pdev)
 
 	unregister_netdev(dev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	iounmap(priv->reg_base);
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -1426,7 +1422,7 @@ static struct platform_driver at91_can_driver = {
 	.driver = {
 		.name = KBUILD_MODNAME,
 		.owner = THIS_MODULE,
-		.of_match_table = at91_can_dt_ids,
+		.of_match_table = of_match_ptr(at91_can_dt_ids),
 	},
 	.id_table = at91_can_id_table,
 };
diff --git a/drivers/net/can/bfin_can.c b/drivers/net/can/bfin_can.c
index d4a15e8..a2700d2 100644
--- a/drivers/net/can/bfin_can.c
+++ b/drivers/net/can/bfin_can.c
@@ -580,7 +580,7 @@ static int bfin_can_probe(struct platform_device *pdev)
 	priv->pin_list = pdata;
 	priv->can.clock.freq = get_sclk();
 
-	dev_set_drvdata(&pdev->dev, dev);
+	platform_set_drvdata(pdev, dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
 	dev->flags |= IFF_ECHO;	/* we support local echo */
@@ -613,7 +613,7 @@ exit:
 
 static int bfin_can_remove(struct platform_device *pdev)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct bfin_can_priv *priv = netdev_priv(dev);
 	struct resource *res;
 
@@ -621,8 +621,6 @@ static int bfin_can_remove(struct platform_device *pdev)
 
 	unregister_candev(dev);
 
-	dev_set_drvdata(&pdev->dev, NULL);
-
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	release_mem_region(res->start, resource_size(res));
 
@@ -635,7 +633,7 @@ static int bfin_can_remove(struct platform_device *pdev)
 #ifdef CONFIG_PM
 static int bfin_can_suspend(struct platform_device *pdev, pm_message_t mesg)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct bfin_can_priv *priv = netdev_priv(dev);
 	struct bfin_can_regs __iomem *reg = priv->membase;
 	int timeout = BFIN_CAN_TIMEOUT;
@@ -658,7 +656,7 @@ static int bfin_can_suspend(struct platform_device *pdev, pm_message_t mesg)
 
 static int bfin_can_resume(struct platform_device *pdev)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct bfin_can_priv *priv = netdev_priv(dev);
 	struct bfin_can_regs __iomem *reg = priv->membase;
 
diff --git a/drivers/net/can/c_can/c_can_platform.c b/drivers/net/can/c_can/c_can_platform.c
index d63b919..b918c73 100644
--- a/drivers/net/can/c_can/c_can_platform.c
+++ b/drivers/net/can/c_can/c_can_platform.c
@@ -201,8 +201,8 @@ static int c_can_plat_probe(struct platform_device *pdev)
 			priv->instance = pdev->id;
 
 		res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
-		priv->raminit_ctrlreg =	devm_request_and_ioremap(&pdev->dev, res);
-		if (!priv->raminit_ctrlreg || priv->instance < 0)
+		priv->raminit_ctrlreg = devm_ioremap_resource(&pdev->dev, res);
+		if (IS_ERR(priv->raminit_ctrlreg) || priv->instance < 0)
 			dev_info(&pdev->dev, "control memory is not used for raminit\n");
 		else
 			priv->raminit = c_can_hw_raminit;
@@ -234,7 +234,6 @@ static int c_can_plat_probe(struct platform_device *pdev)
 	return 0;
 
 exit_free_device:
-	platform_set_drvdata(pdev, NULL);
 	free_c_can_dev(dev);
 exit_iounmap:
 	iounmap(addr);
@@ -255,7 +254,6 @@ static int c_can_plat_remove(struct platform_device *pdev)
 	struct resource *mem;
 
 	unregister_c_can_dev(dev);
-	platform_set_drvdata(pdev, NULL);
 
 	free_c_can_dev(dev);
 	iounmap(priv->base);
diff --git a/drivers/net/can/cc770/cc770_isa.c b/drivers/net/can/cc770/cc770_isa.c
index 8eaaac8..87a47c0 100644
--- a/drivers/net/can/cc770/cc770_isa.c
+++ b/drivers/net/can/cc770/cc770_isa.c
@@ -265,7 +265,7 @@ static int cc770_isa_probe(struct platform_device *pdev)
 	else
 		priv->clkout = COR_DEFAULT;
 
-	dev_set_drvdata(&pdev->dev, dev);
+	platform_set_drvdata(pdev, dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
 	err = register_cc770dev(dev);
@@ -293,12 +293,11 @@ static int cc770_isa_probe(struct platform_device *pdev)
 
 static int cc770_isa_remove(struct platform_device *pdev)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct cc770_priv *priv = netdev_priv(dev);
 	int idx = pdev->id;
 
 	unregister_cc770dev(dev);
-	dev_set_drvdata(&pdev->dev, NULL);
 
 	if (mem[idx]) {
 		iounmap(priv->reg_base);
diff --git a/drivers/net/can/cc770/cc770_platform.c b/drivers/net/can/cc770/cc770_platform.c
index d0f6bfc..034bdd8 100644
--- a/drivers/net/can/cc770/cc770_platform.c
+++ b/drivers/net/can/cc770/cc770_platform.c
@@ -216,7 +216,7 @@ static int cc770_platform_probe(struct platform_device *pdev)
 		 priv->reg_base, dev->irq, priv->can.clock.freq,
 		 priv->cpu_interface, priv->bus_config, priv->clkout);
 
-	dev_set_drvdata(&pdev->dev, dev);
+	platform_set_drvdata(pdev, dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
 	err = register_cc770dev(dev);
@@ -240,7 +240,7 @@ exit_release_mem:
 
 static int cc770_platform_remove(struct platform_device *pdev)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct cc770_priv *priv = netdev_priv(dev);
 	struct resource *mem;
 
diff --git a/drivers/net/can/flexcan.c b/drivers/net/can/flexcan.c
index 769d29e..7b0be09 100644
--- a/drivers/net/can/flexcan.c
+++ b/drivers/net/can/flexcan.c
@@ -24,7 +24,6 @@
 #include <linux/can/dev.h>
 #include <linux/can/error.h>
 #include <linux/can/led.h>
-#include <linux/can/platform/flexcan.h>
 #include <linux/clk.h>
 #include <linux/delay.h>
 #include <linux/if_arp.h>
@@ -37,7 +36,7 @@
 #include <linux/of.h>
 #include <linux/of_device.h>
 #include <linux/platform_device.h>
-#include <linux/pinctrl/consumer.h>
+#include <linux/regulator/consumer.h>
 
 #define DRV_NAME			"flexcan"
 
@@ -212,6 +211,7 @@ struct flexcan_priv {
 	struct clk *clk_per;
 	struct flexcan_platform_data *pdata;
 	const struct flexcan_devtype_data *devtype_data;
+	struct regulator *reg_xceiver;
 };
 
 static struct flexcan_devtype_data fsl_p1010_devtype_data = {
@@ -259,15 +259,6 @@ static inline void flexcan_write(u32 val, void __iomem *addr)
 }
 #endif
 
-/*
- * Swtich transceiver on or off
- */
-static void flexcan_transceiver_switch(const struct flexcan_priv *priv, int on)
-{
-	if (priv->pdata && priv->pdata->transceiver_switch)
-		priv->pdata->transceiver_switch(on);
-}
-
 static inline int flexcan_has_and_handle_berr(const struct flexcan_priv *priv,
 					      u32 reg_esr)
 {
@@ -800,7 +791,11 @@ static int flexcan_chip_start(struct net_device *dev)
 	if (priv->devtype_data->features & FLEXCAN_HAS_V10_FEATURES)
 		flexcan_write(0x0, &regs->rxfgmask);
 
-	flexcan_transceiver_switch(priv, 1);
+	if (priv->reg_xceiver)	{
+		err = regulator_enable(priv->reg_xceiver);
+		if (err)
+			goto out;
+	}
 
 	/* synchronize with the can bus */
 	reg_mcr = flexcan_read(&regs->mcr);
@@ -843,7 +838,8 @@ static void flexcan_chip_stop(struct net_device *dev)
 	reg |= FLEXCAN_MCR_MDIS | FLEXCAN_MCR_HALT;
 	flexcan_write(reg, &regs->mcr);
 
-	flexcan_transceiver_switch(priv, 0);
+	if (priv->reg_xceiver)
+		regulator_disable(priv->reg_xceiver);
 	priv->can.state = CAN_STATE_STOPPED;
 
 	return;
@@ -1004,16 +1000,11 @@ static int flexcan_probe(struct platform_device *pdev)
 	struct flexcan_priv *priv;
 	struct resource *mem;
 	struct clk *clk_ipg = NULL, *clk_per = NULL;
-	struct pinctrl *pinctrl;
 	void __iomem *base;
 	resource_size_t mem_size;
 	int err, irq;
 	u32 clock_freq = 0;
 
-	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
-	if (IS_ERR(pinctrl))
-		return PTR_ERR(pinctrl);
-
 	if (pdev->dev.of_node)
 		of_property_read_u32(pdev->dev.of_node,
 						"clock-frequency", &clock_freq);
@@ -1090,6 +1081,10 @@ static int flexcan_probe(struct platform_device *pdev)
 	priv->pdata = pdev->dev.platform_data;
 	priv->devtype_data = devtype_data;
 
+	priv->reg_xceiver = devm_regulator_get(&pdev->dev, "xceiver");
+	if (IS_ERR(priv->reg_xceiver))
+		priv->reg_xceiver = NULL;
+
 	netif_napi_add(dev, &priv->napi, flexcan_poll, FLEXCAN_NAPI_WEIGHT);
 
 	dev_set_drvdata(&pdev->dev, dev);
@@ -1127,7 +1122,6 @@ static int flexcan_remove(struct platform_device *pdev)
 	struct resource *mem;
 
 	unregister_flexcandev(dev);
-	platform_set_drvdata(pdev, NULL);
 	iounmap(priv->base);
 
 	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -1138,10 +1132,10 @@ static int flexcan_remove(struct platform_device *pdev)
 	return 0;
 }
 
-#ifdef CONFIG_PM
-static int flexcan_suspend(struct platform_device *pdev, pm_message_t state)
+#ifdef CONFIG_PM_SLEEP
+static int flexcan_suspend(struct device *device)
 {
-	struct net_device *dev = platform_get_drvdata(pdev);
+	struct net_device *dev = dev_get_drvdata(device);
 	struct flexcan_priv *priv = netdev_priv(dev);
 
 	flexcan_chip_disable(priv);
@@ -1155,9 +1149,9 @@ static int flexcan_suspend(struct platform_device *pdev, pm_message_t state)
 	return 0;
 }
 
-static int flexcan_resume(struct platform_device *pdev)
+static int flexcan_resume(struct device *device)
 {
-	struct net_device *dev = platform_get_drvdata(pdev);
+	struct net_device *dev = dev_get_drvdata(device);
 	struct flexcan_priv *priv = netdev_priv(dev);
 
 	priv->can.state = CAN_STATE_ERROR_ACTIVE;
@@ -1169,21 +1163,19 @@ static int flexcan_resume(struct platform_device *pdev)
 
 	return 0;
 }
-#else
-#define flexcan_suspend NULL
-#define flexcan_resume NULL
-#endif
+#endif /* CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(flexcan_pm_ops, flexcan_suspend, flexcan_resume);
 
 static struct platform_driver flexcan_driver = {
 	.driver = {
 		.name = DRV_NAME,
 		.owner = THIS_MODULE,
+		.pm = &flexcan_pm_ops,
 		.of_match_table = flexcan_of_match,
 	},
 	.probe = flexcan_probe,
 	.remove = flexcan_remove,
-	.suspend = flexcan_suspend,
-	.resume = flexcan_resume,
 	.id_table = flexcan_id_table,
 };
 
diff --git a/drivers/net/can/grcan.c b/drivers/net/can/grcan.c
index 17fbc7a..6aa737a 100644
--- a/drivers/net/can/grcan.c
+++ b/drivers/net/can/grcan.c
@@ -1646,7 +1646,7 @@ static int grcan_setup_netdev(struct platform_device *ofdev,
 	if (err)
 		goto exit_free_candev;
 
-	dev_set_drvdata(&ofdev->dev, dev);
+	platform_set_drvdata(ofdev, dev);
 
 	/* Reset device to allow bit-timing to be set. No need to call
 	 * grcan_reset at this stage. That is done in grcan_open.
@@ -1683,10 +1683,9 @@ static int grcan_probe(struct platform_device *ofdev)
 	}
 
 	res = platform_get_resource(ofdev, IORESOURCE_MEM, 0);
-	base = devm_request_and_ioremap(&ofdev->dev, res);
-	if (!base) {
-		dev_err(&ofdev->dev, "couldn't map IO resource\n");
-		err = -EADDRNOTAVAIL;
+	base = devm_ioremap_resource(&ofdev->dev, res);
+	if (IS_ERR(base)) {
+		err = PTR_ERR(base);
 		goto exit_error;
 	}
 
@@ -1716,13 +1715,12 @@ exit_error:
 
 static int grcan_remove(struct platform_device *ofdev)
 {
-	struct net_device *dev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *dev = platform_get_drvdata(ofdev);
 	struct grcan_priv *priv = netdev_priv(dev);
 
 	unregister_candev(dev); /* Will in turn call grcan_close */
 
 	irq_dispose_mapping(dev->irq);
-	dev_set_drvdata(&ofdev->dev, NULL);
 	netif_napi_del(&priv->napi);
 	free_candev(dev);
 
diff --git a/drivers/net/can/janz-ican3.c b/drivers/net/can/janz-ican3.c
index c4bc1d2..36bd6fa 100644
--- a/drivers/net/can/janz-ican3.c
+++ b/drivers/net/can/janz-ican3.c
@@ -1734,7 +1734,7 @@ static ssize_t ican3_sysfs_set_term(struct device *dev,
 	unsigned long enable;
 	int ret;
 
-	if (strict_strtoul(buf, 0, &enable))
+	if (kstrtoul(buf, 0, &enable))
 		return -EINVAL;
 
 	ret = ican3_set_termination(mod, enable);
diff --git a/drivers/net/can/led.c b/drivers/net/can/led.c
index f27fca6..a3d99a8 100644
--- a/drivers/net/can/led.c
+++ b/drivers/net/can/led.c
@@ -88,9 +88,9 @@ EXPORT_SYMBOL_GPL(devm_can_led_init);
 
 /* NETDEV rename notifier to rename the associated led triggers too */
 static int can_led_notifier(struct notifier_block *nb, unsigned long msg,
-			void *data)
+			    void *ptr)
 {
-	struct net_device *netdev = data;
+	struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
 	struct can_priv *priv = safe_candev_priv(netdev);
 	char name[CAN_LED_NAME_SZ];
 
diff --git a/drivers/net/can/mscan/mpc5xxx_can.c b/drivers/net/can/mscan/mpc5xxx_can.c
index 668850e..5b0ee8e 100644
--- a/drivers/net/can/mscan/mpc5xxx_can.c
+++ b/drivers/net/can/mscan/mpc5xxx_can.c
@@ -302,7 +302,7 @@ static int mpc5xxx_can_probe(struct platform_device *ofdev)
 		goto exit_free_mscan;
 	}
 
-	dev_set_drvdata(&ofdev->dev, dev);
+	platform_set_drvdata(ofdev, dev);
 
 	dev_info(&ofdev->dev, "MSCAN at 0x%p, irq %d, clock %d Hz\n",
 		 priv->reg_base, dev->irq, priv->can.clock.freq);
@@ -321,11 +321,9 @@ exit_unmap_mem:
 
 static int mpc5xxx_can_remove(struct platform_device *ofdev)
 {
-	struct net_device *dev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *dev = platform_get_drvdata(ofdev);
 	struct mscan_priv *priv = netdev_priv(dev);
 
-	dev_set_drvdata(&ofdev->dev, NULL);
-
 	unregister_mscandev(dev);
 	iounmap(priv->reg_base);
 	irq_dispose_mapping(dev->irq);
@@ -338,7 +336,7 @@ static int mpc5xxx_can_remove(struct platform_device *ofdev)
 static struct mscan_regs saved_regs;
 static int mpc5xxx_can_suspend(struct platform_device *ofdev, pm_message_t state)
 {
-	struct net_device *dev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *dev = platform_get_drvdata(ofdev);
 	struct mscan_priv *priv = netdev_priv(dev);
 	struct mscan_regs *regs = (struct mscan_regs *)priv->reg_base;
 
@@ -349,7 +347,7 @@ static int mpc5xxx_can_suspend(struct platform_device *ofdev, pm_message_t state
 
 static int mpc5xxx_can_resume(struct platform_device *ofdev)
 {
-	struct net_device *dev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *dev = platform_get_drvdata(ofdev);
 	struct mscan_priv *priv = netdev_priv(dev);
 	struct mscan_regs *regs = (struct mscan_regs *)priv->reg_base;
 
diff --git a/drivers/net/can/sja1000/sja1000_isa.c b/drivers/net/can/sja1000/sja1000_isa.c
index 5c8da46..06a2823 100644
--- a/drivers/net/can/sja1000/sja1000_isa.c
+++ b/drivers/net/can/sja1000/sja1000_isa.c
@@ -197,7 +197,7 @@ static int sja1000_isa_probe(struct platform_device *pdev)
 	else
 		priv->cdr = CDR_DEFAULT;
 
-	dev_set_drvdata(&pdev->dev, dev);
+	platform_set_drvdata(pdev, dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
 	err = register_sja1000dev(dev);
@@ -225,12 +225,11 @@ static int sja1000_isa_probe(struct platform_device *pdev)
 
 static int sja1000_isa_remove(struct platform_device *pdev)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct sja1000_priv *priv = netdev_priv(dev);
 	int idx = pdev->id;
 
 	unregister_sja1000dev(dev);
-	dev_set_drvdata(&pdev->dev, NULL);
 
 	if (mem[idx]) {
 		iounmap(priv->reg_base);
diff --git a/drivers/net/can/sja1000/sja1000_of_platform.c b/drivers/net/can/sja1000/sja1000_of_platform.c
index 8e0c4a0..31ad339 100644
--- a/drivers/net/can/sja1000/sja1000_of_platform.c
+++ b/drivers/net/can/sja1000/sja1000_of_platform.c
@@ -72,13 +72,11 @@ static void sja1000_ofp_write_reg(const struct sja1000_priv *priv,
 
 static int sja1000_ofp_remove(struct platform_device *ofdev)
 {
-	struct net_device *dev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *dev = platform_get_drvdata(ofdev);
 	struct sja1000_priv *priv = netdev_priv(dev);
 	struct device_node *np = ofdev->dev.of_node;
 	struct resource res;
 
-	dev_set_drvdata(&ofdev->dev, NULL);
-
 	unregister_sja1000dev(dev);
 	free_sja1000dev(dev);
 	iounmap(priv->reg_base);
@@ -181,7 +179,7 @@ static int sja1000_ofp_probe(struct platform_device *ofdev)
 		 priv->reg_base, dev->irq, priv->can.clock.freq,
 		 priv->ocr, priv->cdr);
 
-	dev_set_drvdata(&ofdev->dev, dev);
+	platform_set_drvdata(ofdev, dev);
 	SET_NETDEV_DEV(dev, &ofdev->dev);
 
 	err = register_sja1000dev(dev);
diff --git a/drivers/net/can/sja1000/sja1000_platform.c b/drivers/net/can/sja1000/sja1000_platform.c
index 21619bb..8e259c5 100644
--- a/drivers/net/can/sja1000/sja1000_platform.c
+++ b/drivers/net/can/sja1000/sja1000_platform.c
@@ -135,7 +135,7 @@ static int sp_probe(struct platform_device *pdev)
 		break;
 	}
 
-	dev_set_drvdata(&pdev->dev, dev);
+	platform_set_drvdata(pdev, dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 
 	err = register_sja1000dev(dev);
@@ -161,12 +161,11 @@ static int sp_probe(struct platform_device *pdev)
 
 static int sp_remove(struct platform_device *pdev)
 {
-	struct net_device *dev = dev_get_drvdata(&pdev->dev);
+	struct net_device *dev = platform_get_drvdata(pdev);
 	struct sja1000_priv *priv = netdev_priv(dev);
 	struct resource *res;
 
 	unregister_sja1000dev(dev);
-	dev_set_drvdata(&pdev->dev, NULL);
 
 	if (priv->reg_base)
 		iounmap(priv->reg_base);
diff --git a/drivers/net/can/slcan.c b/drivers/net/can/slcan.c
index 06b7e09..874188b 100644
--- a/drivers/net/can/slcan.c
+++ b/drivers/net/can/slcan.c
@@ -161,7 +161,7 @@ static void slc_bump(struct slcan *sl)
 
 	sl->rbuff[dlc_pos] = 0; /* terminate can_id string */
 
-	if (strict_strtoul(sl->rbuff+1, 16, &ultmp))
+	if (kstrtoul(sl->rbuff+1, 16, &ultmp))
 		return;
 
 	cf.can_id = ultmp;
diff --git a/drivers/net/can/softing/softing_main.c b/drivers/net/can/softing/softing_main.c
index 3a2b456..65eef1e 100644
--- a/drivers/net/can/softing/softing_main.c
+++ b/drivers/net/can/softing/softing_main.c
@@ -594,7 +594,7 @@ static ssize_t store_output(struct device *dev, struct device_attribute *attr,
 	unsigned long val;
 	int ret;
 
-	ret = strict_strtoul(buf, 0, &val);
+	ret = kstrtoul(buf, 0, &val);
 	if (ret < 0)
 		return ret;
 	val &= 0xFF;
diff --git a/drivers/net/can/ti_hecc.c b/drivers/net/can/ti_hecc.c
index f21fc37..3a349a2 100644
--- a/drivers/net/can/ti_hecc.c
+++ b/drivers/net/can/ti_hecc.c
@@ -1001,7 +1001,6 @@ static int ti_hecc_remove(struct platform_device *pdev)
 	iounmap(priv->base);
 	release_mem_region(res->start, resource_size(res));
 	free_candev(ndev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/3com/3c509.c b/drivers/net/ethernet/3com/3c509.c
index adb4bf5..ede8daa 100644
--- a/drivers/net/ethernet/3com/3c509.c
+++ b/drivers/net/ethernet/3com/3c509.c
@@ -723,25 +723,6 @@ el3_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		pr_debug("%s: el3_start_xmit(length = %u) called, status %4.4x.\n",
 			   dev->name, skb->len, inw(ioaddr + EL3_STATUS));
 	}
-#if 0
-#ifndef final_version
-	{	/* Error-checking code, delete someday. */
-		ushort status = inw(ioaddr + EL3_STATUS);
-		if (status & 0x0001 && 		/* IRQ line active, missed one. */
-		    inw(ioaddr + EL3_STATUS) & 1) { 			/* Make sure. */
-			pr_debug("%s: Missed interrupt, status then %04x now %04x"
-				   "  Tx %2.2x Rx %4.4x.\n", dev->name, status,
-				   inw(ioaddr + EL3_STATUS), inb(ioaddr + TX_STATUS),
-				   inw(ioaddr + RX_STATUS));
-			/* Fake interrupt trigger by masking, acknowledge interrupts. */
-			outw(SetStatusEnb | 0x00, ioaddr + EL3_CMD);
-			outw(AckIntr | IntLatch | TxAvailable | RxEarly | IntReq,
-				 ioaddr + EL3_CMD);
-			outw(SetStatusEnb | 0xff, ioaddr + EL3_CMD);
-		}
-	}
-#endif
-#endif
 	/*
 	 *	We lock the driver against other processors. Note
 	 *	we don't need to lock versus the IRQ as we suspended
diff --git a/drivers/net/ethernet/3com/3c59x.c b/drivers/net/ethernet/3com/3c59x.c
index 072c6f1..ad5272b 100644
--- a/drivers/net/ethernet/3com/3c59x.c
+++ b/drivers/net/ethernet/3com/3c59x.c
@@ -1012,10 +1012,8 @@ static int vortex_init_one(struct pci_dev *pdev,
 		goto out;
 
 	rc = pci_request_regions(pdev, DRV_NAME);
-	if (rc < 0) {
-		pci_disable_device(pdev);
-		goto out;
-	}
+	if (rc < 0)
+		goto out_disable;
 
 	unit = vortex_cards_found;
 
@@ -1032,23 +1030,24 @@ static int vortex_init_one(struct pci_dev *pdev,
 	if (!ioaddr) /* If mapping fails, fall-back to BAR 0... */
 		ioaddr = pci_iomap(pdev, 0, 0);
 	if (!ioaddr) {
-		pci_release_regions(pdev);
-		pci_disable_device(pdev);
 		rc = -ENOMEM;
-		goto out;
+		goto out_release;
 	}
 
 	rc = vortex_probe1(&pdev->dev, ioaddr, pdev->irq,
 			   ent->driver_data, unit);
-	if (rc < 0) {
-		pci_iounmap(pdev, ioaddr);
-		pci_release_regions(pdev);
-		pci_disable_device(pdev);
-		goto out;
-	}
+	if (rc < 0)
+		goto out_iounmap;
 
 	vortex_cards_found++;
+	goto out;
 
+out_iounmap:
+	pci_iounmap(pdev, ioaddr);
+out_release:
+	pci_release_regions(pdev);
+out_disable:
+	pci_disable_device(pdev);
 out:
 	return rc;
 }
@@ -1473,7 +1472,7 @@ static int vortex_probe1(struct device *gendev, void __iomem *ioaddr, int irq,
 
 	if (pdev) {
 		vp->pm_state_valid = 1;
- 		pci_save_state(VORTEX_PCI(vp));
+		pci_save_state(pdev);
  		acpi_set_WOL(dev);
 	}
 	retval = register_netdev(dev);
@@ -3233,21 +3232,20 @@ static void vortex_remove_one(struct pci_dev *pdev)
 	vp = netdev_priv(dev);
 
 	if (vp->cb_fn_base)
-		pci_iounmap(VORTEX_PCI(vp), vp->cb_fn_base);
+		pci_iounmap(pdev, vp->cb_fn_base);
 
 	unregister_netdev(dev);
 
-	if (VORTEX_PCI(vp)) {
-		pci_set_power_state(VORTEX_PCI(vp), PCI_D0);	/* Go active */
-		if (vp->pm_state_valid)
-			pci_restore_state(VORTEX_PCI(vp));
-		pci_disable_device(VORTEX_PCI(vp));
-	}
+	pci_set_power_state(pdev, PCI_D0);	/* Go active */
+	if (vp->pm_state_valid)
+		pci_restore_state(pdev);
+	pci_disable_device(pdev);
+
 	/* Should really use issue_and_wait() here */
 	iowrite16(TotalReset | ((vp->drv_flags & EEPROM_RESET) ? 0x04 : 0x14),
 	     vp->ioaddr + EL3_CMD);
 
-	pci_iounmap(VORTEX_PCI(vp), vp->ioaddr);
+	pci_iounmap(pdev, vp->ioaddr);
 
 	pci_free_consistent(pdev,
 						sizeof(struct boom_rx_desc) * RX_RING_SIZE
diff --git a/drivers/net/ethernet/3com/Kconfig b/drivers/net/ethernet/3com/Kconfig
index 1c71c76..f00c763 100644
--- a/drivers/net/ethernet/3com/Kconfig
+++ b/drivers/net/ethernet/3com/Kconfig
@@ -67,7 +67,6 @@ config PCMCIA_3C589
 config VORTEX
 	tristate "3c590/3c900 series (592/595/597) \"Vortex/Boomerang\" support"
 	depends on (PCI || EISA) && HAS_IOPORT
-	select NET_CORE
 	select MII
 	---help---
 	  This option enables driver support for a large number of 10Mbps and
diff --git a/drivers/net/ethernet/8390/ne.c b/drivers/net/ethernet/8390/ne.c
index 47618e5..b2e8405 100644
--- a/drivers/net/ethernet/8390/ne.c
+++ b/drivers/net/ethernet/8390/ne.c
@@ -849,7 +849,6 @@ static int ne_drv_remove(struct platform_device *pdev)
 		free_irq(dev->irq, dev);
 		release_region(dev->base_addr, NE_IO_EXTENT);
 		free_netdev(dev);
-		platform_set_drvdata(pdev, NULL);
 	}
 	return 0;
 }
diff --git a/drivers/net/ethernet/8390/ne2k-pci.c b/drivers/net/ethernet/8390/ne2k-pci.c
index 587a885..9220108 100644
--- a/drivers/net/ethernet/8390/ne2k-pci.c
+++ b/drivers/net/ethernet/8390/ne2k-pci.c
@@ -676,7 +676,7 @@ static int ne2k_pci_resume (struct pci_dev *pdev)
 	struct net_device *dev = pci_get_drvdata (pdev);
 	int rc;
 
-	pci_set_power_state(pdev, 0);
+	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 
 	rc = pci_enable_device(pdev);
diff --git a/drivers/net/ethernet/Kconfig b/drivers/net/ethernet/Kconfig
index ed956e0..2037080 100644
--- a/drivers/net/ethernet/Kconfig
+++ b/drivers/net/ethernet/Kconfig
@@ -20,9 +20,11 @@ config SUNGEM_PHY
 source "drivers/net/ethernet/3com/Kconfig"
 source "drivers/net/ethernet/adaptec/Kconfig"
 source "drivers/net/ethernet/aeroflex/Kconfig"
+source "drivers/net/ethernet/allwinner/Kconfig"
 source "drivers/net/ethernet/alteon/Kconfig"
 source "drivers/net/ethernet/amd/Kconfig"
 source "drivers/net/ethernet/apple/Kconfig"
+source "drivers/net/ethernet/arc/Kconfig"
 source "drivers/net/ethernet/atheros/Kconfig"
 source "drivers/net/ethernet/cadence/Kconfig"
 source "drivers/net/ethernet/adi/Kconfig"
@@ -63,7 +65,6 @@ config JME
 	tristate "JMicron(R) PCI-Express Gigabit Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports the PCI-Express gigabit ethernet adapters
@@ -95,7 +96,6 @@ config FEALNX
 	tristate "Myson MTD-8xx PCI Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here to support the Myson MTD-800 family of PCI-based Ethernet
@@ -106,7 +106,6 @@ source "drivers/net/ethernet/8390/Kconfig"
 
 config NET_NETX
 	tristate "NetX Ethernet support"
-	select NET_CORE
 	select MII
 	depends on ARCH_NETX
 	---help---
@@ -124,7 +123,6 @@ source "drivers/net/ethernet/oki-semi/Kconfig"
 config ETHOC
 	tristate "OpenCores 10/100 Mbps Ethernet MAC support"
 	depends on HAS_IOMEM && HAS_DMA
-	select NET_CORE
 	select MII
 	select PHYLIB
 	select CRC32
diff --git a/drivers/net/ethernet/Makefile b/drivers/net/ethernet/Makefile
index 8268d85..390bd0b 100644
--- a/drivers/net/ethernet/Makefile
+++ b/drivers/net/ethernet/Makefile
@@ -6,9 +6,11 @@ obj-$(CONFIG_NET_VENDOR_3COM) += 3com/
 obj-$(CONFIG_NET_VENDOR_8390) += 8390/
 obj-$(CONFIG_NET_VENDOR_ADAPTEC) += adaptec/
 obj-$(CONFIG_GRETH) += aeroflex/
+obj-$(CONFIG_NET_VENDOR_ALLWINNER) += allwinner/
 obj-$(CONFIG_NET_VENDOR_ALTEON) += alteon/
 obj-$(CONFIG_NET_VENDOR_AMD) += amd/
 obj-$(CONFIG_NET_VENDOR_APPLE) += apple/
+obj-$(CONFIG_NET_VENDOR_ARC) += arc/
 obj-$(CONFIG_NET_VENDOR_ATHEROS) += atheros/
 obj-$(CONFIG_NET_CADENCE) += cadence/
 obj-$(CONFIG_NET_BFIN) += adi/
diff --git a/drivers/net/ethernet/adaptec/Kconfig b/drivers/net/ethernet/adaptec/Kconfig
index 0bff571..5c804bb 100644
--- a/drivers/net/ethernet/adaptec/Kconfig
+++ b/drivers/net/ethernet/adaptec/Kconfig
@@ -22,7 +22,6 @@ config ADAPTEC_STARFIRE
 	tristate "Adaptec Starfire/DuraLAN support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here if you have an Adaptec Starfire (or DuraLAN) PCI network
diff --git a/drivers/net/ethernet/adi/Kconfig b/drivers/net/ethernet/adi/Kconfig
index a948160..f952fff 100644
--- a/drivers/net/ethernet/adi/Kconfig
+++ b/drivers/net/ethernet/adi/Kconfig
@@ -23,7 +23,6 @@ config BFIN_MAC
 	tristate "Blackfin on-chip MAC support"
 	depends on (BF516 || BF518 || BF526 || BF527 || BF536 || BF537)
 	select CRC32
-	select NET_CORE
 	select MII
 	select PHYLIB
 	select BFIN_MAC_USE_L1 if DMA_UNCACHED_NONE
diff --git a/drivers/net/ethernet/adi/bfin_mac.c b/drivers/net/ethernet/adi/bfin_mac.c
index dada66b..e904b38 100644
--- a/drivers/net/ethernet/adi/bfin_mac.c
+++ b/drivers/net/ethernet/adi/bfin_mac.c
@@ -1719,7 +1719,6 @@ out_err_mii_probe:
 	mdiobus_unregister(lp->mii_bus);
 	mdiobus_free(lp->mii_bus);
 out_err_probe_mac:
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(ndev);
 
 	return rc;
@@ -1732,8 +1731,6 @@ static int bfin_mac_remove(struct platform_device *pdev)
 
 	bfin_phc_release(lp);
 
-	platform_set_drvdata(pdev, NULL);
-
 	lp->mii_bus->priv = NULL;
 
 	unregister_netdev(ndev);
@@ -1868,7 +1865,6 @@ static int bfin_mii_bus_remove(struct platform_device *pdev)
 	struct bfin_mii_bus_platform_data *mii_bus_pd =
 		dev_get_platdata(&pdev->dev);
 
-	platform_set_drvdata(pdev, NULL);
 	mdiobus_unregister(miibus);
 	kfree(miibus->irq);
 	mdiobus_free(miibus);
diff --git a/drivers/net/ethernet/aeroflex/greth.c b/drivers/net/ethernet/aeroflex/greth.c
index 2692954..7ff4b30 100644
--- a/drivers/net/ethernet/aeroflex/greth.c
+++ b/drivers/net/ethernet/aeroflex/greth.c
@@ -1565,7 +1565,7 @@ error1:
 
 static int greth_of_remove(struct platform_device *of_dev)
 {
-	struct net_device *ndev = dev_get_drvdata(&of_dev->dev);
+	struct net_device *ndev = platform_get_drvdata(of_dev);
 	struct greth_private *greth = netdev_priv(ndev);
 
 	/* Free descriptor areas */
@@ -1573,8 +1573,6 @@ static int greth_of_remove(struct platform_device *of_dev)
 
 	dma_free_coherent(&of_dev->dev, 1024, greth->tx_bd_base, greth->tx_bd_base_phys);
 
-	dev_set_drvdata(&of_dev->dev, NULL);
-
 	if (greth->phy)
 		phy_stop(greth->phy);
 	mdiobus_unregister(greth->mdio);
diff --git a/drivers/net/ethernet/allwinner/Kconfig b/drivers/net/ethernet/allwinner/Kconfig
new file mode 100644
index 0000000..53ad213
--- /dev/null
+++ b/drivers/net/ethernet/allwinner/Kconfig
@@ -0,0 +1,35 @@
+#
+# Allwinner device configuration
+#
+
+config NET_VENDOR_ALLWINNER
+       bool "Allwinner devices"
+       default y
+       depends on ARCH_SUNXI
+       ---help---
+         If you have a network (Ethernet) card belonging to this
+	 class, say Y and read the Ethernet-HOWTO, available from
+	 <http://www.tldp.org/docs.html#howto>.
+
+	 Note that the answer to this question doesn't directly
+	 affect the kernel: saying N will just cause the configurator
+	 to skip all the questions about Allwinner cards. If you say Y,
+	 you will be asked for your specific card in the following
+	 questions.
+
+if NET_VENDOR_ALLWINNER
+
+config SUN4I_EMAC
+        tristate "Allwinner A10 EMAC support"
+	depends on ARCH_SUNXI
+	depends on OF
+	select CRC32
+	select MII
+	select PHYLIB
+        ---help---
+          Support for Allwinner A10 EMAC ethernet driver.
+
+          To compile this driver as a module, choose M here.  The module
+          will be called sun4i-emac.
+
+endif # NET_VENDOR_ALLWINNER
diff --git a/drivers/net/ethernet/allwinner/Makefile b/drivers/net/ethernet/allwinner/Makefile
new file mode 100644
index 0000000..03129f7
--- /dev/null
+++ b/drivers/net/ethernet/allwinner/Makefile
@@ -0,0 +1,5 @@
+#
+# Makefile for the Allwinner device drivers.
+#
+
+obj-$(CONFIG_SUN4I_EMAC) += sun4i-emac.o
diff --git a/drivers/net/ethernet/allwinner/sun4i-emac.c b/drivers/net/ethernet/allwinner/sun4i-emac.c
new file mode 100644
index 0000000..50b853a
--- /dev/null
+++ b/drivers/net/ethernet/allwinner/sun4i-emac.c
@@ -0,0 +1,954 @@
+/*
+ * Allwinner EMAC Fast Ethernet driver for Linux.
+ *
+ * Copyright 2012-2013 Stefan Roese <sr@denx.de>
+ * Copyright 2013 Maxime Ripard <maxime.ripard@free-electrons.com>
+ *
+ * Based on the Linux driver provided by Allwinner:
+ * Copyright (C) 1997  Sten Wang
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <linux/clk.h>
+#include <linux/etherdevice.h>
+#include <linux/ethtool.h>
+#include <linux/gpio.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/mii.h>
+#include <linux/module.h>
+#include <linux/netdevice.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_mdio.h>
+#include <linux/of_net.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/phy.h>
+
+#include "sun4i-emac.h"
+
+#define DRV_NAME		"sun4i-emac"
+#define DRV_VERSION		"1.02"
+
+#define EMAC_MAX_FRAME_LEN	0x0600
+
+/* Transmit timeout, default 5 seconds. */
+static int watchdog = 5000;
+module_param(watchdog, int, 0400);
+MODULE_PARM_DESC(watchdog, "transmit timeout in milliseconds");
+
+/* EMAC register address locking.
+ *
+ * The EMAC uses an address register to control where data written
+ * to the data register goes. This means that the address register
+ * must be preserved over interrupts or similar calls.
+ *
+ * During interrupt and other critical calls, a spinlock is used to
+ * protect the system, but the calls themselves save the address
+ * in the address register in case they are interrupting another
+ * access to the device.
+ *
+ * For general accesses a lock is provided so that calls which are
+ * allowed to sleep are serialised so that the address register does
+ * not need to be saved. This lock also serves to serialise access
+ * to the EEPROM and PHY access registers which are shared between
+ * these two devices.
+ */
+
+/* The driver supports the original EMACE, and now the two newer
+ * devices, EMACA and EMACB.
+ */
+
+struct emac_board_info {
+	struct clk		*clk;
+	struct device		*dev;
+	struct platform_device	*pdev;
+	spinlock_t		lock;
+	void __iomem		*membase;
+	u32			msg_enable;
+	struct net_device	*ndev;
+	struct sk_buff		*skb_last;
+	u16			tx_fifo_stat;
+
+	int			emacrx_completed_flag;
+
+	struct phy_device	*phy_dev;
+	struct device_node	*phy_node;
+	unsigned int		link;
+	unsigned int		speed;
+	unsigned int		duplex;
+
+	phy_interface_t		phy_interface;
+};
+
+static void emac_update_speed(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	unsigned int reg_val;
+
+	/* set EMAC SPEED, depend on PHY  */
+	reg_val = readl(db->membase + EMAC_MAC_SUPP_REG);
+	reg_val &= ~(0x1 << 8);
+	if (db->speed == SPEED_100)
+		reg_val |= 1 << 8;
+	writel(reg_val, db->membase + EMAC_MAC_SUPP_REG);
+}
+
+static void emac_update_duplex(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	unsigned int reg_val;
+
+	/* set duplex depend on phy */
+	reg_val = readl(db->membase + EMAC_MAC_CTL1_REG);
+	reg_val &= ~EMAC_MAC_CTL1_DUPLEX_EN;
+	if (db->duplex)
+		reg_val |= EMAC_MAC_CTL1_DUPLEX_EN;
+	writel(reg_val, db->membase + EMAC_MAC_CTL1_REG);
+}
+
+static void emac_handle_link_change(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	struct phy_device *phydev = db->phy_dev;
+	unsigned long flags;
+	int status_change = 0;
+
+	if (phydev->link) {
+		if (db->speed != phydev->speed) {
+			spin_lock_irqsave(&db->lock, flags);
+			db->speed = phydev->speed;
+			emac_update_speed(dev);
+			spin_unlock_irqrestore(&db->lock, flags);
+			status_change = 1;
+		}
+
+		if (db->duplex != phydev->duplex) {
+			spin_lock_irqsave(&db->lock, flags);
+			db->duplex = phydev->duplex;
+			emac_update_duplex(dev);
+			spin_unlock_irqrestore(&db->lock, flags);
+			status_change = 1;
+		}
+	}
+
+	if (phydev->link != db->link) {
+		if (!phydev->link) {
+			db->speed = 0;
+			db->duplex = -1;
+		}
+		db->link = phydev->link;
+
+		status_change = 1;
+	}
+
+	if (status_change)
+		phy_print_status(phydev);
+}
+
+static int emac_mdio_probe(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+
+	/* to-do: PHY interrupts are currently not supported */
+
+	/* attach the mac to the phy */
+	db->phy_dev = of_phy_connect(db->ndev, db->phy_node,
+				     &emac_handle_link_change, 0,
+				     db->phy_interface);
+	if (!db->phy_dev) {
+		netdev_err(db->ndev, "could not find the PHY\n");
+		return -ENODEV;
+	}
+
+	/* mask with MAC supported features */
+	db->phy_dev->supported &= PHY_BASIC_FEATURES;
+	db->phy_dev->advertising = db->phy_dev->supported;
+
+	db->link = 0;
+	db->speed = 0;
+	db->duplex = -1;
+
+	return 0;
+}
+
+static void emac_mdio_remove(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+
+	phy_disconnect(db->phy_dev);
+	db->phy_dev = NULL;
+}
+
+static void emac_reset(struct emac_board_info *db)
+{
+	dev_dbg(db->dev, "resetting device\n");
+
+	/* RESET device */
+	writel(0, db->membase + EMAC_CTL_REG);
+	udelay(200);
+	writel(EMAC_CTL_RESET, db->membase + EMAC_CTL_REG);
+	udelay(200);
+}
+
+static void emac_outblk_32bit(void __iomem *reg, void *data, int count)
+{
+	writesl(reg, data, round_up(count, 4) / 4);
+}
+
+static void emac_inblk_32bit(void __iomem *reg, void *data, int count)
+{
+	readsl(reg, data, round_up(count, 4) / 4);
+}
+
+static int emac_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
+{
+	struct emac_board_info *dm = netdev_priv(dev);
+	struct phy_device *phydev = dm->phy_dev;
+
+	if (!netif_running(dev))
+		return -EINVAL;
+
+	if (!phydev)
+		return -ENODEV;
+
+	return phy_mii_ioctl(phydev, rq, cmd);
+}
+
+/* ethtool ops */
+static void emac_get_drvinfo(struct net_device *dev,
+			      struct ethtool_drvinfo *info)
+{
+	strlcpy(info->driver, DRV_NAME, sizeof(DRV_NAME));
+	strlcpy(info->version, DRV_VERSION, sizeof(DRV_VERSION));
+	strlcpy(info->bus_info, dev_name(&dev->dev), sizeof(info->bus_info));
+}
+
+static int emac_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+	struct emac_board_info *dm = netdev_priv(dev);
+	struct phy_device *phydev = dm->phy_dev;
+
+	if (!phydev)
+		return -ENODEV;
+
+	return phy_ethtool_gset(phydev, cmd);
+}
+
+static int emac_set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
+{
+	struct emac_board_info *dm = netdev_priv(dev);
+	struct phy_device *phydev = dm->phy_dev;
+
+	if (!phydev)
+		return -ENODEV;
+
+	return phy_ethtool_sset(phydev, cmd);
+}
+
+static const struct ethtool_ops emac_ethtool_ops = {
+	.get_drvinfo	= emac_get_drvinfo,
+	.get_settings	= emac_get_settings,
+	.set_settings	= emac_set_settings,
+	.get_link	= ethtool_op_get_link,
+};
+
+static unsigned int emac_setup(struct net_device *ndev)
+{
+	struct emac_board_info *db = netdev_priv(ndev);
+	unsigned int reg_val;
+
+	/* set up TX */
+	reg_val = readl(db->membase + EMAC_TX_MODE_REG);
+
+	writel(reg_val | EMAC_TX_MODE_ABORTED_FRAME_EN,
+		db->membase + EMAC_TX_MODE_REG);
+
+	/* set up RX */
+	reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+
+	writel(reg_val | EMAC_RX_CTL_PASS_LEN_OOR_EN |
+		EMAC_RX_CTL_ACCEPT_UNICAST_EN | EMAC_RX_CTL_DA_FILTER_EN |
+		EMAC_RX_CTL_ACCEPT_MULTICAST_EN |
+		EMAC_RX_CTL_ACCEPT_BROADCAST_EN,
+		db->membase + EMAC_RX_CTL_REG);
+
+	/* set MAC */
+	/* set MAC CTL0 */
+	reg_val = readl(db->membase + EMAC_MAC_CTL0_REG);
+	writel(reg_val | EMAC_MAC_CTL0_RX_FLOW_CTL_EN |
+		EMAC_MAC_CTL0_TX_FLOW_CTL_EN,
+		db->membase + EMAC_MAC_CTL0_REG);
+
+	/* set MAC CTL1 */
+	reg_val = readl(db->membase + EMAC_MAC_CTL1_REG);
+	reg_val |= EMAC_MAC_CTL1_LEN_CHECK_EN;
+	reg_val |= EMAC_MAC_CTL1_CRC_EN;
+	reg_val |= EMAC_MAC_CTL1_PAD_EN;
+	writel(reg_val, db->membase + EMAC_MAC_CTL1_REG);
+
+	/* set up IPGT */
+	writel(EMAC_MAC_IPGT_FULL_DUPLEX, db->membase + EMAC_MAC_IPGT_REG);
+
+	/* set up IPGR */
+	writel((EMAC_MAC_IPGR_IPG1 << 8) | EMAC_MAC_IPGR_IPG2,
+		db->membase + EMAC_MAC_IPGR_REG);
+
+	/* set up Collison window */
+	writel((EMAC_MAC_CLRT_COLLISION_WINDOW << 8) | EMAC_MAC_CLRT_RM,
+		db->membase + EMAC_MAC_CLRT_REG);
+
+	/* set up Max Frame Length */
+	writel(EMAC_MAX_FRAME_LEN,
+		db->membase + EMAC_MAC_MAXF_REG);
+
+	return 0;
+}
+
+static unsigned int emac_powerup(struct net_device *ndev)
+{
+	struct emac_board_info *db = netdev_priv(ndev);
+	unsigned int reg_val;
+
+	/* initial EMAC */
+	/* flush RX FIFO */
+	reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+	reg_val |= 0x8;
+	writel(reg_val, db->membase + EMAC_RX_CTL_REG);
+	udelay(1);
+
+	/* initial MAC */
+	/* soft reset MAC */
+	reg_val = readl(db->membase + EMAC_MAC_CTL0_REG);
+	reg_val &= ~EMAC_MAC_CTL0_SOFT_RESET;
+	writel(reg_val, db->membase + EMAC_MAC_CTL0_REG);
+
+	/* set MII clock */
+	reg_val = readl(db->membase + EMAC_MAC_MCFG_REG);
+	reg_val &= (~(0xf << 2));
+	reg_val |= (0xD << 2);
+	writel(reg_val, db->membase + EMAC_MAC_MCFG_REG);
+
+	/* clear RX counter */
+	writel(0x0, db->membase + EMAC_RX_FBC_REG);
+
+	/* disable all interrupt and clear interrupt status */
+	writel(0, db->membase + EMAC_INT_CTL_REG);
+	reg_val = readl(db->membase + EMAC_INT_STA_REG);
+	writel(reg_val, db->membase + EMAC_INT_STA_REG);
+
+	udelay(1);
+
+	/* set up EMAC */
+	emac_setup(ndev);
+
+	/* set mac_address to chip */
+	writel(ndev->dev_addr[0] << 16 | ndev->dev_addr[1] << 8 | ndev->
+	       dev_addr[2], db->membase + EMAC_MAC_A1_REG);
+	writel(ndev->dev_addr[3] << 16 | ndev->dev_addr[4] << 8 | ndev->
+	       dev_addr[5], db->membase + EMAC_MAC_A0_REG);
+
+	mdelay(1);
+
+	return 0;
+}
+
+static int emac_set_mac_address(struct net_device *dev, void *p)
+{
+	struct sockaddr *addr = p;
+	struct emac_board_info *db = netdev_priv(dev);
+
+	if (netif_running(dev))
+		return -EBUSY;
+
+	memcpy(dev->dev_addr, addr->sa_data, ETH_ALEN);
+
+	writel(dev->dev_addr[0] << 16 | dev->dev_addr[1] << 8 | dev->
+	       dev_addr[2], db->membase + EMAC_MAC_A1_REG);
+	writel(dev->dev_addr[3] << 16 | dev->dev_addr[4] << 8 | dev->
+	       dev_addr[5], db->membase + EMAC_MAC_A0_REG);
+
+	return 0;
+}
+
+/* Initialize emac board */
+static void emac_init_device(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	unsigned long flags;
+	unsigned int reg_val;
+
+	spin_lock_irqsave(&db->lock, flags);
+
+	emac_update_speed(dev);
+	emac_update_duplex(dev);
+
+	/* enable RX/TX */
+	reg_val = readl(db->membase + EMAC_CTL_REG);
+	writel(reg_val | EMAC_CTL_RESET | EMAC_CTL_TX_EN | EMAC_CTL_RX_EN,
+		db->membase + EMAC_CTL_REG);
+
+	/* enable RX/TX0/RX Hlevel interrup */
+	reg_val = readl(db->membase + EMAC_INT_CTL_REG);
+	reg_val |= (0xf << 0) | (0x01 << 8);
+	writel(reg_val, db->membase + EMAC_INT_CTL_REG);
+
+	spin_unlock_irqrestore(&db->lock, flags);
+}
+
+/* Our watchdog timed out. Called by the networking layer */
+static void emac_timeout(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	unsigned long flags;
+
+	if (netif_msg_timer(db))
+		dev_err(db->dev, "tx time out.\n");
+
+	/* Save previous register address */
+	spin_lock_irqsave(&db->lock, flags);
+
+	netif_stop_queue(dev);
+	emac_reset(db);
+	emac_init_device(dev);
+	/* We can accept TX packets again */
+	dev->trans_start = jiffies;
+	netif_wake_queue(dev);
+
+	/* Restore previous register address */
+	spin_unlock_irqrestore(&db->lock, flags);
+}
+
+/* Hardware start transmission.
+ * Send a packet to media from the upper layer.
+ */
+static int emac_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	unsigned long channel;
+	unsigned long flags;
+
+	channel = db->tx_fifo_stat & 3;
+	if (channel == 3)
+		return 1;
+
+	channel = (channel == 1 ? 1 : 0);
+
+	spin_lock_irqsave(&db->lock, flags);
+
+	writel(channel, db->membase + EMAC_TX_INS_REG);
+
+	emac_outblk_32bit(db->membase + EMAC_TX_IO_DATA_REG,
+			skb->data, skb->len);
+	dev->stats.tx_bytes += skb->len;
+
+	db->tx_fifo_stat |= 1 << channel;
+	/* TX control: First packet immediately send, second packet queue */
+	if (channel == 0) {
+		/* set TX len */
+		writel(skb->len, db->membase + EMAC_TX_PL0_REG);
+		/* start translate from fifo to phy */
+		writel(readl(db->membase + EMAC_TX_CTL0_REG) | 1,
+		       db->membase + EMAC_TX_CTL0_REG);
+
+		/* save the time stamp */
+		dev->trans_start = jiffies;
+	} else if (channel == 1) {
+		/* set TX len */
+		writel(skb->len, db->membase + EMAC_TX_PL1_REG);
+		/* start translate from fifo to phy */
+		writel(readl(db->membase + EMAC_TX_CTL1_REG) | 1,
+		       db->membase + EMAC_TX_CTL1_REG);
+
+		/* save the time stamp */
+		dev->trans_start = jiffies;
+	}
+
+	if ((db->tx_fifo_stat & 3) == 3) {
+		/* Second packet */
+		netif_stop_queue(dev);
+	}
+
+	spin_unlock_irqrestore(&db->lock, flags);
+
+	/* free this SKB */
+	dev_kfree_skb(skb);
+
+	return NETDEV_TX_OK;
+}
+
+/* EMAC interrupt handler
+ * receive the packet to upper layer, free the transmitted packet
+ */
+static void emac_tx_done(struct net_device *dev, struct emac_board_info *db,
+			  unsigned int tx_status)
+{
+	/* One packet sent complete */
+	db->tx_fifo_stat &= ~(tx_status & 3);
+	if (3 == (tx_status & 3))
+		dev->stats.tx_packets += 2;
+	else
+		dev->stats.tx_packets++;
+
+	if (netif_msg_tx_done(db))
+		dev_dbg(db->dev, "tx done, NSR %02x\n", tx_status);
+
+	netif_wake_queue(dev);
+}
+
+/* Received a packet and pass to upper layer
+ */
+static void emac_rx(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	struct sk_buff *skb;
+	u8 *rdptr;
+	bool good_packet;
+	static int rxlen_last;
+	unsigned int reg_val;
+	u32 rxhdr, rxstatus, rxcount, rxlen;
+
+	/* Check packet ready or not */
+	while (1) {
+		/* race warning: the first packet might arrive with
+		 * the interrupts disabled, but the second will fix
+		 * it
+		 */
+		rxcount = readl(db->membase + EMAC_RX_FBC_REG);
+
+		if (netif_msg_rx_status(db))
+			dev_dbg(db->dev, "RXCount: %x\n", rxcount);
+
+		if ((db->skb_last != NULL) && (rxlen_last > 0)) {
+			dev->stats.rx_bytes += rxlen_last;
+
+			/* Pass to upper layer */
+			db->skb_last->protocol = eth_type_trans(db->skb_last,
+								dev);
+			netif_rx(db->skb_last);
+			dev->stats.rx_packets++;
+			db->skb_last = NULL;
+			rxlen_last = 0;
+
+			reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+			reg_val &= ~EMAC_RX_CTL_DMA_EN;
+			writel(reg_val, db->membase + EMAC_RX_CTL_REG);
+		}
+
+		if (!rxcount) {
+			db->emacrx_completed_flag = 1;
+			reg_val = readl(db->membase + EMAC_INT_CTL_REG);
+			reg_val |= (0xf << 0) | (0x01 << 8);
+			writel(reg_val, db->membase + EMAC_INT_CTL_REG);
+
+			/* had one stuck? */
+			rxcount = readl(db->membase + EMAC_RX_FBC_REG);
+			if (!rxcount)
+				return;
+		}
+
+		reg_val = readl(db->membase + EMAC_RX_IO_DATA_REG);
+		if (netif_msg_rx_status(db))
+			dev_dbg(db->dev, "receive header: %x\n", reg_val);
+		if (reg_val != EMAC_UNDOCUMENTED_MAGIC) {
+			/* disable RX */
+			reg_val = readl(db->membase + EMAC_CTL_REG);
+			writel(reg_val & ~EMAC_CTL_RX_EN,
+			       db->membase + EMAC_CTL_REG);
+
+			/* Flush RX FIFO */
+			reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+			writel(reg_val | (1 << 3),
+			       db->membase + EMAC_RX_CTL_REG);
+
+			do {
+				reg_val = readl(db->membase + EMAC_RX_CTL_REG);
+			} while (reg_val & (1 << 3));
+
+			/* enable RX */
+			reg_val = readl(db->membase + EMAC_CTL_REG);
+			writel(reg_val | EMAC_CTL_RX_EN,
+			       db->membase + EMAC_CTL_REG);
+			reg_val = readl(db->membase + EMAC_INT_CTL_REG);
+			reg_val |= (0xf << 0) | (0x01 << 8);
+			writel(reg_val, db->membase + EMAC_INT_CTL_REG);
+
+			db->emacrx_completed_flag = 1;
+
+			return;
+		}
+
+		/* A packet ready now  & Get status/length */
+		good_packet = true;
+
+		emac_inblk_32bit(db->membase + EMAC_RX_IO_DATA_REG,
+				&rxhdr, sizeof(rxhdr));
+
+		if (netif_msg_rx_status(db))
+			dev_dbg(db->dev, "rxhdr: %x\n", *((int *)(&rxhdr)));
+
+		rxlen = EMAC_RX_IO_DATA_LEN(rxhdr);
+		rxstatus = EMAC_RX_IO_DATA_STATUS(rxhdr);
+
+		if (netif_msg_rx_status(db))
+			dev_dbg(db->dev, "RX: status %02x, length %04x\n",
+				rxstatus, rxlen);
+
+		/* Packet Status check */
+		if (rxlen < 0x40) {
+			good_packet = false;
+			if (netif_msg_rx_err(db))
+				dev_dbg(db->dev, "RX: Bad Packet (runt)\n");
+		}
+
+		if (unlikely(!(rxstatus & EMAC_RX_IO_DATA_STATUS_OK))) {
+			good_packet = false;
+
+			if (rxstatus & EMAC_RX_IO_DATA_STATUS_CRC_ERR) {
+				if (netif_msg_rx_err(db))
+					dev_dbg(db->dev, "crc error\n");
+				dev->stats.rx_crc_errors++;
+			}
+
+			if (rxstatus & EMAC_RX_IO_DATA_STATUS_LEN_ERR) {
+				if (netif_msg_rx_err(db))
+					dev_dbg(db->dev, "length error\n");
+				dev->stats.rx_length_errors++;
+			}
+		}
+
+		/* Move data from EMAC */
+		skb = dev_alloc_skb(rxlen + 4);
+		if (good_packet && skb) {
+			skb_reserve(skb, 2);
+			rdptr = (u8 *) skb_put(skb, rxlen - 4);
+
+			/* Read received packet from RX SRAM */
+			if (netif_msg_rx_status(db))
+				dev_dbg(db->dev, "RxLen %x\n", rxlen);
+
+			emac_inblk_32bit(db->membase + EMAC_RX_IO_DATA_REG,
+					rdptr, rxlen);
+			dev->stats.rx_bytes += rxlen;
+
+			/* Pass to upper layer */
+			skb->protocol = eth_type_trans(skb, dev);
+			netif_rx(skb);
+			dev->stats.rx_packets++;
+		}
+	}
+}
+
+static irqreturn_t emac_interrupt(int irq, void *dev_id)
+{
+	struct net_device *dev = dev_id;
+	struct emac_board_info *db = netdev_priv(dev);
+	int int_status;
+	unsigned long flags;
+	unsigned int reg_val;
+
+	/* A real interrupt coming */
+
+	/* holders of db->lock must always block IRQs */
+	spin_lock_irqsave(&db->lock, flags);
+
+	/* Disable all interrupts */
+	writel(0, db->membase + EMAC_INT_CTL_REG);
+
+	/* Got EMAC interrupt status */
+	/* Got ISR */
+	int_status = readl(db->membase + EMAC_INT_STA_REG);
+	/* Clear ISR status */
+	writel(int_status, db->membase + EMAC_INT_STA_REG);
+
+	if (netif_msg_intr(db))
+		dev_dbg(db->dev, "emac interrupt %02x\n", int_status);
+
+	/* Received the coming packet */
+	if ((int_status & 0x100) && (db->emacrx_completed_flag == 1)) {
+		/* carrier lost */
+		db->emacrx_completed_flag = 0;
+		emac_rx(dev);
+	}
+
+	/* Transmit Interrupt check */
+	if (int_status & (0x01 | 0x02))
+		emac_tx_done(dev, db, int_status);
+
+	if (int_status & (0x04 | 0x08))
+		netdev_info(dev, " ab : %x\n", int_status);
+
+	/* Re-enable interrupt mask */
+	if (db->emacrx_completed_flag == 1) {
+		reg_val = readl(db->membase + EMAC_INT_CTL_REG);
+		reg_val |= (0xf << 0) | (0x01 << 8);
+		writel(reg_val, db->membase + EMAC_INT_CTL_REG);
+	}
+	spin_unlock_irqrestore(&db->lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+/*
+ * Used by netconsole
+ */
+static void emac_poll_controller(struct net_device *dev)
+{
+	disable_irq(dev->irq);
+	emac_interrupt(dev->irq, dev);
+	enable_irq(dev->irq);
+}
+#endif
+
+/*  Open the interface.
+ *  The interface is opened whenever "ifconfig" actives it.
+ */
+static int emac_open(struct net_device *dev)
+{
+	struct emac_board_info *db = netdev_priv(dev);
+	int ret;
+
+	if (netif_msg_ifup(db))
+		dev_dbg(db->dev, "enabling %s\n", dev->name);
+
+	if (devm_request_irq(db->dev, dev->irq, &emac_interrupt,
+			     0, dev->name, dev))
+		return -EAGAIN;
+
+	/* Initialize EMAC board */
+	emac_reset(db);
+	emac_init_device(dev);
+
+	ret = emac_mdio_probe(dev);
+	if (ret < 0) {
+		netdev_err(dev, "cannot probe MDIO bus\n");
+		return ret;
+	}
+
+	phy_start(db->phy_dev);
+	netif_start_queue(dev);
+
+	return 0;
+}
+
+static void emac_shutdown(struct net_device *dev)
+{
+	unsigned int reg_val;
+	struct emac_board_info *db = netdev_priv(dev);
+
+	/* Disable all interrupt */
+	writel(0, db->membase + EMAC_INT_CTL_REG);
+
+	/* clear interupt status */
+	reg_val = readl(db->membase + EMAC_INT_STA_REG);
+	writel(reg_val, db->membase + EMAC_INT_STA_REG);
+
+	/* Disable RX/TX */
+	reg_val = readl(db->membase + EMAC_CTL_REG);
+	reg_val &= ~(EMAC_CTL_TX_EN | EMAC_CTL_RX_EN | EMAC_CTL_RESET);
+	writel(reg_val, db->membase + EMAC_CTL_REG);
+}
+
+/* Stop the interface.
+ * The interface is stopped when it is brought.
+ */
+static int emac_stop(struct net_device *ndev)
+{
+	struct emac_board_info *db = netdev_priv(ndev);
+
+	if (netif_msg_ifdown(db))
+		dev_dbg(db->dev, "shutting down %s\n", ndev->name);
+
+	netif_stop_queue(ndev);
+	netif_carrier_off(ndev);
+
+	phy_stop(db->phy_dev);
+
+	emac_mdio_remove(ndev);
+
+	emac_shutdown(ndev);
+
+	return 0;
+}
+
+static const struct net_device_ops emac_netdev_ops = {
+	.ndo_open		= emac_open,
+	.ndo_stop		= emac_stop,
+	.ndo_start_xmit		= emac_start_xmit,
+	.ndo_tx_timeout		= emac_timeout,
+	.ndo_do_ioctl		= emac_ioctl,
+	.ndo_change_mtu		= eth_change_mtu,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_set_mac_address	= emac_set_mac_address,
+#ifdef CONFIG_NET_POLL_CONTROLLER
+	.ndo_poll_controller	= emac_poll_controller,
+#endif
+};
+
+/* Search EMAC board, allocate space and register it
+ */
+static int emac_probe(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct emac_board_info *db;
+	struct net_device *ndev;
+	int ret = 0;
+	const char *mac_addr;
+
+	ndev = alloc_etherdev(sizeof(struct emac_board_info));
+	if (!ndev) {
+		dev_err(&pdev->dev, "could not allocate device.\n");
+		return -ENOMEM;
+	}
+
+	SET_NETDEV_DEV(ndev, &pdev->dev);
+
+	db = netdev_priv(ndev);
+	memset(db, 0, sizeof(*db));
+
+	db->dev = &pdev->dev;
+	db->ndev = ndev;
+	db->pdev = pdev;
+
+	spin_lock_init(&db->lock);
+
+	db->membase = of_iomap(np, 0);
+	if (!db->membase) {
+		dev_err(&pdev->dev, "failed to remap registers\n");
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	/* fill in parameters for net-dev structure */
+	ndev->base_addr = (unsigned long)db->membase;
+	ndev->irq = irq_of_parse_and_map(np, 0);
+	if (ndev->irq == -ENXIO) {
+		netdev_err(ndev, "No irq resource\n");
+		ret = ndev->irq;
+		goto out;
+	}
+
+	db->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(db->clk))
+		goto out;
+
+	clk_prepare_enable(db->clk);
+
+	db->phy_node = of_parse_phandle(np, "phy", 0);
+	if (!db->phy_node) {
+		dev_err(&pdev->dev, "no associated PHY\n");
+		ret = -ENODEV;
+		goto out;
+	}
+
+	/* Read MAC-address from DT */
+	mac_addr = of_get_mac_address(np);
+	if (mac_addr)
+		memcpy(ndev->dev_addr, mac_addr, ETH_ALEN);
+
+	/* Check if the MAC address is valid, if not get a random one */
+	if (!is_valid_ether_addr(ndev->dev_addr)) {
+		eth_hw_addr_random(ndev);
+		dev_warn(&pdev->dev, "using random MAC address %pM\n",
+			 ndev->dev_addr);
+	}
+
+	db->emacrx_completed_flag = 1;
+	emac_powerup(ndev);
+	emac_reset(db);
+
+	ether_setup(ndev);
+
+	ndev->netdev_ops = &emac_netdev_ops;
+	ndev->watchdog_timeo = msecs_to_jiffies(watchdog);
+	ndev->ethtool_ops = &emac_ethtool_ops;
+
+	platform_set_drvdata(pdev, ndev);
+
+	/* Carrier starts down, phylib will bring it up */
+	netif_carrier_off(ndev);
+
+	ret = register_netdev(ndev);
+	if (ret) {
+		dev_err(&pdev->dev, "Registering netdev failed!\n");
+		ret = -ENODEV;
+		goto out;
+	}
+
+	dev_info(&pdev->dev, "%s: at %p, IRQ %d MAC: %pM\n",
+		 ndev->name, db->membase, ndev->irq, ndev->dev_addr);
+
+	return 0;
+
+out:
+	dev_err(db->dev, "not found (%d).\n", ret);
+
+	free_netdev(ndev);
+
+	return ret;
+}
+
+static int emac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+
+	unregister_netdev(ndev);
+	free_netdev(ndev);
+
+	dev_dbg(&pdev->dev, "released and freed device\n");
+	return 0;
+}
+
+static int emac_suspend(struct platform_device *dev, pm_message_t state)
+{
+	struct net_device *ndev = platform_get_drvdata(dev);
+
+	netif_carrier_off(ndev);
+	netif_device_detach(ndev);
+	emac_shutdown(ndev);
+
+	return 0;
+}
+
+static int emac_resume(struct platform_device *dev)
+{
+	struct net_device *ndev = platform_get_drvdata(dev);
+	struct emac_board_info *db = netdev_priv(ndev);
+
+	emac_reset(db);
+	emac_init_device(ndev);
+	netif_device_attach(ndev);
+
+	return 0;
+}
+
+static const struct of_device_id emac_of_match[] = {
+	{.compatible = "allwinner,sun4i-emac",},
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, emac_of_match);
+
+static struct platform_driver emac_driver = {
+	.driver = {
+		.name = "sun4i-emac",
+		.of_match_table = emac_of_match,
+	},
+	.probe = emac_probe,
+	.remove = emac_remove,
+	.suspend = emac_suspend,
+	.resume = emac_resume,
+};
+
+module_platform_driver(emac_driver);
+
+MODULE_AUTHOR("Stefan Roese <sr@denx.de>");
+MODULE_AUTHOR("Maxime Ripard <maxime.ripard@free-electrons.com>");
+MODULE_DESCRIPTION("Allwinner A10 emac network driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/allwinner/sun4i-emac.h b/drivers/net/ethernet/allwinner/sun4i-emac.h
new file mode 100644
index 0000000..38c72d9
--- /dev/null
+++ b/drivers/net/ethernet/allwinner/sun4i-emac.h
@@ -0,0 +1,108 @@
+/*
+ * Allwinner EMAC Fast Ethernet driver for Linux.
+ *
+ * Copyright 2012 Stefan Roese <sr@denx.de>
+ * Copyright 2013 Maxime Ripard <maxime.ripard@free-electrons.com>
+ *
+ * Based on the Linux driver provided by Allwinner:
+ * Copyright (C) 1997  Sten Wang
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#ifndef _SUN4I_EMAC_H_
+#define _SUN4I_EMAC_H_
+
+#define EMAC_CTL_REG		(0x00)
+#define EMAC_CTL_RESET			(1 << 0)
+#define EMAC_CTL_TX_EN			(1 << 1)
+#define EMAC_CTL_RX_EN			(1 << 2)
+#define EMAC_TX_MODE_REG	(0x04)
+#define EMAC_TX_MODE_ABORTED_FRAME_EN	(1 << 0)
+#define EMAC_TX_MODE_DMA_EN		(1 << 1)
+#define EMAC_TX_FLOW_REG	(0x08)
+#define EMAC_TX_CTL0_REG	(0x0c)
+#define EMAC_TX_CTL1_REG	(0x10)
+#define EMAC_TX_INS_REG		(0x14)
+#define EMAC_TX_PL0_REG		(0x18)
+#define EMAC_TX_PL1_REG		(0x1c)
+#define EMAC_TX_STA_REG		(0x20)
+#define EMAC_TX_IO_DATA_REG	(0x24)
+#define EMAC_TX_IO_DATA1_REG	(0x28)
+#define EMAC_TX_TSVL0_REG	(0x2c)
+#define EMAC_TX_TSVH0_REG	(0x30)
+#define EMAC_TX_TSVL1_REG	(0x34)
+#define EMAC_TX_TSVH1_REG	(0x38)
+#define EMAC_RX_CTL_REG		(0x3c)
+#define EMAC_RX_CTL_AUTO_DRQ_EN		(1 << 1)
+#define EMAC_RX_CTL_DMA_EN		(1 << 2)
+#define EMAC_RX_CTL_PASS_ALL_EN		(1 << 4)
+#define EMAC_RX_CTL_PASS_CTL_EN		(1 << 5)
+#define EMAC_RX_CTL_PASS_CRC_ERR_EN	(1 << 6)
+#define EMAC_RX_CTL_PASS_LEN_ERR_EN	(1 << 7)
+#define EMAC_RX_CTL_PASS_LEN_OOR_EN	(1 << 8)
+#define EMAC_RX_CTL_ACCEPT_UNICAST_EN	(1 << 16)
+#define EMAC_RX_CTL_DA_FILTER_EN	(1 << 17)
+#define EMAC_RX_CTL_ACCEPT_MULTICAST_EN	(1 << 20)
+#define EMAC_RX_CTL_HASH_FILTER_EN	(1 << 21)
+#define EMAC_RX_CTL_ACCEPT_BROADCAST_EN	(1 << 22)
+#define EMAC_RX_CTL_SA_FILTER_EN	(1 << 24)
+#define EMAC_RX_CTL_SA_FILTER_INVERT_EN	(1 << 25)
+#define EMAC_RX_HASH0_REG	(0x40)
+#define EMAC_RX_HASH1_REG	(0x44)
+#define EMAC_RX_STA_REG		(0x48)
+#define EMAC_RX_IO_DATA_REG	(0x4c)
+#define EMAC_RX_IO_DATA_LEN(x)		(x & 0xffff)
+#define EMAC_RX_IO_DATA_STATUS(x)	((x >> 16) & 0xffff)
+#define EMAC_RX_IO_DATA_STATUS_CRC_ERR	(1 << 4)
+#define EMAC_RX_IO_DATA_STATUS_LEN_ERR	(3 << 5)
+#define EMAC_RX_IO_DATA_STATUS_OK	(1 << 7)
+#define EMAC_RX_FBC_REG		(0x50)
+#define EMAC_INT_CTL_REG	(0x54)
+#define EMAC_INT_STA_REG	(0x58)
+#define EMAC_MAC_CTL0_REG	(0x5c)
+#define EMAC_MAC_CTL0_RX_FLOW_CTL_EN	(1 << 2)
+#define EMAC_MAC_CTL0_TX_FLOW_CTL_EN	(1 << 3)
+#define EMAC_MAC_CTL0_SOFT_RESET	(1 << 15)
+#define EMAC_MAC_CTL1_REG	(0x60)
+#define EMAC_MAC_CTL1_DUPLEX_EN		(1 << 0)
+#define EMAC_MAC_CTL1_LEN_CHECK_EN	(1 << 1)
+#define EMAC_MAC_CTL1_HUGE_FRAME_EN	(1 << 2)
+#define EMAC_MAC_CTL1_DELAYED_CRC_EN	(1 << 3)
+#define EMAC_MAC_CTL1_CRC_EN		(1 << 4)
+#define EMAC_MAC_CTL1_PAD_EN		(1 << 5)
+#define EMAC_MAC_CTL1_PAD_CRC_EN	(1 << 6)
+#define EMAC_MAC_CTL1_AD_SHORT_FRAME_EN	(1 << 7)
+#define EMAC_MAC_CTL1_BACKOFF_DIS	(1 << 12)
+#define EMAC_MAC_IPGT_REG	(0x64)
+#define EMAC_MAC_IPGT_HALF_DUPLEX	(0x12)
+#define EMAC_MAC_IPGT_FULL_DUPLEX	(0x15)
+#define EMAC_MAC_IPGR_REG	(0x68)
+#define EMAC_MAC_IPGR_IPG1		(0x0c)
+#define EMAC_MAC_IPGR_IPG2		(0x12)
+#define EMAC_MAC_CLRT_REG	(0x6c)
+#define EMAC_MAC_CLRT_COLLISION_WINDOW	(0x37)
+#define EMAC_MAC_CLRT_RM		(0x0f)
+#define EMAC_MAC_MAXF_REG	(0x70)
+#define EMAC_MAC_SUPP_REG	(0x74)
+#define EMAC_MAC_TEST_REG	(0x78)
+#define EMAC_MAC_MCFG_REG	(0x7c)
+#define EMAC_MAC_A0_REG		(0x98)
+#define EMAC_MAC_A1_REG		(0x9c)
+#define EMAC_MAC_A2_REG		(0xa0)
+#define EMAC_SAFX_L_REG0	(0xa4)
+#define EMAC_SAFX_H_REG0	(0xa8)
+#define EMAC_SAFX_L_REG1	(0xac)
+#define EMAC_SAFX_H_REG1	(0xb0)
+#define EMAC_SAFX_L_REG2	(0xb4)
+#define EMAC_SAFX_H_REG2	(0xb8)
+#define EMAC_SAFX_L_REG3	(0xbc)
+#define EMAC_SAFX_H_REG3	(0xc0)
+
+#define EMAC_PHY_DUPLEX		(1 << 8)
+
+#define EMAC_EEPROM_MAGIC	(0x444d394b)
+#define EMAC_UNDOCUMENTED_MAGIC	(0x0143414d)
+#endif /* _SUN4I_EMAC_H_ */
diff --git a/drivers/net/ethernet/alteon/acenic.c b/drivers/net/ethernet/alteon/acenic.c
index b7894f8..219be1b 100644
--- a/drivers/net/ethernet/alteon/acenic.c
+++ b/drivers/net/ethernet/alteon/acenic.c
@@ -702,19 +702,6 @@ static struct pci_driver acenic_pci_driver = {
 	.remove		= acenic_remove_one,
 };
 
-static int __init acenic_init(void)
-{
-	return pci_register_driver(&acenic_pci_driver);
-}
-
-static void __exit acenic_exit(void)
-{
-	pci_unregister_driver(&acenic_pci_driver);
-}
-
-module_init(acenic_init);
-module_exit(acenic_exit);
-
 static void ace_free_descriptors(struct net_device *dev)
 {
 	struct ace_private *ap = netdev_priv(dev);
@@ -3199,3 +3186,5 @@ static int read_eeprom_byte(struct net_device *dev, unsigned long offset)
 	       ap->name, offset);
 	goto out;
 }
+
+module_pci_driver(acenic_pci_driver);
diff --git a/drivers/net/ethernet/amd/Kconfig b/drivers/net/ethernet/amd/Kconfig
index 13d74aa..562df46 100644
--- a/drivers/net/ethernet/amd/Kconfig
+++ b/drivers/net/ethernet/amd/Kconfig
@@ -34,7 +34,6 @@ config AMD8111_ETH
 	tristate "AMD 8111 (new PCI LANCE) support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  If you have an AMD 8111-based PCI LANCE ethernet card,
@@ -60,7 +59,6 @@ config PCNET32
 	tristate "AMD PCnet32 PCI support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a PCnet32 or PCnetPCI based network (Ethernet) card,
diff --git a/drivers/net/ethernet/amd/amd8111e.c b/drivers/net/ethernet/amd/amd8111e.c
index 8e6b665..1b1429d 100644
--- a/drivers/net/ethernet/amd/amd8111e.c
+++ b/drivers/net/ethernet/amd/amd8111e.c
@@ -1813,7 +1813,7 @@ static const struct net_device_ops amd8111e_netdev_ops = {
 static int amd8111e_probe_one(struct pci_dev *pdev,
 				  const struct pci_device_id *ent)
 {
-	int err,i,pm_cap;
+	int err, i;
 	unsigned long reg_addr,reg_len;
 	struct amd8111e_priv* lp;
 	struct net_device* dev;
@@ -1842,7 +1842,7 @@ static int amd8111e_probe_one(struct pci_dev *pdev,
 	pci_set_master(pdev);
 
 	/* Find power-management capability. */
-	if((pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM))==0){
+	if (!pdev->pm_cap) {
 		printk(KERN_ERR "amd8111e: No Power Management capability, "
 		       "exiting.\n");
 		err = -ENODEV;
@@ -1875,7 +1875,7 @@ static int amd8111e_probe_one(struct pci_dev *pdev,
 	lp = netdev_priv(dev);
 	lp->pci_dev = pdev;
 	lp->amd8111e_net_dev = dev;
-	lp->pm_cap = pm_cap;
+	lp->pm_cap = pdev->pm_cap;
 
 	spin_lock_init(&lp->lock);
 
@@ -1981,15 +1981,4 @@ static struct pci_driver amd8111e_driver = {
 	.resume		= amd8111e_resume
 };
 
-static int __init amd8111e_init(void)
-{
-	return pci_register_driver(&amd8111e_driver);
-}
-
-static void __exit amd8111e_cleanup(void)
-{
-	pci_unregister_driver(&amd8111e_driver);
-}
-
-module_init(amd8111e_init);
-module_exit(amd8111e_cleanup);
+module_pci_driver(amd8111e_driver);
diff --git a/drivers/net/ethernet/amd/au1000_eth.c b/drivers/net/ethernet/amd/au1000_eth.c
index 688aede..ceb45bc 100644
--- a/drivers/net/ethernet/amd/au1000_eth.c
+++ b/drivers/net/ethernet/amd/au1000_eth.c
@@ -1301,8 +1301,6 @@ static int au1000_remove(struct platform_device *pdev)
 	int i;
 	struct resource *base, *macen;
 
-	platform_set_drvdata(pdev, NULL);
-
 	unregister_netdev(dev);
 	mdiobus_unregister(aup->mii_bus);
 	mdiobus_free(aup->mii_bus);
diff --git a/drivers/net/ethernet/amd/sunlance.c b/drivers/net/ethernet/amd/sunlance.c
index f47b780..ece5683 100644
--- a/drivers/net/ethernet/amd/sunlance.c
+++ b/drivers/net/ethernet/amd/sunlance.c
@@ -1470,7 +1470,7 @@ no_link_test:
 		goto fail;
 	}
 
-	dev_set_drvdata(&op->dev, lp);
+	platform_set_drvdata(op, lp);
 
 	printk(KERN_INFO "%s: LANCE %pM\n",
 	       dev->name, dev->dev_addr);
@@ -1501,7 +1501,7 @@ static int sunlance_sbus_probe(struct platform_device *op)
 
 static int sunlance_sbus_remove(struct platform_device *op)
 {
-	struct lance_private *lp = dev_get_drvdata(&op->dev);
+	struct lance_private *lp = platform_get_drvdata(op);
 	struct net_device *net_dev = lp->dev;
 
 	unregister_netdev(net_dev);
@@ -1510,8 +1510,6 @@ static int sunlance_sbus_remove(struct platform_device *op)
 
 	free_netdev(net_dev);
 
-	dev_set_drvdata(&op->dev, NULL);
-
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/apple/bmac.c b/drivers/net/ethernet/apple/bmac.c
index f36bbd6..a597b76 100644
--- a/drivers/net/ethernet/apple/bmac.c
+++ b/drivers/net/ethernet/apple/bmac.c
@@ -1016,7 +1016,6 @@ static void bmac_set_multicast(struct net_device *dev)
 static void bmac_set_multicast(struct net_device *dev)
 {
 	struct netdev_hw_addr *ha;
-	int i;
 	unsigned short rx_cfg;
 	u32 crc;
 
@@ -1030,14 +1029,12 @@ static void bmac_set_multicast(struct net_device *dev)
 		rx_cfg |= RxPromiscEnable;
 		bmwrite(dev, RXCFG, rx_cfg);
 	} else {
-		u16 hash_table[4];
+		u16 hash_table[4] = { 0 };
 
 		rx_cfg = bmread(dev, RXCFG);
 		rx_cfg &= ~RxPromiscEnable;
 		bmwrite(dev, RXCFG, rx_cfg);
 
-		for(i = 0; i < 4; i++) hash_table[i] = 0;
-
 		netdev_for_each_mc_addr(ha, dev) {
 			crc = ether_crc_le(6, ha->addr);
 			crc >>= 26;
diff --git a/drivers/net/ethernet/arc/Kconfig b/drivers/net/ethernet/arc/Kconfig
new file mode 100644
index 0000000..514c57f
--- /dev/null
+++ b/drivers/net/ethernet/arc/Kconfig
@@ -0,0 +1,31 @@
+#
+# ARC EMAC network device configuration
+#
+
+config NET_VENDOR_ARC
+	bool "ARC devices"
+	default y
+	---help---
+	  If you have a network (Ethernet) card belonging to this class, say Y
+	  and read the Ethernet-HOWTO, available from
+	  <http://www.tldp.org/docs.html#howto>.
+
+	  Note that the answer to this question doesn't directly affect the
+	  kernel: saying N will just cause the configurator to skip all
+	  the questions about ARC cards. If you say Y, you will be asked for
+	  your specific card in the following questions.
+
+if NET_VENDOR_ARC
+
+config ARC_EMAC
+	tristate "ARC EMAC support"
+	select MII
+	select PHYLIB
+	depends on OF_IRQ
+	depends on OF_NET
+	---help---
+	  On some legacy ARC (Synopsys) FPGA boards such as ARCAngel4/ML50x
+	  non-standard on-chip ethernet device ARC EMAC 10/100 is used.
+	  Say Y here if you have such a board.  If unsure, say N.
+
+endif # NET_VENDOR_ARC
diff --git a/drivers/net/ethernet/arc/Makefile b/drivers/net/ethernet/arc/Makefile
new file mode 100644
index 0000000..00c8657
--- /dev/null
+++ b/drivers/net/ethernet/arc/Makefile
@@ -0,0 +1,6 @@
+#
+# Makefile for the ARC network device drivers.
+#
+
+arc_emac-objs := emac_main.o emac_mdio.o
+obj-$(CONFIG_ARC_EMAC) += arc_emac.o
diff --git a/drivers/net/ethernet/arc/emac.h b/drivers/net/ethernet/arc/emac.h
new file mode 100644
index 0000000..dc08678
--- /dev/null
+++ b/drivers/net/ethernet/arc/emac.h
@@ -0,0 +1,214 @@
+/*
+ * Copyright (C) 2004-2013 Synopsys, Inc. (www.synopsys.com)
+ *
+ * Registers and bits definitions of ARC EMAC
+ */
+
+#ifndef ARC_EMAC_H
+#define ARC_EMAC_H
+
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/netdevice.h>
+#include <linux/phy.h>
+
+/* STATUS and ENABLE Register bit masks */
+#define TXINT_MASK	(1<<0)	/* Transmit interrupt */
+#define RXINT_MASK	(1<<1)	/* Receive interrupt */
+#define ERR_MASK	(1<<2)	/* Error interrupt */
+#define TXCH_MASK	(1<<3)	/* Transmit chaining error interrupt */
+#define MSER_MASK	(1<<4)	/* Missed packet counter error */
+#define RXCR_MASK	(1<<8)	/* RXCRCERR counter rolled over  */
+#define RXFR_MASK	(1<<9)	/* RXFRAMEERR counter rolled over */
+#define RXFL_MASK	(1<<10)	/* RXOFLOWERR counter rolled over */
+#define MDIO_MASK	(1<<12)	/* MDIO complete interrupt */
+#define TXPL_MASK	(1<<31)	/* Force polling of BD by EMAC */
+
+/* CONTROL Register bit masks */
+#define EN_MASK		(1<<0)	/* VMAC enable */
+#define TXRN_MASK	(1<<3)	/* TX enable */
+#define RXRN_MASK	(1<<4)	/* RX enable */
+#define DSBC_MASK	(1<<8)	/* Disable receive broadcast */
+#define ENFL_MASK	(1<<10)	/* Enable Full-duplex */
+#define PROM_MASK	(1<<11)	/* Promiscuous mode */
+
+/* Buffer descriptor INFO bit masks */
+#define OWN_MASK	(1<<31)	/* 0-CPU owns buffer, 1-EMAC owns buffer */
+#define FIRST_MASK	(1<<16)	/* First buffer in chain */
+#define LAST_MASK	(1<<17)	/* Last buffer in chain */
+#define LEN_MASK	0x000007FF	/* last 11 bits */
+#define CRLS		(1<<21)
+#define DEFR		(1<<22)
+#define DROP		(1<<23)
+#define RTRY		(1<<24)
+#define LTCL		(1<<28)
+#define UFLO		(1<<29)
+
+#define FOR_EMAC	OWN_MASK
+#define FOR_CPU		0
+
+/* ARC EMAC register set combines entries for MAC and MDIO */
+enum {
+	R_ID = 0,
+	R_STATUS,
+	R_ENABLE,
+	R_CTRL,
+	R_POLLRATE,
+	R_RXERR,
+	R_MISS,
+	R_TX_RING,
+	R_RX_RING,
+	R_ADDRL,
+	R_ADDRH,
+	R_LAFL,
+	R_LAFH,
+	R_MDIO,
+};
+
+#define TX_TIMEOUT		(400*HZ/1000)	/* Transmission timeout */
+
+#define ARC_EMAC_NAPI_WEIGHT	40		/* Workload for NAPI */
+
+#define EMAC_BUFFER_SIZE	1536		/* EMAC buffer size */
+
+/**
+ * struct arc_emac_bd - EMAC buffer descriptor (BD).
+ *
+ * @info:	Contains status information on the buffer itself.
+ * @data:	32-bit byte addressable pointer to the packet data.
+ */
+struct arc_emac_bd {
+	__le32 info;
+	dma_addr_t data;
+};
+
+/* Number of Rx/Tx BD's */
+#define RX_BD_NUM	128
+#define TX_BD_NUM	128
+
+#define RX_RING_SZ	(RX_BD_NUM * sizeof(struct arc_emac_bd))
+#define TX_RING_SZ	(TX_BD_NUM * sizeof(struct arc_emac_bd))
+
+/**
+ * struct buffer_state - Stores Rx/Tx buffer state.
+ * @sk_buff:	Pointer to socket buffer.
+ * @addr:	Start address of DMA-mapped memory region.
+ * @len:	Length of DMA-mapped memory region.
+ */
+struct buffer_state {
+	struct sk_buff *skb;
+	DEFINE_DMA_UNMAP_ADDR(addr);
+	DEFINE_DMA_UNMAP_LEN(len);
+};
+
+/**
+ * struct arc_emac_priv - Storage of EMAC's private information.
+ * @dev:	Pointer to the current device.
+ * @ndev:	Pointer to the current network device.
+ * @phy_dev:	Pointer to attached PHY device.
+ * @bus:	Pointer to the current MII bus.
+ * @regs:	Base address of EMAC memory-mapped control registers.
+ * @napi:	Structure for NAPI.
+ * @stats:	Network device statistics.
+ * @rxbd:	Pointer to Rx BD ring.
+ * @txbd:	Pointer to Tx BD ring.
+ * @rxbd_dma:	DMA handle for Rx BD ring.
+ * @txbd_dma:	DMA handle for Tx BD ring.
+ * @rx_buff:	Storage for Rx buffers states.
+ * @tx_buff:	Storage for Tx buffers states.
+ * @txbd_curr:	Index of Tx BD to use on the next "ndo_start_xmit".
+ * @txbd_dirty:	Index of Tx BD to free on the next Tx interrupt.
+ * @last_rx_bd:	Index of the last Rx BD we've got from EMAC.
+ * @link:	PHY's last seen link state.
+ * @duplex:	PHY's last set duplex mode.
+ * @speed:	PHY's last set speed.
+ * @max_speed:	Maximum supported by current system network data-rate.
+ */
+struct arc_emac_priv {
+	/* Devices */
+	struct device *dev;
+	struct net_device *ndev;
+	struct phy_device *phy_dev;
+	struct mii_bus *bus;
+
+	void __iomem *regs;
+
+	struct napi_struct napi;
+	struct net_device_stats stats;
+
+	struct arc_emac_bd *rxbd;
+	struct arc_emac_bd *txbd;
+
+	dma_addr_t rxbd_dma;
+	dma_addr_t txbd_dma;
+
+	struct buffer_state rx_buff[RX_BD_NUM];
+	struct buffer_state tx_buff[TX_BD_NUM];
+	unsigned int txbd_curr;
+	unsigned int txbd_dirty;
+
+	unsigned int last_rx_bd;
+
+	unsigned int link;
+	unsigned int duplex;
+	unsigned int speed;
+	unsigned int max_speed;
+};
+
+/**
+ * arc_reg_set - Sets EMAC register with provided value.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ * @value:	Value to set in register.
+ */
+static inline void arc_reg_set(struct arc_emac_priv *priv, int reg, int value)
+{
+	iowrite32(value, priv->regs + reg * sizeof(int));
+}
+
+/**
+ * arc_reg_get - Gets value of specified EMAC register.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ *
+ * returns:	Value of requested register.
+ */
+static inline unsigned int arc_reg_get(struct arc_emac_priv *priv, int reg)
+{
+	return ioread32(priv->regs + reg * sizeof(int));
+}
+
+/**
+ * arc_reg_or - Applies mask to specified EMAC register - ("reg" | "mask").
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ * @mask:	Mask to apply to specified register.
+ *
+ * This function reads initial register value, then applies provided mask
+ * to it and then writes register back.
+ */
+static inline void arc_reg_or(struct arc_emac_priv *priv, int reg, int mask)
+{
+	unsigned int value = arc_reg_get(priv, reg);
+	arc_reg_set(priv, reg, value | mask);
+}
+
+/**
+ * arc_reg_clr - Applies mask to specified EMAC register - ("reg" & ~"mask").
+ * @priv:	Pointer to ARC EMAC private data structure.
+ * @reg:	Register offset from base address.
+ * @mask:	Mask to apply to specified register.
+ *
+ * This function reads initial register value, then applies provided mask
+ * to it and then writes register back.
+ */
+static inline void arc_reg_clr(struct arc_emac_priv *priv, int reg, int mask)
+{
+	unsigned int value = arc_reg_get(priv, reg);
+	arc_reg_set(priv, reg, value & ~mask);
+}
+
+int arc_mdio_probe(struct platform_device *pdev, struct arc_emac_priv *priv);
+int arc_mdio_remove(struct arc_emac_priv *priv);
+
+#endif /* ARC_EMAC_H */
diff --git a/drivers/net/ethernet/arc/emac_main.c b/drivers/net/ethernet/arc/emac_main.c
new file mode 100644
index 0000000..f1b121e
--- /dev/null
+++ b/drivers/net/ethernet/arc/emac_main.c
@@ -0,0 +1,819 @@
+/*
+ * Copyright (C) 2004-2013 Synopsys, Inc. (www.synopsys.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Driver for the ARC EMAC 10100 (hardware revision 5)
+ *
+ * Contributors:
+ *		Amit Bhor
+ *		Sameer Dhavale
+ *		Vineet Gupta
+ */
+
+#include <linux/etherdevice.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_mdio.h>
+#include <linux/of_net.h>
+#include <linux/of_platform.h>
+
+#include "emac.h"
+
+#define DRV_NAME	"arc_emac"
+#define DRV_VERSION	"1.0"
+
+/**
+ * arc_emac_adjust_link - Adjust the PHY link duplex.
+ * @ndev:	Pointer to the net_device structure.
+ *
+ * This function is called to change the duplex setting after auto negotiation
+ * is done by the PHY.
+ */
+static void arc_emac_adjust_link(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct phy_device *phy_dev = priv->phy_dev;
+	unsigned int reg, state_changed = 0;
+
+	if (priv->link != phy_dev->link) {
+		priv->link = phy_dev->link;
+		state_changed = 1;
+	}
+
+	if (priv->speed != phy_dev->speed) {
+		priv->speed = phy_dev->speed;
+		state_changed = 1;
+	}
+
+	if (priv->duplex != phy_dev->duplex) {
+		reg = arc_reg_get(priv, R_CTRL);
+
+		if (DUPLEX_FULL == phy_dev->duplex)
+			reg |= ENFL_MASK;
+		else
+			reg &= ~ENFL_MASK;
+
+		arc_reg_set(priv, R_CTRL, reg);
+		priv->duplex = phy_dev->duplex;
+		state_changed = 1;
+	}
+
+	if (state_changed)
+		phy_print_status(phy_dev);
+}
+
+/**
+ * arc_emac_get_settings - Get PHY settings.
+ * @ndev:	Pointer to net_device structure.
+ * @cmd:	Pointer to ethtool_cmd structure.
+ *
+ * This implements ethtool command for getting PHY settings. If PHY could
+ * not be found, the function returns -ENODEV. This function calls the
+ * relevant PHY ethtool API to get the PHY settings.
+ * Issue "ethtool ethX" under linux prompt to execute this function.
+ */
+static int arc_emac_get_settings(struct net_device *ndev,
+				 struct ethtool_cmd *cmd)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	return phy_ethtool_gset(priv->phy_dev, cmd);
+}
+
+/**
+ * arc_emac_set_settings - Set PHY settings as passed in the argument.
+ * @ndev:	Pointer to net_device structure.
+ * @cmd:	Pointer to ethtool_cmd structure.
+ *
+ * This implements ethtool command for setting various PHY settings. If PHY
+ * could not be found, the function returns -ENODEV. This function calls the
+ * relevant PHY ethtool API to set the PHY.
+ * Issue e.g. "ethtool -s ethX speed 1000" under linux prompt to execute this
+ * function.
+ */
+static int arc_emac_set_settings(struct net_device *ndev,
+				 struct ethtool_cmd *cmd)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EPERM;
+
+	return phy_ethtool_sset(priv->phy_dev, cmd);
+}
+
+/**
+ * arc_emac_get_drvinfo - Get EMAC driver information.
+ * @ndev:	Pointer to net_device structure.
+ * @info:	Pointer to ethtool_drvinfo structure.
+ *
+ * This implements ethtool command for getting the driver information.
+ * Issue "ethtool -i ethX" under linux prompt to execute this function.
+ */
+static void arc_emac_get_drvinfo(struct net_device *ndev,
+				 struct ethtool_drvinfo *info)
+{
+	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
+	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
+}
+
+static const struct ethtool_ops arc_emac_ethtool_ops = {
+	.get_settings	= arc_emac_get_settings,
+	.set_settings	= arc_emac_set_settings,
+	.get_drvinfo	= arc_emac_get_drvinfo,
+	.get_link	= ethtool_op_get_link,
+};
+
+#define FIRST_OR_LAST_MASK	(FIRST_MASK | LAST_MASK)
+
+/**
+ * arc_emac_tx_clean - clears processed by EMAC Tx BDs.
+ * @ndev:	Pointer to the network device.
+ */
+static void arc_emac_tx_clean(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct net_device_stats *stats = &priv->stats;
+	unsigned int i;
+
+	for (i = 0; i < TX_BD_NUM; i++) {
+		unsigned int *txbd_dirty = &priv->txbd_dirty;
+		struct arc_emac_bd *txbd = &priv->txbd[*txbd_dirty];
+		struct buffer_state *tx_buff = &priv->tx_buff[*txbd_dirty];
+		struct sk_buff *skb = tx_buff->skb;
+		unsigned int info = le32_to_cpu(txbd->info);
+
+		*txbd_dirty = (*txbd_dirty + 1) % TX_BD_NUM;
+
+		if ((info & FOR_EMAC) || !txbd->data)
+			break;
+
+		if (unlikely(info & (DROP | DEFR | LTCL | UFLO))) {
+			stats->tx_errors++;
+			stats->tx_dropped++;
+
+			if (info & DEFR)
+				stats->tx_carrier_errors++;
+
+			if (info & LTCL)
+				stats->collisions++;
+
+			if (info & UFLO)
+				stats->tx_fifo_errors++;
+		} else if (likely(info & FIRST_OR_LAST_MASK)) {
+			stats->tx_packets++;
+			stats->tx_bytes += skb->len;
+		}
+
+		dma_unmap_single(&ndev->dev, dma_unmap_addr(tx_buff, addr),
+				 dma_unmap_len(tx_buff, len), DMA_TO_DEVICE);
+
+		/* return the sk_buff to system */
+		dev_kfree_skb_irq(skb);
+
+		txbd->data = 0;
+		txbd->info = 0;
+
+		if (netif_queue_stopped(ndev))
+			netif_wake_queue(ndev);
+	}
+}
+
+/**
+ * arc_emac_rx - processing of Rx packets.
+ * @ndev:	Pointer to the network device.
+ * @budget:	How many BDs to process on 1 call.
+ *
+ * returns:	Number of processed BDs
+ *
+ * Iterate through Rx BDs and deliver received packages to upper layer.
+ */
+static int arc_emac_rx(struct net_device *ndev, int budget)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	unsigned int work_done;
+
+	for (work_done = 0; work_done <= budget; work_done++) {
+		unsigned int *last_rx_bd = &priv->last_rx_bd;
+		struct net_device_stats *stats = &priv->stats;
+		struct buffer_state *rx_buff = &priv->rx_buff[*last_rx_bd];
+		struct arc_emac_bd *rxbd = &priv->rxbd[*last_rx_bd];
+		unsigned int pktlen, info = le32_to_cpu(rxbd->info);
+		struct sk_buff *skb;
+		dma_addr_t addr;
+
+		if (unlikely((info & OWN_MASK) == FOR_EMAC))
+			break;
+
+		/* Make a note that we saw a packet at this BD.
+		 * So next time, driver starts from this + 1
+		 */
+		*last_rx_bd = (*last_rx_bd + 1) % RX_BD_NUM;
+
+		if (unlikely((info & FIRST_OR_LAST_MASK) !=
+			     FIRST_OR_LAST_MASK)) {
+			/* We pre-allocate buffers of MTU size so incoming
+			 * packets won't be split/chained.
+			 */
+			if (net_ratelimit())
+				netdev_err(ndev, "incomplete packet received\n");
+
+			/* Return ownership to EMAC */
+			rxbd->info = cpu_to_le32(FOR_EMAC | EMAC_BUFFER_SIZE);
+			stats->rx_errors++;
+			stats->rx_length_errors++;
+			continue;
+		}
+
+		pktlen = info & LEN_MASK;
+		stats->rx_packets++;
+		stats->rx_bytes += pktlen;
+		skb = rx_buff->skb;
+		skb_put(skb, pktlen);
+		skb->dev = ndev;
+		skb->protocol = eth_type_trans(skb, ndev);
+
+		dma_unmap_single(&ndev->dev, dma_unmap_addr(rx_buff, addr),
+				 dma_unmap_len(rx_buff, len), DMA_FROM_DEVICE);
+
+		/* Prepare the BD for next cycle */
+		rx_buff->skb = netdev_alloc_skb_ip_align(ndev,
+							 EMAC_BUFFER_SIZE);
+		if (unlikely(!rx_buff->skb)) {
+			stats->rx_errors++;
+			/* Because receive_skb is below, increment rx_dropped */
+			stats->rx_dropped++;
+			continue;
+		}
+
+		/* receive_skb only if new skb was allocated to avoid holes */
+		netif_receive_skb(skb);
+
+		addr = dma_map_single(&ndev->dev, (void *)rx_buff->skb->data,
+				      EMAC_BUFFER_SIZE, DMA_FROM_DEVICE);
+		if (dma_mapping_error(&ndev->dev, addr)) {
+			if (net_ratelimit())
+				netdev_err(ndev, "cannot dma map\n");
+			dev_kfree_skb(rx_buff->skb);
+			stats->rx_errors++;
+			continue;
+		}
+		dma_unmap_addr_set(rx_buff, addr, addr);
+		dma_unmap_len_set(rx_buff, len, EMAC_BUFFER_SIZE);
+
+		rxbd->data = cpu_to_le32(addr);
+
+		/* Make sure pointer to data buffer is set */
+		wmb();
+
+		/* Return ownership to EMAC */
+		rxbd->info = cpu_to_le32(FOR_EMAC | EMAC_BUFFER_SIZE);
+	}
+
+	return work_done;
+}
+
+/**
+ * arc_emac_poll - NAPI poll handler.
+ * @napi:	Pointer to napi_struct structure.
+ * @budget:	How many BDs to process on 1 call.
+ *
+ * returns:	Number of processed BDs
+ */
+static int arc_emac_poll(struct napi_struct *napi, int budget)
+{
+	struct net_device *ndev = napi->dev;
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	unsigned int work_done;
+
+	arc_emac_tx_clean(ndev);
+
+	work_done = arc_emac_rx(ndev, budget);
+	if (work_done < budget) {
+		napi_complete(napi);
+		arc_reg_or(priv, R_ENABLE, RXINT_MASK);
+	}
+
+	return work_done;
+}
+
+/**
+ * arc_emac_intr - Global interrupt handler for EMAC.
+ * @irq:		irq number.
+ * @dev_instance:	device instance.
+ *
+ * returns: IRQ_HANDLED for all cases.
+ *
+ * ARC EMAC has only 1 interrupt line, and depending on bits raised in
+ * STATUS register we may tell what is a reason for interrupt to fire.
+ */
+static irqreturn_t arc_emac_intr(int irq, void *dev_instance)
+{
+	struct net_device *ndev = dev_instance;
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct net_device_stats *stats = &priv->stats;
+	unsigned int status;
+
+	status = arc_reg_get(priv, R_STATUS);
+	status &= ~MDIO_MASK;
+
+	/* Reset all flags except "MDIO complete" */
+	arc_reg_set(priv, R_STATUS, status);
+
+	if (status & RXINT_MASK) {
+		if (likely(napi_schedule_prep(&priv->napi))) {
+			arc_reg_clr(priv, R_ENABLE, RXINT_MASK);
+			__napi_schedule(&priv->napi);
+		}
+	}
+
+	if (status & ERR_MASK) {
+		/* MSER/RXCR/RXFR/RXFL interrupt fires on corresponding
+		 * 8-bit error counter overrun.
+		 */
+
+		if (status & MSER_MASK) {
+			stats->rx_missed_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+
+		if (status & RXCR_MASK) {
+			stats->rx_crc_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+
+		if (status & RXFR_MASK) {
+			stats->rx_frame_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+
+		if (status & RXFL_MASK) {
+			stats->rx_over_errors += 0x100;
+			stats->rx_errors += 0x100;
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+/**
+ * arc_emac_open - Open the network device.
+ * @ndev:	Pointer to the network device.
+ *
+ * returns: 0, on success or non-zero error value on failure.
+ *
+ * This function sets the MAC address, requests and enables an IRQ
+ * for the EMAC device and starts the Tx queue.
+ * It also connects to the phy device.
+ */
+static int arc_emac_open(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct phy_device *phy_dev = priv->phy_dev;
+	int i;
+
+	phy_dev->autoneg = AUTONEG_ENABLE;
+	phy_dev->speed = 0;
+	phy_dev->duplex = 0;
+	phy_dev->advertising = phy_dev->supported;
+
+	if (priv->max_speed > 100) {
+		phy_dev->advertising &= PHY_GBIT_FEATURES;
+	} else if (priv->max_speed <= 100) {
+		phy_dev->advertising &= PHY_BASIC_FEATURES;
+		if (priv->max_speed <= 10) {
+			phy_dev->advertising &= ~SUPPORTED_100baseT_Half;
+			phy_dev->advertising &= ~SUPPORTED_100baseT_Full;
+		}
+	}
+
+	priv->last_rx_bd = 0;
+
+	/* Allocate and set buffers for Rx BD's */
+	for (i = 0; i < RX_BD_NUM; i++) {
+		dma_addr_t addr;
+		unsigned int *last_rx_bd = &priv->last_rx_bd;
+		struct arc_emac_bd *rxbd = &priv->rxbd[*last_rx_bd];
+		struct buffer_state *rx_buff = &priv->rx_buff[*last_rx_bd];
+
+		rx_buff->skb = netdev_alloc_skb_ip_align(ndev,
+							 EMAC_BUFFER_SIZE);
+		if (unlikely(!rx_buff->skb))
+			return -ENOMEM;
+
+		addr = dma_map_single(&ndev->dev, (void *)rx_buff->skb->data,
+				      EMAC_BUFFER_SIZE, DMA_FROM_DEVICE);
+		if (dma_mapping_error(&ndev->dev, addr)) {
+			netdev_err(ndev, "cannot dma map\n");
+			dev_kfree_skb(rx_buff->skb);
+			return -ENOMEM;
+		}
+		dma_unmap_addr_set(rx_buff, addr, addr);
+		dma_unmap_len_set(rx_buff, len, EMAC_BUFFER_SIZE);
+
+		rxbd->data = cpu_to_le32(addr);
+
+		/* Make sure pointer to data buffer is set */
+		wmb();
+
+		/* Return ownership to EMAC */
+		rxbd->info = cpu_to_le32(FOR_EMAC | EMAC_BUFFER_SIZE);
+
+		*last_rx_bd = (*last_rx_bd + 1) % RX_BD_NUM;
+	}
+
+	/* Clean Tx BD's */
+	memset(priv->txbd, 0, TX_RING_SZ);
+
+	/* Initialize logical address filter */
+	arc_reg_set(priv, R_LAFL, 0);
+	arc_reg_set(priv, R_LAFH, 0);
+
+	/* Set BD ring pointers for device side */
+	arc_reg_set(priv, R_RX_RING, (unsigned int)priv->rxbd_dma);
+	arc_reg_set(priv, R_TX_RING, (unsigned int)priv->txbd_dma);
+
+	/* Enable interrupts */
+	arc_reg_set(priv, R_ENABLE, RXINT_MASK | ERR_MASK);
+
+	/* Set CONTROL */
+	arc_reg_set(priv, R_CTRL,
+		     (RX_BD_NUM << 24) |	/* RX BD table length */
+		     (TX_BD_NUM << 16) |	/* TX BD table length */
+		     TXRN_MASK | RXRN_MASK);
+
+	napi_enable(&priv->napi);
+
+	/* Enable EMAC */
+	arc_reg_or(priv, R_CTRL, EN_MASK);
+
+	phy_start_aneg(priv->phy_dev);
+
+	netif_start_queue(ndev);
+
+	return 0;
+}
+
+/**
+ * arc_emac_stop - Close the network device.
+ * @ndev:	Pointer to the network device.
+ *
+ * This function stops the Tx queue, disables interrupts and frees the IRQ for
+ * the EMAC device.
+ * It also disconnects the PHY device associated with the EMAC device.
+ */
+static int arc_emac_stop(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	napi_disable(&priv->napi);
+	netif_stop_queue(ndev);
+
+	/* Disable interrupts */
+	arc_reg_clr(priv, R_ENABLE, RXINT_MASK | ERR_MASK);
+
+	/* Disable EMAC */
+	arc_reg_clr(priv, R_CTRL, EN_MASK);
+
+	return 0;
+}
+
+/**
+ * arc_emac_stats - Get system network statistics.
+ * @ndev:	Pointer to net_device structure.
+ *
+ * Returns the address of the device statistics structure.
+ * Statistics are updated in interrupt handler.
+ */
+static struct net_device_stats *arc_emac_stats(struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct net_device_stats *stats = &priv->stats;
+	unsigned long miss, rxerr;
+	u8 rxcrc, rxfram, rxoflow;
+
+	rxerr = arc_reg_get(priv, R_RXERR);
+	miss = arc_reg_get(priv, R_MISS);
+
+	rxcrc = rxerr;
+	rxfram = rxerr >> 8;
+	rxoflow = rxerr >> 16;
+
+	stats->rx_errors += miss;
+	stats->rx_errors += rxcrc + rxfram + rxoflow;
+
+	stats->rx_over_errors += rxoflow;
+	stats->rx_frame_errors += rxfram;
+	stats->rx_crc_errors += rxcrc;
+	stats->rx_missed_errors += miss;
+
+	return stats;
+}
+
+/**
+ * arc_emac_tx - Starts the data transmission.
+ * @skb:	sk_buff pointer that contains data to be Transmitted.
+ * @ndev:	Pointer to net_device structure.
+ *
+ * returns: NETDEV_TX_OK, on success
+ *		NETDEV_TX_BUSY, if any of the descriptors are not free.
+ *
+ * This function is invoked from upper layers to initiate transmission.
+ */
+static int arc_emac_tx(struct sk_buff *skb, struct net_device *ndev)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	unsigned int len, *txbd_curr = &priv->txbd_curr;
+	struct net_device_stats *stats = &priv->stats;
+	__le32 *info = &priv->txbd[*txbd_curr].info;
+	dma_addr_t addr;
+
+	if (skb_padto(skb, ETH_ZLEN))
+		return NETDEV_TX_OK;
+
+	len = max_t(unsigned int, ETH_ZLEN, skb->len);
+
+	/* EMAC still holds this buffer in its possession.
+	 * CPU must not modify this buffer descriptor
+	 */
+	if (unlikely((le32_to_cpu(*info) & OWN_MASK) == FOR_EMAC)) {
+		netif_stop_queue(ndev);
+		return NETDEV_TX_BUSY;
+	}
+
+	addr = dma_map_single(&ndev->dev, (void *)skb->data, len,
+			      DMA_TO_DEVICE);
+
+	if (unlikely(dma_mapping_error(&ndev->dev, addr))) {
+		stats->tx_dropped++;
+		stats->tx_errors++;
+		dev_kfree_skb(skb);
+		return NETDEV_TX_OK;
+	}
+	dma_unmap_addr_set(&priv->tx_buff[*txbd_curr], addr, addr);
+	dma_unmap_len_set(&priv->tx_buff[*txbd_curr], len, len);
+
+	priv->tx_buff[*txbd_curr].skb = skb;
+	priv->txbd[*txbd_curr].data = cpu_to_le32(addr);
+
+	/* Make sure pointer to data buffer is set */
+	wmb();
+
+	*info = cpu_to_le32(FOR_EMAC | FIRST_OR_LAST_MASK | len);
+
+	/* Increment index to point to the next BD */
+	*txbd_curr = (*txbd_curr + 1) % TX_BD_NUM;
+
+	/* Get "info" of the next BD */
+	info = &priv->txbd[*txbd_curr].info;
+
+	/* Check if if Tx BD ring is full - next BD is still owned by EMAC */
+	if (unlikely((le32_to_cpu(*info) & OWN_MASK) == FOR_EMAC))
+		netif_stop_queue(ndev);
+
+	arc_reg_set(priv, R_STATUS, TXPL_MASK);
+
+	skb_tx_timestamp(skb);
+
+	return NETDEV_TX_OK;
+}
+
+/**
+ * arc_emac_set_address - Set the MAC address for this device.
+ * @ndev:	Pointer to net_device structure.
+ * @p:		6 byte Address to be written as MAC address.
+ *
+ * This function copies the HW address from the sockaddr structure to the
+ * net_device structure and updates the address in HW.
+ *
+ * returns:	-EBUSY if the net device is busy or 0 if the address is set
+ *		successfully.
+ */
+static int arc_emac_set_address(struct net_device *ndev, void *p)
+{
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+	struct sockaddr *addr = p;
+	unsigned int addr_low, addr_hi;
+
+	if (netif_running(ndev))
+		return -EBUSY;
+
+	if (!is_valid_ether_addr(addr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	memcpy(ndev->dev_addr, addr->sa_data, ndev->addr_len);
+
+	addr_low = le32_to_cpu(*(__le32 *) &ndev->dev_addr[0]);
+	addr_hi = le16_to_cpu(*(__le16 *) &ndev->dev_addr[4]);
+
+	arc_reg_set(priv, R_ADDRL, addr_low);
+	arc_reg_set(priv, R_ADDRH, addr_hi);
+
+	return 0;
+}
+
+static const struct net_device_ops arc_emac_netdev_ops = {
+	.ndo_open		= arc_emac_open,
+	.ndo_stop		= arc_emac_stop,
+	.ndo_start_xmit		= arc_emac_tx,
+	.ndo_set_mac_address	= arc_emac_set_address,
+	.ndo_get_stats		= arc_emac_stats,
+};
+
+static int arc_emac_probe(struct platform_device *pdev)
+{
+	struct resource res_regs, res_irq;
+	struct device_node *phy_node;
+	struct arc_emac_priv *priv;
+	struct net_device *ndev;
+	const char *mac_addr;
+	unsigned int id, clock_frequency;
+	int err;
+
+	if (!pdev->dev.of_node)
+		return -ENODEV;
+
+	/* Get PHY from device tree */
+	phy_node = of_parse_phandle(pdev->dev.of_node, "phy", 0);
+	if (!phy_node) {
+		dev_err(&pdev->dev, "failed to retrieve phy description from device tree\n");
+		return -ENODEV;
+	}
+
+	/* Get EMAC registers base address from device tree */
+	err = of_address_to_resource(pdev->dev.of_node, 0, &res_regs);
+	if (err) {
+		dev_err(&pdev->dev, "failed to retrieve registers base from device tree\n");
+		return -ENODEV;
+	}
+
+	/* Get CPU clock frequency from device tree */
+	if (of_property_read_u32(pdev->dev.of_node, "clock-frequency",
+				 &clock_frequency)) {
+		dev_err(&pdev->dev, "failed to retrieve <clock-frequency> from device tree\n");
+		return -EINVAL;
+	}
+
+	/* Get IRQ from device tree */
+	err = of_irq_to_resource(pdev->dev.of_node, 0, &res_irq);
+	if (!err) {
+		dev_err(&pdev->dev, "failed to retrieve <irq> value from device tree\n");
+		return -ENODEV;
+	}
+
+	ndev = alloc_etherdev(sizeof(struct arc_emac_priv));
+	if (!ndev)
+		return -ENOMEM;
+
+	SET_NETDEV_DEV(ndev, &pdev->dev);
+
+	ndev->netdev_ops = &arc_emac_netdev_ops;
+	ndev->ethtool_ops = &arc_emac_ethtool_ops;
+	ndev->watchdog_timeo = TX_TIMEOUT;
+	/* FIXME :: no multicast support yet */
+	ndev->flags &= ~IFF_MULTICAST;
+
+	priv = netdev_priv(ndev);
+	priv->dev = &pdev->dev;
+	priv->ndev = ndev;
+
+	priv->regs = devm_ioremap_resource(&pdev->dev, &res_regs);
+	if (IS_ERR(priv->regs)) {
+		err = PTR_ERR(priv->regs);
+		goto out;
+	}
+	dev_dbg(&pdev->dev, "Registers base address is 0x%p\n", priv->regs);
+
+	id = arc_reg_get(priv, R_ID);
+
+	/* Check for EMAC revision 5 or 7, magic number */
+	if (!(id == 0x0005fd02 || id == 0x0007fd02)) {
+		dev_err(&pdev->dev, "ARC EMAC not detected, id=0x%x\n", id);
+		err = -ENODEV;
+		goto out;
+	}
+	dev_info(&pdev->dev, "ARC EMAC detected with id: 0x%x\n", id);
+
+	/* Set poll rate so that it polls every 1 ms */
+	arc_reg_set(priv, R_POLLRATE, clock_frequency / 1000000);
+
+	/* Get max speed of operation from device tree */
+	if (of_property_read_u32(pdev->dev.of_node, "max-speed",
+				 &priv->max_speed)) {
+		dev_err(&pdev->dev, "failed to retrieve <max-speed> from device tree\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	ndev->irq = res_irq.start;
+	dev_info(&pdev->dev, "IRQ is %d\n", ndev->irq);
+
+	/* Register interrupt handler for device */
+	err = devm_request_irq(&pdev->dev, ndev->irq, arc_emac_intr, 0,
+			       ndev->name, ndev);
+	if (err) {
+		dev_err(&pdev->dev, "could not allocate IRQ\n");
+		goto out;
+	}
+
+	/* Get MAC address from device tree */
+	mac_addr = of_get_mac_address(pdev->dev.of_node);
+
+	if (!mac_addr || !is_valid_ether_addr(mac_addr))
+		eth_hw_addr_random(ndev);
+	else
+		memcpy(ndev->dev_addr, mac_addr, ETH_ALEN);
+
+	dev_info(&pdev->dev, "MAC address is now %pM\n", ndev->dev_addr);
+
+	/* Do 1 allocation instead of 2 separate ones for Rx and Tx BD rings */
+	priv->rxbd = dmam_alloc_coherent(&pdev->dev, RX_RING_SZ + TX_RING_SZ,
+					 &priv->rxbd_dma, GFP_KERNEL);
+
+	if (!priv->rxbd) {
+		dev_err(&pdev->dev, "failed to allocate data buffers\n");
+		err = -ENOMEM;
+		goto out;
+	}
+
+	priv->txbd = priv->rxbd + RX_BD_NUM;
+
+	priv->txbd_dma = priv->rxbd_dma + RX_RING_SZ;
+	dev_dbg(&pdev->dev, "EMAC Device addr: Rx Ring [0x%x], Tx Ring[%x]\n",
+		(unsigned int)priv->rxbd_dma, (unsigned int)priv->txbd_dma);
+
+	err = arc_mdio_probe(pdev, priv);
+	if (err) {
+		dev_err(&pdev->dev, "failed to probe MII bus\n");
+		goto out;
+	}
+
+	priv->phy_dev = of_phy_connect(ndev, phy_node, arc_emac_adjust_link, 0,
+				       PHY_INTERFACE_MODE_MII);
+	if (!priv->phy_dev) {
+		dev_err(&pdev->dev, "of_phy_connect() failed\n");
+		err = -ENODEV;
+		goto out;
+	}
+
+	dev_info(&pdev->dev, "connected to %s phy with id 0x%x\n",
+		 priv->phy_dev->drv->name, priv->phy_dev->phy_id);
+
+	netif_napi_add(ndev, &priv->napi, arc_emac_poll, ARC_EMAC_NAPI_WEIGHT);
+
+	err = register_netdev(ndev);
+	if (err) {
+		netif_napi_del(&priv->napi);
+		dev_err(&pdev->dev, "failed to register network device\n");
+		goto out;
+	}
+
+	return 0;
+
+out:
+	free_netdev(ndev);
+	return err;
+}
+
+static int arc_emac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct arc_emac_priv *priv = netdev_priv(ndev);
+
+	phy_disconnect(priv->phy_dev);
+	priv->phy_dev = NULL;
+	arc_mdio_remove(priv);
+	unregister_netdev(ndev);
+	netif_napi_del(&priv->napi);
+	free_netdev(ndev);
+
+	return 0;
+}
+
+static const struct of_device_id arc_emac_dt_ids[] = {
+	{ .compatible = "snps,arc-emac" },
+	{ /* Sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, arc_emac_dt_ids);
+
+static struct platform_driver arc_emac_driver = {
+	.probe = arc_emac_probe,
+	.remove = arc_emac_remove,
+	.driver = {
+		.name = DRV_NAME,
+		.owner = THIS_MODULE,
+		.of_match_table  = arc_emac_dt_ids,
+		},
+};
+
+module_platform_driver(arc_emac_driver);
+
+MODULE_AUTHOR("Alexey Brodkin <abrodkin@synopsys.com>");
+MODULE_DESCRIPTION("ARC EMAC driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/arc/emac_mdio.c b/drivers/net/ethernet/arc/emac_mdio.c
new file mode 100644
index 0000000..26ba242
--- /dev/null
+++ b/drivers/net/ethernet/arc/emac_mdio.c
@@ -0,0 +1,152 @@
+/*
+ * Copyright (C) 2004-2013 Synopsys, Inc. (www.synopsys.com)
+ *
+ * MDIO implementation for ARC EMAC
+ */
+
+#include <linux/delay.h>
+#include <linux/of_mdio.h>
+#include <linux/platform_device.h>
+
+#include "emac.h"
+
+/* Number of seconds we wait for "MDIO complete" flag to appear */
+#define ARC_MDIO_COMPLETE_POLL_COUNT	1
+
+/**
+ * arc_mdio_complete_wait - Waits until MDIO transaction is completed.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ *
+ * returns:	0 on success, -ETIMEDOUT on a timeout.
+ */
+static int arc_mdio_complete_wait(struct arc_emac_priv *priv)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARC_MDIO_COMPLETE_POLL_COUNT * 40; i++) {
+		unsigned int status = arc_reg_get(priv, R_STATUS);
+
+		status &= MDIO_MASK;
+
+		if (status) {
+			/* Reset "MDIO complete" flag */
+			arc_reg_set(priv, R_STATUS, status);
+			return 0;
+		}
+
+		msleep(25);
+	}
+
+	return -ETIMEDOUT;
+}
+
+/**
+ * arc_mdio_read - MDIO interface read function.
+ * @bus:	Pointer to MII bus structure.
+ * @phy_addr:	Address of the PHY device.
+ * @reg_num:	PHY register to read.
+ *
+ * returns:	The register contents on success, -ETIMEDOUT on a timeout.
+ *
+ * Reads the contents of the requested register from the requested PHY
+ * address.
+ */
+static int arc_mdio_read(struct mii_bus *bus, int phy_addr, int reg_num)
+{
+	struct arc_emac_priv *priv = bus->priv;
+	unsigned int value;
+	int error;
+
+	arc_reg_set(priv, R_MDIO,
+		    0x60020000 | (phy_addr << 23) | (reg_num << 18));
+
+	error = arc_mdio_complete_wait(priv);
+	if (error < 0)
+		return error;
+
+	value = arc_reg_get(priv, R_MDIO) & 0xffff;
+
+	dev_dbg(priv->dev, "arc_mdio_read(phy_addr=%i, reg_num=%x) = %x\n",
+		phy_addr, reg_num, value);
+
+	return value;
+}
+
+/**
+ * arc_mdio_write - MDIO interface write function.
+ * @bus:	Pointer to MII bus structure.
+ * @phy_addr:	Address of the PHY device.
+ * @reg_num:	PHY register to write to.
+ * @value:	Value to be written into the register.
+ *
+ * returns:	0 on success, -ETIMEDOUT on a timeout.
+ *
+ * Writes the value to the requested register.
+ */
+static int arc_mdio_write(struct mii_bus *bus, int phy_addr,
+			  int reg_num, u16 value)
+{
+	struct arc_emac_priv *priv = bus->priv;
+
+	dev_dbg(priv->dev,
+		"arc_mdio_write(phy_addr=%i, reg_num=%x, value=%x)\n",
+		phy_addr, reg_num, value);
+
+	arc_reg_set(priv, R_MDIO,
+		     0x50020000 | (phy_addr << 23) | (reg_num << 18) | value);
+
+	return arc_mdio_complete_wait(priv);
+}
+
+/**
+ * arc_mdio_probe - MDIO probe function.
+ * @pdev:	Pointer to platform device.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ *
+ * returns:	0 on success, -ENOMEM when mdiobus_alloc
+ * (to allocate memory for MII bus structure) fails.
+ *
+ * Sets up and registers the MDIO interface.
+ */
+int arc_mdio_probe(struct platform_device *pdev, struct arc_emac_priv *priv)
+{
+	struct mii_bus *bus;
+	int error;
+
+	bus = mdiobus_alloc();
+	if (!bus)
+		return -ENOMEM;
+
+	priv->bus = bus;
+	bus->priv = priv;
+	bus->parent = priv->dev;
+	bus->name = "Synopsys MII Bus",
+	bus->read = &arc_mdio_read;
+	bus->write = &arc_mdio_write;
+
+	snprintf(bus->id, MII_BUS_ID_SIZE, "%s", pdev->name);
+
+	error = of_mdiobus_register(bus, pdev->dev.of_node);
+	if (error) {
+		dev_err(priv->dev, "cannot register MDIO bus %s\n", bus->name);
+		mdiobus_free(bus);
+		return error;
+	}
+
+	return 0;
+}
+
+/**
+ * arc_mdio_remove - MDIO remove function.
+ * @priv:	Pointer to ARC EMAC private data structure.
+ *
+ * Unregisters the MDIO and frees any associate memory for MII bus.
+ */
+int arc_mdio_remove(struct arc_emac_priv *priv)
+{
+	mdiobus_unregister(priv->bus);
+	mdiobus_free(priv->bus);
+	priv->bus = NULL;
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/atheros/Kconfig b/drivers/net/ethernet/atheros/Kconfig
index ad6aa1e9..58ad37c 100644
--- a/drivers/net/ethernet/atheros/Kconfig
+++ b/drivers/net/ethernet/atheros/Kconfig
@@ -22,7 +22,6 @@ config ATL2
 	tristate "Atheros L2 Fast Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports the Atheros L2 fast ethernet adapter.
@@ -34,7 +33,6 @@ config ATL1
 	tristate "Atheros/Attansic L1 Gigabit Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports the Atheros/Attansic L1 gigabit ethernet
@@ -47,7 +45,6 @@ config ATL1E
 	tristate "Atheros L1E Gigabit Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports the Atheros L1E gigabit ethernet adapter.
@@ -59,7 +56,6 @@ config ATL1C
 	tristate "Atheros L1C Gigabit Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports the Atheros L1C gigabit ethernet adapter.
@@ -71,7 +67,6 @@ config ALX
 	tristate "Qualcomm Atheros AR816x/AR817x support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MDIO
 	help
 	  This driver supports the Qualcomm Atheros L1F ethernet adapter,
diff --git a/drivers/net/ethernet/atheros/alx/alx.h b/drivers/net/ethernet/atheros/alx/alx.h
index 50b3ae2..d71103d 100644
--- a/drivers/net/ethernet/atheros/alx/alx.h
+++ b/drivers/net/ethernet/atheros/alx/alx.h
@@ -85,16 +85,16 @@ struct alx_priv {
 	struct {
 		dma_addr_t dma;
 		void *virt;
-		int size;
+		unsigned int size;
 	} descmem;
 
 	/* protect int_mask updates */
 	spinlock_t irq_lock;
 	u32 int_mask;
 
-	int tx_ringsz;
-	int rx_ringsz;
-	int rxbuf_size;
+	unsigned int tx_ringsz;
+	unsigned int rx_ringsz;
+	unsigned int rxbuf_size;
 
 	struct napi_struct napi;
 	struct alx_tx_queue txq;
diff --git a/drivers/net/ethernet/atheros/alx/ethtool.c b/drivers/net/ethernet/atheros/alx/ethtool.c
index 6fa2aec..45b3650 100644
--- a/drivers/net/ethernet/atheros/alx/ethtool.c
+++ b/drivers/net/ethernet/atheros/alx/ethtool.c
@@ -46,21 +46,37 @@
 #include "reg.h"
 #include "hw.h"
 
+static u32 alx_get_supported_speeds(struct alx_hw *hw)
+{
+	u32 supported = SUPPORTED_10baseT_Half |
+			SUPPORTED_10baseT_Full |
+			SUPPORTED_100baseT_Half |
+			SUPPORTED_100baseT_Full;
+
+	if (alx_hw_giga(hw))
+		supported |= SUPPORTED_1000baseT_Full;
+
+	BUILD_BUG_ON(SUPPORTED_10baseT_Half != ADVERTISED_10baseT_Half);
+	BUILD_BUG_ON(SUPPORTED_10baseT_Full != ADVERTISED_10baseT_Full);
+	BUILD_BUG_ON(SUPPORTED_100baseT_Half != ADVERTISED_100baseT_Half);
+	BUILD_BUG_ON(SUPPORTED_100baseT_Full != ADVERTISED_100baseT_Full);
+	BUILD_BUG_ON(SUPPORTED_1000baseT_Full != ADVERTISED_1000baseT_Full);
+
+	return supported;
+}
 
 static int alx_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 {
 	struct alx_priv *alx = netdev_priv(netdev);
 	struct alx_hw *hw = &alx->hw;
 
-	ecmd->supported = SUPPORTED_10baseT_Half |
-			  SUPPORTED_10baseT_Full |
-			  SUPPORTED_100baseT_Half |
-			  SUPPORTED_100baseT_Full |
-			  SUPPORTED_Autoneg |
+	ecmd->supported = SUPPORTED_Autoneg |
 			  SUPPORTED_TP |
-			  SUPPORTED_Pause;
+			  SUPPORTED_Pause |
+			  SUPPORTED_Asym_Pause;
 	if (alx_hw_giga(hw))
 		ecmd->supported |= SUPPORTED_1000baseT_Full;
+	ecmd->supported |= alx_get_supported_speeds(hw);
 
 	ecmd->advertising = ADVERTISED_TP;
 	if (hw->adv_cfg & ADVERTISED_Autoneg)
@@ -68,6 +84,7 @@ static int alx_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 
 	ecmd->port = PORT_TP;
 	ecmd->phy_address = 0;
+
 	if (hw->adv_cfg & ADVERTISED_Autoneg)
 		ecmd->autoneg = AUTONEG_ENABLE;
 	else
@@ -85,14 +102,8 @@ static int alx_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 		}
 	}
 
-	if (hw->link_speed != SPEED_UNKNOWN) {
-		ethtool_cmd_speed_set(ecmd,
-				      hw->link_speed - hw->link_speed % 10);
-		ecmd->duplex = hw->link_speed % 10;
-	} else {
-		ethtool_cmd_speed_set(ecmd, SPEED_UNKNOWN);
-		ecmd->duplex = DUPLEX_UNKNOWN;
-	}
+	ethtool_cmd_speed_set(ecmd, hw->link_speed);
+	ecmd->duplex = hw->duplex;
 
 	return 0;
 }
@@ -106,28 +117,15 @@ static int alx_set_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 	ASSERT_RTNL();
 
 	if (ecmd->autoneg == AUTONEG_ENABLE) {
-		if (ecmd->advertising & ADVERTISED_1000baseT_Half)
+		if (ecmd->advertising & ~alx_get_supported_speeds(hw))
 			return -EINVAL;
 		adv_cfg = ecmd->advertising | ADVERTISED_Autoneg;
 	} else {
-		int speed = ethtool_cmd_speed(ecmd);
-
-		switch (speed + ecmd->duplex) {
-		case SPEED_10 + DUPLEX_HALF:
-			adv_cfg = ADVERTISED_10baseT_Half;
-			break;
-		case SPEED_10 + DUPLEX_FULL:
-			adv_cfg = ADVERTISED_10baseT_Full;
-			break;
-		case SPEED_100 + DUPLEX_HALF:
-			adv_cfg = ADVERTISED_100baseT_Half;
-			break;
-		case SPEED_100 + DUPLEX_FULL:
-			adv_cfg = ADVERTISED_100baseT_Full;
-			break;
-		default:
+		adv_cfg = alx_speed_to_ethadv(ethtool_cmd_speed(ecmd),
+					      ecmd->duplex);
+
+		if (!adv_cfg || adv_cfg == ADVERTISED_1000baseT_Full)
 			return -EINVAL;
-		}
 	}
 
 	hw->adv_cfg = adv_cfg;
@@ -140,21 +138,10 @@ static void alx_get_pauseparam(struct net_device *netdev,
 	struct alx_priv *alx = netdev_priv(netdev);
 	struct alx_hw *hw = &alx->hw;
 
-	if (hw->flowctrl & ALX_FC_ANEG &&
-	    hw->adv_cfg & ADVERTISED_Autoneg)
-		pause->autoneg = AUTONEG_ENABLE;
-	else
-		pause->autoneg = AUTONEG_DISABLE;
-
-	if (hw->flowctrl & ALX_FC_TX)
-		pause->tx_pause = 1;
-	else
-		pause->tx_pause = 0;
-
-	if (hw->flowctrl & ALX_FC_RX)
-		pause->rx_pause = 1;
-	else
-		pause->rx_pause = 0;
+	pause->autoneg = !!(hw->flowctrl & ALX_FC_ANEG &&
+			    hw->adv_cfg & ADVERTISED_Autoneg);
+	pause->tx_pause = !!(hw->flowctrl & ALX_FC_TX);
+	pause->rx_pause = !!(hw->flowctrl & ALX_FC_RX);
 }
 
 
@@ -187,7 +174,8 @@ static int alx_set_pauseparam(struct net_device *netdev,
 
 	if (reconfig_phy) {
 		err = alx_setup_speed_duplex(hw, hw->adv_cfg, fc);
-		return err;
+		if (err)
+			return err;
 	}
 
 	/* flow control on mac */
@@ -213,60 +201,12 @@ static void alx_set_msglevel(struct net_device *netdev, u32 data)
 	alx->msg_enable = data;
 }
 
-static void alx_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
-{
-	struct alx_priv *alx = netdev_priv(netdev);
-	struct alx_hw *hw = &alx->hw;
-
-	wol->supported = WAKE_MAGIC | WAKE_PHY;
-	wol->wolopts = 0;
-
-	if (hw->sleep_ctrl & ALX_SLEEP_WOL_MAGIC)
-		wol->wolopts |= WAKE_MAGIC;
-	if (hw->sleep_ctrl & ALX_SLEEP_WOL_PHY)
-		wol->wolopts |= WAKE_PHY;
-}
-
-static int alx_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
-{
-	struct alx_priv *alx = netdev_priv(netdev);
-	struct alx_hw *hw = &alx->hw;
-
-	if (wol->wolopts & (WAKE_ARP | WAKE_MAGICSECURE |
-			    WAKE_UCAST | WAKE_BCAST | WAKE_MCAST))
-		return -EOPNOTSUPP;
-
-	hw->sleep_ctrl = 0;
-
-	if (wol->wolopts & WAKE_MAGIC)
-		hw->sleep_ctrl |= ALX_SLEEP_WOL_MAGIC;
-	if (wol->wolopts & WAKE_PHY)
-		hw->sleep_ctrl |= ALX_SLEEP_WOL_PHY;
-
-	device_set_wakeup_enable(&alx->hw.pdev->dev, hw->sleep_ctrl);
-
-	return 0;
-}
-
-static void alx_get_drvinfo(struct net_device *netdev,
-			    struct ethtool_drvinfo *drvinfo)
-{
-	struct alx_priv *alx = netdev_priv(netdev);
-
-	strlcpy(drvinfo->driver, alx_drv_name, sizeof(drvinfo->driver));
-	strlcpy(drvinfo->bus_info, pci_name(alx->hw.pdev),
-		sizeof(drvinfo->bus_info));
-}
-
 const struct ethtool_ops alx_ethtool_ops = {
 	.get_settings	= alx_get_settings,
 	.set_settings	= alx_set_settings,
 	.get_pauseparam	= alx_get_pauseparam,
 	.set_pauseparam	= alx_set_pauseparam,
-	.get_drvinfo	= alx_get_drvinfo,
 	.get_msglevel	= alx_get_msglevel,
 	.set_msglevel	= alx_set_msglevel,
-	.get_wol	= alx_get_wol,
-	.set_wol	= alx_set_wol,
 	.get_link	= ethtool_op_get_link,
 };
diff --git a/drivers/net/ethernet/atheros/alx/hw.c b/drivers/net/ethernet/atheros/alx/hw.c
index 220a16a..1e8c24a 100644
--- a/drivers/net/ethernet/atheros/alx/hw.c
+++ b/drivers/net/ethernet/atheros/alx/hw.c
@@ -282,8 +282,8 @@ static bool alx_read_macaddr(struct alx_hw *hw, u8 *addr)
 	mac1 = alx_read_mem32(hw, ALX_STAD1);
 
 	/* addr should be big-endian */
-	*(__be32 *)(addr + 2) = cpu_to_be32(mac0);
-	*(__be16 *)addr = cpu_to_be16(mac1);
+	put_unaligned(cpu_to_be32(mac0), (__be32 *)(addr + 2));
+	put_unaligned(cpu_to_be16(mac1), (__be16 *)addr);
 
 	return is_valid_ether_addr(addr);
 }
@@ -326,22 +326,12 @@ void alx_set_macaddr(struct alx_hw *hw, const u8 *addr)
 	u32 val;
 
 	/* for example: 00-0B-6A-F6-00-DC * STAD0=6AF600DC, STAD1=000B */
-	val = be32_to_cpu(*(__be32 *)(addr + 2));
+	val = be32_to_cpu(get_unaligned((__be32 *)(addr + 2)));
 	alx_write_mem32(hw, ALX_STAD0, val);
-	val = be16_to_cpu(*(__be16 *)addr);
+	val = be16_to_cpu(get_unaligned((__be16 *)addr));
 	alx_write_mem32(hw, ALX_STAD1, val);
 }
 
-static void alx_enable_osc(struct alx_hw *hw)
-{
-	u32 val;
-
-	/* rising edge */
-	val = alx_read_mem32(hw, ALX_MISC);
-	alx_write_mem32(hw, ALX_MISC, val & ~ALX_MISC_INTNLOSC_OPEN);
-	alx_write_mem32(hw, ALX_MISC, val | ALX_MISC_INTNLOSC_OPEN);
-}
-
 static void alx_reset_osc(struct alx_hw *hw, u8 rev)
 {
 	u32 val, val2;
@@ -624,12 +614,12 @@ void alx_start_mac(struct alx_hw *hw)
 	alx_write_mem32(hw, ALX_TXQ0, txq | ALX_TXQ0_EN);
 
 	mac = hw->rx_ctrl;
-	if (hw->link_speed % 10 == DUPLEX_FULL)
+	if (hw->duplex == DUPLEX_FULL)
 		mac |= ALX_MAC_CTRL_FULLD;
 	else
 		mac &= ~ALX_MAC_CTRL_FULLD;
 	ALX_SET_FIELD(mac, ALX_MAC_CTRL_SPEED,
-		      hw->link_speed >= SPEED_1000 ? ALX_MAC_CTRL_SPEED_1000 :
+		      hw->link_speed == SPEED_1000 ? ALX_MAC_CTRL_SPEED_1000 :
 						     ALX_MAC_CTRL_SPEED_10_100);
 	mac |= ALX_MAC_CTRL_TX_EN | ALX_MAC_CTRL_RX_EN;
 	hw->rx_ctrl = mac;
@@ -790,28 +780,22 @@ void alx_post_phy_link(struct alx_hw *hw)
 	u16 phy_val, len, agc;
 	u8 revid = alx_hw_revision(hw);
 	bool adj_th = revid == ALX_REV_B0;
-	int speed;
-
-	if (hw->link_speed == SPEED_UNKNOWN)
-		speed = SPEED_UNKNOWN;
-	else
-		speed = hw->link_speed - hw->link_speed % 10;
 
 	if (revid != ALX_REV_B0 && !alx_is_rev_a(revid))
 		return;
 
 	/* 1000BT/AZ, wrong cable length */
-	if (speed != SPEED_UNKNOWN) {
+	if (hw->link_speed != SPEED_UNKNOWN) {
 		alx_read_phy_ext(hw, ALX_MIIEXT_PCS, ALX_MIIEXT_CLDCTRL6,
 				 &phy_val);
 		len = ALX_GET_FIELD(phy_val, ALX_CLDCTRL6_CAB_LEN);
 		alx_read_phy_dbg(hw, ALX_MIIDBG_AGC, &phy_val);
 		agc = ALX_GET_FIELD(phy_val, ALX_AGC_2_VGA);
 
-		if ((speed == SPEED_1000 &&
+		if ((hw->link_speed == SPEED_1000 &&
 		     (len > ALX_CLDCTRL6_CAB_LEN_SHORT1G ||
 		      (len == 0 && agc > ALX_AGC_LONG1G_LIMT))) ||
-		    (speed == SPEED_100 &&
+		    (hw->link_speed == SPEED_100 &&
 		     (len > ALX_CLDCTRL6_CAB_LEN_SHORT100M ||
 		      (len == 0 && agc > ALX_AGC_LONG100M_LIMT)))) {
 			alx_write_phy_dbg(hw, ALX_MIIDBG_AZ_ANADECT,
@@ -831,10 +815,10 @@ void alx_post_phy_link(struct alx_hw *hw)
 
 		/* threshold adjust */
 		if (adj_th && hw->lnk_patch) {
-			if (speed == SPEED_100) {
+			if (hw->link_speed == SPEED_100) {
 				alx_write_phy_dbg(hw, ALX_MIIDBG_MSE16DB,
 						  ALX_MSE16DB_UP);
-			} else if (speed == SPEED_1000) {
+			} else if (hw->link_speed == SPEED_1000) {
 				/*
 				 * Giga link threshold, raise the tolerance of
 				 * noise 50%
@@ -864,66 +848,6 @@ void alx_post_phy_link(struct alx_hw *hw)
 	}
 }
 
-
-/* NOTE:
- *    1. phy link must be established before calling this function
- *    2. wol option (pattern,magic,link,etc.) is configed before call it.
- */
-int alx_pre_suspend(struct alx_hw *hw, int speed)
-{
-	u32 master, mac, phy, val;
-	int err = 0;
-
-	master = alx_read_mem32(hw, ALX_MASTER);
-	master &= ~ALX_MASTER_PCLKSEL_SRDS;
-	mac = hw->rx_ctrl;
-	/* 10/100 half */
-	ALX_SET_FIELD(mac, ALX_MAC_CTRL_SPEED,  ALX_MAC_CTRL_SPEED_10_100);
-	mac &= ~(ALX_MAC_CTRL_FULLD | ALX_MAC_CTRL_RX_EN | ALX_MAC_CTRL_TX_EN);
-
-	phy = alx_read_mem32(hw, ALX_PHY_CTRL);
-	phy &= ~(ALX_PHY_CTRL_DSPRST_OUT | ALX_PHY_CTRL_CLS);
-	phy |= ALX_PHY_CTRL_RST_ANALOG | ALX_PHY_CTRL_HIB_PULSE |
-	       ALX_PHY_CTRL_HIB_EN;
-
-	/* without any activity  */
-	if (!(hw->sleep_ctrl & ALX_SLEEP_ACTIVE)) {
-		err = alx_write_phy_reg(hw, ALX_MII_IER, 0);
-		if (err)
-			return err;
-		phy |= ALX_PHY_CTRL_IDDQ | ALX_PHY_CTRL_POWER_DOWN;
-	} else {
-		if (hw->sleep_ctrl & (ALX_SLEEP_WOL_MAGIC | ALX_SLEEP_CIFS))
-			mac |= ALX_MAC_CTRL_RX_EN | ALX_MAC_CTRL_BRD_EN;
-		if (hw->sleep_ctrl & ALX_SLEEP_CIFS)
-			mac |= ALX_MAC_CTRL_TX_EN;
-		if (speed % 10 == DUPLEX_FULL)
-			mac |= ALX_MAC_CTRL_FULLD;
-		if (speed >= SPEED_1000)
-			ALX_SET_FIELD(mac, ALX_MAC_CTRL_SPEED,
-				      ALX_MAC_CTRL_SPEED_1000);
-		phy |= ALX_PHY_CTRL_DSPRST_OUT;
-		err = alx_write_phy_ext(hw, ALX_MIIEXT_ANEG,
-					ALX_MIIEXT_S3DIG10,
-					ALX_MIIEXT_S3DIG10_SL);
-		if (err)
-			return err;
-	}
-
-	alx_enable_osc(hw);
-	hw->rx_ctrl = mac;
-	alx_write_mem32(hw, ALX_MASTER, master);
-	alx_write_mem32(hw, ALX_MAC_CTRL, mac);
-	alx_write_mem32(hw, ALX_PHY_CTRL, phy);
-
-	/* set val of PDLL D3PLLOFF */
-	val = alx_read_mem32(hw, ALX_PDLL_TRNS1);
-	val |= ALX_PDLL_TRNS1_D3PLLOFF_EN;
-	alx_write_mem32(hw, ALX_PDLL_TRNS1, val);
-
-	return 0;
-}
-
 bool alx_phy_configured(struct alx_hw *hw)
 {
 	u32 cfg, hw_cfg;
@@ -938,7 +862,7 @@ bool alx_phy_configured(struct alx_hw *hw)
 	return cfg == hw_cfg;
 }
 
-int alx_get_phy_link(struct alx_hw *hw, int *speed)
+int alx_read_phy_link(struct alx_hw *hw)
 {
 	struct pci_dev *pdev = hw->pdev;
 	u16 bmsr, giga;
@@ -953,7 +877,8 @@ int alx_get_phy_link(struct alx_hw *hw, int *speed)
 		return err;
 
 	if (!(bmsr & BMSR_LSTATUS)) {
-		*speed = SPEED_UNKNOWN;
+		hw->link_speed = SPEED_UNKNOWN;
+		hw->duplex = DUPLEX_UNKNOWN;
 		return 0;
 	}
 
@@ -967,20 +892,20 @@ int alx_get_phy_link(struct alx_hw *hw, int *speed)
 
 	switch (giga & ALX_GIGA_PSSR_SPEED) {
 	case ALX_GIGA_PSSR_1000MBS:
-		*speed = SPEED_1000;
+		hw->link_speed = SPEED_1000;
 		break;
 	case ALX_GIGA_PSSR_100MBS:
-		*speed = SPEED_100;
+		hw->link_speed = SPEED_100;
 		break;
 	case ALX_GIGA_PSSR_10MBS:
-		*speed = SPEED_10;
+		hw->link_speed = SPEED_10;
 		break;
 	default:
 		goto wrong_speed;
 	}
 
-	*speed += (giga & ALX_GIGA_PSSR_DPLX) ? DUPLEX_FULL : DUPLEX_HALF;
-	return 1;
+	hw->duplex = (giga & ALX_GIGA_PSSR_DPLX) ? DUPLEX_FULL : DUPLEX_HALF;
+	return 0;
 
 wrong_speed:
 	dev_err(&pdev->dev, "invalid PHY speed/duplex: 0x%x\n", giga);
@@ -995,26 +920,6 @@ int alx_clear_phy_intr(struct alx_hw *hw)
 	return alx_read_phy_reg(hw, ALX_MII_ISR, &isr);
 }
 
-int alx_config_wol(struct alx_hw *hw)
-{
-	u32 wol = 0;
-	int err = 0;
-
-	/* turn on magic packet event */
-	if (hw->sleep_ctrl & ALX_SLEEP_WOL_MAGIC)
-		wol |= ALX_WOL0_MAGIC_EN | ALX_WOL0_PME_MAGIC_EN;
-
-	/* turn on link up event */
-	if (hw->sleep_ctrl & ALX_SLEEP_WOL_PHY) {
-		wol |=  ALX_WOL0_LINK_EN | ALX_WOL0_PME_LINK;
-		/* only link up can wake up */
-		err = alx_write_phy_reg(hw, ALX_MII_IER, ALX_IER_LINK_UP);
-	}
-	alx_write_mem32(hw, ALX_WOL0, wol);
-
-	return err;
-}
-
 void alx_disable_rss(struct alx_hw *hw)
 {
 	u32 ctrl = alx_read_mem32(hw, ALX_RXQ0);
@@ -1126,85 +1031,6 @@ void alx_configure_basic(struct alx_hw *hw)
 	alx_write_mem32(hw, ALX_WRR, val);
 }
 
-static inline u32 alx_speed_to_ethadv(int speed)
-{
-	switch (speed) {
-	case SPEED_1000 + DUPLEX_FULL:
-		return ADVERTISED_1000baseT_Full;
-	case SPEED_100 + DUPLEX_FULL:
-		return ADVERTISED_100baseT_Full;
-	case SPEED_100 + DUPLEX_HALF:
-		return ADVERTISED_10baseT_Half;
-	case SPEED_10 + DUPLEX_FULL:
-		return ADVERTISED_10baseT_Full;
-	case SPEED_10 + DUPLEX_HALF:
-		return ADVERTISED_10baseT_Half;
-	default:
-		return 0;
-	}
-}
-
-int alx_select_powersaving_speed(struct alx_hw *hw, int *speed)
-{
-	int i, err, spd;
-	u16 lpa;
-
-	err = alx_get_phy_link(hw, &spd);
-	if (err < 0)
-		return err;
-
-	if (spd == SPEED_UNKNOWN)
-		return 0;
-
-	err = alx_read_phy_reg(hw, MII_LPA, &lpa);
-	if (err)
-		return err;
-
-	if (!(lpa & LPA_LPACK)) {
-		*speed = spd;
-		return 0;
-	}
-
-	if (lpa & LPA_10FULL)
-		*speed = SPEED_10 + DUPLEX_FULL;
-	else if (lpa & LPA_10HALF)
-		*speed = SPEED_10 + DUPLEX_HALF;
-	else if (lpa & LPA_100FULL)
-		*speed = SPEED_100 + DUPLEX_FULL;
-	else
-		*speed = SPEED_100 + DUPLEX_HALF;
-
-	if (*speed != spd) {
-		err = alx_write_phy_reg(hw, ALX_MII_IER, 0);
-		if (err)
-			return err;
-		err = alx_setup_speed_duplex(hw,
-					     alx_speed_to_ethadv(*speed) |
-					     ADVERTISED_Autoneg,
-					     ALX_FC_ANEG | ALX_FC_RX |
-					     ALX_FC_TX);
-		if (err)
-			return err;
-
-		/* wait for linkup */
-		for (i = 0; i < ALX_MAX_SETUP_LNK_CYCLE; i++) {
-			int speed2;
-
-			msleep(100);
-
-			err = alx_get_phy_link(hw, &speed2);
-			if (err < 0)
-				return err;
-			if (speed2 != SPEED_UNKNOWN)
-				break;
-		}
-		if (i == ALX_MAX_SETUP_LNK_CYCLE)
-			return -ETIMEDOUT;
-	}
-
-	return 0;
-}
-
 bool alx_get_phy_info(struct alx_hw *hw)
 {
 	u16  devs1, devs2;
diff --git a/drivers/net/ethernet/atheros/alx/hw.h b/drivers/net/ethernet/atheros/alx/hw.h
index 65e723d..96f3b43 100644
--- a/drivers/net/ethernet/atheros/alx/hw.h
+++ b/drivers/net/ethernet/atheros/alx/hw.h
@@ -412,12 +412,11 @@ struct alx_hw {
 	u32 smb_timer;
 	/* SPEED_* + DUPLEX_*, SPEED_UNKNOWN if link is down */
 	int link_speed;
+	u8 duplex;
 
 	/* auto-neg advertisement or force mode config */
-	u32 adv_cfg;
 	u8 flowctrl;
-
-	u32 sleep_ctrl;
+	u32 adv_cfg;
 
 	spinlock_t mdio_lock;
 	struct mdio_if_info mdio;
@@ -478,14 +477,12 @@ void alx_reset_pcie(struct alx_hw *hw);
 void alx_enable_aspm(struct alx_hw *hw, bool l0s_en, bool l1_en);
 int alx_setup_speed_duplex(struct alx_hw *hw, u32 ethadv, u8 flowctrl);
 void alx_post_phy_link(struct alx_hw *hw);
-int alx_pre_suspend(struct alx_hw *hw, int speed);
 int alx_read_phy_reg(struct alx_hw *hw, u16 reg, u16 *phy_data);
 int alx_write_phy_reg(struct alx_hw *hw, u16 reg, u16 phy_data);
 int alx_read_phy_ext(struct alx_hw *hw, u8 dev, u16 reg, u16 *pdata);
 int alx_write_phy_ext(struct alx_hw *hw, u8 dev, u16 reg, u16 data);
-int alx_get_phy_link(struct alx_hw *hw, int *speed);
+int alx_read_phy_link(struct alx_hw *hw);
 int alx_clear_phy_intr(struct alx_hw *hw);
-int alx_config_wol(struct alx_hw *hw);
 void alx_cfg_mac_flowcontrol(struct alx_hw *hw, u8 fc);
 void alx_start_mac(struct alx_hw *hw);
 int alx_reset_mac(struct alx_hw *hw);
@@ -493,7 +490,21 @@ void alx_set_macaddr(struct alx_hw *hw, const u8 *addr);
 bool alx_phy_configured(struct alx_hw *hw);
 void alx_configure_basic(struct alx_hw *hw);
 void alx_disable_rss(struct alx_hw *hw);
-int alx_select_powersaving_speed(struct alx_hw *hw, int *speed);
 bool alx_get_phy_info(struct alx_hw *hw);
 
+static inline u32 alx_speed_to_ethadv(int speed, u8 duplex)
+{
+	if (speed == SPEED_1000 && duplex == DUPLEX_FULL)
+		return ADVERTISED_1000baseT_Full;
+	if (speed == SPEED_100 && duplex == DUPLEX_FULL)
+		return ADVERTISED_100baseT_Full;
+	if (speed == SPEED_100 && duplex== DUPLEX_HALF)
+		return ADVERTISED_100baseT_Half;
+	if (speed == SPEED_10 && duplex == DUPLEX_FULL)
+		return ADVERTISED_10baseT_Full;
+	if (speed == SPEED_10 && duplex == DUPLEX_HALF)
+		return ADVERTISED_10baseT_Half;
+	return 0;
+}
+
 #endif
diff --git a/drivers/net/ethernet/atheros/alx/main.c b/drivers/net/ethernet/atheros/alx/main.c
index 418de8b..0e0b242 100644
--- a/drivers/net/ethernet/atheros/alx/main.c
+++ b/drivers/net/ethernet/atheros/alx/main.c
@@ -706,12 +706,12 @@ static int alx_init_sw(struct alx_priv *alx)
 	alx->rxbuf_size = ALIGN(ALX_RAW_MTU(hw->mtu), 8);
 	alx->tx_ringsz = 256;
 	alx->rx_ringsz = 512;
-	hw->sleep_ctrl = ALX_SLEEP_WOL_MAGIC | ALX_SLEEP_WOL_PHY;
 	hw->imt = 200;
 	alx->int_mask = ALX_ISR_MISC;
 	hw->dma_chnl = hw->max_dma_chnl;
 	hw->ith_tpd = alx->tx_ringsz / 3;
 	hw->link_speed = SPEED_UNKNOWN;
+	hw->duplex = DUPLEX_UNKNOWN;
 	hw->adv_cfg = ADVERTISED_Autoneg |
 		      ADVERTISED_10baseT_Half |
 		      ADVERTISED_10baseT_Full |
@@ -758,6 +758,7 @@ static void alx_halt(struct alx_priv *alx)
 
 	alx_netif_stop(alx);
 	hw->link_speed = SPEED_UNKNOWN;
+	hw->duplex = DUPLEX_UNKNOWN;
 
 	alx_reset_mac(hw);
 
@@ -869,18 +870,18 @@ static void __alx_stop(struct alx_priv *alx)
 	alx_free_rings(alx);
 }
 
-static const char *alx_speed_desc(u16 speed)
+static const char *alx_speed_desc(struct alx_hw *hw)
 {
-	switch (speed) {
-	case SPEED_1000 + DUPLEX_FULL:
+	switch (alx_speed_to_ethadv(hw->link_speed, hw->duplex)) {
+	case ADVERTISED_1000baseT_Full:
 		return "1 Gbps Full";
-	case SPEED_100 + DUPLEX_FULL:
+	case ADVERTISED_100baseT_Full:
 		return "100 Mbps Full";
-	case SPEED_100 + DUPLEX_HALF:
+	case ADVERTISED_100baseT_Half:
 		return "100 Mbps Half";
-	case SPEED_10 + DUPLEX_FULL:
+	case ADVERTISED_10baseT_Full:
 		return "10 Mbps Full";
-	case SPEED_10 + DUPLEX_HALF:
+	case ADVERTISED_10baseT_Half:
 		return "10 Mbps Half";
 	default:
 		return "Unknown speed";
@@ -891,7 +892,8 @@ static void alx_check_link(struct alx_priv *alx)
 {
 	struct alx_hw *hw = &alx->hw;
 	unsigned long flags;
-	int speed, old_speed;
+	int old_speed;
+	u8 old_duplex;
 	int err;
 
 	/* clear PHY internal interrupt status, otherwise the main
@@ -899,7 +901,9 @@ static void alx_check_link(struct alx_priv *alx)
 	 */
 	alx_clear_phy_intr(hw);
 
-	err = alx_get_phy_link(hw, &speed);
+	old_speed = hw->link_speed;
+	old_duplex = hw->duplex;
+	err = alx_read_phy_link(hw);
 	if (err < 0)
 		goto reset;
 
@@ -908,15 +912,12 @@ static void alx_check_link(struct alx_priv *alx)
 	alx_write_mem32(hw, ALX_IMR, alx->int_mask);
 	spin_unlock_irqrestore(&alx->irq_lock, flags);
 
-	old_speed = hw->link_speed;
-
-	if (old_speed == speed)
+	if (old_speed == hw->link_speed)
 		return;
-	hw->link_speed = speed;
 
-	if (speed != SPEED_UNKNOWN) {
+	if (hw->link_speed != SPEED_UNKNOWN) {
 		netif_info(alx, link, alx->dev,
-			   "NIC Up: %s\n", alx_speed_desc(speed));
+			   "NIC Up: %s\n", alx_speed_desc(hw));
 		alx_post_phy_link(hw);
 		alx_enable_aspm(hw, true, true);
 		alx_start_mac(hw);
@@ -959,65 +960,6 @@ static int alx_stop(struct net_device *netdev)
 	return 0;
 }
 
-static int __alx_shutdown(struct pci_dev *pdev, bool *wol_en)
-{
-	struct alx_priv *alx = pci_get_drvdata(pdev);
-	struct net_device *netdev = alx->dev;
-	struct alx_hw *hw = &alx->hw;
-	int err, speed;
-
-	netif_device_detach(netdev);
-
-	if (netif_running(netdev))
-		__alx_stop(alx);
-
-#ifdef CONFIG_PM_SLEEP
-	err = pci_save_state(pdev);
-	if (err)
-		return err;
-#endif
-
-	err = alx_select_powersaving_speed(hw, &speed);
-	if (err)
-		return err;
-	err = alx_clear_phy_intr(hw);
-	if (err)
-		return err;
-	err = alx_pre_suspend(hw, speed);
-	if (err)
-		return err;
-	err = alx_config_wol(hw);
-	if (err)
-		return err;
-
-	*wol_en = false;
-	if (hw->sleep_ctrl & ALX_SLEEP_ACTIVE) {
-		netif_info(alx, wol, netdev,
-			   "wol: ctrl=%X, speed=%X\n",
-			   hw->sleep_ctrl, speed);
-		device_set_wakeup_enable(&pdev->dev, true);
-		*wol_en = true;
-	}
-
-	pci_disable_device(pdev);
-
-	return 0;
-}
-
-static void alx_shutdown(struct pci_dev *pdev)
-{
-	int err;
-	bool wol_en;
-
-	err = __alx_shutdown(pdev, &wol_en);
-	if (!err) {
-		pci_wake_from_d3(pdev, wol_en);
-		pci_set_power_state(pdev, PCI_D3hot);
-	} else {
-		dev_err(&pdev->dev, "shutdown fail %d\n", err);
-	}
-}
-
 static void alx_link_check(struct work_struct *work)
 {
 	struct alx_priv *alx;
@@ -1396,8 +1338,6 @@ static int alx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto out_unmap;
 	}
 
-	device_set_wakeup_enable(&pdev->dev, hw->sleep_ctrl);
-
 	netdev_info(netdev,
 		    "Qualcomm Atheros AR816x/AR817x Ethernet [%pM]\n",
 		    netdev->dev_addr);
@@ -1442,22 +1382,12 @@ static void alx_remove(struct pci_dev *pdev)
 static int alx_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
-	int err;
-	bool wol_en;
-
-	err = __alx_shutdown(pdev, &wol_en);
-	if (err) {
-		dev_err(&pdev->dev, "shutdown fail in suspend %d\n", err);
-		return err;
-	}
-
-	if (wol_en) {
-		pci_prepare_to_sleep(pdev);
-	} else {
-		pci_wake_from_d3(pdev, false);
-		pci_set_power_state(pdev, PCI_D3hot);
-	}
+	struct alx_priv *alx = pci_get_drvdata(pdev);
 
+	if (!netif_running(alx->dev))
+		return 0;
+	netif_device_detach(alx->dev);
+	__alx_stop(alx);
 	return 0;
 }
 
@@ -1465,49 +1395,20 @@ static int alx_resume(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct alx_priv *alx = pci_get_drvdata(pdev);
-	struct net_device *netdev = alx->dev;
-	struct alx_hw *hw = &alx->hw;
-	int err;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	pci_save_state(pdev);
-
-	pci_enable_wake(pdev, PCI_D3hot, 0);
-	pci_enable_wake(pdev, PCI_D3cold, 0);
 
-	hw->link_speed = SPEED_UNKNOWN;
-	alx->int_mask = ALX_ISR_MISC;
-
-	alx_reset_pcie(hw);
-	alx_reset_phy(hw);
-
-	err = alx_reset_mac(hw);
-	if (err) {
-		netif_err(alx, hw, alx->dev,
-			  "resume:reset_mac fail %d\n", err);
-		return -EIO;
-	}
-
-	err = alx_setup_speed_duplex(hw, hw->adv_cfg, hw->flowctrl);
-	if (err) {
-		netif_err(alx, hw, alx->dev,
-			  "resume:setup_speed_duplex fail %d\n", err);
-		return -EIO;
-	}
-
-	if (netif_running(netdev)) {
-		err = __alx_open(alx, true);
-		if (err)
-			return err;
-	}
-
-	netif_device_attach(netdev);
-
-	return err;
+	if (!netif_running(alx->dev))
+		return 0;
+	netif_device_attach(alx->dev);
+	return __alx_open(alx, true);
 }
+
+static SIMPLE_DEV_PM_OPS(alx_pm_ops, alx_suspend, alx_resume);
+#define ALX_PM_OPS      (&alx_pm_ops)
+#else
+#define ALX_PM_OPS      NULL
 #endif
 
+
 static pci_ers_result_t alx_pci_error_detected(struct pci_dev *pdev,
 					       pci_channel_state_t state)
 {
@@ -1550,8 +1451,6 @@ static pci_ers_result_t alx_pci_error_slot_reset(struct pci_dev *pdev)
 	}
 
 	pci_set_master(pdev);
-	pci_enable_wake(pdev, PCI_D3hot, 0);
-	pci_enable_wake(pdev, PCI_D3cold, 0);
 
 	alx_reset_pcie(hw);
 	if (!alx_reset_mac(hw))
@@ -1587,13 +1486,6 @@ static const struct pci_error_handlers alx_err_handlers = {
 	.resume         = alx_pci_error_resume,
 };
 
-#ifdef CONFIG_PM_SLEEP
-static SIMPLE_DEV_PM_OPS(alx_pm_ops, alx_suspend, alx_resume);
-#define ALX_PM_OPS      (&alx_pm_ops)
-#else
-#define ALX_PM_OPS      NULL
-#endif
-
 static DEFINE_PCI_DEVICE_TABLE(alx_pci_tbl) = {
 	{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_AR8161),
 	  .driver_data = ALX_DEV_QUIRK_MSI_INTX_DISABLE_BUG },
@@ -1611,7 +1503,6 @@ static struct pci_driver alx_driver = {
 	.id_table    = alx_pci_tbl,
 	.probe       = alx_probe,
 	.remove      = alx_remove,
-	.shutdown    = alx_shutdown,
 	.err_handler = &alx_err_handlers,
 	.driver.pm   = ALX_PM_OPS,
 };
diff --git a/drivers/net/ethernet/atheros/atl1c/atl1c_main.c b/drivers/net/ethernet/atheros/atl1c/atl1c_main.c
index 0ba9007..786a874 100644
--- a/drivers/net/ethernet/atheros/atl1c/atl1c_main.c
+++ b/drivers/net/ethernet/atheros/atl1c/atl1c_main.c
@@ -2755,27 +2755,4 @@ static struct pci_driver atl1c_driver = {
 	.driver.pm = &atl1c_pm_ops,
 };
 
-/**
- * atl1c_init_module - Driver Registration Routine
- *
- * atl1c_init_module is the first routine called when the driver is
- * loaded. All it does is register with the PCI subsystem.
- */
-static int __init atl1c_init_module(void)
-{
-	return pci_register_driver(&atl1c_driver);
-}
-
-/**
- * atl1c_exit_module - Driver Exit Cleanup Routine
- *
- * atl1c_exit_module is called just before the driver is removed
- * from memory.
- */
-static void __exit atl1c_exit_module(void)
-{
-	pci_unregister_driver(&atl1c_driver);
-}
-
-module_init(atl1c_init_module);
-module_exit(atl1c_exit_module);
+module_pci_driver(atl1c_driver);
diff --git a/drivers/net/ethernet/atheros/atl1e/atl1e_main.c b/drivers/net/ethernet/atheros/atl1e/atl1e_main.c
index 0688bb8..895f537 100644
--- a/drivers/net/ethernet/atheros/atl1e/atl1e_main.c
+++ b/drivers/net/ethernet/atheros/atl1e/atl1e_main.c
@@ -2489,27 +2489,4 @@ static struct pci_driver atl1e_driver = {
 	.err_handler = &atl1e_err_handler
 };
 
-/**
- * atl1e_init_module - Driver Registration Routine
- *
- * atl1e_init_module is the first routine called when the driver is
- * loaded. All it does is register with the PCI subsystem.
- */
-static int __init atl1e_init_module(void)
-{
-	return pci_register_driver(&atl1e_driver);
-}
-
-/**
- * atl1e_exit_module - Driver Exit Cleanup Routine
- *
- * atl1e_exit_module is called just before the driver is removed
- * from memory.
- */
-static void __exit atl1e_exit_module(void)
-{
-	pci_unregister_driver(&atl1e_driver);
-}
-
-module_init(atl1e_init_module);
-module_exit(atl1e_exit_module);
+module_pci_driver(atl1e_driver);
diff --git a/drivers/net/ethernet/atheros/atlx/atl1.c b/drivers/net/ethernet/atheros/atlx/atl1.c
index fa0915f..538211d 100644
--- a/drivers/net/ethernet/atheros/atlx/atl1.c
+++ b/drivers/net/ethernet/atheros/atlx/atl1.c
@@ -3145,31 +3145,6 @@ static struct pci_driver atl1_driver = {
 	.driver.pm = &atl1_pm_ops,
 };
 
-/**
- * atl1_exit_module - Driver Exit Cleanup Routine
- *
- * atl1_exit_module is called just before the driver is removed
- * from memory.
- */
-static void __exit atl1_exit_module(void)
-{
-	pci_unregister_driver(&atl1_driver);
-}
-
-/**
- * atl1_init_module - Driver Registration Routine
- *
- * atl1_init_module is the first routine called when the driver is
- * loaded. All it does is register with the PCI subsystem.
- */
-static int __init atl1_init_module(void)
-{
-	return pci_register_driver(&atl1_driver);
-}
-
-module_init(atl1_init_module);
-module_exit(atl1_exit_module);
-
 struct atl1_stats {
 	char stat_string[ETH_GSTRING_LEN];
 	int sizeof_stat;
@@ -3705,3 +3680,5 @@ static const struct ethtool_ops atl1_ethtool_ops = {
 	.get_ethtool_stats	= atl1_get_ethtool_stats,
 	.get_sset_count		= atl1_get_sset_count,
 };
+
+module_pci_driver(atl1_driver);
diff --git a/drivers/net/ethernet/broadcom/Kconfig b/drivers/net/ethernet/broadcom/Kconfig
index 3e69b3f..1d680ba 100644
--- a/drivers/net/ethernet/broadcom/Kconfig
+++ b/drivers/net/ethernet/broadcom/Kconfig
@@ -22,7 +22,6 @@ config B44
 	tristate "Broadcom 440x/47xx ethernet support"
 	depends on SSB_POSSIBLE && HAS_DMA
 	select SSB
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a network (Ethernet) controller of this type, say Y
@@ -54,7 +53,6 @@ config B44_PCI
 config BCM63XX_ENET
 	tristate "Broadcom 63xx internal mac support"
 	depends on BCM63XX
-	select NET_CORE
 	select MII
 	select PHYLIB
 	help
diff --git a/drivers/net/ethernet/broadcom/bcm63xx_enet.c b/drivers/net/ethernet/broadcom/bcm63xx_enet.c
index 0b3e23e..b1bcd4b 100644
--- a/drivers/net/ethernet/broadcom/bcm63xx_enet.c
+++ b/drivers/net/ethernet/broadcom/bcm63xx_enet.c
@@ -41,8 +41,8 @@ static int copybreak __read_mostly = 128;
 module_param(copybreak, int, 0);
 MODULE_PARM_DESC(copybreak, "Receive copy threshold");
 
-/* io memory shared between all devices */
-static void __iomem *bcm_enet_shared_base;
+/* io registers memory shared between all devices */
+static void __iomem *bcm_enet_shared_base[3];
 
 /*
  * io helpers to access mac registers
@@ -59,17 +59,76 @@ static inline void enet_writel(struct bcm_enet_priv *priv,
 }
 
 /*
- * io helpers to access shared registers
+ * io helpers to access switch registers
  */
+static inline u32 enetsw_readl(struct bcm_enet_priv *priv, u32 off)
+{
+	return bcm_readl(priv->base + off);
+}
+
+static inline void enetsw_writel(struct bcm_enet_priv *priv,
+				 u32 val, u32 off)
+{
+	bcm_writel(val, priv->base + off);
+}
+
+static inline u16 enetsw_readw(struct bcm_enet_priv *priv, u32 off)
+{
+	return bcm_readw(priv->base + off);
+}
+
+static inline void enetsw_writew(struct bcm_enet_priv *priv,
+				 u16 val, u32 off)
+{
+	bcm_writew(val, priv->base + off);
+}
+
+static inline u8 enetsw_readb(struct bcm_enet_priv *priv, u32 off)
+{
+	return bcm_readb(priv->base + off);
+}
+
+static inline void enetsw_writeb(struct bcm_enet_priv *priv,
+				 u8 val, u32 off)
+{
+	bcm_writeb(val, priv->base + off);
+}
+
+
+/* io helpers to access shared registers */
 static inline u32 enet_dma_readl(struct bcm_enet_priv *priv, u32 off)
 {
-	return bcm_readl(bcm_enet_shared_base + off);
+	return bcm_readl(bcm_enet_shared_base[0] + off);
 }
 
 static inline void enet_dma_writel(struct bcm_enet_priv *priv,
 				       u32 val, u32 off)
 {
-	bcm_writel(val, bcm_enet_shared_base + off);
+	bcm_writel(val, bcm_enet_shared_base[0] + off);
+}
+
+static inline u32 enet_dmac_readl(struct bcm_enet_priv *priv, u32 off, int chan)
+{
+	return bcm_readl(bcm_enet_shared_base[1] +
+		bcm63xx_enetdmacreg(off) + chan * priv->dma_chan_width);
+}
+
+static inline void enet_dmac_writel(struct bcm_enet_priv *priv,
+				       u32 val, u32 off, int chan)
+{
+	bcm_writel(val, bcm_enet_shared_base[1] +
+		bcm63xx_enetdmacreg(off) + chan * priv->dma_chan_width);
+}
+
+static inline u32 enet_dmas_readl(struct bcm_enet_priv *priv, u32 off, int chan)
+{
+	return bcm_readl(bcm_enet_shared_base[2] + off + chan * priv->dma_chan_width);
+}
+
+static inline void enet_dmas_writel(struct bcm_enet_priv *priv,
+				       u32 val, u32 off, int chan)
+{
+	bcm_writel(val, bcm_enet_shared_base[2] + off + chan * priv->dma_chan_width);
 }
 
 /*
@@ -196,7 +255,6 @@ static int bcm_enet_refill_rx(struct net_device *dev)
 			if (!skb)
 				break;
 			priv->rx_skb[desc_idx] = skb;
-
 			p = dma_map_single(&priv->pdev->dev, skb->data,
 					   priv->rx_skb_size,
 					   DMA_FROM_DEVICE);
@@ -206,7 +264,7 @@ static int bcm_enet_refill_rx(struct net_device *dev)
 		len_stat = priv->rx_skb_size << DMADESC_LENGTH_SHIFT;
 		len_stat |= DMADESC_OWNER_MASK;
 		if (priv->rx_dirty_desc == priv->rx_ring_size - 1) {
-			len_stat |= DMADESC_WRAP_MASK;
+			len_stat |= (DMADESC_WRAP_MASK >> priv->dma_desc_shift);
 			priv->rx_dirty_desc = 0;
 		} else {
 			priv->rx_dirty_desc++;
@@ -217,7 +275,10 @@ static int bcm_enet_refill_rx(struct net_device *dev)
 		priv->rx_desc_count++;
 
 		/* tell dma engine we allocated one buffer */
-		enet_dma_writel(priv, 1, ENETDMA_BUFALLOC_REG(priv->rx_chan));
+		if (priv->dma_has_sram)
+			enet_dma_writel(priv, 1, ENETDMA_BUFALLOC_REG(priv->rx_chan));
+		else
+			enet_dmac_writel(priv, 1, ENETDMAC_BUFALLOC, priv->rx_chan);
 	}
 
 	/* If rx ring is still empty, set a timer to try allocating
@@ -293,13 +354,15 @@ static int bcm_enet_receive_queue(struct net_device *dev, int budget)
 
 		/* if the packet does not have start of packet _and_
 		 * end of packet flag set, then just recycle it */
-		if ((len_stat & DMADESC_ESOP_MASK) != DMADESC_ESOP_MASK) {
+		if ((len_stat & (DMADESC_ESOP_MASK >> priv->dma_desc_shift)) !=
+			(DMADESC_ESOP_MASK >> priv->dma_desc_shift)) {
 			dev->stats.rx_dropped++;
 			continue;
 		}
 
 		/* recycle packet if it's marked as bad */
-		if (unlikely(len_stat & DMADESC_ERR_MASK)) {
+		if (!priv->enet_is_sw &&
+		    unlikely(len_stat & DMADESC_ERR_MASK)) {
 			dev->stats.rx_errors++;
 
 			if (len_stat & DMADESC_OVSIZE_MASK)
@@ -353,8 +416,8 @@ static int bcm_enet_receive_queue(struct net_device *dev, int budget)
 		bcm_enet_refill_rx(dev);
 
 		/* kick rx dma */
-		enet_dma_writel(priv, ENETDMA_CHANCFG_EN_MASK,
-				ENETDMA_CHANCFG_REG(priv->rx_chan));
+		enet_dmac_writel(priv, priv->dma_chan_en_mask,
+					 ENETDMAC_CHANCFG, priv->rx_chan);
 	}
 
 	return processed;
@@ -429,10 +492,10 @@ static int bcm_enet_poll(struct napi_struct *napi, int budget)
 	dev = priv->net_dev;
 
 	/* ack interrupts */
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IR_REG(priv->rx_chan));
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IR_REG(priv->tx_chan));
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IR, priv->rx_chan);
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IR, priv->tx_chan);
 
 	/* reclaim sent skb */
 	tx_work_done = bcm_enet_tx_reclaim(dev, 0);
@@ -451,10 +514,10 @@ static int bcm_enet_poll(struct napi_struct *napi, int budget)
 	napi_complete(napi);
 
 	/* restore rx/tx interrupt */
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IRMASK_REG(priv->rx_chan));
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IRMASK_REG(priv->tx_chan));
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IRMASK, priv->tx_chan);
 
 	return rx_work_done;
 }
@@ -497,8 +560,8 @@ static irqreturn_t bcm_enet_isr_dma(int irq, void *dev_id)
 	priv = netdev_priv(dev);
 
 	/* mask rx/tx interrupts */
-	enet_dma_writel(priv, 0, ENETDMA_IRMASK_REG(priv->rx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_IRMASK_REG(priv->tx_chan));
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->tx_chan);
 
 	napi_schedule(&priv->napi);
 
@@ -530,6 +593,26 @@ static int bcm_enet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		goto out_unlock;
 	}
 
+	/* pad small packets sent on a switch device */
+	if (priv->enet_is_sw && skb->len < 64) {
+		int needed = 64 - skb->len;
+		char *data;
+
+		if (unlikely(skb_tailroom(skb) < needed)) {
+			struct sk_buff *nskb;
+
+			nskb = skb_copy_expand(skb, 0, needed, GFP_ATOMIC);
+			if (!nskb) {
+				ret = NETDEV_TX_BUSY;
+				goto out_unlock;
+			}
+			dev_kfree_skb(skb);
+			skb = nskb;
+		}
+		data = skb_put(skb, needed);
+		memset(data, 0, needed);
+	}
+
 	/* point to the next available desc */
 	desc = &priv->tx_desc_cpu[priv->tx_curr_desc];
 	priv->tx_skb[priv->tx_curr_desc] = skb;
@@ -539,14 +622,14 @@ static int bcm_enet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 				       DMA_TO_DEVICE);
 
 	len_stat = (skb->len << DMADESC_LENGTH_SHIFT) & DMADESC_LENGTH_MASK;
-	len_stat |= DMADESC_ESOP_MASK |
+	len_stat |= (DMADESC_ESOP_MASK >> priv->dma_desc_shift) |
 		DMADESC_APPEND_CRC |
 		DMADESC_OWNER_MASK;
 
 	priv->tx_curr_desc++;
 	if (priv->tx_curr_desc == priv->tx_ring_size) {
 		priv->tx_curr_desc = 0;
-		len_stat |= DMADESC_WRAP_MASK;
+		len_stat |= (DMADESC_WRAP_MASK >> priv->dma_desc_shift);
 	}
 	priv->tx_desc_count--;
 
@@ -557,8 +640,8 @@ static int bcm_enet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	wmb();
 
 	/* kick tx dma */
-	enet_dma_writel(priv, ENETDMA_CHANCFG_EN_MASK,
-			ENETDMA_CHANCFG_REG(priv->tx_chan));
+	enet_dmac_writel(priv, priv->dma_chan_en_mask,
+				 ENETDMAC_CHANCFG, priv->tx_chan);
 
 	/* stop queue if no more desc available */
 	if (!priv->tx_desc_count)
@@ -686,6 +769,9 @@ static void bcm_enet_set_flow(struct bcm_enet_priv *priv, int rx_en, int tx_en)
 		val &= ~ENET_RXCFG_ENFLOW_MASK;
 	enet_writel(priv, val, ENET_RXCFG_REG);
 
+	if (!priv->dma_has_sram)
+		return;
+
 	/* tx flow control (pause frame generation) */
 	val = enet_dma_readl(priv, ENETDMA_CFG_REG);
 	if (tx_en)
@@ -833,8 +919,8 @@ static int bcm_enet_open(struct net_device *dev)
 
 	/* mask all interrupts and request them */
 	enet_writel(priv, 0, ENET_IRMASK_REG);
-	enet_dma_writel(priv, 0, ENETDMA_IRMASK_REG(priv->rx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_IRMASK_REG(priv->tx_chan));
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->tx_chan);
 
 	ret = request_irq(dev->irq, bcm_enet_isr_mac, 0, dev->name, dev);
 	if (ret)
@@ -909,8 +995,12 @@ static int bcm_enet_open(struct net_device *dev)
 	priv->rx_curr_desc = 0;
 
 	/* initialize flow control buffer allocation */
-	enet_dma_writel(priv, ENETDMA_BUFALLOC_FORCE_MASK | 0,
-			ENETDMA_BUFALLOC_REG(priv->rx_chan));
+	if (priv->dma_has_sram)
+		enet_dma_writel(priv, ENETDMA_BUFALLOC_FORCE_MASK | 0,
+				ENETDMA_BUFALLOC_REG(priv->rx_chan));
+	else
+		enet_dmac_writel(priv, ENETDMA_BUFALLOC_FORCE_MASK | 0,
+				ENETDMAC_BUFALLOC, priv->rx_chan);
 
 	if (bcm_enet_refill_rx(dev)) {
 		dev_err(kdev, "cannot allocate rx skb queue\n");
@@ -919,37 +1009,55 @@ static int bcm_enet_open(struct net_device *dev)
 	}
 
 	/* write rx & tx ring addresses */
-	enet_dma_writel(priv, priv->rx_desc_dma,
-			ENETDMA_RSTART_REG(priv->rx_chan));
-	enet_dma_writel(priv, priv->tx_desc_dma,
-			ENETDMA_RSTART_REG(priv->tx_chan));
+	if (priv->dma_has_sram) {
+		enet_dmas_writel(priv, priv->rx_desc_dma,
+				 ENETDMAS_RSTART_REG, priv->rx_chan);
+		enet_dmas_writel(priv, priv->tx_desc_dma,
+			 ENETDMAS_RSTART_REG, priv->tx_chan);
+	} else {
+		enet_dmac_writel(priv, priv->rx_desc_dma,
+				ENETDMAC_RSTART, priv->rx_chan);
+		enet_dmac_writel(priv, priv->tx_desc_dma,
+				ENETDMAC_RSTART, priv->tx_chan);
+	}
 
 	/* clear remaining state ram for rx & tx channel */
-	enet_dma_writel(priv, 0, ENETDMA_SRAM2_REG(priv->rx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_SRAM2_REG(priv->tx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_SRAM3_REG(priv->rx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_SRAM3_REG(priv->tx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_SRAM4_REG(priv->rx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_SRAM4_REG(priv->tx_chan));
+	if (priv->dma_has_sram) {
+		enet_dmas_writel(priv, 0, ENETDMAS_SRAM2_REG, priv->rx_chan);
+		enet_dmas_writel(priv, 0, ENETDMAS_SRAM2_REG, priv->tx_chan);
+		enet_dmas_writel(priv, 0, ENETDMAS_SRAM3_REG, priv->rx_chan);
+		enet_dmas_writel(priv, 0, ENETDMAS_SRAM3_REG, priv->tx_chan);
+		enet_dmas_writel(priv, 0, ENETDMAS_SRAM4_REG, priv->rx_chan);
+		enet_dmas_writel(priv, 0, ENETDMAS_SRAM4_REG, priv->tx_chan);
+	} else {
+		enet_dmac_writel(priv, 0, ENETDMAC_FC, priv->rx_chan);
+		enet_dmac_writel(priv, 0, ENETDMAC_FC, priv->tx_chan);
+	}
 
 	/* set max rx/tx length */
 	enet_writel(priv, priv->hw_mtu, ENET_RXMAXLEN_REG);
 	enet_writel(priv, priv->hw_mtu, ENET_TXMAXLEN_REG);
 
 	/* set dma maximum burst len */
-	enet_dma_writel(priv, BCMENET_DMA_MAXBURST,
-			ENETDMA_MAXBURST_REG(priv->rx_chan));
-	enet_dma_writel(priv, BCMENET_DMA_MAXBURST,
-			ENETDMA_MAXBURST_REG(priv->tx_chan));
+	enet_dmac_writel(priv, priv->dma_maxburst,
+			 ENETDMAC_MAXBURST, priv->rx_chan);
+	enet_dmac_writel(priv, priv->dma_maxburst,
+			 ENETDMAC_MAXBURST, priv->tx_chan);
 
 	/* set correct transmit fifo watermark */
 	enet_writel(priv, BCMENET_TX_FIFO_TRESH, ENET_TXWMARK_REG);
 
 	/* set flow control low/high threshold to 1/3 / 2/3 */
-	val = priv->rx_ring_size / 3;
-	enet_dma_writel(priv, val, ENETDMA_FLOWCL_REG(priv->rx_chan));
-	val = (priv->rx_ring_size * 2) / 3;
-	enet_dma_writel(priv, val, ENETDMA_FLOWCH_REG(priv->rx_chan));
+	if (priv->dma_has_sram) {
+		val = priv->rx_ring_size / 3;
+		enet_dma_writel(priv, val, ENETDMA_FLOWCL_REG(priv->rx_chan));
+		val = (priv->rx_ring_size * 2) / 3;
+		enet_dma_writel(priv, val, ENETDMA_FLOWCH_REG(priv->rx_chan));
+	} else {
+		enet_dmac_writel(priv, 5, ENETDMAC_FC, priv->rx_chan);
+		enet_dmac_writel(priv, priv->rx_ring_size, ENETDMAC_LEN, priv->rx_chan);
+		enet_dmac_writel(priv, priv->tx_ring_size, ENETDMAC_LEN, priv->tx_chan);
+	}
 
 	/* all set, enable mac and interrupts, start dma engine and
 	 * kick rx dma channel */
@@ -958,26 +1066,26 @@ static int bcm_enet_open(struct net_device *dev)
 	val |= ENET_CTL_ENABLE_MASK;
 	enet_writel(priv, val, ENET_CTL_REG);
 	enet_dma_writel(priv, ENETDMA_CFG_EN_MASK, ENETDMA_CFG_REG);
-	enet_dma_writel(priv, ENETDMA_CHANCFG_EN_MASK,
-			ENETDMA_CHANCFG_REG(priv->rx_chan));
+	enet_dmac_writel(priv, priv->dma_chan_en_mask,
+			 ENETDMAC_CHANCFG, priv->rx_chan);
 
 	/* watch "mib counters about to overflow" interrupt */
 	enet_writel(priv, ENET_IR_MIB, ENET_IR_REG);
 	enet_writel(priv, ENET_IR_MIB, ENET_IRMASK_REG);
 
 	/* watch "packet transferred" interrupt in rx and tx */
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IR_REG(priv->rx_chan));
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IR_REG(priv->tx_chan));
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IR, priv->rx_chan);
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IR, priv->tx_chan);
 
 	/* make sure we enable napi before rx interrupt  */
 	napi_enable(&priv->napi);
 
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IRMASK_REG(priv->rx_chan));
-	enet_dma_writel(priv, ENETDMA_IR_PKTDONE_MASK,
-			ENETDMA_IRMASK_REG(priv->tx_chan));
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, priv->dma_chan_int_mask,
+			 ENETDMAC_IRMASK, priv->tx_chan);
 
 	if (priv->has_phy)
 		phy_start(priv->phydev);
@@ -1057,14 +1165,14 @@ static void bcm_enet_disable_dma(struct bcm_enet_priv *priv, int chan)
 {
 	int limit;
 
-	enet_dma_writel(priv, 0, ENETDMA_CHANCFG_REG(chan));
+	enet_dmac_writel(priv, 0, ENETDMAC_CHANCFG, chan);
 
 	limit = 1000;
 	do {
 		u32 val;
 
-		val = enet_dma_readl(priv, ENETDMA_CHANCFG_REG(chan));
-		if (!(val & ENETDMA_CHANCFG_EN_MASK))
+		val = enet_dmac_readl(priv, ENETDMAC_CHANCFG, chan);
+		if (!(val & ENETDMAC_CHANCFG_EN_MASK))
 			break;
 		udelay(1);
 	} while (limit--);
@@ -1090,8 +1198,8 @@ static int bcm_enet_stop(struct net_device *dev)
 
 	/* mask all interrupts */
 	enet_writel(priv, 0, ENET_IRMASK_REG);
-	enet_dma_writel(priv, 0, ENETDMA_IRMASK_REG(priv->rx_chan));
-	enet_dma_writel(priv, 0, ENETDMA_IRMASK_REG(priv->tx_chan));
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->tx_chan);
 
 	/* make sure no mib update is scheduled */
 	cancel_work_sync(&priv->mib_update_task);
@@ -1328,6 +1436,20 @@ static void bcm_enet_get_ethtool_stats(struct net_device *netdev,
 	mutex_unlock(&priv->mib_update_lock);
 }
 
+static int bcm_enet_nway_reset(struct net_device *dev)
+{
+	struct bcm_enet_priv *priv;
+
+	priv = netdev_priv(dev);
+	if (priv->has_phy) {
+		if (!priv->phydev)
+			return -ENODEV;
+		return genphy_restart_aneg(priv->phydev);
+	}
+
+	return -EOPNOTSUPP;
+}
+
 static int bcm_enet_get_settings(struct net_device *dev,
 				 struct ethtool_cmd *cmd)
 {
@@ -1470,6 +1592,7 @@ static const struct ethtool_ops bcm_enet_ethtool_ops = {
 	.get_strings		= bcm_enet_get_strings,
 	.get_sset_count		= bcm_enet_get_sset_count,
 	.get_ethtool_stats      = bcm_enet_get_ethtool_stats,
+	.nway_reset		= bcm_enet_nway_reset,
 	.get_settings		= bcm_enet_get_settings,
 	.set_settings		= bcm_enet_set_settings,
 	.get_drvinfo		= bcm_enet_get_drvinfo,
@@ -1530,7 +1653,7 @@ static int compute_hw_mtu(struct bcm_enet_priv *priv, int mtu)
 	 * it's appended
 	 */
 	priv->rx_skb_size = ALIGN(actual_mtu + ETH_FCS_LEN,
-				  BCMENET_DMA_MAXBURST * 4);
+				  priv->dma_maxburst * 4);
 	return 0;
 }
 
@@ -1621,7 +1744,7 @@ static int bcm_enet_probe(struct platform_device *pdev)
 
 	/* stop if shared driver failed, assume driver->probe will be
 	 * called in the same order we register devices (correct ?) */
-	if (!bcm_enet_shared_base)
+	if (!bcm_enet_shared_base[0])
 		return -ENODEV;
 
 	res_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -1637,6 +1760,9 @@ static int bcm_enet_probe(struct platform_device *pdev)
 		return -ENOMEM;
 	priv = netdev_priv(dev);
 
+	priv->enet_is_sw = false;
+	priv->dma_maxburst = BCMENET_DMA_MAXBURST;
+
 	ret = compute_hw_mtu(priv, dev->mtu);
 	if (ret)
 		goto out;
@@ -1687,6 +1813,11 @@ static int bcm_enet_probe(struct platform_device *pdev)
 		priv->pause_tx = pd->pause_tx;
 		priv->force_duplex_full = pd->force_duplex_full;
 		priv->force_speed_100 = pd->force_speed_100;
+		priv->dma_chan_en_mask = pd->dma_chan_en_mask;
+		priv->dma_chan_int_mask = pd->dma_chan_int_mask;
+		priv->dma_chan_width = pd->dma_chan_width;
+		priv->dma_has_sram = pd->dma_has_sram;
+		priv->dma_desc_shift = pd->dma_desc_shift;
 	}
 
 	if (priv->mac_id == 0 && priv->has_phy && !priv->use_external_mii) {
@@ -1847,7 +1978,6 @@ static int bcm_enet_remove(struct platform_device *pdev)
 	clk_disable_unprepare(priv->mac_clk);
 	clk_put(priv->mac_clk);
 
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(dev);
 	return 0;
 }
@@ -1862,19 +1992,881 @@ struct platform_driver bcm63xx_enet_driver = {
 };
 
 /*
- * reserve & remap memory space shared between all macs
+ * switch mii access callbacks
  */
-static int bcm_enet_shared_probe(struct platform_device *pdev)
+static int bcmenet_sw_mdio_read(struct bcm_enet_priv *priv,
+				int ext, int phy_id, int location)
 {
-	struct resource *res;
+	u32 reg;
+	int ret;
 
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!res)
+	spin_lock_bh(&priv->enetsw_mdio_lock);
+	enetsw_writel(priv, 0, ENETSW_MDIOC_REG);
+
+	reg = ENETSW_MDIOC_RD_MASK |
+		(phy_id << ENETSW_MDIOC_PHYID_SHIFT) |
+		(location << ENETSW_MDIOC_REG_SHIFT);
+
+	if (ext)
+		reg |= ENETSW_MDIOC_EXT_MASK;
+
+	enetsw_writel(priv, reg, ENETSW_MDIOC_REG);
+	udelay(50);
+	ret = enetsw_readw(priv, ENETSW_MDIOD_REG);
+	spin_unlock_bh(&priv->enetsw_mdio_lock);
+	return ret;
+}
+
+static void bcmenet_sw_mdio_write(struct bcm_enet_priv *priv,
+				 int ext, int phy_id, int location,
+				 uint16_t data)
+{
+	u32 reg;
+
+	spin_lock_bh(&priv->enetsw_mdio_lock);
+	enetsw_writel(priv, 0, ENETSW_MDIOC_REG);
+
+	reg = ENETSW_MDIOC_WR_MASK |
+		(phy_id << ENETSW_MDIOC_PHYID_SHIFT) |
+		(location << ENETSW_MDIOC_REG_SHIFT);
+
+	if (ext)
+		reg |= ENETSW_MDIOC_EXT_MASK;
+
+	reg |= data;
+
+	enetsw_writel(priv, reg, ENETSW_MDIOC_REG);
+	udelay(50);
+	spin_unlock_bh(&priv->enetsw_mdio_lock);
+}
+
+static inline int bcm_enet_port_is_rgmii(int portid)
+{
+	return portid >= ENETSW_RGMII_PORT0;
+}
+
+/*
+ * enet sw PHY polling
+ */
+static void swphy_poll_timer(unsigned long data)
+{
+	struct bcm_enet_priv *priv = (struct bcm_enet_priv *)data;
+	unsigned int i;
+
+	for (i = 0; i < priv->num_ports; i++) {
+		struct bcm63xx_enetsw_port *port;
+		int val, j, up, advertise, lpa, lpa2, speed, duplex, media;
+		int external_phy = bcm_enet_port_is_rgmii(i);
+		u8 override;
+
+		port = &priv->used_ports[i];
+		if (!port->used)
+			continue;
+
+		if (port->bypass_link)
+			continue;
+
+		/* dummy read to clear */
+		for (j = 0; j < 2; j++)
+			val = bcmenet_sw_mdio_read(priv, external_phy,
+						   port->phy_id, MII_BMSR);
+
+		if (val == 0xffff)
+			continue;
+
+		up = (val & BMSR_LSTATUS) ? 1 : 0;
+		if (!(up ^ priv->sw_port_link[i]))
+			continue;
+
+		priv->sw_port_link[i] = up;
+
+		/* link changed */
+		if (!up) {
+			dev_info(&priv->pdev->dev, "link DOWN on %s\n",
+				 port->name);
+			enetsw_writeb(priv, ENETSW_PORTOV_ENABLE_MASK,
+				      ENETSW_PORTOV_REG(i));
+			enetsw_writeb(priv, ENETSW_PTCTRL_RXDIS_MASK |
+				      ENETSW_PTCTRL_TXDIS_MASK,
+				      ENETSW_PTCTRL_REG(i));
+			continue;
+		}
+
+		advertise = bcmenet_sw_mdio_read(priv, external_phy,
+						 port->phy_id, MII_ADVERTISE);
+
+		lpa = bcmenet_sw_mdio_read(priv, external_phy, port->phy_id,
+					   MII_LPA);
+
+		lpa2 = bcmenet_sw_mdio_read(priv, external_phy, port->phy_id,
+					    MII_STAT1000);
+
+		/* figure out media and duplex from advertise and LPA values */
+		media = mii_nway_result(lpa & advertise);
+		duplex = (media & ADVERTISE_FULL) ? 1 : 0;
+		if (lpa2 & LPA_1000FULL)
+			duplex = 1;
+
+		if (lpa2 & (LPA_1000FULL | LPA_1000HALF))
+			speed = 1000;
+		else {
+			if (media & (ADVERTISE_100FULL | ADVERTISE_100HALF))
+				speed = 100;
+			else
+				speed = 10;
+		}
+
+		dev_info(&priv->pdev->dev,
+			 "link UP on %s, %dMbps, %s-duplex\n",
+			 port->name, speed, duplex ? "full" : "half");
+
+		override = ENETSW_PORTOV_ENABLE_MASK |
+			ENETSW_PORTOV_LINKUP_MASK;
+
+		if (speed == 1000)
+			override |= ENETSW_IMPOV_1000_MASK;
+		else if (speed == 100)
+			override |= ENETSW_IMPOV_100_MASK;
+		if (duplex)
+			override |= ENETSW_IMPOV_FDX_MASK;
+
+		enetsw_writeb(priv, override, ENETSW_PORTOV_REG(i));
+		enetsw_writeb(priv, 0, ENETSW_PTCTRL_REG(i));
+	}
+
+	priv->swphy_poll.expires = jiffies + HZ;
+	add_timer(&priv->swphy_poll);
+}
+
+/*
+ * open callback, allocate dma rings & buffers and start rx operation
+ */
+static int bcm_enetsw_open(struct net_device *dev)
+{
+	struct bcm_enet_priv *priv;
+	struct device *kdev;
+	int i, ret;
+	unsigned int size;
+	void *p;
+	u32 val;
+
+	priv = netdev_priv(dev);
+	kdev = &priv->pdev->dev;
+
+	/* mask all interrupts and request them */
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->tx_chan);
+
+	ret = request_irq(priv->irq_rx, bcm_enet_isr_dma,
+			  IRQF_DISABLED, dev->name, dev);
+	if (ret)
+		goto out_freeirq;
+
+	if (priv->irq_tx != -1) {
+		ret = request_irq(priv->irq_tx, bcm_enet_isr_dma,
+				  IRQF_DISABLED, dev->name, dev);
+		if (ret)
+			goto out_freeirq_rx;
+	}
+
+	/* allocate rx dma ring */
+	size = priv->rx_ring_size * sizeof(struct bcm_enet_desc);
+	p = dma_alloc_coherent(kdev, size, &priv->rx_desc_dma, GFP_KERNEL);
+	if (!p) {
+		dev_err(kdev, "cannot allocate rx ring %u\n", size);
+		ret = -ENOMEM;
+		goto out_freeirq_tx;
+	}
+
+	memset(p, 0, size);
+	priv->rx_desc_alloc_size = size;
+	priv->rx_desc_cpu = p;
+
+	/* allocate tx dma ring */
+	size = priv->tx_ring_size * sizeof(struct bcm_enet_desc);
+	p = dma_alloc_coherent(kdev, size, &priv->tx_desc_dma, GFP_KERNEL);
+	if (!p) {
+		dev_err(kdev, "cannot allocate tx ring\n");
+		ret = -ENOMEM;
+		goto out_free_rx_ring;
+	}
+
+	memset(p, 0, size);
+	priv->tx_desc_alloc_size = size;
+	priv->tx_desc_cpu = p;
+
+	priv->tx_skb = kzalloc(sizeof(struct sk_buff *) * priv->tx_ring_size,
+			       GFP_KERNEL);
+	if (!priv->tx_skb) {
+		dev_err(kdev, "cannot allocate rx skb queue\n");
+		ret = -ENOMEM;
+		goto out_free_tx_ring;
+	}
+
+	priv->tx_desc_count = priv->tx_ring_size;
+	priv->tx_dirty_desc = 0;
+	priv->tx_curr_desc = 0;
+	spin_lock_init(&priv->tx_lock);
+
+	/* init & fill rx ring with skbs */
+	priv->rx_skb = kzalloc(sizeof(struct sk_buff *) * priv->rx_ring_size,
+			       GFP_KERNEL);
+	if (!priv->rx_skb) {
+		dev_err(kdev, "cannot allocate rx skb queue\n");
+		ret = -ENOMEM;
+		goto out_free_tx_skb;
+	}
+
+	priv->rx_desc_count = 0;
+	priv->rx_dirty_desc = 0;
+	priv->rx_curr_desc = 0;
+
+	/* disable all ports */
+	for (i = 0; i < priv->num_ports; i++) {
+		enetsw_writeb(priv, ENETSW_PORTOV_ENABLE_MASK,
+			      ENETSW_PORTOV_REG(i));
+		enetsw_writeb(priv, ENETSW_PTCTRL_RXDIS_MASK |
+			      ENETSW_PTCTRL_TXDIS_MASK,
+			      ENETSW_PTCTRL_REG(i));
+
+		priv->sw_port_link[i] = 0;
+	}
+
+	/* reset mib */
+	val = enetsw_readb(priv, ENETSW_GMCR_REG);
+	val |= ENETSW_GMCR_RST_MIB_MASK;
+	enetsw_writeb(priv, val, ENETSW_GMCR_REG);
+	mdelay(1);
+	val &= ~ENETSW_GMCR_RST_MIB_MASK;
+	enetsw_writeb(priv, val, ENETSW_GMCR_REG);
+	mdelay(1);
+
+	/* force CPU port state */
+	val = enetsw_readb(priv, ENETSW_IMPOV_REG);
+	val |= ENETSW_IMPOV_FORCE_MASK | ENETSW_IMPOV_LINKUP_MASK;
+	enetsw_writeb(priv, val, ENETSW_IMPOV_REG);
+
+	/* enable switch forward engine */
+	val = enetsw_readb(priv, ENETSW_SWMODE_REG);
+	val |= ENETSW_SWMODE_FWD_EN_MASK;
+	enetsw_writeb(priv, val, ENETSW_SWMODE_REG);
+
+	/* enable jumbo on all ports */
+	enetsw_writel(priv, 0x1ff, ENETSW_JMBCTL_PORT_REG);
+	enetsw_writew(priv, 9728, ENETSW_JMBCTL_MAXSIZE_REG);
+
+	/* initialize flow control buffer allocation */
+	enet_dma_writel(priv, ENETDMA_BUFALLOC_FORCE_MASK | 0,
+			ENETDMA_BUFALLOC_REG(priv->rx_chan));
+
+	if (bcm_enet_refill_rx(dev)) {
+		dev_err(kdev, "cannot allocate rx skb queue\n");
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	/* write rx & tx ring addresses */
+	enet_dmas_writel(priv, priv->rx_desc_dma,
+			 ENETDMAS_RSTART_REG, priv->rx_chan);
+	enet_dmas_writel(priv, priv->tx_desc_dma,
+			 ENETDMAS_RSTART_REG, priv->tx_chan);
+
+	/* clear remaining state ram for rx & tx channel */
+	enet_dmas_writel(priv, 0, ENETDMAS_SRAM2_REG, priv->rx_chan);
+	enet_dmas_writel(priv, 0, ENETDMAS_SRAM2_REG, priv->tx_chan);
+	enet_dmas_writel(priv, 0, ENETDMAS_SRAM3_REG, priv->rx_chan);
+	enet_dmas_writel(priv, 0, ENETDMAS_SRAM3_REG, priv->tx_chan);
+	enet_dmas_writel(priv, 0, ENETDMAS_SRAM4_REG, priv->rx_chan);
+	enet_dmas_writel(priv, 0, ENETDMAS_SRAM4_REG, priv->tx_chan);
+
+	/* set dma maximum burst len */
+	enet_dmac_writel(priv, priv->dma_maxburst,
+			 ENETDMAC_MAXBURST, priv->rx_chan);
+	enet_dmac_writel(priv, priv->dma_maxburst,
+			 ENETDMAC_MAXBURST, priv->tx_chan);
+
+	/* set flow control low/high threshold to 1/3 / 2/3 */
+	val = priv->rx_ring_size / 3;
+	enet_dma_writel(priv, val, ENETDMA_FLOWCL_REG(priv->rx_chan));
+	val = (priv->rx_ring_size * 2) / 3;
+	enet_dma_writel(priv, val, ENETDMA_FLOWCH_REG(priv->rx_chan));
+
+	/* all set, enable mac and interrupts, start dma engine and
+	 * kick rx dma channel
+	 */
+	wmb();
+	enet_dma_writel(priv, ENETDMA_CFG_EN_MASK, ENETDMA_CFG_REG);
+	enet_dmac_writel(priv, ENETDMAC_CHANCFG_EN_MASK,
+			 ENETDMAC_CHANCFG, priv->rx_chan);
+
+	/* watch "packet transferred" interrupt in rx and tx */
+	enet_dmac_writel(priv, ENETDMAC_IR_PKTDONE_MASK,
+			 ENETDMAC_IR, priv->rx_chan);
+	enet_dmac_writel(priv, ENETDMAC_IR_PKTDONE_MASK,
+			 ENETDMAC_IR, priv->tx_chan);
+
+	/* make sure we enable napi before rx interrupt  */
+	napi_enable(&priv->napi);
+
+	enet_dmac_writel(priv, ENETDMAC_IR_PKTDONE_MASK,
+			 ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, ENETDMAC_IR_PKTDONE_MASK,
+			 ENETDMAC_IRMASK, priv->tx_chan);
+
+	netif_carrier_on(dev);
+	netif_start_queue(dev);
+
+	/* apply override config for bypass_link ports here. */
+	for (i = 0; i < priv->num_ports; i++) {
+		struct bcm63xx_enetsw_port *port;
+		u8 override;
+		port = &priv->used_ports[i];
+		if (!port->used)
+			continue;
+
+		if (!port->bypass_link)
+			continue;
+
+		override = ENETSW_PORTOV_ENABLE_MASK |
+			ENETSW_PORTOV_LINKUP_MASK;
+
+		switch (port->force_speed) {
+		case 1000:
+			override |= ENETSW_IMPOV_1000_MASK;
+			break;
+		case 100:
+			override |= ENETSW_IMPOV_100_MASK;
+			break;
+		case 10:
+			break;
+		default:
+			pr_warn("invalid forced speed on port %s: assume 10\n",
+			       port->name);
+			break;
+		}
+
+		if (port->force_duplex_full)
+			override |= ENETSW_IMPOV_FDX_MASK;
+
+
+		enetsw_writeb(priv, override, ENETSW_PORTOV_REG(i));
+		enetsw_writeb(priv, 0, ENETSW_PTCTRL_REG(i));
+	}
+
+	/* start phy polling timer */
+	init_timer(&priv->swphy_poll);
+	priv->swphy_poll.function = swphy_poll_timer;
+	priv->swphy_poll.data = (unsigned long)priv;
+	priv->swphy_poll.expires = jiffies;
+	add_timer(&priv->swphy_poll);
+	return 0;
+
+out:
+	for (i = 0; i < priv->rx_ring_size; i++) {
+		struct bcm_enet_desc *desc;
+
+		if (!priv->rx_skb[i])
+			continue;
+
+		desc = &priv->rx_desc_cpu[i];
+		dma_unmap_single(kdev, desc->address, priv->rx_skb_size,
+				 DMA_FROM_DEVICE);
+		kfree_skb(priv->rx_skb[i]);
+	}
+	kfree(priv->rx_skb);
+
+out_free_tx_skb:
+	kfree(priv->tx_skb);
+
+out_free_tx_ring:
+	dma_free_coherent(kdev, priv->tx_desc_alloc_size,
+			  priv->tx_desc_cpu, priv->tx_desc_dma);
+
+out_free_rx_ring:
+	dma_free_coherent(kdev, priv->rx_desc_alloc_size,
+			  priv->rx_desc_cpu, priv->rx_desc_dma);
+
+out_freeirq_tx:
+	if (priv->irq_tx != -1)
+		free_irq(priv->irq_tx, dev);
+
+out_freeirq_rx:
+	free_irq(priv->irq_rx, dev);
+
+out_freeirq:
+	return ret;
+}
+
+/* stop callback */
+static int bcm_enetsw_stop(struct net_device *dev)
+{
+	struct bcm_enet_priv *priv;
+	struct device *kdev;
+	int i;
+
+	priv = netdev_priv(dev);
+	kdev = &priv->pdev->dev;
+
+	del_timer_sync(&priv->swphy_poll);
+	netif_stop_queue(dev);
+	napi_disable(&priv->napi);
+	del_timer_sync(&priv->rx_timeout);
+
+	/* mask all interrupts */
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->rx_chan);
+	enet_dmac_writel(priv, 0, ENETDMAC_IRMASK, priv->tx_chan);
+
+	/* disable dma & mac */
+	bcm_enet_disable_dma(priv, priv->tx_chan);
+	bcm_enet_disable_dma(priv, priv->rx_chan);
+
+	/* force reclaim of all tx buffers */
+	bcm_enet_tx_reclaim(dev, 1);
+
+	/* free the rx skb ring */
+	for (i = 0; i < priv->rx_ring_size; i++) {
+		struct bcm_enet_desc *desc;
+
+		if (!priv->rx_skb[i])
+			continue;
+
+		desc = &priv->rx_desc_cpu[i];
+		dma_unmap_single(kdev, desc->address, priv->rx_skb_size,
+				 DMA_FROM_DEVICE);
+		kfree_skb(priv->rx_skb[i]);
+	}
+
+	/* free remaining allocated memory */
+	kfree(priv->rx_skb);
+	kfree(priv->tx_skb);
+	dma_free_coherent(kdev, priv->rx_desc_alloc_size,
+			  priv->rx_desc_cpu, priv->rx_desc_dma);
+	dma_free_coherent(kdev, priv->tx_desc_alloc_size,
+			  priv->tx_desc_cpu, priv->tx_desc_dma);
+	if (priv->irq_tx != -1)
+		free_irq(priv->irq_tx, dev);
+	free_irq(priv->irq_rx, dev);
+
+	return 0;
+}
+
+/* try to sort out phy external status by walking the used_port field
+ * in the bcm_enet_priv structure. in case the phy address is not
+ * assigned to any physical port on the switch, assume it is external
+ * (and yell at the user).
+ */
+static int bcm_enetsw_phy_is_external(struct bcm_enet_priv *priv, int phy_id)
+{
+	int i;
+
+	for (i = 0; i < priv->num_ports; ++i) {
+		if (!priv->used_ports[i].used)
+			continue;
+		if (priv->used_ports[i].phy_id == phy_id)
+			return bcm_enet_port_is_rgmii(i);
+	}
+
+	printk_once(KERN_WARNING  "bcm63xx_enet: could not find a used port with phy_id %i, assuming phy is external\n",
+		    phy_id);
+	return 1;
+}
+
+/* can't use bcmenet_sw_mdio_read directly as we need to sort out
+ * external/internal status of the given phy_id first.
+ */
+static int bcm_enetsw_mii_mdio_read(struct net_device *dev, int phy_id,
+				    int location)
+{
+	struct bcm_enet_priv *priv;
+
+	priv = netdev_priv(dev);
+	return bcmenet_sw_mdio_read(priv,
+				    bcm_enetsw_phy_is_external(priv, phy_id),
+				    phy_id, location);
+}
+
+/* can't use bcmenet_sw_mdio_write directly as we need to sort out
+ * external/internal status of the given phy_id first.
+ */
+static void bcm_enetsw_mii_mdio_write(struct net_device *dev, int phy_id,
+				      int location,
+				      int val)
+{
+	struct bcm_enet_priv *priv;
+
+	priv = netdev_priv(dev);
+	bcmenet_sw_mdio_write(priv, bcm_enetsw_phy_is_external(priv, phy_id),
+			      phy_id, location, val);
+}
+
+static int bcm_enetsw_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
+{
+	struct mii_if_info mii;
+
+	mii.dev = dev;
+	mii.mdio_read = bcm_enetsw_mii_mdio_read;
+	mii.mdio_write = bcm_enetsw_mii_mdio_write;
+	mii.phy_id = 0;
+	mii.phy_id_mask = 0x3f;
+	mii.reg_num_mask = 0x1f;
+	return generic_mii_ioctl(&mii, if_mii(rq), cmd, NULL);
+
+}
+
+static const struct net_device_ops bcm_enetsw_ops = {
+	.ndo_open		= bcm_enetsw_open,
+	.ndo_stop		= bcm_enetsw_stop,
+	.ndo_start_xmit		= bcm_enet_start_xmit,
+	.ndo_change_mtu		= bcm_enet_change_mtu,
+	.ndo_do_ioctl		= bcm_enetsw_ioctl,
+};
+
+
+static const struct bcm_enet_stats bcm_enetsw_gstrings_stats[] = {
+	{ "rx_packets", DEV_STAT(rx_packets), -1 },
+	{ "tx_packets",	DEV_STAT(tx_packets), -1 },
+	{ "rx_bytes", DEV_STAT(rx_bytes), -1 },
+	{ "tx_bytes", DEV_STAT(tx_bytes), -1 },
+	{ "rx_errors", DEV_STAT(rx_errors), -1 },
+	{ "tx_errors", DEV_STAT(tx_errors), -1 },
+	{ "rx_dropped",	DEV_STAT(rx_dropped), -1 },
+	{ "tx_dropped",	DEV_STAT(tx_dropped), -1 },
+
+	{ "tx_good_octets", GEN_STAT(mib.tx_gd_octets), ETHSW_MIB_RX_GD_OCT },
+	{ "tx_unicast", GEN_STAT(mib.tx_unicast), ETHSW_MIB_RX_BRDCAST },
+	{ "tx_broadcast", GEN_STAT(mib.tx_brdcast), ETHSW_MIB_RX_BRDCAST },
+	{ "tx_multicast", GEN_STAT(mib.tx_mult), ETHSW_MIB_RX_MULT },
+	{ "tx_64_octets", GEN_STAT(mib.tx_64), ETHSW_MIB_RX_64 },
+	{ "tx_65_127_oct", GEN_STAT(mib.tx_65_127), ETHSW_MIB_RX_65_127 },
+	{ "tx_128_255_oct", GEN_STAT(mib.tx_128_255), ETHSW_MIB_RX_128_255 },
+	{ "tx_256_511_oct", GEN_STAT(mib.tx_256_511), ETHSW_MIB_RX_256_511 },
+	{ "tx_512_1023_oct", GEN_STAT(mib.tx_512_1023), ETHSW_MIB_RX_512_1023},
+	{ "tx_1024_1522_oct", GEN_STAT(mib.tx_1024_max),
+	  ETHSW_MIB_RX_1024_1522 },
+	{ "tx_1523_2047_oct", GEN_STAT(mib.tx_1523_2047),
+	  ETHSW_MIB_RX_1523_2047 },
+	{ "tx_2048_4095_oct", GEN_STAT(mib.tx_2048_4095),
+	  ETHSW_MIB_RX_2048_4095 },
+	{ "tx_4096_8191_oct", GEN_STAT(mib.tx_4096_8191),
+	  ETHSW_MIB_RX_4096_8191 },
+	{ "tx_8192_9728_oct", GEN_STAT(mib.tx_8192_9728),
+	  ETHSW_MIB_RX_8192_9728 },
+	{ "tx_oversize", GEN_STAT(mib.tx_ovr), ETHSW_MIB_RX_OVR },
+	{ "tx_oversize_drop", GEN_STAT(mib.tx_ovr), ETHSW_MIB_RX_OVR_DISC },
+	{ "tx_dropped",	GEN_STAT(mib.tx_drop), ETHSW_MIB_RX_DROP },
+	{ "tx_undersize", GEN_STAT(mib.tx_underrun), ETHSW_MIB_RX_UND },
+	{ "tx_pause", GEN_STAT(mib.tx_pause), ETHSW_MIB_RX_PAUSE },
+
+	{ "rx_good_octets", GEN_STAT(mib.rx_gd_octets), ETHSW_MIB_TX_ALL_OCT },
+	{ "rx_broadcast", GEN_STAT(mib.rx_brdcast), ETHSW_MIB_TX_BRDCAST },
+	{ "rx_multicast", GEN_STAT(mib.rx_mult), ETHSW_MIB_TX_MULT },
+	{ "rx_unicast", GEN_STAT(mib.rx_unicast), ETHSW_MIB_TX_MULT },
+	{ "rx_pause", GEN_STAT(mib.rx_pause), ETHSW_MIB_TX_PAUSE },
+	{ "rx_dropped", GEN_STAT(mib.rx_drop), ETHSW_MIB_TX_DROP_PKTS },
+
+};
+
+#define BCM_ENETSW_STATS_LEN	\
+	(sizeof(bcm_enetsw_gstrings_stats) / sizeof(struct bcm_enet_stats))
+
+static void bcm_enetsw_get_strings(struct net_device *netdev,
+				   u32 stringset, u8 *data)
+{
+	int i;
+
+	switch (stringset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < BCM_ENETSW_STATS_LEN; i++) {
+			memcpy(data + i * ETH_GSTRING_LEN,
+			       bcm_enetsw_gstrings_stats[i].stat_string,
+			       ETH_GSTRING_LEN);
+		}
+		break;
+	}
+}
+
+static int bcm_enetsw_get_sset_count(struct net_device *netdev,
+				     int string_set)
+{
+	switch (string_set) {
+	case ETH_SS_STATS:
+		return BCM_ENETSW_STATS_LEN;
+	default:
+		return -EINVAL;
+	}
+}
+
+static void bcm_enetsw_get_drvinfo(struct net_device *netdev,
+				   struct ethtool_drvinfo *drvinfo)
+{
+	strncpy(drvinfo->driver, bcm_enet_driver_name, 32);
+	strncpy(drvinfo->version, bcm_enet_driver_version, 32);
+	strncpy(drvinfo->fw_version, "N/A", 32);
+	strncpy(drvinfo->bus_info, "bcm63xx", 32);
+	drvinfo->n_stats = BCM_ENETSW_STATS_LEN;
+}
+
+static void bcm_enetsw_get_ethtool_stats(struct net_device *netdev,
+					 struct ethtool_stats *stats,
+					 u64 *data)
+{
+	struct bcm_enet_priv *priv;
+	int i;
+
+	priv = netdev_priv(netdev);
+
+	for (i = 0; i < BCM_ENETSW_STATS_LEN; i++) {
+		const struct bcm_enet_stats *s;
+		u32 lo, hi;
+		char *p;
+		int reg;
+
+		s = &bcm_enetsw_gstrings_stats[i];
+
+		reg = s->mib_reg;
+		if (reg == -1)
+			continue;
+
+		lo = enetsw_readl(priv, ENETSW_MIB_REG(reg));
+		p = (char *)priv + s->stat_offset;
+
+		if (s->sizeof_stat == sizeof(u64)) {
+			hi = enetsw_readl(priv, ENETSW_MIB_REG(reg + 1));
+			*(u64 *)p = ((u64)hi << 32 | lo);
+		} else {
+			*(u32 *)p = lo;
+		}
+	}
+
+	for (i = 0; i < BCM_ENETSW_STATS_LEN; i++) {
+		const struct bcm_enet_stats *s;
+		char *p;
+
+		s = &bcm_enetsw_gstrings_stats[i];
+
+		if (s->mib_reg == -1)
+			p = (char *)&netdev->stats + s->stat_offset;
+		else
+			p = (char *)priv + s->stat_offset;
+
+		data[i] = (s->sizeof_stat == sizeof(u64)) ?
+			*(u64 *)p : *(u32 *)p;
+	}
+}
+
+static void bcm_enetsw_get_ringparam(struct net_device *dev,
+				     struct ethtool_ringparam *ering)
+{
+	struct bcm_enet_priv *priv;
+
+	priv = netdev_priv(dev);
+
+	/* rx/tx ring is actually only limited by memory */
+	ering->rx_max_pending = 8192;
+	ering->tx_max_pending = 8192;
+	ering->rx_mini_max_pending = 0;
+	ering->rx_jumbo_max_pending = 0;
+	ering->rx_pending = priv->rx_ring_size;
+	ering->tx_pending = priv->tx_ring_size;
+}
+
+static int bcm_enetsw_set_ringparam(struct net_device *dev,
+				    struct ethtool_ringparam *ering)
+{
+	struct bcm_enet_priv *priv;
+	int was_running;
+
+	priv = netdev_priv(dev);
+
+	was_running = 0;
+	if (netif_running(dev)) {
+		bcm_enetsw_stop(dev);
+		was_running = 1;
+	}
+
+	priv->rx_ring_size = ering->rx_pending;
+	priv->tx_ring_size = ering->tx_pending;
+
+	if (was_running) {
+		int err;
+
+		err = bcm_enetsw_open(dev);
+		if (err)
+			dev_close(dev);
+	}
+	return 0;
+}
+
+static struct ethtool_ops bcm_enetsw_ethtool_ops = {
+	.get_strings		= bcm_enetsw_get_strings,
+	.get_sset_count		= bcm_enetsw_get_sset_count,
+	.get_ethtool_stats      = bcm_enetsw_get_ethtool_stats,
+	.get_drvinfo		= bcm_enetsw_get_drvinfo,
+	.get_ringparam		= bcm_enetsw_get_ringparam,
+	.set_ringparam		= bcm_enetsw_set_ringparam,
+};
+
+/* allocate netdevice, request register memory and register device. */
+static int bcm_enetsw_probe(struct platform_device *pdev)
+{
+	struct bcm_enet_priv *priv;
+	struct net_device *dev;
+	struct bcm63xx_enetsw_platform_data *pd;
+	struct resource *res_mem;
+	int ret, irq_rx, irq_tx;
+
+	/* stop if shared driver failed, assume driver->probe will be
+	 * called in the same order we register devices (correct ?)
+	 */
+	if (!bcm_enet_shared_base[0])
+		return -ENODEV;
+
+	res_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq_rx = platform_get_irq(pdev, 0);
+	irq_tx = platform_get_irq(pdev, 1);
+	if (!res_mem || irq_rx < 0)
 		return -ENODEV;
 
-	bcm_enet_shared_base = devm_request_and_ioremap(&pdev->dev, res);
-	if (!bcm_enet_shared_base)
+	ret = 0;
+	dev = alloc_etherdev(sizeof(*priv));
+	if (!dev)
 		return -ENOMEM;
+	priv = netdev_priv(dev);
+	memset(priv, 0, sizeof(*priv));
+
+	/* initialize default and fetch platform data */
+	priv->enet_is_sw = true;
+	priv->irq_rx = irq_rx;
+	priv->irq_tx = irq_tx;
+	priv->rx_ring_size = BCMENET_DEF_RX_DESC;
+	priv->tx_ring_size = BCMENET_DEF_TX_DESC;
+	priv->dma_maxburst = BCMENETSW_DMA_MAXBURST;
+
+	pd = pdev->dev.platform_data;
+	if (pd) {
+		memcpy(dev->dev_addr, pd->mac_addr, ETH_ALEN);
+		memcpy(priv->used_ports, pd->used_ports,
+		       sizeof(pd->used_ports));
+		priv->num_ports = pd->num_ports;
+		priv->dma_has_sram = pd->dma_has_sram;
+		priv->dma_chan_en_mask = pd->dma_chan_en_mask;
+		priv->dma_chan_int_mask = pd->dma_chan_int_mask;
+		priv->dma_chan_width = pd->dma_chan_width;
+	}
+
+	ret = compute_hw_mtu(priv, dev->mtu);
+	if (ret)
+		goto out;
+
+	if (!request_mem_region(res_mem->start, resource_size(res_mem),
+				"bcm63xx_enetsw")) {
+		ret = -EBUSY;
+		goto out;
+	}
+
+	priv->base = ioremap(res_mem->start, resource_size(res_mem));
+	if (priv->base == NULL) {
+		ret = -ENOMEM;
+		goto out_release_mem;
+	}
+
+	priv->mac_clk = clk_get(&pdev->dev, "enetsw");
+	if (IS_ERR(priv->mac_clk)) {
+		ret = PTR_ERR(priv->mac_clk);
+		goto out_unmap;
+	}
+	clk_enable(priv->mac_clk);
+
+	priv->rx_chan = 0;
+	priv->tx_chan = 1;
+	spin_lock_init(&priv->rx_lock);
+
+	/* init rx timeout (used for oom) */
+	init_timer(&priv->rx_timeout);
+	priv->rx_timeout.function = bcm_enet_refill_rx_timer;
+	priv->rx_timeout.data = (unsigned long)dev;
+
+	/* register netdevice */
+	dev->netdev_ops = &bcm_enetsw_ops;
+	netif_napi_add(dev, &priv->napi, bcm_enet_poll, 16);
+	SET_ETHTOOL_OPS(dev, &bcm_enetsw_ethtool_ops);
+	SET_NETDEV_DEV(dev, &pdev->dev);
+
+	spin_lock_init(&priv->enetsw_mdio_lock);
+
+	ret = register_netdev(dev);
+	if (ret)
+		goto out_put_clk;
+
+	netif_carrier_off(dev);
+	platform_set_drvdata(pdev, dev);
+	priv->pdev = pdev;
+	priv->net_dev = dev;
+
+	return 0;
+
+out_put_clk:
+	clk_put(priv->mac_clk);
+
+out_unmap:
+	iounmap(priv->base);
+
+out_release_mem:
+	release_mem_region(res_mem->start, resource_size(res_mem));
+out:
+	free_netdev(dev);
+	return ret;
+}
+
+
+/* exit func, stops hardware and unregisters netdevice */
+static int bcm_enetsw_remove(struct platform_device *pdev)
+{
+	struct bcm_enet_priv *priv;
+	struct net_device *dev;
+	struct resource *res;
+
+	/* stop netdevice */
+	dev = platform_get_drvdata(pdev);
+	priv = netdev_priv(dev);
+	unregister_netdev(dev);
+
+	/* release device resources */
+	iounmap(priv->base);
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	release_mem_region(res->start, resource_size(res));
+
+	platform_set_drvdata(pdev, NULL);
+	free_netdev(dev);
+	return 0;
+}
+
+struct platform_driver bcm63xx_enetsw_driver = {
+	.probe	= bcm_enetsw_probe,
+	.remove	= bcm_enetsw_remove,
+	.driver	= {
+		.name	= "bcm63xx_enetsw",
+		.owner  = THIS_MODULE,
+	},
+};
+
+/* reserve & remap memory space shared between all macs */
+static int bcm_enet_shared_probe(struct platform_device *pdev)
+{
+	struct resource *res;
+	void __iomem *p[3];
+	unsigned int i;
+
+	memset(bcm_enet_shared_base, 0, sizeof(bcm_enet_shared_base));
+
+	for (i = 0; i < 3; i++) {
+		res = platform_get_resource(pdev, IORESOURCE_MEM, i);
+		p[i] = devm_ioremap_resource(&pdev->dev, res);
+		if (IS_ERR(p[i]))
+			return PTR_ERR(p[i]);
+	}
+
+	memcpy(bcm_enet_shared_base, p, sizeof(bcm_enet_shared_base));
 
 	return 0;
 }
@@ -1884,8 +2876,7 @@ static int bcm_enet_shared_remove(struct platform_device *pdev)
 	return 0;
 }
 
-/*
- * this "shared" driver is needed because both macs share a single
+/* this "shared" driver is needed because both macs share a single
  * address space
  */
 struct platform_driver bcm63xx_enet_shared_driver = {
@@ -1897,9 +2888,7 @@ struct platform_driver bcm63xx_enet_shared_driver = {
 	},
 };
 
-/*
- * entry point
- */
+/* entry point */
 static int __init bcm_enet_init(void)
 {
 	int ret;
@@ -1912,12 +2901,19 @@ static int __init bcm_enet_init(void)
 	if (ret)
 		platform_driver_unregister(&bcm63xx_enet_shared_driver);
 
+	ret = platform_driver_register(&bcm63xx_enetsw_driver);
+	if (ret) {
+		platform_driver_unregister(&bcm63xx_enet_driver);
+		platform_driver_unregister(&bcm63xx_enet_shared_driver);
+	}
+
 	return ret;
 }
 
 static void __exit bcm_enet_exit(void)
 {
 	platform_driver_unregister(&bcm63xx_enet_driver);
+	platform_driver_unregister(&bcm63xx_enetsw_driver);
 	platform_driver_unregister(&bcm63xx_enet_shared_driver);
 }
 
diff --git a/drivers/net/ethernet/broadcom/bcm63xx_enet.h b/drivers/net/ethernet/broadcom/bcm63xx_enet.h
index 133d585..f55af43 100644
--- a/drivers/net/ethernet/broadcom/bcm63xx_enet.h
+++ b/drivers/net/ethernet/broadcom/bcm63xx_enet.h
@@ -18,6 +18,7 @@
 
 /* maximum burst len for dma (4 bytes unit) */
 #define BCMENET_DMA_MAXBURST	16
+#define BCMENETSW_DMA_MAXBURST	8
 
 /* tx transmit threshold (4 bytes unit), fifo is 256 bytes, the value
  * must be low enough so that a DMA transfer of above burst length can
@@ -84,11 +85,60 @@
 #define ETH_MIB_RX_CNTRL			54
 
 
+/*
+ * SW MIB Counters register definitions
+*/
+#define ETHSW_MIB_TX_ALL_OCT			0
+#define ETHSW_MIB_TX_DROP_PKTS			2
+#define ETHSW_MIB_TX_QOS_PKTS			3
+#define ETHSW_MIB_TX_BRDCAST			4
+#define ETHSW_MIB_TX_MULT			5
+#define ETHSW_MIB_TX_UNI			6
+#define ETHSW_MIB_TX_COL			7
+#define ETHSW_MIB_TX_1_COL			8
+#define ETHSW_MIB_TX_M_COL			9
+#define ETHSW_MIB_TX_DEF			10
+#define ETHSW_MIB_TX_LATE			11
+#define ETHSW_MIB_TX_EX_COL			12
+#define ETHSW_MIB_TX_PAUSE			14
+#define ETHSW_MIB_TX_QOS_OCT			15
+
+#define ETHSW_MIB_RX_ALL_OCT			17
+#define ETHSW_MIB_RX_UND			19
+#define ETHSW_MIB_RX_PAUSE			20
+#define ETHSW_MIB_RX_64				21
+#define ETHSW_MIB_RX_65_127			22
+#define ETHSW_MIB_RX_128_255			23
+#define ETHSW_MIB_RX_256_511			24
+#define ETHSW_MIB_RX_512_1023			25
+#define ETHSW_MIB_RX_1024_1522			26
+#define ETHSW_MIB_RX_OVR			27
+#define ETHSW_MIB_RX_JAB			28
+#define ETHSW_MIB_RX_ALIGN			29
+#define ETHSW_MIB_RX_CRC			30
+#define ETHSW_MIB_RX_GD_OCT			31
+#define ETHSW_MIB_RX_DROP			33
+#define ETHSW_MIB_RX_UNI			34
+#define ETHSW_MIB_RX_MULT			35
+#define ETHSW_MIB_RX_BRDCAST			36
+#define ETHSW_MIB_RX_SA_CHANGE			37
+#define ETHSW_MIB_RX_FRAG			38
+#define ETHSW_MIB_RX_OVR_DISC			39
+#define ETHSW_MIB_RX_SYM			40
+#define ETHSW_MIB_RX_QOS_PKTS			41
+#define ETHSW_MIB_RX_QOS_OCT			42
+#define ETHSW_MIB_RX_1523_2047			44
+#define ETHSW_MIB_RX_2048_4095			45
+#define ETHSW_MIB_RX_4096_8191			46
+#define ETHSW_MIB_RX_8192_9728			47
+
+
 struct bcm_enet_mib_counters {
 	u64 tx_gd_octets;
 	u32 tx_gd_pkts;
 	u32 tx_all_octets;
 	u32 tx_all_pkts;
+	u32 tx_unicast;
 	u32 tx_brdcast;
 	u32 tx_mult;
 	u32 tx_64;
@@ -97,7 +147,12 @@ struct bcm_enet_mib_counters {
 	u32 tx_256_511;
 	u32 tx_512_1023;
 	u32 tx_1024_max;
+	u32 tx_1523_2047;
+	u32 tx_2048_4095;
+	u32 tx_4096_8191;
+	u32 tx_8192_9728;
 	u32 tx_jab;
+	u32 tx_drop;
 	u32 tx_ovr;
 	u32 tx_frag;
 	u32 tx_underrun;
@@ -114,6 +169,7 @@ struct bcm_enet_mib_counters {
 	u32 rx_all_octets;
 	u32 rx_all_pkts;
 	u32 rx_brdcast;
+	u32 rx_unicast;
 	u32 rx_mult;
 	u32 rx_64;
 	u32 rx_65_127;
@@ -197,6 +253,9 @@ struct bcm_enet_priv {
 	/* number of dma desc in tx ring */
 	int tx_ring_size;
 
+	/* maximum dma burst size */
+	int dma_maxburst;
+
 	/* cpu view of rx dma ring */
 	struct bcm_enet_desc *tx_desc_cpu;
 
@@ -269,6 +328,33 @@ struct bcm_enet_priv {
 
 	/* maximum hardware transmit/receive size */
 	unsigned int hw_mtu;
+
+	bool enet_is_sw;
+
+	/* port mapping for switch devices */
+	int num_ports;
+	struct bcm63xx_enetsw_port used_ports[ENETSW_MAX_PORT];
+	int sw_port_link[ENETSW_MAX_PORT];
+
+	/* used to poll switch port state */
+	struct timer_list swphy_poll;
+	spinlock_t enetsw_mdio_lock;
+
+	/* dma channel enable mask */
+	u32 dma_chan_en_mask;
+
+	/* dma channel interrupt mask */
+	u32 dma_chan_int_mask;
+
+	/* DMA engine has internal SRAM */
+	bool dma_has_sram;
+
+	/* dma channel width */
+	unsigned int dma_chan_width;
+
+	/* dma descriptor shift value */
+	unsigned int dma_desc_shift;
 };
 
+
 #endif /* ! BCM63XX_ENET_H_ */
diff --git a/drivers/net/ethernet/broadcom/bnx2.c b/drivers/net/ethernet/broadcom/bnx2.c
index 5d20449..6a2de1d 100644
--- a/drivers/net/ethernet/broadcom/bnx2.c
+++ b/drivers/net/ethernet/broadcom/bnx2.c
@@ -8104,7 +8104,7 @@ bnx2_init_board(struct pci_dev *pdev, struct net_device *dev)
 
 	pci_set_master(pdev);
 
-	bp->pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM);
+	bp->pm_cap = pdev->pm_cap;
 	if (bp->pm_cap == 0) {
 		dev_err(&pdev->dev,
 			"Cannot find power management capability, aborting\n");
@@ -8764,18 +8764,4 @@ static struct pci_driver bnx2_pci_driver = {
 	.err_handler	= &bnx2_err_handler,
 };
 
-static int __init bnx2_init(void)
-{
-	return pci_register_driver(&bnx2_pci_driver);
-}
-
-static void __exit bnx2_cleanup(void)
-{
-	pci_unregister_driver(&bnx2_pci_driver);
-}
-
-module_init(bnx2_init);
-module_exit(bnx2_cleanup);
-
-
-
+module_pci_driver(bnx2_pci_driver);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 3dba2a7..dedbd76 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -34,12 +34,10 @@
 #define BCM_DCBNL
 #endif
 
-
 #include "bnx2x_hsi.h"
 
 #include "../cnic_if.h"
 
-
 #define BNX2X_MIN_MSIX_VEC_CNT(bp)		((bp)->min_msix_vec_cnt)
 
 #include <linux/mdio.h>
@@ -114,7 +112,6 @@ do {								\
 #define BNX2X_ERROR(fmt, ...)					\
 	pr_err("[%s:%d]" fmt, __func__, __LINE__, ##__VA_ARGS__)
 
-
 /* before we have a dev->name use dev_info() */
 #define BNX2X_DEV_INFO(fmt, ...)				 \
 do {								 \
@@ -147,7 +144,6 @@ do {						\
 #define U64_HI(x)			((u32)(((u64)(x)) >> 32))
 #define HILO_U64(hi, lo)		((((u64)(hi)) << 32) + (lo))
 
-
 #define REG_ADDR(bp, offset)		((bp->regview) + (offset))
 
 #define REG_RD(bp, offset)		readl(REG_ADDR(bp, offset))
@@ -366,7 +362,7 @@ union db_prod {
 /*
  * Number of required  SGEs is the sum of two:
  * 1. Number of possible opened aggregations (next packet for
- *    these aggregations will probably consume SGE immidiatelly)
+ *    these aggregations will probably consume SGE immediately)
  * 2. Rest of BRB blocks divided by 2 (block will consume new SGE only
  *    after placement on BD for new TPA aggregation)
  *
@@ -387,7 +383,6 @@ union db_prod {
 #define BIT_VEC64_ELEM_SHIFT		6
 #define BIT_VEC64_ELEM_MASK		((u64)BIT_VEC64_ELEM_SZ - 1)
 
-
 #define __BIT_VEC64_SET_BIT(el, bit) \
 	do { \
 		el = ((el) | ((u64)0x1 << (bit))); \
@@ -398,7 +393,6 @@ union db_prod {
 		el = ((el) & (~((u64)0x1 << (bit)))); \
 	} while (0)
 
-
 #define BIT_VEC64_SET_BIT(vec64, idx) \
 	__BIT_VEC64_SET_BIT((vec64)[(idx) >> BIT_VEC64_ELEM_SHIFT], \
 			   (idx) & BIT_VEC64_ELEM_MASK)
@@ -419,8 +413,6 @@ union db_prod {
 
 /*******************************************************/
 
-
-
 /* Number of u64 elements in SGE mask array */
 #define RX_SGE_MASK_LEN			(NUM_RX_SGE / BIT_VEC64_ELEM_SZ)
 #define RX_SGE_MASK_LEN_MASK		(RX_SGE_MASK_LEN - 1)
@@ -493,11 +485,26 @@ struct bnx2x_fastpath {
 	struct bnx2x		*bp; /* parent */
 
 	struct napi_struct	napi;
+
+#ifdef CONFIG_NET_LL_RX_POLL
+	unsigned int state;
+#define BNX2X_FP_STATE_IDLE		      0
+#define BNX2X_FP_STATE_NAPI		(1 << 0)    /* NAPI owns this FP */
+#define BNX2X_FP_STATE_POLL		(1 << 1)    /* poll owns this FP */
+#define BNX2X_FP_STATE_NAPI_YIELD	(1 << 2)    /* NAPI yielded this FP */
+#define BNX2X_FP_STATE_POLL_YIELD	(1 << 3)    /* poll yielded this FP */
+#define BNX2X_FP_YIELD	(BNX2X_FP_STATE_NAPI_YIELD | BNX2X_FP_STATE_POLL_YIELD)
+#define BNX2X_FP_LOCKED	(BNX2X_FP_STATE_NAPI | BNX2X_FP_STATE_POLL)
+#define BNX2X_FP_USER_PEND (BNX2X_FP_STATE_POLL | BNX2X_FP_STATE_POLL_YIELD)
+	/* protect state */
+	spinlock_t lock;
+#endif /* CONFIG_NET_LL_RX_POLL */
+
 	union host_hc_status_block	status_blk;
-	/* chip independed shortcuts into sb structure */
+	/* chip independent shortcuts into sb structure */
 	__le16			*sb_index_values;
 	__le16			*sb_running_index;
-	/* chip independed shortcut into rx_prods_offset memory */
+	/* chip independent shortcut into rx_prods_offset memory */
 	u32			ustorm_rx_prods_offset;
 
 	u32			rx_buf_size;
@@ -565,6 +572,116 @@ struct bnx2x_fastpath {
 #define bnx2x_fp_stats(bp, fp)	(&((bp)->fp_stats[(fp)->index]))
 #define bnx2x_fp_qstats(bp, fp)	(&((bp)->fp_stats[(fp)->index].eth_q_stats))
 
+#ifdef CONFIG_NET_LL_RX_POLL
+static inline void bnx2x_fp_init_lock(struct bnx2x_fastpath *fp)
+{
+	spin_lock_init(&fp->lock);
+	fp->state = BNX2X_FP_STATE_IDLE;
+}
+
+/* called from the device poll routine to get ownership of a FP */
+static inline bool bnx2x_fp_lock_napi(struct bnx2x_fastpath *fp)
+{
+	bool rc = true;
+
+	spin_lock(&fp->lock);
+	if (fp->state & BNX2X_FP_LOCKED) {
+		WARN_ON(fp->state & BNX2X_FP_STATE_NAPI);
+		fp->state |= BNX2X_FP_STATE_NAPI_YIELD;
+		rc = false;
+	} else {
+		/* we don't care if someone yielded */
+		fp->state = BNX2X_FP_STATE_NAPI;
+	}
+	spin_unlock(&fp->lock);
+	return rc;
+}
+
+/* returns true is someone tried to get the FP while napi had it */
+static inline bool bnx2x_fp_unlock_napi(struct bnx2x_fastpath *fp)
+{
+	bool rc = false;
+
+	spin_lock(&fp->lock);
+	WARN_ON(fp->state &
+		(BNX2X_FP_STATE_POLL | BNX2X_FP_STATE_NAPI_YIELD));
+
+	if (fp->state & BNX2X_FP_STATE_POLL_YIELD)
+		rc = true;
+	fp->state = BNX2X_FP_STATE_IDLE;
+	spin_unlock(&fp->lock);
+	return rc;
+}
+
+/* called from bnx2x_low_latency_poll() */
+static inline bool bnx2x_fp_lock_poll(struct bnx2x_fastpath *fp)
+{
+	bool rc = true;
+
+	spin_lock_bh(&fp->lock);
+	if ((fp->state & BNX2X_FP_LOCKED)) {
+		fp->state |= BNX2X_FP_STATE_POLL_YIELD;
+		rc = false;
+	} else {
+		/* preserve yield marks */
+		fp->state |= BNX2X_FP_STATE_POLL;
+	}
+	spin_unlock_bh(&fp->lock);
+	return rc;
+}
+
+/* returns true if someone tried to get the FP while it was locked */
+static inline bool bnx2x_fp_unlock_poll(struct bnx2x_fastpath *fp)
+{
+	bool rc = false;
+
+	spin_lock_bh(&fp->lock);
+	WARN_ON(fp->state & BNX2X_FP_STATE_NAPI);
+
+	if (fp->state & BNX2X_FP_STATE_POLL_YIELD)
+		rc = true;
+	fp->state = BNX2X_FP_STATE_IDLE;
+	spin_unlock_bh(&fp->lock);
+	return rc;
+}
+
+/* true if a socket is polling, even if it did not get the lock */
+static inline bool bnx2x_fp_ll_polling(struct bnx2x_fastpath *fp)
+{
+	WARN_ON(!(fp->state & BNX2X_FP_LOCKED));
+	return fp->state & BNX2X_FP_USER_PEND;
+}
+#else
+static inline void bnx2x_fp_init_lock(struct bnx2x_fastpath *fp)
+{
+}
+
+static inline bool bnx2x_fp_lock_napi(struct bnx2x_fastpath *fp)
+{
+	return true;
+}
+
+static inline bool bnx2x_fp_unlock_napi(struct bnx2x_fastpath *fp)
+{
+	return false;
+}
+
+static inline bool bnx2x_fp_lock_poll(struct bnx2x_fastpath *fp)
+{
+	return false;
+}
+
+static inline bool bnx2x_fp_unlock_poll(struct bnx2x_fastpath *fp)
+{
+	return false;
+}
+
+static inline bool bnx2x_fp_ll_polling(struct bnx2x_fastpath *fp)
+{
+	return false;
+}
+#endif /* CONFIG_NET_LL_RX_POLL */
+
 /* Use 2500 as a mini-jumbo MTU for FCoE */
 #define BNX2X_FCOE_MINI_JUMBO_MTU	2500
 
@@ -580,12 +697,10 @@ struct bnx2x_fastpath {
 						txdata_ptr[FIRST_TX_COS_INDEX] \
 						->var)
 
-
 #define IS_ETH_FP(fp)		((fp)->index < BNX2X_NUM_ETH_QUEUES((fp)->bp))
 #define IS_FCOE_FP(fp)		((fp)->index == FCOE_IDX((fp)->bp))
 #define IS_FCOE_IDX(idx)	((idx) == FCOE_IDX(bp))
 
-
 /* MC hsi */
 #define MAX_FETCH_BD		13	/* HW max BDs per packet */
 #define RX_COPY_THRESH		92
@@ -613,7 +728,7 @@ struct bnx2x_fastpath {
  * START_BD(splitted)	- includes unpaged data segment for GSO
  * PARSING_BD		- for TSO and CSUM data
  * PARSING_BD2		- for encapsulation data
- * Frag BDs		- decribes pages for frags
+ * Frag BDs		- describes pages for frags
  */
 #define BDS_PER_TX_PKT		4
 #define MAX_BDS_PER_TX_PKT	(MAX_SKB_FRAGS + BDS_PER_TX_PKT)
@@ -693,12 +808,10 @@ struct bnx2x_fastpath {
 				 FW_DROP_LEVEL(bp))
 #define RCQ_TH_HI(bp)		(RCQ_TH_LO(bp) + DROPLESS_FC_HEADROOM)
 
-
 /* This is needed for determining of last_max */
 #define SUB_S16(a, b)		(s16)((s16)(a) - (s16)(b))
 #define SUB_S32(a, b)		(s32)((s32)(a) - (s32)(b))
 
-
 #define BNX2X_SWCID_SHIFT	17
 #define BNX2X_SWCID_MASK	((0x1 << BNX2X_SWCID_SHIFT) - 1)
 
@@ -723,7 +836,6 @@ struct bnx2x_fastpath {
 		       DPM_TRIGER_TYPE); \
 	} while (0)
 
-
 /* TX CSUM helpers */
 #define SKB_CS_OFF(skb)		(offsetof(struct tcphdr, check) - \
 				 skb->csum_offset)
@@ -766,7 +878,6 @@ struct bnx2x_fastpath {
 #define BNX2X_RX_SUM_FIX(cqe) \
 	BNX2X_PRS_FLAG_OVERETH_IPV4(cqe->fast_path_cqe.pars_flags.flags)
 
-
 #define FP_USB_FUNC_OFF	\
 			offsetof(struct cstorm_status_block_u, func)
 #define FP_CSB_FUNC_OFF	\
@@ -900,14 +1011,14 @@ struct bnx2x_common {
 #define CHIP_IS_E3A0(bp)		(CHIP_IS_E3(bp) && \
 					 (CHIP_REV(bp) == CHIP_REV_Ax))
 /* This define is used in two main places:
- * 1. In the early stages of nic_load, to know if to configrue Parser / Searcher
+ * 1. In the early stages of nic_load, to know if to configure Parser / Searcher
  * to nic-only mode or to offload mode. Offload mode is configured if either the
  * chip is E1x (where MIC_MODE register is not applicable), or if cnic already
  * registered for this port (which means that the user wants storage services).
  * 2. During cnic-related load, to know if offload mode is already configured in
- * the HW or needs to be configrued.
+ * the HW or needs to be configured.
  * Since the transition from nic-mode to offload-mode in HW causes traffic
- * coruption, nic-mode is configured only in ports on which storage services
+ * corruption, nic-mode is configured only in ports on which storage services
  * where never requested.
  */
 #define CONFIGURE_NIC_MODE(bp)		(!CHIP_IS_E1x(bp) && !CNIC_ENABLED(bp))
@@ -1008,14 +1119,14 @@ extern struct workqueue_struct *bnx2x_wq;
  * If the maximum number of FP-SB available is X then:
  * a. If CNIC is supported it consumes 1 FP-SB thus the max number of
  *    regular L2 queues is Y=X-1
- * b. in MF mode the actual number of L2 queues is Y= (X-1/MF_factor)
+ * b. In MF mode the actual number of L2 queues is Y= (X-1/MF_factor)
  * c. If the FCoE L2 queue is supported the actual number of L2 queues
  *    is Y+1
  * d. The number of irqs (MSIX vectors) is either Y+1 (one extra for
  *    slow-path interrupts) or Y+2 if CNIC is supported (one additional
  *    FP interrupt context for the CNIC).
  * e. The number of HW context (CID count) is always X or X+1 if FCoE
- *    L2 queue is supported. the cid for the FCoE L2 queue is always X.
+ *    L2 queue is supported. The cid for the FCoE L2 queue is always X.
  */
 
 /* fast-path interrupt contexts E1x */
@@ -1068,7 +1179,6 @@ struct bnx2x_slowpath {
 		struct eth_classify_rules_ramrod_data	e2;
 	} mac_rdata;
 
-
 	union {
 		struct tstorm_eth_mac_filter_config	e1x;
 		struct eth_filter_rules_ramrod_data	e2;
@@ -1119,7 +1229,6 @@ struct bnx2x_slowpath {
 #define bnx2x_sp_mapping(bp, var) \
 		(bp->slowpath_mapping + offsetof(struct bnx2x_slowpath, var))
 
-
 /* attn group wiring */
 #define MAX_DYNAMIC_ATTN_GRPS		8
 
@@ -1221,11 +1330,11 @@ enum {
 	BNX2X_SP_RTNL_AFEX_F_UPDATE,
 	BNX2X_SP_RTNL_ENABLE_SRIOV,
 	BNX2X_SP_RTNL_VFPF_MCAST,
+	BNX2X_SP_RTNL_VFPF_CHANNEL_DOWN,
 	BNX2X_SP_RTNL_VFPF_STORM_RX_MODE,
 	BNX2X_SP_RTNL_HYPERVISOR_VLAN,
 };
 
-
 struct bnx2x_prev_path_list {
 	struct list_head list;
 	u8 bus;
@@ -1392,6 +1501,7 @@ struct bnx2x {
 #define USING_SINGLE_MSIX_FLAG		(1 << 20)
 #define BC_SUPPORTS_DCBX_MSG_NON_PMF	(1 << 21)
 #define IS_VF_FLAG			(1 << 22)
+#define INTERRUPTS_ENABLED_FLAG		(1 << 23)
 
 #define BP_NOMCP(bp)			((bp)->flags & NO_MCP_FLAG)
 
@@ -1585,7 +1695,7 @@ struct bnx2x {
 	struct mutex		cnic_mutex;
 	struct bnx2x_vlan_mac_obj iscsi_l2_mac_obj;
 
-	/* Start index of the "special" (CNIC related) L2 cleints */
+	/* Start index of the "special" (CNIC related) L2 clients */
 	u8				cnic_base_cl_id;
 
 	int			dmae_ready;
@@ -1699,7 +1809,7 @@ struct bnx2x {
 	/* operation indication for the sp_rtnl task */
 	unsigned long				sp_rtnl_state;
 
-	/* DCBX Negotation results */
+	/* DCBX Negotiation results */
 	struct dcbx_features			dcbx_local_feat;
 	u32					dcbx_error;
 
@@ -1755,7 +1865,6 @@ extern int num_queues;
 #define FUNC_FLG_SPQ		0x0010
 #define FUNC_FLG_LEADING	0x0020	/* PF only */
 
-
 struct bnx2x_func_init_params {
 	/* dma */
 	dma_addr_t	fw_stat_map;	/* valid iff FUNC_FLG_STATS */
@@ -1853,9 +1962,6 @@ struct bnx2x_func_init_params {
 
 #define skip_queue(bp, idx)	(NO_FCOE(bp) && IS_FCOE_IDX(idx))
 
-
-
-
 /**
  * bnx2x_set_mac_one - configure a single MAC address
  *
@@ -1921,7 +2027,6 @@ u32 bnx2x_dmae_opcode(struct bnx2x *bp, u8 src_type, u8 dst_type,
 void bnx2x_prep_dmae_with_comp(struct bnx2x *bp, struct dmae_command *dmae,
 			       u8 src_type, u8 dst_type);
 int bnx2x_issue_dmae_with_comp(struct bnx2x *bp, struct dmae_command *dmae);
-void bnx2x_dp_dmae(struct bnx2x *bp, struct dmae_command *dmae, int msglvl);
 
 /* FLR related routines */
 u32 bnx2x_flr_clnup_poll_count(struct bnx2x *bp);
@@ -1937,6 +2042,8 @@ int bnx2x_sp_post(struct bnx2x *bp, int command, int cid,
 void bnx2x_update_coalesce(struct bnx2x *bp);
 int bnx2x_get_cur_phy_idx(struct bnx2x *bp);
 
+bool bnx2x_port_after_undi(struct bnx2x *bp);
+
 static inline u32 reg_poll(struct bnx2x *bp, u32 reg, u32 expected, int ms,
 			   int wait)
 {
@@ -1998,7 +2105,6 @@ void bnx2x_igu_clear_sb_gen(struct bnx2x *bp, u8 func, u8 idu_sb_id,
 #define UNLOAD_CLOSE			1
 #define UNLOAD_RECOVERY			2
 
-
 /* DMAE command defines */
 #define DMAE_TIMEOUT			-1
 #define DMAE_PCI_ERROR			-2	/* E2 and onward */
@@ -2062,7 +2168,8 @@ void bnx2x_igu_clear_sb_gen(struct bnx2x *bp, u8 func, u8 idu_sb_id,
 #define DMAE_LEN32_WR_MAX(bp)		(CHIP_IS_E1(bp) ? 0x400 : 0x2000)
 
 #define DMAE_COMP_VAL			0x60d0d0ae /* E2 and on - upper bit
-							indicates eror */
+						    * indicates error
+						    */
 
 #define MAX_DMAE_C_PER_PORT		8
 #define INIT_DMAE_C(bp)			(BP_PORT(bp) * MAX_DMAE_C_PER_PORT + \
@@ -2100,7 +2207,6 @@ void bnx2x_igu_clear_sb_gen(struct bnx2x *bp, u8 func, u8 idu_sb_id,
 #define SP_DESC_CNT		(BCM_PAGE_SIZE / sizeof(struct eth_spe))
 #define MAX_SP_DESC_CNT			(SP_DESC_CNT - 1)
 
-
 #define BNX2X_BTR			4
 #define MAX_SPQ_PENDING			8
 
@@ -2137,6 +2243,8 @@ void bnx2x_igu_clear_sb_gen(struct bnx2x *bp, u8 func, u8 idu_sb_id,
 #define ATTN_HARD_WIRED_MASK		0xff00
 #define ATTENTION_ID			4
 
+#define IS_MF_STORAGE_ONLY(bp) (IS_MF_STORAGE_SD(bp) || \
+				 IS_MF_FCOE_AFEX(bp))
 
 /* stuff added to make the code fit 80Col */
 
@@ -2338,4 +2446,9 @@ enum {
 
 #define NUM_MACS	8
 
+enum bnx2x_pci_bus_speed {
+	BNX2X_PCI_LINK_SPEED_2500 = 2500,
+	BNX2X_PCI_LINK_SPEED_5000 = 5000,
+	BNX2X_PCI_LINK_SPEED_8000 = 8000
+};
 #endif /* bnx2x.h */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index 638e554..ec3aa1d 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -24,6 +24,7 @@
 #include <net/tcp.h>
 #include <net/ipv6.h>
 #include <net/ip6_checksum.h>
+#include <net/ll_poll.h>
 #include <linux/prefetch.h>
 #include "bnx2x_cmn.h"
 #include "bnx2x_init.h"
@@ -124,7 +125,7 @@ static void bnx2x_shrink_eth_fp(struct bnx2x *bp, int delta)
 	int i, cos, old_eth_num = BNX2X_NUM_ETH_QUEUES(bp);
 
 	/* Queue pointer cannot be re-set on an fp-basis, as moving pointer
-	 * backward along the array could cause memory to be overriden
+	 * backward along the array could cause memory to be overridden
 	 */
 	for (cos = 1; cos < bp->max_cos; cos++) {
 		for (i = 0; i < old_eth_num - delta; i++) {
@@ -165,7 +166,6 @@ static u16 bnx2x_free_tx_pkt(struct bnx2x *bp, struct bnx2x_fp_txdata *txdata,
 	dma_unmap_single(&bp->pdev->dev, BD_UNMAP_ADDR(tx_start_bd),
 			 BD_UNMAP_LEN(tx_start_bd), DMA_TO_DEVICE);
 
-
 	nbd = le16_to_cpu(tx_start_bd->nbd) - 1;
 #ifdef BNX2X_STOP_ON_ERROR
 	if ((nbd - 1) > (MAX_SKB_FRAGS + 2)) {
@@ -259,7 +259,7 @@ int bnx2x_tx_int(struct bnx2x *bp, struct bnx2x_fp_txdata *txdata)
 	smp_mb();
 
 	if (unlikely(netif_tx_queue_stopped(txq))) {
-		/* Taking tx_lock() is needed to prevent reenabling the queue
+		/* Taking tx_lock() is needed to prevent re-enabling the queue
 		 * while it's empty. This could have happen if rx_action() gets
 		 * suspended in bnx2x_tx_int() after the condition before
 		 * netif_tx_wake_queue(), while tx_action (bnx2x_start_xmit()):
@@ -572,7 +572,7 @@ static int bnx2x_fill_frag_skb(struct bnx2x *bp, struct bnx2x_fastpath *fp,
 			return err;
 		}
 
-		/* Unmap the page as we r going to pass it to the stack */
+		/* Unmap the page as we're going to pass it to the stack */
 		dma_unmap_page(&bp->pdev->dev,
 			       dma_unmap_addr(&old_rx_pg, mapping),
 			       SGE_PAGES, DMA_FROM_DEVICE);
@@ -733,7 +733,6 @@ static void bnx2x_tpa_stop(struct bnx2x *bp, struct bnx2x_fastpath *fp,
 			dev_kfree_skb_any(skb);
 		}
 
-
 		/* put new data in bin */
 		rx_buf->data = new_data;
 
@@ -805,40 +804,32 @@ int bnx2x_rx_int(struct bnx2x_fastpath *fp, int budget)
 {
 	struct bnx2x *bp = fp->bp;
 	u16 bd_cons, bd_prod, bd_prod_fw, comp_ring_cons;
-	u16 hw_comp_cons, sw_comp_cons, sw_comp_prod;
+	u16 sw_comp_cons, sw_comp_prod;
 	int rx_pkt = 0;
+	union eth_rx_cqe *cqe;
+	struct eth_fast_path_rx_cqe *cqe_fp;
 
 #ifdef BNX2X_STOP_ON_ERROR
 	if (unlikely(bp->panic))
 		return 0;
 #endif
 
-	/* CQ "next element" is of the size of the regular element,
-	   that's why it's ok here */
-	hw_comp_cons = le16_to_cpu(*fp->rx_cons_sb);
-	if ((hw_comp_cons & MAX_RCQ_DESC_CNT) == MAX_RCQ_DESC_CNT)
-		hw_comp_cons++;
-
 	bd_cons = fp->rx_bd_cons;
 	bd_prod = fp->rx_bd_prod;
 	bd_prod_fw = bd_prod;
 	sw_comp_cons = fp->rx_comp_cons;
 	sw_comp_prod = fp->rx_comp_prod;
 
-	/* Memory barrier necessary as speculative reads of the rx
-	 * buffer can be ahead of the index in the status block
-	 */
-	rmb();
+	comp_ring_cons = RCQ_BD(sw_comp_cons);
+	cqe = &fp->rx_comp_ring[comp_ring_cons];
+	cqe_fp = &cqe->fast_path_cqe;
 
 	DP(NETIF_MSG_RX_STATUS,
-	   "queue[%d]:  hw_comp_cons %u  sw_comp_cons %u\n",
-	   fp->index, hw_comp_cons, sw_comp_cons);
+	   "queue[%d]: sw_comp_cons %u\n", fp->index, sw_comp_cons);
 
-	while (sw_comp_cons != hw_comp_cons) {
+	while (BNX2X_IS_CQE_COMPLETED(cqe_fp)) {
 		struct sw_rx_bd *rx_buf = NULL;
 		struct sk_buff *skb;
-		union eth_rx_cqe *cqe;
-		struct eth_fast_path_rx_cqe *cqe_fp;
 		u8 cqe_fp_flags;
 		enum eth_rx_cqe_type cqe_fp_type;
 		u16 len, pad, queue;
@@ -850,12 +841,9 @@ int bnx2x_rx_int(struct bnx2x_fastpath *fp, int budget)
 			return 0;
 #endif
 
-		comp_ring_cons = RCQ_BD(sw_comp_cons);
 		bd_prod = RX_BD(bd_prod);
 		bd_cons = RX_BD(bd_cons);
 
-		cqe = &fp->rx_comp_ring[comp_ring_cons];
-		cqe_fp = &cqe->fast_path_cqe;
 		cqe_fp_flags = cqe_fp->type_error_flags;
 		cqe_fp_type = cqe_fp_flags & ETH_FAST_PATH_RX_CQE_TYPE;
 
@@ -899,7 +887,6 @@ int bnx2x_rx_int(struct bnx2x_fastpath *fp, int budget)
 						cqe_fp);
 
 				goto next_rx;
-
 			}
 			queue = cqe->end_agg_cqe.queue_index;
 			tpa_info = &fp->tpa_info[queue];
@@ -1002,9 +989,13 @@ reuse_rx:
 		    PARSING_FLAGS_VLAN)
 			__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
 					       le16_to_cpu(cqe_fp->vlan_tag));
-		napi_gro_receive(&fp->napi, skb);
 
+		skb_mark_ll(skb, &fp->napi);
 
+		if (bnx2x_fp_ll_polling(fp))
+			netif_receive_skb(skb);
+		else
+			napi_gro_receive(&fp->napi, skb);
 next_rx:
 		rx_buf->data = NULL;
 
@@ -1016,8 +1007,15 @@ next_cqe:
 		sw_comp_prod = NEXT_RCQ_IDX(sw_comp_prod);
 		sw_comp_cons = NEXT_RCQ_IDX(sw_comp_cons);
 
+		/* mark CQE as free */
+		BNX2X_SEED_CQE(cqe_fp);
+
 		if (rx_pkt == budget)
 			break;
+
+		comp_ring_cons = RCQ_BD(sw_comp_cons);
+		cqe = &fp->rx_comp_ring[comp_ring_cons];
+		cqe_fp = &cqe->fast_path_cqe;
 	} /* while */
 
 	fp->rx_bd_cons = bd_cons;
@@ -1053,8 +1051,6 @@ static irqreturn_t bnx2x_msix_fp_int(int irq, void *fp_cookie)
 #endif
 
 	/* Handle Rx and Tx according to MSI-X vector */
-	prefetch(fp->rx_cons_sb);
-
 	for_each_cos_in_tx_queue(fp, cos)
 		prefetch(fp->txdata_ptr[cos]->tx_cons_sb);
 
@@ -1118,7 +1114,7 @@ static void bnx2x_fill_report_data(struct bnx2x *bp,
 
 	memset(data, 0, sizeof(*data));
 
-	/* Fill the report data: efective line speed */
+	/* Fill the report data: effective line speed */
 	data->line_speed = line_speed;
 
 	/* Link is down */
@@ -1161,7 +1157,7 @@ void bnx2x_link_report(struct bnx2x *bp)
  *
  * @bp:		driver handle
  *
- * None atomic inmlementation.
+ * None atomic implementation.
  * Should be called under the phy_lock.
  */
 void __bnx2x_link_report(struct bnx2x *bp)
@@ -1304,7 +1300,7 @@ void bnx2x_init_rx_rings(struct bnx2x *bp)
 		   "mtu %d  rx_buf_size %d\n", bp->dev->mtu, fp->rx_buf_size);
 
 		if (!fp->disable_tpa) {
-			/* Fill the per-aggregtion pool */
+			/* Fill the per-aggregation pool */
 			for (i = 0; i < MAX_AGG_QS(bp); i++) {
 				struct bnx2x_agg_info *tpa_info =
 					&fp->tpa_info[i];
@@ -1726,7 +1722,7 @@ static int bnx2x_req_irq(struct bnx2x *bp)
 	return request_irq(irq, bnx2x_interrupt, flags, bp->dev->name, bp->dev);
 }
 
-int bnx2x_setup_irqs(struct bnx2x *bp)
+static int bnx2x_setup_irqs(struct bnx2x *bp)
 {
 	int rc = 0;
 	if (bp->flags & USING_MSIX_FLAG &&
@@ -1759,32 +1755,46 @@ static void bnx2x_napi_enable_cnic(struct bnx2x *bp)
 {
 	int i;
 
-	for_each_rx_queue_cnic(bp, i)
+	for_each_rx_queue_cnic(bp, i) {
+		bnx2x_fp_init_lock(&bp->fp[i]);
 		napi_enable(&bnx2x_fp(bp, i, napi));
+	}
 }
 
 static void bnx2x_napi_enable(struct bnx2x *bp)
 {
 	int i;
 
-	for_each_eth_queue(bp, i)
+	for_each_eth_queue(bp, i) {
+		bnx2x_fp_init_lock(&bp->fp[i]);
 		napi_enable(&bnx2x_fp(bp, i, napi));
+	}
 }
 
 static void bnx2x_napi_disable_cnic(struct bnx2x *bp)
 {
 	int i;
 
-	for_each_rx_queue_cnic(bp, i)
+	local_bh_disable();
+	for_each_rx_queue_cnic(bp, i) {
 		napi_disable(&bnx2x_fp(bp, i, napi));
+		while (!bnx2x_fp_lock_napi(&bp->fp[i]))
+			mdelay(1);
+	}
+	local_bh_enable();
 }
 
 static void bnx2x_napi_disable(struct bnx2x *bp)
 {
 	int i;
 
-	for_each_eth_queue(bp, i)
+	local_bh_disable();
+	for_each_eth_queue(bp, i) {
 		napi_disable(&bnx2x_fp(bp, i, napi));
+		while (!bnx2x_fp_lock_napi(&bp->fp[i]))
+			mdelay(1);
+	}
+	local_bh_enable();
 }
 
 void bnx2x_netif_start(struct bnx2x *bp)
@@ -1829,7 +1839,7 @@ u16 bnx2x_select_queue(struct net_device *dev, struct sk_buff *skb)
 	}
 
 	/* select a non-FCoE queue */
-	return __skb_tx_hash(dev, skb, BNX2X_NUM_ETH_QUEUES(bp));
+	return __netdev_pick_tx(dev, skb) % BNX2X_NUM_ETH_QUEUES(bp);
 }
 
 void bnx2x_set_num_queues(struct bnx2x *bp)
@@ -1862,7 +1872,7 @@ void bnx2x_set_num_queues(struct bnx2x *bp)
  *
  * If the actual number of Tx queues (for each CoS) is less than 16 then there
  * will be the holes at the end of each group of 16 ETh L2 indices (0..15,
- * 16..31,...) with indicies that are not coupled with any real Tx queue.
+ * 16..31,...) with indices that are not coupled with any real Tx queue.
  *
  * The proper configuration of skb->queue_mapping is handled by
  * bnx2x_select_queue() and __skb_tx_hash().
@@ -1924,7 +1934,7 @@ static void bnx2x_set_rx_buf_size(struct bnx2x *bp)
 				  ETH_OVREHEAD +
 				  mtu +
 				  BNX2X_FW_RX_ALIGN_END;
-		/* Note : rx_buf_size doesnt take into account NET_SKB_PAD */
+		/* Note : rx_buf_size doesn't take into account NET_SKB_PAD */
 		if (fp->rx_buf_size + NET_SKB_PAD <= PAGE_SIZE)
 			fp->rx_frag_size = fp->rx_buf_size + NET_SKB_PAD;
 		else
@@ -1937,7 +1947,7 @@ static int bnx2x_init_rss_pf(struct bnx2x *bp)
 	int i;
 	u8 num_eth_queues = BNX2X_NUM_ETH_QUEUES(bp);
 
-	/* Prepare the initial contents fo the indirection table if RSS is
+	/* Prepare the initial contents for the indirection table if RSS is
 	 * enabled
 	 */
 	for (i = 0; i < sizeof(bp->rss_conf_obj.ind_table); i++)
@@ -2015,7 +2025,7 @@ static int bnx2x_init_hw(struct bnx2x *bp, u32 load_code)
 
 /*
  * Cleans the object that have internal lists without sending
- * ramrods. Should be run when interrutps are disabled.
+ * ramrods. Should be run when interrupts are disabled.
  */
 void bnx2x_squeeze_objects(struct bnx2x *bp)
 {
@@ -2166,10 +2176,10 @@ static int bnx2x_alloc_fw_stats_mem(struct bnx2x *bp)
 	bp->fw_stats_data_mapping = bp->fw_stats_mapping +
 		bp->fw_stats_req_sz;
 
-	DP(BNX2X_MSG_SP, "statistics request base address set to %x %x",
+	DP(BNX2X_MSG_SP, "statistics request base address set to %x %x\n",
 	   U64_HI(bp->fw_stats_req_mapping),
 	   U64_LO(bp->fw_stats_req_mapping));
-	DP(BNX2X_MSG_SP, "statistics data base address set to %x %x",
+	DP(BNX2X_MSG_SP, "statistics data base address set to %x %x\n",
 	   U64_HI(bp->fw_stats_data_mapping),
 	   U64_LO(bp->fw_stats_data_mapping));
 	return 0;
@@ -2183,6 +2193,8 @@ alloc_mem_err:
 /* send load request to mcp and analyze response */
 static int bnx2x_nic_load_request(struct bnx2x *bp, u32 *load_code)
 {
+	u32 param;
+
 	/* init fw_seq */
 	bp->fw_seq =
 		(SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
@@ -2195,9 +2207,13 @@ static int bnx2x_nic_load_request(struct bnx2x *bp, u32 *load_code)
 		 DRV_PULSE_SEQ_MASK);
 	BNX2X_DEV_INFO("drv_pulse 0x%x\n", bp->fw_drv_pulse_wr_seq);
 
+	param = DRV_MSG_CODE_LOAD_REQ_WITH_LFA;
+
+	if (IS_MF_SD(bp) && bnx2x_port_after_undi(bp))
+		param |= DRV_MSG_CODE_LOAD_REQ_FORCE_LFA;
+
 	/* load request */
-	(*load_code) = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_REQ,
-					DRV_MSG_CODE_LOAD_REQ_WITH_LFA);
+	(*load_code) = bnx2x_fw_command(bp, DRV_MSG_CODE_LOAD_REQ, param);
 
 	/* if mcp fails to respond we must abort */
 	if (!(*load_code)) {
@@ -2238,7 +2254,7 @@ int bnx2x_nic_load_analyze_req(struct bnx2x *bp, u32 load_code)
 
 		/* abort nic load if version mismatch */
 		if (my_fw != loaded_fw) {
-			BNX2X_ERR("bnx2x with FW %x was already loaded which mismatches my %x FW. aborting\n",
+			BNX2X_ERR("bnx2x with FW %x was already loaded which mismatches my %x FW. Aborting\n",
 				  loaded_fw, my_fw);
 			return -EBUSY;
 		}
@@ -2316,10 +2332,10 @@ static void bnx2x_nic_load_afex_dcc(struct bnx2x *bp, int load_code)
 static void bnx2x_bz_fp(struct bnx2x *bp, int index)
 {
 	struct bnx2x_fastpath *fp = &bp->fp[index];
-
 	int cos;
 	struct napi_struct orig_napi = fp->napi;
 	struct bnx2x_agg_info *orig_tpa_info = fp->tpa_info;
+
 	/* bzero bnx2x_fastpath contents */
 	if (fp->tpa_info)
 		memset(fp->tpa_info, 0, ETH_MAX_AGGREGATION_QUEUES_E1H_E2 *
@@ -2345,8 +2361,7 @@ static void bnx2x_bz_fp(struct bnx2x *bp, int index)
 			fp->txdata_ptr[cos] = &bp->bnx2x_txq[cos *
 				BNX2X_NUM_ETH_QUEUES(bp) + index];
 
-	/*
-	 * set the tpa flag for each queue. The tpa flag determines the queue
+	/* set the tpa flag for each queue. The tpa flag determines the queue
 	 * minimal size so it must be set prior to queue memory allocation
 	 */
 	fp->disable_tpa = !(bp->flags & TPA_ENABLE_FLAG ||
@@ -2429,7 +2444,6 @@ int bnx2x_load_cnic(struct bnx2x *bp)
 	if (bp->state == BNX2X_STATE_OPEN)
 		bnx2x_cnic_notify(bp, CNIC_CTL_START_CMD);
 
-
 	DP(NETIF_MSG_IFUP, "Ending successfully CNIC-related load\n");
 
 	return 0;
@@ -2472,6 +2486,7 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
 
 	bp->state = BNX2X_STATE_OPENING_WAIT4_LOAD;
 
+	/* zero the structure w/o any lock, before SP handler is initialized */
 	memset(&bp->last_reported_link, 0, sizeof(bp->last_reported_link));
 	__set_bit(BNX2X_LINK_REPORT_LINK_DOWN,
 		&bp->last_reported_link.link_report_flags);
@@ -2536,8 +2551,8 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
 	}
 
 	/* configure multi cos mappings in kernel.
-	 * this configuration may be overriden by a multi class queue discipline
-	 * or by a dcbx negotiation result.
+	 * this configuration may be overridden by a multi class queue
+	 * discipline or by a dcbx negotiation result.
 	 */
 	bnx2x_setup_tc(bp->dev, bp->max_cos);
 
@@ -2696,7 +2711,7 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode)
 	/* Start the Tx */
 	switch (load_mode) {
 	case LOAD_NORMAL:
-		/* Tx queue should be only reenabled */
+		/* Tx queue should be only re-enabled */
 		netif_tx_wake_all_queues(bp->dev);
 		break;
 
@@ -2841,7 +2856,7 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
 	}
 
 	/* Nothing to do during unload if previous bnx2x_nic_load()
-	 * have not completed succesfully - all resourses are released.
+	 * have not completed successfully - all resources are released.
 	 *
 	 * we can get here only after unsuccessful ndo_* callback, during which
 	 * dev->IFF_UP flag is still on.
@@ -2856,6 +2871,9 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
 	bp->state = BNX2X_STATE_CLOSING_WAIT4_HALT;
 	smp_mb();
 
+	/* indicate to VFs that the PF is going down */
+	bnx2x_iov_channel_down(bp);
+
 	if (CNIC_LOADED(bp))
 		bnx2x_cnic_notify(bp, CNIC_CTL_STOP_CMD);
 
@@ -2890,10 +2908,9 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
 		/* Send the UNLOAD_REQUEST to the MCP */
 		bnx2x_send_unload_req(bp, unload_mode);
 
-		/*
-		 * Prevent transactions to host from the functions on the
+		/* Prevent transactions to host from the functions on the
 		 * engine that doesn't reset global blocks in case of global
-		 * attention once gloabl blocks are reset and gates are opened
+		 * attention once global blocks are reset and gates are opened
 		 * (the engine which leader will perform the recovery
 		 * last).
 		 */
@@ -2914,7 +2931,7 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
 	}
 
 	/*
-	 * At this stage no more interrupts will arrive so we may safly clean
+	 * At this stage no more interrupts will arrive so we may safely clean
 	 * the queueable objects here in case they failed to get cleaned so far.
 	 */
 	if (IS_PF(bp))
@@ -2955,7 +2972,6 @@ int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link)
 			bnx2x_set_reset_global(bp);
 	}
 
-
 	/* The last driver must disable a "close the gate" if there is no
 	 * parity attention or "process kill" pending.
 	 */
@@ -3040,6 +3056,8 @@ int bnx2x_poll(struct napi_struct *napi, int budget)
 			return 0;
 		}
 #endif
+		if (!bnx2x_fp_lock_napi(fp))
+			return work_done;
 
 		for_each_cos_in_tx_queue(fp, cos)
 			if (bnx2x_tx_queue_has_work(fp->txdata_ptr[cos]))
@@ -3049,12 +3067,15 @@ int bnx2x_poll(struct napi_struct *napi, int budget)
 			work_done += bnx2x_rx_int(fp, budget - work_done);
 
 			/* must not complete if we consumed full budget */
-			if (work_done >= budget)
+			if (work_done >= budget) {
+				bnx2x_fp_unlock_napi(fp);
 				break;
+			}
 		}
 
 		/* Fall out from the NAPI loop if needed */
-		if (!(bnx2x_has_rx_work(fp) || bnx2x_has_tx_work(fp))) {
+		if (!bnx2x_fp_unlock_napi(fp) &&
+		    !(bnx2x_has_rx_work(fp) || bnx2x_has_tx_work(fp))) {
 
 			/* No need to update SB for FCoE L2 ring as long as
 			 * it's connected to the default SB and the SB
@@ -3096,6 +3117,32 @@ int bnx2x_poll(struct napi_struct *napi, int budget)
 	return work_done;
 }
 
+#ifdef CONFIG_NET_LL_RX_POLL
+/* must be called with local_bh_disable()d */
+int bnx2x_low_latency_recv(struct napi_struct *napi)
+{
+	struct bnx2x_fastpath *fp = container_of(napi, struct bnx2x_fastpath,
+						 napi);
+	struct bnx2x *bp = fp->bp;
+	int found = 0;
+
+	if ((bp->state == BNX2X_STATE_CLOSED) ||
+	    (bp->state == BNX2X_STATE_ERROR) ||
+	    (bp->flags & (TPA_ENABLE_FLAG | GRO_ENABLE_FLAG)))
+		return LL_FLUSH_FAILED;
+
+	if (!bnx2x_fp_lock_poll(fp))
+		return LL_FLUSH_BUSY;
+
+	if (bnx2x_has_rx_work(fp))
+		found = bnx2x_rx_int(fp, 4);
+
+	bnx2x_fp_unlock_poll(fp);
+
+	return found;
+}
+#endif
+
 /* we split the first BD into headers and data BDs
  * to ease the pain of our fellow microcode engineers
  * we use one mapping for both BDs
@@ -3496,9 +3543,12 @@ static void bnx2x_update_pbds_gso_enc(struct sk_buff *skb,
 	/* outer IP header info */
 	if (xmit_type & XMIT_CSUM_V4) {
 		struct iphdr *iph = ip_hdr(skb);
+		u16 csum = (__force u16)(~iph->check) -
+			   (__force u16)iph->tot_len -
+			   (__force u16)iph->frag_off;
+
 		pbd2->fw_ip_csum_wo_len_flags_frag =
-			bswab16(csum_fold((~iph->check) -
-					  iph->tot_len - iph->frag_off));
+			bswab16(csum_fold((__force __wsum)csum));
 	} else {
 		pbd2->fw_ip_hdr_to_payload_w =
 			hlen_w - ((sizeof(struct ipv6hdr)) >> 1);
@@ -3586,7 +3636,7 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	DP(NETIF_MSG_TX_QUEUED, "indices: txq %d, fp %d, txdata %d\n",
 	   txq_index, fp_index, txdata_index); */
 
-	/* enable this debug print to view the tranmission details
+	/* enable this debug print to view the transmission details
 	DP(NETIF_MSG_TX_QUEUED,
 	   "transmitting packet cid %d fp index %d txdata_index %d tx_data ptr %p fp pointer %p\n",
 	   txdata->cid, fp_index, txdata_index, txdata, fp); */
@@ -3968,7 +4018,7 @@ int bnx2x_setup_tc(struct net_device *dev, u8 num_tc)
 	/* setup tc must be called under rtnl lock */
 	ASSERT_RTNL();
 
-	/* no traffic classes requested. aborting */
+	/* no traffic classes requested. Aborting */
 	if (!num_tc) {
 		netdev_reset_tc(dev);
 		return 0;
@@ -3976,7 +4026,7 @@ int bnx2x_setup_tc(struct net_device *dev, u8 num_tc)
 
 	/* requested to support too many traffic classes */
 	if (num_tc > bp->max_cos) {
-		BNX2X_ERR("support for too many traffic classes requested: %d. max supported is %d\n",
+		BNX2X_ERR("support for too many traffic classes requested: %d. Max supported is %d\n",
 			  num_tc, bp->max_cos);
 		return -EINVAL;
 	}
@@ -3995,8 +4045,7 @@ int bnx2x_setup_tc(struct net_device *dev, u8 num_tc)
 		   prio, bp->prio_to_cos[prio]);
 	}
 
-
-	/* Use this configuration to diffrentiate tc0 from other COSes
+	/* Use this configuration to differentiate tc0 from other COSes
 	   This can be used for ets or pfc, and save the effort of setting
 	   up a multio class queue disc or negotiating DCBX with a switch
 	netdev_set_prio_tc_map(dev, 0, 0);
@@ -4288,10 +4337,11 @@ static int bnx2x_alloc_fp_mem_at(struct bnx2x *bp, int index)
 				&bnx2x_fp(bp, index, rx_desc_mapping),
 				sizeof(struct eth_rx_bd) * NUM_RX_BD);
 
-		BNX2X_PCI_ALLOC(bnx2x_fp(bp, index, rx_comp_ring),
-				&bnx2x_fp(bp, index, rx_comp_mapping),
-				sizeof(struct eth_fast_path_rx_cqe) *
-				NUM_RCQ_BD);
+		/* Seed all CQEs by 1s */
+		BNX2X_PCI_FALLOC(bnx2x_fp(bp, index, rx_comp_ring),
+				 &bnx2x_fp(bp, index, rx_comp_mapping),
+				 sizeof(struct eth_fast_path_rx_cqe) *
+				 NUM_RCQ_BD);
 
 		/* SGE ring */
 		BNX2X_ALLOC(bnx2x_fp(bp, index, rx_page_ring),
@@ -4472,7 +4522,6 @@ int bnx2x_alloc_mem_bp(struct bnx2x *bp)
 alloc_err:
 	bnx2x_free_mem_bp(bp);
 	return -ENOMEM;
-
 }
 
 int bnx2x_reload_if_running(struct net_device *dev)
@@ -4514,7 +4563,6 @@ int bnx2x_get_cur_phy_idx(struct bnx2x *bp)
 	}
 
 	return sel_phy_idx;
-
 }
 int bnx2x_get_link_cfg_idx(struct bnx2x *bp)
 {
@@ -4602,6 +4650,7 @@ int bnx2x_set_features(struct net_device *dev, netdev_features_t features)
 {
 	struct bnx2x *bp = netdev_priv(dev);
 	u32 flags = bp->flags;
+	u32 changes;
 	bool bnx2x_reload = false;
 
 	if (features & NETIF_F_LRO)
@@ -4626,10 +4675,16 @@ int bnx2x_set_features(struct net_device *dev, netdev_features_t features)
 		}
 	}
 
-	if (flags ^ bp->flags) {
-		bp->flags = flags;
+	changes = flags ^ bp->flags;
+
+	/* if GRO is changed while LRO is enabled, don't force a reload */
+	if ((changes & GRO_ENABLE_FLAG) && (flags & TPA_ENABLE_FLAG))
+		changes &= ~GRO_ENABLE_FLAG;
+
+	if (changes)
 		bnx2x_reload = true;
-	}
+
+	bp->flags = flags;
 
 	if (bnx2x_reload) {
 		if (bp->recovery_state == BNX2X_RECOVERY_DONE)
@@ -4724,7 +4779,6 @@ int bnx2x_resume(struct pci_dev *pdev)
 	return rc;
 }
 
-
 void bnx2x_set_ctx_validation(struct bnx2x *bp, struct eth_context *cxt,
 			      u32 cid)
 {
@@ -4742,7 +4796,6 @@ static void storm_memset_hc_timeout(struct bnx2x *bp, u8 port,
 				    u8 fw_sb_id, u8 sb_index,
 				    u8 ticks)
 {
-
 	u32 addr = BAR_CSTRORM_INTMEM +
 		   CSTORM_STATUS_BLOCK_DATA_TIMEOUT_OFFSET(fw_sb_id, sb_index);
 	REG_WR8(bp, addr, ticks);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
index 151675d..c07a6d0 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
@@ -22,7 +22,6 @@
 #include <linux/netdevice.h>
 #include <linux/etherdevice.h>
 
-
 #include "bnx2x.h"
 #include "bnx2x_sriov.h"
 
@@ -50,13 +49,25 @@ extern int int_mode;
 		} \
 	} while (0)
 
-#define BNX2X_PCI_ALLOC(x, y, size)				\
-do {								\
-	x = dma_alloc_coherent(&bp->pdev->dev, size, y,		\
-			       GFP_KERNEL | __GFP_ZERO);	\
-	if (x == NULL)						\
-		goto alloc_mem_err;				\
-} while (0)
+#define BNX2X_PCI_ALLOC(x, y, size) \
+	do { \
+		x = dma_alloc_coherent(&bp->pdev->dev, size, y, \
+				       GFP_KERNEL | __GFP_ZERO); \
+		if (x == NULL) \
+			goto alloc_mem_err; \
+		DP(NETIF_MSG_HW, "BNX2X_PCI_ALLOC: Physical %Lx Virtual %p\n", \
+		   (unsigned long long)(*y), x); \
+	} while (0)
+
+#define BNX2X_PCI_FALLOC(x, y, size) \
+	do { \
+		x = dma_alloc_coherent(&bp->pdev->dev, size, y, GFP_KERNEL); \
+		if (x == NULL) \
+			goto alloc_mem_err; \
+		memset((void *)x, 0xFFFFFFFF, size); \
+		DP(NETIF_MSG_HW, "BNX2X_PCI_FALLOC: Physical %Lx Virtual %p\n",\
+		   (unsigned long long)(*y), x); \
+	} while (0)
 
 #define BNX2X_ALLOC(x, size) \
 	do { \
@@ -494,9 +505,6 @@ void bnx2x_update_max_mf_config(struct bnx2x *bp, u32 value);
 /* Error handling */
 void bnx2x_fw_dump_lvl(struct bnx2x *bp, const char *lvl);
 
-/* validate currect fw is loaded */
-bool bnx2x_test_firmware_version(struct bnx2x *bp, bool is_err);
-
 /* dev_close main block */
 int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode, bool keep_link);
 
@@ -607,6 +615,13 @@ int bnx2x_enable_msi(struct bnx2x *bp);
 int bnx2x_poll(struct napi_struct *napi, int budget);
 
 /**
+ * bnx2x_low_latency_recv - LL callback
+ *
+ * @napi:	napi structure
+ */
+int bnx2x_low_latency_recv(struct napi_struct *napi);
+
+/**
  * bnx2x_alloc_mem_bp - allocate memories outsize main driver structure
  *
  * @bp:		driver handle
@@ -800,16 +815,18 @@ static inline bool bnx2x_has_tx_work(struct bnx2x_fastpath *fp)
 	return false;
 }
 
+#define BNX2X_IS_CQE_COMPLETED(cqe_fp) (cqe_fp->marker == 0x0)
+#define BNX2X_SEED_CQE(cqe_fp) (cqe_fp->marker = 0xFFFFFFFF)
 static inline int bnx2x_has_rx_work(struct bnx2x_fastpath *fp)
 {
-	u16 rx_cons_sb;
+	u16 cons;
+	union eth_rx_cqe *cqe;
+	struct eth_fast_path_rx_cqe *cqe_fp;
 
-	/* Tell compiler that status block fields can change */
-	barrier();
-	rx_cons_sb = le16_to_cpu(*fp->rx_cons_sb);
-	if ((rx_cons_sb & MAX_RCQ_DESC_CNT) == MAX_RCQ_DESC_CNT)
-		rx_cons_sb++;
-	return (fp->rx_comp_cons != rx_cons_sb);
+	cons = RCQ_BD(fp->rx_comp_cons);
+	cqe = &fp->rx_comp_ring[cons];
+	cqe_fp = &cqe->fast_path_cqe;
+	return BNX2X_IS_CQE_COMPLETED(cqe_fp);
 }
 
 /**
@@ -848,9 +865,11 @@ static inline void bnx2x_add_all_napi_cnic(struct bnx2x *bp)
 	int i;
 
 	/* Add NAPI objects */
-	for_each_rx_queue_cnic(bp, i)
+	for_each_rx_queue_cnic(bp, i) {
 		netif_napi_add(bp->dev, &bnx2x_fp(bp, i, napi),
 			       bnx2x_poll, NAPI_POLL_WEIGHT);
+		napi_hash_add(&bnx2x_fp(bp, i, napi));
+	}
 }
 
 static inline void bnx2x_add_all_napi(struct bnx2x *bp)
@@ -858,25 +877,31 @@ static inline void bnx2x_add_all_napi(struct bnx2x *bp)
 	int i;
 
 	/* Add NAPI objects */
-	for_each_eth_queue(bp, i)
+	for_each_eth_queue(bp, i) {
 		netif_napi_add(bp->dev, &bnx2x_fp(bp, i, napi),
 			       bnx2x_poll, NAPI_POLL_WEIGHT);
+		napi_hash_add(&bnx2x_fp(bp, i, napi));
+	}
 }
 
 static inline void bnx2x_del_all_napi_cnic(struct bnx2x *bp)
 {
 	int i;
 
-	for_each_rx_queue_cnic(bp, i)
+	for_each_rx_queue_cnic(bp, i) {
+		napi_hash_del(&bnx2x_fp(bp, i, napi));
 		netif_napi_del(&bnx2x_fp(bp, i, napi));
+	}
 }
 
 static inline void bnx2x_del_all_napi(struct bnx2x *bp)
 {
 	int i;
 
-	for_each_eth_queue(bp, i)
+	for_each_eth_queue(bp, i) {
+		napi_hash_del(&bnx2x_fp(bp, i, napi));
 		netif_napi_del(&bnx2x_fp(bp, i, napi));
+	}
 }
 
 int bnx2x_set_int_mode(struct bnx2x *bp);
@@ -1171,7 +1196,6 @@ static inline u8 bnx2x_cnic_eth_cl_id(struct bnx2x *bp, u8 cl_idx)
 
 static inline u8 bnx2x_cnic_fw_sb_id(struct bnx2x *bp)
 {
-
 	/* the 'first' id is allocated for the cnic */
 	return bp->base_fw_ndsb;
 }
@@ -1181,7 +1205,6 @@ static inline u8 bnx2x_cnic_igu_sb_id(struct bnx2x *bp)
 	return bp->igu_base_sb;
 }
 
-
 static inline void bnx2x_init_fcoe_fp(struct bnx2x *bp)
 {
 	struct bnx2x_fastpath *fp = bnx2x_fcoe_fp(bp);
@@ -1334,8 +1357,8 @@ static inline bool bnx2x_mtu_allows_gro(int mtu)
 	int fpp = SGE_PAGE_SIZE / (mtu - ETH_MAX_TPA_HEADER_SIZE);
 
 	/*
-	 * 1. number of frags should not grow above MAX_SKB_FRAGS
-	 * 2. frag must fit the page
+	 * 1. Number of frags should not grow above MAX_SKB_FRAGS
+	 * 2. Frag must fit the page
 	 */
 	return mtu <= SGE_PAGE_SIZE && (U_ETH_SGL_SIZE * fpp) <= MAX_SKB_FRAGS;
 }
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.c
index 4b077a7..0c94df4 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.c
@@ -253,7 +253,6 @@ static void bnx2x_dcbx_get_ets_feature(struct bnx2x *bp,
 
 	memset(&pg_help_data, 0, sizeof(struct pg_help_data));
 
-
 	if (GET_FLAGS(error, DCBX_LOCAL_ETS_ERROR))
 		DP(BNX2X_MSG_DCB, "DCBX_LOCAL_ETS_ERROR\n");
 
@@ -298,7 +297,6 @@ static void bnx2x_dcbx_get_ets_feature(struct bnx2x *bp,
 static void  bnx2x_dcbx_get_pfc_feature(struct bnx2x *bp,
 					struct dcbx_pfc_feature *pfc, u32 error)
 {
-
 	if (GET_FLAGS(error, DCBX_LOCAL_PFC_ERROR))
 		DP(BNX2X_MSG_DCB, "DCBX_LOCAL_PFC_ERROR\n");
 
@@ -367,7 +365,6 @@ static int bnx2x_dcbx_read_mib(struct bnx2x *bp,
 	struct lldp_remote_mib *remote_mib ;
 	struct lldp_local_mib  *local_mib;
 
-
 	switch (read_mib_type) {
 	case DCBX_READ_LOCAL_MIB:
 		mib_size = sizeof(struct lldp_local_mib);
@@ -629,7 +626,6 @@ static int bnx2x_dcbx_read_shmem_neg_results(struct bnx2x *bp)
 	return 0;
 }
 
-
 #ifdef BCM_DCBNL
 static inline
 u8 bnx2x_dcbx_dcbnl_app_up(struct dcbx_app_priority_entry *ent)
@@ -691,7 +687,7 @@ static inline void bnx2x_dcbx_update_tc_mapping(struct bnx2x *bp)
 	}
 
 	/* setup tc must be called under rtnl lock, but we can't take it here
-	 * as we are handling an attetntion on a work queue which must be
+	 * as we are handling an attention on a work queue which must be
 	 * flushed at some rtnl-locked contexts (e.g. if down)
 	 */
 	if (!test_and_set_bit(BNX2X_SP_RTNL_SETUP_TC, &bp->sp_rtnl_state))
@@ -711,7 +707,7 @@ void bnx2x_dcbx_set_params(struct bnx2x *bp, u32 state)
 			 */
 			bnx2x_dcbnl_update_applist(bp, true);
 
-			/* Read rmeote mib if dcbx is in the FW */
+			/* Read remote mib if dcbx is in the FW */
 			if (bnx2x_dcbx_read_shmem_remote_mib(bp))
 				return;
 #endif
@@ -742,7 +738,7 @@ void bnx2x_dcbx_set_params(struct bnx2x *bp, u32 state)
 			bnx2x_dcbx_update_tc_mapping(bp);
 
 			/*
-			 * allow other funtions to update their netdevices
+			 * allow other functions to update their netdevices
 			 * accordingly
 			 */
 			if (IS_MF(bp))
@@ -864,7 +860,7 @@ static void bnx2x_dcbx_admin_mib_updated_params(struct bnx2x *bp,
 			   i, DCBX_PRI_PG_GET(af->ets.pri_pg_tbl, i));
 		}
 
-		/*For IEEE admin_recommendation_bw_precentage
+		/*For IEEE admin_recommendation_bw_percentage
 		 *For IEEE admin_recommendation_ets_pg */
 		af->pfc.pri_en_bitmap = (u8)dp->admin_pfc_bitmap;
 		for (i = 0; i < DCBX_CONFIG_MAX_APP_PROTOCOL; i++) {
@@ -896,13 +892,11 @@ static void bnx2x_dcbx_admin_mib_updated_params(struct bnx2x *bp,
 		}
 
 		af->app.default_pri = (u8)dp->admin_default_priority;
-
 	}
 
 	/* Write the data. */
 	bnx2x_write_data(bp, (u32 *)&admin_mib, offset,
 			 sizeof(struct lldp_admin_mib));
-
 }
 
 void bnx2x_dcbx_set_state(struct bnx2x *bp, bool dcb_on, u32 dcbx_enabled)
@@ -1076,7 +1070,7 @@ static void bnx2x_dcbx_get_num_pg_traf_type(struct bnx2x *bp,
 	bool pg_found  = false;
 	u32 i, traf_type, add_traf_type, add_pg;
 	u32 *ttp = bp->dcbx_port_params.app.traffic_type_priority;
-	struct pg_entry_help_data *data = help_data->data; /*shotcut*/
+	struct pg_entry_help_data *data = help_data->data; /*shortcut*/
 
 	/* Set to invalid */
 	for (i = 0; i < LLFC_DRIVER_TRAFFIC_TYPE_MAX; i++)
@@ -1172,7 +1166,8 @@ static void bnx2x_dcbx_separate_pauseable_from_non(struct bnx2x *bp,
 				DCBX_PG_BW_GET(ets->pg_bw_tbl, pg_entry));
 		else
 			/* If we join a group and one is strict
-			 * than the bw rulls */
+			 * than the bw rules
+			 */
 			cos_data->data[entry].strict =
 						BNX2X_DCBX_STRICT_COS_HIGHEST;
 	}
@@ -1181,7 +1176,6 @@ static void bnx2x_dcbx_separate_pauseable_from_non(struct bnx2x *bp,
 		BNX2X_ERR("dcbx error: Both groups must have priorities\n");
 }
 
-
 #ifndef POWER_OF_2
 #define POWER_OF_2(x)	((0 != x) && (0 == (x & (x-1))))
 #endif
@@ -1284,7 +1278,7 @@ static void bnx2x_dcbx_2cos_limit_cee_single_pg_to_cos_params(struct bnx2x *bp,
 		} else {
 			/* If there are only pauseable priorities or
 			 * only non-pauseable,* the lower priorities go
-			 * to the first queue and the higherpriorities go
+			 * to the first queue and the higher priorities go
 			 * to the second queue.
 			 */
 			cos_data->data[0].pausable =
@@ -1484,7 +1478,7 @@ static void bnx2x_dcbx_2cos_limit_cee_three_pg_to_cos_params(
 		 * queue and one priority goes to the second queue.
 		 *
 		 * We will join this two cases:
-		 * if one is BW limited it will go to the secoend queue
+		 * if one is BW limited it will go to the second queue
 		 * otherwise the last priority will get it
 		 */
 
@@ -1504,7 +1498,8 @@ static void bnx2x_dcbx_2cos_limit_cee_three_pg_to_cos_params(
 				    false == b_found_strict)
 					/* last entry will be handled separately
 					 * If no priority is strict than last
-					 * enty goes to last queue.*/
+					 * entry goes to last queue.
+					 */
 					entry = 1;
 				cos_data->data[entry].pri_join_mask |=
 								pri_tested;
@@ -1516,7 +1511,8 @@ static void bnx2x_dcbx_2cos_limit_cee_three_pg_to_cos_params(
 				b_found_strict = true;
 				cos_data->data[1].pri_join_mask |= pri_tested;
 				/* If we join a group and one is strict
-				 * than the bw rulls */
+				 * than the bw rules
+				 */
 				cos_data->data[1].strict =
 					BNX2X_DCBX_STRICT_COS_HIGHEST;
 			}
@@ -1524,7 +1520,6 @@ static void bnx2x_dcbx_2cos_limit_cee_three_pg_to_cos_params(
 	}
 }
 
-
 static void bnx2x_dcbx_2cos_limit_cee_fill_cos_params(struct bnx2x *bp,
 				       struct pg_help_data *help_data,
 				       struct dcbx_ets_feature *ets,
@@ -1533,7 +1528,6 @@ static void bnx2x_dcbx_2cos_limit_cee_fill_cos_params(struct bnx2x *bp,
 				       u32 pri_join_mask,
 				       u8 num_of_dif_pri)
 {
-
 	/* default E2 settings */
 	cos_data->num_of_cos = DCBX_COS_MAX_NUM_E2;
 
@@ -1629,7 +1623,6 @@ static u8 bnx2x_dcbx_cee_fill_strict_pri(struct bnx2x *bp,
 					 u8 num_spread_of_entries,
 					 u8 strict_app_pris)
 {
-
 	if (bnx2x_dcbx_spread_strict_pri(bp, cos_data, entry,
 					 num_spread_of_entries,
 					 strict_app_pris)) {
@@ -1848,7 +1841,7 @@ static void bnx2x_dcbx_fw_struct(struct bnx2x *bp,
 
 void bnx2x_dcbx_pmf_update(struct bnx2x *bp)
 {
-	/* if we need to syncronize DCBX result from prev PMF
+	/* if we need to synchronize DCBX result from prev PMF
 	 * read it from shmem and update bp and netdev accordingly
 	 */
 	if (SHMEM2_HAS(bp, drv_flags) &&
@@ -1876,7 +1869,6 @@ void bnx2x_dcbx_pmf_update(struct bnx2x *bp)
 		 * dcbx negotiation.
 		 */
 		bnx2x_dcbx_update_tc_mapping(bp);
-
 	}
 }
 
@@ -1943,14 +1935,14 @@ static void bnx2x_dcbnl_set_pg_tccfg_tx(struct net_device *netdev, int prio,
 		return;
 
 	/**
-	 * bw_pct ingnored -	band-width percentage devision between user
+	 * bw_pct ignored -	band-width percentage devision between user
 	 *			priorities within the same group is not
 	 *			standard and hence not supported
 	 *
-	 * prio_type igonred -	priority levels within the same group are not
+	 * prio_type ignored -	priority levels within the same group are not
 	 *			standard and hence are not supported. According
 	 *			to the standard pgid 15 is dedicated to strict
-	 *			prioirty traffic (on the port level).
+	 *			priority traffic (on the port level).
 	 *
 	 * up_map ignored
 	 */
@@ -1995,14 +1987,14 @@ static void bnx2x_dcbnl_get_pg_tccfg_tx(struct net_device *netdev, int prio,
 	DP(BNX2X_MSG_DCB, "prio = %d\n", prio);
 
 	/**
-	 * bw_pct ingnored -	band-width percentage devision between user
+	 * bw_pct ignored -	band-width percentage devision between user
 	 *			priorities within the same group is not
 	 *			standard and hence not supported
 	 *
-	 * prio_type igonred -	priority levels within the same group are not
+	 * prio_type ignored -	priority levels within the same group are not
 	 *			standard and hence are not supported. According
 	 *			to the standard pgid 15 is dedicated to strict
-	 *			prioirty traffic (on the port level).
+	 *			priority traffic (on the port level).
 	 *
 	 * up_map ignored
 	 */
@@ -2389,7 +2381,7 @@ static u8 bnx2x_dcbnl_get_featcfg(struct net_device *netdev, int featid,
 				*flags |= DCB_FEATCFG_ERROR;
 			break;
 		default:
-			BNX2X_ERR("Non valid featrue-ID\n");
+			BNX2X_ERR("Non valid feature-ID\n");
 			rval = 1;
 			break;
 		}
@@ -2430,7 +2422,7 @@ static u8 bnx2x_dcbnl_set_featcfg(struct net_device *netdev, int featid,
 				flags & DCB_FEATCFG_WILLING ? 1 : 0;
 			break;
 		default:
-			BNX2X_ERR("Non valid featrue-ID\n");
+			BNX2X_ERR("Non valid feature-ID\n");
 			rval = 1;
 			break;
 		}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.h
index d153f44..125bd1b 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dcb.h
@@ -134,8 +134,6 @@ enum {
 #define PFC_BRB1_REG_HIGH_LLFC_LOW_THRESHOLD			130
 #define PFC_BRB1_REG_HIGH_LLFC_HIGH_THRESHOLD			170
 
-
-
 struct cos_entry_help_data {
 	u32			pri_join_mask;
 	u32			cos_bw;
@@ -170,7 +168,6 @@ struct cos_help_data {
 			(!(IS_DCBX_PFC_PRI_ONLY_NON_PAUSE((bp), (pg_pri)) || \
 			 IS_DCBX_PFC_PRI_ONLY_PAUSE((bp), (pg_pri))))
 
-
 struct pg_entry_help_data {
 	u8	num_of_dif_pri;
 	u8	pg;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h
index bff5e33..12eb4ba 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_dump.h
@@ -13,12 +13,6 @@
  * consent.
  */
 
-
-/* This struct holds a signature to ensure the dump returned from the driver
- * match the meta data file inserted to grc_dump.tcl
- * The signature is time stamp, diag version and grc_dump version
- */
-
 #ifndef BNX2X_DUMP_H
 #define BNX2X_DUMP_H
 
@@ -28,7 +22,6 @@
 #define DRV_DUMP_USTORM_WAITP_ADDRESS    0x338a80
 #define DRV_DUMP_CSTORM_WAITP_ADDRESS    0x238a80
 
-
 /* Possible Chips */
 #define DUMP_CHIP_E1 1
 #define DUMP_CHIP_E1H 2
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
index ce1a916..c5f2251 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
@@ -320,7 +320,7 @@ static int bnx2x_set_settings(struct net_device *dev, struct ethtool_cmd *cmd)
 
 	speed = ethtool_cmd_speed(cmd);
 
-	/* If recieved a request for an unknown duplex, assume full*/
+	/* If received a request for an unknown duplex, assume full*/
 	if (cmd->duplex == DUPLEX_UNKNOWN)
 		cmd->duplex = DUPLEX_FULL;
 
@@ -733,7 +733,6 @@ static bool bnx2x_is_reg_in_chip(struct bnx2x *bp,
 		return false;
 }
 
-
 static bool bnx2x_is_wreg_in_chip(struct bnx2x *bp,
 	const struct wreg_addr *wreg_info)
 {
@@ -850,7 +849,7 @@ static int __bnx2x_get_preset_regs(struct bnx2x *bp, u32 *p, u32 preset)
 
 	/* Paged registers are supported in E2 & E3 only */
 	if (CHIP_IS_E2(bp) || CHIP_IS_E3(bp)) {
-		/* Read "paged" registes */
+		/* Read "paged" registers */
 		bnx2x_read_pages_regs(bp, p, preset);
 	}
 
@@ -960,6 +959,9 @@ static int bnx2x_set_dump(struct net_device *dev, struct ethtool_dump *val)
 	struct bnx2x *bp = netdev_priv(dev);
 
 	/* Use the ethtool_dump "flag" field as the dump preset index */
+	if (val->flag < 1 || val->flag > DUMP_MAX_PRESETS)
+		return -EINVAL;
+
 	bp->dump_preset_idx = val->flag;
 	return 0;
 }
@@ -969,12 +971,12 @@ static int bnx2x_get_dump_flag(struct net_device *dev,
 {
 	struct bnx2x *bp = netdev_priv(dev);
 
+	dump->version = BNX2X_DUMP_VERSION;
+	dump->flag = bp->dump_preset_idx;
 	/* Calculate the requested preset idx length */
 	dump->len = bnx2x_get_preset_regs_len(dev, bp->dump_preset_idx);
 	DP(BNX2X_MSG_ETHTOOL, "Get dump preset %d length=%d\n",
 	   bp->dump_preset_idx, dump->len);
-
-	dump->flag = ETHTOOL_GET_DUMP_DATA;
 	return 0;
 }
 
@@ -986,8 +988,6 @@ static int bnx2x_get_dump_data(struct net_device *dev,
 	struct bnx2x *bp = netdev_priv(dev);
 	struct dump_header dump_hdr = {0};
 
-	memset(p, 0, dump->len);
-
 	/* Disable parity attentions as long as following dump may
 	 * cause false alarms by reading never written registers. We
 	 * will re-enable parity attentions right after the dump.
@@ -1155,8 +1155,8 @@ static int bnx2x_get_eeprom_len(struct net_device *dev)
 	return bp->common.flash_size;
 }
 
-/* Per pf misc lock must be aquired before the per port mcp lock. Otherwise, had
- * we done things the other way around, if two pfs from the same port would
+/* Per pf misc lock must be acquired before the per port mcp lock. Otherwise,
+ * had we done things the other way around, if two pfs from the same port would
  * attempt to access nvram at the same time, we could run into a scenario such
  * as:
  * pf A takes the port lock.
@@ -1381,12 +1381,29 @@ static int bnx2x_nvram_read32(struct bnx2x *bp, u32 offset, u32 *buf,
 	return rc;
 }
 
+static bool bnx2x_is_nvm_accessible(struct bnx2x *bp)
+{
+	int rc = 1;
+	u16 pm = 0;
+	struct net_device *dev = pci_get_drvdata(bp->pdev);
+
+	if (bp->pm_cap)
+		rc = pci_read_config_word(bp->pdev,
+					  bp->pm_cap + PCI_PM_CTRL, &pm);
+
+	if ((rc && !netif_running(dev)) ||
+	    (!rc && ((pm & PCI_PM_CTRL_STATE_MASK) != (__force u16)PCI_D0)))
+		return false;
+
+	return true;
+}
+
 static int bnx2x_get_eeprom(struct net_device *dev,
 			    struct ethtool_eeprom *eeprom, u8 *eebuf)
 {
 	struct bnx2x *bp = netdev_priv(dev);
 
-	if (!netif_running(dev)) {
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL  | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return -EAGAIN;
@@ -1411,7 +1428,7 @@ static int bnx2x_get_module_eeprom(struct net_device *dev,
 	u8 *user_data = data;
 	unsigned int start_addr = ee->offset, xfer_size = 0;
 
-	if (!netif_running(dev)) {
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return -EAGAIN;
@@ -1474,7 +1491,7 @@ static int bnx2x_get_module_info(struct net_device *dev,
 	int phy_idx, rc;
 	u8 sff8472_comp, diag_type;
 
-	if (!netif_running(dev)) {
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return -EAGAIN;
@@ -1594,8 +1611,10 @@ static int bnx2x_nvram_write1(struct bnx2x *bp, u32 offset, u8 *data_buf,
 		 */
 		val = be32_to_cpu(val_be);
 
-		val &= ~le32_to_cpu(0xff << BYTE_OFFSET(offset));
-		val |= le32_to_cpu(*data_buf << BYTE_OFFSET(offset));
+		val &= ~le32_to_cpu((__force __le32)
+				    (0xff << BYTE_OFFSET(offset)));
+		val |= le32_to_cpu((__force __le32)
+				   (*data_buf << BYTE_OFFSET(offset)));
 
 		rc = bnx2x_nvram_write_dword(bp, align_offset, val,
 					     cmd_flags);
@@ -1676,7 +1695,8 @@ static int bnx2x_set_eeprom(struct net_device *dev,
 	int port = BP_PORT(bp);
 	int rc = 0;
 	u32 ext_phy_config;
-	if (!netif_running(dev)) {
+
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return -EAGAIN;
@@ -1921,6 +1941,19 @@ static const char bnx2x_tests_str_arr[BNX2X_NUM_TESTS_SF][ETH_GSTRING_LEN] = {
 	"link_test (online)         "
 };
 
+enum {
+	BNX2X_PRI_FLAG_ISCSI,
+	BNX2X_PRI_FLAG_FCOE,
+	BNX2X_PRI_FLAG_STORAGE,
+	BNX2X_PRI_FLAG_LEN,
+};
+
+static const char bnx2x_private_arr[BNX2X_PRI_FLAG_LEN][ETH_GSTRING_LEN] = {
+	"iSCSI offload support",
+	"FCoE offload support",
+	"Storage only interface"
+};
+
 static u32 bnx2x_eee_to_adv(u32 eee_adv)
 {
 	u32 modes = 0;
@@ -2041,7 +2074,7 @@ static int bnx2x_set_eee(struct net_device *dev, struct ethtool_eee *edata)
 				    EEE_MODE_OVERRIDE_NVRAM |
 				    EEE_MODE_OUTPUT_TIME;
 
-	/* Restart link to propogate changes */
+	/* Restart link to propagate changes */
 	if (netif_running(dev)) {
 		bnx2x_stats_handle(bp, STATS_EVENT_STOP);
 		bnx2x_force_link_reset(bp);
@@ -2160,7 +2193,7 @@ static int bnx2x_test_registers(struct bnx2x *bp)
 		{ BNX2X_CHIP_MASK_ALL, 0xffffffff, 0, 0x00000000 }
 	};
 
-	if (!netif_running(bp->dev)) {
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return rc;
@@ -2264,7 +2297,7 @@ static int bnx2x_test_memory(struct bnx2x *bp)
 		{ NULL, 0xffffffff, {0, 0, 0, 0} }
 	};
 
-	if (!netif_running(bp->dev)) {
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return rc;
@@ -2978,32 +3011,47 @@ static int bnx2x_num_stat_queues(struct bnx2x *bp)
 static int bnx2x_get_sset_count(struct net_device *dev, int stringset)
 {
 	struct bnx2x *bp = netdev_priv(dev);
-	int i, num_stats;
+	int i, num_strings = 0;
 
 	switch (stringset) {
 	case ETH_SS_STATS:
 		if (is_multi(bp)) {
-			num_stats = bnx2x_num_stat_queues(bp) *
-						BNX2X_NUM_Q_STATS;
+			num_strings = bnx2x_num_stat_queues(bp) *
+				      BNX2X_NUM_Q_STATS;
 		} else
-			num_stats = 0;
+			num_strings = 0;
 		if (IS_MF_MODE_STAT(bp)) {
 			for (i = 0; i < BNX2X_NUM_STATS; i++)
 				if (IS_FUNC_STAT(i))
-					num_stats++;
+					num_strings++;
 		} else
-			num_stats += BNX2X_NUM_STATS;
+			num_strings += BNX2X_NUM_STATS;
 
-		return num_stats;
+		return num_strings;
 
 	case ETH_SS_TEST:
 		return BNX2X_NUM_TESTS(bp);
 
+	case ETH_SS_PRIV_FLAGS:
+		return BNX2X_PRI_FLAG_LEN;
+
 	default:
 		return -EINVAL;
 	}
 }
 
+static u32 bnx2x_get_private_flags(struct net_device *dev)
+{
+	struct bnx2x *bp = netdev_priv(dev);
+	u32 flags = 0;
+
+	flags |= (!(bp->flags & NO_ISCSI_FLAG) ? 1 : 0) << BNX2X_PRI_FLAG_ISCSI;
+	flags |= (!(bp->flags & NO_FCOE_FLAG)  ? 1 : 0) << BNX2X_PRI_FLAG_FCOE;
+	flags |= (!!IS_MF_STORAGE_ONLY(bp)) << BNX2X_PRI_FLAG_STORAGE;
+
+	return flags;
+}
+
 static void bnx2x_get_strings(struct net_device *dev, u32 stringset, u8 *buf)
 {
 	struct bnx2x *bp = netdev_priv(dev);
@@ -3026,7 +3074,6 @@ static void bnx2x_get_strings(struct net_device *dev, u32 stringset, u8 *buf)
 			}
 		}
 
-
 		for (i = 0, j = 0; i < BNX2X_NUM_STATS; i++) {
 			if (IS_MF_MODE_STAT(bp) && IS_PORT_STAT(i))
 				continue;
@@ -3045,6 +3092,12 @@ static void bnx2x_get_strings(struct net_device *dev, u32 stringset, u8 *buf)
 			start = 4;
 		memcpy(buf, bnx2x_tests_str_arr + start,
 		       ETH_GSTRING_LEN * BNX2X_NUM_TESTS(bp));
+		break;
+
+	case ETH_SS_PRIV_FLAGS:
+		memcpy(buf, bnx2x_private_arr,
+		       ETH_GSTRING_LEN * BNX2X_PRI_FLAG_LEN);
+		break;
 	}
 }
 
@@ -3106,17 +3159,12 @@ static int bnx2x_set_phys_id(struct net_device *dev,
 {
 	struct bnx2x *bp = netdev_priv(dev);
 
-	if (!netif_running(dev)) {
+	if (!bnx2x_is_nvm_accessible(bp)) {
 		DP(BNX2X_MSG_ETHTOOL | BNX2X_MSG_NVM,
 		   "cannot access eeprom when the interface is down\n");
 		return -EAGAIN;
 	}
 
-	if (!bp->port.pmf) {
-		DP(BNX2X_MSG_ETHTOOL, "Interface is not pmf\n");
-		return -EOPNOTSUPP;
-	}
-
 	switch (state) {
 	case ETHTOOL_ID_ACTIVE:
 		return 1;	/* cycle on/off once per second */
@@ -3148,7 +3196,6 @@ static int bnx2x_set_phys_id(struct net_device *dev,
 
 static int bnx2x_get_rss_flags(struct bnx2x *bp, struct ethtool_rxnfc *info)
 {
-
 	switch (info->flow_type) {
 	case TCP_V4_FLOW:
 	case TCP_V6_FLOW:
@@ -3384,7 +3431,6 @@ static int bnx2x_set_channels(struct net_device *dev,
 {
 	struct bnx2x *bp = netdev_priv(dev);
 
-
 	DP(BNX2X_MSG_ETHTOOL,
 	   "set-channels command parameters: rx = %d, tx = %d, other = %d, combined = %d\n",
 	   channels->rx_count, channels->tx_count, channels->other_count,
@@ -3445,6 +3491,7 @@ static const struct ethtool_ops bnx2x_ethtool_ops = {
 	.set_pauseparam		= bnx2x_set_pauseparam,
 	.self_test		= bnx2x_self_test,
 	.get_sset_count		= bnx2x_get_sset_count,
+	.get_priv_flags		= bnx2x_get_private_flags,
 	.get_strings		= bnx2x_get_strings,
 	.set_phys_id		= bnx2x_set_phys_id,
 	.get_ethtool_stats	= bnx2x_get_ethtool_stats,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
index 12f00a4..5018e52 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
@@ -1323,6 +1323,8 @@ struct drv_func_mb {
 	#define DRV_MSG_CODE_UNLOAD_SKIP_LINK_RESET     0x00000002
 
 	#define DRV_MSG_CODE_LOAD_REQ_WITH_LFA          0x0000100a
+	#define DRV_MSG_CODE_LOAD_REQ_FORCE_LFA         0x00002000
+
 	u32 fw_mb_header;
 	#define FW_MSG_CODE_MASK                        0xffff0000
 	#define FW_MSG_CODE_DRV_LOAD_COMMON             0x10100000
@@ -3816,7 +3818,8 @@ struct eth_fast_path_rx_cqe {
 	__le16 len_on_bd;
 	struct parsing_flags pars_flags;
 	union eth_sgl_or_raw_data sgl_or_raw_data;
-	__le32 reserved1[8];
+	__le32 reserved1[7];
+	u32 marker;
 };
 
 
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index b4c9dea..15a528b 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -93,7 +93,6 @@ MODULE_FIRMWARE(FW_FILE_NAME_E1);
 MODULE_FIRMWARE(FW_FILE_NAME_E1H);
 MODULE_FIRMWARE(FW_FILE_NAME_E2);
 
-
 int num_queues;
 module_param(num_queues, int, 0);
 MODULE_PARM_DESC(num_queues,
@@ -103,8 +102,6 @@ static int disable_tpa;
 module_param(disable_tpa, int, 0);
 MODULE_PARM_DESC(disable_tpa, " Disable the TPA (LRO) feature");
 
-#define INT_MODE_INTx			1
-#define INT_MODE_MSI			2
 int int_mode;
 module_param(int_mode, int, 0);
 MODULE_PARM_DESC(int_mode, " Force interrupt mode other than MSI-X "
@@ -122,8 +119,6 @@ static int debug;
 module_param(debug, int, 0);
 MODULE_PARM_DESC(debug, " Default debug msglevel");
 
-
-
 struct workqueue_struct *bnx2x_wq;
 
 struct bnx2x_mac_vals {
@@ -376,9 +371,11 @@ static u32 bnx2x_reg_rd_ind(struct bnx2x *bp, u32 addr)
 #define DMAE_DP_DST_PCI		"pci dst_addr [%x:%08x]"
 #define DMAE_DP_DST_NONE	"dst_addr [none]"
 
-void bnx2x_dp_dmae(struct bnx2x *bp, struct dmae_command *dmae, int msglvl)
+static void bnx2x_dp_dmae(struct bnx2x *bp,
+			  struct dmae_command *dmae, int msglvl)
 {
 	u32 src_type = dmae->opcode & DMAE_COMMAND_SRC;
+	int i;
 
 	switch (dmae->opcode & DMAE_COMMAND_DST) {
 	case DMAE_CMD_DST_PCI:
@@ -434,6 +431,10 @@ void bnx2x_dp_dmae(struct bnx2x *bp, struct dmae_command *dmae, int msglvl)
 			   dmae->comp_val);
 		break;
 	}
+
+	for (i = 0; i < (sizeof(struct dmae_command)/4); i++)
+		DP(msglvl, "DMAE RAW [%02d]: 0x%08x\n",
+		   i, *(((u32 *)dmae) + i));
 }
 
 /* copy command into DMAE command memory and set DMAE command go */
@@ -508,8 +509,9 @@ int bnx2x_issue_dmae_with_comp(struct bnx2x *bp, struct dmae_command *dmae)
 	int cnt = CHIP_REV_IS_SLOW(bp) ? (400000) : 4000;
 	int rc = 0;
 
-	/*
-	 * Lock the dmae channel. Disable BHs to prevent a dead-lock
+	bnx2x_dp_dmae(bp, dmae, BNX2X_MSG_DMAE);
+
+	/* Lock the dmae channel. Disable BHs to prevent a dead-lock
 	 * as long as this code is called both from syscall context and
 	 * from ndo_set_rx_mode() flow that may be called from BH.
 	 */
@@ -548,6 +550,7 @@ unlock:
 void bnx2x_write_dmae(struct bnx2x *bp, dma_addr_t dma_addr, u32 dst_addr,
 		      u32 len32)
 {
+	int rc;
 	struct dmae_command dmae;
 
 	if (!bp->dmae_ready) {
@@ -571,11 +574,16 @@ void bnx2x_write_dmae(struct bnx2x *bp, dma_addr_t dma_addr, u32 dst_addr,
 	dmae.len = len32;
 
 	/* issue the command and wait for completion */
-	bnx2x_issue_dmae_with_comp(bp, &dmae);
+	rc = bnx2x_issue_dmae_with_comp(bp, &dmae);
+	if (rc) {
+		BNX2X_ERR("DMAE returned failure %d\n", rc);
+		bnx2x_panic();
+	}
 }
 
 void bnx2x_read_dmae(struct bnx2x *bp, u32 src_addr, u32 len32)
 {
+	int rc;
 	struct dmae_command dmae;
 
 	if (!bp->dmae_ready) {
@@ -603,7 +611,11 @@ void bnx2x_read_dmae(struct bnx2x *bp, u32 src_addr, u32 len32)
 	dmae.len = len32;
 
 	/* issue the command and wait for completion */
-	bnx2x_issue_dmae_with_comp(bp, &dmae);
+	rc = bnx2x_issue_dmae_with_comp(bp, &dmae);
+	if (rc) {
+		BNX2X_ERR("DMAE returned failure %d\n", rc);
+		bnx2x_panic();
+	}
 }
 
 static void bnx2x_write_dmae_phys_len(struct bnx2x *bp, dma_addr_t phys_addr,
@@ -811,8 +823,8 @@ static void bnx2x_hc_int_disable(struct bnx2x *bp)
 	u32 val = REG_RD(bp, addr);
 
 	/* in E1 we must use only PCI configuration space to disable
-	 * MSI/MSIX capablility
-	 * It's forbitten to disable IGU_PF_CONF_MSI_MSIX_EN in HC block
+	 * MSI/MSIX capability
+	 * It's forbidden to disable IGU_PF_CONF_MSI_MSIX_EN in HC block
 	 */
 	if (CHIP_IS_E1(bp)) {
 		/* Since IGU_PF_CONF_MSI_MSIX_EN still always on
@@ -839,7 +851,7 @@ static void bnx2x_hc_int_disable(struct bnx2x *bp)
 
 	REG_WR(bp, addr, val);
 	if (REG_RD(bp, addr) != val)
-		BNX2X_ERR("BUG! proper val not read from IGU!\n");
+		BNX2X_ERR("BUG! Proper val not read from IGU!\n");
 }
 
 static void bnx2x_igu_int_disable(struct bnx2x *bp)
@@ -857,7 +869,7 @@ static void bnx2x_igu_int_disable(struct bnx2x *bp)
 
 	REG_WR(bp, IGU_REG_PF_CONFIGURATION, val);
 	if (REG_RD(bp, IGU_REG_PF_CONFIGURATION) != val)
-		BNX2X_ERR("BUG! proper val not read from IGU!\n");
+		BNX2X_ERR("BUG! Proper val not read from IGU!\n");
 }
 
 static void bnx2x_int_disable(struct bnx2x *bp)
@@ -917,7 +929,6 @@ void bnx2x_panic_dump(struct bnx2x *bp, bool disable_int)
 	       sp_sb_data.p_func.vf_valid,
 	       sp_sb_data.state);
 
-
 	for_each_eth_queue(bp, i) {
 		struct bnx2x_fastpath *fp = &bp->fp[i];
 		int loop;
@@ -1016,7 +1027,7 @@ void bnx2x_panic_dump(struct bnx2x *bp, bool disable_int)
 				hc_sm_p[j].timer_value);
 		}
 
-		/* Indecies data */
+		/* Indices data */
 		for (j = 0; j < loop; j++) {
 			pr_cont("INDEX[%d] flags (0x%x) timeout (0x%x)\n", j,
 			       hc_index_p[j].flags,
@@ -1027,6 +1038,7 @@ void bnx2x_panic_dump(struct bnx2x *bp, bool disable_int)
 #ifdef BNX2X_STOP_ON_ERROR
 
 	/* event queue */
+	BNX2X_ERR("eq cons %x prod %x\n", bp->eq_cons, bp->eq_prod);
 	for (i = 0; i < NUM_EQ_DESC; i++) {
 		u32 *data = (u32 *)&bp->eq_ring[i].message.data;
 
@@ -1111,7 +1123,7 @@ void bnx2x_panic_dump(struct bnx2x *bp, bool disable_int)
  * bnx2x_pf_flr_clnup() is called during nic_load in the per function HW
  * initialization.
  */
-#define FLR_WAIT_USEC		10000	/* 10 miliseconds */
+#define FLR_WAIT_USEC		10000	/* 10 milliseconds */
 #define FLR_WAIT_INTERVAL	50	/* usec */
 #define	FLR_POLL_CNT		(FLR_WAIT_USEC/FLR_WAIT_INTERVAL) /* 200 */
 
@@ -1290,7 +1302,6 @@ void bnx2x_tx_hw_flushed(struct bnx2x *bp, u32 poll_count)
 	for (i = 0; i < ARRAY_SIZE(cmd_regs); i++)
 		bnx2x_pbf_pN_cmd_flushed(bp, &cmd_regs[i], poll_count);
 
-
 	/* Verify the transmission buffers are flushed P0, P1, P4 */
 	for (i = 0; i < ARRAY_SIZE(buf_regs); i++)
 		bnx2x_pbf_pN_buf_flushed(bp, &buf_regs[i], poll_count);
@@ -1305,11 +1316,9 @@ void bnx2x_tx_hw_flushed(struct bnx2x *bp, u32 poll_count)
 #define OP_GEN_AGG_VECT(index) \
 	(((index) << SDM_OP_GEN_AGG_VECT_IDX_SHIFT) & SDM_OP_GEN_AGG_VECT_IDX)
 
-
 int bnx2x_send_final_clnup(struct bnx2x *bp, u8 clnup_func, u32 poll_cnt)
 {
 	u32 op_gen_command = 0;
-
 	u32 comp_addr = BAR_CSTRORM_INTMEM +
 			CSTORM_FINAL_CLEANUP_COMPLETE_OFFSET(clnup_func);
 	int ret = 0;
@@ -1334,7 +1343,7 @@ int bnx2x_send_final_clnup(struct bnx2x *bp, u8 clnup_func, u32 poll_cnt)
 		bnx2x_panic();
 		return 1;
 	}
-	/* Zero completion for nxt FLR */
+	/* Zero completion for next FLR */
 	REG_WR(bp, comp_addr, 0);
 
 	return ret;
@@ -1352,7 +1361,6 @@ u8 bnx2x_is_pcie_pending(struct pci_dev *dev)
 */
 static int bnx2x_poll_hw_usage_counters(struct bnx2x *bp, u32 poll_cnt)
 {
-
 	/* wait for CFC PF usage-counter to zero (includes all the VFs) */
 	if (bnx2x_flr_clnup_poll_hw_counter(bp,
 			CFC_REG_NUM_LCIDS_INSIDE_PF,
@@ -1360,7 +1368,6 @@ static int bnx2x_poll_hw_usage_counters(struct bnx2x *bp, u32 poll_cnt)
 			poll_cnt))
 		return 1;
 
-
 	/* Wait for DQ PF usage-counter to zero (until DQ cleanup) */
 	if (bnx2x_flr_clnup_poll_hw_counter(bp,
 			DORQ_REG_PF_USAGE_CNT,
@@ -1390,7 +1397,7 @@ static int bnx2x_poll_hw_usage_counters(struct bnx2x *bp, u32 poll_cnt)
 	/* Wait DMAE PF usage counter to zero */
 	if (bnx2x_flr_clnup_poll_hw_counter(bp,
 			dmae_reg_go_c[INIT_DMAE_C(bp)],
-			"DMAE dommand register timed out",
+			"DMAE command register timed out",
 			poll_cnt))
 		return 1;
 
@@ -1770,7 +1777,7 @@ void bnx2x_sp_event(struct bnx2x_fastpath *fp, union eth_rx_cqe *rr_cqe)
 		break;
 
 	case (RAMROD_CMD_ID_ETH_TERMINATE):
-		DP(BNX2X_MSG_SP, "got MULTI[%d] teminate ramrod\n", cid);
+		DP(BNX2X_MSG_SP, "got MULTI[%d] terminate ramrod\n", cid);
 		drv_cmd = BNX2X_Q_CMD_TERMINATE;
 		break;
 
@@ -1859,7 +1866,6 @@ irqreturn_t bnx2x_interrupt(int irq, void *dev_instance)
 		mask = 0x2 << (fp->index + CNIC_SUPPORT(bp));
 		if (status & mask) {
 			/* Handle Rx or Tx according to SB id */
-			prefetch(fp->rx_cons_sb);
 			for_each_cos_in_tx_queue(fp, cos)
 				prefetch(fp->txdata_ptr[cos]->tx_cons_sb);
 			prefetch(&fp->sb_running_index[SM_RX_ID]);
@@ -1947,7 +1953,7 @@ int bnx2x_acquire_hw_lock(struct bnx2x *bp, u32 resource)
 		if (lock_status & resource_bit)
 			return 0;
 
-		msleep(5);
+		usleep_range(5000, 10000);
 	}
 	BNX2X_ERR("Timeout\n");
 	return -EAGAIN;
@@ -1982,8 +1988,8 @@ int bnx2x_release_hw_lock(struct bnx2x *bp, u32 resource)
 	/* Validating that the resource is currently taken */
 	lock_status = REG_RD(bp, hw_lock_control_reg);
 	if (!(lock_status & resource_bit)) {
-		BNX2X_ERR("lock_status 0x%x resource_bit 0x%x. unlock was called but lock wasn't taken!\n",
-		   lock_status, resource_bit);
+		BNX2X_ERR("lock_status 0x%x resource_bit 0x%x. Unlock was called but lock wasn't taken!\n",
+			  lock_status, resource_bit);
 		return -EFAULT;
 	}
 
@@ -1991,7 +1997,6 @@ int bnx2x_release_hw_lock(struct bnx2x *bp, u32 resource)
 	return 0;
 }
 
-
 int bnx2x_get_gpio(struct bnx2x *bp, int gpio_num, u8 port)
 {
 	/* The GPIO should be swapped if swap register is set and active */
@@ -2347,14 +2352,13 @@ u8 bnx2x_link_test(struct bnx2x *bp, u8 is_serdes)
 	return rc;
 }
 
-
 /* Calculates the sum of vn_min_rates.
    It's needed for further normalizing of the min_rates.
    Returns:
      sum of vn_min_rates.
        or
      0 - if all the min_rates are 0.
-     In the later case fainess algorithm should be deactivated.
+     In the later case fairness algorithm should be deactivated.
      If not all min_rates are zero then those that are zeroes will be set to 1.
  */
 static void bnx2x_calc_vn_min(struct bnx2x *bp,
@@ -2419,7 +2423,6 @@ static void bnx2x_calc_vn_max(struct bnx2x *bp, int vn,
 	input->vnic_max_rate[vn] = vn_max_rate;
 }
 
-
 static int bnx2x_get_cmng_fns_mode(struct bnx2x *bp)
 {
 	if (CHIP_REV_IS_SLOW(bp))
@@ -2435,7 +2438,7 @@ void bnx2x_read_mf_cfg(struct bnx2x *bp)
 	int vn, n = (CHIP_MODE_IS_4_PORT(bp) ? 2 : 1);
 
 	if (BP_NOMCP(bp))
-		return; /* what should be the default bvalue in this case */
+		return; /* what should be the default value in this case */
 
 	/* For 2 port configuration the absolute function number formula
 	 * is:
@@ -2901,7 +2904,6 @@ u32 bnx2x_fw_command(struct bnx2x *bp, u32 command, u32 param)
 	return rc;
 }
 
-
 static void storm_memset_func_cfg(struct bnx2x *bp,
 				 struct tstorm_eth_function_common_config *tcfg,
 				 u16 abs_fid)
@@ -2935,7 +2937,7 @@ void bnx2x_func_init(struct bnx2x *bp, struct bnx2x_func_init_params *p)
 }
 
 /**
- * bnx2x_get_tx_only_flags - Return common flags
+ * bnx2x_get_common_flags - Return common flags
  *
  * @bp		device handle
  * @fp		queue handle
@@ -3006,7 +3008,6 @@ static unsigned long bnx2x_get_q_flags(struct bnx2x *bp,
 	if (IS_MF_AFEX(bp))
 		__set_bit(BNX2X_Q_FLG_SILENT_VLAN_REM, &flags);
 
-
 	return flags | bnx2x_get_common_flags(bp, fp, true);
 }
 
@@ -3082,7 +3083,7 @@ static void bnx2x_pf_rx_q_prep(struct bnx2x *bp,
 	 * placed on the BD (not including paddings).
 	 */
 	rxq_init->buf_sz = fp->rx_buf_size - BNX2X_FW_RX_ALIGN_START -
-		BNX2X_FW_RX_ALIGN_END -	IP_HEADER_ALIGNMENT_PADDING;
+			   BNX2X_FW_RX_ALIGN_END - IP_HEADER_ALIGNMENT_PADDING;
 
 	rxq_init->cl_qzone_id = fp->cl_qzone_id;
 	rxq_init->tpa_agg_sz = tpa_agg_size;
@@ -3124,7 +3125,7 @@ static void bnx2x_pf_tx_q_prep(struct bnx2x *bp,
 	txq_init->fw_sb_id = fp->fw_sb_id;
 
 	/*
-	 * set the tss leading client id for TX classfication ==
+	 * set the tss leading client id for TX classification ==
 	 * leading RSS client id
 	 */
 	txq_init->tss_leading_cl_id = bnx2x_fp(bp, 0, cl_id);
@@ -3196,7 +3197,6 @@ static void bnx2x_pf_init(struct bnx2x *bp)
 	storm_memset_eq_data(bp, &eq_data, BP_FUNC(bp));
 }
 
-
 static void bnx2x_e1h_disable(struct bnx2x *bp)
 {
 	int port = BP_PORT(bp);
@@ -3212,7 +3212,7 @@ static void bnx2x_e1h_enable(struct bnx2x *bp)
 
 	REG_WR(bp, NIG_REG_LLH0_FUNC_EN + port*8, 1);
 
-	/* Tx queue should be only reenabled */
+	/* Tx queue should be only re-enabled */
 	netif_tx_wake_all_queues(bp->dev);
 
 	/*
@@ -3540,10 +3540,8 @@ static bool bnx2x_is_contextless_ramrod(int cmd, int cmd_type)
 		return true;
 	else
 		return false;
-
 }
 
-
 /**
  * bnx2x_sp_post - place a single command on an SP ring
  *
@@ -3608,14 +3606,13 @@ int bnx2x_sp_post(struct bnx2x *bp, int command, int cid,
 	/*
 	 * It's ok if the actual decrement is issued towards the memory
 	 * somewhere between the spin_lock and spin_unlock. Thus no
-	 * more explict memory barrier is needed.
+	 * more explicit memory barrier is needed.
 	 */
 	if (common)
 		atomic_dec(&bp->eq_spq_left);
 	else
 		atomic_dec(&bp->cq_spq_left);
 
-
 	DP(BNX2X_MSG_SP,
 	   "SPQE[%x] (%x:%x)  (cmd, common?) (%d,%d)  hw_cid %x  data (%x:%x) type(0x%x) left (CQ, EQ) (%x,%x)\n",
 	   bp->spq_prod_idx, (u32)U64_HI(bp->spq_mapping),
@@ -3637,15 +3634,14 @@ static int bnx2x_acquire_alr(struct bnx2x *bp)
 
 	might_sleep();
 	for (j = 0; j < 1000; j++) {
-		val = (1UL << 31);
-		REG_WR(bp, GRCBASE_MCP + 0x9c, val);
-		val = REG_RD(bp, GRCBASE_MCP + 0x9c);
-		if (val & (1L << 31))
+		REG_WR(bp, MCP_REG_MCPR_ACCESS_LOCK, MCPR_ACCESS_LOCK_LOCK);
+		val = REG_RD(bp, MCP_REG_MCPR_ACCESS_LOCK);
+		if (val & MCPR_ACCESS_LOCK_LOCK)
 			break;
 
-		msleep(5);
+		usleep_range(5000, 10000);
 	}
-	if (!(val & (1L << 31))) {
+	if (!(val & MCPR_ACCESS_LOCK_LOCK)) {
 		BNX2X_ERR("Cannot acquire MCP access lock register\n");
 		rc = -EBUSY;
 	}
@@ -3656,7 +3652,7 @@ static int bnx2x_acquire_alr(struct bnx2x *bp)
 /* release split MCP access lock register */
 static void bnx2x_release_alr(struct bnx2x *bp)
 {
-	REG_WR(bp, GRCBASE_MCP + 0x9c, 0);
+	REG_WR(bp, MCP_REG_MCPR_ACCESS_LOCK, 0);
 }
 
 #define BNX2X_DEF_SB_ATT_IDX	0x0001
@@ -3678,7 +3674,7 @@ static u16 bnx2x_update_dsb_idx(struct bnx2x *bp)
 		rc |= BNX2X_DEF_SB_IDX;
 	}
 
-	/* Do not reorder: indecies reading should complete before handling */
+	/* Do not reorder: indices reading should complete before handling */
 	barrier();
 	return rc;
 }
@@ -3827,8 +3823,7 @@ static void bnx2x_fan_failure(struct bnx2x *bp)
 	netdev_err(bp->dev, "Fan Failure on Network Controller has caused the driver to shutdown the card to prevent permanent damage.\n"
 			    "Please contact OEM Support for assistance\n");
 
-	/*
-	 * Schedule device reset (unload)
+	/* Schedule device reset (unload)
 	 * This is due to some boards consuming sufficient power when driver is
 	 * up to overheat if fan fails.
 	 */
@@ -3836,7 +3831,6 @@ static void bnx2x_fan_failure(struct bnx2x *bp)
 	set_bit(BNX2X_SP_RTNL_FAN_FAILURE, &bp->sp_rtnl_state);
 	smp_mb__after_clear_bit();
 	schedule_delayed_work(&bp->sp_rtnl_task, 0);
-
 }
 
 static void bnx2x_attn_int_deasserted0(struct bnx2x *bp, u32 attn)
@@ -4106,7 +4100,7 @@ static void bnx2x_clear_reset_global(struct bnx2x *bp)
  */
 static bool bnx2x_reset_is_global(struct bnx2x *bp)
 {
-	u32 val	= REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
+	u32 val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
 
 	DP(NETIF_MSG_HW, "GEN_REG_VAL=0x%08x\n", val);
 	return (val & BNX2X_GLOBAL_RESET_BIT) ? true : false;
@@ -4157,7 +4151,7 @@ void bnx2x_set_reset_in_progress(struct bnx2x *bp)
  */
 bool bnx2x_reset_is_done(struct bnx2x *bp, int engine)
 {
-	u32 val	= REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
+	u32 val = REG_RD(bp, BNX2X_RECOVERY_GLOB_REG);
 	u32 bit = engine ?
 		BNX2X_PATH1_RST_IN_PROG_BIT : BNX2X_PATH0_RST_IN_PROG_BIT;
 
@@ -4260,13 +4254,18 @@ static bool bnx2x_get_load_status(struct bnx2x *bp, int engine)
 	return val != 0;
 }
 
+static void _print_parity(struct bnx2x *bp, u32 reg)
+{
+	pr_cont(" [0x%08x] ", REG_RD(bp, reg));
+}
+
 static void _print_next_block(int idx, const char *blk)
 {
 	pr_cont("%s%s", idx ? ", " : "", blk);
 }
 
-static int bnx2x_check_blocks_with_parity0(u32 sig, int par_num,
-					   bool print)
+static int bnx2x_check_blocks_with_parity0(struct bnx2x *bp, u32 sig,
+					    int par_num, bool print)
 {
 	int i = 0;
 	u32 cur_bit = 0;
@@ -4275,33 +4274,54 @@ static int bnx2x_check_blocks_with_parity0(u32 sig, int par_num,
 		if (sig & cur_bit) {
 			switch (cur_bit) {
 			case AEU_INPUTS_ATTN_BITS_BRB_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "BRB");
+					_print_parity(bp,
+						      BRB1_REG_BRB1_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_PARSER_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "PARSER");
+					_print_parity(bp, PRS_REG_PRS_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_TSDM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "TSDM");
+					_print_parity(bp,
+						      TSDM_REG_TSDM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_SEARCHER_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++,
 							  "SEARCHER");
+					_print_parity(bp, SRC_REG_SRC_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_TCM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "TCM");
+					_print_parity(bp,
+						      TCM_REG_TCM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_TSEMI_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "TSEMI");
+					_print_parity(bp,
+						      TSEM_REG_TSEM_PRTY_STS_0);
+					_print_parity(bp,
+						      TSEM_REG_TSEM_PRTY_STS_1);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_PBCLIENT_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "XPB");
+					_print_parity(bp, GRCBASE_XPB +
+							  PB_REG_PB_PRTY_STS);
+				}
 				break;
 			}
 
@@ -4313,8 +4333,9 @@ static int bnx2x_check_blocks_with_parity0(u32 sig, int par_num,
 	return par_num;
 }
 
-static int bnx2x_check_blocks_with_parity1(u32 sig, int par_num,
-					   bool *global, bool print)
+static int bnx2x_check_blocks_with_parity1(struct bnx2x *bp, u32 sig,
+					    int par_num, bool *global,
+					    bool print)
 {
 	int i = 0;
 	u32 cur_bit = 0;
@@ -4323,37 +4344,66 @@ static int bnx2x_check_blocks_with_parity1(u32 sig, int par_num,
 		if (sig & cur_bit) {
 			switch (cur_bit) {
 			case AEU_INPUTS_ATTN_BITS_PBF_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "PBF");
+					_print_parity(bp, PBF_REG_PBF_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_QM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "QM");
+					_print_parity(bp, QM_REG_QM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_TIMERS_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "TM");
+					_print_parity(bp, TM_REG_TM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_XSDM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "XSDM");
+					_print_parity(bp,
+						      XSDM_REG_XSDM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_XCM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "XCM");
+					_print_parity(bp, XCM_REG_XCM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_XSEMI_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "XSEMI");
+					_print_parity(bp,
+						      XSEM_REG_XSEM_PRTY_STS_0);
+					_print_parity(bp,
+						      XSEM_REG_XSEM_PRTY_STS_1);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_DOORBELLQ_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++,
 							  "DOORBELLQ");
+					_print_parity(bp,
+						      DORQ_REG_DORQ_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_NIG_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "NIG");
+					if (CHIP_IS_E1x(bp)) {
+						_print_parity(bp,
+							NIG_REG_NIG_PRTY_STS);
+					} else {
+						_print_parity(bp,
+							NIG_REG_NIG_PRTY_STS_0);
+						_print_parity(bp,
+							NIG_REG_NIG_PRTY_STS_1);
+					}
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_VAUX_PCI_CORE_PARITY_ERROR:
 				if (print)
@@ -4362,32 +4412,52 @@ static int bnx2x_check_blocks_with_parity1(u32 sig, int par_num,
 				*global = true;
 				break;
 			case AEU_INPUTS_ATTN_BITS_DEBUG_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "DEBUG");
+					_print_parity(bp, DBG_REG_DBG_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_USDM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "USDM");
+					_print_parity(bp,
+						      USDM_REG_USDM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_UCM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "UCM");
+					_print_parity(bp, UCM_REG_UCM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_USEMI_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "USEMI");
+					_print_parity(bp,
+						      USEM_REG_USEM_PRTY_STS_0);
+					_print_parity(bp,
+						      USEM_REG_USEM_PRTY_STS_1);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_UPB_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "UPB");
+					_print_parity(bp, GRCBASE_UPB +
+							  PB_REG_PB_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_CSDM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "CSDM");
+					_print_parity(bp,
+						      CSDM_REG_CSDM_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_CCM_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "CCM");
+					_print_parity(bp, CCM_REG_CCM_PRTY_STS);
+				}
 				break;
 			}
 
@@ -4399,8 +4469,8 @@ static int bnx2x_check_blocks_with_parity1(u32 sig, int par_num,
 	return par_num;
 }
 
-static int bnx2x_check_blocks_with_parity2(u32 sig, int par_num,
-					   bool print)
+static int bnx2x_check_blocks_with_parity2(struct bnx2x *bp, u32 sig,
+					    int par_num, bool print)
 {
 	int i = 0;
 	u32 cur_bit = 0;
@@ -4409,12 +4479,23 @@ static int bnx2x_check_blocks_with_parity2(u32 sig, int par_num,
 		if (sig & cur_bit) {
 			switch (cur_bit) {
 			case AEU_INPUTS_ATTN_BITS_CSEMI_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "CSEMI");
+					_print_parity(bp,
+						      CSEM_REG_CSEM_PRTY_STS_0);
+					_print_parity(bp,
+						      CSEM_REG_CSEM_PRTY_STS_1);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_PXP_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "PXP");
+					_print_parity(bp, PXP_REG_PXP_PRTY_STS);
+					_print_parity(bp,
+						      PXP2_REG_PXP2_PRTY_STS_0);
+					_print_parity(bp,
+						      PXP2_REG_PXP2_PRTY_STS_1);
+				}
 				break;
 			case AEU_IN_ATTN_BITS_PXPPCICLOCKCLIENT_PARITY_ERROR:
 				if (print)
@@ -4422,24 +4503,42 @@ static int bnx2x_check_blocks_with_parity2(u32 sig, int par_num,
 					"PXPPCICLOCKCLIENT");
 				break;
 			case AEU_INPUTS_ATTN_BITS_CFC_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "CFC");
+					_print_parity(bp,
+						      CFC_REG_CFC_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_CDU_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "CDU");
+					_print_parity(bp, CDU_REG_CDU_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_DMAE_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "DMAE");
+					_print_parity(bp,
+						      DMAE_REG_DMAE_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_IGU_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "IGU");
+					if (CHIP_IS_E1x(bp))
+						_print_parity(bp,
+							HC_REG_HC_PRTY_STS);
+					else
+						_print_parity(bp,
+							IGU_REG_IGU_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_MISC_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "MISC");
+					_print_parity(bp,
+						      MISC_REG_MISC_PRTY_STS);
+				}
 				break;
 			}
 
@@ -4493,8 +4592,8 @@ static int bnx2x_check_blocks_with_parity3(u32 sig, int par_num,
 	return par_num;
 }
 
-static int bnx2x_check_blocks_with_parity4(u32 sig, int par_num,
-					   bool print)
+static int bnx2x_check_blocks_with_parity4(struct bnx2x *bp, u32 sig,
+					    int par_num, bool print)
 {
 	int i = 0;
 	u32 cur_bit = 0;
@@ -4503,12 +4602,18 @@ static int bnx2x_check_blocks_with_parity4(u32 sig, int par_num,
 		if (sig & cur_bit) {
 			switch (cur_bit) {
 			case AEU_INPUTS_ATTN_BITS_PGLUE_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "PGLUE_B");
+					_print_parity(bp,
+						PGLUE_B_REG_PGLUE_B_PRTY_STS);
+				}
 				break;
 			case AEU_INPUTS_ATTN_BITS_ATC_PARITY_ERROR:
-				if (print)
+				if (print) {
 					_print_next_block(par_num++, "ATC");
+					_print_parity(bp,
+						      ATC_REG_ATC_PRTY_STS);
+				}
 				break;
 			}
 
@@ -4539,15 +4644,15 @@ static bool bnx2x_parity_attn(struct bnx2x *bp, bool *global, bool print,
 		if (print)
 			netdev_err(bp->dev,
 				   "Parity errors detected in blocks: ");
-		par_num = bnx2x_check_blocks_with_parity0(
+		par_num = bnx2x_check_blocks_with_parity0(bp,
 			sig[0] & HW_PRTY_ASSERT_SET_0, par_num, print);
-		par_num = bnx2x_check_blocks_with_parity1(
+		par_num = bnx2x_check_blocks_with_parity1(bp,
 			sig[1] & HW_PRTY_ASSERT_SET_1, par_num, global, print);
-		par_num = bnx2x_check_blocks_with_parity2(
+		par_num = bnx2x_check_blocks_with_parity2(bp,
 			sig[2] & HW_PRTY_ASSERT_SET_2, par_num, print);
 		par_num = bnx2x_check_blocks_with_parity3(
 			sig[3] & HW_PRTY_ASSERT_SET_3, par_num, global, print);
-		par_num = bnx2x_check_blocks_with_parity4(
+		par_num = bnx2x_check_blocks_with_parity4(bp,
 			sig[4] & HW_PRTY_ASSERT_SET_4, par_num, print);
 
 		if (print)
@@ -4591,7 +4696,6 @@ bool bnx2x_chk_parity_attn(struct bnx2x *bp, bool *global, bool print)
 	return bnx2x_parity_attn(bp, global, print, attn.sig);
 }
 
-
 static void bnx2x_attn_int_deasserted4(struct bnx2x *bp, u32 attn)
 {
 	u32 val;
@@ -4643,7 +4747,6 @@ static void bnx2x_attn_int_deasserted4(struct bnx2x *bp, u32 attn)
 		(u32)(attn & (AEU_INPUTS_ATTN_BITS_PGLUE_PARITY_ERROR |
 		    AEU_INPUTS_ATTN_BITS_ATC_PARITY_ERROR)));
 	}
-
 }
 
 static void bnx2x_attn_int_deasserted(struct bnx2x *bp, u32 deasserted)
@@ -4878,7 +4981,6 @@ static void bnx2x_handle_classification_eqe(struct bnx2x *bp,
 		BNX2X_ERR("Failed to schedule new commands: %d\n", rc);
 	else if (rc > 0)
 		DP(BNX2X_MSG_SP, "Scheduled next pending commands...\n");
-
 }
 
 static void bnx2x_set_iscsi_eth_rx_mode(struct bnx2x *bp, bool start);
@@ -5009,7 +5111,7 @@ static void bnx2x_eq_int(struct bnx2x *bp)
 	hw_cons = le16_to_cpu(*bp->eq_cons_sb);
 
 	/* The hw_cos range is 1-255, 257 - the sw_cons range is 0-254, 256.
-	 * when we get the the next-page we nned to adjust so the loop
+	 * when we get the next-page we need to adjust so the loop
 	 * condition below will be met. The next element is the size of a
 	 * regular element and hence incrementing by 1
 	 */
@@ -5075,8 +5177,6 @@ static void bnx2x_eq_int(struct bnx2x *bp)
 			if (q_obj->complete_cmd(bp, q_obj, BNX2X_Q_CMD_CFC_DEL))
 				break;
 
-
-
 			goto next_spqe;
 
 		case EVENT_RING_OPCODE_STOP_TRAFFIC:
@@ -5218,7 +5318,7 @@ static void bnx2x_sp_task(struct work_struct *work)
 
 	DP(BNX2X_MSG_SP, "sp task invoked\n");
 
-	/* make sure the atomic interupt_occurred has been written */
+	/* make sure the atomic interrupt_occurred has been written */
 	smp_rmb();
 	if (atomic_read(&bp->interrupt_occurred)) {
 
@@ -5265,7 +5365,6 @@ static void bnx2x_sp_task(struct work_struct *work)
 		/* ack status block only if something was actually handled */
 		bnx2x_ack_sb(bp, bp->igu_dsb_id, ATTENTION_ID,
 			     le16_to_cpu(bp->def_att_idx), IGU_INT_ENABLE, 1);
-
 	}
 
 	/* must be called after the EQ processing (since eq leads to sriov
@@ -5316,7 +5415,6 @@ irqreturn_t bnx2x_msix_sp_int(int irq, void *dev_instance)
 
 /* end of slow path */
 
-
 void bnx2x_drv_pulse(struct bnx2x *bp)
 {
 	SHMEM_WR(bp, func_mb[BP_FW_MB_IDX(bp)].drv_pulse_mb,
@@ -5360,7 +5458,7 @@ static void bnx2x_timer(unsigned long data)
 
 	/* sample pf vf bulletin board for new posts from pf */
 	if (IS_VF(bp))
-		bnx2x_sample_bulletin(bp);
+		bnx2x_timer_sriov(bp);
 
 	mod_timer(&bp->timer, jiffies + bp->current_interval);
 }
@@ -5382,7 +5480,6 @@ static void bnx2x_fill(struct bnx2x *bp, u32 addr, int fill, u32 len)
 	else
 		for (i = 0; i < len; i++)
 			REG_WR8(bp, addr + i, fill);
-
 }
 
 /* helper: writes FP SP data to FW - data_size in dwords */
@@ -5461,10 +5558,8 @@ static void bnx2x_zero_sp_sb(struct bnx2x *bp)
 	bnx2x_fill(bp, BAR_CSTRORM_INTMEM +
 			CSTORM_SP_SYNC_BLOCK_OFFSET(func), 0,
 			CSTORM_SP_SYNC_BLOCK_SIZE);
-
 }
 
-
 static void bnx2x_setup_ndsb_state_machine(struct hc_status_block_sm *hc_sm,
 					   int igu_sb_id, int igu_seg_id)
 {
@@ -5474,7 +5569,6 @@ static void bnx2x_setup_ndsb_state_machine(struct hc_status_block_sm *hc_sm,
 	hc_sm->time_to_expire = 0xFFFFFFFF;
 }
 
-
 /* allocates state machine ids. */
 static void bnx2x_map_sb_state_machines(struct hc_index_data *index_data)
 {
@@ -5700,7 +5794,7 @@ static void bnx2x_init_eq_ring(struct bnx2x *bp)
 	bp->eq_cons = 0;
 	bp->eq_prod = NUM_EQ_DESC;
 	bp->eq_cons_sb = BNX2X_EQ_INDEX;
-	/* we want a warning message before it gets rought... */
+	/* we want a warning message before it gets wrought... */
 	atomic_set(&bp->eq_spq_left,
 		min_t(int, MAX_SP_DESC_CNT - MAX_SPQ_PENDING, NUM_EQ_DESC) - 1);
 }
@@ -5784,7 +5878,7 @@ static int bnx2x_fill_accept_flags(struct bnx2x *bp, u32 rx_mode,
 
 		break;
 	case BNX2X_RX_MODE_PROMISC:
-		/* According to deffinition of SI mode, iface in promisc mode
+		/* According to definition of SI mode, iface in promisc mode
 		 * should receive matched and unmatched (in resolution of port)
 		 * unicast packets.
 		 */
@@ -5927,7 +6021,7 @@ static void bnx2x_init_eth_fp(struct bnx2x *bp, int fp_idx)
 	/* init shortcut */
 	fp->ustorm_rx_prods_offset = bnx2x_rx_ustorm_prods_offset(fp);
 
-	/* Setup SB indicies */
+	/* Setup SB indices */
 	fp->rx_cons_sb = BNX2X_RX_SB_INDEX;
 
 	/* Configure Queue State object */
@@ -5983,6 +6077,8 @@ static void bnx2x_init_tx_ring_one(struct bnx2x_fp_txdata *txdata)
 				    BCM_PAGE_SIZE*(i % NUM_TX_RINGS)));
 	}
 
+	*txdata->tx_cons_sb = cpu_to_le16(0);
+
 	SET_FLAG(txdata->tx_db.data.header.header, DOORBELL_HDR_DB_TYPE, 1);
 	txdata->tx_db.data.zero_fill1 = 0;
 	txdata->tx_db.data.prod = 0;
@@ -6001,6 +6097,7 @@ static void bnx2x_init_tx_rings_cnic(struct bnx2x *bp)
 	for_each_tx_queue_cnic(bp, i)
 		bnx2x_init_tx_ring_one(bp->fp[i].txdata_ptr[0]);
 }
+
 static void bnx2x_init_tx_rings(struct bnx2x *bp)
 {
 	int i;
@@ -6043,11 +6140,6 @@ void bnx2x_pre_irq_nic_init(struct bnx2x *bp)
 	bnx2x_init_rx_rings(bp);
 	bnx2x_init_tx_rings(bp);
 
-	if (IS_VF(bp)) {
-		bnx2x_memset_stats(bp);
-		return;
-	}
-
 	if (IS_PF(bp)) {
 		/* Initialize MOD_ABS interrupts */
 		bnx2x_init_mod_abs_int(bp, &bp->link_vars, bp->common.chip_id,
@@ -6058,6 +6150,8 @@ void bnx2x_pre_irq_nic_init(struct bnx2x *bp)
 		bnx2x_init_def_sb(bp);
 		bnx2x_update_dsb_idx(bp);
 		bnx2x_init_sp_ring(bp);
+	} else {
+		bnx2x_memset_stats(bp);
 	}
 }
 
@@ -6236,7 +6330,7 @@ static int bnx2x_int_mem_test(struct bnx2x *bp)
 		if (val == 0x10)
 			break;
 
-		msleep(10);
+		usleep_range(10000, 20000);
 		count--;
 	}
 	if (val != 0x10) {
@@ -6251,7 +6345,7 @@ static int bnx2x_int_mem_test(struct bnx2x *bp)
 		if (val == 1)
 			break;
 
-		msleep(10);
+		usleep_range(10000, 20000);
 		count--;
 	}
 	if (val != 0x1) {
@@ -6292,7 +6386,7 @@ static int bnx2x_int_mem_test(struct bnx2x *bp)
 		if (val == 0xb0)
 			break;
 
-		msleep(10);
+		usleep_range(10000, 20000);
 		count--;
 	}
 	if (val != 0xb0) {
@@ -6681,7 +6775,7 @@ static int bnx2x_init_hw_common(struct bnx2x *bp)
  *		    stay set)
  *		f.  If this is VNIC 3 of a port then also init
  *		    first_timers_ilt_entry to zero and last_timers_ilt_entry
- *		    to the last enrty in the ILT.
+ *		    to the last entry in the ILT.
  *
  *	Notes:
  *	Currently the PF error in the PGLC is non recoverable.
@@ -6772,7 +6866,6 @@ static int bnx2x_init_hw_common(struct bnx2x *bp)
 
 	bnx2x_init_block(bp, BLOCK_QM, PHASE_COMMON);
 
-
 	/* QM queues pointers table */
 	bnx2x_qm_init_ptr_table(bp, bp->qm_cid_count, INITOP_SET);
 
@@ -7013,7 +7106,6 @@ static int bnx2x_init_hw_port(struct bnx2x *bp)
 	u32 low, high;
 	u32 val;
 
-
 	DP(NETIF_MSG_HW, "starting port init  port %d\n", port);
 
 	REG_WR(bp, NIG_REG_MASK_INTERRUPT_PORT0 + port*4, 0);
@@ -7078,7 +7170,6 @@ static int bnx2x_init_hw_port(struct bnx2x *bp)
 			    BRB1_REG_MAC_GUARANTIED_1 :
 			    BRB1_REG_MAC_GUARANTIED_0), 40);
 
-
 	bnx2x_init_block(bp, BLOCK_PRS, init_phase);
 	if (CHIP_IS_E3B0(bp)) {
 		if (IS_MF_AFEX(bp)) {
@@ -7150,8 +7241,8 @@ static int bnx2x_init_hw_port(struct bnx2x *bp)
 
 	bnx2x_init_block(bp, BLOCK_MISC_AEU, init_phase);
 	/* init aeu_mask_attn_func_0/1:
-	 *  - SF mode: bits 3-7 are masked. only bits 0-2 are in use
-	 *  - MF mode: bit 3 is masked. bits 0-2 are in use as in SF
+	 *  - SF mode: bits 3-7 are masked. Only bits 0-2 are in use
+	 *  - MF mode: bit 3 is masked. Bits 0-2 are in use as in SF
 	 *             bits 4-7 are used for "per vn group attention" */
 	val = IS_MF(bp) ? 0xF7 : 0x7;
 	/* Enable DCBX attention for all but E1 */
@@ -7275,7 +7366,6 @@ void bnx2x_igu_clear_sb_gen(struct bnx2x *bp, u8 func, u8 idu_sb_id, bool is_pf)
 	while (!(REG_RD(bp, igu_addr_ack) & sb_bit) && --cnt)
 		msleep(20);
 
-
 	if (!(REG_RD(bp, igu_addr_ack) & sb_bit)) {
 		DP(NETIF_MSG_HW,
 		   "Unable to finish IGU cleanup: idu_sb_id %d offset %d bit %d (cnt %d)\n",
@@ -7295,7 +7385,6 @@ static void bnx2x_clear_func_ilt(struct bnx2x *bp, u32 func)
 		bnx2x_ilt_wr(bp, i, 0);
 }
 
-
 static void bnx2x_init_searcher(struct bnx2x *bp)
 {
 	int port = BP_PORT(bp);
@@ -7331,7 +7420,6 @@ static int bnx2x_reset_nic_mode(struct bnx2x *bp)
 	int rc, i, port = BP_PORT(bp);
 	int vlan_en = 0, mac_en[NUM_MACS];
 
-
 	/* Close input from network */
 	if (bp->mf_mode == SINGLE_FUNCTION) {
 		bnx2x_set_rx_filter(&bp->link_params, 0);
@@ -7406,7 +7494,7 @@ int bnx2x_init_hw_func_cnic(struct bnx2x *bp)
 	bnx2x_ilt_init_op_cnic(bp, INITOP_SET);
 
 	if (CONFIGURE_NIC_MODE(bp)) {
-		/* Configrue searcher as part of function hw init */
+		/* Configure searcher as part of function hw init */
 		bnx2x_init_searcher(bp);
 
 		/* Reset NIC mode */
@@ -7479,8 +7567,7 @@ static int bnx2x_init_hw_func(struct bnx2x *bp)
 	} else {
 		/* Set NIC mode */
 		REG_WR(bp, PRS_REG_NIC_MODE, 1);
-		DP(NETIF_MSG_IFUP, "NIC MODE configrued\n");
-
+		DP(NETIF_MSG_IFUP, "NIC MODE configured\n");
 	}
 
 	if (!CHIP_IS_E1x(bp)) {
@@ -7677,7 +7764,7 @@ static int bnx2x_init_hw_func(struct bnx2x *bp)
 			}
 			bnx2x_igu_clear_sb(bp, bp->igu_dsb_id);
 
-			/* !!! these should become driver const once
+			/* !!! These should become driver const once
 			   rf-tool supports split-68 const */
 			REG_WR(bp, IGU_REG_SB_INT_BEFORE_MASK_LSB, 0);
 			REG_WR(bp, IGU_REG_SB_INT_BEFORE_MASK_MSB, 0);
@@ -7734,7 +7821,6 @@ static int bnx2x_init_hw_func(struct bnx2x *bp)
 	return 0;
 }
 
-
 void bnx2x_free_mem_cnic(struct bnx2x *bp)
 {
 	bnx2x_ilt_mem_op_cnic(bp, ILT_MEMOP_FREE);
@@ -7779,7 +7865,6 @@ void bnx2x_free_mem(struct bnx2x *bp)
 	bnx2x_iov_free_mem(bp);
 }
 
-
 int bnx2x_alloc_mem_cnic(struct bnx2x *bp)
 {
 	if (!CHIP_IS_E1x(bp))
@@ -7793,7 +7878,7 @@ int bnx2x_alloc_mem_cnic(struct bnx2x *bp)
 				       host_hc_status_block_e1x));
 
 	if (CONFIGURE_NIC_MODE(bp) && !bp->t2)
-		/* allocate searcher T2 table, as it wan't allocated before */
+		/* allocate searcher T2 table, as it wasn't allocated before */
 		BNX2X_PCI_ALLOC(bp->t2, &bp->t2_mapping, SRC_T2_SZ);
 
 	/* write address to which L5 should insert its values */
@@ -8068,7 +8153,6 @@ void bnx2x_ilt_set_info(struct bnx2x *bp)
 		   ilt_client->page_size,
 		   ilt_client->flags,
 		   ilog2(ilt_client->page_size >> 12));
-
 	}
 
 	if (CNIC_SUPPORT(bp)) {
@@ -8124,7 +8208,6 @@ void bnx2x_ilt_set_info(struct bnx2x *bp)
 static void bnx2x_pf_q_prep_init(struct bnx2x *bp,
 	struct bnx2x_fastpath *fp, struct bnx2x_queue_init_params *init_params)
 {
-
 	u8 cos;
 	int cxt_index, cxt_offset;
 
@@ -8133,7 +8216,7 @@ static void bnx2x_pf_q_prep_init(struct bnx2x *bp,
 		__set_bit(BNX2X_Q_FLG_HC, &init_params->rx.flags);
 		__set_bit(BNX2X_Q_FLG_HC, &init_params->tx.flags);
 
-		/* If HC is supporterd, enable host coalescing in the transition
+		/* If HC is supported, enable host coalescing in the transition
 		 * to INIT state.
 		 */
 		__set_bit(BNX2X_Q_FLG_HC_EN, &init_params->rx.flags);
@@ -8205,7 +8288,6 @@ static int bnx2x_setup_tx_only(struct bnx2x *bp, struct bnx2x_fastpath *fp,
 	return bnx2x_queue_state_change(bp, q_params);
 }
 
-
 /**
  * bnx2x_setup_queue - setup queue
  *
@@ -8254,7 +8336,6 @@ int bnx2x_setup_queue(struct bnx2x *bp, struct bnx2x_fastpath *fp,
 
 	DP(NETIF_MSG_IFUP, "init complete\n");
 
-
 	/* Now move the Queue to the SETUP state... */
 	memset(setup_params, 0, sizeof(*setup_params));
 
@@ -8315,7 +8396,6 @@ static int bnx2x_stop_queue(struct bnx2x *bp, int index)
 	/* We want to wait for completion in this context */
 	__set_bit(RAMROD_COMP_WAIT, &q_params.ramrod_flags);
 
-
 	/* close tx-only connections */
 	for (tx_index = FIRST_TX_ONLY_COS_INDEX;
 	     tx_index < fp->max_cos;
@@ -8369,7 +8449,6 @@ static int bnx2x_stop_queue(struct bnx2x *bp, int index)
 	return bnx2x_queue_state_change(bp, &q_params);
 }
 
-
 static void bnx2x_reset_func(struct bnx2x *bp)
 {
 	int port = BP_PORT(bp);
@@ -8422,7 +8501,7 @@ static void bnx2x_reset_func(struct bnx2x *bp)
 		 * scan to complete
 		 */
 		for (i = 0; i < 200; i++) {
-			msleep(10);
+			usleep_range(10000, 20000);
 			if (!REG_RD(bp, TM_REG_LIN0_SCAN_ON + port*4))
 				break;
 		}
@@ -8623,14 +8702,14 @@ static int bnx2x_func_wait_started(struct bnx2x *bp)
 
 	/*
 	 * (assumption: No Attention from MCP at this stage)
-	 * PMF probably in the middle of TXdisable/enable transaction
+	 * PMF probably in the middle of TX disable/enable transaction
 	 * 1. Sync IRS for default SB
-	 * 2. Sync SP queue - this guarantes us that attention handling started
-	 * 3. Wait, that TXdisable/enable transaction completes
+	 * 2. Sync SP queue - this guarantees us that attention handling started
+	 * 3. Wait, that TX disable/enable transaction completes
 	 *
-	 * 1+2 guranty that if DCBx attention was scheduled it already changed
-	 * pending bit of transaction from STARTED-->TX_STOPPED, if we alredy
-	 * received complettion for the transaction the state is TX_STOPPED.
+	 * 1+2 guarantee that if DCBx attention was scheduled it already changed
+	 * pending bit of transaction from STARTED-->TX_STOPPED, if we already
+	 * received completion for the transaction the state is TX_STOPPED.
 	 * State will return to STARTED after completion of TX_STOPPED-->STARTED
 	 * transaction.
 	 */
@@ -8660,7 +8739,7 @@ static int bnx2x_func_wait_started(struct bnx2x *bp)
 		struct bnx2x_func_state_params func_params = {NULL};
 
 		DP(NETIF_MSG_IFDOWN,
-		   "Hmmm... unexpected function state! Forcing STARTED-->TX_ST0PPED-->STARTED\n");
+		   "Hmmm... Unexpected function state! Forcing STARTED-->TX_ST0PPED-->STARTED\n");
 
 		func_params.f_obj = &bp->func_obj;
 		__set_bit(RAMROD_DRV_CLR_ONLY,
@@ -8740,7 +8819,6 @@ void bnx2x_chip_cleanup(struct bnx2x *bp, int unload_mode, bool keep_link)
 
 	bnx2x_iov_chip_cleanup(bp);
 
-
 	/*
 	 * Send the UNLOAD_REQUEST to the MCP. This will return if
 	 * this function should perform FUNC, PORT or COMMON HW
@@ -8750,7 +8828,7 @@ void bnx2x_chip_cleanup(struct bnx2x *bp, int unload_mode, bool keep_link)
 
 	/*
 	 * (assumption: No Attention from MCP at this stage)
-	 * PMF probably in the middle of TXdisable/enable transaction
+	 * PMF probably in the middle of TX disable/enable transaction
 	 */
 	rc = bnx2x_func_wait_started(bp);
 	if (rc) {
@@ -8813,7 +8891,6 @@ unload_error:
 	if (rc)
 		BNX2X_ERR("HW_RESET failed\n");
 
-
 	/* Report UNLOAD_DONE to MCP */
 	bnx2x_send_unload_done(bp, keep_link);
 }
@@ -9179,7 +9256,6 @@ static int bnx2x_process_kill(struct bnx2x *bp, bool global)
 	if (!CHIP_IS_E1x(bp) && bnx2x_er_poll_igu_vq(bp))
 		return -EAGAIN;
 
-
 	/* TBD: Indicate that "process kill" is in progress to MCP */
 
 	/* Clear "unprepared" bit */
@@ -9367,7 +9443,7 @@ static void bnx2x_parity_recover(struct bnx2x *bp)
 				 * the first leader that performs a
 				 * leader_reset() reset the global blocks in
 				 * order to clear global attentions. Otherwise
-				 * the the gates will remain closed for that
+				 * the gates will remain closed for that
 				 * engine.
 				 */
 				if (load_status ||
@@ -9480,14 +9556,12 @@ static void bnx2x_sp_rtnl_task(struct work_struct *work)
 		return;
 	}
 
-	/* if stop on error is defined no recovery flows should be executed */
+	if (unlikely(bp->recovery_state != BNX2X_RECOVERY_DONE)) {
 #ifdef BNX2X_STOP_ON_ERROR
-	BNX2X_ERR("recovery flow called but STOP_ON_ERROR defined so reset not done to allow debug dump,\n"
-		  "you will need to reboot when done\n");
-	goto sp_rtnl_not_reset;
+		BNX2X_ERR("recovery flow called but STOP_ON_ERROR defined so reset not done to allow debug dump,\n"
+			  "you will need to reboot when done\n");
+		goto sp_rtnl_not_reset;
 #endif
-
-	if (unlikely(bp->recovery_state != BNX2X_RECOVERY_DONE)) {
 		/*
 		 * Clear all pending SP commands as we are going to reset the
 		 * function anyway.
@@ -9502,6 +9576,12 @@ static void bnx2x_sp_rtnl_task(struct work_struct *work)
 	}
 
 	if (test_and_clear_bit(BNX2X_SP_RTNL_TX_TIMEOUT, &bp->sp_rtnl_state)) {
+#ifdef BNX2X_STOP_ON_ERROR
+		BNX2X_ERR("recovery flow called but STOP_ON_ERROR defined so reset not done to allow debug dump,\n"
+			  "you will need to reboot when done\n");
+		goto sp_rtnl_not_reset;
+#endif
+
 		/*
 		 * Clear all pending SP commands as we are going to reset the
 		 * function anyway.
@@ -9540,6 +9620,13 @@ sp_rtnl_not_reset:
 		   "sending set mcast vf pf channel message from rtnl sp-task\n");
 		bnx2x_vfpf_set_mcast(bp->dev);
 	}
+	if (test_and_clear_bit(BNX2X_SP_RTNL_VFPF_CHANNEL_DOWN,
+			       &bp->sp_rtnl_state)){
+		if (!test_bit(__LINK_STATE_NOCARRIER, &bp->dev->state)) {
+			bnx2x_tx_disable(bp);
+			BNX2X_ERR("PF indicated channel is not servicable anymore. This means this VF device is no longer operational\n");
+		}
+	}
 
 	if (test_and_clear_bit(BNX2X_SP_RTNL_VFPF_STORM_RX_MODE,
 			       &bp->sp_rtnl_state)) {
@@ -9647,7 +9734,6 @@ static void bnx2x_prev_unload_close_mac(struct bnx2x *bp,
 			wb_data[0] &= ~BMAC_CONTROL_RX_ENABLE;
 			REG_WR(bp, vals->bmac_addr, wb_data[0]);
 			REG_WR(bp, vals->bmac_addr + 0x4, wb_data[1]);
-
 		}
 		BNX2X_DEV_INFO("Disable emac Rx\n");
 		vals->emac_addr = NIG_REG_NIG_EMAC0_EN + BP_PORT(bp)*4;
@@ -9681,7 +9767,6 @@ static void bnx2x_prev_unload_close_mac(struct bnx2x *bp,
 
 	if (mac_stopped)
 		msleep(20);
-
 }
 
 #define BNX2X_PREV_UNDI_PROD_ADDR(p) (BAR_TSTRORM_INTMEM + 0x1508 + ((p) << 4))
@@ -9780,6 +9865,21 @@ static bool bnx2x_prev_is_path_marked(struct bnx2x *bp)
 	return rc;
 }
 
+bool bnx2x_port_after_undi(struct bnx2x *bp)
+{
+	struct bnx2x_prev_path_list *entry;
+	bool val;
+
+	down(&bnx2x_prev_sem);
+
+	entry = bnx2x_prev_path_get_entry(bp);
+	val = !!(entry && (entry->undi & (1 << BP_PORT(bp))));
+
+	up(&bnx2x_prev_sem);
+
+	return val;
+}
+
 static int bnx2x_prev_mark_path(struct bnx2x *bp, bool after_undi)
 {
 	struct bnx2x_prev_path_list *tmp_list;
@@ -9839,7 +9939,6 @@ static int bnx2x_do_flr(struct bnx2x *bp)
 	u16 status;
 	struct pci_dev *dev = bp->pdev;
 
-
 	if (CHIP_IS_E1x(bp)) {
 		BNX2X_DEV_INFO("FLR not supported in E1/E1H\n");
 		return -EINVAL;
@@ -9986,7 +10085,6 @@ static int bnx2x_prev_unload_common(struct bnx2x *bp)
 
 		if (!timer_count)
 			BNX2X_ERR("Failed to empty BRB, hope for the best\n");
-
 	}
 
 	/* No packets are in the pipeline, path is ready for reset */
@@ -10036,7 +10134,6 @@ static int bnx2x_prev_unload(struct bnx2x *bp)
 {
 	int time_counter = 10;
 	u32 rc, fw, hw_lock_reg, hw_lock_val;
-	struct bnx2x_prev_path_list *prev_list;
 	BNX2X_DEV_INFO("Entering Previous Unload Flow\n");
 
 	/* clear hw from errors which may have resulted from an interrupted
@@ -10049,7 +10146,7 @@ static int bnx2x_prev_unload(struct bnx2x *bp)
 		      (MISC_REG_DRIVER_CONTROL_1 + BP_FUNC(bp) * 8) :
 		      (MISC_REG_DRIVER_CONTROL_7 + (BP_FUNC(bp) - 6) * 8);
 
-	hw_lock_val = (REG_RD(bp, hw_lock_reg));
+	hw_lock_val = REG_RD(bp, hw_lock_reg);
 	if (hw_lock_val) {
 		if (hw_lock_val & HW_LOCK_RESOURCE_NVRAM) {
 			BNX2X_DEV_INFO("Release Previously held NVRAM lock\n");
@@ -10064,7 +10161,7 @@ static int bnx2x_prev_unload(struct bnx2x *bp)
 
 	if (MCPR_ACCESS_LOCK_LOCK & REG_RD(bp, MCP_REG_MCPR_ACCESS_LOCK)) {
 		BNX2X_DEV_INFO("Release previously held alr\n");
-		REG_WR(bp, MCP_REG_MCPR_ACCESS_LOCK, 0);
+		bnx2x_release_alr(bp);
 	}
 
 	do {
@@ -10093,7 +10190,7 @@ static int bnx2x_prev_unload(struct bnx2x *bp)
 			break;
 		}
 
-		/* non-common reply from MCP night require looping */
+		/* non-common reply from MCP might require looping */
 		rc = bnx2x_prev_unload_uncommon(bp);
 		if (rc != BNX2X_PREV_WAIT_NEEDED)
 			break;
@@ -10107,8 +10204,7 @@ static int bnx2x_prev_unload(struct bnx2x *bp)
 	}
 
 	/* Mark function if its port was used to boot from SAN */
-	prev_list = bnx2x_prev_path_get_entry(bp);
-	if (prev_list && (prev_list->undi & (1 << BP_PORT(bp))))
+	if (bnx2x_port_after_undi(bp))
 		bp->link_params.feature_config_flags |=
 			FEATURE_CONFIG_BOOT_FROM_SAN;
 
@@ -10192,8 +10288,6 @@ static void bnx2x_get_common_hwinfo(struct bnx2x *bp)
 
 	bnx2x_init_shmem(bp);
 
-
-
 	bp->common.shmem2_base = REG_RD(bp, (BP_PATH(bp) ?
 					MISC_REG_GENERIC_CR_1 :
 					MISC_REG_GENERIC_CR_0));
@@ -10455,6 +10549,9 @@ static void bnx2x_link_settings_supported(struct bnx2x *bp, u32 switch_cfg)
 					PORT_HW_CFG_SPEED_CAPABILITY_D0_10G))
 			bp->port.supported[idx] &= ~SUPPORTED_10000baseT_Full;
 
+		if (!(bp->link_params.speed_cap_mask[idx] &
+					PORT_HW_CFG_SPEED_CAPABILITY_D0_20G))
+			bp->port.supported[idx] &= ~SUPPORTED_20000baseKR2_Full;
 	}
 
 	BNX2X_DEV_INFO("supported 0x%x 0x%x\n", bp->port.supported[0],
@@ -10765,7 +10862,6 @@ void bnx2x_get_iscsi_info(struct bnx2x *bp)
 	 */
 	if (!bp->cnic_eth_dev.max_iscsi_conn)
 		bp->flags |= no_flags;
-
 }
 
 static void bnx2x_get_ext_wwn_info(struct bnx2x *bp, int func)
@@ -10782,12 +10878,56 @@ static void bnx2x_get_ext_wwn_info(struct bnx2x *bp, int func)
 	bp->cnic_eth_dev.fcoe_wwn_node_name_lo =
 		MF_CFG_RD(bp, func_ext_config[func].fcoe_wwn_node_name_lower);
 }
+
+static int bnx2x_shared_fcoe_funcs(struct bnx2x *bp)
+{
+	u8 count = 0;
+
+	if (IS_MF(bp)) {
+		u8 fid;
+
+		/* iterate over absolute function ids for this path: */
+		for (fid = BP_PATH(bp); fid < E2_FUNC_MAX * 2; fid += 2) {
+			if (IS_MF_SD(bp)) {
+				u32 cfg = MF_CFG_RD(bp,
+						    func_mf_config[fid].config);
+
+				if (!(cfg & FUNC_MF_CFG_FUNC_HIDE) &&
+				    ((cfg & FUNC_MF_CFG_PROTOCOL_MASK) ==
+					    FUNC_MF_CFG_PROTOCOL_FCOE))
+					count++;
+			} else {
+				u32 cfg = MF_CFG_RD(bp,
+						    func_ext_config[fid].
+								      func_cfg);
+
+				if ((cfg & MACP_FUNC_CFG_FLAGS_ENABLED) &&
+				    (cfg & MACP_FUNC_CFG_FLAGS_FCOE_OFFLOAD))
+					count++;
+			}
+		}
+	} else { /* SF */
+		int port, port_cnt = CHIP_MODE_IS_4_PORT(bp) ? 2 : 1;
+
+		for (port = 0; port < port_cnt; port++) {
+			u32 lic = SHMEM_RD(bp,
+					   drv_lic_key[port].max_fcoe_conn) ^
+				  FW_ENCODE_32BIT_PATTERN;
+			if (lic)
+				count++;
+		}
+	}
+
+	return count;
+}
+
 static void bnx2x_get_fcoe_info(struct bnx2x *bp)
 {
 	int port = BP_PORT(bp);
 	int func = BP_ABS_FUNC(bp);
 	u32 max_fcoe_conn = FW_ENCODE_32BIT_PATTERN ^ SHMEM_RD(bp,
 				drv_lic_key[port].max_fcoe_conn);
+	u8 num_fcoe_func = bnx2x_shared_fcoe_funcs(bp);
 
 	if (!CNIC_SUPPORT(bp)) {
 		bp->flags |= NO_FCOE_FLAG;
@@ -10801,9 +10941,10 @@ static void bnx2x_get_fcoe_info(struct bnx2x *bp)
 
 	/* Calculate the number of maximum allowed FCoE tasks */
 	bp->cnic_eth_dev.max_fcoe_exchanges = MAX_NUM_FCOE_TASKS_PER_ENGINE;
-	if (IS_MF(bp) || CHIP_MODE_IS_4_PORT(bp))
-		bp->cnic_eth_dev.max_fcoe_exchanges /=
-						MAX_FCOE_FUNCS_PER_ENGINE;
+
+	/* check if FCoE resources must be shared between different functions */
+	if (num_fcoe_func)
+		bp->cnic_eth_dev.max_fcoe_exchanges /= num_fcoe_func;
 
 	/* Read the WWN: */
 	if (!IS_MF(bp)) {
@@ -11031,7 +11172,7 @@ static int bnx2x_get_hwinfo(struct bnx2x *bp)
 	} else {
 		bp->common.int_block = INT_BLOCK_IGU;
 
-		/* do not allow device reset during IGU info preocessing */
+		/* do not allow device reset during IGU info processing */
 		bnx2x_acquire_hw_lock(bp, HW_LOCK_RESOURCE_RESET);
 
 		val = REG_RD(bp, IGU_REG_BLOCK_CONFIGURATION);
@@ -11110,7 +11251,7 @@ static int bnx2x_get_hwinfo(struct bnx2x *bp)
 				E1H_FUNC_MAX * sizeof(struct drv_func_mb);
 		/*
 		 * get mf configuration:
-		 * 1. existence of MF configuration
+		 * 1. Existence of MF configuration
 		 * 2. MAC address must be legal (check only upper bytes)
 		 *    for  Switch-Independent mode;
 		 *    OVLAN must be legal for Switch-Dependent mode
@@ -11384,7 +11525,6 @@ static int bnx2x_init_bp(struct bnx2x *bp)
 	mutex_init(&bp->fw_mb_mutex);
 	spin_lock_init(&bp->stats_lock);
 
-
 	INIT_DELAYED_WORK(&bp->sp_task, bnx2x_sp_task);
 	INIT_DELAYED_WORK(&bp->sp_rtnl_task, bnx2x_sp_rtnl_task);
 	INIT_DELAYED_WORK(&bp->period_task, bnx2x_period_task);
@@ -11393,7 +11533,7 @@ static int bnx2x_init_bp(struct bnx2x *bp)
 		if (rc)
 			return rc;
 	} else {
-		random_ether_addr(bp->dev->dev_addr);
+		eth_zero_addr(bp->dev->dev_addr);
 	}
 
 	bnx2x_set_modes_bitmap(bp);
@@ -11417,7 +11557,6 @@ static int bnx2x_init_bp(struct bnx2x *bp)
 		bnx2x_prev_unload(bp);
 	}
 
-
 	if (CHIP_REV_IS_FPGA(bp))
 		dev_err(&bp->pdev->dev, "FPGA detected\n");
 
@@ -11489,7 +11628,7 @@ static int bnx2x_init_bp(struct bnx2x *bp)
 
 	/* We need at least one default status block for slow-path events,
 	 * second status block for the L2 queue, and a third status block for
-	 * CNIC if supproted.
+	 * CNIC if supported.
 	 */
 	if (CNIC_SUPPORT(bp))
 		bp->min_msix_vec_cnt = 3;
@@ -11497,10 +11636,11 @@ static int bnx2x_init_bp(struct bnx2x *bp)
 		bp->min_msix_vec_cnt = 2;
 	BNX2X_DEV_INFO("bp->min_msix_vec_cnt %d", bp->min_msix_vec_cnt);
 
+	bp->dump_preset_idx = 1;
+
 	return rc;
 }
 
-
 /****************************************************************************
 * General service functions
 ****************************************************************************/
@@ -11585,9 +11725,6 @@ static int bnx2x_close(struct net_device *dev)
 	/* Unload the driver, release IRQs */
 	bnx2x_nic_unload(bp, UNLOAD_CLOSE, false);
 
-	/* Power off */
-	bnx2x_set_power_state(bp, PCI_D3hot);
-
 	return 0;
 }
 
@@ -11852,6 +11989,10 @@ static int bnx2x_validate_addr(struct net_device *dev)
 {
 	struct bnx2x *bp = netdev_priv(dev);
 
+	/* query the bulletin board for mac address configured by the PF */
+	if (IS_VF(bp))
+		bnx2x_sample_bulletin(bp);
+
 	if (!bnx2x_is_valid_ether_addr(bp, dev->dev_addr)) {
 		BNX2X_ERR("Non-valid Ethernet address\n");
 		return -EADDRNOTAVAIL;
@@ -11878,12 +12019,16 @@ static const struct net_device_ops bnx2x_netdev_ops = {
 	.ndo_setup_tc		= bnx2x_setup_tc,
 #ifdef CONFIG_BNX2X_SRIOV
 	.ndo_set_vf_mac		= bnx2x_set_vf_mac,
-	.ndo_set_vf_vlan        = bnx2x_set_vf_vlan,
+	.ndo_set_vf_vlan	= bnx2x_set_vf_vlan,
 	.ndo_get_vf_config	= bnx2x_get_vf_config,
 #endif
 #ifdef NETDEV_FCOE_WWNN
 	.ndo_fcoe_get_wwn	= bnx2x_fcoe_get_wwn,
 #endif
+
+#ifdef CONFIG_NET_LL_RX_POLL
+	.ndo_ll_poll		= bnx2x_low_latency_recv,
+#endif
 };
 
 static int bnx2x_set_coherency_mask(struct bnx2x *bp)
@@ -11959,7 +12104,7 @@ static int bnx2x_init_dev(struct bnx2x *bp, struct pci_dev *pdev,
 	}
 
 	if (IS_PF(bp)) {
-		bp->pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM);
+		bp->pm_cap = pdev->pm_cap;
 		if (bp->pm_cap == 0) {
 			dev_err(&bp->pdev->dev,
 				"Cannot find power management capability, aborting\n");
@@ -12008,8 +12153,6 @@ static int bnx2x_init_dev(struct bnx2x *bp, struct pci_dev *pdev,
 	}
 	BNX2X_DEV_INFO("me reg PF num: %d\n", bp->pf_num);
 
-	bnx2x_set_power_state(bp, PCI_D0);
-
 	/* clean indirect addresses */
 	pci_write_config_dword(bp->pdev, PCICFG_GRC_ADDRESS,
 			       PCICFG_VENDOR_ID_OFFSET);
@@ -12094,15 +12237,26 @@ err_out:
 	return rc;
 }
 
-static void bnx2x_get_pcie_width_speed(struct bnx2x *bp, int *width, int *speed)
+static void bnx2x_get_pcie_width_speed(struct bnx2x *bp, int *width,
+				       enum bnx2x_pci_bus_speed *speed)
 {
-	u32 val = 0;
+	u32 link_speed, val = 0;
 
 	pci_read_config_dword(bp->pdev, PCICFG_LINK_CONTROL, &val);
 	*width = (val & PCICFG_LINK_WIDTH) >> PCICFG_LINK_WIDTH_SHIFT;
 
-	/* return value of 1=2.5GHz 2=5GHz */
-	*speed = (val & PCICFG_LINK_SPEED) >> PCICFG_LINK_SPEED_SHIFT;
+	link_speed = (val & PCICFG_LINK_SPEED) >> PCICFG_LINK_SPEED_SHIFT;
+
+	switch (link_speed) {
+	case 3:
+		*speed = BNX2X_PCI_LINK_SPEED_8000;
+		break;
+	case 2:
+		*speed = BNX2X_PCI_LINK_SPEED_5000;
+		break;
+	default:
+		*speed = BNX2X_PCI_LINK_SPEED_2500;
+	}
 }
 
 static int bnx2x_check_firmware(struct bnx2x *bp)
@@ -12327,7 +12481,6 @@ static void bnx2x_release_firmware(struct bnx2x *bp)
 	bp->firmware = NULL;
 }
 
-
 static struct bnx2x_func_sp_drv_ops bnx2x_func_sp_drv = {
 	.init_hw_cmn_chip = bnx2x_init_hw_common_chip,
 	.init_hw_cmn      = bnx2x_init_hw_common,
@@ -12465,7 +12618,8 @@ static int bnx2x_init_one(struct pci_dev *pdev,
 {
 	struct net_device *dev = NULL;
 	struct bnx2x *bp;
-	int pcie_width, pcie_speed;
+	int pcie_width;
+	enum bnx2x_pci_bus_speed pcie_speed;
 	int rc, max_non_def_sbs;
 	int rx_count, tx_count, rss_count, doorbell_size;
 	int max_cos_est;
@@ -12605,7 +12759,6 @@ static int bnx2x_init_one(struct pci_dev *pdev,
 	}
 	BNX2X_DEV_INFO("device name after netdev register %s\n", dev->name);
 
-
 	if (!NO_FCOE(bp)) {
 		/* Add storage MAC address */
 		rtnl_lock();
@@ -12617,15 +12770,15 @@ static int bnx2x_init_one(struct pci_dev *pdev,
 	BNX2X_DEV_INFO("got pcie width %d and speed %d\n",
 		       pcie_width, pcie_speed);
 
-	BNX2X_DEV_INFO(
-		"%s (%c%d) PCI-E x%d %s found at mem %lx, IRQ %d, node addr %pM\n",
-		    board_info[ent->driver_data].name,
-		    (CHIP_REV(bp) >> 12) + 'A', (CHIP_METAL(bp) >> 4),
-		    pcie_width,
-		    ((!CHIP_IS_E2(bp) && pcie_speed == 2) ||
-		     (CHIP_IS_E2(bp) && pcie_speed == 1)) ?
-		    "5GHz (Gen2)" : "2.5GHz",
-		    dev->base_addr, bp->pdev->irq, dev->dev_addr);
+	BNX2X_DEV_INFO("%s (%c%d) PCI-E x%d %s found at mem %lx, IRQ %d, node addr %pM\n",
+		       board_info[ent->driver_data].name,
+		       (CHIP_REV(bp) >> 12) + 'A', (CHIP_METAL(bp) >> 4),
+		       pcie_width,
+		       pcie_speed == BNX2X_PCI_LINK_SPEED_2500 ? "2.5GHz" :
+		       pcie_speed == BNX2X_PCI_LINK_SPEED_5000 ? "5.0GHz" :
+		       pcie_speed == BNX2X_PCI_LINK_SPEED_8000 ? "8.0GHz" :
+		       "Unknown",
+		       dev->base_addr, bp->pdev->irq, dev->dev_addr);
 
 	return 0;
 
@@ -12647,17 +12800,11 @@ init_one_exit:
 	return rc;
 }
 
-static void bnx2x_remove_one(struct pci_dev *pdev)
+static void __bnx2x_remove(struct pci_dev *pdev,
+			   struct net_device *dev,
+			   struct bnx2x *bp,
+			   bool remove_netdev)
 {
-	struct net_device *dev = pci_get_drvdata(pdev);
-	struct bnx2x *bp;
-
-	if (!dev) {
-		dev_err(&pdev->dev, "BAD net device from bnx2x_init_one\n");
-		return;
-	}
-	bp = netdev_priv(dev);
-
 	/* Delete storage MAC address */
 	if (!NO_FCOE(bp)) {
 		rtnl_lock();
@@ -12670,7 +12817,17 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
 	bnx2x_dcbnl_update_applist(bp, true);
 #endif
 
-	unregister_netdev(dev);
+	/* Close the interface - either directly or implicitly */
+	if (remove_netdev) {
+		unregister_netdev(dev);
+	} else {
+		rtnl_lock();
+		if (netif_running(dev))
+			bnx2x_close(dev);
+		rtnl_unlock();
+	}
+
+	bnx2x_iov_remove_one(bp);
 
 	/* Power on: we can't let PCI layer write to us while we are in D3 */
 	if (IS_PF(bp))
@@ -12686,12 +12843,16 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
 	/* Make sure RESET task is not scheduled before continuing */
 	cancel_delayed_work_sync(&bp->sp_rtnl_task);
 
-	bnx2x_iov_remove_one(bp);
-
 	/* send message via vfpf channel to release the resources of this vf */
 	if (IS_VF(bp))
 		bnx2x_vfpf_release(bp);
 
+	/* Assumes no further PCIe PM changes will occur */
+	if (system_state == SYSTEM_POWER_OFF) {
+		pci_wake_from_d3(pdev, bp->wol);
+		pci_set_power_state(pdev, PCI_D3hot);
+	}
+
 	if (bp->regview)
 		iounmap(bp->regview);
 
@@ -12706,7 +12867,8 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
 	}
 	bnx2x_free_mem_bp(bp);
 
-	free_netdev(dev);
+	if (remove_netdev)
+		free_netdev(dev);
 
 	if (atomic_read(&pdev->enable_cnt) == 1)
 		pci_release_regions(pdev);
@@ -12715,6 +12877,20 @@ static void bnx2x_remove_one(struct pci_dev *pdev)
 	pci_set_drvdata(pdev, NULL);
 }
 
+static void bnx2x_remove_one(struct pci_dev *pdev)
+{
+	struct net_device *dev = pci_get_drvdata(pdev);
+	struct bnx2x *bp;
+
+	if (!dev) {
+		dev_err(&pdev->dev, "BAD net device from bnx2x_init_one\n");
+		return;
+	}
+	bp = netdev_priv(dev);
+
+	__bnx2x_remove(pdev, dev, bp, true);
+}
+
 static int bnx2x_eeh_nic_unload(struct bnx2x *bp)
 {
 	bp->state = BNX2X_STATE_CLOSING_WAIT4_HALT;
@@ -12747,19 +12923,6 @@ static int bnx2x_eeh_nic_unload(struct bnx2x *bp)
 	return 0;
 }
 
-static void bnx2x_eeh_recover(struct bnx2x *bp)
-{
-	u32 val;
-
-	mutex_init(&bp->port.phy_mutex);
-
-
-	val = SHMEM_RD(bp, validity_map[BP_PORT(bp)]);
-	if ((val & (SHR_MEM_VALIDITY_DEV_INFO | SHR_MEM_VALIDITY_MB))
-		!= (SHR_MEM_VALIDITY_DEV_INFO | SHR_MEM_VALIDITY_MB))
-		BNX2X_ERR("BAD MCP validity signature\n");
-}
-
 /**
  * bnx2x_io_error_detected - called when PCI error is detected
  * @pdev: Pointer to PCI device
@@ -12828,6 +12991,10 @@ static pci_ers_result_t bnx2x_io_slot_reset(struct pci_dev *pdev)
 
 	if (netif_running(dev)) {
 		BNX2X_ERR("IO slot reset --> driver unload\n");
+
+		/* MCP should have been reset; Need to wait for validity */
+		bnx2x_init_shmem(bp);
+
 		if (IS_PF(bp) && SHMEM2_HAS(bp, drv_capabilities_flag)) {
 			u32 v;
 
@@ -12849,7 +13016,7 @@ static pci_ers_result_t bnx2x_io_slot_reset(struct pci_dev *pdev)
 
 		bnx2x_prev_unload(bp);
 
-		/* We should have resetted the engine, so It's fair to
+		/* We should have reseted the engine, so It's fair to
 		 * assume the FW will no longer write to the bnx2x driver.
 		 */
 		bnx2x_squeeze_objects(bp);
@@ -12886,8 +13053,6 @@ static void bnx2x_io_resume(struct pci_dev *pdev)
 
 	rtnl_lock();
 
-	bnx2x_eeh_recover(bp);
-
 	bp->fw_seq = SHMEM_RD(bp, func_mb[BP_FW_MB_IDX(bp)].drv_mb_header) &
 							DRV_MSG_SEQ_NUMBER_MASK;
 
@@ -12905,6 +13070,29 @@ static const struct pci_error_handlers bnx2x_err_handler = {
 	.resume         = bnx2x_io_resume,
 };
 
+static void bnx2x_shutdown(struct pci_dev *pdev)
+{
+	struct net_device *dev = pci_get_drvdata(pdev);
+	struct bnx2x *bp;
+
+	if (!dev)
+		return;
+
+	bp = netdev_priv(dev);
+	if (!bp)
+		return;
+
+	rtnl_lock();
+	netif_device_detach(dev);
+	rtnl_unlock();
+
+	/* Don't remove the netdevice, as there are scenarios which will cause
+	 * the kernel to hang, e.g., when trying to remove bnx2i while the
+	 * rootfs is mounted from SAN.
+	 */
+	__bnx2x_remove(pdev, dev, bp, false);
+}
+
 static struct pci_driver bnx2x_pci_driver = {
 	.name        = DRV_MODULE_NAME,
 	.id_table    = bnx2x_pci_tbl,
@@ -12916,6 +13104,7 @@ static struct pci_driver bnx2x_pci_driver = {
 #ifdef CONFIG_BNX2X_SRIOV
 	.sriov_configure = bnx2x_sriov_configure,
 #endif
+	.shutdown    = bnx2x_shutdown,
 };
 
 static int __init bnx2x_init(void)
@@ -12941,11 +13130,12 @@ static int __init bnx2x_init(void)
 static void __exit bnx2x_cleanup(void)
 {
 	struct list_head *pos, *q;
+
 	pci_unregister_driver(&bnx2x_pci_driver);
 
 	destroy_workqueue(bnx2x_wq);
 
-	/* Free globablly allocated resources */
+	/* Free globally allocated resources */
 	list_for_each_safe(pos, q, &bnx2x_prev_list) {
 		struct bnx2x_prev_path_list *tmp =
 			list_entry(pos, struct bnx2x_prev_path_list, list);
@@ -12968,7 +13158,7 @@ module_exit(bnx2x_cleanup);
  * @bp:		driver handle
  * @set:	set or clear the CAM entry
  *
- * This function will wait until the ramdord completion returns.
+ * This function will wait until the ramrod completion returns.
  * Return 0 if success, -ENODEV if ramrod doesn't return.
  */
 static int bnx2x_set_iscsi_eth_mac_addr(struct bnx2x *bp)
@@ -12996,7 +13186,6 @@ static void bnx2x_cnic_sp_post(struct bnx2x *bp, int count)
 	BUG_ON(bp->cnic_spq_pending < count);
 	bp->cnic_spq_pending -= count;
 
-
 	for (; bp->cnic_kwq_pending; bp->cnic_kwq_pending--) {
 		u16 type =  (le16_to_cpu(bp->cnic_kwq_cons->hdr.type)
 				& SPE_HDR_CONN_TYPE) >>
@@ -13169,7 +13358,6 @@ static void bnx2x_cnic_cfc_comp(struct bnx2x *bp, int cid, u8 err)
 	bnx2x_cnic_sp_post(bp, 0);
 }
 
-
 /* Called with netif_addr_lock_bh() taken.
  * Sets an rx_mode config for an iSCSI ETH client.
  * Doesn't block.
@@ -13210,7 +13398,6 @@ static void bnx2x_set_iscsi_eth_rx_mode(struct bnx2x *bp, bool start)
 	}
 }
 
-
 static int bnx2x_drv_ctl(struct net_device *dev, struct drv_ctl_info *ctl)
 {
 	struct bnx2x *bp = netdev_priv(dev);
@@ -13398,7 +13585,6 @@ void bnx2x_setup_cnic_info(struct bnx2x *bp)
 {
 	struct cnic_eth_dev *cp = &bp->cnic_eth_dev;
 
-
 	cp->ctx_tbl_offset = FUNC_ILT_BASE(BP_FUNC(bp)) +
 			     bnx2x_cid_ilt_lines(bp);
 	cp->starting_cid = bnx2x_cid_ilt_lines(bp) * ILT_PAGE_CIDS;
@@ -13434,7 +13620,6 @@ static int bnx2x_register_cnic(struct net_device *dev, struct cnic_ops *ops,
 			BNX2X_ERR("CNIC-related load failed\n");
 			return rc;
 		}
-
 	}
 
 	bp->cnic_enabled = true;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
index d22bc40..8e627b8 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
@@ -35,6 +35,8 @@
 #define ATC_REG_ATC_INT_STS_CLR					 0x1101c0
 /* [RW 5] Parity mask register #0 read/write */
 #define ATC_REG_ATC_PRTY_MASK					 0x1101d8
+/* [R 5] Parity register #0 read */
+#define ATC_REG_ATC_PRTY_STS					 0x1101cc
 /* [RC 5] Parity register #0 read clear */
 #define ATC_REG_ATC_PRTY_STS_CLR				 0x1101d0
 /* [RW 19] Interrupt mask register #0 read/write */
@@ -2750,6 +2752,8 @@
 #define PBF_REG_PBF_INT_STS					 0x1401c8
 /* [RW 20] Parity mask register #0 read/write */
 #define PBF_REG_PBF_PRTY_MASK					 0x1401e4
+/* [R 28] Parity register #0 read */
+#define PBF_REG_PBF_PRTY_STS					 0x1401d8
 /* [RC 20] Parity register #0 read clear */
 #define PBF_REG_PBF_PRTY_STS_CLR				 0x1401dc
 /* [RW 16] The Ethernet type value for L2 tag 0 */
@@ -4517,6 +4521,8 @@
 #define TM_REG_TM_INT_STS					 0x1640f0
 /* [RW 7] Parity mask register #0 read/write */
 #define TM_REG_TM_PRTY_MASK					 0x16410c
+/* [R 7] Parity register #0 read */
+#define TM_REG_TM_PRTY_STS					 0x164100
 /* [RC 7] Parity register #0 read clear */
 #define TM_REG_TM_PRTY_STS_CLR					 0x164104
 /* [RW 8] The event id for aggregated interrupt 0 */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
index 32a9609..8f03c98 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
@@ -35,9 +35,9 @@
 /**
  * bnx2x_exe_queue_init - init the Exe Queue object
  *
- * @o:		poiter to the object
+ * @o:		pointer to the object
  * @exe_len:	length
- * @owner:	poiter to the owner
+ * @owner:	pointer to the owner
  * @validate:	validate function pointer
  * @optimize:	optimize function pointer
  * @exec:	execute function pointer
@@ -142,7 +142,6 @@ free_and_exit:
 	spin_unlock_bh(&o->lock);
 
 	return rc;
-
 }
 
 static inline void __bnx2x_exe_queue_reset_pending(
@@ -163,13 +162,11 @@ static inline void __bnx2x_exe_queue_reset_pending(
 static inline void bnx2x_exe_queue_reset_pending(struct bnx2x *bp,
 						 struct bnx2x_exe_queue_obj *o)
 {
-
 	spin_lock_bh(&o->lock);
 
 	__bnx2x_exe_queue_reset_pending(bp, o);
 
 	spin_unlock_bh(&o->lock);
-
 }
 
 /**
@@ -179,7 +176,7 @@ static inline void bnx2x_exe_queue_reset_pending(struct bnx2x *bp,
  * @o:			queue
  * @ramrod_flags:	flags
  *
- * (Atomicy is ensured using the exe_queue->lock).
+ * (Atomicity is ensured using the exe_queue->lock).
  */
 static inline int bnx2x_exe_queue_step(struct bnx2x *bp,
 				       struct bnx2x_exe_queue_obj *o,
@@ -192,8 +189,7 @@ static inline int bnx2x_exe_queue_step(struct bnx2x *bp,
 
 	spin_lock_bh(&o->lock);
 
-	/*
-	 * Next step should not be performed until the current is finished,
+	/* Next step should not be performed until the current is finished,
 	 * unless a DRV_CLEAR_ONLY bit is set. In this case we just want to
 	 * properly clear object internals without sending any command to the FW
 	 * which also implies there won't be any completion to clear the
@@ -209,8 +205,7 @@ static inline int bnx2x_exe_queue_step(struct bnx2x *bp,
 		}
 	}
 
-	/*
-	 * Run through the pending commands list and create a next
+	/* Run through the pending commands list and create a next
 	 * execution chunk.
 	 */
 	while (!list_empty(&o->exe_queue)) {
@@ -220,8 +215,7 @@ static inline int bnx2x_exe_queue_step(struct bnx2x *bp,
 
 		if (cur_len + elem->cmd_len <= o->exe_chunk_len) {
 			cur_len += elem->cmd_len;
-			/*
-			 * Prevent from both lists being empty when moving an
+			/* Prevent from both lists being empty when moving an
 			 * element. This will allow the call of
 			 * bnx2x_exe_queue_empty() without locking.
 			 */
@@ -241,14 +235,12 @@ static inline int bnx2x_exe_queue_step(struct bnx2x *bp,
 
 	rc = o->execute(bp, o->owner, &o->pending_comp, ramrod_flags);
 	if (rc < 0)
-		/*
-		 *  In case of an error return the commands back to the queue
-		 *  and reset the pending_comp.
+		/* In case of an error return the commands back to the queue
+		 * and reset the pending_comp.
 		 */
 		list_splice_init(&o->pending_comp, &o->exe_queue);
 	else if (!rc)
-		/*
-		 * If zero is returned, means there are no outstanding pending
+		/* If zero is returned, means there are no outstanding pending
 		 * completions and we may dismiss the pending list.
 		 */
 		__bnx2x_exe_queue_reset_pending(bp, o);
@@ -308,7 +300,6 @@ static inline int bnx2x_state_wait(struct bnx2x *bp, int state,
 	/* can take a while if any port is running */
 	int cnt = 5000;
 
-
 	if (CHIP_REV_IS_EMUL(bp))
 		cnt *= 20;
 
@@ -456,7 +447,6 @@ static int bnx2x_get_n_elements(struct bnx2x *bp, struct bnx2x_vlan_mac_obj *o,
 			DP(BNX2X_MSG_SP, "copied element number %d to address %p element was:\n",
 			   counter, next);
 			next += stride + size;
-
 		}
 	}
 	return counter * ETH_ALEN;
@@ -518,7 +508,6 @@ static int bnx2x_check_vlan_mac_add(struct bnx2x *bp,
 	return 0;
 }
 
-
 /* check_del() callbacks */
 static struct bnx2x_vlan_mac_registry_elem *
 	bnx2x_check_mac_del(struct bnx2x *bp,
@@ -609,7 +598,6 @@ static bool bnx2x_check_move_always_err(
 	return false;
 }
 
-
 static inline u8 bnx2x_vlan_mac_get_rx_tx_flag(struct bnx2x_vlan_mac_obj *o)
 {
 	struct bnx2x_raw_obj *raw = &o->raw;
@@ -626,7 +614,6 @@ static inline u8 bnx2x_vlan_mac_get_rx_tx_flag(struct bnx2x_vlan_mac_obj *o)
 	return rx_tx_flag;
 }
 
-
 void bnx2x_set_mac_in_nig(struct bnx2x *bp,
 			  bool add, unsigned char *dev_addr, int index)
 {
@@ -693,7 +680,7 @@ static inline void bnx2x_vlan_mac_set_cmd_hdr_e2(struct bnx2x *bp,
  *
  * @cid:	connection id
  * @type:	BNX2X_FILTER_XXX_PENDING
- * @hdr:	poiter to header to setup
+ * @hdr:	pointer to header to setup
  * @rule_cnt:
  *
  * currently we always configure one rule and echo field to contain a CID and an
@@ -707,7 +694,6 @@ static inline void bnx2x_vlan_mac_set_rdata_hdr_e2(u32 cid, int type,
 	hdr->rule_cnt = (u8)rule_cnt;
 }
 
-
 /* hw_config() callbacks */
 static void bnx2x_set_one_mac_e2(struct bnx2x *bp,
 				 struct bnx2x_vlan_mac_obj *o,
@@ -723,8 +709,7 @@ static void bnx2x_set_one_mac_e2(struct bnx2x *bp,
 	unsigned long *vlan_mac_flags = &elem->cmd_data.vlan_mac.vlan_mac_flags;
 	u8 *mac = elem->cmd_data.vlan_mac.u.mac.mac;
 
-	/*
-	 * Set LLH CAM entry: currently only iSCSI and ETH macs are
+	/* Set LLH CAM entry: currently only iSCSI and ETH macs are
 	 * relevant. In addition, current implementation is tuned for a
 	 * single ETH MAC.
 	 *
@@ -879,8 +864,7 @@ static void bnx2x_set_one_mac_e1x(struct bnx2x *bp,
 	struct bnx2x_raw_obj *raw = &o->raw;
 	struct mac_configuration_cmd *config =
 		(struct mac_configuration_cmd *)(raw->rdata);
-	/*
-	 * 57710 and 57711 do not support MOVE command,
+	/* 57710 and 57711 do not support MOVE command,
 	 * so it's either ADD or DEL
 	 */
 	bool add = (elem->cmd_data.vlan_mac.cmd == BNX2X_VLAN_MAC_ADD) ?
@@ -960,7 +944,6 @@ static void bnx2x_set_one_vlan_mac_e2(struct bnx2x *bp,
 	u16 vlan = elem->cmd_data.vlan_mac.u.vlan_mac.vlan;
 	u8 *mac = elem->cmd_data.vlan_mac.u.vlan_mac.mac;
 
-
 	/* Reset the ramrod data buffer for the first rule */
 	if (rule_idx == 0)
 		memset(data, 0, sizeof(*data));
@@ -969,7 +952,7 @@ static void bnx2x_set_one_vlan_mac_e2(struct bnx2x *bp,
 	bnx2x_vlan_mac_set_cmd_hdr_e2(bp, o, add, CLASSIFY_RULE_OPCODE_PAIR,
 				      &rule_entry->pair.header);
 
-	/* Set VLAN and MAC themselvs */
+	/* Set VLAN and MAC themselves */
 	rule_entry->pair.vlan = cpu_to_le16(vlan);
 	bnx2x_set_fw_mac_addr(&rule_entry->pair.mac_msb,
 			      &rule_entry->pair.mac_mid,
@@ -1021,8 +1004,7 @@ static void bnx2x_set_one_vlan_mac_e1h(struct bnx2x *bp,
 	struct bnx2x_raw_obj *raw = &o->raw;
 	struct mac_configuration_cmd *config =
 		(struct mac_configuration_cmd *)(raw->rdata);
-	/*
-	 * 57710 and 57711 do not support MOVE command,
+	/* 57710 and 57711 do not support MOVE command,
 	 * so it's either ADD or DEL
 	 */
 	bool add = (elem->cmd_data.vlan_mac.cmd == BNX2X_VLAN_MAC_ADD) ?
@@ -1046,7 +1028,7 @@ static void bnx2x_set_one_vlan_mac_e1h(struct bnx2x *bp,
  *
  * @bp:		device handle
  * @p:		command parameters
- * @ppos:	pointer to the cooky
+ * @ppos:	pointer to the cookie
  *
  * reconfigure next MAC/VLAN/VLAN-MAC element from the
  * previously configured elements list.
@@ -1054,7 +1036,7 @@ static void bnx2x_set_one_vlan_mac_e1h(struct bnx2x *bp,
  * from command parameters only RAMROD_COMP_WAIT bit in ramrod_flags is	taken
  * into an account
  *
- * pointer to the cooky  - that should be given back in the next call to make
+ * pointer to the cookie  - that should be given back in the next call to make
  * function handle the next element. If *ppos is set to NULL it will restart the
  * iterator. If returned *ppos == NULL this means that the last element has been
  * handled.
@@ -1102,8 +1084,7 @@ static int bnx2x_vlan_mac_restore(struct bnx2x *bp,
 	return bnx2x_config_vlan_mac(bp, p);
 }
 
-/*
- * bnx2x_exeq_get_mac/bnx2x_exeq_get_vlan/bnx2x_exeq_get_vlan_mac return a
+/* bnx2x_exeq_get_mac/bnx2x_exeq_get_vlan/bnx2x_exeq_get_vlan_mac return a
  * pointer to an element with a specific criteria and NULL if such an element
  * hasn't been found.
  */
@@ -1187,8 +1168,7 @@ static inline int bnx2x_validate_vlan_mac_add(struct bnx2x *bp,
 		return rc;
 	}
 
-	/*
-	 * Check if there is a pending ADD command for this
+	/* Check if there is a pending ADD command for this
 	 * MAC/VLAN/VLAN-MAC. Return an error if there is.
 	 */
 	if (exeq->get(exeq, elem)) {
@@ -1196,8 +1176,7 @@ static inline int bnx2x_validate_vlan_mac_add(struct bnx2x *bp,
 		return -EEXIST;
 	}
 
-	/*
-	 * TODO: Check the pending MOVE from other objects where this
+	/* TODO: Check the pending MOVE from other objects where this
 	 * object is a destination object.
 	 */
 
@@ -1240,8 +1219,7 @@ static inline int bnx2x_validate_vlan_mac_del(struct bnx2x *bp,
 		return -EEXIST;
 	}
 
-	/*
-	 * Check if there are pending DEL or MOVE commands for this
+	/* Check if there are pending DEL or MOVE commands for this
 	 * MAC/VLAN/VLAN-MAC. Return an error if so.
 	 */
 	memcpy(&query_elem, elem, sizeof(query_elem));
@@ -1292,8 +1270,7 @@ static inline int bnx2x_validate_vlan_mac_move(struct bnx2x *bp,
 	struct bnx2x_exe_queue_obj *src_exeq = &src_o->exe_queue;
 	struct bnx2x_exe_queue_obj *dest_exeq = &dest_o->exe_queue;
 
-	/*
-	 * Check if we can perform this operation based on the current registry
+	/* Check if we can perform this operation based on the current registry
 	 * state.
 	 */
 	if (!src_o->check_move(bp, src_o, dest_o,
@@ -1302,8 +1279,7 @@ static inline int bnx2x_validate_vlan_mac_move(struct bnx2x *bp,
 		return -EINVAL;
 	}
 
-	/*
-	 * Check if there is an already pending DEL or MOVE command for the
+	/* Check if there is an already pending DEL or MOVE command for the
 	 * source object or ADD command for a destination object. Return an
 	 * error if so.
 	 */
@@ -1392,7 +1368,7 @@ static int bnx2x_remove_vlan_mac(struct bnx2x *bp,
 }
 
 /**
- * bnx2x_wait_vlan_mac - passivly wait for 5 seconds until all work completes.
+ * bnx2x_wait_vlan_mac - passively wait for 5 seconds until all work completes.
  *
  * @bp:		device handle
  * @o:		bnx2x_vlan_mac_obj
@@ -1550,9 +1526,8 @@ static inline int bnx2x_vlan_mac_get_registry_elem(
 
 		/* Get a new CAM offset */
 		if (!o->get_cam_offset(o, &reg_elem->cam_offset)) {
-			/*
-			 * This shell never happen, because we have checked the
-			 * CAM availiability in the 'validate'.
+			/* This shall never happen, because we have checked the
+			 * CAM availability in the 'validate'.
 			 */
 			WARN_ON(1);
 			kfree(reg_elem);
@@ -1599,8 +1574,7 @@ static int bnx2x_execute_vlan_mac(struct bnx2x *bp,
 	struct bnx2x_vlan_mac_registry_elem *reg_elem;
 	enum bnx2x_vlan_mac_cmd cmd;
 
-	/*
-	 * If DRIVER_ONLY execution is requested, cleanup a registry
+	/* If DRIVER_ONLY execution is requested, cleanup a registry
 	 * and exit. Otherwise send a ramrod to FW.
 	 */
 	if (!drv_only) {
@@ -1609,11 +1583,10 @@ static int bnx2x_execute_vlan_mac(struct bnx2x *bp,
 		/* Set pending */
 		r->set_pending(r);
 
-		/* Fill tha ramrod data */
+		/* Fill the ramrod data */
 		list_for_each_entry(elem, exe_chunk, link) {
 			cmd = elem->cmd_data.vlan_mac.cmd;
-			/*
-			 * We will add to the target object in MOVE command, so
+			/* We will add to the target object in MOVE command, so
 			 * change the object for a CAM search.
 			 */
 			if (cmd == BNX2X_VLAN_MAC_MOVE)
@@ -1646,12 +1619,11 @@ static int bnx2x_execute_vlan_mac(struct bnx2x *bp,
 				idx++;
 		}
 
-		/*
-		 *  No need for an explicit memory barrier here as long we would
-		 *  need to ensure the ordering of writing to the SPQ element
-		 *  and updating of the SPQ producer which involves a memory
-		 *  read and we will have to put a full memory barrier there
-		 *  (inside bnx2x_sp_post()).
+		/* No need for an explicit memory barrier here as long we would
+		 * need to ensure the ordering of writing to the SPQ element
+		 * and updating of the SPQ producer which involves a memory
+		 * read and we will have to put a full memory barrier there
+		 * (inside bnx2x_sp_post()).
 		 */
 
 		rc = bnx2x_sp_post(bp, o->ramrod_cmd, r->cid,
@@ -1766,8 +1738,7 @@ int bnx2x_config_vlan_mac(
 			return rc;
 	}
 
-	/*
-	 * If nothing will be executed further in this iteration we want to
+	/* If nothing will be executed further in this iteration we want to
 	 * return PENDING if there are pending commands
 	 */
 	if (!bnx2x_exe_queue_empty(&o->exe_queue))
@@ -1786,13 +1757,11 @@ int bnx2x_config_vlan_mac(
 			return rc;
 	}
 
-	/*
-	 * RAMROD_COMP_WAIT is a superset of RAMROD_EXEC. If it was set
+	/* RAMROD_COMP_WAIT is a superset of RAMROD_EXEC. If it was set
 	 * then user want to wait until the last command is done.
 	 */
 	if (test_bit(RAMROD_COMP_WAIT, &p->ramrod_flags)) {
-		/*
-		 * Wait maximum for the current exe_queue length iterations plus
+		/* Wait maximum for the current exe_queue length iterations plus
 		 * one (for the current pending command).
 		 */
 		int max_iterations = bnx2x_exe_queue_length(&o->exe_queue) + 1;
@@ -1818,8 +1787,6 @@ int bnx2x_config_vlan_mac(
 	return rc;
 }
 
-
-
 /**
  * bnx2x_vlan_mac_del_all - delete elements with given vlan_mac_flags spec
  *
@@ -1829,7 +1796,7 @@ int bnx2x_config_vlan_mac(
  * @ramrod_flags:	execution flags to be used for this deletion
  *
  * if the last operation has completed successfully and there are no
- * moreelements left, positive value if the last operation has completed
+ * more elements left, positive value if the last operation has completed
  * successfully and there are more previously configured elements, negative
  * value is current operation has failed.
  */
@@ -1870,8 +1837,7 @@ static int bnx2x_vlan_mac_del_all(struct bnx2x *bp,
 	p.ramrod_flags = *ramrod_flags;
 	p.user_req.cmd = BNX2X_VLAN_MAC_DEL;
 
-	/*
-	 * Add all but the last VLAN-MAC to the execution queue without actually
+	/* Add all but the last VLAN-MAC to the execution queue without actually
 	 * execution anything.
 	 */
 	__clear_bit(RAMROD_COMP_WAIT, &p.ramrod_flags);
@@ -1934,7 +1900,6 @@ static inline void bnx2x_init_vlan_mac_common(struct bnx2x_vlan_mac_obj *o,
 			   state, pstate, type);
 }
 
-
 void bnx2x_init_mac_obj(struct bnx2x *bp,
 			struct bnx2x_vlan_mac_obj *mac_obj,
 			u8 cl_id, u32 cid, u8 func_id, void *rdata,
@@ -2048,8 +2013,7 @@ void bnx2x_init_vlan_mac_obj(struct bnx2x *bp,
 	/* CAM pool handling */
 	vlan_mac_obj->get_credit = bnx2x_get_credit_vlan_mac;
 	vlan_mac_obj->put_credit = bnx2x_put_credit_vlan_mac;
-	/*
-	 * CAM offset is relevant for 57710 and 57711 chips only which have a
+	/* CAM offset is relevant for 57710 and 57711 chips only which have a
 	 * single CAM for both MACs and VLAN-MAC pairs. So the offset
 	 * will be taken from MACs' pool object only.
 	 */
@@ -2092,7 +2056,6 @@ void bnx2x_init_vlan_mac_obj(struct bnx2x *bp,
 				     bnx2x_execute_vlan_mac,
 				     bnx2x_exeq_get_vlan_mac);
 	}
-
 }
 
 /* RX_MODE verbs: DROP_ALL/ACCEPT_ALL/ACCEPT_ALL_MULTI/ACCEPT_ALL_VLAN/NORMAL */
@@ -2117,12 +2080,12 @@ static int bnx2x_set_rx_mode_e1x(struct bnx2x *bp,
 	struct tstorm_eth_mac_filter_config *mac_filters =
 		(struct tstorm_eth_mac_filter_config *)p->rdata;
 
-	/* initial seeting is drop-all */
+	/* initial setting is drop-all */
 	u8 drop_all_ucast = 1, drop_all_mcast = 1;
 	u8 accp_all_ucast = 0, accp_all_bcast = 0, accp_all_mcast = 0;
 	u8 unmatched_unicast = 0;
 
-    /* In e1x there we only take into account rx acceot flag since tx switching
+    /* In e1x there we only take into account rx accept flag since tx switching
      * isn't enabled. */
 	if (test_bit(BNX2X_ACCEPT_UNICAST, &p->rx_accept_flags))
 		/* accept matched ucast */
@@ -2245,7 +2208,6 @@ static inline void bnx2x_rx_mode_set_cmd_state_e2(struct bnx2x *bp,
 	}
 
 	cmd->state = cpu_to_le16(state);
-
 }
 
 static int bnx2x_set_rx_mode_e2(struct bnx2x *bp,
@@ -2286,9 +2248,7 @@ static int bnx2x_set_rx_mode_e2(struct bnx2x *bp,
 					       false);
 	}
 
-
-	/*
-	 * If FCoE Queue configuration has been requested configure the Rx and
+	/* If FCoE Queue configuration has been requested configure the Rx and
 	 * internal switching modes for this queue in separate rules.
 	 *
 	 * FCoE queue shell never be set to ACCEPT_ALL packets of any sort:
@@ -2324,8 +2284,7 @@ static int bnx2x_set_rx_mode_e2(struct bnx2x *bp,
 		}
 	}
 
-	/*
-	 * Set the ramrod header (most importantly - number of rules to
+	/* Set the ramrod header (most importantly - number of rules to
 	 * configure).
 	 */
 	bnx2x_rx_mode_set_rdata_hdr_e2(p->cid, &data->header, rule_idx);
@@ -2334,12 +2293,11 @@ static int bnx2x_set_rx_mode_e2(struct bnx2x *bp,
 			 data->header.rule_cnt, p->rx_accept_flags,
 			 p->tx_accept_flags);
 
-	/*
-	 *  No need for an explicit memory barrier here as long we would
-	 *  need to ensure the ordering of writing to the SPQ element
-	 *  and updating of the SPQ producer which involves a memory
-	 *  read and we will have to put a full memory barrier there
-	 *  (inside bnx2x_sp_post()).
+	/* No need for an explicit memory barrier here as long we would
+	 * need to ensure the ordering of writing to the SPQ element
+	 * and updating of the SPQ producer which involves a memory
+	 * read and we will have to put a full memory barrier there
+	 * (inside bnx2x_sp_post()).
 	 */
 
 	/* Send a ramrod */
@@ -2476,7 +2434,7 @@ static int bnx2x_mcast_enqueue_cmd(struct bnx2x *bp,
 		cur_mac = (struct bnx2x_mcast_mac_elem *)
 			  ((u8 *)new_cmd + sizeof(*new_cmd));
 
-		/* Push the MACs of the current command into the pendig command
+		/* Push the MACs of the current command into the pending command
 		 * MACs list: FIFO
 		 */
 		list_for_each_entry(pos, &p->mcast_list, link) {
@@ -2909,7 +2867,6 @@ static int bnx2x_mcast_validate_e2(struct bnx2x *bp,
 	default:
 		BNX2X_ERR("Unknown command: %d\n", cmd);
 		return -EINVAL;
-
 	}
 
 	/* Increase the total number of MACs pending to be configured */
@@ -3034,20 +2991,18 @@ static int bnx2x_mcast_setup_e2(struct bnx2x *bp,
 	if (!o->total_pending_num)
 		bnx2x_mcast_refresh_registry_e2(bp, o);
 
-	/*
-	 * If CLEAR_ONLY was requested - don't send a ramrod and clear
+	/* If CLEAR_ONLY was requested - don't send a ramrod and clear
 	 * RAMROD_PENDING status immediately.
 	 */
 	if (test_bit(RAMROD_DRV_CLR_ONLY, &p->ramrod_flags)) {
 		raw->clear_pending(raw);
 		return 0;
 	} else {
-		/*
-		 *  No need for an explicit memory barrier here as long we would
-		 *  need to ensure the ordering of writing to the SPQ element
-		 *  and updating of the SPQ producer which involves a memory
-		 *  read and we will have to put a full memory barrier there
-		 *  (inside bnx2x_sp_post()).
+		/* No need for an explicit memory barrier here as long we would
+		 * need to ensure the ordering of writing to the SPQ element
+		 * and updating of the SPQ producer which involves a memory
+		 * read and we will have to put a full memory barrier there
+		 * (inside bnx2x_sp_post()).
 		 */
 
 		/* Send a ramrod */
@@ -3121,7 +3076,7 @@ static inline void bnx2x_mcast_hdl_restore_e1h(struct bnx2x *bp,
 	}
 }
 
-/* On 57711 we write the multicast MACs' aproximate match
+/* On 57711 we write the multicast MACs' approximate match
  * table by directly into the TSTORM's internal RAM. So we don't
  * really need to handle any tricks to make it work.
  */
@@ -3223,7 +3178,6 @@ static int bnx2x_mcast_validate_e1(struct bnx2x *bp,
 	default:
 		BNX2X_ERR("Unknown command: %d\n", cmd);
 		return -EINVAL;
-
 	}
 
 	/* We want to ensure that commands are executed one by one for 57710.
@@ -3245,7 +3199,7 @@ static void bnx2x_mcast_revert_e1(struct bnx2x *bp,
 
 	/* If current command hasn't been handled yet and we are
 	 * here means that it's meant to be dropped and we have to
-	 * update the number of outstandling MACs accordingly.
+	 * update the number of outstanding MACs accordingly.
 	 */
 	if (p->mcast_list_len)
 		o->total_pending_num -= o->max_cmd_len;
@@ -3342,7 +3296,6 @@ static inline int bnx2x_mcast_handle_restore_cmd_e1(
 	return -1;
 }
 
-
 static inline int bnx2x_mcast_handle_pending_cmds_e1(
 	struct bnx2x *bp, struct bnx2x_mcast_ramrod_params *p)
 {
@@ -3352,7 +3305,6 @@ static inline int bnx2x_mcast_handle_pending_cmds_e1(
 	union bnx2x_mcast_config_data cfg_data = {NULL};
 	int cnt = 0;
 
-
 	/* If nothing to be done - return */
 	if (list_empty(&o->pending_cmds_head))
 		return 0;
@@ -3523,20 +3475,18 @@ static int bnx2x_mcast_setup_e1(struct bnx2x *bp,
 	if (rc)
 		return rc;
 
-	/*
-	 * If CLEAR_ONLY was requested - don't send a ramrod and clear
+	/* If CLEAR_ONLY was requested - don't send a ramrod and clear
 	 * RAMROD_PENDING status immediately.
 	 */
 	if (test_bit(RAMROD_DRV_CLR_ONLY, &p->ramrod_flags)) {
 		raw->clear_pending(raw);
 		return 0;
 	} else {
-		/*
-		 *  No need for an explicit memory barrier here as long we would
-		 *  need to ensure the ordering of writing to the SPQ element
-		 *  and updating of the SPQ producer which involves a memory
-		 *  read and we will have to put a full memory barrier there
-		 *  (inside bnx2x_sp_post()).
+		/* No need for an explicit memory barrier here as long we would
+		 * need to ensure the ordering of writing to the SPQ element
+		 * and updating of the SPQ producer which involves a memory
+		 * read and we will have to put a full memory barrier there
+		 * (inside bnx2x_sp_post()).
 		 */
 
 		/* Send a ramrod */
@@ -3550,7 +3500,6 @@ static int bnx2x_mcast_setup_e1(struct bnx2x *bp,
 		/* Ramrod completion is pending */
 		return 1;
 	}
-
 }
 
 static int bnx2x_mcast_get_registry_size_exact(struct bnx2x_mcast_obj *o)
@@ -3848,7 +3797,6 @@ static bool bnx2x_credit_pool_always_true(struct bnx2x_credit_pool_obj *o,
 	return true;
 }
 
-
 static bool bnx2x_credit_pool_get_entry(
 	struct bnx2x_credit_pool_obj *o,
 	int *offset)
@@ -3999,8 +3947,7 @@ void bnx2x_init_mac_credit_pool(struct bnx2x *bp,
 
 	} else {
 
-		/*
-		 * CAM credit is equaly divided between all active functions
+		/* CAM credit is equaly divided between all active functions
 		 * on the PATH.
 		 */
 		if ((func_num > 0)) {
@@ -4009,8 +3956,7 @@ void bnx2x_init_mac_credit_pool(struct bnx2x *bp,
 			else
 				cam_sz = BNX2X_CAM_SIZE_EMUL;
 
-			/*
-			 * No need for CAM entries handling for 57712 and
+			/* No need for CAM entries handling for 57712 and
 			 * newer.
 			 */
 			bnx2x_init_credit_pool(p, -1, cam_sz);
@@ -4018,7 +3964,6 @@ void bnx2x_init_mac_credit_pool(struct bnx2x *bp,
 			/* this should never happen! Block MAC operations. */
 			bnx2x_init_credit_pool(p, 0, 0);
 		}
-
 	}
 }
 
@@ -4028,14 +3973,12 @@ void bnx2x_init_vlan_credit_pool(struct bnx2x *bp,
 				 u8 func_num)
 {
 	if (CHIP_IS_E1x(bp)) {
-		/*
-		 * There is no VLAN credit in HW on 57710 and 57711 only
+		/* There is no VLAN credit in HW on 57710 and 57711 only
 		 * MAC / MAC-VLAN can be set
 		 */
 		bnx2x_init_credit_pool(p, 0, -1);
 	} else {
-		/*
-		 * CAM credit is equaly divided between all active functions
+		/* CAM credit is equally divided between all active functions
 		 * on the PATH.
 		 */
 		if (func_num > 0) {
@@ -4051,7 +3994,7 @@ void bnx2x_init_vlan_credit_pool(struct bnx2x *bp,
 /**
  * bnx2x_debug_print_ind_table - prints the indirection table configuration.
  *
- * @bp:		driver hanlde
+ * @bp:		driver handle
  * @p:		pointer to rss configuration
  *
  * Prints it when NETIF_MSG_IFUP debug level is configured.
@@ -4164,12 +4107,11 @@ static int bnx2x_setup_rss(struct bnx2x *bp,
 		data->capabilities |= ETH_RSS_UPDATE_RAMROD_DATA_UPDATE_RSS_KEY;
 	}
 
-	/*
-	 *  No need for an explicit memory barrier here as long we would
-	 *  need to ensure the ordering of writing to the SPQ element
-	 *  and updating of the SPQ producer which involves a memory
-	 *  read and we will have to put a full memory barrier there
-	 *  (inside bnx2x_sp_post()).
+	/* No need for an explicit memory barrier here as long we would
+	 * need to ensure the ordering of writing to the SPQ element
+	 * and updating of the SPQ producer which involves a memory
+	 * read and we will have to put a full memory barrier there
+	 * (inside bnx2x_sp_post()).
 	 */
 
 	/* Send a ramrod */
@@ -4215,7 +4157,6 @@ int bnx2x_config_rss(struct bnx2x *bp,
 	return rc;
 }
 
-
 void bnx2x_init_rss_config_obj(struct bnx2x *bp,
 			       struct bnx2x_rss_config_obj *rss_obj,
 			       u8 cl_id, u32 cid, u8 func_id, u8 engine_id,
@@ -4288,7 +4229,6 @@ int bnx2x_queue_state_change(struct bnx2x *bp,
 	return !!test_bit(pending_bit, pending);
 }
 
-
 static int bnx2x_queue_set_pending(struct bnx2x_queue_sp_obj *obj,
 				   struct bnx2x_queue_state_params *params)
 {
@@ -4337,7 +4277,7 @@ static int bnx2x_queue_comp_cmd(struct bnx2x *bp,
 	}
 
 	if (o->next_tx_only >= o->max_cos)
-		/* >= becuase tx only must always be smaller than cos since the
+		/* >= because tx only must always be smaller than cos since the
 		 * primary connection supports COS 0
 		 */
 		BNX2X_ERR("illegal value for next tx_only: %d. max cos was %d",
@@ -4403,7 +4343,6 @@ static void bnx2x_q_fill_init_general_data(struct bnx2x *bp,
 	gen_data->mtu = cpu_to_le16(params->mtu);
 	gen_data->func_id = o->func_id;
 
-
 	gen_data->cos = params->cos;
 
 	gen_data->traffic_type =
@@ -4530,7 +4469,6 @@ static void bnx2x_q_fill_init_rx_data(struct bnx2x_queue_sp_obj *o,
 		cpu_to_le16(params->silent_removal_value);
 	rx_data->silent_vlan_mask =
 		cpu_to_le16(params->silent_removal_mask);
-
 }
 
 /* initialize the general, tx and rx parts of a queue object */
@@ -4652,12 +4590,11 @@ static inline int bnx2x_q_send_setup_e1x(struct bnx2x *bp,
 	/* Fill the ramrod data */
 	bnx2x_q_fill_setup_data_cmn(bp, params, rdata);
 
-	/*
-	 *  No need for an explicit memory barrier here as long we would
-	 *  need to ensure the ordering of writing to the SPQ element
-	 *  and updating of the SPQ producer which involves a memory
-	 *  read and we will have to put a full memory barrier there
-	 *  (inside bnx2x_sp_post()).
+	/* No need for an explicit memory barrier here as long we would
+	 * need to ensure the ordering of writing to the SPQ element
+	 * and updating of the SPQ producer which involves a memory
+	 * read and we will have to put a full memory barrier there
+	 * (inside bnx2x_sp_post()).
 	 */
 
 	return bnx2x_sp_post(bp, ramrod, o->cids[BNX2X_PRIMARY_CID_INDEX],
@@ -4681,12 +4618,11 @@ static inline int bnx2x_q_send_setup_e2(struct bnx2x *bp,
 	bnx2x_q_fill_setup_data_cmn(bp, params, rdata);
 	bnx2x_q_fill_setup_data_e2(bp, params, rdata);
 
-	/*
-	 *  No need for an explicit memory barrier here as long we would
-	 *  need to ensure the ordering of writing to the SPQ element
-	 *  and updating of the SPQ producer which involves a memory
-	 *  read and we will have to put a full memory barrier there
-	 *  (inside bnx2x_sp_post()).
+	/* No need for an explicit memory barrier here as long we would
+	 * need to ensure the ordering of writing to the SPQ element
+	 * and updating of the SPQ producer which involves a memory
+	 * read and we will have to put a full memory barrier there
+	 * (inside bnx2x_sp_post()).
 	 */
 
 	return bnx2x_sp_post(bp, ramrod, o->cids[BNX2X_PRIMARY_CID_INDEX],
@@ -4706,7 +4642,6 @@ static inline int bnx2x_q_send_setup_tx_only(struct bnx2x *bp,
 		&params->params.tx_only;
 	u8 cid_index = tx_only_params->cid_index;
 
-
 	if (cid_index >= o->max_cos) {
 		BNX2X_ERR("queue[%d]: cid_index (%d) is out of range\n",
 			  o->cl_id, cid_index);
@@ -4727,12 +4662,11 @@ static inline int bnx2x_q_send_setup_tx_only(struct bnx2x *bp,
 			 o->cids[cid_index], rdata->general.client_id,
 			 rdata->general.sp_client_id, rdata->general.cos);
 
-	/*
-	 *  No need for an explicit memory barrier here as long we would
-	 *  need to ensure the ordering of writing to the SPQ element
-	 *  and updating of the SPQ producer which involves a memory
-	 *  read and we will have to put a full memory barrier there
-	 *  (inside bnx2x_sp_post()).
+	/* No need for an explicit memory barrier here as long we would
+	 * need to ensure the ordering of writing to the SPQ element
+	 * and updating of the SPQ producer which involves a memory
+	 * read and we will have to put a full memory barrier there
+	 * (inside bnx2x_sp_post()).
 	 */
 
 	return bnx2x_sp_post(bp, ramrod, o->cids[cid_index],
@@ -4761,7 +4695,7 @@ static void bnx2x_q_fill_update_data(struct bnx2x *bp,
 		test_bit(BNX2X_Q_UPDATE_IN_VLAN_REM_CHNG,
 			 &params->update_flags);
 
-	/* Outer VLAN sripping */
+	/* Outer VLAN stripping */
 	data->outer_vlan_removal_enable_flg =
 		test_bit(BNX2X_Q_UPDATE_OUT_VLAN_REM, &params->update_flags);
 	data->outer_vlan_removal_change_flg =
@@ -4816,19 +4750,17 @@ static inline int bnx2x_q_send_update(struct bnx2x *bp,
 		return -EINVAL;
 	}
 
-
 	/* Clear the ramrod data */
 	memset(rdata, 0, sizeof(*rdata));
 
 	/* Fill the ramrod data */
 	bnx2x_q_fill_update_data(bp, o, update_params, rdata);
 
-	/*
-	 *  No need for an explicit memory barrier here as long we would
-	 *  need to ensure the ordering of writing to the SPQ element
-	 *  and updating of the SPQ producer which involves a memory
-	 *  read and we will have to put a full memory barrier there
-	 *  (inside bnx2x_sp_post()).
+	/* No need for an explicit memory barrier here as long we would
+	 * need to ensure the ordering of writing to the SPQ element
+	 * and updating of the SPQ producer which involves a memory
+	 * read and we will have to put a full memory barrier there
+	 * (inside bnx2x_sp_post()).
 	 */
 
 	return bnx2x_sp_post(bp, RAMROD_CMD_ID_ETH_CLIENT_UPDATE,
@@ -5038,8 +4970,7 @@ static int bnx2x_queue_chk_transition(struct bnx2x *bp,
 		 &params->params.update;
 	u8 next_tx_only = o->num_tx_only;
 
-	/*
-	 * Forget all pending for completion commands if a driver only state
+	/* Forget all pending for completion commands if a driver only state
 	 * transition has been requested.
 	 */
 	if (test_bit(RAMROD_DRV_CLR_ONLY, &params->ramrod_flags)) {
@@ -5047,8 +4978,7 @@ static int bnx2x_queue_chk_transition(struct bnx2x *bp,
 		o->next_state = BNX2X_Q_STATE_MAX;
 	}
 
-	/*
-	 * Don't allow a next state transition if we are in the middle of
+	/* Don't allow a next state transition if we are in the middle of
 	 * the previous one.
 	 */
 	if (o->pending) {
@@ -5257,8 +5187,7 @@ enum bnx2x_func_state bnx2x_func_get_state(struct bnx2x *bp,
 	if (o->pending)
 		return BNX2X_F_STATE_MAX;
 
-	/*
-	 * unsure the order of reading of o->pending and o->state
+	/* unsure the order of reading of o->pending and o->state
 	 * o->pending should be read first
 	 */
 	rmb();
@@ -5356,8 +5285,7 @@ static int bnx2x_func_chk_transition(struct bnx2x *bp,
 	enum bnx2x_func_state state = o->state, next_state = BNX2X_F_STATE_MAX;
 	enum bnx2x_func_cmd cmd = params->cmd;
 
-	/*
-	 * Forget all pending for completion commands if a driver only state
+	/* Forget all pending for completion commands if a driver only state
 	 * transition has been requested.
 	 */
 	if (test_bit(RAMROD_DRV_CLR_ONLY, &params->ramrod_flags)) {
@@ -5365,8 +5293,7 @@ static int bnx2x_func_chk_transition(struct bnx2x *bp,
 		o->next_state = BNX2X_F_STATE_MAX;
 	}
 
-	/*
-	 * Don't allow a next state transition if we are in the middle of
+	/* Don't allow a next state transition if we are in the middle of
 	 * the previous one.
 	 */
 	if (o->pending)
@@ -5539,7 +5466,7 @@ static int bnx2x_func_hw_init(struct bnx2x *bp,
 		goto init_err;
 	}
 
-	/* Handle the beginning of COMMON_XXX pases separatelly... */
+	/* Handle the beginning of COMMON_XXX pases separately... */
 	switch (load_code) {
 	case FW_MSG_CODE_DRV_LOAD_COMMON_CHIP:
 		rc = bnx2x_func_init_cmn_chip(bp, drv);
@@ -5573,7 +5500,7 @@ static int bnx2x_func_hw_init(struct bnx2x *bp,
 init_err:
 	drv->gunzip_end(bp);
 
-	/* In case of success, complete the comand immediatelly: no ramrods
+	/* In case of success, complete the command immediately: no ramrods
 	 * have been sent.
 	 */
 	if (!rc)
@@ -5598,7 +5525,7 @@ static inline void bnx2x_func_reset_func(struct bnx2x *bp,
 }
 
 /**
- * bnx2x_func_reset_port - reser HW at port stage
+ * bnx2x_func_reset_port - reset HW at port stage
  *
  * @bp:		device handle
  * @drv:
@@ -5620,7 +5547,7 @@ static inline void bnx2x_func_reset_port(struct bnx2x *bp,
 }
 
 /**
- * bnx2x_func_reset_cmn - reser HW at common stage
+ * bnx2x_func_reset_cmn - reset HW at common stage
  *
  * @bp:		device handle
  * @drv:
@@ -5636,7 +5563,6 @@ static inline void bnx2x_func_reset_cmn(struct bnx2x *bp,
 	drv->reset_hw_cmn(bp);
 }
 
-
 static inline int bnx2x_func_hw_reset(struct bnx2x *bp,
 				      struct bnx2x_func_state_params *params)
 {
@@ -5663,7 +5589,7 @@ static inline int bnx2x_func_hw_reset(struct bnx2x *bp,
 		break;
 	}
 
-	/* Complete the comand immediatelly: no ramrods have been sent. */
+	/* Complete the command immediately: no ramrods have been sent. */
 	o->complete_cmd(bp, o, BNX2X_F_CMD_HW_RESET);
 
 	return 0;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
index 43c00bc..798dfe9 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
@@ -34,8 +34,7 @@ enum {
 	RAMROD_RESTORE,
 	 /* Execute the next command now */
 	RAMROD_EXEC,
-	/*
-	 * Don't add a new command and continue execution of posponed
+	/* Don't add a new command and continue execution of postponed
 	 * commands. If not set a new command will be added to the
 	 * pending commands list.
 	 */
@@ -129,8 +128,7 @@ enum bnx2x_vlan_mac_cmd {
 struct bnx2x_vlan_mac_data {
 	/* Requested command: BNX2X_VLAN_MAC_XX */
 	enum bnx2x_vlan_mac_cmd cmd;
-	/*
-	 * used to contain the data related vlan_mac_flags bits from
+	/* used to contain the data related vlan_mac_flags bits from
 	 * ramrod parameters.
 	 */
 	unsigned long vlan_mac_flags;
@@ -190,14 +188,10 @@ typedef struct bnx2x_exeq_elem *
 				     struct bnx2x_exeq_elem *elem);
 
 struct bnx2x_exe_queue_obj {
-	/*
-	 * Commands pending for an execution.
-	 */
+	/* Commands pending for an execution. */
 	struct list_head	exe_queue;
 
-	/*
-	 * Commands pending for an completion.
-	 */
+	/* Commands pending for an completion. */
 	struct list_head	pending_comp;
 
 	spinlock_t		lock;
@@ -245,14 +239,13 @@ struct bnx2x_exe_queue_obj {
 };
 /***************** Classification verbs: Set/Del MAC/VLAN/VLAN-MAC ************/
 /*
- * Element in the VLAN_MAC registry list having all currenty configured
+ * Element in the VLAN_MAC registry list having all currently configured
  * rules.
  */
 struct bnx2x_vlan_mac_registry_elem {
 	struct list_head	link;
 
-	/*
-	 * Used to store the cam offset used for the mac/vlan/vlan-mac.
+	/* Used to store the cam offset used for the mac/vlan/vlan-mac.
 	 * Relevant for 57710 and 57711 only. VLANs and MACs share the
 	 * same CAM for these chips.
 	 */
@@ -310,7 +303,7 @@ struct bnx2x_vlan_mac_obj {
 	 * @param n number of elements to get
 	 * @param buf buffer preallocated by caller into which elements
 	 *            will be copied. Note elements are 4-byte aligned
-	 *            so buffer size must be able to accomodate the
+	 *            so buffer size must be able to accommodate the
 	 *            aligned elements.
 	 *
 	 * @return number of copied bytes
@@ -395,7 +388,7 @@ struct bnx2x_vlan_mac_obj {
 	 * @param bp
 	 * @param p Command parameters (RAMROD_COMP_WAIT bit in
 	 *          ramrod_flags is only taken into an account)
-	 * @param ppos a pointer to the cooky that should be given back in the
+	 * @param ppos a pointer to the cookie that should be given back in the
 	 *        next call to make function handle the next element. If
 	 *        *ppos is set to NULL it will restart the iterator.
 	 *        If returned *ppos == NULL this means that the last
@@ -408,7 +401,7 @@ struct bnx2x_vlan_mac_obj {
 		       struct bnx2x_vlan_mac_registry_elem **ppos);
 
 	/**
-	 * Should be called on a completion arival.
+	 * Should be called on a completion arrival.
 	 *
 	 * @param bp
 	 * @param o
@@ -447,7 +440,7 @@ void bnx2x_set_mac_in_nig(struct bnx2x *bp,
 
 /** RX_MODE verbs:DROP_ALL/ACCEPT_ALL/ACCEPT_ALL_MULTI/ACCEPT_ALL_VLAN/NORMAL */
 
-/* RX_MODE ramrod spesial flags: set in rx_mode_flags field in
+/* RX_MODE ramrod special flags: set in rx_mode_flags field in
  * a bnx2x_rx_mode_ramrod_params.
  */
 enum {
@@ -475,8 +468,7 @@ struct bnx2x_rx_mode_ramrod_params {
 	unsigned long ramrod_flags;
 	unsigned long rx_mode_flags;
 
-	/*
-	 * rdata is either a pointer to eth_filter_rules_ramrod_data(e2) or to
+	/* rdata is either a pointer to eth_filter_rules_ramrod_data(e2) or to
 	 * a tstorm_eth_mac_filter_config (e1x).
 	 */
 	void *rdata;
@@ -646,12 +638,11 @@ struct bnx2x_credit_pool_obj {
 	/* Maximum allowed credit. put() will check against it. */
 	int		pool_sz;
 
-	/*
-	 *  Allocate a pool table statically.
+	/* Allocate a pool table statically.
 	 *
-	 *  Currently the mamimum allowed size is MAX_MAC_CREDIT_E2(272)
+	 * Currently the maximum allowed size is MAX_MAC_CREDIT_E2(272)
 	 *
-	 *  The set bit in the table will mean that the entry is available.
+	 * The set bit in the table will mean that the entry is available.
 	 */
 #define BNX2X_POOL_VEC_SIZE	(MAX_MAC_CREDIT_E2 / 64)
 	u64		pool_mirror[BNX2X_POOL_VEC_SIZE];
@@ -832,7 +823,7 @@ enum {
 	BNX2X_Q_FLG_TUN_INC_INNER_IP_ID
 };
 
-/* Queue type options: queue type may be a compination of below. */
+/* Queue type options: queue type may be a combination of below. */
 enum bnx2x_q_type {
 	/** TODO: Consider moving both these flags into the init()
 	 *        ramrod params.
@@ -1002,10 +993,9 @@ struct bnx2x_queue_sp_obj {
 	u8		cl_id;
 	u8		func_id;
 
-	/*
-	 * number of traffic classes supported by queue.
-	 * The primary connection of the queue suppotrs the first traffic
-	 * class. Any further traffic class is suppoted by a tx-only
+	/* number of traffic classes supported by queue.
+	 * The primary connection of the queue supports the first traffic
+	 * class. Any further traffic class is supported by a tx-only
 	 * connection.
 	 *
 	 * Therefore max_cos is also a number of valid entries in the cids
@@ -1021,7 +1011,7 @@ struct bnx2x_queue_sp_obj {
 
 	/* BNX2X_Q_CMD_XX bits. This object implements "one
 	 * pending" paradigm but for debug and tracing purposes it's
-	 * more convinient to have different bits for different
+	 * more convenient to have different bits for different
 	 * commands.
 	 */
 	unsigned long	pending;
@@ -1210,7 +1200,7 @@ struct bnx2x_func_sp_obj {
 
 	/* BNX2X_FUNC_CMD_XX bits. This object implements "one
 	 * pending" paradigm but for debug and tracing purposes it's
-	 * more convinient to have different bits for different
+	 * more convenient to have different bits for different
 	 * commands.
 	 */
 	unsigned long		pending;
@@ -1329,7 +1319,7 @@ void bnx2x_init_rx_mode_obj(struct bnx2x *bp,
  *
  * @p: Command parameters
  *
- * Return: 0 - if operation was successfull and there is no pending completions,
+ * Return: 0 - if operation was successful and there is no pending completions,
  *         positive number - if there are pending completions,
  *         negative - if there were errors
  */
@@ -1361,7 +1351,7 @@ void bnx2x_init_mcast_obj(struct bnx2x *bp,
  * the current command will be enqueued to the tail of the
  * pending commands list.
  *
- * Return: 0 is operation was successfull and there are no pending completions,
+ * Return: 0 is operation was successful and there are no pending completions,
  *         negative if there were errors, positive if there are pending
  *         completions.
  */
@@ -1377,7 +1367,6 @@ void bnx2x_init_vlan_credit_pool(struct bnx2x *bp,
 				 struct bnx2x_credit_pool_obj *p, u8 func_id,
 				 u8 func_num);
 
-
 /****************** RSS CONFIGURATION ****************/
 void bnx2x_init_rss_config_obj(struct bnx2x *bp,
 			       struct bnx2x_rss_config_obj *rss_obj,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
index 2ce7c74..95861ef 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
@@ -1341,7 +1341,7 @@ int bnx2x_vfop_qdown_cmd(struct bnx2x *bp,
  */
 
 /* internal vf enable - until vf is enabled internally all transactions
- * are blocked. this routine should always be called last with pretend.
+ * are blocked. This routine should always be called last with pretend.
  */
 static void bnx2x_vf_enable_internal(struct bnx2x *bp, u8 enable)
 {
@@ -1459,21 +1459,16 @@ static u8 bnx2x_vf_is_pcie_pending(struct bnx2x *bp, u8 abs_vfid)
 	struct bnx2x_virtf *vf = bnx2x_vf_by_abs_fid(bp, abs_vfid);
 
 	if (!vf)
-		goto unknown_dev;
+		return false;
 
 	dev = pci_get_bus_and_slot(vf->bus, vf->devfn);
 	if (dev)
 		return bnx2x_is_pcie_pending(dev);
-
-unknown_dev:
 	return false;
 }
 
 int bnx2x_vf_flr_clnup_epilog(struct bnx2x *bp, u8 abs_vfid)
 {
-	/* Wait 100ms */
-	msleep(100);
-
 	/* Verify no pending pci transactions */
 	if (bnx2x_vf_is_pcie_pending(bp, abs_vfid))
 		BNX2X_ERR("PCIE Transactions still pending\n");
@@ -1620,7 +1615,7 @@ next_vf_to_clean:
 	     i++)
 		;
 
-	DP(BNX2X_MSG_IOV, "next vf to cleanup: %d. num of vfs: %d\n", i,
+	DP(BNX2X_MSG_IOV, "next vf to cleanup: %d. Num of vfs: %d\n", i,
 	   BNX2X_NR_VIRTFN(bp));
 
 	if (i < BNX2X_NR_VIRTFN(bp)) {
@@ -1743,7 +1738,7 @@ void bnx2x_iov_init_dq(struct bnx2x *bp)
 	REG_WR(bp, DORQ_REG_VF_TYPE_MIN_MCID_0, 0);
 	REG_WR(bp, DORQ_REG_VF_TYPE_MAX_MCID_0, 0x1ffff);
 
-	/* set the number of VF alllowed doorbells to the full DQ range */
+	/* set the number of VF allowed doorbells to the full DQ range */
 	REG_WR(bp, DORQ_REG_VF_NORM_MAX_CID_COUNT, 0x20000);
 
 	/* set the VF doorbell threshold */
@@ -2176,6 +2171,9 @@ int bnx2x_iov_nic_init(struct bnx2x *bp)
 
 	DP(BNX2X_MSG_IOV, "num of vfs: %d\n", (bp)->vfdb->sriov.nr_virtfn);
 
+	/* let FLR complete ... */
+	msleep(100);
+
 	/* initialize vf database */
 	for_each_vf(bp, vfid) {
 		struct bnx2x_virtf *vf = BP_VF(bp, vfid);
@@ -2403,7 +2401,7 @@ int bnx2x_iov_eq_sp_event(struct bnx2x *bp, union event_ring_elem *elem)
 
 	/* extract vf and rxq index from vf_cid - relies on the following:
 	 * 1. vfid on cid reflects the true abs_vfid
-	 * 2. the max number of VFs (per path) is 64
+	 * 2. The max number of VFs (per path) is 64
 	 */
 	qidx = cid & ((1 << BNX2X_VF_CID_WND)-1);
 	abs_vfid = (cid >> BNX2X_VF_CID_WND) & (BNX2X_MAX_NUM_OF_VFS-1);
@@ -2461,7 +2459,7 @@ static struct bnx2x_virtf *bnx2x_vf_by_cid(struct bnx2x *bp, int vf_cid)
 {
 	/* extract the vf from vf_cid - relies on the following:
 	 * 1. vfid on cid reflects the true abs_vfid
-	 * 2. the max number of VFs (per path) is 64
+	 * 2. The max number of VFs (per path) is 64
 	 */
 	int abs_vfid = (vf_cid >> BNX2X_VF_CID_WND) & (BNX2X_MAX_NUM_OF_VFS-1);
 	return bnx2x_vf_by_abs_fid(bp, abs_vfid);
@@ -2480,7 +2478,7 @@ void bnx2x_iov_set_queue_sp_obj(struct bnx2x *bp, int vf_cid,
 	if (vf) {
 		/* extract queue index from vf_cid - relies on the following:
 		 * 1. vfid on cid reflects the true abs_vfid
-		 * 2. the max number of VFs (per path) is 64
+		 * 2. The max number of VFs (per path) is 64
 		 */
 		int q_index = vf_cid & ((1 << BNX2X_VF_CID_WND)-1);
 		*q_obj = &bnx2x_vfq(vf, q_index, sp_obj);
@@ -2705,7 +2703,7 @@ int bnx2x_vf_acquire(struct bnx2x *bp, struct bnx2x_virtf *vf,
 	}
 
 	/* static allocation:
-	 * the global maximum number are fixed per VF. fail the request if
+	 * the global maximum number are fixed per VF. Fail the request if
 	 * requested number exceed these globals
 	 */
 	if (!bnx2x_vf_chk_avail_resc(bp, vf, resc)) {
@@ -2777,6 +2775,10 @@ int bnx2x_vf_init(struct bnx2x *bp, struct bnx2x_virtf *vf, dma_addr_t *sb_map)
 		   vf->abs_vfid, vf->state);
 		return -EINVAL;
 	}
+
+	/* let FLR complete ... */
+	msleep(100);
+
 	/* FLR cleanup epilogue */
 	if (bnx2x_vf_flr_clnup_epilog(bp, vf->abs_vfid))
 		return -EBUSY;
@@ -2890,7 +2892,7 @@ int bnx2x_vfop_close_cmd(struct bnx2x *bp,
 	return -ENOMEM;
 }
 
-/* VF release can be called either: 1. the VF was acquired but
+/* VF release can be called either: 1. The VF was acquired but
  * not enabled 2. the vf was enabled or in the process of being
  * enabled
  */
@@ -3024,7 +3026,6 @@ void bnx2x_unlock_vf_pf_channel(struct bnx2x *bp, struct bnx2x_virtf *vf,
 
 int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs_param)
 {
-
 	struct bnx2x *bp = netdev_priv(pci_get_drvdata(dev));
 
 	DP(BNX2X_MSG_IOV, "bnx2x_sriov_configure called with %d, BNX2X_NR_VIRTFN(bp) was %d\n",
@@ -3032,7 +3033,7 @@ int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs_param)
 
 	/* HW channel is only operational when PF is up */
 	if (bp->state != BNX2X_STATE_OPEN) {
-		BNX2X_ERR("VF num configurtion via sysfs not supported while PF is down");
+		BNX2X_ERR("VF num configuration via sysfs not supported while PF is down\n");
 		return -EINVAL;
 	}
 
@@ -3086,6 +3087,11 @@ void bnx2x_disable_sriov(struct bnx2x *bp)
 static int bnx2x_vf_ndo_sanity(struct bnx2x *bp, int vfidx,
 			       struct bnx2x_virtf *vf)
 {
+	if (bp->state != BNX2X_STATE_OPEN) {
+		BNX2X_ERR("vf ndo called though PF is down\n");
+		return -EINVAL;
+	}
+
 	if (!IS_SRIOV(bp)) {
 		BNX2X_ERR("vf ndo called though sriov is disabled\n");
 		return -EINVAL;
@@ -3141,7 +3147,7 @@ int bnx2x_get_vf_config(struct net_device *dev, int vfidx,
 			/* mac configured by ndo so its in bulletin board */
 			memcpy(&ivi->mac, bulletin->mac, ETH_ALEN);
 		else
-			/* funtion has not been loaded yet. Show mac as 0s */
+			/* function has not been loaded yet. Show mac as 0s */
 			memset(&ivi->mac, 0, ETH_ALEN);
 
 		/* vlan */
@@ -3149,7 +3155,7 @@ int bnx2x_get_vf_config(struct net_device *dev, int vfidx,
 			/* vlan configured by ndo so its in bulletin board */
 			memcpy(&ivi->vlan, &bulletin->vlan, VLAN_HLEN);
 		else
-			/* funtion has not been loaded yet. Show vlans as 0s */
+			/* function has not been loaded yet. Show vlans as 0s */
 			memset(&ivi->vlan, 0, VLAN_HLEN);
 	}
 
@@ -3189,7 +3195,7 @@ int bnx2x_set_vf_mac(struct net_device *dev, int vfidx, u8 *mac)
 		return -EINVAL;
 	}
 
-	/* update PF's copy of the VF's bulletin. will no longer accept mac
+	/* update PF's copy of the VF's bulletin. Will no longer accept mac
 	 * configuration requests from vf unless match this mac
 	 */
 	bulletin->valid_bitmap |= 1 << MAC_ADDR_VALID;
@@ -3358,8 +3364,11 @@ int bnx2x_set_vf_vlan(struct net_device *dev, int vfidx, u16 vlan, u8 qos)
 	return 0;
 }
 
-/* crc is the first field in the bulletin board. compute the crc over the
- * entire bulletin board excluding the crc field itself
+/* crc is the first field in the bulletin board. Compute the crc over the
+ * entire bulletin board excluding the crc field itself. Use the length field
+ * as the Bulletin Board was posted by a PF with possibly a different version
+ * from the vf which will sample it. Therefore, the length is computed by the
+ * PF and the used blindly by the VF.
  */
 u32 bnx2x_crc_vf_bulletin(struct bnx2x *bp,
 			  struct pf_vf_bulletin_content *bulletin)
@@ -3389,7 +3398,7 @@ enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp)
 			if (bulletin.crc == bnx2x_crc_vf_bulletin(bp,
 								  &bulletin))
 				break;
-			BNX2X_ERR("bad crc on bulletin board. contained %x computed %x\n",
+			BNX2X_ERR("bad crc on bulletin board. Contained %x computed %x\n",
 				  bulletin.crc,
 				  bnx2x_crc_vf_bulletin(bp, &bulletin));
 		}
@@ -3417,6 +3426,20 @@ enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp)
 	return PFVF_BULLETIN_UPDATED;
 }
 
+void bnx2x_timer_sriov(struct bnx2x *bp)
+{
+	bnx2x_sample_bulletin(bp);
+
+	/* if channel is down we need to self destruct */
+	if (bp->old_bulletin.valid_bitmap & 1 << CHANNEL_DOWN) {
+		smp_mb__before_clear_bit();
+		set_bit(BNX2X_SP_RTNL_VFPF_CHANNEL_DOWN,
+			&bp->sp_rtnl_state);
+		smp_mb__after_clear_bit();
+		schedule_delayed_work(&bp->sp_rtnl_task, 0);
+	}
+}
+
 void __iomem *bnx2x_vf_doorbells(struct bnx2x *bp)
 {
 	/* vf doorbells are embedded within the regview */
@@ -3452,7 +3475,7 @@ int bnx2x_open_epilog(struct bnx2x *bp)
 	 * register_netdevice which must have rtnl lock taken. As we are holding
 	 * the lock right now, that could only work if the probe would not take
 	 * the lock. However, as the probe of the vf may be called from other
-	 * contexts as well (such as passthrough to vm failes) it can't assume
+	 * contexts as well (such as passthrough to vm fails) it can't assume
 	 * the lock is being held for it. Using delayed work here allows the
 	 * probe code to simply take the lock (i.e. wait for it to be released
 	 * if it is being held). We only want to do this if the number of VFs
@@ -3467,3 +3490,23 @@ int bnx2x_open_epilog(struct bnx2x *bp)
 
 	return 0;
 }
+
+void bnx2x_iov_channel_down(struct bnx2x *bp)
+{
+	int vf_idx;
+	struct pf_vf_bulletin_content *bulletin;
+
+	if (!IS_SRIOV(bp))
+		return;
+
+	for_each_vf(bp, vf_idx) {
+		/* locate this VFs bulletin board and update the channel down
+		 * bit
+		 */
+		bulletin = BP_VF_BULLETIN(bp, vf_idx);
+		bulletin->valid_bitmap |= 1 << CHANNEL_DOWN;
+
+		/* update vf bulletin board */
+		bnx2x_post_vf_bulletin(bp, vf_idx);
+	}
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
index d67ddc5..d143a7c 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
@@ -197,7 +197,7 @@ struct bnx2x_virtf {
 
 	u8 state;
 #define VF_FREE		0	/* VF ready to be acquired holds no resc */
-#define VF_ACQUIRED	1	/* VF aquired, but not initalized */
+#define VF_ACQUIRED	1	/* VF acquired, but not initialized */
 #define VF_ENABLED	2	/* VF Enabled */
 #define VF_RESET	3	/* VF FLR'd, pending cleanup */
 
@@ -496,7 +496,7 @@ enum {
 		else if ((next) == VFOP_VERIFY_PEND)			\
 			BNX2X_ERR("expected pending\n");		\
 		else {							\
-			DP(BNX2X_MSG_IOV, "no ramrod. scheduling\n");	\
+			DP(BNX2X_MSG_IOV, "no ramrod. Scheduling\n");	\
 			atomic_set(&vf->op_in_progress, 1);		\
 			queue_delayed_work(bnx2x_wq, &bp->sp_task, 0);  \
 			return;						\
@@ -722,7 +722,6 @@ u32 bnx2x_crc_vf_bulletin(struct bnx2x *bp,
 			  struct pf_vf_bulletin_content *bulletin);
 int bnx2x_post_vf_bulletin(struct bnx2x *bp, int vf);
 
-
 enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp);
 
 /* VF side vfpf channel functions */
@@ -752,6 +751,7 @@ static inline int bnx2x_vf_ustorm_prods_offset(struct bnx2x *bp,
 }
 
 enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp);
+void bnx2x_timer_sriov(struct bnx2x *bp);
 void __iomem *bnx2x_vf_doorbells(struct bnx2x *bp);
 int bnx2x_vf_pci_alloc(struct bnx2x *bp);
 int bnx2x_enable_sriov(struct bnx2x *bp);
@@ -762,6 +762,7 @@ static inline int bnx2x_vf_headroom(struct bnx2x *bp)
 }
 void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp);
 int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs);
+void bnx2x_iov_channel_down(struct bnx2x *bp);
 int bnx2x_open_epilog(struct bnx2x *bp);
 
 #else /* CONFIG_BNX2X_SRIOV */
@@ -809,6 +810,7 @@ static inline enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp
 {
 	return PFVF_BULLETIN_UNCHANGED;
 }
+static inline void bnx2x_timer_sriov(struct bnx2x *bp) {}
 
 static inline void __iomem *bnx2x_vf_doorbells(struct bnx2x *bp)
 {
@@ -818,6 +820,7 @@ static inline void __iomem *bnx2x_vf_doorbells(struct bnx2x *bp)
 static inline int bnx2x_vf_pci_alloc(struct bnx2x *bp) {return 0; }
 static inline void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp) {}
 static inline int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs) {return 0; }
+static inline void bnx2x_iov_channel_down(struct bnx2x *bp) {}
 static inline int bnx2x_open_epilog(struct bnx2x *bp) {return 0; }
 
 #endif /* CONFIG_BNX2X_SRIOV */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.c
index 2ca3d94..98366ab 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.c
@@ -1002,7 +1002,6 @@ static int bnx2x_storm_stats_update(struct bnx2x *bp)
 		qstats->valid_bytes_received_lo =
 					qstats->total_bytes_received_lo;
 
-
 		UPDATE_EXTEND_TSTAT(rcv_ucast_pkts,
 					total_unicast_packets_received);
 		UPDATE_EXTEND_TSTAT(rcv_mcast_pkts,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.h
index d117f47..853824d 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_stats.h
@@ -40,7 +40,6 @@ struct nig_stats {
 	u32 egress_mac_pkt1_hi;
 };
 
-
 enum bnx2x_stats_event {
 	STATS_EVENT_PMF = 0,
 	STATS_EVENT_LINK_UP,
@@ -208,7 +207,6 @@ struct bnx2x_eth_stats {
 	u32 eee_tx_lpi;
 };
 
-
 struct bnx2x_eth_q_stats {
 	u32 total_unicast_bytes_received_hi;
 	u32 total_unicast_bytes_received_lo;
@@ -331,7 +329,6 @@ struct bnx2x_fw_port_stats_old {
 	 u32 mac_discard;
 };
 
-
 /****************************************************************************
 * Macros
 ****************************************************************************/
@@ -536,7 +533,6 @@ struct bnx2x_fw_port_stats_old {
 		SUB_EXTEND_64(qstats->t##_hi, qstats->t##_lo, diff); \
 	} while (0)
 
-
 /* forward */
 struct bnx2x;
 
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c
index 928b074..2088063 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c
@@ -113,7 +113,7 @@ static int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping)
 {
 	struct cstorm_vf_zone_data __iomem *zone_data =
 		REG_ADDR(bp, PXP_VF_ADDR_CSDM_GLOBAL_START);
-	int tout = 600, interval = 100; /* wait for 60 seconds */
+	int tout = 100, interval = 100; /* wait for 10 seconds */
 
 	if (*done) {
 		BNX2X_ERR("done was non zero before message to pf was sent\n");
@@ -121,6 +121,16 @@ static int bnx2x_send_msg2pf(struct bnx2x *bp, u8 *done, dma_addr_t msg_mapping)
 		return -EINVAL;
 	}
 
+	/* if PF indicated channel is down avoid sending message. Return success
+	 * so calling flow can continue
+	 */
+	bnx2x_sample_bulletin(bp);
+	if (bp->old_bulletin.valid_bitmap & 1 << CHANNEL_DOWN) {
+		DP(BNX2X_MSG_IOV, "detecting channel down. Aborting message\n");
+		*done = PFVF_STATUS_SUCCESS;
+		return 0;
+	}
+
 	/* Write message address */
 	writel(U64_LO(msg_mapping),
 	       &zone_data->non_trigger.vf_pf_channel.msg_addr_lo);
@@ -233,7 +243,7 @@ int bnx2x_vfpf_acquire(struct bnx2x *bp, u8 tx_count, u8 rx_count)
 
 		attempts++;
 
-		/* test whether the PF accepted our request. If not, humble the
+		/* test whether the PF accepted our request. If not, humble
 		 * the request and try again.
 		 */
 		if (bp->acquire_resp.hdr.status == PFVF_STATUS_SUCCESS) {
@@ -333,7 +343,7 @@ int bnx2x_vfpf_release(struct bnx2x *bp)
 		DP(BNX2X_MSG_SP, "vf released\n");
 	} else {
 		/* PF reports error */
-		BNX2X_ERR("PF failed our release request - are we out of sync? response status: %d\n",
+		BNX2X_ERR("PF failed our release request - are we out of sync? Response status: %d\n",
 			  resp->hdr.status);
 		rc = -EAGAIN;
 		goto out;
@@ -787,7 +797,7 @@ static inline void bnx2x_set_vf_mbxs_valid(struct bnx2x *bp)
 		storm_memset_vf_mbx_valid(bp, bnx2x_vf(bp, i, abs_vfid));
 }
 
-/* enable vf_pf mailbox (aka vf-pf-chanell) */
+/* enable vf_pf mailbox (aka vf-pf-channel) */
 void bnx2x_vf_enable_mbx(struct bnx2x *bp, u8 abs_vfid)
 {
 	bnx2x_vf_flr_clnup_epilog(bp, abs_vfid);
@@ -844,7 +854,6 @@ static int bnx2x_copy32_vf_dmae(struct bnx2x *bp, u8 from_vf,
 		dmae.dst_addr_hi = vf_addr_hi;
 	}
 	dmae.len = len32;
-	bnx2x_dp_dmae(bp, &dmae, BNX2X_MSG_DMAE);
 
 	/* issue the command and wait for completion */
 	return bnx2x_issue_dmae_with_comp(bp, &dmae);
@@ -1072,7 +1081,7 @@ static void bnx2x_vf_mbx_set_q_flags(struct bnx2x *bp, u32 mbx_q_flags,
 	if (mbx_q_flags & VFPF_QUEUE_FLG_DHC)
 		__set_bit(BNX2X_Q_FLG_DHC, sp_q_flags);
 
-	/* outer vlan removal is set according to the PF's multi fuction mode */
+	/* outer vlan removal is set according to PF's multi function mode */
 	if (IS_MF_SD(bp))
 		__set_bit(BNX2X_Q_FLG_OV, sp_q_flags);
 }
@@ -1104,7 +1113,7 @@ static void bnx2x_vf_mbx_setup_q(struct bnx2x *bp, struct bnx2x_virtf *vf,
 		struct bnx2x_queue_init_params *init_p;
 		struct bnx2x_queue_setup_params *setup_p;
 
-		/* reinit the VF operation context */
+		/* re-init the VF operation context */
 		memset(&vf->op_params.qctor, 0 , sizeof(vf->op_params.qctor));
 		setup_p = &vf->op_params.qctor.prep_qsetup;
 		init_p =  &vf->op_params.qctor.qstate.params.init;
@@ -1588,8 +1597,9 @@ static void bnx2x_vf_mbx_request(struct bnx2x *bp, struct bnx2x_virtf *vf,
 		 * support them. Or this may be because someone wrote a crappy
 		 * VF driver and is sending garbage over the channel.
 		 */
-		BNX2X_ERR("unknown TLV. type %d length %d. first 20 bytes of mailbox buffer:\n",
-			  mbx->first_tlv.tl.type, mbx->first_tlv.tl.length);
+		BNX2X_ERR("unknown TLV. type %d length %d vf->state was %d. first 20 bytes of mailbox buffer:\n",
+			  mbx->first_tlv.tl.type, mbx->first_tlv.tl.length,
+			  vf->state);
 		for (i = 0; i < 20; i++)
 			DP_CONT(BNX2X_MSG_IOV, "%x ",
 				mbx->msg->req.tlv_buf_size.tlv_buffer[i]);
@@ -1605,8 +1615,11 @@ static void bnx2x_vf_mbx_request(struct bnx2x *bp, struct bnx2x_virtf *vf,
 			bnx2x_vf_mbx_resp(bp, vf);
 		} else {
 			/* can't send a response since this VF is unknown to us
-			 * just unlock the channel and be done with.
+			 * just ack the FW to release the mailbox and unlock
+			 * the channel.
 			 */
+			storm_memset_vf_mbx_ack(bp, vf->abs_vfid);
+			mmiowb();
 			bnx2x_unlock_vf_pf_channel(bp, vf,
 						   mbx->first_tlv.tl.type);
 		}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index 41708fa..f3ad174 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -331,7 +331,10 @@ struct pf_vf_bulletin_content {
 #define VLAN_VALID		1	/* when set, the vf should not access
 					 * the vfpf channel
 					 */
-
+#define CHANNEL_DOWN		2	/* vfpf channel is disabled. VFs are not
+					 * to attempt to send messages on the
+					 * channel after this bit is set
+					 */
 	u8 mac[ETH_ALEN];
 	u8 mac_padding[2];
 
diff --git a/drivers/net/ethernet/broadcom/cnic.c b/drivers/net/ethernet/broadcom/cnic.c
index 6b0dc13..d78d4cf 100644
--- a/drivers/net/ethernet/broadcom/cnic.c
+++ b/drivers/net/ethernet/broadcom/cnic.c
@@ -5622,7 +5622,7 @@ static void cnic_rcv_netevent(struct cnic_local *cp, unsigned long event,
 static int cnic_netdev_event(struct notifier_block *this, unsigned long event,
 							 void *ptr)
 {
-	struct net_device *netdev = ptr;
+	struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
 	struct cnic_dev *dev;
 	int new_dev = 0;
 
diff --git a/drivers/net/ethernet/broadcom/sb1250-mac.c b/drivers/net/ethernet/broadcom/sb1250-mac.c
index e80bfb6..c277771 100644
--- a/drivers/net/ethernet/broadcom/sb1250-mac.c
+++ b/drivers/net/ethernet/broadcom/sb1250-mac.c
@@ -2197,7 +2197,7 @@ static const struct net_device_ops sbmac_netdev_ops = {
 
 static int sbmac_init(struct platform_device *pldev, long long base)
 {
-	struct net_device *dev = dev_get_drvdata(&pldev->dev);
+	struct net_device *dev = platform_get_drvdata(pldev);
 	int idx = pldev->id;
 	struct sbmac_softc *sc = netdev_priv(dev);
 	unsigned char *eaddr;
@@ -2275,7 +2275,7 @@ static int sbmac_init(struct platform_device *pldev, long long base)
 		       dev->name);
 		goto free_mdio;
 	}
-	dev_set_drvdata(&pldev->dev, sc->mii_bus);
+	platform_set_drvdata(pldev, sc->mii_bus);
 
 	err = register_netdev(dev);
 	if (err) {
@@ -2300,7 +2300,6 @@ static int sbmac_init(struct platform_device *pldev, long long base)
 	return 0;
 unreg_mdio:
 	mdiobus_unregister(sc->mii_bus);
-	dev_set_drvdata(&pldev->dev, NULL);
 free_mdio:
 	mdiobus_free(sc->mii_bus);
 uninit_ctx:
@@ -2624,7 +2623,7 @@ static int sbmac_probe(struct platform_device *pldev)
 		goto out_unmap;
 	}
 
-	dev_set_drvdata(&pldev->dev, dev);
+	platform_set_drvdata(pldev, dev);
 	SET_NETDEV_DEV(dev, &pldev->dev);
 
 	sc = netdev_priv(dev);
@@ -2649,7 +2648,7 @@ out_out:
 
 static int __exit sbmac_remove(struct platform_device *pldev)
 {
-	struct net_device *dev = dev_get_drvdata(&pldev->dev);
+	struct net_device *dev = platform_get_drvdata(pldev);
 	struct sbmac_softc *sc = netdev_priv(dev);
 
 	unregister_netdev(dev);
diff --git a/drivers/net/ethernet/broadcom/tg3.c b/drivers/net/ethernet/broadcom/tg3.c
index a13463e..d964f30 100644
--- a/drivers/net/ethernet/broadcom/tg3.c
+++ b/drivers/net/ethernet/broadcom/tg3.c
@@ -968,9 +968,6 @@ static void tg3_ape_driver_state_change(struct tg3 *tp, int kind)
 
 		event = APE_EVENT_STATUS_STATE_UNLOAD;
 		break;
-	case RESET_KIND_SUSPEND:
-		event = APE_EVENT_STATUS_STATE_SUSPEND;
-		break;
 	default:
 		return;
 	}
@@ -1317,8 +1314,8 @@ static int tg3_phy_toggle_auxctl_smdsp(struct tg3 *tp, bool enable)
 
 	if (err)
 		return err;
-	if (enable)
 
+	if (enable)
 		val |= MII_TG3_AUXCTL_ACTL_SMDSP_ENA;
 	else
 		val &= ~MII_TG3_AUXCTL_ACTL_SMDSP_ENA;
@@ -1745,10 +1742,6 @@ static void tg3_write_sig_pre_reset(struct tg3 *tp, int kind)
 			break;
 		}
 	}
-
-	if (kind == RESET_KIND_INIT ||
-	    kind == RESET_KIND_SUSPEND)
-		tg3_ape_driver_state_change(tp, kind);
 }
 
 /* tp->lock is held. */
@@ -1770,9 +1763,6 @@ static void tg3_write_sig_post_reset(struct tg3 *tp, int kind)
 			break;
 		}
 	}
-
-	if (kind == RESET_KIND_SHUTDOWN)
-		tg3_ape_driver_state_change(tp, kind);
 }
 
 /* tp->lock is held. */
@@ -2341,6 +2331,46 @@ static void tg3_phy_apply_otp(struct tg3 *tp)
 	tg3_phy_toggle_auxctl_smdsp(tp, false);
 }
 
+static void tg3_eee_pull_config(struct tg3 *tp, struct ethtool_eee *eee)
+{
+	u32 val;
+	struct ethtool_eee *dest = &tp->eee;
+
+	if (!(tp->phy_flags & TG3_PHYFLG_EEE_CAP))
+		return;
+
+	if (eee)
+		dest = eee;
+
+	if (tg3_phy_cl45_read(tp, MDIO_MMD_AN, TG3_CL45_D7_EEERES_STAT, &val))
+		return;
+
+	/* Pull eee_active */
+	if (val == TG3_CL45_D7_EEERES_STAT_LP_1000T ||
+	    val == TG3_CL45_D7_EEERES_STAT_LP_100TX) {
+		dest->eee_active = 1;
+	} else
+		dest->eee_active = 0;
+
+	/* Pull lp advertised settings */
+	if (tg3_phy_cl45_read(tp, MDIO_MMD_AN, MDIO_AN_EEE_LPABLE, &val))
+		return;
+	dest->lp_advertised = mmd_eee_adv_to_ethtool_adv_t(val);
+
+	/* Pull advertised and eee_enabled settings */
+	if (tg3_phy_cl45_read(tp, MDIO_MMD_AN, MDIO_AN_EEE_ADV, &val))
+		return;
+	dest->eee_enabled = !!val;
+	dest->advertised = mmd_eee_adv_to_ethtool_adv_t(val);
+
+	/* Pull tx_lpi_enabled */
+	val = tr32(TG3_CPMU_EEE_MODE);
+	dest->tx_lpi_enabled = !!(val & TG3_CPMU_EEEMD_LPI_IN_TX);
+
+	/* Pull lpi timer value */
+	dest->tx_lpi_timer = tr32(TG3_CPMU_EEE_DBTMR1) & 0xffff;
+}
+
 static void tg3_phy_eee_adjust(struct tg3 *tp, bool current_link_up)
 {
 	u32 val;
@@ -2364,11 +2394,8 @@ static void tg3_phy_eee_adjust(struct tg3 *tp, bool current_link_up)
 
 		tw32(TG3_CPMU_EEE_CTRL, eeectl);
 
-		tg3_phy_cl45_read(tp, MDIO_MMD_AN,
-				  TG3_CL45_D7_EEERES_STAT, &val);
-
-		if (val == TG3_CL45_D7_EEERES_STAT_LP_1000T ||
-		    val == TG3_CL45_D7_EEERES_STAT_LP_100TX)
+		tg3_eee_pull_config(tp, NULL);
+		if (tp->eee.eee_active)
 			tp->setlpicnt = 2;
 	}
 
@@ -4192,6 +4219,8 @@ static int tg3_power_down_prepare(struct tg3 *tp)
 
 	tg3_write_sig_post_reset(tp, RESET_KIND_SHUTDOWN);
 
+	tg3_ape_driver_state_change(tp, RESET_KIND_SHUTDOWN);
+
 	return 0;
 }
 
@@ -4292,6 +4321,16 @@ static int tg3_phy_autoneg_cfg(struct tg3 *tp, u32 advertise, u32 flowctrl)
 		/* Advertise 1000-BaseT EEE ability */
 		if (advertise & ADVERTISED_1000baseT_Full)
 			val |= MDIO_AN_EEE_ADV_1000T;
+
+		if (!tp->eee.eee_enabled) {
+			val = 0;
+			tp->eee.advertised = 0;
+		} else {
+			tp->eee.advertised = advertise &
+					     (ADVERTISED_100baseT_Full |
+					      ADVERTISED_1000baseT_Full);
+		}
+
 		err = tg3_phy_cl45_write(tp, MDIO_MMD_AN, MDIO_AN_EEE_ADV, val);
 		if (err)
 			val = 0;
@@ -4536,26 +4575,23 @@ static int tg3_init_5401phy_dsp(struct tg3 *tp)
 
 static bool tg3_phy_eee_config_ok(struct tg3 *tp)
 {
-	u32 val;
-	u32 tgtadv = 0;
-	u32 advertising = tp->link_config.advertising;
+	struct ethtool_eee eee;
 
 	if (!(tp->phy_flags & TG3_PHYFLG_EEE_CAP))
 		return true;
 
-	if (tg3_phy_cl45_read(tp, MDIO_MMD_AN, MDIO_AN_EEE_ADV, &val))
-		return false;
-
-	val &= (MDIO_AN_EEE_ADV_100TX | MDIO_AN_EEE_ADV_1000T);
-
-
-	if (advertising & ADVERTISED_100baseT_Full)
-		tgtadv |= MDIO_AN_EEE_ADV_100TX;
-	if (advertising & ADVERTISED_1000baseT_Full)
-		tgtadv |= MDIO_AN_EEE_ADV_1000T;
+	tg3_eee_pull_config(tp, &eee);
 
-	if (val != tgtadv)
-		return false;
+	if (tp->eee.eee_enabled) {
+		if (tp->eee.advertised != eee.advertised ||
+		    tp->eee.tx_lpi_timer != eee.tx_lpi_timer ||
+		    tp->eee.tx_lpi_enabled != eee.tx_lpi_enabled)
+			return false;
+	} else {
+		/* EEE is disabled but we're advertising */
+		if (eee.advertised)
+			return false;
+	}
 
 	return true;
 }
@@ -4656,6 +4692,42 @@ static void tg3_clear_mac_status(struct tg3 *tp)
 	udelay(40);
 }
 
+static void tg3_setup_eee(struct tg3 *tp)
+{
+	u32 val;
+
+	val = TG3_CPMU_EEE_LNKIDL_PCIE_NL0 |
+	      TG3_CPMU_EEE_LNKIDL_UART_IDL;
+	if (tg3_chip_rev_id(tp) == CHIPREV_ID_57765_A0)
+		val |= TG3_CPMU_EEE_LNKIDL_APE_TX_MT;
+
+	tw32_f(TG3_CPMU_EEE_LNKIDL_CTRL, val);
+
+	tw32_f(TG3_CPMU_EEE_CTRL,
+	       TG3_CPMU_EEE_CTRL_EXIT_20_1_US);
+
+	val = TG3_CPMU_EEEMD_ERLY_L1_XIT_DET |
+	      (tp->eee.tx_lpi_enabled ? TG3_CPMU_EEEMD_LPI_IN_TX : 0) |
+	      TG3_CPMU_EEEMD_LPI_IN_RX |
+	      TG3_CPMU_EEEMD_EEE_ENABLE;
+
+	if (tg3_asic_rev(tp) != ASIC_REV_5717)
+		val |= TG3_CPMU_EEEMD_SND_IDX_DET_EN;
+
+	if (tg3_flag(tp, ENABLE_APE))
+		val |= TG3_CPMU_EEEMD_APE_TX_DET_EN;
+
+	tw32_f(TG3_CPMU_EEE_MODE, tp->eee.eee_enabled ? val : 0);
+
+	tw32_f(TG3_CPMU_EEE_DBTMR1,
+	       TG3_CPMU_DBTMR1_PCIEXIT_2047US |
+	       (tp->eee.tx_lpi_timer & 0xffff));
+
+	tw32_f(TG3_CPMU_EEE_DBTMR2,
+	       TG3_CPMU_DBTMR2_APE_TX_2047US |
+	       TG3_CPMU_DBTMR2_TXIDXEQ_2047US);
+}
+
 static int tg3_setup_copper_phy(struct tg3 *tp, bool force_reset)
 {
 	bool current_link_up;
@@ -4822,8 +4894,10 @@ static int tg3_setup_copper_phy(struct tg3 *tp, bool force_reset)
 			 */
 			if (!eee_config_ok &&
 			    (tp->phy_flags & TG3_PHYFLG_KEEP_LINK_ON_PWRDN) &&
-			    !force_reset)
+			    !force_reset) {
+				tg3_setup_eee(tp);
 				tg3_phy_reset(tp);
+			}
 		} else {
 			if (!(bmcr & BMCR_ANENABLE) &&
 			    tp->link_config.speed == current_speed &&
@@ -6335,9 +6409,7 @@ static void tg3_tx_recover(struct tg3 *tp)
 		    "Please report the problem to the driver maintainer "
 		    "and include system chipset information.\n");
 
-	spin_lock(&tp->lock);
 	tg3_flag_set(tp, TX_RECOVERY_PENDING);
-	spin_unlock(&tp->lock);
 }
 
 static inline u32 tg3_tx_avail(struct tg3_napi *tnapi)
@@ -9205,11 +9277,9 @@ static void __tg3_set_coalesce(struct tg3 *tp, struct ethtool_coalesce *ec)
 }
 
 /* tp->lock is held. */
-static void tg3_rings_reset(struct tg3 *tp)
+static void tg3_tx_rcbs_disable(struct tg3 *tp)
 {
-	int i;
-	u32 stblk, txrcb, rxrcb, limit;
-	struct tg3_napi *tnapi = &tp->napi[0];
+	u32 txrcb, limit;
 
 	/* Disable all transmit rings but the first. */
 	if (!tg3_flag(tp, 5705_PLUS))
@@ -9226,7 +9296,33 @@ static void tg3_rings_reset(struct tg3 *tp)
 	     txrcb < limit; txrcb += TG3_BDINFO_SIZE)
 		tg3_write_mem(tp, txrcb + TG3_BDINFO_MAXLEN_FLAGS,
 			      BDINFO_FLAGS_DISABLED);
+}
+
+/* tp->lock is held. */
+static void tg3_tx_rcbs_init(struct tg3 *tp)
+{
+	int i = 0;
+	u32 txrcb = NIC_SRAM_SEND_RCB;
+
+	if (tg3_flag(tp, ENABLE_TSS))
+		i++;
+
+	for (; i < tp->irq_max; i++, txrcb += TG3_BDINFO_SIZE) {
+		struct tg3_napi *tnapi = &tp->napi[i];
+
+		if (!tnapi->tx_ring)
+			continue;
+
+		tg3_set_bdinfo(tp, txrcb, tnapi->tx_desc_mapping,
+			       (TG3_TX_RING_SIZE << BDINFO_FLAGS_MAXLEN_SHIFT),
+			       NIC_SRAM_TX_BUFFER_DESC);
+	}
+}
 
+/* tp->lock is held. */
+static void tg3_rx_ret_rcbs_disable(struct tg3 *tp)
+{
+	u32 rxrcb, limit;
 
 	/* Disable all receive return rings but the first. */
 	if (tg3_flag(tp, 5717_PLUS))
@@ -9244,6 +9340,39 @@ static void tg3_rings_reset(struct tg3 *tp)
 	     rxrcb < limit; rxrcb += TG3_BDINFO_SIZE)
 		tg3_write_mem(tp, rxrcb + TG3_BDINFO_MAXLEN_FLAGS,
 			      BDINFO_FLAGS_DISABLED);
+}
+
+/* tp->lock is held. */
+static void tg3_rx_ret_rcbs_init(struct tg3 *tp)
+{
+	int i = 0;
+	u32 rxrcb = NIC_SRAM_RCV_RET_RCB;
+
+	if (tg3_flag(tp, ENABLE_RSS))
+		i++;
+
+	for (; i < tp->irq_max; i++, rxrcb += TG3_BDINFO_SIZE) {
+		struct tg3_napi *tnapi = &tp->napi[i];
+
+		if (!tnapi->rx_rcb)
+			continue;
+
+		tg3_set_bdinfo(tp, rxrcb, tnapi->rx_rcb_mapping,
+			       (tp->rx_ret_ring_mask + 1) <<
+				BDINFO_FLAGS_MAXLEN_SHIFT, 0);
+	}
+}
+
+/* tp->lock is held. */
+static void tg3_rings_reset(struct tg3 *tp)
+{
+	int i;
+	u32 stblk;
+	struct tg3_napi *tnapi = &tp->napi[0];
+
+	tg3_tx_rcbs_disable(tp);
+
+	tg3_rx_ret_rcbs_disable(tp);
 
 	/* Disable interrupts */
 	tw32_mailbox_f(tp->napi[0].int_mbox, 1);
@@ -9280,9 +9409,6 @@ static void tg3_rings_reset(struct tg3 *tp)
 			tw32_tx_mbox(mbox + i * 8, 0);
 	}
 
-	txrcb = NIC_SRAM_SEND_RCB;
-	rxrcb = NIC_SRAM_RCV_RET_RCB;
-
 	/* Clear status block in ram. */
 	memset(tnapi->hw_status, 0, TG3_HW_STATUS_SIZE);
 
@@ -9292,46 +9418,20 @@ static void tg3_rings_reset(struct tg3 *tp)
 	tw32(HOSTCC_STATUS_BLK_HOST_ADDR + TG3_64BIT_REG_LOW,
 	     ((u64) tnapi->status_mapping & 0xffffffff));
 
-	if (tnapi->tx_ring) {
-		tg3_set_bdinfo(tp, txrcb, tnapi->tx_desc_mapping,
-			       (TG3_TX_RING_SIZE <<
-				BDINFO_FLAGS_MAXLEN_SHIFT),
-			       NIC_SRAM_TX_BUFFER_DESC);
-		txrcb += TG3_BDINFO_SIZE;
-	}
-
-	if (tnapi->rx_rcb) {
-		tg3_set_bdinfo(tp, rxrcb, tnapi->rx_rcb_mapping,
-			       (tp->rx_ret_ring_mask + 1) <<
-				BDINFO_FLAGS_MAXLEN_SHIFT, 0);
-		rxrcb += TG3_BDINFO_SIZE;
-	}
-
 	stblk = HOSTCC_STATBLCK_RING1;
 
 	for (i = 1, tnapi++; i < tp->irq_cnt; i++, tnapi++) {
 		u64 mapping = (u64)tnapi->status_mapping;
 		tw32(stblk + TG3_64BIT_REG_HIGH, mapping >> 32);
 		tw32(stblk + TG3_64BIT_REG_LOW, mapping & 0xffffffff);
+		stblk += 8;
 
 		/* Clear status block in ram. */
 		memset(tnapi->hw_status, 0, TG3_HW_STATUS_SIZE);
-
-		if (tnapi->tx_ring) {
-			tg3_set_bdinfo(tp, txrcb, tnapi->tx_desc_mapping,
-				       (TG3_TX_RING_SIZE <<
-					BDINFO_FLAGS_MAXLEN_SHIFT),
-				       NIC_SRAM_TX_BUFFER_DESC);
-			txrcb += TG3_BDINFO_SIZE;
-		}
-
-		tg3_set_bdinfo(tp, rxrcb, tnapi->rx_rcb_mapping,
-			       ((tp->rx_ret_ring_mask + 1) <<
-				BDINFO_FLAGS_MAXLEN_SHIFT), 0);
-
-		stblk += 8;
-		rxrcb += TG3_BDINFO_SIZE;
 	}
+
+	tg3_tx_rcbs_init(tp);
+	tg3_rx_ret_rcbs_init(tp);
 }
 
 static void tg3_setup_rxbd_thresholds(struct tg3 *tp)
@@ -9531,46 +9631,17 @@ static int tg3_reset_hw(struct tg3 *tp, bool reset_phy)
 	if (tg3_flag(tp, INIT_COMPLETE))
 		tg3_abort_hw(tp, 1);
 
-	/* Enable MAC control of LPI */
-	if (tp->phy_flags & TG3_PHYFLG_EEE_CAP) {
-		val = TG3_CPMU_EEE_LNKIDL_PCIE_NL0 |
-		      TG3_CPMU_EEE_LNKIDL_UART_IDL;
-		if (tg3_chip_rev_id(tp) == CHIPREV_ID_57765_A0)
-			val |= TG3_CPMU_EEE_LNKIDL_APE_TX_MT;
-
-		tw32_f(TG3_CPMU_EEE_LNKIDL_CTRL, val);
-
-		tw32_f(TG3_CPMU_EEE_CTRL,
-		       TG3_CPMU_EEE_CTRL_EXIT_20_1_US);
-
-		val = TG3_CPMU_EEEMD_ERLY_L1_XIT_DET |
-		      TG3_CPMU_EEEMD_LPI_IN_TX |
-		      TG3_CPMU_EEEMD_LPI_IN_RX |
-		      TG3_CPMU_EEEMD_EEE_ENABLE;
-
-		if (tg3_asic_rev(tp) != ASIC_REV_5717)
-			val |= TG3_CPMU_EEEMD_SND_IDX_DET_EN;
-
-		if (tg3_flag(tp, ENABLE_APE))
-			val |= TG3_CPMU_EEEMD_APE_TX_DET_EN;
-
-		tw32_f(TG3_CPMU_EEE_MODE, val);
-
-		tw32_f(TG3_CPMU_EEE_DBTMR1,
-		       TG3_CPMU_DBTMR1_PCIEXIT_2047US |
-		       TG3_CPMU_DBTMR1_LNKIDLE_2047US);
-
-		tw32_f(TG3_CPMU_EEE_DBTMR2,
-		       TG3_CPMU_DBTMR2_APE_TX_2047US |
-		       TG3_CPMU_DBTMR2_TXIDXEQ_2047US);
-	}
-
 	if ((tp->phy_flags & TG3_PHYFLG_KEEP_LINK_ON_PWRDN) &&
 	    !(tp->phy_flags & TG3_PHYFLG_USER_CONFIGURED)) {
 		tg3_phy_pull_config(tp);
+		tg3_eee_pull_config(tp, NULL);
 		tp->phy_flags |= TG3_PHYFLG_USER_CONFIGURED;
 	}
 
+	/* Enable MAC control of LPI */
+	if (tp->phy_flags & TG3_PHYFLG_EEE_CAP)
+		tg3_setup_eee(tp);
+
 	if (reset_phy)
 		tg3_phy_reset(tp);
 
@@ -11226,7 +11297,7 @@ static int tg3_start(struct tg3 *tp, bool reset_phy, bool test_irq,
 	 */
 	err = tg3_alloc_consistent(tp);
 	if (err)
-		goto err_out1;
+		goto out_ints_fini;
 
 	tg3_napi_init(tp);
 
@@ -11240,12 +11311,15 @@ static int tg3_start(struct tg3 *tp, bool reset_phy, bool test_irq,
 				tnapi = &tp->napi[i];
 				free_irq(tnapi->irq_vec, tnapi);
 			}
-			goto err_out2;
+			goto out_napi_fini;
 		}
 	}
 
 	tg3_full_lock(tp, 0);
 
+	if (init)
+		tg3_ape_driver_state_change(tp, RESET_KIND_INIT);
+
 	err = tg3_init_hw(tp, reset_phy);
 	if (err) {
 		tg3_halt(tp, RESET_KIND_SHUTDOWN, 1);
@@ -11255,7 +11329,7 @@ static int tg3_start(struct tg3 *tp, bool reset_phy, bool test_irq,
 	tg3_full_unlock(tp);
 
 	if (err)
-		goto err_out3;
+		goto out_free_irq;
 
 	if (test_irq && tg3_flag(tp, USING_MSI)) {
 		err = tg3_test_msi(tp);
@@ -11266,7 +11340,7 @@ static int tg3_start(struct tg3 *tp, bool reset_phy, bool test_irq,
 			tg3_free_rings(tp);
 			tg3_full_unlock(tp);
 
-			goto err_out2;
+			goto out_napi_fini;
 		}
 
 		if (!tg3_flag(tp, 57765_PLUS) && tg3_flag(tp, USING_MSI)) {
@@ -11306,18 +11380,18 @@ static int tg3_start(struct tg3 *tp, bool reset_phy, bool test_irq,
 
 	return 0;
 
-err_out3:
+out_free_irq:
 	for (i = tp->irq_cnt - 1; i >= 0; i--) {
 		struct tg3_napi *tnapi = &tp->napi[i];
 		free_irq(tnapi->irq_vec, tnapi);
 	}
 
-err_out2:
+out_napi_fini:
 	tg3_napi_disable(tp);
 	tg3_napi_fini(tp);
 	tg3_free_consistent(tp);
 
-err_out1:
+out_ints_fini:
 	tg3_ints_fini(tp);
 
 	return err;
@@ -13362,11 +13436,13 @@ static void tg3_self_test(struct net_device *dev, struct ethtool_test *etest,
 	struct tg3 *tp = netdev_priv(dev);
 	bool doextlpbk = etest->flags & ETH_TEST_FL_EXTERNAL_LB;
 
-	if ((tp->phy_flags & TG3_PHYFLG_IS_LOW_POWER) &&
-	    tg3_power_up(tp)) {
-		etest->flags |= ETH_TEST_FL_FAILED;
-		memset(data, 1, sizeof(u64) * TG3_NUM_TEST);
-		return;
+	if (tp->phy_flags & TG3_PHYFLG_IS_LOW_POWER) {
+		if (tg3_power_up(tp)) {
+			etest->flags |= ETH_TEST_FL_FAILED;
+			memset(data, 1, sizeof(u64) * TG3_NUM_TEST);
+			return;
+		}
+		tg3_ape_driver_state_change(tp, RESET_KIND_INIT);
 	}
 
 	memset(data, 0, sizeof(u64) * TG3_NUM_TEST);
@@ -13657,6 +13733,57 @@ static int tg3_set_coalesce(struct net_device *dev, struct ethtool_coalesce *ec)
 	return 0;
 }
 
+static int tg3_set_eee(struct net_device *dev, struct ethtool_eee *edata)
+{
+	struct tg3 *tp = netdev_priv(dev);
+
+	if (!(tp->phy_flags & TG3_PHYFLG_EEE_CAP)) {
+		netdev_warn(tp->dev, "Board does not support EEE!\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (edata->advertised != tp->eee.advertised) {
+		netdev_warn(tp->dev,
+			    "Direct manipulation of EEE advertisement is not supported\n");
+		return -EINVAL;
+	}
+
+	if (edata->tx_lpi_timer > TG3_CPMU_DBTMR1_LNKIDLE_MAX) {
+		netdev_warn(tp->dev,
+			    "Maximal Tx Lpi timer supported is %#x(u)\n",
+			    TG3_CPMU_DBTMR1_LNKIDLE_MAX);
+		return -EINVAL;
+	}
+
+	tp->eee = *edata;
+
+	tp->phy_flags |= TG3_PHYFLG_USER_CONFIGURED;
+	tg3_warn_mgmt_link_flap(tp);
+
+	if (netif_running(tp->dev)) {
+		tg3_full_lock(tp, 0);
+		tg3_setup_eee(tp);
+		tg3_phy_reset(tp);
+		tg3_full_unlock(tp);
+	}
+
+	return 0;
+}
+
+static int tg3_get_eee(struct net_device *dev, struct ethtool_eee *edata)
+{
+	struct tg3 *tp = netdev_priv(dev);
+
+	if (!(tp->phy_flags & TG3_PHYFLG_EEE_CAP)) {
+		netdev_warn(tp->dev,
+			    "Board does not support EEE!\n");
+		return -EOPNOTSUPP;
+	}
+
+	*edata = tp->eee;
+	return 0;
+}
+
 static const struct ethtool_ops tg3_ethtool_ops = {
 	.get_settings		= tg3_get_settings,
 	.set_settings		= tg3_set_settings,
@@ -13690,6 +13817,8 @@ static const struct ethtool_ops tg3_ethtool_ops = {
 	.get_channels		= tg3_get_channels,
 	.set_channels		= tg3_set_channels,
 	.get_ts_info		= tg3_get_ts_info,
+	.get_eee		= tg3_get_eee,
+	.set_eee		= tg3_set_eee,
 };
 
 static struct rtnl_link_stats64 *tg3_get_stats64(struct net_device *dev,
@@ -15038,9 +15167,18 @@ static int tg3_phy_probe(struct tg3 *tp)
 	     (tg3_asic_rev(tp) == ASIC_REV_5717 &&
 	      tg3_chip_rev_id(tp) != CHIPREV_ID_5717_A0) ||
 	     (tg3_asic_rev(tp) == ASIC_REV_57765 &&
-	      tg3_chip_rev_id(tp) != CHIPREV_ID_57765_A0)))
+	      tg3_chip_rev_id(tp) != CHIPREV_ID_57765_A0))) {
 		tp->phy_flags |= TG3_PHYFLG_EEE_CAP;
 
+		tp->eee.supported = SUPPORTED_100baseT_Full |
+				    SUPPORTED_1000baseT_Full;
+		tp->eee.advertised = ADVERTISED_100baseT_Full |
+				     ADVERTISED_1000baseT_Full;
+		tp->eee.eee_enabled = 1;
+		tp->eee.tx_lpi_enabled = 1;
+		tp->eee.tx_lpi_timer = TG3_CPMU_DBTMR1_LNKIDLE_2047US;
+	}
+
 	tg3_phy_init_link_config(tp);
 
 	if (!(tp->phy_flags & TG3_PHYFLG_KEEP_LINK_ON_PWRDN) &&
@@ -17112,7 +17250,7 @@ static int tg3_init_one(struct pci_dev *pdev,
 {
 	struct net_device *dev;
 	struct tg3 *tp;
-	int i, err, pm_cap;
+	int i, err;
 	u32 sndmbx, rcvmbx, intmbx;
 	char str[40];
 	u64 dma_mask, persist_dma_mask;
@@ -17134,25 +17272,10 @@ static int tg3_init_one(struct pci_dev *pdev,
 
 	pci_set_master(pdev);
 
-	/* Find power-management capability. */
-	pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM);
-	if (pm_cap == 0) {
-		dev_err(&pdev->dev,
-			"Cannot find Power Management capability, aborting\n");
-		err = -EIO;
-		goto err_out_free_res;
-	}
-
-	err = pci_set_power_state(pdev, PCI_D0);
-	if (err) {
-		dev_err(&pdev->dev, "Transition to D0 failed, aborting\n");
-		goto err_out_free_res;
-	}
-
 	dev = alloc_etherdev_mq(sizeof(*tp), TG3_IRQ_MAX_VECS);
 	if (!dev) {
 		err = -ENOMEM;
-		goto err_out_power_down;
+		goto err_out_free_res;
 	}
 
 	SET_NETDEV_DEV(dev, &pdev->dev);
@@ -17160,7 +17283,7 @@ static int tg3_init_one(struct pci_dev *pdev,
 	tp = netdev_priv(dev);
 	tp->pdev = pdev;
 	tp->dev = dev;
-	tp->pm_cap = pm_cap;
+	tp->pm_cap = pdev->pm_cap;
 	tp->rx_mode = TG3_DEF_RX_MODE;
 	tp->tx_mode = TG3_DEF_TX_MODE;
 	tp->irq_sync = 1;
@@ -17498,9 +17621,6 @@ err_out_iounmap:
 err_out_free_dev:
 	free_netdev(dev);
 
-err_out_power_down:
-	pci_set_power_state(pdev, PCI_D3hot);
-
 err_out_free_res:
 	pci_release_regions(pdev);
 
@@ -17610,6 +17730,8 @@ static int tg3_resume(struct device *device)
 
 	tg3_full_lock(tp, 0);
 
+	tg3_ape_driver_state_change(tp, RESET_KIND_INIT);
+
 	tg3_flag_set(tp, INIT_COMPLETE);
 	err = tg3_restart_hw(tp,
 			     !(tp->phy_flags & TG3_PHYFLG_KEEP_LINK_ON_PWRDN));
@@ -17671,10 +17793,13 @@ static pci_ers_result_t tg3_io_error_detected(struct pci_dev *pdev,
 	tg3_full_unlock(tp);
 
 done:
-	if (state == pci_channel_io_perm_failure)
+	if (state == pci_channel_io_perm_failure) {
+		tg3_napi_enable(tp);
+		dev_close(netdev);
 		err = PCI_ERS_RESULT_DISCONNECT;
-	else
+	} else {
 		pci_disable_device(pdev);
+	}
 
 	rtnl_unlock();
 
@@ -17720,6 +17845,10 @@ static pci_ers_result_t tg3_io_slot_reset(struct pci_dev *pdev)
 	rc = PCI_ERS_RESULT_RECOVERED;
 
 done:
+	if (rc != PCI_ERS_RESULT_RECOVERED && netif_running(netdev)) {
+		tg3_napi_enable(tp);
+		dev_close(netdev);
+	}
 	rtnl_unlock();
 
 	return rc;
@@ -17744,6 +17873,7 @@ static void tg3_io_resume(struct pci_dev *pdev)
 		goto done;
 
 	tg3_full_lock(tp, 0);
+	tg3_ape_driver_state_change(tp, RESET_KIND_INIT);
 	tg3_flag_set(tp, INIT_COMPLETE);
 	err = tg3_restart_hw(tp, true);
 	if (err) {
@@ -17781,15 +17911,4 @@ static struct pci_driver tg3_driver = {
 	.driver.pm	= &tg3_pm_ops,
 };
 
-static int __init tg3_init(void)
-{
-	return pci_register_driver(&tg3_driver);
-}
-
-static void __exit tg3_cleanup(void)
-{
-	pci_unregister_driver(&tg3_driver);
-}
-
-module_init(tg3_init);
-module_exit(tg3_cleanup);
+module_pci_driver(tg3_driver);
diff --git a/drivers/net/ethernet/broadcom/tg3.h b/drivers/net/ethernet/broadcom/tg3.h
index ff6e30e..cd63d11 100644
--- a/drivers/net/ethernet/broadcom/tg3.h
+++ b/drivers/net/ethernet/broadcom/tg3.h
@@ -1175,6 +1175,7 @@
 #define TG3_CPMU_EEE_DBTMR1		0x000036b4
 #define  TG3_CPMU_DBTMR1_PCIEXIT_2047US	 0x07ff0000
 #define  TG3_CPMU_DBTMR1_LNKIDLE_2047US	 0x000007ff
+#define  TG3_CPMU_DBTMR1_LNKIDLE_MAX	 0x0000ffff
 #define TG3_CPMU_EEE_DBTMR2		0x000036b8
 #define  TG3_CPMU_DBTMR2_APE_TX_2047US	 0x07ff0000
 #define  TG3_CPMU_DBTMR2_TXIDXEQ_2047US	 0x000007ff
@@ -3372,6 +3373,7 @@ struct tg3 {
 	unsigned int			irq_cnt;
 
 	struct ethtool_coalesce		coal;
+	struct ethtool_eee		eee;
 
 	/* firmware info */
 	const char			*fw_needed;
diff --git a/drivers/net/ethernet/brocade/bna/bfa_defs.h b/drivers/net/ethernet/brocade/bna/bfa_defs.h
index e423f82..b7d8127 100644
--- a/drivers/net/ethernet/brocade/bna/bfa_defs.h
+++ b/drivers/net/ethernet/brocade/bna/bfa_defs.h
@@ -164,7 +164,8 @@ struct bfa_ioc_attr {
 	u8				port_mode;	/*!< enum bfa_mode */
 	u8				cap_bm;		/*!< capability */
 	u8				port_mode_cfg;	/*!< enum bfa_mode */
-	u8				rsvd[4];	/*!< 64bit align */
+	u8				def_fn;		/*!< 1 if default fn */
+	u8				rsvd[3];	/*!< 64bit align */
 };
 
 /* Adapter capability mask definition */
diff --git a/drivers/net/ethernet/brocade/bna/bfa_ioc.c b/drivers/net/ethernet/brocade/bna/bfa_ioc.c
index f2b73ff..6f3cac0 100644
--- a/drivers/net/ethernet/brocade/bna/bfa_ioc.c
+++ b/drivers/net/ethernet/brocade/bna/bfa_ioc.c
@@ -2371,7 +2371,7 @@ bfa_nw_ioc_get_attr(struct bfa_ioc *ioc, struct bfa_ioc_attr *ioc_attr)
 	memset((void *)ioc_attr, 0, sizeof(struct bfa_ioc_attr));
 
 	ioc_attr->state = bfa_ioc_get_state(ioc);
-	ioc_attr->port_id = ioc->port_id;
+	ioc_attr->port_id = bfa_ioc_portid(ioc);
 	ioc_attr->port_mode = ioc->port_mode;
 
 	ioc_attr->port_mode_cfg = ioc->port_mode_cfg;
@@ -2381,8 +2381,9 @@ bfa_nw_ioc_get_attr(struct bfa_ioc *ioc, struct bfa_ioc_attr *ioc_attr)
 
 	bfa_ioc_get_adapter_attr(ioc, &ioc_attr->adapter_attr);
 
-	ioc_attr->pci_attr.device_id = ioc->pcidev.device_id;
-	ioc_attr->pci_attr.pcifn = ioc->pcidev.pci_func;
+	ioc_attr->pci_attr.device_id = bfa_ioc_devid(ioc);
+	ioc_attr->pci_attr.pcifn = bfa_ioc_pcifn(ioc);
+	ioc_attr->def_fn = bfa_ioc_is_default(ioc);
 	bfa_ioc_get_pci_chip_rev(ioc, ioc_attr->pci_attr.chip_rev);
 }
 
diff --git a/drivers/net/ethernet/brocade/bna/bfa_ioc.h b/drivers/net/ethernet/brocade/bna/bfa_ioc.h
index 63a85e5..f04e0aa 100644
--- a/drivers/net/ethernet/brocade/bna/bfa_ioc.h
+++ b/drivers/net/ethernet/brocade/bna/bfa_ioc.h
@@ -222,6 +222,8 @@ struct bfa_ioc_hwif {
 #define bfa_ioc_bar0(__ioc)		((__ioc)->pcidev.pci_bar_kva)
 #define bfa_ioc_portid(__ioc)		((__ioc)->port_id)
 #define bfa_ioc_asic_gen(__ioc)		((__ioc)->asic_gen)
+#define bfa_ioc_is_default(__ioc)	\
+	(bfa_ioc_pcifn(__ioc) == bfa_ioc_portid(__ioc))
 #define bfa_ioc_fetch_stats(__ioc, __stats) \
 		(((__stats)->drv_stats) = (__ioc)->stats)
 #define bfa_ioc_clr_stats(__ioc)	\
diff --git a/drivers/net/ethernet/brocade/bna/bna.h b/drivers/net/ethernet/brocade/bna/bna.h
index 25dae75..f1eafc4 100644
--- a/drivers/net/ethernet/brocade/bna/bna.h
+++ b/drivers/net/ethernet/brocade/bna/bna.h
@@ -455,6 +455,8 @@ void bna_bfi_rx_enet_stop_rsp(struct bna_rx *rx,
 void bna_bfi_rxf_cfg_rsp(struct bna_rxf *rxf, struct bfi_msgq_mhdr *msghdr);
 void bna_bfi_rxf_mcast_add_rsp(struct bna_rxf *rxf,
 			       struct bfi_msgq_mhdr *msghdr);
+void bna_bfi_rxf_ucast_set_rsp(struct bna_rxf *rxf,
+			       struct bfi_msgq_mhdr *msghdr);
 
 /* APIs for BNA */
 void bna_rx_mod_init(struct bna_rx_mod *rx_mod, struct bna *bna,
diff --git a/drivers/net/ethernet/brocade/bna/bna_enet.c b/drivers/net/ethernet/brocade/bna/bna_enet.c
index db14f69..3ca77fa 100644
--- a/drivers/net/ethernet/brocade/bna/bna_enet.c
+++ b/drivers/net/ethernet/brocade/bna/bna_enet.c
@@ -298,7 +298,6 @@ bna_msgq_rsp_handler(void *arg, struct bfi_msgq_mhdr *msghdr)
 	case BFI_ENET_I2H_RSS_ENABLE_RSP:
 	case BFI_ENET_I2H_RX_PROMISCUOUS_RSP:
 	case BFI_ENET_I2H_RX_DEFAULT_RSP:
-	case BFI_ENET_I2H_MAC_UCAST_SET_RSP:
 	case BFI_ENET_I2H_MAC_UCAST_CLR_RSP:
 	case BFI_ENET_I2H_MAC_UCAST_ADD_RSP:
 	case BFI_ENET_I2H_MAC_UCAST_DEL_RSP:
@@ -311,6 +310,12 @@ bna_msgq_rsp_handler(void *arg, struct bfi_msgq_mhdr *msghdr)
 			bna_bfi_rxf_cfg_rsp(&rx->rxf, msghdr);
 		break;
 
+	case BFI_ENET_I2H_MAC_UCAST_SET_RSP:
+		bna_rx_from_rid(bna, msghdr->enet_id, rx);
+		if (rx)
+			bna_bfi_rxf_ucast_set_rsp(&rx->rxf, msghdr);
+		break;
+
 	case BFI_ENET_I2H_MAC_MCAST_ADD_RSP:
 		bna_rx_from_rid(bna, msghdr->enet_id, rx);
 		if (rx)
diff --git a/drivers/net/ethernet/brocade/bna/bna_tx_rx.c b/drivers/net/ethernet/brocade/bna/bna_tx_rx.c
index ea6f4a0..57cd1bf 100644
--- a/drivers/net/ethernet/brocade/bna/bna_tx_rx.c
+++ b/drivers/net/ethernet/brocade/bna/bna_tx_rx.c
@@ -711,6 +711,21 @@ bna_bfi_rxf_cfg_rsp(struct bna_rxf *rxf, struct bfi_msgq_mhdr *msghdr)
 }
 
 void
+bna_bfi_rxf_ucast_set_rsp(struct bna_rxf *rxf,
+			struct bfi_msgq_mhdr *msghdr)
+{
+	struct bfi_enet_rsp *rsp =
+		(struct bfi_enet_rsp *)msghdr;
+
+	if (rsp->error) {
+		/* Clear ucast from cache */
+		rxf->ucast_active_set = 0;
+	}
+
+	bfa_fsm_send_event(rxf, RXF_E_FW_RESP);
+}
+
+void
 bna_bfi_rxf_mcast_add_rsp(struct bna_rxf *rxf,
 			struct bfi_msgq_mhdr *msghdr)
 {
diff --git a/drivers/net/ethernet/brocade/bna/bnad.c b/drivers/net/ethernet/brocade/bna/bnad.c
index 07f7ef0..b78e69e 100644
--- a/drivers/net/ethernet/brocade/bna/bnad.c
+++ b/drivers/net/ethernet/brocade/bna/bnad.c
@@ -2624,6 +2624,9 @@ bnad_stop(struct net_device *netdev)
 	bnad_destroy_tx(bnad, 0);
 	bnad_destroy_rx(bnad, 0);
 
+	/* These config flags are cleared in the hardware */
+	bnad->cfg_flags &= ~(BNAD_CF_ALLMULTI | BNAD_CF_PROMISC);
+
 	/* Synchronize mailbox IRQ */
 	bnad_mbox_irq_sync(bnad);
 
diff --git a/drivers/net/ethernet/brocade/bna/bnad.h b/drivers/net/ethernet/brocade/bna/bnad.h
index c1d0bc0..aefee77 100644
--- a/drivers/net/ethernet/brocade/bna/bnad.h
+++ b/drivers/net/ethernet/brocade/bna/bnad.h
@@ -71,7 +71,7 @@ struct bnad_rx_ctrl {
 #define BNAD_NAME			"bna"
 #define BNAD_NAME_LEN			64
 
-#define BNAD_VERSION			"3.1.2.1"
+#define BNAD_VERSION			"3.2.21.1"
 
 #define BNAD_MAILBOX_MSIX_INDEX		0
 #define BNAD_MAILBOX_MSIX_VECTORS	1
diff --git a/drivers/net/ethernet/brocade/bna/cna.h b/drivers/net/ethernet/brocade/bna/cna.h
index 14ca931..c37f706 100644
--- a/drivers/net/ethernet/brocade/bna/cna.h
+++ b/drivers/net/ethernet/brocade/bna/cna.h
@@ -37,8 +37,8 @@
 
 extern char bfa_version[];
 
-#define CNA_FW_FILE_CT	"ctfw-3.1.0.0.bin"
-#define CNA_FW_FILE_CT2	"ct2fw-3.1.0.0.bin"
+#define CNA_FW_FILE_CT	"ctfw-3.2.1.0.bin"
+#define CNA_FW_FILE_CT2	"ct2fw-3.2.1.0.bin"
 #define FC_SYMNAME_MAX	256	/*!< max name server symbolic name size */
 
 #pragma pack(1)
diff --git a/drivers/net/ethernet/cadence/Kconfig b/drivers/net/ethernet/cadence/Kconfig
index 768285e..8030cc0 100644
--- a/drivers/net/ethernet/cadence/Kconfig
+++ b/drivers/net/ethernet/cadence/Kconfig
@@ -23,7 +23,6 @@ if NET_CADENCE
 config ARM_AT91_ETHER
 	tristate "AT91RM9200 Ethernet support"
 	depends on GENERIC_HARDIRQS && HAS_DMA
-	select NET_CORE
 	select MACB
 	---help---
 	  If you wish to compile a kernel for the AT91RM9200 and enable
diff --git a/drivers/net/ethernet/cadence/at91_ether.c b/drivers/net/ethernet/cadence/at91_ether.c
index cc9a185..3f19571 100644
--- a/drivers/net/ethernet/cadence/at91_ether.c
+++ b/drivers/net/ethernet/cadence/at91_ether.c
@@ -435,7 +435,6 @@ static int at91ether_remove(struct platform_device *pdev)
 	unregister_netdev(dev);
 	clk_disable(lp->pclk);
 	free_netdev(dev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c
index c89aa41..e866608 100644
--- a/drivers/net/ethernet/cadence/macb.c
+++ b/drivers/net/ethernet/cadence/macb.c
@@ -32,7 +32,8 @@
 
 #include "macb.h"
 
-#define RX_BUFFER_SIZE		128
+#define MACB_RX_BUFFER_SIZE	128
+#define RX_BUFFER_MULTIPLE	64  /* bytes */
 #define RX_RING_SIZE		512 /* must be power of 2 */
 #define RX_RING_BYTES		(sizeof(struct macb_dma_desc) * RX_RING_SIZE)
 
@@ -92,7 +93,7 @@ static struct macb_dma_desc *macb_rx_desc(struct macb *bp, unsigned int index)
 
 static void *macb_rx_buffer(struct macb *bp, unsigned int index)
 {
-	return bp->rx_buffers + RX_BUFFER_SIZE * macb_rx_ring_wrap(index);
+	return bp->rx_buffers + bp->rx_buffer_size * macb_rx_ring_wrap(index);
 }
 
 void macb_set_hwaddr(struct macb *bp)
@@ -528,6 +529,155 @@ static void macb_tx_interrupt(struct macb *bp)
 		netif_wake_queue(bp->dev);
 }
 
+static void gem_rx_refill(struct macb *bp)
+{
+	unsigned int		entry;
+	struct sk_buff		*skb;
+	struct macb_dma_desc	*desc;
+	dma_addr_t		paddr;
+
+	while (CIRC_SPACE(bp->rx_prepared_head, bp->rx_tail, RX_RING_SIZE) > 0) {
+		u32 addr, ctrl;
+
+		entry = macb_rx_ring_wrap(bp->rx_prepared_head);
+		desc = &bp->rx_ring[entry];
+
+		/* Make hw descriptor updates visible to CPU */
+		rmb();
+
+		addr = desc->addr;
+		ctrl = desc->ctrl;
+		bp->rx_prepared_head++;
+
+		if ((addr & MACB_BIT(RX_USED)))
+			continue;
+
+		if (bp->rx_skbuff[entry] == NULL) {
+			/* allocate sk_buff for this free entry in ring */
+			skb = netdev_alloc_skb(bp->dev, bp->rx_buffer_size);
+			if (unlikely(skb == NULL)) {
+				netdev_err(bp->dev,
+					   "Unable to allocate sk_buff\n");
+				break;
+			}
+			bp->rx_skbuff[entry] = skb;
+
+			/* now fill corresponding descriptor entry */
+			paddr = dma_map_single(&bp->pdev->dev, skb->data,
+					       bp->rx_buffer_size, DMA_FROM_DEVICE);
+
+			if (entry == RX_RING_SIZE - 1)
+				paddr |= MACB_BIT(RX_WRAP);
+			bp->rx_ring[entry].addr = paddr;
+			bp->rx_ring[entry].ctrl = 0;
+
+			/* properly align Ethernet header */
+			skb_reserve(skb, NET_IP_ALIGN);
+		}
+	}
+
+	/* Make descriptor updates visible to hardware */
+	wmb();
+
+	netdev_vdbg(bp->dev, "rx ring: prepared head %d, tail %d\n",
+		   bp->rx_prepared_head, bp->rx_tail);
+}
+
+/* Mark DMA descriptors from begin up to and not including end as unused */
+static void discard_partial_frame(struct macb *bp, unsigned int begin,
+				  unsigned int end)
+{
+	unsigned int frag;
+
+	for (frag = begin; frag != end; frag++) {
+		struct macb_dma_desc *desc = macb_rx_desc(bp, frag);
+		desc->addr &= ~MACB_BIT(RX_USED);
+	}
+
+	/* Make descriptor updates visible to hardware */
+	wmb();
+
+	/*
+	 * When this happens, the hardware stats registers for
+	 * whatever caused this is updated, so we don't have to record
+	 * anything.
+	 */
+}
+
+static int gem_rx(struct macb *bp, int budget)
+{
+	unsigned int		len;
+	unsigned int		entry;
+	struct sk_buff		*skb;
+	struct macb_dma_desc	*desc;
+	int			count = 0;
+
+	while (count < budget) {
+		u32 addr, ctrl;
+
+		entry = macb_rx_ring_wrap(bp->rx_tail);
+		desc = &bp->rx_ring[entry];
+
+		/* Make hw descriptor updates visible to CPU */
+		rmb();
+
+		addr = desc->addr;
+		ctrl = desc->ctrl;
+
+		if (!(addr & MACB_BIT(RX_USED)))
+			break;
+
+		desc->addr &= ~MACB_BIT(RX_USED);
+		bp->rx_tail++;
+		count++;
+
+		if (!(ctrl & MACB_BIT(RX_SOF) && ctrl & MACB_BIT(RX_EOF))) {
+			netdev_err(bp->dev,
+				   "not whole frame pointed by descriptor\n");
+			bp->stats.rx_dropped++;
+			break;
+		}
+		skb = bp->rx_skbuff[entry];
+		if (unlikely(!skb)) {
+			netdev_err(bp->dev,
+				   "inconsistent Rx descriptor chain\n");
+			bp->stats.rx_dropped++;
+			break;
+		}
+		/* now everything is ready for receiving packet */
+		bp->rx_skbuff[entry] = NULL;
+		len = MACB_BFEXT(RX_FRMLEN, ctrl);
+
+		netdev_vdbg(bp->dev, "gem_rx %u (len %u)\n", entry, len);
+
+		skb_put(skb, len);
+		addr = MACB_BF(RX_WADDR, MACB_BFEXT(RX_WADDR, addr));
+		dma_unmap_single(&bp->pdev->dev, addr,
+				 len, DMA_FROM_DEVICE);
+
+		skb->protocol = eth_type_trans(skb, bp->dev);
+		skb_checksum_none_assert(skb);
+
+		bp->stats.rx_packets++;
+		bp->stats.rx_bytes += skb->len;
+
+#if defined(DEBUG) && defined(VERBOSE_DEBUG)
+		netdev_vdbg(bp->dev, "received skb of length %u, csum: %08x\n",
+			    skb->len, skb->csum);
+		print_hex_dump(KERN_DEBUG, " mac: ", DUMP_PREFIX_ADDRESS, 16, 1,
+			       skb->mac_header, 16, true);
+		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_ADDRESS, 16, 1,
+			       skb->data, 32, true);
+#endif
+
+		netif_receive_skb(skb);
+	}
+
+	gem_rx_refill(bp);
+
+	return count;
+}
+
 static int macb_rx_frame(struct macb *bp, unsigned int first_frag,
 			 unsigned int last_frag)
 {
@@ -575,7 +725,7 @@ static int macb_rx_frame(struct macb *bp, unsigned int first_frag,
 	skb_put(skb, len);
 
 	for (frag = first_frag; ; frag++) {
-		unsigned int frag_len = RX_BUFFER_SIZE;
+		unsigned int frag_len = bp->rx_buffer_size;
 
 		if (offset + frag_len > len) {
 			BUG_ON(frag != last_frag);
@@ -583,7 +733,7 @@ static int macb_rx_frame(struct macb *bp, unsigned int first_frag,
 		}
 		skb_copy_to_linear_data_offset(skb, offset,
 				macb_rx_buffer(bp, frag), frag_len);
-		offset += RX_BUFFER_SIZE;
+		offset += bp->rx_buffer_size;
 		desc = macb_rx_desc(bp, frag);
 		desc->addr &= ~MACB_BIT(RX_USED);
 
@@ -606,27 +756,6 @@ static int macb_rx_frame(struct macb *bp, unsigned int first_frag,
 	return 0;
 }
 
-/* Mark DMA descriptors from begin up to and not including end as unused */
-static void discard_partial_frame(struct macb *bp, unsigned int begin,
-				  unsigned int end)
-{
-	unsigned int frag;
-
-	for (frag = begin; frag != end; frag++) {
-		struct macb_dma_desc *desc = macb_rx_desc(bp, frag);
-		desc->addr &= ~MACB_BIT(RX_USED);
-	}
-
-	/* Make descriptor updates visible to hardware */
-	wmb();
-
-	/*
-	 * When this happens, the hardware stats registers for
-	 * whatever caused this is updated, so we don't have to record
-	 * anything.
-	 */
-}
-
 static int macb_rx(struct macb *bp, int budget)
 {
 	int received = 0;
@@ -687,7 +816,7 @@ static int macb_poll(struct napi_struct *napi, int budget)
 	netdev_vdbg(bp->dev, "poll: status = %08lx, budget = %d\n",
 		   (unsigned long)status, budget);
 
-	work_done = macb_rx(bp, budget);
+	work_done = bp->macbgem_ops.mog_rx(bp, budget);
 	if (work_done < budget) {
 		napi_complete(napi);
 
@@ -870,12 +999,71 @@ static int macb_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	return NETDEV_TX_OK;
 }
 
+static void macb_init_rx_buffer_size(struct macb *bp, size_t size)
+{
+	if (!macb_is_gem(bp)) {
+		bp->rx_buffer_size = MACB_RX_BUFFER_SIZE;
+	} else {
+		bp->rx_buffer_size = size;
+
+		if (bp->rx_buffer_size % RX_BUFFER_MULTIPLE) {
+			netdev_dbg(bp->dev,
+				    "RX buffer must be multiple of %d bytes, expanding\n",
+				    RX_BUFFER_MULTIPLE);
+			bp->rx_buffer_size =
+				roundup(bp->rx_buffer_size, RX_BUFFER_MULTIPLE);
+		}
+	}
+
+	netdev_dbg(bp->dev, "mtu [%u] rx_buffer_size [%Zu]\n",
+		   bp->dev->mtu, bp->rx_buffer_size);
+}
+
+static void gem_free_rx_buffers(struct macb *bp)
+{
+	struct sk_buff		*skb;
+	struct macb_dma_desc	*desc;
+	dma_addr_t		addr;
+	int i;
+
+	if (!bp->rx_skbuff)
+		return;
+
+	for (i = 0; i < RX_RING_SIZE; i++) {
+		skb = bp->rx_skbuff[i];
+
+		if (skb == NULL)
+			continue;
+
+		desc = &bp->rx_ring[i];
+		addr = MACB_BF(RX_WADDR, MACB_BFEXT(RX_WADDR, desc->addr));
+		dma_unmap_single(&bp->pdev->dev, addr, skb->len,
+				 DMA_FROM_DEVICE);
+		dev_kfree_skb_any(skb);
+		skb = NULL;
+	}
+
+	kfree(bp->rx_skbuff);
+	bp->rx_skbuff = NULL;
+}
+
+static void macb_free_rx_buffers(struct macb *bp)
+{
+	if (bp->rx_buffers) {
+		dma_free_coherent(&bp->pdev->dev,
+				  RX_RING_SIZE * bp->rx_buffer_size,
+				  bp->rx_buffers, bp->rx_buffers_dma);
+		bp->rx_buffers = NULL;
+	}
+}
+
 static void macb_free_consistent(struct macb *bp)
 {
 	if (bp->tx_skb) {
 		kfree(bp->tx_skb);
 		bp->tx_skb = NULL;
 	}
+	bp->macbgem_ops.mog_free_rx_buffers(bp);
 	if (bp->rx_ring) {
 		dma_free_coherent(&bp->pdev->dev, RX_RING_BYTES,
 				  bp->rx_ring, bp->rx_ring_dma);
@@ -886,12 +1074,37 @@ static void macb_free_consistent(struct macb *bp)
 				  bp->tx_ring, bp->tx_ring_dma);
 		bp->tx_ring = NULL;
 	}
-	if (bp->rx_buffers) {
-		dma_free_coherent(&bp->pdev->dev,
-				  RX_RING_SIZE * RX_BUFFER_SIZE,
-				  bp->rx_buffers, bp->rx_buffers_dma);
-		bp->rx_buffers = NULL;
-	}
+}
+
+static int gem_alloc_rx_buffers(struct macb *bp)
+{
+	int size;
+
+	size = RX_RING_SIZE * sizeof(struct sk_buff *);
+	bp->rx_skbuff = kzalloc(size, GFP_KERNEL);
+	if (!bp->rx_skbuff)
+		return -ENOMEM;
+	else
+		netdev_dbg(bp->dev,
+			   "Allocated %d RX struct sk_buff entries at %p\n",
+			   RX_RING_SIZE, bp->rx_skbuff);
+	return 0;
+}
+
+static int macb_alloc_rx_buffers(struct macb *bp)
+{
+	int size;
+
+	size = RX_RING_SIZE * bp->rx_buffer_size;
+	bp->rx_buffers = dma_alloc_coherent(&bp->pdev->dev, size,
+					    &bp->rx_buffers_dma, GFP_KERNEL);
+	if (!bp->rx_buffers)
+		return -ENOMEM;
+	else
+		netdev_dbg(bp->dev,
+			   "Allocated RX buffers of %d bytes at %08lx (mapped %p)\n",
+			   size, (unsigned long)bp->rx_buffers_dma, bp->rx_buffers);
+	return 0;
 }
 
 static int macb_alloc_consistent(struct macb *bp)
@@ -921,14 +1134,8 @@ static int macb_alloc_consistent(struct macb *bp)
 		   "Allocated TX ring of %d bytes at %08lx (mapped %p)\n",
 		   size, (unsigned long)bp->tx_ring_dma, bp->tx_ring);
 
-	size = RX_RING_SIZE * RX_BUFFER_SIZE;
-	bp->rx_buffers = dma_alloc_coherent(&bp->pdev->dev, size,
-					    &bp->rx_buffers_dma, GFP_KERNEL);
-	if (!bp->rx_buffers)
+	if (bp->macbgem_ops.mog_alloc_rx_buffers(bp))
 		goto out_err;
-	netdev_dbg(bp->dev,
-		   "Allocated RX buffers of %d bytes at %08lx (mapped %p)\n",
-		   size, (unsigned long)bp->rx_buffers_dma, bp->rx_buffers);
 
 	return 0;
 
@@ -937,6 +1144,21 @@ out_err:
 	return -ENOMEM;
 }
 
+static void gem_init_rings(struct macb *bp)
+{
+	int i;
+
+	for (i = 0; i < TX_RING_SIZE; i++) {
+		bp->tx_ring[i].addr = 0;
+		bp->tx_ring[i].ctrl = MACB_BIT(TX_USED);
+	}
+	bp->tx_ring[TX_RING_SIZE - 1].ctrl |= MACB_BIT(TX_WRAP);
+
+	bp->rx_tail = bp->rx_prepared_head = bp->tx_head = bp->tx_tail = 0;
+
+	gem_rx_refill(bp);
+}
+
 static void macb_init_rings(struct macb *bp)
 {
 	int i;
@@ -946,7 +1168,7 @@ static void macb_init_rings(struct macb *bp)
 	for (i = 0; i < RX_RING_SIZE; i++) {
 		bp->rx_ring[i].addr = addr;
 		bp->rx_ring[i].ctrl = 0;
-		addr += RX_BUFFER_SIZE;
+		addr += bp->rx_buffer_size;
 	}
 	bp->rx_ring[RX_RING_SIZE - 1].addr |= MACB_BIT(RX_WRAP);
 
@@ -1056,7 +1278,7 @@ static void macb_configure_dma(struct macb *bp)
 
 	if (macb_is_gem(bp)) {
 		dmacfg = gem_readl(bp, DMACFG) & ~GEM_BF(RXBS, -1L);
-		dmacfg |= GEM_BF(RXBS, RX_BUFFER_SIZE / 64);
+		dmacfg |= GEM_BF(RXBS, bp->rx_buffer_size / RX_BUFFER_MULTIPLE);
 		dmacfg |= GEM_BF(FBLDO, 16);
 		dmacfg |= GEM_BIT(TXPBMS) | GEM_BF(RXBMS, -1L);
 		dmacfg &= ~GEM_BIT(ENDIA);
@@ -1070,7 +1292,7 @@ static void macb_configure_dma(struct macb *bp)
 static void macb_configure_caps(struct macb *bp)
 {
 	if (macb_is_gem(bp)) {
-		if (GEM_BF(IRQCOR, gem_readl(bp, DCFG1)) == 0)
+		if (GEM_BFEXT(IRQCOR, gem_readl(bp, DCFG1)) == 0)
 			bp->caps |= MACB_CAPS_ISR_CLEAR_ON_WRITE;
 	}
 }
@@ -1233,6 +1455,7 @@ EXPORT_SYMBOL_GPL(macb_set_rx_mode);
 static int macb_open(struct net_device *dev)
 {
 	struct macb *bp = netdev_priv(dev);
+	size_t bufsz = dev->mtu + ETH_HLEN + ETH_FCS_LEN + NET_IP_ALIGN;
 	int err;
 
 	netdev_dbg(bp->dev, "open\n");
@@ -1244,6 +1467,9 @@ static int macb_open(struct net_device *dev)
 	if (!bp->phy_dev)
 		return -EAGAIN;
 
+	/* RX buffers initialization */
+	macb_init_rx_buffer_size(bp, bufsz);
+
 	err = macb_alloc_consistent(bp);
 	if (err) {
 		netdev_err(dev, "Unable to allocate DMA memory (error %d)\n",
@@ -1253,7 +1479,7 @@ static int macb_open(struct net_device *dev)
 
 	napi_enable(&bp->napi);
 
-	macb_init_rings(bp);
+	bp->macbgem_ops.mog_init_rings(bp);
 	macb_init_hw(bp);
 
 	/* schedule a link state check */
@@ -1572,6 +1798,19 @@ static int __init macb_probe(struct platform_device *pdev)
 
 	dev->base_addr = regs->start;
 
+	/* setup appropriated routines according to adapter type */
+	if (macb_is_gem(bp)) {
+		bp->macbgem_ops.mog_alloc_rx_buffers = gem_alloc_rx_buffers;
+		bp->macbgem_ops.mog_free_rx_buffers = gem_free_rx_buffers;
+		bp->macbgem_ops.mog_init_rings = gem_init_rings;
+		bp->macbgem_ops.mog_rx = gem_rx;
+	} else {
+		bp->macbgem_ops.mog_alloc_rx_buffers = macb_alloc_rx_buffers;
+		bp->macbgem_ops.mog_free_rx_buffers = macb_free_rx_buffers;
+		bp->macbgem_ops.mog_init_rings = macb_init_rings;
+		bp->macbgem_ops.mog_rx = macb_rx;
+	}
+
 	/* Set MII management clock divider */
 	config = macb_mdc_clk_div(bp);
 	config |= macb_dbw(bp);
@@ -1649,7 +1888,6 @@ err_out_put_pclk:
 err_out_free_dev:
 	free_netdev(dev);
 err_out:
-	platform_set_drvdata(pdev, NULL);
 	return err;
 }
 
@@ -1675,7 +1913,6 @@ static int __exit macb_remove(struct platform_device *pdev)
 		clk_disable_unprepare(bp->pclk);
 		clk_put(bp->pclk);
 		free_netdev(dev);
-		platform_set_drvdata(pdev, NULL);
 	}
 
 	return 0;
diff --git a/drivers/net/ethernet/cadence/macb.h b/drivers/net/ethernet/cadence/macb.h
index 548c0ec..f407615 100644
--- a/drivers/net/ethernet/cadence/macb.h
+++ b/drivers/net/ethernet/cadence/macb.h
@@ -545,12 +545,24 @@ struct gem_stats {
 	u32	rx_udp_checksum_errors;
 };
 
+struct macb;
+
+struct macb_or_gem_ops {
+	int	(*mog_alloc_rx_buffers)(struct macb *bp);
+	void	(*mog_free_rx_buffers)(struct macb *bp);
+	void	(*mog_init_rings)(struct macb *bp);
+	int	(*mog_rx)(struct macb *bp, int budget);
+};
+
 struct macb {
 	void __iomem		*regs;
 
 	unsigned int		rx_tail;
+	unsigned int		rx_prepared_head;
 	struct macb_dma_desc	*rx_ring;
+	struct sk_buff		**rx_skbuff;
 	void			*rx_buffers;
+	size_t			rx_buffer_size;
 
 	unsigned int		tx_head, tx_tail;
 	struct macb_dma_desc	*tx_ring;
@@ -573,6 +585,8 @@ struct macb {
 	dma_addr_t		tx_ring_dma;
 	dma_addr_t		rx_buffers_dma;
 
+	struct macb_or_gem_ops	macbgem_ops;
+
 	struct mii_bus		*mii_bus;
 	struct phy_device	*phy_dev;
 	unsigned int 		link;
diff --git a/drivers/net/ethernet/calxeda/xgmac.c b/drivers/net/ethernet/calxeda/xgmac.c
index 4a1f2fa..7cb148c 100644
--- a/drivers/net/ethernet/calxeda/xgmac.c
+++ b/drivers/net/ethernet/calxeda/xgmac.c
@@ -1790,7 +1790,6 @@ err_io:
 	free_netdev(ndev);
 err_alloc:
 	release_mem_region(res->start, resource_size(res));
-	platform_set_drvdata(pdev, NULL);
 	return ret;
 }
 
@@ -1813,7 +1812,6 @@ static int xgmac_remove(struct platform_device *pdev)
 	free_irq(ndev->irq, ndev);
 	free_irq(priv->pmt_irq, ndev);
 
-	platform_set_drvdata(pdev, NULL);
 	unregister_netdev(ndev);
 	netif_napi_del(&priv->napi);
 
diff --git a/drivers/net/ethernet/chelsio/cxgb/cxgb2.c b/drivers/net/ethernet/chelsio/cxgb/cxgb2.c
index 9624cfe..d7048db 100644
--- a/drivers/net/ethernet/chelsio/cxgb/cxgb2.c
+++ b/drivers/net/ethernet/chelsio/cxgb/cxgb2.c
@@ -1351,22 +1351,11 @@ static void remove_one(struct pci_dev *pdev)
 	t1_sw_reset(pdev);
 }
 
-static struct pci_driver driver = {
+static struct pci_driver cxgb_pci_driver = {
 	.name     = DRV_NAME,
 	.id_table = t1_pci_tbl,
 	.probe    = init_one,
 	.remove   = remove_one,
 };
 
-static int __init t1_init_module(void)
-{
-	return pci_register_driver(&driver);
-}
-
-static void __exit t1_cleanup_module(void)
-{
-	pci_unregister_driver(&driver);
-}
-
-module_init(t1_init_module);
-module_exit(t1_cleanup_module);
+module_pci_driver(cxgb_pci_driver);
diff --git a/drivers/net/ethernet/chelsio/cxgb3/cxgb3_main.c b/drivers/net/ethernet/chelsio/cxgb3/cxgb3_main.c
index 71497e8..b650951 100644
--- a/drivers/net/ethernet/chelsio/cxgb3/cxgb3_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb3/cxgb3_main.c
@@ -3037,7 +3037,9 @@ static void t3_io_resume(struct pci_dev *pdev)
 	CH_ALERT(adapter, "adapter recovering, PEX ERR 0x%x\n",
 		 t3_read_reg(adapter, A_PCIE_PEX_ERR));
 
+	rtnl_lock();
 	t3_resume_ports(adapter);
+	rtnl_unlock();
 }
 
 static const struct pci_error_handlers t3_err_handler = {
diff --git a/drivers/net/ethernet/chelsio/cxgb3/cxgb3_offload.c b/drivers/net/ethernet/chelsio/cxgb3/cxgb3_offload.c
index 0c96e5f..4058b85 100644
--- a/drivers/net/ethernet/chelsio/cxgb3/cxgb3_offload.c
+++ b/drivers/net/ethernet/chelsio/cxgb3/cxgb3_offload.c
@@ -1246,6 +1246,7 @@ int cxgb3_offload_activate(struct adapter *adapter)
 	struct tid_range stid_range, tid_range;
 	struct mtutab mtutab;
 	unsigned int l2t_capacity;
+	struct l2t_data *l2td;
 
 	t = kzalloc(sizeof(*t), GFP_KERNEL);
 	if (!t)
@@ -1261,8 +1262,8 @@ int cxgb3_offload_activate(struct adapter *adapter)
 		goto out_free;
 
 	err = -ENOMEM;
-	RCU_INIT_POINTER(dev->l2opt, t3_init_l2t(l2t_capacity));
-	if (!L2DATA(dev))
+	l2td = t3_init_l2t(l2t_capacity);
+	if (!l2td)
 		goto out_free;
 
 	natids = min(tid_range.num / 2, MAX_ATIDS);
@@ -1279,6 +1280,7 @@ int cxgb3_offload_activate(struct adapter *adapter)
 	INIT_LIST_HEAD(&t->list_node);
 	t->dev = dev;
 
+	RCU_INIT_POINTER(dev->l2opt, l2td);
 	T3C_DATA(dev) = t;
 	dev->recv = process_rx;
 	dev->neigh_update = t3_l2t_update;
@@ -1294,8 +1296,7 @@ int cxgb3_offload_activate(struct adapter *adapter)
 	return 0;
 
 out_free_l2t:
-	t3_free_l2t(L2DATA(dev));
-	RCU_INIT_POINTER(dev->l2opt, NULL);
+	t3_free_l2t(l2td);
 out_free:
 	kfree(t);
 	return err;
diff --git a/drivers/net/ethernet/chelsio/cxgb3/sge.c b/drivers/net/ethernet/chelsio/cxgb3/sge.c
index f12e6b8..687ec4a 100644
--- a/drivers/net/ethernet/chelsio/cxgb3/sge.c
+++ b/drivers/net/ethernet/chelsio/cxgb3/sge.c
@@ -455,6 +455,11 @@ static int alloc_pg_chunk(struct adapter *adapter, struct sge_fl *q,
 		q->pg_chunk.offset = 0;
 		mapping = pci_map_page(adapter->pdev, q->pg_chunk.page,
 				       0, q->alloc_size, PCI_DMA_FROMDEVICE);
+		if (unlikely(pci_dma_mapping_error(adapter->pdev, mapping))) {
+			__free_pages(q->pg_chunk.page, order);
+			q->pg_chunk.page = NULL;
+			return -EIO;
+		}
 		q->pg_chunk.mapping = mapping;
 	}
 	sd->pg_chunk = q->pg_chunk;
@@ -949,40 +954,75 @@ static inline unsigned int calc_tx_descs(const struct sk_buff *skb)
 	return flits_to_desc(flits);
 }
 
+
+/*	map_skb - map a packet main body and its page fragments
+ *	@pdev: the PCI device
+ *	@skb: the packet
+ *	@addr: placeholder to save the mapped addresses
+ *
+ *	map the main body of an sk_buff and its page fragments, if any.
+ */
+static int map_skb(struct pci_dev *pdev, const struct sk_buff *skb,
+		   dma_addr_t *addr)
+{
+	const skb_frag_t *fp, *end;
+	const struct skb_shared_info *si;
+
+	*addr = pci_map_single(pdev, skb->data, skb_headlen(skb),
+			       PCI_DMA_TODEVICE);
+	if (pci_dma_mapping_error(pdev, *addr))
+		goto out_err;
+
+	si = skb_shinfo(skb);
+	end = &si->frags[si->nr_frags];
+
+	for (fp = si->frags; fp < end; fp++) {
+		*++addr = skb_frag_dma_map(&pdev->dev, fp, 0, skb_frag_size(fp),
+					   DMA_TO_DEVICE);
+		if (pci_dma_mapping_error(pdev, *addr))
+			goto unwind;
+	}
+	return 0;
+
+unwind:
+	while (fp-- > si->frags)
+		dma_unmap_page(&pdev->dev, *--addr, skb_frag_size(fp),
+			       DMA_TO_DEVICE);
+
+	pci_unmap_single(pdev, addr[-1], skb_headlen(skb), PCI_DMA_TODEVICE);
+out_err:
+	return -ENOMEM;
+}
+
 /**
- *	make_sgl - populate a scatter/gather list for a packet
+ *	write_sgl - populate a scatter/gather list for a packet
  *	@skb: the packet
  *	@sgp: the SGL to populate
  *	@start: start address of skb main body data to include in the SGL
  *	@len: length of skb main body data to include in the SGL
- *	@pdev: the PCI device
+ *	@addr: the list of the mapped addresses
  *
- *	Generates a scatter/gather list for the buffers that make up a packet
+ *	Copies the scatter/gather list for the buffers that make up a packet
  *	and returns the SGL size in 8-byte words.  The caller must size the SGL
  *	appropriately.
  */
-static inline unsigned int make_sgl(const struct sk_buff *skb,
+static inline unsigned int write_sgl(const struct sk_buff *skb,
 				    struct sg_ent *sgp, unsigned char *start,
-				    unsigned int len, struct pci_dev *pdev)
+				    unsigned int len, const dma_addr_t *addr)
 {
-	dma_addr_t mapping;
-	unsigned int i, j = 0, nfrags;
+	unsigned int i, j = 0, k = 0, nfrags;
 
 	if (len) {
-		mapping = pci_map_single(pdev, start, len, PCI_DMA_TODEVICE);
 		sgp->len[0] = cpu_to_be32(len);
-		sgp->addr[0] = cpu_to_be64(mapping);
-		j = 1;
+		sgp->addr[j++] = cpu_to_be64(addr[k++]);
 	}
 
 	nfrags = skb_shinfo(skb)->nr_frags;
 	for (i = 0; i < nfrags; i++) {
 		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 
-		mapping = skb_frag_dma_map(&pdev->dev, frag, 0, skb_frag_size(frag),
-					   DMA_TO_DEVICE);
 		sgp->len[j] = cpu_to_be32(skb_frag_size(frag));
-		sgp->addr[j] = cpu_to_be64(mapping);
+		sgp->addr[j] = cpu_to_be64(addr[k++]);
 		j ^= 1;
 		if (j == 0)
 			++sgp;
@@ -1138,7 +1178,7 @@ static void write_tx_pkt_wr(struct adapter *adap, struct sk_buff *skb,
 			    const struct port_info *pi,
 			    unsigned int pidx, unsigned int gen,
 			    struct sge_txq *q, unsigned int ndesc,
-			    unsigned int compl)
+			    unsigned int compl, const dma_addr_t *addr)
 {
 	unsigned int flits, sgl_flits, cntrl, tso_info;
 	struct sg_ent *sgp, sgl[MAX_SKB_FRAGS / 2 + 1];
@@ -1196,7 +1236,7 @@ static void write_tx_pkt_wr(struct adapter *adap, struct sk_buff *skb,
 	}
 
 	sgp = ndesc == 1 ? (struct sg_ent *)&d->flit[flits] : sgl;
-	sgl_flits = make_sgl(skb, sgp, skb->data, skb_headlen(skb), adap->pdev);
+	sgl_flits = write_sgl(skb, sgp, skb->data, skb_headlen(skb), addr);
 
 	write_wr_hdr_sgl(ndesc, skb, d, pidx, q, sgl, flits, sgl_flits, gen,
 			 htonl(V_WR_OP(FW_WROPCODE_TUNNEL_TX_PKT) | compl),
@@ -1227,6 +1267,7 @@ netdev_tx_t t3_eth_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct netdev_queue *txq;
 	struct sge_qset *qs;
 	struct sge_txq *q;
+	dma_addr_t addr[MAX_SKB_FRAGS + 1];
 
 	/*
 	 * The chip min packet length is 9 octets but play safe and reject
@@ -1255,6 +1296,11 @@ netdev_tx_t t3_eth_xmit(struct sk_buff *skb, struct net_device *dev)
 		return NETDEV_TX_BUSY;
 	}
 
+	if (unlikely(map_skb(adap->pdev, skb, addr) < 0)) {
+		dev_kfree_skb(skb);
+		return NETDEV_TX_OK;
+	}
+
 	q->in_use += ndesc;
 	if (unlikely(credits - ndesc < q->stop_thres)) {
 		t3_stop_tx_queue(txq, qs, q);
@@ -1312,7 +1358,7 @@ netdev_tx_t t3_eth_xmit(struct sk_buff *skb, struct net_device *dev)
 	if (likely(!skb_shared(skb)))
 		skb_orphan(skb);
 
-	write_tx_pkt_wr(adap, skb, pi, pidx, gen, q, ndesc, compl);
+	write_tx_pkt_wr(adap, skb, pi, pidx, gen, q, ndesc, compl, addr);
 	check_ring_tx_db(adap, q);
 	return NETDEV_TX_OK;
 }
@@ -1537,10 +1583,9 @@ static void deferred_unmap_destructor(struct sk_buff *skb)
 	dui = (struct deferred_unmap_info *)skb->head;
 	p = dui->addr;
 
-	if (skb->tail - skb->transport_header)
-		pci_unmap_single(dui->pdev, *p++,
-				 skb->tail - skb->transport_header,
-				 PCI_DMA_TODEVICE);
+	if (skb_tail_pointer(skb) - skb_transport_header(skb))
+		pci_unmap_single(dui->pdev, *p++, skb_tail_pointer(skb) -
+				 skb_transport_header(skb), PCI_DMA_TODEVICE);
 
 	si = skb_shinfo(skb);
 	for (i = 0; i < si->nr_frags; i++)
@@ -1578,7 +1623,8 @@ static void setup_deferred_unmapping(struct sk_buff *skb, struct pci_dev *pdev,
  */
 static void write_ofld_wr(struct adapter *adap, struct sk_buff *skb,
 			  struct sge_txq *q, unsigned int pidx,
-			  unsigned int gen, unsigned int ndesc)
+			  unsigned int gen, unsigned int ndesc,
+			  const dma_addr_t *addr)
 {
 	unsigned int sgl_flits, flits;
 	struct work_request_hdr *from;
@@ -1599,9 +1645,9 @@ static void write_ofld_wr(struct adapter *adap, struct sk_buff *skb,
 
 	flits = skb_transport_offset(skb) / 8;
 	sgp = ndesc == 1 ? (struct sg_ent *)&d->flit[flits] : sgl;
-	sgl_flits = make_sgl(skb, sgp, skb_transport_header(skb),
-			     skb->tail - skb->transport_header,
-			     adap->pdev);
+	sgl_flits = write_sgl(skb, sgp, skb_transport_header(skb),
+			     skb_tail_pointer(skb) -
+			     skb_transport_header(skb), addr);
 	if (need_skb_unmap()) {
 		setup_deferred_unmapping(skb, adap->pdev, sgp, sgl_flits);
 		skb->destructor = deferred_unmap_destructor;
@@ -1627,7 +1673,7 @@ static inline unsigned int calc_tx_descs_ofld(const struct sk_buff *skb)
 
 	flits = skb_transport_offset(skb) / 8;	/* headers */
 	cnt = skb_shinfo(skb)->nr_frags;
-	if (skb->tail != skb->transport_header)
+	if (skb_tail_pointer(skb) != skb_transport_header(skb))
 		cnt++;
 	return flits_to_desc(flits + sgl_len(cnt));
 }
@@ -1659,6 +1705,11 @@ again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 		goto again;
 	}
 
+	if (map_skb(adap->pdev, skb, (dma_addr_t *)skb->head)) {
+		spin_unlock(&q->lock);
+		return NET_XMIT_SUCCESS;
+	}
+
 	gen = q->gen;
 	q->in_use += ndesc;
 	pidx = q->pidx;
@@ -1669,7 +1720,7 @@ again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 	}
 	spin_unlock(&q->lock);
 
-	write_ofld_wr(adap, skb, q, pidx, gen, ndesc);
+	write_ofld_wr(adap, skb, q, pidx, gen, ndesc, (dma_addr_t *)skb->head);
 	check_ring_tx_db(adap, q);
 	return NET_XMIT_SUCCESS;
 }
@@ -1687,6 +1738,7 @@ static void restart_offloadq(unsigned long data)
 	struct sge_txq *q = &qs->txq[TXQ_OFLD];
 	const struct port_info *pi = netdev_priv(qs->netdev);
 	struct adapter *adap = pi->adapter;
+	unsigned int written = 0;
 
 	spin_lock(&q->lock);
 again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
@@ -1706,10 +1758,14 @@ again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 			break;
 		}
 
+		if (map_skb(adap->pdev, skb, (dma_addr_t *)skb->head))
+			break;
+
 		gen = q->gen;
 		q->in_use += ndesc;
 		pidx = q->pidx;
 		q->pidx += ndesc;
+		written += ndesc;
 		if (q->pidx >= q->size) {
 			q->pidx -= q->size;
 			q->gen ^= 1;
@@ -1717,7 +1773,8 @@ again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 		__skb_unlink(skb, &q->sendq);
 		spin_unlock(&q->lock);
 
-		write_ofld_wr(adap, skb, q, pidx, gen, ndesc);
+		write_ofld_wr(adap, skb, q, pidx, gen, ndesc,
+			     (dma_addr_t *)skb->head);
 		spin_lock(&q->lock);
 	}
 	spin_unlock(&q->lock);
@@ -1727,8 +1784,9 @@ again:	reclaim_completed_tx(adap, q, TX_RECLAIM_CHUNK);
 	set_bit(TXQ_LAST_PKT_DB, &q->flags);
 #endif
 	wmb();
-	t3_write_reg(adap, A_SG_KDOORBELL,
-		     F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id));
+	if (likely(written))
+		t3_write_reg(adap, A_SG_KDOORBELL,
+			     F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id));
 }
 
 /**
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
index 681804b..2aafb80 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
@@ -51,7 +51,7 @@
 #include "t4_hw.h"
 
 #define FW_VERSION_MAJOR 1
-#define FW_VERSION_MINOR 1
+#define FW_VERSION_MINOR 4
 #define FW_VERSION_MICRO 0
 
 #define FW_VERSION_MAJOR_T5 0
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index 3cd397d..5a3256b 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -4842,8 +4842,17 @@ static int adap_init0(struct adapter *adap)
 	 * is excessively mismatched relative to the driver.)
 	 */
 	ret = t4_check_fw_version(adap);
+
+	/* The error code -EFAULT is returned by t4_check_fw_version() if
+	 * firmware on adapter < supported firmware. If firmware on adapter
+	 * is too old (not supported by driver) and we're the MASTER_PF set
+	 * adapter state to DEV_STATE_UNINIT to force firmware upgrade
+	 * and reinitialization.
+	 */
+	if ((adap->flags & MASTER_PF) && ret == -EFAULT)
+		state = DEV_STATE_UNINIT;
 	if ((adap->flags & MASTER_PF) && state != DEV_STATE_INIT) {
-		if (ret == -EINVAL || ret > 0) {
+		if (ret == -EINVAL || ret == -EFAULT || ret > 0) {
 			if (upgrade_fw(adap) >= 0) {
 				/*
 				 * Note that the chip was reset as part of the
@@ -4852,7 +4861,21 @@ static int adap_init0(struct adapter *adap)
 				 */
 				reset = 0;
 				ret = t4_check_fw_version(adap);
-			}
+			} else
+				if (ret == -EFAULT) {
+					/*
+					 * Firmware is old but still might
+					 * work if we force reinitialization
+					 * of the adapter. Ignoring FW upgrade
+					 * failure.
+					 */
+					dev_warn(adap->pdev_dev,
+						 "Ignoring firmware upgrade "
+						 "failure, and forcing driver "
+						 "to reinitialize the "
+						 "adapter.\n");
+					ret = 0;
+				}
 		}
 		if (ret < 0)
 			return ret;
diff --git a/drivers/net/ethernet/chelsio/cxgb4/sge.c b/drivers/net/ethernet/chelsio/cxgb4/sge.c
index 2bfbb20..ac311f5 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/sge.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/sge.c
@@ -1294,7 +1294,7 @@ static inline unsigned int calc_tx_flits_ofld(const struct sk_buff *skb)
 
 	flits = skb_transport_offset(skb) / 8U;   /* headers */
 	cnt = skb_shinfo(skb)->nr_frags;
-	if (skb->tail != skb->transport_header)
+	if (skb_tail_pointer(skb) != skb_transport_header(skb))
 		cnt++;
 	return flits + sgl_len(cnt);
 }
diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c b/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c
index d02d4e8..4cbb2f9 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c
@@ -938,6 +938,15 @@ int t4_check_fw_version(struct adapter *adapter)
 	memcpy(adapter->params.api_vers, api_vers,
 	       sizeof(adapter->params.api_vers));
 
+	if (major < exp_major || (major == exp_major && minor < exp_minor) ||
+	    (major == exp_major && minor == exp_minor && micro < exp_micro)) {
+		dev_err(adapter->pdev_dev,
+			"Card has firmware version %u.%u.%u, minimum "
+			"supported firmware is %u.%u.%u.\n", major, minor,
+			micro, exp_major, exp_minor, exp_micro);
+		return -EFAULT;
+	}
+
 	if (major != exp_major) {            /* major mismatch - fail */
 		dev_err(adapter->pdev_dev,
 			"card FW has major version %u, driver wants %u\n",
@@ -3773,7 +3782,6 @@ int t4_port_init(struct adapter *adap, int mbox, int pf, int vf)
 		p->lport = j;
 		p->rss_size = rss_size;
 		memcpy(adap->port[i]->dev_addr, addr, ETH_ALEN);
-		adap->port[i]->dev_id = j;
 
 		ret = ntohl(c.u.info.lstatus_to_modtype);
 		p->mdio_addr = (ret & FW_PORT_CMD_MDIOCAP) ?
diff --git a/drivers/net/ethernet/cirrus/Kconfig b/drivers/net/ethernet/cirrus/Kconfig
index 8388e36..7403dff 100644
--- a/drivers/net/ethernet/cirrus/Kconfig
+++ b/drivers/net/ethernet/cirrus/Kconfig
@@ -44,7 +44,6 @@ config CS89x0_PLATFORM
 config EP93XX_ETH
 	tristate "EP93xx Ethernet support"
 	depends on ARM && ARCH_EP93XX
-	select NET_CORE
 	select MII
 	help
 	  This is a driver for the ethernet hardware included in EP93xx CPUs.
diff --git a/drivers/net/ethernet/cirrus/ep93xx_eth.c b/drivers/net/ethernet/cirrus/ep93xx_eth.c
index 67b0388..e3d4ec8 100644
--- a/drivers/net/ethernet/cirrus/ep93xx_eth.c
+++ b/drivers/net/ethernet/cirrus/ep93xx_eth.c
@@ -783,7 +783,6 @@ static int ep93xx_eth_remove(struct platform_device *pdev)
 	dev = platform_get_drvdata(pdev);
 	if (dev == NULL)
 		return 0;
-	platform_set_drvdata(pdev, NULL);
 
 	ep = netdev_priv(dev);
 
diff --git a/drivers/net/ethernet/cisco/enic/enic_main.c b/drivers/net/ethernet/cisco/enic/enic_main.c
index 635f559..992ec2e 100644
--- a/drivers/net/ethernet/cisco/enic/enic_main.c
+++ b/drivers/net/ethernet/cisco/enic/enic_main.c
@@ -1761,6 +1761,7 @@ static void enic_change_mtu_work(struct work_struct *work)
 	enic_synchronize_irqs(enic);
 	err = vnic_rq_disable(&enic->rq[0]);
 	if (err) {
+		rtnl_unlock();
 		netdev_err(netdev, "Unable to disable RQ.\n");
 		return;
 	}
@@ -1773,6 +1774,7 @@ static void enic_change_mtu_work(struct work_struct *work)
 	vnic_rq_fill(&enic->rq[0], enic_rq_alloc_buf);
 	/* Need at least one buffer on ring to get going */
 	if (vnic_rq_desc_used(&enic->rq[0]) == 0) {
+		rtnl_unlock();
 		netdev_err(netdev, "Unable to alloc receive buffers.\n");
 		return;
 	}
diff --git a/drivers/net/ethernet/davicom/Kconfig b/drivers/net/ethernet/davicom/Kconfig
index 9745fe5..316c5e5 100644
--- a/drivers/net/ethernet/davicom/Kconfig
+++ b/drivers/net/ethernet/davicom/Kconfig
@@ -6,7 +6,6 @@ config DM9000
 	tristate "DM9000 support"
 	depends on ARM || BLACKFIN || MIPS || COLDFIRE
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  Support for DM9000 chipset.
diff --git a/drivers/net/ethernet/davicom/dm9000.c b/drivers/net/ethernet/davicom/dm9000.c
index 9105465..a13b312 100644
--- a/drivers/net/ethernet/davicom/dm9000.c
+++ b/drivers/net/ethernet/davicom/dm9000.c
@@ -29,6 +29,8 @@
 #include <linux/spinlock.h>
 #include <linux/crc32.h>
 #include <linux/mii.h>
+#include <linux/of.h>
+#include <linux/of_net.h>
 #include <linux/ethtool.h>
 #include <linux/dm9000.h>
 #include <linux/delay.h>
@@ -827,7 +829,7 @@ dm9000_hash_table_unlocked(struct net_device *dev)
 	struct netdev_hw_addr *ha;
 	int i, oft;
 	u32 hash_val;
-	u16 hash_table[4];
+	u16 hash_table[4] = { 0, 0, 0, 0x8000 }; /* broadcast address */
 	u8 rcr = RCR_DIS_LONG | RCR_DIS_CRC | RCR_RXEN;
 
 	dm9000_dbg(db, 1, "entering %s\n", __func__);
@@ -835,13 +837,6 @@ dm9000_hash_table_unlocked(struct net_device *dev)
 	for (i = 0, oft = DM9000_PAR; i < 6; i++, oft++)
 		iow(db, oft, dev->dev_addr[i]);
 
-	/* Clear Hash Table */
-	for (i = 0; i < 4; i++)
-		hash_table[i] = 0x0;
-
-	/* broadcast address */
-	hash_table[3] = 0x8000;
-
 	if (dev->flags & IFF_PROMISC)
 		rcr |= RCR_PRMSC;
 
@@ -1358,6 +1353,31 @@ static const struct net_device_ops dm9000_netdev_ops = {
 #endif
 };
 
+static struct dm9000_plat_data *dm9000_parse_dt(struct device *dev)
+{
+	struct dm9000_plat_data *pdata;
+	struct device_node *np = dev->of_node;
+	const void *mac_addr;
+
+	if (!IS_ENABLED(CONFIG_OF) || !np)
+		return NULL;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	if (of_find_property(np, "davicom,ext-phy", NULL))
+		pdata->flags |= DM9000_PLATF_EXT_PHY;
+	if (of_find_property(np, "davicom,no-eeprom", NULL))
+		pdata->flags |= DM9000_PLATF_NO_EEPROM;
+
+	mac_addr = of_get_mac_address(np);
+	if (mac_addr)
+		memcpy(pdata->dev_addr, mac_addr, sizeof(pdata->dev_addr));
+
+	return pdata;
+}
+
 /*
  * Search DM9000 board, allocate space and register it
  */
@@ -1373,6 +1393,12 @@ dm9000_probe(struct platform_device *pdev)
 	int i;
 	u32 id_val;
 
+	if (!pdata) {
+		pdata = dm9000_parse_dt(&pdev->dev);
+		if (IS_ERR(pdata))
+			return PTR_ERR(pdata);
+	}
+
 	/* Init network device */
 	ndev = alloc_etherdev(sizeof(struct board_info));
 	if (!ndev)
@@ -1673,8 +1699,6 @@ dm9000_drv_remove(struct platform_device *pdev)
 {
 	struct net_device *ndev = platform_get_drvdata(pdev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	unregister_netdev(ndev);
 	dm9000_release_board(pdev, netdev_priv(ndev));
 	free_netdev(ndev);		/* free device structure */
@@ -1683,11 +1707,20 @@ dm9000_drv_remove(struct platform_device *pdev)
 	return 0;
 }
 
+#ifdef CONFIG_OF
+static const struct of_device_id dm9000_of_matches[] = {
+	{ .compatible = "davicom,dm9000", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, dm9000_of_matches);
+#endif
+
 static struct platform_driver dm9000_driver = {
 	.driver	= {
 		.name    = "dm9000",
 		.owner	 = THIS_MODULE,
 		.pm	 = &dm9000_drv_pm_ops,
+		.of_match_table = of_match_ptr(dm9000_of_matches),
 	},
 	.probe   = dm9000_probe,
 	.remove  = dm9000_drv_remove,
diff --git a/drivers/net/ethernet/dec/tulip/Kconfig b/drivers/net/ethernet/dec/tulip/Kconfig
index 1df33c7..eb9ba6e 100644
--- a/drivers/net/ethernet/dec/tulip/Kconfig
+++ b/drivers/net/ethernet/dec/tulip/Kconfig
@@ -126,7 +126,6 @@ config WINBOND_840
 	tristate "Winbond W89c840 Ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver is for the Winbond W89c840 chip.  It also works with 
diff --git a/drivers/net/ethernet/dec/tulip/tulip_core.c b/drivers/net/ethernet/dec/tulip/tulip_core.c
index 1e9443d..c94152f 100644
--- a/drivers/net/ethernet/dec/tulip/tulip_core.c
+++ b/drivers/net/ethernet/dec/tulip/tulip_core.c
@@ -1410,12 +1410,6 @@ static int tulip_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 		return i;
 	}
 
-	/* The chip will fail to enter a low-power state later unless
-	 * first explicitly commanded into D0 */
-	if (pci_set_power_state(pdev, PCI_D0)) {
-		pr_notice("Failed to set power state to D0\n");
-	}
-
 	irq = pdev->irq;
 
 	/* alloc_etherdev ensures aligned and zeroed private structures */
diff --git a/drivers/net/ethernet/dec/tulip/xircom_cb.c b/drivers/net/ethernet/dec/tulip/xircom_cb.c
index cdbcd16..9b84cb0 100644
--- a/drivers/net/ethernet/dec/tulip/xircom_cb.c
+++ b/drivers/net/ethernet/dec/tulip/xircom_cb.c
@@ -1171,16 +1171,4 @@ investigate_write_descriptor(struct net_device *dev,
 	}
 }
 
-static int __init xircom_init(void)
-{
-	return pci_register_driver(&xircom_ops);
-}
-
-static void __exit xircom_exit(void)
-{
-	pci_unregister_driver(&xircom_ops);
-}
-
-module_init(xircom_init)
-module_exit(xircom_exit)
-
+module_pci_driver(xircom_ops);
diff --git a/drivers/net/ethernet/dlink/Kconfig b/drivers/net/ethernet/dlink/Kconfig
index ee26ce7..c543ac1 100644
--- a/drivers/net/ethernet/dlink/Kconfig
+++ b/drivers/net/ethernet/dlink/Kconfig
@@ -36,7 +36,6 @@ config SUNDANCE
 	tristate "Sundance Alta support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver is for the Sundance "Alta" chip.
diff --git a/drivers/net/ethernet/emulex/benet/be.h b/drivers/net/ethernet/emulex/benet/be.h
index 0a51068..c827b1b 100644
--- a/drivers/net/ethernet/emulex/benet/be.h
+++ b/drivers/net/ethernet/emulex/benet/be.h
@@ -333,6 +333,9 @@ enum vf_state {
 #define BE_VF_UC_PMAC_COUNT		2
 #define BE_FLAGS_QNQ_ASYNC_EVT_RCVD		(1 << 11)
 
+/* Ethtool set_dump flags */
+#define LANCER_INITIATE_FW_DUMP			0x1
+
 struct phy_info {
 	u8 transceiver;
 	u8 autoneg;
@@ -398,6 +401,7 @@ struct be_adapter {
 	u32 cmd_privileges;
 	/* Ethtool knobs and info */
 	char fw_ver[FW_VER_LEN];
+	char fw_on_flash[FW_VER_LEN];
 	int if_handle;		/* Used to configure filtering */
 	u32 *pmac_id;		/* MAC addr handle used by BE card */
 	u32 beacon_state;	/* for set_phys_id */
diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index 1db2df6..6e6e0a1 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c
@@ -3255,6 +3255,72 @@ err:
 	return status;
 }
 
+static int lancer_wait_idle(struct be_adapter *adapter)
+{
+#define SLIPORT_IDLE_TIMEOUT 30
+	u32 reg_val;
+	int status = 0, i;
+
+	for (i = 0; i < SLIPORT_IDLE_TIMEOUT; i++) {
+		reg_val = ioread32(adapter->db + PHYSDEV_CONTROL_OFFSET);
+		if ((reg_val & PHYSDEV_CONTROL_INP_MASK) == 0)
+			break;
+
+		ssleep(1);
+	}
+
+	if (i == SLIPORT_IDLE_TIMEOUT)
+		status = -1;
+
+	return status;
+}
+
+int lancer_physdev_ctrl(struct be_adapter *adapter, u32 mask)
+{
+	int status = 0;
+
+	status = lancer_wait_idle(adapter);
+	if (status)
+		return status;
+
+	iowrite32(mask, adapter->db + PHYSDEV_CONTROL_OFFSET);
+
+	return status;
+}
+
+/* Routine to check whether dump image is present or not */
+bool dump_present(struct be_adapter *adapter)
+{
+	u32 sliport_status = 0;
+
+	sliport_status = ioread32(adapter->db + SLIPORT_STATUS_OFFSET);
+	return !!(sliport_status & SLIPORT_STATUS_DIP_MASK);
+}
+
+int lancer_initiate_dump(struct be_adapter *adapter)
+{
+	int status;
+
+	/* give firmware reset and diagnostic dump */
+	status = lancer_physdev_ctrl(adapter, PHYSDEV_CONTROL_FW_RESET_MASK |
+				     PHYSDEV_CONTROL_DD_MASK);
+	if (status < 0) {
+		dev_err(&adapter->pdev->dev, "Firmware reset failed\n");
+		return status;
+	}
+
+	status = lancer_wait_idle(adapter);
+	if (status)
+		return status;
+
+	if (!dump_present(adapter)) {
+		dev_err(&adapter->pdev->dev, "Dump image not present\n");
+		return -1;
+	}
+
+	return 0;
+}
+
 /* Uses sync mcc */
 int be_cmd_enable_vf(struct be_adapter *adapter, u8 domain)
 {
diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.h b/drivers/net/ethernet/emulex/benet/be_cmds.h
index 025bdb0..5228d88 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.h
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.h
@@ -1937,6 +1937,9 @@ extern int be_cmd_set_ext_fat_capabilites(struct be_adapter *adapter,
 					  struct be_dma_mem *cmd,
 					  struct be_fat_conf_params *cfgs);
 extern int lancer_wait_ready(struct be_adapter *adapter);
+extern int lancer_physdev_ctrl(struct be_adapter *adapter, u32 mask);
+extern int lancer_initiate_dump(struct be_adapter *adapter);
+extern bool dump_present(struct be_adapter *adapter);
 extern int lancer_test_and_set_rdy_state(struct be_adapter *adapter);
 extern int be_cmd_query_port_name(struct be_adapter *adapter, u8 *port_name);
 extern int be_cmd_get_func_config(struct be_adapter *adapter);
diff --git a/drivers/net/ethernet/emulex/benet/be_ethtool.c b/drivers/net/ethernet/emulex/benet/be_ethtool.c
index 3d4461a..4f8c941 100644
--- a/drivers/net/ethernet/emulex/benet/be_ethtool.c
+++ b/drivers/net/ethernet/emulex/benet/be_ethtool.c
@@ -177,19 +177,15 @@ static void be_get_drvinfo(struct net_device *netdev,
 				struct ethtool_drvinfo *drvinfo)
 {
 	struct be_adapter *adapter = netdev_priv(netdev);
-	char fw_on_flash[FW_VER_LEN];
-
-	memset(fw_on_flash, 0 , sizeof(fw_on_flash));
-	be_cmd_get_fw_ver(adapter, adapter->fw_ver, fw_on_flash);
 
 	strlcpy(drvinfo->driver, DRV_NAME, sizeof(drvinfo->driver));
 	strlcpy(drvinfo->version, DRV_VER, sizeof(drvinfo->version));
-	if (!memcmp(adapter->fw_ver, fw_on_flash, FW_VER_LEN))
+	if (!memcmp(adapter->fw_ver, adapter->fw_on_flash, FW_VER_LEN))
 		strlcpy(drvinfo->fw_version, adapter->fw_ver,
 			sizeof(drvinfo->fw_version));
 	else
 		snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
-			 "%s [%s]", adapter->fw_ver, fw_on_flash);
+			 "%s [%s]", adapter->fw_ver, adapter->fw_on_flash);
 
 	strlcpy(drvinfo->bus_info, pci_name(adapter->pdev),
 		sizeof(drvinfo->bus_info));
@@ -673,6 +669,34 @@ be_set_phys_id(struct net_device *netdev,
 	return 0;
 }
 
+static int be_set_dump(struct net_device *netdev, struct ethtool_dump *dump)
+{
+	struct be_adapter *adapter = netdev_priv(netdev);
+	struct device *dev = &adapter->pdev->dev;
+	int status;
+
+	if (!lancer_chip(adapter)) {
+		dev_err(dev, "FW dump not supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (dump_present(adapter)) {
+		dev_err(dev, "Previous dump not cleared, not forcing dump\n");
+		return 0;
+	}
+
+	switch (dump->flag) {
+	case LANCER_INITIATE_FW_DUMP:
+		status = lancer_initiate_dump(adapter);
+		if (!status)
+			dev_info(dev, "F/w dump initiated successfully\n");
+		break;
+	default:
+		dev_err(dev, "Invalid dump level: 0x%x\n", dump->flag);
+		return -EINVAL;
+	}
+	return status;
+}
 
 static void
 be_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
@@ -1110,6 +1134,7 @@ const struct ethtool_ops be_ethtool_ops = {
 	.set_pauseparam = be_set_pauseparam,
 	.get_strings = be_get_stat_strings,
 	.set_phys_id = be_set_phys_id,
+	.set_dump = be_set_dump,
 	.get_msglevel = be_get_msg_level,
 	.set_msglevel = be_set_msg_level,
 	.get_sset_count = be_get_sset_count,
diff --git a/drivers/net/ethernet/emulex/benet/be_hw.h b/drivers/net/ethernet/emulex/benet/be_hw.h
index 8780183..3e21621 100644
--- a/drivers/net/ethernet/emulex/benet/be_hw.h
+++ b/drivers/net/ethernet/emulex/benet/be_hw.h
@@ -53,10 +53,12 @@
 #define PHYSDEV_CONTROL_OFFSET		0x414
 
 #define SLIPORT_STATUS_ERR_MASK		0x80000000
+#define SLIPORT_STATUS_DIP_MASK		0x02000000
 #define SLIPORT_STATUS_RN_MASK		0x01000000
 #define SLIPORT_STATUS_RDY_MASK		0x00800000
 #define SLI_PORT_CONTROL_IP_MASK	0x08000000
 #define PHYSDEV_CONTROL_FW_RESET_MASK	0x00000002
+#define PHYSDEV_CONTROL_DD_MASK		0x00000004
 #define PHYSDEV_CONTROL_INP_MASK	0x40000000
 
 #define SLIPORT_ERROR_NO_RESOURCE1	0x2
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index a0b4be5..2df48bb 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -834,32 +834,39 @@ static int be_vlan_tag_tx_chk(struct be_adapter *adapter, struct sk_buff *skb)
 	return vlan_tx_tag_present(skb) || adapter->pvid || adapter->qnq_vid;
 }
 
-static int be_ipv6_tx_stall_chk(struct be_adapter *adapter, struct sk_buff *skb)
+static int be_ipv6_tx_stall_chk(struct be_adapter *adapter,
+				struct sk_buff *skb)
 {
-	return BE3_chip(adapter) &&
-		be_ipv6_exthdr_check(skb);
+	return BE3_chip(adapter) && be_ipv6_exthdr_check(skb);
 }
 
-static netdev_tx_t be_xmit(struct sk_buff *skb,
-			struct net_device *netdev)
+static struct sk_buff *be_xmit_workarounds(struct be_adapter *adapter,
+					   struct sk_buff *skb,
+					   bool *skip_hw_vlan)
 {
-	struct be_adapter *adapter = netdev_priv(netdev);
-	struct be_tx_obj *txo = &adapter->tx_obj[skb_get_queue_mapping(skb)];
-	struct be_queue_info *txq = &txo->q;
-	struct iphdr *ip = NULL;
-	u32 wrb_cnt = 0, copied = 0;
-	u32 start = txq->head, eth_hdr_len;
-	bool dummy_wrb, stopped = false;
-	bool skip_hw_vlan = false;
 	struct vlan_ethhdr *veh = (struct vlan_ethhdr *)skb->data;
+	unsigned int eth_hdr_len;
+	struct iphdr *ip;
 
-	eth_hdr_len = ntohs(skb->protocol) == ETH_P_8021Q ?
-		VLAN_ETH_HLEN : ETH_HLEN;
+	/* Lancer ASIC has a bug wherein packets that are 32 bytes or less
+	 * may cause a transmit stall on that port. So the work-around is to
+	 * pad such packets to a 36-byte length.
+	 */
+	if (unlikely(lancer_chip(adapter) && skb->len <= 32)) {
+		if (skb_padto(skb, 36))
+			goto tx_drop;
+		skb->len = 36;
+	}
 
 	/* For padded packets, BE HW modifies tot_len field in IP header
 	 * incorrecly when VLAN tag is inserted by HW.
+	 * For padded packets, Lancer computes incorrect checksum.
 	 */
-	if (skb->len <= 60 && vlan_tx_tag_present(skb) && is_ipv4_pkt(skb)) {
+	eth_hdr_len = ntohs(skb->protocol) == ETH_P_8021Q ?
+						VLAN_ETH_HLEN : ETH_HLEN;
+	if (skb->len <= 60 &&
+	    (lancer_chip(adapter) || vlan_tx_tag_present(skb)) &&
+	    is_ipv4_pkt(skb)) {
 		ip = (struct iphdr *)ip_hdr(skb);
 		pskb_trim(skb, eth_hdr_len + ntohs(ip->tot_len));
 	}
@@ -869,15 +876,15 @@ static netdev_tx_t be_xmit(struct sk_buff *skb,
 	 */
 	if ((adapter->function_mode & UMC_ENABLED) &&
 	    veh->h_vlan_proto == htons(ETH_P_8021Q))
-			skip_hw_vlan = true;
+			*skip_hw_vlan = true;
 
 	/* HW has a bug wherein it will calculate CSUM for VLAN
 	 * pkts even though it is disabled.
 	 * Manually insert VLAN in pkt.
 	 */
 	if (skb->ip_summed != CHECKSUM_PARTIAL &&
-			vlan_tx_tag_present(skb)) {
-		skb = be_insert_vlan_in_pkt(adapter, skb, &skip_hw_vlan);
+	    vlan_tx_tag_present(skb)) {
+		skb = be_insert_vlan_in_pkt(adapter, skb, skip_hw_vlan);
 		if (unlikely(!skb))
 			goto tx_drop;
 	}
@@ -887,8 +894,8 @@ static netdev_tx_t be_xmit(struct sk_buff *skb,
 	 * skip HW tagging is not enabled by FW.
 	 */
 	if (unlikely(be_ipv6_tx_stall_chk(adapter, skb) &&
-		     (adapter->pvid || adapter->qnq_vid) &&
-		     !qnq_async_evt_rcvd(adapter)))
+	    (adapter->pvid || adapter->qnq_vid) &&
+	    !qnq_async_evt_rcvd(adapter)))
 		goto tx_drop;
 
 	/* Manual VLAN tag insertion to prevent:
@@ -899,11 +906,31 @@ static netdev_tx_t be_xmit(struct sk_buff *skb,
 	 */
 	if (be_ipv6_tx_stall_chk(adapter, skb) &&
 	    be_vlan_tag_tx_chk(adapter, skb)) {
-		skb = be_insert_vlan_in_pkt(adapter, skb, &skip_hw_vlan);
+		skb = be_insert_vlan_in_pkt(adapter, skb, skip_hw_vlan);
 		if (unlikely(!skb))
 			goto tx_drop;
 	}
 
+	return skb;
+tx_drop:
+	dev_kfree_skb_any(skb);
+	return NULL;
+}
+
+static netdev_tx_t be_xmit(struct sk_buff *skb, struct net_device *netdev)
+{
+	struct be_adapter *adapter = netdev_priv(netdev);
+	struct be_tx_obj *txo = &adapter->tx_obj[skb_get_queue_mapping(skb)];
+	struct be_queue_info *txq = &txo->q;
+	bool dummy_wrb, stopped = false;
+	u32 wrb_cnt = 0, copied = 0;
+	bool skip_hw_vlan = false;
+	u32 start = txq->head;
+
+	skb = be_xmit_workarounds(adapter, skb, &skip_hw_vlan);
+	if (!skb)
+		return NETDEV_TX_OK;
+
 	wrb_cnt = wrb_cnt_for_skb(adapter, skb, &dummy_wrb);
 
 	copied = make_tx_wrbs(adapter, txq, skb, wrb_cnt, dummy_wrb,
@@ -933,7 +960,6 @@ static netdev_tx_t be_xmit(struct sk_buff *skb,
 		txq->head = start;
 		dev_kfree_skb_any(skb);
 	}
-tx_drop:
 	return NETDEV_TX_OK;
 }
 
@@ -1236,30 +1262,6 @@ static int be_set_vf_tx_rate(struct net_device *netdev,
 	return status;
 }
 
-static int be_find_vfs(struct be_adapter *adapter, int vf_state)
-{
-	struct pci_dev *dev, *pdev = adapter->pdev;
-	int vfs = 0, assigned_vfs = 0, pos;
-	u16 offset, stride;
-
-	pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_SRIOV);
-	if (!pos)
-		return 0;
-	pci_read_config_word(pdev, pos + PCI_SRIOV_VF_OFFSET, &offset);
-	pci_read_config_word(pdev, pos + PCI_SRIOV_VF_STRIDE, &stride);
-
-	dev = pci_get_device(pdev->vendor, PCI_ANY_ID, NULL);
-	while (dev) {
-		if (dev->is_virtfn && pci_physfn(dev) == pdev) {
-			vfs++;
-			if (dev->dev_flags & PCI_DEV_FLAGS_ASSIGNED)
-				assigned_vfs++;
-		}
-		dev = pci_get_device(pdev->vendor, PCI_ANY_ID, dev);
-	}
-	return (vf_state == ASSIGNED) ? assigned_vfs : vfs;
-}
-
 static void be_eqd_update(struct be_adapter *adapter, struct be_eq_obj *eqo)
 {
 	struct be_rx_stats *stats = rx_stats(&adapter->rx_obj[eqo->idx]);
@@ -2771,7 +2773,7 @@ static void be_vf_clear(struct be_adapter *adapter)
 	struct be_vf_cfg *vf_cfg;
 	u32 vf;
 
-	if (be_find_vfs(adapter, ASSIGNED)) {
+	if (pci_vfs_assigned(adapter->pdev)) {
 		dev_warn(&adapter->pdev->dev,
 			 "VFs are assigned to VMs: not disabling VFs\n");
 		goto done;
@@ -2873,7 +2875,7 @@ static int be_vf_setup(struct be_adapter *adapter)
 	int status, old_vfs, vf;
 	struct device *dev = &adapter->pdev->dev;
 
-	old_vfs = be_find_vfs(adapter, ENABLED);
+	old_vfs = pci_num_vf(adapter->pdev);
 	if (old_vfs) {
 		dev_info(dev, "%d VFs are already enabled\n", old_vfs);
 		if (old_vfs != num_vfs)
@@ -3184,7 +3186,7 @@ static int be_setup(struct be_adapter *adapter)
 	if (status)
 		goto err;
 
-	be_cmd_get_fw_ver(adapter, adapter->fw_ver, NULL);
+	be_cmd_get_fw_ver(adapter, adapter->fw_ver, adapter->fw_on_flash);
 
 	if (adapter->vlans_added)
 		be_vid_config(adapter);
@@ -3530,40 +3532,6 @@ static int be_flash_skyhawk(struct be_adapter *adapter,
 	return 0;
 }
 
-static int lancer_wait_idle(struct be_adapter *adapter)
-{
-#define SLIPORT_IDLE_TIMEOUT 30
-	u32 reg_val;
-	int status = 0, i;
-
-	for (i = 0; i < SLIPORT_IDLE_TIMEOUT; i++) {
-		reg_val = ioread32(adapter->db + PHYSDEV_CONTROL_OFFSET);
-		if ((reg_val & PHYSDEV_CONTROL_INP_MASK) == 0)
-			break;
-
-		ssleep(1);
-	}
-
-	if (i == SLIPORT_IDLE_TIMEOUT)
-		status = -1;
-
-	return status;
-}
-
-static int lancer_fw_reset(struct be_adapter *adapter)
-{
-	int status = 0;
-
-	status = lancer_wait_idle(adapter);
-	if (status)
-		return status;
-
-	iowrite32(PHYSDEV_CONTROL_FW_RESET_MASK, adapter->db +
-		  PHYSDEV_CONTROL_OFFSET);
-
-	return status;
-}
-
 static int lancer_fw_download(struct be_adapter *adapter,
 				const struct firmware *fw)
 {
@@ -3641,7 +3609,8 @@ static int lancer_fw_download(struct be_adapter *adapter,
 	}
 
 	if (change_status == LANCER_FW_RESET_NEEDED) {
-		status = lancer_fw_reset(adapter);
+		status = lancer_physdev_ctrl(adapter,
+					     PHYSDEV_CONTROL_FW_RESET_MASK);
 		if (status) {
 			dev_err(&adapter->pdev->dev,
 				"Adapter busy for FW reset.\n"
@@ -3776,6 +3745,10 @@ int be_load_fw(struct be_adapter *adapter, u8 *fw_file)
 	else
 		status = be_fw_download(adapter, fw);
 
+	if (!status)
+		be_cmd_get_fw_ver(adapter, adapter->fw_ver,
+				  adapter->fw_on_flash);
+
 fw_exit:
 	release_firmware(fw);
 	return status;
@@ -4203,9 +4176,10 @@ reschedule:
 	schedule_delayed_work(&adapter->work, msecs_to_jiffies(1000));
 }
 
+/* If any VFs are already enabled don't FLR the PF */
 static bool be_reset_required(struct be_adapter *adapter)
 {
-	return be_find_vfs(adapter, ENABLED) > 0 ? false : true;
+	return pci_num_vf(adapter->pdev) ? false : true;
 }
 
 static char *mc_name(struct be_adapter *adapter)
@@ -4390,7 +4364,7 @@ static int be_resume(struct pci_dev *pdev)
 	if (status)
 		return status;
 
-	pci_set_power_state(pdev, 0);
+	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 
 	/* tell fw we're ready to fire cmds */
@@ -4486,7 +4460,7 @@ static pci_ers_result_t be_eeh_reset(struct pci_dev *pdev)
 		return PCI_ERS_RESULT_DISCONNECT;
 
 	pci_set_master(pdev);
-	pci_set_power_state(pdev, 0);
+	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 
 	/* Check if card is ok and fw is ready */
diff --git a/drivers/net/ethernet/ethoc.c b/drivers/net/ethernet/ethoc.c
index 5722bc6..cf579fb 100644
--- a/drivers/net/ethernet/ethoc.c
+++ b/drivers/net/ethernet/ethoc.c
@@ -1147,8 +1147,6 @@ static int ethoc_remove(struct platform_device *pdev)
 	struct net_device *netdev = platform_get_drvdata(pdev);
 	struct ethoc *priv = netdev_priv(netdev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	if (netdev) {
 		netif_napi_del(&priv->napi);
 		phy_disconnect(priv->phy);
diff --git a/drivers/net/ethernet/faraday/Kconfig b/drivers/net/ethernet/faraday/Kconfig
index b8974b9..5918c68 100644
--- a/drivers/net/ethernet/faraday/Kconfig
+++ b/drivers/net/ethernet/faraday/Kconfig
@@ -21,7 +21,6 @@ if NET_VENDOR_FARADAY
 config FTMAC100
 	tristate "Faraday FTMAC100 10/100 Ethernet support"
 	depends on ARM
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports the FTMAC100 10/100 Ethernet controller
diff --git a/drivers/net/ethernet/faraday/ftgmac100.c b/drivers/net/ethernet/faraday/ftgmac100.c
index 21b85fb..934e1ae 100644
--- a/drivers/net/ethernet/faraday/ftgmac100.c
+++ b/drivers/net/ethernet/faraday/ftgmac100.c
@@ -1311,7 +1311,6 @@ err_ioremap:
 	release_resource(priv->res);
 err_req_mem:
 	netif_napi_del(&priv->napi);
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
 err_alloc_etherdev:
 	return err;
@@ -1335,7 +1334,6 @@ static int __exit ftgmac100_remove(struct platform_device *pdev)
 	release_resource(priv->res);
 
 	netif_napi_del(&priv->napi);
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
 	return 0;
 }
diff --git a/drivers/net/ethernet/faraday/ftmac100.c b/drivers/net/ethernet/faraday/ftmac100.c
index a6eda8d..4658f4c 100644
--- a/drivers/net/ethernet/faraday/ftmac100.c
+++ b/drivers/net/ethernet/faraday/ftmac100.c
@@ -1149,7 +1149,6 @@ err_ioremap:
 	release_resource(priv->res);
 err_req_mem:
 	netif_napi_del(&priv->napi);
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
 err_alloc_etherdev:
 	return err;
@@ -1169,7 +1168,6 @@ static int __exit ftmac100_remove(struct platform_device *pdev)
 	release_resource(priv->res);
 
 	netif_napi_del(&priv->napi);
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(netdev);
 	return 0;
 }
diff --git a/drivers/net/ethernet/freescale/fec.h b/drivers/net/ethernet/freescale/fec.h
index 9ce5b71..2b0a0ea 100644
--- a/drivers/net/ethernet/freescale/fec.h
+++ b/drivers/net/ethernet/freescale/fec.h
@@ -60,6 +60,61 @@
 #define BM_MIIGSK_CFGR_RMII		0x01
 #define BM_MIIGSK_CFGR_FRCONT_10M	0x40
 
+#define RMON_T_DROP		0x200 /* Count of frames not cntd correctly */
+#define RMON_T_PACKETS		0x204 /* RMON TX packet count */
+#define RMON_T_BC_PKT		0x208 /* RMON TX broadcast pkts */
+#define RMON_T_MC_PKT		0x20C /* RMON TX multicast pkts */
+#define RMON_T_CRC_ALIGN	0x210 /* RMON TX pkts with CRC align err */
+#define RMON_T_UNDERSIZE	0x214 /* RMON TX pkts < 64 bytes, good CRC */
+#define RMON_T_OVERSIZE		0x218 /* RMON TX pkts > MAX_FL bytes good CRC */
+#define RMON_T_FRAG		0x21C /* RMON TX pkts < 64 bytes, bad CRC */
+#define RMON_T_JAB		0x220 /* RMON TX pkts > MAX_FL bytes, bad CRC */
+#define RMON_T_COL		0x224 /* RMON TX collision count */
+#define RMON_T_P64		0x228 /* RMON TX 64 byte pkts */
+#define RMON_T_P65TO127		0x22C /* RMON TX 65 to 127 byte pkts */
+#define RMON_T_P128TO255	0x230 /* RMON TX 128 to 255 byte pkts */
+#define RMON_T_P256TO511	0x234 /* RMON TX 256 to 511 byte pkts */
+#define RMON_T_P512TO1023	0x238 /* RMON TX 512 to 1023 byte pkts */
+#define RMON_T_P1024TO2047	0x23C /* RMON TX 1024 to 2047 byte pkts */
+#define RMON_T_P_GTE2048	0x240 /* RMON TX pkts > 2048 bytes */
+#define RMON_T_OCTETS		0x244 /* RMON TX octets */
+#define IEEE_T_DROP		0x248 /* Count of frames not counted crtly */
+#define IEEE_T_FRAME_OK		0x24C /* Frames tx'd OK */
+#define IEEE_T_1COL		0x250 /* Frames tx'd with single collision */
+#define IEEE_T_MCOL		0x254 /* Frames tx'd with multiple collision */
+#define IEEE_T_DEF		0x258 /* Frames tx'd after deferral delay */
+#define IEEE_T_LCOL		0x25C /* Frames tx'd with late collision */
+#define IEEE_T_EXCOL		0x260 /* Frames tx'd with excesv collisions */
+#define IEEE_T_MACERR		0x264 /* Frames tx'd with TX FIFO underrun */
+#define IEEE_T_CSERR		0x268 /* Frames tx'd with carrier sense err */
+#define IEEE_T_SQE		0x26C /* Frames tx'd with SQE err */
+#define IEEE_T_FDXFC		0x270 /* Flow control pause frames tx'd */
+#define IEEE_T_OCTETS_OK	0x274 /* Octet count for frames tx'd w/o err */
+#define RMON_R_PACKETS		0x284 /* RMON RX packet count */
+#define RMON_R_BC_PKT		0x288 /* RMON RX broadcast pkts */
+#define RMON_R_MC_PKT		0x28C /* RMON RX multicast pkts */
+#define RMON_R_CRC_ALIGN	0x290 /* RMON RX pkts with CRC alignment err */
+#define RMON_R_UNDERSIZE	0x294 /* RMON RX pkts < 64 bytes, good CRC */
+#define RMON_R_OVERSIZE		0x298 /* RMON RX pkts > MAX_FL bytes good CRC */
+#define RMON_R_FRAG		0x29C /* RMON RX pkts < 64 bytes, bad CRC */
+#define RMON_R_JAB		0x2A0 /* RMON RX pkts > MAX_FL bytes, bad CRC */
+#define RMON_R_RESVD_O		0x2A4 /* Reserved */
+#define RMON_R_P64		0x2A8 /* RMON RX 64 byte pkts */
+#define RMON_R_P65TO127		0x2AC /* RMON RX 65 to 127 byte pkts */
+#define RMON_R_P128TO255	0x2B0 /* RMON RX 128 to 255 byte pkts */
+#define RMON_R_P256TO511	0x2B4 /* RMON RX 256 to 511 byte pkts */
+#define RMON_R_P512TO1023	0x2B8 /* RMON RX 512 to 1023 byte pkts */
+#define RMON_R_P1024TO2047	0x2BC /* RMON RX 1024 to 2047 byte pkts */
+#define RMON_R_P_GTE2048	0x2C0 /* RMON RX pkts > 2048 bytes */
+#define RMON_R_OCTETS		0x2C4 /* RMON RX octets */
+#define IEEE_R_DROP		0x2C8 /* Count frames not counted correctly */
+#define IEEE_R_FRAME_OK		0x2CC /* Frames rx'd OK */
+#define IEEE_R_CRC		0x2D0 /* Frames rx'd with CRC err */
+#define IEEE_R_ALIGN		0x2D4 /* Frames rx'd with alignment err */
+#define IEEE_R_MACERR		0x2D8 /* Receive FIFO overflow count */
+#define IEEE_R_FDXFC		0x2DC /* Flow control pause frames rx'd */
+#define IEEE_R_OCTETS_OK	0x2E0 /* Octet cnt for frames rx'd w/o err */
+
 #else
 
 #define FEC_ECNTRL		0x000 /* Ethernet control reg */
@@ -148,6 +203,9 @@ struct bufdesc_ex {
 #define BD_ENET_RX_CL           ((ushort)0x0001)
 #define BD_ENET_RX_STATS        ((ushort)0x013f)        /* All status bits */
 
+/* Enhanced buffer descriptor control/status used by Ethernet receive */
+#define BD_ENET_RX_VLAN         0x00000004
+
 /* Buffer descriptor control/status used by Ethernet transmit.
 */
 #define BD_ENET_TX_READY        ((ushort)0x8000)
@@ -272,9 +330,10 @@ struct fec_enet_private {
 	int hwts_tx_en;
 	struct timer_list time_keep;
 	struct fec_enet_delayed_work delay_work;
+	struct regulator *reg_phy;
 };
 
-void fec_ptp_init(struct net_device *ndev, struct platform_device *pdev);
+void fec_ptp_init(struct platform_device *pdev);
 void fec_ptp_start_cyclecounter(struct net_device *ndev);
 int fec_ptp_ioctl(struct net_device *ndev, struct ifreq *ifr, int cmd);
 
diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index d48099f..d3ad5ea 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -53,13 +53,15 @@
 #include <linux/of_device.h>
 #include <linux/of_gpio.h>
 #include <linux/of_net.h>
-#include <linux/pinctrl/consumer.h>
 #include <linux/regulator/consumer.h>
+#include <linux/if_vlan.h>
 
 #include <asm/cacheflush.h>
 
 #include "fec.h"
 
+static void set_multicast_list(struct net_device *ndev);
+
 #if defined(CONFIG_ARM)
 #define FEC_ALIGNMENT	0xf
 #else
@@ -89,6 +91,8 @@
 #define FEC_QUIRK_HAS_BUFDESC_EX	(1 << 4)
 /* Controller has hardware checksum support */
 #define FEC_QUIRK_HAS_CSUM		(1 << 5)
+/* Controller has hardware vlan support */
+#define FEC_QUIRK_HAS_VLAN		(1 << 6)
 
 static struct platform_device_id fec_devtype[] = {
 	{
@@ -107,7 +111,8 @@ static struct platform_device_id fec_devtype[] = {
 	}, {
 		.name = "imx6q-fec",
 		.driver_data = FEC_QUIRK_ENET_MAC | FEC_QUIRK_HAS_GBIT |
-				FEC_QUIRK_HAS_BUFDESC_EX | FEC_QUIRK_HAS_CSUM,
+				FEC_QUIRK_HAS_BUFDESC_EX | FEC_QUIRK_HAS_CSUM |
+				FEC_QUIRK_HAS_VLAN,
 	}, {
 		.name = "mvf600-fec",
 		.driver_data = FEC_QUIRK_ENET_MAC,
@@ -178,11 +183,11 @@ MODULE_PARM_DESC(macaddr, "FEC Ethernet MAC address");
 #define FEC_DEFAULT_IMASK (FEC_ENET_TXF | FEC_ENET_RXF | FEC_ENET_MII)
 #define FEC_RX_DISABLED_IMASK (FEC_DEFAULT_IMASK & (~FEC_ENET_RXF))
 
-/* The FEC stores dest/src/type, data, and checksum for receive packets.
+/* The FEC stores dest/src/type/vlan, data, and checksum for receive packets.
  */
-#define PKT_MAXBUF_SIZE		1518
+#define PKT_MAXBUF_SIZE		1522
 #define PKT_MINBUF_SIZE		64
-#define PKT_MAXBLR_SIZE		1520
+#define PKT_MAXBLR_SIZE		1536
 
 /* FEC receive acceleration */
 #define FEC_RACC_IPDIS		(1 << 1)
@@ -243,7 +248,7 @@ static void *swap_buffer(void *bufaddr, int len)
 	int i;
 	unsigned int *buf = bufaddr;
 
-	for (i = 0; i < (len + 3) / 4; i++, buf++)
+	for (i = 0; i < DIV_ROUND_UP(len, 4); i++, buf++)
 		*buf = cpu_to_be32(*buf);
 
 	return bufaddr;
@@ -471,9 +476,8 @@ fec_restart(struct net_device *ndev, int duplex)
 	/* Clear any outstanding interrupt. */
 	writel(0xffc00000, fep->hwp + FEC_IEVENT);
 
-	/* Reset all multicast.	*/
-	writel(0, fep->hwp + FEC_GRP_HASH_TABLE_HIGH);
-	writel(0, fep->hwp + FEC_GRP_HASH_TABLE_LOW);
+	/* Setup multicast filter. */
+	set_multicast_list(ndev);
 #ifndef CONFIG_M5272
 	writel(0, fep->hwp + FEC_HASH_TABLE_HIGH);
 	writel(0, fep->hwp + FEC_HASH_TABLE_LOW);
@@ -609,6 +613,11 @@ fec_restart(struct net_device *ndev, int duplex)
 	if (fep->bufdesc_ex)
 		ecntl |= (1 << 4);
 
+#ifndef CONFIG_M5272
+	/* Enable the MIB statistic event counters */
+	writel(0 << 31, fep->hwp + FEC_MIB_CTRLSTAT);
+#endif
+
 	/* And last, enable the transmit and receive processing */
 	writel(ecntl, fep->hwp + FEC_ECNTRL);
 	writel(0, fep->hwp + FEC_R_DES_ACTIVE);
@@ -735,6 +744,7 @@ fec_enet_tx(struct net_device *ndev)
 				ndev->stats.tx_carrier_errors++;
 		} else {
 			ndev->stats.tx_packets++;
+			ndev->stats.tx_bytes += bdp->cbd_datlen;
 		}
 
 		if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_IN_PROGRESS) &&
@@ -800,6 +810,9 @@ fec_enet_rx(struct net_device *ndev, int budget)
 	ushort	pkt_len;
 	__u8 *data;
 	int	pkt_received = 0;
+	struct	bufdesc_ex *ebdp = NULL;
+	bool	vlan_packet_rcvd = false;
+	u16	vlan_tag;
 
 #ifdef CONFIG_M532x
 	flush_cache_all();
@@ -863,6 +876,24 @@ fec_enet_rx(struct net_device *ndev, int budget)
 		if (id_entry->driver_data & FEC_QUIRK_SWAP_FRAME)
 			swap_buffer(data, pkt_len);
 
+		/* Extract the enhanced buffer descriptor */
+		ebdp = NULL;
+		if (fep->bufdesc_ex)
+			ebdp = (struct bufdesc_ex *)bdp;
+
+		/* If this is a VLAN packet remove the VLAN Tag */
+		vlan_packet_rcvd = false;
+		if ((ndev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
+		    fep->bufdesc_ex && (ebdp->cbd_esc & BD_ENET_RX_VLAN)) {
+			/* Push and remove the vlan tag */
+			struct vlan_hdr *vlan_header =
+					(struct vlan_hdr *) (data + ETH_HLEN);
+			vlan_tag = ntohs(vlan_header->h_vlan_TCI);
+			pkt_len -= VLAN_HLEN;
+
+			vlan_packet_rcvd = true;
+		}
+
 		/* This does 16 byte alignment, exactly what we need.
 		 * The packet length includes FCS, but we don't want to
 		 * include that when passing upstream as it messes up
@@ -873,9 +904,18 @@ fec_enet_rx(struct net_device *ndev, int budget)
 		if (unlikely(!skb)) {
 			ndev->stats.rx_dropped++;
 		} else {
+			int payload_offset = (2 * ETH_ALEN);
 			skb_reserve(skb, NET_IP_ALIGN);
 			skb_put(skb, pkt_len - 4);	/* Make room */
-			skb_copy_to_linear_data(skb, data, pkt_len - 4);
+
+			/* Extract the frame data without the VLAN header. */
+			skb_copy_to_linear_data(skb, data, (2 * ETH_ALEN));
+			if (vlan_packet_rcvd)
+				payload_offset = (2 * ETH_ALEN) + VLAN_HLEN;
+			skb_copy_to_linear_data_offset(skb, (2 * ETH_ALEN),
+						       data + payload_offset,
+						       pkt_len - 4 - (2 * ETH_ALEN));
+
 			skb->protocol = eth_type_trans(skb, ndev);
 
 			/* Get receive timestamp from the skb */
@@ -883,8 +923,6 @@ fec_enet_rx(struct net_device *ndev, int budget)
 				struct skb_shared_hwtstamps *shhwtstamps =
 							    skb_hwtstamps(skb);
 				unsigned long flags;
-				struct bufdesc_ex *ebdp =
-					(struct bufdesc_ex *)bdp;
 
 				memset(shhwtstamps, 0, sizeof(*shhwtstamps));
 
@@ -895,9 +933,7 @@ fec_enet_rx(struct net_device *ndev, int budget)
 			}
 
 			if (fep->bufdesc_ex &&
-				(fep->csum_flags & FLAG_RX_CSUM_ENABLED)) {
-				struct bufdesc_ex *ebdp =
-					(struct bufdesc_ex *)bdp;
+			    (fep->csum_flags & FLAG_RX_CSUM_ENABLED)) {
 				if (!(ebdp->cbd_esc & FLAG_RX_CSUM_ERROR)) {
 					/* don't check it */
 					skb->ip_summed = CHECKSUM_UNNECESSARY;
@@ -906,6 +942,12 @@ fec_enet_rx(struct net_device *ndev, int budget)
 				}
 			}
 
+			/* Handle received VLAN packets */
+			if (vlan_packet_rcvd)
+				__vlan_hwaccel_put_tag(skb,
+						       htons(ETH_P_8021Q),
+						       vlan_tag);
+
 			if (!skb_defer_rx_timestamp(skb))
 				napi_gro_receive(&fep->napi, skb);
 		}
@@ -1444,8 +1486,117 @@ static int fec_enet_set_pauseparam(struct net_device *ndev,
 	return 0;
 }
 
+static const struct fec_stat {
+	char name[ETH_GSTRING_LEN];
+	u16 offset;
+} fec_stats[] = {
+	/* RMON TX */
+	{ "tx_dropped", RMON_T_DROP },
+	{ "tx_packets", RMON_T_PACKETS },
+	{ "tx_broadcast", RMON_T_BC_PKT },
+	{ "tx_multicast", RMON_T_MC_PKT },
+	{ "tx_crc_errors", RMON_T_CRC_ALIGN },
+	{ "tx_undersize", RMON_T_UNDERSIZE },
+	{ "tx_oversize", RMON_T_OVERSIZE },
+	{ "tx_fragment", RMON_T_FRAG },
+	{ "tx_jabber", RMON_T_JAB },
+	{ "tx_collision", RMON_T_COL },
+	{ "tx_64byte", RMON_T_P64 },
+	{ "tx_65to127byte", RMON_T_P65TO127 },
+	{ "tx_128to255byte", RMON_T_P128TO255 },
+	{ "tx_256to511byte", RMON_T_P256TO511 },
+	{ "tx_512to1023byte", RMON_T_P512TO1023 },
+	{ "tx_1024to2047byte", RMON_T_P1024TO2047 },
+	{ "tx_GTE2048byte", RMON_T_P_GTE2048 },
+	{ "tx_octets", RMON_T_OCTETS },
+
+	/* IEEE TX */
+	{ "IEEE_tx_drop", IEEE_T_DROP },
+	{ "IEEE_tx_frame_ok", IEEE_T_FRAME_OK },
+	{ "IEEE_tx_1col", IEEE_T_1COL },
+	{ "IEEE_tx_mcol", IEEE_T_MCOL },
+	{ "IEEE_tx_def", IEEE_T_DEF },
+	{ "IEEE_tx_lcol", IEEE_T_LCOL },
+	{ "IEEE_tx_excol", IEEE_T_EXCOL },
+	{ "IEEE_tx_macerr", IEEE_T_MACERR },
+	{ "IEEE_tx_cserr", IEEE_T_CSERR },
+	{ "IEEE_tx_sqe", IEEE_T_SQE },
+	{ "IEEE_tx_fdxfc", IEEE_T_FDXFC },
+	{ "IEEE_tx_octets_ok", IEEE_T_OCTETS_OK },
+
+	/* RMON RX */
+	{ "rx_packets", RMON_R_PACKETS },
+	{ "rx_broadcast", RMON_R_BC_PKT },
+	{ "rx_multicast", RMON_R_MC_PKT },
+	{ "rx_crc_errors", RMON_R_CRC_ALIGN },
+	{ "rx_undersize", RMON_R_UNDERSIZE },
+	{ "rx_oversize", RMON_R_OVERSIZE },
+	{ "rx_fragment", RMON_R_FRAG },
+	{ "rx_jabber", RMON_R_JAB },
+	{ "rx_64byte", RMON_R_P64 },
+	{ "rx_65to127byte", RMON_R_P65TO127 },
+	{ "rx_128to255byte", RMON_R_P128TO255 },
+	{ "rx_256to511byte", RMON_R_P256TO511 },
+	{ "rx_512to1023byte", RMON_R_P512TO1023 },
+	{ "rx_1024to2047byte", RMON_R_P1024TO2047 },
+	{ "rx_GTE2048byte", RMON_R_P_GTE2048 },
+	{ "rx_octets", RMON_R_OCTETS },
+
+	/* IEEE RX */
+	{ "IEEE_rx_drop", IEEE_R_DROP },
+	{ "IEEE_rx_frame_ok", IEEE_R_FRAME_OK },
+	{ "IEEE_rx_crc", IEEE_R_CRC },
+	{ "IEEE_rx_align", IEEE_R_ALIGN },
+	{ "IEEE_rx_macerr", IEEE_R_MACERR },
+	{ "IEEE_rx_fdxfc", IEEE_R_FDXFC },
+	{ "IEEE_rx_octets_ok", IEEE_R_OCTETS_OK },
+};
+
+static void fec_enet_get_ethtool_stats(struct net_device *dev,
+	struct ethtool_stats *stats, u64 *data)
+{
+	struct fec_enet_private *fep = netdev_priv(dev);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(fec_stats); i++)
+		data[i] = readl(fep->hwp + fec_stats[i].offset);
+}
+
+static void fec_enet_get_strings(struct net_device *netdev,
+	u32 stringset, u8 *data)
+{
+	int i;
+	switch (stringset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < ARRAY_SIZE(fec_stats); i++)
+			memcpy(data + i * ETH_GSTRING_LEN,
+				fec_stats[i].name, ETH_GSTRING_LEN);
+		break;
+	}
+}
+
+static int fec_enet_get_sset_count(struct net_device *dev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return ARRAY_SIZE(fec_stats);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
 #endif /* !defined(CONFIG_M5272) */
 
+static int fec_enet_nway_reset(struct net_device *dev)
+{
+	struct fec_enet_private *fep = netdev_priv(dev);
+	struct phy_device *phydev = fep->phy_dev;
+
+	if (!phydev)
+		return -ENODEV;
+
+	return genphy_restart_aneg(phydev);
+}
+
 static const struct ethtool_ops fec_enet_ethtool_ops = {
 #if !defined(CONFIG_M5272)
 	.get_pauseparam		= fec_enet_get_pauseparam,
@@ -1456,6 +1607,12 @@ static const struct ethtool_ops fec_enet_ethtool_ops = {
 	.get_drvinfo		= fec_enet_get_drvinfo,
 	.get_link		= ethtool_op_get_link,
 	.get_ts_info		= fec_enet_get_ts_info,
+	.nway_reset		= fec_enet_nway_reset,
+#ifndef CONFIG_M5272
+	.get_ethtool_stats	= fec_enet_get_ethtool_stats,
+	.get_strings		= fec_enet_get_strings,
+	.get_sset_count		= fec_enet_get_sset_count,
+#endif
 };
 
 static int fec_enet_ioctl(struct net_device *ndev, struct ifreq *rq, int cmd)
@@ -1803,6 +1960,12 @@ static int fec_enet_init(struct net_device *ndev)
 	writel(FEC_RX_DISABLED_IMASK, fep->hwp + FEC_IMASK);
 	netif_napi_add(ndev, &fep->napi, fec_enet_rx_napi, FEC_NAPI_WEIGHT);
 
+	if (id_entry->driver_data & FEC_QUIRK_HAS_VLAN) {
+		/* enable hw VLAN support */
+		ndev->features |= NETIF_F_HW_VLAN_CTAG_RX;
+		ndev->hw_features |= NETIF_F_HW_VLAN_CTAG_RX;
+	}
+
 	if (id_entry->driver_data & FEC_QUIRK_HAS_CSUM) {
 		/* enable hw accelerator */
 		ndev->features |= (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM
@@ -1865,8 +2028,6 @@ fec_probe(struct platform_device *pdev)
 	struct resource *r;
 	const struct of_device_id *of_id;
 	static int dev_id;
-	struct pinctrl *pinctrl;
-	struct regulator *reg_phy;
 
 	of_id = of_match_device(fec_dt_ids, &pdev->dev);
 	if (of_id)
@@ -1893,17 +2054,17 @@ fec_probe(struct platform_device *pdev)
 		fep->pause_flag |= FEC_PAUSE_FLAG_AUTONEG;
 #endif
 
-	fep->hwp = devm_request_and_ioremap(&pdev->dev, r);
+	fep->hwp = devm_ioremap_resource(&pdev->dev, r);
+	if (IS_ERR(fep->hwp)) {
+		ret = PTR_ERR(fep->hwp);
+		goto failed_ioremap;
+	}
+
 	fep->pdev = pdev;
 	fep->dev_id = dev_id++;
 
 	fep->bufdesc_ex = 0;
 
-	if (!fep->hwp) {
-		ret = -ENOMEM;
-		goto failed_ioremap;
-	}
-
 	platform_set_drvdata(pdev, ndev);
 
 	ret = of_get_phy_mode(pdev->dev.of_node);
@@ -1917,12 +2078,6 @@ fec_probe(struct platform_device *pdev)
 		fep->phy_interface = ret;
 	}
 
-	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
-	if (IS_ERR(pinctrl)) {
-		ret = PTR_ERR(pinctrl);
-		goto failed_pin;
-	}
-
 	fep->clk_ipg = devm_clk_get(&pdev->dev, "ipg");
 	if (IS_ERR(fep->clk_ipg)) {
 		ret = PTR_ERR(fep->clk_ipg);
@@ -1953,20 +2108,22 @@ fec_probe(struct platform_device *pdev)
 	clk_prepare_enable(fep->clk_enet_out);
 	clk_prepare_enable(fep->clk_ptp);
 
-	reg_phy = devm_regulator_get(&pdev->dev, "phy");
-	if (!IS_ERR(reg_phy)) {
-		ret = regulator_enable(reg_phy);
+	fep->reg_phy = devm_regulator_get(&pdev->dev, "phy");
+	if (!IS_ERR(fep->reg_phy)) {
+		ret = regulator_enable(fep->reg_phy);
 		if (ret) {
 			dev_err(&pdev->dev,
 				"Failed to enable phy regulator: %d\n", ret);
 			goto failed_regulator;
 		}
+	} else {
+		fep->reg_phy = NULL;
 	}
 
 	fec_reset_phy(pdev);
 
 	if (fep->bufdesc_ex)
-		fec_ptp_init(ndev, pdev);
+		fec_ptp_init(pdev);
 
 	ret = fec_enet_init(ndev);
 	if (ret)
@@ -2010,19 +2167,20 @@ fec_probe(struct platform_device *pdev)
 failed_register:
 	fec_enet_mii_remove(fep);
 failed_mii_init:
-failed_init:
+failed_irq:
 	for (i = 0; i < FEC_IRQ_NUM; i++) {
 		irq = platform_get_irq(pdev, i);
 		if (irq > 0)
 			free_irq(irq, ndev);
 	}
-failed_irq:
+failed_init:
+	if (fep->reg_phy)
+		regulator_disable(fep->reg_phy);
 failed_regulator:
 	clk_disable_unprepare(fep->clk_ahb);
 	clk_disable_unprepare(fep->clk_ipg);
 	clk_disable_unprepare(fep->clk_enet_out);
 	clk_disable_unprepare(fep->clk_ptp);
-failed_pin:
 failed_clk:
 failed_ioremap:
 	free_netdev(ndev);
@@ -2041,21 +2199,21 @@ fec_drv_remove(struct platform_device *pdev)
 	unregister_netdev(ndev);
 	fec_enet_mii_remove(fep);
 	del_timer_sync(&fep->time_keep);
+	for (i = 0; i < FEC_IRQ_NUM; i++) {
+		int irq = platform_get_irq(pdev, i);
+		if (irq > 0)
+			free_irq(irq, ndev);
+	}
+	if (fep->reg_phy)
+		regulator_disable(fep->reg_phy);
 	clk_disable_unprepare(fep->clk_ptp);
 	if (fep->ptp_clock)
 		ptp_clock_unregister(fep->ptp_clock);
 	clk_disable_unprepare(fep->clk_enet_out);
 	clk_disable_unprepare(fep->clk_ahb);
 	clk_disable_unprepare(fep->clk_ipg);
-	for (i = 0; i < FEC_IRQ_NUM; i++) {
-		int irq = platform_get_irq(pdev, i);
-		if (irq > 0)
-			free_irq(irq, ndev);
-	}
 	free_netdev(ndev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	return 0;
 }
 
@@ -2074,6 +2232,9 @@ fec_suspend(struct device *dev)
 	clk_disable_unprepare(fep->clk_ahb);
 	clk_disable_unprepare(fep->clk_ipg);
 
+	if (fep->reg_phy)
+		regulator_disable(fep->reg_phy);
+
 	return 0;
 }
 
@@ -2082,6 +2243,13 @@ fec_resume(struct device *dev)
 {
 	struct net_device *ndev = dev_get_drvdata(dev);
 	struct fec_enet_private *fep = netdev_priv(ndev);
+	int ret;
+
+	if (fep->reg_phy) {
+		ret = regulator_enable(fep->reg_phy);
+		if (ret)
+			return ret;
+	}
 
 	clk_prepare_enable(fep->clk_enet_out);
 	clk_prepare_enable(fep->clk_ahb);
diff --git a/drivers/net/ethernet/freescale/fec_mpc52xx.c b/drivers/net/ethernet/freescale/fec_mpc52xx.c
index 9bc15e2..9947765 100644
--- a/drivers/net/ethernet/freescale/fec_mpc52xx.c
+++ b/drivers/net/ethernet/freescale/fec_mpc52xx.c
@@ -981,7 +981,7 @@ static int mpc52xx_fec_probe(struct platform_device *op)
 		goto err_node;
 
 	/* We're done ! */
-	dev_set_drvdata(&op->dev, ndev);
+	platform_set_drvdata(op, ndev);
 	netdev_info(ndev, "%s MAC %pM\n",
 		    op->dev.of_node->full_name, ndev->dev_addr);
 
@@ -1010,7 +1010,7 @@ mpc52xx_fec_remove(struct platform_device *op)
 	struct net_device *ndev;
 	struct mpc52xx_fec_priv *priv;
 
-	ndev = dev_get_drvdata(&op->dev);
+	ndev = platform_get_drvdata(op);
 	priv = netdev_priv(ndev);
 
 	unregister_netdev(ndev);
@@ -1030,14 +1030,13 @@ mpc52xx_fec_remove(struct platform_device *op)
 
 	free_netdev(ndev);
 
-	dev_set_drvdata(&op->dev, NULL);
 	return 0;
 }
 
 #ifdef CONFIG_PM
 static int mpc52xx_fec_of_suspend(struct platform_device *op, pm_message_t state)
 {
-	struct net_device *dev = dev_get_drvdata(&op->dev);
+	struct net_device *dev = platform_get_drvdata(op);
 
 	if (netif_running(dev))
 		mpc52xx_fec_close(dev);
@@ -1047,7 +1046,7 @@ static int mpc52xx_fec_of_suspend(struct platform_device *op, pm_message_t state
 
 static int mpc52xx_fec_of_resume(struct platform_device *op)
 {
-	struct net_device *dev = dev_get_drvdata(&op->dev);
+	struct net_device *dev = platform_get_drvdata(op);
 
 	mpc52xx_fec_hw_init(dev);
 	mpc52xx_fec_reset_stats(dev);
diff --git a/drivers/net/ethernet/freescale/fec_ptp.c b/drivers/net/ethernet/freescale/fec_ptp.c
index 25fc960..5007e4f 100644
--- a/drivers/net/ethernet/freescale/fec_ptp.c
+++ b/drivers/net/ethernet/freescale/fec_ptp.c
@@ -347,8 +347,9 @@ static void fec_time_keep(unsigned long _data)
  * cyclecounter init routine and exits.
  */
 
-void fec_ptp_init(struct net_device *ndev, struct platform_device *pdev)
+void fec_ptp_init(struct platform_device *pdev)
 {
+	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct fec_enet_private *fep = netdev_priv(ndev);
 
 	fep->ptp_caps.owner = THIS_MODULE;
diff --git a/drivers/net/ethernet/freescale/fs_enet/Kconfig b/drivers/net/ethernet/freescale/fs_enet/Kconfig
index 268414d..be92229 100644
--- a/drivers/net/ethernet/freescale/fs_enet/Kconfig
+++ b/drivers/net/ethernet/freescale/fs_enet/Kconfig
@@ -1,7 +1,6 @@
 config FS_ENET
        tristate "Freescale Ethernet Driver"
        depends on NET_VENDOR_FREESCALE && (CPM1 || CPM2 || PPC_MPC512x)
-       select NET_CORE
        select MII
        select PHYLIB
 
diff --git a/drivers/net/ethernet/freescale/fs_enet/fs_enet-main.c b/drivers/net/ethernet/freescale/fs_enet/fs_enet-main.c
index edc1200..8de53a1 100644
--- a/drivers/net/ethernet/freescale/fs_enet/fs_enet-main.c
+++ b/drivers/net/ethernet/freescale/fs_enet/fs_enet-main.c
@@ -1048,7 +1048,7 @@ static int fs_enet_probe(struct platform_device *ofdev)
 	}
 
 	SET_NETDEV_DEV(ndev, &ofdev->dev);
-	dev_set_drvdata(&ofdev->dev, ndev);
+	platform_set_drvdata(ofdev, ndev);
 
 	fep = netdev_priv(ndev);
 	fep->dev = &ofdev->dev;
@@ -1106,7 +1106,6 @@ out_cleanup_data:
 	fep->ops->cleanup_data(ndev);
 out_free_dev:
 	free_netdev(ndev);
-	dev_set_drvdata(&ofdev->dev, NULL);
 out_put:
 	of_node_put(fpi->phy_node);
 out_free_fpi:
@@ -1116,7 +1115,7 @@ out_free_fpi:
 
 static int fs_enet_remove(struct platform_device *ofdev)
 {
-	struct net_device *ndev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *ndev = platform_get_drvdata(ofdev);
 	struct fs_enet_private *fep = netdev_priv(ndev);
 
 	unregister_netdev(ndev);
diff --git a/drivers/net/ethernet/freescale/fs_enet/mii-bitbang.c b/drivers/net/ethernet/freescale/fs_enet/mii-bitbang.c
index 2bafbd3..844ecfa 100644
--- a/drivers/net/ethernet/freescale/fs_enet/mii-bitbang.c
+++ b/drivers/net/ethernet/freescale/fs_enet/mii-bitbang.c
@@ -179,7 +179,7 @@ static int fs_enet_mdio_probe(struct platform_device *ofdev)
 	}
 
 	new_bus->parent = &ofdev->dev;
-	dev_set_drvdata(&ofdev->dev, new_bus);
+	platform_set_drvdata(ofdev, new_bus);
 
 	ret = of_mdiobus_register(new_bus, ofdev->dev.of_node);
 	if (ret)
@@ -188,7 +188,6 @@ static int fs_enet_mdio_probe(struct platform_device *ofdev)
 	return 0;
 
 out_free_irqs:
-	dev_set_drvdata(&ofdev->dev, NULL);
 	kfree(new_bus->irq);
 out_unmap_regs:
 	iounmap(bitbang->dir);
@@ -202,11 +201,10 @@ out:
 
 static int fs_enet_mdio_remove(struct platform_device *ofdev)
 {
-	struct mii_bus *bus = dev_get_drvdata(&ofdev->dev);
+	struct mii_bus *bus = platform_get_drvdata(ofdev);
 	struct bb_info *bitbang = bus->priv;
 
 	mdiobus_unregister(bus);
-	dev_set_drvdata(&ofdev->dev, NULL);
 	kfree(bus->irq);
 	free_mdio_bitbang(bus);
 	iounmap(bitbang->dir);
diff --git a/drivers/net/ethernet/freescale/fs_enet/mii-fec.c b/drivers/net/ethernet/freescale/fs_enet/mii-fec.c
index 18e8ef2..2f1c46a 100644
--- a/drivers/net/ethernet/freescale/fs_enet/mii-fec.c
+++ b/drivers/net/ethernet/freescale/fs_enet/mii-fec.c
@@ -180,7 +180,7 @@ static int fs_enet_mdio_probe(struct platform_device *ofdev)
 	}
 
 	new_bus->parent = &ofdev->dev;
-	dev_set_drvdata(&ofdev->dev, new_bus);
+	platform_set_drvdata(ofdev, new_bus);
 
 	ret = of_mdiobus_register(new_bus, ofdev->dev.of_node);
 	if (ret)
@@ -189,7 +189,6 @@ static int fs_enet_mdio_probe(struct platform_device *ofdev)
 	return 0;
 
 out_free_irqs:
-	dev_set_drvdata(&ofdev->dev, NULL);
 	kfree(new_bus->irq);
 out_unmap_regs:
 	iounmap(fec->fecp);
@@ -204,11 +203,10 @@ out:
 
 static int fs_enet_mdio_remove(struct platform_device *ofdev)
 {
-	struct mii_bus *bus = dev_get_drvdata(&ofdev->dev);
+	struct mii_bus *bus = platform_get_drvdata(ofdev);
 	struct fec_info *fec = bus->priv;
 
 	mdiobus_unregister(bus);
-	dev_set_drvdata(&ofdev->dev, NULL);
 	kfree(bus->irq);
 	iounmap(fec->fecp);
 	kfree(fec);
diff --git a/drivers/net/ethernet/freescale/gianfar.c b/drivers/net/ethernet/freescale/gianfar.c
index 2375a01..8d2db7b 100644
--- a/drivers/net/ethernet/freescale/gianfar.c
+++ b/drivers/net/ethernet/freescale/gianfar.c
@@ -128,6 +128,7 @@ static void gfar_set_multi(struct net_device *dev);
 static void gfar_set_hash_for_addr(struct net_device *dev, u8 *addr);
 static void gfar_configure_serdes(struct net_device *dev);
 static int gfar_poll(struct napi_struct *napi, int budget);
+static int gfar_poll_sq(struct napi_struct *napi, int budget);
 #ifdef CONFIG_NET_POLL_CONTROLLER
 static void gfar_netpoll(struct net_device *dev);
 #endif
@@ -1000,7 +1001,7 @@ static int gfar_probe(struct platform_device *ofdev)
 	spin_lock_init(&priv->bflock);
 	INIT_WORK(&priv->reset_task, gfar_reset_task);
 
-	dev_set_drvdata(&ofdev->dev, priv);
+	platform_set_drvdata(ofdev, priv);
 	regs = priv->gfargrp[0].regs;
 
 	gfar_detect_errata(priv);
@@ -1038,9 +1039,13 @@ static int gfar_probe(struct platform_device *ofdev)
 	dev->ethtool_ops = &gfar_ethtool_ops;
 
 	/* Register for napi ...We are registering NAPI for each grp */
-	for (i = 0; i < priv->num_grps; i++)
-		netif_napi_add(dev, &priv->gfargrp[i].napi, gfar_poll,
+	if (priv->mode == SQ_SG_MODE)
+		netif_napi_add(dev, &priv->gfargrp[0].napi, gfar_poll_sq,
 			       GFAR_DEV_WEIGHT);
+	else
+		for (i = 0; i < priv->num_grps; i++)
+			netif_napi_add(dev, &priv->gfargrp[i].napi, gfar_poll,
+				       GFAR_DEV_WEIGHT);
 
 	if (priv->device_flags & FSL_GIANFAR_DEV_HAS_CSUM) {
 		dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_SG |
@@ -1240,15 +1245,13 @@ register_fail:
 
 static int gfar_remove(struct platform_device *ofdev)
 {
-	struct gfar_private *priv = dev_get_drvdata(&ofdev->dev);
+	struct gfar_private *priv = platform_get_drvdata(ofdev);
 
 	if (priv->phy_node)
 		of_node_put(priv->phy_node);
 	if (priv->tbi_node)
 		of_node_put(priv->tbi_node);
 
-	dev_set_drvdata(&ofdev->dev, NULL);
-
 	unregister_netdev(priv->ndev);
 	unmap_group_regs(priv);
 	free_gfar_dev(priv);
@@ -2825,6 +2828,48 @@ int gfar_clean_rx_ring(struct gfar_priv_rx_q *rx_queue, int rx_work_limit)
 	return howmany;
 }
 
+static int gfar_poll_sq(struct napi_struct *napi, int budget)
+{
+	struct gfar_priv_grp *gfargrp =
+		container_of(napi, struct gfar_priv_grp, napi);
+	struct gfar __iomem *regs = gfargrp->regs;
+	struct gfar_priv_tx_q *tx_queue = gfargrp->priv->tx_queue[0];
+	struct gfar_priv_rx_q *rx_queue = gfargrp->priv->rx_queue[0];
+	int work_done = 0;
+
+	/* Clear IEVENT, so interrupts aren't called again
+	 * because of the packets that have already arrived
+	 */
+	gfar_write(&regs->ievent, IEVENT_RTX_MASK);
+
+	/* run Tx cleanup to completion */
+	if (tx_queue->tx_skbuff[tx_queue->skb_dirtytx])
+		gfar_clean_tx_ring(tx_queue);
+
+	work_done = gfar_clean_rx_ring(rx_queue, budget);
+
+	if (work_done < budget) {
+		napi_complete(napi);
+		/* Clear the halt bit in RSTAT */
+		gfar_write(&regs->rstat, gfargrp->rstat);
+
+		gfar_write(&regs->imask, IMASK_DEFAULT);
+
+		/* If we are coalescing interrupts, update the timer
+		 * Otherwise, clear it
+		 */
+		gfar_write(&regs->txic, 0);
+		if (likely(tx_queue->txcoalescing))
+			gfar_write(&regs->txic, tx_queue->txic);
+
+		gfar_write(&regs->rxic, 0);
+		if (unlikely(rx_queue->rxcoalescing))
+			gfar_write(&regs->rxic, rx_queue->rxic);
+	}
+
+	return work_done;
+}
+
 static int gfar_poll(struct napi_struct *napi, int budget)
 {
 	struct gfar_priv_grp *gfargrp =
diff --git a/drivers/net/ethernet/freescale/gianfar_ptp.c b/drivers/net/ethernet/freescale/gianfar_ptp.c
index 083ea2b..098f133 100644
--- a/drivers/net/ethernet/freescale/gianfar_ptp.c
+++ b/drivers/net/ethernet/freescale/gianfar_ptp.c
@@ -519,7 +519,7 @@ static int gianfar_ptp_probe(struct platform_device *dev)
 	}
 	gfar_phc_index = ptp_clock_index(etsects->clock);
 
-	dev_set_drvdata(&dev->dev, etsects);
+	platform_set_drvdata(dev, etsects);
 
 	return 0;
 
@@ -537,7 +537,7 @@ no_memory:
 
 static int gianfar_ptp_remove(struct platform_device *dev)
 {
-	struct etsects *etsects = dev_get_drvdata(&dev->dev);
+	struct etsects *etsects = platform_get_drvdata(dev);
 
 	gfar_write(&etsects->regs->tmr_temask, 0);
 	gfar_write(&etsects->regs->tmr_ctrl,   0);
diff --git a/drivers/net/ethernet/freescale/ucc_geth.c b/drivers/net/ethernet/freescale/ucc_geth.c
index e04c598..3c43dac 100644
--- a/drivers/net/ethernet/freescale/ucc_geth.c
+++ b/drivers/net/ethernet/freescale/ucc_geth.c
@@ -3564,7 +3564,7 @@ static void ucc_geth_timeout(struct net_device *dev)
 
 static int ucc_geth_suspend(struct platform_device *ofdev, pm_message_t state)
 {
-	struct net_device *ndev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *ndev = platform_get_drvdata(ofdev);
 	struct ucc_geth_private *ugeth = netdev_priv(ndev);
 
 	if (!netif_running(ndev))
@@ -3592,7 +3592,7 @@ static int ucc_geth_suspend(struct platform_device *ofdev, pm_message_t state)
 
 static int ucc_geth_resume(struct platform_device *ofdev)
 {
-	struct net_device *ndev = dev_get_drvdata(&ofdev->dev);
+	struct net_device *ndev = platform_get_drvdata(ofdev);
 	struct ucc_geth_private *ugeth = netdev_priv(ndev);
 	int err;
 
diff --git a/drivers/net/ethernet/freescale/xgmac_mdio.c b/drivers/net/ethernet/freescale/xgmac_mdio.c
index 418068b..c1b6e7e 100644
--- a/drivers/net/ethernet/freescale/xgmac_mdio.c
+++ b/drivers/net/ethernet/freescale/xgmac_mdio.c
@@ -227,7 +227,7 @@ static int xgmac_mdio_probe(struct platform_device *pdev)
 		goto err_registration;
 	}
 
-	dev_set_drvdata(&pdev->dev, bus);
+	platform_set_drvdata(pdev, bus);
 
 	return 0;
 
@@ -242,7 +242,7 @@ err_ioremap:
 
 static int xgmac_mdio_remove(struct platform_device *pdev)
 {
-	struct mii_bus *bus = dev_get_drvdata(&pdev->dev);
+	struct mii_bus *bus = platform_get_drvdata(pdev);
 
 	mdiobus_unregister(bus);
 	iounmap(bus->priv);
diff --git a/drivers/net/ethernet/ibm/Kconfig b/drivers/net/ethernet/ibm/Kconfig
index 6529d31..563a1ac 100644
--- a/drivers/net/ethernet/ibm/Kconfig
+++ b/drivers/net/ethernet/ibm/Kconfig
@@ -5,8 +5,7 @@
 config NET_VENDOR_IBM
 	bool "IBM devices"
 	default y
-	depends on MCA || PPC_PSERIES || PPC_PSERIES || PPC_DCR || \
-		   (IBMEBUS && SPARSEMEM)
+	depends on PPC_PSERIES || PPC_DCR || (IBMEBUS && SPARSEMEM)
 	---help---
 	  If you have a network (Ethernet) card belonging to this class, say Y
 	  and read the Ethernet-HOWTO, available from
diff --git a/drivers/net/ethernet/ibm/ehea/ehea_main.c b/drivers/net/ethernet/ibm/ehea/ehea_main.c
index de2969c..35853b4 100644
--- a/drivers/net/ethernet/ibm/ehea/ehea_main.c
+++ b/drivers/net/ethernet/ibm/ehea/ehea_main.c
@@ -3287,7 +3287,7 @@ static int ehea_probe_adapter(struct platform_device *dev)
 
 	adapter->pd = EHEA_PD_ID;
 
-	dev_set_drvdata(&dev->dev, adapter);
+	platform_set_drvdata(dev, adapter);
 
 
 	/* initialize adapter and ports */
@@ -3358,7 +3358,7 @@ out:
 
 static int ehea_remove(struct platform_device *dev)
 {
-	struct ehea_adapter *adapter = dev_get_drvdata(&dev->dev);
+	struct ehea_adapter *adapter = platform_get_drvdata(dev);
 	int i;
 
 	for (i = 0; i < EHEA_MAX_PORTS; i++)
diff --git a/drivers/net/ethernet/ibm/emac/mal.c b/drivers/net/ethernet/ibm/emac/mal.c
index 610ed223..856ea66 100644
--- a/drivers/net/ethernet/ibm/emac/mal.c
+++ b/drivers/net/ethernet/ibm/emac/mal.c
@@ -696,7 +696,7 @@ static int mal_probe(struct platform_device *ofdev)
 
 	/* Advertise this instance to the rest of the world */
 	wmb();
-	dev_set_drvdata(&ofdev->dev, mal);
+	platform_set_drvdata(ofdev, mal);
 
 	mal_dbg_register(mal);
 
@@ -722,7 +722,7 @@ static int mal_probe(struct platform_device *ofdev)
 
 static int mal_remove(struct platform_device *ofdev)
 {
-	struct mal_instance *mal = dev_get_drvdata(&ofdev->dev);
+	struct mal_instance *mal = platform_get_drvdata(ofdev);
 
 	MAL_DBG(mal, "remove" NL);
 
@@ -735,8 +735,6 @@ static int mal_remove(struct platform_device *ofdev)
 		       "mal%d: commac list is not empty on remove!\n",
 		       mal->index);
 
-	dev_set_drvdata(&ofdev->dev, NULL);
-
 	free_irq(mal->serr_irq, mal);
 	free_irq(mal->txde_irq, mal);
 	free_irq(mal->txeob_irq, mal);
diff --git a/drivers/net/ethernet/ibm/emac/rgmii.c b/drivers/net/ethernet/ibm/emac/rgmii.c
index 3925176..c47e23d 100644
--- a/drivers/net/ethernet/ibm/emac/rgmii.c
+++ b/drivers/net/ethernet/ibm/emac/rgmii.c
@@ -95,7 +95,7 @@ static inline u32 rgmii_mode_mask(int mode, int input)
 
 int rgmii_attach(struct platform_device *ofdev, int input, int mode)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 	struct rgmii_regs __iomem *p = dev->base;
 
 	RGMII_DBG(dev, "attach(%d)" NL, input);
@@ -124,7 +124,7 @@ int rgmii_attach(struct platform_device *ofdev, int input, int mode)
 
 void rgmii_set_speed(struct platform_device *ofdev, int input, int speed)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 	struct rgmii_regs __iomem *p = dev->base;
 	u32 ssr;
 
@@ -146,7 +146,7 @@ void rgmii_set_speed(struct platform_device *ofdev, int input, int speed)
 
 void rgmii_get_mdio(struct platform_device *ofdev, int input)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 	struct rgmii_regs __iomem *p = dev->base;
 	u32 fer;
 
@@ -167,7 +167,7 @@ void rgmii_get_mdio(struct platform_device *ofdev, int input)
 
 void rgmii_put_mdio(struct platform_device *ofdev, int input)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 	struct rgmii_regs __iomem *p = dev->base;
 	u32 fer;
 
@@ -188,7 +188,7 @@ void rgmii_put_mdio(struct platform_device *ofdev, int input)
 
 void rgmii_detach(struct platform_device *ofdev, int input)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 	struct rgmii_regs __iomem *p;
 
 	BUG_ON(!dev || dev->users == 0);
@@ -214,7 +214,7 @@ int rgmii_get_regs_len(struct platform_device *ofdev)
 
 void *rgmii_dump_regs(struct platform_device *ofdev, void *buf)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 	struct emac_ethtool_regs_subhdr *hdr = buf;
 	struct rgmii_regs *regs = (struct rgmii_regs *)(hdr + 1);
 
@@ -279,7 +279,7 @@ static int rgmii_probe(struct platform_device *ofdev)
 	       (dev->flags & EMAC_RGMII_FLAG_HAS_MDIO) ? "" : "out");
 
 	wmb();
-	dev_set_drvdata(&ofdev->dev, dev);
+	platform_set_drvdata(ofdev, dev);
 
 	return 0;
 
@@ -291,9 +291,7 @@ static int rgmii_probe(struct platform_device *ofdev)
 
 static int rgmii_remove(struct platform_device *ofdev)
 {
-	struct rgmii_instance *dev = dev_get_drvdata(&ofdev->dev);
-
-	dev_set_drvdata(&ofdev->dev, NULL);
+	struct rgmii_instance *dev = platform_get_drvdata(ofdev);
 
 	WARN_ON(dev->users != 0);
 
diff --git a/drivers/net/ethernet/ibm/emac/tah.c b/drivers/net/ethernet/ibm/emac/tah.c
index 795f139..c231a4a 100644
--- a/drivers/net/ethernet/ibm/emac/tah.c
+++ b/drivers/net/ethernet/ibm/emac/tah.c
@@ -25,7 +25,7 @@
 
 int tah_attach(struct platform_device *ofdev, int channel)
 {
-	struct tah_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct tah_instance *dev = platform_get_drvdata(ofdev);
 
 	mutex_lock(&dev->lock);
 	/* Reset has been done at probe() time... nothing else to do for now */
@@ -37,7 +37,7 @@ int tah_attach(struct platform_device *ofdev, int channel)
 
 void tah_detach(struct platform_device *ofdev, int channel)
 {
-	struct tah_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct tah_instance *dev = platform_get_drvdata(ofdev);
 
 	mutex_lock(&dev->lock);
 	--dev->users;
@@ -46,7 +46,7 @@ void tah_detach(struct platform_device *ofdev, int channel)
 
 void tah_reset(struct platform_device *ofdev)
 {
-	struct tah_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct tah_instance *dev = platform_get_drvdata(ofdev);
 	struct tah_regs __iomem *p = dev->base;
 	int n;
 
@@ -74,7 +74,7 @@ int tah_get_regs_len(struct platform_device *ofdev)
 
 void *tah_dump_regs(struct platform_device *ofdev, void *buf)
 {
-	struct tah_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct tah_instance *dev = platform_get_drvdata(ofdev);
 	struct emac_ethtool_regs_subhdr *hdr = buf;
 	struct tah_regs *regs = (struct tah_regs *)(hdr + 1);
 
@@ -118,7 +118,7 @@ static int tah_probe(struct platform_device *ofdev)
 		goto err_free;
 	}
 
-	dev_set_drvdata(&ofdev->dev, dev);
+	platform_set_drvdata(ofdev, dev);
 
 	/* Initialize TAH and enable IPv4 checksum verification, no TSO yet */
 	tah_reset(ofdev);
@@ -137,9 +137,7 @@ static int tah_probe(struct platform_device *ofdev)
 
 static int tah_remove(struct platform_device *ofdev)
 {
-	struct tah_instance *dev = dev_get_drvdata(&ofdev->dev);
-
-	dev_set_drvdata(&ofdev->dev, NULL);
+	struct tah_instance *dev = platform_get_drvdata(ofdev);
 
 	WARN_ON(dev->users != 0);
 
diff --git a/drivers/net/ethernet/ibm/emac/zmii.c b/drivers/net/ethernet/ibm/emac/zmii.c
index f91202f..4cdf286 100644
--- a/drivers/net/ethernet/ibm/emac/zmii.c
+++ b/drivers/net/ethernet/ibm/emac/zmii.c
@@ -84,7 +84,7 @@ static inline u32 zmii_mode_mask(int mode, int input)
 
 int zmii_attach(struct platform_device *ofdev, int input, int *mode)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 	struct zmii_regs __iomem *p = dev->base;
 
 	ZMII_DBG(dev, "init(%d, %d)" NL, input, *mode);
@@ -150,7 +150,7 @@ int zmii_attach(struct platform_device *ofdev, int input, int *mode)
 
 void zmii_get_mdio(struct platform_device *ofdev, int input)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 	u32 fer;
 
 	ZMII_DBG2(dev, "get_mdio(%d)" NL, input);
@@ -163,7 +163,7 @@ void zmii_get_mdio(struct platform_device *ofdev, int input)
 
 void zmii_put_mdio(struct platform_device *ofdev, int input)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 
 	ZMII_DBG2(dev, "put_mdio(%d)" NL, input);
 	mutex_unlock(&dev->lock);
@@ -172,7 +172,7 @@ void zmii_put_mdio(struct platform_device *ofdev, int input)
 
 void zmii_set_speed(struct platform_device *ofdev, int input, int speed)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 	u32 ssr;
 
 	mutex_lock(&dev->lock);
@@ -193,7 +193,7 @@ void zmii_set_speed(struct platform_device *ofdev, int input, int speed)
 
 void zmii_detach(struct platform_device *ofdev, int input)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 
 	BUG_ON(!dev || dev->users == 0);
 
@@ -218,7 +218,7 @@ int zmii_get_regs_len(struct platform_device *ofdev)
 
 void *zmii_dump_regs(struct platform_device *ofdev, void *buf)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 	struct emac_ethtool_regs_subhdr *hdr = buf;
 	struct zmii_regs *regs = (struct zmii_regs *)(hdr + 1);
 
@@ -272,7 +272,7 @@ static int zmii_probe(struct platform_device *ofdev)
 	printk(KERN_INFO
 	       "ZMII %s initialized\n", ofdev->dev.of_node->full_name);
 	wmb();
-	dev_set_drvdata(&ofdev->dev, dev);
+	platform_set_drvdata(ofdev, dev);
 
 	return 0;
 
@@ -284,9 +284,7 @@ static int zmii_probe(struct platform_device *ofdev)
 
 static int zmii_remove(struct platform_device *ofdev)
 {
-	struct zmii_instance *dev = dev_get_drvdata(&ofdev->dev);
-
-	dev_set_drvdata(&ofdev->dev, NULL);
+	struct zmii_instance *dev = platform_get_drvdata(ofdev);
 
 	WARN_ON(dev->users != 0);
 
diff --git a/drivers/net/ethernet/icplus/Kconfig b/drivers/net/ethernet/icplus/Kconfig
index 5119ef1..14a66e9 100644
--- a/drivers/net/ethernet/icplus/Kconfig
+++ b/drivers/net/ethernet/icplus/Kconfig
@@ -5,7 +5,6 @@
 config IP1000
 	tristate "IP1000 Gigabit Ethernet support"
 	depends on PCI
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports IP1000 gigabit Ethernet cards.
diff --git a/drivers/net/ethernet/icplus/ipg.c b/drivers/net/ethernet/icplus/ipg.c
index 068d781..1fde90b 100644
--- a/drivers/net/ethernet/icplus/ipg.c
+++ b/drivers/net/ethernet/icplus/ipg.c
@@ -2298,15 +2298,4 @@ static struct pci_driver ipg_pci_driver = {
 	.remove		= ipg_remove,
 };
 
-static int __init ipg_init_module(void)
-{
-	return pci_register_driver(&ipg_pci_driver);
-}
-
-static void __exit ipg_exit_module(void)
-{
-	pci_unregister_driver(&ipg_pci_driver);
-}
-
-module_init(ipg_init_module);
-module_exit(ipg_exit_module);
+module_pci_driver(ipg_pci_driver);
diff --git a/drivers/net/ethernet/intel/Kconfig b/drivers/net/ethernet/intel/Kconfig
index 05f7264..f0e7ed2 100644
--- a/drivers/net/ethernet/intel/Kconfig
+++ b/drivers/net/ethernet/intel/Kconfig
@@ -20,7 +20,6 @@ if NET_VENDOR_INTEL
 config E100
 	tristate "Intel(R) PRO/100+ support"
 	depends on PCI
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports Intel(R) PRO/100 family of adapters.
diff --git a/drivers/net/ethernet/intel/e100.c b/drivers/net/ethernet/intel/e100.c
index d2bea3f..5115ae7 100644
--- a/drivers/net/ethernet/intel/e100.c
+++ b/drivers/net/ethernet/intel/e100.c
@@ -3069,7 +3069,7 @@ static int e100_resume(struct pci_dev *pdev)
 	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
 	/* ack any pending wake events, disable PME */
-	pci_enable_wake(pdev, 0, 0);
+	pci_enable_wake(pdev, PCI_D0, 0);
 
 	/* disable reverse auto-negotiation */
 	if (nic->phy == phy_82552_v) {
@@ -3160,7 +3160,7 @@ static void e100_io_resume(struct pci_dev *pdev)
 	struct nic *nic = netdev_priv(netdev);
 
 	/* ack any pending wake events, disable PME */
-	pci_enable_wake(pdev, 0, 0);
+	pci_enable_wake(pdev, PCI_D0, 0);
 
 	netif_device_attach(netdev);
 	if (netif_running(netdev)) {
diff --git a/drivers/net/ethernet/intel/e1000e/80003es2lan.c b/drivers/net/ethernet/intel/e1000e/80003es2lan.c
index b71c850..895450e 100644
--- a/drivers/net/ethernet/intel/e1000e/80003es2lan.c
+++ b/drivers/net/ethernet/intel/e1000e/80003es2lan.c
@@ -66,17 +66,17 @@ static s32 e1000_init_phy_params_80003es2lan(struct e1000_hw *hw)
 	s32 ret_val;
 
 	if (hw->phy.media_type != e1000_media_type_copper) {
-		phy->type	= e1000_phy_none;
+		phy->type = e1000_phy_none;
 		return 0;
 	} else {
 		phy->ops.power_up = e1000_power_up_phy_copper;
 		phy->ops.power_down = e1000_power_down_phy_copper_80003es2lan;
 	}
 
-	phy->addr		= 1;
-	phy->autoneg_mask	= AUTONEG_ADVERTISE_SPEED_DEFAULT;
-	phy->reset_delay_us      = 100;
-	phy->type		= e1000_phy_gg82563;
+	phy->addr = 1;
+	phy->autoneg_mask = AUTONEG_ADVERTISE_SPEED_DEFAULT;
+	phy->reset_delay_us = 100;
+	phy->type = e1000_phy_gg82563;
 
 	/* This can only be done after all function pointers are setup. */
 	ret_val = e1000e_get_phy_id(hw);
@@ -98,19 +98,19 @@ static s32 e1000_init_nvm_params_80003es2lan(struct e1000_hw *hw)
 	u32 eecd = er32(EECD);
 	u16 size;
 
-	nvm->opcode_bits	= 8;
-	nvm->delay_usec	 = 1;
+	nvm->opcode_bits = 8;
+	nvm->delay_usec = 1;
 	switch (nvm->override) {
 	case e1000_nvm_override_spi_large:
-		nvm->page_size    = 32;
+		nvm->page_size = 32;
 		nvm->address_bits = 16;
 		break;
 	case e1000_nvm_override_spi_small:
-		nvm->page_size    = 8;
+		nvm->page_size = 8;
 		nvm->address_bits = 8;
 		break;
 	default:
-		nvm->page_size    = eecd & E1000_EECD_ADDR_BITS ? 32 : 8;
+		nvm->page_size = eecd & E1000_EECD_ADDR_BITS ? 32 : 8;
 		nvm->address_bits = eecd & E1000_EECD_ADDR_BITS ? 16 : 8;
 		break;
 	}
@@ -128,7 +128,7 @@ static s32 e1000_init_nvm_params_80003es2lan(struct e1000_hw *hw)
 	/* EEPROM access above 16k is unsupported */
 	if (size > 14)
 		size = 14;
-	nvm->word_size	= 1 << size;
+	nvm->word_size = 1 << size;
 
 	return 0;
 }
@@ -859,7 +859,7 @@ static void e1000_initialize_hw_bits_80003es2lan(struct e1000_hw *hw)
 
 	/* Transmit Arbitration Control 0 */
 	reg = er32(TARC(0));
-	reg &= ~(0xF << 27); /* 30:27 */
+	reg &= ~(0xF << 27);	/* 30:27 */
 	if (hw->phy.media_type != e1000_media_type_copper)
 		reg &= ~(1 << 20);
 	ew32(TARC(0), reg);
diff --git a/drivers/net/ethernet/intel/e1000e/82571.c b/drivers/net/ethernet/intel/e1000e/82571.c
index 7380442..4c303e2 100644
--- a/drivers/net/ethernet/intel/e1000e/82571.c
+++ b/drivers/net/ethernet/intel/e1000e/82571.c
@@ -77,24 +77,24 @@ static s32 e1000_init_phy_params_82571(struct e1000_hw *hw)
 		return 0;
 	}
 
-	phy->addr			 = 1;
-	phy->autoneg_mask		 = AUTONEG_ADVERTISE_SPEED_DEFAULT;
-	phy->reset_delay_us		 = 100;
+	phy->addr = 1;
+	phy->autoneg_mask = AUTONEG_ADVERTISE_SPEED_DEFAULT;
+	phy->reset_delay_us = 100;
 
-	phy->ops.power_up		 = e1000_power_up_phy_copper;
-	phy->ops.power_down		 = e1000_power_down_phy_copper_82571;
+	phy->ops.power_up = e1000_power_up_phy_copper;
+	phy->ops.power_down = e1000_power_down_phy_copper_82571;
 
 	switch (hw->mac.type) {
 	case e1000_82571:
 	case e1000_82572:
-		phy->type		 = e1000_phy_igp_2;
+		phy->type = e1000_phy_igp_2;
 		break;
 	case e1000_82573:
-		phy->type		 = e1000_phy_m88;
+		phy->type = e1000_phy_m88;
 		break;
 	case e1000_82574:
 	case e1000_82583:
-		phy->type		 = e1000_phy_bm;
+		phy->type = e1000_phy_bm;
 		phy->ops.acquire = e1000_get_hw_semaphore_82574;
 		phy->ops.release = e1000_put_hw_semaphore_82574;
 		phy->ops.set_d0_lplu_state = e1000_set_d0_lplu_state_82574;
@@ -193,7 +193,7 @@ static s32 e1000_init_nvm_params_82571(struct e1000_hw *hw)
 		/* EEPROM access above 16k is unsupported */
 		if (size > 14)
 			size = 14;
-		nvm->word_size	= 1 << size;
+		nvm->word_size = 1 << size;
 		break;
 	}
 
@@ -339,7 +339,7 @@ static s32 e1000_init_mac_params_82571(struct e1000_hw *hw)
 static s32 e1000_get_variants_82571(struct e1000_adapter *adapter)
 {
 	struct e1000_hw *hw = &adapter->hw;
-	static int global_quad_port_a; /* global port a indication */
+	static int global_quad_port_a;	/* global port a indication */
 	struct pci_dev *pdev = adapter->pdev;
 	int is_port_b = er32(STATUS) & E1000_STATUS_FUNC_1;
 	s32 rc;
@@ -1003,8 +1003,6 @@ static s32 e1000_reset_hw_82571(struct e1000_hw *hw)
 	default:
 		break;
 	}
-	if (ret_val)
-		e_dbg("Cannot acquire MDIO ownership\n");
 
 	ctrl = er32(CTRL);
 
@@ -1015,7 +1013,9 @@ static s32 e1000_reset_hw_82571(struct e1000_hw *hw)
 	switch (hw->mac.type) {
 	case e1000_82574:
 	case e1000_82583:
-		e1000_put_hw_semaphore_82574(hw);
+		/* Release mutex only if the hw semaphore is acquired */
+		if (!ret_val)
+			e1000_put_hw_semaphore_82574(hw);
 		break;
 	default:
 		break;
@@ -1178,7 +1178,7 @@ static void e1000_initialize_hw_bits_82571(struct e1000_hw *hw)
 
 	/* Transmit Arbitration Control 0 */
 	reg = er32(TARC(0));
-	reg &= ~(0xF << 27); /* 30:27 */
+	reg &= ~(0xF << 27);	/* 30:27 */
 	switch (hw->mac.type) {
 	case e1000_82571:
 	case e1000_82572:
@@ -1390,7 +1390,7 @@ bool e1000_check_phy_82574(struct e1000_hw *hw)
 	ret_val = e1e_rphy(hw, E1000_RECEIVE_ERROR_COUNTER, &receive_errors);
 	if (ret_val)
 		return false;
-	if (receive_errors == E1000_RECEIVE_ERROR_MAX)  {
+	if (receive_errors == E1000_RECEIVE_ERROR_MAX) {
 		ret_val = e1e_rphy(hw, E1000_BASE1000T_STATUS, &status_1kbt);
 		if (ret_val)
 			return false;
diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/net/ethernet/intel/e1000e/ethtool.c
index 7c8ca65..59c22bf 100644
--- a/drivers/net/ethernet/intel/e1000e/ethtool.c
+++ b/drivers/net/ethernet/intel/e1000e/ethtool.c
@@ -244,7 +244,7 @@ static int e1000_set_spd_dplx(struct e1000_adapter *adapter, u32 spd, u8 dplx)
 		mac->autoneg = 1;
 		adapter->hw.phy.autoneg_advertised = ADVERTISE_1000_FULL;
 		break;
-	case SPEED_1000 + DUPLEX_HALF: /* not supported */
+	case SPEED_1000 + DUPLEX_HALF:	/* not supported */
 	default:
 		goto err_inval;
 	}
@@ -416,7 +416,7 @@ static void e1000_set_msglevel(struct net_device *netdev, u32 data)
 
 static int e1000_get_regs_len(struct net_device __always_unused *netdev)
 {
-#define E1000_REGS_LEN 32 /* overestimate */
+#define E1000_REGS_LEN 32	/* overestimate */
 	return E1000_REGS_LEN * sizeof(u32);
 }
 
@@ -433,22 +433,22 @@ static void e1000_get_regs(struct net_device *netdev,
 	regs->version = (1 << 24) | (adapter->pdev->revision << 16) |
 	    adapter->pdev->device;
 
-	regs_buff[0]  = er32(CTRL);
-	regs_buff[1]  = er32(STATUS);
+	regs_buff[0] = er32(CTRL);
+	regs_buff[1] = er32(STATUS);
 
-	regs_buff[2]  = er32(RCTL);
-	regs_buff[3]  = er32(RDLEN(0));
-	regs_buff[4]  = er32(RDH(0));
-	regs_buff[5]  = er32(RDT(0));
-	regs_buff[6]  = er32(RDTR);
+	regs_buff[2] = er32(RCTL);
+	regs_buff[3] = er32(RDLEN(0));
+	regs_buff[4] = er32(RDH(0));
+	regs_buff[5] = er32(RDT(0));
+	regs_buff[6] = er32(RDTR);
 
-	regs_buff[7]  = er32(TCTL);
-	regs_buff[8]  = er32(TDLEN(0));
-	regs_buff[9]  = er32(TDH(0));
+	regs_buff[7] = er32(TCTL);
+	regs_buff[8] = er32(TDLEN(0));
+	regs_buff[9] = er32(TDH(0));
 	regs_buff[10] = er32(TDT(0));
 	regs_buff[11] = er32(TIDV);
 
-	regs_buff[12] = adapter->hw.phy.type;  /* PHY type (IGP=1, M88=0) */
+	regs_buff[12] = adapter->hw.phy.type;	/* PHY type (IGP=1, M88=0) */
 
 	/* ethtool doesn't use anything past this point, so all this
 	 * code is likely legacy junk for apps that may or may not exist
@@ -1379,7 +1379,7 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter)
 
 	if (hw->phy.media_type == e1000_media_type_copper &&
 	    hw->phy.type == e1000_phy_m88) {
-		ctrl_reg |= E1000_CTRL_ILOS; /* Invert Loss of Signal */
+		ctrl_reg |= E1000_CTRL_ILOS;	/* Invert Loss of Signal */
 	} else {
 		/* Set the ILOS bit on the fiber Nic if half duplex link is
 		 * detected.
@@ -1613,7 +1613,7 @@ static int e1000_run_loopback_test(struct e1000_adapter *adapter)
 		ew32(TDT(0), k);
 		e1e_flush();
 		msleep(200);
-		time = jiffies; /* set the start time for the receive */
+		time = jiffies;	/* set the start time for the receive */
 		good_cnt = 0;
 		/* receive the sent packets */
 		do {
@@ -1636,11 +1636,11 @@ static int e1000_run_loopback_test(struct e1000_adapter *adapter)
 			 */
 		} while ((good_cnt < 64) && !time_after(jiffies, time + 20));
 		if (good_cnt != 64) {
-			ret_val = 13; /* ret_val is the same as mis-compare */
+			ret_val = 13;	/* ret_val is the same as mis-compare */
 			break;
 		}
 		if (jiffies >= (time + 20)) {
-			ret_val = 14; /* error code for time out error */
+			ret_val = 14;	/* error code for time out error */
 			break;
 		}
 	}
diff --git a/drivers/net/ethernet/intel/e1000e/hw.h b/drivers/net/ethernet/intel/e1000e/hw.h
index 84850f7..a6f903a 100644
--- a/drivers/net/ethernet/intel/e1000e/hw.h
+++ b/drivers/net/ethernet/intel/e1000e/hw.h
@@ -402,13 +402,13 @@ struct e1000_phy_stats {
 
 struct e1000_host_mng_dhcp_cookie {
 	u32 signature;
-	u8  status;
-	u8  reserved0;
+	u8 status;
+	u8 reserved0;
 	u16 vlan_id;
 	u32 reserved1;
 	u16 reserved2;
-	u8  reserved3;
-	u8  checksum;
+	u8 reserved3;
+	u8 checksum;
 };
 
 /* Host Interface "Rev 1" */
@@ -427,8 +427,8 @@ struct e1000_host_command_info {
 
 /* Host Interface "Rev 2" */
 struct e1000_host_mng_command_header {
-	u8  command_id;
-	u8  checksum;
+	u8 command_id;
+	u8 checksum;
 	u16 reserved1;
 	u16 reserved2;
 	u16 command_length;
@@ -549,7 +549,7 @@ struct e1000_mac_info {
 	u32 mta_shadow[MAX_MTA_REG];
 	u16 rar_entry_count;
 
-	u8  forced_speed_duplex;
+	u8 forced_speed_duplex;
 
 	bool adaptive_ifs;
 	bool has_fwsm;
@@ -577,7 +577,7 @@ struct e1000_phy_info {
 
 	u32 addr;
 	u32 id;
-	u32 reset_delay_us; /* in usec */
+	u32 reset_delay_us;	/* in usec */
 	u32 revision;
 
 	enum e1000_media_type media_type;
@@ -636,11 +636,11 @@ struct e1000_dev_spec_82571 {
 };
 
 struct e1000_dev_spec_80003es2lan {
-	bool  mdic_wa_enable;
+	bool mdic_wa_enable;
 };
 
 struct e1000_shadow_ram {
-	u16  value;
+	u16 value;
 	bool modified;
 };
 
@@ -660,17 +660,17 @@ struct e1000_hw {
 	void __iomem *hw_addr;
 	void __iomem *flash_address;
 
-	struct e1000_mac_info  mac;
-	struct e1000_fc_info   fc;
-	struct e1000_phy_info  phy;
-	struct e1000_nvm_info  nvm;
-	struct e1000_bus_info  bus;
+	struct e1000_mac_info mac;
+	struct e1000_fc_info fc;
+	struct e1000_phy_info phy;
+	struct e1000_nvm_info nvm;
+	struct e1000_bus_info bus;
 	struct e1000_host_mng_dhcp_cookie mng_cookie;
 
 	union {
-		struct e1000_dev_spec_82571	e82571;
+		struct e1000_dev_spec_82571 e82571;
 		struct e1000_dev_spec_80003es2lan e80003es2lan;
-		struct e1000_dev_spec_ich8lan	ich8lan;
+		struct e1000_dev_spec_ich8lan ich8lan;
 	} dev_spec;
 };
 
diff --git a/drivers/net/ethernet/intel/e1000e/ich8lan.c b/drivers/net/ethernet/intel/e1000e/ich8lan.c
index ad9d8f2..9dde390 100644
--- a/drivers/net/ethernet/intel/e1000e/ich8lan.c
+++ b/drivers/net/ethernet/intel/e1000e/ich8lan.c
@@ -101,12 +101,12 @@ union ich8_hws_flash_regacc {
 /* ICH Flash Protected Region */
 union ich8_flash_protected_range {
 	struct ich8_pr {
-		u32 base:13;     /* 0:12 Protected Range Base */
-		u32 reserved1:2; /* 13:14 Reserved */
-		u32 rpe:1;       /* 15 Read Protection Enable */
-		u32 limit:13;    /* 16:28 Protected Range Limit */
-		u32 reserved2:2; /* 29:30 Reserved */
-		u32 wpe:1;       /* 31 Write Protection Enable */
+		u32 base:13;	/* 0:12 Protected Range Base */
+		u32 reserved1:2;	/* 13:14 Reserved */
+		u32 rpe:1;	/* 15 Read Protection Enable */
+		u32 limit:13;	/* 16:28 Protected Range Limit */
+		u32 reserved2:2;	/* 29:30 Reserved */
+		u32 wpe:1;	/* 31 Write Protection Enable */
 	} range;
 	u32 regval;
 };
@@ -362,21 +362,21 @@ static s32 e1000_init_phy_params_pchlan(struct e1000_hw *hw)
 	struct e1000_phy_info *phy = &hw->phy;
 	s32 ret_val;
 
-	phy->addr                     = 1;
-	phy->reset_delay_us           = 100;
-
-	phy->ops.set_page             = e1000_set_page_igp;
-	phy->ops.read_reg             = e1000_read_phy_reg_hv;
-	phy->ops.read_reg_locked      = e1000_read_phy_reg_hv_locked;
-	phy->ops.read_reg_page        = e1000_read_phy_reg_page_hv;
-	phy->ops.set_d0_lplu_state    = e1000_set_lplu_state_pchlan;
-	phy->ops.set_d3_lplu_state    = e1000_set_lplu_state_pchlan;
-	phy->ops.write_reg            = e1000_write_phy_reg_hv;
-	phy->ops.write_reg_locked     = e1000_write_phy_reg_hv_locked;
-	phy->ops.write_reg_page       = e1000_write_phy_reg_page_hv;
-	phy->ops.power_up             = e1000_power_up_phy_copper;
-	phy->ops.power_down           = e1000_power_down_phy_copper_ich8lan;
-	phy->autoneg_mask             = AUTONEG_ADVERTISE_SPEED_DEFAULT;
+	phy->addr = 1;
+	phy->reset_delay_us = 100;
+
+	phy->ops.set_page = e1000_set_page_igp;
+	phy->ops.read_reg = e1000_read_phy_reg_hv;
+	phy->ops.read_reg_locked = e1000_read_phy_reg_hv_locked;
+	phy->ops.read_reg_page = e1000_read_phy_reg_page_hv;
+	phy->ops.set_d0_lplu_state = e1000_set_lplu_state_pchlan;
+	phy->ops.set_d3_lplu_state = e1000_set_lplu_state_pchlan;
+	phy->ops.write_reg = e1000_write_phy_reg_hv;
+	phy->ops.write_reg_locked = e1000_write_phy_reg_hv_locked;
+	phy->ops.write_reg_page = e1000_write_phy_reg_page_hv;
+	phy->ops.power_up = e1000_power_up_phy_copper;
+	phy->ops.power_down = e1000_power_down_phy_copper_ich8lan;
+	phy->autoneg_mask = AUTONEG_ADVERTISE_SPEED_DEFAULT;
 
 	phy->id = e1000_phy_unknown;
 
@@ -445,11 +445,11 @@ static s32 e1000_init_phy_params_ich8lan(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 i = 0;
 
-	phy->addr			= 1;
-	phy->reset_delay_us		= 100;
+	phy->addr = 1;
+	phy->reset_delay_us = 100;
 
-	phy->ops.power_up               = e1000_power_up_phy_copper;
-	phy->ops.power_down             = e1000_power_down_phy_copper_ich8lan;
+	phy->ops.power_up = e1000_power_up_phy_copper;
+	phy->ops.power_down = e1000_power_down_phy_copper_ich8lan;
 
 	/* We may need to do this twice - once for IGP and if that fails,
 	 * we'll set BM func pointers and try again
@@ -457,7 +457,7 @@ static s32 e1000_init_phy_params_ich8lan(struct e1000_hw *hw)
 	ret_val = e1000e_determine_phy_address(hw);
 	if (ret_val) {
 		phy->ops.write_reg = e1000e_write_phy_reg_bm;
-		phy->ops.read_reg  = e1000e_read_phy_reg_bm;
+		phy->ops.read_reg = e1000e_read_phy_reg_bm;
 		ret_val = e1000e_determine_phy_address(hw);
 		if (ret_val) {
 			e_dbg("Cannot determine PHY addr. Erroring out\n");
@@ -560,7 +560,7 @@ static s32 e1000_init_nvm_params_ich8lan(struct e1000_hw *hw)
 	/* Clear shadow ram */
 	for (i = 0; i < nvm->word_size; i++) {
 		dev_spec->shadow_ram[i].modified = false;
-		dev_spec->shadow_ram[i].value    = 0xFFFF;
+		dev_spec->shadow_ram[i].value = 0xFFFF;
 	}
 
 	return 0;
@@ -1012,7 +1012,7 @@ static s32 e1000_check_for_copper_link_ich8lan(struct e1000_hw *hw)
 	hw->dev_spec.ich8lan.eee_lp_ability = 0;
 
 	if (!link)
-		return 0; /* No link detected */
+		return 0;	/* No link detected */
 
 	mac->get_link_status = false;
 
@@ -2816,7 +2816,7 @@ static s32 e1000_read_flash_data_ich8lan(struct e1000_hw *hw, u32 offset,
 	s32 ret_val = -E1000_ERR_NVM;
 	u8 count = 0;
 
-	if (size < 1  || size > 2 || offset > ICH_FLASH_LINEAR_ADDR_MASK)
+	if (size < 1 || size > 2 || offset > ICH_FLASH_LINEAR_ADDR_MASK)
 		return -E1000_ERR_NVM;
 
 	flash_linear_addr = ((ICH_FLASH_LINEAR_ADDR_MASK & offset) +
@@ -2939,7 +2939,7 @@ static s32 e1000_update_nvm_checksum_ich8lan(struct e1000_hw *hw)
 	 * write to bank 0 etc.  We also need to erase the segment that
 	 * is going to be written
 	 */
-	ret_val =  e1000_valid_nvm_bank_detect_ich8lan(hw, &bank);
+	ret_val = e1000_valid_nvm_bank_detect_ich8lan(hw, &bank);
 	if (ret_val) {
 		e_dbg("Could not detect valid bank, assuming bank 0\n");
 		bank = 0;
@@ -4073,7 +4073,7 @@ void e1000e_igp3_phy_powerdown_workaround_ich8lan(struct e1000_hw *hw)
 {
 	u32 reg;
 	u16 data;
-	u8  retry = 0;
+	u8 retry = 0;
 
 	if (hw->phy.type != e1000_phy_igp_3)
 		return;
diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index a27e3bc..77f81cb 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -1196,7 +1196,7 @@ static bool e1000_clean_tx_irq(struct e1000_ring *tx_ring)
 	while ((eop_desc->upper.data & cpu_to_le32(E1000_TXD_STAT_DD)) &&
 	       (count < tx_ring->count)) {
 		bool cleaned = false;
-		rmb(); /* read buffer_info after eop_desc */
+		rmb();		/* read buffer_info after eop_desc */
 		for (; !cleaned; count++) {
 			tx_desc = E1000_TX_DESC(*tx_ring, i);
 			buffer_info = &tx_ring->buffer_info[i];
@@ -1385,7 +1385,7 @@ static bool e1000_clean_rx_irq_ps(struct e1000_ring *rx_ring, int *work_done,
 
 				skb_put(skb, l1);
 				goto copydone;
-			} /* if */
+			}	/* if */
 		}
 
 		for (j = 0; j < PS_PAGE_BUFFERS; j++) {
@@ -1800,7 +1800,7 @@ static irqreturn_t e1000_intr(int __always_unused irq, void *data)
 	u32 rctl, icr = er32(ICR);
 
 	if (!icr || test_bit(__E1000_DOWN, &adapter->state))
-		return IRQ_NONE;  /* Not our interrupt */
+		return IRQ_NONE;	/* Not our interrupt */
 
 	/* IMS will not auto-mask if INT_ASSERTED is not set, and if it is
 	 * not set, then the adapter didn't send an interrupt
@@ -2487,7 +2487,7 @@ static unsigned int e1000_update_itr(u16 itr_setting, int packets, int bytes)
 		else if ((packets < 5) && (bytes > 512))
 			retval = low_latency;
 		break;
-	case low_latency:  /* 50 usec aka 20000 ints/s */
+	case low_latency:	/* 50 usec aka 20000 ints/s */
 		if (bytes > 10000) {
 			/* this if handles the TSO accounting */
 			if (bytes / packets > 8000)
@@ -2502,7 +2502,7 @@ static unsigned int e1000_update_itr(u16 itr_setting, int packets, int bytes)
 			retval = lowest_latency;
 		}
 		break;
-	case bulk_latency: /* 250 usec aka 4000 ints/s */
+	case bulk_latency:	/* 250 usec aka 4000 ints/s */
 		if (bytes > 25000) {
 			if (packets > 35)
 				retval = low_latency;
@@ -2554,7 +2554,7 @@ static void e1000_set_itr(struct e1000_adapter *adapter)
 		new_itr = 70000;
 		break;
 	case low_latency:
-		new_itr = 20000; /* aka hwitr = ~200 */
+		new_itr = 20000;	/* aka hwitr = ~200 */
 		break;
 	case bulk_latency:
 		new_itr = 4000;
@@ -2673,7 +2673,7 @@ static int e1000e_poll(struct napi_struct *napi, int weight)
 }
 
 static int e1000_vlan_rx_add_vid(struct net_device *netdev,
-				 __be16 proto, u16 vid)
+				 __always_unused __be16 proto, u16 vid)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct e1000_hw *hw = &adapter->hw;
@@ -2699,7 +2699,7 @@ static int e1000_vlan_rx_add_vid(struct net_device *netdev,
 }
 
 static int e1000_vlan_rx_kill_vid(struct net_device *netdev,
-				  __be16 proto, u16 vid)
+				  __always_unused __be16 proto, u16 vid)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct e1000_hw *hw = &adapter->hw;
@@ -3104,13 +3104,13 @@ static void e1000_setup_rctl(struct e1000_adapter *adapter)
 		/* UPE and MPE will be handled by normal PROMISC logic
 		 * in e1000e_set_rx_mode
 		 */
-		rctl |= (E1000_RCTL_SBP | /* Receive bad packets */
-			 E1000_RCTL_BAM | /* RX All Bcast Pkts */
-			 E1000_RCTL_PMCF); /* RX All MAC Ctrl Pkts */
+		rctl |= (E1000_RCTL_SBP |	/* Receive bad packets */
+			 E1000_RCTL_BAM |	/* RX All Bcast Pkts */
+			 E1000_RCTL_PMCF);	/* RX All MAC Ctrl Pkts */
 
-		rctl &= ~(E1000_RCTL_VFE | /* Disable VLAN filter */
-			  E1000_RCTL_DPF | /* Allow filtered pause */
-			  E1000_RCTL_CFIEN); /* Dis VLAN CFIEN Filter */
+		rctl &= ~(E1000_RCTL_VFE |	/* Disable VLAN filter */
+			  E1000_RCTL_DPF |	/* Allow filtered pause */
+			  E1000_RCTL_CFIEN);	/* Dis VLAN CFIEN Filter */
 		/* Do not mess with E1000_CTRL_VME, it affects transmit as well,
 		 * and that breaks VLANs.
 		 */
@@ -3799,7 +3799,7 @@ void e1000e_reset(struct e1000_adapter *adapter)
 		hwm = min(((pba << 10) * 9 / 10),
 			  ((pba << 10) - adapter->max_frame_size));
 
-		fc->high_water = hwm & E1000_FCRTH_RTH; /* 8-byte granularity */
+		fc->high_water = hwm & E1000_FCRTH_RTH;	/* 8-byte granularity */
 		fc->low_water = fc->high_water - 8;
 		break;
 	case e1000_pchlan:
@@ -3808,10 +3808,10 @@ void e1000e_reset(struct e1000_adapter *adapter)
 		 */
 		if (adapter->netdev->mtu > ETH_DATA_LEN) {
 			fc->high_water = 0x3500;
-			fc->low_water  = 0x1500;
+			fc->low_water = 0x1500;
 		} else {
 			fc->high_water = 0x5000;
-			fc->low_water  = 0x3000;
+			fc->low_water = 0x3000;
 		}
 		fc->refresh_time = 0x1000;
 		break;
@@ -4581,7 +4581,7 @@ static void e1000e_update_stats(struct e1000_adapter *adapter)
 	adapter->stats.crcerrs += er32(CRCERRS);
 	adapter->stats.gprc += er32(GPRC);
 	adapter->stats.gorc += er32(GORCL);
-	er32(GORCH); /* Clear gorc */
+	er32(GORCH);		/* Clear gorc */
 	adapter->stats.bprc += er32(BPRC);
 	adapter->stats.mprc += er32(MPRC);
 	adapter->stats.roc += er32(ROC);
@@ -4614,7 +4614,7 @@ static void e1000e_update_stats(struct e1000_adapter *adapter)
 	adapter->stats.xofftxc += er32(XOFFTXC);
 	adapter->stats.gptc += er32(GPTC);
 	adapter->stats.gotc += er32(GOTCL);
-	er32(GOTCH); /* Clear gotc */
+	er32(GOTCH);		/* Clear gotc */
 	adapter->stats.rnbc += er32(RNBC);
 	adapter->stats.ruc += er32(RUC);
 
@@ -5106,13 +5106,13 @@ static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb)
 	context_desc = E1000_CONTEXT_DESC(*tx_ring, i);
 	buffer_info = &tx_ring->buffer_info[i];
 
-	context_desc->lower_setup.ip_fields.ipcss  = ipcss;
-	context_desc->lower_setup.ip_fields.ipcso  = ipcso;
-	context_desc->lower_setup.ip_fields.ipcse  = cpu_to_le16(ipcse);
+	context_desc->lower_setup.ip_fields.ipcss = ipcss;
+	context_desc->lower_setup.ip_fields.ipcso = ipcso;
+	context_desc->lower_setup.ip_fields.ipcse = cpu_to_le16(ipcse);
 	context_desc->upper_setup.tcp_fields.tucss = tucss;
 	context_desc->upper_setup.tcp_fields.tucso = tucso;
 	context_desc->upper_setup.tcp_fields.tucse = 0;
-	context_desc->tcp_seg_setup.fields.mss     = cpu_to_le16(mss);
+	context_desc->tcp_seg_setup.fields.mss = cpu_to_le16(mss);
 	context_desc->tcp_seg_setup.fields.hdr_len = hdr_len;
 	context_desc->cmd_and_length = cpu_to_le32(cmd_length);
 
@@ -5363,7 +5363,7 @@ static void e1000_tx_queue(struct e1000_ring *tx_ring, int tx_flags, int count)
 static int e1000_transfer_dhcp_info(struct e1000_adapter *adapter,
 				    struct sk_buff *skb)
 {
-	struct e1000_hw *hw =  &adapter->hw;
+	struct e1000_hw *hw = &adapter->hw;
 	u16 length, offset;
 
 	if (vlan_tx_tag_present(skb) &&
@@ -6259,7 +6259,7 @@ static void e1000_netpoll(struct net_device *netdev)
 		e1000_intr_msi(adapter->pdev->irq, netdev);
 		enable_irq(adapter->pdev->irq);
 		break;
-	default: /* E1000E_INT_MODE_LEGACY */
+	default:		/* E1000E_INT_MODE_LEGACY */
 		disable_irq(adapter->pdev->irq);
 		e1000_intr(adapter->pdev->irq, netdev);
 		enable_irq(adapter->pdev->irq);
@@ -6589,9 +6589,9 @@ static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		adapter->eee_advert = MDIO_EEE_100TX | MDIO_EEE_1000T;
 
 	/* construct the net_device struct */
-	netdev->netdev_ops		= &e1000e_netdev_ops;
+	netdev->netdev_ops = &e1000e_netdev_ops;
 	e1000e_set_ethtool_ops(netdev);
-	netdev->watchdog_timeo		= 5 * HZ;
+	netdev->watchdog_timeo = 5 * HZ;
 	netif_napi_add(netdev, &adapter->napi, e1000e_poll, 64);
 	strlcpy(netdev->name, pci_name(pdev), sizeof(netdev->name));
 
@@ -7034,7 +7034,6 @@ static void __exit e1000_exit_module(void)
 }
 module_exit(e1000_exit_module);
 
-
 MODULE_AUTHOR("Intel Corporation, <linux.nics@intel.com>");
 MODULE_DESCRIPTION("Intel(R) PRO/1000 Network Driver");
 MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/intel/e1000e/nvm.c b/drivers/net/ethernet/intel/e1000e/nvm.c
index 44ddc0a..d70a039 100644
--- a/drivers/net/ethernet/intel/e1000e/nvm.c
+++ b/drivers/net/ethernet/intel/e1000e/nvm.c
@@ -117,7 +117,6 @@ static u16 e1000_shift_in_eec_bits(struct e1000_hw *hw, u16 count)
 	u16 data;
 
 	eecd = er32(EECD);
-
 	eecd &= ~(E1000_EECD_DO | E1000_EECD_DI);
 	data = 0;
 
diff --git a/drivers/net/ethernet/intel/e1000e/phy.c b/drivers/net/ethernet/intel/e1000e/phy.c
index 59c76a6..da2be59 100644
--- a/drivers/net/ethernet/intel/e1000e/phy.c
+++ b/drivers/net/ethernet/intel/e1000e/phy.c
@@ -1583,13 +1583,13 @@ s32 e1000e_check_downshift(struct e1000_hw *hw)
 	case e1000_phy_gg82563:
 	case e1000_phy_bm:
 	case e1000_phy_82578:
-		offset	= M88E1000_PHY_SPEC_STATUS;
-		mask	= M88E1000_PSSR_DOWNSHIFT;
+		offset = M88E1000_PHY_SPEC_STATUS;
+		mask = M88E1000_PSSR_DOWNSHIFT;
 		break;
 	case e1000_phy_igp_2:
 	case e1000_phy_igp_3:
-		offset	= IGP01E1000_PHY_LINK_HEALTH;
-		mask	= IGP01E1000_PLHR_SS_DOWNGRADE;
+		offset = IGP01E1000_PHY_LINK_HEALTH;
+		mask = IGP01E1000_PLHR_SS_DOWNGRADE;
 		break;
 	default:
 		/* speed downshift not supported */
@@ -1653,14 +1653,14 @@ s32 e1000_check_polarity_igp(struct e1000_hw *hw)
 
 	if ((data & IGP01E1000_PSSR_SPEED_MASK) ==
 	    IGP01E1000_PSSR_SPEED_1000MBPS) {
-		offset	= IGP01E1000_PHY_PCS_INIT_REG;
-		mask	= IGP01E1000_PHY_POLARITY_MASK;
+		offset = IGP01E1000_PHY_PCS_INIT_REG;
+		mask = IGP01E1000_PHY_POLARITY_MASK;
 	} else {
 		/* This really only applies to 10Mbps since
 		 * there is no polarity for 100Mbps (always 0).
 		 */
-		offset	= IGP01E1000_PHY_PORT_STATUS;
-		mask	= IGP01E1000_PSSR_POLARITY_REVERSED;
+		offset = IGP01E1000_PHY_PORT_STATUS;
+		mask = IGP01E1000_PSSR_POLARITY_REVERSED;
 	}
 
 	ret_val = e1e_rphy(hw, offset, &data);
@@ -1900,7 +1900,7 @@ s32 e1000e_get_cable_length_igp_2(struct e1000_hw *hw)
 s32 e1000e_get_phy_info_m88(struct e1000_hw *hw)
 {
 	struct e1000_phy_info *phy = &hw->phy;
-	s32  ret_val;
+	s32 ret_val;
 	u16 phy_data;
 	bool link;
 
@@ -2253,7 +2253,7 @@ enum e1000_phy_type e1000e_get_phy_type_from_id(u32 phy_id)
 	case M88E1011_I_PHY_ID:
 		phy_type = e1000_phy_m88;
 		break;
-	case IGP01E1000_I_PHY_ID: /* IGP 1 & 2 share this */
+	case IGP01E1000_I_PHY_ID:	/* IGP 1 & 2 share this */
 		phy_type = e1000_phy_igp_2;
 		break;
 	case GG82563_E_PHY_ID:
@@ -2317,7 +2317,7 @@ s32 e1000e_determine_phy_address(struct e1000_hw *hw)
 			/* If phy_type is valid, break - we found our
 			 * PHY address
 			 */
-			if (phy_type  != e1000_phy_unknown)
+			if (phy_type != e1000_phy_unknown)
 				return 0;
 
 			usleep_range(1000, 2000);
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.c b/drivers/net/ethernet/intel/igb/e1000_82575.c
index ff6a17c..f21a91a 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.c
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.c
@@ -401,12 +401,82 @@ static s32 igb_init_mac_params_82575(struct e1000_hw *hw)
 	return 0;
 }
 
+/**
+ *  igb_set_sfp_media_type_82575 - derives SFP module media type.
+ *  @hw: pointer to the HW structure
+ *
+ *  The media type is chosen based on SFP module.
+ *  compatibility flags retrieved from SFP ID EEPROM.
+ **/
+static s32 igb_set_sfp_media_type_82575(struct e1000_hw *hw)
+{
+	s32 ret_val = E1000_ERR_CONFIG;
+	u32 ctrl_ext = 0;
+	struct e1000_dev_spec_82575 *dev_spec = &hw->dev_spec._82575;
+	struct e1000_sfp_flags *eth_flags = &dev_spec->eth_flags;
+	u8 tranceiver_type = 0;
+	s32 timeout = 3;
+
+	/* Turn I2C interface ON and power on sfp cage */
+	ctrl_ext = rd32(E1000_CTRL_EXT);
+	ctrl_ext &= ~E1000_CTRL_EXT_SDP3_DATA;
+	wr32(E1000_CTRL_EXT, ctrl_ext | E1000_CTRL_I2C_ENA);
+
+	wrfl();
+
+	/* Read SFP module data */
+	while (timeout) {
+		ret_val = igb_read_sfp_data_byte(hw,
+			E1000_I2CCMD_SFP_DATA_ADDR(E1000_SFF_IDENTIFIER_OFFSET),
+			&tranceiver_type);
+		if (ret_val == 0)
+			break;
+		msleep(100);
+		timeout--;
+	}
+	if (ret_val != 0)
+		goto out;
+
+	ret_val = igb_read_sfp_data_byte(hw,
+			E1000_I2CCMD_SFP_DATA_ADDR(E1000_SFF_ETH_FLAGS_OFFSET),
+			(u8 *)eth_flags);
+	if (ret_val != 0)
+		goto out;
+
+	/* Check if there is some SFP module plugged and powered */
+	if ((tranceiver_type == E1000_SFF_IDENTIFIER_SFP) ||
+	    (tranceiver_type == E1000_SFF_IDENTIFIER_SFF)) {
+		dev_spec->module_plugged = true;
+		if (eth_flags->e1000_base_lx || eth_flags->e1000_base_sx) {
+			hw->phy.media_type = e1000_media_type_internal_serdes;
+		} else if (eth_flags->e100_base_fx) {
+			dev_spec->sgmii_active = true;
+			hw->phy.media_type = e1000_media_type_internal_serdes;
+		} else if (eth_flags->e1000_base_t) {
+			dev_spec->sgmii_active = true;
+			hw->phy.media_type = e1000_media_type_copper;
+		} else {
+			hw->phy.media_type = e1000_media_type_unknown;
+			hw_dbg("PHY module has not been recognized\n");
+			goto out;
+		}
+	} else {
+		hw->phy.media_type = e1000_media_type_unknown;
+	}
+	ret_val = 0;
+out:
+	/* Restore I2C interface setting */
+	wr32(E1000_CTRL_EXT, ctrl_ext);
+	return ret_val;
+}
+
 static s32 igb_get_invariants_82575(struct e1000_hw *hw)
 {
 	struct e1000_mac_info *mac = &hw->mac;
 	struct e1000_dev_spec_82575 * dev_spec = &hw->dev_spec._82575;
 	s32 ret_val;
 	u32 ctrl_ext = 0;
+	u32 link_mode = 0;
 
 	switch (hw->device_id) {
 	case E1000_DEV_ID_82575EB_COPPER:
@@ -470,16 +540,56 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
 	 */
 	hw->phy.media_type = e1000_media_type_copper;
 	dev_spec->sgmii_active = false;
+	dev_spec->module_plugged = false;
 
 	ctrl_ext = rd32(E1000_CTRL_EXT);
-	switch (ctrl_ext & E1000_CTRL_EXT_LINK_MODE_MASK) {
-	case E1000_CTRL_EXT_LINK_MODE_SGMII:
-		dev_spec->sgmii_active = true;
-		break;
+
+	link_mode = ctrl_ext & E1000_CTRL_EXT_LINK_MODE_MASK;
+	switch (link_mode) {
 	case E1000_CTRL_EXT_LINK_MODE_1000BASE_KX:
-	case E1000_CTRL_EXT_LINK_MODE_PCIE_SERDES:
 		hw->phy.media_type = e1000_media_type_internal_serdes;
 		break;
+	case E1000_CTRL_EXT_LINK_MODE_SGMII:
+		/* Get phy control interface type set (MDIO vs. I2C)*/
+		if (igb_sgmii_uses_mdio_82575(hw)) {
+			hw->phy.media_type = e1000_media_type_copper;
+			dev_spec->sgmii_active = true;
+			break;
+		}
+		/* fall through for I2C based SGMII */
+	case E1000_CTRL_EXT_LINK_MODE_PCIE_SERDES:
+		/* read media type from SFP EEPROM */
+		ret_val = igb_set_sfp_media_type_82575(hw);
+		if ((ret_val != 0) ||
+		    (hw->phy.media_type == e1000_media_type_unknown)) {
+			/* If media type was not identified then return media
+			 * type defined by the CTRL_EXT settings.
+			 */
+			hw->phy.media_type = e1000_media_type_internal_serdes;
+
+			if (link_mode == E1000_CTRL_EXT_LINK_MODE_SGMII) {
+				hw->phy.media_type = e1000_media_type_copper;
+				dev_spec->sgmii_active = true;
+			}
+
+			break;
+		}
+
+		/* do not change link mode for 100BaseFX */
+		if (dev_spec->eth_flags.e100_base_fx)
+			break;
+
+		/* change current link mode setting */
+		ctrl_ext &= ~E1000_CTRL_EXT_LINK_MODE_MASK;
+
+		if (hw->phy.media_type == e1000_media_type_copper)
+			ctrl_ext |= E1000_CTRL_EXT_LINK_MODE_SGMII;
+		else
+			ctrl_ext |= E1000_CTRL_EXT_LINK_MODE_PCIE_SERDES;
+
+		wr32(E1000_CTRL_EXT, ctrl_ext);
+
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/net/ethernet/intel/igb/e1000_defines.h b/drivers/net/ethernet/intel/igb/e1000_defines.h
index 31a0f82..aa201ab 100644
--- a/drivers/net/ethernet/intel/igb/e1000_defines.h
+++ b/drivers/net/ethernet/intel/igb/e1000_defines.h
@@ -61,20 +61,22 @@
 /* Clear Interrupt timers after IMS clear */
 /* packet buffer parity error detection enabled */
 /* descriptor FIFO parity error detection enable */
-#define E1000_CTRL_EXT_PBA_CLR        0x80000000 /* PBA Clear */
-#define E1000_I2CCMD_REG_ADDR_SHIFT   16
-#define E1000_I2CCMD_PHY_ADDR_SHIFT   24
-#define E1000_I2CCMD_OPCODE_READ      0x08000000
-#define E1000_I2CCMD_OPCODE_WRITE     0x00000000
-#define E1000_I2CCMD_READY            0x20000000
-#define E1000_I2CCMD_ERROR            0x80000000
-#define E1000_MAX_SGMII_PHY_REG_ADDR  255
-#define E1000_I2CCMD_PHY_TIMEOUT      200
-#define E1000_IVAR_VALID              0x80
-#define E1000_GPIE_NSICR              0x00000001
-#define E1000_GPIE_MSIX_MODE          0x00000010
-#define E1000_GPIE_EIAME              0x40000000
-#define E1000_GPIE_PBA                0x80000000
+#define E1000_CTRL_EXT_PBA_CLR		0x80000000 /* PBA Clear */
+#define E1000_I2CCMD_REG_ADDR_SHIFT	16
+#define E1000_I2CCMD_PHY_ADDR_SHIFT	24
+#define E1000_I2CCMD_OPCODE_READ	0x08000000
+#define E1000_I2CCMD_OPCODE_WRITE	0x00000000
+#define E1000_I2CCMD_READY		0x20000000
+#define E1000_I2CCMD_ERROR		0x80000000
+#define E1000_I2CCMD_SFP_DATA_ADDR(a)	(0x0000 + (a))
+#define E1000_I2CCMD_SFP_DIAG_ADDR(a)	(0x0100 + (a))
+#define E1000_MAX_SGMII_PHY_REG_ADDR	255
+#define E1000_I2CCMD_PHY_TIMEOUT	200
+#define E1000_IVAR_VALID		0x80
+#define E1000_GPIE_NSICR		0x00000001
+#define E1000_GPIE_MSIX_MODE		0x00000010
+#define E1000_GPIE_EIAME		0x40000000
+#define E1000_GPIE_PBA			0x80000000
 
 /* Receive Descriptor bit definitions */
 #define E1000_RXD_STAT_DD       0x01    /* Descriptor Done */
@@ -270,8 +272,10 @@
 #define AUTONEG_ADVERTISE_SPEED_DEFAULT   E1000_ALL_SPEED_DUPLEX
 
 /* LED Control */
-#define E1000_LEDCTL_LED0_MODE_SHIFT      0
-#define E1000_LEDCTL_LED0_BLINK           0x00000080
+#define E1000_LEDCTL_LED0_MODE_SHIFT	0
+#define E1000_LEDCTL_LED0_BLINK		0x00000080
+#define E1000_LEDCTL_LED0_MODE_MASK	0x0000000F
+#define E1000_LEDCTL_LED0_IVRT		0x00000040
 
 #define E1000_LEDCTL_MODE_LED_ON        0xE
 #define E1000_LEDCTL_MODE_LED_OFF       0xF
diff --git a/drivers/net/ethernet/intel/igb/e1000_hw.h b/drivers/net/ethernet/intel/igb/e1000_hw.h
index 488abb2..94d7866 100644
--- a/drivers/net/ethernet/intel/igb/e1000_hw.h
+++ b/drivers/net/ethernet/intel/igb/e1000_hw.h
@@ -528,6 +528,8 @@ struct e1000_dev_spec_82575 {
 	bool global_device_reset;
 	bool eee_disable;
 	bool clear_semaphore_once;
+	struct e1000_sfp_flags eth_flags;
+	bool module_plugged;
 };
 
 struct e1000_hw {
diff --git a/drivers/net/ethernet/intel/igb/e1000_i210.h b/drivers/net/ethernet/intel/igb/e1000_i210.h
index bfc08e0..5caa332 100644
--- a/drivers/net/ethernet/intel/igb/e1000_i210.h
+++ b/drivers/net/ethernet/intel/igb/e1000_i210.h
@@ -82,11 +82,11 @@ enum E1000_INVM_STRUCTURE_TYPE {
 #define E1000_INVM_MAJOR_SHIFT		4
 
 #define ID_LED_DEFAULT_I210		((ID_LED_OFF1_ON2  << 8) | \
-					 (ID_LED_OFF1_OFF2 <<  4) | \
-					 (ID_LED_DEF1_DEF2))
+					 (ID_LED_DEF1_DEF2 <<  4) | \
+					 (ID_LED_OFF1_OFF2))
 #define ID_LED_DEFAULT_I210_SERDES	((ID_LED_DEF1_DEF2 << 8) | \
 					 (ID_LED_DEF1_DEF2 <<  4) | \
-					 (ID_LED_DEF1_DEF2))
+					 (ID_LED_OFF1_ON2))
 
 /* NVM offset defaults for i211 device */
 #define NVM_INIT_CTRL_2_DEFAULT_I211	0X7243
diff --git a/drivers/net/ethernet/intel/igb/e1000_mac.c b/drivers/net/ethernet/intel/igb/e1000_mac.c
index 2559d70..bab556a 100644
--- a/drivers/net/ethernet/intel/igb/e1000_mac.c
+++ b/drivers/net/ethernet/intel/igb/e1000_mac.c
@@ -1332,7 +1332,13 @@ s32 igb_id_led_init(struct e1000_hw *hw)
 	u16 data, i, temp;
 	const u16 led_mask = 0x0F;
 
-	ret_val = igb_valid_led_default(hw, &data);
+	/* i210 and i211 devices have different LED mechanism */
+	if ((hw->mac.type == e1000_i210) ||
+	    (hw->mac.type == e1000_i211))
+		ret_val = igb_valid_led_default_i210(hw, &data);
+	else
+		ret_val = igb_valid_led_default(hw, &data);
+
 	if (ret_val)
 		goto out;
 
@@ -1406,15 +1412,34 @@ s32 igb_blink_led(struct e1000_hw *hw)
 	u32 ledctl_blink = 0;
 	u32 i;
 
-	/* set the blink bit for each LED that's "on" (0x0E)
-	 * in ledctl_mode2
-	 */
-	ledctl_blink = hw->mac.ledctl_mode2;
-	for (i = 0; i < 4; i++)
-		if (((hw->mac.ledctl_mode2 >> (i * 8)) & 0xFF) ==
-		    E1000_LEDCTL_MODE_LED_ON)
-			ledctl_blink |= (E1000_LEDCTL_LED0_BLINK <<
-					 (i * 8));
+	if (hw->phy.media_type == e1000_media_type_fiber) {
+		/* always blink LED0 for PCI-E fiber */
+		ledctl_blink = E1000_LEDCTL_LED0_BLINK |
+		     (E1000_LEDCTL_MODE_LED_ON << E1000_LEDCTL_LED0_MODE_SHIFT);
+	} else {
+		/* Set the blink bit for each LED that's "on" (0x0E)
+		 * (or "off" if inverted) in ledctl_mode2.  The blink
+		 * logic in hardware only works when mode is set to "on"
+		 * so it must be changed accordingly when the mode is
+		 * "off" and inverted.
+		 */
+		ledctl_blink = hw->mac.ledctl_mode2;
+		for (i = 0; i < 32; i += 8) {
+			u32 mode = (hw->mac.ledctl_mode2 >> i) &
+			    E1000_LEDCTL_LED0_MODE_MASK;
+			u32 led_default = hw->mac.ledctl_default >> i;
+
+			if ((!(led_default & E1000_LEDCTL_LED0_IVRT) &&
+			     (mode == E1000_LEDCTL_MODE_LED_ON)) ||
+			    ((led_default & E1000_LEDCTL_LED0_IVRT) &&
+			     (mode == E1000_LEDCTL_MODE_LED_OFF))) {
+				ledctl_blink &=
+				    ~(E1000_LEDCTL_LED0_MODE_MASK << i);
+				ledctl_blink |= (E1000_LEDCTL_LED0_BLINK |
+						 E1000_LEDCTL_MODE_LED_ON) << i;
+			}
+		}
+	}
 
 	wr32(E1000_LEDCTL, ledctl_blink);
 
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.c b/drivers/net/ethernet/intel/igb/e1000_phy.c
index 9979ebc..6046194 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.c
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.c
@@ -341,6 +341,130 @@ s32 igb_write_phy_reg_i2c(struct e1000_hw *hw, u32 offset, u16 data)
 }
 
 /**
+ *  igb_read_sfp_data_byte - Reads SFP module data.
+ *  @hw: pointer to the HW structure
+ *  @offset: byte location offset to be read
+ *  @data: read data buffer pointer
+ *
+ *  Reads one byte from SFP module data stored
+ *  in SFP resided EEPROM memory or SFP diagnostic area.
+ *  Function should be called with
+ *  E1000_I2CCMD_SFP_DATA_ADDR(<byte offset>) for SFP module database access
+ *  E1000_I2CCMD_SFP_DIAG_ADDR(<byte offset>) for SFP diagnostics parameters
+ *  access
+ **/
+s32 igb_read_sfp_data_byte(struct e1000_hw *hw, u16 offset, u8 *data)
+{
+	u32 i = 0;
+	u32 i2ccmd = 0;
+	u32 data_local = 0;
+
+	if (offset > E1000_I2CCMD_SFP_DIAG_ADDR(255)) {
+		hw_dbg("I2CCMD command address exceeds upper limit\n");
+		return -E1000_ERR_PHY;
+	}
+
+	/* Set up Op-code, EEPROM Address,in the I2CCMD
+	 * register. The MAC will take care of interfacing with the
+	 * EEPROM to retrieve the desired data.
+	 */
+	i2ccmd = ((offset << E1000_I2CCMD_REG_ADDR_SHIFT) |
+		  E1000_I2CCMD_OPCODE_READ);
+
+	wr32(E1000_I2CCMD, i2ccmd);
+
+	/* Poll the ready bit to see if the I2C read completed */
+	for (i = 0; i < E1000_I2CCMD_PHY_TIMEOUT; i++) {
+		udelay(50);
+		data_local = rd32(E1000_I2CCMD);
+		if (data_local & E1000_I2CCMD_READY)
+			break;
+	}
+	if (!(data_local & E1000_I2CCMD_READY)) {
+		hw_dbg("I2CCMD Read did not complete\n");
+		return -E1000_ERR_PHY;
+	}
+	if (data_local & E1000_I2CCMD_ERROR) {
+		hw_dbg("I2CCMD Error bit set\n");
+		return -E1000_ERR_PHY;
+	}
+	*data = (u8) data_local & 0xFF;
+
+	return 0;
+}
+
+/**
+ *  e1000_write_sfp_data_byte - Writes SFP module data.
+ *  @hw: pointer to the HW structure
+ *  @offset: byte location offset to write to
+ *  @data: data to write
+ *
+ *  Writes one byte to SFP module data stored
+ *  in SFP resided EEPROM memory or SFP diagnostic area.
+ *  Function should be called with
+ *  E1000_I2CCMD_SFP_DATA_ADDR(<byte offset>) for SFP module database access
+ *  E1000_I2CCMD_SFP_DIAG_ADDR(<byte offset>) for SFP diagnostics parameters
+ *  access
+ **/
+s32 e1000_write_sfp_data_byte(struct e1000_hw *hw, u16 offset, u8 data)
+{
+	u32 i = 0;
+	u32 i2ccmd = 0;
+	u32 data_local = 0;
+
+	if (offset > E1000_I2CCMD_SFP_DIAG_ADDR(255)) {
+		hw_dbg("I2CCMD command address exceeds upper limit\n");
+		return -E1000_ERR_PHY;
+	}
+	/* The programming interface is 16 bits wide
+	 * so we need to read the whole word first
+	 * then update appropriate byte lane and write
+	 * the updated word back.
+	 */
+	/* Set up Op-code, EEPROM Address,in the I2CCMD
+	 * register. The MAC will take care of interfacing
+	 * with an EEPROM to write the data given.
+	 */
+	i2ccmd = ((offset << E1000_I2CCMD_REG_ADDR_SHIFT) |
+		  E1000_I2CCMD_OPCODE_READ);
+	/* Set a command to read single word */
+	wr32(E1000_I2CCMD, i2ccmd);
+	for (i = 0; i < E1000_I2CCMD_PHY_TIMEOUT; i++) {
+		udelay(50);
+		/* Poll the ready bit to see if lastly
+		 * launched I2C operation completed
+		 */
+		i2ccmd = rd32(E1000_I2CCMD);
+		if (i2ccmd & E1000_I2CCMD_READY) {
+			/* Check if this is READ or WRITE phase */
+			if ((i2ccmd & E1000_I2CCMD_OPCODE_READ) ==
+			    E1000_I2CCMD_OPCODE_READ) {
+				/* Write the selected byte
+				 * lane and update whole word
+				 */
+				data_local = i2ccmd & 0xFF00;
+				data_local |= data;
+				i2ccmd = ((offset <<
+					E1000_I2CCMD_REG_ADDR_SHIFT) |
+					E1000_I2CCMD_OPCODE_WRITE | data_local);
+				wr32(E1000_I2CCMD, i2ccmd);
+			} else {
+				break;
+			}
+		}
+	}
+	if (!(i2ccmd & E1000_I2CCMD_READY)) {
+		hw_dbg("I2CCMD Write did not complete\n");
+		return -E1000_ERR_PHY;
+	}
+	if (i2ccmd & E1000_I2CCMD_ERROR) {
+		hw_dbg("I2CCMD Error bit set\n");
+		return -E1000_ERR_PHY;
+	}
+	return 0;
+}
+
+/**
  *  igb_read_phy_reg_igp - Read igp PHY register
  *  @hw: pointer to the HW structure
  *  @offset: register offset to be read
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.h b/drivers/net/ethernet/intel/igb/e1000_phy.h
index 784fd1c..6a0873f 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.h
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.h
@@ -69,6 +69,8 @@ s32  igb_read_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 *data);
 s32  igb_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data);
 s32  igb_read_phy_reg_i2c(struct e1000_hw *hw, u32 offset, u16 *data);
 s32  igb_write_phy_reg_i2c(struct e1000_hw *hw, u32 offset, u16 data);
+s32  igb_read_sfp_data_byte(struct e1000_hw *hw, u16 offset, u8 *data);
+s32  e1000_write_sfp_data_byte(struct e1000_hw *hw, u16 offset, u8 data);
 s32  igb_copper_link_setup_82580(struct e1000_hw *hw);
 s32  igb_get_phy_info_82580(struct e1000_hw *hw);
 s32  igb_phy_force_speed_duplex_82580(struct e1000_hw *hw);
@@ -157,4 +159,22 @@ s32  igb_check_polarity_m88(struct e1000_hw *hw);
 #define GS40G_CS_POWER_DOWN		0x0002
 #define GS40G_LINE_LB			0x4000
 
+/* SFP modules ID memory locations */
+#define E1000_SFF_IDENTIFIER_OFFSET	0x00
+#define E1000_SFF_IDENTIFIER_SFF	0x02
+#define E1000_SFF_IDENTIFIER_SFP	0x03
+
+#define E1000_SFF_ETH_FLAGS_OFFSET	0x06
+/* Flags for SFP modules compatible with ETH up to 1Gb */
+struct e1000_sfp_flags {
+	u8 e1000_base_sx:1;
+	u8 e1000_base_lx:1;
+	u8 e1000_base_cx:1;
+	u8 e1000_base_t:1;
+	u8 e100_base_lx:1;
+	u8 e100_base_fx:1;
+	u8 e10_base_bx10:1;
+	u8 e10_base_px:1;
+};
+
 #endif
diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h
index 9d6c075..15ea8dc 100644
--- a/drivers/net/ethernet/intel/igb/igb.h
+++ b/drivers/net/ethernet/intel/igb/igb.h
@@ -322,11 +322,6 @@ static inline int igb_desc_unused(struct igb_ring *ring)
 	return ring->count + ring->next_to_clean - ring->next_to_use - 1;
 }
 
-struct igb_i2c_client_list {
-	struct i2c_client *client;
-	struct igb_i2c_client_list *next;
-};
-
 #ifdef CONFIG_IGB_HWMON
 
 #define IGB_HWMON_TYPE_LOC	0
@@ -514,13 +509,18 @@ extern void igb_ptp_rx_rgtstamp(struct igb_q_vector *q_vector,
 extern void igb_ptp_rx_pktstamp(struct igb_q_vector *q_vector,
 				unsigned char *va,
 				struct sk_buff *skb);
-static inline void igb_ptp_rx_hwtstamp(struct igb_q_vector *q_vector,
+static inline void igb_ptp_rx_hwtstamp(struct igb_ring *rx_ring,
 				       union e1000_adv_rx_desc *rx_desc,
 				       struct sk_buff *skb)
 {
 	if (igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TS) &&
 	    !igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TSIP))
-		igb_ptp_rx_rgtstamp(q_vector, skb);
+		igb_ptp_rx_rgtstamp(rx_ring->q_vector, skb);
+
+	/* Update the last_rx_timestamp timer in order to enable watchdog check
+	 * for error case of latched timestamp on a dropped packet.
+	 */
+	rx_ring->last_rx_timestamp = jiffies;
 }
 
 extern int igb_ptp_hwtstamp_ioctl(struct net_device *netdev,
diff --git a/drivers/net/ethernet/intel/igb/igb_ethtool.c b/drivers/net/ethernet/intel/igb/igb_ethtool.c
index 7876240f..85fe7b5 100644
--- a/drivers/net/ethernet/intel/igb/igb_ethtool.c
+++ b/drivers/net/ethernet/intel/igb/igb_ethtool.c
@@ -142,6 +142,8 @@ static int igb_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 {
 	struct igb_adapter *adapter = netdev_priv(netdev);
 	struct e1000_hw *hw = &adapter->hw;
+	struct e1000_dev_spec_82575 *dev_spec = &hw->dev_spec._82575;
+	struct e1000_sfp_flags *eth_flags = &dev_spec->eth_flags;
 	u32 status;
 
 	if (hw->phy.media_type == e1000_media_type_copper) {
@@ -162,49 +164,26 @@ static int igb_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 			ecmd->advertising |= hw->phy.autoneg_advertised;
 		}
 
-		if (hw->mac.autoneg != 1)
-			ecmd->advertising &= ~(ADVERTISED_Pause |
-					       ADVERTISED_Asym_Pause);
-
-		if (hw->fc.requested_mode == e1000_fc_full)
-			ecmd->advertising |= ADVERTISED_Pause;
-		else if (hw->fc.requested_mode == e1000_fc_rx_pause)
-			ecmd->advertising |= (ADVERTISED_Pause |
-					      ADVERTISED_Asym_Pause);
-		else if (hw->fc.requested_mode == e1000_fc_tx_pause)
-			ecmd->advertising |=  ADVERTISED_Asym_Pause;
-		else
-			ecmd->advertising &= ~(ADVERTISED_Pause |
-					       ADVERTISED_Asym_Pause);
-
 		ecmd->port = PORT_TP;
 		ecmd->phy_address = hw->phy.addr;
 		ecmd->transceiver = XCVR_INTERNAL;
 	} else {
-		ecmd->supported = (SUPPORTED_1000baseT_Full |
-				   SUPPORTED_100baseT_Full |
-				   SUPPORTED_FIBRE |
+		ecmd->supported = (SUPPORTED_FIBRE |
 				   SUPPORTED_Autoneg |
 				   SUPPORTED_Pause);
-		if (hw->mac.type == e1000_i354)
-				ecmd->supported |= SUPPORTED_2500baseX_Full;
-
 		ecmd->advertising = ADVERTISED_FIBRE;
-
-		switch (adapter->link_speed) {
-		case SPEED_2500:
-			ecmd->advertising = ADVERTISED_2500baseX_Full;
-			break;
-		case SPEED_1000:
-			ecmd->advertising = ADVERTISED_1000baseT_Full;
-			break;
-		case SPEED_100:
-			ecmd->advertising = ADVERTISED_100baseT_Full;
-			break;
-		default:
-			break;
+		if (hw->mac.type == e1000_i354) {
+			ecmd->supported |= SUPPORTED_2500baseX_Full;
+			ecmd->advertising |= ADVERTISED_2500baseX_Full;
+		}
+		if ((eth_flags->e1000_base_lx) || (eth_flags->e1000_base_sx)) {
+			ecmd->supported |= SUPPORTED_1000baseT_Full;
+			ecmd->advertising |= ADVERTISED_1000baseT_Full;
+		}
+		if (eth_flags->e100_base_fx) {
+			ecmd->supported |= SUPPORTED_100baseT_Full;
+			ecmd->advertising |= ADVERTISED_100baseT_Full;
 		}
-
 		if (hw->mac.autoneg == 1)
 			ecmd->advertising |= ADVERTISED_Autoneg;
 
@@ -212,6 +191,21 @@ static int igb_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
 		ecmd->transceiver = XCVR_EXTERNAL;
 	}
 
+	if (hw->mac.autoneg != 1)
+		ecmd->advertising &= ~(ADVERTISED_Pause |
+				       ADVERTISED_Asym_Pause);
+
+	if (hw->fc.requested_mode == e1000_fc_full)
+		ecmd->advertising |= ADVERTISED_Pause;
+	else if (hw->fc.requested_mode == e1000_fc_rx_pause)
+		ecmd->advertising |= (ADVERTISED_Pause |
+				      ADVERTISED_Asym_Pause);
+	else if (hw->fc.requested_mode == e1000_fc_tx_pause)
+		ecmd->advertising |=  ADVERTISED_Asym_Pause;
+	else
+		ecmd->advertising &= ~(ADVERTISED_Pause |
+				       ADVERTISED_Asym_Pause);
+
 	status = rd32(E1000_STATUS);
 
 	if (status & E1000_STATUS_LU) {
@@ -392,6 +386,10 @@ static int igb_set_pauseparam(struct net_device *netdev,
 	struct e1000_hw *hw = &adapter->hw;
 	int retval = 0;
 
+	/* 100basefx does not support setting link flow control */
+	if (hw->dev_spec._82575.eth_flags.e100_base_fx)
+		return -EINVAL;
+
 	adapter->fc_autoneg = pause->autoneg;
 
 	while (test_and_set_bit(__IGB_RESETTING, &adapter->state))
@@ -813,10 +811,8 @@ static int igb_set_eeprom(struct net_device *netdev,
 	ret_val = hw->nvm.ops.write(hw, first_word,
 				    last_word - first_word + 1, eeprom_buff);
 
-	/* Update the checksum over the first part of the EEPROM if needed
-	 * and flush shadow RAM for 82573 controllers
-	 */
-	if ((ret_val == 0) && ((first_word <= NVM_CHECKSUM_REG)))
+	/* Update the checksum if nvm write succeeded */
+	if (ret_val == 0)
 		hw->nvm.ops.update(hw);
 
 	igb_set_fw_version(adapter);
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 64cbe0d..6a0c1b6 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -1667,10 +1667,13 @@ void igb_down(struct igb_adapter *adapter)
 	wrfl();
 	msleep(10);
 
-	for (i = 0; i < adapter->num_q_vectors; i++)
+	igb_irq_disable(adapter);
+
+	for (i = 0; i < adapter->num_q_vectors; i++) {
+		napi_synchronize(&(adapter->q_vector[i]->napi));
 		napi_disable(&(adapter->q_vector[i]->napi));
+	}
 
-	igb_irq_disable(adapter);
 
 	del_timer_sync(&adapter->watchdog_timer);
 	del_timer_sync(&adapter->phy_info_timer);
@@ -6622,7 +6625,7 @@ static void igb_process_skb_fields(struct igb_ring *rx_ring,
 
 	igb_rx_checksum(rx_ring, rx_desc, skb);
 
-	igb_ptp_rx_hwtstamp(rx_ring->q_vector, rx_desc, skb);
+	igb_ptp_rx_hwtstamp(rx_ring, rx_desc, skb);
 
 	if ((dev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
 	    igb_test_staterr(rx_desc, E1000_RXD_STAT_VP)) {
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index ca93238..fb098b4 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -52,6 +52,11 @@
 #include <linux/dca.h>
 #endif
 
+#include <net/ll_poll.h>
+
+#ifdef CONFIG_NET_LL_RX_POLL
+#define LL_EXTENDED_STATS
+#endif
 /* common prefix used by pr_<> macros */
 #undef pr_fmt
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
@@ -182,6 +187,11 @@ struct ixgbe_rx_buffer {
 struct ixgbe_queue_stats {
 	u64 packets;
 	u64 bytes;
+#ifdef LL_EXTENDED_STATS
+	u64 yields;
+	u64 misses;
+	u64 cleaned;
+#endif  /* LL_EXTENDED_STATS */
 };
 
 struct ixgbe_tx_queue_stats {
@@ -356,9 +366,133 @@ struct ixgbe_q_vector {
 	struct rcu_head rcu;	/* to avoid race with update stats on free */
 	char name[IFNAMSIZ + 9];
 
+#ifdef CONFIG_NET_LL_RX_POLL
+	unsigned int state;
+#define IXGBE_QV_STATE_IDLE        0
+#define IXGBE_QV_STATE_NAPI	   1    /* NAPI owns this QV */
+#define IXGBE_QV_STATE_POLL	   2    /* poll owns this QV */
+#define IXGBE_QV_LOCKED (IXGBE_QV_STATE_NAPI | IXGBE_QV_STATE_POLL)
+#define IXGBE_QV_STATE_NAPI_YIELD  4    /* NAPI yielded this QV */
+#define IXGBE_QV_STATE_POLL_YIELD  8    /* poll yielded this QV */
+#define IXGBE_QV_YIELD (IXGBE_QV_STATE_NAPI_YIELD | IXGBE_QV_STATE_POLL_YIELD)
+#define IXGBE_QV_USER_PEND (IXGBE_QV_STATE_POLL | IXGBE_QV_STATE_POLL_YIELD)
+	spinlock_t lock;
+#endif  /* CONFIG_NET_LL_RX_POLL */
+
 	/* for dynamic allocation of rings associated with this q_vector */
 	struct ixgbe_ring ring[0] ____cacheline_internodealigned_in_smp;
 };
+#ifdef CONFIG_NET_LL_RX_POLL
+static inline void ixgbe_qv_init_lock(struct ixgbe_q_vector *q_vector)
+{
+
+	spin_lock_init(&q_vector->lock);
+	q_vector->state = IXGBE_QV_STATE_IDLE;
+}
+
+/* called from the device poll routine to get ownership of a q_vector */
+static inline bool ixgbe_qv_lock_napi(struct ixgbe_q_vector *q_vector)
+{
+	int rc = true;
+	spin_lock(&q_vector->lock);
+	if (q_vector->state & IXGBE_QV_LOCKED) {
+		WARN_ON(q_vector->state & IXGBE_QV_STATE_NAPI);
+		q_vector->state |= IXGBE_QV_STATE_NAPI_YIELD;
+		rc = false;
+#ifdef LL_EXTENDED_STATS
+		q_vector->tx.ring->stats.yields++;
+#endif
+	} else
+		/* we don't care if someone yielded */
+		q_vector->state = IXGBE_QV_STATE_NAPI;
+	spin_unlock(&q_vector->lock);
+	return rc;
+}
+
+/* returns true is someone tried to get the qv while napi had it */
+static inline bool ixgbe_qv_unlock_napi(struct ixgbe_q_vector *q_vector)
+{
+	int rc = false;
+	spin_lock(&q_vector->lock);
+	WARN_ON(q_vector->state & (IXGBE_QV_STATE_POLL |
+			       IXGBE_QV_STATE_NAPI_YIELD));
+
+	if (q_vector->state & IXGBE_QV_STATE_POLL_YIELD)
+		rc = true;
+	q_vector->state = IXGBE_QV_STATE_IDLE;
+	spin_unlock(&q_vector->lock);
+	return rc;
+}
+
+/* called from ixgbe_low_latency_poll() */
+static inline bool ixgbe_qv_lock_poll(struct ixgbe_q_vector *q_vector)
+{
+	int rc = true;
+	spin_lock_bh(&q_vector->lock);
+	if ((q_vector->state & IXGBE_QV_LOCKED)) {
+		q_vector->state |= IXGBE_QV_STATE_POLL_YIELD;
+		rc = false;
+#ifdef LL_EXTENDED_STATS
+		q_vector->rx.ring->stats.yields++;
+#endif
+	} else
+		/* preserve yield marks */
+		q_vector->state |= IXGBE_QV_STATE_POLL;
+	spin_unlock_bh(&q_vector->lock);
+	return rc;
+}
+
+/* returns true if someone tried to get the qv while it was locked */
+static inline bool ixgbe_qv_unlock_poll(struct ixgbe_q_vector *q_vector)
+{
+	int rc = false;
+	spin_lock_bh(&q_vector->lock);
+	WARN_ON(q_vector->state & (IXGBE_QV_STATE_NAPI));
+
+	if (q_vector->state & IXGBE_QV_STATE_POLL_YIELD)
+		rc = true;
+	q_vector->state = IXGBE_QV_STATE_IDLE;
+	spin_unlock_bh(&q_vector->lock);
+	return rc;
+}
+
+/* true if a socket is polling, even if it did not get the lock */
+static inline bool ixgbe_qv_ll_polling(struct ixgbe_q_vector *q_vector)
+{
+	WARN_ON(!(q_vector->state & IXGBE_QV_LOCKED));
+	return q_vector->state & IXGBE_QV_USER_PEND;
+}
+#else /* CONFIG_NET_LL_RX_POLL */
+static inline void ixgbe_qv_init_lock(struct ixgbe_q_vector *q_vector)
+{
+}
+
+static inline bool ixgbe_qv_lock_napi(struct ixgbe_q_vector *q_vector)
+{
+	return true;
+}
+
+static inline bool ixgbe_qv_unlock_napi(struct ixgbe_q_vector *q_vector)
+{
+	return false;
+}
+
+static inline bool ixgbe_qv_lock_poll(struct ixgbe_q_vector *q_vector)
+{
+	return false;
+}
+
+static inline bool ixgbe_qv_unlock_poll(struct ixgbe_q_vector *q_vector)
+{
+	return false;
+}
+
+static inline bool ixgbe_qv_ll_polling(struct ixgbe_q_vector *q_vector)
+{
+	return false;
+}
+#endif /* CONFIG_NET_LL_RX_POLL */
+
 #ifdef CONFIG_IXGBE_HWMON
 
 #define IXGBE_HWMON_TYPE_LOC		0
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.c
index 1f2c805..e055e00 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.c
@@ -380,3 +380,26 @@ s32 ixgbe_dcb_hw_ets_config(struct ixgbe_hw *hw,
 	}
 	return 0;
 }
+
+static void ixgbe_dcb_read_rtrup2tc_82599(struct ixgbe_hw *hw, u8 *map)
+{
+	u32 reg, i;
+
+	reg = IXGBE_READ_REG(hw, IXGBE_RTRUP2TC);
+	for (i = 0; i < MAX_USER_PRIORITY; i++)
+		map[i] = IXGBE_RTRUP2TC_UP_MASK &
+			(reg >> (i * IXGBE_RTRUP2TC_UP_SHIFT));
+	return;
+}
+
+void ixgbe_dcb_read_rtrup2tc(struct ixgbe_hw *hw, u8 *map)
+{
+	switch (hw->mac.type) {
+	case ixgbe_mac_82599EB:
+	case ixgbe_mac_X540:
+		ixgbe_dcb_read_rtrup2tc_82599(hw, map);
+		break;
+	default:
+		break;
+	}
+}
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.h
index 1634de8..fc0a2dd 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.h
@@ -159,6 +159,8 @@ s32 ixgbe_dcb_hw_ets_config(struct ixgbe_hw *hw, u16 *refill, u16 *max,
 s32 ixgbe_dcb_hw_pfc_config(struct ixgbe_hw *hw, u8 pfc_en, u8 *tc_prio);
 s32 ixgbe_dcb_hw_config(struct ixgbe_hw *, struct ixgbe_dcb_config *);
 
+void ixgbe_dcb_read_rtrup2tc(struct ixgbe_hw *hw, u8 *map);
+
 /* DCB definitions for credit calculation */
 #define DCB_CREDIT_QUANTUM	64   /* DCB Quantum */
 #define MAX_CREDIT_REFILL       511  /* 0x1FF * 64B = 32704B */
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.h
index a4ef076..d71d9ce 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.h
@@ -45,6 +45,7 @@
 
 /* Receive UP2TC mapping */
 #define IXGBE_RTRUP2TC_UP_SHIFT 3
+#define IXGBE_RTRUP2TC_UP_MASK	7
 /* Transmit UP2TC mapping */
 #define IXGBE_RTTUP2TC_UP_SHIFT 3
 
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
index f3d68f9..edd89a1 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
@@ -554,6 +554,9 @@ static int ixgbe_dcbnl_ieee_setets(struct net_device *dev,
 		for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++)
 			adapter->ixgbe_ieee_ets->prio_tc[i] =
 				IEEE_8021QAZ_MAX_TCS;
+		/* if possible update UP2TC mappings from HW */
+		ixgbe_dcb_read_rtrup2tc(&adapter->hw,
+					adapter->ixgbe_ieee_ets->prio_tc);
 	}
 
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
index d375472..24e2e7a 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
@@ -1054,6 +1054,12 @@ static void ixgbe_get_ethtool_stats(struct net_device *netdev,
 			data[i] = 0;
 			data[i+1] = 0;
 			i += 2;
+#ifdef LL_EXTENDED_STATS
+			data[i] = 0;
+			data[i+1] = 0;
+			data[i+2] = 0;
+			i += 3;
+#endif
 			continue;
 		}
 
@@ -1063,6 +1069,12 @@ static void ixgbe_get_ethtool_stats(struct net_device *netdev,
 			data[i+1] = ring->stats.bytes;
 		} while (u64_stats_fetch_retry_bh(&ring->syncp, start));
 		i += 2;
+#ifdef LL_EXTENDED_STATS
+		data[i] = ring->stats.yields;
+		data[i+1] = ring->stats.misses;
+		data[i+2] = ring->stats.cleaned;
+		i += 3;
+#endif
 	}
 	for (j = 0; j < IXGBE_NUM_RX_QUEUES; j++) {
 		ring = adapter->rx_ring[j];
@@ -1070,6 +1082,12 @@ static void ixgbe_get_ethtool_stats(struct net_device *netdev,
 			data[i] = 0;
 			data[i+1] = 0;
 			i += 2;
+#ifdef LL_EXTENDED_STATS
+			data[i] = 0;
+			data[i+1] = 0;
+			data[i+2] = 0;
+			i += 3;
+#endif
 			continue;
 		}
 
@@ -1079,6 +1097,12 @@ static void ixgbe_get_ethtool_stats(struct net_device *netdev,
 			data[i+1] = ring->stats.bytes;
 		} while (u64_stats_fetch_retry_bh(&ring->syncp, start));
 		i += 2;
+#ifdef LL_EXTENDED_STATS
+		data[i] = ring->stats.yields;
+		data[i+1] = ring->stats.misses;
+		data[i+2] = ring->stats.cleaned;
+		i += 3;
+#endif
 	}
 
 	for (j = 0; j < IXGBE_MAX_PACKET_BUFFERS; j++) {
@@ -1115,12 +1139,28 @@ static void ixgbe_get_strings(struct net_device *netdev, u32 stringset,
 			p += ETH_GSTRING_LEN;
 			sprintf(p, "tx_queue_%u_bytes", i);
 			p += ETH_GSTRING_LEN;
+#ifdef LL_EXTENDED_STATS
+			sprintf(p, "tx_q_%u_napi_yield", i);
+			p += ETH_GSTRING_LEN;
+			sprintf(p, "tx_q_%u_misses", i);
+			p += ETH_GSTRING_LEN;
+			sprintf(p, "tx_q_%u_cleaned", i);
+			p += ETH_GSTRING_LEN;
+#endif /* LL_EXTENDED_STATS */
 		}
 		for (i = 0; i < IXGBE_NUM_RX_QUEUES; i++) {
 			sprintf(p, "rx_queue_%u_packets", i);
 			p += ETH_GSTRING_LEN;
 			sprintf(p, "rx_queue_%u_bytes", i);
 			p += ETH_GSTRING_LEN;
+#ifdef LL_EXTENDED_STATS
+			sprintf(p, "rx_q_%u_ll_poll_yield", i);
+			p += ETH_GSTRING_LEN;
+			sprintf(p, "rx_q_%u_misses", i);
+			p += ETH_GSTRING_LEN;
+			sprintf(p, "rx_q_%u_cleaned", i);
+			p += ETH_GSTRING_LEN;
+#endif /* LL_EXTENDED_STATS */
 		}
 		for (i = 0; i < IXGBE_MAX_PACKET_BUFFERS; i++) {
 			sprintf(p, "tx_pb_%u_pxon", i);
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c
index ef5f7a6..90b4e10 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c
@@ -811,6 +811,7 @@ static int ixgbe_alloc_q_vector(struct ixgbe_adapter *adapter,
 	/* initialize NAPI */
 	netif_napi_add(adapter->netdev, &q_vector->napi,
 		       ixgbe_poll, 64);
+	napi_hash_add(&q_vector->napi);
 
 	/* tie q_vector and adapter together */
 	adapter->q_vector[v_idx] = q_vector;
@@ -931,6 +932,7 @@ static void ixgbe_free_q_vector(struct ixgbe_adapter *adapter, int v_idx)
 		adapter->rx_ring[ring->queue_index] = NULL;
 
 	adapter->q_vector[v_idx] = NULL;
+	napi_hash_del(&q_vector->napi);
 	netif_napi_del(&q_vector->napi);
 
 	/*
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index d30fbdd..047ebaa 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -1504,7 +1504,9 @@ static void ixgbe_rx_skb(struct ixgbe_q_vector *q_vector,
 {
 	struct ixgbe_adapter *adapter = q_vector->adapter;
 
-	if (!(adapter->flags & IXGBE_FLAG_IN_NETPOLL))
+	if (ixgbe_qv_ll_polling(q_vector))
+		netif_receive_skb(skb);
+	else if (!(adapter->flags & IXGBE_FLAG_IN_NETPOLL))
 		napi_gro_receive(&q_vector->napi, skb);
 	else
 		netif_rx(skb);
@@ -1892,9 +1894,9 @@ dma_sync:
  * expensive overhead for IOMMU access this provides a means of avoiding
  * it by maintaining the mapping of the page to the syste.
  *
- * Returns true if all work is completed without reaching budget
+ * Returns amount of work completed
  **/
-static bool ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
+static int ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 			       struct ixgbe_ring *rx_ring,
 			       const int budget)
 {
@@ -1976,6 +1978,7 @@ static bool ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 		}
 
 #endif /* IXGBE_FCOE */
+		skb_mark_ll(skb, &q_vector->napi);
 		ixgbe_rx_skb(q_vector, skb);
 
 		/* update budget accounting */
@@ -1992,9 +1995,43 @@ static bool ixgbe_clean_rx_irq(struct ixgbe_q_vector *q_vector,
 	if (cleaned_count)
 		ixgbe_alloc_rx_buffers(rx_ring, cleaned_count);
 
-	return (total_rx_packets < budget);
+	return total_rx_packets;
 }
 
+#ifdef CONFIG_NET_LL_RX_POLL
+/* must be called with local_bh_disable()d */
+static int ixgbe_low_latency_recv(struct napi_struct *napi)
+{
+	struct ixgbe_q_vector *q_vector =
+			container_of(napi, struct ixgbe_q_vector, napi);
+	struct ixgbe_adapter *adapter = q_vector->adapter;
+	struct ixgbe_ring  *ring;
+	int found = 0;
+
+	if (test_bit(__IXGBE_DOWN, &adapter->state))
+		return LL_FLUSH_FAILED;
+
+	if (!ixgbe_qv_lock_poll(q_vector))
+		return LL_FLUSH_BUSY;
+
+	ixgbe_for_each_ring(ring, q_vector->rx) {
+		found = ixgbe_clean_rx_irq(q_vector, ring, 4);
+#ifdef LL_EXTENDED_STATS
+		if (found)
+			ring->stats.cleaned += found;
+		else
+			ring->stats.misses++;
+#endif
+		if (found)
+			break;
+	}
+
+	ixgbe_qv_unlock_poll(q_vector);
+
+	return found;
+}
+#endif	/* CONFIG_NET_LL_RX_POLL */
+
 /**
  * ixgbe_configure_msix - Configure MSI-X hardware
  * @adapter: board private structure
@@ -2550,6 +2587,9 @@ int ixgbe_poll(struct napi_struct *napi, int budget)
 	ixgbe_for_each_ring(ring, q_vector->tx)
 		clean_complete &= !!ixgbe_clean_tx_irq(q_vector, ring);
 
+	if (!ixgbe_qv_lock_napi(q_vector))
+		return budget;
+
 	/* attempt to distribute budget to each queue fairly, but don't allow
 	 * the budget to go below 1 because we'll exit polling */
 	if (q_vector->rx.count > 1)
@@ -2558,9 +2598,10 @@ int ixgbe_poll(struct napi_struct *napi, int budget)
 		per_ring_budget = budget;
 
 	ixgbe_for_each_ring(ring, q_vector->rx)
-		clean_complete &= ixgbe_clean_rx_irq(q_vector, ring,
-						     per_ring_budget);
+		clean_complete &= (ixgbe_clean_rx_irq(q_vector, ring,
+				   per_ring_budget) < per_ring_budget);
 
+	ixgbe_qv_unlock_napi(q_vector);
 	/* If all work not completed, return budget and keep polling */
 	if (!clean_complete)
 		return budget;
@@ -3747,16 +3788,25 @@ static void ixgbe_napi_enable_all(struct ixgbe_adapter *adapter)
 {
 	int q_idx;
 
-	for (q_idx = 0; q_idx < adapter->num_q_vectors; q_idx++)
+	for (q_idx = 0; q_idx < adapter->num_q_vectors; q_idx++) {
+		ixgbe_qv_init_lock(adapter->q_vector[q_idx]);
 		napi_enable(&adapter->q_vector[q_idx]->napi);
+	}
 }
 
 static void ixgbe_napi_disable_all(struct ixgbe_adapter *adapter)
 {
 	int q_idx;
 
-	for (q_idx = 0; q_idx < adapter->num_q_vectors; q_idx++)
+	local_bh_disable(); /* for ixgbe_qv_lock_napi() */
+	for (q_idx = 0; q_idx < adapter->num_q_vectors; q_idx++) {
 		napi_disable(&adapter->q_vector[q_idx]->napi);
+		while (!ixgbe_qv_lock_napi(adapter->q_vector[q_idx])) {
+			pr_info("QV %d locked\n", q_idx);
+			mdelay(1);
+		}
+	}
+	local_bh_enable();
 }
 
 #ifdef CONFIG_IXGBE_DCB
@@ -7177,6 +7227,9 @@ static const struct net_device_ops ixgbe_netdev_ops = {
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= ixgbe_netpoll,
 #endif
+#ifdef CONFIG_NET_LL_RX_POLL
+	.ndo_ll_poll		= ixgbe_low_latency_recv,
+#endif
 #ifdef IXGBE_FCOE
 	.ndo_fcoe_ddp_setup = ixgbe_fcoe_ddp_get,
 	.ndo_fcoe_ddp_target = ixgbe_fcoe_ddp_target,
diff --git a/drivers/net/ethernet/jme.c b/drivers/net/ethernet/jme.c
index 070a6f1..7fbe6ab 100644
--- a/drivers/net/ethernet/jme.c
+++ b/drivers/net/ethernet/jme.c
@@ -3148,7 +3148,6 @@ jme_init_one(struct pci_dev *pdev,
 	jme->mii_if.mdio_write = jme_mdio_write;
 
 	jme_clear_pm(jme);
-	pci_set_power_state(jme->pdev, PCI_D0);
 	device_set_wakeup_enable(&pdev->dev, true);
 
 	jme_set_phyfifo_5level(jme);
diff --git a/drivers/net/ethernet/korina.c b/drivers/net/ethernet/korina.c
index 5409fe8..270e65f 100644
--- a/drivers/net/ethernet/korina.c
+++ b/drivers/net/ethernet/korina.c
@@ -483,7 +483,6 @@ static void korina_multicast_list(struct net_device *dev)
 	unsigned long flags;
 	struct netdev_hw_addr *ha;
 	u32 recognise = ETH_ARC_AB;	/* always accept broadcasts */
-	int i;
 
 	/* Set promiscuous mode */
 	if (dev->flags & IFF_PROMISC)
@@ -495,12 +494,9 @@ static void korina_multicast_list(struct net_device *dev)
 
 	/* Build the hash table */
 	if (netdev_mc_count(dev) > 4) {
-		u16 hash_table[4];
+		u16 hash_table[4] = { 0 };
 		u32 crc;
 
-		for (i = 0; i < 4; i++)
-			hash_table[i] = 0;
-
 		netdev_for_each_mc_addr(ha, dev) {
 			crc = ether_crc_le(6, ha->addr);
 			crc >>= 26;
@@ -1214,7 +1210,6 @@ static int korina_remove(struct platform_device *pdev)
 	iounmap(lp->rx_dma_regs);
 	iounmap(lp->tx_dma_regs);
 
-	platform_set_drvdata(pdev, NULL);
 	unregister_netdev(bif->dev);
 	free_netdev(bif->dev);
 
diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c
index d1cbfb1..c35db73 100644
--- a/drivers/net/ethernet/marvell/mv643xx_eth.c
+++ b/drivers/net/ethernet/marvell/mv643xx_eth.c
@@ -60,6 +60,10 @@
 #include <linux/types.h>
 #include <linux/slab.h>
 #include <linux/clk.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/of_net.h>
+#include <linux/of_mdio.h>
 
 static char mv643xx_eth_driver_name[] = "mv643xx_eth";
 static char mv643xx_eth_driver_version[] = "1.4";
@@ -115,6 +119,8 @@ static char mv643xx_eth_driver_version[] = "1.4";
 #define  LINK_UP			0x00000002
 #define TXQ_COMMAND			0x0048
 #define TXQ_FIX_PRIO_CONF		0x004c
+#define PORT_SERIAL_CONTROL1		0x004c
+#define  CLK125_BYPASS_EN		0x00000010
 #define TX_BW_RATE			0x0050
 #define TX_BW_MTU			0x0058
 #define TX_BW_BURST			0x005c
@@ -615,7 +621,7 @@ static int rxq_refill(struct rx_queue *rxq, int budget)
 
 		rx_desc = rxq->rx_desc_area + rx;
 
-		size = skb->end - skb->data;
+		size = skb_end_pointer(skb) - skb->data;
 		rx_desc->buf_ptr = dma_map_single(mp->dev->dev.parent,
 						  skb->data, size,
 						  DMA_FROM_DEVICE);
@@ -2450,13 +2456,159 @@ static void infer_hw_params(struct mv643xx_eth_shared_private *msp)
 	}
 }
 
+#if defined(CONFIG_OF)
+static const struct of_device_id mv643xx_eth_shared_ids[] = {
+	{ .compatible = "marvell,orion-eth", },
+	{ .compatible = "marvell,kirkwood-eth", },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, mv643xx_eth_shared_ids);
+#endif
+
+#if defined(CONFIG_OF) && !defined(CONFIG_MV64X60)
+#define mv643xx_eth_property(_np, _name, _v)				\
+	do {								\
+		u32 tmp;						\
+		if (!of_property_read_u32(_np, "marvell," _name, &tmp))	\
+			_v = tmp;					\
+	} while (0)
+
+static struct platform_device *port_platdev[3];
+
+static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev,
+					  struct device_node *pnp)
+{
+	struct platform_device *ppdev;
+	struct mv643xx_eth_platform_data ppd;
+	struct resource res;
+	const char *mac_addr;
+	int ret;
+	int dev_num = 0;
+
+	memset(&ppd, 0, sizeof(ppd));
+	ppd.shared = pdev;
+
+	memset(&res, 0, sizeof(res));
+	if (!of_irq_to_resource(pnp, 0, &res)) {
+		dev_err(&pdev->dev, "missing interrupt on %s\n", pnp->name);
+		return -EINVAL;
+	}
+
+	if (of_property_read_u32(pnp, "reg", &ppd.port_number)) {
+		dev_err(&pdev->dev, "missing reg property on %s\n", pnp->name);
+		return -EINVAL;
+	}
+
+	if (ppd.port_number >= 3) {
+		dev_err(&pdev->dev, "invalid reg property on %s\n", pnp->name);
+		return -EINVAL;
+	}
+
+	while (dev_num < 3 && port_platdev[dev_num])
+		dev_num++;
+
+	if (dev_num == 3) {
+		dev_err(&pdev->dev, "too many ports registered\n");
+		return -EINVAL;
+	}
+
+	mac_addr = of_get_mac_address(pnp);
+	if (mac_addr)
+		memcpy(ppd.mac_addr, mac_addr, 6);
+
+	mv643xx_eth_property(pnp, "tx-queue-size", ppd.tx_queue_size);
+	mv643xx_eth_property(pnp, "tx-sram-addr", ppd.tx_sram_addr);
+	mv643xx_eth_property(pnp, "tx-sram-size", ppd.tx_sram_size);
+	mv643xx_eth_property(pnp, "rx-queue-size", ppd.rx_queue_size);
+	mv643xx_eth_property(pnp, "rx-sram-addr", ppd.rx_sram_addr);
+	mv643xx_eth_property(pnp, "rx-sram-size", ppd.rx_sram_size);
+
+	ppd.phy_node = of_parse_phandle(pnp, "phy-handle", 0);
+	if (!ppd.phy_node) {
+		ppd.phy_addr = MV643XX_ETH_PHY_NONE;
+		of_property_read_u32(pnp, "speed", &ppd.speed);
+		of_property_read_u32(pnp, "duplex", &ppd.duplex);
+	}
+
+	ppdev = platform_device_alloc(MV643XX_ETH_NAME, dev_num);
+	if (!ppdev)
+		return -ENOMEM;
+	ppdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
+
+	ret = platform_device_add_resources(ppdev, &res, 1);
+	if (ret)
+		goto port_err;
+
+	ret = platform_device_add_data(ppdev, &ppd, sizeof(ppd));
+	if (ret)
+		goto port_err;
+
+	ret = platform_device_add(ppdev);
+	if (ret)
+		goto port_err;
+
+	port_platdev[dev_num] = ppdev;
+
+	return 0;
+
+port_err:
+	platform_device_put(ppdev);
+	return ret;
+}
+
+static int mv643xx_eth_shared_of_probe(struct platform_device *pdev)
+{
+	struct mv643xx_eth_shared_platform_data *pd;
+	struct device_node *pnp, *np = pdev->dev.of_node;
+	int ret;
+
+	/* bail out if not registered from DT */
+	if (!np)
+		return 0;
+
+	pd = devm_kzalloc(&pdev->dev, sizeof(*pd), GFP_KERNEL);
+	if (!pd)
+		return -ENOMEM;
+	pdev->dev.platform_data = pd;
+
+	mv643xx_eth_property(np, "tx-checksum-limit", pd->tx_csum_limit);
+
+	for_each_available_child_of_node(np, pnp) {
+		ret = mv643xx_eth_shared_of_add_port(pdev, pnp);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+static void mv643xx_eth_shared_of_remove(void)
+{
+	int n;
+
+	for (n = 0; n < 3; n++) {
+		platform_device_del(port_platdev[n]);
+		port_platdev[n] = NULL;
+	}
+}
+#else
+static inline int mv643xx_eth_shared_of_probe(struct platform_device *pdev)
+{
+	return 0;
+}
+
+static inline void mv643xx_eth_shared_of_remove(void)
+{
+}
+#endif
+
 static int mv643xx_eth_shared_probe(struct platform_device *pdev)
 {
 	static int mv643xx_eth_version_printed;
-	struct mv643xx_eth_shared_platform_data *pd = pdev->dev.platform_data;
+	struct mv643xx_eth_shared_platform_data *pd;
 	struct mv643xx_eth_shared_private *msp;
 	const struct mbus_dram_target_info *dram;
 	struct resource *res;
+	int ret;
 
 	if (!mv643xx_eth_version_printed++)
 		pr_notice("MV-643xx 10/100/1000 ethernet driver version %s\n",
@@ -2469,8 +2621,9 @@ static int mv643xx_eth_shared_probe(struct platform_device *pdev)
 	msp = devm_kzalloc(&pdev->dev, sizeof(*msp), GFP_KERNEL);
 	if (msp == NULL)
 		return -ENOMEM;
+	platform_set_drvdata(pdev, msp);
 
-	msp->base = ioremap(res->start, resource_size(res));
+	msp->base = devm_ioremap(&pdev->dev, res->start, resource_size(res));
 	if (msp->base == NULL)
 		return -ENOMEM;
 
@@ -2485,12 +2638,15 @@ static int mv643xx_eth_shared_probe(struct platform_device *pdev)
 	if (dram)
 		mv643xx_eth_conf_mbus_windows(msp, dram);
 
+	ret = mv643xx_eth_shared_of_probe(pdev);
+	if (ret)
+		return ret;
+	pd = pdev->dev.platform_data;
+
 	msp->tx_csum_limit = (pd != NULL && pd->tx_csum_limit) ?
 					pd->tx_csum_limit : 9 * 1024;
 	infer_hw_params(msp);
 
-	platform_set_drvdata(pdev, msp);
-
 	return 0;
 }
 
@@ -2498,10 +2654,9 @@ static int mv643xx_eth_shared_remove(struct platform_device *pdev)
 {
 	struct mv643xx_eth_shared_private *msp = platform_get_drvdata(pdev);
 
-	iounmap(msp->base);
+	mv643xx_eth_shared_of_remove();
 	if (!IS_ERR(msp->clk))
 		clk_disable_unprepare(msp->clk);
-
 	return 0;
 }
 
@@ -2511,6 +2666,7 @@ static struct platform_driver mv643xx_eth_shared_driver = {
 	.driver = {
 		.name	= MV643XX_ETH_SHARED_NAME,
 		.owner	= THIS_MODULE,
+		.of_match_table = of_match_ptr(mv643xx_eth_shared_ids),
 	},
 };
 
@@ -2701,6 +2857,15 @@ static int mv643xx_eth_probe(struct platform_device *pdev)
 
 	mp->dev = dev;
 
+	/* Kirkwood resets some registers on gated clocks. Especially
+	 * CLK125_BYPASS_EN must be cleared but is not available on
+	 * all other SoCs/System Controllers using this driver.
+	 */
+	if (of_device_is_compatible(pdev->dev.of_node,
+				    "marvell,kirkwood-eth-port"))
+		wrlp(mp, PORT_SERIAL_CONTROL1,
+		     rdlp(mp, PORT_SERIAL_CONTROL1) & ~CLK125_BYPASS_EN);
+
 	/*
 	 * Start with a default rate, and if there is a clock, allow
 	 * it to override the default.
@@ -2710,23 +2875,35 @@ static int mv643xx_eth_probe(struct platform_device *pdev)
 	if (!IS_ERR(mp->clk)) {
 		clk_prepare_enable(mp->clk);
 		mp->t_clk = clk_get_rate(mp->clk);
+	} else if (!IS_ERR(mp->shared->clk)) {
+		mp->t_clk = clk_get_rate(mp->shared->clk);
 	}
 
 	set_params(mp, pd);
 	netif_set_real_num_tx_queues(dev, mp->txq_count);
 	netif_set_real_num_rx_queues(dev, mp->rxq_count);
 
-	if (pd->phy_addr != MV643XX_ETH_PHY_NONE) {
+	err = 0;
+	if (pd->phy_node) {
+		mp->phy = of_phy_connect(mp->dev, pd->phy_node,
+					 mv643xx_eth_adjust_link, 0,
+					 PHY_INTERFACE_MODE_GMII);
+		if (!mp->phy)
+			err = -ENODEV;
+	} else if (pd->phy_addr != MV643XX_ETH_PHY_NONE) {
 		mp->phy = phy_scan(mp, pd->phy_addr);
 
-		if (IS_ERR(mp->phy)) {
+		if (IS_ERR(mp->phy))
 			err = PTR_ERR(mp->phy);
-			if (err == -ENODEV)
-				err = -EPROBE_DEFER;
-			goto out;
-		}
-		phy_init(mp, pd->speed, pd->duplex);
+		else
+			phy_init(mp, pd->speed, pd->duplex);
 	}
+	if (err == -ENODEV) {
+		err = -EPROBE_DEFER;
+		goto out;
+	}
+	if (err)
+		goto out;
 
 	SET_ETHTOOL_OPS(dev, &mv643xx_eth_ethtool_ops);
 
@@ -2805,7 +2982,7 @@ static int mv643xx_eth_remove(struct platform_device *pdev)
 
 	unregister_netdev(mp->dev);
 	if (mp->phy != NULL)
-		phy_detach(mp->phy);
+		phy_disconnect(mp->phy);
 	cancel_work_sync(&mp->tx_timeout_task);
 
 	if (!IS_ERR(mp->clk))
@@ -2813,8 +2990,6 @@ static int mv643xx_eth_remove(struct platform_device *pdev)
 
 	free_netdev(mp->dev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index c966785..712779f 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -2251,6 +2251,21 @@ static int mvneta_change_mtu(struct net_device *dev, int mtu)
 	return 0;
 }
 
+/* Get mac address */
+static void mvneta_get_mac_addr(struct mvneta_port *pp, unsigned char *addr)
+{
+	u32 mac_addr_l, mac_addr_h;
+
+	mac_addr_l = mvreg_read(pp, MVNETA_MAC_ADDR_LOW);
+	mac_addr_h = mvreg_read(pp, MVNETA_MAC_ADDR_HIGH);
+	addr[0] = (mac_addr_h >> 24) & 0xFF;
+	addr[1] = (mac_addr_h >> 16) & 0xFF;
+	addr[2] = (mac_addr_h >> 8) & 0xFF;
+	addr[3] = mac_addr_h & 0xFF;
+	addr[4] = (mac_addr_l >> 8) & 0xFF;
+	addr[5] = mac_addr_l & 0xFF;
+}
+
 /* Handle setting mac address */
 static int mvneta_set_mac_addr(struct net_device *dev, void *addr)
 {
@@ -2667,7 +2682,9 @@ static int mvneta_probe(struct platform_device *pdev)
 	u32 phy_addr;
 	struct mvneta_port *pp;
 	struct net_device *dev;
-	const char *mac_addr;
+	const char *dt_mac_addr;
+	char hw_mac_addr[ETH_ALEN];
+	const char *mac_from;
 	int phy_mode;
 	int err;
 
@@ -2703,13 +2720,6 @@ static int mvneta_probe(struct platform_device *pdev)
 		goto err_free_irq;
 	}
 
-	mac_addr = of_get_mac_address(dn);
-
-	if (!mac_addr || !is_valid_ether_addr(mac_addr))
-		eth_hw_addr_random(dev);
-	else
-		memcpy(dev->dev_addr, mac_addr, ETH_ALEN);
-
 	dev->tx_queue_len = MVNETA_MAX_TXD;
 	dev->watchdog_timeo = 5 * HZ;
 	dev->netdev_ops = &mvneta_netdev_ops;
@@ -2740,6 +2750,21 @@ static int mvneta_probe(struct platform_device *pdev)
 
 	clk_prepare_enable(pp->clk);
 
+	dt_mac_addr = of_get_mac_address(dn);
+	if (dt_mac_addr && is_valid_ether_addr(dt_mac_addr)) {
+		mac_from = "device tree";
+		memcpy(dev->dev_addr, dt_mac_addr, ETH_ALEN);
+	} else {
+		mvneta_get_mac_addr(pp, hw_mac_addr);
+		if (is_valid_ether_addr(hw_mac_addr)) {
+			mac_from = "hardware";
+			memcpy(dev->dev_addr, hw_mac_addr, ETH_ALEN);
+		} else {
+			mac_from = "random";
+			eth_hw_addr_random(dev);
+		}
+	}
+
 	pp->tx_done_timer.data = (unsigned long)dev;
 
 	pp->tx_ring_size = MVNETA_MAX_TXD;
@@ -2772,7 +2797,8 @@ static int mvneta_probe(struct platform_device *pdev)
 		goto err_deinit;
 	}
 
-	netdev_info(dev, "mac: %pM\n", dev->dev_addr);
+	netdev_info(dev, "Using %s mac address %pM\n", mac_from,
+		    dev->dev_addr);
 
 	platform_set_drvdata(pdev, pp->dev);
 
@@ -2804,8 +2830,6 @@ static int mvneta_remove(struct platform_device *pdev)
 	irq_dispose_mapping(dev->irq);
 	free_netdev(dev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/marvell/pxa168_eth.c b/drivers/net/ethernet/marvell/pxa168_eth.c
index 1c8af8b..db48147 100644
--- a/drivers/net/ethernet/marvell/pxa168_eth.c
+++ b/drivers/net/ethernet/marvell/pxa168_eth.c
@@ -357,7 +357,7 @@ static void rxq_refill(struct net_device *dev)
 		/* Get 'used' Rx descriptor */
 		used_rx_desc = pep->rx_used_desc_q;
 		p_used_rx_desc = &pep->p_rx_desc_area[used_rx_desc];
-		size = skb->end - skb->data;
+		size = skb_end_pointer(skb) - skb->data;
 		p_used_rx_desc->buf_ptr = dma_map_single(NULL,
 							 skb->data,
 							 size,
@@ -1602,7 +1602,6 @@ static int pxa168_eth_remove(struct platform_device *pdev)
 	unregister_netdev(dev);
 	cancel_work_sync(&pep->tx_timeout_task);
 	free_netdev(dev);
-	platform_set_drvdata(pdev, NULL);
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/marvell/skge.c b/drivers/net/ethernet/marvell/skge.c
index 171f4b3..c896079 100644
--- a/drivers/net/ethernet/marvell/skge.c
+++ b/drivers/net/ethernet/marvell/skge.c
@@ -3706,7 +3706,7 @@ static const struct file_operations skge_debug_fops = {
 static int skge_device_event(struct notifier_block *unused,
 			     unsigned long event, void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct skge_port *skge;
 	struct dentry *d;
 
diff --git a/drivers/net/ethernet/marvell/sky2.c b/drivers/net/ethernet/marvell/sky2.c
index d175bbd..e09a8c6 100644
--- a/drivers/net/ethernet/marvell/sky2.c
+++ b/drivers/net/ethernet/marvell/sky2.c
@@ -4642,7 +4642,7 @@ static const struct file_operations sky2_debug_fops = {
 static int sky2_device_event(struct notifier_block *unused,
 			     unsigned long event, void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct sky2_port *sky2 = netdev_priv(dev);
 
 	if (dev->netdev_ops->ndo_open != sky2_open || !sky2_debug)
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 0e572a5..299d018 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -39,6 +39,7 @@
 #include <linux/errno.h>
 
 #include <linux/mlx4/cmd.h>
+#include <linux/mlx4/device.h>
 #include <linux/semaphore.h>
 #include <rdma/ib_smi.h>
 
@@ -111,6 +112,14 @@ enum {
 	GO_BIT_TIMEOUT_MSECS	= 10000
 };
 
+enum mlx4_vlan_transition {
+	MLX4_VLAN_TRANSITION_VST_VST = 0,
+	MLX4_VLAN_TRANSITION_VST_VGT = 1,
+	MLX4_VLAN_TRANSITION_VGT_VST = 2,
+	MLX4_VLAN_TRANSITION_VGT_VGT = 3,
+};
+
+
 struct mlx4_cmd_context {
 	struct completion	done;
 	int			result;
@@ -256,6 +265,8 @@ static int mlx4_comm_cmd_wait(struct mlx4_dev *dev, u8 op,
 
 	if (!wait_for_completion_timeout(&context->done,
 					 msecs_to_jiffies(timeout))) {
+		mlx4_warn(dev, "communication channel command 0x%x timed out\n",
+			  op);
 		err = -EBUSY;
 		goto out;
 	}
@@ -485,6 +496,8 @@ static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
 	}
 
 	if (cmd_pending(dev)) {
+		mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
+			  op);
 		err = -ETIMEDOUT;
 		goto out;
 	}
@@ -548,6 +561,8 @@ static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
 
 	if (!wait_for_completion_timeout(&context->done,
 					 msecs_to_jiffies(timeout))) {
+		mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
+			  op);
 		err = -EBUSY;
 		goto out;
 	}
@@ -785,6 +800,15 @@ static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
 				    vhcr->op, MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
 }
 
+int MLX4_CMD_UPDATE_QP_wrapper(struct mlx4_dev *dev, int slave,
+		     struct mlx4_vhcr *vhcr,
+		     struct mlx4_cmd_mailbox *inbox,
+		     struct mlx4_cmd_mailbox *outbox,
+		     struct mlx4_cmd_info *cmd)
+{
+	return -EPERM;
+}
+
 int mlx4_DMA_wrapper(struct mlx4_dev *dev, int slave,
 		     struct mlx4_vhcr *vhcr,
 		     struct mlx4_cmd_mailbox *inbox,
@@ -1219,6 +1243,15 @@ static struct mlx4_cmd_info cmd_info[] = {
 		.wrapper = mlx4_GEN_QP_wrapper
 	},
 	{
+		.opcode = MLX4_CMD_UPDATE_QP,
+		.has_inbox = false,
+		.has_outbox = false,
+		.out_is_imm = false,
+		.encode_slave_id = false,
+		.verify = NULL,
+		.wrapper = MLX4_CMD_UPDATE_QP_wrapper
+	},
+	{
 		.opcode = MLX4_CMD_CONF_SPECIAL_QP,
 		.has_inbox = false,
 		.has_outbox = false,
@@ -1488,6 +1521,102 @@ out:
 	return ret;
 }
 
+static int calculate_transition(u16 oper_vlan, u16 admin_vlan)
+{
+	return (2 * (oper_vlan == MLX4_VGT) + (admin_vlan == MLX4_VGT));
+}
+
+int mlx4_master_immediate_activate_vlan_qos(struct mlx4_priv *priv,
+					    int slave, int port)
+{
+	struct mlx4_vport_oper_state *vp_oper;
+	struct mlx4_vport_state *vp_admin;
+	struct mlx4_vf_immed_vlan_work *work;
+	struct mlx4_dev *dev = &(priv->dev);
+	int err;
+	int admin_vlan_ix = NO_INDX;
+	enum mlx4_vlan_transition vlan_trans;
+
+	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+	vp_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+
+	if (vp_oper->state.default_vlan == vp_admin->default_vlan &&
+	    vp_oper->state.default_qos == vp_admin->default_qos &&
+	    vp_oper->state.link_state == vp_admin->link_state)
+		return 0;
+
+	vlan_trans = calculate_transition(vp_oper->state.default_vlan,
+					  vp_admin->default_vlan);
+
+	if (!(priv->mfunc.master.slave_state[slave].active &&
+	      dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP &&
+	      vlan_trans == MLX4_VLAN_TRANSITION_VST_VST)) {
+		/* even if the UPDATE_QP command isn't supported, we still want
+		 * to set this VF link according to the admin directive
+		 */
+		vp_oper->state.link_state = vp_admin->link_state;
+		return -1;
+	}
+
+	mlx4_dbg(dev, "updating immediately admin params slave %d port %d\n",
+		 slave, port);
+	mlx4_dbg(dev, "vlan %d QoS %d link down %d\n", vp_admin->default_vlan,
+		 vp_admin->default_qos, vp_admin->link_state);
+
+	work = kzalloc(sizeof(*work), GFP_KERNEL);
+	if (!work)
+		return -ENOMEM;
+
+	if (vp_oper->state.default_vlan != vp_admin->default_vlan) {
+		err = __mlx4_register_vlan(&priv->dev, port,
+					   vp_admin->default_vlan,
+					   &admin_vlan_ix);
+		if (err) {
+			kfree(work);
+			mlx4_warn((&priv->dev),
+				  "No vlan resources slave %d, port %d\n",
+				  slave, port);
+			return err;
+		}
+		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_VLAN;
+		mlx4_dbg((&(priv->dev)),
+			 "alloc vlan %d idx  %d slave %d port %d\n",
+			 (int)(vp_admin->default_vlan),
+			 admin_vlan_ix, slave, port);
+	}
+
+	/* save original vlan ix and vlan id */
+	work->orig_vlan_id = vp_oper->state.default_vlan;
+	work->orig_vlan_ix = vp_oper->vlan_idx;
+
+	/* handle new qos */
+	if (vp_oper->state.default_qos != vp_admin->default_qos)
+		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_QOS;
+
+	if (work->flags & MLX4_VF_IMMED_VLAN_FLAG_VLAN)
+		vp_oper->vlan_idx = admin_vlan_ix;
+
+	vp_oper->state.default_vlan = vp_admin->default_vlan;
+	vp_oper->state.default_qos = vp_admin->default_qos;
+	vp_oper->state.link_state = vp_admin->link_state;
+
+	if (vp_admin->link_state == IFLA_VF_LINK_STATE_DISABLE)
+		work->flags |= MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE;
+
+	/* iterate over QPs owned by this slave, using UPDATE_QP */
+	work->port = port;
+	work->slave = slave;
+	work->qos = vp_oper->state.default_qos;
+	work->vlan_id = vp_oper->state.default_vlan;
+	work->vlan_ix = vp_oper->vlan_idx;
+	work->priv = priv;
+	INIT_WORK(&work->work, mlx4_vf_immed_vlan_work_handler);
+	queue_work(priv->mfunc.master.comm_wq, &work->work);
+
+	return 0;
+}
+
+
 static int mlx4_master_activate_admin_state(struct mlx4_priv *priv, int slave)
 {
 	int port, err;
@@ -2102,10 +2231,12 @@ int mlx4_set_vf_mac(struct mlx4_dev *dev, int port, int vf, u64 mac)
 }
 EXPORT_SYMBOL_GPL(mlx4_set_vf_mac);
 
+
 int mlx4_set_vf_vlan(struct mlx4_dev *dev, int port, int vf, u16 vlan, u8 qos)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
-	struct mlx4_vport_state *s_info;
+	struct mlx4_vport_oper_state *vf_oper;
+	struct mlx4_vport_state *vf_admin;
 	int slave;
 
 	if ((!mlx4_is_master(dev)) ||
@@ -2119,12 +2250,19 @@ int mlx4_set_vf_vlan(struct mlx4_dev *dev, int port, int vf, u16 vlan, u8 qos)
 	if (slave < 0)
 		return -EINVAL;
 
-	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
+	vf_admin = &priv->mfunc.master.vf_admin[slave].vport[port];
+	vf_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
+
 	if ((0 == vlan) && (0 == qos))
-		s_info->default_vlan = MLX4_VGT;
+		vf_admin->default_vlan = MLX4_VGT;
 	else
-		s_info->default_vlan = vlan;
-	s_info->default_qos = qos;
+		vf_admin->default_vlan = vlan;
+	vf_admin->default_qos = qos;
+
+	if (mlx4_master_immediate_activate_vlan_qos(priv, slave, port))
+		mlx4_info(dev,
+			  "updating vf %d port %d config will take effect on next VF restart\n",
+			  vf, port);
 	return 0;
 }
 EXPORT_SYMBOL_GPL(mlx4_set_vf_vlan);
@@ -2178,7 +2316,55 @@ int mlx4_get_vf_config(struct mlx4_dev *dev, int port, int vf, struct ifla_vf_in
 	ivf->qos	= s_info->default_qos;
 	ivf->tx_rate	= s_info->tx_rate;
 	ivf->spoofchk	= s_info->spoofchk;
+	ivf->linkstate	= s_info->link_state;
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(mlx4_get_vf_config);
+
+int mlx4_set_vf_link_state(struct mlx4_dev *dev, int port, int vf, int link_state)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_vport_state *s_info;
+	int slave;
+	u8 link_stat_event;
+
+	slave = mlx4_get_slave_indx(dev, vf);
+	if (slave < 0)
+		return -EINVAL;
+
+	switch (link_state) {
+	case IFLA_VF_LINK_STATE_AUTO:
+		/* get current link state */
+		if (!priv->sense.do_sense_port[port])
+			link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_ACTIVE;
+		else
+			link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_DOWN;
+	    break;
+
+	case IFLA_VF_LINK_STATE_ENABLE:
+		link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_ACTIVE;
+	    break;
+
+	case IFLA_VF_LINK_STATE_DISABLE:
+		link_stat_event = MLX4_PORT_CHANGE_SUBTYPE_DOWN;
+	    break;
+
+	default:
+		mlx4_warn(dev, "unknown value for link_state %02x on slave %d port %d\n",
+			  link_state, slave, port);
+		return -EINVAL;
+	};
+	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
+	s_info->link_state = link_state;
+
+	/* send event */
+	mlx4_gen_port_state_change_eqe(dev, slave, port, link_stat_event);
+
+	if (mlx4_master_immediate_activate_vlan_qos(priv, slave, port))
+		mlx4_dbg(dev,
+			 "updating vf %d port %d no link state HW enforcment\n",
+			 vf, port);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_set_vf_link_state);
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_cq.c b/drivers/net/ethernet/mellanox/mlx4/en_cq.c
index 1e6c594..3e2d504 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_cq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_cq.c
@@ -139,6 +139,7 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
 
 	if (!cq->is_tx) {
 		netif_napi_add(cq->dev, &cq->napi, mlx4_en_poll_rx_cq, 64);
+		napi_hash_add(&cq->napi);
 		napi_enable(&cq->napi);
 	}
 
@@ -162,6 +163,8 @@ void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
 {
 	if (!cq->is_tx) {
 		napi_disable(&cq->napi);
+		napi_hash_del(&cq->napi);
+		synchronize_rcu();
 		netif_napi_del(&cq->napi);
 	}
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
index 0f91222..9d4a1ea 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_dcb_nl.c
@@ -207,9 +207,6 @@ static int mlx4_en_dcbnl_ieee_getmaxrate(struct net_device *dev,
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 	int i;
 
-	if (!priv->maxrate)
-		return -EINVAL;
-
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++)
 		maxrate->tc_maxrate[i] =
 			priv->maxrate[i] * MLX4_RATELIMIT_UNITS_IN_KB;
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
index c9e6b62..727874f 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
@@ -222,7 +222,12 @@ static int mlx4_en_get_sset_count(struct net_device *dev, int sset)
 	switch (sset) {
 	case ETH_SS_STATS:
 		return (priv->stats_bitmap ? bit_count : NUM_ALL_STATS) +
-			(priv->tx_ring_num + priv->rx_ring_num) * 2;
+			(priv->tx_ring_num * 2) +
+#ifdef CONFIG_NET_LL_RX_POLL
+			(priv->rx_ring_num * 5);
+#else
+			(priv->rx_ring_num * 2);
+#endif
 	case ETH_SS_TEST:
 		return MLX4_EN_NUM_SELF_TEST - !(priv->mdev->dev->caps.flags
 					& MLX4_DEV_CAP_FLAG_UC_LOOPBACK) * 2;
@@ -271,6 +276,11 @@ static void mlx4_en_get_ethtool_stats(struct net_device *dev,
 	for (i = 0; i < priv->rx_ring_num; i++) {
 		data[index++] = priv->rx_ring[i].packets;
 		data[index++] = priv->rx_ring[i].bytes;
+#ifdef CONFIG_NET_LL_RX_POLL
+		data[index++] = priv->rx_ring[i].yields;
+		data[index++] = priv->rx_ring[i].misses;
+		data[index++] = priv->rx_ring[i].cleaned;
+#endif
 	}
 	spin_unlock_bh(&priv->stats_lock);
 
@@ -334,6 +344,14 @@ static void mlx4_en_get_strings(struct net_device *dev,
 				"rx%d_packets", i);
 			sprintf(data + (index++) * ETH_GSTRING_LEN,
 				"rx%d_bytes", i);
+#ifdef CONFIG_NET_LL_RX_POLL
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_napi_yield", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_misses", i);
+			sprintf(data + (index++) * ETH_GSTRING_LEN,
+				"rx%d_cleaned", i);
+#endif
 		}
 		break;
 	}
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_main.c b/drivers/net/ethernet/mellanox/mlx4/en_main.c
index a5c9df07..a071cda 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_main.c
@@ -310,7 +310,7 @@ static void *mlx4_en_add(struct mlx4_dev *dev)
 err_mr:
 	(void) mlx4_mr_free(dev, &mdev->mr);
 err_map:
-	if (!mdev->uar_map)
+	if (mdev->uar_map)
 		iounmap(mdev->uar_map);
 err_uar:
 	mlx4_uar_free(dev, &mdev->priv_uar);
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 89c47ea..caf2047 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -38,6 +38,7 @@
 #include <linux/slab.h>
 #include <linux/hash.h>
 #include <net/ip.h>
+#include <net/ll_poll.h>
 
 #include <linux/mlx4/driver.h>
 #include <linux/mlx4/device.h>
@@ -67,6 +68,34 @@ int mlx4_en_setup_tc(struct net_device *dev, u8 up)
 	return 0;
 }
 
+#ifdef CONFIG_NET_LL_RX_POLL
+/* must be called with local_bh_disable()d */
+static int mlx4_en_low_latency_recv(struct napi_struct *napi)
+{
+	struct mlx4_en_cq *cq = container_of(napi, struct mlx4_en_cq, napi);
+	struct net_device *dev = cq->dev;
+	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_rx_ring *rx_ring = &priv->rx_ring[cq->ring];
+	int done;
+
+	if (!priv->port_up)
+		return LL_FLUSH_FAILED;
+
+	if (!mlx4_en_cq_lock_poll(cq))
+		return LL_FLUSH_BUSY;
+
+	done = mlx4_en_process_rx_cq(dev, cq, 4);
+	if (likely(done))
+		rx_ring->cleaned += done;
+	else
+		rx_ring->misses++;
+
+	mlx4_en_cq_unlock_poll(cq);
+
+	return done;
+}
+#endif	/* CONFIG_NET_LL_RX_POLL */
+
 #ifdef CONFIG_RFS_ACCEL
 
 struct mlx4_en_filter {
@@ -376,7 +405,7 @@ static int mlx4_en_vlan_rx_add_vid(struct net_device *dev,
 			en_err(priv, "Failed configuring VLAN filter\n");
 	}
 	if (mlx4_register_vlan(mdev->dev, priv->port, vid, &idx))
-		en_err(priv, "failed adding vlan %d\n", vid);
+		en_dbg(HW, priv, "failed adding vlan %d\n", vid);
 	mutex_unlock(&mdev->state_lock);
 
 	return 0;
@@ -399,7 +428,7 @@ static int mlx4_en_vlan_rx_kill_vid(struct net_device *dev,
 	if (!mlx4_find_cached_vlan(mdev->dev, priv->port, vid, &idx))
 		mlx4_unregister_vlan(mdev->dev, priv->port, idx);
 	else
-		en_err(priv, "could not find vid %d in cache\n", vid);
+		en_dbg(HW, priv, "could not find vid %d in cache\n", vid);
 
 	if (mdev->device_up && priv->port_up) {
 		err = mlx4_SET_VLAN_FLTR(mdev->dev, priv);
@@ -1207,10 +1236,19 @@ static void mlx4_en_tx_timeout(struct net_device *dev)
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 	struct mlx4_en_dev *mdev = priv->mdev;
+	int i;
 
 	if (netif_msg_timer(priv))
 		en_warn(priv, "Tx timeout called on port:%d\n", priv->port);
 
+	for (i = 0; i < priv->tx_ring_num; i++) {
+		if (!netif_tx_queue_stopped(netdev_get_tx_queue(dev, i)))
+			continue;
+		en_warn(priv, "TX timeout on queue: %d, QP: 0x%x, CQ: 0x%x, Cons: 0x%x, Prod: 0x%x\n",
+			i, priv->tx_ring[i].qpn, priv->tx_ring[i].cqn,
+			priv->tx_ring[i].cons, priv->tx_ring[i].prod);
+	}
+
 	priv->port_stats.tx_timeout++;
 	en_dbg(DRV, priv, "Scheduling watchdog\n");
 	queue_work(mdev->workqueue, &priv->watchdog_task);
@@ -1346,12 +1384,13 @@ static void mlx4_en_do_get_stats(struct work_struct *work)
 
 	mutex_lock(&mdev->state_lock);
 	if (mdev->device_up) {
-		err = mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 0);
-		if (err)
-			en_dbg(HW, priv, "Could not update stats\n");
+		if (priv->port_up) {
+			err = mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 0);
+			if (err)
+				en_dbg(HW, priv, "Could not update stats\n");
 
-		if (priv->port_up)
 			mlx4_en_auto_moderation(priv);
+		}
 
 		queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY);
 	}
@@ -1445,6 +1484,8 @@ int mlx4_en_start_port(struct net_device *dev)
 	for (i = 0; i < priv->rx_ring_num; i++) {
 		cq = &priv->rx_cq[i];
 
+		mlx4_en_cq_init_lock(cq);
+
 		err = mlx4_en_activate_cq(priv, cq, i);
 		if (err) {
 			en_err(priv, "Failed activating Rx CQ\n");
@@ -1603,6 +1644,9 @@ void mlx4_en_stop_port(struct net_device *dev, int detach)
 		return;
 	}
 
+	/* close port*/
+	mlx4_CLOSE_PORT(mdev->dev, priv->port);
+
 	/* Synchronize with tx routine */
 	netif_tx_lock_bh(dev);
 	if (detach)
@@ -1694,14 +1738,20 @@ void mlx4_en_stop_port(struct net_device *dev, int detach)
 
 	/* Free RX Rings */
 	for (i = 0; i < priv->rx_ring_num; i++) {
-		mlx4_en_deactivate_rx_ring(priv, &priv->rx_ring[i]);
-		while (test_bit(NAPI_STATE_SCHED, &priv->rx_cq[i].napi.state))
+		struct mlx4_en_cq *cq = &priv->rx_cq[i];
+
+		local_bh_disable();
+		while (!mlx4_en_cq_lock_napi(cq)) {
+			pr_info("CQ %d locked\n", i);
+			mdelay(1);
+		}
+		local_bh_enable();
+
+		while (test_bit(NAPI_STATE_SCHED, &cq->napi.state))
 			msleep(1);
-		mlx4_en_deactivate_cq(priv, &priv->rx_cq[i]);
+		mlx4_en_deactivate_rx_ring(priv, &priv->rx_ring[i]);
+		mlx4_en_deactivate_cq(priv, cq);
 	}
-
-	/* close port*/
-	mlx4_CLOSE_PORT(mdev->dev, priv->port);
 }
 
 static void mlx4_en_restart(struct work_struct *work)
@@ -2061,6 +2111,13 @@ static int mlx4_en_get_vf_config(struct net_device *dev, int vf, struct ifla_vf_
 	return mlx4_get_vf_config(mdev->dev, en_priv->port, vf, ivf);
 }
 
+static int mlx4_en_set_vf_link_state(struct net_device *dev, int vf, int link_state)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_set_vf_link_state(mdev->dev, en_priv->port, vf, link_state);
+}
 static const struct net_device_ops mlx4_netdev_ops = {
 	.ndo_open		= mlx4_en_open,
 	.ndo_stop		= mlx4_en_close,
@@ -2083,6 +2140,9 @@ static const struct net_device_ops mlx4_netdev_ops = {
 #ifdef CONFIG_RFS_ACCEL
 	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
 #endif
+#ifdef CONFIG_NET_LL_RX_POLL
+	.ndo_ll_poll		= mlx4_en_low_latency_recv,
+#endif
 };
 
 static const struct net_device_ops mlx4_netdev_ops_master = {
@@ -2101,6 +2161,7 @@ static const struct net_device_ops mlx4_netdev_ops_master = {
 	.ndo_set_vf_mac		= mlx4_en_set_vf_mac,
 	.ndo_set_vf_vlan	= mlx4_en_set_vf_vlan,
 	.ndo_set_vf_spoofchk	= mlx4_en_set_vf_spoofchk,
+	.ndo_set_vf_link_state	= mlx4_en_set_vf_link_state,
 	.ndo_get_vf_config	= mlx4_en_get_vf_config,
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= mlx4_en_netpoll,
@@ -2271,6 +2332,8 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	mdev->pndev[port] = dev;
 
 	netif_carrier_off(dev);
+	mlx4_en_set_default_moderation(priv);
+
 	err = register_netdev(dev);
 	if (err) {
 		en_err(priv, "Netdev registration failed for port %d\n", port);
@@ -2302,7 +2365,6 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 		en_err(priv, "Failed Initializing port\n");
 		goto out;
 	}
-	mlx4_en_set_default_moderation(priv);
 	queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY);
 
 	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index 02aee1e..76997b9 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -31,6 +31,7 @@
  *
  */
 
+#include <net/ll_poll.h>
 #include <linux/mlx4/cq.h>
 #include <linux/slab.h>
 #include <linux/mlx4/qp.h>
@@ -42,40 +43,64 @@
 
 #include "mlx4_en.h"
 
+static int mlx4_alloc_pages(struct mlx4_en_priv *priv,
+			    struct mlx4_en_rx_alloc *page_alloc,
+			    const struct mlx4_en_frag_info *frag_info,
+			    gfp_t _gfp)
+{
+	int order;
+	struct page *page;
+	dma_addr_t dma;
+
+	for (order = MLX4_EN_ALLOC_PREFER_ORDER; ;) {
+		gfp_t gfp = _gfp;
+
+		if (order)
+			gfp |= __GFP_COMP | __GFP_NOWARN;
+		page = alloc_pages(gfp, order);
+		if (likely(page))
+			break;
+		if (--order < 0 ||
+		    ((PAGE_SIZE << order) < frag_info->frag_size))
+			return -ENOMEM;
+	}
+	dma = dma_map_page(priv->ddev, page, 0, PAGE_SIZE << order,
+			   PCI_DMA_FROMDEVICE);
+	if (dma_mapping_error(priv->ddev, dma)) {
+		put_page(page);
+		return -ENOMEM;
+	}
+	page_alloc->size = PAGE_SIZE << order;
+	page_alloc->page = page;
+	page_alloc->dma = dma;
+	page_alloc->offset = frag_info->frag_align;
+	/* Not doing get_page() for each frag is a big win
+	 * on asymetric workloads.
+	 */
+	atomic_set(&page->_count, page_alloc->size / frag_info->frag_stride);
+	return 0;
+}
+
 static int mlx4_en_alloc_frags(struct mlx4_en_priv *priv,
 			       struct mlx4_en_rx_desc *rx_desc,
 			       struct mlx4_en_rx_alloc *frags,
-			       struct mlx4_en_rx_alloc *ring_alloc)
+			       struct mlx4_en_rx_alloc *ring_alloc,
+			       gfp_t gfp)
 {
 	struct mlx4_en_rx_alloc page_alloc[MLX4_EN_MAX_RX_FRAGS];
-	struct mlx4_en_frag_info *frag_info;
+	const struct mlx4_en_frag_info *frag_info;
 	struct page *page;
 	dma_addr_t dma;
 	int i;
 
 	for (i = 0; i < priv->num_frags; i++) {
 		frag_info = &priv->frag_info[i];
-		if (ring_alloc[i].offset == frag_info->last_offset) {
-			page = alloc_pages(GFP_ATOMIC | __GFP_COMP,
-					MLX4_EN_ALLOC_ORDER);
-			if (!page)
-				goto out;
-			dma = dma_map_page(priv->ddev, page, 0,
-				MLX4_EN_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
-			if (dma_mapping_error(priv->ddev, dma)) {
-				put_page(page);
-				goto out;
-			}
-			page_alloc[i].page = page;
-			page_alloc[i].dma = dma;
-			page_alloc[i].offset = frag_info->frag_align;
-		} else {
-			page_alloc[i].page = ring_alloc[i].page;
-			get_page(ring_alloc[i].page);
-			page_alloc[i].dma = ring_alloc[i].dma;
-			page_alloc[i].offset = ring_alloc[i].offset +
-						frag_info->frag_stride;
-		}
+		page_alloc[i] = ring_alloc[i];
+		page_alloc[i].offset += frag_info->frag_stride;
+		if (page_alloc[i].offset + frag_info->frag_stride <= ring_alloc[i].size)
+			continue;
+		if (mlx4_alloc_pages(priv, &page_alloc[i], frag_info, gfp))
+			goto out;
 	}
 
 	for (i = 0; i < priv->num_frags; i++) {
@@ -87,14 +112,16 @@ static int mlx4_en_alloc_frags(struct mlx4_en_priv *priv,
 
 	return 0;
 
-
 out:
 	while (i--) {
 		frag_info = &priv->frag_info[i];
-		if (ring_alloc[i].offset == frag_info->last_offset)
+		if (page_alloc[i].page != ring_alloc[i].page) {
 			dma_unmap_page(priv->ddev, page_alloc[i].dma,
-				MLX4_EN_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
-		put_page(page_alloc[i].page);
+				page_alloc[i].size, PCI_DMA_FROMDEVICE);
+			page = page_alloc[i].page;
+			atomic_set(&page->_count, 1);
+			put_page(page);
+		}
 	}
 	return -ENOMEM;
 }
@@ -103,12 +130,12 @@ static void mlx4_en_free_frag(struct mlx4_en_priv *priv,
 			      struct mlx4_en_rx_alloc *frags,
 			      int i)
 {
-	struct mlx4_en_frag_info *frag_info = &priv->frag_info[i];
+	const struct mlx4_en_frag_info *frag_info = &priv->frag_info[i];
 
-	if (frags[i].offset == frag_info->last_offset) {
-		dma_unmap_page(priv->ddev, frags[i].dma, MLX4_EN_ALLOC_SIZE,
+	if (frags[i].offset + frag_info->frag_stride > frags[i].size)
+		dma_unmap_page(priv->ddev, frags[i].dma, frags[i].size,
 					 PCI_DMA_FROMDEVICE);
-	}
+
 	if (frags[i].page)
 		put_page(frags[i].page);
 }
@@ -116,35 +143,28 @@ static void mlx4_en_free_frag(struct mlx4_en_priv *priv,
 static int mlx4_en_init_allocator(struct mlx4_en_priv *priv,
 				  struct mlx4_en_rx_ring *ring)
 {
-	struct mlx4_en_rx_alloc *page_alloc;
 	int i;
+	struct mlx4_en_rx_alloc *page_alloc;
 
 	for (i = 0; i < priv->num_frags; i++) {
-		page_alloc = &ring->page_alloc[i];
-		page_alloc->page = alloc_pages(GFP_ATOMIC | __GFP_COMP,
-					       MLX4_EN_ALLOC_ORDER);
-		if (!page_alloc->page)
-			goto out;
+		const struct mlx4_en_frag_info *frag_info = &priv->frag_info[i];
 
-		page_alloc->dma = dma_map_page(priv->ddev, page_alloc->page, 0,
-					MLX4_EN_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
-		if (dma_mapping_error(priv->ddev, page_alloc->dma)) {
-			put_page(page_alloc->page);
-			page_alloc->page = NULL;
+		if (mlx4_alloc_pages(priv, &ring->page_alloc[i],
+				     frag_info, GFP_KERNEL))
 			goto out;
-		}
-		page_alloc->offset = priv->frag_info[i].frag_align;
-		en_dbg(DRV, priv, "Initialized allocator:%d with page:%p\n",
-		       i, page_alloc->page);
 	}
 	return 0;
 
 out:
 	while (i--) {
+		struct page *page;
+
 		page_alloc = &ring->page_alloc[i];
 		dma_unmap_page(priv->ddev, page_alloc->dma,
-				MLX4_EN_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
-		put_page(page_alloc->page);
+			       page_alloc->size, PCI_DMA_FROMDEVICE);
+		page = page_alloc->page;
+		atomic_set(&page->_count, 1);
+		put_page(page);
 		page_alloc->page = NULL;
 	}
 	return -ENOMEM;
@@ -157,13 +177,18 @@ static void mlx4_en_destroy_allocator(struct mlx4_en_priv *priv,
 	int i;
 
 	for (i = 0; i < priv->num_frags; i++) {
+		const struct mlx4_en_frag_info *frag_info = &priv->frag_info[i];
+
 		page_alloc = &ring->page_alloc[i];
 		en_dbg(DRV, priv, "Freeing allocator:%d count:%d\n",
 		       i, page_count(page_alloc->page));
 
 		dma_unmap_page(priv->ddev, page_alloc->dma,
-				MLX4_EN_ALLOC_SIZE, PCI_DMA_FROMDEVICE);
-		put_page(page_alloc->page);
+				page_alloc->size, PCI_DMA_FROMDEVICE);
+		while (page_alloc->offset + frag_info->frag_stride < page_alloc->size) {
+			put_page(page_alloc->page);
+			page_alloc->offset += frag_info->frag_stride;
+		}
 		page_alloc->page = NULL;
 	}
 }
@@ -194,13 +219,14 @@ static void mlx4_en_init_rx_desc(struct mlx4_en_priv *priv,
 }
 
 static int mlx4_en_prepare_rx_desc(struct mlx4_en_priv *priv,
-				   struct mlx4_en_rx_ring *ring, int index)
+				   struct mlx4_en_rx_ring *ring, int index,
+				   gfp_t gfp)
 {
 	struct mlx4_en_rx_desc *rx_desc = ring->buf + (index * ring->stride);
 	struct mlx4_en_rx_alloc *frags = ring->rx_info +
 					(index << priv->log_rx_info);
 
-	return mlx4_en_alloc_frags(priv, rx_desc, frags, ring->page_alloc);
+	return mlx4_en_alloc_frags(priv, rx_desc, frags, ring->page_alloc, gfp);
 }
 
 static inline void mlx4_en_update_rx_prod_db(struct mlx4_en_rx_ring *ring)
@@ -234,7 +260,8 @@ static int mlx4_en_fill_rx_buffers(struct mlx4_en_priv *priv)
 			ring = &priv->rx_ring[ring_ind];
 
 			if (mlx4_en_prepare_rx_desc(priv, ring,
-						    ring->actual_size)) {
+						    ring->actual_size,
+						    GFP_KERNEL)) {
 				if (ring->actual_size < MLX4_EN_MIN_RX_SIZE) {
 					en_err(priv, "Failed to allocate "
 						     "enough rx buffers\n");
@@ -449,11 +476,11 @@ static int mlx4_en_complete_rx_desc(struct mlx4_en_priv *priv,
 					DMA_FROM_DEVICE);
 
 		/* Save page reference in skb */
-		get_page(frags[nr].page);
 		__skb_frag_set_page(&skb_frags_rx[nr], frags[nr].page);
 		skb_frag_size_set(&skb_frags_rx[nr], frag_info->frag_size);
 		skb_frags_rx[nr].page_offset = frags[nr].offset;
 		skb->truesize += frag_info->frag_stride;
+		frags[nr].page = NULL;
 	}
 	/* Adjust size of last fragment to match actual length */
 	if (nr > 0)
@@ -546,7 +573,7 @@ static void mlx4_en_refill_rx_buffers(struct mlx4_en_priv *priv,
 	int index = ring->prod & ring->size_mask;
 
 	while ((u32) (ring->prod - ring->cons) < ring->actual_size) {
-		if (mlx4_en_prepare_rx_desc(priv, ring, index))
+		if (mlx4_en_prepare_rx_desc(priv, ring, index, GFP_ATOMIC))
 			break;
 		ring->prod++;
 		index = ring->prod & ring->size_mask;
@@ -656,8 +683,11 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 				 * - DIX Ethernet (type interpretation)
 				 * - TCP/IP (v4)
 				 * - without IP options
-				 * - not an IP fragment */
-				if (dev->features & NETIF_F_GRO) {
+				 * - not an IP fragment
+				 * - no LLS polling in progress
+				 */
+				if (!mlx4_en_cq_ll_polling(cq) &&
+				    (dev->features & NETIF_F_GRO)) {
 					struct sk_buff *gro_skb = napi_get_frags(&cq->napi);
 					if (!gro_skb)
 						goto next;
@@ -737,6 +767,8 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 					       timestamp);
 		}
 
+		skb_mark_ll(skb, &cq->napi);
+
 		/* Push it up the stack */
 		netif_receive_skb(skb);
 
@@ -781,8 +813,13 @@ int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget)
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 	int done;
 
+	if (!mlx4_en_cq_lock_napi(cq))
+		return budget;
+
 	done = mlx4_en_process_rx_cq(dev, cq, budget);
 
+	mlx4_en_cq_unlock_napi(cq);
+
 	/* If we used up all the quota - we're probably not done yet... */
 	if (done == budget)
 		INC_PERF_COUNTER(priv->pstats.napi_quota);
@@ -794,21 +831,7 @@ int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget)
 	return done;
 }
 
-
-/* Calculate the last offset position that accommodates a full fragment
- * (assuming fagment size = stride-align) */
-static int mlx4_en_last_alloc_offset(struct mlx4_en_priv *priv, u16 stride, u16 align)
-{
-	u16 res = MLX4_EN_ALLOC_SIZE % stride;
-	u16 offset = MLX4_EN_ALLOC_SIZE - stride - res + align;
-
-	en_dbg(DRV, priv, "Calculated last offset for stride:%d align:%d "
-			    "res:%d offset:%d\n", stride, align, res, offset);
-	return offset;
-}
-
-
-static int frag_sizes[] = {
+static const int frag_sizes[] = {
 	FRAG_SZ0,
 	FRAG_SZ1,
 	FRAG_SZ2,
@@ -836,9 +859,6 @@ void mlx4_en_calc_rx_buf(struct net_device *dev)
 			priv->frag_info[i].frag_stride =
 				ALIGN(frag_sizes[i], SMP_CACHE_BYTES);
 		}
-		priv->frag_info[i].last_offset = mlx4_en_last_alloc_offset(
-						priv, priv->frag_info[i].frag_stride,
-						priv->frag_info[i].frag_align);
 		buf_size += priv->frag_info[i].frag_size;
 		i++;
 	}
@@ -850,13 +870,13 @@ void mlx4_en_calc_rx_buf(struct net_device *dev)
 	en_dbg(DRV, priv, "Rx buffer scatter-list (effective-mtu:%d "
 		  "num_frags:%d):\n", eff_mtu, priv->num_frags);
 	for (i = 0; i < priv->num_frags; i++) {
-		en_dbg(DRV, priv, "  frag:%d - size:%d prefix:%d align:%d "
-				"stride:%d last_offset:%d\n", i,
-				priv->frag_info[i].frag_size,
-				priv->frag_info[i].frag_prefix_size,
-				priv->frag_info[i].frag_align,
-				priv->frag_info[i].frag_stride,
-				priv->frag_info[i].last_offset);
+		en_err(priv,
+		       "  frag:%d - size:%d prefix:%d align:%d stride:%d\n",
+		       i,
+		       priv->frag_info[i].frag_size,
+		       priv->frag_info[i].frag_prefix_size,
+		       priv->frag_info[i].frag_align,
+		       priv->frag_info[i].frag_stride);
 	}
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index 4e6877a..7c49238 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -544,7 +544,7 @@ u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb)
 	if (vlan_tx_tag_present(skb))
 		up = vlan_tx_tag_get(skb) >> VLAN_PRIO_SHIFT;
 
-	return __skb_tx_hash(dev, skb, rings_p_up) + up * rings_p_up;
+	return __netdev_pick_tx(dev, skb) % rings_p_up + up * rings_p_up;
 }
 
 static void mlx4_bf_copy(void __iomem *dst, unsigned long *src, unsigned bytecnt)
diff --git a/drivers/net/ethernet/mellanox/mlx4/eq.c b/drivers/net/ethernet/mellanox/mlx4/eq.c
index 6000342..7e04286 100644
--- a/drivers/net/ethernet/mellanox/mlx4/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/eq.c
@@ -448,6 +448,7 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
 	int i;
 	enum slave_port_gen_event gen_event;
 	unsigned long flags;
+	struct mlx4_vport_state *s_info;
 
 	while ((eqe = next_eqe_sw(eq, dev->caps.eqe_factor))) {
 		/*
@@ -556,7 +557,9 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
 						mlx4_dbg(dev, "%s: Sending MLX4_PORT_CHANGE_SUBTYPE_DOWN"
 							 " to slave: %d, port:%d\n",
 							 __func__, i, port);
-						mlx4_slave_event(dev, i, eqe);
+						s_info = &priv->mfunc.master.vf_oper[slave].vport[port].state;
+						if (IFLA_VF_LINK_STATE_AUTO == s_info->link_state)
+							mlx4_slave_event(dev, i, eqe);
 					} else {  /* IB port */
 						set_and_calc_slave_port_state(dev, i, port,
 									      MLX4_PORT_STATE_DEV_EVENT_PORT_DOWN,
@@ -580,7 +583,9 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
 					for (i = 0; i < dev->num_slaves; i++) {
 						if (i == mlx4_master_func_num(dev))
 							continue;
-						mlx4_slave_event(dev, i, eqe);
+						s_info = &priv->mfunc.master.vf_oper[slave].vport[port].state;
+						if (IFLA_VF_LINK_STATE_AUTO == s_info->link_state)
+							mlx4_slave_event(dev, i, eqe);
 					}
 				else /* IB port */
 					/* port-up event will be sent to a slave when the
diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c
index 2c97901..8873d68 100644
--- a/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -133,7 +133,8 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[4] = "Automatic MAC reassignment support",
 		[5] = "Time stamping support",
 		[6] = "VST (control vlan insertion/stripping) support",
-		[7] = "FSM (MAC anti-spoofing) support"
+		[7] = "FSM (MAC anti-spoofing) support",
+		[8] = "Dynamic QP updates support"
 	};
 	int i;
 
@@ -659,6 +660,8 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 			 QUERY_DEV_CAP_MAX_COUNTERS_OFFSET);
 
 	MLX4_GET(field32, outbox, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
+	if (field32 & (1 << 16))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_UPDATE_QP;
 	if (field32 & (1 << 26))
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_VLAN_CONTROL;
 	if (field32 & (1 << 20))
@@ -830,8 +833,10 @@ int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave,
 	u8 port_type;
 	u16 short_field;
 	int err;
+	int admin_link_state;
 
 #define MLX4_VF_PORT_NO_LINK_SENSE_MASK	0xE0
+#define MLX4_PORT_LINK_UP_MASK		0x80
 #define QUERY_PORT_CUR_MAX_PKEY_OFFSET	0x0c
 #define QUERY_PORT_CUR_MAX_GID_OFFSET	0x0e
 
@@ -861,6 +866,12 @@ int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave,
 		/* set port type to currently operating port type */
 		port_type |= (dev->caps.port_type[vhcr->in_modifier] & 0x3);
 
+		admin_link_state = priv->mfunc.master.vf_oper[slave].vport[vhcr->in_modifier].state.link_state;
+		if (IFLA_VF_LINK_STATE_ENABLE == admin_link_state)
+			port_type |= MLX4_PORT_LINK_UP_MASK;
+		else if (IFLA_VF_LINK_STATE_DISABLE == admin_link_state)
+			port_type &= ~MLX4_PORT_LINK_UP_MASK;
+
 		MLX4_PUT(outbox->buf, port_type,
 			 QUERY_PORT_SUPPORTED_TYPE_OFFSET);
 
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 264ddeb..e85af92 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -842,11 +842,11 @@ static ssize_t set_port_ib_mtu(struct device *dev,
 		return -EINVAL;
 	}
 
-	err = sscanf(buf, "%d", &mtu);
-	if (err > 0)
+	err = kstrtoint(buf, 0, &mtu);
+	if (!err)
 		ibta_mtu = int_to_ibta_mtu(mtu);
 
-	if (err <= 0 || ibta_mtu < 0) {
+	if (err || ibta_mtu < 0) {
 		mlx4_err(mdev, "%s is invalid IBTA mtu\n", buf);
 		return -EINVAL;
 	}
@@ -2080,6 +2080,11 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data)
 		       num_vfs, MLX4_MAX_NUM_VF);
 		return -EINVAL;
 	}
+
+	if (num_vfs < 0) {
+		pr_err("num_vfs module parameter cannot be negative\n");
+		return -EINVAL;
+	}
 	/*
 	 * Check for BARs.
 	 */
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index df15bb6..17d9277 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -482,6 +482,7 @@ struct mlx4_vport_state {
 	u8  default_qos;
 	u32 tx_rate;
 	bool spoofchk;
+	u32 link_state;
 };
 
 struct mlx4_vf_admin_state {
@@ -570,6 +571,25 @@ struct mlx4_cmd {
 	u8			comm_toggle;
 };
 
+enum {
+	MLX4_VF_IMMED_VLAN_FLAG_VLAN = 1 << 0,
+	MLX4_VF_IMMED_VLAN_FLAG_QOS = 1 << 1,
+	MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE = 1 << 2,
+};
+struct mlx4_vf_immed_vlan_work {
+	struct work_struct	work;
+	struct mlx4_priv	*priv;
+	int			flags;
+	int			slave;
+	int			vlan_ix;
+	int			orig_vlan_ix;
+	u8			port;
+	u8			qos;
+	u16			vlan_id;
+	u16			orig_vlan_id;
+};
+
+
 struct mlx4_uar_table {
 	struct mlx4_bitmap	bitmap;
 };
@@ -1217,4 +1237,6 @@ static inline spinlock_t *mlx4_tlock(struct mlx4_dev *dev)
 
 #define NOT_MASKED_PD_BITS 17
 
+void mlx4_vf_immed_vlan_work_handler(struct work_struct *_work);
+
 #endif /* MLX4_H */
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index b1d7657..35fb60e 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -96,13 +96,14 @@
 
 /* Use the maximum between 16384 and a single page */
 #define MLX4_EN_ALLOC_SIZE	PAGE_ALIGN(16384)
-#define MLX4_EN_ALLOC_ORDER	get_order(MLX4_EN_ALLOC_SIZE)
 
-/* Receive fragment sizes; we use at most 4 fragments (for 9600 byte MTU
+#define MLX4_EN_ALLOC_PREFER_ORDER	PAGE_ALLOC_COSTLY_ORDER
+
+/* Receive fragment sizes; we use at most 3 fragments (for 9600 byte MTU
  * and 4K allocations) */
 enum {
-	FRAG_SZ0 = 512 - NET_IP_ALIGN,
-	FRAG_SZ1 = 1024,
+	FRAG_SZ0 = 1536 - NET_IP_ALIGN,
+	FRAG_SZ1 = 4096,
 	FRAG_SZ2 = 4096,
 	FRAG_SZ3 = MLX4_EN_ALLOC_SIZE
 };
@@ -234,9 +235,10 @@ struct mlx4_en_tx_desc {
 #define MLX4_EN_CX3_HIGH_ID	0x1005
 
 struct mlx4_en_rx_alloc {
-	struct page *page;
-	dma_addr_t dma;
-	u16 offset;
+	struct page	*page;
+	dma_addr_t	dma;
+	u32		offset;
+	u32		size;
 };
 
 struct mlx4_en_tx_ring {
@@ -290,6 +292,11 @@ struct mlx4_en_rx_ring {
 	void *rx_info;
 	unsigned long bytes;
 	unsigned long packets;
+#ifdef CONFIG_NET_LL_RX_POLL
+	unsigned long yields;
+	unsigned long misses;
+	unsigned long cleaned;
+#endif
 	unsigned long csum_ok;
 	unsigned long csum_none;
 	int hwtstamp_rx_filter;
@@ -310,6 +317,19 @@ struct mlx4_en_cq {
 	u16 moder_cnt;
 	struct mlx4_cqe *buf;
 #define MLX4_EN_OPCODE_ERROR	0x1e
+
+#ifdef CONFIG_NET_LL_RX_POLL
+	unsigned int state;
+#define MLX4_EN_CQ_STATE_IDLE        0
+#define MLX4_EN_CQ_STATE_NAPI     1    /* NAPI owns this CQ */
+#define MLX4_EN_CQ_STATE_POLL     2    /* poll owns this CQ */
+#define MLX4_CQ_LOCKED (MLX4_EN_CQ_STATE_NAPI | MLX4_EN_CQ_STATE_POLL)
+#define MLX4_EN_CQ_STATE_NAPI_YIELD  4    /* NAPI yielded this CQ */
+#define MLX4_EN_CQ_STATE_POLL_YIELD  8    /* poll yielded this CQ */
+#define CQ_YIELD (MLX4_EN_CQ_STATE_NAPI_YIELD | MLX4_EN_CQ_STATE_POLL_YIELD)
+#define CQ_USER_PEND (MLX4_EN_CQ_STATE_POLL | MLX4_EN_CQ_STATE_POLL_YIELD)
+	spinlock_t poll_lock; /* protects from LLS/napi conflicts */
+#endif  /* CONFIG_NET_LL_RX_POLL */
 };
 
 struct mlx4_en_port_profile {
@@ -421,8 +441,6 @@ struct mlx4_en_frag_info {
 	u16 frag_prefix_size;
 	u16 frag_stride;
 	u16 frag_align;
-	u16 last_offset;
-
 };
 
 #ifdef CONFIG_MLX4_EN_DCB
@@ -562,6 +580,115 @@ struct mlx4_mac_entry {
 	struct rcu_head rcu;
 };
 
+#ifdef CONFIG_NET_LL_RX_POLL
+static inline void mlx4_en_cq_init_lock(struct mlx4_en_cq *cq)
+{
+	spin_lock_init(&cq->poll_lock);
+	cq->state = MLX4_EN_CQ_STATE_IDLE;
+}
+
+/* called from the device poll rutine to get ownership of a cq */
+static inline bool mlx4_en_cq_lock_napi(struct mlx4_en_cq *cq)
+{
+	int rc = true;
+	spin_lock(&cq->poll_lock);
+	if (cq->state & MLX4_CQ_LOCKED) {
+		WARN_ON(cq->state & MLX4_EN_CQ_STATE_NAPI);
+		cq->state |= MLX4_EN_CQ_STATE_NAPI_YIELD;
+		rc = false;
+	} else
+		/* we don't care if someone yielded */
+		cq->state = MLX4_EN_CQ_STATE_NAPI;
+	spin_unlock(&cq->poll_lock);
+	return rc;
+}
+
+/* returns true is someone tried to get the cq while napi had it */
+static inline bool mlx4_en_cq_unlock_napi(struct mlx4_en_cq *cq)
+{
+	int rc = false;
+	spin_lock(&cq->poll_lock);
+	WARN_ON(cq->state & (MLX4_EN_CQ_STATE_POLL |
+			       MLX4_EN_CQ_STATE_NAPI_YIELD));
+
+	if (cq->state & MLX4_EN_CQ_STATE_POLL_YIELD)
+		rc = true;
+	cq->state = MLX4_EN_CQ_STATE_IDLE;
+	spin_unlock(&cq->poll_lock);
+	return rc;
+}
+
+/* called from mlx4_en_low_latency_poll() */
+static inline bool mlx4_en_cq_lock_poll(struct mlx4_en_cq *cq)
+{
+	int rc = true;
+	spin_lock_bh(&cq->poll_lock);
+	if ((cq->state & MLX4_CQ_LOCKED)) {
+		struct net_device *dev = cq->dev;
+		struct mlx4_en_priv *priv = netdev_priv(dev);
+		struct mlx4_en_rx_ring *rx_ring = &priv->rx_ring[cq->ring];
+
+		cq->state |= MLX4_EN_CQ_STATE_POLL_YIELD;
+		rc = false;
+		rx_ring->yields++;
+	} else
+		/* preserve yield marks */
+		cq->state |= MLX4_EN_CQ_STATE_POLL;
+	spin_unlock_bh(&cq->poll_lock);
+	return rc;
+}
+
+/* returns true if someone tried to get the cq while it was locked */
+static inline bool mlx4_en_cq_unlock_poll(struct mlx4_en_cq *cq)
+{
+	int rc = false;
+	spin_lock_bh(&cq->poll_lock);
+	WARN_ON(cq->state & (MLX4_EN_CQ_STATE_NAPI));
+
+	if (cq->state & MLX4_EN_CQ_STATE_POLL_YIELD)
+		rc = true;
+	cq->state = MLX4_EN_CQ_STATE_IDLE;
+	spin_unlock_bh(&cq->poll_lock);
+	return rc;
+}
+
+/* true if a socket is polling, even if it did not get the lock */
+static inline bool mlx4_en_cq_ll_polling(struct mlx4_en_cq *cq)
+{
+	WARN_ON(!(cq->state & MLX4_CQ_LOCKED));
+	return cq->state & CQ_USER_PEND;
+}
+#else
+static inline void mlx4_en_cq_init_lock(struct mlx4_en_cq *cq)
+{
+}
+
+static inline bool mlx4_en_cq_lock_napi(struct mlx4_en_cq *cq)
+{
+	return true;
+}
+
+static inline bool mlx4_en_cq_unlock_napi(struct mlx4_en_cq *cq)
+{
+	return false;
+}
+
+static inline bool mlx4_en_cq_lock_poll(struct mlx4_en_cq *cq)
+{
+	return false;
+}
+
+static inline bool mlx4_en_cq_unlock_poll(struct mlx4_en_cq *cq)
+{
+	return false;
+}
+
+static inline bool mlx4_en_cq_ll_polling(struct mlx4_en_cq *cq)
+{
+	return false;
+}
+#endif /* CONFIG_NET_LL_RX_POLL */
+
 #define MLX4_EN_WOL_DO_MODIFY (1ULL << 63)
 
 void mlx4_en_update_loopback_state(struct net_device *dev,
diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
index 1157f02..f984a89 100644
--- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
+++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -101,6 +101,8 @@ struct res_qp {
 	spinlock_t		mcg_spl;
 	int			local_qpn;
 	atomic_t		ref_count;
+	u32			qpc_flags;
+	u8			sched_queue;
 };
 
 enum res_mtt_states {
@@ -355,7 +357,7 @@ static void update_gid(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *inbox,
 
 static int update_vport_qp_param(struct mlx4_dev *dev,
 				 struct mlx4_cmd_mailbox *inbox,
-				 u8 slave)
+				 u8 slave, u32 qpn)
 {
 	struct mlx4_qp_context	*qpc = inbox->buf + 8;
 	struct mlx4_vport_oper_state *vp_oper;
@@ -369,12 +371,30 @@ static int update_vport_qp_param(struct mlx4_dev *dev,
 
 	if (MLX4_VGT != vp_oper->state.default_vlan) {
 		qp_type	= (be32_to_cpu(qpc->flags) >> 16) & 0xff;
-		if (MLX4_QP_ST_RC == qp_type)
+		if (MLX4_QP_ST_RC == qp_type ||
+		    (MLX4_QP_ST_UD == qp_type &&
+		     !mlx4_is_qp_reserved(dev, qpn)))
 			return -EINVAL;
 
+		/* the reserved QPs (special, proxy, tunnel)
+		 * do not operate over vlans
+		 */
+		if (mlx4_is_qp_reserved(dev, qpn))
+			return 0;
+
 		/* force strip vlan by clear vsd */
 		qpc->param3 &= ~cpu_to_be32(MLX4_STRIP_VLAN);
-		if (0 != vp_oper->state.default_vlan) {
+
+		if (vp_oper->state.link_state == IFLA_VF_LINK_STATE_DISABLE &&
+		    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP) {
+			qpc->pri_path.vlan_control =
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
+				MLX4_VLAN_CTRL_ETH_TX_BLOCK_UNTAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED |
+				MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+		} else if (0 != vp_oper->state.default_vlan) {
 			qpc->pri_path.vlan_control =
 				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
 				MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
@@ -2114,6 +2134,8 @@ int mlx4_RST2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
 	if (err)
 		return err;
 	qp->local_qpn = local_qpn;
+	qp->sched_queue = 0;
+	qp->qpc_flags = be32_to_cpu(qpc->flags);
 
 	err = get_res(dev, slave, mtt_base, RES_MTT, &mtt);
 	if (err)
@@ -2836,6 +2858,9 @@ int mlx4_INIT2RTR_QP_wrapper(struct mlx4_dev *dev, int slave,
 {
 	int err;
 	struct mlx4_qp_context *qpc = inbox->buf + 8;
+	int qpn = vhcr->in_modifier & 0x7fffff;
+	struct res_qp *qp;
+	u8 orig_sched_queue;
 
 	err = verify_qp_parameters(dev, inbox, QP_TRANS_INIT2RTR, slave);
 	if (err)
@@ -2844,11 +2869,30 @@ int mlx4_INIT2RTR_QP_wrapper(struct mlx4_dev *dev, int slave,
 	update_pkey_index(dev, slave, inbox);
 	update_gid(dev, inbox, (u8)slave);
 	adjust_proxy_tun_qkey(dev, vhcr, qpc);
-	err = update_vport_qp_param(dev, inbox, slave);
+	orig_sched_queue = qpc->pri_path.sched_queue;
+	err = update_vport_qp_param(dev, inbox, slave, qpn);
 	if (err)
 		return err;
 
-	return mlx4_GEN_QP_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+	err = get_res(dev, slave, qpn, RES_QP, &qp);
+	if (err)
+		return err;
+	if (qp->com.from_state != RES_QP_HW) {
+		err = -EBUSY;
+		goto out;
+	}
+
+	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
+out:
+	/* if no error, save sched queue value passed in by VF. This is
+	 * essentially the QOS value provided by the VF. This will be useful
+	 * if we allow dynamic changes from VST back to VGT
+	 */
+	if (!err)
+		qp->sched_queue = orig_sched_queue;
+
+	put_res(dev, slave, qpn, RES_QP);
+	return err;
 }
 
 int mlx4_RTR2RTS_QP_wrapper(struct mlx4_dev *dev, int slave,
@@ -3932,3 +3976,112 @@ void mlx4_delete_all_resources_for_slave(struct mlx4_dev *dev, int slave)
 	rem_slave_xrcdns(dev, slave);
 	mutex_unlock(&priv->mfunc.master.res_tracker.slave_list[slave].mutex);
 }
+
+void mlx4_vf_immed_vlan_work_handler(struct work_struct *_work)
+{
+	struct mlx4_vf_immed_vlan_work *work =
+		container_of(_work, struct mlx4_vf_immed_vlan_work, work);
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_update_qp_context *upd_context;
+	struct mlx4_dev *dev = &work->priv->dev;
+	struct mlx4_resource_tracker *tracker =
+		&work->priv->mfunc.master.res_tracker;
+	struct list_head *qp_list =
+		&tracker->slave_list[work->slave].res_list[RES_QP];
+	struct res_qp *qp;
+	struct res_qp *tmp;
+	u64 qp_mask = ((1ULL << MLX4_UPD_QP_PATH_MASK_ETH_TX_BLOCK_UNTAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_TX_BLOCK_1P) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_TX_BLOCK_TAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_RX_BLOCK_UNTAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_RX_BLOCK_1P) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_RX_BLOCK_TAGGED) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_VLAN_INDEX) |
+		       (1ULL << MLX4_UPD_QP_PATH_MASK_SCHED_QUEUE));
+
+	int err;
+	int port, errors = 0;
+	u8 vlan_control;
+
+	if (mlx4_is_slave(dev)) {
+		mlx4_warn(dev, "Trying to update-qp in slave %d\n",
+			  work->slave);
+		goto out;
+	}
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		goto out;
+	if (work->flags & MLX4_VF_IMMED_VLAN_FLAG_LINK_DISABLE) /* block all */
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_TX_BLOCK_UNTAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+	else if (!work->vlan_id)
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
+	else
+		vlan_control = MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
+			MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED;
+
+	upd_context = mailbox->buf;
+	upd_context->primary_addr_path_mask = cpu_to_be64(qp_mask);
+	upd_context->qp_context.pri_path.vlan_control = vlan_control;
+	upd_context->qp_context.pri_path.vlan_index = work->vlan_ix;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry_safe(qp, tmp, qp_list, com.list) {
+		spin_unlock_irq(mlx4_tlock(dev));
+		if (qp->com.owner == work->slave) {
+			if (qp->com.from_state != RES_QP_HW ||
+			    !qp->sched_queue ||  /* no INIT2RTR trans yet */
+			    mlx4_is_qp_reserved(dev, qp->local_qpn) ||
+			    qp->qpc_flags & (1 << MLX4_RSS_QPC_FLAG_OFFSET)) {
+				spin_lock_irq(mlx4_tlock(dev));
+				continue;
+			}
+			port = (qp->sched_queue >> 6 & 1) + 1;
+			if (port != work->port) {
+				spin_lock_irq(mlx4_tlock(dev));
+				continue;
+			}
+			upd_context->qp_context.pri_path.sched_queue =
+				qp->sched_queue & 0xC7;
+			upd_context->qp_context.pri_path.sched_queue |=
+				((work->qos & 0x7) << 3);
+
+			err = mlx4_cmd(dev, mailbox->dma,
+				       qp->local_qpn & 0xffffff,
+				       0, MLX4_CMD_UPDATE_QP,
+				       MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
+			if (err) {
+				mlx4_info(dev, "UPDATE_QP failed for slave %d, "
+					  "port %d, qpn %d (%d)\n",
+					  work->slave, port, qp->local_qpn,
+					  err);
+				errors++;
+			}
+		}
+		spin_lock_irq(mlx4_tlock(dev));
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	if (errors)
+		mlx4_err(dev, "%d UPDATE_QP failures for slave %d, port %d\n",
+			 errors, work->slave, work->port);
+
+	/* unregister previous vlan_id if needed and we had no errors
+	 * while updating the QPs
+	 */
+	if (work->flags & MLX4_VF_IMMED_VLAN_FLAG_VLAN && !errors &&
+	    NO_INDX != work->orig_vlan_ix)
+		__mlx4_unregister_vlan(&work->priv->dev, work->port,
+				       work->orig_vlan_ix);
+out:
+	kfree(work);
+	return;
+}
diff --git a/drivers/net/ethernet/micrel/Kconfig b/drivers/net/ethernet/micrel/Kconfig
index fe42fc0..d16b11e 100644
--- a/drivers/net/ethernet/micrel/Kconfig
+++ b/drivers/net/ethernet/micrel/Kconfig
@@ -22,7 +22,6 @@ if NET_VENDOR_MICREL
 config ARM_KS8695_ETHER
 	tristate "KS8695 Ethernet support"
 	depends on ARM && ARCH_KS8695
-	select NET_CORE
 	select MII
 	---help---
 	  If you wish to compile a kernel for the KS8695 and want to
@@ -39,7 +38,6 @@ config KS8842
 config KS8851
 	tristate "Micrel KS8851 SPI"
 	depends on SPI
-	select NET_CORE
 	select MII
 	select CRC32
 	select EEPROM_93CX6
@@ -49,7 +47,6 @@ config KS8851
 config KS8851_MLL
 	tristate "Micrel KS8851 MLL"
 	depends on HAS_IOMEM
-	select NET_CORE
 	select MII
 	---help---
 	  This platform driver is for Micrel KS8851 Address/data bus
@@ -58,7 +55,6 @@ config KS8851_MLL
 config KSZ884X_PCI
 	tristate "Micrel KSZ8841/2 PCI"
 	depends on PCI
-	select NET_CORE
 	select MII
 	select CRC32
 	---help---
diff --git a/drivers/net/ethernet/micrel/ks8695net.c b/drivers/net/ethernet/micrel/ks8695net.c
index b6c60fd..106eb97 100644
--- a/drivers/net/ethernet/micrel/ks8695net.c
+++ b/drivers/net/ethernet/micrel/ks8695net.c
@@ -1600,7 +1600,6 @@ ks8695_drv_remove(struct platform_device *pdev)
 	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct ks8695_priv *ksp = netdev_priv(ndev);
 
-	platform_set_drvdata(pdev, NULL);
 	netif_napi_del(&ksp->napi);
 
 	unregister_netdev(ndev);
diff --git a/drivers/net/ethernet/micrel/ks8842.c b/drivers/net/ethernet/micrel/ks8842.c
index fbcb9e7..e393d99 100644
--- a/drivers/net/ethernet/micrel/ks8842.c
+++ b/drivers/net/ethernet/micrel/ks8842.c
@@ -1250,7 +1250,6 @@ static int ks8842_remove(struct platform_device *pdev)
 	iounmap(adapter->hw_addr);
 	free_netdev(netdev);
 	release_mem_region(iomem->start, resource_size(iomem));
-	platform_set_drvdata(pdev, NULL);
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/micrel/ks8851_mll.c b/drivers/net/ethernet/micrel/ks8851_mll.c
index ddaf138..ac20098 100644
--- a/drivers/net/ethernet/micrel/ks8851_mll.c
+++ b/drivers/net/ethernet/micrel/ks8851_mll.c
@@ -35,6 +35,9 @@
 #include <linux/delay.h>
 #include <linux/slab.h>
 #include <linux/ks8851_mll.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_net.h>
 
 #define	DRV_NAME	"ks8851_mll"
 
@@ -1524,6 +1527,13 @@ static int ks_hw_init(struct ks_net *ks)
 	return true;
 }
 
+#if defined(CONFIG_OF)
+static const struct of_device_id ks8851_ml_dt_ids[] = {
+	{ .compatible = "micrel,ks8851-mll" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, ks8851_ml_dt_ids);
+#endif
 
 static int ks8851_probe(struct platform_device *pdev)
 {
@@ -1532,7 +1542,7 @@ static int ks8851_probe(struct platform_device *pdev)
 	struct net_device *netdev;
 	struct ks_net *ks;
 	u16 id, data;
-	struct ks8851_mll_platform_data *pdata;
+	const char *mac;
 
 	io_d = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	io_c = platform_get_resource(pdev, IORESOURCE_MEM, 1);
@@ -1619,13 +1629,21 @@ static int ks8851_probe(struct platform_device *pdev)
 	ks_wrreg16(ks, KS_OBCR, data | OBCR_ODS_16MA);
 
 	/* overwriting the default MAC address */
-	pdata = pdev->dev.platform_data;
-	if (!pdata) {
-		netdev_err(netdev, "No platform data\n");
-		err = -ENODEV;
-		goto err_pdata;
+	if (pdev->dev.of_node) {
+		mac = of_get_mac_address(pdev->dev.of_node);
+		if (mac)
+			memcpy(ks->mac_addr, mac, ETH_ALEN);
+	} else {
+		struct ks8851_mll_platform_data *pdata;
+
+		pdata = pdev->dev.platform_data;
+		if (!pdata) {
+			netdev_err(netdev, "No platform data\n");
+			err = -ENODEV;
+			goto err_pdata;
+		}
+		memcpy(ks->mac_addr, pdata->mac_addr, ETH_ALEN);
 	}
-	memcpy(ks->mac_addr, pdata->mac_addr, 6);
 	if (!is_valid_ether_addr(ks->mac_addr)) {
 		/* Use random MAC address if none passed */
 		eth_random_addr(ks->mac_addr);
@@ -1671,7 +1689,6 @@ static int ks8851_remove(struct platform_device *pdev)
 	iounmap(ks->hw_addr);
 	free_netdev(netdev);
 	release_mem_region(iomem->start, resource_size(iomem));
-	platform_set_drvdata(pdev, NULL);
 	return 0;
 
 }
@@ -1680,6 +1697,7 @@ static struct platform_driver ks8851_platform_driver = {
 	.driver = {
 		.name = DRV_NAME,
 		.owner = THIS_MODULE,
+		.of_match_table	= of_match_ptr(ks8851_ml_dt_ids),
 	},
 	.probe = ks8851_probe,
 	.remove = ks8851_remove,
diff --git a/drivers/net/ethernet/myricom/myri10ge/myri10ge.c b/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
index 7be9788..967bae8 100644
--- a/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
+++ b/drivers/net/ethernet/myricom/myri10ge/myri10ge.c
@@ -3299,7 +3299,7 @@ static int myri10ge_resume(struct pci_dev *pdev)
 	if (mgp == NULL)
 		return -EINVAL;
 	netdev = mgp->dev;
-	pci_set_power_state(pdev, 0);	/* zeros conf space as a side effect */
+	pci_set_power_state(pdev, PCI_D0);	/* zeros conf space as a side effect */
 	msleep(5);		/* give card time to respond */
 	pci_read_config_word(mgp->pdev, PCI_VENDOR_ID, &vendor);
 	if (vendor == 0xffff) {
diff --git a/drivers/net/ethernet/netx-eth.c b/drivers/net/ethernet/netx-eth.c
index cb9e638..dc2c6f5 100644
--- a/drivers/net/ethernet/netx-eth.c
+++ b/drivers/net/ethernet/netx-eth.c
@@ -422,7 +422,6 @@ exit_free_pfifo:
 exit_free_xc:
 	free_xc(priv->xc);
 exit_free_netdev:
-	platform_set_drvdata(pdev, NULL);
 	free_netdev(ndev);
 exit:
 	return ret;
@@ -430,11 +429,9 @@ exit:
 
 static int netx_eth_drv_remove(struct platform_device *pdev)
 {
-	struct net_device *ndev = dev_get_drvdata(&pdev->dev);
+	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct netx_eth_priv *priv = netdev_priv(ndev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	unregister_netdev(ndev);
 	xc_stop(priv->xc);
 	free_xc(priv->xc);
diff --git a/drivers/net/ethernet/nuvoton/Kconfig b/drivers/net/ethernet/nuvoton/Kconfig
index 334c171..01182b5 100644
--- a/drivers/net/ethernet/nuvoton/Kconfig
+++ b/drivers/net/ethernet/nuvoton/Kconfig
@@ -22,7 +22,6 @@ config W90P910_ETH
 	tristate "Nuvoton w90p910 Ethernet support"
 	depends on ARM && ARCH_W90X900
 	select PHYLIB
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here if you want to use built-in Ethernet ports
diff --git a/drivers/net/ethernet/nuvoton/w90p910_ether.c b/drivers/net/ethernet/nuvoton/w90p910_ether.c
index 3df8287..e88bdb1 100644
--- a/drivers/net/ethernet/nuvoton/w90p910_ether.c
+++ b/drivers/net/ethernet/nuvoton/w90p910_ether.c
@@ -1051,7 +1051,6 @@ failed_put_clk:
 	clk_put(ether->clk);
 failed_free_rxirq:
 	free_irq(ether->rxirq, pdev);
-	platform_set_drvdata(pdev, NULL);
 failed_free_txirq:
 	free_irq(ether->txirq, pdev);
 failed_free_io:
@@ -1080,7 +1079,6 @@ static int w90p910_ether_remove(struct platform_device *pdev)
 	free_irq(ether->rxirq, dev);
 
 	del_timer_sync(&ether->check_timer);
-	platform_set_drvdata(pdev, NULL);
 
 	free_netdev(dev);
 	return 0;
diff --git a/drivers/net/ethernet/nvidia/forcedeth.c b/drivers/net/ethernet/nvidia/forcedeth.c
index b003fe5..098b96d 100644
--- a/drivers/net/ethernet/nvidia/forcedeth.c
+++ b/drivers/net/ethernet/nvidia/forcedeth.c
@@ -6340,7 +6340,7 @@ static DEFINE_PCI_DEVICE_TABLE(pci_tbl) = {
 	{0,},
 };
 
-static struct pci_driver driver = {
+static struct pci_driver forcedeth_pci_driver = {
 	.name		= DRV_NAME,
 	.id_table	= pci_tbl,
 	.probe		= nv_probe,
@@ -6349,16 +6349,6 @@ static struct pci_driver driver = {
 	.driver.pm	= NV_PM_OPS,
 };
 
-static int __init init_nic(void)
-{
-	return pci_register_driver(&driver);
-}
-
-static void __exit exit_nic(void)
-{
-	pci_unregister_driver(&driver);
-}
-
 module_param(max_interrupt_work, int, 0);
 MODULE_PARM_DESC(max_interrupt_work, "forcedeth maximum events handled per interrupt");
 module_param(optimization_mode, int, 0);
@@ -6379,11 +6369,8 @@ module_param(debug_tx_timeout, bool, 0);
 MODULE_PARM_DESC(debug_tx_timeout,
 		 "Dump tx related registers and ring when tx_timeout happens");
 
+module_pci_driver(forcedeth_pci_driver);
 MODULE_AUTHOR("Manfred Spraul <manfred@colorfullife.com>");
 MODULE_DESCRIPTION("Reverse Engineered nForce ethernet driver");
 MODULE_LICENSE("GPL");
-
 MODULE_DEVICE_TABLE(pci, pci_tbl);
-
-module_init(init_nic);
-module_exit(exit_nic);
diff --git a/drivers/net/ethernet/nxp/lpc_eth.c b/drivers/net/ethernet/nxp/lpc_eth.c
index 55a5548..a061b93 100644
--- a/drivers/net/ethernet/nxp/lpc_eth.c
+++ b/drivers/net/ethernet/nxp/lpc_eth.c
@@ -1483,7 +1483,6 @@ static int lpc_eth_drv_probe(struct platform_device *pdev)
 	return 0;
 
 err_out_unregister_netdev:
-	platform_set_drvdata(pdev, NULL);
 	unregister_netdev(ndev);
 err_out_dma_unmap:
 	if (!use_iram_for_net(&pldat->pdev->dev) ||
@@ -1511,7 +1510,6 @@ static int lpc_eth_drv_remove(struct platform_device *pdev)
 	struct netdata_local *pldat = netdev_priv(ndev);
 
 	unregister_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 
 	if (!use_iram_for_net(&pldat->pdev->dev) ||
 	    pldat->dma_buff_size > lpc32xx_return_iram_size())
diff --git a/drivers/net/ethernet/octeon/octeon_mgmt.c b/drivers/net/ethernet/octeon/octeon_mgmt.c
index 91a8a5d..622aa75 100644
--- a/drivers/net/ethernet/octeon/octeon_mgmt.c
+++ b/drivers/net/ethernet/octeon/octeon_mgmt.c
@@ -1448,7 +1448,7 @@ static int octeon_mgmt_probe(struct platform_device *pdev)
 
 	SET_NETDEV_DEV(netdev, &pdev->dev);
 
-	dev_set_drvdata(&pdev->dev, netdev);
+	platform_set_drvdata(pdev, netdev);
 	p = netdev_priv(netdev);
 	netif_napi_add(netdev, &p->napi, octeon_mgmt_napi_poll,
 		       OCTEON_MGMT_NAPI_WEIGHT);
@@ -1570,7 +1570,7 @@ err:
 
 static int octeon_mgmt_remove(struct platform_device *pdev)
 {
-	struct net_device *netdev = dev_get_drvdata(&pdev->dev);
+	struct net_device *netdev = platform_get_drvdata(pdev);
 
 	unregister_netdev(netdev);
 	free_netdev(netdev);
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/Kconfig b/drivers/net/ethernet/oki-semi/pch_gbe/Kconfig
index 34d05bf..cb22341 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/Kconfig
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/Kconfig
@@ -5,7 +5,6 @@
 config PCH_GBE
 	tristate "OKI SEMICONDUCTOR IOH(ML7223/ML7831) GbE"
 	depends on PCI
-	select NET_CORE
 	select MII
 	select PTP_1588_CLOCK_PCH
 	---help---
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe.h b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe.h
index 7fb7e17..7779036 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe.h
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe.h
@@ -633,6 +633,8 @@ struct pch_gbe_adapter {
 	struct pci_dev *ptp_pdev;
 };
 
+#define pch_gbe_hw_to_adapter(hw)	container_of(hw, struct pch_gbe_adapter, hw)
+
 extern const char pch_driver_version[];
 
 /* pch_gbe_main.c */
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_api.c b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_api.c
index 5ae03e8..ff3ad70 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_api.c
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_api.c
@@ -19,6 +19,7 @@
  */
 #include "pch_gbe.h"
 #include "pch_gbe_phy.h"
+#include "pch_gbe_api.h"
 
 /* bus type values */
 #define pch_gbe_bus_type_unknown	0
@@ -70,7 +71,9 @@ static s32 pch_gbe_plat_init_hw(struct pch_gbe_hw *hw)
 
 	ret_val = pch_gbe_phy_get_id(hw);
 	if (ret_val) {
-		pr_err("pch_gbe_phy_get_id error\n");
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "pch_gbe_phy_get_id error\n");
 		return ret_val;
 	}
 	pch_gbe_phy_init_setting(hw);
@@ -112,10 +115,12 @@ static void pch_gbe_plat_init_function_pointers(struct pch_gbe_hw *hw)
  *	0:	Successfully
  *	ENOSYS:	Function is not registered
  */
-inline s32 pch_gbe_hal_setup_init_funcs(struct pch_gbe_hw *hw)
+s32 pch_gbe_hal_setup_init_funcs(struct pch_gbe_hw *hw)
 {
 	if (!hw->reg) {
-		pr_err("ERROR: Registers not mapped\n");
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "ERROR: Registers not mapped\n");
 		return -ENOSYS;
 	}
 	pch_gbe_plat_init_function_pointers(hw);
@@ -126,12 +131,15 @@ inline s32 pch_gbe_hal_setup_init_funcs(struct pch_gbe_hw *hw)
  * pch_gbe_hal_get_bus_info - Obtain bus information for adapter
  * @hw:	Pointer to the HW structure
  */
-inline void pch_gbe_hal_get_bus_info(struct pch_gbe_hw *hw)
+void pch_gbe_hal_get_bus_info(struct pch_gbe_hw *hw)
 {
-	if (!hw->func->get_bus_info)
-		pr_err("ERROR: configuration\n");
-	else
-		hw->func->get_bus_info(hw);
+	if (!hw->func->get_bus_info) {
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "ERROR: configuration\n");
+		return;
+	}
+	hw->func->get_bus_info(hw);
 }
 
 /**
@@ -141,10 +149,12 @@ inline void pch_gbe_hal_get_bus_info(struct pch_gbe_hw *hw)
  *	0:	Successfully
  *	ENOSYS:	Function is not registered
  */
-inline s32 pch_gbe_hal_init_hw(struct pch_gbe_hw *hw)
+s32 pch_gbe_hal_init_hw(struct pch_gbe_hw *hw)
 {
 	if (!hw->func->init_hw) {
-		pr_err("ERROR: configuration\n");
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "ERROR: configuration\n");
 		return -ENOSYS;
 	}
 	return hw->func->init_hw(hw);
@@ -159,7 +169,7 @@ inline s32 pch_gbe_hal_init_hw(struct pch_gbe_hw *hw)
  *	0:	Successfully
  *	Negative value:	Failed
  */
-inline s32 pch_gbe_hal_read_phy_reg(struct pch_gbe_hw *hw, u32 offset,
+s32 pch_gbe_hal_read_phy_reg(struct pch_gbe_hw *hw, u32 offset,
 					u16 *data)
 {
 	if (!hw->func->read_phy_reg)
@@ -176,7 +186,7 @@ inline s32 pch_gbe_hal_read_phy_reg(struct pch_gbe_hw *hw, u32 offset,
  *	0:	Successfully
  *	Negative value:	Failed
  */
-inline s32 pch_gbe_hal_write_phy_reg(struct pch_gbe_hw *hw, u32 offset,
+s32 pch_gbe_hal_write_phy_reg(struct pch_gbe_hw *hw, u32 offset,
 					u16 data)
 {
 	if (!hw->func->write_phy_reg)
@@ -188,24 +198,30 @@ inline s32 pch_gbe_hal_write_phy_reg(struct pch_gbe_hw *hw, u32 offset,
  * pch_gbe_hal_phy_hw_reset - Hard PHY reset
  * @hw:	    Pointer to the HW structure
  */
-inline void pch_gbe_hal_phy_hw_reset(struct pch_gbe_hw *hw)
+void pch_gbe_hal_phy_hw_reset(struct pch_gbe_hw *hw)
 {
-	if (!hw->func->reset_phy)
-		pr_err("ERROR: configuration\n");
-	else
-		hw->func->reset_phy(hw);
+	if (!hw->func->reset_phy) {
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "ERROR: configuration\n");
+		return;
+	}
+	hw->func->reset_phy(hw);
 }
 
 /**
  * pch_gbe_hal_phy_sw_reset - Soft PHY reset
  * @hw:	    Pointer to the HW structure
  */
-inline void pch_gbe_hal_phy_sw_reset(struct pch_gbe_hw *hw)
+void pch_gbe_hal_phy_sw_reset(struct pch_gbe_hw *hw)
 {
-	if (!hw->func->sw_reset_phy)
-		pr_err("ERROR: configuration\n");
-	else
-		hw->func->sw_reset_phy(hw);
+	if (!hw->func->sw_reset_phy) {
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "ERROR: configuration\n");
+		return;
+	}
+	hw->func->sw_reset_phy(hw);
 }
 
 /**
@@ -215,10 +231,12 @@ inline void pch_gbe_hal_phy_sw_reset(struct pch_gbe_hw *hw)
  *	0:	Successfully
  *	ENOSYS:	Function is not registered
  */
-inline s32 pch_gbe_hal_read_mac_addr(struct pch_gbe_hw *hw)
+s32 pch_gbe_hal_read_mac_addr(struct pch_gbe_hw *hw)
 {
 	if (!hw->func->read_mac_addr) {
-		pr_err("ERROR: configuration\n");
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "ERROR: configuration\n");
 		return -ENOSYS;
 	}
 	return hw->func->read_mac_addr(hw);
@@ -228,7 +246,7 @@ inline s32 pch_gbe_hal_read_mac_addr(struct pch_gbe_hw *hw)
  * pch_gbe_hal_power_up_phy - Power up PHY
  * @hw:	Pointer to the HW structure
  */
-inline void pch_gbe_hal_power_up_phy(struct pch_gbe_hw *hw)
+void pch_gbe_hal_power_up_phy(struct pch_gbe_hw *hw)
 {
 	if (hw->func->power_up_phy)
 		hw->func->power_up_phy(hw);
@@ -238,7 +256,7 @@ inline void pch_gbe_hal_power_up_phy(struct pch_gbe_hw *hw)
  * pch_gbe_hal_power_down_phy - Power down PHY
  * @hw:	Pointer to the HW structure
  */
-inline void pch_gbe_hal_power_down_phy(struct pch_gbe_hw *hw)
+void pch_gbe_hal_power_down_phy(struct pch_gbe_hw *hw)
 {
 	if (hw->func->power_down_phy)
 		hw->func->power_down_phy(hw);
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_ethtool.c b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_ethtool.c
index 24b787b..1129db0 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_ethtool.c
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_ethtool.c
@@ -122,7 +122,7 @@ static int pch_gbe_set_settings(struct net_device *netdev,
 	}
 	ret = mii_ethtool_sset(&adapter->mii, ecmd);
 	if (ret) {
-		pr_err("Error: mii_ethtool_sset\n");
+		netdev_err(netdev, "Error: mii_ethtool_sset\n");
 		return ret;
 	}
 	hw->mac.link_speed = speed;
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c
index 0c1c65a..ab1039a 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c
@@ -287,7 +287,7 @@ static int hwtstamp_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
 	return copy_to_user(ifr->ifr_data, &cfg, sizeof(cfg)) ? -EFAULT : 0;
 }
 
-inline void pch_gbe_mac_load_mac_addr(struct pch_gbe_hw *hw)
+static inline void pch_gbe_mac_load_mac_addr(struct pch_gbe_hw *hw)
 {
 	iowrite32(0x01, &hw->reg->MAC_ADDR_LOAD);
 }
@@ -300,6 +300,7 @@ inline void pch_gbe_mac_load_mac_addr(struct pch_gbe_hw *hw)
  */
 s32 pch_gbe_mac_read_mac_addr(struct pch_gbe_hw *hw)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	u32  adr1a, adr1b;
 
 	adr1a = ioread32(&hw->reg->mac_adr[0].high);
@@ -312,7 +313,7 @@ s32 pch_gbe_mac_read_mac_addr(struct pch_gbe_hw *hw)
 	hw->mac.addr[4] = (u8)(adr1b & 0xFF);
 	hw->mac.addr[5] = (u8)((adr1b >> 8) & 0xFF);
 
-	pr_debug("hw->mac.addr : %pM\n", hw->mac.addr);
+	netdev_dbg(adapter->netdev, "hw->mac.addr : %pM\n", hw->mac.addr);
 	return 0;
 }
 
@@ -324,6 +325,7 @@ s32 pch_gbe_mac_read_mac_addr(struct pch_gbe_hw *hw)
 static void pch_gbe_wait_clr_bit(void *reg, u32 bit)
 {
 	u32 tmp;
+
 	/* wait busy */
 	tmp = 1000;
 	while ((ioread32(reg) & bit) && --tmp)
@@ -340,9 +342,10 @@ static void pch_gbe_wait_clr_bit(void *reg, u32 bit)
  */
 static void pch_gbe_mac_mar_set(struct pch_gbe_hw *hw, u8 * addr, u32 index)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	u32 mar_low, mar_high, adrmask;
 
-	pr_debug("index : 0x%x\n", index);
+	netdev_dbg(adapter->netdev, "index : 0x%x\n", index);
 
 	/*
 	 * HW expects these in little endian so we reverse the byte order
@@ -468,10 +471,11 @@ static void pch_gbe_mac_mc_addr_list_update(struct pch_gbe_hw *hw,
  */
 s32 pch_gbe_mac_force_mac_fc(struct pch_gbe_hw *hw)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	struct pch_gbe_mac_info *mac = &hw->mac;
 	u32 rx_fctrl;
 
-	pr_debug("mac->fc = %u\n", mac->fc);
+	netdev_dbg(adapter->netdev, "mac->fc = %u\n", mac->fc);
 
 	rx_fctrl = ioread32(&hw->reg->RX_FCTRL);
 
@@ -493,14 +497,16 @@ s32 pch_gbe_mac_force_mac_fc(struct pch_gbe_hw *hw)
 		mac->tx_fc_enable = true;
 		break;
 	default:
-		pr_err("Flow control param set incorrectly\n");
+		netdev_err(adapter->netdev,
+			   "Flow control param set incorrectly\n");
 		return -EINVAL;
 	}
 	if (mac->link_duplex == DUPLEX_HALF)
 		rx_fctrl &= ~PCH_GBE_FL_CTRL_EN;
 	iowrite32(rx_fctrl, &hw->reg->RX_FCTRL);
-	pr_debug("RX_FCTRL reg : 0x%08x  mac->tx_fc_enable : %d\n",
-		 ioread32(&hw->reg->RX_FCTRL), mac->tx_fc_enable);
+	netdev_dbg(adapter->netdev,
+		   "RX_FCTRL reg : 0x%08x  mac->tx_fc_enable : %d\n",
+		   ioread32(&hw->reg->RX_FCTRL), mac->tx_fc_enable);
 	return 0;
 }
 
@@ -511,10 +517,11 @@ s32 pch_gbe_mac_force_mac_fc(struct pch_gbe_hw *hw)
  */
 static void pch_gbe_mac_set_wol_event(struct pch_gbe_hw *hw, u32 wu_evt)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	u32 addr_mask;
 
-	pr_debug("wu_evt : 0x%08x  ADDR_MASK reg : 0x%08x\n",
-		 wu_evt, ioread32(&hw->reg->ADDR_MASK));
+	netdev_dbg(adapter->netdev, "wu_evt : 0x%08x  ADDR_MASK reg : 0x%08x\n",
+		   wu_evt, ioread32(&hw->reg->ADDR_MASK));
 
 	if (wu_evt) {
 		/* Set Wake-On-Lan address mask */
@@ -546,6 +553,7 @@ static void pch_gbe_mac_set_wol_event(struct pch_gbe_hw *hw, u32 wu_evt)
 u16 pch_gbe_mac_ctrl_miim(struct pch_gbe_hw *hw, u32 addr, u32 dir, u32 reg,
 			u16 data)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	u32 data_out = 0;
 	unsigned int i;
 	unsigned long flags;
@@ -558,7 +566,7 @@ u16 pch_gbe_mac_ctrl_miim(struct pch_gbe_hw *hw, u32 addr, u32 dir, u32 reg,
 		udelay(20);
 	}
 	if (i == 0) {
-		pr_err("pch-gbe.miim won't go Ready\n");
+		netdev_err(adapter->netdev, "pch-gbe.miim won't go Ready\n");
 		spin_unlock_irqrestore(&hw->miim_lock, flags);
 		return 0;	/* No way to indicate timeout error */
 	}
@@ -573,9 +581,9 @@ u16 pch_gbe_mac_ctrl_miim(struct pch_gbe_hw *hw, u32 addr, u32 dir, u32 reg,
 	}
 	spin_unlock_irqrestore(&hw->miim_lock, flags);
 
-	pr_debug("PHY %s: reg=%d, data=0x%04X\n",
-		 dir == PCH_GBE_MIIM_OPER_READ ? "READ" : "WRITE", reg,
-		 dir == PCH_GBE_MIIM_OPER_READ ? data_out : data);
+	netdev_dbg(adapter->netdev, "PHY %s: reg=%d, data=0x%04X\n",
+		   dir == PCH_GBE_MIIM_OPER_READ ? "READ" : "WRITE", reg,
+		   dir == PCH_GBE_MIIM_OPER_READ ? data_out : data);
 	return (u16) data_out;
 }
 
@@ -585,6 +593,7 @@ u16 pch_gbe_mac_ctrl_miim(struct pch_gbe_hw *hw, u32 addr, u32 dir, u32 reg,
  */
 static void pch_gbe_mac_set_pause_packet(struct pch_gbe_hw *hw)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	unsigned long tmp2, tmp3;
 
 	/* Set Pause packet */
@@ -606,10 +615,13 @@ static void pch_gbe_mac_set_pause_packet(struct pch_gbe_hw *hw)
 	/* Transmit Pause Packet */
 	iowrite32(PCH_GBE_PS_PKT_RQ, &hw->reg->PAUSE_REQ);
 
-	pr_debug("PAUSE_PKT1-5 reg : 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x\n",
-		 ioread32(&hw->reg->PAUSE_PKT1), ioread32(&hw->reg->PAUSE_PKT2),
-		 ioread32(&hw->reg->PAUSE_PKT3), ioread32(&hw->reg->PAUSE_PKT4),
-		 ioread32(&hw->reg->PAUSE_PKT5));
+	netdev_dbg(adapter->netdev,
+		   "PAUSE_PKT1-5 reg : 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x\n",
+		   ioread32(&hw->reg->PAUSE_PKT1),
+		   ioread32(&hw->reg->PAUSE_PKT2),
+		   ioread32(&hw->reg->PAUSE_PKT3),
+		   ioread32(&hw->reg->PAUSE_PKT4),
+		   ioread32(&hw->reg->PAUSE_PKT5));
 
 	return;
 }
@@ -624,15 +636,15 @@ static void pch_gbe_mac_set_pause_packet(struct pch_gbe_hw *hw)
  */
 static int pch_gbe_alloc_queues(struct pch_gbe_adapter *adapter)
 {
-	adapter->tx_ring = kzalloc(sizeof(*adapter->tx_ring), GFP_KERNEL);
+	adapter->tx_ring = devm_kzalloc(&adapter->pdev->dev,
+					sizeof(*adapter->tx_ring), GFP_KERNEL);
 	if (!adapter->tx_ring)
 		return -ENOMEM;
 
-	adapter->rx_ring = kzalloc(sizeof(*adapter->rx_ring), GFP_KERNEL);
-	if (!adapter->rx_ring) {
-		kfree(adapter->tx_ring);
+	adapter->rx_ring = devm_kzalloc(&adapter->pdev->dev,
+					sizeof(*adapter->rx_ring), GFP_KERNEL);
+	if (!adapter->rx_ring)
 		return -ENOMEM;
-	}
 	return 0;
 }
 
@@ -669,7 +681,7 @@ static int pch_gbe_init_phy(struct pch_gbe_adapter *adapter)
 			break;
 	}
 	adapter->hw.phy.addr = adapter->mii.phy_id;
-	pr_debug("phy_addr = %d\n", adapter->mii.phy_id);
+	netdev_dbg(netdev, "phy_addr = %d\n", adapter->mii.phy_id);
 	if (addr == 32)
 		return -EAGAIN;
 	/* Selected the phy and isolate the rest */
@@ -758,13 +770,15 @@ void pch_gbe_reinit_locked(struct pch_gbe_adapter *adapter)
  */
 void pch_gbe_reset(struct pch_gbe_adapter *adapter)
 {
+	struct net_device *netdev = adapter->netdev;
+
 	pch_gbe_mac_reset_hw(&adapter->hw);
 	/* reprogram multicast address register after reset */
-	pch_gbe_set_multi(adapter->netdev);
+	pch_gbe_set_multi(netdev);
 	/* Setup the receive address. */
 	pch_gbe_mac_init_rx_addrs(&adapter->hw, PCH_GBE_MAR_ENTRIES);
 	if (pch_gbe_hal_init_hw(&adapter->hw))
-		pr_err("Hardware Error\n");
+		netdev_err(netdev, "Hardware Error\n");
 }
 
 /**
@@ -778,7 +792,7 @@ static void pch_gbe_free_irq(struct pch_gbe_adapter *adapter)
 	free_irq(adapter->pdev->irq, netdev);
 	if (adapter->have_msi) {
 		pci_disable_msi(adapter->pdev);
-		pr_debug("call pci_disable_msi\n");
+		netdev_dbg(netdev, "call pci_disable_msi\n");
 	}
 }
 
@@ -795,7 +809,8 @@ static void pch_gbe_irq_disable(struct pch_gbe_adapter *adapter)
 	ioread32(&hw->reg->INT_ST);
 	synchronize_irq(adapter->pdev->irq);
 
-	pr_debug("INT_EN reg : 0x%08x\n", ioread32(&hw->reg->INT_EN));
+	netdev_dbg(adapter->netdev, "INT_EN reg : 0x%08x\n",
+		   ioread32(&hw->reg->INT_EN));
 }
 
 /**
@@ -809,7 +824,8 @@ static void pch_gbe_irq_enable(struct pch_gbe_adapter *adapter)
 	if (likely(atomic_dec_and_test(&adapter->irq_sem)))
 		iowrite32(PCH_GBE_INT_ENABLE_MASK, &hw->reg->INT_EN);
 	ioread32(&hw->reg->INT_ST);
-	pr_debug("INT_EN reg : 0x%08x\n", ioread32(&hw->reg->INT_EN));
+	netdev_dbg(adapter->netdev, "INT_EN reg : 0x%08x\n",
+		   ioread32(&hw->reg->INT_EN));
 }
 
 
@@ -846,9 +862,9 @@ static void pch_gbe_configure_tx(struct pch_gbe_adapter *adapter)
 	struct pch_gbe_hw *hw = &adapter->hw;
 	u32 tdba, tdlen, dctrl;
 
-	pr_debug("dma addr = 0x%08llx  size = 0x%08x\n",
-		 (unsigned long long)adapter->tx_ring->dma,
-		 adapter->tx_ring->size);
+	netdev_dbg(adapter->netdev, "dma addr = 0x%08llx  size = 0x%08x\n",
+		   (unsigned long long)adapter->tx_ring->dma,
+		   adapter->tx_ring->size);
 
 	/* Setup the HW Tx Head and Tail descriptor pointers */
 	tdba = adapter->tx_ring->dma;
@@ -894,9 +910,9 @@ static void pch_gbe_configure_rx(struct pch_gbe_adapter *adapter)
 	struct pch_gbe_hw *hw = &adapter->hw;
 	u32 rdba, rdlen, rxdma;
 
-	pr_debug("dma adr = 0x%08llx  size = 0x%08x\n",
-		 (unsigned long long)adapter->rx_ring->dma,
-		 adapter->rx_ring->size);
+	netdev_dbg(adapter->netdev, "dma adr = 0x%08llx  size = 0x%08x\n",
+		   (unsigned long long)adapter->rx_ring->dma,
+		   adapter->rx_ring->size);
 
 	pch_gbe_mac_force_mac_fc(hw);
 
@@ -907,9 +923,10 @@ static void pch_gbe_configure_rx(struct pch_gbe_adapter *adapter)
 	rxdma &= ~PCH_GBE_RX_DMA_EN;
 	iowrite32(rxdma, &hw->reg->DMA_CTRL);
 
-	pr_debug("MAC_RX_EN reg = 0x%08x  DMA_CTRL reg = 0x%08x\n",
-		 ioread32(&hw->reg->MAC_RX_EN),
-		 ioread32(&hw->reg->DMA_CTRL));
+	netdev_dbg(adapter->netdev,
+		   "MAC_RX_EN reg = 0x%08x  DMA_CTRL reg = 0x%08x\n",
+		   ioread32(&hw->reg->MAC_RX_EN),
+		   ioread32(&hw->reg->DMA_CTRL));
 
 	/* Setup the HW Rx Head and Tail Descriptor Pointers and
 	 * the Base and Length of the Rx Descriptor Ring */
@@ -977,7 +994,8 @@ static void pch_gbe_clean_tx_ring(struct pch_gbe_adapter *adapter,
 		buffer_info = &tx_ring->buffer_info[i];
 		pch_gbe_unmap_and_free_tx_resource(adapter, buffer_info);
 	}
-	pr_debug("call pch_gbe_unmap_and_free_tx_resource() %d count\n", i);
+	netdev_dbg(adapter->netdev,
+		   "call pch_gbe_unmap_and_free_tx_resource() %d count\n", i);
 
 	size = (unsigned long)sizeof(struct pch_gbe_buffer) * tx_ring->count;
 	memset(tx_ring->buffer_info, 0, size);
@@ -1009,7 +1027,8 @@ pch_gbe_clean_rx_ring(struct pch_gbe_adapter *adapter,
 		buffer_info = &rx_ring->buffer_info[i];
 		pch_gbe_unmap_and_free_rx_resource(adapter, buffer_info);
 	}
-	pr_debug("call pch_gbe_unmap_and_free_rx_resource() %d count\n", i);
+	netdev_dbg(adapter->netdev,
+		   "call pch_gbe_unmap_and_free_rx_resource() %d count\n", i);
 	size = (unsigned long)sizeof(struct pch_gbe_buffer) * rx_ring->count;
 	memset(rx_ring->buffer_info, 0, size);
 
@@ -1087,7 +1106,7 @@ static void pch_gbe_watchdog(unsigned long data)
 	struct net_device *netdev = adapter->netdev;
 	struct pch_gbe_hw *hw = &adapter->hw;
 
-	pr_debug("right now = %ld\n", jiffies);
+	netdev_dbg(netdev, "right now = %ld\n", jiffies);
 
 	pch_gbe_update_stats(adapter);
 	if ((mii_link_ok(&adapter->mii)) && (!netif_carrier_ok(netdev))) {
@@ -1095,7 +1114,7 @@ static void pch_gbe_watchdog(unsigned long data)
 		netdev->tx_queue_len = adapter->tx_queue_len;
 		/* mii library handles link maintenance tasks */
 		if (mii_ethtool_gset(&adapter->mii, &cmd)) {
-			pr_err("ethtool get setting Error\n");
+			netdev_err(netdev, "ethtool get setting Error\n");
 			mod_timer(&adapter->watchdog_timer,
 				  round_jiffies(jiffies +
 						PCH_GBE_WATCHDOG_PERIOD));
@@ -1213,7 +1232,7 @@ static void pch_gbe_tx_queue(struct pch_gbe_adapter *adapter,
 					  buffer_info->length,
 					  DMA_TO_DEVICE);
 	if (dma_mapping_error(&adapter->pdev->dev, buffer_info->dma)) {
-		pr_err("TX DMA map failed\n");
+		netdev_err(adapter->netdev, "TX DMA map failed\n");
 		buffer_info->dma = 0;
 		buffer_info->time_stamp = 0;
 		tx_ring->next_to_use = ring_num;
@@ -1333,13 +1352,13 @@ static irqreturn_t pch_gbe_intr(int irq, void *data)
 	/* When request status is no interruption factor */
 	if (unlikely(!int_st))
 		return IRQ_NONE;	/* Not our interrupt. End processing. */
-	pr_debug("%s occur int_st = 0x%08x\n", __func__, int_st);
+	netdev_dbg(netdev, "%s occur int_st = 0x%08x\n", __func__, int_st);
 	if (int_st & PCH_GBE_INT_RX_FRAME_ERR)
 		adapter->stats.intr_rx_frame_err_count++;
 	if (int_st & PCH_GBE_INT_RX_FIFO_ERR)
 		if (!adapter->rx_stop_flag) {
 			adapter->stats.intr_rx_fifo_err_count++;
-			pr_debug("Rx fifo over run\n");
+			netdev_dbg(netdev, "Rx fifo over run\n");
 			adapter->rx_stop_flag = true;
 			int_en = ioread32(&hw->reg->INT_EN);
 			iowrite32((int_en & ~PCH_GBE_INT_RX_FIFO_ERR),
@@ -1359,7 +1378,7 @@ static irqreturn_t pch_gbe_intr(int irq, void *data)
 	/* When Rx descriptor is empty  */
 	if ((int_st & PCH_GBE_INT_RX_DSC_EMP)) {
 		adapter->stats.intr_rx_dsc_empty_count++;
-		pr_debug("Rx descriptor is empty\n");
+		netdev_dbg(netdev, "Rx descriptor is empty\n");
 		int_en = ioread32(&hw->reg->INT_EN);
 		iowrite32((int_en & ~PCH_GBE_INT_RX_DSC_EMP), &hw->reg->INT_EN);
 		if (hw->mac.tx_fc_enable) {
@@ -1382,8 +1401,8 @@ static irqreturn_t pch_gbe_intr(int irq, void *data)
 			__napi_schedule(&adapter->napi);
 		}
 	}
-	pr_debug("return = 0x%08x  INT_EN reg = 0x%08x\n",
-		 IRQ_HANDLED, ioread32(&hw->reg->INT_EN));
+	netdev_dbg(netdev, "return = 0x%08x  INT_EN reg = 0x%08x\n",
+		   IRQ_HANDLED, ioread32(&hw->reg->INT_EN));
 	return IRQ_HANDLED;
 }
 
@@ -1437,9 +1456,10 @@ pch_gbe_alloc_rx_buffers(struct pch_gbe_adapter *adapter,
 		rx_desc->buffer_addr = (buffer_info->dma);
 		rx_desc->gbec_status = DSC_INIT16;
 
-		pr_debug("i = %d  buffer_info->dma = 0x08%llx  buffer_info->length = 0x%x\n",
-			 i, (unsigned long long)buffer_info->dma,
-			 buffer_info->length);
+		netdev_dbg(netdev,
+			   "i = %d  buffer_info->dma = 0x08%llx  buffer_info->length = 0x%x\n",
+			   i, (unsigned long long)buffer_info->dma,
+			   buffer_info->length);
 
 		if (unlikely(++i == rx_ring->count))
 			i = 0;
@@ -1531,12 +1551,13 @@ pch_gbe_clean_tx(struct pch_gbe_adapter *adapter,
 	bool cleaned = false;
 	int unused, thresh;
 
-	pr_debug("next_to_clean : %d\n", tx_ring->next_to_clean);
+	netdev_dbg(adapter->netdev, "next_to_clean : %d\n",
+		   tx_ring->next_to_clean);
 
 	i = tx_ring->next_to_clean;
 	tx_desc = PCH_GBE_TX_DESC(*tx_ring, i);
-	pr_debug("gbec_status:0x%04x  dma_status:0x%04x\n",
-		 tx_desc->gbec_status, tx_desc->dma_status);
+	netdev_dbg(adapter->netdev, "gbec_status:0x%04x  dma_status:0x%04x\n",
+		   tx_desc->gbec_status, tx_desc->dma_status);
 
 	unused = PCH_GBE_DESC_UNUSED(tx_ring);
 	thresh = tx_ring->count - PCH_GBE_TX_WEIGHT;
@@ -1544,8 +1565,10 @@ pch_gbe_clean_tx(struct pch_gbe_adapter *adapter,
 	{  /* current marked clean, tx queue filling up, do extra clean */
 		int j, k;
 		if (unused < 8) {  /* tx queue nearly full */
-			pr_debug("clean_tx: transmit queue warning (%x,%x) unused=%d\n",
-				tx_ring->next_to_clean,tx_ring->next_to_use,unused);
+			netdev_dbg(adapter->netdev,
+				   "clean_tx: transmit queue warning (%x,%x) unused=%d\n",
+				   tx_ring->next_to_clean, tx_ring->next_to_use,
+				   unused);
 		}
 
 		/* current marked clean, scan for more that need cleaning. */
@@ -1557,49 +1580,56 @@ pch_gbe_clean_tx(struct pch_gbe_adapter *adapter,
 			if (++k >= tx_ring->count) k = 0;  /*increment, wrap*/
 		}
 		if (j < PCH_GBE_TX_WEIGHT) {
-			pr_debug("clean_tx: unused=%d loops=%d found tx_desc[%x,%x:%x].gbec_status=%04x\n",
-				unused,j, i,k, tx_ring->next_to_use, tx_desc->gbec_status);
+			netdev_dbg(adapter->netdev,
+				   "clean_tx: unused=%d loops=%d found tx_desc[%x,%x:%x].gbec_status=%04x\n",
+				   unused, j, i, k, tx_ring->next_to_use,
+				   tx_desc->gbec_status);
 			i = k;  /*found one to clean, usu gbec_status==2000.*/
 		}
 	}
 
 	while ((tx_desc->gbec_status & DSC_INIT16) == 0x0000) {
-		pr_debug("gbec_status:0x%04x\n", tx_desc->gbec_status);
+		netdev_dbg(adapter->netdev, "gbec_status:0x%04x\n",
+			   tx_desc->gbec_status);
 		buffer_info = &tx_ring->buffer_info[i];
 		skb = buffer_info->skb;
 		cleaned = true;
 
 		if ((tx_desc->gbec_status & PCH_GBE_TXD_GMAC_STAT_ABT)) {
 			adapter->stats.tx_aborted_errors++;
-			pr_err("Transfer Abort Error\n");
+			netdev_err(adapter->netdev, "Transfer Abort Error\n");
 		} else if ((tx_desc->gbec_status & PCH_GBE_TXD_GMAC_STAT_CRSER)
 			  ) {
 			adapter->stats.tx_carrier_errors++;
-			pr_err("Transfer Carrier Sense Error\n");
+			netdev_err(adapter->netdev,
+				   "Transfer Carrier Sense Error\n");
 		} else if ((tx_desc->gbec_status & PCH_GBE_TXD_GMAC_STAT_EXCOL)
 			  ) {
 			adapter->stats.tx_aborted_errors++;
-			pr_err("Transfer Collision Abort Error\n");
+			netdev_err(adapter->netdev,
+				   "Transfer Collision Abort Error\n");
 		} else if ((tx_desc->gbec_status &
 			    (PCH_GBE_TXD_GMAC_STAT_SNGCOL |
 			     PCH_GBE_TXD_GMAC_STAT_MLTCOL))) {
 			adapter->stats.collisions++;
 			adapter->stats.tx_packets++;
 			adapter->stats.tx_bytes += skb->len;
-			pr_debug("Transfer Collision\n");
+			netdev_dbg(adapter->netdev, "Transfer Collision\n");
 		} else if ((tx_desc->gbec_status & PCH_GBE_TXD_GMAC_STAT_CMPLT)
 			  ) {
 			adapter->stats.tx_packets++;
 			adapter->stats.tx_bytes += skb->len;
 		}
 		if (buffer_info->mapped) {
-			pr_debug("unmap buffer_info->dma : %d\n", i);
+			netdev_dbg(adapter->netdev,
+				   "unmap buffer_info->dma : %d\n", i);
 			dma_unmap_single(&adapter->pdev->dev, buffer_info->dma,
 					 buffer_info->length, DMA_TO_DEVICE);
 			buffer_info->mapped = false;
 		}
 		if (buffer_info->skb) {
-			pr_debug("trim buffer_info->skb : %d\n", i);
+			netdev_dbg(adapter->netdev,
+				   "trim buffer_info->skb : %d\n", i);
 			skb_trim(buffer_info->skb, 0);
 		}
 		tx_desc->gbec_status = DSC_INIT16;
@@ -1613,8 +1643,9 @@ pch_gbe_clean_tx(struct pch_gbe_adapter *adapter,
 			break;
 		}
 	}
-	pr_debug("called pch_gbe_unmap_and_free_tx_resource() %d count\n",
-		 cleaned_count);
+	netdev_dbg(adapter->netdev,
+		   "called pch_gbe_unmap_and_free_tx_resource() %d count\n",
+		   cleaned_count);
 	if (cleaned_count > 0)  { /*skip this if nothing cleaned*/
 		/* Recover from running out of Tx resources in xmit_frame */
 		spin_lock(&tx_ring->tx_lock);
@@ -1622,12 +1653,13 @@ pch_gbe_clean_tx(struct pch_gbe_adapter *adapter,
 		{
 			netif_wake_queue(adapter->netdev);
 			adapter->stats.tx_restart_count++;
-			pr_debug("Tx wake queue\n");
+			netdev_dbg(adapter->netdev, "Tx wake queue\n");
 		}
 
 		tx_ring->next_to_clean = i;
 
-		pr_debug("next_to_clean : %d\n", tx_ring->next_to_clean);
+		netdev_dbg(adapter->netdev, "next_to_clean : %d\n",
+			   tx_ring->next_to_clean);
 		spin_unlock(&tx_ring->tx_lock);
 	}
 	return cleaned;
@@ -1684,22 +1716,22 @@ pch_gbe_clean_rx(struct pch_gbe_adapter *adapter,
 				   buffer_info->length, DMA_FROM_DEVICE);
 		buffer_info->mapped = false;
 
-		pr_debug("RxDecNo = 0x%04x  Status[DMA:0x%02x GBE:0x%04x "
-			 "TCP:0x%08x]  BufInf = 0x%p\n",
-			 i, dma_status, gbec_status, tcp_ip_status,
-			 buffer_info);
+		netdev_dbg(netdev,
+			   "RxDecNo = 0x%04x  Status[DMA:0x%02x GBE:0x%04x TCP:0x%08x]  BufInf = 0x%p\n",
+			   i, dma_status, gbec_status, tcp_ip_status,
+			   buffer_info);
 		/* Error check */
 		if (unlikely(gbec_status & PCH_GBE_RXD_GMAC_STAT_NOTOCTAL)) {
 			adapter->stats.rx_frame_errors++;
-			pr_err("Receive Not Octal Error\n");
+			netdev_err(netdev, "Receive Not Octal Error\n");
 		} else if (unlikely(gbec_status &
 				PCH_GBE_RXD_GMAC_STAT_NBLERR)) {
 			adapter->stats.rx_frame_errors++;
-			pr_err("Receive Nibble Error\n");
+			netdev_err(netdev, "Receive Nibble Error\n");
 		} else if (unlikely(gbec_status &
 				PCH_GBE_RXD_GMAC_STAT_CRCERR)) {
 			adapter->stats.rx_crc_errors++;
-			pr_err("Receive CRC Error\n");
+			netdev_err(netdev, "Receive CRC Error\n");
 		} else {
 			/* get receive length */
 			/* length convert[-3], length includes FCS length */
@@ -1730,8 +1762,9 @@ pch_gbe_clean_rx(struct pch_gbe_adapter *adapter,
 
 			napi_gro_receive(&adapter->napi, skb);
 			(*work_done)++;
-			pr_debug("Receive skb->ip_summed: %d length: %d\n",
-				 skb->ip_summed, length);
+			netdev_dbg(netdev,
+				   "Receive skb->ip_summed: %d length: %d\n",
+				   skb->ip_summed, length);
 		}
 		/* return some buffers to hardware, one at a time is too slow */
 		if (unlikely(cleaned_count >= PCH_GBE_RX_BUFFER_WRITE)) {
@@ -1787,10 +1820,10 @@ int pch_gbe_setup_tx_resources(struct pch_gbe_adapter *adapter,
 		tx_desc = PCH_GBE_TX_DESC(*tx_ring, desNo);
 		tx_desc->gbec_status = DSC_INIT16;
 	}
-	pr_debug("tx_ring->desc = 0x%p  tx_ring->dma = 0x%08llx\n"
-		 "next_to_clean = 0x%08x  next_to_use = 0x%08x\n",
-		 tx_ring->desc, (unsigned long long)tx_ring->dma,
-		 tx_ring->next_to_clean, tx_ring->next_to_use);
+	netdev_dbg(adapter->netdev,
+		   "tx_ring->desc = 0x%p  tx_ring->dma = 0x%08llx next_to_clean = 0x%08x  next_to_use = 0x%08x\n",
+		   tx_ring->desc, (unsigned long long)tx_ring->dma,
+		   tx_ring->next_to_clean, tx_ring->next_to_use);
 	return 0;
 }
 
@@ -1829,10 +1862,10 @@ int pch_gbe_setup_rx_resources(struct pch_gbe_adapter *adapter,
 		rx_desc = PCH_GBE_RX_DESC(*rx_ring, desNo);
 		rx_desc->gbec_status = DSC_INIT16;
 	}
-	pr_debug("rx_ring->desc = 0x%p  rx_ring->dma = 0x%08llx "
-		 "next_to_clean = 0x%08x  next_to_use = 0x%08x\n",
-		 rx_ring->desc, (unsigned long long)rx_ring->dma,
-		 rx_ring->next_to_clean, rx_ring->next_to_use);
+	netdev_dbg(adapter->netdev,
+		   "rx_ring->desc = 0x%p  rx_ring->dma = 0x%08llx next_to_clean = 0x%08x  next_to_use = 0x%08x\n",
+		   rx_ring->desc, (unsigned long long)rx_ring->dma,
+		   rx_ring->next_to_clean, rx_ring->next_to_use);
 	return 0;
 }
 
@@ -1886,9 +1919,9 @@ static int pch_gbe_request_irq(struct pch_gbe_adapter *adapter)
 	flags = IRQF_SHARED;
 	adapter->have_msi = false;
 	err = pci_enable_msi(adapter->pdev);
-	pr_debug("call pci_enable_msi\n");
+	netdev_dbg(netdev, "call pci_enable_msi\n");
 	if (err) {
-		pr_debug("call pci_enable_msi - Error: %d\n", err);
+		netdev_dbg(netdev, "call pci_enable_msi - Error: %d\n", err);
 	} else {
 		flags = 0;
 		adapter->have_msi = true;
@@ -1896,9 +1929,11 @@ static int pch_gbe_request_irq(struct pch_gbe_adapter *adapter)
 	err = request_irq(adapter->pdev->irq, &pch_gbe_intr,
 			  flags, netdev->name, netdev);
 	if (err)
-		pr_err("Unable to allocate interrupt Error: %d\n", err);
-	pr_debug("adapter->have_msi : %d  flags : 0x%04x  return : 0x%04x\n",
-		 adapter->have_msi, flags, err);
+		netdev_err(netdev, "Unable to allocate interrupt Error: %d\n",
+			   err);
+	netdev_dbg(netdev,
+		   "adapter->have_msi : %d  flags : 0x%04x  return : 0x%04x\n",
+		   adapter->have_msi, flags, err);
 	return err;
 }
 
@@ -1919,7 +1954,7 @@ int pch_gbe_up(struct pch_gbe_adapter *adapter)
 
 	/* Ensure we have a valid MAC */
 	if (!is_valid_ether_addr(adapter->hw.mac.addr)) {
-		pr_err("Error: Invalid MAC address\n");
+		netdev_err(netdev, "Error: Invalid MAC address\n");
 		goto out;
 	}
 
@@ -1933,12 +1968,14 @@ int pch_gbe_up(struct pch_gbe_adapter *adapter)
 
 	err = pch_gbe_request_irq(adapter);
 	if (err) {
-		pr_err("Error: can't bring device up - irq request failed\n");
+		netdev_err(netdev,
+			   "Error: can't bring device up - irq request failed\n");
 		goto out;
 	}
 	err = pch_gbe_alloc_rx_buffers_pool(adapter, rx_ring, rx_ring->count);
 	if (err) {
-		pr_err("Error: can't bring device up - alloc rx buffers pool failed\n");
+		netdev_err(netdev,
+			   "Error: can't bring device up - alloc rx buffers pool failed\n");
 		goto freeirq;
 	}
 	pch_gbe_alloc_tx_buffers(adapter, tx_ring);
@@ -2015,11 +2052,11 @@ static int pch_gbe_sw_init(struct pch_gbe_adapter *adapter)
 
 	/* Initialize the hardware-specific values */
 	if (pch_gbe_hal_setup_init_funcs(hw)) {
-		pr_err("Hardware Initialization Failure\n");
+		netdev_err(netdev, "Hardware Initialization Failure\n");
 		return -EIO;
 	}
 	if (pch_gbe_alloc_queues(adapter)) {
-		pr_err("Unable to allocate memory for queues\n");
+		netdev_err(netdev, "Unable to allocate memory for queues\n");
 		return -ENOMEM;
 	}
 	spin_lock_init(&adapter->hw.miim_lock);
@@ -2030,9 +2067,10 @@ static int pch_gbe_sw_init(struct pch_gbe_adapter *adapter)
 
 	pch_gbe_init_stats(adapter);
 
-	pr_debug("rx_buffer_len : %d  mac.min_frame_size : %d  mac.max_frame_size : %d\n",
-		 (u32) adapter->rx_buffer_len,
-		 hw->mac.min_frame_size, hw->mac.max_frame_size);
+	netdev_dbg(netdev,
+		   "rx_buffer_len : %d  mac.min_frame_size : %d  mac.max_frame_size : %d\n",
+		   (u32) adapter->rx_buffer_len,
+		   hw->mac.min_frame_size, hw->mac.max_frame_size);
 	return 0;
 }
 
@@ -2061,7 +2099,7 @@ static int pch_gbe_open(struct net_device *netdev)
 	err = pch_gbe_up(adapter);
 	if (err)
 		goto err_up;
-	pr_debug("Success End\n");
+	netdev_dbg(netdev, "Success End\n");
 	return 0;
 
 err_up:
@@ -2072,7 +2110,7 @@ err_setup_rx:
 	pch_gbe_free_tx_resources(adapter, adapter->tx_ring);
 err_setup_tx:
 	pch_gbe_reset(adapter);
-	pr_err("Error End\n");
+	netdev_err(netdev, "Error End\n");
 	return err;
 }
 
@@ -2116,8 +2154,9 @@ static int pch_gbe_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 	if (unlikely(!PCH_GBE_DESC_UNUSED(tx_ring))) {
 		netif_stop_queue(netdev);
 		spin_unlock_irqrestore(&tx_ring->tx_lock, flags);
-		pr_debug("Return : BUSY  next_to use : 0x%08x  next_to clean : 0x%08x\n",
-			 tx_ring->next_to_use, tx_ring->next_to_clean);
+		netdev_dbg(netdev,
+			   "Return : BUSY  next_to use : 0x%08x  next_to clean : 0x%08x\n",
+			   tx_ring->next_to_use, tx_ring->next_to_clean);
 		return NETDEV_TX_BUSY;
 	}
 
@@ -2152,7 +2191,7 @@ static void pch_gbe_set_multi(struct net_device *netdev)
 	int i;
 	int mc_count;
 
-	pr_debug("netdev->flags : 0x%08x\n", netdev->flags);
+	netdev_dbg(netdev, "netdev->flags : 0x%08x\n", netdev->flags);
 
 	/* Check for Promiscuous and All Multicast modes */
 	rctl = ioread32(&hw->reg->RX_MODE);
@@ -2192,7 +2231,8 @@ static void pch_gbe_set_multi(struct net_device *netdev)
 					PCH_GBE_MAR_ENTRIES);
 	kfree(mta_list);
 
-	pr_debug("RX_MODE reg(check bit31,30 ADD,MLT) : 0x%08x  netdev->mc_count : 0x%08x\n",
+	netdev_dbg(netdev,
+		 "RX_MODE reg(check bit31,30 ADD,MLT) : 0x%08x  netdev->mc_count : 0x%08x\n",
 		 ioread32(&hw->reg->RX_MODE), mc_count);
 }
 
@@ -2218,12 +2258,12 @@ static int pch_gbe_set_mac(struct net_device *netdev, void *addr)
 		pch_gbe_mac_mar_set(&adapter->hw, adapter->hw.mac.addr, 0);
 		ret_val = 0;
 	}
-	pr_debug("ret_val : 0x%08x\n", ret_val);
-	pr_debug("dev_addr : %pM\n", netdev->dev_addr);
-	pr_debug("mac_addr : %pM\n", adapter->hw.mac.addr);
-	pr_debug("MAC_ADR1AB reg : 0x%08x 0x%08x\n",
-		 ioread32(&adapter->hw.reg->mac_adr[0].high),
-		 ioread32(&adapter->hw.reg->mac_adr[0].low));
+	netdev_dbg(netdev, "ret_val : 0x%08x\n", ret_val);
+	netdev_dbg(netdev, "dev_addr : %pM\n", netdev->dev_addr);
+	netdev_dbg(netdev, "mac_addr : %pM\n", adapter->hw.mac.addr);
+	netdev_dbg(netdev, "MAC_ADR1AB reg : 0x%08x 0x%08x\n",
+		   ioread32(&adapter->hw.reg->mac_adr[0].high),
+		   ioread32(&adapter->hw.reg->mac_adr[0].low));
 	return ret_val;
 }
 
@@ -2245,7 +2285,7 @@ static int pch_gbe_change_mtu(struct net_device *netdev, int new_mtu)
 	max_frame = new_mtu + ETH_HLEN + ETH_FCS_LEN;
 	if ((max_frame < ETH_ZLEN + ETH_FCS_LEN) ||
 		(max_frame > PCH_GBE_MAX_JUMBO_FRAME_SIZE)) {
-		pr_err("Invalid MTU setting\n");
+		netdev_err(netdev, "Invalid MTU setting\n");
 		return -EINVAL;
 	}
 	if (max_frame <= PCH_GBE_FRAME_SIZE_2048)
@@ -2274,9 +2314,10 @@ static int pch_gbe_change_mtu(struct net_device *netdev, int new_mtu)
 		adapter->hw.mac.max_frame_size = max_frame;
 	}
 
-	pr_debug("max_frame : %d  rx_buffer_len : %d  mtu : %d  max_frame_size : %d\n",
-		 max_frame, (u32) adapter->rx_buffer_len, netdev->mtu,
-		 adapter->hw.mac.max_frame_size);
+	netdev_dbg(netdev,
+		   "max_frame : %d  rx_buffer_len : %d  mtu : %d  max_frame_size : %d\n",
+		   max_frame, (u32) adapter->rx_buffer_len, netdev->mtu,
+		   adapter->hw.mac.max_frame_size);
 	return 0;
 }
 
@@ -2317,7 +2358,7 @@ static int pch_gbe_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
 {
 	struct pch_gbe_adapter *adapter = netdev_priv(netdev);
 
-	pr_debug("cmd : 0x%04x\n", cmd);
+	netdev_dbg(netdev, "cmd : 0x%04x\n", cmd);
 
 	if (cmd == SIOCSHWTSTAMP)
 		return hwtstamp_ioctl(netdev, ifr, cmd);
@@ -2354,7 +2395,7 @@ static int pch_gbe_napi_poll(struct napi_struct *napi, int budget)
 	bool poll_end_flag = false;
 	bool cleaned = false;
 
-	pr_debug("budget : %d\n", budget);
+	netdev_dbg(adapter->netdev, "budget : %d\n", budget);
 
 	pch_gbe_clean_rx(adapter, adapter->rx_ring, &work_done, budget);
 	cleaned = pch_gbe_clean_tx(adapter, adapter->tx_ring);
@@ -2377,8 +2418,9 @@ static int pch_gbe_napi_poll(struct napi_struct *napi, int budget)
 		pch_gbe_enable_dma_rx(&adapter->hw);
 	}
 
-	pr_debug("poll_end_flag : %d  work_done : %d  budget : %d\n",
-		 poll_end_flag, work_done, budget);
+	netdev_dbg(adapter->netdev,
+		   "poll_end_flag : %d  work_done : %d  budget : %d\n",
+		   poll_end_flag, work_done, budget);
 
 	return work_done;
 }
@@ -2435,7 +2477,7 @@ static pci_ers_result_t pch_gbe_io_slot_reset(struct pci_dev *pdev)
 	struct pch_gbe_hw *hw = &adapter->hw;
 
 	if (pci_enable_device(pdev)) {
-		pr_err("Cannot re-enable PCI device after reset\n");
+		netdev_err(netdev, "Cannot re-enable PCI device after reset\n");
 		return PCI_ERS_RESULT_DISCONNECT;
 	}
 	pci_set_master(pdev);
@@ -2455,7 +2497,8 @@ static void pch_gbe_io_resume(struct pci_dev *pdev)
 
 	if (netif_running(netdev)) {
 		if (pch_gbe_up(adapter)) {
-			pr_debug("can't bring device back up after reset\n");
+			netdev_dbg(netdev,
+				   "can't bring device back up after reset\n");
 			return;
 		}
 	}
@@ -2509,7 +2552,7 @@ static int pch_gbe_resume(struct device *device)
 
 	err = pci_enable_device(pdev);
 	if (err) {
-		pr_err("Cannot enable PCI device from suspend\n");
+		netdev_err(netdev, "Cannot enable PCI device from suspend\n");
 		return err;
 	}
 	pci_set_master(pdev);
@@ -2545,13 +2588,7 @@ static void pch_gbe_remove(struct pci_dev *pdev)
 
 	pch_gbe_hal_phy_hw_reset(&adapter->hw);
 
-	kfree(adapter->tx_ring);
-	kfree(adapter->rx_ring);
-
-	iounmap(adapter->hw.reg);
-	pci_release_regions(pdev);
 	free_netdev(netdev);
-	pci_disable_device(pdev);
 }
 
 static int pch_gbe_probe(struct pci_dev *pdev,
@@ -2561,7 +2598,7 @@ static int pch_gbe_probe(struct pci_dev *pdev,
 	struct pch_gbe_adapter *adapter;
 	int ret;
 
-	ret = pci_enable_device(pdev);
+	ret = pcim_enable_device(pdev);
 	if (ret)
 		return ret;
 
@@ -2574,24 +2611,22 @@ static int pch_gbe_probe(struct pci_dev *pdev,
 			if (ret) {
 				dev_err(&pdev->dev, "ERR: No usable DMA "
 					"configuration, aborting\n");
-				goto err_disable_device;
+				return ret;
 			}
 		}
 	}
 
-	ret = pci_request_regions(pdev, KBUILD_MODNAME);
+	ret = pcim_iomap_regions(pdev, 1 << PCH_GBE_PCI_BAR, pci_name(pdev));
 	if (ret) {
 		dev_err(&pdev->dev,
 			"ERR: Can't reserve PCI I/O and memory resources\n");
-		goto err_disable_device;
+		return ret;
 	}
 	pci_set_master(pdev);
 
 	netdev = alloc_etherdev((int)sizeof(struct pch_gbe_adapter));
-	if (!netdev) {
-		ret = -ENOMEM;
-		goto err_release_pci;
-	}
+	if (!netdev)
+		return -ENOMEM;
 	SET_NETDEV_DEV(netdev, &pdev->dev);
 
 	pci_set_drvdata(pdev, netdev);
@@ -2599,18 +2634,14 @@ static int pch_gbe_probe(struct pci_dev *pdev,
 	adapter->netdev = netdev;
 	adapter->pdev = pdev;
 	adapter->hw.back = adapter;
-	adapter->hw.reg = pci_iomap(pdev, PCH_GBE_PCI_BAR, 0);
-	if (!adapter->hw.reg) {
-		ret = -EIO;
-		dev_err(&pdev->dev, "Can't ioremap\n");
-		goto err_free_netdev;
-	}
+	adapter->hw.reg = pcim_iomap_table(pdev)[PCH_GBE_PCI_BAR];
 
 	adapter->ptp_pdev = pci_get_bus_and_slot(adapter->pdev->bus->number,
 					       PCI_DEVFN(12, 4));
 	if (ptp_filter_init(ptp_filter, ARRAY_SIZE(ptp_filter))) {
-		pr_err("Bad ptp filter\n");
-		return -EINVAL;
+		dev_err(&pdev->dev, "Bad ptp filter\n");
+		ret = -EINVAL;
+		goto err_free_netdev;
 	}
 
 	netdev->netdev_ops = &pch_gbe_netdev_ops;
@@ -2628,7 +2659,7 @@ static int pch_gbe_probe(struct pci_dev *pdev,
 	/* setup the private structure */
 	ret = pch_gbe_sw_init(adapter);
 	if (ret)
-		goto err_iounmap;
+		goto err_free_netdev;
 
 	/* Initialize PHY */
 	ret = pch_gbe_init_phy(adapter);
@@ -2684,16 +2715,8 @@ static int pch_gbe_probe(struct pci_dev *pdev,
 
 err_free_adapter:
 	pch_gbe_hal_phy_hw_reset(&adapter->hw);
-	kfree(adapter->tx_ring);
-	kfree(adapter->rx_ring);
-err_iounmap:
-	iounmap(adapter->hw.reg);
 err_free_netdev:
 	free_netdev(netdev);
-err_release_pci:
-	pci_release_regions(pdev);
-err_disable_device:
-	pci_disable_device(pdev);
 	return ret;
 }
 
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_param.c b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_param.c
index 8653c3b..cf7c9b3 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_param.c
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_param.c
@@ -237,16 +237,17 @@ static int pch_gbe_validate_option(int *value,
 	case enable_option:
 		switch (*value) {
 		case OPTION_ENABLED:
-			pr_debug("%s Enabled\n", opt->name);
+			netdev_dbg(adapter->netdev, "%s Enabled\n", opt->name);
 			return 0;
 		case OPTION_DISABLED:
-			pr_debug("%s Disabled\n", opt->name);
+			netdev_dbg(adapter->netdev, "%s Disabled\n", opt->name);
 			return 0;
 		}
 		break;
 	case range_option:
 		if (*value >= opt->arg.r.min && *value <= opt->arg.r.max) {
-			pr_debug("%s set to %i\n", opt->name, *value);
+			netdev_dbg(adapter->netdev, "%s set to %i\n",
+				   opt->name, *value);
 			return 0;
 		}
 		break;
@@ -258,7 +259,8 @@ static int pch_gbe_validate_option(int *value,
 			ent = &opt->arg.l.p[i];
 			if (*value == ent->i) {
 				if (ent->str[0] != '\0')
-					pr_debug("%s\n", ent->str);
+					netdev_dbg(adapter->netdev, "%s\n",
+						   ent->str);
 				return 0;
 			}
 		}
@@ -268,8 +270,8 @@ static int pch_gbe_validate_option(int *value,
 		BUG();
 	}
 
-	pr_debug("Invalid %s value specified (%i) %s\n",
-		 opt->name, *value, opt->err);
+	netdev_dbg(adapter->netdev, "Invalid %s value specified (%i) %s\n",
+		   opt->name, *value, opt->err);
 	*value = opt->def;
 	return -1;
 }
@@ -318,7 +320,8 @@ static void pch_gbe_check_copper_options(struct pch_gbe_adapter *adapter)
 					 .p = an_list} }
 		};
 		if (speed || dplx) {
-			pr_debug("AutoNeg specified along with Speed or Duplex, AutoNeg parameter ignored\n");
+			netdev_dbg(adapter->netdev,
+				   "AutoNeg specified along with Speed or Duplex, AutoNeg parameter ignored\n");
 			hw->phy.autoneg_advertised = opt.def;
 		} else {
 			int tmp = AutoNeg;
@@ -332,13 +335,16 @@ static void pch_gbe_check_copper_options(struct pch_gbe_adapter *adapter)
 	case 0:
 		hw->mac.autoneg = hw->mac.fc_autoneg = 1;
 		if ((speed || dplx))
-			pr_debug("Speed and duplex autonegotiation enabled\n");
+			netdev_dbg(adapter->netdev,
+				   "Speed and duplex autonegotiation enabled\n");
 		hw->mac.link_speed = SPEED_10;
 		hw->mac.link_duplex = DUPLEX_HALF;
 		break;
 	case HALF_DUPLEX:
-		pr_debug("Half Duplex specified without Speed\n");
-		pr_debug("Using Autonegotiation at Half Duplex only\n");
+		netdev_dbg(adapter->netdev,
+			   "Half Duplex specified without Speed\n");
+		netdev_dbg(adapter->netdev,
+			   "Using Autonegotiation at Half Duplex only\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 1;
 		hw->phy.autoneg_advertised = PHY_ADVERTISE_10_HALF |
 						PHY_ADVERTISE_100_HALF;
@@ -346,8 +352,10 @@ static void pch_gbe_check_copper_options(struct pch_gbe_adapter *adapter)
 		hw->mac.link_duplex = DUPLEX_HALF;
 		break;
 	case FULL_DUPLEX:
-		pr_debug("Full Duplex specified without Speed\n");
-		pr_debug("Using Autonegotiation at Full Duplex only\n");
+		netdev_dbg(adapter->netdev,
+			   "Full Duplex specified without Speed\n");
+		netdev_dbg(adapter->netdev,
+			   "Using Autonegotiation at Full Duplex only\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 1;
 		hw->phy.autoneg_advertised = PHY_ADVERTISE_10_FULL |
 						PHY_ADVERTISE_100_FULL |
@@ -356,8 +364,10 @@ static void pch_gbe_check_copper_options(struct pch_gbe_adapter *adapter)
 		hw->mac.link_duplex = DUPLEX_FULL;
 		break;
 	case SPEED_10:
-		pr_debug("10 Mbps Speed specified without Duplex\n");
-		pr_debug("Using Autonegotiation at 10 Mbps only\n");
+		netdev_dbg(adapter->netdev,
+			   "10 Mbps Speed specified without Duplex\n");
+		netdev_dbg(adapter->netdev,
+			   "Using Autonegotiation at 10 Mbps only\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 1;
 		hw->phy.autoneg_advertised = PHY_ADVERTISE_10_HALF |
 						PHY_ADVERTISE_10_FULL;
@@ -365,22 +375,24 @@ static void pch_gbe_check_copper_options(struct pch_gbe_adapter *adapter)
 		hw->mac.link_duplex = DUPLEX_HALF;
 		break;
 	case SPEED_10 + HALF_DUPLEX:
-		pr_debug("Forcing to 10 Mbps Half Duplex\n");
+		netdev_dbg(adapter->netdev, "Forcing to 10 Mbps Half Duplex\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 0;
 		hw->phy.autoneg_advertised = 0;
 		hw->mac.link_speed = SPEED_10;
 		hw->mac.link_duplex = DUPLEX_HALF;
 		break;
 	case SPEED_10 + FULL_DUPLEX:
-		pr_debug("Forcing to 10 Mbps Full Duplex\n");
+		netdev_dbg(adapter->netdev, "Forcing to 10 Mbps Full Duplex\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 0;
 		hw->phy.autoneg_advertised = 0;
 		hw->mac.link_speed = SPEED_10;
 		hw->mac.link_duplex = DUPLEX_FULL;
 		break;
 	case SPEED_100:
-		pr_debug("100 Mbps Speed specified without Duplex\n");
-		pr_debug("Using Autonegotiation at 100 Mbps only\n");
+		netdev_dbg(adapter->netdev,
+			   "100 Mbps Speed specified without Duplex\n");
+		netdev_dbg(adapter->netdev,
+			   "Using Autonegotiation at 100 Mbps only\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 1;
 		hw->phy.autoneg_advertised = PHY_ADVERTISE_100_HALF |
 						PHY_ADVERTISE_100_FULL;
@@ -388,28 +400,33 @@ static void pch_gbe_check_copper_options(struct pch_gbe_adapter *adapter)
 		hw->mac.link_duplex = DUPLEX_HALF;
 		break;
 	case SPEED_100 + HALF_DUPLEX:
-		pr_debug("Forcing to 100 Mbps Half Duplex\n");
+		netdev_dbg(adapter->netdev,
+			   "Forcing to 100 Mbps Half Duplex\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 0;
 		hw->phy.autoneg_advertised = 0;
 		hw->mac.link_speed = SPEED_100;
 		hw->mac.link_duplex = DUPLEX_HALF;
 		break;
 	case SPEED_100 + FULL_DUPLEX:
-		pr_debug("Forcing to 100 Mbps Full Duplex\n");
+		netdev_dbg(adapter->netdev,
+			   "Forcing to 100 Mbps Full Duplex\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 0;
 		hw->phy.autoneg_advertised = 0;
 		hw->mac.link_speed = SPEED_100;
 		hw->mac.link_duplex = DUPLEX_FULL;
 		break;
 	case SPEED_1000:
-		pr_debug("1000 Mbps Speed specified without Duplex\n");
+		netdev_dbg(adapter->netdev,
+			   "1000 Mbps Speed specified without Duplex\n");
 		goto full_duplex_only;
 	case SPEED_1000 + HALF_DUPLEX:
-		pr_debug("Half Duplex is not supported at 1000 Mbps\n");
+		netdev_dbg(adapter->netdev,
+			   "Half Duplex is not supported at 1000 Mbps\n");
 		/* fall through */
 	case SPEED_1000 + FULL_DUPLEX:
 full_duplex_only:
-		pr_debug("Using Autonegotiation at 1000 Mbps Full Duplex only\n");
+		netdev_dbg(adapter->netdev,
+			   "Using Autonegotiation at 1000 Mbps Full Duplex only\n");
 		hw->mac.autoneg = hw->mac.fc_autoneg = 1;
 		hw->phy.autoneg_advertised = PHY_ADVERTISE_1000_FULL;
 		hw->mac.link_speed = SPEED_1000;
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_phy.c b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_phy.c
index 28bb960..da07907 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_phy.c
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_phy.c
@@ -97,6 +97,7 @@
  */
 s32 pch_gbe_phy_get_id(struct pch_gbe_hw *hw)
 {
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	struct pch_gbe_phy_info *phy = &hw->phy;
 	s32 ret;
 	u16 phy_id1;
@@ -115,8 +116,9 @@ s32 pch_gbe_phy_get_id(struct pch_gbe_hw *hw)
 	phy->id = (u32)phy_id1;
 	phy->id = ((phy->id << 6) | ((phy_id2 & 0xFC00) >> 10));
 	phy->revision = (u32) (phy_id2 & 0x000F);
-	pr_debug("phy->id : 0x%08x  phy->revision : 0x%08x\n",
-		 phy->id, phy->revision);
+	netdev_dbg(adapter->netdev,
+		   "phy->id : 0x%08x  phy->revision : 0x%08x\n",
+		   phy->id, phy->revision);
 	return 0;
 }
 
@@ -134,7 +136,10 @@ s32 pch_gbe_phy_read_reg_miic(struct pch_gbe_hw *hw, u32 offset, u16 *data)
 	struct pch_gbe_phy_info *phy = &hw->phy;
 
 	if (offset > PHY_MAX_REG_ADDRESS) {
-		pr_err("PHY Address %d is out of range\n", offset);
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "PHY Address %d is out of range\n",
+			   offset);
 		return -EINVAL;
 	}
 	*data = pch_gbe_mac_ctrl_miim(hw, phy->addr, PCH_GBE_HAL_MIIM_READ,
@@ -156,7 +161,10 @@ s32 pch_gbe_phy_write_reg_miic(struct pch_gbe_hw *hw, u32 offset, u16 data)
 	struct pch_gbe_phy_info *phy = &hw->phy;
 
 	if (offset > PHY_MAX_REG_ADDRESS) {
-		pr_err("PHY Address %d is out of range\n", offset);
+		struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
+
+		netdev_err(adapter->netdev, "PHY Address %d is out of range\n",
+			   offset);
 		return -EINVAL;
 	}
 	pch_gbe_mac_ctrl_miim(hw, phy->addr, PCH_GBE_HAL_MIIM_WRITE,
@@ -235,7 +243,7 @@ void pch_gbe_phy_power_down(struct pch_gbe_hw *hw)
  * pch_gbe_phy_set_rgmii - RGMII interface setting
  * @hw:	            Pointer to the HW structure
  */
-inline void pch_gbe_phy_set_rgmii(struct pch_gbe_hw *hw)
+void pch_gbe_phy_set_rgmii(struct pch_gbe_hw *hw)
 {
 	pch_gbe_phy_sw_reset(hw);
 }
@@ -246,15 +254,14 @@ inline void pch_gbe_phy_set_rgmii(struct pch_gbe_hw *hw)
  */
 void pch_gbe_phy_init_setting(struct pch_gbe_hw *hw)
 {
-	struct pch_gbe_adapter *adapter;
+	struct pch_gbe_adapter *adapter = pch_gbe_hw_to_adapter(hw);
 	struct ethtool_cmd     cmd = { .cmd = ETHTOOL_GSET };
 	int ret;
 	u16 mii_reg;
 
-	adapter = container_of(hw, struct pch_gbe_adapter, hw);
 	ret = mii_ethtool_gset(&adapter->mii, &cmd);
 	if (ret)
-		pr_err("Error: mii_ethtool_gset\n");
+		netdev_err(adapter->netdev, "Error: mii_ethtool_gset\n");
 
 	ethtool_cmd_speed_set(&cmd, hw->mac.link_speed);
 	cmd.duplex = hw->mac.link_duplex;
@@ -263,12 +270,11 @@ void pch_gbe_phy_init_setting(struct pch_gbe_hw *hw)
 	pch_gbe_phy_write_reg_miic(hw, MII_BMCR, BMCR_RESET);
 	ret = mii_ethtool_sset(&adapter->mii, &cmd);
 	if (ret)
-		pr_err("Error: mii_ethtool_sset\n");
+		netdev_err(adapter->netdev, "Error: mii_ethtool_sset\n");
 
 	pch_gbe_phy_sw_reset(hw);
 
 	pch_gbe_phy_read_reg_miic(hw, PHY_PHYSP_CONTROL, &mii_reg);
 	mii_reg |= PHYSP_CTRL_ASSERT_CRS_TX;
 	pch_gbe_phy_write_reg_miic(hw, PHY_PHYSP_CONTROL, mii_reg);
-
 }
diff --git a/drivers/net/ethernet/packetengines/Kconfig b/drivers/net/ethernet/packetengines/Kconfig
index cbbeca3..8d51800 100644
--- a/drivers/net/ethernet/packetengines/Kconfig
+++ b/drivers/net/ethernet/packetengines/Kconfig
@@ -21,7 +21,6 @@ if NET_PACKET_ENGINE
 config HAMACHI
 	tristate "Packet Engines Hamachi GNIC-II support"
 	depends on PCI
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a Gigabit Ethernet card of this type, say Y and read
diff --git a/drivers/net/ethernet/qlogic/netxen/netxen_nic.h b/drivers/net/ethernet/qlogic/netxen/netxen_nic.h
index 322a36b..3fe09ab 100644
--- a/drivers/net/ethernet/qlogic/netxen/netxen_nic.h
+++ b/drivers/net/ethernet/qlogic/netxen/netxen_nic.h
@@ -53,8 +53,8 @@
 
 #define _NETXEN_NIC_LINUX_MAJOR 4
 #define _NETXEN_NIC_LINUX_MINOR 0
-#define _NETXEN_NIC_LINUX_SUBVERSION 80
-#define NETXEN_NIC_LINUX_VERSIONID  "4.0.80"
+#define _NETXEN_NIC_LINUX_SUBVERSION 81
+#define NETXEN_NIC_LINUX_VERSIONID  "4.0.81"
 
 #define NETXEN_VERSION_CODE(a, b, c)	(((a) << 24) + ((b) << 16) + (c))
 #define _major(v)	(((v) >> 24) & 0xff)
@@ -1855,7 +1855,7 @@ static const struct netxen_brdinfo netxen_boards[] = {
 
 #define NUM_SUPPORTED_BOARDS ARRAY_SIZE(netxen_boards)
 
-static inline void get_brd_name_by_type(u32 type, char *name)
+static inline int netxen_nic_get_brd_name_by_type(u32 type, char *name)
 {
 	int i, found = 0;
 	for (i = 0; i < NUM_SUPPORTED_BOARDS; ++i) {
@@ -1864,10 +1864,14 @@ static inline void get_brd_name_by_type(u32 type, char *name)
 			found = 1;
 			break;
 		}
+	}
 
+	if (!found) {
+		strcpy(name, "Unknown");
+		return -EINVAL;
 	}
-	if (!found)
-		name = "Unknown";
+
+	return 0;
 }
 
 static inline u32 netxen_tx_avail(struct nx_host_tx_ring *tx_ring)
diff --git a/drivers/net/ethernet/qlogic/netxen/netxen_nic_hdr.h b/drivers/net/ethernet/qlogic/netxen/netxen_nic_hdr.h
index 28e0769..32c7906 100644
--- a/drivers/net/ethernet/qlogic/netxen/netxen_nic_hdr.h
+++ b/drivers/net/ethernet/qlogic/netxen/netxen_nic_hdr.h
@@ -734,6 +734,9 @@ enum {
 #define NIC_CRB_BASE_2		(NETXEN_CAM_RAM(0x700))
 #define NETXEN_NIC_REG(X)	(NIC_CRB_BASE+(X))
 #define NETXEN_NIC_REG_2(X)	(NIC_CRB_BASE_2+(X))
+#define NETXEN_INTR_MODE_REG	NETXEN_NIC_REG(0x44)
+#define NETXEN_MSI_MODE		0x1
+#define NETXEN_INTX_MODE	0x2
 
 #define NX_CDRP_CRB_OFFSET		(NETXEN_NIC_REG(0x18))
 #define NX_ARG1_CRB_OFFSET		(NETXEN_NIC_REG(0x1c))
diff --git a/drivers/net/ethernet/qlogic/netxen/netxen_nic_main.c b/drivers/net/ethernet/qlogic/netxen/netxen_nic_main.c
index af951f3..c401b0b 100644
--- a/drivers/net/ethernet/qlogic/netxen/netxen_nic_main.c
+++ b/drivers/net/ethernet/qlogic/netxen/netxen_nic_main.c
@@ -592,48 +592,60 @@ static const struct net_device_ops netxen_netdev_ops = {
 #endif
 };
 
-static void
-netxen_setup_intr(struct netxen_adapter *adapter)
+static inline bool netxen_function_zero(struct pci_dev *pdev)
 {
-	struct netxen_legacy_intr_set *legacy_intrp;
-	struct pci_dev *pdev = adapter->pdev;
-	int err, num_msix;
+	return (PCI_FUNC(pdev->devfn) == 0) ? true : false;
+}
 
-	if (adapter->rss_supported) {
-		num_msix = (num_online_cpus() >= MSIX_ENTRIES_PER_ADAPTER) ?
-			MSIX_ENTRIES_PER_ADAPTER : 2;
-	} else
-		num_msix = 1;
+static inline void netxen_set_interrupt_mode(struct netxen_adapter *adapter,
+					     u32 mode)
+{
+	NXWR32(adapter, NETXEN_INTR_MODE_REG, mode);
+}
 
-	adapter->max_sds_rings = 1;
+static inline u32 netxen_get_interrupt_mode(struct netxen_adapter *adapter)
+{
+	return NXRD32(adapter, NETXEN_INTR_MODE_REG);
+}
 
-	adapter->flags &= ~(NETXEN_NIC_MSI_ENABLED | NETXEN_NIC_MSIX_ENABLED);
+static void
+netxen_initialize_interrupt_registers(struct netxen_adapter *adapter)
+{
+	struct netxen_legacy_intr_set *legacy_intrp;
+	u32 tgt_status_reg, int_state_reg;
 
 	if (adapter->ahw.revision_id >= NX_P3_B0)
 		legacy_intrp = &legacy_intr[adapter->ahw.pci_func];
 	else
 		legacy_intrp = &legacy_intr[0];
 
+	tgt_status_reg = legacy_intrp->tgt_status_reg;
+	int_state_reg = ISR_INT_STATE_REG;
+
 	adapter->int_vec_bit = legacy_intrp->int_vec_bit;
-	adapter->tgt_status_reg = netxen_get_ioaddr(adapter,
-			legacy_intrp->tgt_status_reg);
+	adapter->tgt_status_reg = netxen_get_ioaddr(adapter, tgt_status_reg);
 	adapter->tgt_mask_reg = netxen_get_ioaddr(adapter,
-			legacy_intrp->tgt_mask_reg);
+						  legacy_intrp->tgt_mask_reg);
 	adapter->pci_int_reg = netxen_get_ioaddr(adapter,
-			legacy_intrp->pci_int_reg);
+						 legacy_intrp->pci_int_reg);
 	adapter->isr_int_vec = netxen_get_ioaddr(adapter, ISR_INT_VECTOR);
 
 	if (adapter->ahw.revision_id >= NX_P3_B1)
 		adapter->crb_int_state_reg = netxen_get_ioaddr(adapter,
-			ISR_INT_STATE_REG);
+							       int_state_reg);
 	else
 		adapter->crb_int_state_reg = netxen_get_ioaddr(adapter,
-			CRB_INT_VECTOR);
+							       CRB_INT_VECTOR);
+}
 
-	netxen_set_msix_bit(pdev, 0);
+static int netxen_setup_msi_interrupts(struct netxen_adapter *adapter,
+				       int num_msix)
+{
+	struct pci_dev *pdev = adapter->pdev;
+	u32 value;
+	int err;
 
 	if (adapter->msix_supported) {
-
 		netxen_init_msix_entries(adapter, num_msix);
 		err = pci_enable_msix(pdev, adapter->msix_entries, num_msix);
 		if (err == 0) {
@@ -644,26 +656,59 @@ netxen_setup_intr(struct netxen_adapter *adapter)
 				adapter->max_sds_rings = num_msix;
 
 			dev_info(&pdev->dev, "using msi-x interrupts\n");
-			return;
+			return 0;
 		}
-
-		if (err > 0)
-			pci_disable_msix(pdev);
-
 		/* fall through for msi */
 	}
 
 	if (use_msi && !pci_enable_msi(pdev)) {
+		value = msi_tgt_status[adapter->ahw.pci_func];
 		adapter->flags |= NETXEN_NIC_MSI_ENABLED;
-		adapter->tgt_status_reg = netxen_get_ioaddr(adapter,
-				msi_tgt_status[adapter->ahw.pci_func]);
-		dev_info(&pdev->dev, "using msi interrupts\n");
+		adapter->tgt_status_reg = netxen_get_ioaddr(adapter, value);
 		adapter->msix_entries[0].vector = pdev->irq;
-		return;
+		dev_info(&pdev->dev, "using msi interrupts\n");
+		return 0;
 	}
 
-	dev_info(&pdev->dev, "using legacy interrupts\n");
-	adapter->msix_entries[0].vector = pdev->irq;
+	dev_err(&pdev->dev, "Failed to acquire MSI-X/MSI interrupt vector\n");
+	return -EIO;
+}
+
+static int netxen_setup_intr(struct netxen_adapter *adapter)
+{
+	struct pci_dev *pdev = adapter->pdev;
+	int num_msix;
+
+	if (adapter->rss_supported)
+		num_msix = (num_online_cpus() >= MSIX_ENTRIES_PER_ADAPTER) ?
+			    MSIX_ENTRIES_PER_ADAPTER : 2;
+	else
+		num_msix = 1;
+
+	adapter->max_sds_rings = 1;
+	adapter->flags &= ~(NETXEN_NIC_MSI_ENABLED | NETXEN_NIC_MSIX_ENABLED);
+
+	netxen_initialize_interrupt_registers(adapter);
+	netxen_set_msix_bit(pdev, 0);
+
+	if (netxen_function_zero(pdev)) {
+		if (!netxen_setup_msi_interrupts(adapter, num_msix))
+			netxen_set_interrupt_mode(adapter, NETXEN_MSI_MODE);
+		else
+			netxen_set_interrupt_mode(adapter, NETXEN_INTX_MODE);
+	} else {
+		if (netxen_get_interrupt_mode(adapter) == NETXEN_MSI_MODE &&
+		    netxen_setup_msi_interrupts(adapter, num_msix)) {
+			dev_err(&pdev->dev, "Co-existence of MSI-X/MSI and INTx interrupts is not supported\n");
+			return -EIO;
+		}
+	}
+
+	if (!NETXEN_IS_MSI_FAMILY(adapter)) {
+		adapter->msix_entries[0].vector = pdev->irq;
+		dev_info(&pdev->dev, "using legacy interrupts\n");
+	}
+	return 0;
 }
 
 static void
@@ -841,7 +886,9 @@ netxen_check_options(struct netxen_adapter *adapter)
 	}
 
 	if (adapter->portnum == 0) {
-		get_brd_name_by_type(adapter->ahw.board_type, brd_name);
+		if (netxen_nic_get_brd_name_by_type(adapter->ahw.board_type,
+						    brd_name))
+			strcpy(serial_num, "Unknown");
 
 		pr_info("%s: %s Board S/N %s  Chip rev 0x%x\n",
 				module_name(THIS_MODULE),
@@ -860,9 +907,9 @@ netxen_check_options(struct netxen_adapter *adapter)
 		adapter->ahw.cut_through = (i & 0x8000) ? 1 : 0;
 	}
 
-	dev_info(&pdev->dev, "firmware v%d.%d.%d [%s]\n",
-			fw_major, fw_minor, fw_build,
-			adapter->ahw.cut_through ? "cut-through" : "legacy");
+	dev_info(&pdev->dev, "Driver v%s, firmware v%d.%d.%d [%s]\n",
+		 NETXEN_NIC_LINUX_VERSIONID, fw_major, fw_minor, fw_build,
+		 adapter->ahw.cut_through ? "cut-through" : "legacy");
 
 	if (adapter->fw_version >= NETXEN_VERSION_CODE(4, 0, 222))
 		adapter->capabilities = NXRD32(adapter, CRB_FW_CAPABILITIES_1);
@@ -1508,7 +1555,13 @@ netxen_nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	netxen_nic_clear_stats(adapter);
 
-	netxen_setup_intr(adapter);
+	err = netxen_setup_intr(adapter);
+
+	if (err) {
+		dev_err(&adapter->pdev->dev,
+			"Failed to setup interrupts, error = %d\n", err);
+		goto err_out_disable_msi;
+	}
 
 	err = netxen_setup_netdev(adapter, netdev);
 	if (err)
@@ -1596,7 +1649,7 @@ static void netxen_nic_remove(struct pci_dev *pdev)
 	clear_bit(__NX_RESETTING, &adapter->state);
 
 	netxen_teardown_intr(adapter);
-
+	netxen_set_interrupt_mode(adapter, 0);
 	netxen_remove_diag_entries(adapter);
 
 	netxen_cleanup_pci_map(adapter);
@@ -2721,7 +2774,7 @@ netxen_store_bridged_mode(struct device *dev,
 	if (adapter->is_up != NETXEN_ADAPTER_UP_MAGIC)
 		goto err_out;
 
-	if (strict_strtoul(buf, 2, &new))
+	if (kstrtoul(buf, 2, &new))
 		goto err_out;
 
 	if (!netxen_config_bridged_mode(adapter, !!new))
@@ -2760,7 +2813,7 @@ netxen_store_diag_mode(struct device *dev,
 	struct netxen_adapter *adapter = dev_get_drvdata(dev);
 	unsigned long new;
 
-	if (strict_strtoul(buf, 2, &new))
+	if (kstrtoul(buf, 2, &new))
 		return -EINVAL;
 
 	if (!!new != !!(adapter->flags & NETXEN_NIC_DIAG_ENABLED))
@@ -3311,7 +3364,7 @@ static int netxen_netdev_event(struct notifier_block *this,
 				 unsigned long event, void *ptr)
 {
 	struct netxen_adapter *adapter;
-	struct net_device *dev = (struct net_device *)ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct net_device *orig_dev = dev;
 	struct net_device *slave;
 
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic.h b/drivers/net/ethernet/qlogic/qlcnic/qlcnic.h
index c1b693c..b00cf56 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic.h
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic.h
@@ -38,8 +38,8 @@
 
 #define _QLCNIC_LINUX_MAJOR 5
 #define _QLCNIC_LINUX_MINOR 2
-#define _QLCNIC_LINUX_SUBVERSION 42
-#define QLCNIC_LINUX_VERSIONID  "5.2.42"
+#define _QLCNIC_LINUX_SUBVERSION 44
+#define QLCNIC_LINUX_VERSIONID  "5.2.44"
 #define QLCNIC_DRV_IDC_VER  0x01
 #define QLCNIC_DRIVER_VERSION  ((_QLCNIC_LINUX_MAJOR << 16) |\
 		 (_QLCNIC_LINUX_MINOR << 8) | (_QLCNIC_LINUX_SUBVERSION))
@@ -303,7 +303,6 @@ extern int qlcnic_use_msi;
 extern int qlcnic_use_msi_x;
 extern int qlcnic_auto_fw_reset;
 extern int qlcnic_load_fw_file;
-extern int qlcnic_config_npars;
 
 /* Number of status descriptors to handle per interrupt */
 #define MAX_STATUS_HANDLE	(64)
@@ -394,6 +393,9 @@ struct qlcnic_fw_dump {
 	u32	size;	/* total size of the dump */
 	void	*data;	/* dump data area */
 	struct	qlcnic_dump_template_hdr *tmpl_hdr;
+	dma_addr_t phys_addr;
+	void	*dma_buffer;
+	bool	use_pex_dma;
 };
 
 /*
@@ -427,6 +429,7 @@ struct qlcnic_hardware_context {
 	u8 nic_mode;
 	char diag_cnt;
 
+	u16 max_uc_count;
 	u16 port_type;
 	u16 board_type;
 	u16 supported_type;
@@ -443,9 +446,10 @@ struct qlcnic_hardware_context {
 	u16 max_mtu;
 	u32 msg_enable;
 	u16 act_pci_func;
+	u16 max_pci_func;
 
 	u32 capabilities;
-	u32 capabilities2;
+	u32 extra_capability[3];
 	u32 temp;
 	u32 int_vec_bit;
 	u32 fw_hal_version;
@@ -815,7 +819,8 @@ struct qlcnic_mac_list_s {
 
 #define QLCNIC_FW_CAPABILITY_2_LRO_MAX_TCP_SEG	BIT_2
 #define QLCNIC_FW_CAP2_HW_LRO_IPV6		BIT_3
-#define QLCNIC_FW_CAPABILITY_2_OCBB		BIT_5
+#define QLCNIC_FW_CAPABILITY_SET_DRV_VER	BIT_5
+#define QLCNIC_FW_CAPABILITY_2_BEACON		BIT_7
 
 /* module types */
 #define LINKEVENT_MODULE_NOT_PRESENT			1
@@ -913,6 +918,9 @@ struct qlcnic_ipaddr {
 #define QLCNIC_IS_TSO_CAPABLE(adapter)  \
 	((adapter)->ahw->capabilities & QLCNIC_FW_CAPABILITY_TSO)
 
+#define QLCNIC_BEACON_EANBLE		0xC
+#define QLCNIC_BEACON_DISABLE		0xD
+
 #define QLCNIC_DEF_NUM_STS_DESC_RINGS	4
 #define QLCNIC_MSIX_TBL_SPACE		8192
 #define QLCNIC_PCI_REG_MSIX_TBL 	0x44
@@ -932,6 +940,7 @@ struct qlcnic_ipaddr {
 #define __QLCNIC_SRIOV_ENABLE		10
 #define __QLCNIC_SRIOV_CAPABLE		11
 #define __QLCNIC_MBX_POLL_ENABLE	12
+#define __QLCNIC_DIAG_MODE		13
 
 #define QLCNIC_INTERRUPT_TEST		1
 #define QLCNIC_LOOPBACK_TEST		2
@@ -1467,7 +1476,7 @@ int qlcnic_nic_del_mac(struct qlcnic_adapter *, const u8 *);
 void qlcnic_82xx_free_mac_list(struct qlcnic_adapter *adapter);
 
 int qlcnic_fw_cmd_set_mtu(struct qlcnic_adapter *adapter, int mtu);
-int qlcnic_fw_cmd_set_drv_version(struct qlcnic_adapter *);
+int qlcnic_fw_cmd_set_drv_version(struct qlcnic_adapter *, u32);
 int qlcnic_change_mtu(struct net_device *netdev, int new_mtu);
 netdev_features_t qlcnic_fix_features(struct net_device *netdev,
 	netdev_features_t features);
@@ -1489,7 +1498,9 @@ netdev_tx_t qlcnic_xmit_frame(struct sk_buff *skb, struct net_device *netdev);
 int qlcnic_set_max_rss(struct qlcnic_adapter *, u8, size_t);
 int qlcnic_validate_max_rss(struct qlcnic_adapter *, __u32);
 void qlcnic_alloc_lb_filters_mem(struct qlcnic_adapter *adapter);
+void qlcnic_82xx_set_mac_filter_count(struct qlcnic_adapter *);
 int qlcnic_enable_msix(struct qlcnic_adapter *, u32);
+void qlcnic_set_drv_version(struct qlcnic_adapter *);
 
 /*  eSwitch management functions */
 int qlcnic_config_switch_port(struct qlcnic_adapter *,
@@ -1543,6 +1554,7 @@ int qlcnic_set_default_offload_settings(struct qlcnic_adapter *);
 int qlcnic_reset_npar_config(struct qlcnic_adapter *);
 int qlcnic_set_eswitch_port_config(struct qlcnic_adapter *);
 void qlcnic_add_lb_filter(struct qlcnic_adapter *, struct sk_buff *, int, u16);
+int qlcnic_get_beacon_state(struct qlcnic_adapter *, u8 *);
 int qlcnic_83xx_configure_opmode(struct qlcnic_adapter *adapter);
 int qlcnic_read_mac_addr(struct qlcnic_adapter *);
 int qlcnic_setup_netdev(struct qlcnic_adapter *, struct net_device *, int);
@@ -1584,6 +1596,8 @@ struct qlcnic_nic_template {
 	void (*napi_del)(struct qlcnic_adapter *);
 	void (*config_ipaddr)(struct qlcnic_adapter *, __be32, int);
 	irqreturn_t (*clear_legacy_intr)(struct qlcnic_adapter *);
+	int (*shutdown)(struct pci_dev *);
+	int (*resume)(struct qlcnic_adapter *);
 };
 
 /* Adapter hardware abstraction */
@@ -1625,6 +1639,7 @@ struct qlcnic_hardware_ops {
 	int (*config_promisc_mode) (struct qlcnic_adapter *, u32);
 	void (*change_l2_filter) (struct qlcnic_adapter *, u64 *, u16);
 	int (*get_board_info) (struct qlcnic_adapter *);
+	void (*set_mac_filter_count) (struct qlcnic_adapter *);
 	void (*free_mac_list) (struct qlcnic_adapter *);
 };
 
@@ -1787,6 +1802,18 @@ static inline void qlcnic_napi_enable(struct qlcnic_adapter *adapter)
 	adapter->ahw->hw_ops->napi_enable(adapter);
 }
 
+static inline int __qlcnic_shutdown(struct pci_dev *pdev)
+{
+	struct qlcnic_adapter *adapter = pci_get_drvdata(pdev);
+
+	return adapter->nic_ops->shutdown(pdev);
+}
+
+static inline int __qlcnic_resume(struct qlcnic_adapter *adapter)
+{
+	return adapter->nic_ops->resume(adapter);
+}
+
 static inline void qlcnic_napi_disable(struct qlcnic_adapter *adapter)
 {
 	adapter->ahw->hw_ops->napi_disable(adapter);
@@ -1840,6 +1867,11 @@ static inline void qlcnic_free_mac_list(struct qlcnic_adapter *adapter)
 	return adapter->ahw->hw_ops->free_mac_list(adapter);
 }
 
+static inline void qlcnic_set_mac_filter_count(struct qlcnic_adapter *adapter)
+{
+	adapter->ahw->hw_ops->set_mac_filter_count(adapter);
+}
+
 static inline void qlcnic_dev_request_reset(struct qlcnic_adapter *adapter,
 					    u32 key)
 {
@@ -1886,6 +1918,21 @@ static inline void qlcnic_enable_int(struct qlcnic_host_sds_ring *sds_ring)
 		writel(0xfbff, adapter->tgt_mask_reg);
 }
 
+static inline int qlcnic_get_diag_lock(struct qlcnic_adapter *adapter)
+{
+	return test_and_set_bit(__QLCNIC_DIAG_MODE, &adapter->state);
+}
+
+static inline void qlcnic_release_diag_lock(struct qlcnic_adapter *adapter)
+{
+	clear_bit(__QLCNIC_DIAG_MODE, &adapter->state);
+}
+
+static inline int qlcnic_check_diag_status(struct qlcnic_adapter *adapter)
+{
+	return test_bit(__QLCNIC_DIAG_MODE, &adapter->state);
+}
+
 extern const struct ethtool_ops qlcnic_sriov_vf_ethtool_ops;
 extern const struct ethtool_ops qlcnic_ethtool_ops;
 extern const struct ethtool_ops qlcnic_ethtool_failed_ops;
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c
index b4ff1e3..0913c62 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c
@@ -63,6 +63,7 @@ static const struct qlcnic_mailbox_metadata qlcnic_83xx_mbx_tbl[] = {
 	{QLCNIC_CMD_STOP_NIC_FUNC, 2, 1},
 	{QLCNIC_CMD_SET_LED_CONFIG, 5, 1},
 	{QLCNIC_CMD_GET_LED_CONFIG, 1, 5},
+	{QLCNIC_CMD_83XX_SET_DRV_VER, 4, 1},
 	{QLCNIC_CMD_ADD_RCV_RINGS, 130, 26},
 	{QLCNIC_CMD_CONFIG_VPORT, 4, 4},
 	{QLCNIC_CMD_BC_EVENT_SETUP, 2, 1},
@@ -172,6 +173,7 @@ static struct qlcnic_hardware_ops qlcnic_83xx_hw_ops = {
 	.config_promisc_mode		= qlcnic_83xx_nic_set_promisc,
 	.change_l2_filter		= qlcnic_83xx_change_l2_filter,
 	.get_board_info			= qlcnic_83xx_get_port_info,
+	.set_mac_filter_count		= qlcnic_83xx_set_mac_filter_count,
 	.free_mac_list			= qlcnic_82xx_free_mac_list,
 };
 
@@ -184,6 +186,8 @@ static struct qlcnic_nic_template qlcnic_83xx_ops = {
 	.napi_del		= qlcnic_83xx_napi_del,
 	.config_ipaddr		= qlcnic_83xx_config_ipaddr,
 	.clear_legacy_intr	= qlcnic_83xx_clear_legacy_intr,
+	.shutdown		= qlcnic_83xx_shutdown,
+	.resume			= qlcnic_83xx_resume,
 };
 
 void qlcnic_83xx_register_map(struct qlcnic_hardware_context *ahw)
@@ -312,6 +316,11 @@ inline void qlcnic_83xx_clear_legacy_intr_mask(struct qlcnic_adapter *adapter)
 	writel(0, adapter->tgt_mask_reg);
 }
 
+inline void qlcnic_83xx_set_legacy_intr_mask(struct qlcnic_adapter *adapter)
+{
+	writel(1, adapter->tgt_mask_reg);
+}
+
 /* Enable MSI-x and INT-x interrupts */
 void qlcnic_83xx_enable_intr(struct qlcnic_adapter *adapter,
 			     struct qlcnic_host_sds_ring *sds_ring)
@@ -458,6 +467,9 @@ void qlcnic_83xx_free_mbx_intr(struct qlcnic_adapter *adapter)
 {
 	u32 num_msix;
 
+	if (!(adapter->flags & QLCNIC_MSIX_ENABLED))
+		qlcnic_83xx_set_legacy_intr_mask(adapter);
+
 	qlcnic_83xx_disable_mbx_intr(adapter);
 
 	if (adapter->flags & QLCNIC_MSIX_ENABLED)
@@ -474,7 +486,6 @@ int qlcnic_83xx_setup_mbx_intr(struct qlcnic_adapter *adapter)
 {
 	irq_handler_t handler;
 	u32 val;
-	char name[32];
 	int err = 0;
 	unsigned long flags = 0;
 
@@ -485,9 +496,7 @@ int qlcnic_83xx_setup_mbx_intr(struct qlcnic_adapter *adapter)
 	if (adapter->flags & QLCNIC_MSIX_ENABLED) {
 		handler = qlcnic_83xx_handle_aen;
 		val = adapter->msix_entries[adapter->ahw->num_msix - 1].vector;
-		snprintf(name, (IFNAMSIZ + 4),
-			 "%s[%s]", "qlcnic", "aen");
-		err = request_irq(val, handler, flags, name, adapter);
+		err = request_irq(val, handler, flags, "qlcnic-MB", adapter);
 		if (err) {
 			dev_err(&adapter->pdev->dev,
 				"failed to register MBX interrupt\n");
@@ -604,6 +613,22 @@ int qlcnic_83xx_get_port_info(struct qlcnic_adapter *adapter)
 	return status;
 }
 
+void qlcnic_83xx_set_mac_filter_count(struct qlcnic_adapter *adapter)
+{
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	u16 act_pci_fn = ahw->act_pci_func;
+	u16 count;
+
+	ahw->max_mc_count = QLC_83XX_MAX_MC_COUNT;
+	if (act_pci_fn <= 2)
+		count = (QLC_83XX_MAX_UC_COUNT - QLC_83XX_MAX_MC_COUNT) /
+			 act_pci_fn;
+	else
+		count = (QLC_83XX_LB_MAX_FILTERS - QLC_83XX_MAX_MC_COUNT) /
+			 act_pci_fn;
+	ahw->max_uc_count = count;
+}
+
 void qlcnic_83xx_enable_mbx_intrpt(struct qlcnic_adapter *adapter)
 {
 	u32 val;
@@ -839,7 +864,9 @@ void qlcnic_83xx_idc_aen_work(struct work_struct *work)
 	int i, err = 0;
 
 	adapter = container_of(work, struct qlcnic_adapter, idc_aen_work.work);
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_IDC_ACK);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_IDC_ACK);
+	if (err)
+		return;
 
 	for (i = 1; i < QLC_83XX_MBX_AEN_CNT; i++)
 		cmd.req.arg[i] = adapter->ahw->mbox_aen[i];
@@ -1080,8 +1107,10 @@ int qlcnic_83xx_create_rx_ctx(struct qlcnic_adapter *adapter)
 		cap |= QLC_83XX_FW_CAP_LRO_MSS;
 
 	/* set mailbox hdr and capabilities */
-	qlcnic_alloc_mbx_args(&cmd, adapter,
-			      QLCNIC_CMD_CREATE_RX_CTX);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_CREATE_RX_CTX);
+	if (err)
+		return err;
 
 	if (qlcnic_sriov_pf_check(adapter) || qlcnic_sriov_vf_check(adapter))
 		cmd.req.arg[0] |= (0x3 << 29);
@@ -1239,7 +1268,9 @@ int qlcnic_83xx_create_tx_ctx(struct qlcnic_adapter *adapter,
 		mbx.intr_id = 0xffff;
 	mbx.src = 0;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CREATE_TX_CTX);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CREATE_TX_CTX);
+	if (err)
+		return err;
 
 	if (qlcnic_sriov_pf_check(adapter) || qlcnic_sriov_vf_check(adapter))
 		cmd.req.arg[0] |= (0x3 << 29);
@@ -1385,8 +1416,11 @@ int qlcnic_83xx_config_led(struct qlcnic_adapter *adapter, u32 state,
 
 	if (state) {
 		/* Get LED configuration */
-		qlcnic_alloc_mbx_args(&cmd, adapter,
-				      QLCNIC_CMD_GET_LED_CONFIG);
+		status = qlcnic_alloc_mbx_args(&cmd, adapter,
+					       QLCNIC_CMD_GET_LED_CONFIG);
+		if (status)
+			return status;
+
 		status = qlcnic_issue_cmd(adapter, &cmd);
 		if (status) {
 			dev_err(&adapter->pdev->dev,
@@ -1400,8 +1434,11 @@ int qlcnic_83xx_config_led(struct qlcnic_adapter *adapter, u32 state,
 		/* Set LED Configuration */
 		mbx_in = (LSW(QLC_83XX_LED_CONFIG) << 16) |
 			  LSW(QLC_83XX_LED_CONFIG);
-		qlcnic_alloc_mbx_args(&cmd, adapter,
-				      QLCNIC_CMD_SET_LED_CONFIG);
+		status = qlcnic_alloc_mbx_args(&cmd, adapter,
+					       QLCNIC_CMD_SET_LED_CONFIG);
+		if (status)
+			return status;
+
 		cmd.req.arg[1] = mbx_in;
 		cmd.req.arg[2] = mbx_in;
 		cmd.req.arg[3] = mbx_in;
@@ -1418,8 +1455,11 @@ mbx_err:
 
 	} else {
 		/* Restoring default LED configuration */
-		qlcnic_alloc_mbx_args(&cmd, adapter,
-				      QLCNIC_CMD_SET_LED_CONFIG);
+		status = qlcnic_alloc_mbx_args(&cmd, adapter,
+					       QLCNIC_CMD_SET_LED_CONFIG);
+		if (status)
+			return status;
+
 		cmd.req.arg[1] = adapter->ahw->mbox_reg[0];
 		cmd.req.arg[2] = adapter->ahw->mbox_reg[1];
 		cmd.req.arg[3] = adapter->ahw->mbox_reg[2];
@@ -1489,10 +1529,18 @@ void qlcnic_83xx_register_nic_idc_func(struct qlcnic_adapter *adapter,
 		return;
 
 	if (enable) {
-		qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_INIT_NIC_FUNC);
+		status = qlcnic_alloc_mbx_args(&cmd, adapter,
+					       QLCNIC_CMD_INIT_NIC_FUNC);
+		if (status)
+			return;
+
 		cmd.req.arg[1] = BIT_0 | BIT_31;
 	} else {
-		qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_STOP_NIC_FUNC);
+		status = qlcnic_alloc_mbx_args(&cmd, adapter,
+					       QLCNIC_CMD_STOP_NIC_FUNC);
+		if (status)
+			return;
+
 		cmd.req.arg[1] = BIT_0 | BIT_31;
 	}
 	status = qlcnic_issue_cmd(adapter, &cmd);
@@ -1509,7 +1557,10 @@ int qlcnic_83xx_set_port_config(struct qlcnic_adapter *adapter)
 	struct qlcnic_cmd_args cmd;
 	int err;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_PORT_CONFIG);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_PORT_CONFIG);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = adapter->ahw->port_config;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	if (err)
@@ -1523,7 +1574,10 @@ int qlcnic_83xx_get_port_config(struct qlcnic_adapter *adapter)
 	struct qlcnic_cmd_args cmd;
 	int err;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_PORT_CONFIG);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_PORT_CONFIG);
+	if (err)
+		return err;
+
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	if (err)
 		dev_info(&adapter->pdev->dev, "Get Port config failed\n");
@@ -1539,7 +1593,10 @@ int qlcnic_83xx_setup_link_event(struct qlcnic_adapter *adapter, int enable)
 	u32 temp;
 	struct qlcnic_cmd_args cmd;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_LINK_EVENT);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_LINK_EVENT);
+	if (err)
+		return err;
+
 	temp = adapter->recv_ctx->context_id << 16;
 	cmd.req.arg[1] = (enable ? 1 : 0) | BIT_8 | temp;
 	err = qlcnic_issue_cmd(adapter, &cmd);
@@ -1570,7 +1627,11 @@ int qlcnic_83xx_nic_set_promisc(struct qlcnic_adapter *adapter, u32 mode)
 	if (adapter->recv_ctx->state == QLCNIC_HOST_CTX_STATE_FREED)
 		return -EIO;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_MAC_RX_MODE);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_CONFIGURE_MAC_RX_MODE);
+	if (err)
+		return err;
+
 	qlcnic_83xx_set_interface_id_promisc(adapter, &temp);
 	cmd.req.arg[1] = (mode ? 1 : 0) | temp;
 	err = qlcnic_issue_cmd(adapter, &cmd);
@@ -1588,16 +1649,24 @@ int qlcnic_83xx_loopback_test(struct net_device *netdev, u8 mode)
 	struct qlcnic_hardware_context *ahw = adapter->ahw;
 	int ret = 0, loop = 0, max_sds_rings = adapter->max_sds_rings;
 
-	QLCDB(adapter, DRV, "%s loopback test in progress\n",
-	      mode == QLCNIC_ILB_MODE ? "internal" : "external");
 	if (ahw->op_mode == QLCNIC_NON_PRIV_FUNC) {
-		dev_warn(&adapter->pdev->dev,
-			 "Loopback test not supported for non privilege function\n");
+		netdev_warn(netdev,
+			    "Loopback test not supported in non privileged mode\n");
 		return ret;
 	}
 
-	if (test_and_set_bit(__QLCNIC_RESETTING, &adapter->state))
+	if (test_bit(__QLCNIC_RESETTING, &adapter->state)) {
+		netdev_info(netdev, "Device is resetting\n");
 		return -EBUSY;
+	}
+
+	if (qlcnic_get_diag_lock(adapter)) {
+		netdev_info(netdev, "Device is in diagnostics mode\n");
+		return -EBUSY;
+	}
+
+	netdev_info(netdev, "%s loopback test in progress\n",
+		    mode == QLCNIC_ILB_MODE ? "internal" : "external");
 
 	ret = qlcnic_83xx_diag_alloc_res(netdev, QLCNIC_LOOPBACK_TEST,
 					 max_sds_rings);
@@ -1610,13 +1679,19 @@ int qlcnic_83xx_loopback_test(struct net_device *netdev, u8 mode)
 
 	/* Poll for link up event before running traffic */
 	do {
-		msleep(500);
+		msleep(QLC_83XX_LB_MSLEEP_COUNT);
 		if (!(adapter->flags & QLCNIC_MSIX_ENABLED))
 			qlcnic_83xx_process_aen(adapter);
 
-		if (loop++ > QLCNIC_ILB_MAX_RCV_LOOP) {
-			dev_info(&adapter->pdev->dev,
-				 "Firmware didn't sent link up event to loopback request\n");
+		if (test_bit(__QLCNIC_RESETTING, &adapter->state)) {
+			netdev_info(netdev,
+				    "Device is resetting, free LB test resources\n");
+			ret = -EIO;
+			goto free_diag_res;
+		}
+		if (loop++ > QLC_83XX_LB_WAIT_COUNT) {
+			netdev_info(netdev,
+				    "Firmware didn't sent link up event to loopback request\n");
 			ret = -QLCNIC_FW_NOT_RESPOND;
 			qlcnic_83xx_clear_lb_mode(adapter, mode);
 			goto free_diag_res;
@@ -1638,13 +1713,14 @@ free_diag_res:
 
 fail_diag_alloc:
 	adapter->max_sds_rings = max_sds_rings;
-	clear_bit(__QLCNIC_RESETTING, &adapter->state);
+	qlcnic_release_diag_lock(adapter);
 	return ret;
 }
 
 int qlcnic_83xx_set_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 {
 	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	struct net_device *netdev = adapter->netdev;
 	int status = 0, loop = 0;
 	u32 config;
 
@@ -1662,9 +1738,9 @@ int qlcnic_83xx_set_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 
 	status = qlcnic_83xx_set_port_config(adapter);
 	if (status) {
-		dev_err(&adapter->pdev->dev,
-			"Failed to Set Loopback Mode = 0x%x.\n",
-			ahw->port_config);
+		netdev_err(netdev,
+			   "Failed to Set Loopback Mode = 0x%x.\n",
+			   ahw->port_config);
 		ahw->port_config = config;
 		clear_bit(QLC_83XX_IDC_COMP_AEN, &ahw->idc.status);
 		return status;
@@ -1672,13 +1748,19 @@ int qlcnic_83xx_set_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 
 	/* Wait for Link and IDC Completion AEN */
 	do {
-		msleep(300);
+		msleep(QLC_83XX_LB_MSLEEP_COUNT);
 		if (!(adapter->flags & QLCNIC_MSIX_ENABLED))
 			qlcnic_83xx_process_aen(adapter);
 
-		if (loop++ > QLCNIC_ILB_MAX_RCV_LOOP) {
-			dev_err(&adapter->pdev->dev,
-				"FW did not generate IDC completion AEN\n");
+		if (test_bit(__QLCNIC_RESETTING, &adapter->state)) {
+			netdev_info(netdev,
+				    "Device is resetting, free LB test resources\n");
+			clear_bit(QLC_83XX_IDC_COMP_AEN, &ahw->idc.status);
+			return -EIO;
+		}
+		if (loop++ > QLC_83XX_LB_WAIT_COUNT) {
+			netdev_err(netdev,
+				   "Did not receive IDC completion AEN\n");
 			clear_bit(QLC_83XX_IDC_COMP_AEN, &ahw->idc.status);
 			qlcnic_83xx_clear_lb_mode(adapter, mode);
 			return -EIO;
@@ -1693,6 +1775,7 @@ int qlcnic_83xx_set_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 int qlcnic_83xx_clear_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 {
 	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	struct net_device *netdev = adapter->netdev;
 	int status = 0, loop = 0;
 	u32 config = ahw->port_config;
 
@@ -1704,9 +1787,9 @@ int qlcnic_83xx_clear_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 
 	status = qlcnic_83xx_set_port_config(adapter);
 	if (status) {
-		dev_err(&adapter->pdev->dev,
-			"Failed to Clear Loopback Mode = 0x%x.\n",
-			ahw->port_config);
+		netdev_err(netdev,
+			   "Failed to Clear Loopback Mode = 0x%x.\n",
+			   ahw->port_config);
 		ahw->port_config = config;
 		clear_bit(QLC_83XX_IDC_COMP_AEN, &ahw->idc.status);
 		return status;
@@ -1714,13 +1797,20 @@ int qlcnic_83xx_clear_lb_mode(struct qlcnic_adapter *adapter, u8 mode)
 
 	/* Wait for Link and IDC Completion AEN */
 	do {
-		msleep(300);
+		msleep(QLC_83XX_LB_MSLEEP_COUNT);
 		if (!(adapter->flags & QLCNIC_MSIX_ENABLED))
 			qlcnic_83xx_process_aen(adapter);
 
-		if (loop++ > QLCNIC_ILB_MAX_RCV_LOOP) {
-			dev_err(&adapter->pdev->dev,
-				"Firmware didn't sent IDC completion AEN\n");
+		if (test_bit(__QLCNIC_RESETTING, &adapter->state)) {
+			netdev_info(netdev,
+				    "Device is resetting, free LB test resources\n");
+			clear_bit(QLC_83XX_IDC_COMP_AEN, &ahw->idc.status);
+			return -EIO;
+		}
+
+		if (loop++ > QLC_83XX_LB_WAIT_COUNT) {
+			netdev_err(netdev,
+				   "Did not receive IDC completion AEN\n");
 			clear_bit(QLC_83XX_IDC_COMP_AEN, &ahw->idc.status);
 			return -EIO;
 		}
@@ -1749,7 +1839,11 @@ void qlcnic_83xx_config_ipaddr(struct qlcnic_adapter *adapter, __be32 ip,
 	u32 temp = 0, temp_ip;
 	struct qlcnic_cmd_args cmd;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_IP_ADDR);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_CONFIGURE_IP_ADDR);
+	if (err)
+		return;
+
 	qlcnic_83xx_set_interface_id_ipaddr(adapter, &temp);
 
 	if (mode == QLCNIC_IP_UP)
@@ -1788,7 +1882,10 @@ int qlcnic_83xx_config_hw_lro(struct qlcnic_adapter *adapter, int mode)
 	if (adapter->recv_ctx->state == QLCNIC_HOST_CTX_STATE_FREED)
 		return 0;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_HW_LRO);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_HW_LRO);
+	if (err)
+		return err;
+
 	temp = adapter->recv_ctx->context_id << 16;
 	arg1 = lro_bit_mask | temp;
 	cmd.req.arg[1] = arg1;
@@ -1810,8 +1907,9 @@ int qlcnic_83xx_config_rss(struct qlcnic_adapter *adapter, int enable)
 			    0xae7b30b4d0ca2bcbULL, 0x43a38fb04167253dULL,
 			    0x255b0ec26d5a56daULL };
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_RSS);
-
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_RSS);
+	if (err)
+		return err;
 	/*
 	 * RSS request:
 	 * bits 3-0: Rsvd
@@ -1917,7 +2015,10 @@ int qlcnic_83xx_get_mac_address(struct qlcnic_adapter *adapter, u8 *mac)
 	struct qlcnic_cmd_args cmd;
 	u32 mac_low, mac_high;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_MAC_ADDRESS);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_MAC_ADDRESS);
+	if (err)
+		return err;
+
 	qlcnic_83xx_configure_mac(adapter, mac, QLCNIC_GET_CURRENT_MAC, &cmd);
 	err = qlcnic_issue_cmd(adapter, &cmd);
 
@@ -1948,7 +2049,10 @@ void qlcnic_83xx_config_intr_coal(struct qlcnic_adapter *adapter)
 	if (adapter->recv_ctx->state == QLCNIC_HOST_CTX_STATE_FREED)
 		return;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIG_INTR_COAL);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIG_INTR_COAL);
+	if (err)
+		return;
+
 	if (coal->type == QLCNIC_INTR_COAL_TYPE_RX) {
 		temp = adapter->recv_ctx->context_id;
 		cmd.req.arg[1] = QLCNIC_INTR_COAL_TYPE_RX | temp << 16;
@@ -2020,7 +2124,10 @@ int qlcnic_enable_eswitch(struct qlcnic_adapter *adapter, u8 port, u8 enable)
 		return err;
 	}
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_TOGGLE_ESWITCH);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_TOGGLE_ESWITCH);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = (port & 0xf) | BIT_4;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 
@@ -2048,7 +2155,10 @@ int qlcnic_83xx_set_nic_info(struct qlcnic_adapter *adapter,
 		return err;
 	}
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_NIC_INFO);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_NIC_INFO);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = (nic->pci_func << 16);
 	cmd.req.arg[2] = 0x1 << 16;
 	cmd.req.arg[3] = nic->phys_port | (nic->switch_mode << 16);
@@ -2079,13 +2189,17 @@ int qlcnic_83xx_get_nic_info(struct qlcnic_adapter *adapter,
 	u32 temp;
 	u8 op = 0;
 	struct qlcnic_cmd_args cmd;
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_NIC_INFO);
-	if (func_id != adapter->ahw->pci_func) {
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_NIC_INFO);
+	if (err)
+		return err;
+
+	if (func_id != ahw->pci_func) {
 		temp = func_id << 16;
 		cmd.req.arg[1] = op | BIT_31 | temp;
 	} else {
-		cmd.req.arg[1] = adapter->ahw->pci_func << 16;
+		cmd.req.arg[1] = ahw->pci_func << 16;
 	}
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	if (err) {
@@ -2112,6 +2226,9 @@ int qlcnic_83xx_get_nic_info(struct qlcnic_adapter *adapter,
 		temp = (cmd.rsp.arg[8] & 0x7FFE0000) >> 17;
 		npar_info->max_linkspeed_reg_offset = temp;
 	}
+	if (npar_info->capabilities & QLCNIC_FW_CAPABILITY_MORE_CAPS)
+		memcpy(ahw->extra_capability, &cmd.rsp.arg[16],
+		       sizeof(ahw->extra_capability));
 
 out:
 	qlcnic_free_mbx_args(&cmd);
@@ -2121,26 +2238,28 @@ out:
 int qlcnic_83xx_get_pci_info(struct qlcnic_adapter *adapter,
 			     struct qlcnic_pci_info *pci_info)
 {
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	struct device *dev = &adapter->pdev->dev;
+	struct qlcnic_cmd_args cmd;
 	int i, err = 0, j = 0;
 	u32 temp;
-	struct qlcnic_cmd_args cmd;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_PCI_INFO);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_PCI_INFO);
+	if (err)
+		return err;
+
 	err = qlcnic_issue_cmd(adapter, &cmd);
 
-	adapter->ahw->act_pci_func = 0;
+	ahw->act_pci_func = 0;
 	if (err == QLCNIC_RCODE_SUCCESS) {
-		pci_info->func_count = cmd.rsp.arg[1] & 0xFF;
-		dev_info(&adapter->pdev->dev,
-			 "%s: total functions = %d\n",
-			 __func__, pci_info->func_count);
+		ahw->max_pci_func = cmd.rsp.arg[1] & 0xFF;
 		for (i = 2, j = 0; j < QLCNIC_MAX_PCI_FUNC; j++, pci_info++) {
 			pci_info->id = cmd.rsp.arg[i] & 0xFFFF;
 			pci_info->active = (cmd.rsp.arg[i] & 0xFFFF0000) >> 16;
 			i++;
 			pci_info->type = cmd.rsp.arg[i] & 0xFFFF;
 			if (pci_info->type == QLCNIC_TYPE_NIC)
-				adapter->ahw->act_pci_func++;
+				ahw->act_pci_func++;
 			temp = (cmd.rsp.arg[i] & 0xFFFF0000) >> 16;
 			pci_info->default_port = temp;
 			i++;
@@ -2152,18 +2271,21 @@ int qlcnic_83xx_get_pci_info(struct qlcnic_adapter *adapter,
 			i++;
 			memcpy(pci_info->mac + sizeof(u32), &cmd.rsp.arg[i], 2);
 			i = i + 3;
-
-			dev_info(&adapter->pdev->dev, "%s:\n"
-				 "\tid = %d active = %d type = %d\n"
-				 "\tport = %d min bw = %d max bw = %d\n"
-				 "\tmac_addr =  %pM\n", __func__,
-				 pci_info->id, pci_info->active, pci_info->type,
-				 pci_info->default_port, pci_info->tx_min_bw,
-				 pci_info->tx_max_bw, pci_info->mac);
+			if (ahw->op_mode == QLCNIC_MGMT_FUNC)
+				dev_info(dev, "id = %d active = %d type = %d\n"
+					 "\tport = %d min bw = %d max bw = %d\n"
+					 "\tmac_addr =  %pM\n", pci_info->id,
+					 pci_info->active, pci_info->type,
+					 pci_info->default_port,
+					 pci_info->tx_min_bw,
+					 pci_info->tx_max_bw, pci_info->mac);
 		}
+		if (ahw->op_mode == QLCNIC_MGMT_FUNC)
+			dev_info(dev, "Max vNIC functions = %d, active vNIC functions = %d\n",
+				 ahw->max_pci_func, ahw->act_pci_func);
+
 	} else {
-		dev_err(&adapter->pdev->dev, "Failed to get PCI Info%d\n",
-			err);
+		dev_err(dev, "Failed to get PCI Info, error = %d\n", err);
 		err = -EIO;
 	}
 
@@ -2180,7 +2302,10 @@ int qlcnic_83xx_config_intrpt(struct qlcnic_adapter *adapter, bool op_type)
 	struct qlcnic_cmd_args cmd;
 
 	max_ints = adapter->ahw->num_msix - 1;
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIG_INTRPT);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIG_INTRPT);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = max_ints;
 
 	if (qlcnic_sriov_vf_check(adapter))
@@ -2808,7 +2933,11 @@ int qlcnic_83xx_test_link(struct qlcnic_adapter *adapter)
 		dev_info(&adapter->pdev->dev, "link state down\n");
 		return config;
 	}
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_LINK_STATUS);
+
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_LINK_STATUS);
+	if (err)
+		return err;
+
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	if (err) {
 		dev_info(&adapter->pdev->dev,
@@ -3034,7 +3163,9 @@ void qlcnic_83xx_get_stats(struct qlcnic_adapter *adapter, u64 *data)
 	struct net_device *netdev = adapter->netdev;
 	int ret = 0;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_STATISTICS);
+	ret = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_STATISTICS);
+	if (ret)
+		return;
 	/* Get Tx stats */
 	cmd.req.arg[1] = BIT_1 | (adapter->tx_ring->ctx_id << 16);
 	cmd.rsp.num = QLC_83XX_TX_STAT_REGS;
@@ -3113,8 +3244,10 @@ int qlcnic_83xx_interrupt_test(struct net_device *netdev)
 	u8 val;
 	int ret, max_sds_rings = adapter->max_sds_rings;
 
-	if (test_and_set_bit(__QLCNIC_RESETTING, &adapter->state))
-		return -EIO;
+	if (qlcnic_get_diag_lock(adapter)) {
+		netdev_info(netdev, "Device in diagnostics mode\n");
+		return -EBUSY;
+	}
 
 	ret = qlcnic_83xx_diag_alloc_res(netdev, QLCNIC_INTERRUPT_TEST,
 					 max_sds_rings);
@@ -3122,7 +3255,9 @@ int qlcnic_83xx_interrupt_test(struct net_device *netdev)
 		goto fail_diag_irq;
 
 	ahw->diag_cnt = 0;
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_INTRPT_TEST);
+	ret = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_INTRPT_TEST);
+	if (ret)
+		goto fail_diag_irq;
 
 	if (adapter->flags & QLCNIC_MSIX_ENABLED)
 		intrpt_id = ahw->intr_tbl[0].id;
@@ -3156,7 +3291,7 @@ done:
 
 fail_diag_irq:
 	adapter->max_sds_rings = max_sds_rings;
-	clear_bit(__QLCNIC_RESETTING, &adapter->state);
+	qlcnic_release_diag_lock(adapter);
 	return ret;
 }
 
@@ -3260,3 +3395,54 @@ int qlcnic_83xx_flash_test(struct qlcnic_adapter *adapter)
 	}
 	return 0;
 }
+
+int qlcnic_83xx_shutdown(struct pci_dev *pdev)
+{
+	struct qlcnic_adapter *adapter = pci_get_drvdata(pdev);
+	struct net_device *netdev = adapter->netdev;
+	int retval;
+
+	netif_device_detach(netdev);
+	qlcnic_cancel_idc_work(adapter);
+
+	if (netif_running(netdev))
+		qlcnic_down(adapter, netdev);
+
+	qlcnic_83xx_disable_mbx_intr(adapter);
+	cancel_delayed_work_sync(&adapter->idc_aen_work);
+
+	retval = pci_save_state(pdev);
+	if (retval)
+		return retval;
+
+	return 0;
+}
+
+int qlcnic_83xx_resume(struct qlcnic_adapter *adapter)
+{
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	struct qlc_83xx_idc *idc = &ahw->idc;
+	int err = 0;
+
+	err = qlcnic_83xx_idc_init(adapter);
+	if (err)
+		return err;
+
+	if (ahw->nic_mode == QLC_83XX_VIRTUAL_NIC_MODE) {
+		if (ahw->op_mode == QLCNIC_MGMT_FUNC) {
+			qlcnic_83xx_set_vnic_opmode(adapter);
+		} else {
+			err = qlcnic_83xx_check_vnic_state(adapter);
+			if (err)
+				return err;
+		}
+	}
+
+	err = qlcnic_83xx_idc_reattach_driver(adapter);
+	if (err)
+		return err;
+
+	qlcnic_schedule_work(adapter, qlcnic_83xx_idc_poll_dev_state,
+			     idc->delay);
+	return err;
+}
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h
index f5db67f..2548d14 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.h
@@ -36,7 +36,8 @@
 #define QLC_83XX_MAX_DRV_LOCK_RECOVERY_ATTEMPT		3
 #define QLC_83XX_DRV_LOCK_RECOVERY_DELAY		200
 #define QLC_83XX_DRV_LOCK_RECOVERY_STATUS_MASK		0x3
-
+#define QLC_83XX_LB_WAIT_COUNT				250
+#define QLC_83XX_LB_MSLEEP_COUNT			20
 #define QLC_83XX_NO_NIC_RESOURCE	0x5
 #define QLC_83XX_MAC_PRESENT		0xC
 #define QLC_83XX_MAC_ABSENT		0xD
@@ -314,6 +315,7 @@ struct qlc_83xx_idc {
 	u8		vnic_state;
 	u8		vnic_wait_limit;
 	u8		quiesce_req;
+	u8		delay_reset;
 	char		**name;
 };
 
@@ -392,6 +394,8 @@ enum qlcnic_83xx_states {
 #define QLC_83XX_LB_MAX_FILTERS			2048
 #define QLC_83XX_LB_BUCKET_SIZE			256
 #define QLC_83XX_MINIMUM_VECTOR			3
+#define QLC_83XX_MAX_MC_COUNT			38
+#define QLC_83XX_MAX_UC_COUNT			4096
 
 #define QLC_83XX_GET_FUNC_MODE_FROM_NPAR_INFO(val)	(val & 0x80000000)
 #define QLC_83XX_GET_LRO_CAPABILITY(val)		(val & 0x20)
@@ -623,4 +627,11 @@ u32 qlcnic_83xx_mac_rcode(struct qlcnic_adapter *);
 u32 qlcnic_83xx_mbx_poll(struct qlcnic_adapter *, u32 *);
 void qlcnic_83xx_enable_mbx_poll(struct qlcnic_adapter *);
 void qlcnic_83xx_disable_mbx_poll(struct qlcnic_adapter *);
+void qlcnic_83xx_set_mac_filter_count(struct qlcnic_adapter *);
+int qlcnic_83xx_shutdown(struct pci_dev *);
+int qlcnic_83xx_resume(struct qlcnic_adapter *);
+int qlcnic_83xx_idc_init(struct qlcnic_adapter *);
+int qlcnic_83xx_idc_reattach_driver(struct qlcnic_adapter *);
+int qlcnic_83xx_set_vnic_opmode(struct qlcnic_adapter *);
+int qlcnic_83xx_check_vnic_state(struct qlcnic_adapter *);
 #endif
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c
index 5e7fb1d..f41dfab 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_init.c
@@ -606,7 +606,7 @@ static int qlcnic_83xx_idc_check_fan_failure(struct qlcnic_adapter *adapter)
 	return 0;
 }
 
-static int qlcnic_83xx_idc_reattach_driver(struct qlcnic_adapter *adapter)
+int qlcnic_83xx_idc_reattach_driver(struct qlcnic_adapter *adapter)
 {
 	int err;
 
@@ -629,6 +629,7 @@ static int qlcnic_83xx_idc_reattach_driver(struct qlcnic_adapter *adapter)
 		return -EIO;
 	}
 
+	qlcnic_set_drv_version(adapter);
 	qlcnic_83xx_idc_attach_driver(adapter);
 
 	return 0;
@@ -649,6 +650,7 @@ static void qlcnic_83xx_idc_update_idc_params(struct qlcnic_adapter *adapter)
 	ahw->idc.collect_dump = 0;
 	ahw->reset_context = 0;
 	adapter->tx_timeo_cnt = 0;
+	ahw->idc.delay_reset = 0;
 
 	clear_bit(__QLCNIC_RESETTING, &adapter->state);
 }
@@ -883,21 +885,41 @@ static int qlcnic_83xx_idc_need_reset_state(struct qlcnic_adapter *adapter)
 	int ret = 0;
 
 	if (adapter->ahw->idc.prev_state != QLC_83XX_IDC_DEV_NEED_RESET) {
-		qlcnic_83xx_idc_update_drv_ack_reg(adapter, 1, 1);
 		qlcnic_83xx_idc_update_audit_reg(adapter, 0, 1);
 		set_bit(__QLCNIC_RESETTING, &adapter->state);
 		clear_bit(QLC_83XX_MBX_READY, &adapter->ahw->idc.status);
 		if (adapter->ahw->nic_mode == QLC_83XX_VIRTUAL_NIC_MODE)
 			qlcnic_83xx_disable_vnic_mode(adapter, 1);
-		qlcnic_83xx_idc_detach_driver(adapter);
+
+		if (qlcnic_check_diag_status(adapter)) {
+			dev_info(&adapter->pdev->dev,
+				 "%s: Wait for diag completion\n", __func__);
+			adapter->ahw->idc.delay_reset = 1;
+			return 0;
+		} else {
+			qlcnic_83xx_idc_update_drv_ack_reg(adapter, 1, 1);
+			qlcnic_83xx_idc_detach_driver(adapter);
+		}
 	}
 
-	/* Check ACK from other functions */
-	ret = qlcnic_83xx_idc_check_reset_ack_reg(adapter);
-	if (ret) {
+	if (qlcnic_check_diag_status(adapter)) {
 		dev_info(&adapter->pdev->dev,
-			 "%s: Waiting for reset ACK\n", __func__);
-		return 0;
+			 "%s: Wait for diag completion\n", __func__);
+		return  -1;
+	} else {
+		if (adapter->ahw->idc.delay_reset) {
+			qlcnic_83xx_idc_update_drv_ack_reg(adapter, 1, 1);
+			qlcnic_83xx_idc_detach_driver(adapter);
+			adapter->ahw->idc.delay_reset = 0;
+		}
+
+		/* Check for ACK from other functions */
+		ret = qlcnic_83xx_idc_check_reset_ack_reg(adapter);
+		if (ret) {
+			dev_info(&adapter->pdev->dev,
+				 "%s: Waiting for reset ACK\n", __func__);
+			return -1;
+		}
 	}
 
 	/* Transit to INIT state and restart the HW */
@@ -1113,7 +1135,7 @@ qlcnic_83xx_idc_first_to_load_function_handler(struct qlcnic_adapter *adapter)
 	return 0;
 }
 
-static int qlcnic_83xx_idc_init(struct qlcnic_adapter *adapter)
+int qlcnic_83xx_idc_init(struct qlcnic_adapter *adapter)
 {
 	int ret = -EIO;
 
@@ -1532,9 +1554,18 @@ static int qlcnic_83xx_reset_template_checksum(struct qlcnic_adapter *p_dev)
 
 int qlcnic_83xx_get_reset_instruction_template(struct qlcnic_adapter *p_dev)
 {
-	u8 *p_buff;
-	u32 addr, count;
 	struct qlcnic_hardware_context *ahw = p_dev->ahw;
+	u32 addr, count, prev_ver, curr_ver;
+	u8 *p_buff;
+
+	if (ahw->reset.buff != NULL) {
+		prev_ver = p_dev->fw_version;
+		curr_ver = qlcnic_83xx_get_fw_version(p_dev);
+		if (curr_ver > prev_ver)
+			kfree(ahw->reset.buff);
+		else
+			return 0;
+	}
 
 	ahw->reset.seq_error = 0;
 	ahw->reset.buff = kzalloc(QLC_83XX_RESTART_TEMPLATE_SIZE, GFP_KERNEL);
@@ -2062,7 +2093,11 @@ static void qlcnic_83xx_clear_function_resources(struct qlcnic_adapter *adapter)
 	audit_mask = QLCRDX(adapter->ahw, QLC_83XX_IDC_DRV_AUDIT);
 
 	if (IS_QLC_83XX_USED(adapter, presence_mask, audit_mask)) {
-		qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_STOP_NIC_FUNC);
+		status = qlcnic_alloc_mbx_args(&cmd, adapter,
+					       QLCNIC_CMD_STOP_NIC_FUNC);
+		if (status)
+			return;
+
 		cmd.req.arg[1] = BIT_31;
 		status = qlcnic_issue_cmd(adapter, &cmd);
 		if (status)
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_vnic.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_vnic.c
index b0c3de9..599d1fd 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_vnic.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_vnic.c
@@ -39,30 +39,21 @@ int qlcnic_83xx_disable_vnic_mode(struct qlcnic_adapter *adapter, int lock)
 	return 0;
 }
 
-static int qlcnic_83xx_set_vnic_opmode(struct qlcnic_adapter *adapter)
+int qlcnic_83xx_set_vnic_opmode(struct qlcnic_adapter *adapter)
 {
 	u8 id;
-	int i, ret = -EBUSY;
+	int ret = -EBUSY;
 	u32 data = QLCNIC_MGMT_FUNC;
 	struct qlcnic_hardware_context *ahw = adapter->ahw;
 
 	if (qlcnic_83xx_lock_driver(adapter))
 		return ret;
 
-	if (qlcnic_config_npars) {
-		for (i = 0; i < ahw->act_pci_func; i++) {
-			id = adapter->npars[i].pci_func;
-			if (id == ahw->pci_func)
-				continue;
-			data |= qlcnic_config_npars &
-				QLC_83XX_SET_FUNC_OPMODE(0x3, id);
-		}
-	} else {
-		data = QLCRDX(adapter->ahw, QLC_83XX_DRV_OP_MODE);
-		data = (data & ~QLC_83XX_SET_FUNC_OPMODE(0x3, ahw->pci_func)) |
-		       QLC_83XX_SET_FUNC_OPMODE(QLCNIC_MGMT_FUNC,
-						ahw->pci_func);
-	}
+	id = ahw->pci_func;
+	data = QLCRDX(adapter->ahw, QLC_83XX_DRV_OP_MODE);
+	data = (data & ~QLC_83XX_SET_FUNC_OPMODE(0x3, id)) |
+	       QLC_83XX_SET_FUNC_OPMODE(QLCNIC_MGMT_FUNC, id);
+
 	QLCWRX(adapter->ahw, QLC_83XX_DRV_OP_MODE, data);
 
 	qlcnic_83xx_unlock_driver(adapter);
@@ -196,20 +187,24 @@ int qlcnic_83xx_config_vnic_opmode(struct qlcnic_adapter *adapter)
 	else
 		priv_level = QLC_83XX_GET_FUNC_PRIVILEGE(op_mode,
 							 ahw->pci_func);
-
-	if (priv_level == QLCNIC_NON_PRIV_FUNC) {
+	switch (priv_level) {
+	case QLCNIC_NON_PRIV_FUNC:
 		ahw->op_mode = QLCNIC_NON_PRIV_FUNC;
 		ahw->idc.state_entry = qlcnic_83xx_idc_ready_state_entry;
 		nic_ops->init_driver = qlcnic_83xx_init_non_privileged_vnic;
-	} else if (priv_level == QLCNIC_PRIV_FUNC) {
+		break;
+	case QLCNIC_PRIV_FUNC:
 		ahw->op_mode = QLCNIC_PRIV_FUNC;
 		ahw->idc.state_entry = qlcnic_83xx_idc_vnic_pf_entry;
 		nic_ops->init_driver = qlcnic_83xx_init_privileged_vnic;
-	} else if (priv_level == QLCNIC_MGMT_FUNC) {
+		break;
+	case QLCNIC_MGMT_FUNC:
 		ahw->op_mode = QLCNIC_MGMT_FUNC;
 		ahw->idc.state_entry = qlcnic_83xx_idc_ready_state_entry;
 		nic_ops->init_driver = qlcnic_83xx_init_mgmt_vnic;
-	} else {
+		break;
+	default:
+		dev_err(&adapter->pdev->dev, "Invalid Virtual NIC opmode\n");
 		return -EIO;
 	}
 
@@ -218,8 +213,29 @@ int qlcnic_83xx_config_vnic_opmode(struct qlcnic_adapter *adapter)
 	else
 		adapter->flags &= ~QLCNIC_ESWITCH_ENABLED;
 
-	adapter->ahw->idc.vnic_state = QLCNIC_DEV_NPAR_NON_OPER;
-	adapter->ahw->idc.vnic_wait_limit = QLCNIC_DEV_NPAR_OPER_TIMEO;
+	ahw->idc.vnic_state = QLCNIC_DEV_NPAR_NON_OPER;
+	ahw->idc.vnic_wait_limit = QLCNIC_DEV_NPAR_OPER_TIMEO;
+
+	return 0;
+}
+
+int qlcnic_83xx_check_vnic_state(struct qlcnic_adapter *adapter)
+{
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	struct qlc_83xx_idc *idc = &ahw->idc;
+	u32 state;
+
+	state = QLCRDX(ahw, QLC_83XX_VNIC_STATE);
+	while (state != QLCNIC_DEV_NPAR_OPER && idc->vnic_wait_limit--) {
+		msleep(1000);
+		state = QLCRDX(ahw, QLC_83XX_VNIC_STATE);
+	}
+
+	if (!idc->vnic_wait_limit) {
+		dev_err(&adapter->pdev->dev,
+			"vNIC mode not operational, state check timed out.\n");
+		return -EIO;
+	}
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ctx.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ctx.c
index 6acf82b..0581a48 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ctx.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ctx.c
@@ -36,7 +36,8 @@ static const struct qlcnic_mailbox_metadata qlcnic_mbx_tbl[] = {
 	{QLCNIC_CMD_CONFIG_PORT, 4, 1},
 	{QLCNIC_CMD_TEMP_SIZE, 4, 4},
 	{QLCNIC_CMD_GET_TEMP_HDR, 4, 1},
-	{QLCNIC_CMD_SET_DRV_VER, 4, 1},
+	{QLCNIC_CMD_82XX_SET_DRV_VER, 4, 1},
+	{QLCNIC_CMD_GET_LED_STATUS, 4, 2},
 };
 
 static inline u32 qlcnic_get_cmd_signature(struct qlcnic_hardware_context *ahw)
@@ -181,7 +182,7 @@ int qlcnic_82xx_issue_cmd(struct qlcnic_adapter *adapter,
 	return cmd->rsp.arg[0];
 }
 
-int qlcnic_fw_cmd_set_drv_version(struct qlcnic_adapter *adapter)
+int qlcnic_fw_cmd_set_drv_version(struct qlcnic_adapter *adapter, u32 fw_cmd)
 {
 	struct qlcnic_cmd_args cmd;
 	u32 arg1, arg2, arg3;
@@ -193,7 +194,10 @@ int qlcnic_fw_cmd_set_drv_version(struct qlcnic_adapter *adapter)
 		 _QLCNIC_LINUX_MAJOR, _QLCNIC_LINUX_MINOR,
 		 _QLCNIC_LINUX_SUBVERSION);
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_DRV_VER);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, fw_cmd);
+	if (err)
+		return err;
+
 	memcpy(&arg1, drv_string, sizeof(u32));
 	memcpy(&arg2, drv_string + 4, sizeof(u32));
 	memcpy(&arg3, drv_string + 8, sizeof(u32));
@@ -221,7 +225,10 @@ qlcnic_fw_cmd_set_mtu(struct qlcnic_adapter *adapter, int mtu)
 
 	if (recv_ctx->state != QLCNIC_HOST_CTX_STATE_ACTIVE)
 		return err;
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_MTU);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_MTU);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = recv_ctx->context_id;
 	cmd.req.arg[2] = mtu;
 
@@ -335,7 +342,10 @@ int qlcnic_82xx_fw_cmd_create_rx_ctx(struct qlcnic_adapter *adapter)
 	}
 
 	phys_addr = hostrq_phys_addr;
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CREATE_RX_CTX);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CREATE_RX_CTX);
+	if (err)
+		goto out_free_rsp;
+
 	cmd.req.arg[1] = MSD(phys_addr);
 	cmd.req.arg[2] = LSD(phys_addr);
 	cmd.req.arg[3] = rq_size;
@@ -373,10 +383,10 @@ int qlcnic_82xx_fw_cmd_create_rx_ctx(struct qlcnic_adapter *adapter)
 	recv_ctx->context_id = le16_to_cpu(prsp->context_id);
 	recv_ctx->virt_port = prsp->virt_port;
 
+	qlcnic_free_mbx_args(&cmd);
 out_free_rsp:
 	dma_free_coherent(&adapter->pdev->dev, rsp_size, prsp,
-		cardrsp_phys_addr);
-	qlcnic_free_mbx_args(&cmd);
+			  cardrsp_phys_addr);
 out_free_rq:
 	dma_free_coherent(&adapter->pdev->dev, rq_size, prq, hostrq_phys_addr);
 	return err;
@@ -388,7 +398,10 @@ void qlcnic_82xx_fw_cmd_del_rx_ctx(struct qlcnic_adapter *adapter)
 	struct qlcnic_cmd_args cmd;
 	struct qlcnic_recv_context *recv_ctx = adapter->recv_ctx;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_DESTROY_RX_CTX);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_DESTROY_RX_CTX);
+	if (err)
+		return;
+
 	cmd.req.arg[1] = recv_ctx->context_id;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	if (err)
@@ -457,7 +470,10 @@ int qlcnic_82xx_fw_cmd_create_tx_ctx(struct qlcnic_adapter *adapter,
 
 	phys_addr = rq_phys_addr;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CREATE_TX_CTX);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CREATE_TX_CTX);
+	if (err)
+		goto out_free_rsp;
+
 	cmd.req.arg[1] = MSD(phys_addr);
 	cmd.req.arg[2] = LSD(phys_addr);
 	cmd.req.arg[3] = rq_size;
@@ -473,12 +489,13 @@ int qlcnic_82xx_fw_cmd_create_tx_ctx(struct qlcnic_adapter *adapter,
 		err = -EIO;
 	}
 
+	qlcnic_free_mbx_args(&cmd);
+
+out_free_rsp:
 	dma_free_coherent(&adapter->pdev->dev, rsp_size, rsp_addr,
 			  rsp_phys_addr);
-
 out_free_rq:
 	dma_free_coherent(&adapter->pdev->dev, rq_size, rq_addr, rq_phys_addr);
-	qlcnic_free_mbx_args(&cmd);
 
 	return err;
 }
@@ -487,8 +504,11 @@ void qlcnic_82xx_fw_cmd_del_tx_ctx(struct qlcnic_adapter *adapter,
 				   struct qlcnic_host_tx_ring *tx_ring)
 {
 	struct qlcnic_cmd_args cmd;
+	int ret;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_DESTROY_TX_CTX);
+	ret = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_DESTROY_TX_CTX);
+	if (ret)
+		return;
 
 	cmd.req.arg[1] = tx_ring->ctx_id;
 	if (qlcnic_issue_cmd(adapter, &cmd))
@@ -503,7 +523,10 @@ qlcnic_fw_cmd_set_port(struct qlcnic_adapter *adapter, u32 config)
 	int err;
 	struct qlcnic_cmd_args cmd;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIG_PORT);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIG_PORT);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = config;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	qlcnic_free_mbx_args(&cmd);
@@ -707,7 +730,10 @@ int qlcnic_82xx_get_mac_address(struct qlcnic_adapter *adapter, u8 *mac)
 	struct qlcnic_cmd_args cmd;
 	u32 mac_low, mac_high;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_MAC_ADDRESS);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_MAC_ADDRESS);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = adapter->ahw->pci_func | BIT_8;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 
@@ -746,7 +772,10 @@ int qlcnic_82xx_get_nic_info(struct qlcnic_adapter *adapter,
 
 	nic_info = nic_info_addr;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_NIC_INFO);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_NIC_INFO);
+	if (err)
+		goto out_free_dma;
+
 	cmd.req.arg[1] = MSD(nic_dma_t);
 	cmd.req.arg[2] = LSD(nic_dma_t);
 	cmd.req.arg[3] = (func_id << 16 | nic_size);
@@ -768,9 +797,10 @@ int qlcnic_82xx_get_nic_info(struct qlcnic_adapter *adapter,
 		npar_info->max_mtu = le16_to_cpu(nic_info->max_mtu);
 	}
 
+	qlcnic_free_mbx_args(&cmd);
+out_free_dma:
 	dma_free_coherent(&adapter->pdev->dev, nic_size, nic_info_addr,
 			  nic_dma_t);
-	qlcnic_free_mbx_args(&cmd);
 
 	return err;
 }
@@ -807,7 +837,10 @@ int qlcnic_82xx_set_nic_info(struct qlcnic_adapter *adapter,
 	nic_info->min_tx_bw = cpu_to_le16(nic->min_tx_bw);
 	nic_info->max_tx_bw = cpu_to_le16(nic->max_tx_bw);
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_NIC_INFO);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_NIC_INFO);
+	if (err)
+		goto out_free_dma;
+
 	cmd.req.arg[1] = MSD(nic_dma_t);
 	cmd.req.arg[2] = LSD(nic_dma_t);
 	cmd.req.arg[3] = ((nic->pci_func << 16) | nic_size);
@@ -819,9 +852,10 @@ int qlcnic_82xx_set_nic_info(struct qlcnic_adapter *adapter,
 		err = -EIO;
 	}
 
-	dma_free_coherent(&adapter->pdev->dev, nic_size, nic_info_addr,
-		nic_dma_t);
 	qlcnic_free_mbx_args(&cmd);
+out_free_dma:
+	dma_free_coherent(&adapter->pdev->dev, nic_size, nic_info_addr,
+			  nic_dma_t);
 
 	return err;
 }
@@ -845,7 +879,10 @@ int qlcnic_82xx_get_pci_info(struct qlcnic_adapter *adapter,
 		return -ENOMEM;
 
 	npar = pci_info_addr;
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_PCI_INFO);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_PCI_INFO);
+	if (err)
+		goto out_free_dma;
+
 	cmd.req.arg[1] = MSD(pci_info_dma_t);
 	cmd.req.arg[2] = LSD(pci_info_dma_t);
 	cmd.req.arg[3] = pci_size;
@@ -873,20 +910,22 @@ int qlcnic_82xx_get_pci_info(struct qlcnic_adapter *adapter,
 		err = -EIO;
 	}
 
+	qlcnic_free_mbx_args(&cmd);
+out_free_dma:
 	dma_free_coherent(&adapter->pdev->dev, pci_size, pci_info_addr,
 		pci_info_dma_t);
-	qlcnic_free_mbx_args(&cmd);
 
 	return err;
 }
 
 /* Configure eSwitch for port mirroring */
 int qlcnic_config_port_mirroring(struct qlcnic_adapter *adapter, u8 id,
-				u8 enable_mirroring, u8 pci_func)
+				 u8 enable_mirroring, u8 pci_func)
 {
+	struct device *dev = &adapter->pdev->dev;
+	struct qlcnic_cmd_args cmd;
 	int err = -EIO;
 	u32 arg1;
-	struct qlcnic_cmd_args cmd;
 
 	if (adapter->ahw->op_mode != QLCNIC_MGMT_FUNC ||
 	    !(adapter->eswitch[id].flags & QLCNIC_SWITCH_ENABLE))
@@ -895,18 +934,20 @@ int qlcnic_config_port_mirroring(struct qlcnic_adapter *adapter, u8 id,
 	arg1 = id | (enable_mirroring ? BIT_4 : 0);
 	arg1 |= pci_func << 8;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_SET_PORTMIRRORING);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_SET_PORTMIRRORING);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = arg1;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 
 	if (err != QLCNIC_RCODE_SUCCESS)
-		dev_err(&adapter->pdev->dev,
-			"Failed to configure port mirroring%d on eswitch:%d\n",
+		dev_err(dev, "Failed to configure port mirroring for vNIC function %d on eSwitch %d\n",
 			pci_func, id);
 	else
-		dev_info(&adapter->pdev->dev,
-			"Configured eSwitch %d for port mirroring:%d\n",
-			id, pci_func);
+		dev_info(dev, "Configured port mirroring for vNIC function %d on eSwitch %d\n",
+			 pci_func, id);
 	qlcnic_free_mbx_args(&cmd);
 
 	return err;
@@ -941,7 +982,11 @@ int qlcnic_get_port_stats(struct qlcnic_adapter *adapter, const u8 func,
 	arg1 = func | QLCNIC_STATS_VERSION << 8 | QLCNIC_STATS_PORT << 12;
 	arg1 |= rx_tx << 15 | stats_size << 16;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_ESWITCH_STATS);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_GET_ESWITCH_STATS);
+	if (err)
+		goto out_free_dma;
+
 	cmd.req.arg[1] = arg1;
 	cmd.req.arg[2] = MSD(stats_dma_t);
 	cmd.req.arg[3] = LSD(stats_dma_t);
@@ -963,9 +1008,10 @@ int qlcnic_get_port_stats(struct qlcnic_adapter *adapter, const u8 func,
 		esw_stats->numbytes = le64_to_cpu(stats->numbytes);
 	}
 
-	dma_free_coherent(&adapter->pdev->dev, stats_size, stats_addr,
-		stats_dma_t);
 	qlcnic_free_mbx_args(&cmd);
+out_free_dma:
+	dma_free_coherent(&adapter->pdev->dev, stats_size, stats_addr,
+			  stats_dma_t);
 
 	return err;
 }
@@ -989,7 +1035,10 @@ int qlcnic_get_mac_stats(struct qlcnic_adapter *adapter,
 	if (!stats_addr)
 		return -ENOMEM;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_MAC_STATS);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_MAC_STATS);
+	if (err)
+		goto out_free_dma;
+
 	cmd.req.arg[1] = stats_size << 16;
 	cmd.req.arg[2] = MSD(stats_dma_t);
 	cmd.req.arg[3] = LSD(stats_dma_t);
@@ -1020,11 +1069,12 @@ int qlcnic_get_mac_stats(struct qlcnic_adapter *adapter,
 			"%s: Get mac stats failed, err=%d.\n", __func__, err);
 	}
 
-	dma_free_coherent(&adapter->pdev->dev, stats_size, stats_addr,
-		stats_dma_t);
-
 	qlcnic_free_mbx_args(&cmd);
 
+out_free_dma:
+	dma_free_coherent(&adapter->pdev->dev, stats_size, stats_addr,
+			  stats_dma_t);
+
 	return err;
 }
 
@@ -1108,7 +1158,11 @@ int qlcnic_clear_esw_stats(struct qlcnic_adapter *adapter, const u8 func_esw,
 	arg1 = port | QLCNIC_STATS_VERSION << 8 | func_esw << 12;
 	arg1 |= BIT_14 | rx_tx << 15;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_ESWITCH_STATS);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_GET_ESWITCH_STATS);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = arg1;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	qlcnic_free_mbx_args(&cmd);
@@ -1121,17 +1175,19 @@ err_ret:
 	return -EIO;
 }
 
-static int
-__qlcnic_get_eswitch_port_config(struct qlcnic_adapter *adapter,
-					u32 *arg1, u32 *arg2)
+static int __qlcnic_get_eswitch_port_config(struct qlcnic_adapter *adapter,
+					    u32 *arg1, u32 *arg2)
 {
-	int err = -EIO;
+	struct device *dev = &adapter->pdev->dev;
 	struct qlcnic_cmd_args cmd;
-	u8 pci_func;
-	pci_func = (*arg1 >> 8);
+	u8 pci_func = *arg1 >> 8;
+	int err;
+
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_GET_ESWITCH_PORT_CONFIG);
+	if (err)
+		return err;
 
-	qlcnic_alloc_mbx_args(&cmd, adapter,
-			      QLCNIC_CMD_GET_ESWITCH_PORT_CONFIG);
 	cmd.req.arg[1] = *arg1;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	*arg1 = cmd.rsp.arg[1];
@@ -1139,12 +1195,11 @@ __qlcnic_get_eswitch_port_config(struct qlcnic_adapter *adapter,
 	qlcnic_free_mbx_args(&cmd);
 
 	if (err == QLCNIC_RCODE_SUCCESS)
-		dev_info(&adapter->pdev->dev,
-			 "eSwitch port config for pci func %d\n", pci_func);
+		dev_info(dev, "Get eSwitch port config for vNIC function %d\n",
+			 pci_func);
 	else
-		dev_err(&adapter->pdev->dev,
-			"Failed to get eswitch port config for pci func %d\n",
-								pci_func);
+		dev_err(dev, "Failed to get eswitch port config for vNIC function %d\n",
+			pci_func);
 	return err;
 }
 /* Configure eSwitch port
@@ -1157,9 +1212,10 @@ op_type = 1 for port vlan_id
 int qlcnic_config_switch_port(struct qlcnic_adapter *adapter,
 		struct qlcnic_esw_func_cfg *esw_cfg)
 {
+	struct device *dev = &adapter->pdev->dev;
+	struct qlcnic_cmd_args cmd;
 	int err = -EIO, index;
 	u32 arg1, arg2 = 0;
-	struct qlcnic_cmd_args cmd;
 	u8 pci_func;
 
 	if (adapter->ahw->op_mode != QLCNIC_MGMT_FUNC)
@@ -1209,18 +1265,22 @@ int qlcnic_config_switch_port(struct qlcnic_adapter *adapter,
 		return err;
 	}
 
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_CONFIGURE_ESWITCH);
+	err = qlcnic_alloc_mbx_args(&cmd, adapter,
+				    QLCNIC_CMD_CONFIGURE_ESWITCH);
+	if (err)
+		return err;
+
 	cmd.req.arg[1] = arg1;
 	cmd.req.arg[2] = arg2;
 	err = qlcnic_issue_cmd(adapter, &cmd);
 	qlcnic_free_mbx_args(&cmd);
 
 	if (err != QLCNIC_RCODE_SUCCESS)
-		dev_err(&adapter->pdev->dev,
-			"Failed to configure eswitch pci func %d\n", pci_func);
+		dev_err(dev, "Failed to configure eswitch for vNIC function %d\n",
+			pci_func);
 	else
-		dev_info(&adapter->pdev->dev,
-			 "Configured eSwitch for pci func %d\n", pci_func);
+		dev_info(dev, "Configured eSwitch for vNIC function %d\n",
+			 pci_func);
 
 	return err;
 }
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c
index f67652d..700a463 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_ethtool.c
@@ -846,7 +846,9 @@ static int qlcnic_irq_test(struct net_device *netdev)
 		goto clear_diag_irq;
 
 	ahw->diag_cnt = 0;
-	qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_INTRPT_TEST);
+	ret = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_INTRPT_TEST);
+	if (ret)
+		goto free_diag_res;
 
 	cmd.req.arg[1] = ahw->pci_func;
 	ret = qlcnic_issue_cmd(adapter, &cmd);
@@ -858,6 +860,8 @@ static int qlcnic_irq_test(struct net_device *netdev)
 
 done:
 	qlcnic_free_mbx_args(&cmd);
+
+free_diag_res:
 	qlcnic_diag_free_res(netdev, max_sds_rings);
 
 clear_diag_irq:
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hdr.h b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hdr.h
index c0f0c0d..d262211 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hdr.h
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hdr.h
@@ -672,6 +672,7 @@ enum {
 #define QLCNIC_HEARTBEAT_CHECK_RETRY_COUNT	10
 
 #define QLCNIC_MAX_MC_COUNT		38
+#define QLCNIC_MAX_UC_COUNT		512
 #define QLCNIC_WATCHDOG_TIMEOUTVALUE	5
 
 #define	ISR_MSI_INT_TRIGGER(FUNC) (QLCNIC_PCIX_PS_REG(PCIX_MSI_F(FUNC)))
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.c
index 106a12f..5b5d2ed 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.c
@@ -499,6 +499,7 @@ int qlcnic_nic_add_mac(struct qlcnic_adapter *adapter, const u8 *addr, u16 vlan)
 void __qlcnic_set_multi(struct net_device *netdev, u16 vlan)
 {
 	struct qlcnic_adapter *adapter = netdev_priv(netdev);
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
 	struct netdev_hw_addr *ha;
 	static const u8 bcast_addr[ETH_ALEN] = {
 		0xff, 0xff, 0xff, 0xff, 0xff, 0xff
@@ -515,25 +516,30 @@ void __qlcnic_set_multi(struct net_device *netdev, u16 vlan)
 	if (netdev->flags & IFF_PROMISC) {
 		if (!(adapter->flags & QLCNIC_PROMISC_DISABLED))
 			mode = VPORT_MISS_MODE_ACCEPT_ALL;
-		goto send_fw_cmd;
-	}
-
-	if ((netdev->flags & IFF_ALLMULTI) ||
-	    (netdev_mc_count(netdev) > adapter->ahw->max_mc_count)) {
-		mode = VPORT_MISS_MODE_ACCEPT_MULTI;
-		goto send_fw_cmd;
+	} else if (netdev->flags & IFF_ALLMULTI) {
+		if (netdev_mc_count(netdev) > ahw->max_mc_count) {
+			mode = VPORT_MISS_MODE_ACCEPT_MULTI;
+		} else if (!netdev_mc_empty(netdev) &&
+			   !qlcnic_sriov_vf_check(adapter)) {
+				netdev_for_each_mc_addr(ha, netdev)
+					qlcnic_nic_add_mac(adapter, ha->addr,
+							   vlan);
+		}
+		if (mode != VPORT_MISS_MODE_ACCEPT_MULTI &&
+		    qlcnic_sriov_vf_check(adapter))
+			qlcnic_vf_add_mc_list(netdev, vlan);
 	}
 
-	if (!netdev_mc_empty(netdev) && !qlcnic_sriov_vf_check(adapter)) {
-		netdev_for_each_mc_addr(ha, netdev) {
+	/* configure unicast MAC address, if there is not sufficient space
+	 * to store all the unicast addresses then enable promiscuous mode
+	 */
+	if (netdev_uc_count(netdev) > ahw->max_uc_count) {
+		mode = VPORT_MISS_MODE_ACCEPT_ALL;
+	} else if (!netdev_uc_empty(netdev)) {
+		netdev_for_each_uc_addr(ha, netdev)
 			qlcnic_nic_add_mac(adapter, ha->addr, vlan);
-		}
 	}
 
-	if (qlcnic_sriov_vf_check(adapter))
-		qlcnic_vf_add_mc_list(netdev, vlan);
-
-send_fw_cmd:
 	if (!qlcnic_sriov_vf_check(adapter)) {
 		if (mode == VPORT_MISS_MODE_ACCEPT_ALL &&
 		    !adapter->fdb_mac_learn) {
@@ -780,7 +786,8 @@ int qlcnic_82xx_config_hw_lro(struct qlcnic_adapter *adapter, int enable)
 	word = 0;
 	if (enable) {
 		word = QLCNIC_ENABLE_IPV4_LRO | QLCNIC_NO_DEST_IPV4_CHECK;
-		if (adapter->ahw->capabilities2 & QLCNIC_FW_CAP2_HW_LRO_IPV6)
+		if (adapter->ahw->extra_capability[0] &
+		    QLCNIC_FW_CAP2_HW_LRO_IPV6)
 			word |= QLCNIC_ENABLE_IPV6_LRO |
 				QLCNIC_NO_DEST_IPV6_CHECK;
 	}
@@ -1503,6 +1510,21 @@ int qlcnic_82xx_config_led(struct qlcnic_adapter *adapter, u32 state, u32 rate)
 	return rv;
 }
 
+int qlcnic_get_beacon_state(struct qlcnic_adapter *adapter, u8 *h_state)
+{
+	struct qlcnic_cmd_args cmd;
+	int err;
+
+	err = qlcnic_alloc_mbx_args(&cmd, adapter, QLCNIC_CMD_GET_LED_STATUS);
+	if (!err) {
+		err = qlcnic_issue_cmd(adapter, &cmd);
+		if (!err)
+			*h_state = cmd.rsp.arg[1];
+	}
+	qlcnic_free_mbx_args(&cmd);
+	return err;
+}
+
 void qlcnic_82xx_get_func_no(struct qlcnic_adapter *adapter)
 {
 	void __iomem *msix_base_addr;
@@ -1555,3 +1577,54 @@ void qlcnic_82xx_api_unlock(struct qlcnic_adapter *adapter)
 {
 	qlcnic_pcie_sem_unlock(adapter, 5);
 }
+
+int qlcnic_82xx_shutdown(struct pci_dev *pdev)
+{
+	struct qlcnic_adapter *adapter = pci_get_drvdata(pdev);
+	struct net_device *netdev = adapter->netdev;
+	int retval;
+
+	netif_device_detach(netdev);
+
+	qlcnic_cancel_idc_work(adapter);
+
+	if (netif_running(netdev))
+		qlcnic_down(adapter, netdev);
+
+	qlcnic_clr_all_drv_state(adapter, 0);
+
+	clear_bit(__QLCNIC_RESETTING, &adapter->state);
+
+	retval = pci_save_state(pdev);
+	if (retval)
+		return retval;
+
+	if (qlcnic_wol_supported(adapter)) {
+		pci_enable_wake(pdev, PCI_D3cold, 1);
+		pci_enable_wake(pdev, PCI_D3hot, 1);
+	}
+
+	return 0;
+}
+
+int qlcnic_82xx_resume(struct qlcnic_adapter *adapter)
+{
+	struct net_device *netdev = adapter->netdev;
+	int err;
+
+	err = qlcnic_start_firmware(adapter);
+	if (err) {
+		dev_err(&adapter->pdev->dev, "failed to start firmware\n");
+		return err;
+	}
+
+	if (netif_running(netdev)) {
+		err = qlcnic_up(adapter, netdev);
+		if (!err)
+			qlcnic_restore_indev_addr(netdev, NETDEV_UP);
+	}
+
+	netif_device_attach(netdev);
+	qlcnic_schedule_work(adapter, qlcnic_fw_poll_work, FW_POLL_DELAY);
+	return err;
+}
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.h b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.h
index b6818f4..2c22504 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.h
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_hw.h
@@ -86,7 +86,8 @@ enum qlcnic_regs {
 #define QLCNIC_CMD_BC_EVENT_SETUP		0x31
 #define	QLCNIC_CMD_CONFIG_VPORT			0x32
 #define QLCNIC_CMD_GET_MAC_STATS		0x37
-#define QLCNIC_CMD_SET_DRV_VER			0x38
+#define QLCNIC_CMD_82XX_SET_DRV_VER		0x38
+#define QLCNIC_CMD_GET_LED_STATUS		0x3C
 #define QLCNIC_CMD_CONFIGURE_RSS		0x41
 #define QLCNIC_CMD_CONFIG_INTR_COAL		0x43
 #define QLCNIC_CMD_CONFIGURE_LED		0x44
@@ -102,6 +103,7 @@ enum qlcnic_regs {
 #define QLCNIC_CMD_GET_LINK_STATUS		0x68
 #define QLCNIC_CMD_SET_LED_CONFIG		0x69
 #define QLCNIC_CMD_GET_LED_CONFIG		0x6A
+#define QLCNIC_CMD_83XX_SET_DRV_VER		0x6F
 #define QLCNIC_CMD_ADD_RCV_RINGS		0x0B
 
 #define QLCNIC_INTRPT_INTX			1
@@ -197,4 +199,8 @@ void qlcnic_82xx_api_unlock(struct qlcnic_adapter *);
 void qlcnic_82xx_napi_enable(struct qlcnic_adapter *);
 void qlcnic_82xx_napi_disable(struct qlcnic_adapter *);
 void qlcnic_82xx_napi_del(struct qlcnic_adapter *);
+int qlcnic_82xx_shutdown(struct pci_dev *);
+int qlcnic_82xx_resume(struct qlcnic_adapter *);
+void qlcnic_clr_all_drv_state(struct qlcnic_adapter *adapter, u8 failed);
+void qlcnic_fw_poll_work(struct work_struct *work);
 #endif				/* __QLCNIC_HW_H_ */
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c
index aeb26a85..4528f8e 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c
@@ -52,10 +52,6 @@ int qlcnic_load_fw_file;
 MODULE_PARM_DESC(load_fw_file, "Load firmware from (0=flash, 1=file)");
 module_param_named(load_fw_file, qlcnic_load_fw_file, int, 0444);
 
-int qlcnic_config_npars;
-module_param(qlcnic_config_npars, int, 0444);
-MODULE_PARM_DESC(qlcnic_config_npars, "Configure NPARs (0=disabled, 1=enabled)");
-
 static int qlcnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent);
 static void qlcnic_remove(struct pci_dev *pdev);
 static int qlcnic_open(struct net_device *netdev);
@@ -63,13 +59,11 @@ static int qlcnic_close(struct net_device *netdev);
 static void qlcnic_tx_timeout(struct net_device *netdev);
 static void qlcnic_attach_work(struct work_struct *work);
 static void qlcnic_fwinit_work(struct work_struct *work);
-static void qlcnic_fw_poll_work(struct work_struct *work);
 #ifdef CONFIG_NET_POLL_CONTROLLER
 static void qlcnic_poll_controller(struct net_device *netdev);
 #endif
 
 static void qlcnic_idc_debug_info(struct qlcnic_adapter *adapter, u8 encoding);
-static void qlcnic_clr_all_drv_state(struct qlcnic_adapter *adapter, u8);
 static int qlcnic_can_start_firmware(struct qlcnic_adapter *adapter);
 
 static irqreturn_t qlcnic_tmp_intr(int irq, void *data);
@@ -364,12 +358,15 @@ static int qlcnic_fdb_del(struct ndmsg *ndm, struct nlattr *tb[],
 		return ndo_dflt_fdb_del(ndm, tb, netdev, addr);
 
 	if (adapter->flags & QLCNIC_ESWITCH_ENABLED) {
-		if (is_unicast_ether_addr(addr))
-			err = qlcnic_nic_del_mac(adapter, addr);
-		else if (is_multicast_ether_addr(addr))
+		if (is_unicast_ether_addr(addr)) {
+			err = dev_uc_del(netdev, addr);
+			if (!err)
+				err = qlcnic_nic_del_mac(adapter, addr);
+		} else if (is_multicast_ether_addr(addr)) {
 			err = dev_mc_del(netdev, addr);
-		else
+		} else {
 			err =  -EINVAL;
+		}
 	}
 	return err;
 }
@@ -392,12 +389,16 @@ static int qlcnic_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
 	if (ether_addr_equal(addr, adapter->mac_addr))
 		return err;
 
-	if (is_unicast_ether_addr(addr))
-		err = qlcnic_nic_add_mac(adapter, addr, 0);
-	else if (is_multicast_ether_addr(addr))
+	if (is_unicast_ether_addr(addr)) {
+		if (netdev_uc_count(netdev) < adapter->ahw->max_uc_count)
+			err = dev_uc_add_excl(netdev, addr);
+		else
+			err = -ENOMEM;
+	} else if (is_multicast_ether_addr(addr)) {
 		err = dev_mc_add_excl(netdev, addr);
-	else
+	} else {
 		err = -EINVAL;
+	}
 
 	return err;
 }
@@ -449,6 +450,7 @@ static const struct net_device_ops qlcnic_netdev_ops = {
 	.ndo_set_vf_tx_rate	= qlcnic_sriov_set_vf_tx_rate,
 	.ndo_get_vf_config	= qlcnic_sriov_get_vf_config,
 	.ndo_set_vf_vlan	= qlcnic_sriov_set_vf_vlan,
+	.ndo_set_vf_spoofchk	= qlcnic_sriov_set_vf_spoofchk,
 #endif
 };
 
@@ -465,6 +467,8 @@ static struct qlcnic_nic_template qlcnic_ops = {
 	.napi_add		= qlcnic_82xx_napi_add,
 	.napi_del		= qlcnic_82xx_napi_del,
 	.config_ipaddr		= qlcnic_82xx_config_ipaddr,
+	.shutdown		= qlcnic_82xx_shutdown,
+	.resume			= qlcnic_82xx_resume,
 	.clear_legacy_intr	= qlcnic_82xx_clear_legacy_intr,
 };
 
@@ -508,6 +512,7 @@ static struct qlcnic_hardware_ops qlcnic_hw_ops = {
 	.config_promisc_mode		= qlcnic_82xx_nic_set_promisc,
 	.change_l2_filter		= qlcnic_82xx_change_filter,
 	.get_board_info			= qlcnic_82xx_get_board_info,
+	.set_mac_filter_count		= qlcnic_82xx_set_mac_filter_count,
 	.free_mac_list			= qlcnic_82xx_free_mac_list,
 };
 
@@ -768,7 +773,7 @@ static int
 qlcnic_set_function_modes(struct qlcnic_adapter *adapter)
 {
 	u8 id;
-	int i, ret = 1;
+	int ret;
 	u32 data = QLCNIC_MGMT_FUNC;
 	struct qlcnic_hardware_context *ahw = adapter->ahw;
 
@@ -776,20 +781,10 @@ qlcnic_set_function_modes(struct qlcnic_adapter *adapter)
 	if (ret)
 		goto err_lock;
 
-	if (qlcnic_config_npars) {
-		for (i = 0; i < ahw->act_pci_func; i++) {
-			id = adapter->npars[i].pci_func;
-			if (id == ahw->pci_func)
-				continue;
-			data |= (qlcnic_config_npars &
-					QLC_DEV_SET_DRV(0xf, id));
-		}
-	} else {
-		data = QLC_SHARED_REG_RD32(adapter, QLCNIC_DRV_OP_MODE);
-		data = (data & ~QLC_DEV_SET_DRV(0xf, ahw->pci_func)) |
-			(QLC_DEV_SET_DRV(QLCNIC_MGMT_FUNC,
-					 ahw->pci_func));
-	}
+	id = ahw->pci_func;
+	data = QLC_SHARED_REG_RD32(adapter, QLCNIC_DRV_OP_MODE);
+	data = (data & ~QLC_DEV_SET_DRV(0xf, id)) |
+	       QLC_DEV_SET_DRV(QLCNIC_MGMT_FUNC, id);
 	QLC_SHARED_REG_WR32(adapter, QLCNIC_DRV_OP_MODE, data);
 	qlcnic_api_unlock(adapter);
 err_lock:
@@ -875,6 +870,27 @@ static int qlcnic_setup_pci_map(struct pci_dev *pdev,
 	return 0;
 }
 
+static inline bool qlcnic_validate_subsystem_id(struct qlcnic_adapter *adapter,
+						int index)
+{
+	struct pci_dev *pdev = adapter->pdev;
+	unsigned short subsystem_vendor;
+	bool ret = true;
+
+	subsystem_vendor = pdev->subsystem_vendor;
+
+	if (pdev->device == PCI_DEVICE_ID_QLOGIC_QLE824X ||
+	    pdev->device == PCI_DEVICE_ID_QLOGIC_QLE834X) {
+		if (qlcnic_boards[index].sub_vendor == subsystem_vendor &&
+		    qlcnic_boards[index].sub_device == pdev->subsystem_device)
+			ret = true;
+		else
+			ret = false;
+	}
+
+	return ret;
+}
+
 static void qlcnic_get_board_name(struct qlcnic_adapter *adapter, char *name)
 {
 	struct pci_dev *pdev = adapter->pdev;
@@ -882,20 +898,18 @@ static void qlcnic_get_board_name(struct qlcnic_adapter *adapter, char *name)
 
 	for (i = 0; i < NUM_SUPPORTED_BOARDS; ++i) {
 		if (qlcnic_boards[i].vendor == pdev->vendor &&
-			qlcnic_boards[i].device == pdev->device &&
-			qlcnic_boards[i].sub_vendor == pdev->subsystem_vendor &&
-			qlcnic_boards[i].sub_device == pdev->subsystem_device) {
-				sprintf(name, "%pM: %s" ,
-					adapter->mac_addr,
-					qlcnic_boards[i].short_name);
-				found = 1;
-				break;
+		    qlcnic_boards[i].device == pdev->device &&
+		    qlcnic_validate_subsystem_id(adapter, i)) {
+			found = 1;
+			break;
 		}
-
 	}
 
 	if (!found)
 		sprintf(name, "%pM Gigabit Ethernet", adapter->mac_addr);
+	else
+		sprintf(name, "%pM: %s" , adapter->mac_addr,
+			qlcnic_boards[i].short_name);
 }
 
 static void
@@ -980,7 +994,7 @@ qlcnic_initialize_nic(struct qlcnic_adapter *adapter)
 	if (adapter->ahw->capabilities & QLCNIC_FW_CAPABILITY_MORE_CAPS) {
 		u32 temp;
 		temp = QLCRD32(adapter, CRB_FW_CAPABILITIES_2);
-		adapter->ahw->capabilities2 = temp;
+		adapter->ahw->extra_capability[0] = temp;
 	}
 	adapter->ahw->max_mac_filters = nic_info.max_mac_filters;
 	adapter->ahw->max_mtu = nic_info.max_mtu;
@@ -1395,16 +1409,23 @@ qlcnic_request_irq(struct qlcnic_adapter *adapter)
 			for (ring = 0; ring < num_sds_rings; ring++) {
 				sds_ring = &recv_ctx->sds_rings[ring];
 				if (qlcnic_82xx_check(adapter) &&
-				    (ring == (num_sds_rings - 1)))
-					snprintf(sds_ring->name,
-						 sizeof(sds_ring->name),
-						 "qlcnic-%s[Tx0+Rx%d]",
-						 netdev->name, ring);
-				else
+				    (ring == (num_sds_rings - 1))) {
+					if (!(adapter->flags &
+					      QLCNIC_MSIX_ENABLED))
+						snprintf(sds_ring->name,
+							 sizeof(sds_ring->name),
+							 "qlcnic");
+					else
+						snprintf(sds_ring->name,
+							 sizeof(sds_ring->name),
+							 "%s-tx-0-rx-%d",
+							 netdev->name, ring);
+				} else {
 					snprintf(sds_ring->name,
 						 sizeof(sds_ring->name),
-						 "qlcnic-%s[Rx%d]",
+						 "%s-rx-%d",
 						 netdev->name, ring);
+				}
 				err = request_irq(sds_ring->irq, handler, flags,
 						  sds_ring->name, sds_ring);
 				if (err)
@@ -1419,7 +1440,7 @@ qlcnic_request_irq(struct qlcnic_adapter *adapter)
 			     ring++) {
 				tx_ring = &adapter->tx_ring[ring];
 				snprintf(tx_ring->name, sizeof(tx_ring->name),
-					 "qlcnic-%s[Tx%d]", netdev->name, ring);
+					 "%s-tx-%d", netdev->name, ring);
 				err = request_irq(tx_ring->irq, handler, flags,
 						  tx_ring->name, tx_ring);
 				if (err)
@@ -1465,7 +1486,7 @@ static void qlcnic_get_lro_mss_capability(struct qlcnic_adapter *adapter)
 	u32 capab = 0;
 
 	if (qlcnic_82xx_check(adapter)) {
-		if (adapter->ahw->capabilities2 &
+		if (adapter->ahw->extra_capability[0] &
 		    QLCNIC_FW_CAPABILITY_2_LRO_MAX_TCP_SEG)
 			adapter->flags |= QLCNIC_FW_LRO_MSS_CAP;
 	} else {
@@ -1816,6 +1837,22 @@ qlcnic_reset_context(struct qlcnic_adapter *adapter)
 	return err;
 }
 
+void qlcnic_82xx_set_mac_filter_count(struct qlcnic_adapter *adapter)
+{
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	u16 act_pci_fn = ahw->act_pci_func;
+	u16 count;
+
+	ahw->max_mc_count = QLCNIC_MAX_MC_COUNT;
+	if (act_pci_fn <= 2)
+		count = (QLCNIC_MAX_UC_COUNT - QLCNIC_MAX_MC_COUNT) /
+			 act_pci_fn;
+	else
+		count = (QLCNIC_LB_MAX_FILTERS - QLCNIC_MAX_MC_COUNT) /
+			 act_pci_fn;
+	ahw->max_uc_count = count;
+}
+
 int
 qlcnic_setup_netdev(struct qlcnic_adapter *adapter, struct net_device *netdev,
 		    int pci_using_dac)
@@ -1825,7 +1862,7 @@ qlcnic_setup_netdev(struct qlcnic_adapter *adapter, struct net_device *netdev,
 
 	adapter->rx_csum = 1;
 	adapter->ahw->mc_enabled = 0;
-	adapter->ahw->max_mc_count = QLCNIC_MAX_MC_COUNT;
+	qlcnic_set_mac_filter_count(adapter);
 
 	netdev->netdev_ops	   = &qlcnic_netdev_ops;
 	netdev->watchdog_timeo     = QLCNIC_WATCHDOG_TIMEOUTVALUE * HZ;
@@ -1863,6 +1900,7 @@ qlcnic_setup_netdev(struct qlcnic_adapter *adapter, struct net_device *netdev,
 		netdev->features |= NETIF_F_LRO;
 
 	netdev->hw_features = netdev->features;
+	netdev->priv_flags |= IFF_UNICAST_FLT;
 	netdev->irq = adapter->msix_entries[0].vector;
 
 	err = register_netdev(netdev);
@@ -1947,6 +1985,21 @@ int qlcnic_alloc_tx_rings(struct qlcnic_adapter *adapter,
 	return 0;
 }
 
+void qlcnic_set_drv_version(struct qlcnic_adapter *adapter)
+{
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	u32 fw_cmd = 0;
+
+	if (qlcnic_82xx_check(adapter))
+		fw_cmd = QLCNIC_CMD_82XX_SET_DRV_VER;
+	else if (qlcnic_83xx_check(adapter))
+		fw_cmd = QLCNIC_CMD_83XX_SET_DRV_VER;
+
+	if ((ahw->capabilities & QLCNIC_FW_CAPABILITY_MORE_CAPS) &&
+	    (ahw->extra_capability[0] & QLCNIC_FW_CAPABILITY_SET_DRV_VER))
+		qlcnic_fw_cmd_set_drv_version(adapter, fw_cmd);
+}
+
 static int
 qlcnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 {
@@ -1954,7 +2007,6 @@ qlcnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	struct qlcnic_adapter *adapter = NULL;
 	struct qlcnic_hardware_context *ahw;
 	int err, pci_using_dac = -1;
-	u32 capab2;
 	char board_name[QLCNIC_MAX_BOARD_NAME_LEN + 19]; /* MAC + ": " + name */
 
 	if (pdev->is_virtfn)
@@ -2109,13 +2161,7 @@ qlcnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (err)
 		goto err_out_disable_mbx_intr;
 
-	if (qlcnic_82xx_check(adapter)) {
-		if (ahw->capabilities & QLCNIC_FW_CAPABILITY_MORE_CAPS) {
-			capab2 = QLCRD32(adapter, CRB_FW_CAPABILITIES_2);
-			if (capab2 & QLCNIC_FW_CAPABILITY_2_OCBB)
-				qlcnic_fw_cmd_set_drv_version(adapter);
-		}
-	}
+	qlcnic_set_drv_version(adapter);
 
 	pci_set_drvdata(pdev, adapter);
 
@@ -2231,37 +2277,6 @@ static void qlcnic_remove(struct pci_dev *pdev)
 	kfree(ahw);
 	free_netdev(netdev);
 }
-static int __qlcnic_shutdown(struct pci_dev *pdev)
-{
-	struct qlcnic_adapter *adapter = pci_get_drvdata(pdev);
-	struct net_device *netdev = adapter->netdev;
-	int retval;
-
-	netif_device_detach(netdev);
-
-	qlcnic_cancel_idc_work(adapter);
-
-	if (netif_running(netdev))
-		qlcnic_down(adapter, netdev);
-
-	qlcnic_sriov_cleanup(adapter);
-	if (qlcnic_82xx_check(adapter))
-		qlcnic_clr_all_drv_state(adapter, 0);
-
-	clear_bit(__QLCNIC_RESETTING, &adapter->state);
-
-	retval = pci_save_state(pdev);
-	if (retval)
-		return retval;
-	if (qlcnic_82xx_check(adapter)) {
-		if (qlcnic_wol_supported(adapter)) {
-			pci_enable_wake(pdev, PCI_D3cold, 1);
-			pci_enable_wake(pdev, PCI_D3hot, 1);
-		}
-	}
-
-	return 0;
-}
 
 static void qlcnic_shutdown(struct pci_dev *pdev)
 {
@@ -2272,8 +2287,7 @@ static void qlcnic_shutdown(struct pci_dev *pdev)
 }
 
 #ifdef CONFIG_PM
-static int
-qlcnic_suspend(struct pci_dev *pdev, pm_message_t state)
+static int qlcnic_suspend(struct pci_dev *pdev, pm_message_t state)
 {
 	int retval;
 
@@ -2285,11 +2299,9 @@ qlcnic_suspend(struct pci_dev *pdev, pm_message_t state)
 	return 0;
 }
 
-static int
-qlcnic_resume(struct pci_dev *pdev)
+static int qlcnic_resume(struct pci_dev *pdev)
 {
 	struct qlcnic_adapter *adapter = pci_get_drvdata(pdev);
-	struct net_device *netdev = adapter->netdev;
 	int err;
 
 	err = pci_enable_device(pdev);
@@ -2300,23 +2312,7 @@ qlcnic_resume(struct pci_dev *pdev)
 	pci_set_master(pdev);
 	pci_restore_state(pdev);
 
-	err = qlcnic_start_firmware(adapter);
-	if (err) {
-		dev_err(&pdev->dev, "failed to start firmware\n");
-		return err;
-	}
-
-	if (netif_running(netdev)) {
-		err = qlcnic_up(adapter, netdev);
-		if (err)
-			goto done;
-
-		qlcnic_restore_indev_addr(netdev, NETDEV_UP);
-	}
-done:
-	netif_device_attach(netdev);
-	qlcnic_schedule_work(adapter, qlcnic_fw_poll_work, FW_POLL_DELAY);
-	return 0;
+	return  __qlcnic_resume(adapter);
 }
 #endif
 
@@ -2655,8 +2651,7 @@ qlcnic_clr_drv_state(struct qlcnic_adapter *adapter)
 	return 0;
 }
 
-static void
-qlcnic_clr_all_drv_state(struct qlcnic_adapter *adapter, u8 failed)
+void qlcnic_clr_all_drv_state(struct qlcnic_adapter *adapter, u8 failed)
 {
 	u32  val;
 
@@ -3086,6 +3081,7 @@ done:
 	adapter->fw_fail_cnt = 0;
 	adapter->flags &= ~QLCNIC_FW_HANG;
 	clear_bit(__QLCNIC_RESETTING, &adapter->state);
+	qlcnic_set_drv_version(adapter);
 
 	if (!qlcnic_clr_drv_state(adapter))
 		qlcnic_schedule_work(adapter, qlcnic_fw_poll_work,
@@ -3166,8 +3162,7 @@ detach:
 	return 1;
 }
 
-static void
-qlcnic_fw_poll_work(struct work_struct *work)
+void qlcnic_fw_poll_work(struct work_struct *work)
 {
 	struct qlcnic_adapter *adapter = container_of(work,
 				struct qlcnic_adapter, fw_work.work);
@@ -3219,7 +3214,6 @@ static int qlcnic_attach_func(struct pci_dev *pdev)
 	if (err)
 		return err;
 
-	pci_set_power_state(pdev, PCI_D0);
 	pci_set_master(pdev);
 	pci_restore_state(pdev);
 
@@ -3517,7 +3511,7 @@ static int qlcnic_netdev_event(struct notifier_block *this,
 				 unsigned long event, void *ptr)
 {
 	struct qlcnic_adapter *adapter;
-	struct net_device *dev = (struct net_device *)ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 
 recheck:
 	if (dev == NULL)
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_minidump.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_minidump.c
index 4b9bab1..ab8a674 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_minidump.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_minidump.c
@@ -15,6 +15,7 @@
 #define QLC_83XX_MINIDUMP_FLASH		0x520000
 #define QLC_83XX_OCM_INDEX			3
 #define QLC_83XX_PCI_INDEX			0
+#define QLC_83XX_DMA_ENGINE_INDEX		8
 
 static const u32 qlcnic_ms_read_data[] = {
 	0x410000A8, 0x410000AC, 0x410000B8, 0x410000BC
@@ -32,6 +33,16 @@ static const u32 qlcnic_ms_read_data[] = {
 
 #define QLCNIC_DUMP_MASK_MAX	0xff
 
+struct qlcnic_pex_dma_descriptor {
+	u32	read_data_size;
+	u32	dma_desc_cmd;
+	u32	src_addr_low;
+	u32	src_addr_high;
+	u32	dma_bus_addr_low;
+	u32	dma_bus_addr_high;
+	u32	rsvd[6];
+} __packed;
+
 struct qlcnic_common_entry_hdr {
 	u32     type;
 	u32     offset;
@@ -90,7 +101,10 @@ struct __ocm {
 } __packed;
 
 struct __mem {
-	u8	rsvd[24];
+	u32	desc_card_addr;
+	u32	dma_desc_cmd;
+	u32	start_dma_cmd;
+	u32	rsvd[3];
 	u32	addr;
 	u32	size;
 } __packed;
@@ -466,12 +480,12 @@ skip_poll:
 	return l2->no_ops * l2->read_addr_num * sizeof(u32);
 }
 
-static u32 qlcnic_read_memory(struct qlcnic_adapter *adapter,
-			      struct qlcnic_dump_entry *entry, __le32 *buffer)
+static u32 qlcnic_read_memory_test_agent(struct qlcnic_adapter *adapter,
+					 struct __mem *mem, __le32 *buffer,
+					 int *ret)
 {
-	u32 addr, data, test, ret = 0;
+	u32 addr, data, test;
 	int i, reg_read;
-	struct __mem *mem = &entry->region.mem;
 
 	reg_read = mem->size;
 	addr = mem->addr;
@@ -480,7 +494,8 @@ static u32 qlcnic_read_memory(struct qlcnic_adapter *adapter,
 		dev_info(&adapter->pdev->dev,
 			 "Unaligned memory addr:0x%x size:0x%x\n",
 			 addr, reg_read);
-		return -EINVAL;
+		*ret = -EINVAL;
+		return 0;
 	}
 
 	mutex_lock(&adapter->ahw->mem_lock);
@@ -499,7 +514,7 @@ static u32 qlcnic_read_memory(struct qlcnic_adapter *adapter,
 			if (printk_ratelimit()) {
 				dev_err(&adapter->pdev->dev,
 					"failed to read through agent\n");
-				ret = -EINVAL;
+				*ret = -EIO;
 				goto out;
 			}
 		}
@@ -516,6 +531,181 @@ out:
 	return mem->size;
 }
 
+/* DMA register base address */
+#define QLC_DMA_REG_BASE_ADDR(dma_no)	(0x77320000 + (dma_no * 0x10000))
+
+/* DMA register offsets w.r.t base address */
+#define QLC_DMA_CMD_BUFF_ADDR_LOW	0
+#define QLC_DMA_CMD_BUFF_ADDR_HI	4
+#define QLC_DMA_CMD_STATUS_CTRL		8
+
+#define QLC_PEX_DMA_READ_SIZE		(PAGE_SIZE * 16)
+
+static int qlcnic_start_pex_dma(struct qlcnic_adapter *adapter,
+				struct __mem *mem)
+{
+	struct qlcnic_dump_template_hdr *tmpl_hdr;
+	struct device *dev = &adapter->pdev->dev;
+	u32 dma_no, dma_base_addr, temp_addr;
+	int i, ret, dma_sts;
+
+	tmpl_hdr = adapter->ahw->fw_dump.tmpl_hdr;
+	dma_no = tmpl_hdr->saved_state[QLC_83XX_DMA_ENGINE_INDEX];
+	dma_base_addr = QLC_DMA_REG_BASE_ADDR(dma_no);
+
+	temp_addr = dma_base_addr + QLC_DMA_CMD_BUFF_ADDR_LOW;
+	ret = qlcnic_83xx_wrt_reg_indirect(adapter, temp_addr,
+					   mem->desc_card_addr);
+	if (ret)
+		return ret;
+
+	temp_addr = dma_base_addr + QLC_DMA_CMD_BUFF_ADDR_HI;
+	ret = qlcnic_83xx_wrt_reg_indirect(adapter, temp_addr, 0);
+	if (ret)
+		return ret;
+
+	temp_addr = dma_base_addr + QLC_DMA_CMD_STATUS_CTRL;
+	ret = qlcnic_83xx_wrt_reg_indirect(adapter, temp_addr,
+					   mem->start_dma_cmd);
+	if (ret)
+		return ret;
+
+	/* Wait for DMA to complete */
+	temp_addr = dma_base_addr + QLC_DMA_CMD_STATUS_CTRL;
+	for (i = 0; i < 400; i++) {
+		dma_sts = qlcnic_ind_rd(adapter, temp_addr);
+
+		if (dma_sts & BIT_1)
+			usleep_range(250, 500);
+		else
+			break;
+	}
+
+	if (i >= 400) {
+		dev_info(dev, "PEX DMA operation timed out");
+		ret = -EIO;
+	}
+
+	return ret;
+}
+
+static u32 qlcnic_read_memory_pexdma(struct qlcnic_adapter *adapter,
+				     struct __mem *mem,
+				     __le32 *buffer, int *ret)
+{
+	struct qlcnic_fw_dump *fw_dump = &adapter->ahw->fw_dump;
+	u32 temp, dma_base_addr, size = 0, read_size = 0;
+	struct qlcnic_pex_dma_descriptor *dma_descr;
+	struct qlcnic_dump_template_hdr *tmpl_hdr;
+	struct device *dev = &adapter->pdev->dev;
+	dma_addr_t dma_phys_addr;
+	void *dma_buffer;
+
+	tmpl_hdr = fw_dump->tmpl_hdr;
+
+	/* Check if DMA engine is available */
+	temp = tmpl_hdr->saved_state[QLC_83XX_DMA_ENGINE_INDEX];
+	dma_base_addr = QLC_DMA_REG_BASE_ADDR(temp);
+	temp = qlcnic_ind_rd(adapter,
+			     dma_base_addr + QLC_DMA_CMD_STATUS_CTRL);
+
+	if (!(temp & BIT_31)) {
+		dev_info(dev, "%s: DMA engine is not available\n", __func__);
+		*ret = -EIO;
+		return 0;
+	}
+
+	/* Create DMA descriptor */
+	dma_descr = kzalloc(sizeof(struct qlcnic_pex_dma_descriptor),
+			    GFP_KERNEL);
+	if (!dma_descr) {
+		*ret = -ENOMEM;
+		return 0;
+	}
+
+	/* dma_desc_cmd  0:15  = 0
+	 * dma_desc_cmd 16:19  = mem->dma_desc_cmd 0:3
+	 * dma_desc_cmd 20:23  = pci function number
+	 * dma_desc_cmd 24:31  = mem->dma_desc_cmd 8:15
+	 */
+	dma_phys_addr = fw_dump->phys_addr;
+	dma_buffer = fw_dump->dma_buffer;
+	temp = 0;
+	temp = mem->dma_desc_cmd & 0xff0f;
+	temp |= (adapter->ahw->pci_func & 0xf) << 4;
+	dma_descr->dma_desc_cmd = (temp << 16) & 0xffff0000;
+	dma_descr->dma_bus_addr_low = LSD(dma_phys_addr);
+	dma_descr->dma_bus_addr_high = MSD(dma_phys_addr);
+	dma_descr->src_addr_high = 0;
+
+	/* Collect memory dump using multiple DMA operations if required */
+	while (read_size < mem->size) {
+		if (mem->size - read_size >= QLC_PEX_DMA_READ_SIZE)
+			size = QLC_PEX_DMA_READ_SIZE;
+		else
+			size = mem->size - read_size;
+
+		dma_descr->src_addr_low = mem->addr + read_size;
+		dma_descr->read_data_size = size;
+
+		/* Write DMA descriptor to MS memory*/
+		temp = sizeof(struct qlcnic_pex_dma_descriptor) / 16;
+		*ret = qlcnic_83xx_ms_mem_write128(adapter, mem->desc_card_addr,
+						   (u32 *)dma_descr, temp);
+		if (*ret) {
+			dev_info(dev, "Failed to write DMA descriptor to MS memory at address 0x%x\n",
+				 mem->desc_card_addr);
+			goto free_dma_descr;
+		}
+
+		*ret = qlcnic_start_pex_dma(adapter, mem);
+		if (*ret) {
+			dev_info(dev, "Failed to start PEX DMA operation\n");
+			goto free_dma_descr;
+		}
+
+		memcpy(buffer, dma_buffer, size);
+		buffer += size / 4;
+		read_size += size;
+	}
+
+free_dma_descr:
+	kfree(dma_descr);
+
+	return read_size;
+}
+
+static u32 qlcnic_read_memory(struct qlcnic_adapter *adapter,
+			      struct qlcnic_dump_entry *entry, __le32 *buffer)
+{
+	struct qlcnic_fw_dump *fw_dump = &adapter->ahw->fw_dump;
+	struct device *dev = &adapter->pdev->dev;
+	struct __mem *mem = &entry->region.mem;
+	u32 data_size;
+	int ret = 0;
+
+	if (fw_dump->use_pex_dma) {
+		data_size = qlcnic_read_memory_pexdma(adapter, mem, buffer,
+						      &ret);
+		if (ret)
+			dev_info(dev,
+				 "Failed to read memory dump using PEX DMA: mask[0x%x]\n",
+				 entry->hdr.mask);
+		else
+			return data_size;
+	}
+
+	data_size = qlcnic_read_memory_test_agent(adapter, mem, buffer, &ret);
+	if (ret) {
+		dev_info(dev,
+			 "Failed to read memory dump using test agent method: mask[0x%x]\n",
+			 entry->hdr.mask);
+		return 0;
+	} else {
+		return data_size;
+	}
+}
+
 static u32 qlcnic_dump_nop(struct qlcnic_adapter *adapter,
 			   struct qlcnic_dump_entry *entry, __le32 *buffer)
 {
@@ -893,6 +1083,12 @@ flash_temp:
 
 	tmpl_hdr = ahw->fw_dump.tmpl_hdr;
 	tmpl_hdr->drv_cap_mask = QLCNIC_DUMP_MASK_DEF;
+
+	if ((tmpl_hdr->version & 0xffffff) >= 0x20001)
+		ahw->fw_dump.use_pex_dma = true;
+	else
+		ahw->fw_dump.use_pex_dma = false;
+
 	ahw->fw_dump.enable = 1;
 
 	return 0;
@@ -910,7 +1106,9 @@ int qlcnic_dump_fw(struct qlcnic_adapter *adapter)
 	struct qlcnic_fw_dump *fw_dump = &adapter->ahw->fw_dump;
 	struct qlcnic_dump_template_hdr *tmpl_hdr = fw_dump->tmpl_hdr;
 	static const struct qlcnic_dump_operations *fw_dump_ops;
+	struct device *dev = &adapter->pdev->dev;
 	struct qlcnic_hardware_context *ahw;
+	void *temp_buffer;
 
 	ahw = adapter->ahw;
 
@@ -944,6 +1142,16 @@ int qlcnic_dump_fw(struct qlcnic_adapter *adapter)
 	tmpl_hdr->sys_info[0] = QLCNIC_DRIVER_VERSION;
 	tmpl_hdr->sys_info[1] = adapter->fw_version;
 
+	if (fw_dump->use_pex_dma) {
+		temp_buffer = dma_alloc_coherent(dev, QLC_PEX_DMA_READ_SIZE,
+						 &fw_dump->phys_addr,
+						 GFP_KERNEL);
+		if (!temp_buffer)
+			fw_dump->use_pex_dma = false;
+		else
+			fw_dump->dma_buffer = temp_buffer;
+	}
+
 	if (qlcnic_82xx_check(adapter)) {
 		ops_cnt = ARRAY_SIZE(qlcnic_fw_dump_ops);
 		fw_dump_ops = qlcnic_fw_dump_ops;
@@ -1002,6 +1210,9 @@ int qlcnic_dump_fw(struct qlcnic_adapter *adapter)
 		return 0;
 	}
 error:
+	if (fw_dump->use_pex_dma)
+		dma_free_coherent(dev, QLC_PEX_DMA_READ_SIZE,
+				  fw_dump->dma_buffer, fw_dump->phys_addr);
 	vfree(fw_dump->data);
 	return -EINVAL;
 }
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov.h b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov.h
index d85fbb5..0daf660 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov.h
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov.h
@@ -129,6 +129,7 @@ struct qlcnic_vport {
 	u8			vlan_mode;
 	u16			vlan;
 	u8			qos;
+	bool			spoofchk;
 	u8			mac[6];
 };
 
@@ -194,6 +195,8 @@ int __qlcnic_sriov_add_act_list(struct qlcnic_sriov *, struct qlcnic_vf_info *,
 int qlcnic_sriov_get_vf_vport_info(struct qlcnic_adapter *,
 				   struct qlcnic_info *, u16);
 int qlcnic_sriov_cfg_vf_guest_vlan(struct qlcnic_adapter *, u16, u8);
+int qlcnic_sriov_vf_shutdown(struct pci_dev *);
+int qlcnic_sriov_vf_resume(struct qlcnic_adapter *);
 
 static inline bool qlcnic_sriov_enable_check(struct qlcnic_adapter *adapter)
 {
@@ -225,6 +228,7 @@ int qlcnic_sriov_set_vf_tx_rate(struct net_device *, int, int);
 int qlcnic_sriov_get_vf_config(struct net_device *, int ,
 			       struct ifla_vf_info *);
 int qlcnic_sriov_set_vf_vlan(struct net_device *, int, u16, u8);
+int qlcnic_sriov_set_vf_spoofchk(struct net_device *, int, bool);
 #else
 static inline void qlcnic_sriov_pf_disable(struct qlcnic_adapter *adapter) {}
 static inline void qlcnic_sriov_pf_cleanup(struct qlcnic_adapter *adapter) {}
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_common.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_common.c
index 8b59a71..62380ce 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_common.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_common.c
@@ -35,6 +35,7 @@ static void qlcnic_sriov_vf_cancel_fw_work(struct qlcnic_adapter *);
 static void qlcnic_sriov_cleanup_transaction(struct qlcnic_bc_trans *);
 static int qlcnic_sriov_vf_mbx_op(struct qlcnic_adapter *,
 				  struct qlcnic_cmd_args *);
+static void qlcnic_sriov_process_bc_cmd(struct work_struct *);
 
 static struct qlcnic_hardware_ops qlcnic_sriov_vf_hw_ops = {
 	.read_crb			= qlcnic_83xx_read_crb,
@@ -75,6 +76,8 @@ static struct qlcnic_nic_template qlcnic_sriov_vf_ops = {
 	.cancel_idc_work        = qlcnic_sriov_vf_cancel_fw_work,
 	.napi_add		= qlcnic_83xx_napi_add,
 	.napi_del		= qlcnic_83xx_napi_del,
+	.shutdown		= qlcnic_sriov_vf_shutdown,
+	.resume			= qlcnic_sriov_vf_resume,
 	.config_ipaddr		= qlcnic_83xx_config_ipaddr,
 	.clear_legacy_intr	= qlcnic_83xx_clear_legacy_intr,
 };
@@ -179,6 +182,8 @@ int qlcnic_sriov_init(struct qlcnic_adapter *adapter, int num_vfs)
 		spin_lock_init(&vf->rcv_pend.lock);
 		init_completion(&vf->ch_free_cmpl);
 
+		INIT_WORK(&vf->trans_work, qlcnic_sriov_process_bc_cmd);
+
 		if (qlcnic_sriov_pf_check(adapter)) {
 			vp = kzalloc(sizeof(struct qlcnic_vport), GFP_KERNEL);
 			if (!vp) {
@@ -187,6 +192,7 @@ int qlcnic_sriov_init(struct qlcnic_adapter *adapter, int num_vfs)
 			}
 			sriov->vf_info[i].vp = vp;
 			vp->max_tx_bw = MAX_BW;
+			vp->spoofchk = true;
 			random_ether_addr(vp->mac);
 			dev_info(&adapter->pdev->dev,
 				 "MAC Address %pM is configured for VF %d\n",
@@ -652,6 +658,8 @@ int qlcnic_sriov_vf_init(struct qlcnic_adapter *adapter, int pci_using_dac)
 	if (qlcnic_read_mac_addr(adapter))
 		dev_warn(&adapter->pdev->dev, "failed to read mac addr\n");
 
+	INIT_DELAYED_WORK(&adapter->idc_aen_work, qlcnic_83xx_idc_aen_work);
+
 	clear_bit(__QLCNIC_RESETTING, &adapter->state);
 	return 0;
 }
@@ -864,7 +872,6 @@ static void qlcnic_sriov_schedule_bc_cmd(struct qlcnic_sriov *sriov,
 	    vf->adapter->need_fw_reset)
 		return;
 
-	INIT_WORK(&vf->trans_work, func);
 	queue_work(sriov->bc.bc_trans_wq, &vf->trans_work);
 }
 
@@ -1949,3 +1956,54 @@ static void qlcnic_sriov_vf_free_mac_list(struct qlcnic_adapter *adapter)
 		kfree(cur);
 	}
 }
+
+int qlcnic_sriov_vf_shutdown(struct pci_dev *pdev)
+{
+	struct qlcnic_adapter *adapter = pci_get_drvdata(pdev);
+	struct net_device *netdev = adapter->netdev;
+	int retval;
+
+	netif_device_detach(netdev);
+	qlcnic_cancel_idc_work(adapter);
+
+	if (netif_running(netdev))
+		qlcnic_down(adapter, netdev);
+
+	qlcnic_sriov_channel_cfg_cmd(adapter, QLCNIC_BC_CMD_CHANNEL_TERM);
+	qlcnic_sriov_cfg_bc_intr(adapter, 0);
+	qlcnic_83xx_disable_mbx_intr(adapter);
+	cancel_delayed_work_sync(&adapter->idc_aen_work);
+
+	retval = pci_save_state(pdev);
+	if (retval)
+		return retval;
+
+	return 0;
+}
+
+int qlcnic_sriov_vf_resume(struct qlcnic_adapter *adapter)
+{
+	struct qlc_83xx_idc *idc = &adapter->ahw->idc;
+	struct net_device *netdev = adapter->netdev;
+	int err;
+
+	set_bit(QLC_83XX_MODULE_LOADED, &idc->status);
+	qlcnic_83xx_enable_mbx_intrpt(adapter);
+	err = qlcnic_sriov_cfg_bc_intr(adapter, 1);
+	if (err)
+		return err;
+
+	err = qlcnic_sriov_channel_cfg_cmd(adapter, QLCNIC_BC_CMD_CHANNEL_INIT);
+	if (!err) {
+		if (netif_running(netdev)) {
+			err = qlcnic_up(adapter, netdev);
+			if (!err)
+				qlcnic_restore_indev_addr(netdev, NETDEV_UP);
+		}
+	}
+
+	netif_device_attach(netdev);
+	qlcnic_schedule_work(adapter, qlcnic_sriov_vf_poll_dev_state,
+			     idc->delay);
+	return err;
+}
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c
index 1a66ccd..ee0c1d3 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c
@@ -580,6 +580,7 @@ static int qlcnic_sriov_set_vf_acl(struct qlcnic_adapter *adapter, u8 func)
 	struct qlcnic_cmd_args cmd;
 	struct qlcnic_vport *vp;
 	int err, id;
+	u8 *mac;
 
 	id = qlcnic_sriov_func_to_index(adapter, func);
 	if (id < 0)
@@ -591,6 +592,14 @@ static int qlcnic_sriov_set_vf_acl(struct qlcnic_adapter *adapter, u8 func)
 		return err;
 
 	cmd.req.arg[1] = 0x3 | func << 16;
+	if (vp->spoofchk == true) {
+		mac = vp->mac;
+		cmd.req.arg[2] |= BIT_1 | BIT_3 | BIT_8;
+		cmd.req.arg[4] = mac[5] | mac[4] << 8 | mac[3] << 16 |
+				 mac[2] << 24;
+		cmd.req.arg[5] = mac[1] | mac[0] << 8;
+	}
+
 	if (vp->vlan_mode == QLC_PVID_MODE) {
 		cmd.req.arg[2] |= BIT_6;
 		cmd.req.arg[3] |= vp->vlan << 8;
@@ -1767,6 +1776,7 @@ int qlcnic_sriov_get_vf_config(struct net_device *netdev,
 	memcpy(&ivi->mac, vp->mac, ETH_ALEN);
 	ivi->vlan = vp->vlan;
 	ivi->qos = vp->qos;
+	ivi->spoofchk = vp->spoofchk;
 	if (vp->max_tx_bw == MAX_BW)
 		ivi->tx_rate = 0;
 	else
@@ -1775,3 +1785,29 @@ int qlcnic_sriov_get_vf_config(struct net_device *netdev,
 	ivi->vf = vf;
 	return 0;
 }
+
+int qlcnic_sriov_set_vf_spoofchk(struct net_device *netdev, int vf, bool chk)
+{
+	struct qlcnic_adapter *adapter = netdev_priv(netdev);
+	struct qlcnic_sriov *sriov = adapter->ahw->sriov;
+	struct qlcnic_vf_info *vf_info;
+	struct qlcnic_vport *vp;
+
+	if (!qlcnic_sriov_pf_check(adapter))
+		return -EOPNOTSUPP;
+
+	if (vf >= sriov->num_vfs)
+		return -EINVAL;
+
+	vf_info = &sriov->vf_info[vf];
+	vp = vf_info->vp;
+	if (test_bit(QLC_BC_VF_STATE, &vf_info->state)) {
+		netdev_err(netdev,
+			   "Spoof check change failed for VF %d, as VF driver is loaded. Please unload VF driver and retry the operation\n",
+			   vf);
+		return -EOPNOTSUPP;
+	}
+
+	vp->spoofchk = chk;
+	return 0;
+}
diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sysfs.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sysfs.c
index e7a2fe2..10ed82b 100644
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sysfs.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_sysfs.c
@@ -47,7 +47,7 @@ static ssize_t qlcnic_store_bridged_mode(struct device *dev,
 	if (!test_bit(__QLCNIC_DEV_UP, &adapter->state))
 		goto err_out;
 
-	if (strict_strtoul(buf, 2, &new))
+	if (kstrtoul(buf, 2, &new))
 		goto err_out;
 
 	if (!qlcnic_config_bridged_mode(adapter, !!new))
@@ -77,7 +77,7 @@ static ssize_t qlcnic_store_diag_mode(struct device *dev,
 	struct qlcnic_adapter *adapter = dev_get_drvdata(dev);
 	unsigned long new;
 
-	if (strict_strtoul(buf, 2, &new))
+	if (kstrtoul(buf, 2, &new))
 		return -EINVAL;
 
 	if (!!new != !!(adapter->flags & QLCNIC_DIAG_ENABLED))
@@ -114,57 +114,51 @@ static int qlcnic_validate_beacon(struct qlcnic_adapter *adapter, u16 beacon,
 	return 0;
 }
 
-static ssize_t qlcnic_store_beacon(struct device *dev,
-				   struct device_attribute *attr,
-				   const char *buf, size_t len)
+static int qlcnic_83xx_store_beacon(struct qlcnic_adapter *adapter,
+				    const char *buf, size_t len)
 {
-	struct qlcnic_adapter *adapter = dev_get_drvdata(dev);
 	struct qlcnic_hardware_context *ahw = adapter->ahw;
-	int err, max_sds_rings = adapter->max_sds_rings;
-	u16 beacon;
-	u8 b_state, b_rate;
 	unsigned long h_beacon;
+	int err;
 
-	if (adapter->ahw->op_mode == QLCNIC_NON_PRIV_FUNC) {
-		dev_warn(dev,
-			 "LED test not supported in non privileged mode\n");
-		return -EOPNOTSUPP;
-	}
+	if (test_bit(__QLCNIC_RESETTING, &adapter->state))
+		return -EIO;
 
-	if (qlcnic_83xx_check(adapter) &&
-	    !test_bit(__QLCNIC_RESETTING, &adapter->state)) {
-		if (kstrtoul(buf, 2, &h_beacon))
-			return -EINVAL;
+	if (kstrtoul(buf, 2, &h_beacon))
+		return -EINVAL;
 
-		if (ahw->beacon_state == h_beacon)
-			return len;
+	if (ahw->beacon_state == h_beacon)
+		return len;
 
-		rtnl_lock();
-		if (!ahw->beacon_state) {
-			if (test_and_set_bit(__QLCNIC_LED_ENABLE,
-					     &adapter->state)) {
-				rtnl_unlock();
-				return -EBUSY;
-			}
-		}
-		if (h_beacon) {
-			err = qlcnic_83xx_config_led(adapter, 1, h_beacon);
-			if (err)
-				goto beacon_err;
-		} else {
-			err = qlcnic_83xx_config_led(adapter, 0, !h_beacon);
-			if (err)
-				goto beacon_err;
+	rtnl_lock();
+	if (!ahw->beacon_state) {
+		if (test_and_set_bit(__QLCNIC_LED_ENABLE, &adapter->state)) {
+			rtnl_unlock();
+			return -EBUSY;
 		}
-		/* set the current beacon state */
+	}
+
+	if (h_beacon)
+		err = qlcnic_83xx_config_led(adapter, 1, h_beacon);
+	else
+		err = qlcnic_83xx_config_led(adapter, 0, !h_beacon);
+	if (!err)
 		ahw->beacon_state = h_beacon;
-beacon_err:
-		if (!ahw->beacon_state)
-			clear_bit(__QLCNIC_LED_ENABLE, &adapter->state);
 
-		rtnl_unlock();
-		return len;
-	}
+	if (!ahw->beacon_state)
+		clear_bit(__QLCNIC_LED_ENABLE, &adapter->state);
+
+	rtnl_unlock();
+	return len;
+}
+
+static int qlcnic_82xx_store_beacon(struct qlcnic_adapter *adapter,
+				    const char *buf, size_t len)
+{
+	struct qlcnic_hardware_context *ahw = adapter->ahw;
+	int err, max_sds_rings = adapter->max_sds_rings;
+	u16 beacon;
+	u8 h_beacon_state, b_state, b_rate;
 
 	if (len != sizeof(u16))
 		return QL_STATUS_INVALID_PARAM;
@@ -174,16 +168,29 @@ beacon_err:
 	if (err)
 		return err;
 
-	if (adapter->ahw->beacon_state == b_state)
+	if (ahw->extra_capability[0] & QLCNIC_FW_CAPABILITY_2_BEACON) {
+		err = qlcnic_get_beacon_state(adapter, &h_beacon_state);
+		if (!err) {
+			dev_info(&adapter->pdev->dev,
+				 "Failed to get current beacon state\n");
+		} else {
+			if (h_beacon_state == QLCNIC_BEACON_DISABLE)
+				ahw->beacon_state = 0;
+			else if (h_beacon_state == QLCNIC_BEACON_EANBLE)
+				ahw->beacon_state = 2;
+		}
+	}
+
+	if (ahw->beacon_state == b_state)
 		return len;
 
 	rtnl_lock();
-
-	if (!adapter->ahw->beacon_state)
+	if (!ahw->beacon_state) {
 		if (test_and_set_bit(__QLCNIC_LED_ENABLE, &adapter->state)) {
 			rtnl_unlock();
 			return -EBUSY;
 		}
+	}
 
 	if (test_bit(__QLCNIC_RESETTING, &adapter->state)) {
 		err = -EIO;
@@ -206,14 +213,37 @@ beacon_err:
 	if (test_and_clear_bit(__QLCNIC_DIAG_RES_ALLOC, &adapter->state))
 		qlcnic_diag_free_res(adapter->netdev, max_sds_rings);
 
- out:
-	if (!adapter->ahw->beacon_state)
+out:
+	if (!ahw->beacon_state)
 		clear_bit(__QLCNIC_LED_ENABLE, &adapter->state);
 	rtnl_unlock();
 
 	return err;
 }
 
+static ssize_t qlcnic_store_beacon(struct device *dev,
+				   struct device_attribute *attr,
+				   const char *buf, size_t len)
+{
+	struct qlcnic_adapter *adapter = dev_get_drvdata(dev);
+	int err = 0;
+
+	if (adapter->ahw->op_mode == QLCNIC_NON_PRIV_FUNC) {
+		dev_warn(dev,
+			 "LED test not supported in non privileged mode\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (qlcnic_82xx_check(adapter))
+		err = qlcnic_82xx_store_beacon(adapter, buf, len);
+	else if (qlcnic_83xx_check(adapter))
+		err = qlcnic_83xx_store_beacon(adapter, buf, len);
+	else
+		return -EIO;
+
+	return err;
+}
+
 static ssize_t qlcnic_show_beacon(struct device *dev,
 				  struct device_attribute *attr, char *buf)
 {
diff --git a/drivers/net/ethernet/qlogic/qlge/qlge_main.c b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
index f87cc21..2553cf4 100644
--- a/drivers/net/ethernet/qlogic/qlge/qlge_main.c
+++ b/drivers/net/ethernet/qlogic/qlge/qlge_main.c
@@ -4946,15 +4946,4 @@ static struct pci_driver qlge_driver = {
 	.err_handler = &qlge_err_handler
 };
 
-static int __init qlge_init_module(void)
-{
-	return pci_register_driver(&qlge_driver);
-}
-
-static void __exit qlge_exit(void)
-{
-	pci_unregister_driver(&qlge_driver);
-}
-
-module_init(qlge_init_module);
-module_exit(qlge_exit);
+module_pci_driver(qlge_driver);
diff --git a/drivers/net/ethernet/rdc/Kconfig b/drivers/net/ethernet/rdc/Kconfig
index c8ba4b3..2055f7e 100644
--- a/drivers/net/ethernet/rdc/Kconfig
+++ b/drivers/net/ethernet/rdc/Kconfig
@@ -22,7 +22,6 @@ config R6040
 	tristate "RDC R6040 Fast Ethernet Adapter support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	select PHYLIB
 	---help---
diff --git a/drivers/net/ethernet/realtek/8139cp.c b/drivers/net/ethernet/realtek/8139cp.c
index 0352345..e6acb9f 100644
--- a/drivers/net/ethernet/realtek/8139cp.c
+++ b/drivers/net/ethernet/realtek/8139cp.c
@@ -1817,7 +1817,7 @@ static int cp_set_eeprom(struct net_device *dev,
 /* Put the board into D3cold state and wait for WakeUp signal */
 static void cp_set_d3_state (struct cp_private *cp)
 {
-	pci_enable_wake (cp->pdev, 0, 1); /* Enable PME# generation */
+	pci_enable_wake(cp->pdev, PCI_D0, 1); /* Enable PME# generation */
 	pci_set_power_state (cp->pdev, PCI_D3hot);
 }
 
diff --git a/drivers/net/ethernet/realtek/Kconfig b/drivers/net/ethernet/realtek/Kconfig
index 783fa8b..ae5d027 100644
--- a/drivers/net/ethernet/realtek/Kconfig
+++ b/drivers/net/ethernet/realtek/Kconfig
@@ -37,7 +37,6 @@ config 8139CP
 	tristate "RealTek RTL-8139 C+ PCI Fast Ethernet Adapter support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This is a driver for the Fast Ethernet PCI network cards based on
@@ -52,7 +51,6 @@ config 8139TOO
 	tristate "RealTek RTL-8129/8130/8139 PCI Fast Ethernet Adapter support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This is a driver for the Fast Ethernet PCI network cards based on
@@ -107,7 +105,6 @@ config R8169
 	depends on PCI
 	select FW_LOADER
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here if you have a Realtek 8169 PCI Gigabit Ethernet adapter.
diff --git a/drivers/net/ethernet/renesas/Kconfig b/drivers/net/ethernet/renesas/Kconfig
index bed9841..544514e 100644
--- a/drivers/net/ethernet/renesas/Kconfig
+++ b/drivers/net/ethernet/renesas/Kconfig
@@ -4,14 +4,7 @@
 
 config SH_ETH
 	tristate "Renesas SuperH Ethernet support"
-	depends on (SUPERH || ARCH_SHMOBILE) && \
-		(CPU_SUBTYPE_SH7710 || CPU_SUBTYPE_SH7712 || \
-		 CPU_SUBTYPE_SH7763 || CPU_SUBTYPE_SH7619 || \
-		 CPU_SUBTYPE_SH7724 || CPU_SUBTYPE_SH7734 || \
-		 CPU_SUBTYPE_SH7757 || ARCH_R8A7740 || \
-		 ARCH_R8A7778 || ARCH_R8A7779)
 	select CRC32
-	select NET_CORE
 	select MII
 	select MDIO_BITBANG
 	select PHYLIB
diff --git a/drivers/net/ethernet/renesas/sh_eth.c b/drivers/net/ethernet/renesas/sh_eth.c
index e29fe8d..a753928 100644
--- a/drivers/net/ethernet/renesas/sh_eth.c
+++ b/drivers/net/ethernet/renesas/sh_eth.c
@@ -313,9 +313,14 @@ static const u16 sh_eth_offset_fast_sh3_sh2[SH_ETH_MAX_REGISTER_OFFSET] = {
 	[TSU_ADRL31]	= 0x01fc,
 };
 
-#if defined(CONFIG_CPU_SUBTYPE_SH7734) || \
-	defined(CONFIG_CPU_SUBTYPE_SH7763) || \
-	defined(CONFIG_ARCH_R8A7740)
+static int sh_eth_is_gether(struct sh_eth_private *mdp)
+{
+	if (mdp->reg_offset == sh_eth_offset_gigabit)
+		return 1;
+	else
+		return 0;
+}
+
 static void sh_eth_select_mii(struct net_device *ndev)
 {
 	u32 value = 0x0;
@@ -339,11 +344,7 @@ static void sh_eth_select_mii(struct net_device *ndev)
 
 	sh_eth_write(ndev, value, RMII_MII);
 }
-#endif
 
-/* There is CPU dependent code */
-#if defined(CONFIG_ARCH_R8A7778) || defined(CONFIG_ARCH_R8A7779)
-#define SH_ETH_RESET_DEFAULT	1
 static void sh_eth_set_duplex(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
@@ -354,7 +355,8 @@ static void sh_eth_set_duplex(struct net_device *ndev)
 		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) & ~ECMR_DM, ECMR);
 }
 
-static void sh_eth_set_rate(struct net_device *ndev)
+/* There is CPU dependent code */
+static void sh_eth_set_rate_r8a777x(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 
@@ -371,9 +373,9 @@ static void sh_eth_set_rate(struct net_device *ndev)
 }
 
 /* R8A7778/9 */
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
+static struct sh_eth_cpu_data r8a777x_data = {
 	.set_duplex	= sh_eth_set_duplex,
-	.set_rate	= sh_eth_set_rate,
+	.set_rate	= sh_eth_set_rate_r8a777x,
 
 	.ecsr_value	= ECSR_PSRTO | ECSR_LCHNG | ECSR_ICD,
 	.ecsipr_value	= ECSIPR_PSRTOIP | ECSIPR_LCHNGIP | ECSIPR_ICDIP,
@@ -383,26 +385,14 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.eesr_err_check	= EESR_TWB | EESR_TABT | EESR_RABT | EESR_RFE |
 			  EESR_RDE | EESR_RFRMER | EESR_TFE | EESR_TDE |
 			  EESR_ECI,
-	.tx_error_check	= EESR_TWB | EESR_TABT | EESR_TDE | EESR_TFE,
 
 	.apr		= 1,
 	.mpr		= 1,
 	.tpauser	= 1,
 	.hw_swap	= 1,
 };
-#elif defined(CONFIG_CPU_SUBTYPE_SH7724)
-#define SH_ETH_RESET_DEFAULT	1
-static void sh_eth_set_duplex(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
 
-	if (mdp->duplex) /* Full */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) | ECMR_DM, ECMR);
-	else		/* Half */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) & ~ECMR_DM, ECMR);
-}
-
-static void sh_eth_set_rate(struct net_device *ndev)
+static void sh_eth_set_rate_sh7724(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 
@@ -419,19 +409,18 @@ static void sh_eth_set_rate(struct net_device *ndev)
 }
 
 /* SH7724 */
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
+static struct sh_eth_cpu_data sh7724_data = {
 	.set_duplex	= sh_eth_set_duplex,
-	.set_rate	= sh_eth_set_rate,
+	.set_rate	= sh_eth_set_rate_sh7724,
 
 	.ecsr_value	= ECSR_PSRTO | ECSR_LCHNG | ECSR_ICD,
 	.ecsipr_value	= ECSIPR_PSRTOIP | ECSIPR_LCHNGIP | ECSIPR_ICDIP,
-	.eesipr_value	= DMAC_M_RFRMER | DMAC_M_ECI | 0x01ff009f,
+	.eesipr_value	= 0x01ff009f,
 
 	.tx_check	= EESR_FTC | EESR_CND | EESR_DLC | EESR_CD | EESR_RTO,
 	.eesr_err_check	= EESR_TWB | EESR_TABT | EESR_RABT | EESR_RFE |
 			  EESR_RDE | EESR_RFRMER | EESR_TFE | EESR_TDE |
 			  EESR_ECI,
-	.tx_error_check	= EESR_TWB | EESR_TABT | EESR_TDE | EESR_TFE,
 
 	.apr		= 1,
 	.mpr		= 1,
@@ -440,22 +429,8 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.rpadir		= 1,
 	.rpadir_value	= 0x00020000, /* NET_IP_ALIGN assumed to be 2 */
 };
-#elif defined(CONFIG_CPU_SUBTYPE_SH7757)
-#define SH_ETH_HAS_BOTH_MODULES	1
-#define SH_ETH_HAS_TSU	1
-static int sh_eth_check_reset(struct net_device *ndev);
-
-static void sh_eth_set_duplex(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
 
-	if (mdp->duplex) /* Full */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) | ECMR_DM, ECMR);
-	else		/* Half */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) & ~ECMR_DM, ECMR);
-}
-
-static void sh_eth_set_rate(struct net_device *ndev)
+static void sh_eth_set_rate_sh7757(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 
@@ -472,9 +447,9 @@ static void sh_eth_set_rate(struct net_device *ndev)
 }
 
 /* SH7757 */
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
-	.set_duplex		= sh_eth_set_duplex,
-	.set_rate		= sh_eth_set_rate,
+static struct sh_eth_cpu_data sh7757_data = {
+	.set_duplex	= sh_eth_set_duplex,
+	.set_rate	= sh_eth_set_rate_sh7757,
 
 	.eesipr_value	= DMAC_M_RFRMER | DMAC_M_ECI | 0x003fffff,
 	.rmcr_value	= 0x00000001,
@@ -483,8 +458,8 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.eesr_err_check	= EESR_TWB | EESR_TABT | EESR_RABT | EESR_RFE |
 			  EESR_RDE | EESR_RFRMER | EESR_TFE | EESR_TDE |
 			  EESR_ECI,
-	.tx_error_check	= EESR_TWB | EESR_TABT | EESR_TDE | EESR_TFE,
 
+	.irq_flags	= IRQF_SHARED,
 	.apr		= 1,
 	.mpr		= 1,
 	.tpauser	= 1,
@@ -494,7 +469,7 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.rpadir_value   = 2 << 16,
 };
 
-#define SH_GIGA_ETH_BASE	0xfee00000
+#define SH_GIGA_ETH_BASE	0xfee00000UL
 #define GIGA_MALR(port)		(SH_GIGA_ETH_BASE + 0x800 * (port) + 0x05c8)
 #define GIGA_MAHR(port)		(SH_GIGA_ETH_BASE + 0x800 * (port) + 0x05c0)
 static void sh_eth_chip_reset_giga(struct net_device *ndev)
@@ -519,52 +494,6 @@ static void sh_eth_chip_reset_giga(struct net_device *ndev)
 	}
 }
 
-static int sh_eth_is_gether(struct sh_eth_private *mdp);
-static int sh_eth_reset(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
-	int ret = 0;
-
-	if (sh_eth_is_gether(mdp)) {
-		sh_eth_write(ndev, 0x03, EDSR);
-		sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_GETHER,
-				EDMR);
-
-		ret = sh_eth_check_reset(ndev);
-		if (ret)
-			goto out;
-
-		/* Table Init */
-		sh_eth_write(ndev, 0x0, TDLAR);
-		sh_eth_write(ndev, 0x0, TDFAR);
-		sh_eth_write(ndev, 0x0, TDFXR);
-		sh_eth_write(ndev, 0x0, TDFFR);
-		sh_eth_write(ndev, 0x0, RDLAR);
-		sh_eth_write(ndev, 0x0, RDFAR);
-		sh_eth_write(ndev, 0x0, RDFXR);
-		sh_eth_write(ndev, 0x0, RDFFR);
-	} else {
-		sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_ETHER,
-				EDMR);
-		mdelay(3);
-		sh_eth_write(ndev, sh_eth_read(ndev, EDMR) & ~EDMR_SRST_ETHER,
-				EDMR);
-	}
-
-out:
-	return ret;
-}
-
-static void sh_eth_set_duplex_giga(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
-
-	if (mdp->duplex) /* Full */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) | ECMR_DM, ECMR);
-	else		/* Half */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) & ~ECMR_DM, ECMR);
-}
-
 static void sh_eth_set_rate_giga(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
@@ -585,9 +514,9 @@ static void sh_eth_set_rate_giga(struct net_device *ndev)
 }
 
 /* SH7757(GETHERC) */
-static struct sh_eth_cpu_data sh_eth_my_cpu_data_giga = {
+static struct sh_eth_cpu_data sh7757_data_giga = {
 	.chip_reset	= sh_eth_chip_reset_giga,
-	.set_duplex	= sh_eth_set_duplex_giga,
+	.set_duplex	= sh_eth_set_duplex,
 	.set_rate	= sh_eth_set_rate_giga,
 
 	.ecsr_value	= ECSR_ICD | ECSR_MPD,
@@ -598,11 +527,10 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data_giga = {
 	.eesr_err_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_RABT |
 			  EESR_RFE | EESR_RDE | EESR_RFRMER | EESR_TFE |
 			  EESR_TDE | EESR_ECI,
-	.tx_error_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_TDE | \
-			  EESR_TFE,
 	.fdr_value	= 0x0000072f,
 	.rmcr_value	= 0x00000001,
 
+	.irq_flags	= IRQF_SHARED,
 	.apr		= 1,
 	.mpr		= 1,
 	.tpauser	= 1,
@@ -615,19 +543,6 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data_giga = {
 	.tsu		= 1,
 };
 
-static struct sh_eth_cpu_data *sh_eth_get_cpu_data(struct sh_eth_private *mdp)
-{
-	if (sh_eth_is_gether(mdp))
-		return &sh_eth_my_cpu_data_giga;
-	else
-		return &sh_eth_my_cpu_data;
-}
-
-#elif defined(CONFIG_CPU_SUBTYPE_SH7734) || defined(CONFIG_CPU_SUBTYPE_SH7763)
-#define SH_ETH_HAS_TSU	1
-static int sh_eth_check_reset(struct net_device *ndev);
-static void sh_eth_reset_hw_crc(struct net_device *ndev);
-
 static void sh_eth_chip_reset(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
@@ -637,17 +552,7 @@ static void sh_eth_chip_reset(struct net_device *ndev)
 	mdelay(1);
 }
 
-static void sh_eth_set_duplex(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
-
-	if (mdp->duplex) /* Full */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) | ECMR_DM, ECMR);
-	else		/* Half */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) & ~ECMR_DM, ECMR);
-}
-
-static void sh_eth_set_rate(struct net_device *ndev)
+static void sh_eth_set_rate_gether(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 
@@ -666,11 +571,11 @@ static void sh_eth_set_rate(struct net_device *ndev)
 	}
 }
 
-/* sh7763 */
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
+/* SH7734 */
+static struct sh_eth_cpu_data sh7734_data = {
 	.chip_reset	= sh_eth_chip_reset,
 	.set_duplex	= sh_eth_set_duplex,
-	.set_rate	= sh_eth_set_rate,
+	.set_rate	= sh_eth_set_rate_gether,
 
 	.ecsr_value	= ECSR_ICD | ECSR_MPD,
 	.ecsipr_value	= ECSIPR_LCHNGIP | ECSIPR_ICDIP | ECSIPR_MPDIP,
@@ -680,8 +585,6 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.eesr_err_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_RABT |
 			  EESR_RFE | EESR_RDE | EESR_RFRMER | EESR_TFE |
 			  EESR_TDE | EESR_ECI,
-	.tx_error_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_TDE | \
-			  EESR_TFE,
 
 	.apr		= 1,
 	.mpr		= 1,
@@ -691,54 +594,37 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.no_trimd	= 1,
 	.no_ade		= 1,
 	.tsu		= 1,
-#if defined(CONFIG_CPU_SUBTYPE_SH7734)
-	.hw_crc     = 1,
-	.select_mii = 1,
-#endif
+	.hw_crc		= 1,
+	.select_mii	= 1,
 };
 
-static int sh_eth_reset(struct net_device *ndev)
-{
-	int ret = 0;
-
-	sh_eth_write(ndev, EDSR_ENALL, EDSR);
-	sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_GETHER, EDMR);
-
-	ret = sh_eth_check_reset(ndev);
-	if (ret)
-		goto out;
+/* SH7763 */
+static struct sh_eth_cpu_data sh7763_data = {
+	.chip_reset	= sh_eth_chip_reset,
+	.set_duplex	= sh_eth_set_duplex,
+	.set_rate	= sh_eth_set_rate_gether,
 
-	/* Table Init */
-	sh_eth_write(ndev, 0x0, TDLAR);
-	sh_eth_write(ndev, 0x0, TDFAR);
-	sh_eth_write(ndev, 0x0, TDFXR);
-	sh_eth_write(ndev, 0x0, TDFFR);
-	sh_eth_write(ndev, 0x0, RDLAR);
-	sh_eth_write(ndev, 0x0, RDFAR);
-	sh_eth_write(ndev, 0x0, RDFXR);
-	sh_eth_write(ndev, 0x0, RDFFR);
-
-	/* Reset HW CRC register */
-	sh_eth_reset_hw_crc(ndev);
-
-	/* Select MII mode */
-	if (sh_eth_my_cpu_data.select_mii)
-		sh_eth_select_mii(ndev);
-out:
-	return ret;
-}
+	.ecsr_value	= ECSR_ICD | ECSR_MPD,
+	.ecsipr_value	= ECSIPR_LCHNGIP | ECSIPR_ICDIP | ECSIPR_MPDIP,
+	.eesipr_value	= DMAC_M_RFRMER | DMAC_M_ECI | 0x003fffff,
 
-static void sh_eth_reset_hw_crc(struct net_device *ndev)
-{
-	if (sh_eth_my_cpu_data.hw_crc)
-		sh_eth_write(ndev, 0x0, CSMR);
-}
+	.tx_check	= EESR_TC1 | EESR_FTC,
+	.eesr_err_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_RABT | \
+			  EESR_RDE | EESR_RFRMER | EESR_TFE | EESR_TDE | \
+			  EESR_ECI,
 
-#elif defined(CONFIG_ARCH_R8A7740)
-#define SH_ETH_HAS_TSU	1
-static int sh_eth_check_reset(struct net_device *ndev);
+	.apr		= 1,
+	.mpr		= 1,
+	.tpauser	= 1,
+	.bculr		= 1,
+	.hw_swap	= 1,
+	.no_trimd	= 1,
+	.no_ade		= 1,
+	.tsu		= 1,
+	.irq_flags	= IRQF_SHARED,
+};
 
-static void sh_eth_chip_reset(struct net_device *ndev)
+static void sh_eth_chip_reset_r8a7740(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 
@@ -749,65 +635,11 @@ static void sh_eth_chip_reset(struct net_device *ndev)
 	sh_eth_select_mii(ndev);
 }
 
-static int sh_eth_reset(struct net_device *ndev)
-{
-	int ret = 0;
-
-	sh_eth_write(ndev, EDSR_ENALL, EDSR);
-	sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_GETHER, EDMR);
-
-	ret = sh_eth_check_reset(ndev);
-	if (ret)
-		goto out;
-
-	/* Table Init */
-	sh_eth_write(ndev, 0x0, TDLAR);
-	sh_eth_write(ndev, 0x0, TDFAR);
-	sh_eth_write(ndev, 0x0, TDFXR);
-	sh_eth_write(ndev, 0x0, TDFFR);
-	sh_eth_write(ndev, 0x0, RDLAR);
-	sh_eth_write(ndev, 0x0, RDFAR);
-	sh_eth_write(ndev, 0x0, RDFXR);
-	sh_eth_write(ndev, 0x0, RDFFR);
-
-out:
-	return ret;
-}
-
-static void sh_eth_set_duplex(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
-
-	if (mdp->duplex) /* Full */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) | ECMR_DM, ECMR);
-	else		/* Half */
-		sh_eth_write(ndev, sh_eth_read(ndev, ECMR) & ~ECMR_DM, ECMR);
-}
-
-static void sh_eth_set_rate(struct net_device *ndev)
-{
-	struct sh_eth_private *mdp = netdev_priv(ndev);
-
-	switch (mdp->speed) {
-	case 10: /* 10BASE */
-		sh_eth_write(ndev, GECMR_10, GECMR);
-		break;
-	case 100:/* 100BASE */
-		sh_eth_write(ndev, GECMR_100, GECMR);
-		break;
-	case 1000: /* 1000BASE */
-		sh_eth_write(ndev, GECMR_1000, GECMR);
-		break;
-	default:
-		break;
-	}
-}
-
 /* R8A7740 */
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
-	.chip_reset	= sh_eth_chip_reset,
+static struct sh_eth_cpu_data r8a7740_data = {
+	.chip_reset	= sh_eth_chip_reset_r8a7740,
 	.set_duplex	= sh_eth_set_duplex,
-	.set_rate	= sh_eth_set_rate,
+	.set_rate	= sh_eth_set_rate_gether,
 
 	.ecsr_value	= ECSR_ICD | ECSR_MPD,
 	.ecsipr_value	= ECSIPR_LCHNGIP | ECSIPR_ICDIP | ECSIPR_MPDIP,
@@ -817,8 +649,6 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.eesr_err_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_RABT |
 			  EESR_RFE | EESR_RDE | EESR_RFRMER | EESR_TFE |
 			  EESR_TDE | EESR_ECI,
-	.tx_error_check	= EESR_TWB1 | EESR_TWB | EESR_TABT | EESR_TDE | \
-			  EESR_TFE,
 
 	.apr		= 1,
 	.mpr		= 1,
@@ -829,11 +659,10 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.no_ade		= 1,
 	.tsu		= 1,
 	.select_mii	= 1,
+	.shift_rd0	= 1,
 };
 
-#elif defined(CONFIG_CPU_SUBTYPE_SH7619)
-#define SH_ETH_RESET_DEFAULT	1
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
+static struct sh_eth_cpu_data sh7619_data = {
 	.eesipr_value	= DMAC_M_RFRMER | DMAC_M_ECI | 0x003fffff,
 
 	.apr		= 1,
@@ -841,14 +670,11 @@ static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
 	.tpauser	= 1,
 	.hw_swap	= 1,
 };
-#elif defined(CONFIG_CPU_SUBTYPE_SH7710) || defined(CONFIG_CPU_SUBTYPE_SH7712)
-#define SH_ETH_RESET_DEFAULT	1
-#define SH_ETH_HAS_TSU	1
-static struct sh_eth_cpu_data sh_eth_my_cpu_data = {
+
+static struct sh_eth_cpu_data sh771x_data = {
 	.eesipr_value	= DMAC_M_RFRMER | DMAC_M_ECI | 0x003fffff,
 	.tsu		= 1,
 };
-#endif
 
 static void sh_eth_set_default_cpu_data(struct sh_eth_cpu_data *cd)
 {
@@ -873,22 +699,8 @@ static void sh_eth_set_default_cpu_data(struct sh_eth_cpu_data *cd)
 
 	if (!cd->eesr_err_check)
 		cd->eesr_err_check = DEFAULT_EESR_ERR_CHECK;
-
-	if (!cd->tx_error_check)
-		cd->tx_error_check = DEFAULT_TX_ERROR_CHECK;
 }
 
-#if defined(SH_ETH_RESET_DEFAULT)
-/* Chip Reset */
-static int  sh_eth_reset(struct net_device *ndev)
-{
-	sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_ETHER, EDMR);
-	mdelay(3);
-	sh_eth_write(ndev, sh_eth_read(ndev, EDMR) & ~EDMR_SRST_ETHER, EDMR);
-
-	return 0;
-}
-#else
 static int sh_eth_check_reset(struct net_device *ndev)
 {
 	int ret = 0;
@@ -906,7 +718,49 @@ static int sh_eth_check_reset(struct net_device *ndev)
 	}
 	return ret;
 }
-#endif
+
+static int sh_eth_reset(struct net_device *ndev)
+{
+	struct sh_eth_private *mdp = netdev_priv(ndev);
+	int ret = 0;
+
+	if (sh_eth_is_gether(mdp)) {
+		sh_eth_write(ndev, EDSR_ENALL, EDSR);
+		sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_GETHER,
+			     EDMR);
+
+		ret = sh_eth_check_reset(ndev);
+		if (ret)
+			goto out;
+
+		/* Table Init */
+		sh_eth_write(ndev, 0x0, TDLAR);
+		sh_eth_write(ndev, 0x0, TDFAR);
+		sh_eth_write(ndev, 0x0, TDFXR);
+		sh_eth_write(ndev, 0x0, TDFFR);
+		sh_eth_write(ndev, 0x0, RDLAR);
+		sh_eth_write(ndev, 0x0, RDFAR);
+		sh_eth_write(ndev, 0x0, RDFXR);
+		sh_eth_write(ndev, 0x0, RDFFR);
+
+		/* Reset HW CRC register */
+		if (mdp->cd->hw_crc)
+			sh_eth_write(ndev, 0x0, CSMR);
+
+		/* Select MII mode */
+		if (mdp->cd->select_mii)
+			sh_eth_select_mii(ndev);
+	} else {
+		sh_eth_write(ndev, sh_eth_read(ndev, EDMR) | EDMR_SRST_ETHER,
+			     EDMR);
+		mdelay(3);
+		sh_eth_write(ndev, sh_eth_read(ndev, EDMR) & ~EDMR_SRST_ETHER,
+			     EDMR);
+	}
+
+out:
+	return ret;
+}
 
 #if defined(CONFIG_CPU_SH4) || defined(CONFIG_ARCH_SHMOBILE)
 static void sh_eth_set_receive_align(struct sk_buff *skb)
@@ -982,14 +836,6 @@ static void read_mac_address(struct net_device *ndev, unsigned char *mac)
 	}
 }
 
-static int sh_eth_is_gether(struct sh_eth_private *mdp)
-{
-	if (mdp->reg_offset == sh_eth_offset_gigabit)
-		return 1;
-	else
-		return 0;
-}
-
 static unsigned long sh_eth_get_edtrr_trns(struct sh_eth_private *mdp)
 {
 	if (sh_eth_is_gether(mdp))
@@ -1388,7 +1234,7 @@ static int sh_eth_txfree(struct net_device *ndev)
 }
 
 /* Packet receive function */
-static int sh_eth_rx(struct net_device *ndev, u32 intr_status)
+static int sh_eth_rx(struct net_device *ndev, u32 intr_status, int *quota)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 	struct sh_eth_rxdesc *rxdesc;
@@ -1396,6 +1242,7 @@ static int sh_eth_rx(struct net_device *ndev, u32 intr_status)
 	int entry = mdp->cur_rx % mdp->num_rx_ring;
 	int boguscnt = (mdp->dirty_rx + mdp->num_rx_ring) - mdp->cur_rx;
 	struct sk_buff *skb;
+	int exceeded = 0;
 	u16 pkt_len = 0;
 	u32 desc_status;
 
@@ -1407,10 +1254,15 @@ static int sh_eth_rx(struct net_device *ndev, u32 intr_status)
 		if (--boguscnt < 0)
 			break;
 
+		if (*quota <= 0) {
+			exceeded = 1;
+			break;
+		}
+		(*quota)--;
+
 		if (!(desc_status & RDFEND))
 			ndev->stats.rx_length_errors++;
 
-#if defined(CONFIG_ARCH_R8A7740)
 		/*
 		 * In case of almost all GETHER/ETHERs, the Receive Frame State
 		 * (RFS) bits in the Receive Descriptor 0 are from bit 9 to
@@ -1418,8 +1270,8 @@ static int sh_eth_rx(struct net_device *ndev, u32 intr_status)
 		 * bits are from bit 25 to bit 16. So, the driver needs right
 		 * shifting by 16.
 		 */
-		desc_status >>= 16;
-#endif
+		if (mdp->cd->shift_rd0)
+			desc_status >>= 16;
 
 		if (desc_status & (RD_RFS1 | RD_RFS2 | RD_RFS3 | RD_RFS4 |
 				   RD_RFS5 | RD_RFS6 | RD_RFS10)) {
@@ -1494,7 +1346,7 @@ static int sh_eth_rx(struct net_device *ndev, u32 intr_status)
 		sh_eth_write(ndev, EDRRR_R, EDRRR);
 	}
 
-	return 0;
+	return exceeded;
 }
 
 static void sh_eth_rcv_snd_disable(struct net_device *ndev)
@@ -1636,7 +1488,7 @@ static irqreturn_t sh_eth_interrupt(int irq, void *netdev)
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 	struct sh_eth_cpu_data *cd = mdp->cd;
 	irqreturn_t ret = IRQ_NONE;
-	unsigned long intr_status;
+	unsigned long intr_status, intr_enable;
 
 	spin_lock(&mdp->lock);
 
@@ -1647,34 +1499,41 @@ static irqreturn_t sh_eth_interrupt(int irq, void *netdev)
 	 * and we need to fully handle it in sh_eth_error() in order to quench
 	 * it as it doesn't get cleared by just writing 1 to the ECI bit...
 	 */
-	intr_status &= sh_eth_read(ndev, EESIPR) | DMAC_M_ECI;
-	/* Clear interrupt */
-	if (intr_status & (EESR_FRC | EESR_RMAF | EESR_RRF |
-			EESR_RTLF | EESR_RTSF | EESR_PRE | EESR_CERF |
-			cd->tx_check | cd->eesr_err_check)) {
-		sh_eth_write(ndev, intr_status, EESR);
+	intr_enable = sh_eth_read(ndev, EESIPR);
+	intr_status &= intr_enable | DMAC_M_ECI;
+	if (intr_status & (EESR_RX_CHECK | cd->tx_check | cd->eesr_err_check))
 		ret = IRQ_HANDLED;
-	} else
+	else
 		goto other_irq;
 
-	if (intr_status & (EESR_FRC | /* Frame recv*/
-			EESR_RMAF | /* Multi cast address recv*/
-			EESR_RRF  | /* Bit frame recv */
-			EESR_RTLF | /* Long frame recv*/
-			EESR_RTSF | /* short frame recv */
-			EESR_PRE  | /* PHY-LSI recv error */
-			EESR_CERF)){ /* recv frame CRC error */
-		sh_eth_rx(ndev, intr_status);
+	if (intr_status & EESR_RX_CHECK) {
+		if (napi_schedule_prep(&mdp->napi)) {
+			/* Mask Rx interrupts */
+			sh_eth_write(ndev, intr_enable & ~EESR_RX_CHECK,
+				     EESIPR);
+			__napi_schedule(&mdp->napi);
+		} else {
+			dev_warn(&ndev->dev,
+				 "ignoring interrupt, status 0x%08lx, mask 0x%08lx.\n",
+				 intr_status, intr_enable);
+		}
 	}
 
 	/* Tx Check */
 	if (intr_status & cd->tx_check) {
+		/* Clear Tx interrupts */
+		sh_eth_write(ndev, intr_status & cd->tx_check, EESR);
+
 		sh_eth_txfree(ndev);
 		netif_wake_queue(ndev);
 	}
 
-	if (intr_status & cd->eesr_err_check)
+	if (intr_status & cd->eesr_err_check) {
+		/* Clear error interrupts */
+		sh_eth_write(ndev, intr_status & cd->eesr_err_check, EESR);
+
 		sh_eth_error(ndev, intr_status);
+	}
 
 other_irq:
 	spin_unlock(&mdp->lock);
@@ -1682,6 +1541,33 @@ other_irq:
 	return ret;
 }
 
+static int sh_eth_poll(struct napi_struct *napi, int budget)
+{
+	struct sh_eth_private *mdp = container_of(napi, struct sh_eth_private,
+						  napi);
+	struct net_device *ndev = napi->dev;
+	int quota = budget;
+	unsigned long intr_status;
+
+	for (;;) {
+		intr_status = sh_eth_read(ndev, EESR);
+		if (!(intr_status & EESR_RX_CHECK))
+			break;
+		/* Clear Rx interrupts */
+		sh_eth_write(ndev, intr_status & EESR_RX_CHECK, EESR);
+
+		if (sh_eth_rx(ndev, intr_status, &quota))
+			goto out;
+	}
+
+	napi_complete(napi);
+
+	/* Reenable Rx interrupts */
+	sh_eth_write(ndev, mdp->cd->eesipr_value, EESIPR);
+out:
+	return budget - quota;
+}
+
 /* PHY state control function */
 static void sh_eth_adjust_link(struct net_device *ndev)
 {
@@ -1972,14 +1858,7 @@ static int sh_eth_open(struct net_device *ndev)
 	pm_runtime_get_sync(&mdp->pdev->dev);
 
 	ret = request_irq(ndev->irq, sh_eth_interrupt,
-#if defined(CONFIG_CPU_SUBTYPE_SH7763) || \
-	defined(CONFIG_CPU_SUBTYPE_SH7764) || \
-	defined(CONFIG_CPU_SUBTYPE_SH7757)
-				IRQF_SHARED,
-#else
-				0,
-#endif
-				ndev->name, ndev);
+			  mdp->cd->irq_flags, ndev->name, ndev);
 	if (ret) {
 		dev_err(&ndev->dev, "Can not assign IRQ number\n");
 		return ret;
@@ -2000,6 +1879,8 @@ static int sh_eth_open(struct net_device *ndev)
 	if (ret)
 		goto out_free_irq;
 
+	napi_enable(&mdp->napi);
+
 	return ret;
 
 out_free_irq:
@@ -2095,6 +1976,8 @@ static int sh_eth_close(struct net_device *ndev)
 {
 	struct sh_eth_private *mdp = netdev_priv(ndev);
 
+	napi_disable(&mdp->napi);
+
 	netif_stop_queue(ndev);
 
 	/* Disable interrupts by clearing the interrupt mask. */
@@ -2165,7 +2048,6 @@ static int sh_eth_do_ioctl(struct net_device *ndev, struct ifreq *rq,
 	return phy_mii_ioctl(phydev, rq, cmd);
 }
 
-#if defined(SH_ETH_HAS_TSU)
 /* For TSU_POSTn. Please refer to the manual about this (strange) bitfields */
 static void *sh_eth_tsu_get_post_reg_offset(struct sh_eth_private *mdp,
 					    int entry)
@@ -2508,7 +2390,6 @@ static int sh_eth_vlan_rx_kill_vid(struct net_device *ndev,
 
 	return 0;
 }
-#endif /* SH_ETH_HAS_TSU */
 
 /* SuperH's TSU register init function */
 static void sh_eth_tsu_init(struct sh_eth_private *mdp)
@@ -2652,11 +2533,21 @@ static const struct net_device_ops sh_eth_netdev_ops = {
 	.ndo_stop		= sh_eth_close,
 	.ndo_start_xmit		= sh_eth_start_xmit,
 	.ndo_get_stats		= sh_eth_get_stats,
-#if defined(SH_ETH_HAS_TSU)
+	.ndo_tx_timeout		= sh_eth_tx_timeout,
+	.ndo_do_ioctl		= sh_eth_do_ioctl,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_set_mac_address	= eth_mac_addr,
+	.ndo_change_mtu		= eth_change_mtu,
+};
+
+static const struct net_device_ops sh_eth_netdev_ops_tsu = {
+	.ndo_open		= sh_eth_open,
+	.ndo_stop		= sh_eth_close,
+	.ndo_start_xmit		= sh_eth_start_xmit,
+	.ndo_get_stats		= sh_eth_get_stats,
 	.ndo_set_rx_mode	= sh_eth_set_multicast_list,
 	.ndo_vlan_rx_add_vid	= sh_eth_vlan_rx_add_vid,
 	.ndo_vlan_rx_kill_vid	= sh_eth_vlan_rx_kill_vid,
-#endif
 	.ndo_tx_timeout		= sh_eth_tx_timeout,
 	.ndo_do_ioctl		= sh_eth_do_ioctl,
 	.ndo_validate_addr	= eth_validate_addr,
@@ -2671,6 +2562,7 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
 	struct net_device *ndev = NULL;
 	struct sh_eth_private *mdp = NULL;
 	struct sh_eth_plat_data *pd = pdev->dev.platform_data;
+	const struct platform_device_id *id = platform_get_device_id(pdev);
 
 	/* get base addr */
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
@@ -2729,15 +2621,14 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
 	mdp->reg_offset = sh_eth_get_register_offset(pd->register_type);
 
 	/* set cpu data */
-#if defined(SH_ETH_HAS_BOTH_MODULES)
-	mdp->cd = sh_eth_get_cpu_data(mdp);
-#else
-	mdp->cd = &sh_eth_my_cpu_data;
-#endif
+	mdp->cd = (struct sh_eth_cpu_data *)id->driver_data;
 	sh_eth_set_default_cpu_data(mdp->cd);
 
 	/* set function */
-	ndev->netdev_ops = &sh_eth_netdev_ops;
+	if (mdp->cd->tsu)
+		ndev->netdev_ops = &sh_eth_netdev_ops_tsu;
+	else
+		ndev->netdev_ops = &sh_eth_netdev_ops;
 	SET_ETHTOOL_OPS(ndev, &sh_eth_ethtool_ops);
 	ndev->watchdog_timeo = TX_TIMEOUT;
 
@@ -2776,10 +2667,12 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
 		}
 	}
 
+	netif_napi_add(ndev, &mdp->napi, sh_eth_poll, 64);
+
 	/* network device register */
 	ret = register_netdev(ndev);
 	if (ret)
-		goto out_release;
+		goto out_napi_del;
 
 	/* mdio bus init */
 	ret = sh_mdio_init(ndev, pdev->id, pd);
@@ -2797,6 +2690,9 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
 out_unregister:
 	unregister_netdev(ndev);
 
+out_napi_del:
+	netif_napi_del(&mdp->napi);
+
 out_release:
 	/* net_dev free */
 	if (ndev)
@@ -2809,16 +2705,18 @@ out:
 static int sh_eth_drv_remove(struct platform_device *pdev)
 {
 	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct sh_eth_private *mdp = netdev_priv(ndev);
 
 	sh_mdio_release(ndev);
 	unregister_netdev(ndev);
+	netif_napi_del(&mdp->napi);
 	pm_runtime_disable(&pdev->dev);
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
 
+#ifdef CONFIG_PM
 static int sh_eth_runtime_nop(struct device *dev)
 {
 	/*
@@ -2832,17 +2730,36 @@ static int sh_eth_runtime_nop(struct device *dev)
 	return 0;
 }
 
-static struct dev_pm_ops sh_eth_dev_pm_ops = {
+static const struct dev_pm_ops sh_eth_dev_pm_ops = {
 	.runtime_suspend = sh_eth_runtime_nop,
 	.runtime_resume = sh_eth_runtime_nop,
 };
+#define SH_ETH_PM_OPS (&sh_eth_dev_pm_ops)
+#else
+#define SH_ETH_PM_OPS NULL
+#endif
+
+static struct platform_device_id sh_eth_id_table[] = {
+	{ "sh7619-ether", (kernel_ulong_t)&sh7619_data },
+	{ "sh771x-ether", (kernel_ulong_t)&sh771x_data },
+	{ "sh7724-ether", (kernel_ulong_t)&sh7724_data },
+	{ "sh7734-gether", (kernel_ulong_t)&sh7734_data },
+	{ "sh7757-ether", (kernel_ulong_t)&sh7757_data },
+	{ "sh7757-gether", (kernel_ulong_t)&sh7757_data_giga },
+	{ "sh7763-gether", (kernel_ulong_t)&sh7763_data },
+	{ "r8a7740-gether", (kernel_ulong_t)&r8a7740_data },
+	{ "r8a777x-ether", (kernel_ulong_t)&r8a777x_data },
+	{ }
+};
+MODULE_DEVICE_TABLE(platform, sh_eth_id_table);
 
 static struct platform_driver sh_eth_driver = {
 	.probe = sh_eth_drv_probe,
 	.remove = sh_eth_drv_remove,
+	.id_table = sh_eth_id_table,
 	.driver = {
 		   .name = CARDNAME,
-		   .pm = &sh_eth_dev_pm_ops,
+		   .pm = SH_ETH_PM_OPS,
 	},
 };
 
diff --git a/drivers/net/ethernet/renesas/sh_eth.h b/drivers/net/ethernet/renesas/sh_eth.h
index 62689a5..99995bf 100644
--- a/drivers/net/ethernet/renesas/sh_eth.h
+++ b/drivers/net/ethernet/renesas/sh_eth.h
@@ -166,19 +166,16 @@ enum {
 /*
  * Register's bits
  */
-#if defined(CONFIG_CPU_SUBTYPE_SH7734) || defined(CONFIG_CPU_SUBTYPE_SH7763) ||\
-    defined(CONFIG_ARCH_R8A7740)
-/* EDSR */
+/* EDSR : sh7734, sh7757, sh7763, and r8a7740 only */
 enum EDSR_BIT {
 	EDSR_ENT = 0x01, EDSR_ENR = 0x02,
 };
 #define EDSR_ENALL (EDSR_ENT|EDSR_ENR)
 
-/* GECMR */
+/* GECMR : sh7734, sh7763 and r8a7740 only */
 enum GECMR_BIT {
 	GECMR_10 = 0x0, GECMR_100 = 0x04, GECMR_1000 = 0x01,
 };
-#endif
 
 /* EDMR */
 enum DMAC_M_BIT {
@@ -251,13 +248,19 @@ enum EESR_BIT {
 	EESR_CERF	= 0x00000001,
 };
 
+#define EESR_RX_CHECK		(EESR_FRC  | /* Frame recv */		\
+				 EESR_RMAF | /* Multicast address recv */ \
+				 EESR_RRF  | /* Bit frame recv */	\
+				 EESR_RTLF | /* Long frame recv */	\
+				 EESR_RTSF | /* Short frame recv */	\
+				 EESR_PRE  | /* PHY-LSI recv error */	\
+				 EESR_CERF)  /* Recv frame CRC error */
+
 #define DEFAULT_TX_CHECK	(EESR_FTC | EESR_CND | EESR_DLC | EESR_CD | \
 				 EESR_RTO)
 #define DEFAULT_EESR_ERR_CHECK	(EESR_TWB | EESR_TABT | EESR_RABT | EESR_RFE | \
 				 EESR_RDE | EESR_RFRMER | EESR_ADE | \
 				 EESR_TFE | EESR_TDE | EESR_ECI)
-#define DEFAULT_TX_ERROR_CHECK	(EESR_TWB | EESR_TABT | EESR_ADE | EESR_TDE | \
-				 EESR_TFE)
 
 /* EESIPR */
 enum DMAC_IM_BIT {
@@ -299,11 +302,11 @@ enum FCFTR_BIT {
 #define DEFAULT_FIFO_F_D_RFF	(FCFTR_RFF2 | FCFTR_RFF1 | FCFTR_RFF0)
 #define DEFAULT_FIFO_F_D_RFD	(FCFTR_RFD2 | FCFTR_RFD1 | FCFTR_RFD0)
 
-/* Transfer descriptor bit */
+/* Transmit descriptor bit */
 enum TD_STS_BIT {
-	TD_TACT = 0x80000000,
-	TD_TDLE = 0x40000000, TD_TFP1 = 0x20000000,
-	TD_TFP0 = 0x10000000,
+	TD_TACT = 0x80000000, TD_TDLE = 0x40000000,
+	TD_TFP1 = 0x20000000, TD_TFP0 = 0x10000000,
+	TD_TFE  = 0x08000000, TD_TWBI = 0x04000000,
 };
 #define TDF1ST	TD_TFP1
 #define TDFEND	TD_TFP0
@@ -463,9 +466,9 @@ struct sh_eth_cpu_data {
 	/* interrupt checking mask */
 	unsigned long tx_check;
 	unsigned long eesr_err_check;
-	unsigned long tx_error_check;
 
 	/* hardware features */
+	unsigned long irq_flags;	/* IRQ configuration flags */
 	unsigned no_psr:1;		/* EtherC DO NOT have PSR */
 	unsigned apr:1;			/* EtherC have APR */
 	unsigned mpr:1;			/* EtherC have MPR */
@@ -478,6 +481,7 @@ struct sh_eth_cpu_data {
 	unsigned no_ade:1;	/* E-DMAC DO NOT have ADE bit in EESR */
 	unsigned hw_crc:1;	/* E-DMAC have CSMR */
 	unsigned select_mii:1;	/* EtherC have RMII_MII (MII select register) */
+	unsigned shift_rd0:1;	/* shift Rx descriptor word 0 right by 16 */
 };
 
 struct sh_eth_private {
@@ -499,6 +503,7 @@ struct sh_eth_private {
 	u32 cur_tx, dirty_tx;
 	u32 rx_buf_sz;		/* Based on MTU+slack. */
 	int edmac_endian;
+	struct napi_struct napi;
 	/* MII transceiver section. */
 	u32 phy_id;					/* PHY ID */
 	struct mii_bus *mii_bus;	/* MDIO bus control */
diff --git a/drivers/net/ethernet/s6gmac.c b/drivers/net/ethernet/s6gmac.c
index b6739af..a99739c 100644
--- a/drivers/net/ethernet/s6gmac.c
+++ b/drivers/net/ethernet/s6gmac.c
@@ -1040,7 +1040,6 @@ static int s6gmac_remove(struct platform_device *pdev)
 		unregister_netdev(dev);
 		free_irq(dev->irq, dev);
 		free_netdev(dev);
-		platform_set_drvdata(pdev, NULL);
 	}
 	return 0;
 }
diff --git a/drivers/net/ethernet/seeq/sgiseeq.c b/drivers/net/ethernet/seeq/sgiseeq.c
index 0ad5694..856e523 100644
--- a/drivers/net/ethernet/seeq/sgiseeq.c
+++ b/drivers/net/ethernet/seeq/sgiseeq.c
@@ -818,7 +818,6 @@ static int __exit sgiseeq_remove(struct platform_device *pdev)
 	dma_free_noncoherent(&pdev->dev, sizeof(*sp->srings), sp->srings,
 			     sp->srings_dma);
 	free_netdev(dev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index 4a14a94..c729688 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -21,8 +21,8 @@
 #include <linux/ethtool.h>
 #include <linux/topology.h>
 #include <linux/gfp.h>
-#include <linux/cpu_rmap.h>
 #include <linux/aer.h>
+#include <linux/interrupt.h>
 #include "net_driver.h"
 #include "efx.h"
 #include "nic.h"
@@ -1283,29 +1283,6 @@ static unsigned int efx_wanted_parallelism(struct efx_nic *efx)
 	return count;
 }
 
-static int
-efx_init_rx_cpu_rmap(struct efx_nic *efx, struct msix_entry *xentries)
-{
-#ifdef CONFIG_RFS_ACCEL
-	unsigned int i;
-	int rc;
-
-	efx->net_dev->rx_cpu_rmap = alloc_irq_cpu_rmap(efx->n_rx_channels);
-	if (!efx->net_dev->rx_cpu_rmap)
-		return -ENOMEM;
-	for (i = 0; i < efx->n_rx_channels; i++) {
-		rc = irq_cpu_rmap_add(efx->net_dev->rx_cpu_rmap,
-				      xentries[i].vector);
-		if (rc) {
-			free_irq_cpu_rmap(efx->net_dev->rx_cpu_rmap);
-			efx->net_dev->rx_cpu_rmap = NULL;
-			return rc;
-		}
-	}
-#endif
-	return 0;
-}
-
 /* Probe the number and type of interrupts we are able to obtain, and
  * the resulting numbers of channels and RX queues.
  */
@@ -1359,11 +1336,6 @@ static int efx_probe_interrupts(struct efx_nic *efx)
 				efx->n_tx_channels = n_channels;
 				efx->n_rx_channels = n_channels;
 			}
-			rc = efx_init_rx_cpu_rmap(efx, xentries);
-			if (rc) {
-				pci_disable_msix(efx->pci_dev);
-				return rc;
-			}
 			for (i = 0; i < efx->n_channels; i++)
 				efx_get_channel(efx, i)->irq =
 					xentries[i].vector;
@@ -1427,6 +1399,10 @@ static void efx_start_interrupts(struct efx_nic *efx, bool may_keep_eventq)
 
 	BUG_ON(efx->state == STATE_DISABLED);
 
+	if (efx->eeh_disabled_legacy_irq) {
+		enable_irq(efx->legacy_irq);
+		efx->eeh_disabled_legacy_irq = false;
+	}
 	if (efx->legacy_irq)
 		efx->legacy_irq_enabled = true;
 	efx_nic_enable_interrupts(efx);
@@ -2120,7 +2096,7 @@ static void efx_update_name(struct efx_nic *efx)
 static int efx_netdev_event(struct notifier_block *this,
 			    unsigned long event, void *ptr)
 {
-	struct net_device *net_dev = ptr;
+	struct net_device *net_dev = netdev_notifier_info_to_dev(ptr);
 
 	if (net_dev->netdev_ops == &efx_netdev_ops &&
 	    event == NETDEV_CHANGENAME)
@@ -2365,7 +2341,7 @@ out:
  * Returns 0 if the recovery mechanisms are unsuccessful.
  * Returns a non-zero value otherwise.
  */
-static int efx_try_recovery(struct efx_nic *efx)
+int efx_try_recovery(struct efx_nic *efx)
 {
 #ifdef CONFIG_EEH
 	/* A PCI error can occur and not be seen by EEH because nothing
@@ -2603,10 +2579,6 @@ static void efx_pci_remove_main(struct efx_nic *efx)
 	BUG_ON(efx->state == STATE_READY);
 	cancel_work_sync(&efx->reset_work);
 
-#ifdef CONFIG_RFS_ACCEL
-	free_irq_cpu_rmap(efx->net_dev->rx_cpu_rmap);
-	efx->net_dev->rx_cpu_rmap = NULL;
-#endif
 	efx_stop_interrupts(efx, false);
 	efx_nic_fini_interrupt(efx);
 	efx_fini_port(efx);
diff --git a/drivers/net/ethernet/sfc/efx.h b/drivers/net/ethernet/sfc/efx.h
index 8372da2..bdb30bb 100644
--- a/drivers/net/ethernet/sfc/efx.h
+++ b/drivers/net/ethernet/sfc/efx.h
@@ -124,6 +124,7 @@ extern const struct ethtool_ops efx_ethtool_ops;
 extern int efx_reset(struct efx_nic *efx, enum reset_type method);
 extern void efx_reset_down(struct efx_nic *efx, enum reset_type method);
 extern int efx_reset_up(struct efx_nic *efx, enum reset_type method, bool ok);
+extern int efx_try_recovery(struct efx_nic *efx);
 
 /* Global */
 extern void efx_schedule_reset(struct efx_nic *efx, enum reset_type type);
diff --git a/drivers/net/ethernet/sfc/ethtool.c b/drivers/net/ethernet/sfc/ethtool.c
index 6e76817..1fc2145 100644
--- a/drivers/net/ethernet/sfc/ethtool.c
+++ b/drivers/net/ethernet/sfc/ethtool.c
@@ -1114,6 +1114,20 @@ static int efx_ethtool_set_rxfh_indir(struct net_device *net_dev,
 	return 0;
 }
 
+int efx_ethtool_get_ts_info(struct net_device *net_dev,
+			    struct ethtool_ts_info *ts_info)
+{
+	struct efx_nic *efx = netdev_priv(net_dev);
+
+	/* Software capabilities */
+	ts_info->so_timestamping = (SOF_TIMESTAMPING_RX_SOFTWARE |
+				    SOF_TIMESTAMPING_SOFTWARE);
+	ts_info->phc_index = -1;
+
+	efx_ptp_get_ts_info(efx, ts_info);
+	return 0;
+}
+
 static int efx_ethtool_get_module_eeprom(struct net_device *net_dev,
 					 struct ethtool_eeprom *ee,
 					 u8 *data)
@@ -1176,7 +1190,7 @@ const struct ethtool_ops efx_ethtool_ops = {
 	.get_rxfh_indir_size	= efx_ethtool_get_rxfh_indir_size,
 	.get_rxfh_indir		= efx_ethtool_get_rxfh_indir,
 	.set_rxfh_indir		= efx_ethtool_set_rxfh_indir,
-	.get_ts_info		= efx_ptp_get_ts_info,
+	.get_ts_info		= efx_ethtool_get_ts_info,
 	.get_module_info	= efx_ethtool_get_module_info,
 	.get_module_eeprom	= efx_ethtool_get_module_eeprom,
 };
diff --git a/drivers/net/ethernet/sfc/filter.c b/drivers/net/ethernet/sfc/filter.c
index 2397f0e..b74a60a 100644
--- a/drivers/net/ethernet/sfc/filter.c
+++ b/drivers/net/ethernet/sfc/filter.c
@@ -1185,8 +1185,21 @@ int efx_filter_rfs(struct net_device *net_dev, const struct sk_buff *skb,
 
 	nhoff = skb_network_offset(skb);
 
-	if (skb->protocol != htons(ETH_P_IP))
+	if (skb->protocol == htons(ETH_P_8021Q)) {
+		EFX_BUG_ON_PARANOID(skb_headlen(skb) <
+				    nhoff + sizeof(struct vlan_hdr));
+		if (((const struct vlan_hdr *)skb->data + nhoff)->
+		    h_vlan_encapsulated_proto != htons(ETH_P_IP))
+			return -EPROTONOSUPPORT;
+
+		/* This is IP over 802.1q VLAN.  We can't filter on the
+		 * IP 5-tuple and the vlan together, so just strip the
+		 * vlan header and filter on the IP part.
+		 */
+		nhoff += sizeof(struct vlan_hdr);
+	} else if (skb->protocol != htons(ETH_P_IP)) {
 		return -EPROTONOSUPPORT;
+	}
 
 	/* RFS must validate the IP header length before calling us */
 	EFX_BUG_ON_PARANOID(skb_headlen(skb) < nhoff + sizeof(*ip));
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index 39d6bd77..f4c7e6b 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -243,6 +243,7 @@ struct efx_rx_buffer {
 #define EFX_RX_BUF_LAST_IN_PAGE	0x0001
 #define EFX_RX_PKT_CSUMMED	0x0002
 #define EFX_RX_PKT_DISCARD	0x0004
+#define EFX_RX_PKT_TCP		0x0040
 
 /**
  * struct efx_rx_page_state - Page-based rx buffer state
@@ -784,9 +785,11 @@ struct efx_nic {
 
 	char name[IFNAMSIZ];
 	struct pci_dev *pci_dev;
+	unsigned int port_num;
 	const struct efx_nic_type *type;
 	int legacy_irq;
 	bool legacy_irq_enabled;
+	bool eeh_disabled_legacy_irq;
 	struct workqueue_struct *workqueue;
 	char workqueue_name[16];
 	struct work_struct reset_work;
@@ -916,7 +919,7 @@ static inline int efx_dev_registered(struct efx_nic *efx)
 
 static inline unsigned int efx_port_num(struct efx_nic *efx)
 {
-	return efx->net_dev->dev_id;
+	return efx->port_num;
 }
 
 /**
diff --git a/drivers/net/ethernet/sfc/nic.c b/drivers/net/ethernet/sfc/nic.c
index b0503cd..56ed3bc 100644
--- a/drivers/net/ethernet/sfc/nic.c
+++ b/drivers/net/ethernet/sfc/nic.c
@@ -14,6 +14,7 @@
 #include <linux/pci.h>
 #include <linux/module.h>
 #include <linux/seq_file.h>
+#include <linux/cpu_rmap.h>
 #include "net_driver.h"
 #include "bitfield.h"
 #include "efx.h"
@@ -1080,12 +1081,21 @@ efx_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
 	rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
 
 	if (likely(rx_ev_pkt_ok)) {
-		/* If packet is marked as OK and packet type is TCP/IP or
-		 * UDP/IP, then we can rely on the hardware checksum.
+		/* If packet is marked as OK then we can rely on the
+		 * hardware checksum and classification.
 		 */
-		flags = (rx_ev_hdr_type == FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_TCP ||
-			 rx_ev_hdr_type == FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_UDP) ?
-			EFX_RX_PKT_CSUMMED : 0;
+		flags = 0;
+		switch (rx_ev_hdr_type) {
+		case FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_TCP:
+			flags |= EFX_RX_PKT_TCP;
+			/* fall through */
+		case FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_UDP:
+			flags |= EFX_RX_PKT_CSUMMED;
+			/* fall through */
+		case FSE_CZ_RX_EV_HDR_TYPE_IPV4V6_OTHER:
+		case FSE_AZ_RX_EV_HDR_TYPE_OTHER:
+			break;
+		}
 	} else {
 		flags = efx_handle_rx_not_ok(rx_queue, event);
 	}
@@ -1579,6 +1589,16 @@ static irqreturn_t efx_legacy_interrupt(int irq, void *dev_id)
 	efx_readd(efx, &reg, FR_BZ_INT_ISR0);
 	queues = EFX_EXTRACT_DWORD(reg, 0, 31);
 
+	/* Legacy interrupts are disabled too late by the EEH kernel
+	 * code. Disable them earlier.
+	 * If an EEH error occurred, the read will have returned all ones.
+	 */
+	if (EFX_DWORD_IS_ALL_ONES(reg) && efx_try_recovery(efx) &&
+	    !efx->eeh_disabled_legacy_irq) {
+		disable_irq_nosync(efx->legacy_irq);
+		efx->eeh_disabled_legacy_irq = true;
+	}
+
 	/* Handle non-event-queue sources */
 	if (queues & (1U << efx->irq_level)) {
 		syserr = EFX_OWORD_FIELD(*int_ker, FSF_AZ_NET_IVEC_FATAL_INT);
@@ -1687,6 +1707,7 @@ void efx_nic_push_rx_indir_table(struct efx_nic *efx)
 int efx_nic_init_interrupt(struct efx_nic *efx)
 {
 	struct efx_channel *channel;
+	unsigned int n_irqs;
 	int rc;
 
 	if (!EFX_INT_MODE_USE_MSI(efx)) {
@@ -1707,7 +1728,19 @@ int efx_nic_init_interrupt(struct efx_nic *efx)
 		return 0;
 	}
 
+#ifdef CONFIG_RFS_ACCEL
+	if (efx->interrupt_mode == EFX_INT_MODE_MSIX) {
+		efx->net_dev->rx_cpu_rmap =
+			alloc_irq_cpu_rmap(efx->n_rx_channels);
+		if (!efx->net_dev->rx_cpu_rmap) {
+			rc = -ENOMEM;
+			goto fail1;
+		}
+	}
+#endif
+
 	/* Hook MSI or MSI-X interrupt */
+	n_irqs = 0;
 	efx_for_each_channel(channel, efx) {
 		rc = request_irq(channel->irq, efx_msi_interrupt,
 				 IRQF_PROBE_SHARED, /* Not shared */
@@ -1718,13 +1751,31 @@ int efx_nic_init_interrupt(struct efx_nic *efx)
 				  "failed to hook IRQ %d\n", channel->irq);
 			goto fail2;
 		}
+		++n_irqs;
+
+#ifdef CONFIG_RFS_ACCEL
+		if (efx->interrupt_mode == EFX_INT_MODE_MSIX &&
+		    channel->channel < efx->n_rx_channels) {
+			rc = irq_cpu_rmap_add(efx->net_dev->rx_cpu_rmap,
+					      channel->irq);
+			if (rc)
+				goto fail2;
+		}
+#endif
 	}
 
 	return 0;
 
  fail2:
-	efx_for_each_channel(channel, efx)
+#ifdef CONFIG_RFS_ACCEL
+	free_irq_cpu_rmap(efx->net_dev->rx_cpu_rmap);
+	efx->net_dev->rx_cpu_rmap = NULL;
+#endif
+	efx_for_each_channel(channel, efx) {
+		if (n_irqs-- == 0)
+			break;
 		free_irq(channel->irq, &efx->channel[channel->channel]);
+	}
  fail1:
 	return rc;
 }
@@ -1734,11 +1785,14 @@ void efx_nic_fini_interrupt(struct efx_nic *efx)
 	struct efx_channel *channel;
 	efx_oword_t reg;
 
+#ifdef CONFIG_RFS_ACCEL
+	free_irq_cpu_rmap(efx->net_dev->rx_cpu_rmap);
+	efx->net_dev->rx_cpu_rmap = NULL;
+#endif
+
 	/* Disable MSI/MSI-X interrupts */
-	efx_for_each_channel(channel, efx) {
-		if (channel->irq)
-			free_irq(channel->irq, &efx->channel[channel->channel]);
-	}
+	efx_for_each_channel(channel, efx)
+		free_irq(channel->irq, &efx->channel[channel->channel]);
 
 	/* ACK legacy interrupt */
 	if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0)
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index 1b00033..d63c299 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -254,8 +254,8 @@ extern int efx_sriov_set_vf_spoofchk(struct net_device *net_dev, int vf,
 struct ethtool_ts_info;
 extern void efx_ptp_probe(struct efx_nic *efx);
 extern int efx_ptp_ioctl(struct efx_nic *efx, struct ifreq *ifr, int cmd);
-extern int efx_ptp_get_ts_info(struct net_device *net_dev,
-			       struct ethtool_ts_info *ts_info);
+extern void efx_ptp_get_ts_info(struct efx_nic *efx,
+				struct ethtool_ts_info *ts_info);
 extern bool efx_ptp_is_ptp_tx(struct efx_nic *efx, struct sk_buff *skb);
 extern int efx_ptp_tx(struct efx_nic *efx, struct sk_buff *skb);
 extern void efx_ptp_event(struct efx_nic *efx, efx_qword_t *ev);
diff --git a/drivers/net/ethernet/sfc/ptp.c b/drivers/net/ethernet/sfc/ptp.c
index 9a95abf..b495394 100644
--- a/drivers/net/ethernet/sfc/ptp.c
+++ b/drivers/net/ethernet/sfc/ptp.c
@@ -1203,18 +1203,16 @@ static int efx_ptp_ts_init(struct efx_nic *efx, struct hwtstamp_config *init)
 	return 0;
 }
 
-int
-efx_ptp_get_ts_info(struct net_device *net_dev, struct ethtool_ts_info *ts_info)
+void efx_ptp_get_ts_info(struct efx_nic *efx, struct ethtool_ts_info *ts_info)
 {
-	struct efx_nic *efx = netdev_priv(net_dev);
 	struct efx_ptp_data *ptp = efx->ptp_data;
 
 	if (!ptp)
-		return -EOPNOTSUPP;
+		return;
 
-	ts_info->so_timestamping = (SOF_TIMESTAMPING_TX_HARDWARE |
-				    SOF_TIMESTAMPING_RX_HARDWARE |
-				    SOF_TIMESTAMPING_RAW_HARDWARE);
+	ts_info->so_timestamping |= (SOF_TIMESTAMPING_TX_HARDWARE |
+				     SOF_TIMESTAMPING_RX_HARDWARE |
+				     SOF_TIMESTAMPING_RAW_HARDWARE);
 	ts_info->phc_index = ptp_clock_index(ptp->phc_clock);
 	ts_info->tx_types = 1 << HWTSTAMP_TX_OFF | 1 << HWTSTAMP_TX_ON;
 	ts_info->rx_filters = (1 << HWTSTAMP_FILTER_NONE |
@@ -1224,7 +1222,6 @@ efx_ptp_get_ts_info(struct net_device *net_dev, struct ethtool_ts_info *ts_info)
 			       1 << HWTSTAMP_FILTER_PTP_V2_L4_EVENT |
 			       1 << HWTSTAMP_FILTER_PTP_V2_L4_SYNC |
 			       1 << HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ);
-	return 0;
 }
 
 int efx_ptp_ioctl(struct efx_nic *efx, struct ifreq *ifr, int cmd)
diff --git a/drivers/net/ethernet/sfc/rx.c b/drivers/net/ethernet/sfc/rx.c
index a7dfe36..6af9cfd 100644
--- a/drivers/net/ethernet/sfc/rx.c
+++ b/drivers/net/ethernet/sfc/rx.c
@@ -36,7 +36,7 @@
 #define EFX_RECYCLE_RING_SIZE_NOIOMMU (2 * EFX_RX_PREFERRED_BATCH)
 
 /* Size of buffer allocated for skb header area. */
-#define EFX_SKB_HEADERS  64u
+#define EFX_SKB_HEADERS  128u
 
 /* This is the percentage fill level below which new RX descriptors
  * will be added to the RX descriptor ring.
@@ -282,9 +282,9 @@ static void efx_fini_rx_buffer(struct efx_rx_queue *rx_queue,
 }
 
 /* Recycle the pages that are used by buffers that have just been received. */
-static void efx_recycle_rx_buffers(struct efx_channel *channel,
-				   struct efx_rx_buffer *rx_buf,
-				   unsigned int n_frags)
+static void efx_recycle_rx_pages(struct efx_channel *channel,
+				 struct efx_rx_buffer *rx_buf,
+				 unsigned int n_frags)
 {
 	struct efx_rx_queue *rx_queue = efx_channel_get_rx_queue(channel);
 
@@ -294,6 +294,20 @@ static void efx_recycle_rx_buffers(struct efx_channel *channel,
 	} while (--n_frags);
 }
 
+static void efx_discard_rx_packet(struct efx_channel *channel,
+				  struct efx_rx_buffer *rx_buf,
+				  unsigned int n_frags)
+{
+	struct efx_rx_queue *rx_queue = efx_channel_get_rx_queue(channel);
+
+	efx_recycle_rx_pages(channel, rx_buf, n_frags);
+
+	do {
+		efx_free_rx_buffer(rx_buf);
+		rx_buf = efx_rx_buf_next(rx_queue, rx_buf);
+	} while (--n_frags);
+}
+
 /**
  * efx_fast_push_rx_descriptors - push new RX descriptors quickly
  * @rx_queue:		RX descriptor queue
@@ -533,8 +547,7 @@ void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
 	 */
 	if (unlikely(rx_buf->flags & EFX_RX_PKT_DISCARD)) {
 		efx_rx_flush_packet(channel);
-		put_page(rx_buf->page);
-		efx_recycle_rx_buffers(channel, rx_buf, n_frags);
+		efx_discard_rx_packet(channel, rx_buf, n_frags);
 		return;
 	}
 
@@ -570,9 +583,9 @@ void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
 		efx_sync_rx_buffer(efx, rx_buf, rx_buf->len);
 	}
 
-	/* All fragments have been DMA-synced, so recycle buffers and pages. */
+	/* All fragments have been DMA-synced, so recycle pages. */
 	rx_buf = efx_rx_buffer(rx_queue, index);
-	efx_recycle_rx_buffers(channel, rx_buf, n_frags);
+	efx_recycle_rx_pages(channel, rx_buf, n_frags);
 
 	/* Pipeline receives so that we give time for packet headers to be
 	 * prefetched into cache.
@@ -598,6 +611,8 @@ static void efx_rx_deliver(struct efx_channel *channel, u8 *eh,
 
 	/* Set the SKB flags */
 	skb_checksum_none_assert(skb);
+	if (likely(rx_buf->flags & EFX_RX_PKT_CSUMMED))
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
 
 	if (channel->type->receive_skb)
 		if (channel->type->receive_skb(channel, skb))
@@ -627,7 +642,7 @@ void __efx_rx_packet(struct efx_channel *channel)
 	if (unlikely(!(efx->net_dev->features & NETIF_F_RXCSUM)))
 		rx_buf->flags &= ~EFX_RX_PKT_CSUMMED;
 
-	if (!channel->type->receive_skb)
+	if ((rx_buf->flags & EFX_RX_PKT_TCP) && !channel->type->receive_skb)
 		efx_rx_packet_gro(channel, rx_buf, channel->rx_pkt_n_frags, eh);
 	else
 		efx_rx_deliver(channel, eh, rx_buf, channel->rx_pkt_n_frags);
@@ -675,7 +690,7 @@ static void efx_init_rx_recycle_ring(struct efx_nic *efx,
 #ifdef CONFIG_PPC64
 	bufs_in_recycle_ring = EFX_RECYCLE_RING_SIZE_IOMMU;
 #else
-	if (efx->pci_dev->dev.iommu_group)
+	if (iommu_present(&pci_bus_type))
 		bufs_in_recycle_ring = EFX_RECYCLE_RING_SIZE_IOMMU;
 	else
 		bufs_in_recycle_ring = EFX_RECYCLE_RING_SIZE_NOIOMMU;
diff --git a/drivers/net/ethernet/sfc/siena.c b/drivers/net/ethernet/sfc/siena.c
index 5166924..8c91775 100644
--- a/drivers/net/ethernet/sfc/siena.c
+++ b/drivers/net/ethernet/sfc/siena.c
@@ -304,7 +304,7 @@ static int siena_probe_nic(struct efx_nic *efx)
 	}
 
 	efx_reado(efx, &reg, FR_AZ_CS_DEBUG);
-	efx->net_dev->dev_id = EFX_OWORD_FIELD(reg, FRF_CZ_CS_PORT_NUM) - 1;
+	efx->port_num = EFX_OWORD_FIELD(reg, FRF_CZ_CS_PORT_NUM) - 1;
 
 	efx_mcdi_init(efx);
 
diff --git a/drivers/net/ethernet/sgi/Kconfig b/drivers/net/ethernet/sgi/Kconfig
index c1c4bb8..e832f46 100644
--- a/drivers/net/ethernet/sgi/Kconfig
+++ b/drivers/net/ethernet/sgi/Kconfig
@@ -22,7 +22,6 @@ config SGI_IOC3_ETH
 	bool "SGI IOC3 Ethernet"
 	depends on PCI && SGI_IP27
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a network (Ethernet) card of this type, say Y and read
diff --git a/drivers/net/ethernet/sgi/ioc3-eth.c b/drivers/net/ethernet/sgi/ioc3-eth.c
index 7ed08c3..ffa7843 100644
--- a/drivers/net/ethernet/sgi/ioc3-eth.c
+++ b/drivers/net/ethernet/sgi/ioc3-eth.c
@@ -1398,16 +1398,6 @@ static struct pci_driver ioc3_driver = {
 	.remove		= ioc3_remove_one,
 };
 
-static int __init ioc3_init_module(void)
-{
-	return pci_register_driver(&ioc3_driver);
-}
-
-static void __exit ioc3_cleanup_module(void)
-{
-	pci_unregister_driver(&ioc3_driver);
-}
-
 static int ioc3_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	unsigned long data;
@@ -1677,9 +1667,7 @@ static void ioc3_set_multicast_list(struct net_device *dev)
 	netif_wake_queue(dev);			/* Let us get going again. */
 }
 
+module_pci_driver(ioc3_driver);
 MODULE_AUTHOR("Ralf Baechle <ralf@linux-mips.org>");
 MODULE_DESCRIPTION("SGI IOC3 Ethernet driver");
 MODULE_LICENSE("GPL");
-
-module_init(ioc3_init_module);
-module_exit(ioc3_cleanup_module);
diff --git a/drivers/net/ethernet/sgi/meth.c b/drivers/net/ethernet/sgi/meth.c
index 4bdbaad9..9f5f35e 100644
--- a/drivers/net/ethernet/sgi/meth.c
+++ b/drivers/net/ethernet/sgi/meth.c
@@ -863,7 +863,6 @@ static int __exit meth_remove(struct platform_device *pdev)
 
 	unregister_netdev(dev);
 	free_netdev(dev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/silan/sc92031.c b/drivers/net/ethernet/silan/sc92031.c
index 28f7268..5eb933c 100644
--- a/drivers/net/ethernet/silan/sc92031.c
+++ b/drivers/net/ethernet/silan/sc92031.c
@@ -1578,19 +1578,7 @@ static struct pci_driver sc92031_pci_driver = {
 	.resume		= sc92031_resume,
 };
 
-static int __init sc92031_init(void)
-{
-	return pci_register_driver(&sc92031_pci_driver);
-}
-
-static void __exit sc92031_exit(void)
-{
-	pci_unregister_driver(&sc92031_pci_driver);
-}
-
-module_init(sc92031_init);
-module_exit(sc92031_exit);
-
+module_pci_driver(sc92031_pci_driver);
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Cesar Eduardo Barros <cesarb@cesarb.net>");
 MODULE_DESCRIPTION("Silan SC92031 PCI Fast Ethernet Adapter driver");
diff --git a/drivers/net/ethernet/sis/Kconfig b/drivers/net/ethernet/sis/Kconfig
index f1135cc..68d052b 100644
--- a/drivers/net/ethernet/sis/Kconfig
+++ b/drivers/net/ethernet/sis/Kconfig
@@ -22,7 +22,6 @@ config SIS900
 	tristate "SiS 900/7016 PCI Fast Ethernet Adapter support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This is a driver for the Fast Ethernet PCI network cards based on
@@ -39,7 +38,6 @@ config SIS190
 	tristate "SiS190/SiS191 gigabit ethernet support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here if you have a SiS 190 PCI Fast Ethernet adapter or
diff --git a/drivers/net/ethernet/sis/sis190.c b/drivers/net/ethernet/sis/sis190.c
index 9a9c379..02df089 100644
--- a/drivers/net/ethernet/sis/sis190.c
+++ b/drivers/net/ethernet/sis/sis190.c
@@ -1934,15 +1934,4 @@ static struct pci_driver sis190_pci_driver = {
 	.remove		= sis190_remove_one,
 };
 
-static int __init sis190_init_module(void)
-{
-	return pci_register_driver(&sis190_pci_driver);
-}
-
-static void __exit sis190_cleanup_module(void)
-{
-	pci_unregister_driver(&sis190_pci_driver);
-}
-
-module_init(sis190_init_module);
-module_exit(sis190_cleanup_module);
+module_pci_driver(sis190_pci_driver);
diff --git a/drivers/net/ethernet/smsc/Kconfig b/drivers/net/ethernet/smsc/Kconfig
index bb4c167..068fc44 100644
--- a/drivers/net/ethernet/smsc/Kconfig
+++ b/drivers/net/ethernet/smsc/Kconfig
@@ -37,7 +37,6 @@ config SMC9194
 config SMC91X
 	tristate "SMC 91C9x/91C1xxx support"
 	select CRC32
-	select NET_CORE
 	select MII
 	depends on (ARM || M32R || SUPERH || MIPS || BLACKFIN || \
 		    MN10300 || COLDFIRE || ARM64)
@@ -57,7 +56,6 @@ config PCMCIA_SMC91C92
 	tristate "SMC 91Cxx PCMCIA support"
 	depends on PCMCIA
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here if you intend to attach an SMC 91Cxx compatible PCMCIA
@@ -70,7 +68,6 @@ config EPIC100
 	tristate "SMC EtherPower II"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  This driver is for the SMC EtherPower II 9432 PCI Ethernet NIC,
@@ -81,7 +78,6 @@ config EPIC100
 config SMC911X
 	tristate "SMSC LAN911[5678] support"
 	select CRC32
-	select NET_CORE
 	select MII
 	depends on (ARM || SUPERH || MN10300)
 	---help---
@@ -97,9 +93,8 @@ config SMC911X
 
 config SMSC911X
 	tristate "SMSC LAN911x/LAN921x families embedded ethernet support"
-	depends on (ARM || SUPERH || BLACKFIN || MIPS || MN10300)
+	depends on HAS_IOMEM
 	select CRC32
-	select NET_CORE
 	select MII
 	select PHYLIB
 	---help---
diff --git a/drivers/net/ethernet/smsc/smc911x.c b/drivers/net/ethernet/smsc/smc911x.c
index 9dd842d..345558f 100644
--- a/drivers/net/ethernet/smsc/smc911x.c
+++ b/drivers/net/ethernet/smsc/smc911x.c
@@ -2087,7 +2087,6 @@ static int smc911x_drv_probe(struct platform_device *pdev)
 	ndev->base_addr = res->start;
 	ret = smc911x_probe(ndev);
 	if (ret != 0) {
-		platform_set_drvdata(pdev, NULL);
 		iounmap(addr);
 release_both:
 		free_netdev(ndev);
@@ -2113,7 +2112,6 @@ static int smc911x_drv_remove(struct platform_device *pdev)
 	struct resource *res;
 
 	DBG(SMC_DEBUG_FUNC, "--> %s\n", __func__);
-	platform_set_drvdata(pdev, NULL);
 
 	unregister_netdev(ndev);
 
diff --git a/drivers/net/ethernet/smsc/smc91x.c b/drivers/net/ethernet/smsc/smc91x.c
index dfbf978..cde13be 100644
--- a/drivers/net/ethernet/smsc/smc91x.c
+++ b/drivers/net/ethernet/smsc/smc91x.c
@@ -2299,7 +2299,6 @@ static int smc_drv_probe(struct platform_device *pdev)
 	return 0;
 
  out_iounmap:
-	platform_set_drvdata(pdev, NULL);
 	iounmap(addr);
  out_release_attrib:
 	smc_release_attrib(pdev, ndev);
@@ -2319,8 +2318,6 @@ static int smc_drv_remove(struct platform_device *pdev)
 	struct smc_local *lp = netdev_priv(ndev);
 	struct resource *res;
 
-	platform_set_drvdata(pdev, NULL);
-
 	unregister_netdev(ndev);
 
 	free_irq(ndev->irq, ndev);
diff --git a/drivers/net/ethernet/smsc/smsc911x.c b/drivers/net/ethernet/smsc/smsc911x.c
index 3663b9e..a141921 100644
--- a/drivers/net/ethernet/smsc/smsc911x.c
+++ b/drivers/net/ethernet/smsc/smsc911x.c
@@ -2284,7 +2284,6 @@ static int smsc911x_drv_remove(struct platform_device *pdev)
 	mdiobus_unregister(pdata->mii_bus);
 	mdiobus_free(pdata->mii_bus);
 
-	platform_set_drvdata(pdev, NULL);
 	unregister_netdev(dev);
 	free_irq(dev->irq, dev);
 	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
@@ -2539,7 +2538,6 @@ out_disable_resources:
 out_enable_resources_fail:
 	smsc911x_free_resources(pdev);
 out_request_resources_fail:
-	platform_set_drvdata(pdev, NULL);
 	iounmap(pdata->ioaddr);
 	free_netdev(dev);
 out_release_io_1:
diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 43c1f32..6e52c0f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -1,7 +1,6 @@
 config STMMAC_ETH
 	tristate "STMicroelectronics 10/100/1000 Ethernet driver"
 	depends on HAS_IOMEM && HAS_DMA
-	select NET_CORE
 	select MII
 	select PHYLIB
 	select CRC32
diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index 9517697..7eb8bab 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -38,16 +38,6 @@
 #include "descs.h"
 #include "mmc.h"
 
-#undef CHIP_DEBUG_PRINT
-/* Turn-on extra printk debug for MAC core, dma and descriptors */
-/* #define CHIP_DEBUG_PRINT */
-
-#ifdef CHIP_DEBUG_PRINT
-#define CHIP_DBG(fmt, args...)  printk(fmt, ## args)
-#else
-#define CHIP_DBG(fmt, args...)  do { } while (0)
-#endif
-
 /* Synopsys Core versions */
 #define	DWMAC_CORE_3_40	0x34
 #define	DWMAC_CORE_3_50	0x35
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
index 7e05e8d..cdd9268 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
@@ -91,8 +91,8 @@ static void dwmac1000_set_filter(struct net_device *dev, int id)
 	unsigned int value = 0;
 	unsigned int perfect_addr_number;
 
-	CHIP_DBG(KERN_INFO "%s: # mcasts %d, # unicast %d\n",
-		 __func__, netdev_mc_count(dev), netdev_uc_count(dev));
+	pr_debug("%s: # mcasts %d, # unicast %d\n", __func__,
+		 netdev_mc_count(dev), netdev_uc_count(dev));
 
 	if (dev->flags & IFF_PROMISC)
 		value = GMAC_FRAME_FILTER_PR;
@@ -152,7 +152,7 @@ static void dwmac1000_set_filter(struct net_device *dev, int id)
 #endif
 	writel(value, ioaddr + GMAC_FRAME_FILTER);
 
-	CHIP_DBG(KERN_INFO "\tFilter: 0x%08x\n\tHash: HI 0x%08x, LO 0x%08x\n",
+	pr_debug("\tFilter: 0x%08x\n\tHash: HI 0x%08x, LO 0x%08x\n",
 		 readl(ioaddr + GMAC_FRAME_FILTER),
 		 readl(ioaddr + GMAC_HASH_HIGH), readl(ioaddr + GMAC_HASH_LOW));
 }
@@ -162,18 +162,18 @@ static void dwmac1000_flow_ctrl(void __iomem *ioaddr, unsigned int duplex,
 {
 	unsigned int flow = 0;
 
-	CHIP_DBG(KERN_DEBUG "GMAC Flow-Control:\n");
+	pr_debug("GMAC Flow-Control:\n");
 	if (fc & FLOW_RX) {
-		CHIP_DBG(KERN_DEBUG "\tReceive Flow-Control ON\n");
+		pr_debug("\tReceive Flow-Control ON\n");
 		flow |= GMAC_FLOW_CTRL_RFE;
 	}
 	if (fc & FLOW_TX) {
-		CHIP_DBG(KERN_DEBUG "\tTransmit Flow-Control ON\n");
+		pr_debug("\tTransmit Flow-Control ON\n");
 		flow |= GMAC_FLOW_CTRL_TFE;
 	}
 
 	if (duplex) {
-		CHIP_DBG(KERN_DEBUG "\tduplex mode: PAUSE %d\n", pause_time);
+		pr_debug("\tduplex mode: PAUSE %d\n", pause_time);
 		flow |= (pause_time << GMAC_FLOW_CTRL_PT_SHIFT);
 	}
 
@@ -185,11 +185,11 @@ static void dwmac1000_pmt(void __iomem *ioaddr, unsigned long mode)
 	unsigned int pmt = 0;
 
 	if (mode & WAKE_MAGIC) {
-		CHIP_DBG(KERN_DEBUG "GMAC: WOL Magic frame\n");
+		pr_debug("GMAC: WOL Magic frame\n");
 		pmt |= power_down | magic_pkt_en;
 	}
 	if (mode & WAKE_UCAST) {
-		CHIP_DBG(KERN_DEBUG "GMAC: WOL on global unicast\n");
+		pr_debug("GMAC: WOL on global unicast\n");
 		pmt |= global_unicast;
 	}
 
@@ -203,23 +203,13 @@ static int dwmac1000_irq_status(void __iomem *ioaddr,
 	int ret = 0;
 
 	/* Not used events (e.g. MMC interrupts) are not handled. */
-	if ((intr_status & mmc_tx_irq)) {
-		CHIP_DBG(KERN_INFO "GMAC: MMC tx interrupt: 0x%08x\n",
-			 readl(ioaddr + GMAC_MMC_TX_INTR));
+	if ((intr_status & mmc_tx_irq))
 		x->mmc_tx_irq_n++;
-	}
-	if (unlikely(intr_status & mmc_rx_irq)) {
-		CHIP_DBG(KERN_INFO "GMAC: MMC rx interrupt: 0x%08x\n",
-			 readl(ioaddr + GMAC_MMC_RX_INTR));
+	if (unlikely(intr_status & mmc_rx_irq))
 		x->mmc_rx_irq_n++;
-	}
-	if (unlikely(intr_status & mmc_rx_csum_offload_irq)) {
-		CHIP_DBG(KERN_INFO "GMAC: MMC rx csum offload: 0x%08x\n",
-			 readl(ioaddr + GMAC_MMC_RX_CSUM_OFFLOAD));
+	if (unlikely(intr_status & mmc_rx_csum_offload_irq))
 		x->mmc_rx_csum_offload_irq_n++;
-	}
 	if (unlikely(intr_status & pmt_irq)) {
-		CHIP_DBG(KERN_INFO "GMAC: received Magic frame\n");
 		/* clear the PMT bits 5 and 6 by reading the PMT status reg */
 		readl(ioaddr + GMAC_PMT);
 		x->irq_receive_pmt_irq_n++;
@@ -229,32 +219,22 @@ static int dwmac1000_irq_status(void __iomem *ioaddr,
 		/* Clean LPI interrupt by reading the Reg 12 */
 		ret = readl(ioaddr + LPI_CTRL_STATUS);
 
-		if (ret & LPI_CTRL_STATUS_TLPIEN) {
-			CHIP_DBG(KERN_INFO "GMAC TX entered in LPI\n");
+		if (ret & LPI_CTRL_STATUS_TLPIEN)
 			x->irq_tx_path_in_lpi_mode_n++;
-		}
-		if (ret & LPI_CTRL_STATUS_TLPIEX) {
-			CHIP_DBG(KERN_INFO "GMAC TX exit from LPI\n");
+		if (ret & LPI_CTRL_STATUS_TLPIEX)
 			x->irq_tx_path_exit_lpi_mode_n++;
-		}
-		if (ret & LPI_CTRL_STATUS_RLPIEN) {
-			CHIP_DBG(KERN_INFO "GMAC RX entered in LPI\n");
+		if (ret & LPI_CTRL_STATUS_RLPIEN)
 			x->irq_rx_path_in_lpi_mode_n++;
-		}
-		if (ret & LPI_CTRL_STATUS_RLPIEX) {
-			CHIP_DBG(KERN_INFO "GMAC RX exit from LPI\n");
+		if (ret & LPI_CTRL_STATUS_RLPIEX)
 			x->irq_rx_path_exit_lpi_mode_n++;
-		}
 	}
 
 	if ((intr_status & pcs_ane_irq) || (intr_status & pcs_link_irq)) {
-		CHIP_DBG(KERN_INFO "GMAC PCS ANE IRQ\n");
 		readl(ioaddr + GMAC_AN_STATUS);
 		x->irq_pcs_ane_n++;
 	}
 	if (intr_status & rgmii_irq) {
 		u32 status = readl(ioaddr + GMAC_S_R_GMII);
-		CHIP_DBG(KERN_INFO "GMAC RGMII/SGMII interrupt\n");
 		x->irq_rgmii_n++;
 
 		/* Save and dump the link status. */
@@ -271,11 +251,12 @@ static int dwmac1000_irq_status(void __iomem *ioaddr,
 				x->pcs_speed = SPEED_10;
 
 			x->pcs_link = 1;
-			pr_debug("Link is Up - %d/%s\n", (int)x->pcs_speed,
+			pr_debug("%s: Link is Up - %d/%s\n", __func__,
+				 (int)x->pcs_speed,
 				 x->pcs_duplex ? "Full" : "Half");
 		} else {
 			x->pcs_link = 0;
-			pr_debug("Link is Down\n");
+			pr_debug("%s: Link is Down\n", __func__);
 		}
 	}
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_dma.c b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_dma.c
index 2c431b6..0c2058a 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_dma.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_dma.c
@@ -116,7 +116,7 @@ static void dwmac1000_dma_operation_mode(void __iomem *ioaddr, int txmode,
 	u32 csr6 = readl(ioaddr + DMA_CONTROL);
 
 	if (txmode == SF_DMA_MODE) {
-		CHIP_DBG(KERN_DEBUG "GMAC: enable TX store and forward mode\n");
+		pr_debug("GMAC: enable TX store and forward mode\n");
 		/* Transmit COE type 2 cannot be done in cut-through mode. */
 		csr6 |= DMA_CONTROL_TSF;
 		/* Operating on second frame increase the performance
@@ -124,8 +124,7 @@ static void dwmac1000_dma_operation_mode(void __iomem *ioaddr, int txmode,
 		 */
 		csr6 |= DMA_CONTROL_OSF;
 	} else {
-		CHIP_DBG(KERN_DEBUG "GMAC: disabling TX SF (threshold %d)\n",
-			 txmode);
+		pr_debug("GMAC: disabling TX SF (threshold %d)\n", txmode);
 		csr6 &= ~DMA_CONTROL_TSF;
 		csr6 &= DMA_CONTROL_TC_TX_MASK;
 		/* Set the transmit threshold */
@@ -142,11 +141,10 @@ static void dwmac1000_dma_operation_mode(void __iomem *ioaddr, int txmode,
 	}
 
 	if (rxmode == SF_DMA_MODE) {
-		CHIP_DBG(KERN_DEBUG "GMAC: enable RX store and forward mode\n");
+		pr_debug("GMAC: enable RX store and forward mode\n");
 		csr6 |= DMA_CONTROL_RSF;
 	} else {
-		CHIP_DBG(KERN_DEBUG "GMAC: disable RX SF mode (threshold %d)\n",
-			 rxmode);
+		pr_debug("GMAC: disable RX SF mode (threshold %d)\n", rxmode);
 		csr6 &= ~DMA_CONTROL_RSF;
 		csr6 &= DMA_CONTROL_TC_RX_MASK;
 		if (rxmode <= 32)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac100_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac100_core.c
index 007bb2b..5857d67 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac100_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac100_core.c
@@ -135,10 +135,6 @@ static void dwmac100_set_filter(struct net_device *dev, int id)
 	}
 
 	writel(value, ioaddr + MAC_CONTROL);
-
-	CHIP_DBG(KERN_INFO "%s: Filter: 0x%08x Hash: HI 0x%08x, LO 0x%08x\n",
-		 __func__, readl(ioaddr + MAC_CONTROL),
-		 readl(ioaddr + MAC_HASH_HIGH), readl(ioaddr + MAC_HASH_LOW));
 }
 
 static void dwmac100_flow_ctrl(void __iomem *ioaddr, unsigned int duplex,
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac100_dma.c b/drivers/net/ethernet/stmicro/stmmac/dwmac100_dma.c
index 67551c1..7d1dce9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac100_dma.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac100_dma.c
@@ -90,14 +90,14 @@ static void dwmac100_dump_dma_regs(void __iomem *ioaddr)
 {
 	int i;
 
-	CHIP_DBG(KERN_DEBUG "DWMAC 100 DMA CSR\n");
+	pr_debug("DWMAC 100 DMA CSR\n");
 	for (i = 0; i < 9; i++)
 		pr_debug("\t CSR%d (offset 0x%x): 0x%08x\n", i,
 			 (DMA_BUS_MODE + i * 4),
 			 readl(ioaddr + DMA_BUS_MODE + i * 4));
-	CHIP_DBG(KERN_DEBUG "\t CSR20 (offset 0x%x): 0x%08x\n",
-		 DMA_CUR_TX_BUF_ADDR, readl(ioaddr + DMA_CUR_TX_BUF_ADDR));
-	CHIP_DBG(KERN_DEBUG "\t CSR21 (offset 0x%x): 0x%08x\n",
+
+	pr_debug("\tCSR20 (0x%x): 0x%08x, CSR21 (0x%x): 0x%08x\n",
+		 DMA_CUR_TX_BUF_ADDR, readl(ioaddr + DMA_CUR_TX_BUF_ADDR),
 		 DMA_CUR_RX_BUF_ADDR, readl(ioaddr + DMA_CUR_RX_BUF_ADDR));
 }
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac_lib.c b/drivers/net/ethernet/stmicro/stmmac/dwmac_lib.c
index 491d7e9..484e3cf 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac_lib.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac_lib.c
@@ -24,13 +24,6 @@
 #include "common.h"
 #include "dwmac_dma.h"
 
-#undef DWMAC_DMA_DEBUG
-#ifdef DWMAC_DMA_DEBUG
-#define DWMAC_LIB_DBG(fmt, args...)  printk(fmt, ## args)
-#else
-#define DWMAC_LIB_DBG(fmt, args...)  do { } while (0)
-#endif
-
 #define GMAC_HI_REG_AE		0x80000000
 
 /* CSR1 enables the transmit DMA to check for new descriptor */
@@ -85,24 +78,24 @@ static void show_tx_process_state(unsigned int status)
 
 	switch (state) {
 	case 0:
-		pr_info("- TX (Stopped): Reset or Stop command\n");
+		pr_debug("- TX (Stopped): Reset or Stop command\n");
 		break;
 	case 1:
-		pr_info("- TX (Running):Fetching the Tx desc\n");
+		pr_debug("- TX (Running):Fetching the Tx desc\n");
 		break;
 	case 2:
-		pr_info("- TX (Running): Waiting for end of tx\n");
+		pr_debug("- TX (Running): Waiting for end of tx\n");
 		break;
 	case 3:
-		pr_info("- TX (Running): Reading the data "
+		pr_debug("- TX (Running): Reading the data "
 		       "and queuing the data into the Tx buf\n");
 		break;
 	case 6:
-		pr_info("- TX (Suspended): Tx Buff Underflow "
+		pr_debug("- TX (Suspended): Tx Buff Underflow "
 		       "or an unavailable Transmit descriptor\n");
 		break;
 	case 7:
-		pr_info("- TX (Running): Closing Tx descriptor\n");
+		pr_debug("- TX (Running): Closing Tx descriptor\n");
 		break;
 	default:
 		break;
@@ -116,29 +109,29 @@ static void show_rx_process_state(unsigned int status)
 
 	switch (state) {
 	case 0:
-		pr_info("- RX (Stopped): Reset or Stop command\n");
+		pr_debug("- RX (Stopped): Reset or Stop command\n");
 		break;
 	case 1:
-		pr_info("- RX (Running): Fetching the Rx desc\n");
+		pr_debug("- RX (Running): Fetching the Rx desc\n");
 		break;
 	case 2:
-		pr_info("- RX (Running):Checking for end of pkt\n");
+		pr_debug("- RX (Running):Checking for end of pkt\n");
 		break;
 	case 3:
-		pr_info("- RX (Running): Waiting for Rx pkt\n");
+		pr_debug("- RX (Running): Waiting for Rx pkt\n");
 		break;
 	case 4:
-		pr_info("- RX (Suspended): Unavailable Rx buf\n");
+		pr_debug("- RX (Suspended): Unavailable Rx buf\n");
 		break;
 	case 5:
-		pr_info("- RX (Running): Closing Rx descriptor\n");
+		pr_debug("- RX (Running): Closing Rx descriptor\n");
 		break;
 	case 6:
-		pr_info("- RX(Running): Flushing the current frame"
+		pr_debug("- RX(Running): Flushing the current frame"
 		       " from the Rx buf\n");
 		break;
 	case 7:
-		pr_info("- RX (Running): Queuing the Rx frame"
+		pr_debug("- RX (Running): Queuing the Rx frame"
 		       " from the Rx buf into memory\n");
 		break;
 	default:
@@ -154,51 +147,37 @@ int dwmac_dma_interrupt(void __iomem *ioaddr,
 	/* read the status register (CSR5) */
 	u32 intr_status = readl(ioaddr + DMA_STATUS);
 
-	DWMAC_LIB_DBG(KERN_INFO "%s: [CSR5: 0x%08x]\n", __func__, intr_status);
 #ifdef DWMAC_DMA_DEBUG
-	/* It displays the DMA process states (CSR5 register) */
+	/* Enable it to monitor DMA rx/tx status in case of critical problems */
+	pr_debug("%s: [CSR5: 0x%08x]\n", __func__, intr_status);
 	show_tx_process_state(intr_status);
 	show_rx_process_state(intr_status);
 #endif
 	/* ABNORMAL interrupts */
 	if (unlikely(intr_status & DMA_STATUS_AIS)) {
-		DWMAC_LIB_DBG(KERN_INFO "CSR5[15] DMA ABNORMAL IRQ: ");
 		if (unlikely(intr_status & DMA_STATUS_UNF)) {
-			DWMAC_LIB_DBG(KERN_INFO "transmit underflow\n");
 			ret = tx_hard_error_bump_tc;
 			x->tx_undeflow_irq++;
 		}
-		if (unlikely(intr_status & DMA_STATUS_TJT)) {
-			DWMAC_LIB_DBG(KERN_INFO "transmit jabber\n");
+		if (unlikely(intr_status & DMA_STATUS_TJT))
 			x->tx_jabber_irq++;
-		}
-		if (unlikely(intr_status & DMA_STATUS_OVF)) {
-			DWMAC_LIB_DBG(KERN_INFO "recv overflow\n");
+
+		if (unlikely(intr_status & DMA_STATUS_OVF))
 			x->rx_overflow_irq++;
-		}
-		if (unlikely(intr_status & DMA_STATUS_RU)) {
-			DWMAC_LIB_DBG(KERN_INFO "receive buffer unavailable\n");
+
+		if (unlikely(intr_status & DMA_STATUS_RU))
 			x->rx_buf_unav_irq++;
-		}
-		if (unlikely(intr_status & DMA_STATUS_RPS)) {
-			DWMAC_LIB_DBG(KERN_INFO "receive process stopped\n");
+		if (unlikely(intr_status & DMA_STATUS_RPS))
 			x->rx_process_stopped_irq++;
-		}
-		if (unlikely(intr_status & DMA_STATUS_RWT)) {
-			DWMAC_LIB_DBG(KERN_INFO "receive watchdog\n");
+		if (unlikely(intr_status & DMA_STATUS_RWT))
 			x->rx_watchdog_irq++;
-		}
-		if (unlikely(intr_status & DMA_STATUS_ETI)) {
-			DWMAC_LIB_DBG(KERN_INFO "transmit early interrupt\n");
+		if (unlikely(intr_status & DMA_STATUS_ETI))
 			x->tx_early_irq++;
-		}
 		if (unlikely(intr_status & DMA_STATUS_TPS)) {
-			DWMAC_LIB_DBG(KERN_INFO "transmit process stopped\n");
 			x->tx_process_stopped_irq++;
 			ret = tx_hard_error;
 		}
 		if (unlikely(intr_status & DMA_STATUS_FBI)) {
-			DWMAC_LIB_DBG(KERN_INFO "fatal bus error\n");
 			x->fatal_bus_error_irq++;
 			ret = tx_hard_error;
 		}
@@ -224,12 +203,11 @@ int dwmac_dma_interrupt(void __iomem *ioaddr,
 	/* Optional hardware blocks, interrupts should be disabled */
 	if (unlikely(intr_status &
 		     (DMA_STATUS_GPI | DMA_STATUS_GMI | DMA_STATUS_GLI)))
-		pr_info("%s: unexpected status %08x\n", __func__, intr_status);
+		pr_warn("%s: unexpected status %08x\n", __func__, intr_status);
 
 	/* Clear the interrupt by writing a logic 1 to the CSR5[15-0] */
 	writel((intr_status & 0x1ffff), ioaddr + DMA_STATUS);
 
-	DWMAC_LIB_DBG(KERN_INFO "\n\n");
 	return ret;
 }
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/enh_desc.c b/drivers/net/ethernet/stmicro/stmmac/enh_desc.c
index 0fbc8fa..7e6628a 100644
--- a/drivers/net/ethernet/stmicro/stmmac/enh_desc.c
+++ b/drivers/net/ethernet/stmicro/stmmac/enh_desc.c
@@ -33,54 +33,40 @@ static int enh_desc_get_tx_status(void *data, struct stmmac_extra_stats *x,
 	struct net_device_stats *stats = (struct net_device_stats *)data;
 
 	if (unlikely(p->des01.etx.error_summary)) {
-		CHIP_DBG(KERN_ERR "GMAC TX error... 0x%08x\n", p->des01.etx);
-		if (unlikely(p->des01.etx.jabber_timeout)) {
-			CHIP_DBG(KERN_ERR "\tjabber_timeout error\n");
+		if (unlikely(p->des01.etx.jabber_timeout))
 			x->tx_jabber++;
-		}
 
 		if (unlikely(p->des01.etx.frame_flushed)) {
-			CHIP_DBG(KERN_ERR "\tframe_flushed error\n");
 			x->tx_frame_flushed++;
 			dwmac_dma_flush_tx_fifo(ioaddr);
 		}
 
 		if (unlikely(p->des01.etx.loss_carrier)) {
-			CHIP_DBG(KERN_ERR "\tloss_carrier error\n");
 			x->tx_losscarrier++;
 			stats->tx_carrier_errors++;
 		}
 		if (unlikely(p->des01.etx.no_carrier)) {
-			CHIP_DBG(KERN_ERR "\tno_carrier error\n");
 			x->tx_carrier++;
 			stats->tx_carrier_errors++;
 		}
-		if (unlikely(p->des01.etx.late_collision)) {
-			CHIP_DBG(KERN_ERR "\tlate_collision error\n");
+		if (unlikely(p->des01.etx.late_collision))
 			stats->collisions += p->des01.etx.collision_count;
-		}
-		if (unlikely(p->des01.etx.excessive_collisions)) {
-			CHIP_DBG(KERN_ERR "\texcessive_collisions\n");
+
+		if (unlikely(p->des01.etx.excessive_collisions))
 			stats->collisions += p->des01.etx.collision_count;
-		}
-		if (unlikely(p->des01.etx.excessive_deferral)) {
-			CHIP_DBG(KERN_INFO "\texcessive tx_deferral\n");
+
+		if (unlikely(p->des01.etx.excessive_deferral))
 			x->tx_deferred++;
-		}
 
 		if (unlikely(p->des01.etx.underflow_error)) {
-			CHIP_DBG(KERN_ERR "\tunderflow error\n");
 			dwmac_dma_flush_tx_fifo(ioaddr);
 			x->tx_underflow++;
 		}
 
-		if (unlikely(p->des01.etx.ip_header_error)) {
-			CHIP_DBG(KERN_ERR "\tTX IP header csum error\n");
+		if (unlikely(p->des01.etx.ip_header_error))
 			x->tx_ip_header_error++;
-		}
 
 		if (unlikely(p->des01.etx.payload_error)) {
-			CHIP_DBG(KERN_ERR "\tAddr/Payload csum error\n");
 			x->tx_payload_error++;
 			dwmac_dma_flush_tx_fifo(ioaddr);
 		}
@@ -88,15 +74,12 @@ static int enh_desc_get_tx_status(void *data, struct stmmac_extra_stats *x,
 		ret = -1;
 	}
 
-	if (unlikely(p->des01.etx.deferred)) {
-		CHIP_DBG(KERN_INFO "GMAC TX status: tx deferred\n");
+	if (unlikely(p->des01.etx.deferred))
 		x->tx_deferred++;
-	}
+
 #ifdef STMMAC_VLAN_TAG_USED
-	if (p->des01.etx.vlan_frame) {
-		CHIP_DBG(KERN_INFO "GMAC TX status: VLAN frame\n");
+	if (p->des01.etx.vlan_frame)
 		x->tx_vlan++;
-	}
 #endif
 
 	return ret;
@@ -123,30 +106,20 @@ static int enh_desc_coe_rdes0(int ipc_err, int type, int payload_err)
 	 *      0 1 1 | COE bypassed.. no IPv4/6 frame
 	 *      0 1 0 | Reserved.
 	 */
-	if (status == 0x0) {
-		CHIP_DBG(KERN_INFO "RX Des0 status: IEEE 802.3 Type frame.\n");
+	if (status == 0x0)
 		ret = llc_snap;
-	} else if (status == 0x4) {
-		CHIP_DBG(KERN_INFO "RX Des0 status: IPv4/6 No CSUM errorS.\n");
+	else if (status == 0x4)
 		ret = good_frame;
-	} else if (status == 0x5) {
-		CHIP_DBG(KERN_ERR "RX Des0 status: IPv4/6 Payload Error.\n");
+	else if (status == 0x5)
 		ret = csum_none;
-	} else if (status == 0x6) {
-		CHIP_DBG(KERN_ERR "RX Des0 status: IPv4/6 Header Error.\n");
+	else if (status == 0x6)
 		ret = csum_none;
-	} else if (status == 0x7) {
-		CHIP_DBG(KERN_ERR
-		    "RX Des0 status: IPv4/6 Header and Payload Error.\n");
+	else if (status == 0x7)
 		ret = csum_none;
-	} else if (status == 0x1) {
-		CHIP_DBG(KERN_ERR
-		    "RX Des0 status: IPv4/6 unsupported IP PAYLOAD.\n");
+	else if (status == 0x1)
 		ret = discard_frame;
-	} else if (status == 0x3) {
-		CHIP_DBG(KERN_ERR "RX Des0 status: No IPv4, IPv6 frame.\n");
+	else if (status == 0x3)
 		ret = discard_frame;
-	}
 	return ret;
 }
 
@@ -208,36 +181,26 @@ static int enh_desc_get_rx_status(void *data, struct stmmac_extra_stats *x,
 	struct net_device_stats *stats = (struct net_device_stats *)data;
 
 	if (unlikely(p->des01.erx.error_summary)) {
-		CHIP_DBG(KERN_ERR "GMAC RX Error Summary 0x%08x\n",
-				  p->des01.erx);
 		if (unlikely(p->des01.erx.descriptor_error)) {
-			CHIP_DBG(KERN_ERR "\tdescriptor error\n");
 			x->rx_desc++;
 			stats->rx_length_errors++;
 		}
-		if (unlikely(p->des01.erx.overflow_error)) {
-			CHIP_DBG(KERN_ERR "\toverflow error\n");
+		if (unlikely(p->des01.erx.overflow_error))
 			x->rx_gmac_overflow++;
-		}
 
 		if (unlikely(p->des01.erx.ipc_csum_error))
-			CHIP_DBG(KERN_ERR "\tIPC Csum Error/Giant frame\n");
+			pr_err("\tIPC Csum Error/Giant frame\n");
 
 		if (unlikely(p->des01.erx.late_collision)) {
-			CHIP_DBG(KERN_ERR "\tlate_collision error\n");
-			stats->collisions++;
 			stats->collisions++;
 		}
-		if (unlikely(p->des01.erx.receive_watchdog)) {
-			CHIP_DBG(KERN_ERR "\treceive_watchdog error\n");
+		if (unlikely(p->des01.erx.receive_watchdog))
 			x->rx_watchdog++;
-		}
-		if (unlikely(p->des01.erx.error_gmii)) {
-			CHIP_DBG(KERN_ERR "\tReceive Error\n");
+
+		if (unlikely(p->des01.erx.error_gmii))
 			x->rx_mii++;
-		}
+
 		if (unlikely(p->des01.erx.crc_error)) {
-			CHIP_DBG(KERN_ERR "\tCRC error\n");
 			x->rx_crc++;
 			stats->rx_crc_errors++;
 		}
@@ -251,30 +214,24 @@ static int enh_desc_get_rx_status(void *data, struct stmmac_extra_stats *x,
 	ret = enh_desc_coe_rdes0(p->des01.erx.ipc_csum_error,
 		p->des01.erx.frame_type, p->des01.erx.rx_mac_addr);
 
-	if (unlikely(p->des01.erx.dribbling)) {
-		CHIP_DBG(KERN_ERR "GMAC RX: dribbling error\n");
+	if (unlikely(p->des01.erx.dribbling))
 		x->dribbling_bit++;
-	}
+
 	if (unlikely(p->des01.erx.sa_filter_fail)) {
-		CHIP_DBG(KERN_ERR "GMAC RX : Source Address filter fail\n");
 		x->sa_rx_filter_fail++;
 		ret = discard_frame;
 	}
 	if (unlikely(p->des01.erx.da_filter_fail)) {
-		CHIP_DBG(KERN_ERR "GMAC RX : Dest Address filter fail\n");
 		x->da_rx_filter_fail++;
 		ret = discard_frame;
 	}
 	if (unlikely(p->des01.erx.length_error)) {
-		CHIP_DBG(KERN_ERR "GMAC RX: length_error error\n");
 		x->rx_length++;
 		ret = discard_frame;
 	}
 #ifdef STMMAC_VLAN_TAG_USED
-	if (p->des01.erx.vlan_tag) {
-		CHIP_DBG(KERN_INFO "GMAC RX: VLAN frame tagged\n");
+	if (p->des01.erx.vlan_tag)
 		x->rx_vlan++;
-	}
 #endif
 
 	return ret;
diff --git a/drivers/net/ethernet/stmicro/stmmac/norm_desc.c b/drivers/net/ethernet/stmicro/stmmac/norm_desc.c
index 11775b9..35ad4f4 100644
--- a/drivers/net/ethernet/stmicro/stmmac/norm_desc.c
+++ b/drivers/net/ethernet/stmicro/stmmac/norm_desc.c
@@ -52,10 +52,8 @@ static int ndesc_get_tx_status(void *data, struct stmmac_extra_stats *x,
 		ret = -1;
 	}
 
-	if (p->des01.etx.vlan_frame) {
-		CHIP_DBG(KERN_INFO "GMAC TX status: VLAN frame\n");
+	if (p->des01.etx.vlan_frame)
 		x->tx_vlan++;
-	}
 
 	if (unlikely(p->des01.tx.deferred))
 		x->tx_deferred++;
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index e9eab29..f2ccb36 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -51,32 +51,6 @@
 #include "stmmac_ptp.h"
 #include "stmmac.h"
 
-#undef STMMAC_DEBUG
-/*#define STMMAC_DEBUG*/
-#ifdef STMMAC_DEBUG
-#define DBG(nlevel, klevel, fmt, args...) \
-		((void)(netif_msg_##nlevel(priv) && \
-		printk(KERN_##klevel fmt, ## args)))
-#else
-#define DBG(nlevel, klevel, fmt, args...) do { } while (0)
-#endif
-
-#undef STMMAC_RX_DEBUG
-/*#define STMMAC_RX_DEBUG*/
-#ifdef STMMAC_RX_DEBUG
-#define RX_DBG(fmt, args...)  printk(fmt, ## args)
-#else
-#define RX_DBG(fmt, args...)  do { } while (0)
-#endif
-
-#undef STMMAC_XMIT_DEBUG
-/*#define STMMAC_XMIT_DEBUG*/
-#ifdef STMMAC_XMIT_DEBUG
-#define TX_DBG(fmt, args...)  printk(fmt, ## args)
-#else
-#define TX_DBG(fmt, args...)  do { } while (0)
-#endif
-
 #define STMMAC_ALIGN(x)	L1_CACHE_ALIGN(x)
 #define JUMBO_LEN	9000
 
@@ -214,19 +188,17 @@ static void stmmac_clk_csr_set(struct stmmac_priv *priv)
 	}
 }
 
-#if defined(STMMAC_XMIT_DEBUG) || defined(STMMAC_RX_DEBUG)
 static void print_pkt(unsigned char *buf, int len)
 {
 	int j;
-	pr_info("len = %d byte, buf addr: 0x%p", len, buf);
+	pr_debug("len = %d byte, buf addr: 0x%p", len, buf);
 	for (j = 0; j < len; j++) {
 		if ((j % 16) == 0)
-			pr_info("\n %03x:", j);
-		pr_info(" %02x", buf[j]);
+			pr_debug("\n %03x:", j);
+		pr_debug(" %02x", buf[j]);
 	}
-	pr_info("\n");
+	pr_debug("\n");
 }
-#endif
 
 /* minimum number of free TX descriptors required to wake up TX process */
 #define STMMAC_TX_THRESH(x)	(x->dma_tx_size/4)
@@ -696,9 +668,6 @@ static void stmmac_adjust_link(struct net_device *dev)
 	if (phydev == NULL)
 		return;
 
-	DBG(probe, DEBUG, "stmmac_adjust_link: called.  address %d link %d\n",
-	    phydev->addr, phydev->link);
-
 	spin_lock_irqsave(&priv->lock, flags);
 
 	if (phydev->link) {
@@ -773,8 +742,6 @@ static void stmmac_adjust_link(struct net_device *dev)
 	priv->eee_enabled = stmmac_eee_init(priv);
 
 	spin_unlock_irqrestore(&priv->lock, flags);
-
-	DBG(probe, DEBUG, "stmmac_adjust_link: exiting\n");
 }
 
 /**
@@ -789,13 +756,13 @@ static void stmmac_check_pcs_mode(struct stmmac_priv *priv)
 	int interface = priv->plat->interface;
 
 	if (priv->dma_cap.pcs) {
-		if ((interface & PHY_INTERFACE_MODE_RGMII) ||
-		    (interface & PHY_INTERFACE_MODE_RGMII_ID) ||
-		    (interface & PHY_INTERFACE_MODE_RGMII_RXID) ||
-		    (interface & PHY_INTERFACE_MODE_RGMII_TXID)) {
+		if ((interface == PHY_INTERFACE_MODE_RGMII) ||
+		    (interface == PHY_INTERFACE_MODE_RGMII_ID) ||
+		    (interface == PHY_INTERFACE_MODE_RGMII_RXID) ||
+		    (interface == PHY_INTERFACE_MODE_RGMII_TXID)) {
 			pr_debug("STMMAC: PCS RGMII support enable\n");
 			priv->pcs = STMMAC_PCS_RGMII;
-		} else if (interface & PHY_INTERFACE_MODE_SGMII) {
+		} else if (interface == PHY_INTERFACE_MODE_SGMII) {
 			pr_debug("STMMAC: PCS SGMII support enable\n");
 			priv->pcs = STMMAC_PCS_SGMII;
 		}
@@ -1015,8 +982,9 @@ static void init_dma_desc_rings(struct net_device *dev)
 	if (bfsize < BUF_SIZE_16KiB)
 		bfsize = stmmac_set_bfsize(dev->mtu, priv->dma_buf_sz);
 
-	DBG(probe, INFO, "stmmac: txsize %d, rxsize %d, bfsize %d\n",
-	    txsize, rxsize, bfsize);
+	if (netif_msg_probe(priv))
+		pr_debug("%s: txsize %d, rxsize %d, bfsize %d\n", __func__,
+			 txsize, rxsize, bfsize);
 
 	if (priv->extend_desc) {
 		priv->dma_erx = dma_alloc_coherent(priv->device, rxsize *
@@ -1052,12 +1020,13 @@ static void init_dma_desc_rings(struct net_device *dev)
 					    GFP_KERNEL);
 	priv->tx_skbuff = kmalloc_array(txsize, sizeof(struct sk_buff *),
 					GFP_KERNEL);
-	if (netif_msg_drv(priv))
+	if (netif_msg_probe(priv)) {
 		pr_debug("(%s) dma_rx_phy=0x%08x dma_tx_phy=0x%08x\n", __func__,
 			 (u32) priv->dma_rx_phy, (u32) priv->dma_tx_phy);
 
-	/* RX INITIALIZATION */
-	DBG(probe, INFO, "stmmac: SKB addresses:\nskb\t\tskb data\tdma data\n");
+		/* RX INITIALIZATION */
+		pr_debug("\tSKB addresses:\nskb\t\tskb data\tdma data\n");
+	}
 	for (i = 0; i < rxsize; i++) {
 		struct dma_desc *p;
 		if (priv->extend_desc)
@@ -1068,8 +1037,10 @@ static void init_dma_desc_rings(struct net_device *dev)
 		if (stmmac_init_rx_buffers(priv, p, i))
 			break;
 
-		DBG(probe, INFO, "[%p]\t[%p]\t[%x]\n", priv->rx_skbuff[i],
-		    priv->rx_skbuff[i]->data, priv->rx_skbuff_dma[i]);
+		if (netif_msg_probe(priv))
+			pr_debug("[%p]\t[%p]\t[%x]\n", priv->rx_skbuff[i],
+				 priv->rx_skbuff[i]->data,
+				 (unsigned int)priv->rx_skbuff_dma[i]);
 	}
 	priv->cur_rx = 0;
 	priv->dirty_rx = (unsigned int)(i - rxsize);
@@ -1244,8 +1215,9 @@ static void stmmac_tx_clean(struct stmmac_priv *priv)
 
 			stmmac_get_tx_hwtstamp(priv, entry, skb);
 		}
-		TX_DBG("%s: curr %d, dirty %d\n", __func__,
-		       priv->cur_tx, priv->dirty_tx);
+		if (netif_msg_tx_done(priv))
+			pr_debug("%s: curr %d, dirty %d\n", __func__,
+				 priv->cur_tx, priv->dirty_tx);
 
 		if (likely(priv->tx_skbuff_dma[entry])) {
 			dma_unmap_single(priv->device,
@@ -1270,7 +1242,8 @@ static void stmmac_tx_clean(struct stmmac_priv *priv)
 		netif_tx_lock(priv->dev);
 		if (netif_queue_stopped(priv->dev) &&
 		    stmmac_tx_avail(priv) > STMMAC_TX_THRESH(priv)) {
-			TX_DBG("%s: restart transmit\n", __func__);
+			if (netif_msg_tx_done(priv))
+				pr_debug("%s: restart transmit\n", __func__);
 			netif_wake_queue(priv->dev);
 		}
 		netif_tx_unlock(priv->dev);
@@ -1579,7 +1552,7 @@ static int stmmac_open(struct net_device *dev)
 		if (ret) {
 			pr_err("%s: Cannot attach to PHY (error: %d)\n",
 			       __func__, ret);
-			goto open_error;
+			goto phy_error;
 		}
 	}
 
@@ -1593,7 +1566,7 @@ static int stmmac_open(struct net_device *dev)
 	ret = stmmac_init_dma_engine(priv);
 	if (ret < 0) {
 		pr_err("%s: DMA initialization failed\n", __func__);
-		goto open_error;
+		goto init_error;
 	}
 
 	/* Copy the MAC addr into the HW  */
@@ -1612,7 +1585,7 @@ static int stmmac_open(struct net_device *dev)
 	if (unlikely(ret < 0)) {
 		pr_err("%s: ERROR: allocating the IRQ %d (error: %d)\n",
 		       __func__, dev->irq, ret);
-		goto open_error;
+		goto init_error;
 	}
 
 	/* Request the Wake IRQ in case of another line is used for WoL */
@@ -1622,7 +1595,7 @@ static int stmmac_open(struct net_device *dev)
 		if (unlikely(ret < 0)) {
 			pr_err("%s: ERROR: allocating the WoL IRQ %d (%d)\n",
 			       __func__, priv->wol_irq, ret);
-			goto open_error_wolirq;
+			goto wolirq_error;
 		}
 	}
 
@@ -1633,7 +1606,7 @@ static int stmmac_open(struct net_device *dev)
 		if (unlikely(ret < 0)) {
 			pr_err("%s: ERROR: allocating the LPI IRQ %d (%d)\n",
 			       __func__, priv->lpi_irq, ret);
-			goto open_error_lpiirq;
+			goto lpiirq_error;
 		}
 	}
 
@@ -1659,7 +1632,7 @@ static int stmmac_open(struct net_device *dev)
 		pr_warn("%s: failed debugFS registration\n", __func__);
 #endif
 	/* Start the ball rolling... */
-	DBG(probe, DEBUG, "%s: DMA RX/TX processes started...\n", dev->name);
+	pr_debug("%s: DMA RX/TX processes started...\n", dev->name);
 	priv->hw->dma->start_tx(priv->ioaddr);
 	priv->hw->dma->start_rx(priv->ioaddr);
 
@@ -1691,17 +1664,17 @@ static int stmmac_open(struct net_device *dev)
 
 	return 0;
 
-open_error_lpiirq:
+lpiirq_error:
 	if (priv->wol_irq != dev->irq)
 		free_irq(priv->wol_irq, dev);
-
-open_error_wolirq:
+wolirq_error:
 	free_irq(dev->irq, dev);
 
-open_error:
+init_error:
+	free_dma_desc_resources(priv);
 	if (priv->phydev)
 		phy_disconnect(priv->phydev);
-
+phy_error:
 	clk_disable_unprepare(priv->stmmac_clk);
 
 	return ret;
@@ -1796,16 +1769,6 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	entry = priv->cur_tx % txsize;
 
-#ifdef STMMAC_XMIT_DEBUG
-	if ((skb->len > ETH_FRAME_LEN) || nfrags)
-		pr_debug("%s: [entry %d]: skb addr %p len: %d nopagedlen: %d\n"
-			 "\tn_frags: %d - ip_summed: %d - %s gso\n"
-			 "\ttx_count_frames %d\n", __func__, entry,
-			 skb, skb->len, nopaged_len, nfrags, skb->ip_summed,
-			 !skb_is_gso(skb) ? "isn't" : "is",
-			 priv->tx_count_frames);
-#endif
-
 	csum_insertion = (skb->ip_summed == CHECKSUM_PARTIAL);
 
 	if (priv->extend_desc)
@@ -1815,12 +1778,6 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	first = desc;
 
-#ifdef STMMAC_XMIT_DEBUG
-	if ((nfrags > 0) || (skb->len > ETH_FRAME_LEN))
-		pr_debug("\tskb len: %d, nopaged_len: %d,\n"
-			 "\t\tn_frags: %d, ip_summed: %d\n",
-			 skb->len, nopaged_len, nfrags, skb->ip_summed);
-#endif
 	priv->tx_skbuff[entry] = skb;
 
 	/* To program the descriptors according to the size of the frame */
@@ -1856,7 +1813,6 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 		else
 			desc = priv->dma_tx + entry;
 
-		TX_DBG("\t[entry %d] segment len: %d\n", entry, len);
 		desc->des2 = skb_frag_dma_map(priv->device, frag, 0, len,
 					      DMA_TO_DEVICE);
 		priv->tx_skbuff_dma[entry] = desc->des2;
@@ -1880,8 +1836,6 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 	if (priv->tx_coal_frames > priv->tx_count_frames) {
 		priv->hw->desc->clear_tx_ic(desc);
 		priv->xstats.tx_reset_ic_bit++;
-		TX_DBG("\t[entry %d]: tx_count_frames %d\n", entry,
-		       priv->tx_count_frames);
 		mod_timer(&priv->txtimer,
 			  STMMAC_COAL_TIMER(priv->tx_coal_timer));
 	} else
@@ -1893,22 +1847,22 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	priv->cur_tx++;
 
-#ifdef STMMAC_XMIT_DEBUG
 	if (netif_msg_pktdata(priv)) {
-		pr_info("%s: curr %d dirty=%d entry=%d, first=%p, nfrags=%d",
+		pr_debug("%s: curr %d dirty=%d entry=%d, first=%p, nfrags=%d",
 			__func__, (priv->cur_tx % txsize),
 			(priv->dirty_tx % txsize), entry, first, nfrags);
+
 		if (priv->extend_desc)
 			stmmac_display_ring((void *)priv->dma_etx, txsize, 1);
 		else
 			stmmac_display_ring((void *)priv->dma_tx, txsize, 0);
 
-		pr_info(">>> frame to be transmitted: ");
+		pr_debug(">>> frame to be transmitted: ");
 		print_pkt(skb->data, skb->len);
 	}
-#endif
 	if (unlikely(stmmac_tx_avail(priv) <= (MAX_SKB_FRAGS + 1))) {
-		TX_DBG("%s: stop transmitted packets\n", __func__);
+		if (netif_msg_hw(priv))
+			pr_debug("%s: stop transmitted packets\n", __func__);
 		netif_stop_queue(dev);
 	}
 
@@ -1968,7 +1922,8 @@ static inline void stmmac_rx_refill(struct stmmac_priv *priv)
 
 			priv->hw->ring->refill_desc3(priv, p);
 
-			RX_DBG(KERN_INFO "\trefill entry #%d\n", entry);
+			if (netif_msg_rx_status(priv))
+				pr_debug("\trefill entry #%d\n", entry);
 		}
 		wmb();
 		priv->hw->desc->set_rx_owner(p);
@@ -1991,15 +1946,13 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit)
 	unsigned int count = 0;
 	int coe = priv->plat->rx_coe;
 
-#ifdef STMMAC_RX_DEBUG
-	if (netif_msg_hw(priv)) {
-		pr_debug(">>> stmmac_rx: descriptor ring:\n");
+	if (netif_msg_rx_status(priv)) {
+		pr_debug("%s: descriptor ring:\n", __func__);
 		if (priv->extend_desc)
 			stmmac_display_ring((void *)priv->dma_erx, rxsize, 1);
 		else
 			stmmac_display_ring((void *)priv->dma_rx, rxsize, 0);
 	}
-#endif
 	while (count < limit) {
 		int status;
 		struct dma_desc *p;
@@ -2053,15 +2006,14 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit)
 			 */
 			if (unlikely(status != llc_snap))
 				frame_len -= ETH_FCS_LEN;
-#ifdef STMMAC_RX_DEBUG
-			if (frame_len > ETH_FRAME_LEN)
-				pr_debug("\tRX frame size %d, COE status: %d\n",
-					 frame_len, status);
 
-			if (netif_msg_hw(priv))
+			if (netif_msg_rx_status(priv)) {
 				pr_debug("\tdesc: %p [entry %d] buff=0x%x\n",
 					 p, entry, p->des2);
-#endif
+				if (frame_len > ETH_FRAME_LEN)
+					pr_debug("\tframe size %d, COE: %d\n",
+						 frame_len, status);
+			}
 			skb = priv->rx_skbuff[entry];
 			if (unlikely(!skb)) {
 				pr_err("%s: Inconsistent Rx descriptor chain\n",
@@ -2078,12 +2030,12 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit)
 			dma_unmap_single(priv->device,
 					 priv->rx_skbuff_dma[entry],
 					 priv->dma_buf_sz, DMA_FROM_DEVICE);
-#ifdef STMMAC_RX_DEBUG
+
 			if (netif_msg_pktdata(priv)) {
-				pr_info(" frame received (%dbytes)", frame_len);
+				pr_debug("frame received (%dbytes)", frame_len);
 				print_pkt(skb->data, frame_len);
 			}
-#endif
+
 			skb->protocol = eth_type_trans(skb, priv->dev);
 
 			if (unlikely(!coe))
@@ -2562,9 +2514,6 @@ static int stmmac_hw_init(struct stmmac_priv *priv)
 	/* Get and dump the chip ID */
 	priv->synopsys_id = stmmac_get_synopsys_id(priv);
 
-	/* To use alternate (extended) or normal descriptor structures */
-	stmmac_selec_desc_mode(priv);
-
 	/* To use the chained or ring mode */
 	if (chain_mode) {
 		priv->hw->chain = &chain_mode_ops;
@@ -2599,6 +2548,9 @@ static int stmmac_hw_init(struct stmmac_priv *priv)
 	} else
 		pr_info(" No HW DMA feature register supported");
 
+	/* To use alternate (extended) or normal descriptor structures */
+	stmmac_selec_desc_mode(priv);
+
 	ret = priv->hw->mac->rx_ipc(priv->ioaddr);
 	if (!ret) {
 		pr_warn(" RX IPC Checksum Offload not configured.\n");
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
index cc15039..fe7bc99 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_mdio.c
@@ -27,6 +27,9 @@
 #include <linux/mii.h>
 #include <linux/phy.h>
 #include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_gpio.h>
+
 #include <asm/io.h>
 
 #include "stmmac.h"
@@ -131,10 +134,46 @@ static int stmmac_mdio_reset(struct mii_bus *bus)
 	struct net_device *ndev = bus->priv;
 	struct stmmac_priv *priv = netdev_priv(ndev);
 	unsigned int mii_address = priv->hw->mii.addr;
+	struct stmmac_mdio_bus_data *data = priv->plat->mdio_bus_data;
+
+#ifdef CONFIG_OF
+	if (priv->device->of_node) {
+		int reset_gpio, active_low;
+
+		if (data->reset_gpio < 0) {
+			struct device_node *np = priv->device->of_node;
+			if (!np)
+				return 0;
+
+			data->reset_gpio = of_get_named_gpio(np,
+						"snps,reset-gpio", 0);
+			if (data->reset_gpio < 0)
+				return 0;
+
+			data->active_low = of_property_read_bool(np,
+						"snps,reset-active-low");
+			of_property_read_u32_array(np,
+				"snps,reset-delays-us", data->delays, 3);
+		}
+
+		reset_gpio = data->reset_gpio;
+		active_low = data->active_low;
+
+		if (!gpio_request(reset_gpio, "mdio-reset")) {
+			gpio_direction_output(reset_gpio, active_low ? 1 : 0);
+			udelay(data->delays[0]);
+			gpio_set_value(reset_gpio, active_low ? 0 : 1);
+			udelay(data->delays[1]);
+			gpio_set_value(reset_gpio, active_low ? 1 : 0);
+			udelay(data->delays[2]);
+			gpio_free(reset_gpio);
+		}
+	}
+#endif
 
-	if (priv->plat->mdio_bus_data->phy_reset) {
+	if (data->phy_reset) {
 		pr_debug("stmmac_mdio_reset: calling phy_reset\n");
-		priv->plat->mdio_bus_data->phy_reset(priv->plat->bsp_priv);
+		data->phy_reset(priv->plat->bsp_priv);
 	}
 
 	/* This is a workaround for problems with the STE101P PHY.
@@ -172,6 +211,11 @@ int stmmac_mdio_register(struct net_device *ndev)
 	else
 		irqlist = priv->mii_irq;
 
+#ifdef CONFIG_OF
+	if (priv->device->of_node)
+		mdio_bus_data->reset_gpio = -1;
+#endif
+
 	new_bus->name = "stmmac";
 	new_bus->read = &stmmac_mdio_read;
 	new_bus->write = &stmmac_mdio_write;
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
index 1d3780f..03de76c 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
@@ -34,12 +34,20 @@ static int stmmac_probe_config_dt(struct platform_device *pdev,
 				  const char **mac)
 {
 	struct device_node *np = pdev->dev.of_node;
+	struct stmmac_dma_cfg *dma_cfg;
 
 	if (!np)
 		return -ENODEV;
 
 	*mac = of_get_mac_address(np);
 	plat->interface = of_get_phy_mode(np);
+
+	plat->bus_id = of_alias_get_id(np, "ethernet");
+	if (plat->bus_id < 0)
+		plat->bus_id = 0;
+
+	of_property_read_u32(np, "snps,phy-addr", &plat->phy_addr);
+
 	plat->mdio_bus_data = devm_kzalloc(&pdev->dev,
 					   sizeof(struct stmmac_mdio_bus_data),
 					   GFP_KERNEL);
@@ -56,6 +64,22 @@ static int stmmac_probe_config_dt(struct platform_device *pdev,
 		plat->pmt = 1;
 	}
 
+	if (of_device_is_compatible(np, "snps,dwmac-3.610") ||
+		of_device_is_compatible(np, "snps,dwmac-3.710")) {
+		plat->enh_desc = 1;
+		plat->bugged_jumbo = 1;
+		plat->force_sf_dma_mode = 1;
+	}
+
+	dma_cfg = devm_kzalloc(&pdev->dev, sizeof(*dma_cfg), GFP_KERNEL);
+	if (!dma_cfg)
+		return -ENOMEM;
+
+	plat->dma_cfg = dma_cfg;
+	of_property_read_u32(np, "snps,pbl", &dma_cfg->pbl);
+	dma_cfg->fixed_burst = of_property_read_bool(np, "snps,fixed-burst");
+	dma_cfg->mixed_burst = of_property_read_bool(np, "snps,mixed-burst");
+
 	return 0;
 }
 #else
@@ -92,8 +116,10 @@ static int stmmac_pltfr_probe(struct platform_device *pdev)
 	if (IS_ERR(addr))
 		return PTR_ERR(addr);
 
+	plat_dat = pdev->dev.platform_data;
 	if (pdev->dev.of_node) {
-		plat_dat = devm_kzalloc(&pdev->dev,
+		if (!plat_dat)
+			plat_dat = devm_kzalloc(&pdev->dev,
 					sizeof(struct plat_stmmacenet_data),
 					GFP_KERNEL);
 		if (!plat_dat) {
@@ -106,8 +132,6 @@ static int stmmac_pltfr_probe(struct platform_device *pdev)
 			pr_err("%s: main dt probe failed", __func__);
 			return ret;
 		}
-	} else {
-		plat_dat = pdev->dev.platform_data;
 	}
 
 	/* Custom initialisation (if needed)*/
@@ -171,8 +195,6 @@ static int stmmac_pltfr_remove(struct platform_device *pdev)
 	if (priv->plat->exit)
 		priv->plat->exit(pdev);
 
-	platform_set_drvdata(pdev, NULL);
-
 	return ret;
 }
 
@@ -230,7 +252,9 @@ static const struct dev_pm_ops stmmac_pltfr_pm_ops;
 
 static const struct of_device_id stmmac_dt_ids[] = {
 	{ .compatible = "st,spear600-gmac"},
+	{ .compatible = "snps,dwmac-3.610"},
 	{ .compatible = "snps,dwmac-3.70a"},
+	{ .compatible = "snps,dwmac-3.710"},
 	{ .compatible = "snps,dwmac"},
 	{ /* sentinel */ }
 };
diff --git a/drivers/net/ethernet/sun/cassini.c b/drivers/net/ethernet/sun/cassini.c
index 4c682a3..759441b 100644
--- a/drivers/net/ethernet/sun/cassini.c
+++ b/drivers/net/ethernet/sun/cassini.c
@@ -808,44 +808,43 @@ static int cas_reset_mii_phy(struct cas *cp)
 	return limit <= 0;
 }
 
-static int cas_saturn_firmware_init(struct cas *cp)
+static void cas_saturn_firmware_init(struct cas *cp)
 {
 	const struct firmware *fw;
 	const char fw_name[] = "sun/cassini.bin";
 	int err;
 
 	if (PHY_NS_DP83065 != cp->phy_id)
-		return 0;
+		return;
 
 	err = request_firmware(&fw, fw_name, &cp->pdev->dev);
 	if (err) {
 		pr_err("Failed to load firmware \"%s\"\n",
 		       fw_name);
-		return err;
+		return;
 	}
 	if (fw->size < 2) {
 		pr_err("bogus length %zu in \"%s\"\n",
 		       fw->size, fw_name);
-		err = -EINVAL;
 		goto out;
 	}
 	cp->fw_load_addr= fw->data[1] << 8 | fw->data[0];
 	cp->fw_size = fw->size - 2;
 	cp->fw_data = vmalloc(cp->fw_size);
-	if (!cp->fw_data) {
-		err = -ENOMEM;
+	if (!cp->fw_data)
 		goto out;
-	}
 	memcpy(cp->fw_data, &fw->data[2], cp->fw_size);
 out:
 	release_firmware(fw);
-	return err;
 }
 
 static void cas_saturn_firmware_load(struct cas *cp)
 {
 	int i;
 
+	if (!cp->fw_data)
+		return;
+
 	cas_phy_powerdown(cp);
 
 	/* expanded memory access mode */
@@ -5083,8 +5082,7 @@ static int cas_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (cas_check_invariants(cp))
 		goto err_out_iounmap;
 	if (cp->cas_flags & CAS_FLAG_SATURN)
-		if (cas_saturn_firmware_init(cp))
-			goto err_out_iounmap;
+		cas_saturn_firmware_init(cp);
 
 	cp->init_block = (struct cas_init_block *)
 		pci_alloc_consistent(pdev, sizeof(struct cas_init_block),
diff --git a/drivers/net/ethernet/sun/niu.c b/drivers/net/ethernet/sun/niu.c
index 95cff98..fa32240 100644
--- a/drivers/net/ethernet/sun/niu.c
+++ b/drivers/net/ethernet/sun/niu.c
@@ -10108,7 +10108,7 @@ static int niu_of_probe(struct platform_device *op)
 		goto err_out_iounmap;
 	}
 
-	dev_set_drvdata(&op->dev, dev);
+	platform_set_drvdata(op, dev);
 
 	niu_device_announce(np);
 
@@ -10145,7 +10145,7 @@ err_out:
 
 static int niu_of_remove(struct platform_device *op)
 {
-	struct net_device *dev = dev_get_drvdata(&op->dev);
+	struct net_device *dev = platform_get_drvdata(op);
 
 	if (dev) {
 		struct niu *np = netdev_priv(dev);
@@ -10175,7 +10175,6 @@ static int niu_of_remove(struct platform_device *op)
 		niu_put_parent(np);
 
 		free_netdev(dev);
-		dev_set_drvdata(&op->dev, NULL);
 	}
 	return 0;
 }
diff --git a/drivers/net/ethernet/sun/sunbmac.c b/drivers/net/ethernet/sun/sunbmac.c
index 0549759..0d43fa9 100644
--- a/drivers/net/ethernet/sun/sunbmac.c
+++ b/drivers/net/ethernet/sun/sunbmac.c
@@ -995,7 +995,6 @@ static void bigmac_set_multicast(struct net_device *dev)
 	struct bigmac *bp = netdev_priv(dev);
 	void __iomem *bregs = bp->bregs;
 	struct netdev_hw_addr *ha;
-	int i;
 	u32 tmp, crc;
 
 	/* Disable the receiver.  The bit self-clears when
@@ -1017,10 +1016,7 @@ static void bigmac_set_multicast(struct net_device *dev)
 		tmp |= BIGMAC_RXCFG_PMISC;
 		sbus_writel(tmp, bregs + BMAC_RXCFG);
 	} else {
-		u16 hash_table[4];
-
-		for (i = 0; i < 4; i++)
-			hash_table[i] = 0;
+		u16 hash_table[4] = { 0 };
 
 		netdev_for_each_mc_addr(ha, dev) {
 			crc = ether_crc_le(6, ha->addr);
diff --git a/drivers/net/ethernet/sun/sungem.c b/drivers/net/ethernet/sun/sungem.c
index 5f3f9d5..e62df2b 100644
--- a/drivers/net/ethernet/sun/sungem.c
+++ b/drivers/net/ethernet/sun/sungem.c
@@ -3028,15 +3028,4 @@ static struct pci_driver gem_driver = {
 #endif /* CONFIG_PM */
 };
 
-static int __init gem_init(void)
-{
-	return pci_register_driver(&gem_driver);
-}
-
-static void __exit gem_cleanup(void)
-{
-	pci_unregister_driver(&gem_driver);
-}
-
-module_init(gem_init);
-module_exit(gem_cleanup);
+module_pci_driver(gem_driver);
diff --git a/drivers/net/ethernet/sun/sunhme.c b/drivers/net/ethernet/sun/sunhme.c
index 436fa9d..171f5b0 100644
--- a/drivers/net/ethernet/sun/sunhme.c
+++ b/drivers/net/ethernet/sun/sunhme.c
@@ -2506,7 +2506,7 @@ static struct quattro *quattro_sbus_find(struct platform_device *child)
 	struct quattro *qp;
 
 	op = to_platform_device(parent);
-	qp = dev_get_drvdata(&op->dev);
+	qp = platform_get_drvdata(op);
 	if (qp)
 		return qp;
 
@@ -2521,7 +2521,7 @@ static struct quattro *quattro_sbus_find(struct platform_device *child)
 		qp->next = qfe_sbus_list;
 		qfe_sbus_list = qp;
 
-		dev_set_drvdata(&op->dev, qp);
+		platform_set_drvdata(op, qp);
 	}
 	return qp;
 }
diff --git a/drivers/net/ethernet/sun/sunqe.c b/drivers/net/ethernet/sun/sunqe.c
index 8182591b..b072f4d 100644
--- a/drivers/net/ethernet/sun/sunqe.c
+++ b/drivers/net/ethernet/sun/sunqe.c
@@ -767,7 +767,7 @@ static struct sunqec *get_qec(struct platform_device *child)
 	struct platform_device *op = to_platform_device(child->dev.parent);
 	struct sunqec *qecp;
 
-	qecp = dev_get_drvdata(&op->dev);
+	qecp = platform_get_drvdata(op);
 	if (!qecp) {
 		qecp = kzalloc(sizeof(struct sunqec), GFP_KERNEL);
 		if (qecp) {
@@ -801,7 +801,7 @@ static struct sunqec *get_qec(struct platform_device *child)
 				goto fail;
 			}
 
-			dev_set_drvdata(&op->dev, qecp);
+			platform_set_drvdata(op, qecp);
 
 			qecp->next_module = root_qec_dev;
 			root_qec_dev = qecp;
@@ -902,7 +902,7 @@ static int qec_ether_init(struct platform_device *op)
 	if (res)
 		goto fail;
 
-	dev_set_drvdata(&op->dev, qe);
+	platform_set_drvdata(op, qe);
 
 	printk(KERN_INFO "%s: qe channel[%d] %pM\n", dev->name, qe->channel,
 	       dev->dev_addr);
@@ -934,7 +934,7 @@ static int qec_sbus_probe(struct platform_device *op)
 
 static int qec_sbus_remove(struct platform_device *op)
 {
-	struct sunqe *qp = dev_get_drvdata(&op->dev);
+	struct sunqe *qp = platform_get_drvdata(op);
 	struct net_device *net_dev = qp->dev;
 
 	unregister_netdev(net_dev);
@@ -948,8 +948,6 @@ static int qec_sbus_remove(struct platform_device *op)
 
 	free_netdev(net_dev);
 
-	dev_set_drvdata(&op->dev, NULL);
-
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/ti/cpsw.c b/drivers/net/ethernet/ti/cpsw.c
index da4415d..05a1674 100644
--- a/drivers/net/ethernet/ti/cpsw.c
+++ b/drivers/net/ethernet/ti/cpsw.c
@@ -1555,6 +1555,8 @@ static int cpsw_probe_dt(struct cpsw_platform_data *data,
 		if (mac_addr)
 			memcpy(slave_data->mac_addr, mac_addr, ETH_ALEN);
 
+		slave_data->phy_if = of_get_phy_mode(slave_node);
+
 		if (data->dual_emac) {
 			if (of_property_read_u32(slave_node, "dual_emac_res_vlan",
 						 &prop)) {
@@ -1702,10 +1704,10 @@ static int cpsw_probe(struct platform_device *pdev)
 
 	if (is_valid_ether_addr(data->slave_data[0].mac_addr)) {
 		memcpy(priv->mac_addr, data->slave_data[0].mac_addr, ETH_ALEN);
-		pr_info("Detected MACID = %pM", priv->mac_addr);
+		pr_info("Detected MACID = %pM\n", priv->mac_addr);
 	} else {
 		eth_random_addr(priv->mac_addr);
-		pr_info("Random MACID = %pM", priv->mac_addr);
+		pr_info("Random MACID = %pM\n", priv->mac_addr);
 	}
 
 	memcpy(ndev->dev_addr, priv->mac_addr, ETH_ALEN);
@@ -1944,7 +1946,6 @@ static int cpsw_remove(struct platform_device *pdev)
 	struct cpsw_priv *priv = netdev_priv(ndev);
 	int i;
 
-	platform_set_drvdata(pdev, NULL);
 	if (priv->data.dual_emac)
 		unregister_netdev(cpsw_get_slave_ndev(priv, 1));
 	unregister_netdev(ndev);
diff --git a/drivers/net/ethernet/ti/davinci_cpdma.c b/drivers/net/ethernet/ti/davinci_cpdma.c
index 053c84f..031ebc8 100644
--- a/drivers/net/ethernet/ti/davinci_cpdma.c
+++ b/drivers/net/ethernet/ti/davinci_cpdma.c
@@ -64,6 +64,7 @@
 #define CPDMA_DESC_TO_PORT_EN	BIT(20)
 #define CPDMA_TO_PORT_SHIFT	16
 #define CPDMA_DESC_PORT_MASK	(BIT(18) | BIT(17) | BIT(16))
+#define CPDMA_DESC_CRC_LEN	4
 
 #define CPDMA_TEARDOWN_VALUE	0xfffffffc
 
@@ -805,6 +806,10 @@ static int __cpdma_chan_process(struct cpdma_chan *chan)
 		status = -EBUSY;
 		goto unlock_ret;
 	}
+
+	if (status & CPDMA_DESC_PASS_CRC)
+		outlen -= CPDMA_DESC_CRC_LEN;
+
 	status	= status & (CPDMA_DESC_EOQ | CPDMA_DESC_TD_COMPLETE |
 			    CPDMA_DESC_PORT_MASK);
 
diff --git a/drivers/net/ethernet/ti/davinci_emac.c b/drivers/net/ethernet/ti/davinci_emac.c
index 860e15d..07b176b 100644
--- a/drivers/net/ethernet/ti/davinci_emac.c
+++ b/drivers/net/ethernet/ti/davinci_emac.c
@@ -1532,7 +1532,7 @@ static int emac_dev_open(struct net_device *ndev)
 	struct device *emac_dev = &ndev->dev;
 	u32 cnt;
 	struct resource *res;
-	int q, m, ret;
+	int ret;
 	int i = 0;
 	int k = 0;
 	struct emac_priv *priv = netdev_priv(ndev);
@@ -1567,8 +1567,9 @@ static int emac_dev_open(struct net_device *ndev)
 
 	while ((res = platform_get_resource(priv->pdev, IORESOURCE_IRQ, k))) {
 		for (i = res->start; i <= res->end; i++) {
-			if (request_irq(i, emac_irq, IRQF_DISABLED,
-					ndev->name, ndev))
+			if (devm_request_irq(&priv->pdev->dev, i, emac_irq,
+					     IRQF_DISABLED,
+					     ndev->name, ndev))
 				goto rollback;
 		}
 		k++;
@@ -1641,15 +1642,7 @@ static int emac_dev_open(struct net_device *ndev)
 
 rollback:
 
-	dev_err(emac_dev, "DaVinci EMAC: request_irq() failed");
-
-	for (q = k; k >= 0; k--) {
-		for (m = i; m >= res->start; m--)
-			free_irq(m, ndev);
-		res = platform_get_resource(priv->pdev, IORESOURCE_IRQ, k-1);
-		m = res->end;
-	}
-
+	dev_err(emac_dev, "DaVinci EMAC: devm_request_irq() failed");
 	ret = -EBUSY;
 err:
 	pm_runtime_put(&priv->pdev->dev);
@@ -1667,9 +1660,6 @@ err:
  */
 static int emac_dev_stop(struct net_device *ndev)
 {
-	struct resource *res;
-	int i = 0;
-	int irq_num;
 	struct emac_priv *priv = netdev_priv(ndev);
 	struct device *emac_dev = &ndev->dev;
 
@@ -1685,13 +1675,6 @@ static int emac_dev_stop(struct net_device *ndev)
 	if (priv->phydev)
 		phy_disconnect(priv->phydev);
 
-	/* Free IRQ */
-	while ((res = platform_get_resource(priv->pdev, IORESOURCE_IRQ, i))) {
-		for (irq_num = res->start; irq_num <= res->end; irq_num++)
-			free_irq(irq_num, priv->ndev);
-		i++;
-	}
-
 	if (netif_msg_drv(priv))
 		dev_notice(emac_dev, "DaVinci EMAC: %s stopped\n", ndev->name);
 
@@ -1771,29 +1754,22 @@ static const struct net_device_ops emac_netdev_ops = {
 #endif
 };
 
-#ifdef CONFIG_OF
-static struct emac_platform_data
-	*davinci_emac_of_get_pdata(struct platform_device *pdev,
-	struct emac_priv *priv)
+static struct emac_platform_data *
+davinci_emac_of_get_pdata(struct platform_device *pdev, struct emac_priv *priv)
 {
 	struct device_node *np;
 	struct emac_platform_data *pdata = NULL;
 	const u8 *mac_addr;
-	u32 data;
-	int ret;
 
-	pdata = pdev->dev.platform_data;
-	if (!pdata) {
-		pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
-		if (!pdata)
-			goto nodata;
-	}
+	if (!IS_ENABLED(CONFIG_OF) || !pdev->dev.of_node)
+		return pdev->dev.platform_data;
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return NULL;
 
 	np = pdev->dev.of_node;
-	if (!np)
-		goto nodata;
-	else
-		pdata->version = EMAC_VERSION_2;
+	pdata->version = EMAC_VERSION_2;
 
 	if (!is_valid_ether_addr(pdata->mac_addr)) {
 		mac_addr = of_get_mac_address(np);
@@ -1801,47 +1777,31 @@ static struct emac_platform_data
 			memcpy(pdata->mac_addr, mac_addr, ETH_ALEN);
 	}
 
-	ret = of_property_read_u32(np, "ti,davinci-ctrl-reg-offset", &data);
-	if (!ret)
-		pdata->ctrl_reg_offset = data;
+	of_property_read_u32(np, "ti,davinci-ctrl-reg-offset",
+			     &pdata->ctrl_reg_offset);
 
-	ret = of_property_read_u32(np, "ti,davinci-ctrl-mod-reg-offset",
-		&data);
-	if (!ret)
-		pdata->ctrl_mod_reg_offset = data;
+	of_property_read_u32(np, "ti,davinci-ctrl-mod-reg-offset",
+			     &pdata->ctrl_mod_reg_offset);
 
-	ret = of_property_read_u32(np, "ti,davinci-ctrl-ram-offset", &data);
-	if (!ret)
-		pdata->ctrl_ram_offset = data;
+	of_property_read_u32(np, "ti,davinci-ctrl-ram-offset",
+			     &pdata->ctrl_ram_offset);
 
-	ret = of_property_read_u32(np, "ti,davinci-ctrl-ram-size", &data);
-	if (!ret)
-		pdata->ctrl_ram_size = data;
+	of_property_read_u32(np, "ti,davinci-ctrl-ram-size",
+			     &pdata->ctrl_ram_size);
 
-	ret = of_property_read_u32(np, "ti,davinci-rmii-en", &data);
-	if (!ret)
-		pdata->rmii_en = data;
+	of_property_read_u8(np, "ti,davinci-rmii-en", &pdata->rmii_en);
 
-	ret = of_property_read_u32(np, "ti,davinci-no-bd-ram", &data);
-	if (!ret)
-		pdata->no_bd_ram = data;
+	pdata->no_bd_ram = of_property_read_bool(np, "ti,davinci-no-bd-ram");
 
 	priv->phy_node = of_parse_phandle(np, "phy-handle", 0);
 	if (!priv->phy_node)
 		pdata->phy_id = "";
 
 	pdev->dev.platform_data = pdata;
-nodata:
+
 	return  pdata;
 }
-#else
-static struct emac_platform_data
-	*davinci_emac_of_get_pdata(struct platform_device *pdev,
-	struct emac_priv *priv)
-{
-	return  pdev->dev.platform_data;
-}
-#endif
+
 /**
  * davinci_emac_probe - EMAC device probe
  * @pdev: The DaVinci EMAC device that we are removing
@@ -1856,7 +1816,7 @@ static int davinci_emac_probe(struct platform_device *pdev)
 	struct resource *res;
 	struct net_device *ndev;
 	struct emac_priv *priv;
-	unsigned long size, hw_ram_addr;
+	unsigned long hw_ram_addr;
 	struct emac_platform_data *pdata;
 	struct device *emac_dev;
 	struct cpdma_params dma_params;
@@ -1907,25 +1867,10 @@ static int davinci_emac_probe(struct platform_device *pdev)
 	emac_dev = &ndev->dev;
 	/* Get EMAC platform data */
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!res) {
-		dev_err(&pdev->dev,"error getting res\n");
-		rc = -ENOENT;
-		goto no_pdata;
-	}
-
 	priv->emac_base_phys = res->start + pdata->ctrl_reg_offset;
-	size = resource_size(res);
-	if (!devm_request_mem_region(&pdev->dev, res->start,
-				     size, ndev->name)) {
-		dev_err(&pdev->dev, "failed request_mem_region() for regs\n");
-		rc = -ENXIO;
-		goto no_pdata;
-	}
-
-	priv->remap_addr = devm_ioremap(&pdev->dev, res->start, size);
-	if (!priv->remap_addr) {
-		dev_err(&pdev->dev, "unable to map IO\n");
-		rc = -ENOMEM;
+	priv->remap_addr = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(priv->remap_addr)) {
+		rc = PTR_ERR(priv->remap_addr);
 		goto no_pdata;
 	}
 	priv->emac_base = priv->remap_addr + pdata->ctrl_reg_offset;
@@ -2037,8 +1982,6 @@ static int davinci_emac_remove(struct platform_device *pdev)
 
 	dev_notice(&ndev->dev, "DaVinci EMAC: davinci_emac_remove()\n");
 
-	platform_set_drvdata(pdev, NULL);
-
 	if (priv->txchan)
 		cpdma_chan_destroy(priv->txchan);
 	if (priv->rxchan)
@@ -2078,11 +2021,13 @@ static const struct dev_pm_ops davinci_emac_pm_ops = {
 	.resume		= davinci_emac_resume,
 };
 
+#if IS_ENABLED(CONFIG_OF)
 static const struct of_device_id davinci_emac_of_match[] = {
 	{.compatible = "ti,davinci-dm6467-emac", },
 	{},
 };
 MODULE_DEVICE_TABLE(of, davinci_emac_of_match);
+#endif
 
 /* davinci_emac_driver: EMAC platform driver structure */
 static struct platform_driver davinci_emac_driver = {
diff --git a/drivers/net/ethernet/ti/davinci_mdio.c b/drivers/net/ethernet/ti/davinci_mdio.c
index ce7c499..16ddfc3 100644
--- a/drivers/net/ethernet/ti/davinci_mdio.c
+++ b/drivers/net/ethernet/ti/davinci_mdio.c
@@ -292,6 +292,7 @@ static int davinci_mdio_write(struct mii_bus *bus, int phy_id,
 	return 0;
 }
 
+#if IS_ENABLED(CONFIG_OF)
 static int davinci_mdio_probe_dt(struct mdio_platform_data *data,
 			 struct platform_device *pdev)
 {
@@ -309,7 +310,7 @@ static int davinci_mdio_probe_dt(struct mdio_platform_data *data,
 
 	return 0;
 }
-
+#endif
 
 static int davinci_mdio_probe(struct platform_device *pdev)
 {
@@ -487,11 +488,13 @@ static const struct dev_pm_ops davinci_mdio_pm_ops = {
 	.resume_early	= davinci_mdio_resume,
 };
 
+#if IS_ENABLED(CONFIG_OF)
 static const struct of_device_id davinci_mdio_of_mtable[] = {
 	{ .compatible = "ti,davinci_mdio", },
 	{ /* sentinel */ },
 };
 MODULE_DEVICE_TABLE(of, davinci_mdio_of_mtable);
+#endif
 
 static struct platform_driver davinci_mdio_driver = {
 	.driver = {
diff --git a/drivers/net/ethernet/ti/tlan.c b/drivers/net/ethernet/ti/tlan.c
index 60c400f..591437e 100644
--- a/drivers/net/ethernet/ti/tlan.c
+++ b/drivers/net/ethernet/ti/tlan.c
@@ -372,7 +372,7 @@ static int tlan_resume(struct pci_dev *pdev)
 
 	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
-	pci_enable_wake(pdev, 0, 0);
+	pci_enable_wake(pdev, PCI_D0, 0);
 	netif_device_attach(dev);
 
 	if (netif_running(dev))
@@ -533,7 +533,6 @@ static int tlan_probe1(struct pci_dev *pdev, long ioaddr, int irq, int rev,
 		/* This is a hack. We need to know which board structure
 		 * is suited for this adapter */
 		device_id = inw(ioaddr + EISA_ID2);
-		priv->is_eisa = 1;
 		if (device_id == 0x20F1) {
 			priv->adapter = &board_info[13]; /* NetFlex-3/E */
 			priv->adapter_rev = 23;		/* TLAN 2.3 */
diff --git a/drivers/net/ethernet/ti/tlan.h b/drivers/net/ethernet/ti/tlan.h
index 5fc98a8..2eb33a2 100644
--- a/drivers/net/ethernet/ti/tlan.h
+++ b/drivers/net/ethernet/ti/tlan.h
@@ -207,7 +207,6 @@ struct tlan_priv {
 	u8			tlan_full_duplex;
 	spinlock_t		lock;
 	u8			link;
-	u8			is_eisa;
 	struct work_struct			tlan_tqueue;
 	u8			neg_be_verbose;
 };
diff --git a/drivers/net/ethernet/toshiba/tc35815.c b/drivers/net/ethernet/toshiba/tc35815.c
index fe25609..a971b9c 100644
--- a/drivers/net/ethernet/toshiba/tc35815.c
+++ b/drivers/net/ethernet/toshiba/tc35815.c
@@ -2209,18 +2209,6 @@ MODULE_PARM_DESC(speed, "0:auto, 10:10Mbps, 100:100Mbps");
 module_param_named(duplex, options.duplex, int, 0);
 MODULE_PARM_DESC(duplex, "0:auto, 1:half, 2:full");
 
-static int __init tc35815_init_module(void)
-{
-	return pci_register_driver(&tc35815_pci_driver);
-}
-
-static void __exit tc35815_cleanup_module(void)
-{
-	pci_unregister_driver(&tc35815_pci_driver);
-}
-
-module_init(tc35815_init_module);
-module_exit(tc35815_cleanup_module);
-
+module_pci_driver(tc35815_pci_driver);
 MODULE_DESCRIPTION("TOSHIBA TC35815 PCI 10M/100M Ethernet driver");
 MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/tundra/tsi108_eth.c b/drivers/net/ethernet/tundra/tsi108_eth.c
index 3c69a04..01bdc6c 100644
--- a/drivers/net/ethernet/tundra/tsi108_eth.c
+++ b/drivers/net/ethernet/tundra/tsi108_eth.c
@@ -1682,7 +1682,6 @@ static int tsi108_ether_remove(struct platform_device *pdev)
 
 	unregister_netdev(dev);
 	tsi108_stop_ethernet(dev);
-	platform_set_drvdata(pdev, NULL);
 	iounmap(priv->regs);
 	iounmap(priv->phyregs);
 	free_netdev(dev);
diff --git a/drivers/net/ethernet/via/Kconfig b/drivers/net/ethernet/via/Kconfig
index 68a9ba6..8a049a2 100644
--- a/drivers/net/ethernet/via/Kconfig
+++ b/drivers/net/ethernet/via/Kconfig
@@ -5,7 +5,6 @@
 config NET_VENDOR_VIA
 	bool "VIA devices"
 	default y
-	depends on PCI
 	---help---
 	  If you have a network (Ethernet) card belonging to this class, say Y
 	  and read the Ethernet-HOWTO, available from
@@ -22,7 +21,6 @@ config VIA_RHINE
 	tristate "VIA Rhine support"
 	depends on PCI
 	select CRC32
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a VIA "Rhine" based network card (Rhine-I (VT86C100A),
@@ -45,10 +43,9 @@ config VIA_RHINE_MMIO
 
 config VIA_VELOCITY
 	tristate "VIA Velocity support"
-	depends on PCI
+	depends on (PCI || USE_OF)
 	select CRC32
 	select CRC_CCITT
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a VIA "Velocity" based network card say Y here.
diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index fb62489..1d6dc41 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -46,6 +46,7 @@
 #include <linux/types.h>
 #include <linux/bitops.h>
 #include <linux/init.h>
+#include <linux/dma-mapping.h>
 #include <linux/mm.h>
 #include <linux/errno.h>
 #include <linux/ioport.h>
@@ -64,7 +65,11 @@
 #include <linux/if.h>
 #include <linux/uaccess.h>
 #include <linux/proc_fs.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_irq.h>
 #include <linux/inetdevice.h>
+#include <linux/platform_device.h>
 #include <linux/reboot.h>
 #include <linux/ethtool.h>
 #include <linux/mii.h>
@@ -79,10 +84,24 @@
 
 #include "via-velocity.h"
 
+enum velocity_bus_type {
+	BUS_PCI,
+	BUS_PLATFORM,
+};
 
 static int velocity_nics;
 static int msglevel = MSG_LEVEL_INFO;
 
+static void velocity_set_power_state(struct velocity_info *vptr, char state)
+{
+	void *addr = vptr->mac_regs;
+
+	if (vptr->pdev)
+		pci_set_power_state(vptr->pdev, state);
+	else
+		writeb(state, addr + 0x154);
+}
+
 /**
  *	mac_get_cam_mask	-	Read a CAM mask
  *	@regs: register block for this velocity
@@ -361,12 +380,23 @@ static struct velocity_info_tbl chip_info_table[] = {
  *	Describe the PCI device identifiers that we support in this
  *	device driver. Used for hotplug autoloading.
  */
-static DEFINE_PCI_DEVICE_TABLE(velocity_id_table) = {
+
+static DEFINE_PCI_DEVICE_TABLE(velocity_pci_id_table) = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_VIA, PCI_DEVICE_ID_VIA_612X) },
 	{ }
 };
 
-MODULE_DEVICE_TABLE(pci, velocity_id_table);
+MODULE_DEVICE_TABLE(pci, velocity_pci_id_table);
+
+/**
+ *	Describe the OF device identifiers that we support in this
+ *	device driver. Used for devicetree nodes.
+ */
+static struct of_device_id velocity_of_ids[] = {
+	{ .compatible = "via,velocity-vt6110", .data = &chip_info_table[0] },
+	{ /* Sentinel */ },
+};
+MODULE_DEVICE_TABLE(of, velocity_of_ids);
 
 /**
  *	get_chip_name	- 	identifier to name
@@ -385,29 +415,6 @@ static const char *get_chip_name(enum chip_type chip_id)
 }
 
 /**
- *	velocity_remove1	-	device unplug
- *	@pdev: PCI device being removed
- *
- *	Device unload callback. Called on an unplug or on module
- *	unload for each active device that is present. Disconnects
- *	the device from the network layer and frees all the resources
- */
-static void velocity_remove1(struct pci_dev *pdev)
-{
-	struct net_device *dev = pci_get_drvdata(pdev);
-	struct velocity_info *vptr = netdev_priv(dev);
-
-	unregister_netdev(dev);
-	iounmap(vptr->mac_regs);
-	pci_release_regions(pdev);
-	pci_disable_device(pdev);
-	pci_set_drvdata(pdev, NULL);
-	free_netdev(dev);
-
-	velocity_nics--;
-}
-
-/**
  *	velocity_set_int_opt	-	parser for integer options
  *	@opt: pointer to option value
  *	@val: value the user requested (or -1 for default)
@@ -998,9 +1005,9 @@ static void velocity_print_link_status(struct velocity_info *vptr)
 {
 
 	if (vptr->mii_status & VELOCITY_LINK_FAIL) {
-		VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: failed to detect cable link\n", vptr->dev->name);
+		VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: failed to detect cable link\n", vptr->netdev->name);
 	} else if (vptr->options.spd_dpx == SPD_DPX_AUTO) {
-		VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link auto-negotiation", vptr->dev->name);
+		VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link auto-negotiation", vptr->netdev->name);
 
 		if (vptr->mii_status & VELOCITY_SPEED_1000)
 			VELOCITY_PRT(MSG_LEVEL_INFO, " speed 1000M bps");
@@ -1014,7 +1021,7 @@ static void velocity_print_link_status(struct velocity_info *vptr)
 		else
 			VELOCITY_PRT(MSG_LEVEL_INFO, " half duplex\n");
 	} else {
-		VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link forced", vptr->dev->name);
+		VELOCITY_PRT(MSG_LEVEL_INFO, KERN_NOTICE "%s: Link forced", vptr->netdev->name);
 		switch (vptr->options.spd_dpx) {
 		case SPD_DPX_1000_FULL:
 			VELOCITY_PRT(MSG_LEVEL_INFO, " speed 1000M bps full duplex\n");
@@ -1180,6 +1187,17 @@ static void mii_init(struct velocity_info *vptr, u32 mii_status)
 	u16 BMCR;
 
 	switch (PHYID_GET_PHY_ID(vptr->phy_id)) {
+	case PHYID_ICPLUS_IP101A:
+		MII_REG_BITS_ON((ADVERTISE_PAUSE_ASYM | ADVERTISE_PAUSE_CAP),
+						MII_ADVERTISE, vptr->mac_regs);
+		if (vptr->mii_status & VELOCITY_DUPLEX_FULL)
+			MII_REG_BITS_ON(TCSR_ECHODIS, MII_SREVISION,
+								vptr->mac_regs);
+		else
+			MII_REG_BITS_OFF(TCSR_ECHODIS, MII_SREVISION,
+								vptr->mac_regs);
+		MII_REG_BITS_ON(PLED_LALBE, MII_TPISTATUS, vptr->mac_regs);
+		break;
 	case PHYID_CICADA_CS8201:
 		/*
 		 *	Reset to hardware default
@@ -1311,6 +1329,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
 				    enum velocity_init_type type)
 {
 	struct mac_regs __iomem *regs = vptr->mac_regs;
+	struct net_device *netdev = vptr->netdev;
 	int i, mii_status;
 
 	mac_wol_reset(regs);
@@ -1319,7 +1338,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
 	case VELOCITY_INIT_RESET:
 	case VELOCITY_INIT_WOL:
 
-		netif_stop_queue(vptr->dev);
+		netif_stop_queue(netdev);
 
 		/*
 		 *	Reset RX to prevent RX pointer not on the 4X location
@@ -1332,7 +1351,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
 		if (velocity_set_media_mode(vptr, mii_status) != VELOCITY_LINK_CHANGE) {
 			velocity_print_link_status(vptr);
 			if (!(vptr->mii_status & VELOCITY_LINK_FAIL))
-				netif_wake_queue(vptr->dev);
+				netif_wake_queue(netdev);
 		}
 
 		enable_flow_control_ability(vptr);
@@ -1352,9 +1371,11 @@ static void velocity_init_registers(struct velocity_info *vptr,
 		velocity_soft_reset(vptr);
 		mdelay(5);
 
-		mac_eeprom_reload(regs);
-		for (i = 0; i < 6; i++)
-			writeb(vptr->dev->dev_addr[i], &(regs->PAR[i]));
+		if (!vptr->no_eeprom) {
+			mac_eeprom_reload(regs);
+			for (i = 0; i < 6; i++)
+				writeb(netdev->dev_addr[i], regs->PAR + i);
+		}
 
 		/*
 		 *	clear Pre_ACPI bit.
@@ -1377,7 +1398,7 @@ static void velocity_init_registers(struct velocity_info *vptr,
 		/*
 		 *	Set packet filter: Receive directed and broadcast address
 		 */
-		velocity_set_multi(vptr->dev);
+		velocity_set_multi(netdev);
 
 		/*
 		 *	Enable MII auto-polling
@@ -1404,14 +1425,14 @@ static void velocity_init_registers(struct velocity_info *vptr,
 		writel((CR0_DPOLL | CR0_TXON | CR0_RXON | CR0_STRT), &regs->CR0Set);
 
 		mii_status = velocity_get_opt_media_mode(vptr);
-		netif_stop_queue(vptr->dev);
+		netif_stop_queue(netdev);
 
 		mii_init(vptr, mii_status);
 
 		if (velocity_set_media_mode(vptr, mii_status) != VELOCITY_LINK_CHANGE) {
 			velocity_print_link_status(vptr);
 			if (!(vptr->mii_status & VELOCITY_LINK_FAIL))
-				netif_wake_queue(vptr->dev);
+				netif_wake_queue(netdev);
 		}
 
 		enable_flow_control_ability(vptr);
@@ -1459,7 +1480,6 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
 	struct velocity_opt *opt = &vptr->options;
 	const unsigned int rx_ring_size = opt->numrx * sizeof(struct rx_desc);
 	const unsigned int tx_ring_size = opt->numtx * sizeof(struct tx_desc);
-	struct pci_dev *pdev = vptr->pdev;
 	dma_addr_t pool_dma;
 	void *pool;
 	unsigned int i;
@@ -1467,14 +1487,14 @@ static int velocity_init_dma_rings(struct velocity_info *vptr)
 	/*
 	 * Allocate all RD/TD rings a single pool.
 	 *
-	 * pci_alloc_consistent() fulfills the requirement for 64 bytes
+	 * dma_alloc_coherent() fulfills the requirement for 64 bytes
 	 * alignment
 	 */
-	pool = pci_alloc_consistent(pdev, tx_ring_size * vptr->tx.numq +
-				    rx_ring_size, &pool_dma);
+	pool = dma_alloc_coherent(vptr->dev, tx_ring_size * vptr->tx.numq +
+				    rx_ring_size, &pool_dma, GFP_ATOMIC);
 	if (!pool) {
-		dev_err(&pdev->dev, "%s : DMA memory allocation failed.\n",
-			vptr->dev->name);
+		dev_err(vptr->dev, "%s : DMA memory allocation failed.\n",
+			vptr->netdev->name);
 		return -ENOMEM;
 	}
 
@@ -1514,7 +1534,7 @@ static int velocity_alloc_rx_buf(struct velocity_info *vptr, int idx)
 	struct rx_desc *rd = &(vptr->rx.ring[idx]);
 	struct velocity_rd_info *rd_info = &(vptr->rx.info[idx]);
 
-	rd_info->skb = netdev_alloc_skb(vptr->dev, vptr->rx.buf_sz + 64);
+	rd_info->skb = netdev_alloc_skb(vptr->netdev, vptr->rx.buf_sz + 64);
 	if (rd_info->skb == NULL)
 		return -ENOMEM;
 
@@ -1524,8 +1544,8 @@ static int velocity_alloc_rx_buf(struct velocity_info *vptr, int idx)
 	 */
 	skb_reserve(rd_info->skb,
 			64 - ((unsigned long) rd_info->skb->data & 63));
-	rd_info->skb_dma = pci_map_single(vptr->pdev, rd_info->skb->data,
-					vptr->rx.buf_sz, PCI_DMA_FROMDEVICE);
+	rd_info->skb_dma = dma_map_single(vptr->dev, rd_info->skb->data,
+					vptr->rx.buf_sz, DMA_FROM_DEVICE);
 
 	/*
 	 *	Fill in the descriptor to match
@@ -1588,8 +1608,8 @@ static void velocity_free_rd_ring(struct velocity_info *vptr)
 
 		if (!rd_info->skb)
 			continue;
-		pci_unmap_single(vptr->pdev, rd_info->skb_dma, vptr->rx.buf_sz,
-				 PCI_DMA_FROMDEVICE);
+		dma_unmap_single(vptr->dev, rd_info->skb_dma, vptr->rx.buf_sz,
+				 DMA_FROM_DEVICE);
 		rd_info->skb_dma = 0;
 
 		dev_kfree_skb(rd_info->skb);
@@ -1620,7 +1640,7 @@ static int velocity_init_rd_ring(struct velocity_info *vptr)
 
 	if (velocity_rx_refill(vptr) != vptr->options.numrx) {
 		VELOCITY_PRT(MSG_LEVEL_ERR, KERN_ERR
-			"%s: failed to allocate RX buffer.\n", vptr->dev->name);
+			"%s: failed to allocate RX buffer.\n", vptr->netdev->name);
 		velocity_free_rd_ring(vptr);
 		goto out;
 	}
@@ -1670,7 +1690,7 @@ static void velocity_free_dma_rings(struct velocity_info *vptr)
 	const int size = vptr->options.numrx * sizeof(struct rx_desc) +
 		vptr->options.numtx * sizeof(struct tx_desc) * vptr->tx.numq;
 
-	pci_free_consistent(vptr->pdev, size, vptr->rx.ring, vptr->rx.pool_dma);
+	dma_free_coherent(vptr->dev, size, vptr->rx.ring, vptr->rx.pool_dma);
 }
 
 static int velocity_init_rings(struct velocity_info *vptr, int mtu)
@@ -1727,8 +1747,8 @@ static void velocity_free_tx_buf(struct velocity_info *vptr,
 				pktlen = max_t(size_t, pktlen,
 						td->td_buf[i].size & ~TD_QUEUE);
 
-			pci_unmap_single(vptr->pdev, tdinfo->skb_dma[i],
-					le16_to_cpu(pktlen), PCI_DMA_TODEVICE);
+			dma_unmap_single(vptr->dev, tdinfo->skb_dma[i],
+					le16_to_cpu(pktlen), DMA_TO_DEVICE);
 		}
 	}
 	dev_kfree_skb_irq(skb);
@@ -1750,8 +1770,8 @@ static void velocity_free_td_ring_entry(struct velocity_info *vptr,
 	if (td_info->skb) {
 		for (i = 0; i < td_info->nskb_dma; i++) {
 			if (td_info->skb_dma[i]) {
-				pci_unmap_single(vptr->pdev, td_info->skb_dma[i],
-					td_info->skb->len, PCI_DMA_TODEVICE);
+				dma_unmap_single(vptr->dev, td_info->skb_dma[i],
+					td_info->skb->len, DMA_TO_DEVICE);
 				td_info->skb_dma[i] = 0;
 			}
 		}
@@ -1809,7 +1829,7 @@ static void velocity_error(struct velocity_info *vptr, int status)
 		printk(KERN_ERR "TD structure error TDindex=%hx\n", readw(&regs->TDIdx[0]));
 		BYTE_REG_BITS_ON(TXESR_TDSTR, &regs->TXESR);
 		writew(TRDCSR_RUN, &regs->TDCSRClr);
-		netif_stop_queue(vptr->dev);
+		netif_stop_queue(vptr->netdev);
 
 		/* FIXME: port over the pci_device_failed code and use it
 		   here */
@@ -1850,10 +1870,10 @@ static void velocity_error(struct velocity_info *vptr, int status)
 
 		if (linked) {
 			vptr->mii_status &= ~VELOCITY_LINK_FAIL;
-			netif_carrier_on(vptr->dev);
+			netif_carrier_on(vptr->netdev);
 		} else {
 			vptr->mii_status |= VELOCITY_LINK_FAIL;
-			netif_carrier_off(vptr->dev);
+			netif_carrier_off(vptr->netdev);
 		}
 
 		velocity_print_link_status(vptr);
@@ -1867,9 +1887,9 @@ static void velocity_error(struct velocity_info *vptr, int status)
 		enable_mii_autopoll(regs);
 
 		if (vptr->mii_status & VELOCITY_LINK_FAIL)
-			netif_stop_queue(vptr->dev);
+			netif_stop_queue(vptr->netdev);
 		else
-			netif_wake_queue(vptr->dev);
+			netif_wake_queue(vptr->netdev);
 
 	}
 	if (status & ISR_MIBFI)
@@ -1894,7 +1914,7 @@ static int velocity_tx_srv(struct velocity_info *vptr)
 	int idx;
 	int works = 0;
 	struct velocity_td_info *tdinfo;
-	struct net_device_stats *stats = &vptr->dev->stats;
+	struct net_device_stats *stats = &vptr->netdev->stats;
 
 	for (qnum = 0; qnum < vptr->tx.numq; qnum++) {
 		for (idx = vptr->tx.tail[qnum]; vptr->tx.used[qnum] > 0;
@@ -1939,9 +1959,9 @@ static int velocity_tx_srv(struct velocity_info *vptr)
 	 *	Look to see if we should kick the transmit network
 	 *	layer for more work.
 	 */
-	if (netif_queue_stopped(vptr->dev) && (full == 0) &&
+	if (netif_queue_stopped(vptr->netdev) && (full == 0) &&
 	    (!(vptr->mii_status & VELOCITY_LINK_FAIL))) {
-		netif_wake_queue(vptr->dev);
+		netif_wake_queue(vptr->netdev);
 	}
 	return works;
 }
@@ -1989,7 +2009,7 @@ static int velocity_rx_copy(struct sk_buff **rx_skb, int pkt_size,
 	if (pkt_size < rx_copybreak) {
 		struct sk_buff *new_skb;
 
-		new_skb = netdev_alloc_skb_ip_align(vptr->dev, pkt_size);
+		new_skb = netdev_alloc_skb_ip_align(vptr->netdev, pkt_size);
 		if (new_skb) {
 			new_skb->ip_summed = rx_skb[0]->ip_summed;
 			skb_copy_from_linear_data(*rx_skb, new_skb->data, pkt_size);
@@ -2029,15 +2049,14 @@ static inline void velocity_iph_realign(struct velocity_info *vptr,
  */
 static int velocity_receive_frame(struct velocity_info *vptr, int idx)
 {
-	void (*pci_action)(struct pci_dev *, dma_addr_t, size_t, int);
-	struct net_device_stats *stats = &vptr->dev->stats;
+	struct net_device_stats *stats = &vptr->netdev->stats;
 	struct velocity_rd_info *rd_info = &(vptr->rx.info[idx]);
 	struct rx_desc *rd = &(vptr->rx.ring[idx]);
 	int pkt_len = le16_to_cpu(rd->rdesc0.len) & 0x3fff;
 	struct sk_buff *skb;
 
 	if (rd->rdesc0.RSR & (RSR_STP | RSR_EDP)) {
-		VELOCITY_PRT(MSG_LEVEL_VERBOSE, KERN_ERR " %s : the received frame span multple RDs.\n", vptr->dev->name);
+		VELOCITY_PRT(MSG_LEVEL_VERBOSE, KERN_ERR " %s : the received frame span multple RDs.\n", vptr->netdev->name);
 		stats->rx_length_errors++;
 		return -EINVAL;
 	}
@@ -2047,8 +2066,8 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
 
 	skb = rd_info->skb;
 
-	pci_dma_sync_single_for_cpu(vptr->pdev, rd_info->skb_dma,
-				    vptr->rx.buf_sz, PCI_DMA_FROMDEVICE);
+	dma_sync_single_for_cpu(vptr->dev, rd_info->skb_dma,
+				    vptr->rx.buf_sz, DMA_FROM_DEVICE);
 
 	/*
 	 *	Drop frame not meeting IEEE 802.3
@@ -2061,21 +2080,20 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
 		}
 	}
 
-	pci_action = pci_dma_sync_single_for_device;
-
 	velocity_rx_csum(rd, skb);
 
 	if (velocity_rx_copy(&skb, pkt_len, vptr) < 0) {
 		velocity_iph_realign(vptr, skb, pkt_len);
-		pci_action = pci_unmap_single;
 		rd_info->skb = NULL;
+		dma_unmap_single(vptr->dev, rd_info->skb_dma, vptr->rx.buf_sz,
+				 DMA_FROM_DEVICE);
+	} else {
+		dma_sync_single_for_device(vptr->dev, rd_info->skb_dma,
+					   vptr->rx.buf_sz, DMA_FROM_DEVICE);
 	}
 
-	pci_action(vptr->pdev, rd_info->skb_dma, vptr->rx.buf_sz,
-		   PCI_DMA_FROMDEVICE);
-
 	skb_put(skb, pkt_len - 4);
-	skb->protocol = eth_type_trans(skb, vptr->dev);
+	skb->protocol = eth_type_trans(skb, vptr->netdev);
 
 	if (rd->rdesc0.RSR & RSR_DETAG) {
 		u16 vid = swab16(le16_to_cpu(rd->rdesc1.PQTAG));
@@ -2100,7 +2118,7 @@ static int velocity_receive_frame(struct velocity_info *vptr, int idx)
  */
 static int velocity_rx_srv(struct velocity_info *vptr, int budget_left)
 {
-	struct net_device_stats *stats = &vptr->dev->stats;
+	struct net_device_stats *stats = &vptr->netdev->stats;
 	int rd_curr = vptr->rx.curr;
 	int works = 0;
 
@@ -2235,15 +2253,15 @@ static int velocity_open(struct net_device *dev)
 		goto out;
 
 	/* Ensure chip is running */
-	pci_set_power_state(vptr->pdev, PCI_D0);
+	velocity_set_power_state(vptr, PCI_D0);
 
 	velocity_init_registers(vptr, VELOCITY_INIT_COLD);
 
-	ret = request_irq(vptr->pdev->irq, velocity_intr, IRQF_SHARED,
+	ret = request_irq(dev->irq, velocity_intr, IRQF_SHARED,
 			  dev->name, dev);
 	if (ret < 0) {
 		/* Power down the chip */
-		pci_set_power_state(vptr->pdev, PCI_D3hot);
+		velocity_set_power_state(vptr, PCI_D3hot);
 		velocity_free_rings(vptr);
 		goto out;
 	}
@@ -2292,7 +2310,7 @@ static int velocity_change_mtu(struct net_device *dev, int new_mtu)
 
 	if ((new_mtu < VELOCITY_MIN_MTU) || new_mtu > (VELOCITY_MAX_MTU)) {
 		VELOCITY_PRT(MSG_LEVEL_ERR, KERN_NOTICE "%s: Invalid MTU.\n",
-				vptr->dev->name);
+				vptr->netdev->name);
 		ret = -EINVAL;
 		goto out_0;
 	}
@@ -2314,8 +2332,9 @@ static int velocity_change_mtu(struct net_device *dev, int new_mtu)
 			goto out_0;
 		}
 
-		tmp_vptr->dev = dev;
+		tmp_vptr->netdev = dev;
 		tmp_vptr->pdev = vptr->pdev;
+		tmp_vptr->dev = vptr->dev;
 		tmp_vptr->options = vptr->options;
 		tmp_vptr->tx.numq = vptr->tx.numq;
 
@@ -2415,7 +2434,7 @@ static int velocity_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
 	   saving then we need to bring the device back up to talk to it */
 
 	if (!netif_running(dev))
-		pci_set_power_state(vptr->pdev, PCI_D0);
+		velocity_set_power_state(vptr, PCI_D0);
 
 	switch (cmd) {
 	case SIOCGMIIPHY:	/* Get address of MII PHY in use. */
@@ -2428,7 +2447,7 @@ static int velocity_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
 		ret = -EOPNOTSUPP;
 	}
 	if (!netif_running(dev))
-		pci_set_power_state(vptr->pdev, PCI_D3hot);
+		velocity_set_power_state(vptr, PCI_D3hot);
 
 
 	return ret;
@@ -2494,7 +2513,7 @@ static int velocity_close(struct net_device *dev)
 	if (vptr->flags & VELOCITY_FLAGS_WOL_ENABLED)
 		velocity_get_ip(vptr);
 
-	free_irq(vptr->pdev->irq, dev);
+	free_irq(dev->irq, dev);
 
 	velocity_free_rings(vptr);
 
@@ -2550,7 +2569,8 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
 	 *	add it to the transmit ring.
 	 */
 	tdinfo->skb = skb;
-	tdinfo->skb_dma[0] = pci_map_single(vptr->pdev, skb->data, pktlen, PCI_DMA_TODEVICE);
+	tdinfo->skb_dma[0] = dma_map_single(vptr->dev, skb->data, pktlen,
+								DMA_TO_DEVICE);
 	td_ptr->tdesc0.len = cpu_to_le16(pktlen);
 	td_ptr->td_buf[0].pa_low = cpu_to_le32(tdinfo->skb_dma[0]);
 	td_ptr->td_buf[0].pa_high = 0;
@@ -2560,7 +2580,7 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
 	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
 		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 
-		tdinfo->skb_dma[i + 1] = skb_frag_dma_map(&vptr->pdev->dev,
+		tdinfo->skb_dma[i + 1] = skb_frag_dma_map(vptr->dev,
 							  frag, 0,
 							  skb_frag_size(frag),
 							  DMA_TO_DEVICE);
@@ -2632,12 +2652,9 @@ static const struct net_device_ops velocity_netdev_ops = {
  *	Set up the initial velocity_info struct for the device that has been
  *	discovered.
  */
-static void velocity_init_info(struct pci_dev *pdev, struct velocity_info *vptr,
-			       const struct velocity_info_tbl *info)
+static void velocity_init_info(struct velocity_info *vptr,
+				const struct velocity_info_tbl *info)
 {
-	memset(vptr, 0, sizeof(struct velocity_info));
-
-	vptr->pdev = pdev;
 	vptr->chip_id = info->chip_id;
 	vptr->tx.numq = info->txqueue;
 	vptr->multicast_limit = MCAM_SIZE;
@@ -2652,10 +2669,9 @@ static void velocity_init_info(struct pci_dev *pdev, struct velocity_info *vptr,
  *	Retrieve the PCI configuration space data that interests us from
  *	the kernel PCI layer
  */
-static int velocity_get_pci_info(struct velocity_info *vptr,
-				 struct pci_dev *pdev)
+static int velocity_get_pci_info(struct velocity_info *vptr)
 {
-	vptr->rev_id = pdev->revision;
+	struct pci_dev *pdev = vptr->pdev;
 
 	pci_set_master(pdev);
 
@@ -2678,7 +2694,37 @@ static int velocity_get_pci_info(struct velocity_info *vptr,
 		dev_err(&pdev->dev, "region #1 is too small.\n");
 		return -EINVAL;
 	}
-	vptr->pdev = pdev;
+
+	return 0;
+}
+
+/**
+ *	velocity_get_platform_info - retrieve platform info for device
+ *	@vptr: velocity device
+ *	@pdev: platform device it matches
+ *
+ *	Retrieve the Platform configuration data that interests us
+ */
+static int velocity_get_platform_info(struct velocity_info *vptr)
+{
+	struct resource res;
+	int ret;
+
+	if (of_get_property(vptr->dev->of_node, "no-eeprom", NULL))
+		vptr->no_eeprom = 1;
+
+	ret = of_address_to_resource(vptr->dev->of_node, 0, &res);
+	if (ret) {
+		dev_err(vptr->dev, "unable to find memory address\n");
+		return ret;
+	}
+
+	vptr->memaddr = res.start;
+
+	if (resource_size(&res) < VELOCITY_IO_SIZE) {
+		dev_err(vptr->dev, "memory region is too small.\n");
+		return -EINVAL;
+	}
 
 	return 0;
 }
@@ -2692,7 +2738,7 @@ static int velocity_get_pci_info(struct velocity_info *vptr,
  */
 static void velocity_print_info(struct velocity_info *vptr)
 {
-	struct net_device *dev = vptr->dev;
+	struct net_device *dev = vptr->netdev;
 
 	printk(KERN_INFO "%s: %s\n", dev->name, get_chip_name(vptr->chip_id));
 	printk(KERN_INFO "%s: Ethernet Address: %pM\n",
@@ -2707,21 +2753,22 @@ static u32 velocity_get_link(struct net_device *dev)
 }
 
 /**
- *	velocity_found1		-	set up discovered velocity card
+ *	velocity_probe - set up discovered velocity device
  *	@pdev: PCI device
  *	@ent: PCI device table entry that matched
+ *	@bustype: bus that device is connected to
  *
  *	Configure a discovered adapter from scratch. Return a negative
  *	errno error code on failure paths.
  */
-static int velocity_found1(struct pci_dev *pdev,
-			   const struct pci_device_id *ent)
+static int velocity_probe(struct device *dev, int irq,
+			   const struct velocity_info_tbl *info,
+			   enum velocity_bus_type bustype)
 {
 	static int first = 1;
-	struct net_device *dev;
+	struct net_device *netdev;
 	int i;
 	const char *drv_string;
-	const struct velocity_info_tbl *info = &chip_info_table[ent->driver_data];
 	struct velocity_info *vptr;
 	struct mac_regs __iomem *regs;
 	int ret = -ENOMEM;
@@ -2730,20 +2777,18 @@ static int velocity_found1(struct pci_dev *pdev,
 	 * can support more than MAX_UNITS.
 	 */
 	if (velocity_nics >= MAX_UNITS) {
-		dev_notice(&pdev->dev, "already found %d NICs.\n",
-			   velocity_nics);
+		dev_notice(dev, "already found %d NICs.\n", velocity_nics);
 		return -ENODEV;
 	}
 
-	dev = alloc_etherdev(sizeof(struct velocity_info));
-	if (!dev)
+	netdev = alloc_etherdev(sizeof(struct velocity_info));
+	if (!netdev)
 		goto out;
 
 	/* Chain it all together */
 
-	SET_NETDEV_DEV(dev, &pdev->dev);
-	vptr = netdev_priv(dev);
-
+	SET_NETDEV_DEV(netdev, dev);
+	vptr = netdev_priv(netdev);
 
 	if (first) {
 		printk(KERN_INFO "%s Ver. %s\n",
@@ -2753,41 +2798,41 @@ static int velocity_found1(struct pci_dev *pdev,
 		first = 0;
 	}
 
-	velocity_init_info(pdev, vptr, info);
-
+	netdev->irq = irq;
+	vptr->netdev = netdev;
 	vptr->dev = dev;
 
-	ret = pci_enable_device(pdev);
-	if (ret < 0)
-		goto err_free_dev;
+	velocity_init_info(vptr, info);
 
-	ret = velocity_get_pci_info(vptr, pdev);
-	if (ret < 0) {
-		/* error message already printed */
-		goto err_disable;
-	}
+	if (bustype == BUS_PCI) {
+		vptr->pdev = to_pci_dev(dev);
 
-	ret = pci_request_regions(pdev, VELOCITY_NAME);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "No PCI resources.\n");
-		goto err_disable;
+		ret = velocity_get_pci_info(vptr);
+		if (ret < 0)
+			goto err_free_dev;
+	} else {
+		vptr->pdev = NULL;
+		ret = velocity_get_platform_info(vptr);
+		if (ret < 0)
+			goto err_free_dev;
 	}
 
 	regs = ioremap(vptr->memaddr, VELOCITY_IO_SIZE);
 	if (regs == NULL) {
 		ret = -EIO;
-		goto err_release_res;
+		goto err_free_dev;
 	}
 
 	vptr->mac_regs = regs;
+	vptr->rev_id = readb(&regs->rev_id);
 
 	mac_wol_reset(regs);
 
 	for (i = 0; i < 6; i++)
-		dev->dev_addr[i] = readb(&regs->PAR[i]);
+		netdev->dev_addr[i] = readb(&regs->PAR[i]);
 
 
-	drv_string = dev_driver_string(&pdev->dev);
+	drv_string = dev_driver_string(dev);
 
 	velocity_get_options(&vptr->options, velocity_nics, drv_string);
 
@@ -2808,46 +2853,125 @@ static int velocity_found1(struct pci_dev *pdev,
 
 	vptr->phy_id = MII_GET_PHY_ID(vptr->mac_regs);
 
-	dev->netdev_ops = &velocity_netdev_ops;
-	dev->ethtool_ops = &velocity_ethtool_ops;
-	netif_napi_add(dev, &vptr->napi, velocity_poll, VELOCITY_NAPI_WEIGHT);
+	netdev->netdev_ops = &velocity_netdev_ops;
+	netdev->ethtool_ops = &velocity_ethtool_ops;
+	netif_napi_add(netdev, &vptr->napi, velocity_poll,
+							VELOCITY_NAPI_WEIGHT);
 
-	dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_SG |
+	netdev->hw_features = NETIF_F_IP_CSUM | NETIF_F_SG |
 			   NETIF_F_HW_VLAN_CTAG_TX;
-	dev->features |= NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_FILTER |
-			 NETIF_F_HW_VLAN_CTAG_RX | NETIF_F_IP_CSUM;
+	netdev->features |= NETIF_F_HW_VLAN_CTAG_TX |
+			NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_HW_VLAN_CTAG_RX |
+			NETIF_F_IP_CSUM;
 
-	ret = register_netdev(dev);
+	ret = register_netdev(netdev);
 	if (ret < 0)
 		goto err_iounmap;
 
-	if (!velocity_get_link(dev)) {
-		netif_carrier_off(dev);
+	if (!velocity_get_link(netdev)) {
+		netif_carrier_off(netdev);
 		vptr->mii_status |= VELOCITY_LINK_FAIL;
 	}
 
 	velocity_print_info(vptr);
-	pci_set_drvdata(pdev, dev);
+	dev_set_drvdata(vptr->dev, netdev);
 
 	/* and leave the chip powered down */
 
-	pci_set_power_state(pdev, PCI_D3hot);
+	velocity_set_power_state(vptr, PCI_D3hot);
 	velocity_nics++;
 out:
 	return ret;
 
 err_iounmap:
 	iounmap(regs);
-err_release_res:
-	pci_release_regions(pdev);
-err_disable:
-	pci_disable_device(pdev);
 err_free_dev:
-	free_netdev(dev);
+	free_netdev(netdev);
 	goto out;
 }
 
-#ifdef CONFIG_PM
+/**
+ *	velocity_remove	- device unplug
+ *	@dev: device being removed
+ *
+ *	Device unload callback. Called on an unplug or on module
+ *	unload for each active device that is present. Disconnects
+ *	the device from the network layer and frees all the resources
+ */
+static int velocity_remove(struct device *dev)
+{
+	struct net_device *netdev = dev_get_drvdata(dev);
+	struct velocity_info *vptr = netdev_priv(netdev);
+
+	unregister_netdev(netdev);
+	iounmap(vptr->mac_regs);
+	free_netdev(netdev);
+	velocity_nics--;
+
+	return 0;
+}
+
+static int velocity_pci_probe(struct pci_dev *pdev,
+			       const struct pci_device_id *ent)
+{
+	const struct velocity_info_tbl *info =
+					&chip_info_table[ent->driver_data];
+	int ret;
+
+	ret = pci_enable_device(pdev);
+	if (ret < 0)
+		return ret;
+
+	ret = pci_request_regions(pdev, VELOCITY_NAME);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "No PCI resources.\n");
+		goto fail1;
+	}
+
+	ret = velocity_probe(&pdev->dev, pdev->irq, info, BUS_PCI);
+	if (ret == 0)
+		return 0;
+
+	pci_release_regions(pdev);
+fail1:
+	pci_disable_device(pdev);
+	return ret;
+}
+
+static void velocity_pci_remove(struct pci_dev *pdev)
+{
+	velocity_remove(&pdev->dev);
+
+	pci_release_regions(pdev);
+	pci_disable_device(pdev);
+}
+
+static int velocity_platform_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *of_id;
+	const struct velocity_info_tbl *info;
+	int irq;
+
+	of_id = of_match_device(velocity_of_ids, &pdev->dev);
+	if (!of_id)
+		return -EINVAL;
+	info = of_id->data;
+
+	irq = irq_of_parse_and_map(pdev->dev.of_node, 0);
+	if (!irq)
+		return -EINVAL;
+
+	return velocity_probe(&pdev->dev, irq, info, BUS_PLATFORM);
+}
+
+static int velocity_platform_remove(struct platform_device *pdev)
+{
+	velocity_remove(&pdev->dev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
 /**
  *	wol_calc_crc		-	WOL CRC
  *	@pattern: data pattern
@@ -3004,32 +3128,35 @@ static void velocity_save_context(struct velocity_info *vptr, struct velocity_co
 
 }
 
-static int velocity_suspend(struct pci_dev *pdev, pm_message_t state)
+static int velocity_suspend(struct device *dev)
 {
-	struct net_device *dev = pci_get_drvdata(pdev);
-	struct velocity_info *vptr = netdev_priv(dev);
+	struct net_device *netdev = dev_get_drvdata(dev);
+	struct velocity_info *vptr = netdev_priv(netdev);
 	unsigned long flags;
 
-	if (!netif_running(vptr->dev))
+	if (!netif_running(vptr->netdev))
 		return 0;
 
-	netif_device_detach(vptr->dev);
+	netif_device_detach(vptr->netdev);
 
 	spin_lock_irqsave(&vptr->lock, flags);
-	pci_save_state(pdev);
+	if (vptr->pdev)
+		pci_save_state(vptr->pdev);
 
 	if (vptr->flags & VELOCITY_FLAGS_WOL_ENABLED) {
 		velocity_get_ip(vptr);
 		velocity_save_context(vptr, &vptr->context);
 		velocity_shutdown(vptr);
 		velocity_set_wol(vptr);
-		pci_enable_wake(pdev, PCI_D3hot, 1);
-		pci_set_power_state(pdev, PCI_D3hot);
+		if (vptr->pdev)
+			pci_enable_wake(vptr->pdev, PCI_D3hot, 1);
+		velocity_set_power_state(vptr, PCI_D3hot);
 	} else {
 		velocity_save_context(vptr, &vptr->context);
 		velocity_shutdown(vptr);
-		pci_disable_device(pdev);
-		pci_set_power_state(pdev, pci_choose_state(pdev, state));
+		if (vptr->pdev)
+			pci_disable_device(vptr->pdev);
+		velocity_set_power_state(vptr, PCI_D3hot);
 	}
 
 	spin_unlock_irqrestore(&vptr->lock, flags);
@@ -3071,19 +3198,22 @@ static void velocity_restore_context(struct velocity_info *vptr, struct velocity
 		writeb(*((u8 *) (context->mac_reg + i)), ptr + i);
 }
 
-static int velocity_resume(struct pci_dev *pdev)
+static int velocity_resume(struct device *dev)
 {
-	struct net_device *dev = pci_get_drvdata(pdev);
-	struct velocity_info *vptr = netdev_priv(dev);
+	struct net_device *netdev = dev_get_drvdata(dev);
+	struct velocity_info *vptr = netdev_priv(netdev);
 	unsigned long flags;
 	int i;
 
-	if (!netif_running(vptr->dev))
+	if (!netif_running(vptr->netdev))
 		return 0;
 
-	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, 0, 0);
-	pci_restore_state(pdev);
+	velocity_set_power_state(vptr, PCI_D0);
+
+	if (vptr->pdev) {
+		pci_enable_wake(vptr->pdev, PCI_D0, 0);
+		pci_restore_state(vptr->pdev);
+	}
 
 	mac_wol_reset(vptr->mac_regs);
 
@@ -3101,27 +3231,38 @@ static int velocity_resume(struct pci_dev *pdev)
 
 	mac_enable_int(vptr->mac_regs);
 	spin_unlock_irqrestore(&vptr->lock, flags);
-	netif_device_attach(vptr->dev);
+	netif_device_attach(vptr->netdev);
 
 	return 0;
 }
-#endif
+#endif	/* CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(velocity_pm_ops, velocity_suspend, velocity_resume);
 
 /*
  *	Definition for our device driver. The PCI layer interface
  *	uses this to handle all our card discover and plugging
  */
-static struct pci_driver velocity_driver = {
+static struct pci_driver velocity_pci_driver = {
 	.name		= VELOCITY_NAME,
-	.id_table	= velocity_id_table,
-	.probe		= velocity_found1,
-	.remove		= velocity_remove1,
-#ifdef CONFIG_PM
-	.suspend	= velocity_suspend,
-	.resume		= velocity_resume,
-#endif
+	.id_table	= velocity_pci_id_table,
+	.probe		= velocity_pci_probe,
+	.remove		= velocity_pci_remove,
+	.driver = {
+		.pm = &velocity_pm_ops,
+	},
 };
 
+static struct platform_driver velocity_platform_driver = {
+	.probe		= velocity_platform_probe,
+	.remove		= velocity_platform_remove,
+	.driver = {
+		.name = "via-velocity",
+		.owner = THIS_MODULE,
+		.of_match_table = velocity_of_ids,
+		.pm = &velocity_pm_ops,
+	},
+};
 
 /**
  *	velocity_ethtool_up	-	pre hook for ethtool
@@ -3134,7 +3275,7 @@ static int velocity_ethtool_up(struct net_device *dev)
 {
 	struct velocity_info *vptr = netdev_priv(dev);
 	if (!netif_running(dev))
-		pci_set_power_state(vptr->pdev, PCI_D0);
+		velocity_set_power_state(vptr, PCI_D0);
 	return 0;
 }
 
@@ -3149,7 +3290,7 @@ static void velocity_ethtool_down(struct net_device *dev)
 {
 	struct velocity_info *vptr = netdev_priv(dev);
 	if (!netif_running(dev))
-		pci_set_power_state(vptr->pdev, PCI_D3hot);
+		velocity_set_power_state(vptr, PCI_D3hot);
 }
 
 static int velocity_get_settings(struct net_device *dev,
@@ -3269,9 +3410,14 @@ static int velocity_set_settings(struct net_device *dev,
 static void velocity_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
 {
 	struct velocity_info *vptr = netdev_priv(dev);
+
 	strlcpy(info->driver, VELOCITY_NAME, sizeof(info->driver));
 	strlcpy(info->version, VELOCITY_VERSION, sizeof(info->version));
-	strlcpy(info->bus_info, pci_name(vptr->pdev), sizeof(info->bus_info));
+	if (vptr->pdev)
+		strlcpy(info->bus_info, pci_name(vptr->pdev),
+						sizeof(info->bus_info));
+	else
+		strlcpy(info->bus_info, "platform", sizeof(info->bus_info));
 }
 
 static void velocity_ethtool_get_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
@@ -3561,13 +3707,20 @@ static void velocity_unregister_notifier(void)
  */
 static int __init velocity_init_module(void)
 {
-	int ret;
+	int ret_pci, ret_platform;
 
 	velocity_register_notifier();
-	ret = pci_register_driver(&velocity_driver);
-	if (ret < 0)
+
+	ret_pci = pci_register_driver(&velocity_pci_driver);
+	ret_platform = platform_driver_register(&velocity_platform_driver);
+
+	/* if both_registers failed, remove the notifier */
+	if ((ret_pci < 0) && (ret_platform < 0)) {
 		velocity_unregister_notifier();
-	return ret;
+		return ret_pci;
+	}
+
+	return 0;
 }
 
 /**
@@ -3581,7 +3734,9 @@ static int __init velocity_init_module(void)
 static void __exit velocity_cleanup_module(void)
 {
 	velocity_unregister_notifier();
-	pci_unregister_driver(&velocity_driver);
+
+	pci_unregister_driver(&velocity_pci_driver);
+	platform_driver_unregister(&velocity_platform_driver);
 }
 
 module_init(velocity_init_module);
diff --git a/drivers/net/ethernet/via/via-velocity.h b/drivers/net/ethernet/via/via-velocity.h
index 4cb9f13..9453bfa 100644
--- a/drivers/net/ethernet/via/via-velocity.h
+++ b/drivers/net/ethernet/via/via-velocity.h
@@ -1265,7 +1265,7 @@ struct velocity_context {
 #define PHYID_VT3216_64BIT  0x000FC600UL
 #define PHYID_MARVELL_1000  0x01410C50UL
 #define PHYID_MARVELL_1000S 0x01410C40UL
-
+#define PHYID_ICPLUS_IP101A 0x02430C54UL
 #define PHYID_REV_ID_MASK   0x0000000FUL
 
 #define PHYID_GET_PHY_ID(i)         ((i) & ~PHYID_REV_ID_MASK)
@@ -1434,8 +1434,10 @@ struct velocity_opt {
 #define GET_RD_BY_IDX(vptr, idx)   (vptr->rd_ring[idx])
 
 struct velocity_info {
+	struct device *dev;
 	struct pci_dev *pdev;
-	struct net_device *dev;
+	struct net_device *netdev;
+	int no_eeprom;
 
 	unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
 	u8 ip_addr[4];
@@ -1514,7 +1516,7 @@ static inline int velocity_get_ip(struct velocity_info *vptr)
 	int res = -ENOENT;
 
 	rcu_read_lock();
-	in_dev = __in_dev_get_rcu(vptr->dev);
+	in_dev = __in_dev_get_rcu(vptr->netdev);
 	if (in_dev != NULL) {
 		ifa = (struct in_ifaddr *) in_dev->ifa_list;
 		if (ifa != NULL) {
diff --git a/drivers/net/ethernet/wiznet/w5100.c b/drivers/net/ethernet/wiznet/w5100.c
index a518dca..30fed08 100644
--- a/drivers/net/ethernet/wiznet/w5100.c
+++ b/drivers/net/ethernet/wiznet/w5100.c
@@ -734,7 +734,6 @@ err_hw_probe:
 	unregister_netdev(ndev);
 err_register:
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 	return err;
 }
 
@@ -750,7 +749,6 @@ static int w5100_remove(struct platform_device *pdev)
 
 	unregister_netdev(ndev);
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/wiznet/w5300.c b/drivers/net/ethernet/wiznet/w5300.c
index 6e00e3f..e928845 100644
--- a/drivers/net/ethernet/wiznet/w5300.c
+++ b/drivers/net/ethernet/wiznet/w5300.c
@@ -646,7 +646,6 @@ err_hw_probe:
 	unregister_netdev(ndev);
 err_register:
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 	return err;
 }
 
@@ -662,7 +661,6 @@ static int w5300_remove(struct platform_device *pdev)
 
 	unregister_netdev(ndev);
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/xilinx/Kconfig b/drivers/net/ethernet/xilinx/Kconfig
index 122d60c..7b90a5e 100644
--- a/drivers/net/ethernet/xilinx/Kconfig
+++ b/drivers/net/ethernet/xilinx/Kconfig
@@ -5,7 +5,7 @@
 config NET_VENDOR_XILINX
 	bool "Xilinx devices"
 	default y
-	depends on PPC || PPC32 || MICROBLAZE
+	depends on PPC || PPC32 || MICROBLAZE || ARCH_ZYNQ
 	---help---
 	  If you have a network (Ethernet) card belonging to this class, say Y
 	  and read the Ethernet-HOWTO, available from
@@ -20,7 +20,7 @@ if NET_VENDOR_XILINX
 
 config XILINX_EMACLITE
 	tristate "Xilinx 10/100 Ethernet Lite support"
-	depends on (PPC32 || MICROBLAZE)
+	depends on (PPC32 || MICROBLAZE || ARCH_ZYNQ)
 	select PHYLIB
 	---help---
 	  This driver supports the 10/100 Ethernet Lite from Xilinx.
diff --git a/drivers/net/ethernet/xilinx/ll_temac_main.c b/drivers/net/ethernet/xilinx/ll_temac_main.c
index 57c2e5e..58eb448 100644
--- a/drivers/net/ethernet/xilinx/ll_temac_main.c
+++ b/drivers/net/ethernet/xilinx/ll_temac_main.c
@@ -1007,7 +1007,7 @@ static int temac_of_probe(struct platform_device *op)
 		return -ENOMEM;
 
 	ether_setup(ndev);
-	dev_set_drvdata(&op->dev, ndev);
+	platform_set_drvdata(op, ndev);
 	SET_NETDEV_DEV(ndev, &op->dev);
 	ndev->flags &= ~IFF_MULTICAST;  /* clear multicast */
 	ndev->features = NETIF_F_SG | NETIF_F_FRAGLIST;
@@ -1136,7 +1136,7 @@ static int temac_of_probe(struct platform_device *op)
 
 static int temac_of_remove(struct platform_device *op)
 {
-	struct net_device *ndev = dev_get_drvdata(&op->dev);
+	struct net_device *ndev = platform_get_drvdata(op);
 	struct temac_local *lp = netdev_priv(ndev);
 
 	temac_mdio_teardown(lp);
@@ -1145,7 +1145,6 @@ static int temac_of_remove(struct platform_device *op)
 	if (lp->phy_node)
 		of_node_put(lp->phy_node);
 	lp->phy_node = NULL;
-	dev_set_drvdata(&op->dev, NULL);
 	iounmap(lp->regs);
 	if (lp->sdma_regs)
 		iounmap(lp->sdma_regs);
diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
index 24748e8..fb7d1c2 100644
--- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
+++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
@@ -1484,7 +1484,7 @@ static int axienet_of_probe(struct platform_device *op)
 		return -ENOMEM;
 
 	ether_setup(ndev);
-	dev_set_drvdata(&op->dev, ndev);
+	platform_set_drvdata(op, ndev);
 
 	SET_NETDEV_DEV(ndev, &op->dev);
 	ndev->flags &= ~IFF_MULTICAST;  /* clear multicast */
@@ -1622,7 +1622,7 @@ nodev:
 
 static int axienet_of_remove(struct platform_device *op)
 {
-	struct net_device *ndev = dev_get_drvdata(&op->dev);
+	struct net_device *ndev = platform_get_drvdata(op);
 	struct axienet_local *lp = netdev_priv(ndev);
 
 	axienet_mdio_teardown(lp);
@@ -1632,8 +1632,6 @@ static int axienet_of_remove(struct platform_device *op)
 		of_node_put(lp->phy_node);
 	lp->phy_node = NULL;
 
-	dev_set_drvdata(&op->dev, NULL);
-
 	iounmap(lp->regs);
 	if (lp->dma_regs)
 		iounmap(lp->dma_regs);
diff --git a/drivers/net/ethernet/xilinx/xilinx_emaclite.c b/drivers/net/ethernet/xilinx/xilinx_emaclite.c
index b7268b3..fd4dbda 100644
--- a/drivers/net/ethernet/xilinx/xilinx_emaclite.c
+++ b/drivers/net/ethernet/xilinx/xilinx_emaclite.c
@@ -2,9 +2,9 @@
  * Xilinx EmacLite Linux driver for the Xilinx Ethernet MAC Lite device.
  *
  * This is a new flat driver which is based on the original emac_lite
- * driver from John Williams <john.williams@petalogix.com>.
+ * driver from John Williams <john.williams@xilinx.com>.
  *
- * 2007-2009 (c) Xilinx, Inc.
+ * 2007 - 2013 (c) Xilinx, Inc.
  *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms of the GNU General Public License as published by the
@@ -159,34 +159,32 @@ static void xemaclite_enable_interrupts(struct net_local *drvdata)
 	u32 reg_data;
 
 	/* Enable the Tx interrupts for the first Buffer */
-	reg_data = in_be32(drvdata->base_addr + XEL_TSR_OFFSET);
-	out_be32(drvdata->base_addr + XEL_TSR_OFFSET,
-		 reg_data | XEL_TSR_XMIT_IE_MASK);
+	reg_data = __raw_readl(drvdata->base_addr + XEL_TSR_OFFSET);
+	__raw_writel(reg_data | XEL_TSR_XMIT_IE_MASK,
+		     drvdata->base_addr + XEL_TSR_OFFSET);
 
 	/* Enable the Tx interrupts for the second Buffer if
 	 * configured in HW */
 	if (drvdata->tx_ping_pong != 0) {
-		reg_data = in_be32(drvdata->base_addr +
+		reg_data = __raw_readl(drvdata->base_addr +
 				   XEL_BUFFER_OFFSET + XEL_TSR_OFFSET);
-		out_be32(drvdata->base_addr + XEL_BUFFER_OFFSET +
-			 XEL_TSR_OFFSET,
-			 reg_data | XEL_TSR_XMIT_IE_MASK);
+		__raw_writel(reg_data | XEL_TSR_XMIT_IE_MASK,
+			     drvdata->base_addr + XEL_BUFFER_OFFSET +
+			     XEL_TSR_OFFSET);
 	}
 
 	/* Enable the Rx interrupts for the first buffer */
-	out_be32(drvdata->base_addr + XEL_RSR_OFFSET,
-		 XEL_RSR_RECV_IE_MASK);
+	__raw_writel(XEL_RSR_RECV_IE_MASK, drvdata->base_addr + XEL_RSR_OFFSET);
 
 	/* Enable the Rx interrupts for the second Buffer if
 	 * configured in HW */
 	if (drvdata->rx_ping_pong != 0) {
-		out_be32(drvdata->base_addr + XEL_BUFFER_OFFSET +
-			 XEL_RSR_OFFSET,
-			 XEL_RSR_RECV_IE_MASK);
+		__raw_writel(XEL_RSR_RECV_IE_MASK, drvdata->base_addr +
+			     XEL_BUFFER_OFFSET + XEL_RSR_OFFSET);
 	}
 
 	/* Enable the Global Interrupt Enable */
-	out_be32(drvdata->base_addr + XEL_GIER_OFFSET, XEL_GIER_GIE_MASK);
+	__raw_writel(XEL_GIER_GIE_MASK, drvdata->base_addr + XEL_GIER_OFFSET);
 }
 
 /**
@@ -201,37 +199,37 @@ static void xemaclite_disable_interrupts(struct net_local *drvdata)
 	u32 reg_data;
 
 	/* Disable the Global Interrupt Enable */
-	out_be32(drvdata->base_addr + XEL_GIER_OFFSET, XEL_GIER_GIE_MASK);
+	__raw_writel(XEL_GIER_GIE_MASK, drvdata->base_addr + XEL_GIER_OFFSET);
 
 	/* Disable the Tx interrupts for the first buffer */
-	reg_data = in_be32(drvdata->base_addr + XEL_TSR_OFFSET);
-	out_be32(drvdata->base_addr + XEL_TSR_OFFSET,
-		 reg_data & (~XEL_TSR_XMIT_IE_MASK));
+	reg_data = __raw_readl(drvdata->base_addr + XEL_TSR_OFFSET);
+	__raw_writel(reg_data & (~XEL_TSR_XMIT_IE_MASK),
+		     drvdata->base_addr + XEL_TSR_OFFSET);
 
 	/* Disable the Tx interrupts for the second Buffer
 	 * if configured in HW */
 	if (drvdata->tx_ping_pong != 0) {
-		reg_data = in_be32(drvdata->base_addr + XEL_BUFFER_OFFSET +
+		reg_data = __raw_readl(drvdata->base_addr + XEL_BUFFER_OFFSET +
 				   XEL_TSR_OFFSET);
-		out_be32(drvdata->base_addr + XEL_BUFFER_OFFSET +
-			 XEL_TSR_OFFSET,
-			 reg_data & (~XEL_TSR_XMIT_IE_MASK));
+		__raw_writel(reg_data & (~XEL_TSR_XMIT_IE_MASK),
+			     drvdata->base_addr + XEL_BUFFER_OFFSET +
+			     XEL_TSR_OFFSET);
 	}
 
 	/* Disable the Rx interrupts for the first buffer */
-	reg_data = in_be32(drvdata->base_addr + XEL_RSR_OFFSET);
-	out_be32(drvdata->base_addr + XEL_RSR_OFFSET,
-		 reg_data & (~XEL_RSR_RECV_IE_MASK));
+	reg_data = __raw_readl(drvdata->base_addr + XEL_RSR_OFFSET);
+	__raw_writel(reg_data & (~XEL_RSR_RECV_IE_MASK),
+		     drvdata->base_addr + XEL_RSR_OFFSET);
 
 	/* Disable the Rx interrupts for the second buffer
 	 * if configured in HW */
 	if (drvdata->rx_ping_pong != 0) {
 
-		reg_data = in_be32(drvdata->base_addr + XEL_BUFFER_OFFSET +
+		reg_data = __raw_readl(drvdata->base_addr + XEL_BUFFER_OFFSET +
 				   XEL_RSR_OFFSET);
-		out_be32(drvdata->base_addr + XEL_BUFFER_OFFSET +
-			 XEL_RSR_OFFSET,
-			 reg_data & (~XEL_RSR_RECV_IE_MASK));
+		__raw_writel(reg_data & (~XEL_RSR_RECV_IE_MASK),
+			     drvdata->base_addr + XEL_BUFFER_OFFSET +
+			     XEL_RSR_OFFSET);
 	}
 }
 
@@ -351,7 +349,7 @@ static int xemaclite_send_data(struct net_local *drvdata, u8 *data,
 		byte_count = ETH_FRAME_LEN;
 
 	/* Check if the expected buffer is available */
-	reg_data = in_be32(addr + XEL_TSR_OFFSET);
+	reg_data = __raw_readl(addr + XEL_TSR_OFFSET);
 	if ((reg_data & (XEL_TSR_XMIT_BUSY_MASK |
 	     XEL_TSR_XMIT_ACTIVE_MASK)) == 0) {
 
@@ -364,7 +362,7 @@ static int xemaclite_send_data(struct net_local *drvdata, u8 *data,
 
 		addr = (void __iomem __force *)((u32 __force)addr ^
 						 XEL_BUFFER_OFFSET);
-		reg_data = in_be32(addr + XEL_TSR_OFFSET);
+		reg_data = __raw_readl(addr + XEL_TSR_OFFSET);
 
 		if ((reg_data & (XEL_TSR_XMIT_BUSY_MASK |
 		     XEL_TSR_XMIT_ACTIVE_MASK)) != 0)
@@ -375,15 +373,16 @@ static int xemaclite_send_data(struct net_local *drvdata, u8 *data,
 	/* Write the frame to the buffer */
 	xemaclite_aligned_write(data, (u32 __force *) addr, byte_count);
 
-	out_be32(addr + XEL_TPLR_OFFSET, (byte_count & XEL_TPLR_LENGTH_MASK));
+	__raw_writel((byte_count & XEL_TPLR_LENGTH_MASK),
+		     addr + XEL_TPLR_OFFSET);
 
 	/* Update the Tx Status Register to indicate that there is a
 	 * frame to send. Set the XEL_TSR_XMIT_ACTIVE_MASK flag which
 	 * is used by the interrupt handler to check whether a frame
 	 * has been transmitted */
-	reg_data = in_be32(addr + XEL_TSR_OFFSET);
+	reg_data = __raw_readl(addr + XEL_TSR_OFFSET);
 	reg_data |= (XEL_TSR_XMIT_BUSY_MASK | XEL_TSR_XMIT_ACTIVE_MASK);
-	out_be32(addr + XEL_TSR_OFFSET, reg_data);
+	__raw_writel(reg_data, addr + XEL_TSR_OFFSET);
 
 	return 0;
 }
@@ -408,7 +407,7 @@ static u16 xemaclite_recv_data(struct net_local *drvdata, u8 *data)
 	addr = (drvdata->base_addr + drvdata->next_rx_buf_to_use);
 
 	/* Verify which buffer has valid data */
-	reg_data = in_be32(addr + XEL_RSR_OFFSET);
+	reg_data = __raw_readl(addr + XEL_RSR_OFFSET);
 
 	if ((reg_data & XEL_RSR_RECV_DONE_MASK) == XEL_RSR_RECV_DONE_MASK) {
 		if (drvdata->rx_ping_pong != 0)
@@ -425,14 +424,14 @@ static u16 xemaclite_recv_data(struct net_local *drvdata, u8 *data)
 			return 0;	/* No data was available */
 
 		/* Verify that buffer has valid data */
-		reg_data = in_be32(addr + XEL_RSR_OFFSET);
+		reg_data = __raw_readl(addr + XEL_RSR_OFFSET);
 		if ((reg_data & XEL_RSR_RECV_DONE_MASK) !=
 		     XEL_RSR_RECV_DONE_MASK)
 			return 0;	/* No data was available */
 	}
 
 	/* Get the protocol type of the ethernet frame that arrived */
-	proto_type = ((ntohl(in_be32(addr + XEL_HEADER_OFFSET +
+	proto_type = ((ntohl(__raw_readl(addr + XEL_HEADER_OFFSET +
 			XEL_RXBUFF_OFFSET)) >> XEL_HEADER_SHIFT) &
 			XEL_RPLR_LENGTH_MASK);
 
@@ -441,7 +440,7 @@ static u16 xemaclite_recv_data(struct net_local *drvdata, u8 *data)
 	if (proto_type > (ETH_FRAME_LEN + ETH_FCS_LEN)) {
 
 		if (proto_type == ETH_P_IP) {
-			length = ((ntohl(in_be32(addr +
+			length = ((ntohl(__raw_readl(addr +
 					XEL_HEADER_IP_LENGTH_OFFSET +
 					XEL_RXBUFF_OFFSET)) >>
 					XEL_HEADER_SHIFT) &
@@ -463,9 +462,9 @@ static u16 xemaclite_recv_data(struct net_local *drvdata, u8 *data)
 				data, length);
 
 	/* Acknowledge the frame */
-	reg_data = in_be32(addr + XEL_RSR_OFFSET);
+	reg_data = __raw_readl(addr + XEL_RSR_OFFSET);
 	reg_data &= ~XEL_RSR_RECV_DONE_MASK;
-	out_be32(addr + XEL_RSR_OFFSET, reg_data);
+	__raw_writel(reg_data, addr + XEL_RSR_OFFSET);
 
 	return length;
 }
@@ -492,14 +491,14 @@ static void xemaclite_update_address(struct net_local *drvdata,
 
 	xemaclite_aligned_write(address_ptr, (u32 __force *) addr, ETH_ALEN);
 
-	out_be32(addr + XEL_TPLR_OFFSET, ETH_ALEN);
+	__raw_writel(ETH_ALEN, addr + XEL_TPLR_OFFSET);
 
 	/* Update the MAC address in the EmacLite */
-	reg_data = in_be32(addr + XEL_TSR_OFFSET);
-	out_be32(addr + XEL_TSR_OFFSET, reg_data | XEL_TSR_PROG_MAC_ADDR);
+	reg_data = __raw_readl(addr + XEL_TSR_OFFSET);
+	__raw_writel(reg_data | XEL_TSR_PROG_MAC_ADDR, addr + XEL_TSR_OFFSET);
 
 	/* Wait for EmacLite to finish with the MAC address update */
-	while ((in_be32(addr + XEL_TSR_OFFSET) &
+	while ((__raw_readl(addr + XEL_TSR_OFFSET) &
 		XEL_TSR_PROG_MAC_ADDR) != 0)
 		;
 }
@@ -669,31 +668,32 @@ static irqreturn_t xemaclite_interrupt(int irq, void *dev_id)
 	u32 tx_status;
 
 	/* Check if there is Rx Data available */
-	if ((in_be32(base_addr + XEL_RSR_OFFSET) & XEL_RSR_RECV_DONE_MASK) ||
-			(in_be32(base_addr + XEL_BUFFER_OFFSET + XEL_RSR_OFFSET)
+	if ((__raw_readl(base_addr + XEL_RSR_OFFSET) &
+			 XEL_RSR_RECV_DONE_MASK) ||
+	    (__raw_readl(base_addr + XEL_BUFFER_OFFSET + XEL_RSR_OFFSET)
 			 & XEL_RSR_RECV_DONE_MASK))
 
 		xemaclite_rx_handler(dev);
 
 	/* Check if the Transmission for the first buffer is completed */
-	tx_status = in_be32(base_addr + XEL_TSR_OFFSET);
+	tx_status = __raw_readl(base_addr + XEL_TSR_OFFSET);
 	if (((tx_status & XEL_TSR_XMIT_BUSY_MASK) == 0) &&
 		(tx_status & XEL_TSR_XMIT_ACTIVE_MASK) != 0) {
 
 		tx_status &= ~XEL_TSR_XMIT_ACTIVE_MASK;
-		out_be32(base_addr + XEL_TSR_OFFSET, tx_status);
+		__raw_writel(tx_status, base_addr + XEL_TSR_OFFSET);
 
 		tx_complete = true;
 	}
 
 	/* Check if the Transmission for the second buffer is completed */
-	tx_status = in_be32(base_addr + XEL_BUFFER_OFFSET + XEL_TSR_OFFSET);
+	tx_status = __raw_readl(base_addr + XEL_BUFFER_OFFSET + XEL_TSR_OFFSET);
 	if (((tx_status & XEL_TSR_XMIT_BUSY_MASK) == 0) &&
 		(tx_status & XEL_TSR_XMIT_ACTIVE_MASK) != 0) {
 
 		tx_status &= ~XEL_TSR_XMIT_ACTIVE_MASK;
-		out_be32(base_addr + XEL_BUFFER_OFFSET + XEL_TSR_OFFSET,
-			 tx_status);
+		__raw_writel(tx_status, base_addr + XEL_BUFFER_OFFSET +
+			     XEL_TSR_OFFSET);
 
 		tx_complete = true;
 	}
@@ -726,7 +726,7 @@ static int xemaclite_mdio_wait(struct net_local *lp)
 	/* wait for the MDIO interface to not be busy or timeout
 	   after some time.
 	*/
-	while (in_be32(lp->base_addr + XEL_MDIOCTRL_OFFSET) &
+	while (__raw_readl(lp->base_addr + XEL_MDIOCTRL_OFFSET) &
 			XEL_MDIOCTRL_MDIOSTS_MASK) {
 		if (end - jiffies <= 0) {
 			WARN_ON(1);
@@ -762,17 +762,17 @@ static int xemaclite_mdio_read(struct mii_bus *bus, int phy_id, int reg)
 	 * MDIO Address register. Set the Status bit in the MDIO Control
 	 * register to start a MDIO read transaction.
 	 */
-	ctrl_reg = in_be32(lp->base_addr + XEL_MDIOCTRL_OFFSET);
-	out_be32(lp->base_addr + XEL_MDIOADDR_OFFSET,
-		 XEL_MDIOADDR_OP_MASK |
-		 ((phy_id << XEL_MDIOADDR_PHYADR_SHIFT) | reg));
-	out_be32(lp->base_addr + XEL_MDIOCTRL_OFFSET,
-		 ctrl_reg | XEL_MDIOCTRL_MDIOSTS_MASK);
+	ctrl_reg = __raw_readl(lp->base_addr + XEL_MDIOCTRL_OFFSET);
+	__raw_writel(XEL_MDIOADDR_OP_MASK |
+		     ((phy_id << XEL_MDIOADDR_PHYADR_SHIFT) | reg),
+		     lp->base_addr + XEL_MDIOADDR_OFFSET);
+	__raw_writel(ctrl_reg | XEL_MDIOCTRL_MDIOSTS_MASK,
+		     lp->base_addr + XEL_MDIOCTRL_OFFSET);
 
 	if (xemaclite_mdio_wait(lp))
 		return -ETIMEDOUT;
 
-	rc = in_be32(lp->base_addr + XEL_MDIORD_OFFSET);
+	rc = __raw_readl(lp->base_addr + XEL_MDIORD_OFFSET);
 
 	dev_dbg(&lp->ndev->dev,
 		"xemaclite_mdio_read(phy_id=%i, reg=%x) == %x\n",
@@ -809,13 +809,13 @@ static int xemaclite_mdio_write(struct mii_bus *bus, int phy_id, int reg,
 	 * Data register. Finally, set the Status bit in the MDIO Control
 	 * register to start a MDIO write transaction.
 	 */
-	ctrl_reg = in_be32(lp->base_addr + XEL_MDIOCTRL_OFFSET);
-	out_be32(lp->base_addr + XEL_MDIOADDR_OFFSET,
-		 ~XEL_MDIOADDR_OP_MASK &
-		 ((phy_id << XEL_MDIOADDR_PHYADR_SHIFT) | reg));
-	out_be32(lp->base_addr + XEL_MDIOWR_OFFSET, val);
-	out_be32(lp->base_addr + XEL_MDIOCTRL_OFFSET,
-		 ctrl_reg | XEL_MDIOCTRL_MDIOSTS_MASK);
+	ctrl_reg = __raw_readl(lp->base_addr + XEL_MDIOCTRL_OFFSET);
+	__raw_writel(~XEL_MDIOADDR_OP_MASK &
+		     ((phy_id << XEL_MDIOADDR_PHYADR_SHIFT) | reg),
+		     lp->base_addr + XEL_MDIOADDR_OFFSET);
+	__raw_writel(val, lp->base_addr + XEL_MDIOWR_OFFSET);
+	__raw_writel(ctrl_reg | XEL_MDIOCTRL_MDIOSTS_MASK,
+		     lp->base_addr + XEL_MDIOCTRL_OFFSET);
 
 	return 0;
 }
@@ -848,24 +848,39 @@ static int xemaclite_mdio_setup(struct net_local *lp, struct device *dev)
 	int rc;
 	struct resource res;
 	struct device_node *np = of_get_parent(lp->phy_node);
+	struct device_node *npp;
 
 	/* Don't register the MDIO bus if the phy_node or its parent node
 	 * can't be found.
 	 */
-	if (!np)
+	if (!np) {
+		dev_err(dev, "Failed to register mdio bus.\n");
 		return -ENODEV;
+	}
+	npp = of_get_parent(np);
+
+	of_address_to_resource(npp, 0, &res);
+	if (lp->ndev->mem_start != res.start) {
+		struct phy_device *phydev;
+		phydev = of_phy_find_device(lp->phy_node);
+		if (!phydev)
+			dev_info(dev,
+				 "MDIO of the phy is not registered yet\n");
+		return 0;
+	}
 
 	/* Enable the MDIO bus by asserting the enable bit in MDIO Control
 	 * register.
 	 */
-	out_be32(lp->base_addr + XEL_MDIOCTRL_OFFSET,
-		 XEL_MDIOCTRL_MDIOEN_MASK);
+	__raw_writel(XEL_MDIOCTRL_MDIOEN_MASK,
+		     lp->base_addr + XEL_MDIOCTRL_OFFSET);
 
 	bus = mdiobus_alloc();
-	if (!bus)
+	if (!bus) {
+		dev_err(dev, "Failed to allocate mdiobus\n");
 		return -ENOMEM;
+	}
 
-	of_address_to_resource(np, 0, &res);
 	snprintf(bus->id, MII_BUS_ID_SIZE, "%.8llx",
 		 (unsigned long long)res.start);
 	bus->priv = lp;
@@ -879,8 +894,10 @@ static int xemaclite_mdio_setup(struct net_local *lp, struct device *dev)
 	lp->mii_bus = bus;
 
 	rc = of_mdiobus_register(bus, np);
-	if (rc)
+	if (rc) {
+		dev_err(dev, "Failed to register mdio bus.\n");
 		goto err_register;
+	}
 
 	return 0;
 
@@ -896,7 +913,7 @@ err_register:
  * There's nothing in the Emaclite device to be configured when the link
  * state changes. We just print the status.
  */
-void xemaclite_adjust_link(struct net_device *ndev)
+static void xemaclite_adjust_link(struct net_device *ndev)
 {
 	struct net_local *lp = netdev_priv(ndev);
 	struct phy_device *phy = lp->phy_dev;
@@ -1058,13 +1075,14 @@ static int xemaclite_send(struct sk_buff *orig_skb, struct net_device *dev)
  * This function un maps the IO region of the Emaclite device and frees the net
  * device.
  */
-static void xemaclite_remove_ndev(struct net_device *ndev)
+static void xemaclite_remove_ndev(struct net_device *ndev,
+				  struct platform_device *pdev)
 {
 	if (ndev) {
 		struct net_local *lp = netdev_priv(ndev);
 
 		if (lp->base_addr)
-			iounmap((void __iomem __force *) (lp->base_addr));
+			devm_iounmap(&pdev->dev, lp->base_addr);
 		free_netdev(ndev);
 	}
 }
@@ -1110,8 +1128,7 @@ static struct net_device_ops xemaclite_netdev_ops;
  */
 static int xemaclite_of_probe(struct platform_device *ofdev)
 {
-	struct resource r_irq; /* Interrupt resources */
-	struct resource r_mem; /* IO mem resources */
+	struct resource *res;
 	struct net_device *ndev = NULL;
 	struct net_local *lp = NULL;
 	struct device *dev = &ofdev->dev;
@@ -1121,20 +1138,6 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 
 	dev_info(dev, "Device Tree Probing\n");
 
-	/* Get iospace for the device */
-	rc = of_address_to_resource(ofdev->dev.of_node, 0, &r_mem);
-	if (rc) {
-		dev_err(dev, "invalid address\n");
-		return rc;
-	}
-
-	/* Get IRQ for the device */
-	rc = of_irq_to_resource(ofdev->dev.of_node, 0, &r_irq);
-	if (!rc) {
-		dev_err(dev, "no IRQ found\n");
-		return rc;
-	}
-
 	/* Create an ethernet device instance */
 	ndev = alloc_etherdev(sizeof(struct net_local));
 	if (!ndev)
@@ -1143,30 +1146,28 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 	dev_set_drvdata(dev, ndev);
 	SET_NETDEV_DEV(ndev, &ofdev->dev);
 
-	ndev->irq = r_irq.start;
-	ndev->mem_start = r_mem.start;
-	ndev->mem_end = r_mem.end;
-
 	lp = netdev_priv(ndev);
 	lp->ndev = ndev;
 
-	if (!request_mem_region(ndev->mem_start,
-				ndev->mem_end - ndev->mem_start + 1,
-				DRIVER_NAME)) {
-		dev_err(dev, "Couldn't lock memory region at %p\n",
-			(void *)ndev->mem_start);
-		rc = -EBUSY;
-		goto error2;
+	/* Get IRQ for the device */
+	res = platform_get_resource(ofdev, IORESOURCE_IRQ, 0);
+	if (!res) {
+		dev_err(dev, "no IRQ found\n");
+		goto error;
 	}
 
-	/* Get the virtual base address for the device */
-	lp->base_addr = ioremap(r_mem.start, resource_size(&r_mem));
-	if (NULL == lp->base_addr) {
-		dev_err(dev, "EmacLite: Could not allocate iomem\n");
-		rc = -EIO;
-		goto error1;
+	ndev->irq = res->start;
+
+	res = platform_get_resource(ofdev, IORESOURCE_MEM, 0);
+	lp->base_addr = devm_ioremap_resource(&ofdev->dev, res);
+	if (IS_ERR(lp->base_addr)) {
+		rc = PTR_ERR(lp->base_addr);
+		goto error;
 	}
 
+	ndev->mem_start = res->start;
+	ndev->mem_end = res->end;
+
 	spin_lock_init(&lp->reset_lock);
 	lp->next_tx_buf_to_use = 0x0;
 	lp->next_rx_buf_to_use = 0x0;
@@ -1181,8 +1182,8 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 		dev_warn(dev, "No MAC address found\n");
 
 	/* Clear the Tx CSR's in case this is a restart */
-	out_be32(lp->base_addr + XEL_TSR_OFFSET, 0);
-	out_be32(lp->base_addr + XEL_BUFFER_OFFSET + XEL_TSR_OFFSET, 0);
+	__raw_writel(0, lp->base_addr + XEL_TSR_OFFSET);
+	__raw_writel(0, lp->base_addr + XEL_BUFFER_OFFSET + XEL_TSR_OFFSET);
 
 	/* Set the MAC address in the EmacLite device */
 	xemaclite_update_address(lp, ndev->dev_addr);
@@ -1203,7 +1204,7 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 	if (rc) {
 		dev_err(dev,
 			"Cannot register network device, aborting\n");
-		goto error1;
+		goto error;
 	}
 
 	dev_info(dev,
@@ -1212,11 +1213,8 @@ static int xemaclite_of_probe(struct platform_device *ofdev)
 		 (unsigned int __force)lp->base_addr, ndev->irq);
 	return 0;
 
-error1:
-	release_mem_region(ndev->mem_start, resource_size(&r_mem));
-
-error2:
-	xemaclite_remove_ndev(ndev);
+error:
+	xemaclite_remove_ndev(ndev, ofdev);
 	return rc;
 }
 
@@ -1251,9 +1249,7 @@ static int xemaclite_of_remove(struct platform_device *of_dev)
 		of_node_put(lp->phy_node);
 	lp->phy_node = NULL;
 
-	release_mem_region(ndev->mem_start, ndev->mem_end-ndev->mem_start + 1);
-
-	xemaclite_remove_ndev(ndev);
+	xemaclite_remove_ndev(ndev, of_dev);
 	dev_set_drvdata(dev, NULL);
 
 	return 0;
diff --git a/drivers/net/ethernet/xscale/ixp4xx_eth.c b/drivers/net/ethernet/xscale/ixp4xx_eth.c
index 6958a5e..3d689fc 100644
--- a/drivers/net/ethernet/xscale/ixp4xx_eth.c
+++ b/drivers/net/ethernet/xscale/ixp4xx_eth.c
@@ -1472,7 +1472,6 @@ err_phy_dis:
 	phy_disconnect(port->phydev);
 err_free_mem:
 	npe_port_tab[NPE_ID(port->id)] = NULL;
-	platform_set_drvdata(pdev, NULL);
 	release_resource(port->mem_res);
 err_npe_rel:
 	npe_release(port->npe);
@@ -1489,7 +1488,6 @@ static int eth_remove_one(struct platform_device *pdev)
 	unregister_netdev(dev);
 	phy_disconnect(port->phydev);
 	npe_port_tab[NPE_ID(port->id)] = NULL;
-	platform_set_drvdata(pdev, NULL);
 	npe_release(port->npe);
 	release_resource(port->mem_res);
 	free_netdev(dev);
diff --git a/drivers/net/fddi/skfp/skfddi.c b/drivers/net/fddi/skfp/skfddi.c
index d5bd563..f5d7305 100644
--- a/drivers/net/fddi/skfp/skfddi.c
+++ b/drivers/net/fddi/skfp/skfddi.c
@@ -2246,15 +2246,4 @@ static struct pci_driver skfddi_pci_driver = {
 	.remove		= skfp_remove_one,
 };
 
-static int __init skfd_init(void)
-{
-	return pci_register_driver(&skfddi_pci_driver);
-}
-
-static void __exit skfd_exit(void)
-{
-	pci_unregister_driver(&skfddi_pci_driver);
-}
-
-module_init(skfd_init);
-module_exit(skfd_exit);
+module_pci_driver(skfddi_pci_driver);
diff --git a/drivers/net/hamradio/bpqether.c b/drivers/net/hamradio/bpqether.c
index 02de6c8..f91bf0d 100644
--- a/drivers/net/hamradio/bpqether.c
+++ b/drivers/net/hamradio/bpqether.c
@@ -103,7 +103,7 @@ static struct packet_type bpq_packet_type __read_mostly = {
 };
 
 static struct notifier_block bpq_dev_notifier = {
-	.notifier_call =bpq_device_event,
+	.notifier_call = bpq_device_event,
 };
 
 
@@ -544,9 +544,10 @@ static void bpq_free_device(struct net_device *ndev)
 /*
  *	Handle device status changes.
  */
-static int bpq_device_event(struct notifier_block *this,unsigned long event, void *ptr)
+static int bpq_device_event(struct notifier_block *this,
+			    unsigned long event, void *ptr)
 {
-	struct net_device *dev = (struct net_device *)ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 
 	if (!net_eq(dev_net(dev), &init_net))
 		return NOTIFY_DONE;
diff --git a/drivers/net/hippi/rrunner.c b/drivers/net/hippi/rrunner.c
index 3c4d627..00ed751 100644
--- a/drivers/net/hippi/rrunner.c
+++ b/drivers/net/hippi/rrunner.c
@@ -1686,15 +1686,4 @@ static struct pci_driver rr_driver = {
 	.remove		= rr_remove_one,
 };
 
-static int __init rr_init_module(void)
-{
-	return pci_register_driver(&rr_driver);
-}
-
-static void __exit rr_cleanup_module(void)
-{
-	pci_unregister_driver(&rr_driver);
-}
-
-module_init(rr_init_module);
-module_exit(rr_cleanup_module);
+module_pci_driver(rr_driver);
diff --git a/drivers/net/irda/bfin_sir.c b/drivers/net/irda/bfin_sir.c
index 22b4527..c74f384 100644
--- a/drivers/net/irda/bfin_sir.c
+++ b/drivers/net/irda/bfin_sir.c
@@ -794,7 +794,6 @@ static int bfin_sir_remove(struct platform_device *pdev)
 	kfree(self->rx_buff.head);
 	free_netdev(dev);
 	kfree(sir_port);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/irda/sh_irda.c b/drivers/net/irda/sh_irda.c
index 9448587..4455425 100644
--- a/drivers/net/irda/sh_irda.c
+++ b/drivers/net/irda/sh_irda.c
@@ -838,7 +838,6 @@ static int sh_irda_remove(struct platform_device *pdev)
 	sh_irda_remove_iobuf(self);
 	iounmap(self->membase);
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/irda/sh_sir.c b/drivers/net/irda/sh_sir.c
index 24aefcd..89682b4 100644
--- a/drivers/net/irda/sh_sir.c
+++ b/drivers/net/irda/sh_sir.c
@@ -796,7 +796,6 @@ static int sh_sir_remove(struct platform_device *pdev)
 	sh_sir_remove_iobuf(self);
 	iounmap(self->membase);
 	free_netdev(ndev);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index 6e91931..18373b6 100644
--- a/drivers/net/macvlan.c
+++ b/drivers/net/macvlan.c
@@ -638,6 +638,14 @@ static int macvlan_ethtool_get_settings(struct net_device *dev,
 	return __ethtool_get_settings(vlan->lowerdev, cmd);
 }
 
+static netdev_features_t macvlan_fix_features(struct net_device *dev,
+					      netdev_features_t features)
+{
+	struct macvlan_dev *vlan = netdev_priv(dev);
+
+	return features & (vlan->set_features | ~MACVLAN_FEATURES);
+}
+
 static const struct ethtool_ops macvlan_ethtool_ops = {
 	.get_link		= ethtool_op_get_link,
 	.get_settings		= macvlan_ethtool_get_settings,
@@ -651,6 +659,7 @@ static const struct net_device_ops macvlan_netdev_ops = {
 	.ndo_stop		= macvlan_stop,
 	.ndo_start_xmit		= macvlan_start_xmit,
 	.ndo_change_mtu		= macvlan_change_mtu,
+	.ndo_fix_features	= macvlan_fix_features,
 	.ndo_change_rx_flags	= macvlan_change_rx_flags,
 	.ndo_set_mac_address	= macvlan_set_mac_address,
 	.ndo_set_rx_mode	= macvlan_set_mac_lists,
@@ -791,6 +800,7 @@ int macvlan_common_newlink(struct net *src_net, struct net_device *dev,
 	vlan->port     = port;
 	vlan->receive  = receive;
 	vlan->forward  = forward;
+	vlan->set_features = MACVLAN_FEATURES;
 
 	vlan->mode     = MACVLAN_MODE_VEPA;
 	if (data && data[IFLA_MACVLAN_MODE])
@@ -927,7 +937,7 @@ static struct rtnl_link_ops macvlan_link_ops = {
 static int macvlan_device_event(struct notifier_block *unused,
 				unsigned long event, void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct macvlan_dev *vlan, *next;
 	struct macvlan_port *port;
 	LIST_HEAD(list_kill);
diff --git a/drivers/net/macvtap.c b/drivers/net/macvtap.c
index b6dd6a7..f2c4a3b 100644
--- a/drivers/net/macvtap.c
+++ b/drivers/net/macvtap.c
@@ -31,10 +31,6 @@
  * macvtap_proto is used to allocate queues through the sock allocation
  * mechanism.
  *
- * TODO: multiqueue support is currently not implemented, even though
- * macvtap is basically prepared for that. We will need to add this
- * here as well as in virtio-net and qemu to get line rate on 10gbit
- * adapters from a guest.
  */
 struct macvtap_queue {
 	struct sock sk;
@@ -44,6 +40,9 @@ struct macvtap_queue {
 	struct macvlan_dev __rcu *vlan;
 	struct file *file;
 	unsigned int flags;
+	u16 queue_index;
+	bool enabled;
+	struct list_head next;
 };
 
 static struct proto macvtap_proto = {
@@ -66,11 +65,14 @@ static struct cdev macvtap_cdev;
 
 static const struct proto_ops macvtap_socket_ops;
 
+#define TUN_OFFLOADS (NETIF_F_HW_CSUM | NETIF_F_TSO_ECN | NETIF_F_TSO | \
+		      NETIF_F_TSO6 | NETIF_F_UFO)
+#define RX_OFFLOADS (NETIF_F_GRO | NETIF_F_LRO)
 /*
  * RCU usage:
  * The macvtap_queue and the macvlan_dev are loosely coupled, the
  * pointers from one to the other can only be read while rcu_read_lock
- * or macvtap_lock is held.
+ * or rtnl is held.
  *
  * Both the file and the macvlan_dev hold a reference on the macvtap_queue
  * through sock_hold(&q->sk). When the macvlan_dev goes away first,
@@ -82,54 +84,84 @@ static const struct proto_ops macvtap_socket_ops;
  * file or the dev. The data structure is freed through __sk_free
  * when both our references and any pending SKBs are gone.
  */
-static DEFINE_SPINLOCK(macvtap_lock);
 
-/*
- * get_slot: return a [unused/occupied] slot in vlan->taps[]:
- *	- if 'q' is NULL, return the first empty slot;
- *	- otherwise, return the slot this pointer occupies.
- */
-static int get_slot(struct macvlan_dev *vlan, struct macvtap_queue *q)
+static int macvtap_enable_queue(struct net_device *dev, struct file *file,
+				struct macvtap_queue *q)
 {
-	int i;
+	struct macvlan_dev *vlan = netdev_priv(dev);
+	int err = -EINVAL;
 
-	for (i = 0; i < MAX_MACVTAP_QUEUES; i++) {
-		if (rcu_dereference_protected(vlan->taps[i],
-					      lockdep_is_held(&macvtap_lock)) == q)
-			return i;
-	}
+	ASSERT_RTNL();
+
+	if (q->enabled)
+		goto out;
 
-	/* Should never happen */
-	BUG_ON(1);
+	err = 0;
+	rcu_assign_pointer(vlan->taps[vlan->numvtaps], q);
+	q->queue_index = vlan->numvtaps;
+	q->enabled = true;
+
+	vlan->numvtaps++;
+out:
+	return err;
 }
 
 static int macvtap_set_queue(struct net_device *dev, struct file *file,
-				struct macvtap_queue *q)
+			     struct macvtap_queue *q)
 {
 	struct macvlan_dev *vlan = netdev_priv(dev);
-	int index;
 	int err = -EBUSY;
 
-	spin_lock(&macvtap_lock);
-	if (vlan->numvtaps == MAX_MACVTAP_QUEUES)
+	rtnl_lock();
+	if (vlan->numqueues == MAX_MACVTAP_QUEUES)
 		goto out;
 
 	err = 0;
-	index = get_slot(vlan, NULL);
 	rcu_assign_pointer(q->vlan, vlan);
-	rcu_assign_pointer(vlan->taps[index], q);
+	rcu_assign_pointer(vlan->taps[vlan->numvtaps], q);
 	sock_hold(&q->sk);
 
 	q->file = file;
+	q->queue_index = vlan->numvtaps;
+	q->enabled = true;
 	file->private_data = q;
+	list_add_tail(&q->next, &vlan->queue_list);
 
 	vlan->numvtaps++;
+	vlan->numqueues++;
 
 out:
-	spin_unlock(&macvtap_lock);
+	rtnl_unlock();
 	return err;
 }
 
+static int macvtap_disable_queue(struct macvtap_queue *q)
+{
+	struct macvlan_dev *vlan;
+	struct macvtap_queue *nq;
+
+	ASSERT_RTNL();
+	if (!q->enabled)
+		return -EINVAL;
+
+	vlan = rtnl_dereference(q->vlan);
+
+	if (vlan) {
+		int index = q->queue_index;
+		BUG_ON(index >= vlan->numvtaps);
+		nq = rtnl_dereference(vlan->taps[vlan->numvtaps - 1]);
+		nq->queue_index = index;
+
+		rcu_assign_pointer(vlan->taps[index], nq);
+		RCU_INIT_POINTER(vlan->taps[vlan->numvtaps - 1], NULL);
+		q->enabled = false;
+
+		vlan->numvtaps--;
+	}
+
+	return 0;
+}
+
 /*
  * The file owning the queue got closed, give up both
  * the reference that the files holds as well as the
@@ -142,19 +174,20 @@ static void macvtap_put_queue(struct macvtap_queue *q)
 {
 	struct macvlan_dev *vlan;
 
-	spin_lock(&macvtap_lock);
-	vlan = rcu_dereference_protected(q->vlan,
-					 lockdep_is_held(&macvtap_lock));
+	rtnl_lock();
+	vlan = rtnl_dereference(q->vlan);
+
 	if (vlan) {
-		int index = get_slot(vlan, q);
+		if (q->enabled)
+			BUG_ON(macvtap_disable_queue(q));
 
-		RCU_INIT_POINTER(vlan->taps[index], NULL);
+		vlan->numqueues--;
 		RCU_INIT_POINTER(q->vlan, NULL);
 		sock_put(&q->sk);
-		--vlan->numvtaps;
+		list_del_init(&q->next);
 	}
 
-	spin_unlock(&macvtap_lock);
+	rtnl_unlock();
 
 	synchronize_rcu();
 	sock_put(&q->sk);
@@ -172,7 +205,12 @@ static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
 {
 	struct macvlan_dev *vlan = netdev_priv(dev);
 	struct macvtap_queue *tap = NULL;
-	int numvtaps = vlan->numvtaps;
+	/* Access to taps array is protected by rcu, but access to numvtaps
+	 * isn't. Below we use it to lookup a queue, but treat it as a hint
+	 * and validate that the result isn't NULL - in case we are
+	 * racing against queue removal.
+	 */
+	int numvtaps = ACCESS_ONCE(vlan->numvtaps);
 	__u32 rxq;
 
 	if (!numvtaps)
@@ -182,8 +220,7 @@ static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
 	rxq = skb_get_rxhash(skb);
 	if (rxq) {
 		tap = rcu_dereference(vlan->taps[rxq % numvtaps]);
-		if (tap)
-			goto out;
+		goto out;
 	}
 
 	if (likely(skb_rx_queue_recorded(skb))) {
@@ -193,17 +230,10 @@ static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
 			rxq -= numvtaps;
 
 		tap = rcu_dereference(vlan->taps[rxq]);
-		if (tap)
-			goto out;
-	}
-
-	/* Everything failed - find first available queue */
-	for (rxq = 0; rxq < MAX_MACVTAP_QUEUES; rxq++) {
-		tap = rcu_dereference(vlan->taps[rxq]);
-		if (tap)
-			break;
+		goto out;
 	}
 
+	tap = rcu_dereference(vlan->taps[0]);
 out:
 	return tap;
 }
@@ -216,27 +246,24 @@ out:
 static void macvtap_del_queues(struct net_device *dev)
 {
 	struct macvlan_dev *vlan = netdev_priv(dev);
-	struct macvtap_queue *q, *qlist[MAX_MACVTAP_QUEUES];
+	struct macvtap_queue *q, *tmp, *qlist[MAX_MACVTAP_QUEUES];
 	int i, j = 0;
 
-	/* macvtap_put_queue can free some slots, so go through all slots */
-	spin_lock(&macvtap_lock);
-	for (i = 0; i < MAX_MACVTAP_QUEUES && vlan->numvtaps; i++) {
-		q = rcu_dereference_protected(vlan->taps[i],
-					      lockdep_is_held(&macvtap_lock));
-		if (q) {
-			qlist[j++] = q;
-			RCU_INIT_POINTER(vlan->taps[i], NULL);
-			RCU_INIT_POINTER(q->vlan, NULL);
+	ASSERT_RTNL();
+	list_for_each_entry_safe(q, tmp, &vlan->queue_list, next) {
+		list_del_init(&q->next);
+		qlist[j++] = q;
+		RCU_INIT_POINTER(q->vlan, NULL);
+		if (q->enabled)
 			vlan->numvtaps--;
-		}
+		vlan->numqueues--;
 	}
-	BUG_ON(vlan->numvtaps != 0);
+	for (i = 0; i < vlan->numvtaps; i++)
+		RCU_INIT_POINTER(vlan->taps[i], NULL);
+	BUG_ON(vlan->numvtaps);
+	BUG_ON(vlan->numqueues);
 	/* guarantee that any future macvtap_set_queue will fail */
 	vlan->numvtaps = MAX_MACVTAP_QUEUES;
-	spin_unlock(&macvtap_lock);
-
-	synchronize_rcu();
 
 	for (--j; j >= 0; j--)
 		sock_put(&qlist[j]->sk);
@@ -249,14 +276,44 @@ static void macvtap_del_queues(struct net_device *dev)
  */
 static int macvtap_forward(struct net_device *dev, struct sk_buff *skb)
 {
+	struct macvlan_dev *vlan = netdev_priv(dev);
 	struct macvtap_queue *q = macvtap_get_queue(dev, skb);
+	netdev_features_t features;
 	if (!q)
 		goto drop;
 
 	if (skb_queue_len(&q->sk.sk_receive_queue) >= dev->tx_queue_len)
 		goto drop;
 
-	skb_queue_tail(&q->sk.sk_receive_queue, skb);
+	skb->dev = dev;
+	/* Apply the forward feature mask so that we perform segmentation
+	 * according to users wishes.
+	 */
+	features = netif_skb_features(skb) & vlan->tap_features;
+	if (netif_needs_gso(skb, features)) {
+		struct sk_buff *segs = __skb_gso_segment(skb, features, false);
+
+		if (IS_ERR(segs))
+			goto drop;
+
+		if (!segs) {
+			skb_queue_tail(&q->sk.sk_receive_queue, skb);
+			goto wake_up;
+		}
+
+		kfree_skb(skb);
+		while (segs) {
+			struct sk_buff *nskb = segs->next;
+
+			segs->next = NULL;
+			skb_queue_tail(&q->sk.sk_receive_queue, segs);
+			segs = nskb;
+		}
+	} else {
+		skb_queue_tail(&q->sk.sk_receive_queue, skb);
+	}
+
+wake_up:
 	wake_up_interruptible_poll(sk_sleep(&q->sk), POLLIN | POLLRDNORM | POLLRDBAND);
 	return NET_RX_SUCCESS;
 
@@ -322,6 +379,14 @@ static int macvtap_newlink(struct net *src_net,
 			   struct nlattr *tb[],
 			   struct nlattr *data[])
 {
+	struct macvlan_dev *vlan = netdev_priv(dev);
+	INIT_LIST_HEAD(&vlan->queue_list);
+
+	/* Since macvlan supports all offloads by default, make
+	 * tap support all offloads also.
+	 */
+	vlan->tap_features = TUN_OFFLOADS;
+
 	/* Don't put anything that may fail after macvlan_common_newlink
 	 * because we can't undo what it does.
 	 */
@@ -385,7 +450,7 @@ static int macvtap_open(struct inode *inode, struct file *file)
 	if (!q)
 		goto out;
 
-	q->sock.wq = &q->wq;
+	RCU_INIT_POINTER(q->sock.wq, &q->wq);
 	init_waitqueue_head(&q->wq.wait);
 	q->sock.type = SOCK_RAW;
 	q->sock.state = SS_CONNECTED;
@@ -729,8 +794,8 @@ static ssize_t macvtap_get_user(struct macvtap_queue *q, struct msghdr *m,
 
 	skb_probe_transport_header(skb, ETH_HLEN);
 
-	rcu_read_lock_bh();
-	vlan = rcu_dereference_bh(q->vlan);
+	rcu_read_lock();
+	vlan = rcu_dereference(q->vlan);
 	/* copy skb_ubuf_info for callback when skb has no error */
 	if (zerocopy) {
 		skb_shinfo(skb)->destructor_arg = m->msg_control;
@@ -741,7 +806,7 @@ static ssize_t macvtap_get_user(struct macvtap_queue *q, struct msghdr *m,
 		macvlan_start_xmit(skb, vlan->dev);
 	else
 		kfree_skb(skb);
-	rcu_read_unlock_bh();
+	rcu_read_unlock();
 
 	return total_len;
 
@@ -749,11 +814,11 @@ err_kfree:
 	kfree_skb(skb);
 
 err:
-	rcu_read_lock_bh();
-	vlan = rcu_dereference_bh(q->vlan);
+	rcu_read_lock();
+	vlan = rcu_dereference(q->vlan);
 	if (vlan)
 		vlan->dev->stats.tx_dropped++;
-	rcu_read_unlock_bh();
+	rcu_read_unlock();
 
 	return err;
 }
@@ -829,11 +894,11 @@ static ssize_t macvtap_put_user(struct macvtap_queue *q,
 	copied += len;
 
 done:
-	rcu_read_lock_bh();
-	vlan = rcu_dereference_bh(q->vlan);
+	rcu_read_lock();
+	vlan = rcu_dereference(q->vlan);
 	if (vlan)
 		macvlan_count_rx(vlan, copied - vnet_hdr_len, ret == 0, 0);
-	rcu_read_unlock_bh();
+	rcu_read_unlock();
 
 	return ret ? ret : copied;
 }
@@ -847,7 +912,9 @@ static ssize_t macvtap_do_read(struct macvtap_queue *q, struct kiocb *iocb,
 	ssize_t ret = 0;
 
 	while (len) {
-		prepare_to_wait(sk_sleep(&q->sk), &wait, TASK_INTERRUPTIBLE);
+		if (!noblock)
+			prepare_to_wait(sk_sleep(&q->sk), &wait,
+					TASK_INTERRUPTIBLE);
 
 		/* Read frames from the queue */
 		skb = skb_dequeue(&q->sk.sk_receive_queue);
@@ -869,7 +936,8 @@ static ssize_t macvtap_do_read(struct macvtap_queue *q, struct kiocb *iocb,
 		break;
 	}
 
-	finish_wait(sk_sleep(&q->sk), &wait);
+	if (!noblock)
+		finish_wait(sk_sleep(&q->sk), &wait);
 	return ret;
 }
 
@@ -892,6 +960,96 @@ out:
 	return ret;
 }
 
+static struct macvlan_dev *macvtap_get_vlan(struct macvtap_queue *q)
+{
+	struct macvlan_dev *vlan;
+
+	ASSERT_RTNL();
+	vlan = rtnl_dereference(q->vlan);
+	if (vlan)
+		dev_hold(vlan->dev);
+
+	return vlan;
+}
+
+static void macvtap_put_vlan(struct macvlan_dev *vlan)
+{
+	dev_put(vlan->dev);
+}
+
+static int macvtap_ioctl_set_queue(struct file *file, unsigned int flags)
+{
+	struct macvtap_queue *q = file->private_data;
+	struct macvlan_dev *vlan;
+	int ret;
+
+	vlan = macvtap_get_vlan(q);
+	if (!vlan)
+		return -EINVAL;
+
+	if (flags & IFF_ATTACH_QUEUE)
+		ret = macvtap_enable_queue(vlan->dev, file, q);
+	else if (flags & IFF_DETACH_QUEUE)
+		ret = macvtap_disable_queue(q);
+	else
+		ret = -EINVAL;
+
+	macvtap_put_vlan(vlan);
+	return ret;
+}
+
+static int set_offload(struct macvtap_queue *q, unsigned long arg)
+{
+	struct macvlan_dev *vlan;
+	netdev_features_t features;
+	netdev_features_t feature_mask = 0;
+
+	vlan = rtnl_dereference(q->vlan);
+	if (!vlan)
+		return -ENOLINK;
+
+	features = vlan->dev->features;
+
+	if (arg & TUN_F_CSUM) {
+		feature_mask = NETIF_F_HW_CSUM;
+
+		if (arg & (TUN_F_TSO4 | TUN_F_TSO6)) {
+			if (arg & TUN_F_TSO_ECN)
+				feature_mask |= NETIF_F_TSO_ECN;
+			if (arg & TUN_F_TSO4)
+				feature_mask |= NETIF_F_TSO;
+			if (arg & TUN_F_TSO6)
+				feature_mask |= NETIF_F_TSO6;
+		}
+
+		if (arg & TUN_F_UFO)
+			feature_mask |= NETIF_F_UFO;
+	}
+
+	/* tun/tap driver inverts the usage for TSO offloads, where
+	 * setting the TSO bit means that the userspace wants to
+	 * accept TSO frames and turning it off means that user space
+	 * does not support TSO.
+	 * For macvtap, we have to invert it to mean the same thing.
+	 * When user space turns off TSO, we turn off GSO/LRO so that
+	 * user-space will not receive TSO frames.
+	 */
+	if (feature_mask & (NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_UFO))
+		features |= RX_OFFLOADS;
+	else
+		features &= ~RX_OFFLOADS;
+
+	/* tap_features are the same as features on tun/tap and
+	 * reflect user expectations.
+	 */
+	vlan->tap_features = vlan->dev->features &
+			    (feature_mask | ~TUN_OFFLOADS);
+	vlan->set_features = features;
+	netdev_update_features(vlan->dev);
+
+	return 0;
+}
+
 /*
  * provide compatibility with generic tun/tap interface
  */
@@ -915,7 +1073,8 @@ static long macvtap_ioctl(struct file *file, unsigned int cmd,
 			return -EFAULT;
 
 		ret = 0;
-		if ((u & ~IFF_VNET_HDR) != (IFF_NO_PI | IFF_TAP))
+		if ((u & ~(IFF_VNET_HDR | IFF_MULTI_QUEUE)) !=
+		    (IFF_NO_PI | IFF_TAP))
 			ret = -EINVAL;
 		else
 			q->flags = u;
@@ -923,24 +1082,31 @@ static long macvtap_ioctl(struct file *file, unsigned int cmd,
 		return ret;
 
 	case TUNGETIFF:
-		rcu_read_lock_bh();
-		vlan = rcu_dereference_bh(q->vlan);
-		if (vlan)
-			dev_hold(vlan->dev);
-		rcu_read_unlock_bh();
-
-		if (!vlan)
+		rtnl_lock();
+		vlan = macvtap_get_vlan(q);
+		if (!vlan) {
+			rtnl_unlock();
 			return -ENOLINK;
+		}
 
 		ret = 0;
 		if (copy_to_user(&ifr->ifr_name, vlan->dev->name, IFNAMSIZ) ||
 		    put_user(q->flags, &ifr->ifr_flags))
 			ret = -EFAULT;
-		dev_put(vlan->dev);
+		macvtap_put_vlan(vlan);
+		rtnl_unlock();
 		return ret;
 
+	case TUNSETQUEUE:
+		if (get_user(u, &ifr->ifr_flags))
+			return -EFAULT;
+		rtnl_lock();
+		ret = macvtap_ioctl_set_queue(file, u);
+		rtnl_unlock();
+
 	case TUNGETFEATURES:
-		if (put_user(IFF_TAP | IFF_NO_PI | IFF_VNET_HDR, up))
+		if (put_user(IFF_TAP | IFF_NO_PI | IFF_VNET_HDR |
+			     IFF_MULTI_QUEUE, up))
 			return -EFAULT;
 		return 0;
 
@@ -976,7 +1142,10 @@ static long macvtap_ioctl(struct file *file, unsigned int cmd,
 			 got enabled for forwarded frames */
 		if (!(q->flags & IFF_VNET_HDR))
 			return  -EINVAL;
-		return 0;
+		rtnl_lock();
+		ret = set_offload(q, arg);
+		rtnl_unlock();
+		return ret;
 
 	default:
 		return -EINVAL;
@@ -1055,7 +1224,7 @@ EXPORT_SYMBOL_GPL(macvtap_get_socket);
 static int macvtap_device_event(struct notifier_block *unused,
 				unsigned long event, void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct macvlan_dev *vlan;
 	struct device *classdev;
 	dev_t devt;
diff --git a/drivers/net/netconsole.c b/drivers/net/netconsole.c
index 4f777ed9..4822aaf 100644
--- a/drivers/net/netconsole.c
+++ b/drivers/net/netconsole.c
@@ -654,12 +654,11 @@ static struct configfs_subsystem netconsole_subsys = {
 
 /* Handle network interface device notifications */
 static int netconsole_netdev_event(struct notifier_block *this,
-				   unsigned long event,
-				   void *ptr)
+				   unsigned long event, void *ptr)
 {
 	unsigned long flags;
 	struct netconsole_target *nt;
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	bool stopped = false;
 
 	if (!(event == NETDEV_CHANGENAME || event == NETDEV_UNREGISTER ||
diff --git a/drivers/net/nlmon.c b/drivers/net/nlmon.c
new file mode 100644
index 0000000..b57ce5f
--- /dev/null
+++ b/drivers/net/nlmon.c
@@ -0,0 +1,181 @@
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/netdevice.h>
+#include <linux/netlink.h>
+#include <net/net_namespace.h>
+#include <linux/if_arp.h>
+#include <net/rtnetlink.h>
+
+struct pcpu_lstats {
+	u64 packets;
+	u64 bytes;
+	struct u64_stats_sync syncp;
+};
+
+static netdev_tx_t nlmon_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	int len = skb->len;
+	struct pcpu_lstats *stats = this_cpu_ptr(dev->lstats);
+
+	u64_stats_update_begin(&stats->syncp);
+	stats->bytes += len;
+	stats->packets++;
+	u64_stats_update_end(&stats->syncp);
+
+	dev_kfree_skb(skb);
+
+	return NETDEV_TX_OK;
+}
+
+static int nlmon_is_valid_mtu(int new_mtu)
+{
+	/* Note that in netlink we do not really have an upper limit. On
+	 * default, we use NLMSG_GOODSIZE. Here at least we should make
+	 * sure that it's at least the header size.
+	 */
+	return new_mtu >= (int) sizeof(struct nlmsghdr);
+}
+
+static int nlmon_change_mtu(struct net_device *dev, int new_mtu)
+{
+	if (!nlmon_is_valid_mtu(new_mtu))
+		return -EINVAL;
+
+	dev->mtu = new_mtu;
+	return 0;
+}
+
+static int nlmon_dev_init(struct net_device *dev)
+{
+	dev->lstats = alloc_percpu(struct pcpu_lstats);
+
+	return dev->lstats == NULL ? -ENOMEM : 0;
+}
+
+static void nlmon_dev_uninit(struct net_device *dev)
+{
+	free_percpu(dev->lstats);
+}
+
+struct nlmon {
+	struct netlink_tap nt;
+};
+
+static int nlmon_open(struct net_device *dev)
+{
+	struct nlmon *nlmon = netdev_priv(dev);
+
+	nlmon->nt.dev = dev;
+	nlmon->nt.module = THIS_MODULE;
+	return netlink_add_tap(&nlmon->nt);
+}
+
+static int nlmon_close(struct net_device *dev)
+{
+	struct nlmon *nlmon = netdev_priv(dev);
+
+	return netlink_remove_tap(&nlmon->nt);
+}
+
+static struct rtnl_link_stats64 *
+nlmon_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+	int i;
+	u64 bytes = 0, packets = 0;
+
+	for_each_possible_cpu(i) {
+		const struct pcpu_lstats *nl_stats;
+		u64 tbytes, tpackets;
+		unsigned int start;
+
+		nl_stats = per_cpu_ptr(dev->lstats, i);
+
+		do {
+			start = u64_stats_fetch_begin_bh(&nl_stats->syncp);
+			tbytes = nl_stats->bytes;
+			tpackets = nl_stats->packets;
+		} while (u64_stats_fetch_retry_bh(&nl_stats->syncp, start));
+
+		packets += tpackets;
+		bytes += tbytes;
+	}
+
+	stats->rx_packets = packets;
+	stats->tx_packets = 0;
+
+	stats->rx_bytes = bytes;
+	stats->tx_bytes = 0;
+
+	return stats;
+}
+
+static u32 always_on(struct net_device *dev)
+{
+	return 1;
+}
+
+static const struct ethtool_ops nlmon_ethtool_ops = {
+	.get_link = always_on,
+};
+
+static const struct net_device_ops nlmon_ops = {
+	.ndo_init = nlmon_dev_init,
+	.ndo_uninit = nlmon_dev_uninit,
+	.ndo_open = nlmon_open,
+	.ndo_stop = nlmon_close,
+	.ndo_start_xmit = nlmon_xmit,
+	.ndo_get_stats64 = nlmon_get_stats64,
+	.ndo_change_mtu = nlmon_change_mtu,
+};
+
+static void nlmon_setup(struct net_device *dev)
+{
+	dev->type = ARPHRD_NETLINK;
+	dev->tx_queue_len = 0;
+
+	dev->netdev_ops	= &nlmon_ops;
+	dev->ethtool_ops = &nlmon_ethtool_ops;
+	dev->destructor	= free_netdev;
+
+	dev->features = NETIF_F_FRAGLIST | NETIF_F_HIGHDMA;
+	dev->flags = IFF_NOARP;
+
+	/* That's rather a softlimit here, which, of course,
+	 * can be altered. Not a real MTU, but what is to be
+	 * expected in most cases.
+	 */
+	dev->mtu = NLMSG_GOODSIZE;
+}
+
+static int nlmon_validate(struct nlattr *tb[], struct nlattr *data[])
+{
+	if (tb[IFLA_ADDRESS])
+		return -EINVAL;
+	return 0;
+}
+
+static struct rtnl_link_ops nlmon_link_ops __read_mostly = {
+	.kind			= "nlmon",
+	.priv_size		= sizeof(struct nlmon),
+	.setup			= nlmon_setup,
+	.validate		= nlmon_validate,
+};
+
+static __init int nlmon_register(void)
+{
+	return rtnl_link_register(&nlmon_link_ops);
+}
+
+static __exit void nlmon_unregister(void)
+{
+	rtnl_link_unregister(&nlmon_link_ops);
+}
+
+module_init(nlmon_register);
+module_exit(nlmon_unregister);
+
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Daniel Borkmann <dborkman@redhat.com>");
+MODULE_AUTHOR("Mathieu Geli <geli@enseirb.fr>");
+MODULE_DESCRIPTION("Netlink monitoring device");
+MODULE_ALIAS_RTNL_LINK("nlmon");
diff --git a/drivers/net/phy/Kconfig b/drivers/net/phy/Kconfig
index 1e11f2b..3a316b3 100644
--- a/drivers/net/phy/Kconfig
+++ b/drivers/net/phy/Kconfig
@@ -144,6 +144,16 @@ config MDIO_OCTEON
 
 	  If in doubt, say Y.
 
+config MDIO_SUN4I
+	tristate "Allwinner sun4i MDIO interface support"
+	depends on ARCH_SUNXI
+	select REGULATOR
+	select REGULATOR_FIXED_VOLTAGE
+	help
+	  This driver supports the MDIO interface found in the network
+	  interface units of the Allwinner SoC that have an EMAC (A10,
+	  A12, A10s, etc.)
+
 config MDIO_BUS_MUX
 	tristate
 	depends on OF_MDIO
diff --git a/drivers/net/phy/Makefile b/drivers/net/phy/Makefile
index 9645e38..23a2ab2 100644
--- a/drivers/net/phy/Makefile
+++ b/drivers/net/phy/Makefile
@@ -30,3 +30,4 @@ obj-$(CONFIG_AMD_PHY)		+= amd.o
 obj-$(CONFIG_MDIO_BUS_MUX)	+= mdio-mux.o
 obj-$(CONFIG_MDIO_BUS_MUX_GPIO)	+= mdio-mux-gpio.o
 obj-$(CONFIG_MDIO_BUS_MUX_MMIOREG) += mdio-mux-mmioreg.o
+obj-$(CONFIG_MDIO_SUN4I)	+= mdio-sun4i.o
diff --git a/drivers/net/phy/at803x.c b/drivers/net/phy/at803x.c
index 45cbc10..1f7091b 100644
--- a/drivers/net/phy/at803x.c
+++ b/drivers/net/phy/at803x.c
@@ -27,15 +27,22 @@
 #define AT803X_MMD_ACCESS_CONTROL		0x0D
 #define AT803X_MMD_ACCESS_CONTROL_DATA		0x0E
 #define AT803X_FUNC_DATA			0x4003
+#define AT803X_DEBUG_ADDR			0x1D
+#define AT803X_DEBUG_DATA			0x1E
+#define AT803X_DEBUG_SYSTEM_MODE_CTRL		0x05
+#define AT803X_DEBUG_RGMII_TX_CLK_DLY		BIT(8)
 
 MODULE_DESCRIPTION("Atheros 803x PHY driver");
 MODULE_AUTHOR("Matus Ujhelyi");
 MODULE_LICENSE("GPL");
 
-static void at803x_set_wol_mac_addr(struct phy_device *phydev)
+static int at803x_set_wol(struct phy_device *phydev,
+			  struct ethtool_wolinfo *wol)
 {
 	struct net_device *ndev = phydev->attached_dev;
 	const u8 *mac;
+	int ret;
+	u32 value;
 	unsigned int i, offsets[] = {
 		AT803X_LOC_MAC_ADDR_32_47_OFFSET,
 		AT803X_LOC_MAC_ADDR_16_31_OFFSET,
@@ -43,30 +50,61 @@ static void at803x_set_wol_mac_addr(struct phy_device *phydev)
 	};
 
 	if (!ndev)
-		return;
+		return -ENODEV;
 
-	mac = (const u8 *) ndev->dev_addr;
+	if (wol->wolopts & WAKE_MAGIC) {
+		mac = (const u8 *) ndev->dev_addr;
 
-	if (!is_valid_ether_addr(mac))
-		return;
+		if (!is_valid_ether_addr(mac))
+			return -EFAULT;
 
-	for (i = 0; i < 3; i++) {
-		phy_write(phydev, AT803X_MMD_ACCESS_CONTROL,
+		for (i = 0; i < 3; i++) {
+			phy_write(phydev, AT803X_MMD_ACCESS_CONTROL,
 				  AT803X_DEVICE_ADDR);
-		phy_write(phydev, AT803X_MMD_ACCESS_CONTROL_DATA,
+			phy_write(phydev, AT803X_MMD_ACCESS_CONTROL_DATA,
 				  offsets[i]);
-		phy_write(phydev, AT803X_MMD_ACCESS_CONTROL,
+			phy_write(phydev, AT803X_MMD_ACCESS_CONTROL,
 				  AT803X_FUNC_DATA);
-		phy_write(phydev, AT803X_MMD_ACCESS_CONTROL_DATA,
+			phy_write(phydev, AT803X_MMD_ACCESS_CONTROL_DATA,
 				  mac[(i * 2) + 1] | (mac[(i * 2)] << 8));
+		}
+
+		value = phy_read(phydev, AT803X_INTR_ENABLE);
+		value |= AT803X_WOL_ENABLE;
+		ret = phy_write(phydev, AT803X_INTR_ENABLE, value);
+		if (ret)
+			return ret;
+		value = phy_read(phydev, AT803X_INTR_STATUS);
+	} else {
+		value = phy_read(phydev, AT803X_INTR_ENABLE);
+		value &= (~AT803X_WOL_ENABLE);
+		ret = phy_write(phydev, AT803X_INTR_ENABLE, value);
+		if (ret)
+			return ret;
+		value = phy_read(phydev, AT803X_INTR_STATUS);
 	}
+
+	return ret;
+}
+
+static void at803x_get_wol(struct phy_device *phydev,
+			   struct ethtool_wolinfo *wol)
+{
+	u32 value;
+
+	wol->supported = WAKE_MAGIC;
+	wol->wolopts = 0;
+
+	value = phy_read(phydev, AT803X_INTR_ENABLE);
+	if (value & AT803X_WOL_ENABLE)
+		wol->wolopts |= WAKE_MAGIC;
 }
 
 static int at803x_config_init(struct phy_device *phydev)
 {
 	int val;
+	int ret;
 	u32 features;
-	int status;
 
 	features = SUPPORTED_TP | SUPPORTED_MII | SUPPORTED_AUI |
 		   SUPPORTED_FIBRE | SUPPORTED_BNC;
@@ -100,20 +138,29 @@ static int at803x_config_init(struct phy_device *phydev)
 	phydev->supported = features;
 	phydev->advertising = features;
 
-	/* enable WOL */
-	at803x_set_wol_mac_addr(phydev);
-	status = phy_write(phydev, AT803X_INTR_ENABLE, AT803X_WOL_ENABLE);
-	status = phy_read(phydev, AT803X_INTR_STATUS);
+	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_TXID) {
+		ret = phy_write(phydev, AT803X_DEBUG_ADDR,
+				AT803X_DEBUG_SYSTEM_MODE_CTRL);
+		if (ret)
+			return ret;
+		ret = phy_write(phydev, AT803X_DEBUG_DATA,
+				AT803X_DEBUG_RGMII_TX_CLK_DLY);
+		if (ret)
+			return ret;
+	}
 
 	return 0;
 }
 
-/* ATHEROS 8035 */
-static struct phy_driver at8035_driver = {
+static struct phy_driver at803x_driver[] = {
+{
+	/* ATHEROS 8035 */
 	.phy_id		= 0x004dd072,
 	.name		= "Atheros 8035 ethernet",
 	.phy_id_mask	= 0xffffffef,
 	.config_init	= at803x_config_init,
+	.set_wol	= at803x_set_wol,
+	.get_wol	= at803x_get_wol,
 	.features	= PHY_GBIT_FEATURES,
 	.flags		= PHY_HAS_INTERRUPT,
 	.config_aneg	= &genphy_config_aneg,
@@ -121,14 +168,14 @@ static struct phy_driver at8035_driver = {
 	.driver		= {
 		.owner = THIS_MODULE,
 	},
-};
-
-/* ATHEROS 8030 */
-static struct phy_driver at8030_driver = {
+}, {
+	/* ATHEROS 8030 */
 	.phy_id		= 0x004dd076,
 	.name		= "Atheros 8030 ethernet",
 	.phy_id_mask	= 0xffffffef,
 	.config_init	= at803x_config_init,
+	.set_wol	= at803x_set_wol,
+	.get_wol	= at803x_get_wol,
 	.features	= PHY_GBIT_FEATURES,
 	.flags		= PHY_HAS_INTERRUPT,
 	.config_aneg	= &genphy_config_aneg,
@@ -136,32 +183,33 @@ static struct phy_driver at8030_driver = {
 	.driver		= {
 		.owner = THIS_MODULE,
 	},
-};
+}, {
+	/* ATHEROS 8031 */
+	.phy_id		= 0x004dd074,
+	.name		= "Atheros 8031 ethernet",
+	.phy_id_mask	= 0xffffffef,
+	.config_init	= at803x_config_init,
+	.set_wol	= at803x_set_wol,
+	.get_wol	= at803x_get_wol,
+	.features	= PHY_GBIT_FEATURES,
+	.flags		= PHY_HAS_INTERRUPT,
+	.config_aneg	= &genphy_config_aneg,
+	.read_status	= &genphy_read_status,
+	.driver		= {
+		.owner = THIS_MODULE,
+	},
+} };
 
 static int __init atheros_init(void)
 {
-	int ret;
-
-	ret = phy_driver_register(&at8035_driver);
-	if (ret)
-		goto err1;
-
-	ret = phy_driver_register(&at8030_driver);
-	if (ret)
-		goto err2;
-
-	return 0;
-
-err2:
-	phy_driver_unregister(&at8035_driver);
-err1:
-	return ret;
+	return phy_drivers_register(at803x_driver,
+				    ARRAY_SIZE(at803x_driver));
 }
 
 static void __exit atheros_exit(void)
 {
-	phy_driver_unregister(&at8035_driver);
-	phy_driver_unregister(&at8030_driver);
+	return phy_drivers_unregister(at803x_driver,
+				      ARRAY_SIZE(at803x_driver));
 }
 
 module_init(atheros_init);
diff --git a/drivers/net/phy/bcm63xx.c b/drivers/net/phy/bcm63xx.c
index 84c7a39..ac55b08 100644
--- a/drivers/net/phy/bcm63xx.c
+++ b/drivers/net/phy/bcm63xx.c
@@ -78,7 +78,7 @@ static struct phy_driver bcm63xx_driver[] = {
 	.name		= "Broadcom BCM63XX (1)",
 	/* ASYM_PAUSE bit is marked RO in datasheet, so don't cheat */
 	.features	= (PHY_BASIC_FEATURES | SUPPORTED_Pause),
-	.flags		= PHY_HAS_INTERRUPT,
+	.flags		= PHY_HAS_INTERRUPT | PHY_IS_INTERNAL,
 	.config_init	= bcm63xx_config_init,
 	.config_aneg	= genphy_config_aneg,
 	.read_status	= genphy_read_status,
@@ -91,7 +91,7 @@ static struct phy_driver bcm63xx_driver[] = {
 	.phy_id_mask	= 0xfffffc00,
 	.name		= "Broadcom BCM63XX (2)",
 	.features	= (PHY_BASIC_FEATURES | SUPPORTED_Pause),
-	.flags		= PHY_HAS_INTERRUPT,
+	.flags		= PHY_HAS_INTERRUPT | PHY_IS_INTERNAL,
 	.config_init	= bcm63xx_config_init,
 	.config_aneg	= genphy_config_aneg,
 	.read_status	= genphy_read_status,
diff --git a/drivers/net/phy/marvell.c b/drivers/net/phy/marvell.c
index 202fe1f..2e91477 100644
--- a/drivers/net/phy/marvell.c
+++ b/drivers/net/phy/marvell.c
@@ -116,6 +116,8 @@
 #define MII_M1011_PHY_STATUS_RESOLVED	0x0800
 #define MII_M1011_PHY_STATUS_LINK	0x0400
 
+#define MII_M1116R_CONTROL_REG_MAC	21
+
 
 MODULE_DESCRIPTION("Marvell PHY driver");
 MODULE_AUTHOR("Andy Fleming");
@@ -372,6 +374,66 @@ static int m88e1318_config_aneg(struct phy_device *phydev)
 	return m88e1121_config_aneg(phydev);
 }
 
+static int m88e1510_config_aneg(struct phy_device *phydev)
+{
+	int err;
+
+	err = m88e1318_config_aneg(phydev);
+	if (err < 0)
+		return err;
+
+	return marvell_of_reg_init(phydev);
+}
+
+static int m88e1116r_config_init(struct phy_device *phydev)
+{
+	int temp;
+	int err;
+
+	temp = phy_read(phydev, MII_BMCR);
+	temp |= BMCR_RESET;
+	err = phy_write(phydev, MII_BMCR, temp);
+	if (err < 0)
+		return err;
+
+	mdelay(500);
+
+	err = phy_write(phydev, MII_MARVELL_PHY_PAGE, 0);
+	if (err < 0)
+		return err;
+
+	temp = phy_read(phydev, MII_M1011_PHY_SCR);
+	temp |= (7 << 12);	/* max number of gigabit attempts */
+	temp |= (1 << 11);	/* enable downshift */
+	temp |= MII_M1011_PHY_SCR_AUTO_CROSS;
+	err = phy_write(phydev, MII_M1011_PHY_SCR, temp);
+	if (err < 0)
+		return err;
+
+	err = phy_write(phydev, MII_MARVELL_PHY_PAGE, 2);
+	if (err < 0)
+		return err;
+	temp = phy_read(phydev, MII_M1116R_CONTROL_REG_MAC);
+	temp |= (1 << 5);
+	temp |= (1 << 4);
+	err = phy_write(phydev, MII_M1116R_CONTROL_REG_MAC, temp);
+	if (err < 0)
+		return err;
+	err = phy_write(phydev, MII_MARVELL_PHY_PAGE, 0);
+	if (err < 0)
+		return err;
+
+	temp = phy_read(phydev, MII_BMCR);
+	temp |= BMCR_RESET;
+	err = phy_write(phydev, MII_BMCR, temp);
+	if (err < 0)
+		return err;
+
+	mdelay(500);
+
+	return 0;
+}
+
 static int m88e1111_config_init(struct phy_device *phydev)
 {
 	int err;
@@ -940,6 +1002,32 @@ static struct phy_driver marvell_drivers[] = {
 		.config_intr = &marvell_config_intr,
 		.driver = { .owner = THIS_MODULE },
 	},
+	{
+		.phy_id = MARVELL_PHY_ID_88E1116R,
+		.phy_id_mask = MARVELL_PHY_ID_MASK,
+		.name = "Marvell 88E1116R",
+		.features = PHY_GBIT_FEATURES,
+		.flags = PHY_HAS_INTERRUPT,
+		.config_init = &m88e1116r_config_init,
+		.config_aneg = &genphy_config_aneg,
+		.read_status = &genphy_read_status,
+		.ack_interrupt = &marvell_ack_interrupt,
+		.config_intr = &marvell_config_intr,
+		.driver = { .owner = THIS_MODULE },
+	},
+	{
+		.phy_id = MARVELL_PHY_ID_88E1510,
+		.phy_id_mask = MARVELL_PHY_ID_MASK,
+		.name = "Marvell 88E1510",
+		.features = PHY_GBIT_FEATURES,
+		.flags = PHY_HAS_INTERRUPT,
+		.config_aneg = &m88e1510_config_aneg,
+		.read_status = &marvell_read_status,
+		.ack_interrupt = &marvell_ack_interrupt,
+		.config_intr = &marvell_config_intr,
+		.did_interrupt = &m88e1121_did_interrupt,
+		.driver = { .owner = THIS_MODULE },
+	},
 };
 
 static int __init marvell_init(void)
@@ -958,15 +1046,17 @@ module_init(marvell_init);
 module_exit(marvell_exit);
 
 static struct mdio_device_id __maybe_unused marvell_tbl[] = {
-	{ 0x01410c60, 0xfffffff0 },
-	{ 0x01410c90, 0xfffffff0 },
-	{ 0x01410cc0, 0xfffffff0 },
-	{ 0x01410e10, 0xfffffff0 },
-	{ 0x01410cb0, 0xfffffff0 },
-	{ 0x01410cd0, 0xfffffff0 },
-	{ 0x01410e50, 0xfffffff0 },
-	{ 0x01410e30, 0xfffffff0 },
-	{ 0x01410e90, 0xfffffff0 },
+	{ MARVELL_PHY_ID_88E1101, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1112, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1111, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1118, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1121R, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1145, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1149R, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1240, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1318S, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1116R, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E1510, MARVELL_PHY_ID_MASK },
 	{ }
 };
 
diff --git a/drivers/net/phy/mdio-sun4i.c b/drivers/net/phy/mdio-sun4i.c
new file mode 100644
index 0000000..61d3f4e
--- /dev/null
+++ b/drivers/net/phy/mdio-sun4i.c
@@ -0,0 +1,194 @@
+/*
+ * Allwinner EMAC MDIO interface driver
+ *
+ * Copyright 2012-2013 Stefan Roese <sr@denx.de>
+ * Copyright 2013 Maxime Ripard <maxime.ripard@free-electrons.com>
+ *
+ * Based on the Linux driver provided by Allwinner:
+ * Copyright (C) 1997  Sten Wang
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/of_address.h>
+#include <linux/of_mdio.h>
+#include <linux/phy.h>
+#include <linux/platform_device.h>
+#include <linux/regulator/consumer.h>
+
+#define EMAC_MAC_MCMD_REG	(0x00)
+#define EMAC_MAC_MADR_REG	(0x04)
+#define EMAC_MAC_MWTD_REG	(0x08)
+#define EMAC_MAC_MRDD_REG	(0x0c)
+#define EMAC_MAC_MIND_REG	(0x10)
+#define EMAC_MAC_SSRR_REG	(0x14)
+
+#define MDIO_TIMEOUT		(msecs_to_jiffies(100))
+
+struct sun4i_mdio_data {
+	void __iomem		*membase;
+	struct regulator	*regulator;
+};
+
+static int sun4i_mdio_read(struct mii_bus *bus, int mii_id, int regnum)
+{
+	struct sun4i_mdio_data *data = bus->priv;
+	unsigned long start_jiffies;
+	int value;
+
+	/* issue the phy address and reg */
+	writel((mii_id << 8) | regnum, data->membase + EMAC_MAC_MADR_REG);
+	/* pull up the phy io line */
+	writel(0x1, data->membase + EMAC_MAC_MCMD_REG);
+
+	/* Wait read complete */
+	start_jiffies = jiffies;
+	while (readl(data->membase + EMAC_MAC_MIND_REG) & 0x1) {
+		if (time_after(start_jiffies,
+			       start_jiffies + MDIO_TIMEOUT))
+			return -ETIMEDOUT;
+		msleep(1);
+	}
+
+	/* push down the phy io line */
+	writel(0x0, data->membase + EMAC_MAC_MCMD_REG);
+	/* and read data */
+	value = readl(data->membase + EMAC_MAC_MRDD_REG);
+
+	return value;
+}
+
+static int sun4i_mdio_write(struct mii_bus *bus, int mii_id, int regnum,
+			    u16 value)
+{
+	struct sun4i_mdio_data *data = bus->priv;
+	unsigned long start_jiffies;
+
+	/* issue the phy address and reg */
+	writel((mii_id << 8) | regnum, data->membase + EMAC_MAC_MADR_REG);
+	/* pull up the phy io line */
+	writel(0x1, data->membase + EMAC_MAC_MCMD_REG);
+
+	/* Wait read complete */
+	start_jiffies = jiffies;
+	while (readl(data->membase + EMAC_MAC_MIND_REG) & 0x1) {
+		if (time_after(start_jiffies,
+			       start_jiffies + MDIO_TIMEOUT))
+			return -ETIMEDOUT;
+		msleep(1);
+	}
+
+	/* push down the phy io line */
+	writel(0x0, data->membase + EMAC_MAC_MCMD_REG);
+	/* and write data */
+	writel(value, data->membase + EMAC_MAC_MWTD_REG);
+
+	return 0;
+}
+
+static int sun4i_mdio_reset(struct mii_bus *bus)
+{
+	return 0;
+}
+
+static int sun4i_mdio_probe(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct mii_bus *bus;
+	struct sun4i_mdio_data *data;
+	int ret, i;
+
+	bus = mdiobus_alloc_size(sizeof(*data));
+	if (!bus)
+		return -ENOMEM;
+
+	bus->name = "sun4i_mii_bus";
+	bus->read = &sun4i_mdio_read;
+	bus->write = &sun4i_mdio_write;
+	bus->reset = &sun4i_mdio_reset;
+	snprintf(bus->id, MII_BUS_ID_SIZE, "%s-mii", dev_name(&pdev->dev));
+	bus->parent = &pdev->dev;
+
+	bus->irq = kmalloc(sizeof(int) * PHY_MAX_ADDR, GFP_KERNEL);
+	if (!bus->irq) {
+		ret = -ENOMEM;
+		goto err_out_free_mdiobus;
+	}
+
+	for (i = 0; i < PHY_MAX_ADDR; i++)
+		bus->irq[i] = PHY_POLL;
+
+	data = bus->priv;
+	data->membase = of_iomap(np, 0);
+	if (!data->membase) {
+		ret = -ENOMEM;
+		goto err_out_free_mdio_irq;
+	}
+
+	data->regulator = devm_regulator_get(&pdev->dev, "phy");
+	if (IS_ERR(data->regulator)) {
+		if (PTR_ERR(data->regulator) == -EPROBE_DEFER)
+			return -EPROBE_DEFER;
+
+		dev_info(&pdev->dev, "no regulator found\n");
+	} else {
+		ret = regulator_enable(data->regulator);
+		if (ret)
+			goto err_out_free_mdio_irq;
+	}
+
+	ret = of_mdiobus_register(bus, np);
+	if (ret < 0)
+		goto err_out_disable_regulator;
+
+	platform_set_drvdata(pdev, bus);
+
+	return 0;
+
+err_out_disable_regulator:
+	regulator_disable(data->regulator);
+err_out_free_mdio_irq:
+	kfree(bus->irq);
+err_out_free_mdiobus:
+	mdiobus_free(bus);
+	return ret;
+}
+
+static int sun4i_mdio_remove(struct platform_device *pdev)
+{
+	struct mii_bus *bus = platform_get_drvdata(pdev);
+
+	mdiobus_unregister(bus);
+	kfree(bus->irq);
+	mdiobus_free(bus);
+
+	return 0;
+}
+
+static const struct of_device_id sun4i_mdio_dt_ids[] = {
+	{ .compatible = "allwinner,sun4i-mdio" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sun4i_mdio_dt_ids);
+
+static struct platform_driver sun4i_mdio_driver = {
+	.probe = sun4i_mdio_probe,
+	.remove = sun4i_mdio_remove,
+	.driver = {
+		.name = "sun4i-mdio",
+		.of_match_table = sun4i_mdio_dt_ids,
+	},
+};
+
+module_platform_driver(sun4i_mdio_driver);
+
+MODULE_DESCRIPTION("Allwinner EMAC MDIO interface driver");
+MODULE_AUTHOR("Maxime Ripard <maxime.ripard@free-electrons.com>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
index 663d2d0..36c6994 100644
--- a/drivers/net/phy/phy.c
+++ b/drivers/net/phy/phy.c
@@ -294,7 +294,8 @@ int phy_ethtool_gset(struct phy_device *phydev, struct ethtool_cmd *cmd)
 	cmd->duplex = phydev->duplex;
 	cmd->port = PORT_MII;
 	cmd->phy_address = phydev->addr;
-	cmd->transceiver = XCVR_EXTERNAL;
+	cmd->transceiver = phy_is_internal(phydev) ?
+		XCVR_INTERNAL : XCVR_EXTERNAL;
 	cmd->autoneg = phydev->autoneg;
 
 	return 0;
@@ -419,8 +420,6 @@ out_unlock:
 EXPORT_SYMBOL(phy_start_aneg);
 
 
-static void phy_change(struct work_struct *work);
-
 /**
  * phy_start_machine - start PHY state machine tracking
  * @phydev: the phy_device struct
@@ -565,8 +564,6 @@ int phy_start_interrupts(struct phy_device *phydev)
 {
 	int err = 0;
 
-	INIT_WORK(&phydev->phy_queue, phy_change);
-
 	atomic_set(&phydev->irq_disable, 0);
 	if (request_irq(phydev->irq, phy_interrupt,
 				IRQF_SHARED,
@@ -623,7 +620,7 @@ EXPORT_SYMBOL(phy_stop_interrupts);
  * phy_change - Scheduled by the phy_interrupt/timer to handle PHY changes
  * @work: work_struct that describes the work to be done
  */
-static void phy_change(struct work_struct *work)
+void phy_change(struct work_struct *work)
 {
 	int err;
 	struct phy_device *phydev =
@@ -682,7 +679,7 @@ void phy_stop(struct phy_device *phydev)
 	if (PHY_HALTED == phydev->state)
 		goto out_unlock;
 
-	if (phydev->irq != PHY_POLL) {
+	if (phy_interrupt_is_valid(phydev)) {
 		/* Disable PHY Interrupts */
 		phy_config_interrupt(phydev, PHY_INTERRUPT_DISABLED);
 
@@ -828,8 +825,9 @@ void phy_state_machine(struct work_struct *work)
 			break;
 		case PHY_RUNNING:
 			/* Only register a CHANGE if we are
-			 * polling */
-			if (PHY_POLL == phydev->irq)
+			 * polling or ignoring interrupts
+			 */
+			if (!phy_interrupt_is_valid(phydev))
 				phydev->state = PHY_CHANGELINK;
 			break;
 		case PHY_CHANGELINK:
@@ -848,7 +846,7 @@ void phy_state_machine(struct work_struct *work)
 
 			phydev->adjust_link(phydev->attached_dev);
 
-			if (PHY_POLL != phydev->irq)
+			if (phy_interrupt_is_valid(phydev))
 				err = phy_config_interrupt(phydev,
 						PHY_INTERRUPT_ENABLED);
 			break;
@@ -922,6 +920,14 @@ void phy_state_machine(struct work_struct *work)
 			PHY_STATE_TIME * HZ);
 }
 
+void phy_mac_interrupt(struct phy_device *phydev, int new_link)
+{
+	cancel_work_sync(&phydev->phy_queue);
+	phydev->link = new_link;
+	schedule_work(&phydev->phy_queue);
+}
+EXPORT_SYMBOL(phy_mac_interrupt);
+
 static inline void mmd_phy_indirect(struct mii_bus *bus, int prtad, int devad,
 				    int addr)
 {
diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index 3657b4a..74630e9 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c
@@ -189,6 +189,7 @@ struct phy_device *phy_device_create(struct mii_bus *bus, int addr, int phy_id,
 
 	mutex_init(&dev->lock);
 	INIT_DELAYED_WORK(&dev->state_queue, phy_state_machine);
+	INIT_WORK(&dev->phy_queue, phy_change);
 
 	/* Request the appropriate module unconditionally; don't
 	   bother trying to do so only if it isn't already loaded,
@@ -1009,10 +1010,16 @@ static int phy_probe(struct device *dev)
 	phydrv = to_phy_driver(drv);
 	phydev->drv = phydrv;
 
-	/* Disable the interrupt if the PHY doesn't support it */
-	if (!(phydrv->flags & PHY_HAS_INTERRUPT))
+	/* Disable the interrupt if the PHY doesn't support it
+	 * but the interrupt is still a valid one
+	 */
+	if (!(phydrv->flags & PHY_HAS_INTERRUPT) &&
+			phy_interrupt_is_valid(phydev))
 		phydev->irq = PHY_POLL;
 
+	if (phydrv->flags & PHY_IS_INTERNAL)
+		phydev->is_internal = true;
+
 	mutex_lock(&phydev->lock);
 
 	/* Start out supporting everything. Eventually,
diff --git a/drivers/net/phy/spi_ks8995.c b/drivers/net/phy/spi_ks8995.c
index d11c93e..f3bea13 100644
--- a/drivers/net/phy/spi_ks8995.c
+++ b/drivers/net/phy/spi_ks8995.c
@@ -354,19 +354,7 @@ static struct spi_driver ks8995_driver = {
 	.remove	  = ks8995_remove,
 };
 
-static int __init ks8995_init(void)
-{
-	pr_info(DRV_DESC " version " DRV_VERSION "\n");
-
-	return spi_register_driver(&ks8995_driver);
-}
-module_init(ks8995_init);
-
-static void __exit ks8995_exit(void)
-{
-	spi_unregister_driver(&ks8995_driver);
-}
-module_exit(ks8995_exit);
+module_spi_driver(ks8995_driver);
 
 MODULE_DESCRIPTION(DRV_DESC);
 MODULE_VERSION(DRV_VERSION);
diff --git a/drivers/net/phy/vitesse.c b/drivers/net/phy/vitesse.c
index 3492b53..69b482b 100644
--- a/drivers/net/phy/vitesse.c
+++ b/drivers/net/phy/vitesse.c
@@ -44,18 +44,19 @@
 #define MII_VSC8244_ISTAT_DUPLEX	0x1000
 
 /* Vitesse Auxiliary Control/Status Register */
-#define MII_VSC8244_AUX_CONSTAT        	0x1c
-#define MII_VSC8244_AUXCONSTAT_INIT    	0x0000
-#define MII_VSC8244_AUXCONSTAT_DUPLEX  	0x0020
-#define MII_VSC8244_AUXCONSTAT_SPEED   	0x0018
-#define MII_VSC8244_AUXCONSTAT_GBIT    	0x0010
-#define MII_VSC8244_AUXCONSTAT_100     	0x0008
+#define MII_VSC8244_AUX_CONSTAT		0x1c
+#define MII_VSC8244_AUXCONSTAT_INIT	0x0000
+#define MII_VSC8244_AUXCONSTAT_DUPLEX	0x0020
+#define MII_VSC8244_AUXCONSTAT_SPEED	0x0018
+#define MII_VSC8244_AUXCONSTAT_GBIT	0x0010
+#define MII_VSC8244_AUXCONSTAT_100	0x0008
 
 #define MII_VSC8221_AUXCONSTAT_INIT	0x0004 /* need to set this bit? */
 #define MII_VSC8221_AUXCONSTAT_RESERVED	0x0004
 
 #define PHY_ID_VSC8244			0x000fc6c0
 #define PHY_ID_VSC8221			0x000fc550
+#define PHY_ID_VSC8211			0x000fc4b0
 
 MODULE_DESCRIPTION("Vitesse PHY driver");
 MODULE_AUTHOR("Kriston Carson");
@@ -100,9 +101,8 @@ static int vsc824x_config_init(struct phy_device *phydev)
 static int vsc824x_ack_interrupt(struct phy_device *phydev)
 {
 	int err = 0;
-	
-	/*
-	 * Don't bother to ACK the interrupts if interrupts
+
+	/* Don't bother to ACK the interrupts if interrupts
 	 * are disabled.  The 824x cannot clear the interrupts
 	 * if they are disabled.
 	 */
@@ -122,8 +122,7 @@ static int vsc82xx_config_intr(struct phy_device *phydev)
 				MII_VSC8244_IMASK_MASK :
 				MII_VSC8221_IMASK_MASK);
 	else {
-		/*
-		 * The Vitesse PHY cannot clear the interrupt
+		/* The Vitesse PHY cannot clear the interrupt
 		 * once it has disabled them, so we clear them first
 		 */
 		err = phy_read(phydev, MII_VSC8244_ISTAT);
@@ -146,7 +145,8 @@ static int vsc8221_config_init(struct phy_device *phydev)
 	return err;
 
 	/* Perhaps we should set EXT_CON1 based on the interface?
-	   Options are 802.3Z SerDes or SGMII */
+	 * Options are 802.3Z SerDes or SGMII
+	 */
 }
 
 /* Vitesse 824x */
@@ -176,6 +176,19 @@ static struct phy_driver vsc82xx_driver[] = {
 	.ack_interrupt	= &vsc824x_ack_interrupt,
 	.config_intr	= &vsc82xx_config_intr,
 	.driver		= { .owner = THIS_MODULE,},
+}, {
+	/* Vitesse 8211 */
+	.phy_id		= PHY_ID_VSC8211,
+	.phy_id_mask	= 0x000ffff0,
+	.name		= "Vitesse VSC8211",
+	.features	= PHY_GBIT_FEATURES,
+	.flags		= PHY_HAS_INTERRUPT,
+	.config_init	= &vsc8221_config_init,
+	.config_aneg	= &genphy_config_aneg,
+	.read_status	= &genphy_read_status,
+	.ack_interrupt	= &vsc824x_ack_interrupt,
+	.config_intr	= &vsc82xx_config_intr,
+	.driver		= { .owner = THIS_MODULE,},
 } };
 
 static int __init vsc82xx_init(void)
@@ -196,6 +209,7 @@ module_exit(vsc82xx_exit);
 static struct mdio_device_id __maybe_unused vitesse_tbl[] = {
 	{ PHY_ID_VSC8244, 0x000fffc0 },
 	{ PHY_ID_VSC8221, 0x000ffff0 },
+	{ PHY_ID_VSC8211, 0x000ffff0 },
 	{ }
 };
 
diff --git a/drivers/net/ppp/pppoe.c b/drivers/net/ppp/pppoe.c
index bb07ba9..5f66e30 100644
--- a/drivers/net/ppp/pppoe.c
+++ b/drivers/net/ppp/pppoe.c
@@ -338,7 +338,7 @@ static void pppoe_flush_dev(struct net_device *dev)
 static int pppoe_device_event(struct notifier_block *this,
 			      unsigned long event, void *ptr)
 {
-	struct net_device *dev = (struct net_device *)ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 
 	/* Only look at sockets that are using this specific device. */
 	switch (event) {
diff --git a/drivers/net/team/team.c b/drivers/net/team/team.c
index b305105..bff7e0b 100644
--- a/drivers/net/team/team.c
+++ b/drivers/net/team/team.c
@@ -525,31 +525,26 @@ static void team_set_no_mode(struct team *team)
 	team->mode = &__team_no_mode;
 }
 
-static void __team_adjust_ops(struct team *team, int en_port_count)
+static void team_adjust_ops(struct team *team)
 {
 	/*
 	 * To avoid checks in rx/tx skb paths, ensure here that non-null and
 	 * correct ops are always set.
 	 */
 
-	if (!en_port_count || !team_is_mode_set(team) ||
+	if (!team->en_port_count || !team_is_mode_set(team) ||
 	    !team->mode->ops->transmit)
 		team->ops.transmit = team_dummy_transmit;
 	else
 		team->ops.transmit = team->mode->ops->transmit;
 
-	if (!en_port_count || !team_is_mode_set(team) ||
+	if (!team->en_port_count || !team_is_mode_set(team) ||
 	    !team->mode->ops->receive)
 		team->ops.receive = team_dummy_receive;
 	else
 		team->ops.receive = team->mode->ops->receive;
 }
 
-static void team_adjust_ops(struct team *team)
-{
-	__team_adjust_ops(team, team->en_port_count);
-}
-
 /*
  * We can benefit from the fact that it's ensured no port is present
  * at the time of mode change. Therefore no packets are in fly so there's no
@@ -725,9 +720,9 @@ static bool team_queue_override_transmit(struct team *team, struct sk_buff *skb)
 static void __team_queue_override_port_del(struct team *team,
 					   struct team_port *port)
 {
+	if (!port->queue_id)
+		return;
 	list_del_rcu(&port->qom_list);
-	synchronize_rcu();
-	INIT_LIST_HEAD(&port->qom_list);
 }
 
 static bool team_queue_override_port_has_gt_prio_than(struct team_port *port,
@@ -749,9 +744,8 @@ static void __team_queue_override_port_add(struct team *team,
 	struct list_head *qom_list;
 	struct list_head *node;
 
-	if (!port->queue_id || !team_port_enabled(port))
+	if (!port->queue_id)
 		return;
-
 	qom_list = __team_get_qom_list(team, port->queue_id);
 	node = qom_list;
 	list_for_each_entry(cur, qom_list, qom_list) {
@@ -768,7 +762,7 @@ static void __team_queue_override_enabled_check(struct team *team)
 	bool enabled = false;
 
 	list_for_each_entry(port, &team->port_list, list) {
-		if (!list_empty(&port->qom_list)) {
+		if (port->queue_id) {
 			enabled = true;
 			break;
 		}
@@ -780,14 +774,44 @@ static void __team_queue_override_enabled_check(struct team *team)
 	team->queue_override_enabled = enabled;
 }
 
-static void team_queue_override_port_refresh(struct team *team,
-					     struct team_port *port)
+static void team_queue_override_port_prio_changed(struct team *team,
+						  struct team_port *port)
 {
+	if (!port->queue_id || team_port_enabled(port))
+		return;
 	__team_queue_override_port_del(team, port);
 	__team_queue_override_port_add(team, port);
 	__team_queue_override_enabled_check(team);
 }
 
+static void team_queue_override_port_change_queue_id(struct team *team,
+						     struct team_port *port,
+						     u16 new_queue_id)
+{
+	if (team_port_enabled(port)) {
+		__team_queue_override_port_del(team, port);
+		port->queue_id = new_queue_id;
+		__team_queue_override_port_add(team, port);
+		__team_queue_override_enabled_check(team);
+	} else {
+		port->queue_id = new_queue_id;
+	}
+}
+
+static void team_queue_override_port_add(struct team *team,
+					 struct team_port *port)
+{
+	__team_queue_override_port_add(team, port);
+	__team_queue_override_enabled_check(team);
+}
+
+static void team_queue_override_port_del(struct team *team,
+					 struct team_port *port)
+{
+	__team_queue_override_port_del(team, port);
+	__team_queue_override_enabled_check(team);
+}
+
 
 /****************
  * Port handling
@@ -819,7 +843,7 @@ static void team_port_enable(struct team *team,
 	hlist_add_head_rcu(&port->hlist,
 			   team_port_index_hash(team, port->index));
 	team_adjust_ops(team);
-	team_queue_override_port_refresh(team, port);
+	team_queue_override_port_add(team, port);
 	if (team->ops.port_enabled)
 		team->ops.port_enabled(team, port);
 }
@@ -848,14 +872,9 @@ static void team_port_disable(struct team *team,
 	hlist_del_rcu(&port->hlist);
 	__reconstruct_port_hlist(team, port->index);
 	port->index = -1;
-	team_queue_override_port_refresh(team, port);
-	__team_adjust_ops(team, team->en_port_count - 1);
-	/*
-	 * Wait until readers see adjusted ops. This ensures that
-	 * readers never see team->en_port_count == 0
-	 */
-	synchronize_rcu();
 	team->en_port_count--;
+	team_queue_override_port_del(team, port);
+	team_adjust_ops(team);
 }
 
 #define TEAM_VLAN_FEATURES (NETIF_F_ALL_CSUM | NETIF_F_SG | \
@@ -1163,8 +1182,7 @@ static int team_port_del(struct team *team, struct net_device *port_dev)
 
 	team_port_set_orig_dev_addr(port);
 	dev_set_mtu(port_dev, port->orig.mtu);
-	synchronize_rcu();
-	kfree(port);
+	kfree_rcu(port, rcu);
 	netdev_info(dev, "Port device %s removed\n", portname);
 	__team_compute_features(team);
 
@@ -1259,9 +1277,12 @@ static int team_priority_option_set(struct team *team,
 				    struct team_gsetter_ctx *ctx)
 {
 	struct team_port *port = ctx->info->port;
+	s32 priority = ctx->data.s32_val;
 
-	port->priority = ctx->data.s32_val;
-	team_queue_override_port_refresh(team, port);
+	if (port->priority == priority)
+		return 0;
+	port->priority = priority;
+	team_queue_override_port_prio_changed(team, port);
 	return 0;
 }
 
@@ -1278,17 +1299,16 @@ static int team_queue_id_option_set(struct team *team,
 				    struct team_gsetter_ctx *ctx)
 {
 	struct team_port *port = ctx->info->port;
+	u16 new_queue_id = ctx->data.u32_val;
 
-	if (port->queue_id == ctx->data.u32_val)
+	if (port->queue_id == new_queue_id)
 		return 0;
-	if (ctx->data.u32_val >= team->dev->real_num_tx_queues)
+	if (new_queue_id >= team->dev->real_num_tx_queues)
 		return -EINVAL;
-	port->queue_id = ctx->data.u32_val;
-	team_queue_override_port_refresh(team, port);
+	team_queue_override_port_change_queue_id(team, port, new_queue_id);
 	return 0;
 }
 
-
 static const struct team_option team_options[] = {
 	{
 		.name = "mode",
@@ -2648,7 +2668,7 @@ static void team_port_change_check(struct team_port *port, bool linkup)
 static int team_device_event(struct notifier_block *unused,
 			     unsigned long event, void *ptr)
 {
-	struct net_device *dev = (struct net_device *) ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct team_port *port;
 
 	port = team_port_get_rtnl(dev);
diff --git a/drivers/net/team/team_mode_loadbalance.c b/drivers/net/team/team_mode_loadbalance.c
index cdc31b5..829a9cd 100644
--- a/drivers/net/team/team_mode_loadbalance.c
+++ b/drivers/net/team/team_mode_loadbalance.c
@@ -112,9 +112,8 @@ static struct team_port *lb_hash_select_tx_port(struct team *team,
 						struct sk_buff *skb,
 						unsigned char hash)
 {
-	int port_index;
+	int port_index = team_num_to_port_index(team, hash);
 
-	port_index = hash % team->en_port_count;
 	return team_get_port_by_index_rcu(team, port_index);
 }
 
diff --git a/drivers/net/team/team_mode_roundrobin.c b/drivers/net/team/team_mode_roundrobin.c
index 472623f..5366585 100644
--- a/drivers/net/team/team_mode_roundrobin.c
+++ b/drivers/net/team/team_mode_roundrobin.c
@@ -30,7 +30,8 @@ static bool rr_transmit(struct team *team, struct sk_buff *skb)
 	struct team_port *port;
 	int port_index;
 
-	port_index = rr_priv(team)->sent_packets++ % team->en_port_count;
+	port_index = team_num_to_port_index(team,
+					    rr_priv(team)->sent_packets++);
 	port = team_get_port_by_index_rcu(team, port_index);
 	if (unlikely(!port))
 		goto drop;
diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index 9c61f87..7eab5fc 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -841,7 +841,7 @@ static const struct net_device_ops tap_netdev_ops = {
 #endif
 };
 
-static int tun_flow_init(struct tun_struct *tun)
+static void tun_flow_init(struct tun_struct *tun)
 {
 	int i;
 
@@ -852,8 +852,6 @@ static int tun_flow_init(struct tun_struct *tun)
 	setup_timer(&tun->flow_gc_timer, tun_flow_cleanup, (unsigned long)tun);
 	mod_timer(&tun->flow_gc_timer,
 		  round_jiffies_up(jiffies + tun->ageing_time));
-
-	return 0;
 }
 
 static void tun_flow_uninit(struct tun_struct *tun)
@@ -1532,6 +1530,9 @@ static int tun_flags(struct tun_struct *tun)
 	if (tun->flags & TUN_TAP_MQ)
 		flags |= IFF_MULTI_QUEUE;
 
+	if (tun->flags & TUN_PERSIST)
+		flags |= IFF_PERSIST;
+
 	return flags;
 }
 
@@ -1661,10 +1662,7 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
 			goto err_free_dev;
 
 		tun_net_init(dev);
-
-		err = tun_flow_init(tun);
-		if (err < 0)
-			goto err_free_dev;
+		tun_flow_init(tun);
 
 		dev->hw_features = NETIF_F_SG | NETIF_F_FRAGLIST |
 			TUN_USER_FEATURES;
diff --git a/drivers/net/usb/Kconfig b/drivers/net/usb/Kconfig
index 287cc62..d84bfd4 100644
--- a/drivers/net/usb/Kconfig
+++ b/drivers/net/usb/Kconfig
@@ -67,7 +67,6 @@ config USB_KAWETH
 
 config USB_PEGASUS
 	tristate "USB Pegasus/Pegasus-II based ethernet device support"
-	select NET_CORE
 	select MII
 	---help---
 	  Say Y here if you know you have Pegasus or Pegasus-II based adapter.
@@ -83,7 +82,6 @@ config USB_PEGASUS
 
 config USB_RTL8150
 	tristate "USB RTL8150 based ethernet device support"
-	select NET_CORE
 	select MII
 	help
 	  Say Y here if you have RTL8150 based usb-ethernet adapter.
@@ -95,7 +93,6 @@ config USB_RTL8150
 
 config USB_RTL8152
 	tristate "Realtek RTL8152 Based USB 2.0 Ethernet Adapters"
-	select NET_CORE
 	select MII
 	help
 	  This option adds support for Realtek RTL8152 based USB 2.0
@@ -106,7 +103,6 @@ config USB_RTL8152
 
 config USB_USBNET
 	tristate "Multi-purpose USB Networking Framework"
-	select NET_CORE
 	select MII
 	---help---
 	  This driver supports several kinds of network links over USB,
diff --git a/drivers/net/usb/ax88179_178a.c b/drivers/net/usb/ax88179_178a.c
index bd8758f..1e3c302 100644
--- a/drivers/net/usb/ax88179_178a.c
+++ b/drivers/net/usb/ax88179_178a.c
@@ -1371,7 +1371,7 @@ static int ax88179_stop(struct usbnet *dev)
 }
 
 static const struct driver_info ax88179_info = {
-	.description = "ASIX AX88179 USB 3.0 Gigibit Ethernet",
+	.description = "ASIX AX88179 USB 3.0 Gigabit Ethernet",
 	.bind = ax88179_bind,
 	.unbind = ax88179_unbind,
 	.status = ax88179_status,
@@ -1384,7 +1384,7 @@ static const struct driver_info ax88179_info = {
 };
 
 static const struct driver_info ax88178a_info = {
-	.description = "ASIX AX88178A USB 2.0 Gigibit Ethernet",
+	.description = "ASIX AX88178A USB 2.0 Gigabit Ethernet",
 	.bind = ax88179_bind,
 	.unbind = ax88179_unbind,
 	.status = ax88179_status,
@@ -1433,6 +1433,7 @@ static struct usb_driver ax88179_178a_driver = {
 	.probe =	usbnet_probe,
 	.suspend =	ax88179_suspend,
 	.resume =	ax88179_resume,
+	.reset_resume =	ax88179_resume,
 	.disconnect =	usbnet_disconnect,
 	.supports_autosuspend = 1,
 	.disable_hub_initiated_lpm = 1,
diff --git a/drivers/net/usb/cdc_ether.c b/drivers/net/usb/cdc_ether.c
index 04ee044..4393f14 100644
--- a/drivers/net/usb/cdc_ether.c
+++ b/drivers/net/usb/cdc_ether.c
@@ -215,6 +215,10 @@ int usbnet_generic_cdc_bind(struct usbnet *dev, struct usb_interface *intf)
 					goto bad_desc;
 			}
 
+			/* some devices merge these - skip class check */
+			if (info->control == info->data)
+				goto next_desc;
+
 			/* a data interface altsetting does the real i/o */
 			d = &info->data->cur_altsetting->desc;
 			if (d->bInterfaceClass != USB_CLASS_CDC_DATA) {
@@ -304,19 +308,23 @@ next_desc:
 	/* claim data interface and set it up ... with side effects.
 	 * network traffic can't flow until an altsetting is enabled.
 	 */
-	status = usb_driver_claim_interface(driver, info->data, dev);
-	if (status < 0)
-		return status;
+	if (info->data != info->control) {
+		status = usb_driver_claim_interface(driver, info->data, dev);
+		if (status < 0)
+			return status;
+	}
 	status = usbnet_get_endpoints(dev, info->data);
 	if (status < 0) {
 		/* ensure immediate exit from usbnet_disconnect */
 		usb_set_intfdata(info->data, NULL);
-		usb_driver_release_interface(driver, info->data);
+		if (info->data != info->control)
+			usb_driver_release_interface(driver, info->data);
 		return status;
 	}
 
 	/* status endpoint: optional for CDC Ethernet, not RNDIS (or ACM) */
-	dev->status = NULL;
+	if (info->data != info->control)
+		dev->status = NULL;
 	if (info->control->cur_altsetting->desc.bNumEndpoints == 1) {
 		struct usb_endpoint_descriptor	*desc;
 
@@ -349,6 +357,10 @@ void usbnet_cdc_unbind(struct usbnet *dev, struct usb_interface *intf)
 	struct cdc_state		*info = (void *) &dev->data;
 	struct usb_driver		*driver = driver_of(intf);
 
+	/* combined interface - nothing  to do */
+	if (info->data == info->control)
+		return;
+
 	/* disconnect master --> disconnect slave */
 	if (intf == info->control && info->data) {
 		/* ensure immediate exit from usbnet_disconnect */
diff --git a/drivers/net/usb/ipheth.c b/drivers/net/usb/ipheth.c
index 534d8be..ff8594d 100644
--- a/drivers/net/usb/ipheth.c
+++ b/drivers/net/usb/ipheth.c
@@ -60,6 +60,7 @@
 #define USB_PRODUCT_IPHONE_3GS  0x1294
 #define USB_PRODUCT_IPHONE_4	0x1297
 #define USB_PRODUCT_IPAD 0x129a
+#define USB_PRODUCT_IPAD_MINI    0x12ab
 #define USB_PRODUCT_IPHONE_4_VZW 0x129c
 #define USB_PRODUCT_IPHONE_4S	0x12a0
 #define USB_PRODUCT_IPHONE_5	0x12a8
@@ -107,6 +108,10 @@ static struct usb_device_id ipheth_table[] = {
 		IPHETH_USBINTF_CLASS, IPHETH_USBINTF_SUBCLASS,
 		IPHETH_USBINTF_PROTO) },
 	{ USB_DEVICE_AND_INTERFACE_INFO(
+		USB_VENDOR_APPLE, USB_PRODUCT_IPAD_MINI,
+		IPHETH_USBINTF_CLASS, IPHETH_USBINTF_SUBCLASS,
+		IPHETH_USBINTF_PROTO) },
+	{ USB_DEVICE_AND_INTERFACE_INFO(
 		USB_VENDOR_APPLE, USB_PRODUCT_IPHONE_4_VZW,
 		IPHETH_USBINTF_CLASS, IPHETH_USBINTF_SUBCLASS,
 		IPHETH_USBINTF_PROTO) },
diff --git a/drivers/net/usb/kalmia.c b/drivers/net/usb/kalmia.c
index 0192073..6866eae 100644
--- a/drivers/net/usb/kalmia.c
+++ b/drivers/net/usb/kalmia.c
@@ -221,12 +221,9 @@ done:
 		memset(skb_put(skb, padlen), 0, padlen);
 	}
 
-	netdev_dbg(
-		dev->net,
-		"Sending package with length %i and padding %i. Header: %02x:%02x:%02x:%02x:%02x:%02x.",
-		content_len, padlen, header_start[0], header_start[1],
-		header_start[2], header_start[3], header_start[4],
-		header_start[5]);
+	netdev_dbg(dev->net,
+		"Sending package with length %i and padding %i. Header: %6phC.",
+		content_len, padlen, header_start);
 
 	return skb;
 }
@@ -263,32 +260,23 @@ kalmia_rx_fixup(struct usbnet *dev, struct sk_buff *skb)
 				sizeof(EXPECTED_UNKNOWN_HEADER_1)) || !memcmp(
 				header_start, EXPECTED_UNKNOWN_HEADER_2,
 				sizeof(EXPECTED_UNKNOWN_HEADER_2))) {
-				netdev_dbg(
-					dev->net,
-					"Received expected unknown frame header: %02x:%02x:%02x:%02x:%02x:%02x. Package length: %i\n",
-					header_start[0], header_start[1],
-					header_start[2], header_start[3],
-					header_start[4], header_start[5],
+				netdev_dbg(dev->net,
+					"Received expected unknown frame header: %6phC. Package length: %i\n",
+					header_start,
 					skb->len - KALMIA_HEADER_LENGTH);
 			}
 			else {
-				netdev_err(
-					dev->net,
-					"Received unknown frame header: %02x:%02x:%02x:%02x:%02x:%02x. Package length: %i\n",
-					header_start[0], header_start[1],
-					header_start[2], header_start[3],
-					header_start[4], header_start[5],
+				netdev_err(dev->net,
+					"Received unknown frame header: %6phC. Package length: %i\n",
+					header_start,
 					skb->len - KALMIA_HEADER_LENGTH);
 				return 0;
 			}
 		}
 		else
-			netdev_dbg(
-				dev->net,
-				"Received header: %02x:%02x:%02x:%02x:%02x:%02x. Package length: %i\n",
-				header_start[0], header_start[1], header_start[2],
-				header_start[3], header_start[4], header_start[5],
-				skb->len - KALMIA_HEADER_LENGTH);
+			netdev_dbg(dev->net,
+				"Received header: %6phC. Package length: %i\n",
+				header_start, skb->len - KALMIA_HEADER_LENGTH);
 
 		/* subtract start header and end header */
 		usb_packet_length = skb->len - (2 * KALMIA_HEADER_LENGTH);
@@ -310,12 +298,9 @@ kalmia_rx_fixup(struct usbnet *dev, struct sk_buff *skb)
 				sizeof(HEADER_END_OF_USB_PACKET)) == 0);
 			if (!is_last) {
 				header_start = skb->data + ether_packet_length;
-				netdev_dbg(
-					dev->net,
-					"End header: %02x:%02x:%02x:%02x:%02x:%02x. Package length: %i\n",
-					header_start[0], header_start[1],
-					header_start[2], header_start[3],
-					header_start[4], header_start[5],
+				netdev_dbg(dev->net,
+					"End header: %6phC. Package length: %i\n",
+					header_start,
 					skb->len - KALMIA_HEADER_LENGTH);
 			}
 		}
diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c
index 5645921..606eba2 100644
--- a/drivers/net/usb/qmi_wwan.c
+++ b/drivers/net/usb/qmi_wwan.c
@@ -523,6 +523,7 @@ static const struct usb_device_id products[] = {
 	{QMI_FIXED_INTF(0x19d2, 0x0002, 1)},
 	{QMI_FIXED_INTF(0x19d2, 0x0012, 1)},
 	{QMI_FIXED_INTF(0x19d2, 0x0017, 3)},
+	{QMI_FIXED_INTF(0x19d2, 0x0019, 3)},	/* ONDA MT689DC */
 	{QMI_FIXED_INTF(0x19d2, 0x0021, 4)},
 	{QMI_FIXED_INTF(0x19d2, 0x0025, 1)},
 	{QMI_FIXED_INTF(0x19d2, 0x0031, 4)},
@@ -582,6 +583,7 @@ static const struct usb_device_id products[] = {
 	{QMI_FIXED_INTF(0x1199, 0x901c, 8)},    /* Sierra Wireless EM7700 */
 	{QMI_FIXED_INTF(0x1bbb, 0x011e, 4)},	/* Telekom Speedstick LTE II (Alcatel One Touch L100V LTE) */
 	{QMI_FIXED_INTF(0x2357, 0x0201, 4)},	/* TP-LINK HSUPA Modem MA180 */
+	{QMI_FIXED_INTF(0x2357, 0x9000, 4)},	/* TP-LINK MA260 */
 	{QMI_FIXED_INTF(0x1bc7, 0x1200, 5)},	/* Telit LE920 */
 	{QMI_FIXED_INTF(0x1e2d, 0x12d1, 4)},	/* Cinterion PLxx */
 
@@ -618,6 +620,7 @@ static const struct usb_device_id products[] = {
 	{QMI_GOBI_DEVICE(0x05c6, 0x9265)},	/* Asus Gobi 2000 Modem device (VR305) */
 	{QMI_GOBI_DEVICE(0x05c6, 0x9235)},	/* Top Global Gobi 2000 Modem device (VR306) */
 	{QMI_GOBI_DEVICE(0x05c6, 0x9275)},	/* iRex Technologies Gobi 2000 Modem device (VR307) */
+	{QMI_GOBI_DEVICE(0x0af0, 0x8120)},	/* Option GTM681W */
 	{QMI_GOBI_DEVICE(0x1199, 0x68a5)},	/* Sierra Wireless Modem */
 	{QMI_GOBI_DEVICE(0x1199, 0x68a9)},	/* Sierra Wireless Modem */
 	{QMI_GOBI_DEVICE(0x1199, 0x9001)},	/* Sierra Wireless Gobi 2000 Modem device (VT773) */
@@ -631,7 +634,6 @@ static const struct usb_device_id products[] = {
 	{QMI_GOBI_DEVICE(0x1199, 0x9009)},	/* Sierra Wireless Gobi 2000 Modem device (VT773) */
 	{QMI_GOBI_DEVICE(0x1199, 0x900a)},	/* Sierra Wireless Gobi 2000 Modem device (VT773) */
 	{QMI_GOBI_DEVICE(0x1199, 0x9011)},	/* Sierra Wireless Gobi 2000 Modem device (MC8305) */
-	{QMI_FIXED_INTF(0x1199, 0x9011, 5)},	/* alternate interface number!? */
 	{QMI_GOBI_DEVICE(0x16d8, 0x8002)},	/* CMDTech Gobi 2000 Modem device (VU922) */
 	{QMI_GOBI_DEVICE(0x05c6, 0x9205)},	/* Gobi 2000 Modem device */
 	{QMI_GOBI_DEVICE(0x1199, 0x9013)},	/* Sierra Wireless Gobi 3000 Modem device (MC8355) */
diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
index 14e5198..d02bac8 100644
--- a/drivers/net/usb/r8152.c
+++ b/drivers/net/usb/r8152.c
@@ -11,7 +11,6 @@
 #include <linux/signal.h>
 #include <linux/slab.h>
 #include <linux/module.h>
-#include <linux/version.h>
 #include <linux/netdevice.h>
 #include <linux/etherdevice.h>
 #include <linux/mii.h>
@@ -1749,18 +1748,7 @@ static struct usb_driver rtl8152_driver = {
 	.resume =	rtl8152_resume
 };
 
-static int __init usb_rtl8152_init(void)
-{
-	return usb_register(&rtl8152_driver);
-}
-
-static void __exit usb_rtl8152_exit(void)
-{
-	usb_deregister(&rtl8152_driver);
-}
-
-module_init(usb_rtl8152_init);
-module_exit(usb_rtl8152_exit);
+module_usb_driver(rtl8152_driver);
 
 MODULE_AUTHOR(DRIVER_AUTHOR);
 MODULE_DESCRIPTION(DRIVER_DESC);
diff --git a/drivers/net/veth.c b/drivers/net/veth.c
index 177f911..da86652 100644
--- a/drivers/net/veth.c
+++ b/drivers/net/veth.c
@@ -379,12 +379,6 @@ static int veth_newlink(struct net *src_net, struct net_device *dev,
 	else
 		snprintf(dev->name, IFNAMSIZ, DRV_NAME "%%d");
 
-	if (strchr(dev->name, '%')) {
-		err = dev_alloc_name(dev, dev->name);
-		if (err < 0)
-			goto err_alloc_name;
-	}
-
 	err = register_netdevice(dev);
 	if (err < 0)
 		goto err_register_dev;
@@ -404,7 +398,6 @@ static int veth_newlink(struct net *src_net, struct net_device *dev,
 
 err_register_dev:
 	/* nothing to do */
-err_alloc_name:
 err_configure_peer:
 	unregister_netdevice(peer);
 	return err;
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index c9e0038..42d670a 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -602,7 +602,7 @@ static int virtnet_poll(struct napi_struct *napi, int budget)
 		container_of(napi, struct receive_queue, napi);
 	struct virtnet_info *vi = rq->vq->vdev->priv;
 	void *buf;
-	unsigned int len, received = 0;
+	unsigned int r, len, received = 0;
 
 again:
 	while (received < budget &&
@@ -619,8 +619,9 @@ again:
 
 	/* Out of packets? */
 	if (received < budget) {
+		r = virtqueue_enable_cb_prepare(rq->vq);
 		napi_complete(napi);
-		if (unlikely(!virtqueue_enable_cb(rq->vq)) &&
+		if (unlikely(virtqueue_poll(rq->vq, r)) &&
 		    napi_schedule_prep(napi)) {
 			virtqueue_disable_cb(rq->vq);
 			__napi_schedule(napi);
diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 57325f3..227b54a 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -44,6 +44,8 @@
 
 #define VXLAN_VERSION	"0.1"
 
+#define PORT_HASH_BITS	8
+#define PORT_HASH_SIZE  (1<<PORT_HASH_BITS)
 #define VNI_HASH_BITS	10
 #define VNI_HASH_SIZE	(1<<VNI_HASH_BITS)
 #define FDB_HASH_BITS	8
@@ -66,30 +68,44 @@ struct vxlanhdr {
 
 /* UDP port for VXLAN traffic.
  * The IANA assigned port is 4789, but the Linux default is 8472
- * for compatability with early adopters.
+ * for compatibility with early adopters.
  */
-static unsigned int vxlan_port __read_mostly = 8472;
-module_param_named(udp_port, vxlan_port, uint, 0444);
+static unsigned short vxlan_port __read_mostly = 8472;
+module_param_named(udp_port, vxlan_port, ushort, 0444);
 MODULE_PARM_DESC(udp_port, "Destination UDP port");
 
 static bool log_ecn_error = true;
 module_param(log_ecn_error, bool, 0644);
 MODULE_PARM_DESC(log_ecn_error, "Log packets received with corrupted ECN");
 
-/* per-net private data for this module */
-static unsigned int vxlan_net_id;
-struct vxlan_net {
-	struct socket	  *sock;	/* UDP encap socket */
+static int vxlan_net_id;
+
+static const u8 all_zeros_mac[ETH_ALEN];
+
+/* per UDP socket information */
+struct vxlan_sock {
+	struct hlist_node hlist;
+	struct rcu_head	  rcu;
+	struct work_struct del_work;
+	atomic_t	  refcnt;
+	struct socket	  *sock;
 	struct hlist_head vni_list[VNI_HASH_SIZE];
 };
 
+/* per-network namespace private data for this module */
+struct vxlan_net {
+	struct list_head  vxlan_list;
+	struct hlist_head sock_list[PORT_HASH_SIZE];
+	spinlock_t	  sock_lock;
+};
+
 struct vxlan_rdst {
-	struct rcu_head		 rcu;
 	__be32			 remote_ip;
 	__be16			 remote_port;
 	u32			 remote_vni;
 	u32			 remote_ifindex;
-	struct vxlan_rdst	*remote_next;
+	struct list_head	 list;
+	struct rcu_head		 rcu;
 };
 
 /* Forwarding table entry */
@@ -98,7 +114,7 @@ struct vxlan_fdb {
 	struct rcu_head	  rcu;
 	unsigned long	  updated;	/* jiffies */
 	unsigned long	  used;
-	struct vxlan_rdst remote;
+	struct list_head  remotes;
 	u16		  state;	/* see ndm_state */
 	u8		  flags;	/* see ndm_flags */
 	u8		  eth_addr[ETH_ALEN];
@@ -106,7 +122,9 @@ struct vxlan_fdb {
 
 /* Pseudo network device */
 struct vxlan_dev {
-	struct hlist_node hlist;
+	struct hlist_node hlist;	/* vni hash table */
+	struct list_head  next;		/* vxlan's per namespace list */
+	struct vxlan_sock *vn_sock;	/* listening socket */
 	struct net_device *dev;
 	struct vxlan_rdst default_dst;	/* default destination */
 	__be32		  saddr;	/* source address */
@@ -117,6 +135,9 @@ struct vxlan_dev {
 	__u8		  ttl;
 	u32		  flags;	/* VXLAN_F_* below */
 
+	struct work_struct sock_work;
+	struct work_struct igmp_work;
+
 	unsigned long	  age_interval;
 	struct timer_list age_timer;
 	spinlock_t	  hash_lock;
@@ -134,20 +155,55 @@ struct vxlan_dev {
 
 /* salt for hash table */
 static u32 vxlan_salt __read_mostly;
+static struct workqueue_struct *vxlan_wq;
+
+static void vxlan_sock_work(struct work_struct *work);
+
+/* Virtual Network hash table head */
+static inline struct hlist_head *vni_head(struct vxlan_sock *vs, u32 id)
+{
+	return &vs->vni_list[hash_32(id, VNI_HASH_BITS)];
+}
 
-static inline struct hlist_head *vni_head(struct net *net, u32 id)
+/* Socket hash table head */
+static inline struct hlist_head *vs_head(struct net *net, __be16 port)
 {
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 
-	return &vn->vni_list[hash_32(id, VNI_HASH_BITS)];
+	return &vn->sock_list[hash_32(ntohs(port), PORT_HASH_BITS)];
+}
+
+/* First remote destination for a forwarding entry.
+ * Guaranteed to be non-NULL because remotes are never deleted.
+ */
+static inline struct vxlan_rdst *first_remote(struct vxlan_fdb *fdb)
+{
+	return list_first_or_null_rcu(&fdb->remotes, struct vxlan_rdst, list);
+}
+
+/* Find VXLAN socket based on network namespace and UDP port */
+static struct vxlan_sock *vxlan_find_port(struct net *net, __be16 port)
+{
+	struct vxlan_sock *vs;
+
+	hlist_for_each_entry_rcu(vs, vs_head(net, port), hlist) {
+		if (inet_sk(vs->sock->sk)->inet_sport == port)
+			return vs;
+	}
+	return NULL;
 }
 
 /* Look up VNI in a per net namespace table */
-static struct vxlan_dev *vxlan_find_vni(struct net *net, u32 id)
+static struct vxlan_dev *vxlan_find_vni(struct net *net, u32 id, __be16 port)
 {
+	struct vxlan_sock *vs;
 	struct vxlan_dev *vxlan;
 
-	hlist_for_each_entry_rcu(vxlan, vni_head(net, id), hlist) {
+	vs = vxlan_find_port(net, port);
+	if (!vs)
+		return NULL;
+
+	hlist_for_each_entry_rcu(vxlan, vni_head(vs, id), hlist) {
 		if (vxlan->default_dst.remote_vni == id)
 			return vxlan;
 	}
@@ -157,9 +213,9 @@ static struct vxlan_dev *vxlan_find_vni(struct net *net, u32 id)
 
 /* Fill in neighbour message in skbuff. */
 static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
-			   const struct vxlan_fdb *fdb,
-			   u32 portid, u32 seq, int type, unsigned int flags,
-			   const struct vxlan_rdst *rdst)
+			  const struct vxlan_fdb *fdb,
+			  u32 portid, u32 seq, int type, unsigned int flags,
+			  const struct vxlan_rdst *rdst)
 {
 	unsigned long now = jiffies;
 	struct nda_cacheinfo ci;
@@ -197,7 +253,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 	    nla_put_be16(skb, NDA_PORT, rdst->remote_port))
 		goto nla_put_failure;
 	if (rdst->remote_vni != vxlan->default_dst.remote_vni &&
-	    nla_put_be32(skb, NDA_VNI, rdst->remote_vni))
+	    nla_put_u32(skb, NDA_VNI, rdst->remote_vni))
 		goto nla_put_failure;
 	if (rdst->remote_ifindex &&
 	    nla_put_u32(skb, NDA_IFINDEX, rdst->remote_ifindex))
@@ -230,7 +286,7 @@ static inline size_t vxlan_nlmsg_size(void)
 }
 
 static void vxlan_fdb_notify(struct vxlan_dev *vxlan,
-			     const struct vxlan_fdb *fdb, int type)
+			     struct vxlan_fdb *fdb, int type)
 {
 	struct net *net = dev_net(vxlan->dev);
 	struct sk_buff *skb;
@@ -240,7 +296,7 @@ static void vxlan_fdb_notify(struct vxlan_dev *vxlan,
 	if (skb == NULL)
 		goto errout;
 
-	err = vxlan_fdb_info(skb, vxlan, fdb, 0, 0, type, 0, &fdb->remote);
+	err = vxlan_fdb_info(skb, vxlan, fdb, 0, 0, type, 0, first_remote(fdb));
 	if (err < 0) {
 		/* -EMSGSIZE implies BUG in vxlan_nlmsg_size() */
 		WARN_ON(err == -EMSGSIZE);
@@ -258,22 +314,27 @@ errout:
 static void vxlan_ip_miss(struct net_device *dev, __be32 ipa)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_fdb f;
+	struct vxlan_fdb f = {
+		.state = NUD_STALE,
+	};
+	struct vxlan_rdst remote = {
+		.remote_ip = ipa, /* goes to NDA_DST */
+		.remote_vni = VXLAN_N_VID,
+	};
 
-	memset(&f, 0, sizeof f);
-	f.state = NUD_STALE;
-	f.remote.remote_ip = ipa; /* goes to NDA_DST */
-	f.remote.remote_vni = VXLAN_N_VID;
+	INIT_LIST_HEAD(&f.remotes);
+	list_add_rcu(&remote.list, &f.remotes);
 
 	vxlan_fdb_notify(vxlan, &f, RTM_GETNEIGH);
 }
 
 static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8 eth_addr[ETH_ALEN])
 {
-	struct vxlan_fdb	f;
+	struct vxlan_fdb f = {
+		.state = NUD_STALE,
+	};
 
-	memset(&f, 0, sizeof f);
-	f.state = NUD_STALE;
+	INIT_LIST_HEAD(&f.remotes);
 	memcpy(f.eth_addr, eth_addr, ETH_ALEN);
 
 	vxlan_fdb_notify(vxlan, &f, RTM_GETNEIGH);
@@ -328,21 +389,34 @@ static struct vxlan_fdb *vxlan_find_mac(struct vxlan_dev *vxlan,
 	return f;
 }
 
-/* Add/update destinations for multicast */
-static int vxlan_fdb_append(struct vxlan_fdb *f,
-			    __be32 ip, __be16 port, __u32 vni, __u32 ifindex)
+/* caller should hold vxlan->hash_lock */
+static struct vxlan_rdst *vxlan_fdb_find_rdst(struct vxlan_fdb *f,
+					      __be32 ip, __be16 port,
+					      __u32 vni, __u32 ifindex)
 {
-	struct vxlan_rdst *rd_prev, *rd;
+	struct vxlan_rdst *rd;
 
-	rd_prev = NULL;
-	for (rd = &f->remote; rd; rd = rd->remote_next) {
+	list_for_each_entry(rd, &f->remotes, list) {
 		if (rd->remote_ip == ip &&
 		    rd->remote_port == port &&
 		    rd->remote_vni == vni &&
 		    rd->remote_ifindex == ifindex)
-			return 0;
-		rd_prev = rd;
+			return rd;
 	}
+
+	return NULL;
+}
+
+/* Add/update destinations for multicast */
+static int vxlan_fdb_append(struct vxlan_fdb *f,
+			    __be32 ip, __be16 port, __u32 vni, __u32 ifindex)
+{
+	struct vxlan_rdst *rd;
+
+	rd = vxlan_fdb_find_rdst(f, ip, port, vni, ifindex);
+	if (rd)
+		return 0;
+
 	rd = kmalloc(sizeof(*rd), GFP_ATOMIC);
 	if (rd == NULL)
 		return -ENOBUFS;
@@ -350,8 +424,9 @@ static int vxlan_fdb_append(struct vxlan_fdb *f,
 	rd->remote_port = port;
 	rd->remote_vni = vni;
 	rd->remote_ifindex = ifindex;
-	rd->remote_next = NULL;
-	rd_prev->remote_next = rd;
+
+	list_add_tail_rcu(&rd->list, &f->remotes);
+
 	return 1;
 }
 
@@ -383,7 +458,8 @@ static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 			notify = 1;
 		}
 		if ((flags & NLM_F_APPEND) &&
-		    is_multicast_ether_addr(f->eth_addr)) {
+		    (is_multicast_ether_addr(f->eth_addr) ||
+		     is_zero_ether_addr(f->eth_addr))) {
 			int rc = vxlan_fdb_append(f, ip, port, vni, ifindex);
 
 			if (rc < 0)
@@ -403,16 +479,14 @@ static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 			return -ENOMEM;
 
 		notify = 1;
-		f->remote.remote_ip = ip;
-		f->remote.remote_port = port;
-		f->remote.remote_vni = vni;
-		f->remote.remote_ifindex = ifindex;
-		f->remote.remote_next = NULL;
 		f->state = state;
 		f->flags = ndm_flags;
 		f->updated = f->used = jiffies;
+		INIT_LIST_HEAD(&f->remotes);
 		memcpy(f->eth_addr, mac, ETH_ALEN);
 
+		vxlan_fdb_append(f, ip, port, vni, ifindex);
+
 		++vxlan->addrcnt;
 		hlist_add_head_rcu(&f->hlist,
 				   vxlan_fdb_head(vxlan, mac));
@@ -424,16 +498,19 @@ static int vxlan_fdb_create(struct vxlan_dev *vxlan,
 	return 0;
 }
 
+static void vxlan_fdb_free_rdst(struct rcu_head *head)
+{
+	struct vxlan_rdst *rd = container_of(head, struct vxlan_rdst, rcu);
+	kfree(rd);
+}
+
 static void vxlan_fdb_free(struct rcu_head *head)
 {
 	struct vxlan_fdb *f = container_of(head, struct vxlan_fdb, rcu);
+	struct vxlan_rdst *rd, *nd;
 
-	while (f->remote.remote_next) {
-		struct vxlan_rdst *rd = f->remote.remote_next;
-
-		f->remote.remote_next = rd->remote_next;
+	list_for_each_entry_safe(rd, nd, &f->remotes, list)
 		kfree(rd);
-	}
 	kfree(f);
 }
 
@@ -449,58 +526,77 @@ static void vxlan_fdb_destroy(struct vxlan_dev *vxlan, struct vxlan_fdb *f)
 	call_rcu(&f->rcu, vxlan_fdb_free);
 }
 
-/* Add static entry (via netlink) */
-static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
-			 struct net_device *dev,
-			 const unsigned char *addr, u16 flags)
+static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
+			   __be32 *ip, __be16 *port, u32 *vni, u32 *ifindex)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct net *net = dev_net(vxlan->dev);
-	__be32 ip;
-	__be16 port;
-	u32 vni, ifindex;
-	int err;
 
-	if (!(ndm->ndm_state & (NUD_PERMANENT|NUD_REACHABLE))) {
-		pr_info("RTM_NEWNEIGH with invalid state %#x\n",
-			ndm->ndm_state);
-		return -EINVAL;
-	}
-
-	if (tb[NDA_DST] == NULL)
-		return -EINVAL;
-
-	if (nla_len(tb[NDA_DST]) != sizeof(__be32))
-		return -EAFNOSUPPORT;
+	if (tb[NDA_DST]) {
+		if (nla_len(tb[NDA_DST]) != sizeof(__be32))
+			return -EAFNOSUPPORT;
 
-	ip = nla_get_be32(tb[NDA_DST]);
+		*ip = nla_get_be32(tb[NDA_DST]);
+	} else {
+		*ip = htonl(INADDR_ANY);
+	}
 
 	if (tb[NDA_PORT]) {
 		if (nla_len(tb[NDA_PORT]) != sizeof(__be16))
 			return -EINVAL;
-		port = nla_get_be16(tb[NDA_PORT]);
-	} else
-		port = vxlan->dst_port;
+		*port = nla_get_be16(tb[NDA_PORT]);
+	} else {
+		*port = vxlan->dst_port;
+	}
 
 	if (tb[NDA_VNI]) {
 		if (nla_len(tb[NDA_VNI]) != sizeof(u32))
 			return -EINVAL;
-		vni = nla_get_u32(tb[NDA_VNI]);
-	} else
-		vni = vxlan->default_dst.remote_vni;
+		*vni = nla_get_u32(tb[NDA_VNI]);
+	} else {
+		*vni = vxlan->default_dst.remote_vni;
+	}
 
 	if (tb[NDA_IFINDEX]) {
 		struct net_device *tdev;
 
 		if (nla_len(tb[NDA_IFINDEX]) != sizeof(u32))
 			return -EINVAL;
-		ifindex = nla_get_u32(tb[NDA_IFINDEX]);
-		tdev = dev_get_by_index(net, ifindex);
+		*ifindex = nla_get_u32(tb[NDA_IFINDEX]);
+		tdev = dev_get_by_index(net, *ifindex);
 		if (!tdev)
 			return -EADDRNOTAVAIL;
 		dev_put(tdev);
-	} else
-		ifindex = 0;
+	} else {
+		*ifindex = 0;
+	}
+
+	return 0;
+}
+
+/* Add static entry (via netlink) */
+static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
+			 struct net_device *dev,
+			 const unsigned char *addr, u16 flags)
+{
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	/* struct net *net = dev_net(vxlan->dev); */
+	__be32 ip;
+	__be16 port;
+	u32 vni, ifindex;
+	int err;
+
+	if (!(ndm->ndm_state & (NUD_PERMANENT|NUD_REACHABLE))) {
+		pr_info("RTM_NEWNEIGH with invalid state %#x\n",
+			ndm->ndm_state);
+		return -EINVAL;
+	}
+
+	if (tb[NDA_DST] == NULL)
+		return -EINVAL;
+
+	err = vxlan_fdb_parse(tb, vxlan, &ip, &port, &vni, &ifindex);
+	if (err)
+		return err;
 
 	spin_lock_bh(&vxlan->hash_lock);
 	err = vxlan_fdb_create(vxlan, addr, ip, ndm->ndm_state, flags,
@@ -517,14 +613,43 @@ static int vxlan_fdb_delete(struct ndmsg *ndm, struct nlattr *tb[],
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_fdb *f;
-	int err = -ENOENT;
+	struct vxlan_rdst *rd = NULL;
+	__be32 ip;
+	__be16 port;
+	u32 vni, ifindex;
+	int err;
+
+	err = vxlan_fdb_parse(tb, vxlan, &ip, &port, &vni, &ifindex);
+	if (err)
+		return err;
+
+	err = -ENOENT;
 
 	spin_lock_bh(&vxlan->hash_lock);
 	f = vxlan_find_mac(vxlan, addr);
-	if (f) {
-		vxlan_fdb_destroy(vxlan, f);
-		err = 0;
+	if (!f)
+		goto out;
+
+	if (ip != htonl(INADDR_ANY)) {
+		rd = vxlan_fdb_find_rdst(f, ip, port, vni, ifindex);
+		if (!rd)
+			goto out;
+	}
+
+	err = 0;
+
+	/* remove a destination if it's not the only one on the list,
+	 * otherwise destroy the fdb entry
+	 */
+	if (rd && !list_is_singular(&f->remotes)) {
+		list_del_rcu(&rd->list);
+		call_rcu(&rd->rcu, vxlan_fdb_free_rdst);
+		goto out;
 	}
+
+	vxlan_fdb_destroy(vxlan, f);
+
+out:
 	spin_unlock_bh(&vxlan->hash_lock);
 
 	return err;
@@ -543,23 +668,24 @@ static int vxlan_fdb_dump(struct sk_buff *skb, struct netlink_callback *cb,
 
 		hlist_for_each_entry_rcu(f, &vxlan->fdb_head[h], hlist) {
 			struct vxlan_rdst *rd;
-			for (rd = &f->remote; rd; rd = rd->remote_next) {
-				if (idx < cb->args[0])
-					goto skip;
 
+			if (idx < cb->args[0])
+				goto skip;
+
+			list_for_each_entry_rcu(rd, &f->remotes, list) {
 				err = vxlan_fdb_info(skb, vxlan, f,
 						     NETLINK_CB(cb->skb).portid,
 						     cb->nlh->nlmsg_seq,
 						     RTM_NEWNEIGH,
 						     NLM_F_MULTI, rd);
 				if (err < 0)
-					break;
-skip:
-				++idx;
+					goto out;
 			}
+skip:
+			++idx;
 		}
 	}
-
+out:
 	return idx;
 }
 
@@ -575,7 +701,9 @@ static bool vxlan_snoop(struct net_device *dev,
 
 	f = vxlan_find_mac(vxlan, src_mac);
 	if (likely(f)) {
-		if (likely(f->remote.remote_ip == src_ip))
+		struct vxlan_rdst *rdst = first_remote(f);
+
+		if (likely(rdst->remote_ip == src_ip))
 			return false;
 
 		/* Don't migrate static entries, drop packets */
@@ -585,10 +713,11 @@ static bool vxlan_snoop(struct net_device *dev,
 		if (net_ratelimit())
 			netdev_info(dev,
 				    "%pM migrated from %pI4 to %pI4\n",
-				    src_mac, &f->remote.remote_ip, &src_ip);
+				    src_mac, &rdst->remote_ip, &src_ip);
 
-		f->remote.remote_ip = src_ip;
+		rdst->remote_ip = src_ip;
 		f->updated = jiffies;
+		vxlan_fdb_notify(vxlan, f, RTM_NEWNEIGH);
 	} else {
 		/* learned new entry */
 		spin_lock(&vxlan->hash_lock);
@@ -609,78 +738,61 @@ static bool vxlan_snoop(struct net_device *dev,
 
 
 /* See if multicast group is already in use by other ID */
-static bool vxlan_group_used(struct vxlan_net *vn,
-			     const struct vxlan_dev *this)
+static bool vxlan_group_used(struct vxlan_net *vn, __be32 remote_ip)
 {
-	const struct vxlan_dev *vxlan;
-	unsigned h;
-
-	for (h = 0; h < VNI_HASH_SIZE; ++h)
-		hlist_for_each_entry(vxlan, &vn->vni_list[h], hlist) {
-			if (vxlan == this)
-				continue;
+	struct vxlan_dev *vxlan;
 
-			if (!netif_running(vxlan->dev))
-				continue;
+	list_for_each_entry(vxlan, &vn->vxlan_list, next) {
+		if (!netif_running(vxlan->dev))
+			continue;
 
-			if (vxlan->default_dst.remote_ip == this->default_dst.remote_ip)
-				return true;
-		}
+		if (vxlan->default_dst.remote_ip == remote_ip)
+			return true;
+	}
 
 	return false;
 }
 
-/* kernel equivalent to IP_ADD_MEMBERSHIP */
-static int vxlan_join_group(struct net_device *dev)
+static void vxlan_sock_hold(struct vxlan_sock *vs)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
-	struct sock *sk = vn->sock->sk;
-	struct ip_mreqn mreq = {
-		.imr_multiaddr.s_addr	= vxlan->default_dst.remote_ip,
-		.imr_ifindex		= vxlan->default_dst.remote_ifindex,
-	};
-	int err;
+	atomic_inc(&vs->refcnt);
+}
 
-	/* Already a member of group */
-	if (vxlan_group_used(vn, vxlan))
-		return 0;
+static void vxlan_sock_release(struct vxlan_net *vn, struct vxlan_sock *vs)
+{
+	if (!atomic_dec_and_test(&vs->refcnt))
+		return;
 
-	/* Need to drop RTNL to call multicast join */
-	rtnl_unlock();
-	lock_sock(sk);
-	err = ip_mc_join_group(sk, &mreq);
-	release_sock(sk);
-	rtnl_lock();
+	spin_lock(&vn->sock_lock);
+	hlist_del_rcu(&vs->hlist);
+	spin_unlock(&vn->sock_lock);
 
-	return err;
+	queue_work(vxlan_wq, &vs->del_work);
 }
 
-
-/* kernel equivalent to IP_DROP_MEMBERSHIP */
-static int vxlan_leave_group(struct net_device *dev)
+/* Callback to update multicast group membership.
+ * Scheduled when vxlan goes up/down.
+ */
+static void vxlan_igmp_work(struct work_struct *work)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
-	int err = 0;
-	struct sock *sk = vn->sock->sk;
+	struct vxlan_dev *vxlan = container_of(work, struct vxlan_dev, igmp_work);
+	struct vxlan_net *vn = net_generic(dev_net(vxlan->dev), vxlan_net_id);
+	struct vxlan_sock *vs = vxlan->vn_sock;
+	struct sock *sk = vs->sock->sk;
 	struct ip_mreqn mreq = {
 		.imr_multiaddr.s_addr	= vxlan->default_dst.remote_ip,
 		.imr_ifindex		= vxlan->default_dst.remote_ifindex,
 	};
 
-	/* Only leave group when last vxlan is done. */
-	if (vxlan_group_used(vn, vxlan))
-		return 0;
-
-	/* Need to drop RTNL to call multicast leave */
-	rtnl_unlock();
 	lock_sock(sk);
-	err = ip_mc_leave_group(sk, &mreq);
+	if (vxlan_group_used(vn, vxlan->default_dst.remote_ip))
+		ip_mc_join_group(sk, &mreq);
+	else
+		ip_mc_leave_group(sk, &mreq);
 	release_sock(sk);
-	rtnl_lock();
 
-	return err;
+	vxlan_sock_release(vn, vs);
+	dev_put(vxlan->dev);
 }
 
 /* Callback from net/ipv4/udp.c to receive packets */
@@ -690,6 +802,7 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
 	struct vxlanhdr *vxh;
 	struct vxlan_dev *vxlan;
 	struct pcpu_tstats *stats;
+	__be16 port;
 	__u32 vni;
 	int err;
 
@@ -713,9 +826,11 @@ static int vxlan_udp_encap_recv(struct sock *sk, struct sk_buff *skb)
 
 	/* Is this VNI defined? */
 	vni = ntohl(vxh->vx_vni) >> 8;
-	vxlan = vxlan_find_vni(sock_net(sk), vni);
+	port = inet_sk(sk)->inet_sport;
+	vxlan = vxlan_find_vni(sock_net(sk), vni, port);
 	if (!vxlan) {
-		netdev_dbg(skb->dev, "unknown vni %d\n", vni);
+		netdev_dbg(skb->dev, "unknown vni %d port %u\n",
+			   vni, ntohs(port));
 		goto drop;
 	}
 
@@ -834,7 +949,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb)
 		}
 
 		f = vxlan_find_mac(vxlan, n->ha);
-		if (f && f->remote.remote_ip == htonl(INADDR_ANY)) {
+		if (f && first_remote(f)->remote_ip == htonl(INADDR_ANY)) {
 			/* bridge-local neighbor */
 			neigh_release(n);
 			goto out;
@@ -896,7 +1011,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
 	return false;
 }
 
-static void vxlan_sock_free(struct sk_buff *skb)
+static void vxlan_sock_put(struct sk_buff *skb)
 {
 	sock_put(skb->sk);
 }
@@ -904,13 +1019,13 @@ static void vxlan_sock_free(struct sk_buff *skb)
 /* On transmit, associate with the tunnel socket */
 static void vxlan_set_owner(struct net_device *dev, struct sk_buff *skb)
 {
-	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
-	struct sock *sk = vn->sock->sk;
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct sock *sk = vxlan->vn_sock->sock->sk;
 
 	skb_orphan(skb);
 	sock_hold(sk);
 	skb->sk = sk;
-	skb->destructor = vxlan_sock_free;
+	skb->destructor = vxlan_sock_put;
 }
 
 /* Compute source port for outgoing packet
@@ -976,21 +1091,21 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
 	}
 }
 
-static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
-				  struct vxlan_rdst *rdst, bool did_rsc)
+static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
+			   struct vxlan_rdst *rdst, bool did_rsc)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct rtable *rt;
 	const struct iphdr *old_iph;
-	struct iphdr *iph;
 	struct vxlanhdr *vxh;
 	struct udphdr *uh;
 	struct flowi4 fl4;
 	__be32 dst;
 	__be16 src_port, dst_port;
-        u32 vni;
+	u32 vni;
 	__be16 df = 0;
 	__u8 tos, ttl;
+	int err;
 
 	dst_port = rdst->remote_port ? rdst->remote_port : vxlan->dst_port;
 	vni = rdst->remote_vni;
@@ -1000,7 +1115,7 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (did_rsc) {
 			/* short-circuited back to local bridge */
 			vxlan_encap_bypass(skb, vxlan, vxlan);
-			return NETDEV_TX_OK;
+			return;
 		}
 		goto drop;
 	}
@@ -1052,19 +1167,12 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		struct vxlan_dev *dst_vxlan;
 
 		ip_rt_put(rt);
-		dst_vxlan = vxlan_find_vni(dev_net(dev), vni);
+		dst_vxlan = vxlan_find_vni(dev_net(dev), vni, dst_port);
 		if (!dst_vxlan)
 			goto tx_error;
 		vxlan_encap_bypass(skb, vxlan, dst_vxlan);
-		return NETDEV_TX_OK;
+		return;
 	}
-
-	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
-	IPCB(skb)->flags &= ~(IPSKB_XFRM_TUNNEL_SIZE | IPSKB_XFRM_TRANSFORMED |
-			      IPSKB_REROUTED);
-	skb_dst_drop(skb);
-	skb_dst_set(skb, &rt->dst);
-
 	vxh = (struct vxlanhdr *) __skb_push(skb, sizeof(*vxh));
 	vxh->vx_flags = htonl(VXLAN_FLAGS);
 	vxh->vx_vni = htonl(vni << 8);
@@ -1079,28 +1187,19 @@ static netdev_tx_t vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 	uh->len = htons(skb->len);
 	uh->check = 0;
 
-	__skb_push(skb, sizeof(*iph));
-	skb_reset_network_header(skb);
-	iph		= ip_hdr(skb);
-	iph->version	= 4;
-	iph->ihl	= sizeof(struct iphdr) >> 2;
-	iph->frag_off	= df;
-	iph->protocol	= IPPROTO_UDP;
-	iph->tos	= ip_tunnel_ecn_encap(tos, old_iph, skb);
-	iph->daddr	= dst;
-	iph->saddr	= fl4.saddr;
-	iph->ttl	= ttl ? : ip4_dst_hoplimit(&rt->dst);
-	tunnel_ip_select_ident(skb, old_iph, &rt->dst);
-
-	nf_reset(skb);
-
 	vxlan_set_owner(dev, skb);
 
 	if (handle_offloads(skb))
 		goto drop;
 
-	iptunnel_xmit(skb, dev);
-	return NETDEV_TX_OK;
+	tos = ip_tunnel_ecn_encap(tos, old_iph, skb);
+	ttl = ttl ? : ip4_dst_hoplimit(&rt->dst);
+
+	err = iptunnel_xmit(dev_net(dev), rt, skb, fl4.saddr, dst,
+			    IPPROTO_UDP, tos, ttl, df);
+	iptunnel_xmit_stats(err, &dev->stats, dev->tstats);
+
+	return;
 
 drop:
 	dev->stats.tx_dropped++;
@@ -1110,7 +1209,6 @@ tx_error:
 	dev->stats.tx_errors++;
 tx_free:
 	dev_kfree_skb(skb);
-	return NETDEV_TX_OK;
 }
 
 /* Transmit local packets over Vxlan
@@ -1124,9 +1222,8 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct ethhdr *eth;
 	bool did_rsc = false;
-	struct vxlan_rdst *rdst0, *rdst;
+	struct vxlan_rdst *rdst;
 	struct vxlan_fdb *f;
-	int rc1, rc;
 
 	skb_reset_mac_header(skb);
 	eth = eth_hdr(skb);
@@ -1145,33 +1242,28 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 	}
 
 	if (f == NULL) {
-		rdst0 = &vxlan->default_dst;
-
-		if (rdst0->remote_ip == htonl(INADDR_ANY) &&
-		    (vxlan->flags & VXLAN_F_L2MISS) &&
-		    !is_multicast_ether_addr(eth->h_dest))
-			vxlan_fdb_miss(vxlan, eth->h_dest);
-	} else
-		rdst0 = &f->remote;
-
-	rc = NETDEV_TX_OK;
+		f = vxlan_find_mac(vxlan, all_zeros_mac);
+		if (f == NULL) {
+			if ((vxlan->flags & VXLAN_F_L2MISS) &&
+			    !is_multicast_ether_addr(eth->h_dest))
+				vxlan_fdb_miss(vxlan, eth->h_dest);
+
+			dev->stats.tx_dropped++;
+			dev_kfree_skb(skb);
+			return NETDEV_TX_OK;
+		}
+	}
 
-	/* if there are multiple destinations, send copies */
-	for (rdst = rdst0->remote_next; rdst; rdst = rdst->remote_next) {
+	list_for_each_entry_rcu(rdst, &f->remotes, list) {
 		struct sk_buff *skb1;
 
 		skb1 = skb_clone(skb, GFP_ATOMIC);
-		if (skb1) {
-			rc1 = vxlan_xmit_one(skb1, dev, rdst, did_rsc);
-			if (rc == NETDEV_TX_OK)
-				rc = rc1;
-		}
+		if (skb1)
+			vxlan_xmit_one(skb1, dev, rdst, did_rsc);
 	}
 
-	rc1 = vxlan_xmit_one(skb, dev, rdst0, did_rsc);
-	if (rc == NETDEV_TX_OK)
-		rc = rc1;
-	return rc;
+	dev_kfree_skb(skb);
+	return NETDEV_TX_OK;
 }
 
 /* Walk the forwarding table and purge stale entries */
@@ -1214,23 +1306,70 @@ static void vxlan_cleanup(unsigned long arg)
 /* Setup stats when device is created */
 static int vxlan_init(struct net_device *dev)
 {
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
+	struct vxlan_sock *vs;
+	__u32 vni = vxlan->default_dst.remote_vni;
+
 	dev->tstats = alloc_percpu(struct pcpu_tstats);
 	if (!dev->tstats)
 		return -ENOMEM;
 
+	spin_lock(&vn->sock_lock);
+	vs = vxlan_find_port(dev_net(dev), vxlan->dst_port);
+	if (vs) {
+		/* If we have a socket with same port already, reuse it */
+		atomic_inc(&vs->refcnt);
+		vxlan->vn_sock = vs;
+		hlist_add_head_rcu(&vxlan->hlist, vni_head(vs, vni));
+	} else {
+		/* otherwise make new socket outside of RTNL */
+		dev_hold(dev);
+		queue_work(vxlan_wq, &vxlan->sock_work);
+	}
+	spin_unlock(&vn->sock_lock);
+
 	return 0;
 }
 
+static void vxlan_fdb_delete_default(struct vxlan_dev *vxlan)
+{
+	struct vxlan_fdb *f;
+
+	spin_lock_bh(&vxlan->hash_lock);
+	f = __vxlan_find_mac(vxlan, all_zeros_mac);
+	if (f)
+		vxlan_fdb_destroy(vxlan, f);
+	spin_unlock_bh(&vxlan->hash_lock);
+}
+
+static void vxlan_uninit(struct net_device *dev)
+{
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
+	struct vxlan_sock *vs = vxlan->vn_sock;
+
+	vxlan_fdb_delete_default(vxlan);
+
+	if (vs)
+		vxlan_sock_release(vn, vs);
+	free_percpu(dev->tstats);
+}
+
 /* Start ageing timer and join group when device is brought up */
 static int vxlan_open(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	int err;
+	struct vxlan_sock *vs = vxlan->vn_sock;
+
+	/* socket hasn't been created */
+	if (!vs)
+		return -ENOTCONN;
 
 	if (IN_MULTICAST(ntohl(vxlan->default_dst.remote_ip))) {
-		err = vxlan_join_group(dev);
-		if (err)
-			return err;
+		vxlan_sock_hold(vs);
+		dev_hold(dev);
+		queue_work(vxlan_wq, &vxlan->igmp_work);
 	}
 
 	if (vxlan->age_interval)
@@ -1242,7 +1381,7 @@ static int vxlan_open(struct net_device *dev)
 /* Purge the forwarding table */
 static void vxlan_flush(struct vxlan_dev *vxlan)
 {
-	unsigned h;
+	unsigned int h;
 
 	spin_lock_bh(&vxlan->hash_lock);
 	for (h = 0; h < FDB_HASH_SIZE; ++h) {
@@ -1250,7 +1389,9 @@ static void vxlan_flush(struct vxlan_dev *vxlan)
 		hlist_for_each_safe(p, n, &vxlan->fdb_head[h]) {
 			struct vxlan_fdb *f
 				= container_of(p, struct vxlan_fdb, hlist);
-			vxlan_fdb_destroy(vxlan, f);
+			/* the all_zeros_mac entry is deleted at vxlan_uninit */
+			if (!is_zero_ether_addr(f->eth_addr))
+				vxlan_fdb_destroy(vxlan, f);
 		}
 	}
 	spin_unlock_bh(&vxlan->hash_lock);
@@ -1260,9 +1401,13 @@ static void vxlan_flush(struct vxlan_dev *vxlan)
 static int vxlan_stop(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct vxlan_sock *vs = vxlan->vn_sock;
 
-	if (IN_MULTICAST(ntohl(vxlan->default_dst.remote_ip)))
-		vxlan_leave_group(dev);
+	if (vs && IN_MULTICAST(ntohl(vxlan->default_dst.remote_ip))) {
+		vxlan_sock_hold(vs);
+		dev_hold(dev);
+		queue_work(vxlan_wq, &vxlan->igmp_work);
+	}
 
 	del_timer_sync(&vxlan->age_timer);
 
@@ -1278,6 +1423,7 @@ static void vxlan_set_multicast_list(struct net_device *dev)
 
 static const struct net_device_ops vxlan_netdev_ops = {
 	.ndo_init		= vxlan_init,
+	.ndo_uninit		= vxlan_uninit,
 	.ndo_open		= vxlan_open,
 	.ndo_stop		= vxlan_stop,
 	.ndo_start_xmit		= vxlan_xmit,
@@ -1296,17 +1442,11 @@ static struct device_type vxlan_type = {
 	.name = "vxlan",
 };
 
-static void vxlan_free(struct net_device *dev)
-{
-	free_percpu(dev->tstats);
-	free_netdev(dev);
-}
-
 /* Initialize the device structure. */
 static void vxlan_setup(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	unsigned h;
+	unsigned int h;
 	int low, high;
 
 	eth_hw_addr_random(dev);
@@ -1314,7 +1454,7 @@ static void vxlan_setup(struct net_device *dev)
 	dev->hard_header_len = ETH_HLEN + VXLAN_HEADROOM;
 
 	dev->netdev_ops = &vxlan_netdev_ops;
-	dev->destructor = vxlan_free;
+	dev->destructor = free_netdev;
 	SET_NETDEV_DEVTYPE(dev, &vxlan_type);
 
 	dev->tx_queue_len = 0;
@@ -1329,7 +1469,10 @@ static void vxlan_setup(struct net_device *dev)
 	dev->priv_flags	&= ~IFF_XMIT_DST_RELEASE;
 	dev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
 
+	INIT_LIST_HEAD(&vxlan->next);
 	spin_lock_init(&vxlan->hash_lock);
+	INIT_WORK(&vxlan->igmp_work, vxlan_igmp_work);
+	INIT_WORK(&vxlan->sock_work, vxlan_sock_work);
 
 	init_timer_deferrable(&vxlan->age_timer);
 	vxlan->age_timer.function = vxlan_cleanup;
@@ -1413,9 +1556,113 @@ static const struct ethtool_ops vxlan_ethtool_ops = {
 	.get_link	= ethtool_op_get_link,
 };
 
+static void vxlan_del_work(struct work_struct *work)
+{
+	struct vxlan_sock *vs = container_of(work, struct vxlan_sock, del_work);
+
+	sk_release_kernel(vs->sock->sk);
+	kfree_rcu(vs, rcu);
+}
+
+static struct vxlan_sock *vxlan_socket_create(struct net *net, __be16 port)
+{
+	struct vxlan_sock *vs;
+	struct sock *sk;
+	struct sockaddr_in vxlan_addr = {
+		.sin_family = AF_INET,
+		.sin_addr.s_addr = htonl(INADDR_ANY),
+		.sin_port = port,
+	};
+	int rc;
+	unsigned int h;
+
+	vs = kmalloc(sizeof(*vs), GFP_KERNEL);
+	if (!vs)
+		return ERR_PTR(-ENOMEM);
+
+	for (h = 0; h < VNI_HASH_SIZE; ++h)
+		INIT_HLIST_HEAD(&vs->vni_list[h]);
+
+	INIT_WORK(&vs->del_work, vxlan_del_work);
+
+	/* Create UDP socket for encapsulation receive. */
+	rc = sock_create_kern(AF_INET, SOCK_DGRAM, IPPROTO_UDP, &vs->sock);
+	if (rc < 0) {
+		pr_debug("UDP socket create failed\n");
+		kfree(vs);
+		return ERR_PTR(rc);
+	}
+
+	/* Put in proper namespace */
+	sk = vs->sock->sk;
+	sk_change_net(sk, net);
+
+	rc = kernel_bind(vs->sock, (struct sockaddr *) &vxlan_addr,
+			 sizeof(vxlan_addr));
+	if (rc < 0) {
+		pr_debug("bind for UDP socket %pI4:%u (%d)\n",
+			 &vxlan_addr.sin_addr, ntohs(vxlan_addr.sin_port), rc);
+		sk_release_kernel(sk);
+		kfree(vs);
+		return ERR_PTR(rc);
+	}
+
+	/* Disable multicast loopback */
+	inet_sk(sk)->mc_loop = 0;
+
+	/* Mark socket as an encapsulation socket. */
+	udp_sk(sk)->encap_type = 1;
+	udp_sk(sk)->encap_rcv = vxlan_udp_encap_recv;
+	udp_encap_enable();
+	atomic_set(&vs->refcnt, 1);
+
+	return vs;
+}
+
+/* Scheduled at device creation to bind to a socket */
+static void vxlan_sock_work(struct work_struct *work)
+{
+	struct vxlan_dev *vxlan
+		= container_of(work, struct vxlan_dev, sock_work);
+	struct net_device *dev = vxlan->dev;
+	struct net *net = dev_net(dev);
+	__u32 vni = vxlan->default_dst.remote_vni;
+	__be16 port = vxlan->dst_port;
+	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
+	struct vxlan_sock *nvs, *ovs;
+
+	nvs = vxlan_socket_create(net, port);
+	if (IS_ERR(nvs)) {
+		netdev_err(vxlan->dev, "Can not create UDP socket, %ld\n",
+			   PTR_ERR(nvs));
+		goto out;
+	}
+
+	spin_lock(&vn->sock_lock);
+	/* Look again to see if can reuse socket */
+	ovs = vxlan_find_port(net, port);
+	if (ovs) {
+		atomic_inc(&ovs->refcnt);
+		vxlan->vn_sock = ovs;
+		hlist_add_head_rcu(&vxlan->hlist, vni_head(ovs, vni));
+		spin_unlock(&vn->sock_lock);
+
+		sk_release_kernel(nvs->sock->sk);
+		kfree(nvs);
+	} else {
+		vxlan->vn_sock = nvs;
+		hlist_add_head_rcu(&nvs->hlist, vs_head(net, port));
+		hlist_add_head_rcu(&vxlan->hlist, vni_head(nvs, vni));
+		spin_unlock(&vn->sock_lock);
+	}
+out:
+	dev_put(dev);
+}
+
 static int vxlan_newlink(struct net *net, struct net_device *dev,
 			 struct nlattr *tb[], struct nlattr *data[])
 {
+	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_rdst *dst = &vxlan->default_dst;
 	__u32 vni;
@@ -1425,10 +1672,6 @@ static int vxlan_newlink(struct net *net, struct net_device *dev,
 		return -EINVAL;
 
 	vni = nla_get_u32(data[IFLA_VXLAN_ID]);
-	if (vxlan_find_vni(net, vni)) {
-		pr_info("duplicate VNI %u\n", vni);
-		return -EEXIST;
-	}
 	dst->remote_vni = vni;
 
 	if (data[IFLA_VXLAN_GROUP])
@@ -1494,13 +1737,32 @@ static int vxlan_newlink(struct net *net, struct net_device *dev,
 	if (data[IFLA_VXLAN_PORT])
 		vxlan->dst_port = nla_get_be16(data[IFLA_VXLAN_PORT]);
 
+	if (vxlan_find_vni(net, vni, vxlan->dst_port)) {
+		pr_info("duplicate VNI %u\n", vni);
+		return -EEXIST;
+	}
+
 	SET_ETHTOOL_OPS(dev, &vxlan_ethtool_ops);
 
+	/* create an fdb entry for default destination */
+	err = vxlan_fdb_create(vxlan, all_zeros_mac,
+			       vxlan->default_dst.remote_ip,
+			       NUD_REACHABLE|NUD_PERMANENT,
+			       NLM_F_EXCL|NLM_F_CREATE,
+			       vxlan->dst_port, vxlan->default_dst.remote_vni,
+			       vxlan->default_dst.remote_ifindex, NTF_SELF);
+	if (err)
+		return err;
+
 	err = register_netdevice(dev);
-	if (!err)
-		hlist_add_head_rcu(&vxlan->hlist, vni_head(net, dst->remote_vni));
+	if (err) {
+		vxlan_fdb_delete_default(vxlan);
+		return err;
+	}
 
-	return err;
+	list_add(&vxlan->next, &vn->vxlan_list);
+
+	return 0;
 }
 
 static void vxlan_dellink(struct net_device *dev, struct list_head *head)
@@ -1508,7 +1770,7 @@ static void vxlan_dellink(struct net_device *dev, struct list_head *head)
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 
 	hlist_del_rcu(&vxlan->hlist);
-
+	list_del(&vxlan->next);
 	unregister_netdevice_queue(dev, head);
 }
 
@@ -1595,46 +1857,13 @@ static struct rtnl_link_ops vxlan_link_ops __read_mostly = {
 static __net_init int vxlan_init_net(struct net *net)
 {
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
-	struct sock *sk;
-	struct sockaddr_in vxlan_addr = {
-		.sin_family = AF_INET,
-		.sin_addr.s_addr = htonl(INADDR_ANY),
-	};
-	int rc;
-	unsigned h;
-
-	/* Create UDP socket for encapsulation receive. */
-	rc = sock_create_kern(AF_INET, SOCK_DGRAM, IPPROTO_UDP, &vn->sock);
-	if (rc < 0) {
-		pr_debug("UDP socket create failed\n");
-		return rc;
-	}
-	/* Put in proper namespace */
-	sk = vn->sock->sk;
-	sk_change_net(sk, net);
-
-	vxlan_addr.sin_port = htons(vxlan_port);
-
-	rc = kernel_bind(vn->sock, (struct sockaddr *) &vxlan_addr,
-			 sizeof(vxlan_addr));
-	if (rc < 0) {
-		pr_debug("bind for UDP socket %pI4:%u (%d)\n",
-			 &vxlan_addr.sin_addr, ntohs(vxlan_addr.sin_port), rc);
-		sk_release_kernel(sk);
-		vn->sock = NULL;
-		return rc;
-	}
-
-	/* Disable multicast loopback */
-	inet_sk(sk)->mc_loop = 0;
+	unsigned int h;
 
-	/* Mark socket as an encapsulation socket. */
-	udp_sk(sk)->encap_type = 1;
-	udp_sk(sk)->encap_rcv = vxlan_udp_encap_recv;
-	udp_encap_enable();
+	INIT_LIST_HEAD(&vn->vxlan_list);
+	spin_lock_init(&vn->sock_lock);
 
-	for (h = 0; h < VNI_HASH_SIZE; ++h)
-		INIT_HLIST_HEAD(&vn->vni_list[h]);
+	for (h = 0; h < PORT_HASH_SIZE; ++h)
+		INIT_HLIST_HEAD(&vn->sock_list[h]);
 
 	return 0;
 }
@@ -1643,18 +1872,11 @@ static __net_exit void vxlan_exit_net(struct net *net)
 {
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan;
-	unsigned h;
 
 	rtnl_lock();
-	for (h = 0; h < VNI_HASH_SIZE; ++h)
-		hlist_for_each_entry(vxlan, &vn->vni_list[h], hlist)
-			dev_close(vxlan->dev);
+	list_for_each_entry(vxlan, &vn->vxlan_list, next)
+		dev_close(vxlan->dev);
 	rtnl_unlock();
-
-	if (vn->sock) {
-		sk_release_kernel(vn->sock->sk);
-		vn->sock = NULL;
-	}
 }
 
 static struct pernet_operations vxlan_net_ops = {
@@ -1668,6 +1890,10 @@ static int __init vxlan_init_module(void)
 {
 	int rc;
 
+	vxlan_wq = alloc_workqueue("vxlan", 0, 0);
+	if (!vxlan_wq)
+		return -ENOMEM;
+
 	get_random_bytes(&vxlan_salt, sizeof(vxlan_salt));
 
 	rc = register_pernet_device(&vxlan_net_ops);
@@ -1683,14 +1909,16 @@ static int __init vxlan_init_module(void)
 out2:
 	unregister_pernet_device(&vxlan_net_ops);
 out1:
+	destroy_workqueue(vxlan_wq);
 	return rc;
 }
-module_init(vxlan_init_module);
+late_initcall(vxlan_init_module);
 
 static void __exit vxlan_cleanup_module(void)
 {
-	rtnl_link_unregister(&vxlan_link_ops);
 	unregister_pernet_device(&vxlan_net_ops);
+	rtnl_link_unregister(&vxlan_link_ops);
+	destroy_workqueue(vxlan_wq);
 	rcu_barrier();
 }
 module_exit(vxlan_cleanup_module);
diff --git a/drivers/net/wan/dlci.c b/drivers/net/wan/dlci.c
index 6a8a382..0d1c759 100644
--- a/drivers/net/wan/dlci.c
+++ b/drivers/net/wan/dlci.c
@@ -493,7 +493,7 @@ static void dlci_setup(struct net_device *dev)
 static int dlci_dev_event(struct notifier_block *unused,
 			  unsigned long event, void *ptr)
 {
-	struct net_device *dev = (struct net_device *) ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 
 	if (dev_net(dev) != &init_net)
 		return NOTIFY_DONE;
diff --git a/drivers/net/wan/hdlc.c b/drivers/net/wan/hdlc.c
index a0a932c..9c33ca9 100644
--- a/drivers/net/wan/hdlc.c
+++ b/drivers/net/wan/hdlc.c
@@ -99,7 +99,7 @@ static inline void hdlc_proto_stop(struct net_device *dev)
 static int hdlc_device_event(struct notifier_block *this, unsigned long event,
 			     void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	hdlc_device *hdlc;
 	unsigned long flags;
 	int on;
diff --git a/drivers/net/wan/ixp4xx_hss.c b/drivers/net/wan/ixp4xx_hss.c
index fc9d11d..e7bbdb7 100644
--- a/drivers/net/wan/ixp4xx_hss.c
+++ b/drivers/net/wan/ixp4xx_hss.c
@@ -1384,7 +1384,6 @@ static int hss_remove_one(struct platform_device *pdev)
 	unregister_hdlc_device(port->netdev);
 	free_netdev(port->netdev);
 	npe_release(port->npe);
-	platform_set_drvdata(pdev, NULL);
 	kfree(port);
 	return 0;
 }
diff --git a/drivers/net/wan/lapbether.c b/drivers/net/wan/lapbether.c
index a73b49e..a33a46f 100644
--- a/drivers/net/wan/lapbether.c
+++ b/drivers/net/wan/lapbether.c
@@ -370,7 +370,7 @@ static int lapbeth_device_event(struct notifier_block *this,
 				unsigned long event, void *ptr)
 {
 	struct lapbethdev *lapbeth;
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 
 	if (dev_net(dev) != &init_net)
 		return NOTIFY_DONE;
diff --git a/drivers/net/wireless/Kconfig b/drivers/net/wireless/Kconfig
index f8f0156..200020e 100644
--- a/drivers/net/wireless/Kconfig
+++ b/drivers/net/wireless/Kconfig
@@ -280,5 +280,6 @@ source "drivers/net/wireless/rtlwifi/Kconfig"
 source "drivers/net/wireless/ti/Kconfig"
 source "drivers/net/wireless/zd1211rw/Kconfig"
 source "drivers/net/wireless/mwifiex/Kconfig"
+source "drivers/net/wireless/cw1200/Kconfig"
 
 endif # WLAN
diff --git a/drivers/net/wireless/Makefile b/drivers/net/wireless/Makefile
index 67156ef..0fab227 100644
--- a/drivers/net/wireless/Makefile
+++ b/drivers/net/wireless/Makefile
@@ -57,3 +57,5 @@ obj-$(CONFIG_MWIFIEX)	+= mwifiex/
 
 obj-$(CONFIG_BRCMFMAC)	+= brcm80211/
 obj-$(CONFIG_BRCMSMAC)	+= brcm80211/
+
+obj-$(CONFIG_CW1200)	+= cw1200/
diff --git a/drivers/net/wireless/ath/Kconfig b/drivers/net/wireless/ath/Kconfig
index 2c02b4e..1abf1d4 100644
--- a/drivers/net/wireless/ath/Kconfig
+++ b/drivers/net/wireless/ath/Kconfig
@@ -31,5 +31,6 @@ source "drivers/net/wireless/ath/carl9170/Kconfig"
 source "drivers/net/wireless/ath/ath6kl/Kconfig"
 source "drivers/net/wireless/ath/ar5523/Kconfig"
 source "drivers/net/wireless/ath/wil6210/Kconfig"
+source "drivers/net/wireless/ath/ath10k/Kconfig"
 
 endif
diff --git a/drivers/net/wireless/ath/Makefile b/drivers/net/wireless/ath/Makefile
index 97b964d..fb05cfd 100644
--- a/drivers/net/wireless/ath/Makefile
+++ b/drivers/net/wireless/ath/Makefile
@@ -4,6 +4,7 @@ obj-$(CONFIG_CARL9170)		+= carl9170/
 obj-$(CONFIG_ATH6KL)		+= ath6kl/
 obj-$(CONFIG_AR5523)		+= ar5523/
 obj-$(CONFIG_WIL6210)		+= wil6210/
+obj-$(CONFIG_ATH10K)		+= ath10k/
 
 obj-$(CONFIG_ATH_COMMON)	+= ath.o
 
diff --git a/drivers/net/wireless/ath/ath.h b/drivers/net/wireless/ath/ath.h
index 4521342..daeafef 100644
--- a/drivers/net/wireless/ath/ath.h
+++ b/drivers/net/wireless/ath/ath.h
@@ -239,13 +239,12 @@ enum ATH_DEBUG {
 	ATH_DBG_CONFIG		= 0x00000200,
 	ATH_DBG_FATAL		= 0x00000400,
 	ATH_DBG_PS		= 0x00000800,
-	ATH_DBG_HWTIMER		= 0x00001000,
-	ATH_DBG_BTCOEX		= 0x00002000,
-	ATH_DBG_WMI		= 0x00004000,
-	ATH_DBG_BSTUCK		= 0x00008000,
-	ATH_DBG_MCI		= 0x00010000,
-	ATH_DBG_DFS		= 0x00020000,
-	ATH_DBG_WOW		= 0x00040000,
+	ATH_DBG_BTCOEX		= 0x00001000,
+	ATH_DBG_WMI		= 0x00002000,
+	ATH_DBG_BSTUCK		= 0x00004000,
+	ATH_DBG_MCI		= 0x00008000,
+	ATH_DBG_DFS		= 0x00010000,
+	ATH_DBG_WOW		= 0x00020000,
 	ATH_DBG_ANY		= 0xffffffff
 };
 
diff --git a/drivers/net/wireless/ath/ath10k/Kconfig b/drivers/net/wireless/ath/ath10k/Kconfig
new file mode 100644
index 0000000..cde58fe
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/Kconfig
@@ -0,0 +1,39 @@
+config ATH10K
+        tristate "Atheros 802.11ac wireless cards support"
+        depends on MAC80211
+	select ATH_COMMON
+        ---help---
+          This module adds support for wireless adapters based on
+          Atheros IEEE 802.11ac family of chipsets.
+
+          If you choose to build a module, it'll be called ath10k.
+
+config ATH10K_PCI
+	tristate "Atheros ath10k PCI support"
+	depends on ATH10K && PCI
+	---help---
+	  This module adds support for PCIE bus
+
+config ATH10K_DEBUG
+	bool "Atheros ath10k debugging"
+	depends on ATH10K
+	---help---
+	  Enables debug support
+
+	  If unsure, say Y to make it easier to debug problems.
+
+config ATH10K_DEBUGFS
+	bool "Atheros ath10k debugfs support"
+	depends on ATH10K
+	---help---
+	  Enabled debugfs support
+
+	  If unsure, say Y to make it easier to debug problems.
+
+config ATH10K_TRACING
+	bool "Atheros ath10k tracing support"
+	depends on ATH10K
+	depends on EVENT_TRACING
+	---help---
+	  Select this to ath10k use tracing infrastructure.
+
diff --git a/drivers/net/wireless/ath/ath10k/Makefile b/drivers/net/wireless/ath/ath10k/Makefile
new file mode 100644
index 0000000..a4179f4
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/Makefile
@@ -0,0 +1,20 @@
+obj-$(CONFIG_ATH10K) += ath10k_core.o
+ath10k_core-y += mac.o \
+		 debug.o \
+		 core.o \
+		 htc.o \
+		 htt.o \
+		 htt_rx.o \
+		 htt_tx.o \
+		 txrx.o \
+		 wmi.o \
+		 bmi.o
+
+ath10k_core-$(CONFIG_ATH10K_TRACING) += trace.o
+
+obj-$(CONFIG_ATH10K_PCI) += ath10k_pci.o
+ath10k_pci-y += pci.o \
+		ce.o
+
+# for tracing framework to find trace.h
+CFLAGS_trace.o := -I$(src)
diff --git a/drivers/net/wireless/ath/ath10k/bmi.c b/drivers/net/wireless/ath/ath10k/bmi.c
new file mode 100644
index 0000000..1a2ef51
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/bmi.c
@@ -0,0 +1,295 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "bmi.h"
+#include "hif.h"
+#include "debug.h"
+#include "htc.h"
+
+int ath10k_bmi_done(struct ath10k *ar)
+{
+	struct bmi_cmd cmd;
+	u32 cmdlen = sizeof(cmd.id) + sizeof(cmd.done);
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_dbg(ATH10K_DBG_CORE, "%s skipped\n", __func__);
+		return 0;
+	}
+
+	ar->bmi.done_sent = true;
+	cmd.id = __cpu_to_le32(BMI_DONE);
+
+	ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, cmdlen, NULL, NULL);
+	if (ret) {
+		ath10k_warn("unable to write to the device: %d\n", ret);
+		return ret;
+	}
+
+	ath10k_dbg(ATH10K_DBG_CORE, "BMI done\n");
+	return 0;
+}
+
+int ath10k_bmi_get_target_info(struct ath10k *ar,
+			       struct bmi_target_info *target_info)
+{
+	struct bmi_cmd cmd;
+	union bmi_resp resp;
+	u32 cmdlen = sizeof(cmd.id) + sizeof(cmd.get_target_info);
+	u32 resplen = sizeof(resp.get_target_info);
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_warn("BMI Get Target Info Command disallowed\n");
+		return -EBUSY;
+	}
+
+	cmd.id = __cpu_to_le32(BMI_GET_TARGET_INFO);
+
+	ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, cmdlen, &resp, &resplen);
+	if (ret) {
+		ath10k_warn("unable to get target info from device\n");
+		return ret;
+	}
+
+	if (resplen < sizeof(resp.get_target_info)) {
+		ath10k_warn("invalid get_target_info response length (%d)\n",
+			    resplen);
+		return -EIO;
+	}
+
+	target_info->version = __le32_to_cpu(resp.get_target_info.version);
+	target_info->type    = __le32_to_cpu(resp.get_target_info.type);
+	return 0;
+}
+
+int ath10k_bmi_read_memory(struct ath10k *ar,
+			   u32 address, void *buffer, u32 length)
+{
+	struct bmi_cmd cmd;
+	union bmi_resp resp;
+	u32 cmdlen = sizeof(cmd.id) + sizeof(cmd.read_mem);
+	u32 rxlen;
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_warn("command disallowed\n");
+		return -EBUSY;
+	}
+
+	ath10k_dbg(ATH10K_DBG_CORE,
+		   "%s: (device: 0x%p, address: 0x%x, length: %d)\n",
+		   __func__, ar, address, length);
+
+	while (length) {
+		rxlen = min_t(u32, length, BMI_MAX_DATA_SIZE);
+
+		cmd.id            = __cpu_to_le32(BMI_READ_MEMORY);
+		cmd.read_mem.addr = __cpu_to_le32(address);
+		cmd.read_mem.len  = __cpu_to_le32(rxlen);
+
+		ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, cmdlen,
+						  &resp, &rxlen);
+		if (ret) {
+			ath10k_warn("unable to read from the device\n");
+			return ret;
+		}
+
+		memcpy(buffer, resp.read_mem.payload, rxlen);
+		address += rxlen;
+		buffer  += rxlen;
+		length  -= rxlen;
+	}
+
+	return 0;
+}
+
+int ath10k_bmi_write_memory(struct ath10k *ar,
+			    u32 address, const void *buffer, u32 length)
+{
+	struct bmi_cmd cmd;
+	u32 hdrlen = sizeof(cmd.id) + sizeof(cmd.write_mem);
+	u32 txlen;
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_warn("command disallowed\n");
+		return -EBUSY;
+	}
+
+	ath10k_dbg(ATH10K_DBG_CORE,
+		   "%s: (device: 0x%p, address: 0x%x, length: %d)\n",
+		   __func__, ar, address, length);
+
+	while (length) {
+		txlen = min(length, BMI_MAX_DATA_SIZE - hdrlen);
+
+		/* copy before roundup to avoid reading beyond buffer*/
+		memcpy(cmd.write_mem.payload, buffer, txlen);
+		txlen = roundup(txlen, 4);
+
+		cmd.id             = __cpu_to_le32(BMI_WRITE_MEMORY);
+		cmd.write_mem.addr = __cpu_to_le32(address);
+		cmd.write_mem.len  = __cpu_to_le32(txlen);
+
+		ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, hdrlen + txlen,
+						  NULL, NULL);
+		if (ret) {
+			ath10k_warn("unable to write to the device\n");
+			return ret;
+		}
+
+		/* fixup roundup() so `length` zeroes out for last chunk */
+		txlen = min(txlen, length);
+
+		address += txlen;
+		buffer  += txlen;
+		length  -= txlen;
+	}
+
+	return 0;
+}
+
+int ath10k_bmi_execute(struct ath10k *ar, u32 address, u32 *param)
+{
+	struct bmi_cmd cmd;
+	union bmi_resp resp;
+	u32 cmdlen = sizeof(cmd.id) + sizeof(cmd.execute);
+	u32 resplen = sizeof(resp.execute);
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_warn("command disallowed\n");
+		return -EBUSY;
+	}
+
+	ath10k_dbg(ATH10K_DBG_CORE,
+		   "%s: (device: 0x%p, address: 0x%x, param: %d)\n",
+		   __func__, ar, address, *param);
+
+	cmd.id            = __cpu_to_le32(BMI_EXECUTE);
+	cmd.execute.addr  = __cpu_to_le32(address);
+	cmd.execute.param = __cpu_to_le32(*param);
+
+	ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, cmdlen, &resp, &resplen);
+	if (ret) {
+		ath10k_warn("unable to read from the device\n");
+		return ret;
+	}
+
+	if (resplen < sizeof(resp.execute)) {
+		ath10k_warn("invalid execute response length (%d)\n",
+			    resplen);
+		return ret;
+	}
+
+	*param = __le32_to_cpu(resp.execute.result);
+	return 0;
+}
+
+int ath10k_bmi_lz_data(struct ath10k *ar, const void *buffer, u32 length)
+{
+	struct bmi_cmd cmd;
+	u32 hdrlen = sizeof(cmd.id) + sizeof(cmd.lz_data);
+	u32 txlen;
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_warn("command disallowed\n");
+		return -EBUSY;
+	}
+
+	while (length) {
+		txlen = min(length, BMI_MAX_DATA_SIZE - hdrlen);
+
+		WARN_ON_ONCE(txlen & 3);
+
+		cmd.id          = __cpu_to_le32(BMI_LZ_DATA);
+		cmd.lz_data.len = __cpu_to_le32(txlen);
+		memcpy(cmd.lz_data.payload, buffer, txlen);
+
+		ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, hdrlen + txlen,
+						  NULL, NULL);
+		if (ret) {
+			ath10k_warn("unable to write to the device\n");
+			return ret;
+		}
+
+		buffer += txlen;
+		length -= txlen;
+	}
+
+	return 0;
+}
+
+int ath10k_bmi_lz_stream_start(struct ath10k *ar, u32 address)
+{
+	struct bmi_cmd cmd;
+	u32 cmdlen = sizeof(cmd.id) + sizeof(cmd.lz_start);
+	int ret;
+
+	if (ar->bmi.done_sent) {
+		ath10k_warn("command disallowed\n");
+		return -EBUSY;
+	}
+
+	cmd.id            = __cpu_to_le32(BMI_LZ_STREAM_START);
+	cmd.lz_start.addr = __cpu_to_le32(address);
+
+	ret = ath10k_hif_exchange_bmi_msg(ar, &cmd, cmdlen, NULL, NULL);
+	if (ret) {
+		ath10k_warn("unable to Start LZ Stream to the device\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+int ath10k_bmi_fast_download(struct ath10k *ar,
+			     u32 address, const void *buffer, u32 length)
+{
+	u8 trailer[4] = {};
+	u32 head_len = rounddown(length, 4);
+	u32 trailer_len = length - head_len;
+	int ret;
+
+	ret = ath10k_bmi_lz_stream_start(ar, address);
+	if (ret)
+		return ret;
+
+	/* copy the last word into a zero padded buffer */
+	if (trailer_len > 0)
+		memcpy(trailer, buffer + head_len, trailer_len);
+
+	ret = ath10k_bmi_lz_data(ar, buffer, head_len);
+	if (ret)
+		return ret;
+
+	if (trailer_len > 0)
+		ret = ath10k_bmi_lz_data(ar, trailer, 4);
+
+	if (ret != 0)
+		return ret;
+
+	/*
+	 * Close compressed stream and open a new (fake) one.
+	 * This serves mainly to flush Target caches.
+	 */
+	ret = ath10k_bmi_lz_stream_start(ar, 0x00);
+
+	return ret;
+}
diff --git a/drivers/net/wireless/ath/ath10k/bmi.h b/drivers/net/wireless/ath/ath10k/bmi.h
new file mode 100644
index 0000000..32c56aa
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/bmi.h
@@ -0,0 +1,224 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _BMI_H_
+#define _BMI_H_
+
+#include "core.h"
+
+/*
+ * Bootloader Messaging Interface (BMI)
+ *
+ * BMI is a very simple messaging interface used during initialization
+ * to read memory, write memory, execute code, and to define an
+ * application entry PC.
+ *
+ * It is used to download an application to QCA988x, to provide
+ * patches to code that is already resident on QCA988x, and generally
+ * to examine and modify state.  The Host has an opportunity to use
+ * BMI only once during bootup.  Once the Host issues a BMI_DONE
+ * command, this opportunity ends.
+ *
+ * The Host writes BMI requests to mailbox0, and reads BMI responses
+ * from mailbox0.   BMI requests all begin with a command
+ * (see below for specific commands), and are followed by
+ * command-specific data.
+ *
+ * Flow control:
+ * The Host can only issue a command once the Target gives it a
+ * "BMI Command Credit", using AR8K Counter #4.  As soon as the
+ * Target has completed a command, it issues another BMI Command
+ * Credit (so the Host can issue the next command).
+ *
+ * BMI handles all required Target-side cache flushing.
+ */
+
+/* Maximum data size used for BMI transfers */
+#define BMI_MAX_DATA_SIZE	256
+
+/* len = cmd + addr + length */
+#define BMI_MAX_CMDBUF_SIZE (BMI_MAX_DATA_SIZE + \
+			sizeof(u32) + \
+			sizeof(u32) + \
+			sizeof(u32))
+
+/* BMI Commands */
+
+enum bmi_cmd_id {
+	BMI_NO_COMMAND          = 0,
+	BMI_DONE                = 1,
+	BMI_READ_MEMORY         = 2,
+	BMI_WRITE_MEMORY        = 3,
+	BMI_EXECUTE             = 4,
+	BMI_SET_APP_START       = 5,
+	BMI_READ_SOC_REGISTER   = 6,
+	BMI_READ_SOC_WORD       = 6,
+	BMI_WRITE_SOC_REGISTER  = 7,
+	BMI_WRITE_SOC_WORD      = 7,
+	BMI_GET_TARGET_ID       = 8,
+	BMI_GET_TARGET_INFO     = 8,
+	BMI_ROMPATCH_INSTALL    = 9,
+	BMI_ROMPATCH_UNINSTALL  = 10,
+	BMI_ROMPATCH_ACTIVATE   = 11,
+	BMI_ROMPATCH_DEACTIVATE = 12,
+	BMI_LZ_STREAM_START     = 13, /* should be followed by LZ_DATA */
+	BMI_LZ_DATA             = 14,
+	BMI_NVRAM_PROCESS       = 15,
+};
+
+#define BMI_NVRAM_SEG_NAME_SZ 16
+
+struct bmi_cmd {
+	__le32 id; /* enum bmi_cmd_id */
+	union {
+		struct {
+		} done;
+		struct {
+			__le32 addr;
+			__le32 len;
+		} read_mem;
+		struct {
+			__le32 addr;
+			__le32 len;
+			u8 payload[0];
+		} write_mem;
+		struct {
+			__le32 addr;
+			__le32 param;
+		} execute;
+		struct {
+			__le32 addr;
+		} set_app_start;
+		struct {
+			__le32 addr;
+		} read_soc_reg;
+		struct {
+			__le32 addr;
+			__le32 value;
+		} write_soc_reg;
+		struct {
+		} get_target_info;
+		struct {
+			__le32 rom_addr;
+			__le32 ram_addr; /* or value */
+			__le32 size;
+			__le32 activate; /* 0=install, but dont activate */
+		} rompatch_install;
+		struct {
+			__le32 patch_id;
+		} rompatch_uninstall;
+		struct {
+			__le32 count;
+			__le32 patch_ids[0]; /* length of @count */
+		} rompatch_activate;
+		struct {
+			__le32 count;
+			__le32 patch_ids[0]; /* length of @count */
+		} rompatch_deactivate;
+		struct {
+			__le32 addr;
+		} lz_start;
+		struct {
+			__le32 len; /* max BMI_MAX_DATA_SIZE */
+			u8 payload[0]; /* length of @len */
+		} lz_data;
+		struct {
+			u8 name[BMI_NVRAM_SEG_NAME_SZ];
+		} nvram_process;
+		u8 payload[BMI_MAX_CMDBUF_SIZE];
+	};
+} __packed;
+
+union bmi_resp {
+	struct {
+		u8 payload[0];
+	} read_mem;
+	struct {
+		__le32 result;
+	} execute;
+	struct {
+		__le32 value;
+	} read_soc_reg;
+	struct {
+		__le32 len;
+		__le32 version;
+		__le32 type;
+	} get_target_info;
+	struct {
+		__le32 patch_id;
+	} rompatch_install;
+	struct {
+		__le32 patch_id;
+	} rompatch_uninstall;
+	struct {
+		/* 0 = nothing executed
+		 * otherwise = NVRAM segment return value */
+		__le32 result;
+	} nvram_process;
+	u8 payload[BMI_MAX_CMDBUF_SIZE];
+} __packed;
+
+struct bmi_target_info {
+	u32 version;
+	u32 type;
+};
+
+
+/* in msec */
+#define BMI_COMMUNICATION_TIMEOUT_HZ (1*HZ)
+
+#define BMI_CE_NUM_TO_TARG 0
+#define BMI_CE_NUM_TO_HOST 1
+
+int ath10k_bmi_done(struct ath10k *ar);
+int ath10k_bmi_get_target_info(struct ath10k *ar,
+			       struct bmi_target_info *target_info);
+int ath10k_bmi_read_memory(struct ath10k *ar, u32 address,
+			   void *buffer, u32 length);
+int ath10k_bmi_write_memory(struct ath10k *ar, u32 address,
+			    const void *buffer, u32 length);
+
+#define ath10k_bmi_read32(ar, item, val)				\
+	({								\
+		int ret;						\
+		u32 addr;						\
+		__le32 tmp;						\
+									\
+		addr = host_interest_item_address(HI_ITEM(item));	\
+		ret = ath10k_bmi_read_memory(ar, addr, (u8 *)&tmp, 4); \
+		*val = __le32_to_cpu(tmp);				\
+		ret;							\
+	 })
+
+#define ath10k_bmi_write32(ar, item, val)				\
+	({								\
+		int ret;						\
+		u32 address;						\
+		__le32 v = __cpu_to_le32(val);				\
+									\
+		address = host_interest_item_address(HI_ITEM(item));	\
+		ret = ath10k_bmi_write_memory(ar, address,		\
+					      (u8 *)&v, sizeof(v));	\
+		ret;							\
+	})
+
+int ath10k_bmi_execute(struct ath10k *ar, u32 address, u32 *param);
+int ath10k_bmi_lz_stream_start(struct ath10k *ar, u32 address);
+int ath10k_bmi_lz_data(struct ath10k *ar, const void *buffer, u32 length);
+int ath10k_bmi_fast_download(struct ath10k *ar, u32 address,
+			     const void *buffer, u32 length);
+#endif /* _BMI_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/ce.c b/drivers/net/wireless/ath/ath10k/ce.c
new file mode 100644
index 0000000..61a8ac7
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/ce.c
@@ -0,0 +1,1189 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "hif.h"
+#include "pci.h"
+#include "ce.h"
+#include "debug.h"
+
+/*
+ * Support for Copy Engine hardware, which is mainly used for
+ * communication between Host and Target over a PCIe interconnect.
+ */
+
+/*
+ * A single CopyEngine (CE) comprises two "rings":
+ *   a source ring
+ *   a destination ring
+ *
+ * Each ring consists of a number of descriptors which specify
+ * an address, length, and meta-data.
+ *
+ * Typically, one side of the PCIe interconnect (Host or Target)
+ * controls one ring and the other side controls the other ring.
+ * The source side chooses when to initiate a transfer and it
+ * chooses what to send (buffer address, length). The destination
+ * side keeps a supply of "anonymous receive buffers" available and
+ * it handles incoming data as it arrives (when the destination
+ * recieves an interrupt).
+ *
+ * The sender may send a simple buffer (address/length) or it may
+ * send a small list of buffers.  When a small list is sent, hardware
+ * "gathers" these and they end up in a single destination buffer
+ * with a single interrupt.
+ *
+ * There are several "contexts" managed by this layer -- more, it
+ * may seem -- than should be needed. These are provided mainly for
+ * maximum flexibility and especially to facilitate a simpler HIF
+ * implementation. There are per-CopyEngine recv, send, and watermark
+ * contexts. These are supplied by the caller when a recv, send,
+ * or watermark handler is established and they are echoed back to
+ * the caller when the respective callbacks are invoked. There is
+ * also a per-transfer context supplied by the caller when a buffer
+ * (or sendlist) is sent and when a buffer is enqueued for recv.
+ * These per-transfer contexts are echoed back to the caller when
+ * the buffer is sent/received.
+ */
+
+static inline void ath10k_ce_dest_ring_write_index_set(struct ath10k *ar,
+						       u32 ce_ctrl_addr,
+						       unsigned int n)
+{
+	ath10k_pci_write32(ar, ce_ctrl_addr + DST_WR_INDEX_ADDRESS, n);
+}
+
+static inline u32 ath10k_ce_dest_ring_write_index_get(struct ath10k *ar,
+						      u32 ce_ctrl_addr)
+{
+	return ath10k_pci_read32(ar, ce_ctrl_addr + DST_WR_INDEX_ADDRESS);
+}
+
+static inline void ath10k_ce_src_ring_write_index_set(struct ath10k *ar,
+						      u32 ce_ctrl_addr,
+						      unsigned int n)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	void __iomem *indicator_addr;
+
+	if (!test_bit(ATH10K_PCI_FEATURE_HW_1_0_WARKAROUND, ar_pci->features)) {
+		ath10k_pci_write32(ar, ce_ctrl_addr + SR_WR_INDEX_ADDRESS, n);
+		return;
+	}
+
+	/* workaround for QCA988x_1.0 HW CE */
+	indicator_addr = ar_pci->mem + ce_ctrl_addr + DST_WATERMARK_ADDRESS;
+
+	if (ce_ctrl_addr == ath10k_ce_base_address(CDC_WAR_DATA_CE)) {
+		iowrite32((CDC_WAR_MAGIC_STR | n), indicator_addr);
+	} else {
+		unsigned long irq_flags;
+		local_irq_save(irq_flags);
+		iowrite32(1, indicator_addr);
+
+		/*
+		 * PCIE write waits for ACK in IPQ8K, there is no
+		 * need to read back value.
+		 */
+		(void)ioread32(indicator_addr);
+		(void)ioread32(indicator_addr); /* conservative */
+
+		ath10k_pci_write32(ar, ce_ctrl_addr + SR_WR_INDEX_ADDRESS, n);
+
+		iowrite32(0, indicator_addr);
+		local_irq_restore(irq_flags);
+	}
+}
+
+static inline u32 ath10k_ce_src_ring_write_index_get(struct ath10k *ar,
+						     u32 ce_ctrl_addr)
+{
+	return ath10k_pci_read32(ar, ce_ctrl_addr + SR_WR_INDEX_ADDRESS);
+}
+
+static inline u32 ath10k_ce_src_ring_read_index_get(struct ath10k *ar,
+						    u32 ce_ctrl_addr)
+{
+	return ath10k_pci_read32(ar, ce_ctrl_addr + CURRENT_SRRI_ADDRESS);
+}
+
+static inline void ath10k_ce_src_ring_base_addr_set(struct ath10k *ar,
+						    u32 ce_ctrl_addr,
+						    unsigned int addr)
+{
+	ath10k_pci_write32(ar, ce_ctrl_addr + SR_BA_ADDRESS, addr);
+}
+
+static inline void ath10k_ce_src_ring_size_set(struct ath10k *ar,
+					       u32 ce_ctrl_addr,
+					       unsigned int n)
+{
+	ath10k_pci_write32(ar, ce_ctrl_addr + SR_SIZE_ADDRESS, n);
+}
+
+static inline void ath10k_ce_src_ring_dmax_set(struct ath10k *ar,
+					       u32 ce_ctrl_addr,
+					       unsigned int n)
+{
+	u32 ctrl1_addr = ath10k_pci_read32((ar),
+					   (ce_ctrl_addr) + CE_CTRL1_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + CE_CTRL1_ADDRESS,
+			   (ctrl1_addr &  ~CE_CTRL1_DMAX_LENGTH_MASK) |
+			   CE_CTRL1_DMAX_LENGTH_SET(n));
+}
+
+static inline void ath10k_ce_src_ring_byte_swap_set(struct ath10k *ar,
+						    u32 ce_ctrl_addr,
+						    unsigned int n)
+{
+	u32 ctrl1_addr = ath10k_pci_read32(ar, ce_ctrl_addr + CE_CTRL1_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + CE_CTRL1_ADDRESS,
+			   (ctrl1_addr & ~CE_CTRL1_SRC_RING_BYTE_SWAP_EN_MASK) |
+			   CE_CTRL1_SRC_RING_BYTE_SWAP_EN_SET(n));
+}
+
+static inline void ath10k_ce_dest_ring_byte_swap_set(struct ath10k *ar,
+						     u32 ce_ctrl_addr,
+						     unsigned int n)
+{
+	u32 ctrl1_addr = ath10k_pci_read32(ar, ce_ctrl_addr + CE_CTRL1_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + CE_CTRL1_ADDRESS,
+			   (ctrl1_addr & ~CE_CTRL1_DST_RING_BYTE_SWAP_EN_MASK) |
+			   CE_CTRL1_DST_RING_BYTE_SWAP_EN_SET(n));
+}
+
+static inline u32 ath10k_ce_dest_ring_read_index_get(struct ath10k *ar,
+						     u32 ce_ctrl_addr)
+{
+	return ath10k_pci_read32(ar, ce_ctrl_addr + CURRENT_DRRI_ADDRESS);
+}
+
+static inline void ath10k_ce_dest_ring_base_addr_set(struct ath10k *ar,
+						     u32 ce_ctrl_addr,
+						     u32 addr)
+{
+	ath10k_pci_write32(ar, ce_ctrl_addr + DR_BA_ADDRESS, addr);
+}
+
+static inline void ath10k_ce_dest_ring_size_set(struct ath10k *ar,
+						u32 ce_ctrl_addr,
+						unsigned int n)
+{
+	ath10k_pci_write32(ar, ce_ctrl_addr + DR_SIZE_ADDRESS, n);
+}
+
+static inline void ath10k_ce_src_ring_highmark_set(struct ath10k *ar,
+						   u32 ce_ctrl_addr,
+						   unsigned int n)
+{
+	u32 addr = ath10k_pci_read32(ar, ce_ctrl_addr + SRC_WATERMARK_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + SRC_WATERMARK_ADDRESS,
+			   (addr & ~SRC_WATERMARK_HIGH_MASK) |
+			   SRC_WATERMARK_HIGH_SET(n));
+}
+
+static inline void ath10k_ce_src_ring_lowmark_set(struct ath10k *ar,
+						  u32 ce_ctrl_addr,
+						  unsigned int n)
+{
+	u32 addr = ath10k_pci_read32(ar, ce_ctrl_addr + SRC_WATERMARK_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + SRC_WATERMARK_ADDRESS,
+			   (addr & ~SRC_WATERMARK_LOW_MASK) |
+			   SRC_WATERMARK_LOW_SET(n));
+}
+
+static inline void ath10k_ce_dest_ring_highmark_set(struct ath10k *ar,
+						    u32 ce_ctrl_addr,
+						    unsigned int n)
+{
+	u32 addr = ath10k_pci_read32(ar, ce_ctrl_addr + DST_WATERMARK_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + DST_WATERMARK_ADDRESS,
+			   (addr & ~DST_WATERMARK_HIGH_MASK) |
+			   DST_WATERMARK_HIGH_SET(n));
+}
+
+static inline void ath10k_ce_dest_ring_lowmark_set(struct ath10k *ar,
+						   u32 ce_ctrl_addr,
+						   unsigned int n)
+{
+	u32 addr = ath10k_pci_read32(ar, ce_ctrl_addr + DST_WATERMARK_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + DST_WATERMARK_ADDRESS,
+			   (addr & ~DST_WATERMARK_LOW_MASK) |
+			   DST_WATERMARK_LOW_SET(n));
+}
+
+static inline void ath10k_ce_copy_complete_inter_enable(struct ath10k *ar,
+							u32 ce_ctrl_addr)
+{
+	u32 host_ie_addr = ath10k_pci_read32(ar,
+					     ce_ctrl_addr + HOST_IE_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + HOST_IE_ADDRESS,
+			   host_ie_addr | HOST_IE_COPY_COMPLETE_MASK);
+}
+
+static inline void ath10k_ce_copy_complete_intr_disable(struct ath10k *ar,
+							u32 ce_ctrl_addr)
+{
+	u32 host_ie_addr = ath10k_pci_read32(ar,
+					     ce_ctrl_addr + HOST_IE_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + HOST_IE_ADDRESS,
+			   host_ie_addr & ~HOST_IE_COPY_COMPLETE_MASK);
+}
+
+static inline void ath10k_ce_watermark_intr_disable(struct ath10k *ar,
+						    u32 ce_ctrl_addr)
+{
+	u32 host_ie_addr = ath10k_pci_read32(ar,
+					     ce_ctrl_addr + HOST_IE_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + HOST_IE_ADDRESS,
+			   host_ie_addr & ~CE_WATERMARK_MASK);
+}
+
+static inline void ath10k_ce_error_intr_enable(struct ath10k *ar,
+					       u32 ce_ctrl_addr)
+{
+	u32 misc_ie_addr = ath10k_pci_read32(ar,
+					     ce_ctrl_addr + MISC_IE_ADDRESS);
+
+	ath10k_pci_write32(ar, ce_ctrl_addr + MISC_IE_ADDRESS,
+			   misc_ie_addr | CE_ERROR_MASK);
+}
+
+static inline void ath10k_ce_engine_int_status_clear(struct ath10k *ar,
+						     u32 ce_ctrl_addr,
+						     unsigned int mask)
+{
+	ath10k_pci_write32(ar, ce_ctrl_addr + HOST_IS_ADDRESS, mask);
+}
+
+
+/*
+ * Guts of ath10k_ce_send, used by both ath10k_ce_send and
+ * ath10k_ce_sendlist_send.
+ * The caller takes responsibility for any needed locking.
+ */
+static int ath10k_ce_send_nolock(struct ce_state *ce_state,
+				 void *per_transfer_context,
+				 u32 buffer,
+				 unsigned int nbytes,
+				 unsigned int transfer_id,
+				 unsigned int flags)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ce_ring_state *src_ring = ce_state->src_ring;
+	struct ce_desc *desc, *sdesc;
+	unsigned int nentries_mask = src_ring->nentries_mask;
+	unsigned int sw_index = src_ring->sw_index;
+	unsigned int write_index = src_ring->write_index;
+	u32 ctrl_addr = ce_state->ctrl_addr;
+	u32 desc_flags = 0;
+	int ret = 0;
+
+	if (nbytes > ce_state->src_sz_max)
+		ath10k_warn("%s: send more we can (nbytes: %d, max: %d)\n",
+			    __func__, nbytes, ce_state->src_sz_max);
+
+	ath10k_pci_wake(ar);
+
+	if (unlikely(CE_RING_DELTA(nentries_mask,
+				   write_index, sw_index - 1) <= 0)) {
+		ret = -EIO;
+		goto exit;
+	}
+
+	desc = CE_SRC_RING_TO_DESC(src_ring->base_addr_owner_space,
+				   write_index);
+	sdesc = CE_SRC_RING_TO_DESC(src_ring->shadow_base, write_index);
+
+	desc_flags |= SM(transfer_id, CE_DESC_FLAGS_META_DATA);
+
+	if (flags & CE_SEND_FLAG_GATHER)
+		desc_flags |= CE_DESC_FLAGS_GATHER;
+	if (flags & CE_SEND_FLAG_BYTE_SWAP)
+		desc_flags |= CE_DESC_FLAGS_BYTE_SWAP;
+
+	sdesc->addr   = __cpu_to_le32(buffer);
+	sdesc->nbytes = __cpu_to_le16(nbytes);
+	sdesc->flags  = __cpu_to_le16(desc_flags);
+
+	*desc = *sdesc;
+
+	src_ring->per_transfer_context[write_index] = per_transfer_context;
+
+	/* Update Source Ring Write Index */
+	write_index = CE_RING_IDX_INCR(nentries_mask, write_index);
+
+	/* WORKAROUND */
+	if (!(flags & CE_SEND_FLAG_GATHER))
+		ath10k_ce_src_ring_write_index_set(ar, ctrl_addr, write_index);
+
+	src_ring->write_index = write_index;
+exit:
+	ath10k_pci_sleep(ar);
+	return ret;
+}
+
+int ath10k_ce_send(struct ce_state *ce_state,
+		   void *per_transfer_context,
+		   u32 buffer,
+		   unsigned int nbytes,
+		   unsigned int transfer_id,
+		   unsigned int flags)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+
+	spin_lock_bh(&ar_pci->ce_lock);
+	ret = ath10k_ce_send_nolock(ce_state, per_transfer_context,
+				    buffer, nbytes, transfer_id, flags);
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+void ath10k_ce_sendlist_buf_add(struct ce_sendlist *sendlist, u32 buffer,
+				unsigned int nbytes, u32 flags)
+{
+	unsigned int num_items = sendlist->num_items;
+	struct ce_sendlist_item *item;
+
+	item = &sendlist->item[num_items];
+	item->data = buffer;
+	item->u.nbytes = nbytes;
+	item->flags = flags;
+	sendlist->num_items++;
+}
+
+int ath10k_ce_sendlist_send(struct ce_state *ce_state,
+			    void *per_transfer_context,
+			    struct ce_sendlist *sendlist,
+			    unsigned int transfer_id)
+{
+	struct ce_ring_state *src_ring = ce_state->src_ring;
+	struct ce_sendlist_item *item;
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	unsigned int nentries_mask = src_ring->nentries_mask;
+	unsigned int num_items = sendlist->num_items;
+	unsigned int sw_index;
+	unsigned int write_index;
+	int i, delta, ret = -ENOMEM;
+
+	spin_lock_bh(&ar_pci->ce_lock);
+
+	sw_index = src_ring->sw_index;
+	write_index = src_ring->write_index;
+
+	delta = CE_RING_DELTA(nentries_mask, write_index, sw_index - 1);
+
+	if (delta >= num_items) {
+		/*
+		 * Handle all but the last item uniformly.
+		 */
+		for (i = 0; i < num_items - 1; i++) {
+			item = &sendlist->item[i];
+			ret = ath10k_ce_send_nolock(ce_state,
+						    CE_SENDLIST_ITEM_CTXT,
+						    (u32) item->data,
+						    item->u.nbytes, transfer_id,
+						    item->flags |
+						    CE_SEND_FLAG_GATHER);
+			if (ret)
+				ath10k_warn("CE send failed for item: %d\n", i);
+		}
+		/*
+		 * Provide valid context pointer for final item.
+		 */
+		item = &sendlist->item[i];
+		ret = ath10k_ce_send_nolock(ce_state, per_transfer_context,
+					    (u32) item->data, item->u.nbytes,
+					    transfer_id, item->flags);
+		if (ret)
+			ath10k_warn("CE send failed for last item: %d\n", i);
+	}
+
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+int ath10k_ce_recv_buf_enqueue(struct ce_state *ce_state,
+			       void *per_recv_context,
+			       u32 buffer)
+{
+	struct ce_ring_state *dest_ring = ce_state->dest_ring;
+	u32 ctrl_addr = ce_state->ctrl_addr;
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	unsigned int nentries_mask = dest_ring->nentries_mask;
+	unsigned int write_index;
+	unsigned int sw_index;
+	int ret;
+
+	spin_lock_bh(&ar_pci->ce_lock);
+	write_index = dest_ring->write_index;
+	sw_index = dest_ring->sw_index;
+
+	ath10k_pci_wake(ar);
+
+	if (CE_RING_DELTA(nentries_mask, write_index, sw_index - 1) > 0) {
+		struct ce_desc *base = dest_ring->base_addr_owner_space;
+		struct ce_desc *desc = CE_DEST_RING_TO_DESC(base, write_index);
+
+		/* Update destination descriptor */
+		desc->addr    = __cpu_to_le32(buffer);
+		desc->nbytes = 0;
+
+		dest_ring->per_transfer_context[write_index] =
+							per_recv_context;
+
+		/* Update Destination Ring Write Index */
+		write_index = CE_RING_IDX_INCR(nentries_mask, write_index);
+		ath10k_ce_dest_ring_write_index_set(ar, ctrl_addr, write_index);
+		dest_ring->write_index = write_index;
+		ret = 0;
+	} else {
+		ret = -EIO;
+	}
+	ath10k_pci_sleep(ar);
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+/*
+ * Guts of ath10k_ce_completed_recv_next.
+ * The caller takes responsibility for any necessary locking.
+ */
+static int ath10k_ce_completed_recv_next_nolock(struct ce_state *ce_state,
+						void **per_transfer_contextp,
+						u32 *bufferp,
+						unsigned int *nbytesp,
+						unsigned int *transfer_idp,
+						unsigned int *flagsp)
+{
+	struct ce_ring_state *dest_ring = ce_state->dest_ring;
+	unsigned int nentries_mask = dest_ring->nentries_mask;
+	unsigned int sw_index = dest_ring->sw_index;
+
+	struct ce_desc *base = dest_ring->base_addr_owner_space;
+	struct ce_desc *desc = CE_DEST_RING_TO_DESC(base, sw_index);
+	struct ce_desc sdesc;
+	u16 nbytes;
+
+	/* Copy in one go for performance reasons */
+	sdesc = *desc;
+
+	nbytes = __le16_to_cpu(sdesc.nbytes);
+	if (nbytes == 0) {
+		/*
+		 * This closes a relatively unusual race where the Host
+		 * sees the updated DRRI before the update to the
+		 * corresponding descriptor has completed. We treat this
+		 * as a descriptor that is not yet done.
+		 */
+		return -EIO;
+	}
+
+	desc->nbytes = 0;
+
+	/* Return data from completed destination descriptor */
+	*bufferp = __le32_to_cpu(sdesc.addr);
+	*nbytesp = nbytes;
+	*transfer_idp = MS(__le16_to_cpu(sdesc.flags), CE_DESC_FLAGS_META_DATA);
+
+	if (__le16_to_cpu(sdesc.flags) & CE_DESC_FLAGS_BYTE_SWAP)
+		*flagsp = CE_RECV_FLAG_SWAPPED;
+	else
+		*flagsp = 0;
+
+	if (per_transfer_contextp)
+		*per_transfer_contextp =
+			dest_ring->per_transfer_context[sw_index];
+
+	/* sanity */
+	dest_ring->per_transfer_context[sw_index] = NULL;
+
+	/* Update sw_index */
+	sw_index = CE_RING_IDX_INCR(nentries_mask, sw_index);
+	dest_ring->sw_index = sw_index;
+
+	return 0;
+}
+
+int ath10k_ce_completed_recv_next(struct ce_state *ce_state,
+				  void **per_transfer_contextp,
+				  u32 *bufferp,
+				  unsigned int *nbytesp,
+				  unsigned int *transfer_idp,
+				  unsigned int *flagsp)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+
+	spin_lock_bh(&ar_pci->ce_lock);
+	ret = ath10k_ce_completed_recv_next_nolock(ce_state,
+						   per_transfer_contextp,
+						   bufferp, nbytesp,
+						   transfer_idp, flagsp);
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+int ath10k_ce_revoke_recv_next(struct ce_state *ce_state,
+			       void **per_transfer_contextp,
+			       u32 *bufferp)
+{
+	struct ce_ring_state *dest_ring;
+	unsigned int nentries_mask;
+	unsigned int sw_index;
+	unsigned int write_index;
+	int ret;
+	struct ath10k *ar;
+	struct ath10k_pci *ar_pci;
+
+	dest_ring = ce_state->dest_ring;
+
+	if (!dest_ring)
+		return -EIO;
+
+	ar = ce_state->ar;
+	ar_pci = ath10k_pci_priv(ar);
+
+	spin_lock_bh(&ar_pci->ce_lock);
+
+	nentries_mask = dest_ring->nentries_mask;
+	sw_index = dest_ring->sw_index;
+	write_index = dest_ring->write_index;
+	if (write_index != sw_index) {
+		struct ce_desc *base = dest_ring->base_addr_owner_space;
+		struct ce_desc *desc = CE_DEST_RING_TO_DESC(base, sw_index);
+
+		/* Return data from completed destination descriptor */
+		*bufferp = __le32_to_cpu(desc->addr);
+
+		if (per_transfer_contextp)
+			*per_transfer_contextp =
+				dest_ring->per_transfer_context[sw_index];
+
+		/* sanity */
+		dest_ring->per_transfer_context[sw_index] = NULL;
+
+		/* Update sw_index */
+		sw_index = CE_RING_IDX_INCR(nentries_mask, sw_index);
+		dest_ring->sw_index = sw_index;
+		ret = 0;
+	} else {
+		ret = -EIO;
+	}
+
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+/*
+ * Guts of ath10k_ce_completed_send_next.
+ * The caller takes responsibility for any necessary locking.
+ */
+static int ath10k_ce_completed_send_next_nolock(struct ce_state *ce_state,
+						void **per_transfer_contextp,
+						u32 *bufferp,
+						unsigned int *nbytesp,
+						unsigned int *transfer_idp)
+{
+	struct ce_ring_state *src_ring = ce_state->src_ring;
+	u32 ctrl_addr = ce_state->ctrl_addr;
+	struct ath10k *ar = ce_state->ar;
+	unsigned int nentries_mask = src_ring->nentries_mask;
+	unsigned int sw_index = src_ring->sw_index;
+	unsigned int read_index;
+	int ret = -EIO;
+
+	if (src_ring->hw_index == sw_index) {
+		/*
+		 * The SW completion index has caught up with the cached
+		 * version of the HW completion index.
+		 * Update the cached HW completion index to see whether
+		 * the SW has really caught up to the HW, or if the cached
+		 * value of the HW index has become stale.
+		 */
+		ath10k_pci_wake(ar);
+		src_ring->hw_index =
+			ath10k_ce_src_ring_read_index_get(ar, ctrl_addr);
+		ath10k_pci_sleep(ar);
+	}
+	read_index = src_ring->hw_index;
+
+	if ((read_index != sw_index) && (read_index != 0xffffffff)) {
+		struct ce_desc *sbase = src_ring->shadow_base;
+		struct ce_desc *sdesc = CE_SRC_RING_TO_DESC(sbase, sw_index);
+
+		/* Return data from completed source descriptor */
+		*bufferp = __le32_to_cpu(sdesc->addr);
+		*nbytesp = __le16_to_cpu(sdesc->nbytes);
+		*transfer_idp = MS(__le16_to_cpu(sdesc->flags),
+						CE_DESC_FLAGS_META_DATA);
+
+		if (per_transfer_contextp)
+			*per_transfer_contextp =
+				src_ring->per_transfer_context[sw_index];
+
+		/* sanity */
+		src_ring->per_transfer_context[sw_index] = NULL;
+
+		/* Update sw_index */
+		sw_index = CE_RING_IDX_INCR(nentries_mask, sw_index);
+		src_ring->sw_index = sw_index;
+		ret = 0;
+	}
+
+	return ret;
+}
+
+/* NB: Modeled after ath10k_ce_completed_send_next */
+int ath10k_ce_cancel_send_next(struct ce_state *ce_state,
+			       void **per_transfer_contextp,
+			       u32 *bufferp,
+			       unsigned int *nbytesp,
+			       unsigned int *transfer_idp)
+{
+	struct ce_ring_state *src_ring;
+	unsigned int nentries_mask;
+	unsigned int sw_index;
+	unsigned int write_index;
+	int ret;
+	struct ath10k *ar;
+	struct ath10k_pci *ar_pci;
+
+	src_ring = ce_state->src_ring;
+
+	if (!src_ring)
+		return -EIO;
+
+	ar = ce_state->ar;
+	ar_pci = ath10k_pci_priv(ar);
+
+	spin_lock_bh(&ar_pci->ce_lock);
+
+	nentries_mask = src_ring->nentries_mask;
+	sw_index = src_ring->sw_index;
+	write_index = src_ring->write_index;
+
+	if (write_index != sw_index) {
+		struct ce_desc *base = src_ring->base_addr_owner_space;
+		struct ce_desc *desc = CE_SRC_RING_TO_DESC(base, sw_index);
+
+		/* Return data from completed source descriptor */
+		*bufferp = __le32_to_cpu(desc->addr);
+		*nbytesp = __le16_to_cpu(desc->nbytes);
+		*transfer_idp = MS(__le16_to_cpu(desc->flags),
+						CE_DESC_FLAGS_META_DATA);
+
+		if (per_transfer_contextp)
+			*per_transfer_contextp =
+				src_ring->per_transfer_context[sw_index];
+
+		/* sanity */
+		src_ring->per_transfer_context[sw_index] = NULL;
+
+		/* Update sw_index */
+		sw_index = CE_RING_IDX_INCR(nentries_mask, sw_index);
+		src_ring->sw_index = sw_index;
+		ret = 0;
+	} else {
+		ret = -EIO;
+	}
+
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+int ath10k_ce_completed_send_next(struct ce_state *ce_state,
+				  void **per_transfer_contextp,
+				  u32 *bufferp,
+				  unsigned int *nbytesp,
+				  unsigned int *transfer_idp)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+
+	spin_lock_bh(&ar_pci->ce_lock);
+	ret = ath10k_ce_completed_send_next_nolock(ce_state,
+						   per_transfer_contextp,
+						   bufferp, nbytesp,
+						   transfer_idp);
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ret;
+}
+
+/*
+ * Guts of interrupt handler for per-engine interrupts on a particular CE.
+ *
+ * Invokes registered callbacks for recv_complete,
+ * send_complete, and watermarks.
+ */
+void ath10k_ce_per_engine_service(struct ath10k *ar, unsigned int ce_id)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_state *ce_state = ar_pci->ce_id_to_state[ce_id];
+	u32 ctrl_addr = ce_state->ctrl_addr;
+	void *transfer_context;
+	u32 buf;
+	unsigned int nbytes;
+	unsigned int id;
+	unsigned int flags;
+
+	ath10k_pci_wake(ar);
+	spin_lock_bh(&ar_pci->ce_lock);
+
+	/* Clear the copy-complete interrupts that will be handled here. */
+	ath10k_ce_engine_int_status_clear(ar, ctrl_addr,
+					  HOST_IS_COPY_COMPLETE_MASK);
+
+	if (ce_state->recv_cb) {
+		/*
+		 * Pop completed recv buffers and call the registered
+		 * recv callback for each
+		 */
+		while (ath10k_ce_completed_recv_next_nolock(ce_state,
+							    &transfer_context,
+							    &buf, &nbytes,
+							    &id, &flags) == 0) {
+			spin_unlock_bh(&ar_pci->ce_lock);
+			ce_state->recv_cb(ce_state, transfer_context, buf,
+					  nbytes, id, flags);
+			spin_lock_bh(&ar_pci->ce_lock);
+		}
+	}
+
+	if (ce_state->send_cb) {
+		/*
+		 * Pop completed send buffers and call the registered
+		 * send callback for each
+		 */
+		while (ath10k_ce_completed_send_next_nolock(ce_state,
+							    &transfer_context,
+							    &buf,
+							    &nbytes,
+							    &id) == 0) {
+			spin_unlock_bh(&ar_pci->ce_lock);
+			ce_state->send_cb(ce_state, transfer_context,
+					  buf, nbytes, id);
+			spin_lock_bh(&ar_pci->ce_lock);
+		}
+	}
+
+	/*
+	 * Misc CE interrupts are not being handled, but still need
+	 * to be cleared.
+	 */
+	ath10k_ce_engine_int_status_clear(ar, ctrl_addr, CE_WATERMARK_MASK);
+
+	spin_unlock_bh(&ar_pci->ce_lock);
+	ath10k_pci_sleep(ar);
+}
+
+/*
+ * Handler for per-engine interrupts on ALL active CEs.
+ * This is used in cases where the system is sharing a
+ * single interrput for all CEs
+ */
+
+void ath10k_ce_per_engine_service_any(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ce_id;
+	u32 intr_summary;
+
+	ath10k_pci_wake(ar);
+	intr_summary = CE_INTERRUPT_SUMMARY(ar);
+
+	for (ce_id = 0; intr_summary && (ce_id < ar_pci->ce_count); ce_id++) {
+		if (intr_summary & (1 << ce_id))
+			intr_summary &= ~(1 << ce_id);
+		else
+			/* no intr pending on this CE */
+			continue;
+
+		ath10k_ce_per_engine_service(ar, ce_id);
+	}
+
+	ath10k_pci_sleep(ar);
+}
+
+/*
+ * Adjust interrupts for the copy complete handler.
+ * If it's needed for either send or recv, then unmask
+ * this interrupt; otherwise, mask it.
+ *
+ * Called with ce_lock held.
+ */
+static void ath10k_ce_per_engine_handler_adjust(struct ce_state *ce_state,
+						int disable_copy_compl_intr)
+{
+	u32 ctrl_addr = ce_state->ctrl_addr;
+	struct ath10k *ar = ce_state->ar;
+
+	ath10k_pci_wake(ar);
+
+	if ((!disable_copy_compl_intr) &&
+	    (ce_state->send_cb || ce_state->recv_cb))
+		ath10k_ce_copy_complete_inter_enable(ar, ctrl_addr);
+	else
+		ath10k_ce_copy_complete_intr_disable(ar, ctrl_addr);
+
+	ath10k_ce_watermark_intr_disable(ar, ctrl_addr);
+
+	ath10k_pci_sleep(ar);
+}
+
+void ath10k_ce_disable_interrupts(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ce_id;
+
+	ath10k_pci_wake(ar);
+	for (ce_id = 0; ce_id < ar_pci->ce_count; ce_id++) {
+		struct ce_state *ce_state = ar_pci->ce_id_to_state[ce_id];
+		u32 ctrl_addr = ce_state->ctrl_addr;
+
+		ath10k_ce_copy_complete_intr_disable(ar, ctrl_addr);
+	}
+	ath10k_pci_sleep(ar);
+}
+
+void ath10k_ce_send_cb_register(struct ce_state *ce_state,
+				void (*send_cb) (struct ce_state *ce_state,
+						 void *transfer_context,
+						 u32 buffer,
+						 unsigned int nbytes,
+						 unsigned int transfer_id),
+				int disable_interrupts)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	spin_lock_bh(&ar_pci->ce_lock);
+	ce_state->send_cb = send_cb;
+	ath10k_ce_per_engine_handler_adjust(ce_state, disable_interrupts);
+	spin_unlock_bh(&ar_pci->ce_lock);
+}
+
+void ath10k_ce_recv_cb_register(struct ce_state *ce_state,
+				void (*recv_cb) (struct ce_state *ce_state,
+						 void *transfer_context,
+						 u32 buffer,
+						 unsigned int nbytes,
+						 unsigned int transfer_id,
+						 unsigned int flags))
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	spin_lock_bh(&ar_pci->ce_lock);
+	ce_state->recv_cb = recv_cb;
+	ath10k_ce_per_engine_handler_adjust(ce_state, 0);
+	spin_unlock_bh(&ar_pci->ce_lock);
+}
+
+static int ath10k_ce_init_src_ring(struct ath10k *ar,
+				   unsigned int ce_id,
+				   struct ce_state *ce_state,
+				   const struct ce_attr *attr)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_ring_state *src_ring;
+	unsigned int nentries = attr->src_nentries;
+	unsigned int ce_nbytes;
+	u32 ctrl_addr = ath10k_ce_base_address(ce_id);
+	dma_addr_t base_addr;
+	char *ptr;
+
+	nentries = roundup_pow_of_two(nentries);
+
+	if (ce_state->src_ring) {
+		WARN_ON(ce_state->src_ring->nentries != nentries);
+		return 0;
+	}
+
+	ce_nbytes = sizeof(struct ce_ring_state) + (nentries * sizeof(void *));
+	ptr = kzalloc(ce_nbytes, GFP_KERNEL);
+	if (ptr == NULL)
+		return -ENOMEM;
+
+	ce_state->src_ring = (struct ce_ring_state *)ptr;
+	src_ring = ce_state->src_ring;
+
+	ptr += sizeof(struct ce_ring_state);
+	src_ring->nentries = nentries;
+	src_ring->nentries_mask = nentries - 1;
+
+	ath10k_pci_wake(ar);
+	src_ring->sw_index = ath10k_ce_src_ring_read_index_get(ar, ctrl_addr);
+	src_ring->hw_index = src_ring->sw_index;
+
+	src_ring->write_index =
+		ath10k_ce_src_ring_write_index_get(ar, ctrl_addr);
+	ath10k_pci_sleep(ar);
+
+	src_ring->per_transfer_context = (void **)ptr;
+
+	/*
+	 * Legacy platforms that do not support cache
+	 * coherent DMA are unsupported
+	 */
+	src_ring->base_addr_owner_space_unaligned =
+		pci_alloc_consistent(ar_pci->pdev,
+				     (nentries * sizeof(struct ce_desc) +
+				      CE_DESC_RING_ALIGN),
+				     &base_addr);
+	src_ring->base_addr_ce_space_unaligned = base_addr;
+
+	src_ring->base_addr_owner_space = PTR_ALIGN(
+			src_ring->base_addr_owner_space_unaligned,
+			CE_DESC_RING_ALIGN);
+	src_ring->base_addr_ce_space = ALIGN(
+			src_ring->base_addr_ce_space_unaligned,
+			CE_DESC_RING_ALIGN);
+
+	/*
+	 * Also allocate a shadow src ring in regular
+	 * mem to use for faster access.
+	 */
+	src_ring->shadow_base_unaligned =
+		kmalloc((nentries * sizeof(struct ce_desc) +
+			 CE_DESC_RING_ALIGN), GFP_KERNEL);
+
+	src_ring->shadow_base = PTR_ALIGN(
+			src_ring->shadow_base_unaligned,
+			CE_DESC_RING_ALIGN);
+
+	ath10k_pci_wake(ar);
+	ath10k_ce_src_ring_base_addr_set(ar, ctrl_addr,
+					 src_ring->base_addr_ce_space);
+	ath10k_ce_src_ring_size_set(ar, ctrl_addr, nentries);
+	ath10k_ce_src_ring_dmax_set(ar, ctrl_addr, attr->src_sz_max);
+	ath10k_ce_src_ring_byte_swap_set(ar, ctrl_addr, 0);
+	ath10k_ce_src_ring_lowmark_set(ar, ctrl_addr, 0);
+	ath10k_ce_src_ring_highmark_set(ar, ctrl_addr, nentries);
+	ath10k_pci_sleep(ar);
+
+	return 0;
+}
+
+static int ath10k_ce_init_dest_ring(struct ath10k *ar,
+				    unsigned int ce_id,
+				    struct ce_state *ce_state,
+				    const struct ce_attr *attr)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_ring_state *dest_ring;
+	unsigned int nentries = attr->dest_nentries;
+	unsigned int ce_nbytes;
+	u32 ctrl_addr = ath10k_ce_base_address(ce_id);
+	dma_addr_t base_addr;
+	char *ptr;
+
+	nentries = roundup_pow_of_two(nentries);
+
+	if (ce_state->dest_ring) {
+		WARN_ON(ce_state->dest_ring->nentries != nentries);
+		return 0;
+	}
+
+	ce_nbytes = sizeof(struct ce_ring_state) + (nentries * sizeof(void *));
+	ptr = kzalloc(ce_nbytes, GFP_KERNEL);
+	if (ptr == NULL)
+		return -ENOMEM;
+
+	ce_state->dest_ring = (struct ce_ring_state *)ptr;
+	dest_ring = ce_state->dest_ring;
+
+	ptr += sizeof(struct ce_ring_state);
+	dest_ring->nentries = nentries;
+	dest_ring->nentries_mask = nentries - 1;
+
+	ath10k_pci_wake(ar);
+	dest_ring->sw_index = ath10k_ce_dest_ring_read_index_get(ar, ctrl_addr);
+	dest_ring->write_index =
+		ath10k_ce_dest_ring_write_index_get(ar, ctrl_addr);
+	ath10k_pci_sleep(ar);
+
+	dest_ring->per_transfer_context = (void **)ptr;
+
+	/*
+	 * Legacy platforms that do not support cache
+	 * coherent DMA are unsupported
+	 */
+	dest_ring->base_addr_owner_space_unaligned =
+		pci_alloc_consistent(ar_pci->pdev,
+				     (nentries * sizeof(struct ce_desc) +
+				      CE_DESC_RING_ALIGN),
+				     &base_addr);
+	dest_ring->base_addr_ce_space_unaligned = base_addr;
+
+	/*
+	 * Correctly initialize memory to 0 to prevent garbage
+	 * data crashing system when download firmware
+	 */
+	memset(dest_ring->base_addr_owner_space_unaligned, 0,
+	       nentries * sizeof(struct ce_desc) + CE_DESC_RING_ALIGN);
+
+	dest_ring->base_addr_owner_space = PTR_ALIGN(
+			dest_ring->base_addr_owner_space_unaligned,
+			CE_DESC_RING_ALIGN);
+	dest_ring->base_addr_ce_space = ALIGN(
+			dest_ring->base_addr_ce_space_unaligned,
+			CE_DESC_RING_ALIGN);
+
+	ath10k_pci_wake(ar);
+	ath10k_ce_dest_ring_base_addr_set(ar, ctrl_addr,
+					  dest_ring->base_addr_ce_space);
+	ath10k_ce_dest_ring_size_set(ar, ctrl_addr, nentries);
+	ath10k_ce_dest_ring_byte_swap_set(ar, ctrl_addr, 0);
+	ath10k_ce_dest_ring_lowmark_set(ar, ctrl_addr, 0);
+	ath10k_ce_dest_ring_highmark_set(ar, ctrl_addr, nentries);
+	ath10k_pci_sleep(ar);
+
+	return 0;
+}
+
+static struct ce_state *ath10k_ce_init_state(struct ath10k *ar,
+					     unsigned int ce_id,
+					     const struct ce_attr *attr)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_state *ce_state = NULL;
+	u32 ctrl_addr = ath10k_ce_base_address(ce_id);
+
+	spin_lock_bh(&ar_pci->ce_lock);
+
+	if (!ar_pci->ce_id_to_state[ce_id]) {
+		ce_state = kzalloc(sizeof(*ce_state), GFP_ATOMIC);
+		if (ce_state == NULL) {
+			spin_unlock_bh(&ar_pci->ce_lock);
+			return NULL;
+		}
+
+		ar_pci->ce_id_to_state[ce_id] = ce_state;
+		ce_state->ar = ar;
+		ce_state->id = ce_id;
+		ce_state->ctrl_addr = ctrl_addr;
+		ce_state->state = CE_RUNNING;
+		/* Save attribute flags */
+		ce_state->attr_flags = attr->flags;
+		ce_state->src_sz_max = attr->src_sz_max;
+	}
+
+	spin_unlock_bh(&ar_pci->ce_lock);
+
+	return ce_state;
+}
+
+/*
+ * Initialize a Copy Engine based on caller-supplied attributes.
+ * This may be called once to initialize both source and destination
+ * rings or it may be called twice for separate source and destination
+ * initialization. It may be that only one side or the other is
+ * initialized by software/firmware.
+ */
+struct ce_state *ath10k_ce_init(struct ath10k *ar,
+				unsigned int ce_id,
+				const struct ce_attr *attr)
+{
+	struct ce_state *ce_state;
+	u32 ctrl_addr = ath10k_ce_base_address(ce_id);
+
+	ce_state = ath10k_ce_init_state(ar, ce_id, attr);
+	if (!ce_state) {
+		ath10k_err("Failed to initialize CE state for ID: %d\n", ce_id);
+		return NULL;
+	}
+
+	if (attr->src_nentries) {
+		if (ath10k_ce_init_src_ring(ar, ce_id, ce_state, attr)) {
+			ath10k_err("Failed to initialize CE src ring for ID: %d\n",
+				   ce_id);
+			ath10k_ce_deinit(ce_state);
+			return NULL;
+		}
+	}
+
+	if (attr->dest_nentries) {
+		if (ath10k_ce_init_dest_ring(ar, ce_id, ce_state, attr)) {
+			ath10k_err("Failed to initialize CE dest ring for ID: %d\n",
+				   ce_id);
+			ath10k_ce_deinit(ce_state);
+			return NULL;
+		}
+	}
+
+	/* Enable CE error interrupts */
+	ath10k_pci_wake(ar);
+	ath10k_ce_error_intr_enable(ar, ctrl_addr);
+	ath10k_pci_sleep(ar);
+
+	return ce_state;
+}
+
+void ath10k_ce_deinit(struct ce_state *ce_state)
+{
+	unsigned int ce_id = ce_state->id;
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	ce_state->state = CE_UNUSED;
+	ar_pci->ce_id_to_state[ce_id] = NULL;
+
+	if (ce_state->src_ring) {
+		kfree(ce_state->src_ring->shadow_base_unaligned);
+		pci_free_consistent(ar_pci->pdev,
+				    (ce_state->src_ring->nentries *
+				     sizeof(struct ce_desc) +
+				     CE_DESC_RING_ALIGN),
+				    ce_state->src_ring->base_addr_owner_space,
+				    ce_state->src_ring->base_addr_ce_space);
+		kfree(ce_state->src_ring);
+	}
+
+	if (ce_state->dest_ring) {
+		pci_free_consistent(ar_pci->pdev,
+				    (ce_state->dest_ring->nentries *
+				     sizeof(struct ce_desc) +
+				     CE_DESC_RING_ALIGN),
+				    ce_state->dest_ring->base_addr_owner_space,
+				    ce_state->dest_ring->base_addr_ce_space);
+		kfree(ce_state->dest_ring);
+	}
+	kfree(ce_state);
+}
diff --git a/drivers/net/wireless/ath/ath10k/ce.h b/drivers/net/wireless/ath/ath10k/ce.h
new file mode 100644
index 0000000..c17f07c
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/ce.h
@@ -0,0 +1,516 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _CE_H_
+#define _CE_H_
+
+#include "hif.h"
+
+
+/* Maximum number of Copy Engine's supported */
+#define CE_COUNT_MAX 8
+#define CE_HTT_H2T_MSG_SRC_NENTRIES 2048
+
+/* Descriptor rings must be aligned to this boundary */
+#define CE_DESC_RING_ALIGN	8
+#define CE_SENDLIST_ITEMS_MAX	12
+#define CE_SEND_FLAG_GATHER	0x00010000
+
+/*
+ * Copy Engine support: low-level Target-side Copy Engine API.
+ * This is a hardware access layer used by code that understands
+ * how to use copy engines.
+ */
+
+struct ce_state;
+
+
+/* Copy Engine operational state */
+enum ce_op_state {
+	CE_UNUSED,
+	CE_PAUSED,
+	CE_RUNNING,
+};
+
+#define CE_DESC_FLAGS_GATHER         (1 << 0)
+#define CE_DESC_FLAGS_BYTE_SWAP      (1 << 1)
+#define CE_DESC_FLAGS_META_DATA_MASK 0xFFFC
+#define CE_DESC_FLAGS_META_DATA_LSB  3
+
+struct ce_desc {
+	__le32 addr;
+	__le16 nbytes;
+	__le16 flags; /* %CE_DESC_FLAGS_ */
+};
+
+/* Copy Engine Ring internal state */
+struct ce_ring_state {
+	/* Number of entries in this ring; must be power of 2 */
+	unsigned int nentries;
+	unsigned int nentries_mask;
+
+	/*
+	 * For dest ring, this is the next index to be processed
+	 * by software after it was/is received into.
+	 *
+	 * For src ring, this is the last descriptor that was sent
+	 * and completion processed by software.
+	 *
+	 * Regardless of src or dest ring, this is an invariant
+	 * (modulo ring size):
+	 *     write index >= read index >= sw_index
+	 */
+	unsigned int sw_index;
+	/* cached copy */
+	unsigned int write_index;
+	/*
+	 * For src ring, this is the next index not yet processed by HW.
+	 * This is a cached copy of the real HW index (read index), used
+	 * for avoiding reading the HW index register more often than
+	 * necessary.
+	 * This extends the invariant:
+	 *     write index >= read index >= hw_index >= sw_index
+	 *
+	 * For dest ring, this is currently unused.
+	 */
+	/* cached copy */
+	unsigned int hw_index;
+
+	/* Start of DMA-coherent area reserved for descriptors */
+	/* Host address space */
+	void *base_addr_owner_space_unaligned;
+	/* CE address space */
+	u32 base_addr_ce_space_unaligned;
+
+	/*
+	 * Actual start of descriptors.
+	 * Aligned to descriptor-size boundary.
+	 * Points into reserved DMA-coherent area, above.
+	 */
+	/* Host address space */
+	void *base_addr_owner_space;
+
+	/* CE address space */
+	u32 base_addr_ce_space;
+	/*
+	 * Start of shadow copy of descriptors, within regular memory.
+	 * Aligned to descriptor-size boundary.
+	 */
+	void *shadow_base_unaligned;
+	struct ce_desc *shadow_base;
+
+	void **per_transfer_context;
+};
+
+/* Copy Engine internal state */
+struct ce_state {
+	struct ath10k *ar;
+	unsigned int id;
+
+	unsigned int attr_flags;
+
+	u32 ctrl_addr;
+	enum ce_op_state state;
+
+	void (*send_cb) (struct ce_state *ce_state,
+			 void *per_transfer_send_context,
+			 u32 buffer,
+			 unsigned int nbytes,
+			 unsigned int transfer_id);
+	void (*recv_cb) (struct ce_state *ce_state,
+			 void *per_transfer_recv_context,
+			 u32 buffer,
+			 unsigned int nbytes,
+			 unsigned int transfer_id,
+			 unsigned int flags);
+
+	unsigned int src_sz_max;
+	struct ce_ring_state *src_ring;
+	struct ce_ring_state *dest_ring;
+};
+
+struct ce_sendlist_item {
+	/* e.g. buffer or desc list */
+	dma_addr_t data;
+	union {
+		/* simple buffer */
+		unsigned int nbytes;
+		/* Rx descriptor list */
+		unsigned int ndesc;
+	} u;
+	/* externally-specified flags; OR-ed with internal flags */
+	u32 flags;
+};
+
+struct ce_sendlist {
+	unsigned int num_items;
+	struct ce_sendlist_item item[CE_SENDLIST_ITEMS_MAX];
+};
+
+/* Copy Engine settable attributes */
+struct ce_attr;
+
+/*==================Send====================*/
+
+/* ath10k_ce_send flags */
+#define CE_SEND_FLAG_BYTE_SWAP 1
+
+/*
+ * Queue a source buffer to be sent to an anonymous destination buffer.
+ *   ce         - which copy engine to use
+ *   buffer          - address of buffer
+ *   nbytes          - number of bytes to send
+ *   transfer_id     - arbitrary ID; reflected to destination
+ *   flags           - CE_SEND_FLAG_* values
+ * Returns 0 on success; otherwise an error status.
+ *
+ * Note: If no flags are specified, use CE's default data swap mode.
+ *
+ * Implementation note: pushes 1 buffer to Source ring
+ */
+int ath10k_ce_send(struct ce_state *ce_state,
+		   void *per_transfer_send_context,
+		   u32 buffer,
+		   unsigned int nbytes,
+		   /* 14 bits */
+		   unsigned int transfer_id,
+		   unsigned int flags);
+
+void ath10k_ce_send_cb_register(struct ce_state *ce_state,
+				void (*send_cb) (struct ce_state *ce_state,
+						 void *transfer_context,
+						 u32 buffer,
+						 unsigned int nbytes,
+						 unsigned int transfer_id),
+				int disable_interrupts);
+
+/* Append a simple buffer (address/length) to a sendlist. */
+void ath10k_ce_sendlist_buf_add(struct ce_sendlist *sendlist,
+				u32 buffer,
+				unsigned int nbytes,
+				/* OR-ed with internal flags */
+				u32 flags);
+
+/*
+ * Queue a "sendlist" of buffers to be sent using gather to a single
+ * anonymous destination buffer
+ *   ce         - which copy engine to use
+ *   sendlist        - list of simple buffers to send using gather
+ *   transfer_id     - arbitrary ID; reflected to destination
+ * Returns 0 on success; otherwise an error status.
+ *
+ * Implemenation note: Pushes multiple buffers with Gather to Source ring.
+ */
+int ath10k_ce_sendlist_send(struct ce_state *ce_state,
+			    void *per_transfer_send_context,
+			    struct ce_sendlist *sendlist,
+			    /* 14 bits */
+			    unsigned int transfer_id);
+
+/*==================Recv=======================*/
+
+/*
+ * Make a buffer available to receive. The buffer must be at least of a
+ * minimal size appropriate for this copy engine (src_sz_max attribute).
+ *   ce                    - which copy engine to use
+ *   per_transfer_recv_context  - context passed back to caller's recv_cb
+ *   buffer                     - address of buffer in CE space
+ * Returns 0 on success; otherwise an error status.
+ *
+ * Implemenation note: Pushes a buffer to Dest ring.
+ */
+int ath10k_ce_recv_buf_enqueue(struct ce_state *ce_state,
+			       void *per_transfer_recv_context,
+			       u32 buffer);
+
+void ath10k_ce_recv_cb_register(struct ce_state *ce_state,
+				void (*recv_cb) (struct ce_state *ce_state,
+						 void *transfer_context,
+						 u32 buffer,
+						 unsigned int nbytes,
+						 unsigned int transfer_id,
+						 unsigned int flags));
+
+/* recv flags */
+/* Data is byte-swapped */
+#define CE_RECV_FLAG_SWAPPED	1
+
+/*
+ * Supply data for the next completed unprocessed receive descriptor.
+ * Pops buffer from Dest ring.
+ */
+int ath10k_ce_completed_recv_next(struct ce_state *ce_state,
+				  void **per_transfer_contextp,
+				  u32 *bufferp,
+				  unsigned int *nbytesp,
+				  unsigned int *transfer_idp,
+				  unsigned int *flagsp);
+/*
+ * Supply data for the next completed unprocessed send descriptor.
+ * Pops 1 completed send buffer from Source ring.
+ */
+int ath10k_ce_completed_send_next(struct ce_state *ce_state,
+			   void **per_transfer_contextp,
+			   u32 *bufferp,
+			   unsigned int *nbytesp,
+			   unsigned int *transfer_idp);
+
+/*==================CE Engine Initialization=======================*/
+
+/* Initialize an instance of a CE */
+struct ce_state *ath10k_ce_init(struct ath10k *ar,
+				unsigned int ce_id,
+				const struct ce_attr *attr);
+
+/*==================CE Engine Shutdown=======================*/
+/*
+ * Support clean shutdown by allowing the caller to revoke
+ * receive buffers.  Target DMA must be stopped before using
+ * this API.
+ */
+int ath10k_ce_revoke_recv_next(struct ce_state *ce_state,
+			       void **per_transfer_contextp,
+			       u32 *bufferp);
+
+/*
+ * Support clean shutdown by allowing the caller to cancel
+ * pending sends.  Target DMA must be stopped before using
+ * this API.
+ */
+int ath10k_ce_cancel_send_next(struct ce_state *ce_state,
+			       void **per_transfer_contextp,
+			       u32 *bufferp,
+			       unsigned int *nbytesp,
+			       unsigned int *transfer_idp);
+
+void ath10k_ce_deinit(struct ce_state *ce_state);
+
+/*==================CE Interrupt Handlers====================*/
+void ath10k_ce_per_engine_service_any(struct ath10k *ar);
+void ath10k_ce_per_engine_service(struct ath10k *ar, unsigned int ce_id);
+void ath10k_ce_disable_interrupts(struct ath10k *ar);
+
+/* ce_attr.flags values */
+/* Use NonSnooping PCIe accesses? */
+#define CE_ATTR_NO_SNOOP		1
+
+/* Byte swap data words */
+#define CE_ATTR_BYTE_SWAP_DATA		2
+
+/* Swizzle descriptors? */
+#define CE_ATTR_SWIZZLE_DESCRIPTORS	4
+
+/* no interrupt on copy completion */
+#define CE_ATTR_DIS_INTR		8
+
+/* Attributes of an instance of a Copy Engine */
+struct ce_attr {
+	/* CE_ATTR_* values */
+	unsigned int flags;
+
+	/* currently not in use */
+	unsigned int priority;
+
+	/* #entries in source ring - Must be a power of 2 */
+	unsigned int src_nentries;
+
+	/*
+	 * Max source send size for this CE.
+	 * This is also the minimum size of a destination buffer.
+	 */
+	unsigned int src_sz_max;
+
+	/* #entries in destination ring - Must be a power of 2 */
+	unsigned int dest_nentries;
+
+	/* Future use */
+	void *reserved;
+};
+
+/*
+ * When using sendlist_send to transfer multiple buffer fragments, the
+ * transfer context of each fragment, except last one, will be filled
+ * with CE_SENDLIST_ITEM_CTXT. ce_completed_send will return success for
+ * each fragment done with send and the transfer context would be
+ * CE_SENDLIST_ITEM_CTXT. Upper layer could use this to identify the
+ * status of a send completion.
+ */
+#define CE_SENDLIST_ITEM_CTXT	((void *)0xcecebeef)
+
+#define SR_BA_ADDRESS		0x0000
+#define SR_SIZE_ADDRESS		0x0004
+#define DR_BA_ADDRESS		0x0008
+#define DR_SIZE_ADDRESS		0x000c
+#define CE_CMD_ADDRESS		0x0018
+
+#define CE_CTRL1_DST_RING_BYTE_SWAP_EN_MSB	17
+#define CE_CTRL1_DST_RING_BYTE_SWAP_EN_LSB	17
+#define CE_CTRL1_DST_RING_BYTE_SWAP_EN_MASK	0x00020000
+#define CE_CTRL1_DST_RING_BYTE_SWAP_EN_SET(x) \
+	(((0 | (x)) << CE_CTRL1_DST_RING_BYTE_SWAP_EN_LSB) & \
+	CE_CTRL1_DST_RING_BYTE_SWAP_EN_MASK)
+
+#define CE_CTRL1_SRC_RING_BYTE_SWAP_EN_MSB	16
+#define CE_CTRL1_SRC_RING_BYTE_SWAP_EN_LSB	16
+#define CE_CTRL1_SRC_RING_BYTE_SWAP_EN_MASK	0x00010000
+#define CE_CTRL1_SRC_RING_BYTE_SWAP_EN_GET(x) \
+	(((x) & CE_CTRL1_SRC_RING_BYTE_SWAP_EN_MASK) >> \
+	 CE_CTRL1_SRC_RING_BYTE_SWAP_EN_LSB)
+#define CE_CTRL1_SRC_RING_BYTE_SWAP_EN_SET(x) \
+	(((0 | (x)) << CE_CTRL1_SRC_RING_BYTE_SWAP_EN_LSB) & \
+	 CE_CTRL1_SRC_RING_BYTE_SWAP_EN_MASK)
+
+#define CE_CTRL1_DMAX_LENGTH_MSB		15
+#define CE_CTRL1_DMAX_LENGTH_LSB		0
+#define CE_CTRL1_DMAX_LENGTH_MASK		0x0000ffff
+#define CE_CTRL1_DMAX_LENGTH_GET(x) \
+	(((x) & CE_CTRL1_DMAX_LENGTH_MASK) >> CE_CTRL1_DMAX_LENGTH_LSB)
+#define CE_CTRL1_DMAX_LENGTH_SET(x) \
+	(((0 | (x)) << CE_CTRL1_DMAX_LENGTH_LSB) & CE_CTRL1_DMAX_LENGTH_MASK)
+
+#define CE_CTRL1_ADDRESS			0x0010
+#define CE_CTRL1_HW_MASK			0x0007ffff
+#define CE_CTRL1_SW_MASK			0x0007ffff
+#define CE_CTRL1_HW_WRITE_MASK			0x00000000
+#define CE_CTRL1_SW_WRITE_MASK			0x0007ffff
+#define CE_CTRL1_RSTMASK			0xffffffff
+#define CE_CTRL1_RESET				0x00000080
+
+#define CE_CMD_HALT_STATUS_MSB			3
+#define CE_CMD_HALT_STATUS_LSB			3
+#define CE_CMD_HALT_STATUS_MASK			0x00000008
+#define CE_CMD_HALT_STATUS_GET(x) \
+	(((x) & CE_CMD_HALT_STATUS_MASK) >> CE_CMD_HALT_STATUS_LSB)
+#define CE_CMD_HALT_STATUS_SET(x) \
+	(((0 | (x)) << CE_CMD_HALT_STATUS_LSB) & CE_CMD_HALT_STATUS_MASK)
+#define CE_CMD_HALT_STATUS_RESET		0
+#define CE_CMD_HALT_MSB				0
+#define CE_CMD_HALT_MASK			0x00000001
+
+#define HOST_IE_COPY_COMPLETE_MSB		0
+#define HOST_IE_COPY_COMPLETE_LSB		0
+#define HOST_IE_COPY_COMPLETE_MASK		0x00000001
+#define HOST_IE_COPY_COMPLETE_GET(x) \
+	(((x) & HOST_IE_COPY_COMPLETE_MASK) >> HOST_IE_COPY_COMPLETE_LSB)
+#define HOST_IE_COPY_COMPLETE_SET(x) \
+	(((0 | (x)) << HOST_IE_COPY_COMPLETE_LSB) & HOST_IE_COPY_COMPLETE_MASK)
+#define HOST_IE_COPY_COMPLETE_RESET		0
+#define HOST_IE_ADDRESS				0x002c
+
+#define HOST_IS_DST_RING_LOW_WATERMARK_MASK	0x00000010
+#define HOST_IS_DST_RING_HIGH_WATERMARK_MASK	0x00000008
+#define HOST_IS_SRC_RING_LOW_WATERMARK_MASK	0x00000004
+#define HOST_IS_SRC_RING_HIGH_WATERMARK_MASK	0x00000002
+#define HOST_IS_COPY_COMPLETE_MASK		0x00000001
+#define HOST_IS_ADDRESS				0x0030
+
+#define MISC_IE_ADDRESS				0x0034
+
+#define MISC_IS_AXI_ERR_MASK			0x00000400
+
+#define MISC_IS_DST_ADDR_ERR_MASK		0x00000200
+#define MISC_IS_SRC_LEN_ERR_MASK		0x00000100
+#define MISC_IS_DST_MAX_LEN_VIO_MASK		0x00000080
+#define MISC_IS_DST_RING_OVERFLOW_MASK		0x00000040
+#define MISC_IS_SRC_RING_OVERFLOW_MASK		0x00000020
+
+#define MISC_IS_ADDRESS				0x0038
+
+#define SR_WR_INDEX_ADDRESS			0x003c
+
+#define DST_WR_INDEX_ADDRESS			0x0040
+
+#define CURRENT_SRRI_ADDRESS			0x0044
+
+#define CURRENT_DRRI_ADDRESS			0x0048
+
+#define SRC_WATERMARK_LOW_MSB			31
+#define SRC_WATERMARK_LOW_LSB			16
+#define SRC_WATERMARK_LOW_MASK			0xffff0000
+#define SRC_WATERMARK_LOW_GET(x) \
+	(((x) & SRC_WATERMARK_LOW_MASK) >> SRC_WATERMARK_LOW_LSB)
+#define SRC_WATERMARK_LOW_SET(x) \
+	(((0 | (x)) << SRC_WATERMARK_LOW_LSB) & SRC_WATERMARK_LOW_MASK)
+#define SRC_WATERMARK_LOW_RESET			0
+#define SRC_WATERMARK_HIGH_MSB			15
+#define SRC_WATERMARK_HIGH_LSB			0
+#define SRC_WATERMARK_HIGH_MASK			0x0000ffff
+#define SRC_WATERMARK_HIGH_GET(x) \
+	(((x) & SRC_WATERMARK_HIGH_MASK) >> SRC_WATERMARK_HIGH_LSB)
+#define SRC_WATERMARK_HIGH_SET(x) \
+	(((0 | (x)) << SRC_WATERMARK_HIGH_LSB) & SRC_WATERMARK_HIGH_MASK)
+#define SRC_WATERMARK_HIGH_RESET		0
+#define SRC_WATERMARK_ADDRESS			0x004c
+
+#define DST_WATERMARK_LOW_LSB			16
+#define DST_WATERMARK_LOW_MASK			0xffff0000
+#define DST_WATERMARK_LOW_SET(x) \
+	(((0 | (x)) << DST_WATERMARK_LOW_LSB) & DST_WATERMARK_LOW_MASK)
+#define DST_WATERMARK_LOW_RESET			0
+#define DST_WATERMARK_HIGH_MSB			15
+#define DST_WATERMARK_HIGH_LSB			0
+#define DST_WATERMARK_HIGH_MASK			0x0000ffff
+#define DST_WATERMARK_HIGH_GET(x) \
+	(((x) & DST_WATERMARK_HIGH_MASK) >> DST_WATERMARK_HIGH_LSB)
+#define DST_WATERMARK_HIGH_SET(x) \
+	(((0 | (x)) << DST_WATERMARK_HIGH_LSB) & DST_WATERMARK_HIGH_MASK)
+#define DST_WATERMARK_HIGH_RESET		0
+#define DST_WATERMARK_ADDRESS			0x0050
+
+
+static inline u32 ath10k_ce_base_address(unsigned int ce_id)
+{
+	return CE0_BASE_ADDRESS + (CE1_BASE_ADDRESS - CE0_BASE_ADDRESS) * ce_id;
+}
+
+#define CE_WATERMARK_MASK (HOST_IS_SRC_RING_LOW_WATERMARK_MASK  | \
+			   HOST_IS_SRC_RING_HIGH_WATERMARK_MASK | \
+			   HOST_IS_DST_RING_LOW_WATERMARK_MASK  | \
+			   HOST_IS_DST_RING_HIGH_WATERMARK_MASK)
+
+#define CE_ERROR_MASK	(MISC_IS_AXI_ERR_MASK           | \
+			 MISC_IS_DST_ADDR_ERR_MASK      | \
+			 MISC_IS_SRC_LEN_ERR_MASK       | \
+			 MISC_IS_DST_MAX_LEN_VIO_MASK   | \
+			 MISC_IS_DST_RING_OVERFLOW_MASK | \
+			 MISC_IS_SRC_RING_OVERFLOW_MASK)
+
+#define CE_SRC_RING_TO_DESC(baddr, idx) \
+	(&(((struct ce_desc *)baddr)[idx]))
+
+#define CE_DEST_RING_TO_DESC(baddr, idx) \
+	(&(((struct ce_desc *)baddr)[idx]))
+
+/* Ring arithmetic (modulus number of entries in ring, which is a pwr of 2). */
+#define CE_RING_DELTA(nentries_mask, fromidx, toidx) \
+	(((int)(toidx)-(int)(fromidx)) & (nentries_mask))
+
+#define CE_RING_IDX_INCR(nentries_mask, idx) (((idx) + 1) & (nentries_mask))
+
+#define CE_WRAPPER_INTERRUPT_SUMMARY_HOST_MSI_LSB		8
+#define CE_WRAPPER_INTERRUPT_SUMMARY_HOST_MSI_MASK		0x0000ff00
+#define CE_WRAPPER_INTERRUPT_SUMMARY_HOST_MSI_GET(x) \
+	(((x) & CE_WRAPPER_INTERRUPT_SUMMARY_HOST_MSI_MASK) >> \
+		CE_WRAPPER_INTERRUPT_SUMMARY_HOST_MSI_LSB)
+#define CE_WRAPPER_INTERRUPT_SUMMARY_ADDRESS			0x0000
+
+#define CE_INTERRUPT_SUMMARY(ar) \
+	CE_WRAPPER_INTERRUPT_SUMMARY_HOST_MSI_GET( \
+		ath10k_pci_read32((ar), CE_WRAPPER_BASE_ADDRESS + \
+		CE_WRAPPER_INTERRUPT_SUMMARY_ADDRESS))
+
+#endif /* _CE_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/core.c b/drivers/net/wireless/ath/ath10k/core.c
new file mode 100644
index 0000000..2b3426b
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/core.c
@@ -0,0 +1,665 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/firmware.h>
+
+#include "core.h"
+#include "mac.h"
+#include "htc.h"
+#include "hif.h"
+#include "wmi.h"
+#include "bmi.h"
+#include "debug.h"
+#include "htt.h"
+
+unsigned int ath10k_debug_mask;
+static bool uart_print;
+static unsigned int ath10k_p2p;
+module_param_named(debug_mask, ath10k_debug_mask, uint, 0644);
+module_param(uart_print, bool, 0644);
+module_param_named(p2p, ath10k_p2p, uint, 0644);
+MODULE_PARM_DESC(debug_mask, "Debugging mask");
+MODULE_PARM_DESC(uart_print, "Uart target debugging");
+MODULE_PARM_DESC(p2p, "Enable ath10k P2P support");
+
+static const struct ath10k_hw_params ath10k_hw_params_list[] = {
+	{
+		.id = QCA988X_HW_1_0_VERSION,
+		.name = "qca988x hw1.0",
+		.patch_load_addr = QCA988X_HW_1_0_PATCH_LOAD_ADDR,
+		.fw = {
+			.dir = QCA988X_HW_1_0_FW_DIR,
+			.fw = QCA988X_HW_1_0_FW_FILE,
+			.otp = QCA988X_HW_1_0_OTP_FILE,
+			.board = QCA988X_HW_1_0_BOARD_DATA_FILE,
+		},
+	},
+	{
+		.id = QCA988X_HW_2_0_VERSION,
+		.name = "qca988x hw2.0",
+		.patch_load_addr = QCA988X_HW_2_0_PATCH_LOAD_ADDR,
+		.fw = {
+			.dir = QCA988X_HW_2_0_FW_DIR,
+			.fw = QCA988X_HW_2_0_FW_FILE,
+			.otp = QCA988X_HW_2_0_OTP_FILE,
+			.board = QCA988X_HW_2_0_BOARD_DATA_FILE,
+		},
+	},
+};
+
+static void ath10k_send_suspend_complete(struct ath10k *ar)
+{
+	ath10k_dbg(ATH10K_DBG_CORE, "%s\n", __func__);
+
+	ar->is_target_paused = true;
+	wake_up(&ar->event_queue);
+}
+
+static int ath10k_check_fw_version(struct ath10k *ar)
+{
+	char version[32];
+
+	if (ar->fw_version_major >= SUPPORTED_FW_MAJOR &&
+	    ar->fw_version_minor >= SUPPORTED_FW_MINOR &&
+	    ar->fw_version_release >= SUPPORTED_FW_RELEASE &&
+	    ar->fw_version_build >= SUPPORTED_FW_BUILD)
+		return 0;
+
+	snprintf(version, sizeof(version), "%u.%u.%u.%u",
+		 SUPPORTED_FW_MAJOR, SUPPORTED_FW_MINOR,
+		 SUPPORTED_FW_RELEASE, SUPPORTED_FW_BUILD);
+
+	ath10k_warn("WARNING: Firmware version %s is not officially supported.\n",
+		    ar->hw->wiphy->fw_version);
+	ath10k_warn("Please upgrade to version %s (or newer)\n", version);
+
+	return 0;
+}
+
+static int ath10k_init_connect_htc(struct ath10k *ar)
+{
+	int status;
+
+	status = ath10k_wmi_connect_htc_service(ar);
+	if (status)
+		goto conn_fail;
+
+	/* Start HTC */
+	status = ath10k_htc_start(ar->htc);
+	if (status)
+		goto conn_fail;
+
+	/* Wait for WMI event to be ready */
+	status = ath10k_wmi_wait_for_service_ready(ar);
+	if (status <= 0) {
+		ath10k_warn("wmi service ready event not received");
+		status = -ETIMEDOUT;
+		goto timeout;
+	}
+
+	ath10k_dbg(ATH10K_DBG_CORE, "core wmi ready\n");
+	return 0;
+
+timeout:
+	ath10k_htc_stop(ar->htc);
+conn_fail:
+	return status;
+}
+
+static int ath10k_init_configure_target(struct ath10k *ar)
+{
+	u32 param_host;
+	int ret;
+
+	/* tell target which HTC version it is used*/
+	ret = ath10k_bmi_write32(ar, hi_app_host_interest,
+				 HTC_PROTOCOL_VERSION);
+	if (ret) {
+		ath10k_err("settings HTC version failed\n");
+		return ret;
+	}
+
+	/* set the firmware mode to STA/IBSS/AP */
+	ret = ath10k_bmi_read32(ar, hi_option_flag, &param_host);
+	if (ret) {
+		ath10k_err("setting firmware mode (1/2) failed\n");
+		return ret;
+	}
+
+	/* TODO following parameters need to be re-visited. */
+	/* num_device */
+	param_host |= (1 << HI_OPTION_NUM_DEV_SHIFT);
+	/* Firmware mode */
+	/* FIXME: Why FW_MODE_AP ??.*/
+	param_host |= (HI_OPTION_FW_MODE_AP << HI_OPTION_FW_MODE_SHIFT);
+	/* mac_addr_method */
+	param_host |= (1 << HI_OPTION_MAC_ADDR_METHOD_SHIFT);
+	/* firmware_bridge */
+	param_host |= (0 << HI_OPTION_FW_BRIDGE_SHIFT);
+	/* fwsubmode */
+	param_host |= (0 << HI_OPTION_FW_SUBMODE_SHIFT);
+
+	ret = ath10k_bmi_write32(ar, hi_option_flag, param_host);
+	if (ret) {
+		ath10k_err("setting firmware mode (2/2) failed\n");
+		return ret;
+	}
+
+	/* We do all byte-swapping on the host */
+	ret = ath10k_bmi_write32(ar, hi_be, 0);
+	if (ret) {
+		ath10k_err("setting host CPU BE mode failed\n");
+		return ret;
+	}
+
+	/* FW descriptor/Data swap flags */
+	ret = ath10k_bmi_write32(ar, hi_fw_swap, 0);
+
+	if (ret) {
+		ath10k_err("setting FW data/desc swap flags failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+static const struct firmware *ath10k_fetch_fw_file(struct ath10k *ar,
+						   const char *dir,
+						   const char *file)
+{
+	char filename[100];
+	const struct firmware *fw;
+	int ret;
+
+	if (file == NULL)
+		return ERR_PTR(-ENOENT);
+
+	if (dir == NULL)
+		dir = ".";
+
+	snprintf(filename, sizeof(filename), "%s/%s", dir, file);
+	ret = request_firmware(&fw, filename, ar->dev);
+	if (ret)
+		return ERR_PTR(ret);
+
+	return fw;
+}
+
+static int ath10k_push_board_ext_data(struct ath10k *ar,
+				      const struct firmware *fw)
+{
+	u32 board_data_size = QCA988X_BOARD_DATA_SZ;
+	u32 board_ext_data_size = QCA988X_BOARD_EXT_DATA_SZ;
+	u32 board_ext_data_addr;
+	int ret;
+
+	ret = ath10k_bmi_read32(ar, hi_board_ext_data, &board_ext_data_addr);
+	if (ret) {
+		ath10k_err("could not read board ext data addr (%d)\n", ret);
+		return ret;
+	}
+
+	ath10k_dbg(ATH10K_DBG_CORE,
+		   "ath10k: Board extended Data download addr: 0x%x\n",
+		   board_ext_data_addr);
+
+	if (board_ext_data_addr == 0)
+		return 0;
+
+	if (fw->size != (board_data_size + board_ext_data_size)) {
+		ath10k_err("invalid board (ext) data sizes %zu != %d+%d\n",
+			   fw->size, board_data_size, board_ext_data_size);
+		return -EINVAL;
+	}
+
+	ret = ath10k_bmi_write_memory(ar, board_ext_data_addr,
+				      fw->data + board_data_size,
+				      board_ext_data_size);
+	if (ret) {
+		ath10k_err("could not write board ext data (%d)\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_bmi_write32(ar, hi_board_ext_data_config,
+				 (board_ext_data_size << 16) | 1);
+	if (ret) {
+		ath10k_err("could not write board ext data bit (%d)\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int ath10k_download_board_data(struct ath10k *ar)
+{
+	u32 board_data_size = QCA988X_BOARD_DATA_SZ;
+	u32 address;
+	const struct firmware *fw;
+	int ret;
+
+	fw = ath10k_fetch_fw_file(ar, ar->hw_params.fw.dir,
+				  ar->hw_params.fw.board);
+	if (IS_ERR(fw)) {
+		ath10k_err("could not fetch board data fw file (%ld)\n",
+			   PTR_ERR(fw));
+		return PTR_ERR(fw);
+	}
+
+	ret = ath10k_push_board_ext_data(ar, fw);
+	if (ret) {
+		ath10k_err("could not push board ext data (%d)\n", ret);
+		goto exit;
+	}
+
+	ret = ath10k_bmi_read32(ar, hi_board_data, &address);
+	if (ret) {
+		ath10k_err("could not read board data addr (%d)\n", ret);
+		goto exit;
+	}
+
+	ret = ath10k_bmi_write_memory(ar, address, fw->data,
+				      min_t(u32, board_data_size, fw->size));
+	if (ret) {
+		ath10k_err("could not write board data (%d)\n", ret);
+		goto exit;
+	}
+
+	ret = ath10k_bmi_write32(ar, hi_board_data_initialized, 1);
+	if (ret) {
+		ath10k_err("could not write board data bit (%d)\n", ret);
+		goto exit;
+	}
+
+exit:
+	release_firmware(fw);
+	return ret;
+}
+
+static int ath10k_download_and_run_otp(struct ath10k *ar)
+{
+	const struct firmware *fw;
+	u32 address;
+	u32 exec_param;
+	int ret;
+
+	/* OTP is optional */
+
+	if (ar->hw_params.fw.otp == NULL) {
+		ath10k_info("otp file not defined\n");
+		return 0;
+	}
+
+	address = ar->hw_params.patch_load_addr;
+
+	fw = ath10k_fetch_fw_file(ar, ar->hw_params.fw.dir,
+				  ar->hw_params.fw.otp);
+	if (IS_ERR(fw)) {
+		ath10k_warn("could not fetch otp (%ld)\n", PTR_ERR(fw));
+		return 0;
+	}
+
+	ret = ath10k_bmi_fast_download(ar, address, fw->data, fw->size);
+	if (ret) {
+		ath10k_err("could not write otp (%d)\n", ret);
+		goto exit;
+	}
+
+	exec_param = 0;
+	ret = ath10k_bmi_execute(ar, address, &exec_param);
+	if (ret) {
+		ath10k_err("could not execute otp (%d)\n", ret);
+		goto exit;
+	}
+
+exit:
+	release_firmware(fw);
+	return ret;
+}
+
+static int ath10k_download_fw(struct ath10k *ar)
+{
+	const struct firmware *fw;
+	u32 address;
+	int ret;
+
+	if (ar->hw_params.fw.fw == NULL)
+		return -EINVAL;
+
+	address = ar->hw_params.patch_load_addr;
+
+	fw = ath10k_fetch_fw_file(ar, ar->hw_params.fw.dir,
+				  ar->hw_params.fw.fw);
+	if (IS_ERR(fw)) {
+		ath10k_err("could not fetch fw (%ld)\n", PTR_ERR(fw));
+		return PTR_ERR(fw);
+	}
+
+	ret = ath10k_bmi_fast_download(ar, address, fw->data, fw->size);
+	if (ret) {
+		ath10k_err("could not write fw (%d)\n", ret);
+		goto exit;
+	}
+
+exit:
+	release_firmware(fw);
+	return ret;
+}
+
+static int ath10k_init_download_firmware(struct ath10k *ar)
+{
+	int ret;
+
+	ret = ath10k_download_board_data(ar);
+	if (ret)
+		return ret;
+
+	ret = ath10k_download_and_run_otp(ar);
+	if (ret)
+		return ret;
+
+	ret = ath10k_download_fw(ar);
+	if (ret)
+		return ret;
+
+	return ret;
+}
+
+static int ath10k_init_uart(struct ath10k *ar)
+{
+	int ret;
+
+	/*
+	 * Explicitly setting UART prints to zero as target turns it on
+	 * based on scratch registers.
+	 */
+	ret = ath10k_bmi_write32(ar, hi_serial_enable, 0);
+	if (ret) {
+		ath10k_warn("could not disable UART prints (%d)\n", ret);
+		return ret;
+	}
+
+	if (!uart_print) {
+		ath10k_info("UART prints disabled\n");
+		return 0;
+	}
+
+	ret = ath10k_bmi_write32(ar, hi_dbg_uart_txpin, 7);
+	if (ret) {
+		ath10k_warn("could not enable UART prints (%d)\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_bmi_write32(ar, hi_serial_enable, 1);
+	if (ret) {
+		ath10k_warn("could not enable UART prints (%d)\n", ret);
+		return ret;
+	}
+
+	ath10k_info("UART prints enabled\n");
+	return 0;
+}
+
+static int ath10k_init_hw_params(struct ath10k *ar)
+{
+	const struct ath10k_hw_params *uninitialized_var(hw_params);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ath10k_hw_params_list); i++) {
+		hw_params = &ath10k_hw_params_list[i];
+
+		if (hw_params->id == ar->target_version)
+			break;
+	}
+
+	if (i == ARRAY_SIZE(ath10k_hw_params_list)) {
+		ath10k_err("Unsupported hardware version: 0x%x\n",
+			   ar->target_version);
+		return -EINVAL;
+	}
+
+	ar->hw_params = *hw_params;
+
+	ath10k_info("Hardware name %s version 0x%x\n",
+		    ar->hw_params.name, ar->target_version);
+
+	return 0;
+}
+
+struct ath10k *ath10k_core_create(void *hif_priv, struct device *dev,
+				  enum ath10k_bus bus,
+				  const struct ath10k_hif_ops *hif_ops)
+{
+	struct ath10k *ar;
+
+	ar = ath10k_mac_create();
+	if (!ar)
+		return NULL;
+
+	ar->ath_common.priv = ar;
+	ar->ath_common.hw = ar->hw;
+
+	ar->p2p = !!ath10k_p2p;
+	ar->dev = dev;
+
+	ar->hif.priv = hif_priv;
+	ar->hif.ops = hif_ops;
+	ar->hif.bus = bus;
+
+	ar->free_vdev_map = 0xFF; /* 8 vdevs */
+
+	init_completion(&ar->scan.started);
+	init_completion(&ar->scan.completed);
+	init_completion(&ar->scan.on_channel);
+
+	init_completion(&ar->install_key_done);
+	init_completion(&ar->vdev_setup_done);
+
+	setup_timer(&ar->scan.timeout, ath10k_reset_scan, (unsigned long)ar);
+
+	ar->workqueue = create_singlethread_workqueue("ath10k_wq");
+	if (!ar->workqueue)
+		goto err_wq;
+
+	mutex_init(&ar->conf_mutex);
+	spin_lock_init(&ar->data_lock);
+
+	INIT_LIST_HEAD(&ar->peers);
+	init_waitqueue_head(&ar->peer_mapping_wq);
+
+	init_completion(&ar->offchan_tx_completed);
+	INIT_WORK(&ar->offchan_tx_work, ath10k_offchan_tx_work);
+	skb_queue_head_init(&ar->offchan_tx_queue);
+
+	init_waitqueue_head(&ar->event_queue);
+
+	return ar;
+
+err_wq:
+	ath10k_mac_destroy(ar);
+	return NULL;
+}
+EXPORT_SYMBOL(ath10k_core_create);
+
+void ath10k_core_destroy(struct ath10k *ar)
+{
+	flush_workqueue(ar->workqueue);
+	destroy_workqueue(ar->workqueue);
+
+	ath10k_mac_destroy(ar);
+}
+EXPORT_SYMBOL(ath10k_core_destroy);
+
+
+int ath10k_core_register(struct ath10k *ar)
+{
+	struct ath10k_htc_ops htc_ops;
+	struct bmi_target_info target_info;
+	int status;
+
+	memset(&target_info, 0, sizeof(target_info));
+	status = ath10k_bmi_get_target_info(ar, &target_info);
+	if (status)
+		goto err;
+
+	ar->target_version = target_info.version;
+	ar->hw->wiphy->hw_version = target_info.version;
+
+	status = ath10k_init_hw_params(ar);
+	if (status)
+		goto err;
+
+	if (ath10k_init_configure_target(ar)) {
+		status = -EINVAL;
+		goto err;
+	}
+
+	status = ath10k_init_download_firmware(ar);
+	if (status)
+		goto err;
+
+	status = ath10k_init_uart(ar);
+	if (status)
+		goto err;
+
+	htc_ops.target_send_suspend_complete = ath10k_send_suspend_complete;
+
+	ar->htc = ath10k_htc_create(ar, &htc_ops);
+	if (IS_ERR(ar->htc)) {
+		status = PTR_ERR(ar->htc);
+		ath10k_err("could not create HTC (%d)\n", status);
+		goto err;
+	}
+
+	status = ath10k_bmi_done(ar);
+	if (status)
+		goto err_htc_destroy;
+
+	status = ath10k_wmi_attach(ar);
+	if (status) {
+		ath10k_err("WMI attach failed: %d\n", status);
+		goto err_htc_destroy;
+	}
+
+	status = ath10k_htc_wait_target(ar->htc);
+	if (status)
+		goto err_wmi_detach;
+
+	ar->htt = ath10k_htt_attach(ar);
+	if (!ar->htt) {
+		status = -ENOMEM;
+		goto err_wmi_detach;
+	}
+
+	status = ath10k_init_connect_htc(ar);
+	if (status)
+		goto err_htt_detach;
+
+	ath10k_info("firmware %s booted\n", ar->hw->wiphy->fw_version);
+
+	status = ath10k_check_fw_version(ar);
+	if (status)
+		goto err_disconnect_htc;
+
+	status = ath10k_wmi_cmd_init(ar);
+	if (status) {
+		ath10k_err("could not send WMI init command (%d)\n", status);
+		goto err_disconnect_htc;
+	}
+
+	status = ath10k_wmi_wait_for_unified_ready(ar);
+	if (status <= 0) {
+		ath10k_err("wmi unified ready event not received\n");
+		status = -ETIMEDOUT;
+		goto err_disconnect_htc;
+	}
+
+	status = ath10k_htt_attach_target(ar->htt);
+	if (status)
+		goto err_disconnect_htc;
+
+	status = ath10k_mac_register(ar);
+	if (status)
+		goto err_disconnect_htc;
+
+	status = ath10k_debug_create(ar);
+	if (status) {
+		ath10k_err("unable to initialize debugfs\n");
+		goto err_unregister_mac;
+	}
+
+	return 0;
+
+err_unregister_mac:
+	ath10k_mac_unregister(ar);
+err_disconnect_htc:
+	ath10k_htc_stop(ar->htc);
+err_htt_detach:
+	ath10k_htt_detach(ar->htt);
+err_wmi_detach:
+	ath10k_wmi_detach(ar);
+err_htc_destroy:
+	ath10k_htc_destroy(ar->htc);
+err:
+	return status;
+}
+EXPORT_SYMBOL(ath10k_core_register);
+
+void ath10k_core_unregister(struct ath10k *ar)
+{
+	/* We must unregister from mac80211 before we stop HTC and HIF.
+	 * Otherwise we will fail to submit commands to FW and mac80211 will be
+	 * unhappy about callback failures. */
+	ath10k_mac_unregister(ar);
+	ath10k_htc_stop(ar->htc);
+	ath10k_htt_detach(ar->htt);
+	ath10k_wmi_detach(ar);
+	ath10k_htc_destroy(ar->htc);
+}
+EXPORT_SYMBOL(ath10k_core_unregister);
+
+int ath10k_core_target_suspend(struct ath10k *ar)
+{
+	int ret;
+
+	ath10k_dbg(ATH10K_DBG_CORE, "%s: called", __func__);
+
+	ret = ath10k_wmi_pdev_suspend_target(ar);
+	if (ret)
+		ath10k_warn("could not suspend target (%d)\n", ret);
+
+	return ret;
+}
+EXPORT_SYMBOL(ath10k_core_target_suspend);
+
+int ath10k_core_target_resume(struct ath10k *ar)
+{
+	int ret;
+
+	ath10k_dbg(ATH10K_DBG_CORE, "%s: called", __func__);
+
+	ret = ath10k_wmi_pdev_resume_target(ar);
+	if (ret)
+		ath10k_warn("could not resume target (%d)\n", ret);
+
+	return ret;
+}
+EXPORT_SYMBOL(ath10k_core_target_resume);
+
+MODULE_AUTHOR("Qualcomm Atheros");
+MODULE_DESCRIPTION("Core module for QCA988X PCIe devices.");
+MODULE_LICENSE("Dual BSD/GPL");
diff --git a/drivers/net/wireless/ath/ath10k/core.h b/drivers/net/wireless/ath/ath10k/core.h
new file mode 100644
index 0000000..539336d
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/core.h
@@ -0,0 +1,369 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _CORE_H_
+#define _CORE_H_
+
+#include <linux/completion.h>
+#include <linux/if_ether.h>
+#include <linux/types.h>
+#include <linux/pci.h>
+
+#include "htc.h"
+#include "hw.h"
+#include "targaddrs.h"
+#include "wmi.h"
+#include "../ath.h"
+#include "../regd.h"
+
+#define MS(_v, _f) (((_v) & _f##_MASK) >> _f##_LSB)
+#define SM(_v, _f) (((_v) << _f##_LSB) & _f##_MASK)
+#define WO(_f)      ((_f##_OFFSET) >> 2)
+
+#define ATH10K_SCAN_ID 0
+#define WMI_READY_TIMEOUT (5 * HZ)
+#define ATH10K_FLUSH_TIMEOUT_HZ (5*HZ)
+
+/* Antenna noise floor */
+#define ATH10K_DEFAULT_NOISE_FLOOR -95
+
+struct ath10k;
+
+enum ath10k_bus {
+	ATH10K_BUS_PCI,
+};
+
+struct ath10k_skb_cb {
+	dma_addr_t paddr;
+	bool is_mapped;
+	bool is_aborted;
+
+	struct {
+		u8 vdev_id;
+		u16 msdu_id;
+		u8 tid;
+		bool is_offchan;
+		bool is_conf;
+		bool discard;
+		bool no_ack;
+		u8 refcount;
+		struct sk_buff *txfrag;
+		struct sk_buff *msdu;
+	} __packed htt;
+
+	/* 4 bytes left on 64bit arch */
+} __packed;
+
+static inline struct ath10k_skb_cb *ATH10K_SKB_CB(struct sk_buff *skb)
+{
+	BUILD_BUG_ON(sizeof(struct ath10k_skb_cb) >
+		     IEEE80211_TX_INFO_DRIVER_DATA_SIZE);
+	return (struct ath10k_skb_cb *)&IEEE80211_SKB_CB(skb)->driver_data;
+}
+
+static inline int ath10k_skb_map(struct device *dev, struct sk_buff *skb)
+{
+	if (ATH10K_SKB_CB(skb)->is_mapped)
+		return -EINVAL;
+
+	ATH10K_SKB_CB(skb)->paddr = dma_map_single(dev, skb->data, skb->len,
+						   DMA_TO_DEVICE);
+
+	if (unlikely(dma_mapping_error(dev, ATH10K_SKB_CB(skb)->paddr)))
+		return -EIO;
+
+	ATH10K_SKB_CB(skb)->is_mapped = true;
+	return 0;
+}
+
+static inline int ath10k_skb_unmap(struct device *dev, struct sk_buff *skb)
+{
+	if (!ATH10K_SKB_CB(skb)->is_mapped)
+		return -EINVAL;
+
+	dma_unmap_single(dev, ATH10K_SKB_CB(skb)->paddr, skb->len,
+			 DMA_TO_DEVICE);
+	ATH10K_SKB_CB(skb)->is_mapped = false;
+	return 0;
+}
+
+static inline u32 host_interest_item_address(u32 item_offset)
+{
+	return QCA988X_HOST_INTEREST_ADDRESS + item_offset;
+}
+
+struct ath10k_bmi {
+	bool done_sent;
+};
+
+struct ath10k_wmi {
+	enum ath10k_htc_ep_id eid;
+	struct completion service_ready;
+	struct completion unified_ready;
+	atomic_t pending_tx_count;
+	wait_queue_head_t wq;
+
+	struct sk_buff_head wmi_event_list;
+	struct work_struct wmi_event_work;
+};
+
+struct ath10k_peer_stat {
+	u8 peer_macaddr[ETH_ALEN];
+	u32 peer_rssi;
+	u32 peer_tx_rate;
+};
+
+struct ath10k_target_stats {
+	/* PDEV stats */
+	s32 ch_noise_floor;
+	u32 tx_frame_count;
+	u32 rx_frame_count;
+	u32 rx_clear_count;
+	u32 cycle_count;
+	u32 phy_err_count;
+	u32 chan_tx_power;
+
+	/* PDEV TX stats */
+	s32 comp_queued;
+	s32 comp_delivered;
+	s32 msdu_enqued;
+	s32 mpdu_enqued;
+	s32 wmm_drop;
+	s32 local_enqued;
+	s32 local_freed;
+	s32 hw_queued;
+	s32 hw_reaped;
+	s32 underrun;
+	s32 tx_abort;
+	s32 mpdus_requed;
+	u32 tx_ko;
+	u32 data_rc;
+	u32 self_triggers;
+	u32 sw_retry_failure;
+	u32 illgl_rate_phy_err;
+	u32 pdev_cont_xretry;
+	u32 pdev_tx_timeout;
+	u32 pdev_resets;
+	u32 phy_underrun;
+	u32 txop_ovf;
+
+	/* PDEV RX stats */
+	s32 mid_ppdu_route_change;
+	s32 status_rcvd;
+	s32 r0_frags;
+	s32 r1_frags;
+	s32 r2_frags;
+	s32 r3_frags;
+	s32 htt_msdus;
+	s32 htt_mpdus;
+	s32 loc_msdus;
+	s32 loc_mpdus;
+	s32 oversize_amsdu;
+	s32 phy_errs;
+	s32 phy_err_drop;
+	s32 mpdu_errs;
+
+	/* VDEV STATS */
+
+	/* PEER STATS */
+	u8 peers;
+	struct ath10k_peer_stat peer_stat[TARGET_NUM_PEERS];
+
+	/* TODO: Beacon filter stats */
+
+};
+
+#define ATH10K_MAX_NUM_PEER_IDS (1 << 11) /* htt rx_desc limit */
+
+struct ath10k_peer {
+	struct list_head list;
+	int vdev_id;
+	u8 addr[ETH_ALEN];
+	DECLARE_BITMAP(peer_ids, ATH10K_MAX_NUM_PEER_IDS);
+	struct ieee80211_key_conf *keys[WMI_MAX_KEY_INDEX + 1];
+};
+
+#define ATH10K_VDEV_SETUP_TIMEOUT_HZ (5*HZ)
+
+struct ath10k_vif {
+	u32 vdev_id;
+	enum wmi_vdev_type vdev_type;
+	enum wmi_vdev_subtype vdev_subtype;
+	u32 beacon_interval;
+	u32 dtim_period;
+
+	struct ath10k *ar;
+	struct ieee80211_vif *vif;
+
+	struct ieee80211_key_conf *wep_keys[WMI_MAX_KEY_INDEX + 1];
+	u8 def_wep_key_index;
+
+	u16 tx_seq_no;
+
+	union {
+		struct {
+			u8 bssid[ETH_ALEN];
+			u32 uapsd;
+		} sta;
+		struct {
+			/* 127 stations; wmi limit */
+			u8 tim_bitmap[16];
+			u8 tim_len;
+			u32 ssid_len;
+			u8 ssid[IEEE80211_MAX_SSID_LEN];
+			bool hidden_ssid;
+			/* P2P_IE with NoA attribute for P2P_GO case */
+			u32 noa_len;
+			u8 *noa_data;
+		} ap;
+		struct {
+			u8 bssid[ETH_ALEN];
+		} ibss;
+	} u;
+};
+
+struct ath10k_vif_iter {
+	u32 vdev_id;
+	struct ath10k_vif *arvif;
+};
+
+struct ath10k_debug {
+	struct dentry *debugfs_phy;
+
+	struct ath10k_target_stats target_stats;
+	u32 wmi_service_bitmap[WMI_SERVICE_BM_SIZE];
+
+	struct completion event_stats_compl;
+};
+
+struct ath10k {
+	struct ath_common ath_common;
+	struct ieee80211_hw *hw;
+	struct device *dev;
+	u8 mac_addr[ETH_ALEN];
+
+	u32 target_version;
+	u8 fw_version_major;
+	u32 fw_version_minor;
+	u16 fw_version_release;
+	u16 fw_version_build;
+	u32 phy_capability;
+	u32 hw_min_tx_power;
+	u32 hw_max_tx_power;
+	u32 ht_cap_info;
+	u32 vht_cap_info;
+
+	struct targetdef *targetdef;
+	struct hostdef *hostdef;
+
+	bool p2p;
+
+	struct {
+		void *priv;
+		enum ath10k_bus bus;
+		const struct ath10k_hif_ops *ops;
+	} hif;
+
+	struct ath10k_wmi wmi;
+
+	wait_queue_head_t event_queue;
+	bool is_target_paused;
+
+	struct ath10k_bmi bmi;
+
+	struct ath10k_htc *htc;
+	struct ath10k_htt *htt;
+
+	struct ath10k_hw_params {
+		u32 id;
+		const char *name;
+		u32 patch_load_addr;
+
+		struct ath10k_hw_params_fw {
+			const char *dir;
+			const char *fw;
+			const char *otp;
+			const char *board;
+		} fw;
+	} hw_params;
+
+	struct {
+		struct completion started;
+		struct completion completed;
+		struct completion on_channel;
+		struct timer_list timeout;
+		bool is_roc;
+		bool in_progress;
+		bool aborting;
+		int vdev_id;
+		int roc_freq;
+	} scan;
+
+	struct {
+		struct ieee80211_supported_band sbands[IEEE80211_NUM_BANDS];
+	} mac;
+
+	/* should never be NULL; needed for regular htt rx */
+	struct ieee80211_channel *rx_channel;
+
+	/* valid during scan; needed for mgmt rx during scan */
+	struct ieee80211_channel *scan_channel;
+
+	int free_vdev_map;
+	int monitor_vdev_id;
+	bool monitor_enabled;
+	bool monitor_present;
+	unsigned int filter_flags;
+
+	struct wmi_pdev_set_wmm_params_arg wmm_params;
+	struct completion install_key_done;
+
+	struct completion vdev_setup_done;
+
+	struct workqueue_struct *workqueue;
+
+	/* prevents concurrent FW reconfiguration */
+	struct mutex conf_mutex;
+
+	/* protects shared structure data */
+	spinlock_t data_lock;
+
+	struct list_head peers;
+	wait_queue_head_t peer_mapping_wq;
+
+	struct work_struct offchan_tx_work;
+	struct sk_buff_head offchan_tx_queue;
+	struct completion offchan_tx_completed;
+	struct sk_buff *offchan_tx_skb;
+
+#ifdef CONFIG_ATH10K_DEBUGFS
+	struct ath10k_debug debug;
+#endif
+};
+
+struct ath10k *ath10k_core_create(void *hif_priv, struct device *dev,
+				  enum ath10k_bus bus,
+				  const struct ath10k_hif_ops *hif_ops);
+void ath10k_core_destroy(struct ath10k *ar);
+
+int ath10k_core_register(struct ath10k *ar);
+void ath10k_core_unregister(struct ath10k *ar);
+
+int ath10k_core_target_suspend(struct ath10k *ar);
+int ath10k_core_target_resume(struct ath10k *ar);
+
+#endif /* _CORE_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/debug.c b/drivers/net/wireless/ath/ath10k/debug.c
new file mode 100644
index 0000000..499034b
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/debug.c
@@ -0,0 +1,503 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/debugfs.h>
+
+#include "core.h"
+#include "debug.h"
+
+static int ath10k_printk(const char *level, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+	int rtn;
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	rtn = printk("%sath10k: %pV", level, &vaf);
+
+	va_end(args);
+
+	return rtn;
+}
+
+int ath10k_info(const char *fmt, ...)
+{
+	struct va_format vaf = {
+		.fmt = fmt,
+	};
+	va_list args;
+	int ret;
+
+	va_start(args, fmt);
+	vaf.va = &args;
+	ret = ath10k_printk(KERN_INFO, "%pV", &vaf);
+	trace_ath10k_log_info(&vaf);
+	va_end(args);
+
+	return ret;
+}
+EXPORT_SYMBOL(ath10k_info);
+
+int ath10k_err(const char *fmt, ...)
+{
+	struct va_format vaf = {
+		.fmt = fmt,
+	};
+	va_list args;
+	int ret;
+
+	va_start(args, fmt);
+	vaf.va = &args;
+	ret = ath10k_printk(KERN_ERR, "%pV", &vaf);
+	trace_ath10k_log_err(&vaf);
+	va_end(args);
+
+	return ret;
+}
+EXPORT_SYMBOL(ath10k_err);
+
+int ath10k_warn(const char *fmt, ...)
+{
+	struct va_format vaf = {
+		.fmt = fmt,
+	};
+	va_list args;
+	int ret = 0;
+
+	va_start(args, fmt);
+	vaf.va = &args;
+
+	if (net_ratelimit())
+		ret = ath10k_printk(KERN_WARNING, "%pV", &vaf);
+
+	trace_ath10k_log_warn(&vaf);
+
+	va_end(args);
+
+	return ret;
+}
+EXPORT_SYMBOL(ath10k_warn);
+
+#ifdef CONFIG_ATH10K_DEBUGFS
+
+void ath10k_debug_read_service_map(struct ath10k *ar,
+				   void *service_map,
+				   size_t map_size)
+{
+	memcpy(ar->debug.wmi_service_bitmap, service_map, map_size);
+}
+
+static ssize_t ath10k_read_wmi_services(struct file *file,
+					char __user *user_buf,
+					size_t count, loff_t *ppos)
+{
+	struct ath10k *ar = file->private_data;
+	char *buf;
+	unsigned int len = 0, buf_len = 1500;
+	const char *status;
+	ssize_t ret_cnt;
+	int i;
+
+	buf = kzalloc(buf_len, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	mutex_lock(&ar->conf_mutex);
+
+	if (len > buf_len)
+		len = buf_len;
+
+	for (i = 0; i < WMI_SERVICE_LAST; i++) {
+		if (WMI_SERVICE_IS_ENABLED(ar->debug.wmi_service_bitmap, i))
+			status = "enabled";
+		else
+			status = "disabled";
+
+		len += scnprintf(buf + len, buf_len - len,
+				 "0x%02x - %20s - %s\n",
+				 i, wmi_service_name(i), status);
+	}
+
+	ret_cnt = simple_read_from_buffer(user_buf, count, ppos, buf, len);
+
+	mutex_unlock(&ar->conf_mutex);
+
+	kfree(buf);
+	return ret_cnt;
+}
+
+static const struct file_operations fops_wmi_services = {
+	.read = ath10k_read_wmi_services,
+	.open = simple_open,
+	.owner = THIS_MODULE,
+	.llseek = default_llseek,
+};
+
+void ath10k_debug_read_target_stats(struct ath10k *ar,
+				    struct wmi_stats_event *ev)
+{
+	u8 *tmp = ev->data;
+	struct ath10k_target_stats *stats;
+	int num_pdev_stats, num_vdev_stats, num_peer_stats;
+	struct wmi_pdev_stats *ps;
+	int i;
+
+	mutex_lock(&ar->conf_mutex);
+
+	stats = &ar->debug.target_stats;
+
+	num_pdev_stats = __le32_to_cpu(ev->num_pdev_stats); /* 0 or 1 */
+	num_vdev_stats = __le32_to_cpu(ev->num_vdev_stats); /* 0 or max vdevs */
+	num_peer_stats = __le32_to_cpu(ev->num_peer_stats); /* 0 or max peers */
+
+	if (num_pdev_stats) {
+		ps = (struct wmi_pdev_stats *)tmp;
+
+		stats->ch_noise_floor = __le32_to_cpu(ps->chan_nf);
+		stats->tx_frame_count = __le32_to_cpu(ps->tx_frame_count);
+		stats->rx_frame_count = __le32_to_cpu(ps->rx_frame_count);
+		stats->rx_clear_count = __le32_to_cpu(ps->rx_clear_count);
+		stats->cycle_count = __le32_to_cpu(ps->cycle_count);
+		stats->phy_err_count = __le32_to_cpu(ps->phy_err_count);
+		stats->chan_tx_power = __le32_to_cpu(ps->chan_tx_pwr);
+
+		stats->comp_queued = __le32_to_cpu(ps->wal.tx.comp_queued);
+		stats->comp_delivered =
+			__le32_to_cpu(ps->wal.tx.comp_delivered);
+		stats->msdu_enqued = __le32_to_cpu(ps->wal.tx.msdu_enqued);
+		stats->mpdu_enqued = __le32_to_cpu(ps->wal.tx.mpdu_enqued);
+		stats->wmm_drop = __le32_to_cpu(ps->wal.tx.wmm_drop);
+		stats->local_enqued = __le32_to_cpu(ps->wal.tx.local_enqued);
+		stats->local_freed = __le32_to_cpu(ps->wal.tx.local_freed);
+		stats->hw_queued = __le32_to_cpu(ps->wal.tx.hw_queued);
+		stats->hw_reaped = __le32_to_cpu(ps->wal.tx.hw_reaped);
+		stats->underrun = __le32_to_cpu(ps->wal.tx.underrun);
+		stats->tx_abort = __le32_to_cpu(ps->wal.tx.tx_abort);
+		stats->mpdus_requed = __le32_to_cpu(ps->wal.tx.mpdus_requed);
+		stats->tx_ko = __le32_to_cpu(ps->wal.tx.tx_ko);
+		stats->data_rc = __le32_to_cpu(ps->wal.tx.data_rc);
+		stats->self_triggers = __le32_to_cpu(ps->wal.tx.self_triggers);
+		stats->sw_retry_failure =
+			__le32_to_cpu(ps->wal.tx.sw_retry_failure);
+		stats->illgl_rate_phy_err =
+			__le32_to_cpu(ps->wal.tx.illgl_rate_phy_err);
+		stats->pdev_cont_xretry =
+			__le32_to_cpu(ps->wal.tx.pdev_cont_xretry);
+		stats->pdev_tx_timeout =
+			__le32_to_cpu(ps->wal.tx.pdev_tx_timeout);
+		stats->pdev_resets = __le32_to_cpu(ps->wal.tx.pdev_resets);
+		stats->phy_underrun = __le32_to_cpu(ps->wal.tx.phy_underrun);
+		stats->txop_ovf = __le32_to_cpu(ps->wal.tx.txop_ovf);
+
+		stats->mid_ppdu_route_change =
+			__le32_to_cpu(ps->wal.rx.mid_ppdu_route_change);
+		stats->status_rcvd = __le32_to_cpu(ps->wal.rx.status_rcvd);
+		stats->r0_frags = __le32_to_cpu(ps->wal.rx.r0_frags);
+		stats->r1_frags = __le32_to_cpu(ps->wal.rx.r1_frags);
+		stats->r2_frags = __le32_to_cpu(ps->wal.rx.r2_frags);
+		stats->r3_frags = __le32_to_cpu(ps->wal.rx.r3_frags);
+		stats->htt_msdus = __le32_to_cpu(ps->wal.rx.htt_msdus);
+		stats->htt_mpdus = __le32_to_cpu(ps->wal.rx.htt_mpdus);
+		stats->loc_msdus = __le32_to_cpu(ps->wal.rx.loc_msdus);
+		stats->loc_mpdus = __le32_to_cpu(ps->wal.rx.loc_mpdus);
+		stats->oversize_amsdu =
+			__le32_to_cpu(ps->wal.rx.oversize_amsdu);
+		stats->phy_errs = __le32_to_cpu(ps->wal.rx.phy_errs);
+		stats->phy_err_drop = __le32_to_cpu(ps->wal.rx.phy_err_drop);
+		stats->mpdu_errs = __le32_to_cpu(ps->wal.rx.mpdu_errs);
+
+		tmp += sizeof(struct wmi_pdev_stats);
+	}
+
+	/* 0 or max vdevs */
+	/* Currently firmware does not support VDEV stats */
+	if (num_vdev_stats) {
+		struct wmi_vdev_stats *vdev_stats;
+
+		for (i = 0; i < num_vdev_stats; i++) {
+			vdev_stats = (struct wmi_vdev_stats *)tmp;
+			tmp += sizeof(struct wmi_vdev_stats);
+		}
+	}
+
+	if (num_peer_stats) {
+		struct wmi_peer_stats *peer_stats;
+		struct ath10k_peer_stat *s;
+
+		stats->peers = num_peer_stats;
+
+		for (i = 0; i < num_peer_stats; i++) {
+			peer_stats = (struct wmi_peer_stats *)tmp;
+			s = &stats->peer_stat[i];
+
+			WMI_MAC_ADDR_TO_CHAR_ARRAY(&peer_stats->peer_macaddr,
+						   s->peer_macaddr);
+			s->peer_rssi = __le32_to_cpu(peer_stats->peer_rssi);
+			s->peer_tx_rate =
+				__le32_to_cpu(peer_stats->peer_tx_rate);
+
+			tmp += sizeof(struct wmi_peer_stats);
+		}
+	}
+
+	mutex_unlock(&ar->conf_mutex);
+	complete(&ar->debug.event_stats_compl);
+}
+
+static ssize_t ath10k_read_fw_stats(struct file *file, char __user *user_buf,
+				    size_t count, loff_t *ppos)
+{
+	struct ath10k *ar = file->private_data;
+	struct ath10k_target_stats *fw_stats;
+	char *buf;
+	unsigned int len = 0, buf_len = 2500;
+	ssize_t ret_cnt;
+	long left;
+	int i;
+	int ret;
+
+	fw_stats = &ar->debug.target_stats;
+
+	buf = kzalloc(buf_len, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	ret = ath10k_wmi_request_stats(ar, WMI_REQUEST_PEER_STAT);
+	if (ret) {
+		ath10k_warn("could not request stats (%d)\n", ret);
+		kfree(buf);
+		return -EIO;
+	}
+
+	left = wait_for_completion_timeout(&ar->debug.event_stats_compl, 1*HZ);
+
+	if (left <= 0) {
+		kfree(buf);
+		return -ETIMEDOUT;
+	}
+
+	mutex_lock(&ar->conf_mutex);
+
+	len += scnprintf(buf + len, buf_len - len, "\n");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n",
+			 "ath10k PDEV stats");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n\n",
+				 "=================");
+
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Channel noise floor", fw_stats->ch_noise_floor);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10u\n",
+			 "Channel TX power", fw_stats->chan_tx_power);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10u\n",
+			 "TX frame count", fw_stats->tx_frame_count);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10u\n",
+			 "RX frame count", fw_stats->rx_frame_count);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10u\n",
+			 "RX clear count", fw_stats->rx_clear_count);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10u\n",
+			 "Cycle count", fw_stats->cycle_count);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10u\n",
+			 "PHY error count", fw_stats->phy_err_count);
+
+	len += scnprintf(buf + len, buf_len - len, "\n");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n",
+			 "ath10k PDEV TX stats");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n\n",
+				 "=================");
+
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "HTT cookies queued", fw_stats->comp_queued);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "HTT cookies disp.", fw_stats->comp_delivered);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MSDU queued", fw_stats->msdu_enqued);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MPDU queued", fw_stats->mpdu_enqued);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MSDUs dropped", fw_stats->wmm_drop);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Local enqued", fw_stats->local_enqued);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Local freed", fw_stats->local_freed);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "HW queued", fw_stats->hw_queued);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "PPDUs reaped", fw_stats->hw_reaped);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Num underruns", fw_stats->underrun);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "PPDUs cleaned", fw_stats->tx_abort);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MPDUs requed", fw_stats->mpdus_requed);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Excessive retries", fw_stats->tx_ko);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "HW rate", fw_stats->data_rc);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Sched self tiggers", fw_stats->self_triggers);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Dropped due to SW retries",
+			 fw_stats->sw_retry_failure);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Illegal rate phy errors",
+			 fw_stats->illgl_rate_phy_err);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Pdev continous xretry", fw_stats->pdev_cont_xretry);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "TX timeout", fw_stats->pdev_tx_timeout);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "PDEV resets", fw_stats->pdev_resets);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "PHY underrun", fw_stats->phy_underrun);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MPDU is more than txop limit", fw_stats->txop_ovf);
+
+	len += scnprintf(buf + len, buf_len - len, "\n");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n",
+			 "ath10k PDEV RX stats");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n\n",
+				 "=================");
+
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Mid PPDU route change",
+			 fw_stats->mid_ppdu_route_change);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Tot. number of statuses", fw_stats->status_rcvd);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Extra frags on rings 0", fw_stats->r0_frags);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Extra frags on rings 1", fw_stats->r1_frags);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Extra frags on rings 2", fw_stats->r2_frags);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Extra frags on rings 3", fw_stats->r3_frags);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MSDUs delivered to HTT", fw_stats->htt_msdus);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MPDUs delivered to HTT", fw_stats->htt_mpdus);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MSDUs delivered to stack", fw_stats->loc_msdus);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MPDUs delivered to stack", fw_stats->loc_mpdus);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "Oversized AMSUs", fw_stats->oversize_amsdu);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "PHY errors", fw_stats->phy_errs);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "PHY errors drops", fw_stats->phy_err_drop);
+	len += scnprintf(buf + len, buf_len - len, "%30s %10d\n",
+			 "MPDU errors (FCS, MIC, ENC)", fw_stats->mpdu_errs);
+
+	len += scnprintf(buf + len, buf_len - len, "\n");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n",
+			 "ath10k PEER stats");
+	len += scnprintf(buf + len, buf_len - len, "%30s\n\n",
+				 "=================");
+
+	for (i = 0; i < fw_stats->peers; i++) {
+		len += scnprintf(buf + len, buf_len - len, "%30s %pM\n",
+				 "Peer MAC address",
+				 fw_stats->peer_stat[i].peer_macaddr);
+		len += scnprintf(buf + len, buf_len - len, "%30s %u\n",
+				 "Peer RSSI", fw_stats->peer_stat[i].peer_rssi);
+		len += scnprintf(buf + len, buf_len - len, "%30s %u\n",
+				 "Peer TX rate",
+				 fw_stats->peer_stat[i].peer_tx_rate);
+		len += scnprintf(buf + len, buf_len - len, "\n");
+	}
+
+	if (len > buf_len)
+		len = buf_len;
+
+	ret_cnt = simple_read_from_buffer(user_buf, count, ppos, buf, len);
+
+	mutex_unlock(&ar->conf_mutex);
+
+	kfree(buf);
+	return ret_cnt;
+}
+
+static const struct file_operations fops_fw_stats = {
+	.read = ath10k_read_fw_stats,
+	.open = simple_open,
+	.owner = THIS_MODULE,
+	.llseek = default_llseek,
+};
+
+int ath10k_debug_create(struct ath10k *ar)
+{
+	ar->debug.debugfs_phy = debugfs_create_dir("ath10k",
+						   ar->hw->wiphy->debugfsdir);
+
+	if (!ar->debug.debugfs_phy)
+		return -ENOMEM;
+
+	init_completion(&ar->debug.event_stats_compl);
+
+	debugfs_create_file("fw_stats", S_IRUSR, ar->debug.debugfs_phy, ar,
+			    &fops_fw_stats);
+
+	debugfs_create_file("wmi_services", S_IRUSR, ar->debug.debugfs_phy, ar,
+			    &fops_wmi_services);
+
+	return 0;
+}
+#endif /* CONFIG_ATH10K_DEBUGFS */
+
+#ifdef CONFIG_ATH10K_DEBUG
+void ath10k_dbg(enum ath10k_debug_mask mask, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	if (ath10k_debug_mask & mask)
+		ath10k_printk(KERN_DEBUG, "%pV", &vaf);
+
+	trace_ath10k_log_dbg(mask, &vaf);
+
+	va_end(args);
+}
+EXPORT_SYMBOL(ath10k_dbg);
+
+void ath10k_dbg_dump(enum ath10k_debug_mask mask,
+		     const char *msg, const char *prefix,
+		     const void *buf, size_t len)
+{
+	if (ath10k_debug_mask & mask) {
+		if (msg)
+			ath10k_dbg(mask, "%s\n", msg);
+
+		print_hex_dump_bytes(prefix, DUMP_PREFIX_OFFSET, buf, len);
+	}
+
+	/* tracing code doesn't like null strings :/ */
+	trace_ath10k_log_dbg_dump(msg ? msg : "", prefix ? prefix : "",
+				  buf, len);
+}
+EXPORT_SYMBOL(ath10k_dbg_dump);
+
+#endif /* CONFIG_ATH10K_DEBUG */
diff --git a/drivers/net/wireless/ath/ath10k/debug.h b/drivers/net/wireless/ath/ath10k/debug.h
new file mode 100644
index 0000000..168140c
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/debug.h
@@ -0,0 +1,90 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _DEBUG_H_
+#define _DEBUG_H_
+
+#include <linux/types.h>
+#include "trace.h"
+
+enum ath10k_debug_mask {
+	ATH10K_DBG_PCI		= 0x00000001,
+	ATH10K_DBG_WMI		= 0x00000002,
+	ATH10K_DBG_HTC		= 0x00000004,
+	ATH10K_DBG_HTT		= 0x00000008,
+	ATH10K_DBG_MAC		= 0x00000010,
+	ATH10K_DBG_CORE		= 0x00000020,
+	ATH10K_DBG_PCI_DUMP	= 0x00000040,
+	ATH10K_DBG_HTT_DUMP	= 0x00000080,
+	ATH10K_DBG_MGMT		= 0x00000100,
+	ATH10K_DBG_DATA		= 0x00000200,
+	ATH10K_DBG_ANY		= 0xffffffff,
+};
+
+extern unsigned int ath10k_debug_mask;
+
+extern __printf(1, 2) int ath10k_info(const char *fmt, ...);
+extern __printf(1, 2) int ath10k_err(const char *fmt, ...);
+extern __printf(1, 2) int ath10k_warn(const char *fmt, ...);
+
+#ifdef CONFIG_ATH10K_DEBUGFS
+int ath10k_debug_create(struct ath10k *ar);
+void ath10k_debug_read_service_map(struct ath10k *ar,
+				   void *service_map,
+				   size_t map_size);
+void ath10k_debug_read_target_stats(struct ath10k *ar,
+				    struct wmi_stats_event *ev);
+
+#else
+static inline int ath10k_debug_create(struct ath10k *ar)
+{
+	return 0;
+}
+
+static inline void ath10k_debug_read_service_map(struct ath10k *ar,
+						 void *service_map,
+						 size_t map_size)
+{
+}
+
+static inline void ath10k_debug_read_target_stats(struct ath10k *ar,
+						  struct wmi_stats_event *ev)
+{
+}
+#endif /* CONFIG_ATH10K_DEBUGFS */
+
+#ifdef CONFIG_ATH10K_DEBUG
+extern __printf(2, 3) void ath10k_dbg(enum ath10k_debug_mask mask,
+				      const char *fmt, ...);
+void ath10k_dbg_dump(enum ath10k_debug_mask mask,
+		     const char *msg, const char *prefix,
+		     const void *buf, size_t len);
+#else /* CONFIG_ATH10K_DEBUG */
+
+static inline int ath10k_dbg(enum ath10k_debug_mask dbg_mask,
+			     const char *fmt, ...)
+{
+	return 0;
+}
+
+static inline void ath10k_dbg_dump(enum ath10k_debug_mask mask,
+				   const char *msg, const char *prefix,
+				   const void *buf, size_t len)
+{
+}
+#endif /* CONFIG_ATH10K_DEBUG */
+#endif /* _DEBUG_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/hif.h b/drivers/net/wireless/ath/ath10k/hif.h
new file mode 100644
index 0000000..73a24d4
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/hif.h
@@ -0,0 +1,137 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _HIF_H_
+#define _HIF_H_
+
+#include <linux/kernel.h>
+#include "core.h"
+
+struct ath10k_hif_cb {
+	int (*tx_completion)(struct ath10k *ar,
+			     struct sk_buff *wbuf,
+			     unsigned transfer_id);
+	int (*rx_completion)(struct ath10k *ar,
+			     struct sk_buff *wbuf,
+			     u8 pipe_id);
+};
+
+struct ath10k_hif_ops {
+	/* Send the head of a buffer to HIF for transmission to the target. */
+	int (*send_head)(struct ath10k *ar, u8 pipe_id,
+			 unsigned int transfer_id,
+			 unsigned int nbytes,
+			 struct sk_buff *buf);
+
+	/*
+	 * API to handle HIF-specific BMI message exchanges, this API is
+	 * synchronous and only allowed to be called from a context that
+	 * can block (sleep)
+	 */
+	int (*exchange_bmi_msg)(struct ath10k *ar,
+				void *request, u32 request_len,
+				void *response, u32 *response_len);
+
+	int (*start)(struct ath10k *ar);
+
+	void (*stop)(struct ath10k *ar);
+
+	int (*map_service_to_pipe)(struct ath10k *ar, u16 service_id,
+				   u8 *ul_pipe, u8 *dl_pipe,
+				   int *ul_is_polled, int *dl_is_polled);
+
+	void (*get_default_pipe)(struct ath10k *ar, u8 *ul_pipe, u8 *dl_pipe);
+
+	/*
+	 * Check if prior sends have completed.
+	 *
+	 * Check whether the pipe in question has any completed
+	 * sends that have not yet been processed.
+	 * This function is only relevant for HIF pipes that are configured
+	 * to be polled rather than interrupt-driven.
+	 */
+	void (*send_complete_check)(struct ath10k *ar, u8 pipe_id, int force);
+
+	void (*init)(struct ath10k *ar,
+		     struct ath10k_hif_cb *callbacks);
+
+	u16 (*get_free_queue_number)(struct ath10k *ar, u8 pipe_id);
+};
+
+
+static inline int ath10k_hif_send_head(struct ath10k *ar, u8 pipe_id,
+				       unsigned int transfer_id,
+				       unsigned int nbytes,
+				       struct sk_buff *buf)
+{
+	return ar->hif.ops->send_head(ar, pipe_id, transfer_id, nbytes, buf);
+}
+
+static inline int ath10k_hif_exchange_bmi_msg(struct ath10k *ar,
+					      void *request, u32 request_len,
+					      void *response, u32 *response_len)
+{
+	return ar->hif.ops->exchange_bmi_msg(ar, request, request_len,
+					     response, response_len);
+}
+
+static inline int ath10k_hif_start(struct ath10k *ar)
+{
+	return ar->hif.ops->start(ar);
+}
+
+static inline void ath10k_hif_stop(struct ath10k *ar)
+{
+	return ar->hif.ops->stop(ar);
+}
+
+static inline int ath10k_hif_map_service_to_pipe(struct ath10k *ar,
+						 u16 service_id,
+						 u8 *ul_pipe, u8 *dl_pipe,
+						 int *ul_is_polled,
+						 int *dl_is_polled)
+{
+	return ar->hif.ops->map_service_to_pipe(ar, service_id,
+						ul_pipe, dl_pipe,
+						ul_is_polled, dl_is_polled);
+}
+
+static inline void ath10k_hif_get_default_pipe(struct ath10k *ar,
+					       u8 *ul_pipe, u8 *dl_pipe)
+{
+	ar->hif.ops->get_default_pipe(ar, ul_pipe, dl_pipe);
+}
+
+static inline void ath10k_hif_send_complete_check(struct ath10k *ar,
+						  u8 pipe_id, int force)
+{
+	ar->hif.ops->send_complete_check(ar, pipe_id, force);
+}
+
+static inline void ath10k_hif_init(struct ath10k *ar,
+				   struct ath10k_hif_cb *callbacks)
+{
+	ar->hif.ops->init(ar, callbacks);
+}
+
+static inline u16 ath10k_hif_get_free_queue_number(struct ath10k *ar,
+						   u8 pipe_id)
+{
+	return ar->hif.ops->get_free_queue_number(ar, pipe_id);
+}
+
+#endif /* _HIF_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/htc.c b/drivers/net/wireless/ath/ath10k/htc.c
new file mode 100644
index 0000000..74363c9
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/htc.c
@@ -0,0 +1,1000 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "core.h"
+#include "hif.h"
+#include "debug.h"
+
+/********/
+/* Send */
+/********/
+
+static inline void ath10k_htc_send_complete_check(struct ath10k_htc_ep *ep,
+						  int force)
+{
+	/*
+	 * Check whether HIF has any prior sends that have finished,
+	 * have not had the post-processing done.
+	 */
+	ath10k_hif_send_complete_check(ep->htc->ar, ep->ul_pipe_id, force);
+}
+
+static void ath10k_htc_control_tx_complete(struct ath10k *ar,
+					   struct sk_buff *skb)
+{
+	kfree_skb(skb);
+}
+
+static struct sk_buff *ath10k_htc_build_tx_ctrl_skb(void *ar)
+{
+	struct sk_buff *skb;
+	struct ath10k_skb_cb *skb_cb;
+
+	skb = dev_alloc_skb(ATH10K_HTC_CONTROL_BUFFER_SIZE);
+	if (!skb) {
+		ath10k_warn("Unable to allocate ctrl skb\n");
+		return NULL;
+	}
+
+	skb_reserve(skb, 20); /* FIXME: why 20 bytes? */
+	WARN_ONCE((unsigned long)skb->data & 3, "unaligned skb");
+
+	skb_cb = ATH10K_SKB_CB(skb);
+	memset(skb_cb, 0, sizeof(*skb_cb));
+
+	ath10k_dbg(ATH10K_DBG_HTC, "%s: skb %p\n", __func__, skb);
+	return skb;
+}
+
+static inline void ath10k_htc_restore_tx_skb(struct ath10k_htc *htc,
+					     struct sk_buff *skb)
+{
+	ath10k_skb_unmap(htc->ar->dev, skb);
+	skb_pull(skb, sizeof(struct ath10k_htc_hdr));
+}
+
+static void ath10k_htc_notify_tx_completion(struct ath10k_htc_ep *ep,
+					    struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_HTC, "%s: ep %d skb %p\n", __func__,
+		   ep->eid, skb);
+
+	ath10k_htc_restore_tx_skb(ep->htc, skb);
+
+	if (!ep->ep_ops.ep_tx_complete) {
+		ath10k_warn("no tx handler for eid %d\n", ep->eid);
+		dev_kfree_skb_any(skb);
+		return;
+	}
+
+	ep->ep_ops.ep_tx_complete(ep->htc->ar, skb);
+}
+
+/* assumes tx_lock is held */
+static bool ath10k_htc_ep_need_credit_update(struct ath10k_htc_ep *ep)
+{
+	if (!ep->tx_credit_flow_enabled)
+		return false;
+	if (ep->tx_credits >= ep->tx_credits_per_max_message)
+		return false;
+
+	ath10k_dbg(ATH10K_DBG_HTC, "HTC: endpoint %d needs credit update\n",
+		   ep->eid);
+	return true;
+}
+
+static void ath10k_htc_prepare_tx_skb(struct ath10k_htc_ep *ep,
+				      struct sk_buff *skb)
+{
+	struct ath10k_htc_hdr *hdr;
+
+	hdr = (struct ath10k_htc_hdr *)skb->data;
+	memset(hdr, 0, sizeof(*hdr));
+
+	hdr->eid = ep->eid;
+	hdr->len = __cpu_to_le16(skb->len - sizeof(*hdr));
+
+	spin_lock_bh(&ep->htc->tx_lock);
+	hdr->seq_no = ep->seq_no++;
+
+	if (ath10k_htc_ep_need_credit_update(ep))
+		hdr->flags |= ATH10K_HTC_FLAG_NEED_CREDIT_UPDATE;
+
+	spin_unlock_bh(&ep->htc->tx_lock);
+}
+
+static int ath10k_htc_issue_skb(struct ath10k_htc *htc,
+				struct ath10k_htc_ep *ep,
+				struct sk_buff *skb,
+				u8 credits)
+{
+	struct ath10k_skb_cb *skb_cb = ATH10K_SKB_CB(skb);
+	int ret;
+
+	ath10k_dbg(ATH10K_DBG_HTC, "%s: ep %d skb %p\n", __func__,
+		   ep->eid, skb);
+
+	ath10k_htc_prepare_tx_skb(ep, skb);
+
+	ret = ath10k_skb_map(htc->ar->dev, skb);
+	if (ret)
+		goto err;
+
+	ret = ath10k_hif_send_head(htc->ar,
+				   ep->ul_pipe_id,
+				   ep->eid,
+				   skb->len,
+				   skb);
+	if (unlikely(ret))
+		goto err;
+
+	return 0;
+err:
+	ath10k_warn("HTC issue failed: %d\n", ret);
+
+	spin_lock_bh(&htc->tx_lock);
+	ep->tx_credits += credits;
+	spin_unlock_bh(&htc->tx_lock);
+
+	/* this is the simplest way to handle out-of-resources for non-credit
+	 * based endpoints. credit based endpoints can still get -ENOSR, but
+	 * this is highly unlikely as credit reservation should prevent that */
+	if (ret == -ENOSR) {
+		spin_lock_bh(&htc->tx_lock);
+		__skb_queue_head(&ep->tx_queue, skb);
+		spin_unlock_bh(&htc->tx_lock);
+
+		return ret;
+	}
+
+	skb_cb->is_aborted = true;
+	ath10k_htc_notify_tx_completion(ep, skb);
+
+	return ret;
+}
+
+static struct sk_buff *ath10k_htc_get_skb_credit_based(struct ath10k_htc *htc,
+						       struct ath10k_htc_ep *ep,
+						       u8 *credits)
+{
+	struct sk_buff *skb;
+	struct ath10k_skb_cb *skb_cb;
+	int credits_required;
+	int remainder;
+	unsigned int transfer_len;
+
+	lockdep_assert_held(&htc->tx_lock);
+
+	skb = __skb_dequeue(&ep->tx_queue);
+	if (!skb)
+		return NULL;
+
+	skb_cb = ATH10K_SKB_CB(skb);
+	transfer_len = skb->len;
+
+	if (likely(transfer_len <= htc->target_credit_size)) {
+		credits_required = 1;
+	} else {
+		/* figure out how many credits this message requires */
+		credits_required = transfer_len / htc->target_credit_size;
+		remainder = transfer_len % htc->target_credit_size;
+
+		if (remainder)
+			credits_required++;
+	}
+
+	ath10k_dbg(ATH10K_DBG_HTC, "Credits required %d got %d\n",
+		   credits_required, ep->tx_credits);
+
+	if (ep->tx_credits < credits_required) {
+		__skb_queue_head(&ep->tx_queue, skb);
+		return NULL;
+	}
+
+	ep->tx_credits -= credits_required;
+	*credits = credits_required;
+	return skb;
+}
+
+static void ath10k_htc_send_work(struct work_struct *work)
+{
+	struct ath10k_htc_ep *ep = container_of(work,
+					struct ath10k_htc_ep, send_work);
+	struct ath10k_htc *htc = ep->htc;
+	struct sk_buff *skb;
+	u8 credits = 0;
+	int ret;
+
+	while (true) {
+		if (ep->ul_is_polled)
+			ath10k_htc_send_complete_check(ep, 0);
+
+		spin_lock_bh(&htc->tx_lock);
+		if (ep->tx_credit_flow_enabled)
+			skb = ath10k_htc_get_skb_credit_based(htc, ep,
+							      &credits);
+		else
+			skb = __skb_dequeue(&ep->tx_queue);
+		spin_unlock_bh(&htc->tx_lock);
+
+		if (!skb)
+			break;
+
+		ret = ath10k_htc_issue_skb(htc, ep, skb, credits);
+		if (ret == -ENOSR)
+			break;
+	}
+}
+
+int ath10k_htc_send(struct ath10k_htc *htc,
+		    enum ath10k_htc_ep_id eid,
+		    struct sk_buff *skb)
+{
+	struct ath10k_htc_ep *ep = &htc->endpoint[eid];
+
+	if (eid >= ATH10K_HTC_EP_COUNT) {
+		ath10k_warn("Invalid endpoint id: %d\n", eid);
+		return -ENOENT;
+	}
+
+	skb_push(skb, sizeof(struct ath10k_htc_hdr));
+
+	spin_lock_bh(&htc->tx_lock);
+	__skb_queue_tail(&ep->tx_queue, skb);
+	spin_unlock_bh(&htc->tx_lock);
+
+	queue_work(htc->ar->workqueue, &ep->send_work);
+	return 0;
+}
+
+static int ath10k_htc_tx_completion_handler(struct ath10k *ar,
+					    struct sk_buff *skb,
+					    unsigned int eid)
+{
+	struct ath10k_htc *htc = ar->htc;
+	struct ath10k_htc_ep *ep = &htc->endpoint[eid];
+	bool stopping;
+
+	ath10k_htc_notify_tx_completion(ep, skb);
+	/* the skb now belongs to the completion handler */
+
+	spin_lock_bh(&htc->tx_lock);
+	stopping = htc->stopping;
+	spin_unlock_bh(&htc->tx_lock);
+
+	if (!ep->tx_credit_flow_enabled && !stopping)
+		/*
+		 * note: when using TX credit flow, the re-checking of
+		 * queues happens when credits flow back from the target.
+		 * in the non-TX credit case, we recheck after the packet
+		 * completes
+		 */
+		queue_work(ar->workqueue, &ep->send_work);
+
+	return 0;
+}
+
+/* flush endpoint TX queue */
+static void ath10k_htc_flush_endpoint_tx(struct ath10k_htc *htc,
+					 struct ath10k_htc_ep *ep)
+{
+	struct sk_buff *skb;
+	struct ath10k_skb_cb *skb_cb;
+
+	spin_lock_bh(&htc->tx_lock);
+	for (;;) {
+		skb = __skb_dequeue(&ep->tx_queue);
+		if (!skb)
+			break;
+
+		skb_cb = ATH10K_SKB_CB(skb);
+		skb_cb->is_aborted = true;
+		ath10k_htc_notify_tx_completion(ep, skb);
+	}
+	spin_unlock_bh(&htc->tx_lock);
+
+	cancel_work_sync(&ep->send_work);
+}
+
+/***********/
+/* Receive */
+/***********/
+
+static void
+ath10k_htc_process_credit_report(struct ath10k_htc *htc,
+				 const struct ath10k_htc_credit_report *report,
+				 int len,
+				 enum ath10k_htc_ep_id eid)
+{
+	struct ath10k_htc_ep *ep;
+	int i, n_reports;
+
+	if (len % sizeof(*report))
+		ath10k_warn("Uneven credit report len %d", len);
+
+	n_reports = len / sizeof(*report);
+
+	spin_lock_bh(&htc->tx_lock);
+	for (i = 0; i < n_reports; i++, report++) {
+		if (report->eid >= ATH10K_HTC_EP_COUNT)
+			break;
+
+		ath10k_dbg(ATH10K_DBG_HTC, "ep %d got %d credits\n",
+			   report->eid, report->credits);
+
+		ep = &htc->endpoint[report->eid];
+		ep->tx_credits += report->credits;
+
+		if (ep->tx_credits && !skb_queue_empty(&ep->tx_queue))
+			queue_work(htc->ar->workqueue, &ep->send_work);
+	}
+	spin_unlock_bh(&htc->tx_lock);
+}
+
+static int ath10k_htc_process_trailer(struct ath10k_htc *htc,
+				      u8 *buffer,
+				      int length,
+				      enum ath10k_htc_ep_id src_eid)
+{
+	int status = 0;
+	struct ath10k_htc_record *record;
+	u8 *orig_buffer;
+	int orig_length;
+	size_t len;
+
+	orig_buffer = buffer;
+	orig_length = length;
+
+	while (length > 0) {
+		record = (struct ath10k_htc_record *)buffer;
+
+		if (length < sizeof(record->hdr)) {
+			status = -EINVAL;
+			break;
+		}
+
+		if (record->hdr.len > length) {
+			/* no room left in buffer for record */
+			ath10k_warn("Invalid record length: %d\n",
+				    record->hdr.len);
+			status = -EINVAL;
+			break;
+		}
+
+		switch (record->hdr.id) {
+		case ATH10K_HTC_RECORD_CREDITS:
+			len = sizeof(struct ath10k_htc_credit_report);
+			if (record->hdr.len < len) {
+				ath10k_warn("Credit report too long\n");
+				status = -EINVAL;
+				break;
+			}
+			ath10k_htc_process_credit_report(htc,
+							 record->credit_report,
+							 record->hdr.len,
+							 src_eid);
+			break;
+		default:
+			ath10k_warn("Unhandled record: id:%d length:%d\n",
+				    record->hdr.id, record->hdr.len);
+			break;
+		}
+
+		if (status)
+			break;
+
+		/* multiple records may be present in a trailer */
+		buffer += sizeof(record->hdr) + record->hdr.len;
+		length -= sizeof(record->hdr) + record->hdr.len;
+	}
+
+	if (status)
+		ath10k_dbg_dump(ATH10K_DBG_HTC, "htc rx bad trailer", "",
+				orig_buffer, orig_length);
+
+	return status;
+}
+
+static int ath10k_htc_rx_completion_handler(struct ath10k *ar,
+					    struct sk_buff *skb,
+					    u8 pipe_id)
+{
+	int status = 0;
+	struct ath10k_htc *htc = ar->htc;
+	struct ath10k_htc_hdr *hdr;
+	struct ath10k_htc_ep *ep;
+	u16 payload_len;
+	u32 trailer_len = 0;
+	size_t min_len;
+	u8 eid;
+	bool trailer_present;
+
+	hdr = (struct ath10k_htc_hdr *)skb->data;
+	skb_pull(skb, sizeof(*hdr));
+
+	eid = hdr->eid;
+
+	if (eid >= ATH10K_HTC_EP_COUNT) {
+		ath10k_warn("HTC Rx: invalid eid %d\n", eid);
+		ath10k_dbg_dump(ATH10K_DBG_HTC, "htc bad header", "",
+				hdr, sizeof(*hdr));
+		status = -EINVAL;
+		goto out;
+	}
+
+	ep = &htc->endpoint[eid];
+
+	/*
+	 * If this endpoint that received a message from the target has
+	 * a to-target HIF pipe whose send completions are polled rather
+	 * than interrupt-driven, this is a good point to ask HIF to check
+	 * whether it has any completed sends to handle.
+	 */
+	if (ep->ul_is_polled)
+		ath10k_htc_send_complete_check(ep, 1);
+
+	payload_len = __le16_to_cpu(hdr->len);
+
+	if (payload_len + sizeof(*hdr) > ATH10K_HTC_MAX_LEN) {
+		ath10k_warn("HTC rx frame too long, len: %zu\n",
+			    payload_len + sizeof(*hdr));
+		ath10k_dbg_dump(ATH10K_DBG_HTC, "htc bad rx pkt len", "",
+				hdr, sizeof(*hdr));
+		status = -EINVAL;
+		goto out;
+	}
+
+	if (skb->len < payload_len) {
+		ath10k_dbg(ATH10K_DBG_HTC,
+			   "HTC Rx: insufficient length, got %d, expected %d\n",
+			   skb->len, payload_len);
+		ath10k_dbg_dump(ATH10K_DBG_HTC, "htc bad rx pkt len",
+				"", hdr, sizeof(*hdr));
+		status = -EINVAL;
+		goto out;
+	}
+
+	/* get flags to check for trailer */
+	trailer_present = hdr->flags & ATH10K_HTC_FLAG_TRAILER_PRESENT;
+	if (trailer_present) {
+		u8 *trailer;
+
+		trailer_len = hdr->trailer_len;
+		min_len = sizeof(struct ath10k_ath10k_htc_record_hdr);
+
+		if ((trailer_len < min_len) ||
+		    (trailer_len > payload_len)) {
+			ath10k_warn("Invalid trailer length: %d\n",
+				    trailer_len);
+			status = -EPROTO;
+			goto out;
+		}
+
+		trailer = (u8 *)hdr;
+		trailer += sizeof(*hdr);
+		trailer += payload_len;
+		trailer -= trailer_len;
+		status = ath10k_htc_process_trailer(htc, trailer,
+						    trailer_len, hdr->eid);
+		if (status)
+			goto out;
+
+		skb_trim(skb, skb->len - trailer_len);
+	}
+
+	if (((int)payload_len - (int)trailer_len) <= 0)
+		/* zero length packet with trailer data, just drop these */
+		goto out;
+
+	if (eid == ATH10K_HTC_EP_0) {
+		struct ath10k_htc_msg *msg = (struct ath10k_htc_msg *)skb->data;
+
+		switch (__le16_to_cpu(msg->hdr.message_id)) {
+		default:
+			/* handle HTC control message */
+			if (completion_done(&htc->ctl_resp)) {
+				/*
+				 * this is a fatal error, target should not be
+				 * sending unsolicited messages on the ep 0
+				 */
+				ath10k_warn("HTC rx ctrl still processing\n");
+				status = -EINVAL;
+				complete(&htc->ctl_resp);
+				goto out;
+			}
+
+			htc->control_resp_len =
+				min_t(int, skb->len,
+				      ATH10K_HTC_MAX_CTRL_MSG_LEN);
+
+			memcpy(htc->control_resp_buffer, skb->data,
+			       htc->control_resp_len);
+
+			complete(&htc->ctl_resp);
+			break;
+		case ATH10K_HTC_MSG_SEND_SUSPEND_COMPLETE:
+			htc->htc_ops.target_send_suspend_complete(ar);
+		}
+		goto out;
+	}
+
+	ath10k_dbg(ATH10K_DBG_HTC, "htc rx completion ep %d skb %p\n",
+		   eid, skb);
+	ep->ep_ops.ep_rx_complete(ar, skb);
+
+	/* skb is now owned by the rx completion handler */
+	skb = NULL;
+out:
+	kfree_skb(skb);
+
+	return status;
+}
+
+static void ath10k_htc_control_rx_complete(struct ath10k *ar,
+					   struct sk_buff *skb)
+{
+	/* This is unexpected. FW is not supposed to send regular rx on this
+	 * endpoint. */
+	ath10k_warn("unexpected htc rx\n");
+	kfree_skb(skb);
+}
+
+/***************/
+/* Init/Deinit */
+/***************/
+
+static const char *htc_service_name(enum ath10k_htc_svc_id id)
+{
+	switch (id) {
+	case ATH10K_HTC_SVC_ID_RESERVED:
+		return "Reserved";
+	case ATH10K_HTC_SVC_ID_RSVD_CTRL:
+		return "Control";
+	case ATH10K_HTC_SVC_ID_WMI_CONTROL:
+		return "WMI";
+	case ATH10K_HTC_SVC_ID_WMI_DATA_BE:
+		return "DATA BE";
+	case ATH10K_HTC_SVC_ID_WMI_DATA_BK:
+		return "DATA BK";
+	case ATH10K_HTC_SVC_ID_WMI_DATA_VI:
+		return "DATA VI";
+	case ATH10K_HTC_SVC_ID_WMI_DATA_VO:
+		return "DATA VO";
+	case ATH10K_HTC_SVC_ID_NMI_CONTROL:
+		return "NMI Control";
+	case ATH10K_HTC_SVC_ID_NMI_DATA:
+		return "NMI Data";
+	case ATH10K_HTC_SVC_ID_HTT_DATA_MSG:
+		return "HTT Data";
+	case ATH10K_HTC_SVC_ID_TEST_RAW_STREAMS:
+		return "RAW";
+	}
+
+	return "Unknown";
+}
+
+static void ath10k_htc_reset_endpoint_states(struct ath10k_htc *htc)
+{
+	struct ath10k_htc_ep *ep;
+	int i;
+
+	for (i = ATH10K_HTC_EP_0; i < ATH10K_HTC_EP_COUNT; i++) {
+		ep = &htc->endpoint[i];
+		ep->service_id = ATH10K_HTC_SVC_ID_UNUSED;
+		ep->max_ep_message_len = 0;
+		ep->max_tx_queue_depth = 0;
+		ep->eid = i;
+		skb_queue_head_init(&ep->tx_queue);
+		ep->htc = htc;
+		ep->tx_credit_flow_enabled = true;
+		INIT_WORK(&ep->send_work, ath10k_htc_send_work);
+	}
+}
+
+static void ath10k_htc_setup_target_buffer_assignments(struct ath10k_htc *htc)
+{
+	struct ath10k_htc_svc_tx_credits *entry;
+
+	entry = &htc->service_tx_alloc[0];
+
+	/*
+	 * for PCIE allocate all credists/HTC buffers to WMI.
+	 * no buffers are used/required for data. data always
+	 * remains on host.
+	 */
+	entry++;
+	entry->service_id = ATH10K_HTC_SVC_ID_WMI_CONTROL;
+	entry->credit_allocation = htc->total_transmit_credits;
+}
+
+static u8 ath10k_htc_get_credit_allocation(struct ath10k_htc *htc,
+					   u16 service_id)
+{
+	u8 allocation = 0;
+	int i;
+
+	for (i = 0; i < ATH10K_HTC_EP_COUNT; i++) {
+		if (htc->service_tx_alloc[i].service_id == service_id)
+			allocation =
+			    htc->service_tx_alloc[i].credit_allocation;
+	}
+
+	return allocation;
+}
+
+int ath10k_htc_wait_target(struct ath10k_htc *htc)
+{
+	int status = 0;
+	struct ath10k_htc_svc_conn_req conn_req;
+	struct ath10k_htc_svc_conn_resp conn_resp;
+	struct ath10k_htc_msg *msg;
+	u16 message_id;
+	u16 credit_count;
+	u16 credit_size;
+
+	INIT_COMPLETION(htc->ctl_resp);
+
+	status = ath10k_hif_start(htc->ar);
+	if (status) {
+		ath10k_err("could not start HIF (%d)\n", status);
+		goto err_start;
+	}
+
+	status = wait_for_completion_timeout(&htc->ctl_resp,
+					     ATH10K_HTC_WAIT_TIMEOUT_HZ);
+	if (status <= 0) {
+		if (status == 0)
+			status = -ETIMEDOUT;
+
+		ath10k_err("ctl_resp never came in (%d)\n", status);
+		goto err_target;
+	}
+
+	if (htc->control_resp_len < sizeof(msg->hdr) + sizeof(msg->ready)) {
+		ath10k_err("Invalid HTC ready msg len:%d\n",
+			   htc->control_resp_len);
+
+		status = -ECOMM;
+		goto err_target;
+	}
+
+	msg = (struct ath10k_htc_msg *)htc->control_resp_buffer;
+	message_id   = __le16_to_cpu(msg->hdr.message_id);
+	credit_count = __le16_to_cpu(msg->ready.credit_count);
+	credit_size  = __le16_to_cpu(msg->ready.credit_size);
+
+	if (message_id != ATH10K_HTC_MSG_READY_ID) {
+		ath10k_err("Invalid HTC ready msg: 0x%x\n", message_id);
+		status = -ECOMM;
+		goto err_target;
+	}
+
+	htc->total_transmit_credits = credit_count;
+	htc->target_credit_size = credit_size;
+
+	ath10k_dbg(ATH10K_DBG_HTC,
+		   "Target ready! transmit resources: %d size:%d\n",
+		   htc->total_transmit_credits,
+		   htc->target_credit_size);
+
+	if ((htc->total_transmit_credits == 0) ||
+	    (htc->target_credit_size == 0)) {
+		status = -ECOMM;
+		ath10k_err("Invalid credit size received\n");
+		goto err_target;
+	}
+
+	ath10k_htc_setup_target_buffer_assignments(htc);
+
+	/* setup our pseudo HTC control endpoint connection */
+	memset(&conn_req, 0, sizeof(conn_req));
+	memset(&conn_resp, 0, sizeof(conn_resp));
+	conn_req.ep_ops.ep_tx_complete = ath10k_htc_control_tx_complete;
+	conn_req.ep_ops.ep_rx_complete = ath10k_htc_control_rx_complete;
+	conn_req.max_send_queue_depth = ATH10K_NUM_CONTROL_TX_BUFFERS;
+	conn_req.service_id = ATH10K_HTC_SVC_ID_RSVD_CTRL;
+
+	/* connect fake service */
+	status = ath10k_htc_connect_service(htc, &conn_req, &conn_resp);
+	if (status) {
+		ath10k_err("could not connect to htc service (%d)\n", status);
+		goto err_target;
+	}
+
+	return 0;
+err_target:
+	ath10k_hif_stop(htc->ar);
+err_start:
+	return status;
+}
+
+int ath10k_htc_connect_service(struct ath10k_htc *htc,
+			       struct ath10k_htc_svc_conn_req *conn_req,
+			       struct ath10k_htc_svc_conn_resp *conn_resp)
+{
+	struct ath10k_htc_msg *msg;
+	struct ath10k_htc_conn_svc *req_msg;
+	struct ath10k_htc_conn_svc_response resp_msg_dummy;
+	struct ath10k_htc_conn_svc_response *resp_msg = &resp_msg_dummy;
+	enum ath10k_htc_ep_id assigned_eid = ATH10K_HTC_EP_COUNT;
+	struct ath10k_htc_ep *ep;
+	struct sk_buff *skb;
+	unsigned int max_msg_size = 0;
+	int length, status;
+	bool disable_credit_flow_ctrl = false;
+	u16 message_id, service_id, flags = 0;
+	u8 tx_alloc = 0;
+
+	/* special case for HTC pseudo control service */
+	if (conn_req->service_id == ATH10K_HTC_SVC_ID_RSVD_CTRL) {
+		disable_credit_flow_ctrl = true;
+		assigned_eid = ATH10K_HTC_EP_0;
+		max_msg_size = ATH10K_HTC_MAX_CTRL_MSG_LEN;
+		memset(&resp_msg_dummy, 0, sizeof(resp_msg_dummy));
+		goto setup;
+	}
+
+	tx_alloc = ath10k_htc_get_credit_allocation(htc,
+						    conn_req->service_id);
+	if (!tx_alloc)
+		ath10k_warn("HTC Service %s does not allocate target credits\n",
+			    htc_service_name(conn_req->service_id));
+
+	skb = ath10k_htc_build_tx_ctrl_skb(htc->ar);
+	if (!skb) {
+		ath10k_err("Failed to allocate HTC packet\n");
+		return -ENOMEM;
+	}
+
+	length = sizeof(msg->hdr) + sizeof(msg->connect_service);
+	skb_put(skb, length);
+	memset(skb->data, 0, length);
+
+	msg = (struct ath10k_htc_msg *)skb->data;
+	msg->hdr.message_id =
+		__cpu_to_le16(ATH10K_HTC_MSG_CONNECT_SERVICE_ID);
+
+	flags |= SM(tx_alloc, ATH10K_HTC_CONN_FLAGS_RECV_ALLOC);
+
+	req_msg = &msg->connect_service;
+	req_msg->flags = __cpu_to_le16(flags);
+	req_msg->service_id = __cpu_to_le16(conn_req->service_id);
+
+	/* Only enable credit flow control for WMI ctrl service */
+	if (conn_req->service_id != ATH10K_HTC_SVC_ID_WMI_CONTROL) {
+		flags |= ATH10K_HTC_CONN_FLAGS_DISABLE_CREDIT_FLOW_CTRL;
+		disable_credit_flow_ctrl = true;
+	}
+
+	INIT_COMPLETION(htc->ctl_resp);
+
+	status = ath10k_htc_send(htc, ATH10K_HTC_EP_0, skb);
+	if (status) {
+		kfree_skb(skb);
+		return status;
+	}
+
+	/* wait for response */
+	status = wait_for_completion_timeout(&htc->ctl_resp,
+					     ATH10K_HTC_CONN_SVC_TIMEOUT_HZ);
+	if (status <= 0) {
+		if (status == 0)
+			status = -ETIMEDOUT;
+		ath10k_err("Service connect timeout: %d\n", status);
+		return status;
+	}
+
+	/* we controlled the buffer creation, it's aligned */
+	msg = (struct ath10k_htc_msg *)htc->control_resp_buffer;
+	resp_msg = &msg->connect_service_response;
+	message_id = __le16_to_cpu(msg->hdr.message_id);
+	service_id = __le16_to_cpu(resp_msg->service_id);
+
+	if ((message_id != ATH10K_HTC_MSG_CONNECT_SERVICE_RESP_ID) ||
+	    (htc->control_resp_len < sizeof(msg->hdr) +
+	     sizeof(msg->connect_service_response))) {
+		ath10k_err("Invalid resp message ID 0x%x", message_id);
+		return -EPROTO;
+	}
+
+	ath10k_dbg(ATH10K_DBG_HTC,
+		   "HTC Service %s connect response: status: 0x%x, assigned ep: 0x%x\n",
+		   htc_service_name(service_id),
+		   resp_msg->status, resp_msg->eid);
+
+	conn_resp->connect_resp_code = resp_msg->status;
+
+	/* check response status */
+	if (resp_msg->status != ATH10K_HTC_CONN_SVC_STATUS_SUCCESS) {
+		ath10k_err("HTC Service %s connect request failed: 0x%x)\n",
+			   htc_service_name(service_id),
+			   resp_msg->status);
+		return -EPROTO;
+	}
+
+	assigned_eid = (enum ath10k_htc_ep_id)resp_msg->eid;
+	max_msg_size = __le16_to_cpu(resp_msg->max_msg_size);
+
+setup:
+
+	if (assigned_eid >= ATH10K_HTC_EP_COUNT)
+		return -EPROTO;
+
+	if (max_msg_size == 0)
+		return -EPROTO;
+
+	ep = &htc->endpoint[assigned_eid];
+	ep->eid = assigned_eid;
+
+	if (ep->service_id != ATH10K_HTC_SVC_ID_UNUSED)
+		return -EPROTO;
+
+	/* return assigned endpoint to caller */
+	conn_resp->eid = assigned_eid;
+	conn_resp->max_msg_len = __le16_to_cpu(resp_msg->max_msg_size);
+
+	/* setup the endpoint */
+	ep->service_id = conn_req->service_id;
+	ep->max_tx_queue_depth = conn_req->max_send_queue_depth;
+	ep->max_ep_message_len = __le16_to_cpu(resp_msg->max_msg_size);
+	ep->tx_credits = tx_alloc;
+	ep->tx_credit_size = htc->target_credit_size;
+	ep->tx_credits_per_max_message = ep->max_ep_message_len /
+					 htc->target_credit_size;
+
+	if (ep->max_ep_message_len % htc->target_credit_size)
+		ep->tx_credits_per_max_message++;
+
+	/* copy all the callbacks */
+	ep->ep_ops = conn_req->ep_ops;
+
+	status = ath10k_hif_map_service_to_pipe(htc->ar,
+						ep->service_id,
+						&ep->ul_pipe_id,
+						&ep->dl_pipe_id,
+						&ep->ul_is_polled,
+						&ep->dl_is_polled);
+	if (status)
+		return status;
+
+	ath10k_dbg(ATH10K_DBG_HTC,
+		   "HTC service: %s UL pipe: %d DL pipe: %d eid: %d ready\n",
+		   htc_service_name(ep->service_id), ep->ul_pipe_id,
+		   ep->dl_pipe_id, ep->eid);
+
+	ath10k_dbg(ATH10K_DBG_HTC,
+		   "EP %d UL polled: %d, DL polled: %d\n",
+		   ep->eid, ep->ul_is_polled, ep->dl_is_polled);
+
+	if (disable_credit_flow_ctrl && ep->tx_credit_flow_enabled) {
+		ep->tx_credit_flow_enabled = false;
+		ath10k_dbg(ATH10K_DBG_HTC,
+			   "HTC service: %s eid: %d TX flow control disabled\n",
+			   htc_service_name(ep->service_id), assigned_eid);
+	}
+
+	return status;
+}
+
+struct sk_buff *ath10k_htc_alloc_skb(int size)
+{
+	struct sk_buff *skb;
+
+	skb = dev_alloc_skb(size + sizeof(struct ath10k_htc_hdr));
+	if (!skb) {
+		ath10k_warn("could not allocate HTC tx skb\n");
+		return NULL;
+	}
+
+	skb_reserve(skb, sizeof(struct ath10k_htc_hdr));
+
+	/* FW/HTC requires 4-byte aligned streams */
+	if (!IS_ALIGNED((unsigned long)skb->data, 4))
+		ath10k_warn("Unaligned HTC tx skb\n");
+
+	return skb;
+}
+
+int ath10k_htc_start(struct ath10k_htc *htc)
+{
+	struct sk_buff *skb;
+	int status = 0;
+	struct ath10k_htc_msg *msg;
+
+	skb = ath10k_htc_build_tx_ctrl_skb(htc->ar);
+	if (!skb)
+		return -ENOMEM;
+
+	skb_put(skb, sizeof(msg->hdr) + sizeof(msg->setup_complete_ext));
+	memset(skb->data, 0, skb->len);
+
+	msg = (struct ath10k_htc_msg *)skb->data;
+	msg->hdr.message_id =
+		__cpu_to_le16(ATH10K_HTC_MSG_SETUP_COMPLETE_EX_ID);
+
+	ath10k_dbg(ATH10K_DBG_HTC, "HTC is using TX credit flow control\n");
+
+	status = ath10k_htc_send(htc, ATH10K_HTC_EP_0, skb);
+	if (status) {
+		kfree_skb(skb);
+		return status;
+	}
+
+	return 0;
+}
+
+/*
+ * stop HTC communications, i.e. stop interrupt reception, and flush all
+ * queued buffers
+ */
+void ath10k_htc_stop(struct ath10k_htc *htc)
+{
+	int i;
+	struct ath10k_htc_ep *ep;
+
+	spin_lock_bh(&htc->tx_lock);
+	htc->stopping = true;
+	spin_unlock_bh(&htc->tx_lock);
+
+	for (i = ATH10K_HTC_EP_0; i < ATH10K_HTC_EP_COUNT; i++) {
+		ep = &htc->endpoint[i];
+		ath10k_htc_flush_endpoint_tx(htc, ep);
+	}
+
+	ath10k_hif_stop(htc->ar);
+	ath10k_htc_reset_endpoint_states(htc);
+}
+
+/* registered target arrival callback from the HIF layer */
+struct ath10k_htc *ath10k_htc_create(struct ath10k *ar,
+				     struct ath10k_htc_ops *htc_ops)
+{
+	struct ath10k_hif_cb htc_callbacks;
+	struct ath10k_htc_ep *ep = NULL;
+	struct ath10k_htc *htc = NULL;
+
+	/* FIXME: use struct ath10k instead */
+	htc = kzalloc(sizeof(struct ath10k_htc), GFP_KERNEL);
+	if (!htc)
+		return ERR_PTR(-ENOMEM);
+
+	spin_lock_init(&htc->tx_lock);
+
+	memcpy(&htc->htc_ops, htc_ops, sizeof(struct ath10k_htc_ops));
+
+	ath10k_htc_reset_endpoint_states(htc);
+
+	/* setup HIF layer callbacks */
+	htc_callbacks.rx_completion = ath10k_htc_rx_completion_handler;
+	htc_callbacks.tx_completion = ath10k_htc_tx_completion_handler;
+	htc->ar = ar;
+
+	/* Get HIF default pipe for HTC message exchange */
+	ep = &htc->endpoint[ATH10K_HTC_EP_0];
+
+	ath10k_hif_init(ar, &htc_callbacks);
+	ath10k_hif_get_default_pipe(ar, &ep->ul_pipe_id, &ep->dl_pipe_id);
+
+	init_completion(&htc->ctl_resp);
+
+	return htc;
+}
+
+void ath10k_htc_destroy(struct ath10k_htc *htc)
+{
+	kfree(htc);
+}
diff --git a/drivers/net/wireless/ath/ath10k/htc.h b/drivers/net/wireless/ath/ath10k/htc.h
new file mode 100644
index 0000000..fa45844
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/htc.h
@@ -0,0 +1,368 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _HTC_H_
+#define _HTC_H_
+
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/bug.h>
+#include <linux/skbuff.h>
+#include <linux/semaphore.h>
+#include <linux/timer.h>
+
+struct ath10k;
+
+/****************/
+/* HTC protocol */
+/****************/
+
+/*
+ * HTC - host-target control protocol
+ *
+ * tx packets are generally <htc_hdr><payload>
+ * rx packets are more complex: <htc_hdr><payload><trailer>
+ *
+ * The payload + trailer length is stored in len.
+ * To get payload-only length one needs to payload - trailer_len.
+ *
+ * Trailer contains (possibly) multiple <htc_record>.
+ * Each record is a id-len-value.
+ *
+ * HTC header flags, control_byte0, control_byte1
+ * have different meaning depending whether its tx
+ * or rx.
+ *
+ * Alignment: htc_hdr, payload and trailer are
+ * 4-byte aligned.
+ */
+
+enum ath10k_htc_tx_flags {
+	ATH10K_HTC_FLAG_NEED_CREDIT_UPDATE = 0x01,
+	ATH10K_HTC_FLAG_SEND_BUNDLE        = 0x02
+};
+
+enum ath10k_htc_rx_flags {
+	ATH10K_HTC_FLAG_TRAILER_PRESENT = 0x02,
+	ATH10K_HTC_FLAG_BUNDLE_MASK     = 0xF0
+};
+
+struct ath10k_htc_hdr {
+	u8 eid; /* @enum ath10k_htc_ep_id */
+	u8 flags; /* @enum ath10k_htc_tx_flags, ath10k_htc_rx_flags */
+	__le16 len;
+	union {
+		u8 trailer_len; /* for rx */
+		u8 control_byte0;
+	} __packed;
+	union {
+		u8 seq_no; /* for tx */
+		u8 control_byte1;
+	} __packed;
+	u8 pad0;
+	u8 pad1;
+} __packed __aligned(4);
+
+enum ath10k_ath10k_htc_msg_id {
+	ATH10K_HTC_MSG_READY_ID                = 1,
+	ATH10K_HTC_MSG_CONNECT_SERVICE_ID      = 2,
+	ATH10K_HTC_MSG_CONNECT_SERVICE_RESP_ID = 3,
+	ATH10K_HTC_MSG_SETUP_COMPLETE_ID       = 4,
+	ATH10K_HTC_MSG_SETUP_COMPLETE_EX_ID    = 5,
+	ATH10K_HTC_MSG_SEND_SUSPEND_COMPLETE   = 6
+};
+
+enum ath10k_htc_version {
+	ATH10K_HTC_VERSION_2P0 = 0x00, /* 2.0 */
+	ATH10K_HTC_VERSION_2P1 = 0x01, /* 2.1 */
+};
+
+enum ath10k_htc_conn_flags {
+	ATH10K_HTC_CONN_FLAGS_THRESHOLD_LEVEL_ONE_FOURTH    = 0x0,
+	ATH10K_HTC_CONN_FLAGS_THRESHOLD_LEVEL_ONE_HALF      = 0x1,
+	ATH10K_HTC_CONN_FLAGS_THRESHOLD_LEVEL_THREE_FOURTHS = 0x2,
+	ATH10K_HTC_CONN_FLAGS_THRESHOLD_LEVEL_UNITY         = 0x3,
+#define ATH10K_HTC_CONN_FLAGS_THRESHOLD_LEVEL_MASK 0x3
+	ATH10K_HTC_CONN_FLAGS_REDUCE_CREDIT_DRIBBLE    = 1 << 2,
+	ATH10K_HTC_CONN_FLAGS_DISABLE_CREDIT_FLOW_CTRL = 1 << 3
+#define ATH10K_HTC_CONN_FLAGS_RECV_ALLOC_MASK 0xFF00
+#define ATH10K_HTC_CONN_FLAGS_RECV_ALLOC_LSB  8
+};
+
+enum ath10k_htc_conn_svc_status {
+	ATH10K_HTC_CONN_SVC_STATUS_SUCCESS      = 0,
+	ATH10K_HTC_CONN_SVC_STATUS_NOT_FOUND    = 1,
+	ATH10K_HTC_CONN_SVC_STATUS_FAILED       = 2,
+	ATH10K_HTC_CONN_SVC_STATUS_NO_RESOURCES = 3,
+	ATH10K_HTC_CONN_SVC_STATUS_NO_MORE_EP   = 4
+};
+
+struct ath10k_ath10k_htc_msg_hdr {
+	__le16 message_id; /* @enum htc_message_id */
+} __packed;
+
+struct ath10k_htc_unknown {
+	u8 pad0;
+	u8 pad1;
+} __packed;
+
+struct ath10k_htc_ready {
+	__le16 credit_count;
+	__le16 credit_size;
+	u8 max_endpoints;
+	u8 pad0;
+} __packed;
+
+struct ath10k_htc_ready_extended {
+	struct ath10k_htc_ready base;
+	u8 htc_version; /* @enum ath10k_htc_version */
+	u8 max_msgs_per_htc_bundle;
+	u8 pad0;
+	u8 pad1;
+} __packed;
+
+struct ath10k_htc_conn_svc {
+	__le16 service_id;
+	__le16 flags; /* @enum ath10k_htc_conn_flags */
+	u8 pad0;
+	u8 pad1;
+} __packed;
+
+struct ath10k_htc_conn_svc_response {
+	__le16 service_id;
+	u8 status; /* @enum ath10k_htc_conn_svc_status */
+	u8 eid;
+	__le16 max_msg_size;
+} __packed;
+
+struct ath10k_htc_setup_complete_extended {
+	u8 pad0;
+	u8 pad1;
+	__le32 flags; /* @enum htc_setup_complete_flags */
+	u8 max_msgs_per_bundled_recv;
+	u8 pad2;
+	u8 pad3;
+	u8 pad4;
+} __packed;
+
+struct ath10k_htc_msg {
+	struct ath10k_ath10k_htc_msg_hdr hdr;
+	union {
+		/* host-to-target */
+		struct ath10k_htc_conn_svc connect_service;
+		struct ath10k_htc_ready ready;
+		struct ath10k_htc_ready_extended ready_ext;
+		struct ath10k_htc_unknown unknown;
+		struct ath10k_htc_setup_complete_extended setup_complete_ext;
+
+		/* target-to-host */
+		struct ath10k_htc_conn_svc_response connect_service_response;
+	};
+} __packed __aligned(4);
+
+enum ath10k_ath10k_htc_record_id {
+	ATH10K_HTC_RECORD_NULL    = 0,
+	ATH10K_HTC_RECORD_CREDITS = 1
+};
+
+struct ath10k_ath10k_htc_record_hdr {
+	u8 id; /* @enum ath10k_ath10k_htc_record_id */
+	u8 len;
+	u8 pad0;
+	u8 pad1;
+} __packed;
+
+struct ath10k_htc_credit_report {
+	u8 eid; /* @enum ath10k_htc_ep_id */
+	u8 credits;
+	u8 pad0;
+	u8 pad1;
+} __packed;
+
+struct ath10k_htc_record {
+	struct ath10k_ath10k_htc_record_hdr hdr;
+	union {
+		struct ath10k_htc_credit_report credit_report[0];
+		u8 pauload[0];
+	};
+} __packed __aligned(4);
+
+/*
+ * note: the trailer offset is dynamic depending
+ * on payload length. this is only a struct layout draft
+ */
+struct ath10k_htc_frame {
+	struct ath10k_htc_hdr hdr;
+	union {
+		struct ath10k_htc_msg msg;
+		u8 payload[0];
+	};
+	struct ath10k_htc_record trailer[0];
+} __packed __aligned(4);
+
+
+/*******************/
+/* Host-side stuff */
+/*******************/
+
+enum ath10k_htc_svc_gid {
+	ATH10K_HTC_SVC_GRP_RSVD = 0,
+	ATH10K_HTC_SVC_GRP_WMI = 1,
+	ATH10K_HTC_SVC_GRP_NMI = 2,
+	ATH10K_HTC_SVC_GRP_HTT = 3,
+
+	ATH10K_HTC_SVC_GRP_TEST = 254,
+	ATH10K_HTC_SVC_GRP_LAST = 255,
+};
+
+#define SVC(group, idx) \
+	(int)(((int)(group) << 8) | (int)(idx))
+
+enum ath10k_htc_svc_id {
+	/* NOTE: service ID of 0x0000 is reserved and should never be used */
+	ATH10K_HTC_SVC_ID_RESERVED	= 0x0000,
+	ATH10K_HTC_SVC_ID_UNUSED	= ATH10K_HTC_SVC_ID_RESERVED,
+
+	ATH10K_HTC_SVC_ID_RSVD_CTRL	= SVC(ATH10K_HTC_SVC_GRP_RSVD, 1),
+	ATH10K_HTC_SVC_ID_WMI_CONTROL	= SVC(ATH10K_HTC_SVC_GRP_WMI, 0),
+	ATH10K_HTC_SVC_ID_WMI_DATA_BE	= SVC(ATH10K_HTC_SVC_GRP_WMI, 1),
+	ATH10K_HTC_SVC_ID_WMI_DATA_BK	= SVC(ATH10K_HTC_SVC_GRP_WMI, 2),
+	ATH10K_HTC_SVC_ID_WMI_DATA_VI	= SVC(ATH10K_HTC_SVC_GRP_WMI, 3),
+	ATH10K_HTC_SVC_ID_WMI_DATA_VO	= SVC(ATH10K_HTC_SVC_GRP_WMI, 4),
+
+	ATH10K_HTC_SVC_ID_NMI_CONTROL	= SVC(ATH10K_HTC_SVC_GRP_NMI, 0),
+	ATH10K_HTC_SVC_ID_NMI_DATA	= SVC(ATH10K_HTC_SVC_GRP_NMI, 1),
+
+	ATH10K_HTC_SVC_ID_HTT_DATA_MSG	= SVC(ATH10K_HTC_SVC_GRP_HTT, 0),
+
+	/* raw stream service (i.e. flash, tcmd, calibration apps) */
+	ATH10K_HTC_SVC_ID_TEST_RAW_STREAMS = SVC(ATH10K_HTC_SVC_GRP_TEST, 0),
+};
+
+#undef SVC
+
+enum ath10k_htc_ep_id {
+	ATH10K_HTC_EP_UNUSED = -1,
+	ATH10K_HTC_EP_0 = 0,
+	ATH10K_HTC_EP_1 = 1,
+	ATH10K_HTC_EP_2,
+	ATH10K_HTC_EP_3,
+	ATH10K_HTC_EP_4,
+	ATH10K_HTC_EP_5,
+	ATH10K_HTC_EP_6,
+	ATH10K_HTC_EP_7,
+	ATH10K_HTC_EP_8,
+	ATH10K_HTC_EP_COUNT,
+};
+
+struct ath10k_htc_ops {
+	void (*target_send_suspend_complete)(struct ath10k *ar);
+};
+
+struct ath10k_htc_ep_ops {
+	void (*ep_tx_complete)(struct ath10k *, struct sk_buff *);
+	void (*ep_rx_complete)(struct ath10k *, struct sk_buff *);
+};
+
+/* service connection information */
+struct ath10k_htc_svc_conn_req {
+	u16 service_id;
+	struct ath10k_htc_ep_ops ep_ops;
+	int max_send_queue_depth;
+};
+
+/* service connection response information */
+struct ath10k_htc_svc_conn_resp {
+	u8 buffer_len;
+	u8 actual_len;
+	enum ath10k_htc_ep_id eid;
+	unsigned int max_msg_len;
+	u8 connect_resp_code;
+};
+
+#define ATH10K_NUM_CONTROL_TX_BUFFERS 2
+#define ATH10K_HTC_MAX_LEN 4096
+#define ATH10K_HTC_MAX_CTRL_MSG_LEN 256
+#define ATH10K_HTC_WAIT_TIMEOUT_HZ (1*HZ)
+#define ATH10K_HTC_CONTROL_BUFFER_SIZE (ATH10K_HTC_MAX_CTRL_MSG_LEN + \
+					sizeof(struct ath10k_htc_hdr))
+#define ATH10K_HTC_CONN_SVC_TIMEOUT_HZ (1*HZ)
+
+struct ath10k_htc_ep {
+	struct ath10k_htc *htc;
+	enum ath10k_htc_ep_id eid;
+	enum ath10k_htc_svc_id service_id;
+	struct ath10k_htc_ep_ops ep_ops;
+
+	int max_tx_queue_depth;
+	int max_ep_message_len;
+	u8 ul_pipe_id;
+	u8 dl_pipe_id;
+	int ul_is_polled; /* call HIF to get tx completions */
+	int dl_is_polled; /* call HIF to fetch rx (not implemented) */
+
+	struct sk_buff_head tx_queue;
+
+	u8 seq_no; /* for debugging */
+	int tx_credits;
+	int tx_credit_size;
+	int tx_credits_per_max_message;
+	bool tx_credit_flow_enabled;
+
+	struct work_struct send_work;
+};
+
+struct ath10k_htc_svc_tx_credits {
+	u16 service_id;
+	u8  credit_allocation;
+};
+
+struct ath10k_htc {
+	struct ath10k *ar;
+	struct ath10k_htc_ep endpoint[ATH10K_HTC_EP_COUNT];
+
+	/* protects endpoint and stopping fields */
+	spinlock_t tx_lock;
+
+	struct ath10k_htc_ops htc_ops;
+
+	u8 control_resp_buffer[ATH10K_HTC_MAX_CTRL_MSG_LEN];
+	int control_resp_len;
+
+	struct completion ctl_resp;
+
+	int total_transmit_credits;
+	struct ath10k_htc_svc_tx_credits service_tx_alloc[ATH10K_HTC_EP_COUNT];
+	int target_credit_size;
+
+	bool stopping;
+};
+
+struct ath10k_htc *ath10k_htc_create(struct ath10k *ar,
+				     struct ath10k_htc_ops *htc_ops);
+int ath10k_htc_wait_target(struct ath10k_htc *htc);
+int ath10k_htc_start(struct ath10k_htc *htc);
+int ath10k_htc_connect_service(struct ath10k_htc *htc,
+			       struct ath10k_htc_svc_conn_req  *conn_req,
+			       struct ath10k_htc_svc_conn_resp *conn_resp);
+int ath10k_htc_send(struct ath10k_htc *htc, enum ath10k_htc_ep_id eid,
+		    struct sk_buff *packet);
+void ath10k_htc_stop(struct ath10k_htc *htc);
+void ath10k_htc_destroy(struct ath10k_htc *htc);
+struct sk_buff *ath10k_htc_alloc_skb(int size);
+
+#endif
diff --git a/drivers/net/wireless/ath/ath10k/htt.c b/drivers/net/wireless/ath/ath10k/htt.c
new file mode 100644
index 0000000..185a546
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/htt.c
@@ -0,0 +1,152 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/slab.h>
+
+#include "htt.h"
+#include "core.h"
+#include "debug.h"
+
+static int ath10k_htt_htc_attach(struct ath10k_htt *htt)
+{
+	struct ath10k_htc_svc_conn_req conn_req;
+	struct ath10k_htc_svc_conn_resp conn_resp;
+	int status;
+
+	memset(&conn_req, 0, sizeof(conn_req));
+	memset(&conn_resp, 0, sizeof(conn_resp));
+
+	conn_req.ep_ops.ep_tx_complete = ath10k_htt_htc_tx_complete;
+	conn_req.ep_ops.ep_rx_complete = ath10k_htt_t2h_msg_handler;
+
+	/* connect to control service */
+	conn_req.service_id = ATH10K_HTC_SVC_ID_HTT_DATA_MSG;
+
+	status = ath10k_htc_connect_service(htt->ar->htc, &conn_req,
+					    &conn_resp);
+
+	if (status)
+		return status;
+
+	htt->eid = conn_resp.eid;
+
+	return 0;
+}
+
+struct ath10k_htt *ath10k_htt_attach(struct ath10k *ar)
+{
+	struct ath10k_htt *htt;
+	int ret;
+
+	htt = kzalloc(sizeof(*htt), GFP_KERNEL);
+	if (!htt)
+		return NULL;
+
+	htt->ar = ar;
+	htt->max_throughput_mbps = 800;
+
+	/*
+	 * Connect to HTC service.
+	 * This has to be done before calling ath10k_htt_rx_attach,
+	 * since ath10k_htt_rx_attach involves sending a rx ring configure
+	 * message to the target.
+	 */
+	if (ath10k_htt_htc_attach(htt))
+		goto err_htc_attach;
+
+	ret = ath10k_htt_tx_attach(htt);
+	if (ret) {
+		ath10k_err("could not attach htt tx (%d)\n", ret);
+		goto err_htc_attach;
+	}
+
+	if (ath10k_htt_rx_attach(htt))
+		goto err_rx_attach;
+
+	/*
+	 * Prefetch enough data to satisfy target
+	 * classification engine.
+	 * This is for LL chips. HL chips will probably
+	 * transfer all frame in the tx fragment.
+	 */
+	htt->prefetch_len =
+		36 + /* 802.11 + qos + ht */
+		4 + /* 802.1q */
+		8 + /* llc snap */
+		2; /* ip4 dscp or ip6 priority */
+
+	return htt;
+
+err_rx_attach:
+	ath10k_htt_tx_detach(htt);
+err_htc_attach:
+	kfree(htt);
+	return NULL;
+}
+
+#define HTT_TARGET_VERSION_TIMEOUT_HZ (3*HZ)
+
+static int ath10k_htt_verify_version(struct ath10k_htt *htt)
+{
+	ath10k_dbg(ATH10K_DBG_HTT,
+		   "htt target version %d.%d; host version %d.%d\n",
+		    htt->target_version_major,
+		    htt->target_version_minor,
+		    HTT_CURRENT_VERSION_MAJOR,
+		    HTT_CURRENT_VERSION_MINOR);
+
+	if (htt->target_version_major != HTT_CURRENT_VERSION_MAJOR) {
+		ath10k_err("htt major versions are incompatible!\n");
+		return -ENOTSUPP;
+	}
+
+	if (htt->target_version_minor != HTT_CURRENT_VERSION_MINOR)
+		ath10k_warn("htt minor version differ but still compatible\n");
+
+	return 0;
+}
+
+int ath10k_htt_attach_target(struct ath10k_htt *htt)
+{
+	int status;
+
+	init_completion(&htt->target_version_received);
+
+	status = ath10k_htt_h2t_ver_req_msg(htt);
+	if (status)
+		return status;
+
+	status = wait_for_completion_timeout(&htt->target_version_received,
+						HTT_TARGET_VERSION_TIMEOUT_HZ);
+	if (status <= 0) {
+		ath10k_warn("htt version request timed out\n");
+		return -ETIMEDOUT;
+	}
+
+	status = ath10k_htt_verify_version(htt);
+	if (status)
+		return status;
+
+	return ath10k_htt_send_rx_ring_cfg_ll(htt);
+}
+
+void ath10k_htt_detach(struct ath10k_htt *htt)
+{
+	ath10k_htt_rx_detach(htt);
+	ath10k_htt_tx_detach(htt);
+	kfree(htt);
+}
diff --git a/drivers/net/wireless/ath/ath10k/htt.h b/drivers/net/wireless/ath/ath10k/htt.h
new file mode 100644
index 0000000..a7a7aa0
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/htt.h
@@ -0,0 +1,1338 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _HTT_H_
+#define _HTT_H_
+
+#include <linux/bug.h>
+
+#include "core.h"
+#include "htc.h"
+#include "rx_desc.h"
+
+#define HTT_CURRENT_VERSION_MAJOR	2
+#define HTT_CURRENT_VERSION_MINOR	1
+
+enum htt_dbg_stats_type {
+	HTT_DBG_STATS_WAL_PDEV_TXRX = 1 << 0,
+	HTT_DBG_STATS_RX_REORDER    = 1 << 1,
+	HTT_DBG_STATS_RX_RATE_INFO  = 1 << 2,
+	HTT_DBG_STATS_TX_PPDU_LOG   = 1 << 3,
+	HTT_DBG_STATS_TX_RATE_INFO  = 1 << 4,
+	/* bits 5-23 currently reserved */
+
+	HTT_DBG_NUM_STATS /* keep this last */
+};
+
+enum htt_h2t_msg_type { /* host-to-target */
+	HTT_H2T_MSG_TYPE_VERSION_REQ        = 0,
+	HTT_H2T_MSG_TYPE_TX_FRM             = 1,
+	HTT_H2T_MSG_TYPE_RX_RING_CFG        = 2,
+	HTT_H2T_MSG_TYPE_STATS_REQ          = 3,
+	HTT_H2T_MSG_TYPE_SYNC               = 4,
+	HTT_H2T_MSG_TYPE_AGGR_CFG           = 5,
+	HTT_H2T_MSG_TYPE_FRAG_DESC_BANK_CFG = 6,
+	HTT_H2T_MSG_TYPE_MGMT_TX            = 7,
+
+	HTT_H2T_NUM_MSGS /* keep this last */
+};
+
+struct htt_cmd_hdr {
+	u8 msg_type;
+} __packed;
+
+struct htt_ver_req {
+	u8 pad[sizeof(u32) - sizeof(struct htt_cmd_hdr)];
+} __packed;
+
+/*
+ * HTT tx MSDU descriptor
+ *
+ * The HTT tx MSDU descriptor is created by the host HTT SW for each
+ * tx MSDU.  The HTT tx MSDU descriptor contains the information that
+ * the target firmware needs for the FW's tx processing, particularly
+ * for creating the HW msdu descriptor.
+ * The same HTT tx descriptor is used for HL and LL systems, though
+ * a few fields within the tx descriptor are used only by LL or
+ * only by HL.
+ * The HTT tx descriptor is defined in two manners: by a struct with
+ * bitfields, and by a series of [dword offset, bit mask, bit shift]
+ * definitions.
+ * The target should use the struct def, for simplicitly and clarity,
+ * but the host shall use the bit-mast + bit-shift defs, to be endian-
+ * neutral.  Specifically, the host shall use the get/set macros built
+ * around the mask + shift defs.
+ */
+struct htt_data_tx_desc_frag {
+	__le32 paddr;
+	__le32 len;
+} __packed;
+
+enum htt_data_tx_desc_flags0 {
+	HTT_DATA_TX_DESC_FLAGS0_MAC_HDR_PRESENT = 1 << 0,
+	HTT_DATA_TX_DESC_FLAGS0_NO_AGGR         = 1 << 1,
+	HTT_DATA_TX_DESC_FLAGS0_NO_ENCRYPT      = 1 << 2,
+	HTT_DATA_TX_DESC_FLAGS0_NO_CLASSIFY     = 1 << 3,
+	HTT_DATA_TX_DESC_FLAGS0_RSVD0           = 1 << 4
+#define HTT_DATA_TX_DESC_FLAGS0_PKT_TYPE_MASK 0xE0
+#define HTT_DATA_TX_DESC_FLAGS0_PKT_TYPE_LSB 5
+};
+
+enum htt_data_tx_desc_flags1 {
+#define HTT_DATA_TX_DESC_FLAGS1_VDEV_ID_BITS 6
+#define HTT_DATA_TX_DESC_FLAGS1_VDEV_ID_MASK 0x003F
+#define HTT_DATA_TX_DESC_FLAGS1_VDEV_ID_LSB  0
+#define HTT_DATA_TX_DESC_FLAGS1_EXT_TID_BITS 5
+#define HTT_DATA_TX_DESC_FLAGS1_EXT_TID_MASK 0x07C0
+#define HTT_DATA_TX_DESC_FLAGS1_EXT_TID_LSB  6
+	HTT_DATA_TX_DESC_FLAGS1_POSTPONED        = 1 << 11,
+	HTT_DATA_TX_DESC_FLAGS1_MORE_IN_BATCH    = 1 << 12,
+	HTT_DATA_TX_DESC_FLAGS1_CKSUM_L3_OFFLOAD = 1 << 13,
+	HTT_DATA_TX_DESC_FLAGS1_CKSUM_L4_OFFLOAD = 1 << 14,
+	HTT_DATA_TX_DESC_FLAGS1_RSVD1            = 1 << 15
+};
+
+enum htt_data_tx_ext_tid {
+	HTT_DATA_TX_EXT_TID_NON_QOS_MCAST_BCAST = 16,
+	HTT_DATA_TX_EXT_TID_MGMT                = 17,
+	HTT_DATA_TX_EXT_TID_INVALID             = 31
+};
+
+#define HTT_INVALID_PEERID 0xFFFF
+
+/*
+ * htt_data_tx_desc - used for data tx path
+ *
+ * Note: vdev_id irrelevant for pkt_type == raw and no_classify == 1.
+ *       ext_tid: for qos-data frames (0-15), see %HTT_DATA_TX_EXT_TID_
+ *                for special kinds of tids
+ *       postponed: only for HL hosts. indicates if this is a resend
+ *                  (HL hosts manage queues on the host )
+ *       more_in_batch: only for HL hosts. indicates if more packets are
+ *                      pending. this allows target to wait and aggregate
+ */
+struct htt_data_tx_desc {
+	u8 flags0; /* %HTT_DATA_TX_DESC_FLAGS0_ */
+	__le16 flags1; /* %HTT_DATA_TX_DESC_FLAGS1_ */
+	__le16 len;
+	__le16 id;
+	__le32 frags_paddr;
+	__le32 peerid;
+	u8 prefetch[0]; /* start of frame, for FW classification engine */
+} __packed;
+
+enum htt_rx_ring_flags {
+	HTT_RX_RING_FLAGS_MAC80211_HDR = 1 << 0,
+	HTT_RX_RING_FLAGS_MSDU_PAYLOAD = 1 << 1,
+	HTT_RX_RING_FLAGS_PPDU_START   = 1 << 2,
+	HTT_RX_RING_FLAGS_PPDU_END     = 1 << 3,
+	HTT_RX_RING_FLAGS_MPDU_START   = 1 << 4,
+	HTT_RX_RING_FLAGS_MPDU_END     = 1 << 5,
+	HTT_RX_RING_FLAGS_MSDU_START   = 1 << 6,
+	HTT_RX_RING_FLAGS_MSDU_END     = 1 << 7,
+	HTT_RX_RING_FLAGS_RX_ATTENTION = 1 << 8,
+	HTT_RX_RING_FLAGS_FRAG_INFO    = 1 << 9,
+	HTT_RX_RING_FLAGS_UNICAST_RX   = 1 << 10,
+	HTT_RX_RING_FLAGS_MULTICAST_RX = 1 << 11,
+	HTT_RX_RING_FLAGS_CTRL_RX      = 1 << 12,
+	HTT_RX_RING_FLAGS_MGMT_RX      = 1 << 13,
+	HTT_RX_RING_FLAGS_NULL_RX      = 1 << 14,
+	HTT_RX_RING_FLAGS_PHY_DATA_RX  = 1 << 15
+};
+
+struct htt_rx_ring_setup_ring {
+	__le32 fw_idx_shadow_reg_paddr;
+	__le32 rx_ring_base_paddr;
+	__le16 rx_ring_len; /* in 4-byte words */
+	__le16 rx_ring_bufsize; /* rx skb size - in bytes */
+	__le16 flags; /* %HTT_RX_RING_FLAGS_ */
+	__le16 fw_idx_init_val;
+
+	/* the following offsets are in 4-byte units */
+	__le16 mac80211_hdr_offset;
+	__le16 msdu_payload_offset;
+	__le16 ppdu_start_offset;
+	__le16 ppdu_end_offset;
+	__le16 mpdu_start_offset;
+	__le16 mpdu_end_offset;
+	__le16 msdu_start_offset;
+	__le16 msdu_end_offset;
+	__le16 rx_attention_offset;
+	__le16 frag_info_offset;
+} __packed;
+
+struct htt_rx_ring_setup_hdr {
+	u8 num_rings; /* supported values: 1, 2 */
+	__le16 rsvd0;
+} __packed;
+
+struct htt_rx_ring_setup {
+	struct htt_rx_ring_setup_hdr hdr;
+	struct htt_rx_ring_setup_ring rings[0];
+} __packed;
+
+/*
+ * htt_stats_req - request target to send specified statistics
+ *
+ * @msg_type: hardcoded %HTT_H2T_MSG_TYPE_STATS_REQ
+ * @upload_types: see %htt_dbg_stats_type. this is 24bit field actually
+ *	so make sure its little-endian.
+ * @reset_types: see %htt_dbg_stats_type. this is 24bit field actually
+ *	so make sure its little-endian.
+ * @cfg_val: stat_type specific configuration
+ * @stat_type: see %htt_dbg_stats_type
+ * @cookie_lsb: used for confirmation message from target->host
+ * @cookie_msb: ditto as %cookie
+ */
+struct htt_stats_req {
+	u8 upload_types[3];
+	u8 rsvd0;
+	u8 reset_types[3];
+	struct {
+		u8 mpdu_bytes;
+		u8 mpdu_num_msdus;
+		u8 msdu_bytes;
+	} __packed;
+	u8 stat_type;
+	__le32 cookie_lsb;
+	__le32 cookie_msb;
+} __packed;
+
+#define HTT_STATS_REQ_CFG_STAT_TYPE_INVALID 0xff
+
+/*
+ * htt_oob_sync_req - request out-of-band sync
+ *
+ * The HTT SYNC tells the target to suspend processing of subsequent
+ * HTT host-to-target messages until some other target agent locally
+ * informs the target HTT FW that the current sync counter is equal to
+ * or greater than (in a modulo sense) the sync counter specified in
+ * the SYNC message.
+ *
+ * This allows other host-target components to synchronize their operation
+ * with HTT, e.g. to ensure that tx frames don't get transmitted until a
+ * security key has been downloaded to and activated by the target.
+ * In the absence of any explicit synchronization counter value
+ * specification, the target HTT FW will use zero as the default current
+ * sync value.
+ *
+ * The HTT target FW will suspend its host->target message processing as long
+ * as 0 < (in-band sync counter - out-of-band sync counter) & 0xff < 128.
+ */
+struct htt_oob_sync_req {
+	u8 sync_count;
+	__le16 rsvd0;
+} __packed;
+
+#define HTT_AGGR_CONF_MAX_NUM_AMSDU_SUBFRAMES_MASK 0x1F
+#define HTT_AGGR_CONF_MAX_NUM_AMSDU_SUBFRAMES_LSB  0
+
+struct htt_aggr_conf {
+	u8 max_num_ampdu_subframes;
+	union {
+		/* dont use bitfields; undefined behaviour */
+		u8 flags; /* see %HTT_AGGR_CONF_MAX_NUM_AMSDU_SUBFRAMES_ */
+		u8 max_num_amsdu_subframes:5;
+	} __packed;
+} __packed;
+
+#define HTT_MGMT_FRM_HDR_DOWNLOAD_LEN 32
+
+struct htt_mgmt_tx_desc {
+	u8 pad[sizeof(u32) - sizeof(struct htt_cmd_hdr)];
+	__le32 msdu_paddr;
+	__le32 desc_id;
+	__le32 len;
+	__le32 vdev_id;
+	u8 hdr[HTT_MGMT_FRM_HDR_DOWNLOAD_LEN];
+} __packed;
+
+enum htt_mgmt_tx_status {
+	HTT_MGMT_TX_STATUS_OK    = 0,
+	HTT_MGMT_TX_STATUS_RETRY = 1,
+	HTT_MGMT_TX_STATUS_DROP  = 2
+};
+
+/*=== target -> host messages ===============================================*/
+
+
+enum htt_t2h_msg_type {
+	HTT_T2H_MSG_TYPE_VERSION_CONF		= 0x0,
+	HTT_T2H_MSG_TYPE_RX_IND			= 0x1,
+	HTT_T2H_MSG_TYPE_RX_FLUSH		= 0x2,
+	HTT_T2H_MSG_TYPE_PEER_MAP		= 0x3,
+	HTT_T2H_MSG_TYPE_PEER_UNMAP		= 0x4,
+	HTT_T2H_MSG_TYPE_RX_ADDBA		= 0x5,
+	HTT_T2H_MSG_TYPE_RX_DELBA		= 0x6,
+	HTT_T2H_MSG_TYPE_TX_COMPL_IND		= 0x7,
+	HTT_T2H_MSG_TYPE_PKTLOG			= 0x8,
+	HTT_T2H_MSG_TYPE_STATS_CONF		= 0x9,
+	HTT_T2H_MSG_TYPE_RX_FRAG_IND		= 0xa,
+	HTT_T2H_MSG_TYPE_SEC_IND		= 0xb,
+	HTT_T2H_MSG_TYPE_RC_UPDATE_IND		= 0xc,
+	HTT_T2H_MSG_TYPE_TX_INSPECT_IND		= 0xd,
+	HTT_T2H_MSG_TYPE_MGMT_TX_COMPLETION	= 0xe,
+	HTT_T2H_MSG_TYPE_TEST,
+	/* keep this last */
+	HTT_T2H_NUM_MSGS
+};
+
+/*
+ * htt_resp_hdr - header for target-to-host messages
+ *
+ * msg_type: see htt_t2h_msg_type
+ */
+struct htt_resp_hdr {
+	u8 msg_type;
+} __packed;
+
+#define HTT_RESP_HDR_MSG_TYPE_OFFSET 0
+#define HTT_RESP_HDR_MSG_TYPE_MASK   0xff
+#define HTT_RESP_HDR_MSG_TYPE_LSB    0
+
+/* htt_ver_resp - response sent for htt_ver_req */
+struct htt_ver_resp {
+	u8 minor;
+	u8 major;
+	u8 rsvd0;
+} __packed;
+
+struct htt_mgmt_tx_completion {
+	u8 rsvd0;
+	u8 rsvd1;
+	u8 rsvd2;
+	__le32 desc_id;
+	__le32 status;
+} __packed;
+
+#define HTT_RX_INDICATION_INFO0_EXT_TID_MASK  (0x3F)
+#define HTT_RX_INDICATION_INFO0_EXT_TID_LSB   (0)
+#define HTT_RX_INDICATION_INFO0_FLUSH_VALID   (1 << 6)
+#define HTT_RX_INDICATION_INFO0_RELEASE_VALID (1 << 7)
+
+#define HTT_RX_INDICATION_INFO1_FLUSH_START_SEQNO_MASK   0x0000003F
+#define HTT_RX_INDICATION_INFO1_FLUSH_START_SEQNO_LSB    0
+#define HTT_RX_INDICATION_INFO1_FLUSH_END_SEQNO_MASK     0x00000FC0
+#define HTT_RX_INDICATION_INFO1_FLUSH_END_SEQNO_LSB      6
+#define HTT_RX_INDICATION_INFO1_RELEASE_START_SEQNO_MASK 0x0003F000
+#define HTT_RX_INDICATION_INFO1_RELEASE_START_SEQNO_LSB  12
+#define HTT_RX_INDICATION_INFO1_RELEASE_END_SEQNO_MASK   0x00FC0000
+#define HTT_RX_INDICATION_INFO1_RELEASE_END_SEQNO_LSB    18
+#define HTT_RX_INDICATION_INFO1_NUM_MPDU_RANGES_MASK     0xFF000000
+#define HTT_RX_INDICATION_INFO1_NUM_MPDU_RANGES_LSB      24
+
+struct htt_rx_indication_hdr {
+	u8 info0; /* %HTT_RX_INDICATION_INFO0_ */
+	__le16 peer_id;
+	__le32 info1; /* %HTT_RX_INDICATION_INFO1_ */
+} __packed;
+
+#define HTT_RX_INDICATION_INFO0_PHY_ERR_VALID    (1 << 0)
+#define HTT_RX_INDICATION_INFO0_LEGACY_RATE_MASK (0x1E)
+#define HTT_RX_INDICATION_INFO0_LEGACY_RATE_LSB  (1)
+#define HTT_RX_INDICATION_INFO0_LEGACY_RATE_CCK  (1 << 5)
+#define HTT_RX_INDICATION_INFO0_END_VALID        (1 << 6)
+#define HTT_RX_INDICATION_INFO0_START_VALID      (1 << 7)
+
+#define HTT_RX_INDICATION_INFO1_VHT_SIG_A1_MASK    0x00FFFFFF
+#define HTT_RX_INDICATION_INFO1_VHT_SIG_A1_LSB     0
+#define HTT_RX_INDICATION_INFO1_PREAMBLE_TYPE_MASK 0xFF000000
+#define HTT_RX_INDICATION_INFO1_PREAMBLE_TYPE_LSB  24
+
+#define HTT_RX_INDICATION_INFO2_VHT_SIG_A1_MASK 0x00FFFFFF
+#define HTT_RX_INDICATION_INFO2_VHT_SIG_A1_LSB  0
+#define HTT_RX_INDICATION_INFO2_SERVICE_MASK    0xFF000000
+#define HTT_RX_INDICATION_INFO2_SERVICE_LSB     24
+
+enum htt_rx_legacy_rate {
+	HTT_RX_OFDM_48 = 0,
+	HTT_RX_OFDM_24 = 1,
+	HTT_RX_OFDM_12,
+	HTT_RX_OFDM_6,
+	HTT_RX_OFDM_54,
+	HTT_RX_OFDM_36,
+	HTT_RX_OFDM_18,
+	HTT_RX_OFDM_9,
+
+	/* long preamble */
+	HTT_RX_CCK_11_LP = 0,
+	HTT_RX_CCK_5_5_LP = 1,
+	HTT_RX_CCK_2_LP,
+	HTT_RX_CCK_1_LP,
+	/* short preamble */
+	HTT_RX_CCK_11_SP,
+	HTT_RX_CCK_5_5_SP,
+	HTT_RX_CCK_2_SP
+};
+
+enum htt_rx_legacy_rate_type {
+	HTT_RX_LEGACY_RATE_OFDM = 0,
+	HTT_RX_LEGACY_RATE_CCK
+};
+
+enum htt_rx_preamble_type {
+	HTT_RX_LEGACY        = 0x4,
+	HTT_RX_HT            = 0x8,
+	HTT_RX_HT_WITH_TXBF  = 0x9,
+	HTT_RX_VHT           = 0xC,
+	HTT_RX_VHT_WITH_TXBF = 0xD,
+};
+
+/*
+ * Fields: phy_err_valid, phy_err_code, tsf,
+ * usec_timestamp, sub_usec_timestamp
+ * ..are valid only if end_valid == 1.
+ *
+ * Fields: rssi_chains, legacy_rate_type,
+ * legacy_rate_cck, preamble_type, service,
+ * vht_sig_*
+ * ..are valid only if start_valid == 1;
+ */
+struct htt_rx_indication_ppdu {
+	u8 combined_rssi;
+	u8 sub_usec_timestamp;
+	u8 phy_err_code;
+	u8 info0; /* HTT_RX_INDICATION_INFO0_ */
+	struct {
+		u8 pri20_db;
+		u8 ext20_db;
+		u8 ext40_db;
+		u8 ext80_db;
+	} __packed rssi_chains[4];
+	__le32 tsf;
+	__le32 usec_timestamp;
+	__le32 info1; /* HTT_RX_INDICATION_INFO1_ */
+	__le32 info2; /* HTT_RX_INDICATION_INFO2_ */
+} __packed;
+
+enum htt_rx_mpdu_status {
+	HTT_RX_IND_MPDU_STATUS_UNKNOWN = 0x0,
+	HTT_RX_IND_MPDU_STATUS_OK,
+	HTT_RX_IND_MPDU_STATUS_ERR_FCS,
+	HTT_RX_IND_MPDU_STATUS_ERR_DUP,
+	HTT_RX_IND_MPDU_STATUS_ERR_REPLAY,
+	HTT_RX_IND_MPDU_STATUS_ERR_INV_PEER,
+	/* only accept EAPOL frames */
+	HTT_RX_IND_MPDU_STATUS_UNAUTH_PEER,
+	HTT_RX_IND_MPDU_STATUS_OUT_OF_SYNC,
+	/* Non-data in promiscous mode */
+	HTT_RX_IND_MPDU_STATUS_MGMT_CTRL,
+	HTT_RX_IND_MPDU_STATUS_TKIP_MIC_ERR,
+	HTT_RX_IND_MPDU_STATUS_DECRYPT_ERR,
+	HTT_RX_IND_MPDU_STATUS_MPDU_LENGTH_ERR,
+	HTT_RX_IND_MPDU_STATUS_ENCRYPT_REQUIRED_ERR,
+	HTT_RX_IND_MPDU_STATUS_PRIVACY_ERR,
+
+	/*
+	 * MISC: discard for unspecified reasons.
+	 * Leave this enum value last.
+	 */
+	HTT_RX_IND_MPDU_STATUS_ERR_MISC = 0xFF
+};
+
+struct htt_rx_indication_mpdu_range {
+	u8 mpdu_count;
+	u8 mpdu_range_status; /* %htt_rx_mpdu_status */
+	u8 pad0;
+	u8 pad1;
+} __packed;
+
+struct htt_rx_indication_prefix {
+	__le16 fw_rx_desc_bytes;
+	u8 pad0;
+	u8 pad1;
+};
+
+struct htt_rx_indication {
+	struct htt_rx_indication_hdr hdr;
+	struct htt_rx_indication_ppdu ppdu;
+	struct htt_rx_indication_prefix prefix;
+
+	/*
+	 * the following fields are both dynamically sized, so
+	 * take care addressing them
+	 */
+
+	/* the size of this is %fw_rx_desc_bytes */
+	struct fw_rx_desc_base fw_desc;
+
+	/*
+	 * %mpdu_ranges starts after &%prefix + roundup(%fw_rx_desc_bytes, 4)
+	 * and has %num_mpdu_ranges elements.
+	 */
+	struct htt_rx_indication_mpdu_range mpdu_ranges[0];
+} __packed;
+
+static inline struct htt_rx_indication_mpdu_range *
+		htt_rx_ind_get_mpdu_ranges(struct htt_rx_indication *rx_ind)
+{
+	void *ptr = rx_ind;
+
+	ptr += sizeof(rx_ind->hdr)
+	     + sizeof(rx_ind->ppdu)
+	     + sizeof(rx_ind->prefix)
+	     + roundup(__le16_to_cpu(rx_ind->prefix.fw_rx_desc_bytes), 4);
+	return ptr;
+}
+
+enum htt_rx_flush_mpdu_status {
+	HTT_RX_FLUSH_MPDU_DISCARD = 0,
+	HTT_RX_FLUSH_MPDU_REORDER = 1,
+};
+
+/*
+ * htt_rx_flush - discard or reorder given range of mpdus
+ *
+ * Note: host must check if all sequence numbers between
+ *	[seq_num_start, seq_num_end-1] are valid.
+ */
+struct htt_rx_flush {
+	__le16 peer_id;
+	u8 tid;
+	u8 rsvd0;
+	u8 mpdu_status; /* %htt_rx_flush_mpdu_status */
+	u8 seq_num_start; /* it is 6 LSBs of 802.11 seq no */
+	u8 seq_num_end; /* it is 6 LSBs of 802.11 seq no */
+};
+
+struct htt_rx_peer_map {
+	u8 vdev_id;
+	__le16 peer_id;
+	u8 addr[6];
+	u8 rsvd0;
+	u8 rsvd1;
+} __packed;
+
+struct htt_rx_peer_unmap {
+	u8 rsvd0;
+	__le16 peer_id;
+} __packed;
+
+enum htt_security_types {
+	HTT_SECURITY_NONE,
+	HTT_SECURITY_WEP128,
+	HTT_SECURITY_WEP104,
+	HTT_SECURITY_WEP40,
+	HTT_SECURITY_TKIP,
+	HTT_SECURITY_TKIP_NOMIC,
+	HTT_SECURITY_AES_CCMP,
+	HTT_SECURITY_WAPI,
+
+	HTT_NUM_SECURITY_TYPES /* keep this last! */
+};
+
+enum htt_security_flags {
+#define HTT_SECURITY_TYPE_MASK 0x7F
+#define HTT_SECURITY_TYPE_LSB  0
+	HTT_SECURITY_IS_UNICAST = 1 << 7
+};
+
+struct htt_security_indication {
+	union {
+		/* dont use bitfields; undefined behaviour */
+		u8 flags; /* %htt_security_flags */
+		struct {
+			u8 security_type:7, /* %htt_security_types */
+			   is_unicast:1;
+		} __packed;
+	} __packed;
+	__le16 peer_id;
+	u8 michael_key[8];
+	u8 wapi_rsc[16];
+} __packed;
+
+#define HTT_RX_BA_INFO0_TID_MASK     0x000F
+#define HTT_RX_BA_INFO0_TID_LSB      0
+#define HTT_RX_BA_INFO0_PEER_ID_MASK 0xFFF0
+#define HTT_RX_BA_INFO0_PEER_ID_LSB  4
+
+struct htt_rx_addba {
+	u8 window_size;
+	__le16 info0; /* %HTT_RX_BA_INFO0_ */
+} __packed;
+
+struct htt_rx_delba {
+	u8 rsvd0;
+	__le16 info0; /* %HTT_RX_BA_INFO0_ */
+} __packed;
+
+enum htt_data_tx_status {
+	HTT_DATA_TX_STATUS_OK            = 0,
+	HTT_DATA_TX_STATUS_DISCARD       = 1,
+	HTT_DATA_TX_STATUS_NO_ACK        = 2,
+	HTT_DATA_TX_STATUS_POSTPONE      = 3, /* HL only */
+	HTT_DATA_TX_STATUS_DOWNLOAD_FAIL = 128
+};
+
+enum htt_data_tx_flags {
+#define HTT_DATA_TX_STATUS_MASK 0x07
+#define HTT_DATA_TX_STATUS_LSB  0
+#define HTT_DATA_TX_TID_MASK    0x78
+#define HTT_DATA_TX_TID_LSB     3
+	HTT_DATA_TX_TID_INVALID = 1 << 7
+};
+
+#define HTT_TX_COMPL_INV_MSDU_ID 0xFFFF
+
+struct htt_data_tx_completion {
+	union {
+		u8 flags;
+		struct {
+			u8 status:3,
+			   tid:4,
+			   tid_invalid:1;
+		} __packed;
+	} __packed;
+	u8 num_msdus;
+	u8 rsvd0;
+	__le16 msdus[0]; /* variable length based on %num_msdus */
+} __packed;
+
+struct htt_tx_compl_ind_base {
+	u32 hdr;
+	u16 payload[1/*or more*/];
+} __packed;
+
+struct htt_rc_tx_done_params {
+	u32 rate_code;
+	u32 rate_code_flags;
+	u32 flags;
+	u32 num_enqued; /* 1 for non-AMPDU */
+	u32 num_retries;
+	u32 num_failed; /* for AMPDU */
+	u32 ack_rssi;
+	u32 time_stamp;
+	u32 is_probe;
+};
+
+struct htt_rc_update {
+	u8 vdev_id;
+	__le16 peer_id;
+	u8 addr[6];
+	u8 num_elems;
+	u8 rsvd0;
+	struct htt_rc_tx_done_params params[0]; /* variable length %num_elems */
+} __packed;
+
+/* see htt_rx_indication for similar fields and descriptions */
+struct htt_rx_fragment_indication {
+	union {
+		u8 info0; /* %HTT_RX_FRAG_IND_INFO0_ */
+		struct {
+			u8 ext_tid:5,
+			   flush_valid:1;
+		} __packed;
+	} __packed;
+	__le16 peer_id;
+	__le32 info1; /* %HTT_RX_FRAG_IND_INFO1_ */
+	__le16 fw_rx_desc_bytes;
+	__le16 rsvd0;
+
+	u8 fw_msdu_rx_desc[0];
+} __packed;
+
+#define HTT_RX_FRAG_IND_INFO0_EXT_TID_MASK     0x1F
+#define HTT_RX_FRAG_IND_INFO0_EXT_TID_LSB      0
+#define HTT_RX_FRAG_IND_INFO0_FLUSH_VALID_MASK 0x20
+#define HTT_RX_FRAG_IND_INFO0_FLUSH_VALID_LSB  5
+
+#define HTT_RX_FRAG_IND_INFO1_FLUSH_SEQ_NUM_START_MASK 0x0000003F
+#define HTT_RX_FRAG_IND_INFO1_FLUSH_SEQ_NUM_START_LSB  0
+#define HTT_RX_FRAG_IND_INFO1_FLUSH_SEQ_NUM_END_MASK   0x00000FC0
+#define HTT_RX_FRAG_IND_INFO1_FLUSH_SEQ_NUM_END_LSB    6
+
+/*
+ * target -> host test message definition
+ *
+ * The following field definitions describe the format of the test
+ * message sent from the target to the host.
+ * The message consists of a 4-octet header, followed by a variable
+ * number of 32-bit integer values, followed by a variable number
+ * of 8-bit character values.
+ *
+ * |31                         16|15           8|7            0|
+ * |-----------------------------------------------------------|
+ * |          num chars          |   num ints   |   msg type   |
+ * |-----------------------------------------------------------|
+ * |                           int 0                           |
+ * |-----------------------------------------------------------|
+ * |                           int 1                           |
+ * |-----------------------------------------------------------|
+ * |                            ...                            |
+ * |-----------------------------------------------------------|
+ * |    char 3    |    char 2    |    char 1    |    char 0    |
+ * |-----------------------------------------------------------|
+ * |              |              |      ...     |    char 4    |
+ * |-----------------------------------------------------------|
+ *   - MSG_TYPE
+ *     Bits 7:0
+ *     Purpose: identifies this as a test message
+ *     Value: HTT_MSG_TYPE_TEST
+ *   - NUM_INTS
+ *     Bits 15:8
+ *     Purpose: indicate how many 32-bit integers follow the message header
+ *   - NUM_CHARS
+ *     Bits 31:16
+ *     Purpose: indicate how many 8-bit charaters follow the series of integers
+ */
+struct htt_rx_test {
+	u8 num_ints;
+	__le16 num_chars;
+
+	/* payload consists of 2 lists:
+	 *  a) num_ints * sizeof(__le32)
+	 *  b) num_chars * sizeof(u8) aligned to 4bytes */
+	u8 payload[0];
+} __packed;
+
+static inline __le32 *htt_rx_test_get_ints(struct htt_rx_test *rx_test)
+{
+	return (__le32 *)rx_test->payload;
+}
+
+static inline u8 *htt_rx_test_get_chars(struct htt_rx_test *rx_test)
+{
+	return rx_test->payload + (rx_test->num_ints * sizeof(__le32));
+}
+
+/*
+ * target -> host packet log message
+ *
+ * The following field definitions describe the format of the packet log
+ * message sent from the target to the host.
+ * The message consists of a 4-octet header,followed by a variable number
+ * of 32-bit character values.
+ *
+ * |31          24|23          16|15           8|7            0|
+ * |-----------------------------------------------------------|
+ * |              |              |              |   msg type   |
+ * |-----------------------------------------------------------|
+ * |                        payload                            |
+ * |-----------------------------------------------------------|
+ *   - MSG_TYPE
+ *     Bits 7:0
+ *     Purpose: identifies this as a test message
+ *     Value: HTT_MSG_TYPE_PACKETLOG
+ */
+struct htt_pktlog_msg {
+	u8 pad[3];
+	__le32 payload[1 /* or more */];
+} __packed;
+
+struct htt_dbg_stats_rx_reorder_stats {
+	/* Non QoS MPDUs received */
+	__le32 deliver_non_qos;
+
+	/* MPDUs received in-order */
+	__le32 deliver_in_order;
+
+	/* Flush due to reorder timer expired */
+	__le32 deliver_flush_timeout;
+
+	/* Flush due to move out of window */
+	__le32 deliver_flush_oow;
+
+	/* Flush due to DELBA */
+	__le32 deliver_flush_delba;
+
+	/* MPDUs dropped due to FCS error */
+	__le32 fcs_error;
+
+	/* MPDUs dropped due to monitor mode non-data packet */
+	__le32 mgmt_ctrl;
+
+	/* MPDUs dropped due to invalid peer */
+	__le32 invalid_peer;
+
+	/* MPDUs dropped due to duplication (non aggregation) */
+	__le32 dup_non_aggr;
+
+	/* MPDUs dropped due to processed before */
+	__le32 dup_past;
+
+	/* MPDUs dropped due to duplicate in reorder queue */
+	__le32 dup_in_reorder;
+
+	/* Reorder timeout happened */
+	__le32 reorder_timeout;
+
+	/* invalid bar ssn */
+	__le32 invalid_bar_ssn;
+
+	/* reorder reset due to bar ssn */
+	__le32 ssn_reset;
+};
+
+struct htt_dbg_stats_wal_tx_stats {
+	/* Num HTT cookies queued to dispatch list */
+	__le32 comp_queued;
+
+	/* Num HTT cookies dispatched */
+	__le32 comp_delivered;
+
+	/* Num MSDU queued to WAL */
+	__le32 msdu_enqued;
+
+	/* Num MPDU queue to WAL */
+	__le32 mpdu_enqued;
+
+	/* Num MSDUs dropped by WMM limit */
+	__le32 wmm_drop;
+
+	/* Num Local frames queued */
+	__le32 local_enqued;
+
+	/* Num Local frames done */
+	__le32 local_freed;
+
+	/* Num queued to HW */
+	__le32 hw_queued;
+
+	/* Num PPDU reaped from HW */
+	__le32 hw_reaped;
+
+	/* Num underruns */
+	__le32 underrun;
+
+	/* Num PPDUs cleaned up in TX abort */
+	__le32 tx_abort;
+
+	/* Num MPDUs requed by SW */
+	__le32 mpdus_requed;
+
+	/* excessive retries */
+	__le32 tx_ko;
+
+	/* data hw rate code */
+	__le32 data_rc;
+
+	/* Scheduler self triggers */
+	__le32 self_triggers;
+
+	/* frames dropped due to excessive sw retries */
+	__le32 sw_retry_failure;
+
+	/* illegal rate phy errors  */
+	__le32 illgl_rate_phy_err;
+
+	/* wal pdev continous xretry */
+	__le32 pdev_cont_xretry;
+
+	/* wal pdev continous xretry */
+	__le32 pdev_tx_timeout;
+
+	/* wal pdev resets  */
+	__le32 pdev_resets;
+
+	__le32 phy_underrun;
+
+	/* MPDU is more than txop limit */
+	__le32 txop_ovf;
+} __packed;
+
+struct htt_dbg_stats_wal_rx_stats {
+	/* Cnts any change in ring routing mid-ppdu */
+	__le32 mid_ppdu_route_change;
+
+	/* Total number of statuses processed */
+	__le32 status_rcvd;
+
+	/* Extra frags on rings 0-3 */
+	__le32 r0_frags;
+	__le32 r1_frags;
+	__le32 r2_frags;
+	__le32 r3_frags;
+
+	/* MSDUs / MPDUs delivered to HTT */
+	__le32 htt_msdus;
+	__le32 htt_mpdus;
+
+	/* MSDUs / MPDUs delivered to local stack */
+	__le32 loc_msdus;
+	__le32 loc_mpdus;
+
+	/* AMSDUs that have more MSDUs than the status ring size */
+	__le32 oversize_amsdu;
+
+	/* Number of PHY errors */
+	__le32 phy_errs;
+
+	/* Number of PHY errors drops */
+	__le32 phy_err_drop;
+
+	/* Number of mpdu errors - FCS, MIC, ENC etc. */
+	__le32 mpdu_errs;
+} __packed;
+
+struct htt_dbg_stats_wal_peer_stats {
+	__le32 dummy; /* REMOVE THIS ONCE REAL PEER STAT COUNTERS ARE ADDED */
+} __packed;
+
+struct htt_dbg_stats_wal_pdev_txrx {
+	struct htt_dbg_stats_wal_tx_stats tx_stats;
+	struct htt_dbg_stats_wal_rx_stats rx_stats;
+	struct htt_dbg_stats_wal_peer_stats peer_stats;
+} __packed;
+
+struct htt_dbg_stats_rx_rate_info {
+	__le32 mcs[10];
+	__le32 sgi[10];
+	__le32 nss[4];
+	__le32 stbc[10];
+	__le32 bw[3];
+	__le32 pream[6];
+	__le32 ldpc;
+	__le32 txbf;
+};
+
+/*
+ * htt_dbg_stats_status -
+ * present -     The requested stats have been delivered in full.
+ *               This indicates that either the stats information was contained
+ *               in its entirety within this message, or else this message
+ *               completes the delivery of the requested stats info that was
+ *               partially delivered through earlier STATS_CONF messages.
+ * partial -     The requested stats have been delivered in part.
+ *               One or more subsequent STATS_CONF messages with the same
+ *               cookie value will be sent to deliver the remainder of the
+ *               information.
+ * error -       The requested stats could not be delivered, for example due
+ *               to a shortage of memory to construct a message holding the
+ *               requested stats.
+ * invalid -     The requested stat type is either not recognized, or the
+ *               target is configured to not gather the stats type in question.
+ * - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+ * series_done - This special value indicates that no further stats info
+ *               elements are present within a series of stats info elems
+ *               (within a stats upload confirmation message).
+ */
+enum htt_dbg_stats_status {
+	HTT_DBG_STATS_STATUS_PRESENT     = 0,
+	HTT_DBG_STATS_STATUS_PARTIAL     = 1,
+	HTT_DBG_STATS_STATUS_ERROR       = 2,
+	HTT_DBG_STATS_STATUS_INVALID     = 3,
+	HTT_DBG_STATS_STATUS_SERIES_DONE = 7
+};
+
+/*
+ * target -> host statistics upload
+ *
+ * The following field definitions describe the format of the HTT target
+ * to host stats upload confirmation message.
+ * The message contains a cookie echoed from the HTT host->target stats
+ * upload request, which identifies which request the confirmation is
+ * for, and a series of tag-length-value stats information elements.
+ * The tag-length header for each stats info element also includes a
+ * status field, to indicate whether the request for the stat type in
+ * question was fully met, partially met, unable to be met, or invalid
+ * (if the stat type in question is disabled in the target).
+ * A special value of all 1's in this status field is used to indicate
+ * the end of the series of stats info elements.
+ *
+ *
+ * |31                         16|15           8|7   5|4       0|
+ * |------------------------------------------------------------|
+ * |                  reserved                  |    msg type   |
+ * |------------------------------------------------------------|
+ * |                        cookie LSBs                         |
+ * |------------------------------------------------------------|
+ * |                        cookie MSBs                         |
+ * |------------------------------------------------------------|
+ * |      stats entry length     |   reserved   |  S  |stat type|
+ * |------------------------------------------------------------|
+ * |                                                            |
+ * |                  type-specific stats info                  |
+ * |                                                            |
+ * |------------------------------------------------------------|
+ * |      stats entry length     |   reserved   |  S  |stat type|
+ * |------------------------------------------------------------|
+ * |                                                            |
+ * |                  type-specific stats info                  |
+ * |                                                            |
+ * |------------------------------------------------------------|
+ * |              n/a            |   reserved   | 111 |   n/a   |
+ * |------------------------------------------------------------|
+ * Header fields:
+ *  - MSG_TYPE
+ *    Bits 7:0
+ *    Purpose: identifies this is a statistics upload confirmation message
+ *    Value: 0x9
+ *  - COOKIE_LSBS
+ *    Bits 31:0
+ *    Purpose: Provide a mechanism to match a target->host stats confirmation
+ *        message with its preceding host->target stats request message.
+ *    Value: LSBs of the opaque cookie specified by the host-side requestor
+ *  - COOKIE_MSBS
+ *    Bits 31:0
+ *    Purpose: Provide a mechanism to match a target->host stats confirmation
+ *        message with its preceding host->target stats request message.
+ *    Value: MSBs of the opaque cookie specified by the host-side requestor
+ *
+ * Stats Information Element tag-length header fields:
+ *  - STAT_TYPE
+ *    Bits 4:0
+ *    Purpose: identifies the type of statistics info held in the
+ *        following information element
+ *    Value: htt_dbg_stats_type
+ *  - STATUS
+ *    Bits 7:5
+ *    Purpose: indicate whether the requested stats are present
+ *    Value: htt_dbg_stats_status, including a special value (0x7) to mark
+ *        the completion of the stats entry series
+ *  - LENGTH
+ *    Bits 31:16
+ *    Purpose: indicate the stats information size
+ *    Value: This field specifies the number of bytes of stats information
+ *       that follows the element tag-length header.
+ *       It is expected but not required that this length is a multiple of
+ *       4 bytes.  Even if the length is not an integer multiple of 4, the
+ *       subsequent stats entry header will begin on a 4-byte aligned
+ *       boundary.
+ */
+
+#define HTT_STATS_CONF_ITEM_INFO_STAT_TYPE_MASK 0x1F
+#define HTT_STATS_CONF_ITEM_INFO_STAT_TYPE_LSB  0
+#define HTT_STATS_CONF_ITEM_INFO_STATUS_MASK    0xE0
+#define HTT_STATS_CONF_ITEM_INFO_STATUS_LSB     5
+
+struct htt_stats_conf_item {
+	union {
+		u8 info;
+		struct {
+			u8 stat_type:5; /* %HTT_DBG_STATS_ */
+			u8 status:3; /* %HTT_DBG_STATS_STATUS_ */
+		} __packed;
+	} __packed;
+	u8 pad;
+	__le16 length;
+	u8 payload[0]; /* roundup(length, 4) long */
+} __packed;
+
+struct htt_stats_conf {
+	u8 pad[3];
+	__le32 cookie_lsb;
+	__le32 cookie_msb;
+
+	/* each item has variable length! */
+	struct htt_stats_conf_item items[0];
+} __packed;
+
+static inline struct htt_stats_conf_item *htt_stats_conf_next_item(
+					const struct htt_stats_conf_item *item)
+{
+	return (void *)item + sizeof(*item) + roundup(item->length, 4);
+}
+/*
+ * host -> target FRAG DESCRIPTOR/MSDU_EXT DESC bank
+ *
+ * The following field definitions describe the format of the HTT host
+ * to target frag_desc/msdu_ext bank configuration message.
+ * The message contains the based address and the min and max id of the
+ * MSDU_EXT/FRAG_DESC that will be used by the HTT to map MSDU DESC and
+ * MSDU_EXT/FRAG_DESC.
+ * HTT will use id in HTT descriptor instead sending the frag_desc_ptr.
+ * For QCA988X HW the firmware will use fragment_desc_ptr but in WIFI2.0
+ * the hardware does the mapping/translation.
+ *
+ * Total banks that can be configured is configured to 16.
+ *
+ * This should be called before any TX has be initiated by the HTT
+ *
+ * |31                         16|15           8|7   5|4       0|
+ * |------------------------------------------------------------|
+ * | DESC_SIZE    |  NUM_BANKS   | RES |SWP|pdev|    msg type   |
+ * |------------------------------------------------------------|
+ * |                     BANK0_BASE_ADDRESS                     |
+ * |------------------------------------------------------------|
+ * |                            ...                             |
+ * |------------------------------------------------------------|
+ * |                    BANK15_BASE_ADDRESS                     |
+ * |------------------------------------------------------------|
+ * |       BANK0_MAX_ID          |       BANK0_MIN_ID           |
+ * |------------------------------------------------------------|
+ * |                            ...                             |
+ * |------------------------------------------------------------|
+ * |       BANK15_MAX_ID         |       BANK15_MIN_ID          |
+ * |------------------------------------------------------------|
+ * Header fields:
+ *  - MSG_TYPE
+ *    Bits 7:0
+ *    Value: 0x6
+ *  - BANKx_BASE_ADDRESS
+ *    Bits 31:0
+ *    Purpose: Provide a mechanism to specify the base address of the MSDU_EXT
+ *         bank physical/bus address.
+ *  - BANKx_MIN_ID
+ *    Bits 15:0
+ *    Purpose: Provide a mechanism to specify the min index that needs to
+ *          mapped.
+ *  - BANKx_MAX_ID
+ *    Bits 31:16
+ *    Purpose: Provide a mechanism to specify the max index that needs to
+ *
+ */
+struct htt_frag_desc_bank_id {
+	__le16 bank_min_id;
+	__le16 bank_max_id;
+} __packed;
+
+/* real is 16 but it wouldn't fit in the max htt message size
+ * so we use a conservatively safe value for now */
+#define HTT_FRAG_DESC_BANK_MAX 4
+
+#define HTT_FRAG_DESC_BANK_CFG_INFO_PDEV_ID_MASK 0x03
+#define HTT_FRAG_DESC_BANK_CFG_INFO_PDEV_ID_LSB  0
+#define HTT_FRAG_DESC_BANK_CFG_INFO_SWAP         (1 << 2)
+
+struct htt_frag_desc_bank_cfg {
+	u8 info; /* HTT_FRAG_DESC_BANK_CFG_INFO_ */
+	u8 num_banks;
+	u8 desc_size;
+	__le32 bank_base_addrs[HTT_FRAG_DESC_BANK_MAX];
+	struct htt_frag_desc_bank_id bank_id[HTT_FRAG_DESC_BANK_MAX];
+} __packed;
+
+union htt_rx_pn_t {
+	/* WEP: 24-bit PN */
+	u32 pn24;
+
+	/* TKIP or CCMP: 48-bit PN */
+	u_int64_t pn48;
+
+	/* WAPI: 128-bit PN */
+	u_int64_t pn128[2];
+};
+
+struct htt_cmd {
+	struct htt_cmd_hdr hdr;
+	union {
+		struct htt_ver_req ver_req;
+		struct htt_mgmt_tx_desc mgmt_tx;
+		struct htt_data_tx_desc data_tx;
+		struct htt_rx_ring_setup rx_setup;
+		struct htt_stats_req stats_req;
+		struct htt_oob_sync_req oob_sync_req;
+		struct htt_aggr_conf aggr_conf;
+		struct htt_frag_desc_bank_cfg frag_desc_bank_cfg;
+	};
+} __packed;
+
+struct htt_resp {
+	struct htt_resp_hdr hdr;
+	union {
+		struct htt_ver_resp ver_resp;
+		struct htt_mgmt_tx_completion mgmt_tx_completion;
+		struct htt_data_tx_completion data_tx_completion;
+		struct htt_rx_indication rx_ind;
+		struct htt_rx_fragment_indication rx_frag_ind;
+		struct htt_rx_peer_map peer_map;
+		struct htt_rx_peer_unmap peer_unmap;
+		struct htt_rx_flush rx_flush;
+		struct htt_rx_addba rx_addba;
+		struct htt_rx_delba rx_delba;
+		struct htt_security_indication security_indication;
+		struct htt_rc_update rc_update;
+		struct htt_rx_test rx_test;
+		struct htt_pktlog_msg pktlog_msg;
+		struct htt_stats_conf stats_conf;
+	};
+} __packed;
+
+
+/*** host side structures follow ***/
+
+struct htt_tx_done {
+	u32 msdu_id;
+	bool discard;
+	bool no_ack;
+};
+
+struct htt_peer_map_event {
+	u8 vdev_id;
+	u16 peer_id;
+	u8 addr[ETH_ALEN];
+};
+
+struct htt_peer_unmap_event {
+	u16 peer_id;
+};
+
+struct htt_rx_info {
+	struct sk_buff *skb;
+	enum htt_rx_mpdu_status status;
+	enum htt_rx_mpdu_encrypt_type encrypt_type;
+	s8 signal;
+	struct {
+		u8 info0;
+		u32 info1;
+		u32 info2;
+	} rate;
+	bool fcs_err;
+};
+
+struct ath10k_htt {
+	struct ath10k *ar;
+	enum ath10k_htc_ep_id eid;
+
+	int max_throughput_mbps;
+	u8 target_version_major;
+	u8 target_version_minor;
+	struct completion target_version_received;
+
+	struct {
+		/*
+		 * Ring of network buffer objects - This ring is
+		 * used exclusively by the host SW. This ring
+		 * mirrors the dev_addrs_ring that is shared
+		 * between the host SW and the MAC HW. The host SW
+		 * uses this netbufs ring to locate the network
+		 * buffer objects whose data buffers the HW has
+		 * filled.
+		 */
+		struct sk_buff **netbufs_ring;
+		/*
+		 * Ring of buffer addresses -
+		 * This ring holds the "physical" device address of the
+		 * rx buffers the host SW provides for the MAC HW to
+		 * fill.
+		 */
+		__le32 *paddrs_ring;
+
+		/*
+		 * Base address of ring, as a "physical" device address
+		 * rather than a CPU address.
+		 */
+		dma_addr_t base_paddr;
+
+		/* how many elems in the ring (power of 2) */
+		int size;
+
+		/* size - 1 */
+		unsigned size_mask;
+
+		/* how many rx buffers to keep in the ring */
+		int fill_level;
+
+		/* how many rx buffers (full+empty) are in the ring */
+		int fill_cnt;
+
+		/*
+		 * alloc_idx - where HTT SW has deposited empty buffers
+		 * This is allocated in consistent mem, so that the FW can
+		 * read this variable, and program the HW's FW_IDX reg with
+		 * the value of this shadow register.
+		 */
+		struct {
+			__le32 *vaddr;
+			dma_addr_t paddr;
+		} alloc_idx;
+
+		/* where HTT SW has processed bufs filled by rx MAC DMA */
+		struct {
+			unsigned msdu_payld;
+		} sw_rd_idx;
+
+		/*
+		 * refill_retry_timer - timer triggered when the ring is
+		 * not refilled to the level expected
+		 */
+		struct timer_list refill_retry_timer;
+
+		/* Protects access to all rx ring buffer state variables */
+		spinlock_t lock;
+	} rx_ring;
+
+	unsigned int prefetch_len;
+
+	/* Protects access to %pending_tx, %used_msdu_ids */
+	spinlock_t tx_lock;
+	int max_num_pending_tx;
+	int num_pending_tx;
+	struct sk_buff **pending_tx;
+	unsigned long *used_msdu_ids; /* bitmap */
+	wait_queue_head_t empty_tx_wq;
+
+	/* set if host-fw communication goes haywire
+	 * used to avoid further failures */
+	bool rx_confused;
+};
+
+#define RX_HTT_HDR_STATUS_LEN 64
+
+/* This structure layout is programmed via rx ring setup
+ * so that FW knows how to transfer the rx descriptor to the host.
+ * Buffers like this are placed on the rx ring. */
+struct htt_rx_desc {
+	union {
+		/* This field is filled on the host using the msdu buffer
+		 * from htt_rx_indication */
+		struct fw_rx_desc_base fw_desc;
+		u32 pad;
+	} __packed;
+	struct {
+		struct rx_attention attention;
+		struct rx_frag_info frag_info;
+		struct rx_mpdu_start mpdu_start;
+		struct rx_msdu_start msdu_start;
+		struct rx_msdu_end msdu_end;
+		struct rx_mpdu_end mpdu_end;
+		struct rx_ppdu_start ppdu_start;
+		struct rx_ppdu_end ppdu_end;
+	} __packed;
+	u8 rx_hdr_status[RX_HTT_HDR_STATUS_LEN];
+	u8 msdu_payload[0];
+};
+
+#define HTT_RX_DESC_ALIGN 8
+
+#define HTT_MAC_ADDR_LEN 6
+
+/*
+ * FIX THIS
+ * Should be: sizeof(struct htt_host_rx_desc) + max rx MSDU size,
+ * rounded up to a cache line size.
+ */
+#define HTT_RX_BUF_SIZE 1920
+#define HTT_RX_MSDU_SIZE (HTT_RX_BUF_SIZE - (int)sizeof(struct htt_rx_desc))
+
+/*
+ * DMA_MAP expects the buffer to be an integral number of cache lines.
+ * Rather than checking the actual cache line size, this code makes a
+ * conservative estimate of what the cache line size could be.
+ */
+#define HTT_LOG2_MAX_CACHE_LINE_SIZE 7	/* 2^7 = 128 */
+#define HTT_MAX_CACHE_LINE_SIZE_MASK ((1 << HTT_LOG2_MAX_CACHE_LINE_SIZE) - 1)
+
+struct ath10k_htt *ath10k_htt_attach(struct ath10k *ar);
+int ath10k_htt_attach_target(struct ath10k_htt *htt);
+void ath10k_htt_detach(struct ath10k_htt *htt);
+
+int ath10k_htt_tx_attach(struct ath10k_htt *htt);
+void ath10k_htt_tx_detach(struct ath10k_htt *htt);
+int ath10k_htt_rx_attach(struct ath10k_htt *htt);
+void ath10k_htt_rx_detach(struct ath10k_htt *htt);
+void ath10k_htt_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb);
+void ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb);
+int ath10k_htt_h2t_ver_req_msg(struct ath10k_htt *htt);
+int ath10k_htt_send_rx_ring_cfg_ll(struct ath10k_htt *htt);
+
+void __ath10k_htt_tx_dec_pending(struct ath10k_htt *htt);
+int ath10k_htt_tx_alloc_msdu_id(struct ath10k_htt *htt);
+void ath10k_htt_tx_free_msdu_id(struct ath10k_htt *htt, u16 msdu_id);
+int ath10k_htt_mgmt_tx(struct ath10k_htt *htt, struct sk_buff *);
+int ath10k_htt_tx(struct ath10k_htt *htt, struct sk_buff *);
+#endif
diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
new file mode 100644
index 0000000..de058d7
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -0,0 +1,1167 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "htc.h"
+#include "htt.h"
+#include "txrx.h"
+#include "debug.h"
+
+#include <linux/log2.h>
+
+/* slightly larger than one large A-MPDU */
+#define HTT_RX_RING_SIZE_MIN 128
+
+/* roughly 20 ms @ 1 Gbps of 1500B MSDUs */
+#define HTT_RX_RING_SIZE_MAX 2048
+
+#define HTT_RX_AVG_FRM_BYTES 1000
+
+/* ms, very conservative */
+#define HTT_RX_HOST_LATENCY_MAX_MS 20
+
+/* ms, conservative */
+#define HTT_RX_HOST_LATENCY_WORST_LIKELY_MS 10
+
+/* when under memory pressure rx ring refill may fail and needs a retry */
+#define HTT_RX_RING_REFILL_RETRY_MS 50
+
+static int ath10k_htt_rx_ring_size(struct ath10k_htt *htt)
+{
+	int size;
+
+	/*
+	 * It is expected that the host CPU will typically be able to
+	 * service the rx indication from one A-MPDU before the rx
+	 * indication from the subsequent A-MPDU happens, roughly 1-2 ms
+	 * later. However, the rx ring should be sized very conservatively,
+	 * to accomodate the worst reasonable delay before the host CPU
+	 * services a rx indication interrupt.
+	 *
+	 * The rx ring need not be kept full of empty buffers. In theory,
+	 * the htt host SW can dynamically track the low-water mark in the
+	 * rx ring, and dynamically adjust the level to which the rx ring
+	 * is filled with empty buffers, to dynamically meet the desired
+	 * low-water mark.
+	 *
+	 * In contrast, it's difficult to resize the rx ring itself, once
+	 * it's in use. Thus, the ring itself should be sized very
+	 * conservatively, while the degree to which the ring is filled
+	 * with empty buffers should be sized moderately conservatively.
+	 */
+
+	/* 1e6 bps/mbps / 1e3 ms per sec = 1000 */
+	size =
+	    htt->max_throughput_mbps +
+	    1000  /
+	    (8 * HTT_RX_AVG_FRM_BYTES) * HTT_RX_HOST_LATENCY_MAX_MS;
+
+	if (size < HTT_RX_RING_SIZE_MIN)
+		size = HTT_RX_RING_SIZE_MIN;
+
+	if (size > HTT_RX_RING_SIZE_MAX)
+		size = HTT_RX_RING_SIZE_MAX;
+
+	size = roundup_pow_of_two(size);
+
+	return size;
+}
+
+static int ath10k_htt_rx_ring_fill_level(struct ath10k_htt *htt)
+{
+	int size;
+
+	/* 1e6 bps/mbps / 1e3 ms per sec = 1000 */
+	size =
+	    htt->max_throughput_mbps *
+	    1000  /
+	    (8 * HTT_RX_AVG_FRM_BYTES) * HTT_RX_HOST_LATENCY_WORST_LIKELY_MS;
+
+	/*
+	 * Make sure the fill level is at least 1 less than the ring size.
+	 * Leaving 1 element empty allows the SW to easily distinguish
+	 * between a full ring vs. an empty ring.
+	 */
+	if (size >= htt->rx_ring.size)
+		size = htt->rx_ring.size - 1;
+
+	return size;
+}
+
+static void ath10k_htt_rx_ring_free(struct ath10k_htt *htt)
+{
+	struct sk_buff *skb;
+	struct ath10k_skb_cb *cb;
+	int i;
+
+	for (i = 0; i < htt->rx_ring.fill_cnt; i++) {
+		skb = htt->rx_ring.netbufs_ring[i];
+		cb = ATH10K_SKB_CB(skb);
+		dma_unmap_single(htt->ar->dev, cb->paddr,
+				 skb->len + skb_tailroom(skb),
+				 DMA_FROM_DEVICE);
+		dev_kfree_skb_any(skb);
+	}
+
+	htt->rx_ring.fill_cnt = 0;
+}
+
+static int __ath10k_htt_rx_ring_fill_n(struct ath10k_htt *htt, int num)
+{
+	struct htt_rx_desc *rx_desc;
+	struct sk_buff *skb;
+	dma_addr_t paddr;
+	int ret = 0, idx;
+
+	idx = __le32_to_cpu(*(htt->rx_ring.alloc_idx.vaddr));
+	while (num > 0) {
+		skb = dev_alloc_skb(HTT_RX_BUF_SIZE + HTT_RX_DESC_ALIGN);
+		if (!skb) {
+			ret = -ENOMEM;
+			goto fail;
+		}
+
+		if (!IS_ALIGNED((unsigned long)skb->data, HTT_RX_DESC_ALIGN))
+			skb_pull(skb,
+				 PTR_ALIGN(skb->data, HTT_RX_DESC_ALIGN) -
+				 skb->data);
+
+		/* Clear rx_desc attention word before posting to Rx ring */
+		rx_desc = (struct htt_rx_desc *)skb->data;
+		rx_desc->attention.flags = __cpu_to_le32(0);
+
+		paddr = dma_map_single(htt->ar->dev, skb->data,
+				       skb->len + skb_tailroom(skb),
+				       DMA_FROM_DEVICE);
+
+		if (unlikely(dma_mapping_error(htt->ar->dev, paddr))) {
+			dev_kfree_skb_any(skb);
+			ret = -ENOMEM;
+			goto fail;
+		}
+
+		ATH10K_SKB_CB(skb)->paddr = paddr;
+		htt->rx_ring.netbufs_ring[idx] = skb;
+		htt->rx_ring.paddrs_ring[idx] = __cpu_to_le32(paddr);
+		htt->rx_ring.fill_cnt++;
+
+		num--;
+		idx++;
+		idx &= htt->rx_ring.size_mask;
+	}
+
+fail:
+	*(htt->rx_ring.alloc_idx.vaddr) = __cpu_to_le32(idx);
+	return ret;
+}
+
+static int ath10k_htt_rx_ring_fill_n(struct ath10k_htt *htt, int num)
+{
+	lockdep_assert_held(&htt->rx_ring.lock);
+	return __ath10k_htt_rx_ring_fill_n(htt, num);
+}
+
+static void ath10k_htt_rx_msdu_buff_replenish(struct ath10k_htt *htt)
+{
+	int ret, num_to_fill;
+
+	spin_lock_bh(&htt->rx_ring.lock);
+	num_to_fill = htt->rx_ring.fill_level - htt->rx_ring.fill_cnt;
+	ret = ath10k_htt_rx_ring_fill_n(htt, num_to_fill);
+	if (ret == -ENOMEM) {
+		/*
+		 * Failed to fill it to the desired level -
+		 * we'll start a timer and try again next time.
+		 * As long as enough buffers are left in the ring for
+		 * another A-MPDU rx, no special recovery is needed.
+		 */
+		mod_timer(&htt->rx_ring.refill_retry_timer, jiffies +
+			  msecs_to_jiffies(HTT_RX_RING_REFILL_RETRY_MS));
+	}
+	spin_unlock_bh(&htt->rx_ring.lock);
+}
+
+static void ath10k_htt_rx_ring_refill_retry(unsigned long arg)
+{
+	struct ath10k_htt *htt = (struct ath10k_htt *)arg;
+	ath10k_htt_rx_msdu_buff_replenish(htt);
+}
+
+static unsigned ath10k_htt_rx_ring_elems(struct ath10k_htt *htt)
+{
+	return (__le32_to_cpu(*htt->rx_ring.alloc_idx.vaddr) -
+		htt->rx_ring.sw_rd_idx.msdu_payld) & htt->rx_ring.size_mask;
+}
+
+void ath10k_htt_rx_detach(struct ath10k_htt *htt)
+{
+	int sw_rd_idx = htt->rx_ring.sw_rd_idx.msdu_payld;
+
+	del_timer_sync(&htt->rx_ring.refill_retry_timer);
+
+	while (sw_rd_idx != __le32_to_cpu(*(htt->rx_ring.alloc_idx.vaddr))) {
+		struct sk_buff *skb =
+				htt->rx_ring.netbufs_ring[sw_rd_idx];
+		struct ath10k_skb_cb *cb = ATH10K_SKB_CB(skb);
+
+		dma_unmap_single(htt->ar->dev, cb->paddr,
+				 skb->len + skb_tailroom(skb),
+				 DMA_FROM_DEVICE);
+		dev_kfree_skb_any(htt->rx_ring.netbufs_ring[sw_rd_idx]);
+		sw_rd_idx++;
+		sw_rd_idx &= htt->rx_ring.size_mask;
+	}
+
+	dma_free_coherent(htt->ar->dev,
+			  (htt->rx_ring.size *
+			   sizeof(htt->rx_ring.paddrs_ring)),
+			  htt->rx_ring.paddrs_ring,
+			  htt->rx_ring.base_paddr);
+
+	dma_free_coherent(htt->ar->dev,
+			  sizeof(*htt->rx_ring.alloc_idx.vaddr),
+			  htt->rx_ring.alloc_idx.vaddr,
+			  htt->rx_ring.alloc_idx.paddr);
+
+	kfree(htt->rx_ring.netbufs_ring);
+}
+
+static inline struct sk_buff *ath10k_htt_rx_netbuf_pop(struct ath10k_htt *htt)
+{
+	int idx;
+	struct sk_buff *msdu;
+
+	spin_lock_bh(&htt->rx_ring.lock);
+
+	if (ath10k_htt_rx_ring_elems(htt) == 0)
+		ath10k_warn("htt rx ring is empty!\n");
+
+	idx = htt->rx_ring.sw_rd_idx.msdu_payld;
+	msdu = htt->rx_ring.netbufs_ring[idx];
+
+	idx++;
+	idx &= htt->rx_ring.size_mask;
+	htt->rx_ring.sw_rd_idx.msdu_payld = idx;
+	htt->rx_ring.fill_cnt--;
+
+	spin_unlock_bh(&htt->rx_ring.lock);
+	return msdu;
+}
+
+static void ath10k_htt_rx_free_msdu_chain(struct sk_buff *skb)
+{
+	struct sk_buff *next;
+
+	while (skb) {
+		next = skb->next;
+		dev_kfree_skb_any(skb);
+		skb = next;
+	}
+}
+
+static int ath10k_htt_rx_amsdu_pop(struct ath10k_htt *htt,
+				   u8 **fw_desc, int *fw_desc_len,
+				   struct sk_buff **head_msdu,
+				   struct sk_buff **tail_msdu)
+{
+	int msdu_len, msdu_chaining = 0;
+	struct sk_buff *msdu;
+	struct htt_rx_desc *rx_desc;
+
+	if (ath10k_htt_rx_ring_elems(htt) == 0)
+		ath10k_warn("htt rx ring is empty!\n");
+
+	if (htt->rx_confused) {
+		ath10k_warn("htt is confused. refusing rx\n");
+		return 0;
+	}
+
+	msdu = *head_msdu = ath10k_htt_rx_netbuf_pop(htt);
+	while (msdu) {
+		int last_msdu, msdu_len_invalid, msdu_chained;
+
+		dma_unmap_single(htt->ar->dev,
+				 ATH10K_SKB_CB(msdu)->paddr,
+				 msdu->len + skb_tailroom(msdu),
+				 DMA_FROM_DEVICE);
+
+		ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt rx: ",
+				msdu->data, msdu->len + skb_tailroom(msdu));
+
+		rx_desc = (struct htt_rx_desc *)msdu->data;
+
+		/* FIXME: we must report msdu payload since this is what caller
+		 *        expects now */
+		skb_put(msdu, offsetof(struct htt_rx_desc, msdu_payload));
+		skb_pull(msdu, offsetof(struct htt_rx_desc, msdu_payload));
+
+		/*
+		 * Sanity check - confirm the HW is finished filling in the
+		 * rx data.
+		 * If the HW and SW are working correctly, then it's guaranteed
+		 * that the HW's MAC DMA is done before this point in the SW.
+		 * To prevent the case that we handle a stale Rx descriptor,
+		 * just assert for now until we have a way to recover.
+		 */
+		if (!(__le32_to_cpu(rx_desc->attention.flags)
+				& RX_ATTENTION_FLAGS_MSDU_DONE)) {
+			ath10k_htt_rx_free_msdu_chain(*head_msdu);
+			*head_msdu = NULL;
+			msdu = NULL;
+			ath10k_err("htt rx stopped. cannot recover\n");
+			htt->rx_confused = true;
+			break;
+		}
+
+		/*
+		 * Copy the FW rx descriptor for this MSDU from the rx
+		 * indication message into the MSDU's netbuf. HL uses the
+		 * same rx indication message definition as LL, and simply
+		 * appends new info (fields from the HW rx desc, and the
+		 * MSDU payload itself). So, the offset into the rx
+		 * indication message only has to account for the standard
+		 * offset of the per-MSDU FW rx desc info within the
+		 * message, and how many bytes of the per-MSDU FW rx desc
+		 * info have already been consumed. (And the endianness of
+		 * the host, since for a big-endian host, the rx ind
+		 * message contents, including the per-MSDU rx desc bytes,
+		 * were byteswapped during upload.)
+		 */
+		if (*fw_desc_len > 0) {
+			rx_desc->fw_desc.info0 = **fw_desc;
+			/*
+			 * The target is expected to only provide the basic
+			 * per-MSDU rx descriptors. Just to be sure, verify
+			 * that the target has not attached extension data
+			 * (e.g. LRO flow ID).
+			 */
+
+			/* or more, if there's extension data */
+			(*fw_desc)++;
+			(*fw_desc_len)--;
+		} else {
+			/*
+			 * When an oversized AMSDU happened, FW will lost
+			 * some of MSDU status - in this case, the FW
+			 * descriptors provided will be less than the
+			 * actual MSDUs inside this MPDU. Mark the FW
+			 * descriptors so that it will still deliver to
+			 * upper stack, if no CRC error for this MPDU.
+			 *
+			 * FIX THIS - the FW descriptors are actually for
+			 * MSDUs in the end of this A-MSDU instead of the
+			 * beginning.
+			 */
+			rx_desc->fw_desc.info0 = 0;
+		}
+
+		msdu_len_invalid = !!(__le32_to_cpu(rx_desc->attention.flags)
+					& (RX_ATTENTION_FLAGS_MPDU_LENGTH_ERR |
+					   RX_ATTENTION_FLAGS_MSDU_LENGTH_ERR));
+		msdu_len = MS(__le32_to_cpu(rx_desc->msdu_start.info0),
+			      RX_MSDU_START_INFO0_MSDU_LENGTH);
+		msdu_chained = rx_desc->frag_info.ring2_more_count;
+
+		if (msdu_len_invalid)
+			msdu_len = 0;
+
+		skb_trim(msdu, 0);
+		skb_put(msdu, min(msdu_len, HTT_RX_MSDU_SIZE));
+		msdu_len -= msdu->len;
+
+		/* FIXME: Do chained buffers include htt_rx_desc or not? */
+		while (msdu_chained--) {
+			struct sk_buff *next = ath10k_htt_rx_netbuf_pop(htt);
+
+			dma_unmap_single(htt->ar->dev,
+					 ATH10K_SKB_CB(next)->paddr,
+					 next->len + skb_tailroom(next),
+					 DMA_FROM_DEVICE);
+
+			ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt rx: ",
+					next->data,
+					next->len + skb_tailroom(next));
+
+			skb_trim(next, 0);
+			skb_put(next, min(msdu_len, HTT_RX_BUF_SIZE));
+			msdu_len -= next->len;
+
+			msdu->next = next;
+			msdu = next;
+			msdu_chaining = 1;
+		}
+
+		if (msdu_len > 0) {
+			/* This may suggest FW bug? */
+			ath10k_warn("htt rx msdu len not consumed (%d)\n",
+				    msdu_len);
+		}
+
+		last_msdu = __le32_to_cpu(rx_desc->msdu_end.info0) &
+				RX_MSDU_END_INFO0_LAST_MSDU;
+
+		if (last_msdu) {
+			msdu->next = NULL;
+			break;
+		} else {
+			struct sk_buff *next = ath10k_htt_rx_netbuf_pop(htt);
+			msdu->next = next;
+			msdu = next;
+		}
+	}
+	*tail_msdu = msdu;
+
+	/*
+	 * Don't refill the ring yet.
+	 *
+	 * First, the elements popped here are still in use - it is not
+	 * safe to overwrite them until the matching call to
+	 * mpdu_desc_list_next. Second, for efficiency it is preferable to
+	 * refill the rx ring with 1 PPDU's worth of rx buffers (something
+	 * like 32 x 3 buffers), rather than one MPDU's worth of rx buffers
+	 * (something like 3 buffers). Consequently, we'll rely on the txrx
+	 * SW to tell us when it is done pulling all the PPDU's rx buffers
+	 * out of the rx ring, and then refill it just once.
+	 */
+
+	return msdu_chaining;
+}
+
+int ath10k_htt_rx_attach(struct ath10k_htt *htt)
+{
+	dma_addr_t paddr;
+	void *vaddr;
+	struct timer_list *timer = &htt->rx_ring.refill_retry_timer;
+
+	htt->rx_ring.size = ath10k_htt_rx_ring_size(htt);
+	if (!is_power_of_2(htt->rx_ring.size)) {
+		ath10k_warn("htt rx ring size is not power of 2\n");
+		return -EINVAL;
+	}
+
+	htt->rx_ring.size_mask = htt->rx_ring.size - 1;
+
+	/*
+	 * Set the initial value for the level to which the rx ring
+	 * should be filled, based on the max throughput and the
+	 * worst likely latency for the host to fill the rx ring
+	 * with new buffers. In theory, this fill level can be
+	 * dynamically adjusted from the initial value set here, to
+	 * reflect the actual host latency rather than a
+	 * conservative assumption about the host latency.
+	 */
+	htt->rx_ring.fill_level = ath10k_htt_rx_ring_fill_level(htt);
+
+	htt->rx_ring.netbufs_ring =
+		kmalloc(htt->rx_ring.size * sizeof(struct sk_buff *),
+			GFP_KERNEL);
+	if (!htt->rx_ring.netbufs_ring)
+		goto err_netbuf;
+
+	vaddr = dma_alloc_coherent(htt->ar->dev,
+		   (htt->rx_ring.size * sizeof(htt->rx_ring.paddrs_ring)),
+		   &paddr, GFP_DMA);
+	if (!vaddr)
+		goto err_dma_ring;
+
+	htt->rx_ring.paddrs_ring = vaddr;
+	htt->rx_ring.base_paddr = paddr;
+
+	vaddr = dma_alloc_coherent(htt->ar->dev,
+				   sizeof(*htt->rx_ring.alloc_idx.vaddr),
+				   &paddr, GFP_DMA);
+	if (!vaddr)
+		goto err_dma_idx;
+
+	htt->rx_ring.alloc_idx.vaddr = vaddr;
+	htt->rx_ring.alloc_idx.paddr = paddr;
+	htt->rx_ring.sw_rd_idx.msdu_payld = 0;
+	*htt->rx_ring.alloc_idx.vaddr = 0;
+
+	/* Initialize the Rx refill retry timer */
+	setup_timer(timer, ath10k_htt_rx_ring_refill_retry, (unsigned long)htt);
+
+	spin_lock_init(&htt->rx_ring.lock);
+
+	htt->rx_ring.fill_cnt = 0;
+	if (__ath10k_htt_rx_ring_fill_n(htt, htt->rx_ring.fill_level))
+		goto err_fill_ring;
+
+	ath10k_dbg(ATH10K_DBG_HTT, "HTT RX ring size: %d, fill_level: %d\n",
+		   htt->rx_ring.size, htt->rx_ring.fill_level);
+	return 0;
+
+err_fill_ring:
+	ath10k_htt_rx_ring_free(htt);
+	dma_free_coherent(htt->ar->dev,
+			  sizeof(*htt->rx_ring.alloc_idx.vaddr),
+			  htt->rx_ring.alloc_idx.vaddr,
+			  htt->rx_ring.alloc_idx.paddr);
+err_dma_idx:
+	dma_free_coherent(htt->ar->dev,
+			  (htt->rx_ring.size *
+			   sizeof(htt->rx_ring.paddrs_ring)),
+			  htt->rx_ring.paddrs_ring,
+			  htt->rx_ring.base_paddr);
+err_dma_ring:
+	kfree(htt->rx_ring.netbufs_ring);
+err_netbuf:
+	return -ENOMEM;
+}
+
+static int ath10k_htt_rx_crypto_param_len(enum htt_rx_mpdu_encrypt_type type)
+{
+	switch (type) {
+	case HTT_RX_MPDU_ENCRYPT_WEP40:
+	case HTT_RX_MPDU_ENCRYPT_WEP104:
+		return 4;
+	case HTT_RX_MPDU_ENCRYPT_TKIP_WITHOUT_MIC:
+	case HTT_RX_MPDU_ENCRYPT_WEP128: /* not tested */
+	case HTT_RX_MPDU_ENCRYPT_TKIP_WPA:
+	case HTT_RX_MPDU_ENCRYPT_WAPI: /* not tested */
+	case HTT_RX_MPDU_ENCRYPT_AES_CCM_WPA2:
+		return 8;
+	case HTT_RX_MPDU_ENCRYPT_NONE:
+		return 0;
+	}
+
+	ath10k_warn("unknown encryption type %d\n", type);
+	return 0;
+}
+
+static int ath10k_htt_rx_crypto_tail_len(enum htt_rx_mpdu_encrypt_type type)
+{
+	switch (type) {
+	case HTT_RX_MPDU_ENCRYPT_NONE:
+	case HTT_RX_MPDU_ENCRYPT_WEP40:
+	case HTT_RX_MPDU_ENCRYPT_WEP104:
+	case HTT_RX_MPDU_ENCRYPT_WEP128:
+	case HTT_RX_MPDU_ENCRYPT_WAPI:
+		return 0;
+	case HTT_RX_MPDU_ENCRYPT_TKIP_WITHOUT_MIC:
+	case HTT_RX_MPDU_ENCRYPT_TKIP_WPA:
+		return 4;
+	case HTT_RX_MPDU_ENCRYPT_AES_CCM_WPA2:
+		return 8;
+	}
+
+	ath10k_warn("unknown encryption type %d\n", type);
+	return 0;
+}
+
+/* Applies for first msdu in chain, before altering it. */
+static struct ieee80211_hdr *ath10k_htt_rx_skb_get_hdr(struct sk_buff *skb)
+{
+	struct htt_rx_desc *rxd;
+	enum rx_msdu_decap_format fmt;
+
+	rxd = (void *)skb->data - sizeof(*rxd);
+	fmt = MS(__le32_to_cpu(rxd->msdu_start.info1),
+			RX_MSDU_START_INFO1_DECAP_FORMAT);
+
+	if (fmt == RX_MSDU_DECAP_RAW)
+		return (void *)skb->data;
+	else
+		return (void *)skb->data - RX_HTT_HDR_STATUS_LEN;
+}
+
+/* This function only applies for first msdu in an msdu chain */
+static bool ath10k_htt_rx_hdr_is_amsdu(struct ieee80211_hdr *hdr)
+{
+	if (ieee80211_is_data_qos(hdr->frame_control)) {
+		u8 *qc = ieee80211_get_qos_ctl(hdr);
+		if (qc[0] & 0x80)
+			return true;
+	}
+	return false;
+}
+
+static int ath10k_htt_rx_amsdu(struct ath10k_htt *htt,
+			struct htt_rx_info *info)
+{
+	struct htt_rx_desc *rxd;
+	struct sk_buff *amsdu;
+	struct sk_buff *first;
+	struct ieee80211_hdr *hdr;
+	struct sk_buff *skb = info->skb;
+	enum rx_msdu_decap_format fmt;
+	enum htt_rx_mpdu_encrypt_type enctype;
+	unsigned int hdr_len;
+	int crypto_len;
+
+	rxd = (void *)skb->data - sizeof(*rxd);
+	fmt = MS(__le32_to_cpu(rxd->msdu_start.info1),
+			RX_MSDU_START_INFO1_DECAP_FORMAT);
+	enctype = MS(__le32_to_cpu(rxd->mpdu_start.info0),
+			RX_MPDU_START_INFO0_ENCRYPT_TYPE);
+
+	/* FIXME: No idea what assumptions are safe here. Need logs */
+	if ((fmt == RX_MSDU_DECAP_RAW && skb->next) ||
+	    (fmt == RX_MSDU_DECAP_8023_SNAP_LLC)) {
+		ath10k_htt_rx_free_msdu_chain(skb->next);
+		skb->next = NULL;
+		return -ENOTSUPP;
+	}
+
+	/* A-MSDU max is a little less than 8K */
+	amsdu = dev_alloc_skb(8*1024);
+	if (!amsdu) {
+		ath10k_warn("A-MSDU allocation failed\n");
+		ath10k_htt_rx_free_msdu_chain(skb->next);
+		skb->next = NULL;
+		return -ENOMEM;
+	}
+
+	if (fmt >= RX_MSDU_DECAP_NATIVE_WIFI) {
+		int hdrlen;
+
+		hdr = (void *)rxd->rx_hdr_status;
+		hdrlen = ieee80211_hdrlen(hdr->frame_control);
+		memcpy(skb_put(amsdu, hdrlen), hdr, hdrlen);
+	}
+
+	first = skb;
+	while (skb) {
+		void *decap_hdr;
+		int decap_len = 0;
+
+		rxd = (void *)skb->data - sizeof(*rxd);
+		fmt = MS(__le32_to_cpu(rxd->msdu_start.info1),
+				RX_MSDU_START_INFO1_DECAP_FORMAT);
+		decap_hdr = (void *)rxd->rx_hdr_status;
+
+		if (skb == first) {
+			/* We receive linked A-MSDU subframe skbuffs. The
+			 * first one contains the original 802.11 header (and
+			 * possible crypto param) in the RX descriptor. The
+			 * A-MSDU subframe header follows that. Each part is
+			 * aligned to 4 byte boundary. */
+
+			hdr = (void *)amsdu->data;
+			hdr_len = ieee80211_hdrlen(hdr->frame_control);
+			crypto_len = ath10k_htt_rx_crypto_param_len(enctype);
+
+			decap_hdr += roundup(hdr_len, 4);
+			decap_hdr += roundup(crypto_len, 4);
+		}
+
+		if (fmt == RX_MSDU_DECAP_ETHERNET2_DIX) {
+			/* Ethernet2 decap inserts ethernet header in place of
+			 * A-MSDU subframe header. */
+			skb_pull(skb, 6 + 6 + 2);
+
+			/* A-MSDU subframe header length */
+			decap_len += 6 + 6 + 2;
+
+			/* Ethernet2 decap also strips the LLC/SNAP so we need
+			 * to re-insert it. The LLC/SNAP follows A-MSDU
+			 * subframe header. */
+			/* FIXME: Not all LLCs are 8 bytes long */
+			decap_len += 8;
+
+			memcpy(skb_put(amsdu, decap_len), decap_hdr, decap_len);
+		}
+
+		if (fmt == RX_MSDU_DECAP_NATIVE_WIFI) {
+			/* Native Wifi decap inserts regular 802.11 header
+			 * in place of A-MSDU subframe header. */
+			hdr = (struct ieee80211_hdr *)skb->data;
+			skb_pull(skb, ieee80211_hdrlen(hdr->frame_control));
+
+			/* A-MSDU subframe header length */
+			decap_len += 6 + 6 + 2;
+
+			memcpy(skb_put(amsdu, decap_len), decap_hdr, decap_len);
+		}
+
+		if (fmt == RX_MSDU_DECAP_RAW)
+			skb_trim(skb, skb->len - 4); /* remove FCS */
+
+		memcpy(skb_put(amsdu, skb->len), skb->data, skb->len);
+
+		/* A-MSDU subframes are padded to 4bytes
+		 * but relative to first subframe, not the whole MPDU */
+		if (skb->next && ((decap_len + skb->len) & 3)) {
+			int padlen = 4 - ((decap_len + skb->len) & 3);
+			memset(skb_put(amsdu, padlen), 0, padlen);
+		}
+
+		skb = skb->next;
+	}
+
+	info->skb = amsdu;
+	info->encrypt_type = enctype;
+
+	ath10k_htt_rx_free_msdu_chain(first);
+
+	return 0;
+}
+
+static int ath10k_htt_rx_msdu(struct ath10k_htt *htt, struct htt_rx_info *info)
+{
+	struct sk_buff *skb = info->skb;
+	struct htt_rx_desc *rxd;
+	struct ieee80211_hdr *hdr;
+	enum rx_msdu_decap_format fmt;
+	enum htt_rx_mpdu_encrypt_type enctype;
+
+	/* This shouldn't happen. If it does than it may be a FW bug. */
+	if (skb->next) {
+		ath10k_warn("received chained non A-MSDU frame\n");
+		ath10k_htt_rx_free_msdu_chain(skb->next);
+		skb->next = NULL;
+	}
+
+	rxd = (void *)skb->data - sizeof(*rxd);
+	fmt = MS(__le32_to_cpu(rxd->msdu_start.info1),
+			RX_MSDU_START_INFO1_DECAP_FORMAT);
+	enctype = MS(__le32_to_cpu(rxd->mpdu_start.info0),
+			RX_MPDU_START_INFO0_ENCRYPT_TYPE);
+	hdr = (void *)skb->data - RX_HTT_HDR_STATUS_LEN;
+
+	switch (fmt) {
+	case RX_MSDU_DECAP_RAW:
+		/* remove trailing FCS */
+		skb_trim(skb, skb->len - 4);
+		break;
+	case RX_MSDU_DECAP_NATIVE_WIFI:
+		/* nothing to do here */
+		break;
+	case RX_MSDU_DECAP_ETHERNET2_DIX:
+		/* macaddr[6] + macaddr[6] + ethertype[2] */
+		skb_pull(skb, 6 + 6 + 2);
+		break;
+	case RX_MSDU_DECAP_8023_SNAP_LLC:
+		/* macaddr[6] + macaddr[6] + len[2] */
+		/* we don't need this for non-A-MSDU */
+		skb_pull(skb, 6 + 6 + 2);
+		break;
+	}
+
+	if (fmt == RX_MSDU_DECAP_ETHERNET2_DIX) {
+		void *llc;
+		int llclen;
+
+		llclen = 8;
+		llc  = hdr;
+		llc += roundup(ieee80211_hdrlen(hdr->frame_control), 4);
+		llc += roundup(ath10k_htt_rx_crypto_param_len(enctype), 4);
+
+		skb_push(skb, llclen);
+		memcpy(skb->data, llc, llclen);
+	}
+
+	if (fmt >= RX_MSDU_DECAP_ETHERNET2_DIX) {
+		int len = ieee80211_hdrlen(hdr->frame_control);
+		skb_push(skb, len);
+		memcpy(skb->data, hdr, len);
+	}
+
+	info->skb = skb;
+	info->encrypt_type = enctype;
+	return 0;
+}
+
+static bool ath10k_htt_rx_has_decrypt_err(struct sk_buff *skb)
+{
+	struct htt_rx_desc *rxd;
+	u32 flags;
+
+	rxd = (void *)skb->data - sizeof(*rxd);
+	flags = __le32_to_cpu(rxd->attention.flags);
+
+	if (flags & RX_ATTENTION_FLAGS_DECRYPT_ERR)
+		return true;
+
+	return false;
+}
+
+static bool ath10k_htt_rx_has_fcs_err(struct sk_buff *skb)
+{
+	struct htt_rx_desc *rxd;
+	u32 flags;
+
+	rxd = (void *)skb->data - sizeof(*rxd);
+	flags = __le32_to_cpu(rxd->attention.flags);
+
+	if (flags & RX_ATTENTION_FLAGS_FCS_ERR)
+		return true;
+
+	return false;
+}
+
+static void ath10k_htt_rx_handler(struct ath10k_htt *htt,
+				  struct htt_rx_indication *rx)
+{
+	struct htt_rx_info info;
+	struct htt_rx_indication_mpdu_range *mpdu_ranges;
+	struct ieee80211_hdr *hdr;
+	int num_mpdu_ranges;
+	int fw_desc_len;
+	u8 *fw_desc;
+	int i, j;
+	int ret;
+
+	memset(&info, 0, sizeof(info));
+
+	fw_desc_len = __le16_to_cpu(rx->prefix.fw_rx_desc_bytes);
+	fw_desc = (u8 *)&rx->fw_desc;
+
+	num_mpdu_ranges = MS(__le32_to_cpu(rx->hdr.info1),
+			     HTT_RX_INDICATION_INFO1_NUM_MPDU_RANGES);
+	mpdu_ranges = htt_rx_ind_get_mpdu_ranges(rx);
+
+	ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt rx ind: ",
+			rx, sizeof(*rx) +
+			(sizeof(struct htt_rx_indication_mpdu_range) *
+				num_mpdu_ranges));
+
+	for (i = 0; i < num_mpdu_ranges; i++) {
+		info.status = mpdu_ranges[i].mpdu_range_status;
+
+		for (j = 0; j < mpdu_ranges[i].mpdu_count; j++) {
+			struct sk_buff *msdu_head, *msdu_tail;
+			enum htt_rx_mpdu_status status;
+			int msdu_chaining;
+
+			msdu_head = NULL;
+			msdu_tail = NULL;
+			msdu_chaining = ath10k_htt_rx_amsdu_pop(htt,
+							 &fw_desc,
+							 &fw_desc_len,
+							 &msdu_head,
+							 &msdu_tail);
+
+			if (!msdu_head) {
+				ath10k_warn("htt rx no data!\n");
+				continue;
+			}
+
+			if (msdu_head->len == 0) {
+				ath10k_dbg(ATH10K_DBG_HTT,
+					   "htt rx dropping due to zero-len\n");
+				ath10k_htt_rx_free_msdu_chain(msdu_head);
+				continue;
+			}
+
+			if (ath10k_htt_rx_has_decrypt_err(msdu_head)) {
+				ath10k_htt_rx_free_msdu_chain(msdu_head);
+				continue;
+			}
+
+			status = info.status;
+
+			/* Skip mgmt frames while we handle this in WMI */
+			if (status == HTT_RX_IND_MPDU_STATUS_MGMT_CTRL) {
+				ath10k_htt_rx_free_msdu_chain(msdu_head);
+				continue;
+			}
+
+			if (status != HTT_RX_IND_MPDU_STATUS_OK &&
+			    status != HTT_RX_IND_MPDU_STATUS_TKIP_MIC_ERR &&
+			    !htt->ar->monitor_enabled) {
+				ath10k_dbg(ATH10K_DBG_HTT,
+					   "htt rx ignoring frame w/ status %d\n",
+					   status);
+				ath10k_htt_rx_free_msdu_chain(msdu_head);
+				continue;
+			}
+
+			/* FIXME: we do not support chaining yet.
+			 * this needs investigation */
+			if (msdu_chaining) {
+				ath10k_warn("msdu_chaining is true\n");
+				ath10k_htt_rx_free_msdu_chain(msdu_head);
+				continue;
+			}
+
+			info.skb     = msdu_head;
+			info.fcs_err = ath10k_htt_rx_has_fcs_err(msdu_head);
+			info.signal  = ATH10K_DEFAULT_NOISE_FLOOR;
+			info.signal += rx->ppdu.combined_rssi;
+
+			info.rate.info0 = rx->ppdu.info0;
+			info.rate.info1 = __le32_to_cpu(rx->ppdu.info1);
+			info.rate.info2 = __le32_to_cpu(rx->ppdu.info2);
+
+			hdr = ath10k_htt_rx_skb_get_hdr(msdu_head);
+
+			if (ath10k_htt_rx_hdr_is_amsdu(hdr))
+				ret = ath10k_htt_rx_amsdu(htt, &info);
+			else
+				ret = ath10k_htt_rx_msdu(htt, &info);
+
+			if (ret && !info.fcs_err) {
+				ath10k_warn("error processing msdus %d\n", ret);
+				dev_kfree_skb_any(info.skb);
+				continue;
+			}
+
+			if (ath10k_htt_rx_hdr_is_amsdu((void *)info.skb->data))
+				ath10k_dbg(ATH10K_DBG_HTT, "htt mpdu is amsdu\n");
+
+			ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt mpdu: ",
+					info.skb->data, info.skb->len);
+			ath10k_process_rx(htt->ar, &info);
+		}
+	}
+
+	ath10k_htt_rx_msdu_buff_replenish(htt);
+}
+
+static void ath10k_htt_rx_frag_handler(struct ath10k_htt *htt,
+				struct htt_rx_fragment_indication *frag)
+{
+	struct sk_buff *msdu_head, *msdu_tail;
+	struct htt_rx_desc *rxd;
+	enum rx_msdu_decap_format fmt;
+	struct htt_rx_info info = {};
+	struct ieee80211_hdr *hdr;
+	int msdu_chaining;
+	bool tkip_mic_err;
+	bool decrypt_err;
+	u8 *fw_desc;
+	int fw_desc_len, hdrlen, paramlen;
+	int trim;
+
+	fw_desc_len = __le16_to_cpu(frag->fw_rx_desc_bytes);
+	fw_desc = (u8 *)frag->fw_msdu_rx_desc;
+
+	msdu_head = NULL;
+	msdu_tail = NULL;
+	msdu_chaining = ath10k_htt_rx_amsdu_pop(htt, &fw_desc, &fw_desc_len,
+						&msdu_head, &msdu_tail);
+
+	ath10k_dbg(ATH10K_DBG_HTT_DUMP, "htt rx frag ahead\n");
+
+	if (!msdu_head) {
+		ath10k_warn("htt rx frag no data\n");
+		return;
+	}
+
+	if (msdu_chaining || msdu_head != msdu_tail) {
+		ath10k_warn("aggregation with fragmentation?!\n");
+		ath10k_htt_rx_free_msdu_chain(msdu_head);
+		return;
+	}
+
+	/* FIXME: implement signal strength */
+
+	hdr = (struct ieee80211_hdr *)msdu_head->data;
+	rxd = (void *)msdu_head->data - sizeof(*rxd);
+	tkip_mic_err = !!(__le32_to_cpu(rxd->attention.flags) &
+				RX_ATTENTION_FLAGS_TKIP_MIC_ERR);
+	decrypt_err = !!(__le32_to_cpu(rxd->attention.flags) &
+				RX_ATTENTION_FLAGS_DECRYPT_ERR);
+	fmt = MS(__le32_to_cpu(rxd->msdu_start.info1),
+			RX_MSDU_START_INFO1_DECAP_FORMAT);
+
+	if (fmt != RX_MSDU_DECAP_RAW) {
+		ath10k_warn("we dont support non-raw fragmented rx yet\n");
+		dev_kfree_skb_any(msdu_head);
+		goto end;
+	}
+
+	info.skb = msdu_head;
+	info.status = HTT_RX_IND_MPDU_STATUS_OK;
+	info.encrypt_type = MS(__le32_to_cpu(rxd->mpdu_start.info0),
+				RX_MPDU_START_INFO0_ENCRYPT_TYPE);
+
+	if (tkip_mic_err) {
+		ath10k_warn("tkip mic error\n");
+		info.status = HTT_RX_IND_MPDU_STATUS_TKIP_MIC_ERR;
+	}
+
+	if (decrypt_err) {
+		ath10k_warn("decryption err in fragmented rx\n");
+		dev_kfree_skb_any(info.skb);
+		goto end;
+	}
+
+	if (info.encrypt_type != HTT_RX_MPDU_ENCRYPT_NONE) {
+		hdrlen = ieee80211_hdrlen(hdr->frame_control);
+		paramlen = ath10k_htt_rx_crypto_param_len(info.encrypt_type);
+
+		/* It is more efficient to move the header than the payload */
+		memmove((void *)info.skb->data + paramlen,
+			(void *)info.skb->data,
+			hdrlen);
+		skb_pull(info.skb, paramlen);
+		hdr = (struct ieee80211_hdr *)info.skb->data;
+	}
+
+	/* remove trailing FCS */
+	trim  = 4;
+
+	/* remove crypto trailer */
+	trim += ath10k_htt_rx_crypto_tail_len(info.encrypt_type);
+
+	/* last fragment of TKIP frags has MIC */
+	if (!ieee80211_has_morefrags(hdr->frame_control) &&
+	    info.encrypt_type == HTT_RX_MPDU_ENCRYPT_TKIP_WPA)
+		trim += 8;
+
+	if (trim > info.skb->len) {
+		ath10k_warn("htt rx fragment: trailer longer than the frame itself? drop\n");
+		dev_kfree_skb_any(info.skb);
+		goto end;
+	}
+
+	skb_trim(info.skb, info.skb->len - trim);
+
+	ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt frag mpdu: ",
+			info.skb->data, info.skb->len);
+	ath10k_process_rx(htt->ar, &info);
+
+end:
+	if (fw_desc_len > 0) {
+		ath10k_dbg(ATH10K_DBG_HTT,
+			   "expecting more fragmented rx in one indication %d\n",
+			   fw_desc_len);
+	}
+}
+
+void ath10k_htt_t2h_msg_handler(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct ath10k_htt *htt = ar->htt;
+	struct htt_resp *resp = (struct htt_resp *)skb->data;
+
+	/* confirm alignment */
+	if (!IS_ALIGNED((unsigned long)skb->data, 4))
+		ath10k_warn("unaligned htt message, expect trouble\n");
+
+	ath10k_dbg(ATH10K_DBG_HTT, "HTT RX, msg_type: 0x%0X\n",
+		   resp->hdr.msg_type);
+	switch (resp->hdr.msg_type) {
+	case HTT_T2H_MSG_TYPE_VERSION_CONF: {
+		htt->target_version_major = resp->ver_resp.major;
+		htt->target_version_minor = resp->ver_resp.minor;
+		complete(&htt->target_version_received);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_RX_IND: {
+		ath10k_htt_rx_handler(htt, &resp->rx_ind);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_PEER_MAP: {
+		struct htt_peer_map_event ev = {
+			.vdev_id = resp->peer_map.vdev_id,
+			.peer_id = __le16_to_cpu(resp->peer_map.peer_id),
+		};
+		memcpy(ev.addr, resp->peer_map.addr, sizeof(ev.addr));
+		ath10k_peer_map_event(htt, &ev);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_PEER_UNMAP: {
+		struct htt_peer_unmap_event ev = {
+			.peer_id = __le16_to_cpu(resp->peer_unmap.peer_id),
+		};
+		ath10k_peer_unmap_event(htt, &ev);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_MGMT_TX_COMPLETION: {
+		struct htt_tx_done tx_done = {};
+		int status = __le32_to_cpu(resp->mgmt_tx_completion.status);
+
+		tx_done.msdu_id =
+			__le32_to_cpu(resp->mgmt_tx_completion.desc_id);
+
+		switch (status) {
+		case HTT_MGMT_TX_STATUS_OK:
+			break;
+		case HTT_MGMT_TX_STATUS_RETRY:
+			tx_done.no_ack = true;
+			break;
+		case HTT_MGMT_TX_STATUS_DROP:
+			tx_done.discard = true;
+			break;
+		}
+
+		ath10k_txrx_tx_completed(htt, &tx_done);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_TX_COMPL_IND: {
+		struct htt_tx_done tx_done = {};
+		int status = MS(resp->data_tx_completion.flags,
+				HTT_DATA_TX_STATUS);
+		__le16 msdu_id;
+		int i;
+
+		switch (status) {
+		case HTT_DATA_TX_STATUS_NO_ACK:
+			tx_done.no_ack = true;
+			break;
+		case HTT_DATA_TX_STATUS_OK:
+			break;
+		case HTT_DATA_TX_STATUS_DISCARD:
+		case HTT_DATA_TX_STATUS_POSTPONE:
+		case HTT_DATA_TX_STATUS_DOWNLOAD_FAIL:
+			tx_done.discard = true;
+			break;
+		default:
+			ath10k_warn("unhandled tx completion status %d\n",
+				    status);
+			tx_done.discard = true;
+			break;
+		}
+
+		ath10k_dbg(ATH10K_DBG_HTT, "htt tx completion num_msdus %d\n",
+			   resp->data_tx_completion.num_msdus);
+
+		for (i = 0; i < resp->data_tx_completion.num_msdus; i++) {
+			msdu_id = resp->data_tx_completion.msdus[i];
+			tx_done.msdu_id = __le16_to_cpu(msdu_id);
+			ath10k_txrx_tx_completed(htt, &tx_done);
+		}
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_SEC_IND: {
+		struct ath10k *ar = htt->ar;
+		struct htt_security_indication *ev = &resp->security_indication;
+
+		ath10k_dbg(ATH10K_DBG_HTT,
+			   "sec ind peer_id %d unicast %d type %d\n",
+			  __le16_to_cpu(ev->peer_id),
+			  !!(ev->flags & HTT_SECURITY_IS_UNICAST),
+			  MS(ev->flags, HTT_SECURITY_TYPE));
+		complete(&ar->install_key_done);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_RX_FRAG_IND: {
+		ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt event: ",
+				skb->data, skb->len);
+		ath10k_htt_rx_frag_handler(htt, &resp->rx_frag_ind);
+		break;
+	}
+	case HTT_T2H_MSG_TYPE_TEST:
+		/* FIX THIS */
+		break;
+	case HTT_T2H_MSG_TYPE_TX_INSPECT_IND:
+	case HTT_T2H_MSG_TYPE_STATS_CONF:
+	case HTT_T2H_MSG_TYPE_RX_ADDBA:
+	case HTT_T2H_MSG_TYPE_RX_DELBA:
+	case HTT_T2H_MSG_TYPE_RX_FLUSH:
+	default:
+		ath10k_dbg(ATH10K_DBG_HTT, "htt event (%d) not handled\n",
+			   resp->hdr.msg_type);
+		ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "htt event: ",
+				skb->data, skb->len);
+		break;
+	};
+
+	/* Free the indication buffer */
+	dev_kfree_skb_any(skb);
+}
diff --git a/drivers/net/wireless/ath/ath10k/htt_tx.c b/drivers/net/wireless/ath/ath10k/htt_tx.c
new file mode 100644
index 0000000..ef79106
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/htt_tx.c
@@ -0,0 +1,510 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/etherdevice.h>
+#include "htt.h"
+#include "mac.h"
+#include "hif.h"
+#include "txrx.h"
+#include "debug.h"
+
+void __ath10k_htt_tx_dec_pending(struct ath10k_htt *htt)
+{
+	htt->num_pending_tx--;
+	if (htt->num_pending_tx == htt->max_num_pending_tx - 1)
+		ieee80211_wake_queues(htt->ar->hw);
+}
+
+static void ath10k_htt_tx_dec_pending(struct ath10k_htt *htt)
+{
+	spin_lock_bh(&htt->tx_lock);
+	__ath10k_htt_tx_dec_pending(htt);
+	spin_unlock_bh(&htt->tx_lock);
+}
+
+static int ath10k_htt_tx_inc_pending(struct ath10k_htt *htt)
+{
+	int ret = 0;
+
+	spin_lock_bh(&htt->tx_lock);
+
+	if (htt->num_pending_tx >= htt->max_num_pending_tx) {
+		ret = -EBUSY;
+		goto exit;
+	}
+
+	htt->num_pending_tx++;
+	if (htt->num_pending_tx == htt->max_num_pending_tx)
+		ieee80211_stop_queues(htt->ar->hw);
+
+exit:
+	spin_unlock_bh(&htt->tx_lock);
+	return ret;
+}
+
+int ath10k_htt_tx_alloc_msdu_id(struct ath10k_htt *htt)
+{
+	int msdu_id;
+
+	lockdep_assert_held(&htt->tx_lock);
+
+	msdu_id = find_first_zero_bit(htt->used_msdu_ids,
+				      htt->max_num_pending_tx);
+	if (msdu_id == htt->max_num_pending_tx)
+		return -ENOBUFS;
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt tx alloc msdu_id %d\n", msdu_id);
+	__set_bit(msdu_id, htt->used_msdu_ids);
+	return msdu_id;
+}
+
+void ath10k_htt_tx_free_msdu_id(struct ath10k_htt *htt, u16 msdu_id)
+{
+	lockdep_assert_held(&htt->tx_lock);
+
+	if (!test_bit(msdu_id, htt->used_msdu_ids))
+		ath10k_warn("trying to free unallocated msdu_id %d\n", msdu_id);
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt tx free msdu_id %hu\n", msdu_id);
+	__clear_bit(msdu_id, htt->used_msdu_ids);
+}
+
+int ath10k_htt_tx_attach(struct ath10k_htt *htt)
+{
+	u8 pipe;
+
+	spin_lock_init(&htt->tx_lock);
+	init_waitqueue_head(&htt->empty_tx_wq);
+
+	/* At the beginning free queue number should hint us the maximum
+	 * queue length */
+	pipe = htt->ar->htc->endpoint[htt->eid].ul_pipe_id;
+	htt->max_num_pending_tx = ath10k_hif_get_free_queue_number(htt->ar,
+								   pipe);
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt tx max num pending tx %d\n",
+		   htt->max_num_pending_tx);
+
+	htt->pending_tx = kzalloc(sizeof(*htt->pending_tx) *
+				  htt->max_num_pending_tx, GFP_KERNEL);
+	if (!htt->pending_tx)
+		return -ENOMEM;
+
+	htt->used_msdu_ids = kzalloc(sizeof(unsigned long) *
+				     BITS_TO_LONGS(htt->max_num_pending_tx),
+				     GFP_KERNEL);
+	if (!htt->used_msdu_ids) {
+		kfree(htt->pending_tx);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void ath10k_htt_tx_cleanup_pending(struct ath10k_htt *htt)
+{
+	struct sk_buff *txdesc;
+	int msdu_id;
+
+	/* No locks needed. Called after communication with the device has
+	 * been stopped. */
+
+	for (msdu_id = 0; msdu_id < htt->max_num_pending_tx; msdu_id++) {
+		if (!test_bit(msdu_id, htt->used_msdu_ids))
+			continue;
+
+		txdesc = htt->pending_tx[msdu_id];
+		if (!txdesc)
+			continue;
+
+		ath10k_dbg(ATH10K_DBG_HTT, "force cleanup msdu_id %hu\n",
+			   msdu_id);
+
+		if (ATH10K_SKB_CB(txdesc)->htt.refcount > 0)
+			ATH10K_SKB_CB(txdesc)->htt.refcount = 1;
+
+		ATH10K_SKB_CB(txdesc)->htt.discard = true;
+		ath10k_txrx_tx_unref(htt, txdesc);
+	}
+}
+
+void ath10k_htt_tx_detach(struct ath10k_htt *htt)
+{
+	ath10k_htt_tx_cleanup_pending(htt);
+	kfree(htt->pending_tx);
+	kfree(htt->used_msdu_ids);
+	return;
+}
+
+void ath10k_htt_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct ath10k_skb_cb *skb_cb = ATH10K_SKB_CB(skb);
+	struct ath10k_htt *htt = ar->htt;
+
+	if (skb_cb->htt.is_conf) {
+		dev_kfree_skb_any(skb);
+		return;
+	}
+
+	if (skb_cb->is_aborted) {
+		skb_cb->htt.discard = true;
+
+		/* if the skbuff is aborted we need to make sure we'll free up
+		 * the tx resources, we can't simply run tx_unref() 2 times
+		 * because if htt tx completion came in earlier we'd access
+		 * unallocated memory */
+		if (skb_cb->htt.refcount > 1)
+			skb_cb->htt.refcount = 1;
+	}
+
+	ath10k_txrx_tx_unref(htt, skb);
+}
+
+int ath10k_htt_h2t_ver_req_msg(struct ath10k_htt *htt)
+{
+	struct sk_buff *skb;
+	struct htt_cmd *cmd;
+	int len = 0;
+	int ret;
+
+	len += sizeof(cmd->hdr);
+	len += sizeof(cmd->ver_req);
+
+	skb = ath10k_htc_alloc_skb(len);
+	if (!skb)
+		return -ENOMEM;
+
+	skb_put(skb, len);
+	cmd = (struct htt_cmd *)skb->data;
+	cmd->hdr.msg_type = HTT_H2T_MSG_TYPE_VERSION_REQ;
+
+	ATH10K_SKB_CB(skb)->htt.is_conf = true;
+
+	ret = ath10k_htc_send(htt->ar->htc, htt->eid, skb);
+	if (ret) {
+		dev_kfree_skb_any(skb);
+		return ret;
+	}
+
+	return 0;
+}
+
+int ath10k_htt_send_rx_ring_cfg_ll(struct ath10k_htt *htt)
+{
+	struct sk_buff *skb;
+	struct htt_cmd *cmd;
+	struct htt_rx_ring_setup_ring *ring;
+	const int num_rx_ring = 1;
+	u16 flags;
+	u32 fw_idx;
+	int len;
+	int ret;
+
+	/*
+	 * the HW expects the buffer to be an integral number of 4-byte
+	 * "words"
+	 */
+	BUILD_BUG_ON(!IS_ALIGNED(HTT_RX_BUF_SIZE, 4));
+	BUILD_BUG_ON((HTT_RX_BUF_SIZE & HTT_MAX_CACHE_LINE_SIZE_MASK) != 0);
+
+	len = sizeof(cmd->hdr) + sizeof(cmd->rx_setup.hdr)
+	    + (sizeof(*ring) * num_rx_ring);
+	skb = ath10k_htc_alloc_skb(len);
+	if (!skb)
+		return -ENOMEM;
+
+	skb_put(skb, len);
+
+	cmd = (struct htt_cmd *)skb->data;
+	ring = &cmd->rx_setup.rings[0];
+
+	cmd->hdr.msg_type = HTT_H2T_MSG_TYPE_RX_RING_CFG;
+	cmd->rx_setup.hdr.num_rings = 1;
+
+	/* FIXME: do we need all of this? */
+	flags = 0;
+	flags |= HTT_RX_RING_FLAGS_MAC80211_HDR;
+	flags |= HTT_RX_RING_FLAGS_MSDU_PAYLOAD;
+	flags |= HTT_RX_RING_FLAGS_PPDU_START;
+	flags |= HTT_RX_RING_FLAGS_PPDU_END;
+	flags |= HTT_RX_RING_FLAGS_MPDU_START;
+	flags |= HTT_RX_RING_FLAGS_MPDU_END;
+	flags |= HTT_RX_RING_FLAGS_MSDU_START;
+	flags |= HTT_RX_RING_FLAGS_MSDU_END;
+	flags |= HTT_RX_RING_FLAGS_RX_ATTENTION;
+	flags |= HTT_RX_RING_FLAGS_FRAG_INFO;
+	flags |= HTT_RX_RING_FLAGS_UNICAST_RX;
+	flags |= HTT_RX_RING_FLAGS_MULTICAST_RX;
+	flags |= HTT_RX_RING_FLAGS_CTRL_RX;
+	flags |= HTT_RX_RING_FLAGS_MGMT_RX;
+	flags |= HTT_RX_RING_FLAGS_NULL_RX;
+	flags |= HTT_RX_RING_FLAGS_PHY_DATA_RX;
+
+	fw_idx = __le32_to_cpu(*htt->rx_ring.alloc_idx.vaddr);
+
+	ring->fw_idx_shadow_reg_paddr =
+		__cpu_to_le32(htt->rx_ring.alloc_idx.paddr);
+	ring->rx_ring_base_paddr = __cpu_to_le32(htt->rx_ring.base_paddr);
+	ring->rx_ring_len = __cpu_to_le16(htt->rx_ring.size);
+	ring->rx_ring_bufsize = __cpu_to_le16(HTT_RX_BUF_SIZE);
+	ring->flags = __cpu_to_le16(flags);
+	ring->fw_idx_init_val = __cpu_to_le16(fw_idx);
+
+#define desc_offset(x) (offsetof(struct htt_rx_desc, x) / 4)
+
+	ring->mac80211_hdr_offset = __cpu_to_le16(desc_offset(rx_hdr_status));
+	ring->msdu_payload_offset = __cpu_to_le16(desc_offset(msdu_payload));
+	ring->ppdu_start_offset = __cpu_to_le16(desc_offset(ppdu_start));
+	ring->ppdu_end_offset = __cpu_to_le16(desc_offset(ppdu_end));
+	ring->mpdu_start_offset = __cpu_to_le16(desc_offset(mpdu_start));
+	ring->mpdu_end_offset = __cpu_to_le16(desc_offset(mpdu_end));
+	ring->msdu_start_offset = __cpu_to_le16(desc_offset(msdu_start));
+	ring->msdu_end_offset = __cpu_to_le16(desc_offset(msdu_end));
+	ring->rx_attention_offset = __cpu_to_le16(desc_offset(attention));
+	ring->frag_info_offset = __cpu_to_le16(desc_offset(frag_info));
+
+#undef desc_offset
+
+	ATH10K_SKB_CB(skb)->htt.is_conf = true;
+
+	ret = ath10k_htc_send(htt->ar->htc, htt->eid, skb);
+	if (ret) {
+		dev_kfree_skb_any(skb);
+		return ret;
+	}
+
+	return 0;
+}
+
+int ath10k_htt_mgmt_tx(struct ath10k_htt *htt, struct sk_buff *msdu)
+{
+	struct device *dev = htt->ar->dev;
+	struct ath10k_skb_cb *skb_cb;
+	struct sk_buff *txdesc = NULL;
+	struct htt_cmd *cmd;
+	u8 vdev_id = ATH10K_SKB_CB(msdu)->htt.vdev_id;
+	int len = 0;
+	int msdu_id = -1;
+	int res;
+
+
+	res = ath10k_htt_tx_inc_pending(htt);
+	if (res)
+		return res;
+
+	len += sizeof(cmd->hdr);
+	len += sizeof(cmd->mgmt_tx);
+
+	txdesc = ath10k_htc_alloc_skb(len);
+	if (!txdesc) {
+		res = -ENOMEM;
+		goto err;
+	}
+
+	spin_lock_bh(&htt->tx_lock);
+	msdu_id = ath10k_htt_tx_alloc_msdu_id(htt);
+	if (msdu_id < 0) {
+		spin_unlock_bh(&htt->tx_lock);
+		res = msdu_id;
+		goto err;
+	}
+	htt->pending_tx[msdu_id] = txdesc;
+	spin_unlock_bh(&htt->tx_lock);
+
+	res = ath10k_skb_map(dev, msdu);
+	if (res)
+		goto err;
+
+	skb_put(txdesc, len);
+	cmd = (struct htt_cmd *)txdesc->data;
+	cmd->hdr.msg_type         = HTT_H2T_MSG_TYPE_MGMT_TX;
+	cmd->mgmt_tx.msdu_paddr = __cpu_to_le32(ATH10K_SKB_CB(msdu)->paddr);
+	cmd->mgmt_tx.len        = __cpu_to_le32(msdu->len);
+	cmd->mgmt_tx.desc_id    = __cpu_to_le32(msdu_id);
+	cmd->mgmt_tx.vdev_id    = __cpu_to_le32(vdev_id);
+	memcpy(cmd->mgmt_tx.hdr, msdu->data,
+	       min_t(int, msdu->len, HTT_MGMT_FRM_HDR_DOWNLOAD_LEN));
+
+	/* refcount is decremented by HTC and HTT completions until it reaches
+	 * zero and is freed */
+	skb_cb = ATH10K_SKB_CB(txdesc);
+	skb_cb->htt.msdu_id = msdu_id;
+	skb_cb->htt.refcount = 2;
+	skb_cb->htt.msdu = msdu;
+
+	res = ath10k_htc_send(htt->ar->htc, htt->eid, txdesc);
+	if (res)
+		goto err;
+
+	return 0;
+
+err:
+	ath10k_skb_unmap(dev, msdu);
+
+	if (txdesc)
+		dev_kfree_skb_any(txdesc);
+	if (msdu_id >= 0) {
+		spin_lock_bh(&htt->tx_lock);
+		htt->pending_tx[msdu_id] = NULL;
+		ath10k_htt_tx_free_msdu_id(htt, msdu_id);
+		spin_unlock_bh(&htt->tx_lock);
+	}
+	ath10k_htt_tx_dec_pending(htt);
+	return res;
+}
+
+int ath10k_htt_tx(struct ath10k_htt *htt, struct sk_buff *msdu)
+{
+	struct device *dev = htt->ar->dev;
+	struct htt_cmd *cmd;
+	struct htt_data_tx_desc_frag *tx_frags;
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)msdu->data;
+	struct ath10k_skb_cb *skb_cb;
+	struct sk_buff *txdesc = NULL;
+	struct sk_buff *txfrag = NULL;
+	u8 vdev_id = ATH10K_SKB_CB(msdu)->htt.vdev_id;
+	u8 tid;
+	int prefetch_len, desc_len, frag_len;
+	dma_addr_t frags_paddr;
+	int msdu_id = -1;
+	int res;
+	u8 flags0;
+	u16 flags1;
+
+	res = ath10k_htt_tx_inc_pending(htt);
+	if (res)
+		return res;
+
+	prefetch_len = min(htt->prefetch_len, msdu->len);
+	prefetch_len = roundup(prefetch_len, 4);
+
+	desc_len = sizeof(cmd->hdr) + sizeof(cmd->data_tx) + prefetch_len;
+	frag_len = sizeof(*tx_frags) * 2;
+
+	txdesc = ath10k_htc_alloc_skb(desc_len);
+	if (!txdesc) {
+		res = -ENOMEM;
+		goto err;
+	}
+
+	txfrag = dev_alloc_skb(frag_len);
+	if (!txfrag) {
+		res = -ENOMEM;
+		goto err;
+	}
+
+	if (!IS_ALIGNED((unsigned long)txdesc->data, 4)) {
+		ath10k_warn("htt alignment check failed. dropping packet.\n");
+		res = -EIO;
+		goto err;
+	}
+
+	spin_lock_bh(&htt->tx_lock);
+	msdu_id = ath10k_htt_tx_alloc_msdu_id(htt);
+	if (msdu_id < 0) {
+		spin_unlock_bh(&htt->tx_lock);
+		res = msdu_id;
+		goto err;
+	}
+	htt->pending_tx[msdu_id] = txdesc;
+	spin_unlock_bh(&htt->tx_lock);
+
+	res = ath10k_skb_map(dev, msdu);
+	if (res)
+		goto err;
+
+	/* tx fragment list must be terminated with zero-entry */
+	skb_put(txfrag, frag_len);
+	tx_frags = (struct htt_data_tx_desc_frag *)txfrag->data;
+	tx_frags[0].paddr = __cpu_to_le32(ATH10K_SKB_CB(msdu)->paddr);
+	tx_frags[0].len   = __cpu_to_le32(msdu->len);
+	tx_frags[1].paddr = __cpu_to_le32(0);
+	tx_frags[1].len   = __cpu_to_le32(0);
+
+	res = ath10k_skb_map(dev, txfrag);
+	if (res)
+		goto err;
+
+	ath10k_dbg(ATH10K_DBG_HTT, "txfrag 0x%llx msdu 0x%llx\n",
+		   (unsigned long long) ATH10K_SKB_CB(txfrag)->paddr,
+		   (unsigned long long) ATH10K_SKB_CB(msdu)->paddr);
+	ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "txfrag: ",
+			txfrag->data, frag_len);
+	ath10k_dbg_dump(ATH10K_DBG_HTT_DUMP, NULL, "msdu: ",
+			msdu->data, msdu->len);
+
+	skb_put(txdesc, desc_len);
+	cmd = (struct htt_cmd *)txdesc->data;
+	memset(cmd, 0, desc_len);
+
+	tid = ATH10K_SKB_CB(msdu)->htt.tid;
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt data tx using tid %hhu\n", tid);
+
+	flags0  = 0;
+	if (!ieee80211_has_protected(hdr->frame_control))
+		flags0 |= HTT_DATA_TX_DESC_FLAGS0_NO_ENCRYPT;
+	flags0 |= HTT_DATA_TX_DESC_FLAGS0_MAC_HDR_PRESENT;
+	flags0 |= SM(ATH10K_HW_TXRX_NATIVE_WIFI,
+		     HTT_DATA_TX_DESC_FLAGS0_PKT_TYPE);
+
+	flags1  = 0;
+	flags1 |= SM((u16)vdev_id, HTT_DATA_TX_DESC_FLAGS1_VDEV_ID);
+	flags1 |= SM((u16)tid, HTT_DATA_TX_DESC_FLAGS1_EXT_TID);
+
+	frags_paddr = ATH10K_SKB_CB(txfrag)->paddr;
+
+	cmd->hdr.msg_type        = HTT_H2T_MSG_TYPE_TX_FRM;
+	cmd->data_tx.flags0      = flags0;
+	cmd->data_tx.flags1      = __cpu_to_le16(flags1);
+	cmd->data_tx.len         = __cpu_to_le16(msdu->len);
+	cmd->data_tx.id          = __cpu_to_le16(msdu_id);
+	cmd->data_tx.frags_paddr = __cpu_to_le32(frags_paddr);
+	cmd->data_tx.peerid      = __cpu_to_le32(HTT_INVALID_PEERID);
+
+	memcpy(cmd->data_tx.prefetch, msdu->data, prefetch_len);
+
+	/* refcount is decremented by HTC and HTT completions until it reaches
+	 * zero and is freed */
+	skb_cb = ATH10K_SKB_CB(txdesc);
+	skb_cb->htt.msdu_id = msdu_id;
+	skb_cb->htt.refcount = 2;
+	skb_cb->htt.txfrag = txfrag;
+	skb_cb->htt.msdu = msdu;
+
+	res = ath10k_htc_send(htt->ar->htc, htt->eid, txdesc);
+	if (res)
+		goto err;
+
+	return 0;
+err:
+	if (txfrag)
+		ath10k_skb_unmap(dev, txfrag);
+	if (txdesc)
+		dev_kfree_skb_any(txdesc);
+	if (txfrag)
+		dev_kfree_skb_any(txfrag);
+	if (msdu_id >= 0) {
+		spin_lock_bh(&htt->tx_lock);
+		htt->pending_tx[msdu_id] = NULL;
+		ath10k_htt_tx_free_msdu_id(htt, msdu_id);
+		spin_unlock_bh(&htt->tx_lock);
+	}
+	ath10k_htt_tx_dec_pending(htt);
+	ath10k_skb_unmap(dev, msdu);
+	return res;
+}
diff --git a/drivers/net/wireless/ath/ath10k/hw.h b/drivers/net/wireless/ath/ath10k/hw.h
new file mode 100644
index 0000000..44ed5af
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/hw.h
@@ -0,0 +1,304 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _HW_H_
+#define _HW_H_
+
+#include "targaddrs.h"
+
+/* Supported FW version */
+#define SUPPORTED_FW_MAJOR	1
+#define SUPPORTED_FW_MINOR	0
+#define SUPPORTED_FW_RELEASE	0
+#define SUPPORTED_FW_BUILD	629
+
+/* QCA988X 1.0 definitions */
+#define QCA988X_HW_1_0_VERSION		0x4000002c
+#define QCA988X_HW_1_0_FW_DIR		"ath10k/QCA988X/hw1.0"
+#define QCA988X_HW_1_0_FW_FILE		"firmware.bin"
+#define QCA988X_HW_1_0_OTP_FILE		"otp.bin"
+#define QCA988X_HW_1_0_BOARD_DATA_FILE	"board.bin"
+#define QCA988X_HW_1_0_PATCH_LOAD_ADDR	0x1234
+
+/* QCA988X 2.0 definitions */
+#define QCA988X_HW_2_0_VERSION		0x4100016c
+#define QCA988X_HW_2_0_FW_DIR		"ath10k/QCA988X/hw2.0"
+#define QCA988X_HW_2_0_FW_FILE		"firmware.bin"
+#define QCA988X_HW_2_0_OTP_FILE		"otp.bin"
+#define QCA988X_HW_2_0_BOARD_DATA_FILE	"board.bin"
+#define QCA988X_HW_2_0_PATCH_LOAD_ADDR	0x1234
+
+/* Known pecularities:
+ *  - current FW doesn't support raw rx mode (last tested v599)
+ *  - current FW dumps upon raw tx mode (last tested v599)
+ *  - raw appears in nwifi decap, raw and nwifi appear in ethernet decap
+ *  - raw have FCS, nwifi doesn't
+ *  - ethernet frames have 802.11 header decapped and parts (base hdr, cipher
+ *    param, llc/snap) are aligned to 4byte boundaries each */
+enum ath10k_hw_txrx_mode {
+	ATH10K_HW_TXRX_RAW = 0,
+	ATH10K_HW_TXRX_NATIVE_WIFI = 1,
+	ATH10K_HW_TXRX_ETHERNET = 2,
+};
+
+enum ath10k_mcast2ucast_mode {
+	ATH10K_MCAST2UCAST_DISABLED = 0,
+	ATH10K_MCAST2UCAST_ENABLED = 1,
+};
+
+#define TARGET_NUM_VDEVS			8
+#define TARGET_NUM_PEER_AST			2
+#define TARGET_NUM_WDS_ENTRIES			32
+#define TARGET_DMA_BURST_SIZE			0
+#define TARGET_MAC_AGGR_DELIM			0
+#define TARGET_AST_SKID_LIMIT			16
+#define TARGET_NUM_PEERS			16
+#define TARGET_NUM_OFFLOAD_PEERS		0
+#define TARGET_NUM_OFFLOAD_REORDER_BUFS         0
+#define TARGET_NUM_PEER_KEYS			2
+#define TARGET_NUM_TIDS		(2 * ((TARGET_NUM_PEERS) + (TARGET_NUM_VDEVS)))
+#define TARGET_TX_CHAIN_MASK			(BIT(0) | BIT(1) | BIT(2))
+#define TARGET_RX_CHAIN_MASK			(BIT(0) | BIT(1) | BIT(2))
+#define TARGET_RX_TIMEOUT_LO_PRI		100
+#define TARGET_RX_TIMEOUT_HI_PRI		40
+#define TARGET_RX_DECAP_MODE			ATH10K_HW_TXRX_ETHERNET
+#define TARGET_SCAN_MAX_PENDING_REQS		4
+#define TARGET_BMISS_OFFLOAD_MAX_VDEV		3
+#define TARGET_ROAM_OFFLOAD_MAX_VDEV		3
+#define TARGET_ROAM_OFFLOAD_MAX_AP_PROFILES	8
+#define TARGET_GTK_OFFLOAD_MAX_VDEV		3
+#define TARGET_NUM_MCAST_GROUPS			0
+#define TARGET_NUM_MCAST_TABLE_ELEMS		0
+#define TARGET_MCAST2UCAST_MODE			ATH10K_MCAST2UCAST_DISABLED
+#define TARGET_TX_DBG_LOG_SIZE			1024
+#define TARGET_RX_SKIP_DEFRAG_TIMEOUT_DUP_DETECTION_CHECK 0
+#define TARGET_VOW_CONFIG			0
+#define TARGET_NUM_MSDU_DESC			(1024 + 400)
+#define TARGET_MAX_FRAG_ENTRIES			0
+
+
+/* Number of Copy Engines supported */
+#define CE_COUNT 8
+
+/*
+ * Total number of PCIe MSI interrupts requested for all interrupt sources.
+ * PCIe standard forces this to be a power of 2.
+ * Some Host OS's limit MSI requests that can be granted to 8
+ * so for now we abide by this limit and avoid requesting more
+ * than that.
+ */
+#define MSI_NUM_REQUEST_LOG2	3
+#define MSI_NUM_REQUEST		(1<<MSI_NUM_REQUEST_LOG2)
+
+/*
+ * Granted MSIs are assigned as follows:
+ * Firmware uses the first
+ * Remaining MSIs, if any, are used by Copy Engines
+ * This mapping is known to both Target firmware and Host software.
+ * It may be changed as long as Host and Target are kept in sync.
+ */
+/* MSI for firmware (errors, etc.) */
+#define MSI_ASSIGN_FW		0
+
+/* MSIs for Copy Engines */
+#define MSI_ASSIGN_CE_INITIAL	1
+#define MSI_ASSIGN_CE_MAX	7
+
+/* as of IP3.7.1 */
+#define RTC_STATE_V_ON				3
+
+#define RTC_STATE_COLD_RESET_MASK		0x00000400
+#define RTC_STATE_V_LSB				0
+#define RTC_STATE_V_MASK			0x00000007
+#define RTC_STATE_ADDRESS			0x0000
+#define PCIE_SOC_WAKE_V_MASK			0x00000001
+#define PCIE_SOC_WAKE_ADDRESS			0x0004
+#define PCIE_SOC_WAKE_RESET			0x00000000
+#define SOC_GLOBAL_RESET_ADDRESS		0x0008
+
+#define RTC_SOC_BASE_ADDRESS			0x00004000
+#define RTC_WMAC_BASE_ADDRESS			0x00005000
+#define MAC_COEX_BASE_ADDRESS			0x00006000
+#define BT_COEX_BASE_ADDRESS			0x00007000
+#define SOC_PCIE_BASE_ADDRESS			0x00008000
+#define SOC_CORE_BASE_ADDRESS			0x00009000
+#define WLAN_UART_BASE_ADDRESS			0x0000c000
+#define WLAN_SI_BASE_ADDRESS			0x00010000
+#define WLAN_GPIO_BASE_ADDRESS			0x00014000
+#define WLAN_ANALOG_INTF_BASE_ADDRESS		0x0001c000
+#define WLAN_MAC_BASE_ADDRESS			0x00020000
+#define EFUSE_BASE_ADDRESS			0x00030000
+#define FPGA_REG_BASE_ADDRESS			0x00039000
+#define WLAN_UART2_BASE_ADDRESS			0x00054c00
+#define CE_WRAPPER_BASE_ADDRESS			0x00057000
+#define CE0_BASE_ADDRESS			0x00057400
+#define CE1_BASE_ADDRESS			0x00057800
+#define CE2_BASE_ADDRESS			0x00057c00
+#define CE3_BASE_ADDRESS			0x00058000
+#define CE4_BASE_ADDRESS			0x00058400
+#define CE5_BASE_ADDRESS			0x00058800
+#define CE6_BASE_ADDRESS			0x00058c00
+#define CE7_BASE_ADDRESS			0x00059000
+#define DBI_BASE_ADDRESS			0x00060000
+#define WLAN_ANALOG_INTF_PCIE_BASE_ADDRESS	0x0006c000
+#define PCIE_LOCAL_BASE_ADDRESS			0x00080000
+
+#define SOC_RESET_CONTROL_OFFSET		0x00000000
+#define SOC_RESET_CONTROL_SI0_RST_MASK		0x00000001
+#define SOC_CPU_CLOCK_OFFSET			0x00000020
+#define SOC_CPU_CLOCK_STANDARD_LSB		0
+#define SOC_CPU_CLOCK_STANDARD_MASK		0x00000003
+#define SOC_CLOCK_CONTROL_OFFSET		0x00000028
+#define SOC_CLOCK_CONTROL_SI0_CLK_MASK		0x00000001
+#define SOC_SYSTEM_SLEEP_OFFSET			0x000000c4
+#define SOC_LPO_CAL_OFFSET			0x000000e0
+#define SOC_LPO_CAL_ENABLE_LSB			20
+#define SOC_LPO_CAL_ENABLE_MASK			0x00100000
+
+#define WLAN_RESET_CONTROL_COLD_RST_MASK	0x00000008
+#define WLAN_RESET_CONTROL_WARM_RST_MASK	0x00000004
+#define WLAN_SYSTEM_SLEEP_DISABLE_LSB		0
+#define WLAN_SYSTEM_SLEEP_DISABLE_MASK		0x00000001
+
+#define WLAN_GPIO_PIN0_ADDRESS			0x00000028
+#define WLAN_GPIO_PIN0_CONFIG_MASK		0x00007800
+#define WLAN_GPIO_PIN1_ADDRESS			0x0000002c
+#define WLAN_GPIO_PIN1_CONFIG_MASK		0x00007800
+#define WLAN_GPIO_PIN10_ADDRESS			0x00000050
+#define WLAN_GPIO_PIN11_ADDRESS			0x00000054
+#define WLAN_GPIO_PIN12_ADDRESS			0x00000058
+#define WLAN_GPIO_PIN13_ADDRESS			0x0000005c
+
+#define CLOCK_GPIO_OFFSET			0xffffffff
+#define CLOCK_GPIO_BT_CLK_OUT_EN_LSB		0
+#define CLOCK_GPIO_BT_CLK_OUT_EN_MASK		0
+
+#define SI_CONFIG_OFFSET			0x00000000
+#define SI_CONFIG_BIDIR_OD_DATA_LSB		18
+#define SI_CONFIG_BIDIR_OD_DATA_MASK		0x00040000
+#define SI_CONFIG_I2C_LSB			16
+#define SI_CONFIG_I2C_MASK			0x00010000
+#define SI_CONFIG_POS_SAMPLE_LSB		7
+#define SI_CONFIG_POS_SAMPLE_MASK		0x00000080
+#define SI_CONFIG_INACTIVE_DATA_LSB		5
+#define SI_CONFIG_INACTIVE_DATA_MASK		0x00000020
+#define SI_CONFIG_INACTIVE_CLK_LSB		4
+#define SI_CONFIG_INACTIVE_CLK_MASK		0x00000010
+#define SI_CONFIG_DIVIDER_LSB			0
+#define SI_CONFIG_DIVIDER_MASK			0x0000000f
+#define SI_CS_OFFSET				0x00000004
+#define SI_CS_DONE_ERR_MASK			0x00000400
+#define SI_CS_DONE_INT_MASK			0x00000200
+#define SI_CS_START_LSB				8
+#define SI_CS_START_MASK			0x00000100
+#define SI_CS_RX_CNT_LSB			4
+#define SI_CS_RX_CNT_MASK			0x000000f0
+#define SI_CS_TX_CNT_LSB			0
+#define SI_CS_TX_CNT_MASK			0x0000000f
+
+#define SI_TX_DATA0_OFFSET			0x00000008
+#define SI_TX_DATA1_OFFSET			0x0000000c
+#define SI_RX_DATA0_OFFSET			0x00000010
+#define SI_RX_DATA1_OFFSET			0x00000014
+
+#define CORE_CTRL_CPU_INTR_MASK			0x00002000
+#define CORE_CTRL_ADDRESS			0x0000
+#define PCIE_INTR_ENABLE_ADDRESS		0x0008
+#define PCIE_INTR_CLR_ADDRESS			0x0014
+#define SCRATCH_3_ADDRESS			0x0030
+
+/* Firmware indications to the Host via SCRATCH_3 register. */
+#define FW_INDICATOR_ADDRESS	(SOC_CORE_BASE_ADDRESS + SCRATCH_3_ADDRESS)
+#define FW_IND_EVENT_PENDING			1
+#define FW_IND_INITIALIZED			2
+
+/* HOST_REG interrupt from firmware */
+#define PCIE_INTR_FIRMWARE_MASK			0x00000400
+#define PCIE_INTR_CE_MASK_ALL			0x0007f800
+
+#define DRAM_BASE_ADDRESS			0x00400000
+
+#define MISSING 0
+
+#define SYSTEM_SLEEP_OFFSET			SOC_SYSTEM_SLEEP_OFFSET
+#define WLAN_SYSTEM_SLEEP_OFFSET		SOC_SYSTEM_SLEEP_OFFSET
+#define WLAN_RESET_CONTROL_OFFSET		SOC_RESET_CONTROL_OFFSET
+#define CLOCK_CONTROL_OFFSET			SOC_CLOCK_CONTROL_OFFSET
+#define CLOCK_CONTROL_SI0_CLK_MASK		SOC_CLOCK_CONTROL_SI0_CLK_MASK
+#define RESET_CONTROL_MBOX_RST_MASK		MISSING
+#define RESET_CONTROL_SI0_RST_MASK		SOC_RESET_CONTROL_SI0_RST_MASK
+#define GPIO_BASE_ADDRESS			WLAN_GPIO_BASE_ADDRESS
+#define GPIO_PIN0_OFFSET			WLAN_GPIO_PIN0_ADDRESS
+#define GPIO_PIN1_OFFSET			WLAN_GPIO_PIN1_ADDRESS
+#define GPIO_PIN0_CONFIG_MASK			WLAN_GPIO_PIN0_CONFIG_MASK
+#define GPIO_PIN1_CONFIG_MASK			WLAN_GPIO_PIN1_CONFIG_MASK
+#define SI_BASE_ADDRESS				WLAN_SI_BASE_ADDRESS
+#define SCRATCH_BASE_ADDRESS			SOC_CORE_BASE_ADDRESS
+#define LOCAL_SCRATCH_OFFSET			0x18
+#define CPU_CLOCK_OFFSET			SOC_CPU_CLOCK_OFFSET
+#define LPO_CAL_OFFSET				SOC_LPO_CAL_OFFSET
+#define GPIO_PIN10_OFFSET			WLAN_GPIO_PIN10_ADDRESS
+#define GPIO_PIN11_OFFSET			WLAN_GPIO_PIN11_ADDRESS
+#define GPIO_PIN12_OFFSET			WLAN_GPIO_PIN12_ADDRESS
+#define GPIO_PIN13_OFFSET			WLAN_GPIO_PIN13_ADDRESS
+#define CPU_CLOCK_STANDARD_LSB			SOC_CPU_CLOCK_STANDARD_LSB
+#define CPU_CLOCK_STANDARD_MASK			SOC_CPU_CLOCK_STANDARD_MASK
+#define LPO_CAL_ENABLE_LSB			SOC_LPO_CAL_ENABLE_LSB
+#define LPO_CAL_ENABLE_MASK			SOC_LPO_CAL_ENABLE_MASK
+#define ANALOG_INTF_BASE_ADDRESS		WLAN_ANALOG_INTF_BASE_ADDRESS
+#define MBOX_BASE_ADDRESS			MISSING
+#define INT_STATUS_ENABLE_ERROR_LSB		MISSING
+#define INT_STATUS_ENABLE_ERROR_MASK		MISSING
+#define INT_STATUS_ENABLE_CPU_LSB		MISSING
+#define INT_STATUS_ENABLE_CPU_MASK		MISSING
+#define INT_STATUS_ENABLE_COUNTER_LSB		MISSING
+#define INT_STATUS_ENABLE_COUNTER_MASK		MISSING
+#define INT_STATUS_ENABLE_MBOX_DATA_LSB		MISSING
+#define INT_STATUS_ENABLE_MBOX_DATA_MASK	MISSING
+#define ERROR_STATUS_ENABLE_RX_UNDERFLOW_LSB	MISSING
+#define ERROR_STATUS_ENABLE_RX_UNDERFLOW_MASK	MISSING
+#define ERROR_STATUS_ENABLE_TX_OVERFLOW_LSB	MISSING
+#define ERROR_STATUS_ENABLE_TX_OVERFLOW_MASK	MISSING
+#define COUNTER_INT_STATUS_ENABLE_BIT_LSB	MISSING
+#define COUNTER_INT_STATUS_ENABLE_BIT_MASK	MISSING
+#define INT_STATUS_ENABLE_ADDRESS		MISSING
+#define CPU_INT_STATUS_ENABLE_BIT_LSB		MISSING
+#define CPU_INT_STATUS_ENABLE_BIT_MASK		MISSING
+#define HOST_INT_STATUS_ADDRESS			MISSING
+#define CPU_INT_STATUS_ADDRESS			MISSING
+#define ERROR_INT_STATUS_ADDRESS		MISSING
+#define ERROR_INT_STATUS_WAKEUP_MASK		MISSING
+#define ERROR_INT_STATUS_WAKEUP_LSB		MISSING
+#define ERROR_INT_STATUS_RX_UNDERFLOW_MASK	MISSING
+#define ERROR_INT_STATUS_RX_UNDERFLOW_LSB	MISSING
+#define ERROR_INT_STATUS_TX_OVERFLOW_MASK	MISSING
+#define ERROR_INT_STATUS_TX_OVERFLOW_LSB	MISSING
+#define COUNT_DEC_ADDRESS			MISSING
+#define HOST_INT_STATUS_CPU_MASK		MISSING
+#define HOST_INT_STATUS_CPU_LSB			MISSING
+#define HOST_INT_STATUS_ERROR_MASK		MISSING
+#define HOST_INT_STATUS_ERROR_LSB		MISSING
+#define HOST_INT_STATUS_COUNTER_MASK		MISSING
+#define HOST_INT_STATUS_COUNTER_LSB		MISSING
+#define RX_LOOKAHEAD_VALID_ADDRESS		MISSING
+#define WINDOW_DATA_ADDRESS			MISSING
+#define WINDOW_READ_ADDR_ADDRESS		MISSING
+#define WINDOW_WRITE_ADDR_ADDRESS		MISSING
+
+#define RTC_STATE_V_GET(x) (((x) & RTC_STATE_V_MASK) >> RTC_STATE_V_LSB)
+
+#endif /* _HW_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/mac.c b/drivers/net/wireless/ath/ath10k/mac.c
new file mode 100644
index 0000000..da5c333
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/mac.c
@@ -0,0 +1,3069 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "mac.h"
+
+#include <net/mac80211.h>
+#include <linux/etherdevice.h>
+
+#include "core.h"
+#include "debug.h"
+#include "wmi.h"
+#include "htt.h"
+#include "txrx.h"
+
+/**********/
+/* Crypto */
+/**********/
+
+static int ath10k_send_key(struct ath10k_vif *arvif,
+			   struct ieee80211_key_conf *key,
+			   enum set_key_cmd cmd,
+			   const u8 *macaddr)
+{
+	struct wmi_vdev_install_key_arg arg = {
+		.vdev_id = arvif->vdev_id,
+		.key_idx = key->keyidx,
+		.key_len = key->keylen,
+		.key_data = key->key,
+		.macaddr = macaddr,
+	};
+
+	if (key->flags & IEEE80211_KEY_FLAG_PAIRWISE)
+		arg.key_flags = WMI_KEY_PAIRWISE;
+	else
+		arg.key_flags = WMI_KEY_GROUP;
+
+	switch (key->cipher) {
+	case WLAN_CIPHER_SUITE_CCMP:
+		arg.key_cipher = WMI_CIPHER_AES_CCM;
+		key->flags |= IEEE80211_KEY_FLAG_SW_MGMT_TX;
+		break;
+	case WLAN_CIPHER_SUITE_TKIP:
+		arg.key_cipher = WMI_CIPHER_TKIP;
+		arg.key_txmic_len = 8;
+		arg.key_rxmic_len = 8;
+		break;
+	case WLAN_CIPHER_SUITE_WEP40:
+	case WLAN_CIPHER_SUITE_WEP104:
+		arg.key_cipher = WMI_CIPHER_WEP;
+		/* AP/IBSS mode requires self-key to be groupwise
+		 * Otherwise pairwise key must be set */
+		if (memcmp(macaddr, arvif->vif->addr, ETH_ALEN))
+			arg.key_flags = WMI_KEY_PAIRWISE;
+		break;
+	default:
+		ath10k_warn("cipher %d is not supported\n", key->cipher);
+		return -EOPNOTSUPP;
+	}
+
+	if (cmd == DISABLE_KEY) {
+		arg.key_cipher = WMI_CIPHER_NONE;
+		arg.key_data = NULL;
+	}
+
+	return ath10k_wmi_vdev_install_key(arvif->ar, &arg);
+}
+
+static int ath10k_install_key(struct ath10k_vif *arvif,
+			      struct ieee80211_key_conf *key,
+			      enum set_key_cmd cmd,
+			      const u8 *macaddr)
+{
+	struct ath10k *ar = arvif->ar;
+	int ret;
+
+	INIT_COMPLETION(ar->install_key_done);
+
+	ret = ath10k_send_key(arvif, key, cmd, macaddr);
+	if (ret)
+		return ret;
+
+	ret = wait_for_completion_timeout(&ar->install_key_done, 3*HZ);
+	if (ret == 0)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static int ath10k_install_peer_wep_keys(struct ath10k_vif *arvif,
+					const u8 *addr)
+{
+	struct ath10k *ar = arvif->ar;
+	struct ath10k_peer *peer;
+	int ret;
+	int i;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	spin_lock_bh(&ar->data_lock);
+	peer = ath10k_peer_find(ar, arvif->vdev_id, addr);
+	spin_unlock_bh(&ar->data_lock);
+
+	if (!peer)
+		return -ENOENT;
+
+	for (i = 0; i < ARRAY_SIZE(arvif->wep_keys); i++) {
+		if (arvif->wep_keys[i] == NULL)
+			continue;
+
+		ret = ath10k_install_key(arvif, arvif->wep_keys[i], SET_KEY,
+					 addr);
+		if (ret)
+			return ret;
+
+		peer->keys[i] = arvif->wep_keys[i];
+	}
+
+	return 0;
+}
+
+static int ath10k_clear_peer_keys(struct ath10k_vif *arvif,
+				  const u8 *addr)
+{
+	struct ath10k *ar = arvif->ar;
+	struct ath10k_peer *peer;
+	int first_errno = 0;
+	int ret;
+	int i;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	spin_lock_bh(&ar->data_lock);
+	peer = ath10k_peer_find(ar, arvif->vdev_id, addr);
+	spin_unlock_bh(&ar->data_lock);
+
+	if (!peer)
+		return -ENOENT;
+
+	for (i = 0; i < ARRAY_SIZE(peer->keys); i++) {
+		if (peer->keys[i] == NULL)
+			continue;
+
+		ret = ath10k_install_key(arvif, peer->keys[i],
+					 DISABLE_KEY, addr);
+		if (ret && first_errno == 0)
+			first_errno = ret;
+
+		if (ret)
+			ath10k_warn("could not remove peer wep key %d (%d)\n",
+				    i, ret);
+
+		peer->keys[i] = NULL;
+	}
+
+	return first_errno;
+}
+
+static int ath10k_clear_vdev_key(struct ath10k_vif *arvif,
+				 struct ieee80211_key_conf *key)
+{
+	struct ath10k *ar = arvif->ar;
+	struct ath10k_peer *peer;
+	u8 addr[ETH_ALEN];
+	int first_errno = 0;
+	int ret;
+	int i;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	for (;;) {
+		/* since ath10k_install_key we can't hold data_lock all the
+		 * time, so we try to remove the keys incrementally */
+		spin_lock_bh(&ar->data_lock);
+		i = 0;
+		list_for_each_entry(peer, &ar->peers, list) {
+			for (i = 0; i < ARRAY_SIZE(peer->keys); i++) {
+				if (peer->keys[i] == key) {
+					memcpy(addr, peer->addr, ETH_ALEN);
+					peer->keys[i] = NULL;
+					break;
+				}
+			}
+
+			if (i < ARRAY_SIZE(peer->keys))
+				break;
+		}
+		spin_unlock_bh(&ar->data_lock);
+
+		if (i == ARRAY_SIZE(peer->keys))
+			break;
+
+		ret = ath10k_install_key(arvif, key, DISABLE_KEY, addr);
+		if (ret && first_errno == 0)
+			first_errno = ret;
+
+		if (ret)
+			ath10k_warn("could not remove key for %pM\n", addr);
+	}
+
+	return first_errno;
+}
+
+
+/*********************/
+/* General utilities */
+/*********************/
+
+static inline enum wmi_phy_mode
+chan_to_phymode(const struct cfg80211_chan_def *chandef)
+{
+	enum wmi_phy_mode phymode = MODE_UNKNOWN;
+
+	switch (chandef->chan->band) {
+	case IEEE80211_BAND_2GHZ:
+		switch (chandef->width) {
+		case NL80211_CHAN_WIDTH_20_NOHT:
+			phymode = MODE_11G;
+			break;
+		case NL80211_CHAN_WIDTH_20:
+			phymode = MODE_11NG_HT20;
+			break;
+		case NL80211_CHAN_WIDTH_40:
+			phymode = MODE_11NG_HT40;
+			break;
+		case NL80211_CHAN_WIDTH_5:
+		case NL80211_CHAN_WIDTH_10:
+		case NL80211_CHAN_WIDTH_80:
+		case NL80211_CHAN_WIDTH_80P80:
+		case NL80211_CHAN_WIDTH_160:
+			phymode = MODE_UNKNOWN;
+			break;
+		}
+		break;
+	case IEEE80211_BAND_5GHZ:
+		switch (chandef->width) {
+		case NL80211_CHAN_WIDTH_20_NOHT:
+			phymode = MODE_11A;
+			break;
+		case NL80211_CHAN_WIDTH_20:
+			phymode = MODE_11NA_HT20;
+			break;
+		case NL80211_CHAN_WIDTH_40:
+			phymode = MODE_11NA_HT40;
+			break;
+		case NL80211_CHAN_WIDTH_80:
+			phymode = MODE_11AC_VHT80;
+			break;
+		case NL80211_CHAN_WIDTH_5:
+		case NL80211_CHAN_WIDTH_10:
+		case NL80211_CHAN_WIDTH_80P80:
+		case NL80211_CHAN_WIDTH_160:
+			phymode = MODE_UNKNOWN;
+			break;
+		}
+		break;
+	default:
+		break;
+	}
+
+	WARN_ON(phymode == MODE_UNKNOWN);
+	return phymode;
+}
+
+static u8 ath10k_parse_mpdudensity(u8 mpdudensity)
+{
+/*
+ * 802.11n D2.0 defined values for "Minimum MPDU Start Spacing":
+ *   0 for no restriction
+ *   1 for 1/4 us
+ *   2 for 1/2 us
+ *   3 for 1 us
+ *   4 for 2 us
+ *   5 for 4 us
+ *   6 for 8 us
+ *   7 for 16 us
+ */
+	switch (mpdudensity) {
+	case 0:
+		return 0;
+	case 1:
+	case 2:
+	case 3:
+	/* Our lower layer calculations limit our precision to
+	   1 microsecond */
+		return 1;
+	case 4:
+		return 2;
+	case 5:
+		return 4;
+	case 6:
+		return 8;
+	case 7:
+		return 16;
+	default:
+		return 0;
+	}
+}
+
+static int ath10k_peer_create(struct ath10k *ar, u32 vdev_id, const u8 *addr)
+{
+	int ret;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	ret = ath10k_wmi_peer_create(ar, vdev_id, addr);
+	if (ret)
+		return ret;
+
+	ret = ath10k_wait_for_peer_created(ar, vdev_id, addr);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int ath10k_peer_delete(struct ath10k *ar, u32 vdev_id, const u8 *addr)
+{
+	int ret;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	ret = ath10k_wmi_peer_delete(ar, vdev_id, addr);
+	if (ret)
+		return ret;
+
+	ret = ath10k_wait_for_peer_deleted(ar, vdev_id, addr);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void ath10k_peer_cleanup(struct ath10k *ar, u32 vdev_id)
+{
+	struct ath10k_peer *peer, *tmp;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	spin_lock_bh(&ar->data_lock);
+	list_for_each_entry_safe(peer, tmp, &ar->peers, list) {
+		if (peer->vdev_id != vdev_id)
+			continue;
+
+		ath10k_warn("removing stale peer %pM from vdev_id %d\n",
+			    peer->addr, vdev_id);
+
+		list_del(&peer->list);
+		kfree(peer);
+	}
+	spin_unlock_bh(&ar->data_lock);
+}
+
+/************************/
+/* Interface management */
+/************************/
+
+static inline int ath10k_vdev_setup_sync(struct ath10k *ar)
+{
+	int ret;
+
+	ret = wait_for_completion_timeout(&ar->vdev_setup_done,
+					  ATH10K_VDEV_SETUP_TIMEOUT_HZ);
+	if (ret == 0)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static int ath10k_vdev_start(struct ath10k_vif *arvif)
+{
+	struct ath10k *ar = arvif->ar;
+	struct ieee80211_conf *conf = &ar->hw->conf;
+	struct ieee80211_channel *channel = conf->chandef.chan;
+	struct wmi_vdev_start_request_arg arg = {};
+	int ret = 0;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	INIT_COMPLETION(ar->vdev_setup_done);
+
+	arg.vdev_id = arvif->vdev_id;
+	arg.dtim_period = arvif->dtim_period;
+	arg.bcn_intval = arvif->beacon_interval;
+
+	arg.channel.freq = channel->center_freq;
+
+	arg.channel.band_center_freq1 = conf->chandef.center_freq1;
+
+	arg.channel.mode = chan_to_phymode(&conf->chandef);
+
+	arg.channel.min_power = channel->max_power * 3;
+	arg.channel.max_power = channel->max_power * 4;
+	arg.channel.max_reg_power = channel->max_reg_power * 4;
+	arg.channel.max_antenna_gain = channel->max_antenna_gain;
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_AP) {
+		arg.ssid = arvif->u.ap.ssid;
+		arg.ssid_len = arvif->u.ap.ssid_len;
+		arg.hidden_ssid = arvif->u.ap.hidden_ssid;
+	} else if (arvif->vdev_type == WMI_VDEV_TYPE_IBSS) {
+		arg.ssid = arvif->vif->bss_conf.ssid;
+		arg.ssid_len = arvif->vif->bss_conf.ssid_len;
+	}
+
+	ret = ath10k_wmi_vdev_start(ar, &arg);
+	if (ret) {
+		ath10k_warn("WMI vdev start failed: ret %d\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_vdev_setup_sync(ar);
+	if (ret) {
+		ath10k_warn("vdev setup failed %d\n", ret);
+		return ret;
+	}
+
+	return ret;
+}
+
+static int ath10k_vdev_stop(struct ath10k_vif *arvif)
+{
+	struct ath10k *ar = arvif->ar;
+	int ret;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	INIT_COMPLETION(ar->vdev_setup_done);
+
+	ret = ath10k_wmi_vdev_stop(ar, arvif->vdev_id);
+	if (ret) {
+		ath10k_warn("WMI vdev stop failed: ret %d\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_vdev_setup_sync(ar);
+	if (ret) {
+		ath10k_warn("vdev setup failed %d\n", ret);
+		return ret;
+	}
+
+	return ret;
+}
+
+static int ath10k_monitor_start(struct ath10k *ar, int vdev_id)
+{
+	struct ieee80211_channel *channel = ar->hw->conf.chandef.chan;
+	struct wmi_vdev_start_request_arg arg = {};
+	enum nl80211_channel_type type;
+	int ret = 0;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	type = cfg80211_get_chandef_type(&ar->hw->conf.chandef);
+
+	arg.vdev_id = vdev_id;
+	arg.channel.freq = channel->center_freq;
+	arg.channel.band_center_freq1 = ar->hw->conf.chandef.center_freq1;
+
+	/* TODO setup this dynamically, what in case we
+	   don't have any vifs? */
+	arg.channel.mode = chan_to_phymode(&ar->hw->conf.chandef);
+
+	arg.channel.min_power = channel->max_power * 3;
+	arg.channel.max_power = channel->max_power * 4;
+	arg.channel.max_reg_power = channel->max_reg_power * 4;
+	arg.channel.max_antenna_gain = channel->max_antenna_gain;
+
+	ret = ath10k_wmi_vdev_start(ar, &arg);
+	if (ret) {
+		ath10k_warn("Monitor vdev start failed: ret %d\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_vdev_setup_sync(ar);
+	if (ret) {
+		ath10k_warn("Monitor vdev setup failed %d\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_wmi_vdev_up(ar, vdev_id, 0, ar->mac_addr);
+	if (ret) {
+		ath10k_warn("Monitor vdev up failed: %d\n", ret);
+		goto vdev_stop;
+	}
+
+	ar->monitor_vdev_id = vdev_id;
+	ar->monitor_enabled = true;
+
+	return 0;
+
+vdev_stop:
+	ret = ath10k_wmi_vdev_stop(ar, ar->monitor_vdev_id);
+	if (ret)
+		ath10k_warn("Monitor vdev stop failed: %d\n", ret);
+
+	return ret;
+}
+
+static int ath10k_monitor_stop(struct ath10k *ar)
+{
+	int ret = 0;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	/* For some reasons, ath10k_wmi_vdev_down() here couse
+	 * often ath10k_wmi_vdev_stop() to fail. Next we could
+	 * not run monitor vdev and driver reload
+	 * required. Don't see such problems we skip
+	 * ath10k_wmi_vdev_down() here.
+	 */
+
+	ret = ath10k_wmi_vdev_stop(ar, ar->monitor_vdev_id);
+	if (ret)
+		ath10k_warn("Monitor vdev stop failed: %d\n", ret);
+
+	ret = ath10k_vdev_setup_sync(ar);
+	if (ret)
+		ath10k_warn("Monitor_down sync failed: %d\n", ret);
+
+	ar->monitor_enabled = false;
+	return ret;
+}
+
+static int ath10k_monitor_create(struct ath10k *ar)
+{
+	int bit, ret = 0;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	if (ar->monitor_present) {
+		ath10k_warn("Monitor mode already enabled\n");
+		return 0;
+	}
+
+	bit = ffs(ar->free_vdev_map);
+	if (bit == 0) {
+		ath10k_warn("No free VDEV slots\n");
+		return -ENOMEM;
+	}
+
+	ar->monitor_vdev_id = bit - 1;
+	ar->free_vdev_map &= ~(1 << ar->monitor_vdev_id);
+
+	ret = ath10k_wmi_vdev_create(ar, ar->monitor_vdev_id,
+				     WMI_VDEV_TYPE_MONITOR,
+				     0, ar->mac_addr);
+	if (ret) {
+		ath10k_warn("WMI vdev monitor create failed: ret %d\n", ret);
+		goto vdev_fail;
+	}
+
+	ath10k_dbg(ATH10K_DBG_MAC, "Monitor interface created, vdev id: %d\n",
+		   ar->monitor_vdev_id);
+
+	ar->monitor_present = true;
+	return 0;
+
+vdev_fail:
+	/*
+	 * Restore the ID to the global map.
+	 */
+	ar->free_vdev_map |= 1 << (ar->monitor_vdev_id);
+	return ret;
+}
+
+static int ath10k_monitor_destroy(struct ath10k *ar)
+{
+	int ret = 0;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	if (!ar->monitor_present)
+		return 0;
+
+	ret = ath10k_wmi_vdev_delete(ar, ar->monitor_vdev_id);
+	if (ret) {
+		ath10k_warn("WMI vdev monitor delete failed: %d\n", ret);
+		return ret;
+	}
+
+	ar->free_vdev_map |= 1 << (ar->monitor_vdev_id);
+	ar->monitor_present = false;
+
+	ath10k_dbg(ATH10K_DBG_MAC, "Monitor interface destroyed, vdev id: %d\n",
+		   ar->monitor_vdev_id);
+	return ret;
+}
+
+static void ath10k_control_beaconing(struct ath10k_vif *arvif,
+				struct ieee80211_bss_conf *info)
+{
+	int ret = 0;
+
+	if (!info->enable_beacon) {
+		ath10k_vdev_stop(arvif);
+		return;
+	}
+
+	arvif->tx_seq_no = 0x1000;
+
+	ret = ath10k_vdev_start(arvif);
+	if (ret)
+		return;
+
+	ret = ath10k_wmi_vdev_up(arvif->ar, arvif->vdev_id, 0, info->bssid);
+	if (ret) {
+		ath10k_warn("Failed to bring up VDEV: %d\n",
+			    arvif->vdev_id);
+		return;
+	}
+	ath10k_dbg(ATH10K_DBG_MAC, "VDEV: %d up\n", arvif->vdev_id);
+}
+
+static void ath10k_control_ibss(struct ath10k_vif *arvif,
+				struct ieee80211_bss_conf *info,
+				const u8 self_peer[ETH_ALEN])
+{
+	int ret = 0;
+
+	if (!info->ibss_joined) {
+		ret = ath10k_peer_delete(arvif->ar, arvif->vdev_id, self_peer);
+		if (ret)
+			ath10k_warn("Failed to delete IBSS self peer:%pM for VDEV:%d ret:%d\n",
+				    self_peer, arvif->vdev_id, ret);
+
+		if (is_zero_ether_addr(arvif->u.ibss.bssid))
+			return;
+
+		ret = ath10k_peer_delete(arvif->ar, arvif->vdev_id,
+					 arvif->u.ibss.bssid);
+		if (ret) {
+			ath10k_warn("Failed to delete IBSS BSSID peer:%pM for VDEV:%d ret:%d\n",
+				    arvif->u.ibss.bssid, arvif->vdev_id, ret);
+			return;
+		}
+
+		memset(arvif->u.ibss.bssid, 0, ETH_ALEN);
+
+		return;
+	}
+
+	ret = ath10k_peer_create(arvif->ar, arvif->vdev_id, self_peer);
+	if (ret) {
+		ath10k_warn("Failed to create IBSS self peer:%pM for VDEV:%d ret:%d\n",
+			    self_peer, arvif->vdev_id, ret);
+		return;
+	}
+
+	ret = ath10k_wmi_vdev_set_param(arvif->ar, arvif->vdev_id,
+					WMI_VDEV_PARAM_ATIM_WINDOW,
+					ATH10K_DEFAULT_ATIM);
+	if (ret)
+		ath10k_warn("Failed to set IBSS ATIM for VDEV:%d ret:%d\n",
+			    arvif->vdev_id, ret);
+}
+
+/*
+ * Review this when mac80211 gains per-interface powersave support.
+ */
+static void ath10k_ps_iter(void *data, u8 *mac, struct ieee80211_vif *vif)
+{
+	struct ath10k_generic_iter *ar_iter = data;
+	struct ieee80211_conf *conf = &ar_iter->ar->hw->conf;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	enum wmi_sta_powersave_param param;
+	enum wmi_sta_ps_mode psmode;
+	int ret;
+
+	if (vif->type != NL80211_IFTYPE_STATION)
+		return;
+
+	if (conf->flags & IEEE80211_CONF_PS) {
+		psmode = WMI_STA_PS_MODE_ENABLED;
+		param = WMI_STA_PS_PARAM_INACTIVITY_TIME;
+
+		ret = ath10k_wmi_set_sta_ps_param(ar_iter->ar,
+						  arvif->vdev_id,
+						  param,
+						  conf->dynamic_ps_timeout);
+		if (ret) {
+			ath10k_warn("Failed to set inactivity time for VDEV: %d\n",
+				    arvif->vdev_id);
+			return;
+		}
+
+		ar_iter->ret = ret;
+	} else {
+		psmode = WMI_STA_PS_MODE_DISABLED;
+	}
+
+	ar_iter->ret = ath10k_wmi_set_psmode(ar_iter->ar, arvif->vdev_id,
+					     psmode);
+	if (ar_iter->ret)
+		ath10k_warn("Failed to set PS Mode: %d for VDEV: %d\n",
+			    psmode, arvif->vdev_id);
+	else
+		ath10k_dbg(ATH10K_DBG_MAC, "Set PS Mode: %d for VDEV: %d\n",
+			   psmode, arvif->vdev_id);
+}
+
+/**********************/
+/* Station management */
+/**********************/
+
+static void ath10k_peer_assoc_h_basic(struct ath10k *ar,
+				      struct ath10k_vif *arvif,
+				      struct ieee80211_sta *sta,
+				      struct ieee80211_bss_conf *bss_conf,
+				      struct wmi_peer_assoc_complete_arg *arg)
+{
+	memcpy(arg->addr, sta->addr, ETH_ALEN);
+	arg->vdev_id = arvif->vdev_id;
+	arg->peer_aid = sta->aid;
+	arg->peer_flags |= WMI_PEER_AUTH;
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_STA)
+		/*
+		 * Seems FW have problems with Power Save in STA
+		 * mode when we setup this parameter to high (eg. 5).
+		 * Often we see that FW don't send NULL (with clean P flags)
+		 * frame even there is info about buffered frames in beacons.
+		 * Sometimes we have to wait more than 10 seconds before FW
+		 * will wakeup. Often sending one ping from AP to our device
+		 * just fail (more than 50%).
+		 *
+		 * Seems setting this FW parameter to 1 couse FW
+		 * will check every beacon and will wakup immediately
+		 * after detection buffered data.
+		 */
+		arg->peer_listen_intval = 1;
+	else
+		arg->peer_listen_intval = ar->hw->conf.listen_interval;
+
+	arg->peer_num_spatial_streams = 1;
+
+	/*
+	 * The assoc capabilities are available only in managed mode.
+	 */
+	if (arvif->vdev_type == WMI_VDEV_TYPE_STA && bss_conf)
+		arg->peer_caps = bss_conf->assoc_capability;
+}
+
+static void ath10k_peer_assoc_h_crypto(struct ath10k *ar,
+				       struct ath10k_vif *arvif,
+				       struct wmi_peer_assoc_complete_arg *arg)
+{
+	struct ieee80211_vif *vif = arvif->vif;
+	struct ieee80211_bss_conf *info = &vif->bss_conf;
+	struct cfg80211_bss *bss;
+	const u8 *rsnie = NULL;
+	const u8 *wpaie = NULL;
+
+	bss = cfg80211_get_bss(ar->hw->wiphy, ar->hw->conf.chandef.chan,
+			       info->bssid, NULL, 0, 0, 0);
+	if (bss) {
+		const struct cfg80211_bss_ies *ies;
+
+		rcu_read_lock();
+		rsnie = ieee80211_bss_get_ie(bss, WLAN_EID_RSN);
+
+		ies = rcu_dereference(bss->ies);
+
+		wpaie = cfg80211_find_vendor_ie(WLAN_OUI_MICROSOFT,
+				WLAN_OUI_TYPE_MICROSOFT_WPA,
+				ies->data,
+				ies->len);
+		rcu_read_unlock();
+		cfg80211_put_bss(ar->hw->wiphy, bss);
+	}
+
+	/* FIXME: base on RSN IE/WPA IE is a correct idea? */
+	if (rsnie || wpaie) {
+		ath10k_dbg(ATH10K_DBG_WMI, "%s: rsn ie found\n", __func__);
+		arg->peer_flags |= WMI_PEER_NEED_PTK_4_WAY;
+	}
+
+	if (wpaie) {
+		ath10k_dbg(ATH10K_DBG_WMI, "%s: wpa ie found\n", __func__);
+		arg->peer_flags |= WMI_PEER_NEED_GTK_2_WAY;
+	}
+}
+
+static void ath10k_peer_assoc_h_rates(struct ath10k *ar,
+				      struct ieee80211_sta *sta,
+				      struct wmi_peer_assoc_complete_arg *arg)
+{
+	struct wmi_rate_set_arg *rateset = &arg->peer_legacy_rates;
+	const struct ieee80211_supported_band *sband;
+	const struct ieee80211_rate *rates;
+	u32 ratemask;
+	int i;
+
+	sband = ar->hw->wiphy->bands[ar->hw->conf.chandef.chan->band];
+	ratemask = sta->supp_rates[ar->hw->conf.chandef.chan->band];
+	rates = sband->bitrates;
+
+	rateset->num_rates = 0;
+
+	for (i = 0; i < 32; i++, ratemask >>= 1, rates++) {
+		if (!(ratemask & 1))
+			continue;
+
+		rateset->rates[rateset->num_rates] = rates->hw_value;
+		rateset->num_rates++;
+	}
+}
+
+static void ath10k_peer_assoc_h_ht(struct ath10k *ar,
+				   struct ieee80211_sta *sta,
+				   struct wmi_peer_assoc_complete_arg *arg)
+{
+	const struct ieee80211_sta_ht_cap *ht_cap = &sta->ht_cap;
+	int smps;
+	int i, n;
+
+	if (!ht_cap->ht_supported)
+		return;
+
+	arg->peer_flags |= WMI_PEER_HT;
+	arg->peer_max_mpdu = (1 << (IEEE80211_HT_MAX_AMPDU_FACTOR +
+				    ht_cap->ampdu_factor)) - 1;
+
+	arg->peer_mpdu_density =
+		ath10k_parse_mpdudensity(ht_cap->ampdu_density);
+
+	arg->peer_ht_caps = ht_cap->cap;
+	arg->peer_rate_caps |= WMI_RC_HT_FLAG;
+
+	if (ht_cap->cap & IEEE80211_HT_CAP_LDPC_CODING)
+		arg->peer_flags |= WMI_PEER_LDPC;
+
+	if (sta->bandwidth >= IEEE80211_STA_RX_BW_40) {
+		arg->peer_flags |= WMI_PEER_40MHZ;
+		arg->peer_rate_caps |= WMI_RC_CW40_FLAG;
+	}
+
+	if (ht_cap->cap & IEEE80211_HT_CAP_SGI_20)
+		arg->peer_rate_caps |= WMI_RC_SGI_FLAG;
+
+	if (ht_cap->cap & IEEE80211_HT_CAP_SGI_40)
+		arg->peer_rate_caps |= WMI_RC_SGI_FLAG;
+
+	if (ht_cap->cap & IEEE80211_HT_CAP_TX_STBC) {
+		arg->peer_rate_caps |= WMI_RC_TX_STBC_FLAG;
+		arg->peer_flags |= WMI_PEER_STBC;
+	}
+
+	if (ht_cap->cap & IEEE80211_HT_CAP_RX_STBC) {
+		u32 stbc;
+		stbc = ht_cap->cap & IEEE80211_HT_CAP_RX_STBC;
+		stbc = stbc >> IEEE80211_HT_CAP_RX_STBC_SHIFT;
+		stbc = stbc << WMI_RC_RX_STBC_FLAG_S;
+		arg->peer_rate_caps |= stbc;
+		arg->peer_flags |= WMI_PEER_STBC;
+	}
+
+	smps = ht_cap->cap & IEEE80211_HT_CAP_SM_PS;
+	smps >>= IEEE80211_HT_CAP_SM_PS_SHIFT;
+
+	if (smps == WLAN_HT_CAP_SM_PS_STATIC) {
+		arg->peer_flags |= WMI_PEER_SPATIAL_MUX;
+		arg->peer_flags |= WMI_PEER_STATIC_MIMOPS;
+	} else if (smps == WLAN_HT_CAP_SM_PS_DYNAMIC) {
+		arg->peer_flags |= WMI_PEER_SPATIAL_MUX;
+		arg->peer_flags |= WMI_PEER_DYN_MIMOPS;
+	}
+
+	if (ht_cap->mcs.rx_mask[1] && ht_cap->mcs.rx_mask[2])
+		arg->peer_rate_caps |= WMI_RC_TS_FLAG;
+	else if (ht_cap->mcs.rx_mask[1])
+		arg->peer_rate_caps |= WMI_RC_DS_FLAG;
+
+	for (i = 0, n = 0; i < IEEE80211_HT_MCS_MASK_LEN*8; i++)
+		if (ht_cap->mcs.rx_mask[i/8] & (1 << i%8))
+			arg->peer_ht_rates.rates[n++] = i;
+
+	arg->peer_ht_rates.num_rates = n;
+	arg->peer_num_spatial_streams = max((n+7) / 8, 1);
+
+	ath10k_dbg(ATH10K_DBG_MAC, "mcs cnt %d nss %d\n",
+		   arg->peer_ht_rates.num_rates,
+		   arg->peer_num_spatial_streams);
+}
+
+static void ath10k_peer_assoc_h_qos_ap(struct ath10k *ar,
+				       struct ath10k_vif *arvif,
+				       struct ieee80211_sta *sta,
+				       struct ieee80211_bss_conf *bss_conf,
+				       struct wmi_peer_assoc_complete_arg *arg)
+{
+	u32 uapsd = 0;
+	u32 max_sp = 0;
+
+	if (sta->wme)
+		arg->peer_flags |= WMI_PEER_QOS;
+
+	if (sta->wme && sta->uapsd_queues) {
+		ath10k_dbg(ATH10K_DBG_MAC, "uapsd_queues: 0x%X, max_sp: %d\n",
+			   sta->uapsd_queues, sta->max_sp);
+
+		arg->peer_flags |= WMI_PEER_APSD;
+		arg->peer_flags |= WMI_RC_UAPSD_FLAG;
+
+		if (sta->uapsd_queues & IEEE80211_WMM_IE_STA_QOSINFO_AC_VO)
+			uapsd |= WMI_AP_PS_UAPSD_AC3_DELIVERY_EN |
+				 WMI_AP_PS_UAPSD_AC3_TRIGGER_EN;
+		if (sta->uapsd_queues & IEEE80211_WMM_IE_STA_QOSINFO_AC_VI)
+			uapsd |= WMI_AP_PS_UAPSD_AC2_DELIVERY_EN |
+				 WMI_AP_PS_UAPSD_AC2_TRIGGER_EN;
+		if (sta->uapsd_queues & IEEE80211_WMM_IE_STA_QOSINFO_AC_BK)
+			uapsd |= WMI_AP_PS_UAPSD_AC1_DELIVERY_EN |
+				 WMI_AP_PS_UAPSD_AC1_TRIGGER_EN;
+		if (sta->uapsd_queues & IEEE80211_WMM_IE_STA_QOSINFO_AC_BE)
+			uapsd |= WMI_AP_PS_UAPSD_AC0_DELIVERY_EN |
+				 WMI_AP_PS_UAPSD_AC0_TRIGGER_EN;
+
+
+		if (sta->max_sp < MAX_WMI_AP_PS_PEER_PARAM_MAX_SP)
+			max_sp = sta->max_sp;
+
+		ath10k_wmi_set_ap_ps_param(ar, arvif->vdev_id,
+					   sta->addr,
+					   WMI_AP_PS_PEER_PARAM_UAPSD,
+					   uapsd);
+
+		ath10k_wmi_set_ap_ps_param(ar, arvif->vdev_id,
+					   sta->addr,
+					   WMI_AP_PS_PEER_PARAM_MAX_SP,
+					   max_sp);
+
+		/* TODO setup this based on STA listen interval and
+		   beacon interval. Currently we don't know
+		   sta->listen_interval - mac80211 patch required.
+		   Currently use 10 seconds */
+		ath10k_wmi_set_ap_ps_param(ar, arvif->vdev_id,
+					   sta->addr,
+					   WMI_AP_PS_PEER_PARAM_AGEOUT_TIME,
+					   10);
+	}
+}
+
+static void ath10k_peer_assoc_h_qos_sta(struct ath10k *ar,
+					struct ath10k_vif *arvif,
+					struct ieee80211_sta *sta,
+					struct ieee80211_bss_conf *bss_conf,
+					struct wmi_peer_assoc_complete_arg *arg)
+{
+	if (bss_conf->qos)
+		arg->peer_flags |= WMI_PEER_QOS;
+}
+
+static void ath10k_peer_assoc_h_vht(struct ath10k *ar,
+				    struct ieee80211_sta *sta,
+				    struct wmi_peer_assoc_complete_arg *arg)
+{
+	const struct ieee80211_sta_vht_cap *vht_cap = &sta->vht_cap;
+
+	if (!vht_cap->vht_supported)
+		return;
+
+	arg->peer_flags |= WMI_PEER_VHT;
+
+	arg->peer_vht_caps = vht_cap->cap;
+
+	if (sta->bandwidth == IEEE80211_STA_RX_BW_80)
+		arg->peer_flags |= WMI_PEER_80MHZ;
+
+	arg->peer_vht_rates.rx_max_rate =
+		__le16_to_cpu(vht_cap->vht_mcs.rx_highest);
+	arg->peer_vht_rates.rx_mcs_set =
+		__le16_to_cpu(vht_cap->vht_mcs.rx_mcs_map);
+	arg->peer_vht_rates.tx_max_rate =
+		__le16_to_cpu(vht_cap->vht_mcs.tx_highest);
+	arg->peer_vht_rates.tx_mcs_set =
+		__le16_to_cpu(vht_cap->vht_mcs.tx_mcs_map);
+
+	ath10k_dbg(ATH10K_DBG_MAC, "mac vht peer\n");
+}
+
+static void ath10k_peer_assoc_h_qos(struct ath10k *ar,
+				    struct ath10k_vif *arvif,
+				    struct ieee80211_sta *sta,
+				    struct ieee80211_bss_conf *bss_conf,
+				    struct wmi_peer_assoc_complete_arg *arg)
+{
+	switch (arvif->vdev_type) {
+	case WMI_VDEV_TYPE_AP:
+		ath10k_peer_assoc_h_qos_ap(ar, arvif, sta, bss_conf, arg);
+		break;
+	case WMI_VDEV_TYPE_STA:
+		ath10k_peer_assoc_h_qos_sta(ar, arvif, sta, bss_conf, arg);
+		break;
+	default:
+		break;
+	}
+}
+
+static void ath10k_peer_assoc_h_phymode(struct ath10k *ar,
+					struct ath10k_vif *arvif,
+					struct ieee80211_sta *sta,
+					struct wmi_peer_assoc_complete_arg *arg)
+{
+	enum wmi_phy_mode phymode = MODE_UNKNOWN;
+
+	/* FIXME: add VHT */
+
+	switch (ar->hw->conf.chandef.chan->band) {
+	case IEEE80211_BAND_2GHZ:
+		if (sta->ht_cap.ht_supported) {
+			if (sta->bandwidth == IEEE80211_STA_RX_BW_40)
+				phymode = MODE_11NG_HT40;
+			else
+				phymode = MODE_11NG_HT20;
+		} else {
+			phymode = MODE_11G;
+		}
+
+		break;
+	case IEEE80211_BAND_5GHZ:
+		if (sta->ht_cap.ht_supported) {
+			if (sta->bandwidth == IEEE80211_STA_RX_BW_40)
+				phymode = MODE_11NA_HT40;
+			else
+				phymode = MODE_11NA_HT20;
+		} else {
+			phymode = MODE_11A;
+		}
+
+		break;
+	default:
+		break;
+	}
+
+	arg->peer_phymode = phymode;
+	WARN_ON(phymode == MODE_UNKNOWN);
+}
+
+static int ath10k_peer_assoc(struct ath10k *ar,
+			     struct ath10k_vif *arvif,
+			     struct ieee80211_sta *sta,
+			     struct ieee80211_bss_conf *bss_conf)
+{
+	struct wmi_peer_assoc_complete_arg arg;
+
+	memset(&arg, 0, sizeof(struct wmi_peer_assoc_complete_arg));
+
+	ath10k_peer_assoc_h_basic(ar, arvif, sta, bss_conf, &arg);
+	ath10k_peer_assoc_h_crypto(ar, arvif, &arg);
+	ath10k_peer_assoc_h_rates(ar, sta, &arg);
+	ath10k_peer_assoc_h_ht(ar, sta, &arg);
+	ath10k_peer_assoc_h_vht(ar, sta, &arg);
+	ath10k_peer_assoc_h_qos(ar, arvif, sta, bss_conf, &arg);
+	ath10k_peer_assoc_h_phymode(ar, arvif, sta, &arg);
+
+	return ath10k_wmi_peer_assoc(ar, &arg);
+}
+
+/* can be called only in mac80211 callbacks due to `key_count` usage */
+static void ath10k_bss_assoc(struct ieee80211_hw *hw,
+			     struct ieee80211_vif *vif,
+			     struct ieee80211_bss_conf *bss_conf)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	struct ieee80211_sta *ap_sta;
+	int ret;
+
+	rcu_read_lock();
+
+	ap_sta = ieee80211_find_sta(vif, bss_conf->bssid);
+	if (!ap_sta) {
+		ath10k_warn("Failed to find station entry for %pM\n",
+			    bss_conf->bssid);
+		rcu_read_unlock();
+		return;
+	}
+
+	ret = ath10k_peer_assoc(ar, arvif, ap_sta, bss_conf);
+	if (ret) {
+		ath10k_warn("Peer assoc failed for %pM\n", bss_conf->bssid);
+		rcu_read_unlock();
+		return;
+	}
+
+	rcu_read_unlock();
+
+	ret = ath10k_wmi_vdev_up(ar, arvif->vdev_id, bss_conf->aid,
+				 bss_conf->bssid);
+	if (ret)
+		ath10k_warn("VDEV: %d up failed: ret %d\n",
+			    arvif->vdev_id, ret);
+	else
+		ath10k_dbg(ATH10K_DBG_MAC,
+			   "VDEV: %d associated, BSSID: %pM, AID: %d\n",
+			   arvif->vdev_id, bss_conf->bssid, bss_conf->aid);
+}
+
+/*
+ * FIXME: flush TIDs
+ */
+static void ath10k_bss_disassoc(struct ieee80211_hw *hw,
+				struct ieee80211_vif *vif)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	int ret;
+
+	/*
+	 * For some reason, calling VDEV-DOWN before VDEV-STOP
+	 * makes the FW to send frames via HTT after disassociation.
+	 * No idea why this happens, even though VDEV-DOWN is supposed
+	 * to be analogous to link down, so just stop the VDEV.
+	 */
+	ret = ath10k_vdev_stop(arvif);
+	if (!ret)
+		ath10k_dbg(ATH10K_DBG_MAC, "VDEV: %d stopped\n",
+			   arvif->vdev_id);
+
+	/*
+	 * If we don't call VDEV-DOWN after VDEV-STOP FW will remain active and
+	 * report beacons from previously associated network through HTT.
+	 * This in turn would spam mac80211 WARN_ON if we bring down all
+	 * interfaces as it expects there is no rx when no interface is
+	 * running.
+	 */
+	ret = ath10k_wmi_vdev_down(ar, arvif->vdev_id);
+	if (ret)
+		ath10k_dbg(ATH10K_DBG_MAC, "VDEV: %d ath10k_wmi_vdev_down failed (%d)\n",
+			   arvif->vdev_id, ret);
+
+	ath10k_wmi_flush_tx(ar);
+
+	arvif->def_wep_key_index = 0;
+}
+
+static int ath10k_station_assoc(struct ath10k *ar, struct ath10k_vif *arvif,
+				struct ieee80211_sta *sta)
+{
+	int ret = 0;
+
+	ret = ath10k_peer_assoc(ar, arvif, sta, NULL);
+	if (ret) {
+		ath10k_warn("WMI peer assoc failed for %pM\n", sta->addr);
+		return ret;
+	}
+
+	ret = ath10k_install_peer_wep_keys(arvif, sta->addr);
+	if (ret) {
+		ath10k_warn("could not install peer wep keys (%d)\n", ret);
+		return ret;
+	}
+
+	return ret;
+}
+
+static int ath10k_station_disassoc(struct ath10k *ar, struct ath10k_vif *arvif,
+				   struct ieee80211_sta *sta)
+{
+	int ret = 0;
+
+	ret = ath10k_clear_peer_keys(arvif, sta->addr);
+	if (ret) {
+		ath10k_warn("could not clear all peer wep keys (%d)\n", ret);
+		return ret;
+	}
+
+	return ret;
+}
+
+/**************/
+/* Regulatory */
+/**************/
+
+static int ath10k_update_channel_list(struct ath10k *ar)
+{
+	struct ieee80211_hw *hw = ar->hw;
+	struct ieee80211_supported_band **bands;
+	enum ieee80211_band band;
+	struct ieee80211_channel *channel;
+	struct wmi_scan_chan_list_arg arg = {0};
+	struct wmi_channel_arg *ch;
+	bool passive;
+	int len;
+	int ret;
+	int i;
+
+	bands = hw->wiphy->bands;
+	for (band = 0; band < IEEE80211_NUM_BANDS; band++) {
+		if (!bands[band])
+			continue;
+
+		for (i = 0; i < bands[band]->n_channels; i++) {
+			if (bands[band]->channels[i].flags &
+			    IEEE80211_CHAN_DISABLED)
+				continue;
+
+			arg.n_channels++;
+		}
+	}
+
+	len = sizeof(struct wmi_channel_arg) * arg.n_channels;
+	arg.channels = kzalloc(len, GFP_KERNEL);
+	if (!arg.channels)
+		return -ENOMEM;
+
+	ch = arg.channels;
+	for (band = 0; band < IEEE80211_NUM_BANDS; band++) {
+		if (!bands[band])
+			continue;
+
+		for (i = 0; i < bands[band]->n_channels; i++) {
+			channel = &bands[band]->channels[i];
+
+			if (channel->flags & IEEE80211_CHAN_DISABLED)
+				continue;
+
+			ch->allow_ht   = true;
+
+			/* FIXME: when should we really allow VHT? */
+			ch->allow_vht = true;
+
+			ch->allow_ibss =
+				!(channel->flags & IEEE80211_CHAN_NO_IBSS);
+
+			ch->ht40plus =
+				!(channel->flags & IEEE80211_CHAN_NO_HT40PLUS);
+
+			passive = channel->flags & IEEE80211_CHAN_PASSIVE_SCAN;
+			ch->passive = passive;
+
+			ch->freq = channel->center_freq;
+			ch->min_power = channel->max_power * 3;
+			ch->max_power = channel->max_power * 4;
+			ch->max_reg_power = channel->max_reg_power * 4;
+			ch->max_antenna_gain = channel->max_antenna_gain;
+			ch->reg_class_id = 0; /* FIXME */
+
+			/* FIXME: why use only legacy modes, why not any
+			 * HT/VHT modes? Would that even make any
+			 * difference? */
+			if (channel->band == IEEE80211_BAND_2GHZ)
+				ch->mode = MODE_11G;
+			else
+				ch->mode = MODE_11A;
+
+			if (WARN_ON_ONCE(ch->mode == MODE_UNKNOWN))
+				continue;
+
+			ath10k_dbg(ATH10K_DBG_WMI,
+				   "%s: [%zd/%d] freq %d maxpower %d regpower %d antenna %d mode %d\n",
+				   __func__, ch - arg.channels, arg.n_channels,
+				   ch->freq, ch->max_power, ch->max_reg_power,
+				   ch->max_antenna_gain, ch->mode);
+
+			ch++;
+		}
+	}
+
+	ret = ath10k_wmi_scan_chan_list(ar, &arg);
+	kfree(arg.channels);
+
+	return ret;
+}
+
+static void ath10k_reg_notifier(struct wiphy *wiphy,
+				struct regulatory_request *request)
+{
+	struct ieee80211_hw *hw = wiphy_to_ieee80211_hw(wiphy);
+	struct reg_dmn_pair_mapping *regpair;
+	struct ath10k *ar = hw->priv;
+	int ret;
+
+	ath_reg_notifier_apply(wiphy, request, &ar->ath_common.regulatory);
+
+	ret = ath10k_update_channel_list(ar);
+	if (ret)
+		ath10k_warn("could not update channel list (%d)\n", ret);
+
+	regpair = ar->ath_common.regulatory.regpair;
+	/* Target allows setting up per-band regdomain but ath_common provides
+	 * a combined one only */
+	ret = ath10k_wmi_pdev_set_regdomain(ar,
+					    regpair->regDmnEnum,
+					    regpair->regDmnEnum, /* 2ghz */
+					    regpair->regDmnEnum, /* 5ghz */
+					    regpair->reg_2ghz_ctl,
+					    regpair->reg_5ghz_ctl);
+	if (ret)
+		ath10k_warn("could not set pdev regdomain (%d)\n", ret);
+}
+
+/***************/
+/* TX handlers */
+/***************/
+
+/*
+ * Frames sent to the FW have to be in "Native Wifi" format.
+ * Strip the QoS field from the 802.11 header.
+ */
+static void ath10k_tx_h_qos_workaround(struct ieee80211_hw *hw,
+				       struct ieee80211_tx_control *control,
+				       struct sk_buff *skb)
+{
+	struct ieee80211_hdr *hdr = (void *)skb->data;
+	u8 *qos_ctl;
+
+	if (!ieee80211_is_data_qos(hdr->frame_control))
+		return;
+
+	qos_ctl = ieee80211_get_qos_ctl(hdr);
+	memmove(qos_ctl, qos_ctl + IEEE80211_QOS_CTL_LEN,
+		skb->len - ieee80211_hdrlen(hdr->frame_control));
+	skb_trim(skb, skb->len - IEEE80211_QOS_CTL_LEN);
+}
+
+static void ath10k_tx_h_update_wep_key(struct sk_buff *skb)
+{
+	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
+	struct ieee80211_vif *vif = info->control.vif;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	struct ath10k *ar = arvif->ar;
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
+	struct ieee80211_key_conf *key = info->control.hw_key;
+	int ret;
+
+	/* TODO AP mode should be implemented */
+	if (vif->type != NL80211_IFTYPE_STATION)
+		return;
+
+	if (!ieee80211_has_protected(hdr->frame_control))
+		return;
+
+	if (!key)
+		return;
+
+	if (key->cipher != WLAN_CIPHER_SUITE_WEP40 &&
+	    key->cipher != WLAN_CIPHER_SUITE_WEP104)
+		return;
+
+	if (key->keyidx == arvif->def_wep_key_index)
+		return;
+
+	ath10k_dbg(ATH10K_DBG_MAC, "new wep keyidx will be %d\n", key->keyidx);
+
+	ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+					WMI_VDEV_PARAM_DEF_KEYID,
+					key->keyidx);
+	if (ret) {
+		ath10k_warn("could not update wep keyidx (%d)\n", ret);
+		return;
+	}
+
+	arvif->def_wep_key_index = key->keyidx;
+}
+
+static void ath10k_tx_h_add_p2p_noa_ie(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
+	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
+	struct ieee80211_vif *vif = info->control.vif;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+
+	/* This is case only for P2P_GO */
+	if (arvif->vdev_type != WMI_VDEV_TYPE_AP ||
+	    arvif->vdev_subtype != WMI_VDEV_SUBTYPE_P2P_GO)
+		return;
+
+	if (unlikely(ieee80211_is_probe_resp(hdr->frame_control))) {
+		spin_lock_bh(&ar->data_lock);
+		if (arvif->u.ap.noa_data)
+			if (!pskb_expand_head(skb, 0, arvif->u.ap.noa_len,
+					      GFP_ATOMIC))
+				memcpy(skb_put(skb, arvif->u.ap.noa_len),
+				       arvif->u.ap.noa_data,
+				       arvif->u.ap.noa_len);
+		spin_unlock_bh(&ar->data_lock);
+	}
+}
+
+static void ath10k_tx_htt(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
+	int ret;
+
+	if (ieee80211_is_mgmt(hdr->frame_control))
+		ret = ath10k_htt_mgmt_tx(ar->htt, skb);
+	else if (ieee80211_is_nullfunc(hdr->frame_control))
+		/* FW does not report tx status properly for NullFunc frames
+		 * unless they are sent through mgmt tx path. mac80211 sends
+		 * those frames when it detects link/beacon loss and depends on
+		 * the tx status to be correct. */
+		ret = ath10k_htt_mgmt_tx(ar->htt, skb);
+	else
+		ret = ath10k_htt_tx(ar->htt, skb);
+
+	if (ret) {
+		ath10k_warn("tx failed (%d). dropping packet.\n", ret);
+		ieee80211_free_txskb(ar->hw, skb);
+	}
+}
+
+void ath10k_offchan_tx_purge(struct ath10k *ar)
+{
+	struct sk_buff *skb;
+
+	for (;;) {
+		skb = skb_dequeue(&ar->offchan_tx_queue);
+		if (!skb)
+			break;
+
+		ieee80211_free_txskb(ar->hw, skb);
+	}
+}
+
+void ath10k_offchan_tx_work(struct work_struct *work)
+{
+	struct ath10k *ar = container_of(work, struct ath10k, offchan_tx_work);
+	struct ath10k_peer *peer;
+	struct ieee80211_hdr *hdr;
+	struct sk_buff *skb;
+	const u8 *peer_addr;
+	int vdev_id;
+	int ret;
+
+	/* FW requirement: We must create a peer before FW will send out
+	 * an offchannel frame. Otherwise the frame will be stuck and
+	 * never transmitted. We delete the peer upon tx completion.
+	 * It is unlikely that a peer for offchannel tx will already be
+	 * present. However it may be in some rare cases so account for that.
+	 * Otherwise we might remove a legitimate peer and break stuff. */
+
+	for (;;) {
+		skb = skb_dequeue(&ar->offchan_tx_queue);
+		if (!skb)
+			break;
+
+		mutex_lock(&ar->conf_mutex);
+
+		ath10k_dbg(ATH10K_DBG_MAC, "processing offchannel skb %p\n",
+			   skb);
+
+		hdr = (struct ieee80211_hdr *)skb->data;
+		peer_addr = ieee80211_get_DA(hdr);
+		vdev_id = ATH10K_SKB_CB(skb)->htt.vdev_id;
+
+		spin_lock_bh(&ar->data_lock);
+		peer = ath10k_peer_find(ar, vdev_id, peer_addr);
+		spin_unlock_bh(&ar->data_lock);
+
+		if (peer)
+			ath10k_dbg(ATH10K_DBG_MAC, "peer %pM on vdev %d already present\n",
+				   peer_addr, vdev_id);
+
+		if (!peer) {
+			ret = ath10k_peer_create(ar, vdev_id, peer_addr);
+			if (ret)
+				ath10k_warn("peer %pM on vdev %d not created (%d)\n",
+					    peer_addr, vdev_id, ret);
+		}
+
+		spin_lock_bh(&ar->data_lock);
+		INIT_COMPLETION(ar->offchan_tx_completed);
+		ar->offchan_tx_skb = skb;
+		spin_unlock_bh(&ar->data_lock);
+
+		ath10k_tx_htt(ar, skb);
+
+		ret = wait_for_completion_timeout(&ar->offchan_tx_completed,
+						  3 * HZ);
+		if (ret <= 0)
+			ath10k_warn("timed out waiting for offchannel skb %p\n",
+				    skb);
+
+		if (!peer) {
+			ret = ath10k_peer_delete(ar, vdev_id, peer_addr);
+			if (ret)
+				ath10k_warn("peer %pM on vdev %d not deleted (%d)\n",
+					    peer_addr, vdev_id, ret);
+		}
+
+		mutex_unlock(&ar->conf_mutex);
+	}
+}
+
+/************/
+/* Scanning */
+/************/
+
+/*
+ * This gets called if we dont get a heart-beat during scan.
+ * This may indicate the FW has hung and we need to abort the
+ * scan manually to prevent cancel_hw_scan() from deadlocking
+ */
+void ath10k_reset_scan(unsigned long ptr)
+{
+	struct ath10k *ar = (struct ath10k *)ptr;
+
+	spin_lock_bh(&ar->data_lock);
+	if (!ar->scan.in_progress) {
+		spin_unlock_bh(&ar->data_lock);
+		return;
+	}
+
+	ath10k_warn("scan timeout. resetting. fw issue?\n");
+
+	if (ar->scan.is_roc)
+		ieee80211_remain_on_channel_expired(ar->hw);
+	else
+		ieee80211_scan_completed(ar->hw, 1 /* aborted */);
+
+	ar->scan.in_progress = false;
+	complete_all(&ar->scan.completed);
+	spin_unlock_bh(&ar->data_lock);
+}
+
+static int ath10k_abort_scan(struct ath10k *ar)
+{
+	struct wmi_stop_scan_arg arg = {
+		.req_id = 1, /* FIXME */
+		.req_type = WMI_SCAN_STOP_ONE,
+		.u.scan_id = ATH10K_SCAN_ID,
+	};
+	int ret;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	del_timer_sync(&ar->scan.timeout);
+
+	spin_lock_bh(&ar->data_lock);
+	if (!ar->scan.in_progress) {
+		spin_unlock_bh(&ar->data_lock);
+		return 0;
+	}
+
+	ar->scan.aborting = true;
+	spin_unlock_bh(&ar->data_lock);
+
+	ret = ath10k_wmi_stop_scan(ar, &arg);
+	if (ret) {
+		ath10k_warn("could not submit wmi stop scan (%d)\n", ret);
+		return -EIO;
+	}
+
+	ath10k_wmi_flush_tx(ar);
+
+	ret = wait_for_completion_timeout(&ar->scan.completed, 3*HZ);
+	if (ret == 0)
+		ath10k_warn("timed out while waiting for scan to stop\n");
+
+	/* scan completion may be done right after we timeout here, so let's
+	 * check the in_progress and tell mac80211 scan is completed. if we
+	 * don't do that and FW fails to send us scan completion indication
+	 * then userspace won't be able to scan anymore */
+	ret = 0;
+
+	spin_lock_bh(&ar->data_lock);
+	if (ar->scan.in_progress) {
+		ath10k_warn("could not stop scan. its still in progress\n");
+		ar->scan.in_progress = false;
+		ath10k_offchan_tx_purge(ar);
+		ret = -ETIMEDOUT;
+	}
+	spin_unlock_bh(&ar->data_lock);
+
+	return ret;
+}
+
+static int ath10k_start_scan(struct ath10k *ar,
+			     const struct wmi_start_scan_arg *arg)
+{
+	int ret;
+
+	lockdep_assert_held(&ar->conf_mutex);
+
+	ret = ath10k_wmi_start_scan(ar, arg);
+	if (ret)
+		return ret;
+
+	/* make sure we submit the command so the completion
+	* timeout makes sense */
+	ath10k_wmi_flush_tx(ar);
+
+	ret = wait_for_completion_timeout(&ar->scan.started, 1*HZ);
+	if (ret == 0) {
+		ath10k_abort_scan(ar);
+		return ret;
+	}
+
+	/* the scan can complete earlier, before we even
+	 * start the timer. in that case the timer handler
+	 * checks ar->scan.in_progress and bails out if its
+	 * false. Add a 200ms margin to account event/command
+	 * processing. */
+	mod_timer(&ar->scan.timeout, jiffies +
+		  msecs_to_jiffies(arg->max_scan_time+200));
+	return 0;
+}
+
+/**********************/
+/* mac80211 callbacks */
+/**********************/
+
+static void ath10k_tx(struct ieee80211_hw *hw,
+		      struct ieee80211_tx_control *control,
+		      struct sk_buff *skb)
+{
+	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = NULL;
+	u32 vdev_id = 0;
+	u8 tid;
+
+	if (info->control.vif) {
+		arvif = ath10k_vif_to_arvif(info->control.vif);
+		vdev_id = arvif->vdev_id;
+	} else if (ar->monitor_enabled) {
+		vdev_id = ar->monitor_vdev_id;
+	}
+
+	/* We should disable CCK RATE due to P2P */
+	if (info->flags & IEEE80211_TX_CTL_NO_CCK_RATE)
+		ath10k_dbg(ATH10K_DBG_MAC, "IEEE80211_TX_CTL_NO_CCK_RATE\n");
+
+	/* we must calculate tid before we apply qos workaround
+	 * as we'd lose the qos control field */
+	tid = HTT_DATA_TX_EXT_TID_NON_QOS_MCAST_BCAST;
+	if (ieee80211_is_data_qos(hdr->frame_control) &&
+	    is_unicast_ether_addr(ieee80211_get_DA(hdr))) {
+		u8 *qc = ieee80211_get_qos_ctl(hdr);
+		tid = qc[0] & IEEE80211_QOS_CTL_TID_MASK;
+	}
+
+	ath10k_tx_h_qos_workaround(hw, control, skb);
+	ath10k_tx_h_update_wep_key(skb);
+	ath10k_tx_h_add_p2p_noa_ie(ar, skb);
+	ath10k_tx_h_seq_no(skb);
+
+	memset(ATH10K_SKB_CB(skb), 0, sizeof(*ATH10K_SKB_CB(skb)));
+	ATH10K_SKB_CB(skb)->htt.vdev_id = vdev_id;
+	ATH10K_SKB_CB(skb)->htt.tid = tid;
+
+	if (info->flags & IEEE80211_TX_CTL_TX_OFFCHAN) {
+		spin_lock_bh(&ar->data_lock);
+		ATH10K_SKB_CB(skb)->htt.is_offchan = true;
+		ATH10K_SKB_CB(skb)->htt.vdev_id = ar->scan.vdev_id;
+		spin_unlock_bh(&ar->data_lock);
+
+		ath10k_dbg(ATH10K_DBG_MAC, "queued offchannel skb %p\n", skb);
+
+		skb_queue_tail(&ar->offchan_tx_queue, skb);
+		ieee80211_queue_work(hw, &ar->offchan_tx_work);
+		return;
+	}
+
+	ath10k_tx_htt(ar, skb);
+}
+
+/*
+ * Initialize various parameters with default vaules.
+ */
+static int ath10k_start(struct ieee80211_hw *hw)
+{
+	struct ath10k *ar = hw->priv;
+	int ret;
+
+	ret = ath10k_wmi_pdev_set_param(ar, WMI_PDEV_PARAM_PMF_QOS, 1);
+	if (ret)
+		ath10k_warn("could not enable WMI_PDEV_PARAM_PMF_QOS (%d)\n",
+			    ret);
+
+	ret = ath10k_wmi_pdev_set_param(ar, WMI_PDEV_PARAM_DYNAMIC_BW, 0);
+	if (ret)
+		ath10k_warn("could not init WMI_PDEV_PARAM_DYNAMIC_BW (%d)\n",
+			    ret);
+
+	return 0;
+}
+
+static void ath10k_stop(struct ieee80211_hw *hw)
+{
+	struct ath10k *ar = hw->priv;
+
+	/* avoid leaks in case FW never confirms scan for offchannel */
+	cancel_work_sync(&ar->offchan_tx_work);
+	ath10k_offchan_tx_purge(ar);
+}
+
+static int ath10k_config(struct ieee80211_hw *hw, u32 changed)
+{
+	struct ath10k_generic_iter ar_iter;
+	struct ath10k *ar = hw->priv;
+	struct ieee80211_conf *conf = &hw->conf;
+	int ret = 0;
+	u32 flags;
+
+	mutex_lock(&ar->conf_mutex);
+
+	if (changed & IEEE80211_CONF_CHANGE_CHANNEL) {
+		ath10k_dbg(ATH10K_DBG_MAC, "Config channel %d mhz\n",
+			   conf->chandef.chan->center_freq);
+		spin_lock_bh(&ar->data_lock);
+		ar->rx_channel = conf->chandef.chan;
+		spin_unlock_bh(&ar->data_lock);
+	}
+
+	if (changed & IEEE80211_CONF_CHANGE_PS) {
+		memset(&ar_iter, 0, sizeof(struct ath10k_generic_iter));
+		ar_iter.ar = ar;
+		flags = IEEE80211_IFACE_ITER_RESUME_ALL;
+
+		ieee80211_iterate_active_interfaces_atomic(hw,
+							   flags,
+							   ath10k_ps_iter,
+							   &ar_iter);
+
+		ret = ar_iter.ret;
+	}
+
+	if (changed & IEEE80211_CONF_CHANGE_MONITOR) {
+		if (conf->flags & IEEE80211_CONF_MONITOR)
+			ret = ath10k_monitor_create(ar);
+		else
+			ret = ath10k_monitor_destroy(ar);
+	}
+
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+/*
+ * TODO:
+ * Figure out how to handle WMI_VDEV_SUBTYPE_P2P_DEVICE,
+ * because we will send mgmt frames without CCK. This requirement
+ * for P2P_FIND/GO_NEG should be handled by checking CCK flag
+ * in the TX packet.
+ */
+static int ath10k_add_interface(struct ieee80211_hw *hw,
+				struct ieee80211_vif *vif)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	enum wmi_sta_powersave_param param;
+	int ret = 0;
+	u32 value;
+	int bit;
+
+	mutex_lock(&ar->conf_mutex);
+
+	arvif->ar = ar;
+	arvif->vif = vif;
+
+	if ((vif->type == NL80211_IFTYPE_MONITOR) && ar->monitor_present) {
+		ath10k_warn("Only one monitor interface allowed\n");
+		ret = -EBUSY;
+		goto exit;
+	}
+
+	bit = ffs(ar->free_vdev_map);
+	if (bit == 0) {
+		ret = -EBUSY;
+		goto exit;
+	}
+
+	arvif->vdev_id = bit - 1;
+	arvif->vdev_subtype = WMI_VDEV_SUBTYPE_NONE;
+	ar->free_vdev_map &= ~(1 << arvif->vdev_id);
+
+	if (ar->p2p)
+		arvif->vdev_subtype = WMI_VDEV_SUBTYPE_P2P_DEVICE;
+
+	switch (vif->type) {
+	case NL80211_IFTYPE_UNSPECIFIED:
+	case NL80211_IFTYPE_STATION:
+		arvif->vdev_type = WMI_VDEV_TYPE_STA;
+		if (vif->p2p)
+			arvif->vdev_subtype = WMI_VDEV_SUBTYPE_P2P_CLIENT;
+		break;
+	case NL80211_IFTYPE_ADHOC:
+		arvif->vdev_type = WMI_VDEV_TYPE_IBSS;
+		break;
+	case NL80211_IFTYPE_AP:
+		arvif->vdev_type = WMI_VDEV_TYPE_AP;
+
+		if (vif->p2p)
+			arvif->vdev_subtype = WMI_VDEV_SUBTYPE_P2P_GO;
+		break;
+	case NL80211_IFTYPE_MONITOR:
+		arvif->vdev_type = WMI_VDEV_TYPE_MONITOR;
+		break;
+	default:
+		WARN_ON(1);
+		break;
+	}
+
+	ath10k_dbg(ATH10K_DBG_MAC, "Add interface: id %d type %d subtype %d\n",
+		   arvif->vdev_id, arvif->vdev_type, arvif->vdev_subtype);
+
+	ret = ath10k_wmi_vdev_create(ar, arvif->vdev_id, arvif->vdev_type,
+				     arvif->vdev_subtype, vif->addr);
+	if (ret) {
+		ath10k_warn("WMI vdev create failed: ret %d\n", ret);
+		goto exit;
+	}
+
+	ret = ath10k_wmi_vdev_set_param(ar, 0, WMI_VDEV_PARAM_DEF_KEYID,
+					arvif->def_wep_key_index);
+	if (ret)
+		ath10k_warn("Failed to set default keyid: %d\n", ret);
+
+	ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+					WMI_VDEV_PARAM_TX_ENCAP_TYPE,
+					ATH10K_HW_TXRX_NATIVE_WIFI);
+	if (ret)
+		ath10k_warn("Failed to set TX encap: %d\n", ret);
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_AP) {
+		ret = ath10k_peer_create(ar, arvif->vdev_id, vif->addr);
+		if (ret) {
+			ath10k_warn("Failed to create peer for AP: %d\n", ret);
+			goto exit;
+		}
+	}
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_STA) {
+		param = WMI_STA_PS_PARAM_RX_WAKE_POLICY;
+		value = WMI_STA_PS_RX_WAKE_POLICY_WAKE;
+		ret = ath10k_wmi_set_sta_ps_param(ar, arvif->vdev_id,
+						  param, value);
+		if (ret)
+			ath10k_warn("Failed to set RX wake policy: %d\n", ret);
+
+		param = WMI_STA_PS_PARAM_TX_WAKE_THRESHOLD;
+		value = WMI_STA_PS_TX_WAKE_THRESHOLD_ALWAYS;
+		ret = ath10k_wmi_set_sta_ps_param(ar, arvif->vdev_id,
+						  param, value);
+		if (ret)
+			ath10k_warn("Failed to set TX wake thresh: %d\n", ret);
+
+		param = WMI_STA_PS_PARAM_PSPOLL_COUNT;
+		value = WMI_STA_PS_PSPOLL_COUNT_NO_MAX;
+		ret = ath10k_wmi_set_sta_ps_param(ar, arvif->vdev_id,
+						  param, value);
+		if (ret)
+			ath10k_warn("Failed to set PSPOLL count: %d\n", ret);
+	}
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_MONITOR)
+		ar->monitor_present = true;
+
+exit:
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+static void ath10k_remove_interface(struct ieee80211_hw *hw,
+				    struct ieee80211_vif *vif)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	int ret;
+
+	mutex_lock(&ar->conf_mutex);
+
+	ath10k_dbg(ATH10K_DBG_MAC, "Remove interface: id %d\n", arvif->vdev_id);
+
+	ar->free_vdev_map |= 1 << (arvif->vdev_id);
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_AP) {
+		ret = ath10k_peer_delete(arvif->ar, arvif->vdev_id, vif->addr);
+		if (ret)
+			ath10k_warn("Failed to remove peer for AP: %d\n", ret);
+
+		kfree(arvif->u.ap.noa_data);
+	}
+
+	ret = ath10k_wmi_vdev_delete(ar, arvif->vdev_id);
+	if (ret)
+		ath10k_warn("WMI vdev delete failed: %d\n", ret);
+
+	if (arvif->vdev_type == WMI_VDEV_TYPE_MONITOR)
+		ar->monitor_present = false;
+
+	ath10k_peer_cleanup(ar, arvif->vdev_id);
+
+	mutex_unlock(&ar->conf_mutex);
+}
+
+/*
+ * FIXME: Has to be verified.
+ */
+#define SUPPORTED_FILTERS			\
+	(FIF_PROMISC_IN_BSS |			\
+	FIF_ALLMULTI |				\
+	FIF_CONTROL |				\
+	FIF_PSPOLL |				\
+	FIF_OTHER_BSS |				\
+	FIF_BCN_PRBRESP_PROMISC |		\
+	FIF_PROBE_REQ |				\
+	FIF_FCSFAIL)
+
+static void ath10k_configure_filter(struct ieee80211_hw *hw,
+				    unsigned int changed_flags,
+				    unsigned int *total_flags,
+				    u64 multicast)
+{
+	struct ath10k *ar = hw->priv;
+	int ret;
+
+	mutex_lock(&ar->conf_mutex);
+
+	changed_flags &= SUPPORTED_FILTERS;
+	*total_flags &= SUPPORTED_FILTERS;
+	ar->filter_flags = *total_flags;
+
+	if ((ar->filter_flags & FIF_PROMISC_IN_BSS) &&
+	    !ar->monitor_enabled) {
+		ret = ath10k_monitor_start(ar, ar->monitor_vdev_id);
+		if (ret)
+			ath10k_warn("Unable to start monitor mode\n");
+		else
+			ath10k_dbg(ATH10K_DBG_MAC, "Monitor mode started\n");
+	} else if (!(ar->filter_flags & FIF_PROMISC_IN_BSS) &&
+		   ar->monitor_enabled) {
+		ret = ath10k_monitor_stop(ar);
+		if (ret)
+			ath10k_warn("Unable to stop monitor mode\n");
+		else
+			ath10k_dbg(ATH10K_DBG_MAC, "Monitor mode stopped\n");
+	}
+
+	mutex_unlock(&ar->conf_mutex);
+}
+
+static void ath10k_bss_info_changed(struct ieee80211_hw *hw,
+				    struct ieee80211_vif *vif,
+				    struct ieee80211_bss_conf *info,
+				    u32 changed)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	int ret = 0;
+
+	mutex_lock(&ar->conf_mutex);
+
+	if (changed & BSS_CHANGED_IBSS)
+		ath10k_control_ibss(arvif, info, vif->addr);
+
+	if (changed & BSS_CHANGED_BEACON_INT) {
+		arvif->beacon_interval = info->beacon_int;
+		ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+						WMI_VDEV_PARAM_BEACON_INTERVAL,
+						arvif->beacon_interval);
+		if (ret)
+			ath10k_warn("Failed to set beacon interval for VDEV: %d\n",
+				    arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Beacon interval: %d set for VDEV: %d\n",
+				   arvif->beacon_interval, arvif->vdev_id);
+	}
+
+	if (changed & BSS_CHANGED_BEACON) {
+		ret = ath10k_wmi_pdev_set_param(ar,
+						WMI_PDEV_PARAM_BEACON_TX_MODE,
+						WMI_BEACON_STAGGERED_MODE);
+		if (ret)
+			ath10k_warn("Failed to set beacon mode for VDEV: %d\n",
+				    arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Set staggered beacon mode for VDEV: %d\n",
+				   arvif->vdev_id);
+	}
+
+	if (changed & BSS_CHANGED_BEACON_INFO) {
+		arvif->dtim_period = info->dtim_period;
+
+		ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+						WMI_VDEV_PARAM_DTIM_PERIOD,
+						arvif->dtim_period);
+		if (ret)
+			ath10k_warn("Failed to set dtim period for VDEV: %d\n",
+				    arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Set dtim period: %d for VDEV: %d\n",
+				   arvif->dtim_period, arvif->vdev_id);
+	}
+
+	if (changed & BSS_CHANGED_SSID &&
+	    vif->type == NL80211_IFTYPE_AP) {
+		arvif->u.ap.ssid_len = info->ssid_len;
+		if (info->ssid_len)
+			memcpy(arvif->u.ap.ssid, info->ssid, info->ssid_len);
+		arvif->u.ap.hidden_ssid = info->hidden_ssid;
+	}
+
+	if (changed & BSS_CHANGED_BSSID) {
+		if (!is_zero_ether_addr(info->bssid)) {
+			ret = ath10k_peer_create(ar, arvif->vdev_id,
+						 info->bssid);
+			if (ret)
+				ath10k_warn("Failed to add peer: %pM for VDEV: %d\n",
+					    info->bssid, arvif->vdev_id);
+			else
+				ath10k_dbg(ATH10K_DBG_MAC,
+					   "Added peer: %pM for VDEV: %d\n",
+					   info->bssid, arvif->vdev_id);
+
+
+			if (vif->type == NL80211_IFTYPE_STATION) {
+				/*
+				 * this is never erased as we it for crypto key
+				 * clearing; this is FW requirement
+				 */
+				memcpy(arvif->u.sta.bssid, info->bssid,
+				       ETH_ALEN);
+
+				ret = ath10k_vdev_start(arvif);
+				if (!ret)
+					ath10k_dbg(ATH10K_DBG_MAC,
+						   "VDEV: %d started with BSSID: %pM\n",
+						   arvif->vdev_id, info->bssid);
+			}
+
+			/*
+			 * Mac80211 does not keep IBSS bssid when leaving IBSS,
+			 * so driver need to store it. It is needed when leaving
+			 * IBSS in order to remove BSSID peer.
+			 */
+			if (vif->type == NL80211_IFTYPE_ADHOC)
+				memcpy(arvif->u.ibss.bssid, info->bssid,
+				       ETH_ALEN);
+		}
+	}
+
+	if (changed & BSS_CHANGED_BEACON_ENABLED)
+		ath10k_control_beaconing(arvif, info);
+
+	if (changed & BSS_CHANGED_ERP_CTS_PROT) {
+		u32 cts_prot;
+		if (info->use_cts_prot)
+			cts_prot = 1;
+		else
+			cts_prot = 0;
+
+		ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+						WMI_VDEV_PARAM_ENABLE_RTSCTS,
+						cts_prot);
+		if (ret)
+			ath10k_warn("Failed to set CTS prot for VDEV: %d\n",
+				    arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Set CTS prot: %d for VDEV: %d\n",
+				   cts_prot, arvif->vdev_id);
+	}
+
+	if (changed & BSS_CHANGED_ERP_SLOT) {
+		u32 slottime;
+		if (info->use_short_slot)
+			slottime = WMI_VDEV_SLOT_TIME_SHORT; /* 9us */
+
+		else
+			slottime = WMI_VDEV_SLOT_TIME_LONG; /* 20us */
+
+		ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+						WMI_VDEV_PARAM_SLOT_TIME,
+						slottime);
+		if (ret)
+			ath10k_warn("Failed to set erp slot for VDEV: %d\n",
+				    arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Set slottime: %d for VDEV: %d\n",
+				   slottime, arvif->vdev_id);
+	}
+
+	if (changed & BSS_CHANGED_ERP_PREAMBLE) {
+		u32 preamble;
+		if (info->use_short_preamble)
+			preamble = WMI_VDEV_PREAMBLE_SHORT;
+		else
+			preamble = WMI_VDEV_PREAMBLE_LONG;
+
+		ret = ath10k_wmi_vdev_set_param(ar, arvif->vdev_id,
+						WMI_VDEV_PARAM_PREAMBLE,
+						preamble);
+		if (ret)
+			ath10k_warn("Failed to set preamble for VDEV: %d\n",
+				    arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Set preamble: %d for VDEV: %d\n",
+				   preamble, arvif->vdev_id);
+	}
+
+	if (changed & BSS_CHANGED_ASSOC) {
+		if (info->assoc)
+			ath10k_bss_assoc(hw, vif, info);
+	}
+
+	mutex_unlock(&ar->conf_mutex);
+}
+
+static int ath10k_hw_scan(struct ieee80211_hw *hw,
+			  struct ieee80211_vif *vif,
+			  struct cfg80211_scan_request *req)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	struct wmi_start_scan_arg arg;
+	int ret = 0;
+	int i;
+
+	mutex_lock(&ar->conf_mutex);
+
+	spin_lock_bh(&ar->data_lock);
+	if (ar->scan.in_progress) {
+		spin_unlock_bh(&ar->data_lock);
+		ret = -EBUSY;
+		goto exit;
+	}
+
+	INIT_COMPLETION(ar->scan.started);
+	INIT_COMPLETION(ar->scan.completed);
+	ar->scan.in_progress = true;
+	ar->scan.aborting = false;
+	ar->scan.is_roc = false;
+	ar->scan.vdev_id = arvif->vdev_id;
+	spin_unlock_bh(&ar->data_lock);
+
+	memset(&arg, 0, sizeof(arg));
+	ath10k_wmi_start_scan_init(ar, &arg);
+	arg.vdev_id = arvif->vdev_id;
+	arg.scan_id = ATH10K_SCAN_ID;
+
+	if (!req->no_cck)
+		arg.scan_ctrl_flags |= WMI_SCAN_ADD_CCK_RATES;
+
+	if (req->ie_len) {
+		arg.ie_len = req->ie_len;
+		memcpy(arg.ie, req->ie, arg.ie_len);
+	}
+
+	if (req->n_ssids) {
+		arg.n_ssids = req->n_ssids;
+		for (i = 0; i < arg.n_ssids; i++) {
+			arg.ssids[i].len  = req->ssids[i].ssid_len;
+			arg.ssids[i].ssid = req->ssids[i].ssid;
+		}
+	}
+
+	if (req->n_channels) {
+		arg.n_channels = req->n_channels;
+		for (i = 0; i < arg.n_channels; i++)
+			arg.channels[i] = req->channels[i]->center_freq;
+	}
+
+	ret = ath10k_start_scan(ar, &arg);
+	if (ret) {
+		ath10k_warn("could not start hw scan (%d)\n", ret);
+		spin_lock_bh(&ar->data_lock);
+		ar->scan.in_progress = false;
+		spin_unlock_bh(&ar->data_lock);
+	}
+
+exit:
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+static void ath10k_cancel_hw_scan(struct ieee80211_hw *hw,
+				  struct ieee80211_vif *vif)
+{
+	struct ath10k *ar = hw->priv;
+	int ret;
+
+	mutex_lock(&ar->conf_mutex);
+	ret = ath10k_abort_scan(ar);
+	if (ret) {
+		ath10k_warn("couldn't abort scan (%d). forcefully sending scan completion to mac80211\n",
+			    ret);
+		ieee80211_scan_completed(hw, 1 /* aborted */);
+	}
+	mutex_unlock(&ar->conf_mutex);
+}
+
+static int ath10k_set_key(struct ieee80211_hw *hw, enum set_key_cmd cmd,
+			  struct ieee80211_vif *vif, struct ieee80211_sta *sta,
+			  struct ieee80211_key_conf *key)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	struct ath10k_peer *peer;
+	const u8 *peer_addr;
+	bool is_wep = key->cipher == WLAN_CIPHER_SUITE_WEP40 ||
+		      key->cipher == WLAN_CIPHER_SUITE_WEP104;
+	int ret = 0;
+
+	if (key->keyidx > WMI_MAX_KEY_INDEX)
+		return -ENOSPC;
+
+	mutex_lock(&ar->conf_mutex);
+
+	if (sta)
+		peer_addr = sta->addr;
+	else if (arvif->vdev_type == WMI_VDEV_TYPE_STA)
+		peer_addr = vif->bss_conf.bssid;
+	else
+		peer_addr = vif->addr;
+
+	key->hw_key_idx = key->keyidx;
+
+	/* the peer should not disappear in mid-way (unless FW goes awry) since
+	 * we already hold conf_mutex. we just make sure its there now. */
+	spin_lock_bh(&ar->data_lock);
+	peer = ath10k_peer_find(ar, arvif->vdev_id, peer_addr);
+	spin_unlock_bh(&ar->data_lock);
+
+	if (!peer) {
+		if (cmd == SET_KEY) {
+			ath10k_warn("cannot install key for non-existent peer %pM\n",
+				    peer_addr);
+			ret = -EOPNOTSUPP;
+			goto exit;
+		} else {
+			/* if the peer doesn't exist there is no key to disable
+			 * anymore */
+			goto exit;
+		}
+	}
+
+	if (is_wep) {
+		if (cmd == SET_KEY)
+			arvif->wep_keys[key->keyidx] = key;
+		else
+			arvif->wep_keys[key->keyidx] = NULL;
+
+		if (cmd == DISABLE_KEY)
+			ath10k_clear_vdev_key(arvif, key);
+	}
+
+	ret = ath10k_install_key(arvif, key, cmd, peer_addr);
+	if (ret) {
+		ath10k_warn("ath10k_install_key failed (%d)\n", ret);
+		goto exit;
+	}
+
+	spin_lock_bh(&ar->data_lock);
+	peer = ath10k_peer_find(ar, arvif->vdev_id, peer_addr);
+	if (peer && cmd == SET_KEY)
+		peer->keys[key->keyidx] = key;
+	else if (peer && cmd == DISABLE_KEY)
+		peer->keys[key->keyidx] = NULL;
+	else if (peer == NULL)
+		/* impossible unless FW goes crazy */
+		ath10k_warn("peer %pM disappeared!\n", peer_addr);
+	spin_unlock_bh(&ar->data_lock);
+
+exit:
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+static int ath10k_sta_state(struct ieee80211_hw *hw,
+			    struct ieee80211_vif *vif,
+			    struct ieee80211_sta *sta,
+			    enum ieee80211_sta_state old_state,
+			    enum ieee80211_sta_state new_state)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	int ret = 0;
+
+	mutex_lock(&ar->conf_mutex);
+
+	if (old_state == IEEE80211_STA_NOTEXIST &&
+	    new_state == IEEE80211_STA_NONE &&
+	    vif->type != NL80211_IFTYPE_STATION) {
+		/*
+		 * New station addition.
+		 */
+		ret = ath10k_peer_create(ar, arvif->vdev_id, sta->addr);
+		if (ret)
+			ath10k_warn("Failed to add peer: %pM for VDEV: %d\n",
+				    sta->addr, arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Added peer: %pM for VDEV: %d\n",
+				   sta->addr, arvif->vdev_id);
+	} else if ((old_state == IEEE80211_STA_NONE &&
+		    new_state == IEEE80211_STA_NOTEXIST)) {
+		/*
+		 * Existing station deletion.
+		 */
+		ret = ath10k_peer_delete(ar, arvif->vdev_id, sta->addr);
+		if (ret)
+			ath10k_warn("Failed to delete peer: %pM for VDEV: %d\n",
+				    sta->addr, arvif->vdev_id);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Removed peer: %pM for VDEV: %d\n",
+				   sta->addr, arvif->vdev_id);
+
+		if (vif->type == NL80211_IFTYPE_STATION)
+			ath10k_bss_disassoc(hw, vif);
+	} else if (old_state == IEEE80211_STA_AUTH &&
+		   new_state == IEEE80211_STA_ASSOC &&
+		   (vif->type == NL80211_IFTYPE_AP ||
+		    vif->type == NL80211_IFTYPE_ADHOC)) {
+		/*
+		 * New association.
+		 */
+		ret = ath10k_station_assoc(ar, arvif, sta);
+		if (ret)
+			ath10k_warn("Failed to associate station: %pM\n",
+				    sta->addr);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Station %pM moved to assoc state\n",
+				   sta->addr);
+	} else if (old_state == IEEE80211_STA_ASSOC &&
+		   new_state == IEEE80211_STA_AUTH &&
+		   (vif->type == NL80211_IFTYPE_AP ||
+		    vif->type == NL80211_IFTYPE_ADHOC)) {
+		/*
+		 * Disassociation.
+		 */
+		ret = ath10k_station_disassoc(ar, arvif, sta);
+		if (ret)
+			ath10k_warn("Failed to disassociate station: %pM\n",
+				    sta->addr);
+		else
+			ath10k_dbg(ATH10K_DBG_MAC,
+				   "Station %pM moved to disassociated state\n",
+				   sta->addr);
+	}
+
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+static int ath10k_conf_tx_uapsd(struct ath10k *ar, struct ieee80211_vif *vif,
+				 u16 ac, bool enable)
+{
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	u32 value = 0;
+	int ret = 0;
+
+	if (arvif->vdev_type != WMI_VDEV_TYPE_STA)
+		return 0;
+
+	switch (ac) {
+	case IEEE80211_AC_VO:
+		value = WMI_STA_PS_UAPSD_AC3_DELIVERY_EN |
+			WMI_STA_PS_UAPSD_AC3_TRIGGER_EN;
+		break;
+	case IEEE80211_AC_VI:
+		value = WMI_STA_PS_UAPSD_AC2_DELIVERY_EN |
+			WMI_STA_PS_UAPSD_AC2_TRIGGER_EN;
+		break;
+	case IEEE80211_AC_BE:
+		value = WMI_STA_PS_UAPSD_AC1_DELIVERY_EN |
+			WMI_STA_PS_UAPSD_AC1_TRIGGER_EN;
+		break;
+	case IEEE80211_AC_BK:
+		value = WMI_STA_PS_UAPSD_AC0_DELIVERY_EN |
+			WMI_STA_PS_UAPSD_AC0_TRIGGER_EN;
+		break;
+	}
+
+	if (enable)
+		arvif->u.sta.uapsd |= value;
+	else
+		arvif->u.sta.uapsd &= ~value;
+
+	ret = ath10k_wmi_set_sta_ps_param(ar, arvif->vdev_id,
+					  WMI_STA_PS_PARAM_UAPSD,
+					  arvif->u.sta.uapsd);
+	if (ret) {
+		ath10k_warn("could not set uapsd params %d\n", ret);
+		goto exit;
+	}
+
+	if (arvif->u.sta.uapsd)
+		value = WMI_STA_PS_RX_WAKE_POLICY_POLL_UAPSD;
+	else
+		value = WMI_STA_PS_RX_WAKE_POLICY_WAKE;
+
+	ret = ath10k_wmi_set_sta_ps_param(ar, arvif->vdev_id,
+					  WMI_STA_PS_PARAM_RX_WAKE_POLICY,
+					  value);
+	if (ret)
+		ath10k_warn("could not set rx wake param %d\n", ret);
+
+exit:
+	return ret;
+}
+
+static int ath10k_conf_tx(struct ieee80211_hw *hw,
+			  struct ieee80211_vif *vif, u16 ac,
+			  const struct ieee80211_tx_queue_params *params)
+{
+	struct ath10k *ar = hw->priv;
+	struct wmi_wmm_params_arg *p = NULL;
+	int ret;
+
+	mutex_lock(&ar->conf_mutex);
+
+	switch (ac) {
+	case IEEE80211_AC_VO:
+		p = &ar->wmm_params.ac_vo;
+		break;
+	case IEEE80211_AC_VI:
+		p = &ar->wmm_params.ac_vi;
+		break;
+	case IEEE80211_AC_BE:
+		p = &ar->wmm_params.ac_be;
+		break;
+	case IEEE80211_AC_BK:
+		p = &ar->wmm_params.ac_bk;
+		break;
+	}
+
+	if (WARN_ON(!p)) {
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	p->cwmin = params->cw_min;
+	p->cwmax = params->cw_max;
+	p->aifs = params->aifs;
+
+	/*
+	 * The channel time duration programmed in the HW is in absolute
+	 * microseconds, while mac80211 gives the txop in units of
+	 * 32 microseconds.
+	 */
+	p->txop = params->txop * 32;
+
+	/* FIXME: FW accepts wmm params per hw, not per vif */
+	ret = ath10k_wmi_pdev_set_wmm_params(ar, &ar->wmm_params);
+	if (ret) {
+		ath10k_warn("could not set wmm params %d\n", ret);
+		goto exit;
+	}
+
+	ret = ath10k_conf_tx_uapsd(ar, vif, ac, params->uapsd);
+	if (ret)
+		ath10k_warn("could not set sta uapsd %d\n", ret);
+
+exit:
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+#define ATH10K_ROC_TIMEOUT_HZ (2*HZ)
+
+static int ath10k_remain_on_channel(struct ieee80211_hw *hw,
+				    struct ieee80211_vif *vif,
+				    struct ieee80211_channel *chan,
+				    int duration,
+				    enum ieee80211_roc_type type)
+{
+	struct ath10k *ar = hw->priv;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	struct wmi_start_scan_arg arg;
+	int ret;
+
+	mutex_lock(&ar->conf_mutex);
+
+	spin_lock_bh(&ar->data_lock);
+	if (ar->scan.in_progress) {
+		spin_unlock_bh(&ar->data_lock);
+		ret = -EBUSY;
+		goto exit;
+	}
+
+	INIT_COMPLETION(ar->scan.started);
+	INIT_COMPLETION(ar->scan.completed);
+	INIT_COMPLETION(ar->scan.on_channel);
+	ar->scan.in_progress = true;
+	ar->scan.aborting = false;
+	ar->scan.is_roc = true;
+	ar->scan.vdev_id = arvif->vdev_id;
+	ar->scan.roc_freq = chan->center_freq;
+	spin_unlock_bh(&ar->data_lock);
+
+	memset(&arg, 0, sizeof(arg));
+	ath10k_wmi_start_scan_init(ar, &arg);
+	arg.vdev_id = arvif->vdev_id;
+	arg.scan_id = ATH10K_SCAN_ID;
+	arg.n_channels = 1;
+	arg.channels[0] = chan->center_freq;
+	arg.dwell_time_active = duration;
+	arg.dwell_time_passive = duration;
+	arg.max_scan_time = 2 * duration;
+	arg.scan_ctrl_flags |= WMI_SCAN_FLAG_PASSIVE;
+	arg.scan_ctrl_flags |= WMI_SCAN_FILTER_PROBE_REQ;
+
+	ret = ath10k_start_scan(ar, &arg);
+	if (ret) {
+		ath10k_warn("could not start roc scan (%d)\n", ret);
+		spin_lock_bh(&ar->data_lock);
+		ar->scan.in_progress = false;
+		spin_unlock_bh(&ar->data_lock);
+		goto exit;
+	}
+
+	ret = wait_for_completion_timeout(&ar->scan.on_channel, 3*HZ);
+	if (ret == 0) {
+		ath10k_warn("could not switch to channel for roc scan\n");
+		ath10k_abort_scan(ar);
+		ret = -ETIMEDOUT;
+		goto exit;
+	}
+
+	ret = 0;
+exit:
+	mutex_unlock(&ar->conf_mutex);
+	return ret;
+}
+
+static int ath10k_cancel_remain_on_channel(struct ieee80211_hw *hw)
+{
+	struct ath10k *ar = hw->priv;
+
+	mutex_lock(&ar->conf_mutex);
+	ath10k_abort_scan(ar);
+	mutex_unlock(&ar->conf_mutex);
+
+	return 0;
+}
+
+/*
+ * Both RTS and Fragmentation threshold are interface-specific
+ * in ath10k, but device-specific in mac80211.
+ */
+static void ath10k_set_rts_iter(void *data, u8 *mac, struct ieee80211_vif *vif)
+{
+	struct ath10k_generic_iter *ar_iter = data;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	u32 rts = ar_iter->ar->hw->wiphy->rts_threshold;
+
+	rts = min_t(u32, rts, ATH10K_RTS_MAX);
+
+	ar_iter->ret = ath10k_wmi_vdev_set_param(ar_iter->ar, arvif->vdev_id,
+						 WMI_VDEV_PARAM_RTS_THRESHOLD,
+						 rts);
+	if (ar_iter->ret)
+		ath10k_warn("Failed to set RTS threshold for VDEV: %d\n",
+			    arvif->vdev_id);
+	else
+		ath10k_dbg(ATH10K_DBG_MAC,
+			   "Set RTS threshold: %d for VDEV: %d\n",
+			   rts, arvif->vdev_id);
+}
+
+static int ath10k_set_rts_threshold(struct ieee80211_hw *hw, u32 value)
+{
+	struct ath10k_generic_iter ar_iter;
+	struct ath10k *ar = hw->priv;
+
+	memset(&ar_iter, 0, sizeof(struct ath10k_generic_iter));
+	ar_iter.ar = ar;
+
+	mutex_lock(&ar->conf_mutex);
+	ieee80211_iterate_active_interfaces(hw, IEEE80211_IFACE_ITER_RESUME_ALL,
+					    ath10k_set_rts_iter, &ar_iter);
+	mutex_unlock(&ar->conf_mutex);
+
+	return ar_iter.ret;
+}
+
+static void ath10k_set_frag_iter(void *data, u8 *mac, struct ieee80211_vif *vif)
+{
+	struct ath10k_generic_iter *ar_iter = data;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+	u32 frag = ar_iter->ar->hw->wiphy->frag_threshold;
+	int ret;
+
+	frag = clamp_t(u32, frag,
+		       ATH10K_FRAGMT_THRESHOLD_MIN,
+		       ATH10K_FRAGMT_THRESHOLD_MAX);
+
+	ret = ath10k_wmi_vdev_set_param(ar_iter->ar, arvif->vdev_id,
+					WMI_VDEV_PARAM_FRAGMENTATION_THRESHOLD,
+					frag);
+
+	ar_iter->ret = ret;
+	if (ar_iter->ret)
+		ath10k_warn("Failed to set frag threshold for VDEV: %d\n",
+			    arvif->vdev_id);
+	else
+		ath10k_dbg(ATH10K_DBG_MAC,
+			   "Set frag threshold: %d for VDEV: %d\n",
+			   frag, arvif->vdev_id);
+}
+
+static int ath10k_set_frag_threshold(struct ieee80211_hw *hw, u32 value)
+{
+	struct ath10k_generic_iter ar_iter;
+	struct ath10k *ar = hw->priv;
+
+	memset(&ar_iter, 0, sizeof(struct ath10k_generic_iter));
+	ar_iter.ar = ar;
+
+	mutex_lock(&ar->conf_mutex);
+	ieee80211_iterate_active_interfaces(hw, IEEE80211_IFACE_ITER_RESUME_ALL,
+					    ath10k_set_frag_iter, &ar_iter);
+	mutex_unlock(&ar->conf_mutex);
+
+	return ar_iter.ret;
+}
+
+static void ath10k_flush(struct ieee80211_hw *hw, u32 queues, bool drop)
+{
+	struct ath10k *ar = hw->priv;
+	int ret;
+
+	/* mac80211 doesn't care if we really xmit queued frames or not
+	 * we'll collect those frames either way if we stop/delete vdevs */
+	if (drop)
+		return;
+
+	ret = wait_event_timeout(ar->htt->empty_tx_wq, ({
+			bool empty;
+			spin_lock_bh(&ar->htt->tx_lock);
+			empty = bitmap_empty(ar->htt->used_msdu_ids,
+					     ar->htt->max_num_pending_tx);
+			spin_unlock_bh(&ar->htt->tx_lock);
+			(empty);
+		}), ATH10K_FLUSH_TIMEOUT_HZ);
+	if (ret <= 0)
+		ath10k_warn("tx not flushed\n");
+}
+
+/* TODO: Implement this function properly
+ * For now it is needed to reply to Probe Requests in IBSS mode.
+ * Propably we need this information from FW.
+ */
+static int ath10k_tx_last_beacon(struct ieee80211_hw *hw)
+{
+	return 1;
+}
+
+static const struct ieee80211_ops ath10k_ops = {
+	.tx				= ath10k_tx,
+	.start				= ath10k_start,
+	.stop				= ath10k_stop,
+	.config				= ath10k_config,
+	.add_interface			= ath10k_add_interface,
+	.remove_interface		= ath10k_remove_interface,
+	.configure_filter		= ath10k_configure_filter,
+	.bss_info_changed		= ath10k_bss_info_changed,
+	.hw_scan			= ath10k_hw_scan,
+	.cancel_hw_scan			= ath10k_cancel_hw_scan,
+	.set_key			= ath10k_set_key,
+	.sta_state			= ath10k_sta_state,
+	.conf_tx			= ath10k_conf_tx,
+	.remain_on_channel		= ath10k_remain_on_channel,
+	.cancel_remain_on_channel	= ath10k_cancel_remain_on_channel,
+	.set_rts_threshold		= ath10k_set_rts_threshold,
+	.set_frag_threshold		= ath10k_set_frag_threshold,
+	.flush				= ath10k_flush,
+	.tx_last_beacon			= ath10k_tx_last_beacon,
+};
+
+#define RATETAB_ENT(_rate, _rateid, _flags) { \
+	.bitrate		= (_rate), \
+	.flags			= (_flags), \
+	.hw_value		= (_rateid), \
+}
+
+#define CHAN2G(_channel, _freq, _flags) { \
+	.band			= IEEE80211_BAND_2GHZ, \
+	.hw_value		= (_channel), \
+	.center_freq		= (_freq), \
+	.flags			= (_flags), \
+	.max_antenna_gain	= 0, \
+	.max_power		= 30, \
+}
+
+#define CHAN5G(_channel, _freq, _flags) { \
+	.band			= IEEE80211_BAND_5GHZ, \
+	.hw_value		= (_channel), \
+	.center_freq		= (_freq), \
+	.flags			= (_flags), \
+	.max_antenna_gain	= 0, \
+	.max_power		= 30, \
+}
+
+static const struct ieee80211_channel ath10k_2ghz_channels[] = {
+	CHAN2G(1, 2412, 0),
+	CHAN2G(2, 2417, 0),
+	CHAN2G(3, 2422, 0),
+	CHAN2G(4, 2427, 0),
+	CHAN2G(5, 2432, 0),
+	CHAN2G(6, 2437, 0),
+	CHAN2G(7, 2442, 0),
+	CHAN2G(8, 2447, 0),
+	CHAN2G(9, 2452, 0),
+	CHAN2G(10, 2457, 0),
+	CHAN2G(11, 2462, 0),
+	CHAN2G(12, 2467, 0),
+	CHAN2G(13, 2472, 0),
+	CHAN2G(14, 2484, 0),
+};
+
+static const struct ieee80211_channel ath10k_5ghz_channels[] = {
+	CHAN5G(36, 5180, 0),
+	CHAN5G(40, 5200, 0),
+	CHAN5G(44, 5220, 0),
+	CHAN5G(48, 5240, 0),
+	CHAN5G(52, 5260, 0),
+	CHAN5G(56, 5280, 0),
+	CHAN5G(60, 5300, 0),
+	CHAN5G(64, 5320, 0),
+	CHAN5G(100, 5500, 0),
+	CHAN5G(104, 5520, 0),
+	CHAN5G(108, 5540, 0),
+	CHAN5G(112, 5560, 0),
+	CHAN5G(116, 5580, 0),
+	CHAN5G(120, 5600, 0),
+	CHAN5G(124, 5620, 0),
+	CHAN5G(128, 5640, 0),
+	CHAN5G(132, 5660, 0),
+	CHAN5G(136, 5680, 0),
+	CHAN5G(140, 5700, 0),
+	CHAN5G(149, 5745, 0),
+	CHAN5G(153, 5765, 0),
+	CHAN5G(157, 5785, 0),
+	CHAN5G(161, 5805, 0),
+	CHAN5G(165, 5825, 0),
+};
+
+static struct ieee80211_rate ath10k_rates[] = {
+	/* CCK */
+	RATETAB_ENT(10,  0x82, 0),
+	RATETAB_ENT(20,  0x84, 0),
+	RATETAB_ENT(55,  0x8b, 0),
+	RATETAB_ENT(110, 0x96, 0),
+	/* OFDM */
+	RATETAB_ENT(60,  0x0c, 0),
+	RATETAB_ENT(90,  0x12, 0),
+	RATETAB_ENT(120, 0x18, 0),
+	RATETAB_ENT(180, 0x24, 0),
+	RATETAB_ENT(240, 0x30, 0),
+	RATETAB_ENT(360, 0x48, 0),
+	RATETAB_ENT(480, 0x60, 0),
+	RATETAB_ENT(540, 0x6c, 0),
+};
+
+#define ath10k_a_rates (ath10k_rates + 4)
+#define ath10k_a_rates_size (ARRAY_SIZE(ath10k_rates) - 4)
+#define ath10k_g_rates (ath10k_rates + 0)
+#define ath10k_g_rates_size (ARRAY_SIZE(ath10k_rates))
+
+struct ath10k *ath10k_mac_create(void)
+{
+	struct ieee80211_hw *hw;
+	struct ath10k *ar;
+
+	hw = ieee80211_alloc_hw(sizeof(struct ath10k), &ath10k_ops);
+	if (!hw)
+		return NULL;
+
+	ar = hw->priv;
+	ar->hw = hw;
+
+	return ar;
+}
+
+void ath10k_mac_destroy(struct ath10k *ar)
+{
+	ieee80211_free_hw(ar->hw);
+}
+
+static const struct ieee80211_iface_limit ath10k_if_limits[] = {
+	{
+	.max	= 8,
+	.types	= BIT(NL80211_IFTYPE_STATION)
+		| BIT(NL80211_IFTYPE_P2P_CLIENT)
+		| BIT(NL80211_IFTYPE_P2P_GO)
+		| BIT(NL80211_IFTYPE_AP)
+	}
+};
+
+static const struct ieee80211_iface_combination ath10k_if_comb = {
+	.limits = ath10k_if_limits,
+	.n_limits = ARRAY_SIZE(ath10k_if_limits),
+	.max_interfaces = 8,
+	.num_different_channels = 1,
+	.beacon_int_infra_match = true,
+};
+
+static struct ieee80211_sta_vht_cap ath10k_create_vht_cap(struct ath10k *ar)
+{
+	struct ieee80211_sta_vht_cap vht_cap = {0};
+	u16 mcs_map;
+
+	vht_cap.vht_supported = 1;
+	vht_cap.cap = ar->vht_cap_info;
+
+	/* FIXME: check dynamically how many streams board supports */
+	mcs_map = IEEE80211_VHT_MCS_SUPPORT_0_9 << 0 |
+		IEEE80211_VHT_MCS_SUPPORT_0_9 << 2 |
+		IEEE80211_VHT_MCS_SUPPORT_0_9 << 4 |
+		IEEE80211_VHT_MCS_NOT_SUPPORTED << 6 |
+		IEEE80211_VHT_MCS_NOT_SUPPORTED << 8 |
+		IEEE80211_VHT_MCS_NOT_SUPPORTED << 10 |
+		IEEE80211_VHT_MCS_NOT_SUPPORTED << 12 |
+		IEEE80211_VHT_MCS_NOT_SUPPORTED << 14;
+
+	vht_cap.vht_mcs.rx_mcs_map = cpu_to_le16(mcs_map);
+	vht_cap.vht_mcs.tx_mcs_map = cpu_to_le16(mcs_map);
+
+	return vht_cap;
+}
+
+static struct ieee80211_sta_ht_cap ath10k_get_ht_cap(struct ath10k *ar)
+{
+	int i;
+	struct ieee80211_sta_ht_cap ht_cap = {0};
+
+	if (!(ar->ht_cap_info & WMI_HT_CAP_ENABLED))
+		return ht_cap;
+
+	ht_cap.ht_supported = 1;
+	ht_cap.ampdu_factor = IEEE80211_HT_MAX_AMPDU_64K;
+	ht_cap.ampdu_density = IEEE80211_HT_MPDU_DENSITY_8;
+	ht_cap.cap |= IEEE80211_HT_CAP_SUP_WIDTH_20_40;
+	ht_cap.cap |= IEEE80211_HT_CAP_DSSSCCK40;
+	ht_cap.cap |= WLAN_HT_CAP_SM_PS_STATIC << IEEE80211_HT_CAP_SM_PS_SHIFT;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_HT20_SGI)
+		ht_cap.cap |= IEEE80211_HT_CAP_SGI_20;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_HT40_SGI)
+		ht_cap.cap |= IEEE80211_HT_CAP_SGI_40;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_DYNAMIC_SMPS) {
+		u32 smps;
+
+		smps   = WLAN_HT_CAP_SM_PS_DYNAMIC;
+		smps <<= IEEE80211_HT_CAP_SM_PS_SHIFT;
+
+		ht_cap.cap |= smps;
+	}
+
+	if (ar->ht_cap_info & WMI_HT_CAP_TX_STBC)
+		ht_cap.cap |= IEEE80211_HT_CAP_TX_STBC;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_RX_STBC) {
+		u32 stbc;
+
+		stbc   = ar->ht_cap_info;
+		stbc  &= WMI_HT_CAP_RX_STBC;
+		stbc >>= WMI_HT_CAP_RX_STBC_MASK_SHIFT;
+		stbc <<= IEEE80211_HT_CAP_RX_STBC_SHIFT;
+		stbc  &= IEEE80211_HT_CAP_RX_STBC;
+
+		ht_cap.cap |= stbc;
+	}
+
+	if (ar->ht_cap_info & WMI_HT_CAP_LDPC)
+		ht_cap.cap |= IEEE80211_HT_CAP_LDPC_CODING;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_L_SIG_TXOP_PROT)
+		ht_cap.cap |= IEEE80211_HT_CAP_LSIG_TXOP_PROT;
+
+	/* max AMSDU is implicitly taken from vht_cap_info */
+	if (ar->vht_cap_info & WMI_VHT_CAP_MAX_MPDU_LEN_MASK)
+		ht_cap.cap |= IEEE80211_HT_CAP_MAX_AMSDU;
+
+	for (i = 0; i < WMI_MAX_SPATIAL_STREAM; i++)
+		ht_cap.mcs.rx_mask[i] = 0xFF;
+
+	ht_cap.mcs.tx_params |= IEEE80211_HT_MCS_TX_DEFINED;
+
+	return ht_cap;
+}
+
+
+static void ath10k_get_arvif_iter(void *data, u8 *mac,
+				  struct ieee80211_vif *vif)
+{
+	struct ath10k_vif_iter *arvif_iter = data;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+
+	if (arvif->vdev_id == arvif_iter->vdev_id)
+		arvif_iter->arvif = arvif;
+}
+
+struct ath10k_vif *ath10k_get_arvif(struct ath10k *ar, u32 vdev_id)
+{
+	struct ath10k_vif_iter arvif_iter;
+	u32 flags;
+
+	memset(&arvif_iter, 0, sizeof(struct ath10k_vif_iter));
+	arvif_iter.vdev_id = vdev_id;
+
+	flags = IEEE80211_IFACE_ITER_RESUME_ALL;
+	ieee80211_iterate_active_interfaces_atomic(ar->hw,
+						   flags,
+						   ath10k_get_arvif_iter,
+						   &arvif_iter);
+	if (!arvif_iter.arvif) {
+		ath10k_warn("No VIF found for VDEV: %d\n", vdev_id);
+		return NULL;
+	}
+
+	return arvif_iter.arvif;
+}
+
+int ath10k_mac_register(struct ath10k *ar)
+{
+	struct ieee80211_supported_band *band;
+	struct ieee80211_sta_vht_cap vht_cap;
+	struct ieee80211_sta_ht_cap ht_cap;
+	void *channels;
+	int ret;
+
+	SET_IEEE80211_PERM_ADDR(ar->hw, ar->mac_addr);
+
+	SET_IEEE80211_DEV(ar->hw, ar->dev);
+
+	ht_cap = ath10k_get_ht_cap(ar);
+	vht_cap = ath10k_create_vht_cap(ar);
+
+	if (ar->phy_capability & WHAL_WLAN_11G_CAPABILITY) {
+		channels = kmemdup(ath10k_2ghz_channels,
+				   sizeof(ath10k_2ghz_channels),
+				   GFP_KERNEL);
+		if (!channels)
+			return -ENOMEM;
+
+		band = &ar->mac.sbands[IEEE80211_BAND_2GHZ];
+		band->n_channels = ARRAY_SIZE(ath10k_2ghz_channels);
+		band->channels = channels;
+		band->n_bitrates = ath10k_g_rates_size;
+		band->bitrates = ath10k_g_rates;
+		band->ht_cap = ht_cap;
+
+		/* vht is not supported in 2.4 GHz */
+
+		ar->hw->wiphy->bands[IEEE80211_BAND_2GHZ] = band;
+	}
+
+	if (ar->phy_capability & WHAL_WLAN_11A_CAPABILITY) {
+		channels = kmemdup(ath10k_5ghz_channels,
+				   sizeof(ath10k_5ghz_channels),
+				   GFP_KERNEL);
+		if (!channels) {
+			if (ar->phy_capability & WHAL_WLAN_11G_CAPABILITY) {
+				band = &ar->mac.sbands[IEEE80211_BAND_2GHZ];
+				kfree(band->channels);
+			}
+			return -ENOMEM;
+		}
+
+		band = &ar->mac.sbands[IEEE80211_BAND_5GHZ];
+		band->n_channels = ARRAY_SIZE(ath10k_5ghz_channels);
+		band->channels = channels;
+		band->n_bitrates = ath10k_a_rates_size;
+		band->bitrates = ath10k_a_rates;
+		band->ht_cap = ht_cap;
+		band->vht_cap = vht_cap;
+		ar->hw->wiphy->bands[IEEE80211_BAND_5GHZ] = band;
+	}
+
+	ar->hw->wiphy->interface_modes =
+		BIT(NL80211_IFTYPE_STATION) |
+		BIT(NL80211_IFTYPE_ADHOC) |
+		BIT(NL80211_IFTYPE_AP) |
+		BIT(NL80211_IFTYPE_P2P_CLIENT) |
+		BIT(NL80211_IFTYPE_P2P_GO);
+
+	ar->hw->flags = IEEE80211_HW_SIGNAL_DBM |
+			IEEE80211_HW_SUPPORTS_PS |
+			IEEE80211_HW_SUPPORTS_DYNAMIC_PS |
+			IEEE80211_HW_SUPPORTS_UAPSD |
+			IEEE80211_HW_MFP_CAPABLE |
+			IEEE80211_HW_REPORTS_TX_ACK_STATUS |
+			IEEE80211_HW_HAS_RATE_CONTROL |
+			IEEE80211_HW_SUPPORTS_STATIC_SMPS |
+			IEEE80211_HW_WANT_MONITOR_VIF |
+			IEEE80211_HW_AP_LINK_PS;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_DYNAMIC_SMPS)
+		ar->hw->flags |= IEEE80211_HW_SUPPORTS_DYNAMIC_SMPS;
+
+	if (ar->ht_cap_info & WMI_HT_CAP_ENABLED) {
+		ar->hw->flags |= IEEE80211_HW_AMPDU_AGGREGATION;
+		ar->hw->flags |= IEEE80211_HW_TX_AMPDU_SETUP_IN_HW;
+	}
+
+	ar->hw->wiphy->max_scan_ssids = WLAN_SCAN_PARAMS_MAX_SSID;
+	ar->hw->wiphy->max_scan_ie_len = WLAN_SCAN_PARAMS_MAX_IE_LEN;
+
+	ar->hw->vif_data_size = sizeof(struct ath10k_vif);
+
+	ar->hw->channel_change_time = 5000;
+	ar->hw->max_listen_interval = ATH10K_MAX_HW_LISTEN_INTERVAL;
+
+	ar->hw->wiphy->flags |= WIPHY_FLAG_HAS_REMAIN_ON_CHANNEL;
+	ar->hw->wiphy->max_remain_on_channel_duration = 5000;
+
+	ar->hw->wiphy->flags |= WIPHY_FLAG_AP_UAPSD;
+	/*
+	 * on LL hardware queues are managed entirely by the FW
+	 * so we only advertise to mac we can do the queues thing
+	 */
+	ar->hw->queues = 4;
+
+	ar->hw->wiphy->iface_combinations = &ath10k_if_comb;
+	ar->hw->wiphy->n_iface_combinations = 1;
+
+	ret = ath_regd_init(&ar->ath_common.regulatory, ar->hw->wiphy,
+			    ath10k_reg_notifier);
+	if (ret) {
+		ath10k_err("Regulatory initialization failed\n");
+		return ret;
+	}
+
+	ret = ieee80211_register_hw(ar->hw);
+	if (ret) {
+		ath10k_err("ieee80211 registration failed: %d\n", ret);
+		return ret;
+	}
+
+	if (!ath_is_world_regd(&ar->ath_common.regulatory)) {
+		ret = regulatory_hint(ar->hw->wiphy,
+				      ar->ath_common.regulatory.alpha2);
+		if (ret)
+			goto exit;
+	}
+
+	return 0;
+exit:
+	ieee80211_unregister_hw(ar->hw);
+	return ret;
+}
+
+void ath10k_mac_unregister(struct ath10k *ar)
+{
+	ieee80211_unregister_hw(ar->hw);
+
+	kfree(ar->mac.sbands[IEEE80211_BAND_2GHZ].channels);
+	kfree(ar->mac.sbands[IEEE80211_BAND_5GHZ].channels);
+
+	SET_IEEE80211_DEV(ar->hw, NULL);
+}
diff --git a/drivers/net/wireless/ath/ath10k/mac.h b/drivers/net/wireless/ath/ath10k/mac.h
new file mode 100644
index 0000000..27fc92e
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/mac.h
@@ -0,0 +1,61 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _MAC_H_
+#define _MAC_H_
+
+#include <net/mac80211.h>
+#include "core.h"
+
+struct ath10k_generic_iter {
+	struct ath10k *ar;
+	int ret;
+};
+
+struct ath10k *ath10k_mac_create(void);
+void ath10k_mac_destroy(struct ath10k *ar);
+int ath10k_mac_register(struct ath10k *ar);
+void ath10k_mac_unregister(struct ath10k *ar);
+struct ath10k_vif *ath10k_get_arvif(struct ath10k *ar, u32 vdev_id);
+void ath10k_reset_scan(unsigned long ptr);
+void ath10k_offchan_tx_purge(struct ath10k *ar);
+void ath10k_offchan_tx_work(struct work_struct *work);
+
+static inline struct ath10k_vif *ath10k_vif_to_arvif(struct ieee80211_vif *vif)
+{
+	return (struct ath10k_vif *)vif->drv_priv;
+}
+
+static inline void ath10k_tx_h_seq_no(struct sk_buff *skb)
+{
+	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
+	struct ieee80211_vif *vif = info->control.vif;
+	struct ath10k_vif *arvif = ath10k_vif_to_arvif(vif);
+
+	if (info->flags  & IEEE80211_TX_CTL_ASSIGN_SEQ) {
+		if (arvif->tx_seq_no == 0)
+			arvif->tx_seq_no = 0x1000;
+
+		if (info->flags & IEEE80211_TX_CTL_FIRST_FRAGMENT)
+			arvif->tx_seq_no += 0x10;
+		hdr->seq_ctrl &= cpu_to_le16(IEEE80211_SCTL_FRAG);
+		hdr->seq_ctrl |= cpu_to_le16(arvif->tx_seq_no);
+	}
+}
+
+#endif /* _MAC_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/pci.c b/drivers/net/wireless/ath/ath10k/pci.c
new file mode 100644
index 0000000..33af467
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/pci.c
@@ -0,0 +1,2507 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+
+#include "core.h"
+#include "debug.h"
+
+#include "targaddrs.h"
+#include "bmi.h"
+
+#include "hif.h"
+#include "htc.h"
+
+#include "ce.h"
+#include "pci.h"
+
+unsigned int ath10k_target_ps;
+module_param(ath10k_target_ps, uint, 0644);
+MODULE_PARM_DESC(ath10k_target_ps, "Enable ath10k Target (SoC) PS option");
+
+#define QCA988X_1_0_DEVICE_ID	(0xabcd)
+#define QCA988X_2_0_DEVICE_ID	(0x003c)
+
+static DEFINE_PCI_DEVICE_TABLE(ath10k_pci_id_table) = {
+	{ PCI_VDEVICE(ATHEROS, QCA988X_1_0_DEVICE_ID) }, /* PCI-E QCA988X V1 */
+	{ PCI_VDEVICE(ATHEROS, QCA988X_2_0_DEVICE_ID) }, /* PCI-E QCA988X V2 */
+	{0}
+};
+
+static int ath10k_pci_diag_read_access(struct ath10k *ar, u32 address,
+				       u32 *data);
+
+static void ath10k_pci_process_ce(struct ath10k *ar);
+static int ath10k_pci_post_rx(struct ath10k *ar);
+static int ath10k_pci_post_rx_pipe(struct hif_ce_pipe_info *pipe_info,
+					     int num);
+static void ath10k_pci_rx_pipe_cleanup(struct hif_ce_pipe_info *pipe_info);
+static void ath10k_pci_stop_ce(struct ath10k *ar);
+
+static const struct ce_attr host_ce_config_wlan[] = {
+	/* host->target HTC control and raw streams */
+	{ /* CE0 */ CE_ATTR_FLAGS, 0, 16, 256, 0, NULL,},
+	/* could be moved to share CE3 */
+	/* target->host HTT + HTC control */
+	{ /* CE1 */ CE_ATTR_FLAGS, 0, 0, 512, 512, NULL,},
+	/* target->host WMI */
+	{ /* CE2 */ CE_ATTR_FLAGS, 0, 0, 2048, 32, NULL,},
+	/* host->target WMI */
+	{ /* CE3 */ CE_ATTR_FLAGS, 0, 32, 2048, 0, NULL,},
+	/* host->target HTT */
+	{ /* CE4 */ CE_ATTR_FLAGS | CE_ATTR_DIS_INTR, 0,
+		    CE_HTT_H2T_MSG_SRC_NENTRIES, 256, 0, NULL,},
+	/* unused */
+	{ /* CE5 */ CE_ATTR_FLAGS, 0, 0, 0, 0, NULL,},
+	/* Target autonomous hif_memcpy */
+	{ /* CE6 */ CE_ATTR_FLAGS, 0, 0, 0, 0, NULL,},
+	/* ce_diag, the Diagnostic Window */
+	{ /* CE7 */ CE_ATTR_FLAGS, 0, 2, DIAG_TRANSFER_LIMIT, 2, NULL,},
+};
+
+/* Target firmware's Copy Engine configuration. */
+static const struct ce_pipe_config target_ce_config_wlan[] = {
+	/* host->target HTC control and raw streams */
+	{ /* CE0 */ 0, PIPEDIR_OUT, 32, 256, CE_ATTR_FLAGS, 0,},
+	/* target->host HTT + HTC control */
+	{ /* CE1 */ 1, PIPEDIR_IN, 32, 512, CE_ATTR_FLAGS, 0,},
+	/* target->host WMI */
+	{ /* CE2 */ 2, PIPEDIR_IN, 32, 2048, CE_ATTR_FLAGS, 0,},
+	/* host->target WMI */
+	{ /* CE3 */ 3, PIPEDIR_OUT, 32, 2048, CE_ATTR_FLAGS, 0,},
+	/* host->target HTT */
+	{ /* CE4 */ 4, PIPEDIR_OUT, 256, 256, CE_ATTR_FLAGS, 0,},
+	/* NB: 50% of src nentries, since tx has 2 frags */
+	/* unused */
+	{ /* CE5 */ 5, PIPEDIR_OUT, 32, 2048, CE_ATTR_FLAGS, 0,},
+	/* Reserved for target autonomous hif_memcpy */
+	{ /* CE6 */ 6, PIPEDIR_INOUT, 32, 4096, CE_ATTR_FLAGS, 0,},
+	/* CE7 used only by Host */
+};
+
+/*
+ * Diagnostic read/write access is provided for startup/config/debug usage.
+ * Caller must guarantee proper alignment, when applicable, and single user
+ * at any moment.
+ */
+static int ath10k_pci_diag_read_mem(struct ath10k *ar, u32 address, void *data,
+				    int nbytes)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret = 0;
+	u32 buf;
+	unsigned int completed_nbytes, orig_nbytes, remaining_bytes;
+	unsigned int id;
+	unsigned int flags;
+	struct ce_state *ce_diag;
+	/* Host buffer address in CE space */
+	u32 ce_data;
+	dma_addr_t ce_data_base = 0;
+	void *data_buf = NULL;
+	int i;
+
+	/*
+	 * This code cannot handle reads to non-memory space. Redirect to the
+	 * register read fn but preserve the multi word read capability of
+	 * this fn
+	 */
+	if (address < DRAM_BASE_ADDRESS) {
+		if (!IS_ALIGNED(address, 4) ||
+		    !IS_ALIGNED((unsigned long)data, 4))
+			return -EIO;
+
+		while ((nbytes >= 4) &&  ((ret = ath10k_pci_diag_read_access(
+					   ar, address, (u32 *)data)) == 0)) {
+			nbytes -= sizeof(u32);
+			address += sizeof(u32);
+			data += sizeof(u32);
+		}
+		return ret;
+	}
+
+	ce_diag = ar_pci->ce_diag;
+
+	/*
+	 * Allocate a temporary bounce buffer to hold caller's data
+	 * to be DMA'ed from Target. This guarantees
+	 *   1) 4-byte alignment
+	 *   2) Buffer in DMA-able space
+	 */
+	orig_nbytes = nbytes;
+	data_buf = (unsigned char *)pci_alloc_consistent(ar_pci->pdev,
+							 orig_nbytes,
+							 &ce_data_base);
+
+	if (!data_buf) {
+		ret = -ENOMEM;
+		goto done;
+	}
+	memset(data_buf, 0, orig_nbytes);
+
+	remaining_bytes = orig_nbytes;
+	ce_data = ce_data_base;
+	while (remaining_bytes) {
+		nbytes = min_t(unsigned int, remaining_bytes,
+			       DIAG_TRANSFER_LIMIT);
+
+		ret = ath10k_ce_recv_buf_enqueue(ce_diag, NULL, ce_data);
+		if (ret != 0)
+			goto done;
+
+		/* Request CE to send from Target(!) address to Host buffer */
+		/*
+		 * The address supplied by the caller is in the
+		 * Target CPU virtual address space.
+		 *
+		 * In order to use this address with the diagnostic CE,
+		 * convert it from Target CPU virtual address space
+		 * to CE address space
+		 */
+		ath10k_pci_wake(ar);
+		address = TARG_CPU_SPACE_TO_CE_SPACE(ar, ar_pci->mem,
+						     address);
+		ath10k_pci_sleep(ar);
+
+		ret = ath10k_ce_send(ce_diag, NULL, (u32)address, nbytes, 0,
+				 0);
+		if (ret)
+			goto done;
+
+		i = 0;
+		while (ath10k_ce_completed_send_next(ce_diag, NULL, &buf,
+						     &completed_nbytes,
+						     &id) != 0) {
+			mdelay(1);
+			if (i++ > DIAG_ACCESS_CE_TIMEOUT_MS) {
+				ret = -EBUSY;
+				goto done;
+			}
+		}
+
+		if (nbytes != completed_nbytes) {
+			ret = -EIO;
+			goto done;
+		}
+
+		if (buf != (u32) address) {
+			ret = -EIO;
+			goto done;
+		}
+
+		i = 0;
+		while (ath10k_ce_completed_recv_next(ce_diag, NULL, &buf,
+						     &completed_nbytes,
+						     &id, &flags) != 0) {
+			mdelay(1);
+
+			if (i++ > DIAG_ACCESS_CE_TIMEOUT_MS) {
+				ret = -EBUSY;
+				goto done;
+			}
+		}
+
+		if (nbytes != completed_nbytes) {
+			ret = -EIO;
+			goto done;
+		}
+
+		if (buf != ce_data) {
+			ret = -EIO;
+			goto done;
+		}
+
+		remaining_bytes -= nbytes;
+		address += nbytes;
+		ce_data += nbytes;
+	}
+
+done:
+	if (ret == 0) {
+		/* Copy data from allocated DMA buf to caller's buf */
+		WARN_ON_ONCE(orig_nbytes & 3);
+		for (i = 0; i < orig_nbytes / sizeof(__le32); i++) {
+			((u32 *)data)[i] =
+				__le32_to_cpu(((__le32 *)data_buf)[i]);
+		}
+	} else
+		ath10k_dbg(ATH10K_DBG_PCI, "%s failure (0x%x)\n",
+			   __func__, address);
+
+	if (data_buf)
+		pci_free_consistent(ar_pci->pdev, orig_nbytes,
+				    data_buf, ce_data_base);
+
+	return ret;
+}
+
+/* Read 4-byte aligned data from Target memory or register */
+static int ath10k_pci_diag_read_access(struct ath10k *ar, u32 address,
+				       u32 *data)
+{
+	/* Assume range doesn't cross this boundary */
+	if (address >= DRAM_BASE_ADDRESS)
+		return ath10k_pci_diag_read_mem(ar, address, data, sizeof(u32));
+
+	ath10k_pci_wake(ar);
+	*data = ath10k_pci_read32(ar, address);
+	ath10k_pci_sleep(ar);
+	return 0;
+}
+
+static int ath10k_pci_diag_write_mem(struct ath10k *ar, u32 address,
+				     const void *data, int nbytes)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret = 0;
+	u32 buf;
+	unsigned int completed_nbytes, orig_nbytes, remaining_bytes;
+	unsigned int id;
+	unsigned int flags;
+	struct ce_state *ce_diag;
+	void *data_buf = NULL;
+	u32 ce_data;	/* Host buffer address in CE space */
+	dma_addr_t ce_data_base = 0;
+	int i;
+
+	ce_diag = ar_pci->ce_diag;
+
+	/*
+	 * Allocate a temporary bounce buffer to hold caller's data
+	 * to be DMA'ed to Target. This guarantees
+	 *   1) 4-byte alignment
+	 *   2) Buffer in DMA-able space
+	 */
+	orig_nbytes = nbytes;
+	data_buf = (unsigned char *)pci_alloc_consistent(ar_pci->pdev,
+							 orig_nbytes,
+							 &ce_data_base);
+	if (!data_buf) {
+		ret = -ENOMEM;
+		goto done;
+	}
+
+	/* Copy caller's data to allocated DMA buf */
+	WARN_ON_ONCE(orig_nbytes & 3);
+	for (i = 0; i < orig_nbytes / sizeof(__le32); i++)
+		((__le32 *)data_buf)[i] = __cpu_to_le32(((u32 *)data)[i]);
+
+	/*
+	 * The address supplied by the caller is in the
+	 * Target CPU virtual address space.
+	 *
+	 * In order to use this address with the diagnostic CE,
+	 * convert it from
+	 *    Target CPU virtual address space
+	 * to
+	 *    CE address space
+	 */
+	ath10k_pci_wake(ar);
+	address = TARG_CPU_SPACE_TO_CE_SPACE(ar, ar_pci->mem, address);
+	ath10k_pci_sleep(ar);
+
+	remaining_bytes = orig_nbytes;
+	ce_data = ce_data_base;
+	while (remaining_bytes) {
+		/* FIXME: check cast */
+		nbytes = min_t(int, remaining_bytes, DIAG_TRANSFER_LIMIT);
+
+		/* Set up to receive directly into Target(!) address */
+		ret = ath10k_ce_recv_buf_enqueue(ce_diag, NULL, address);
+		if (ret != 0)
+			goto done;
+
+		/*
+		 * Request CE to send caller-supplied data that
+		 * was copied to bounce buffer to Target(!) address.
+		 */
+		ret = ath10k_ce_send(ce_diag, NULL, (u32) ce_data,
+				     nbytes, 0, 0);
+		if (ret != 0)
+			goto done;
+
+		i = 0;
+		while (ath10k_ce_completed_send_next(ce_diag, NULL, &buf,
+						     &completed_nbytes,
+						     &id) != 0) {
+			mdelay(1);
+
+			if (i++ > DIAG_ACCESS_CE_TIMEOUT_MS) {
+				ret = -EBUSY;
+				goto done;
+			}
+		}
+
+		if (nbytes != completed_nbytes) {
+			ret = -EIO;
+			goto done;
+		}
+
+		if (buf != ce_data) {
+			ret = -EIO;
+			goto done;
+		}
+
+		i = 0;
+		while (ath10k_ce_completed_recv_next(ce_diag, NULL, &buf,
+						     &completed_nbytes,
+						     &id, &flags) != 0) {
+			mdelay(1);
+
+			if (i++ > DIAG_ACCESS_CE_TIMEOUT_MS) {
+				ret = -EBUSY;
+				goto done;
+			}
+		}
+
+		if (nbytes != completed_nbytes) {
+			ret = -EIO;
+			goto done;
+		}
+
+		if (buf != address) {
+			ret = -EIO;
+			goto done;
+		}
+
+		remaining_bytes -= nbytes;
+		address += nbytes;
+		ce_data += nbytes;
+	}
+
+done:
+	if (data_buf) {
+		pci_free_consistent(ar_pci->pdev, orig_nbytes, data_buf,
+				    ce_data_base);
+	}
+
+	if (ret != 0)
+		ath10k_dbg(ATH10K_DBG_PCI, "%s failure (0x%x)\n", __func__,
+			   address);
+
+	return ret;
+}
+
+/* Write 4B data to Target memory or register */
+static int ath10k_pci_diag_write_access(struct ath10k *ar, u32 address,
+					u32 data)
+{
+	/* Assume range doesn't cross this boundary */
+	if (address >= DRAM_BASE_ADDRESS)
+		return ath10k_pci_diag_write_mem(ar, address, &data,
+						 sizeof(u32));
+
+	ath10k_pci_wake(ar);
+	ath10k_pci_write32(ar, address, data);
+	ath10k_pci_sleep(ar);
+	return 0;
+}
+
+static bool ath10k_pci_target_is_awake(struct ath10k *ar)
+{
+	void __iomem *mem = ath10k_pci_priv(ar)->mem;
+	u32 val;
+	val = ioread32(mem + PCIE_LOCAL_BASE_ADDRESS +
+		       RTC_STATE_ADDRESS);
+	return (RTC_STATE_V_GET(val) == RTC_STATE_V_ON);
+}
+
+static void ath10k_pci_wait(struct ath10k *ar)
+{
+	int n = 100;
+
+	while (n-- && !ath10k_pci_target_is_awake(ar))
+		msleep(10);
+
+	if (n < 0)
+		ath10k_warn("Unable to wakeup target\n");
+}
+
+void ath10k_do_pci_wake(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	void __iomem *pci_addr = ar_pci->mem;
+	int tot_delay = 0;
+	int curr_delay = 5;
+
+	if (atomic_read(&ar_pci->keep_awake_count) == 0) {
+		/* Force AWAKE */
+		iowrite32(PCIE_SOC_WAKE_V_MASK,
+			  pci_addr + PCIE_LOCAL_BASE_ADDRESS +
+			  PCIE_SOC_WAKE_ADDRESS);
+	}
+	atomic_inc(&ar_pci->keep_awake_count);
+
+	if (ar_pci->verified_awake)
+		return;
+
+	for (;;) {
+		if (ath10k_pci_target_is_awake(ar)) {
+			ar_pci->verified_awake = true;
+			break;
+		}
+
+		if (tot_delay > PCIE_WAKE_TIMEOUT) {
+			ath10k_warn("target takes too long to wake up (awake count %d)\n",
+				    atomic_read(&ar_pci->keep_awake_count));
+			break;
+		}
+
+		udelay(curr_delay);
+		tot_delay += curr_delay;
+
+		if (curr_delay < 50)
+			curr_delay += 5;
+	}
+}
+
+void ath10k_do_pci_sleep(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	void __iomem *pci_addr = ar_pci->mem;
+
+	if (atomic_dec_and_test(&ar_pci->keep_awake_count)) {
+		/* Allow sleep */
+		ar_pci->verified_awake = false;
+		iowrite32(PCIE_SOC_WAKE_RESET,
+			  pci_addr + PCIE_LOCAL_BASE_ADDRESS +
+			  PCIE_SOC_WAKE_ADDRESS);
+	}
+}
+
+/*
+ * FIXME: Handle OOM properly.
+ */
+static inline
+struct ath10k_pci_compl *get_free_compl(struct hif_ce_pipe_info *pipe_info)
+{
+	struct ath10k_pci_compl *compl = NULL;
+
+	spin_lock_bh(&pipe_info->pipe_lock);
+	if (list_empty(&pipe_info->compl_free)) {
+		ath10k_warn("Completion buffers are full\n");
+		goto exit;
+	}
+	compl = list_first_entry(&pipe_info->compl_free,
+				 struct ath10k_pci_compl, list);
+	list_del(&compl->list);
+exit:
+	spin_unlock_bh(&pipe_info->pipe_lock);
+	return compl;
+}
+
+/* Called by lower (CE) layer when a send to Target completes. */
+static void ath10k_pci_ce_send_done(struct ce_state *ce_state,
+				    void *transfer_context,
+				    u32 ce_data,
+				    unsigned int nbytes,
+				    unsigned int transfer_id)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info =  &ar_pci->pipe_info[ce_state->id];
+	struct ath10k_pci_compl *compl;
+	bool process = false;
+
+	do {
+		/*
+		 * For the send completion of an item in sendlist, just
+		 * increment num_sends_allowed. The upper layer callback will
+		 * be triggered when last fragment is done with send.
+		 */
+		if (transfer_context == CE_SENDLIST_ITEM_CTXT) {
+			spin_lock_bh(&pipe_info->pipe_lock);
+			pipe_info->num_sends_allowed++;
+			spin_unlock_bh(&pipe_info->pipe_lock);
+			continue;
+		}
+
+		compl = get_free_compl(pipe_info);
+		if (!compl)
+			break;
+
+		compl->send_or_recv = HIF_CE_COMPLETE_SEND;
+		compl->ce_state = ce_state;
+		compl->pipe_info = pipe_info;
+		compl->transfer_context = transfer_context;
+		compl->nbytes = nbytes;
+		compl->transfer_id = transfer_id;
+		compl->flags = 0;
+
+		/*
+		 * Add the completion to the processing queue.
+		 */
+		spin_lock_bh(&ar_pci->compl_lock);
+		list_add_tail(&compl->list, &ar_pci->compl_process);
+		spin_unlock_bh(&ar_pci->compl_lock);
+
+		process = true;
+	} while (ath10k_ce_completed_send_next(ce_state,
+							   &transfer_context,
+							   &ce_data, &nbytes,
+							   &transfer_id) == 0);
+
+	/*
+	 * If only some of the items within a sendlist have completed,
+	 * don't invoke completion processing until the entire sendlist
+	 * has been sent.
+	 */
+	if (!process)
+		return;
+
+	ath10k_pci_process_ce(ar);
+}
+
+/* Called by lower (CE) layer when data is received from the Target. */
+static void ath10k_pci_ce_recv_data(struct ce_state *ce_state,
+				    void *transfer_context, u32 ce_data,
+				    unsigned int nbytes,
+				    unsigned int transfer_id,
+				    unsigned int flags)
+{
+	struct ath10k *ar = ce_state->ar;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info =  &ar_pci->pipe_info[ce_state->id];
+	struct ath10k_pci_compl *compl;
+	struct sk_buff *skb;
+
+	do {
+		compl = get_free_compl(pipe_info);
+		if (!compl)
+			break;
+
+		compl->send_or_recv = HIF_CE_COMPLETE_RECV;
+		compl->ce_state = ce_state;
+		compl->pipe_info = pipe_info;
+		compl->transfer_context = transfer_context;
+		compl->nbytes = nbytes;
+		compl->transfer_id = transfer_id;
+		compl->flags = flags;
+
+		skb = transfer_context;
+		dma_unmap_single(ar->dev, ATH10K_SKB_CB(skb)->paddr,
+				 skb->len + skb_tailroom(skb),
+				 DMA_FROM_DEVICE);
+		/*
+		 * Add the completion to the processing queue.
+		 */
+		spin_lock_bh(&ar_pci->compl_lock);
+		list_add_tail(&compl->list, &ar_pci->compl_process);
+		spin_unlock_bh(&ar_pci->compl_lock);
+
+	} while (ath10k_ce_completed_recv_next(ce_state,
+							   &transfer_context,
+							   &ce_data, &nbytes,
+							   &transfer_id,
+							   &flags) == 0);
+
+	ath10k_pci_process_ce(ar);
+}
+
+/* Send the first nbytes bytes of the buffer */
+static int ath10k_pci_hif_send_head(struct ath10k *ar, u8 pipe_id,
+				    unsigned int transfer_id,
+				    unsigned int bytes, struct sk_buff *nbuf)
+{
+	struct ath10k_skb_cb *skb_cb = ATH10K_SKB_CB(nbuf);
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info = &(ar_pci->pipe_info[pipe_id]);
+	struct ce_state *ce_hdl = pipe_info->ce_hdl;
+	struct ce_sendlist sendlist;
+	unsigned int len;
+	u32 flags = 0;
+	int ret;
+
+	memset(&sendlist, 0, sizeof(struct ce_sendlist));
+
+	len = min(bytes, nbuf->len);
+	bytes -= len;
+
+	if (len & 3)
+		ath10k_warn("skb not aligned to 4-byte boundary (%d)\n", len);
+
+	ath10k_dbg(ATH10K_DBG_PCI,
+		   "pci send data vaddr %p paddr 0x%llx len %d as %d bytes\n",
+		   nbuf->data, (unsigned long long) skb_cb->paddr,
+		   nbuf->len, len);
+	ath10k_dbg_dump(ATH10K_DBG_PCI_DUMP, NULL,
+			"ath10k tx: data: ",
+			nbuf->data, nbuf->len);
+
+	ath10k_ce_sendlist_buf_add(&sendlist, skb_cb->paddr, len, flags);
+
+	/* Make sure we have resources to handle this request */
+	spin_lock_bh(&pipe_info->pipe_lock);
+	if (!pipe_info->num_sends_allowed) {
+		ath10k_warn("Pipe: %d is full\n", pipe_id);
+		spin_unlock_bh(&pipe_info->pipe_lock);
+		return -ENOSR;
+	}
+	pipe_info->num_sends_allowed--;
+	spin_unlock_bh(&pipe_info->pipe_lock);
+
+	ret = ath10k_ce_sendlist_send(ce_hdl, nbuf, &sendlist, transfer_id);
+	if (ret)
+		ath10k_warn("CE send failed: %p\n", nbuf);
+
+	return ret;
+}
+
+static u16 ath10k_pci_hif_get_free_queue_number(struct ath10k *ar, u8 pipe)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info = &(ar_pci->pipe_info[pipe]);
+	int ret;
+
+	spin_lock_bh(&pipe_info->pipe_lock);
+	ret = pipe_info->num_sends_allowed;
+	spin_unlock_bh(&pipe_info->pipe_lock);
+
+	return ret;
+}
+
+static void ath10k_pci_hif_dump_area(struct ath10k *ar)
+{
+	u32 reg_dump_area = 0;
+	u32 reg_dump_values[REG_DUMP_COUNT_QCA988X] = {};
+	u32 host_addr;
+	int ret;
+	u32 i;
+
+	ath10k_err("firmware crashed!\n");
+	ath10k_err("hardware name %s version 0x%x\n",
+		   ar->hw_params.name, ar->target_version);
+	ath10k_err("firmware version: %u.%u.%u.%u\n", ar->fw_version_major,
+		   ar->fw_version_minor, ar->fw_version_release,
+		   ar->fw_version_build);
+
+	host_addr = host_interest_item_address(HI_ITEM(hi_failure_state));
+	if (ath10k_pci_diag_read_mem(ar, host_addr,
+				     &reg_dump_area, sizeof(u32)) != 0) {
+		ath10k_warn("could not read hi_failure_state\n");
+		return;
+	}
+
+	ath10k_err("target register Dump Location: 0x%08X\n", reg_dump_area);
+
+	ret = ath10k_pci_diag_read_mem(ar, reg_dump_area,
+				       &reg_dump_values[0],
+				       REG_DUMP_COUNT_QCA988X * sizeof(u32));
+	if (ret != 0) {
+		ath10k_err("could not dump FW Dump Area\n");
+		return;
+	}
+
+	BUILD_BUG_ON(REG_DUMP_COUNT_QCA988X % 4);
+
+	ath10k_err("target Register Dump\n");
+	for (i = 0; i < REG_DUMP_COUNT_QCA988X; i += 4)
+		ath10k_err("[%02d]: 0x%08X 0x%08X 0x%08X 0x%08X\n",
+			   i,
+			   reg_dump_values[i],
+			   reg_dump_values[i + 1],
+			   reg_dump_values[i + 2],
+			   reg_dump_values[i + 3]);
+}
+
+static void ath10k_pci_hif_send_complete_check(struct ath10k *ar, u8 pipe,
+					       int force)
+{
+	if (!force) {
+		int resources;
+		/*
+		 * Decide whether to actually poll for completions, or just
+		 * wait for a later chance.
+		 * If there seem to be plenty of resources left, then just wait
+		 * since checking involves reading a CE register, which is a
+		 * relatively expensive operation.
+		 */
+		resources = ath10k_pci_hif_get_free_queue_number(ar, pipe);
+
+		/*
+		 * If at least 50% of the total resources are still available,
+		 * don't bother checking again yet.
+		 */
+		if (resources > (host_ce_config_wlan[pipe].src_nentries >> 1))
+			return;
+	}
+	ath10k_ce_per_engine_service(ar, pipe);
+}
+
+static void ath10k_pci_hif_post_init(struct ath10k *ar,
+				     struct ath10k_hif_cb *callbacks)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	ath10k_dbg(ATH10K_DBG_PCI, "%s\n", __func__);
+
+	memcpy(&ar_pci->msg_callbacks_current, callbacks,
+	       sizeof(ar_pci->msg_callbacks_current));
+}
+
+static int ath10k_pci_start_ce(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_state *ce_diag = ar_pci->ce_diag;
+	const struct ce_attr *attr;
+	struct hif_ce_pipe_info *pipe_info;
+	struct ath10k_pci_compl *compl;
+	int i, pipe_num, completions, disable_interrupts;
+
+	spin_lock_init(&ar_pci->compl_lock);
+	INIT_LIST_HEAD(&ar_pci->compl_process);
+
+	for (pipe_num = 0; pipe_num < ar_pci->ce_count; pipe_num++) {
+		pipe_info = &ar_pci->pipe_info[pipe_num];
+
+		spin_lock_init(&pipe_info->pipe_lock);
+		INIT_LIST_HEAD(&pipe_info->compl_free);
+
+		/* Handle Diagnostic CE specially */
+		if (pipe_info->ce_hdl == ce_diag)
+			continue;
+
+		attr = &host_ce_config_wlan[pipe_num];
+		completions = 0;
+
+		if (attr->src_nentries) {
+			disable_interrupts = attr->flags & CE_ATTR_DIS_INTR;
+			ath10k_ce_send_cb_register(pipe_info->ce_hdl,
+						   ath10k_pci_ce_send_done,
+						   disable_interrupts);
+			completions += attr->src_nentries;
+			pipe_info->num_sends_allowed = attr->src_nentries - 1;
+		}
+
+		if (attr->dest_nentries) {
+			ath10k_ce_recv_cb_register(pipe_info->ce_hdl,
+						   ath10k_pci_ce_recv_data);
+			completions += attr->dest_nentries;
+		}
+
+		if (completions == 0)
+			continue;
+
+		for (i = 0; i < completions; i++) {
+			compl = kmalloc(sizeof(struct ath10k_pci_compl),
+					GFP_KERNEL);
+			if (!compl) {
+				ath10k_warn("No memory for completion state\n");
+				ath10k_pci_stop_ce(ar);
+				return -ENOMEM;
+			}
+
+			compl->send_or_recv = HIF_CE_COMPLETE_FREE;
+			list_add_tail(&compl->list, &pipe_info->compl_free);
+		}
+	}
+
+	return 0;
+}
+
+static void ath10k_pci_stop_ce(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ath10k_pci_compl *compl;
+	struct sk_buff *skb;
+	int i;
+
+	ath10k_ce_disable_interrupts(ar);
+
+	/* Cancel the pending tasklet */
+	tasklet_kill(&ar_pci->intr_tq);
+
+	for (i = 0; i < CE_COUNT; i++)
+		tasklet_kill(&ar_pci->pipe_info[i].intr);
+
+	/* Mark pending completions as aborted, so that upper layers free up
+	 * their associated resources */
+	spin_lock_bh(&ar_pci->compl_lock);
+	list_for_each_entry(compl, &ar_pci->compl_process, list) {
+		skb = (struct sk_buff *)compl->transfer_context;
+		ATH10K_SKB_CB(skb)->is_aborted = true;
+	}
+	spin_unlock_bh(&ar_pci->compl_lock);
+}
+
+static void ath10k_pci_cleanup_ce(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ath10k_pci_compl *compl, *tmp;
+	struct hif_ce_pipe_info *pipe_info;
+	struct sk_buff *netbuf;
+	int pipe_num;
+
+	/* Free pending completions. */
+	spin_lock_bh(&ar_pci->compl_lock);
+	if (!list_empty(&ar_pci->compl_process))
+		ath10k_warn("pending completions still present! possible memory leaks.\n");
+
+	list_for_each_entry_safe(compl, tmp, &ar_pci->compl_process, list) {
+		list_del(&compl->list);
+		netbuf = (struct sk_buff *)compl->transfer_context;
+		dev_kfree_skb_any(netbuf);
+		kfree(compl);
+	}
+	spin_unlock_bh(&ar_pci->compl_lock);
+
+	/* Free unused completions for each pipe. */
+	for (pipe_num = 0; pipe_num < ar_pci->ce_count; pipe_num++) {
+		pipe_info = &ar_pci->pipe_info[pipe_num];
+
+		spin_lock_bh(&pipe_info->pipe_lock);
+		list_for_each_entry_safe(compl, tmp,
+					 &pipe_info->compl_free, list) {
+			list_del(&compl->list);
+			kfree(compl);
+		}
+		spin_unlock_bh(&pipe_info->pipe_lock);
+	}
+}
+
+static void ath10k_pci_process_ce(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ar->hif.priv;
+	struct ath10k_hif_cb *cb = &ar_pci->msg_callbacks_current;
+	struct ath10k_pci_compl *compl;
+	struct sk_buff *skb;
+	unsigned int nbytes;
+	int ret, send_done = 0;
+
+	/* Upper layers aren't ready to handle tx/rx completions in parallel so
+	 * we must serialize all completion processing. */
+
+	spin_lock_bh(&ar_pci->compl_lock);
+	if (ar_pci->compl_processing) {
+		spin_unlock_bh(&ar_pci->compl_lock);
+		return;
+	}
+	ar_pci->compl_processing = true;
+	spin_unlock_bh(&ar_pci->compl_lock);
+
+	for (;;) {
+		spin_lock_bh(&ar_pci->compl_lock);
+		if (list_empty(&ar_pci->compl_process)) {
+			spin_unlock_bh(&ar_pci->compl_lock);
+			break;
+		}
+		compl = list_first_entry(&ar_pci->compl_process,
+					 struct ath10k_pci_compl, list);
+		list_del(&compl->list);
+		spin_unlock_bh(&ar_pci->compl_lock);
+
+		if (compl->send_or_recv == HIF_CE_COMPLETE_SEND) {
+			cb->tx_completion(ar,
+					  compl->transfer_context,
+					  compl->transfer_id);
+			send_done = 1;
+		} else {
+			ret = ath10k_pci_post_rx_pipe(compl->pipe_info, 1);
+			if (ret) {
+				ath10k_warn("Unable to post recv buffer for pipe: %d\n",
+					    compl->pipe_info->pipe_num);
+				break;
+			}
+
+			skb = (struct sk_buff *)compl->transfer_context;
+			nbytes = compl->nbytes;
+
+			ath10k_dbg(ATH10K_DBG_PCI,
+				   "ath10k_pci_ce_recv_data netbuf=%p  nbytes=%d\n",
+				   skb, nbytes);
+			ath10k_dbg_dump(ATH10K_DBG_PCI_DUMP, NULL,
+					"ath10k rx: ", skb->data, nbytes);
+
+			if (skb->len + skb_tailroom(skb) >= nbytes) {
+				skb_trim(skb, 0);
+				skb_put(skb, nbytes);
+				cb->rx_completion(ar, skb,
+						  compl->pipe_info->pipe_num);
+			} else {
+				ath10k_warn("rxed more than expected (nbytes %d, max %d)",
+					    nbytes,
+					    skb->len + skb_tailroom(skb));
+			}
+		}
+
+		compl->send_or_recv = HIF_CE_COMPLETE_FREE;
+
+		/*
+		 * Add completion back to the pipe's free list.
+		 */
+		spin_lock_bh(&compl->pipe_info->pipe_lock);
+		list_add_tail(&compl->list, &compl->pipe_info->compl_free);
+		compl->pipe_info->num_sends_allowed += send_done;
+		spin_unlock_bh(&compl->pipe_info->pipe_lock);
+	}
+
+	spin_lock_bh(&ar_pci->compl_lock);
+	ar_pci->compl_processing = false;
+	spin_unlock_bh(&ar_pci->compl_lock);
+}
+
+/* TODO - temporary mapping while we have too few CE's */
+static int ath10k_pci_hif_map_service_to_pipe(struct ath10k *ar,
+					      u16 service_id, u8 *ul_pipe,
+					      u8 *dl_pipe, int *ul_is_polled,
+					      int *dl_is_polled)
+{
+	int ret = 0;
+
+	/* polling for received messages not supported */
+	*dl_is_polled = 0;
+
+	switch (service_id) {
+	case ATH10K_HTC_SVC_ID_HTT_DATA_MSG:
+		/*
+		 * Host->target HTT gets its own pipe, so it can be polled
+		 * while other pipes are interrupt driven.
+		 */
+		*ul_pipe = 4;
+		/*
+		 * Use the same target->host pipe for HTC ctrl, HTC raw
+		 * streams, and HTT.
+		 */
+		*dl_pipe = 1;
+		break;
+
+	case ATH10K_HTC_SVC_ID_RSVD_CTRL:
+	case ATH10K_HTC_SVC_ID_TEST_RAW_STREAMS:
+		/*
+		 * Note: HTC_RAW_STREAMS_SVC is currently unused, and
+		 * HTC_CTRL_RSVD_SVC could share the same pipe as the
+		 * WMI services.  So, if another CE is needed, change
+		 * this to *ul_pipe = 3, which frees up CE 0.
+		 */
+		/* *ul_pipe = 3; */
+		*ul_pipe = 0;
+		*dl_pipe = 1;
+		break;
+
+	case ATH10K_HTC_SVC_ID_WMI_DATA_BK:
+	case ATH10K_HTC_SVC_ID_WMI_DATA_BE:
+	case ATH10K_HTC_SVC_ID_WMI_DATA_VI:
+	case ATH10K_HTC_SVC_ID_WMI_DATA_VO:
+
+	case ATH10K_HTC_SVC_ID_WMI_CONTROL:
+		*ul_pipe = 3;
+		*dl_pipe = 2;
+		break;
+
+		/* pipe 5 unused   */
+		/* pipe 6 reserved */
+		/* pipe 7 reserved */
+
+	default:
+		ret = -1;
+		break;
+	}
+	*ul_is_polled =
+		(host_ce_config_wlan[*ul_pipe].flags & CE_ATTR_DIS_INTR) != 0;
+
+	return ret;
+}
+
+static void ath10k_pci_hif_get_default_pipe(struct ath10k *ar,
+						u8 *ul_pipe, u8 *dl_pipe)
+{
+	int ul_is_polled, dl_is_polled;
+
+	(void)ath10k_pci_hif_map_service_to_pipe(ar,
+						 ATH10K_HTC_SVC_ID_RSVD_CTRL,
+						 ul_pipe,
+						 dl_pipe,
+						 &ul_is_polled,
+						 &dl_is_polled);
+}
+
+static int ath10k_pci_post_rx_pipe(struct hif_ce_pipe_info *pipe_info,
+				   int num)
+{
+	struct ath10k *ar = pipe_info->hif_ce_state;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_state *ce_state = pipe_info->ce_hdl;
+	struct sk_buff *skb;
+	dma_addr_t ce_data;
+	int i, ret = 0;
+
+	if (pipe_info->buf_sz == 0)
+		return 0;
+
+	for (i = 0; i < num; i++) {
+		skb = dev_alloc_skb(pipe_info->buf_sz);
+		if (!skb) {
+			ath10k_warn("could not allocate skbuff for pipe %d\n",
+				    num);
+			ret = -ENOMEM;
+			goto err;
+		}
+
+		WARN_ONCE((unsigned long)skb->data & 3, "unaligned skb");
+
+		ce_data = dma_map_single(ar->dev, skb->data,
+					 skb->len + skb_tailroom(skb),
+					 DMA_FROM_DEVICE);
+
+		if (unlikely(dma_mapping_error(ar->dev, ce_data))) {
+			ath10k_warn("could not dma map skbuff\n");
+			dev_kfree_skb_any(skb);
+			ret = -EIO;
+			goto err;
+		}
+
+		ATH10K_SKB_CB(skb)->paddr = ce_data;
+
+		pci_dma_sync_single_for_device(ar_pci->pdev, ce_data,
+					       pipe_info->buf_sz,
+					       PCI_DMA_FROMDEVICE);
+
+		ret = ath10k_ce_recv_buf_enqueue(ce_state, (void *)skb,
+						 ce_data);
+		if (ret) {
+			ath10k_warn("could not enqueue to pipe %d (%d)\n",
+				    num, ret);
+			goto err;
+		}
+	}
+
+	return ret;
+
+err:
+	ath10k_pci_rx_pipe_cleanup(pipe_info);
+	return ret;
+}
+
+static int ath10k_pci_post_rx(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info;
+	const struct ce_attr *attr;
+	int pipe_num, ret = 0;
+
+	for (pipe_num = 0; pipe_num < ar_pci->ce_count; pipe_num++) {
+		pipe_info = &ar_pci->pipe_info[pipe_num];
+		attr = &host_ce_config_wlan[pipe_num];
+
+		if (attr->dest_nentries == 0)
+			continue;
+
+		ret = ath10k_pci_post_rx_pipe(pipe_info,
+					      attr->dest_nentries - 1);
+		if (ret) {
+			ath10k_warn("Unable to replenish recv buffers for pipe: %d\n",
+				    pipe_num);
+
+			for (; pipe_num >= 0; pipe_num--) {
+				pipe_info = &ar_pci->pipe_info[pipe_num];
+				ath10k_pci_rx_pipe_cleanup(pipe_info);
+			}
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static int ath10k_pci_hif_start(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+
+	ret = ath10k_pci_start_ce(ar);
+	if (ret) {
+		ath10k_warn("could not start CE (%d)\n", ret);
+		return ret;
+	}
+
+	/* Post buffers once to start things off. */
+	ret = ath10k_pci_post_rx(ar);
+	if (ret) {
+		ath10k_warn("could not post rx pipes (%d)\n", ret);
+		return ret;
+	}
+
+	ar_pci->started = 1;
+	return 0;
+}
+
+static void ath10k_pci_rx_pipe_cleanup(struct hif_ce_pipe_info *pipe_info)
+{
+	struct ath10k *ar;
+	struct ath10k_pci *ar_pci;
+	struct ce_state *ce_hdl;
+	u32 buf_sz;
+	struct sk_buff *netbuf;
+	u32 ce_data;
+
+	buf_sz = pipe_info->buf_sz;
+
+	/* Unused Copy Engine */
+	if (buf_sz == 0)
+		return;
+
+	ar = pipe_info->hif_ce_state;
+	ar_pci = ath10k_pci_priv(ar);
+
+	if (!ar_pci->started)
+		return;
+
+	ce_hdl = pipe_info->ce_hdl;
+
+	while (ath10k_ce_revoke_recv_next(ce_hdl, (void **)&netbuf,
+					  &ce_data) == 0) {
+		dma_unmap_single(ar->dev, ATH10K_SKB_CB(netbuf)->paddr,
+				 netbuf->len + skb_tailroom(netbuf),
+				 DMA_FROM_DEVICE);
+		dev_kfree_skb_any(netbuf);
+	}
+}
+
+static void ath10k_pci_tx_pipe_cleanup(struct hif_ce_pipe_info *pipe_info)
+{
+	struct ath10k *ar;
+	struct ath10k_pci *ar_pci;
+	struct ce_state *ce_hdl;
+	struct sk_buff *netbuf;
+	u32 ce_data;
+	unsigned int nbytes;
+	unsigned int id;
+	u32 buf_sz;
+
+	buf_sz = pipe_info->buf_sz;
+
+	/* Unused Copy Engine */
+	if (buf_sz == 0)
+		return;
+
+	ar = pipe_info->hif_ce_state;
+	ar_pci = ath10k_pci_priv(ar);
+
+	if (!ar_pci->started)
+		return;
+
+	ce_hdl = pipe_info->ce_hdl;
+
+	while (ath10k_ce_cancel_send_next(ce_hdl, (void **)&netbuf,
+					  &ce_data, &nbytes, &id) == 0) {
+		if (netbuf != CE_SENDLIST_ITEM_CTXT)
+			/*
+			 * Indicate the completion to higer layer to free
+			 * the buffer
+			 */
+			ATH10K_SKB_CB(netbuf)->is_aborted = true;
+			ar_pci->msg_callbacks_current.tx_completion(ar,
+								    netbuf,
+								    id);
+	}
+}
+
+/*
+ * Cleanup residual buffers for device shutdown:
+ *    buffers that were enqueued for receive
+ *    buffers that were to be sent
+ * Note: Buffers that had completed but which were
+ * not yet processed are on a completion queue. They
+ * are handled when the completion thread shuts down.
+ */
+static void ath10k_pci_buffer_cleanup(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int pipe_num;
+
+	for (pipe_num = 0; pipe_num < ar_pci->ce_count; pipe_num++) {
+		struct hif_ce_pipe_info *pipe_info;
+
+		pipe_info = &ar_pci->pipe_info[pipe_num];
+		ath10k_pci_rx_pipe_cleanup(pipe_info);
+		ath10k_pci_tx_pipe_cleanup(pipe_info);
+	}
+}
+
+static void ath10k_pci_ce_deinit(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info;
+	int pipe_num;
+
+	for (pipe_num = 0; pipe_num < ar_pci->ce_count; pipe_num++) {
+		pipe_info = &ar_pci->pipe_info[pipe_num];
+		if (pipe_info->ce_hdl) {
+			ath10k_ce_deinit(pipe_info->ce_hdl);
+			pipe_info->ce_hdl = NULL;
+			pipe_info->buf_sz = 0;
+		}
+	}
+}
+
+static void ath10k_pci_hif_stop(struct ath10k *ar)
+{
+	ath10k_dbg(ATH10K_DBG_PCI, "%s\n", __func__);
+
+	ath10k_pci_stop_ce(ar);
+
+	/* At this point, asynchronous threads are stopped, the target should
+	 * not DMA nor interrupt. We process the leftovers and then free
+	 * everything else up. */
+
+	ath10k_pci_process_ce(ar);
+	ath10k_pci_cleanup_ce(ar);
+	ath10k_pci_buffer_cleanup(ar);
+	ath10k_pci_ce_deinit(ar);
+}
+
+static int ath10k_pci_hif_exchange_bmi_msg(struct ath10k *ar,
+					   void *req, u32 req_len,
+					   void *resp, u32 *resp_len)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct ce_state *ce_tx = ar_pci->pipe_info[BMI_CE_NUM_TO_TARG].ce_hdl;
+	struct ce_state *ce_rx = ar_pci->pipe_info[BMI_CE_NUM_TO_HOST].ce_hdl;
+	dma_addr_t req_paddr = 0;
+	dma_addr_t resp_paddr = 0;
+	struct bmi_xfer xfer = {};
+	void *treq, *tresp = NULL;
+	int ret = 0;
+
+	if (resp && !resp_len)
+		return -EINVAL;
+
+	if (resp && resp_len && *resp_len == 0)
+		return -EINVAL;
+
+	treq = kmemdup(req, req_len, GFP_KERNEL);
+	if (!treq)
+		return -ENOMEM;
+
+	req_paddr = dma_map_single(ar->dev, treq, req_len, DMA_TO_DEVICE);
+	ret = dma_mapping_error(ar->dev, req_paddr);
+	if (ret)
+		goto err_dma;
+
+	if (resp && resp_len) {
+		tresp = kzalloc(*resp_len, GFP_KERNEL);
+		if (!tresp) {
+			ret = -ENOMEM;
+			goto err_req;
+		}
+
+		resp_paddr = dma_map_single(ar->dev, tresp, *resp_len,
+					    DMA_FROM_DEVICE);
+		ret = dma_mapping_error(ar->dev, resp_paddr);
+		if (ret)
+			goto err_req;
+
+		xfer.wait_for_resp = true;
+		xfer.resp_len = 0;
+
+		ath10k_ce_recv_buf_enqueue(ce_rx, &xfer, resp_paddr);
+	}
+
+	init_completion(&xfer.done);
+
+	ret = ath10k_ce_send(ce_tx, &xfer, req_paddr, req_len, -1, 0);
+	if (ret)
+		goto err_resp;
+
+	ret = wait_for_completion_timeout(&xfer.done,
+					  BMI_COMMUNICATION_TIMEOUT_HZ);
+	if (ret <= 0) {
+		u32 unused_buffer;
+		unsigned int unused_nbytes;
+		unsigned int unused_id;
+
+		ret = -ETIMEDOUT;
+		ath10k_ce_cancel_send_next(ce_tx, NULL, &unused_buffer,
+					   &unused_nbytes, &unused_id);
+	} else {
+		/* non-zero means we did not time out */
+		ret = 0;
+	}
+
+err_resp:
+	if (resp) {
+		u32 unused_buffer;
+
+		ath10k_ce_revoke_recv_next(ce_rx, NULL, &unused_buffer);
+		dma_unmap_single(ar->dev, resp_paddr,
+				 *resp_len, DMA_FROM_DEVICE);
+	}
+err_req:
+	dma_unmap_single(ar->dev, req_paddr, req_len, DMA_TO_DEVICE);
+
+	if (ret == 0 && resp_len) {
+		*resp_len = min(*resp_len, xfer.resp_len);
+		memcpy(resp, tresp, xfer.resp_len);
+	}
+err_dma:
+	kfree(treq);
+	kfree(tresp);
+
+	return ret;
+}
+
+static void ath10k_pci_bmi_send_done(struct ce_state *ce_state,
+				     void *transfer_context,
+				     u32 data,
+				     unsigned int nbytes,
+				     unsigned int transfer_id)
+{
+	struct bmi_xfer *xfer = transfer_context;
+
+	if (xfer->wait_for_resp)
+		return;
+
+	complete(&xfer->done);
+}
+
+static void ath10k_pci_bmi_recv_data(struct ce_state *ce_state,
+				     void *transfer_context,
+				     u32 data,
+				     unsigned int nbytes,
+				     unsigned int transfer_id,
+				     unsigned int flags)
+{
+	struct bmi_xfer *xfer = transfer_context;
+
+	if (!xfer->wait_for_resp) {
+		ath10k_warn("unexpected: BMI data received; ignoring\n");
+		return;
+	}
+
+	xfer->resp_len = nbytes;
+	complete(&xfer->done);
+}
+
+/*
+ * Map from service/endpoint to Copy Engine.
+ * This table is derived from the CE_PCI TABLE, above.
+ * It is passed to the Target at startup for use by firmware.
+ */
+static const struct service_to_pipe target_service_to_ce_map_wlan[] = {
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_VO,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 3,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_VO,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 2,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_BK,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 3,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_BK,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 2,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_BE,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 3,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_BE,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 2,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_VI,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 3,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_DATA_VI,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 2,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_CONTROL,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 3,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_WMI_CONTROL,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 2,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_RSVD_CTRL,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 0,		/* could be moved to 3 (share with WMI) */
+	},
+	{
+		 ATH10K_HTC_SVC_ID_RSVD_CTRL,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 1,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_TEST_RAW_STREAMS,	/* not currently used */
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 0,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_TEST_RAW_STREAMS,	/* not currently used */
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 1,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_HTT_DATA_MSG,
+		 PIPEDIR_OUT,		/* out = UL = host -> target */
+		 4,
+	},
+	{
+		 ATH10K_HTC_SVC_ID_HTT_DATA_MSG,
+		 PIPEDIR_IN,		/* in = DL = target -> host */
+		 1,
+	},
+
+	/* (Additions here) */
+
+	{				/* Must be last */
+		 0,
+		 0,
+		 0,
+	},
+};
+
+/*
+ * Send an interrupt to the device to wake up the Target CPU
+ * so it has an opportunity to notice any changed state.
+ */
+static int ath10k_pci_wake_target_cpu(struct ath10k *ar)
+{
+	int ret;
+	u32 core_ctrl;
+
+	ret = ath10k_pci_diag_read_access(ar, SOC_CORE_BASE_ADDRESS |
+					      CORE_CTRL_ADDRESS,
+					  &core_ctrl);
+	if (ret) {
+		ath10k_warn("Unable to read core ctrl\n");
+		return ret;
+	}
+
+	/* A_INUM_FIRMWARE interrupt to Target CPU */
+	core_ctrl |= CORE_CTRL_CPU_INTR_MASK;
+
+	ret = ath10k_pci_diag_write_access(ar, SOC_CORE_BASE_ADDRESS |
+					       CORE_CTRL_ADDRESS,
+					   core_ctrl);
+	if (ret)
+		ath10k_warn("Unable to set interrupt mask\n");
+
+	return ret;
+}
+
+static int ath10k_pci_init_config(struct ath10k *ar)
+{
+	u32 interconnect_targ_addr;
+	u32 pcie_state_targ_addr = 0;
+	u32 pipe_cfg_targ_addr = 0;
+	u32 svc_to_pipe_map = 0;
+	u32 pcie_config_flags = 0;
+	u32 ealloc_value;
+	u32 ealloc_targ_addr;
+	u32 flag2_value;
+	u32 flag2_targ_addr;
+	int ret = 0;
+
+	/* Download to Target the CE Config and the service-to-CE map */
+	interconnect_targ_addr =
+		host_interest_item_address(HI_ITEM(hi_interconnect_state));
+
+	/* Supply Target-side CE configuration */
+	ret = ath10k_pci_diag_read_access(ar, interconnect_targ_addr,
+					  &pcie_state_targ_addr);
+	if (ret != 0) {
+		ath10k_err("Failed to get pcie state addr: %d\n", ret);
+		return ret;
+	}
+
+	if (pcie_state_targ_addr == 0) {
+		ret = -EIO;
+		ath10k_err("Invalid pcie state addr\n");
+		return ret;
+	}
+
+	ret = ath10k_pci_diag_read_access(ar, pcie_state_targ_addr +
+					  offsetof(struct pcie_state,
+						   pipe_cfg_addr),
+					  &pipe_cfg_targ_addr);
+	if (ret != 0) {
+		ath10k_err("Failed to get pipe cfg addr: %d\n", ret);
+		return ret;
+	}
+
+	if (pipe_cfg_targ_addr == 0) {
+		ret = -EIO;
+		ath10k_err("Invalid pipe cfg addr\n");
+		return ret;
+	}
+
+	ret = ath10k_pci_diag_write_mem(ar, pipe_cfg_targ_addr,
+				 target_ce_config_wlan,
+				 sizeof(target_ce_config_wlan));
+
+	if (ret != 0) {
+		ath10k_err("Failed to write pipe cfg: %d\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_pci_diag_read_access(ar, pcie_state_targ_addr +
+					  offsetof(struct pcie_state,
+						   svc_to_pipe_map),
+					  &svc_to_pipe_map);
+	if (ret != 0) {
+		ath10k_err("Failed to get svc/pipe map: %d\n", ret);
+		return ret;
+	}
+
+	if (svc_to_pipe_map == 0) {
+		ret = -EIO;
+		ath10k_err("Invalid svc_to_pipe map\n");
+		return ret;
+	}
+
+	ret = ath10k_pci_diag_write_mem(ar, svc_to_pipe_map,
+				 target_service_to_ce_map_wlan,
+				 sizeof(target_service_to_ce_map_wlan));
+	if (ret != 0) {
+		ath10k_err("Failed to write svc/pipe map: %d\n", ret);
+		return ret;
+	}
+
+	ret = ath10k_pci_diag_read_access(ar, pcie_state_targ_addr +
+					  offsetof(struct pcie_state,
+						   config_flags),
+					  &pcie_config_flags);
+	if (ret != 0) {
+		ath10k_err("Failed to get pcie config_flags: %d\n", ret);
+		return ret;
+	}
+
+	pcie_config_flags &= ~PCIE_CONFIG_FLAG_ENABLE_L1;
+
+	ret = ath10k_pci_diag_write_mem(ar, pcie_state_targ_addr +
+				 offsetof(struct pcie_state, config_flags),
+				 &pcie_config_flags,
+				 sizeof(pcie_config_flags));
+	if (ret != 0) {
+		ath10k_err("Failed to write pcie config_flags: %d\n", ret);
+		return ret;
+	}
+
+	/* configure early allocation */
+	ealloc_targ_addr = host_interest_item_address(HI_ITEM(hi_early_alloc));
+
+	ret = ath10k_pci_diag_read_access(ar, ealloc_targ_addr, &ealloc_value);
+	if (ret != 0) {
+		ath10k_err("Faile to get early alloc val: %d\n", ret);
+		return ret;
+	}
+
+	/* first bank is switched to IRAM */
+	ealloc_value |= ((HI_EARLY_ALLOC_MAGIC << HI_EARLY_ALLOC_MAGIC_SHIFT) &
+			 HI_EARLY_ALLOC_MAGIC_MASK);
+	ealloc_value |= ((1 << HI_EARLY_ALLOC_IRAM_BANKS_SHIFT) &
+			 HI_EARLY_ALLOC_IRAM_BANKS_MASK);
+
+	ret = ath10k_pci_diag_write_access(ar, ealloc_targ_addr, ealloc_value);
+	if (ret != 0) {
+		ath10k_err("Failed to set early alloc val: %d\n", ret);
+		return ret;
+	}
+
+	/* Tell Target to proceed with initialization */
+	flag2_targ_addr = host_interest_item_address(HI_ITEM(hi_option_flag2));
+
+	ret = ath10k_pci_diag_read_access(ar, flag2_targ_addr, &flag2_value);
+	if (ret != 0) {
+		ath10k_err("Failed to get option val: %d\n", ret);
+		return ret;
+	}
+
+	flag2_value |= HI_OPTION_EARLY_CFG_DONE;
+
+	ret = ath10k_pci_diag_write_access(ar, flag2_targ_addr, flag2_value);
+	if (ret != 0) {
+		ath10k_err("Failed to set option val: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+
+
+static int ath10k_pci_ce_init(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	struct hif_ce_pipe_info *pipe_info;
+	const struct ce_attr *attr;
+	int pipe_num;
+
+	for (pipe_num = 0; pipe_num < ar_pci->ce_count; pipe_num++) {
+		pipe_info = &ar_pci->pipe_info[pipe_num];
+		pipe_info->pipe_num = pipe_num;
+		pipe_info->hif_ce_state = ar;
+		attr = &host_ce_config_wlan[pipe_num];
+
+		pipe_info->ce_hdl = ath10k_ce_init(ar, pipe_num, attr);
+		if (pipe_info->ce_hdl == NULL) {
+			ath10k_err("Unable to initialize CE for pipe: %d\n",
+				   pipe_num);
+
+			/* It is safe to call it here. It checks if ce_hdl is
+			 * valid for each pipe */
+			ath10k_pci_ce_deinit(ar);
+			return -1;
+		}
+
+		if (pipe_num == ar_pci->ce_count - 1) {
+			/*
+			 * Reserve the ultimate CE for
+			 * diagnostic Window support
+			 */
+			ar_pci->ce_diag =
+			ar_pci->pipe_info[ar_pci->ce_count - 1].ce_hdl;
+			continue;
+		}
+
+		pipe_info->buf_sz = (size_t) (attr->src_sz_max);
+	}
+
+	/*
+	 * Initially, establish CE completion handlers for use with BMI.
+	 * These are overwritten with generic handlers after we exit BMI phase.
+	 */
+	pipe_info = &ar_pci->pipe_info[BMI_CE_NUM_TO_TARG];
+	ath10k_ce_send_cb_register(pipe_info->ce_hdl,
+				   ath10k_pci_bmi_send_done, 0);
+
+	pipe_info = &ar_pci->pipe_info[BMI_CE_NUM_TO_HOST];
+	ath10k_ce_recv_cb_register(pipe_info->ce_hdl,
+				   ath10k_pci_bmi_recv_data);
+
+	return 0;
+}
+
+static void ath10k_pci_fw_interrupt_handler(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	u32 fw_indicator_address, fw_indicator;
+
+	ath10k_pci_wake(ar);
+
+	fw_indicator_address = ar_pci->fw_indicator_address;
+	fw_indicator = ath10k_pci_read32(ar, fw_indicator_address);
+
+	if (fw_indicator & FW_IND_EVENT_PENDING) {
+		/* ACK: clear Target-side pending event */
+		ath10k_pci_write32(ar, fw_indicator_address,
+				   fw_indicator & ~FW_IND_EVENT_PENDING);
+
+		if (ar_pci->started) {
+			ath10k_pci_hif_dump_area(ar);
+		} else {
+			/*
+			 * Probable Target failure before we're prepared
+			 * to handle it.  Generally unexpected.
+			 */
+			ath10k_warn("early firmware event indicated\n");
+		}
+	}
+
+	ath10k_pci_sleep(ar);
+}
+
+static const struct ath10k_hif_ops ath10k_pci_hif_ops = {
+	.send_head		= ath10k_pci_hif_send_head,
+	.exchange_bmi_msg	= ath10k_pci_hif_exchange_bmi_msg,
+	.start			= ath10k_pci_hif_start,
+	.stop			= ath10k_pci_hif_stop,
+	.map_service_to_pipe	= ath10k_pci_hif_map_service_to_pipe,
+	.get_default_pipe	= ath10k_pci_hif_get_default_pipe,
+	.send_complete_check	= ath10k_pci_hif_send_complete_check,
+	.init			= ath10k_pci_hif_post_init,
+	.get_free_queue_number	= ath10k_pci_hif_get_free_queue_number,
+};
+
+static void ath10k_pci_ce_tasklet(unsigned long ptr)
+{
+	struct hif_ce_pipe_info *pipe = (struct hif_ce_pipe_info *)ptr;
+	struct ath10k_pci *ar_pci = pipe->ar_pci;
+
+	ath10k_ce_per_engine_service(ar_pci->ar, pipe->pipe_num);
+}
+
+static void ath10k_msi_err_tasklet(unsigned long data)
+{
+	struct ath10k *ar = (struct ath10k *)data;
+
+	ath10k_pci_fw_interrupt_handler(ar);
+}
+
+/*
+ * Handler for a per-engine interrupt on a PARTICULAR CE.
+ * This is used in cases where each CE has a private MSI interrupt.
+ */
+static irqreturn_t ath10k_pci_per_engine_handler(int irq, void *arg)
+{
+	struct ath10k *ar = arg;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ce_id = irq - ar_pci->pdev->irq - MSI_ASSIGN_CE_INITIAL;
+
+	if (ce_id < 0 || ce_id >= ARRAY_SIZE(ar_pci->pipe_info)) {
+		ath10k_warn("unexpected/invalid irq %d ce_id %d\n", irq, ce_id);
+		return IRQ_HANDLED;
+	}
+
+	/*
+	 * NOTE: We are able to derive ce_id from irq because we
+	 * use a one-to-one mapping for CE's 0..5.
+	 * CE's 6 & 7 do not use interrupts at all.
+	 *
+	 * This mapping must be kept in sync with the mapping
+	 * used by firmware.
+	 */
+	tasklet_schedule(&ar_pci->pipe_info[ce_id].intr);
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t ath10k_pci_msi_fw_handler(int irq, void *arg)
+{
+	struct ath10k *ar = arg;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	tasklet_schedule(&ar_pci->msi_fw_err);
+	return IRQ_HANDLED;
+}
+
+/*
+ * Top-level interrupt handler for all PCI interrupts from a Target.
+ * When a block of MSI interrupts is allocated, this top-level handler
+ * is not used; instead, we directly call the correct sub-handler.
+ */
+static irqreturn_t ath10k_pci_interrupt_handler(int irq, void *arg)
+{
+	struct ath10k *ar = arg;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	if (ar_pci->num_msi_intrs == 0) {
+		/*
+		 * IMPORTANT: INTR_CLR regiser has to be set after
+		 * INTR_ENABLE is set to 0, otherwise interrupt can not be
+		 * really cleared.
+		 */
+		iowrite32(0, ar_pci->mem +
+			  (SOC_CORE_BASE_ADDRESS |
+			   PCIE_INTR_ENABLE_ADDRESS));
+		iowrite32(PCIE_INTR_FIRMWARE_MASK |
+			  PCIE_INTR_CE_MASK_ALL,
+			  ar_pci->mem + (SOC_CORE_BASE_ADDRESS |
+					 PCIE_INTR_CLR_ADDRESS));
+		/*
+		 * IMPORTANT: this extra read transaction is required to
+		 * flush the posted write buffer.
+		 */
+		(void) ioread32(ar_pci->mem +
+				(SOC_CORE_BASE_ADDRESS |
+				 PCIE_INTR_ENABLE_ADDRESS));
+	}
+
+	tasklet_schedule(&ar_pci->intr_tq);
+
+	return IRQ_HANDLED;
+}
+
+static void ath10k_pci_tasklet(unsigned long data)
+{
+	struct ath10k *ar = (struct ath10k *)data;
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	ath10k_pci_fw_interrupt_handler(ar); /* FIXME: Handle FW error */
+	ath10k_ce_per_engine_service_any(ar);
+
+	if (ar_pci->num_msi_intrs == 0) {
+		/* Enable Legacy PCI line interrupts */
+		iowrite32(PCIE_INTR_FIRMWARE_MASK |
+			  PCIE_INTR_CE_MASK_ALL,
+			  ar_pci->mem + (SOC_CORE_BASE_ADDRESS |
+					 PCIE_INTR_ENABLE_ADDRESS));
+		/*
+		 * IMPORTANT: this extra read transaction is required to
+		 * flush the posted write buffer
+		 */
+		(void) ioread32(ar_pci->mem +
+				(SOC_CORE_BASE_ADDRESS |
+				 PCIE_INTR_ENABLE_ADDRESS));
+	}
+}
+
+static int ath10k_pci_start_intr_msix(struct ath10k *ar, int num)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+	int i;
+
+	ret = pci_enable_msi_block(ar_pci->pdev, num);
+	if (ret)
+		return ret;
+
+	ret = request_irq(ar_pci->pdev->irq + MSI_ASSIGN_FW,
+			  ath10k_pci_msi_fw_handler,
+			  IRQF_SHARED, "ath10k_pci", ar);
+	if (ret)
+		return ret;
+
+	for (i = MSI_ASSIGN_CE_INITIAL; i <= MSI_ASSIGN_CE_MAX; i++) {
+		ret = request_irq(ar_pci->pdev->irq + i,
+				  ath10k_pci_per_engine_handler,
+				  IRQF_SHARED, "ath10k_pci", ar);
+		if (ret) {
+			ath10k_warn("request_irq(%d) failed %d\n",
+				    ar_pci->pdev->irq + i, ret);
+
+			for (i--; i >= MSI_ASSIGN_CE_INITIAL; i--)
+				free_irq(ar_pci->pdev->irq + i, ar);
+
+			free_irq(ar_pci->pdev->irq + MSI_ASSIGN_FW, ar);
+			pci_disable_msi(ar_pci->pdev);
+			return ret;
+		}
+	}
+
+	ath10k_info("MSI-X interrupt handling (%d intrs)\n", num);
+	return 0;
+}
+
+static int ath10k_pci_start_intr_msi(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+
+	ret = pci_enable_msi(ar_pci->pdev);
+	if (ret < 0)
+		return ret;
+
+	ret = request_irq(ar_pci->pdev->irq,
+			  ath10k_pci_interrupt_handler,
+			  IRQF_SHARED, "ath10k_pci", ar);
+	if (ret < 0) {
+		pci_disable_msi(ar_pci->pdev);
+		return ret;
+	}
+
+	ath10k_info("MSI interrupt handling\n");
+	return 0;
+}
+
+static int ath10k_pci_start_intr_legacy(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int ret;
+
+	ret = request_irq(ar_pci->pdev->irq,
+			  ath10k_pci_interrupt_handler,
+			  IRQF_SHARED, "ath10k_pci", ar);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * Make sure to wake the Target before enabling Legacy
+	 * Interrupt.
+	 */
+	iowrite32(PCIE_SOC_WAKE_V_MASK,
+		  ar_pci->mem + PCIE_LOCAL_BASE_ADDRESS +
+		  PCIE_SOC_WAKE_ADDRESS);
+
+	ath10k_pci_wait(ar);
+
+	/*
+	 * A potential race occurs here: The CORE_BASE write
+	 * depends on target correctly decoding AXI address but
+	 * host won't know when target writes BAR to CORE_CTRL.
+	 * This write might get lost if target has NOT written BAR.
+	 * For now, fix the race by repeating the write in below
+	 * synchronization checking.
+	 */
+	iowrite32(PCIE_INTR_FIRMWARE_MASK |
+		  PCIE_INTR_CE_MASK_ALL,
+		  ar_pci->mem + (SOC_CORE_BASE_ADDRESS |
+				 PCIE_INTR_ENABLE_ADDRESS));
+	iowrite32(PCIE_SOC_WAKE_RESET,
+		  ar_pci->mem + PCIE_LOCAL_BASE_ADDRESS +
+		  PCIE_SOC_WAKE_ADDRESS);
+
+	ath10k_info("legacy interrupt handling\n");
+	return 0;
+}
+
+static int ath10k_pci_start_intr(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int num = MSI_NUM_REQUEST;
+	int ret;
+	int i;
+
+	tasklet_init(&ar_pci->intr_tq, ath10k_pci_tasklet, (unsigned long) ar);
+	tasklet_init(&ar_pci->msi_fw_err, ath10k_msi_err_tasklet,
+		     (unsigned long) ar);
+
+	for (i = 0; i < CE_COUNT; i++) {
+		ar_pci->pipe_info[i].ar_pci = ar_pci;
+		tasklet_init(&ar_pci->pipe_info[i].intr,
+			     ath10k_pci_ce_tasklet,
+			     (unsigned long)&ar_pci->pipe_info[i]);
+	}
+
+	if (!test_bit(ATH10K_PCI_FEATURE_MSI_X, ar_pci->features))
+		num = 1;
+
+	if (num > 1) {
+		ret = ath10k_pci_start_intr_msix(ar, num);
+		if (ret == 0)
+			goto exit;
+
+		ath10k_warn("MSI-X didn't succeed (%d), trying MSI\n", ret);
+		num = 1;
+	}
+
+	if (num == 1) {
+		ret = ath10k_pci_start_intr_msi(ar);
+		if (ret == 0)
+			goto exit;
+
+		ath10k_warn("MSI didn't succeed (%d), trying legacy INTR\n",
+			    ret);
+		num = 0;
+	}
+
+	ret = ath10k_pci_start_intr_legacy(ar);
+
+exit:
+	ar_pci->num_msi_intrs = num;
+	ar_pci->ce_count = CE_COUNT;
+	return ret;
+}
+
+static void ath10k_pci_stop_intr(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int i;
+
+	/* There's at least one interrupt irregardless whether its legacy INTR
+	 * or MSI or MSI-X */
+	for (i = 0; i < max(1, ar_pci->num_msi_intrs); i++)
+		free_irq(ar_pci->pdev->irq + i, ar);
+
+	if (ar_pci->num_msi_intrs > 0)
+		pci_disable_msi(ar_pci->pdev);
+}
+
+static int ath10k_pci_reset_target(struct ath10k *ar)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	int wait_limit = 300; /* 3 sec */
+
+	/* Wait for Target to finish initialization before we proceed. */
+	iowrite32(PCIE_SOC_WAKE_V_MASK,
+		  ar_pci->mem + PCIE_LOCAL_BASE_ADDRESS +
+		  PCIE_SOC_WAKE_ADDRESS);
+
+	ath10k_pci_wait(ar);
+
+	while (wait_limit-- &&
+	       !(ioread32(ar_pci->mem + FW_INDICATOR_ADDRESS) &
+		 FW_IND_INITIALIZED)) {
+		if (ar_pci->num_msi_intrs == 0)
+			/* Fix potential race by repeating CORE_BASE writes */
+			iowrite32(PCIE_INTR_FIRMWARE_MASK |
+				  PCIE_INTR_CE_MASK_ALL,
+				  ar_pci->mem + (SOC_CORE_BASE_ADDRESS |
+						 PCIE_INTR_ENABLE_ADDRESS));
+		mdelay(10);
+	}
+
+	if (wait_limit < 0) {
+		ath10k_err("Target stalled\n");
+		iowrite32(PCIE_SOC_WAKE_RESET,
+			  ar_pci->mem + PCIE_LOCAL_BASE_ADDRESS +
+			  PCIE_SOC_WAKE_ADDRESS);
+		return -EIO;
+	}
+
+	iowrite32(PCIE_SOC_WAKE_RESET,
+		  ar_pci->mem + PCIE_LOCAL_BASE_ADDRESS +
+		  PCIE_SOC_WAKE_ADDRESS);
+
+	return 0;
+}
+
+static void ath10k_pci_device_reset(struct ath10k_pci *ar_pci)
+{
+	struct ath10k *ar = ar_pci->ar;
+	void __iomem *mem = ar_pci->mem;
+	int i;
+	u32 val;
+
+	if (!SOC_GLOBAL_RESET_ADDRESS)
+		return;
+
+	if (!mem)
+		return;
+
+	ath10k_pci_reg_write32(mem, PCIE_SOC_WAKE_ADDRESS,
+			       PCIE_SOC_WAKE_V_MASK);
+	for (i = 0; i < ATH_PCI_RESET_WAIT_MAX; i++) {
+		if (ath10k_pci_target_is_awake(ar))
+			break;
+		msleep(1);
+	}
+
+	/* Put Target, including PCIe, into RESET. */
+	val = ath10k_pci_reg_read32(mem, SOC_GLOBAL_RESET_ADDRESS);
+	val |= 1;
+	ath10k_pci_reg_write32(mem, SOC_GLOBAL_RESET_ADDRESS, val);
+
+	for (i = 0; i < ATH_PCI_RESET_WAIT_MAX; i++) {
+		if (ath10k_pci_reg_read32(mem, RTC_STATE_ADDRESS) &
+					  RTC_STATE_COLD_RESET_MASK)
+			break;
+		msleep(1);
+	}
+
+	/* Pull Target, including PCIe, out of RESET. */
+	val &= ~1;
+	ath10k_pci_reg_write32(mem, SOC_GLOBAL_RESET_ADDRESS, val);
+
+	for (i = 0; i < ATH_PCI_RESET_WAIT_MAX; i++) {
+		if (!(ath10k_pci_reg_read32(mem, RTC_STATE_ADDRESS) &
+					    RTC_STATE_COLD_RESET_MASK))
+			break;
+		msleep(1);
+	}
+
+	ath10k_pci_reg_write32(mem, PCIE_SOC_WAKE_ADDRESS, PCIE_SOC_WAKE_RESET);
+}
+
+static void ath10k_pci_dump_features(struct ath10k_pci *ar_pci)
+{
+	int i;
+
+	for (i = 0; i < ATH10K_PCI_FEATURE_COUNT; i++) {
+		if (!test_bit(i, ar_pci->features))
+			continue;
+
+		switch (i) {
+		case ATH10K_PCI_FEATURE_MSI_X:
+			ath10k_dbg(ATH10K_DBG_PCI, "device supports MSI-X\n");
+			break;
+		case ATH10K_PCI_FEATURE_HW_1_0_WARKAROUND:
+			ath10k_dbg(ATH10K_DBG_PCI, "QCA988X_1.0 workaround enabled\n");
+			break;
+		}
+	}
+}
+
+static int ath10k_pci_probe(struct pci_dev *pdev,
+			    const struct pci_device_id *pci_dev)
+{
+	void __iomem *mem;
+	int ret = 0;
+	struct ath10k *ar;
+	struct ath10k_pci *ar_pci;
+	u32 lcr_val;
+
+	ath10k_dbg(ATH10K_DBG_PCI, "%s\n", __func__);
+
+	ar_pci = kzalloc(sizeof(*ar_pci), GFP_KERNEL);
+	if (ar_pci == NULL)
+		return -ENOMEM;
+
+	ar_pci->pdev = pdev;
+	ar_pci->dev = &pdev->dev;
+
+	switch (pci_dev->device) {
+	case QCA988X_1_0_DEVICE_ID:
+		set_bit(ATH10K_PCI_FEATURE_HW_1_0_WARKAROUND, ar_pci->features);
+		break;
+	case QCA988X_2_0_DEVICE_ID:
+		set_bit(ATH10K_PCI_FEATURE_MSI_X, ar_pci->features);
+		break;
+	default:
+		ret = -ENODEV;
+		ath10k_err("Unkown device ID: %d\n", pci_dev->device);
+		goto err_ar_pci;
+	}
+
+	ath10k_pci_dump_features(ar_pci);
+
+	ar = ath10k_core_create(ar_pci, ar_pci->dev, ATH10K_BUS_PCI,
+				&ath10k_pci_hif_ops);
+	if (!ar) {
+		ath10k_err("ath10k_core_create failed!\n");
+		ret = -EINVAL;
+		goto err_ar_pci;
+	}
+
+	/* Enable QCA988X_1.0 HW workarounds */
+	if (test_bit(ATH10K_PCI_FEATURE_HW_1_0_WARKAROUND, ar_pci->features))
+		spin_lock_init(&ar_pci->hw_v1_workaround_lock);
+
+	ar_pci->ar = ar;
+	ar_pci->fw_indicator_address = FW_INDICATOR_ADDRESS;
+	atomic_set(&ar_pci->keep_awake_count, 0);
+
+	pci_set_drvdata(pdev, ar);
+
+	/*
+	 * Without any knowledge of the Host, the Target may have been reset or
+	 * power cycled and its Config Space may no longer reflect the PCI
+	 * address space that was assigned earlier by the PCI infrastructure.
+	 * Refresh it now.
+	 */
+	ret = pci_assign_resource(pdev, BAR_NUM);
+	if (ret) {
+		ath10k_err("cannot assign PCI space: %d\n", ret);
+		goto err_ar;
+	}
+
+	ret = pci_enable_device(pdev);
+	if (ret) {
+		ath10k_err("cannot enable PCI device: %d\n", ret);
+		goto err_ar;
+	}
+
+	/* Request MMIO resources */
+	ret = pci_request_region(pdev, BAR_NUM, "ath");
+	if (ret) {
+		ath10k_err("PCI MMIO reservation error: %d\n", ret);
+		goto err_device;
+	}
+
+	/*
+	 * Target structures have a limit of 32 bit DMA pointers.
+	 * DMA pointers can be wider than 32 bits by default on some systems.
+	 */
+	ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (ret) {
+		ath10k_err("32-bit DMA not available: %d\n", ret);
+		goto err_region;
+	}
+
+	ret = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (ret) {
+		ath10k_err("cannot enable 32-bit consistent DMA\n");
+		goto err_region;
+	}
+
+	/* Set bus master bit in PCI_COMMAND to enable DMA */
+	pci_set_master(pdev);
+
+	/*
+	 * Temporary FIX: disable ASPM
+	 * Will be removed after the OTP is programmed
+	 */
+	pci_read_config_dword(pdev, 0x80, &lcr_val);
+	pci_write_config_dword(pdev, 0x80, (lcr_val & 0xffffff00));
+
+	/* Arrange for access to Target SoC registers. */
+	mem = pci_iomap(pdev, BAR_NUM, 0);
+	if (!mem) {
+		ath10k_err("PCI iomap error\n");
+		ret = -EIO;
+		goto err_master;
+	}
+
+	ar_pci->mem = mem;
+
+	spin_lock_init(&ar_pci->ce_lock);
+
+	ar_pci->cacheline_sz = dma_get_cache_alignment();
+
+	ret = ath10k_pci_start_intr(ar);
+	if (ret) {
+		ath10k_err("could not start interrupt handling (%d)\n", ret);
+		goto err_iomap;
+	}
+
+	/*
+	 * Bring the target up cleanly.
+	 *
+	 * The target may be in an undefined state with an AUX-powered Target
+	 * and a Host in WoW mode. If the Host crashes, loses power, or is
+	 * restarted (without unloading the driver) then the Target is left
+	 * (aux) powered and running. On a subsequent driver load, the Target
+	 * is in an unexpected state. We try to catch that here in order to
+	 * reset the Target and retry the probe.
+	 */
+	ath10k_pci_device_reset(ar_pci);
+
+	ret = ath10k_pci_reset_target(ar);
+	if (ret)
+		goto err_intr;
+
+	if (ath10k_target_ps) {
+		ath10k_dbg(ATH10K_DBG_PCI, "on-chip power save enabled\n");
+	} else {
+		/* Force AWAKE forever */
+		ath10k_dbg(ATH10K_DBG_PCI, "on-chip power save disabled\n");
+		ath10k_do_pci_wake(ar);
+	}
+
+	ret = ath10k_pci_ce_init(ar);
+	if (ret)
+		goto err_intr;
+
+	ret = ath10k_pci_init_config(ar);
+	if (ret)
+		goto err_ce;
+
+	ret = ath10k_pci_wake_target_cpu(ar);
+	if (ret) {
+		ath10k_err("could not wake up target CPU (%d)\n", ret);
+		goto err_ce;
+	}
+
+	ret = ath10k_core_register(ar);
+	if (ret) {
+		ath10k_err("could not register driver core (%d)\n", ret);
+		goto err_ce;
+	}
+
+	return 0;
+
+err_ce:
+	ath10k_pci_ce_deinit(ar);
+err_intr:
+	ath10k_pci_stop_intr(ar);
+err_iomap:
+	pci_iounmap(pdev, mem);
+err_master:
+	pci_clear_master(pdev);
+err_region:
+	pci_release_region(pdev, BAR_NUM);
+err_device:
+	pci_disable_device(pdev);
+err_ar:
+	pci_set_drvdata(pdev, NULL);
+	ath10k_core_destroy(ar);
+err_ar_pci:
+	/* call HIF PCI free here */
+	kfree(ar_pci);
+
+	return ret;
+}
+
+static void ath10k_pci_remove(struct pci_dev *pdev)
+{
+	struct ath10k *ar = pci_get_drvdata(pdev);
+	struct ath10k_pci *ar_pci;
+
+	ath10k_dbg(ATH10K_DBG_PCI, "%s\n", __func__);
+
+	if (!ar)
+		return;
+
+	ar_pci = ath10k_pci_priv(ar);
+
+	if (!ar_pci)
+		return;
+
+	tasklet_kill(&ar_pci->msi_fw_err);
+
+	ath10k_core_unregister(ar);
+	ath10k_pci_stop_intr(ar);
+
+	pci_set_drvdata(pdev, NULL);
+	pci_iounmap(pdev, ar_pci->mem);
+	pci_release_region(pdev, BAR_NUM);
+	pci_clear_master(pdev);
+	pci_disable_device(pdev);
+
+	ath10k_core_destroy(ar);
+	kfree(ar_pci);
+}
+
+#if defined(CONFIG_PM_SLEEP)
+
+#define ATH10K_PCI_PM_CONTROL 0x44
+
+static int ath10k_pci_suspend(struct device *device)
+{
+	struct pci_dev *pdev = to_pci_dev(device);
+	struct ath10k *ar = pci_get_drvdata(pdev);
+	struct ath10k_pci *ar_pci;
+	u32 val;
+	int ret, retval;
+
+	ath10k_dbg(ATH10K_DBG_PCI, "%s\n", __func__);
+
+	if (!ar)
+		return -ENODEV;
+
+	ar_pci = ath10k_pci_priv(ar);
+	if (!ar_pci)
+		return -ENODEV;
+
+	if (ath10k_core_target_suspend(ar))
+		return -EBUSY;
+
+	ret = wait_event_interruptible_timeout(ar->event_queue,
+						ar->is_target_paused == true,
+						1 * HZ);
+	if (ret < 0) {
+		ath10k_warn("suspend interrupted (%d)\n", ret);
+		retval = ret;
+		goto resume;
+	} else if (ret == 0) {
+		ath10k_warn("suspend timed out - target pause event never came\n");
+		retval = EIO;
+		goto resume;
+	}
+
+	/*
+	 * reset is_target_paused and host can check that in next time,
+	 * or it will always be TRUE and host just skip the waiting
+	 * condition, it causes target assert due to host already
+	 * suspend
+	 */
+	ar->is_target_paused = false;
+
+	pci_read_config_dword(pdev, ATH10K_PCI_PM_CONTROL, &val);
+
+	if ((val & 0x000000ff) != 0x3) {
+		pci_save_state(pdev);
+		pci_disable_device(pdev);
+		pci_write_config_dword(pdev, ATH10K_PCI_PM_CONTROL,
+				       (val & 0xffffff00) | 0x03);
+	}
+
+	return 0;
+resume:
+	ret = ath10k_core_target_resume(ar);
+	if (ret)
+		ath10k_warn("could not resume (%d)\n", ret);
+
+	return retval;
+}
+
+static int ath10k_pci_resume(struct device *device)
+{
+	struct pci_dev *pdev = to_pci_dev(device);
+	struct ath10k *ar = pci_get_drvdata(pdev);
+	struct ath10k_pci *ar_pci;
+	int ret;
+	u32 val;
+
+	ath10k_dbg(ATH10K_DBG_PCI, "%s\n", __func__);
+
+	if (!ar)
+		return -ENODEV;
+	ar_pci = ath10k_pci_priv(ar);
+
+	if (!ar_pci)
+		return -ENODEV;
+
+	ret = pci_enable_device(pdev);
+	if (ret) {
+		ath10k_warn("cannot enable PCI device: %d\n", ret);
+		return ret;
+	}
+
+	pci_read_config_dword(pdev, ATH10K_PCI_PM_CONTROL, &val);
+
+	if ((val & 0x000000ff) != 0) {
+		pci_restore_state(pdev);
+		pci_write_config_dword(pdev, ATH10K_PCI_PM_CONTROL,
+				       val & 0xffffff00);
+		/*
+		 * Suspend/Resume resets the PCI configuration space,
+		 * so we have to re-disable the RETRY_TIMEOUT register (0x41)
+		 * to keep PCI Tx retries from interfering with C3 CPU state
+		 */
+		pci_read_config_dword(pdev, 0x40, &val);
+
+		if ((val & 0x0000ff00) != 0)
+			pci_write_config_dword(pdev, 0x40, val & 0xffff00ff);
+	}
+
+	ret = ath10k_core_target_resume(ar);
+	if (ret)
+		ath10k_warn("target resume failed: %d\n", ret);
+
+	return ret;
+}
+
+static SIMPLE_DEV_PM_OPS(ath10k_dev_pm_ops,
+			 ath10k_pci_suspend,
+			 ath10k_pci_resume);
+
+#define ATH10K_PCI_PM_OPS (&ath10k_dev_pm_ops)
+
+#else
+
+#define ATH10K_PCI_PM_OPS NULL
+
+#endif /* CONFIG_PM_SLEEP */
+
+MODULE_DEVICE_TABLE(pci, ath10k_pci_id_table);
+
+static struct pci_driver ath10k_pci_driver = {
+	.name = "ath10k_pci",
+	.id_table = ath10k_pci_id_table,
+	.probe = ath10k_pci_probe,
+	.remove = ath10k_pci_remove,
+	.driver.pm = ATH10K_PCI_PM_OPS,
+};
+
+static int __init ath10k_pci_init(void)
+{
+	int ret;
+
+	ret = pci_register_driver(&ath10k_pci_driver);
+	if (ret)
+		ath10k_err("pci_register_driver failed [%d]\n", ret);
+
+	return ret;
+}
+module_init(ath10k_pci_init);
+
+static void __exit ath10k_pci_exit(void)
+{
+	pci_unregister_driver(&ath10k_pci_driver);
+}
+
+module_exit(ath10k_pci_exit);
+
+MODULE_AUTHOR("Qualcomm Atheros");
+MODULE_DESCRIPTION("Driver support for Atheros QCA988X PCIe devices");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_FIRMWARE(QCA988X_HW_1_0_FW_DIR "/" QCA988X_HW_1_0_FW_FILE);
+MODULE_FIRMWARE(QCA988X_HW_1_0_FW_DIR "/" QCA988X_HW_1_0_OTP_FILE);
+MODULE_FIRMWARE(QCA988X_HW_1_0_FW_DIR "/" QCA988X_HW_1_0_BOARD_DATA_FILE);
+MODULE_FIRMWARE(QCA988X_HW_2_0_FW_DIR "/" QCA988X_HW_2_0_FW_FILE);
+MODULE_FIRMWARE(QCA988X_HW_2_0_FW_DIR "/" QCA988X_HW_2_0_OTP_FILE);
+MODULE_FIRMWARE(QCA988X_HW_2_0_FW_DIR "/" QCA988X_HW_2_0_BOARD_DATA_FILE);
diff --git a/drivers/net/wireless/ath/ath10k/pci.h b/drivers/net/wireless/ath/ath10k/pci.h
new file mode 100644
index 0000000..d2a055a
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/pci.h
@@ -0,0 +1,355 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _PCI_H_
+#define _PCI_H_
+
+#include <linux/interrupt.h>
+
+#include "hw.h"
+#include "ce.h"
+
+/* FW dump area */
+#define REG_DUMP_COUNT_QCA988X 60
+
+/*
+ * maximum number of bytes that can be handled atomically by DiagRead/DiagWrite
+ */
+#define DIAG_TRANSFER_LIMIT 2048
+
+/*
+ * maximum number of bytes that can be
+ * handled atomically by DiagRead/DiagWrite
+ */
+#define DIAG_TRANSFER_LIMIT 2048
+
+struct bmi_xfer {
+	struct completion done;
+	bool wait_for_resp;
+	u32 resp_len;
+};
+
+struct ath10k_pci_compl {
+	struct list_head list;
+	int send_or_recv;
+	struct ce_state *ce_state;
+	struct hif_ce_pipe_info *pipe_info;
+	void *transfer_context;
+	unsigned int nbytes;
+	unsigned int transfer_id;
+	unsigned int flags;
+};
+
+/* compl_state.send_or_recv */
+#define HIF_CE_COMPLETE_FREE 0
+#define HIF_CE_COMPLETE_SEND 1
+#define HIF_CE_COMPLETE_RECV 2
+
+/*
+ * PCI-specific Target state
+ *
+ * NOTE: Structure is shared between Host software and Target firmware!
+ *
+ * Much of this may be of interest to the Host so
+ * HOST_INTEREST->hi_interconnect_state points here
+ * (and all members are 32-bit quantities in order to
+ * facilitate Host access). In particular, Host software is
+ * required to initialize pipe_cfg_addr and svc_to_pipe_map.
+ */
+struct pcie_state {
+	/* Pipe configuration Target address */
+	/* NB: ce_pipe_config[CE_COUNT] */
+	u32 pipe_cfg_addr;
+
+	/* Service to pipe map Target address */
+	/* NB: service_to_pipe[PIPE_TO_CE_MAP_CN] */
+	u32 svc_to_pipe_map;
+
+	/* number of MSI interrupts requested */
+	u32 msi_requested;
+
+	/* number of MSI interrupts granted */
+	u32 msi_granted;
+
+	/* Message Signalled Interrupt address */
+	u32 msi_addr;
+
+	/* Base data */
+	u32 msi_data;
+
+	/*
+	 * Data for firmware interrupt;
+	 * MSI data for other interrupts are
+	 * in various SoC registers
+	 */
+	u32 msi_fw_intr_data;
+
+	/* PCIE_PWR_METHOD_* */
+	u32 power_mgmt_method;
+
+	/* PCIE_CONFIG_FLAG_* */
+	u32 config_flags;
+};
+
+/* PCIE_CONFIG_FLAG definitions */
+#define PCIE_CONFIG_FLAG_ENABLE_L1  0x0000001
+
+/* Host software's Copy Engine configuration. */
+#define CE_ATTR_FLAGS 0
+
+/*
+ * Configuration information for a Copy Engine pipe.
+ * Passed from Host to Target during startup (one per CE).
+ *
+ * NOTE: Structure is shared between Host software and Target firmware!
+ */
+struct ce_pipe_config {
+	u32 pipenum;
+	u32 pipedir;
+	u32 nentries;
+	u32 nbytes_max;
+	u32 flags;
+	u32 reserved;
+};
+
+/*
+ * Directions for interconnect pipe configuration.
+ * These definitions may be used during configuration and are shared
+ * between Host and Target.
+ *
+ * Pipe Directions are relative to the Host, so PIPEDIR_IN means
+ * "coming IN over air through Target to Host" as with a WiFi Rx operation.
+ * Conversely, PIPEDIR_OUT means "going OUT from Host through Target over air"
+ * as with a WiFi Tx operation. This is somewhat awkward for the "middle-man"
+ * Target since things that are "PIPEDIR_OUT" are coming IN to the Target
+ * over the interconnect.
+ */
+#define PIPEDIR_NONE    0
+#define PIPEDIR_IN      1  /* Target-->Host, WiFi Rx direction */
+#define PIPEDIR_OUT     2  /* Host->Target, WiFi Tx direction */
+#define PIPEDIR_INOUT   3  /* bidirectional */
+
+/* Establish a mapping between a service/direction and a pipe. */
+struct service_to_pipe {
+	u32 service_id;
+	u32 pipedir;
+	u32 pipenum;
+};
+
+enum ath10k_pci_features {
+	ATH10K_PCI_FEATURE_MSI_X		= 0,
+	ATH10K_PCI_FEATURE_HW_1_0_WARKAROUND	= 1,
+
+	/* keep last */
+	ATH10K_PCI_FEATURE_COUNT
+};
+
+/* Per-pipe state. */
+struct hif_ce_pipe_info {
+	/* Handle of underlying Copy Engine */
+	struct ce_state *ce_hdl;
+
+	/* Our pipe number; facilitiates use of pipe_info ptrs. */
+	u8 pipe_num;
+
+	/* Convenience back pointer to hif_ce_state. */
+	struct ath10k *hif_ce_state;
+
+	size_t buf_sz;
+
+	/* protects compl_free and num_send_allowed */
+	spinlock_t pipe_lock;
+
+	/* List of free CE completion slots */
+	struct list_head compl_free;
+
+	/* Limit the number of outstanding send requests. */
+	int num_sends_allowed;
+
+	struct ath10k_pci *ar_pci;
+	struct tasklet_struct intr;
+};
+
+struct ath10k_pci {
+	struct pci_dev *pdev;
+	struct device *dev;
+	struct ath10k *ar;
+	void __iomem *mem;
+	int cacheline_sz;
+
+	DECLARE_BITMAP(features, ATH10K_PCI_FEATURE_COUNT);
+
+	/*
+	 * Number of MSI interrupts granted, 0 --> using legacy PCI line
+	 * interrupts.
+	 */
+	int num_msi_intrs;
+
+	struct tasklet_struct intr_tq;
+	struct tasklet_struct msi_fw_err;
+
+	/* Number of Copy Engines supported */
+	unsigned int ce_count;
+
+	int started;
+
+	atomic_t keep_awake_count;
+	bool verified_awake;
+
+	/* List of CE completions to be processed */
+	struct list_head compl_process;
+
+	/* protects compl_processing and compl_process */
+	spinlock_t compl_lock;
+
+	bool compl_processing;
+
+	struct hif_ce_pipe_info pipe_info[CE_COUNT_MAX];
+
+	struct ath10k_hif_cb msg_callbacks_current;
+
+	/* Target address used to signal a pending firmware event */
+	u32 fw_indicator_address;
+
+	/* Copy Engine used for Diagnostic Accesses */
+	struct ce_state *ce_diag;
+
+	/* FIXME: document what this really protects */
+	spinlock_t ce_lock;
+
+	/* Map CE id to ce_state */
+	struct ce_state *ce_id_to_state[CE_COUNT_MAX];
+
+	/* makes sure that dummy reads are atomic */
+	spinlock_t hw_v1_workaround_lock;
+};
+
+static inline struct ath10k_pci *ath10k_pci_priv(struct ath10k *ar)
+{
+	return ar->hif.priv;
+}
+
+static inline u32 ath10k_pci_reg_read32(void __iomem *mem, u32 addr)
+{
+	return ioread32(mem + PCIE_LOCAL_BASE_ADDRESS + addr);
+}
+
+static inline void ath10k_pci_reg_write32(void __iomem *mem, u32 addr, u32 val)
+{
+	iowrite32(val, mem + PCIE_LOCAL_BASE_ADDRESS + addr);
+}
+
+#define ATH_PCI_RESET_WAIT_MAX 10 /* ms */
+#define PCIE_WAKE_TIMEOUT 5000	/* 5ms */
+
+#define BAR_NUM 0
+
+#define CDC_WAR_MAGIC_STR   0xceef0000
+#define CDC_WAR_DATA_CE     4
+
+/*
+ * TODO: Should be a function call specific to each Target-type.
+ * This convoluted macro converts from Target CPU Virtual Address Space to CE
+ * Address Space. As part of this process, we conservatively fetch the current
+ * PCIE_BAR. MOST of the time, this should match the upper bits of PCI space
+ * for this device; but that's not guaranteed.
+ */
+#define TARG_CPU_SPACE_TO_CE_SPACE(ar, pci_addr, addr)			\
+	(((ioread32((pci_addr)+(SOC_CORE_BASE_ADDRESS|			\
+	  CORE_CTRL_ADDRESS)) & 0x7ff) << 21) |				\
+	 0x100000 | ((addr) & 0xfffff))
+
+/* Wait up to this many Ms for a Diagnostic Access CE operation to complete */
+#define DIAG_ACCESS_CE_TIMEOUT_MS 10
+
+/*
+ * This API allows the Host to access Target registers directly
+ * and relatively efficiently over PCIe.
+ * This allows the Host to avoid extra overhead associated with
+ * sending a message to firmware and waiting for a response message
+ * from firmware, as is done on other interconnects.
+ *
+ * Yet there is some complexity with direct accesses because the
+ * Target's power state is not known a priori. The Host must issue
+ * special PCIe reads/writes in order to explicitly wake the Target
+ * and to verify that it is awake and will remain awake.
+ *
+ * Usage:
+ *
+ *   Use ath10k_pci_read32 and ath10k_pci_write32 to access Target space.
+ *   These calls must be bracketed by ath10k_pci_wake and
+ *   ath10k_pci_sleep.  A single BEGIN/END pair is adequate for
+ *   multiple READ/WRITE operations.
+ *
+ *   Use ath10k_pci_wake to put the Target in a state in
+ *   which it is legal for the Host to directly access it. This
+ *   may involve waking the Target from a low power state, which
+ *   may take up to 2Ms!
+ *
+ *   Use ath10k_pci_sleep to tell the Target that as far as
+ *   this code path is concerned, it no longer needs to remain
+ *   directly accessible.  BEGIN/END is under a reference counter;
+ *   multiple code paths may issue BEGIN/END on a single targid.
+ */
+static inline void ath10k_pci_write32(struct ath10k *ar, u32 offset,
+				      u32 value)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+	void __iomem *addr = ar_pci->mem;
+
+	if (test_bit(ATH10K_PCI_FEATURE_HW_1_0_WARKAROUND, ar_pci->features)) {
+		unsigned long irq_flags;
+
+		spin_lock_irqsave(&ar_pci->hw_v1_workaround_lock, irq_flags);
+
+		ioread32(addr+offset+4); /* 3rd read prior to write */
+		ioread32(addr+offset+4); /* 2nd read prior to write */
+		ioread32(addr+offset+4); /* 1st read prior to write */
+		iowrite32(value, addr+offset);
+
+		spin_unlock_irqrestore(&ar_pci->hw_v1_workaround_lock,
+				       irq_flags);
+	} else {
+		iowrite32(value, addr+offset);
+	}
+}
+
+static inline u32 ath10k_pci_read32(struct ath10k *ar, u32 offset)
+{
+	struct ath10k_pci *ar_pci = ath10k_pci_priv(ar);
+
+	return ioread32(ar_pci->mem + offset);
+}
+
+extern unsigned int ath10k_target_ps;
+
+void ath10k_do_pci_wake(struct ath10k *ar);
+void ath10k_do_pci_sleep(struct ath10k *ar);
+
+static inline void ath10k_pci_wake(struct ath10k *ar)
+{
+	if (ath10k_target_ps)
+		ath10k_do_pci_wake(ar);
+}
+
+static inline void ath10k_pci_sleep(struct ath10k *ar)
+{
+	if (ath10k_target_ps)
+		ath10k_do_pci_sleep(ar);
+}
+
+#endif /* _PCI_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/rx_desc.h b/drivers/net/wireless/ath/ath10k/rx_desc.h
new file mode 100644
index 0000000..bfec6c8
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/rx_desc.h
@@ -0,0 +1,990 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _RX_DESC_H_
+#define _RX_DESC_H_
+
+enum rx_attention_flags {
+	RX_ATTENTION_FLAGS_FIRST_MPDU          = 1 << 0,
+	RX_ATTENTION_FLAGS_LAST_MPDU           = 1 << 1,
+	RX_ATTENTION_FLAGS_MCAST_BCAST         = 1 << 2,
+	RX_ATTENTION_FLAGS_PEER_IDX_INVALID    = 1 << 3,
+	RX_ATTENTION_FLAGS_PEER_IDX_TIMEOUT    = 1 << 4,
+	RX_ATTENTION_FLAGS_POWER_MGMT          = 1 << 5,
+	RX_ATTENTION_FLAGS_NON_QOS             = 1 << 6,
+	RX_ATTENTION_FLAGS_NULL_DATA           = 1 << 7,
+	RX_ATTENTION_FLAGS_MGMT_TYPE           = 1 << 8,
+	RX_ATTENTION_FLAGS_CTRL_TYPE           = 1 << 9,
+	RX_ATTENTION_FLAGS_MORE_DATA           = 1 << 10,
+	RX_ATTENTION_FLAGS_EOSP                = 1 << 11,
+	RX_ATTENTION_FLAGS_U_APSD_TRIGGER      = 1 << 12,
+	RX_ATTENTION_FLAGS_FRAGMENT            = 1 << 13,
+	RX_ATTENTION_FLAGS_ORDER               = 1 << 14,
+	RX_ATTENTION_FLAGS_CLASSIFICATION      = 1 << 15,
+	RX_ATTENTION_FLAGS_OVERFLOW_ERR        = 1 << 16,
+	RX_ATTENTION_FLAGS_MSDU_LENGTH_ERR     = 1 << 17,
+	RX_ATTENTION_FLAGS_TCP_UDP_CHKSUM_FAIL = 1 << 18,
+	RX_ATTENTION_FLAGS_IP_CHKSUM_FAIL      = 1 << 19,
+	RX_ATTENTION_FLAGS_SA_IDX_INVALID      = 1 << 20,
+	RX_ATTENTION_FLAGS_DA_IDX_INVALID      = 1 << 21,
+	RX_ATTENTION_FLAGS_SA_IDX_TIMEOUT      = 1 << 22,
+	RX_ATTENTION_FLAGS_DA_IDX_TIMEOUT      = 1 << 23,
+	RX_ATTENTION_FLAGS_ENCRYPT_REQUIRED    = 1 << 24,
+	RX_ATTENTION_FLAGS_DIRECTED            = 1 << 25,
+	RX_ATTENTION_FLAGS_BUFFER_FRAGMENT     = 1 << 26,
+	RX_ATTENTION_FLAGS_MPDU_LENGTH_ERR     = 1 << 27,
+	RX_ATTENTION_FLAGS_TKIP_MIC_ERR        = 1 << 28,
+	RX_ATTENTION_FLAGS_DECRYPT_ERR         = 1 << 29,
+	RX_ATTENTION_FLAGS_FCS_ERR             = 1 << 30,
+	RX_ATTENTION_FLAGS_MSDU_DONE           = 1 << 31,
+};
+
+struct rx_attention {
+	__le32 flags; /* %RX_ATTENTION_FLAGS_ */
+} __packed;
+
+/*
+ * first_mpdu
+ *		Indicates the first MSDU of the PPDU.  If both first_mpdu
+ *		and last_mpdu are set in the MSDU then this is a not an
+ *		A-MPDU frame but a stand alone MPDU.  Interior MPDU in an
+ *		A-MPDU shall have both first_mpdu and last_mpdu bits set to
+ *		0.  The PPDU start status will only be valid when this bit
+ *		is set.
+ *
+ * last_mpdu
+ *		Indicates the last MSDU of the last MPDU of the PPDU.  The
+ *		PPDU end status will only be valid when this bit is set.
+ *
+ * mcast_bcast
+ *		Multicast / broadcast indicator.  Only set when the MAC
+ *		address 1 bit 0 is set indicating mcast/bcast and the BSSID
+ *		matches one of the 4 BSSID registers. Only set when
+ *		first_msdu is set.
+ *
+ * peer_idx_invalid
+ *		Indicates no matching entries within the the max search
+ *		count.  Only set when first_msdu is set.
+ *
+ * peer_idx_timeout
+ *		Indicates an unsuccessful search for the peer index due to
+ *		timeout.  Only set when first_msdu is set.
+ *
+ * power_mgmt
+ *		Power management bit set in the 802.11 header.  Only set
+ *		when first_msdu is set.
+ *
+ * non_qos
+ *		Set if packet is not a non-QoS data frame.  Only set when
+ *		first_msdu is set.
+ *
+ * null_data
+ *		Set if frame type indicates either null data or QoS null
+ *		data format.  Only set when first_msdu is set.
+ *
+ * mgmt_type
+ *		Set if packet is a management packet.  Only set when
+ *		first_msdu is set.
+ *
+ * ctrl_type
+ *		Set if packet is a control packet.  Only set when first_msdu
+ *		is set.
+ *
+ * more_data
+ *		Set if more bit in frame control is set.  Only set when
+ *		first_msdu is set.
+ *
+ * eosp
+ *		Set if the EOSP (end of service period) bit in the QoS
+ *		control field is set.  Only set when first_msdu is set.
+ *
+ * u_apsd_trigger
+ *		Set if packet is U-APSD trigger.  Key table will have bits
+ *		per TID to indicate U-APSD trigger.
+ *
+ * fragment
+ *		Indicates that this is an 802.11 fragment frame.  This is
+ *		set when either the more_frag bit is set in the frame
+ *		control or the fragment number is not zero.  Only set when
+ *		first_msdu is set.
+ *
+ * order
+ *		Set if the order bit in the frame control is set.  Only set
+ *		when first_msdu is set.
+ *
+ * classification
+ *		Indicates that this status has a corresponding MSDU that
+ *		requires FW processing.  The OLE will have classification
+ *		ring mask registers which will indicate the ring(s) for
+ *		packets and descriptors which need FW attention.
+ *
+ * overflow_err
+ *		PCU Receive FIFO does not have enough space to store the
+ *		full receive packet.  Enough space is reserved in the
+ *		receive FIFO for the status is written.  This MPDU remaining
+ *		packets in the PPDU will be filtered and no Ack response
+ *		will be transmitted.
+ *
+ * msdu_length_err
+ *		Indicates that the MSDU length from the 802.3 encapsulated
+ *		length field extends beyond the MPDU boundary.
+ *
+ * tcp_udp_chksum_fail
+ *		Indicates that the computed checksum (tcp_udp_chksum) did
+ *		not match the checksum in the TCP/UDP header.
+ *
+ * ip_chksum_fail
+ *		Indicates that the computed checksum did not match the
+ *		checksum in the IP header.
+ *
+ * sa_idx_invalid
+ *		Indicates no matching entry was found in the address search
+ *		table for the source MAC address.
+ *
+ * da_idx_invalid
+ *		Indicates no matching entry was found in the address search
+ *		table for the destination MAC address.
+ *
+ * sa_idx_timeout
+ *		Indicates an unsuccessful search for the source MAC address
+ *		due to the expiring of the search timer.
+ *
+ * da_idx_timeout
+ *		Indicates an unsuccessful search for the destination MAC
+ *		address due to the expiring of the search timer.
+ *
+ * encrypt_required
+ *		Indicates that this data type frame is not encrypted even if
+ *		the policy for this MPDU requires encryption as indicated in
+ *		the peer table key type.
+ *
+ * directed
+ *		MPDU is a directed packet which means that the RA matched
+ *		our STA addresses.  In proxySTA it means that the TA matched
+ *		an entry in our address search table with the corresponding
+ *		'no_ack' bit is the address search entry cleared.
+ *
+ * buffer_fragment
+ *		Indicates that at least one of the rx buffers has been
+ *		fragmented.  If set the FW should look at the rx_frag_info
+ *		descriptor described below.
+ *
+ * mpdu_length_err
+ *		Indicates that the MPDU was pre-maturely terminated
+ *		resulting in a truncated MPDU.  Don't trust the MPDU length
+ *		field.
+ *
+ * tkip_mic_err
+ *		Indicates that the MPDU Michael integrity check failed
+ *
+ * decrypt_err
+ *		Indicates that the MPDU decrypt integrity check failed
+ *
+ * fcs_err
+ *		Indicates that the MPDU FCS check failed
+ *
+ * msdu_done
+ *		If set indicates that the RX packet data, RX header data, RX
+ *		PPDU start descriptor, RX MPDU start/end descriptor, RX MSDU
+ *		start/end descriptors and RX Attention descriptor are all
+ *		valid.  This bit must be in the last octet of the
+ *		descriptor.
+ */
+
+struct rx_frag_info {
+	u8 ring0_more_count;
+	u8 ring1_more_count;
+	u8 ring2_more_count;
+	u8 ring3_more_count;
+} __packed;
+
+/*
+ * ring0_more_count
+ *		Indicates the number of more buffers associated with RX DMA
+ *		ring 0.  Field is filled in by the RX_DMA.
+ *
+ * ring1_more_count
+ *		Indicates the number of more buffers associated with RX DMA
+ *		ring 1. Field is filled in by the RX_DMA.
+ *
+ * ring2_more_count
+ *		Indicates the number of more buffers associated with RX DMA
+ *		ring 2. Field is filled in by the RX_DMA.
+ *
+ * ring3_more_count
+ *		Indicates the number of more buffers associated with RX DMA
+ *		ring 3. Field is filled in by the RX_DMA.
+ */
+
+enum htt_rx_mpdu_encrypt_type {
+	HTT_RX_MPDU_ENCRYPT_WEP40            = 0,
+	HTT_RX_MPDU_ENCRYPT_WEP104           = 1,
+	HTT_RX_MPDU_ENCRYPT_TKIP_WITHOUT_MIC = 2,
+	HTT_RX_MPDU_ENCRYPT_WEP128           = 3,
+	HTT_RX_MPDU_ENCRYPT_TKIP_WPA         = 4,
+	HTT_RX_MPDU_ENCRYPT_WAPI             = 5,
+	HTT_RX_MPDU_ENCRYPT_AES_CCM_WPA2     = 6,
+	HTT_RX_MPDU_ENCRYPT_NONE             = 7,
+};
+
+#define RX_MPDU_START_INFO0_PEER_IDX_MASK     0x000007ff
+#define RX_MPDU_START_INFO0_PEER_IDX_LSB      0
+#define RX_MPDU_START_INFO0_SEQ_NUM_MASK      0x0fff0000
+#define RX_MPDU_START_INFO0_SEQ_NUM_LSB       16
+#define RX_MPDU_START_INFO0_ENCRYPT_TYPE_MASK 0xf0000000
+#define RX_MPDU_START_INFO0_ENCRYPT_TYPE_LSB  28
+#define RX_MPDU_START_INFO0_FROM_DS           (1 << 11)
+#define RX_MPDU_START_INFO0_TO_DS             (1 << 12)
+#define RX_MPDU_START_INFO0_ENCRYPTED         (1 << 13)
+#define RX_MPDU_START_INFO0_RETRY             (1 << 14)
+#define RX_MPDU_START_INFO0_TXBF_H_INFO       (1 << 15)
+
+#define RX_MPDU_START_INFO1_TID_MASK 0xf0000000
+#define RX_MPDU_START_INFO1_TID_LSB  28
+#define RX_MPDU_START_INFO1_DIRECTED (1 << 16)
+
+struct rx_mpdu_start {
+	__le32 info0;
+	union {
+		struct {
+			__le32 pn31_0;
+			__le32 info1; /* %RX_MPDU_START_INFO1_ */
+		} __packed;
+		struct {
+			u8 pn[6];
+		} __packed;
+	} __packed;
+} __packed;
+
+/*
+ * peer_idx
+ *		The index of the address search table which associated with
+ *		the peer table entry corresponding to this MPDU.  Only valid
+ *		when first_msdu is set.
+ *
+ * fr_ds
+ *		Set if the from DS bit is set in the frame control.  Only
+ *		valid when first_msdu is set.
+ *
+ * to_ds
+ *		Set if the to DS bit is set in the frame control.  Only
+ *		valid when first_msdu is set.
+ *
+ * encrypted
+ *		Protected bit from the frame control.  Only valid when
+ *		first_msdu is set.
+ *
+ * retry
+ *		Retry bit from the frame control.  Only valid when
+ *		first_msdu is set.
+ *
+ * txbf_h_info
+ *		The MPDU data will contain H information.  Primarily used
+ *		for debug.
+ *
+ * seq_num
+ *		The sequence number from the 802.11 header.  Only valid when
+ *		first_msdu is set.
+ *
+ * encrypt_type
+ *		Indicates type of decrypt cipher used (as defined in the
+ *		peer table)
+ *		0: WEP40
+ *		1: WEP104
+ *		2: TKIP without MIC
+ *		3: WEP128
+ *		4: TKIP (WPA)
+ *		5: WAPI
+ *		6: AES-CCM (WPA2)
+ *		7: No cipher
+ *		Only valid when first_msdu_is set
+ *
+ * pn_31_0
+ *		Bits [31:0] of the PN number extracted from the IV field
+ *		WEP: IV = {key_id_octet, pn2, pn1, pn0}.  Only pn[23:0] is
+ *		valid.
+ *		TKIP: IV = {pn5, pn4, pn3, pn2, key_id_octet, pn0,
+ *		WEPSeed[1], pn1}.  Only pn[47:0] is valid.
+ *		AES-CCM: IV = {pn5, pn4, pn3, pn2, key_id_octet, 0x0, pn1,
+ *		pn0}.  Only pn[47:0] is valid.
+ *		WAPI: IV = {key_id_octet, 0x0, pn15, pn14, pn13, pn12, pn11,
+ *		pn10, pn9, pn8, pn7, pn6, pn5, pn4, pn3, pn2, pn1, pn0}.
+ *		The ext_wapi_pn[127:48] in the rx_msdu_misc descriptor and
+ *		pn[47:0] are valid.
+ *		Only valid when first_msdu is set.
+ *
+ * pn_47_32
+ *		Bits [47:32] of the PN number.   See description for
+ *		pn_31_0.  The remaining PN fields are in the rx_msdu_end
+ *		descriptor
+ *
+ * pn
+ *		Use this field to access the pn without worrying about
+ *		byte-order and bitmasking/bitshifting.
+ *
+ * directed
+ *		See definition in RX attention descriptor
+ *
+ * reserved_2
+ *		Reserved: HW should fill with zero.  FW should ignore.
+ *
+ * tid
+ *		The TID field in the QoS control field
+ */
+
+#define RX_MPDU_END_INFO0_RESERVED_0_MASK     0x00001fff
+#define RX_MPDU_END_INFO0_RESERVED_0_LSB      0
+#define RX_MPDU_END_INFO0_POST_DELIM_CNT_MASK 0x0fff0000
+#define RX_MPDU_END_INFO0_POST_DELIM_CNT_LSB  16
+#define RX_MPDU_END_INFO0_OVERFLOW_ERR        (1 << 13)
+#define RX_MPDU_END_INFO0_LAST_MPDU           (1 << 14)
+#define RX_MPDU_END_INFO0_POST_DELIM_ERR      (1 << 15)
+#define RX_MPDU_END_INFO0_MPDU_LENGTH_ERR     (1 << 28)
+#define RX_MPDU_END_INFO0_TKIP_MIC_ERR        (1 << 29)
+#define RX_MPDU_END_INFO0_DECRYPT_ERR         (1 << 30)
+#define RX_MPDU_END_INFO0_FCS_ERR             (1 << 31)
+
+struct rx_mpdu_end {
+	__le32 info0;
+} __packed;
+
+/*
+ * reserved_0
+ *		Reserved
+ *
+ * overflow_err
+ *		PCU Receive FIFO does not have enough space to store the
+ *		full receive packet.  Enough space is reserved in the
+ *		receive FIFO for the status is written.  This MPDU remaining
+ *		packets in the PPDU will be filtered and no Ack response
+ *		will be transmitted.
+ *
+ * last_mpdu
+ *		Indicates that this is the last MPDU of a PPDU.
+ *
+ * post_delim_err
+ *		Indicates that a delimiter FCS error occurred after this
+ *		MPDU before the next MPDU.  Only valid when last_msdu is
+ *		set.
+ *
+ * post_delim_cnt
+ *		Count of the delimiters after this MPDU.  This requires the
+ *		last MPDU to be held until all the EOF descriptors have been
+ *		received.  This may be inefficient in the future when
+ *		ML-MIMO is used.  Only valid when last_mpdu is set.
+ *
+ * mpdu_length_err
+ *		See definition in RX attention descriptor
+ *
+ * tkip_mic_err
+ *		See definition in RX attention descriptor
+ *
+ * decrypt_err
+ *		See definition in RX attention descriptor
+ *
+ * fcs_err
+ *		See definition in RX attention descriptor
+ */
+
+#define RX_MSDU_START_INFO0_MSDU_LENGTH_MASK    0x00003fff
+#define RX_MSDU_START_INFO0_MSDU_LENGTH_LSB     0
+#define RX_MSDU_START_INFO0_IP_OFFSET_MASK      0x000fc000
+#define RX_MSDU_START_INFO0_IP_OFFSET_LSB       14
+#define RX_MSDU_START_INFO0_RING_MASK_MASK      0x00f00000
+#define RX_MSDU_START_INFO0_RING_MASK_LSB       20
+#define RX_MSDU_START_INFO0_TCP_UDP_OFFSET_MASK 0x7f000000
+#define RX_MSDU_START_INFO0_TCP_UDP_OFFSET_LSB  24
+
+#define RX_MSDU_START_INFO1_MSDU_NUMBER_MASK    0x000000ff
+#define RX_MSDU_START_INFO1_MSDU_NUMBER_LSB     0
+#define RX_MSDU_START_INFO1_DECAP_FORMAT_MASK   0x00000300
+#define RX_MSDU_START_INFO1_DECAP_FORMAT_LSB    8
+#define RX_MSDU_START_INFO1_SA_IDX_MASK         0x07ff0000
+#define RX_MSDU_START_INFO1_SA_IDX_LSB          16
+#define RX_MSDU_START_INFO1_IPV4_PROTO          (1 << 10)
+#define RX_MSDU_START_INFO1_IPV6_PROTO          (1 << 11)
+#define RX_MSDU_START_INFO1_TCP_PROTO           (1 << 12)
+#define RX_MSDU_START_INFO1_UDP_PROTO           (1 << 13)
+#define RX_MSDU_START_INFO1_IP_FRAG             (1 << 14)
+#define RX_MSDU_START_INFO1_TCP_ONLY_ACK        (1 << 15)
+
+enum rx_msdu_decap_format {
+	RX_MSDU_DECAP_RAW           = 0,
+	RX_MSDU_DECAP_NATIVE_WIFI   = 1,
+	RX_MSDU_DECAP_ETHERNET2_DIX = 2,
+	RX_MSDU_DECAP_8023_SNAP_LLC = 3
+};
+
+struct rx_msdu_start {
+	__le32 info0; /* %RX_MSDU_START_INFO0_ */
+	__le32 flow_id_crc;
+	__le32 info1; /* %RX_MSDU_START_INFO1_ */
+} __packed;
+
+/*
+ * msdu_length
+ *		MSDU length in bytes after decapsulation.  This field is
+ *		still valid for MPDU frames without A-MSDU.  It still
+ *		represents MSDU length after decapsulation
+ *
+ * ip_offset
+ *		Indicates the IP offset in bytes from the start of the
+ *		packet after decapsulation.  Only valid if ipv4_proto or
+ *		ipv6_proto is set.
+ *
+ * ring_mask
+ *		Indicates the destination RX rings for this MSDU.
+ *
+ * tcp_udp_offset
+ *		Indicates the offset in bytes to the start of TCP or UDP
+ *		header from the start of the IP header after decapsulation.
+ *		Only valid if tcp_prot or udp_prot is set.  The value 0
+ *		indicates that the offset is longer than 127 bytes.
+ *
+ * reserved_0c
+ *		Reserved: HW should fill with zero.  FW should ignore.
+ *
+ * flow_id_crc
+ *		The flow_id_crc runs CRC32 on the following information:
+ *		IPv4 option: dest_addr[31:0], src_addr [31:0], {24'b0,
+ *		protocol[7:0]}.
+ *		IPv6 option: dest_addr[127:0], src_addr [127:0], {24'b0,
+ *		next_header[7:0]}
+ *		UDP case: sort_port[15:0], dest_port[15:0]
+ *		TCP case: sort_port[15:0], dest_port[15:0],
+ *		{header_length[3:0], 6'b0, flags[5:0], window_size[15:0]},
+ *		{16'b0, urgent_ptr[15:0]}, all options except 32-bit
+ *		timestamp.
+ *
+ * msdu_number
+ *		Indicates the MSDU number within a MPDU.  This value is
+ *		reset to zero at the start of each MPDU.  If the number of
+ *		MSDU exceeds 255 this number will wrap using modulo 256.
+ *
+ * decap_format
+ *		Indicates the format after decapsulation:
+ *		0: RAW: No decapsulation
+ *		1: Native WiFi
+ *		2: Ethernet 2 (DIX)
+ *		3: 802.3 (SNAP/LLC)
+ *
+ * ipv4_proto
+ *		Set if L2 layer indicates IPv4 protocol.
+ *
+ * ipv6_proto
+ *		Set if L2 layer indicates IPv6 protocol.
+ *
+ * tcp_proto
+ *		Set if the ipv4_proto or ipv6_proto are set and the IP
+ *		protocol indicates TCP.
+ *
+ * udp_proto
+ *		Set if the ipv4_proto or ipv6_proto are set and the IP
+ *			protocol indicates UDP.
+ *
+ * ip_frag
+ *		Indicates that either the IP More frag bit is set or IP frag
+ *		number is non-zero.  If set indicates that this is a
+ *		fragmented IP packet.
+ *
+ * tcp_only_ack
+ *		Set if only the TCP Ack bit is set in the TCP flags and if
+ *		the TCP payload is 0.
+ *
+ * sa_idx
+ *		The offset in the address table which matches the MAC source
+ *		address.
+ *
+ * reserved_2b
+ *		Reserved: HW should fill with zero.  FW should ignore.
+ */
+
+#define RX_MSDU_END_INFO0_REPORTED_MPDU_LENGTH_MASK 0x00003fff
+#define RX_MSDU_END_INFO0_REPORTED_MPDU_LENGTH_LSB  0
+#define RX_MSDU_END_INFO0_FIRST_MSDU                (1 << 14)
+#define RX_MSDU_END_INFO0_LAST_MSDU                 (1 << 15)
+#define RX_MSDU_END_INFO0_PRE_DELIM_ERR             (1 << 30)
+#define RX_MSDU_END_INFO0_RESERVED_3B               (1 << 31)
+
+struct rx_msdu_end {
+	__le16 ip_hdr_cksum;
+	__le16 tcp_hdr_cksum;
+	u8 key_id_octet;
+	u8 classification_filter;
+	u8 wapi_pn[10];
+	__le32 info0;
+} __packed;
+
+/*
+ *ip_hdr_chksum
+ *		This can include the IP header checksum or the pseudo header
+ *		checksum used by TCP/UDP checksum.
+ *
+ *tcp_udp_chksum
+ *		The value of the computed TCP/UDP checksum.  A mode bit
+ *		selects whether this checksum is the full checksum or the
+ *		partial checksum which does not include the pseudo header.
+ *
+ *key_id_octet
+ *		The key ID octet from the IV.  Only valid when first_msdu is
+ *		set.
+ *
+ *classification_filter
+ *		Indicates the number classification filter rule
+ *
+ *ext_wapi_pn_63_48
+ *		Extension PN (packet number) which is only used by WAPI.
+ *		This corresponds to WAPI PN bits [63:48] (pn6 and pn7).  The
+ *		WAPI PN bits [63:0] are in the pn field of the rx_mpdu_start
+ *		descriptor.
+ *
+ *ext_wapi_pn_95_64
+ *		Extension PN (packet number) which is only used by WAPI.
+ *		This corresponds to WAPI PN bits [95:64] (pn8, pn9, pn10 and
+ *		pn11).
+ *
+ *ext_wapi_pn_127_96
+ *		Extension PN (packet number) which is only used by WAPI.
+ *		This corresponds to WAPI PN bits [127:96] (pn12, pn13, pn14,
+ *		pn15).
+ *
+ *reported_mpdu_length
+ *		MPDU length before decapsulation.  Only valid when
+ *		first_msdu is set.  This field is taken directly from the
+ *		length field of the A-MPDU delimiter or the preamble length
+ *		field for non-A-MPDU frames.
+ *
+ *first_msdu
+ *		Indicates the first MSDU of A-MSDU.  If both first_msdu and
+ *		last_msdu are set in the MSDU then this is a non-aggregated
+ *		MSDU frame: normal MPDU.  Interior MSDU in an A-MSDU shall
+ *		have both first_mpdu and last_mpdu bits set to 0.
+ *
+ *last_msdu
+ *		Indicates the last MSDU of the A-MSDU.  MPDU end status is
+ *		only valid when last_msdu is set.
+ *
+ *reserved_3a
+ *		Reserved: HW should fill with zero.  FW should ignore.
+ *
+ *pre_delim_err
+ *		Indicates that the first delimiter had a FCS failure.  Only
+ *		valid when first_mpdu and first_msdu are set.
+ *
+ *reserved_3b
+ *		Reserved: HW should fill with zero.  FW should ignore.
+ */
+
+#define RX_PPDU_START_SIG_RATE_SELECT_OFDM 0
+#define RX_PPDU_START_SIG_RATE_SELECT_CCK  1
+
+#define RX_PPDU_START_SIG_RATE_OFDM_48 0
+#define RX_PPDU_START_SIG_RATE_OFDM_24 1
+#define RX_PPDU_START_SIG_RATE_OFDM_12 2
+#define RX_PPDU_START_SIG_RATE_OFDM_6  3
+#define RX_PPDU_START_SIG_RATE_OFDM_54 4
+#define RX_PPDU_START_SIG_RATE_OFDM_36 5
+#define RX_PPDU_START_SIG_RATE_OFDM_18 6
+#define RX_PPDU_START_SIG_RATE_OFDM_9  7
+
+#define RX_PPDU_START_SIG_RATE_CCK_LP_11  0
+#define RX_PPDU_START_SIG_RATE_CCK_LP_5_5 1
+#define RX_PPDU_START_SIG_RATE_CCK_LP_2   2
+#define RX_PPDU_START_SIG_RATE_CCK_LP_1   3
+#define RX_PPDU_START_SIG_RATE_CCK_SP_11  4
+#define RX_PPDU_START_SIG_RATE_CCK_SP_5_5 5
+#define RX_PPDU_START_SIG_RATE_CCK_SP_2   6
+
+#define HTT_RX_PPDU_START_PREAMBLE_LEGACY        0x04
+#define HTT_RX_PPDU_START_PREAMBLE_HT            0x08
+#define HTT_RX_PPDU_START_PREAMBLE_HT_WITH_TXBF  0x09
+#define HTT_RX_PPDU_START_PREAMBLE_VHT           0x0C
+#define HTT_RX_PPDU_START_PREAMBLE_VHT_WITH_TXBF 0x0D
+
+#define RX_PPDU_START_INFO0_IS_GREENFIELD (1 << 0)
+
+#define RX_PPDU_START_INFO1_L_SIG_RATE_MASK    0x0000000f
+#define RX_PPDU_START_INFO1_L_SIG_RATE_LSB     0
+#define RX_PPDU_START_INFO1_L_SIG_LENGTH_MASK  0x0001ffe0
+#define RX_PPDU_START_INFO1_L_SIG_LENGTH_LSB   5
+#define RX_PPDU_START_INFO1_L_SIG_TAIL_MASK    0x00fc0000
+#define RX_PPDU_START_INFO1_L_SIG_TAIL_LSB     18
+#define RX_PPDU_START_INFO1_PREAMBLE_TYPE_MASK 0xff000000
+#define RX_PPDU_START_INFO1_PREAMBLE_TYPE_LSB  24
+#define RX_PPDU_START_INFO1_L_SIG_RATE_SELECT  (1 << 4)
+#define RX_PPDU_START_INFO1_L_SIG_PARITY       (1 << 17)
+
+#define RX_PPDU_START_INFO2_HT_SIG_VHT_SIG_A_1_MASK 0x00ffffff
+#define RX_PPDU_START_INFO2_HT_SIG_VHT_SIG_A_1_LSB  0
+
+#define RX_PPDU_START_INFO3_HT_SIG_VHT_SIG_A_2_MASK 0x00ffffff
+#define RX_PPDU_START_INFO3_HT_SIG_VHT_SIG_A_2_LSB  0
+#define RX_PPDU_START_INFO3_TXBF_H_INFO             (1 << 24)
+
+#define RX_PPDU_START_INFO4_VHT_SIG_B_MASK 0x1fffffff
+#define RX_PPDU_START_INFO4_VHT_SIG_B_LSB  0
+
+#define RX_PPDU_START_INFO5_SERVICE_MASK 0x0000ffff
+#define RX_PPDU_START_INFO5_SERVICE_LSB  0
+
+struct rx_ppdu_start {
+	struct {
+		u8 pri20_mhz;
+		u8 ext20_mhz;
+		u8 ext40_mhz;
+		u8 ext80_mhz;
+	} rssi_chains[4];
+	u8 rssi_comb;
+	__le16 rsvd0;
+	u8 info0; /* %RX_PPDU_START_INFO0_ */
+	__le32 info1; /* %RX_PPDU_START_INFO1_ */
+	__le32 info2; /* %RX_PPDU_START_INFO2_ */
+	__le32 info3; /* %RX_PPDU_START_INFO3_ */
+	__le32 info4; /* %RX_PPDU_START_INFO4_ */
+	__le32 info5; /* %RX_PPDU_START_INFO5_ */
+} __packed;
+
+/*
+ * rssi_chain0_pri20
+ *		RSSI of RX PPDU on chain 0 of primary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain0_sec20
+ *		RSSI of RX PPDU on chain 0 of secondary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain0_sec40
+ *		RSSI of RX PPDU on chain 0 of secondary 40 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain0_sec80
+ *		RSSI of RX PPDU on chain 0 of secondary 80 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain1_pri20
+ *		RSSI of RX PPDU on chain 1 of primary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain1_sec20
+ *		RSSI of RX PPDU on chain 1 of secondary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain1_sec40
+ *		RSSI of RX PPDU on chain 1 of secondary 40 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain1_sec80
+ *		RSSI of RX PPDU on chain 1 of secondary 80 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain2_pri20
+ *		RSSI of RX PPDU on chain 2 of primary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain2_sec20
+ *		RSSI of RX PPDU on chain 2 of secondary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain2_sec40
+ *		RSSI of RX PPDU on chain 2 of secondary 40 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain2_sec80
+ *		RSSI of RX PPDU on chain 2 of secondary 80 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain3_pri20
+ *		RSSI of RX PPDU on chain 3 of primary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain3_sec20
+ *		RSSI of RX PPDU on chain 3 of secondary 20 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain3_sec40
+ *		RSSI of RX PPDU on chain 3 of secondary 40 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_chain3_sec80
+ *		RSSI of RX PPDU on chain 3 of secondary 80 MHz bandwidth.
+ *		Value of 0x80 indicates invalid.
+ *
+ * rssi_comb
+ *		The combined RSSI of RX PPDU of all active chains and
+ *		bandwidths.  Value of 0x80 indicates invalid.
+ *
+ * reserved_4a
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * is_greenfield
+ *		Do we really support this?
+ *
+ * reserved_4b
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * l_sig_rate
+ *		If l_sig_rate_select is 0:
+ *		0x8: OFDM 48 Mbps
+ *		0x9: OFDM 24 Mbps
+ *		0xA: OFDM 12 Mbps
+ *		0xB: OFDM 6 Mbps
+ *		0xC: OFDM 54 Mbps
+ *		0xD: OFDM 36 Mbps
+ *		0xE: OFDM 18 Mbps
+ *		0xF: OFDM 9 Mbps
+ *		If l_sig_rate_select is 1:
+ *		0x8: CCK 11 Mbps long preamble
+ *		0x9: CCK 5.5 Mbps long preamble
+ *		0xA: CCK 2 Mbps long preamble
+ *		0xB: CCK 1 Mbps long preamble
+ *		0xC: CCK 11 Mbps short preamble
+ *		0xD: CCK 5.5 Mbps short preamble
+ *		0xE: CCK 2 Mbps short preamble
+ *
+ * l_sig_rate_select
+ *		Legacy signal rate select.  If set then l_sig_rate indicates
+ *		CCK rates.  If clear then l_sig_rate indicates OFDM rates.
+ *
+ * l_sig_length
+ *		Length of legacy frame in octets.
+ *
+ * l_sig_parity
+ *		Odd parity over l_sig_rate and l_sig_length
+ *
+ * l_sig_tail
+ *		Tail bits for Viterbi decoder
+ *
+ * preamble_type
+ *		Indicates the type of preamble ahead:
+ *		0x4: Legacy (OFDM/CCK)
+ *		0x8: HT
+ *		0x9: HT with TxBF
+ *		0xC: VHT
+ *		0xD: VHT with TxBF
+ *		0x80 - 0xFF: Reserved for special baseband data types such
+ *		as radar and spectral scan.
+ *
+ * ht_sig_vht_sig_a_1
+ *		If preamble_type == 0x8 or 0x9
+ *		HT-SIG (first 24 bits)
+ *		If preamble_type == 0xC or 0xD
+ *		VHT-SIG A (first 24 bits)
+ *		Else
+ *		Reserved
+ *
+ * reserved_6
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * ht_sig_vht_sig_a_2
+ *		If preamble_type == 0x8 or 0x9
+ *		HT-SIG (last 24 bits)
+ *		If preamble_type == 0xC or 0xD
+ *		VHT-SIG A (last 24 bits)
+ *		Else
+ *		Reserved
+ *
+ * txbf_h_info
+ *		Indicates that the packet data carries H information which
+ *		is used for TxBF debug.
+ *
+ * reserved_7
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * vht_sig_b
+ *		WiFi 1.0 and WiFi 2.0 will likely have this field to be all
+ *		0s since the BB does not plan on decoding VHT SIG-B.
+ *
+ * reserved_8
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * service
+ *		Service field from BB for OFDM, HT and VHT packets.  CCK
+ *		packets will have service field of 0.
+ *
+ * reserved_9
+ *		Reserved: HW should fill with 0, FW should ignore.
+*/
+
+
+#define RX_PPDU_END_FLAGS_PHY_ERR             (1 << 0)
+#define RX_PPDU_END_FLAGS_RX_LOCATION         (1 << 1)
+#define RX_PPDU_END_FLAGS_TXBF_H_INFO         (1 << 2)
+
+#define RX_PPDU_END_INFO0_RX_ANTENNA_MASK     0x00ffffff
+#define RX_PPDU_END_INFO0_RX_ANTENNA_LSB      0
+#define RX_PPDU_END_INFO0_FLAGS_TX_HT_VHT_ACK (1 << 24)
+#define RX_PPDU_END_INFO0_BB_CAPTURED_CHANNEL (1 << 25)
+
+#define RX_PPDU_END_INFO1_PPDU_DONE (1 << 15)
+
+struct rx_ppdu_end {
+	__le32 evm_p0;
+	__le32 evm_p1;
+	__le32 evm_p2;
+	__le32 evm_p3;
+	__le32 evm_p4;
+	__le32 evm_p5;
+	__le32 evm_p6;
+	__le32 evm_p7;
+	__le32 evm_p8;
+	__le32 evm_p9;
+	__le32 evm_p10;
+	__le32 evm_p11;
+	__le32 evm_p12;
+	__le32 evm_p13;
+	__le32 evm_p14;
+	__le32 evm_p15;
+	__le32 tsf_timestamp;
+	__le32 wb_timestamp;
+	u8 locationing_timestamp;
+	u8 phy_err_code;
+	__le16 flags; /* %RX_PPDU_END_FLAGS_ */
+	__le32 info0; /* %RX_PPDU_END_INFO0_ */
+	__le16 bb_length;
+	__le16 info1; /* %RX_PPDU_END_INFO1_ */
+} __packed;
+
+/*
+ * evm_p0
+ *		EVM for pilot 0.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p1
+ *		EVM for pilot 1.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p2
+ *		EVM for pilot 2.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p3
+ *		EVM for pilot 3.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p4
+ *		EVM for pilot 4.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p5
+ *		EVM for pilot 5.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p6
+ *		EVM for pilot 6.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p7
+ *		EVM for pilot 7.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p8
+ *		EVM for pilot 8.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p9
+ *		EVM for pilot 9.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p10
+ *		EVM for pilot 10.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p11
+ *		EVM for pilot 11.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p12
+ *		EVM for pilot 12.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p13
+ *		EVM for pilot 13.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p14
+ *		EVM for pilot 14.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * evm_p15
+ *		EVM for pilot 15.  Contain EVM for streams: 0, 1, 2 and 3.
+ *
+ * tsf_timestamp
+ *		Receive TSF timestamp sampled on the rising edge of
+ *		rx_clear.  For PHY errors this may be the current TSF when
+ *		phy_error is asserted if the rx_clear does not assert before
+ *		the end of the PHY error.
+ *
+ * wb_timestamp
+ *		WLAN/BT timestamp is a 1 usec resolution timestamp which
+ *		does not get updated based on receive beacon like TSF.  The
+ *		same rules for capturing tsf_timestamp are used to capture
+ *		the wb_timestamp.
+ *
+ * locationing_timestamp
+ *		Timestamp used for locationing.  This timestamp is used to
+ *		indicate fractions of usec.  For example if the MAC clock is
+ *		running at 80 MHz, the timestamp will increment every 12.5
+ *		nsec.  The value starts at 0 and increments to 79 and
+ *		returns to 0 and repeats.  This information is valid for
+ *		every PPDU.  This information can be used in conjunction
+ *		with wb_timestamp to capture large delta times.
+ *
+ * phy_err_code
+ *		See the 1.10.8.1.2 for the list of the PHY error codes.
+ *
+ * phy_err
+ *		Indicates a PHY error was detected for this PPDU.
+ *
+ * rx_location
+ *		Indicates that location information was requested.
+ *
+ * txbf_h_info
+ *		Indicates that the packet data carries H information which
+ *		is used for TxBF debug.
+ *
+ * reserved_18
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * rx_antenna
+ *		Receive antenna value
+ *
+ * tx_ht_vht_ack
+ *		Indicates that a HT or VHT Ack/BA frame was transmitted in
+ *		response to this receive packet.
+ *
+ * bb_captured_channel
+ *		Indicates that the BB has captured a channel dump.  FW can
+ *		then read the channel dump memory.  This may indicate that
+ *		the channel was captured either based on PCU setting the
+ *		capture_channel bit  BB descriptor or FW setting the
+ *		capture_channel mode bit.
+ *
+ * reserved_19
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * bb_length
+ *		Indicates the number of bytes of baseband information for
+ *		PPDUs where the BB descriptor preamble type is 0x80 to 0xFF
+ *		which indicates that this is not a normal PPDU but rather
+ *		contains baseband debug information.
+ *
+ * reserved_20
+ *		Reserved: HW should fill with 0, FW should ignore.
+ *
+ * ppdu_done
+ *		PPDU end status is only valid when ppdu_done bit is set.
+ *		Every time HW sets this bit in memory FW/SW must clear this
+ *		bit in memory.  FW will initialize all the ppdu_done dword
+ *		to 0.
+*/
+
+#define FW_RX_DESC_INFO0_DISCARD  (1 << 0)
+#define FW_RX_DESC_INFO0_FORWARD  (1 << 1)
+#define FW_RX_DESC_INFO0_INSPECT  (1 << 5)
+#define FW_RX_DESC_INFO0_EXT_MASK 0xC0
+#define FW_RX_DESC_INFO0_EXT_LSB  6
+
+struct fw_rx_desc_base {
+	u8 info0;
+} __packed;
+
+#endif /* _RX_DESC_H_ */
diff --git a/drivers/net/wireless/ath/ath10k/targaddrs.h b/drivers/net/wireless/ath/ath10k/targaddrs.h
new file mode 100644
index 0000000..be7ba1e
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/targaddrs.h
@@ -0,0 +1,449 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef __TARGADDRS_H__
+#define __TARGADDRS_H__
+
+/*
+ * xxx_HOST_INTEREST_ADDRESS is the address in Target RAM of the
+ * host_interest structure.  It must match the address of the _host_interest
+ * symbol (see linker script).
+ *
+ * Host Interest is shared between Host and Target in order to coordinate
+ * between the two, and is intended to remain constant (with additions only
+ * at the end) across software releases.
+ *
+ * All addresses are available here so that it's possible to
+ * write a single binary that works with all Target Types.
+ * May be used in assembler code as well as C.
+ */
+#define QCA988X_HOST_INTEREST_ADDRESS    0x00400800
+#define HOST_INTEREST_MAX_SIZE          0x200
+
+/*
+ * These are items that the Host may need to access via BMI or via the
+ * Diagnostic Window. The position of items in this structure must remain
+ * constant across firmware revisions! Types for each item must be fixed
+ * size across target and host platforms. More items may be added at the end.
+ */
+struct host_interest {
+	/*
+	 * Pointer to application-defined area, if any.
+	 * Set by Target application during startup.
+	 */
+	u32 hi_app_host_interest;			/* 0x00 */
+
+	/* Pointer to register dump area, valid after Target crash. */
+	u32 hi_failure_state;				/* 0x04 */
+
+	/* Pointer to debug logging header */
+	u32 hi_dbglog_hdr;				/* 0x08 */
+
+	u32 hi_unused0c;				/* 0x0c */
+
+	/*
+	 * General-purpose flag bits, similar to SOC_OPTION_* flags.
+	 * Can be used by application rather than by OS.
+	 */
+	u32 hi_option_flag;				/* 0x10 */
+
+	/*
+	 * Boolean that determines whether or not to
+	 * display messages on the serial port.
+	 */
+	u32 hi_serial_enable;				/* 0x14 */
+
+	/* Start address of DataSet index, if any */
+	u32 hi_dset_list_head;				/* 0x18 */
+
+	/* Override Target application start address */
+	u32 hi_app_start;				/* 0x1c */
+
+	/* Clock and voltage tuning */
+	u32 hi_skip_clock_init;				/* 0x20 */
+	u32 hi_core_clock_setting;			/* 0x24 */
+	u32 hi_cpu_clock_setting;			/* 0x28 */
+	u32 hi_system_sleep_setting;			/* 0x2c */
+	u32 hi_xtal_control_setting;			/* 0x30 */
+	u32 hi_pll_ctrl_setting_24ghz;			/* 0x34 */
+	u32 hi_pll_ctrl_setting_5ghz;			/* 0x38 */
+	u32 hi_ref_voltage_trim_setting;		/* 0x3c */
+	u32 hi_clock_info;				/* 0x40 */
+
+	/* Host uses BE CPU or not */
+	u32 hi_be;					/* 0x44 */
+
+	u32 hi_stack;	/* normal stack */			/* 0x48 */
+	u32 hi_err_stack; /* error stack */		/* 0x4c */
+	u32 hi_desired_cpu_speed_hz;			/* 0x50 */
+
+	/* Pointer to Board Data  */
+	u32 hi_board_data;				/* 0x54 */
+
+	/*
+	 * Indication of Board Data state:
+	 *    0: board data is not yet initialized.
+	 *    1: board data is initialized; unknown size
+	 *   >1: number of bytes of initialized board data
+	 */
+	u32 hi_board_data_initialized;			/* 0x58 */
+
+	u32 hi_dset_ram_index_table;			/* 0x5c */
+
+	u32 hi_desired_baud_rate;			/* 0x60 */
+	u32 hi_dbglog_config;				/* 0x64 */
+	u32 hi_end_ram_reserve_sz;			/* 0x68 */
+	u32 hi_mbox_io_block_sz;			/* 0x6c */
+
+	u32 hi_num_bpatch_streams;			/* 0x70 -- unused */
+	u32 hi_mbox_isr_yield_limit;			/* 0x74 */
+
+	u32 hi_refclk_hz;				/* 0x78 */
+	u32 hi_ext_clk_detected;			/* 0x7c */
+	u32 hi_dbg_uart_txpin;				/* 0x80 */
+	u32 hi_dbg_uart_rxpin;				/* 0x84 */
+	u32 hi_hci_uart_baud;				/* 0x88 */
+	u32 hi_hci_uart_pin_assignments;		/* 0x8C */
+
+	u32 hi_hci_uart_baud_scale_val;			/* 0x90 */
+	u32 hi_hci_uart_baud_step_val;			/* 0x94 */
+
+	u32 hi_allocram_start;				/* 0x98 */
+	u32 hi_allocram_sz;				/* 0x9c */
+	u32 hi_hci_bridge_flags;			/* 0xa0 */
+	u32 hi_hci_uart_support_pins;			/* 0xa4 */
+
+	u32 hi_hci_uart_pwr_mgmt_params;		/* 0xa8 */
+
+	/*
+	 * 0xa8 - [1]: 0 = UART FC active low, 1 = UART FC active high
+	 *        [31:16]: wakeup timeout in ms
+	 */
+	/* Pointer to extended board Data  */
+	u32 hi_board_ext_data;				/* 0xac */
+	u32 hi_board_ext_data_config;			/* 0xb0 */
+	/*
+	 * Bit [0]  :   valid
+	 * Bit[31:16:   size
+	 */
+	/*
+	 * hi_reset_flag is used to do some stuff when target reset.
+	 * such as restore app_start after warm reset or
+	 * preserve host Interest area, or preserve ROM data, literals etc.
+	 */
+	u32  hi_reset_flag;				/* 0xb4 */
+	/* indicate hi_reset_flag is valid */
+	u32  hi_reset_flag_valid;			/* 0xb8 */
+	u32 hi_hci_uart_pwr_mgmt_params_ext;		/* 0xbc */
+	/* 0xbc - [31:0]: idle timeout in ms */
+	/* ACS flags */
+	u32 hi_acs_flags;				/* 0xc0 */
+	u32 hi_console_flags;				/* 0xc4 */
+	u32 hi_nvram_state;				/* 0xc8 */
+	u32 hi_option_flag2;				/* 0xcc */
+
+	/* If non-zero, override values sent to Host in WMI_READY event. */
+	u32 hi_sw_version_override;			/* 0xd0 */
+	u32 hi_abi_version_override;			/* 0xd4 */
+
+	/*
+	 * Percentage of high priority RX traffic to total expected RX traffic
+	 * applicable only to ar6004
+	 */
+	u32 hi_hp_rx_traffic_ratio;			/* 0xd8 */
+
+	/* test applications flags */
+	u32 hi_test_apps_related;			/* 0xdc */
+	/* location of test script */
+	u32 hi_ota_testscript;				/* 0xe0 */
+	/* location of CAL data */
+	u32 hi_cal_data;				/* 0xe4 */
+
+	/* Number of packet log buffers */
+	u32 hi_pktlog_num_buffers;			/* 0xe8 */
+
+	/* wow extension configuration */
+	u32 hi_wow_ext_config;				/* 0xec */
+	u32 hi_pwr_save_flags;				/* 0xf0 */
+
+	/* Spatial Multiplexing Power Save (SMPS) options */
+	u32 hi_smps_options;				/* 0xf4 */
+
+	/* Interconnect-specific state */
+	u32 hi_interconnect_state;			/* 0xf8 */
+
+	/* Coex configuration flags */
+	u32 hi_coex_config;				/* 0xfc */
+
+	/* Early allocation support */
+	u32 hi_early_alloc;				/* 0x100 */
+	/* FW swap field */
+	/*
+	 * Bits of this 32bit word will be used to pass specific swap
+	 * instruction to FW
+	 */
+	/*
+	 * Bit 0 -- AP Nart descriptor no swap. When this bit is set
+	 * FW will not swap TX descriptor. Meaning packets are formed
+	 * on the target processor.
+	 */
+	/* Bit 1 - unused */
+	u32 hi_fw_swap;					/* 0x104 */
+} __packed;
+
+#define HI_ITEM(item)  offsetof(struct host_interest, item)
+
+/* Bits defined in hi_option_flag */
+
+/* Enable timer workaround */
+#define HI_OPTION_TIMER_WAR         0x01
+/* Limit BMI command credits */
+#define HI_OPTION_BMI_CRED_LIMIT    0x02
+/* Relay Dot11 hdr to/from host */
+#define HI_OPTION_RELAY_DOT11_HDR   0x04
+/* MAC addr method 0-locally administred 1-globally unique addrs */
+#define HI_OPTION_MAC_ADDR_METHOD   0x08
+/* Firmware Bridging */
+#define HI_OPTION_FW_BRIDGE         0x10
+/* Enable CPU profiling */
+#define HI_OPTION_ENABLE_PROFILE    0x20
+/* Disable debug logging */
+#define HI_OPTION_DISABLE_DBGLOG    0x40
+/* Skip Era Tracking */
+#define HI_OPTION_SKIP_ERA_TRACKING 0x80
+/* Disable PAPRD (debug) */
+#define HI_OPTION_PAPRD_DISABLE     0x100
+#define HI_OPTION_NUM_DEV_LSB       0x200
+#define HI_OPTION_NUM_DEV_MSB       0x800
+#define HI_OPTION_DEV_MODE_LSB      0x1000
+#define HI_OPTION_DEV_MODE_MSB      0x8000000
+/* Disable LowFreq Timer Stabilization */
+#define HI_OPTION_NO_LFT_STBL       0x10000000
+/* Skip regulatory scan */
+#define HI_OPTION_SKIP_REG_SCAN     0x20000000
+/*
+ * Do regulatory scan during init before
+ * sending WMI ready event to host
+ */
+#define HI_OPTION_INIT_REG_SCAN     0x40000000
+
+/* REV6: Do not adjust memory map */
+#define HI_OPTION_SKIP_MEMMAP       0x80000000
+
+#define HI_OPTION_MAC_ADDR_METHOD_SHIFT 3
+
+/* 2 bits of hi_option_flag are used to represent 3 modes */
+#define HI_OPTION_FW_MODE_IBSS    0x0 /* IBSS Mode */
+#define HI_OPTION_FW_MODE_BSS_STA 0x1 /* STA Mode */
+#define HI_OPTION_FW_MODE_AP      0x2 /* AP Mode */
+#define HI_OPTION_FW_MODE_BT30AMP 0x3 /* BT30 AMP Mode */
+
+/* 2 bits of hi_option flag are usedto represent 4 submodes */
+#define HI_OPTION_FW_SUBMODE_NONE    0x0  /* Normal mode */
+#define HI_OPTION_FW_SUBMODE_P2PDEV  0x1  /* p2p device mode */
+#define HI_OPTION_FW_SUBMODE_P2PCLIENT 0x2 /* p2p client mode */
+#define HI_OPTION_FW_SUBMODE_P2PGO   0x3 /* p2p go mode */
+
+/* Num dev Mask */
+#define HI_OPTION_NUM_DEV_MASK    0x7
+#define HI_OPTION_NUM_DEV_SHIFT   0x9
+
+/* firmware bridging */
+#define HI_OPTION_FW_BRIDGE_SHIFT 0x04
+
+/*
+Fw Mode/SubMode Mask
+|-----------------------------------------------------------------------------|
+|  SUB   |   SUB   |   SUB   |  SUB    |         |         |         |        |
+|MODE[3] | MODE[2] | MODE[1] | MODE[0] | MODE[3] | MODE[2] | MODE[1] | MODE[0]|
+|  (2)   |   (2)   |   (2)   |   (2)   |   (2)   |   (2)   |   (2)   |   (2)  |
+|-----------------------------------------------------------------------------|
+*/
+#define HI_OPTION_FW_MODE_BITS         0x2
+#define HI_OPTION_FW_MODE_MASK         0x3
+#define HI_OPTION_FW_MODE_SHIFT        0xC
+#define HI_OPTION_ALL_FW_MODE_MASK     0xFF
+
+#define HI_OPTION_FW_SUBMODE_BITS      0x2
+#define HI_OPTION_FW_SUBMODE_MASK      0x3
+#define HI_OPTION_FW_SUBMODE_SHIFT     0x14
+#define HI_OPTION_ALL_FW_SUBMODE_MASK  0xFF00
+#define HI_OPTION_ALL_FW_SUBMODE_SHIFT 0x8
+
+
+/* hi_option_flag2 options */
+#define HI_OPTION_OFFLOAD_AMSDU     0x01
+#define HI_OPTION_DFS_SUPPORT       0x02 /* Enable DFS support */
+#define HI_OPTION_ENABLE_RFKILL     0x04 /* RFKill Enable Feature*/
+#define HI_OPTION_RADIO_RETENTION_DISABLE 0x08 /* Disable radio retention */
+#define HI_OPTION_EARLY_CFG_DONE    0x10 /* Early configuration is complete */
+
+#define HI_OPTION_RF_KILL_SHIFT     0x2
+#define HI_OPTION_RF_KILL_MASK      0x1
+
+/* hi_reset_flag */
+/* preserve App Start address */
+#define HI_RESET_FLAG_PRESERVE_APP_START         0x01
+/* preserve host interest */
+#define HI_RESET_FLAG_PRESERVE_HOST_INTEREST     0x02
+/* preserve ROM data */
+#define HI_RESET_FLAG_PRESERVE_ROMDATA           0x04
+#define HI_RESET_FLAG_PRESERVE_NVRAM_STATE       0x08
+#define HI_RESET_FLAG_PRESERVE_BOOT_INFO         0x10
+#define HI_RESET_FLAG_WARM_RESET	0x20
+
+/* define hi_fw_swap bits */
+#define HI_DESC_IN_FW_BIT	0x01
+
+/* indicate the reset flag is valid */
+#define HI_RESET_FLAG_IS_VALID  0x12345678
+
+/* ACS is enabled */
+#define HI_ACS_FLAGS_ENABLED        (1 << 0)
+/* Use physical WWAN device */
+#define HI_ACS_FLAGS_USE_WWAN       (1 << 1)
+/* Use test VAP */
+#define HI_ACS_FLAGS_TEST_VAP       (1 << 2)
+
+/*
+ * CONSOLE FLAGS
+ *
+ * Bit Range  Meaning
+ * ---------  --------------------------------
+ *   2..0     UART ID (0 = Default)
+ *    3       Baud Select (0 = 9600, 1 = 115200)
+ *   30..4    Reserved
+ *    31      Enable Console
+ *
+ */
+
+#define HI_CONSOLE_FLAGS_ENABLE       (1 << 31)
+#define HI_CONSOLE_FLAGS_UART_MASK    (0x7)
+#define HI_CONSOLE_FLAGS_UART_SHIFT   0
+#define HI_CONSOLE_FLAGS_BAUD_SELECT  (1 << 3)
+
+/* SM power save options */
+#define HI_SMPS_ALLOW_MASK            (0x00000001)
+#define HI_SMPS_MODE_MASK             (0x00000002)
+#define HI_SMPS_MODE_STATIC           (0x00000000)
+#define HI_SMPS_MODE_DYNAMIC          (0x00000002)
+#define HI_SMPS_DISABLE_AUTO_MODE     (0x00000004)
+#define HI_SMPS_DATA_THRESH_MASK      (0x000007f8)
+#define HI_SMPS_DATA_THRESH_SHIFT     (3)
+#define HI_SMPS_RSSI_THRESH_MASK      (0x0007f800)
+#define HI_SMPS_RSSI_THRESH_SHIFT     (11)
+#define HI_SMPS_LOWPWR_CM_MASK        (0x00380000)
+#define HI_SMPS_LOWPWR_CM_SHIFT       (15)
+#define HI_SMPS_HIPWR_CM_MASK         (0x03c00000)
+#define HI_SMPS_HIPWR_CM_SHIFT        (19)
+
+/*
+ * WOW Extension configuration
+ *
+ * Bit Range  Meaning
+ * ---------  --------------------------------
+ *   8..0     Size of each WOW pattern (max 511)
+ *   15..9    Number of patterns per list (max 127)
+ *   17..16   Number of lists (max 4)
+ *   30..18   Reserved
+ *   31       Enabled
+ *
+ *  set values (except enable) to zeros for default settings
+ */
+
+#define HI_WOW_EXT_ENABLED_MASK        (1 << 31)
+#define HI_WOW_EXT_NUM_LIST_SHIFT      16
+#define HI_WOW_EXT_NUM_LIST_MASK       (0x3 << HI_WOW_EXT_NUM_LIST_SHIFT)
+#define HI_WOW_EXT_NUM_PATTERNS_SHIFT  9
+#define HI_WOW_EXT_NUM_PATTERNS_MASK   (0x7F << HI_WOW_EXT_NUM_PATTERNS_SHIFT)
+#define HI_WOW_EXT_PATTERN_SIZE_SHIFT  0
+#define HI_WOW_EXT_PATTERN_SIZE_MASK   (0x1FF << HI_WOW_EXT_PATTERN_SIZE_SHIFT)
+
+#define HI_WOW_EXT_MAKE_CONFIG(num_lists, count, size) \
+	((((num_lists) << HI_WOW_EXT_NUM_LIST_SHIFT) & \
+		HI_WOW_EXT_NUM_LIST_MASK) | \
+	(((count) << HI_WOW_EXT_NUM_PATTERNS_SHIFT) & \
+		HI_WOW_EXT_NUM_PATTERNS_MASK) | \
+	(((size) << HI_WOW_EXT_PATTERN_SIZE_SHIFT) & \
+		HI_WOW_EXT_PATTERN_SIZE_MASK))
+
+#define HI_WOW_EXT_GET_NUM_LISTS(config) \
+	(((config) & HI_WOW_EXT_NUM_LIST_MASK) >> HI_WOW_EXT_NUM_LIST_SHIFT)
+#define HI_WOW_EXT_GET_NUM_PATTERNS(config) \
+	(((config) & HI_WOW_EXT_NUM_PATTERNS_MASK) >> \
+		HI_WOW_EXT_NUM_PATTERNS_SHIFT)
+#define HI_WOW_EXT_GET_PATTERN_SIZE(config) \
+	(((config) & HI_WOW_EXT_PATTERN_SIZE_MASK) >> \
+		HI_WOW_EXT_PATTERN_SIZE_SHIFT)
+
+/*
+ * Early allocation configuration
+ * Support RAM bank configuration before BMI done and this eases the memory
+ * allocation at very early stage
+ * Bit Range  Meaning
+ * ---------  ----------------------------------
+ * [0:3]      number of bank assigned to be IRAM
+ * [4:15]     reserved
+ * [16:31]    magic number
+ *
+ * Note:
+ * 1. target firmware would check magic number and if it's a match, firmware
+ *    would consider the bits[0:15] are valid and base on that to calculate
+ *    the end of DRAM. Early allocation would be located at that area and
+ *    may be reclaimed when necesary
+ * 2. if no magic number is found, early allocation would happen at "_end"
+ *    symbol of ROM which is located before the app-data and might NOT be
+ *    re-claimable. If this is adopted, link script should keep this in
+ *    mind to avoid data corruption.
+ */
+#define HI_EARLY_ALLOC_MAGIC		0x6d8a
+#define HI_EARLY_ALLOC_MAGIC_MASK	0xffff0000
+#define HI_EARLY_ALLOC_MAGIC_SHIFT	16
+#define HI_EARLY_ALLOC_IRAM_BANKS_MASK	0x0000000f
+#define HI_EARLY_ALLOC_IRAM_BANKS_SHIFT	0
+
+#define HI_EARLY_ALLOC_VALID() \
+	((((HOST_INTEREST->hi_early_alloc) & HI_EARLY_ALLOC_MAGIC_MASK) >> \
+	HI_EARLY_ALLOC_MAGIC_SHIFT) == (HI_EARLY_ALLOC_MAGIC))
+#define HI_EARLY_ALLOC_GET_IRAM_BANKS() \
+	(((HOST_INTEREST->hi_early_alloc) & HI_EARLY_ALLOC_IRAM_BANKS_MASK) \
+	>> HI_EARLY_ALLOC_IRAM_BANKS_SHIFT)
+
+/*power save flag bit definitions*/
+#define HI_PWR_SAVE_LPL_ENABLED   0x1
+/*b1-b3 reserved*/
+/*b4-b5 : dev0 LPL type : 0 - none
+			  1- Reduce Pwr Search
+			  2- Reduce Pwr Listen*/
+/*b6-b7 : dev1 LPL type and so on for Max 8 devices*/
+#define HI_PWR_SAVE_LPL_DEV0_LSB   4
+#define HI_PWR_SAVE_LPL_DEV_MASK   0x3
+/*power save related utility macros*/
+#define HI_LPL_ENABLED() \
+	((HOST_INTEREST->hi_pwr_save_flags & HI_PWR_SAVE_LPL_ENABLED))
+#define HI_DEV_LPL_TYPE_GET(_devix) \
+	(HOST_INTEREST->hi_pwr_save_flags & ((HI_PWR_SAVE_LPL_DEV_MASK) << \
+	 (HI_PWR_SAVE_LPL_DEV0_LSB + (_devix)*2)))
+
+#define HOST_INTEREST_SMPS_IS_ALLOWED() \
+	((HOST_INTEREST->hi_smps_options & HI_SMPS_ALLOW_MASK))
+
+/* Reserve 1024 bytes for extended board data */
+#define QCA988X_BOARD_DATA_SZ     7168
+#define QCA988X_BOARD_EXT_DATA_SZ 0
+
+#endif /* __TARGADDRS_H__ */
diff --git a/drivers/net/wireless/ath/ath10k/trace.c b/drivers/net/wireless/ath/ath10k/trace.c
new file mode 100644
index 0000000..4a31e2c
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/trace.c
@@ -0,0 +1,20 @@
+/*
+ * Copyright (c) 2012 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/module.h>
+
+#define CREATE_TRACE_POINTS
+#include "trace.h"
diff --git a/drivers/net/wireless/ath/ath10k/trace.h b/drivers/net/wireless/ath/ath10k/trace.h
new file mode 100644
index 0000000..85e806b
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/trace.h
@@ -0,0 +1,170 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#if !defined(_TRACE_H_) || defined(TRACE_HEADER_MULTI_READ)
+
+#include <linux/tracepoint.h>
+
+#define _TRACE_H_
+
+/* create empty functions when tracing is disabled */
+#if !defined(CONFIG_ATH10K_TRACING)
+#undef TRACE_EVENT
+#define TRACE_EVENT(name, proto, ...) \
+static inline void trace_ ## name(proto) {}
+#undef DECLARE_EVENT_CLASS
+#define DECLARE_EVENT_CLASS(...)
+#undef DEFINE_EVENT
+#define DEFINE_EVENT(evt_class, name, proto, ...) \
+static inline void trace_ ## name(proto) {}
+#endif /* !CONFIG_ATH10K_TRACING || __CHECKER__ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM ath10k
+
+#define ATH10K_MSG_MAX 200
+
+DECLARE_EVENT_CLASS(ath10k_log_event,
+	TP_PROTO(struct va_format *vaf),
+	TP_ARGS(vaf),
+	TP_STRUCT__entry(
+		__dynamic_array(char, msg, ATH10K_MSG_MAX)
+	),
+	TP_fast_assign(
+		WARN_ON_ONCE(vsnprintf(__get_dynamic_array(msg),
+				       ATH10K_MSG_MAX,
+				       vaf->fmt,
+				       *vaf->va) >= ATH10K_MSG_MAX);
+	),
+	TP_printk("%s", __get_str(msg))
+);
+
+DEFINE_EVENT(ath10k_log_event, ath10k_log_err,
+	     TP_PROTO(struct va_format *vaf),
+	     TP_ARGS(vaf)
+);
+
+DEFINE_EVENT(ath10k_log_event, ath10k_log_warn,
+	     TP_PROTO(struct va_format *vaf),
+	     TP_ARGS(vaf)
+);
+
+DEFINE_EVENT(ath10k_log_event, ath10k_log_info,
+	     TP_PROTO(struct va_format *vaf),
+	     TP_ARGS(vaf)
+);
+
+TRACE_EVENT(ath10k_log_dbg,
+	TP_PROTO(unsigned int level, struct va_format *vaf),
+	TP_ARGS(level, vaf),
+	TP_STRUCT__entry(
+		__field(unsigned int, level)
+		__dynamic_array(char, msg, ATH10K_MSG_MAX)
+	),
+	TP_fast_assign(
+		__entry->level = level;
+		WARN_ON_ONCE(vsnprintf(__get_dynamic_array(msg),
+				       ATH10K_MSG_MAX,
+				       vaf->fmt,
+				       *vaf->va) >= ATH10K_MSG_MAX);
+	),
+	TP_printk("%s", __get_str(msg))
+);
+
+TRACE_EVENT(ath10k_log_dbg_dump,
+	TP_PROTO(const char *msg, const char *prefix,
+		 const void *buf, size_t buf_len),
+
+	TP_ARGS(msg, prefix, buf, buf_len),
+
+	TP_STRUCT__entry(
+		__string(msg, msg)
+		__string(prefix, prefix)
+		__field(size_t, buf_len)
+		__dynamic_array(u8, buf, buf_len)
+	),
+
+	TP_fast_assign(
+		__assign_str(msg, msg);
+		__assign_str(prefix, prefix);
+		__entry->buf_len = buf_len;
+		memcpy(__get_dynamic_array(buf), buf, buf_len);
+	),
+
+	TP_printk(
+		"%s/%s\n", __get_str(prefix), __get_str(msg)
+	)
+);
+
+TRACE_EVENT(ath10k_wmi_cmd,
+	TP_PROTO(int id, void *buf, size_t buf_len),
+
+	TP_ARGS(id, buf, buf_len),
+
+	TP_STRUCT__entry(
+		__field(unsigned int, id)
+		__field(size_t, buf_len)
+		__dynamic_array(u8, buf, buf_len)
+	),
+
+	TP_fast_assign(
+		__entry->id = id;
+		__entry->buf_len = buf_len;
+		memcpy(__get_dynamic_array(buf), buf, buf_len);
+	),
+
+	TP_printk(
+		"id %d len %zu",
+		__entry->id,
+		__entry->buf_len
+	)
+);
+
+TRACE_EVENT(ath10k_wmi_event,
+	TP_PROTO(int id, void *buf, size_t buf_len),
+
+	TP_ARGS(id, buf, buf_len),
+
+	TP_STRUCT__entry(
+		__field(unsigned int, id)
+		__field(size_t, buf_len)
+		__dynamic_array(u8, buf, buf_len)
+	),
+
+	TP_fast_assign(
+		__entry->id = id;
+		__entry->buf_len = buf_len;
+		memcpy(__get_dynamic_array(buf), buf, buf_len);
+	),
+
+	TP_printk(
+		"id %d len %zu",
+		__entry->id,
+		__entry->buf_len
+	)
+);
+
+#endif /* _TRACE_H_ || TRACE_HEADER_MULTI_READ*/
+
+/* we don't want to use include/trace/events */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>
diff --git a/drivers/net/wireless/ath/ath10k/txrx.c b/drivers/net/wireless/ath/ath10k/txrx.c
new file mode 100644
index 0000000..68b6fae
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/txrx.c
@@ -0,0 +1,417 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "core.h"
+#include "txrx.h"
+#include "htt.h"
+#include "mac.h"
+#include "debug.h"
+
+static void ath10k_report_offchan_tx(struct ath10k *ar, struct sk_buff *skb)
+{
+	if (!ATH10K_SKB_CB(skb)->htt.is_offchan)
+		return;
+
+	/* If the original wait_for_completion() timed out before
+	 * {data,mgmt}_tx_completed() was called then we could complete
+	 * offchan_tx_completed for a different skb. Prevent this by using
+	 * offchan_tx_skb. */
+	spin_lock_bh(&ar->data_lock);
+	if (ar->offchan_tx_skb != skb) {
+		ath10k_warn("completed old offchannel frame\n");
+		goto out;
+	}
+
+	complete(&ar->offchan_tx_completed);
+	ar->offchan_tx_skb = NULL; /* just for sanity */
+
+	ath10k_dbg(ATH10K_DBG_HTT, "completed offchannel skb %p\n", skb);
+out:
+	spin_unlock_bh(&ar->data_lock);
+}
+
+void ath10k_txrx_tx_unref(struct ath10k_htt *htt, struct sk_buff *txdesc)
+{
+	struct device *dev = htt->ar->dev;
+	struct ieee80211_tx_info *info;
+	struct sk_buff *txfrag = ATH10K_SKB_CB(txdesc)->htt.txfrag;
+	struct sk_buff *msdu = ATH10K_SKB_CB(txdesc)->htt.msdu;
+	int ret;
+
+	if (ATH10K_SKB_CB(txdesc)->htt.refcount == 0)
+		return;
+
+	ATH10K_SKB_CB(txdesc)->htt.refcount--;
+
+	if (ATH10K_SKB_CB(txdesc)->htt.refcount > 0)
+		return;
+
+	if (txfrag) {
+		ret = ath10k_skb_unmap(dev, txfrag);
+		if (ret)
+			ath10k_warn("txfrag unmap failed (%d)\n", ret);
+
+		dev_kfree_skb_any(txfrag);
+	}
+
+	ret = ath10k_skb_unmap(dev, msdu);
+	if (ret)
+		ath10k_warn("data skb unmap failed (%d)\n", ret);
+
+	ath10k_report_offchan_tx(htt->ar, msdu);
+
+	info = IEEE80211_SKB_CB(msdu);
+	memset(&info->status, 0, sizeof(info->status));
+
+	if (ATH10K_SKB_CB(txdesc)->htt.discard) {
+		ieee80211_free_txskb(htt->ar->hw, msdu);
+		goto exit;
+	}
+
+	if (!(info->flags & IEEE80211_TX_CTL_NO_ACK))
+		info->flags |= IEEE80211_TX_STAT_ACK;
+
+	if (ATH10K_SKB_CB(txdesc)->htt.no_ack)
+		info->flags &= ~IEEE80211_TX_STAT_ACK;
+
+	ieee80211_tx_status(htt->ar->hw, msdu);
+	/* we do not own the msdu anymore */
+
+exit:
+	spin_lock_bh(&htt->tx_lock);
+	htt->pending_tx[ATH10K_SKB_CB(txdesc)->htt.msdu_id] = NULL;
+	ath10k_htt_tx_free_msdu_id(htt, ATH10K_SKB_CB(txdesc)->htt.msdu_id);
+	__ath10k_htt_tx_dec_pending(htt);
+	if (bitmap_empty(htt->used_msdu_ids, htt->max_num_pending_tx))
+		wake_up(&htt->empty_tx_wq);
+	spin_unlock_bh(&htt->tx_lock);
+
+	dev_kfree_skb_any(txdesc);
+}
+
+void ath10k_txrx_tx_completed(struct ath10k_htt *htt,
+			      const struct htt_tx_done *tx_done)
+{
+	struct sk_buff *txdesc;
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt tx completion msdu_id %u discard %d no_ack %d\n",
+		   tx_done->msdu_id, !!tx_done->discard, !!tx_done->no_ack);
+
+	if (tx_done->msdu_id >= htt->max_num_pending_tx) {
+		ath10k_warn("warning: msdu_id %d too big, ignoring\n",
+			    tx_done->msdu_id);
+		return;
+	}
+
+	txdesc = htt->pending_tx[tx_done->msdu_id];
+
+	ATH10K_SKB_CB(txdesc)->htt.discard = tx_done->discard;
+	ATH10K_SKB_CB(txdesc)->htt.no_ack = tx_done->no_ack;
+
+	ath10k_txrx_tx_unref(htt, txdesc);
+}
+
+static const u8 rx_legacy_rate_idx[] = {
+	3,	/* 0x00  - 11Mbps  */
+	2,	/* 0x01  - 5.5Mbps */
+	1,	/* 0x02  - 2Mbps   */
+	0,	/* 0x03  - 1Mbps   */
+	3,	/* 0x04  - 11Mbps  */
+	2,	/* 0x05  - 5.5Mbps */
+	1,	/* 0x06  - 2Mbps   */
+	0,	/* 0x07  - 1Mbps   */
+	10,	/* 0x08  - 48Mbps  */
+	8,	/* 0x09  - 24Mbps  */
+	6,	/* 0x0A  - 12Mbps  */
+	4,	/* 0x0B  - 6Mbps   */
+	11,	/* 0x0C  - 54Mbps  */
+	9,	/* 0x0D  - 36Mbps  */
+	7,	/* 0x0E  - 18Mbps  */
+	5,	/* 0x0F  - 9Mbps   */
+};
+
+static void process_rx_rates(struct ath10k *ar, struct htt_rx_info *info,
+			     enum ieee80211_band band,
+			     struct ieee80211_rx_status *status)
+{
+	u8 cck, rate, rate_idx, bw, sgi, mcs, nss;
+	u8 info0 = info->rate.info0;
+	u32 info1 = info->rate.info1;
+	u32 info2 = info->rate.info2;
+	u8 preamble = 0;
+
+	/* Check if valid fields */
+	if (!(info0 & HTT_RX_INDICATION_INFO0_START_VALID))
+		return;
+
+	preamble = MS(info1, HTT_RX_INDICATION_INFO1_PREAMBLE_TYPE);
+
+	switch (preamble) {
+	case HTT_RX_LEGACY:
+		cck = info0 & HTT_RX_INDICATION_INFO0_LEGACY_RATE_CCK;
+		rate = MS(info0, HTT_RX_INDICATION_INFO0_LEGACY_RATE);
+		rate_idx = 0;
+
+		if (rate < 0x08 || rate > 0x0F)
+			break;
+
+		switch (band) {
+		case IEEE80211_BAND_2GHZ:
+			if (cck)
+				rate &= ~BIT(3);
+			rate_idx = rx_legacy_rate_idx[rate];
+			break;
+		case IEEE80211_BAND_5GHZ:
+			rate_idx = rx_legacy_rate_idx[rate];
+			/* We are using same rate table registering
+			   HW - ath10k_rates[]. In case of 5GHz skip
+			   CCK rates, so -4 here */
+			rate_idx -= 4;
+			break;
+		default:
+			break;
+		}
+
+		status->rate_idx = rate_idx;
+		break;
+	case HTT_RX_HT:
+	case HTT_RX_HT_WITH_TXBF:
+		/* HT-SIG - Table 20-11 in info1 and info2 */
+		mcs = info1 & 0x1F;
+		nss = mcs >> 3;
+		bw = (info1 >> 7) & 1;
+		sgi = (info2 >> 7) & 1;
+
+		status->rate_idx = mcs;
+		status->flag |= RX_FLAG_HT;
+		if (sgi)
+			status->flag |= RX_FLAG_SHORT_GI;
+		if (bw)
+			status->flag |= RX_FLAG_40MHZ;
+		break;
+	case HTT_RX_VHT:
+	case HTT_RX_VHT_WITH_TXBF:
+		/* VHT-SIG-A1 in info 1, VHT-SIG-A2 in info2
+		   TODO check this */
+		mcs = (info2 >> 4) & 0x0F;
+		nss = (info1 >> 10) & 0x07;
+		bw = info1 & 3;
+		sgi = info2 & 1;
+
+		status->rate_idx = mcs;
+		status->vht_nss = nss;
+
+		if (sgi)
+			status->flag |= RX_FLAG_SHORT_GI;
+
+		switch (bw) {
+		/* 20MHZ */
+		case 0:
+			break;
+		/* 40MHZ */
+		case 1:
+			status->flag |= RX_FLAG_40MHZ;
+			break;
+		/* 80MHZ */
+		case 2:
+			status->flag |= RX_FLAG_80MHZ;
+		}
+
+		status->flag |= RX_FLAG_VHT;
+		break;
+	default:
+		break;
+	}
+}
+
+void ath10k_process_rx(struct ath10k *ar, struct htt_rx_info *info)
+{
+	struct ieee80211_rx_status *status;
+	struct ieee80211_channel *ch;
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)info->skb->data;
+
+	status = IEEE80211_SKB_RXCB(info->skb);
+	memset(status, 0, sizeof(*status));
+
+	if (info->encrypt_type != HTT_RX_MPDU_ENCRYPT_NONE) {
+		status->flag |= RX_FLAG_DECRYPTED | RX_FLAG_IV_STRIPPED |
+				RX_FLAG_MMIC_STRIPPED;
+		hdr->frame_control = __cpu_to_le16(
+				__le16_to_cpu(hdr->frame_control) &
+				~IEEE80211_FCTL_PROTECTED);
+	}
+
+	if (info->status == HTT_RX_IND_MPDU_STATUS_TKIP_MIC_ERR)
+		status->flag |= RX_FLAG_MMIC_ERROR;
+
+	if (info->fcs_err)
+		status->flag |= RX_FLAG_FAILED_FCS_CRC;
+
+	status->signal = info->signal;
+
+	spin_lock_bh(&ar->data_lock);
+	ch = ar->scan_channel;
+	if (!ch)
+		ch = ar->rx_channel;
+	spin_unlock_bh(&ar->data_lock);
+
+	if (!ch) {
+		ath10k_warn("no channel configured; ignoring frame!\n");
+		dev_kfree_skb_any(info->skb);
+		return;
+	}
+
+	process_rx_rates(ar, info, ch->band, status);
+	status->band = ch->band;
+	status->freq = ch->center_freq;
+
+	ath10k_dbg(ATH10K_DBG_DATA,
+		   "rx skb %p len %u %s%s%s%s%s %srate_idx %u vht_nss %u freq %u band %u\n",
+		   info->skb,
+		   info->skb->len,
+		   status->flag == 0 ? "legacy" : "",
+		   status->flag & RX_FLAG_HT ? "ht" : "",
+		   status->flag & RX_FLAG_VHT ? "vht" : "",
+		   status->flag & RX_FLAG_40MHZ ? "40" : "",
+		   status->flag & RX_FLAG_80MHZ ? "80" : "",
+		   status->flag & RX_FLAG_SHORT_GI ? "sgi " : "",
+		   status->rate_idx,
+		   status->vht_nss,
+		   status->freq,
+		   status->band);
+
+	ieee80211_rx(ar->hw, info->skb);
+}
+
+struct ath10k_peer *ath10k_peer_find(struct ath10k *ar, int vdev_id,
+				     const u8 *addr)
+{
+	struct ath10k_peer *peer;
+
+	lockdep_assert_held(&ar->data_lock);
+
+	list_for_each_entry(peer, &ar->peers, list) {
+		if (peer->vdev_id != vdev_id)
+			continue;
+		if (memcmp(peer->addr, addr, ETH_ALEN))
+			continue;
+
+		return peer;
+	}
+
+	return NULL;
+}
+
+static struct ath10k_peer *ath10k_peer_find_by_id(struct ath10k *ar,
+						  int peer_id)
+{
+	struct ath10k_peer *peer;
+
+	lockdep_assert_held(&ar->data_lock);
+
+	list_for_each_entry(peer, &ar->peers, list)
+		if (test_bit(peer_id, peer->peer_ids))
+			return peer;
+
+	return NULL;
+}
+
+static int ath10k_wait_for_peer_common(struct ath10k *ar, int vdev_id,
+				       const u8 *addr, bool expect_mapped)
+{
+	int ret;
+
+	ret = wait_event_timeout(ar->peer_mapping_wq, ({
+			bool mapped;
+
+			spin_lock_bh(&ar->data_lock);
+			mapped = !!ath10k_peer_find(ar, vdev_id, addr);
+			spin_unlock_bh(&ar->data_lock);
+
+			mapped == expect_mapped;
+		}), 3*HZ);
+
+	if (ret <= 0)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+int ath10k_wait_for_peer_created(struct ath10k *ar, int vdev_id, const u8 *addr)
+{
+	return ath10k_wait_for_peer_common(ar, vdev_id, addr, true);
+}
+
+int ath10k_wait_for_peer_deleted(struct ath10k *ar, int vdev_id, const u8 *addr)
+{
+	return ath10k_wait_for_peer_common(ar, vdev_id, addr, false);
+}
+
+void ath10k_peer_map_event(struct ath10k_htt *htt,
+			   struct htt_peer_map_event *ev)
+{
+	struct ath10k *ar = htt->ar;
+	struct ath10k_peer *peer;
+
+	spin_lock_bh(&ar->data_lock);
+	peer = ath10k_peer_find(ar, ev->vdev_id, ev->addr);
+	if (!peer) {
+		peer = kzalloc(sizeof(*peer), GFP_ATOMIC);
+		if (!peer)
+			goto exit;
+
+		peer->vdev_id = ev->vdev_id;
+		memcpy(peer->addr, ev->addr, ETH_ALEN);
+		list_add(&peer->list, &ar->peers);
+		wake_up(&ar->peer_mapping_wq);
+	}
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt peer map vdev %d peer %pM id %d\n",
+		   ev->vdev_id, ev->addr, ev->peer_id);
+
+	set_bit(ev->peer_id, peer->peer_ids);
+exit:
+	spin_unlock_bh(&ar->data_lock);
+}
+
+void ath10k_peer_unmap_event(struct ath10k_htt *htt,
+			     struct htt_peer_unmap_event *ev)
+{
+	struct ath10k *ar = htt->ar;
+	struct ath10k_peer *peer;
+
+	spin_lock_bh(&ar->data_lock);
+	peer = ath10k_peer_find_by_id(ar, ev->peer_id);
+	if (!peer) {
+		ath10k_warn("unknown peer id %d\n", ev->peer_id);
+		goto exit;
+	}
+
+	ath10k_dbg(ATH10K_DBG_HTT, "htt peer unmap vdev %d peer %pM id %d\n",
+		   peer->vdev_id, peer->addr, ev->peer_id);
+
+	clear_bit(ev->peer_id, peer->peer_ids);
+
+	if (bitmap_empty(peer->peer_ids, ATH10K_MAX_NUM_PEER_IDS)) {
+		list_del(&peer->list);
+		kfree(peer);
+		wake_up(&ar->peer_mapping_wq);
+	}
+
+exit:
+	spin_unlock_bh(&ar->data_lock);
+}
diff --git a/drivers/net/wireless/ath/ath10k/txrx.h b/drivers/net/wireless/ath/ath10k/txrx.h
new file mode 100644
index 0000000..e78632a
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/txrx.h
@@ -0,0 +1,39 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+#ifndef _TXRX_H_
+#define _TXRX_H_
+
+#include "htt.h"
+
+void ath10k_txrx_tx_unref(struct ath10k_htt *htt, struct sk_buff *txdesc);
+void ath10k_txrx_tx_completed(struct ath10k_htt *htt,
+			      const struct htt_tx_done *tx_done);
+void ath10k_process_rx(struct ath10k *ar, struct htt_rx_info *info);
+
+struct ath10k_peer *ath10k_peer_find(struct ath10k *ar, int vdev_id,
+				     const u8 *addr);
+int ath10k_wait_for_peer_created(struct ath10k *ar, int vdev_id,
+				 const u8 *addr);
+int ath10k_wait_for_peer_deleted(struct ath10k *ar, int vdev_id,
+				 const u8 *addr);
+
+void ath10k_peer_map_event(struct ath10k_htt *htt,
+			   struct htt_peer_map_event *ev);
+void ath10k_peer_unmap_event(struct ath10k_htt *htt,
+			     struct htt_peer_unmap_event *ev);
+
+#endif
diff --git a/drivers/net/wireless/ath/ath10k/wmi.c b/drivers/net/wireless/ath/ath10k/wmi.c
new file mode 100644
index 0000000..7d4b798
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/wmi.c
@@ -0,0 +1,2081 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/skbuff.h>
+
+#include "core.h"
+#include "htc.h"
+#include "debug.h"
+#include "wmi.h"
+#include "mac.h"
+
+void ath10k_wmi_flush_tx(struct ath10k *ar)
+{
+	int ret;
+
+	ret = wait_event_timeout(ar->wmi.wq,
+				 atomic_read(&ar->wmi.pending_tx_count) == 0,
+				 5*HZ);
+	if (atomic_read(&ar->wmi.pending_tx_count) == 0)
+		return;
+
+	if (ret == 0)
+		ret = -ETIMEDOUT;
+
+	if (ret < 0)
+		ath10k_warn("wmi flush failed (%d)\n", ret);
+}
+
+int ath10k_wmi_wait_for_service_ready(struct ath10k *ar)
+{
+	int ret;
+	ret = wait_for_completion_timeout(&ar->wmi.service_ready,
+					  WMI_SERVICE_READY_TIMEOUT_HZ);
+	return ret;
+}
+
+int ath10k_wmi_wait_for_unified_ready(struct ath10k *ar)
+{
+	int ret;
+	ret = wait_for_completion_timeout(&ar->wmi.unified_ready,
+					  WMI_UNIFIED_READY_TIMEOUT_HZ);
+	return ret;
+}
+
+static struct sk_buff *ath10k_wmi_alloc_skb(u32 len)
+{
+	struct sk_buff *skb;
+	u32 round_len = roundup(len, 4);
+
+	skb = ath10k_htc_alloc_skb(WMI_SKB_HEADROOM + round_len);
+	if (!skb)
+		return NULL;
+
+	skb_reserve(skb, WMI_SKB_HEADROOM);
+	if (!IS_ALIGNED((unsigned long)skb->data, 4))
+		ath10k_warn("Unaligned WMI skb\n");
+
+	skb_put(skb, round_len);
+	memset(skb->data, 0, round_len);
+
+	return skb;
+}
+
+static void ath10k_wmi_htc_tx_complete(struct ath10k *ar, struct sk_buff *skb)
+{
+	dev_kfree_skb(skb);
+
+	if (atomic_sub_return(1, &ar->wmi.pending_tx_count) == 0)
+		wake_up(&ar->wmi.wq);
+}
+
+/* WMI command API */
+static int ath10k_wmi_cmd_send(struct ath10k *ar, struct sk_buff *skb,
+			       enum wmi_cmd_id cmd_id)
+{
+	struct ath10k_skb_cb *skb_cb = ATH10K_SKB_CB(skb);
+	struct wmi_cmd_hdr *cmd_hdr;
+	int status;
+	u32 cmd = 0;
+
+	if (skb_push(skb, sizeof(struct wmi_cmd_hdr)) == NULL)
+		return -ENOMEM;
+
+	cmd |= SM(cmd_id, WMI_CMD_HDR_CMD_ID);
+
+	cmd_hdr = (struct wmi_cmd_hdr *)skb->data;
+	cmd_hdr->cmd_id = __cpu_to_le32(cmd);
+
+	if (atomic_add_return(1, &ar->wmi.pending_tx_count) >
+	    WMI_MAX_PENDING_TX_COUNT) {
+		/* avoid using up memory when FW hangs */
+		atomic_dec(&ar->wmi.pending_tx_count);
+		return -EBUSY;
+	}
+
+	memset(skb_cb, 0, sizeof(*skb_cb));
+
+	trace_ath10k_wmi_cmd(cmd_id, skb->data, skb->len);
+
+	status = ath10k_htc_send(ar->htc, ar->wmi.eid, skb);
+	if (status) {
+		dev_kfree_skb_any(skb);
+		atomic_dec(&ar->wmi.pending_tx_count);
+		return status;
+	}
+
+	return 0;
+}
+
+static int ath10k_wmi_event_scan(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct wmi_scan_event *event = (struct wmi_scan_event *)skb->data;
+	enum wmi_scan_event_type event_type;
+	enum wmi_scan_completion_reason reason;
+	u32 freq;
+	u32 req_id;
+	u32 scan_id;
+	u32 vdev_id;
+
+	event_type = __le32_to_cpu(event->event_type);
+	reason     = __le32_to_cpu(event->reason);
+	freq       = __le32_to_cpu(event->channel_freq);
+	req_id     = __le32_to_cpu(event->scan_req_id);
+	scan_id    = __le32_to_cpu(event->scan_id);
+	vdev_id    = __le32_to_cpu(event->vdev_id);
+
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_SCAN_EVENTID\n");
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "scan event type %d reason %d freq %d req_id %d "
+		   "scan_id %d vdev_id %d\n",
+		   event_type, reason, freq, req_id, scan_id, vdev_id);
+
+	spin_lock_bh(&ar->data_lock);
+
+	switch (event_type) {
+	case WMI_SCAN_EVENT_STARTED:
+		ath10k_dbg(ATH10K_DBG_WMI, "SCAN_EVENT_STARTED\n");
+		if (ar->scan.in_progress && ar->scan.is_roc)
+			ieee80211_ready_on_channel(ar->hw);
+
+		complete(&ar->scan.started);
+		break;
+	case WMI_SCAN_EVENT_COMPLETED:
+		ath10k_dbg(ATH10K_DBG_WMI, "SCAN_EVENT_COMPLETED\n");
+		switch (reason) {
+		case WMI_SCAN_REASON_COMPLETED:
+			ath10k_dbg(ATH10K_DBG_WMI, "SCAN_REASON_COMPLETED\n");
+			break;
+		case WMI_SCAN_REASON_CANCELLED:
+			ath10k_dbg(ATH10K_DBG_WMI, "SCAN_REASON_CANCELED\n");
+			break;
+		case WMI_SCAN_REASON_PREEMPTED:
+			ath10k_dbg(ATH10K_DBG_WMI, "SCAN_REASON_PREEMPTED\n");
+			break;
+		case WMI_SCAN_REASON_TIMEDOUT:
+			ath10k_dbg(ATH10K_DBG_WMI, "SCAN_REASON_TIMEDOUT\n");
+			break;
+		default:
+			break;
+		}
+
+		ar->scan_channel = NULL;
+		if (!ar->scan.in_progress) {
+			ath10k_warn("no scan requested, ignoring\n");
+			break;
+		}
+
+		if (ar->scan.is_roc) {
+			ath10k_offchan_tx_purge(ar);
+
+			if (!ar->scan.aborting)
+				ieee80211_remain_on_channel_expired(ar->hw);
+		} else {
+			ieee80211_scan_completed(ar->hw, ar->scan.aborting);
+		}
+
+		del_timer(&ar->scan.timeout);
+		complete_all(&ar->scan.completed);
+		ar->scan.in_progress = false;
+		break;
+	case WMI_SCAN_EVENT_BSS_CHANNEL:
+		ath10k_dbg(ATH10K_DBG_WMI, "SCAN_EVENT_BSS_CHANNEL\n");
+		ar->scan_channel = NULL;
+		break;
+	case WMI_SCAN_EVENT_FOREIGN_CHANNEL:
+		ath10k_dbg(ATH10K_DBG_WMI, "SCAN_EVENT_FOREIGN_CHANNEL\n");
+		ar->scan_channel = ieee80211_get_channel(ar->hw->wiphy, freq);
+		if (ar->scan.in_progress && ar->scan.is_roc &&
+		    ar->scan.roc_freq == freq) {
+			complete(&ar->scan.on_channel);
+		}
+		break;
+	case WMI_SCAN_EVENT_DEQUEUED:
+		ath10k_dbg(ATH10K_DBG_WMI, "SCAN_EVENT_DEQUEUED\n");
+		break;
+	case WMI_SCAN_EVENT_PREEMPTED:
+		ath10k_dbg(ATH10K_DBG_WMI, "WMI_SCAN_EVENT_PREEMPTED\n");
+		break;
+	case WMI_SCAN_EVENT_START_FAILED:
+		ath10k_dbg(ATH10K_DBG_WMI, "WMI_SCAN_EVENT_START_FAILED\n");
+		break;
+	default:
+		break;
+	}
+
+	spin_unlock_bh(&ar->data_lock);
+	return 0;
+}
+
+static inline enum ieee80211_band phy_mode_to_band(u32 phy_mode)
+{
+	enum ieee80211_band band;
+
+	switch (phy_mode) {
+	case MODE_11A:
+	case MODE_11NA_HT20:
+	case MODE_11NA_HT40:
+	case MODE_11AC_VHT20:
+	case MODE_11AC_VHT40:
+	case MODE_11AC_VHT80:
+		band = IEEE80211_BAND_5GHZ;
+		break;
+	case MODE_11G:
+	case MODE_11B:
+	case MODE_11GONLY:
+	case MODE_11NG_HT20:
+	case MODE_11NG_HT40:
+	case MODE_11AC_VHT20_2G:
+	case MODE_11AC_VHT40_2G:
+	case MODE_11AC_VHT80_2G:
+	default:
+		band = IEEE80211_BAND_2GHZ;
+	}
+
+	return band;
+}
+
+static inline u8 get_rate_idx(u32 rate, enum ieee80211_band band)
+{
+	u8 rate_idx = 0;
+
+	/* rate in Kbps */
+	switch (rate) {
+	case 1000:
+		rate_idx = 0;
+		break;
+	case 2000:
+		rate_idx = 1;
+		break;
+	case 5500:
+		rate_idx = 2;
+		break;
+	case 11000:
+		rate_idx = 3;
+		break;
+	case 6000:
+		rate_idx = 4;
+		break;
+	case 9000:
+		rate_idx = 5;
+		break;
+	case 12000:
+		rate_idx = 6;
+		break;
+	case 18000:
+		rate_idx = 7;
+		break;
+	case 24000:
+		rate_idx = 8;
+		break;
+	case 36000:
+		rate_idx = 9;
+		break;
+	case 48000:
+		rate_idx = 10;
+		break;
+	case 54000:
+		rate_idx = 11;
+		break;
+	default:
+		break;
+	}
+
+	if (band == IEEE80211_BAND_5GHZ) {
+		if (rate_idx > 3)
+			/* Omit CCK rates */
+			rate_idx -= 4;
+		else
+			rate_idx = 0;
+	}
+
+	return rate_idx;
+}
+
+static int ath10k_wmi_event_mgmt_rx(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct wmi_mgmt_rx_event *event = (struct wmi_mgmt_rx_event *)skb->data;
+	struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(skb);
+	struct ieee80211_hdr *hdr;
+	u32 rx_status;
+	u32 channel;
+	u32 phy_mode;
+	u32 snr;
+	u32 rate;
+	u32 buf_len;
+	u16 fc;
+
+	channel   = __le32_to_cpu(event->hdr.channel);
+	buf_len   = __le32_to_cpu(event->hdr.buf_len);
+	rx_status = __le32_to_cpu(event->hdr.status);
+	snr       = __le32_to_cpu(event->hdr.snr);
+	phy_mode  = __le32_to_cpu(event->hdr.phy_mode);
+	rate	  = __le32_to_cpu(event->hdr.rate);
+
+	memset(status, 0, sizeof(*status));
+
+	ath10k_dbg(ATH10K_DBG_MGMT,
+		   "event mgmt rx status %08x\n", rx_status);
+
+	if (rx_status & WMI_RX_STATUS_ERR_DECRYPT) {
+		dev_kfree_skb(skb);
+		return 0;
+	}
+
+	if (rx_status & WMI_RX_STATUS_ERR_KEY_CACHE_MISS) {
+		dev_kfree_skb(skb);
+		return 0;
+	}
+
+	if (rx_status & WMI_RX_STATUS_ERR_CRC)
+		status->flag |= RX_FLAG_FAILED_FCS_CRC;
+	if (rx_status & WMI_RX_STATUS_ERR_MIC)
+		status->flag |= RX_FLAG_MMIC_ERROR;
+
+	status->band = phy_mode_to_band(phy_mode);
+	status->freq = ieee80211_channel_to_frequency(channel, status->band);
+	status->signal = snr + ATH10K_DEFAULT_NOISE_FLOOR;
+	status->rate_idx = get_rate_idx(rate, status->band);
+
+	skb_pull(skb, sizeof(event->hdr));
+
+	hdr = (struct ieee80211_hdr *)skb->data;
+	fc = le16_to_cpu(hdr->frame_control);
+
+	if (fc & IEEE80211_FCTL_PROTECTED) {
+		status->flag |= RX_FLAG_DECRYPTED | RX_FLAG_IV_STRIPPED |
+				RX_FLAG_MMIC_STRIPPED;
+		hdr->frame_control = __cpu_to_le16(fc &
+					~IEEE80211_FCTL_PROTECTED);
+	}
+
+	ath10k_dbg(ATH10K_DBG_MGMT,
+		   "event mgmt rx skb %p len %d ftype %02x stype %02x\n",
+		   skb, skb->len,
+		   fc & IEEE80211_FCTL_FTYPE, fc & IEEE80211_FCTL_STYPE);
+
+	ath10k_dbg(ATH10K_DBG_MGMT,
+		   "event mgmt rx freq %d band %d snr %d, rate_idx %d\n",
+		   status->freq, status->band, status->signal,
+		   status->rate_idx);
+
+	/*
+	 * packets from HTC come aligned to 4byte boundaries
+	 * because they can originally come in along with a trailer
+	 */
+	skb_trim(skb, buf_len);
+
+	ieee80211_rx(ar->hw, skb);
+	return 0;
+}
+
+static void ath10k_wmi_event_chan_info(struct ath10k *ar, struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_CHAN_INFO_EVENTID\n");
+}
+
+static void ath10k_wmi_event_echo(struct ath10k *ar, struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_ECHO_EVENTID\n");
+}
+
+static void ath10k_wmi_event_debug_mesg(struct ath10k *ar, struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_DEBUG_MESG_EVENTID\n");
+}
+
+static void ath10k_wmi_event_update_stats(struct ath10k *ar,
+					  struct sk_buff *skb)
+{
+	struct wmi_stats_event *ev = (struct wmi_stats_event *)skb->data;
+
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_UPDATE_STATS_EVENTID\n");
+
+	ath10k_debug_read_target_stats(ar, ev);
+}
+
+static void ath10k_wmi_event_vdev_start_resp(struct ath10k *ar,
+					     struct sk_buff *skb)
+{
+	struct wmi_vdev_start_response_event *ev;
+
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_VDEV_START_RESP_EVENTID\n");
+
+	ev = (struct wmi_vdev_start_response_event *)skb->data;
+
+	if (WARN_ON(__le32_to_cpu(ev->status)))
+		return;
+
+	complete(&ar->vdev_setup_done);
+}
+
+static void ath10k_wmi_event_vdev_stopped(struct ath10k *ar,
+					  struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_VDEV_STOPPED_EVENTID\n");
+	complete(&ar->vdev_setup_done);
+}
+
+static void ath10k_wmi_event_peer_sta_kickout(struct ath10k *ar,
+					      struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_PEER_STA_KICKOUT_EVENTID\n");
+}
+
+/*
+ * FIXME
+ *
+ * We don't report to mac80211 sleep state of connected
+ * stations. Due to this mac80211 can't fill in TIM IE
+ * correctly.
+ *
+ * I know of no way of getting nullfunc frames that contain
+ * sleep transition from connected stations - these do not
+ * seem to be sent from the target to the host. There also
+ * doesn't seem to be a dedicated event for that. So the
+ * only way left to do this would be to read tim_bitmap
+ * during SWBA.
+ *
+ * We could probably try using tim_bitmap from SWBA to tell
+ * mac80211 which stations are asleep and which are not. The
+ * problem here is calling mac80211 functions so many times
+ * could take too long and make us miss the time to submit
+ * the beacon to the target.
+ *
+ * So as a workaround we try to extend the TIM IE if there
+ * is unicast buffered for stations with aid > 7 and fill it
+ * in ourselves.
+ */
+static void ath10k_wmi_update_tim(struct ath10k *ar,
+				  struct ath10k_vif *arvif,
+				  struct sk_buff *bcn,
+				  struct wmi_bcn_info *bcn_info)
+{
+	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)bcn->data;
+	struct ieee80211_tim_ie *tim;
+	u8 *ies, *ie;
+	u8 ie_len, pvm_len;
+
+	/* if next SWBA has no tim_changed the tim_bitmap is garbage.
+	 * we must copy the bitmap upon change and reuse it later */
+	if (__le32_to_cpu(bcn_info->tim_info.tim_changed)) {
+		int i;
+
+		BUILD_BUG_ON(sizeof(arvif->u.ap.tim_bitmap) !=
+			     sizeof(bcn_info->tim_info.tim_bitmap));
+
+		for (i = 0; i < sizeof(arvif->u.ap.tim_bitmap); i++) {
+			__le32 t = bcn_info->tim_info.tim_bitmap[i / 4];
+			u32 v = __le32_to_cpu(t);
+			arvif->u.ap.tim_bitmap[i] = (v >> ((i % 4) * 8)) & 0xFF;
+		}
+
+		/* FW reports either length 0 or 16
+		 * so we calculate this on our own */
+		arvif->u.ap.tim_len = 0;
+		for (i = 0; i < sizeof(arvif->u.ap.tim_bitmap); i++)
+			if (arvif->u.ap.tim_bitmap[i])
+				arvif->u.ap.tim_len = i;
+
+		arvif->u.ap.tim_len++;
+	}
+
+	ies = bcn->data;
+	ies += ieee80211_hdrlen(hdr->frame_control);
+	ies += 12; /* fixed parameters */
+
+	ie = (u8 *)cfg80211_find_ie(WLAN_EID_TIM, ies,
+				    (u8 *)skb_tail_pointer(bcn) - ies);
+	if (!ie) {
+		/* highly unlikely for mac80211 */
+		ath10k_warn("no tim ie found;\n");
+		return;
+	}
+
+	tim = (void *)ie + 2;
+	ie_len = ie[1];
+	pvm_len = ie_len - 3; /* exclude dtim count, dtim period, bmap ctl */
+
+	if (pvm_len < arvif->u.ap.tim_len) {
+		int expand_size = sizeof(arvif->u.ap.tim_bitmap) - pvm_len;
+		int move_size = skb_tail_pointer(bcn) - (ie + 2 + ie_len);
+		void *next_ie = ie + 2 + ie_len;
+
+		if (skb_put(bcn, expand_size)) {
+			memmove(next_ie + expand_size, next_ie, move_size);
+
+			ie[1] += expand_size;
+			ie_len += expand_size;
+			pvm_len += expand_size;
+		} else {
+			ath10k_warn("tim expansion failed\n");
+		}
+	}
+
+	if (pvm_len > sizeof(arvif->u.ap.tim_bitmap)) {
+		ath10k_warn("tim pvm length is too great (%d)\n", pvm_len);
+		return;
+	}
+
+	tim->bitmap_ctrl = !!__le32_to_cpu(bcn_info->tim_info.tim_mcast);
+	memcpy(tim->virtual_map, arvif->u.ap.tim_bitmap, pvm_len);
+
+	ath10k_dbg(ATH10K_DBG_MGMT, "dtim %d/%d mcast %d pvmlen %d\n",
+		   tim->dtim_count, tim->dtim_period,
+		   tim->bitmap_ctrl, pvm_len);
+}
+
+static void ath10k_p2p_fill_noa_ie(u8 *data, u32 len,
+				   struct wmi_p2p_noa_info *noa)
+{
+	struct ieee80211_p2p_noa_attr *noa_attr;
+	u8  ctwindow_oppps = noa->ctwindow_oppps;
+	u8 ctwindow = ctwindow_oppps >> WMI_P2P_OPPPS_CTWINDOW_OFFSET;
+	bool oppps = !!(ctwindow_oppps & WMI_P2P_OPPPS_ENABLE_BIT);
+	__le16 *noa_attr_len;
+	u16 attr_len;
+	u8 noa_descriptors = noa->num_descriptors;
+	int i;
+
+	/* P2P IE */
+	data[0] = WLAN_EID_VENDOR_SPECIFIC;
+	data[1] = len - 2;
+	data[2] = (WLAN_OUI_WFA >> 16) & 0xff;
+	data[3] = (WLAN_OUI_WFA >> 8) & 0xff;
+	data[4] = (WLAN_OUI_WFA >> 0) & 0xff;
+	data[5] = WLAN_OUI_TYPE_WFA_P2P;
+
+	/* NOA ATTR */
+	data[6] = IEEE80211_P2P_ATTR_ABSENCE_NOTICE;
+	noa_attr_len = (__le16 *)&data[7]; /* 2 bytes */
+	noa_attr = (struct ieee80211_p2p_noa_attr *)&data[9];
+
+	noa_attr->index = noa->index;
+	noa_attr->oppps_ctwindow = ctwindow;
+	if (oppps)
+		noa_attr->oppps_ctwindow |= IEEE80211_P2P_OPPPS_ENABLE_BIT;
+
+	for (i = 0; i < noa_descriptors; i++) {
+		noa_attr->desc[i].count =
+			__le32_to_cpu(noa->descriptors[i].type_count);
+		noa_attr->desc[i].duration = noa->descriptors[i].duration;
+		noa_attr->desc[i].interval = noa->descriptors[i].interval;
+		noa_attr->desc[i].start_time = noa->descriptors[i].start_time;
+	}
+
+	attr_len = 2; /* index + oppps_ctwindow */
+	attr_len += noa_descriptors * sizeof(struct ieee80211_p2p_noa_desc);
+	*noa_attr_len = __cpu_to_le16(attr_len);
+}
+
+static u32 ath10k_p2p_calc_noa_ie_len(struct wmi_p2p_noa_info *noa)
+{
+	u32 len = 0;
+	u8 noa_descriptors = noa->num_descriptors;
+	u8 opp_ps_info = noa->ctwindow_oppps;
+	bool opps_enabled = !!(opp_ps_info & WMI_P2P_OPPPS_ENABLE_BIT);
+
+
+	if (!noa_descriptors && !opps_enabled)
+		return len;
+
+	len += 1 + 1 + 4; /* EID + len + OUI */
+	len += 1 + 2; /* noa attr  + attr len */
+	len += 1 + 1; /* index + oppps_ctwindow */
+	len += noa_descriptors * sizeof(struct ieee80211_p2p_noa_desc);
+
+	return len;
+}
+
+static void ath10k_wmi_update_noa(struct ath10k *ar, struct ath10k_vif *arvif,
+				  struct sk_buff *bcn,
+				  struct wmi_bcn_info *bcn_info)
+{
+	struct wmi_p2p_noa_info *noa = &bcn_info->p2p_noa_info;
+	u8 *new_data, *old_data = arvif->u.ap.noa_data;
+	u32 new_len;
+
+	if (arvif->vdev_subtype != WMI_VDEV_SUBTYPE_P2P_GO)
+		return;
+
+	ath10k_dbg(ATH10K_DBG_MGMT, "noa changed: %d\n", noa->changed);
+	if (noa->changed & WMI_P2P_NOA_CHANGED_BIT) {
+		new_len = ath10k_p2p_calc_noa_ie_len(noa);
+		if (!new_len)
+			goto cleanup;
+
+		new_data = kmalloc(new_len, GFP_ATOMIC);
+		if (!new_data)
+			goto cleanup;
+
+		ath10k_p2p_fill_noa_ie(new_data, new_len, noa);
+
+		spin_lock_bh(&ar->data_lock);
+		arvif->u.ap.noa_data = new_data;
+		arvif->u.ap.noa_len = new_len;
+		spin_unlock_bh(&ar->data_lock);
+		kfree(old_data);
+	}
+
+	if (arvif->u.ap.noa_data)
+		if (!pskb_expand_head(bcn, 0, arvif->u.ap.noa_len, GFP_ATOMIC))
+			memcpy(skb_put(bcn, arvif->u.ap.noa_len),
+			       arvif->u.ap.noa_data,
+			       arvif->u.ap.noa_len);
+	return;
+
+cleanup:
+	spin_lock_bh(&ar->data_lock);
+	arvif->u.ap.noa_data = NULL;
+	arvif->u.ap.noa_len = 0;
+	spin_unlock_bh(&ar->data_lock);
+	kfree(old_data);
+}
+
+
+static void ath10k_wmi_event_host_swba(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct wmi_host_swba_event *ev;
+	u32 map;
+	int i = -1;
+	struct wmi_bcn_info *bcn_info;
+	struct ath10k_vif *arvif;
+	struct wmi_bcn_tx_arg arg;
+	struct sk_buff *bcn;
+	int vdev_id = 0;
+	int ret;
+
+	ath10k_dbg(ATH10K_DBG_MGMT, "WMI_HOST_SWBA_EVENTID\n");
+
+	ev = (struct wmi_host_swba_event *)skb->data;
+	map = __le32_to_cpu(ev->vdev_map);
+
+	ath10k_dbg(ATH10K_DBG_MGMT, "host swba:\n"
+		   "-vdev map 0x%x\n",
+		   ev->vdev_map);
+
+	for (; map; map >>= 1, vdev_id++) {
+		if (!(map & 0x1))
+			continue;
+
+		i++;
+
+		if (i >= WMI_MAX_AP_VDEV) {
+			ath10k_warn("swba has corrupted vdev map\n");
+			break;
+		}
+
+		bcn_info = &ev->bcn_info[i];
+
+		ath10k_dbg(ATH10K_DBG_MGMT,
+			   "-bcn_info[%d]:\n"
+			   "--tim_len %d\n"
+			   "--tim_mcast %d\n"
+			   "--tim_changed %d\n"
+			   "--tim_num_ps_pending %d\n"
+			   "--tim_bitmap 0x%08x%08x%08x%08x\n",
+			   i,
+			   __le32_to_cpu(bcn_info->tim_info.tim_len),
+			   __le32_to_cpu(bcn_info->tim_info.tim_mcast),
+			   __le32_to_cpu(bcn_info->tim_info.tim_changed),
+			   __le32_to_cpu(bcn_info->tim_info.tim_num_ps_pending),
+			   __le32_to_cpu(bcn_info->tim_info.tim_bitmap[3]),
+			   __le32_to_cpu(bcn_info->tim_info.tim_bitmap[2]),
+			   __le32_to_cpu(bcn_info->tim_info.tim_bitmap[1]),
+			   __le32_to_cpu(bcn_info->tim_info.tim_bitmap[0]));
+
+		arvif = ath10k_get_arvif(ar, vdev_id);
+		if (arvif == NULL) {
+			ath10k_warn("no vif for vdev_id %d found\n", vdev_id);
+			continue;
+		}
+
+		bcn = ieee80211_beacon_get(ar->hw, arvif->vif);
+		if (!bcn) {
+			ath10k_warn("could not get mac80211 beacon\n");
+			continue;
+		}
+
+		ath10k_tx_h_seq_no(bcn);
+		ath10k_wmi_update_tim(ar, arvif, bcn, bcn_info);
+		ath10k_wmi_update_noa(ar, arvif, bcn, bcn_info);
+
+		arg.vdev_id = arvif->vdev_id;
+		arg.tx_rate = 0;
+		arg.tx_power = 0;
+		arg.bcn = bcn->data;
+		arg.bcn_len = bcn->len;
+
+		ret = ath10k_wmi_beacon_send(ar, &arg);
+		if (ret)
+			ath10k_warn("could not send beacon (%d)\n", ret);
+
+		dev_kfree_skb_any(bcn);
+	}
+}
+
+static void ath10k_wmi_event_tbttoffset_update(struct ath10k *ar,
+					       struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_TBTTOFFSET_UPDATE_EVENTID\n");
+}
+
+static void ath10k_wmi_event_phyerr(struct ath10k *ar, struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_PHYERR_EVENTID\n");
+}
+
+static void ath10k_wmi_event_roam(struct ath10k *ar, struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_ROAM_EVENTID\n");
+}
+
+static void ath10k_wmi_event_profile_match(struct ath10k *ar,
+				    struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_PROFILE_MATCH\n");
+}
+
+static void ath10k_wmi_event_debug_print(struct ath10k *ar,
+				  struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_DEBUG_PRINT_EVENTID\n");
+}
+
+static void ath10k_wmi_event_pdev_qvit(struct ath10k *ar, struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_PDEV_QVIT_EVENTID\n");
+}
+
+static void ath10k_wmi_event_wlan_profile_data(struct ath10k *ar,
+					       struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_WLAN_PROFILE_DATA_EVENTID\n");
+}
+
+static void ath10k_wmi_event_rtt_measurement_report(struct ath10k *ar,
+					     struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_RTT_MEASUREMENT_REPORT_EVENTID\n");
+}
+
+static void ath10k_wmi_event_tsf_measurement_report(struct ath10k *ar,
+					     struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_TSF_MEASUREMENT_REPORT_EVENTID\n");
+}
+
+static void ath10k_wmi_event_rtt_error_report(struct ath10k *ar,
+					      struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_RTT_ERROR_REPORT_EVENTID\n");
+}
+
+static void ath10k_wmi_event_wow_wakeup_host(struct ath10k *ar,
+					     struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_WOW_WAKEUP_HOST_EVENTID\n");
+}
+
+static void ath10k_wmi_event_dcs_interference(struct ath10k *ar,
+					      struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_DCS_INTERFERENCE_EVENTID\n");
+}
+
+static void ath10k_wmi_event_pdev_tpc_config(struct ath10k *ar,
+					     struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_PDEV_TPC_CONFIG_EVENTID\n");
+}
+
+static void ath10k_wmi_event_pdev_ftm_intg(struct ath10k *ar,
+					   struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_PDEV_FTM_INTG_EVENTID\n");
+}
+
+static void ath10k_wmi_event_gtk_offload_status(struct ath10k *ar,
+					 struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_GTK_OFFLOAD_STATUS_EVENTID\n");
+}
+
+static void ath10k_wmi_event_gtk_rekey_fail(struct ath10k *ar,
+					    struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_GTK_REKEY_FAIL_EVENTID\n");
+}
+
+static void ath10k_wmi_event_delba_complete(struct ath10k *ar,
+					    struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_TX_DELBA_COMPLETE_EVENTID\n");
+}
+
+static void ath10k_wmi_event_addba_complete(struct ath10k *ar,
+					    struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_TX_ADDBA_COMPLETE_EVENTID\n");
+}
+
+static void ath10k_wmi_event_vdev_install_key_complete(struct ath10k *ar,
+						struct sk_buff *skb)
+{
+	ath10k_dbg(ATH10K_DBG_WMI, "WMI_VDEV_INSTALL_KEY_COMPLETE_EVENTID\n");
+}
+
+static void ath10k_wmi_service_ready_event_rx(struct ath10k *ar,
+					      struct sk_buff *skb)
+{
+	struct wmi_service_ready_event *ev = (void *)skb->data;
+
+	if (skb->len < sizeof(*ev)) {
+		ath10k_warn("Service ready event was %d B but expected %zu B. Wrong firmware version?\n",
+			    skb->len, sizeof(*ev));
+		return;
+	}
+
+	ar->hw_min_tx_power = __le32_to_cpu(ev->hw_min_tx_power);
+	ar->hw_max_tx_power = __le32_to_cpu(ev->hw_max_tx_power);
+	ar->ht_cap_info = __le32_to_cpu(ev->ht_cap_info);
+	ar->vht_cap_info = __le32_to_cpu(ev->vht_cap_info);
+	ar->fw_version_major =
+		(__le32_to_cpu(ev->sw_version) & 0xff000000) >> 24;
+	ar->fw_version_minor = (__le32_to_cpu(ev->sw_version) & 0x00ffffff);
+	ar->fw_version_release =
+		(__le32_to_cpu(ev->sw_version_1) & 0xffff0000) >> 16;
+	ar->fw_version_build = (__le32_to_cpu(ev->sw_version_1) & 0x0000ffff);
+	ar->phy_capability = __le32_to_cpu(ev->phy_capability);
+
+	ar->ath_common.regulatory.current_rd =
+		__le32_to_cpu(ev->hal_reg_capabilities.eeprom_rd);
+
+	ath10k_debug_read_service_map(ar, ev->wmi_service_bitmap,
+				      sizeof(ev->wmi_service_bitmap));
+
+	if (strlen(ar->hw->wiphy->fw_version) == 0) {
+		snprintf(ar->hw->wiphy->fw_version,
+			 sizeof(ar->hw->wiphy->fw_version),
+			 "%u.%u.%u.%u",
+			 ar->fw_version_major,
+			 ar->fw_version_minor,
+			 ar->fw_version_release,
+			 ar->fw_version_build);
+	}
+
+	/* FIXME: it probably should be better to support this */
+	if (__le32_to_cpu(ev->num_mem_reqs) > 0) {
+		ath10k_warn("target requested %d memory chunks; ignoring\n",
+			    __le32_to_cpu(ev->num_mem_reqs));
+	}
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi event service ready sw_ver 0x%08x sw_ver1 0x%08x abi_ver %u phy_cap 0x%08x ht_cap 0x%08x vht_cap 0x%08x vht_supp_msc 0x%08x sys_cap_info 0x%08x mem_reqs %u\n",
+		   __le32_to_cpu(ev->sw_version),
+		   __le32_to_cpu(ev->sw_version_1),
+		   __le32_to_cpu(ev->abi_version),
+		   __le32_to_cpu(ev->phy_capability),
+		   __le32_to_cpu(ev->ht_cap_info),
+		   __le32_to_cpu(ev->vht_cap_info),
+		   __le32_to_cpu(ev->vht_supp_mcs),
+		   __le32_to_cpu(ev->sys_cap_info),
+		   __le32_to_cpu(ev->num_mem_reqs));
+
+	complete(&ar->wmi.service_ready);
+}
+
+static int ath10k_wmi_ready_event_rx(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct wmi_ready_event *ev = (struct wmi_ready_event *)skb->data;
+
+	if (WARN_ON(skb->len < sizeof(*ev)))
+		return -EINVAL;
+
+	memcpy(ar->mac_addr, ev->mac_addr.addr, ETH_ALEN);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi event ready sw_version %u abi_version %u mac_addr %pM status %d\n",
+		   __le32_to_cpu(ev->sw_version),
+		   __le32_to_cpu(ev->abi_version),
+		   ev->mac_addr.addr,
+		   __le32_to_cpu(ev->status));
+
+	complete(&ar->wmi.unified_ready);
+	return 0;
+}
+
+static void ath10k_wmi_event_process(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct wmi_cmd_hdr *cmd_hdr;
+	enum wmi_event_id id;
+	u16 len;
+
+	cmd_hdr = (struct wmi_cmd_hdr *)skb->data;
+	id = MS(__le32_to_cpu(cmd_hdr->cmd_id), WMI_CMD_HDR_CMD_ID);
+
+	if (skb_pull(skb, sizeof(struct wmi_cmd_hdr)) == NULL)
+		return;
+
+	len = skb->len;
+
+	trace_ath10k_wmi_event(id, skb->data, skb->len);
+
+	switch (id) {
+	case WMI_MGMT_RX_EVENTID:
+		ath10k_wmi_event_mgmt_rx(ar, skb);
+		/* mgmt_rx() owns the skb now! */
+		return;
+	case WMI_SCAN_EVENTID:
+		ath10k_wmi_event_scan(ar, skb);
+		break;
+	case WMI_CHAN_INFO_EVENTID:
+		ath10k_wmi_event_chan_info(ar, skb);
+		break;
+	case WMI_ECHO_EVENTID:
+		ath10k_wmi_event_echo(ar, skb);
+		break;
+	case WMI_DEBUG_MESG_EVENTID:
+		ath10k_wmi_event_debug_mesg(ar, skb);
+		break;
+	case WMI_UPDATE_STATS_EVENTID:
+		ath10k_wmi_event_update_stats(ar, skb);
+		break;
+	case WMI_VDEV_START_RESP_EVENTID:
+		ath10k_wmi_event_vdev_start_resp(ar, skb);
+		break;
+	case WMI_VDEV_STOPPED_EVENTID:
+		ath10k_wmi_event_vdev_stopped(ar, skb);
+		break;
+	case WMI_PEER_STA_KICKOUT_EVENTID:
+		ath10k_wmi_event_peer_sta_kickout(ar, skb);
+		break;
+	case WMI_HOST_SWBA_EVENTID:
+		ath10k_wmi_event_host_swba(ar, skb);
+		break;
+	case WMI_TBTTOFFSET_UPDATE_EVENTID:
+		ath10k_wmi_event_tbttoffset_update(ar, skb);
+		break;
+	case WMI_PHYERR_EVENTID:
+		ath10k_wmi_event_phyerr(ar, skb);
+		break;
+	case WMI_ROAM_EVENTID:
+		ath10k_wmi_event_roam(ar, skb);
+		break;
+	case WMI_PROFILE_MATCH:
+		ath10k_wmi_event_profile_match(ar, skb);
+		break;
+	case WMI_DEBUG_PRINT_EVENTID:
+		ath10k_wmi_event_debug_print(ar, skb);
+		break;
+	case WMI_PDEV_QVIT_EVENTID:
+		ath10k_wmi_event_pdev_qvit(ar, skb);
+		break;
+	case WMI_WLAN_PROFILE_DATA_EVENTID:
+		ath10k_wmi_event_wlan_profile_data(ar, skb);
+		break;
+	case WMI_RTT_MEASUREMENT_REPORT_EVENTID:
+		ath10k_wmi_event_rtt_measurement_report(ar, skb);
+		break;
+	case WMI_TSF_MEASUREMENT_REPORT_EVENTID:
+		ath10k_wmi_event_tsf_measurement_report(ar, skb);
+		break;
+	case WMI_RTT_ERROR_REPORT_EVENTID:
+		ath10k_wmi_event_rtt_error_report(ar, skb);
+		break;
+	case WMI_WOW_WAKEUP_HOST_EVENTID:
+		ath10k_wmi_event_wow_wakeup_host(ar, skb);
+		break;
+	case WMI_DCS_INTERFERENCE_EVENTID:
+		ath10k_wmi_event_dcs_interference(ar, skb);
+		break;
+	case WMI_PDEV_TPC_CONFIG_EVENTID:
+		ath10k_wmi_event_pdev_tpc_config(ar, skb);
+		break;
+	case WMI_PDEV_FTM_INTG_EVENTID:
+		ath10k_wmi_event_pdev_ftm_intg(ar, skb);
+		break;
+	case WMI_GTK_OFFLOAD_STATUS_EVENTID:
+		ath10k_wmi_event_gtk_offload_status(ar, skb);
+		break;
+	case WMI_GTK_REKEY_FAIL_EVENTID:
+		ath10k_wmi_event_gtk_rekey_fail(ar, skb);
+		break;
+	case WMI_TX_DELBA_COMPLETE_EVENTID:
+		ath10k_wmi_event_delba_complete(ar, skb);
+		break;
+	case WMI_TX_ADDBA_COMPLETE_EVENTID:
+		ath10k_wmi_event_addba_complete(ar, skb);
+		break;
+	case WMI_VDEV_INSTALL_KEY_COMPLETE_EVENTID:
+		ath10k_wmi_event_vdev_install_key_complete(ar, skb);
+		break;
+	case WMI_SERVICE_READY_EVENTID:
+		ath10k_wmi_service_ready_event_rx(ar, skb);
+		break;
+	case WMI_READY_EVENTID:
+		ath10k_wmi_ready_event_rx(ar, skb);
+		break;
+	default:
+		ath10k_warn("Unknown eventid: %d\n", id);
+		break;
+	}
+
+	dev_kfree_skb(skb);
+}
+
+static void ath10k_wmi_event_work(struct work_struct *work)
+{
+	struct ath10k *ar = container_of(work, struct ath10k,
+					 wmi.wmi_event_work);
+	struct sk_buff *skb;
+
+	for (;;) {
+		skb = skb_dequeue(&ar->wmi.wmi_event_list);
+		if (!skb)
+			break;
+
+		ath10k_wmi_event_process(ar, skb);
+	}
+}
+
+static void ath10k_wmi_process_rx(struct ath10k *ar, struct sk_buff *skb)
+{
+	struct wmi_cmd_hdr *cmd_hdr = (struct wmi_cmd_hdr *)skb->data;
+	enum wmi_event_id event_id;
+
+	event_id = MS(__le32_to_cpu(cmd_hdr->cmd_id), WMI_CMD_HDR_CMD_ID);
+
+	/* some events require to be handled ASAP
+	 * thus can't be defered to a worker thread */
+	switch (event_id) {
+	case WMI_HOST_SWBA_EVENTID:
+	case WMI_MGMT_RX_EVENTID:
+		ath10k_wmi_event_process(ar, skb);
+		return;
+	default:
+		break;
+	}
+
+	skb_queue_tail(&ar->wmi.wmi_event_list, skb);
+	queue_work(ar->workqueue, &ar->wmi.wmi_event_work);
+}
+
+/* WMI Initialization functions */
+int ath10k_wmi_attach(struct ath10k *ar)
+{
+	init_completion(&ar->wmi.service_ready);
+	init_completion(&ar->wmi.unified_ready);
+	init_waitqueue_head(&ar->wmi.wq);
+
+	skb_queue_head_init(&ar->wmi.wmi_event_list);
+	INIT_WORK(&ar->wmi.wmi_event_work, ath10k_wmi_event_work);
+
+	return 0;
+}
+
+void ath10k_wmi_detach(struct ath10k *ar)
+{
+	/* HTC should've drained the packets already */
+	if (WARN_ON(atomic_read(&ar->wmi.pending_tx_count) > 0))
+		ath10k_warn("there are still pending packets\n");
+
+	cancel_work_sync(&ar->wmi.wmi_event_work);
+	skb_queue_purge(&ar->wmi.wmi_event_list);
+}
+
+int ath10k_wmi_connect_htc_service(struct ath10k *ar)
+{
+	int status;
+	struct ath10k_htc_svc_conn_req conn_req;
+	struct ath10k_htc_svc_conn_resp conn_resp;
+
+	memset(&conn_req, 0, sizeof(conn_req));
+	memset(&conn_resp, 0, sizeof(conn_resp));
+
+	/* these fields are the same for all service endpoints */
+	conn_req.ep_ops.ep_tx_complete = ath10k_wmi_htc_tx_complete;
+	conn_req.ep_ops.ep_rx_complete = ath10k_wmi_process_rx;
+
+	/* connect to control service */
+	conn_req.service_id = ATH10K_HTC_SVC_ID_WMI_CONTROL;
+
+	status = ath10k_htc_connect_service(ar->htc, &conn_req, &conn_resp);
+	if (status) {
+		ath10k_warn("failed to connect to WMI CONTROL service status: %d\n",
+			    status);
+		return status;
+	}
+
+	ar->wmi.eid = conn_resp.eid;
+	return 0;
+}
+
+int ath10k_wmi_pdev_set_regdomain(struct ath10k *ar, u16 rd, u16 rd2g,
+				  u16 rd5g, u16 ctl2g, u16 ctl5g)
+{
+	struct wmi_pdev_set_regdomain_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_pdev_set_regdomain_cmd *)skb->data;
+	cmd->reg_domain = __cpu_to_le32(rd);
+	cmd->reg_domain_2G = __cpu_to_le32(rd2g);
+	cmd->reg_domain_5G = __cpu_to_le32(rd5g);
+	cmd->conformance_test_limit_2G = __cpu_to_le32(ctl2g);
+	cmd->conformance_test_limit_5G = __cpu_to_le32(ctl5g);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi pdev regdomain rd %x rd2g %x rd5g %x ctl2g %x ctl5g %x\n",
+		   rd, rd2g, rd5g, ctl2g, ctl5g);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PDEV_SET_REGDOMAIN_CMDID);
+}
+
+int ath10k_wmi_pdev_set_channel(struct ath10k *ar,
+				const struct wmi_channel_arg *arg)
+{
+	struct wmi_set_channel_cmd *cmd;
+	struct sk_buff *skb;
+
+	if (arg->passive)
+		return -EINVAL;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_set_channel_cmd *)skb->data;
+	cmd->chan.mhz               = __cpu_to_le32(arg->freq);
+	cmd->chan.band_center_freq1 = __cpu_to_le32(arg->freq);
+	cmd->chan.mode              = arg->mode;
+	cmd->chan.min_power         = arg->min_power;
+	cmd->chan.max_power         = arg->max_power;
+	cmd->chan.reg_power         = arg->max_reg_power;
+	cmd->chan.reg_classid       = arg->reg_class_id;
+	cmd->chan.antenna_max       = arg->max_antenna_gain;
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi set channel mode %d freq %d\n",
+		   arg->mode, arg->freq);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PDEV_SET_CHANNEL_CMDID);
+}
+
+int ath10k_wmi_pdev_suspend_target(struct ath10k *ar)
+{
+	struct wmi_pdev_suspend_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_pdev_suspend_cmd *)skb->data;
+	cmd->suspend_opt = WMI_PDEV_SUSPEND;
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PDEV_SUSPEND_CMDID);
+}
+
+int ath10k_wmi_pdev_resume_target(struct ath10k *ar)
+{
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(0);
+	if (skb == NULL)
+		return -ENOMEM;
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PDEV_RESUME_CMDID);
+}
+
+int ath10k_wmi_pdev_set_param(struct ath10k *ar, enum wmi_pdev_param id,
+			      u32 value)
+{
+	struct wmi_pdev_set_param_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_pdev_set_param_cmd *)skb->data;
+	cmd->param_id    = __cpu_to_le32(id);
+	cmd->param_value = __cpu_to_le32(value);
+
+	ath10k_dbg(ATH10K_DBG_WMI, "wmi pdev set param %d value %d\n",
+		   id, value);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PDEV_SET_PARAM_CMDID);
+}
+
+int ath10k_wmi_cmd_init(struct ath10k *ar)
+{
+	struct wmi_init_cmd *cmd;
+	struct sk_buff *buf;
+	struct wmi_resource_config config = {};
+	u32 val;
+
+	config.num_vdevs = __cpu_to_le32(TARGET_NUM_VDEVS);
+	config.num_peers = __cpu_to_le32(TARGET_NUM_PEERS + TARGET_NUM_VDEVS);
+	config.num_offload_peers = __cpu_to_le32(TARGET_NUM_OFFLOAD_PEERS);
+
+	config.num_offload_reorder_bufs =
+		__cpu_to_le32(TARGET_NUM_OFFLOAD_REORDER_BUFS);
+
+	config.num_peer_keys = __cpu_to_le32(TARGET_NUM_PEER_KEYS);
+	config.num_tids = __cpu_to_le32(TARGET_NUM_TIDS);
+	config.ast_skid_limit = __cpu_to_le32(TARGET_AST_SKID_LIMIT);
+	config.tx_chain_mask = __cpu_to_le32(TARGET_TX_CHAIN_MASK);
+	config.rx_chain_mask = __cpu_to_le32(TARGET_RX_CHAIN_MASK);
+	config.rx_timeout_pri_vo = __cpu_to_le32(TARGET_RX_TIMEOUT_LO_PRI);
+	config.rx_timeout_pri_vi = __cpu_to_le32(TARGET_RX_TIMEOUT_LO_PRI);
+	config.rx_timeout_pri_be = __cpu_to_le32(TARGET_RX_TIMEOUT_LO_PRI);
+	config.rx_timeout_pri_bk = __cpu_to_le32(TARGET_RX_TIMEOUT_HI_PRI);
+	config.rx_decap_mode = __cpu_to_le32(TARGET_RX_DECAP_MODE);
+
+	config.scan_max_pending_reqs =
+		__cpu_to_le32(TARGET_SCAN_MAX_PENDING_REQS);
+
+	config.bmiss_offload_max_vdev =
+		__cpu_to_le32(TARGET_BMISS_OFFLOAD_MAX_VDEV);
+
+	config.roam_offload_max_vdev =
+		__cpu_to_le32(TARGET_ROAM_OFFLOAD_MAX_VDEV);
+
+	config.roam_offload_max_ap_profiles =
+		__cpu_to_le32(TARGET_ROAM_OFFLOAD_MAX_AP_PROFILES);
+
+	config.num_mcast_groups = __cpu_to_le32(TARGET_NUM_MCAST_GROUPS);
+	config.num_mcast_table_elems =
+		__cpu_to_le32(TARGET_NUM_MCAST_TABLE_ELEMS);
+
+	config.mcast2ucast_mode = __cpu_to_le32(TARGET_MCAST2UCAST_MODE);
+	config.tx_dbg_log_size = __cpu_to_le32(TARGET_TX_DBG_LOG_SIZE);
+	config.num_wds_entries = __cpu_to_le32(TARGET_NUM_WDS_ENTRIES);
+	config.dma_burst_size = __cpu_to_le32(TARGET_DMA_BURST_SIZE);
+	config.mac_aggr_delim = __cpu_to_le32(TARGET_MAC_AGGR_DELIM);
+
+	val = TARGET_RX_SKIP_DEFRAG_TIMEOUT_DUP_DETECTION_CHECK;
+	config.rx_skip_defrag_timeout_dup_detection_check = __cpu_to_le32(val);
+
+	config.vow_config = __cpu_to_le32(TARGET_VOW_CONFIG);
+
+	config.gtk_offload_max_vdev =
+		__cpu_to_le32(TARGET_GTK_OFFLOAD_MAX_VDEV);
+
+	config.num_msdu_desc = __cpu_to_le32(TARGET_NUM_MSDU_DESC);
+	config.max_frag_entries = __cpu_to_le32(TARGET_MAX_FRAG_ENTRIES);
+
+	buf = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!buf)
+		return -ENOMEM;
+
+	cmd = (struct wmi_init_cmd *)buf->data;
+	cmd->num_host_mem_chunks = 0;
+	memcpy(&cmd->resource_config, &config, sizeof(config));
+
+	ath10k_dbg(ATH10K_DBG_WMI, "wmi init\n");
+	return ath10k_wmi_cmd_send(ar, buf, WMI_INIT_CMDID);
+}
+
+static int ath10k_wmi_start_scan_calc_len(const struct wmi_start_scan_arg *arg)
+{
+	int len;
+
+	len = sizeof(struct wmi_start_scan_cmd);
+
+	if (arg->ie_len) {
+		if (!arg->ie)
+			return -EINVAL;
+		if (arg->ie_len > WLAN_SCAN_PARAMS_MAX_IE_LEN)
+			return -EINVAL;
+
+		len += sizeof(struct wmi_ie_data);
+		len += roundup(arg->ie_len, 4);
+	}
+
+	if (arg->n_channels) {
+		if (!arg->channels)
+			return -EINVAL;
+		if (arg->n_channels > ARRAY_SIZE(arg->channels))
+			return -EINVAL;
+
+		len += sizeof(struct wmi_chan_list);
+		len += sizeof(__le32) * arg->n_channels;
+	}
+
+	if (arg->n_ssids) {
+		if (!arg->ssids)
+			return -EINVAL;
+		if (arg->n_ssids > WLAN_SCAN_PARAMS_MAX_SSID)
+			return -EINVAL;
+
+		len += sizeof(struct wmi_ssid_list);
+		len += sizeof(struct wmi_ssid) * arg->n_ssids;
+	}
+
+	if (arg->n_bssids) {
+		if (!arg->bssids)
+			return -EINVAL;
+		if (arg->n_bssids > WLAN_SCAN_PARAMS_MAX_BSSID)
+			return -EINVAL;
+
+		len += sizeof(struct wmi_bssid_list);
+		len += sizeof(struct wmi_mac_addr) * arg->n_bssids;
+	}
+
+	return len;
+}
+
+int ath10k_wmi_start_scan(struct ath10k *ar,
+			  const struct wmi_start_scan_arg *arg)
+{
+	struct wmi_start_scan_cmd *cmd;
+	struct sk_buff *skb;
+	struct wmi_ie_data *ie;
+	struct wmi_chan_list *channels;
+	struct wmi_ssid_list *ssids;
+	struct wmi_bssid_list *bssids;
+	u32 scan_id;
+	u32 scan_req_id;
+	int off;
+	int len = 0;
+	int i;
+
+	len = ath10k_wmi_start_scan_calc_len(arg);
+	if (len < 0)
+		return len; /* len contains error code here */
+
+	skb = ath10k_wmi_alloc_skb(len);
+	if (!skb)
+		return -ENOMEM;
+
+	scan_id  = WMI_HOST_SCAN_REQ_ID_PREFIX;
+	scan_id |= arg->scan_id;
+
+	scan_req_id  = WMI_HOST_SCAN_REQUESTOR_ID_PREFIX;
+	scan_req_id |= arg->scan_req_id;
+
+	cmd = (struct wmi_start_scan_cmd *)skb->data;
+	cmd->scan_id            = __cpu_to_le32(scan_id);
+	cmd->scan_req_id        = __cpu_to_le32(scan_req_id);
+	cmd->vdev_id            = __cpu_to_le32(arg->vdev_id);
+	cmd->scan_priority      = __cpu_to_le32(arg->scan_priority);
+	cmd->notify_scan_events = __cpu_to_le32(arg->notify_scan_events);
+	cmd->dwell_time_active  = __cpu_to_le32(arg->dwell_time_active);
+	cmd->dwell_time_passive = __cpu_to_le32(arg->dwell_time_passive);
+	cmd->min_rest_time      = __cpu_to_le32(arg->min_rest_time);
+	cmd->max_rest_time      = __cpu_to_le32(arg->max_rest_time);
+	cmd->repeat_probe_time  = __cpu_to_le32(arg->repeat_probe_time);
+	cmd->probe_spacing_time = __cpu_to_le32(arg->probe_spacing_time);
+	cmd->idle_time          = __cpu_to_le32(arg->idle_time);
+	cmd->max_scan_time      = __cpu_to_le32(arg->max_scan_time);
+	cmd->probe_delay        = __cpu_to_le32(arg->probe_delay);
+	cmd->scan_ctrl_flags    = __cpu_to_le32(arg->scan_ctrl_flags);
+
+	/* TLV list starts after fields included in the struct */
+	off = sizeof(*cmd);
+
+	if (arg->n_channels) {
+		channels = (void *)skb->data + off;
+		channels->tag = __cpu_to_le32(WMI_CHAN_LIST_TAG);
+		channels->num_chan = __cpu_to_le32(arg->n_channels);
+
+		for (i = 0; i < arg->n_channels; i++)
+			channels->channel_list[i] =
+				__cpu_to_le32(arg->channels[i]);
+
+		off += sizeof(*channels);
+		off += sizeof(__le32) * arg->n_channels;
+	}
+
+	if (arg->n_ssids) {
+		ssids = (void *)skb->data + off;
+		ssids->tag = __cpu_to_le32(WMI_SSID_LIST_TAG);
+		ssids->num_ssids = __cpu_to_le32(arg->n_ssids);
+
+		for (i = 0; i < arg->n_ssids; i++) {
+			ssids->ssids[i].ssid_len =
+				__cpu_to_le32(arg->ssids[i].len);
+			memcpy(&ssids->ssids[i].ssid,
+			       arg->ssids[i].ssid,
+			       arg->ssids[i].len);
+		}
+
+		off += sizeof(*ssids);
+		off += sizeof(struct wmi_ssid) * arg->n_ssids;
+	}
+
+	if (arg->n_bssids) {
+		bssids = (void *)skb->data + off;
+		bssids->tag = __cpu_to_le32(WMI_BSSID_LIST_TAG);
+		bssids->num_bssid = __cpu_to_le32(arg->n_bssids);
+
+		for (i = 0; i < arg->n_bssids; i++)
+			memcpy(&bssids->bssid_list[i],
+			       arg->bssids[i].bssid,
+			       ETH_ALEN);
+
+		off += sizeof(*bssids);
+		off += sizeof(struct wmi_mac_addr) * arg->n_bssids;
+	}
+
+	if (arg->ie_len) {
+		ie = (void *)skb->data + off;
+		ie->tag = __cpu_to_le32(WMI_IE_TAG);
+		ie->ie_len = __cpu_to_le32(arg->ie_len);
+		memcpy(ie->ie_data, arg->ie, arg->ie_len);
+
+		off += sizeof(*ie);
+		off += roundup(arg->ie_len, 4);
+	}
+
+	if (off != skb->len) {
+		dev_kfree_skb(skb);
+		return -EINVAL;
+	}
+
+	ath10k_dbg(ATH10K_DBG_WMI, "wmi start scan\n");
+	return ath10k_wmi_cmd_send(ar, skb, WMI_START_SCAN_CMDID);
+}
+
+void ath10k_wmi_start_scan_init(struct ath10k *ar,
+				struct wmi_start_scan_arg *arg)
+{
+	/* setup commonly used values */
+	arg->scan_req_id = 1;
+	arg->scan_priority = WMI_SCAN_PRIORITY_LOW;
+	arg->dwell_time_active = 50;
+	arg->dwell_time_passive = 150;
+	arg->min_rest_time = 50;
+	arg->max_rest_time = 500;
+	arg->repeat_probe_time = 0;
+	arg->probe_spacing_time = 0;
+	arg->idle_time = 0;
+	arg->max_scan_time = 5000;
+	arg->probe_delay = 5;
+	arg->notify_scan_events = WMI_SCAN_EVENT_STARTED
+		| WMI_SCAN_EVENT_COMPLETED
+		| WMI_SCAN_EVENT_BSS_CHANNEL
+		| WMI_SCAN_EVENT_FOREIGN_CHANNEL
+		| WMI_SCAN_EVENT_DEQUEUED;
+	arg->scan_ctrl_flags |= WMI_SCAN_ADD_OFDM_RATES;
+	arg->scan_ctrl_flags |= WMI_SCAN_CHAN_STAT_EVENT;
+	arg->n_bssids = 1;
+	arg->bssids[0].bssid = "\xFF\xFF\xFF\xFF\xFF\xFF";
+}
+
+int ath10k_wmi_stop_scan(struct ath10k *ar, const struct wmi_stop_scan_arg *arg)
+{
+	struct wmi_stop_scan_cmd *cmd;
+	struct sk_buff *skb;
+	u32 scan_id;
+	u32 req_id;
+
+	if (arg->req_id > 0xFFF)
+		return -EINVAL;
+	if (arg->req_type == WMI_SCAN_STOP_ONE && arg->u.scan_id > 0xFFF)
+		return -EINVAL;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	scan_id = arg->u.scan_id;
+	scan_id |= WMI_HOST_SCAN_REQ_ID_PREFIX;
+
+	req_id = arg->req_id;
+	req_id |= WMI_HOST_SCAN_REQUESTOR_ID_PREFIX;
+
+	cmd = (struct wmi_stop_scan_cmd *)skb->data;
+	cmd->req_type    = __cpu_to_le32(arg->req_type);
+	cmd->vdev_id     = __cpu_to_le32(arg->u.vdev_id);
+	cmd->scan_id     = __cpu_to_le32(scan_id);
+	cmd->scan_req_id = __cpu_to_le32(req_id);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi stop scan reqid %d req_type %d vdev/scan_id %d\n",
+		   arg->req_id, arg->req_type, arg->u.scan_id);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_STOP_SCAN_CMDID);
+}
+
+int ath10k_wmi_vdev_create(struct ath10k *ar, u32 vdev_id,
+			   enum wmi_vdev_type type,
+			   enum wmi_vdev_subtype subtype,
+			   const u8 macaddr[ETH_ALEN])
+{
+	struct wmi_vdev_create_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_create_cmd *)skb->data;
+	cmd->vdev_id      = __cpu_to_le32(vdev_id);
+	cmd->vdev_type    = __cpu_to_le32(type);
+	cmd->vdev_subtype = __cpu_to_le32(subtype);
+	memcpy(cmd->vdev_macaddr.addr, macaddr, ETH_ALEN);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "WMI vdev create: id %d type %d subtype %d macaddr %pM\n",
+		   vdev_id, type, subtype, macaddr);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_CREATE_CMDID);
+}
+
+int ath10k_wmi_vdev_delete(struct ath10k *ar, u32 vdev_id)
+{
+	struct wmi_vdev_delete_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_delete_cmd *)skb->data;
+	cmd->vdev_id = __cpu_to_le32(vdev_id);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "WMI vdev delete id %d\n", vdev_id);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_DELETE_CMDID);
+}
+
+static int ath10k_wmi_vdev_start_restart(struct ath10k *ar,
+				const struct wmi_vdev_start_request_arg *arg,
+				enum wmi_cmd_id cmd_id)
+{
+	struct wmi_vdev_start_request_cmd *cmd;
+	struct sk_buff *skb;
+	const char *cmdname;
+	u32 flags = 0;
+
+	if (cmd_id != WMI_VDEV_START_REQUEST_CMDID &&
+	    cmd_id != WMI_VDEV_RESTART_REQUEST_CMDID)
+		return -EINVAL;
+	if (WARN_ON(arg->ssid && arg->ssid_len == 0))
+		return -EINVAL;
+	if (WARN_ON(arg->hidden_ssid && !arg->ssid))
+		return -EINVAL;
+	if (WARN_ON(arg->ssid_len > sizeof(cmd->ssid.ssid)))
+		return -EINVAL;
+
+	if (cmd_id == WMI_VDEV_START_REQUEST_CMDID)
+		cmdname = "start";
+	else if (cmd_id == WMI_VDEV_RESTART_REQUEST_CMDID)
+		cmdname = "restart";
+	else
+		return -EINVAL; /* should not happen, we already check cmd_id */
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	if (arg->hidden_ssid)
+		flags |= WMI_VDEV_START_HIDDEN_SSID;
+	if (arg->pmf_enabled)
+		flags |= WMI_VDEV_START_PMF_ENABLED;
+
+	cmd = (struct wmi_vdev_start_request_cmd *)skb->data;
+	cmd->vdev_id         = __cpu_to_le32(arg->vdev_id);
+	cmd->disable_hw_ack  = __cpu_to_le32(arg->disable_hw_ack);
+	cmd->beacon_interval = __cpu_to_le32(arg->bcn_intval);
+	cmd->dtim_period     = __cpu_to_le32(arg->dtim_period);
+	cmd->flags           = __cpu_to_le32(flags);
+	cmd->bcn_tx_rate     = __cpu_to_le32(arg->bcn_tx_rate);
+	cmd->bcn_tx_power    = __cpu_to_le32(arg->bcn_tx_power);
+
+	if (arg->ssid) {
+		cmd->ssid.ssid_len = __cpu_to_le32(arg->ssid_len);
+		memcpy(cmd->ssid.ssid, arg->ssid, arg->ssid_len);
+	}
+
+	cmd->chan.mhz = __cpu_to_le32(arg->channel.freq);
+
+	cmd->chan.band_center_freq1 =
+		__cpu_to_le32(arg->channel.band_center_freq1);
+
+	cmd->chan.mode = arg->channel.mode;
+	cmd->chan.min_power = arg->channel.min_power;
+	cmd->chan.max_power = arg->channel.max_power;
+	cmd->chan.reg_power = arg->channel.max_reg_power;
+	cmd->chan.reg_classid = arg->channel.reg_class_id;
+	cmd->chan.antenna_max = arg->channel.max_antenna_gain;
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi vdev %s id 0x%x freq %d, mode %d, ch_flags: 0x%0X,"
+		   "max_power: %d\n", cmdname, arg->vdev_id, arg->channel.freq,
+		   arg->channel.mode, flags, arg->channel.max_power);
+
+	return ath10k_wmi_cmd_send(ar, skb, cmd_id);
+}
+
+int ath10k_wmi_vdev_start(struct ath10k *ar,
+			  const struct wmi_vdev_start_request_arg *arg)
+{
+	return ath10k_wmi_vdev_start_restart(ar, arg,
+					     WMI_VDEV_START_REQUEST_CMDID);
+}
+
+int ath10k_wmi_vdev_restart(struct ath10k *ar,
+		     const struct wmi_vdev_start_request_arg *arg)
+{
+	return ath10k_wmi_vdev_start_restart(ar, arg,
+					     WMI_VDEV_RESTART_REQUEST_CMDID);
+}
+
+int ath10k_wmi_vdev_stop(struct ath10k *ar, u32 vdev_id)
+{
+	struct wmi_vdev_stop_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_stop_cmd *)skb->data;
+	cmd->vdev_id = __cpu_to_le32(vdev_id);
+
+	ath10k_dbg(ATH10K_DBG_WMI, "wmi vdev stop id 0x%x\n", vdev_id);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_STOP_CMDID);
+}
+
+int ath10k_wmi_vdev_up(struct ath10k *ar, u32 vdev_id, u32 aid, const u8 *bssid)
+{
+	struct wmi_vdev_up_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_up_cmd *)skb->data;
+	cmd->vdev_id       = __cpu_to_le32(vdev_id);
+	cmd->vdev_assoc_id = __cpu_to_le32(aid);
+	memcpy(&cmd->vdev_bssid.addr, bssid, 6);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi mgmt vdev up id 0x%x assoc id %d bssid %pM\n",
+		   vdev_id, aid, bssid);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_UP_CMDID);
+}
+
+int ath10k_wmi_vdev_down(struct ath10k *ar, u32 vdev_id)
+{
+	struct wmi_vdev_down_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_down_cmd *)skb->data;
+	cmd->vdev_id = __cpu_to_le32(vdev_id);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi mgmt vdev down id 0x%x\n", vdev_id);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_DOWN_CMDID);
+}
+
+int ath10k_wmi_vdev_set_param(struct ath10k *ar, u32 vdev_id,
+			      enum wmi_vdev_param param_id, u32 param_value)
+{
+	struct wmi_vdev_set_param_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_set_param_cmd *)skb->data;
+	cmd->vdev_id     = __cpu_to_le32(vdev_id);
+	cmd->param_id    = __cpu_to_le32(param_id);
+	cmd->param_value = __cpu_to_le32(param_value);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi vdev id 0x%x set param %d value %d\n",
+		   vdev_id, param_id, param_value);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_SET_PARAM_CMDID);
+}
+
+int ath10k_wmi_vdev_install_key(struct ath10k *ar,
+				const struct wmi_vdev_install_key_arg *arg)
+{
+	struct wmi_vdev_install_key_cmd *cmd;
+	struct sk_buff *skb;
+
+	if (arg->key_cipher == WMI_CIPHER_NONE && arg->key_data != NULL)
+		return -EINVAL;
+	if (arg->key_cipher != WMI_CIPHER_NONE && arg->key_data == NULL)
+		return -EINVAL;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd) + arg->key_len);
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_vdev_install_key_cmd *)skb->data;
+	cmd->vdev_id       = __cpu_to_le32(arg->vdev_id);
+	cmd->key_idx       = __cpu_to_le32(arg->key_idx);
+	cmd->key_flags     = __cpu_to_le32(arg->key_flags);
+	cmd->key_cipher    = __cpu_to_le32(arg->key_cipher);
+	cmd->key_len       = __cpu_to_le32(arg->key_len);
+	cmd->key_txmic_len = __cpu_to_le32(arg->key_txmic_len);
+	cmd->key_rxmic_len = __cpu_to_le32(arg->key_rxmic_len);
+
+	if (arg->macaddr)
+		memcpy(cmd->peer_macaddr.addr, arg->macaddr, ETH_ALEN);
+	if (arg->key_data)
+		memcpy(cmd->key_data, arg->key_data, arg->key_len);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_VDEV_INSTALL_KEY_CMDID);
+}
+
+int ath10k_wmi_peer_create(struct ath10k *ar, u32 vdev_id,
+			   const u8 peer_addr[ETH_ALEN])
+{
+	struct wmi_peer_create_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_peer_create_cmd *)skb->data;
+	cmd->vdev_id = __cpu_to_le32(vdev_id);
+	memcpy(cmd->peer_macaddr.addr, peer_addr, ETH_ALEN);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi peer create vdev_id %d peer_addr %pM\n",
+		   vdev_id, peer_addr);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PEER_CREATE_CMDID);
+}
+
+int ath10k_wmi_peer_delete(struct ath10k *ar, u32 vdev_id,
+			   const u8 peer_addr[ETH_ALEN])
+{
+	struct wmi_peer_delete_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_peer_delete_cmd *)skb->data;
+	cmd->vdev_id = __cpu_to_le32(vdev_id);
+	memcpy(cmd->peer_macaddr.addr, peer_addr, ETH_ALEN);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi peer delete vdev_id %d peer_addr %pM\n",
+		   vdev_id, peer_addr);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PEER_DELETE_CMDID);
+}
+
+int ath10k_wmi_peer_flush(struct ath10k *ar, u32 vdev_id,
+			  const u8 peer_addr[ETH_ALEN], u32 tid_bitmap)
+{
+	struct wmi_peer_flush_tids_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_peer_flush_tids_cmd *)skb->data;
+	cmd->vdev_id         = __cpu_to_le32(vdev_id);
+	cmd->peer_tid_bitmap = __cpu_to_le32(tid_bitmap);
+	memcpy(cmd->peer_macaddr.addr, peer_addr, ETH_ALEN);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi peer flush vdev_id %d peer_addr %pM tids %08x\n",
+		   vdev_id, peer_addr, tid_bitmap);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PEER_FLUSH_TIDS_CMDID);
+}
+
+int ath10k_wmi_peer_set_param(struct ath10k *ar, u32 vdev_id,
+			      const u8 *peer_addr, enum wmi_peer_param param_id,
+			      u32 param_value)
+{
+	struct wmi_peer_set_param_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_peer_set_param_cmd *)skb->data;
+	cmd->vdev_id     = __cpu_to_le32(vdev_id);
+	cmd->param_id    = __cpu_to_le32(param_id);
+	cmd->param_value = __cpu_to_le32(param_value);
+	memcpy(&cmd->peer_macaddr.addr, peer_addr, 6);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi vdev %d peer 0x%pM set param %d value %d\n",
+		   vdev_id, peer_addr, param_id, param_value);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PEER_SET_PARAM_CMDID);
+}
+
+int ath10k_wmi_set_psmode(struct ath10k *ar, u32 vdev_id,
+			  enum wmi_sta_ps_mode psmode)
+{
+	struct wmi_sta_powersave_mode_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_sta_powersave_mode_cmd *)skb->data;
+	cmd->vdev_id     = __cpu_to_le32(vdev_id);
+	cmd->sta_ps_mode = __cpu_to_le32(psmode);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi set powersave id 0x%x mode %d\n",
+		   vdev_id, psmode);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_STA_POWERSAVE_MODE_CMDID);
+}
+
+int ath10k_wmi_set_sta_ps_param(struct ath10k *ar, u32 vdev_id,
+				enum wmi_sta_powersave_param param_id,
+				u32 value)
+{
+	struct wmi_sta_powersave_param_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_sta_powersave_param_cmd *)skb->data;
+	cmd->vdev_id     = __cpu_to_le32(vdev_id);
+	cmd->param_id    = __cpu_to_le32(param_id);
+	cmd->param_value = __cpu_to_le32(value);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi sta ps param vdev_id 0x%x param %d value %d\n",
+		   vdev_id, param_id, value);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_STA_POWERSAVE_PARAM_CMDID);
+}
+
+int ath10k_wmi_set_ap_ps_param(struct ath10k *ar, u32 vdev_id, const u8 *mac,
+			       enum wmi_ap_ps_peer_param param_id, u32 value)
+{
+	struct wmi_ap_ps_peer_cmd *cmd;
+	struct sk_buff *skb;
+
+	if (!mac)
+		return -EINVAL;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_ap_ps_peer_cmd *)skb->data;
+	cmd->vdev_id = __cpu_to_le32(vdev_id);
+	cmd->param_id = __cpu_to_le32(param_id);
+	cmd->param_value = __cpu_to_le32(value);
+	memcpy(&cmd->peer_macaddr, mac, ETH_ALEN);
+
+	ath10k_dbg(ATH10K_DBG_WMI,
+		   "wmi ap ps param vdev_id 0x%X param %d value %d mac_addr %pM\n",
+		   vdev_id, param_id, value, mac);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_AP_PS_PEER_PARAM_CMDID);
+}
+
+int ath10k_wmi_scan_chan_list(struct ath10k *ar,
+			      const struct wmi_scan_chan_list_arg *arg)
+{
+	struct wmi_scan_chan_list_cmd *cmd;
+	struct sk_buff *skb;
+	struct wmi_channel_arg *ch;
+	struct wmi_channel *ci;
+	int len;
+	int i;
+
+	len = sizeof(*cmd) + arg->n_channels * sizeof(struct wmi_channel);
+
+	skb = ath10k_wmi_alloc_skb(len);
+	if (!skb)
+		return -EINVAL;
+
+	cmd = (struct wmi_scan_chan_list_cmd *)skb->data;
+	cmd->num_scan_chans = __cpu_to_le32(arg->n_channels);
+
+	for (i = 0; i < arg->n_channels; i++) {
+		u32 flags = 0;
+
+		ch = &arg->channels[i];
+		ci = &cmd->chan_info[i];
+
+		if (ch->passive)
+			flags |= WMI_CHAN_FLAG_PASSIVE;
+		if (ch->allow_ibss)
+			flags |= WMI_CHAN_FLAG_ADHOC_ALLOWED;
+		if (ch->allow_ht)
+			flags |= WMI_CHAN_FLAG_ALLOW_HT;
+		if (ch->allow_vht)
+			flags |= WMI_CHAN_FLAG_ALLOW_VHT;
+		if (ch->ht40plus)
+			flags |= WMI_CHAN_FLAG_HT40_PLUS;
+
+		ci->mhz               = __cpu_to_le32(ch->freq);
+		ci->band_center_freq1 = __cpu_to_le32(ch->freq);
+		ci->band_center_freq2 = 0;
+		ci->min_power         = ch->min_power;
+		ci->max_power         = ch->max_power;
+		ci->reg_power         = ch->max_reg_power;
+		ci->antenna_max       = ch->max_antenna_gain;
+		ci->antenna_max       = 0;
+
+		/* mode & flags share storage */
+		ci->mode              = ch->mode;
+		ci->flags            |= __cpu_to_le32(flags);
+	}
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_SCAN_CHAN_LIST_CMDID);
+}
+
+int ath10k_wmi_peer_assoc(struct ath10k *ar,
+			  const struct wmi_peer_assoc_complete_arg *arg)
+{
+	struct wmi_peer_assoc_complete_cmd *cmd;
+	struct sk_buff *skb;
+
+	if (arg->peer_mpdu_density > 16)
+		return -EINVAL;
+	if (arg->peer_legacy_rates.num_rates > MAX_SUPPORTED_RATES)
+		return -EINVAL;
+	if (arg->peer_ht_rates.num_rates > MAX_SUPPORTED_RATES)
+		return -EINVAL;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_peer_assoc_complete_cmd *)skb->data;
+	cmd->vdev_id            = __cpu_to_le32(arg->vdev_id);
+	cmd->peer_new_assoc     = __cpu_to_le32(arg->peer_reassoc ? 0 : 1);
+	cmd->peer_associd       = __cpu_to_le32(arg->peer_aid);
+	cmd->peer_flags         = __cpu_to_le32(arg->peer_flags);
+	cmd->peer_caps          = __cpu_to_le32(arg->peer_caps);
+	cmd->peer_listen_intval = __cpu_to_le32(arg->peer_listen_intval);
+	cmd->peer_ht_caps       = __cpu_to_le32(arg->peer_ht_caps);
+	cmd->peer_max_mpdu      = __cpu_to_le32(arg->peer_max_mpdu);
+	cmd->peer_mpdu_density  = __cpu_to_le32(arg->peer_mpdu_density);
+	cmd->peer_rate_caps     = __cpu_to_le32(arg->peer_rate_caps);
+	cmd->peer_nss           = __cpu_to_le32(arg->peer_num_spatial_streams);
+	cmd->peer_vht_caps      = __cpu_to_le32(arg->peer_vht_caps);
+	cmd->peer_phymode       = __cpu_to_le32(arg->peer_phymode);
+
+	memcpy(cmd->peer_macaddr.addr, arg->addr, ETH_ALEN);
+
+	cmd->peer_legacy_rates.num_rates =
+		__cpu_to_le32(arg->peer_legacy_rates.num_rates);
+	memcpy(cmd->peer_legacy_rates.rates, arg->peer_legacy_rates.rates,
+	       arg->peer_legacy_rates.num_rates);
+
+	cmd->peer_ht_rates.num_rates =
+		__cpu_to_le32(arg->peer_ht_rates.num_rates);
+	memcpy(cmd->peer_ht_rates.rates, arg->peer_ht_rates.rates,
+	       arg->peer_ht_rates.num_rates);
+
+	cmd->peer_vht_rates.rx_max_rate =
+		__cpu_to_le32(arg->peer_vht_rates.rx_max_rate);
+	cmd->peer_vht_rates.rx_mcs_set =
+		__cpu_to_le32(arg->peer_vht_rates.rx_mcs_set);
+	cmd->peer_vht_rates.tx_max_rate =
+		__cpu_to_le32(arg->peer_vht_rates.tx_max_rate);
+	cmd->peer_vht_rates.tx_mcs_set =
+		__cpu_to_le32(arg->peer_vht_rates.tx_mcs_set);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PEER_ASSOC_CMDID);
+}
+
+int ath10k_wmi_beacon_send(struct ath10k *ar, const struct wmi_bcn_tx_arg *arg)
+{
+	struct wmi_bcn_tx_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd) + arg->bcn_len);
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_bcn_tx_cmd *)skb->data;
+	cmd->hdr.vdev_id  = __cpu_to_le32(arg->vdev_id);
+	cmd->hdr.tx_rate  = __cpu_to_le32(arg->tx_rate);
+	cmd->hdr.tx_power = __cpu_to_le32(arg->tx_power);
+	cmd->hdr.bcn_len  = __cpu_to_le32(arg->bcn_len);
+	memcpy(cmd->bcn, arg->bcn, arg->bcn_len);
+
+	return ath10k_wmi_cmd_send(ar, skb, WMI_BCN_TX_CMDID);
+}
+
+static void ath10k_wmi_pdev_set_wmm_param(struct wmi_wmm_params *params,
+					  const struct wmi_wmm_params_arg *arg)
+{
+	params->cwmin  = __cpu_to_le32(arg->cwmin);
+	params->cwmax  = __cpu_to_le32(arg->cwmax);
+	params->aifs   = __cpu_to_le32(arg->aifs);
+	params->txop   = __cpu_to_le32(arg->txop);
+	params->acm    = __cpu_to_le32(arg->acm);
+	params->no_ack = __cpu_to_le32(arg->no_ack);
+}
+
+int ath10k_wmi_pdev_set_wmm_params(struct ath10k *ar,
+			const struct wmi_pdev_set_wmm_params_arg *arg)
+{
+	struct wmi_pdev_set_wmm_params *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_pdev_set_wmm_params *)skb->data;
+	ath10k_wmi_pdev_set_wmm_param(&cmd->ac_be, &arg->ac_be);
+	ath10k_wmi_pdev_set_wmm_param(&cmd->ac_bk, &arg->ac_bk);
+	ath10k_wmi_pdev_set_wmm_param(&cmd->ac_vi, &arg->ac_vi);
+	ath10k_wmi_pdev_set_wmm_param(&cmd->ac_vo, &arg->ac_vo);
+
+	ath10k_dbg(ATH10K_DBG_WMI, "wmi pdev set wmm params\n");
+	return ath10k_wmi_cmd_send(ar, skb, WMI_PDEV_SET_WMM_PARAMS_CMDID);
+}
+
+int ath10k_wmi_request_stats(struct ath10k *ar, enum wmi_stats_id stats_id)
+{
+	struct wmi_request_stats_cmd *cmd;
+	struct sk_buff *skb;
+
+	skb = ath10k_wmi_alloc_skb(sizeof(*cmd));
+	if (!skb)
+		return -ENOMEM;
+
+	cmd = (struct wmi_request_stats_cmd *)skb->data;
+	cmd->stats_id = __cpu_to_le32(stats_id);
+
+	ath10k_dbg(ATH10K_DBG_WMI, "wmi request stats %d\n", (int)stats_id);
+	return ath10k_wmi_cmd_send(ar, skb, WMI_REQUEST_STATS_CMDID);
+}
diff --git a/drivers/net/wireless/ath/ath10k/wmi.h b/drivers/net/wireless/ath/ath10k/wmi.h
new file mode 100644
index 0000000..9555f5a
--- /dev/null
+++ b/drivers/net/wireless/ath/ath10k/wmi.h
@@ -0,0 +1,3052 @@
+/*
+ * Copyright (c) 2005-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef _WMI_H_
+#define _WMI_H_
+
+#include <linux/types.h>
+#include <net/mac80211.h>
+
+/*
+ * This file specifies the WMI interface for the Unified Software
+ * Architecture.
+ *
+ * It includes definitions of all the commands and events. Commands are
+ * messages from the host to the target. Events and Replies are messages
+ * from the target to the host.
+ *
+ * Ownership of correctness in regards to WMI commands belongs to the host
+ * driver and the target is not required to validate parameters for value,
+ * proper range, or any other checking.
+ *
+ * Guidelines for extending this interface are below.
+ *
+ * 1. Add new WMI commands ONLY within the specified range - 0x9000 - 0x9fff
+ *
+ * 2. Use ONLY u32 type for defining member variables within WMI
+ *    command/event structures. Do not use u8, u16, bool or
+ *    enum types within these structures.
+ *
+ * 3. DO NOT define bit fields within structures. Implement bit fields
+ *    using masks if necessary. Do not use the programming language's bit
+ *    field definition.
+ *
+ * 4. Define macros for encode/decode of u8, u16 fields within
+ *    the u32 variables. Use these macros for set/get of these fields.
+ *    Try to use this to optimize the structure without bloating it with
+ *    u32 variables for every lower sized field.
+ *
+ * 5. Do not use PACK/UNPACK attributes for the structures as each member
+ *    variable is already 4-byte aligned by virtue of being a u32
+ *    type.
+ *
+ * 6. Comment each parameter part of the WMI command/event structure by
+ *    using the 2 stars at the begining of C comment instead of one star to
+ *    enable HTML document generation using Doxygen.
+ *
+ */
+
+/* Control Path */
+struct wmi_cmd_hdr {
+	__le32 cmd_id;
+} __packed;
+
+#define WMI_CMD_HDR_CMD_ID_MASK   0x00FFFFFF
+#define WMI_CMD_HDR_CMD_ID_LSB    0
+#define WMI_CMD_HDR_PLT_PRIV_MASK 0xFF000000
+#define WMI_CMD_HDR_PLT_PRIV_LSB  24
+
+#define HTC_PROTOCOL_VERSION    0x0002
+#define WMI_PROTOCOL_VERSION    0x0002
+
+enum wmi_service_id {
+	WMI_SERVICE_BEACON_OFFLOAD = 0,   /* beacon offload */
+	WMI_SERVICE_SCAN_OFFLOAD,	  /* scan offload */
+	WMI_SERVICE_ROAM_OFFLOAD,	  /* roam offload */
+	WMI_SERVICE_BCN_MISS_OFFLOAD,     /* beacon miss offload */
+	WMI_SERVICE_STA_PWRSAVE,	  /* fake sleep + basic power save */
+	WMI_SERVICE_STA_ADVANCED_PWRSAVE, /* uapsd, pspoll, force sleep */
+	WMI_SERVICE_AP_UAPSD,		  /* uapsd on AP */
+	WMI_SERVICE_AP_DFS,		  /* DFS on AP */
+	WMI_SERVICE_11AC,		  /* supports 11ac */
+	WMI_SERVICE_BLOCKACK,	/* Supports triggering ADDBA/DELBA from host*/
+	WMI_SERVICE_PHYERR,		  /* PHY error */
+	WMI_SERVICE_BCN_FILTER,		  /* Beacon filter support */
+	WMI_SERVICE_RTT,		  /* RTT (round trip time) support */
+	WMI_SERVICE_RATECTRL,		  /* Rate-control */
+	WMI_SERVICE_WOW,		  /* WOW Support */
+	WMI_SERVICE_RATECTRL_CACHE,       /* Rate-control caching */
+	WMI_SERVICE_IRAM_TIDS,            /* TIDs in IRAM */
+	WMI_SERVICE_ARPNS_OFFLOAD,	  /* ARP NS Offload support */
+	WMI_SERVICE_NLO,		  /* Network list offload service */
+	WMI_SERVICE_GTK_OFFLOAD,	  /* GTK offload */
+	WMI_SERVICE_SCAN_SCH,		  /* Scan Scheduler Service */
+	WMI_SERVICE_CSA_OFFLOAD,	  /* CSA offload service */
+	WMI_SERVICE_CHATTER,		  /* Chatter service */
+	WMI_SERVICE_COEX_FREQAVOID,	  /* FW report freq range to avoid */
+	WMI_SERVICE_PACKET_POWER_SAVE,	  /* packet power save service */
+	WMI_SERVICE_FORCE_FW_HANG,        /* To test fw recovery mechanism */
+	WMI_SERVICE_GPIO,                 /* GPIO service */
+	WMI_SERVICE_STA_DTIM_PS_MODULATED_DTIM, /* Modulated DTIM support */
+	WMI_STA_UAPSD_BASIC_AUTO_TRIG,    /* UAPSD AC Trigger Generation  */
+	WMI_STA_UAPSD_VAR_AUTO_TRIG,      /* -do- */
+	WMI_SERVICE_STA_KEEP_ALIVE,       /* STA keep alive mechanism support */
+	WMI_SERVICE_TX_ENCAP,             /* Packet type for TX encapsulation */
+
+	WMI_SERVICE_LAST,
+	WMI_MAX_SERVICE = 64		  /* max service */
+};
+
+static inline char *wmi_service_name(int service_id)
+{
+	switch (service_id) {
+	case WMI_SERVICE_BEACON_OFFLOAD:
+		return "BEACON_OFFLOAD";
+	case WMI_SERVICE_SCAN_OFFLOAD:
+		return "SCAN_OFFLOAD";
+	case WMI_SERVICE_ROAM_OFFLOAD:
+		return "ROAM_OFFLOAD";
+	case WMI_SERVICE_BCN_MISS_OFFLOAD:
+		return "BCN_MISS_OFFLOAD";
+	case WMI_SERVICE_STA_PWRSAVE:
+		return "STA_PWRSAVE";
+	case WMI_SERVICE_STA_ADVANCED_PWRSAVE:
+		return "STA_ADVANCED_PWRSAVE";
+	case WMI_SERVICE_AP_UAPSD:
+		return "AP_UAPSD";
+	case WMI_SERVICE_AP_DFS:
+		return "AP_DFS";
+	case WMI_SERVICE_11AC:
+		return "11AC";
+	case WMI_SERVICE_BLOCKACK:
+		return "BLOCKACK";
+	case WMI_SERVICE_PHYERR:
+		return "PHYERR";
+	case WMI_SERVICE_BCN_FILTER:
+		return "BCN_FILTER";
+	case WMI_SERVICE_RTT:
+		return "RTT";
+	case WMI_SERVICE_RATECTRL:
+		return "RATECTRL";
+	case WMI_SERVICE_WOW:
+		return "WOW";
+	case WMI_SERVICE_RATECTRL_CACHE:
+		return "RATECTRL CACHE";
+	case WMI_SERVICE_IRAM_TIDS:
+		return "IRAM TIDS";
+	case WMI_SERVICE_ARPNS_OFFLOAD:
+		return "ARPNS_OFFLOAD";
+	case WMI_SERVICE_NLO:
+		return "NLO";
+	case WMI_SERVICE_GTK_OFFLOAD:
+		return "GTK_OFFLOAD";
+	case WMI_SERVICE_SCAN_SCH:
+		return "SCAN_SCH";
+	case WMI_SERVICE_CSA_OFFLOAD:
+		return "CSA_OFFLOAD";
+	case WMI_SERVICE_CHATTER:
+		return "CHATTER";
+	case WMI_SERVICE_COEX_FREQAVOID:
+		return "COEX_FREQAVOID";
+	case WMI_SERVICE_PACKET_POWER_SAVE:
+		return "PACKET_POWER_SAVE";
+	case WMI_SERVICE_FORCE_FW_HANG:
+		return "FORCE FW HANG";
+	case WMI_SERVICE_GPIO:
+		return "GPIO";
+	case WMI_SERVICE_STA_DTIM_PS_MODULATED_DTIM:
+		return "MODULATED DTIM";
+	case WMI_STA_UAPSD_BASIC_AUTO_TRIG:
+		return "BASIC UAPSD";
+	case WMI_STA_UAPSD_VAR_AUTO_TRIG:
+		return "VAR UAPSD";
+	case WMI_SERVICE_STA_KEEP_ALIVE:
+		return "STA KEEP ALIVE";
+	case WMI_SERVICE_TX_ENCAP:
+		return "TX ENCAP";
+	default:
+		return "UNKNOWN SERVICE\n";
+	}
+}
+
+
+#define WMI_SERVICE_BM_SIZE \
+	((WMI_MAX_SERVICE + sizeof(u32) - 1)/sizeof(u32))
+
+/* 2 word representation of MAC addr */
+struct wmi_mac_addr {
+	union {
+		u8 addr[6];
+		struct {
+			u32 word0;
+			u32 word1;
+		} __packed;
+	} __packed;
+} __packed;
+
+/* macro to convert MAC address from WMI word format to char array */
+#define WMI_MAC_ADDR_TO_CHAR_ARRAY(pwmi_mac_addr, c_macaddr) do { \
+	(c_macaddr)[0] =  ((pwmi_mac_addr)->word0) & 0xff; \
+	(c_macaddr)[1] = (((pwmi_mac_addr)->word0) >> 8) & 0xff; \
+	(c_macaddr)[2] = (((pwmi_mac_addr)->word0) >> 16) & 0xff; \
+	(c_macaddr)[3] = (((pwmi_mac_addr)->word0) >> 24) & 0xff; \
+	(c_macaddr)[4] =  ((pwmi_mac_addr)->word1) & 0xff; \
+	(c_macaddr)[5] = (((pwmi_mac_addr)->word1) >> 8) & 0xff; \
+	} while (0)
+
+/*
+ * wmi command groups.
+ */
+enum wmi_cmd_group {
+	/* 0 to 2 are reserved */
+	WMI_GRP_START = 0x3,
+	WMI_GRP_SCAN = WMI_GRP_START,
+	WMI_GRP_PDEV,
+	WMI_GRP_VDEV,
+	WMI_GRP_PEER,
+	WMI_GRP_MGMT,
+	WMI_GRP_BA_NEG,
+	WMI_GRP_STA_PS,
+	WMI_GRP_DFS,
+	WMI_GRP_ROAM,
+	WMI_GRP_OFL_SCAN,
+	WMI_GRP_P2P,
+	WMI_GRP_AP_PS,
+	WMI_GRP_RATE_CTRL,
+	WMI_GRP_PROFILE,
+	WMI_GRP_SUSPEND,
+	WMI_GRP_BCN_FILTER,
+	WMI_GRP_WOW,
+	WMI_GRP_RTT,
+	WMI_GRP_SPECTRAL,
+	WMI_GRP_STATS,
+	WMI_GRP_ARP_NS_OFL,
+	WMI_GRP_NLO_OFL,
+	WMI_GRP_GTK_OFL,
+	WMI_GRP_CSA_OFL,
+	WMI_GRP_CHATTER,
+	WMI_GRP_TID_ADDBA,
+	WMI_GRP_MISC,
+	WMI_GRP_GPIO,
+};
+
+#define WMI_CMD_GRP(grp_id) (((grp_id) << 12) | 0x1)
+#define WMI_EVT_GRP_START_ID(grp_id) (((grp_id) << 12) | 0x1)
+
+/* Command IDs and commande events. */
+enum wmi_cmd_id {
+	WMI_INIT_CMDID = 0x1,
+
+	/* Scan specific commands */
+	WMI_START_SCAN_CMDID = WMI_CMD_GRP(WMI_GRP_SCAN),
+	WMI_STOP_SCAN_CMDID,
+	WMI_SCAN_CHAN_LIST_CMDID,
+	WMI_SCAN_SCH_PRIO_TBL_CMDID,
+
+	/* PDEV (physical device) specific commands */
+	WMI_PDEV_SET_REGDOMAIN_CMDID = WMI_CMD_GRP(WMI_GRP_PDEV),
+	WMI_PDEV_SET_CHANNEL_CMDID,
+	WMI_PDEV_SET_PARAM_CMDID,
+	WMI_PDEV_PKTLOG_ENABLE_CMDID,
+	WMI_PDEV_PKTLOG_DISABLE_CMDID,
+	WMI_PDEV_SET_WMM_PARAMS_CMDID,
+	WMI_PDEV_SET_HT_CAP_IE_CMDID,
+	WMI_PDEV_SET_VHT_CAP_IE_CMDID,
+	WMI_PDEV_SET_DSCP_TID_MAP_CMDID,
+	WMI_PDEV_SET_QUIET_MODE_CMDID,
+	WMI_PDEV_GREEN_AP_PS_ENABLE_CMDID,
+	WMI_PDEV_GET_TPC_CONFIG_CMDID,
+	WMI_PDEV_SET_BASE_MACADDR_CMDID,
+
+	/* VDEV (virtual device) specific commands */
+	WMI_VDEV_CREATE_CMDID = WMI_CMD_GRP(WMI_GRP_VDEV),
+	WMI_VDEV_DELETE_CMDID,
+	WMI_VDEV_START_REQUEST_CMDID,
+	WMI_VDEV_RESTART_REQUEST_CMDID,
+	WMI_VDEV_UP_CMDID,
+	WMI_VDEV_STOP_CMDID,
+	WMI_VDEV_DOWN_CMDID,
+	WMI_VDEV_SET_PARAM_CMDID,
+	WMI_VDEV_INSTALL_KEY_CMDID,
+
+	/* peer specific commands */
+	WMI_PEER_CREATE_CMDID = WMI_CMD_GRP(WMI_GRP_PEER),
+	WMI_PEER_DELETE_CMDID,
+	WMI_PEER_FLUSH_TIDS_CMDID,
+	WMI_PEER_SET_PARAM_CMDID,
+	WMI_PEER_ASSOC_CMDID,
+	WMI_PEER_ADD_WDS_ENTRY_CMDID,
+	WMI_PEER_REMOVE_WDS_ENTRY_CMDID,
+	WMI_PEER_MCAST_GROUP_CMDID,
+
+	/* beacon/management specific commands */
+	WMI_BCN_TX_CMDID = WMI_CMD_GRP(WMI_GRP_MGMT),
+	WMI_PDEV_SEND_BCN_CMDID,
+	WMI_BCN_TMPL_CMDID,
+	WMI_BCN_FILTER_RX_CMDID,
+	WMI_PRB_REQ_FILTER_RX_CMDID,
+	WMI_MGMT_TX_CMDID,
+	WMI_PRB_TMPL_CMDID,
+
+	/* commands to directly control BA negotiation directly from host. */
+	WMI_ADDBA_CLEAR_RESP_CMDID = WMI_CMD_GRP(WMI_GRP_BA_NEG),
+	WMI_ADDBA_SEND_CMDID,
+	WMI_ADDBA_STATUS_CMDID,
+	WMI_DELBA_SEND_CMDID,
+	WMI_ADDBA_SET_RESP_CMDID,
+	WMI_SEND_SINGLEAMSDU_CMDID,
+
+	/* Station power save specific config */
+	WMI_STA_POWERSAVE_MODE_CMDID = WMI_CMD_GRP(WMI_GRP_STA_PS),
+	WMI_STA_POWERSAVE_PARAM_CMDID,
+	WMI_STA_MIMO_PS_MODE_CMDID,
+
+	/** DFS-specific commands */
+	WMI_PDEV_DFS_ENABLE_CMDID = WMI_CMD_GRP(WMI_GRP_DFS),
+	WMI_PDEV_DFS_DISABLE_CMDID,
+
+	/* Roaming specific  commands */
+	WMI_ROAM_SCAN_MODE = WMI_CMD_GRP(WMI_GRP_ROAM),
+	WMI_ROAM_SCAN_RSSI_THRESHOLD,
+	WMI_ROAM_SCAN_PERIOD,
+	WMI_ROAM_SCAN_RSSI_CHANGE_THRESHOLD,
+	WMI_ROAM_AP_PROFILE,
+
+	/* offload scan specific commands */
+	WMI_OFL_SCAN_ADD_AP_PROFILE = WMI_CMD_GRP(WMI_GRP_OFL_SCAN),
+	WMI_OFL_SCAN_REMOVE_AP_PROFILE,
+	WMI_OFL_SCAN_PERIOD,
+
+	/* P2P specific commands */
+	WMI_P2P_DEV_SET_DEVICE_INFO = WMI_CMD_GRP(WMI_GRP_P2P),
+	WMI_P2P_DEV_SET_DISCOVERABILITY,
+	WMI_P2P_GO_SET_BEACON_IE,
+	WMI_P2P_GO_SET_PROBE_RESP_IE,
+	WMI_P2P_SET_VENDOR_IE_DATA_CMDID,
+
+	/* AP power save specific config */
+	WMI_AP_PS_PEER_PARAM_CMDID = WMI_CMD_GRP(WMI_GRP_AP_PS),
+	WMI_AP_PS_PEER_UAPSD_COEX_CMDID,
+
+	/* Rate-control specific commands */
+	WMI_PEER_RATE_RETRY_SCHED_CMDID =
+	WMI_CMD_GRP(WMI_GRP_RATE_CTRL),
+
+	/* WLAN Profiling commands. */
+	WMI_WLAN_PROFILE_TRIGGER_CMDID = WMI_CMD_GRP(WMI_GRP_PROFILE),
+	WMI_WLAN_PROFILE_SET_HIST_INTVL_CMDID,
+	WMI_WLAN_PROFILE_GET_PROFILE_DATA_CMDID,
+	WMI_WLAN_PROFILE_ENABLE_PROFILE_ID_CMDID,
+	WMI_WLAN_PROFILE_LIST_PROFILE_ID_CMDID,
+
+	/* Suspend resume command Ids */
+	WMI_PDEV_SUSPEND_CMDID = WMI_CMD_GRP(WMI_GRP_SUSPEND),
+	WMI_PDEV_RESUME_CMDID,
+
+	/* Beacon filter commands */
+	WMI_ADD_BCN_FILTER_CMDID = WMI_CMD_GRP(WMI_GRP_BCN_FILTER),
+	WMI_RMV_BCN_FILTER_CMDID,
+
+	/* WOW Specific WMI commands*/
+	WMI_WOW_ADD_WAKE_PATTERN_CMDID = WMI_CMD_GRP(WMI_GRP_WOW),
+	WMI_WOW_DEL_WAKE_PATTERN_CMDID,
+	WMI_WOW_ENABLE_DISABLE_WAKE_EVENT_CMDID,
+	WMI_WOW_ENABLE_CMDID,
+	WMI_WOW_HOSTWAKEUP_FROM_SLEEP_CMDID,
+
+	/* RTT measurement related cmd */
+	WMI_RTT_MEASREQ_CMDID = WMI_CMD_GRP(WMI_GRP_RTT),
+	WMI_RTT_TSF_CMDID,
+
+	/* spectral scan commands */
+	WMI_VDEV_SPECTRAL_SCAN_CONFIGURE_CMDID = WMI_CMD_GRP(WMI_GRP_SPECTRAL),
+	WMI_VDEV_SPECTRAL_SCAN_ENABLE_CMDID,
+
+	/* F/W stats */
+	WMI_REQUEST_STATS_CMDID = WMI_CMD_GRP(WMI_GRP_STATS),
+
+	/* ARP OFFLOAD REQUEST*/
+	WMI_SET_ARP_NS_OFFLOAD_CMDID = WMI_CMD_GRP(WMI_GRP_ARP_NS_OFL),
+
+	/* NS offload confid*/
+	WMI_NETWORK_LIST_OFFLOAD_CONFIG_CMDID = WMI_CMD_GRP(WMI_GRP_NLO_OFL),
+
+	/* GTK offload Specific WMI commands*/
+	WMI_GTK_OFFLOAD_CMDID = WMI_CMD_GRP(WMI_GRP_GTK_OFL),
+
+	/* CSA offload Specific WMI commands*/
+	WMI_CSA_OFFLOAD_ENABLE_CMDID = WMI_CMD_GRP(WMI_GRP_CSA_OFL),
+	WMI_CSA_OFFLOAD_CHANSWITCH_CMDID,
+
+	/* Chatter commands*/
+	WMI_CHATTER_SET_MODE_CMDID = WMI_CMD_GRP(WMI_GRP_CHATTER),
+
+	/* addba specific commands */
+	WMI_PEER_TID_ADDBA_CMDID = WMI_CMD_GRP(WMI_GRP_TID_ADDBA),
+	WMI_PEER_TID_DELBA_CMDID,
+
+	/* set station mimo powersave method */
+	WMI_STA_DTIM_PS_METHOD_CMDID,
+	/* Configure the Station UAPSD AC Auto Trigger Parameters */
+	WMI_STA_UAPSD_AUTO_TRIG_CMDID,
+
+	/* STA Keep alive parameter configuration,
+	   Requires WMI_SERVICE_STA_KEEP_ALIVE */
+	WMI_STA_KEEPALIVE_CMD,
+
+	/* misc command group */
+	WMI_ECHO_CMDID = WMI_CMD_GRP(WMI_GRP_MISC),
+	WMI_PDEV_UTF_CMDID,
+	WMI_DBGLOG_CFG_CMDID,
+	WMI_PDEV_QVIT_CMDID,
+	WMI_PDEV_FTM_INTG_CMDID,
+	WMI_VDEV_SET_KEEPALIVE_CMDID,
+	WMI_VDEV_GET_KEEPALIVE_CMDID,
+
+	/* GPIO Configuration */
+	WMI_GPIO_CONFIG_CMDID = WMI_CMD_GRP(WMI_GRP_GPIO),
+	WMI_GPIO_OUTPUT_CMDID,
+};
+
+enum wmi_event_id {
+	WMI_SERVICE_READY_EVENTID = 0x1,
+	WMI_READY_EVENTID,
+
+	/* Scan specific events */
+	WMI_SCAN_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_SCAN),
+
+	/* PDEV specific events */
+	WMI_PDEV_TPC_CONFIG_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_PDEV),
+	WMI_CHAN_INFO_EVENTID,
+	WMI_PHYERR_EVENTID,
+
+	/* VDEV specific events */
+	WMI_VDEV_START_RESP_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_VDEV),
+	WMI_VDEV_STOPPED_EVENTID,
+	WMI_VDEV_INSTALL_KEY_COMPLETE_EVENTID,
+
+	/* peer specific events */
+	WMI_PEER_STA_KICKOUT_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_PEER),
+
+	/* beacon/mgmt specific events */
+	WMI_MGMT_RX_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_MGMT),
+	WMI_HOST_SWBA_EVENTID,
+	WMI_TBTTOFFSET_UPDATE_EVENTID,
+
+	/* ADDBA Related WMI Events*/
+	WMI_TX_DELBA_COMPLETE_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_BA_NEG),
+	WMI_TX_ADDBA_COMPLETE_EVENTID,
+
+	/* Roam event to trigger roaming on host */
+	WMI_ROAM_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_ROAM),
+	WMI_PROFILE_MATCH,
+
+	/* WoW */
+	WMI_WOW_WAKEUP_HOST_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_WOW),
+
+	/* RTT */
+	WMI_RTT_MEASUREMENT_REPORT_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_RTT),
+	WMI_TSF_MEASUREMENT_REPORT_EVENTID,
+	WMI_RTT_ERROR_REPORT_EVENTID,
+
+	/* GTK offload */
+	WMI_GTK_OFFLOAD_STATUS_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_GTK_OFL),
+	WMI_GTK_REKEY_FAIL_EVENTID,
+
+	/* CSA IE received event */
+	WMI_CSA_HANDLING_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_CSA_OFL),
+
+	/* Misc events */
+	WMI_ECHO_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_MISC),
+	WMI_PDEV_UTF_EVENTID,
+	WMI_DEBUG_MESG_EVENTID,
+	WMI_UPDATE_STATS_EVENTID,
+	WMI_DEBUG_PRINT_EVENTID,
+	WMI_DCS_INTERFERENCE_EVENTID,
+	WMI_PDEV_QVIT_EVENTID,
+	WMI_WLAN_PROFILE_DATA_EVENTID,
+	WMI_PDEV_FTM_INTG_EVENTID,
+	WMI_WLAN_FREQ_AVOID_EVENTID,
+	WMI_VDEV_GET_KEEPALIVE_EVENTID,
+
+	/* GPIO Event */
+	WMI_GPIO_INPUT_EVENTID = WMI_EVT_GRP_START_ID(WMI_GRP_GPIO),
+};
+
+enum wmi_phy_mode {
+	MODE_11A        = 0,   /* 11a Mode */
+	MODE_11G        = 1,   /* 11b/g Mode */
+	MODE_11B        = 2,   /* 11b Mode */
+	MODE_11GONLY    = 3,   /* 11g only Mode */
+	MODE_11NA_HT20   = 4,  /* 11a HT20 mode */
+	MODE_11NG_HT20   = 5,  /* 11g HT20 mode */
+	MODE_11NA_HT40   = 6,  /* 11a HT40 mode */
+	MODE_11NG_HT40   = 7,  /* 11g HT40 mode */
+	MODE_11AC_VHT20 = 8,
+	MODE_11AC_VHT40 = 9,
+	MODE_11AC_VHT80 = 10,
+	/*    MODE_11AC_VHT160 = 11, */
+	MODE_11AC_VHT20_2G = 11,
+	MODE_11AC_VHT40_2G = 12,
+	MODE_11AC_VHT80_2G = 13,
+	MODE_UNKNOWN    = 14,
+	MODE_MAX        = 14
+};
+
+#define WMI_CHAN_LIST_TAG	0x1
+#define WMI_SSID_LIST_TAG	0x2
+#define WMI_BSSID_LIST_TAG	0x3
+#define WMI_IE_TAG		0x4
+
+struct wmi_channel {
+	__le32 mhz;
+	__le32 band_center_freq1;
+	__le32 band_center_freq2; /* valid for 11ac, 80plus80 */
+	union {
+		__le32 flags; /* WMI_CHAN_FLAG_ */
+		struct {
+			u8 mode; /* only 6 LSBs */
+		} __packed;
+	} __packed;
+	union {
+		__le32 reginfo0;
+		struct {
+			u8 min_power;
+			u8 max_power;
+			u8 reg_power;
+			u8 reg_classid;
+		} __packed;
+	} __packed;
+	union {
+		__le32 reginfo1;
+		struct {
+			u8 antenna_max;
+		} __packed;
+	} __packed;
+} __packed;
+
+struct wmi_channel_arg {
+	u32 freq;
+	u32 band_center_freq1;
+	bool passive;
+	bool allow_ibss;
+	bool allow_ht;
+	bool allow_vht;
+	bool ht40plus;
+	/* note: power unit is 1/4th of dBm */
+	u32 min_power;
+	u32 max_power;
+	u32 max_reg_power;
+	u32 max_antenna_gain;
+	u32 reg_class_id;
+	enum wmi_phy_mode mode;
+};
+
+enum wmi_channel_change_cause {
+	WMI_CHANNEL_CHANGE_CAUSE_NONE = 0,
+	WMI_CHANNEL_CHANGE_CAUSE_CSA,
+};
+
+#define WMI_CHAN_FLAG_HT40_PLUS      (1 << 6)
+#define WMI_CHAN_FLAG_PASSIVE        (1 << 7)
+#define WMI_CHAN_FLAG_ADHOC_ALLOWED  (1 << 8)
+#define WMI_CHAN_FLAG_AP_DISABLED    (1 << 9)
+#define WMI_CHAN_FLAG_DFS            (1 << 10)
+#define WMI_CHAN_FLAG_ALLOW_HT       (1 << 11)
+#define WMI_CHAN_FLAG_ALLOW_VHT      (1 << 12)
+
+/* Indicate reason for channel switch */
+#define WMI_CHANNEL_CHANGE_CAUSE_CSA (1 << 13)
+
+#define WMI_MAX_SPATIAL_STREAM   3
+
+/* HT Capabilities*/
+#define WMI_HT_CAP_ENABLED                0x0001   /* HT Enabled/ disabled */
+#define WMI_HT_CAP_HT20_SGI       0x0002   /* Short Guard Interval with HT20 */
+#define WMI_HT_CAP_DYNAMIC_SMPS           0x0004   /* Dynamic MIMO powersave */
+#define WMI_HT_CAP_TX_STBC                0x0008   /* B3 TX STBC */
+#define WMI_HT_CAP_TX_STBC_MASK_SHIFT     3
+#define WMI_HT_CAP_RX_STBC                0x0030   /* B4-B5 RX STBC */
+#define WMI_HT_CAP_RX_STBC_MASK_SHIFT     4
+#define WMI_HT_CAP_LDPC                   0x0040   /* LDPC supported */
+#define WMI_HT_CAP_L_SIG_TXOP_PROT        0x0080   /* L-SIG TXOP Protection */
+#define WMI_HT_CAP_MPDU_DENSITY           0x0700   /* MPDU Density */
+#define WMI_HT_CAP_MPDU_DENSITY_MASK_SHIFT 8
+#define WMI_HT_CAP_HT40_SGI               0x0800
+
+#define WMI_HT_CAP_DEFAULT_ALL (WMI_HT_CAP_ENABLED       | \
+				WMI_HT_CAP_HT20_SGI      | \
+				WMI_HT_CAP_HT40_SGI      | \
+				WMI_HT_CAP_TX_STBC       | \
+				WMI_HT_CAP_RX_STBC       | \
+				WMI_HT_CAP_LDPC)
+
+
+/*
+ * WMI_VHT_CAP_* these maps to ieee 802.11ac vht capability information
+ * field. The fields not defined here are not supported, or reserved.
+ * Do not change these masks and if you have to add new one follow the
+ * bitmask as specified by 802.11ac draft.
+ */
+
+#define WMI_VHT_CAP_MAX_MPDU_LEN_MASK            0x00000003
+#define WMI_VHT_CAP_RX_LDPC                      0x00000010
+#define WMI_VHT_CAP_SGI_80MHZ                    0x00000020
+#define WMI_VHT_CAP_TX_STBC                      0x00000080
+#define WMI_VHT_CAP_RX_STBC_MASK                 0x00000300
+#define WMI_VHT_CAP_RX_STBC_MASK_SHIFT           8
+#define WMI_VHT_CAP_MAX_AMPDU_LEN_EXP            0x03800000
+#define WMI_VHT_CAP_MAX_AMPDU_LEN_EXP_SHIFT      23
+#define WMI_VHT_CAP_RX_FIXED_ANT                 0x10000000
+#define WMI_VHT_CAP_TX_FIXED_ANT                 0x20000000
+
+/* The following also refer for max HT AMSDU */
+#define WMI_VHT_CAP_MAX_MPDU_LEN_3839            0x00000000
+#define WMI_VHT_CAP_MAX_MPDU_LEN_7935            0x00000001
+#define WMI_VHT_CAP_MAX_MPDU_LEN_11454           0x00000002
+
+#define WMI_VHT_CAP_DEFAULT_ALL (WMI_VHT_CAP_MAX_MPDU_LEN_11454  | \
+				 WMI_VHT_CAP_RX_LDPC             | \
+				 WMI_VHT_CAP_SGI_80MHZ           | \
+				 WMI_VHT_CAP_TX_STBC             | \
+				 WMI_VHT_CAP_RX_STBC_MASK        | \
+				 WMI_VHT_CAP_MAX_AMPDU_LEN_EXP   | \
+				 WMI_VHT_CAP_RX_FIXED_ANT        | \
+				 WMI_VHT_CAP_TX_FIXED_ANT)
+
+/*
+ * Interested readers refer to Rx/Tx MCS Map definition as defined in
+ * 802.11ac
+ */
+#define WMI_VHT_MAX_MCS_4_SS_MASK(r, ss)      ((3 & (r)) << (((ss) - 1) << 1))
+#define WMI_VHT_MAX_SUPP_RATE_MASK           0x1fff0000
+#define WMI_VHT_MAX_SUPP_RATE_MASK_SHIFT     16
+
+enum {
+	REGDMN_MODE_11A              = 0x00001, /* 11a channels */
+	REGDMN_MODE_TURBO            = 0x00002, /* 11a turbo-only channels */
+	REGDMN_MODE_11B              = 0x00004, /* 11b channels */
+	REGDMN_MODE_PUREG            = 0x00008, /* 11g channels (OFDM only) */
+	REGDMN_MODE_11G              = 0x00008, /* XXX historical */
+	REGDMN_MODE_108G             = 0x00020, /* 11a+Turbo channels */
+	REGDMN_MODE_108A             = 0x00040, /* 11g+Turbo channels */
+	REGDMN_MODE_XR               = 0x00100, /* XR channels */
+	REGDMN_MODE_11A_HALF_RATE    = 0x00200, /* 11A half rate channels */
+	REGDMN_MODE_11A_QUARTER_RATE = 0x00400, /* 11A quarter rate channels */
+	REGDMN_MODE_11NG_HT20        = 0x00800, /* 11N-G HT20 channels */
+	REGDMN_MODE_11NA_HT20        = 0x01000, /* 11N-A HT20 channels */
+	REGDMN_MODE_11NG_HT40PLUS    = 0x02000, /* 11N-G HT40 + channels */
+	REGDMN_MODE_11NG_HT40MINUS   = 0x04000, /* 11N-G HT40 - channels */
+	REGDMN_MODE_11NA_HT40PLUS    = 0x08000, /* 11N-A HT40 + channels */
+	REGDMN_MODE_11NA_HT40MINUS   = 0x10000, /* 11N-A HT40 - channels */
+	REGDMN_MODE_11AC_VHT20       = 0x20000, /* 5Ghz, VHT20 */
+	REGDMN_MODE_11AC_VHT40PLUS   = 0x40000, /* 5Ghz, VHT40 + channels */
+	REGDMN_MODE_11AC_VHT40MINUS  = 0x80000, /* 5Ghz  VHT40 - channels */
+	REGDMN_MODE_11AC_VHT80       = 0x100000, /* 5Ghz, VHT80 channels */
+	REGDMN_MODE_ALL              = 0xffffffff
+};
+
+#define REGDMN_CAP1_CHAN_HALF_RATE        0x00000001
+#define REGDMN_CAP1_CHAN_QUARTER_RATE     0x00000002
+#define REGDMN_CAP1_CHAN_HAL49GHZ         0x00000004
+
+/* regulatory capabilities */
+#define REGDMN_EEPROM_EEREGCAP_EN_FCC_MIDBAND   0x0040
+#define REGDMN_EEPROM_EEREGCAP_EN_KK_U1_EVEN    0x0080
+#define REGDMN_EEPROM_EEREGCAP_EN_KK_U2         0x0100
+#define REGDMN_EEPROM_EEREGCAP_EN_KK_MIDBAND    0x0200
+#define REGDMN_EEPROM_EEREGCAP_EN_KK_U1_ODD     0x0400
+#define REGDMN_EEPROM_EEREGCAP_EN_KK_NEW_11A    0x0800
+
+struct hal_reg_capabilities {
+	/* regdomain value specified in EEPROM */
+	__le32 eeprom_rd;
+	/*regdomain */
+	__le32 eeprom_rd_ext;
+	/* CAP1 capabilities bit map. */
+	__le32 regcap1;
+	/* REGDMN EEPROM CAP. */
+	__le32 regcap2;
+	/* REGDMN MODE */
+	__le32 wireless_modes;
+	__le32 low_2ghz_chan;
+	__le32 high_2ghz_chan;
+	__le32 low_5ghz_chan;
+	__le32 high_5ghz_chan;
+} __packed;
+
+enum wlan_mode_capability {
+	WHAL_WLAN_11A_CAPABILITY   = 0x1,
+	WHAL_WLAN_11G_CAPABILITY   = 0x2,
+	WHAL_WLAN_11AG_CAPABILITY  = 0x3,
+};
+
+/* structure used by FW for requesting host memory */
+struct wlan_host_mem_req {
+	/* ID of the request */
+	__le32 req_id;
+	/* size of the  of each unit */
+	__le32 unit_size;
+	/* flags to  indicate that
+	 * the number units is dependent
+	 * on number of resources(num vdevs num peers .. etc)
+	 */
+	__le32 num_unit_info;
+	/*
+	 * actual number of units to allocate . if flags in the num_unit_info
+	 * indicate that number of units is tied to number of a particular
+	 * resource to allocate then  num_units filed is set to 0 and host
+	 * will derive the number units from number of the resources it is
+	 * requesting.
+	 */
+	__le32 num_units;
+} __packed;
+
+#define WMI_SERVICE_IS_ENABLED(wmi_svc_bmap, svc_id) \
+	((((wmi_svc_bmap)[(svc_id)/(sizeof(u32))]) & \
+	(1 << ((svc_id)%(sizeof(u32))))) != 0)
+
+/*
+ * The following struct holds optional payload for
+ * wmi_service_ready_event,e.g., 11ac pass some of the
+ * device capability to the host.
+ */
+struct wmi_service_ready_event {
+	__le32 sw_version;
+	__le32 sw_version_1;
+	__le32 abi_version;
+	/* WMI_PHY_CAPABILITY */
+	__le32 phy_capability;
+	/* Maximum number of frag table entries that SW will populate less 1 */
+	__le32 max_frag_entry;
+	__le32 wmi_service_bitmap[WMI_SERVICE_BM_SIZE];
+	__le32 num_rf_chains;
+	/*
+	 * The following field is only valid for service type
+	 * WMI_SERVICE_11AC
+	 */
+	__le32 ht_cap_info; /* WMI HT Capability */
+	__le32 vht_cap_info; /* VHT capability info field of 802.11ac */
+	__le32 vht_supp_mcs; /* VHT Supported MCS Set field Rx/Tx same */
+	__le32 hw_min_tx_power;
+	__le32 hw_max_tx_power;
+	struct hal_reg_capabilities hal_reg_capabilities;
+	__le32 sys_cap_info;
+	__le32 min_pkt_size_enable; /* Enterprise mode short pkt enable */
+	/*
+	 * Max beacon and Probe Response IE offload size
+	 * (includes optional P2P IEs)
+	 */
+	__le32 max_bcn_ie_size;
+	/*
+	 * request to host to allocate a chuck of memory and pss it down to FW
+	 * via WM_INIT. FW uses this as FW extesnsion memory for saving its
+	 * data structures. Only valid for low latency interfaces like PCIE
+	 * where FW can access this memory directly (or) by DMA.
+	 */
+	__le32 num_mem_reqs;
+	struct wlan_host_mem_req mem_reqs[1];
+} __packed;
+
+/*
+ * status consists of  upper 16 bits fo int status and lower 16 bits of
+ * module ID that retuned status
+ */
+#define WLAN_INIT_STATUS_SUCCESS   0x0
+#define WLAN_GET_INIT_STATUS_REASON(status)    ((status) & 0xffff)
+#define WLAN_GET_INIT_STATUS_MODULE_ID(status) (((status) >> 16) & 0xffff)
+
+#define WMI_SERVICE_READY_TIMEOUT_HZ (5*HZ)
+#define WMI_UNIFIED_READY_TIMEOUT_HZ (5*HZ)
+
+struct wmi_ready_event {
+	__le32 sw_version;
+	__le32 abi_version;
+	struct wmi_mac_addr mac_addr;
+	__le32 status;
+} __packed;
+
+struct wmi_resource_config {
+	/* number of virtual devices (VAPs) to support */
+	__le32 num_vdevs;
+
+	/* number of peer nodes to support */
+	__le32 num_peers;
+
+	/*
+	 * In offload mode target supports features like WOW, chatter and
+	 * other protocol offloads. In order to support them some
+	 * functionalities like reorder buffering, PN checking need to be
+	 * done in target. This determines maximum number of peers suported
+	 * by target in offload mode
+	 */
+	__le32 num_offload_peers;
+
+	/* For target-based RX reordering */
+	__le32 num_offload_reorder_bufs;
+
+	/* number of keys per peer */
+	__le32 num_peer_keys;
+
+	/* total number of TX/RX data TIDs */
+	__le32 num_tids;
+
+	/*
+	 * max skid for resolving hash collisions
+	 *
+	 *   The address search table is sparse, so that if two MAC addresses
+	 *   result in the same hash value, the second of these conflicting
+	 *   entries can slide to the next index in the address search table,
+	 *   and use it, if it is unoccupied.  This ast_skid_limit parameter
+	 *   specifies the upper bound on how many subsequent indices to search
+	 *   over to find an unoccupied space.
+	 */
+	__le32 ast_skid_limit;
+
+	/*
+	 * the nominal chain mask for transmit
+	 *
+	 *   The chain mask may be modified dynamically, e.g. to operate AP
+	 *   tx with a reduced number of chains if no clients are associated.
+	 *   This configuration parameter specifies the nominal chain-mask that
+	 *   should be used when not operating with a reduced set of tx chains.
+	 */
+	__le32 tx_chain_mask;
+
+	/*
+	 * the nominal chain mask for receive
+	 *
+	 *   The chain mask may be modified dynamically, e.g. for a client
+	 *   to use a reduced number of chains for receive if the traffic to
+	 *   the client is low enough that it doesn't require downlink MIMO
+	 *   or antenna diversity.
+	 *   This configuration parameter specifies the nominal chain-mask that
+	 *   should be used when not operating with a reduced set of rx chains.
+	 */
+	__le32 rx_chain_mask;
+
+	/*
+	 * what rx reorder timeout (ms) to use for the AC
+	 *
+	 *   Each WMM access class (voice, video, best-effort, background) will
+	 *   have its own timeout value to dictate how long to wait for missing
+	 *   rx MPDUs to arrive before flushing subsequent MPDUs that have
+	 *   already been received.
+	 *   This parameter specifies the timeout in milliseconds for each
+	 *   class.
+	 */
+	__le32 rx_timeout_pri_vi;
+	__le32 rx_timeout_pri_vo;
+	__le32 rx_timeout_pri_be;
+	__le32 rx_timeout_pri_bk;
+
+	/*
+	 * what mode the rx should decap packets to
+	 *
+	 *   MAC can decap to RAW (no decap), native wifi or Ethernet types
+	 *   THis setting also determines the default TX behavior, however TX
+	 *   behavior can be modified on a per VAP basis during VAP init
+	 */
+	__le32 rx_decap_mode;
+
+	/* what is the maximum scan requests than can be queued */
+	__le32 scan_max_pending_reqs;
+
+	/* maximum VDEV that could use BMISS offload */
+	__le32 bmiss_offload_max_vdev;
+
+	/* maximum VDEV that could use offload roaming */
+	__le32 roam_offload_max_vdev;
+
+	/* maximum AP profiles that would push to offload roaming */
+	__le32 roam_offload_max_ap_profiles;
+
+	/*
+	 * how many groups to use for mcast->ucast conversion
+	 *
+	 *   The target's WAL maintains a table to hold information regarding
+	 *   which peers belong to a given multicast group, so that if
+	 *   multicast->unicast conversion is enabled, the target can convert
+	 *   multicast tx frames to a series of unicast tx frames, to each
+	 *   peer within the multicast group.
+	     This num_mcast_groups configuration parameter tells the target how
+	 *   many multicast groups to provide storage for within its multicast
+	 *   group membership table.
+	 */
+	__le32 num_mcast_groups;
+
+	/*
+	 * size to alloc for the mcast membership table
+	 *
+	 *   This num_mcast_table_elems configuration parameter tells the
+	 *   target how many peer elements it needs to provide storage for in
+	 *   its multicast group membership table.
+	 *   These multicast group membership table elements are shared by the
+	 *   multicast groups stored within the table.
+	 */
+	__le32 num_mcast_table_elems;
+
+	/*
+	 * whether/how to do multicast->unicast conversion
+	 *
+	 *   This configuration parameter specifies whether the target should
+	 *   perform multicast --> unicast conversion on transmit, and if so,
+	 *   what to do if it finds no entries in its multicast group
+	 *   membership table for the multicast IP address in the tx frame.
+	 *   Configuration value:
+	 *   0 -> Do not perform multicast to unicast conversion.
+	 *   1 -> Convert multicast frames to unicast, if the IP multicast
+	 *        address from the tx frame is found in the multicast group
+	 *        membership table.  If the IP multicast address is not found,
+	 *        drop the frame.
+	 *   2 -> Convert multicast frames to unicast, if the IP multicast
+	 *        address from the tx frame is found in the multicast group
+	 *        membership table.  If the IP multicast address is not found,
+	 *        transmit the frame as multicast.
+	 */
+	__le32 mcast2ucast_mode;
+
+	/*
+	 * how much memory to allocate for a tx PPDU dbg log
+	 *
+	 *   This parameter controls how much memory the target will allocate
+	 *   to store a log of tx PPDU meta-information (how large the PPDU
+	 *   was, when it was sent, whether it was successful, etc.)
+	 */
+	__le32 tx_dbg_log_size;
+
+	/* how many AST entries to be allocated for WDS */
+	__le32 num_wds_entries;
+
+	/*
+	 * MAC DMA burst size, e.g., For target PCI limit can be
+	 * 0 -default, 1 256B
+	 */
+	__le32 dma_burst_size;
+
+	/*
+	 * Fixed delimiters to be inserted after every MPDU to
+	 * account for interface latency to avoid underrun.
+	 */
+	__le32 mac_aggr_delim;
+
+	/*
+	 *   determine whether target is responsible for detecting duplicate
+	 *   non-aggregate MPDU and timing out stale fragments.
+	 *
+	 *   A-MPDU reordering is always performed on the target.
+	 *
+	 *   0: target responsible for frag timeout and dup checking
+	 *   1: host responsible for frag timeout and dup checking
+	 */
+	__le32 rx_skip_defrag_timeout_dup_detection_check;
+
+	/*
+	 * Configuration for VoW :
+	 * No of Video Nodes to be supported
+	 * and Max no of descriptors for each Video link (node).
+	 */
+	__le32 vow_config;
+
+	/* maximum VDEV that could use GTK offload */
+	__le32 gtk_offload_max_vdev;
+
+	/* Number of msdu descriptors target should use */
+	__le32 num_msdu_desc;
+
+	/*
+	 * Max. number of Tx fragments per MSDU
+	 *  This parameter controls the max number of Tx fragments per MSDU.
+	 *  This is sent by the target as part of the WMI_SERVICE_READY event
+	 *  and is overriden by the OS shim as required.
+	 */
+	__le32 max_frag_entries;
+} __packed;
+
+/* strucutre describing host memory chunk. */
+struct host_memory_chunk {
+	/* id of the request that is passed up in service ready */
+	__le32 req_id;
+	/* the physical address the memory chunk */
+	__le32 ptr;
+	/* size of the chunk */
+	__le32 size;
+} __packed;
+
+struct wmi_init_cmd {
+	struct wmi_resource_config resource_config;
+	__le32 num_host_mem_chunks;
+
+	/*
+	 * variable number of host memory chunks.
+	 * This should be the last element in the structure
+	 */
+	struct host_memory_chunk host_mem_chunks[1];
+} __packed;
+
+/* TLV for channel list */
+struct wmi_chan_list {
+	__le32 tag; /* WMI_CHAN_LIST_TAG */
+	__le32 num_chan;
+	__le32 channel_list[0];
+} __packed;
+
+struct wmi_bssid_list {
+	__le32 tag; /* WMI_BSSID_LIST_TAG */
+	__le32 num_bssid;
+	struct wmi_mac_addr bssid_list[0];
+} __packed;
+
+struct wmi_ie_data {
+	__le32 tag; /* WMI_IE_TAG */
+	__le32 ie_len;
+	u8 ie_data[0];
+} __packed;
+
+struct wmi_ssid {
+	__le32 ssid_len;
+	u8 ssid[32];
+} __packed;
+
+struct wmi_ssid_list {
+	__le32 tag; /* WMI_SSID_LIST_TAG */
+	__le32 num_ssids;
+	struct wmi_ssid ssids[0];
+} __packed;
+
+/* prefix used by scan requestor ids on the host */
+#define WMI_HOST_SCAN_REQUESTOR_ID_PREFIX 0xA000
+
+/* prefix used by scan request ids generated on the host */
+/* host cycles through the lower 12 bits to generate ids */
+#define WMI_HOST_SCAN_REQ_ID_PREFIX 0xA000
+
+#define WLAN_SCAN_PARAMS_MAX_SSID    16
+#define WLAN_SCAN_PARAMS_MAX_BSSID   4
+#define WLAN_SCAN_PARAMS_MAX_IE_LEN  256
+
+/* Scan priority numbers must be sequential, starting with 0 */
+enum wmi_scan_priority {
+	WMI_SCAN_PRIORITY_VERY_LOW = 0,
+	WMI_SCAN_PRIORITY_LOW,
+	WMI_SCAN_PRIORITY_MEDIUM,
+	WMI_SCAN_PRIORITY_HIGH,
+	WMI_SCAN_PRIORITY_VERY_HIGH,
+	WMI_SCAN_PRIORITY_COUNT   /* number of priorities supported */
+};
+
+struct wmi_start_scan_cmd {
+	/* Scan ID */
+	__le32 scan_id;
+	/* Scan requestor ID */
+	__le32 scan_req_id;
+	/* VDEV id(interface) that is requesting scan */
+	__le32 vdev_id;
+	/* Scan Priority, input to scan scheduler */
+	__le32 scan_priority;
+	/* Scan events subscription */
+	__le32 notify_scan_events;
+	/* dwell time in msec on active channels */
+	__le32 dwell_time_active;
+	/* dwell time in msec on passive channels */
+	__le32 dwell_time_passive;
+	/*
+	 * min time in msec on the BSS channel,only valid if atleast one
+	 * VDEV is active
+	 */
+	__le32 min_rest_time;
+	/*
+	 * max rest time in msec on the BSS channel,only valid if at least
+	 * one VDEV is active
+	 */
+	/*
+	 * the scanner will rest on the bss channel at least min_rest_time
+	 * after min_rest_time the scanner will start checking for tx/rx
+	 * activity on all VDEVs. if there is no activity the scanner will
+	 * switch to off channel. if there is activity the scanner will let
+	 * the radio on the bss channel until max_rest_time expires.at
+	 * max_rest_time scanner will switch to off channel irrespective of
+	 * activity. activity is determined by the idle_time parameter.
+	 */
+	__le32 max_rest_time;
+	/*
+	 * time before sending next set of probe requests.
+	 * The scanner keeps repeating probe requests transmission with
+	 * period specified by repeat_probe_time.
+	 * The number of probe requests specified depends on the ssid_list
+	 * and bssid_list
+	 */
+	__le32 repeat_probe_time;
+	/* time in msec between 2 consequetive probe requests with in a set. */
+	__le32 probe_spacing_time;
+	/*
+	 * data inactivity time in msec on bss channel that will be used by
+	 * scanner for measuring the inactivity.
+	 */
+	__le32 idle_time;
+	/* maximum time in msec allowed for scan  */
+	__le32 max_scan_time;
+	/*
+	 * delay in msec before sending first probe request after switching
+	 * to a channel
+	 */
+	__le32 probe_delay;
+	/* Scan control flags */
+	__le32 scan_ctrl_flags;
+
+	/* Burst duration time in msecs */
+	__le32 burst_duration;
+	/*
+	 * TLV (tag length value )  paramerters follow the scan_cmd structure.
+	 * TLV can contain channel list, bssid list, ssid list and
+	 * ie. the TLV tags are defined above;
+	 */
+} __packed;
+
+struct wmi_ssid_arg {
+	int len;
+	const u8 *ssid;
+};
+
+struct wmi_bssid_arg {
+	const u8 *bssid;
+};
+
+struct wmi_start_scan_arg {
+	u32 scan_id;
+	u32 scan_req_id;
+	u32 vdev_id;
+	u32 scan_priority;
+	u32 notify_scan_events;
+	u32 dwell_time_active;
+	u32 dwell_time_passive;
+	u32 min_rest_time;
+	u32 max_rest_time;
+	u32 repeat_probe_time;
+	u32 probe_spacing_time;
+	u32 idle_time;
+	u32 max_scan_time;
+	u32 probe_delay;
+	u32 scan_ctrl_flags;
+
+	u32 ie_len;
+	u32 n_channels;
+	u32 n_ssids;
+	u32 n_bssids;
+
+	u8 ie[WLAN_SCAN_PARAMS_MAX_IE_LEN];
+	u32 channels[64];
+	struct wmi_ssid_arg ssids[WLAN_SCAN_PARAMS_MAX_SSID];
+	struct wmi_bssid_arg bssids[WLAN_SCAN_PARAMS_MAX_BSSID];
+};
+
+/* scan control flags */
+
+/* passively scan all channels including active channels */
+#define WMI_SCAN_FLAG_PASSIVE        0x1
+/* add wild card ssid probe request even though ssid_list is specified. */
+#define WMI_SCAN_ADD_BCAST_PROBE_REQ 0x2
+/* add cck rates to rates/xrate ie for the generated probe request */
+#define WMI_SCAN_ADD_CCK_RATES 0x4
+/* add ofdm rates to rates/xrate ie for the generated probe request */
+#define WMI_SCAN_ADD_OFDM_RATES 0x8
+/* To enable indication of Chan load and Noise floor to host */
+#define WMI_SCAN_CHAN_STAT_EVENT 0x10
+/* Filter Probe request frames  */
+#define WMI_SCAN_FILTER_PROBE_REQ 0x20
+/* When set, DFS channels will not be scanned */
+#define WMI_SCAN_BYPASS_DFS_CHN 0x40
+/* Different FW scan engine may choose to bail out on errors.
+ * Allow the driver to have influence over that. */
+#define WMI_SCAN_CONTINUE_ON_ERROR 0x80
+
+/* WMI_SCAN_CLASS_MASK must be the same value as IEEE80211_SCAN_CLASS_MASK */
+#define WMI_SCAN_CLASS_MASK 0xFF000000
+
+
+enum wmi_stop_scan_type {
+	WMI_SCAN_STOP_ONE	= 0x00000000, /* stop by scan_id */
+	WMI_SCAN_STOP_VDEV_ALL	= 0x01000000, /* stop by vdev_id */
+	WMI_SCAN_STOP_ALL	= 0x04000000, /* stop all scans */
+};
+
+struct wmi_stop_scan_cmd {
+	__le32 scan_req_id;
+	__le32 scan_id;
+	__le32 req_type;
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_stop_scan_arg {
+	u32 req_id;
+	enum wmi_stop_scan_type req_type;
+	union {
+		u32 scan_id;
+		u32 vdev_id;
+	} u;
+};
+
+struct wmi_scan_chan_list_cmd {
+	__le32 num_scan_chans;
+	struct wmi_channel chan_info[0];
+} __packed;
+
+struct wmi_scan_chan_list_arg {
+	u32 n_channels;
+	struct wmi_channel_arg *channels;
+};
+
+enum wmi_bss_filter {
+	WMI_BSS_FILTER_NONE = 0,        /* no beacons forwarded */
+	WMI_BSS_FILTER_ALL,             /* all beacons forwarded */
+	WMI_BSS_FILTER_PROFILE,         /* only beacons matching profile */
+	WMI_BSS_FILTER_ALL_BUT_PROFILE, /* all but beacons matching profile */
+	WMI_BSS_FILTER_CURRENT_BSS,     /* only beacons matching current BSS */
+	WMI_BSS_FILTER_ALL_BUT_BSS,     /* all but beacons matching BSS */
+	WMI_BSS_FILTER_PROBED_SSID,     /* beacons matching probed ssid */
+	WMI_BSS_FILTER_LAST_BSS,        /* marker only */
+};
+
+enum wmi_scan_event_type {
+	WMI_SCAN_EVENT_STARTED         = 0x1,
+	WMI_SCAN_EVENT_COMPLETED       = 0x2,
+	WMI_SCAN_EVENT_BSS_CHANNEL     = 0x4,
+	WMI_SCAN_EVENT_FOREIGN_CHANNEL = 0x8,
+	WMI_SCAN_EVENT_DEQUEUED        = 0x10,
+	WMI_SCAN_EVENT_PREEMPTED       = 0x20, /* possibly by high-prio scan */
+	WMI_SCAN_EVENT_START_FAILED    = 0x40,
+	WMI_SCAN_EVENT_RESTARTED       = 0x80,
+	WMI_SCAN_EVENT_MAX             = 0x8000
+};
+
+enum wmi_scan_completion_reason {
+	WMI_SCAN_REASON_COMPLETED,
+	WMI_SCAN_REASON_CANCELLED,
+	WMI_SCAN_REASON_PREEMPTED,
+	WMI_SCAN_REASON_TIMEDOUT,
+	WMI_SCAN_REASON_MAX,
+};
+
+struct wmi_scan_event {
+	__le32 event_type; /* %WMI_SCAN_EVENT_ */
+	__le32 reason; /* %WMI_SCAN_REASON_ */
+	__le32 channel_freq; /* only valid for WMI_SCAN_EVENT_FOREIGN_CHANNEL */
+	__le32 scan_req_id;
+	__le32 scan_id;
+	__le32 vdev_id;
+} __packed;
+
+/*
+ * This defines how much headroom is kept in the
+ * receive frame between the descriptor and the
+ * payload, in order for the WMI PHY error and
+ * management handler to insert header contents.
+ *
+ * This is in bytes.
+ */
+#define WMI_MGMT_RX_HDR_HEADROOM    52
+
+/*
+ * This event will be used for sending scan results
+ * as well as rx mgmt frames to the host. The rx buffer
+ * will be sent as part of this WMI event. It would be a
+ * good idea to pass all the fields in the RX status
+ * descriptor up to the host.
+ */
+struct wmi_mgmt_rx_hdr {
+	__le32 channel;
+	__le32 snr;
+	__le32 rate;
+	__le32 phy_mode;
+	__le32 buf_len;
+	__le32 status; /* %WMI_RX_STATUS_ */
+} __packed;
+
+struct wmi_mgmt_rx_event {
+	struct wmi_mgmt_rx_hdr hdr;
+	u8 buf[0];
+} __packed;
+
+#define WMI_RX_STATUS_OK			0x00
+#define WMI_RX_STATUS_ERR_CRC			0x01
+#define WMI_RX_STATUS_ERR_DECRYPT		0x08
+#define WMI_RX_STATUS_ERR_MIC			0x10
+#define WMI_RX_STATUS_ERR_KEY_CACHE_MISS	0x20
+
+struct wmi_single_phyerr_rx_hdr {
+	/* TSF timestamp */
+	__le32 tsf_timestamp;
+
+	/*
+	 * Current freq1, freq2
+	 *
+	 * [7:0]:    freq1[lo]
+	 * [15:8] :   freq1[hi]
+	 * [23:16]:   freq2[lo]
+	 * [31:24]:   freq2[hi]
+	 */
+	__le16 freq1;
+	__le16 freq2;
+
+	/*
+	 * Combined RSSI over all chains and channel width for this PHY error
+	 *
+	 * [7:0]: RSSI combined
+	 * [15:8]: Channel width (MHz)
+	 * [23:16]: PHY error code
+	 * [24:16]: reserved (future use)
+	 */
+	u8 rssi_combined;
+	u8 chan_width_mhz;
+	u8 phy_err_code;
+	u8 rsvd0;
+
+	/*
+	 * RSSI on chain 0 through 3
+	 *
+	 * This is formatted the same as the PPDU_START RX descriptor
+	 * field:
+	 *
+	 * [7:0]:   pri20
+	 * [15:8]:  sec20
+	 * [23:16]: sec40
+	 * [31:24]: sec80
+	 */
+
+	__le32 rssi_chain0;
+	__le32 rssi_chain1;
+	__le32 rssi_chain2;
+	__le32 rssi_chain3;
+
+	/*
+	 * Last calibrated NF value for chain 0 through 3
+	 *
+	 * nf_list_1:
+	 *
+	 * + [15:0] - chain 0
+	 * + [31:16] - chain 1
+	 *
+	 * nf_list_2:
+	 *
+	 * + [15:0] - chain 2
+	 * + [31:16] - chain 3
+	 */
+	__le32 nf_list_1;
+	__le32 nf_list_2;
+
+
+	/* Length of the frame */
+	__le32 buf_len;
+} __packed;
+
+struct wmi_single_phyerr_rx_event {
+	/* Phy error event header */
+	struct wmi_single_phyerr_rx_hdr hdr;
+	/* frame buffer */
+	u8 bufp[0];
+} __packed;
+
+struct wmi_comb_phyerr_rx_hdr {
+	/* Phy error phy error count */
+	__le32 num_phyerr_events;
+	__le32 tsf_l32;
+	__le32 tsf_u32;
+} __packed;
+
+struct wmi_comb_phyerr_rx_event {
+	/* Phy error phy error count */
+	struct wmi_comb_phyerr_rx_hdr hdr;
+	/*
+	 * frame buffer - contains multiple payloads in the order:
+	 *                    header - payload, header - payload...
+	 *  (The header is of type: wmi_single_phyerr_rx_hdr)
+	 */
+	u8 bufp[0];
+} __packed;
+
+struct wmi_mgmt_tx_hdr {
+	__le32 vdev_id;
+	struct wmi_mac_addr peer_macaddr;
+	__le32 tx_rate;
+	__le32 tx_power;
+	__le32 buf_len;
+} __packed;
+
+struct wmi_mgmt_tx_cmd {
+	struct wmi_mgmt_tx_hdr hdr;
+	u8 buf[0];
+} __packed;
+
+struct wmi_echo_event {
+	__le32 value;
+} __packed;
+
+struct wmi_echo_cmd {
+	__le32 value;
+} __packed;
+
+
+struct wmi_pdev_set_regdomain_cmd {
+	__le32 reg_domain;
+	__le32 reg_domain_2G;
+	__le32 reg_domain_5G;
+	__le32 conformance_test_limit_2G;
+	__le32 conformance_test_limit_5G;
+} __packed;
+
+/* Command to set/unset chip in quiet mode */
+struct wmi_pdev_set_quiet_cmd {
+	/* period in TUs */
+	__le32 period;
+
+	/* duration in TUs */
+	__le32 duration;
+
+	/* offset in TUs */
+	__le32 next_start;
+
+	/* enable/disable */
+	__le32 enabled;
+} __packed;
+
+
+/*
+ * 802.11g protection mode.
+ */
+enum ath10k_protmode {
+	ATH10K_PROT_NONE     = 0,    /* no protection */
+	ATH10K_PROT_CTSONLY  = 1,    /* CTS to self */
+	ATH10K_PROT_RTSCTS   = 2,    /* RTS-CTS */
+};
+
+enum wmi_beacon_gen_mode {
+	WMI_BEACON_STAGGERED_MODE = 0,
+	WMI_BEACON_BURST_MODE = 1
+};
+
+enum wmi_csa_event_ies_present_flag {
+	WMI_CSA_IE_PRESENT = 0x00000001,
+	WMI_XCSA_IE_PRESENT = 0x00000002,
+	WMI_WBW_IE_PRESENT = 0x00000004,
+	WMI_CSWARP_IE_PRESENT = 0x00000008,
+};
+
+/* wmi CSA receive event from beacon frame */
+struct wmi_csa_event {
+	__le32 i_fc_dur;
+	/* Bit 0-15: FC */
+	/* Bit 16-31: DUR */
+	struct wmi_mac_addr i_addr1;
+	struct wmi_mac_addr i_addr2;
+	__le32 csa_ie[2];
+	__le32 xcsa_ie[2];
+	__le32 wb_ie[2];
+	__le32 cswarp_ie;
+	__le32 ies_present_flag; /* wmi_csa_event_ies_present_flag */
+} __packed;
+
+/* the definition of different PDEV parameters */
+#define PDEV_DEFAULT_STATS_UPDATE_PERIOD    500
+#define VDEV_DEFAULT_STATS_UPDATE_PERIOD    500
+#define PEER_DEFAULT_STATS_UPDATE_PERIOD    500
+
+enum wmi_pdev_param {
+	/* TX chian mask */
+	WMI_PDEV_PARAM_TX_CHAIN_MASK = 0x1,
+	/* RX chian mask */
+	WMI_PDEV_PARAM_RX_CHAIN_MASK,
+	/* TX power limit for 2G Radio */
+	WMI_PDEV_PARAM_TXPOWER_LIMIT2G,
+	/* TX power limit for 5G Radio */
+	WMI_PDEV_PARAM_TXPOWER_LIMIT5G,
+	/* TX power scale */
+	WMI_PDEV_PARAM_TXPOWER_SCALE,
+	/* Beacon generation mode . 0: host, 1: target   */
+	WMI_PDEV_PARAM_BEACON_GEN_MODE,
+	/* Beacon generation mode . 0: staggered 1: bursted   */
+	WMI_PDEV_PARAM_BEACON_TX_MODE,
+	/*
+	 * Resource manager off chan mode .
+	 * 0: turn off off chan mode. 1: turn on offchan mode
+	 */
+	WMI_PDEV_PARAM_RESMGR_OFFCHAN_MODE,
+	/*
+	 * Protection mode:
+	 * 0: no protection 1:use CTS-to-self 2: use RTS/CTS
+	 */
+	WMI_PDEV_PARAM_PROTECTION_MODE,
+	/* Dynamic bandwidth 0: disable 1: enable */
+	WMI_PDEV_PARAM_DYNAMIC_BW,
+	/* Non aggregrate/ 11g sw retry threshold.0-disable */
+	WMI_PDEV_PARAM_NON_AGG_SW_RETRY_TH,
+	/* aggregrate sw retry threshold. 0-disable*/
+	WMI_PDEV_PARAM_AGG_SW_RETRY_TH,
+	/* Station kickout threshold (non of consecutive failures).0-disable */
+	WMI_PDEV_PARAM_STA_KICKOUT_TH,
+	/* Aggerate size scaling configuration per AC */
+	WMI_PDEV_PARAM_AC_AGGRSIZE_SCALING,
+	/* LTR enable */
+	WMI_PDEV_PARAM_LTR_ENABLE,
+	/* LTR latency for BE, in us */
+	WMI_PDEV_PARAM_LTR_AC_LATENCY_BE,
+	/* LTR latency for BK, in us */
+	WMI_PDEV_PARAM_LTR_AC_LATENCY_BK,
+	/* LTR latency for VI, in us */
+	WMI_PDEV_PARAM_LTR_AC_LATENCY_VI,
+	/* LTR latency for VO, in us  */
+	WMI_PDEV_PARAM_LTR_AC_LATENCY_VO,
+	/* LTR AC latency timeout, in ms */
+	WMI_PDEV_PARAM_LTR_AC_LATENCY_TIMEOUT,
+	/* LTR platform latency override, in us */
+	WMI_PDEV_PARAM_LTR_SLEEP_OVERRIDE,
+	/* LTR-RX override, in us */
+	WMI_PDEV_PARAM_LTR_RX_OVERRIDE,
+	/* Tx activity timeout for LTR, in us */
+	WMI_PDEV_PARAM_LTR_TX_ACTIVITY_TIMEOUT,
+	/* L1SS state machine enable */
+	WMI_PDEV_PARAM_L1SS_ENABLE,
+	/* Deep sleep state machine enable */
+	WMI_PDEV_PARAM_DSLEEP_ENABLE,
+	/* RX buffering flush enable */
+	WMI_PDEV_PARAM_PCIELP_TXBUF_FLUSH,
+	/* RX buffering matermark */
+	WMI_PDEV_PARAM_PCIELP_TXBUF_WATERMARK,
+	/* RX buffering timeout enable */
+	WMI_PDEV_PARAM_PCIELP_TXBUF_TMO_EN,
+	/* RX buffering timeout value */
+	WMI_PDEV_PARAM_PCIELP_TXBUF_TMO_VALUE,
+	/* pdev level stats update period in ms */
+	WMI_PDEV_PARAM_PDEV_STATS_UPDATE_PERIOD,
+	/* vdev level stats update period in ms */
+	WMI_PDEV_PARAM_VDEV_STATS_UPDATE_PERIOD,
+	/* peer level stats update period in ms */
+	WMI_PDEV_PARAM_PEER_STATS_UPDATE_PERIOD,
+	/* beacon filter status update period */
+	WMI_PDEV_PARAM_BCNFLT_STATS_UPDATE_PERIOD,
+	/* QOS Mgmt frame protection MFP/PMF 0: disable, 1: enable */
+	WMI_PDEV_PARAM_PMF_QOS,
+	/* Access category on which ARP frames are sent */
+	WMI_PDEV_PARAM_ARP_AC_OVERRIDE,
+	/* DCS configuration */
+	WMI_PDEV_PARAM_DCS,
+	/* Enable/Disable ANI on target */
+	WMI_PDEV_PARAM_ANI_ENABLE,
+	/* configure the ANI polling period */
+	WMI_PDEV_PARAM_ANI_POLL_PERIOD,
+	/* configure the ANI listening period */
+	WMI_PDEV_PARAM_ANI_LISTEN_PERIOD,
+	/* configure OFDM immunity level */
+	WMI_PDEV_PARAM_ANI_OFDM_LEVEL,
+	/* configure CCK immunity level */
+	WMI_PDEV_PARAM_ANI_CCK_LEVEL,
+	/* Enable/Disable CDD for 1x1 STAs in rate control module */
+	WMI_PDEV_PARAM_DYNTXCHAIN,
+	/* Enable/Disable proxy STA */
+	WMI_PDEV_PARAM_PROXY_STA,
+	/* Enable/Disable low power state when all VDEVs are inactive/idle. */
+	WMI_PDEV_PARAM_IDLE_PS_CONFIG,
+	/* Enable/Disable power gating sleep */
+	WMI_PDEV_PARAM_POWER_GATING_SLEEP,
+};
+
+struct wmi_pdev_set_param_cmd {
+	__le32 param_id;
+	__le32 param_value;
+} __packed;
+
+struct wmi_pdev_get_tpc_config_cmd {
+	/* parameter   */
+	__le32 param;
+} __packed;
+
+#define WMI_TPC_RATE_MAX		160
+#define WMI_TPC_TX_N_CHAIN		4
+
+enum wmi_tpc_config_event_flag {
+	WMI_TPC_CONFIG_EVENT_FLAG_TABLE_CDD	= 0x1,
+	WMI_TPC_CONFIG_EVENT_FLAG_TABLE_STBC	= 0x2,
+	WMI_TPC_CONFIG_EVENT_FLAG_TABLE_TXBF	= 0x4,
+};
+
+struct wmi_pdev_tpc_config_event {
+	__le32 reg_domain;
+	__le32 chan_freq;
+	__le32 phy_mode;
+	__le32 twice_antenna_reduction;
+	__le32 twice_max_rd_power;
+	s32 twice_antenna_gain;
+	__le32 power_limit;
+	__le32 rate_max;
+	__le32 num_tx_chain;
+	__le32 ctl;
+	__le32 flags;
+	s8 max_reg_allow_pow[WMI_TPC_TX_N_CHAIN];
+	s8 max_reg_allow_pow_agcdd[WMI_TPC_TX_N_CHAIN][WMI_TPC_TX_N_CHAIN];
+	s8 max_reg_allow_pow_agstbc[WMI_TPC_TX_N_CHAIN][WMI_TPC_TX_N_CHAIN];
+	s8 max_reg_allow_pow_agtxbf[WMI_TPC_TX_N_CHAIN][WMI_TPC_TX_N_CHAIN];
+	u8 rates_array[WMI_TPC_RATE_MAX];
+} __packed;
+
+/* Transmit power scale factor. */
+enum wmi_tp_scale {
+	WMI_TP_SCALE_MAX    = 0,	/* no scaling (default) */
+	WMI_TP_SCALE_50     = 1,	/* 50% of max (-3 dBm) */
+	WMI_TP_SCALE_25     = 2,	/* 25% of max (-6 dBm) */
+	WMI_TP_SCALE_12     = 3,	/* 12% of max (-9 dBm) */
+	WMI_TP_SCALE_MIN    = 4,	/* min, but still on   */
+	WMI_TP_SCALE_SIZE   = 5,	/* max num of enum     */
+};
+
+struct wmi_set_channel_cmd {
+	/* channel (only frequency and mode info are used) */
+	struct wmi_channel chan;
+} __packed;
+
+struct wmi_pdev_chanlist_update_event {
+	/* number of channels */
+	__le32 num_chan;
+	/* array of channels */
+	struct wmi_channel channel_list[1];
+} __packed;
+
+#define WMI_MAX_DEBUG_MESG (sizeof(u32) * 32)
+
+struct wmi_debug_mesg_event {
+	/* message buffer, NULL terminated */
+	char bufp[WMI_MAX_DEBUG_MESG];
+} __packed;
+
+enum {
+	/* P2P device */
+	VDEV_SUBTYPE_P2PDEV = 0,
+	/* P2P client */
+	VDEV_SUBTYPE_P2PCLI,
+	/* P2P GO */
+	VDEV_SUBTYPE_P2PGO,
+	/* BT3.0 HS */
+	VDEV_SUBTYPE_BT,
+};
+
+struct wmi_pdev_set_channel_cmd {
+	/* idnore power , only use flags , mode and freq */
+	struct wmi_channel chan;
+} __packed;
+
+/* Customize the DSCP (bit) to TID (0-7) mapping for QOS */
+#define WMI_DSCP_MAP_MAX    (64)
+struct wmi_pdev_set_dscp_tid_map_cmd {
+	/* map indicating DSCP to TID conversion */
+	__le32 dscp_to_tid_map[WMI_DSCP_MAP_MAX];
+} __packed;
+
+enum mcast_bcast_rate_id {
+	WMI_SET_MCAST_RATE,
+	WMI_SET_BCAST_RATE
+};
+
+struct mcast_bcast_rate {
+	enum mcast_bcast_rate_id rate_id;
+	__le32 rate;
+} __packed;
+
+struct wmi_wmm_params {
+	__le32 cwmin;
+	__le32 cwmax;
+	__le32 aifs;
+	__le32 txop;
+	__le32 acm;
+	__le32 no_ack;
+} __packed;
+
+struct wmi_pdev_set_wmm_params {
+	struct wmi_wmm_params ac_be;
+	struct wmi_wmm_params ac_bk;
+	struct wmi_wmm_params ac_vi;
+	struct wmi_wmm_params ac_vo;
+} __packed;
+
+struct wmi_wmm_params_arg {
+	u32 cwmin;
+	u32 cwmax;
+	u32 aifs;
+	u32 txop;
+	u32 acm;
+	u32 no_ack;
+};
+
+struct wmi_pdev_set_wmm_params_arg {
+	struct wmi_wmm_params_arg ac_be;
+	struct wmi_wmm_params_arg ac_bk;
+	struct wmi_wmm_params_arg ac_vi;
+	struct wmi_wmm_params_arg ac_vo;
+};
+
+struct wal_dbg_tx_stats {
+	/* Num HTT cookies queued to dispatch list */
+	__le32 comp_queued;
+
+	/* Num HTT cookies dispatched */
+	__le32 comp_delivered;
+
+	/* Num MSDU queued to WAL */
+	__le32 msdu_enqued;
+
+	/* Num MPDU queue to WAL */
+	__le32 mpdu_enqued;
+
+	/* Num MSDUs dropped by WMM limit */
+	__le32 wmm_drop;
+
+	/* Num Local frames queued */
+	__le32 local_enqued;
+
+	/* Num Local frames done */
+	__le32 local_freed;
+
+	/* Num queued to HW */
+	__le32 hw_queued;
+
+	/* Num PPDU reaped from HW */
+	__le32 hw_reaped;
+
+	/* Num underruns */
+	__le32 underrun;
+
+	/* Num PPDUs cleaned up in TX abort */
+	__le32 tx_abort;
+
+	/* Num MPDUs requed by SW */
+	__le32 mpdus_requed;
+
+	/* excessive retries */
+	__le32 tx_ko;
+
+	/* data hw rate code */
+	__le32 data_rc;
+
+	/* Scheduler self triggers */
+	__le32 self_triggers;
+
+	/* frames dropped due to excessive sw retries */
+	__le32 sw_retry_failure;
+
+	/* illegal rate phy errors  */
+	__le32 illgl_rate_phy_err;
+
+	/* wal pdev continous xretry */
+	__le32 pdev_cont_xretry;
+
+	/* wal pdev continous xretry */
+	__le32 pdev_tx_timeout;
+
+	/* wal pdev resets  */
+	__le32 pdev_resets;
+
+	__le32 phy_underrun;
+
+	/* MPDU is more than txop limit */
+	__le32 txop_ovf;
+} __packed;
+
+struct wal_dbg_rx_stats {
+	/* Cnts any change in ring routing mid-ppdu */
+	__le32 mid_ppdu_route_change;
+
+	/* Total number of statuses processed */
+	__le32 status_rcvd;
+
+	/* Extra frags on rings 0-3 */
+	__le32 r0_frags;
+	__le32 r1_frags;
+	__le32 r2_frags;
+	__le32 r3_frags;
+
+	/* MSDUs / MPDUs delivered to HTT */
+	__le32 htt_msdus;
+	__le32 htt_mpdus;
+
+	/* MSDUs / MPDUs delivered to local stack */
+	__le32 loc_msdus;
+	__le32 loc_mpdus;
+
+	/* AMSDUs that have more MSDUs than the status ring size */
+	__le32 oversize_amsdu;
+
+	/* Number of PHY errors */
+	__le32 phy_errs;
+
+	/* Number of PHY errors drops */
+	__le32 phy_err_drop;
+
+	/* Number of mpdu errors - FCS, MIC, ENC etc. */
+	__le32 mpdu_errs;
+} __packed;
+
+struct wal_dbg_peer_stats {
+	/* REMOVE THIS ONCE REAL PEER STAT COUNTERS ARE ADDED */
+	__le32 dummy;
+} __packed;
+
+struct wal_dbg_stats {
+	struct wal_dbg_tx_stats tx;
+	struct wal_dbg_rx_stats rx;
+	struct wal_dbg_peer_stats peer;
+} __packed;
+
+enum wmi_stats_id {
+	WMI_REQUEST_PEER_STAT	= 0x01,
+	WMI_REQUEST_AP_STAT	= 0x02
+};
+
+struct wmi_request_stats_cmd {
+	__le32 stats_id;
+
+	/*
+	 * Space to add parameters like
+	 * peer mac addr
+	 */
+} __packed;
+
+/* Suspend option */
+enum {
+	/* suspend */
+	WMI_PDEV_SUSPEND,
+
+	/* suspend and disable all interrupts */
+	WMI_PDEV_SUSPEND_AND_DISABLE_INTR,
+};
+
+struct wmi_pdev_suspend_cmd {
+	/* suspend option sent to target */
+	__le32 suspend_opt;
+} __packed;
+
+struct wmi_stats_event {
+	__le32 stats_id; /* %WMI_REQUEST_ */
+	/*
+	 * number of pdev stats event structures
+	 * (wmi_pdev_stats) 0 or 1
+	 */
+	__le32 num_pdev_stats;
+	/*
+	 * number of vdev stats event structures
+	 * (wmi_vdev_stats) 0 or max vdevs
+	 */
+	__le32 num_vdev_stats;
+	/*
+	 * number of peer stats event structures
+	 * (wmi_peer_stats) 0 or max peers
+	 */
+	__le32 num_peer_stats;
+	__le32 num_bcnflt_stats;
+	/*
+	 * followed by
+	 *   num_pdev_stats * size of(struct wmi_pdev_stats)
+	 *   num_vdev_stats * size of(struct wmi_vdev_stats)
+	 *   num_peer_stats * size of(struct wmi_peer_stats)
+	 *
+	 *  By having a zero sized array, the pointer to data area
+	 *  becomes available without increasing the struct size
+	 */
+	u8 data[0];
+} __packed;
+
+/*
+ * PDEV statistics
+ * TODO: add all PDEV stats here
+ */
+struct wmi_pdev_stats {
+	__le32 chan_nf;        /* Channel noise floor */
+	__le32 tx_frame_count; /* TX frame count */
+	__le32 rx_frame_count; /* RX frame count */
+	__le32 rx_clear_count; /* rx clear count */
+	__le32 cycle_count;    /* cycle count */
+	__le32 phy_err_count;  /* Phy error count */
+	__le32 chan_tx_pwr;    /* channel tx power */
+	struct wal_dbg_stats wal; /* WAL dbg stats */
+} __packed;
+
+/*
+ * VDEV statistics
+ * TODO: add all VDEV stats here
+ */
+struct wmi_vdev_stats {
+	__le32 vdev_id;
+} __packed;
+
+/*
+ * peer statistics.
+ * TODO: add more stats
+ */
+struct wmi_peer_stats {
+	struct wmi_mac_addr peer_macaddr;
+	__le32 peer_rssi;
+	__le32 peer_tx_rate;
+} __packed;
+
+struct wmi_vdev_create_cmd {
+	__le32 vdev_id;
+	__le32 vdev_type;
+	__le32 vdev_subtype;
+	struct wmi_mac_addr vdev_macaddr;
+} __packed;
+
+enum wmi_vdev_type {
+	WMI_VDEV_TYPE_AP      = 1,
+	WMI_VDEV_TYPE_STA     = 2,
+	WMI_VDEV_TYPE_IBSS    = 3,
+	WMI_VDEV_TYPE_MONITOR = 4,
+};
+
+enum wmi_vdev_subtype {
+	WMI_VDEV_SUBTYPE_NONE       = 0,
+	WMI_VDEV_SUBTYPE_P2P_DEVICE = 1,
+	WMI_VDEV_SUBTYPE_P2P_CLIENT = 2,
+	WMI_VDEV_SUBTYPE_P2P_GO     = 3,
+};
+
+/* values for vdev_subtype */
+
+/* values for vdev_start_request flags */
+/*
+ * Indicates that AP VDEV uses hidden ssid. only valid for
+ *  AP/GO */
+#define WMI_VDEV_START_HIDDEN_SSID  (1<<0)
+/*
+ * Indicates if robust management frame/management frame
+ *  protection is enabled. For GO/AP vdevs, it indicates that
+ *  it may support station/client associations with RMF enabled.
+ *  For STA/client vdevs, it indicates that sta will
+ *  associate with AP with RMF enabled. */
+#define WMI_VDEV_START_PMF_ENABLED  (1<<1)
+
+struct wmi_p2p_noa_descriptor {
+	__le32 type_count; /* 255: continuous schedule, 0: reserved */
+	__le32 duration;  /* Absent period duration in micro seconds */
+	__le32 interval;   /* Absent period interval in micro seconds */
+	__le32 start_time; /* 32 bit tsf time when in starts */
+} __packed;
+
+struct wmi_vdev_start_request_cmd {
+	/* WMI channel */
+	struct wmi_channel chan;
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* requestor id identifying the caller module */
+	__le32 requestor_id;
+	/* beacon interval from received beacon */
+	__le32 beacon_interval;
+	/* DTIM Period from the received beacon */
+	__le32 dtim_period;
+	/* Flags */
+	__le32 flags;
+	/* ssid field. Only valid for AP/GO/IBSS/BTAmp VDEV type. */
+	struct wmi_ssid ssid;
+	/* beacon/probe reponse xmit rate. Applicable for SoftAP. */
+	__le32 bcn_tx_rate;
+	/* beacon/probe reponse xmit power. Applicable for SoftAP. */
+	__le32 bcn_tx_power;
+	/* number of p2p NOA descriptor(s) from scan entry */
+	__le32 num_noa_descriptors;
+	/*
+	 * Disable H/W ack. This used by WMI_VDEV_RESTART_REQUEST_CMDID.
+	 * During CAC, Our HW shouldn't ack ditected frames
+	 */
+	__le32 disable_hw_ack;
+	/* actual p2p NOA descriptor from scan entry */
+	struct wmi_p2p_noa_descriptor noa_descriptors[2];
+} __packed;
+
+struct wmi_vdev_restart_request_cmd {
+	struct wmi_vdev_start_request_cmd vdev_start_request_cmd;
+} __packed;
+
+struct wmi_vdev_start_request_arg {
+	u32 vdev_id;
+	struct wmi_channel_arg channel;
+	u32 bcn_intval;
+	u32 dtim_period;
+	u8 *ssid;
+	u32 ssid_len;
+	u32 bcn_tx_rate;
+	u32 bcn_tx_power;
+	bool disable_hw_ack;
+	bool hidden_ssid;
+	bool pmf_enabled;
+};
+
+struct wmi_vdev_delete_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_up_cmd {
+	__le32 vdev_id;
+	__le32 vdev_assoc_id;
+	struct wmi_mac_addr vdev_bssid;
+} __packed;
+
+struct wmi_vdev_stop_cmd {
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_down_cmd {
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_standby_response_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_resume_response_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_set_param_cmd {
+	__le32 vdev_id;
+	__le32 param_id;
+	__le32 param_value;
+} __packed;
+
+#define WMI_MAX_KEY_INDEX   3
+#define WMI_MAX_KEY_LEN     32
+
+#define WMI_KEY_PAIRWISE 0x00
+#define WMI_KEY_GROUP    0x01
+#define WMI_KEY_TX_USAGE 0x02 /* default tx key - static wep */
+
+struct wmi_key_seq_counter {
+	__le32 key_seq_counter_l;
+	__le32 key_seq_counter_h;
+} __packed;
+
+#define WMI_CIPHER_NONE     0x0 /* clear key */
+#define WMI_CIPHER_WEP      0x1
+#define WMI_CIPHER_TKIP     0x2
+#define WMI_CIPHER_AES_OCB  0x3
+#define WMI_CIPHER_AES_CCM  0x4
+#define WMI_CIPHER_WAPI     0x5
+#define WMI_CIPHER_CKIP     0x6
+#define WMI_CIPHER_AES_CMAC 0x7
+
+struct wmi_vdev_install_key_cmd {
+	__le32 vdev_id;
+	struct wmi_mac_addr peer_macaddr;
+	__le32 key_idx;
+	__le32 key_flags;
+	__le32 key_cipher; /* %WMI_CIPHER_ */
+	struct wmi_key_seq_counter key_rsc_counter;
+	struct wmi_key_seq_counter key_global_rsc_counter;
+	struct wmi_key_seq_counter key_tsc_counter;
+	u8 wpi_key_rsc_counter[16];
+	u8 wpi_key_tsc_counter[16];
+	__le32 key_len;
+	__le32 key_txmic_len;
+	__le32 key_rxmic_len;
+
+	/* contains key followed by tx mic followed by rx mic */
+	u8 key_data[0];
+} __packed;
+
+struct wmi_vdev_install_key_arg {
+	u32 vdev_id;
+	const u8 *macaddr;
+	u32 key_idx;
+	u32 key_flags;
+	u32 key_cipher;
+	u32 key_len;
+	u32 key_txmic_len;
+	u32 key_rxmic_len;
+	const void *key_data;
+};
+
+/* Preamble types to be used with VDEV fixed rate configuration */
+enum wmi_rate_preamble {
+	WMI_RATE_PREAMBLE_OFDM,
+	WMI_RATE_PREAMBLE_CCK,
+	WMI_RATE_PREAMBLE_HT,
+	WMI_RATE_PREAMBLE_VHT,
+};
+
+/* Value to disable fixed rate setting */
+#define WMI_FIXED_RATE_NONE    (0xff)
+
+/* the definition of different VDEV parameters */
+enum wmi_vdev_param {
+	/* RTS Threshold */
+	WMI_VDEV_PARAM_RTS_THRESHOLD = 0x1,
+	/* Fragmentation threshold */
+	WMI_VDEV_PARAM_FRAGMENTATION_THRESHOLD,
+	/* beacon interval in TUs */
+	WMI_VDEV_PARAM_BEACON_INTERVAL,
+	/* Listen interval in TUs */
+	WMI_VDEV_PARAM_LISTEN_INTERVAL,
+	/* muticast rate in Mbps */
+	WMI_VDEV_PARAM_MULTICAST_RATE,
+	/* management frame rate in Mbps */
+	WMI_VDEV_PARAM_MGMT_TX_RATE,
+	/* slot time (long vs short) */
+	WMI_VDEV_PARAM_SLOT_TIME,
+	/* preamble (long vs short) */
+	WMI_VDEV_PARAM_PREAMBLE,
+	/* SWBA time (time before tbtt in msec) */
+	WMI_VDEV_PARAM_SWBA_TIME,
+	/* time period for updating VDEV stats */
+	WMI_VDEV_STATS_UPDATE_PERIOD,
+	/* age out time in msec for frames queued for station in power save */
+	WMI_VDEV_PWRSAVE_AGEOUT_TIME,
+	/*
+	 * Host SWBA interval (time in msec before tbtt for SWBA event
+	 * generation).
+	 */
+	WMI_VDEV_HOST_SWBA_INTERVAL,
+	/* DTIM period (specified in units of num beacon intervals) */
+	WMI_VDEV_PARAM_DTIM_PERIOD,
+	/*
+	 * scheduler air time limit for this VDEV. used by off chan
+	 * scheduler.
+	 */
+	WMI_VDEV_OC_SCHEDULER_AIR_TIME_LIMIT,
+	/* enable/dsiable WDS for this VDEV  */
+	WMI_VDEV_PARAM_WDS,
+	/* ATIM Window */
+	WMI_VDEV_PARAM_ATIM_WINDOW,
+	/* BMISS max */
+	WMI_VDEV_PARAM_BMISS_COUNT_MAX,
+	/* BMISS first time */
+	WMI_VDEV_PARAM_BMISS_FIRST_BCNT,
+	/* BMISS final time */
+	WMI_VDEV_PARAM_BMISS_FINAL_BCNT,
+	/* WMM enables/disabled */
+	WMI_VDEV_PARAM_FEATURE_WMM,
+	/* Channel width */
+	WMI_VDEV_PARAM_CHWIDTH,
+	/* Channel Offset */
+	WMI_VDEV_PARAM_CHEXTOFFSET,
+	/* Disable HT Protection */
+	WMI_VDEV_PARAM_DISABLE_HTPROTECTION,
+	/* Quick STA Kickout */
+	WMI_VDEV_PARAM_STA_QUICKKICKOUT,
+	/* Rate to be used with Management frames */
+	WMI_VDEV_PARAM_MGMT_RATE,
+	/* Protection Mode */
+	WMI_VDEV_PARAM_PROTECTION_MODE,
+	/* Fixed rate setting */
+	WMI_VDEV_PARAM_FIXED_RATE,
+	/* Short GI Enable/Disable */
+	WMI_VDEV_PARAM_SGI,
+	/* Enable LDPC */
+	WMI_VDEV_PARAM_LDPC,
+	/* Enable Tx STBC */
+	WMI_VDEV_PARAM_TX_STBC,
+	/* Enable Rx STBC */
+	WMI_VDEV_PARAM_RX_STBC,
+	/* Intra BSS forwarding  */
+	WMI_VDEV_PARAM_INTRA_BSS_FWD,
+	/* Setting Default xmit key for Vdev */
+	WMI_VDEV_PARAM_DEF_KEYID,
+	/* NSS width */
+	WMI_VDEV_PARAM_NSS,
+	/* Set the custom rate for the broadcast data frames */
+	WMI_VDEV_PARAM_BCAST_DATA_RATE,
+	/* Set the custom rate (rate-code) for multicast data frames */
+	WMI_VDEV_PARAM_MCAST_DATA_RATE,
+	/* Tx multicast packet indicate Enable/Disable */
+	WMI_VDEV_PARAM_MCAST_INDICATE,
+	/* Tx DHCP packet indicate Enable/Disable */
+	WMI_VDEV_PARAM_DHCP_INDICATE,
+	/* Enable host inspection of Tx unicast packet to unknown destination */
+	WMI_VDEV_PARAM_UNKNOWN_DEST_INDICATE,
+
+	/* The minimum amount of time AP begins to consider STA inactive */
+	WMI_VDEV_PARAM_AP_KEEPALIVE_MIN_IDLE_INACTIVE_TIME_SECS,
+
+	/*
+	 * An associated STA is considered inactive when there is no recent
+	 * TX/RX activity and no downlink frames are buffered for it. Once a
+	 * STA exceeds the maximum idle inactive time, the AP will send an
+	 * 802.11 data-null as a keep alive to verify the STA is still
+	 * associated. If the STA does ACK the data-null, or if the data-null
+	 * is buffered and the STA does not retrieve it, the STA will be
+	 * considered unresponsive
+	 * (see WMI_VDEV_AP_KEEPALIVE_MAX_UNRESPONSIVE_TIME_SECS).
+	 */
+	WMI_VDEV_PARAM_AP_KEEPALIVE_MAX_IDLE_INACTIVE_TIME_SECS,
+
+	/*
+	 * An associated STA is considered unresponsive if there is no recent
+	 * TX/RX activity and downlink frames are buffered for it. Once a STA
+	 * exceeds the maximum unresponsive time, the AP will send a
+	 * WMI_STA_KICKOUT event to the host so the STA can be deleted. */
+	WMI_VDEV_PARAM_AP_KEEPALIVE_MAX_UNRESPONSIVE_TIME_SECS,
+
+	/* Enable NAWDS : MCAST INSPECT Enable, NAWDS Flag set */
+	WMI_VDEV_PARAM_AP_ENABLE_NAWDS,
+	/* Enable/Disable RTS-CTS */
+	WMI_VDEV_PARAM_ENABLE_RTSCTS,
+	/* Enable TXBFee/er */
+	WMI_VDEV_PARAM_TXBF,
+
+	/* Set packet power save */
+	WMI_VDEV_PARAM_PACKET_POWERSAVE,
+
+	/*
+	 * Drops un-encrypted packets if eceived in an encrypted connection
+	 * otherwise forwards to host.
+	 */
+	WMI_VDEV_PARAM_DROP_UNENCRY,
+
+	/*
+	 * Set the encapsulation type for frames.
+	 */
+	WMI_VDEV_PARAM_TX_ENCAP_TYPE,
+};
+
+/* slot time long */
+#define WMI_VDEV_SLOT_TIME_LONG		0x1
+/* slot time short */
+#define WMI_VDEV_SLOT_TIME_SHORT	0x2
+/* preablbe long */
+#define WMI_VDEV_PREAMBLE_LONG		0x1
+/* preablbe short */
+#define WMI_VDEV_PREAMBLE_SHORT		0x2
+
+enum wmi_start_event_param {
+	WMI_VDEV_RESP_START_EVENT = 0,
+	WMI_VDEV_RESP_RESTART_EVENT,
+};
+
+struct wmi_vdev_start_response_event {
+	__le32 vdev_id;
+	__le32 req_id;
+	__le32 resp_type; /* %WMI_VDEV_RESP_ */
+	__le32 status;
+} __packed;
+
+struct wmi_vdev_standby_req_event {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_resume_req_event {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+struct wmi_vdev_stopped_event {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+/*
+ * common structure used for simple events
+ * (stopped, resume_req, standby response)
+ */
+struct wmi_vdev_simple_event {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+} __packed;
+
+/* VDEV start response status codes */
+/* VDEV succesfully started */
+#define WMI_INIFIED_VDEV_START_RESPONSE_STATUS_SUCCESS	0x0
+
+/* requested VDEV not found */
+#define WMI_INIFIED_VDEV_START_RESPONSE_INVALID_VDEVID	0x1
+
+/* unsupported VDEV combination */
+#define WMI_INIFIED_VDEV_START_RESPONSE_NOT_SUPPORTED	0x2
+
+/* Beacon processing related command and event structures */
+struct wmi_bcn_tx_hdr {
+	__le32 vdev_id;
+	__le32 tx_rate;
+	__le32 tx_power;
+	__le32 bcn_len;
+} __packed;
+
+struct wmi_bcn_tx_cmd {
+	struct wmi_bcn_tx_hdr hdr;
+	u8 *bcn[0];
+} __packed;
+
+struct wmi_bcn_tx_arg {
+	u32 vdev_id;
+	u32 tx_rate;
+	u32 tx_power;
+	u32 bcn_len;
+	const void *bcn;
+};
+
+/* Beacon filter */
+#define WMI_BCN_FILTER_ALL   0 /* Filter all beacons */
+#define WMI_BCN_FILTER_NONE  1 /* Pass all beacons */
+#define WMI_BCN_FILTER_RSSI  2 /* Pass Beacons RSSI >= RSSI threshold */
+#define WMI_BCN_FILTER_BSSID 3 /* Pass Beacons with matching BSSID */
+#define WMI_BCN_FILTER_SSID  4 /* Pass Beacons with matching SSID */
+
+struct wmi_bcn_filter_rx_cmd {
+	/* Filter ID */
+	__le32 bcn_filter_id;
+	/* Filter type - wmi_bcn_filter */
+	__le32 bcn_filter;
+	/* Buffer len */
+	__le32 bcn_filter_len;
+	/* Filter info (threshold, BSSID, RSSI) */
+	u8 *bcn_filter_buf;
+} __packed;
+
+/* Capabilities and IEs to be passed to firmware */
+struct wmi_bcn_prb_info {
+	/* Capabilities */
+	__le32 caps;
+	/* ERP info */
+	__le32 erp;
+	/* Advanced capabilities */
+	/* HT capabilities */
+	/* HT Info */
+	/* ibss_dfs */
+	/* wpa Info */
+	/* rsn Info */
+	/* rrm info */
+	/* ath_ext */
+	/* app IE */
+} __packed;
+
+struct wmi_bcn_tmpl_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* TIM IE offset from the beginning of the template. */
+	__le32 tim_ie_offset;
+	/* beacon probe capabilities and IEs */
+	struct wmi_bcn_prb_info bcn_prb_info;
+	/* beacon buffer length */
+	__le32 buf_len;
+	/* variable length data */
+	u8 data[1];
+} __packed;
+
+struct wmi_prb_tmpl_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* beacon probe capabilities and IEs */
+	struct wmi_bcn_prb_info bcn_prb_info;
+	/* beacon buffer length */
+	__le32 buf_len;
+	/* Variable length data */
+	u8 data[1];
+} __packed;
+
+enum wmi_sta_ps_mode {
+	/* enable power save for the given STA VDEV */
+	WMI_STA_PS_MODE_DISABLED = 0,
+	/* disable power save  for a given STA VDEV */
+	WMI_STA_PS_MODE_ENABLED = 1,
+};
+
+struct wmi_sta_powersave_mode_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+
+	/*
+	 * Power save mode
+	 * (see enum wmi_sta_ps_mode)
+	 */
+	__le32 sta_ps_mode;
+} __packed;
+
+enum wmi_csa_offload_en {
+	WMI_CSA_OFFLOAD_DISABLE = 0,
+	WMI_CSA_OFFLOAD_ENABLE = 1,
+};
+
+struct wmi_csa_offload_enable_cmd {
+	__le32 vdev_id;
+	__le32 csa_offload_enable;
+} __packed;
+
+struct wmi_csa_offload_chanswitch_cmd {
+	__le32 vdev_id;
+	struct wmi_channel chan;
+} __packed;
+
+/*
+ * This parameter controls the policy for retrieving frames from AP while the
+ * STA is in sleep state.
+ *
+ * Only takes affect if the sta_ps_mode is enabled
+ */
+enum wmi_sta_ps_param_rx_wake_policy {
+	/*
+	 * Wake up when ever there is an  RX activity on the VDEV. In this mode
+	 * the Power save SM(state machine) will come out of sleep by either
+	 * sending null frame (or) a data frame (with PS==0) in response to TIM
+	 * bit set in the received beacon frame from AP.
+	 */
+	WMI_STA_PS_RX_WAKE_POLICY_WAKE = 0,
+
+	/*
+	 * Here the power save state machine will not wakeup in response to TIM
+	 * bit, instead it will send a PSPOLL (or) UASPD trigger based on UAPSD
+	 * configuration setup by WMISET_PS_SET_UAPSD  WMI command.  When all
+	 * access categories are delivery-enabled, the station will send a
+	 * UAPSD trigger frame, otherwise it will send a PS-Poll.
+	 */
+	WMI_STA_PS_RX_WAKE_POLICY_POLL_UAPSD = 1,
+};
+
+/*
+ * Number of tx frames/beacon  that cause the power save SM to wake up.
+ *
+ * Value 1 causes the SM to wake up for every TX. Value 0 has a special
+ * meaning, It will cause the SM to never wake up. This is useful if you want
+ * to keep the system to sleep all the time for some kind of test mode . host
+ * can change this parameter any time.  It will affect at the next tx frame.
+ */
+enum wmi_sta_ps_param_tx_wake_threshold {
+	WMI_STA_PS_TX_WAKE_THRESHOLD_NEVER = 0,
+	WMI_STA_PS_TX_WAKE_THRESHOLD_ALWAYS = 1,
+
+	/*
+	 * Values greater than one indicate that many TX attempts per beacon
+	 * interval before the STA will wake up
+	 */
+};
+
+/*
+ * The maximum number of PS-Poll frames the FW will send in response to
+ * traffic advertised in TIM before waking up (by sending a null frame with PS
+ * = 0). Value 0 has a special meaning: there is no maximum count and the FW
+ * will send as many PS-Poll as are necessary to retrieve buffered BU. This
+ * parameter is used when the RX wake policy is
+ * WMI_STA_PS_RX_WAKE_POLICY_POLL_UAPSD and ignored when the RX wake
+ * policy is WMI_STA_PS_RX_WAKE_POLICY_WAKE.
+ */
+enum wmi_sta_ps_param_pspoll_count {
+	WMI_STA_PS_PSPOLL_COUNT_NO_MAX = 0,
+	/*
+	 * Values greater than 0 indicate the maximum numer of PS-Poll frames
+	 * FW will send before waking up.
+	 */
+};
+
+/*
+ * This will include the delivery and trigger enabled state for every AC.
+ * This is the negotiated state with AP. The host MLME needs to set this based
+ * on AP capability and the state Set in the association request by the
+ * station MLME.Lower 8 bits of the value specify the UAPSD configuration.
+ */
+#define WMI_UAPSD_AC_TYPE_DELI 0
+#define WMI_UAPSD_AC_TYPE_TRIG 1
+
+#define WMI_UAPSD_AC_BIT_MASK(ac, type) \
+	((type ==  WMI_UAPSD_AC_TYPE_DELI) ? (1<<(ac<<1)) : (1<<((ac<<1)+1)))
+
+enum wmi_sta_ps_param_uapsd {
+	WMI_STA_PS_UAPSD_AC0_DELIVERY_EN = (1 << 0),
+	WMI_STA_PS_UAPSD_AC0_TRIGGER_EN  = (1 << 1),
+	WMI_STA_PS_UAPSD_AC1_DELIVERY_EN = (1 << 2),
+	WMI_STA_PS_UAPSD_AC1_TRIGGER_EN  = (1 << 3),
+	WMI_STA_PS_UAPSD_AC2_DELIVERY_EN = (1 << 4),
+	WMI_STA_PS_UAPSD_AC2_TRIGGER_EN  = (1 << 5),
+	WMI_STA_PS_UAPSD_AC3_DELIVERY_EN = (1 << 6),
+	WMI_STA_PS_UAPSD_AC3_TRIGGER_EN  = (1 << 7),
+};
+
+enum wmi_sta_powersave_param {
+	/*
+	 * Controls how frames are retrievd from AP while STA is sleeping
+	 *
+	 * (see enum wmi_sta_ps_param_rx_wake_policy)
+	 */
+	WMI_STA_PS_PARAM_RX_WAKE_POLICY = 0,
+
+	/*
+	 * The STA will go active after this many TX
+	 *
+	 * (see enum wmi_sta_ps_param_tx_wake_threshold)
+	 */
+	WMI_STA_PS_PARAM_TX_WAKE_THRESHOLD = 1,
+
+	/*
+	 * Number of PS-Poll to send before STA wakes up
+	 *
+	 * (see enum wmi_sta_ps_param_pspoll_count)
+	 *
+	 */
+	WMI_STA_PS_PARAM_PSPOLL_COUNT = 2,
+
+	/*
+	 * TX/RX inactivity time in msec before going to sleep.
+	 *
+	 * The power save SM will monitor tx/rx activity on the VDEV, if no
+	 * activity for the specified msec of the parameter the Power save
+	 * SM will go to sleep.
+	 */
+	WMI_STA_PS_PARAM_INACTIVITY_TIME = 3,
+
+	/*
+	 * Set uapsd configuration.
+	 *
+	 * (see enum wmi_sta_ps_param_uapsd)
+	 */
+	WMI_STA_PS_PARAM_UAPSD = 4,
+};
+
+struct wmi_sta_powersave_param_cmd {
+	__le32 vdev_id;
+	__le32 param_id; /* %WMI_STA_PS_PARAM_ */
+	__le32 param_value;
+} __packed;
+
+/* No MIMO power save */
+#define WMI_STA_MIMO_PS_MODE_DISABLE
+/* mimo powersave mode static*/
+#define WMI_STA_MIMO_PS_MODE_STATIC
+/* mimo powersave mode dynamic */
+#define WMI_STA_MIMO_PS_MODE_DYNAMIC
+
+struct wmi_sta_mimo_ps_mode_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* mimo powersave mode as defined above */
+	__le32 mimo_pwrsave_mode;
+} __packed;
+
+/* U-APSD configuration of peer station from (re)assoc request and TSPECs */
+enum wmi_ap_ps_param_uapsd {
+	WMI_AP_PS_UAPSD_AC0_DELIVERY_EN = (1 << 0),
+	WMI_AP_PS_UAPSD_AC0_TRIGGER_EN  = (1 << 1),
+	WMI_AP_PS_UAPSD_AC1_DELIVERY_EN = (1 << 2),
+	WMI_AP_PS_UAPSD_AC1_TRIGGER_EN  = (1 << 3),
+	WMI_AP_PS_UAPSD_AC2_DELIVERY_EN = (1 << 4),
+	WMI_AP_PS_UAPSD_AC2_TRIGGER_EN  = (1 << 5),
+	WMI_AP_PS_UAPSD_AC3_DELIVERY_EN = (1 << 6),
+	WMI_AP_PS_UAPSD_AC3_TRIGGER_EN  = (1 << 7),
+};
+
+/* U-APSD maximum service period of peer station */
+enum wmi_ap_ps_peer_param_max_sp {
+	WMI_AP_PS_PEER_PARAM_MAX_SP_UNLIMITED = 0,
+	WMI_AP_PS_PEER_PARAM_MAX_SP_2 = 1,
+	WMI_AP_PS_PEER_PARAM_MAX_SP_4 = 2,
+	WMI_AP_PS_PEER_PARAM_MAX_SP_6 = 3,
+	MAX_WMI_AP_PS_PEER_PARAM_MAX_SP,
+};
+
+/*
+ * AP power save parameter
+ * Set a power save specific parameter for a peer station
+ */
+enum wmi_ap_ps_peer_param {
+	/* Set uapsd configuration for a given peer.
+	 *
+	 * Include the delivery and trigger enabled state for every AC.
+	 * The host  MLME needs to set this based on AP capability and stations
+	 * request Set in the association request  received from the station.
+	 *
+	 * Lower 8 bits of the value specify the UAPSD configuration.
+	 *
+	 * (see enum wmi_ap_ps_param_uapsd)
+	 * The default value is 0.
+	 */
+	WMI_AP_PS_PEER_PARAM_UAPSD = 0,
+
+	/*
+	 * Set the service period for a UAPSD capable station
+	 *
+	 * The service period from wme ie in the (re)assoc request frame.
+	 *
+	 * (see enum wmi_ap_ps_peer_param_max_sp)
+	 */
+	WMI_AP_PS_PEER_PARAM_MAX_SP = 1,
+
+	/* Time in seconds for aging out buffered frames for STA in PS */
+	WMI_AP_PS_PEER_PARAM_AGEOUT_TIME = 2,
+};
+
+struct wmi_ap_ps_peer_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+
+	/* AP powersave param (see enum wmi_ap_ps_peer_param) */
+	__le32 param_id;
+
+	/* AP powersave param value */
+	__le32 param_value;
+} __packed;
+
+/* 128 clients = 4 words */
+#define WMI_TIM_BITMAP_ARRAY_SIZE 4
+
+struct wmi_tim_info {
+	__le32 tim_len;
+	__le32 tim_mcast;
+	__le32 tim_bitmap[WMI_TIM_BITMAP_ARRAY_SIZE];
+	__le32 tim_changed;
+	__le32 tim_num_ps_pending;
+} __packed;
+
+/* Maximum number of NOA Descriptors supported */
+#define WMI_P2P_MAX_NOA_DESCRIPTORS 4
+#define WMI_P2P_OPPPS_ENABLE_BIT	BIT(0)
+#define WMI_P2P_OPPPS_CTWINDOW_OFFSET	1
+#define WMI_P2P_NOA_CHANGED_BIT	BIT(0)
+
+struct wmi_p2p_noa_info {
+	/* Bit 0 - Flag to indicate an update in NOA schedule
+	   Bits 7-1 - Reserved */
+	u8 changed;
+	/* NOA index */
+	u8 index;
+	/* Bit 0 - Opp PS state of the AP
+	   Bits 1-7 - Ctwindow in TUs */
+	u8 ctwindow_oppps;
+	/* Number of NOA descriptors */
+	u8 num_descriptors;
+
+	struct wmi_p2p_noa_descriptor descriptors[WMI_P2P_MAX_NOA_DESCRIPTORS];
+} __packed;
+
+struct wmi_bcn_info {
+	struct wmi_tim_info tim_info;
+	struct wmi_p2p_noa_info p2p_noa_info;
+} __packed;
+
+struct wmi_host_swba_event {
+	__le32 vdev_map;
+	struct wmi_bcn_info bcn_info[1];
+} __packed;
+
+#define WMI_MAX_AP_VDEV 16
+
+struct wmi_tbtt_offset_event {
+	__le32 vdev_map;
+	__le32 tbttoffset_list[WMI_MAX_AP_VDEV];
+} __packed;
+
+
+struct wmi_peer_create_cmd {
+	__le32 vdev_id;
+	struct wmi_mac_addr peer_macaddr;
+} __packed;
+
+struct wmi_peer_delete_cmd {
+	__le32 vdev_id;
+	struct wmi_mac_addr peer_macaddr;
+} __packed;
+
+struct wmi_peer_flush_tids_cmd {
+	__le32 vdev_id;
+	struct wmi_mac_addr peer_macaddr;
+	__le32 peer_tid_bitmap;
+} __packed;
+
+struct wmi_fixed_rate {
+	/*
+	 * rate mode . 0: disable fixed rate (auto rate)
+	 *   1: legacy (non 11n) rate  specified as ieee rate 2*Mbps
+	 *   2: ht20 11n rate  specified as mcs index
+	 *   3: ht40 11n rate  specified as mcs index
+	 */
+	__le32  rate_mode;
+	/*
+	 * 4 rate values for 4 rate series. series 0 is stored in byte 0 (LSB)
+	 * and series 3 is stored at byte 3 (MSB)
+	 */
+	__le32  rate_series;
+	/*
+	 * 4 retry counts for 4 rate series. retry count for rate 0 is stored
+	 * in byte 0 (LSB) and retry count for rate 3 is stored at byte 3
+	 * (MSB)
+	 */
+	__le32  rate_retries;
+} __packed;
+
+struct wmi_peer_fixed_rate_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+	/* fixed rate */
+	struct wmi_fixed_rate peer_fixed_rate;
+} __packed;
+
+#define WMI_MGMT_TID    17
+
+struct wmi_addba_clear_resp_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+} __packed;
+
+struct wmi_addba_send_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+	/* Tid number */
+	__le32 tid;
+	/* Buffer/Window size*/
+	__le32 buffersize;
+} __packed;
+
+struct wmi_delba_send_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+	/* Tid number */
+	__le32 tid;
+	/* Is Initiator */
+	__le32 initiator;
+	/* Reason code */
+	__le32 reasoncode;
+} __packed;
+
+struct wmi_addba_setresponse_cmd {
+	/* unique id identifying the vdev, generated by the caller */
+	__le32 vdev_id;
+	/* peer mac address */
+	struct wmi_mac_addr peer_macaddr;
+	/* Tid number */
+	__le32 tid;
+	/* status code */
+	__le32 statuscode;
+} __packed;
+
+struct wmi_send_singleamsdu_cmd {
+	/* unique id identifying the vdev, generated by the caller */
+	__le32 vdev_id;
+	/* peer mac address */
+	struct wmi_mac_addr peer_macaddr;
+	/* Tid number */
+	__le32 tid;
+} __packed;
+
+enum wmi_peer_smps_state {
+	WMI_PEER_SMPS_PS_NONE = 0x0,
+	WMI_PEER_SMPS_STATIC  = 0x1,
+	WMI_PEER_SMPS_DYNAMIC = 0x2
+};
+
+enum wmi_peer_param {
+	WMI_PEER_SMPS_STATE = 0x1, /* see %wmi_peer_smps_state */
+	WMI_PEER_AMPDU      = 0x2,
+	WMI_PEER_AUTHORIZE  = 0x3,
+	WMI_PEER_CHAN_WIDTH = 0x4,
+	WMI_PEER_NSS        = 0x5,
+	WMI_PEER_USE_4ADDR  = 0x6
+};
+
+struct wmi_peer_set_param_cmd {
+	__le32 vdev_id;
+	struct wmi_mac_addr peer_macaddr;
+	__le32 param_id;
+	__le32 param_value;
+} __packed;
+
+#define MAX_SUPPORTED_RATES 128
+
+struct wmi_rate_set {
+	/* total number of rates */
+	__le32 num_rates;
+	/*
+	 * rates (each 8bit value) packed into a 32 bit word.
+	 * the rates are filled from least significant byte to most
+	 * significant byte.
+	 */
+	__le32 rates[(MAX_SUPPORTED_RATES/4)+1];
+} __packed;
+
+struct wmi_rate_set_arg {
+	unsigned int num_rates;
+	u8 rates[MAX_SUPPORTED_RATES];
+};
+
+/*
+ * NOTE: It would bea good idea to represent the Tx MCS
+ * info in one word and Rx in another word. This is split
+ * into multiple words for convenience
+ */
+struct wmi_vht_rate_set {
+	__le32 rx_max_rate; /* Max Rx data rate */
+	__le32 rx_mcs_set;  /* Negotiated RX VHT rates */
+	__le32 tx_max_rate; /* Max Tx data rate */
+	__le32 tx_mcs_set;  /* Negotiated TX VHT rates */
+} __packed;
+
+struct wmi_vht_rate_set_arg {
+	u32 rx_max_rate;
+	u32 rx_mcs_set;
+	u32 tx_max_rate;
+	u32 tx_mcs_set;
+};
+
+struct wmi_peer_set_rates_cmd {
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+	/* legacy rate set */
+	struct wmi_rate_set peer_legacy_rates;
+	/* ht rate set */
+	struct wmi_rate_set peer_ht_rates;
+} __packed;
+
+struct wmi_peer_set_q_empty_callback_cmd {
+	/* unique id identifying the VDEV, generated by the caller */
+	__le32 vdev_id;
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+	__le32 callback_enable;
+} __packed;
+
+#define WMI_PEER_AUTH           0x00000001
+#define WMI_PEER_QOS            0x00000002
+#define WMI_PEER_NEED_PTK_4_WAY 0x00000004
+#define WMI_PEER_NEED_GTK_2_WAY 0x00000010
+#define WMI_PEER_APSD           0x00000800
+#define WMI_PEER_HT             0x00001000
+#define WMI_PEER_40MHZ          0x00002000
+#define WMI_PEER_STBC           0x00008000
+#define WMI_PEER_LDPC           0x00010000
+#define WMI_PEER_DYN_MIMOPS     0x00020000
+#define WMI_PEER_STATIC_MIMOPS  0x00040000
+#define WMI_PEER_SPATIAL_MUX    0x00200000
+#define WMI_PEER_VHT            0x02000000
+#define WMI_PEER_80MHZ          0x04000000
+#define WMI_PEER_PMF            0x08000000
+
+/*
+ * Peer rate capabilities.
+ *
+ * This is of interest to the ratecontrol
+ * module which resides in the firmware. The bit definitions are
+ * consistent with that defined in if_athrate.c.
+ */
+#define WMI_RC_DS_FLAG          0x01
+#define WMI_RC_CW40_FLAG        0x02
+#define WMI_RC_SGI_FLAG         0x04
+#define WMI_RC_HT_FLAG          0x08
+#define WMI_RC_RTSCTS_FLAG      0x10
+#define WMI_RC_TX_STBC_FLAG     0x20
+#define WMI_RC_RX_STBC_FLAG     0xC0
+#define WMI_RC_RX_STBC_FLAG_S   6
+#define WMI_RC_WEP_TKIP_FLAG    0x100
+#define WMI_RC_TS_FLAG          0x200
+#define WMI_RC_UAPSD_FLAG       0x400
+
+/* Maximum listen interval supported by hw in units of beacon interval */
+#define ATH10K_MAX_HW_LISTEN_INTERVAL 5
+
+struct wmi_peer_assoc_complete_cmd {
+	struct wmi_mac_addr peer_macaddr;
+	__le32 vdev_id;
+	__le32 peer_new_assoc; /* 1=assoc, 0=reassoc */
+	__le32 peer_associd; /* 16 LSBs */
+	__le32 peer_flags;
+	__le32 peer_caps; /* 16 LSBs */
+	__le32 peer_listen_intval;
+	__le32 peer_ht_caps;
+	__le32 peer_max_mpdu;
+	__le32 peer_mpdu_density; /* 0..16 */
+	__le32 peer_rate_caps;
+	struct wmi_rate_set peer_legacy_rates;
+	struct wmi_rate_set peer_ht_rates;
+	__le32 peer_nss; /* num of spatial streams */
+	__le32 peer_vht_caps;
+	__le32 peer_phymode;
+	struct wmi_vht_rate_set peer_vht_rates;
+	/* HT Operation Element of the peer. Five bytes packed in 2
+	 *  INT32 array and filled from lsb to msb. */
+	__le32 peer_ht_info[2];
+} __packed;
+
+struct wmi_peer_assoc_complete_arg {
+	u8 addr[ETH_ALEN];
+	u32 vdev_id;
+	bool peer_reassoc;
+	u16 peer_aid;
+	u32 peer_flags; /* see %WMI_PEER_ */
+	u16 peer_caps;
+	u32 peer_listen_intval;
+	u32 peer_ht_caps;
+	u32 peer_max_mpdu;
+	u32 peer_mpdu_density; /* 0..16 */
+	u32 peer_rate_caps; /* see %WMI_RC_ */
+	struct wmi_rate_set_arg peer_legacy_rates;
+	struct wmi_rate_set_arg peer_ht_rates;
+	u32 peer_num_spatial_streams;
+	u32 peer_vht_caps;
+	enum wmi_phy_mode peer_phymode;
+	struct wmi_vht_rate_set_arg peer_vht_rates;
+};
+
+struct wmi_peer_add_wds_entry_cmd {
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+	/* wds MAC addr */
+	struct wmi_mac_addr wds_macaddr;
+} __packed;
+
+struct wmi_peer_remove_wds_entry_cmd {
+	/* wds MAC addr */
+	struct wmi_mac_addr wds_macaddr;
+} __packed;
+
+struct wmi_peer_q_empty_callback_event {
+	/* peer MAC address */
+	struct wmi_mac_addr peer_macaddr;
+} __packed;
+
+/*
+ * Channel info WMI event
+ */
+struct wmi_chan_info_event {
+	__le32 err_code;
+	__le32 freq;
+	__le32 cmd_flags;
+	__le32 noise_floor;
+	__le32 rx_clear_count;
+	__le32 cycle_count;
+} __packed;
+
+/* Beacon filter wmi command info */
+#define BCN_FLT_MAX_SUPPORTED_IES	256
+#define BCN_FLT_MAX_ELEMS_IE_LIST	(BCN_FLT_MAX_SUPPORTED_IES / 32)
+
+struct bss_bcn_stats {
+	__le32 vdev_id;
+	__le32 bss_bcnsdropped;
+	__le32 bss_bcnsdelivered;
+} __packed;
+
+struct bcn_filter_stats {
+	__le32 bcns_dropped;
+	__le32 bcns_delivered;
+	__le32 activefilters;
+	struct bss_bcn_stats bss_stats;
+} __packed;
+
+struct wmi_add_bcn_filter_cmd {
+	u32 vdev_id;
+	u32 ie_map[BCN_FLT_MAX_ELEMS_IE_LIST];
+} __packed;
+
+enum wmi_sta_keepalive_method {
+	WMI_STA_KEEPALIVE_METHOD_NULL_FRAME = 1,
+	WMI_STA_KEEPALIVE_METHOD_UNSOLICITATED_ARP_RESPONSE = 2,
+};
+
+/* note: ip4 addresses are in network byte order, i.e. big endian */
+struct wmi_sta_keepalive_arp_resp {
+	__be32 src_ip4_addr;
+	__be32 dest_ip4_addr;
+	struct wmi_mac_addr dest_mac_addr;
+} __packed;
+
+struct wmi_sta_keepalive_cmd {
+	__le32 vdev_id;
+	__le32 enabled;
+	__le32 method; /* WMI_STA_KEEPALIVE_METHOD_ */
+	__le32 interval; /* in seconds */
+	struct wmi_sta_keepalive_arp_resp arp_resp;
+} __packed;
+
+#define ATH10K_RTS_MAX		2347
+#define ATH10K_FRAGMT_THRESHOLD_MIN	540
+#define ATH10K_FRAGMT_THRESHOLD_MAX	2346
+
+#define WMI_MAX_EVENT 0x1000
+/* Maximum number of pending TXed WMI packets */
+#define WMI_MAX_PENDING_TX_COUNT 128
+#define WMI_SKB_HEADROOM sizeof(struct wmi_cmd_hdr)
+
+/* By default disable power save for IBSS */
+#define ATH10K_DEFAULT_ATIM 0
+
+struct ath10k;
+struct ath10k_vif;
+
+int ath10k_wmi_attach(struct ath10k *ar);
+void ath10k_wmi_detach(struct ath10k *ar);
+int ath10k_wmi_wait_for_service_ready(struct ath10k *ar);
+int ath10k_wmi_wait_for_unified_ready(struct ath10k *ar);
+void ath10k_wmi_flush_tx(struct ath10k *ar);
+
+int ath10k_wmi_connect_htc_service(struct ath10k *ar);
+int ath10k_wmi_pdev_set_channel(struct ath10k *ar,
+				const struct wmi_channel_arg *);
+int ath10k_wmi_pdev_suspend_target(struct ath10k *ar);
+int ath10k_wmi_pdev_resume_target(struct ath10k *ar);
+int ath10k_wmi_pdev_set_regdomain(struct ath10k *ar, u16 rd, u16 rd2g,
+				  u16 rd5g, u16 ctl2g, u16 ctl5g);
+int ath10k_wmi_pdev_set_param(struct ath10k *ar, enum wmi_pdev_param id,
+			      u32 value);
+int ath10k_wmi_cmd_init(struct ath10k *ar);
+int ath10k_wmi_start_scan(struct ath10k *ar, const struct wmi_start_scan_arg *);
+void ath10k_wmi_start_scan_init(struct ath10k *ar, struct wmi_start_scan_arg *);
+int ath10k_wmi_stop_scan(struct ath10k *ar,
+			 const struct wmi_stop_scan_arg *arg);
+int ath10k_wmi_vdev_create(struct ath10k *ar, u32 vdev_id,
+			   enum wmi_vdev_type type,
+			   enum wmi_vdev_subtype subtype,
+			   const u8 macaddr[ETH_ALEN]);
+int ath10k_wmi_vdev_delete(struct ath10k *ar, u32 vdev_id);
+int ath10k_wmi_vdev_start(struct ath10k *ar,
+			  const struct wmi_vdev_start_request_arg *);
+int ath10k_wmi_vdev_restart(struct ath10k *ar,
+			    const struct wmi_vdev_start_request_arg *);
+int ath10k_wmi_vdev_stop(struct ath10k *ar, u32 vdev_id);
+int ath10k_wmi_vdev_up(struct ath10k *ar, u32 vdev_id, u32 aid,
+		       const u8 *bssid);
+int ath10k_wmi_vdev_down(struct ath10k *ar, u32 vdev_id);
+int ath10k_wmi_vdev_set_param(struct ath10k *ar, u32 vdev_id,
+			      enum wmi_vdev_param param_id, u32 param_value);
+int ath10k_wmi_vdev_install_key(struct ath10k *ar,
+				const struct wmi_vdev_install_key_arg *arg);
+int ath10k_wmi_peer_create(struct ath10k *ar, u32 vdev_id,
+		    const u8 peer_addr[ETH_ALEN]);
+int ath10k_wmi_peer_delete(struct ath10k *ar, u32 vdev_id,
+		    const u8 peer_addr[ETH_ALEN]);
+int ath10k_wmi_peer_flush(struct ath10k *ar, u32 vdev_id,
+		   const u8 peer_addr[ETH_ALEN], u32 tid_bitmap);
+int ath10k_wmi_peer_set_param(struct ath10k *ar, u32 vdev_id,
+			      const u8 *peer_addr,
+			      enum wmi_peer_param param_id, u32 param_value);
+int ath10k_wmi_peer_assoc(struct ath10k *ar,
+			  const struct wmi_peer_assoc_complete_arg *arg);
+int ath10k_wmi_set_psmode(struct ath10k *ar, u32 vdev_id,
+			  enum wmi_sta_ps_mode psmode);
+int ath10k_wmi_set_sta_ps_param(struct ath10k *ar, u32 vdev_id,
+				enum wmi_sta_powersave_param param_id,
+				u32 value);
+int ath10k_wmi_set_ap_ps_param(struct ath10k *ar, u32 vdev_id, const u8 *mac,
+			       enum wmi_ap_ps_peer_param param_id, u32 value);
+int ath10k_wmi_scan_chan_list(struct ath10k *ar,
+			      const struct wmi_scan_chan_list_arg *arg);
+int ath10k_wmi_beacon_send(struct ath10k *ar, const struct wmi_bcn_tx_arg *arg);
+int ath10k_wmi_pdev_set_wmm_params(struct ath10k *ar,
+			const struct wmi_pdev_set_wmm_params_arg *arg);
+int ath10k_wmi_request_stats(struct ath10k *ar, enum wmi_stats_id stats_id);
+
+#endif /* _WMI_H_ */
diff --git a/drivers/net/wireless/ath/ath5k/ahb.c b/drivers/net/wireless/ath/ath5k/ahb.c
index 8e8bcc7a..e9bc9e6 100644
--- a/drivers/net/wireless/ath/ath5k/ahb.c
+++ b/drivers/net/wireless/ath/ath5k/ahb.c
@@ -185,7 +185,6 @@ static int ath_ahb_probe(struct platform_device *pdev)
 
  err_free_hw:
 	ieee80211_free_hw(hw);
-	platform_set_drvdata(pdev, NULL);
  err_iounmap:
         iounmap(mem);
  err_out:
@@ -221,7 +220,6 @@ static int ath_ahb_remove(struct platform_device *pdev)
 
 	ath5k_deinit_ah(ah);
 	iounmap(ah->iobase);
-	platform_set_drvdata(pdev, NULL);
 	ieee80211_free_hw(hw);
 
 	return 0;
diff --git a/drivers/net/wireless/ath/ath5k/base.c b/drivers/net/wireless/ath/ath5k/base.c
index 7f702fe..ce67ab7 100644
--- a/drivers/net/wireless/ath/ath5k/base.c
+++ b/drivers/net/wireless/ath/ath5k/base.c
@@ -60,6 +60,7 @@
 
 #include <asm/unaligned.h>
 
+#include <net/mac80211.h>
 #include "base.h"
 #include "reg.h"
 #include "debug.h"
@@ -666,9 +667,46 @@ static enum ath5k_pkt_type get_hw_packet_type(struct sk_buff *skb)
 	return htype;
 }
 
+static struct ieee80211_rate *
+ath5k_get_rate(const struct ieee80211_hw *hw,
+	       const struct ieee80211_tx_info *info,
+	       struct ath5k_buf *bf, int idx)
+{
+	/*
+	* convert a ieee80211_tx_rate RC-table entry to
+	* the respective ieee80211_rate struct
+	*/
+	if (bf->rates[idx].idx < 0) {
+		return NULL;
+	}
+
+	return &hw->wiphy->bands[info->band]->bitrates[ bf->rates[idx].idx ];
+}
+
+static u16
+ath5k_get_rate_hw_value(const struct ieee80211_hw *hw,
+			const struct ieee80211_tx_info *info,
+			struct ath5k_buf *bf, int idx)
+{
+	struct ieee80211_rate *rate;
+	u16 hw_rate;
+	u8 rc_flags;
+
+	rate = ath5k_get_rate(hw, info, bf, idx);
+	if (!rate)
+		return 0;
+
+	rc_flags = bf->rates[idx].flags;
+	hw_rate = (rc_flags & IEEE80211_TX_RC_USE_SHORT_PREAMBLE) ?
+		   rate->hw_value_short : rate->hw_value;
+
+	return hw_rate;
+}
+
 static int
 ath5k_txbuf_setup(struct ath5k_hw *ah, struct ath5k_buf *bf,
-		  struct ath5k_txq *txq, int padsize)
+		  struct ath5k_txq *txq, int padsize,
+		  struct ieee80211_tx_control *control)
 {
 	struct ath5k_desc *ds = bf->desc;
 	struct sk_buff *skb = bf->skb;
@@ -688,7 +726,11 @@ ath5k_txbuf_setup(struct ath5k_hw *ah, struct ath5k_buf *bf,
 	bf->skbaddr = dma_map_single(ah->dev, skb->data, skb->len,
 			DMA_TO_DEVICE);
 
-	rate = ieee80211_get_tx_rate(ah->hw, info);
+	ieee80211_get_tx_rates(info->control.vif, (control) ? control->sta : NULL, skb, bf->rates,
+			       ARRAY_SIZE(bf->rates));
+
+	rate = ath5k_get_rate(ah->hw, info, bf, 0);
+
 	if (!rate) {
 		ret = -EINVAL;
 		goto err_unmap;
@@ -698,8 +740,8 @@ ath5k_txbuf_setup(struct ath5k_hw *ah, struct ath5k_buf *bf,
 		flags |= AR5K_TXDESC_NOACK;
 
 	rc_flags = info->control.rates[0].flags;
-	hw_rate = (rc_flags & IEEE80211_TX_RC_USE_SHORT_PREAMBLE) ?
-		rate->hw_value_short : rate->hw_value;
+
+	hw_rate = ath5k_get_rate_hw_value(ah->hw, info, bf, 0);
 
 	pktlen = skb->len;
 
@@ -722,12 +764,13 @@ ath5k_txbuf_setup(struct ath5k_hw *ah, struct ath5k_buf *bf,
 		duration = le16_to_cpu(ieee80211_ctstoself_duration(ah->hw,
 			info->control.vif, pktlen, info));
 	}
+
 	ret = ah->ah_setup_tx_desc(ah, ds, pktlen,
 		ieee80211_get_hdrlen_from_skb(skb), padsize,
 		get_hw_packet_type(skb),
 		(ah->ah_txpower.txp_requested * 2),
 		hw_rate,
-		info->control.rates[0].count, keyidx, ah->ah_tx_ant, flags,
+		bf->rates[0].count, keyidx, ah->ah_tx_ant, flags,
 		cts_rate, duration);
 	if (ret)
 		goto err_unmap;
@@ -736,13 +779,15 @@ ath5k_txbuf_setup(struct ath5k_hw *ah, struct ath5k_buf *bf,
 	if (ah->ah_capabilities.cap_has_mrr_support) {
 		memset(mrr_rate, 0, sizeof(mrr_rate));
 		memset(mrr_tries, 0, sizeof(mrr_tries));
+
 		for (i = 0; i < 3; i++) {
-			rate = ieee80211_get_alt_retry_rate(ah->hw, info, i);
+
+			rate = ath5k_get_rate(ah->hw, info, bf, i);
 			if (!rate)
 				break;
 
-			mrr_rate[i] = rate->hw_value;
-			mrr_tries[i] = info->control.rates[i + 1].count;
+			mrr_rate[i] = ath5k_get_rate_hw_value(ah->hw, info, bf, i);
+			mrr_tries[i] = bf->rates[i].count;
 		}
 
 		ath5k_hw_setup_mrr_tx_desc(ah, ds,
@@ -1515,7 +1560,7 @@ unlock:
 
 void
 ath5k_tx_queue(struct ieee80211_hw *hw, struct sk_buff *skb,
-	       struct ath5k_txq *txq)
+	       struct ath5k_txq *txq, struct ieee80211_tx_control *control)
 {
 	struct ath5k_hw *ah = hw->priv;
 	struct ath5k_buf *bf;
@@ -1555,7 +1600,7 @@ ath5k_tx_queue(struct ieee80211_hw *hw, struct sk_buff *skb,
 
 	bf->skb = skb;
 
-	if (ath5k_txbuf_setup(ah, bf, txq, padsize)) {
+	if (ath5k_txbuf_setup(ah, bf, txq, padsize, control)) {
 		bf->skb = NULL;
 		spin_lock_irqsave(&ah->txbuflock, flags);
 		list_add_tail(&bf->list, &ah->txbuf);
@@ -1571,11 +1616,13 @@ drop_packet:
 
 static void
 ath5k_tx_frame_completed(struct ath5k_hw *ah, struct sk_buff *skb,
-			 struct ath5k_txq *txq, struct ath5k_tx_status *ts)
+			 struct ath5k_txq *txq, struct ath5k_tx_status *ts,
+			 struct ath5k_buf *bf)
 {
 	struct ieee80211_tx_info *info;
 	u8 tries[3];
 	int i;
+	int size = 0;
 
 	ah->stats.tx_all_count++;
 	ah->stats.tx_bytes_count += skb->len;
@@ -1587,6 +1634,9 @@ ath5k_tx_frame_completed(struct ath5k_hw *ah, struct sk_buff *skb,
 
 	ieee80211_tx_info_clear_status(info);
 
+	size = min_t(int, sizeof(info->status.rates), sizeof(bf->rates));
+	memcpy(info->status.rates, bf->rates, size);
+
 	for (i = 0; i < ts->ts_final_idx; i++) {
 		struct ieee80211_tx_rate *r =
 			&info->status.rates[i];
@@ -1663,7 +1713,7 @@ ath5k_tx_processq(struct ath5k_hw *ah, struct ath5k_txq *txq)
 
 			dma_unmap_single(ah->dev, bf->skbaddr, skb->len,
 					DMA_TO_DEVICE);
-			ath5k_tx_frame_completed(ah, skb, txq, &ts);
+			ath5k_tx_frame_completed(ah, skb, txq, &ts, bf);
 		}
 
 		/*
@@ -1917,7 +1967,7 @@ ath5k_beacon_send(struct ath5k_hw *ah)
 
 	skb = ieee80211_get_buffered_bc(ah->hw, vif);
 	while (skb) {
-		ath5k_tx_queue(ah->hw, skb, ah->cabq);
+		ath5k_tx_queue(ah->hw, skb, ah->cabq, NULL);
 
 		if (ah->cabq->txq_len >= ah->cabq->txq_max)
 			break;
@@ -2442,7 +2492,8 @@ ath5k_init_ah(struct ath5k_hw *ah, const struct ath_bus_ops *bus_ops)
 			IEEE80211_HW_HOST_BROADCAST_PS_BUFFERING |
 			IEEE80211_HW_SIGNAL_DBM |
 			IEEE80211_HW_MFP_CAPABLE |
-			IEEE80211_HW_REPORTS_TX_ACK_STATUS;
+			IEEE80211_HW_REPORTS_TX_ACK_STATUS |
+			IEEE80211_HW_SUPPORTS_RC_TABLE;
 
 	hw->wiphy->interface_modes =
 		BIT(NL80211_IFTYPE_AP) |
diff --git a/drivers/net/wireless/ath/ath5k/base.h b/drivers/net/wireless/ath/ath5k/base.h
index 6c94c7f..ca9a83c 100644
--- a/drivers/net/wireless/ath/ath5k/base.h
+++ b/drivers/net/wireless/ath/ath5k/base.h
@@ -47,6 +47,7 @@ struct ath5k_hw;
 struct ath5k_txq;
 struct ieee80211_channel;
 struct ath_bus_ops;
+struct ieee80211_tx_control;
 enum nl80211_iftype;
 
 enum ath5k_srev_type {
@@ -61,11 +62,12 @@ struct ath5k_srev_name {
 };
 
 struct ath5k_buf {
-	struct list_head	list;
-	struct ath5k_desc	*desc;	/* virtual addr of desc */
-	dma_addr_t		daddr;	/* physical addr of desc */
-	struct sk_buff		*skb;	/* skbuff for buf */
-	dma_addr_t		skbaddr;/* physical addr of skb data */
+	struct list_head		list;
+	struct ath5k_desc		*desc;		/* virtual addr of desc */
+	dma_addr_t			daddr;		/* physical addr of desc */
+	struct sk_buff			*skb;		/* skbuff for buf */
+	dma_addr_t			skbaddr;	/* physical addr of skb data */
+	struct ieee80211_tx_rate	rates[4];	/* number of multi-rate stages */
 };
 
 struct ath5k_vif {
@@ -103,7 +105,7 @@ int ath5k_chan_set(struct ath5k_hw *ah, struct ieee80211_channel *chan);
 void ath5k_txbuf_free_skb(struct ath5k_hw *ah, struct ath5k_buf *bf);
 void ath5k_rxbuf_free_skb(struct ath5k_hw *ah, struct ath5k_buf *bf);
 void ath5k_tx_queue(struct ieee80211_hw *hw, struct sk_buff *skb,
-		    struct ath5k_txq *txq);
+		    struct ath5k_txq *txq, struct ieee80211_tx_control *control);
 
 const char *ath5k_chip_name(enum ath5k_srev_type type, u_int16_t val);
 
diff --git a/drivers/net/wireless/ath/ath5k/mac80211-ops.c b/drivers/net/wireless/ath/ath5k/mac80211-ops.c
index 06f86f4..81b686c 100644
--- a/drivers/net/wireless/ath/ath5k/mac80211-ops.c
+++ b/drivers/net/wireless/ath/ath5k/mac80211-ops.c
@@ -66,7 +66,7 @@ ath5k_tx(struct ieee80211_hw *hw, struct ieee80211_tx_control *control,
 		return;
 	}
 
-	ath5k_tx_queue(hw, skb, &ah->txqs[qnum]);
+	ath5k_tx_queue(hw, skb, &ah->txqs[qnum], control);
 }
 
 
diff --git a/drivers/net/wireless/ath/ath6kl/cfg80211.c b/drivers/net/wireless/ath/ath6kl/cfg80211.c
index 5c9736a..2437ad2 100644
--- a/drivers/net/wireless/ath/ath6kl/cfg80211.c
+++ b/drivers/net/wireless/ath/ath6kl/cfg80211.c
@@ -3175,10 +3175,21 @@ static int ath6kl_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 {
 	struct ath6kl_vif *vif = ath6kl_vif_from_wdev(wdev);
 	struct ath6kl *ar = ath6kl_priv(vif->ndev);
-	u32 id;
+	u32 id, freq;
 	const struct ieee80211_mgmt *mgmt;
 	bool more_data, queued;
 
+	/* default to the current channel, but use the one specified as argument
+	 * if any
+	 */
+	freq = vif->ch_hint;
+	if (chan)
+		freq = chan->center_freq;
+
+	/* never send freq zero to the firmware */
+	if (WARN_ON(freq == 0))
+		return -EINVAL;
+
 	mgmt = (const struct ieee80211_mgmt *) buf;
 	if (vif->nw_type == AP_NETWORK && test_bit(CONNECTED, &vif->flags) &&
 	    ieee80211_is_probe_resp(mgmt->frame_control) &&
@@ -3188,8 +3199,7 @@ static int ath6kl_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 		 * command to allow the target to fill in the generic IEs.
 		 */
 		*cookie = 0; /* TX status not supported */
-		return ath6kl_send_go_probe_resp(vif, buf, len,
-						 chan->center_freq);
+		return ath6kl_send_go_probe_resp(vif, buf, len, freq);
 	}
 
 	id = vif->send_action_id++;
@@ -3205,17 +3215,14 @@ static int ath6kl_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 
 	/* AP mode Power saving processing */
 	if (vif->nw_type == AP_NETWORK) {
-		queued = ath6kl_mgmt_powersave_ap(vif,
-					id, chan->center_freq,
-					wait, buf,
-					len, &more_data, no_cck);
+		queued = ath6kl_mgmt_powersave_ap(vif, id, freq, wait, buf, len,
+						  &more_data, no_cck);
 		if (queued)
 			return 0;
 	}
 
-	return ath6kl_wmi_send_mgmt_cmd(ar->wmi, vif->fw_vif_idx, id,
-					chan->center_freq, wait,
-					buf, len, no_cck);
+	return ath6kl_wmi_send_mgmt_cmd(ar->wmi, vif->fw_vif_idx, id, freq,
+					wait, buf, len, no_cck);
 }
 
 static void ath6kl_mgmt_frame_register(struct wiphy *wiphy,
@@ -3679,6 +3686,20 @@ err:
 	return NULL;
 }
 
+#ifdef CONFIG_PM
+static const struct wiphy_wowlan_support ath6kl_wowlan_support = {
+	.flags = WIPHY_WOWLAN_MAGIC_PKT |
+		 WIPHY_WOWLAN_DISCONNECT |
+		 WIPHY_WOWLAN_GTK_REKEY_FAILURE  |
+		 WIPHY_WOWLAN_SUPPORTS_GTK_REKEY |
+		 WIPHY_WOWLAN_EAP_IDENTITY_REQ   |
+		 WIPHY_WOWLAN_4WAY_HANDSHAKE,
+	.n_patterns = WOW_MAX_FILTERS_PER_LIST,
+	.pattern_min_len = 1,
+	.pattern_max_len = WOW_PATTERN_SIZE,
+};
+#endif
+
 int ath6kl_cfg80211_init(struct ath6kl *ar)
 {
 	struct wiphy *wiphy = ar->wiphy;
@@ -3772,15 +3793,7 @@ int ath6kl_cfg80211_init(struct ath6kl *ar)
 	wiphy->n_cipher_suites = ARRAY_SIZE(cipher_suites);
 
 #ifdef CONFIG_PM
-	wiphy->wowlan.flags = WIPHY_WOWLAN_MAGIC_PKT |
-			      WIPHY_WOWLAN_DISCONNECT |
-			      WIPHY_WOWLAN_GTK_REKEY_FAILURE  |
-			      WIPHY_WOWLAN_SUPPORTS_GTK_REKEY |
-			      WIPHY_WOWLAN_EAP_IDENTITY_REQ   |
-			      WIPHY_WOWLAN_4WAY_HANDSHAKE;
-	wiphy->wowlan.n_patterns = WOW_MAX_FILTERS_PER_LIST;
-	wiphy->wowlan.pattern_min_len = 1;
-	wiphy->wowlan.pattern_max_len = WOW_PATTERN_SIZE;
+	wiphy->wowlan = &ath6kl_wowlan_support;
 #endif
 
 	wiphy->max_sched_scan_ssids = MAX_PROBED_SSIDS;
diff --git a/drivers/net/wireless/ath/ath6kl/debug.c b/drivers/net/wireless/ath/ath6kl/debug.c
index fe38b83..dbfd17d 100644
--- a/drivers/net/wireless/ath/ath6kl/debug.c
+++ b/drivers/net/wireless/ath/ath6kl/debug.c
@@ -1240,20 +1240,14 @@ static ssize_t ath6kl_force_roam_write(struct file *file,
 	char buf[20];
 	size_t len;
 	u8 bssid[ETH_ALEN];
-	int i;
-	int addr[ETH_ALEN];
 
 	len = min(count, sizeof(buf) - 1);
 	if (copy_from_user(buf, user_buf, len))
 		return -EFAULT;
 	buf[len] = '\0';
 
-	if (sscanf(buf, "%02x:%02x:%02x:%02x:%02x:%02x",
-		   &addr[0], &addr[1], &addr[2], &addr[3], &addr[4], &addr[5])
-	    != ETH_ALEN)
+	if (!mac_pton(buf, bssid))
 		return -EINVAL;
-	for (i = 0; i < ETH_ALEN; i++)
-		bssid[i] = addr[i];
 
 	ret = ath6kl_wmi_force_roam_cmd(ar->wmi, bssid);
 	if (ret)
diff --git a/drivers/net/wireless/ath/ath6kl/init.c b/drivers/net/wireless/ath/ath6kl/init.c
index 40ffee6..6a67881 100644
--- a/drivers/net/wireless/ath/ath6kl/init.c
+++ b/drivers/net/wireless/ath/ath6kl/init.c
@@ -1696,10 +1696,16 @@ static int __ath6kl_init_hw_start(struct ath6kl *ar)
 						    test_bit(WMI_READY,
 							     &ar->flag),
 						    WMI_TIMEOUT);
+	if (timeleft <= 0) {
+		clear_bit(WMI_READY, &ar->flag);
+		ath6kl_err("wmi is not ready or wait was interrupted: %ld\n",
+			   timeleft);
+		ret = -EIO;
+		goto err_htc_stop;
+	}
 
 	ath6kl_dbg(ATH6KL_DBG_BOOT, "firmware booted\n");
 
-
 	if (test_and_clear_bit(FIRST_BOOT, &ar->flag)) {
 		ath6kl_info("%s %s fw %s api %d%s\n",
 			    ar->hw.name,
@@ -1718,12 +1724,6 @@ static int __ath6kl_init_hw_start(struct ath6kl *ar)
 		goto err_htc_stop;
 	}
 
-	if (!timeleft || signal_pending(current)) {
-		ath6kl_err("wmi is not ready or wait was interrupted\n");
-		ret = -EIO;
-		goto err_htc_stop;
-	}
-
 	ath6kl_dbg(ATH6KL_DBG_TRC, "%s: wmi is ready\n", __func__);
 
 	/* communicate the wmi protocol verision to the target */
diff --git a/drivers/net/wireless/ath/ath6kl/sdio.c b/drivers/net/wireless/ath/ath6kl/sdio.c
index fb14145..7126bdd 100644
--- a/drivers/net/wireless/ath/ath6kl/sdio.c
+++ b/drivers/net/wireless/ath/ath6kl/sdio.c
@@ -345,17 +345,17 @@ static int ath6kl_sdio_alloc_prep_scat_req(struct ath6kl_sdio *ar_sdio,
 {
 	struct hif_scatter_req *s_req;
 	struct bus_request *bus_req;
-	int i, scat_req_sz, scat_list_sz, sg_sz, buf_sz;
+	int i, scat_req_sz, scat_list_sz, size;
 	u8 *virt_buf;
 
 	scat_list_sz = (n_scat_entry - 1) * sizeof(struct hif_scatter_item);
 	scat_req_sz = sizeof(*s_req) + scat_list_sz;
 
 	if (!virt_scat)
-		sg_sz = sizeof(struct scatterlist) * n_scat_entry;
+		size = sizeof(struct scatterlist) * n_scat_entry;
 	else
-		buf_sz =  2 * L1_CACHE_BYTES +
-			  ATH6KL_MAX_TRANSFER_SIZE_PER_SCATTER;
+		size =  2 * L1_CACHE_BYTES +
+			ATH6KL_MAX_TRANSFER_SIZE_PER_SCATTER;
 
 	for (i = 0; i < n_scat_req; i++) {
 		/* allocate the scatter request */
@@ -364,7 +364,7 @@ static int ath6kl_sdio_alloc_prep_scat_req(struct ath6kl_sdio *ar_sdio,
 			return -ENOMEM;
 
 		if (virt_scat) {
-			virt_buf = kzalloc(buf_sz, GFP_KERNEL);
+			virt_buf = kzalloc(size, GFP_KERNEL);
 			if (!virt_buf) {
 				kfree(s_req);
 				return -ENOMEM;
@@ -374,7 +374,7 @@ static int ath6kl_sdio_alloc_prep_scat_req(struct ath6kl_sdio *ar_sdio,
 				(u8 *)L1_CACHE_ALIGN((unsigned long)virt_buf);
 		} else {
 			/* allocate sglist */
-			s_req->sgentries = kzalloc(sg_sz, GFP_KERNEL);
+			s_req->sgentries = kzalloc(size, GFP_KERNEL);
 
 			if (!s_req->sgentries) {
 				kfree(s_req);
diff --git a/drivers/net/wireless/ath/ath6kl/usb.c b/drivers/net/wireless/ath/ath6kl/usb.c
index bed0d33..f38ff6a 100644
--- a/drivers/net/wireless/ath/ath6kl/usb.c
+++ b/drivers/net/wireless/ath/ath6kl/usb.c
@@ -1061,6 +1061,22 @@ static void ath6kl_usb_cleanup_scatter(struct ath6kl *ar)
 	return;
 }
 
+static int ath6kl_usb_suspend(struct ath6kl *ar, struct cfg80211_wowlan *wow)
+{
+	/*
+	 * cfg80211 suspend/WOW currently not supported for USB.
+	 */
+	return 0;
+}
+
+static int ath6kl_usb_resume(struct ath6kl *ar)
+{
+	/*
+	 * cfg80211 resume currently not supported for USB.
+	 */
+	return 0;
+}
+
 static const struct ath6kl_hif_ops ath6kl_usb_ops = {
 	.diag_read32 = ath6kl_usb_diag_read32,
 	.diag_write32 = ath6kl_usb_diag_write32,
@@ -1074,6 +1090,8 @@ static const struct ath6kl_hif_ops ath6kl_usb_ops = {
 	.pipe_map_service = ath6kl_usb_map_service_pipe,
 	.pipe_get_free_queue_number = ath6kl_usb_get_free_queue_number,
 	.cleanup_scatter = ath6kl_usb_cleanup_scatter,
+	.suspend = ath6kl_usb_suspend,
+	.resume = ath6kl_usb_resume,
 };
 
 /* ath6kl usb driver registered functions */
@@ -1152,7 +1170,7 @@ static void ath6kl_usb_remove(struct usb_interface *interface)
 
 #ifdef CONFIG_PM
 
-static int ath6kl_usb_suspend(struct usb_interface *interface,
+static int ath6kl_usb_pm_suspend(struct usb_interface *interface,
 			      pm_message_t message)
 {
 	struct ath6kl_usb *device;
@@ -1162,7 +1180,7 @@ static int ath6kl_usb_suspend(struct usb_interface *interface,
 	return 0;
 }
 
-static int ath6kl_usb_resume(struct usb_interface *interface)
+static int ath6kl_usb_pm_resume(struct usb_interface *interface)
 {
 	struct ath6kl_usb *device;
 	device = usb_get_intfdata(interface);
@@ -1175,7 +1193,7 @@ static int ath6kl_usb_resume(struct usb_interface *interface)
 	return 0;
 }
 
-static int ath6kl_usb_reset_resume(struct usb_interface *intf)
+static int ath6kl_usb_pm_reset_resume(struct usb_interface *intf)
 {
 	if (usb_get_intfdata(intf))
 		ath6kl_usb_remove(intf);
@@ -1184,9 +1202,9 @@ static int ath6kl_usb_reset_resume(struct usb_interface *intf)
 
 #else
 
-#define ath6kl_usb_suspend NULL
-#define ath6kl_usb_resume NULL
-#define ath6kl_usb_reset_resume NULL
+#define ath6kl_usb_pm_suspend NULL
+#define ath6kl_usb_pm_resume NULL
+#define ath6kl_usb_pm_reset_resume NULL
 
 #endif
 
@@ -1201,9 +1219,9 @@ MODULE_DEVICE_TABLE(usb, ath6kl_usb_ids);
 static struct usb_driver ath6kl_usb_driver = {
 	.name = "ath6kl_usb",
 	.probe = ath6kl_usb_probe,
-	.suspend = ath6kl_usb_suspend,
-	.resume = ath6kl_usb_resume,
-	.reset_resume = ath6kl_usb_reset_resume,
+	.suspend = ath6kl_usb_pm_suspend,
+	.resume = ath6kl_usb_pm_resume,
+	.reset_resume = ath6kl_usb_pm_reset_resume,
 	.disconnect = ath6kl_usb_remove,
 	.id_table = ath6kl_usb_ids,
 	.supports_autosuspend = true,
diff --git a/drivers/net/wireless/ath/ath9k/Kconfig b/drivers/net/wireless/ath/ath9k/Kconfig
index f985cf3..d491a31 100644
--- a/drivers/net/wireless/ath/ath9k/Kconfig
+++ b/drivers/net/wireless/ath/ath9k/Kconfig
@@ -84,14 +84,6 @@ config ATH9K_DFS_CERTIFIED
 	  developed. At this point enabling this option won't do anything
 	  except increase code size.
 
-config ATH9K_MAC_DEBUG
-	bool "Atheros MAC statistics"
-	depends on ATH9K_DEBUGFS
-	default y
-	---help---
-	  This option enables collection of statistics for Rx/Tx status
-	  data and some other MAC related statistics
-
 config ATH9K_LEGACY_RATE_CONTROL
 	bool "Atheros ath9k rate control"
 	depends on ATH9K
diff --git a/drivers/net/wireless/ath/ath9k/ahb.c b/drivers/net/wireless/ath/ath9k/ahb.c
index d1ff3c2..072e4b5 100644
--- a/drivers/net/wireless/ath/ath9k/ahb.c
+++ b/drivers/net/wireless/ath/ath9k/ahb.c
@@ -150,7 +150,6 @@ static int ath_ahb_probe(struct platform_device *pdev)
 	free_irq(irq, sc);
  err_free_hw:
 	ieee80211_free_hw(hw);
-	platform_set_drvdata(pdev, NULL);
 	return ret;
 }
 
@@ -164,7 +163,6 @@ static int ath_ahb_remove(struct platform_device *pdev)
 		ath9k_deinit_device(sc);
 		free_irq(sc->irq, sc);
 		ieee80211_free_hw(sc->hw);
-		platform_set_drvdata(pdev, NULL);
 	}
 
 	return 0;
diff --git a/drivers/net/wireless/ath/ath9k/ani.c b/drivers/net/wireless/ath/ath9k/ani.c
index 7ecd40f..4994bea 100644
--- a/drivers/net/wireless/ath/ath9k/ani.c
+++ b/drivers/net/wireless/ath/ath9k/ani.c
@@ -46,8 +46,8 @@ static const struct ani_ofdm_level_entry ofdm_level_table[] = {
 	{  5,  4,  1  }, /* lvl 5 */
 	{  6,  5,  1  }, /* lvl 6 */
 	{  7,  6,  1  }, /* lvl 7 */
-	{  7,  6,  0  }, /* lvl 8 */
-	{  7,  7,  0  }  /* lvl 9 */
+	{  7,  7,  1  }, /* lvl 8 */
+	{  7,  8,  0  }  /* lvl 9 */
 };
 #define ATH9K_ANI_OFDM_NUM_LEVEL \
 	ARRAY_SIZE(ofdm_level_table)
@@ -91,8 +91,8 @@ static const struct ani_cck_level_entry cck_level_table[] = {
 	{  4,  0  }, /* lvl 4 */
 	{  5,  0  }, /* lvl 5 */
 	{  6,  0  }, /* lvl 6 */
-	{  6,  0  }, /* lvl 7 (only for high rssi) */
-	{  7,  0  }  /* lvl 8 (only for high rssi) */
+	{  7,  0  }, /* lvl 7 (only for high rssi) */
+	{  8,  0  }  /* lvl 8 (only for high rssi) */
 };
 
 #define ATH9K_ANI_CCK_NUM_LEVEL \
@@ -118,10 +118,10 @@ static void ath9k_ani_restart(struct ath_hw *ah)
 {
 	struct ar5416AniState *aniState;
 
-	if (!DO_ANI(ah))
+	if (!ah->curchan)
 		return;
 
-	aniState = &ah->curchan->ani;
+	aniState = &ah->ani;
 	aniState->listenTime = 0;
 
 	ENABLE_REGWRITE_BUFFER(ah);
@@ -143,7 +143,7 @@ static void ath9k_ani_restart(struct ath_hw *ah)
 static void ath9k_hw_set_ofdm_nil(struct ath_hw *ah, u8 immunityLevel,
 				  bool scan)
 {
-	struct ar5416AniState *aniState = &ah->curchan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
 	struct ath_common *common = ath9k_hw_common(ah);
 	const struct ani_ofdm_level_entry *entry_ofdm;
 	const struct ani_cck_level_entry *entry_cck;
@@ -177,10 +177,15 @@ static void ath9k_hw_set_ofdm_nil(struct ath_hw *ah, u8 immunityLevel,
 	    BEACON_RSSI(ah) <= ATH9K_ANI_RSSI_THR_HIGH)
 		weak_sig = true;
 
-	if (aniState->ofdmWeakSigDetect != weak_sig)
-			ath9k_hw_ani_control(ah,
-				ATH9K_ANI_OFDM_WEAK_SIGNAL_DETECTION,
-				entry_ofdm->ofdm_weak_signal_on);
+	/*
+	 * OFDM Weak signal detection is always enabled for AP mode.
+	 */
+	if (ah->opmode != NL80211_IFTYPE_AP &&
+	    aniState->ofdmWeakSigDetect != weak_sig) {
+		ath9k_hw_ani_control(ah,
+				     ATH9K_ANI_OFDM_WEAK_SIGNAL_DETECTION,
+				     entry_ofdm->ofdm_weak_signal_on);
+	}
 
 	if (aniState->ofdmNoiseImmunityLevel >= ATH9K_ANI_OFDM_DEF_LEVEL) {
 		ah->config.ofdm_trig_high = ATH9K_ANI_OFDM_TRIG_HIGH;
@@ -195,10 +200,10 @@ static void ath9k_hw_ani_ofdm_err_trigger(struct ath_hw *ah)
 {
 	struct ar5416AniState *aniState;
 
-	if (!DO_ANI(ah))
+	if (!ah->curchan)
 		return;
 
-	aniState = &ah->curchan->ani;
+	aniState = &ah->ani;
 
 	if (aniState->ofdmNoiseImmunityLevel < ATH9K_ANI_OFDM_MAX_LEVEL)
 		ath9k_hw_set_ofdm_nil(ah, aniState->ofdmNoiseImmunityLevel + 1, false);
@@ -210,7 +215,7 @@ static void ath9k_hw_ani_ofdm_err_trigger(struct ath_hw *ah)
 static void ath9k_hw_set_cck_nil(struct ath_hw *ah, u_int8_t immunityLevel,
 				 bool scan)
 {
-	struct ar5416AniState *aniState = &ah->curchan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
 	struct ath_common *common = ath9k_hw_common(ah);
 	const struct ani_ofdm_level_entry *entry_ofdm;
 	const struct ani_cck_level_entry *entry_cck;
@@ -251,10 +256,10 @@ static void ath9k_hw_ani_cck_err_trigger(struct ath_hw *ah)
 {
 	struct ar5416AniState *aniState;
 
-	if (!DO_ANI(ah))
+	if (!ah->curchan)
 		return;
 
-	aniState = &ah->curchan->ani;
+	aniState = &ah->ani;
 
 	if (aniState->cckNoiseImmunityLevel < ATH9K_ANI_CCK_MAX_LEVEL)
 		ath9k_hw_set_cck_nil(ah, aniState->cckNoiseImmunityLevel + 1,
@@ -269,7 +274,7 @@ static void ath9k_hw_ani_lower_immunity(struct ath_hw *ah)
 {
 	struct ar5416AniState *aniState;
 
-	aniState = &ah->curchan->ani;
+	aniState = &ah->ani;
 
 	/* lower OFDM noise immunity */
 	if (aniState->ofdmNoiseImmunityLevel > 0 &&
@@ -292,12 +297,12 @@ static void ath9k_hw_ani_lower_immunity(struct ath_hw *ah)
  */
 void ath9k_ani_reset(struct ath_hw *ah, bool is_scanning)
 {
-	struct ar5416AniState *aniState = &ah->curchan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
 	struct ath9k_channel *chan = ah->curchan;
 	struct ath_common *common = ath9k_hw_common(ah);
 	int ofdm_nil, cck_nil;
 
-	if (!DO_ANI(ah))
+	if (!ah->curchan)
 		return;
 
 	BUG_ON(aniState == NULL);
@@ -363,24 +368,13 @@ void ath9k_ani_reset(struct ath_hw *ah, bool is_scanning)
 	ath9k_hw_set_ofdm_nil(ah, ofdm_nil, is_scanning);
 	ath9k_hw_set_cck_nil(ah, cck_nil, is_scanning);
 
-	/*
-	 * enable phy counters if hw supports or if not, enable phy
-	 * interrupts (so we can count each one)
-	 */
 	ath9k_ani_restart(ah);
-
-	ENABLE_REGWRITE_BUFFER(ah);
-
-	REG_WRITE(ah, AR_PHY_ERR_MASK_1, AR_PHY_ERR_OFDM_TIMING);
-	REG_WRITE(ah, AR_PHY_ERR_MASK_2, AR_PHY_ERR_CCK_TIMING);
-
-	REGWRITE_BUFFER_FLUSH(ah);
 }
 
 static bool ath9k_hw_ani_read_counters(struct ath_hw *ah)
 {
 	struct ath_common *common = ath9k_hw_common(ah);
-	struct ar5416AniState *aniState = &ah->curchan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
 	u32 phyCnt1, phyCnt2;
 	int32_t listenTime;
 
@@ -415,10 +409,10 @@ void ath9k_hw_ani_monitor(struct ath_hw *ah, struct ath9k_channel *chan)
 	struct ath_common *common = ath9k_hw_common(ah);
 	u32 ofdmPhyErrRate, cckPhyErrRate;
 
-	if (!DO_ANI(ah))
+	if (!ah->curchan)
 		return;
 
-	aniState = &ah->curchan->ani;
+	aniState = &ah->ani;
 	if (!ath9k_hw_ani_read_counters(ah))
 		return;
 
@@ -490,32 +484,22 @@ EXPORT_SYMBOL(ath9k_hw_disable_mib_counters);
 void ath9k_hw_ani_init(struct ath_hw *ah)
 {
 	struct ath_common *common = ath9k_hw_common(ah);
-	int i;
+	struct ar5416AniState *ani = &ah->ani;
 
 	ath_dbg(common, ANI, "Initialize ANI\n");
 
 	ah->config.ofdm_trig_high = ATH9K_ANI_OFDM_TRIG_HIGH;
 	ah->config.ofdm_trig_low = ATH9K_ANI_OFDM_TRIG_LOW;
-
 	ah->config.cck_trig_high = ATH9K_ANI_CCK_TRIG_HIGH;
 	ah->config.cck_trig_low = ATH9K_ANI_CCK_TRIG_LOW;
 
-	for (i = 0; i < ARRAY_SIZE(ah->channels); i++) {
-		struct ath9k_channel *chan = &ah->channels[i];
-		struct ar5416AniState *ani = &chan->ani;
-
-		ani->spurImmunityLevel = ATH9K_ANI_SPUR_IMMUNE_LVL;
-
-		ani->firstepLevel = ATH9K_ANI_FIRSTEP_LVL;
-
-		ani->mrcCCK = AR_SREV_9300_20_OR_LATER(ah) ? true : false;
-
-		ani->ofdmsTurn = true;
-
-		ani->ofdmWeakSigDetect = ATH9K_ANI_USE_OFDM_WEAK_SIG;
-		ani->cckNoiseImmunityLevel = ATH9K_ANI_CCK_DEF_LEVEL;
-		ani->ofdmNoiseImmunityLevel = ATH9K_ANI_OFDM_DEF_LEVEL;
-	}
+	ani->spurImmunityLevel = ATH9K_ANI_SPUR_IMMUNE_LVL;
+	ani->firstepLevel = ATH9K_ANI_FIRSTEP_LVL;
+	ani->mrcCCK = AR_SREV_9300_20_OR_LATER(ah) ? true : false;
+	ani->ofdmsTurn = true;
+	ani->ofdmWeakSigDetect = true;
+	ani->cckNoiseImmunityLevel = ATH9K_ANI_CCK_DEF_LEVEL;
+	ani->ofdmNoiseImmunityLevel = ATH9K_ANI_OFDM_DEF_LEVEL;
 
 	/*
 	 * since we expect some ongoing maintenance on the tables, let's sanity
@@ -524,9 +508,6 @@ void ath9k_hw_ani_init(struct ath_hw *ah)
 	ah->aniperiod = ATH9K_ANI_PERIOD;
 	ah->config.ani_poll_interval = ATH9K_ANI_POLLINTERVAL;
 
-	if (ah->config.enable_ani)
-		ah->proc_phyerr |= HAL_PROCESS_ANI;
-
 	ath9k_ani_restart(ah);
 	ath9k_enable_mib_counters(ah);
 }
diff --git a/drivers/net/wireless/ath/ath9k/ani.h b/drivers/net/wireless/ath/ath9k/ani.h
index dddb136..b54a3fb 100644
--- a/drivers/net/wireless/ath/ath9k/ani.h
+++ b/drivers/net/wireless/ath/ath9k/ani.h
@@ -17,32 +17,19 @@
 #ifndef ANI_H
 #define ANI_H
 
-#define HAL_PROCESS_ANI           0x00000001
-
-#define DO_ANI(ah) (((ah)->proc_phyerr & HAL_PROCESS_ANI) && ah->curchan)
-
 #define BEACON_RSSI(ahp) (ahp->stats.avgbrssi)
 
 /* units are errors per second */
-#define ATH9K_ANI_OFDM_TRIG_HIGH          3500
+#define ATH9K_ANI_OFDM_TRIG_HIGH           3500
 #define ATH9K_ANI_OFDM_TRIG_HIGH_BELOW_INI 1000
 
-/* units are errors per second */
 #define ATH9K_ANI_OFDM_TRIG_LOW           400
 #define ATH9K_ANI_OFDM_TRIG_LOW_ABOVE_INI 900
 
-/* units are errors per second */
 #define ATH9K_ANI_CCK_TRIG_HIGH           600
-
-/* units are errors per second */
 #define ATH9K_ANI_CCK_TRIG_LOW            300
 
-#define ATH9K_ANI_NOISE_IMMUNE_LVL        4
-#define ATH9K_ANI_USE_OFDM_WEAK_SIG       true
-#define ATH9K_ANI_CCK_WEAK_SIG_THR        false
-
 #define ATH9K_ANI_SPUR_IMMUNE_LVL         3
-
 #define ATH9K_ANI_FIRSTEP_LVL             2
 
 #define ATH9K_ANI_RSSI_THR_HIGH           40
@@ -53,10 +40,6 @@
 /* in ms */
 #define ATH9K_ANI_POLLINTERVAL            1000
 
-#define HAL_NOISE_IMMUNE_MAX              4
-#define HAL_SPUR_IMMUNE_MAX               7
-#define HAL_FIRST_STEP_MAX                2
-
 #define ATH9K_SIG_FIRSTEP_SETTING_MIN     0
 #define ATH9K_SIG_FIRSTEP_SETTING_MAX     20
 #define ATH9K_SIG_SPUR_IMM_SETTING_MIN    0
@@ -111,7 +94,7 @@ struct ar5416AniState {
 	u8 mrcCCK;
 	u8 spurImmunityLevel;
 	u8 firstepLevel;
-	u8 ofdmWeakSigDetect;
+	bool ofdmWeakSigDetect;
 	u32 listenTime;
 	u32 ofdmPhyErrCount;
 	u32 cckPhyErrCount;
@@ -119,8 +102,6 @@ struct ar5416AniState {
 };
 
 struct ar5416Stats {
-	u32 ast_ani_niup;
-	u32 ast_ani_nidown;
 	u32 ast_ani_spurup;
 	u32 ast_ani_spurdown;
 	u32 ast_ani_ofdmon;
diff --git a/drivers/net/wireless/ath/ath9k/ar5008_phy.c b/drivers/net/wireless/ath/ath9k/ar5008_phy.c
index 391da5a..d1acfe9 100644
--- a/drivers/net/wireless/ath/ath9k/ar5008_phy.c
+++ b/drivers/net/wireless/ath/ath9k/ar5008_phy.c
@@ -931,7 +931,7 @@ static bool ar5008_hw_ani_control_new(struct ath_hw *ah,
 {
 	struct ath_common *common = ath9k_hw_common(ah);
 	struct ath9k_channel *chan = ah->curchan;
-	struct ar5416AniState *aniState = &chan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
 	s32 value, value2;
 
 	switch (cmd & ah->ani_function) {
@@ -1207,7 +1207,7 @@ static void ar5008_hw_ani_cache_ini_regs(struct ath_hw *ah)
 {
 	struct ath_common *common = ath9k_hw_common(ah);
 	struct ath9k_channel *chan = ah->curchan;
-	struct ar5416AniState *aniState = &chan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
 	struct ath9k_ani_default *iniDef;
 	u32 val;
 
@@ -1251,7 +1251,7 @@ static void ar5008_hw_ani_cache_ini_regs(struct ath_hw *ah)
 	/* these levels just got reset to defaults by the INI */
 	aniState->spurImmunityLevel = ATH9K_ANI_SPUR_IMMUNE_LVL;
 	aniState->firstepLevel = ATH9K_ANI_FIRSTEP_LVL;
-	aniState->ofdmWeakSigDetect = ATH9K_ANI_USE_OFDM_WEAK_SIG;
+	aniState->ofdmWeakSigDetect = true;
 	aniState->mrcCCK = false; /* not available on pre AR9003 */
 }
 
diff --git a/drivers/net/wireless/ath/ath9k/ar9002_hw.c b/drivers/net/wireless/ath/ath9k/ar9002_hw.c
index 830daa1..8dc2d08 100644
--- a/drivers/net/wireless/ath/ath9k/ar9002_hw.c
+++ b/drivers/net/wireless/ath/ath9k/ar9002_hw.c
@@ -38,10 +38,6 @@ static int ar9002_hw_init_mode_regs(struct ath_hw *ah)
 	else
 		INIT_INI_ARRAY(&ah->iniPcieSerdes,
 			   ar9280PciePhy_clkreq_always_on_L1_9280);
-#ifdef CONFIG_PM_SLEEP
-		INIT_INI_ARRAY(&ah->iniPcieSerdesWow,
-			       ar9280PciePhy_awow);
-#endif
 
 	if (AR_SREV_9287_11_OR_LATER(ah)) {
 		INIT_INI_ARRAY(&ah->iniModes, ar9287Modes_9287_1_1);
diff --git a/drivers/net/wireless/ath/ath9k/ar9002_initvals.h b/drivers/net/wireless/ath/ath9k/ar9002_initvals.h
index beb6162..4d18c66 100644
--- a/drivers/net/wireless/ath/ath9k/ar9002_initvals.h
+++ b/drivers/net/wireless/ath/ath9k/ar9002_initvals.h
@@ -925,20 +925,6 @@ static const u32 ar9280PciePhy_clkreq_always_on_L1_9280[][2] = {
 	{0x00004044, 0x00000000},
 };
 
-static const u32 ar9280PciePhy_awow[][2] = {
-	/* Addr      allmodes  */
-	{0x00004040, 0x9248fd00},
-	{0x00004040, 0x24924924},
-	{0x00004040, 0xa8000019},
-	{0x00004040, 0x13160820},
-	{0x00004040, 0xe5980560},
-	{0x00004040, 0xc01dcffd},
-	{0x00004040, 0x1aaabe41},
-	{0x00004040, 0xbe105554},
-	{0x00004040, 0x00043007},
-	{0x00004044, 0x00000000},
-};
-
 static const u32 ar9285Modes_9285_1_2[][5] = {
 	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
 	{0x00001030, 0x00000230, 0x00000460, 0x000002c0, 0x00000160},
diff --git a/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c b/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c
index e6b92ff..d105e43 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c
@@ -3563,14 +3563,24 @@ static void ar9003_hw_ant_ctrl_apply(struct ath_hw *ah, bool is2ghz)
 {
 	struct ath9k_hw_capabilities *pCap = &ah->caps;
 	int chain;
-	u32 regval;
+	u32 regval, value, gpio;
 	static const u32 switch_chain_reg[AR9300_MAX_CHAINS] = {
 			AR_PHY_SWITCH_CHAIN_0,
 			AR_PHY_SWITCH_CHAIN_1,
 			AR_PHY_SWITCH_CHAIN_2,
 	};
 
-	u32 value = ar9003_hw_ant_ctrl_common_get(ah, is2ghz);
+	if (AR_SREV_9485(ah) && (ar9003_hw_get_rx_gain_idx(ah) == 0)) {
+		if (ah->config.xlna_gpio)
+			gpio = ah->config.xlna_gpio;
+		else
+			gpio = AR9300_EXT_LNA_CTL_GPIO_AR9485;
+
+		ath9k_hw_cfg_output(ah, gpio,
+				    AR_GPIO_OUTPUT_MUX_AS_PCIE_ATTENTION_LED);
+	}
+
+	value = ar9003_hw_ant_ctrl_common_get(ah, is2ghz);
 
 	if (AR_SREV_9462(ah) || AR_SREV_9565(ah)) {
 		REG_RMW_FIELD(ah, AR_PHY_SWITCH_COM,
@@ -3596,7 +3606,7 @@ static void ar9003_hw_ant_ctrl_apply(struct ath_hw *ah, bool is2ghz)
 	 *   7:4 R/W  SWITCH_TABLE_COM_SPDT_WLAN_IDLE
 	 * SWITCH_TABLE_COM_SPDT_WLAN_IDLE
 	 */
-	if (AR_SREV_9462_20(ah) || AR_SREV_9565(ah)) {
+	if (AR_SREV_9462_20_OR_LATER(ah) || AR_SREV_9565(ah)) {
 		value = ar9003_switch_com_spdt_get(ah, is2ghz);
 		REG_RMW_FIELD(ah, AR_PHY_GLB_CONTROL,
 				AR_SWITCH_TABLE_COM_SPDT_ALL, value);
@@ -3796,7 +3806,13 @@ static void ar9003_hw_atten_apply(struct ath_hw *ah, struct ath9k_channel *chan)
 			REG_RMW_FIELD(ah, ext_atten_reg[i],
 				      AR_PHY_EXT_ATTEN_CTL_XATTEN1_DB, value);
 
-			value = ar9003_hw_atten_chain_get_margin(ah, i, chan);
+			if (AR_SREV_9485(ah) &&
+			    (ar9003_hw_get_rx_gain_idx(ah) == 0) &&
+			    ah->config.xatten_margin_cfg)
+				value = 5;
+			else
+				value = ar9003_hw_atten_chain_get_margin(ah, i, chan);
+
 			REG_RMW_FIELD(ah, ext_atten_reg[i],
 				      AR_PHY_EXT_ATTEN_CTL_XATTEN1_MARGIN,
 				      value);
@@ -4043,8 +4059,9 @@ static void ar9003_hw_thermo_cal_apply(struct ath_hw *ah)
 {
 	u32 data, ko, kg;
 
-	if (!AR_SREV_9462_20(ah))
+	if (!AR_SREV_9462_20_OR_LATER(ah))
 		return;
+
 	ar9300_otp_read_word(ah, 1, &data);
 	ko = data & 0xff;
 	kg = (data >> 8) & 0xff;
@@ -4546,7 +4563,7 @@ static void ar9003_hw_get_target_power_eeprom(struct ath_hw *ah,
 						 is2GHz);
 
 	for (i = 0; i < ar9300RateSize; i++) {
-		ath_dbg(common, EEPROM, "TPC[%02d] 0x%08x\n",
+		ath_dbg(common, REGULATORY, "TPC[%02d] 0x%08x\n",
 			i, targetPowerValT2[i]);
 	}
 }
@@ -4736,7 +4753,7 @@ tempslope:
 			      AR_PHY_TPC_19_ALPHA_THERM, temp_slope);
 	}
 
-	if (AR_SREV_9462_20(ah))
+	if (AR_SREV_9462_20_OR_LATER(ah))
 		REG_RMW_FIELD(ah, AR_PHY_TPC_19_B1,
 			      AR_PHY_TPC_19_B1_ALPHA_THERM, temp_slope);
 
@@ -5272,7 +5289,7 @@ static void ath9k_hw_ar9300_set_txpower(struct ath_hw *ah,
 		return;
 
 	for (i = 0; i < ar9300RateSize; i++) {
-		ath_dbg(common, EEPROM, "TPC[%02d] 0x%08x\n",
+		ath_dbg(common, REGULATORY, "TPC[%02d] 0x%08x\n",
 			i, targetPowerValT2[i]);
 	}
 
diff --git a/drivers/net/wireless/ath/ath9k/ar9003_hw.c b/drivers/net/wireless/ath/ath9k/ar9003_hw.c
index a3523c9..d402cb3 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_hw.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_hw.c
@@ -24,6 +24,7 @@
 #include "ar955x_1p0_initvals.h"
 #include "ar9580_1p0_initvals.h"
 #include "ar9462_2p0_initvals.h"
+#include "ar9462_2p1_initvals.h"
 #include "ar9565_1p0_initvals.h"
 
 /* General hardware code for the AR9003 hadware family */
@@ -197,6 +198,31 @@ static void ar9003_hw_init_mode_regs(struct ath_hw *ah)
 
 		INIT_INI_ARRAY(&ah->iniPcieSerdesLowPower,
 				ar9485_1_1_pcie_phy_clkreq_disable_L1);
+	} else if (AR_SREV_9462_21(ah)) {
+		INIT_INI_ARRAY(&ah->iniMac[ATH_INI_CORE],
+			       ar9462_2p1_mac_core);
+		INIT_INI_ARRAY(&ah->iniMac[ATH_INI_POST],
+			       ar9462_2p1_mac_postamble);
+		INIT_INI_ARRAY(&ah->iniBB[ATH_INI_CORE],
+			       ar9462_2p1_baseband_core);
+		INIT_INI_ARRAY(&ah->iniBB[ATH_INI_POST],
+			       ar9462_2p1_baseband_postamble);
+		INIT_INI_ARRAY(&ah->iniRadio[ATH_INI_CORE],
+			       ar9462_2p1_radio_core);
+		INIT_INI_ARRAY(&ah->iniRadio[ATH_INI_POST],
+			       ar9462_2p1_radio_postamble);
+		INIT_INI_ARRAY(&ah->ini_radio_post_sys2ant,
+			       ar9462_2p1_radio_postamble_sys2ant);
+		INIT_INI_ARRAY(&ah->iniSOC[ATH_INI_PRE],
+			       ar9462_2p1_soc_preamble);
+		INIT_INI_ARRAY(&ah->iniSOC[ATH_INI_POST],
+			       ar9462_2p1_soc_postamble);
+		INIT_INI_ARRAY(&ah->iniModesRxGain,
+			       ar9462_2p1_common_rx_gain);
+		INIT_INI_ARRAY(&ah->iniModesFastClock,
+			       ar9462_2p1_modes_fast_clock);
+		INIT_INI_ARRAY(&ah->iniCckfirJapan2484,
+			       ar9462_2p1_baseband_core_txfir_coeff_japan_2484);
 	} else if (AR_SREV_9462_20(ah)) {
 
 		INIT_INI_ARRAY(&ah->iniMac[ATH_INI_CORE], ar9462_2p0_mac_core);
@@ -407,6 +433,9 @@ static void ar9003_tx_gain_table_mode0(struct ath_hw *ah)
 	else if (AR_SREV_9580(ah))
 		INIT_INI_ARRAY(&ah->iniModesTxGain,
 			ar9580_1p0_lowest_ob_db_tx_gain_table);
+	else if (AR_SREV_9462_21(ah))
+		INIT_INI_ARRAY(&ah->iniModesTxGain,
+			ar9462_2p1_modes_low_ob_db_tx_gain);
 	else if (AR_SREV_9462_20(ah))
 		INIT_INI_ARRAY(&ah->iniModesTxGain,
 			ar9462_modes_low_ob_db_tx_gain_table_2p0);
@@ -438,6 +467,9 @@ static void ar9003_tx_gain_table_mode1(struct ath_hw *ah)
 	else if (AR_SREV_9550(ah))
 		INIT_INI_ARRAY(&ah->iniModesTxGain,
 			ar955x_1p0_modes_no_xpa_tx_gain_table);
+	else if (AR_SREV_9462_21(ah))
+		INIT_INI_ARRAY(&ah->iniModesTxGain,
+			ar9462_2p1_modes_high_ob_db_tx_gain);
 	else if (AR_SREV_9462_20(ah))
 		INIT_INI_ARRAY(&ah->iniModesTxGain,
 			ar9462_modes_high_ob_db_tx_gain_table_2p0);
@@ -507,6 +539,12 @@ static void ar9003_tx_gain_table_mode4(struct ath_hw *ah)
 	else if (AR_SREV_9580(ah))
 		INIT_INI_ARRAY(&ah->iniModesTxGain,
 			ar9580_1p0_mixed_ob_db_tx_gain_table);
+	else if (AR_SREV_9462_21(ah))
+		INIT_INI_ARRAY(&ah->iniModesTxGain,
+		       ar9462_2p1_modes_mix_ob_db_tx_gain);
+	else if (AR_SREV_9462_20(ah))
+		INIT_INI_ARRAY(&ah->iniModesTxGain,
+		       ar9462_modes_mix_ob_db_tx_gain_table_2p0);
 	else
 		INIT_INI_ARRAY(&ah->iniModesTxGain,
 			ar9300Modes_mixed_ob_db_tx_gain_table_2p2);
@@ -584,6 +622,9 @@ static void ar9003_rx_gain_table_mode0(struct ath_hw *ah)
 	} else if (AR_SREV_9580(ah))
 		INIT_INI_ARRAY(&ah->iniModesRxGain,
 				ar9580_1p0_rx_gain_table);
+	else if (AR_SREV_9462_21(ah))
+		INIT_INI_ARRAY(&ah->iniModesRxGain,
+				ar9462_2p1_common_rx_gain);
 	else if (AR_SREV_9462_20(ah))
 		INIT_INI_ARRAY(&ah->iniModesRxGain,
 				ar9462_common_rx_gain_table_2p0);
@@ -606,6 +647,9 @@ static void ar9003_rx_gain_table_mode1(struct ath_hw *ah)
 	else if (AR_SREV_9485_11(ah))
 		INIT_INI_ARRAY(&ah->iniModesRxGain,
 			ar9485Common_wo_xlna_rx_gain_1_1);
+	else if (AR_SREV_9462_21(ah))
+		INIT_INI_ARRAY(&ah->iniModesRxGain,
+			ar9462_2p1_common_wo_xlna_rx_gain);
 	else if (AR_SREV_9462_20(ah))
 		INIT_INI_ARRAY(&ah->iniModesRxGain,
 			ar9462_common_wo_xlna_rx_gain_table_2p0);
@@ -627,9 +671,40 @@ static void ar9003_rx_gain_table_mode1(struct ath_hw *ah)
 
 static void ar9003_rx_gain_table_mode2(struct ath_hw *ah)
 {
-	if (AR_SREV_9462_20(ah))
+	if (AR_SREV_9462_21(ah)) {
+		INIT_INI_ARRAY(&ah->iniModesRxGain,
+			       ar9462_2p1_common_mixed_rx_gain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_bb_core,
+			       ar9462_2p1_baseband_core_mix_rxgain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_bb_postamble,
+			       ar9462_2p1_baseband_postamble_mix_rxgain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_5g_xlna,
+			       ar9462_2p1_baseband_postamble_5g_xlna);
+	} else if (AR_SREV_9462_20(ah)) {
 		INIT_INI_ARRAY(&ah->iniModesRxGain,
 			       ar9462_common_mixed_rx_gain_table_2p0);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_bb_core,
+			       ar9462_2p0_baseband_core_mix_rxgain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_bb_postamble,
+			       ar9462_2p0_baseband_postamble_mix_rxgain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_5g_xlna,
+			       ar9462_2p0_baseband_postamble_5g_xlna);
+	}
+}
+
+static void ar9003_rx_gain_table_mode3(struct ath_hw *ah)
+{
+	if (AR_SREV_9462_21(ah)) {
+		INIT_INI_ARRAY(&ah->iniModesRxGain,
+			       ar9462_2p1_common_5g_xlna_only_rx_gain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_5g_xlna,
+			       ar9462_2p1_baseband_postamble_5g_xlna);
+	} else if (AR_SREV_9462_20(ah)) {
+		INIT_INI_ARRAY(&ah->iniModesRxGain,
+			       ar9462_2p0_5g_xlna_only_rxgain);
+		INIT_INI_ARRAY(&ah->ini_modes_rxgain_5g_xlna,
+			       ar9462_2p0_baseband_postamble_5g_xlna);
+	}
 }
 
 static void ar9003_rx_gain_table_apply(struct ath_hw *ah)
@@ -645,6 +720,9 @@ static void ar9003_rx_gain_table_apply(struct ath_hw *ah)
 	case 2:
 		ar9003_rx_gain_table_mode2(ah);
 		break;
+	case 3:
+		ar9003_rx_gain_table_mode3(ah);
+		break;
 	}
 }
 
diff --git a/drivers/net/wireless/ath/ath9k/ar9003_mac.c b/drivers/net/wireless/ath/ath9k/ar9003_mac.c
index 301bf72..5163abd 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_mac.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_mac.c
@@ -469,6 +469,7 @@ int ath9k_hw_process_rxdesc_edma(struct ath_hw *ah, struct ath_rx_status *rxs,
 
 	rxs->rs_status = 0;
 	rxs->rs_flags =  0;
+	rxs->flag =  0;
 
 	rxs->rs_datalen = rxsp->status2 & AR_DataLen;
 	rxs->rs_tstamp =  rxsp->status3;
@@ -493,8 +494,8 @@ int ath9k_hw_process_rxdesc_edma(struct ath_hw *ah, struct ath_rx_status *rxs,
 	rxs->rs_isaggr = (rxsp->status11 & AR_RxAggr) ? 1 : 0;
 	rxs->rs_moreaggr = (rxsp->status11 & AR_RxMoreAggr) ? 1 : 0;
 	rxs->rs_antenna = (MS(rxsp->status4, AR_RxAntenna) & 0x7);
-	rxs->rs_flags  = (rxsp->status4 & AR_GI) ? ATH9K_RX_GI : 0;
-	rxs->rs_flags  |= (rxsp->status4 & AR_2040) ? ATH9K_RX_2040 : 0;
+	rxs->flag  |= (rxsp->status4 & AR_GI) ? RX_FLAG_SHORT_GI : 0;
+	rxs->flag  |= (rxsp->status4 & AR_2040) ? RX_FLAG_40MHZ : 0;
 
 	rxs->evm0 = rxsp->status6;
 	rxs->evm1 = rxsp->status7;
diff --git a/drivers/net/wireless/ath/ath9k/ar9003_paprd.c b/drivers/net/wireless/ath/ath9k/ar9003_paprd.c
index 09c1f9d..6343cc9 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_paprd.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_paprd.c
@@ -454,6 +454,8 @@ static bool create_pa_curve(u32 *data_L, u32 *data_U, u32 *pa_table, u16 *gain)
 		if (accum_cnt <= thresh_accum_cnt)
 			continue;
 
+		max_index++;
+
 		/* sum(tx amplitude) */
 		accum_tx = ((data_L[i] >> 16) & 0xffff) |
 		    ((data_U[i] & 0x7ff) << 16);
@@ -468,20 +470,21 @@ static bool create_pa_curve(u32 *data_L, u32 *data_U, u32 *pa_table, u16 *gain)
 
 		accum_tx <<= scale_factor;
 		accum_rx <<= scale_factor;
-		x_est[i + 1] = (((accum_tx + accum_cnt) / accum_cnt) + 32) >>
-		    scale_factor;
+		x_est[max_index] =
+			(((accum_tx + accum_cnt) / accum_cnt) + 32) >>
+			scale_factor;
 
-		Y[i + 1] = ((((accum_rx + accum_cnt) / accum_cnt) + 32) >>
+		Y[max_index] =
+			((((accum_rx + accum_cnt) / accum_cnt) + 32) >>
 			    scale_factor) +
-			    (1 << scale_factor) * max_index + 16;
+			(1 << scale_factor) * i + 16;
 
 		if (accum_ang >= (1 << 26))
 			accum_ang -= 1 << 27;
 
-		theta[i + 1] = ((accum_ang * (1 << scale_factor)) + accum_cnt) /
-		    accum_cnt;
-
-		max_index++;
+		theta[max_index] =
+			((accum_ang * (1 << scale_factor)) + accum_cnt) /
+			accum_cnt;
 	}
 
 	/*
diff --git a/drivers/net/wireless/ath/ath9k/ar9003_phy.c b/drivers/net/wireless/ath/ath9k/ar9003_phy.c
index e1714d7..1f694ab 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_phy.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_phy.c
@@ -735,22 +735,53 @@ static int ar9003_hw_process_ini(struct ath_hw *ah,
 		return -EINVAL;
 	}
 
+	/*
+	 * SOC, MAC, BB, RADIO initvals.
+	 */
 	for (i = 0; i < ATH_INI_NUM_SPLIT; i++) {
 		ar9003_hw_prog_ini(ah, &ah->iniSOC[i], modesIndex);
 		ar9003_hw_prog_ini(ah, &ah->iniMac[i], modesIndex);
 		ar9003_hw_prog_ini(ah, &ah->iniBB[i], modesIndex);
 		ar9003_hw_prog_ini(ah, &ah->iniRadio[i], modesIndex);
-		if (i == ATH_INI_POST && AR_SREV_9462_20(ah))
+		if (i == ATH_INI_POST && AR_SREV_9462_20_OR_LATER(ah))
 			ar9003_hw_prog_ini(ah,
 					   &ah->ini_radio_post_sys2ant,
 					   modesIndex);
 	}
 
+	/*
+	 * RXGAIN initvals.
+	 */
 	REG_WRITE_ARRAY(&ah->iniModesRxGain, 1, regWrites);
+
+	if (AR_SREV_9462_20_OR_LATER(ah)) {
+		/*
+		 * CUS217 mix LNA mode.
+		 */
+		if (ar9003_hw_get_rx_gain_idx(ah) == 2) {
+			REG_WRITE_ARRAY(&ah->ini_modes_rxgain_bb_core,
+					1, regWrites);
+			REG_WRITE_ARRAY(&ah->ini_modes_rxgain_bb_postamble,
+					modesIndex, regWrites);
+		}
+
+		/*
+		 * 5G-XLNA
+		 */
+		if ((ar9003_hw_get_rx_gain_idx(ah) == 2) ||
+		    (ar9003_hw_get_rx_gain_idx(ah) == 3)) {
+			REG_WRITE_ARRAY(&ah->ini_modes_rxgain_5g_xlna,
+					modesIndex, regWrites);
+		}
+	}
+
 	if (AR_SREV_9550(ah))
 		REG_WRITE_ARRAY(&ah->ini_modes_rx_gain_bounds, modesIndex,
 				regWrites);
 
+	/*
+	 * TXGAIN initvals.
+	 */
 	if (AR_SREV_9550(ah)) {
 		int modes_txgain_index;
 
@@ -772,8 +803,14 @@ static int ar9003_hw_process_ini(struct ath_hw *ah,
 		REG_WRITE_ARRAY(&ah->iniModesFastClock,
 				modesIndex, regWrites);
 
+	/*
+	 * Clock frequency initvals.
+	 */
 	REG_WRITE_ARRAY(&ah->iniAdditional, 1, regWrites);
 
+	/*
+	 * JAPAN regulatory.
+	 */
 	if (chan->channel == 2484)
 		ar9003_hw_prog_ini(ah, &ah->iniCckfirJapan2484, 1);
 
@@ -905,7 +942,12 @@ static bool ar9003_hw_ani_control(struct ath_hw *ah,
 {
 	struct ath_common *common = ath9k_hw_common(ah);
 	struct ath9k_channel *chan = ah->curchan;
-	struct ar5416AniState *aniState = &chan->ani;
+	struct ar5416AniState *aniState = &ah->ani;
+	int m1ThreshLow, m2ThreshLow;
+	int m1Thresh, m2Thresh;
+	int m2CountThr, m2CountThrLow;
+	int m1ThreshLowExt, m2ThreshLowExt;
+	int m1ThreshExt, m2ThreshExt;
 	s32 value, value2;
 
 	switch (cmd & ah->ani_function) {
@@ -919,6 +961,61 @@ static bool ar9003_hw_ani_control(struct ath_hw *ah,
 		 */
 		u32 on = param ? 1 : 0;
 
+		if (AR_SREV_9462(ah) || AR_SREV_9565(ah))
+			goto skip_ws_det;
+
+		m1ThreshLow = on ?
+			aniState->iniDef.m1ThreshLow : m1ThreshLow_off;
+		m2ThreshLow = on ?
+			aniState->iniDef.m2ThreshLow : m2ThreshLow_off;
+		m1Thresh = on ?
+			aniState->iniDef.m1Thresh : m1Thresh_off;
+		m2Thresh = on ?
+			aniState->iniDef.m2Thresh : m2Thresh_off;
+		m2CountThr = on ?
+			aniState->iniDef.m2CountThr : m2CountThr_off;
+		m2CountThrLow = on ?
+			aniState->iniDef.m2CountThrLow : m2CountThrLow_off;
+		m1ThreshLowExt = on ?
+			aniState->iniDef.m1ThreshLowExt : m1ThreshLowExt_off;
+		m2ThreshLowExt = on ?
+			aniState->iniDef.m2ThreshLowExt : m2ThreshLowExt_off;
+		m1ThreshExt = on ?
+			aniState->iniDef.m1ThreshExt : m1ThreshExt_off;
+		m2ThreshExt = on ?
+			aniState->iniDef.m2ThreshExt : m2ThreshExt_off;
+
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_LOW,
+			      AR_PHY_SFCORR_LOW_M1_THRESH_LOW,
+			      m1ThreshLow);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_LOW,
+			      AR_PHY_SFCORR_LOW_M2_THRESH_LOW,
+			      m2ThreshLow);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR,
+			      AR_PHY_SFCORR_M1_THRESH,
+			      m1Thresh);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR,
+			      AR_PHY_SFCORR_M2_THRESH,
+			      m2Thresh);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR,
+			      AR_PHY_SFCORR_M2COUNT_THR,
+			      m2CountThr);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_LOW,
+			      AR_PHY_SFCORR_LOW_M2COUNT_THR_LOW,
+			      m2CountThrLow);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_EXT,
+			      AR_PHY_SFCORR_EXT_M1_THRESH_LOW,
+			      m1ThreshLowExt);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_EXT,
+			      AR_PHY_SFCORR_EXT_M2_THRESH_LOW,
+			      m2ThreshLowExt);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_EXT,
+			      AR_PHY_SFCORR_EXT_M1_THRESH,
+			      m1ThreshExt);
+		REG_RMW_FIELD(ah, AR_PHY_SFCORR_EXT,
+			      AR_PHY_SFCORR_EXT_M2_THRESH,
+			      m2ThreshExt);
+skip_ws_det:
 		if (on)
 			REG_SET_BIT(ah, AR_PHY_SFCORR_LOW,
 				    AR_PHY_SFCORR_LOW_USE_SELF_CORR_LOW);
@@ -1173,7 +1270,7 @@ static void ar9003_hw_ani_cache_ini_regs(struct ath_hw *ah)
 	struct ath9k_ani_default *iniDef;
 	u32 val;
 
-	aniState = &ah->curchan->ani;
+	aniState = &ah->ani;
 	iniDef = &aniState->iniDef;
 
 	ath_dbg(common, ANI, "ver %d.%d opmode %u chan %d Mhz/0x%x\n",
@@ -1214,7 +1311,7 @@ static void ar9003_hw_ani_cache_ini_regs(struct ath_hw *ah)
 	/* these levels just got reset to defaults by the INI */
 	aniState->spurImmunityLevel = ATH9K_ANI_SPUR_IMMUNE_LVL;
 	aniState->firstepLevel = ATH9K_ANI_FIRSTEP_LVL;
-	aniState->ofdmWeakSigDetect = ATH9K_ANI_USE_OFDM_WEAK_SIG;
+	aniState->ofdmWeakSigDetect = true;
 	aniState->mrcCCK = true;
 }
 
@@ -1415,7 +1512,7 @@ static int ar9003_hw_fast_chan_change(struct ath_hw *ah,
 	ar9003_hw_prog_ini(ah, &ah->iniBB[ATH_INI_POST], modesIndex);
 	ar9003_hw_prog_ini(ah, &ah->iniRadio[ATH_INI_POST], modesIndex);
 
-	if (AR_SREV_9462_20(ah))
+	if (AR_SREV_9462_20_OR_LATER(ah))
 		ar9003_hw_prog_ini(ah, &ah->ini_radio_post_sys2ant,
 				   modesIndex);
 
diff --git a/drivers/net/wireless/ath/ath9k/ar9003_phy.h b/drivers/net/wireless/ath/ath9k/ar9003_phy.h
index e717741..d4d39f3 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_phy.h
+++ b/drivers/net/wireless/ath/ath9k/ar9003_phy.h
@@ -351,6 +351,8 @@
 
 #define AR_PHY_CCA_NOM_VAL_9330_2GHZ          -118
 
+#define AR9300_EXT_LNA_CTL_GPIO_AR9485 9
+
 /*
  * AGC Field Definitions
  */
@@ -952,7 +954,7 @@
 #define AR_PHY_TPC_5_B1         (AR_SM1_BASE + 0x208)
 #define AR_PHY_TPC_6_B1         (AR_SM1_BASE + 0x20c)
 #define AR_PHY_TPC_11_B1        (AR_SM1_BASE + 0x220)
-#define AR_PHY_PDADC_TAB_1	(AR_SM1_BASE + (AR_SREV_AR9462(ah) ? \
+#define AR_PHY_PDADC_TAB_1	(AR_SM1_BASE + (AR_SREV_9462_20_OR_LATER(ah) ? \
 					0x280 : 0x240))
 #define AR_PHY_TPC_19_B1	(AR_SM1_BASE + 0x240)
 #define AR_PHY_TPC_19_B1_ALPHA_THERM		0xff
@@ -1046,7 +1048,7 @@
 #define AR_GLB_GPIO_CONTROL	(AR_GLB_BASE)
 #define AR_PHY_GLB_CONTROL	(AR_GLB_BASE + 0x44)
 #define AR_GLB_SCRATCH(_ah)	(AR_GLB_BASE + \
-					(AR_SREV_9462_20(_ah) ? 0x4c : 0x50))
+					(AR_SREV_9462_20_OR_LATER(_ah) ? 0x4c : 0x50))
 #define AR_GLB_STATUS		(AR_GLB_BASE + 0x48)
 
 /*
diff --git a/drivers/net/wireless/ath/ath9k/ar9462_2p0_initvals.h b/drivers/net/wireless/ath/ath9k/ar9462_2p0_initvals.h
index 999ab08..092b9d4 100644
--- a/drivers/net/wireless/ath/ath9k/ar9462_2p0_initvals.h
+++ b/drivers/net/wireless/ath/ath9k/ar9462_2p0_initvals.h
@@ -78,7 +78,7 @@ static const u32 ar9462_2p0_baseband_postamble[][5] = {
 	{0x0000a284, 0x00000000, 0x00000000, 0x00000150, 0x00000150},
 	{0x0000a288, 0x00000110, 0x00000110, 0x00000110, 0x00000110},
 	{0x0000a28c, 0x00022222, 0x00022222, 0x00022222, 0x00022222},
-	{0x0000a2c4, 0x00058d18, 0x00058d18, 0x00058d18, 0x00058d18},
+	{0x0000a2c4, 0x00158d18, 0x00158d18, 0x00158d18, 0x00158d18},
 	{0x0000a2d0, 0x00041981, 0x00041981, 0x00041981, 0x00041982},
 	{0x0000a2d8, 0x7999a83b, 0x7999a83b, 0x7999a83b, 0x7999a83b},
 	{0x0000a358, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
@@ -879,6 +879,69 @@ static const u32 ar9462_2p0_radio_postamble[][5] = {
 	{0x0001650c, 0x48000000, 0x40000000, 0x40000000, 0x40000000},
 };
 
+static const u32 ar9462_modes_mix_ob_db_tx_gain_table_2p0[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x000098bc, 0x00000002, 0x00000002, 0x00000002, 0x00000002},
+	{0x0000a2dc, 0x01feee00, 0x01feee00, 0x03aaa352, 0x03aaa352},
+	{0x0000a2e0, 0x0000f000, 0x0000f000, 0x03ccc584, 0x03ccc584},
+	{0x0000a2e4, 0x01ff0000, 0x01ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000a2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+	{0x0000a410, 0x0000d0da, 0x0000d0da, 0x0000d0de, 0x0000d0de},
+	{0x0000a458, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a500, 0x00002220, 0x00002220, 0x00000000, 0x00000000},
+	{0x0000a504, 0x06002223, 0x06002223, 0x04000002, 0x04000002},
+	{0x0000a508, 0x0a022220, 0x0a022220, 0x08000004, 0x08000004},
+	{0x0000a50c, 0x0f022223, 0x0f022223, 0x0b000200, 0x0b000200},
+	{0x0000a510, 0x14022620, 0x14022620, 0x0f000202, 0x0f000202},
+	{0x0000a514, 0x18022622, 0x18022622, 0x12000400, 0x12000400},
+	{0x0000a518, 0x1b022822, 0x1b022822, 0x16000402, 0x16000402},
+	{0x0000a51c, 0x20022842, 0x20022842, 0x19000404, 0x19000404},
+	{0x0000a520, 0x22022c41, 0x22022c41, 0x1c000603, 0x1c000603},
+	{0x0000a524, 0x28023042, 0x28023042, 0x21000a02, 0x21000a02},
+	{0x0000a528, 0x2c023044, 0x2c023044, 0x25000a04, 0x25000a04},
+	{0x0000a52c, 0x2f023644, 0x2f023644, 0x28000a20, 0x28000a20},
+	{0x0000a530, 0x34025643, 0x34025643, 0x2c000e20, 0x2c000e20},
+	{0x0000a534, 0x38025a44, 0x38025a44, 0x30000e22, 0x30000e22},
+	{0x0000a538, 0x3b025e45, 0x3b025e45, 0x34000e24, 0x34000e24},
+	{0x0000a53c, 0x41025e4a, 0x41025e4a, 0x38001640, 0x38001640},
+	{0x0000a540, 0x48025e6c, 0x48025e6c, 0x3c001660, 0x3c001660},
+	{0x0000a544, 0x4e025e8e, 0x4e025e8e, 0x3f001861, 0x3f001861},
+	{0x0000a548, 0x55025eb3, 0x55025eb3, 0x43001a81, 0x43001a81},
+	{0x0000a54c, 0x58025ef3, 0x58025ef3, 0x47001a83, 0x47001a83},
+	{0x0000a550, 0x5d025ef6, 0x5d025ef6, 0x4a001c84, 0x4a001c84},
+	{0x0000a554, 0x62025f56, 0x62025f56, 0x4e001ce3, 0x4e001ce3},
+	{0x0000a558, 0x66027f56, 0x66027f56, 0x52001ce5, 0x52001ce5},
+	{0x0000a55c, 0x6a029f56, 0x6a029f56, 0x56001ce9, 0x56001ce9},
+	{0x0000a560, 0x70049f56, 0x70049f56, 0x5a001ceb, 0x5a001ceb},
+	{0x0000a564, 0x751ffff6, 0x751ffff6, 0x5c001eec, 0x5c001eec},
+	{0x0000a568, 0x751ffff6, 0x751ffff6, 0x5e001ef0, 0x5e001ef0},
+	{0x0000a56c, 0x751ffff6, 0x751ffff6, 0x60001ef4, 0x60001ef4},
+	{0x0000a570, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a574, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a578, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a57c, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a600, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a604, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a608, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a60c, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a610, 0x00804000, 0x00804000, 0x00000000, 0x00000000},
+	{0x0000a614, 0x00804201, 0x00804201, 0x01404000, 0x01404000},
+	{0x0000a618, 0x0280c802, 0x0280c802, 0x01404501, 0x01404501},
+	{0x0000a61c, 0x0280ca03, 0x0280ca03, 0x02008501, 0x02008501},
+	{0x0000a620, 0x04c15104, 0x04c15104, 0x0280ca03, 0x0280ca03},
+	{0x0000a624, 0x04c15305, 0x04c15305, 0x03010c04, 0x03010c04},
+	{0x0000a628, 0x04c15305, 0x04c15305, 0x04014c04, 0x04014c04},
+	{0x0000a62c, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a630, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a634, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a638, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a63c, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000b2dc, 0x01feee00, 0x01feee00, 0x03aaa352, 0x03aaa352},
+	{0x0000b2e0, 0x0000f000, 0x0000f000, 0x03ccc584, 0x03ccc584},
+	{0x0000b2e4, 0x01ff0000, 0x01ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000b2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+};
+
 static const u32 ar9462_modes_high_ob_db_tx_gain_table_2p0[][5] = {
 	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
 	{0x000098bc, 0x00000002, 0x00000002, 0x00000002, 0x00000002},
@@ -1449,4 +1512,284 @@ static const u32 ar9462_common_mixed_rx_gain_table_2p0[][2] = {
 	{0x0000b1fc, 0x00000196},
 };
 
+static const u32 ar9462_2p0_baseband_postamble_5g_xlna[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00009e3c, 0xcf946220, 0xcf946220, 0xcfd5c782, 0xcfd5c282},
+};
+
+static const u32 ar9462_2p0_5g_xlna_only_rxgain[][2] = {
+	/* Addr      allmodes  */
+	{0x0000a000, 0x00010000},
+	{0x0000a004, 0x00030002},
+	{0x0000a008, 0x00050004},
+	{0x0000a00c, 0x00810080},
+	{0x0000a010, 0x00830082},
+	{0x0000a014, 0x01810180},
+	{0x0000a018, 0x01830182},
+	{0x0000a01c, 0x01850184},
+	{0x0000a020, 0x01890188},
+	{0x0000a024, 0x018b018a},
+	{0x0000a028, 0x018d018c},
+	{0x0000a02c, 0x03820190},
+	{0x0000a030, 0x03840383},
+	{0x0000a034, 0x03880385},
+	{0x0000a038, 0x038a0389},
+	{0x0000a03c, 0x038c038b},
+	{0x0000a040, 0x0390038d},
+	{0x0000a044, 0x03920391},
+	{0x0000a048, 0x03940393},
+	{0x0000a04c, 0x03960395},
+	{0x0000a050, 0x00000000},
+	{0x0000a054, 0x00000000},
+	{0x0000a058, 0x00000000},
+	{0x0000a05c, 0x00000000},
+	{0x0000a060, 0x00000000},
+	{0x0000a064, 0x00000000},
+	{0x0000a068, 0x00000000},
+	{0x0000a06c, 0x00000000},
+	{0x0000a070, 0x00000000},
+	{0x0000a074, 0x00000000},
+	{0x0000a078, 0x00000000},
+	{0x0000a07c, 0x00000000},
+	{0x0000a080, 0x29292929},
+	{0x0000a084, 0x29292929},
+	{0x0000a088, 0x29292929},
+	{0x0000a08c, 0x29292929},
+	{0x0000a090, 0x22292929},
+	{0x0000a094, 0x1d1d2222},
+	{0x0000a098, 0x0c111117},
+	{0x0000a09c, 0x00030303},
+	{0x0000a0a0, 0x00000000},
+	{0x0000a0a4, 0x00000000},
+	{0x0000a0a8, 0x00000000},
+	{0x0000a0ac, 0x00000000},
+	{0x0000a0b0, 0x00000000},
+	{0x0000a0b4, 0x00000000},
+	{0x0000a0b8, 0x00000000},
+	{0x0000a0bc, 0x00000000},
+	{0x0000a0c0, 0x001f0000},
+	{0x0000a0c4, 0x01000101},
+	{0x0000a0c8, 0x011e011f},
+	{0x0000a0cc, 0x011c011d},
+	{0x0000a0d0, 0x02030204},
+	{0x0000a0d4, 0x02010202},
+	{0x0000a0d8, 0x021f0200},
+	{0x0000a0dc, 0x0302021e},
+	{0x0000a0e0, 0x03000301},
+	{0x0000a0e4, 0x031e031f},
+	{0x0000a0e8, 0x0402031d},
+	{0x0000a0ec, 0x04000401},
+	{0x0000a0f0, 0x041e041f},
+	{0x0000a0f4, 0x0502041d},
+	{0x0000a0f8, 0x05000501},
+	{0x0000a0fc, 0x051e051f},
+	{0x0000a100, 0x06010602},
+	{0x0000a104, 0x061f0600},
+	{0x0000a108, 0x061d061e},
+	{0x0000a10c, 0x07020703},
+	{0x0000a110, 0x07000701},
+	{0x0000a114, 0x00000000},
+	{0x0000a118, 0x00000000},
+	{0x0000a11c, 0x00000000},
+	{0x0000a120, 0x00000000},
+	{0x0000a124, 0x00000000},
+	{0x0000a128, 0x00000000},
+	{0x0000a12c, 0x00000000},
+	{0x0000a130, 0x00000000},
+	{0x0000a134, 0x00000000},
+	{0x0000a138, 0x00000000},
+	{0x0000a13c, 0x00000000},
+	{0x0000a140, 0x001f0000},
+	{0x0000a144, 0x01000101},
+	{0x0000a148, 0x011e011f},
+	{0x0000a14c, 0x011c011d},
+	{0x0000a150, 0x02030204},
+	{0x0000a154, 0x02010202},
+	{0x0000a158, 0x021f0200},
+	{0x0000a15c, 0x0302021e},
+	{0x0000a160, 0x03000301},
+	{0x0000a164, 0x031e031f},
+	{0x0000a168, 0x0402031d},
+	{0x0000a16c, 0x04000401},
+	{0x0000a170, 0x041e041f},
+	{0x0000a174, 0x0502041d},
+	{0x0000a178, 0x05000501},
+	{0x0000a17c, 0x051e051f},
+	{0x0000a180, 0x06010602},
+	{0x0000a184, 0x061f0600},
+	{0x0000a188, 0x061d061e},
+	{0x0000a18c, 0x07020703},
+	{0x0000a190, 0x07000701},
+	{0x0000a194, 0x00000000},
+	{0x0000a198, 0x00000000},
+	{0x0000a19c, 0x00000000},
+	{0x0000a1a0, 0x00000000},
+	{0x0000a1a4, 0x00000000},
+	{0x0000a1a8, 0x00000000},
+	{0x0000a1ac, 0x00000000},
+	{0x0000a1b0, 0x00000000},
+	{0x0000a1b4, 0x00000000},
+	{0x0000a1b8, 0x00000000},
+	{0x0000a1bc, 0x00000000},
+	{0x0000a1c0, 0x00000000},
+	{0x0000a1c4, 0x00000000},
+	{0x0000a1c8, 0x00000000},
+	{0x0000a1cc, 0x00000000},
+	{0x0000a1d0, 0x00000000},
+	{0x0000a1d4, 0x00000000},
+	{0x0000a1d8, 0x00000000},
+	{0x0000a1dc, 0x00000000},
+	{0x0000a1e0, 0x00000000},
+	{0x0000a1e4, 0x00000000},
+	{0x0000a1e8, 0x00000000},
+	{0x0000a1ec, 0x00000000},
+	{0x0000a1f0, 0x00000396},
+	{0x0000a1f4, 0x00000396},
+	{0x0000a1f8, 0x00000396},
+	{0x0000a1fc, 0x00000196},
+	{0x0000b000, 0x00010000},
+	{0x0000b004, 0x00030002},
+	{0x0000b008, 0x00050004},
+	{0x0000b00c, 0x00810080},
+	{0x0000b010, 0x00830082},
+	{0x0000b014, 0x01810180},
+	{0x0000b018, 0x01830182},
+	{0x0000b01c, 0x01850184},
+	{0x0000b020, 0x02810280},
+	{0x0000b024, 0x02830282},
+	{0x0000b028, 0x02850284},
+	{0x0000b02c, 0x02890288},
+	{0x0000b030, 0x028b028a},
+	{0x0000b034, 0x0388028c},
+	{0x0000b038, 0x038a0389},
+	{0x0000b03c, 0x038c038b},
+	{0x0000b040, 0x0390038d},
+	{0x0000b044, 0x03920391},
+	{0x0000b048, 0x03940393},
+	{0x0000b04c, 0x03960395},
+	{0x0000b050, 0x00000000},
+	{0x0000b054, 0x00000000},
+	{0x0000b058, 0x00000000},
+	{0x0000b05c, 0x00000000},
+	{0x0000b060, 0x00000000},
+	{0x0000b064, 0x00000000},
+	{0x0000b068, 0x00000000},
+	{0x0000b06c, 0x00000000},
+	{0x0000b070, 0x00000000},
+	{0x0000b074, 0x00000000},
+	{0x0000b078, 0x00000000},
+	{0x0000b07c, 0x00000000},
+	{0x0000b080, 0x2a2d2f32},
+	{0x0000b084, 0x21232328},
+	{0x0000b088, 0x19191c1e},
+	{0x0000b08c, 0x12141417},
+	{0x0000b090, 0x07070e0e},
+	{0x0000b094, 0x03030305},
+	{0x0000b098, 0x00000003},
+	{0x0000b09c, 0x00000000},
+	{0x0000b0a0, 0x00000000},
+	{0x0000b0a4, 0x00000000},
+	{0x0000b0a8, 0x00000000},
+	{0x0000b0ac, 0x00000000},
+	{0x0000b0b0, 0x00000000},
+	{0x0000b0b4, 0x00000000},
+	{0x0000b0b8, 0x00000000},
+	{0x0000b0bc, 0x00000000},
+	{0x0000b0c0, 0x003f0020},
+	{0x0000b0c4, 0x00400041},
+	{0x0000b0c8, 0x0140005f},
+	{0x0000b0cc, 0x0160015f},
+	{0x0000b0d0, 0x017e017f},
+	{0x0000b0d4, 0x02410242},
+	{0x0000b0d8, 0x025f0240},
+	{0x0000b0dc, 0x027f0260},
+	{0x0000b0e0, 0x0341027e},
+	{0x0000b0e4, 0x035f0340},
+	{0x0000b0e8, 0x037f0360},
+	{0x0000b0ec, 0x04400441},
+	{0x0000b0f0, 0x0460045f},
+	{0x0000b0f4, 0x0541047f},
+	{0x0000b0f8, 0x055f0540},
+	{0x0000b0fc, 0x057f0560},
+	{0x0000b100, 0x06400641},
+	{0x0000b104, 0x0660065f},
+	{0x0000b108, 0x067e067f},
+	{0x0000b10c, 0x07410742},
+	{0x0000b110, 0x075f0740},
+	{0x0000b114, 0x077f0760},
+	{0x0000b118, 0x07800781},
+	{0x0000b11c, 0x07a0079f},
+	{0x0000b120, 0x07c107bf},
+	{0x0000b124, 0x000007c0},
+	{0x0000b128, 0x00000000},
+	{0x0000b12c, 0x00000000},
+	{0x0000b130, 0x00000000},
+	{0x0000b134, 0x00000000},
+	{0x0000b138, 0x00000000},
+	{0x0000b13c, 0x00000000},
+	{0x0000b140, 0x003f0020},
+	{0x0000b144, 0x00400041},
+	{0x0000b148, 0x0140005f},
+	{0x0000b14c, 0x0160015f},
+	{0x0000b150, 0x017e017f},
+	{0x0000b154, 0x02410242},
+	{0x0000b158, 0x025f0240},
+	{0x0000b15c, 0x027f0260},
+	{0x0000b160, 0x0341027e},
+	{0x0000b164, 0x035f0340},
+	{0x0000b168, 0x037f0360},
+	{0x0000b16c, 0x04400441},
+	{0x0000b170, 0x0460045f},
+	{0x0000b174, 0x0541047f},
+	{0x0000b178, 0x055f0540},
+	{0x0000b17c, 0x057f0560},
+	{0x0000b180, 0x06400641},
+	{0x0000b184, 0x0660065f},
+	{0x0000b188, 0x067e067f},
+	{0x0000b18c, 0x07410742},
+	{0x0000b190, 0x075f0740},
+	{0x0000b194, 0x077f0760},
+	{0x0000b198, 0x07800781},
+	{0x0000b19c, 0x07a0079f},
+	{0x0000b1a0, 0x07c107bf},
+	{0x0000b1a4, 0x000007c0},
+	{0x0000b1a8, 0x00000000},
+	{0x0000b1ac, 0x00000000},
+	{0x0000b1b0, 0x00000000},
+	{0x0000b1b4, 0x00000000},
+	{0x0000b1b8, 0x00000000},
+	{0x0000b1bc, 0x00000000},
+	{0x0000b1c0, 0x00000000},
+	{0x0000b1c4, 0x00000000},
+	{0x0000b1c8, 0x00000000},
+	{0x0000b1cc, 0x00000000},
+	{0x0000b1d0, 0x00000000},
+	{0x0000b1d4, 0x00000000},
+	{0x0000b1d8, 0x00000000},
+	{0x0000b1dc, 0x00000000},
+	{0x0000b1e0, 0x00000000},
+	{0x0000b1e4, 0x00000000},
+	{0x0000b1e8, 0x00000000},
+	{0x0000b1ec, 0x00000000},
+	{0x0000b1f0, 0x00000396},
+	{0x0000b1f4, 0x00000396},
+	{0x0000b1f8, 0x00000396},
+	{0x0000b1fc, 0x00000196},
+};
+
+static const u32 ar9462_2p0_baseband_core_mix_rxgain[][2] = {
+	/* Addr      allmodes  */
+	{0x00009fd0, 0x0a2d6b93},
+};
+
+static const u32 ar9462_2p0_baseband_postamble_mix_rxgain[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00009820, 0x206a022e, 0x206a022e, 0x206a01ae, 0x206a01ae},
+	{0x00009824, 0x63c640de, 0x5ac640d0, 0x63c640da, 0x63c640da},
+	{0x00009828, 0x0796be89, 0x0696b081, 0x0916be81, 0x0916be81},
+	{0x00009e0c, 0x6c4000e2, 0x6d4000e2, 0x6d4000d8, 0x6c4000d8},
+	{0x00009e10, 0x92c88d2e, 0x7ec88d2e, 0x7ec86d2e, 0x7ec86d2e},
+	{0x00009e14, 0x37b95d5e, 0x37b9605e, 0x3236605e, 0x32395c5e},
+};
+
 #endif /* INITVALS_9462_2P0_H */
diff --git a/drivers/net/wireless/ath/ath9k/ar9462_2p1_initvals.h b/drivers/net/wireless/ath/ath9k/ar9462_2p1_initvals.h
new file mode 100644
index 0000000..4dbc294
--- /dev/null
+++ b/drivers/net/wireless/ath/ath9k/ar9462_2p1_initvals.h
@@ -0,0 +1,1774 @@
+/*
+ * Copyright (c) 2010-2011 Atheros Communications Inc.
+ * Copyright (c) 2011-2012 Qualcomm Atheros Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#ifndef INITVALS_9462_2P1_H
+#define INITVALS_9462_2P1_H
+
+/* AR9462 2.1 */
+
+static const u32 ar9462_2p1_mac_core[][2] = {
+	/* Addr      allmodes  */
+	{0x00000008, 0x00000000},
+	{0x00000030, 0x000e0085},
+	{0x00000034, 0x00000005},
+	{0x00000040, 0x00000000},
+	{0x00000044, 0x00000000},
+	{0x00000048, 0x00000008},
+	{0x0000004c, 0x00000010},
+	{0x00000050, 0x00000000},
+	{0x00001040, 0x002ffc0f},
+	{0x00001044, 0x002ffc0f},
+	{0x00001048, 0x002ffc0f},
+	{0x0000104c, 0x002ffc0f},
+	{0x00001050, 0x002ffc0f},
+	{0x00001054, 0x002ffc0f},
+	{0x00001058, 0x002ffc0f},
+	{0x0000105c, 0x002ffc0f},
+	{0x00001060, 0x002ffc0f},
+	{0x00001064, 0x002ffc0f},
+	{0x000010f0, 0x00000100},
+	{0x00001270, 0x00000000},
+	{0x000012b0, 0x00000000},
+	{0x000012f0, 0x00000000},
+	{0x0000143c, 0x00000000},
+	{0x0000147c, 0x00000000},
+	{0x00001810, 0x0f000003},
+	{0x00008000, 0x00000000},
+	{0x00008004, 0x00000000},
+	{0x00008008, 0x00000000},
+	{0x0000800c, 0x00000000},
+	{0x00008018, 0x00000000},
+	{0x00008020, 0x00000000},
+	{0x00008038, 0x00000000},
+	{0x0000803c, 0x00080000},
+	{0x00008040, 0x00000000},
+	{0x00008044, 0x00000000},
+	{0x00008048, 0x00000000},
+	{0x0000804c, 0xffffffff},
+	{0x00008054, 0x00000000},
+	{0x00008058, 0x00000000},
+	{0x0000805c, 0x000fc78f},
+	{0x00008060, 0x0000000f},
+	{0x00008064, 0x00000000},
+	{0x00008070, 0x00000310},
+	{0x00008074, 0x00000020},
+	{0x00008078, 0x00000000},
+	{0x0000809c, 0x0000000f},
+	{0x000080a0, 0x00000000},
+	{0x000080a4, 0x02ff0000},
+	{0x000080a8, 0x0e070605},
+	{0x000080ac, 0x0000000d},
+	{0x000080b0, 0x00000000},
+	{0x000080b4, 0x00000000},
+	{0x000080b8, 0x00000000},
+	{0x000080bc, 0x00000000},
+	{0x000080c0, 0x2a800000},
+	{0x000080c4, 0x06900168},
+	{0x000080c8, 0x13881c20},
+	{0x000080cc, 0x01f40000},
+	{0x000080d0, 0x00252500},
+	{0x000080d4, 0x00b00005},
+	{0x000080d8, 0x00400002},
+	{0x000080dc, 0x00000000},
+	{0x000080e0, 0xffffffff},
+	{0x000080e4, 0x0000ffff},
+	{0x000080e8, 0x3f3f3f3f},
+	{0x000080ec, 0x00000000},
+	{0x000080f0, 0x00000000},
+	{0x000080f4, 0x00000000},
+	{0x000080fc, 0x00020000},
+	{0x00008100, 0x00000000},
+	{0x00008108, 0x00000052},
+	{0x0000810c, 0x00000000},
+	{0x00008110, 0x00000000},
+	{0x00008114, 0x000007ff},
+	{0x00008118, 0x000000aa},
+	{0x0000811c, 0x00003210},
+	{0x00008124, 0x00000000},
+	{0x00008128, 0x00000000},
+	{0x0000812c, 0x00000000},
+	{0x00008130, 0x00000000},
+	{0x00008134, 0x00000000},
+	{0x00008138, 0x00000000},
+	{0x0000813c, 0x0000ffff},
+	{0x00008144, 0xffffffff},
+	{0x00008168, 0x00000000},
+	{0x0000816c, 0x00000000},
+	{0x00008170, 0x18486e00},
+	{0x00008174, 0x33332210},
+	{0x00008178, 0x00000000},
+	{0x0000817c, 0x00020000},
+	{0x000081c4, 0x33332210},
+	{0x000081c8, 0x00000000},
+	{0x000081cc, 0x00000000},
+	{0x000081d4, 0x00000000},
+	{0x000081ec, 0x00000000},
+	{0x000081f0, 0x00000000},
+	{0x000081f4, 0x00000000},
+	{0x000081f8, 0x00000000},
+	{0x000081fc, 0x00000000},
+	{0x00008240, 0x00100000},
+	{0x00008244, 0x0010f400},
+	{0x00008248, 0x00000800},
+	{0x0000824c, 0x0001e800},
+	{0x00008250, 0x00000000},
+	{0x00008254, 0x00000000},
+	{0x00008258, 0x00000000},
+	{0x0000825c, 0x40000000},
+	{0x00008260, 0x00080922},
+	{0x00008264, 0x99c00010},
+	{0x00008268, 0xffffffff},
+	{0x0000826c, 0x0000ffff},
+	{0x00008270, 0x00000000},
+	{0x00008274, 0x40000000},
+	{0x00008278, 0x003e4180},
+	{0x0000827c, 0x00000004},
+	{0x00008284, 0x0000002c},
+	{0x00008288, 0x0000002c},
+	{0x0000828c, 0x000000ff},
+	{0x00008294, 0x00000000},
+	{0x00008298, 0x00000000},
+	{0x0000829c, 0x00000000},
+	{0x00008300, 0x00000140},
+	{0x00008314, 0x00000000},
+	{0x0000831c, 0x0000010d},
+	{0x00008328, 0x00000000},
+	{0x0000832c, 0x0000001f},
+	{0x00008330, 0x00000302},
+	{0x00008334, 0x00000700},
+	{0x00008338, 0xffff0000},
+	{0x0000833c, 0x02400000},
+	{0x00008340, 0x000107ff},
+	{0x00008344, 0xaa48107b},
+	{0x00008348, 0x008f0000},
+	{0x0000835c, 0x00000000},
+	{0x00008360, 0xffffffff},
+	{0x00008364, 0xffffffff},
+	{0x00008368, 0x00000000},
+	{0x00008370, 0x00000000},
+	{0x00008374, 0x000000ff},
+	{0x00008378, 0x00000000},
+	{0x0000837c, 0x00000000},
+	{0x00008380, 0xffffffff},
+	{0x00008384, 0xffffffff},
+	{0x00008390, 0xffffffff},
+	{0x00008394, 0xffffffff},
+	{0x00008398, 0x00000000},
+	{0x0000839c, 0x00000000},
+	{0x000083a4, 0x0000fa14},
+	{0x000083a8, 0x000f0c00},
+	{0x000083ac, 0x33332210},
+	{0x000083b0, 0x33332210},
+	{0x000083b4, 0x33332210},
+	{0x000083b8, 0x33332210},
+	{0x000083bc, 0x00000000},
+	{0x000083c0, 0x00000000},
+	{0x000083c4, 0x00000000},
+	{0x000083c8, 0x00000000},
+	{0x000083cc, 0x00000200},
+	{0x000083d0, 0x000301ff},
+};
+
+static const u32 ar9462_2p1_mac_postamble[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00001030, 0x00000230, 0x00000460, 0x000002c0, 0x00000160},
+	{0x00001070, 0x00000168, 0x000002d0, 0x00000318, 0x0000018c},
+	{0x000010b0, 0x00000e60, 0x00001cc0, 0x00007c70, 0x00003e38},
+	{0x00008014, 0x03e803e8, 0x07d007d0, 0x10801600, 0x08400b00},
+	{0x0000801c, 0x128d8027, 0x128d804f, 0x12e00057, 0x12e0002b},
+	{0x00008120, 0x08f04800, 0x08f04800, 0x08f04810, 0x08f04810},
+	{0x000081d0, 0x00003210, 0x00003210, 0x0000320a, 0x0000320a},
+	{0x00008318, 0x00003e80, 0x00007d00, 0x00006880, 0x00003440},
+};
+
+static const u32 ar9462_2p1_baseband_core[][2] = {
+	/* Addr      allmodes  */
+	{0x00009800, 0xafe68e30},
+	{0x00009804, 0xfd14e000},
+	{0x00009808, 0x9c0a9f6b},
+	{0x0000980c, 0x04900000},
+	{0x00009814, 0x9280c00a},
+	{0x00009818, 0x00000000},
+	{0x0000981c, 0x00020028},
+	{0x00009834, 0x6400a290},
+	{0x00009838, 0x0108ecff},
+	{0x0000983c, 0x0d000600},
+	{0x00009880, 0x201fff00},
+	{0x00009884, 0x00001042},
+	{0x000098a4, 0x00200400},
+	{0x000098b0, 0x32440bbe},
+	{0x000098d0, 0x004b6a8e},
+	{0x000098d4, 0x00000820},
+	{0x000098dc, 0x00000000},
+	{0x000098e4, 0x01ffffff},
+	{0x000098e8, 0x01ffffff},
+	{0x000098ec, 0x01ffffff},
+	{0x000098f0, 0x00000000},
+	{0x000098f4, 0x00000000},
+	{0x00009bf0, 0x80000000},
+	{0x00009c04, 0xff55ff55},
+	{0x00009c08, 0x0320ff55},
+	{0x00009c0c, 0x00000000},
+	{0x00009c10, 0x00000000},
+	{0x00009c14, 0x00046384},
+	{0x00009c18, 0x05b6b440},
+	{0x00009c1c, 0x00b6b440},
+	{0x00009d00, 0xc080a333},
+	{0x00009d04, 0x40206c10},
+	{0x00009d08, 0x009c4060},
+	{0x00009d0c, 0x9883800a},
+	{0x00009d10, 0x01834061},
+	{0x00009d14, 0x00c0040b},
+	{0x00009d18, 0x00000000},
+	{0x00009e08, 0x0038230c},
+	{0x00009e24, 0x990bb515},
+	{0x00009e28, 0x0c6f0000},
+	{0x00009e30, 0x06336f77},
+	{0x00009e34, 0x6af6532f},
+	{0x00009e38, 0x0cc80c00},
+	{0x00009e40, 0x15262820},
+	{0x00009e4c, 0x00001004},
+	{0x00009e50, 0x00ff03f1},
+	{0x00009e54, 0xe4c555c2},
+	{0x00009e58, 0xfd857722},
+	{0x00009e5c, 0xe9198724},
+	{0x00009fc0, 0x803e4788},
+	{0x00009fc4, 0x0001efb5},
+	{0x00009fcc, 0x40000014},
+	{0x00009fd0, 0x0a193b93},
+	{0x0000a20c, 0x00000000},
+	{0x0000a220, 0x00000000},
+	{0x0000a224, 0x00000000},
+	{0x0000a228, 0x10002310},
+	{0x0000a23c, 0x00000000},
+	{0x0000a244, 0x0c000000},
+	{0x0000a2a0, 0x00000001},
+	{0x0000a2c0, 0x00000001},
+	{0x0000a2c8, 0x00000000},
+	{0x0000a2cc, 0x18c43433},
+	{0x0000a2d4, 0x00000000},
+	{0x0000a2ec, 0x00000000},
+	{0x0000a2f0, 0x00000000},
+	{0x0000a2f4, 0x00000000},
+	{0x0000a2f8, 0x00000000},
+	{0x0000a344, 0x00000000},
+	{0x0000a34c, 0x00000000},
+	{0x0000a350, 0x0000a000},
+	{0x0000a364, 0x00000000},
+	{0x0000a370, 0x00000000},
+	{0x0000a390, 0x00000001},
+	{0x0000a394, 0x00000444},
+	{0x0000a398, 0x001f0e0f},
+	{0x0000a39c, 0x0075393f},
+	{0x0000a3a0, 0xb79f6427},
+	{0x0000a3c0, 0x20202020},
+	{0x0000a3c4, 0x22222220},
+	{0x0000a3c8, 0x20200020},
+	{0x0000a3cc, 0x20202020},
+	{0x0000a3d0, 0x20202020},
+	{0x0000a3d4, 0x20202020},
+	{0x0000a3d8, 0x20202020},
+	{0x0000a3dc, 0x20202020},
+	{0x0000a3e0, 0x20202020},
+	{0x0000a3e4, 0x20202020},
+	{0x0000a3e8, 0x20202020},
+	{0x0000a3ec, 0x20202020},
+	{0x0000a3f0, 0x00000000},
+	{0x0000a3f4, 0x00000006},
+	{0x0000a3f8, 0x0c9bd380},
+	{0x0000a3fc, 0x000f0f01},
+	{0x0000a400, 0x8fa91f01},
+	{0x0000a404, 0x00000000},
+	{0x0000a408, 0x0e79e5c6},
+	{0x0000a40c, 0x00820820},
+	{0x0000a414, 0x1ce739ce},
+	{0x0000a418, 0x2d001dce},
+	{0x0000a434, 0x00000000},
+	{0x0000a438, 0x00001801},
+	{0x0000a43c, 0x00100000},
+	{0x0000a444, 0x00000000},
+	{0x0000a448, 0x05000080},
+	{0x0000a44c, 0x00000001},
+	{0x0000a450, 0x00010000},
+	{0x0000a454, 0x07000000},
+	{0x0000a644, 0xbfad9d74},
+	{0x0000a648, 0x0048060a},
+	{0x0000a64c, 0x00002037},
+	{0x0000a670, 0x03020100},
+	{0x0000a674, 0x09080504},
+	{0x0000a678, 0x0d0c0b0a},
+	{0x0000a67c, 0x13121110},
+	{0x0000a680, 0x31301514},
+	{0x0000a684, 0x35343332},
+	{0x0000a688, 0x00000036},
+	{0x0000a690, 0x00000838},
+	{0x0000a6b0, 0x0000000a},
+	{0x0000a6b4, 0x00512c01},
+	{0x0000a7c0, 0x00000000},
+	{0x0000a7c4, 0xfffffffc},
+	{0x0000a7c8, 0x00000000},
+	{0x0000a7cc, 0x00000000},
+	{0x0000a7d0, 0x00000000},
+	{0x0000a7d4, 0x00000004},
+	{0x0000a7dc, 0x00000000},
+	{0x0000a7f0, 0x80000000},
+	{0x0000a8d0, 0x004b6a8e},
+	{0x0000a8d4, 0x00000820},
+	{0x0000a8dc, 0x00000000},
+	{0x0000a8f0, 0x00000000},
+	{0x0000a8f4, 0x00000000},
+	{0x0000abf0, 0x80000000},
+	{0x0000b2d0, 0x00000080},
+	{0x0000b2d4, 0x00000000},
+	{0x0000b2ec, 0x00000000},
+	{0x0000b2f0, 0x00000000},
+	{0x0000b2f4, 0x00000000},
+	{0x0000b2f8, 0x00000000},
+	{0x0000b408, 0x0e79e5c0},
+	{0x0000b40c, 0x00820820},
+	{0x0000b420, 0x00000000},
+	{0x0000b6b0, 0x0000000a},
+	{0x0000b6b4, 0x00000001},
+};
+
+static const u32 ar9462_2p1_baseband_postamble[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00009810, 0xd00a8005, 0xd00a8005, 0xd00a8011, 0xd00a800d},
+	{0x00009820, 0x206a022e, 0x206a022e, 0x206a012e, 0x206a01ae},
+	{0x00009824, 0x63c640de, 0x5ac640d0, 0x5ac640d0, 0x63c640da},
+	{0x00009828, 0x0796be89, 0x0696b081, 0x0696b881, 0x09143e81},
+	{0x0000982c, 0x05eea6d4, 0x05eea6d4, 0x05eea6d4, 0x05eea6d4},
+	{0x00009830, 0x0000059c, 0x0000059c, 0x0000119c, 0x0000119c},
+	{0x00009c00, 0x000000c4, 0x000000c4, 0x000000c4, 0x000000c4},
+	{0x00009e00, 0x0372111a, 0x0372111a, 0x037216a0, 0x037216a2},
+	{0x00009e04, 0x001c2020, 0x001c2020, 0x001c2020, 0x001c2020},
+	{0x00009e0c, 0x6c4000e2, 0x6d4000e2, 0x6d4000e2, 0x6c4000d8},
+	{0x00009e10, 0x92c88d2e, 0x7ec88d2e, 0x7ec84d2e, 0x7ec86d2e},
+	{0x00009e14, 0x37b95d5e, 0x37b9605e, 0x3236605e, 0x32365a5e},
+	{0x00009e18, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x00009e1c, 0x0001cf9c, 0x0001cf9c, 0x00021f9c, 0x00021f9c},
+	{0x00009e20, 0x000003b5, 0x000003b5, 0x000003ce, 0x000003ce},
+	{0x00009e2c, 0x0000001c, 0x0000001c, 0x00000021, 0x00000021},
+	{0x00009e3c, 0xcf946220, 0xcf946220, 0xcfd5c782, 0xcfd5c282},
+	{0x00009e44, 0x62321e27, 0x62321e27, 0xfe291e27, 0xfe291e27},
+	{0x00009e48, 0x5030201a, 0x5030201a, 0x50302012, 0x50302012},
+	{0x00009fc8, 0x0003f000, 0x0003f000, 0x0001a000, 0x0001a000},
+	{0x0000a204, 0x01318fc0, 0x01318fc4, 0x01318fc4, 0x01318fc0},
+	{0x0000a208, 0x00000104, 0x00000104, 0x00000004, 0x00000004},
+	{0x0000a22c, 0x01026a2f, 0x01026a27, 0x01026a2f, 0x01026a2f},
+	{0x0000a230, 0x0000400a, 0x00004014, 0x00004016, 0x0000400b},
+	{0x0000a234, 0x00000fff, 0x10000fff, 0x10000fff, 0x00000fff},
+	{0x0000a238, 0xffb81018, 0xffb81018, 0xffb81018, 0xffb81018},
+	{0x0000a250, 0x00000000, 0x00000000, 0x00000210, 0x00000108},
+	{0x0000a254, 0x000007d0, 0x00000fa0, 0x00001130, 0x00000898},
+	{0x0000a258, 0x02020002, 0x02020002, 0x02020002, 0x02020002},
+	{0x0000a25c, 0x01000e0e, 0x01000e0e, 0x01000e0e, 0x01000e0e},
+	{0x0000a260, 0x0a021501, 0x0a021501, 0x3a021501, 0x3a021501},
+	{0x0000a264, 0x00000e0e, 0x00000e0e, 0x00000e0e, 0x00000e0e},
+	{0x0000a280, 0x00000007, 0x00000007, 0x0000000b, 0x0000000b},
+	{0x0000a284, 0x00000000, 0x00000000, 0x00000150, 0x00000150},
+	{0x0000a288, 0x00000110, 0x00000110, 0x00000110, 0x00000110},
+	{0x0000a28c, 0x00022222, 0x00022222, 0x00022222, 0x00022222},
+	{0x0000a2c4, 0x00158d18, 0x00158d18, 0x00158d18, 0x00158d18},
+	{0x0000a2d0, 0x00041981, 0x00041981, 0x00041981, 0x00041982},
+	{0x0000a2d8, 0x7999a83b, 0x7999a83b, 0x7999a83b, 0x7999a83b},
+	{0x0000a358, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a3a4, 0x00000050, 0x00000050, 0x00000000, 0x00000000},
+	{0x0000a3a8, 0xaaaaaaaa, 0xaaaaaaaa, 0xaaaaaaaa, 0xaaaaaaaa},
+	{0x0000a3ac, 0xaaaaaa00, 0xaa30aa30, 0xaaaaaa00, 0xaaaaaa00},
+	{0x0000a41c, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce},
+	{0x0000a420, 0x000001ce, 0x000001ce, 0x000001ce, 0x000001ce},
+	{0x0000a424, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce},
+	{0x0000a428, 0x000001ce, 0x000001ce, 0x000001ce, 0x000001ce},
+	{0x0000a42c, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce},
+	{0x0000a430, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce, 0x1ce739ce},
+	{0x0000a830, 0x0000019c, 0x0000019c, 0x0000019c, 0x0000019c},
+	{0x0000ae04, 0x001c0000, 0x001c0000, 0x001c0000, 0x00100000},
+	{0x0000ae18, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000ae1c, 0x0000019c, 0x0000019c, 0x0000019c, 0x0000019c},
+	{0x0000ae20, 0x000001b5, 0x000001b5, 0x000001ce, 0x000001ce},
+	{0x0000b284, 0x00000000, 0x00000000, 0x00000550, 0x00000550},
+};
+
+static const u32 ar9462_2p1_radio_core[][2] = {
+	/* Addr      allmodes  */
+	{0x00016000, 0x36db6db6},
+	{0x00016004, 0x6db6db40},
+	{0x00016008, 0x73f00000},
+	{0x0001600c, 0x00000000},
+	{0x00016010, 0x6d820001},
+	{0x00016040, 0x7f80fff8},
+	{0x0001604c, 0x2699e04f},
+	{0x00016050, 0x6db6db6c},
+	{0x00016058, 0x6c200000},
+	{0x00016080, 0x000c0000},
+	{0x00016084, 0x9a68048c},
+	{0x00016088, 0x54214514},
+	{0x0001608c, 0x1203040b},
+	{0x00016090, 0x24926490},
+	{0x00016098, 0xd2888888},
+	{0x000160a0, 0x0a108ffe},
+	{0x000160a4, 0x812fc491},
+	{0x000160a8, 0x423c8000},
+	{0x000160b4, 0x92000000},
+	{0x000160b8, 0x0285dddc},
+	{0x000160bc, 0x02908888},
+	{0x000160c0, 0x00adb6d0},
+	{0x000160c4, 0x6db6db60},
+	{0x000160c8, 0x6db6db6c},
+	{0x000160cc, 0x0de6c1b0},
+	{0x00016100, 0x3fffbe04},
+	{0x00016104, 0xfff80000},
+	{0x00016108, 0x00200400},
+	{0x00016110, 0x00000000},
+	{0x00016144, 0x02084080},
+	{0x00016148, 0x000080c0},
+	{0x00016280, 0x050a0001},
+	{0x00016284, 0x3d841418},
+	{0x00016288, 0x00000000},
+	{0x0001628c, 0xe3000000},
+	{0x00016290, 0xa1005080},
+	{0x00016294, 0x00000020},
+	{0x00016298, 0x54a82900},
+	{0x00016340, 0x121e4276},
+	{0x00016344, 0x00300000},
+	{0x00016400, 0x36db6db6},
+	{0x00016404, 0x6db6db40},
+	{0x00016408, 0x73f00000},
+	{0x0001640c, 0x00000000},
+	{0x00016410, 0x6c800001},
+	{0x00016440, 0x7f80fff8},
+	{0x0001644c, 0x4699e04f},
+	{0x00016450, 0x6db6db6c},
+	{0x00016500, 0x3fffbe04},
+	{0x00016504, 0xfff80000},
+	{0x00016508, 0x00200400},
+	{0x00016510, 0x00000000},
+	{0x00016544, 0x02084080},
+	{0x00016548, 0x000080c0},
+};
+
+static const u32 ar9462_2p1_radio_postamble[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x0001609c, 0x0b8ee524, 0x0b8ee524, 0x0b8ee524, 0x0b8ee524},
+	{0x000160b0, 0x01d67f70, 0x01d67f70, 0x01d67f70, 0x01d67f70},
+	{0x0001610c, 0x48000000, 0x40000000, 0x40000000, 0x40000000},
+	{0x0001650c, 0x48000000, 0x40000000, 0x40000000, 0x40000000},
+};
+
+static const u32 ar9462_2p1_soc_preamble[][2] = {
+	/* Addr      allmodes  */
+	{0x000040a4, 0x00a0c1c9},
+	{0x00007020, 0x00000000},
+	{0x00007034, 0x00000002},
+	{0x00007038, 0x000004c2},
+};
+
+static const u32 ar9462_2p1_soc_postamble[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00007010, 0x00000033, 0x00000033, 0x00000033, 0x00000033},
+};
+
+static const u32 ar9462_2p1_radio_postamble_sys2ant[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x000160ac, 0xa4646c08, 0xa4646c08, 0x24645808, 0x24645808},
+	{0x00016140, 0x10804008, 0x10804008, 0x50804008, 0x50804008},
+	{0x00016540, 0x10804008, 0x10804008, 0x50804008, 0x50804008},
+};
+
+static const u32 ar9462_2p1_common_rx_gain[][2] = {
+	/* Addr      allmodes  */
+	{0x0000a000, 0x00010000},
+	{0x0000a004, 0x00030002},
+	{0x0000a008, 0x00050004},
+	{0x0000a00c, 0x00810080},
+	{0x0000a010, 0x00830082},
+	{0x0000a014, 0x01810180},
+	{0x0000a018, 0x01830182},
+	{0x0000a01c, 0x01850184},
+	{0x0000a020, 0x01890188},
+	{0x0000a024, 0x018b018a},
+	{0x0000a028, 0x018d018c},
+	{0x0000a02c, 0x01910190},
+	{0x0000a030, 0x01930192},
+	{0x0000a034, 0x01950194},
+	{0x0000a038, 0x038a0196},
+	{0x0000a03c, 0x038c038b},
+	{0x0000a040, 0x0390038d},
+	{0x0000a044, 0x03920391},
+	{0x0000a048, 0x03940393},
+	{0x0000a04c, 0x03960395},
+	{0x0000a050, 0x00000000},
+	{0x0000a054, 0x00000000},
+	{0x0000a058, 0x00000000},
+	{0x0000a05c, 0x00000000},
+	{0x0000a060, 0x00000000},
+	{0x0000a064, 0x00000000},
+	{0x0000a068, 0x00000000},
+	{0x0000a06c, 0x00000000},
+	{0x0000a070, 0x00000000},
+	{0x0000a074, 0x00000000},
+	{0x0000a078, 0x00000000},
+	{0x0000a07c, 0x00000000},
+	{0x0000a080, 0x22222229},
+	{0x0000a084, 0x1d1d1d1d},
+	{0x0000a088, 0x1d1d1d1d},
+	{0x0000a08c, 0x1d1d1d1d},
+	{0x0000a090, 0x171d1d1d},
+	{0x0000a094, 0x11111717},
+	{0x0000a098, 0x00030311},
+	{0x0000a09c, 0x00000000},
+	{0x0000a0a0, 0x00000000},
+	{0x0000a0a4, 0x00000000},
+	{0x0000a0a8, 0x00000000},
+	{0x0000a0ac, 0x00000000},
+	{0x0000a0b0, 0x00000000},
+	{0x0000a0b4, 0x00000000},
+	{0x0000a0b8, 0x00000000},
+	{0x0000a0bc, 0x00000000},
+	{0x0000a0c0, 0x001f0000},
+	{0x0000a0c4, 0x01000101},
+	{0x0000a0c8, 0x011e011f},
+	{0x0000a0cc, 0x011c011d},
+	{0x0000a0d0, 0x02030204},
+	{0x0000a0d4, 0x02010202},
+	{0x0000a0d8, 0x021f0200},
+	{0x0000a0dc, 0x0302021e},
+	{0x0000a0e0, 0x03000301},
+	{0x0000a0e4, 0x031e031f},
+	{0x0000a0e8, 0x0402031d},
+	{0x0000a0ec, 0x04000401},
+	{0x0000a0f0, 0x041e041f},
+	{0x0000a0f4, 0x0502041d},
+	{0x0000a0f8, 0x05000501},
+	{0x0000a0fc, 0x051e051f},
+	{0x0000a100, 0x06010602},
+	{0x0000a104, 0x061f0600},
+	{0x0000a108, 0x061d061e},
+	{0x0000a10c, 0x07020703},
+	{0x0000a110, 0x07000701},
+	{0x0000a114, 0x00000000},
+	{0x0000a118, 0x00000000},
+	{0x0000a11c, 0x00000000},
+	{0x0000a120, 0x00000000},
+	{0x0000a124, 0x00000000},
+	{0x0000a128, 0x00000000},
+	{0x0000a12c, 0x00000000},
+	{0x0000a130, 0x00000000},
+	{0x0000a134, 0x00000000},
+	{0x0000a138, 0x00000000},
+	{0x0000a13c, 0x00000000},
+	{0x0000a140, 0x001f0000},
+	{0x0000a144, 0x01000101},
+	{0x0000a148, 0x011e011f},
+	{0x0000a14c, 0x011c011d},
+	{0x0000a150, 0x02030204},
+	{0x0000a154, 0x02010202},
+	{0x0000a158, 0x021f0200},
+	{0x0000a15c, 0x0302021e},
+	{0x0000a160, 0x03000301},
+	{0x0000a164, 0x031e031f},
+	{0x0000a168, 0x0402031d},
+	{0x0000a16c, 0x04000401},
+	{0x0000a170, 0x041e041f},
+	{0x0000a174, 0x0502041d},
+	{0x0000a178, 0x05000501},
+	{0x0000a17c, 0x051e051f},
+	{0x0000a180, 0x06010602},
+	{0x0000a184, 0x061f0600},
+	{0x0000a188, 0x061d061e},
+	{0x0000a18c, 0x07020703},
+	{0x0000a190, 0x07000701},
+	{0x0000a194, 0x00000000},
+	{0x0000a198, 0x00000000},
+	{0x0000a19c, 0x00000000},
+	{0x0000a1a0, 0x00000000},
+	{0x0000a1a4, 0x00000000},
+	{0x0000a1a8, 0x00000000},
+	{0x0000a1ac, 0x00000000},
+	{0x0000a1b0, 0x00000000},
+	{0x0000a1b4, 0x00000000},
+	{0x0000a1b8, 0x00000000},
+	{0x0000a1bc, 0x00000000},
+	{0x0000a1c0, 0x00000000},
+	{0x0000a1c4, 0x00000000},
+	{0x0000a1c8, 0x00000000},
+	{0x0000a1cc, 0x00000000},
+	{0x0000a1d0, 0x00000000},
+	{0x0000a1d4, 0x00000000},
+	{0x0000a1d8, 0x00000000},
+	{0x0000a1dc, 0x00000000},
+	{0x0000a1e0, 0x00000000},
+	{0x0000a1e4, 0x00000000},
+	{0x0000a1e8, 0x00000000},
+	{0x0000a1ec, 0x00000000},
+	{0x0000a1f0, 0x00000396},
+	{0x0000a1f4, 0x00000396},
+	{0x0000a1f8, 0x00000396},
+	{0x0000a1fc, 0x00000196},
+	{0x0000b000, 0x00010000},
+	{0x0000b004, 0x00030002},
+	{0x0000b008, 0x00050004},
+	{0x0000b00c, 0x00810080},
+	{0x0000b010, 0x00830082},
+	{0x0000b014, 0x01810180},
+	{0x0000b018, 0x01830182},
+	{0x0000b01c, 0x01850184},
+	{0x0000b020, 0x02810280},
+	{0x0000b024, 0x02830282},
+	{0x0000b028, 0x02850284},
+	{0x0000b02c, 0x02890288},
+	{0x0000b030, 0x028b028a},
+	{0x0000b034, 0x0388028c},
+	{0x0000b038, 0x038a0389},
+	{0x0000b03c, 0x038c038b},
+	{0x0000b040, 0x0390038d},
+	{0x0000b044, 0x03920391},
+	{0x0000b048, 0x03940393},
+	{0x0000b04c, 0x03960395},
+	{0x0000b050, 0x00000000},
+	{0x0000b054, 0x00000000},
+	{0x0000b058, 0x00000000},
+	{0x0000b05c, 0x00000000},
+	{0x0000b060, 0x00000000},
+	{0x0000b064, 0x00000000},
+	{0x0000b068, 0x00000000},
+	{0x0000b06c, 0x00000000},
+	{0x0000b070, 0x00000000},
+	{0x0000b074, 0x00000000},
+	{0x0000b078, 0x00000000},
+	{0x0000b07c, 0x00000000},
+	{0x0000b080, 0x2a2d2f32},
+	{0x0000b084, 0x21232328},
+	{0x0000b088, 0x19191c1e},
+	{0x0000b08c, 0x12141417},
+	{0x0000b090, 0x07070e0e},
+	{0x0000b094, 0x03030305},
+	{0x0000b098, 0x00000003},
+	{0x0000b09c, 0x00000000},
+	{0x0000b0a0, 0x00000000},
+	{0x0000b0a4, 0x00000000},
+	{0x0000b0a8, 0x00000000},
+	{0x0000b0ac, 0x00000000},
+	{0x0000b0b0, 0x00000000},
+	{0x0000b0b4, 0x00000000},
+	{0x0000b0b8, 0x00000000},
+	{0x0000b0bc, 0x00000000},
+	{0x0000b0c0, 0x003f0020},
+	{0x0000b0c4, 0x00400041},
+	{0x0000b0c8, 0x0140005f},
+	{0x0000b0cc, 0x0160015f},
+	{0x0000b0d0, 0x017e017f},
+	{0x0000b0d4, 0x02410242},
+	{0x0000b0d8, 0x025f0240},
+	{0x0000b0dc, 0x027f0260},
+	{0x0000b0e0, 0x0341027e},
+	{0x0000b0e4, 0x035f0340},
+	{0x0000b0e8, 0x037f0360},
+	{0x0000b0ec, 0x04400441},
+	{0x0000b0f0, 0x0460045f},
+	{0x0000b0f4, 0x0541047f},
+	{0x0000b0f8, 0x055f0540},
+	{0x0000b0fc, 0x057f0560},
+	{0x0000b100, 0x06400641},
+	{0x0000b104, 0x0660065f},
+	{0x0000b108, 0x067e067f},
+	{0x0000b10c, 0x07410742},
+	{0x0000b110, 0x075f0740},
+	{0x0000b114, 0x077f0760},
+	{0x0000b118, 0x07800781},
+	{0x0000b11c, 0x07a0079f},
+	{0x0000b120, 0x07c107bf},
+	{0x0000b124, 0x000007c0},
+	{0x0000b128, 0x00000000},
+	{0x0000b12c, 0x00000000},
+	{0x0000b130, 0x00000000},
+	{0x0000b134, 0x00000000},
+	{0x0000b138, 0x00000000},
+	{0x0000b13c, 0x00000000},
+	{0x0000b140, 0x003f0020},
+	{0x0000b144, 0x00400041},
+	{0x0000b148, 0x0140005f},
+	{0x0000b14c, 0x0160015f},
+	{0x0000b150, 0x017e017f},
+	{0x0000b154, 0x02410242},
+	{0x0000b158, 0x025f0240},
+	{0x0000b15c, 0x027f0260},
+	{0x0000b160, 0x0341027e},
+	{0x0000b164, 0x035f0340},
+	{0x0000b168, 0x037f0360},
+	{0x0000b16c, 0x04400441},
+	{0x0000b170, 0x0460045f},
+	{0x0000b174, 0x0541047f},
+	{0x0000b178, 0x055f0540},
+	{0x0000b17c, 0x057f0560},
+	{0x0000b180, 0x06400641},
+	{0x0000b184, 0x0660065f},
+	{0x0000b188, 0x067e067f},
+	{0x0000b18c, 0x07410742},
+	{0x0000b190, 0x075f0740},
+	{0x0000b194, 0x077f0760},
+	{0x0000b198, 0x07800781},
+	{0x0000b19c, 0x07a0079f},
+	{0x0000b1a0, 0x07c107bf},
+	{0x0000b1a4, 0x000007c0},
+	{0x0000b1a8, 0x00000000},
+	{0x0000b1ac, 0x00000000},
+	{0x0000b1b0, 0x00000000},
+	{0x0000b1b4, 0x00000000},
+	{0x0000b1b8, 0x00000000},
+	{0x0000b1bc, 0x00000000},
+	{0x0000b1c0, 0x00000000},
+	{0x0000b1c4, 0x00000000},
+	{0x0000b1c8, 0x00000000},
+	{0x0000b1cc, 0x00000000},
+	{0x0000b1d0, 0x00000000},
+	{0x0000b1d4, 0x00000000},
+	{0x0000b1d8, 0x00000000},
+	{0x0000b1dc, 0x00000000},
+	{0x0000b1e0, 0x00000000},
+	{0x0000b1e4, 0x00000000},
+	{0x0000b1e8, 0x00000000},
+	{0x0000b1ec, 0x00000000},
+	{0x0000b1f0, 0x00000396},
+	{0x0000b1f4, 0x00000396},
+	{0x0000b1f8, 0x00000396},
+	{0x0000b1fc, 0x00000196},
+};
+
+static const u32 ar9462_2p1_common_mixed_rx_gain[][2] = {
+	/* Addr      allmodes  */
+	{0x0000a000, 0x00010000},
+	{0x0000a004, 0x00030002},
+	{0x0000a008, 0x00050004},
+	{0x0000a00c, 0x00810080},
+	{0x0000a010, 0x00830082},
+	{0x0000a014, 0x01810180},
+	{0x0000a018, 0x01830182},
+	{0x0000a01c, 0x01850184},
+	{0x0000a020, 0x01890188},
+	{0x0000a024, 0x018b018a},
+	{0x0000a028, 0x018d018c},
+	{0x0000a02c, 0x03820190},
+	{0x0000a030, 0x03840383},
+	{0x0000a034, 0x03880385},
+	{0x0000a038, 0x038a0389},
+	{0x0000a03c, 0x038c038b},
+	{0x0000a040, 0x0390038d},
+	{0x0000a044, 0x03920391},
+	{0x0000a048, 0x03940393},
+	{0x0000a04c, 0x03960395},
+	{0x0000a050, 0x00000000},
+	{0x0000a054, 0x00000000},
+	{0x0000a058, 0x00000000},
+	{0x0000a05c, 0x00000000},
+	{0x0000a060, 0x00000000},
+	{0x0000a064, 0x00000000},
+	{0x0000a068, 0x00000000},
+	{0x0000a06c, 0x00000000},
+	{0x0000a070, 0x00000000},
+	{0x0000a074, 0x00000000},
+	{0x0000a078, 0x00000000},
+	{0x0000a07c, 0x00000000},
+	{0x0000a080, 0x29292929},
+	{0x0000a084, 0x29292929},
+	{0x0000a088, 0x29292929},
+	{0x0000a08c, 0x29292929},
+	{0x0000a090, 0x22292929},
+	{0x0000a094, 0x1d1d2222},
+	{0x0000a098, 0x0c111117},
+	{0x0000a09c, 0x00030303},
+	{0x0000a0a0, 0x00000000},
+	{0x0000a0a4, 0x00000000},
+	{0x0000a0a8, 0x00000000},
+	{0x0000a0ac, 0x00000000},
+	{0x0000a0b0, 0x00000000},
+	{0x0000a0b4, 0x00000000},
+	{0x0000a0b8, 0x00000000},
+	{0x0000a0bc, 0x00000000},
+	{0x0000a0c0, 0x001f0000},
+	{0x0000a0c4, 0x01000101},
+	{0x0000a0c8, 0x011e011f},
+	{0x0000a0cc, 0x011c011d},
+	{0x0000a0d0, 0x02030204},
+	{0x0000a0d4, 0x02010202},
+	{0x0000a0d8, 0x021f0200},
+	{0x0000a0dc, 0x0302021e},
+	{0x0000a0e0, 0x03000301},
+	{0x0000a0e4, 0x031e031f},
+	{0x0000a0e8, 0x0402031d},
+	{0x0000a0ec, 0x04000401},
+	{0x0000a0f0, 0x041e041f},
+	{0x0000a0f4, 0x0502041d},
+	{0x0000a0f8, 0x05000501},
+	{0x0000a0fc, 0x051e051f},
+	{0x0000a100, 0x06010602},
+	{0x0000a104, 0x061f0600},
+	{0x0000a108, 0x061d061e},
+	{0x0000a10c, 0x07020703},
+	{0x0000a110, 0x07000701},
+	{0x0000a114, 0x00000000},
+	{0x0000a118, 0x00000000},
+	{0x0000a11c, 0x00000000},
+	{0x0000a120, 0x00000000},
+	{0x0000a124, 0x00000000},
+	{0x0000a128, 0x00000000},
+	{0x0000a12c, 0x00000000},
+	{0x0000a130, 0x00000000},
+	{0x0000a134, 0x00000000},
+	{0x0000a138, 0x00000000},
+	{0x0000a13c, 0x00000000},
+	{0x0000a140, 0x001f0000},
+	{0x0000a144, 0x01000101},
+	{0x0000a148, 0x011e011f},
+	{0x0000a14c, 0x011c011d},
+	{0x0000a150, 0x02030204},
+	{0x0000a154, 0x02010202},
+	{0x0000a158, 0x021f0200},
+	{0x0000a15c, 0x0302021e},
+	{0x0000a160, 0x03000301},
+	{0x0000a164, 0x031e031f},
+	{0x0000a168, 0x0402031d},
+	{0x0000a16c, 0x04000401},
+	{0x0000a170, 0x041e041f},
+	{0x0000a174, 0x0502041d},
+	{0x0000a178, 0x05000501},
+	{0x0000a17c, 0x051e051f},
+	{0x0000a180, 0x06010602},
+	{0x0000a184, 0x061f0600},
+	{0x0000a188, 0x061d061e},
+	{0x0000a18c, 0x07020703},
+	{0x0000a190, 0x07000701},
+	{0x0000a194, 0x00000000},
+	{0x0000a198, 0x00000000},
+	{0x0000a19c, 0x00000000},
+	{0x0000a1a0, 0x00000000},
+	{0x0000a1a4, 0x00000000},
+	{0x0000a1a8, 0x00000000},
+	{0x0000a1ac, 0x00000000},
+	{0x0000a1b0, 0x00000000},
+	{0x0000a1b4, 0x00000000},
+	{0x0000a1b8, 0x00000000},
+	{0x0000a1bc, 0x00000000},
+	{0x0000a1c0, 0x00000000},
+	{0x0000a1c4, 0x00000000},
+	{0x0000a1c8, 0x00000000},
+	{0x0000a1cc, 0x00000000},
+	{0x0000a1d0, 0x00000000},
+	{0x0000a1d4, 0x00000000},
+	{0x0000a1d8, 0x00000000},
+	{0x0000a1dc, 0x00000000},
+	{0x0000a1e0, 0x00000000},
+	{0x0000a1e4, 0x00000000},
+	{0x0000a1e8, 0x00000000},
+	{0x0000a1ec, 0x00000000},
+	{0x0000a1f0, 0x00000396},
+	{0x0000a1f4, 0x00000396},
+	{0x0000a1f8, 0x00000396},
+	{0x0000a1fc, 0x00000196},
+	{0x0000b000, 0x00010000},
+	{0x0000b004, 0x00030002},
+	{0x0000b008, 0x00050004},
+	{0x0000b00c, 0x00810080},
+	{0x0000b010, 0x00830082},
+	{0x0000b014, 0x01810180},
+	{0x0000b018, 0x01830182},
+	{0x0000b01c, 0x01850184},
+	{0x0000b020, 0x02810280},
+	{0x0000b024, 0x02830282},
+	{0x0000b028, 0x02850284},
+	{0x0000b02c, 0x02890288},
+	{0x0000b030, 0x028b028a},
+	{0x0000b034, 0x0388028c},
+	{0x0000b038, 0x038a0389},
+	{0x0000b03c, 0x038c038b},
+	{0x0000b040, 0x0390038d},
+	{0x0000b044, 0x03920391},
+	{0x0000b048, 0x03940393},
+	{0x0000b04c, 0x03960395},
+	{0x0000b050, 0x00000000},
+	{0x0000b054, 0x00000000},
+	{0x0000b058, 0x00000000},
+	{0x0000b05c, 0x00000000},
+	{0x0000b060, 0x00000000},
+	{0x0000b064, 0x00000000},
+	{0x0000b068, 0x00000000},
+	{0x0000b06c, 0x00000000},
+	{0x0000b070, 0x00000000},
+	{0x0000b074, 0x00000000},
+	{0x0000b078, 0x00000000},
+	{0x0000b07c, 0x00000000},
+	{0x0000b080, 0x2a2d2f32},
+	{0x0000b084, 0x21232328},
+	{0x0000b088, 0x19191c1e},
+	{0x0000b08c, 0x12141417},
+	{0x0000b090, 0x07070e0e},
+	{0x0000b094, 0x03030305},
+	{0x0000b098, 0x00000003},
+	{0x0000b09c, 0x00000000},
+	{0x0000b0a0, 0x00000000},
+	{0x0000b0a4, 0x00000000},
+	{0x0000b0a8, 0x00000000},
+	{0x0000b0ac, 0x00000000},
+	{0x0000b0b0, 0x00000000},
+	{0x0000b0b4, 0x00000000},
+	{0x0000b0b8, 0x00000000},
+	{0x0000b0bc, 0x00000000},
+	{0x0000b0c0, 0x003f0020},
+	{0x0000b0c4, 0x00400041},
+	{0x0000b0c8, 0x0140005f},
+	{0x0000b0cc, 0x0160015f},
+	{0x0000b0d0, 0x017e017f},
+	{0x0000b0d4, 0x02410242},
+	{0x0000b0d8, 0x025f0240},
+	{0x0000b0dc, 0x027f0260},
+	{0x0000b0e0, 0x0341027e},
+	{0x0000b0e4, 0x035f0340},
+	{0x0000b0e8, 0x037f0360},
+	{0x0000b0ec, 0x04400441},
+	{0x0000b0f0, 0x0460045f},
+	{0x0000b0f4, 0x0541047f},
+	{0x0000b0f8, 0x055f0540},
+	{0x0000b0fc, 0x057f0560},
+	{0x0000b100, 0x06400641},
+	{0x0000b104, 0x0660065f},
+	{0x0000b108, 0x067e067f},
+	{0x0000b10c, 0x07410742},
+	{0x0000b110, 0x075f0740},
+	{0x0000b114, 0x077f0760},
+	{0x0000b118, 0x07800781},
+	{0x0000b11c, 0x07a0079f},
+	{0x0000b120, 0x07c107bf},
+	{0x0000b124, 0x000007c0},
+	{0x0000b128, 0x00000000},
+	{0x0000b12c, 0x00000000},
+	{0x0000b130, 0x00000000},
+	{0x0000b134, 0x00000000},
+	{0x0000b138, 0x00000000},
+	{0x0000b13c, 0x00000000},
+	{0x0000b140, 0x003f0020},
+	{0x0000b144, 0x00400041},
+	{0x0000b148, 0x0140005f},
+	{0x0000b14c, 0x0160015f},
+	{0x0000b150, 0x017e017f},
+	{0x0000b154, 0x02410242},
+	{0x0000b158, 0x025f0240},
+	{0x0000b15c, 0x027f0260},
+	{0x0000b160, 0x0341027e},
+	{0x0000b164, 0x035f0340},
+	{0x0000b168, 0x037f0360},
+	{0x0000b16c, 0x04400441},
+	{0x0000b170, 0x0460045f},
+	{0x0000b174, 0x0541047f},
+	{0x0000b178, 0x055f0540},
+	{0x0000b17c, 0x057f0560},
+	{0x0000b180, 0x06400641},
+	{0x0000b184, 0x0660065f},
+	{0x0000b188, 0x067e067f},
+	{0x0000b18c, 0x07410742},
+	{0x0000b190, 0x075f0740},
+	{0x0000b194, 0x077f0760},
+	{0x0000b198, 0x07800781},
+	{0x0000b19c, 0x07a0079f},
+	{0x0000b1a0, 0x07c107bf},
+	{0x0000b1a4, 0x000007c0},
+	{0x0000b1a8, 0x00000000},
+	{0x0000b1ac, 0x00000000},
+	{0x0000b1b0, 0x00000000},
+	{0x0000b1b4, 0x00000000},
+	{0x0000b1b8, 0x00000000},
+	{0x0000b1bc, 0x00000000},
+	{0x0000b1c0, 0x00000000},
+	{0x0000b1c4, 0x00000000},
+	{0x0000b1c8, 0x00000000},
+	{0x0000b1cc, 0x00000000},
+	{0x0000b1d0, 0x00000000},
+	{0x0000b1d4, 0x00000000},
+	{0x0000b1d8, 0x00000000},
+	{0x0000b1dc, 0x00000000},
+	{0x0000b1e0, 0x00000000},
+	{0x0000b1e4, 0x00000000},
+	{0x0000b1e8, 0x00000000},
+	{0x0000b1ec, 0x00000000},
+	{0x0000b1f0, 0x00000396},
+	{0x0000b1f4, 0x00000396},
+	{0x0000b1f8, 0x00000396},
+	{0x0000b1fc, 0x00000196},
+};
+
+static const u32 ar9462_2p1_baseband_core_mix_rxgain[][2] = {
+	/* Addr      allmodes  */
+	{0x00009fd0, 0x0a2d6b93},
+};
+
+static const u32 ar9462_2p1_baseband_postamble_mix_rxgain[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00009820, 0x206a022e, 0x206a022e, 0x206a01ae, 0x206a01ae},
+	{0x00009824, 0x63c640de, 0x5ac640d0, 0x63c640da, 0x63c640da},
+	{0x00009828, 0x0796be89, 0x0696b081, 0x0916be81, 0x0916be81},
+	{0x00009e0c, 0x6c4000e2, 0x6d4000e2, 0x6d4000d8, 0x6c4000d8},
+	{0x00009e10, 0x92c88d2e, 0x7ec88d2e, 0x7ec86d2e, 0x7ec86d2e},
+	{0x00009e14, 0x37b95d5e, 0x37b9605e, 0x3236605e, 0x32395c5e},
+};
+
+static const u32 ar9462_2p1_baseband_postamble_5g_xlna[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x00009e3c, 0xcf946220, 0xcf946220, 0xcfd5c782, 0xcfd5c282},
+};
+
+static const u32 ar9462_2p1_common_wo_xlna_rx_gain[][2] = {
+	/* Addr      allmodes  */
+	{0x0000a000, 0x00010000},
+	{0x0000a004, 0x00030002},
+	{0x0000a008, 0x00050004},
+	{0x0000a00c, 0x00810080},
+	{0x0000a010, 0x00830082},
+	{0x0000a014, 0x01810180},
+	{0x0000a018, 0x01830182},
+	{0x0000a01c, 0x01850184},
+	{0x0000a020, 0x01890188},
+	{0x0000a024, 0x018b018a},
+	{0x0000a028, 0x018d018c},
+	{0x0000a02c, 0x03820190},
+	{0x0000a030, 0x03840383},
+	{0x0000a034, 0x03880385},
+	{0x0000a038, 0x038a0389},
+	{0x0000a03c, 0x038c038b},
+	{0x0000a040, 0x0390038d},
+	{0x0000a044, 0x03920391},
+	{0x0000a048, 0x03940393},
+	{0x0000a04c, 0x03960395},
+	{0x0000a050, 0x00000000},
+	{0x0000a054, 0x00000000},
+	{0x0000a058, 0x00000000},
+	{0x0000a05c, 0x00000000},
+	{0x0000a060, 0x00000000},
+	{0x0000a064, 0x00000000},
+	{0x0000a068, 0x00000000},
+	{0x0000a06c, 0x00000000},
+	{0x0000a070, 0x00000000},
+	{0x0000a074, 0x00000000},
+	{0x0000a078, 0x00000000},
+	{0x0000a07c, 0x00000000},
+	{0x0000a080, 0x29292929},
+	{0x0000a084, 0x29292929},
+	{0x0000a088, 0x29292929},
+	{0x0000a08c, 0x29292929},
+	{0x0000a090, 0x22292929},
+	{0x0000a094, 0x1d1d2222},
+	{0x0000a098, 0x0c111117},
+	{0x0000a09c, 0x00030303},
+	{0x0000a0a0, 0x00000000},
+	{0x0000a0a4, 0x00000000},
+	{0x0000a0a8, 0x00000000},
+	{0x0000a0ac, 0x00000000},
+	{0x0000a0b0, 0x00000000},
+	{0x0000a0b4, 0x00000000},
+	{0x0000a0b8, 0x00000000},
+	{0x0000a0bc, 0x00000000},
+	{0x0000a0c0, 0x001f0000},
+	{0x0000a0c4, 0x01000101},
+	{0x0000a0c8, 0x011e011f},
+	{0x0000a0cc, 0x011c011d},
+	{0x0000a0d0, 0x02030204},
+	{0x0000a0d4, 0x02010202},
+	{0x0000a0d8, 0x021f0200},
+	{0x0000a0dc, 0x0302021e},
+	{0x0000a0e0, 0x03000301},
+	{0x0000a0e4, 0x031e031f},
+	{0x0000a0e8, 0x0402031d},
+	{0x0000a0ec, 0x04000401},
+	{0x0000a0f0, 0x041e041f},
+	{0x0000a0f4, 0x0502041d},
+	{0x0000a0f8, 0x05000501},
+	{0x0000a0fc, 0x051e051f},
+	{0x0000a100, 0x06010602},
+	{0x0000a104, 0x061f0600},
+	{0x0000a108, 0x061d061e},
+	{0x0000a10c, 0x07020703},
+	{0x0000a110, 0x07000701},
+	{0x0000a114, 0x00000000},
+	{0x0000a118, 0x00000000},
+	{0x0000a11c, 0x00000000},
+	{0x0000a120, 0x00000000},
+	{0x0000a124, 0x00000000},
+	{0x0000a128, 0x00000000},
+	{0x0000a12c, 0x00000000},
+	{0x0000a130, 0x00000000},
+	{0x0000a134, 0x00000000},
+	{0x0000a138, 0x00000000},
+	{0x0000a13c, 0x00000000},
+	{0x0000a140, 0x001f0000},
+	{0x0000a144, 0x01000101},
+	{0x0000a148, 0x011e011f},
+	{0x0000a14c, 0x011c011d},
+	{0x0000a150, 0x02030204},
+	{0x0000a154, 0x02010202},
+	{0x0000a158, 0x021f0200},
+	{0x0000a15c, 0x0302021e},
+	{0x0000a160, 0x03000301},
+	{0x0000a164, 0x031e031f},
+	{0x0000a168, 0x0402031d},
+	{0x0000a16c, 0x04000401},
+	{0x0000a170, 0x041e041f},
+	{0x0000a174, 0x0502041d},
+	{0x0000a178, 0x05000501},
+	{0x0000a17c, 0x051e051f},
+	{0x0000a180, 0x06010602},
+	{0x0000a184, 0x061f0600},
+	{0x0000a188, 0x061d061e},
+	{0x0000a18c, 0x07020703},
+	{0x0000a190, 0x07000701},
+	{0x0000a194, 0x00000000},
+	{0x0000a198, 0x00000000},
+	{0x0000a19c, 0x00000000},
+	{0x0000a1a0, 0x00000000},
+	{0x0000a1a4, 0x00000000},
+	{0x0000a1a8, 0x00000000},
+	{0x0000a1ac, 0x00000000},
+	{0x0000a1b0, 0x00000000},
+	{0x0000a1b4, 0x00000000},
+	{0x0000a1b8, 0x00000000},
+	{0x0000a1bc, 0x00000000},
+	{0x0000a1c0, 0x00000000},
+	{0x0000a1c4, 0x00000000},
+	{0x0000a1c8, 0x00000000},
+	{0x0000a1cc, 0x00000000},
+	{0x0000a1d0, 0x00000000},
+	{0x0000a1d4, 0x00000000},
+	{0x0000a1d8, 0x00000000},
+	{0x0000a1dc, 0x00000000},
+	{0x0000a1e0, 0x00000000},
+	{0x0000a1e4, 0x00000000},
+	{0x0000a1e8, 0x00000000},
+	{0x0000a1ec, 0x00000000},
+	{0x0000a1f0, 0x00000396},
+	{0x0000a1f4, 0x00000396},
+	{0x0000a1f8, 0x00000396},
+	{0x0000a1fc, 0x00000196},
+	{0x0000b000, 0x00010000},
+	{0x0000b004, 0x00030002},
+	{0x0000b008, 0x00050004},
+	{0x0000b00c, 0x00810080},
+	{0x0000b010, 0x00830082},
+	{0x0000b014, 0x01810180},
+	{0x0000b018, 0x01830182},
+	{0x0000b01c, 0x01850184},
+	{0x0000b020, 0x02810280},
+	{0x0000b024, 0x02830282},
+	{0x0000b028, 0x02850284},
+	{0x0000b02c, 0x02890288},
+	{0x0000b030, 0x028b028a},
+	{0x0000b034, 0x0388028c},
+	{0x0000b038, 0x038a0389},
+	{0x0000b03c, 0x038c038b},
+	{0x0000b040, 0x0390038d},
+	{0x0000b044, 0x03920391},
+	{0x0000b048, 0x03940393},
+	{0x0000b04c, 0x03960395},
+	{0x0000b050, 0x00000000},
+	{0x0000b054, 0x00000000},
+	{0x0000b058, 0x00000000},
+	{0x0000b05c, 0x00000000},
+	{0x0000b060, 0x00000000},
+	{0x0000b064, 0x00000000},
+	{0x0000b068, 0x00000000},
+	{0x0000b06c, 0x00000000},
+	{0x0000b070, 0x00000000},
+	{0x0000b074, 0x00000000},
+	{0x0000b078, 0x00000000},
+	{0x0000b07c, 0x00000000},
+	{0x0000b080, 0x32323232},
+	{0x0000b084, 0x2f2f3232},
+	{0x0000b088, 0x23282a2d},
+	{0x0000b08c, 0x1c1e2123},
+	{0x0000b090, 0x14171919},
+	{0x0000b094, 0x0e0e1214},
+	{0x0000b098, 0x03050707},
+	{0x0000b09c, 0x00030303},
+	{0x0000b0a0, 0x00000000},
+	{0x0000b0a4, 0x00000000},
+	{0x0000b0a8, 0x00000000},
+	{0x0000b0ac, 0x00000000},
+	{0x0000b0b0, 0x00000000},
+	{0x0000b0b4, 0x00000000},
+	{0x0000b0b8, 0x00000000},
+	{0x0000b0bc, 0x00000000},
+	{0x0000b0c0, 0x003f0020},
+	{0x0000b0c4, 0x00400041},
+	{0x0000b0c8, 0x0140005f},
+	{0x0000b0cc, 0x0160015f},
+	{0x0000b0d0, 0x017e017f},
+	{0x0000b0d4, 0x02410242},
+	{0x0000b0d8, 0x025f0240},
+	{0x0000b0dc, 0x027f0260},
+	{0x0000b0e0, 0x0341027e},
+	{0x0000b0e4, 0x035f0340},
+	{0x0000b0e8, 0x037f0360},
+	{0x0000b0ec, 0x04400441},
+	{0x0000b0f0, 0x0460045f},
+	{0x0000b0f4, 0x0541047f},
+	{0x0000b0f8, 0x055f0540},
+	{0x0000b0fc, 0x057f0560},
+	{0x0000b100, 0x06400641},
+	{0x0000b104, 0x0660065f},
+	{0x0000b108, 0x067e067f},
+	{0x0000b10c, 0x07410742},
+	{0x0000b110, 0x075f0740},
+	{0x0000b114, 0x077f0760},
+	{0x0000b118, 0x07800781},
+	{0x0000b11c, 0x07a0079f},
+	{0x0000b120, 0x07c107bf},
+	{0x0000b124, 0x000007c0},
+	{0x0000b128, 0x00000000},
+	{0x0000b12c, 0x00000000},
+	{0x0000b130, 0x00000000},
+	{0x0000b134, 0x00000000},
+	{0x0000b138, 0x00000000},
+	{0x0000b13c, 0x00000000},
+	{0x0000b140, 0x003f0020},
+	{0x0000b144, 0x00400041},
+	{0x0000b148, 0x0140005f},
+	{0x0000b14c, 0x0160015f},
+	{0x0000b150, 0x017e017f},
+	{0x0000b154, 0x02410242},
+	{0x0000b158, 0x025f0240},
+	{0x0000b15c, 0x027f0260},
+	{0x0000b160, 0x0341027e},
+	{0x0000b164, 0x035f0340},
+	{0x0000b168, 0x037f0360},
+	{0x0000b16c, 0x04400441},
+	{0x0000b170, 0x0460045f},
+	{0x0000b174, 0x0541047f},
+	{0x0000b178, 0x055f0540},
+	{0x0000b17c, 0x057f0560},
+	{0x0000b180, 0x06400641},
+	{0x0000b184, 0x0660065f},
+	{0x0000b188, 0x067e067f},
+	{0x0000b18c, 0x07410742},
+	{0x0000b190, 0x075f0740},
+	{0x0000b194, 0x077f0760},
+	{0x0000b198, 0x07800781},
+	{0x0000b19c, 0x07a0079f},
+	{0x0000b1a0, 0x07c107bf},
+	{0x0000b1a4, 0x000007c0},
+	{0x0000b1a8, 0x00000000},
+	{0x0000b1ac, 0x00000000},
+	{0x0000b1b0, 0x00000000},
+	{0x0000b1b4, 0x00000000},
+	{0x0000b1b8, 0x00000000},
+	{0x0000b1bc, 0x00000000},
+	{0x0000b1c0, 0x00000000},
+	{0x0000b1c4, 0x00000000},
+	{0x0000b1c8, 0x00000000},
+	{0x0000b1cc, 0x00000000},
+	{0x0000b1d0, 0x00000000},
+	{0x0000b1d4, 0x00000000},
+	{0x0000b1d8, 0x00000000},
+	{0x0000b1dc, 0x00000000},
+	{0x0000b1e0, 0x00000000},
+	{0x0000b1e4, 0x00000000},
+	{0x0000b1e8, 0x00000000},
+	{0x0000b1ec, 0x00000000},
+	{0x0000b1f0, 0x00000396},
+	{0x0000b1f4, 0x00000396},
+	{0x0000b1f8, 0x00000396},
+	{0x0000b1fc, 0x00000196},
+};
+
+static const u32 ar9462_2p1_common_5g_xlna_only_rx_gain[][2] = {
+	/* Addr      allmodes  */
+	{0x0000a000, 0x00010000},
+	{0x0000a004, 0x00030002},
+	{0x0000a008, 0x00050004},
+	{0x0000a00c, 0x00810080},
+	{0x0000a010, 0x00830082},
+	{0x0000a014, 0x01810180},
+	{0x0000a018, 0x01830182},
+	{0x0000a01c, 0x01850184},
+	{0x0000a020, 0x01890188},
+	{0x0000a024, 0x018b018a},
+	{0x0000a028, 0x018d018c},
+	{0x0000a02c, 0x03820190},
+	{0x0000a030, 0x03840383},
+	{0x0000a034, 0x03880385},
+	{0x0000a038, 0x038a0389},
+	{0x0000a03c, 0x038c038b},
+	{0x0000a040, 0x0390038d},
+	{0x0000a044, 0x03920391},
+	{0x0000a048, 0x03940393},
+	{0x0000a04c, 0x03960395},
+	{0x0000a050, 0x00000000},
+	{0x0000a054, 0x00000000},
+	{0x0000a058, 0x00000000},
+	{0x0000a05c, 0x00000000},
+	{0x0000a060, 0x00000000},
+	{0x0000a064, 0x00000000},
+	{0x0000a068, 0x00000000},
+	{0x0000a06c, 0x00000000},
+	{0x0000a070, 0x00000000},
+	{0x0000a074, 0x00000000},
+	{0x0000a078, 0x00000000},
+	{0x0000a07c, 0x00000000},
+	{0x0000a080, 0x29292929},
+	{0x0000a084, 0x29292929},
+	{0x0000a088, 0x29292929},
+	{0x0000a08c, 0x29292929},
+	{0x0000a090, 0x22292929},
+	{0x0000a094, 0x1d1d2222},
+	{0x0000a098, 0x0c111117},
+	{0x0000a09c, 0x00030303},
+	{0x0000a0a0, 0x00000000},
+	{0x0000a0a4, 0x00000000},
+	{0x0000a0a8, 0x00000000},
+	{0x0000a0ac, 0x00000000},
+	{0x0000a0b0, 0x00000000},
+	{0x0000a0b4, 0x00000000},
+	{0x0000a0b8, 0x00000000},
+	{0x0000a0bc, 0x00000000},
+	{0x0000a0c0, 0x001f0000},
+	{0x0000a0c4, 0x01000101},
+	{0x0000a0c8, 0x011e011f},
+	{0x0000a0cc, 0x011c011d},
+	{0x0000a0d0, 0x02030204},
+	{0x0000a0d4, 0x02010202},
+	{0x0000a0d8, 0x021f0200},
+	{0x0000a0dc, 0x0302021e},
+	{0x0000a0e0, 0x03000301},
+	{0x0000a0e4, 0x031e031f},
+	{0x0000a0e8, 0x0402031d},
+	{0x0000a0ec, 0x04000401},
+	{0x0000a0f0, 0x041e041f},
+	{0x0000a0f4, 0x0502041d},
+	{0x0000a0f8, 0x05000501},
+	{0x0000a0fc, 0x051e051f},
+	{0x0000a100, 0x06010602},
+	{0x0000a104, 0x061f0600},
+	{0x0000a108, 0x061d061e},
+	{0x0000a10c, 0x07020703},
+	{0x0000a110, 0x07000701},
+	{0x0000a114, 0x00000000},
+	{0x0000a118, 0x00000000},
+	{0x0000a11c, 0x00000000},
+	{0x0000a120, 0x00000000},
+	{0x0000a124, 0x00000000},
+	{0x0000a128, 0x00000000},
+	{0x0000a12c, 0x00000000},
+	{0x0000a130, 0x00000000},
+	{0x0000a134, 0x00000000},
+	{0x0000a138, 0x00000000},
+	{0x0000a13c, 0x00000000},
+	{0x0000a140, 0x001f0000},
+	{0x0000a144, 0x01000101},
+	{0x0000a148, 0x011e011f},
+	{0x0000a14c, 0x011c011d},
+	{0x0000a150, 0x02030204},
+	{0x0000a154, 0x02010202},
+	{0x0000a158, 0x021f0200},
+	{0x0000a15c, 0x0302021e},
+	{0x0000a160, 0x03000301},
+	{0x0000a164, 0x031e031f},
+	{0x0000a168, 0x0402031d},
+	{0x0000a16c, 0x04000401},
+	{0x0000a170, 0x041e041f},
+	{0x0000a174, 0x0502041d},
+	{0x0000a178, 0x05000501},
+	{0x0000a17c, 0x051e051f},
+	{0x0000a180, 0x06010602},
+	{0x0000a184, 0x061f0600},
+	{0x0000a188, 0x061d061e},
+	{0x0000a18c, 0x07020703},
+	{0x0000a190, 0x07000701},
+	{0x0000a194, 0x00000000},
+	{0x0000a198, 0x00000000},
+	{0x0000a19c, 0x00000000},
+	{0x0000a1a0, 0x00000000},
+	{0x0000a1a4, 0x00000000},
+	{0x0000a1a8, 0x00000000},
+	{0x0000a1ac, 0x00000000},
+	{0x0000a1b0, 0x00000000},
+	{0x0000a1b4, 0x00000000},
+	{0x0000a1b8, 0x00000000},
+	{0x0000a1bc, 0x00000000},
+	{0x0000a1c0, 0x00000000},
+	{0x0000a1c4, 0x00000000},
+	{0x0000a1c8, 0x00000000},
+	{0x0000a1cc, 0x00000000},
+	{0x0000a1d0, 0x00000000},
+	{0x0000a1d4, 0x00000000},
+	{0x0000a1d8, 0x00000000},
+	{0x0000a1dc, 0x00000000},
+	{0x0000a1e0, 0x00000000},
+	{0x0000a1e4, 0x00000000},
+	{0x0000a1e8, 0x00000000},
+	{0x0000a1ec, 0x00000000},
+	{0x0000a1f0, 0x00000396},
+	{0x0000a1f4, 0x00000396},
+	{0x0000a1f8, 0x00000396},
+	{0x0000a1fc, 0x00000196},
+	{0x0000b000, 0x00010000},
+	{0x0000b004, 0x00030002},
+	{0x0000b008, 0x00050004},
+	{0x0000b00c, 0x00810080},
+	{0x0000b010, 0x00830082},
+	{0x0000b014, 0x01810180},
+	{0x0000b018, 0x01830182},
+	{0x0000b01c, 0x01850184},
+	{0x0000b020, 0x02810280},
+	{0x0000b024, 0x02830282},
+	{0x0000b028, 0x02850284},
+	{0x0000b02c, 0x02890288},
+	{0x0000b030, 0x028b028a},
+	{0x0000b034, 0x0388028c},
+	{0x0000b038, 0x038a0389},
+	{0x0000b03c, 0x038c038b},
+	{0x0000b040, 0x0390038d},
+	{0x0000b044, 0x03920391},
+	{0x0000b048, 0x03940393},
+	{0x0000b04c, 0x03960395},
+	{0x0000b050, 0x00000000},
+	{0x0000b054, 0x00000000},
+	{0x0000b058, 0x00000000},
+	{0x0000b05c, 0x00000000},
+	{0x0000b060, 0x00000000},
+	{0x0000b064, 0x00000000},
+	{0x0000b068, 0x00000000},
+	{0x0000b06c, 0x00000000},
+	{0x0000b070, 0x00000000},
+	{0x0000b074, 0x00000000},
+	{0x0000b078, 0x00000000},
+	{0x0000b07c, 0x00000000},
+	{0x0000b080, 0x2a2d2f32},
+	{0x0000b084, 0x21232328},
+	{0x0000b088, 0x19191c1e},
+	{0x0000b08c, 0x12141417},
+	{0x0000b090, 0x07070e0e},
+	{0x0000b094, 0x03030305},
+	{0x0000b098, 0x00000003},
+	{0x0000b09c, 0x00000000},
+	{0x0000b0a0, 0x00000000},
+	{0x0000b0a4, 0x00000000},
+	{0x0000b0a8, 0x00000000},
+	{0x0000b0ac, 0x00000000},
+	{0x0000b0b0, 0x00000000},
+	{0x0000b0b4, 0x00000000},
+	{0x0000b0b8, 0x00000000},
+	{0x0000b0bc, 0x00000000},
+	{0x0000b0c0, 0x003f0020},
+	{0x0000b0c4, 0x00400041},
+	{0x0000b0c8, 0x0140005f},
+	{0x0000b0cc, 0x0160015f},
+	{0x0000b0d0, 0x017e017f},
+	{0x0000b0d4, 0x02410242},
+	{0x0000b0d8, 0x025f0240},
+	{0x0000b0dc, 0x027f0260},
+	{0x0000b0e0, 0x0341027e},
+	{0x0000b0e4, 0x035f0340},
+	{0x0000b0e8, 0x037f0360},
+	{0x0000b0ec, 0x04400441},
+	{0x0000b0f0, 0x0460045f},
+	{0x0000b0f4, 0x0541047f},
+	{0x0000b0f8, 0x055f0540},
+	{0x0000b0fc, 0x057f0560},
+	{0x0000b100, 0x06400641},
+	{0x0000b104, 0x0660065f},
+	{0x0000b108, 0x067e067f},
+	{0x0000b10c, 0x07410742},
+	{0x0000b110, 0x075f0740},
+	{0x0000b114, 0x077f0760},
+	{0x0000b118, 0x07800781},
+	{0x0000b11c, 0x07a0079f},
+	{0x0000b120, 0x07c107bf},
+	{0x0000b124, 0x000007c0},
+	{0x0000b128, 0x00000000},
+	{0x0000b12c, 0x00000000},
+	{0x0000b130, 0x00000000},
+	{0x0000b134, 0x00000000},
+	{0x0000b138, 0x00000000},
+	{0x0000b13c, 0x00000000},
+	{0x0000b140, 0x003f0020},
+	{0x0000b144, 0x00400041},
+	{0x0000b148, 0x0140005f},
+	{0x0000b14c, 0x0160015f},
+	{0x0000b150, 0x017e017f},
+	{0x0000b154, 0x02410242},
+	{0x0000b158, 0x025f0240},
+	{0x0000b15c, 0x027f0260},
+	{0x0000b160, 0x0341027e},
+	{0x0000b164, 0x035f0340},
+	{0x0000b168, 0x037f0360},
+	{0x0000b16c, 0x04400441},
+	{0x0000b170, 0x0460045f},
+	{0x0000b174, 0x0541047f},
+	{0x0000b178, 0x055f0540},
+	{0x0000b17c, 0x057f0560},
+	{0x0000b180, 0x06400641},
+	{0x0000b184, 0x0660065f},
+	{0x0000b188, 0x067e067f},
+	{0x0000b18c, 0x07410742},
+	{0x0000b190, 0x075f0740},
+	{0x0000b194, 0x077f0760},
+	{0x0000b198, 0x07800781},
+	{0x0000b19c, 0x07a0079f},
+	{0x0000b1a0, 0x07c107bf},
+	{0x0000b1a4, 0x000007c0},
+	{0x0000b1a8, 0x00000000},
+	{0x0000b1ac, 0x00000000},
+	{0x0000b1b0, 0x00000000},
+	{0x0000b1b4, 0x00000000},
+	{0x0000b1b8, 0x00000000},
+	{0x0000b1bc, 0x00000000},
+	{0x0000b1c0, 0x00000000},
+	{0x0000b1c4, 0x00000000},
+	{0x0000b1c8, 0x00000000},
+	{0x0000b1cc, 0x00000000},
+	{0x0000b1d0, 0x00000000},
+	{0x0000b1d4, 0x00000000},
+	{0x0000b1d8, 0x00000000},
+	{0x0000b1dc, 0x00000000},
+	{0x0000b1e0, 0x00000000},
+	{0x0000b1e4, 0x00000000},
+	{0x0000b1e8, 0x00000000},
+	{0x0000b1ec, 0x00000000},
+	{0x0000b1f0, 0x00000396},
+	{0x0000b1f4, 0x00000396},
+	{0x0000b1f8, 0x00000396},
+	{0x0000b1fc, 0x00000196},
+};
+
+static const u32 ar9462_2p1_modes_low_ob_db_tx_gain[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x000098bc, 0x00000002, 0x00000002, 0x00000002, 0x00000002},
+	{0x0000a2dc, 0x0380c7fc, 0x0380c7fc, 0x03aaa352, 0x03aaa352},
+	{0x0000a2e0, 0x0000f800, 0x0000f800, 0x03ccc584, 0x03ccc584},
+	{0x0000a2e4, 0x03ff0000, 0x03ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000a2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+	{0x0000a410, 0x000050d9, 0x000050d9, 0x000050d9, 0x000050d9},
+	{0x0000a458, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a500, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a504, 0x06000003, 0x06000003, 0x04000002, 0x04000002},
+	{0x0000a508, 0x0a000020, 0x0a000020, 0x08000004, 0x08000004},
+	{0x0000a50c, 0x10000023, 0x10000023, 0x0b000200, 0x0b000200},
+	{0x0000a510, 0x16000220, 0x16000220, 0x0f000202, 0x0f000202},
+	{0x0000a514, 0x1c000223, 0x1c000223, 0x12000400, 0x12000400},
+	{0x0000a518, 0x21020220, 0x21020220, 0x16000402, 0x16000402},
+	{0x0000a51c, 0x27020223, 0x27020223, 0x19000404, 0x19000404},
+	{0x0000a520, 0x2b022220, 0x2b022220, 0x1c000603, 0x1c000603},
+	{0x0000a524, 0x2f022222, 0x2f022222, 0x21000a02, 0x21000a02},
+	{0x0000a528, 0x34022225, 0x34022225, 0x25000a04, 0x25000a04},
+	{0x0000a52c, 0x3a02222a, 0x3a02222a, 0x28000a20, 0x28000a20},
+	{0x0000a530, 0x3e02222c, 0x3e02222c, 0x2c000e20, 0x2c000e20},
+	{0x0000a534, 0x4202242a, 0x4202242a, 0x30000e22, 0x30000e22},
+	{0x0000a538, 0x4702244a, 0x4702244a, 0x34000e24, 0x34000e24},
+	{0x0000a53c, 0x4b02244c, 0x4b02244c, 0x38001640, 0x38001640},
+	{0x0000a540, 0x4e02246c, 0x4e02246c, 0x3c001660, 0x3c001660},
+	{0x0000a544, 0x5302266c, 0x5302266c, 0x3f001861, 0x3f001861},
+	{0x0000a548, 0x5702286c, 0x5702286c, 0x43001a81, 0x43001a81},
+	{0x0000a54c, 0x5c04286b, 0x5c04286b, 0x47001a83, 0x47001a83},
+	{0x0000a550, 0x61042a6c, 0x61042a6c, 0x4a001c84, 0x4a001c84},
+	{0x0000a554, 0x66062a6c, 0x66062a6c, 0x4e001ce3, 0x4e001ce3},
+	{0x0000a558, 0x6b062e6c, 0x6b062e6c, 0x52001ce5, 0x52001ce5},
+	{0x0000a55c, 0x7006308c, 0x7006308c, 0x56001ce9, 0x56001ce9},
+	{0x0000a560, 0x730a308a, 0x730a308a, 0x5a001ceb, 0x5a001ceb},
+	{0x0000a564, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a568, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a56c, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a570, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a574, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a578, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a57c, 0x770a308c, 0x770a308c, 0x5d001eec, 0x5d001eec},
+	{0x0000a600, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a604, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a608, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a60c, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a610, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a614, 0x01404000, 0x01404000, 0x01404000, 0x01404000},
+	{0x0000a618, 0x01404501, 0x01404501, 0x01404501, 0x01404501},
+	{0x0000a61c, 0x02008802, 0x02008802, 0x02008501, 0x02008501},
+	{0x0000a620, 0x0300cc03, 0x0300cc03, 0x0280ca03, 0x0280ca03},
+	{0x0000a624, 0x0300cc03, 0x0300cc03, 0x03010c04, 0x03010c04},
+	{0x0000a628, 0x0300cc03, 0x0300cc03, 0x04014c04, 0x04014c04},
+	{0x0000a62c, 0x03810c03, 0x03810c03, 0x04015005, 0x04015005},
+	{0x0000a630, 0x03810e04, 0x03810e04, 0x04015005, 0x04015005},
+	{0x0000a634, 0x03810e04, 0x03810e04, 0x04015005, 0x04015005},
+	{0x0000a638, 0x03810e04, 0x03810e04, 0x04015005, 0x04015005},
+	{0x0000a63c, 0x03810e04, 0x03810e04, 0x04015005, 0x04015005},
+	{0x0000b2dc, 0x0380c7fc, 0x0380c7fc, 0x03aaa352, 0x03aaa352},
+	{0x0000b2e0, 0x0000f800, 0x0000f800, 0x03ccc584, 0x03ccc584},
+	{0x0000b2e4, 0x03ff0000, 0x03ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000b2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+	{0x00016044, 0x012482d4, 0x012482d4, 0x012482d4, 0x012482d4},
+	{0x00016048, 0x64992060, 0x64992060, 0x64992060, 0x64992060},
+	{0x00016054, 0x6db60000, 0x6db60000, 0x6db60000, 0x6db60000},
+	{0x00016444, 0x012482d4, 0x012482d4, 0x012482d4, 0x012482d4},
+	{0x00016448, 0x64992000, 0x64992000, 0x64992000, 0x64992000},
+	{0x00016454, 0x6db60000, 0x6db60000, 0x6db60000, 0x6db60000},
+};
+
+static const u32 ar9462_2p1_modes_high_ob_db_tx_gain[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x000098bc, 0x00000002, 0x00000002, 0x00000002, 0x00000002},
+	{0x0000a2dc, 0x01feee00, 0x01feee00, 0x03aaa352, 0x03aaa352},
+	{0x0000a2e0, 0x0000f000, 0x0000f000, 0x03ccc584, 0x03ccc584},
+	{0x0000a2e4, 0x01ff0000, 0x01ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000a2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+	{0x0000a410, 0x000050da, 0x000050da, 0x000050de, 0x000050de},
+	{0x0000a458, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a500, 0x00002220, 0x00002220, 0x00000000, 0x00000000},
+	{0x0000a504, 0x06002223, 0x06002223, 0x04000002, 0x04000002},
+	{0x0000a508, 0x0a022220, 0x0a022220, 0x08000004, 0x08000004},
+	{0x0000a50c, 0x0f022223, 0x0f022223, 0x0b000200, 0x0b000200},
+	{0x0000a510, 0x14022620, 0x14022620, 0x0f000202, 0x0f000202},
+	{0x0000a514, 0x18022622, 0x18022622, 0x11000400, 0x11000400},
+	{0x0000a518, 0x1b022822, 0x1b022822, 0x15000402, 0x15000402},
+	{0x0000a51c, 0x20022842, 0x20022842, 0x19000404, 0x19000404},
+	{0x0000a520, 0x22022c41, 0x22022c41, 0x1b000603, 0x1b000603},
+	{0x0000a524, 0x28023042, 0x28023042, 0x1f000a02, 0x1f000a02},
+	{0x0000a528, 0x2c023044, 0x2c023044, 0x23000a04, 0x23000a04},
+	{0x0000a52c, 0x2f023644, 0x2f023644, 0x26000a20, 0x26000a20},
+	{0x0000a530, 0x34025643, 0x34025643, 0x2a000e20, 0x2a000e20},
+	{0x0000a534, 0x38025a44, 0x38025a44, 0x2e000e22, 0x2e000e22},
+	{0x0000a538, 0x3b025e45, 0x3b025e45, 0x31000e24, 0x31000e24},
+	{0x0000a53c, 0x41025e4a, 0x41025e4a, 0x34001640, 0x34001640},
+	{0x0000a540, 0x48025e6c, 0x48025e6c, 0x38001660, 0x38001660},
+	{0x0000a544, 0x4e025e8e, 0x4e025e8e, 0x3b001861, 0x3b001861},
+	{0x0000a548, 0x55025eb3, 0x55025eb3, 0x3e001a81, 0x3e001a81},
+	{0x0000a54c, 0x58025ef3, 0x58025ef3, 0x42001a83, 0x42001a83},
+	{0x0000a550, 0x5d025ef6, 0x5d025ef6, 0x44001a84, 0x44001a84},
+	{0x0000a554, 0x62025f56, 0x62025f56, 0x48001ce3, 0x48001ce3},
+	{0x0000a558, 0x66027f56, 0x66027f56, 0x4c001ce5, 0x4c001ce5},
+	{0x0000a55c, 0x6a029f56, 0x6a029f56, 0x50001ce9, 0x50001ce9},
+	{0x0000a560, 0x70049f56, 0x70049f56, 0x54001ceb, 0x54001ceb},
+	{0x0000a564, 0x751ffff6, 0x751ffff6, 0x56001eec, 0x56001eec},
+	{0x0000a568, 0x751ffff6, 0x751ffff6, 0x58001ef0, 0x58001ef0},
+	{0x0000a56c, 0x751ffff6, 0x751ffff6, 0x5a001ef4, 0x5a001ef4},
+	{0x0000a570, 0x751ffff6, 0x751ffff6, 0x5c001ff6, 0x5c001ff6},
+	{0x0000a574, 0x751ffff6, 0x751ffff6, 0x5c001ff6, 0x5c001ff6},
+	{0x0000a578, 0x751ffff6, 0x751ffff6, 0x5c001ff6, 0x5c001ff6},
+	{0x0000a57c, 0x751ffff6, 0x751ffff6, 0x5c001ff6, 0x5c001ff6},
+	{0x0000a600, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a604, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a608, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a60c, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a610, 0x00804000, 0x00804000, 0x00000000, 0x00000000},
+	{0x0000a614, 0x00804201, 0x00804201, 0x01404000, 0x01404000},
+	{0x0000a618, 0x0280c802, 0x0280c802, 0x01404501, 0x01404501},
+	{0x0000a61c, 0x0280ca03, 0x0280ca03, 0x02008501, 0x02008501},
+	{0x0000a620, 0x04c15104, 0x04c15104, 0x0280ca03, 0x0280ca03},
+	{0x0000a624, 0x04c15305, 0x04c15305, 0x03010c04, 0x03010c04},
+	{0x0000a628, 0x04c15305, 0x04c15305, 0x04014c04, 0x04014c04},
+	{0x0000a62c, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a630, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a634, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a638, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a63c, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000b2dc, 0x01feee00, 0x01feee00, 0x03aaa352, 0x03aaa352},
+	{0x0000b2e0, 0x0000f000, 0x0000f000, 0x03ccc584, 0x03ccc584},
+	{0x0000b2e4, 0x01ff0000, 0x01ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000b2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+	{0x00016044, 0x056d82e4, 0x056d82e4, 0x056d82e4, 0x056d82e4},
+	{0x00016048, 0x8db49060, 0x8db49060, 0x8db49060, 0x8db49060},
+	{0x00016054, 0x6db60000, 0x6db60000, 0x6db60000, 0x6db60000},
+	{0x00016444, 0x056d82e4, 0x056d82e4, 0x056d82e4, 0x056d82e4},
+	{0x00016448, 0x8db49000, 0x8db49000, 0x8db49000, 0x8db49000},
+	{0x00016454, 0x6db60000, 0x6db60000, 0x6db60000, 0x6db60000},
+};
+
+static const u32 ar9462_2p1_modes_mix_ob_db_tx_gain[][5] = {
+	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */
+	{0x000098bc, 0x00000002, 0x00000002, 0x00000002, 0x00000002},
+	{0x0000a2dc, 0x01feee00, 0x01feee00, 0x03aaa352, 0x03aaa352},
+	{0x0000a2e0, 0x0000f000, 0x0000f000, 0x03ccc584, 0x03ccc584},
+	{0x0000a2e4, 0x01ff0000, 0x01ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000a2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+	{0x0000a410, 0x0000d0da, 0x0000d0da, 0x0000d0de, 0x0000d0de},
+	{0x0000a458, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a500, 0x00002220, 0x00002220, 0x00000000, 0x00000000},
+	{0x0000a504, 0x06002223, 0x06002223, 0x04000002, 0x04000002},
+	{0x0000a508, 0x0a022220, 0x0a022220, 0x08000004, 0x08000004},
+	{0x0000a50c, 0x0f022223, 0x0f022223, 0x0b000200, 0x0b000200},
+	{0x0000a510, 0x14022620, 0x14022620, 0x0f000202, 0x0f000202},
+	{0x0000a514, 0x18022622, 0x18022622, 0x12000400, 0x12000400},
+	{0x0000a518, 0x1b022822, 0x1b022822, 0x16000402, 0x16000402},
+	{0x0000a51c, 0x20022842, 0x20022842, 0x19000404, 0x19000404},
+	{0x0000a520, 0x22022c41, 0x22022c41, 0x1c000603, 0x1c000603},
+	{0x0000a524, 0x28023042, 0x28023042, 0x21000a02, 0x21000a02},
+	{0x0000a528, 0x2c023044, 0x2c023044, 0x25000a04, 0x25000a04},
+	{0x0000a52c, 0x2f023644, 0x2f023644, 0x28000a20, 0x28000a20},
+	{0x0000a530, 0x34025643, 0x34025643, 0x2c000e20, 0x2c000e20},
+	{0x0000a534, 0x38025a44, 0x38025a44, 0x30000e22, 0x30000e22},
+	{0x0000a538, 0x3b025e45, 0x3b025e45, 0x34000e24, 0x34000e24},
+	{0x0000a53c, 0x41025e4a, 0x41025e4a, 0x38001640, 0x38001640},
+	{0x0000a540, 0x48025e6c, 0x48025e6c, 0x3c001660, 0x3c001660},
+	{0x0000a544, 0x4e025e8e, 0x4e025e8e, 0x3f001861, 0x3f001861},
+	{0x0000a548, 0x55025eb3, 0x55025eb3, 0x43001a81, 0x43001a81},
+	{0x0000a54c, 0x58025ef3, 0x58025ef3, 0x47001a83, 0x47001a83},
+	{0x0000a550, 0x5d025ef6, 0x5d025ef6, 0x4a001c84, 0x4a001c84},
+	{0x0000a554, 0x62025f56, 0x62025f56, 0x4e001ce3, 0x4e001ce3},
+	{0x0000a558, 0x66027f56, 0x66027f56, 0x52001ce5, 0x52001ce5},
+	{0x0000a55c, 0x6a029f56, 0x6a029f56, 0x56001ce9, 0x56001ce9},
+	{0x0000a560, 0x70049f56, 0x70049f56, 0x5a001ceb, 0x5a001ceb},
+	{0x0000a564, 0x751ffff6, 0x751ffff6, 0x5c001eec, 0x5c001eec},
+	{0x0000a568, 0x751ffff6, 0x751ffff6, 0x5e001ef0, 0x5e001ef0},
+	{0x0000a56c, 0x751ffff6, 0x751ffff6, 0x60001ef4, 0x60001ef4},
+	{0x0000a570, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a574, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a578, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a57c, 0x751ffff6, 0x751ffff6, 0x62001ff6, 0x62001ff6},
+	{0x0000a600, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a604, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a608, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a60c, 0x00000000, 0x00000000, 0x00000000, 0x00000000},
+	{0x0000a610, 0x00804000, 0x00804000, 0x00000000, 0x00000000},
+	{0x0000a614, 0x00804201, 0x00804201, 0x01404000, 0x01404000},
+	{0x0000a618, 0x0280c802, 0x0280c802, 0x01404501, 0x01404501},
+	{0x0000a61c, 0x0280ca03, 0x0280ca03, 0x02008501, 0x02008501},
+	{0x0000a620, 0x04c15104, 0x04c15104, 0x0280ca03, 0x0280ca03},
+	{0x0000a624, 0x04c15305, 0x04c15305, 0x03010c04, 0x03010c04},
+	{0x0000a628, 0x04c15305, 0x04c15305, 0x04014c04, 0x04014c04},
+	{0x0000a62c, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a630, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a634, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a638, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000a63c, 0x04c15305, 0x04c15305, 0x04015005, 0x04015005},
+	{0x0000b2dc, 0x01feee00, 0x01feee00, 0x03aaa352, 0x03aaa352},
+	{0x0000b2e0, 0x0000f000, 0x0000f000, 0x03ccc584, 0x03ccc584},
+	{0x0000b2e4, 0x01ff0000, 0x01ff0000, 0x03f0f800, 0x03f0f800},
+	{0x0000b2e8, 0x00000000, 0x00000000, 0x03ff0000, 0x03ff0000},
+};
+
+static const u32 ar9462_2p1_modes_fast_clock[][3] = {
+	/* Addr      5G_HT20     5G_HT40   */
+	{0x00001030, 0x00000268, 0x000004d0},
+	{0x00001070, 0x0000018c, 0x00000318},
+	{0x000010b0, 0x00000fd0, 0x00001fa0},
+	{0x00008014, 0x044c044c, 0x08980898},
+	{0x0000801c, 0x148ec02b, 0x148ec057},
+	{0x00008318, 0x000044c0, 0x00008980},
+	{0x00009e00, 0x0372131c, 0x0372131c},
+	{0x0000a230, 0x0000400b, 0x00004016},
+	{0x0000a254, 0x00000898, 0x00001130},
+};
+
+static const u32 ar9462_2p1_baseband_core_txfir_coeff_japan_2484[][2] = {
+	/* Addr      allmodes  */
+	{0x0000a398, 0x00000000},
+	{0x0000a39c, 0x6f7f0301},
+	{0x0000a3a0, 0xca9228ee},
+};
+
+#endif /* INITVALS_9462_2P1_H */
diff --git a/drivers/net/wireless/ath/ath9k/ath9k.h b/drivers/net/wireless/ath/ath9k/ath9k.h
index 42b03dc..c1224b5 100644
--- a/drivers/net/wireless/ath/ath9k/ath9k.h
+++ b/drivers/net/wireless/ath/ath9k/ath9k.h
@@ -296,6 +296,7 @@ struct ath_tx {
 	struct ath_txq txq[ATH9K_NUM_TX_QUEUES];
 	struct ath_descdma txdma;
 	struct ath_txq *txq_map[IEEE80211_NUM_ACS];
+	struct ath_txq *uapsdq;
 	u32 txq_max_pending[IEEE80211_NUM_ACS];
 	u16 max_aggr_framelen[IEEE80211_NUM_ACS][4][32];
 };
@@ -343,6 +344,8 @@ int ath_txq_update(struct ath_softc *sc, int qnum,
 void ath_update_max_aggr_framelen(struct ath_softc *sc, int queue, int txop);
 int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 		 struct ath_tx_control *txctl);
+void ath_tx_cabq(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
+		 struct sk_buff *skb);
 void ath_tx_tasklet(struct ath_softc *sc);
 void ath_tx_edma_tasklet(struct ath_softc *sc);
 int ath_tx_aggr_start(struct ath_softc *sc, struct ieee80211_sta *sta,
@@ -353,6 +356,11 @@ void ath_tx_aggr_resume(struct ath_softc *sc, struct ieee80211_sta *sta, u16 tid
 void ath_tx_aggr_wakeup(struct ath_softc *sc, struct ath_node *an);
 void ath_tx_aggr_sleep(struct ieee80211_sta *sta, struct ath_softc *sc,
 		       struct ath_node *an);
+void ath9k_release_buffered_frames(struct ieee80211_hw *hw,
+				   struct ieee80211_sta *sta,
+				   u16 tids, int nframes,
+				   enum ieee80211_frame_release_type reason,
+				   bool more_data);
 
 /********/
 /* VIFs */
@@ -623,6 +631,11 @@ void ath_ant_comb_update(struct ath_softc *sc);
 /* Main driver core */
 /********************/
 
+#define ATH9K_PCI_CUS198 0x0001
+#define ATH9K_PCI_CUS230 0x0002
+#define ATH9K_PCI_CUS217 0x0004
+#define ATH9K_PCI_WOW    0x0008
+
 /*
  * Default cache line size, in bytes.
  * Used when PCI device not fully initialized by bootrom/BIOS
@@ -642,6 +655,7 @@ enum sc_op_flags {
 	SC_OP_ANI_RUN,
 	SC_OP_PRIM_STA_VIF,
 	SC_OP_HW_RESET,
+	SC_OP_SCANNING,
 };
 
 /* Powersave flags */
@@ -706,6 +720,7 @@ struct ath_softc {
 
 	unsigned int hw_busy_count;
 	unsigned long sc_flags;
+	unsigned long driver_data;
 
 	u32 intrstatus;
 	u16 ps_flags; /* PS_* */
@@ -755,7 +770,6 @@ struct ath_softc {
 	struct rchan *rfs_chan_spec_scan;
 	enum spectral_mode spectral_mode;
 	struct ath_spec_scan spec_config;
-	int scanning;
 
 #ifdef CONFIG_PM_SLEEP
 	atomic_t wow_got_bmiss_intr;
diff --git a/drivers/net/wireless/ath/ath9k/beacon.c b/drivers/net/wireless/ath/ath9k/beacon.c
index 2ff570f..1a17732 100644
--- a/drivers/net/wireless/ath/ath9k/beacon.c
+++ b/drivers/net/wireless/ath/ath9k/beacon.c
@@ -39,7 +39,8 @@ static void ath9k_beaconq_config(struct ath_softc *sc)
 
 	ath9k_hw_get_txq_props(ah, sc->beacon.beaconq, &qi);
 
-	if (sc->sc_ah->opmode == NL80211_IFTYPE_AP) {
+	if (sc->sc_ah->opmode == NL80211_IFTYPE_AP ||
+	    sc->sc_ah->opmode == NL80211_IFTYPE_MESH_POINT) {
 		/* Always burst out beacon and CAB traffic. */
 		qi.tqi_aifs = 1;
 		qi.tqi_cwmin = 0;
@@ -107,23 +108,6 @@ static void ath9k_beacon_setup(struct ath_softc *sc, struct ieee80211_vif *vif,
 	ath9k_hw_set_txdesc(ah, bf->bf_desc, &info);
 }
 
-static void ath9k_tx_cabq(struct ieee80211_hw *hw, struct sk_buff *skb)
-{
-	struct ath_softc *sc = hw->priv;
-	struct ath_common *common = ath9k_hw_common(sc->sc_ah);
-	struct ath_tx_control txctl;
-
-	memset(&txctl, 0, sizeof(struct ath_tx_control));
-	txctl.txq = sc->beacon.cabq;
-
-	ath_dbg(common, XMIT, "transmitting CABQ packet, skb: %p\n", skb);
-
-	if (ath_tx_start(hw, skb, &txctl) != 0) {
-		ath_dbg(common, XMIT, "CABQ TX failed\n");
-		ieee80211_free_txskb(hw, skb);
-	}
-}
-
 static struct ath_buf *ath9k_beacon_generate(struct ieee80211_hw *hw,
 					     struct ieee80211_vif *vif)
 {
@@ -205,10 +189,8 @@ static struct ath_buf *ath9k_beacon_generate(struct ieee80211_hw *hw,
 
 	ath9k_beacon_setup(sc, vif, bf, info->control.rates[0].idx);
 
-	while (skb) {
-		ath9k_tx_cabq(hw, skb);
-		skb = ieee80211_get_buffered_bc(hw, vif);
-	}
+	if (skb)
+		ath_tx_cabq(hw, vif, skb);
 
 	return bf;
 }
@@ -273,7 +255,8 @@ static int ath9k_beacon_choose_slot(struct ath_softc *sc)
 	u64 tsf;
 	int slot;
 
-	if (sc->sc_ah->opmode != NL80211_IFTYPE_AP) {
+	if (sc->sc_ah->opmode != NL80211_IFTYPE_AP &&
+	    sc->sc_ah->opmode != NL80211_IFTYPE_MESH_POINT) {
 		ath_dbg(common, BEACON, "slot 0, tsf: %llu\n",
 			ath9k_hw_gettsf64(sc->sc_ah));
 		return 0;
@@ -765,10 +748,10 @@ void ath9k_set_beacon(struct ath_softc *sc)
 
 	switch (sc->sc_ah->opmode) {
 	case NL80211_IFTYPE_AP:
+	case NL80211_IFTYPE_MESH_POINT:
 		ath9k_beacon_config_ap(sc, cur_conf);
 		break;
 	case NL80211_IFTYPE_ADHOC:
-	case NL80211_IFTYPE_MESH_POINT:
 		ath9k_beacon_config_adhoc(sc, cur_conf);
 		break;
 	case NL80211_IFTYPE_STATION:
diff --git a/drivers/net/wireless/ath/ath9k/calib.c b/drivers/net/wireless/ath/ath9k/calib.c
index 7304e75..5e8219a 100644
--- a/drivers/net/wireless/ath/ath9k/calib.c
+++ b/drivers/net/wireless/ath/ath9k/calib.c
@@ -387,7 +387,6 @@ bool ath9k_hw_getnf(struct ath_hw *ah, struct ath9k_channel *chan)
 
 	if (!caldata) {
 		chan->noisefloor = nf;
-		ah->noise = ath9k_hw_getchan_noise(ah, chan);
 		return false;
 	}
 
diff --git a/drivers/net/wireless/ath/ath9k/debug.c b/drivers/net/wireless/ath/ath9k/debug.c
index b37eb8d..87454f6 100644
--- a/drivers/net/wireless/ath/ath9k/debug.c
+++ b/drivers/net/wireless/ath/ath9k/debug.c
@@ -69,7 +69,7 @@ static ssize_t write_file_debug(struct file *file, const char __user *user_buf,
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &mask))
+	if (kstrtoul(buf, 0, &mask))
 		return -EINVAL;
 
 	common->debug_mask = mask;
@@ -114,7 +114,7 @@ static ssize_t write_file_tx_chainmask(struct file *file, const char __user *use
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &mask))
+	if (kstrtoul(buf, 0, &mask))
 		return -EINVAL;
 
 	ah->txchainmask = mask;
@@ -157,7 +157,7 @@ static ssize_t write_file_rx_chainmask(struct file *file, const char __user *use
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &mask))
+	if (kstrtoul(buf, 0, &mask))
 		return -EINVAL;
 
 	ah->rxchainmask = mask;
@@ -173,25 +173,69 @@ static const struct file_operations fops_rx_chainmask = {
 	.llseek = default_llseek,
 };
 
-static ssize_t read_file_disable_ani(struct file *file, char __user *user_buf,
+static ssize_t read_file_ani(struct file *file, char __user *user_buf,
 			     size_t count, loff_t *ppos)
 {
 	struct ath_softc *sc = file->private_data;
 	struct ath_common *common = ath9k_hw_common(sc->sc_ah);
-	char buf[32];
-	unsigned int len;
+	struct ath_hw *ah = sc->sc_ah;
+	unsigned int len = 0, size = 1024;
+	ssize_t retval = 0;
+	char *buf;
 
-	len = sprintf(buf, "%d\n", common->disable_ani);
-	return simple_read_from_buffer(user_buf, count, ppos, buf, len);
+	buf = kzalloc(size, GFP_KERNEL);
+	if (buf == NULL)
+		return -ENOMEM;
+
+	if (common->disable_ani) {
+		len += snprintf(buf + len, size - len, "%s: %s\n",
+				"ANI", "DISABLED");
+		goto exit;
+	}
+
+	len += snprintf(buf + len, size - len, "%15s: %s\n",
+			"ANI", "ENABLED");
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"ANI RESET", ah->stats.ast_ani_reset);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"SPUR UP", ah->stats.ast_ani_spurup);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"SPUR DOWN", ah->stats.ast_ani_spurup);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"OFDM WS-DET ON", ah->stats.ast_ani_ofdmon);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"OFDM WS-DET OFF", ah->stats.ast_ani_ofdmoff);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"MRC-CCK ON", ah->stats.ast_ani_ccklow);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"MRC-CCK OFF", ah->stats.ast_ani_cckhigh);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"FIR-STEP UP", ah->stats.ast_ani_stepup);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"FIR-STEP DOWN", ah->stats.ast_ani_stepdown);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"INV LISTENTIME", ah->stats.ast_ani_lneg_or_lzero);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"OFDM ERRORS", ah->stats.ast_ani_ofdmerrs);
+	len += snprintf(buf + len, size - len, "%15s: %u\n",
+			"CCK ERRORS", ah->stats.ast_ani_cckerrs);
+exit:
+	if (len > size)
+		len = size;
+
+	retval = simple_read_from_buffer(user_buf, count, ppos, buf, len);
+	kfree(buf);
+
+	return retval;
 }
 
-static ssize_t write_file_disable_ani(struct file *file,
-				      const char __user *user_buf,
-				      size_t count, loff_t *ppos)
+static ssize_t write_file_ani(struct file *file,
+			      const char __user *user_buf,
+			      size_t count, loff_t *ppos)
 {
 	struct ath_softc *sc = file->private_data;
 	struct ath_common *common = ath9k_hw_common(sc->sc_ah);
-	unsigned long disable_ani;
+	unsigned long ani;
 	char buf[32];
 	ssize_t len;
 
@@ -200,12 +244,15 @@ static ssize_t write_file_disable_ani(struct file *file,
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &disable_ani))
+	if (kstrtoul(buf, 0, &ani))
 		return -EINVAL;
 
-	common->disable_ani = !!disable_ani;
+	if (ani < 0 || ani > 1)
+		return -EINVAL;
+
+	common->disable_ani = !ani;
 
-	if (disable_ani) {
+	if (common->disable_ani) {
 		clear_bit(SC_OP_ANI_RUN, &sc->sc_flags);
 		ath_stop_ani(sc);
 	} else {
@@ -215,9 +262,9 @@ static ssize_t write_file_disable_ani(struct file *file,
 	return count;
 }
 
-static const struct file_operations fops_disable_ani = {
-	.read = read_file_disable_ani,
-	.write = write_file_disable_ani,
+static const struct file_operations fops_ani = {
+	.read = read_file_ani,
+	.write = write_file_ani,
 	.open = simple_open,
 	.owner = THIS_MODULE,
 	.llseek = default_llseek,
@@ -253,7 +300,7 @@ static ssize_t write_file_ant_diversity(struct file *file,
 		goto exit;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &antenna_diversity))
+	if (kstrtoul(buf, 0, &antenna_diversity))
 		return -EINVAL;
 
 	common->antenna_diversity = !!antenna_diversity;
@@ -738,8 +785,6 @@ void ath_debug_stat_tx(struct ath_softc *sc, struct ath_buf *bf,
 		       struct ath_tx_status *ts, struct ath_txq *txq,
 		       unsigned int flags)
 {
-#define TX_SAMP_DBG(c) (sc->debug.bb_mac_samp[sc->debug.sampidx].ts\
-			[sc->debug.tsidx].c)
 	int qnum = txq->axq_qnum;
 
 	TX_STAT_INC(qnum, tx_pkts_all);
@@ -771,37 +816,6 @@ void ath_debug_stat_tx(struct ath_softc *sc, struct ath_buf *bf,
 		TX_STAT_INC(qnum, data_underrun);
 	if (ts->ts_flags & ATH9K_TX_DELIM_UNDERRUN)
 		TX_STAT_INC(qnum, delim_underrun);
-
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-	spin_lock(&sc->debug.samp_lock);
-	TX_SAMP_DBG(jiffies) = jiffies;
-	TX_SAMP_DBG(rssi_ctl0) = ts->ts_rssi_ctl0;
-	TX_SAMP_DBG(rssi_ctl1) = ts->ts_rssi_ctl1;
-	TX_SAMP_DBG(rssi_ctl2) = ts->ts_rssi_ctl2;
-	TX_SAMP_DBG(rssi_ext0) = ts->ts_rssi_ext0;
-	TX_SAMP_DBG(rssi_ext1) = ts->ts_rssi_ext1;
-	TX_SAMP_DBG(rssi_ext2) = ts->ts_rssi_ext2;
-	TX_SAMP_DBG(rateindex) = ts->ts_rateindex;
-	TX_SAMP_DBG(isok) = !!(ts->ts_status & ATH9K_TXERR_MASK);
-	TX_SAMP_DBG(rts_fail_cnt) = ts->ts_shortretry;
-	TX_SAMP_DBG(data_fail_cnt) = ts->ts_longretry;
-	TX_SAMP_DBG(rssi) = ts->ts_rssi;
-	TX_SAMP_DBG(tid) = ts->tid;
-	TX_SAMP_DBG(qid) = ts->qid;
-
-	if (ts->ts_flags & ATH9K_TX_BA) {
-		TX_SAMP_DBG(ba_low) = ts->ba_low;
-		TX_SAMP_DBG(ba_high) = ts->ba_high;
-	} else {
-		TX_SAMP_DBG(ba_low) = 0;
-		TX_SAMP_DBG(ba_high) = 0;
-	}
-
-	sc->debug.tsidx = (sc->debug.tsidx + 1) % ATH_DBG_MAX_SAMPLES;
-	spin_unlock(&sc->debug.samp_lock);
-#endif
-
-#undef TX_SAMP_DBG
 }
 
 static const struct file_operations fops_xmit = {
@@ -915,8 +929,6 @@ static ssize_t read_file_recv(struct file *file, char __user *user_buf,
 void ath_debug_stat_rx(struct ath_softc *sc, struct ath_rx_status *rs)
 {
 #define RX_PHY_ERR_INC(c) sc->debug.stats.rxstats.phy_err_stats[c]++
-#define RX_SAMP_DBG(c) (sc->debug.bb_mac_samp[sc->debug.sampidx].rs\
-			[sc->debug.rsidx].c)
 
 	RX_STAT_INC(rx_pkts_all);
 	sc->debug.stats.rxstats.rx_bytes_all += rs->rs_datalen;
@@ -940,27 +952,7 @@ void ath_debug_stat_rx(struct ath_softc *sc, struct ath_rx_status *rs)
 			RX_PHY_ERR_INC(rs->rs_phyerr);
 	}
 
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-	spin_lock(&sc->debug.samp_lock);
-	RX_SAMP_DBG(jiffies) = jiffies;
-	RX_SAMP_DBG(rssi_ctl0) = rs->rs_rssi_ctl0;
-	RX_SAMP_DBG(rssi_ctl1) = rs->rs_rssi_ctl1;
-	RX_SAMP_DBG(rssi_ctl2) = rs->rs_rssi_ctl2;
-	RX_SAMP_DBG(rssi_ext0) = rs->rs_rssi_ext0;
-	RX_SAMP_DBG(rssi_ext1) = rs->rs_rssi_ext1;
-	RX_SAMP_DBG(rssi_ext2) = rs->rs_rssi_ext2;
-	RX_SAMP_DBG(antenna) = rs->rs_antenna;
-	RX_SAMP_DBG(rssi) = rs->rs_rssi;
-	RX_SAMP_DBG(rate) = rs->rs_rate;
-	RX_SAMP_DBG(is_mybeacon) = rs->is_mybeacon;
-
-	sc->debug.rsidx = (sc->debug.rsidx + 1) % ATH_DBG_MAX_SAMPLES;
-	spin_unlock(&sc->debug.samp_lock);
-
-#endif
-
 #undef RX_PHY_ERR_INC
-#undef RX_SAMP_DBG
 }
 
 static const struct file_operations fops_recv = {
@@ -1278,7 +1270,7 @@ static ssize_t write_file_regidx(struct file *file, const char __user *user_buf,
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &regidx))
+	if (kstrtoul(buf, 0, &regidx))
 		return -EINVAL;
 
 	sc->debug.regidx = regidx;
@@ -1323,7 +1315,7 @@ static ssize_t write_file_regval(struct file *file, const char __user *user_buf,
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &regval))
+	if (kstrtoul(buf, 0, &regval))
 		return -EINVAL;
 
 	ath9k_ps_wakeup(sc);
@@ -1485,283 +1477,6 @@ static const struct file_operations fops_modal_eeprom = {
 	.llseek = default_llseek,
 };
 
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-
-void ath9k_debug_samp_bb_mac(struct ath_softc *sc)
-{
-#define ATH_SAMP_DBG(c) (sc->debug.bb_mac_samp[sc->debug.sampidx].c)
-	struct ath_hw *ah = sc->sc_ah;
-	struct ath_common *common = ath9k_hw_common(ah);
-	unsigned long flags;
-	int i;
-
-	ath9k_ps_wakeup(sc);
-
-	spin_lock_bh(&sc->debug.samp_lock);
-
-	spin_lock_irqsave(&common->cc_lock, flags);
-	ath_hw_cycle_counters_update(common);
-
-	ATH_SAMP_DBG(cc.cycles) = common->cc_ani.cycles;
-	ATH_SAMP_DBG(cc.rx_busy) = common->cc_ani.rx_busy;
-	ATH_SAMP_DBG(cc.rx_frame) = common->cc_ani.rx_frame;
-	ATH_SAMP_DBG(cc.tx_frame) = common->cc_ani.tx_frame;
-	spin_unlock_irqrestore(&common->cc_lock, flags);
-
-	ATH_SAMP_DBG(noise) = ah->noise;
-
-	REG_WRITE_D(ah, AR_MACMISC,
-		  ((AR_MACMISC_DMA_OBS_LINE_8 << AR_MACMISC_DMA_OBS_S) |
-		   (AR_MACMISC_MISC_OBS_BUS_1 <<
-		    AR_MACMISC_MISC_OBS_BUS_MSB_S)));
-
-	for (i = 0; i < ATH9K_NUM_DMA_DEBUG_REGS; i++)
-		ATH_SAMP_DBG(dma_dbg_reg_vals[i]) = REG_READ_D(ah,
-				AR_DMADBG_0 + (i * sizeof(u32)));
-
-	ATH_SAMP_DBG(pcu_obs) = REG_READ_D(ah, AR_OBS_BUS_1);
-	ATH_SAMP_DBG(pcu_cr) = REG_READ_D(ah, AR_CR);
-
-	memcpy(ATH_SAMP_DBG(nfCalHist), sc->caldata.nfCalHist,
-			sizeof(ATH_SAMP_DBG(nfCalHist)));
-
-	sc->debug.sampidx = (sc->debug.sampidx + 1) % ATH_DBG_MAX_SAMPLES;
-	spin_unlock_bh(&sc->debug.samp_lock);
-	ath9k_ps_restore(sc);
-
-#undef ATH_SAMP_DBG
-}
-
-static int open_file_bb_mac_samps(struct inode *inode, struct file *file)
-{
-#define ATH_SAMP_DBG(c) bb_mac_samp[sampidx].c
-	struct ath_softc *sc = inode->i_private;
-	struct ath_hw *ah = sc->sc_ah;
-	struct ath_common *common = ath9k_hw_common(ah);
-	struct ieee80211_conf *conf = &common->hw->conf;
-	struct ath_dbg_bb_mac_samp *bb_mac_samp;
-	struct ath9k_nfcal_hist *h;
-	int i, j, qcuOffset = 0, dcuOffset = 0;
-	u32 *qcuBase, *dcuBase, size = 30000, len = 0;
-	u32 sampidx = 0;
-	u8 *buf;
-	u8 chainmask = (ah->rxchainmask << 3) | ah->rxchainmask;
-	u8 nread;
-
-	if (test_bit(SC_OP_INVALID, &sc->sc_flags))
-		return -EAGAIN;
-
-	buf = vmalloc(size);
-	if (!buf)
-		return -ENOMEM;
-	bb_mac_samp = vmalloc(sizeof(*bb_mac_samp) * ATH_DBG_MAX_SAMPLES);
-	if (!bb_mac_samp) {
-		vfree(buf);
-		return -ENOMEM;
-	}
-	/* Account the current state too */
-	ath9k_debug_samp_bb_mac(sc);
-
-	spin_lock_bh(&sc->debug.samp_lock);
-	memcpy(bb_mac_samp, sc->debug.bb_mac_samp,
-			sizeof(*bb_mac_samp) * ATH_DBG_MAX_SAMPLES);
-	len += snprintf(buf + len, size - len,
-			"Current Sample Index: %d\n", sc->debug.sampidx);
-	spin_unlock_bh(&sc->debug.samp_lock);
-
-	len += snprintf(buf + len, size - len,
-			"Raw DMA Debug Dump:\n");
-	len += snprintf(buf + len, size - len, "Sample |\t");
-	for (i = 0; i < ATH9K_NUM_DMA_DEBUG_REGS; i++)
-		len += snprintf(buf + len, size - len, " DMA Reg%d |\t", i);
-	len += snprintf(buf + len, size - len, "\n");
-
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		len += snprintf(buf + len, size - len, "%d\t", sampidx);
-
-		for (i = 0; i < ATH9K_NUM_DMA_DEBUG_REGS; i++)
-			len += snprintf(buf + len, size - len, " %08x\t",
-					ATH_SAMP_DBG(dma_dbg_reg_vals[i]));
-		len += snprintf(buf + len, size - len, "\n");
-	}
-	len += snprintf(buf + len, size - len, "\n");
-
-	len += snprintf(buf + len, size - len,
-			"Sample Num QCU: chain_st fsp_ok fsp_st DCU: chain_st\n");
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		qcuBase = &ATH_SAMP_DBG(dma_dbg_reg_vals[0]);
-		dcuBase = &ATH_SAMP_DBG(dma_dbg_reg_vals[4]);
-
-		for (i = 0; i < ATH9K_NUM_QUEUES; i++,
-				qcuOffset += 4, dcuOffset += 5) {
-			if (i == 8) {
-				qcuOffset = 0;
-				qcuBase++;
-			}
-
-			if (i == 6) {
-				dcuOffset = 0;
-				dcuBase++;
-			}
-			if (!sc->debug.stats.txstats[i].queued)
-				continue;
-
-			len += snprintf(buf + len, size - len,
-				"%4d %7d    %2x      %1x     %2x         %2x\n",
-				sampidx, i,
-				(*qcuBase & (0x7 << qcuOffset)) >> qcuOffset,
-				(*qcuBase & (0x8 << qcuOffset)) >>
-				(qcuOffset + 3),
-				ATH_SAMP_DBG(dma_dbg_reg_vals[2]) &
-				(0x7 << (i * 3)) >> (i * 3),
-				(*dcuBase & (0x1f << dcuOffset)) >> dcuOffset);
-		}
-		len += snprintf(buf + len, size - len, "\n");
-	}
-	len += snprintf(buf + len, size - len,
-			"samp qcu_sh qcu_fh qcu_comp dcu_comp dcu_arb dcu_fp "
-			"ch_idle_dur ch_idle_dur_val txfifo_val0 txfifo_val1 "
-			"txfifo_dcu0 txfifo_dcu1 pcu_obs AR_CR\n");
-
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		qcuBase = &ATH_SAMP_DBG(dma_dbg_reg_vals[0]);
-		dcuBase = &ATH_SAMP_DBG(dma_dbg_reg_vals[4]);
-
-		len += snprintf(buf + len, size - len, "%4d %5x %5x ", sampidx,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[3]) & 0x003c0000) >> 18,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[3]) & 0x03c00000) >> 22);
-		len += snprintf(buf + len, size - len, "%7x %8x ",
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[3]) & 0x1c000000) >> 26,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x3));
-		len += snprintf(buf + len, size - len, "%7x %7x ",
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[5]) & 0x06000000) >> 25,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[5]) & 0x38000000) >> 27);
-		len += snprintf(buf + len, size - len, "%7d %12d ",
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x000003fc) >> 2,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x00000400) >> 10);
-		len += snprintf(buf + len, size - len, "%12d %12d ",
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x00000800) >> 11,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x00001000) >> 12);
-		len += snprintf(buf + len, size - len, "%12d %12d ",
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x0001e000) >> 13,
-			(ATH_SAMP_DBG(dma_dbg_reg_vals[6]) & 0x001e0000) >> 17);
-		len += snprintf(buf + len, size - len, "0x%07x 0x%07x\n",
-				ATH_SAMP_DBG(pcu_obs), ATH_SAMP_DBG(pcu_cr));
-	}
-
-	len += snprintf(buf + len, size - len,
-			"Sample ChNoise Chain privNF #Reading Readings\n");
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		h = ATH_SAMP_DBG(nfCalHist);
-		if (!ATH_SAMP_DBG(noise))
-			continue;
-
-		for (i = 0; i < NUM_NF_READINGS; i++) {
-			if (!(chainmask & (1 << i)) ||
-			    ((i >= AR5416_MAX_CHAINS) && !conf_is_ht40(conf)))
-				continue;
-
-			nread = AR_PHY_CCA_FILTERWINDOW_LENGTH -
-				h[i].invalidNFcount;
-			len += snprintf(buf + len, size - len,
-					"%4d %5d %4d\t   %d\t %d\t",
-					sampidx, ATH_SAMP_DBG(noise),
-					i, h[i].privNF, nread);
-			for (j = 0; j < nread; j++)
-				len += snprintf(buf + len, size - len,
-					" %d", h[i].nfCalBuffer[j]);
-			len += snprintf(buf + len, size - len, "\n");
-		}
-	}
-	len += snprintf(buf + len, size - len, "\nCycle counters:\n"
-			"Sample Total    Rxbusy   Rxframes Txframes\n");
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		if (!ATH_SAMP_DBG(cc.cycles))
-			continue;
-		len += snprintf(buf + len, size - len,
-				"%4d %08x %08x %08x %08x\n",
-				sampidx, ATH_SAMP_DBG(cc.cycles),
-				ATH_SAMP_DBG(cc.rx_busy),
-				ATH_SAMP_DBG(cc.rx_frame),
-				ATH_SAMP_DBG(cc.tx_frame));
-	}
-
-	len += snprintf(buf + len, size - len, "Tx status Dump :\n");
-	len += snprintf(buf + len, size - len,
-			"Sample rssi:- ctl0 ctl1 ctl2 ext0 ext1 ext2 comb "
-			"isok rts_fail data_fail rate tid qid "
-					"ba_low  ba_high tx_before(ms)\n");
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		for (i = 0; i < ATH_DBG_MAX_SAMPLES; i++) {
-			if (!ATH_SAMP_DBG(ts[i].jiffies))
-				continue;
-			len += snprintf(buf + len, size - len, "%-14d"
-				"%-4d %-4d %-4d %-4d %-4d %-4d %-4d %-4d %-8d "
-				"%-9d %-4d %-3d %-3d %08x %08x %-11d\n",
-				sampidx,
-				ATH_SAMP_DBG(ts[i].rssi_ctl0),
-				ATH_SAMP_DBG(ts[i].rssi_ctl1),
-				ATH_SAMP_DBG(ts[i].rssi_ctl2),
-				ATH_SAMP_DBG(ts[i].rssi_ext0),
-				ATH_SAMP_DBG(ts[i].rssi_ext1),
-				ATH_SAMP_DBG(ts[i].rssi_ext2),
-				ATH_SAMP_DBG(ts[i].rssi),
-				ATH_SAMP_DBG(ts[i].isok),
-				ATH_SAMP_DBG(ts[i].rts_fail_cnt),
-				ATH_SAMP_DBG(ts[i].data_fail_cnt),
-				ATH_SAMP_DBG(ts[i].rateindex),
-				ATH_SAMP_DBG(ts[i].tid),
-				ATH_SAMP_DBG(ts[i].qid),
-				ATH_SAMP_DBG(ts[i].ba_low),
-				ATH_SAMP_DBG(ts[i].ba_high),
-				jiffies_to_msecs(jiffies -
-					ATH_SAMP_DBG(ts[i].jiffies)));
-		}
-	}
-
-	len += snprintf(buf + len, size - len, "Rx status Dump :\n");
-	len += snprintf(buf + len, size - len, "Sample rssi:- ctl0 ctl1 ctl2 "
-			"ext0 ext1 ext2 comb beacon ant rate rx_before(ms)\n");
-	for (sampidx = 0; sampidx < ATH_DBG_MAX_SAMPLES; sampidx++) {
-		for (i = 0; i < ATH_DBG_MAX_SAMPLES; i++) {
-			if (!ATH_SAMP_DBG(rs[i].jiffies))
-				continue;
-			len += snprintf(buf + len, size - len, "%-14d"
-				"%-4d %-4d %-4d %-4d %-4d %-4d %-4d %-9s %-2d %02x %-13d\n",
-				sampidx,
-				ATH_SAMP_DBG(rs[i].rssi_ctl0),
-				ATH_SAMP_DBG(rs[i].rssi_ctl1),
-				ATH_SAMP_DBG(rs[i].rssi_ctl2),
-				ATH_SAMP_DBG(rs[i].rssi_ext0),
-				ATH_SAMP_DBG(rs[i].rssi_ext1),
-				ATH_SAMP_DBG(rs[i].rssi_ext2),
-				ATH_SAMP_DBG(rs[i].rssi),
-				ATH_SAMP_DBG(rs[i].is_mybeacon) ?
-				"True" : "False",
-				ATH_SAMP_DBG(rs[i].antenna),
-				ATH_SAMP_DBG(rs[i].rate),
-				jiffies_to_msecs(jiffies -
-					ATH_SAMP_DBG(rs[i].jiffies)));
-		}
-	}
-
-	vfree(bb_mac_samp);
-	file->private_data = buf;
-
-	return 0;
-#undef ATH_SAMP_DBG
-}
-
-static const struct file_operations fops_samps = {
-	.open = open_file_bb_mac_samps,
-	.read = ath9k_debugfs_read_buf,
-	.release = ath9k_debugfs_release_buf,
-	.owner = THIS_MODULE,
-	.llseek = default_llseek,
-};
-
-#endif
-
 #ifdef CONFIG_ATH9K_BTCOEX_SUPPORT
 static ssize_t read_file_btcoex(struct file *file, char __user *user_buf,
 				size_t count, loff_t *ppos)
@@ -2059,8 +1774,8 @@ int ath9k_init_debug(struct ath_hw *ah)
 			    sc->debug.debugfs_phy, sc, &fops_rx_chainmask);
 	debugfs_create_file("tx_chainmask", S_IRUSR | S_IWUSR,
 			    sc->debug.debugfs_phy, sc, &fops_tx_chainmask);
-	debugfs_create_file("disable_ani", S_IRUSR | S_IWUSR,
-			    sc->debug.debugfs_phy, sc, &fops_disable_ani);
+	debugfs_create_file("ani", S_IRUSR | S_IWUSR,
+			    sc->debug.debugfs_phy, sc, &fops_ani);
 	debugfs_create_bool("paprd", S_IRUSR | S_IWUSR, sc->debug.debugfs_phy,
 			    &sc->sc_ah->config.enable_paprd);
 	debugfs_create_file("regidx", S_IRUSR | S_IWUSR, sc->debug.debugfs_phy,
@@ -2095,11 +1810,6 @@ int ath9k_init_debug(struct ath_hw *ah)
 	debugfs_create_file("spectral_fft_period", S_IRUSR | S_IWUSR,
 			    sc->debug.debugfs_phy, sc,
 			    &fops_spectral_fft_period);
-
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-	debugfs_create_file("samples", S_IRUSR, sc->debug.debugfs_phy, sc,
-			    &fops_samps);
-#endif
 	debugfs_create_u32("gpio_mask", S_IRUSR | S_IWUSR,
 			   sc->debug.debugfs_phy, &sc->sc_ah->gpio_mask);
 	debugfs_create_u32("gpio_val", S_IRUSR | S_IWUSR,
diff --git a/drivers/net/wireless/ath/ath9k/debug.h b/drivers/net/wireless/ath/ath9k/debug.h
index 9d49aab..fc67919 100644
--- a/drivers/net/wireless/ath/ath9k/debug.h
+++ b/drivers/net/wireless/ath/ath9k/debug.h
@@ -251,56 +251,10 @@ struct ath_stats {
 	u32 reset[__RESET_TYPE_MAX];
 };
 
-#define ATH_DBG_MAX_SAMPLES	10
-struct ath_dbg_bb_mac_samp {
-	u32 dma_dbg_reg_vals[ATH9K_NUM_DMA_DEBUG_REGS];
-	u32 pcu_obs, pcu_cr, noise;
-	struct {
-		u64 jiffies;
-		int8_t rssi_ctl0;
-		int8_t rssi_ctl1;
-		int8_t rssi_ctl2;
-		int8_t rssi_ext0;
-		int8_t rssi_ext1;
-		int8_t rssi_ext2;
-		int8_t rssi;
-		bool isok;
-		u8 rts_fail_cnt;
-		u8 data_fail_cnt;
-		u8 rateindex;
-		u8 qid;
-		u8 tid;
-		u32 ba_low;
-		u32 ba_high;
-	} ts[ATH_DBG_MAX_SAMPLES];
-	struct {
-		u64 jiffies;
-		int8_t rssi_ctl0;
-		int8_t rssi_ctl1;
-		int8_t rssi_ctl2;
-		int8_t rssi_ext0;
-		int8_t rssi_ext1;
-		int8_t rssi_ext2;
-		int8_t rssi;
-		bool is_mybeacon;
-		u8 antenna;
-		u8 rate;
-	} rs[ATH_DBG_MAX_SAMPLES];
-	struct ath_cycle_counters cc;
-	struct ath9k_nfcal_hist nfCalHist[NUM_NF_READINGS];
-};
-
 struct ath9k_debug {
 	struct dentry *debugfs_phy;
 	u32 regidx;
 	struct ath_stats stats;
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-	spinlock_t samp_lock;
-	struct ath_dbg_bb_mac_samp bb_mac_samp[ATH_DBG_MAX_SAMPLES];
-	u8 sampidx;
-	u8 tsidx;
-	u8 rsidx;
-#endif
 };
 
 int ath9k_init_debug(struct ath_hw *ah);
@@ -364,17 +318,4 @@ static inline void ath_debug_stat_rx(struct ath_softc *sc,
 
 #endif /* CONFIG_ATH9K_DEBUGFS */
 
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-
-void ath9k_debug_samp_bb_mac(struct ath_softc *sc);
-
-#else
-
-static inline void ath9k_debug_samp_bb_mac(struct ath_softc *sc)
-{
-}
-
-#endif
-
-
 #endif /* DEBUG_H */
diff --git a/drivers/net/wireless/ath/ath9k/dfs_debug.c b/drivers/net/wireless/ath/ath9k/dfs_debug.c
index b7611b7..3c6e413 100644
--- a/drivers/net/wireless/ath/ath9k/dfs_debug.c
+++ b/drivers/net/wireless/ath/ath9k/dfs_debug.c
@@ -96,7 +96,7 @@ static ssize_t write_file_dfs(struct file *file, const char __user *user_buf,
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &val))
+	if (kstrtoul(buf, 0, &val))
 		return -EINVAL;
 
 	if (val == DFS_STATS_RESET_MAGIC)
diff --git a/drivers/net/wireless/ath/ath9k/hif_usb.c b/drivers/net/wireless/ath/ath9k/hif_usb.c
index f5dda84..9e582e1 100644
--- a/drivers/net/wireless/ath/ath9k/hif_usb.c
+++ b/drivers/net/wireless/ath/ath9k/hif_usb.c
@@ -234,10 +234,15 @@ static inline void ath9k_skb_queue_complete(struct hif_device_usb *hif_dev,
 	struct sk_buff *skb;
 
 	while ((skb = __skb_dequeue(queue)) != NULL) {
+#ifdef CONFIG_ATH9K_HTC_DEBUGFS
+		int ln = skb->len;
+#endif
 		ath9k_htc_txcompletion_cb(hif_dev->htc_handle,
 					  skb, txok);
-		if (txok)
+		if (txok) {
 			TX_STAT_INC(skb_success);
+			TX_STAT_ADD(skb_success_bytes, ln);
+		}
 		else
 			TX_STAT_INC(skb_failed);
 	}
@@ -620,6 +625,7 @@ static void ath9k_hif_usb_rx_stream(struct hif_device_usb *hif_dev,
 
 err:
 	for (i = 0; i < pool_index; i++) {
+		RX_STAT_ADD(skb_completed_bytes, skb_pool[i]->len);
 		ath9k_htc_rx_msg(hif_dev->htc_handle, skb_pool[i],
 				 skb_pool[i]->len, USB_WLAN_RX_PIPE);
 		RX_STAT_INC(skb_completed);
diff --git a/drivers/net/wireless/ath/ath9k/htc.h b/drivers/net/wireless/ath/ath9k/htc.h
index d3b099d..055d7c2 100644
--- a/drivers/net/wireless/ath/ath9k/htc.h
+++ b/drivers/net/wireless/ath/ath9k/htc.h
@@ -142,6 +142,7 @@ struct ath9k_htc_target_aggr {
 #define WLAN_RC_40_FLAG  0x02
 #define WLAN_RC_SGI_FLAG 0x04
 #define WLAN_RC_HT_FLAG  0x08
+#define ATH_RC_TX_STBC_FLAG 0x20
 
 struct ath9k_htc_rateset {
 	u8 rs_nrates;
@@ -208,6 +209,9 @@ struct ath9k_htc_target_rx_stats {
 		case NL80211_IFTYPE_AP:		\
 			_priv->num_ap_vif++;	\
 			break;			\
+		case NL80211_IFTYPE_MESH_POINT:	\
+			_priv->num_mbss_vif++;	\
+			break;			\
 		default:			\
 			break;			\
 		}				\
@@ -224,6 +228,9 @@ struct ath9k_htc_target_rx_stats {
 		case NL80211_IFTYPE_AP:		\
 			_priv->num_ap_vif--;	\
 			break;			\
+		case NL80211_IFTYPE_MESH_POINT:	\
+			_priv->num_mbss_vif--;	\
+			break;			\
 		default:			\
 			break;			\
 		}				\
@@ -317,7 +324,9 @@ static inline struct ath9k_htc_tx_ctl *HTC_SKB_CB(struct sk_buff *skb)
 #ifdef CONFIG_ATH9K_HTC_DEBUGFS
 
 #define TX_STAT_INC(c) (hif_dev->htc_handle->drv_priv->debug.tx_stats.c++)
+#define TX_STAT_ADD(c, a) (hif_dev->htc_handle->drv_priv->debug.tx_stats.c += a)
 #define RX_STAT_INC(c) (hif_dev->htc_handle->drv_priv->debug.rx_stats.c++)
+#define RX_STAT_ADD(c, a) (hif_dev->htc_handle->drv_priv->debug.rx_stats.c += a)
 #define CAB_STAT_INC   priv->debug.tx_stats.cab_queued++
 
 #define TX_QSTAT_INC(q) (priv->debug.tx_stats.queue_stats[q]++)
@@ -330,6 +339,7 @@ struct ath_tx_stats {
 	u32 buf_completed;
 	u32 skb_queued;
 	u32 skb_success;
+	u32 skb_success_bytes;
 	u32 skb_failed;
 	u32 cab_queued;
 	u32 queue_stats[IEEE80211_NUM_ACS];
@@ -338,6 +348,7 @@ struct ath_tx_stats {
 struct ath_rx_stats {
 	u32 skb_allocated;
 	u32 skb_completed;
+	u32 skb_completed_bytes;
 	u32 skb_dropped;
 	u32 err_crc;
 	u32 err_decrypt_crc;
@@ -355,10 +366,20 @@ struct ath9k_debug {
 	struct ath_rx_stats rx_stats;
 };
 
+void ath9k_htc_get_et_strings(struct ieee80211_hw *hw,
+			      struct ieee80211_vif *vif,
+			      u32 sset, u8 *data);
+int ath9k_htc_get_et_sset_count(struct ieee80211_hw *hw,
+				struct ieee80211_vif *vif, int sset);
+void ath9k_htc_get_et_stats(struct ieee80211_hw *hw,
+			    struct ieee80211_vif *vif,
+			    struct ethtool_stats *stats, u64 *data);
 #else
 
 #define TX_STAT_INC(c) do { } while (0)
+#define TX_STAT_ADD(c, a) do { } while (0)
 #define RX_STAT_INC(c) do { } while (0)
+#define RX_STAT_ADD(c, a) do { } while (0)
 #define CAB_STAT_INC   do { } while (0)
 
 #define TX_QSTAT_INC(c) do { } while (0)
@@ -450,6 +471,7 @@ struct ath9k_htc_priv {
 	u8 sta_slot;
 	u8 vif_sta_pos[ATH9K_HTC_MAX_VIF];
 	u8 num_ibss_vif;
+	u8 num_mbss_vif;
 	u8 num_sta_vif;
 	u8 num_sta_assoc_vif;
 	u8 num_ap_vif;
@@ -575,6 +597,8 @@ bool ath9k_htc_setpower(struct ath9k_htc_priv *priv,
 void ath9k_start_rfkill_poll(struct ath9k_htc_priv *priv);
 void ath9k_htc_rfkill_poll_state(struct ieee80211_hw *hw);
 
+struct base_eep_header *ath9k_htc_get_eeprom_base(struct ath9k_htc_priv *priv);
+
 #ifdef CONFIG_MAC80211_LEDS
 void ath9k_init_leds(struct ath9k_htc_priv *priv);
 void ath9k_deinit_leds(struct ath9k_htc_priv *priv);
diff --git a/drivers/net/wireless/ath/ath9k/htc_drv_beacon.c b/drivers/net/wireless/ath/ath9k/htc_drv_beacon.c
index f13f458..e0c03bd 100644
--- a/drivers/net/wireless/ath/ath9k/htc_drv_beacon.c
+++ b/drivers/net/wireless/ath/ath9k/htc_drv_beacon.c
@@ -28,7 +28,8 @@ void ath9k_htc_beaconq_config(struct ath9k_htc_priv *priv)
 
 	ath9k_hw_get_txq_props(ah, priv->beaconq, &qi);
 
-	if (priv->ah->opmode == NL80211_IFTYPE_AP) {
+	if (priv->ah->opmode == NL80211_IFTYPE_AP ||
+	    priv->ah->opmode == NL80211_IFTYPE_MESH_POINT) {
 		qi.tqi_aifs = 1;
 		qi.tqi_cwmin = 0;
 		qi.tqi_cwmax = 0;
@@ -628,6 +629,7 @@ void ath9k_htc_beacon_config(struct ath9k_htc_priv *priv,
 	case NL80211_IFTYPE_ADHOC:
 		ath9k_htc_beacon_config_adhoc(priv, cur_conf);
 		break;
+	case NL80211_IFTYPE_MESH_POINT:
 	case NL80211_IFTYPE_AP:
 		ath9k_htc_beacon_config_ap(priv, cur_conf);
 		break;
@@ -649,6 +651,7 @@ void ath9k_htc_beacon_reconfig(struct ath9k_htc_priv *priv)
 	case NL80211_IFTYPE_ADHOC:
 		ath9k_htc_beacon_config_adhoc(priv, cur_conf);
 		break;
+	case NL80211_IFTYPE_MESH_POINT:
 	case NL80211_IFTYPE_AP:
 		ath9k_htc_beacon_config_ap(priv, cur_conf);
 		break;
diff --git a/drivers/net/wireless/ath/ath9k/htc_drv_debug.c b/drivers/net/wireless/ath/ath9k/htc_drv_debug.c
index 87110de5..c1b45e2 100644
--- a/drivers/net/wireless/ath/ath9k/htc_drv_debug.c
+++ b/drivers/net/wireless/ath/ath9k/htc_drv_debug.c
@@ -471,7 +471,7 @@ static ssize_t write_file_debug(struct file *file, const char __user *user_buf,
 		return -EFAULT;
 
 	buf[len] = '\0';
-	if (strict_strtoul(buf, 0, &mask))
+	if (kstrtoul(buf, 0, &mask))
 		return -EINVAL;
 
 	common->debug_mask = mask;
@@ -496,21 +496,7 @@ static ssize_t read_file_base_eeprom(struct file *file, char __user *user_buf,
 	ssize_t retval = 0;
 	char *buf;
 
-	/*
-	 * This can be done since all the 3 EEPROM families have the
-	 * same base header upto a certain point, and we are interested in
-	 * the data only upto that point.
-	 */
-
-	if (AR_SREV_9271(priv->ah))
-		pBase = (struct base_eep_header *)
-			&priv->ah->eeprom.map4k.baseEepHeader;
-	else if (priv->ah->hw_version.usbdev == AR9280_USB)
-		pBase = (struct base_eep_header *)
-			&priv->ah->eeprom.def.baseEepHeader;
-	else if (priv->ah->hw_version.usbdev == AR9287_USB)
-		pBase = (struct base_eep_header *)
-			&priv->ah->eeprom.map9287.baseEepHeader;
+	pBase = ath9k_htc_get_eeprom_base(priv);
 
 	if (pBase == NULL) {
 		ath_err(common, "Unknown EEPROM type\n");
@@ -916,6 +902,87 @@ static const struct file_operations fops_modal_eeprom = {
 	.llseek = default_llseek,
 };
 
+
+/* Ethtool support for get-stats */
+#define AMKSTR(nm) #nm "_BE", #nm "_BK", #nm "_VI", #nm "_VO"
+static const char ath9k_htc_gstrings_stats[][ETH_GSTRING_LEN] = {
+	"tx_pkts_nic",
+	"tx_bytes_nic",
+	"rx_pkts_nic",
+	"rx_bytes_nic",
+
+	AMKSTR(d_tx_pkts),
+
+	"d_rx_crc_err",
+	"d_rx_decrypt_crc_err",
+	"d_rx_phy_err",
+	"d_rx_mic_err",
+	"d_rx_pre_delim_crc_err",
+	"d_rx_post_delim_crc_err",
+	"d_rx_decrypt_busy_err",
+
+	"d_rx_phyerr_radar",
+	"d_rx_phyerr_ofdm_timing",
+	"d_rx_phyerr_cck_timing",
+
+};
+#define ATH9K_HTC_SSTATS_LEN ARRAY_SIZE(ath9k_htc_gstrings_stats)
+
+void ath9k_htc_get_et_strings(struct ieee80211_hw *hw,
+			      struct ieee80211_vif *vif,
+			      u32 sset, u8 *data)
+{
+	if (sset == ETH_SS_STATS)
+		memcpy(data, *ath9k_htc_gstrings_stats,
+		       sizeof(ath9k_htc_gstrings_stats));
+}
+
+int ath9k_htc_get_et_sset_count(struct ieee80211_hw *hw,
+				struct ieee80211_vif *vif, int sset)
+{
+	if (sset == ETH_SS_STATS)
+		return ATH9K_HTC_SSTATS_LEN;
+	return 0;
+}
+
+#define STXBASE priv->debug.tx_stats
+#define SRXBASE priv->debug.rx_stats
+#define ASTXQ(a)					\
+	data[i++] = STXBASE.a[IEEE80211_AC_BE];		\
+	data[i++] = STXBASE.a[IEEE80211_AC_BK];		\
+	data[i++] = STXBASE.a[IEEE80211_AC_VI];		\
+	data[i++] = STXBASE.a[IEEE80211_AC_VO]
+
+void ath9k_htc_get_et_stats(struct ieee80211_hw *hw,
+			    struct ieee80211_vif *vif,
+			    struct ethtool_stats *stats, u64 *data)
+{
+	struct ath9k_htc_priv *priv = hw->priv;
+	int i = 0;
+
+	data[i++] = STXBASE.skb_success;
+	data[i++] = STXBASE.skb_success_bytes;
+	data[i++] = SRXBASE.skb_completed;
+	data[i++] = SRXBASE.skb_completed_bytes;
+
+	ASTXQ(queue_stats);
+
+	data[i++] = SRXBASE.err_crc;
+	data[i++] = SRXBASE.err_decrypt_crc;
+	data[i++] = SRXBASE.err_phy;
+	data[i++] = SRXBASE.err_mic;
+	data[i++] = SRXBASE.err_pre_delim;
+	data[i++] = SRXBASE.err_post_delim;
+	data[i++] = SRXBASE.err_decrypt_busy;
+
+	data[i++] = SRXBASE.err_phy_stats[ATH9K_PHYERR_RADAR];
+	data[i++] = SRXBASE.err_phy_stats[ATH9K_PHYERR_OFDM_TIMING];
+	data[i++] = SRXBASE.err_phy_stats[ATH9K_PHYERR_CCK_TIMING];
+
+	WARN_ON(i != ATH9K_HTC_SSTATS_LEN);
+}
+
+
 int ath9k_htc_init_debug(struct ath_hw *ah)
 {
 	struct ath_common *common = ath9k_hw_common(ah);
diff --git a/drivers/net/wireless/ath/ath9k/htc_drv_init.c b/drivers/net/wireless/ath/ath9k/htc_drv_init.c
index a47f5e0..71a183f 100644
--- a/drivers/net/wireless/ath/ath9k/htc_drv_init.c
+++ b/drivers/net/wireless/ath/ath9k/htc_drv_init.c
@@ -517,6 +517,9 @@ static void setup_ht_cap(struct ath9k_htc_priv *priv,
 	ath_dbg(common, CONFIG, "TX streams %d, RX streams: %d\n",
 		tx_streams, rx_streams);
 
+	if (tx_streams >= 2)
+		ht_info->cap |= IEEE80211_HT_CAP_TX_STBC;
+
 	if (tx_streams != rx_streams) {
 		ht_info->mcs.tx_params |= IEEE80211_HT_MCS_TX_RX_DIFF;
 		ht_info->mcs.tx_params |= ((tx_streams - 1) <<
@@ -698,6 +701,9 @@ static const struct ieee80211_iface_limit if_limits[] = {
 	{ .max = 2,	.types = BIT(NL80211_IFTYPE_STATION) |
 				 BIT(NL80211_IFTYPE_P2P_CLIENT) },
 	{ .max = 2,	.types = BIT(NL80211_IFTYPE_AP) |
+#ifdef CONFIG_MAC80211_MESH
+				 BIT(NL80211_IFTYPE_MESH_POINT) |
+#endif
 				 BIT(NL80211_IFTYPE_P2P_GO) },
 };
 
@@ -712,6 +718,7 @@ static void ath9k_set_hw_capab(struct ath9k_htc_priv *priv,
 			       struct ieee80211_hw *hw)
 {
 	struct ath_common *common = ath9k_hw_common(priv->ah);
+	struct base_eep_header *pBase;
 
 	hw->flags = IEEE80211_HW_SIGNAL_DBM |
 		IEEE80211_HW_AMPDU_AGGREGATION |
@@ -721,6 +728,7 @@ static void ath9k_set_hw_capab(struct ath9k_htc_priv *priv,
 		IEEE80211_HW_SUPPORTS_PS |
 		IEEE80211_HW_PS_NULLFUNC_STACK |
 		IEEE80211_HW_REPORTS_TX_ACK_STATUS |
+		IEEE80211_HW_MFP_CAPABLE |
 		IEEE80211_HW_HOST_BROADCAST_PS_BUFFERING;
 
 	hw->wiphy->interface_modes =
@@ -728,7 +736,8 @@ static void ath9k_set_hw_capab(struct ath9k_htc_priv *priv,
 		BIT(NL80211_IFTYPE_ADHOC) |
 		BIT(NL80211_IFTYPE_AP) |
 		BIT(NL80211_IFTYPE_P2P_GO) |
-		BIT(NL80211_IFTYPE_P2P_CLIENT);
+		BIT(NL80211_IFTYPE_P2P_CLIENT) |
+		BIT(NL80211_IFTYPE_MESH_POINT);
 
 	hw->wiphy->iface_combinations = &if_comb;
 	hw->wiphy->n_iface_combinations = 1;
@@ -765,6 +774,12 @@ static void ath9k_set_hw_capab(struct ath9k_htc_priv *priv,
 				     &priv->sbands[IEEE80211_BAND_5GHZ].ht_cap);
 	}
 
+	pBase = ath9k_htc_get_eeprom_base(priv);
+	if (pBase) {
+		hw->wiphy->available_antennas_rx = pBase->rxMask;
+		hw->wiphy->available_antennas_tx = pBase->txMask;
+	}
+
 	SET_IEEE80211_PERM_ADDR(hw, common->macaddr);
 }
 
diff --git a/drivers/net/wireless/ath/ath9k/htc_drv_main.c b/drivers/net/wireless/ath/ath9k/htc_drv_main.c
index 62f1b76..5c1bec1 100644
--- a/drivers/net/wireless/ath/ath9k/htc_drv_main.c
+++ b/drivers/net/wireless/ath/ath9k/htc_drv_main.c
@@ -113,7 +113,9 @@ static void ath9k_htc_vif_iter(void *data, u8 *mac, struct ieee80211_vif *vif)
 	struct ath9k_htc_priv *priv = data;
 	struct ieee80211_bss_conf *bss_conf = &vif->bss_conf;
 
-	if ((vif->type == NL80211_IFTYPE_AP) && bss_conf->enable_beacon)
+	if ((vif->type == NL80211_IFTYPE_AP ||
+	     vif->type == NL80211_IFTYPE_MESH_POINT) &&
+	    bss_conf->enable_beacon)
 		priv->reconfig_beacon = true;
 
 	if (bss_conf->assoc) {
@@ -180,6 +182,8 @@ static void ath9k_htc_set_opmode(struct ath9k_htc_priv *priv)
 		priv->ah->opmode = NL80211_IFTYPE_ADHOC;
 	else if (priv->num_ap_vif)
 		priv->ah->opmode = NL80211_IFTYPE_AP;
+	else if (priv->num_mbss_vif)
+		priv->ah->opmode = NL80211_IFTYPE_MESH_POINT;
 	else
 		priv->ah->opmode = NL80211_IFTYPE_STATION;
 
@@ -623,6 +627,8 @@ static void ath9k_htc_setup_rate(struct ath9k_htc_priv *priv,
 		trate->rates.ht_rates.rs_nrates = j;
 
 		caps = WLAN_RC_HT_FLAG;
+		if (sta->ht_cap.cap & IEEE80211_HT_CAP_RX_STBC)
+			caps |= ATH_RC_TX_STBC_FLAG;
 		if (sta->ht_cap.mcs.rx_mask[1])
 			caps |= WLAN_RC_DS_FLAG;
 		if ((sta->ht_cap.cap & IEEE80211_HT_CAP_SUP_WIDTH_20_40) &&
@@ -810,8 +816,7 @@ void ath9k_htc_ani_work(struct work_struct *work)
 	}
 
 	/* Verify whether we must check ANI */
-	if (ah->config.enable_ani &&
-	    (timestamp - common->ani.checkani_timer) >= ATH_ANI_POLLINTERVAL) {
+	if ((timestamp - common->ani.checkani_timer) >= ATH_ANI_POLLINTERVAL) {
 		aniflag = true;
 		common->ani.checkani_timer = timestamp;
 	}
@@ -841,8 +846,7 @@ set_timer:
 	* short calibration and long calibration.
 	*/
 	cal_interval = ATH_LONG_CALINTERVAL;
-	if (ah->config.enable_ani)
-		cal_interval = min(cal_interval, (u32)ATH_ANI_POLLINTERVAL);
+	cal_interval = min(cal_interval, (u32)ATH_ANI_POLLINTERVAL);
 	if (!common->ani.caldone)
 		cal_interval = min(cal_interval, (u32)short_cal_interval);
 
@@ -1052,6 +1056,9 @@ static int ath9k_htc_add_interface(struct ieee80211_hw *hw,
 	case NL80211_IFTYPE_AP:
 		hvif.opmode = HTC_M_HOSTAP;
 		break;
+	case NL80211_IFTYPE_MESH_POINT:
+		hvif.opmode = HTC_M_WDS;	/* close enough */
+		break;
 	default:
 		ath_err(common,
 			"Interface type %d not yet supported\n", vif->type);
@@ -1084,6 +1091,7 @@ static int ath9k_htc_add_interface(struct ieee80211_hw *hw,
 	INC_VIF(priv, vif->type);
 
 	if ((vif->type == NL80211_IFTYPE_AP) ||
+	    (vif->type == NL80211_IFTYPE_MESH_POINT) ||
 	    (vif->type == NL80211_IFTYPE_ADHOC))
 		ath9k_htc_assign_bslot(priv, vif);
 
@@ -1134,6 +1142,7 @@ static void ath9k_htc_remove_interface(struct ieee80211_hw *hw,
 	DEC_VIF(priv, vif->type);
 
 	if ((vif->type == NL80211_IFTYPE_AP) ||
+	     vif->type == NL80211_IFTYPE_MESH_POINT ||
 	    (vif->type == NL80211_IFTYPE_ADHOC))
 		ath9k_htc_remove_bslot(priv, vif);
 
@@ -1525,9 +1534,10 @@ static void ath9k_htc_bss_info_changed(struct ieee80211_hw *hw,
 	if ((changed & BSS_CHANGED_BEACON_ENABLED) && !bss_conf->enable_beacon) {
 		/*
 		 * Disable SWBA interrupt only if there are no
-		 * AP/IBSS interfaces.
+		 * concurrent AP/mesh or IBSS interfaces.
 		 */
-		if ((priv->num_ap_vif <= 1) || priv->num_ibss_vif) {
+		if ((priv->num_ap_vif + priv->num_mbss_vif <= 1) ||
+		     priv->num_ibss_vif) {
 			ath_dbg(common, CONFIG,
 				"Beacon disabled for BSS: %pM\n",
 				bss_conf->bssid);
@@ -1538,12 +1548,15 @@ static void ath9k_htc_bss_info_changed(struct ieee80211_hw *hw,
 
 	if (changed & BSS_CHANGED_BEACON_INT) {
 		/*
-		 * Reset the HW TSF for the first AP interface.
+		 * Reset the HW TSF for the first AP or mesh interface.
 		 */
-		if ((priv->ah->opmode == NL80211_IFTYPE_AP) &&
-		    (priv->nvifs == 1) &&
-		    (priv->num_ap_vif == 1) &&
-		    (vif->type == NL80211_IFTYPE_AP)) {
+		if (priv->nvifs == 1 &&
+		    ((priv->ah->opmode == NL80211_IFTYPE_AP &&
+		      vif->type == NL80211_IFTYPE_AP &&
+		      priv->num_ap_vif == 1) ||
+		    (priv->ah->opmode == NL80211_IFTYPE_MESH_POINT &&
+		      vif->type == NL80211_IFTYPE_MESH_POINT &&
+		      priv->num_mbss_vif == 1))) {
 			set_bit(OP_TSF_RESET, &priv->op_flags);
 		}
 		ath_dbg(common, CONFIG,
@@ -1761,6 +1774,43 @@ static int ath9k_htc_get_stats(struct ieee80211_hw *hw,
 	return 0;
 }
 
+struct base_eep_header *ath9k_htc_get_eeprom_base(struct ath9k_htc_priv *priv)
+{
+	struct base_eep_header *pBase = NULL;
+	/*
+	 * This can be done since all the 3 EEPROM families have the
+	 * same base header upto a certain point, and we are interested in
+	 * the data only upto that point.
+	 */
+
+	if (AR_SREV_9271(priv->ah))
+		pBase = (struct base_eep_header *)
+			&priv->ah->eeprom.map4k.baseEepHeader;
+	else if (priv->ah->hw_version.usbdev == AR9280_USB)
+		pBase = (struct base_eep_header *)
+			&priv->ah->eeprom.def.baseEepHeader;
+	else if (priv->ah->hw_version.usbdev == AR9287_USB)
+		pBase = (struct base_eep_header *)
+			&priv->ah->eeprom.map9287.baseEepHeader;
+	return pBase;
+}
+
+
+static int ath9k_htc_get_antenna(struct ieee80211_hw *hw, u32 *tx_ant,
+				 u32 *rx_ant)
+{
+	struct ath9k_htc_priv *priv = hw->priv;
+	struct base_eep_header *pBase = ath9k_htc_get_eeprom_base(priv);
+	if (pBase) {
+		*tx_ant = pBase->txMask;
+		*rx_ant = pBase->rxMask;
+	} else {
+		*tx_ant = 0;
+		*rx_ant = 0;
+	}
+	return 0;
+}
+
 struct ieee80211_ops ath9k_htc_ops = {
 	.tx                 = ath9k_htc_tx,
 	.start              = ath9k_htc_start,
@@ -1786,4 +1836,11 @@ struct ieee80211_ops ath9k_htc_ops = {
 	.set_coverage_class = ath9k_htc_set_coverage_class,
 	.set_bitrate_mask   = ath9k_htc_set_bitrate_mask,
 	.get_stats	    = ath9k_htc_get_stats,
+	.get_antenna	    = ath9k_htc_get_antenna,
+
+#ifdef CONFIG_ATH9K_HTC_DEBUGFS
+	.get_et_sset_count  = ath9k_htc_get_et_sset_count,
+	.get_et_stats       = ath9k_htc_get_et_stats,
+	.get_et_strings     = ath9k_htc_get_et_strings,
+#endif
 };
diff --git a/drivers/net/wireless/ath/ath9k/htc_drv_txrx.c b/drivers/net/wireless/ath/ath9k/htc_drv_txrx.c
index 6bd0e92..e602c95 100644
--- a/drivers/net/wireless/ath/ath9k/htc_drv_txrx.c
+++ b/drivers/net/wireless/ath/ath9k/htc_drv_txrx.c
@@ -887,7 +887,7 @@ u32 ath9k_htc_calcrxfilter(struct ath9k_htc_priv *priv)
 	if (priv->rxfilter & FIF_PSPOLL)
 		rfilt |= ATH9K_RX_FILTER_PSPOLL;
 
-	if (priv->nvifs > 1)
+	if (priv->nvifs > 1 || priv->rxfilter & FIF_OTHER_BSS)
 		rfilt |= ATH9K_RX_FILTER_MCAST_BCAST_ALL;
 
 	return rfilt;
diff --git a/drivers/net/wireless/ath/ath9k/hw.c b/drivers/net/wireless/ath/ath9k/hw.c
index 15dfefc..4ca0cb0 100644
--- a/drivers/net/wireless/ath/ath9k/hw.c
+++ b/drivers/net/wireless/ath/ath9k/hw.c
@@ -452,7 +452,6 @@ static void ath9k_hw_init_config(struct ath_hw *ah)
 	ah->config.pcie_clock_req = 0;
 	ah->config.pcie_waen = 0;
 	ah->config.analog_shiftreg = 1;
-	ah->config.enable_ani = true;
 
 	for (i = 0; i < AR_EEPROM_MODAL_SPURS; i++) {
 		ah->config.spurchans[i][0] = AR_NO_SPUR;
@@ -549,8 +548,7 @@ static int ath9k_hw_post_init(struct ath_hw *ah)
 		ah->eep_ops->get_eeprom_ver(ah),
 		ah->eep_ops->get_eeprom_rev(ah));
 
-	if (ah->config.enable_ani)
-		ath9k_hw_ani_init(ah);
+	ath9k_hw_ani_init(ah);
 
 	return 0;
 }
@@ -1250,10 +1248,10 @@ static void ath9k_hw_set_operating_mode(struct ath_hw *ah, int opmode)
 
 	switch (opmode) {
 	case NL80211_IFTYPE_ADHOC:
-	case NL80211_IFTYPE_MESH_POINT:
 		set |= AR_STA_ID1_ADHOC;
 		REG_SET_BIT(ah, AR_CFG, AR_CFG_AP_ADHOC_INDICATION);
 		break;
+	case NL80211_IFTYPE_MESH_POINT:
 	case NL80211_IFTYPE_AP:
 		set |= AR_STA_ID1_STA_AP;
 		/* fall through */
@@ -1872,7 +1870,8 @@ int ath9k_hw_reset(struct ath_hw *ah, struct ath9k_channel *chan,
 
 	ah->caldata = caldata;
 	if (caldata && (chan->channel != caldata->channel ||
-			chan->channelFlags != caldata->channelFlags)) {
+			chan->channelFlags != caldata->channelFlags ||
+			chan->chanmode != caldata->chanmode)) {
 		/* Operating channel changed, reset channel calibration data */
 		memset(caldata, 0, sizeof(*caldata));
 		ath9k_init_nfcal_hist_buffer(ah, chan);
@@ -2255,12 +2254,12 @@ void ath9k_hw_beaconinit(struct ath_hw *ah, u32 next_beacon, u32 beacon_period)
 
 	switch (ah->opmode) {
 	case NL80211_IFTYPE_ADHOC:
-	case NL80211_IFTYPE_MESH_POINT:
 		REG_SET_BIT(ah, AR_TXCFG,
 			    AR_TXCFG_ADHOC_BEACON_ATIM_TX_POLICY);
 		REG_WRITE(ah, AR_NEXT_NDP_TIMER, next_beacon +
 			  TU_TO_USEC(ah->atim_window ? ah->atim_window : 1));
 		flags |= AR_NDP_TIMER_EN;
+	case NL80211_IFTYPE_MESH_POINT:
 	case NL80211_IFTYPE_AP:
 		REG_WRITE(ah, AR_NEXT_TBTT_TIMER, next_beacon);
 		REG_WRITE(ah, AR_NEXT_DMA_BEACON_ALERT, next_beacon -
@@ -2600,17 +2599,12 @@ int ath9k_hw_fill_cap_info(struct ath_hw *ah)
 		if (!(ah->ent_mode & AR_ENT_OTP_49GHZ_DISABLE))
 			pCap->hw_caps |= ATH9K_HW_CAP_MCI;
 
-		if (AR_SREV_9462_20(ah))
+		if (AR_SREV_9462_20_OR_LATER(ah))
 			pCap->hw_caps |= ATH9K_HW_CAP_RTT;
 	}
 
-	if (AR_SREV_9280_20_OR_LATER(ah)) {
-		pCap->hw_caps |= ATH9K_HW_WOW_DEVICE_CAPABLE |
-				 ATH9K_HW_WOW_PATTERN_MATCH_EXACT;
-
-		if (AR_SREV_9280(ah))
-			pCap->hw_caps |= ATH9K_HW_WOW_PATTERN_MATCH_DWORD;
-	}
+	if (AR_SREV_9462(ah))
+		pCap->hw_caps |= ATH9K_HW_WOW_DEVICE_CAPABLE;
 
 	if (AR_SREV_9300_20_OR_LATER(ah) &&
 	    ah->eep_ops->get_eeprom(ah, EEP_PAPRD))
@@ -3048,7 +3042,7 @@ void ath9k_hw_gen_timer_start(struct ath_hw *ah,
 
 	timer_next = tsf + trig_timeout;
 
-	ath_dbg(ath9k_hw_common(ah), HWTIMER,
+	ath_dbg(ath9k_hw_common(ah), BTCOEX,
 		"current tsf %x period %x timer_next %x\n",
 		tsf, timer_period, timer_next);
 
@@ -3147,7 +3141,7 @@ void ath_gen_timer_isr(struct ath_hw *ah)
 		index = rightmost_index(timer_table, &thresh_mask);
 		timer = timer_table->timers[index];
 		BUG_ON(!timer);
-		ath_dbg(common, HWTIMER, "TSF overflow for Gen timer %d\n",
+		ath_dbg(common, BTCOEX, "TSF overflow for Gen timer %d\n",
 			index);
 		timer->overflow(timer->arg);
 	}
@@ -3156,7 +3150,7 @@ void ath_gen_timer_isr(struct ath_hw *ah)
 		index = rightmost_index(timer_table, &trigger_mask);
 		timer = timer_table->timers[index];
 		BUG_ON(!timer);
-		ath_dbg(common, HWTIMER,
+		ath_dbg(common, BTCOEX,
 			"Gen timer[%d] trigger\n", index);
 		timer->trigger(timer->arg);
 	}
diff --git a/drivers/net/wireless/ath/ath9k/hw.h b/drivers/net/wireless/ath/ath9k/hw.h
index ae30343..cd74b3a 100644
--- a/drivers/net/wireless/ath/ath9k/hw.h
+++ b/drivers/net/wireless/ath/ath9k/hw.h
@@ -246,9 +246,7 @@ enum ath9k_hw_caps {
 	ATH9K_HW_CAP_MCI			= BIT(15),
 	ATH9K_HW_CAP_DFS			= BIT(16),
 	ATH9K_HW_WOW_DEVICE_CAPABLE		= BIT(17),
-	ATH9K_HW_WOW_PATTERN_MATCH_EXACT	= BIT(18),
-	ATH9K_HW_WOW_PATTERN_MATCH_DWORD	= BIT(19),
-	ATH9K_HW_CAP_PAPRD			= BIT(20),
+	ATH9K_HW_CAP_PAPRD			= BIT(18),
 };
 
 /*
@@ -291,7 +289,6 @@ struct ath9k_ops_config {
 	u32 ofdm_trig_high;
 	u32 cck_trig_high;
 	u32 cck_trig_low;
-	u32 enable_ani;
 	u32 enable_paprd;
 	int serialize_regmode;
 	bool rx_intr_mitigation;
@@ -310,6 +307,10 @@ struct ath9k_ops_config {
 	u16 spurchans[AR_EEPROM_MODAL_SPURS][2];
 	u8 max_txtrig_level;
 	u16 ani_poll_interval; /* ANI poll interval in ms */
+
+	/* Platform specific config */
+	u32 xlna_gpio;
+	bool xatten_margin_cfg;
 };
 
 enum ath9k_int {
@@ -423,7 +424,6 @@ struct ath9k_hw_cal_data {
 
 struct ath9k_channel {
 	struct ieee80211_channel *chan;
-	struct ar5416AniState ani;
 	u16 channel;
 	u32 channelFlags;
 	u32 chanmode;
@@ -854,10 +854,10 @@ struct ath_hw {
 	u32 globaltxtimeout;
 
 	/* ANI */
-	u32 proc_phyerr;
 	u32 aniperiod;
 	enum ath9k_ani_cmd ani_function;
 	u32 ani_skip_count;
+	struct ar5416AniState ani;
 
 #ifdef CONFIG_ATH9K_BTCOEX_SUPPORT
 	struct ath_btcoex_hw btcoex_hw;
@@ -882,9 +882,6 @@ struct ath_hw {
 	struct ar5416IniArray iniBank6;
 	struct ar5416IniArray iniAddac;
 	struct ar5416IniArray iniPcieSerdes;
-#ifdef CONFIG_PM_SLEEP
-	struct ar5416IniArray iniPcieSerdesWow;
-#endif
 	struct ar5416IniArray iniPcieSerdesLowPower;
 	struct ar5416IniArray iniModesFastClock;
 	struct ar5416IniArray iniAdditional;
@@ -895,6 +892,9 @@ struct ath_hw {
 	struct ar5416IniArray iniCckfirJapan2484;
 	struct ar5416IniArray iniModes_9271_ANI_reg;
 	struct ar5416IniArray ini_radio_post_sys2ant;
+	struct ar5416IniArray ini_modes_rxgain_5g_xlna;
+	struct ar5416IniArray ini_modes_rxgain_bb_core;
+	struct ar5416IniArray ini_modes_rxgain_bb_postamble;
 
 	struct ar5416IniArray iniMac[ATH_INI_NUM_SPLIT];
 	struct ar5416IniArray iniBB[ATH_INI_NUM_SPLIT];
@@ -1165,8 +1165,6 @@ static inline void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 }
 #endif
 
-
-
 #define ATH9K_CLOCK_RATE_CCK		22
 #define ATH9K_CLOCK_RATE_5GHZ_OFDM	40
 #define ATH9K_CLOCK_RATE_2GHZ_OFDM	44
diff --git a/drivers/net/wireless/ath/ath9k/init.c b/drivers/net/wireless/ath/ath9k/init.c
index 2ba4945..16f8b20 100644
--- a/drivers/net/wireless/ath/ath9k/init.c
+++ b/drivers/net/wireless/ath/ath9k/init.c
@@ -21,6 +21,7 @@
 #include <linux/ath9k_platform.h>
 #include <linux/module.h>
 #include <linux/relay.h>
+#include <net/ieee80211_radiotap.h>
 
 #include "ath9k.h"
 
@@ -431,6 +432,8 @@ static int ath9k_init_queues(struct ath_softc *sc)
 	sc->config.cabqReadytime = ATH_CABQ_READY_TIME;
 	ath_cabq_update(sc);
 
+	sc->tx.uapsdq = ath_txq_setup(sc, ATH9K_TX_QUEUE_UAPSD, 0);
+
 	for (i = 0; i < IEEE80211_NUM_ACS; i++) {
 		sc->tx.txq_map[i] = ath_txq_setup(sc, ATH9K_TX_QUEUE_DATA, i);
 		sc->tx.txq_map[i]->mac80211_qnum = i;
@@ -510,6 +513,27 @@ static void ath9k_init_misc(struct ath_softc *sc)
 	sc->spec_config.fft_period = 0xF;
 }
 
+static void ath9k_init_platform(struct ath_softc *sc)
+{
+	struct ath_hw *ah = sc->sc_ah;
+	struct ath_common *common = ath9k_hw_common(ah);
+
+	if (common->bus_ops->ath_bus_type != ATH_PCI)
+		return;
+
+	if (sc->driver_data & (ATH9K_PCI_CUS198 |
+			       ATH9K_PCI_CUS230)) {
+		ah->config.xlna_gpio = 9;
+		ah->config.xatten_margin_cfg = true;
+
+		ath_info(common, "Set parameters for %s\n",
+			 (sc->driver_data & ATH9K_PCI_CUS198) ?
+			 "CUS198" : "CUS230");
+	} else if (sc->driver_data & ATH9K_PCI_CUS217) {
+		ath_info(common, "CUS217 card detected\n");
+	}
+}
+
 static void ath9k_eeprom_request_cb(const struct firmware *eeprom_blob,
 				    void *ctx)
 {
@@ -602,6 +626,11 @@ static int ath9k_init_softc(u16 devid, struct ath_softc *sc,
 	common->disable_ani = false;
 
 	/*
+	 * Platform quirks.
+	 */
+	ath9k_init_platform(sc);
+
+	/*
 	 * Enable Antenna diversity only when BTCOEX is disabled
 	 * and the user manually requests the feature.
 	 */
@@ -613,9 +642,6 @@ static int ath9k_init_softc(u16 devid, struct ath_softc *sc,
 	spin_lock_init(&sc->sc_serial_rw);
 	spin_lock_init(&sc->sc_pm_lock);
 	mutex_init(&sc->mutex);
-#ifdef CONFIG_ATH9K_MAC_DEBUG
-	spin_lock_init(&sc->debug.samp_lock);
-#endif
 	tasklet_init(&sc->intr_tq, ath9k_tasklet, (unsigned long)sc);
 	tasklet_init(&sc->bcon_tasklet, ath9k_beacon_tasklet,
 		     (unsigned long)sc);
@@ -755,6 +781,15 @@ static const struct ieee80211_iface_combination if_comb[] = {
 	}
 };
 
+#ifdef CONFIG_PM
+static const struct wiphy_wowlan_support ath9k_wowlan_support = {
+	.flags = WIPHY_WOWLAN_MAGIC_PKT | WIPHY_WOWLAN_DISCONNECT,
+	.n_patterns = MAX_NUM_USER_PATTERN,
+	.pattern_min_len = 1,
+	.pattern_max_len = MAX_PATTERN_SIZE,
+};
+#endif
+
 void ath9k_set_hw_capab(struct ath_softc *sc, struct ieee80211_hw *hw)
 {
 	struct ath_hw *ah = sc->sc_ah;
@@ -769,12 +804,19 @@ void ath9k_set_hw_capab(struct ath_softc *sc, struct ieee80211_hw *hw)
 		IEEE80211_HW_REPORTS_TX_ACK_STATUS |
 		IEEE80211_HW_SUPPORTS_RC_TABLE;
 
-	if (sc->sc_ah->caps.hw_caps & ATH9K_HW_CAP_HT)
-		 hw->flags |= IEEE80211_HW_AMPDU_AGGREGATION;
+	if (sc->sc_ah->caps.hw_caps & ATH9K_HW_CAP_HT) {
+		hw->flags |= IEEE80211_HW_AMPDU_AGGREGATION;
+
+		if (AR_SREV_9280_20_OR_LATER(ah))
+			hw->radiotap_mcs_details |=
+				IEEE80211_RADIOTAP_MCS_HAVE_STBC;
+	}
 
 	if (AR_SREV_9160_10_OR_LATER(sc->sc_ah) || ath9k_modparam_nohwcrypt)
 		hw->flags |= IEEE80211_HW_MFP_CAPABLE;
 
+	hw->wiphy->features |= NL80211_FEATURE_ACTIVE_MONITOR;
+
 	hw->wiphy->interface_modes =
 		BIT(NL80211_IFTYPE_P2P_GO) |
 		BIT(NL80211_IFTYPE_P2P_CLIENT) |
@@ -794,21 +836,13 @@ void ath9k_set_hw_capab(struct ath_softc *sc, struct ieee80211_hw *hw)
 	hw->wiphy->flags |= WIPHY_FLAG_HAS_REMAIN_ON_CHANNEL;
 
 #ifdef CONFIG_PM_SLEEP
-
 	if ((ah->caps.hw_caps & ATH9K_HW_WOW_DEVICE_CAPABLE) &&
-	    device_can_wakeup(sc->dev)) {
-
-		hw->wiphy->wowlan.flags = WIPHY_WOWLAN_MAGIC_PKT |
-					  WIPHY_WOWLAN_DISCONNECT;
-		hw->wiphy->wowlan.n_patterns = MAX_NUM_USER_PATTERN;
-		hw->wiphy->wowlan.pattern_min_len = 1;
-		hw->wiphy->wowlan.pattern_max_len = MAX_PATTERN_SIZE;
-
-	}
+	    (sc->driver_data & ATH9K_PCI_WOW) &&
+	    device_can_wakeup(sc->dev))
+		hw->wiphy->wowlan = &ath9k_wowlan_support;
 
 	atomic_set(&sc->wow_sleep_proc_intr, -1);
 	atomic_set(&sc->wow_got_bmiss_intr, -1);
-
 #endif
 
 	hw->queues = 4;
diff --git a/drivers/net/wireless/ath/ath9k/link.c b/drivers/net/wireless/ath/ath9k/link.c
index 849259b..fff5d3c 100644
--- a/drivers/net/wireless/ath/ath9k/link.c
+++ b/drivers/net/wireless/ath/ath9k/link.c
@@ -390,9 +390,7 @@ void ath_ani_calibrate(unsigned long data)
 	}
 
 	/* Verify whether we must check ANI */
-	if (sc->sc_ah->config.enable_ani
-	    && (timestamp - common->ani.checkani_timer) >=
-	    ah->config.ani_poll_interval) {
+	if ((timestamp - common->ani.checkani_timer) >= ah->config.ani_poll_interval) {
 		aniflag = true;
 		common->ani.checkani_timer = timestamp;
 	}
@@ -418,7 +416,6 @@ void ath_ani_calibrate(unsigned long data)
 		longcal ? "long" : "", shortcal ? "short" : "",
 		aniflag ? "ani" : "", common->ani.caldone ? "true" : "false");
 
-	ath9k_debug_samp_bb_mac(sc);
 	ath9k_ps_restore(sc);
 
 set_timer:
@@ -428,9 +425,7 @@ set_timer:
 	* short calibration and long calibration.
 	*/
 	cal_interval = ATH_LONG_CALINTERVAL;
-	if (sc->sc_ah->config.enable_ani)
-		cal_interval = min(cal_interval,
-				   (u32)ah->config.ani_poll_interval);
+	cal_interval = min(cal_interval, (u32)ah->config.ani_poll_interval);
 	if (!common->ani.caldone)
 		cal_interval = min(cal_interval, (u32)short_cal_interval);
 
diff --git a/drivers/net/wireless/ath/ath9k/mac.c b/drivers/net/wireless/ath/ath9k/mac.c
index 566109a..2ef05eb 100644
--- a/drivers/net/wireless/ath/ath9k/mac.c
+++ b/drivers/net/wireless/ath/ath9k/mac.c
@@ -547,6 +547,7 @@ int ath9k_hw_rxprocdesc(struct ath_hw *ah, struct ath_desc *ds,
 
 	rs->rs_status = 0;
 	rs->rs_flags = 0;
+	rs->flag = 0;
 
 	rs->rs_datalen = ads.ds_rxstatus1 & AR_DataLen;
 	rs->rs_tstamp = ads.AR_RcvTimestamp;
@@ -586,10 +587,17 @@ int ath9k_hw_rxprocdesc(struct ath_hw *ah, struct ath_desc *ds,
 	rs->rs_moreaggr =
 		(ads.ds_rxstatus8 & AR_RxMoreAggr) ? 1 : 0;
 	rs->rs_antenna = MS(ads.ds_rxstatus3, AR_RxAntenna);
-	rs->rs_flags =
-		(ads.ds_rxstatus3 & AR_GI) ? ATH9K_RX_GI : 0;
-	rs->rs_flags |=
-		(ads.ds_rxstatus3 & AR_2040) ? ATH9K_RX_2040 : 0;
+
+	/* directly mapped flags for ieee80211_rx_status */
+	rs->flag |=
+		(ads.ds_rxstatus3 & AR_GI) ? RX_FLAG_SHORT_GI : 0;
+	rs->flag |=
+		(ads.ds_rxstatus3 & AR_2040) ? RX_FLAG_40MHZ : 0;
+	if (AR_SREV_9280_20_OR_LATER(ah))
+		rs->flag |=
+			(ads.ds_rxstatus3 & AR_STBC) ?
+				/* we can only Nss=1 STBC */
+				(1 << RX_FLAG_STBC_SHIFT) : 0;
 
 	if (ads.ds_rxstatus8 & AR_PreDelimCRCErr)
 		rs->rs_flags |= ATH9K_RX_DELIM_CRC_PRE;
diff --git a/drivers/net/wireless/ath/ath9k/mac.h b/drivers/net/wireless/ath/ath9k/mac.h
index 5865f92..b02dfce 100644
--- a/drivers/net/wireless/ath/ath9k/mac.h
+++ b/drivers/net/wireless/ath/ath9k/mac.h
@@ -149,6 +149,7 @@ struct ath_rx_status {
 	u32 evm2;
 	u32 evm3;
 	u32 evm4;
+	u32 flag; /* see enum mac80211_rx_flags */
 };
 
 struct ath_htc_rx_status {
@@ -533,7 +534,8 @@ struct ar5416_desc {
 #define AR_2040             0x00000002
 #define AR_Parallel40       0x00000004
 #define AR_Parallel40_S     2
-#define AR_RxStatusRsvd30   0x000000f8
+#define AR_STBC             0x00000008 /* on ar9280 and later */
+#define AR_RxStatusRsvd30   0x000000f0
 #define AR_RxAntenna	    0xffffff00
 #define AR_RxAntenna_S	    8
 
diff --git a/drivers/net/wireless/ath/ath9k/main.c b/drivers/net/wireless/ath/ath9k/main.c
index 5092eca..1737a3e 100644
--- a/drivers/net/wireless/ath/ath9k/main.c
+++ b/drivers/net/wireless/ath/ath9k/main.c
@@ -193,7 +193,6 @@ static bool ath_prepare_reset(struct ath_softc *sc)
 	ath_stop_ani(sc);
 	del_timer_sync(&sc->rx_poll_timer);
 
-	ath9k_debug_samp_bb_mac(sc);
 	ath9k_hw_disable_interrupts(ah);
 
 	if (!ath_drain_all_txq(sc))
@@ -1211,13 +1210,6 @@ static int ath9k_config(struct ieee80211_hw *hw, u32 changed)
 		ath_update_survey_stats(sc);
 		spin_unlock_irqrestore(&common->cc_lock, flags);
 
-		/*
-		 * Preserve the current channel values, before updating
-		 * the same channel
-		 */
-		if (ah->curchan && (old_pos == pos))
-			ath9k_hw_getnf(ah, ah->curchan);
-
 		ath9k_cmn_update_ichannel(&sc->sc_ah->channels[pos],
 					  curchan, channel_type);
 
@@ -1273,7 +1265,7 @@ static int ath9k_config(struct ieee80211_hw *hw, u32 changed)
 				curchan->center_freq);
 		} else {
 			/* perform spectral scan if requested. */
-			if (sc->scanning &&
+			if (test_bit(SC_OP_SCANNING, &sc->sc_flags) &&
 			    sc->spectral_mode == SPECTRAL_CHANSCAN)
 				ath9k_spectral_scan_trigger(hw);
 		}
@@ -1690,7 +1682,7 @@ static int ath9k_ampdu_action(struct ieee80211_hw *hw,
 	bool flush = false;
 	int ret = 0;
 
-	local_bh_disable();
+	mutex_lock(&sc->mutex);
 
 	switch (action) {
 	case IEEE80211_AMPDU_RX_START:
@@ -1723,7 +1715,7 @@ static int ath9k_ampdu_action(struct ieee80211_hw *hw,
 		ath_err(ath9k_hw_common(sc->sc_ah), "Unknown AMPDU action\n");
 	}
 
-	local_bh_enable();
+	mutex_unlock(&sc->mutex);
 
 	return ret;
 }
@@ -2007,7 +1999,6 @@ static void ath9k_wow_add_disassoc_deauth_pattern(struct ath_softc *sc)
 {
 	struct ath_hw *ah = sc->sc_ah;
 	struct ath_common *common = ath9k_hw_common(ah);
-	struct ath9k_hw_capabilities *pcaps = &ah->caps;
 	int pattern_count = 0;
 	int i, byte_cnt;
 	u8 dis_deauth_pattern[MAX_PATTERN_SIZE];
@@ -2077,36 +2068,9 @@ static void ath9k_wow_add_disassoc_deauth_pattern(struct ath_softc *sc)
 
 	/* Create Disassociate pattern mask */
 
-	if (pcaps->hw_caps & ATH9K_HW_WOW_PATTERN_MATCH_EXACT) {
-
-		if (pcaps->hw_caps & ATH9K_HW_WOW_PATTERN_MATCH_DWORD) {
-			/*
-			 * for AR9280, because of hardware limitation, the
-			 * first 4 bytes have to be matched for all patterns.
-			 * the mask for disassociation and de-auth pattern
-			 * matching need to enable the first 4 bytes.
-			 * also the duration field needs to be filled.
-			 */
-			dis_deauth_mask[0] = 0xf0;
-
-			/*
-			 * fill in duration field
-			 FIXME: what is the exact value ?
-			 */
-			dis_deauth_pattern[2] = 0xff;
-			dis_deauth_pattern[3] = 0xff;
-		} else {
-			dis_deauth_mask[0] = 0xfe;
-		}
-
-		dis_deauth_mask[1] = 0x03;
-		dis_deauth_mask[2] = 0xc0;
-	} else {
-		dis_deauth_mask[0] = 0xef;
-		dis_deauth_mask[1] = 0x3f;
-		dis_deauth_mask[2] = 0x00;
-		dis_deauth_mask[3] = 0xfc;
-	}
+	dis_deauth_mask[0] = 0xfe;
+	dis_deauth_mask[1] = 0x03;
+	dis_deauth_mask[2] = 0xc0;
 
 	ath_dbg(common, WOW, "Adding disassoc/deauth patterns for WoW\n");
 
@@ -2342,15 +2306,13 @@ static void ath9k_set_wakeup(struct ieee80211_hw *hw, bool enabled)
 static void ath9k_sw_scan_start(struct ieee80211_hw *hw)
 {
 	struct ath_softc *sc = hw->priv;
-
-	sc->scanning = 1;
+	set_bit(SC_OP_SCANNING, &sc->sc_flags);
 }
 
 static void ath9k_sw_scan_complete(struct ieee80211_hw *hw)
 {
 	struct ath_softc *sc = hw->priv;
-
-	sc->scanning = 0;
+	clear_bit(SC_OP_SCANNING, &sc->sc_flags);
 }
 
 struct ieee80211_ops ath9k_ops = {
@@ -2378,6 +2340,7 @@ struct ieee80211_ops ath9k_ops = {
 	.flush		    = ath9k_flush,
 	.tx_frames_pending  = ath9k_tx_frames_pending,
 	.tx_last_beacon     = ath9k_tx_last_beacon,
+	.release_buffered_frames = ath9k_release_buffered_frames,
 	.get_stats	    = ath9k_get_stats,
 	.set_antenna	    = ath9k_set_antenna,
 	.get_antenna	    = ath9k_get_antenna,
diff --git a/drivers/net/wireless/ath/ath9k/pci.c b/drivers/net/wireless/ath/ath9k/pci.c
index 0e0d395..c585c9b 100644
--- a/drivers/net/wireless/ath/ath9k/pci.c
+++ b/drivers/net/wireless/ath/ath9k/pci.c
@@ -34,8 +34,108 @@ static DEFINE_PCI_DEVICE_TABLE(ath_pci_id_table) = {
 	{ PCI_VDEVICE(ATHEROS, 0x002D) }, /* PCI   */
 	{ PCI_VDEVICE(ATHEROS, 0x002E) }, /* PCI-E */
 	{ PCI_VDEVICE(ATHEROS, 0x0030) }, /* PCI-E  AR9300 */
+
+	/* PCI-E CUS198 */
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0032,
+			 PCI_VENDOR_ID_AZWAVE,
+			 0x2086),
+	  .driver_data = ATH9K_PCI_CUS198 },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0032,
+			 PCI_VENDOR_ID_AZWAVE,
+			 0x1237),
+	  .driver_data = ATH9K_PCI_CUS198 },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0032,
+			 PCI_VENDOR_ID_AZWAVE,
+			 0x2126),
+	  .driver_data = ATH9K_PCI_CUS198 },
+
+	/* PCI-E CUS230 */
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0032,
+			 PCI_VENDOR_ID_AZWAVE,
+			 0x2152),
+	  .driver_data = ATH9K_PCI_CUS230 },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0032,
+			 PCI_VENDOR_ID_FOXCONN,
+			 0xE075),
+	  .driver_data = ATH9K_PCI_CUS230 },
+
 	{ PCI_VDEVICE(ATHEROS, 0x0032) }, /* PCI-E  AR9485 */
 	{ PCI_VDEVICE(ATHEROS, 0x0033) }, /* PCI-E  AR9580 */
+
+	/* PCI-E CUS217 */
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_AZWAVE,
+			 0x2116),
+	  .driver_data = ATH9K_PCI_CUS217 },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 0x11AD, /* LITEON */
+			 0x6661),
+	  .driver_data = ATH9K_PCI_CUS217 },
+
+	/* AR9462 with WoW support */
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_ATHEROS,
+			 0x3117),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_LENOVO,
+			 0x3214),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_ATTANSIC,
+			 0x0091),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_AZWAVE,
+			 0x2110),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_ASUSTEK,
+			 0x850E),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 0x11AD, /* LITEON */
+			 0x6631),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 0x11AD, /* LITEON */
+			 0x6641),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 PCI_VENDOR_ID_HP,
+			 0x1864),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 0x14CD, /* USI */
+			 0x0063),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 0x14CD, /* USI */
+			 0x0064),
+	  .driver_data = ATH9K_PCI_WOW },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_ATHEROS,
+			 0x0034,
+			 0x10CF, /* Fujitsu */
+			 0x1783),
+	  .driver_data = ATH9K_PCI_WOW },
+
 	{ PCI_VDEVICE(ATHEROS, 0x0034) }, /* PCI-E  AR9462 */
 	{ PCI_VDEVICE(ATHEROS, 0x0037) }, /* PCI-E  AR1111/AR9485 */
 	{ PCI_VDEVICE(ATHEROS, 0x0036) }, /* PCI-E  AR9565 */
@@ -221,6 +321,7 @@ static int ath_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	sc->hw = hw;
 	sc->dev = &pdev->dev;
 	sc->mem = pcim_iomap_table(pdev)[0];
+	sc->driver_data = id->driver_data;
 
 	/* Will be cleared in ath9k_start() */
 	set_bit(SC_OP_INVALID, &sc->sc_flags);
diff --git a/drivers/net/wireless/ath/ath9k/recv.c b/drivers/net/wireless/ath/ath9k/recv.c
index 8be2b5d..865e043 100644
--- a/drivers/net/wireless/ath/ath9k/recv.c
+++ b/drivers/net/wireless/ath/ath9k/recv.c
@@ -868,10 +868,7 @@ static int ath9k_process_rate(struct ath_common *common,
 	if (rx_stats->rs_rate & 0x80) {
 		/* HT rate */
 		rxs->flag |= RX_FLAG_HT;
-		if (rx_stats->rs_flags & ATH9K_RX_2040)
-			rxs->flag |= RX_FLAG_40MHZ;
-		if (rx_stats->rs_flags & ATH9K_RX_GI)
-			rxs->flag |= RX_FLAG_SHORT_GI;
+		rxs->flag |= rx_stats->flag;
 		rxs->rate_idx = rx_stats->rs_rate & 0x7f;
 		return 0;
 	}
@@ -958,11 +955,11 @@ static int ath9k_rx_skb_preprocess(struct ath_softc *sc,
 	if (rx_stats->rs_more)
 		return 0;
 
-	ath9k_process_rssi(common, hw, hdr, rx_stats);
-
 	if (ath9k_process_rate(common, hw, rx_stats, rx_status))
 		return -EINVAL;
 
+	ath9k_process_rssi(common, hw, hdr, rx_stats);
+
 	rx_status->band = hw->conf.chandef.chan->band;
 	rx_status->freq = hw->conf.chandef.chan->center_freq;
 	rx_status->signal = ah->noise + rx_stats->rs_rssi;
diff --git a/drivers/net/wireless/ath/ath9k/reg.h b/drivers/net/wireless/ath/ath9k/reg.h
index f7c90cc..5af9744 100644
--- a/drivers/net/wireless/ath/ath9k/reg.h
+++ b/drivers/net/wireless/ath/ath9k/reg.h
@@ -806,6 +806,7 @@
 #define AR_SREV_REVISION_9580_10	4 /* AR9580 1.0 */
 #define AR_SREV_VERSION_9462		0x280
 #define AR_SREV_REVISION_9462_20	2
+#define AR_SREV_REVISION_9462_21	3
 #define AR_SREV_VERSION_9565            0x2C0
 #define AR_SREV_REVISION_9565_10        0
 #define AR_SREV_VERSION_9550		0x400
@@ -911,10 +912,18 @@
 
 #define AR_SREV_9462(_ah) \
 	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9462))
-
 #define AR_SREV_9462_20(_ah) \
 	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9462) && \
-	((_ah)->hw_version.macRev == AR_SREV_REVISION_9462_20))
+	 ((_ah)->hw_version.macRev == AR_SREV_REVISION_9462_20))
+#define AR_SREV_9462_21(_ah) \
+	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9462) && \
+	 ((_ah)->hw_version.macRev == AR_SREV_REVISION_9462_21))
+#define AR_SREV_9462_20_OR_LATER(_ah) \
+	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9462) && \
+	 ((_ah)->hw_version.macRev >= AR_SREV_REVISION_9462_20))
+#define AR_SREV_9462_21_OR_LATER(_ah) \
+	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9462) && \
+	 ((_ah)->hw_version.macRev >= AR_SREV_REVISION_9462_21))
 
 #define AR_SREV_9565(_ah) \
 	(((_ah)->hw_version.macVersion == AR_SREV_VERSION_9565))
diff --git a/drivers/net/wireless/ath/ath9k/wow.c b/drivers/net/wireless/ath/ath9k/wow.c
index 9f85630..81c88dd 100644
--- a/drivers/net/wireless/ath/ath9k/wow.c
+++ b/drivers/net/wireless/ath/ath9k/wow.c
@@ -34,17 +34,6 @@ const char *ath9k_hw_wow_event_to_string(u32 wow_event)
 }
 EXPORT_SYMBOL(ath9k_hw_wow_event_to_string);
 
-static void ath9k_hw_config_serdes_wow_sleep(struct ath_hw *ah)
-{
-	int i;
-
-	for (i = 0; i < ah->iniPcieSerdesWow.ia_rows; i++)
-		REG_WRITE(ah, INI_RA(&ah->iniPcieSerdesWow, i, 0),
-			  INI_RA(&ah->iniPcieSerdesWow, i, 1));
-
-	usleep_range(1000, 1500);
-}
-
 static void ath9k_hw_set_powermode_wow_sleep(struct ath_hw *ah)
 {
 	struct ath_common *common = ath9k_hw_common(ah);
@@ -58,15 +47,8 @@ static void ath9k_hw_set_powermode_wow_sleep(struct ath_hw *ah)
 		ath_err(common, "Failed to stop Rx DMA in 10ms AR_CR=0x%08x AR_DIAG_SW=0x%08x\n",
 			REG_READ(ah, AR_CR), REG_READ(ah, AR_DIAG_SW));
 		return;
-	} else {
-		if (!AR_SREV_9300_20_OR_LATER(ah))
-			REG_WRITE(ah, AR_RXDP, 0x0);
 	}
 
-	/* AR9280 WoW has sleep issue, do not set it to sleep */
-	if (AR_SREV_9280_20(ah))
-		return;
-
 	REG_WRITE(ah, AR_RTC_FORCE_WAKE, AR_RTC_FORCE_WAKE_ON_INT);
 }
 
@@ -84,27 +66,16 @@ static void ath9k_wow_create_keep_alive_pattern(struct ath_hw *ah)
 
 	/* set the transmit buffer */
 	ctl[0] = (KAL_FRAME_LEN | (MAX_RATE_POWER << 16));
-
-	if (!(AR_SREV_9300_20_OR_LATER(ah)))
-		ctl[0] += (KAL_ANTENNA_MODE << 25);
-
 	ctl[1] = 0;
 	ctl[3] = 0xb;	/* OFDM_6M hardware value for this rate */
 	ctl[4] = 0;
 	ctl[7] = (ah->txchainmask) << 2;
-
-	if (AR_SREV_9300_20_OR_LATER(ah))
-		ctl[2] = 0xf << 16; /* tx_tries 0 */
-	else
-		ctl[2] = 0x7 << 16; /* tx_tries 0 */
-
+	ctl[2] = 0xf << 16; /* tx_tries 0 */
 
 	for (i = 0; i < KAL_NUM_DESC_WORDS; i++)
 		REG_WRITE(ah, (AR_WOW_KA_DESC_WORD2 + i * 4), ctl[i]);
 
-	/* for AR9300 family 13 descriptor words */
-	if (AR_SREV_9300_20_OR_LATER(ah))
-		REG_WRITE(ah, (AR_WOW_KA_DESC_WORD2 + i * 4), ctl[i]);
+	REG_WRITE(ah, (AR_WOW_KA_DESC_WORD2 + i * 4), ctl[i]);
 
 	data_word[0] = (KAL_FRAME_TYPE << 2) | (KAL_FRAME_SUB_TYPE << 4) |
 		       (KAL_TO_DS << 8) | (KAL_DURATION_ID << 16);
@@ -183,9 +154,6 @@ void ath9k_hw_wow_apply_pattern(struct ath_hw *ah, u8 *user_pattern,
 
 	ah->wow_event_mask |= BIT(pattern_count + AR_WOW_PAT_FOUND_SHIFT);
 
-	if (!AR_SREV_9285_12_OR_LATER(ah))
-		return;
-
 	if (pattern_count < 4) {
 		/* Pattern 0-3 uses AR_WOW_LENGTH1 register */
 		set = (pattern_len & AR_WOW_LENGTH_MAX) <<
@@ -207,6 +175,7 @@ u32 ath9k_hw_wow_wakeup(struct ath_hw *ah)
 {
 	u32 wow_status = 0;
 	u32 val = 0, rval;
+
 	/*
 	 * read the WoW status register to know
 	 * the wakeup reason
@@ -223,19 +192,14 @@ u32 ath9k_hw_wow_wakeup(struct ath_hw *ah)
 	val &= ah->wow_event_mask;
 
 	if (val) {
-
 		if (val & AR_WOW_MAGIC_PAT_FOUND)
 			wow_status |= AH_WOW_MAGIC_PATTERN_EN;
-
 		if (AR_WOW_PATTERN_FOUND(val))
 			wow_status |= AH_WOW_USER_PATTERN_EN;
-
 		if (val & AR_WOW_KEEP_ALIVE_FAIL)
 			wow_status |= AH_WOW_LINK_CHANGE;
-
 		if (val & AR_WOW_BEACON_FAIL)
 			wow_status |= AH_WOW_BEACON_MISS;
-
 	}
 
 	/*
@@ -255,17 +219,6 @@ u32 ath9k_hw_wow_wakeup(struct ath_hw *ah)
 		  AR_WOW_CLEAR_EVENTS(REG_READ(ah, AR_WOW_PATTERN)));
 
 	/*
-	 * tie reset register for AR9002 family of chipsets
-	 * NB: not tieing it back might have some repurcussions.
-	 */
-
-	if (!AR_SREV_9300_20_OR_LATER(ah)) {
-		REG_SET_BIT(ah, AR_WA, AR_WA_UNTIE_RESET_EN |
-			    AR_WA_POR_SHORT | AR_WA_RESET_EN);
-	}
-
-
-	/*
 	 * restore the beacon threshold to init value
 	 */
 	REG_WRITE(ah, AR_RSSI_THR, INIT_RSSI_THR);
@@ -277,8 +230,7 @@ u32 ath9k_hw_wow_wakeup(struct ath_hw *ah)
 	 * reset to our Chip's Power On Reset so that any PCI-E
 	 * reset from the bus will not reset our chip
 	 */
-
-	if (AR_SREV_9280_20_OR_LATER(ah) && ah->is_pciexpress)
+	if (ah->is_pciexpress)
 		ath9k_hw_configpcipowersave(ah, false);
 
 	ah->wow_event_mask = 0;
@@ -298,7 +250,6 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	 * are from the 'pattern_enable' in this function and
 	 * 'pattern_count' of ath9k_hw_wow_apply_pattern()
 	 */
-
 	wow_event_mask = ah->wow_event_mask;
 
 	/*
@@ -306,50 +257,15 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	 * WOW sleep, we do want the Reset from the PCI-E to disturb
 	 * our hw state
 	 */
-
 	if (ah->is_pciexpress) {
-
 		/*
 		 * we need to untie the internal POR (power-on-reset)
 		 * to the external PCI-E reset. We also need to tie
 		 * the PCI-E Phy reset to the PCI-E reset.
 		 */
-
-		if (AR_SREV_9300_20_OR_LATER(ah)) {
-			set = AR_WA_RESET_EN | AR_WA_POR_SHORT;
-			clr = AR_WA_UNTIE_RESET_EN | AR_WA_D3_L1_DISABLE;
-			REG_RMW(ah, AR_WA, set, clr);
-		} else {
-			if (AR_SREV_9285(ah) || AR_SREV_9287(ah))
-				set = AR9285_WA_DEFAULT;
-			else
-				set = AR9280_WA_DEFAULT;
-
-			/*
-			 * In AR9280 and AR9285, bit 14 in WA register
-			 * (disable L1) should only be set when device
-			 * enters D3 state and be cleared when device
-			 * comes back to D0
-			 */
-
-			if (ah->config.pcie_waen & AR_WA_D3_L1_DISABLE)
-				set |= AR_WA_D3_L1_DISABLE;
-
-			clr = AR_WA_UNTIE_RESET_EN;
-			set |= AR_WA_RESET_EN | AR_WA_POR_SHORT;
-			REG_RMW(ah, AR_WA, set, clr);
-
-			/*
-			 * for WoW sleep, we reprogram the SerDes so that the
-			 * PLL and CLK REQ are both enabled. This uses more
-			 * power but otherwise WoW sleep is unstable and the
-			 * chip may disappear.
-			 */
-
-			if (AR_SREV_9285_12_OR_LATER(ah))
-				ath9k_hw_config_serdes_wow_sleep(ah);
-
-		}
+		set = AR_WA_RESET_EN | AR_WA_POR_SHORT;
+		clr = AR_WA_UNTIE_RESET_EN | AR_WA_D3_L1_DISABLE;
+		REG_RMW(ah, AR_WA, set, clr);
 	}
 
 	/*
@@ -378,7 +294,6 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	 * Program default values for pattern backoff, aifs/slot/KAL count,
 	 * beacon miss timeout, KAL timeout, etc.
 	 */
-
 	set = AR_WOW_BACK_OFF_SHIFT(AR_WOW_PAT_BACKOFF);
 	REG_SET_BIT(ah, AR_WOW_PATTERN, set);
 
@@ -398,7 +313,7 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	/*
 	 * Keep alive timo in ms except AR9280
 	 */
-	if (!pattern_enable || AR_SREV_9280(ah))
+	if (!pattern_enable)
 		set = AR_WOW_KEEP_ALIVE_NEVER;
 	else
 		set = KAL_TIMEOUT * 32;
@@ -420,7 +335,6 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	/*
 	 * Configure MAC WoW Registers
 	 */
-
 	set = 0;
 	/* Send keep alive timeouts anyway */
 	clr = AR_WOW_KEEP_ALIVE_AUTO_DIS;
@@ -430,16 +344,9 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	else
 		set = AR_WOW_KEEP_ALIVE_FAIL_DIS;
 
-	/*
-	 * FIXME: For now disable keep alive frame
-	 * failure. This seems to sometimes trigger
-	 * unnecessary wake up with AR9485 chipsets.
-	 */
 	set = AR_WOW_KEEP_ALIVE_FAIL_DIS;
-
 	REG_RMW(ah, AR_WOW_KEEP_ALIVE, set, clr);
 
-
 	/*
 	 * we are relying on a bmiss failure. ensure we have
 	 * enough threshold to prevent false positives
@@ -473,14 +380,8 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	set |= AR_WOW_MAC_INTR_EN;
 	REG_RMW(ah, AR_WOW_PATTERN, set, clr);
 
-	/*
-	 * For AR9285 and later version of chipsets
-	 * enable WoW pattern match for packets less
-	 * than 256 bytes for all patterns
-	 */
-	if (AR_SREV_9285_12_OR_LATER(ah))
-		REG_WRITE(ah, AR_WOW_PATTERN_MATCH_LT_256B,
-			  AR_WOW_PATTERN_SUPPORTED);
+	REG_WRITE(ah, AR_WOW_PATTERN_MATCH_LT_256B,
+		  AR_WOW_PATTERN_SUPPORTED);
 
 	/*
 	 * Set the power states appropriately and enable PME
@@ -488,43 +389,32 @@ void ath9k_hw_wow_enable(struct ath_hw *ah, u32 pattern_enable)
 	clr = 0;
 	set = AR_PMCTRL_PWR_STATE_D1D3 | AR_PMCTRL_HOST_PME_EN |
 	      AR_PMCTRL_PWR_PM_CTRL_ENA;
-	/*
-	 * This is needed for AR9300 chipsets to wake-up
-	 * the host.
-	 */
-	if (AR_SREV_9300_20_OR_LATER(ah))
-		clr = AR_PCIE_PM_CTRL_ENA;
 
+	clr = AR_PCIE_PM_CTRL_ENA;
 	REG_RMW(ah, AR_PCIE_PM_CTRL, set, clr);
 
-	if (AR_SREV_9462(ah) || AR_SREV_9565(ah)) {
-		/*
-		 * this is needed to prevent the chip waking up
-		 * the host within 3-4 seconds with certain
-		 * platform/BIOS. The fix is to enable
-		 * D1 & D3 to match original definition and
-		 * also match the OTP value. Anyway this
-		 * is more related to SW WOW.
-		 */
-		clr = AR_PMCTRL_PWR_STATE_D1D3;
-		REG_CLR_BIT(ah, AR_PCIE_PM_CTRL, clr);
-
-		set = AR_PMCTRL_PWR_STATE_D1D3_REAL;
-		REG_SET_BIT(ah, AR_PCIE_PM_CTRL, set);
-	}
-
+	/*
+	 * this is needed to prevent the chip waking up
+	 * the host within 3-4 seconds with certain
+	 * platform/BIOS. The fix is to enable
+	 * D1 & D3 to match original definition and
+	 * also match the OTP value. Anyway this
+	 * is more related to SW WOW.
+	 */
+	clr = AR_PMCTRL_PWR_STATE_D1D3;
+	REG_CLR_BIT(ah, AR_PCIE_PM_CTRL, clr);
 
+	set = AR_PMCTRL_PWR_STATE_D1D3_REAL;
+	REG_SET_BIT(ah, AR_PCIE_PM_CTRL, set);
 
 	REG_CLR_BIT(ah, AR_STA_ID1, AR_STA_ID1_PRESERVE_SEQNUM);
 
-	if (AR_SREV_9300_20_OR_LATER(ah)) {
-		/* to bring down WOW power low margin */
-		set = BIT(13);
-		REG_SET_BIT(ah, AR_PCIE_PHY_REG3, set);
-		/* HW WoW */
-		clr = BIT(5);
-		REG_CLR_BIT(ah, AR_PCU_MISC_MODE3, clr);
-	}
+	/* to bring down WOW power low margin */
+	set = BIT(13);
+	REG_SET_BIT(ah, AR_PCIE_PHY_REG3, set);
+	/* HW WoW */
+	clr = BIT(5);
+	REG_CLR_BIT(ah, AR_PCU_MISC_MODE3, clr);
 
 	ath9k_hw_set_powermode_wow_sleep(ah);
 	ah->wow_event_mask = wow_event_mask;
diff --git a/drivers/net/wireless/ath/ath9k/xmit.c b/drivers/net/wireless/ath/ath9k/xmit.c
index 83ab6be..c59ae43 100644
--- a/drivers/net/wireless/ath/ath9k/xmit.c
+++ b/drivers/net/wireless/ath/ath9k/xmit.c
@@ -518,6 +518,10 @@ static void ath_tx_complete_aggr(struct ath_softc *sc, struct ath_txq *txq,
 			ath_tx_complete_buf(sc, bf, txq, &bf_head, ts,
 				!txfail);
 		} else {
+			if (tx_info->flags & IEEE80211_TX_STATUS_EOSP) {
+				tx_info->flags &= ~IEEE80211_TX_STATUS_EOSP;
+				ieee80211_sta_eosp(sta);
+			}
 			/* retry the un-acked ones */
 			if (bf->bf_next == NULL && bf_last->bf_stale) {
 				struct ath_buf *tbf;
@@ -786,25 +790,20 @@ static int ath_compute_num_delims(struct ath_softc *sc, struct ath_atx_tid *tid,
 	return ndelim;
 }
 
-static enum ATH_AGGR_STATUS ath_tx_form_aggr(struct ath_softc *sc,
-					     struct ath_txq *txq,
-					     struct ath_atx_tid *tid,
-					     struct list_head *bf_q,
-					     int *aggr_len)
+static struct ath_buf *
+ath_tx_get_tid_subframe(struct ath_softc *sc, struct ath_txq *txq,
+			struct ath_atx_tid *tid)
 {
-#define PADBYTES(_len) ((4 - ((_len) % 4)) % 4)
-	struct ath_buf *bf, *bf_first = NULL, *bf_prev = NULL;
-	int rl = 0, nframes = 0, ndelim, prev_al = 0;
-	u16 aggr_limit = 0, al = 0, bpad = 0,
-		al_delta, h_baw = tid->baw_size / 2;
-	enum ATH_AGGR_STATUS status = ATH_AGGR_DONE;
-	struct ieee80211_tx_info *tx_info;
 	struct ath_frame_info *fi;
 	struct sk_buff *skb;
+	struct ath_buf *bf;
 	u16 seqno;
 
-	do {
+	while (1) {
 		skb = skb_peek(&tid->buf_q);
+		if (!skb)
+			break;
+
 		fi = get_frame_info(skb);
 		bf = fi->bf;
 		if (!fi->bf)
@@ -820,10 +819,8 @@ static enum ATH_AGGR_STATUS ath_tx_form_aggr(struct ath_softc *sc,
 		seqno = bf->bf_state.seqno;
 
 		/* do not step over block-ack window */
-		if (!BAW_WITHIN(tid->seq_start, tid->baw_size, seqno)) {
-			status = ATH_AGGR_BAW_CLOSED;
+		if (!BAW_WITHIN(tid->seq_start, tid->baw_size, seqno))
 			break;
-		}
 
 		if (tid->bar_index > ATH_BA_INDEX(tid->seq_start, seqno)) {
 			struct ath_tx_status ts = {};
@@ -837,6 +834,40 @@ static enum ATH_AGGR_STATUS ath_tx_form_aggr(struct ath_softc *sc,
 			continue;
 		}
 
+		bf->bf_next = NULL;
+		bf->bf_lastbf = bf;
+		return bf;
+	}
+
+	return NULL;
+}
+
+static enum ATH_AGGR_STATUS ath_tx_form_aggr(struct ath_softc *sc,
+					     struct ath_txq *txq,
+					     struct ath_atx_tid *tid,
+					     struct list_head *bf_q,
+					     int *aggr_len)
+{
+#define PADBYTES(_len) ((4 - ((_len) % 4)) % 4)
+	struct ath_buf *bf, *bf_first = NULL, *bf_prev = NULL;
+	int rl = 0, nframes = 0, ndelim, prev_al = 0;
+	u16 aggr_limit = 0, al = 0, bpad = 0,
+		al_delta, h_baw = tid->baw_size / 2;
+	enum ATH_AGGR_STATUS status = ATH_AGGR_DONE;
+	struct ieee80211_tx_info *tx_info;
+	struct ath_frame_info *fi;
+	struct sk_buff *skb;
+
+	do {
+		bf = ath_tx_get_tid_subframe(sc, txq, tid);
+		if (!bf) {
+			status = ATH_AGGR_BAW_CLOSED;
+			break;
+		}
+
+		skb = bf->bf_mpdu;
+		fi = get_frame_info(skb);
+
 		if (!bf_first)
 			bf_first = bf;
 
@@ -882,7 +913,7 @@ static enum ATH_AGGR_STATUS ath_tx_form_aggr(struct ath_softc *sc,
 
 		/* link buffers of this frame to the aggregate */
 		if (!fi->retries)
-			ath_tx_addto_baw(sc, tid, seqno);
+			ath_tx_addto_baw(sc, tid, bf->bf_state.seqno);
 		bf->bf_state.ndelim = ndelim;
 
 		__skb_unlink(skb, &tid->buf_q);
@@ -1090,10 +1121,8 @@ static void ath_tx_fill_desc(struct ath_softc *sc, struct ath_buf *bf,
 			     struct ath_txq *txq, int len)
 {
 	struct ath_hw *ah = sc->sc_ah;
-	struct ieee80211_tx_info *tx_info = IEEE80211_SKB_CB(bf->bf_mpdu);
-	struct ath_buf *bf_first = bf;
+	struct ath_buf *bf_first = NULL;
 	struct ath_tx_info info;
-	bool aggr = !!(bf->bf_state.bf_type & BUF_AGGR);
 
 	memset(&info, 0, sizeof(info));
 	info.is_first = true;
@@ -1101,24 +1130,11 @@ static void ath_tx_fill_desc(struct ath_softc *sc, struct ath_buf *bf,
 	info.txpower = MAX_RATE_POWER;
 	info.qcu = txq->axq_qnum;
 
-	info.flags = ATH9K_TXDESC_INTREQ;
-	if (tx_info->flags & IEEE80211_TX_CTL_NO_ACK)
-		info.flags |= ATH9K_TXDESC_NOACK;
-	if (tx_info->flags & IEEE80211_TX_CTL_LDPC)
-		info.flags |= ATH9K_TXDESC_LDPC;
-
-	ath_buf_set_rate(sc, bf, &info, len);
-
-	if (tx_info->flags & IEEE80211_TX_CTL_CLEAR_PS_FILT)
-		info.flags |= ATH9K_TXDESC_CLRDMASK;
-
-	if (bf->bf_state.bfs_paprd)
-		info.flags |= (u32) bf->bf_state.bfs_paprd << ATH9K_TXDESC_PAPRD_S;
-
-
 	while (bf) {
 		struct sk_buff *skb = bf->bf_mpdu;
+		struct ieee80211_tx_info *tx_info = IEEE80211_SKB_CB(skb);
 		struct ath_frame_info *fi = get_frame_info(skb);
+		bool aggr = !!(bf->bf_state.bf_type & BUF_AGGR);
 
 		info.type = get_hw_packet_type(skb);
 		if (bf->bf_next)
@@ -1126,6 +1142,26 @@ static void ath_tx_fill_desc(struct ath_softc *sc, struct ath_buf *bf,
 		else
 			info.link = 0;
 
+		if (!bf_first) {
+			bf_first = bf;
+
+			info.flags = ATH9K_TXDESC_INTREQ;
+			if ((tx_info->flags & IEEE80211_TX_CTL_CLEAR_PS_FILT) ||
+			    txq == sc->tx.uapsdq)
+				info.flags |= ATH9K_TXDESC_CLRDMASK;
+
+			if (tx_info->flags & IEEE80211_TX_CTL_NO_ACK)
+				info.flags |= ATH9K_TXDESC_NOACK;
+			if (tx_info->flags & IEEE80211_TX_CTL_LDPC)
+				info.flags |= ATH9K_TXDESC_LDPC;
+
+			if (bf->bf_state.bfs_paprd)
+				info.flags |= (u32) bf->bf_state.bfs_paprd <<
+					      ATH9K_TXDESC_PAPRD_S;
+
+			ath_buf_set_rate(sc, bf, &info, len);
+		}
+
 		info.buf_addr[0] = bf->bf_buf_addr;
 		info.buf_len[0] = skb->len;
 		info.pkt_len = fi->framelen;
@@ -1135,7 +1171,7 @@ static void ath_tx_fill_desc(struct ath_softc *sc, struct ath_buf *bf,
 		if (aggr) {
 			if (bf == bf_first)
 				info.aggr = AGGR_BUF_FIRST;
-			else if (!bf->bf_next)
+			else if (bf == bf_first->bf_lastbf)
 				info.aggr = AGGR_BUF_LAST;
 			else
 				info.aggr = AGGR_BUF_MIDDLE;
@@ -1144,6 +1180,9 @@ static void ath_tx_fill_desc(struct ath_softc *sc, struct ath_buf *bf,
 			info.aggr_len = len;
 		}
 
+		if (bf == bf_first->bf_lastbf)
+			bf_first = NULL;
+
 		ath9k_hw_set_txdesc(ah, bf->bf_desc, &info);
 		bf = bf->bf_next;
 	}
@@ -1328,6 +1367,70 @@ void ath_tx_aggr_resume(struct ath_softc *sc, struct ieee80211_sta *sta,
 	ath_txq_unlock_complete(sc, txq);
 }
 
+void ath9k_release_buffered_frames(struct ieee80211_hw *hw,
+				   struct ieee80211_sta *sta,
+				   u16 tids, int nframes,
+				   enum ieee80211_frame_release_type reason,
+				   bool more_data)
+{
+	struct ath_softc *sc = hw->priv;
+	struct ath_node *an = (struct ath_node *)sta->drv_priv;
+	struct ath_txq *txq = sc->tx.uapsdq;
+	struct ieee80211_tx_info *info;
+	struct list_head bf_q;
+	struct ath_buf *bf_tail = NULL, *bf;
+	int sent = 0;
+	int i;
+
+	INIT_LIST_HEAD(&bf_q);
+	for (i = 0; tids && nframes; i++, tids >>= 1) {
+		struct ath_atx_tid *tid;
+
+		if (!(tids & 1))
+			continue;
+
+		tid = ATH_AN_2_TID(an, i);
+		if (tid->paused)
+			continue;
+
+		ath_txq_lock(sc, tid->ac->txq);
+		while (!skb_queue_empty(&tid->buf_q) && nframes > 0) {
+			bf = ath_tx_get_tid_subframe(sc, sc->tx.uapsdq, tid);
+			if (!bf)
+				break;
+
+			__skb_unlink(bf->bf_mpdu, &tid->buf_q);
+			list_add_tail(&bf->list, &bf_q);
+			ath_set_rates(tid->an->vif, tid->an->sta, bf);
+			ath_tx_addto_baw(sc, tid, bf->bf_state.seqno);
+			bf->bf_state.bf_type &= ~BUF_AGGR;
+			if (bf_tail)
+				bf_tail->bf_next = bf;
+
+			bf_tail = bf;
+			nframes--;
+			sent++;
+			TX_STAT_INC(txq->axq_qnum, a_queued_hw);
+
+			if (skb_queue_empty(&tid->buf_q))
+				ieee80211_sta_set_buffered(an->sta, i, false);
+		}
+		ath_txq_unlock_complete(sc, tid->ac->txq);
+	}
+
+	if (list_empty(&bf_q))
+		return;
+
+	info = IEEE80211_SKB_CB(bf_tail->bf_mpdu);
+	info->flags |= IEEE80211_TX_STATUS_EOSP;
+
+	bf = list_first_entry(&bf_q, struct ath_buf, list);
+	ath_txq_lock(sc, txq);
+	ath_tx_fill_desc(sc, bf, txq, 0);
+	ath_tx_txqaddbuf(sc, txq, &bf_q, false);
+	ath_txq_unlock(sc, txq);
+}
+
 /********************/
 /* Queue Management */
 /********************/
@@ -1679,14 +1782,19 @@ static void ath_tx_txqaddbuf(struct ath_softc *sc, struct ath_txq *txq,
 	}
 
 	if (!internal) {
-		txq->axq_depth++;
-		if (bf_is_ampdu_not_probing(bf))
-			txq->axq_ampdu_depth++;
+		while (bf) {
+			txq->axq_depth++;
+			if (bf_is_ampdu_not_probing(bf))
+				txq->axq_ampdu_depth++;
+
+			bf = bf->bf_lastbf->bf_next;
+		}
 	}
 }
 
-static void ath_tx_send_ampdu(struct ath_softc *sc, struct ath_atx_tid *tid,
-			      struct sk_buff *skb, struct ath_tx_control *txctl)
+static void ath_tx_send_ampdu(struct ath_softc *sc, struct ath_txq *txq,
+			      struct ath_atx_tid *tid, struct sk_buff *skb,
+			      struct ath_tx_control *txctl)
 {
 	struct ath_frame_info *fi = get_frame_info(skb);
 	struct list_head bf_head;
@@ -1699,21 +1807,22 @@ static void ath_tx_send_ampdu(struct ath_softc *sc, struct ath_atx_tid *tid,
 	 * - seqno is not within block-ack window
 	 * - h/w queue depth exceeds low water mark
 	 */
-	if (!skb_queue_empty(&tid->buf_q) || tid->paused ||
-	    !BAW_WITHIN(tid->seq_start, tid->baw_size, tid->seq_next) ||
-	    txctl->txq->axq_ampdu_depth >= ATH_AGGR_MIN_QDEPTH) {
+	if ((!skb_queue_empty(&tid->buf_q) || tid->paused ||
+	     !BAW_WITHIN(tid->seq_start, tid->baw_size, tid->seq_next) ||
+	     txq->axq_ampdu_depth >= ATH_AGGR_MIN_QDEPTH) &&
+	    txq != sc->tx.uapsdq) {
 		/*
 		 * Add this frame to software queue for scheduling later
 		 * for aggregation.
 		 */
-		TX_STAT_INC(txctl->txq->axq_qnum, a_queued_sw);
+		TX_STAT_INC(txq->axq_qnum, a_queued_sw);
 		__skb_queue_tail(&tid->buf_q, skb);
 		if (!txctl->an || !txctl->an->sleeping)
-			ath_tx_queue_tid(txctl->txq, tid);
+			ath_tx_queue_tid(txq, tid);
 		return;
 	}
 
-	bf = ath_tx_setup_buffer(sc, txctl->txq, tid, skb);
+	bf = ath_tx_setup_buffer(sc, txq, tid, skb);
 	if (!bf) {
 		ieee80211_free_txskb(sc->hw, skb);
 		return;
@@ -1728,10 +1837,10 @@ static void ath_tx_send_ampdu(struct ath_softc *sc, struct ath_atx_tid *tid,
 	ath_tx_addto_baw(sc, tid, bf->bf_state.seqno);
 
 	/* Queue to h/w without aggregation */
-	TX_STAT_INC(txctl->txq->axq_qnum, a_queued_hw);
+	TX_STAT_INC(txq->axq_qnum, a_queued_hw);
 	bf->bf_lastbf = bf;
-	ath_tx_fill_desc(sc, bf, txctl->txq, fi->framelen);
-	ath_tx_txqaddbuf(sc, txctl->txq, &bf_head, false);
+	ath_tx_fill_desc(sc, bf, txq, fi->framelen);
+	ath_tx_txqaddbuf(sc, txq, &bf_head, false);
 }
 
 static void ath_tx_send_normal(struct ath_softc *sc, struct ath_txq *txq,
@@ -1869,22 +1978,16 @@ static struct ath_buf *ath_tx_setup_buffer(struct ath_softc *sc,
 	return bf;
 }
 
-/* Upon failure caller should free skb */
-int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
-		 struct ath_tx_control *txctl)
+static int ath_tx_prepare(struct ieee80211_hw *hw, struct sk_buff *skb,
+			  struct ath_tx_control *txctl)
 {
 	struct ieee80211_hdr *hdr = (struct ieee80211_hdr *) skb->data;
 	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
 	struct ieee80211_sta *sta = txctl->sta;
 	struct ieee80211_vif *vif = info->control.vif;
 	struct ath_softc *sc = hw->priv;
-	struct ath_txq *txq = txctl->txq;
-	struct ath_atx_tid *tid = NULL;
-	struct ath_buf *bf;
-	int padpos, padsize;
 	int frmlen = skb->len + FCS_LEN;
-	u8 tidno;
-	int q;
+	int padpos, padsize;
 
 	/* NOTE:  sta can be NULL according to net/mac80211.h */
 	if (sta)
@@ -1905,6 +2008,11 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 		hdr->seq_ctrl |= cpu_to_le16(sc->tx.seq_no);
 	}
 
+	if ((vif && vif->type != NL80211_IFTYPE_AP &&
+	            vif->type != NL80211_IFTYPE_AP_VLAN) ||
+	    !ieee80211_is_data(hdr->frame_control))
+		info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT;
+
 	/* Add the padding after the header if this is not already done */
 	padpos = ieee80211_hdrlen(hdr->frame_control);
 	padsize = padpos & 3;
@@ -1914,16 +2022,34 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 
 		skb_push(skb, padsize);
 		memmove(skb->data, skb->data + padsize, padpos);
-		hdr = (struct ieee80211_hdr *) skb->data;
 	}
 
-	if ((vif && vif->type != NL80211_IFTYPE_AP &&
-	            vif->type != NL80211_IFTYPE_AP_VLAN) ||
-	    !ieee80211_is_data(hdr->frame_control))
-		info->flags |= IEEE80211_TX_CTL_CLEAR_PS_FILT;
-
 	setup_frame_info(hw, sta, skb, frmlen);
+	return 0;
+}
+
+
+/* Upon failure caller should free skb */
+int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
+		 struct ath_tx_control *txctl)
+{
+	struct ieee80211_hdr *hdr;
+	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
+	struct ieee80211_sta *sta = txctl->sta;
+	struct ieee80211_vif *vif = info->control.vif;
+	struct ath_softc *sc = hw->priv;
+	struct ath_txq *txq = txctl->txq;
+	struct ath_atx_tid *tid = NULL;
+	struct ath_buf *bf;
+	u8 tidno;
+	int q;
+	int ret;
+
+	ret = ath_tx_prepare(hw, skb, txctl);
+	if (ret)
+	    return ret;
 
+	hdr = (struct ieee80211_hdr *) skb->data;
 	/*
 	 * At this point, the vif, hw_key and sta pointers in the tx control
 	 * info are no longer valid (overwritten by the ath_frame_info data.
@@ -1939,6 +2065,12 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 		txq->stopped = true;
 	}
 
+	if (info->flags & IEEE80211_TX_CTL_PS_RESPONSE) {
+		ath_txq_unlock(sc, txq);
+		txq = sc->tx.uapsdq;
+		ath_txq_lock(sc, txq);
+	}
+
 	if (txctl->an && ieee80211_is_data_qos(hdr->frame_control)) {
 		tidno = ieee80211_get_qos_ctl(hdr)[0] &
 			IEEE80211_QOS_CTL_TID_MASK;
@@ -1952,11 +2084,11 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 		 * Try aggregation if it's a unicast data frame
 		 * and the destination is HT capable.
 		 */
-		ath_tx_send_ampdu(sc, tid, skb, txctl);
+		ath_tx_send_ampdu(sc, txq, tid, skb, txctl);
 		goto out;
 	}
 
-	bf = ath_tx_setup_buffer(sc, txctl->txq, tid, skb);
+	bf = ath_tx_setup_buffer(sc, txq, tid, skb);
 	if (!bf) {
 		if (txctl->paprd)
 			dev_kfree_skb_any(skb);
@@ -1971,7 +2103,7 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 		bf->bf_state.bfs_paprd_timestamp = jiffies;
 
 	ath_set_rates(vif, sta, bf);
-	ath_tx_send_normal(sc, txctl->txq, tid, skb);
+	ath_tx_send_normal(sc, txq, tid, skb);
 
 out:
 	ath_txq_unlock(sc, txq);
@@ -1979,6 +2111,74 @@ out:
 	return 0;
 }
 
+void ath_tx_cabq(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
+		 struct sk_buff *skb)
+{
+	struct ath_softc *sc = hw->priv;
+	struct ath_tx_control txctl = {
+		.txq = sc->beacon.cabq
+	};
+	struct ath_tx_info info = {};
+	struct ieee80211_hdr *hdr;
+	struct ath_buf *bf_tail = NULL;
+	struct ath_buf *bf;
+	LIST_HEAD(bf_q);
+	int duration = 0;
+	int max_duration;
+
+	max_duration =
+		sc->cur_beacon_conf.beacon_interval * 1000 *
+		sc->cur_beacon_conf.dtim_period / ATH_BCBUF;
+
+	do {
+		struct ath_frame_info *fi = get_frame_info(skb);
+
+		if (ath_tx_prepare(hw, skb, &txctl))
+			break;
+
+		bf = ath_tx_setup_buffer(sc, txctl.txq, NULL, skb);
+		if (!bf)
+			break;
+
+		bf->bf_lastbf = bf;
+		ath_set_rates(vif, NULL, bf);
+		ath_buf_set_rate(sc, bf, &info, fi->framelen);
+		duration += info.rates[0].PktDuration;
+		if (bf_tail)
+			bf_tail->bf_next = bf;
+
+		list_add_tail(&bf->list, &bf_q);
+		bf_tail = bf;
+		skb = NULL;
+
+		if (duration > max_duration)
+			break;
+
+		skb = ieee80211_get_buffered_bc(hw, vif);
+	} while(skb);
+
+	if (skb)
+		ieee80211_free_txskb(hw, skb);
+
+	if (list_empty(&bf_q))
+		return;
+
+	bf = list_first_entry(&bf_q, struct ath_buf, list);
+	hdr = (struct ieee80211_hdr *) bf->bf_mpdu->data;
+
+	if (hdr->frame_control & IEEE80211_FCTL_MOREDATA) {
+		hdr->frame_control &= ~IEEE80211_FCTL_MOREDATA;
+		dma_sync_single_for_device(sc->dev, bf->bf_buf_addr,
+			sizeof(*hdr), DMA_TO_DEVICE);
+	}
+
+	ath_txq_lock(sc, txctl.txq);
+	ath_tx_fill_desc(sc, bf, txctl.txq, 0);
+	ath_tx_txqaddbuf(sc, txctl.txq, &bf_q, false);
+	TX_STAT_INC(txctl.txq->axq_qnum, queued);
+	ath_txq_unlock(sc, txctl.txq);
+}
+
 /*****************/
 /* TX Completion */
 /*****************/
@@ -2024,7 +2224,12 @@ static void ath_tx_complete(struct ath_softc *sc, struct sk_buff *skb,
 	}
 	spin_unlock_irqrestore(&sc->sc_pm_lock, flags);
 
+	__skb_queue_tail(&txq->complete_q, skb);
+
 	q = skb_get_queue_mapping(skb);
+	if (txq == sc->tx.uapsdq)
+		txq = sc->tx.txq_map[q];
+
 	if (txq == sc->tx.txq_map[q]) {
 		if (WARN_ON(--txq->pending_frames < 0))
 			txq->pending_frames = 0;
@@ -2035,8 +2240,6 @@ static void ath_tx_complete(struct ath_softc *sc, struct sk_buff *skb,
 			txq->stopped = false;
 		}
 	}
-
-	__skb_queue_tail(&txq->complete_q, skb);
 }
 
 static void ath_tx_complete_buf(struct ath_softc *sc, struct ath_buf *bf,
diff --git a/drivers/net/wireless/ath/carl9170/carl9170.h b/drivers/net/wireless/ath/carl9170/carl9170.h
index 9dce106..8596aba 100644
--- a/drivers/net/wireless/ath/carl9170/carl9170.h
+++ b/drivers/net/wireless/ath/carl9170/carl9170.h
@@ -133,6 +133,9 @@ struct carl9170_sta_tid {
 
 	/* Preaggregation reorder queue */
 	struct sk_buff_head queue;
+
+	struct ieee80211_sta *sta;
+	struct ieee80211_vif *vif;
 };
 
 #define CARL9170_QUEUE_TIMEOUT		256
diff --git a/drivers/net/wireless/ath/carl9170/main.c b/drivers/net/wireless/ath/carl9170/main.c
index e9010a4..4a33c6e 100644
--- a/drivers/net/wireless/ath/carl9170/main.c
+++ b/drivers/net/wireless/ath/carl9170/main.c
@@ -1448,6 +1448,8 @@ static int carl9170_op_ampdu_action(struct ieee80211_hw *hw,
 		tid_info->state = CARL9170_TID_STATE_PROGRESS;
 		tid_info->tid = tid;
 		tid_info->max = sta_info->ampdu_max_len;
+		tid_info->sta = sta;
+		tid_info->vif = vif;
 
 		INIT_LIST_HEAD(&tid_info->list);
 		INIT_LIST_HEAD(&tid_info->tmp_list);
@@ -1857,6 +1859,7 @@ void *carl9170_alloc(size_t priv_size)
 		     IEEE80211_HW_SUPPORTS_PS |
 		     IEEE80211_HW_PS_NULLFUNC_STACK |
 		     IEEE80211_HW_NEED_DTIM_BEFORE_ASSOC |
+		     IEEE80211_HW_SUPPORTS_RC_TABLE |
 		     IEEE80211_HW_SIGNAL_DBM;
 
 	if (!modparam_noht) {
diff --git a/drivers/net/wireless/ath/carl9170/tx.c b/drivers/net/wireless/ath/carl9170/tx.c
index c61cafa..e3f696e 100644
--- a/drivers/net/wireless/ath/carl9170/tx.c
+++ b/drivers/net/wireless/ath/carl9170/tx.c
@@ -625,7 +625,7 @@ static void carl9170_tx_ampdu_timeout(struct ar9170 *ar)
 		    msecs_to_jiffies(CARL9170_QUEUE_TIMEOUT)))
 			goto unlock;
 
-		sta = __carl9170_get_tx_sta(ar, skb);
+		sta = iter->sta;
 		if (WARN_ON(!sta))
 			goto unlock;
 
@@ -866,6 +866,93 @@ static bool carl9170_tx_cts_check(struct ar9170 *ar,
 	return false;
 }
 
+static void carl9170_tx_get_rates(struct ar9170 *ar,
+				  struct ieee80211_vif *vif,
+				  struct ieee80211_sta *sta,
+				  struct sk_buff *skb)
+{
+	struct ieee80211_tx_info *info;
+
+	BUILD_BUG_ON(IEEE80211_TX_MAX_RATES < CARL9170_TX_MAX_RATES);
+	BUILD_BUG_ON(IEEE80211_TX_MAX_RATES > IEEE80211_TX_RATE_TABLE_SIZE);
+
+	info = IEEE80211_SKB_CB(skb);
+
+	ieee80211_get_tx_rates(vif, sta, skb,
+			       info->control.rates,
+			       IEEE80211_TX_MAX_RATES);
+}
+
+static void carl9170_tx_apply_rateset(struct ar9170 *ar,
+				      struct ieee80211_tx_info *sinfo,
+				      struct sk_buff *skb)
+{
+	struct ieee80211_tx_rate *txrate;
+	struct ieee80211_tx_info *info;
+	struct _carl9170_tx_superframe *txc = (void *) skb->data;
+	int i;
+	bool ampdu;
+	bool no_ack;
+
+	info = IEEE80211_SKB_CB(skb);
+	ampdu = !!(info->flags & IEEE80211_TX_CTL_AMPDU);
+	no_ack = !!(info->flags & IEEE80211_TX_CTL_NO_ACK);
+
+	/* Set the rate control probe flag for all (sub-) frames.
+	 * This is because the TX_STATS_AMPDU flag is only set on
+	 * the last frame, so it has to be inherited.
+	 */
+	info->flags |= (sinfo->flags & IEEE80211_TX_CTL_RATE_CTRL_PROBE);
+
+	/* NOTE: For the first rate, the ERP & AMPDU flags are directly
+	 * taken from mac_control. For all fallback rate, the firmware
+	 * updates the mac_control flags from the rate info field.
+	 */
+	for (i = 0; i < CARL9170_TX_MAX_RATES; i++) {
+		__le32 phy_set;
+
+		txrate = &sinfo->control.rates[i];
+		if (txrate->idx < 0)
+			break;
+
+		phy_set = carl9170_tx_physet(ar, info, txrate);
+		if (i == 0) {
+			__le16 mac_tmp = cpu_to_le16(0);
+
+			/* first rate - part of the hw's frame header */
+			txc->f.phy_control = phy_set;
+
+			if (ampdu && txrate->flags & IEEE80211_TX_RC_MCS)
+				mac_tmp |= cpu_to_le16(AR9170_TX_MAC_AGGR);
+
+			if (carl9170_tx_rts_check(ar, txrate, ampdu, no_ack))
+				mac_tmp |= cpu_to_le16(AR9170_TX_MAC_PROT_RTS);
+			else if (carl9170_tx_cts_check(ar, txrate))
+				mac_tmp |= cpu_to_le16(AR9170_TX_MAC_PROT_CTS);
+
+			txc->f.mac_control |= mac_tmp;
+		} else {
+			/* fallback rates are stored in the firmware's
+			 * retry rate set array.
+			 */
+			txc->s.rr[i - 1] = phy_set;
+		}
+
+		SET_VAL(CARL9170_TX_SUPER_RI_TRIES, txc->s.ri[i],
+			txrate->count);
+
+		if (carl9170_tx_rts_check(ar, txrate, ampdu, no_ack))
+			txc->s.ri[i] |= (AR9170_TX_MAC_PROT_RTS <<
+				CARL9170_TX_SUPER_RI_ERP_PROT_S);
+		else if (carl9170_tx_cts_check(ar, txrate))
+			txc->s.ri[i] |= (AR9170_TX_MAC_PROT_CTS <<
+				CARL9170_TX_SUPER_RI_ERP_PROT_S);
+
+		if (ampdu && (txrate->flags & IEEE80211_TX_RC_MCS))
+			txc->s.ri[i] |= CARL9170_TX_SUPER_RI_AMPDU;
+	}
+}
+
 static int carl9170_tx_prepare(struct ar9170 *ar,
 			       struct ieee80211_sta *sta,
 			       struct sk_buff *skb)
@@ -874,13 +961,10 @@ static int carl9170_tx_prepare(struct ar9170 *ar,
 	struct _carl9170_tx_superframe *txc;
 	struct carl9170_vif_info *cvif;
 	struct ieee80211_tx_info *info;
-	struct ieee80211_tx_rate *txrate;
 	struct carl9170_tx_info *arinfo;
 	unsigned int hw_queue;
-	int i;
 	__le16 mac_tmp;
 	u16 len;
-	bool ampdu, no_ack;
 
 	BUILD_BUG_ON(sizeof(*arinfo) > sizeof(info->rate_driver_data));
 	BUILD_BUG_ON(sizeof(struct _carl9170_tx_superdesc) !=
@@ -889,8 +973,6 @@ static int carl9170_tx_prepare(struct ar9170 *ar,
 	BUILD_BUG_ON(sizeof(struct _ar9170_tx_hwdesc) !=
 		     AR9170_TX_HWDESC_LEN);
 
-	BUILD_BUG_ON(IEEE80211_TX_MAX_RATES < CARL9170_TX_MAX_RATES);
-
 	BUILD_BUG_ON(AR9170_MAX_VIRTUAL_MAC >
 		((CARL9170_TX_SUPER_MISC_VIF_ID >>
 		 CARL9170_TX_SUPER_MISC_VIF_ID_S) + 1));
@@ -932,8 +1014,7 @@ static int carl9170_tx_prepare(struct ar9170 *ar,
 	mac_tmp |= cpu_to_le16((hw_queue << AR9170_TX_MAC_QOS_S) &
 			       AR9170_TX_MAC_QOS);
 
-	no_ack = !!(info->flags & IEEE80211_TX_CTL_NO_ACK);
-	if (unlikely(no_ack))
+	if (unlikely(info->flags & IEEE80211_TX_CTL_NO_ACK))
 		mac_tmp |= cpu_to_le16(AR9170_TX_MAC_NO_ACK);
 
 	if (info->control.hw_key) {
@@ -954,8 +1035,7 @@ static int carl9170_tx_prepare(struct ar9170 *ar,
 		}
 	}
 
-	ampdu = !!(info->flags & IEEE80211_TX_CTL_AMPDU);
-	if (ampdu) {
+	if (info->flags & IEEE80211_TX_CTL_AMPDU) {
 		unsigned int density, factor;
 
 		if (unlikely(!sta || !cvif))
@@ -982,50 +1062,6 @@ static int carl9170_tx_prepare(struct ar9170 *ar,
 			txc->s.ampdu_settings, factor);
 	}
 
-	/*
-	 * NOTE: For the first rate, the ERP & AMPDU flags are directly
-	 * taken from mac_control. For all fallback rate, the firmware
-	 * updates the mac_control flags from the rate info field.
-	 */
-	for (i = 0; i < CARL9170_TX_MAX_RATES; i++) {
-		__le32 phy_set;
-		txrate = &info->control.rates[i];
-		if (txrate->idx < 0)
-			break;
-
-		phy_set = carl9170_tx_physet(ar, info, txrate);
-		if (i == 0) {
-			/* first rate - part of the hw's frame header */
-			txc->f.phy_control = phy_set;
-
-			if (ampdu && txrate->flags & IEEE80211_TX_RC_MCS)
-				mac_tmp |= cpu_to_le16(AR9170_TX_MAC_AGGR);
-			if (carl9170_tx_rts_check(ar, txrate, ampdu, no_ack))
-				mac_tmp |= cpu_to_le16(AR9170_TX_MAC_PROT_RTS);
-			else if (carl9170_tx_cts_check(ar, txrate))
-				mac_tmp |= cpu_to_le16(AR9170_TX_MAC_PROT_CTS);
-
-		} else {
-			/* fallback rates are stored in the firmware's
-			 * retry rate set array.
-			 */
-			txc->s.rr[i - 1] = phy_set;
-		}
-
-		SET_VAL(CARL9170_TX_SUPER_RI_TRIES, txc->s.ri[i],
-			txrate->count);
-
-		if (carl9170_tx_rts_check(ar, txrate, ampdu, no_ack))
-			txc->s.ri[i] |= (AR9170_TX_MAC_PROT_RTS <<
-				CARL9170_TX_SUPER_RI_ERP_PROT_S);
-		else if (carl9170_tx_cts_check(ar, txrate))
-			txc->s.ri[i] |= (AR9170_TX_MAC_PROT_CTS <<
-				CARL9170_TX_SUPER_RI_ERP_PROT_S);
-
-		if (ampdu && (txrate->flags & IEEE80211_TX_RC_MCS))
-			txc->s.ri[i] |= CARL9170_TX_SUPER_RI_AMPDU;
-	}
-
 	txc->s.len = cpu_to_le16(skb->len);
 	txc->f.length = cpu_to_le16(len + FCS_LEN);
 	txc->f.mac_control = mac_tmp;
@@ -1086,31 +1122,12 @@ static void carl9170_set_ampdu_params(struct ar9170 *ar, struct sk_buff *skb)
 	}
 }
 
-static bool carl9170_tx_rate_check(struct ar9170 *ar, struct sk_buff *_dest,
-				   struct sk_buff *_src)
-{
-	struct _carl9170_tx_superframe *dest, *src;
-
-	dest = (void *) _dest->data;
-	src = (void *) _src->data;
-
-	/*
-	 * The mac80211 rate control algorithm expects that all MPDUs in
-	 * an AMPDU share the same tx vectors.
-	 * This is not really obvious right now, because the hardware
-	 * does the AMPDU setup according to its own rulebook.
-	 * Our nicely assembled, strictly monotonic increasing mpdu
-	 * chains will be broken up, mashed back together...
-	 */
-
-	return (dest->f.phy_control == src->f.phy_control);
-}
-
 static void carl9170_tx_ampdu(struct ar9170 *ar)
 {
 	struct sk_buff_head agg;
 	struct carl9170_sta_tid *tid_info;
 	struct sk_buff *skb, *first;
+	struct ieee80211_tx_info *tx_info_first;
 	unsigned int i = 0, done_ampdus = 0;
 	u16 seq, queue, tmpssn;
 
@@ -1156,6 +1173,7 @@ retry:
 			goto processed;
 		}
 
+		tx_info_first = NULL;
 		while ((skb = skb_peek(&tid_info->queue))) {
 			/* strict 0, 1, ..., n - 1, n frame sequence order */
 			if (unlikely(carl9170_get_seq(skb) != seq))
@@ -1166,8 +1184,13 @@ retry:
 			    (tid_info->max - 1)))
 				break;
 
-			if (!carl9170_tx_rate_check(ar, skb, first))
-				break;
+			if (!tx_info_first) {
+				carl9170_tx_get_rates(ar, tid_info->vif,
+						      tid_info->sta, first);
+				tx_info_first = IEEE80211_SKB_CB(first);
+			}
+
+			carl9170_tx_apply_rateset(ar, tx_info_first, skb);
 
 			atomic_inc(&ar->tx_ampdu_upload);
 			tid_info->snx = seq = SEQ_NEXT(seq);
@@ -1182,8 +1205,7 @@ retry:
 		if (skb_queue_empty(&tid_info->queue) ||
 		    carl9170_get_seq(skb_peek(&tid_info->queue)) !=
 		    tid_info->snx) {
-			/*
-			 * stop TID, if A-MPDU frames are still missing,
+			/* stop TID, if A-MPDU frames are still missing,
 			 * or whenever the queue is empty.
 			 */
 
@@ -1450,12 +1472,14 @@ void carl9170_op_tx(struct ieee80211_hw *hw,
 	struct ar9170 *ar = hw->priv;
 	struct ieee80211_tx_info *info;
 	struct ieee80211_sta *sta = control->sta;
+	struct ieee80211_vif *vif;
 	bool run;
 
 	if (unlikely(!IS_STARTED(ar)))
 		goto err_free;
 
 	info = IEEE80211_SKB_CB(skb);
+	vif = info->control.vif;
 
 	if (unlikely(carl9170_tx_prepare(ar, sta, skb)))
 		goto err_free;
@@ -1486,6 +1510,8 @@ void carl9170_op_tx(struct ieee80211_hw *hw,
 	} else {
 		unsigned int queue = skb_get_queue_mapping(skb);
 
+		carl9170_tx_get_rates(ar, vif, sta, skb);
+		carl9170_tx_apply_rateset(ar, info, skb);
 		skb_queue_tail(&ar->tx_pending[queue], skb);
 	}
 
diff --git a/drivers/net/wireless/ath/regd.c b/drivers/net/wireless/ath/regd.c
index ccc4c71..7d077c7 100644
--- a/drivers/net/wireless/ath/regd.c
+++ b/drivers/net/wireless/ath/regd.c
@@ -42,11 +42,11 @@ static int __ath_regd_init(struct ath_regulatory *reg);
 				NL80211_RRF_PASSIVE_SCAN | NL80211_RRF_NO_OFDM)
 
 /* We allow IBSS on these on a case by case basis by regulatory domain */
-#define ATH9K_5GHZ_5150_5350	REG_RULE(5150-10, 5350+10, 40, 0, 30,\
+#define ATH9K_5GHZ_5150_5350	REG_RULE(5150-10, 5350+10, 80, 0, 30,\
 				NL80211_RRF_PASSIVE_SCAN | NL80211_RRF_NO_IBSS)
-#define ATH9K_5GHZ_5470_5850	REG_RULE(5470-10, 5850+10, 40, 0, 30,\
+#define ATH9K_5GHZ_5470_5850	REG_RULE(5470-10, 5850+10, 80, 0, 30,\
 				NL80211_RRF_PASSIVE_SCAN | NL80211_RRF_NO_IBSS)
-#define ATH9K_5GHZ_5725_5850	REG_RULE(5725-10, 5850+10, 40, 0, 30,\
+#define ATH9K_5GHZ_5725_5850	REG_RULE(5725-10, 5850+10, 80, 0, 30,\
 				NL80211_RRF_PASSIVE_SCAN | NL80211_RRF_NO_IBSS)
 
 #define ATH9K_2GHZ_ALL		ATH9K_2GHZ_CH01_11, \
diff --git a/drivers/net/wireless/ath/wil6210/Kconfig b/drivers/net/wireless/ath/wil6210/Kconfig
index bac3d98..ce8c038 100644
--- a/drivers/net/wireless/ath/wil6210/Kconfig
+++ b/drivers/net/wireless/ath/wil6210/Kconfig
@@ -27,3 +27,15 @@ config WIL6210_ISR_COR
 	  self-clear when accessed for debug purposes, it makes
 	  such monitoring impossible.
 	  Say y unless you debug interrupts
+
+config WIL6210_TRACING
+	bool "wil6210 tracing support"
+	depends on WIL6210
+	depends on EVENT_TRACING
+	default y
+	---help---
+	  Say Y here to enable tracepoints for the wil6210 driver
+	  using the kernel tracing infrastructure.  Select this
+	  option if you are interested in debugging the driver.
+
+	  If unsure, say Y to make it easier to debug problems.
diff --git a/drivers/net/wireless/ath/wil6210/Makefile b/drivers/net/wireless/ath/wil6210/Makefile
index d288eea..f891d51 100644
--- a/drivers/net/wireless/ath/wil6210/Makefile
+++ b/drivers/net/wireless/ath/wil6210/Makefile
@@ -1,15 +1,20 @@
 obj-$(CONFIG_WIL6210) += wil6210.o
 
-wil6210-objs := main.o
-wil6210-objs += netdev.o
-wil6210-objs += cfg80211.o
-wil6210-objs += pcie_bus.o
-wil6210-objs += debugfs.o
-wil6210-objs += wmi.o
-wil6210-objs += interrupt.o
-wil6210-objs += txrx.o
+wil6210-y := main.o
+wil6210-y += netdev.o
+wil6210-y += cfg80211.o
+wil6210-y += pcie_bus.o
+wil6210-y += debugfs.o
+wil6210-y += wmi.o
+wil6210-y += interrupt.o
+wil6210-y += txrx.o
+wil6210-y += debug.o
+wil6210-$(CONFIG_WIL6210_TRACING) += trace.o
 
 ifeq (, $(findstring -W,$(EXTRA_CFLAGS)))
 	subdir-ccflags-y += -Werror
 endif
+# for tracing framework to find trace.h
+CFLAGS_trace.o := -I$(src)
+
 subdir-ccflags-y += -D__CHECK_ENDIAN__
diff --git a/drivers/net/wireless/ath/wil6210/cfg80211.c b/drivers/net/wireless/ath/wil6210/cfg80211.c
index c5d4a87..61c302a 100644
--- a/drivers/net/wireless/ath/wil6210/cfg80211.c
+++ b/drivers/net/wireless/ath/wil6210/cfg80211.c
@@ -322,12 +322,16 @@ static int wil_cfg80211_connect(struct wiphy *wiphy,
 	 * FW don't support scan after connection attempt
 	 */
 	set_bit(wil_status_dontscan, &wil->status);
+	set_bit(wil_status_fwconnecting, &wil->status);
 
 	rc = wmi_send(wil, WMI_CONNECT_CMDID, &conn, sizeof(conn));
 	if (rc == 0) {
 		/* Connect can take lots of time */
 		mod_timer(&wil->connect_timer,
 			  jiffies + msecs_to_jiffies(2000));
+	} else {
+		clear_bit(wil_status_dontscan, &wil->status);
+		clear_bit(wil_status_fwconnecting, &wil->status);
 	}
 
  out:
@@ -398,6 +402,30 @@ static int wil_cfg80211_set_default_key(struct wiphy *wiphy,
 	return 0;
 }
 
+static int wil_fix_bcon(struct wil6210_priv *wil,
+			struct cfg80211_beacon_data *bcon)
+{
+	struct ieee80211_mgmt *f = (struct ieee80211_mgmt *)bcon->probe_resp;
+	size_t hlen = offsetof(struct ieee80211_mgmt, u.probe_resp.variable);
+	int rc = 0;
+
+	if (bcon->probe_resp_len <= hlen)
+		return 0;
+
+	if (!bcon->proberesp_ies) {
+		bcon->proberesp_ies = f->u.probe_resp.variable;
+		bcon->proberesp_ies_len = bcon->probe_resp_len - hlen;
+		rc = 1;
+	}
+	if (!bcon->assocresp_ies) {
+		bcon->assocresp_ies = f->u.probe_resp.variable;
+		bcon->assocresp_ies_len = bcon->probe_resp_len - hlen;
+		rc = 1;
+	}
+
+	return rc;
+}
+
 static int wil_cfg80211_start_ap(struct wiphy *wiphy,
 				 struct net_device *ndev,
 				 struct cfg80211_ap_settings *info)
@@ -419,10 +447,18 @@ static int wil_cfg80211_start_ap(struct wiphy *wiphy,
 	print_hex_dump_bytes("SSID ", DUMP_PREFIX_OFFSET,
 			     info->ssid, info->ssid_len);
 
+	if (wil_fix_bcon(wil, bcon))
+		wil_dbg_misc(wil, "Fixed bcon\n");
+
 	rc = wil_reset(wil);
 	if (rc)
 		return rc;
 
+	/* Rx VRING. */
+	rc = wil_rx_init(wil);
+	if (rc)
+		return rc;
+
 	rc = wmi_set_ssid(wil, info->ssid_len, info->ssid);
 	if (rc)
 		return rc;
@@ -451,8 +487,6 @@ static int wil_cfg80211_start_ap(struct wiphy *wiphy,
 	if (rc)
 		return rc;
 
-	/* Rx VRING. After MAC and beacon */
-	rc = wil_rx_init(wil);
 
 	netif_carrier_on(ndev);
 
diff --git a/drivers/net/wireless/ath/wil6210/debug.c b/drivers/net/wireless/ath/wil6210/debug.c
new file mode 100644
index 0000000..9eeabf4
--- /dev/null
+++ b/drivers/net/wireless/ath/wil6210/debug.c
@@ -0,0 +1,69 @@
+/*
+ * Copyright (c) 2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include "wil6210.h"
+#include "trace.h"
+
+int wil_err(struct wil6210_priv *wil, const char *fmt, ...)
+{
+	struct net_device *ndev = wil_to_ndev(wil);
+	struct va_format vaf = {
+		.fmt = fmt,
+	};
+	va_list args;
+	int ret;
+
+	va_start(args, fmt);
+	vaf.va = &args;
+	ret = netdev_err(ndev, "%pV", &vaf);
+	trace_wil6210_log_err(&vaf);
+	va_end(args);
+
+	return ret;
+}
+
+int wil_info(struct wil6210_priv *wil, const char *fmt, ...)
+{
+	struct net_device *ndev = wil_to_ndev(wil);
+	struct va_format vaf = {
+		.fmt = fmt,
+	};
+	va_list args;
+	int ret;
+
+	va_start(args, fmt);
+	vaf.va = &args;
+	ret = netdev_info(ndev, "%pV", &vaf);
+	trace_wil6210_log_info(&vaf);
+	va_end(args);
+
+	return ret;
+}
+
+int wil_dbg_trace(struct wil6210_priv *wil, const char *fmt, ...)
+{
+	struct va_format vaf = {
+		.fmt = fmt,
+	};
+	va_list args;
+
+	va_start(args, fmt);
+	vaf.va = &args;
+	trace_wil6210_log_dbg(&vaf);
+	va_end(args);
+
+	return 0;
+}
diff --git a/drivers/net/wireless/ath/wil6210/debugfs.c b/drivers/net/wireless/ath/wil6210/debugfs.c
index 727b1f5..e8308ec 100644
--- a/drivers/net/wireless/ath/wil6210/debugfs.c
+++ b/drivers/net/wireless/ath/wil6210/debugfs.c
@@ -418,9 +418,15 @@ static int wil_txdesc_debugfs_show(struct seq_file *s, void *data)
 		if (skb) {
 			unsigned char printbuf[16 * 3 + 2];
 			int i = 0;
-			int len = skb_headlen(skb);
+			int len = le16_to_cpu(d->dma.length);
 			void *p = skb->data;
 
+			if (len != skb_headlen(skb)) {
+				seq_printf(s, "!!! len: desc = %d skb = %d\n",
+					   len, skb_headlen(skb));
+				len = min_t(int, len, skb_headlen(skb));
+			}
+
 			seq_printf(s, "    len = %d\n", len);
 
 			while (i < len) {
diff --git a/drivers/net/wireless/ath/wil6210/interrupt.c b/drivers/net/wireless/ath/wil6210/interrupt.c
index e3c1e76..8205d3e 100644
--- a/drivers/net/wireless/ath/wil6210/interrupt.c
+++ b/drivers/net/wireless/ath/wil6210/interrupt.c
@@ -17,6 +17,7 @@
 #include <linux/interrupt.h>
 
 #include "wil6210.h"
+#include "trace.h"
 
 /**
  * Theory of operation:
@@ -103,14 +104,14 @@ static void wil6210_mask_irq_pseudo(struct wil6210_priv *wil)
 	clear_bit(wil_status_irqen, &wil->status);
 }
 
-static void wil6210_unmask_irq_tx(struct wil6210_priv *wil)
+void wil6210_unmask_irq_tx(struct wil6210_priv *wil)
 {
 	iowrite32(WIL6210_IMC_TX, wil->csr +
 		  HOSTADDR(RGF_DMA_EP_TX_ICR) +
 		  offsetof(struct RGF_ICR, IMC));
 }
 
-static void wil6210_unmask_irq_rx(struct wil6210_priv *wil)
+void wil6210_unmask_irq_rx(struct wil6210_priv *wil)
 {
 	iowrite32(WIL6210_IMC_RX, wil->csr +
 		  HOSTADDR(RGF_DMA_EP_RX_ICR) +
@@ -168,6 +169,7 @@ static irqreturn_t wil6210_irq_rx(int irq, void *cookie)
 					 HOSTADDR(RGF_DMA_EP_RX_ICR) +
 					 offsetof(struct RGF_ICR, ICR));
 
+	trace_wil6210_irq_rx(isr);
 	wil_dbg_irq(wil, "ISR RX 0x%08x\n", isr);
 
 	if (!isr) {
@@ -180,13 +182,14 @@ static irqreturn_t wil6210_irq_rx(int irq, void *cookie)
 	if (isr & BIT_DMA_EP_RX_ICR_RX_DONE) {
 		wil_dbg_irq(wil, "RX done\n");
 		isr &= ~BIT_DMA_EP_RX_ICR_RX_DONE;
-		wil_rx_handle(wil);
+		wil_dbg_txrx(wil, "NAPI schedule\n");
+		napi_schedule(&wil->napi_rx);
 	}
 
 	if (isr)
 		wil_err(wil, "un-handled RX ISR bits 0x%08x\n", isr);
 
-	wil6210_unmask_irq_rx(wil);
+	/* Rx IRQ will be enabled when NAPI processing finished */
 
 	return IRQ_HANDLED;
 }
@@ -198,6 +201,7 @@ static irqreturn_t wil6210_irq_tx(int irq, void *cookie)
 					 HOSTADDR(RGF_DMA_EP_TX_ICR) +
 					 offsetof(struct RGF_ICR, ICR));
 
+	trace_wil6210_irq_tx(isr);
 	wil_dbg_irq(wil, "ISR TX 0x%08x\n", isr);
 
 	if (!isr) {
@@ -208,23 +212,17 @@ static irqreturn_t wil6210_irq_tx(int irq, void *cookie)
 	wil6210_mask_irq_tx(wil);
 
 	if (isr & BIT_DMA_EP_TX_ICR_TX_DONE) {
-		uint i;
 		wil_dbg_irq(wil, "TX done\n");
+		napi_schedule(&wil->napi_tx);
 		isr &= ~BIT_DMA_EP_TX_ICR_TX_DONE;
-		for (i = 0; i < 24; i++) {
-			u32 mask = BIT_DMA_EP_TX_ICR_TX_DONE_N(i);
-			if (isr & mask) {
-				isr &= ~mask;
-				wil_dbg_irq(wil, "TX done(%i)\n", i);
-				wil_tx_complete(wil, i);
-			}
-		}
+		/* clear also all VRING interrupts */
+		isr &= ~(BIT(25) - 1UL);
 	}
 
 	if (isr)
 		wil_err(wil, "un-handled TX ISR bits 0x%08x\n", isr);
 
-	wil6210_unmask_irq_tx(wil);
+	/* Tx IRQ will be enabled when NAPI processing finished */
 
 	return IRQ_HANDLED;
 }
@@ -256,6 +254,7 @@ static irqreturn_t wil6210_irq_misc(int irq, void *cookie)
 					 HOSTADDR(RGF_DMA_EP_MISC_ICR) +
 					 offsetof(struct RGF_ICR, ICR));
 
+	trace_wil6210_irq_misc(isr);
 	wil_dbg_irq(wil, "ISR MISC 0x%08x\n", isr);
 
 	if (!isr) {
@@ -301,6 +300,7 @@ static irqreturn_t wil6210_irq_misc_thread(int irq, void *cookie)
 	struct wil6210_priv *wil = cookie;
 	u32 isr = wil->isr_misc;
 
+	trace_wil6210_irq_misc_thread(isr);
 	wil_dbg_irq(wil, "Thread ISR MISC 0x%08x\n", isr);
 
 	if (isr & ISR_MISC_FW_ERROR) {
@@ -408,6 +408,7 @@ static irqreturn_t wil6210_hardirq(int irq, void *cookie)
 	if (wil6210_debug_irq_mask(wil, pseudo_cause))
 		return IRQ_NONE;
 
+	trace_wil6210_irq_pseudo(pseudo_cause);
 	wil_dbg_irq(wil, "Pseudo IRQ 0x%08x\n", pseudo_cause);
 
 	wil6210_mask_irq_pseudo(wil);
diff --git a/drivers/net/wireless/ath/wil6210/main.c b/drivers/net/wireless/ath/wil6210/main.c
index a0478e2..0a2844c 100644
--- a/drivers/net/wireless/ath/wil6210/main.c
+++ b/drivers/net/wireless/ath/wil6210/main.c
@@ -56,27 +56,21 @@ static void _wil6210_disconnect(struct wil6210_priv *wil, void *bssid)
 {
 	uint i;
 	struct net_device *ndev = wil_to_ndev(wil);
-	struct wireless_dev *wdev = wil->wdev;
 
 	wil_dbg_misc(wil, "%s()\n", __func__);
 
 	wil_link_off(wil);
-	clear_bit(wil_status_fwconnected, &wil->status);
-
-	switch (wdev->sme_state) {
-	case CFG80211_SME_CONNECTED:
-		cfg80211_disconnected(ndev, WLAN_STATUS_UNSPECIFIED_FAILURE,
+	if (test_bit(wil_status_fwconnected, &wil->status)) {
+		clear_bit(wil_status_fwconnected, &wil->status);
+		cfg80211_disconnected(ndev,
+				      WLAN_STATUS_UNSPECIFIED_FAILURE,
 				      NULL, 0, GFP_KERNEL);
-		break;
-	case CFG80211_SME_CONNECTING:
+	} else if (test_bit(wil_status_fwconnecting, &wil->status)) {
 		cfg80211_connect_result(ndev, bssid, NULL, 0, NULL, 0,
 					WLAN_STATUS_UNSPECIFIED_FAILURE,
 					GFP_KERNEL);
-		break;
-	default:
-		break;
 	}
-
+	clear_bit(wil_status_fwconnecting, &wil->status);
 	for (i = 0; i < ARRAY_SIZE(wil->vring_tx); i++)
 		wil_vring_fini_tx(wil, i);
 
@@ -292,41 +286,36 @@ static int __wil_up(struct wil6210_priv *wil)
 {
 	struct net_device *ndev = wil_to_ndev(wil);
 	struct wireless_dev *wdev = wil->wdev;
-	struct ieee80211_channel *channel = wdev->preset_chandef.chan;
 	int rc;
-	int bi;
-	u16 wmi_nettype = wil_iftype_nl2wmi(wdev->iftype);
 
 	rc = wil_reset(wil);
 	if (rc)
 		return rc;
 
-	/* FIXME Firmware works now in PBSS mode(ToDS=0, FromDS=0) */
-	wmi_nettype = wil_iftype_nl2wmi(NL80211_IFTYPE_ADHOC);
+	/* Rx VRING. After MAC and beacon */
+	rc = wil_rx_init(wil);
+	if (rc)
+		return rc;
+
 	switch (wdev->iftype) {
 	case NL80211_IFTYPE_STATION:
 		wil_dbg_misc(wil, "type: STATION\n");
-		bi = 0;
 		ndev->type = ARPHRD_ETHER;
 		break;
 	case NL80211_IFTYPE_AP:
 		wil_dbg_misc(wil, "type: AP\n");
-		bi = 100;
 		ndev->type = ARPHRD_ETHER;
 		break;
 	case NL80211_IFTYPE_P2P_CLIENT:
 		wil_dbg_misc(wil, "type: P2P_CLIENT\n");
-		bi = 0;
 		ndev->type = ARPHRD_ETHER;
 		break;
 	case NL80211_IFTYPE_P2P_GO:
 		wil_dbg_misc(wil, "type: P2P_GO\n");
-		bi = 100;
 		ndev->type = ARPHRD_ETHER;
 		break;
 	case NL80211_IFTYPE_MONITOR:
 		wil_dbg_misc(wil, "type: Monitor\n");
-		bi = 0;
 		ndev->type = ARPHRD_IEEE80211_RADIOTAP;
 		/* ARPHRD_IEEE80211 or ARPHRD_IEEE80211_RADIOTAP ? */
 		break;
@@ -334,36 +323,12 @@ static int __wil_up(struct wil6210_priv *wil)
 		return -EOPNOTSUPP;
 	}
 
-	/* Apply profile in the following order: */
-	/* SSID and channel for the AP */
-	switch (wdev->iftype) {
-	case NL80211_IFTYPE_AP:
-	case NL80211_IFTYPE_P2P_GO:
-		if (wdev->ssid_len == 0) {
-			wil_err(wil, "SSID not set\n");
-			return -EINVAL;
-		}
-		rc = wmi_set_ssid(wil, wdev->ssid_len, wdev->ssid);
-		if (rc)
-			return rc;
-		break;
-	default:
-		break;
-	}
-
 	/* MAC address - pre-requisite for other commands */
 	wmi_set_mac_address(wil, ndev->dev_addr);
 
-	/* Set up beaconing if required. */
-	if (bi > 0) {
-		rc = wmi_pcp_start(wil, bi, wmi_nettype,
-				   (channel ? channel->hw_value : 0));
-		if (rc)
-			return rc;
-	}
 
-	/* Rx VRING. After MAC and beacon */
-	wil_rx_init(wil);
+	napi_enable(&wil->napi_rx);
+	napi_enable(&wil->napi_tx);
 
 	return 0;
 }
@@ -381,6 +346,9 @@ int wil_up(struct wil6210_priv *wil)
 
 static int __wil_down(struct wil6210_priv *wil)
 {
+	napi_disable(&wil->napi_rx);
+	napi_disable(&wil->napi_tx);
+
 	if (wil->scan_request) {
 		cfg80211_scan_done(wil->scan_request, true);
 		wil->scan_request = NULL;
diff --git a/drivers/net/wireless/ath/wil6210/netdev.c b/drivers/net/wireless/ath/wil6210/netdev.c
index 098a8ec..29dd1e5 100644
--- a/drivers/net/wireless/ath/wil6210/netdev.c
+++ b/drivers/net/wireless/ath/wil6210/netdev.c
@@ -40,6 +40,55 @@ static const struct net_device_ops wil_netdev_ops = {
 	.ndo_validate_addr	= eth_validate_addr,
 };
 
+static int wil6210_netdev_poll_rx(struct napi_struct *napi, int budget)
+{
+	struct wil6210_priv *wil = container_of(napi, struct wil6210_priv,
+						napi_rx);
+	int quota = budget;
+	int done;
+
+	wil_rx_handle(wil, &quota);
+	done = budget - quota;
+
+	if (done <= 1) { /* burst ends - only one packet processed */
+		napi_complete(napi);
+		wil6210_unmask_irq_rx(wil);
+		wil_dbg_txrx(wil, "NAPI RX complete\n");
+	}
+
+	wil_dbg_txrx(wil, "NAPI RX poll(%d) done %d\n", budget, done);
+
+	return done;
+}
+
+static int wil6210_netdev_poll_tx(struct napi_struct *napi, int budget)
+{
+	struct wil6210_priv *wil = container_of(napi, struct wil6210_priv,
+						napi_tx);
+	int tx_done = 0;
+	uint i;
+
+	/* always process ALL Tx complete, regardless budget - it is fast */
+	for (i = 0; i < WIL6210_MAX_TX_RINGS; i++) {
+		struct vring *vring = &wil->vring_tx[i];
+
+		if (!vring->va)
+			continue;
+
+		tx_done += wil_tx_complete(wil, i);
+	}
+
+	if (tx_done <= 1) { /* burst ends - only one packet processed */
+		napi_complete(napi);
+		wil6210_unmask_irq_tx(wil);
+		wil_dbg_txrx(wil, "NAPI TX complete\n");
+	}
+
+	wil_dbg_txrx(wil, "NAPI TX poll(%d) done %d\n", budget, tx_done);
+
+	return min(tx_done, budget);
+}
+
 void *wil_if_alloc(struct device *dev, void __iomem *csr)
 {
 	struct net_device *ndev;
@@ -81,6 +130,11 @@ void *wil_if_alloc(struct device *dev, void __iomem *csr)
 	SET_NETDEV_DEV(ndev, wiphy_dev(wdev->wiphy));
 	wdev->netdev = ndev;
 
+	netif_napi_add(ndev, &wil->napi_rx, wil6210_netdev_poll_rx,
+		       WIL6210_NAPI_BUDGET);
+	netif_napi_add(ndev, &wil->napi_tx, wil6210_netdev_poll_tx,
+		       WIL6210_NAPI_BUDGET);
+
 	wil_link_off(wil);
 
 	return wil;
diff --git a/drivers/net/wireless/ath/wil6210/trace.c b/drivers/net/wireless/ath/wil6210/trace.c
new file mode 100644
index 0000000..cd2534b
--- /dev/null
+++ b/drivers/net/wireless/ath/wil6210/trace.c
@@ -0,0 +1,20 @@
+/*
+ * Copyright (c) 2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#include <linux/module.h>
+
+#define CREATE_TRACE_POINTS
+#include "trace.h"
diff --git a/drivers/net/wireless/ath/wil6210/trace.h b/drivers/net/wireless/ath/wil6210/trace.h
new file mode 100644
index 0000000..eff1239
--- /dev/null
+++ b/drivers/net/wireless/ath/wil6210/trace.h
@@ -0,0 +1,235 @@
+/*
+ * Copyright (c) 2013 Qualcomm Atheros, Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM wil6210
+#if !defined(WIL6210_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define WIL6210_TRACE_H
+
+#include <linux/tracepoint.h>
+#include "wil6210.h"
+#include "txrx.h"
+
+/* create empty functions when tracing is disabled */
+#if !defined(CONFIG_WIL6210_TRACING) || defined(__CHECKER__)
+
+#undef TRACE_EVENT
+#define TRACE_EVENT(name, proto, ...) \
+static inline void trace_ ## name(proto) {}
+#undef DECLARE_EVENT_CLASS
+#define DECLARE_EVENT_CLASS(...)
+#undef DEFINE_EVENT
+#define DEFINE_EVENT(evt_class, name, proto, ...) \
+static inline void trace_ ## name(proto) {}
+#endif /* !CONFIG_WIL6210_TRACING || defined(__CHECKER__) */
+
+DECLARE_EVENT_CLASS(wil6210_wmi,
+	TP_PROTO(u16 id, void *buf, u16 buf_len),
+
+	TP_ARGS(id, buf, buf_len),
+
+	TP_STRUCT__entry(
+		__field(u16, id)
+		__field(u16, buf_len)
+		__dynamic_array(u8, buf, buf_len)
+	),
+
+	TP_fast_assign(
+		__entry->id = id;
+		__entry->buf_len = buf_len;
+		memcpy(__get_dynamic_array(buf), buf, buf_len);
+	),
+
+	TP_printk(
+		"id 0x%04x len %d",
+		__entry->id, __entry->buf_len
+	)
+);
+
+DEFINE_EVENT(wil6210_wmi, wil6210_wmi_cmd,
+	TP_PROTO(u16 id, void *buf, u16 buf_len),
+	TP_ARGS(id, buf, buf_len)
+);
+
+DEFINE_EVENT(wil6210_wmi, wil6210_wmi_event,
+	TP_PROTO(u16 id, void *buf, u16 buf_len),
+	TP_ARGS(id, buf, buf_len)
+);
+
+#define WIL6210_MSG_MAX (200)
+
+DECLARE_EVENT_CLASS(wil6210_log_event,
+	TP_PROTO(struct va_format *vaf),
+	TP_ARGS(vaf),
+	TP_STRUCT__entry(
+		__dynamic_array(char, msg, WIL6210_MSG_MAX)
+	),
+	TP_fast_assign(
+		WARN_ON_ONCE(vsnprintf(__get_dynamic_array(msg),
+				       WIL6210_MSG_MAX,
+				       vaf->fmt,
+				       *vaf->va) >= WIL6210_MSG_MAX);
+	),
+	TP_printk("%s", __get_str(msg))
+);
+
+DEFINE_EVENT(wil6210_log_event, wil6210_log_err,
+	TP_PROTO(struct va_format *vaf),
+	TP_ARGS(vaf)
+);
+
+DEFINE_EVENT(wil6210_log_event, wil6210_log_info,
+	TP_PROTO(struct va_format *vaf),
+	TP_ARGS(vaf)
+);
+
+DEFINE_EVENT(wil6210_log_event, wil6210_log_dbg,
+	TP_PROTO(struct va_format *vaf),
+	TP_ARGS(vaf)
+);
+
+#define wil_pseudo_irq_cause(x) __print_flags(x, "|",	\
+	{BIT_DMA_PSEUDO_CAUSE_RX,	"Rx" },		\
+	{BIT_DMA_PSEUDO_CAUSE_TX,	"Tx" },		\
+	{BIT_DMA_PSEUDO_CAUSE_MISC,	"Misc" })
+
+TRACE_EVENT(wil6210_irq_pseudo,
+	TP_PROTO(u32 x),
+	TP_ARGS(x),
+	TP_STRUCT__entry(
+		__field(u32, x)
+	),
+	TP_fast_assign(
+		__entry->x = x;
+	),
+	TP_printk("cause 0x%08x : %s", __entry->x,
+		  wil_pseudo_irq_cause(__entry->x))
+);
+
+DECLARE_EVENT_CLASS(wil6210_irq,
+	TP_PROTO(u32 x),
+	TP_ARGS(x),
+	TP_STRUCT__entry(
+		__field(u32, x)
+	),
+	TP_fast_assign(
+		__entry->x = x;
+	),
+	TP_printk("cause 0x%08x", __entry->x)
+);
+
+DEFINE_EVENT(wil6210_irq, wil6210_irq_rx,
+	TP_PROTO(u32 x),
+	TP_ARGS(x)
+);
+
+DEFINE_EVENT(wil6210_irq, wil6210_irq_tx,
+	TP_PROTO(u32 x),
+	TP_ARGS(x)
+);
+
+DEFINE_EVENT(wil6210_irq, wil6210_irq_misc,
+	TP_PROTO(u32 x),
+	TP_ARGS(x)
+);
+
+DEFINE_EVENT(wil6210_irq, wil6210_irq_misc_thread,
+	TP_PROTO(u32 x),
+	TP_ARGS(x)
+);
+
+TRACE_EVENT(wil6210_rx,
+	TP_PROTO(u16 index, struct vring_rx_desc *d),
+	TP_ARGS(index, d),
+	TP_STRUCT__entry(
+		__field(u16, index)
+		__field(unsigned int, len)
+		__field(u8, mid)
+		__field(u8, cid)
+		__field(u8, tid)
+		__field(u8, type)
+		__field(u8, subtype)
+		__field(u16, seq)
+		__field(u8, mcs)
+	),
+	TP_fast_assign(
+		__entry->index = index;
+		__entry->len = d->dma.length;
+		__entry->mid = wil_rxdesc_mid(d);
+		__entry->cid = wil_rxdesc_cid(d);
+		__entry->tid = wil_rxdesc_tid(d);
+		__entry->type = wil_rxdesc_ftype(d);
+		__entry->subtype = wil_rxdesc_subtype(d);
+		__entry->seq = wil_rxdesc_seq(d);
+		__entry->mcs = wil_rxdesc_mcs(d);
+	),
+	TP_printk("index %d len %d mid %d cid %d tid %d mcs %d seq 0x%03x"
+		  " type 0x%1x subtype 0x%1x", __entry->index, __entry->len,
+		  __entry->mid, __entry->cid, __entry->tid, __entry->mcs,
+		  __entry->seq, __entry->type, __entry->subtype)
+);
+
+TRACE_EVENT(wil6210_tx,
+	TP_PROTO(u8 vring, u16 index, unsigned int len, u8 frags),
+	TP_ARGS(vring, index, len, frags),
+	TP_STRUCT__entry(
+		__field(u8, vring)
+		__field(u8, frags)
+		__field(u16, index)
+		__field(unsigned int, len)
+	),
+	TP_fast_assign(
+		__entry->vring = vring;
+		__entry->frags = frags;
+		__entry->index = index;
+		__entry->len = len;
+	),
+	TP_printk("vring %d index %d len %d frags %d",
+		  __entry->vring, __entry->index, __entry->len, __entry->frags)
+);
+
+TRACE_EVENT(wil6210_tx_done,
+	TP_PROTO(u8 vring, u16 index, unsigned int len, u8 err),
+	TP_ARGS(vring, index, len, err),
+	TP_STRUCT__entry(
+		__field(u8, vring)
+		__field(u8, err)
+		__field(u16, index)
+		__field(unsigned int, len)
+	),
+	TP_fast_assign(
+		__entry->vring = vring;
+		__entry->index = index;
+		__entry->len = len;
+		__entry->err = err;
+	),
+	TP_printk("vring %d index %d len %d err 0x%02x",
+		  __entry->vring, __entry->index, __entry->len,
+		  __entry->err)
+);
+
+#endif /* WIL6210_TRACE_H || TRACE_HEADER_MULTI_READ*/
+
+#if defined(CONFIG_WIL6210_TRACING) && !defined(__CHECKER__)
+/* we don't want to use include/trace/events */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>
+#endif /* defined(CONFIG_WIL6210_TRACING) && !defined(__CHECKER__) */
diff --git a/drivers/net/wireless/ath/wil6210/txrx.c b/drivers/net/wireless/ath/wil6210/txrx.c
index 7970245..d240b24 100644
--- a/drivers/net/wireless/ath/wil6210/txrx.c
+++ b/drivers/net/wireless/ath/wil6210/txrx.c
@@ -22,6 +22,7 @@
 #include "wil6210.h"
 #include "wmi.h"
 #include "txrx.h"
+#include "trace.h"
 
 static bool rtap_include_phy_info;
 module_param(rtap_include_phy_info, bool, S_IRUGO);
@@ -89,8 +90,8 @@ static int wil_vring_alloc(struct wil6210_priv *wil, struct vring *vring)
 	 * we can use any
 	 */
 	for (i = 0; i < vring->size; i++) {
-		volatile struct vring_tx_desc *d = &(vring->va[i].tx);
-		d->dma.status = TX_DMA_STATUS_DU;
+		volatile struct vring_tx_desc *_d = &(vring->va[i].tx);
+		_d->dma.status = TX_DMA_STATUS_DU;
 	}
 
 	wil_dbg_misc(wil, "vring[%d] 0x%p:0x%016llx 0x%p\n", vring->size,
@@ -106,30 +107,39 @@ static void wil_vring_free(struct wil6210_priv *wil, struct vring *vring,
 	size_t sz = vring->size * sizeof(vring->va[0]);
 
 	while (!wil_vring_is_empty(vring)) {
+		dma_addr_t pa;
+		struct sk_buff *skb;
+		u16 dmalen;
+
 		if (tx) {
-			volatile struct vring_tx_desc *d =
+			struct vring_tx_desc dd, *d = &dd;
+			volatile struct vring_tx_desc *_d =
 					&vring->va[vring->swtail].tx;
-			dma_addr_t pa = d->dma.addr_low |
-					((u64)d->dma.addr_high << 32);
-			struct sk_buff *skb = vring->ctx[vring->swtail];
+
+			*d = *_d;
+			pa = wil_desc_addr(&d->dma.addr);
+			dmalen = le16_to_cpu(d->dma.length);
+			skb = vring->ctx[vring->swtail];
 			if (skb) {
-				dma_unmap_single(dev, pa, d->dma.length,
+				dma_unmap_single(dev, pa, dmalen,
 						 DMA_TO_DEVICE);
 				dev_kfree_skb_any(skb);
 				vring->ctx[vring->swtail] = NULL;
 			} else {
-				dma_unmap_page(dev, pa, d->dma.length,
+				dma_unmap_page(dev, pa, dmalen,
 					       DMA_TO_DEVICE);
 			}
 			vring->swtail = wil_vring_next_tail(vring);
 		} else { /* rx */
-			volatile struct vring_rx_desc *d =
+			struct vring_rx_desc dd, *d = &dd;
+			volatile struct vring_rx_desc *_d =
 					&vring->va[vring->swtail].rx;
-			dma_addr_t pa = d->dma.addr_low |
-					((u64)d->dma.addr_high << 32);
-			struct sk_buff *skb = vring->ctx[vring->swhead];
-			dma_unmap_single(dev, pa, d->dma.length,
-					 DMA_FROM_DEVICE);
+
+			*d = *_d;
+			pa = wil_desc_addr(&d->dma.addr);
+			dmalen = le16_to_cpu(d->dma.length);
+			skb = vring->ctx[vring->swhead];
+			dma_unmap_single(dev, pa, dmalen, DMA_FROM_DEVICE);
 			kfree_skb(skb);
 			wil_vring_advance_head(vring, 1);
 		}
@@ -151,7 +161,8 @@ static int wil_vring_alloc_skb(struct wil6210_priv *wil, struct vring *vring,
 {
 	struct device *dev = wil_to_dev(wil);
 	unsigned int sz = RX_BUF_LEN;
-	volatile struct vring_rx_desc *d = &(vring->va[i].rx);
+	struct vring_rx_desc dd, *d = &dd;
+	volatile struct vring_rx_desc *_d = &(vring->va[i].rx);
 	dma_addr_t pa;
 
 	/* TODO align */
@@ -169,13 +180,13 @@ static int wil_vring_alloc_skb(struct wil6210_priv *wil, struct vring *vring,
 	}
 
 	d->dma.d0 = BIT(9) | RX_DMA_D0_CMD_DMA_IT;
-	d->dma.addr_low = lower_32_bits(pa);
-	d->dma.addr_high = (u16)upper_32_bits(pa);
+	wil_desc_addr_set(&d->dma.addr, pa);
 	/* ip_length don't care */
 	/* b11 don't care */
 	/* error don't care */
 	d->dma.status = 0; /* BIT(0) should be 0 for HW_OWNED */
-	d->dma.length = sz;
+	d->dma.length = cpu_to_le16(sz);
+	*_d = *d;
 	vring->ctx[i] = skb;
 
 	return 0;
@@ -321,11 +332,12 @@ static struct sk_buff *wil_vring_reap_rx(struct wil6210_priv *wil,
 {
 	struct device *dev = wil_to_dev(wil);
 	struct net_device *ndev = wil_to_ndev(wil);
-	volatile struct vring_rx_desc *d;
-	struct vring_rx_desc *d1;
+	volatile struct vring_rx_desc *_d;
+	struct vring_rx_desc *d;
 	struct sk_buff *skb;
 	dma_addr_t pa;
 	unsigned int sz = RX_BUF_LEN;
+	u16 dmalen;
 	u8 ftype;
 	u8 ds_bits;
 
@@ -334,32 +346,44 @@ static struct sk_buff *wil_vring_reap_rx(struct wil6210_priv *wil,
 	if (wil_vring_is_empty(vring))
 		return NULL;
 
-	d = &(vring->va[vring->swhead].rx);
-	if (!(d->dma.status & RX_DMA_STATUS_DU)) {
+	_d = &(vring->va[vring->swhead].rx);
+	if (!(_d->dma.status & RX_DMA_STATUS_DU)) {
 		/* it is not error, we just reached end of Rx done area */
 		return NULL;
 	}
 
-	pa = d->dma.addr_low | ((u64)d->dma.addr_high << 32);
 	skb = vring->ctx[vring->swhead];
+	d = wil_skb_rxdesc(skb);
+	*d = *_d;
+	pa = wil_desc_addr(&d->dma.addr);
+	vring->ctx[vring->swhead] = NULL;
+	wil_vring_advance_head(vring, 1);
+
 	dma_unmap_single(dev, pa, sz, DMA_FROM_DEVICE);
-	skb_trim(skb, d->dma.length);
+	dmalen = le16_to_cpu(d->dma.length);
+
+	trace_wil6210_rx(vring->swhead, d);
+	wil_dbg_txrx(wil, "Rx[%3d] : %d bytes\n", vring->swhead, dmalen);
+	wil_hex_dump_txrx("Rx ", DUMP_PREFIX_NONE, 32, 4,
+			  (const void *)d, sizeof(*d), false);
+
+	if (dmalen > sz) {
+		wil_err(wil, "Rx size too large: %d bytes!\n", dmalen);
+		kfree_skb(skb);
+		return NULL;
+	}
+	skb_trim(skb, dmalen);
+
+	wil_hex_dump_txrx("Rx ", DUMP_PREFIX_OFFSET, 16, 1,
+			  skb->data, skb_headlen(skb), false);
 
-	d1 = wil_skb_rxdesc(skb);
-	*d1 = *d;
 
-	wil->stats.last_mcs_rx = wil_rxdesc_mcs(d1);
+	wil->stats.last_mcs_rx = wil_rxdesc_mcs(d);
 
 	/* use radiotap header only if required */
 	if (ndev->type == ARPHRD_IEEE80211_RADIOTAP)
 		wil_rx_add_radiotap_header(wil, skb);
 
-	wil_dbg_txrx(wil, "Rx[%3d] : %d bytes\n", vring->swhead, d->dma.length);
-	wil_hex_dump_txrx("Rx ", DUMP_PREFIX_NONE, 32, 4,
-			  (const void *)d, sizeof(*d), false);
-
-	wil_vring_advance_head(vring, 1);
-
 	/* no extra checks if in sniffer mode */
 	if (ndev->type != ARPHRD_ETHER)
 		return skb;
@@ -368,7 +392,7 @@ static struct sk_buff *wil_vring_reap_rx(struct wil6210_priv *wil,
 	 * Driver should recognize it by frame type, that is found
 	 * in Rx descriptor. If type is not data, it is 802.11 frame as is
 	 */
-	ftype = wil_rxdesc_ftype(d1) << 2;
+	ftype = wil_rxdesc_ftype(d) << 2;
 	if (ftype != IEEE80211_FTYPE_DATA) {
 		wil_dbg_txrx(wil, "Non-data frame ftype 0x%08x\n", ftype);
 		/* TODO: process it */
@@ -383,7 +407,7 @@ static struct sk_buff *wil_vring_reap_rx(struct wil6210_priv *wil,
 		return NULL;
 	}
 
-	ds_bits = wil_rxdesc_ds_bits(d1);
+	ds_bits = wil_rxdesc_ds_bits(d);
 	if (ds_bits == 1) {
 		/*
 		 * HW bug - in ToDS mode, i.e. Rx on AP side,
@@ -425,6 +449,7 @@ static int wil_rx_refill(struct wil6210_priv *wil, int count)
 
 /*
  * Pass Rx packet to the netif. Update statistics.
+ * Called in softirq context (NAPI poll).
  */
 static void wil_netif_rx_any(struct sk_buff *skb, struct net_device *ndev)
 {
@@ -433,10 +458,7 @@ static void wil_netif_rx_any(struct sk_buff *skb, struct net_device *ndev)
 
 	skb_orphan(skb);
 
-	if (in_interrupt())
-		rc = netif_rx(skb);
-	else
-		rc = netif_rx_ni(skb);
+	rc = netif_receive_skb(skb);
 
 	if (likely(rc == NET_RX_SUCCESS)) {
 		ndev->stats.rx_packets++;
@@ -450,9 +472,9 @@ static void wil_netif_rx_any(struct sk_buff *skb, struct net_device *ndev)
 /**
  * Proceed all completed skb's from Rx VRING
  *
- * Safe to call from IRQ
+ * Safe to call from NAPI poll, i.e. softirq with interrupts enabled
  */
-void wil_rx_handle(struct wil6210_priv *wil)
+void wil_rx_handle(struct wil6210_priv *wil, int *quota)
 {
 	struct net_device *ndev = wil_to_ndev(wil);
 	struct vring *v = &wil->vring_rx;
@@ -463,9 +485,8 @@ void wil_rx_handle(struct wil6210_priv *wil)
 		return;
 	}
 	wil_dbg_txrx(wil, "%s()\n", __func__);
-	while (NULL != (skb = wil_vring_reap_rx(wil, v))) {
-		wil_hex_dump_txrx("Rx ", DUMP_PREFIX_OFFSET, 16, 1,
-				  skb->data, skb_headlen(skb), false);
+	while ((*quota > 0) && (NULL != (skb = wil_vring_reap_rx(wil, v)))) {
+		(*quota)--;
 
 		if (wil->wdev->iftype == NL80211_IFTYPE_MONITOR) {
 			skb->dev = ndev;
@@ -600,18 +621,17 @@ static struct vring *wil_find_tx_vring(struct wil6210_priv *wil,
 	return NULL;
 }
 
-static int wil_tx_desc_map(volatile struct vring_tx_desc *d,
-			   dma_addr_t pa, u32 len)
+static int wil_tx_desc_map(struct vring_tx_desc *d, dma_addr_t pa, u32 len,
+			   int vring_index)
 {
-	d->dma.addr_low = lower_32_bits(pa);
-	d->dma.addr_high = (u16)upper_32_bits(pa);
+	wil_desc_addr_set(&d->dma.addr, pa);
 	d->dma.ip_length = 0;
 	/* 0..6: mac_length; 7:ip_version 0-IP6 1-IP4*/
 	d->dma.b11 = 0/*14 | BIT(7)*/;
 	d->dma.error = 0;
 	d->dma.status = 0; /* BIT(0) should be 0 for HW_OWNED */
-	d->dma.length = len;
-	d->dma.d0 = 0;
+	d->dma.length = cpu_to_le16((u16)len);
+	d->dma.d0 = (vring_index << DMA_CFG_DESC_TX_0_QID_POS);
 	d->mac.d[0] = 0;
 	d->mac.d[1] = 0;
 	d->mac.d[2] = 0;
@@ -630,7 +650,8 @@ static int wil_tx_vring(struct wil6210_priv *wil, struct vring *vring,
 			struct sk_buff *skb)
 {
 	struct device *dev = wil_to_dev(wil);
-	volatile struct vring_tx_desc *d;
+	struct vring_tx_desc dd, *d = &dd;
+	volatile struct vring_tx_desc *_d;
 	u32 swhead = vring->swhead;
 	int avail = wil_vring_avail_tx(vring);
 	int nr_frags = skb_shinfo(skb)->nr_frags;
@@ -648,7 +669,7 @@ static int wil_tx_vring(struct wil6210_priv *wil, struct vring *vring,
 			1 + nr_frags);
 		return -ENOMEM;
 	}
-	d = &(vring->va[i].tx);
+	_d = &(vring->va[i].tx);
 
 	/* FIXME FW can accept only unicast frames for the peer */
 	memcpy(skb->data, wil->dst_addr[vring_index], ETH_ALEN);
@@ -664,28 +685,32 @@ static int wil_tx_vring(struct wil6210_priv *wil, struct vring *vring,
 	if (unlikely(dma_mapping_error(dev, pa)))
 		return -EINVAL;
 	/* 1-st segment */
-	wil_tx_desc_map(d, pa, skb_headlen(skb));
+	wil_tx_desc_map(d, pa, skb_headlen(skb), vring_index);
 	d->mac.d[2] |= ((nr_frags + 1) <<
 		       MAC_CFG_DESC_TX_2_NUM_OF_DESCRIPTORS_POS);
+	if (nr_frags)
+		*_d = *d;
+
 	/* middle segments */
 	for (f = 0; f < nr_frags; f++) {
 		const struct skb_frag_struct *frag =
 				&skb_shinfo(skb)->frags[f];
 		int len = skb_frag_size(frag);
 		i = (swhead + f + 1) % vring->size;
-		d = &(vring->va[i].tx);
+		_d = &(vring->va[i].tx);
 		pa = skb_frag_dma_map(dev, frag, 0, skb_frag_size(frag),
 				DMA_TO_DEVICE);
 		if (unlikely(dma_mapping_error(dev, pa)))
 			goto dma_error;
-		wil_tx_desc_map(d, pa, len);
+		wil_tx_desc_map(d, pa, len, vring_index);
 		vring->ctx[i] = NULL;
+		*_d = *d;
 	}
 	/* for the last seg only */
 	d->dma.d0 |= BIT(DMA_CFG_DESC_TX_0_CMD_EOP_POS);
-	d->dma.d0 |= BIT(9); /* BUG: undocumented bit */
+	d->dma.d0 |= BIT(DMA_CFG_DESC_TX_0_CMD_MARK_WB_POS);
 	d->dma.d0 |= BIT(DMA_CFG_DESC_TX_0_CMD_DMA_IT_POS);
-	d->dma.d0 |= (vring_index << DMA_CFG_DESC_TX_0_QID_POS);
+	*_d = *d;
 
 	wil_hex_dump_txrx("Tx ", DUMP_PREFIX_NONE, 32, 4,
 			  (const void *)d, sizeof(*d), false);
@@ -693,6 +718,7 @@ static int wil_tx_vring(struct wil6210_priv *wil, struct vring *vring,
 	/* advance swhead */
 	wil_vring_advance_head(vring, nr_frags + 1);
 	wil_dbg_txrx(wil, "Tx swhead %d -> %d\n", swhead, vring->swhead);
+	trace_wil6210_tx(vring_index, swhead, skb->len, nr_frags);
 	iowrite32(vring->swhead, wil->csr + HOSTADDR(vring->hwtail));
 	/* hold reference to skb
 	 * to prevent skb release before accounting
@@ -705,14 +731,18 @@ static int wil_tx_vring(struct wil6210_priv *wil, struct vring *vring,
 	/* unmap what we have mapped */
 	/* Note: increment @f to operate with positive index */
 	for (f++; f > 0; f--) {
+		u16 dmalen;
+
 		i = (swhead + f) % vring->size;
-		d = &(vring->va[i].tx);
-		d->dma.status = TX_DMA_STATUS_DU;
-		pa = d->dma.addr_low | ((u64)d->dma.addr_high << 32);
+		_d = &(vring->va[i].tx);
+		*d = *_d;
+		_d->dma.status = TX_DMA_STATUS_DU;
+		pa = wil_desc_addr(&d->dma.addr);
+		dmalen = le16_to_cpu(d->dma.length);
 		if (vring->ctx[i])
-			dma_unmap_single(dev, pa, d->dma.length, DMA_TO_DEVICE);
+			dma_unmap_single(dev, pa, dmalen, DMA_TO_DEVICE);
 		else
-			dma_unmap_page(dev, pa, d->dma.length, DMA_TO_DEVICE);
+			dma_unmap_page(dev, pa, dmalen, DMA_TO_DEVICE);
 	}
 
 	return -EINVAL;
@@ -738,18 +768,16 @@ netdev_tx_t wil_start_xmit(struct sk_buff *skb, struct net_device *ndev)
 		wil_err(wil, "Xmit in monitor mode not supported\n");
 		goto drop;
 	}
-	if (skb->protocol == cpu_to_be16(ETH_P_PAE)) {
-		rc = wmi_tx_eapol(wil, skb);
-	} else {
-		/* find vring */
-		vring = wil_find_tx_vring(wil, skb);
-		if (!vring) {
-			wil_err(wil, "No Tx VRING available\n");
-			goto drop;
-		}
-		/* set up vring entry */
-		rc = wil_tx_vring(wil, vring, skb);
+
+	/* find vring */
+	vring = wil_find_tx_vring(wil, skb);
+	if (!vring) {
+		wil_err(wil, "No Tx VRING available\n");
+		goto drop;
 	}
+	/* set up vring entry */
+	rc = wil_tx_vring(wil, vring, skb);
+
 	switch (rc) {
 	case 0:
 		/* statistics will be updated on the tx_complete */
@@ -761,7 +789,6 @@ netdev_tx_t wil_start_xmit(struct sk_buff *skb, struct net_device *ndev)
 		break; /* goto drop; */
 	}
  drop:
-	netif_tx_stop_all_queues(ndev);
 	ndev->stats.tx_dropped++;
 	dev_kfree_skb_any(skb);
 
@@ -771,41 +798,48 @@ netdev_tx_t wil_start_xmit(struct sk_buff *skb, struct net_device *ndev)
 /**
  * Clean up transmitted skb's from the Tx VRING
  *
+ * Return number of descriptors cleared
+ *
  * Safe to call from IRQ
  */
-void wil_tx_complete(struct wil6210_priv *wil, int ringid)
+int wil_tx_complete(struct wil6210_priv *wil, int ringid)
 {
 	struct net_device *ndev = wil_to_ndev(wil);
 	struct device *dev = wil_to_dev(wil);
 	struct vring *vring = &wil->vring_tx[ringid];
+	int done = 0;
 
 	if (!vring->va) {
 		wil_err(wil, "Tx irq[%d]: vring not initialized\n", ringid);
-		return;
+		return 0;
 	}
 
 	wil_dbg_txrx(wil, "%s(%d)\n", __func__, ringid);
 
 	while (!wil_vring_is_empty(vring)) {
-		volatile struct vring_tx_desc *d1 =
+		volatile struct vring_tx_desc *_d =
 					      &vring->va[vring->swtail].tx;
 		struct vring_tx_desc dd, *d = &dd;
 		dma_addr_t pa;
 		struct sk_buff *skb;
+		u16 dmalen;
 
-		dd = *d1;
+		*d = *_d;
 
 		if (!(d->dma.status & TX_DMA_STATUS_DU))
 			break;
 
+		dmalen = le16_to_cpu(d->dma.length);
+		trace_wil6210_tx_done(ringid, vring->swtail, dmalen,
+				      d->dma.error);
 		wil_dbg_txrx(wil,
 			     "Tx[%3d] : %d bytes, status 0x%02x err 0x%02x\n",
-			     vring->swtail, d->dma.length, d->dma.status,
+			     vring->swtail, dmalen, d->dma.status,
 			     d->dma.error);
 		wil_hex_dump_txrx("TxC ", DUMP_PREFIX_NONE, 32, 4,
 				  (const void *)d, sizeof(*d), false);
 
-		pa = d->dma.addr_low | ((u64)d->dma.addr_high << 32);
+		pa = wil_desc_addr(&d->dma.addr);
 		skb = vring->ctx[vring->swtail];
 		if (skb) {
 			if (d->dma.error == 0) {
@@ -815,18 +849,21 @@ void wil_tx_complete(struct wil6210_priv *wil, int ringid)
 				ndev->stats.tx_errors++;
 			}
 
-			dma_unmap_single(dev, pa, d->dma.length, DMA_TO_DEVICE);
+			dma_unmap_single(dev, pa, dmalen, DMA_TO_DEVICE);
 			dev_kfree_skb_any(skb);
 			vring->ctx[vring->swtail] = NULL;
 		} else {
-			dma_unmap_page(dev, pa, d->dma.length, DMA_TO_DEVICE);
+			dma_unmap_page(dev, pa, dmalen, DMA_TO_DEVICE);
 		}
-		d->dma.addr_low = 0;
-		d->dma.addr_high = 0;
+		d->dma.addr.addr_low = 0;
+		d->dma.addr.addr_high = 0;
 		d->dma.length = 0;
 		d->dma.status = TX_DMA_STATUS_DU;
 		vring->swtail = wil_vring_next_tail(vring);
+		done++;
 	}
 	if (wil_vring_avail_tx(vring) > vring->size/4)
 		netif_tx_wake_all_queues(wil_to_ndev(wil));
+
+	return done;
 }
diff --git a/drivers/net/wireless/ath/wil6210/txrx.h b/drivers/net/wireless/ath/wil6210/txrx.h
index adef12f..859aea6 100644
--- a/drivers/net/wireless/ath/wil6210/txrx.h
+++ b/drivers/net/wireless/ath/wil6210/txrx.h
@@ -27,6 +27,28 @@
 #define WIL6210_RTAP_SIZE (128)
 
 /* Tx/Rx path */
+
+/*
+ * Common representation of physical address in Vring
+ */
+struct vring_dma_addr {
+	__le32 addr_low;
+	__le16 addr_high;
+} __packed;
+
+static inline dma_addr_t wil_desc_addr(struct vring_dma_addr *addr)
+{
+	return le32_to_cpu(addr->addr_low) |
+			   ((u64)le16_to_cpu(addr->addr_high) << 32);
+}
+
+static inline void wil_desc_addr_set(struct vring_dma_addr *addr,
+				     dma_addr_t pa)
+{
+	addr->addr_low = cpu_to_le32(lower_32_bits(pa));
+	addr->addr_high = cpu_to_le16((u16)upper_32_bits(pa));
+}
+
 /*
  * Tx descriptor - MAC part
  * [dword 0]
@@ -179,6 +201,10 @@ struct vring_tx_mac {
 #define DMA_CFG_DESC_TX_0_CMD_EOP_LEN 1
 #define DMA_CFG_DESC_TX_0_CMD_EOP_MSK 0x100
 
+#define DMA_CFG_DESC_TX_0_CMD_MARK_WB_POS 9
+#define DMA_CFG_DESC_TX_0_CMD_MARK_WB_LEN 1
+#define DMA_CFG_DESC_TX_0_CMD_MARK_WB_MSK 0x200
+
 #define DMA_CFG_DESC_TX_0_CMD_DMA_IT_POS 10
 #define DMA_CFG_DESC_TX_0_CMD_DMA_IT_LEN 1
 #define DMA_CFG_DESC_TX_0_CMD_DMA_IT_MSK 0x400
@@ -216,13 +242,12 @@ struct vring_tx_mac {
 
 struct vring_tx_dma {
 	u32 d0;
-	u32 addr_low;
-	u16 addr_high;
+	struct vring_dma_addr addr;
 	u8  ip_length;
 	u8  b11;       /* 0..6: mac_length; 7:ip_version */
 	u8  error;     /* 0..2: err; 3..7: reserved; */
 	u8  status;    /* 0: used; 1..7; reserved */
-	u16 length;
+	__le16 length;
 } __packed;
 
 /*
@@ -315,13 +340,12 @@ struct vring_rx_mac {
 
 struct vring_rx_dma {
 	u32 d0;
-	u32 addr_low;
-	u16 addr_high;
+	struct vring_dma_addr addr;
 	u8  ip_length;
 	u8  b11;
 	u8  error;
 	u8  status;
-	u16 length;
+	__le16 length;
 } __packed;
 
 struct vring_tx_desc {
diff --git a/drivers/net/wireless/ath/wil6210/wil6210.h b/drivers/net/wireless/ath/wil6210/wil6210.h
index 8f76ecd..44fdab5 100644
--- a/drivers/net/wireless/ath/wil6210/wil6210.h
+++ b/drivers/net/wireless/ath/wil6210/wil6210.h
@@ -34,9 +34,11 @@ static inline u32 WIL_GET_BITS(u32 x, int b0, int b1)
 
 #define WIL6210_MEM_SIZE (2*1024*1024UL)
 
-#define WIL6210_RX_RING_SIZE (128)
-#define WIL6210_TX_RING_SIZE (128)
-#define WIL6210_MAX_TX_RINGS (24)
+#define WIL6210_RX_RING_SIZE	(128)
+#define WIL6210_TX_RING_SIZE	(128)
+#define WIL6210_MAX_TX_RINGS	(24) /* HW limit */
+#define WIL6210_MAX_CID		(8) /* HW limit */
+#define WIL6210_NAPI_BUDGET	(16) /* arbitrary */
 
 /* Hardware definitions begin */
 
@@ -184,6 +186,7 @@ struct vring {
 
 enum { /* for wil6210_priv.status */
 	wil_status_fwready = 0,
+	wil_status_fwconnecting,
 	wil_status_fwconnected,
 	wil_status_dontscan,
 	wil_status_reset_done,
@@ -239,6 +242,8 @@ struct wil6210_priv {
 	 * - consumed in thread by wmi_event_worker
 	 */
 	spinlock_t wmi_ev_lock;
+	struct napi_struct napi_rx;
+	struct napi_struct napi_tx;
 	/* DMA related */
 	struct vring vring_rx;
 	struct vring vring_tx[WIL6210_MAX_TX_RINGS];
@@ -267,9 +272,13 @@ struct wil6210_priv {
 #define wil_to_ndev(i) (wil_to_wdev(i)->netdev)
 #define ndev_to_wil(n) (wdev_to_wil(n->ieee80211_ptr))
 
-#define wil_dbg(wil, fmt, arg...) netdev_dbg(wil_to_ndev(wil), fmt, ##arg)
-#define wil_info(wil, fmt, arg...) netdev_info(wil_to_ndev(wil), fmt, ##arg)
-#define wil_err(wil, fmt, arg...) netdev_err(wil_to_ndev(wil), fmt, ##arg)
+int wil_dbg_trace(struct wil6210_priv *wil, const char *fmt, ...);
+int wil_err(struct wil6210_priv *wil, const char *fmt, ...);
+int wil_info(struct wil6210_priv *wil, const char *fmt, ...);
+#define wil_dbg(wil, fmt, arg...) do { \
+	netdev_dbg(wil_to_ndev(wil), fmt, ##arg); \
+	wil_dbg_trace(wil, fmt, ##arg); \
+} while (0)
 
 #define wil_dbg_irq(wil, fmt, arg...) wil_dbg(wil, "DBG[ IRQ]" fmt, ##arg)
 #define wil_dbg_txrx(wil, fmt, arg...) wil_dbg(wil, "DBG[TXRX]" fmt, ##arg)
@@ -320,7 +329,6 @@ int wmi_set_ssid(struct wil6210_priv *wil, u8 ssid_len, const void *ssid);
 int wmi_get_ssid(struct wil6210_priv *wil, u8 *ssid_len, void *ssid);
 int wmi_set_channel(struct wil6210_priv *wil, int channel);
 int wmi_get_channel(struct wil6210_priv *wil, int *channel);
-int wmi_tx_eapol(struct wil6210_priv *wil, struct sk_buff *skb);
 int wmi_del_cipher_key(struct wil6210_priv *wil, u8 key_index,
 		       const void *mac_addr);
 int wmi_add_cipher_key(struct wil6210_priv *wil, u8 key_index,
@@ -356,10 +364,12 @@ int wil_vring_init_tx(struct wil6210_priv *wil, int id, int size,
 void wil_vring_fini_tx(struct wil6210_priv *wil, int id);
 
 netdev_tx_t wil_start_xmit(struct sk_buff *skb, struct net_device *ndev);
-void wil_tx_complete(struct wil6210_priv *wil, int ringid);
+int wil_tx_complete(struct wil6210_priv *wil, int ringid);
+void wil6210_unmask_irq_tx(struct wil6210_priv *wil);
 
 /* RX API */
-void wil_rx_handle(struct wil6210_priv *wil);
+void wil_rx_handle(struct wil6210_priv *wil, int *quota);
+void wil6210_unmask_irq_rx(struct wil6210_priv *wil);
 
 int wil_iftype_nl2wmi(enum nl80211_iftype type);
 
diff --git a/drivers/net/wireless/ath/wil6210/wmi.c b/drivers/net/wireless/ath/wil6210/wmi.c
index 45b04e3..dc8059a 100644
--- a/drivers/net/wireless/ath/wil6210/wmi.c
+++ b/drivers/net/wireless/ath/wil6210/wmi.c
@@ -20,6 +20,7 @@
 #include "wil6210.h"
 #include "txrx.h"
 #include "wmi.h"
+#include "trace.h"
 
 /**
  * WMI event receiving - theory of operations
@@ -74,10 +75,11 @@ static const struct {
 	{0x800000, 0x808000, 0x900000}, /* FW data RAM 32k */
 	{0x840000, 0x860000, 0x908000}, /* peripheral data RAM 128k/96k used */
 	{0x880000, 0x88a000, 0x880000}, /* various RGF */
-	{0x8c0000, 0x932000, 0x8c0000}, /* trivial mapping for upper area */
+	{0x8c0000, 0x949000, 0x8c0000}, /* trivial mapping for upper area */
 	/*
 	 * 920000..930000 ucode code RAM
 	 * 930000..932000 ucode data RAM
+	 * 932000..949000 back-door debug data
 	 */
 };
 
@@ -246,6 +248,8 @@ static int __wmi_send(struct wil6210_priv *wil, u16 cmdid, void *buf, u16 len)
 	iowrite32(r->head = next_head, wil->csr + HOST_MBOX +
 		  offsetof(struct wil6210_mbox_ctl, tx.head));
 
+	trace_wil6210_wmi_cmd(cmdid, buf, len);
+
 	/* interrupt to FW */
 	iowrite32(SW_INT_MBOX, wil->csr + HOST_SW_INT);
 
@@ -311,8 +315,8 @@ static void wmi_evt_rx_mgmt(struct wil6210_priv *wil, int id, void *d, int len)
 
 	wil_dbg_wmi(wil, "MGMT: channel %d MCS %d SNR %d\n",
 		    data->info.channel, data->info.mcs, data->info.snr);
-	wil_dbg_wmi(wil, "status 0x%04x len %d stype %04x\n", d_status, d_len,
-		    le16_to_cpu(data->info.stype));
+	wil_dbg_wmi(wil, "status 0x%04x len %d fc 0x%04x\n", d_status, d_len,
+		    le16_to_cpu(fc));
 	wil_dbg_wmi(wil, "qid %d mid %d cid %d\n",
 		    data->info.qid, data->info.mid, data->info.cid);
 
@@ -406,7 +410,7 @@ static void wmi_evt_connect(struct wil6210_priv *wil, int id, void *d, int len)
 
 	if ((wdev->iftype == NL80211_IFTYPE_STATION) ||
 	    (wdev->iftype == NL80211_IFTYPE_P2P_CLIENT)) {
-		if (wdev->sme_state != CFG80211_SME_CONNECTING) {
+		if (!test_bit(wil_status_fwconnecting, &wil->status)) {
 			wil_err(wil, "Not in connecting state\n");
 			return;
 		}
@@ -430,6 +434,7 @@ static void wmi_evt_connect(struct wil6210_priv *wil, int id, void *d, int len)
 
 		cfg80211_new_sta(ndev, evt->bssid, &sinfo, GFP_KERNEL);
 	}
+	clear_bit(wil_status_fwconnecting, &wil->status);
 	set_bit(wil_status_fwconnected, &wil->status);
 
 	/* FIXME FW can transmit only ucast frames to peer */
@@ -635,8 +640,9 @@ void wmi_recv_cmd(struct wil6210_priv *wil)
 			    hdr.flags);
 		if ((hdr.type == WIL_MBOX_HDR_TYPE_WMI) &&
 		    (len >= sizeof(struct wil6210_mbox_hdr_wmi))) {
-			wil_dbg_wmi(wil, "WMI event 0x%04x\n",
-				    evt->event.wmi.id);
+			u16 id = le16_to_cpu(evt->event.wmi.id);
+			wil_dbg_wmi(wil, "WMI event 0x%04x\n", id);
+			trace_wil6210_wmi_event(id, &evt->event.wmi, len);
 		}
 		wil_hex_dump_wmi("evt ", DUMP_PREFIX_OFFSET, 16, 1,
 				 &evt->event.hdr, sizeof(hdr) + len, true);
@@ -724,7 +730,7 @@ int wmi_pcp_start(struct wil6210_priv *wil, int bi, u8 wmi_nettype, u8 chan)
 		.bcon_interval = cpu_to_le16(bi),
 		.network_type = wmi_nettype,
 		.disable_sec_offload = 1,
-		.channel = chan,
+		.channel = chan - 1,
 	};
 	struct {
 		struct wil6210_mbox_hdr_wmi wmi;
@@ -734,8 +740,12 @@ int wmi_pcp_start(struct wil6210_priv *wil, int bi, u8 wmi_nettype, u8 chan)
 	if (!wil->secure_pcp)
 		cmd.disable_sec = 1;
 
+	/*
+	 * Processing time may be huge, in case of secure AP it takes about
+	 * 3500ms for FW to start AP
+	 */
 	rc = wmi_call(wil, WMI_PCP_START_CMDID, &cmd, sizeof(cmd),
-		      WMI_PCP_STARTED_EVENTID, &reply, sizeof(reply), 100);
+		      WMI_PCP_STARTED_EVENTID, &reply, sizeof(reply), 5000);
 	if (rc)
 		return rc;
 
@@ -829,40 +839,6 @@ int wmi_p2p_cfg(struct wil6210_priv *wil, int channel)
 	return wmi_send(wil, WMI_P2P_CFG_CMDID, &cmd, sizeof(cmd));
 }
 
-int wmi_tx_eapol(struct wil6210_priv *wil, struct sk_buff *skb)
-{
-	struct wmi_eapol_tx_cmd *cmd;
-	struct ethhdr *eth;
-	u16 eapol_len = skb->len - ETH_HLEN;
-	void *eapol = skb->data + ETH_HLEN;
-	uint i;
-	int rc;
-
-	skb_set_mac_header(skb, 0);
-	eth = eth_hdr(skb);
-	wil_dbg_wmi(wil, "EAPOL %d bytes to %pM\n", eapol_len, eth->h_dest);
-	for (i = 0; i < ARRAY_SIZE(wil->vring_tx); i++) {
-		if (memcmp(wil->dst_addr[i], eth->h_dest, ETH_ALEN) == 0)
-			goto found_dest;
-	}
-
-	return -EINVAL;
-
- found_dest:
-	/* find out eapol data & len */
-	cmd = kzalloc(sizeof(*cmd) + eapol_len, GFP_KERNEL);
-	if (!cmd)
-		return -EINVAL;
-
-	memcpy(cmd->dst_mac, eth->h_dest, ETH_ALEN);
-	cmd->eapol_len = cpu_to_le16(eapol_len);
-	memcpy(cmd->eapol, eapol, eapol_len);
-	rc = wmi_send(wil, WMI_EAPOL_TX_CMDID, cmd, sizeof(*cmd) + eapol_len);
-	kfree(cmd);
-
-	return rc;
-}
-
 int wmi_del_cipher_key(struct wil6210_priv *wil, u8 key_index,
 		       const void *mac_addr)
 {
diff --git a/drivers/net/wireless/b43/Kconfig b/drivers/net/wireless/b43/Kconfig
index 078e6f3..51ff0b1 100644
--- a/drivers/net/wireless/b43/Kconfig
+++ b/drivers/net/wireless/b43/Kconfig
@@ -28,18 +28,12 @@ config B43
 
 config B43_BCMA
 	bool "Support for BCMA bus"
-	depends on B43 && BCMA
-	default y
-
-config B43_BCMA_EXTRA
-	bool "Hardware support that overlaps with the brcmsmac driver"
-	depends on B43_BCMA
-	default n if BRCMSMAC
+	depends on B43 && (BCMA = y || BCMA = B43)
 	default y
 
 config B43_SSB
 	bool
-	depends on B43 && SSB
+	depends on B43 && (SSB = y || SSB = B43)
 	default y
 
 # Auto-select SSB PCI-HOST support, if possible
@@ -111,6 +105,7 @@ config B43_PIO
 config B43_PHY_N
 	bool "Support for 802.11n (N-PHY) devices"
 	depends on B43
+	default y
 	---help---
 	  Support for the N-PHY.
 
@@ -132,6 +127,7 @@ config B43_PHY_LP
 config B43_PHY_HT
 	bool "Support for HT-PHY (high throughput) devices"
 	depends on B43 && B43_BCMA
+	default y
 	---help---
 	  Support for the HT-PHY.
 
diff --git a/drivers/net/wireless/b43/main.c b/drivers/net/wireless/b43/main.c
index a95b77a..0e933bb 100644
--- a/drivers/net/wireless/b43/main.c
+++ b/drivers/net/wireless/b43/main.c
@@ -113,13 +113,15 @@ static int b43_modparam_pio = 0;
 module_param_named(pio, b43_modparam_pio, int, 0644);
 MODULE_PARM_DESC(pio, "Use PIO accesses by default: 0=DMA, 1=PIO");
 
+static int modparam_allhwsupport = !IS_ENABLED(CONFIG_BRCMSMAC);
+module_param_named(allhwsupport, modparam_allhwsupport, int, 0444);
+MODULE_PARM_DESC(allhwsupport, "Enable support for all hardware (even it if overlaps with the brcmsmac driver)");
+
 #ifdef CONFIG_B43_BCMA
 static const struct bcma_device_id b43_bcma_tbl[] = {
 	BCMA_CORE(BCMA_MANUF_BCM, BCMA_CORE_80211, 0x11, BCMA_ANY_CLASS),
-#ifdef CONFIG_B43_BCMA_EXTRA
 	BCMA_CORE(BCMA_MANUF_BCM, BCMA_CORE_80211, 0x17, BCMA_ANY_CLASS),
 	BCMA_CORE(BCMA_MANUF_BCM, BCMA_CORE_80211, 0x18, BCMA_ANY_CLASS),
-#endif
 	BCMA_CORE(BCMA_MANUF_BCM, BCMA_CORE_80211, 0x1D, BCMA_ANY_CLASS),
 	BCMA_CORETABLE_END
 };
@@ -5396,6 +5398,12 @@ static int b43_bcma_probe(struct bcma_device *core)
 	struct b43_wl *wl;
 	int err;
 
+	if (!modparam_allhwsupport &&
+	    (core->id.rev == 0x17 || core->id.rev == 0x18)) {
+		pr_err("Support for cores revisions 0x17 and 0x18 disabled by module param allhwsupport=0. Try b43.allhwsupport=1\n");
+		return -ENOTSUPP;
+	}
+
 	dev = b43_bus_dev_bcma_init(core);
 	if (!dev)
 		return -ENODEV;
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh.c b/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh.c
index 4891e3d..e3f3c48 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh.c
@@ -22,9 +22,11 @@
 #include <linux/pci_ids.h>
 #include <linux/sched.h>
 #include <linux/completion.h>
+#include <linux/scatterlist.h>
 #include <linux/mmc/sdio.h>
 #include <linux/mmc/sdio_func.h>
 #include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
 #include <linux/platform_data/brcmfmac-sdio.h>
 
 #include <defs.h>
@@ -160,7 +162,7 @@ int brcmf_sdio_intr_unregister(struct brcmf_sdio_dev *sdiodev)
 	return 0;
 }
 
-int
+static int
 brcmf_sdcard_set_sbaddr_window(struct brcmf_sdio_dev *sdiodev, u32 address)
 {
 	int err = 0, i;
@@ -191,12 +193,33 @@ brcmf_sdcard_set_sbaddr_window(struct brcmf_sdio_dev *sdiodev, u32 address)
 	return err;
 }
 
+static int
+brcmf_sdio_addrprep(struct brcmf_sdio_dev *sdiodev, uint width, u32 *addr)
+{
+	uint bar0 = *addr & ~SBSDIO_SB_OFT_ADDR_MASK;
+	int err = 0;
+
+	if (bar0 != sdiodev->sbwad) {
+		err = brcmf_sdcard_set_sbaddr_window(sdiodev, bar0);
+		if (err)
+			return err;
+
+		sdiodev->sbwad = bar0;
+	}
+
+	*addr &= SBSDIO_SB_OFT_ADDR_MASK;
+
+	if (width == 4)
+		*addr |= SBSDIO_SB_ACCESS_2_4B_FLAG;
+
+	return 0;
+}
+
 int
 brcmf_sdio_regrw_helper(struct brcmf_sdio_dev *sdiodev, u32 addr,
 			void *data, bool write)
 {
 	u8 func_num, reg_size;
-	u32 bar;
 	s32 retry = 0;
 	int ret;
 
@@ -216,18 +239,7 @@ brcmf_sdio_regrw_helper(struct brcmf_sdio_dev *sdiodev, u32 addr,
 		func_num = SDIO_FUNC_1;
 		reg_size = 4;
 
-		/* Set the window for SB core register */
-		bar = addr & ~SBSDIO_SB_OFT_ADDR_MASK;
-		if (bar != sdiodev->sbwad) {
-			ret = brcmf_sdcard_set_sbaddr_window(sdiodev, bar);
-			if (ret != 0) {
-				memset(data, 0xFF, reg_size);
-				return ret;
-			}
-			sdiodev->sbwad = bar;
-		}
-		addr &= SBSDIO_SB_OFT_ADDR_MASK;
-		addr |= SBSDIO_SB_ACCESS_2_4B_FLAG;
+		brcmf_sdio_addrprep(sdiodev, reg_size, &addr);
 	}
 
 	do {
@@ -303,30 +315,207 @@ void brcmf_sdio_regwl(struct brcmf_sdio_dev *sdiodev, u32 addr,
 		*ret = retval;
 }
 
-static int brcmf_sdcard_recv_prepare(struct brcmf_sdio_dev *sdiodev, uint fn,
-				     uint flags, uint width, u32 *addr)
+/**
+ * brcmf_sdio_buffrw - SDIO interface function for block data access
+ * @sdiodev: brcmfmac sdio device
+ * @fn: SDIO function number
+ * @write: direction flag
+ * @addr: dongle memory address as source/destination
+ * @pkt: skb pointer
+ *
+ * This function takes the respbonsibility as the interface function to MMC
+ * stack for block data access. It assumes that the skb passed down by the
+ * caller has already been padded and aligned.
+ */
+static int brcmf_sdio_buffrw(struct brcmf_sdio_dev *sdiodev, uint fn,
+			     bool write, u32 addr, struct sk_buff_head *pktlist)
 {
-	uint bar0 = *addr & ~SBSDIO_SB_OFT_ADDR_MASK;
-	int err = 0;
+	unsigned int req_sz, func_blk_sz, sg_cnt, sg_data_sz, pkt_offset;
+	unsigned int max_blks, max_req_sz, orig_offset, dst_offset;
+	unsigned short max_seg_sz, seg_sz;
+	unsigned char *pkt_data, *orig_data, *dst_data;
+	struct sk_buff *pkt_next = NULL, *local_pkt_next;
+	struct sk_buff_head local_list, *target_list;
+	struct mmc_request mmc_req;
+	struct mmc_command mmc_cmd;
+	struct mmc_data mmc_dat;
+	struct sg_table st;
+	struct scatterlist *sgl;
+	struct mmc_host *host;
+	int ret = 0;
 
-	/* Async not implemented yet */
-	if (flags & SDIO_REQ_ASYNC)
-		return -ENOTSUPP;
+	if (!pktlist->qlen)
+		return -EINVAL;
 
-	if (bar0 != sdiodev->sbwad) {
-		err = brcmf_sdcard_set_sbaddr_window(sdiodev, bar0);
-		if (err)
-			return err;
+	brcmf_pm_resume_wait(sdiodev, &sdiodev->request_buffer_wait);
+	if (brcmf_pm_resume_error(sdiodev))
+		return -EIO;
 
-		sdiodev->sbwad = bar0;
+	/* Single skb use the standard mmc interface */
+	if (pktlist->qlen == 1) {
+		pkt_next = pktlist->next;
+		req_sz = pkt_next->len + 3;
+		req_sz &= (uint)~3;
+
+		if (write)
+			return sdio_memcpy_toio(sdiodev->func[fn], addr,
+						((u8 *)(pkt_next->data)),
+						req_sz);
+		else if (fn == 1)
+			return sdio_memcpy_fromio(sdiodev->func[fn],
+						  ((u8 *)(pkt_next->data)),
+						  addr, req_sz);
+		else
+			/* function 2 read is FIFO operation */
+			return sdio_readsb(sdiodev->func[fn],
+					   ((u8 *)(pkt_next->data)), addr,
+					   req_sz);
 	}
 
-	*addr &= SBSDIO_SB_OFT_ADDR_MASK;
+	target_list = pktlist;
+	/* for host with broken sg support, prepare a page aligned list */
+	__skb_queue_head_init(&local_list);
+	if (sdiodev->pdata && sdiodev->pdata->broken_sg_support && !write) {
+		req_sz = 0;
+		skb_queue_walk(pktlist, pkt_next)
+			req_sz += pkt_next->len;
+		req_sz = ALIGN(req_sz, sdiodev->func[fn]->cur_blksize);
+		while (req_sz > PAGE_SIZE) {
+			pkt_next = brcmu_pkt_buf_get_skb(PAGE_SIZE);
+			if (pkt_next == NULL) {
+				ret = -ENOMEM;
+				goto exit;
+			}
+			__skb_queue_tail(&local_list, pkt_next);
+			req_sz -= PAGE_SIZE;
+		}
+		pkt_next = brcmu_pkt_buf_get_skb(req_sz);
+		if (pkt_next == NULL) {
+			ret = -ENOMEM;
+			goto exit;
+		}
+		__skb_queue_tail(&local_list, pkt_next);
+		target_list = &local_list;
+	}
 
-	if (width == 4)
-		*addr |= SBSDIO_SB_ACCESS_2_4B_FLAG;
+	host = sdiodev->func[fn]->card->host;
+	func_blk_sz = sdiodev->func[fn]->cur_blksize;
+	/* Blocks per command is limited by host count, host transfer
+	 * size and the maximum for IO_RW_EXTENDED of 511 blocks.
+	 */
+	max_blks = min_t(unsigned int, host->max_blk_count, 511u);
+	max_req_sz = min_t(unsigned int, host->max_req_size,
+			   max_blks * func_blk_sz);
+	max_seg_sz = min_t(unsigned short, host->max_segs, SG_MAX_SINGLE_ALLOC);
+	max_seg_sz = min_t(unsigned short, max_seg_sz, target_list->qlen);
+	seg_sz = target_list->qlen;
+	pkt_offset = 0;
+	pkt_next = target_list->next;
+
+	if (sg_alloc_table(&st, max_seg_sz, GFP_KERNEL)) {
+		ret = -ENOMEM;
+		goto exit;
+	}
 
-	return 0;
+	while (seg_sz) {
+		req_sz = 0;
+		sg_cnt = 0;
+		memset(&mmc_req, 0, sizeof(struct mmc_request));
+		memset(&mmc_cmd, 0, sizeof(struct mmc_command));
+		memset(&mmc_dat, 0, sizeof(struct mmc_data));
+		sgl = st.sgl;
+		/* prep sg table */
+		while (pkt_next != (struct sk_buff *)target_list) {
+			pkt_data = pkt_next->data + pkt_offset;
+			sg_data_sz = pkt_next->len - pkt_offset;
+			if (sg_data_sz > host->max_seg_size)
+				sg_data_sz = host->max_seg_size;
+			if (sg_data_sz > max_req_sz - req_sz)
+				sg_data_sz = max_req_sz - req_sz;
+
+			sg_set_buf(sgl, pkt_data, sg_data_sz);
+
+			sg_cnt++;
+			sgl = sg_next(sgl);
+			req_sz += sg_data_sz;
+			pkt_offset += sg_data_sz;
+			if (pkt_offset == pkt_next->len) {
+				pkt_offset = 0;
+				pkt_next = pkt_next->next;
+			}
+
+			if (req_sz >= max_req_sz || sg_cnt >= max_seg_sz)
+				break;
+		}
+		seg_sz -= sg_cnt;
+
+		if (req_sz % func_blk_sz != 0) {
+			brcmf_err("sg request length %u is not %u aligned\n",
+				  req_sz, func_blk_sz);
+			ret = -ENOTBLK;
+			goto exit;
+		}
+		mmc_dat.sg = st.sgl;
+		mmc_dat.sg_len = sg_cnt;
+		mmc_dat.blksz = func_blk_sz;
+		mmc_dat.blocks = req_sz / func_blk_sz;
+		mmc_dat.flags = write ? MMC_DATA_WRITE : MMC_DATA_READ;
+		mmc_cmd.opcode = SD_IO_RW_EXTENDED;
+		mmc_cmd.arg = write ? 1<<31 : 0;	/* write flag  */
+		mmc_cmd.arg |= (fn & 0x7) << 28;	/* SDIO func num */
+		mmc_cmd.arg |= 1<<27;			/* block mode */
+		/* incrementing addr for function 1 */
+		mmc_cmd.arg |= (fn == 1) ? 1<<26 : 0;
+		mmc_cmd.arg |= (addr & 0x1FFFF) << 9;	/* address */
+		mmc_cmd.arg |= mmc_dat.blocks & 0x1FF;	/* block count */
+		mmc_cmd.flags = MMC_RSP_SPI_R5 | MMC_RSP_R5 | MMC_CMD_ADTC;
+		mmc_req.cmd = &mmc_cmd;
+		mmc_req.data = &mmc_dat;
+		if (fn == 1)
+			addr += req_sz;
+
+		mmc_set_data_timeout(&mmc_dat, sdiodev->func[fn]->card);
+		mmc_wait_for_req(host, &mmc_req);
+
+		ret = mmc_cmd.error ? mmc_cmd.error : mmc_dat.error;
+		if (ret != 0) {
+			brcmf_err("CMD53 sg block %s failed %d\n",
+				  write ? "write" : "read", ret);
+			ret = -EIO;
+			break;
+		}
+	}
+
+	if (sdiodev->pdata && sdiodev->pdata->broken_sg_support && !write) {
+		local_pkt_next = local_list.next;
+		orig_offset = 0;
+		skb_queue_walk(pktlist, pkt_next) {
+			dst_offset = 0;
+			do {
+				req_sz = local_pkt_next->len - orig_offset;
+				req_sz = min_t(uint, pkt_next->len - dst_offset,
+					       req_sz);
+				orig_data = local_pkt_next->data + orig_offset;
+				dst_data = pkt_next->data + dst_offset;
+				memcpy(dst_data, orig_data, req_sz);
+				orig_offset += req_sz;
+				dst_offset += req_sz;
+				if (orig_offset == local_pkt_next->len) {
+					orig_offset = 0;
+					local_pkt_next = local_pkt_next->next;
+				}
+				if (dst_offset == pkt_next->len)
+					break;
+			} while (!skb_queue_empty(&local_list));
+		}
+	}
+
+exit:
+	sg_free_table(&st);
+	while ((pkt_next = __skb_dequeue(&local_list)) != NULL)
+		brcmu_pkt_buf_free_skb(pkt_next);
+
+	return ret;
 }
 
 int
@@ -355,21 +544,22 @@ int
 brcmf_sdcard_recv_pkt(struct brcmf_sdio_dev *sdiodev, u32 addr, uint fn,
 		      uint flags, struct sk_buff *pkt)
 {
-	uint incr_fix;
 	uint width;
 	int err = 0;
+	struct sk_buff_head pkt_list;
 
 	brcmf_dbg(SDIO, "fun = %d, addr = 0x%x, size = %d\n",
 		  fn, addr, pkt->len);
 
 	width = (flags & SDIO_REQ_4BYTE) ? 4 : 2;
-	err = brcmf_sdcard_recv_prepare(sdiodev, fn, flags, width, &addr);
+	err = brcmf_sdio_addrprep(sdiodev, width, &addr);
 	if (err)
 		goto done;
 
-	incr_fix = (flags & SDIO_REQ_FIXED) ? SDIOH_DATA_FIX : SDIOH_DATA_INC;
-	err = brcmf_sdioh_request_buffer(sdiodev, incr_fix, SDIOH_READ,
-					 fn, addr, pkt);
+	skb_queue_head_init(&pkt_list);
+	skb_queue_tail(&pkt_list, pkt);
+	err = brcmf_sdio_buffrw(sdiodev, fn, false, addr, &pkt_list);
+	skb_dequeue_tail(&pkt_list);
 
 done:
 	return err;
@@ -386,13 +576,12 @@ int brcmf_sdcard_recv_chain(struct brcmf_sdio_dev *sdiodev, u32 addr, uint fn,
 		  fn, addr, pktq->qlen);
 
 	width = (flags & SDIO_REQ_4BYTE) ? 4 : 2;
-	err = brcmf_sdcard_recv_prepare(sdiodev, fn, flags, width, &addr);
+	err = brcmf_sdio_addrprep(sdiodev, width, &addr);
 	if (err)
 		goto done;
 
 	incr_fix = (flags & SDIO_REQ_FIXED) ? SDIOH_DATA_FIX : SDIOH_DATA_INC;
-	err = brcmf_sdioh_request_chain(sdiodev, incr_fix, SDIOH_READ, fn, addr,
-					pktq);
+	err = brcmf_sdio_buffrw(sdiodev, fn, false, addr, pktq);
 
 done:
 	return err;
@@ -424,37 +613,21 @@ int
 brcmf_sdcard_send_pkt(struct brcmf_sdio_dev *sdiodev, u32 addr, uint fn,
 		      uint flags, struct sk_buff *pkt)
 {
-	uint incr_fix;
 	uint width;
-	uint bar0 = addr & ~SBSDIO_SB_OFT_ADDR_MASK;
 	int err = 0;
+	struct sk_buff_head pkt_list;
 
 	brcmf_dbg(SDIO, "fun = %d, addr = 0x%x, size = %d\n",
 		  fn, addr, pkt->len);
 
-	/* Async not implemented yet */
-	if (flags & SDIO_REQ_ASYNC)
-		return -ENOTSUPP;
-
-	if (bar0 != sdiodev->sbwad) {
-		err = brcmf_sdcard_set_sbaddr_window(sdiodev, bar0);
-		if (err)
-			goto done;
-
-		sdiodev->sbwad = bar0;
-	}
-
-	addr &= SBSDIO_SB_OFT_ADDR_MASK;
-
-	incr_fix = (flags & SDIO_REQ_FIXED) ? SDIOH_DATA_FIX : SDIOH_DATA_INC;
 	width = (flags & SDIO_REQ_4BYTE) ? 4 : 2;
-	if (width == 4)
-		addr |= SBSDIO_SB_ACCESS_2_4B_FLAG;
+	brcmf_sdio_addrprep(sdiodev, width, &addr);
 
-	err = brcmf_sdioh_request_buffer(sdiodev, incr_fix, SDIOH_WRITE, fn,
-					 addr, pkt);
+	skb_queue_head_init(&pkt_list);
+	skb_queue_tail(&pkt_list, pkt);
+	err = brcmf_sdio_buffrw(sdiodev, fn, true, addr, &pkt_list);
+	skb_dequeue_tail(&pkt_list);
 
-done:
 	return err;
 }
 
@@ -466,6 +639,7 @@ brcmf_sdio_ramrw(struct brcmf_sdio_dev *sdiodev, bool write, u32 address,
 	struct sk_buff *pkt;
 	u32 sdaddr;
 	uint dsize;
+	struct sk_buff_head pkt_list;
 
 	dsize = min_t(uint, SBSDIO_SB_OFT_ADDR_LIMIT, size);
 	pkt = dev_alloc_skb(dsize);
@@ -474,6 +648,7 @@ brcmf_sdio_ramrw(struct brcmf_sdio_dev *sdiodev, bool write, u32 address,
 		return -EIO;
 	}
 	pkt->priority = 0;
+	skb_queue_head_init(&pkt_list);
 
 	/* Determine initial transfer parameters */
 	sdaddr = address & SBSDIO_SB_OFT_ADDR_MASK;
@@ -501,9 +676,10 @@ brcmf_sdio_ramrw(struct brcmf_sdio_dev *sdiodev, bool write, u32 address,
 		skb_put(pkt, dsize);
 		if (write)
 			memcpy(pkt->data, data, dsize);
-		bcmerror = brcmf_sdioh_request_buffer(sdiodev, SDIOH_DATA_INC,
-						      write, SDIO_FUNC_1,
-						      sdaddr, pkt);
+		skb_queue_tail(&pkt_list, pkt);
+		bcmerror = brcmf_sdio_buffrw(sdiodev, SDIO_FUNC_1, write,
+					     sdaddr, &pkt_list);
+		skb_dequeue_tail(&pkt_list);
 		if (bcmerror) {
 			brcmf_err("membytes transfer failed\n");
 			break;
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh_sdmmc.c b/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh_sdmmc.c
index 44fa0cd..289e386 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh_sdmmc.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/bcmsdh_sdmmc.c
@@ -66,7 +66,7 @@ MODULE_DEVICE_TABLE(sdio, brcmf_sdmmc_ids);
 static struct brcmfmac_sdio_platform_data *brcmfmac_sdio_pdata;
 
 
-static bool
+bool
 brcmf_pm_resume_error(struct brcmf_sdio_dev *sdiodev)
 {
 	bool is_err = false;
@@ -76,7 +76,7 @@ brcmf_pm_resume_error(struct brcmf_sdio_dev *sdiodev)
 	return is_err;
 }
 
-static void
+void
 brcmf_pm_resume_wait(struct brcmf_sdio_dev *sdiodev, wait_queue_head_t *wq)
 {
 #ifdef CONFIG_PM_SLEEP
@@ -211,115 +211,6 @@ int brcmf_sdioh_request_word(struct brcmf_sdio_dev *sdiodev,
 	return err_ret;
 }
 
-/* precondition: host controller is claimed */
-static int
-brcmf_sdioh_request_data(struct brcmf_sdio_dev *sdiodev, uint write, bool fifo,
-			 uint func, uint addr, struct sk_buff *pkt, uint pktlen)
-{
-	int err_ret = 0;
-
-	if ((write) && (!fifo)) {
-		err_ret = sdio_memcpy_toio(sdiodev->func[func], addr,
-					   ((u8 *) (pkt->data)), pktlen);
-	} else if (write) {
-		err_ret = sdio_memcpy_toio(sdiodev->func[func], addr,
-					   ((u8 *) (pkt->data)), pktlen);
-	} else if (fifo) {
-		err_ret = sdio_readsb(sdiodev->func[func],
-				      ((u8 *) (pkt->data)), addr, pktlen);
-	} else {
-		err_ret = sdio_memcpy_fromio(sdiodev->func[func],
-					     ((u8 *) (pkt->data)),
-					     addr, pktlen);
-	}
-
-	return err_ret;
-}
-
-/*
- * This function takes a queue of packets. The packets on the queue
- * are assumed to be properly aligned by the caller.
- */
-int
-brcmf_sdioh_request_chain(struct brcmf_sdio_dev *sdiodev, uint fix_inc,
-			  uint write, uint func, uint addr,
-			  struct sk_buff_head *pktq)
-{
-	bool fifo = (fix_inc == SDIOH_DATA_FIX);
-	u32 SGCount = 0;
-	int err_ret = 0;
-
-	struct sk_buff *pkt;
-
-	brcmf_dbg(SDIO, "Enter\n");
-
-	brcmf_pm_resume_wait(sdiodev, &sdiodev->request_chain_wait);
-	if (brcmf_pm_resume_error(sdiodev))
-		return -EIO;
-
-	skb_queue_walk(pktq, pkt) {
-		uint pkt_len = pkt->len;
-		pkt_len += 3;
-		pkt_len &= 0xFFFFFFFC;
-
-		err_ret = brcmf_sdioh_request_data(sdiodev, write, fifo, func,
-						   addr, pkt, pkt_len);
-		if (err_ret) {
-			brcmf_err("%s FAILED %p[%d], addr=0x%05x, pkt_len=%d, ERR=0x%08x\n",
-				  write ? "TX" : "RX", pkt, SGCount, addr,
-				  pkt_len, err_ret);
-		} else {
-			brcmf_dbg(SDIO, "%s xfr'd %p[%d], addr=0x%05x, len=%d\n",
-				  write ? "TX" : "RX", pkt, SGCount, addr,
-				  pkt_len);
-		}
-		if (!fifo)
-			addr += pkt_len;
-
-		SGCount++;
-	}
-
-	brcmf_dbg(SDIO, "Exit\n");
-	return err_ret;
-}
-
-/*
- * This function takes a single DMA-able packet.
- */
-int brcmf_sdioh_request_buffer(struct brcmf_sdio_dev *sdiodev,
-			       uint fix_inc, uint write, uint func, uint addr,
-			       struct sk_buff *pkt)
-{
-	int status;
-	uint pkt_len;
-	bool fifo = (fix_inc == SDIOH_DATA_FIX);
-
-	brcmf_dbg(SDIO, "Enter\n");
-
-	if (pkt == NULL)
-		return -EINVAL;
-	pkt_len = pkt->len;
-
-	brcmf_pm_resume_wait(sdiodev, &sdiodev->request_buffer_wait);
-	if (brcmf_pm_resume_error(sdiodev))
-		return -EIO;
-
-	pkt_len += 3;
-	pkt_len &= (uint)~3;
-
-	status = brcmf_sdioh_request_data(sdiodev, write, fifo, func,
-					   addr, pkt, pkt_len);
-	if (status) {
-		brcmf_err("%s FAILED %p, addr=0x%05x, pkt_len=%d, ERR=0x%08x\n",
-			  write ? "TX" : "RX", pkt, addr, pkt_len, status);
-	} else {
-		brcmf_dbg(SDIO, "%s xfr'd %p, addr=0x%05x, len=%d\n",
-			  write ? "TX" : "RX", pkt, addr, pkt_len);
-	}
-
-	return status;
-}
-
 static int brcmf_sdioh_get_cisaddr(struct brcmf_sdio_dev *sdiodev, u32 regaddr)
 {
 	/* read 24 bits and return valid 17 bit addr */
@@ -468,7 +359,6 @@ static int brcmf_ops_sdio_probe(struct sdio_func *func,
 	atomic_set(&sdiodev->suspend, false);
 	init_waitqueue_head(&sdiodev->request_byte_wait);
 	init_waitqueue_head(&sdiodev->request_word_wait);
-	init_waitqueue_head(&sdiodev->request_chain_wait);
 	init_waitqueue_head(&sdiodev->request_buffer_wait);
 
 	brcmf_dbg(SDIO, "F2 found, calling brcmf_sdio_probe...\n");
@@ -606,7 +496,8 @@ static int brcmf_sdio_pd_remove(struct platform_device *pdev)
 static struct platform_driver brcmf_sdio_pd = {
 	.remove		= brcmf_sdio_pd_remove,
 	.driver		= {
-		.name	= BRCMFMAC_SDIO_PDATA_NAME
+		.name	= BRCMFMAC_SDIO_PDATA_NAME,
+		.owner	= THIS_MODULE,
 	}
 };
 
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/dhd.h b/drivers/net/wireless/brcm80211/brcmfmac/dhd.h
index 28db9cf..86cbfe2 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/dhd.h
+++ b/drivers/net/wireless/brcm80211/brcmfmac/dhd.h
@@ -583,6 +583,7 @@ enum brcmf_netif_stop_reason {
  * @bssidx: index of bss associated with this interface.
  * @mac_addr: assigned mac address.
  * @netif_stop: bitmap indicates reason why netif queues are stopped.
+ * @netif_stop_lock: spinlock for update netif_stop from multiple sources.
  * @pend_8021x_cnt: tracks outstanding number of 802.1x frames.
  * @pend_8021x_wait: used for signalling change in count.
  */
@@ -598,6 +599,7 @@ struct brcmf_if {
 	s32 bssidx;
 	u8 mac_addr[ETH_ALEN];
 	u8 netif_stop;
+	spinlock_t netif_stop_lock;
 	atomic_t pend_8021x_cnt;
 	wait_queue_head_t pend_8021x_wait;
 };
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/dhd_cdc.c b/drivers/net/wireless/brcm80211/brcmfmac/dhd_cdc.c
index 59c77aa..dd85401 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/dhd_cdc.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/dhd_cdc.c
@@ -30,6 +30,7 @@
 #include "dhd_bus.h"
 #include "fwsignal.h"
 #include "dhd_dbg.h"
+#include "tracepoint.h"
 
 struct brcmf_proto_cdc_dcmd {
 	__le32 cmd;	/* dongle command value */
@@ -292,6 +293,7 @@ void brcmf_proto_hdrpush(struct brcmf_pub *drvr, int ifidx, u8 offset,
 	h->flags2 = 0;
 	h->data_offset = offset;
 	BDC_SET_IF_IDX(h, ifidx);
+	trace_brcmf_bdchdr(pktbuf->data);
 }
 
 int brcmf_proto_hdrpull(struct brcmf_pub *drvr, bool do_fws, u8 *ifidx,
@@ -309,6 +311,7 @@ int brcmf_proto_hdrpull(struct brcmf_pub *drvr, bool do_fws, u8 *ifidx,
 		return -EBADE;
 	}
 
+	trace_brcmf_bdchdr(pktbuf->data);
 	h = (struct brcmf_proto_bdc_header *)(pktbuf->data);
 
 	*ifidx = BDC_GET_IF_IDX(h);
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.c b/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.c
index 202869c..c37b9d6 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.c
@@ -156,8 +156,11 @@ ssize_t brcmf_debugfs_fws_stats_read(struct file *f, char __user *data,
 			"txs_suppr_core:    %u\n"
 			"txs_suppr_ps:      %u\n"
 			"txs_tossed:        %u\n"
+			"txs_host_tossed:   %u\n"
+			"bus_flow_block:    %u\n"
+			"fws_flow_block:    %u\n"
 			"send_pkts:         BK:%u BE:%u VO:%u VI:%u BCMC:%u\n"
-			"fifo_credits_sent: BK:%u BE:%u VO:%u VI:%u BCMC:%u\n",
+			"requested_sent:    BK:%u BE:%u VO:%u VI:%u BCMC:%u\n",
 			fwstats->header_pulls,
 			fwstats->header_only_pkt,
 			fwstats->tlv_parse_failed,
@@ -176,14 +179,17 @@ ssize_t brcmf_debugfs_fws_stats_read(struct file *f, char __user *data,
 			fwstats->txs_supp_core,
 			fwstats->txs_supp_ps,
 			fwstats->txs_tossed,
+			fwstats->txs_host_tossed,
+			fwstats->bus_flow_block,
+			fwstats->fws_flow_block,
 			fwstats->send_pkts[0], fwstats->send_pkts[1],
 			fwstats->send_pkts[2], fwstats->send_pkts[3],
 			fwstats->send_pkts[4],
-			fwstats->fifo_credits_sent[0],
-			fwstats->fifo_credits_sent[1],
-			fwstats->fifo_credits_sent[2],
-			fwstats->fifo_credits_sent[3],
-			fwstats->fifo_credits_sent[4]);
+			fwstats->requested_sent[0],
+			fwstats->requested_sent[1],
+			fwstats->requested_sent[2],
+			fwstats->requested_sent[3],
+			fwstats->requested_sent[4]);
 
 	return simple_read_from_buffer(data, count, ppos, buf, res);
 }
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.h b/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.h
index 009c87b..0af1f5d 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.h
+++ b/drivers/net/wireless/brcm80211/brcmfmac/dhd_dbg.h
@@ -141,8 +141,7 @@ struct brcmf_fws_stats {
 	u32 header_pulls;
 	u32 pkt2bus;
 	u32 send_pkts[5];
-	u32 fifo_credits_sent[5];
-	u32 fifo_credits_back[6];
+	u32 requested_sent[5];
 	u32 generic_error;
 	u32 mac_update_failed;
 	u32 mac_ps_update_failed;
@@ -158,6 +157,9 @@ struct brcmf_fws_stats {
 	u32 txs_supp_core;
 	u32 txs_supp_ps;
 	u32 txs_tossed;
+	u32 txs_host_tossed;
+	u32 bus_flow_block;
+	u32 fws_flow_block;
 };
 
 struct brcmf_pub;
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/dhd_linux.c b/drivers/net/wireless/brcm80211/brcmfmac/dhd_linux.c
index 2c59357..8e89755 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/dhd_linux.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/dhd_linux.c
@@ -179,7 +179,7 @@ static netdev_tx_t brcmf_netdev_start_xmit(struct sk_buff *skb,
 	struct brcmf_pub *drvr = ifp->drvr;
 	struct ethhdr *eh;
 
-	brcmf_dbg(TRACE, "Enter, idx=%d\n", ifp->bssidx);
+	brcmf_dbg(DATA, "Enter, idx=%d\n", ifp->bssidx);
 
 	/* Can the device send data? */
 	if (drvr->bus_if->state != BRCMF_BUS_DATA) {
@@ -240,11 +240,15 @@ done:
 void brcmf_txflowblock_if(struct brcmf_if *ifp,
 			  enum brcmf_netif_stop_reason reason, bool state)
 {
+	unsigned long flags;
+
 	if (!ifp)
 		return;
 
 	brcmf_dbg(TRACE, "enter: idx=%d stop=0x%X reason=%d state=%d\n",
 		  ifp->bssidx, ifp->netif_stop, reason, state);
+
+	spin_lock_irqsave(&ifp->netif_stop_lock, flags);
 	if (state) {
 		if (!ifp->netif_stop)
 			netif_stop_queue(ifp->ndev);
@@ -254,6 +258,7 @@ void brcmf_txflowblock_if(struct brcmf_if *ifp,
 		if (!ifp->netif_stop)
 			netif_wake_queue(ifp->ndev);
 	}
+	spin_unlock_irqrestore(&ifp->netif_stop_lock, flags);
 }
 
 void brcmf_txflowblock(struct device *dev, bool state)
@@ -264,15 +269,18 @@ void brcmf_txflowblock(struct device *dev, bool state)
 
 	brcmf_dbg(TRACE, "Enter\n");
 
-	for (i = 0; i < BRCMF_MAX_IFS; i++)
-		brcmf_txflowblock_if(drvr->iflist[i],
-				     BRCMF_NETIF_STOP_REASON_BLOCK_BUS, state);
+	if (brcmf_fws_fc_active(drvr->fws)) {
+		brcmf_fws_bus_blocked(drvr, state);
+	} else {
+		for (i = 0; i < BRCMF_MAX_IFS; i++)
+			brcmf_txflowblock_if(drvr->iflist[i],
+					     BRCMF_NETIF_STOP_REASON_BLOCK_BUS,
+					     state);
+	}
 }
 
 void brcmf_rx_frames(struct device *dev, struct sk_buff_head *skb_list)
 {
-	unsigned char *eth;
-	uint len;
 	struct sk_buff *skb, *pnext;
 	struct brcmf_if *ifp;
 	struct brcmf_bus *bus_if = dev_get_drvdata(dev);
@@ -280,7 +288,7 @@ void brcmf_rx_frames(struct device *dev, struct sk_buff_head *skb_list)
 	u8 ifidx;
 	int ret;
 
-	brcmf_dbg(TRACE, "Enter\n");
+	brcmf_dbg(DATA, "Enter\n");
 
 	skb_queue_walk_safe(skb_list, skb, pnext) {
 		skb_unlink(skb, skb_list);
@@ -296,33 +304,12 @@ void brcmf_rx_frames(struct device *dev, struct sk_buff_head *skb_list)
 			continue;
 		}
 
-		/* Get the protocol, maintain skb around eth_type_trans()
-		 * The main reason for this hack is for the limitation of
-		 * Linux 2.4 where 'eth_type_trans' uses the
-		 * 'net->hard_header_len'
-		 * to perform skb_pull inside vs ETH_HLEN. Since to avoid
-		 * coping of the packet coming from the network stack to add
-		 * BDC, Hardware header etc, during network interface
-		 * registration
-		 * we set the 'net->hard_header_len' to ETH_HLEN + extra space
-		 * required
-		 * for BDC, Hardware header etc. and not just the ETH_HLEN
-		 */
-		eth = skb->data;
-		len = skb->len;
-
 		skb->dev = ifp->ndev;
 		skb->protocol = eth_type_trans(skb, skb->dev);
 
 		if (skb->pkt_type == PACKET_MULTICAST)
 			ifp->stats.multicast++;
 
-		skb->data = eth;
-		skb->len = len;
-
-		/* Strip header, count, deliver upward */
-		skb_pull(skb, ETH_HLEN);
-
 		/* Process special event packets */
 		brcmf_fweh_process_skb(drvr, skb);
 
@@ -338,10 +325,8 @@ void brcmf_rx_frames(struct device *dev, struct sk_buff_head *skb_list)
 			netif_rx(skb);
 		else
 			/* If the receive is not processed inside an ISR,
-			 * the softirqd must be woken explicitly to service
-			 * the NET_RX_SOFTIRQ.  In 2.6 kernels, this is handled
-			 * by netif_rx_ni(), but in earlier kernels, we need
-			 * to do it manually.
+			 * the softirqd must be woken explicitly to service the
+			 * NET_RX_SOFTIRQ. This is handled by netif_rx_ni().
 			 */
 			netif_rx_ni(skb);
 	}
@@ -630,7 +615,7 @@ int brcmf_net_attach(struct brcmf_if *ifp, bool rtnl_locked)
 	/* set appropriate operations */
 	ndev->netdev_ops = &brcmf_netdev_ops_pri;
 
-	ndev->hard_header_len = ETH_HLEN + drvr->hdrlen;
+	ndev->hard_header_len += drvr->hdrlen;
 	ndev->ethtool_ops = &brcmf_ethtool_ops;
 
 	drvr->rxsz = ndev->mtu + ndev->hard_header_len +
@@ -779,6 +764,7 @@ struct brcmf_if *brcmf_add_if(struct brcmf_pub *drvr, s32 bssidx, s32 ifidx,
 	ifp->bssidx = bssidx;
 
 	init_waitqueue_head(&ifp->pend_8021x_wait);
+	spin_lock_init(&ifp->netif_stop_lock);
 
 	if (mac_addr != NULL)
 		memcpy(ifp->mac_addr, mac_addr, ETH_ALEN);
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/dhd_sdio.c b/drivers/net/wireless/brcm80211/brcmfmac/dhd_sdio.c
index d248751..2641119 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/dhd_sdio.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/dhd_sdio.c
@@ -448,8 +448,6 @@ struct brcmf_sdio {
 	uint rxblen;		/* Allocated length of rxbuf */
 	u8 *rxctl;		/* Aligned pointer into rxbuf */
 	u8 *rxctl_orig;		/* pointer for freeing rxctl */
-	u8 *databuf;		/* Buffer for receiving big glom packet */
-	u8 *dataptr;		/* Aligned pointer into databuf */
 	uint rxlen;		/* Length of valid data in buffer */
 	spinlock_t rxctl_lock;	/* protection lock for ctrl frame resources */
 
@@ -473,8 +471,6 @@ struct brcmf_sdio {
 	s32 idletime;		/* Control for activity timeout */
 	s32 idlecount;	/* Activity timeout counter */
 	s32 idleclock;	/* How to set bus driver when idle */
-	s32 sd_rxchain;
-	bool use_rxchain;	/* If brcmf should use PKT chains */
 	bool rxflow_mode;	/* Rx flow control mode */
 	bool rxflow;		/* Is rx flow control on */
 	bool alp_only;		/* Don't use HT clock (ALP only) */
@@ -495,8 +491,7 @@ struct brcmf_sdio {
 
 	struct workqueue_struct *brcmf_wq;
 	struct work_struct datawork;
-	struct list_head dpc_tsklst;
-	spinlock_t dpc_tl_lock;
+	atomic_t dpc_tskcnt;
 
 	const struct firmware *firmware;
 	u32 fw_ptr;
@@ -1026,29 +1021,6 @@ static void brcmf_sdbrcm_rxfail(struct brcmf_sdio *bus, bool abort, bool rtx)
 		bus->sdiodev->bus_if->state = BRCMF_BUS_DOWN;
 }
 
-/* copy a buffer into a pkt buffer chain */
-static uint brcmf_sdbrcm_glom_from_buf(struct brcmf_sdio *bus, uint len)
-{
-	uint n, ret = 0;
-	struct sk_buff *p;
-	u8 *buf;
-
-	buf = bus->dataptr;
-
-	/* copy the data */
-	skb_queue_walk(&bus->glom, p) {
-		n = min_t(uint, p->len, len);
-		memcpy(p->data, buf, n);
-		buf += n;
-		len -= n;
-		ret += n;
-		if (!len)
-			break;
-	}
-
-	return ret;
-}
-
 /* return total length of buffer chain */
 static uint brcmf_sdbrcm_glom_len(struct brcmf_sdio *bus)
 {
@@ -1202,8 +1174,6 @@ static u8 brcmf_sdbrcm_rxglom(struct brcmf_sdio *bus, u8 rxseq)
 	int errcode;
 	u8 doff, sfdoff;
 
-	bool usechain = bus->use_rxchain;
-
 	struct brcmf_sdio_read rd_new;
 
 	/* If packets, issue read(s) and send up packet chain */
@@ -1238,7 +1208,6 @@ static u8 brcmf_sdbrcm_rxglom(struct brcmf_sdio *bus, u8 rxseq)
 			if (sublen % BRCMF_SDALIGN) {
 				brcmf_err("sublen %d not multiple of %d\n",
 					  sublen, BRCMF_SDALIGN);
-				usechain = false;
 			}
 			totlen += sublen;
 
@@ -1305,27 +1274,9 @@ static u8 brcmf_sdbrcm_rxglom(struct brcmf_sdio *bus, u8 rxseq)
 		 * packet and and copy into the chain.
 		 */
 		sdio_claim_host(bus->sdiodev->func[1]);
-		if (usechain) {
-			errcode = brcmf_sdcard_recv_chain(bus->sdiodev,
-					bus->sdiodev->sbwad,
-					SDIO_FUNC_2, F2SYNC, &bus->glom);
-		} else if (bus->dataptr) {
-			errcode = brcmf_sdcard_recv_buf(bus->sdiodev,
-					bus->sdiodev->sbwad,
-					SDIO_FUNC_2, F2SYNC,
-					bus->dataptr, dlen);
-			sublen = (u16) brcmf_sdbrcm_glom_from_buf(bus, dlen);
-			if (sublen != dlen) {
-				brcmf_err("FAILED TO COPY, dlen %d sublen %d\n",
-					  dlen, sublen);
-				errcode = -1;
-			}
-			pnext = NULL;
-		} else {
-			brcmf_err("COULDN'T ALLOC %d-BYTE GLOM, FORCE FAILURE\n",
-				  dlen);
-			errcode = -1;
-		}
+		errcode = brcmf_sdcard_recv_chain(bus->sdiodev,
+				bus->sdiodev->sbwad,
+				SDIO_FUNC_2, F2SYNC, &bus->glom);
 		sdio_release_host(bus->sdiodev->func[1]);
 		bus->sdcnt.f2rxdata++;
 
@@ -2061,23 +2012,6 @@ static inline void brcmf_sdbrcm_clrintr(struct brcmf_sdio *bus)
 	}
 }
 
-static inline void brcmf_sdbrcm_adddpctsk(struct brcmf_sdio *bus)
-{
-	struct list_head *new_hd;
-	unsigned long flags;
-
-	if (in_interrupt())
-		new_hd = kzalloc(sizeof(struct list_head), GFP_ATOMIC);
-	else
-		new_hd = kzalloc(sizeof(struct list_head), GFP_KERNEL);
-	if (new_hd == NULL)
-		return;
-
-	spin_lock_irqsave(&bus->dpc_tl_lock, flags);
-	list_add_tail(new_hd, &bus->dpc_tsklst);
-	spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
-}
-
 static int brcmf_sdio_intr_rstatus(struct brcmf_sdio *bus)
 {
 	u8 idx;
@@ -2312,7 +2246,7 @@ static void brcmf_sdbrcm_dpc(struct brcmf_sdio *bus)
 		   (!atomic_read(&bus->fcstate) &&
 		    brcmu_pktq_mlen(&bus->txq, ~bus->flowcontrol) &&
 		    data_ok(bus)) || PKT_AVAILABLE()) {
-		brcmf_sdbrcm_adddpctsk(bus);
+		atomic_inc(&bus->dpc_tskcnt);
 	}
 
 	/* If we're done for now, turn off clock request. */
@@ -2342,7 +2276,6 @@ static int brcmf_sdbrcm_bus_txdata(struct device *dev, struct sk_buff *pkt)
 	struct brcmf_bus *bus_if = dev_get_drvdata(dev);
 	struct brcmf_sdio_dev *sdiodev = bus_if->bus_priv.sdio;
 	struct brcmf_sdio *bus = sdiodev->bus;
-	unsigned long flags;
 
 	brcmf_dbg(TRACE, "Enter\n");
 
@@ -2369,26 +2302,21 @@ static int brcmf_sdbrcm_bus_txdata(struct device *dev, struct sk_buff *pkt)
 	} else {
 		ret = 0;
 	}
-	spin_unlock_bh(&bus->txqlock);
 
 	if (pktq_len(&bus->txq) >= TXHI) {
 		bus->txoff = true;
 		brcmf_txflowblock(bus->sdiodev->dev, true);
 	}
+	spin_unlock_bh(&bus->txqlock);
 
 #ifdef DEBUG
 	if (pktq_plen(&bus->txq, prec) > qcount[prec])
 		qcount[prec] = pktq_plen(&bus->txq, prec);
 #endif
 
-	spin_lock_irqsave(&bus->dpc_tl_lock, flags);
-	if (list_empty(&bus->dpc_tsklst)) {
-		spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
-
-		brcmf_sdbrcm_adddpctsk(bus);
+	if (atomic_read(&bus->dpc_tskcnt) == 0) {
+		atomic_inc(&bus->dpc_tskcnt);
 		queue_work(bus->brcmf_wq, &bus->datawork);
-	} else {
-		spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
 	}
 
 	return ret;
@@ -2525,7 +2453,6 @@ brcmf_sdbrcm_bus_txctl(struct device *dev, unsigned char *msg, uint msglen)
 	struct brcmf_bus *bus_if = dev_get_drvdata(dev);
 	struct brcmf_sdio_dev *sdiodev = bus_if->bus_priv.sdio;
 	struct brcmf_sdio *bus = sdiodev->bus;
-	unsigned long flags;
 
 	brcmf_dbg(TRACE, "Enter\n");
 
@@ -2612,18 +2539,13 @@ brcmf_sdbrcm_bus_txctl(struct device *dev, unsigned char *msg, uint msglen)
 		} while (ret < 0 && retries++ < TXRETRIES);
 	}
 
-	spin_lock_irqsave(&bus->dpc_tl_lock, flags);
 	if ((bus->idletime == BRCMF_IDLE_IMMEDIATE) &&
-	    list_empty(&bus->dpc_tsklst)) {
-		spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
-
+	    atomic_read(&bus->dpc_tskcnt) == 0) {
 		bus->activity = false;
 		sdio_claim_host(bus->sdiodev->func[1]);
 		brcmf_dbg(INFO, "idle\n");
 		brcmf_sdbrcm_clkctl(bus, CLK_NONE, true);
 		sdio_release_host(bus->sdiodev->func[1]);
-	} else {
-		spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
 	}
 
 	if (ret)
@@ -3451,7 +3373,7 @@ void brcmf_sdbrcm_isr(void *arg)
 	if (!bus->intr)
 		brcmf_err("isr w/o interrupt configured!\n");
 
-	brcmf_sdbrcm_adddpctsk(bus);
+	atomic_inc(&bus->dpc_tskcnt);
 	queue_work(bus->brcmf_wq, &bus->datawork);
 }
 
@@ -3460,7 +3382,6 @@ static bool brcmf_sdbrcm_bus_watchdog(struct brcmf_sdio *bus)
 #ifdef DEBUG
 	struct brcmf_bus *bus_if = dev_get_drvdata(bus->sdiodev->dev);
 #endif	/* DEBUG */
-	unsigned long flags;
 
 	brcmf_dbg(TIMER, "Enter\n");
 
@@ -3476,11 +3397,9 @@ static bool brcmf_sdbrcm_bus_watchdog(struct brcmf_sdio *bus)
 		if (!bus->intr ||
 		    (bus->sdcnt.intrcount == bus->sdcnt.lastintrs)) {
 
-			spin_lock_irqsave(&bus->dpc_tl_lock, flags);
-			if (list_empty(&bus->dpc_tsklst)) {
+			if (atomic_read(&bus->dpc_tskcnt) == 0) {
 				u8 devpend;
-				spin_unlock_irqrestore(&bus->dpc_tl_lock,
-						       flags);
+
 				sdio_claim_host(bus->sdiodev->func[1]);
 				devpend = brcmf_sdio_regrb(bus->sdiodev,
 							   SDIO_CCCR_INTx,
@@ -3489,9 +3408,6 @@ static bool brcmf_sdbrcm_bus_watchdog(struct brcmf_sdio *bus)
 				intstatus =
 				    devpend & (INTR_STATUS_FUNC1 |
 					       INTR_STATUS_FUNC2);
-			} else {
-				spin_unlock_irqrestore(&bus->dpc_tl_lock,
-						       flags);
 			}
 
 			/* If there is something, make like the ISR and
@@ -3500,7 +3416,7 @@ static bool brcmf_sdbrcm_bus_watchdog(struct brcmf_sdio *bus)
 				bus->sdcnt.pollcnt++;
 				atomic_set(&bus->ipend, 1);
 
-				brcmf_sdbrcm_adddpctsk(bus);
+				atomic_inc(&bus->dpc_tskcnt);
 				queue_work(bus->brcmf_wq, &bus->datawork);
 			}
 		}
@@ -3545,41 +3461,15 @@ static bool brcmf_sdbrcm_bus_watchdog(struct brcmf_sdio *bus)
 	return (atomic_read(&bus->ipend) > 0);
 }
 
-static bool brcmf_sdbrcm_chipmatch(u16 chipid)
-{
-	if (chipid == BCM43143_CHIP_ID)
-		return true;
-	if (chipid == BCM43241_CHIP_ID)
-		return true;
-	if (chipid == BCM4329_CHIP_ID)
-		return true;
-	if (chipid == BCM4330_CHIP_ID)
-		return true;
-	if (chipid == BCM4334_CHIP_ID)
-		return true;
-	if (chipid == BCM4335_CHIP_ID)
-		return true;
-	return false;
-}
-
 static void brcmf_sdio_dataworker(struct work_struct *work)
 {
 	struct brcmf_sdio *bus = container_of(work, struct brcmf_sdio,
 					      datawork);
-	struct list_head *cur_hd, *tmp_hd;
-	unsigned long flags;
-
-	spin_lock_irqsave(&bus->dpc_tl_lock, flags);
-	list_for_each_safe(cur_hd, tmp_hd, &bus->dpc_tsklst) {
-		spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
 
+	while (atomic_read(&bus->dpc_tskcnt)) {
 		brcmf_sdbrcm_dpc(bus);
-
-		spin_lock_irqsave(&bus->dpc_tl_lock, flags);
-		list_del(cur_hd);
-		kfree(cur_hd);
+		atomic_dec(&bus->dpc_tskcnt);
 	}
-	spin_unlock_irqrestore(&bus->dpc_tl_lock, flags);
 }
 
 static void brcmf_sdbrcm_release_malloc(struct brcmf_sdio *bus)
@@ -3589,9 +3479,6 @@ static void brcmf_sdbrcm_release_malloc(struct brcmf_sdio *bus)
 	kfree(bus->rxbuf);
 	bus->rxctl = bus->rxbuf = NULL;
 	bus->rxlen = 0;
-
-	kfree(bus->databuf);
-	bus->databuf = NULL;
 }
 
 static bool brcmf_sdbrcm_probe_malloc(struct brcmf_sdio *bus)
@@ -3604,29 +3491,10 @@ static bool brcmf_sdbrcm_probe_malloc(struct brcmf_sdio *bus)
 			    ALIGNMENT) + BRCMF_SDALIGN;
 		bus->rxbuf = kmalloc(bus->rxblen, GFP_ATOMIC);
 		if (!(bus->rxbuf))
-			goto fail;
-	}
-
-	/* Allocate buffer to receive glomed packet */
-	bus->databuf = kmalloc(MAX_DATA_BUF, GFP_ATOMIC);
-	if (!(bus->databuf)) {
-		/* release rxbuf which was already located as above */
-		if (!bus->rxblen)
-			kfree(bus->rxbuf);
-		goto fail;
+			return false;
 	}
 
-	/* Align the buffer */
-	if ((unsigned long)bus->databuf % BRCMF_SDALIGN)
-		bus->dataptr = bus->databuf + (BRCMF_SDALIGN -
-			       ((unsigned long)bus->databuf % BRCMF_SDALIGN));
-	else
-		bus->dataptr = bus->databuf;
-
 	return true;
-
-fail:
-	return false;
 }
 
 static bool
@@ -3667,11 +3535,6 @@ brcmf_sdbrcm_probe_attach(struct brcmf_sdio *bus, u32 regsva)
 		goto fail;
 	}
 
-	if (!brcmf_sdbrcm_chipmatch((u16) bus->ci->chip)) {
-		brcmf_err("unsupported chip: 0x%04x\n", bus->ci->chip);
-		goto fail;
-	}
-
 	if (brcmf_sdbrcm_kso_init(bus)) {
 		brcmf_err("error enabling KSO\n");
 		goto fail;
@@ -3770,10 +3633,6 @@ static bool brcmf_sdbrcm_probe_init(struct brcmf_sdio *bus)
 	bus->blocksize = bus->sdiodev->func[2]->cur_blksize;
 	bus->roundup = min(max_roundup, bus->blocksize);
 
-	/* bus module does not support packet chaining */
-	bus->use_rxchain = false;
-	bus->sd_rxchain = false;
-
 	/* SR state */
 	bus->sleeping = false;
 	bus->sr_enabled = false;
@@ -3927,8 +3786,7 @@ void *brcmf_sdbrcm_probe(u32 regsva, struct brcmf_sdio_dev *sdiodev)
 		bus->watchdog_tsk = NULL;
 	}
 	/* Initialize DPC thread */
-	INIT_LIST_HEAD(&bus->dpc_tsklst);
-	spin_lock_init(&bus->dpc_tl_lock);
+	atomic_set(&bus->dpc_tskcnt, 0);
 
 	/* Assign bus interface call back */
 	bus->sdiodev->bus_if->dev = bus->sdiodev->dev;
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/fweh.h b/drivers/net/wireless/brcm80211/brcmfmac/fweh.h
index 6ec5db9..e679214 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/fweh.h
+++ b/drivers/net/wireless/brcm80211/brcmfmac/fweh.h
@@ -101,7 +101,8 @@ struct brcmf_event;
 	BRCMF_ENUM_DEF(P2P_PROBEREQ_MSG, 72) \
 	BRCMF_ENUM_DEF(DCS_REQUEST, 73) \
 	BRCMF_ENUM_DEF(FIFO_CREDIT_MAP, 74) \
-	BRCMF_ENUM_DEF(ACTION_FRAME_RX, 75)
+	BRCMF_ENUM_DEF(ACTION_FRAME_RX, 75) \
+	BRCMF_ENUM_DEF(BCMC_CREDIT_SUPPORT, 127)
 
 #define BRCMF_ENUM_DEF(id, val) \
 	BRCMF_E_##id = (val),
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.c b/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.c
index 5352dc1..f0d9f7f 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.c
@@ -22,7 +22,6 @@
 #include <linux/etherdevice.h>
 #include <linux/err.h>
 #include <linux/jiffies.h>
-#include <uapi/linux/nl80211.h>
 #include <net/cfg80211.h>
 
 #include <brcmu_utils.h>
@@ -142,7 +141,7 @@ static const char *brcmf_fws_get_tlv_name(enum brcmf_fws_tlv_type id)
 #define BRCMF_FWS_FLOWCONTROL_HIWATER			128
 #define BRCMF_FWS_FLOWCONTROL_LOWATER			64
 
-#define BRCMF_FWS_PSQ_PREC_COUNT		((NL80211_NUM_ACS + 1) * 2)
+#define BRCMF_FWS_PSQ_PREC_COUNT		((BRCMF_FWS_FIFO_COUNT + 1) * 2)
 #define BRCMF_FWS_PSQ_LEN				256
 
 #define BRCMF_FWS_HTOD_FLAG_PKTFROMHOST			0x01
@@ -157,11 +156,13 @@ static const char *brcmf_fws_get_tlv_name(enum brcmf_fws_tlv_type id)
  * @BRCMF_FWS_SKBSTATE_NEW: sk_buff is newly arrived in the driver.
  * @BRCMF_FWS_SKBSTATE_DELAYED: sk_buff had to wait on queue.
  * @BRCMF_FWS_SKBSTATE_SUPPRESSED: sk_buff has been suppressed by firmware.
+ * @BRCMF_FWS_SKBSTATE_TIM: allocated for TIM update info.
  */
 enum brcmf_fws_skb_state {
 	BRCMF_FWS_SKBSTATE_NEW,
 	BRCMF_FWS_SKBSTATE_DELAYED,
-	BRCMF_FWS_SKBSTATE_SUPPRESSED
+	BRCMF_FWS_SKBSTATE_SUPPRESSED,
+	BRCMF_FWS_SKBSTATE_TIM
 };
 
 /**
@@ -193,9 +194,8 @@ struct brcmf_skbuff_cb {
  *	b[11]  - packet sent upon firmware request.
  *	b[10]  - packet only contains signalling data.
  *	b[9]   - packet is a tx packet.
- *	b[8]   - packet uses FIFO credit (non-pspoll).
+ *	b[8]   - packet used requested credit
  *	b[7]   - interface in AP mode.
- *	b[6:4] - AC FIFO number.
  *	b[3:0] - interface index.
  */
 #define BRCMF_SKB_IF_FLAGS_REQUESTED_MASK	0x0800
@@ -204,12 +204,10 @@ struct brcmf_skbuff_cb {
 #define BRCMF_SKB_IF_FLAGS_SIGNAL_ONLY_SHIFT	10
 #define BRCMF_SKB_IF_FLAGS_TRANSMIT_MASK        0x0200
 #define BRCMF_SKB_IF_FLAGS_TRANSMIT_SHIFT	9
-#define BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK	0x0100
-#define BRCMF_SKB_IF_FLAGS_CREDITCHECK_SHIFT	8
+#define BRCMF_SKB_IF_FLAGS_REQ_CREDIT_MASK	0x0100
+#define BRCMF_SKB_IF_FLAGS_REQ_CREDIT_SHIFT	8
 #define BRCMF_SKB_IF_FLAGS_IF_AP_MASK		0x0080
 #define BRCMF_SKB_IF_FLAGS_IF_AP_SHIFT		7
-#define BRCMF_SKB_IF_FLAGS_FIFO_MASK		0x0070
-#define BRCMF_SKB_IF_FLAGS_FIFO_SHIFT		4
 #define BRCMF_SKB_IF_FLAGS_INDEX_MASK		0x000f
 #define BRCMF_SKB_IF_FLAGS_INDEX_SHIFT		0
 
@@ -246,7 +244,7 @@ struct brcmf_skbuff_cb {
 #define BRCMF_SKB_HTOD_TAG_HSLOT_MASK			0x00ffff00
 #define BRCMF_SKB_HTOD_TAG_HSLOT_SHIFT			8
 #define BRCMF_SKB_HTOD_TAG_FREERUN_MASK			0x000000ff
-#define BRCMF_SKB_HTOD_TAG_FREERUN_SHIFT			0
+#define BRCMF_SKB_HTOD_TAG_FREERUN_SHIFT		0
 
 #define brcmf_skb_htod_tag_set_field(skb, field, value) \
 	brcmu_maskset32(&(brcmf_skbcb(skb)->htod), \
@@ -278,6 +276,7 @@ struct brcmf_skbuff_cb {
 /**
  * enum brcmf_fws_fifo - fifo indices used by dongle firmware.
  *
+ * @BRCMF_FWS_FIFO_FIRST: first fifo, ie. background.
  * @BRCMF_FWS_FIFO_AC_BK: fifo for background traffic.
  * @BRCMF_FWS_FIFO_AC_BE: fifo for best-effort traffic.
  * @BRCMF_FWS_FIFO_AC_VI: fifo for video traffic.
@@ -287,7 +286,8 @@ struct brcmf_skbuff_cb {
  * @BRCMF_FWS_FIFO_COUNT: number of fifos.
  */
 enum brcmf_fws_fifo {
-	BRCMF_FWS_FIFO_AC_BK,
+	BRCMF_FWS_FIFO_FIRST,
+	BRCMF_FWS_FIFO_AC_BK = BRCMF_FWS_FIFO_FIRST,
 	BRCMF_FWS_FIFO_AC_BE,
 	BRCMF_FWS_FIFO_AC_VI,
 	BRCMF_FWS_FIFO_AC_VO,
@@ -307,12 +307,15 @@ enum brcmf_fws_fifo {
  *	firmware suppress the packet as device is already in PS mode.
  * @BRCMF_FWS_TXSTATUS_FW_TOSSED:
  *	firmware tossed the packet.
+ * @BRCMF_FWS_TXSTATUS_HOST_TOSSED:
+ *	host tossed the packet.
  */
 enum brcmf_fws_txstatus {
 	BRCMF_FWS_TXSTATUS_DISCARD,
 	BRCMF_FWS_TXSTATUS_CORE_SUPPRESS,
 	BRCMF_FWS_TXSTATUS_FW_PS_SUPPRESS,
-	BRCMF_FWS_TXSTATUS_FW_TOSSED
+	BRCMF_FWS_TXSTATUS_FW_TOSSED,
+	BRCMF_FWS_TXSTATUS_HOST_TOSSED
 };
 
 enum brcmf_fws_fcmode {
@@ -343,6 +346,7 @@ enum brcmf_fws_mac_desc_state {
  * @transit_count: packet in transit to firmware.
  */
 struct brcmf_fws_mac_descriptor {
+	char name[16];
 	u8 occupied;
 	u8 mac_handle;
 	u8 interface_id;
@@ -356,7 +360,6 @@ struct brcmf_fws_mac_descriptor {
 	u8 seq[BRCMF_FWS_FIFO_COUNT];
 	struct pktq psq;
 	int transit_count;
-	int suppress_count;
 	int suppr_transit_count;
 	bool send_tim_signal;
 	u8 traffic_pending_bmp;
@@ -383,12 +386,10 @@ enum brcmf_fws_hanger_item_state {
  * struct brcmf_fws_hanger_item - single entry for tx pending packet.
  *
  * @state: entry is either free or occupied.
- * @gen: generation.
  * @pkt: packet itself.
  */
 struct brcmf_fws_hanger_item {
 	enum brcmf_fws_hanger_item_state state;
-	u8 gen;
 	struct sk_buff *pkt;
 };
 
@@ -424,6 +425,7 @@ struct brcmf_fws_info {
 	struct brcmf_fws_stats stats;
 	struct brcmf_fws_hanger hanger;
 	enum brcmf_fws_fcmode fcmode;
+	bool bcmc_credit_check;
 	struct brcmf_fws_macdesc_table desc;
 	struct workqueue_struct *fws_wq;
 	struct work_struct fws_dequeue_work;
@@ -434,6 +436,8 @@ struct brcmf_fws_info {
 	u32 fifo_credit_map;
 	u32 fifo_delay_map;
 	unsigned long borrow_defer_timestamp;
+	bool bus_flow_blocked;
+	bool creditmap_received;
 };
 
 /*
@@ -507,7 +511,6 @@ static void brcmf_fws_hanger_init(struct brcmf_fws_hanger *hanger)
 {
 	int i;
 
-	brcmf_dbg(TRACE, "enter\n");
 	memset(hanger, 0, sizeof(*hanger));
 	for (i = 0; i < ARRAY_SIZE(hanger->items); i++)
 		hanger->items[i].state = BRCMF_FWS_HANGER_ITEM_STATE_FREE;
@@ -517,7 +520,6 @@ static u32 brcmf_fws_hanger_get_free_slot(struct brcmf_fws_hanger *h)
 {
 	u32 i;
 
-	brcmf_dbg(TRACE, "enter\n");
 	i = (h->slot_pos + 1) % BRCMF_FWS_HANGER_MAXITEMS;
 
 	while (i != h->slot_pos) {
@@ -533,14 +535,12 @@ static u32 brcmf_fws_hanger_get_free_slot(struct brcmf_fws_hanger *h)
 	h->failed_slotfind++;
 	i = BRCMF_FWS_HANGER_MAXITEMS;
 done:
-	brcmf_dbg(TRACE, "exit: %d\n", i);
 	return i;
 }
 
 static int brcmf_fws_hanger_pushpkt(struct brcmf_fws_hanger *h,
-					   struct sk_buff *pkt, u32 slot_id)
+				    struct sk_buff *pkt, u32 slot_id)
 {
-	brcmf_dbg(TRACE, "enter\n");
 	if (slot_id >= BRCMF_FWS_HANGER_MAXITEMS)
 		return -ENOENT;
 
@@ -560,7 +560,6 @@ static int brcmf_fws_hanger_poppkt(struct brcmf_fws_hanger *h,
 					  u32 slot_id, struct sk_buff **pktout,
 					  bool remove_item)
 {
-	brcmf_dbg(TRACE, "enter\n");
 	if (slot_id >= BRCMF_FWS_HANGER_MAXITEMS)
 		return -ENOENT;
 
@@ -574,23 +573,18 @@ static int brcmf_fws_hanger_poppkt(struct brcmf_fws_hanger *h,
 	if (remove_item) {
 		h->items[slot_id].state = BRCMF_FWS_HANGER_ITEM_STATE_FREE;
 		h->items[slot_id].pkt = NULL;
-		h->items[slot_id].gen = 0xff;
 		h->popped++;
 	}
 	return 0;
 }
 
 static int brcmf_fws_hanger_mark_suppressed(struct brcmf_fws_hanger *h,
-						   u32 slot_id, u8 gen)
+					    u32 slot_id)
 {
-	brcmf_dbg(TRACE, "enter\n");
-
 	if (slot_id >= BRCMF_FWS_HANGER_MAXITEMS)
 		return -ENOENT;
 
-	h->items[slot_id].gen = gen;
-
-	if (h->items[slot_id].state != BRCMF_FWS_HANGER_ITEM_STATE_INUSE) {
+	if (h->items[slot_id].state == BRCMF_FWS_HANGER_ITEM_STATE_FREE) {
 		brcmf_err("entry not in use\n");
 		return -EINVAL;
 	}
@@ -599,25 +593,6 @@ static int brcmf_fws_hanger_mark_suppressed(struct brcmf_fws_hanger *h,
 	return 0;
 }
 
-static int brcmf_fws_hanger_get_genbit(struct brcmf_fws_hanger *hanger,
-					      struct sk_buff *pkt, u32 slot_id,
-					      int *gen)
-{
-	brcmf_dbg(TRACE, "enter\n");
-	*gen = 0xff;
-
-	if (slot_id >= BRCMF_FWS_HANGER_MAXITEMS)
-		return -ENOENT;
-
-	if (hanger->items[slot_id].state == BRCMF_FWS_HANGER_ITEM_STATE_FREE) {
-		brcmf_err("slot not in use\n");
-		return -EINVAL;
-	}
-
-	*gen = hanger->items[slot_id].gen;
-	return 0;
-}
-
 static void brcmf_fws_hanger_cleanup(struct brcmf_fws_info *fws,
 				     bool (*fn)(struct sk_buff *, void *),
 				     int ifidx)
@@ -627,7 +602,6 @@ static void brcmf_fws_hanger_cleanup(struct brcmf_fws_info *fws,
 	int i;
 	enum brcmf_fws_hanger_item_state s;
 
-	brcmf_dbg(TRACE, "enter: ifidx=%d\n", ifidx);
 	for (i = 0; i < ARRAY_SIZE(h->items); i++) {
 		s = h->items[i].state;
 		if (s == BRCMF_FWS_HANGER_ITEM_STATE_INUSE ||
@@ -644,14 +618,28 @@ static void brcmf_fws_hanger_cleanup(struct brcmf_fws_info *fws,
 	}
 }
 
-static void brcmf_fws_init_mac_descriptor(struct brcmf_fws_mac_descriptor *desc,
-					  u8 *addr, u8 ifidx)
+static void brcmf_fws_macdesc_set_name(struct brcmf_fws_info *fws,
+				       struct brcmf_fws_mac_descriptor *desc)
+{
+	if (desc == &fws->desc.other)
+		strlcpy(desc->name, "MAC-OTHER", sizeof(desc->name));
+	else if (desc->mac_handle)
+		scnprintf(desc->name, sizeof(desc->name), "MAC-%d:%d",
+			  desc->mac_handle, desc->interface_id);
+	else
+		scnprintf(desc->name, sizeof(desc->name), "MACIF:%d",
+			  desc->interface_id);
+}
+
+static void brcmf_fws_macdesc_init(struct brcmf_fws_mac_descriptor *desc,
+				   u8 *addr, u8 ifidx)
 {
 	brcmf_dbg(TRACE,
 		  "enter: desc %p ea=%pM, ifidx=%u\n", desc, addr, ifidx);
 	desc->occupied = 1;
 	desc->state = BRCMF_FWS_STATE_OPEN;
 	desc->requested_credit = 0;
+	desc->requested_packet = 0;
 	/* depending on use may need ifp->bssidx instead */
 	desc->interface_id = ifidx;
 	desc->ac_bitmap = 0xff; /* update this when handling APSD */
@@ -660,22 +648,22 @@ static void brcmf_fws_init_mac_descriptor(struct brcmf_fws_mac_descriptor *desc,
 }
 
 static
-void brcmf_fws_clear_mac_descriptor(struct brcmf_fws_mac_descriptor *desc)
+void brcmf_fws_macdesc_deinit(struct brcmf_fws_mac_descriptor *desc)
 {
 	brcmf_dbg(TRACE,
 		  "enter: ea=%pM, ifidx=%u\n", desc->ea, desc->interface_id);
 	desc->occupied = 0;
 	desc->state = BRCMF_FWS_STATE_CLOSE;
 	desc->requested_credit = 0;
+	desc->requested_packet = 0;
 }
 
 static struct brcmf_fws_mac_descriptor *
-brcmf_fws_mac_descriptor_lookup(struct brcmf_fws_info *fws, u8 *ea)
+brcmf_fws_macdesc_lookup(struct brcmf_fws_info *fws, u8 *ea)
 {
 	struct brcmf_fws_mac_descriptor *entry;
 	int i;
 
-	brcmf_dbg(TRACE, "enter: ea=%pM\n", ea);
 	if (ea == NULL)
 		return ERR_PTR(-EINVAL);
 
@@ -690,42 +678,33 @@ brcmf_fws_mac_descriptor_lookup(struct brcmf_fws_info *fws, u8 *ea)
 }
 
 static struct brcmf_fws_mac_descriptor*
-brcmf_fws_find_mac_desc(struct brcmf_fws_info *fws, struct brcmf_if *ifp,
-			u8 *da)
+brcmf_fws_macdesc_find(struct brcmf_fws_info *fws, struct brcmf_if *ifp, u8 *da)
 {
 	struct brcmf_fws_mac_descriptor *entry = &fws->desc.other;
 	bool multicast;
-	enum nl80211_iftype iftype;
-
-	brcmf_dbg(TRACE, "enter: idx=%d\n", ifp->bssidx);
 
 	multicast = is_multicast_ether_addr(da);
-	iftype = brcmf_cfg80211_get_iftype(ifp);
 
-	/* Multicast destination and P2P clients get the interface entry.
-	 * STA gets the interface entry if there is no exact match. For
-	 * example, TDLS destinations have their own entry.
+	/* Multicast destination, STA and P2P clients get the interface entry.
+	 * STA/GC gets the Mac Entry for TDLS destinations, TDLS destinations
+	 * have their own entry.
 	 */
-	entry = NULL;
-	if ((multicast || iftype == NL80211_IFTYPE_STATION ||
-	     iftype == NL80211_IFTYPE_P2P_CLIENT) && ifp->fws_desc)
+	if (multicast && ifp->fws_desc) {
 		entry = ifp->fws_desc;
-
-	if (entry != NULL && iftype != NL80211_IFTYPE_STATION)
 		goto done;
+	}
 
-	entry = brcmf_fws_mac_descriptor_lookup(fws, da);
+	entry = brcmf_fws_macdesc_lookup(fws, da);
 	if (IS_ERR(entry))
-		entry = &fws->desc.other;
+		entry = ifp->fws_desc;
 
 done:
-	brcmf_dbg(TRACE, "exit: entry=%p\n", entry);
 	return entry;
 }
 
-static bool brcmf_fws_mac_desc_closed(struct brcmf_fws_info *fws,
-				      struct brcmf_fws_mac_descriptor *entry,
-				      int fifo)
+static bool brcmf_fws_macdesc_closed(struct brcmf_fws_info *fws,
+				     struct brcmf_fws_mac_descriptor *entry,
+				     int fifo)
 {
 	struct brcmf_fws_mac_descriptor *if_entry;
 	bool closed;
@@ -748,15 +727,11 @@ static bool brcmf_fws_mac_desc_closed(struct brcmf_fws_info *fws,
 	return closed || !(entry->ac_bitmap & BIT(fifo));
 }
 
-static void brcmf_fws_mac_desc_cleanup(struct brcmf_fws_info *fws,
-				       struct brcmf_fws_mac_descriptor *entry,
-				       int ifidx)
+static void brcmf_fws_macdesc_cleanup(struct brcmf_fws_info *fws,
+				      struct brcmf_fws_mac_descriptor *entry,
+				      int ifidx)
 {
-	brcmf_dbg(TRACE, "enter: entry=(ea=%pM, ifid=%d), ifidx=%d\n",
-		  entry->ea, entry->interface_id, ifidx);
 	if (entry->occupied && (ifidx == -1 || ifidx == entry->interface_id)) {
-		brcmf_dbg(TRACE, "flush psq: ifidx=%d, qlen=%d\n",
-			  ifidx, entry->psq.len);
 		brcmf_fws_psq_flush(fws, &entry->psq, ifidx);
 		entry->occupied = !!(entry->psq.len);
 	}
@@ -772,7 +747,6 @@ static void brcmf_fws_bus_txq_cleanup(struct brcmf_fws_info *fws,
 	int prec;
 	u32 hslot;
 
-	brcmf_dbg(TRACE, "enter: ifidx=%d\n", ifidx);
 	txq = brcmf_bus_gettxq(fws->drvr->bus_if);
 	if (IS_ERR(txq)) {
 		brcmf_dbg(TRACE, "no txq to clean up\n");
@@ -798,7 +772,6 @@ static void brcmf_fws_cleanup(struct brcmf_fws_info *fws, int ifidx)
 	struct brcmf_fws_mac_descriptor *table;
 	bool (*matchfn)(struct sk_buff *, void *) = NULL;
 
-	brcmf_dbg(TRACE, "enter: ifidx=%d\n", ifidx);
 	if (fws == NULL)
 		return;
 
@@ -808,51 +781,121 @@ static void brcmf_fws_cleanup(struct brcmf_fws_info *fws, int ifidx)
 	/* cleanup individual nodes */
 	table = &fws->desc.nodes[0];
 	for (i = 0; i < ARRAY_SIZE(fws->desc.nodes); i++)
-		brcmf_fws_mac_desc_cleanup(fws, &table[i], ifidx);
+		brcmf_fws_macdesc_cleanup(fws, &table[i], ifidx);
 
-	brcmf_fws_mac_desc_cleanup(fws, &fws->desc.other, ifidx);
+	brcmf_fws_macdesc_cleanup(fws, &fws->desc.other, ifidx);
 	brcmf_fws_bus_txq_cleanup(fws, matchfn, ifidx);
 	brcmf_fws_hanger_cleanup(fws, matchfn, ifidx);
 }
 
-static void brcmf_fws_tim_update(struct brcmf_fws_info *ctx,
-				 struct brcmf_fws_mac_descriptor *entry,
-				 int prec)
+static int brcmf_fws_hdrpush(struct brcmf_fws_info *fws, struct sk_buff *skb)
 {
-	brcmf_dbg(TRACE, "enter: ea=%pM\n", entry->ea);
-	if (entry->state == BRCMF_FWS_STATE_CLOSE) {
-		/* check delayedQ and suppressQ in one call using bitmap */
-		if (brcmu_pktq_mlen(&entry->psq, 3 << (prec * 2)) == 0)
-			entry->traffic_pending_bmp =
-				entry->traffic_pending_bmp & ~NBITVAL(prec);
-		else
-			entry->traffic_pending_bmp =
-				entry->traffic_pending_bmp | NBITVAL(prec);
+	struct brcmf_fws_mac_descriptor *entry = brcmf_skbcb(skb)->mac;
+	u8 *wlh;
+	u16 data_offset = 0;
+	u8 fillers;
+	__le32 pkttag = cpu_to_le32(brcmf_skbcb(skb)->htod);
+
+	brcmf_dbg(TRACE, "enter: %s, idx=%d pkttag=0x%08X, hslot=%d\n",
+		  entry->name, brcmf_skb_if_flags_get_field(skb, INDEX),
+		  le32_to_cpu(pkttag), (le32_to_cpu(pkttag) >> 8) & 0xffff);
+	if (entry->send_tim_signal)
+		data_offset += 2 + BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN;
+
+	/* +2 is for Type[1] and Len[1] in TLV, plus TIM signal */
+	data_offset += 2 + BRCMF_FWS_TYPE_PKTTAG_LEN;
+	fillers = round_up(data_offset, 4) - data_offset;
+	data_offset += fillers;
+
+	skb_push(skb, data_offset);
+	wlh = skb->data;
+
+	wlh[0] = BRCMF_FWS_TYPE_PKTTAG;
+	wlh[1] = BRCMF_FWS_TYPE_PKTTAG_LEN;
+	memcpy(&wlh[2], &pkttag, sizeof(pkttag));
+	wlh += BRCMF_FWS_TYPE_PKTTAG_LEN + 2;
+
+	if (entry->send_tim_signal) {
+		entry->send_tim_signal = 0;
+		wlh[0] = BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP;
+		wlh[1] = BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN;
+		wlh[2] = entry->mac_handle;
+		wlh[3] = entry->traffic_pending_bmp;
+		brcmf_dbg(TRACE, "adding TIM info: handle %d bmp 0x%X\n",
+			  entry->mac_handle, entry->traffic_pending_bmp);
+		wlh += BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN + 2;
+		entry->traffic_lastreported_bmp = entry->traffic_pending_bmp;
 	}
-	/* request a TIM update to firmware at the next piggyback opportunity */
+	if (fillers)
+		memset(wlh, BRCMF_FWS_TYPE_FILLER, fillers);
+
+	brcmf_proto_hdrpush(fws->drvr, brcmf_skb_if_flags_get_field(skb, INDEX),
+			    data_offset >> 2, skb);
+	return 0;
+}
+
+static bool brcmf_fws_tim_update(struct brcmf_fws_info *fws,
+				 struct brcmf_fws_mac_descriptor *entry,
+				 int fifo, bool send_immediately)
+{
+	struct sk_buff *skb;
+	struct brcmf_bus *bus;
+	struct brcmf_skbuff_cb *skcb;
+	s32 err;
+	u32 len;
+
+	/* check delayedQ and suppressQ in one call using bitmap */
+	if (brcmu_pktq_mlen(&entry->psq, 3 << (fifo * 2)) == 0)
+		entry->traffic_pending_bmp &= ~NBITVAL(fifo);
+	else
+		entry->traffic_pending_bmp |= NBITVAL(fifo);
+
+	entry->send_tim_signal = false;
 	if (entry->traffic_lastreported_bmp != entry->traffic_pending_bmp)
 		entry->send_tim_signal = true;
+	if (send_immediately && entry->send_tim_signal &&
+	    entry->state == BRCMF_FWS_STATE_CLOSE) {
+		/* create a dummy packet and sent that. The traffic          */
+		/* bitmap info will automatically be attached to that packet */
+		len = BRCMF_FWS_TYPE_PKTTAG_LEN + 2 +
+		      BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN + 2 +
+		      4 + fws->drvr->hdrlen;
+		skb = brcmu_pkt_buf_get_skb(len);
+		if (skb == NULL)
+			return false;
+		skb_pull(skb, len);
+		skcb = brcmf_skbcb(skb);
+		skcb->mac = entry;
+		skcb->state = BRCMF_FWS_SKBSTATE_TIM;
+		bus = fws->drvr->bus_if;
+		err = brcmf_fws_hdrpush(fws, skb);
+		if (err == 0)
+			err = brcmf_bus_txdata(bus, skb);
+		if (err)
+			brcmu_pkt_buf_free_skb(skb);
+		return true;
+	}
+	return false;
 }
 
 static void
 brcmf_fws_flow_control_check(struct brcmf_fws_info *fws, struct pktq *pq,
 			     u8 if_id)
 {
-	struct brcmf_if *ifp = fws->drvr->iflist[if_id];
+	struct brcmf_if *ifp = fws->drvr->iflist[!if_id ? 0 : if_id + 1];
 
 	if (WARN_ON(!ifp))
 		return;
 
-	brcmf_dbg(TRACE,
-		  "enter: bssidx=%d, ifidx=%d\n", ifp->bssidx, ifp->ifidx);
-
 	if ((ifp->netif_stop & BRCMF_NETIF_STOP_REASON_FWS_FC) &&
 	    pq->len <= BRCMF_FWS_FLOWCONTROL_LOWATER)
 		brcmf_txflowblock_if(ifp,
 				     BRCMF_NETIF_STOP_REASON_FWS_FC, false);
 	if (!(ifp->netif_stop & BRCMF_NETIF_STOP_REASON_FWS_FC) &&
-	    pq->len >= BRCMF_FWS_FLOWCONTROL_HIWATER)
+	    pq->len >= BRCMF_FWS_FLOWCONTROL_HIWATER) {
+		fws->stats.fws_flow_block++;
 		brcmf_txflowblock_if(ifp, BRCMF_NETIF_STOP_REASON_FWS_FC, true);
+	}
 	return;
 }
 
@@ -862,10 +905,26 @@ static int brcmf_fws_rssi_indicate(struct brcmf_fws_info *fws, s8 rssi)
 	return 0;
 }
 
+/* using macro so sparse checking does not complain
+ * about locking imbalance.
+ */
+#define brcmf_fws_lock(drvr, flags)				\
+do {								\
+	flags = 0;						\
+	spin_lock_irqsave(&((drvr)->fws_spinlock), (flags));	\
+} while (0)
+
+/* using macro so sparse checking does not complain
+ * about locking imbalance.
+ */
+#define brcmf_fws_unlock(drvr, flags) \
+	spin_unlock_irqrestore(&((drvr)->fws_spinlock), (flags))
+
 static
 int brcmf_fws_macdesc_indicate(struct brcmf_fws_info *fws, u8 type, u8 *data)
 {
 	struct brcmf_fws_mac_descriptor *entry, *existing;
+	ulong flags;
 	u8 mac_handle;
 	u8 ifidx;
 	u8 *addr;
@@ -876,34 +935,44 @@ int brcmf_fws_macdesc_indicate(struct brcmf_fws_info *fws, u8 type, u8 *data)
 
 	entry = &fws->desc.nodes[mac_handle & 0x1F];
 	if (type == BRCMF_FWS_TYPE_MACDESC_DEL) {
-		brcmf_dbg(TRACE, "deleting mac %pM idx %d\n", addr, ifidx);
 		if (entry->occupied) {
-			brcmf_fws_mac_desc_cleanup(fws, entry, -1);
-			brcmf_fws_clear_mac_descriptor(entry);
+			brcmf_dbg(TRACE, "deleting %s mac %pM\n",
+				  entry->name, addr);
+			brcmf_fws_lock(fws->drvr, flags);
+			brcmf_fws_macdesc_cleanup(fws, entry, -1);
+			brcmf_fws_macdesc_deinit(entry);
+			brcmf_fws_unlock(fws->drvr, flags);
 		} else
 			fws->stats.mac_update_failed++;
 		return 0;
 	}
 
-	brcmf_dbg(TRACE,
-		  "add mac %pM handle %u idx %d\n", addr, mac_handle, ifidx);
-	existing = brcmf_fws_mac_descriptor_lookup(fws, addr);
+	existing = brcmf_fws_macdesc_lookup(fws, addr);
 	if (IS_ERR(existing)) {
 		if (!entry->occupied) {
+			brcmf_fws_lock(fws->drvr, flags);
 			entry->mac_handle = mac_handle;
-			brcmf_fws_init_mac_descriptor(entry, addr, ifidx);
+			brcmf_fws_macdesc_init(entry, addr, ifidx);
+			brcmf_fws_macdesc_set_name(fws, entry);
 			brcmu_pktq_init(&entry->psq, BRCMF_FWS_PSQ_PREC_COUNT,
 					BRCMF_FWS_PSQ_LEN);
+			brcmf_fws_unlock(fws->drvr, flags);
+			brcmf_dbg(TRACE, "add %s mac %pM\n", entry->name, addr);
 		} else {
 			fws->stats.mac_update_failed++;
 		}
 	} else {
 		if (entry != existing) {
-			brcmf_dbg(TRACE, "relocate mac\n");
+			brcmf_dbg(TRACE, "copy mac %s\n", existing->name);
+			brcmf_fws_lock(fws->drvr, flags);
 			memcpy(entry, existing,
 			       offsetof(struct brcmf_fws_mac_descriptor, psq));
 			entry->mac_handle = mac_handle;
-			brcmf_fws_clear_mac_descriptor(existing);
+			brcmf_fws_macdesc_deinit(existing);
+			brcmf_fws_macdesc_set_name(fws, entry);
+			brcmf_fws_unlock(fws->drvr, flags);
+			brcmf_dbg(TRACE, "relocate %s mac %pM\n", entry->name,
+				  addr);
 		} else {
 			brcmf_dbg(TRACE, "use existing\n");
 			WARN_ON(entry->mac_handle != mac_handle);
@@ -917,8 +986,9 @@ static int brcmf_fws_macdesc_state_indicate(struct brcmf_fws_info *fws,
 					    u8 type, u8 *data)
 {
 	struct brcmf_fws_mac_descriptor *entry;
+	ulong flags;
 	u8 mac_handle;
-	int i;
+	int ret;
 
 	mac_handle = data[0];
 	entry = &fws->desc.nodes[mac_handle & 0x1F];
@@ -926,30 +996,35 @@ static int brcmf_fws_macdesc_state_indicate(struct brcmf_fws_info *fws,
 		fws->stats.mac_ps_update_failed++;
 		return -ESRCH;
 	}
-
-	/* a state update should wipe old credits? */
+	brcmf_fws_lock(fws->drvr, flags);
+	/* a state update should wipe old credits */
 	entry->requested_credit = 0;
+	entry->requested_packet = 0;
 	if (type == BRCMF_FWS_TYPE_MAC_OPEN) {
 		entry->state = BRCMF_FWS_STATE_OPEN;
-		return BRCMF_FWS_RET_OK_SCHEDULE;
+		ret = BRCMF_FWS_RET_OK_SCHEDULE;
 	} else {
 		entry->state = BRCMF_FWS_STATE_CLOSE;
-		for (i = BRCMF_FWS_FIFO_AC_BE; i < NL80211_NUM_ACS; i++)
-			brcmf_fws_tim_update(fws, entry, i);
+		brcmf_fws_tim_update(fws, entry, BRCMF_FWS_FIFO_AC_BK, false);
+		brcmf_fws_tim_update(fws, entry, BRCMF_FWS_FIFO_AC_BE, false);
+		brcmf_fws_tim_update(fws, entry, BRCMF_FWS_FIFO_AC_VI, false);
+		brcmf_fws_tim_update(fws, entry, BRCMF_FWS_FIFO_AC_VO, true);
+		ret = BRCMF_FWS_RET_OK_NOSCHEDULE;
 	}
-	return BRCMF_FWS_RET_OK_NOSCHEDULE;
+	brcmf_fws_unlock(fws->drvr, flags);
+	return ret;
 }
 
 static int brcmf_fws_interface_state_indicate(struct brcmf_fws_info *fws,
 					      u8 type, u8 *data)
 {
 	struct brcmf_fws_mac_descriptor *entry;
+	ulong flags;
 	u8 ifidx;
 	int ret;
 
 	ifidx = data[0];
 
-	brcmf_dbg(TRACE, "enter: ifidx=%d\n", ifidx);
 	if (ifidx >= BRCMF_MAX_IFS) {
 		ret = -ERANGE;
 		goto fail;
@@ -961,17 +1036,26 @@ static int brcmf_fws_interface_state_indicate(struct brcmf_fws_info *fws,
 		goto fail;
 	}
 
+	brcmf_dbg(TRACE, "%s (%d): %s\n", brcmf_fws_get_tlv_name(type), type,
+		  entry->name);
+	brcmf_fws_lock(fws->drvr, flags);
 	switch (type) {
 	case BRCMF_FWS_TYPE_INTERFACE_OPEN:
 		entry->state = BRCMF_FWS_STATE_OPEN;
-		return BRCMF_FWS_RET_OK_SCHEDULE;
+		ret = BRCMF_FWS_RET_OK_SCHEDULE;
+		break;
 	case BRCMF_FWS_TYPE_INTERFACE_CLOSE:
 		entry->state = BRCMF_FWS_STATE_CLOSE;
-		return BRCMF_FWS_RET_OK_NOSCHEDULE;
+		ret = BRCMF_FWS_RET_OK_NOSCHEDULE;
+		break;
 	default:
 		ret = -EINVAL;
-		break;
+		brcmf_fws_unlock(fws->drvr, flags);
+		goto fail;
 	}
+	brcmf_fws_unlock(fws->drvr, flags);
+	return ret;
+
 fail:
 	fws->stats.if_update_failed++;
 	return ret;
@@ -981,6 +1065,7 @@ static int brcmf_fws_request_indicate(struct brcmf_fws_info *fws, u8 type,
 				      u8 *data)
 {
 	struct brcmf_fws_mac_descriptor *entry;
+	ulong flags;
 
 	entry = &fws->desc.nodes[data[1] & 0x1F];
 	if (!entry->occupied) {
@@ -991,15 +1076,51 @@ static int brcmf_fws_request_indicate(struct brcmf_fws_info *fws, u8 type,
 		return -ESRCH;
 	}
 
+	brcmf_dbg(TRACE, "%s (%d): %s cnt %d bmp %d\n",
+		  brcmf_fws_get_tlv_name(type), type, entry->name,
+		  data[0], data[2]);
+	brcmf_fws_lock(fws->drvr, flags);
 	if (type == BRCMF_FWS_TYPE_MAC_REQUEST_CREDIT)
 		entry->requested_credit = data[0];
 	else
 		entry->requested_packet = data[0];
 
 	entry->ac_bitmap = data[2];
+	brcmf_fws_unlock(fws->drvr, flags);
 	return BRCMF_FWS_RET_OK_SCHEDULE;
 }
 
+static void
+brcmf_fws_macdesc_use_req_credit(struct brcmf_fws_mac_descriptor *entry,
+				 struct sk_buff *skb)
+{
+	if (entry->requested_credit > 0) {
+		entry->requested_credit--;
+		brcmf_skb_if_flags_set_field(skb, REQUESTED, 1);
+		brcmf_skb_if_flags_set_field(skb, REQ_CREDIT, 1);
+		if (entry->state != BRCMF_FWS_STATE_CLOSE)
+			brcmf_err("requested credit set while mac not closed!\n");
+	} else if (entry->requested_packet > 0) {
+		entry->requested_packet--;
+		brcmf_skb_if_flags_set_field(skb, REQUESTED, 1);
+		brcmf_skb_if_flags_set_field(skb, REQ_CREDIT, 0);
+		if (entry->state != BRCMF_FWS_STATE_CLOSE)
+			brcmf_err("requested packet set while mac not closed!\n");
+	} else {
+		brcmf_skb_if_flags_set_field(skb, REQUESTED, 0);
+		brcmf_skb_if_flags_set_field(skb, REQ_CREDIT, 0);
+	}
+}
+
+static void brcmf_fws_macdesc_return_req_credit(struct sk_buff *skb)
+{
+	struct brcmf_fws_mac_descriptor *entry = brcmf_skbcb(skb)->mac;
+
+	if ((brcmf_skb_if_flags_get_field(skb, REQ_CREDIT)) &&
+	    (entry->state == BRCMF_FWS_STATE_CLOSE))
+		entry->requested_credit++;
+}
+
 static void brcmf_fws_return_credits(struct brcmf_fws_info *fws,
 				     u8 fifo, u8 credits)
 {
@@ -1010,6 +1131,8 @@ static void brcmf_fws_return_credits(struct brcmf_fws_info *fws,
 	if (!credits)
 		return;
 
+	fws->fifo_credit_map |= 1 << fifo;
+
 	if ((fifo == BRCMF_FWS_FIFO_AC_BE) &&
 	    (fws->credits_borrowed[0])) {
 		for (lender_ac = BRCMF_FWS_FIFO_AC_VO; lender_ac >= 0;
@@ -1031,7 +1154,6 @@ static void brcmf_fws_return_credits(struct brcmf_fws_info *fws,
 		}
 	}
 
-	fws->fifo_credit_map |= 1 << fifo;
 	fws->fifo_credit[fifo] += credits;
 }
 
@@ -1042,27 +1164,6 @@ static void brcmf_fws_schedule_deq(struct brcmf_fws_info *fws)
 		queue_work(fws->fws_wq, &fws->fws_dequeue_work);
 }
 
-static void brcmf_skb_pick_up_credit(struct brcmf_fws_info *fws, int fifo,
-				     struct sk_buff *p)
-{
-	struct brcmf_fws_mac_descriptor *entry = brcmf_skbcb(p)->mac;
-
-	if (brcmf_skbcb(p)->if_flags & BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK) {
-		if (fws->fcmode != BRCMF_FWS_FCMODE_IMPLIED_CREDIT)
-			return;
-		brcmf_fws_return_credits(fws, fifo, 1);
-	} else {
-		/*
-		 * if this packet did not count against FIFO credit, it
-		 * must have taken a requested_credit from the destination
-		 * entry (for pspoll etc.)
-		 */
-		if (!brcmf_skb_if_flags_get_field(p, REQUESTED))
-			entry->requested_credit++;
-	}
-	brcmf_fws_schedule_deq(fws);
-}
-
 static int brcmf_fws_enq(struct brcmf_fws_info *fws,
 			 enum brcmf_fws_skb_state state, int fifo,
 			 struct sk_buff *p)
@@ -1078,7 +1179,7 @@ static int brcmf_fws_enq(struct brcmf_fws_info *fws,
 		return -ENOENT;
 	}
 
-	brcmf_dbg(TRACE, "enter: ea=%pM, qlen=%d\n", entry->ea, entry->psq.len);
+	brcmf_dbg(DATA, "enter: fifo %d skb %p\n", fifo, p);
 	if (state == BRCMF_FWS_SKBSTATE_SUPPRESSED) {
 		prec += 1;
 		qfull_stat = &fws->stats.supprq_full_error;
@@ -1095,14 +1196,12 @@ static int brcmf_fws_enq(struct brcmf_fws_info *fws,
 
 	/* update the sk_buff state */
 	brcmf_skbcb(p)->state = state;
-	if (state == BRCMF_FWS_SKBSTATE_SUPPRESSED)
-		entry->suppress_count++;
 
 	/*
 	 * A packet has been pushed so update traffic
 	 * availability bitmap, if applicable
 	 */
-	brcmf_fws_tim_update(fws, entry, fifo);
+	brcmf_fws_tim_update(fws, entry, fifo, true);
 	brcmf_fws_flow_control_check(fws, &entry->psq,
 				     brcmf_skb_if_flags_get_field(p, INDEX));
 	return 0;
@@ -1113,7 +1212,6 @@ static struct sk_buff *brcmf_fws_deq(struct brcmf_fws_info *fws, int fifo)
 	struct brcmf_fws_mac_descriptor *table;
 	struct brcmf_fws_mac_descriptor *entry;
 	struct sk_buff *p;
-	int use_credit = 1;
 	int num_nodes;
 	int node_pos;
 	int prec_out;
@@ -1127,7 +1225,7 @@ static struct sk_buff *brcmf_fws_deq(struct brcmf_fws_info *fws, int fifo)
 	for (i = 0; i < num_nodes; i++) {
 		entry = &table[(node_pos + i) % num_nodes];
 		if (!entry->occupied ||
-		    brcmf_fws_mac_desc_closed(fws, entry, fifo))
+		    brcmf_fws_macdesc_closed(fws, entry, fifo))
 			continue;
 
 		if (entry->suppressed)
@@ -1137,9 +1235,8 @@ static struct sk_buff *brcmf_fws_deq(struct brcmf_fws_info *fws, int fifo)
 		p = brcmu_pktq_mdeq(&entry->psq, pmsk << (fifo * 2), &prec_out);
 		if (p == NULL) {
 			if (entry->suppressed) {
-				if (entry->suppr_transit_count >
-				    entry->suppress_count)
-					return NULL;
+				if (entry->suppr_transit_count)
+					continue;
 				entry->suppressed = false;
 				p = brcmu_pktq_mdeq(&entry->psq,
 						    1 << (fifo * 2), &prec_out);
@@ -1148,26 +1245,7 @@ static struct sk_buff *brcmf_fws_deq(struct brcmf_fws_info *fws, int fifo)
 		if  (p == NULL)
 			continue;
 
-		/* did the packet come from suppress sub-queue? */
-		if (entry->requested_credit > 0) {
-			entry->requested_credit--;
-			/*
-			 * if the packet was pulled out while destination is in
-			 * closed state but had a non-zero packets requested,
-			 * then this should not count against the FIFO credit.
-			 * That is due to the fact that the firmware will
-			 * most likely hold onto this packet until a suitable
-			 * time later to push it to the appropriate AC FIFO.
-			 */
-			if (entry->state == BRCMF_FWS_STATE_CLOSE)
-				use_credit = 0;
-		} else if (entry->requested_packet > 0) {
-			entry->requested_packet--;
-			brcmf_skb_if_flags_set_field(p, REQUESTED, 1);
-			if (entry->state == BRCMF_FWS_STATE_CLOSE)
-				use_credit = 0;
-		}
-		brcmf_skb_if_flags_set_field(p, CREDITCHECK, use_credit);
+		brcmf_fws_macdesc_use_req_credit(entry, p);
 
 		/* move dequeue position to ensure fair round-robin */
 		fws->deq_node_pos[fifo] = (node_pos + i + 1) % num_nodes;
@@ -1179,7 +1257,7 @@ static struct sk_buff *brcmf_fws_deq(struct brcmf_fws_info *fws, int fifo)
 		 * A packet has been picked up, update traffic
 		 * availability bitmap, if applicable
 		 */
-		brcmf_fws_tim_update(fws, entry, fifo);
+		brcmf_fws_tim_update(fws, entry, fifo, false);
 
 		/*
 		 * decrement total enqueued fifo packets and
@@ -1192,7 +1270,7 @@ static struct sk_buff *brcmf_fws_deq(struct brcmf_fws_info *fws, int fifo)
 	}
 	p = NULL;
 done:
-	brcmf_dbg(TRACE, "exit: fifo %d skb %p\n", fifo, p);
+	brcmf_dbg(DATA, "exit: fifo %d skb %p\n", fifo, p);
 	return p;
 }
 
@@ -1202,22 +1280,26 @@ static int brcmf_fws_txstatus_suppressed(struct brcmf_fws_info *fws, int fifo,
 	struct brcmf_fws_mac_descriptor *entry = brcmf_skbcb(skb)->mac;
 	u32 hslot;
 	int ret;
+	u8 ifidx;
 
 	hslot = brcmf_skb_htod_tag_get_field(skb, HSLOT);
 
 	/* this packet was suppressed */
-	if (!entry->suppressed || entry->generation != genbit) {
+	if (!entry->suppressed) {
 		entry->suppressed = true;
-		entry->suppress_count = brcmu_pktq_mlen(&entry->psq,
-							1 << (fifo * 2 + 1));
 		entry->suppr_transit_count = entry->transit_count;
+		brcmf_dbg(DATA, "suppress %s: transit %d\n",
+			  entry->name, entry->transit_count);
 	}
 
 	entry->generation = genbit;
 
-	ret = brcmf_fws_enq(fws, BRCMF_FWS_SKBSTATE_SUPPRESSED, fifo, skb);
+	ret = brcmf_proto_hdrpull(fws->drvr, false, &ifidx, skb);
+	if (ret == 0)
+		ret = brcmf_fws_enq(fws, BRCMF_FWS_SKBSTATE_SUPPRESSED, fifo,
+				    skb);
 	if (ret != 0) {
-		/* suppress q is full, drop this packet */
+		/* suppress q is full or hdrpull failed, drop this packet */
 		brcmf_fws_hanger_poppkt(&fws->hanger, hslot, &skb,
 					true);
 	} else {
@@ -1225,26 +1307,24 @@ static int brcmf_fws_txstatus_suppressed(struct brcmf_fws_info *fws, int fifo,
 		 * Mark suppressed to avoid a double free during
 		 * wlfc cleanup
 		 */
-		brcmf_fws_hanger_mark_suppressed(&fws->hanger, hslot,
-						 genbit);
-		entry->suppress_count++;
+		brcmf_fws_hanger_mark_suppressed(&fws->hanger, hslot);
 	}
 
 	return ret;
 }
 
 static int
-brcmf_fws_txstatus_process(struct brcmf_fws_info *fws, u8 flags, u32 hslot,
+brcmf_fws_txs_process(struct brcmf_fws_info *fws, u8 flags, u32 hslot,
 			   u32 genbit)
 {
 	u32 fifo;
 	int ret;
 	bool remove_from_hanger = true;
 	struct sk_buff *skb;
+	struct brcmf_skbuff_cb *skcb;
 	struct brcmf_fws_mac_descriptor *entry = NULL;
 
-	brcmf_dbg(TRACE, "status: flags=0x%X, hslot=%d\n",
-		  flags, hslot);
+	brcmf_dbg(DATA, "flags %d\n", flags);
 
 	if (flags == BRCMF_FWS_TXSTATUS_DISCARD)
 		fws->stats.txs_discard++;
@@ -1256,6 +1336,8 @@ brcmf_fws_txstatus_process(struct brcmf_fws_info *fws, u8 flags, u32 hslot,
 		remove_from_hanger = false;
 	} else if (flags == BRCMF_FWS_TXSTATUS_FW_TOSSED)
 		fws->stats.txs_tossed++;
+	else if (flags == BRCMF_FWS_TXSTATUS_HOST_TOSSED)
+		fws->stats.txs_host_tossed++;
 	else
 		brcmf_err("unexpected txstatus\n");
 
@@ -1266,32 +1348,42 @@ brcmf_fws_txstatus_process(struct brcmf_fws_info *fws, u8 flags, u32 hslot,
 		return ret;
 	}
 
-	entry = brcmf_skbcb(skb)->mac;
+	skcb = brcmf_skbcb(skb);
+	entry = skcb->mac;
 	if (WARN_ON(!entry)) {
 		brcmu_pkt_buf_free_skb(skb);
 		return -EINVAL;
 	}
+	entry->transit_count--;
+	if (entry->suppressed && entry->suppr_transit_count)
+		entry->suppr_transit_count--;
+
+	brcmf_dbg(DATA, "%s flags %X htod %X\n", entry->name, skcb->if_flags,
+		  skcb->htod);
 
 	/* pick up the implicit credit from this packet */
 	fifo = brcmf_skb_htod_tag_get_field(skb, FIFO);
-	brcmf_skb_pick_up_credit(fws, fifo, skb);
+	if ((fws->fcmode == BRCMF_FWS_FCMODE_IMPLIED_CREDIT) ||
+	    (brcmf_skb_if_flags_get_field(skb, REQ_CREDIT)) ||
+	    (flags == BRCMF_FWS_TXSTATUS_HOST_TOSSED)) {
+		brcmf_fws_return_credits(fws, fifo, 1);
+		brcmf_fws_schedule_deq(fws);
+	}
+	brcmf_fws_macdesc_return_req_credit(skb);
 
 	if (!remove_from_hanger)
 		ret = brcmf_fws_txstatus_suppressed(fws, fifo, skb, genbit);
 
-	if (remove_from_hanger || ret) {
-		entry->transit_count--;
-		if (entry->suppressed)
-			entry->suppr_transit_count--;
-
+	if (remove_from_hanger || ret)
 		brcmf_txfinalize(fws->drvr, skb, true);
-	}
+
 	return 0;
 }
 
 static int brcmf_fws_fifocreditback_indicate(struct brcmf_fws_info *fws,
 					     u8 *data)
 {
+	ulong flags;
 	int i;
 
 	if (fws->fcmode != BRCMF_FWS_FCMODE_EXPLICIT_CREDIT) {
@@ -1299,17 +1391,20 @@ static int brcmf_fws_fifocreditback_indicate(struct brcmf_fws_info *fws,
 		return BRCMF_FWS_RET_OK_NOSCHEDULE;
 	}
 
-	brcmf_dbg(TRACE, "enter: data %pM\n", data);
+	brcmf_dbg(DATA, "enter: data %pM\n", data);
+	brcmf_fws_lock(fws->drvr, flags);
 	for (i = 0; i < BRCMF_FWS_FIFO_COUNT; i++)
 		brcmf_fws_return_credits(fws, i, data[i]);
 
-	brcmf_dbg(INFO, "map: credit %x delay %x\n", fws->fifo_credit_map,
+	brcmf_dbg(DATA, "map: credit %x delay %x\n", fws->fifo_credit_map,
 		  fws->fifo_delay_map);
+	brcmf_fws_unlock(fws->drvr, flags);
 	return BRCMF_FWS_RET_OK_SCHEDULE;
 }
 
 static int brcmf_fws_txstatus_indicate(struct brcmf_fws_info *fws, u8 *data)
 {
+	ulong lflags;
 	__le32 status_le;
 	u32 status;
 	u32 hslot;
@@ -1323,7 +1418,10 @@ static int brcmf_fws_txstatus_indicate(struct brcmf_fws_info *fws, u8 *data)
 	hslot = brcmf_txstatus_get_field(status, HSLOT);
 	genbit = brcmf_txstatus_get_field(status, GENERATION);
 
-	return brcmf_fws_txstatus_process(fws, flags, hslot, genbit);
+	brcmf_fws_lock(fws->drvr, lflags);
+	brcmf_fws_txs_process(fws, flags, hslot, genbit);
+	brcmf_fws_unlock(fws->drvr, lflags);
+	return BRCMF_FWS_RET_OK_NOSCHEDULE;
 }
 
 static int brcmf_fws_dbg_seqnum_check(struct brcmf_fws_info *fws, u8 *data)
@@ -1331,26 +1429,11 @@ static int brcmf_fws_dbg_seqnum_check(struct brcmf_fws_info *fws, u8 *data)
 	__le32 timestamp;
 
 	memcpy(&timestamp, &data[2], sizeof(timestamp));
-	brcmf_dbg(INFO, "received: seq %d, timestamp %d\n", data[1],
+	brcmf_dbg(CTL, "received: seq %d, timestamp %d\n", data[1],
 		  le32_to_cpu(timestamp));
 	return 0;
 }
 
-/* using macro so sparse checking does not complain
- * about locking imbalance.
- */
-#define brcmf_fws_lock(drvr, flags)				\
-do {								\
-	flags = 0;						\
-	spin_lock_irqsave(&((drvr)->fws_spinlock), (flags));	\
-} while (0)
-
-/* using macro so sparse checking does not complain
- * about locking imbalance.
- */
-#define brcmf_fws_unlock(drvr, flags) \
-	spin_unlock_irqrestore(&((drvr)->fws_spinlock), (flags))
-
 static int brcmf_fws_notify_credit_map(struct brcmf_if *ifp,
 				       const struct brcmf_event_msg *e,
 				       void *data)
@@ -1364,6 +1447,10 @@ static int brcmf_fws_notify_credit_map(struct brcmf_if *ifp,
 		brcmf_err("event payload too small (%d)\n", e->datalen);
 		return -EINVAL;
 	}
+	if (fws->creditmap_received)
+		return 0;
+
+	fws->creditmap_received = true;
 
 	brcmf_dbg(TRACE, "enter: credits %pM\n", credits);
 	brcmf_fws_lock(ifp->drvr, flags);
@@ -1379,11 +1466,24 @@ static int brcmf_fws_notify_credit_map(struct brcmf_if *ifp,
 	return 0;
 }
 
+static int brcmf_fws_notify_bcmc_credit_support(struct brcmf_if *ifp,
+						const struct brcmf_event_msg *e,
+						void *data)
+{
+	struct brcmf_fws_info *fws = ifp->drvr->fws;
+	ulong flags;
+
+	brcmf_fws_lock(ifp->drvr, flags);
+	if (fws)
+		fws->bcmc_credit_check = true;
+	brcmf_fws_unlock(ifp->drvr, flags);
+	return 0;
+}
+
 int brcmf_fws_hdrpull(struct brcmf_pub *drvr, int ifidx, s16 signal_len,
 		      struct sk_buff *skb)
 {
 	struct brcmf_fws_info *fws = drvr->fws;
-	ulong flags;
 	u8 *signal_data;
 	s16 data_len;
 	u8 type;
@@ -1392,7 +1492,7 @@ int brcmf_fws_hdrpull(struct brcmf_pub *drvr, int ifidx, s16 signal_len,
 	s32 status;
 	s32 err;
 
-	brcmf_dbg(TRACE, "enter: ifidx %d, skblen %u, sig %d\n",
+	brcmf_dbg(HDRS, "enter: ifidx %d, skblen %u, sig %d\n",
 		  ifidx, skb->len, signal_len);
 
 	WARN_ON(signal_len > skb->len);
@@ -1403,9 +1503,6 @@ int brcmf_fws_hdrpull(struct brcmf_pub *drvr, int ifidx, s16 signal_len,
 		return 0;
 	}
 
-	/* lock during tlv parsing */
-	brcmf_fws_lock(drvr, flags);
-
 	fws->stats.header_pulls++;
 	data_len = signal_len;
 	signal_data = skb->data;
@@ -1426,14 +1523,15 @@ int brcmf_fws_hdrpull(struct brcmf_pub *drvr, int ifidx, s16 signal_len,
 		len = signal_data[1];
 		data = signal_data + 2;
 
-		brcmf_dbg(INFO, "tlv type=%d (%s), len=%d, data[0]=%d\n", type,
-			  brcmf_fws_get_tlv_name(type), len, *data);
+		brcmf_dbg(HDRS, "tlv type=%s (%d), len=%d (%d)\n",
+			  brcmf_fws_get_tlv_name(type), type, len,
+			  brcmf_fws_get_tlv_len(fws, type));
 
 		/* abort parsing when length invalid */
 		if (data_len < len + 2)
 			break;
 
-		if (len != brcmf_fws_get_tlv_len(fws, type))
+		if (len < brcmf_fws_get_tlv_len(fws, type))
 			break;
 
 		err = BRCMF_FWS_RET_OK_NOSCHEDULE;
@@ -1498,203 +1596,74 @@ int brcmf_fws_hdrpull(struct brcmf_pub *drvr, int ifidx, s16 signal_len,
 	if (skb->len == 0)
 		fws->stats.header_only_pkt++;
 
-	brcmf_fws_unlock(drvr, flags);
-	return 0;
-}
-
-static int brcmf_fws_hdrpush(struct brcmf_fws_info *fws, struct sk_buff *skb)
-{
-	struct brcmf_fws_mac_descriptor *entry = brcmf_skbcb(skb)->mac;
-	u8 *wlh;
-	u16 data_offset = 0;
-	u8 fillers;
-	__le32 pkttag = cpu_to_le32(brcmf_skbcb(skb)->htod);
-
-	brcmf_dbg(TRACE, "enter: ea=%pM, ifidx=%u, pkttag=0x%08X\n",
-		  entry->ea, entry->interface_id, le32_to_cpu(pkttag));
-	if (entry->send_tim_signal)
-		data_offset += 2 + BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN;
-
-	/* +2 is for Type[1] and Len[1] in TLV, plus TIM signal */
-	data_offset += 2 + BRCMF_FWS_TYPE_PKTTAG_LEN;
-	fillers = round_up(data_offset, 4) - data_offset;
-	data_offset += fillers;
-
-	skb_push(skb, data_offset);
-	wlh = skb->data;
-
-	wlh[0] = BRCMF_FWS_TYPE_PKTTAG;
-	wlh[1] = BRCMF_FWS_TYPE_PKTTAG_LEN;
-	memcpy(&wlh[2], &pkttag, sizeof(pkttag));
-	wlh += BRCMF_FWS_TYPE_PKTTAG_LEN + 2;
-
-	if (entry->send_tim_signal) {
-		entry->send_tim_signal = 0;
-		wlh[0] = BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP;
-		wlh[1] = BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN;
-		wlh[2] = entry->mac_handle;
-		wlh[3] = entry->traffic_pending_bmp;
-		wlh += BRCMF_FWS_TYPE_PENDING_TRAFFIC_BMP_LEN + 2;
-		entry->traffic_lastreported_bmp = entry->traffic_pending_bmp;
-	}
-	if (fillers)
-		memset(wlh, BRCMF_FWS_TYPE_FILLER, fillers);
-
-	brcmf_proto_hdrpush(fws->drvr, brcmf_skb_if_flags_get_field(skb, INDEX),
-			    data_offset >> 2, skb);
 	return 0;
 }
 
-static int brcmf_fws_precommit_skb(struct brcmf_fws_info *fws, int fifo,
+static void brcmf_fws_precommit_skb(struct brcmf_fws_info *fws, int fifo,
 				   struct sk_buff *p)
 {
 	struct brcmf_skbuff_cb *skcb = brcmf_skbcb(p);
 	struct brcmf_fws_mac_descriptor *entry = skcb->mac;
-	int rc = 0;
-	bool header_needed;
-	int hslot = BRCMF_FWS_HANGER_MAXITEMS;
-	u8 free_ctr;
-	u8 ifidx;
 	u8 flags;
 
-	header_needed = skcb->state != BRCMF_FWS_SKBSTATE_SUPPRESSED;
-
-	if (header_needed) {
-		/* obtaining free slot may fail, but that will be caught
-		 * by the hanger push. This assures the packet has a BDC
-		 * header upon return.
-		 */
-		hslot = brcmf_fws_hanger_get_free_slot(&fws->hanger);
-		free_ctr = entry->seq[fifo];
-		brcmf_skb_htod_tag_set_field(p, HSLOT, hslot);
-		brcmf_skb_htod_tag_set_field(p, FREERUN, free_ctr);
-		brcmf_skb_htod_tag_set_field(p, GENERATION, 1);
-		entry->transit_count++;
-	}
 	brcmf_skb_if_flags_set_field(p, TRANSMIT, 1);
-	brcmf_skb_htod_tag_set_field(p, FIFO, fifo);
-
+	brcmf_skb_htod_tag_set_field(p, GENERATION, entry->generation);
 	flags = BRCMF_FWS_HTOD_FLAG_PKTFROMHOST;
-	if (!(skcb->if_flags & BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK)) {
+	if (brcmf_skb_if_flags_get_field(p, REQUESTED)) {
 		/*
-		Indicate that this packet is being sent in response to an
-		explicit request from the firmware side.
-		*/
+		 * Indicate that this packet is being sent in response to an
+		 * explicit request from the firmware side.
+		 */
 		flags |= BRCMF_FWS_HTOD_FLAG_PKT_REQUESTED;
 	}
 	brcmf_skb_htod_tag_set_field(p, FLAGS, flags);
-	if (header_needed) {
-		brcmf_fws_hdrpush(fws, p);
-		rc = brcmf_fws_hanger_pushpkt(&fws->hanger, p, hslot);
-		if (rc)
-			brcmf_err("hanger push failed: rc=%d\n", rc);
-	} else {
-		int gen;
-
-		/* remove old header */
-		rc = brcmf_proto_hdrpull(fws->drvr, false, &ifidx, p);
-		if (rc == 0) {
-			hslot = brcmf_skb_htod_tag_get_field(p, HSLOT);
-			brcmf_fws_hanger_get_genbit(&fws->hanger, p,
-						    hslot, &gen);
-			brcmf_skb_htod_tag_set_field(p, GENERATION, gen);
-
-			/* push new header */
-			brcmf_fws_hdrpush(fws, p);
-		}
-	}
-
-	return rc;
+	brcmf_fws_hdrpush(fws, p);
 }
 
-static void
-brcmf_fws_rollback_toq(struct brcmf_fws_info *fws, struct sk_buff *skb)
+static void brcmf_fws_rollback_toq(struct brcmf_fws_info *fws,
+				   struct sk_buff *skb, int fifo)
 {
-	/*
-	put the packet back to the head of queue
-
-	- suppressed packet goes back to suppress sub-queue
-	- pull out the header, if new or delayed packet
-
-	Note: hslot is used only when header removal is done.
-	*/
 	struct brcmf_fws_mac_descriptor *entry;
-	enum brcmf_fws_skb_state state;
 	struct sk_buff *pktout;
+	int qidx, hslot;
 	int rc = 0;
-	int fifo;
-	int hslot;
-	u8 ifidx;
 
-	fifo = brcmf_skb_if_flags_get_field(skb, FIFO);
-	state = brcmf_skbcb(skb)->state;
 	entry = brcmf_skbcb(skb)->mac;
-
-	if (entry != NULL) {
-		if (state == BRCMF_FWS_SKBSTATE_SUPPRESSED) {
-			/* wl-header is saved for suppressed packets */
-			pktout = brcmu_pktq_penq_head(&entry->psq, 2 * fifo + 1,
-						      skb);
-			if (pktout == NULL) {
-				brcmf_err("suppress queue full\n");
-				rc = -ENOSPC;
-			}
-		} else {
-			hslot = brcmf_skb_htod_tag_get_field(skb, HSLOT);
-
-			/* remove header first */
-			rc = brcmf_proto_hdrpull(fws->drvr, false, &ifidx, skb);
-			if (rc) {
-				brcmf_err("header removal failed\n");
-				/* free the hanger slot */
-				brcmf_fws_hanger_poppkt(&fws->hanger, hslot,
-							&pktout, true);
-				rc = -EINVAL;
-				goto fail;
-			}
-
-			/* delay-q packets are going to delay-q */
-			pktout = brcmu_pktq_penq_head(&entry->psq,
-						      2 * fifo, skb);
-			if (pktout == NULL) {
-				brcmf_err("delay queue full\n");
-				rc = -ENOSPC;
-			}
-
-			/* free the hanger slot */
-			brcmf_fws_hanger_poppkt(&fws->hanger, hslot, &pktout,
-						true);
-
-			/* decrement sequence count */
-			entry->seq[fifo]--;
+	if (entry->occupied) {
+		qidx = 2 * fifo;
+		if (brcmf_skbcb(skb)->state == BRCMF_FWS_SKBSTATE_SUPPRESSED)
+			qidx++;
+
+		pktout = brcmu_pktq_penq_head(&entry->psq, qidx, skb);
+		if (pktout == NULL) {
+			brcmf_err("%s queue %d full\n", entry->name, qidx);
+			rc = -ENOSPC;
 		}
-		/*
-		if this packet did not count against FIFO credit, it must have
-		taken a requested_credit from the firmware (for pspoll etc.)
-		*/
-		if (!(brcmf_skbcb(skb)->if_flags &
-		      BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK))
-			entry->requested_credit++;
 	} else {
-		brcmf_err("no mac entry linked\n");
+		brcmf_err("%s entry removed\n", entry->name);
 		rc = -ENOENT;
 	}
 
-
-fail:
 	if (rc) {
-		brcmf_txfinalize(fws->drvr, skb, false);
 		fws->stats.rollback_failed++;
-	} else
+		hslot = brcmf_skb_htod_tag_get_field(skb, HSLOT);
+		brcmf_fws_txs_process(fws, BRCMF_FWS_TXSTATUS_HOST_TOSSED,
+				      hslot, 0);
+	} else {
 		fws->stats.rollback_success++;
+		brcmf_fws_return_credits(fws, fifo, 1);
+		brcmf_fws_macdesc_return_req_credit(skb);
+	}
 }
 
 static int brcmf_fws_borrow_credit(struct brcmf_fws_info *fws)
 {
 	int lender_ac;
 
-	if (time_after(fws->borrow_defer_timestamp, jiffies))
+	if (time_after(fws->borrow_defer_timestamp, jiffies)) {
+		fws->fifo_credit_map &= ~(1 << BRCMF_FWS_FIFO_AC_BE);
 		return -ENAVAIL;
+	}
 
 	for (lender_ac = 0; lender_ac <= BRCMF_FWS_FIFO_AC_VO; lender_ac++) {
 		if (fws->fifo_credit[lender_ac]) {
@@ -1702,66 +1671,15 @@ static int brcmf_fws_borrow_credit(struct brcmf_fws_info *fws)
 			fws->fifo_credit[lender_ac]--;
 			if (fws->fifo_credit[lender_ac] == 0)
 				fws->fifo_credit_map &= ~(1 << lender_ac);
-			brcmf_dbg(TRACE, "borrow credit from: %d\n", lender_ac);
+			fws->fifo_credit_map |= (1 << BRCMF_FWS_FIFO_AC_BE);
+			brcmf_dbg(DATA, "borrow credit from: %d\n", lender_ac);
 			return 0;
 		}
 	}
+	fws->fifo_credit_map &= ~(1 << BRCMF_FWS_FIFO_AC_BE);
 	return -ENAVAIL;
 }
 
-static int brcmf_fws_consume_credit(struct brcmf_fws_info *fws, int fifo,
-				    struct sk_buff *skb)
-{
-	struct brcmf_fws_mac_descriptor *entry = brcmf_skbcb(skb)->mac;
-	int *credit = &fws->fifo_credit[fifo];
-	int use_credit = 1;
-
-	brcmf_dbg(TRACE, "enter: ac=%d, credits=%d\n", fifo, *credit);
-
-	if (entry->requested_credit > 0) {
-		/*
-		 * if the packet was pulled out while destination is in
-		 * closed state but had a non-zero packets requested,
-		 * then this should not count against the FIFO credit.
-		 * That is due to the fact that the firmware will
-		 * most likely hold onto this packet until a suitable
-		 * time later to push it to the appropriate AC FIFO.
-		 */
-		entry->requested_credit--;
-		if (entry->state == BRCMF_FWS_STATE_CLOSE)
-			use_credit = 0;
-	} else if (entry->requested_packet > 0) {
-		entry->requested_packet--;
-		brcmf_skb_if_flags_set_field(skb, REQUESTED, 1);
-		if (entry->state == BRCMF_FWS_STATE_CLOSE)
-			use_credit = 0;
-	}
-	brcmf_skb_if_flags_set_field(skb, CREDITCHECK, use_credit);
-	if (!use_credit) {
-		brcmf_dbg(TRACE, "exit: no creditcheck set\n");
-		return 0;
-	}
-
-	if (fifo != BRCMF_FWS_FIFO_AC_BE)
-		fws->borrow_defer_timestamp = jiffies +
-					      BRCMF_FWS_BORROW_DEFER_PERIOD;
-
-	if (!(*credit)) {
-		/* Try to borrow a credit from other queue */
-		if (fifo == BRCMF_FWS_FIFO_AC_BE &&
-		    brcmf_fws_borrow_credit(fws) == 0)
-			return 0;
-
-		brcmf_dbg(TRACE, "exit: ac=%d, credits depleted\n", fifo);
-		return -ENAVAIL;
-	}
-	(*credit)--;
-	if (!(*credit))
-		fws->fifo_credit_map &= ~(1 << fifo);
-	brcmf_dbg(TRACE, "exit: ac=%d, credits=%d\n", fifo, *credit);
-	return 0;
-}
-
 static int brcmf_fws_commit_skb(struct brcmf_fws_info *fws, int fifo,
 				struct sk_buff *skb)
 {
@@ -1769,32 +1687,51 @@ static int brcmf_fws_commit_skb(struct brcmf_fws_info *fws, int fifo,
 	struct brcmf_fws_mac_descriptor *entry;
 	struct brcmf_bus *bus = fws->drvr->bus_if;
 	int rc;
+	u8 ifidx;
 
 	entry = skcb->mac;
 	if (IS_ERR(entry))
 		return PTR_ERR(entry);
 
-	rc = brcmf_fws_precommit_skb(fws, fifo, skb);
+	brcmf_fws_precommit_skb(fws, fifo, skb);
+	rc = brcmf_bus_txdata(bus, skb);
+	brcmf_dbg(DATA, "%s flags %X htod %X bus_tx %d\n", entry->name,
+		  skcb->if_flags, skcb->htod, rc);
 	if (rc < 0) {
-		fws->stats.generic_error++;
+		brcmf_proto_hdrpull(fws->drvr, false, &ifidx, skb);
 		goto rollback;
 	}
 
-	rc = brcmf_bus_txdata(bus, skb);
-	if (rc < 0)
-		goto rollback;
-
-	entry->seq[fifo]++;
+	entry->transit_count++;
+	if (entry->suppressed)
+		entry->suppr_transit_count++;
 	fws->stats.pkt2bus++;
-	if (brcmf_skbcb(skb)->if_flags & BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK) {
-		fws->stats.send_pkts[fifo]++;
-		fws->stats.fifo_credits_sent[fifo]++;
-	}
+	fws->stats.send_pkts[fifo]++;
+	if (brcmf_skb_if_flags_get_field(skb, REQUESTED))
+		fws->stats.requested_sent[fifo]++;
 
 	return rc;
 
 rollback:
-	brcmf_fws_rollback_toq(fws, skb);
+	brcmf_fws_rollback_toq(fws, skb, fifo);
+	return rc;
+}
+
+static int brcmf_fws_assign_htod(struct brcmf_fws_info *fws, struct sk_buff *p,
+				  int fifo)
+{
+	struct brcmf_skbuff_cb *skcb = brcmf_skbcb(p);
+	int rc, hslot;
+
+	hslot = brcmf_fws_hanger_get_free_slot(&fws->hanger);
+	brcmf_skb_htod_tag_set_field(p, HSLOT, hslot);
+	brcmf_skb_htod_tag_set_field(p, FREERUN, skcb->mac->seq[fifo]);
+	brcmf_skb_htod_tag_set_field(p, FIFO, fifo);
+	rc = brcmf_fws_hanger_pushpkt(&fws->hanger, p, hslot);
+	if (!rc)
+		skcb->mac->seq[fifo]++;
+	else
+		fws->stats.generic_error++;
 	return rc;
 }
 
@@ -1826,29 +1763,25 @@ int brcmf_fws_process_skb(struct brcmf_if *ifp, struct sk_buff *skb)
 
 	/* set control buffer information */
 	skcb->if_flags = 0;
-	skcb->mac = brcmf_fws_find_mac_desc(fws, ifp, eh->h_dest);
 	skcb->state = BRCMF_FWS_SKBSTATE_NEW;
 	brcmf_skb_if_flags_set_field(skb, INDEX, ifp->ifidx);
 	if (!multicast)
 		fifo = brcmf_fws_prio2fifo[skb->priority];
-	brcmf_skb_if_flags_set_field(skb, FIFO, fifo);
-
-	brcmf_dbg(TRACE, "ea=%pM, multi=%d, fifo=%d\n", eh->h_dest,
-		  multicast, fifo);
 
 	brcmf_fws_lock(drvr, flags);
-	if (skcb->mac->suppressed ||
-	    brcmf_fws_mac_desc_closed(fws, skcb->mac, fifo) ||
-	    brcmu_pktq_mlen(&skcb->mac->psq, 3 << (fifo * 2)) ||
-	    (!multicast &&
-	     brcmf_fws_consume_credit(fws, fifo, skb) < 0)) {
-		/* enqueue the packet in delayQ */
-		drvr->fws->fifo_delay_map |= 1 << fifo;
+	if (fifo != BRCMF_FWS_FIFO_AC_BE && fifo < BRCMF_FWS_FIFO_BCMC)
+		fws->borrow_defer_timestamp = jiffies +
+					      BRCMF_FWS_BORROW_DEFER_PERIOD;
+
+	skcb->mac = brcmf_fws_macdesc_find(fws, ifp, eh->h_dest);
+	brcmf_dbg(DATA, "%s mac %pM multi %d fifo %d\n", skcb->mac->name,
+		  eh->h_dest, multicast, fifo);
+	if (!brcmf_fws_assign_htod(fws, skb, fifo)) {
 		brcmf_fws_enq(fws, BRCMF_FWS_SKBSTATE_DELAYED, fifo, skb);
+		brcmf_fws_schedule_deq(fws);
 	} else {
-		if (brcmf_fws_commit_skb(fws, fifo, skb))
-			if (!multicast)
-				brcmf_skb_pick_up_credit(fws, fifo, skb);
+		brcmf_err("drop skb: no hanger slot\n");
+		brcmu_pkt_buf_free_skb(skb);
 	}
 	brcmf_fws_unlock(drvr, flags);
 	return 0;
@@ -1862,7 +1795,7 @@ void brcmf_fws_reset_interface(struct brcmf_if *ifp)
 	if (!entry)
 		return;
 
-	brcmf_fws_init_mac_descriptor(entry, ifp->mac_addr, ifp->ifidx);
+	brcmf_fws_macdesc_init(entry, ifp->mac_addr, ifp->ifidx);
 }
 
 void brcmf_fws_add_interface(struct brcmf_if *ifp)
@@ -1870,16 +1803,16 @@ void brcmf_fws_add_interface(struct brcmf_if *ifp)
 	struct brcmf_fws_info *fws = ifp->drvr->fws;
 	struct brcmf_fws_mac_descriptor *entry;
 
-	brcmf_dbg(TRACE, "enter: idx=%d, mac=%pM\n",
-		  ifp->bssidx, ifp->mac_addr);
 	if (!ifp->ndev || !ifp->drvr->fw_signals)
 		return;
 
 	entry = &fws->desc.iface[ifp->ifidx];
 	ifp->fws_desc = entry;
-	brcmf_fws_init_mac_descriptor(entry, ifp->mac_addr, ifp->ifidx);
+	brcmf_fws_macdesc_init(entry, ifp->mac_addr, ifp->ifidx);
+	brcmf_fws_macdesc_set_name(fws, entry);
 	brcmu_pktq_init(&entry->psq, BRCMF_FWS_PSQ_PREC_COUNT,
 			BRCMF_FWS_PSQ_LEN);
+	brcmf_dbg(TRACE, "added %s\n", entry->name);
 }
 
 void brcmf_fws_del_interface(struct brcmf_if *ifp)
@@ -1887,13 +1820,13 @@ void brcmf_fws_del_interface(struct brcmf_if *ifp)
 	struct brcmf_fws_mac_descriptor *entry = ifp->fws_desc;
 	ulong flags;
 
-	brcmf_dbg(TRACE, "enter: idx=%d\n", ifp->bssidx);
 	if (!entry)
 		return;
 
 	brcmf_fws_lock(ifp->drvr, flags);
 	ifp->fws_desc = NULL;
-	brcmf_fws_clear_mac_descriptor(entry);
+	brcmf_dbg(TRACE, "deleting %s\n", entry->name);
+	brcmf_fws_macdesc_deinit(entry);
 	brcmf_fws_cleanup(ifp->drvr->fws, ifp->ifidx);
 	brcmf_fws_unlock(ifp->drvr, flags);
 }
@@ -1904,39 +1837,37 @@ static void brcmf_fws_dequeue_worker(struct work_struct *worker)
 	struct sk_buff *skb;
 	ulong flags;
 	int fifo;
-	int credit;
 
 	fws = container_of(worker, struct brcmf_fws_info, fws_dequeue_work);
 
-	brcmf_dbg(TRACE, "enter: fws=%p\n", fws);
 	brcmf_fws_lock(fws->drvr, flags);
-	for (fifo = NL80211_NUM_ACS; fifo >= 0; fifo--) {
-		brcmf_dbg(TRACE, "fifo %d credit %d\n", fifo,
-			  fws->fifo_credit[fifo]);
-		for (credit = 0; credit < fws->fifo_credit[fifo]; /* nop */) {
+	for (fifo = BRCMF_FWS_FIFO_BCMC; fifo >= 0 && !fws->bus_flow_blocked;
+	     fifo--) {
+		while ((fws->fifo_credit[fifo]) || ((!fws->bcmc_credit_check) &&
+		       (fifo == BRCMF_FWS_FIFO_BCMC))) {
 			skb = brcmf_fws_deq(fws, fifo);
-			if (!skb || brcmf_fws_commit_skb(fws, fifo, skb))
+			if (!skb)
+				break;
+			fws->fifo_credit[fifo]--;
+			if (brcmf_fws_commit_skb(fws, fifo, skb))
+				break;
+			if (fws->bus_flow_blocked)
 				break;
-			if (brcmf_skbcb(skb)->if_flags &
-			    BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK)
-				credit++;
 		}
 		if ((fifo == BRCMF_FWS_FIFO_AC_BE) &&
-		    (credit == fws->fifo_credit[fifo])) {
-			fws->fifo_credit[fifo] -= credit;
+		    (fws->fifo_credit[fifo] == 0) &&
+		    (!fws->bus_flow_blocked)) {
 			while (brcmf_fws_borrow_credit(fws) == 0) {
 				skb = brcmf_fws_deq(fws, fifo);
 				if (!skb) {
 					brcmf_fws_return_credits(fws, fifo, 1);
 					break;
 				}
-				if (brcmf_fws_commit_skb(fws, fifo, skb)) {
-					brcmf_fws_return_credits(fws, fifo, 1);
+				if (brcmf_fws_commit_skb(fws, fifo, skb))
+					break;
+				if (fws->bus_flow_blocked)
 					break;
-				}
 			}
-		} else {
-			fws->fifo_credit[fifo] -= credit;
 		}
 	}
 	brcmf_fws_unlock(fws->drvr, flags);
@@ -1982,6 +1913,13 @@ int brcmf_fws_init(struct brcmf_pub *drvr)
 		brcmf_err("register credit map handler failed\n");
 		goto fail;
 	}
+	rc = brcmf_fweh_register(drvr, BRCMF_E_BCMC_CREDIT_SUPPORT,
+				 brcmf_fws_notify_bcmc_credit_support);
+	if (rc < 0) {
+		brcmf_err("register bcmc credit handler failed\n");
+		brcmf_fweh_unregister(drvr, BRCMF_E_FIFO_CREDIT_MAP);
+		goto fail;
+	}
 
 	/* setting the iovar may fail if feature is unsupported
 	 * so leave the rc as is so driver initialization can
@@ -1993,19 +1931,20 @@ int brcmf_fws_init(struct brcmf_pub *drvr)
 	}
 
 	brcmf_fws_hanger_init(&drvr->fws->hanger);
-	brcmf_fws_init_mac_descriptor(&drvr->fws->desc.other, NULL, 0);
+	brcmf_fws_macdesc_init(&drvr->fws->desc.other, NULL, 0);
+	brcmf_fws_macdesc_set_name(drvr->fws, &drvr->fws->desc.other);
 	brcmu_pktq_init(&drvr->fws->desc.other.psq, BRCMF_FWS_PSQ_PREC_COUNT,
 			BRCMF_FWS_PSQ_LEN);
 
 	/* create debugfs file for statistics */
 	brcmf_debugfs_create_fws_stats(drvr, &drvr->fws->stats);
 
-	/* TODO: remove upon feature delivery */
-	brcmf_err("%s bdcv2 tlv signaling [%x]\n",
+	brcmf_dbg(INFO, "%s bdcv2 tlv signaling [%x]\n",
 		  drvr->fw_signals ? "enabled" : "disabled", tlv);
 	return 0;
 
 fail_event:
+	brcmf_fweh_unregister(drvr, BRCMF_E_BCMC_CREDIT_SUPPORT);
 	brcmf_fweh_unregister(drvr, BRCMF_E_FIFO_CREDIT_MAP);
 fail:
 	brcmf_fws_deinit(drvr);
@@ -2043,25 +1982,31 @@ bool brcmf_fws_fc_active(struct brcmf_fws_info *fws)
 	if (!fws)
 		return false;
 
-	brcmf_dbg(TRACE, "enter: mode=%d\n", fws->fcmode);
 	return fws->fcmode != BRCMF_FWS_FCMODE_NONE;
 }
 
 void brcmf_fws_bustxfail(struct brcmf_fws_info *fws, struct sk_buff *skb)
 {
 	ulong flags;
+	u32 hslot;
 
-	brcmf_fws_lock(fws->drvr, flags);
-	brcmf_fws_txstatus_process(fws, BRCMF_FWS_TXSTATUS_FW_TOSSED,
-				   brcmf_skb_htod_tag_get_field(skb, HSLOT), 0);
-	/* the packet never reached firmware so reclaim credit */
-	if (fws->fcmode == BRCMF_FWS_FCMODE_EXPLICIT_CREDIT &&
-	    brcmf_skbcb(skb)->if_flags & BRCMF_SKB_IF_FLAGS_CREDITCHECK_MASK) {
-		brcmf_fws_return_credits(fws,
-					 brcmf_skb_htod_tag_get_field(skb,
-								      FIFO),
-					 1);
-		brcmf_fws_schedule_deq(fws);
+	if (brcmf_skbcb(skb)->state == BRCMF_FWS_SKBSTATE_TIM) {
+		brcmu_pkt_buf_free_skb(skb);
+		return;
 	}
+	brcmf_fws_lock(fws->drvr, flags);
+	hslot = brcmf_skb_htod_tag_get_field(skb, HSLOT);
+	brcmf_fws_txs_process(fws, BRCMF_FWS_TXSTATUS_HOST_TOSSED, hslot, 0);
 	brcmf_fws_unlock(fws->drvr, flags);
 }
+
+void brcmf_fws_bus_blocked(struct brcmf_pub *drvr, bool flow_blocked)
+{
+	struct brcmf_fws_info *fws = drvr->fws;
+
+	fws->bus_flow_blocked = flow_blocked;
+	if (!flow_blocked)
+		brcmf_fws_schedule_deq(fws);
+	else
+		fws->stats.bus_flow_block++;
+}
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.h b/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.h
index fbe483d..9fc8609 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.h
+++ b/drivers/net/wireless/brcm80211/brcmfmac/fwsignal.h
@@ -29,5 +29,6 @@ void brcmf_fws_reset_interface(struct brcmf_if *ifp);
 void brcmf_fws_add_interface(struct brcmf_if *ifp);
 void brcmf_fws_del_interface(struct brcmf_if *ifp);
 void brcmf_fws_bustxfail(struct brcmf_fws_info *fws, struct sk_buff *skb);
+void brcmf_fws_bus_blocked(struct brcmf_pub *drvr, bool flow_blocked);
 
 #endif /* FWSIGNAL_H_ */
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/sdio_host.h b/drivers/net/wireless/brcm80211/brcmfmac/sdio_host.h
index 7c1b633..09786a5 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/sdio_host.h
+++ b/drivers/net/wireless/brcm80211/brcmfmac/sdio_host.h
@@ -170,7 +170,6 @@ struct brcmf_sdio_dev {
 	atomic_t suspend;		/* suspend flag */
 	wait_queue_head_t request_byte_wait;
 	wait_queue_head_t request_word_wait;
-	wait_queue_head_t request_chain_wait;
 	wait_queue_head_t request_buffer_wait;
 	struct device *dev;
 	struct brcmf_bus *bus_if;
@@ -230,8 +229,6 @@ brcmf_sdcard_recv_chain(struct brcmf_sdio_dev *sdiodev, u32 addr, uint fn,
 #define SDIO_REQ_4BYTE	0x1
 /* Fixed address (FIFO) (vs. incrementing address) */
 #define SDIO_REQ_FIXED	0x2
-/* Async request (vs. sync request) */
-#define SDIO_REQ_ASYNC	0x4
 
 /* Read/write to memory block (F1, no FIFO) via CMD53 (sync only).
  *   rw:       read or write (0/1)
@@ -252,9 +249,6 @@ extern int brcmf_sdcard_abort(struct brcmf_sdio_dev *sdiodev, uint fn);
 extern int brcmf_sdio_probe(struct brcmf_sdio_dev *sdiodev);
 extern int brcmf_sdio_remove(struct brcmf_sdio_dev *sdiodev);
 
-extern int brcmf_sdcard_set_sbaddr_window(struct brcmf_sdio_dev *sdiodev,
-					  u32 address);
-
 /* attach, return handler on success, NULL if failed.
  *  The handler shall be provided by all subsequent calls. No local cache
  *  cfghdl points to the starting address of pci device mapped memory
@@ -272,16 +266,6 @@ brcmf_sdioh_request_word(struct brcmf_sdio_dev *sdiodev,
 			 uint rw, uint fnc, uint addr,
 			 u32 *word, uint nbyte);
 
-/* read or write any buffer using cmd53 */
-extern int
-brcmf_sdioh_request_buffer(struct brcmf_sdio_dev *sdiodev,
-			   uint fix_inc, uint rw, uint fnc_num, u32 addr,
-			   struct sk_buff *pkt);
-extern int
-brcmf_sdioh_request_chain(struct brcmf_sdio_dev *sdiodev, uint fix_inc,
-			  uint write, uint func, uint addr,
-			  struct sk_buff_head *pktq);
-
 /* Watchdog timer interface for pm ops */
 extern void brcmf_sdio_wdtmr_enable(struct brcmf_sdio_dev *sdiodev,
 				    bool enable);
@@ -291,4 +275,8 @@ extern void brcmf_sdbrcm_disconnect(void *ptr);
 extern void brcmf_sdbrcm_isr(void *arg);
 
 extern void brcmf_sdbrcm_wd_timer(struct brcmf_sdio *bus, uint wdtick);
+
+extern void brcmf_pm_resume_wait(struct brcmf_sdio_dev *sdiodev,
+				 wait_queue_head_t *wq);
+extern bool brcmf_pm_resume_error(struct brcmf_sdio_dev *sdiodev);
 #endif				/* _BRCM_SDH_H_ */
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/tracepoint.h b/drivers/net/wireless/brcm80211/brcmfmac/tracepoint.h
index 9df1f7a..bc29171 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/tracepoint.h
+++ b/drivers/net/wireless/brcm80211/brcmfmac/tracepoint.h
@@ -87,6 +87,27 @@ TRACE_EVENT(brcmf_hexdump,
 	TP_printk("hexdump [length=%lu]", __entry->len)
 );
 
+TRACE_EVENT(brcmf_bdchdr,
+	TP_PROTO(void *data),
+	TP_ARGS(data),
+	TP_STRUCT__entry(
+		__field(u8, flags)
+		__field(u8, prio)
+		__field(u8, flags2)
+		__field(u32, siglen)
+		__dynamic_array(u8, signal, *((u8 *)data + 3) * 4)
+	),
+	TP_fast_assign(
+		__entry->flags = *(u8 *)data;
+		__entry->prio = *((u8 *)data + 1);
+		__entry->flags2 = *((u8 *)data + 2);
+		__entry->siglen = *((u8 *)data + 3) * 4;
+		memcpy(__get_dynamic_array(signal),
+		       (u8 *)data + 4, __entry->siglen);
+	),
+	TP_printk("bdc: prio=%d siglen=%d", __entry->prio, __entry->siglen)
+);
+
 #ifdef CONFIG_BRCM_TRACING
 
 #undef TRACE_INCLUDE_PATH
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/usb.c b/drivers/net/wireless/brcm80211/brcmfmac/usb.c
index 01aed7a..322cadc 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/usb.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/usb.c
@@ -82,6 +82,7 @@ struct brcmf_usbdev_info {
 	int tx_high_watermark;
 	int tx_freecount;
 	bool tx_flowblock;
+	spinlock_t tx_flowblock_lock;
 
 	struct brcmf_usbreq *tx_reqs;
 	struct brcmf_usbreq *rx_reqs;
@@ -411,6 +412,7 @@ static void brcmf_usb_tx_complete(struct urb *urb)
 {
 	struct brcmf_usbreq *req = (struct brcmf_usbreq *)urb->context;
 	struct brcmf_usbdev_info *devinfo = req->devinfo;
+	unsigned long flags;
 
 	brcmf_dbg(USB, "Enter, urb->status=%d, skb=%p\n", urb->status,
 		  req->skb);
@@ -419,11 +421,13 @@ static void brcmf_usb_tx_complete(struct urb *urb)
 	brcmf_txcomplete(devinfo->dev, req->skb, urb->status == 0);
 	req->skb = NULL;
 	brcmf_usb_enq(devinfo, &devinfo->tx_freeq, req, &devinfo->tx_freecount);
+	spin_lock_irqsave(&devinfo->tx_flowblock_lock, flags);
 	if (devinfo->tx_freecount > devinfo->tx_high_watermark &&
 		devinfo->tx_flowblock) {
 		brcmf_txflowblock(devinfo->dev, false);
 		devinfo->tx_flowblock = false;
 	}
+	spin_unlock_irqrestore(&devinfo->tx_flowblock_lock, flags);
 }
 
 static void brcmf_usb_rx_complete(struct urb *urb)
@@ -568,6 +572,7 @@ static int brcmf_usb_tx(struct device *dev, struct sk_buff *skb)
 	struct brcmf_usbdev_info *devinfo = brcmf_usb_get_businfo(dev);
 	struct brcmf_usbreq  *req;
 	int ret;
+	unsigned long flags;
 
 	brcmf_dbg(USB, "Enter, skb=%p\n", skb);
 	if (devinfo->bus_pub.state != BRCMFMAC_USB_STATE_UP) {
@@ -599,11 +604,13 @@ static int brcmf_usb_tx(struct device *dev, struct sk_buff *skb)
 		goto fail;
 	}
 
+	spin_lock_irqsave(&devinfo->tx_flowblock_lock, flags);
 	if (devinfo->tx_freecount < devinfo->tx_low_watermark &&
 	    !devinfo->tx_flowblock) {
 		brcmf_txflowblock(dev, true);
 		devinfo->tx_flowblock = true;
 	}
+	spin_unlock_irqrestore(&devinfo->tx_flowblock_lock, flags);
 	return 0;
 
 fail:
@@ -1164,6 +1171,7 @@ struct brcmf_usbdev *brcmf_usb_attach(struct brcmf_usbdev_info *devinfo,
 
 	/* Initialize the spinlocks */
 	spin_lock_init(&devinfo->qlock);
+	spin_lock_init(&devinfo->tx_flowblock_lock);
 
 	INIT_LIST_HEAD(&devinfo->rx_freeq);
 	INIT_LIST_HEAD(&devinfo->rx_postq);
diff --git a/drivers/net/wireless/brcm80211/brcmfmac/wl_cfg80211.c b/drivers/net/wireless/brcm80211/brcmfmac/wl_cfg80211.c
index 301e572..277b37a 100644
--- a/drivers/net/wireless/brcm80211/brcmfmac/wl_cfg80211.c
+++ b/drivers/net/wireless/brcm80211/brcmfmac/wl_cfg80211.c
@@ -3982,6 +3982,7 @@ brcmf_cfg80211_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 	struct brcmf_fil_af_params_le *af_params;
 	bool ack;
 	s32 chan_nr;
+	u32 freq;
 
 	brcmf_dbg(TRACE, "Enter\n");
 
@@ -3994,6 +3995,8 @@ brcmf_cfg80211_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 		return -EPERM;
 	}
 
+	vif = container_of(wdev, struct brcmf_cfg80211_vif, wdev);
+
 	if (ieee80211_is_probe_resp(mgmt->frame_control)) {
 		/* Right now the only reason to get a probe response */
 		/* is for p2p listen response or for p2p GO from     */
@@ -4009,7 +4012,6 @@ brcmf_cfg80211_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 		ie_offset =  DOT11_MGMT_HDR_LEN +
 			     DOT11_BCN_PRB_FIXED_LEN;
 		ie_len = len - ie_offset;
-		vif = container_of(wdev, struct brcmf_cfg80211_vif, wdev);
 		if (vif == cfg->p2p.bss_idx[P2PAPI_BSSCFG_PRIMARY].vif)
 			vif = cfg->p2p.bss_idx[P2PAPI_BSSCFG_DEVICE].vif;
 		err = brcmf_vif_set_mgmt_ie(vif,
@@ -4033,16 +4035,22 @@ brcmf_cfg80211_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 		memcpy(&af_params->bssid[0], &mgmt->bssid[0], ETH_ALEN);
 		/* Add the length exepted for 802.11 header  */
 		action_frame->len = cpu_to_le16(len - DOT11_MGMT_HDR_LEN);
-		/* Add the channel */
-		chan_nr = ieee80211_frequency_to_channel(chan->center_freq);
+		/* Add the channel. Use the one specified as parameter if any or
+		 * the current one (got from the firmware) otherwise
+		 */
+		if (chan)
+			freq = chan->center_freq;
+		else
+			brcmf_fil_cmd_int_get(vif->ifp, BRCMF_C_GET_CHANNEL,
+					      &freq);
+		chan_nr = ieee80211_frequency_to_channel(freq);
 		af_params->channel = cpu_to_le32(chan_nr);
 
 		memcpy(action_frame->data, &buf[DOT11_MGMT_HDR_LEN],
 		       le16_to_cpu(action_frame->len));
 
 		brcmf_dbg(TRACE, "Action frame, cookie=%lld, len=%d, freq=%d\n",
-			  *cookie, le16_to_cpu(action_frame->len),
-			  chan->center_freq);
+			  *cookie, le16_to_cpu(action_frame->len), freq);
 
 		ack = brcmf_p2p_send_action_frame(cfg, cfg_to_ndev(cfg),
 						  af_params);
diff --git a/drivers/net/wireless/brcm80211/brcmsmac/ampdu.c b/drivers/net/wireless/brcm80211/brcmsmac/ampdu.c
index 1585cc5..bd98285 100644
--- a/drivers/net/wireless/brcm80211/brcmsmac/ampdu.c
+++ b/drivers/net/wireless/brcm80211/brcmsmac/ampdu.c
@@ -900,7 +900,7 @@ brcms_c_ampdu_dotxstatus_complete(struct ampdu_info *ampdu, struct scb *scb,
 		if (supr_status) {
 			update_rate = false;
 			if (supr_status == TX_STATUS_SUPR_BADCH) {
-				brcms_err(wlc->hw->d11core,
+				brcms_dbg_ht(wlc->hw->d11core,
 					  "%s: Pkt tx suppressed, illegal channel possibly %d\n",
 					  __func__, CHSPEC_CHANNEL(
 					  wlc->default_bss->chanspec));
diff --git a/drivers/net/wireless/cw1200/Kconfig b/drivers/net/wireless/cw1200/Kconfig
new file mode 100644
index 0000000..0880742
--- /dev/null
+++ b/drivers/net/wireless/cw1200/Kconfig
@@ -0,0 +1,30 @@
+config CW1200
+	tristate "CW1200 WLAN support"
+	depends on MAC80211 && CFG80211
+	help
+	  This is a driver for the ST-E CW1100 & CW1200 WLAN chipsets.
+	  This option just enables the driver core, see below for
+	  specific bus support.
+
+if CW1200
+
+config CW1200_WLAN_SDIO
+	tristate "Support SDIO platforms"
+	depends on CW1200 && MMC
+	help
+	  Enable support for the CW1200 connected via an SDIO bus.
+	  By default this driver only supports the Sagrad SG901-1091/1098 EVK
+	  and similar designs that utilize a hardware reset circuit. To
+	  support different CW1200 SDIO designs you will need to override
+	  the default platform data by calling cw1200_sdio_set_platform_data()
+	  in your board setup file.
+
+config CW1200_WLAN_SPI
+	tristate "Support SPI platforms"
+	depends on CW1200 && SPI
+	help
+	  Enables support for the CW1200 connected via a SPI bus.  You will
+	  need to add appropriate platform data glue in your board setup
+	  file.
+
+endif
diff --git a/drivers/net/wireless/cw1200/Makefile b/drivers/net/wireless/cw1200/Makefile
new file mode 100644
index 0000000..b086aac
--- /dev/null
+++ b/drivers/net/wireless/cw1200/Makefile
@@ -0,0 +1,21 @@
+cw1200_core-y := \
+		fwio.o \
+		txrx.o \
+		main.o \
+		queue.o \
+		hwio.o \
+		bh.o \
+		wsm.o \
+		sta.o \
+		scan.o \
+		debug.o
+cw1200_core-$(CONFIG_PM)	+= pm.o
+
+# CFLAGS_sta.o += -DDEBUG
+
+cw1200_wlan_sdio-y := cw1200_sdio.o
+cw1200_wlan_spi-y := cw1200_spi.o
+
+obj-$(CONFIG_CW1200) += cw1200_core.o
+obj-$(CONFIG_CW1200_WLAN_SDIO) += cw1200_wlan_sdio.o
+obj-$(CONFIG_CW1200_WLAN_SPI) += cw1200_wlan_spi.o
diff --git a/drivers/net/wireless/cw1200/bh.c b/drivers/net/wireless/cw1200/bh.c
new file mode 100644
index 0000000..c1ec2a4
--- /dev/null
+++ b/drivers/net/wireless/cw1200/bh.c
@@ -0,0 +1,619 @@
+/*
+ * Device handling thread implementation for mac80211 ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on:
+ * ST-Ericsson UMAC CW1200 driver, which is
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Ajitpal Singh <ajitpal.singh@stericsson.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <net/mac80211.h>
+#include <linux/kthread.h>
+#include <linux/timer.h>
+
+#include "cw1200.h"
+#include "bh.h"
+#include "hwio.h"
+#include "wsm.h"
+#include "hwbus.h"
+#include "debug.h"
+#include "fwio.h"
+
+static int cw1200_bh(void *arg);
+
+#define DOWNLOAD_BLOCK_SIZE_WR	(0x1000 - 4)
+/* an SPI message cannot be bigger than (2"12-1)*2 bytes
+ * "*2" to cvt to bytes
+ */
+#define MAX_SZ_RD_WR_BUFFERS	(DOWNLOAD_BLOCK_SIZE_WR*2)
+#define PIGGYBACK_CTRL_REG	(2)
+#define EFFECTIVE_BUF_SIZE	(MAX_SZ_RD_WR_BUFFERS - PIGGYBACK_CTRL_REG)
+
+/* Suspend state privates */
+enum cw1200_bh_pm_state {
+	CW1200_BH_RESUMED = 0,
+	CW1200_BH_SUSPEND,
+	CW1200_BH_SUSPENDED,
+	CW1200_BH_RESUME,
+};
+
+typedef int (*cw1200_wsm_handler)(struct cw1200_common *priv,
+	u8 *data, size_t size);
+
+static void cw1200_bh_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+	container_of(work, struct cw1200_common, bh_work);
+	cw1200_bh(priv);
+}
+
+int cw1200_register_bh(struct cw1200_common *priv)
+{
+	int err = 0;
+	/* Realtime workqueue */
+	priv->bh_workqueue = alloc_workqueue("cw1200_bh",
+				WQ_MEM_RECLAIM | WQ_HIGHPRI
+				| WQ_CPU_INTENSIVE, 1);
+
+	if (!priv->bh_workqueue)
+		return -ENOMEM;
+
+	INIT_WORK(&priv->bh_work, cw1200_bh_work);
+
+	pr_debug("[BH] register.\n");
+
+	atomic_set(&priv->bh_rx, 0);
+	atomic_set(&priv->bh_tx, 0);
+	atomic_set(&priv->bh_term, 0);
+	atomic_set(&priv->bh_suspend, CW1200_BH_RESUMED);
+	priv->bh_error = 0;
+	priv->hw_bufs_used = 0;
+	priv->buf_id_tx = 0;
+	priv->buf_id_rx = 0;
+	init_waitqueue_head(&priv->bh_wq);
+	init_waitqueue_head(&priv->bh_evt_wq);
+
+	err = !queue_work(priv->bh_workqueue, &priv->bh_work);
+	WARN_ON(err);
+	return err;
+}
+
+void cw1200_unregister_bh(struct cw1200_common *priv)
+{
+	atomic_add(1, &priv->bh_term);
+	wake_up(&priv->bh_wq);
+
+	flush_workqueue(priv->bh_workqueue);
+
+	destroy_workqueue(priv->bh_workqueue);
+	priv->bh_workqueue = NULL;
+
+	pr_debug("[BH] unregistered.\n");
+}
+
+void cw1200_irq_handler(struct cw1200_common *priv)
+{
+	pr_debug("[BH] irq.\n");
+
+	/* Disable Interrupts! */
+	/* NOTE:  hwbus_ops->lock already held */
+	__cw1200_irq_enable(priv, 0);
+
+	if (/* WARN_ON */(priv->bh_error))
+		return;
+
+	if (atomic_add_return(1, &priv->bh_rx) == 1)
+		wake_up(&priv->bh_wq);
+}
+EXPORT_SYMBOL_GPL(cw1200_irq_handler);
+
+void cw1200_bh_wakeup(struct cw1200_common *priv)
+{
+	pr_debug("[BH] wakeup.\n");
+	if (priv->bh_error) {
+		pr_err("[BH] wakeup failed (BH error)\n");
+		return;
+	}
+
+	if (atomic_add_return(1, &priv->bh_tx) == 1)
+		wake_up(&priv->bh_wq);
+}
+
+int cw1200_bh_suspend(struct cw1200_common *priv)
+{
+	pr_debug("[BH] suspend.\n");
+	if (priv->bh_error) {
+		wiphy_warn(priv->hw->wiphy, "BH error -- can't suspend\n");
+		return -EINVAL;
+	}
+
+	atomic_set(&priv->bh_suspend, CW1200_BH_SUSPEND);
+	wake_up(&priv->bh_wq);
+	return wait_event_timeout(priv->bh_evt_wq, priv->bh_error ||
+		(CW1200_BH_SUSPENDED == atomic_read(&priv->bh_suspend)),
+		 1 * HZ) ? 0 : -ETIMEDOUT;
+}
+
+int cw1200_bh_resume(struct cw1200_common *priv)
+{
+	pr_debug("[BH] resume.\n");
+	if (priv->bh_error) {
+		wiphy_warn(priv->hw->wiphy, "BH error -- can't resume\n");
+		return -EINVAL;
+	}
+
+	atomic_set(&priv->bh_suspend, CW1200_BH_RESUME);
+	wake_up(&priv->bh_wq);
+	return wait_event_timeout(priv->bh_evt_wq, priv->bh_error ||
+		(CW1200_BH_RESUMED == atomic_read(&priv->bh_suspend)),
+		1 * HZ) ? 0 : -ETIMEDOUT;
+}
+
+static inline void wsm_alloc_tx_buffer(struct cw1200_common *priv)
+{
+	++priv->hw_bufs_used;
+}
+
+int wsm_release_tx_buffer(struct cw1200_common *priv, int count)
+{
+	int ret = 0;
+	int hw_bufs_used = priv->hw_bufs_used;
+
+	priv->hw_bufs_used -= count;
+	if (WARN_ON(priv->hw_bufs_used < 0))
+		ret = -1;
+	else if (hw_bufs_used >= priv->wsm_caps.input_buffers)
+		ret = 1;
+	if (!priv->hw_bufs_used)
+		wake_up(&priv->bh_evt_wq);
+	return ret;
+}
+
+static int cw1200_bh_read_ctrl_reg(struct cw1200_common *priv,
+					  u16 *ctrl_reg)
+{
+	int ret;
+
+	ret = cw1200_reg_read_16(priv,
+			ST90TDS_CONTROL_REG_ID, ctrl_reg);
+	if (ret) {
+		ret = cw1200_reg_read_16(priv,
+				ST90TDS_CONTROL_REG_ID, ctrl_reg);
+		if (ret)
+			pr_err("[BH] Failed to read control register.\n");
+	}
+
+	return ret;
+}
+
+static int cw1200_device_wakeup(struct cw1200_common *priv)
+{
+	u16 ctrl_reg;
+	int ret;
+
+	pr_debug("[BH] Device wakeup.\n");
+
+	/* First, set the dpll register */
+	ret = cw1200_reg_write_32(priv, ST90TDS_TSET_GEN_R_W_REG_ID,
+				  cw1200_dpll_from_clk(priv->hw_refclk));
+	if (WARN_ON(ret))
+		return ret;
+
+	/* To force the device to be always-on, the host sets WLAN_UP to 1 */
+	ret = cw1200_reg_write_16(priv, ST90TDS_CONTROL_REG_ID,
+			ST90TDS_CONT_WUP_BIT);
+	if (WARN_ON(ret))
+		return ret;
+
+	ret = cw1200_bh_read_ctrl_reg(priv, &ctrl_reg);
+	if (WARN_ON(ret))
+		return ret;
+
+	/* If the device returns WLAN_RDY as 1, the device is active and will
+	 * remain active.
+	 */
+	if (ctrl_reg & ST90TDS_CONT_RDY_BIT) {
+		pr_debug("[BH] Device awake.\n");
+		return 1;
+	}
+
+	return 0;
+}
+
+/* Must be called from BH thraed. */
+void cw1200_enable_powersave(struct cw1200_common *priv,
+			     bool enable)
+{
+	pr_debug("[BH] Powerave is %s.\n",
+		 enable ? "enabled" : "disabled");
+	priv->powersave_enabled = enable;
+}
+
+static int cw1200_bh_rx_helper(struct cw1200_common *priv,
+			       uint16_t *ctrl_reg,
+			       int *tx)
+{
+	size_t read_len = 0;
+	struct sk_buff *skb_rx = NULL;
+	struct wsm_hdr *wsm;
+	size_t wsm_len;
+	u16 wsm_id;
+	u8 wsm_seq;
+	int rx_resync = 1;
+
+	size_t alloc_len;
+	u8 *data;
+
+	read_len = (*ctrl_reg & ST90TDS_CONT_NEXT_LEN_MASK) * 2;
+	if (!read_len)
+		return 0; /* No more work */
+
+	if (WARN_ON((read_len < sizeof(struct wsm_hdr)) ||
+		    (read_len > EFFECTIVE_BUF_SIZE))) {
+		pr_debug("Invalid read len: %zu (%04x)",
+			 read_len, *ctrl_reg);
+		goto err;
+	}
+
+	/* Add SIZE of PIGGYBACK reg (CONTROL Reg)
+	 * to the NEXT Message length + 2 Bytes for SKB
+	 */
+	read_len = read_len + 2;
+
+	alloc_len = priv->hwbus_ops->align_size(
+		priv->hwbus_priv, read_len);
+
+	/* Check if not exceeding CW1200 capabilities */
+	if (WARN_ON_ONCE(alloc_len > EFFECTIVE_BUF_SIZE)) {
+		pr_debug("Read aligned len: %zu\n",
+			 alloc_len);
+	}
+
+	skb_rx = dev_alloc_skb(alloc_len);
+	if (WARN_ON(!skb_rx))
+		goto err;
+
+	skb_trim(skb_rx, 0);
+	skb_put(skb_rx, read_len);
+	data = skb_rx->data;
+	if (WARN_ON(!data))
+		goto err;
+
+	if (WARN_ON(cw1200_data_read(priv, data, alloc_len))) {
+		pr_err("rx blew up, len %zu\n", alloc_len);
+		goto err;
+	}
+
+	/* Piggyback */
+	*ctrl_reg = __le16_to_cpu(
+		((__le16 *)data)[alloc_len / 2 - 1]);
+
+	wsm = (struct wsm_hdr *)data;
+	wsm_len = __le16_to_cpu(wsm->len);
+	if (WARN_ON(wsm_len > read_len))
+		goto err;
+
+	if (priv->wsm_enable_wsm_dumps)
+		print_hex_dump_bytes("<-- ",
+				     DUMP_PREFIX_NONE,
+				     data, wsm_len);
+
+	wsm_id  = __le16_to_cpu(wsm->id) & 0xFFF;
+	wsm_seq = (__le16_to_cpu(wsm->id) >> 13) & 7;
+
+	skb_trim(skb_rx, wsm_len);
+
+	if (wsm_id == 0x0800) {
+		wsm_handle_exception(priv,
+				     &data[sizeof(*wsm)],
+				     wsm_len - sizeof(*wsm));
+		goto err;
+	} else if (!rx_resync) {
+		if (WARN_ON(wsm_seq != priv->wsm_rx_seq))
+			goto err;
+	}
+	priv->wsm_rx_seq = (wsm_seq + 1) & 7;
+	rx_resync = 0;
+
+	if (wsm_id & 0x0400) {
+		int rc = wsm_release_tx_buffer(priv, 1);
+		if (WARN_ON(rc < 0))
+			return rc;
+		else if (rc > 0)
+			*tx = 1;
+	}
+
+	/* cw1200_wsm_rx takes care on SKB livetime */
+	if (WARN_ON(wsm_handle_rx(priv, wsm_id, wsm, &skb_rx)))
+		goto err;
+
+	if (skb_rx) {
+		dev_kfree_skb(skb_rx);
+		skb_rx = NULL;
+	}
+
+	return 0;
+
+err:
+	if (skb_rx) {
+		dev_kfree_skb(skb_rx);
+		skb_rx = NULL;
+	}
+	return -1;
+}
+
+static int cw1200_bh_tx_helper(struct cw1200_common *priv,
+			       int *pending_tx,
+			       int *tx_burst)
+{
+	size_t tx_len;
+	u8 *data;
+	int ret;
+	struct wsm_hdr *wsm;
+
+	if (priv->device_can_sleep) {
+		ret = cw1200_device_wakeup(priv);
+		if (WARN_ON(ret < 0)) { /* Error in wakeup */
+			*pending_tx = 1;
+			return 0;
+		} else if (ret) { /* Woke up */
+			priv->device_can_sleep = false;
+		} else { /* Did not awake */
+			*pending_tx = 1;
+			return 0;
+		}
+	}
+
+	wsm_alloc_tx_buffer(priv);
+	ret = wsm_get_tx(priv, &data, &tx_len, tx_burst);
+	if (ret <= 0) {
+		wsm_release_tx_buffer(priv, 1);
+		if (WARN_ON(ret < 0))
+			return ret; /* Error */
+		return 0; /* No work */
+	}
+
+	wsm = (struct wsm_hdr *)data;
+	BUG_ON(tx_len < sizeof(*wsm));
+	BUG_ON(__le16_to_cpu(wsm->len) != tx_len);
+
+	atomic_add(1, &priv->bh_tx);
+
+	tx_len = priv->hwbus_ops->align_size(
+		priv->hwbus_priv, tx_len);
+
+	/* Check if not exceeding CW1200 capabilities */
+	if (WARN_ON_ONCE(tx_len > EFFECTIVE_BUF_SIZE))
+		pr_debug("Write aligned len: %zu\n", tx_len);
+
+	wsm->id &= __cpu_to_le16(0xffff ^ WSM_TX_SEQ(WSM_TX_SEQ_MAX));
+	wsm->id |= __cpu_to_le16(WSM_TX_SEQ(priv->wsm_tx_seq));
+
+	if (WARN_ON(cw1200_data_write(priv, data, tx_len))) {
+		pr_err("tx blew up, len %zu\n", tx_len);
+		wsm_release_tx_buffer(priv, 1);
+		return -1; /* Error */
+	}
+
+	if (priv->wsm_enable_wsm_dumps)
+		print_hex_dump_bytes("--> ",
+				     DUMP_PREFIX_NONE,
+				     data,
+				     __le16_to_cpu(wsm->len));
+
+	wsm_txed(priv, data);
+	priv->wsm_tx_seq = (priv->wsm_tx_seq + 1) & WSM_TX_SEQ_MAX;
+
+	if (*tx_burst > 1) {
+		cw1200_debug_tx_burst(priv);
+		return 1; /* Work remains */
+	}
+
+	return 0;
+}
+
+static int cw1200_bh(void *arg)
+{
+	struct cw1200_common *priv = arg;
+	int rx, tx, term, suspend;
+	u16 ctrl_reg = 0;
+	int tx_allowed;
+	int pending_tx = 0;
+	int tx_burst;
+	long status;
+	u32 dummy;
+	int ret;
+
+	for (;;) {
+		if (!priv->hw_bufs_used &&
+		    priv->powersave_enabled &&
+		    !priv->device_can_sleep &&
+		    !atomic_read(&priv->recent_scan)) {
+			status = 1 * HZ;
+			pr_debug("[BH] Device wakedown. No data.\n");
+			cw1200_reg_write_16(priv, ST90TDS_CONTROL_REG_ID, 0);
+			priv->device_can_sleep = true;
+		} else if (priv->hw_bufs_used) {
+			/* Interrupt loss detection */
+			status = 1 * HZ;
+		} else {
+			status = MAX_SCHEDULE_TIMEOUT;
+		}
+
+		/* Dummy Read for SDIO retry mechanism*/
+		if ((priv->hw_type != -1) &&
+		    (atomic_read(&priv->bh_rx) == 0) &&
+		    (atomic_read(&priv->bh_tx) == 0))
+			cw1200_reg_read(priv, ST90TDS_CONFIG_REG_ID,
+					&dummy, sizeof(dummy));
+
+		pr_debug("[BH] waiting ...\n");
+		status = wait_event_interruptible_timeout(priv->bh_wq, ({
+				rx = atomic_xchg(&priv->bh_rx, 0);
+				tx = atomic_xchg(&priv->bh_tx, 0);
+				term = atomic_xchg(&priv->bh_term, 0);
+				suspend = pending_tx ?
+					0 : atomic_read(&priv->bh_suspend);
+				(rx || tx || term || suspend || priv->bh_error);
+			}), status);
+
+		pr_debug("[BH] - rx: %d, tx: %d, term: %d, suspend: %d, status: %ld\n",
+			 rx, tx, term, suspend, status);
+
+		/* Did an error occur? */
+		if ((status < 0 && status != -ERESTARTSYS) ||
+		    term || priv->bh_error) {
+			break;
+		}
+		if (!status) {  /* wait_event timed out */
+			unsigned long timestamp = jiffies;
+			long timeout;
+			int pending = 0;
+			int i;
+
+			/* Check to see if we have any outstanding frames */
+			if (priv->hw_bufs_used && (!rx || !tx)) {
+				wiphy_warn(priv->hw->wiphy,
+					   "Missed interrupt? (%d frames outstanding)\n",
+					   priv->hw_bufs_used);
+				rx = 1;
+
+				/* Get a timestamp of "oldest" frame */
+				for (i = 0; i < 4; ++i)
+					pending += cw1200_queue_get_xmit_timestamp(
+						&priv->tx_queue[i],
+						&timestamp,
+						priv->pending_frame_id);
+
+				/* Check if frame transmission is timed out.
+				 * Add an extra second with respect to possible
+				 * interrupt loss.
+				 */
+				timeout = timestamp +
+					WSM_CMD_LAST_CHANCE_TIMEOUT +
+					1 * HZ  -
+					jiffies;
+
+				/* And terminate BH thread if the frame is "stuck" */
+				if (pending && timeout < 0) {
+					wiphy_warn(priv->hw->wiphy,
+						   "Timeout waiting for TX confirm (%d/%d pending, %ld vs %lu).\n",
+						   priv->hw_bufs_used, pending,
+						   timestamp, jiffies);
+					break;
+				}
+			} else if (!priv->device_can_sleep &&
+				   !atomic_read(&priv->recent_scan)) {
+				pr_debug("[BH] Device wakedown. Timeout.\n");
+				cw1200_reg_write_16(priv,
+						    ST90TDS_CONTROL_REG_ID, 0);
+				priv->device_can_sleep = true;
+			}
+			goto done;
+		} else if (suspend) {
+			pr_debug("[BH] Device suspend.\n");
+			if (priv->powersave_enabled) {
+				pr_debug("[BH] Device wakedown. Suspend.\n");
+				cw1200_reg_write_16(priv,
+						    ST90TDS_CONTROL_REG_ID, 0);
+				priv->device_can_sleep = true;
+			}
+
+			atomic_set(&priv->bh_suspend, CW1200_BH_SUSPENDED);
+			wake_up(&priv->bh_evt_wq);
+			status = wait_event_interruptible(priv->bh_wq,
+							  CW1200_BH_RESUME == atomic_read(&priv->bh_suspend));
+			if (status < 0) {
+				wiphy_err(priv->hw->wiphy,
+					  "Failed to wait for resume: %ld.\n",
+					  status);
+				break;
+			}
+			pr_debug("[BH] Device resume.\n");
+			atomic_set(&priv->bh_suspend, CW1200_BH_RESUMED);
+			wake_up(&priv->bh_evt_wq);
+			atomic_add(1, &priv->bh_rx);
+			goto done;
+		}
+
+	rx:
+		tx += pending_tx;
+		pending_tx = 0;
+
+		if (cw1200_bh_read_ctrl_reg(priv, &ctrl_reg))
+			break;
+
+		/* Don't bother trying to rx unless we have data to read */
+		if (ctrl_reg & ST90TDS_CONT_NEXT_LEN_MASK) {
+			ret = cw1200_bh_rx_helper(priv, &ctrl_reg, &tx);
+			if (ret < 0)
+				break;
+			/* Double up here if there's more data.. */
+			if (ctrl_reg & ST90TDS_CONT_NEXT_LEN_MASK) {
+				ret = cw1200_bh_rx_helper(priv, &ctrl_reg, &tx);
+				if (ret < 0)
+					break;
+			}
+		}
+
+	tx:
+		if (tx) {
+			tx = 0;
+
+			BUG_ON(priv->hw_bufs_used > priv->wsm_caps.input_buffers);
+			tx_burst = priv->wsm_caps.input_buffers - priv->hw_bufs_used;
+			tx_allowed = tx_burst > 0;
+
+			if (!tx_allowed) {
+				/* Buffers full.  Ensure we process tx
+				 * after we handle rx..
+				 */
+				pending_tx = tx;
+				goto done_rx;
+			}
+			ret = cw1200_bh_tx_helper(priv, &pending_tx, &tx_burst);
+			if (ret < 0)
+				break;
+			if (ret > 0) /* More to transmit */
+				tx = ret;
+
+			/* Re-read ctrl reg */
+			if (cw1200_bh_read_ctrl_reg(priv, &ctrl_reg))
+				break;
+		}
+
+	done_rx:
+		if (priv->bh_error)
+			break;
+		if (ctrl_reg & ST90TDS_CONT_NEXT_LEN_MASK)
+			goto rx;
+		if (tx)
+			goto tx;
+
+	done:
+		/* Re-enable device interrupts */
+		priv->hwbus_ops->lock(priv->hwbus_priv);
+		__cw1200_irq_enable(priv, 1);
+		priv->hwbus_ops->unlock(priv->hwbus_priv);
+	}
+
+	/* Explicitly disable device interrupts */
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+	__cw1200_irq_enable(priv, 0);
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+
+	if (!term) {
+		pr_err("[BH] Fatal error, exiting.\n");
+		priv->bh_error = 1;
+		/* TODO: schedule_work(recovery) */
+	}
+	return 0;
+}
diff --git a/drivers/net/wireless/cw1200/bh.h b/drivers/net/wireless/cw1200/bh.h
new file mode 100644
index 0000000..af6a485
--- /dev/null
+++ b/drivers/net/wireless/cw1200/bh.h
@@ -0,0 +1,28 @@
+/*
+ * Device handling thread interface for mac80211 ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_BH_H
+#define CW1200_BH_H
+
+/* extern */ struct cw1200_common;
+
+int cw1200_register_bh(struct cw1200_common *priv);
+void cw1200_unregister_bh(struct cw1200_common *priv);
+void cw1200_irq_handler(struct cw1200_common *priv);
+void cw1200_bh_wakeup(struct cw1200_common *priv);
+int cw1200_bh_suspend(struct cw1200_common *priv);
+int cw1200_bh_resume(struct cw1200_common *priv);
+/* Must be called from BH thread. */
+void cw1200_enable_powersave(struct cw1200_common *priv,
+			     bool enable);
+int wsm_release_tx_buffer(struct cw1200_common *priv, int count);
+
+#endif /* CW1200_BH_H */
diff --git a/drivers/net/wireless/cw1200/cw1200.h b/drivers/net/wireless/cw1200/cw1200.h
new file mode 100644
index 0000000..1ad7d36
--- /dev/null
+++ b/drivers/net/wireless/cw1200/cw1200.h
@@ -0,0 +1,323 @@
+/*
+ * Common private data for ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on the mac80211 Prism54 code, which is
+ * Copyright (c) 2006, Michael Wu <flamingice@sourmilk.net>
+ *
+ * Based on the islsm (softmac prism54) driver, which is:
+ * Copyright 2004-2006 Jean-Baptiste Note <jbnote@gmail.com>, et al.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_H
+#define CW1200_H
+
+#include <linux/wait.h>
+#include <linux/mutex.h>
+#include <linux/workqueue.h>
+#include <net/mac80211.h>
+
+#include "queue.h"
+#include "wsm.h"
+#include "scan.h"
+#include "txrx.h"
+#include "pm.h"
+
+/* Forward declarations */
+struct hwbus_ops;
+struct task_struct;
+struct cw1200_debug_priv;
+struct firmware;
+
+#define CW1200_MAX_CTRL_FRAME_LEN	(0x1000)
+
+#define CW1200_MAX_STA_IN_AP_MODE	(5)
+#define CW1200_LINK_ID_AFTER_DTIM	(CW1200_MAX_STA_IN_AP_MODE + 1)
+#define CW1200_LINK_ID_UAPSD		(CW1200_MAX_STA_IN_AP_MODE + 2)
+#define CW1200_LINK_ID_MAX		(CW1200_MAX_STA_IN_AP_MODE + 3)
+#define CW1200_MAX_REQUEUE_ATTEMPTS	(5)
+
+#define CW1200_MAX_TID			(8)
+
+#define CW1200_BLOCK_ACK_CNT		(30)
+#define CW1200_BLOCK_ACK_THLD		(800)
+#define CW1200_BLOCK_ACK_HIST		(3)
+#define CW1200_BLOCK_ACK_INTERVAL	(1 * HZ / CW1200_BLOCK_ACK_HIST)
+
+#define CW1200_JOIN_TIMEOUT		(1 * HZ)
+#define CW1200_AUTH_TIMEOUT		(5 * HZ)
+
+struct cw1200_ht_info {
+	struct ieee80211_sta_ht_cap     ht_cap;
+	enum nl80211_channel_type       channel_type;
+	u16                             operation_mode;
+};
+
+/* Please keep order */
+enum cw1200_join_status {
+	CW1200_JOIN_STATUS_PASSIVE = 0,
+	CW1200_JOIN_STATUS_MONITOR,
+	CW1200_JOIN_STATUS_JOINING,
+	CW1200_JOIN_STATUS_PRE_STA,
+	CW1200_JOIN_STATUS_STA,
+	CW1200_JOIN_STATUS_IBSS,
+	CW1200_JOIN_STATUS_AP,
+};
+
+enum cw1200_link_status {
+	CW1200_LINK_OFF,
+	CW1200_LINK_RESERVE,
+	CW1200_LINK_SOFT,
+	CW1200_LINK_HARD,
+	CW1200_LINK_RESET,
+	CW1200_LINK_RESET_REMAP,
+};
+
+extern int cw1200_power_mode;
+extern const char * const cw1200_fw_types[];
+
+struct cw1200_link_entry {
+	unsigned long			timestamp;
+	enum cw1200_link_status		status;
+	enum cw1200_link_status		prev_status;
+	u8				mac[ETH_ALEN];
+	u8				buffered[CW1200_MAX_TID];
+	struct sk_buff_head		rx_queue;
+};
+
+struct cw1200_common {
+	/* interfaces to the rest of the stack */
+	struct ieee80211_hw		*hw;
+	struct ieee80211_vif		*vif;
+	struct device			*pdev;
+
+	/* Statistics */
+	struct ieee80211_low_level_stats stats;
+
+	/* Our macaddr */
+	u8 mac_addr[ETH_ALEN];
+
+	/* Hardware interface */
+	const struct hwbus_ops		*hwbus_ops;
+	struct hwbus_priv		*hwbus_priv;
+
+	/* Hardware information */
+	enum {
+		HIF_9000_SILICON_VERSATILE = 0,
+		HIF_8601_VERSATILE,
+		HIF_8601_SILICON,
+	} hw_type;
+	enum {
+		CW1200_HW_REV_CUT10 = 10,
+		CW1200_HW_REV_CUT11 = 11,
+		CW1200_HW_REV_CUT20 = 20,
+		CW1200_HW_REV_CUT22 = 22,
+		CW1X60_HW_REV       = 40,
+	} hw_revision;
+	int                             hw_refclk;
+	bool				hw_have_5ghz;
+	const struct firmware		*sdd;
+	char                            *sdd_path;
+
+	struct cw1200_debug_priv	*debug;
+
+	struct workqueue_struct		*workqueue;
+	struct mutex			conf_mutex;
+
+	struct cw1200_queue		tx_queue[4];
+	struct cw1200_queue_stats	tx_queue_stats;
+	int				tx_burst_idx;
+
+	/* firmware/hardware info */
+	unsigned int tx_hdr_len;
+
+	/* Radio data */
+	int output_power;
+
+	/* BBP/MAC state */
+	struct ieee80211_rate		*rates;
+	struct ieee80211_rate		*mcs_rates;
+	struct ieee80211_channel	*channel;
+	struct wsm_edca_params		edca;
+	struct wsm_tx_queue_params	tx_queue_params;
+	struct wsm_mib_association_mode	association_mode;
+	struct wsm_set_bss_params	bss_params;
+	struct cw1200_ht_info		ht_info;
+	struct wsm_set_pm		powersave_mode;
+	struct wsm_set_pm		firmware_ps_mode;
+	int				cqm_rssi_thold;
+	unsigned			cqm_rssi_hyst;
+	bool				cqm_use_rssi;
+	int				cqm_beacon_loss_count;
+	int				channel_switch_in_progress;
+	wait_queue_head_t		channel_switch_done;
+	u8				long_frame_max_tx_count;
+	u8				short_frame_max_tx_count;
+	int				mode;
+	bool				enable_beacon;
+	int				beacon_int;
+	bool				listening;
+	struct wsm_rx_filter		rx_filter;
+	struct wsm_mib_multicast_filter multicast_filter;
+	bool				has_multicast_subscription;
+	bool				disable_beacon_filter;
+	struct work_struct		update_filtering_work;
+	struct work_struct		set_beacon_wakeup_period_work;
+
+	u8				ba_rx_tid_mask;
+	u8				ba_tx_tid_mask;
+
+	struct cw1200_pm_state		pm_state;
+
+	struct wsm_p2p_ps_modeinfo	p2p_ps_modeinfo;
+	struct wsm_uapsd_info		uapsd_info;
+	bool				setbssparams_done;
+	bool				bt_present;
+	u8				conf_listen_interval;
+	u32				listen_interval;
+	u32				erp_info;
+	u32				rts_threshold;
+
+	/* BH */
+	atomic_t			bh_rx;
+	atomic_t			bh_tx;
+	atomic_t			bh_term;
+	atomic_t			bh_suspend;
+
+	struct workqueue_struct         *bh_workqueue;
+	struct work_struct              bh_work;
+
+	int				bh_error;
+	wait_queue_head_t		bh_wq;
+	wait_queue_head_t		bh_evt_wq;
+	u8				buf_id_tx;
+	u8				buf_id_rx;
+	u8				wsm_rx_seq;
+	u8				wsm_tx_seq;
+	int				hw_bufs_used;
+	bool				powersave_enabled;
+	bool				device_can_sleep;
+
+	/* Scan status */
+	struct cw1200_scan scan;
+	/* Keep cw1200 awake (WUP = 1) 1 second after each scan to avoid
+	 * FW issue with sleeping/waking up.
+	 */
+	atomic_t			recent_scan;
+	struct delayed_work		clear_recent_scan_work;
+
+	/* WSM */
+	struct wsm_startup_ind		wsm_caps;
+	struct mutex			wsm_cmd_mux;
+	struct wsm_buf			wsm_cmd_buf;
+	struct wsm_cmd			wsm_cmd;
+	wait_queue_head_t		wsm_cmd_wq;
+	wait_queue_head_t		wsm_startup_done;
+	int                             firmware_ready;
+	atomic_t			tx_lock;
+
+	/* WSM debug */
+	int				wsm_enable_wsm_dumps;
+
+	/* WSM Join */
+	enum cw1200_join_status	join_status;
+	u32			pending_frame_id;
+	bool			join_pending;
+	struct delayed_work	join_timeout;
+	struct work_struct	unjoin_work;
+	struct work_struct	join_complete_work;
+	int			join_complete_status;
+	int			join_dtim_period;
+	bool			delayed_unjoin;
+
+	/* TX/RX and security */
+	s8			wep_default_key_id;
+	struct work_struct	wep_key_work;
+	u32			key_map;
+	struct wsm_add_key	keys[WSM_KEY_MAX_INDEX + 1];
+
+	/* AP powersave */
+	u32			link_id_map;
+	struct cw1200_link_entry link_id_db[CW1200_MAX_STA_IN_AP_MODE];
+	struct work_struct	link_id_work;
+	struct delayed_work	link_id_gc_work;
+	u32			sta_asleep_mask;
+	u32			pspoll_mask;
+	bool			aid0_bit_set;
+	spinlock_t		ps_state_lock; /* Protect power save state */
+	bool			buffered_multicasts;
+	bool			tx_multicast;
+	struct work_struct	set_tim_work;
+	struct work_struct	set_cts_work;
+	struct work_struct	multicast_start_work;
+	struct work_struct	multicast_stop_work;
+	struct timer_list	mcast_timeout;
+
+	/* WSM events and CQM implementation */
+	spinlock_t		event_queue_lock; /* Protect event queue */
+	struct list_head	event_queue;
+	struct work_struct	event_handler;
+
+	struct delayed_work	bss_loss_work;
+	spinlock_t		bss_loss_lock; /* Protect BSS loss state */
+	int                     bss_loss_state;
+	u32                     bss_loss_confirm_id;
+	int			delayed_link_loss;
+	struct work_struct	bss_params_work;
+
+	/* TX rate policy cache */
+	struct tx_policy_cache tx_policy_cache;
+	struct work_struct tx_policy_upload_work;
+
+	/* legacy PS mode switch in suspend */
+	int			ps_mode_switch_in_progress;
+	wait_queue_head_t	ps_mode_switch_done;
+
+	/* Workaround for WFD testcase 6.1.10*/
+	struct work_struct	linkid_reset_work;
+	u8			action_frame_sa[ETH_ALEN];
+	u8			action_linkid;
+};
+
+struct cw1200_sta_priv {
+	int link_id;
+};
+
+/* interfaces for the drivers */
+int cw1200_core_probe(const struct hwbus_ops *hwbus_ops,
+		      struct hwbus_priv *hwbus,
+		      struct device *pdev,
+		      struct cw1200_common **pself,
+		      int ref_clk, const u8 *macaddr,
+		      const char *sdd_path, bool have_5ghz);
+void cw1200_core_release(struct cw1200_common *self);
+
+#define FWLOAD_BLOCK_SIZE (1024)
+
+static inline int cw1200_is_ht(const struct cw1200_ht_info *ht_info)
+{
+	return ht_info->channel_type != NL80211_CHAN_NO_HT;
+}
+
+static inline int cw1200_ht_greenfield(const struct cw1200_ht_info *ht_info)
+{
+	return cw1200_is_ht(ht_info) &&
+		(ht_info->ht_cap.cap & IEEE80211_HT_CAP_GRN_FLD) &&
+		!(ht_info->operation_mode &
+		  IEEE80211_HT_OP_MODE_NON_GF_STA_PRSNT);
+}
+
+static inline int cw1200_ht_ampdu_density(const struct cw1200_ht_info *ht_info)
+{
+	if (!cw1200_is_ht(ht_info))
+		return 0;
+	return ht_info->ht_cap.ampdu_density;
+}
+
+#endif /* CW1200_H */
diff --git a/drivers/net/wireless/cw1200/cw1200_sdio.c b/drivers/net/wireless/cw1200/cw1200_sdio.c
new file mode 100644
index 0000000..ebdcdf4
--- /dev/null
+++ b/drivers/net/wireless/cw1200/cw1200_sdio.c
@@ -0,0 +1,425 @@
+/*
+ * Mac80211 SDIO driver for ST-Ericsson CW1200 device
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/gpio.h>
+#include <linux/delay.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/sdio_func.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/sdio.h>
+#include <net/mac80211.h>
+
+#include "cw1200.h"
+#include "hwbus.h"
+#include <linux/platform_data/net-cw1200.h>
+#include "hwio.h"
+
+MODULE_AUTHOR("Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>");
+MODULE_DESCRIPTION("mac80211 ST-Ericsson CW1200 SDIO driver");
+MODULE_LICENSE("GPL");
+
+#define SDIO_BLOCK_SIZE (512)
+
+/* Default platform data for Sagrad modules */
+static struct cw1200_platform_data_sdio sagrad_109x_evk_platform_data = {
+	.ref_clk = 38400,
+	.have_5ghz = false,
+	.sdd_file = "sdd_sagrad_1091_1098.bin",
+};
+
+/* Allow platform data to be overridden */
+static struct cw1200_platform_data_sdio *global_plat_data = &sagrad_109x_evk_platform_data;
+
+void __init cw1200_sdio_set_platform_data(struct cw1200_platform_data_sdio *pdata)
+{
+	global_plat_data = pdata;
+}
+
+struct hwbus_priv {
+	struct sdio_func	*func;
+	struct cw1200_common	*core;
+	const struct cw1200_platform_data_sdio *pdata;
+};
+
+#ifndef SDIO_VENDOR_ID_STE
+#define SDIO_VENDOR_ID_STE		0x0020
+#endif
+
+#ifndef SDIO_DEVICE_ID_STE_CW1200
+#define SDIO_DEVICE_ID_STE_CW1200	0x2280
+#endif
+
+static const struct sdio_device_id cw1200_sdio_ids[] = {
+	{ SDIO_DEVICE(SDIO_VENDOR_ID_STE, SDIO_DEVICE_ID_STE_CW1200) },
+	{ /* end: all zeroes */			},
+};
+
+/* hwbus_ops implemetation */
+
+static int cw1200_sdio_memcpy_fromio(struct hwbus_priv *self,
+				     unsigned int addr,
+				     void *dst, int count)
+{
+	return sdio_memcpy_fromio(self->func, dst, addr, count);
+}
+
+static int cw1200_sdio_memcpy_toio(struct hwbus_priv *self,
+				   unsigned int addr,
+				   const void *src, int count)
+{
+	return sdio_memcpy_toio(self->func, addr, (void *)src, count);
+}
+
+static void cw1200_sdio_lock(struct hwbus_priv *self)
+{
+	sdio_claim_host(self->func);
+}
+
+static void cw1200_sdio_unlock(struct hwbus_priv *self)
+{
+	sdio_release_host(self->func);
+}
+
+static void cw1200_sdio_irq_handler(struct sdio_func *func)
+{
+	struct hwbus_priv *self = sdio_get_drvdata(func);
+
+	/* note:  sdio_host already claimed here. */
+	if (self->core)
+		cw1200_irq_handler(self->core);
+}
+
+static irqreturn_t cw1200_gpio_hardirq(int irq, void *dev_id)
+{
+	return IRQ_WAKE_THREAD;
+}
+
+static irqreturn_t cw1200_gpio_irq(int irq, void *dev_id)
+{
+	struct hwbus_priv *self = dev_id;
+
+	if (self->core) {
+		sdio_claim_host(self->func);
+		cw1200_irq_handler(self->core);
+		sdio_release_host(self->func);
+		return IRQ_HANDLED;
+	} else {
+		return IRQ_NONE;
+	}
+}
+
+static int cw1200_request_irq(struct hwbus_priv *self)
+{
+	int ret;
+	u8 cccr;
+
+	cccr = sdio_f0_readb(self->func, SDIO_CCCR_IENx, &ret);
+	if (WARN_ON(ret))
+		goto err;
+
+	/* Master interrupt enable ... */
+	cccr |= BIT(0);
+
+	/* ... for our function */
+	cccr |= BIT(self->func->num);
+
+	sdio_f0_writeb(self->func, cccr, SDIO_CCCR_IENx, &ret);
+	if (WARN_ON(ret))
+		goto err;
+
+	ret = enable_irq_wake(self->pdata->irq);
+	if (WARN_ON(ret))
+		goto err;
+
+	/* Request the IRQ */
+	ret =  request_threaded_irq(self->pdata->irq, cw1200_gpio_hardirq,
+				    cw1200_gpio_irq,
+				    IRQF_TRIGGER_HIGH | IRQF_ONESHOT,
+				    "cw1200_wlan_irq", self);
+	if (WARN_ON(ret))
+		goto err;
+
+	return 0;
+
+err:
+	return ret;
+}
+
+static int cw1200_sdio_irq_subscribe(struct hwbus_priv *self)
+{
+	int ret = 0;
+
+	pr_debug("SW IRQ subscribe\n");
+	sdio_claim_host(self->func);
+	if (self->pdata->irq)
+		ret = cw1200_request_irq(self);
+	else
+		ret = sdio_claim_irq(self->func, cw1200_sdio_irq_handler);
+
+	sdio_release_host(self->func);
+	return ret;
+}
+
+static int cw1200_sdio_irq_unsubscribe(struct hwbus_priv *self)
+{
+	int ret = 0;
+
+	pr_debug("SW IRQ unsubscribe\n");
+
+	if (self->pdata->irq) {
+		disable_irq_wake(self->pdata->irq);
+		free_irq(self->pdata->irq, self);
+	} else {
+		sdio_claim_host(self->func);
+		ret = sdio_release_irq(self->func);
+		sdio_release_host(self->func);
+	}
+	return ret;
+}
+
+static int cw1200_sdio_off(const struct cw1200_platform_data_sdio *pdata)
+{
+	if (pdata->reset) {
+		gpio_set_value(pdata->reset, 0);
+		msleep(30); /* Min is 2 * CLK32K cycles */
+		gpio_free(pdata->reset);
+	}
+
+	if (pdata->power_ctrl)
+		pdata->power_ctrl(pdata, false);
+	if (pdata->clk_ctrl)
+		pdata->clk_ctrl(pdata, false);
+
+	return 0;
+}
+
+static int cw1200_sdio_on(const struct cw1200_platform_data_sdio *pdata)
+{
+	/* Ensure I/Os are pulled low */
+	if (pdata->reset) {
+		gpio_request(pdata->reset, "cw1200_wlan_reset");
+		gpio_direction_output(pdata->reset, 0);
+	}
+	if (pdata->powerup) {
+		gpio_request(pdata->powerup, "cw1200_wlan_powerup");
+		gpio_direction_output(pdata->powerup, 0);
+	}
+	if (pdata->reset || pdata->powerup)
+		msleep(10); /* Settle time? */
+
+	/* Enable 3v3 and 1v8 to hardware */
+	if (pdata->power_ctrl) {
+		if (pdata->power_ctrl(pdata, true)) {
+			pr_err("power_ctrl() failed!\n");
+			return -1;
+		}
+	}
+
+	/* Enable CLK32K */
+	if (pdata->clk_ctrl) {
+		if (pdata->clk_ctrl(pdata, true)) {
+			pr_err("clk_ctrl() failed!\n");
+			return -1;
+		}
+		msleep(10); /* Delay until clock is stable for 2 cycles */
+	}
+
+	/* Enable POWERUP signal */
+	if (pdata->powerup) {
+		gpio_set_value(pdata->powerup, 1);
+		msleep(250); /* or more..? */
+	}
+	/* Enable RSTn signal */
+	if (pdata->reset) {
+		gpio_set_value(pdata->reset, 1);
+		msleep(50); /* Or more..? */
+	}
+	return 0;
+}
+
+static size_t cw1200_sdio_align_size(struct hwbus_priv *self, size_t size)
+{
+	if (self->pdata->no_nptb)
+		size = round_up(size, SDIO_BLOCK_SIZE);
+	else
+		size = sdio_align_size(self->func, size);
+
+	return size;
+}
+
+static int cw1200_sdio_pm(struct hwbus_priv *self, bool suspend)
+{
+	int ret = 0;
+
+	if (self->pdata->irq)
+		ret = irq_set_irq_wake(self->pdata->irq, suspend);
+	return ret;
+}
+
+static struct hwbus_ops cw1200_sdio_hwbus_ops = {
+	.hwbus_memcpy_fromio	= cw1200_sdio_memcpy_fromio,
+	.hwbus_memcpy_toio	= cw1200_sdio_memcpy_toio,
+	.lock			= cw1200_sdio_lock,
+	.unlock			= cw1200_sdio_unlock,
+	.align_size		= cw1200_sdio_align_size,
+	.power_mgmt		= cw1200_sdio_pm,
+};
+
+/* Probe Function to be called by SDIO stack when device is discovered */
+static int cw1200_sdio_probe(struct sdio_func *func,
+			     const struct sdio_device_id *id)
+{
+	struct hwbus_priv *self;
+	int status;
+
+	pr_info("cw1200_wlan_sdio: Probe called\n");
+
+	/* We are only able to handle the wlan function */
+	if (func->num != 0x01)
+		return -ENODEV;
+
+	self = kzalloc(sizeof(*self), GFP_KERNEL);
+	if (!self) {
+		pr_err("Can't allocate SDIO hwbus_priv.\n");
+		return -ENOMEM;
+	}
+
+	func->card->quirks |= MMC_QUIRK_LENIENT_FN0;
+
+	self->pdata = global_plat_data; /* FIXME */
+	self->func = func;
+	sdio_set_drvdata(func, self);
+	sdio_claim_host(func);
+	sdio_enable_func(func);
+	sdio_release_host(func);
+
+	status = cw1200_sdio_irq_subscribe(self);
+
+	status = cw1200_core_probe(&cw1200_sdio_hwbus_ops,
+				   self, &func->dev, &self->core,
+				   self->pdata->ref_clk,
+				   self->pdata->macaddr,
+				   self->pdata->sdd_file,
+				   self->pdata->have_5ghz);
+	if (status) {
+		cw1200_sdio_irq_unsubscribe(self);
+		sdio_claim_host(func);
+		sdio_disable_func(func);
+		sdio_release_host(func);
+		sdio_set_drvdata(func, NULL);
+		kfree(self);
+	}
+
+	return status;
+}
+
+/* Disconnect Function to be called by SDIO stack when
+ * device is disconnected
+ */
+static void cw1200_sdio_disconnect(struct sdio_func *func)
+{
+	struct hwbus_priv *self = sdio_get_drvdata(func);
+
+	if (self) {
+		cw1200_sdio_irq_unsubscribe(self);
+		if (self->core) {
+			cw1200_core_release(self->core);
+			self->core = NULL;
+		}
+		sdio_claim_host(func);
+		sdio_disable_func(func);
+		sdio_release_host(func);
+		sdio_set_drvdata(func, NULL);
+		kfree(self);
+	}
+}
+
+#ifdef CONFIG_PM
+static int cw1200_sdio_suspend(struct device *dev)
+{
+	int ret;
+	struct sdio_func *func = dev_to_sdio_func(dev);
+	struct hwbus_priv *self = sdio_get_drvdata(func);
+
+	if (!cw1200_can_suspend(self->core))
+		return -EAGAIN;
+
+	/* Notify SDIO that CW1200 will remain powered during suspend */
+	ret = sdio_set_host_pm_flags(func, MMC_PM_KEEP_POWER);
+	if (ret)
+		pr_err("Error setting SDIO pm flags: %i\n", ret);
+
+	return ret;
+}
+
+static int cw1200_sdio_resume(struct device *dev)
+{
+	return 0;
+}
+
+static const struct dev_pm_ops cw1200_pm_ops = {
+	.suspend = cw1200_sdio_suspend,
+	.resume = cw1200_sdio_resume,
+};
+#endif
+
+static struct sdio_driver sdio_driver = {
+	.name		= "cw1200_wlan_sdio",
+	.id_table	= cw1200_sdio_ids,
+	.probe		= cw1200_sdio_probe,
+	.remove		= cw1200_sdio_disconnect,
+#ifdef CONFIG_PM
+	.drv = {
+		.pm = &cw1200_pm_ops,
+	}
+#endif
+};
+
+/* Init Module function -> Called by insmod */
+static int __init cw1200_sdio_init(void)
+{
+	const struct cw1200_platform_data_sdio *pdata;
+	int ret;
+
+	/* FIXME -- this won't support multiple devices */
+	pdata = global_plat_data;
+
+	if (cw1200_sdio_on(pdata)) {
+		ret = -1;
+		goto err;
+	}
+
+	ret = sdio_register_driver(&sdio_driver);
+	if (ret)
+		goto err;
+
+	return 0;
+
+err:
+	cw1200_sdio_off(pdata);
+	return ret;
+}
+
+/* Called at Driver Unloading */
+static void __exit cw1200_sdio_exit(void)
+{
+	const struct cw1200_platform_data_sdio *pdata;
+
+	/* FIXME -- this won't support multiple devices */
+	pdata = global_plat_data;
+	sdio_unregister_driver(&sdio_driver);
+	cw1200_sdio_off(pdata);
+}
+
+
+module_init(cw1200_sdio_init);
+module_exit(cw1200_sdio_exit);
diff --git a/drivers/net/wireless/cw1200/cw1200_spi.c b/drivers/net/wireless/cw1200/cw1200_spi.c
new file mode 100644
index 0000000..d063760
--- /dev/null
+++ b/drivers/net/wireless/cw1200/cw1200_spi.c
@@ -0,0 +1,471 @@
+/*
+ * Mac80211 SPI driver for ST-Ericsson CW1200 device
+ *
+ * Copyright (c) 2011, Sagrad Inc.
+ * Author:  Solomon Peachy <speachy@sagrad.com>
+ *
+ * Based on cw1200_sdio.c
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/gpio.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <net/mac80211.h>
+
+#include <linux/spi/spi.h>
+#include <linux/device.h>
+
+#include "cw1200.h"
+#include "hwbus.h"
+#include <linux/platform_data/net-cw1200.h>
+#include "hwio.h"
+
+MODULE_AUTHOR("Solomon Peachy <speachy@sagrad.com>");
+MODULE_DESCRIPTION("mac80211 ST-Ericsson CW1200 SPI driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("spi:cw1200_wlan_spi");
+
+/* #define SPI_DEBUG */
+
+struct hwbus_priv {
+	struct spi_device	*func;
+	struct cw1200_common	*core;
+	const struct cw1200_platform_data_spi *pdata;
+	spinlock_t		lock; /* Serialize all bus operations */
+	int claimed;
+};
+
+#define SDIO_TO_SPI_ADDR(addr) ((addr & 0x1f)>>2)
+#define SET_WRITE 0x7FFF /* usage: and operation */
+#define SET_READ 0x8000  /* usage: or operation */
+
+/* Notes on byte ordering:
+   LE:  B0 B1 B2 B3
+   BE:  B3 B2 B1 B0
+
+   Hardware expects 32-bit data to be written as 16-bit BE words:
+
+   B1 B0 B3 B2
+*/
+
+static int cw1200_spi_memcpy_fromio(struct hwbus_priv *self,
+				     unsigned int addr,
+				     void *dst, int count)
+{
+	int ret, i;
+	u16 regaddr;
+	struct spi_message      m;
+
+	struct spi_transfer     t_addr = {
+		.tx_buf         = &regaddr,
+		.len            = sizeof(regaddr),
+	};
+	struct spi_transfer     t_msg = {
+		.rx_buf         = dst,
+		.len            = count,
+	};
+
+	regaddr = (SDIO_TO_SPI_ADDR(addr))<<12;
+	regaddr |= SET_READ;
+	regaddr |= (count>>1);
+
+#ifdef SPI_DEBUG
+	pr_info("READ : %04d from 0x%02x (%04x)\n", count, addr, regaddr);
+#endif
+
+	/* Header is LE16 */
+	regaddr = cpu_to_le16(regaddr);
+
+	/* We have to byteswap if the SPI bus is limited to 8b operation
+	   or we are running on a Big Endian system
+	*/
+#if defined(__LITTLE_ENDIAN)
+	if (self->func->bits_per_word == 8)
+#endif
+		regaddr = swab16(regaddr);
+
+	spi_message_init(&m);
+	spi_message_add_tail(&t_addr, &m);
+	spi_message_add_tail(&t_msg, &m);
+	ret = spi_sync(self->func, &m);
+
+#ifdef SPI_DEBUG
+	pr_info("READ : ");
+	for (i = 0; i < t_addr.len; i++)
+		printk("%02x ", ((u8 *)t_addr.tx_buf)[i]);
+	printk(" : ");
+	for (i = 0; i < t_msg.len; i++)
+		printk("%02x ", ((u8 *)t_msg.rx_buf)[i]);
+	printk("\n");
+#endif
+
+	/* We have to byteswap if the SPI bus is limited to 8b operation
+	   or we are running on a Big Endian system
+	*/
+#if defined(__LITTLE_ENDIAN)
+	if (self->func->bits_per_word == 8)
+#endif
+	{
+		uint16_t *buf = (uint16_t *)dst;
+		for (i = 0; i < ((count + 1) >> 1); i++)
+			buf[i] = swab16(buf[i]);
+	}
+
+	return ret;
+}
+
+static int cw1200_spi_memcpy_toio(struct hwbus_priv *self,
+				   unsigned int addr,
+				   const void *src, int count)
+{
+	int rval, i;
+	u16 regaddr;
+	struct spi_transfer     t_addr = {
+		.tx_buf         = &regaddr,
+		.len            = sizeof(regaddr),
+	};
+	struct spi_transfer     t_msg = {
+		.tx_buf         = src,
+		.len            = count,
+	};
+	struct spi_message      m;
+
+	regaddr = (SDIO_TO_SPI_ADDR(addr))<<12;
+	regaddr &= SET_WRITE;
+	regaddr |= (count>>1);
+
+#ifdef SPI_DEBUG
+	pr_info("WRITE: %04d  to  0x%02x (%04x)\n", count, addr, regaddr);
+#endif
+
+	/* Header is LE16 */
+	regaddr = cpu_to_le16(regaddr);
+
+	/* We have to byteswap if the SPI bus is limited to 8b operation
+	   or we are running on a Big Endian system
+	*/
+#if defined(__LITTLE_ENDIAN)
+	if (self->func->bits_per_word == 8)
+#endif
+	{
+		uint16_t *buf = (uint16_t *)src;
+	        regaddr = swab16(regaddr);
+		for (i = 0; i < ((count + 1) >> 1); i++)
+			buf[i] = swab16(buf[i]);
+	}
+
+#ifdef SPI_DEBUG
+	pr_info("WRITE: ");
+	for (i = 0; i < t_addr.len; i++)
+		printk("%02x ", ((u8 *)t_addr.tx_buf)[i]);
+	printk(" : ");
+	for (i = 0; i < t_msg.len; i++)
+		printk("%02x ", ((u8 *)t_msg.tx_buf)[i]);
+	printk("\n");
+#endif
+
+	spi_message_init(&m);
+	spi_message_add_tail(&t_addr, &m);
+	spi_message_add_tail(&t_msg, &m);
+	rval = spi_sync(self->func, &m);
+
+#ifdef SPI_DEBUG
+	pr_info("WROTE: %d\n", m.actual_length);
+#endif
+
+#if defined(__LITTLE_ENDIAN)
+	/* We have to byteswap if the SPI bus is limited to 8b operation */
+	if (self->func->bits_per_word == 8)
+#endif
+	{
+		uint16_t *buf = (uint16_t *)src;
+		for (i = 0; i < ((count + 1) >> 1); i++)
+			buf[i] = swab16(buf[i]);
+	}
+	return rval;
+}
+
+static void cw1200_spi_lock(struct hwbus_priv *self)
+{
+	unsigned long flags;
+
+	might_sleep();
+
+	spin_lock_irqsave(&self->lock, flags);
+	while (1) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		if (!self->claimed)
+			break;
+		spin_unlock_irqrestore(&self->lock, flags);
+		schedule();
+		spin_lock_irqsave(&self->lock, flags);
+	}
+	set_current_state(TASK_RUNNING);
+	self->claimed = 1;
+	spin_unlock_irqrestore(&self->lock, flags);
+
+	return;
+}
+
+static void cw1200_spi_unlock(struct hwbus_priv *self)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&self->lock, flags);
+	self->claimed = 0;
+	spin_unlock_irqrestore(&self->lock, flags);
+	return;
+}
+
+static irqreturn_t cw1200_spi_irq_handler(int irq, void *dev_id)
+{
+	struct hwbus_priv *self = dev_id;
+
+	if (self->core) {
+		cw1200_irq_handler(self->core);
+		return IRQ_HANDLED;
+	} else {
+		return IRQ_NONE;
+	}
+}
+
+static int cw1200_spi_irq_subscribe(struct hwbus_priv *self)
+{
+	int ret;
+
+	pr_debug("SW IRQ subscribe\n");
+
+	ret = request_any_context_irq(self->func->irq, cw1200_spi_irq_handler,
+				      IRQF_TRIGGER_HIGH,
+				      "cw1200_wlan_irq", self);
+	if (WARN_ON(ret < 0))
+		goto exit;
+
+	ret = enable_irq_wake(self->func->irq);
+	if (WARN_ON(ret))
+		goto free_irq;
+
+	return 0;
+
+free_irq:
+	free_irq(self->func->irq, self);
+exit:
+	return ret;
+}
+
+static int cw1200_spi_irq_unsubscribe(struct hwbus_priv *self)
+{
+	int ret = 0;
+
+	pr_debug("SW IRQ unsubscribe\n");
+	disable_irq_wake(self->func->irq);
+	free_irq(self->func->irq, self);
+
+	return ret;
+}
+
+static int cw1200_spi_off(const struct cw1200_platform_data_spi *pdata)
+{
+	if (pdata->reset) {
+		gpio_set_value(pdata->reset, 0);
+		msleep(30); /* Min is 2 * CLK32K cycles */
+		gpio_free(pdata->reset);
+	}
+
+	if (pdata->power_ctrl)
+		pdata->power_ctrl(pdata, false);
+	if (pdata->clk_ctrl)
+		pdata->clk_ctrl(pdata, false);
+
+	return 0;
+}
+
+static int cw1200_spi_on(const struct cw1200_platform_data_spi *pdata)
+{
+	/* Ensure I/Os are pulled low */
+	if (pdata->reset) {
+		gpio_request(pdata->reset, "cw1200_wlan_reset");
+		gpio_direction_output(pdata->reset, 0);
+	}
+	if (pdata->powerup) {
+		gpio_request(pdata->powerup, "cw1200_wlan_powerup");
+		gpio_direction_output(pdata->powerup, 0);
+	}
+	if (pdata->reset || pdata->powerup)
+		msleep(10); /* Settle time? */
+
+	/* Enable 3v3 and 1v8 to hardware */
+	if (pdata->power_ctrl) {
+		if (pdata->power_ctrl(pdata, true)) {
+			pr_err("power_ctrl() failed!\n");
+			return -1;
+		}
+	}
+
+	/* Enable CLK32K */
+	if (pdata->clk_ctrl) {
+		if (pdata->clk_ctrl(pdata, true)) {
+			pr_err("clk_ctrl() failed!\n");
+			return -1;
+		}
+		msleep(10); /* Delay until clock is stable for 2 cycles */
+	}
+
+	/* Enable POWERUP signal */
+	if (pdata->powerup) {
+		gpio_set_value(pdata->powerup, 1);
+		msleep(250); /* or more..? */
+	}
+	/* Enable RSTn signal */
+	if (pdata->reset) {
+		gpio_set_value(pdata->reset, 1);
+		msleep(50); /* Or more..? */
+	}
+	return 0;
+}
+
+static size_t cw1200_spi_align_size(struct hwbus_priv *self, size_t size)
+{
+	return size & 1 ? size + 1 : size;
+}
+
+static int cw1200_spi_pm(struct hwbus_priv *self, bool suspend)
+{
+	return irq_set_irq_wake(self->func->irq, suspend);
+}
+
+static struct hwbus_ops cw1200_spi_hwbus_ops = {
+	.hwbus_memcpy_fromio	= cw1200_spi_memcpy_fromio,
+	.hwbus_memcpy_toio	= cw1200_spi_memcpy_toio,
+	.lock			= cw1200_spi_lock,
+	.unlock			= cw1200_spi_unlock,
+	.align_size		= cw1200_spi_align_size,
+	.power_mgmt		= cw1200_spi_pm,
+};
+
+/* Probe Function to be called by SPI stack when device is discovered */
+static int cw1200_spi_probe(struct spi_device *func)
+{
+	const struct cw1200_platform_data_spi *plat_data =
+		func->dev.platform_data;
+	struct hwbus_priv *self;
+	int status;
+
+	/* Sanity check speed */
+	if (func->max_speed_hz > 52000000)
+		func->max_speed_hz = 52000000;
+	if (func->max_speed_hz < 1000000)
+		func->max_speed_hz = 1000000;
+
+	/* Fix up transfer size */
+	if (plat_data->spi_bits_per_word)
+		func->bits_per_word = plat_data->spi_bits_per_word;
+	if (!func->bits_per_word)
+		func->bits_per_word = 16;
+
+	/* And finally.. */
+	func->mode = SPI_MODE_0;
+
+	pr_info("cw1200_wlan_spi: Probe called (CS %d M %d BPW %d CLK %d)\n",
+		func->chip_select, func->mode, func->bits_per_word,
+		func->max_speed_hz);
+
+	if (cw1200_spi_on(plat_data)) {
+		pr_err("spi_on() failed!\n");
+		return -1;
+	}
+
+	if (spi_setup(func)) {
+		pr_err("spi_setup() failed!\n");
+		return -1;
+	}
+
+	self = kzalloc(sizeof(*self), GFP_KERNEL);
+	if (!self) {
+		pr_err("Can't allocate SPI hwbus_priv.");
+		return -ENOMEM;
+	}
+
+	self->pdata = plat_data;
+	self->func = func;
+	spin_lock_init(&self->lock);
+
+	spi_set_drvdata(func, self);
+
+	status = cw1200_spi_irq_subscribe(self);
+
+	status = cw1200_core_probe(&cw1200_spi_hwbus_ops,
+				   self, &func->dev, &self->core,
+				   self->pdata->ref_clk,
+				   self->pdata->macaddr,
+				   self->pdata->sdd_file,
+				   self->pdata->have_5ghz);
+
+	if (status) {
+		cw1200_spi_irq_unsubscribe(self);
+		cw1200_spi_off(plat_data);
+		kfree(self);
+	}
+
+	return status;
+}
+
+/* Disconnect Function to be called by SPI stack when device is disconnected */
+static int cw1200_spi_disconnect(struct spi_device *func)
+{
+	struct hwbus_priv *self = spi_get_drvdata(func);
+
+	if (self) {
+		cw1200_spi_irq_unsubscribe(self);
+		if (self->core) {
+			cw1200_core_release(self->core);
+			self->core = NULL;
+		}
+		kfree(self);
+	}
+	cw1200_spi_off(func->dev.platform_data);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int cw1200_spi_suspend(struct device *dev, pm_message_t state)
+{
+	struct hwbus_priv *self = spi_get_drvdata(to_spi_device(dev));
+
+	if (!cw1200_can_suspend(self->core))
+		return -EAGAIN;
+
+	/* XXX notify host that we have to keep CW1200 powered on? */
+	return 0;
+}
+
+static int cw1200_spi_resume(struct device *dev)
+{
+	return 0;
+}
+#endif
+
+static struct spi_driver spi_driver = {
+	.probe		= cw1200_spi_probe,
+	.remove		= cw1200_spi_disconnect,
+	.driver = {
+		.name		= "cw1200_wlan_spi",
+		.bus            = &spi_bus_type,
+		.owner          = THIS_MODULE,
+#ifdef CONFIG_PM
+		.suspend        = cw1200_spi_suspend,
+		.resume         = cw1200_spi_resume,
+#endif
+	},
+};
+
+module_spi_driver(spi_driver);
diff --git a/drivers/net/wireless/cw1200/debug.c b/drivers/net/wireless/cw1200/debug.c
new file mode 100644
index 0000000..e323b4d
--- /dev/null
+++ b/drivers/net/wireless/cw1200/debug.c
@@ -0,0 +1,428 @@
+/*
+ * mac80211 glue code for mac80211 ST-Ericsson CW1200 drivers
+ * DebugFS code
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include "cw1200.h"
+#include "debug.h"
+#include "fwio.h"
+
+/* join_status */
+static const char * const cw1200_debug_join_status[] = {
+	"passive",
+	"monitor",
+	"station (joining)",
+	"station (not authenticated yet)",
+	"station",
+	"adhoc",
+	"access point",
+};
+
+/* WSM_JOIN_PREAMBLE_... */
+static const char * const cw1200_debug_preamble[] = {
+	"long",
+	"short",
+	"long on 1 and 2 Mbps",
+};
+
+
+static const char * const cw1200_debug_link_id[] = {
+	"OFF",
+	"REQ",
+	"SOFT",
+	"HARD",
+};
+
+static const char *cw1200_debug_mode(int mode)
+{
+	switch (mode) {
+	case NL80211_IFTYPE_UNSPECIFIED:
+		return "unspecified";
+	case NL80211_IFTYPE_MONITOR:
+		return "monitor";
+	case NL80211_IFTYPE_STATION:
+		return "station";
+	case NL80211_IFTYPE_ADHOC:
+		return "adhoc";
+	case NL80211_IFTYPE_MESH_POINT:
+		return "mesh point";
+	case NL80211_IFTYPE_AP:
+		return "access point";
+	case NL80211_IFTYPE_P2P_CLIENT:
+		return "p2p client";
+	case NL80211_IFTYPE_P2P_GO:
+		return "p2p go";
+	default:
+		return "unsupported";
+	}
+}
+
+static void cw1200_queue_status_show(struct seq_file *seq,
+				     struct cw1200_queue *q)
+{
+	int i;
+	seq_printf(seq, "Queue       %d:\n", q->queue_id);
+	seq_printf(seq, "  capacity: %zu\n", q->capacity);
+	seq_printf(seq, "  queued:   %zu\n", q->num_queued);
+	seq_printf(seq, "  pending:  %zu\n", q->num_pending);
+	seq_printf(seq, "  sent:     %zu\n", q->num_sent);
+	seq_printf(seq, "  locked:   %s\n", q->tx_locked_cnt ? "yes" : "no");
+	seq_printf(seq, "  overfull: %s\n", q->overfull ? "yes" : "no");
+	seq_puts(seq,   "  link map: 0-> ");
+	for (i = 0; i < q->stats->map_capacity; ++i)
+		seq_printf(seq, "%.2d ", q->link_map_cache[i]);
+	seq_printf(seq, "<-%zu\n", q->stats->map_capacity);
+}
+
+static void cw1200_debug_print_map(struct seq_file *seq,
+				   struct cw1200_common *priv,
+				   const char *label,
+				   u32 map)
+{
+	int i;
+	seq_printf(seq, "%s0-> ", label);
+	for (i = 0; i < priv->tx_queue_stats.map_capacity; ++i)
+		seq_printf(seq, "%s ", (map & BIT(i)) ? "**" : "..");
+	seq_printf(seq, "<-%zu\n", priv->tx_queue_stats.map_capacity - 1);
+}
+
+static int cw1200_status_show(struct seq_file *seq, void *v)
+{
+	int i;
+	struct list_head *item;
+	struct cw1200_common *priv = seq->private;
+	struct cw1200_debug_priv *d = priv->debug;
+
+	seq_puts(seq,   "CW1200 Wireless LAN driver status\n");
+	seq_printf(seq, "Hardware:   %d.%d\n",
+		   priv->wsm_caps.hw_id,
+		   priv->wsm_caps.hw_subid);
+	seq_printf(seq, "Firmware:   %s %d.%d\n",
+		   cw1200_fw_types[priv->wsm_caps.fw_type],
+		   priv->wsm_caps.fw_ver,
+		   priv->wsm_caps.fw_build);
+	seq_printf(seq, "FW API:     %d\n",
+		   priv->wsm_caps.fw_api);
+	seq_printf(seq, "FW caps:    0x%.4X\n",
+		   priv->wsm_caps.fw_cap);
+	seq_printf(seq, "FW label:  '%s'\n",
+		   priv->wsm_caps.fw_label);
+	seq_printf(seq, "Mode:       %s%s\n",
+		   cw1200_debug_mode(priv->mode),
+		   priv->listening ? " (listening)" : "");
+	seq_printf(seq, "Join state: %s\n",
+		   cw1200_debug_join_status[priv->join_status]);
+	if (priv->channel)
+		seq_printf(seq, "Channel:    %d%s\n",
+			   priv->channel->hw_value,
+			   priv->channel_switch_in_progress ?
+			   " (switching)" : "");
+	if (priv->rx_filter.promiscuous)
+		seq_puts(seq,   "Filter:     promisc\n");
+	else if (priv->rx_filter.fcs)
+		seq_puts(seq,   "Filter:     fcs\n");
+	if (priv->rx_filter.bssid)
+		seq_puts(seq,   "Filter:     bssid\n");
+	if (!priv->disable_beacon_filter)
+		seq_puts(seq,   "Filter:     beacons\n");
+
+	if (priv->enable_beacon ||
+	    priv->mode == NL80211_IFTYPE_AP ||
+	    priv->mode == NL80211_IFTYPE_ADHOC ||
+	    priv->mode == NL80211_IFTYPE_MESH_POINT ||
+	    priv->mode == NL80211_IFTYPE_P2P_GO)
+		seq_printf(seq, "Beaconing:  %s\n",
+			   priv->enable_beacon ?
+			   "enabled" : "disabled");
+
+	for (i = 0; i < 4; ++i)
+		seq_printf(seq, "EDCA(%d):    %d, %d, %d, %d, %d\n", i,
+			   priv->edca.params[i].cwmin,
+			   priv->edca.params[i].cwmax,
+			   priv->edca.params[i].aifns,
+			   priv->edca.params[i].txop_limit,
+			   priv->edca.params[i].max_rx_lifetime);
+
+	if (priv->join_status == CW1200_JOIN_STATUS_STA) {
+		static const char *pm_mode = "unknown";
+		switch (priv->powersave_mode.mode) {
+		case WSM_PSM_ACTIVE:
+			pm_mode = "off";
+			break;
+		case WSM_PSM_PS:
+			pm_mode = "on";
+			break;
+		case WSM_PSM_FAST_PS:
+			pm_mode = "dynamic";
+			break;
+		}
+		seq_printf(seq, "Preamble:   %s\n",
+			   cw1200_debug_preamble[priv->association_mode.preamble]);
+		seq_printf(seq, "AMPDU spcn: %d\n",
+			   priv->association_mode.mpdu_start_spacing);
+		seq_printf(seq, "Basic rate: 0x%.8X\n",
+			   le32_to_cpu(priv->association_mode.basic_rate_set));
+		seq_printf(seq, "Bss lost:   %d beacons\n",
+			   priv->bss_params.beacon_lost_count);
+		seq_printf(seq, "AID:        %d\n",
+			   priv->bss_params.aid);
+		seq_printf(seq, "Rates:      0x%.8X\n",
+			   priv->bss_params.operational_rate_set);
+		seq_printf(seq, "Powersave:  %s\n", pm_mode);
+	}
+	seq_printf(seq, "HT:         %s\n",
+		   cw1200_is_ht(&priv->ht_info) ? "on" : "off");
+	if (cw1200_is_ht(&priv->ht_info)) {
+		seq_printf(seq, "Greenfield: %s\n",
+			   cw1200_ht_greenfield(&priv->ht_info) ? "yes" : "no");
+		seq_printf(seq, "AMPDU dens: %d\n",
+			   cw1200_ht_ampdu_density(&priv->ht_info));
+	}
+	seq_printf(seq, "RSSI thold: %d\n",
+		   priv->cqm_rssi_thold);
+	seq_printf(seq, "RSSI hyst:  %d\n",
+		   priv->cqm_rssi_hyst);
+	seq_printf(seq, "Long retr:  %d\n",
+		   priv->long_frame_max_tx_count);
+	seq_printf(seq, "Short retr: %d\n",
+		   priv->short_frame_max_tx_count);
+	spin_lock_bh(&priv->tx_policy_cache.lock);
+	i = 0;
+	list_for_each(item, &priv->tx_policy_cache.used)
+		++i;
+	spin_unlock_bh(&priv->tx_policy_cache.lock);
+	seq_printf(seq, "RC in use:  %d\n", i);
+
+	seq_puts(seq, "\n");
+	for (i = 0; i < 4; ++i) {
+		cw1200_queue_status_show(seq, &priv->tx_queue[i]);
+		seq_puts(seq, "\n");
+	}
+
+	cw1200_debug_print_map(seq, priv, "Link map:   ",
+			       priv->link_id_map);
+	cw1200_debug_print_map(seq, priv, "Asleep map: ",
+			       priv->sta_asleep_mask);
+	cw1200_debug_print_map(seq, priv, "PSPOLL map: ",
+			       priv->pspoll_mask);
+
+	seq_puts(seq, "\n");
+
+	for (i = 0; i < CW1200_MAX_STA_IN_AP_MODE; ++i) {
+		if (priv->link_id_db[i].status) {
+			seq_printf(seq, "Link %d:     %s, %pM\n",
+				   i + 1,
+				   cw1200_debug_link_id[priv->link_id_db[i].status],
+				   priv->link_id_db[i].mac);
+		}
+	}
+
+	seq_puts(seq, "\n");
+
+	seq_printf(seq, "BH status:  %s\n",
+		   atomic_read(&priv->bh_term) ? "terminated" : "alive");
+	seq_printf(seq, "Pending RX: %d\n",
+		   atomic_read(&priv->bh_rx));
+	seq_printf(seq, "Pending TX: %d\n",
+		   atomic_read(&priv->bh_tx));
+	if (priv->bh_error)
+		seq_printf(seq, "BH errcode: %d\n",
+			   priv->bh_error);
+	seq_printf(seq, "TX bufs:    %d x %d bytes\n",
+		   priv->wsm_caps.input_buffers,
+		   priv->wsm_caps.input_buffer_size);
+	seq_printf(seq, "Used bufs:  %d\n",
+		   priv->hw_bufs_used);
+	seq_printf(seq, "Powermgmt:  %s\n",
+		   priv->powersave_enabled ? "on" : "off");
+	seq_printf(seq, "Device:     %s\n",
+		   priv->device_can_sleep ? "asleep" : "awake");
+
+	spin_lock(&priv->wsm_cmd.lock);
+	seq_printf(seq, "WSM status: %s\n",
+		   priv->wsm_cmd.done ? "idle" : "active");
+	seq_printf(seq, "WSM cmd:    0x%.4X (%td bytes)\n",
+		   priv->wsm_cmd.cmd, priv->wsm_cmd.len);
+	seq_printf(seq, "WSM retval: %d\n",
+		   priv->wsm_cmd.ret);
+	spin_unlock(&priv->wsm_cmd.lock);
+
+	seq_printf(seq, "Datapath:   %s\n",
+		   atomic_read(&priv->tx_lock) ? "locked" : "unlocked");
+	if (atomic_read(&priv->tx_lock))
+		seq_printf(seq, "TXlock cnt: %d\n",
+			   atomic_read(&priv->tx_lock));
+
+	seq_printf(seq, "TXed:       %d\n",
+		   d->tx);
+	seq_printf(seq, "AGG TXed:   %d\n",
+		   d->tx_agg);
+	seq_printf(seq, "MULTI TXed: %d (%d)\n",
+		   d->tx_multi, d->tx_multi_frames);
+	seq_printf(seq, "RXed:       %d\n",
+		   d->rx);
+	seq_printf(seq, "AGG RXed:   %d\n",
+		   d->rx_agg);
+	seq_printf(seq, "TX miss:    %d\n",
+		   d->tx_cache_miss);
+	seq_printf(seq, "TX align:   %d\n",
+		   d->tx_align);
+	seq_printf(seq, "TX burst:   %d\n",
+		   d->tx_burst);
+	seq_printf(seq, "TX TTL:     %d\n",
+		   d->tx_ttl);
+	seq_printf(seq, "Scan:       %s\n",
+		   atomic_read(&priv->scan.in_progress) ? "active" : "idle");
+
+	return 0;
+}
+
+static int cw1200_status_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, &cw1200_status_show,
+		inode->i_private);
+}
+
+static const struct file_operations fops_status = {
+	.open = cw1200_status_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static int cw1200_counters_show(struct seq_file *seq, void *v)
+{
+	int ret;
+	struct cw1200_common *priv = seq->private;
+	struct wsm_mib_counters_table counters;
+
+	ret = wsm_get_counters_table(priv, &counters);
+	if (ret)
+		return ret;
+
+#define PUT_COUNTER(tab, name) \
+	seq_printf(seq, "%s:" tab "%d\n", #name, \
+		__le32_to_cpu(counters.name))
+
+	PUT_COUNTER("\t\t", plcp_errors);
+	PUT_COUNTER("\t\t", fcs_errors);
+	PUT_COUNTER("\t\t", tx_packets);
+	PUT_COUNTER("\t\t", rx_packets);
+	PUT_COUNTER("\t\t", rx_packet_errors);
+	PUT_COUNTER("\t",   rx_decryption_failures);
+	PUT_COUNTER("\t\t", rx_mic_failures);
+	PUT_COUNTER("\t",   rx_no_key_failures);
+	PUT_COUNTER("\t",   tx_multicast_frames);
+	PUT_COUNTER("\t",   tx_frames_success);
+	PUT_COUNTER("\t",   tx_frame_failures);
+	PUT_COUNTER("\t",   tx_frames_retried);
+	PUT_COUNTER("\t",   tx_frames_multi_retried);
+	PUT_COUNTER("\t",   rx_frame_duplicates);
+	PUT_COUNTER("\t\t", rts_success);
+	PUT_COUNTER("\t\t", rts_failures);
+	PUT_COUNTER("\t\t", ack_failures);
+	PUT_COUNTER("\t",   rx_multicast_frames);
+	PUT_COUNTER("\t",   rx_frames_success);
+	PUT_COUNTER("\t",   rx_cmac_icv_errors);
+	PUT_COUNTER("\t\t", rx_cmac_replays);
+	PUT_COUNTER("\t",   rx_mgmt_ccmp_replays);
+
+#undef PUT_COUNTER
+
+	return 0;
+}
+
+static int cw1200_counters_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, &cw1200_counters_show,
+		inode->i_private);
+}
+
+static const struct file_operations fops_counters = {
+	.open = cw1200_counters_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.owner = THIS_MODULE,
+};
+
+static ssize_t cw1200_wsm_dumps(struct file *file,
+	const char __user *user_buf, size_t count, loff_t *ppos)
+{
+	struct cw1200_common *priv = file->private_data;
+	char buf[1];
+
+	if (!count)
+		return -EINVAL;
+	if (copy_from_user(buf, user_buf, 1))
+		return -EFAULT;
+
+	if (buf[0] == '1')
+		priv->wsm_enable_wsm_dumps = 1;
+	else
+		priv->wsm_enable_wsm_dumps = 0;
+
+	return count;
+}
+
+static const struct file_operations fops_wsm_dumps = {
+	.open = simple_open,
+	.write = cw1200_wsm_dumps,
+	.llseek = default_llseek,
+};
+
+int cw1200_debug_init(struct cw1200_common *priv)
+{
+	int ret = -ENOMEM;
+	struct cw1200_debug_priv *d = kzalloc(sizeof(struct cw1200_debug_priv),
+			GFP_KERNEL);
+	priv->debug = d;
+	if (!d)
+		return ret;
+
+	d->debugfs_phy = debugfs_create_dir("cw1200",
+					    priv->hw->wiphy->debugfsdir);
+	if (!d->debugfs_phy)
+		goto err;
+
+	if (!debugfs_create_file("status", S_IRUSR, d->debugfs_phy,
+				 priv, &fops_status))
+		goto err;
+
+	if (!debugfs_create_file("counters", S_IRUSR, d->debugfs_phy,
+				 priv, &fops_counters))
+		goto err;
+
+	if (!debugfs_create_file("wsm_dumps", S_IWUSR, d->debugfs_phy,
+				 priv, &fops_wsm_dumps))
+		goto err;
+
+	return 0;
+
+err:
+	priv->debug = NULL;
+	debugfs_remove_recursive(d->debugfs_phy);
+	kfree(d);
+	return ret;
+}
+
+void cw1200_debug_release(struct cw1200_common *priv)
+{
+	struct cw1200_debug_priv *d = priv->debug;
+	if (d) {
+		debugfs_remove_recursive(d->debugfs_phy);
+		priv->debug = NULL;
+		kfree(d);
+	}
+}
diff --git a/drivers/net/wireless/cw1200/debug.h b/drivers/net/wireless/cw1200/debug.h
new file mode 100644
index 0000000..b525aba
--- /dev/null
+++ b/drivers/net/wireless/cw1200/debug.h
@@ -0,0 +1,93 @@
+/*
+ * DebugFS code for ST-Ericsson CW1200 mac80211 driver
+ *
+ * Copyright (c) 2011, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_DEBUG_H_INCLUDED
+#define CW1200_DEBUG_H_INCLUDED
+
+struct cw1200_debug_priv {
+	struct dentry *debugfs_phy;
+	int tx;
+	int tx_agg;
+	int rx;
+	int rx_agg;
+	int tx_multi;
+	int tx_multi_frames;
+	int tx_cache_miss;
+	int tx_align;
+	int tx_ttl;
+	int tx_burst;
+	int ba_cnt;
+	int ba_acc;
+	int ba_cnt_rx;
+	int ba_acc_rx;
+};
+
+int cw1200_debug_init(struct cw1200_common *priv);
+void cw1200_debug_release(struct cw1200_common *priv);
+
+static inline void cw1200_debug_txed(struct cw1200_common *priv)
+{
+	++priv->debug->tx;
+}
+
+static inline void cw1200_debug_txed_agg(struct cw1200_common *priv)
+{
+	++priv->debug->tx_agg;
+}
+
+static inline void cw1200_debug_txed_multi(struct cw1200_common *priv,
+					   int count)
+{
+	++priv->debug->tx_multi;
+	priv->debug->tx_multi_frames += count;
+}
+
+static inline void cw1200_debug_rxed(struct cw1200_common *priv)
+{
+	++priv->debug->rx;
+}
+
+static inline void cw1200_debug_rxed_agg(struct cw1200_common *priv)
+{
+	++priv->debug->rx_agg;
+}
+
+static inline void cw1200_debug_tx_cache_miss(struct cw1200_common *priv)
+{
+	++priv->debug->tx_cache_miss;
+}
+
+static inline void cw1200_debug_tx_align(struct cw1200_common *priv)
+{
+	++priv->debug->tx_align;
+}
+
+static inline void cw1200_debug_tx_ttl(struct cw1200_common *priv)
+{
+	++priv->debug->tx_ttl;
+}
+
+static inline void cw1200_debug_tx_burst(struct cw1200_common *priv)
+{
+	++priv->debug->tx_burst;
+}
+
+static inline void cw1200_debug_ba(struct cw1200_common *priv,
+				   int ba_cnt, int ba_acc,
+				   int ba_cnt_rx, int ba_acc_rx)
+{
+	priv->debug->ba_cnt = ba_cnt;
+	priv->debug->ba_acc = ba_acc;
+	priv->debug->ba_cnt_rx = ba_cnt_rx;
+	priv->debug->ba_acc_rx = ba_acc_rx;
+}
+
+#endif /* CW1200_DEBUG_H_INCLUDED */
diff --git a/drivers/net/wireless/cw1200/fwio.c b/drivers/net/wireless/cw1200/fwio.c
new file mode 100644
index 0000000..acdff0f
--- /dev/null
+++ b/drivers/net/wireless/cw1200/fwio.c
@@ -0,0 +1,520 @@
+/*
+ * Firmware I/O code for mac80211 ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on:
+ * ST-Ericsson UMAC CW1200 driver which is
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Ajitpal Singh <ajitpal.singh@stericsson.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/init.h>
+#include <linux/vmalloc.h>
+#include <linux/sched.h>
+#include <linux/firmware.h>
+
+#include "cw1200.h"
+#include "fwio.h"
+#include "hwio.h"
+#include "hwbus.h"
+#include "bh.h"
+
+static int cw1200_get_hw_type(u32 config_reg_val, int *major_revision)
+{
+	int hw_type = -1;
+	u32 silicon_type = (config_reg_val >> 24) & 0x7;
+	u32 silicon_vers = (config_reg_val >> 31) & 0x1;
+
+	switch (silicon_type) {
+	case 0x00:
+		*major_revision = 1;
+		hw_type = HIF_9000_SILICON_VERSATILE;
+		break;
+	case 0x01:
+	case 0x02: /* CW1x00 */
+	case 0x04: /* CW1x60 */
+		*major_revision = silicon_type;
+		if (silicon_vers)
+			hw_type = HIF_8601_VERSATILE;
+		else
+			hw_type = HIF_8601_SILICON;
+		break;
+	default:
+		break;
+	}
+
+	return hw_type;
+}
+
+static int cw1200_load_firmware_cw1200(struct cw1200_common *priv)
+{
+	int ret, block, num_blocks;
+	unsigned i;
+	u32 val32;
+	u32 put = 0, get = 0;
+	u8 *buf = NULL;
+	const char *fw_path;
+	const struct firmware *firmware = NULL;
+
+	/* Macroses are local. */
+#define APB_WRITE(reg, val) \
+	do { \
+		ret = cw1200_apb_write_32(priv, CW1200_APB(reg), (val)); \
+		if (ret < 0) \
+			goto error; \
+	} while (0)
+#define APB_READ(reg, val) \
+	do { \
+		ret = cw1200_apb_read_32(priv, CW1200_APB(reg), &(val)); \
+		if (ret < 0) \
+			goto error; \
+	} while (0)
+#define REG_WRITE(reg, val) \
+	do { \
+		ret = cw1200_reg_write_32(priv, (reg), (val)); \
+		if (ret < 0) \
+			goto error; \
+	} while (0)
+#define REG_READ(reg, val) \
+	do { \
+		ret = cw1200_reg_read_32(priv, (reg), &(val)); \
+		if (ret < 0) \
+			goto error; \
+	} while (0)
+
+	switch (priv->hw_revision) {
+	case CW1200_HW_REV_CUT10:
+		fw_path = FIRMWARE_CUT10;
+		if (!priv->sdd_path)
+			priv->sdd_path = SDD_FILE_10;
+		break;
+	case CW1200_HW_REV_CUT11:
+		fw_path = FIRMWARE_CUT11;
+		if (!priv->sdd_path)
+			priv->sdd_path = SDD_FILE_11;
+		break;
+	case CW1200_HW_REV_CUT20:
+		fw_path = FIRMWARE_CUT20;
+		if (!priv->sdd_path)
+			priv->sdd_path = SDD_FILE_20;
+		break;
+	case CW1200_HW_REV_CUT22:
+		fw_path = FIRMWARE_CUT22;
+		if (!priv->sdd_path)
+			priv->sdd_path = SDD_FILE_22;
+		break;
+	case CW1X60_HW_REV:
+		fw_path = FIRMWARE_CW1X60;
+		if (!priv->sdd_path)
+			priv->sdd_path = SDD_FILE_CW1X60;
+		break;
+	default:
+		pr_err("Invalid silicon revision %d.\n", priv->hw_revision);
+		return -EINVAL;
+	}
+
+	/* Initialize common registers */
+	APB_WRITE(DOWNLOAD_IMAGE_SIZE_REG, DOWNLOAD_ARE_YOU_HERE);
+	APB_WRITE(DOWNLOAD_PUT_REG, 0);
+	APB_WRITE(DOWNLOAD_GET_REG, 0);
+	APB_WRITE(DOWNLOAD_STATUS_REG, DOWNLOAD_PENDING);
+	APB_WRITE(DOWNLOAD_FLAGS_REG, 0);
+
+	/* Write the NOP Instruction */
+	REG_WRITE(ST90TDS_SRAM_BASE_ADDR_REG_ID, 0xFFF20000);
+	REG_WRITE(ST90TDS_AHB_DPORT_REG_ID, 0xEAFFFFFE);
+
+	/* Release CPU from RESET */
+	REG_READ(ST90TDS_CONFIG_REG_ID, val32);
+	val32 &= ~ST90TDS_CONFIG_CPU_RESET_BIT;
+	REG_WRITE(ST90TDS_CONFIG_REG_ID, val32);
+
+	/* Enable Clock */
+	val32 &= ~ST90TDS_CONFIG_CPU_CLK_DIS_BIT;
+	REG_WRITE(ST90TDS_CONFIG_REG_ID, val32);
+
+	/* Load a firmware file */
+	ret = request_firmware(&firmware, fw_path, priv->pdev);
+	if (ret) {
+		pr_err("Can't load firmware file %s.\n", fw_path);
+		goto error;
+	}
+
+	buf = kmalloc(DOWNLOAD_BLOCK_SIZE, GFP_KERNEL | GFP_DMA);
+	if (!buf) {
+		pr_err("Can't allocate firmware load buffer.\n");
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	/* Check if the bootloader is ready */
+	for (i = 0; i < 100; i += 1 + i / 2) {
+		APB_READ(DOWNLOAD_IMAGE_SIZE_REG, val32);
+		if (val32 == DOWNLOAD_I_AM_HERE)
+			break;
+		mdelay(i);
+	} /* End of for loop */
+
+	if (val32 != DOWNLOAD_I_AM_HERE) {
+		pr_err("Bootloader is not ready.\n");
+		ret = -ETIMEDOUT;
+		goto error;
+	}
+
+	/* Calculcate number of download blocks */
+	num_blocks = (firmware->size - 1) / DOWNLOAD_BLOCK_SIZE + 1;
+
+	/* Updating the length in Download Ctrl Area */
+	val32 = firmware->size; /* Explicit cast from size_t to u32 */
+	APB_WRITE(DOWNLOAD_IMAGE_SIZE_REG, val32);
+
+	/* Firmware downloading loop */
+	for (block = 0; block < num_blocks; block++) {
+		size_t tx_size;
+		size_t block_size;
+
+		/* check the download status */
+		APB_READ(DOWNLOAD_STATUS_REG, val32);
+		if (val32 != DOWNLOAD_PENDING) {
+			pr_err("Bootloader reported error %d.\n", val32);
+			ret = -EIO;
+			goto error;
+		}
+
+		/* loop until put - get <= 24K */
+		for (i = 0; i < 100; i++) {
+			APB_READ(DOWNLOAD_GET_REG, get);
+			if ((put - get) <=
+			    (DOWNLOAD_FIFO_SIZE - DOWNLOAD_BLOCK_SIZE))
+				break;
+			mdelay(i);
+		}
+
+		if ((put - get) > (DOWNLOAD_FIFO_SIZE - DOWNLOAD_BLOCK_SIZE)) {
+			pr_err("Timeout waiting for FIFO.\n");
+			ret = -ETIMEDOUT;
+			goto error;
+		}
+
+		/* calculate the block size */
+		tx_size = block_size = min((size_t)(firmware->size - put),
+			(size_t)DOWNLOAD_BLOCK_SIZE);
+
+		memcpy(buf, &firmware->data[put], block_size);
+		if (block_size < DOWNLOAD_BLOCK_SIZE) {
+			memset(&buf[block_size], 0,
+			       DOWNLOAD_BLOCK_SIZE - block_size);
+			tx_size = DOWNLOAD_BLOCK_SIZE;
+		}
+
+		/* send the block to sram */
+		ret = cw1200_apb_write(priv,
+			CW1200_APB(DOWNLOAD_FIFO_OFFSET +
+				   (put & (DOWNLOAD_FIFO_SIZE - 1))),
+			buf, tx_size);
+		if (ret < 0) {
+			pr_err("Can't write firmware block @ %d!\n",
+			       put & (DOWNLOAD_FIFO_SIZE - 1));
+			goto error;
+		}
+
+		/* update the put register */
+		put += block_size;
+		APB_WRITE(DOWNLOAD_PUT_REG, put);
+	} /* End of firmware download loop */
+
+	/* Wait for the download completion */
+	for (i = 0; i < 300; i += 1 + i / 2) {
+		APB_READ(DOWNLOAD_STATUS_REG, val32);
+		if (val32 != DOWNLOAD_PENDING)
+			break;
+		mdelay(i);
+	}
+	if (val32 != DOWNLOAD_SUCCESS) {
+		pr_err("Wait for download completion failed: 0x%.8X\n", val32);
+		ret = -ETIMEDOUT;
+		goto error;
+	} else {
+		pr_info("Firmware download completed.\n");
+		ret = 0;
+	}
+
+error:
+	kfree(buf);
+	if (firmware)
+		release_firmware(firmware);
+	return ret;
+
+#undef APB_WRITE
+#undef APB_READ
+#undef REG_WRITE
+#undef REG_READ
+}
+
+
+static int config_reg_read(struct cw1200_common *priv, u32 *val)
+{
+	switch (priv->hw_type) {
+	case HIF_9000_SILICON_VERSATILE: {
+		u16 val16;
+		int ret = cw1200_reg_read_16(priv,
+					     ST90TDS_CONFIG_REG_ID,
+					     &val16);
+		if (ret < 0)
+			return ret;
+		*val = val16;
+		return 0;
+	}
+	case HIF_8601_VERSATILE:
+	case HIF_8601_SILICON:
+	default:
+		cw1200_reg_read_32(priv, ST90TDS_CONFIG_REG_ID, val);
+		break;
+	}
+	return 0;
+}
+
+static int config_reg_write(struct cw1200_common *priv, u32 val)
+{
+	switch (priv->hw_type) {
+	case HIF_9000_SILICON_VERSATILE:
+		return cw1200_reg_write_16(priv,
+					   ST90TDS_CONFIG_REG_ID,
+					   (u16)val);
+	case HIF_8601_VERSATILE:
+	case HIF_8601_SILICON:
+	default:
+		return cw1200_reg_write_32(priv, ST90TDS_CONFIG_REG_ID, val);
+		break;
+	}
+	return 0;
+}
+
+int cw1200_load_firmware(struct cw1200_common *priv)
+{
+	int ret;
+	int i;
+	u32 val32;
+	u16 val16;
+	int major_revision = -1;
+
+	/* Read CONFIG Register */
+	ret = cw1200_reg_read_32(priv, ST90TDS_CONFIG_REG_ID, &val32);
+	if (ret < 0) {
+		pr_err("Can't read config register.\n");
+		goto out;
+	}
+
+	if (val32 == 0 || val32 == 0xffffffff) {
+		pr_err("Bad config register value (0x%08x)\n", val32);
+		ret = -EIO;
+		goto out;
+	}
+
+	priv->hw_type = cw1200_get_hw_type(val32, &major_revision);
+	if (priv->hw_type < 0) {
+		pr_err("Can't deduce hardware type.\n");
+		ret = -ENOTSUPP;
+		goto out;
+	}
+
+	/* Set DPLL Reg value, and read back to confirm writes work */
+	ret = cw1200_reg_write_32(priv, ST90TDS_TSET_GEN_R_W_REG_ID,
+				  cw1200_dpll_from_clk(priv->hw_refclk));
+	if (ret < 0) {
+		pr_err("Can't write DPLL register.\n");
+		goto out;
+	}
+
+	msleep(20);
+
+	ret = cw1200_reg_read_32(priv,
+		ST90TDS_TSET_GEN_R_W_REG_ID, &val32);
+	if (ret < 0) {
+		pr_err("Can't read DPLL register.\n");
+		goto out;
+	}
+
+	if (val32 != cw1200_dpll_from_clk(priv->hw_refclk)) {
+		pr_err("Unable to initialise DPLL register. Wrote 0x%.8X, Read 0x%.8X.\n",
+		       cw1200_dpll_from_clk(priv->hw_refclk), val32);
+		ret = -EIO;
+		goto out;
+	}
+
+	/* Set wakeup bit in device */
+	ret = cw1200_reg_read_16(priv, ST90TDS_CONTROL_REG_ID, &val16);
+	if (ret < 0) {
+		pr_err("set_wakeup: can't read control register.\n");
+		goto out;
+	}
+
+	ret = cw1200_reg_write_16(priv, ST90TDS_CONTROL_REG_ID,
+		val16 | ST90TDS_CONT_WUP_BIT);
+	if (ret < 0) {
+		pr_err("set_wakeup: can't write control register.\n");
+		goto out;
+	}
+
+	/* Wait for wakeup */
+	for (i = 0; i < 300; i += (1 + i / 2)) {
+		ret = cw1200_reg_read_16(priv,
+			ST90TDS_CONTROL_REG_ID, &val16);
+		if (ret < 0) {
+			pr_err("wait_for_wakeup: can't read control register.\n");
+			goto out;
+		}
+
+		if (val16 & ST90TDS_CONT_RDY_BIT)
+			break;
+
+		msleep(i);
+	}
+
+	if ((val16 & ST90TDS_CONT_RDY_BIT) == 0) {
+		pr_err("wait_for_wakeup: device is not responding.\n");
+		ret = -ETIMEDOUT;
+		goto out;
+	}
+
+	switch (major_revision) {
+	case 1:
+		/* CW1200 Hardware detection logic : Check for CUT1.1 */
+		ret = cw1200_ahb_read_32(priv, CW1200_CUT_ID_ADDR, &val32);
+		if (ret) {
+			pr_err("HW detection: can't read CUT ID.\n");
+			goto out;
+		}
+
+		switch (val32) {
+		case CW1200_CUT_11_ID_STR:
+			pr_info("CW1x00 Cut 1.1 silicon detected.\n");
+			priv->hw_revision = CW1200_HW_REV_CUT11;
+			break;
+		default:
+			pr_info("CW1x00 Cut 1.0 silicon detected.\n");
+			priv->hw_revision = CW1200_HW_REV_CUT10;
+			break;
+		}
+
+		/* According to ST-E, CUT<2.0 has busted BA TID0-3.
+		   Just disable it entirely...
+		*/
+		priv->ba_rx_tid_mask = 0;
+		priv->ba_tx_tid_mask = 0;
+		break;
+	case 2: {
+		u32 ar1, ar2, ar3;
+		ret = cw1200_ahb_read_32(priv, CW1200_CUT2_ID_ADDR, &ar1);
+		if (ret) {
+			pr_err("(1) HW detection: can't read CUT ID\n");
+			goto out;
+		}
+		ret = cw1200_ahb_read_32(priv, CW1200_CUT2_ID_ADDR + 4, &ar2);
+		if (ret) {
+			pr_err("(2) HW detection: can't read CUT ID.\n");
+			goto out;
+		}
+
+		ret = cw1200_ahb_read_32(priv, CW1200_CUT2_ID_ADDR + 8, &ar3);
+		if (ret) {
+			pr_err("(3) HW detection: can't read CUT ID.\n");
+			goto out;
+		}
+
+		if (ar1 == CW1200_CUT_22_ID_STR1 &&
+		    ar2 == CW1200_CUT_22_ID_STR2 &&
+		    ar3 == CW1200_CUT_22_ID_STR3) {
+			pr_info("CW1x00 Cut 2.2 silicon detected.\n");
+			priv->hw_revision = CW1200_HW_REV_CUT22;
+		} else {
+			pr_info("CW1x00 Cut 2.0 silicon detected.\n");
+			priv->hw_revision = CW1200_HW_REV_CUT20;
+		}
+		break;
+	}
+	case 4:
+		pr_info("CW1x60 silicon detected.\n");
+		priv->hw_revision = CW1X60_HW_REV;
+		break;
+	default:
+		pr_err("Unsupported silicon major revision %d.\n",
+		       major_revision);
+		ret = -ENOTSUPP;
+		goto out;
+	}
+
+	/* Checking for access mode */
+	ret = config_reg_read(priv, &val32);
+	if (ret < 0) {
+		pr_err("Can't read config register.\n");
+		goto out;
+	}
+
+	if (!(val32 & ST90TDS_CONFIG_ACCESS_MODE_BIT)) {
+		pr_err("Device is already in QUEUE mode!\n");
+			ret = -EINVAL;
+			goto out;
+	}
+
+	switch (priv->hw_type)  {
+	case HIF_8601_SILICON:
+		if (priv->hw_revision == CW1X60_HW_REV) {
+			pr_err("Can't handle CW1160/1260 firmware load yet.\n");
+			ret = -ENOTSUPP;
+			goto out;
+		}
+		ret = cw1200_load_firmware_cw1200(priv);
+		break;
+	default:
+		pr_err("Can't perform firmware load for hw type %d.\n",
+		       priv->hw_type);
+		ret = -ENOTSUPP;
+		goto out;
+	}
+	if (ret < 0) {
+		pr_err("Firmware load error.\n");
+		goto out;
+	}
+
+	/* Enable interrupt signalling */
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+	ret = __cw1200_irq_enable(priv, 1);
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	if (ret < 0)
+		goto unsubscribe;
+
+	/* Configure device for MESSSAGE MODE */
+	ret = config_reg_read(priv, &val32);
+	if (ret < 0) {
+		pr_err("Can't read config register.\n");
+		goto unsubscribe;
+	}
+	ret = config_reg_write(priv, val32 & ~ST90TDS_CONFIG_ACCESS_MODE_BIT);
+	if (ret < 0) {
+		pr_err("Can't write config register.\n");
+		goto unsubscribe;
+	}
+
+	/* Unless we read the CONFIG Register we are
+	 * not able to get an interrupt
+	 */
+	mdelay(10);
+	config_reg_read(priv, &val32);
+
+out:
+	return ret;
+
+unsubscribe:
+	/* Disable interrupt signalling */
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+	ret = __cw1200_irq_enable(priv, 0);
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
diff --git a/drivers/net/wireless/cw1200/fwio.h b/drivers/net/wireless/cw1200/fwio.h
new file mode 100644
index 0000000..ea30993
--- /dev/null
+++ b/drivers/net/wireless/cw1200/fwio.h
@@ -0,0 +1,39 @@
+/*
+ * Firmware API for mac80211 ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on:
+ * ST-Ericsson UMAC CW1200 driver which is
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Ajitpal Singh <ajitpal.singh@stericsson.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef FWIO_H_INCLUDED
+#define FWIO_H_INCLUDED
+
+#define BOOTLOADER_CW1X60       "boot_cw1x60.bin"
+#define FIRMWARE_CW1X60		"wsm_cw1x60.bin"
+#define FIRMWARE_CUT22		"wsm_22.bin"
+#define FIRMWARE_CUT20		"wsm_20.bin"
+#define FIRMWARE_CUT11		"wsm_11.bin"
+#define FIRMWARE_CUT10		"wsm_10.bin"
+#define SDD_FILE_CW1X60		"sdd_cw1x60.bin"
+#define SDD_FILE_22		"sdd_22.bin"
+#define SDD_FILE_20		"sdd_20.bin"
+#define SDD_FILE_11		"sdd_11.bin"
+#define SDD_FILE_10		"sdd_10.bin"
+
+int cw1200_load_firmware(struct cw1200_common *priv);
+
+/* SDD definitions */
+#define SDD_PTA_CFG_ELT_ID 0xEB
+#define SDD_REFERENCE_FREQUENCY_ELT_ID 0xc5
+u32 cw1200_dpll_from_clk(u16 clk);
+
+#endif
diff --git a/drivers/net/wireless/cw1200/hwbus.h b/drivers/net/wireless/cw1200/hwbus.h
new file mode 100644
index 0000000..8b2fc83
--- /dev/null
+++ b/drivers/net/wireless/cw1200/hwbus.h
@@ -0,0 +1,33 @@
+/*
+ * Common hwbus abstraction layer interface for cw1200 wireless driver
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_HWBUS_H
+#define CW1200_HWBUS_H
+
+struct hwbus_priv;
+
+void cw1200_irq_handler(struct cw1200_common *priv);
+
+/* This MUST be wrapped with hwbus_ops->lock/unlock! */
+int __cw1200_irq_enable(struct cw1200_common *priv, int enable);
+
+struct hwbus_ops {
+	int (*hwbus_memcpy_fromio)(struct hwbus_priv *self, unsigned int addr,
+					void *dst, int count);
+	int (*hwbus_memcpy_toio)(struct hwbus_priv *self, unsigned int addr,
+					const void *src, int count);
+	void (*lock)(struct hwbus_priv *self);
+	void (*unlock)(struct hwbus_priv *self);
+	size_t (*align_size)(struct hwbus_priv *self, size_t size);
+	int (*power_mgmt)(struct hwbus_priv *self, bool suspend);
+};
+
+#endif /* CW1200_HWBUS_H */
diff --git a/drivers/net/wireless/cw1200/hwio.c b/drivers/net/wireless/cw1200/hwio.c
new file mode 100644
index 0000000..ff230b7
--- /dev/null
+++ b/drivers/net/wireless/cw1200/hwio.c
@@ -0,0 +1,312 @@
+/*
+ * Low-level device IO routines for ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on:
+ * ST-Ericsson UMAC CW1200 driver, which is
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Ajitpal Singh <ajitpal.singh@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/types.h>
+
+#include "cw1200.h"
+#include "hwio.h"
+#include "hwbus.h"
+
+ /* Sdio addr is 4*spi_addr */
+#define SPI_REG_ADDR_TO_SDIO(spi_reg_addr) ((spi_reg_addr) << 2)
+#define SDIO_ADDR17BIT(buf_id, mpf, rfu, reg_id_ofs) \
+				((((buf_id)    & 0x1F) << 7) \
+				| (((mpf)        & 1) << 6) \
+				| (((rfu)        & 1) << 5) \
+				| (((reg_id_ofs) & 0x1F) << 0))
+#define MAX_RETRY		3
+
+
+static int __cw1200_reg_read(struct cw1200_common *priv, u16 addr,
+			     void *buf, size_t buf_len, int buf_id)
+{
+	u16 addr_sdio;
+	u32 sdio_reg_addr_17bit;
+
+	/* Check if buffer is aligned to 4 byte boundary */
+	if (WARN_ON(((unsigned long)buf & 3) && (buf_len > 4))) {
+		pr_err("buffer is not aligned.\n");
+		return -EINVAL;
+	}
+
+	/* Convert to SDIO Register Address */
+	addr_sdio = SPI_REG_ADDR_TO_SDIO(addr);
+	sdio_reg_addr_17bit = SDIO_ADDR17BIT(buf_id, 0, 0, addr_sdio);
+
+	return priv->hwbus_ops->hwbus_memcpy_fromio(priv->hwbus_priv,
+						  sdio_reg_addr_17bit,
+						  buf, buf_len);
+}
+
+static int __cw1200_reg_write(struct cw1200_common *priv, u16 addr,
+				const void *buf, size_t buf_len, int buf_id)
+{
+	u16 addr_sdio;
+	u32 sdio_reg_addr_17bit;
+
+	/* Convert to SDIO Register Address */
+	addr_sdio = SPI_REG_ADDR_TO_SDIO(addr);
+	sdio_reg_addr_17bit = SDIO_ADDR17BIT(buf_id, 0, 0, addr_sdio);
+
+	return priv->hwbus_ops->hwbus_memcpy_toio(priv->hwbus_priv,
+						sdio_reg_addr_17bit,
+						buf, buf_len);
+}
+
+static inline int __cw1200_reg_read_32(struct cw1200_common *priv,
+					u16 addr, u32 *val)
+{
+	__le32 tmp;
+	int i = __cw1200_reg_read(priv, addr, &tmp, sizeof(tmp), 0);
+	*val = le32_to_cpu(tmp);
+	return i;
+}
+
+static inline int __cw1200_reg_write_32(struct cw1200_common *priv,
+					u16 addr, u32 val)
+{
+	__le32 tmp = cpu_to_le32(val);
+	return __cw1200_reg_write(priv, addr, &tmp, sizeof(tmp), 0);
+}
+
+static inline int __cw1200_reg_read_16(struct cw1200_common *priv,
+					u16 addr, u16 *val)
+{
+	__le16 tmp;
+	int i = __cw1200_reg_read(priv, addr, &tmp, sizeof(tmp), 0);
+	*val = le16_to_cpu(tmp);
+	return i;
+}
+
+static inline int __cw1200_reg_write_16(struct cw1200_common *priv,
+					u16 addr, u16 val)
+{
+	__le16 tmp = cpu_to_le16(val);
+	return __cw1200_reg_write(priv, addr, &tmp, sizeof(tmp), 0);
+}
+
+int cw1200_reg_read(struct cw1200_common *priv, u16 addr, void *buf,
+			size_t buf_len)
+{
+	int ret;
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+	ret = __cw1200_reg_read(priv, addr, buf, buf_len, 0);
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
+
+int cw1200_reg_write(struct cw1200_common *priv, u16 addr, const void *buf,
+			size_t buf_len)
+{
+	int ret;
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+	ret = __cw1200_reg_write(priv, addr, buf, buf_len, 0);
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
+
+int cw1200_data_read(struct cw1200_common *priv, void *buf, size_t buf_len)
+{
+	int ret, retry = 1;
+	int buf_id_rx = priv->buf_id_rx;
+
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+
+	while (retry <= MAX_RETRY) {
+		ret = __cw1200_reg_read(priv,
+					ST90TDS_IN_OUT_QUEUE_REG_ID, buf,
+					buf_len, buf_id_rx + 1);
+		if (!ret) {
+			buf_id_rx = (buf_id_rx + 1) & 3;
+			priv->buf_id_rx = buf_id_rx;
+			break;
+		} else {
+			retry++;
+			mdelay(1);
+			pr_err("error :[%d]\n", ret);
+		}
+	}
+
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
+
+int cw1200_data_write(struct cw1200_common *priv, const void *buf,
+			size_t buf_len)
+{
+	int ret, retry = 1;
+	int buf_id_tx = priv->buf_id_tx;
+
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+
+	while (retry <= MAX_RETRY) {
+		ret = __cw1200_reg_write(priv,
+					 ST90TDS_IN_OUT_QUEUE_REG_ID, buf,
+					 buf_len, buf_id_tx);
+		if (!ret) {
+			buf_id_tx = (buf_id_tx + 1) & 31;
+			priv->buf_id_tx = buf_id_tx;
+			break;
+		} else {
+			retry++;
+			mdelay(1);
+			pr_err("error :[%d]\n", ret);
+		}
+	}
+
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
+
+int cw1200_indirect_read(struct cw1200_common *priv, u32 addr, void *buf,
+			 size_t buf_len, u32 prefetch, u16 port_addr)
+{
+	u32 val32 = 0;
+	int i, ret;
+
+	if ((buf_len / 2) >= 0x1000) {
+		pr_err("Can't read more than 0xfff words.\n");
+		return -EINVAL;
+	}
+
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+	/* Write address */
+	ret = __cw1200_reg_write_32(priv, ST90TDS_SRAM_BASE_ADDR_REG_ID, addr);
+	if (ret < 0) {
+		pr_err("Can't write address register.\n");
+		goto out;
+	}
+
+	/* Read CONFIG Register Value - We will read 32 bits */
+	ret = __cw1200_reg_read_32(priv, ST90TDS_CONFIG_REG_ID, &val32);
+	if (ret < 0) {
+		pr_err("Can't read config register.\n");
+		goto out;
+	}
+
+	/* Set PREFETCH bit */
+	ret = __cw1200_reg_write_32(priv, ST90TDS_CONFIG_REG_ID,
+					val32 | prefetch);
+	if (ret < 0) {
+		pr_err("Can't write prefetch bit.\n");
+		goto out;
+	}
+
+	/* Check for PRE-FETCH bit to be cleared */
+	for (i = 0; i < 20; i++) {
+		ret = __cw1200_reg_read_32(priv, ST90TDS_CONFIG_REG_ID, &val32);
+		if (ret < 0) {
+			pr_err("Can't check prefetch bit.\n");
+			goto out;
+		}
+		if (!(val32 & prefetch))
+			break;
+
+		mdelay(i);
+	}
+
+	if (val32 & prefetch) {
+		pr_err("Prefetch bit is not cleared.\n");
+		goto out;
+	}
+
+	/* Read data port */
+	ret = __cw1200_reg_read(priv, port_addr, buf, buf_len, 0);
+	if (ret < 0) {
+		pr_err("Can't read data port.\n");
+		goto out;
+	}
+
+out:
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
+
+int cw1200_apb_write(struct cw1200_common *priv, u32 addr, const void *buf,
+			size_t buf_len)
+{
+	int ret;
+
+	if ((buf_len / 2) >= 0x1000) {
+		pr_err("Can't write more than 0xfff words.\n");
+		return -EINVAL;
+	}
+
+	priv->hwbus_ops->lock(priv->hwbus_priv);
+
+	/* Write address */
+	ret = __cw1200_reg_write_32(priv, ST90TDS_SRAM_BASE_ADDR_REG_ID, addr);
+	if (ret < 0) {
+		pr_err("Can't write address register.\n");
+		goto out;
+	}
+
+	/* Write data port */
+	ret = __cw1200_reg_write(priv, ST90TDS_SRAM_DPORT_REG_ID,
+					buf, buf_len, 0);
+	if (ret < 0) {
+		pr_err("Can't write data port.\n");
+		goto out;
+	}
+
+out:
+	priv->hwbus_ops->unlock(priv->hwbus_priv);
+	return ret;
+}
+
+int __cw1200_irq_enable(struct cw1200_common *priv, int enable)
+{
+	u32 val32;
+	u16 val16;
+	int ret;
+
+	if (HIF_8601_SILICON == priv->hw_type) {
+		ret = __cw1200_reg_read_32(priv, ST90TDS_CONFIG_REG_ID, &val32);
+		if (ret < 0) {
+			pr_err("Can't read config register.\n");
+			return ret;
+		}
+
+		if (enable)
+			val32 |= ST90TDS_CONF_IRQ_RDY_ENABLE;
+		else
+			val32 &= ~ST90TDS_CONF_IRQ_RDY_ENABLE;
+
+		ret = __cw1200_reg_write_32(priv, ST90TDS_CONFIG_REG_ID, val32);
+		if (ret < 0) {
+			pr_err("Can't write config register.\n");
+			return ret;
+		}
+	} else {
+		ret = __cw1200_reg_read_16(priv, ST90TDS_CONFIG_REG_ID, &val16);
+		if (ret < 0) {
+			pr_err("Can't read control register.\n");
+			return ret;
+		}
+
+		if (enable)
+			val16 |= ST90TDS_CONT_IRQ_RDY_ENABLE;
+		else
+			val16 &= ~ST90TDS_CONT_IRQ_RDY_ENABLE;
+
+		ret = __cw1200_reg_write_16(priv, ST90TDS_CONFIG_REG_ID, val16);
+		if (ret < 0) {
+			pr_err("Can't write control register.\n");
+			return ret;
+		}
+	}
+	return 0;
+}
diff --git a/drivers/net/wireless/cw1200/hwio.h b/drivers/net/wireless/cw1200/hwio.h
new file mode 100644
index 0000000..ddf5266
--- /dev/null
+++ b/drivers/net/wireless/cw1200/hwio.h
@@ -0,0 +1,247 @@
+/*
+ * Low-level API for mac80211 ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on:
+ * ST-Ericsson UMAC CW1200 driver which is
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Ajitpal Singh <ajitpal.singh@stericsson.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_HWIO_H_INCLUDED
+#define CW1200_HWIO_H_INCLUDED
+
+/* extern */ struct cw1200_common;
+
+#define CW1200_CUT_11_ID_STR		(0x302E3830)
+#define CW1200_CUT_22_ID_STR1		(0x302e3132)
+#define CW1200_CUT_22_ID_STR2		(0x32302e30)
+#define CW1200_CUT_22_ID_STR3		(0x3335)
+#define CW1200_CUT_ID_ADDR		(0xFFF17F90)
+#define CW1200_CUT2_ID_ADDR		(0xFFF1FF90)
+
+/* Download control area */
+/* boot loader start address in SRAM */
+#define DOWNLOAD_BOOT_LOADER_OFFSET	(0x00000000)
+/* 32K, 0x4000 to 0xDFFF */
+#define DOWNLOAD_FIFO_OFFSET		(0x00004000)
+/* 32K */
+#define DOWNLOAD_FIFO_SIZE		(0x00008000)
+/* 128 bytes, 0xFF80 to 0xFFFF */
+#define DOWNLOAD_CTRL_OFFSET		(0x0000FF80)
+#define DOWNLOAD_CTRL_DATA_DWORDS	(32-6)
+
+struct download_cntl_t {
+	/* size of whole firmware file (including Cheksum), host init */
+	u32 image_size;
+	/* downloading flags */
+	u32 flags;
+	/* No. of bytes put into the download, init & updated by host */
+	u32 put;
+	/* last traced program counter, last ARM reg_pc */
+	u32 trace_pc;
+	/* No. of bytes read from the download, host init, device updates */
+	u32 get;
+	/* r0, boot losader status, host init to pending, device updates */
+	u32 status;
+	/* Extra debug info, r1 to r14 if status=r0=DOWNLOAD_EXCEPTION */
+	u32 debug_data[DOWNLOAD_CTRL_DATA_DWORDS];
+};
+
+#define	DOWNLOAD_IMAGE_SIZE_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, image_size))
+#define	DOWNLOAD_FLAGS_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, flags))
+#define DOWNLOAD_PUT_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, put))
+#define DOWNLOAD_TRACE_PC_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, trace_pc))
+#define	DOWNLOAD_GET_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, get))
+#define	DOWNLOAD_STATUS_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, status))
+#define DOWNLOAD_DEBUG_DATA_REG		\
+	(DOWNLOAD_CTRL_OFFSET + offsetof(struct download_cntl_t, debug_data))
+#define DOWNLOAD_DEBUG_DATA_LEN		(108)
+
+#define DOWNLOAD_BLOCK_SIZE		(1024)
+
+/* For boot loader detection */
+#define DOWNLOAD_ARE_YOU_HERE		(0x87654321)
+#define DOWNLOAD_I_AM_HERE		(0x12345678)
+
+/* Download error code */
+#define DOWNLOAD_PENDING		(0xFFFFFFFF)
+#define DOWNLOAD_SUCCESS		(0)
+#define DOWNLOAD_EXCEPTION		(1)
+#define DOWNLOAD_ERR_MEM_1		(2)
+#define DOWNLOAD_ERR_MEM_2		(3)
+#define DOWNLOAD_ERR_SOFTWARE		(4)
+#define DOWNLOAD_ERR_FILE_SIZE		(5)
+#define DOWNLOAD_ERR_CHECKSUM		(6)
+#define DOWNLOAD_ERR_OVERFLOW		(7)
+#define DOWNLOAD_ERR_IMAGE		(8)
+#define DOWNLOAD_ERR_HOST		(9)
+#define DOWNLOAD_ERR_ABORT		(10)
+
+
+#define SYS_BASE_ADDR_SILICON		(0)
+#define PAC_BASE_ADDRESS_SILICON	(SYS_BASE_ADDR_SILICON + 0x09000000)
+#define PAC_SHARED_MEMORY_SILICON	(PAC_BASE_ADDRESS_SILICON)
+
+#define CW1200_APB(addr)		(PAC_SHARED_MEMORY_SILICON + (addr))
+
+/* Device register definitions */
+
+/* WBF - SPI Register Addresses */
+#define ST90TDS_ADDR_ID_BASE		(0x0000)
+/* 16/32 bits */
+#define ST90TDS_CONFIG_REG_ID		(0x0000)
+/* 16/32 bits */
+#define ST90TDS_CONTROL_REG_ID		(0x0001)
+/* 16 bits, Q mode W/R */
+#define ST90TDS_IN_OUT_QUEUE_REG_ID	(0x0002)
+/* 32 bits, AHB bus R/W */
+#define ST90TDS_AHB_DPORT_REG_ID	(0x0003)
+/* 16/32 bits */
+#define ST90TDS_SRAM_BASE_ADDR_REG_ID   (0x0004)
+/* 32 bits, APB bus R/W */
+#define ST90TDS_SRAM_DPORT_REG_ID	(0x0005)
+/* 32 bits, t_settle/general */
+#define ST90TDS_TSET_GEN_R_W_REG_ID	(0x0006)
+/* 16 bits, Q mode read, no length */
+#define ST90TDS_FRAME_OUT_REG_ID	(0x0007)
+#define ST90TDS_ADDR_ID_MAX		(ST90TDS_FRAME_OUT_REG_ID)
+
+/* WBF - Control register bit set */
+/* next o/p length, bit 11 to 0 */
+#define ST90TDS_CONT_NEXT_LEN_MASK	(0x0FFF)
+#define ST90TDS_CONT_WUP_BIT		(BIT(12))
+#define ST90TDS_CONT_RDY_BIT		(BIT(13))
+#define ST90TDS_CONT_IRQ_ENABLE		(BIT(14))
+#define ST90TDS_CONT_RDY_ENABLE		(BIT(15))
+#define ST90TDS_CONT_IRQ_RDY_ENABLE	(BIT(14)|BIT(15))
+
+/* SPI Config register bit set */
+#define ST90TDS_CONFIG_FRAME_BIT	(BIT(2))
+#define ST90TDS_CONFIG_WORD_MODE_BITS	(BIT(3)|BIT(4))
+#define ST90TDS_CONFIG_WORD_MODE_1	(BIT(3))
+#define ST90TDS_CONFIG_WORD_MODE_2	(BIT(4))
+#define ST90TDS_CONFIG_ERROR_0_BIT	(BIT(5))
+#define ST90TDS_CONFIG_ERROR_1_BIT	(BIT(6))
+#define ST90TDS_CONFIG_ERROR_2_BIT	(BIT(7))
+/* TBD: Sure??? */
+#define ST90TDS_CONFIG_CSN_FRAME_BIT	(BIT(7))
+#define ST90TDS_CONFIG_ERROR_3_BIT	(BIT(8))
+#define ST90TDS_CONFIG_ERROR_4_BIT	(BIT(9))
+/* QueueM */
+#define ST90TDS_CONFIG_ACCESS_MODE_BIT	(BIT(10))
+/* AHB bus */
+#define ST90TDS_CONFIG_AHB_PRFETCH_BIT	(BIT(11))
+#define ST90TDS_CONFIG_CPU_CLK_DIS_BIT	(BIT(12))
+/* APB bus */
+#define ST90TDS_CONFIG_PRFETCH_BIT	(BIT(13))
+/* cpu reset */
+#define ST90TDS_CONFIG_CPU_RESET_BIT	(BIT(14))
+#define ST90TDS_CONFIG_CLEAR_INT_BIT	(BIT(15))
+
+/* For CW1200 the IRQ Enable and Ready Bits are in CONFIG register */
+#define ST90TDS_CONF_IRQ_ENABLE		(BIT(16))
+#define ST90TDS_CONF_RDY_ENABLE		(BIT(17))
+#define ST90TDS_CONF_IRQ_RDY_ENABLE	(BIT(16)|BIT(17))
+
+int cw1200_data_read(struct cw1200_common *priv,
+		     void *buf, size_t buf_len);
+int cw1200_data_write(struct cw1200_common *priv,
+		      const void *buf, size_t buf_len);
+
+int cw1200_reg_read(struct cw1200_common *priv, u16 addr,
+		    void *buf, size_t buf_len);
+int cw1200_reg_write(struct cw1200_common *priv, u16 addr,
+		     const void *buf, size_t buf_len);
+
+static inline int cw1200_reg_read_16(struct cw1200_common *priv,
+				     u16 addr, u16 *val)
+{
+	__le32 tmp;
+	int i;
+	i = cw1200_reg_read(priv, addr, &tmp, sizeof(tmp));
+	*val = le32_to_cpu(tmp) & 0xfffff;
+	return i;
+}
+
+static inline int cw1200_reg_write_16(struct cw1200_common *priv,
+				      u16 addr, u16 val)
+{
+	__le32 tmp = cpu_to_le32((u32)val);
+	return cw1200_reg_write(priv, addr, &tmp, sizeof(tmp));
+}
+
+static inline int cw1200_reg_read_32(struct cw1200_common *priv,
+				     u16 addr, u32 *val)
+{
+	__le32 tmp;
+	int i = cw1200_reg_read(priv, addr, &tmp, sizeof(tmp));
+	*val = le32_to_cpu(tmp);
+	return i;
+}
+
+static inline int cw1200_reg_write_32(struct cw1200_common *priv,
+				      u16 addr, u32 val)
+{
+	__le32 tmp = cpu_to_le32(val);
+	return cw1200_reg_write(priv, addr, &tmp, sizeof(val));
+}
+
+int cw1200_indirect_read(struct cw1200_common *priv, u32 addr, void *buf,
+			 size_t buf_len, u32 prefetch, u16 port_addr);
+int cw1200_apb_write(struct cw1200_common *priv, u32 addr, const void *buf,
+		     size_t buf_len);
+
+static inline int cw1200_apb_read(struct cw1200_common *priv, u32 addr,
+				  void *buf, size_t buf_len)
+{
+	return cw1200_indirect_read(priv, addr, buf, buf_len,
+				    ST90TDS_CONFIG_PRFETCH_BIT,
+				    ST90TDS_SRAM_DPORT_REG_ID);
+}
+
+static inline int cw1200_ahb_read(struct cw1200_common *priv, u32 addr,
+				  void *buf, size_t buf_len)
+{
+	return cw1200_indirect_read(priv, addr, buf, buf_len,
+				    ST90TDS_CONFIG_AHB_PRFETCH_BIT,
+				    ST90TDS_AHB_DPORT_REG_ID);
+}
+
+static inline int cw1200_apb_read_32(struct cw1200_common *priv,
+				     u32 addr, u32 *val)
+{
+	__le32 tmp;
+	int i = cw1200_apb_read(priv, addr, &tmp, sizeof(tmp));
+	*val = le32_to_cpu(tmp);
+	return i;
+}
+
+static inline int cw1200_apb_write_32(struct cw1200_common *priv,
+				      u32 addr, u32 val)
+{
+	__le32 tmp = cpu_to_le32(val);
+	return cw1200_apb_write(priv, addr, &tmp, sizeof(val));
+}
+static inline int cw1200_ahb_read_32(struct cw1200_common *priv,
+				     u32 addr, u32 *val)
+{
+	__le32 tmp;
+	int i = cw1200_ahb_read(priv, addr, &tmp, sizeof(tmp));
+	*val = le32_to_cpu(tmp);
+	return i;
+}
+
+#endif /* CW1200_HWIO_H_INCLUDED */
diff --git a/drivers/net/wireless/cw1200/main.c b/drivers/net/wireless/cw1200/main.c
new file mode 100644
index 0000000..3724e73
--- /dev/null
+++ b/drivers/net/wireless/cw1200/main.c
@@ -0,0 +1,605 @@
+/*
+ * mac80211 glue code for mac80211 ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on:
+ * Copyright (c) 2006, Michael Wu <flamingice@sourmilk.net>
+ * Copyright (c) 2007-2009, Christian Lamparter <chunkeey@web.de>
+ * Copyright 2008, Johannes Berg <johannes@sipsolutions.net>
+ *
+ * Based on:
+ * - the islsm (softmac prism54) driver, which is:
+ *   Copyright 2004-2006 Jean-Baptiste Note <jbnote@gmail.com>, et al.
+ * - stlc45xx driver
+ *   Copyright (C) 2008 Nokia Corporation and/or its subsidiary(-ies).
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/firmware.h>
+#include <linux/etherdevice.h>
+#include <linux/vmalloc.h>
+#include <linux/random.h>
+#include <linux/sched.h>
+#include <net/mac80211.h>
+
+#include "cw1200.h"
+#include "txrx.h"
+#include "hwbus.h"
+#include "fwio.h"
+#include "hwio.h"
+#include "bh.h"
+#include "sta.h"
+#include "scan.h"
+#include "debug.h"
+#include "pm.h"
+
+MODULE_AUTHOR("Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>");
+MODULE_DESCRIPTION("Softmac ST-Ericsson CW1200 common code");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("cw1200_core");
+
+/* Accept MAC address of the form macaddr=0x00,0x80,0xE1,0x30,0x40,0x50 */
+static u8 cw1200_mac_template[ETH_ALEN] = {0x02, 0x80, 0xe1, 0x00, 0x00, 0x00};
+module_param_array_named(macaddr, cw1200_mac_template, byte, NULL, S_IRUGO);
+MODULE_PARM_DESC(macaddr, "Override platform_data MAC address");
+
+static char *cw1200_sdd_path;
+module_param(cw1200_sdd_path, charp, 0644);
+MODULE_PARM_DESC(cw1200_sdd_path, "Override platform_data SDD file");
+static int cw1200_refclk;
+module_param(cw1200_refclk, int, 0644);
+MODULE_PARM_DESC(cw1200_refclk, "Override platform_data reference clock");
+
+int cw1200_power_mode = wsm_power_mode_quiescent;
+module_param(cw1200_power_mode, int, 0644);
+MODULE_PARM_DESC(cw1200_power_mode, "WSM power mode.  0 == active, 1 == doze, 2 == quiescent (default)");
+
+#define RATETAB_ENT(_rate, _rateid, _flags)		\
+	{						\
+		.bitrate	= (_rate),		\
+		.hw_value	= (_rateid),		\
+		.flags		= (_flags),		\
+	}
+
+static struct ieee80211_rate cw1200_rates[] = {
+	RATETAB_ENT(10,  0,   0),
+	RATETAB_ENT(20,  1,   0),
+	RATETAB_ENT(55,  2,   0),
+	RATETAB_ENT(110, 3,   0),
+	RATETAB_ENT(60,  6,  0),
+	RATETAB_ENT(90,  7,  0),
+	RATETAB_ENT(120, 8,  0),
+	RATETAB_ENT(180, 9,  0),
+	RATETAB_ENT(240, 10, 0),
+	RATETAB_ENT(360, 11, 0),
+	RATETAB_ENT(480, 12, 0),
+	RATETAB_ENT(540, 13, 0),
+};
+
+static struct ieee80211_rate cw1200_mcs_rates[] = {
+	RATETAB_ENT(65,  14, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(130, 15, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(195, 16, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(260, 17, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(390, 18, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(520, 19, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(585, 20, IEEE80211_TX_RC_MCS),
+	RATETAB_ENT(650, 21, IEEE80211_TX_RC_MCS),
+};
+
+#define cw1200_a_rates		(cw1200_rates + 4)
+#define cw1200_a_rates_size	(ARRAY_SIZE(cw1200_rates) - 4)
+#define cw1200_g_rates		(cw1200_rates + 0)
+#define cw1200_g_rates_size	(ARRAY_SIZE(cw1200_rates))
+#define cw1200_n_rates		(cw1200_mcs_rates)
+#define cw1200_n_rates_size	(ARRAY_SIZE(cw1200_mcs_rates))
+
+
+#define CHAN2G(_channel, _freq, _flags) {			\
+	.band			= IEEE80211_BAND_2GHZ,		\
+	.center_freq		= (_freq),			\
+	.hw_value		= (_channel),			\
+	.flags			= (_flags),			\
+	.max_antenna_gain	= 0,				\
+	.max_power		= 30,				\
+}
+
+#define CHAN5G(_channel, _flags) {				\
+	.band			= IEEE80211_BAND_5GHZ,		\
+	.center_freq	= 5000 + (5 * (_channel)),		\
+	.hw_value		= (_channel),			\
+	.flags			= (_flags),			\
+	.max_antenna_gain	= 0,				\
+	.max_power		= 30,				\
+}
+
+static struct ieee80211_channel cw1200_2ghz_chantable[] = {
+	CHAN2G(1, 2412, 0),
+	CHAN2G(2, 2417, 0),
+	CHAN2G(3, 2422, 0),
+	CHAN2G(4, 2427, 0),
+	CHAN2G(5, 2432, 0),
+	CHAN2G(6, 2437, 0),
+	CHAN2G(7, 2442, 0),
+	CHAN2G(8, 2447, 0),
+	CHAN2G(9, 2452, 0),
+	CHAN2G(10, 2457, 0),
+	CHAN2G(11, 2462, 0),
+	CHAN2G(12, 2467, 0),
+	CHAN2G(13, 2472, 0),
+	CHAN2G(14, 2484, 0),
+};
+
+static struct ieee80211_channel cw1200_5ghz_chantable[] = {
+	CHAN5G(34, 0),		CHAN5G(36, 0),
+	CHAN5G(38, 0),		CHAN5G(40, 0),
+	CHAN5G(42, 0),		CHAN5G(44, 0),
+	CHAN5G(46, 0),		CHAN5G(48, 0),
+	CHAN5G(52, 0),		CHAN5G(56, 0),
+	CHAN5G(60, 0),		CHAN5G(64, 0),
+	CHAN5G(100, 0),		CHAN5G(104, 0),
+	CHAN5G(108, 0),		CHAN5G(112, 0),
+	CHAN5G(116, 0),		CHAN5G(120, 0),
+	CHAN5G(124, 0),		CHAN5G(128, 0),
+	CHAN5G(132, 0),		CHAN5G(136, 0),
+	CHAN5G(140, 0),		CHAN5G(149, 0),
+	CHAN5G(153, 0),		CHAN5G(157, 0),
+	CHAN5G(161, 0),		CHAN5G(165, 0),
+	CHAN5G(184, 0),		CHAN5G(188, 0),
+	CHAN5G(192, 0),		CHAN5G(196, 0),
+	CHAN5G(200, 0),		CHAN5G(204, 0),
+	CHAN5G(208, 0),		CHAN5G(212, 0),
+	CHAN5G(216, 0),
+};
+
+static struct ieee80211_supported_band cw1200_band_2ghz = {
+	.channels = cw1200_2ghz_chantable,
+	.n_channels = ARRAY_SIZE(cw1200_2ghz_chantable),
+	.bitrates = cw1200_g_rates,
+	.n_bitrates = cw1200_g_rates_size,
+	.ht_cap = {
+		.cap = IEEE80211_HT_CAP_GRN_FLD |
+			(1 << IEEE80211_HT_CAP_RX_STBC_SHIFT) |
+			IEEE80211_HT_CAP_MAX_AMSDU,
+		.ht_supported = 1,
+		.ampdu_factor = IEEE80211_HT_MAX_AMPDU_8K,
+		.ampdu_density = IEEE80211_HT_MPDU_DENSITY_NONE,
+		.mcs = {
+			.rx_mask[0] = 0xFF,
+			.rx_highest = __cpu_to_le16(0x41),
+			.tx_params = IEEE80211_HT_MCS_TX_DEFINED,
+		},
+	},
+};
+
+static struct ieee80211_supported_band cw1200_band_5ghz = {
+	.channels = cw1200_5ghz_chantable,
+	.n_channels = ARRAY_SIZE(cw1200_5ghz_chantable),
+	.bitrates = cw1200_a_rates,
+	.n_bitrates = cw1200_a_rates_size,
+	.ht_cap = {
+		.cap = IEEE80211_HT_CAP_GRN_FLD |
+			(1 << IEEE80211_HT_CAP_RX_STBC_SHIFT) |
+			IEEE80211_HT_CAP_MAX_AMSDU,
+		.ht_supported = 1,
+		.ampdu_factor = IEEE80211_HT_MAX_AMPDU_8K,
+		.ampdu_density = IEEE80211_HT_MPDU_DENSITY_NONE,
+		.mcs = {
+			.rx_mask[0] = 0xFF,
+			.rx_highest = __cpu_to_le16(0x41),
+			.tx_params = IEEE80211_HT_MCS_TX_DEFINED,
+		},
+	},
+};
+
+static const unsigned long cw1200_ttl[] = {
+	1 * HZ,	/* VO */
+	2 * HZ,	/* VI */
+	5 * HZ, /* BE */
+	10 * HZ	/* BK */
+};
+
+static const struct ieee80211_ops cw1200_ops = {
+	.start			= cw1200_start,
+	.stop			= cw1200_stop,
+	.add_interface		= cw1200_add_interface,
+	.remove_interface	= cw1200_remove_interface,
+	.change_interface	= cw1200_change_interface,
+	.tx			= cw1200_tx,
+	.hw_scan		= cw1200_hw_scan,
+	.set_tim		= cw1200_set_tim,
+	.sta_notify		= cw1200_sta_notify,
+	.sta_add		= cw1200_sta_add,
+	.sta_remove		= cw1200_sta_remove,
+	.set_key		= cw1200_set_key,
+	.set_rts_threshold	= cw1200_set_rts_threshold,
+	.config			= cw1200_config,
+	.bss_info_changed	= cw1200_bss_info_changed,
+	.prepare_multicast	= cw1200_prepare_multicast,
+	.configure_filter	= cw1200_configure_filter,
+	.conf_tx		= cw1200_conf_tx,
+	.get_stats		= cw1200_get_stats,
+	.ampdu_action		= cw1200_ampdu_action,
+	.flush			= cw1200_flush,
+#ifdef CONFIG_PM
+	.suspend		= cw1200_wow_suspend,
+	.resume			= cw1200_wow_resume,
+#endif
+	/* Intentionally not offloaded:					*/
+	/*.channel_switch	= cw1200_channel_switch,		*/
+	/*.remain_on_channel	= cw1200_remain_on_channel,		*/
+	/*.cancel_remain_on_channel = cw1200_cancel_remain_on_channel,	*/
+};
+
+static int cw1200_ba_rx_tids = -1;
+static int cw1200_ba_tx_tids = -1;
+module_param(cw1200_ba_rx_tids, int, 0644);
+module_param(cw1200_ba_tx_tids, int, 0644);
+MODULE_PARM_DESC(cw1200_ba_rx_tids, "Block ACK RX TIDs");
+MODULE_PARM_DESC(cw1200_ba_tx_tids, "Block ACK TX TIDs");
+
+#ifdef CONFIG_PM
+static const struct wiphy_wowlan_support cw1200_wowlan_support = {
+	/* Support only for limited wowlan functionalities */
+	.flags = WIPHY_WOWLAN_ANY | WIPHY_WOWLAN_DISCONNECT,
+};
+#endif
+
+
+static struct ieee80211_hw *cw1200_init_common(const u8 *macaddr,
+						const bool have_5ghz)
+{
+	int i, band;
+	struct ieee80211_hw *hw;
+	struct cw1200_common *priv;
+
+	hw = ieee80211_alloc_hw(sizeof(struct cw1200_common), &cw1200_ops);
+	if (!hw)
+		return NULL;
+
+	priv = hw->priv;
+	priv->hw = hw;
+	priv->hw_type = -1;
+	priv->mode = NL80211_IFTYPE_UNSPECIFIED;
+	priv->rates = cw1200_rates; /* TODO: fetch from FW */
+	priv->mcs_rates = cw1200_n_rates;
+	if (cw1200_ba_rx_tids != -1)
+		priv->ba_rx_tid_mask = cw1200_ba_rx_tids;
+	else
+		priv->ba_rx_tid_mask = 0xFF; /* Enable RX BLKACK for all TIDs */
+	if (cw1200_ba_tx_tids != -1)
+		priv->ba_tx_tid_mask = cw1200_ba_tx_tids;
+	else
+		priv->ba_tx_tid_mask = 0xff; /* Enable TX BLKACK for all TIDs */
+
+	hw->flags = IEEE80211_HW_SIGNAL_DBM |
+		    IEEE80211_HW_SUPPORTS_PS |
+		    IEEE80211_HW_SUPPORTS_DYNAMIC_PS |
+		    IEEE80211_HW_REPORTS_TX_ACK_STATUS |
+		    IEEE80211_HW_SUPPORTS_UAPSD |
+		    IEEE80211_HW_CONNECTION_MONITOR |
+		    IEEE80211_HW_AMPDU_AGGREGATION |
+		    IEEE80211_HW_TX_AMPDU_SETUP_IN_HW |
+		    IEEE80211_HW_NEED_DTIM_BEFORE_ASSOC;
+
+	hw->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION) |
+					  BIT(NL80211_IFTYPE_ADHOC) |
+					  BIT(NL80211_IFTYPE_AP) |
+					  BIT(NL80211_IFTYPE_MESH_POINT) |
+					  BIT(NL80211_IFTYPE_P2P_CLIENT) |
+					  BIT(NL80211_IFTYPE_P2P_GO);
+
+#ifdef CONFIG_PM
+	hw->wiphy->wowlan = &cw1200_wowlan_support;
+#endif
+
+	hw->wiphy->flags |= WIPHY_FLAG_AP_UAPSD;
+
+	hw->channel_change_time = 1000;	/* TODO: find actual value */
+	hw->queues = 4;
+
+	priv->rts_threshold = -1;
+
+	hw->max_rates = 8;
+	hw->max_rate_tries = 15;
+	hw->extra_tx_headroom = WSM_TX_EXTRA_HEADROOM +
+		8;  /* TKIP IV */
+
+	hw->sta_data_size = sizeof(struct cw1200_sta_priv);
+
+	hw->wiphy->bands[IEEE80211_BAND_2GHZ] = &cw1200_band_2ghz;
+	if (have_5ghz)
+		hw->wiphy->bands[IEEE80211_BAND_5GHZ] = &cw1200_band_5ghz;
+
+	/* Channel params have to be cleared before registering wiphy again */
+	for (band = 0; band < IEEE80211_NUM_BANDS; band++) {
+		struct ieee80211_supported_band *sband = hw->wiphy->bands[band];
+		if (!sband)
+			continue;
+		for (i = 0; i < sband->n_channels; i++) {
+			sband->channels[i].flags = 0;
+			sband->channels[i].max_antenna_gain = 0;
+			sband->channels[i].max_power = 30;
+		}
+	}
+
+	hw->wiphy->max_scan_ssids = 2;
+	hw->wiphy->max_scan_ie_len = IEEE80211_MAX_DATA_LEN;
+
+	if (macaddr)
+		SET_IEEE80211_PERM_ADDR(hw, (u8 *)macaddr);
+	else
+		SET_IEEE80211_PERM_ADDR(hw, cw1200_mac_template);
+
+	/* Fix up mac address if necessary */
+	if (hw->wiphy->perm_addr[3] == 0 &&
+	    hw->wiphy->perm_addr[4] == 0 &&
+	    hw->wiphy->perm_addr[5] == 0) {
+		get_random_bytes(&hw->wiphy->perm_addr[3], 3);
+	}
+
+	mutex_init(&priv->wsm_cmd_mux);
+	mutex_init(&priv->conf_mutex);
+	priv->workqueue = create_singlethread_workqueue("cw1200_wq");
+	sema_init(&priv->scan.lock, 1);
+	INIT_WORK(&priv->scan.work, cw1200_scan_work);
+	INIT_DELAYED_WORK(&priv->scan.probe_work, cw1200_probe_work);
+	INIT_DELAYED_WORK(&priv->scan.timeout, cw1200_scan_timeout);
+	INIT_DELAYED_WORK(&priv->clear_recent_scan_work,
+			  cw1200_clear_recent_scan_work);
+	INIT_DELAYED_WORK(&priv->join_timeout, cw1200_join_timeout);
+	INIT_WORK(&priv->unjoin_work, cw1200_unjoin_work);
+	INIT_WORK(&priv->join_complete_work, cw1200_join_complete_work);
+	INIT_WORK(&priv->wep_key_work, cw1200_wep_key_work);
+	INIT_WORK(&priv->tx_policy_upload_work, tx_policy_upload_work);
+	spin_lock_init(&priv->event_queue_lock);
+	INIT_LIST_HEAD(&priv->event_queue);
+	INIT_WORK(&priv->event_handler, cw1200_event_handler);
+	INIT_DELAYED_WORK(&priv->bss_loss_work, cw1200_bss_loss_work);
+	INIT_WORK(&priv->bss_params_work, cw1200_bss_params_work);
+	spin_lock_init(&priv->bss_loss_lock);
+	spin_lock_init(&priv->ps_state_lock);
+	INIT_WORK(&priv->set_cts_work, cw1200_set_cts_work);
+	INIT_WORK(&priv->set_tim_work, cw1200_set_tim_work);
+	INIT_WORK(&priv->multicast_start_work, cw1200_multicast_start_work);
+	INIT_WORK(&priv->multicast_stop_work, cw1200_multicast_stop_work);
+	INIT_WORK(&priv->link_id_work, cw1200_link_id_work);
+	INIT_DELAYED_WORK(&priv->link_id_gc_work, cw1200_link_id_gc_work);
+	INIT_WORK(&priv->linkid_reset_work, cw1200_link_id_reset);
+	INIT_WORK(&priv->update_filtering_work, cw1200_update_filtering_work);
+	INIT_WORK(&priv->set_beacon_wakeup_period_work,
+		  cw1200_set_beacon_wakeup_period_work);
+	init_timer(&priv->mcast_timeout);
+	priv->mcast_timeout.data = (unsigned long)priv;
+	priv->mcast_timeout.function = cw1200_mcast_timeout;
+
+	if (cw1200_queue_stats_init(&priv->tx_queue_stats,
+				    CW1200_LINK_ID_MAX,
+				    cw1200_skb_dtor,
+				    priv)) {
+		ieee80211_free_hw(hw);
+		return NULL;
+	}
+
+	for (i = 0; i < 4; ++i) {
+		if (cw1200_queue_init(&priv->tx_queue[i],
+				      &priv->tx_queue_stats, i, 16,
+				      cw1200_ttl[i])) {
+			for (; i > 0; i--)
+				cw1200_queue_deinit(&priv->tx_queue[i - 1]);
+			cw1200_queue_stats_deinit(&priv->tx_queue_stats);
+			ieee80211_free_hw(hw);
+			return NULL;
+		}
+	}
+
+	init_waitqueue_head(&priv->channel_switch_done);
+	init_waitqueue_head(&priv->wsm_cmd_wq);
+	init_waitqueue_head(&priv->wsm_startup_done);
+	init_waitqueue_head(&priv->ps_mode_switch_done);
+	wsm_buf_init(&priv->wsm_cmd_buf);
+	spin_lock_init(&priv->wsm_cmd.lock);
+	priv->wsm_cmd.done = 1;
+	tx_policy_init(priv);
+
+	return hw;
+}
+
+static int cw1200_register_common(struct ieee80211_hw *dev)
+{
+	struct cw1200_common *priv = dev->priv;
+	int err;
+
+#ifdef CONFIG_PM
+	err = cw1200_pm_init(&priv->pm_state, priv);
+	if (err) {
+		pr_err("Cannot init PM. (%d).\n",
+		       err);
+		return err;
+	}
+#endif
+
+	err = ieee80211_register_hw(dev);
+	if (err) {
+		pr_err("Cannot register device (%d).\n",
+		       err);
+#ifdef CONFIG_PM
+		cw1200_pm_deinit(&priv->pm_state);
+#endif
+		return err;
+	}
+
+	cw1200_debug_init(priv);
+
+	pr_info("Registered as '%s'\n", wiphy_name(dev->wiphy));
+	return 0;
+}
+
+static void cw1200_free_common(struct ieee80211_hw *dev)
+{
+	ieee80211_free_hw(dev);
+}
+
+static void cw1200_unregister_common(struct ieee80211_hw *dev)
+{
+	struct cw1200_common *priv = dev->priv;
+	int i;
+
+	ieee80211_unregister_hw(dev);
+
+	del_timer_sync(&priv->mcast_timeout);
+	cw1200_unregister_bh(priv);
+
+	cw1200_debug_release(priv);
+
+	mutex_destroy(&priv->conf_mutex);
+
+	wsm_buf_deinit(&priv->wsm_cmd_buf);
+
+	destroy_workqueue(priv->workqueue);
+	priv->workqueue = NULL;
+
+	if (priv->sdd) {
+		release_firmware(priv->sdd);
+		priv->sdd = NULL;
+	}
+
+	for (i = 0; i < 4; ++i)
+		cw1200_queue_deinit(&priv->tx_queue[i]);
+
+	cw1200_queue_stats_deinit(&priv->tx_queue_stats);
+#ifdef CONFIG_PM
+	cw1200_pm_deinit(&priv->pm_state);
+#endif
+}
+
+/* Clock is in KHz */
+u32 cw1200_dpll_from_clk(u16 clk_khz)
+{
+	switch (clk_khz) {
+	case 0x32C8: /* 13000 KHz */
+		return 0x1D89D241;
+	case 0x3E80: /* 16000 KHz */
+		return 0x000001E1;
+	case 0x41A0: /* 16800 KHz */
+		return 0x124931C1;
+	case 0x4B00: /* 19200 KHz */
+		return 0x00000191;
+	case 0x5DC0: /* 24000 KHz */
+		return 0x00000141;
+	case 0x6590: /* 26000 KHz */
+		return 0x0EC4F121;
+	case 0x8340: /* 33600 KHz */
+		return 0x092490E1;
+	case 0x9600: /* 38400 KHz */
+		return 0x100010C1;
+	case 0x9C40: /* 40000 KHz */
+		return 0x000000C1;
+	case 0xBB80: /* 48000 KHz */
+		return 0x000000A1;
+	case 0xCB20: /* 52000 KHz */
+		return 0x07627091;
+	default:
+		pr_err("Unknown Refclk freq (0x%04x), using 2600KHz\n",
+		       clk_khz);
+		return 0x0EC4F121;
+	}
+}
+
+int cw1200_core_probe(const struct hwbus_ops *hwbus_ops,
+		      struct hwbus_priv *hwbus,
+		      struct device *pdev,
+		      struct cw1200_common **core,
+		      int ref_clk, const u8 *macaddr,
+		      const char *sdd_path, bool have_5ghz)
+{
+	int err = -EINVAL;
+	struct ieee80211_hw *dev;
+	struct cw1200_common *priv;
+	struct wsm_operational_mode mode = {
+		.power_mode = cw1200_power_mode,
+		.disable_more_flag_usage = true,
+	};
+
+	dev = cw1200_init_common(macaddr, have_5ghz);
+	if (!dev)
+		goto err;
+
+	priv = dev->priv;
+	priv->hw_refclk = ref_clk;
+	if (cw1200_refclk)
+		priv->hw_refclk = cw1200_refclk;
+
+	priv->sdd_path = (char *)sdd_path;
+	if (cw1200_sdd_path)
+		priv->sdd_path = cw1200_sdd_path;
+
+	priv->hwbus_ops = hwbus_ops;
+	priv->hwbus_priv = hwbus;
+	priv->pdev = pdev;
+	SET_IEEE80211_DEV(priv->hw, pdev);
+
+	/* Pass struct cw1200_common back up */
+	*core = priv;
+
+	err = cw1200_register_bh(priv);
+	if (err)
+		goto err1;
+
+	err = cw1200_load_firmware(priv);
+	if (err)
+		goto err2;
+
+	if (wait_event_interruptible_timeout(priv->wsm_startup_done,
+					     priv->firmware_ready,
+					     3*HZ) <= 0) {
+		/* TODO: Need to find how to reset device
+		   in QUEUE mode properly.
+		*/
+		pr_err("Timeout waiting on device startup\n");
+		err = -ETIMEDOUT;
+		goto err2;
+	}
+
+	/* Set low-power mode. */
+	wsm_set_operational_mode(priv, &mode);
+
+	/* Enable multi-TX confirmation */
+	wsm_use_multi_tx_conf(priv, true);
+
+	err = cw1200_register_common(dev);
+	if (err)
+		goto err2;
+
+	return err;
+
+err2:
+	cw1200_unregister_bh(priv);
+err1:
+	cw1200_free_common(dev);
+err:
+	*core = NULL;
+	return err;
+}
+EXPORT_SYMBOL_GPL(cw1200_core_probe);
+
+void cw1200_core_release(struct cw1200_common *self)
+{
+	/* Disable device interrupts */
+	self->hwbus_ops->lock(self->hwbus_priv);
+	__cw1200_irq_enable(self, 0);
+	self->hwbus_ops->unlock(self->hwbus_priv);
+
+	/* And then clean up */
+	cw1200_unregister_common(self->hw);
+	cw1200_free_common(self->hw);
+	return;
+}
+EXPORT_SYMBOL_GPL(cw1200_core_release);
diff --git a/drivers/net/wireless/cw1200/pm.c b/drivers/net/wireless/cw1200/pm.c
new file mode 100644
index 0000000..b37abb9
--- /dev/null
+++ b/drivers/net/wireless/cw1200/pm.c
@@ -0,0 +1,367 @@
+/*
+ * Mac80211 power management API for ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2011, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/if_ether.h>
+#include "cw1200.h"
+#include "pm.h"
+#include "sta.h"
+#include "bh.h"
+#include "hwbus.h"
+
+#define CW1200_BEACON_SKIPPING_MULTIPLIER 3
+
+struct cw1200_udp_port_filter {
+	struct wsm_udp_port_filter_hdr hdr;
+	/* Up to 4 filters are allowed. */
+	struct wsm_udp_port_filter filters[WSM_MAX_FILTER_ELEMENTS];
+} __packed;
+
+struct cw1200_ether_type_filter {
+	struct wsm_ether_type_filter_hdr hdr;
+	/* Up to 4 filters are allowed. */
+	struct wsm_ether_type_filter filters[WSM_MAX_FILTER_ELEMENTS];
+} __packed;
+
+static struct cw1200_udp_port_filter cw1200_udp_port_filter_on = {
+	.hdr.num = 2,
+	.filters = {
+		[0] = {
+			.action = WSM_FILTER_ACTION_FILTER_OUT,
+			.type = WSM_FILTER_PORT_TYPE_DST,
+			.port = __cpu_to_le16(67), /* DHCP Bootps */
+		},
+		[1] = {
+			.action = WSM_FILTER_ACTION_FILTER_OUT,
+			.type = WSM_FILTER_PORT_TYPE_DST,
+			.port = __cpu_to_le16(68), /* DHCP Bootpc */
+		},
+	}
+};
+
+static struct wsm_udp_port_filter_hdr cw1200_udp_port_filter_off = {
+	.num = 0,
+};
+
+#ifndef ETH_P_WAPI
+#define ETH_P_WAPI     0x88B4
+#endif
+
+static struct cw1200_ether_type_filter cw1200_ether_type_filter_on = {
+	.hdr.num = 4,
+	.filters = {
+		[0] = {
+			.action = WSM_FILTER_ACTION_FILTER_IN,
+			.type = __cpu_to_le16(ETH_P_IP),
+		},
+		[1] = {
+			.action = WSM_FILTER_ACTION_FILTER_IN,
+			.type = __cpu_to_le16(ETH_P_PAE),
+		},
+		[2] = {
+			.action = WSM_FILTER_ACTION_FILTER_IN,
+			.type = __cpu_to_le16(ETH_P_WAPI),
+		},
+		[3] = {
+			.action = WSM_FILTER_ACTION_FILTER_IN,
+			.type = __cpu_to_le16(ETH_P_ARP),
+		},
+	},
+};
+
+static struct wsm_ether_type_filter_hdr cw1200_ether_type_filter_off = {
+	.num = 0,
+};
+
+/* private */
+struct cw1200_suspend_state {
+	unsigned long bss_loss_tmo;
+	unsigned long join_tmo;
+	unsigned long direct_probe;
+	unsigned long link_id_gc;
+	bool beacon_skipping;
+	u8 prev_ps_mode;
+};
+
+static void cw1200_pm_stay_awake_tmo(unsigned long arg)
+{
+	/* XXX what's the point of this ? */
+}
+
+int cw1200_pm_init(struct cw1200_pm_state *pm,
+		   struct cw1200_common *priv)
+{
+	spin_lock_init(&pm->lock);
+
+	init_timer(&pm->stay_awake);
+	pm->stay_awake.data = (unsigned long)pm;
+	pm->stay_awake.function = cw1200_pm_stay_awake_tmo;
+
+	return 0;
+}
+
+void cw1200_pm_deinit(struct cw1200_pm_state *pm)
+{
+	del_timer_sync(&pm->stay_awake);
+}
+
+void cw1200_pm_stay_awake(struct cw1200_pm_state *pm,
+			  unsigned long tmo)
+{
+	long cur_tmo;
+	spin_lock_bh(&pm->lock);
+	cur_tmo = pm->stay_awake.expires - jiffies;
+	if (!timer_pending(&pm->stay_awake) || cur_tmo < (long)tmo)
+		mod_timer(&pm->stay_awake, jiffies + tmo);
+	spin_unlock_bh(&pm->lock);
+}
+
+static long cw1200_suspend_work(struct delayed_work *work)
+{
+	int ret = cancel_delayed_work(work);
+	long tmo;
+	if (ret > 0) {
+		/* Timer is pending */
+		tmo = work->timer.expires - jiffies;
+		if (tmo < 0)
+			tmo = 0;
+	} else {
+		tmo = -1;
+	}
+	return tmo;
+}
+
+static int cw1200_resume_work(struct cw1200_common *priv,
+			       struct delayed_work *work,
+			       unsigned long tmo)
+{
+	if ((long)tmo < 0)
+		return 1;
+
+	return queue_delayed_work(priv->workqueue, work, tmo);
+}
+
+int cw1200_can_suspend(struct cw1200_common *priv)
+{
+	if (atomic_read(&priv->bh_rx)) {
+		wiphy_dbg(priv->hw->wiphy, "Suspend interrupted.\n");
+		return 0;
+	}
+	return 1;
+}
+EXPORT_SYMBOL_GPL(cw1200_can_suspend);
+
+int cw1200_wow_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
+{
+	struct cw1200_common *priv = hw->priv;
+	struct cw1200_pm_state *pm_state = &priv->pm_state;
+	struct cw1200_suspend_state *state;
+	int ret;
+
+	spin_lock_bh(&pm_state->lock);
+	ret = timer_pending(&pm_state->stay_awake);
+	spin_unlock_bh(&pm_state->lock);
+	if (ret)
+		return -EAGAIN;
+
+	/* Do not suspend when datapath is not idle */
+	if (priv->tx_queue_stats.num_queued)
+		return -EBUSY;
+
+	/* Make sure there is no configuration requests in progress. */
+	if (!mutex_trylock(&priv->conf_mutex))
+		return -EBUSY;
+
+	/* Ensure pending operations are done.
+	 * Note also that wow_suspend must return in ~2.5sec, before
+	 * watchdog is triggered.
+	 */
+	if (priv->channel_switch_in_progress)
+		goto revert1;
+
+	/* Do not suspend when join is pending */
+	if (priv->join_pending)
+		goto revert1;
+
+	/* Do not suspend when scanning */
+	if (down_trylock(&priv->scan.lock))
+		goto revert1;
+
+	/* Lock TX. */
+	wsm_lock_tx_async(priv);
+
+	/* Wait to avoid possible race with bh code.
+	 * But do not wait too long...
+	 */
+	if (wait_event_timeout(priv->bh_evt_wq,
+			       !priv->hw_bufs_used, HZ / 10) <= 0)
+		goto revert2;
+
+	/* Set UDP filter */
+	wsm_set_udp_port_filter(priv, &cw1200_udp_port_filter_on.hdr);
+
+	/* Set ethernet frame type filter */
+	wsm_set_ether_type_filter(priv, &cw1200_ether_type_filter_on.hdr);
+
+	/* Allocate state */
+	state = kzalloc(sizeof(struct cw1200_suspend_state), GFP_KERNEL);
+	if (!state)
+		goto revert3;
+
+	/* Change to legacy PS while going to suspend */
+	if (!priv->vif->p2p &&
+	    priv->join_status == CW1200_JOIN_STATUS_STA &&
+	    priv->powersave_mode.mode != WSM_PSM_PS) {
+		state->prev_ps_mode = priv->powersave_mode.mode;
+		priv->powersave_mode.mode = WSM_PSM_PS;
+		cw1200_set_pm(priv, &priv->powersave_mode);
+		if (wait_event_interruptible_timeout(priv->ps_mode_switch_done,
+						     !priv->ps_mode_switch_in_progress, 1*HZ) <= 0) {
+			goto revert3;
+		}
+	}
+
+	/* Store delayed work states. */
+	state->bss_loss_tmo =
+		cw1200_suspend_work(&priv->bss_loss_work);
+	state->join_tmo =
+		cw1200_suspend_work(&priv->join_timeout);
+	state->direct_probe =
+		cw1200_suspend_work(&priv->scan.probe_work);
+	state->link_id_gc =
+		cw1200_suspend_work(&priv->link_id_gc_work);
+
+	cancel_delayed_work_sync(&priv->clear_recent_scan_work);
+	atomic_set(&priv->recent_scan, 0);
+
+	/* Enable beacon skipping */
+	if (priv->join_status == CW1200_JOIN_STATUS_STA &&
+	    priv->join_dtim_period &&
+	    !priv->has_multicast_subscription) {
+		state->beacon_skipping = true;
+		wsm_set_beacon_wakeup_period(priv,
+					     priv->join_dtim_period,
+					     CW1200_BEACON_SKIPPING_MULTIPLIER * priv->join_dtim_period);
+	}
+
+	/* Stop serving thread */
+	if (cw1200_bh_suspend(priv))
+		goto revert4;
+
+	ret = timer_pending(&priv->mcast_timeout);
+	if (ret)
+		goto revert5;
+
+	/* Store suspend state */
+	pm_state->suspend_state = state;
+
+	/* Enable IRQ wake */
+	ret = priv->hwbus_ops->power_mgmt(priv->hwbus_priv, true);
+	if (ret) {
+		wiphy_err(priv->hw->wiphy,
+			  "PM request failed: %d. WoW is disabled.\n", ret);
+		cw1200_wow_resume(hw);
+		return -EBUSY;
+	}
+
+	/* Force resume if event is coming from the device. */
+	if (atomic_read(&priv->bh_rx)) {
+		cw1200_wow_resume(hw);
+		return -EAGAIN;
+	}
+
+	return 0;
+
+revert5:
+	WARN_ON(cw1200_bh_resume(priv));
+revert4:
+	cw1200_resume_work(priv, &priv->bss_loss_work,
+			   state->bss_loss_tmo);
+	cw1200_resume_work(priv, &priv->join_timeout,
+			   state->join_tmo);
+	cw1200_resume_work(priv, &priv->scan.probe_work,
+			   state->direct_probe);
+	cw1200_resume_work(priv, &priv->link_id_gc_work,
+			   state->link_id_gc);
+	kfree(state);
+revert3:
+	wsm_set_udp_port_filter(priv, &cw1200_udp_port_filter_off);
+	wsm_set_ether_type_filter(priv, &cw1200_ether_type_filter_off);
+revert2:
+	wsm_unlock_tx(priv);
+	up(&priv->scan.lock);
+revert1:
+	mutex_unlock(&priv->conf_mutex);
+	return -EBUSY;
+}
+
+int cw1200_wow_resume(struct ieee80211_hw *hw)
+{
+	struct cw1200_common *priv = hw->priv;
+	struct cw1200_pm_state *pm_state = &priv->pm_state;
+	struct cw1200_suspend_state *state;
+
+	state = pm_state->suspend_state;
+	pm_state->suspend_state = NULL;
+
+	/* Disable IRQ wake */
+	priv->hwbus_ops->power_mgmt(priv->hwbus_priv, false);
+
+	/* Scan.lock must be released before BH is resumed other way
+	 * in case when BSS_LOST command arrived the processing of the
+	 * command will be delayed.
+	 */
+	up(&priv->scan.lock);
+
+	/* Resume BH thread */
+	WARN_ON(cw1200_bh_resume(priv));
+
+	/* Restores previous PS mode */
+	if (!priv->vif->p2p && priv->join_status == CW1200_JOIN_STATUS_STA) {
+		priv->powersave_mode.mode = state->prev_ps_mode;
+		cw1200_set_pm(priv, &priv->powersave_mode);
+	}
+
+	if (state->beacon_skipping) {
+		wsm_set_beacon_wakeup_period(priv, priv->beacon_int *
+					     priv->join_dtim_period >
+					     MAX_BEACON_SKIP_TIME_MS ? 1 :
+					     priv->join_dtim_period, 0);
+		state->beacon_skipping = false;
+	}
+
+	/* Resume delayed work */
+	cw1200_resume_work(priv, &priv->bss_loss_work,
+			   state->bss_loss_tmo);
+	cw1200_resume_work(priv, &priv->join_timeout,
+			   state->join_tmo);
+	cw1200_resume_work(priv, &priv->scan.probe_work,
+			   state->direct_probe);
+	cw1200_resume_work(priv, &priv->link_id_gc_work,
+			   state->link_id_gc);
+
+	/* Remove UDP port filter */
+	wsm_set_udp_port_filter(priv, &cw1200_udp_port_filter_off);
+
+	/* Remove ethernet frame type filter */
+	wsm_set_ether_type_filter(priv, &cw1200_ether_type_filter_off);
+
+	/* Unlock datapath */
+	wsm_unlock_tx(priv);
+
+	/* Unlock configuration mutex */
+	mutex_unlock(&priv->conf_mutex);
+
+	/* Free memory */
+	kfree(state);
+
+	return 0;
+}
diff --git a/drivers/net/wireless/cw1200/pm.h b/drivers/net/wireless/cw1200/pm.h
new file mode 100644
index 0000000..3ed90ff
--- /dev/null
+++ b/drivers/net/wireless/cw1200/pm.h
@@ -0,0 +1,43 @@
+/*
+ * Mac80211 power management interface for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2011, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef PM_H_INCLUDED
+#define PM_H_INCLUDED
+
+/* ******************************************************************** */
+/* mac80211 API								*/
+
+/* extern */  struct cw1200_common;
+/* private */ struct cw1200_suspend_state;
+
+struct cw1200_pm_state {
+	struct cw1200_suspend_state *suspend_state;
+	struct timer_list stay_awake;
+	struct platform_device *pm_dev;
+	spinlock_t lock; /* Protect access */
+};
+
+#ifdef CONFIG_PM
+int cw1200_pm_init(struct cw1200_pm_state *pm,
+		    struct cw1200_common *priv);
+void cw1200_pm_deinit(struct cw1200_pm_state *pm);
+int cw1200_wow_suspend(struct ieee80211_hw *hw,
+		       struct cfg80211_wowlan *wowlan);
+int cw1200_wow_resume(struct ieee80211_hw *hw);
+int cw1200_can_suspend(struct cw1200_common *priv);
+void cw1200_pm_stay_awake(struct cw1200_pm_state *pm,
+			  unsigned long tmo);
+#else
+static inline void cw1200_pm_stay_awake(struct cw1200_pm_state *pm,
+					unsigned long tmo) {
+}
+#endif
+#endif
diff --git a/drivers/net/wireless/cw1200/queue.c b/drivers/net/wireless/cw1200/queue.c
new file mode 100644
index 0000000..9c3925f
--- /dev/null
+++ b/drivers/net/wireless/cw1200/queue.c
@@ -0,0 +1,583 @@
+/*
+ * O(1) TX queue with built-in allocator for ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <net/mac80211.h>
+#include <linux/sched.h>
+#include "queue.h"
+#include "cw1200.h"
+#include "debug.h"
+
+/* private */ struct cw1200_queue_item
+{
+	struct list_head	head;
+	struct sk_buff		*skb;
+	u32			packet_id;
+	unsigned long		queue_timestamp;
+	unsigned long		xmit_timestamp;
+	struct cw1200_txpriv	txpriv;
+	u8			generation;
+};
+
+static inline void __cw1200_queue_lock(struct cw1200_queue *queue)
+{
+	struct cw1200_queue_stats *stats = queue->stats;
+	if (queue->tx_locked_cnt++ == 0) {
+		pr_debug("[TX] Queue %d is locked.\n",
+			 queue->queue_id);
+		ieee80211_stop_queue(stats->priv->hw, queue->queue_id);
+	}
+}
+
+static inline void __cw1200_queue_unlock(struct cw1200_queue *queue)
+{
+	struct cw1200_queue_stats *stats = queue->stats;
+	BUG_ON(!queue->tx_locked_cnt);
+	if (--queue->tx_locked_cnt == 0) {
+		pr_debug("[TX] Queue %d is unlocked.\n",
+			 queue->queue_id);
+		ieee80211_wake_queue(stats->priv->hw, queue->queue_id);
+	}
+}
+
+static inline void cw1200_queue_parse_id(u32 packet_id, u8 *queue_generation,
+					 u8 *queue_id, u8 *item_generation,
+					 u8 *item_id)
+{
+	*item_id		= (packet_id >>  0) & 0xFF;
+	*item_generation	= (packet_id >>  8) & 0xFF;
+	*queue_id		= (packet_id >> 16) & 0xFF;
+	*queue_generation	= (packet_id >> 24) & 0xFF;
+}
+
+static inline u32 cw1200_queue_mk_packet_id(u8 queue_generation, u8 queue_id,
+					    u8 item_generation, u8 item_id)
+{
+	return ((u32)item_id << 0) |
+		((u32)item_generation << 8) |
+		((u32)queue_id << 16) |
+		((u32)queue_generation << 24);
+}
+
+static void cw1200_queue_post_gc(struct cw1200_queue_stats *stats,
+				 struct list_head *gc_list)
+{
+	struct cw1200_queue_item *item, *tmp;
+
+	list_for_each_entry_safe(item, tmp, gc_list, head) {
+		list_del(&item->head);
+		stats->skb_dtor(stats->priv, item->skb, &item->txpriv);
+		kfree(item);
+	}
+}
+
+static void cw1200_queue_register_post_gc(struct list_head *gc_list,
+					  struct cw1200_queue_item *item)
+{
+	struct cw1200_queue_item *gc_item;
+	gc_item = kmalloc(sizeof(struct cw1200_queue_item),
+			GFP_ATOMIC);
+	BUG_ON(!gc_item);
+	memcpy(gc_item, item, sizeof(struct cw1200_queue_item));
+	list_add_tail(&gc_item->head, gc_list);
+}
+
+static void __cw1200_queue_gc(struct cw1200_queue *queue,
+			      struct list_head *head,
+			      bool unlock)
+{
+	struct cw1200_queue_stats *stats = queue->stats;
+	struct cw1200_queue_item *item = NULL, *tmp;
+	bool wakeup_stats = false;
+
+	list_for_each_entry_safe(item, tmp, &queue->queue, head) {
+		if (jiffies - item->queue_timestamp < queue->ttl)
+			break;
+		--queue->num_queued;
+		--queue->link_map_cache[item->txpriv.link_id];
+		spin_lock_bh(&stats->lock);
+		--stats->num_queued;
+		if (!--stats->link_map_cache[item->txpriv.link_id])
+			wakeup_stats = true;
+		spin_unlock_bh(&stats->lock);
+		cw1200_debug_tx_ttl(stats->priv);
+		cw1200_queue_register_post_gc(head, item);
+		item->skb = NULL;
+		list_move_tail(&item->head, &queue->free_pool);
+	}
+
+	if (wakeup_stats)
+		wake_up(&stats->wait_link_id_empty);
+
+	if (queue->overfull) {
+		if (queue->num_queued <= (queue->capacity >> 1)) {
+			queue->overfull = false;
+			if (unlock)
+				__cw1200_queue_unlock(queue);
+		} else if (item) {
+			unsigned long tmo = item->queue_timestamp + queue->ttl;
+			mod_timer(&queue->gc, tmo);
+			cw1200_pm_stay_awake(&stats->priv->pm_state,
+					     tmo - jiffies);
+		}
+	}
+}
+
+static void cw1200_queue_gc(unsigned long arg)
+{
+	LIST_HEAD(list);
+	struct cw1200_queue *queue =
+		(struct cw1200_queue *)arg;
+
+	spin_lock_bh(&queue->lock);
+	__cw1200_queue_gc(queue, &list, true);
+	spin_unlock_bh(&queue->lock);
+	cw1200_queue_post_gc(queue->stats, &list);
+}
+
+int cw1200_queue_stats_init(struct cw1200_queue_stats *stats,
+			    size_t map_capacity,
+			    cw1200_queue_skb_dtor_t skb_dtor,
+			    struct cw1200_common *priv)
+{
+	memset(stats, 0, sizeof(*stats));
+	stats->map_capacity = map_capacity;
+	stats->skb_dtor = skb_dtor;
+	stats->priv = priv;
+	spin_lock_init(&stats->lock);
+	init_waitqueue_head(&stats->wait_link_id_empty);
+
+	stats->link_map_cache = kzalloc(sizeof(int) * map_capacity,
+					GFP_KERNEL);
+	if (!stats->link_map_cache)
+		return -ENOMEM;
+
+	return 0;
+}
+
+int cw1200_queue_init(struct cw1200_queue *queue,
+		      struct cw1200_queue_stats *stats,
+		      u8 queue_id,
+		      size_t capacity,
+		      unsigned long ttl)
+{
+	size_t i;
+
+	memset(queue, 0, sizeof(*queue));
+	queue->stats = stats;
+	queue->capacity = capacity;
+	queue->queue_id = queue_id;
+	queue->ttl = ttl;
+	INIT_LIST_HEAD(&queue->queue);
+	INIT_LIST_HEAD(&queue->pending);
+	INIT_LIST_HEAD(&queue->free_pool);
+	spin_lock_init(&queue->lock);
+	init_timer(&queue->gc);
+	queue->gc.data = (unsigned long)queue;
+	queue->gc.function = cw1200_queue_gc;
+
+	queue->pool = kzalloc(sizeof(struct cw1200_queue_item) * capacity,
+			GFP_KERNEL);
+	if (!queue->pool)
+		return -ENOMEM;
+
+	queue->link_map_cache = kzalloc(sizeof(int) * stats->map_capacity,
+			GFP_KERNEL);
+	if (!queue->link_map_cache) {
+		kfree(queue->pool);
+		queue->pool = NULL;
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < capacity; ++i)
+		list_add_tail(&queue->pool[i].head, &queue->free_pool);
+
+	return 0;
+}
+
+int cw1200_queue_clear(struct cw1200_queue *queue)
+{
+	int i;
+	LIST_HEAD(gc_list);
+	struct cw1200_queue_stats *stats = queue->stats;
+	struct cw1200_queue_item *item, *tmp;
+
+	spin_lock_bh(&queue->lock);
+	queue->generation++;
+	list_splice_tail_init(&queue->queue, &queue->pending);
+	list_for_each_entry_safe(item, tmp, &queue->pending, head) {
+		WARN_ON(!item->skb);
+		cw1200_queue_register_post_gc(&gc_list, item);
+		item->skb = NULL;
+		list_move_tail(&item->head, &queue->free_pool);
+	}
+	queue->num_queued = 0;
+	queue->num_pending = 0;
+
+	spin_lock_bh(&stats->lock);
+	for (i = 0; i < stats->map_capacity; ++i) {
+		stats->num_queued -= queue->link_map_cache[i];
+		stats->link_map_cache[i] -= queue->link_map_cache[i];
+		queue->link_map_cache[i] = 0;
+	}
+	spin_unlock_bh(&stats->lock);
+	if (queue->overfull) {
+		queue->overfull = false;
+		__cw1200_queue_unlock(queue);
+	}
+	spin_unlock_bh(&queue->lock);
+	wake_up(&stats->wait_link_id_empty);
+	cw1200_queue_post_gc(stats, &gc_list);
+	return 0;
+}
+
+void cw1200_queue_stats_deinit(struct cw1200_queue_stats *stats)
+{
+	kfree(stats->link_map_cache);
+	stats->link_map_cache = NULL;
+}
+
+void cw1200_queue_deinit(struct cw1200_queue *queue)
+{
+	cw1200_queue_clear(queue);
+	del_timer_sync(&queue->gc);
+	INIT_LIST_HEAD(&queue->free_pool);
+	kfree(queue->pool);
+	kfree(queue->link_map_cache);
+	queue->pool = NULL;
+	queue->link_map_cache = NULL;
+	queue->capacity = 0;
+}
+
+size_t cw1200_queue_get_num_queued(struct cw1200_queue *queue,
+				   u32 link_id_map)
+{
+	size_t ret;
+	int i, bit;
+	size_t map_capacity = queue->stats->map_capacity;
+
+	if (!link_id_map)
+		return 0;
+
+	spin_lock_bh(&queue->lock);
+	if (link_id_map == (u32)-1) {
+		ret = queue->num_queued - queue->num_pending;
+	} else {
+		ret = 0;
+		for (i = 0, bit = 1; i < map_capacity; ++i, bit <<= 1) {
+			if (link_id_map & bit)
+				ret += queue->link_map_cache[i];
+		}
+	}
+	spin_unlock_bh(&queue->lock);
+	return ret;
+}
+
+int cw1200_queue_put(struct cw1200_queue *queue,
+		     struct sk_buff *skb,
+		     struct cw1200_txpriv *txpriv)
+{
+	int ret = 0;
+	LIST_HEAD(gc_list);
+	struct cw1200_queue_stats *stats = queue->stats;
+
+	if (txpriv->link_id >= queue->stats->map_capacity)
+		return -EINVAL;
+
+	spin_lock_bh(&queue->lock);
+	if (!WARN_ON(list_empty(&queue->free_pool))) {
+		struct cw1200_queue_item *item = list_first_entry(
+			&queue->free_pool, struct cw1200_queue_item, head);
+		BUG_ON(item->skb);
+
+		list_move_tail(&item->head, &queue->queue);
+		item->skb = skb;
+		item->txpriv = *txpriv;
+		item->generation = 0;
+		item->packet_id = cw1200_queue_mk_packet_id(queue->generation,
+							    queue->queue_id,
+							    item->generation,
+							    item - queue->pool);
+		item->queue_timestamp = jiffies;
+
+		++queue->num_queued;
+		++queue->link_map_cache[txpriv->link_id];
+
+		spin_lock_bh(&stats->lock);
+		++stats->num_queued;
+		++stats->link_map_cache[txpriv->link_id];
+		spin_unlock_bh(&stats->lock);
+
+		/* TX may happen in parallel sometimes.
+		 * Leave extra queue slots so we don't overflow.
+		 */
+		if (queue->overfull == false &&
+		    queue->num_queued >=
+		    (queue->capacity - (num_present_cpus() - 1))) {
+			queue->overfull = true;
+			__cw1200_queue_lock(queue);
+			mod_timer(&queue->gc, jiffies);
+		}
+	} else {
+		ret = -ENOENT;
+	}
+	spin_unlock_bh(&queue->lock);
+	return ret;
+}
+
+int cw1200_queue_get(struct cw1200_queue *queue,
+		     u32 link_id_map,
+		     struct wsm_tx **tx,
+		     struct ieee80211_tx_info **tx_info,
+		     const struct cw1200_txpriv **txpriv)
+{
+	int ret = -ENOENT;
+	struct cw1200_queue_item *item;
+	struct cw1200_queue_stats *stats = queue->stats;
+	bool wakeup_stats = false;
+
+	spin_lock_bh(&queue->lock);
+	list_for_each_entry(item, &queue->queue, head) {
+		if (link_id_map & BIT(item->txpriv.link_id)) {
+			ret = 0;
+			break;
+		}
+	}
+
+	if (!WARN_ON(ret)) {
+		*tx = (struct wsm_tx *)item->skb->data;
+		*tx_info = IEEE80211_SKB_CB(item->skb);
+		*txpriv = &item->txpriv;
+		(*tx)->packet_id = item->packet_id;
+		list_move_tail(&item->head, &queue->pending);
+		++queue->num_pending;
+		--queue->link_map_cache[item->txpriv.link_id];
+		item->xmit_timestamp = jiffies;
+
+		spin_lock_bh(&stats->lock);
+		--stats->num_queued;
+		if (!--stats->link_map_cache[item->txpriv.link_id])
+			wakeup_stats = true;
+		spin_unlock_bh(&stats->lock);
+	}
+	spin_unlock_bh(&queue->lock);
+	if (wakeup_stats)
+		wake_up(&stats->wait_link_id_empty);
+	return ret;
+}
+
+int cw1200_queue_requeue(struct cw1200_queue *queue, u32 packet_id)
+{
+	int ret = 0;
+	u8 queue_generation, queue_id, item_generation, item_id;
+	struct cw1200_queue_item *item;
+	struct cw1200_queue_stats *stats = queue->stats;
+
+	cw1200_queue_parse_id(packet_id, &queue_generation, &queue_id,
+			      &item_generation, &item_id);
+
+	item = &queue->pool[item_id];
+
+	spin_lock_bh(&queue->lock);
+	BUG_ON(queue_id != queue->queue_id);
+	if (queue_generation != queue->generation) {
+		ret = -ENOENT;
+	} else if (item_id >= (unsigned) queue->capacity) {
+		WARN_ON(1);
+		ret = -EINVAL;
+	} else if (item->generation != item_generation) {
+		WARN_ON(1);
+		ret = -ENOENT;
+	} else {
+		--queue->num_pending;
+		++queue->link_map_cache[item->txpriv.link_id];
+
+		spin_lock_bh(&stats->lock);
+		++stats->num_queued;
+		++stats->link_map_cache[item->txpriv.link_id];
+		spin_unlock_bh(&stats->lock);
+
+		item->generation = ++item_generation;
+		item->packet_id = cw1200_queue_mk_packet_id(queue_generation,
+							    queue_id,
+							    item_generation,
+							    item_id);
+		list_move(&item->head, &queue->queue);
+	}
+	spin_unlock_bh(&queue->lock);
+	return ret;
+}
+
+int cw1200_queue_requeue_all(struct cw1200_queue *queue)
+{
+	struct cw1200_queue_item *item, *tmp;
+	struct cw1200_queue_stats *stats = queue->stats;
+	spin_lock_bh(&queue->lock);
+
+	list_for_each_entry_safe_reverse(item, tmp, &queue->pending, head) {
+		--queue->num_pending;
+		++queue->link_map_cache[item->txpriv.link_id];
+
+		spin_lock_bh(&stats->lock);
+		++stats->num_queued;
+		++stats->link_map_cache[item->txpriv.link_id];
+		spin_unlock_bh(&stats->lock);
+
+		++item->generation;
+		item->packet_id = cw1200_queue_mk_packet_id(queue->generation,
+							    queue->queue_id,
+							    item->generation,
+							    item - queue->pool);
+		list_move(&item->head, &queue->queue);
+	}
+	spin_unlock_bh(&queue->lock);
+
+	return 0;
+}
+
+int cw1200_queue_remove(struct cw1200_queue *queue, u32 packet_id)
+{
+	int ret = 0;
+	u8 queue_generation, queue_id, item_generation, item_id;
+	struct cw1200_queue_item *item;
+	struct cw1200_queue_stats *stats = queue->stats;
+	struct sk_buff *gc_skb = NULL;
+	struct cw1200_txpriv gc_txpriv;
+
+	cw1200_queue_parse_id(packet_id, &queue_generation, &queue_id,
+			      &item_generation, &item_id);
+
+	item = &queue->pool[item_id];
+
+	spin_lock_bh(&queue->lock);
+	BUG_ON(queue_id != queue->queue_id);
+	if (queue_generation != queue->generation) {
+		ret = -ENOENT;
+	} else if (item_id >= (unsigned) queue->capacity) {
+		WARN_ON(1);
+		ret = -EINVAL;
+	} else if (item->generation != item_generation) {
+		WARN_ON(1);
+		ret = -ENOENT;
+	} else {
+		gc_txpriv = item->txpriv;
+		gc_skb = item->skb;
+		item->skb = NULL;
+		--queue->num_pending;
+		--queue->num_queued;
+		++queue->num_sent;
+		++item->generation;
+		/* Do not use list_move_tail here, but list_move:
+		 * try to utilize cache row.
+		 */
+		list_move(&item->head, &queue->free_pool);
+
+		if (queue->overfull &&
+		    (queue->num_queued <= (queue->capacity >> 1))) {
+			queue->overfull = false;
+			__cw1200_queue_unlock(queue);
+		}
+	}
+	spin_unlock_bh(&queue->lock);
+
+	if (gc_skb)
+		stats->skb_dtor(stats->priv, gc_skb, &gc_txpriv);
+
+	return ret;
+}
+
+int cw1200_queue_get_skb(struct cw1200_queue *queue, u32 packet_id,
+			 struct sk_buff **skb,
+			 const struct cw1200_txpriv **txpriv)
+{
+	int ret = 0;
+	u8 queue_generation, queue_id, item_generation, item_id;
+	struct cw1200_queue_item *item;
+	cw1200_queue_parse_id(packet_id, &queue_generation, &queue_id,
+			      &item_generation, &item_id);
+
+	item = &queue->pool[item_id];
+
+	spin_lock_bh(&queue->lock);
+	BUG_ON(queue_id != queue->queue_id);
+	if (queue_generation != queue->generation) {
+		ret = -ENOENT;
+	} else if (item_id >= (unsigned) queue->capacity) {
+		WARN_ON(1);
+		ret = -EINVAL;
+	} else if (item->generation != item_generation) {
+		WARN_ON(1);
+		ret = -ENOENT;
+	} else {
+		*skb = item->skb;
+		*txpriv = &item->txpriv;
+	}
+	spin_unlock_bh(&queue->lock);
+	return ret;
+}
+
+void cw1200_queue_lock(struct cw1200_queue *queue)
+{
+	spin_lock_bh(&queue->lock);
+	__cw1200_queue_lock(queue);
+	spin_unlock_bh(&queue->lock);
+}
+
+void cw1200_queue_unlock(struct cw1200_queue *queue)
+{
+	spin_lock_bh(&queue->lock);
+	__cw1200_queue_unlock(queue);
+	spin_unlock_bh(&queue->lock);
+}
+
+bool cw1200_queue_get_xmit_timestamp(struct cw1200_queue *queue,
+				     unsigned long *timestamp,
+				     u32 pending_frame_id)
+{
+	struct cw1200_queue_item *item;
+	bool ret;
+
+	spin_lock_bh(&queue->lock);
+	ret = !list_empty(&queue->pending);
+	if (ret) {
+		list_for_each_entry(item, &queue->pending, head) {
+			if (item->packet_id != pending_frame_id)
+				if (time_before(item->xmit_timestamp,
+						*timestamp))
+					*timestamp = item->xmit_timestamp;
+		}
+	}
+	spin_unlock_bh(&queue->lock);
+	return ret;
+}
+
+bool cw1200_queue_stats_is_empty(struct cw1200_queue_stats *stats,
+				 u32 link_id_map)
+{
+	bool empty = true;
+
+	spin_lock_bh(&stats->lock);
+	if (link_id_map == (u32)-1) {
+		empty = stats->num_queued == 0;
+	} else {
+		int i;
+		for (i = 0; i < stats->map_capacity; ++i) {
+			if (link_id_map & BIT(i)) {
+				if (stats->link_map_cache[i]) {
+					empty = false;
+					break;
+				}
+			}
+		}
+	}
+	spin_unlock_bh(&stats->lock);
+
+	return empty;
+}
diff --git a/drivers/net/wireless/cw1200/queue.h b/drivers/net/wireless/cw1200/queue.h
new file mode 100644
index 0000000..119f9c7
--- /dev/null
+++ b/drivers/net/wireless/cw1200/queue.h
@@ -0,0 +1,116 @@
+/*
+ * O(1) TX queue with built-in allocator for ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_QUEUE_H_INCLUDED
+#define CW1200_QUEUE_H_INCLUDED
+
+/* private */ struct cw1200_queue_item;
+
+/* extern */ struct sk_buff;
+/* extern */ struct wsm_tx;
+/* extern */ struct cw1200_common;
+/* extern */ struct ieee80211_tx_queue_stats;
+/* extern */ struct cw1200_txpriv;
+
+/* forward */ struct cw1200_queue_stats;
+
+typedef void (*cw1200_queue_skb_dtor_t)(struct cw1200_common *priv,
+					struct sk_buff *skb,
+					const struct cw1200_txpriv *txpriv);
+
+struct cw1200_queue {
+	struct cw1200_queue_stats *stats;
+	size_t			capacity;
+	size_t			num_queued;
+	size_t			num_pending;
+	size_t			num_sent;
+	struct cw1200_queue_item *pool;
+	struct list_head	queue;
+	struct list_head	free_pool;
+	struct list_head	pending;
+	int			tx_locked_cnt;
+	int			*link_map_cache;
+	bool			overfull;
+	spinlock_t		lock; /* Protect queue entry */
+	u8			queue_id;
+	u8			generation;
+	struct timer_list	gc;
+	unsigned long		ttl;
+};
+
+struct cw1200_queue_stats {
+	spinlock_t		lock; /* Protect stats entry */
+	int			*link_map_cache;
+	int			num_queued;
+	size_t			map_capacity;
+	wait_queue_head_t	wait_link_id_empty;
+	cw1200_queue_skb_dtor_t	skb_dtor;
+	struct cw1200_common	*priv;
+};
+
+struct cw1200_txpriv {
+	u8 link_id;
+	u8 raw_link_id;
+	u8 tid;
+	u8 rate_id;
+	u8 offset;
+};
+
+int cw1200_queue_stats_init(struct cw1200_queue_stats *stats,
+			    size_t map_capacity,
+			    cw1200_queue_skb_dtor_t skb_dtor,
+			    struct cw1200_common *priv);
+int cw1200_queue_init(struct cw1200_queue *queue,
+		      struct cw1200_queue_stats *stats,
+		      u8 queue_id,
+		      size_t capacity,
+		      unsigned long ttl);
+int cw1200_queue_clear(struct cw1200_queue *queue);
+void cw1200_queue_stats_deinit(struct cw1200_queue_stats *stats);
+void cw1200_queue_deinit(struct cw1200_queue *queue);
+
+size_t cw1200_queue_get_num_queued(struct cw1200_queue *queue,
+				   u32 link_id_map);
+int cw1200_queue_put(struct cw1200_queue *queue,
+		     struct sk_buff *skb,
+		     struct cw1200_txpriv *txpriv);
+int cw1200_queue_get(struct cw1200_queue *queue,
+		     u32 link_id_map,
+		     struct wsm_tx **tx,
+		     struct ieee80211_tx_info **tx_info,
+		     const struct cw1200_txpriv **txpriv);
+int cw1200_queue_requeue(struct cw1200_queue *queue, u32 packet_id);
+int cw1200_queue_requeue_all(struct cw1200_queue *queue);
+int cw1200_queue_remove(struct cw1200_queue *queue,
+			u32 packet_id);
+int cw1200_queue_get_skb(struct cw1200_queue *queue, u32 packet_id,
+			 struct sk_buff **skb,
+			 const struct cw1200_txpriv **txpriv);
+void cw1200_queue_lock(struct cw1200_queue *queue);
+void cw1200_queue_unlock(struct cw1200_queue *queue);
+bool cw1200_queue_get_xmit_timestamp(struct cw1200_queue *queue,
+				     unsigned long *timestamp,
+				     u32 pending_frame_id);
+
+bool cw1200_queue_stats_is_empty(struct cw1200_queue_stats *stats,
+				 u32 link_id_map);
+
+static inline u8 cw1200_queue_get_queue_id(u32 packet_id)
+{
+	return (packet_id >> 16) & 0xFF;
+}
+
+static inline u8 cw1200_queue_get_generation(u32 packet_id)
+{
+	return (packet_id >>  8) & 0xFF;
+}
+
+#endif /* CW1200_QUEUE_H_INCLUDED */
diff --git a/drivers/net/wireless/cw1200/scan.c b/drivers/net/wireless/cw1200/scan.c
new file mode 100644
index 0000000..ee3c190
--- /dev/null
+++ b/drivers/net/wireless/cw1200/scan.c
@@ -0,0 +1,461 @@
+/*
+ * Scan implementation for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/sched.h>
+#include "cw1200.h"
+#include "scan.h"
+#include "sta.h"
+#include "pm.h"
+
+static void cw1200_scan_restart_delayed(struct cw1200_common *priv);
+
+static int cw1200_scan_start(struct cw1200_common *priv, struct wsm_scan *scan)
+{
+	int ret, i;
+	int tmo = 2000;
+
+	switch (priv->join_status) {
+	case CW1200_JOIN_STATUS_PRE_STA:
+	case CW1200_JOIN_STATUS_JOINING:
+		return -EBUSY;
+	default:
+		break;
+	}
+
+	wiphy_dbg(priv->hw->wiphy, "[SCAN] hw req, type %d, %d channels, flags: 0x%x.\n",
+		  scan->type, scan->num_channels, scan->flags);
+
+	for (i = 0; i < scan->num_channels; ++i)
+		tmo += scan->ch[i].max_chan_time + 10;
+
+	cancel_delayed_work_sync(&priv->clear_recent_scan_work);
+	atomic_set(&priv->scan.in_progress, 1);
+	atomic_set(&priv->recent_scan, 1);
+	cw1200_pm_stay_awake(&priv->pm_state, tmo * HZ / 1000);
+	queue_delayed_work(priv->workqueue, &priv->scan.timeout,
+			   tmo * HZ / 1000);
+	ret = wsm_scan(priv, scan);
+	if (ret) {
+		atomic_set(&priv->scan.in_progress, 0);
+		cancel_delayed_work_sync(&priv->scan.timeout);
+		cw1200_scan_restart_delayed(priv);
+	}
+	return ret;
+}
+
+int cw1200_hw_scan(struct ieee80211_hw *hw,
+		   struct ieee80211_vif *vif,
+		   struct cfg80211_scan_request *req)
+{
+	struct cw1200_common *priv = hw->priv;
+	struct wsm_template_frame frame = {
+		.frame_type = WSM_FRAME_TYPE_PROBE_REQUEST,
+	};
+	int i, ret;
+
+	if (!priv->vif)
+		return -EINVAL;
+
+	/* Scan when P2P_GO corrupt firmware MiniAP mode */
+	if (priv->join_status == CW1200_JOIN_STATUS_AP)
+		return -EOPNOTSUPP;
+
+	if (req->n_ssids == 1 && !req->ssids[0].ssid_len)
+		req->n_ssids = 0;
+
+	wiphy_dbg(hw->wiphy, "[SCAN] Scan request for %d SSIDs.\n",
+		  req->n_ssids);
+
+	if (req->n_ssids > WSM_SCAN_MAX_NUM_OF_SSIDS)
+		return -EINVAL;
+
+	frame.skb = ieee80211_probereq_get(hw, priv->vif, NULL, 0,
+		req->ie_len);
+	if (!frame.skb)
+		return -ENOMEM;
+
+	if (req->ie_len)
+		memcpy(skb_put(frame.skb, req->ie_len), req->ie, req->ie_len);
+
+	/* will be unlocked in cw1200_scan_work() */
+	down(&priv->scan.lock);
+	mutex_lock(&priv->conf_mutex);
+
+	ret = wsm_set_template_frame(priv, &frame);
+	if (!ret) {
+		/* Host want to be the probe responder. */
+		ret = wsm_set_probe_responder(priv, true);
+	}
+	if (ret) {
+		mutex_unlock(&priv->conf_mutex);
+		up(&priv->scan.lock);
+		dev_kfree_skb(frame.skb);
+		return ret;
+	}
+
+	wsm_lock_tx(priv);
+
+	BUG_ON(priv->scan.req);
+	priv->scan.req = req;
+	priv->scan.n_ssids = 0;
+	priv->scan.status = 0;
+	priv->scan.begin = &req->channels[0];
+	priv->scan.curr = priv->scan.begin;
+	priv->scan.end = &req->channels[req->n_channels];
+	priv->scan.output_power = priv->output_power;
+
+	for (i = 0; i < req->n_ssids; ++i) {
+		struct wsm_ssid *dst = &priv->scan.ssids[priv->scan.n_ssids];
+		memcpy(&dst->ssid[0], req->ssids[i].ssid, sizeof(dst->ssid));
+		dst->length = req->ssids[i].ssid_len;
+		++priv->scan.n_ssids;
+	}
+
+	mutex_unlock(&priv->conf_mutex);
+
+	if (frame.skb)
+		dev_kfree_skb(frame.skb);
+	queue_work(priv->workqueue, &priv->scan.work);
+	return 0;
+}
+
+void cw1200_scan_work(struct work_struct *work)
+{
+	struct cw1200_common *priv = container_of(work, struct cw1200_common,
+							scan.work);
+	struct ieee80211_channel **it;
+	struct wsm_scan scan = {
+		.type = WSM_SCAN_TYPE_FOREGROUND,
+		.flags = WSM_SCAN_FLAG_SPLIT_METHOD,
+	};
+	bool first_run = (priv->scan.begin == priv->scan.curr &&
+			  priv->scan.begin != priv->scan.end);
+	int i;
+
+	if (first_run) {
+		/* Firmware gets crazy if scan request is sent
+		 * when STA is joined but not yet associated.
+		 * Force unjoin in this case.
+		 */
+		if (cancel_delayed_work_sync(&priv->join_timeout) > 0)
+			cw1200_join_timeout(&priv->join_timeout.work);
+	}
+
+	mutex_lock(&priv->conf_mutex);
+
+	if (first_run) {
+		if (priv->join_status == CW1200_JOIN_STATUS_STA &&
+		    !(priv->powersave_mode.mode & WSM_PSM_PS)) {
+			struct wsm_set_pm pm = priv->powersave_mode;
+			pm.mode = WSM_PSM_PS;
+			cw1200_set_pm(priv, &pm);
+		} else if (priv->join_status == CW1200_JOIN_STATUS_MONITOR) {
+			/* FW bug: driver has to restart p2p-dev mode
+			 * after scan
+			 */
+			cw1200_disable_listening(priv);
+		}
+	}
+
+	if (!priv->scan.req || (priv->scan.curr == priv->scan.end)) {
+		if (priv->scan.output_power != priv->output_power)
+			wsm_set_output_power(priv, priv->output_power * 10);
+		if (priv->join_status == CW1200_JOIN_STATUS_STA &&
+		    !(priv->powersave_mode.mode & WSM_PSM_PS))
+			cw1200_set_pm(priv, &priv->powersave_mode);
+
+		if (priv->scan.status < 0)
+			wiphy_dbg(priv->hw->wiphy, "[SCAN] Scan failed (%d).\n",
+				  priv->scan.status);
+		else if (priv->scan.req)
+			wiphy_dbg(priv->hw->wiphy,
+				  "[SCAN] Scan completed.\n");
+		else
+			wiphy_dbg(priv->hw->wiphy,
+				  "[SCAN] Scan canceled.\n");
+
+		priv->scan.req = NULL;
+		cw1200_scan_restart_delayed(priv);
+		wsm_unlock_tx(priv);
+		mutex_unlock(&priv->conf_mutex);
+		ieee80211_scan_completed(priv->hw, priv->scan.status ? 1 : 0);
+		up(&priv->scan.lock);
+		return;
+	} else {
+		struct ieee80211_channel *first = *priv->scan.curr;
+		for (it = priv->scan.curr + 1, i = 1;
+		     it != priv->scan.end && i < WSM_SCAN_MAX_NUM_OF_CHANNELS;
+		     ++it, ++i) {
+			if ((*it)->band != first->band)
+				break;
+			if (((*it)->flags ^ first->flags) &
+					IEEE80211_CHAN_PASSIVE_SCAN)
+				break;
+			if (!(first->flags & IEEE80211_CHAN_PASSIVE_SCAN) &&
+			    (*it)->max_power != first->max_power)
+				break;
+		}
+		scan.band = first->band;
+
+		if (priv->scan.req->no_cck)
+			scan.max_tx_rate = WSM_TRANSMIT_RATE_6;
+		else
+			scan.max_tx_rate = WSM_TRANSMIT_RATE_1;
+		scan.num_probes =
+			(first->flags & IEEE80211_CHAN_PASSIVE_SCAN) ? 0 : 2;
+		scan.num_ssids = priv->scan.n_ssids;
+		scan.ssids = &priv->scan.ssids[0];
+		scan.num_channels = it - priv->scan.curr;
+		/* TODO: Is it optimal? */
+		scan.probe_delay = 100;
+		/* It is not stated in WSM specification, however
+		 * FW team says that driver may not use FG scan
+		 * when joined.
+		 */
+		if (priv->join_status == CW1200_JOIN_STATUS_STA) {
+			scan.type = WSM_SCAN_TYPE_BACKGROUND;
+			scan.flags = WSM_SCAN_FLAG_FORCE_BACKGROUND;
+		}
+		scan.ch = kzalloc(
+			sizeof(struct wsm_scan_ch) * (it - priv->scan.curr),
+			GFP_KERNEL);
+		if (!scan.ch) {
+			priv->scan.status = -ENOMEM;
+			goto fail;
+		}
+		for (i = 0; i < scan.num_channels; ++i) {
+			scan.ch[i].number = priv->scan.curr[i]->hw_value;
+			if (priv->scan.curr[i]->flags & IEEE80211_CHAN_PASSIVE_SCAN) {
+				scan.ch[i].min_chan_time = 50;
+				scan.ch[i].max_chan_time = 100;
+			} else {
+				scan.ch[i].min_chan_time = 10;
+				scan.ch[i].max_chan_time = 25;
+			}
+		}
+		if (!(first->flags & IEEE80211_CHAN_PASSIVE_SCAN) &&
+		    priv->scan.output_power != first->max_power) {
+			priv->scan.output_power = first->max_power;
+			wsm_set_output_power(priv,
+					     priv->scan.output_power * 10);
+		}
+		priv->scan.status = cw1200_scan_start(priv, &scan);
+		kfree(scan.ch);
+		if (priv->scan.status)
+			goto fail;
+		priv->scan.curr = it;
+	}
+	mutex_unlock(&priv->conf_mutex);
+	return;
+
+fail:
+	priv->scan.curr = priv->scan.end;
+	mutex_unlock(&priv->conf_mutex);
+	queue_work(priv->workqueue, &priv->scan.work);
+	return;
+}
+
+static void cw1200_scan_restart_delayed(struct cw1200_common *priv)
+{
+	/* FW bug: driver has to restart p2p-dev mode after scan. */
+	if (priv->join_status == CW1200_JOIN_STATUS_MONITOR) {
+		cw1200_enable_listening(priv);
+		cw1200_update_filtering(priv);
+	}
+
+	if (priv->delayed_unjoin) {
+		priv->delayed_unjoin = false;
+		if (queue_work(priv->workqueue, &priv->unjoin_work) <= 0)
+			wsm_unlock_tx(priv);
+	} else if (priv->delayed_link_loss) {
+			wiphy_dbg(priv->hw->wiphy, "[CQM] Requeue BSS loss.\n");
+			priv->delayed_link_loss = 0;
+			cw1200_cqm_bssloss_sm(priv, 1, 0, 0);
+	}
+}
+
+static void cw1200_scan_complete(struct cw1200_common *priv)
+{
+	queue_delayed_work(priv->workqueue, &priv->clear_recent_scan_work, HZ);
+	if (priv->scan.direct_probe) {
+		wiphy_dbg(priv->hw->wiphy, "[SCAN] Direct probe complete.\n");
+		cw1200_scan_restart_delayed(priv);
+		priv->scan.direct_probe = 0;
+		up(&priv->scan.lock);
+		wsm_unlock_tx(priv);
+	} else {
+		cw1200_scan_work(&priv->scan.work);
+	}
+}
+
+void cw1200_scan_failed_cb(struct cw1200_common *priv)
+{
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED)
+		/* STA is stopped. */
+		return;
+
+	if (cancel_delayed_work_sync(&priv->scan.timeout) > 0) {
+		priv->scan.status = -EIO;
+		queue_delayed_work(priv->workqueue, &priv->scan.timeout, 0);
+	}
+}
+
+
+void cw1200_scan_complete_cb(struct cw1200_common *priv,
+				struct wsm_scan_complete *arg)
+{
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED)
+		/* STA is stopped. */
+		return;
+
+	if (cancel_delayed_work_sync(&priv->scan.timeout) > 0) {
+		priv->scan.status = 1;
+		queue_delayed_work(priv->workqueue, &priv->scan.timeout, 0);
+	}
+}
+
+void cw1200_clear_recent_scan_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common,
+			     clear_recent_scan_work.work);
+	atomic_xchg(&priv->recent_scan, 0);
+}
+
+void cw1200_scan_timeout(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, scan.timeout.work);
+	if (atomic_xchg(&priv->scan.in_progress, 0)) {
+		if (priv->scan.status > 0) {
+			priv->scan.status = 0;
+		} else if (!priv->scan.status) {
+			wiphy_warn(priv->hw->wiphy,
+				   "Timeout waiting for scan complete notification.\n");
+			priv->scan.status = -ETIMEDOUT;
+			priv->scan.curr = priv->scan.end;
+			wsm_stop_scan(priv);
+		}
+		cw1200_scan_complete(priv);
+	}
+}
+
+void cw1200_probe_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, scan.probe_work.work);
+	u8 queue_id = cw1200_queue_get_queue_id(priv->pending_frame_id);
+	struct cw1200_queue *queue = &priv->tx_queue[queue_id];
+	const struct cw1200_txpriv *txpriv;
+	struct wsm_tx *wsm;
+	struct wsm_template_frame frame = {
+		.frame_type = WSM_FRAME_TYPE_PROBE_REQUEST,
+	};
+	struct wsm_ssid ssids[1] = {{
+		.length = 0,
+	} };
+	struct wsm_scan_ch ch[1] = {{
+		.min_chan_time = 0,
+		.max_chan_time = 10,
+	} };
+	struct wsm_scan scan = {
+		.type = WSM_SCAN_TYPE_FOREGROUND,
+		.num_probes = 1,
+		.probe_delay = 0,
+		.num_channels = 1,
+		.ssids = ssids,
+		.ch = ch,
+	};
+	u8 *ies;
+	size_t ies_len;
+	int ret;
+
+	wiphy_dbg(priv->hw->wiphy, "[SCAN] Direct probe work.\n");
+
+	mutex_lock(&priv->conf_mutex);
+	if (down_trylock(&priv->scan.lock)) {
+		/* Scan is already in progress. Requeue self. */
+		schedule();
+		queue_delayed_work(priv->workqueue,
+				   &priv->scan.probe_work, HZ / 10);
+		mutex_unlock(&priv->conf_mutex);
+		return;
+	}
+
+	/* Make sure we still have a pending probe req */
+	if (cw1200_queue_get_skb(queue,	priv->pending_frame_id,
+				 &frame.skb, &txpriv)) {
+		up(&priv->scan.lock);
+		mutex_unlock(&priv->conf_mutex);
+		wsm_unlock_tx(priv);
+		return;
+	}
+	wsm = (struct wsm_tx *)frame.skb->data;
+	scan.max_tx_rate = wsm->max_tx_rate;
+	scan.band = (priv->channel->band == IEEE80211_BAND_5GHZ) ?
+		WSM_PHY_BAND_5G : WSM_PHY_BAND_2_4G;
+	if (priv->join_status == CW1200_JOIN_STATUS_STA ||
+	    priv->join_status == CW1200_JOIN_STATUS_IBSS) {
+		scan.type = WSM_SCAN_TYPE_BACKGROUND;
+		scan.flags = WSM_SCAN_FLAG_FORCE_BACKGROUND;
+	}
+	ch[0].number = priv->channel->hw_value;
+
+	skb_pull(frame.skb, txpriv->offset);
+
+	ies = &frame.skb->data[sizeof(struct ieee80211_hdr_3addr)];
+	ies_len = frame.skb->len - sizeof(struct ieee80211_hdr_3addr);
+
+	if (ies_len) {
+		u8 *ssidie =
+			(u8 *)cfg80211_find_ie(WLAN_EID_SSID, ies, ies_len);
+		if (ssidie && ssidie[1] && ssidie[1] <= sizeof(ssids[0].ssid)) {
+			u8 *nextie = &ssidie[2 + ssidie[1]];
+			/* Remove SSID from the IE list. It has to be provided
+			 * as a separate argument in cw1200_scan_start call
+			 */
+
+			/* Store SSID localy */
+			ssids[0].length = ssidie[1];
+			memcpy(ssids[0].ssid, &ssidie[2], ssids[0].length);
+			scan.num_ssids = 1;
+
+			/* Remove SSID from IE list */
+			ssidie[1] = 0;
+			memmove(&ssidie[2], nextie, &ies[ies_len] - nextie);
+			skb_trim(frame.skb, frame.skb->len - ssids[0].length);
+		}
+	}
+
+	/* FW bug: driver has to restart p2p-dev mode after scan */
+	if (priv->join_status == CW1200_JOIN_STATUS_MONITOR)
+		cw1200_disable_listening(priv);
+	ret = wsm_set_template_frame(priv, &frame);
+	priv->scan.direct_probe = 1;
+	if (!ret) {
+		wsm_flush_tx(priv);
+		ret = cw1200_scan_start(priv, &scan);
+	}
+	mutex_unlock(&priv->conf_mutex);
+
+	skb_push(frame.skb, txpriv->offset);
+	if (!ret)
+		IEEE80211_SKB_CB(frame.skb)->flags |= IEEE80211_TX_STAT_ACK;
+	BUG_ON(cw1200_queue_remove(queue, priv->pending_frame_id));
+
+	if (ret) {
+		priv->scan.direct_probe = 0;
+		up(&priv->scan.lock);
+		wsm_unlock_tx(priv);
+	}
+
+	return;
+}
diff --git a/drivers/net/wireless/cw1200/scan.h b/drivers/net/wireless/cw1200/scan.h
new file mode 100644
index 0000000..5a8296c
--- /dev/null
+++ b/drivers/net/wireless/cw1200/scan.h
@@ -0,0 +1,56 @@
+/*
+ * Scan interface for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef SCAN_H_INCLUDED
+#define SCAN_H_INCLUDED
+
+#include <linux/semaphore.h>
+#include "wsm.h"
+
+/* external */ struct sk_buff;
+/* external */ struct cfg80211_scan_request;
+/* external */ struct ieee80211_channel;
+/* external */ struct ieee80211_hw;
+/* external */ struct work_struct;
+
+struct cw1200_scan {
+	struct semaphore lock;
+	struct work_struct work;
+	struct delayed_work timeout;
+	struct cfg80211_scan_request *req;
+	struct ieee80211_channel **begin;
+	struct ieee80211_channel **curr;
+	struct ieee80211_channel **end;
+	struct wsm_ssid ssids[WSM_SCAN_MAX_NUM_OF_SSIDS];
+	int output_power;
+	int n_ssids;
+	int status;
+	atomic_t in_progress;
+	/* Direct probe requests workaround */
+	struct delayed_work probe_work;
+	int direct_probe;
+};
+
+int cw1200_hw_scan(struct ieee80211_hw *hw,
+		   struct ieee80211_vif *vif,
+		   struct cfg80211_scan_request *req);
+void cw1200_scan_work(struct work_struct *work);
+void cw1200_scan_timeout(struct work_struct *work);
+void cw1200_clear_recent_scan_work(struct work_struct *work);
+void cw1200_scan_complete_cb(struct cw1200_common *priv,
+			     struct wsm_scan_complete *arg);
+void cw1200_scan_failed_cb(struct cw1200_common *priv);
+
+/* ******************************************************************** */
+/* Raw probe requests TX workaround					*/
+void cw1200_probe_work(struct work_struct *work);
+
+#endif
diff --git a/drivers/net/wireless/cw1200/sta.c b/drivers/net/wireless/cw1200/sta.c
new file mode 100644
index 0000000..7365674
--- /dev/null
+++ b/drivers/net/wireless/cw1200/sta.c
@@ -0,0 +1,2403 @@
+/*
+ * Mac80211 STA API for ST-Ericsson CW1200 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/vmalloc.h>
+#include <linux/sched.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+
+#include "cw1200.h"
+#include "sta.h"
+#include "fwio.h"
+#include "bh.h"
+#include "debug.h"
+
+#ifndef ERP_INFO_BYTE_OFFSET
+#define ERP_INFO_BYTE_OFFSET 2
+#endif
+
+static void cw1200_do_join(struct cw1200_common *priv);
+static void cw1200_do_unjoin(struct cw1200_common *priv);
+
+static int cw1200_upload_beacon(struct cw1200_common *priv);
+static int cw1200_upload_pspoll(struct cw1200_common *priv);
+static int cw1200_upload_null(struct cw1200_common *priv);
+static int cw1200_upload_qosnull(struct cw1200_common *priv);
+static int cw1200_start_ap(struct cw1200_common *priv);
+static int cw1200_update_beaconing(struct cw1200_common *priv);
+static int cw1200_enable_beaconing(struct cw1200_common *priv,
+				   bool enable);
+static void __cw1200_sta_notify(struct ieee80211_hw *dev,
+				struct ieee80211_vif *vif,
+				enum sta_notify_cmd notify_cmd,
+				int link_id);
+static int __cw1200_flush(struct cw1200_common *priv, bool drop);
+
+static inline void __cw1200_free_event_queue(struct list_head *list)
+{
+	struct cw1200_wsm_event *event, *tmp;
+	list_for_each_entry_safe(event, tmp, list, link) {
+		list_del(&event->link);
+		kfree(event);
+	}
+}
+
+/* ******************************************************************** */
+/* STA API								*/
+
+int cw1200_start(struct ieee80211_hw *dev)
+{
+	struct cw1200_common *priv = dev->priv;
+	int ret = 0;
+
+	cw1200_pm_stay_awake(&priv->pm_state, HZ);
+
+	mutex_lock(&priv->conf_mutex);
+
+	/* default EDCA */
+	WSM_EDCA_SET(&priv->edca, 0, 0x0002, 0x0003, 0x0007, 47, 0xc8, false);
+	WSM_EDCA_SET(&priv->edca, 1, 0x0002, 0x0007, 0x000f, 94, 0xc8, false);
+	WSM_EDCA_SET(&priv->edca, 2, 0x0003, 0x000f, 0x03ff, 0, 0xc8, false);
+	WSM_EDCA_SET(&priv->edca, 3, 0x0007, 0x000f, 0x03ff, 0, 0xc8, false);
+	ret = wsm_set_edca_params(priv, &priv->edca);
+	if (ret)
+		goto out;
+
+	ret = cw1200_set_uapsd_param(priv, &priv->edca);
+	if (ret)
+		goto out;
+
+	priv->setbssparams_done = false;
+
+	memcpy(priv->mac_addr, dev->wiphy->perm_addr, ETH_ALEN);
+	priv->mode = NL80211_IFTYPE_MONITOR;
+	priv->wep_default_key_id = -1;
+
+	priv->cqm_beacon_loss_count = 10;
+
+	ret = cw1200_setup_mac(priv);
+	if (ret)
+		goto out;
+
+out:
+	mutex_unlock(&priv->conf_mutex);
+	return ret;
+}
+
+void cw1200_stop(struct ieee80211_hw *dev)
+{
+	struct cw1200_common *priv = dev->priv;
+	LIST_HEAD(list);
+	int i;
+
+	wsm_lock_tx(priv);
+
+	while (down_trylock(&priv->scan.lock)) {
+		/* Scan is in progress. Force it to stop. */
+		priv->scan.req = NULL;
+		schedule();
+	}
+	up(&priv->scan.lock);
+
+	cancel_delayed_work_sync(&priv->scan.probe_work);
+	cancel_delayed_work_sync(&priv->scan.timeout);
+	cancel_delayed_work_sync(&priv->clear_recent_scan_work);
+	cancel_delayed_work_sync(&priv->join_timeout);
+	cw1200_cqm_bssloss_sm(priv, 0, 0, 0);
+	cancel_work_sync(&priv->unjoin_work);
+	cancel_delayed_work_sync(&priv->link_id_gc_work);
+	flush_workqueue(priv->workqueue);
+	del_timer_sync(&priv->mcast_timeout);
+	mutex_lock(&priv->conf_mutex);
+	priv->mode = NL80211_IFTYPE_UNSPECIFIED;
+	priv->listening = false;
+
+	spin_lock(&priv->event_queue_lock);
+	list_splice_init(&priv->event_queue, &list);
+	spin_unlock(&priv->event_queue_lock);
+	__cw1200_free_event_queue(&list);
+
+
+	priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+	priv->join_pending = false;
+
+	for (i = 0; i < 4; i++)
+		cw1200_queue_clear(&priv->tx_queue[i]);
+	mutex_unlock(&priv->conf_mutex);
+	tx_policy_clean(priv);
+
+	/* HACK! */
+	if (atomic_xchg(&priv->tx_lock, 1) != 1)
+		pr_debug("[STA] TX is force-unlocked due to stop request.\n");
+
+	wsm_unlock_tx(priv);
+	atomic_xchg(&priv->tx_lock, 0); /* for recovery to work */
+}
+
+static int cw1200_bssloss_mitigation = 1;
+module_param(cw1200_bssloss_mitigation, int, 0644);
+MODULE_PARM_DESC(cw1200_bssloss_mitigation, "BSS Loss mitigation. 0 == disabled, 1 == enabled (default)");
+
+
+void __cw1200_cqm_bssloss_sm(struct cw1200_common *priv,
+			     int init, int good, int bad)
+{
+	int tx = 0;
+
+	priv->delayed_link_loss = 0;
+	cancel_work_sync(&priv->bss_params_work);
+
+	pr_debug("[STA] CQM BSSLOSS_SM: state: %d init %d good %d bad: %d txlock: %d uj: %d\n",
+		 priv->bss_loss_state,
+		 init, good, bad,
+		 atomic_read(&priv->tx_lock),
+		 priv->delayed_unjoin);
+
+	/* If we have a pending unjoin */
+	if (priv->delayed_unjoin)
+		return;
+
+	if (init) {
+		queue_delayed_work(priv->workqueue,
+				   &priv->bss_loss_work,
+				   HZ);
+		priv->bss_loss_state = 0;
+
+		/* Skip the confimration procedure in P2P case */
+		if (!priv->vif->p2p && !atomic_read(&priv->tx_lock))
+			tx = 1;
+	} else if (good) {
+		cancel_delayed_work_sync(&priv->bss_loss_work);
+		priv->bss_loss_state = 0;
+		queue_work(priv->workqueue, &priv->bss_params_work);
+	} else if (bad) {
+		/* XXX Should we just keep going until we time out? */
+		if (priv->bss_loss_state < 3)
+			tx = 1;
+	} else {
+		cancel_delayed_work_sync(&priv->bss_loss_work);
+		priv->bss_loss_state = 0;
+	}
+
+	/* Bypass mitigation if it's disabled */
+	if (!cw1200_bssloss_mitigation)
+		tx = 0;
+
+	/* Spit out a NULL packet to our AP if necessary */
+	if (tx) {
+		struct sk_buff *skb;
+
+		priv->bss_loss_state++;
+
+		skb = ieee80211_nullfunc_get(priv->hw, priv->vif);
+		WARN_ON(!skb);
+		if (skb)
+			cw1200_tx(priv->hw, NULL, skb);
+	}
+}
+
+int cw1200_add_interface(struct ieee80211_hw *dev,
+			 struct ieee80211_vif *vif)
+{
+	int ret;
+	struct cw1200_common *priv = dev->priv;
+	/* __le32 auto_calibration_mode = __cpu_to_le32(1); */
+
+	vif->driver_flags |= IEEE80211_VIF_BEACON_FILTER |
+			     IEEE80211_VIF_SUPPORTS_CQM_RSSI;
+
+	mutex_lock(&priv->conf_mutex);
+
+	if (priv->mode != NL80211_IFTYPE_MONITOR) {
+		mutex_unlock(&priv->conf_mutex);
+		return -EOPNOTSUPP;
+	}
+
+	switch (vif->type) {
+	case NL80211_IFTYPE_STATION:
+	case NL80211_IFTYPE_ADHOC:
+	case NL80211_IFTYPE_MESH_POINT:
+	case NL80211_IFTYPE_AP:
+		priv->mode = vif->type;
+		break;
+	default:
+		mutex_unlock(&priv->conf_mutex);
+		return -EOPNOTSUPP;
+	}
+
+	priv->vif = vif;
+	memcpy(priv->mac_addr, vif->addr, ETH_ALEN);
+	ret = cw1200_setup_mac(priv);
+	/* Enable auto-calibration */
+	/* Exception in subsequent channel switch; disabled.
+	 *  wsm_write_mib(priv, WSM_MIB_ID_SET_AUTO_CALIBRATION_MODE,
+	 *      &auto_calibration_mode, sizeof(auto_calibration_mode));
+	*/
+
+	mutex_unlock(&priv->conf_mutex);
+	return ret;
+}
+
+void cw1200_remove_interface(struct ieee80211_hw *dev,
+			     struct ieee80211_vif *vif)
+{
+	struct cw1200_common *priv = dev->priv;
+	struct wsm_reset reset = {
+		.reset_statistics = true,
+	};
+	int i;
+
+	mutex_lock(&priv->conf_mutex);
+	switch (priv->join_status) {
+	case CW1200_JOIN_STATUS_JOINING:
+	case CW1200_JOIN_STATUS_PRE_STA:
+	case CW1200_JOIN_STATUS_STA:
+	case CW1200_JOIN_STATUS_IBSS:
+		wsm_lock_tx(priv);
+		if (queue_work(priv->workqueue, &priv->unjoin_work) <= 0)
+			wsm_unlock_tx(priv);
+		break;
+	case CW1200_JOIN_STATUS_AP:
+		for (i = 0; priv->link_id_map; ++i) {
+			if (priv->link_id_map & BIT(i)) {
+				reset.link_id = i;
+				wsm_reset(priv, &reset);
+				priv->link_id_map &= ~BIT(i);
+			}
+		}
+		memset(priv->link_id_db, 0, sizeof(priv->link_id_db));
+		priv->sta_asleep_mask = 0;
+		priv->enable_beacon = false;
+		priv->tx_multicast = false;
+		priv->aid0_bit_set = false;
+		priv->buffered_multicasts = false;
+		priv->pspoll_mask = 0;
+		reset.link_id = 0;
+		wsm_reset(priv, &reset);
+		break;
+	case CW1200_JOIN_STATUS_MONITOR:
+		cw1200_update_listening(priv, false);
+		break;
+	default:
+		break;
+	}
+	priv->vif = NULL;
+	priv->mode = NL80211_IFTYPE_MONITOR;
+	memset(priv->mac_addr, 0, ETH_ALEN);
+	memset(&priv->p2p_ps_modeinfo, 0, sizeof(priv->p2p_ps_modeinfo));
+	cw1200_free_keys(priv);
+	cw1200_setup_mac(priv);
+	priv->listening = false;
+	priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+	if (!__cw1200_flush(priv, true))
+		wsm_unlock_tx(priv);
+
+	mutex_unlock(&priv->conf_mutex);
+}
+
+int cw1200_change_interface(struct ieee80211_hw *dev,
+			    struct ieee80211_vif *vif,
+			    enum nl80211_iftype new_type,
+			    bool p2p)
+{
+	int ret = 0;
+	pr_debug("change_interface new: %d (%d), old: %d (%d)\n", new_type,
+		 p2p, vif->type, vif->p2p);
+
+	if (new_type != vif->type || vif->p2p != p2p) {
+		cw1200_remove_interface(dev, vif);
+		vif->type = new_type;
+		vif->p2p = p2p;
+		ret = cw1200_add_interface(dev, vif);
+	}
+
+	return ret;
+}
+
+int cw1200_config(struct ieee80211_hw *dev, u32 changed)
+{
+	int ret = 0;
+	struct cw1200_common *priv = dev->priv;
+	struct ieee80211_conf *conf = &dev->conf;
+
+	pr_debug("CONFIG CHANGED:  %08x\n", changed);
+
+	down(&priv->scan.lock);
+	mutex_lock(&priv->conf_mutex);
+	/* TODO: IEEE80211_CONF_CHANGE_QOS */
+	/* TODO: IEEE80211_CONF_CHANGE_LISTEN_INTERVAL */
+
+	if (changed & IEEE80211_CONF_CHANGE_POWER) {
+		priv->output_power = conf->power_level;
+		pr_debug("[STA] TX power: %d\n", priv->output_power);
+		wsm_set_output_power(priv, priv->output_power * 10);
+	}
+
+	if ((changed & IEEE80211_CONF_CHANGE_CHANNEL) &&
+	    (priv->channel != conf->chandef.chan)) {
+		struct ieee80211_channel *ch = conf->chandef.chan;
+		struct wsm_switch_channel channel = {
+			.channel_number = ch->hw_value,
+		};
+		pr_debug("[STA] Freq %d (wsm ch: %d).\n",
+			 ch->center_freq, ch->hw_value);
+
+		/* __cw1200_flush() implicitly locks tx, if successful */
+		if (!__cw1200_flush(priv, false)) {
+			if (!wsm_switch_channel(priv, &channel)) {
+				ret = wait_event_timeout(priv->channel_switch_done,
+							 !priv->channel_switch_in_progress,
+							 3 * HZ);
+				if (ret) {
+					/* Already unlocks if successful */
+					priv->channel = ch;
+					ret = 0;
+				} else {
+					ret = -ETIMEDOUT;
+				}
+			} else {
+				/* Unlock if switch channel fails */
+				wsm_unlock_tx(priv);
+			}
+		}
+	}
+
+	if (changed & IEEE80211_CONF_CHANGE_PS) {
+		if (!(conf->flags & IEEE80211_CONF_PS))
+			priv->powersave_mode.mode = WSM_PSM_ACTIVE;
+		else if (conf->dynamic_ps_timeout <= 0)
+			priv->powersave_mode.mode = WSM_PSM_PS;
+		else
+			priv->powersave_mode.mode = WSM_PSM_FAST_PS;
+
+		/* Firmware requires that value for this 1-byte field must
+		 * be specified in units of 500us. Values above the 128ms
+		 * threshold are not supported.
+		 */
+		if (conf->dynamic_ps_timeout >= 0x80)
+			priv->powersave_mode.fast_psm_idle_period = 0xFF;
+		else
+			priv->powersave_mode.fast_psm_idle_period =
+					conf->dynamic_ps_timeout << 1;
+
+		if (priv->join_status == CW1200_JOIN_STATUS_STA &&
+		    priv->bss_params.aid)
+			cw1200_set_pm(priv, &priv->powersave_mode);
+	}
+
+	if (changed & IEEE80211_CONF_CHANGE_MONITOR) {
+		/* TBD: It looks like it's transparent
+		 * there's a monitor interface present -- use this
+		 * to determine for example whether to calculate
+		 * timestamps for packets or not, do not use instead
+		 * of filter flags!
+		 */
+	}
+
+	if (changed & IEEE80211_CONF_CHANGE_IDLE) {
+		struct wsm_operational_mode mode = {
+			.power_mode = cw1200_power_mode,
+			.disable_more_flag_usage = true,
+		};
+
+		wsm_lock_tx(priv);
+		/* Disable p2p-dev mode forced by TX request */
+		if ((priv->join_status == CW1200_JOIN_STATUS_MONITOR) &&
+		    (conf->flags & IEEE80211_CONF_IDLE) &&
+		    !priv->listening) {
+			cw1200_disable_listening(priv);
+			priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+		}
+		wsm_set_operational_mode(priv, &mode);
+		wsm_unlock_tx(priv);
+	}
+
+	if (changed & IEEE80211_CONF_CHANGE_RETRY_LIMITS) {
+		pr_debug("[STA] Retry limits: %d (long), %d (short).\n",
+			 conf->long_frame_max_tx_count,
+			 conf->short_frame_max_tx_count);
+		spin_lock_bh(&priv->tx_policy_cache.lock);
+		priv->long_frame_max_tx_count = conf->long_frame_max_tx_count;
+		priv->short_frame_max_tx_count =
+			(conf->short_frame_max_tx_count < 0x0F) ?
+			conf->short_frame_max_tx_count : 0x0F;
+		priv->hw->max_rate_tries = priv->short_frame_max_tx_count;
+		spin_unlock_bh(&priv->tx_policy_cache.lock);
+	}
+	mutex_unlock(&priv->conf_mutex);
+	up(&priv->scan.lock);
+	return ret;
+}
+
+void cw1200_update_filtering(struct cw1200_common *priv)
+{
+	int ret;
+	bool bssid_filtering = !priv->rx_filter.bssid;
+	bool is_p2p = priv->vif && priv->vif->p2p;
+	bool is_sta = priv->vif && NL80211_IFTYPE_STATION == priv->vif->type;
+
+	static struct wsm_beacon_filter_control bf_ctrl;
+	static struct wsm_mib_beacon_filter_table bf_tbl = {
+		.entry[0].ie_id = WLAN_EID_VENDOR_SPECIFIC,
+		.entry[0].flags = WSM_BEACON_FILTER_IE_HAS_CHANGED |
+					WSM_BEACON_FILTER_IE_NO_LONGER_PRESENT |
+					WSM_BEACON_FILTER_IE_HAS_APPEARED,
+		.entry[0].oui[0] = 0x50,
+		.entry[0].oui[1] = 0x6F,
+		.entry[0].oui[2] = 0x9A,
+		.entry[1].ie_id = WLAN_EID_HT_OPERATION,
+		.entry[1].flags = WSM_BEACON_FILTER_IE_HAS_CHANGED |
+					WSM_BEACON_FILTER_IE_NO_LONGER_PRESENT |
+					WSM_BEACON_FILTER_IE_HAS_APPEARED,
+		.entry[2].ie_id = WLAN_EID_ERP_INFO,
+		.entry[2].flags = WSM_BEACON_FILTER_IE_HAS_CHANGED |
+					WSM_BEACON_FILTER_IE_NO_LONGER_PRESENT |
+					WSM_BEACON_FILTER_IE_HAS_APPEARED,
+	};
+
+	if (priv->join_status == CW1200_JOIN_STATUS_PASSIVE)
+		return;
+	else if (priv->join_status == CW1200_JOIN_STATUS_MONITOR)
+		bssid_filtering = false;
+
+	if (priv->disable_beacon_filter) {
+		bf_ctrl.enabled = 0;
+		bf_ctrl.bcn_count = 1;
+		bf_tbl.num = __cpu_to_le32(0);
+	} else if (is_p2p || !is_sta) {
+		bf_ctrl.enabled = WSM_BEACON_FILTER_ENABLE |
+			WSM_BEACON_FILTER_AUTO_ERP;
+		bf_ctrl.bcn_count = 0;
+		bf_tbl.num = __cpu_to_le32(2);
+	} else {
+		bf_ctrl.enabled = WSM_BEACON_FILTER_ENABLE;
+		bf_ctrl.bcn_count = 0;
+		bf_tbl.num = __cpu_to_le32(3);
+	}
+
+	/* When acting as p2p client being connected to p2p GO, in order to
+	 * receive frames from a different p2p device, turn off bssid filter.
+	 *
+	 * WARNING: FW dependency!
+	 * This can only be used with FW WSM371 and its successors.
+	 * In that FW version even with bssid filter turned off,
+	 * device will block most of the unwanted frames.
+	 */
+	if (is_p2p)
+		bssid_filtering = false;
+
+	ret = wsm_set_rx_filter(priv, &priv->rx_filter);
+	if (!ret)
+		ret = wsm_set_beacon_filter_table(priv, &bf_tbl);
+	if (!ret)
+		ret = wsm_beacon_filter_control(priv, &bf_ctrl);
+	if (!ret)
+		ret = wsm_set_bssid_filtering(priv, bssid_filtering);
+	if (!ret)
+		ret = wsm_set_multicast_filter(priv, &priv->multicast_filter);
+	if (ret)
+		wiphy_err(priv->hw->wiphy,
+			  "Update filtering failed: %d.\n", ret);
+	return;
+}
+
+void cw1200_update_filtering_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common,
+			     update_filtering_work);
+
+	cw1200_update_filtering(priv);
+}
+
+void cw1200_set_beacon_wakeup_period_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common,
+			     set_beacon_wakeup_period_work);
+
+	wsm_set_beacon_wakeup_period(priv,
+				     priv->beacon_int * priv->join_dtim_period >
+				     MAX_BEACON_SKIP_TIME_MS ? 1 :
+				     priv->join_dtim_period, 0);
+}
+
+u64 cw1200_prepare_multicast(struct ieee80211_hw *hw,
+			     struct netdev_hw_addr_list *mc_list)
+{
+	static u8 broadcast_ipv6[ETH_ALEN] = {
+		0x33, 0x33, 0x00, 0x00, 0x00, 0x01
+	};
+	static u8 broadcast_ipv4[ETH_ALEN] = {
+		0x01, 0x00, 0x5e, 0x00, 0x00, 0x01
+	};
+	struct cw1200_common *priv = hw->priv;
+	struct netdev_hw_addr *ha;
+	int count = 0;
+
+	/* Disable multicast filtering */
+	priv->has_multicast_subscription = false;
+	memset(&priv->multicast_filter, 0x00, sizeof(priv->multicast_filter));
+
+	if (netdev_hw_addr_list_count(mc_list) > WSM_MAX_GRP_ADDRTABLE_ENTRIES)
+		return 0;
+
+	/* Enable if requested */
+	netdev_hw_addr_list_for_each(ha, mc_list) {
+		pr_debug("[STA] multicast: %pM\n", ha->addr);
+		memcpy(&priv->multicast_filter.macaddrs[count],
+		       ha->addr, ETH_ALEN);
+		if (memcmp(ha->addr, broadcast_ipv4, ETH_ALEN) &&
+		    memcmp(ha->addr, broadcast_ipv6, ETH_ALEN))
+			priv->has_multicast_subscription = true;
+		count++;
+	}
+
+	if (count) {
+		priv->multicast_filter.enable = __cpu_to_le32(1);
+		priv->multicast_filter.num_addrs = __cpu_to_le32(count);
+	}
+
+	return netdev_hw_addr_list_count(mc_list);
+}
+
+void cw1200_configure_filter(struct ieee80211_hw *dev,
+			     unsigned int changed_flags,
+			     unsigned int *total_flags,
+			     u64 multicast)
+{
+	struct cw1200_common *priv = dev->priv;
+	bool listening = !!(*total_flags &
+			    (FIF_PROMISC_IN_BSS |
+			     FIF_OTHER_BSS |
+			     FIF_BCN_PRBRESP_PROMISC |
+			     FIF_PROBE_REQ));
+
+	*total_flags &= FIF_PROMISC_IN_BSS |
+			FIF_OTHER_BSS |
+			FIF_FCSFAIL |
+			FIF_BCN_PRBRESP_PROMISC |
+			FIF_PROBE_REQ;
+
+	down(&priv->scan.lock);
+	mutex_lock(&priv->conf_mutex);
+
+	priv->rx_filter.promiscuous = (*total_flags & FIF_PROMISC_IN_BSS)
+			? 1 : 0;
+	priv->rx_filter.bssid = (*total_flags & (FIF_OTHER_BSS |
+			FIF_PROBE_REQ)) ? 1 : 0;
+	priv->rx_filter.fcs = (*total_flags & FIF_FCSFAIL) ? 1 : 0;
+	priv->disable_beacon_filter = !(*total_flags &
+					(FIF_BCN_PRBRESP_PROMISC |
+					 FIF_PROMISC_IN_BSS |
+					 FIF_PROBE_REQ));
+	if (priv->listening != listening) {
+		priv->listening = listening;
+		wsm_lock_tx(priv);
+		cw1200_update_listening(priv, listening);
+		wsm_unlock_tx(priv);
+	}
+	cw1200_update_filtering(priv);
+	mutex_unlock(&priv->conf_mutex);
+	up(&priv->scan.lock);
+}
+
+int cw1200_conf_tx(struct ieee80211_hw *dev, struct ieee80211_vif *vif,
+		   u16 queue, const struct ieee80211_tx_queue_params *params)
+{
+	struct cw1200_common *priv = dev->priv;
+	int ret = 0;
+	/* To prevent re-applying PM request OID again and again*/
+	bool old_uapsd_flags;
+
+	mutex_lock(&priv->conf_mutex);
+
+	if (queue < dev->queues) {
+		old_uapsd_flags = le16_to_cpu(priv->uapsd_info.uapsd_flags);
+
+		WSM_TX_QUEUE_SET(&priv->tx_queue_params, queue, 0, 0, 0);
+		ret = wsm_set_tx_queue_params(priv,
+					      &priv->tx_queue_params.params[queue], queue);
+		if (ret) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		WSM_EDCA_SET(&priv->edca, queue, params->aifs,
+			     params->cw_min, params->cw_max,
+			     params->txop, 0xc8,
+			     params->uapsd);
+		ret = wsm_set_edca_params(priv, &priv->edca);
+		if (ret) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		if (priv->mode == NL80211_IFTYPE_STATION) {
+			ret = cw1200_set_uapsd_param(priv, &priv->edca);
+			if (!ret && priv->setbssparams_done &&
+			    (priv->join_status == CW1200_JOIN_STATUS_STA) &&
+			    (old_uapsd_flags != le16_to_cpu(priv->uapsd_info.uapsd_flags)))
+				ret = cw1200_set_pm(priv, &priv->powersave_mode);
+		}
+	} else {
+		ret = -EINVAL;
+	}
+
+out:
+	mutex_unlock(&priv->conf_mutex);
+	return ret;
+}
+
+int cw1200_get_stats(struct ieee80211_hw *dev,
+		     struct ieee80211_low_level_stats *stats)
+{
+	struct cw1200_common *priv = dev->priv;
+
+	memcpy(stats, &priv->stats, sizeof(*stats));
+	return 0;
+}
+
+int cw1200_set_pm(struct cw1200_common *priv, const struct wsm_set_pm *arg)
+{
+	struct wsm_set_pm pm = *arg;
+
+	if (priv->uapsd_info.uapsd_flags != 0)
+		pm.mode &= ~WSM_PSM_FAST_PS_FLAG;
+
+	if (memcmp(&pm, &priv->firmware_ps_mode,
+		   sizeof(struct wsm_set_pm))) {
+		priv->firmware_ps_mode = pm;
+		return wsm_set_pm(priv, &pm);
+	} else {
+		return 0;
+	}
+}
+
+int cw1200_set_key(struct ieee80211_hw *dev, enum set_key_cmd cmd,
+		   struct ieee80211_vif *vif, struct ieee80211_sta *sta,
+		   struct ieee80211_key_conf *key)
+{
+	int ret = -EOPNOTSUPP;
+	struct cw1200_common *priv = dev->priv;
+	struct ieee80211_key_seq seq;
+
+	mutex_lock(&priv->conf_mutex);
+
+	if (cmd == SET_KEY) {
+		u8 *peer_addr = NULL;
+		int pairwise = (key->flags & IEEE80211_KEY_FLAG_PAIRWISE) ?
+			1 : 0;
+		int idx = cw1200_alloc_key(priv);
+		struct wsm_add_key *wsm_key = &priv->keys[idx];
+
+		if (idx < 0) {
+			ret = -EINVAL;
+			goto finally;
+		}
+
+		if (sta)
+			peer_addr = sta->addr;
+
+		key->flags |= IEEE80211_KEY_FLAG_PUT_IV_SPACE;
+
+		switch (key->cipher) {
+		case WLAN_CIPHER_SUITE_WEP40:
+		case WLAN_CIPHER_SUITE_WEP104:
+			if (key->keylen > 16) {
+				cw1200_free_key(priv, idx);
+				ret = -EINVAL;
+				goto finally;
+			}
+
+			if (pairwise) {
+				wsm_key->type = WSM_KEY_TYPE_WEP_PAIRWISE;
+				memcpy(wsm_key->wep_pairwise.peer,
+				       peer_addr, ETH_ALEN);
+				memcpy(wsm_key->wep_pairwise.keydata,
+				       &key->key[0], key->keylen);
+				wsm_key->wep_pairwise.keylen = key->keylen;
+			} else {
+				wsm_key->type = WSM_KEY_TYPE_WEP_DEFAULT;
+				memcpy(wsm_key->wep_group.keydata,
+				       &key->key[0], key->keylen);
+				wsm_key->wep_group.keylen = key->keylen;
+				wsm_key->wep_group.keyid = key->keyidx;
+			}
+			break;
+		case WLAN_CIPHER_SUITE_TKIP:
+			ieee80211_get_key_rx_seq(key, 0, &seq);
+			if (pairwise) {
+				wsm_key->type = WSM_KEY_TYPE_TKIP_PAIRWISE;
+				memcpy(wsm_key->tkip_pairwise.peer,
+				       peer_addr, ETH_ALEN);
+				memcpy(wsm_key->tkip_pairwise.keydata,
+				       &key->key[0], 16);
+				memcpy(wsm_key->tkip_pairwise.tx_mic_key,
+				       &key->key[16], 8);
+				memcpy(wsm_key->tkip_pairwise.rx_mic_key,
+				       &key->key[24], 8);
+			} else {
+				size_t mic_offset =
+					(priv->mode == NL80211_IFTYPE_AP) ?
+					16 : 24;
+				wsm_key->type = WSM_KEY_TYPE_TKIP_GROUP;
+				memcpy(wsm_key->tkip_group.keydata,
+				       &key->key[0], 16);
+				memcpy(wsm_key->tkip_group.rx_mic_key,
+				       &key->key[mic_offset], 8);
+
+				wsm_key->tkip_group.rx_seqnum[0] = seq.tkip.iv16 & 0xff;
+				wsm_key->tkip_group.rx_seqnum[1] = (seq.tkip.iv16 >> 8) & 0xff;
+				wsm_key->tkip_group.rx_seqnum[2] = seq.tkip.iv32 & 0xff;
+				wsm_key->tkip_group.rx_seqnum[3] = (seq.tkip.iv32 >> 8) & 0xff;
+				wsm_key->tkip_group.rx_seqnum[4] = (seq.tkip.iv32 >> 16) & 0xff;
+				wsm_key->tkip_group.rx_seqnum[5] = (seq.tkip.iv32 >> 24) & 0xff;
+				wsm_key->tkip_group.rx_seqnum[6] = 0;
+				wsm_key->tkip_group.rx_seqnum[7] = 0;
+
+				wsm_key->tkip_group.keyid = key->keyidx;
+			}
+			break;
+		case WLAN_CIPHER_SUITE_CCMP:
+			ieee80211_get_key_rx_seq(key, 0, &seq);
+			if (pairwise) {
+				wsm_key->type = WSM_KEY_TYPE_AES_PAIRWISE;
+				memcpy(wsm_key->aes_pairwise.peer,
+				       peer_addr, ETH_ALEN);
+				memcpy(wsm_key->aes_pairwise.keydata,
+				       &key->key[0], 16);
+			} else {
+				wsm_key->type = WSM_KEY_TYPE_AES_GROUP;
+				memcpy(wsm_key->aes_group.keydata,
+				       &key->key[0], 16);
+
+				wsm_key->aes_group.rx_seqnum[0] = seq.ccmp.pn[5];
+				wsm_key->aes_group.rx_seqnum[1] = seq.ccmp.pn[4];
+				wsm_key->aes_group.rx_seqnum[2] = seq.ccmp.pn[3];
+				wsm_key->aes_group.rx_seqnum[3] = seq.ccmp.pn[2];
+				wsm_key->aes_group.rx_seqnum[4] = seq.ccmp.pn[1];
+				wsm_key->aes_group.rx_seqnum[5] = seq.ccmp.pn[0];
+				wsm_key->aes_group.rx_seqnum[6] = 0;
+				wsm_key->aes_group.rx_seqnum[7] = 0;
+				wsm_key->aes_group.keyid = key->keyidx;
+			}
+			break;
+		case WLAN_CIPHER_SUITE_SMS4:
+			if (pairwise) {
+				wsm_key->type = WSM_KEY_TYPE_WAPI_PAIRWISE;
+				memcpy(wsm_key->wapi_pairwise.peer,
+				       peer_addr, ETH_ALEN);
+				memcpy(wsm_key->wapi_pairwise.keydata,
+				       &key->key[0], 16);
+				memcpy(wsm_key->wapi_pairwise.mic_key,
+				       &key->key[16], 16);
+				wsm_key->wapi_pairwise.keyid = key->keyidx;
+			} else {
+				wsm_key->type = WSM_KEY_TYPE_WAPI_GROUP;
+				memcpy(wsm_key->wapi_group.keydata,
+				       &key->key[0],  16);
+				memcpy(wsm_key->wapi_group.mic_key,
+				       &key->key[16], 16);
+				wsm_key->wapi_group.keyid = key->keyidx;
+			}
+			break;
+		default:
+			pr_warn("Unhandled key type %d\n", key->cipher);
+			cw1200_free_key(priv, idx);
+			ret = -EOPNOTSUPP;
+			goto finally;
+		}
+		ret = wsm_add_key(priv, wsm_key);
+		if (!ret)
+			key->hw_key_idx = idx;
+		else
+			cw1200_free_key(priv, idx);
+	} else if (cmd == DISABLE_KEY) {
+		struct wsm_remove_key wsm_key = {
+			.index = key->hw_key_idx,
+		};
+
+		if (wsm_key.index > WSM_KEY_MAX_INDEX) {
+			ret = -EINVAL;
+			goto finally;
+		}
+
+		cw1200_free_key(priv, wsm_key.index);
+		ret = wsm_remove_key(priv, &wsm_key);
+	} else {
+		pr_warn("Unhandled key command %d\n", cmd);
+	}
+
+finally:
+	mutex_unlock(&priv->conf_mutex);
+	return ret;
+}
+
+void cw1200_wep_key_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, wep_key_work);
+	u8 queue_id = cw1200_queue_get_queue_id(priv->pending_frame_id);
+	struct cw1200_queue *queue = &priv->tx_queue[queue_id];
+	__le32 wep_default_key_id = __cpu_to_le32(
+		priv->wep_default_key_id);
+
+	pr_debug("[STA] Setting default WEP key: %d\n",
+		 priv->wep_default_key_id);
+	wsm_flush_tx(priv);
+	wsm_write_mib(priv, WSM_MIB_ID_DOT11_WEP_DEFAULT_KEY_ID,
+		      &wep_default_key_id, sizeof(wep_default_key_id));
+	cw1200_queue_requeue(queue, priv->pending_frame_id);
+	wsm_unlock_tx(priv);
+}
+
+int cw1200_set_rts_threshold(struct ieee80211_hw *hw, u32 value)
+{
+	int ret = 0;
+	__le32 val32;
+	struct cw1200_common *priv = hw->priv;
+
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED)
+		return 0;
+
+	if (value != (u32) -1)
+		val32 = __cpu_to_le32(value);
+	else
+		val32 = 0; /* disabled */
+
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED) {
+		/* device is down, can _not_ set threshold */
+		ret = -ENODEV;
+		goto out;
+	}
+
+	if (priv->rts_threshold == value)
+		goto out;
+
+	pr_debug("[STA] Setting RTS threshold: %d\n",
+		 priv->rts_threshold);
+
+	/* mutex_lock(&priv->conf_mutex); */
+	ret = wsm_write_mib(priv, WSM_MIB_ID_DOT11_RTS_THRESHOLD,
+			    &val32, sizeof(val32));
+	if (!ret)
+		priv->rts_threshold = value;
+	/* mutex_unlock(&priv->conf_mutex); */
+
+out:
+	return ret;
+}
+
+/* If successful, LOCKS the TX queue! */
+static int __cw1200_flush(struct cw1200_common *priv, bool drop)
+{
+	int i, ret;
+
+	for (;;) {
+		/* TODO: correct flush handling is required when dev_stop.
+		 * Temporary workaround: 2s
+		 */
+		if (drop) {
+			for (i = 0; i < 4; ++i)
+				cw1200_queue_clear(&priv->tx_queue[i]);
+		} else {
+			ret = wait_event_timeout(
+				priv->tx_queue_stats.wait_link_id_empty,
+				cw1200_queue_stats_is_empty(
+					&priv->tx_queue_stats, -1),
+				2 * HZ);
+		}
+
+		if (!drop && ret <= 0) {
+			ret = -ETIMEDOUT;
+			break;
+		} else {
+			ret = 0;
+		}
+
+		wsm_lock_tx(priv);
+		if (!cw1200_queue_stats_is_empty(&priv->tx_queue_stats, -1)) {
+			/* Highly unlikely: WSM requeued frames. */
+			wsm_unlock_tx(priv);
+			continue;
+		}
+		break;
+	}
+	return ret;
+}
+
+void cw1200_flush(struct ieee80211_hw *hw, u32 queues, bool drop)
+{
+	struct cw1200_common *priv = hw->priv;
+
+	switch (priv->mode) {
+	case NL80211_IFTYPE_MONITOR:
+		drop = true;
+		break;
+	case NL80211_IFTYPE_AP:
+		if (!priv->enable_beacon)
+			drop = true;
+		break;
+	}
+
+	if (!__cw1200_flush(priv, drop))
+		wsm_unlock_tx(priv);
+
+	return;
+}
+
+/* ******************************************************************** */
+/* WSM callbacks							*/
+
+void cw1200_free_event_queue(struct cw1200_common *priv)
+{
+	LIST_HEAD(list);
+
+	spin_lock(&priv->event_queue_lock);
+	list_splice_init(&priv->event_queue, &list);
+	spin_unlock(&priv->event_queue_lock);
+
+	__cw1200_free_event_queue(&list);
+}
+
+void cw1200_event_handler(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, event_handler);
+	struct cw1200_wsm_event *event;
+	LIST_HEAD(list);
+
+	spin_lock(&priv->event_queue_lock);
+	list_splice_init(&priv->event_queue, &list);
+	spin_unlock(&priv->event_queue_lock);
+
+	list_for_each_entry(event, &list, link) {
+		switch (event->evt.id) {
+		case WSM_EVENT_ERROR:
+			pr_err("Unhandled WSM Error from LMAC\n");
+			break;
+		case WSM_EVENT_BSS_LOST:
+			pr_debug("[CQM] BSS lost.\n");
+			cancel_work_sync(&priv->unjoin_work);
+			if (!down_trylock(&priv->scan.lock)) {
+				cw1200_cqm_bssloss_sm(priv, 1, 0, 0);
+				up(&priv->scan.lock);
+			} else {
+				/* Scan is in progress. Delay reporting.
+				 * Scan complete will trigger bss_loss_work
+				 */
+				priv->delayed_link_loss = 1;
+				/* Also start a watchdog. */
+				queue_delayed_work(priv->workqueue,
+						   &priv->bss_loss_work, 5*HZ);
+			}
+			break;
+		case WSM_EVENT_BSS_REGAINED:
+			pr_debug("[CQM] BSS regained.\n");
+			cw1200_cqm_bssloss_sm(priv, 0, 0, 0);
+			cancel_work_sync(&priv->unjoin_work);
+			break;
+		case WSM_EVENT_RADAR_DETECTED:
+			wiphy_info(priv->hw->wiphy, "radar pulse detected\n");
+			break;
+		case WSM_EVENT_RCPI_RSSI:
+		{
+			/* RSSI: signed Q8.0, RCPI: unsigned Q7.1
+			 * RSSI = RCPI / 2 - 110
+			 */
+			int rcpi_rssi = (int)(event->evt.data & 0xFF);
+			int cqm_evt;
+			if (priv->cqm_use_rssi)
+				rcpi_rssi = (s8)rcpi_rssi;
+			else
+				rcpi_rssi =  rcpi_rssi / 2 - 110;
+
+			cqm_evt = (rcpi_rssi <= priv->cqm_rssi_thold) ?
+				NL80211_CQM_RSSI_THRESHOLD_EVENT_LOW :
+				NL80211_CQM_RSSI_THRESHOLD_EVENT_HIGH;
+			pr_debug("[CQM] RSSI event: %d.\n", rcpi_rssi);
+			ieee80211_cqm_rssi_notify(priv->vif, cqm_evt,
+						  GFP_KERNEL);
+			break;
+		}
+		case WSM_EVENT_BT_INACTIVE:
+			pr_warn("Unhandled BT INACTIVE from LMAC\n");
+			break;
+		case WSM_EVENT_BT_ACTIVE:
+			pr_warn("Unhandled BT ACTIVE from LMAC\n");
+			break;
+		}
+	}
+	__cw1200_free_event_queue(&list);
+}
+
+void cw1200_bss_loss_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, bss_loss_work.work);
+
+	pr_debug("[CQM] Reporting connection loss.\n");
+	wsm_lock_tx(priv);
+	if (queue_work(priv->workqueue, &priv->unjoin_work) <= 0)
+		wsm_unlock_tx(priv);
+}
+
+void cw1200_bss_params_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, bss_params_work);
+	mutex_lock(&priv->conf_mutex);
+
+	priv->bss_params.reset_beacon_loss = 1;
+	wsm_set_bss_params(priv, &priv->bss_params);
+	priv->bss_params.reset_beacon_loss = 0;
+
+	mutex_unlock(&priv->conf_mutex);
+}
+
+/* ******************************************************************** */
+/* Internal API								*/
+
+/* This function is called to Parse the SDD file
+ * to extract listen_interval and PTA related information
+ * sdd is a TLV: u8 id, u8 len, u8 data[]
+ */
+static int cw1200_parse_sdd_file(struct cw1200_common *priv)
+{
+	const u8 *p = priv->sdd->data;
+	int ret = 0;
+
+	while (p + 2 <= priv->sdd->data + priv->sdd->size) {
+		if (p + p[1] + 2 > priv->sdd->data + priv->sdd->size) {
+			pr_warn("Malformed sdd structure\n");
+			return -1;
+		}
+		switch (p[0]) {
+		case SDD_PTA_CFG_ELT_ID: {
+			u16 v;
+			if (p[1] < 4) {
+				pr_warn("SDD_PTA_CFG_ELT_ID malformed\n");
+				ret = -1;
+				break;
+			}
+			v = le16_to_cpu(*((__le16 *)(p + 2)));
+			if (!v)  /* non-zero means this is enabled */
+				break;
+
+			v = le16_to_cpu(*((__le16 *)(p + 4)));
+			priv->conf_listen_interval = (v >> 7) & 0x1F;
+			pr_debug("PTA found; Listen Interval %d\n",
+				 priv->conf_listen_interval);
+			break;
+		}
+		case SDD_REFERENCE_FREQUENCY_ELT_ID: {
+			u16 clk = le16_to_cpu(*((__le16 *)(p + 2)));
+			if (clk != priv->hw_refclk)
+				pr_warn("SDD file doesn't match configured refclk (%d vs %d)\n",
+					clk, priv->hw_refclk);
+			break;
+		}
+		default:
+			break;
+		}
+		p += p[1] + 2;
+	}
+
+	if (!priv->bt_present) {
+		pr_debug("PTA element NOT found.\n");
+		priv->conf_listen_interval = 0;
+	}
+	return ret;
+}
+
+int cw1200_setup_mac(struct cw1200_common *priv)
+{
+	int ret = 0;
+
+	/* NOTE: There is a bug in FW: it reports signal
+	 * as RSSI if RSSI subscription is enabled.
+	 * It's not enough to set WSM_RCPI_RSSI_USE_RSSI.
+	 *
+	 * NOTE2: RSSI based reports have been switched to RCPI, since
+	 * FW has a bug and RSSI reported values are not stable,
+	 * what can leads to signal level oscilations in user-end applications
+	 */
+	struct wsm_rcpi_rssi_threshold threshold = {
+		.rssiRcpiMode = WSM_RCPI_RSSI_THRESHOLD_ENABLE |
+		WSM_RCPI_RSSI_DONT_USE_UPPER |
+		WSM_RCPI_RSSI_DONT_USE_LOWER,
+		.rollingAverageCount = 16,
+	};
+
+	struct wsm_configuration cfg = {
+		.dot11StationId = &priv->mac_addr[0],
+	};
+
+	/* Remember the decission here to make sure, we will handle
+	 * the RCPI/RSSI value correctly on WSM_EVENT_RCPI_RSS
+	 */
+	if (threshold.rssiRcpiMode & WSM_RCPI_RSSI_USE_RSSI)
+		priv->cqm_use_rssi = true;
+
+	if (!priv->sdd) {
+		ret = request_firmware(&priv->sdd, priv->sdd_path, priv->pdev);
+		if (ret) {
+			pr_err("Can't load sdd file %s.\n", priv->sdd_path);
+			return ret;
+		}
+		cw1200_parse_sdd_file(priv);
+	}
+
+	cfg.dpdData = priv->sdd->data;
+	cfg.dpdData_size = priv->sdd->size;
+	ret = wsm_configuration(priv, &cfg);
+	if (ret)
+		return ret;
+
+	/* Configure RSSI/SCPI reporting as RSSI. */
+	wsm_set_rcpi_rssi_threshold(priv, &threshold);
+
+	return 0;
+}
+
+static void cw1200_join_complete(struct cw1200_common *priv)
+{
+	pr_debug("[STA] Join complete (%d)\n", priv->join_complete_status);
+
+	priv->join_pending = false;
+	if (priv->join_complete_status) {
+		priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+		cw1200_update_listening(priv, priv->listening);
+		cw1200_do_unjoin(priv);
+		ieee80211_connection_loss(priv->vif);
+	} else {
+		if (priv->mode == NL80211_IFTYPE_ADHOC)
+			priv->join_status = CW1200_JOIN_STATUS_IBSS;
+		else
+			priv->join_status = CW1200_JOIN_STATUS_PRE_STA;
+	}
+	wsm_unlock_tx(priv); /* Clearing the lock held before do_join() */
+}
+
+void cw1200_join_complete_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, join_complete_work);
+	mutex_lock(&priv->conf_mutex);
+	cw1200_join_complete(priv);
+	mutex_unlock(&priv->conf_mutex);
+}
+
+void cw1200_join_complete_cb(struct cw1200_common *priv,
+			     struct wsm_join_complete *arg)
+{
+	pr_debug("[STA] cw1200_join_complete_cb called, status=%d.\n",
+		 arg->status);
+
+	if (cancel_delayed_work(&priv->join_timeout)) {
+		priv->join_complete_status = arg->status;
+		queue_work(priv->workqueue, &priv->join_complete_work);
+	}
+}
+
+/* MUST be called with tx_lock held!  It will be unlocked for us. */
+static void cw1200_do_join(struct cw1200_common *priv)
+{
+	const u8 *bssid;
+	struct ieee80211_bss_conf *conf = &priv->vif->bss_conf;
+	struct cfg80211_bss *bss = NULL;
+	struct wsm_protected_mgmt_policy mgmt_policy;
+	struct wsm_join join = {
+		.mode = conf->ibss_joined ?
+				WSM_JOIN_MODE_IBSS : WSM_JOIN_MODE_BSS,
+		.preamble_type = WSM_JOIN_PREAMBLE_LONG,
+		.probe_for_join = 1,
+		.atim_window = 0,
+		.basic_rate_set = cw1200_rate_mask_to_wsm(priv,
+							  conf->basic_rates),
+	};
+	if (delayed_work_pending(&priv->join_timeout)) {
+		pr_warn("[STA] - Join request already pending, skipping..\n");
+		wsm_unlock_tx(priv);
+		return;
+	}
+
+	if (priv->join_status)
+		cw1200_do_unjoin(priv);
+
+	bssid = priv->vif->bss_conf.bssid;
+
+	bss = cfg80211_get_bss(priv->hw->wiphy, priv->channel,
+			bssid, NULL, 0, 0, 0);
+
+	if (!bss && !conf->ibss_joined) {
+		wsm_unlock_tx(priv);
+		return;
+	}
+
+	mutex_lock(&priv->conf_mutex);
+
+	/* Under the conf lock: check scan status and
+	 * bail out if it is in progress.
+	 */
+	if (atomic_read(&priv->scan.in_progress)) {
+		wsm_unlock_tx(priv);
+		goto done_put;
+	}
+
+	priv->join_pending = true;
+
+	/* Sanity check basic rates */
+	if (!join.basic_rate_set)
+		join.basic_rate_set = 7;
+
+	/* Sanity check beacon interval */
+	if (!priv->beacon_int)
+		priv->beacon_int = 1;
+
+	join.beacon_interval = priv->beacon_int;
+
+	/* BT Coex related changes */
+	if (priv->bt_present) {
+		if (((priv->conf_listen_interval * 100) %
+		     priv->beacon_int) == 0)
+			priv->listen_interval =
+				((priv->conf_listen_interval * 100) /
+				 priv->beacon_int);
+		else
+			priv->listen_interval =
+				((priv->conf_listen_interval * 100) /
+				 priv->beacon_int + 1);
+	}
+
+	if (priv->hw->conf.ps_dtim_period)
+		priv->join_dtim_period = priv->hw->conf.ps_dtim_period;
+	join.dtim_period = priv->join_dtim_period;
+
+	join.channel_number = priv->channel->hw_value;
+	join.band = (priv->channel->band == IEEE80211_BAND_5GHZ) ?
+		WSM_PHY_BAND_5G : WSM_PHY_BAND_2_4G;
+
+	memcpy(join.bssid, bssid, sizeof(join.bssid));
+
+	pr_debug("[STA] Join BSSID: %pM DTIM: %d, interval: %d\n",
+		 join.bssid,
+		 join.dtim_period, priv->beacon_int);
+
+	if (!conf->ibss_joined) {
+		const u8 *ssidie;
+		rcu_read_lock();
+		ssidie = ieee80211_bss_get_ie(bss, WLAN_EID_SSID);
+		if (ssidie) {
+			join.ssid_len = ssidie[1];
+			memcpy(join.ssid, &ssidie[2], join.ssid_len);
+		}
+		rcu_read_unlock();
+	}
+
+	if (priv->vif->p2p) {
+		join.flags |= WSM_JOIN_FLAGS_P2P_GO;
+		join.basic_rate_set =
+			cw1200_rate_mask_to_wsm(priv, 0xFF0);
+	}
+
+	/* Enable asynchronous join calls */
+	if (!conf->ibss_joined) {
+		join.flags |= WSM_JOIN_FLAGS_FORCE;
+		join.flags |= WSM_JOIN_FLAGS_FORCE_WITH_COMPLETE_IND;
+	}
+
+	wsm_flush_tx(priv);
+
+	/* Stay Awake for Join and Auth Timeouts and a bit more */
+	cw1200_pm_stay_awake(&priv->pm_state,
+			     CW1200_JOIN_TIMEOUT + CW1200_AUTH_TIMEOUT);
+
+	cw1200_update_listening(priv, false);
+
+	/* Turn on Block ACKs */
+	wsm_set_block_ack_policy(priv, priv->ba_tx_tid_mask,
+				 priv->ba_rx_tid_mask);
+
+	/* Set up timeout */
+	if (join.flags & WSM_JOIN_FLAGS_FORCE_WITH_COMPLETE_IND) {
+		priv->join_status = CW1200_JOIN_STATUS_JOINING;
+		queue_delayed_work(priv->workqueue,
+				   &priv->join_timeout,
+				   CW1200_JOIN_TIMEOUT);
+	}
+
+	/* 802.11w protected mgmt frames */
+	mgmt_policy.protectedMgmtEnable = 0;
+	mgmt_policy.unprotectedMgmtFramesAllowed = 1;
+	mgmt_policy.encryptionForAuthFrame = 1;
+	wsm_set_protected_mgmt_policy(priv, &mgmt_policy);
+
+	/* Perform actual join */
+	if (wsm_join(priv, &join)) {
+		pr_err("[STA] cw1200_join_work: wsm_join failed!\n");
+		cancel_delayed_work_sync(&priv->join_timeout);
+		cw1200_update_listening(priv, priv->listening);
+		/* Tx lock still held, unjoin will clear it. */
+		if (queue_work(priv->workqueue, &priv->unjoin_work) <= 0)
+			wsm_unlock_tx(priv);
+	} else {
+		if (!(join.flags & WSM_JOIN_FLAGS_FORCE_WITH_COMPLETE_IND))
+			cw1200_join_complete(priv); /* Will clear tx_lock */
+
+		/* Upload keys */
+		cw1200_upload_keys(priv);
+
+		/* Due to beacon filtering it is possible that the
+		 * AP's beacon is not known for the mac80211 stack.
+		 * Disable filtering temporary to make sure the stack
+		 * receives at least one
+		 */
+		priv->disable_beacon_filter = true;
+	}
+	cw1200_update_filtering(priv);
+
+done_put:
+	mutex_unlock(&priv->conf_mutex);
+	if (bss)
+		cfg80211_put_bss(priv->hw->wiphy, bss);
+}
+
+void cw1200_join_timeout(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, join_timeout.work);
+	pr_debug("[WSM] Join timed out.\n");
+	wsm_lock_tx(priv);
+	if (queue_work(priv->workqueue, &priv->unjoin_work) <= 0)
+		wsm_unlock_tx(priv);
+}
+
+static void cw1200_do_unjoin(struct cw1200_common *priv)
+{
+	struct wsm_reset reset = {
+		.reset_statistics = true,
+	};
+
+	cancel_delayed_work_sync(&priv->join_timeout);
+
+	mutex_lock(&priv->conf_mutex);
+	priv->join_pending = false;
+
+	if (atomic_read(&priv->scan.in_progress)) {
+		if (priv->delayed_unjoin)
+			wiphy_dbg(priv->hw->wiphy, "Delayed unjoin is already scheduled.\n");
+		else
+			priv->delayed_unjoin = true;
+		goto done;
+	}
+
+	priv->delayed_link_loss = false;
+
+	if (!priv->join_status)
+		goto done;
+
+	if (priv->join_status > CW1200_JOIN_STATUS_IBSS) {
+		wiphy_err(priv->hw->wiphy, "Unexpected: join status: %d\n",
+			  priv->join_status);
+		BUG_ON(1);
+	}
+
+	cancel_work_sync(&priv->update_filtering_work);
+	cancel_work_sync(&priv->set_beacon_wakeup_period_work);
+	priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+
+	/* Unjoin is a reset. */
+	wsm_flush_tx(priv);
+	wsm_keep_alive_period(priv, 0);
+	wsm_reset(priv, &reset);
+	wsm_set_output_power(priv, priv->output_power * 10);
+	priv->join_dtim_period = 0;
+	cw1200_setup_mac(priv);
+	cw1200_free_event_queue(priv);
+	cancel_work_sync(&priv->event_handler);
+	cw1200_update_listening(priv, priv->listening);
+	cw1200_cqm_bssloss_sm(priv, 0, 0, 0);
+
+	/* Disable Block ACKs */
+	wsm_set_block_ack_policy(priv, 0, 0);
+
+	priv->disable_beacon_filter = false;
+	cw1200_update_filtering(priv);
+	memset(&priv->association_mode, 0,
+	       sizeof(priv->association_mode));
+	memset(&priv->bss_params, 0, sizeof(priv->bss_params));
+	priv->setbssparams_done = false;
+	memset(&priv->firmware_ps_mode, 0,
+	       sizeof(priv->firmware_ps_mode));
+
+	pr_debug("[STA] Unjoin completed.\n");
+
+done:
+	mutex_unlock(&priv->conf_mutex);
+}
+
+void cw1200_unjoin_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, unjoin_work);
+
+	cw1200_do_unjoin(priv);
+
+	/* Tell the stack we're dead */
+	ieee80211_connection_loss(priv->vif);
+
+	wsm_unlock_tx(priv);
+}
+
+int cw1200_enable_listening(struct cw1200_common *priv)
+{
+	struct wsm_start start = {
+		.mode = WSM_START_MODE_P2P_DEV,
+		.band = WSM_PHY_BAND_2_4G,
+		.beacon_interval = 100,
+		.dtim_period = 1,
+		.probe_delay = 0,
+		.basic_rate_set = 0x0F,
+	};
+
+	if (priv->channel) {
+		start.band = priv->channel->band == IEEE80211_BAND_5GHZ ?
+			     WSM_PHY_BAND_5G : WSM_PHY_BAND_2_4G;
+		start.channel_number = priv->channel->hw_value;
+	} else {
+		start.band = WSM_PHY_BAND_2_4G;
+		start.channel_number = 1;
+	}
+
+	return wsm_start(priv, &start);
+}
+
+int cw1200_disable_listening(struct cw1200_common *priv)
+{
+	int ret;
+	struct wsm_reset reset = {
+		.reset_statistics = true,
+	};
+	ret = wsm_reset(priv, &reset);
+	return ret;
+}
+
+void cw1200_update_listening(struct cw1200_common *priv, bool enabled)
+{
+	if (enabled) {
+		if (priv->join_status == CW1200_JOIN_STATUS_PASSIVE) {
+			if (!cw1200_enable_listening(priv))
+				priv->join_status = CW1200_JOIN_STATUS_MONITOR;
+			wsm_set_probe_responder(priv, true);
+		}
+	} else {
+		if (priv->join_status == CW1200_JOIN_STATUS_MONITOR) {
+			if (!cw1200_disable_listening(priv))
+				priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+			wsm_set_probe_responder(priv, false);
+		}
+	}
+}
+
+int cw1200_set_uapsd_param(struct cw1200_common *priv,
+			   const struct wsm_edca_params *arg)
+{
+	int ret;
+	u16 uapsd_flags = 0;
+
+	/* Here's the mapping AC [queue, bit]
+	 *  VO [0,3], VI [1, 2], BE [2, 1], BK [3, 0]
+	 */
+
+	if (arg->uapsd_enable[0])
+		uapsd_flags |= 1 << 3;
+
+	if (arg->uapsd_enable[1])
+		uapsd_flags |= 1 << 2;
+
+	if (arg->uapsd_enable[2])
+		uapsd_flags |= 1 << 1;
+
+	if (arg->uapsd_enable[3])
+		uapsd_flags |= 1;
+
+	/* Currently pseudo U-APSD operation is not supported, so setting
+	 * MinAutoTriggerInterval, MaxAutoTriggerInterval and
+	 * AutoTriggerStep to 0
+	 */
+
+	priv->uapsd_info.uapsd_flags = cpu_to_le16(uapsd_flags);
+	priv->uapsd_info.min_auto_trigger_interval = 0;
+	priv->uapsd_info.max_auto_trigger_interval = 0;
+	priv->uapsd_info.auto_trigger_step = 0;
+
+	ret = wsm_set_uapsd_info(priv, &priv->uapsd_info);
+	return ret;
+}
+
+/* ******************************************************************** */
+/* AP API								*/
+
+int cw1200_sta_add(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
+		   struct ieee80211_sta *sta)
+{
+	struct cw1200_common *priv = hw->priv;
+	struct cw1200_sta_priv *sta_priv =
+			(struct cw1200_sta_priv *)&sta->drv_priv;
+	struct cw1200_link_entry *entry;
+	struct sk_buff *skb;
+
+	if (priv->mode != NL80211_IFTYPE_AP)
+		return 0;
+
+	sta_priv->link_id = cw1200_find_link_id(priv, sta->addr);
+	if (WARN_ON(!sta_priv->link_id)) {
+		wiphy_info(priv->hw->wiphy,
+			   "[AP] No more link IDs available.\n");
+		return -ENOENT;
+	}
+
+	entry = &priv->link_id_db[sta_priv->link_id - 1];
+	spin_lock_bh(&priv->ps_state_lock);
+	if ((sta->uapsd_queues & IEEE80211_WMM_IE_STA_QOSINFO_AC_MASK) ==
+					IEEE80211_WMM_IE_STA_QOSINFO_AC_MASK)
+		priv->sta_asleep_mask |= BIT(sta_priv->link_id);
+	entry->status = CW1200_LINK_HARD;
+	while ((skb = skb_dequeue(&entry->rx_queue)))
+		ieee80211_rx_irqsafe(priv->hw, skb);
+	spin_unlock_bh(&priv->ps_state_lock);
+	return 0;
+}
+
+int cw1200_sta_remove(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
+		      struct ieee80211_sta *sta)
+{
+	struct cw1200_common *priv = hw->priv;
+	struct cw1200_sta_priv *sta_priv =
+			(struct cw1200_sta_priv *)&sta->drv_priv;
+	struct cw1200_link_entry *entry;
+
+	if (priv->mode != NL80211_IFTYPE_AP || !sta_priv->link_id)
+		return 0;
+
+	entry = &priv->link_id_db[sta_priv->link_id - 1];
+	spin_lock_bh(&priv->ps_state_lock);
+	entry->status = CW1200_LINK_RESERVE;
+	entry->timestamp = jiffies;
+	wsm_lock_tx_async(priv);
+	if (queue_work(priv->workqueue, &priv->link_id_work) <= 0)
+		wsm_unlock_tx(priv);
+	spin_unlock_bh(&priv->ps_state_lock);
+	flush_workqueue(priv->workqueue);
+	return 0;
+}
+
+static void __cw1200_sta_notify(struct ieee80211_hw *dev,
+				struct ieee80211_vif *vif,
+				enum sta_notify_cmd notify_cmd,
+				int link_id)
+{
+	struct cw1200_common *priv = dev->priv;
+	u32 bit, prev;
+
+	/* Zero link id means "for all link IDs" */
+	if (link_id)
+		bit = BIT(link_id);
+	else if (WARN_ON_ONCE(notify_cmd != STA_NOTIFY_AWAKE))
+		bit = 0;
+	else
+		bit = priv->link_id_map;
+	prev = priv->sta_asleep_mask & bit;
+
+	switch (notify_cmd) {
+	case STA_NOTIFY_SLEEP:
+		if (!prev) {
+			if (priv->buffered_multicasts &&
+			    !priv->sta_asleep_mask)
+				queue_work(priv->workqueue,
+					   &priv->multicast_start_work);
+			priv->sta_asleep_mask |= bit;
+		}
+		break;
+	case STA_NOTIFY_AWAKE:
+		if (prev) {
+			priv->sta_asleep_mask &= ~bit;
+			priv->pspoll_mask &= ~bit;
+			if (priv->tx_multicast && link_id &&
+			    !priv->sta_asleep_mask)
+				queue_work(priv->workqueue,
+					   &priv->multicast_stop_work);
+			cw1200_bh_wakeup(priv);
+		}
+		break;
+	}
+}
+
+void cw1200_sta_notify(struct ieee80211_hw *dev,
+		       struct ieee80211_vif *vif,
+		       enum sta_notify_cmd notify_cmd,
+		       struct ieee80211_sta *sta)
+{
+	struct cw1200_common *priv = dev->priv;
+	struct cw1200_sta_priv *sta_priv =
+		(struct cw1200_sta_priv *)&sta->drv_priv;
+
+	spin_lock_bh(&priv->ps_state_lock);
+	__cw1200_sta_notify(dev, vif, notify_cmd, sta_priv->link_id);
+	spin_unlock_bh(&priv->ps_state_lock);
+}
+
+static void cw1200_ps_notify(struct cw1200_common *priv,
+		      int link_id, bool ps)
+{
+	if (link_id > CW1200_MAX_STA_IN_AP_MODE)
+		return;
+
+	pr_debug("%s for LinkId: %d. STAs asleep: %.8X\n",
+		 ps ? "Stop" : "Start",
+		 link_id, priv->sta_asleep_mask);
+
+	__cw1200_sta_notify(priv->hw, priv->vif,
+			    ps ? STA_NOTIFY_SLEEP : STA_NOTIFY_AWAKE, link_id);
+}
+
+static int cw1200_set_tim_impl(struct cw1200_common *priv, bool aid0_bit_set)
+{
+	struct sk_buff *skb;
+	struct wsm_update_ie update_ie = {
+		.what = WSM_UPDATE_IE_BEACON,
+		.count = 1,
+	};
+	u16 tim_offset, tim_length;
+
+	pr_debug("[AP] mcast: %s.\n", aid0_bit_set ? "ena" : "dis");
+
+	skb = ieee80211_beacon_get_tim(priv->hw, priv->vif,
+			&tim_offset, &tim_length);
+	if (!skb) {
+		if (!__cw1200_flush(priv, true))
+			wsm_unlock_tx(priv);
+		return -ENOENT;
+	}
+
+	if (tim_offset && tim_length >= 6) {
+		/* Ignore DTIM count from mac80211:
+		 * firmware handles DTIM internally.
+		 */
+		skb->data[tim_offset + 2] = 0;
+
+		/* Set/reset aid0 bit */
+		if (aid0_bit_set)
+			skb->data[tim_offset + 4] |= 1;
+		else
+			skb->data[tim_offset + 4] &= ~1;
+	}
+
+	update_ie.ies = &skb->data[tim_offset];
+	update_ie.length = tim_length;
+	wsm_update_ie(priv, &update_ie);
+
+	dev_kfree_skb(skb);
+
+	return 0;
+}
+
+void cw1200_set_tim_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, set_tim_work);
+	(void)cw1200_set_tim_impl(priv, priv->aid0_bit_set);
+}
+
+int cw1200_set_tim(struct ieee80211_hw *dev, struct ieee80211_sta *sta,
+		   bool set)
+{
+	struct cw1200_common *priv = dev->priv;
+	queue_work(priv->workqueue, &priv->set_tim_work);
+	return 0;
+}
+
+void cw1200_set_cts_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, set_cts_work);
+
+	u8 erp_ie[3] = {WLAN_EID_ERP_INFO, 0x1, 0};
+	struct wsm_update_ie update_ie = {
+		.what = WSM_UPDATE_IE_BEACON,
+		.count = 1,
+		.ies = erp_ie,
+		.length = 3,
+	};
+	u32 erp_info;
+	__le32 use_cts_prot;
+	mutex_lock(&priv->conf_mutex);
+	erp_info = priv->erp_info;
+	mutex_unlock(&priv->conf_mutex);
+	use_cts_prot =
+		erp_info & WLAN_ERP_USE_PROTECTION ?
+		__cpu_to_le32(1) : 0;
+
+	erp_ie[ERP_INFO_BYTE_OFFSET] = erp_info;
+
+	pr_debug("[STA] ERP information 0x%x\n", erp_info);
+
+	wsm_write_mib(priv, WSM_MIB_ID_NON_ERP_PROTECTION,
+		      &use_cts_prot, sizeof(use_cts_prot));
+	wsm_update_ie(priv, &update_ie);
+
+	return;
+}
+
+static int cw1200_set_btcoexinfo(struct cw1200_common *priv)
+{
+	struct wsm_override_internal_txrate arg;
+	int ret = 0;
+
+	if (priv->mode == NL80211_IFTYPE_STATION) {
+		/* Plumb PSPOLL and NULL template */
+		cw1200_upload_pspoll(priv);
+		cw1200_upload_null(priv);
+		cw1200_upload_qosnull(priv);
+	} else {
+		return 0;
+	}
+
+	memset(&arg, 0, sizeof(struct wsm_override_internal_txrate));
+
+	if (!priv->vif->p2p) {
+		/* STATION mode */
+		if (priv->bss_params.operational_rate_set & ~0xF) {
+			pr_debug("[STA] STA has ERP rates\n");
+			/* G or BG mode */
+			arg.internalTxRate = (__ffs(
+			priv->bss_params.operational_rate_set & ~0xF));
+		} else {
+			pr_debug("[STA] STA has non ERP rates\n");
+			/* B only mode */
+			arg.internalTxRate = (__ffs(le32_to_cpu(priv->association_mode.basic_rate_set)));
+		}
+		arg.nonErpInternalTxRate = (__ffs(le32_to_cpu(priv->association_mode.basic_rate_set)));
+	} else {
+		/* P2P mode */
+		arg.internalTxRate = (__ffs(priv->bss_params.operational_rate_set & ~0xF));
+		arg.nonErpInternalTxRate = (__ffs(priv->bss_params.operational_rate_set & ~0xF));
+	}
+
+	pr_debug("[STA] BTCOEX_INFO MODE %d, internalTxRate : %x, nonErpInternalTxRate: %x\n",
+		 priv->mode,
+		 arg.internalTxRate,
+		 arg.nonErpInternalTxRate);
+
+	ret = wsm_write_mib(priv, WSM_MIB_ID_OVERRIDE_INTERNAL_TX_RATE,
+			    &arg, sizeof(arg));
+
+	return ret;
+}
+
+void cw1200_bss_info_changed(struct ieee80211_hw *dev,
+			     struct ieee80211_vif *vif,
+			     struct ieee80211_bss_conf *info,
+			     u32 changed)
+{
+	struct cw1200_common *priv = dev->priv;
+	bool do_join = false;
+
+	mutex_lock(&priv->conf_mutex);
+
+	pr_debug("BSS CHANGED:  %08x\n", changed);
+
+	/* TODO: BSS_CHANGED_QOS */
+	/* TODO: BSS_CHANGED_TXPOWER */
+
+	if (changed & BSS_CHANGED_ARP_FILTER) {
+		struct wsm_mib_arp_ipv4_filter filter = {0};
+		int i;
+
+		pr_debug("[STA] BSS_CHANGED_ARP_FILTER cnt: %d\n",
+			 info->arp_addr_cnt);
+
+		/* Currently only one IP address is supported by firmware.
+		 * In case of more IPs arp filtering will be disabled.
+		 */
+		if (info->arp_addr_cnt > 0 &&
+		    info->arp_addr_cnt <= WSM_MAX_ARP_IP_ADDRTABLE_ENTRIES) {
+			for (i = 0; i < info->arp_addr_cnt; i++) {
+				filter.ipv4addrs[i] = info->arp_addr_list[i];
+				pr_debug("[STA] addr[%d]: 0x%X\n",
+					 i, filter.ipv4addrs[i]);
+			}
+			filter.enable = __cpu_to_le32(1);
+		}
+
+		pr_debug("[STA] arp ip filter enable: %d\n",
+			 __le32_to_cpu(filter.enable));
+
+		wsm_set_arp_ipv4_filter(priv, &filter);
+	}
+
+	if (changed &
+	    (BSS_CHANGED_BEACON |
+	     BSS_CHANGED_AP_PROBE_RESP |
+	     BSS_CHANGED_BSSID |
+	     BSS_CHANGED_SSID |
+	     BSS_CHANGED_IBSS)) {
+		pr_debug("BSS_CHANGED_BEACON\n");
+		priv->beacon_int = info->beacon_int;
+		cw1200_update_beaconing(priv);
+		cw1200_upload_beacon(priv);
+	}
+
+	if (changed & BSS_CHANGED_BEACON_ENABLED) {
+		pr_debug("BSS_CHANGED_BEACON_ENABLED (%d)\n", info->enable_beacon);
+
+		if (priv->enable_beacon != info->enable_beacon) {
+			cw1200_enable_beaconing(priv, info->enable_beacon);
+			priv->enable_beacon = info->enable_beacon;
+		}
+	}
+
+	if (changed & BSS_CHANGED_BEACON_INT) {
+		pr_debug("CHANGED_BEACON_INT\n");
+		if (info->ibss_joined)
+			do_join = true;
+		else if (priv->join_status == CW1200_JOIN_STATUS_AP)
+			cw1200_update_beaconing(priv);
+	}
+
+	/* assoc/disassoc, or maybe AID changed */
+	if (changed & BSS_CHANGED_ASSOC) {
+		wsm_lock_tx(priv);
+		priv->wep_default_key_id = -1;
+		wsm_unlock_tx(priv);
+	}
+
+	if (changed & BSS_CHANGED_BSSID) {
+		pr_debug("BSS_CHANGED_BSSID\n");
+		do_join = true;
+	}
+
+	if (changed &
+	    (BSS_CHANGED_ASSOC |
+	     BSS_CHANGED_BSSID |
+	     BSS_CHANGED_IBSS |
+	     BSS_CHANGED_BASIC_RATES |
+	     BSS_CHANGED_HT)) {
+		pr_debug("BSS_CHANGED_ASSOC\n");
+		if (info->assoc) {
+			if (priv->join_status < CW1200_JOIN_STATUS_PRE_STA) {
+				ieee80211_connection_loss(vif);
+				mutex_unlock(&priv->conf_mutex);
+				return;
+			} else if (priv->join_status == CW1200_JOIN_STATUS_PRE_STA) {
+				priv->join_status = CW1200_JOIN_STATUS_STA;
+			}
+		} else {
+			do_join = true;
+		}
+
+		if (info->assoc || info->ibss_joined) {
+			struct ieee80211_sta *sta = NULL;
+			__le32 htprot = 0;
+
+			if (info->dtim_period)
+				priv->join_dtim_period = info->dtim_period;
+			priv->beacon_int = info->beacon_int;
+
+			rcu_read_lock();
+
+			if (info->bssid && !info->ibss_joined)
+				sta = ieee80211_find_sta(vif, info->bssid);
+			if (sta) {
+				priv->ht_info.ht_cap = sta->ht_cap;
+				priv->bss_params.operational_rate_set =
+					cw1200_rate_mask_to_wsm(priv,
+								sta->supp_rates[priv->channel->band]);
+				priv->ht_info.channel_type = cfg80211_get_chandef_type(&dev->conf.chandef);
+				priv->ht_info.operation_mode = info->ht_operation_mode;
+			} else {
+				memset(&priv->ht_info, 0,
+				       sizeof(priv->ht_info));
+				priv->bss_params.operational_rate_set = -1;
+			}
+			rcu_read_unlock();
+
+			/* Non Greenfield stations present */
+			if (priv->ht_info.operation_mode &
+			    IEEE80211_HT_OP_MODE_NON_GF_STA_PRSNT)
+				htprot |= cpu_to_le32(WSM_NON_GREENFIELD_STA_PRESENT);
+
+			/* Set HT protection method */
+			htprot |= cpu_to_le32((priv->ht_info.operation_mode & IEEE80211_HT_OP_MODE_PROTECTION) << 2);
+
+			/* TODO:
+			 * STBC_param.dual_cts
+			 *  STBC_param.LSIG_TXOP_FILL
+			 */
+
+			wsm_write_mib(priv, WSM_MIB_ID_SET_HT_PROTECTION,
+				      &htprot, sizeof(htprot));
+
+			priv->association_mode.greenfield =
+				cw1200_ht_greenfield(&priv->ht_info);
+			priv->association_mode.flags =
+				WSM_ASSOCIATION_MODE_SNOOP_ASSOC_FRAMES |
+				WSM_ASSOCIATION_MODE_USE_PREAMBLE_TYPE |
+				WSM_ASSOCIATION_MODE_USE_HT_MODE |
+				WSM_ASSOCIATION_MODE_USE_BASIC_RATE_SET |
+				WSM_ASSOCIATION_MODE_USE_MPDU_START_SPACING;
+			priv->association_mode.preamble =
+				info->use_short_preamble ?
+				WSM_JOIN_PREAMBLE_SHORT :
+				WSM_JOIN_PREAMBLE_LONG;
+			priv->association_mode.basic_rate_set = __cpu_to_le32(
+				cw1200_rate_mask_to_wsm(priv,
+							info->basic_rates));
+			priv->association_mode.mpdu_start_spacing =
+				cw1200_ht_ampdu_density(&priv->ht_info);
+
+			cw1200_cqm_bssloss_sm(priv, 0, 0, 0);
+			cancel_work_sync(&priv->unjoin_work);
+
+			priv->bss_params.beacon_lost_count = priv->cqm_beacon_loss_count;
+			priv->bss_params.aid = info->aid;
+
+			if (priv->join_dtim_period < 1)
+				priv->join_dtim_period = 1;
+
+			pr_debug("[STA] DTIM %d, interval: %d\n",
+				 priv->join_dtim_period, priv->beacon_int);
+			pr_debug("[STA] Preamble: %d, Greenfield: %d, Aid: %d, Rates: 0x%.8X, Basic: 0x%.8X\n",
+				 priv->association_mode.preamble,
+				 priv->association_mode.greenfield,
+				 priv->bss_params.aid,
+				 priv->bss_params.operational_rate_set,
+				 priv->association_mode.basic_rate_set);
+			wsm_set_association_mode(priv, &priv->association_mode);
+
+			if (!info->ibss_joined) {
+				wsm_keep_alive_period(priv, 30 /* sec */);
+				wsm_set_bss_params(priv, &priv->bss_params);
+				priv->setbssparams_done = true;
+				cw1200_set_beacon_wakeup_period_work(&priv->set_beacon_wakeup_period_work);
+				cw1200_set_pm(priv, &priv->powersave_mode);
+			}
+			if (priv->vif->p2p) {
+				pr_debug("[STA] Setting p2p powersave configuration.\n");
+				wsm_set_p2p_ps_modeinfo(priv,
+							&priv->p2p_ps_modeinfo);
+			}
+			if (priv->bt_present)
+				cw1200_set_btcoexinfo(priv);
+		} else {
+			memset(&priv->association_mode, 0,
+			       sizeof(priv->association_mode));
+			memset(&priv->bss_params, 0, sizeof(priv->bss_params));
+		}
+	}
+
+	/* ERP Protection */
+	if (changed & (BSS_CHANGED_ASSOC |
+		       BSS_CHANGED_ERP_CTS_PROT |
+		       BSS_CHANGED_ERP_PREAMBLE)) {
+		u32 prev_erp_info = priv->erp_info;
+		if (info->use_cts_prot)
+			priv->erp_info |= WLAN_ERP_USE_PROTECTION;
+		else if (!(prev_erp_info & WLAN_ERP_NON_ERP_PRESENT))
+			priv->erp_info &= ~WLAN_ERP_USE_PROTECTION;
+
+		if (info->use_short_preamble)
+			priv->erp_info |= WLAN_ERP_BARKER_PREAMBLE;
+		else
+			priv->erp_info &= ~WLAN_ERP_BARKER_PREAMBLE;
+
+		pr_debug("[STA] ERP Protection: %x\n", priv->erp_info);
+
+		if (prev_erp_info != priv->erp_info)
+			queue_work(priv->workqueue, &priv->set_cts_work);
+	}
+
+	/* ERP Slottime */
+	if (changed & (BSS_CHANGED_ASSOC | BSS_CHANGED_ERP_SLOT)) {
+		__le32 slot_time = info->use_short_slot ?
+			__cpu_to_le32(9) : __cpu_to_le32(20);
+		pr_debug("[STA] Slot time: %d us.\n",
+			 __le32_to_cpu(slot_time));
+		wsm_write_mib(priv, WSM_MIB_ID_DOT11_SLOT_TIME,
+			      &slot_time, sizeof(slot_time));
+	}
+
+	if (changed & (BSS_CHANGED_ASSOC | BSS_CHANGED_CQM)) {
+		struct wsm_rcpi_rssi_threshold threshold = {
+			.rollingAverageCount = 8,
+		};
+		pr_debug("[CQM] RSSI threshold subscribe: %d +- %d\n",
+			 info->cqm_rssi_thold, info->cqm_rssi_hyst);
+		priv->cqm_rssi_thold = info->cqm_rssi_thold;
+		priv->cqm_rssi_hyst = info->cqm_rssi_hyst;
+
+		if (info->cqm_rssi_thold || info->cqm_rssi_hyst) {
+			/* RSSI subscription enabled */
+			/* TODO: It's not a correct way of setting threshold.
+			 * Upper and lower must be set equal here and adjusted
+			 * in callback. However current implementation is much
+			 * more relaible and stable.
+			 */
+
+			/* RSSI: signed Q8.0, RCPI: unsigned Q7.1
+			 * RSSI = RCPI / 2 - 110
+			 */
+			if (priv->cqm_use_rssi) {
+				threshold.upperThreshold =
+					info->cqm_rssi_thold + info->cqm_rssi_hyst;
+				threshold.lowerThreshold =
+					info->cqm_rssi_thold;
+				threshold.rssiRcpiMode |= WSM_RCPI_RSSI_USE_RSSI;
+			} else {
+				threshold.upperThreshold = (info->cqm_rssi_thold + info->cqm_rssi_hyst + 110) * 2;
+				threshold.lowerThreshold = (info->cqm_rssi_thold + 110) * 2;
+			}
+			threshold.rssiRcpiMode |= WSM_RCPI_RSSI_THRESHOLD_ENABLE;
+		} else {
+			/* There is a bug in FW, see sta.c. We have to enable
+			 * dummy subscription to get correct RSSI values.
+			 */
+			threshold.rssiRcpiMode |=
+				WSM_RCPI_RSSI_THRESHOLD_ENABLE |
+				WSM_RCPI_RSSI_DONT_USE_UPPER |
+				WSM_RCPI_RSSI_DONT_USE_LOWER;
+			if (priv->cqm_use_rssi)
+				threshold.rssiRcpiMode |= WSM_RCPI_RSSI_USE_RSSI;
+		}
+		wsm_set_rcpi_rssi_threshold(priv, &threshold);
+	}
+	mutex_unlock(&priv->conf_mutex);
+
+	if (do_join) {
+		wsm_lock_tx(priv);
+		cw1200_do_join(priv); /* Will unlock it for us */
+	}
+}
+
+void cw1200_multicast_start_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, multicast_start_work);
+	long tmo = priv->join_dtim_period *
+			(priv->beacon_int + 20) * HZ / 1024;
+
+	cancel_work_sync(&priv->multicast_stop_work);
+
+	if (!priv->aid0_bit_set) {
+		wsm_lock_tx(priv);
+		cw1200_set_tim_impl(priv, true);
+		priv->aid0_bit_set = true;
+		mod_timer(&priv->mcast_timeout, jiffies + tmo);
+		wsm_unlock_tx(priv);
+	}
+}
+
+void cw1200_multicast_stop_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, multicast_stop_work);
+
+	if (priv->aid0_bit_set) {
+		del_timer_sync(&priv->mcast_timeout);
+		wsm_lock_tx(priv);
+		priv->aid0_bit_set = false;
+		cw1200_set_tim_impl(priv, false);
+		wsm_unlock_tx(priv);
+	}
+}
+
+void cw1200_mcast_timeout(unsigned long arg)
+{
+	struct cw1200_common *priv =
+		(struct cw1200_common *)arg;
+
+	wiphy_warn(priv->hw->wiphy,
+		   "Multicast delivery timeout.\n");
+	spin_lock_bh(&priv->ps_state_lock);
+	priv->tx_multicast = priv->aid0_bit_set &&
+			priv->buffered_multicasts;
+	if (priv->tx_multicast)
+		cw1200_bh_wakeup(priv);
+	spin_unlock_bh(&priv->ps_state_lock);
+}
+
+int cw1200_ampdu_action(struct ieee80211_hw *hw,
+			struct ieee80211_vif *vif,
+			enum ieee80211_ampdu_mlme_action action,
+			struct ieee80211_sta *sta, u16 tid, u16 *ssn,
+			u8 buf_size)
+{
+	/* Aggregation is implemented fully in firmware,
+	 * including block ack negotiation. Do not allow
+	 * mac80211 stack to do anything: it interferes with
+	 * the firmware.
+	 */
+
+	/* Note that we still need this function stubbed. */
+	return -ENOTSUPP;
+}
+
+/* ******************************************************************** */
+/* WSM callback								*/
+void cw1200_suspend_resume(struct cw1200_common *priv,
+			  struct wsm_suspend_resume *arg)
+{
+	pr_debug("[AP] %s: %s\n",
+		 arg->stop ? "stop" : "start",
+		 arg->multicast ? "broadcast" : "unicast");
+
+	if (arg->multicast) {
+		bool cancel_tmo = false;
+		spin_lock_bh(&priv->ps_state_lock);
+		if (arg->stop) {
+			priv->tx_multicast = false;
+		} else {
+			/* Firmware sends this indication every DTIM if there
+			 * is a STA in powersave connected. There is no reason
+			 * to suspend, following wakeup will consume much more
+			 * power than it could be saved.
+			 */
+			cw1200_pm_stay_awake(&priv->pm_state,
+					     priv->join_dtim_period *
+					     (priv->beacon_int + 20) * HZ / 1024);
+			priv->tx_multicast = (priv->aid0_bit_set &&
+					      priv->buffered_multicasts);
+			if (priv->tx_multicast) {
+				cancel_tmo = true;
+				cw1200_bh_wakeup(priv);
+			}
+		}
+		spin_unlock_bh(&priv->ps_state_lock);
+		if (cancel_tmo)
+			del_timer_sync(&priv->mcast_timeout);
+	} else {
+		spin_lock_bh(&priv->ps_state_lock);
+		cw1200_ps_notify(priv, arg->link_id, arg->stop);
+		spin_unlock_bh(&priv->ps_state_lock);
+		if (!arg->stop)
+			cw1200_bh_wakeup(priv);
+	}
+	return;
+}
+
+/* ******************************************************************** */
+/* AP privates								*/
+
+static int cw1200_upload_beacon(struct cw1200_common *priv)
+{
+	int ret = 0;
+	struct ieee80211_mgmt *mgmt;
+	struct wsm_template_frame frame = {
+		.frame_type = WSM_FRAME_TYPE_BEACON,
+	};
+
+	u16 tim_offset;
+	u16 tim_len;
+
+	if (priv->mode == NL80211_IFTYPE_STATION ||
+	    priv->mode == NL80211_IFTYPE_MONITOR ||
+	    priv->mode == NL80211_IFTYPE_UNSPECIFIED)
+		goto done;
+
+	if (priv->vif->p2p)
+		frame.rate = WSM_TRANSMIT_RATE_6;
+
+	frame.skb = ieee80211_beacon_get_tim(priv->hw, priv->vif,
+					     &tim_offset, &tim_len);
+	if (!frame.skb)
+		return -ENOMEM;
+
+	ret = wsm_set_template_frame(priv, &frame);
+
+	if (ret)
+		goto done;
+
+	/* TODO: Distill probe resp; remove TIM
+	 * and any other beacon-specific IEs
+	 */
+	mgmt = (void *)frame.skb->data;
+	mgmt->frame_control =
+		__cpu_to_le16(IEEE80211_FTYPE_MGMT |
+			      IEEE80211_STYPE_PROBE_RESP);
+
+	frame.frame_type = WSM_FRAME_TYPE_PROBE_RESPONSE;
+	if (priv->vif->p2p) {
+		ret = wsm_set_probe_responder(priv, true);
+	} else {
+		ret = wsm_set_template_frame(priv, &frame);
+		wsm_set_probe_responder(priv, false);
+	}
+
+done:
+	dev_kfree_skb(frame.skb);
+
+	return ret;
+}
+
+static int cw1200_upload_pspoll(struct cw1200_common *priv)
+{
+	int ret = 0;
+	struct wsm_template_frame frame = {
+		.frame_type = WSM_FRAME_TYPE_PS_POLL,
+		.rate = 0xFF,
+	};
+
+
+	frame.skb = ieee80211_pspoll_get(priv->hw, priv->vif);
+	if (!frame.skb)
+		return -ENOMEM;
+
+	ret = wsm_set_template_frame(priv, &frame);
+
+	dev_kfree_skb(frame.skb);
+
+	return ret;
+}
+
+static int cw1200_upload_null(struct cw1200_common *priv)
+{
+	int ret = 0;
+	struct wsm_template_frame frame = {
+		.frame_type = WSM_FRAME_TYPE_NULL,
+		.rate = 0xFF,
+	};
+
+	frame.skb = ieee80211_nullfunc_get(priv->hw, priv->vif);
+	if (!frame.skb)
+		return -ENOMEM;
+
+	ret = wsm_set_template_frame(priv, &frame);
+
+	dev_kfree_skb(frame.skb);
+
+	return ret;
+}
+
+static int cw1200_upload_qosnull(struct cw1200_common *priv)
+{
+	int ret = 0;
+	/* TODO:  This needs to be implemented
+
+	struct wsm_template_frame frame = {
+		.frame_type = WSM_FRAME_TYPE_QOS_NULL,
+		.rate = 0xFF,
+	};
+
+	frame.skb = ieee80211_qosnullfunc_get(priv->hw, priv->vif);
+	if (!frame.skb)
+		return -ENOMEM;
+
+	ret = wsm_set_template_frame(priv, &frame);
+
+	dev_kfree_skb(frame.skb);
+
+	*/
+	return ret;
+}
+
+static int cw1200_enable_beaconing(struct cw1200_common *priv,
+				   bool enable)
+{
+	struct wsm_beacon_transmit transmit = {
+		.enable_beaconing = enable,
+	};
+
+	return wsm_beacon_transmit(priv, &transmit);
+}
+
+static int cw1200_start_ap(struct cw1200_common *priv)
+{
+	int ret;
+	struct ieee80211_bss_conf *conf = &priv->vif->bss_conf;
+	struct wsm_start start = {
+		.mode = priv->vif->p2p ?
+				WSM_START_MODE_P2P_GO : WSM_START_MODE_AP,
+		.band = (priv->channel->band == IEEE80211_BAND_5GHZ) ?
+				WSM_PHY_BAND_5G : WSM_PHY_BAND_2_4G,
+		.channel_number = priv->channel->hw_value,
+		.beacon_interval = conf->beacon_int,
+		.dtim_period = conf->dtim_period,
+		.preamble = conf->use_short_preamble ?
+				WSM_JOIN_PREAMBLE_SHORT :
+				WSM_JOIN_PREAMBLE_LONG,
+		.probe_delay = 100,
+		.basic_rate_set = cw1200_rate_mask_to_wsm(priv,
+				conf->basic_rates),
+	};
+	struct wsm_operational_mode mode = {
+		.power_mode = cw1200_power_mode,
+		.disable_more_flag_usage = true,
+	};
+
+	memset(start.ssid, 0, sizeof(start.ssid));
+	if (!conf->hidden_ssid) {
+		start.ssid_len = conf->ssid_len;
+		memcpy(start.ssid, conf->ssid, start.ssid_len);
+	}
+
+	priv->beacon_int = conf->beacon_int;
+	priv->join_dtim_period = conf->dtim_period;
+
+	memset(&priv->link_id_db, 0, sizeof(priv->link_id_db));
+
+	pr_debug("[AP] ch: %d(%d), bcn: %d(%d), brt: 0x%.8X, ssid: %.*s.\n",
+		 start.channel_number, start.band,
+		 start.beacon_interval, start.dtim_period,
+		 start.basic_rate_set,
+		 start.ssid_len, start.ssid);
+	ret = wsm_start(priv, &start);
+	if (!ret)
+		ret = cw1200_upload_keys(priv);
+	if (!ret && priv->vif->p2p) {
+		pr_debug("[AP] Setting p2p powersave configuration.\n");
+		wsm_set_p2p_ps_modeinfo(priv, &priv->p2p_ps_modeinfo);
+	}
+	if (!ret) {
+		wsm_set_block_ack_policy(priv, 0, 0);
+		priv->join_status = CW1200_JOIN_STATUS_AP;
+		cw1200_update_filtering(priv);
+	}
+	wsm_set_operational_mode(priv, &mode);
+	return ret;
+}
+
+static int cw1200_update_beaconing(struct cw1200_common *priv)
+{
+	struct ieee80211_bss_conf *conf = &priv->vif->bss_conf;
+	struct wsm_reset reset = {
+		.link_id = 0,
+		.reset_statistics = true,
+	};
+
+	if (priv->mode == NL80211_IFTYPE_AP) {
+		/* TODO: check if changed channel, band */
+		if (priv->join_status != CW1200_JOIN_STATUS_AP ||
+		    priv->beacon_int != conf->beacon_int) {
+			pr_debug("ap restarting\n");
+			wsm_lock_tx(priv);
+			if (priv->join_status != CW1200_JOIN_STATUS_PASSIVE)
+				wsm_reset(priv, &reset);
+			priv->join_status = CW1200_JOIN_STATUS_PASSIVE;
+			cw1200_start_ap(priv);
+			wsm_unlock_tx(priv);
+		} else
+			pr_debug("ap started join_status: %d\n",
+				 priv->join_status);
+	}
+	return 0;
+}
diff --git a/drivers/net/wireless/cw1200/sta.h b/drivers/net/wireless/cw1200/sta.h
new file mode 100644
index 0000000..35babb6
--- /dev/null
+++ b/drivers/net/wireless/cw1200/sta.h
@@ -0,0 +1,123 @@
+/*
+ * Mac80211 STA interface for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef STA_H_INCLUDED
+#define STA_H_INCLUDED
+
+/* ******************************************************************** */
+/* mac80211 API								*/
+
+int cw1200_start(struct ieee80211_hw *dev);
+void cw1200_stop(struct ieee80211_hw *dev);
+int cw1200_add_interface(struct ieee80211_hw *dev,
+			 struct ieee80211_vif *vif);
+void cw1200_remove_interface(struct ieee80211_hw *dev,
+			     struct ieee80211_vif *vif);
+int cw1200_change_interface(struct ieee80211_hw *dev,
+			    struct ieee80211_vif *vif,
+			    enum nl80211_iftype new_type,
+			    bool p2p);
+int cw1200_config(struct ieee80211_hw *dev, u32 changed);
+void cw1200_configure_filter(struct ieee80211_hw *dev,
+			     unsigned int changed_flags,
+			     unsigned int *total_flags,
+			     u64 multicast);
+int cw1200_conf_tx(struct ieee80211_hw *dev, struct ieee80211_vif *vif,
+		   u16 queue, const struct ieee80211_tx_queue_params *params);
+int cw1200_get_stats(struct ieee80211_hw *dev,
+		     struct ieee80211_low_level_stats *stats);
+int cw1200_set_key(struct ieee80211_hw *dev, enum set_key_cmd cmd,
+		   struct ieee80211_vif *vif, struct ieee80211_sta *sta,
+		   struct ieee80211_key_conf *key);
+
+int cw1200_set_rts_threshold(struct ieee80211_hw *hw, u32 value);
+
+void cw1200_flush(struct ieee80211_hw *hw, u32 queues, bool drop);
+
+u64 cw1200_prepare_multicast(struct ieee80211_hw *hw,
+			     struct netdev_hw_addr_list *mc_list);
+
+int cw1200_set_pm(struct cw1200_common *priv, const struct wsm_set_pm *arg);
+
+/* ******************************************************************** */
+/* WSM callbacks							*/
+
+void cw1200_join_complete_cb(struct cw1200_common *priv,
+				struct wsm_join_complete *arg);
+
+/* ******************************************************************** */
+/* WSM events								*/
+
+void cw1200_free_event_queue(struct cw1200_common *priv);
+void cw1200_event_handler(struct work_struct *work);
+void cw1200_bss_loss_work(struct work_struct *work);
+void cw1200_bss_params_work(struct work_struct *work);
+void cw1200_keep_alive_work(struct work_struct *work);
+void cw1200_tx_failure_work(struct work_struct *work);
+
+void __cw1200_cqm_bssloss_sm(struct cw1200_common *priv, int init, int good,
+			     int bad);
+static inline void cw1200_cqm_bssloss_sm(struct cw1200_common *priv,
+					 int init, int good, int bad)
+{
+	spin_lock(&priv->bss_loss_lock);
+	__cw1200_cqm_bssloss_sm(priv, init, good, bad);
+	spin_unlock(&priv->bss_loss_lock);
+}
+
+/* ******************************************************************** */
+/* Internal API								*/
+
+int cw1200_setup_mac(struct cw1200_common *priv);
+void cw1200_join_timeout(struct work_struct *work);
+void cw1200_unjoin_work(struct work_struct *work);
+void cw1200_join_complete_work(struct work_struct *work);
+void cw1200_wep_key_work(struct work_struct *work);
+void cw1200_update_listening(struct cw1200_common *priv, bool enabled);
+void cw1200_update_filtering(struct cw1200_common *priv);
+void cw1200_update_filtering_work(struct work_struct *work);
+void cw1200_set_beacon_wakeup_period_work(struct work_struct *work);
+int cw1200_enable_listening(struct cw1200_common *priv);
+int cw1200_disable_listening(struct cw1200_common *priv);
+int cw1200_set_uapsd_param(struct cw1200_common *priv,
+				const struct wsm_edca_params *arg);
+void cw1200_ba_work(struct work_struct *work);
+void cw1200_ba_timer(unsigned long arg);
+
+/* AP stuffs */
+int cw1200_set_tim(struct ieee80211_hw *dev, struct ieee80211_sta *sta,
+		   bool set);
+int cw1200_sta_add(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
+		   struct ieee80211_sta *sta);
+int cw1200_sta_remove(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
+		      struct ieee80211_sta *sta);
+void cw1200_sta_notify(struct ieee80211_hw *dev, struct ieee80211_vif *vif,
+		       enum sta_notify_cmd notify_cmd,
+		       struct ieee80211_sta *sta);
+void cw1200_bss_info_changed(struct ieee80211_hw *dev,
+			     struct ieee80211_vif *vif,
+			     struct ieee80211_bss_conf *info,
+			     u32 changed);
+int cw1200_ampdu_action(struct ieee80211_hw *hw,
+			struct ieee80211_vif *vif,
+			enum ieee80211_ampdu_mlme_action action,
+			struct ieee80211_sta *sta, u16 tid, u16 *ssn,
+			u8 buf_size);
+
+void cw1200_suspend_resume(struct cw1200_common *priv,
+			  struct wsm_suspend_resume *arg);
+void cw1200_set_tim_work(struct work_struct *work);
+void cw1200_set_cts_work(struct work_struct *work);
+void cw1200_multicast_start_work(struct work_struct *work);
+void cw1200_multicast_stop_work(struct work_struct *work);
+void cw1200_mcast_timeout(unsigned long arg);
+
+#endif
diff --git a/drivers/net/wireless/cw1200/txrx.c b/drivers/net/wireless/cw1200/txrx.c
new file mode 100644
index 0000000..5862c37
--- /dev/null
+++ b/drivers/net/wireless/cw1200/txrx.c
@@ -0,0 +1,1473 @@
+/*
+ * Datapath implementation for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <net/mac80211.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+
+#include "cw1200.h"
+#include "wsm.h"
+#include "bh.h"
+#include "sta.h"
+#include "debug.h"
+
+#define CW1200_INVALID_RATE_ID (0xFF)
+
+static int cw1200_handle_action_rx(struct cw1200_common *priv,
+				   struct sk_buff *skb);
+static const struct ieee80211_rate *
+cw1200_get_tx_rate(const struct cw1200_common *priv,
+		   const struct ieee80211_tx_rate *rate);
+
+/* ******************************************************************** */
+/* TX queue lock / unlock						*/
+
+static inline void cw1200_tx_queues_lock(struct cw1200_common *priv)
+{
+	int i;
+	for (i = 0; i < 4; ++i)
+		cw1200_queue_lock(&priv->tx_queue[i]);
+}
+
+static inline void cw1200_tx_queues_unlock(struct cw1200_common *priv)
+{
+	int i;
+	for (i = 0; i < 4; ++i)
+		cw1200_queue_unlock(&priv->tx_queue[i]);
+}
+
+/* ******************************************************************** */
+/* TX policy cache implementation					*/
+
+static void tx_policy_dump(struct tx_policy *policy)
+{
+	pr_debug("[TX policy] %.1X%.1X%.1X%.1X%.1X%.1X%.1X%.1X %.1X%.1X%.1X%.1X%.1X%.1X%.1X%.1X %.1X%.1X%.1X%.1X%.1X%.1X%.1X%.1X: %d\n",
+		 policy->raw[0] & 0x0F,  policy->raw[0] >> 4,
+		 policy->raw[1] & 0x0F,  policy->raw[1] >> 4,
+		 policy->raw[2] & 0x0F,  policy->raw[2] >> 4,
+		 policy->raw[3] & 0x0F,  policy->raw[3] >> 4,
+		 policy->raw[4] & 0x0F,  policy->raw[4] >> 4,
+		 policy->raw[5] & 0x0F,  policy->raw[5] >> 4,
+		 policy->raw[6] & 0x0F,  policy->raw[6] >> 4,
+		 policy->raw[7] & 0x0F,  policy->raw[7] >> 4,
+		 policy->raw[8] & 0x0F,  policy->raw[8] >> 4,
+		 policy->raw[9] & 0x0F,  policy->raw[9] >> 4,
+		 policy->raw[10] & 0x0F,  policy->raw[10] >> 4,
+		 policy->raw[11] & 0x0F,  policy->raw[11] >> 4,
+		 policy->defined);
+}
+
+static void tx_policy_build(const struct cw1200_common *priv,
+	/* [out] */ struct tx_policy *policy,
+	struct ieee80211_tx_rate *rates, size_t count)
+{
+	int i, j;
+	unsigned limit = priv->short_frame_max_tx_count;
+	unsigned total = 0;
+	BUG_ON(rates[0].idx < 0);
+	memset(policy, 0, sizeof(*policy));
+
+	/* Sort rates in descending order. */
+	for (i = 1; i < count; ++i) {
+		if (rates[i].idx < 0) {
+			count = i;
+			break;
+		}
+		if (rates[i].idx > rates[i - 1].idx) {
+			struct ieee80211_tx_rate tmp = rates[i - 1];
+			rates[i - 1] = rates[i];
+			rates[i] = tmp;
+		}
+	}
+
+	/* Eliminate duplicates. */
+	total = rates[0].count;
+	for (i = 0, j = 1; j < count; ++j) {
+		if (rates[j].idx == rates[i].idx) {
+			rates[i].count += rates[j].count;
+		} else if (rates[j].idx > rates[i].idx) {
+			break;
+		} else {
+			++i;
+			if (i != j)
+				rates[i] = rates[j];
+		}
+		total += rates[j].count;
+	}
+	count = i + 1;
+
+	/* Re-fill policy trying to keep every requested rate and with
+	 * respect to the global max tx retransmission count.
+	 */
+	if (limit < count)
+		limit = count;
+	if (total > limit) {
+		for (i = 0; i < count; ++i) {
+			int left = count - i - 1;
+			if (rates[i].count > limit - left)
+				rates[i].count = limit - left;
+			limit -= rates[i].count;
+		}
+	}
+
+	/* HACK!!! Device has problems (at least) switching from
+	 * 54Mbps CTS to 1Mbps. This switch takes enormous amount
+	 * of time (100-200 ms), leading to valuable throughput drop.
+	 * As a workaround, additional g-rates are injected to the
+	 * policy.
+	 */
+	if (count == 2 && !(rates[0].flags & IEEE80211_TX_RC_MCS) &&
+	    rates[0].idx > 4 && rates[0].count > 2 &&
+	    rates[1].idx < 2) {
+		int mid_rate = (rates[0].idx + 4) >> 1;
+
+		/* Decrease number of retries for the initial rate */
+		rates[0].count -= 2;
+
+		if (mid_rate != 4) {
+			/* Keep fallback rate at 1Mbps. */
+			rates[3] = rates[1];
+
+			/* Inject 1 transmission on lowest g-rate */
+			rates[2].idx = 4;
+			rates[2].count = 1;
+			rates[2].flags = rates[1].flags;
+
+			/* Inject 1 transmission on mid-rate */
+			rates[1].idx = mid_rate;
+			rates[1].count = 1;
+
+			/* Fallback to 1 Mbps is a really bad thing,
+			 * so let's try to increase probability of
+			 * successful transmission on the lowest g rate
+			 * even more
+			 */
+			if (rates[0].count >= 3) {
+				--rates[0].count;
+				++rates[2].count;
+			}
+
+			/* Adjust amount of rates defined */
+			count += 2;
+		} else {
+			/* Keep fallback rate at 1Mbps. */
+			rates[2] = rates[1];
+
+			/* Inject 2 transmissions on lowest g-rate */
+			rates[1].idx = 4;
+			rates[1].count = 2;
+
+			/* Adjust amount of rates defined */
+			count += 1;
+		}
+	}
+
+	policy->defined = cw1200_get_tx_rate(priv, &rates[0])->hw_value + 1;
+
+	for (i = 0; i < count; ++i) {
+		register unsigned rateid, off, shift, retries;
+
+		rateid = cw1200_get_tx_rate(priv, &rates[i])->hw_value;
+		off = rateid >> 3;		/* eq. rateid / 8 */
+		shift = (rateid & 0x07) << 2;	/* eq. (rateid % 8) * 4 */
+
+		retries = rates[i].count;
+		if (retries > 0x0F) {
+			rates[i].count = 0x0f;
+			retries = 0x0F;
+		}
+		policy->tbl[off] |= __cpu_to_le32(retries << shift);
+		policy->retry_count += retries;
+	}
+
+	pr_debug("[TX policy] Policy (%zu): %d:%d, %d:%d, %d:%d, %d:%d\n",
+		 count,
+		 rates[0].idx, rates[0].count,
+		 rates[1].idx, rates[1].count,
+		 rates[2].idx, rates[2].count,
+		 rates[3].idx, rates[3].count);
+}
+
+static inline bool tx_policy_is_equal(const struct tx_policy *wanted,
+					const struct tx_policy *cached)
+{
+	size_t count = wanted->defined >> 1;
+	if (wanted->defined > cached->defined)
+		return false;
+	if (count) {
+		if (memcmp(wanted->raw, cached->raw, count))
+			return false;
+	}
+	if (wanted->defined & 1) {
+		if ((wanted->raw[count] & 0x0F) != (cached->raw[count] & 0x0F))
+			return false;
+	}
+	return true;
+}
+
+static int tx_policy_find(struct tx_policy_cache *cache,
+				const struct tx_policy *wanted)
+{
+	/* O(n) complexity. Not so good, but there's only 8 entries in
+	 * the cache.
+	 * Also lru helps to reduce search time.
+	 */
+	struct tx_policy_cache_entry *it;
+	/* First search for policy in "used" list */
+	list_for_each_entry(it, &cache->used, link) {
+		if (tx_policy_is_equal(wanted, &it->policy))
+			return it - cache->cache;
+	}
+	/* Then - in "free list" */
+	list_for_each_entry(it, &cache->free, link) {
+		if (tx_policy_is_equal(wanted, &it->policy))
+			return it - cache->cache;
+	}
+	return -1;
+}
+
+static inline void tx_policy_use(struct tx_policy_cache *cache,
+				 struct tx_policy_cache_entry *entry)
+{
+	++entry->policy.usage_count;
+	list_move(&entry->link, &cache->used);
+}
+
+static inline int tx_policy_release(struct tx_policy_cache *cache,
+				    struct tx_policy_cache_entry *entry)
+{
+	int ret = --entry->policy.usage_count;
+	if (!ret)
+		list_move(&entry->link, &cache->free);
+	return ret;
+}
+
+void tx_policy_clean(struct cw1200_common *priv)
+{
+	int idx, locked;
+	struct tx_policy_cache *cache = &priv->tx_policy_cache;
+	struct tx_policy_cache_entry *entry;
+
+	cw1200_tx_queues_lock(priv);
+	spin_lock_bh(&cache->lock);
+	locked = list_empty(&cache->free);
+
+	for (idx = 0; idx < TX_POLICY_CACHE_SIZE; idx++) {
+		entry = &cache->cache[idx];
+		/* Policy usage count should be 0 at this time as all queues
+		   should be empty
+		 */
+		if (WARN_ON(entry->policy.usage_count)) {
+			entry->policy.usage_count = 0;
+			list_move(&entry->link, &cache->free);
+		}
+		memset(&entry->policy, 0, sizeof(entry->policy));
+	}
+	if (locked)
+		cw1200_tx_queues_unlock(priv);
+
+	cw1200_tx_queues_unlock(priv);
+	spin_unlock_bh(&cache->lock);
+}
+
+/* ******************************************************************** */
+/* External TX policy cache API						*/
+
+void tx_policy_init(struct cw1200_common *priv)
+{
+	struct tx_policy_cache *cache = &priv->tx_policy_cache;
+	int i;
+
+	memset(cache, 0, sizeof(*cache));
+
+	spin_lock_init(&cache->lock);
+	INIT_LIST_HEAD(&cache->used);
+	INIT_LIST_HEAD(&cache->free);
+
+	for (i = 0; i < TX_POLICY_CACHE_SIZE; ++i)
+		list_add(&cache->cache[i].link, &cache->free);
+}
+
+static int tx_policy_get(struct cw1200_common *priv,
+		  struct ieee80211_tx_rate *rates,
+		  size_t count, bool *renew)
+{
+	int idx;
+	struct tx_policy_cache *cache = &priv->tx_policy_cache;
+	struct tx_policy wanted;
+
+	tx_policy_build(priv, &wanted, rates, count);
+
+	spin_lock_bh(&cache->lock);
+	if (WARN_ON_ONCE(list_empty(&cache->free))) {
+		spin_unlock_bh(&cache->lock);
+		return CW1200_INVALID_RATE_ID;
+	}
+	idx = tx_policy_find(cache, &wanted);
+	if (idx >= 0) {
+		pr_debug("[TX policy] Used TX policy: %d\n", idx);
+		*renew = false;
+	} else {
+		struct tx_policy_cache_entry *entry;
+		*renew = true;
+		/* If policy is not found create a new one
+		 * using the oldest entry in "free" list
+		 */
+		entry = list_entry(cache->free.prev,
+			struct tx_policy_cache_entry, link);
+		entry->policy = wanted;
+		idx = entry - cache->cache;
+		pr_debug("[TX policy] New TX policy: %d\n", idx);
+		tx_policy_dump(&entry->policy);
+	}
+	tx_policy_use(cache, &cache->cache[idx]);
+	if (list_empty(&cache->free)) {
+		/* Lock TX queues. */
+		cw1200_tx_queues_lock(priv);
+	}
+	spin_unlock_bh(&cache->lock);
+	return idx;
+}
+
+static void tx_policy_put(struct cw1200_common *priv, int idx)
+{
+	int usage, locked;
+	struct tx_policy_cache *cache = &priv->tx_policy_cache;
+
+	spin_lock_bh(&cache->lock);
+	locked = list_empty(&cache->free);
+	usage = tx_policy_release(cache, &cache->cache[idx]);
+	if (locked && !usage) {
+		/* Unlock TX queues. */
+		cw1200_tx_queues_unlock(priv);
+	}
+	spin_unlock_bh(&cache->lock);
+}
+
+static int tx_policy_upload(struct cw1200_common *priv)
+{
+	struct tx_policy_cache *cache = &priv->tx_policy_cache;
+	int i;
+	struct wsm_set_tx_rate_retry_policy arg = {
+		.num = 0,
+	};
+	spin_lock_bh(&cache->lock);
+
+	/* Upload only modified entries. */
+	for (i = 0; i < TX_POLICY_CACHE_SIZE; ++i) {
+		struct tx_policy *src = &cache->cache[i].policy;
+		if (src->retry_count && !src->uploaded) {
+			struct wsm_tx_rate_retry_policy *dst =
+				&arg.tbl[arg.num];
+			dst->index = i;
+			dst->short_retries = priv->short_frame_max_tx_count;
+			dst->long_retries = priv->long_frame_max_tx_count;
+
+			dst->flags = WSM_TX_RATE_POLICY_FLAG_TERMINATE_WHEN_FINISHED |
+				WSM_TX_RATE_POLICY_FLAG_COUNT_INITIAL_TRANSMIT;
+			memcpy(dst->rate_count_indices, src->tbl,
+			       sizeof(dst->rate_count_indices));
+			src->uploaded = 1;
+			++arg.num;
+		}
+	}
+	spin_unlock_bh(&cache->lock);
+	cw1200_debug_tx_cache_miss(priv);
+	pr_debug("[TX policy] Upload %d policies\n", arg.num);
+	return wsm_set_tx_rate_retry_policy(priv, &arg);
+}
+
+void tx_policy_upload_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, tx_policy_upload_work);
+
+	pr_debug("[TX] TX policy upload.\n");
+	tx_policy_upload(priv);
+
+	wsm_unlock_tx(priv);
+	cw1200_tx_queues_unlock(priv);
+}
+
+/* ******************************************************************** */
+/* cw1200 TX implementation						*/
+
+struct cw1200_txinfo {
+	struct sk_buff *skb;
+	unsigned queue;
+	struct ieee80211_tx_info *tx_info;
+	const struct ieee80211_rate *rate;
+	struct ieee80211_hdr *hdr;
+	size_t hdrlen;
+	const u8 *da;
+	struct cw1200_sta_priv *sta_priv;
+	struct ieee80211_sta *sta;
+	struct cw1200_txpriv txpriv;
+};
+
+u32 cw1200_rate_mask_to_wsm(struct cw1200_common *priv, u32 rates)
+{
+	u32 ret = 0;
+	int i;
+	for (i = 0; i < 32; ++i) {
+		if (rates & BIT(i))
+			ret |= BIT(priv->rates[i].hw_value);
+	}
+	return ret;
+}
+
+static const struct ieee80211_rate *
+cw1200_get_tx_rate(const struct cw1200_common *priv,
+		   const struct ieee80211_tx_rate *rate)
+{
+	if (rate->idx < 0)
+		return NULL;
+	if (rate->flags & IEEE80211_TX_RC_MCS)
+		return &priv->mcs_rates[rate->idx];
+	return &priv->hw->wiphy->bands[priv->channel->band]->
+		bitrates[rate->idx];
+}
+
+static int
+cw1200_tx_h_calc_link_ids(struct cw1200_common *priv,
+			  struct cw1200_txinfo *t)
+{
+	if (t->sta && t->sta_priv->link_id)
+		t->txpriv.raw_link_id =
+				t->txpriv.link_id =
+				t->sta_priv->link_id;
+	else if (priv->mode != NL80211_IFTYPE_AP)
+		t->txpriv.raw_link_id =
+				t->txpriv.link_id = 0;
+	else if (is_multicast_ether_addr(t->da)) {
+		if (priv->enable_beacon) {
+			t->txpriv.raw_link_id = 0;
+			t->txpriv.link_id = CW1200_LINK_ID_AFTER_DTIM;
+		} else {
+			t->txpriv.raw_link_id = 0;
+			t->txpriv.link_id = 0;
+		}
+	} else {
+		t->txpriv.link_id = cw1200_find_link_id(priv, t->da);
+		if (!t->txpriv.link_id)
+			t->txpriv.link_id = cw1200_alloc_link_id(priv, t->da);
+		if (!t->txpriv.link_id) {
+			wiphy_err(priv->hw->wiphy,
+				  "No more link IDs available.\n");
+			return -ENOENT;
+		}
+		t->txpriv.raw_link_id = t->txpriv.link_id;
+	}
+	if (t->txpriv.raw_link_id)
+		priv->link_id_db[t->txpriv.raw_link_id - 1].timestamp =
+				jiffies;
+	if (t->sta && (t->sta->uapsd_queues & BIT(t->queue)))
+		t->txpriv.link_id = CW1200_LINK_ID_UAPSD;
+	return 0;
+}
+
+static void
+cw1200_tx_h_pm(struct cw1200_common *priv,
+	       struct cw1200_txinfo *t)
+{
+	if (ieee80211_is_auth(t->hdr->frame_control)) {
+		u32 mask = ~BIT(t->txpriv.raw_link_id);
+		spin_lock_bh(&priv->ps_state_lock);
+		priv->sta_asleep_mask &= mask;
+		priv->pspoll_mask &= mask;
+		spin_unlock_bh(&priv->ps_state_lock);
+	}
+}
+
+static void
+cw1200_tx_h_calc_tid(struct cw1200_common *priv,
+		     struct cw1200_txinfo *t)
+{
+	if (ieee80211_is_data_qos(t->hdr->frame_control)) {
+		u8 *qos = ieee80211_get_qos_ctl(t->hdr);
+		t->txpriv.tid = qos[0] & IEEE80211_QOS_CTL_TID_MASK;
+	} else if (ieee80211_is_data(t->hdr->frame_control)) {
+		t->txpriv.tid = 0;
+	}
+}
+
+static int
+cw1200_tx_h_crypt(struct cw1200_common *priv,
+		  struct cw1200_txinfo *t)
+{
+	if (!t->tx_info->control.hw_key ||
+	    !ieee80211_has_protected(t->hdr->frame_control))
+		return 0;
+
+	t->hdrlen += t->tx_info->control.hw_key->iv_len;
+	skb_put(t->skb, t->tx_info->control.hw_key->icv_len);
+
+	if (t->tx_info->control.hw_key->cipher == WLAN_CIPHER_SUITE_TKIP)
+		skb_put(t->skb, 8); /* MIC space */
+
+	return 0;
+}
+
+static int
+cw1200_tx_h_align(struct cw1200_common *priv,
+		  struct cw1200_txinfo *t,
+		  u8 *flags)
+{
+	size_t offset = (size_t)t->skb->data & 3;
+
+	if (!offset)
+		return 0;
+
+	if (offset & 1) {
+		wiphy_err(priv->hw->wiphy,
+			  "Bug: attempt to transmit a frame with wrong alignment: %zu\n",
+			  offset);
+		return -EINVAL;
+	}
+
+	if (skb_headroom(t->skb) < offset) {
+		wiphy_err(priv->hw->wiphy,
+			  "Bug: no space allocated for DMA alignment. headroom: %d\n",
+			  skb_headroom(t->skb));
+		return -ENOMEM;
+	}
+	skb_push(t->skb, offset);
+	t->hdrlen += offset;
+	t->txpriv.offset += offset;
+	*flags |= WSM_TX_2BYTES_SHIFT;
+	cw1200_debug_tx_align(priv);
+	return 0;
+}
+
+static int
+cw1200_tx_h_action(struct cw1200_common *priv,
+		   struct cw1200_txinfo *t)
+{
+	struct ieee80211_mgmt *mgmt =
+		(struct ieee80211_mgmt *)t->hdr;
+	if (ieee80211_is_action(t->hdr->frame_control) &&
+	    mgmt->u.action.category == WLAN_CATEGORY_BACK)
+		return 1;
+	else
+		return 0;
+}
+
+/* Add WSM header */
+static struct wsm_tx *
+cw1200_tx_h_wsm(struct cw1200_common *priv,
+		struct cw1200_txinfo *t)
+{
+	struct wsm_tx *wsm;
+
+	if (skb_headroom(t->skb) < sizeof(struct wsm_tx)) {
+		wiphy_err(priv->hw->wiphy,
+			  "Bug: no space allocated for WSM header. headroom: %d\n",
+			  skb_headroom(t->skb));
+		return NULL;
+	}
+
+	wsm = (struct wsm_tx *)skb_push(t->skb, sizeof(struct wsm_tx));
+	t->txpriv.offset += sizeof(struct wsm_tx);
+	memset(wsm, 0, sizeof(*wsm));
+	wsm->hdr.len = __cpu_to_le16(t->skb->len);
+	wsm->hdr.id = __cpu_to_le16(0x0004);
+	wsm->queue_id = wsm_queue_id_to_wsm(t->queue);
+	return wsm;
+}
+
+/* BT Coex specific handling */
+static void
+cw1200_tx_h_bt(struct cw1200_common *priv,
+	       struct cw1200_txinfo *t,
+	       struct wsm_tx *wsm)
+{
+	u8 priority = 0;
+
+	if (!priv->bt_present)
+		return;
+
+	if (ieee80211_is_nullfunc(t->hdr->frame_control)) {
+		priority = WSM_EPTA_PRIORITY_MGT;
+	} else if (ieee80211_is_data(t->hdr->frame_control)) {
+		/* Skip LLC SNAP header (+6) */
+		u8 *payload = &t->skb->data[t->hdrlen];
+		__be16 *ethertype = (__be16 *)&payload[6];
+		if (be16_to_cpu(*ethertype) == ETH_P_PAE)
+			priority = WSM_EPTA_PRIORITY_EAPOL;
+	} else if (ieee80211_is_assoc_req(t->hdr->frame_control) ||
+		ieee80211_is_reassoc_req(t->hdr->frame_control)) {
+		struct ieee80211_mgmt *mgt_frame =
+				(struct ieee80211_mgmt *)t->hdr;
+
+		if (le16_to_cpu(mgt_frame->u.assoc_req.listen_interval) <
+						priv->listen_interval) {
+			pr_debug("Modified Listen Interval to %d from %d\n",
+				 priv->listen_interval,
+				 mgt_frame->u.assoc_req.listen_interval);
+			/* Replace listen interval derieved from
+			 * the one read from SDD
+			 */
+			mgt_frame->u.assoc_req.listen_interval = cpu_to_le16(priv->listen_interval);
+		}
+	}
+
+	if (!priority) {
+		if (ieee80211_is_action(t->hdr->frame_control))
+			priority = WSM_EPTA_PRIORITY_ACTION;
+		else if (ieee80211_is_mgmt(t->hdr->frame_control))
+			priority = WSM_EPTA_PRIORITY_MGT;
+		else if ((wsm->queue_id == WSM_QUEUE_VOICE))
+			priority = WSM_EPTA_PRIORITY_VOICE;
+		else if ((wsm->queue_id == WSM_QUEUE_VIDEO))
+			priority = WSM_EPTA_PRIORITY_VIDEO;
+		else
+			priority = WSM_EPTA_PRIORITY_DATA;
+	}
+
+	pr_debug("[TX] EPTA priority %d.\n", priority);
+
+	wsm->flags |= priority << 1;
+}
+
+static int
+cw1200_tx_h_rate_policy(struct cw1200_common *priv,
+			struct cw1200_txinfo *t,
+			struct wsm_tx *wsm)
+{
+	bool tx_policy_renew = false;
+
+	t->txpriv.rate_id = tx_policy_get(priv,
+		t->tx_info->control.rates, IEEE80211_TX_MAX_RATES,
+		&tx_policy_renew);
+	if (t->txpriv.rate_id == CW1200_INVALID_RATE_ID)
+		return -EFAULT;
+
+	wsm->flags |= t->txpriv.rate_id << 4;
+
+	t->rate = cw1200_get_tx_rate(priv,
+		&t->tx_info->control.rates[0]),
+	wsm->max_tx_rate = t->rate->hw_value;
+	if (t->rate->flags & IEEE80211_TX_RC_MCS) {
+		if (cw1200_ht_greenfield(&priv->ht_info))
+			wsm->ht_tx_parameters |=
+				__cpu_to_le32(WSM_HT_TX_GREENFIELD);
+		else
+			wsm->ht_tx_parameters |=
+				__cpu_to_le32(WSM_HT_TX_MIXED);
+	}
+
+	if (tx_policy_renew) {
+		pr_debug("[TX] TX policy renew.\n");
+		/* It's not so optimal to stop TX queues every now and then.
+		 * Better to reimplement task scheduling with
+		 * a counter. TODO.
+		 */
+		wsm_lock_tx_async(priv);
+		cw1200_tx_queues_lock(priv);
+		if (queue_work(priv->workqueue,
+			       &priv->tx_policy_upload_work) <= 0) {
+			cw1200_tx_queues_unlock(priv);
+			wsm_unlock_tx(priv);
+		}
+	}
+	return 0;
+}
+
+static bool
+cw1200_tx_h_pm_state(struct cw1200_common *priv,
+		     struct cw1200_txinfo *t)
+{
+	int was_buffered = 1;
+
+	if (t->txpriv.link_id == CW1200_LINK_ID_AFTER_DTIM &&
+	    !priv->buffered_multicasts) {
+		priv->buffered_multicasts = true;
+		if (priv->sta_asleep_mask)
+			queue_work(priv->workqueue,
+				   &priv->multicast_start_work);
+	}
+
+	if (t->txpriv.raw_link_id && t->txpriv.tid < CW1200_MAX_TID)
+		was_buffered = priv->link_id_db[t->txpriv.raw_link_id - 1].buffered[t->txpriv.tid]++;
+
+	return !was_buffered;
+}
+
+/* ******************************************************************** */
+
+void cw1200_tx(struct ieee80211_hw *dev,
+	       struct ieee80211_tx_control *control,
+	       struct sk_buff *skb)
+{
+	struct cw1200_common *priv = dev->priv;
+	struct cw1200_txinfo t = {
+		.skb = skb,
+		.queue = skb_get_queue_mapping(skb),
+		.tx_info = IEEE80211_SKB_CB(skb),
+		.hdr = (struct ieee80211_hdr *)skb->data,
+		.txpriv.tid = CW1200_MAX_TID,
+		.txpriv.rate_id = CW1200_INVALID_RATE_ID,
+	};
+	struct ieee80211_sta *sta;
+	struct wsm_tx *wsm;
+	bool tid_update = 0;
+	u8 flags = 0;
+	int ret;
+
+	if (priv->bh_error)
+		goto drop;
+
+	t.hdrlen = ieee80211_hdrlen(t.hdr->frame_control);
+	t.da = ieee80211_get_DA(t.hdr);
+	if (control) {
+		t.sta = control->sta;
+		t.sta_priv = (struct cw1200_sta_priv *)&t.sta->drv_priv;
+	}
+
+	if (WARN_ON(t.queue >= 4))
+		goto drop;
+
+	ret = cw1200_tx_h_calc_link_ids(priv, &t);
+	if (ret)
+		goto drop;
+
+	pr_debug("[TX] TX %d bytes (queue: %d, link_id: %d (%d)).\n",
+		 skb->len, t.queue, t.txpriv.link_id,
+		 t.txpriv.raw_link_id);
+
+	cw1200_tx_h_pm(priv, &t);
+	cw1200_tx_h_calc_tid(priv, &t);
+	ret = cw1200_tx_h_crypt(priv, &t);
+	if (ret)
+		goto drop;
+	ret = cw1200_tx_h_align(priv, &t, &flags);
+	if (ret)
+		goto drop;
+	ret = cw1200_tx_h_action(priv, &t);
+	if (ret)
+		goto drop;
+	wsm = cw1200_tx_h_wsm(priv, &t);
+	if (!wsm) {
+		ret = -ENOMEM;
+		goto drop;
+	}
+	wsm->flags |= flags;
+	cw1200_tx_h_bt(priv, &t, wsm);
+	ret = cw1200_tx_h_rate_policy(priv, &t, wsm);
+	if (ret)
+		goto drop;
+
+	rcu_read_lock();
+	sta = rcu_dereference(t.sta);
+
+	spin_lock_bh(&priv->ps_state_lock);
+	{
+		tid_update = cw1200_tx_h_pm_state(priv, &t);
+		BUG_ON(cw1200_queue_put(&priv->tx_queue[t.queue],
+					t.skb, &t.txpriv));
+	}
+	spin_unlock_bh(&priv->ps_state_lock);
+
+	if (tid_update && sta)
+		ieee80211_sta_set_buffered(sta, t.txpriv.tid, true);
+
+	rcu_read_unlock();
+
+	cw1200_bh_wakeup(priv);
+
+	return;
+
+drop:
+	cw1200_skb_dtor(priv, skb, &t.txpriv);
+	return;
+}
+
+/* ******************************************************************** */
+
+static int cw1200_handle_action_rx(struct cw1200_common *priv,
+				   struct sk_buff *skb)
+{
+	struct ieee80211_mgmt *mgmt = (void *)skb->data;
+
+	/* Filter block ACK negotiation: fully controlled by firmware */
+	if (mgmt->u.action.category == WLAN_CATEGORY_BACK)
+		return 1;
+
+	return 0;
+}
+
+static int cw1200_handle_pspoll(struct cw1200_common *priv,
+				struct sk_buff *skb)
+{
+	struct ieee80211_sta *sta;
+	struct ieee80211_pspoll *pspoll = (struct ieee80211_pspoll *)skb->data;
+	int link_id = 0;
+	u32 pspoll_mask = 0;
+	int drop = 1;
+	int i;
+
+	if (priv->join_status != CW1200_JOIN_STATUS_AP)
+		goto done;
+	if (memcmp(priv->vif->addr, pspoll->bssid, ETH_ALEN))
+		goto done;
+
+	rcu_read_lock();
+	sta = ieee80211_find_sta(priv->vif, pspoll->ta);
+	if (sta) {
+		struct cw1200_sta_priv *sta_priv;
+		sta_priv = (struct cw1200_sta_priv *)&sta->drv_priv;
+		link_id = sta_priv->link_id;
+		pspoll_mask = BIT(sta_priv->link_id);
+	}
+	rcu_read_unlock();
+	if (!link_id)
+		goto done;
+
+	priv->pspoll_mask |= pspoll_mask;
+	drop = 0;
+
+	/* Do not report pspols if data for given link id is queued already. */
+	for (i = 0; i < 4; ++i) {
+		if (cw1200_queue_get_num_queued(&priv->tx_queue[i],
+						pspoll_mask)) {
+			cw1200_bh_wakeup(priv);
+			drop = 1;
+			break;
+		}
+	}
+	pr_debug("[RX] PSPOLL: %s\n", drop ? "local" : "fwd");
+done:
+	return drop;
+}
+
+/* ******************************************************************** */
+
+void cw1200_tx_confirm_cb(struct cw1200_common *priv,
+			  int link_id,
+			  struct wsm_tx_confirm *arg)
+{
+	u8 queue_id = cw1200_queue_get_queue_id(arg->packet_id);
+	struct cw1200_queue *queue = &priv->tx_queue[queue_id];
+	struct sk_buff *skb;
+	const struct cw1200_txpriv *txpriv;
+
+	pr_debug("[TX] TX confirm: %d, %d.\n",
+		 arg->status, arg->ack_failures);
+
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED) {
+		/* STA is stopped. */
+		return;
+	}
+
+	if (WARN_ON(queue_id >= 4))
+		return;
+
+	if (arg->status)
+		pr_debug("TX failed: %d.\n", arg->status);
+
+	if ((arg->status == WSM_REQUEUE) &&
+	    (arg->flags & WSM_TX_STATUS_REQUEUE)) {
+		/* "Requeue" means "implicit suspend" */
+		struct wsm_suspend_resume suspend = {
+			.link_id = link_id,
+			.stop = 1,
+			.multicast = !link_id,
+		};
+		cw1200_suspend_resume(priv, &suspend);
+		wiphy_warn(priv->hw->wiphy, "Requeue for link_id %d (try %d). STAs asleep: 0x%.8X\n",
+			   link_id,
+			   cw1200_queue_get_generation(arg->packet_id) + 1,
+			   priv->sta_asleep_mask);
+		cw1200_queue_requeue(queue, arg->packet_id);
+		spin_lock_bh(&priv->ps_state_lock);
+		if (!link_id) {
+			priv->buffered_multicasts = true;
+			if (priv->sta_asleep_mask) {
+				queue_work(priv->workqueue,
+					   &priv->multicast_start_work);
+			}
+		}
+		spin_unlock_bh(&priv->ps_state_lock);
+	} else if (!cw1200_queue_get_skb(queue, arg->packet_id,
+					 &skb, &txpriv)) {
+		struct ieee80211_tx_info *tx = IEEE80211_SKB_CB(skb);
+		int tx_count = arg->ack_failures;
+		u8 ht_flags = 0;
+		int i;
+
+		if (cw1200_ht_greenfield(&priv->ht_info))
+			ht_flags |= IEEE80211_TX_RC_GREEN_FIELD;
+
+		spin_lock(&priv->bss_loss_lock);
+		if (priv->bss_loss_state &&
+		    arg->packet_id == priv->bss_loss_confirm_id) {
+			if (arg->status) {
+				/* Recovery failed */
+				__cw1200_cqm_bssloss_sm(priv, 0, 0, 1);
+			} else {
+				/* Recovery succeeded */
+				__cw1200_cqm_bssloss_sm(priv, 0, 1, 0);
+			}
+		}
+		spin_unlock(&priv->bss_loss_lock);
+
+		if (!arg->status) {
+			tx->flags |= IEEE80211_TX_STAT_ACK;
+			++tx_count;
+			cw1200_debug_txed(priv);
+			if (arg->flags & WSM_TX_STATUS_AGGREGATION) {
+				/* Do not report aggregation to mac80211:
+				 * it confuses minstrel a lot.
+				 */
+				/* tx->flags |= IEEE80211_TX_STAT_AMPDU; */
+				cw1200_debug_txed_agg(priv);
+			}
+		} else {
+			if (tx_count)
+				++tx_count;
+		}
+
+		for (i = 0; i < IEEE80211_TX_MAX_RATES; ++i) {
+			if (tx->status.rates[i].count >= tx_count) {
+				tx->status.rates[i].count = tx_count;
+				break;
+			}
+			tx_count -= tx->status.rates[i].count;
+			if (tx->status.rates[i].flags & IEEE80211_TX_RC_MCS)
+				tx->status.rates[i].flags |= ht_flags;
+		}
+
+		for (++i; i < IEEE80211_TX_MAX_RATES; ++i) {
+			tx->status.rates[i].count = 0;
+			tx->status.rates[i].idx = -1;
+		}
+
+		/* Pull off any crypto trailers that we added on */
+		if (tx->control.hw_key) {
+			skb_trim(skb, skb->len - tx->control.hw_key->icv_len);
+			if (tx->control.hw_key->cipher == WLAN_CIPHER_SUITE_TKIP)
+				skb_trim(skb, skb->len - 8); /* MIC space */
+		}
+		cw1200_queue_remove(queue, arg->packet_id);
+	}
+	/* XXX TODO:  Only wake if there are pending transmits.. */
+	cw1200_bh_wakeup(priv);
+}
+
+static void cw1200_notify_buffered_tx(struct cw1200_common *priv,
+			       struct sk_buff *skb, int link_id, int tid)
+{
+	struct ieee80211_sta *sta;
+	struct ieee80211_hdr *hdr;
+	u8 *buffered;
+	u8 still_buffered = 0;
+
+	if (link_id && tid < CW1200_MAX_TID) {
+		buffered = priv->link_id_db
+				[link_id - 1].buffered;
+
+		spin_lock_bh(&priv->ps_state_lock);
+		if (!WARN_ON(!buffered[tid]))
+			still_buffered = --buffered[tid];
+		spin_unlock_bh(&priv->ps_state_lock);
+
+		if (!still_buffered && tid < CW1200_MAX_TID) {
+			hdr = (struct ieee80211_hdr *)skb->data;
+			rcu_read_lock();
+			sta = ieee80211_find_sta(priv->vif, hdr->addr1);
+			if (sta)
+				ieee80211_sta_set_buffered(sta, tid, false);
+			rcu_read_unlock();
+		}
+	}
+}
+
+void cw1200_skb_dtor(struct cw1200_common *priv,
+		     struct sk_buff *skb,
+		     const struct cw1200_txpriv *txpriv)
+{
+	skb_pull(skb, txpriv->offset);
+	if (txpriv->rate_id != CW1200_INVALID_RATE_ID) {
+		cw1200_notify_buffered_tx(priv, skb,
+					  txpriv->raw_link_id, txpriv->tid);
+		tx_policy_put(priv, txpriv->rate_id);
+	}
+	ieee80211_tx_status(priv->hw, skb);
+}
+
+void cw1200_rx_cb(struct cw1200_common *priv,
+		  struct wsm_rx *arg,
+		  int link_id,
+		  struct sk_buff **skb_p)
+{
+	struct sk_buff *skb = *skb_p;
+	struct ieee80211_rx_status *hdr = IEEE80211_SKB_RXCB(skb);
+	struct ieee80211_hdr *frame = (struct ieee80211_hdr *)skb->data;
+	struct ieee80211_mgmt *mgmt = (struct ieee80211_mgmt *)skb->data;
+	struct cw1200_link_entry *entry = NULL;
+	unsigned long grace_period;
+
+	bool early_data = false;
+	bool p2p = priv->vif && priv->vif->p2p;
+	size_t hdrlen;
+	hdr->flag = 0;
+
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED) {
+		/* STA is stopped. */
+		goto drop;
+	}
+
+	if (link_id && link_id <= CW1200_MAX_STA_IN_AP_MODE) {
+		entry =	&priv->link_id_db[link_id - 1];
+		if (entry->status == CW1200_LINK_SOFT &&
+		    ieee80211_is_data(frame->frame_control))
+			early_data = true;
+		entry->timestamp = jiffies;
+	} else if (p2p &&
+		   ieee80211_is_action(frame->frame_control) &&
+		   (mgmt->u.action.category == WLAN_CATEGORY_PUBLIC)) {
+		pr_debug("[RX] Going to MAP&RESET link ID\n");
+		WARN_ON(work_pending(&priv->linkid_reset_work));
+		memcpy(&priv->action_frame_sa[0],
+		       ieee80211_get_SA(frame), ETH_ALEN);
+		priv->action_linkid = 0;
+		schedule_work(&priv->linkid_reset_work);
+	}
+
+	if (link_id && p2p &&
+	    ieee80211_is_action(frame->frame_control) &&
+	    (mgmt->u.action.category == WLAN_CATEGORY_PUBLIC)) {
+		/* Link ID already exists for the ACTION frame.
+		 * Reset and Remap
+		 */
+		WARN_ON(work_pending(&priv->linkid_reset_work));
+		memcpy(&priv->action_frame_sa[0],
+		       ieee80211_get_SA(frame), ETH_ALEN);
+		priv->action_linkid = link_id;
+		schedule_work(&priv->linkid_reset_work);
+	}
+	if (arg->status) {
+		if (arg->status == WSM_STATUS_MICFAILURE) {
+			pr_debug("[RX] MIC failure.\n");
+			hdr->flag |= RX_FLAG_MMIC_ERROR;
+		} else if (arg->status == WSM_STATUS_NO_KEY_FOUND) {
+			pr_debug("[RX] No key found.\n");
+			goto drop;
+		} else {
+			pr_debug("[RX] Receive failure: %d.\n",
+				 arg->status);
+			goto drop;
+		}
+	}
+
+	if (skb->len < sizeof(struct ieee80211_pspoll)) {
+		wiphy_warn(priv->hw->wiphy, "Mailformed SDU rx'ed. Size is lesser than IEEE header.\n");
+		goto drop;
+	}
+
+	if (ieee80211_is_pspoll(frame->frame_control))
+		if (cw1200_handle_pspoll(priv, skb))
+			goto drop;
+
+	hdr->band = ((arg->channel_number & 0xff00) ||
+		     (arg->channel_number > 14)) ?
+			IEEE80211_BAND_5GHZ : IEEE80211_BAND_2GHZ;
+	hdr->freq = ieee80211_channel_to_frequency(
+			arg->channel_number,
+			hdr->band);
+
+	if (arg->rx_rate >= 14) {
+		hdr->flag |= RX_FLAG_HT;
+		hdr->rate_idx = arg->rx_rate - 14;
+	} else if (arg->rx_rate >= 4) {
+		hdr->rate_idx = arg->rx_rate - 2;
+	} else {
+		hdr->rate_idx = arg->rx_rate;
+	}
+
+	hdr->signal = (s8)arg->rcpi_rssi;
+	hdr->antenna = 0;
+
+	hdrlen = ieee80211_hdrlen(frame->frame_control);
+
+	if (WSM_RX_STATUS_ENCRYPTION(arg->flags)) {
+		size_t iv_len = 0, icv_len = 0;
+
+		hdr->flag |= RX_FLAG_DECRYPTED | RX_FLAG_IV_STRIPPED;
+
+		/* Oops... There is no fast way to ask mac80211 about
+		 * IV/ICV lengths. Even defineas are not exposed.
+		 */
+		switch (WSM_RX_STATUS_ENCRYPTION(arg->flags)) {
+		case WSM_RX_STATUS_WEP:
+			iv_len = 4 /* WEP_IV_LEN */;
+			icv_len = 4 /* WEP_ICV_LEN */;
+			break;
+		case WSM_RX_STATUS_TKIP:
+			iv_len = 8 /* TKIP_IV_LEN */;
+			icv_len = 4 /* TKIP_ICV_LEN */
+				+ 8 /*MICHAEL_MIC_LEN*/;
+			hdr->flag |= RX_FLAG_MMIC_STRIPPED;
+			break;
+		case WSM_RX_STATUS_AES:
+			iv_len = 8 /* CCMP_HDR_LEN */;
+			icv_len = 8 /* CCMP_MIC_LEN */;
+			break;
+		case WSM_RX_STATUS_WAPI:
+			iv_len = 18 /* WAPI_HDR_LEN */;
+			icv_len = 16 /* WAPI_MIC_LEN */;
+			break;
+		default:
+			pr_warn("Unknown encryption type %d\n",
+				WSM_RX_STATUS_ENCRYPTION(arg->flags));
+			goto drop;
+		}
+
+		/* Firmware strips ICV in case of MIC failure. */
+		if (arg->status == WSM_STATUS_MICFAILURE)
+			icv_len = 0;
+
+		if (skb->len < hdrlen + iv_len + icv_len) {
+			wiphy_warn(priv->hw->wiphy, "Malformed SDU rx'ed. Size is lesser than crypto headers.\n");
+			goto drop;
+		}
+
+		/* Remove IV, ICV and MIC */
+		skb_trim(skb, skb->len - icv_len);
+		memmove(skb->data + iv_len, skb->data, hdrlen);
+		skb_pull(skb, iv_len);
+	}
+
+	/* Remove TSF from the end of frame */
+	if (arg->flags & WSM_RX_STATUS_TSF_INCLUDED) {
+		memcpy(&hdr->mactime, skb->data + skb->len - 8, 8);
+		hdr->mactime = le64_to_cpu(hdr->mactime);
+		if (skb->len >= 8)
+			skb_trim(skb, skb->len - 8);
+	} else {
+		hdr->mactime = 0;
+	}
+
+	cw1200_debug_rxed(priv);
+	if (arg->flags & WSM_RX_STATUS_AGGREGATE)
+		cw1200_debug_rxed_agg(priv);
+
+	if (ieee80211_is_action(frame->frame_control) &&
+	    (arg->flags & WSM_RX_STATUS_ADDRESS1)) {
+		if (cw1200_handle_action_rx(priv, skb))
+			return;
+	} else if (ieee80211_is_beacon(frame->frame_control) &&
+		   !arg->status &&
+		   !memcmp(ieee80211_get_SA(frame), priv->vif->bss_conf.bssid,
+			   ETH_ALEN)) {
+		const u8 *tim_ie;
+		u8 *ies = ((struct ieee80211_mgmt *)
+			  (skb->data))->u.beacon.variable;
+		size_t ies_len = skb->len - (ies - (u8 *)(skb->data));
+
+		tim_ie = cfg80211_find_ie(WLAN_EID_TIM, ies, ies_len);
+		if (tim_ie) {
+			struct ieee80211_tim_ie *tim =
+				(struct ieee80211_tim_ie *)&tim_ie[2];
+
+			if (priv->join_dtim_period != tim->dtim_period) {
+				priv->join_dtim_period = tim->dtim_period;
+				queue_work(priv->workqueue,
+					   &priv->set_beacon_wakeup_period_work);
+			}
+		}
+
+		/* Disable beacon filter once we're associated... */
+		if (priv->disable_beacon_filter &&
+		    (priv->vif->bss_conf.assoc ||
+		     priv->vif->bss_conf.ibss_joined)) {
+			priv->disable_beacon_filter = false;
+			queue_work(priv->workqueue,
+				   &priv->update_filtering_work);
+		}
+	}
+
+	/* Stay awake after frame is received to give
+	 * userspace chance to react and acquire appropriate
+	 * wakelock.
+	 */
+	if (ieee80211_is_auth(frame->frame_control))
+		grace_period = 5 * HZ;
+	else if (ieee80211_is_deauth(frame->frame_control))
+		grace_period = 5 * HZ;
+	else
+		grace_period = 1 * HZ;
+	cw1200_pm_stay_awake(&priv->pm_state, grace_period);
+
+	if (early_data) {
+		spin_lock_bh(&priv->ps_state_lock);
+		/* Double-check status with lock held */
+		if (entry->status == CW1200_LINK_SOFT)
+			skb_queue_tail(&entry->rx_queue, skb);
+		else
+			ieee80211_rx_irqsafe(priv->hw, skb);
+		spin_unlock_bh(&priv->ps_state_lock);
+	} else {
+		ieee80211_rx_irqsafe(priv->hw, skb);
+	}
+	*skb_p = NULL;
+
+	return;
+
+drop:
+	/* TODO: update failure counters */
+	return;
+}
+
+/* ******************************************************************** */
+/* Security								*/
+
+int cw1200_alloc_key(struct cw1200_common *priv)
+{
+	int idx;
+
+	idx = ffs(~priv->key_map) - 1;
+	if (idx < 0 || idx > WSM_KEY_MAX_INDEX)
+		return -1;
+
+	priv->key_map |= BIT(idx);
+	priv->keys[idx].index = idx;
+	return idx;
+}
+
+void cw1200_free_key(struct cw1200_common *priv, int idx)
+{
+	BUG_ON(!(priv->key_map & BIT(idx)));
+	memset(&priv->keys[idx], 0, sizeof(priv->keys[idx]));
+	priv->key_map &= ~BIT(idx);
+}
+
+void cw1200_free_keys(struct cw1200_common *priv)
+{
+	memset(&priv->keys, 0, sizeof(priv->keys));
+	priv->key_map = 0;
+}
+
+int cw1200_upload_keys(struct cw1200_common *priv)
+{
+	int idx, ret = 0;
+	for (idx = 0; idx <= WSM_KEY_MAX_INDEX; ++idx)
+		if (priv->key_map & BIT(idx)) {
+			ret = wsm_add_key(priv, &priv->keys[idx]);
+			if (ret < 0)
+				break;
+		}
+	return ret;
+}
+
+/* Workaround for WFD test case 6.1.10 */
+void cw1200_link_id_reset(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, linkid_reset_work);
+	int temp_linkid;
+
+	if (!priv->action_linkid) {
+		/* In GO mode we can receive ACTION frames without a linkID */
+		temp_linkid = cw1200_alloc_link_id(priv,
+				&priv->action_frame_sa[0]);
+		WARN_ON(!temp_linkid);
+		if (temp_linkid) {
+			/* Make sure we execute the WQ */
+			flush_workqueue(priv->workqueue);
+			/* Release the link ID */
+			spin_lock_bh(&priv->ps_state_lock);
+			priv->link_id_db[temp_linkid - 1].prev_status =
+				priv->link_id_db[temp_linkid - 1].status;
+			priv->link_id_db[temp_linkid - 1].status =
+				CW1200_LINK_RESET;
+			spin_unlock_bh(&priv->ps_state_lock);
+			wsm_lock_tx_async(priv);
+			if (queue_work(priv->workqueue,
+				       &priv->link_id_work) <= 0)
+				wsm_unlock_tx(priv);
+		}
+	} else {
+		spin_lock_bh(&priv->ps_state_lock);
+		priv->link_id_db[priv->action_linkid - 1].prev_status =
+			priv->link_id_db[priv->action_linkid - 1].status;
+		priv->link_id_db[priv->action_linkid - 1].status =
+			CW1200_LINK_RESET_REMAP;
+		spin_unlock_bh(&priv->ps_state_lock);
+		wsm_lock_tx_async(priv);
+		if (queue_work(priv->workqueue, &priv->link_id_work) <= 0)
+			wsm_unlock_tx(priv);
+		flush_workqueue(priv->workqueue);
+	}
+}
+
+int cw1200_find_link_id(struct cw1200_common *priv, const u8 *mac)
+{
+	int i, ret = 0;
+	spin_lock_bh(&priv->ps_state_lock);
+	for (i = 0; i < CW1200_MAX_STA_IN_AP_MODE; ++i) {
+		if (!memcmp(mac, priv->link_id_db[i].mac, ETH_ALEN) &&
+		    priv->link_id_db[i].status) {
+			priv->link_id_db[i].timestamp = jiffies;
+			ret = i + 1;
+			break;
+		}
+	}
+	spin_unlock_bh(&priv->ps_state_lock);
+	return ret;
+}
+
+int cw1200_alloc_link_id(struct cw1200_common *priv, const u8 *mac)
+{
+	int i, ret = 0;
+	unsigned long max_inactivity = 0;
+	unsigned long now = jiffies;
+
+	spin_lock_bh(&priv->ps_state_lock);
+	for (i = 0; i < CW1200_MAX_STA_IN_AP_MODE; ++i) {
+		if (!priv->link_id_db[i].status) {
+			ret = i + 1;
+			break;
+		} else if (priv->link_id_db[i].status != CW1200_LINK_HARD &&
+			   !priv->tx_queue_stats.link_map_cache[i + 1]) {
+			unsigned long inactivity =
+				now - priv->link_id_db[i].timestamp;
+			if (inactivity < max_inactivity)
+				continue;
+			max_inactivity = inactivity;
+			ret = i + 1;
+		}
+	}
+	if (ret) {
+		struct cw1200_link_entry *entry = &priv->link_id_db[ret - 1];
+		pr_debug("[AP] STA added, link_id: %d\n", ret);
+		entry->status = CW1200_LINK_RESERVE;
+		memcpy(&entry->mac, mac, ETH_ALEN);
+		memset(&entry->buffered, 0, CW1200_MAX_TID);
+		skb_queue_head_init(&entry->rx_queue);
+		wsm_lock_tx_async(priv);
+		if (queue_work(priv->workqueue, &priv->link_id_work) <= 0)
+			wsm_unlock_tx(priv);
+	} else {
+		wiphy_info(priv->hw->wiphy,
+			   "[AP] Early: no more link IDs available.\n");
+	}
+
+	spin_unlock_bh(&priv->ps_state_lock);
+	return ret;
+}
+
+void cw1200_link_id_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, link_id_work);
+	wsm_flush_tx(priv);
+	cw1200_link_id_gc_work(&priv->link_id_gc_work.work);
+	wsm_unlock_tx(priv);
+}
+
+void cw1200_link_id_gc_work(struct work_struct *work)
+{
+	struct cw1200_common *priv =
+		container_of(work, struct cw1200_common, link_id_gc_work.work);
+	struct wsm_reset reset = {
+		.reset_statistics = false,
+	};
+	struct wsm_map_link map_link = {
+		.link_id = 0,
+	};
+	unsigned long now = jiffies;
+	unsigned long next_gc = -1;
+	long ttl;
+	bool need_reset;
+	u32 mask;
+	int i;
+
+	if (priv->join_status != CW1200_JOIN_STATUS_AP)
+		return;
+
+	wsm_lock_tx(priv);
+	spin_lock_bh(&priv->ps_state_lock);
+	for (i = 0; i < CW1200_MAX_STA_IN_AP_MODE; ++i) {
+		need_reset = false;
+		mask = BIT(i + 1);
+		if (priv->link_id_db[i].status == CW1200_LINK_RESERVE ||
+		    (priv->link_id_db[i].status == CW1200_LINK_HARD &&
+		     !(priv->link_id_map & mask))) {
+			if (priv->link_id_map & mask) {
+				priv->sta_asleep_mask &= ~mask;
+				priv->pspoll_mask &= ~mask;
+				need_reset = true;
+			}
+			priv->link_id_map |= mask;
+			if (priv->link_id_db[i].status != CW1200_LINK_HARD)
+				priv->link_id_db[i].status = CW1200_LINK_SOFT;
+			memcpy(map_link.mac_addr, priv->link_id_db[i].mac,
+			       ETH_ALEN);
+			spin_unlock_bh(&priv->ps_state_lock);
+			if (need_reset) {
+				reset.link_id = i + 1;
+				wsm_reset(priv, &reset);
+			}
+			map_link.link_id = i + 1;
+			wsm_map_link(priv, &map_link);
+			next_gc = min(next_gc, CW1200_LINK_ID_GC_TIMEOUT);
+			spin_lock_bh(&priv->ps_state_lock);
+		} else if (priv->link_id_db[i].status == CW1200_LINK_SOFT) {
+			ttl = priv->link_id_db[i].timestamp - now +
+					CW1200_LINK_ID_GC_TIMEOUT;
+			if (ttl <= 0) {
+				need_reset = true;
+				priv->link_id_db[i].status = CW1200_LINK_OFF;
+				priv->link_id_map &= ~mask;
+				priv->sta_asleep_mask &= ~mask;
+				priv->pspoll_mask &= ~mask;
+				memset(map_link.mac_addr, 0, ETH_ALEN);
+				spin_unlock_bh(&priv->ps_state_lock);
+				reset.link_id = i + 1;
+				wsm_reset(priv, &reset);
+				spin_lock_bh(&priv->ps_state_lock);
+			} else {
+				next_gc = min_t(unsigned long, next_gc, ttl);
+			}
+		} else if (priv->link_id_db[i].status == CW1200_LINK_RESET ||
+				priv->link_id_db[i].status ==
+				CW1200_LINK_RESET_REMAP) {
+			int status = priv->link_id_db[i].status;
+			priv->link_id_db[i].status =
+					priv->link_id_db[i].prev_status;
+			priv->link_id_db[i].timestamp = now;
+			reset.link_id = i + 1;
+			spin_unlock_bh(&priv->ps_state_lock);
+			wsm_reset(priv, &reset);
+			if (status == CW1200_LINK_RESET_REMAP) {
+				memcpy(map_link.mac_addr,
+				       priv->link_id_db[i].mac,
+				       ETH_ALEN);
+				map_link.link_id = i + 1;
+				wsm_map_link(priv, &map_link);
+				next_gc = min(next_gc,
+						CW1200_LINK_ID_GC_TIMEOUT);
+			}
+			spin_lock_bh(&priv->ps_state_lock);
+		}
+		if (need_reset) {
+			skb_queue_purge(&priv->link_id_db[i].rx_queue);
+			pr_debug("[AP] STA removed, link_id: %d\n",
+				 reset.link_id);
+		}
+	}
+	spin_unlock_bh(&priv->ps_state_lock);
+	if (next_gc != -1)
+		queue_delayed_work(priv->workqueue,
+				   &priv->link_id_gc_work, next_gc);
+	wsm_unlock_tx(priv);
+}
diff --git a/drivers/net/wireless/cw1200/txrx.h b/drivers/net/wireless/cw1200/txrx.h
new file mode 100644
index 0000000..492a4e1
--- /dev/null
+++ b/drivers/net/wireless/cw1200/txrx.h
@@ -0,0 +1,106 @@
+/*
+ * Datapath interface for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_TXRX_H
+#define CW1200_TXRX_H
+
+#include <linux/list.h>
+
+/* extern */ struct ieee80211_hw;
+/* extern */ struct sk_buff;
+/* extern */ struct wsm_tx;
+/* extern */ struct wsm_rx;
+/* extern */ struct wsm_tx_confirm;
+/* extern */ struct cw1200_txpriv;
+
+struct tx_policy {
+	union {
+		__le32 tbl[3];
+		u8 raw[12];
+	};
+	u8  defined;
+	u8  usage_count;
+	u8  retry_count;
+	u8  uploaded;
+};
+
+struct tx_policy_cache_entry {
+	struct tx_policy policy;
+	struct list_head link;
+};
+
+#define TX_POLICY_CACHE_SIZE	(8)
+struct tx_policy_cache {
+	struct tx_policy_cache_entry cache[TX_POLICY_CACHE_SIZE];
+	struct list_head used;
+	struct list_head free;
+	spinlock_t lock; /* Protect policy cache */
+};
+
+/* ******************************************************************** */
+/* TX policy cache							*/
+/* Intention of TX policy cache is an overcomplicated WSM API.
+ * Device does not accept per-PDU tx retry sequence.
+ * It uses "tx retry policy id" instead, so driver code has to sync
+ * linux tx retry sequences with a retry policy table in the device.
+ */
+void tx_policy_init(struct cw1200_common *priv);
+void tx_policy_upload_work(struct work_struct *work);
+void tx_policy_clean(struct cw1200_common *priv);
+
+/* ******************************************************************** */
+/* TX implementation							*/
+
+u32 cw1200_rate_mask_to_wsm(struct cw1200_common *priv,
+			       u32 rates);
+void cw1200_tx(struct ieee80211_hw *dev,
+	       struct ieee80211_tx_control *control,
+	       struct sk_buff *skb);
+void cw1200_skb_dtor(struct cw1200_common *priv,
+		     struct sk_buff *skb,
+		     const struct cw1200_txpriv *txpriv);
+
+/* ******************************************************************** */
+/* WSM callbacks							*/
+
+void cw1200_tx_confirm_cb(struct cw1200_common *priv,
+			  int link_id,
+			  struct wsm_tx_confirm *arg);
+void cw1200_rx_cb(struct cw1200_common *priv,
+		  struct wsm_rx *arg,
+		  int link_id,
+		  struct sk_buff **skb_p);
+
+/* ******************************************************************** */
+/* Timeout								*/
+
+void cw1200_tx_timeout(struct work_struct *work);
+
+/* ******************************************************************** */
+/* Security								*/
+int cw1200_alloc_key(struct cw1200_common *priv);
+void cw1200_free_key(struct cw1200_common *priv, int idx);
+void cw1200_free_keys(struct cw1200_common *priv);
+int cw1200_upload_keys(struct cw1200_common *priv);
+
+/* ******************************************************************** */
+/* Workaround for WFD test case 6.1.10					*/
+void cw1200_link_id_reset(struct work_struct *work);
+
+#define CW1200_LINK_ID_GC_TIMEOUT ((unsigned long)(10 * HZ))
+
+int cw1200_find_link_id(struct cw1200_common *priv, const u8 *mac);
+int cw1200_alloc_link_id(struct cw1200_common *priv, const u8 *mac);
+void cw1200_link_id_work(struct work_struct *work);
+void cw1200_link_id_gc_work(struct work_struct *work);
+
+
+#endif /* CW1200_TXRX_H */
diff --git a/drivers/net/wireless/cw1200/wsm.c b/drivers/net/wireless/cw1200/wsm.c
new file mode 100644
index 0000000..cbb74d7
--- /dev/null
+++ b/drivers/net/wireless/cw1200/wsm.c
@@ -0,0 +1,1822 @@
+/*
+ * WSM host interface (HI) implementation for
+ * ST-Ericsson CW1200 mac80211 drivers.
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/skbuff.h>
+#include <linux/wait.h>
+#include <linux/delay.h>
+#include <linux/sched.h>
+#include <linux/random.h>
+
+#include "cw1200.h"
+#include "wsm.h"
+#include "bh.h"
+#include "sta.h"
+#include "debug.h"
+
+#define WSM_CMD_TIMEOUT		(2 * HZ) /* With respect to interrupt loss */
+#define WSM_CMD_START_TIMEOUT	(7 * HZ)
+#define WSM_CMD_RESET_TIMEOUT	(3 * HZ) /* 2 sec. timeout was observed.   */
+#define WSM_CMD_MAX_TIMEOUT	(3 * HZ)
+
+#define WSM_SKIP(buf, size)						\
+	do {								\
+		if ((buf)->data + size > (buf)->end)			\
+			goto underflow;					\
+		(buf)->data += size;					\
+	} while (0)
+
+#define WSM_GET(buf, ptr, size)						\
+	do {								\
+		if ((buf)->data + size > (buf)->end)			\
+			goto underflow;					\
+		memcpy(ptr, (buf)->data, size);				\
+		(buf)->data += size;					\
+	} while (0)
+
+#define __WSM_GET(buf, type, type2, cvt)				\
+	({								\
+		type val;						\
+		if ((buf)->data + sizeof(type) > (buf)->end)		\
+			goto underflow;					\
+		val = cvt(*(type2 *)(buf)->data);			\
+		(buf)->data += sizeof(type);				\
+		val;							\
+	})
+
+#define WSM_GET8(buf)  __WSM_GET(buf, u8, u8, (u8))
+#define WSM_GET16(buf) __WSM_GET(buf, u16, __le16, __le16_to_cpu)
+#define WSM_GET32(buf) __WSM_GET(buf, u32, __le32, __le32_to_cpu)
+
+#define WSM_PUT(buf, ptr, size)						\
+	do {								\
+		if ((buf)->data + size > (buf)->end)		\
+			if (wsm_buf_reserve((buf), size))	\
+				goto nomem;				\
+		memcpy((buf)->data, ptr, size);				\
+		(buf)->data += size;					\
+	} while (0)
+
+#define __WSM_PUT(buf, val, type, type2, cvt)				\
+	do {								\
+		if ((buf)->data + sizeof(type) > (buf)->end)		\
+			if (wsm_buf_reserve((buf), sizeof(type))) \
+				goto nomem;				\
+		*(type2 *)(buf)->data = cvt(val);			\
+		(buf)->data += sizeof(type);				\
+	} while (0)
+
+#define WSM_PUT8(buf, val)  __WSM_PUT(buf, val, u8, u8, (u8))
+#define WSM_PUT16(buf, val) __WSM_PUT(buf, val, u16, __le16, __cpu_to_le16)
+#define WSM_PUT32(buf, val) __WSM_PUT(buf, val, u32, __le32, __cpu_to_le32)
+
+static void wsm_buf_reset(struct wsm_buf *buf);
+static int wsm_buf_reserve(struct wsm_buf *buf, size_t extra_size);
+
+static int wsm_cmd_send(struct cw1200_common *priv,
+			struct wsm_buf *buf,
+			void *arg, u16 cmd, long tmo);
+
+#define wsm_cmd_lock(__priv) mutex_lock(&((__priv)->wsm_cmd_mux))
+#define wsm_cmd_unlock(__priv) mutex_unlock(&((__priv)->wsm_cmd_mux))
+
+/* ******************************************************************** */
+/* WSM API implementation						*/
+
+static int wsm_generic_confirm(struct cw1200_common *priv,
+			     void *arg,
+			     struct wsm_buf *buf)
+{
+	u32 status = WSM_GET32(buf);
+	if (status != WSM_STATUS_SUCCESS)
+		return -EINVAL;
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+int wsm_configuration(struct cw1200_common *priv, struct wsm_configuration *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT32(buf, arg->dot11MaxTransmitMsduLifeTime);
+	WSM_PUT32(buf, arg->dot11MaxReceiveLifeTime);
+	WSM_PUT32(buf, arg->dot11RtsThreshold);
+
+	/* DPD block. */
+	WSM_PUT16(buf, arg->dpdData_size + 12);
+	WSM_PUT16(buf, 1); /* DPD version */
+	WSM_PUT(buf, arg->dot11StationId, ETH_ALEN);
+	WSM_PUT16(buf, 5); /* DPD flags */
+	WSM_PUT(buf, arg->dpdData, arg->dpdData_size);
+
+	ret = wsm_cmd_send(priv, buf, arg,
+			   WSM_CONFIGURATION_REQ_ID, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+static int wsm_configuration_confirm(struct cw1200_common *priv,
+				     struct wsm_configuration *arg,
+				     struct wsm_buf *buf)
+{
+	int i;
+	int status;
+
+	status = WSM_GET32(buf);
+	if (WARN_ON(status != WSM_STATUS_SUCCESS))
+		return -EINVAL;
+
+	WSM_GET(buf, arg->dot11StationId, ETH_ALEN);
+	arg->dot11FrequencyBandsSupported = WSM_GET8(buf);
+	WSM_SKIP(buf, 1);
+	arg->supportedRateMask = WSM_GET32(buf);
+	for (i = 0; i < 2; ++i) {
+		arg->txPowerRange[i].min_power_level = WSM_GET32(buf);
+		arg->txPowerRange[i].max_power_level = WSM_GET32(buf);
+		arg->txPowerRange[i].stepping = WSM_GET32(buf);
+	}
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+/* ******************************************************************** */
+
+int wsm_reset(struct cw1200_common *priv, const struct wsm_reset *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	u16 cmd = WSM_RESET_REQ_ID | WSM_TX_LINK_ID(arg->link_id);
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT32(buf, arg->reset_statistics ? 0 : 1);
+	ret = wsm_cmd_send(priv, buf, NULL, cmd, WSM_CMD_RESET_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+struct wsm_mib {
+	u16 mib_id;
+	void *buf;
+	size_t buf_size;
+};
+
+int wsm_read_mib(struct cw1200_common *priv, u16 mib_id, void *_buf,
+			size_t buf_size)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	struct wsm_mib mib_buf = {
+		.mib_id = mib_id,
+		.buf = _buf,
+		.buf_size = buf_size,
+	};
+	wsm_cmd_lock(priv);
+
+	WSM_PUT16(buf, mib_id);
+	WSM_PUT16(buf, 0);
+
+	ret = wsm_cmd_send(priv, buf, &mib_buf,
+			   WSM_READ_MIB_REQ_ID, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+static int wsm_read_mib_confirm(struct cw1200_common *priv,
+				struct wsm_mib *arg,
+				struct wsm_buf *buf)
+{
+	u16 size;
+	if (WARN_ON(WSM_GET32(buf) != WSM_STATUS_SUCCESS))
+		return -EINVAL;
+
+	if (WARN_ON(WSM_GET16(buf) != arg->mib_id))
+		return -EINVAL;
+
+	size = WSM_GET16(buf);
+	if (size > arg->buf_size)
+		size = arg->buf_size;
+
+	WSM_GET(buf, arg->buf, size);
+	arg->buf_size = size;
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+/* ******************************************************************** */
+
+int wsm_write_mib(struct cw1200_common *priv, u16 mib_id, void *_buf,
+			size_t buf_size)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	struct wsm_mib mib_buf = {
+		.mib_id = mib_id,
+		.buf = _buf,
+		.buf_size = buf_size,
+	};
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT16(buf, mib_id);
+	WSM_PUT16(buf, buf_size);
+	WSM_PUT(buf, _buf, buf_size);
+
+	ret = wsm_cmd_send(priv, buf, &mib_buf,
+			   WSM_WRITE_MIB_REQ_ID, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+static int wsm_write_mib_confirm(struct cw1200_common *priv,
+				struct wsm_mib *arg,
+				struct wsm_buf *buf)
+{
+	int ret;
+
+	ret = wsm_generic_confirm(priv, arg, buf);
+	if (ret)
+		return ret;
+
+	if (arg->mib_id == WSM_MIB_ID_OPERATIONAL_POWER_MODE) {
+		/* OperationalMode: update PM status. */
+		const char *p = arg->buf;
+		cw1200_enable_powersave(priv, (p[0] & 0x0F) ? true : false);
+	}
+	return 0;
+}
+
+/* ******************************************************************** */
+
+int wsm_scan(struct cw1200_common *priv, const struct wsm_scan *arg)
+{
+	int i;
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	if (arg->num_channels > 48)
+		return -EINVAL;
+
+	if (arg->num_ssids > 2)
+		return -EINVAL;
+
+	if (arg->band > 1)
+		return -EINVAL;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, arg->band);
+	WSM_PUT8(buf, arg->type);
+	WSM_PUT8(buf, arg->flags);
+	WSM_PUT8(buf, arg->max_tx_rate);
+	WSM_PUT32(buf, arg->auto_scan_interval);
+	WSM_PUT8(buf, arg->num_probes);
+	WSM_PUT8(buf, arg->num_channels);
+	WSM_PUT8(buf, arg->num_ssids);
+	WSM_PUT8(buf, arg->probe_delay);
+
+	for (i = 0; i < arg->num_channels; ++i) {
+		WSM_PUT16(buf, arg->ch[i].number);
+		WSM_PUT16(buf, 0);
+		WSM_PUT32(buf, arg->ch[i].min_chan_time);
+		WSM_PUT32(buf, arg->ch[i].max_chan_time);
+		WSM_PUT32(buf, 0);
+	}
+
+	for (i = 0; i < arg->num_ssids; ++i) {
+		WSM_PUT32(buf, arg->ssids[i].length);
+		WSM_PUT(buf, &arg->ssids[i].ssid[0],
+			sizeof(arg->ssids[i].ssid));
+	}
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_START_SCAN_REQ_ID, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_stop_scan(struct cw1200_common *priv)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	wsm_cmd_lock(priv);
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_STOP_SCAN_REQ_ID, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+}
+
+
+static int wsm_tx_confirm(struct cw1200_common *priv,
+			  struct wsm_buf *buf,
+			  int link_id)
+{
+	struct wsm_tx_confirm tx_confirm;
+
+	tx_confirm.packet_id = WSM_GET32(buf);
+	tx_confirm.status = WSM_GET32(buf);
+	tx_confirm.tx_rate = WSM_GET8(buf);
+	tx_confirm.ack_failures = WSM_GET8(buf);
+	tx_confirm.flags = WSM_GET16(buf);
+	tx_confirm.media_delay = WSM_GET32(buf);
+	tx_confirm.tx_queue_delay = WSM_GET32(buf);
+
+	cw1200_tx_confirm_cb(priv, link_id, &tx_confirm);
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+static int wsm_multi_tx_confirm(struct cw1200_common *priv,
+				struct wsm_buf *buf, int link_id)
+{
+	int ret;
+	int count;
+	int i;
+
+	count = WSM_GET32(buf);
+	if (WARN_ON(count <= 0))
+		return -EINVAL;
+
+	if (count > 1) {
+		/* We already released one buffer, now for the rest */
+		ret = wsm_release_tx_buffer(priv, count - 1);
+		if (ret < 0)
+			return ret;
+		else if (ret > 0)
+			cw1200_bh_wakeup(priv);
+	}
+
+	cw1200_debug_txed_multi(priv, count);
+	for (i = 0; i < count; ++i) {
+		ret = wsm_tx_confirm(priv, buf, link_id);
+		if (ret)
+			return ret;
+	}
+	return ret;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+/* ******************************************************************** */
+
+static int wsm_join_confirm(struct cw1200_common *priv,
+			    struct wsm_join_cnf *arg,
+			    struct wsm_buf *buf)
+{
+	arg->status = WSM_GET32(buf);
+	if (WARN_ON(arg->status) != WSM_STATUS_SUCCESS)
+		return -EINVAL;
+
+	arg->min_power_level = WSM_GET32(buf);
+	arg->max_power_level = WSM_GET32(buf);
+
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+int wsm_join(struct cw1200_common *priv, struct wsm_join *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	struct wsm_join_cnf resp;
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, arg->mode);
+	WSM_PUT8(buf, arg->band);
+	WSM_PUT16(buf, arg->channel_number);
+	WSM_PUT(buf, &arg->bssid[0], sizeof(arg->bssid));
+	WSM_PUT16(buf, arg->atim_window);
+	WSM_PUT8(buf, arg->preamble_type);
+	WSM_PUT8(buf, arg->probe_for_join);
+	WSM_PUT8(buf, arg->dtim_period);
+	WSM_PUT8(buf, arg->flags);
+	WSM_PUT32(buf, arg->ssid_len);
+	WSM_PUT(buf, &arg->ssid[0], sizeof(arg->ssid));
+	WSM_PUT32(buf, arg->beacon_interval);
+	WSM_PUT32(buf, arg->basic_rate_set);
+
+	priv->tx_burst_idx = -1;
+	ret = wsm_cmd_send(priv, buf, &resp,
+			   WSM_JOIN_REQ_ID, WSM_CMD_TIMEOUT);
+	/* TODO:  Update state based on resp.min|max_power_level */
+
+	priv->join_complete_status = resp.status;
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_set_bss_params(struct cw1200_common *priv,
+		       const struct wsm_set_bss_params *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, (arg->reset_beacon_loss ?  0x1 : 0));
+	WSM_PUT8(buf, arg->beacon_lost_count);
+	WSM_PUT16(buf, arg->aid);
+	WSM_PUT32(buf, arg->operational_rate_set);
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_SET_BSS_PARAMS_REQ_ID, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_add_key(struct cw1200_common *priv, const struct wsm_add_key *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT(buf, arg, sizeof(*arg));
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_ADD_KEY_REQ_ID, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_remove_key(struct cw1200_common *priv, const struct wsm_remove_key *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, arg->index);
+	WSM_PUT8(buf, 0);
+	WSM_PUT16(buf, 0);
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_REMOVE_KEY_REQ_ID, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_set_tx_queue_params(struct cw1200_common *priv,
+		const struct wsm_set_tx_queue_params *arg, u8 id)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	u8 queue_id_to_wmm_aci[] = {3, 2, 0, 1};
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, queue_id_to_wmm_aci[id]);
+	WSM_PUT8(buf, 0);
+	WSM_PUT8(buf, arg->ackPolicy);
+	WSM_PUT8(buf, 0);
+	WSM_PUT32(buf, arg->maxTransmitLifetime);
+	WSM_PUT16(buf, arg->allowedMediumTime);
+	WSM_PUT16(buf, 0);
+
+	ret = wsm_cmd_send(priv, buf, NULL, 0x0012, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_set_edca_params(struct cw1200_common *priv,
+				const struct wsm_edca_params *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	/* Implemented according to specification. */
+
+	WSM_PUT16(buf, arg->params[3].cwmin);
+	WSM_PUT16(buf, arg->params[2].cwmin);
+	WSM_PUT16(buf, arg->params[1].cwmin);
+	WSM_PUT16(buf, arg->params[0].cwmin);
+
+	WSM_PUT16(buf, arg->params[3].cwmax);
+	WSM_PUT16(buf, arg->params[2].cwmax);
+	WSM_PUT16(buf, arg->params[1].cwmax);
+	WSM_PUT16(buf, arg->params[0].cwmax);
+
+	WSM_PUT8(buf, arg->params[3].aifns);
+	WSM_PUT8(buf, arg->params[2].aifns);
+	WSM_PUT8(buf, arg->params[1].aifns);
+	WSM_PUT8(buf, arg->params[0].aifns);
+
+	WSM_PUT16(buf, arg->params[3].txop_limit);
+	WSM_PUT16(buf, arg->params[2].txop_limit);
+	WSM_PUT16(buf, arg->params[1].txop_limit);
+	WSM_PUT16(buf, arg->params[0].txop_limit);
+
+	WSM_PUT32(buf, arg->params[3].max_rx_lifetime);
+	WSM_PUT32(buf, arg->params[2].max_rx_lifetime);
+	WSM_PUT32(buf, arg->params[1].max_rx_lifetime);
+	WSM_PUT32(buf, arg->params[0].max_rx_lifetime);
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_EDCA_PARAMS_REQ_ID, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_switch_channel(struct cw1200_common *priv,
+			const struct wsm_switch_channel *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, arg->mode);
+	WSM_PUT8(buf, arg->switch_count);
+	WSM_PUT16(buf, arg->channel_number);
+
+	priv->channel_switch_in_progress = 1;
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_SWITCH_CHANNEL_REQ_ID, WSM_CMD_TIMEOUT);
+	if (ret)
+		priv->channel_switch_in_progress = 0;
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_set_pm(struct cw1200_common *priv, const struct wsm_set_pm *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	priv->ps_mode_switch_in_progress = 1;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, arg->mode);
+	WSM_PUT8(buf, arg->fast_psm_idle_period);
+	WSM_PUT8(buf, arg->ap_psm_change_period);
+	WSM_PUT8(buf, arg->min_auto_pspoll_period);
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_SET_PM_REQ_ID, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_start(struct cw1200_common *priv, const struct wsm_start *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT8(buf, arg->mode);
+	WSM_PUT8(buf, arg->band);
+	WSM_PUT16(buf, arg->channel_number);
+	WSM_PUT32(buf, arg->ct_window);
+	WSM_PUT32(buf, arg->beacon_interval);
+	WSM_PUT8(buf, arg->dtim_period);
+	WSM_PUT8(buf, arg->preamble);
+	WSM_PUT8(buf, arg->probe_delay);
+	WSM_PUT8(buf, arg->ssid_len);
+	WSM_PUT(buf, arg->ssid, sizeof(arg->ssid));
+	WSM_PUT32(buf, arg->basic_rate_set);
+
+	priv->tx_burst_idx = -1;
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_START_REQ_ID, WSM_CMD_START_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_beacon_transmit(struct cw1200_common *priv,
+			const struct wsm_beacon_transmit *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT32(buf, arg->enable_beaconing ? 1 : 0);
+
+	ret = wsm_cmd_send(priv, buf, NULL,
+			   WSM_BEACON_TRANSMIT_REQ_ID, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_start_find(struct cw1200_common *priv)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+	ret = wsm_cmd_send(priv, buf, NULL, 0x0019, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+}
+
+/* ******************************************************************** */
+
+int wsm_stop_find(struct cw1200_common *priv)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+	ret = wsm_cmd_send(priv, buf, NULL, 0x001A, WSM_CMD_TIMEOUT);
+	wsm_cmd_unlock(priv);
+	return ret;
+}
+
+/* ******************************************************************** */
+
+int wsm_map_link(struct cw1200_common *priv, const struct wsm_map_link *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+	u16 cmd = 0x001C | WSM_TX_LINK_ID(arg->link_id);
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT(buf, &arg->mac_addr[0], sizeof(arg->mac_addr));
+	WSM_PUT16(buf, 0);
+
+	ret = wsm_cmd_send(priv, buf, NULL, cmd, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+
+int wsm_update_ie(struct cw1200_common *priv,
+		  const struct wsm_update_ie *arg)
+{
+	int ret;
+	struct wsm_buf *buf = &priv->wsm_cmd_buf;
+
+	wsm_cmd_lock(priv);
+
+	WSM_PUT16(buf, arg->what);
+	WSM_PUT16(buf, arg->count);
+	WSM_PUT(buf, arg->ies, arg->length);
+
+	ret = wsm_cmd_send(priv, buf, NULL, 0x001B, WSM_CMD_TIMEOUT);
+
+	wsm_cmd_unlock(priv);
+	return ret;
+
+nomem:
+	wsm_cmd_unlock(priv);
+	return -ENOMEM;
+}
+
+/* ******************************************************************** */
+int wsm_set_probe_responder(struct cw1200_common *priv, bool enable)
+{
+	priv->rx_filter.probeResponder = enable;
+	return wsm_set_rx_filter(priv, &priv->rx_filter);
+}
+
+/* ******************************************************************** */
+/* WSM indication events implementation					*/
+const char * const cw1200_fw_types[] = {
+	"ETF",
+	"WFM",
+	"WSM",
+	"HI test",
+	"Platform test"
+};
+
+static int wsm_startup_indication(struct cw1200_common *priv,
+					struct wsm_buf *buf)
+{
+	priv->wsm_caps.input_buffers     = WSM_GET16(buf);
+	priv->wsm_caps.input_buffer_size = WSM_GET16(buf);
+	priv->wsm_caps.hw_id	  = WSM_GET16(buf);
+	priv->wsm_caps.hw_subid	  = WSM_GET16(buf);
+	priv->wsm_caps.status	  = WSM_GET16(buf);
+	priv->wsm_caps.fw_cap	  = WSM_GET16(buf);
+	priv->wsm_caps.fw_type	  = WSM_GET16(buf);
+	priv->wsm_caps.fw_api	  = WSM_GET16(buf);
+	priv->wsm_caps.fw_build   = WSM_GET16(buf);
+	priv->wsm_caps.fw_ver     = WSM_GET16(buf);
+	WSM_GET(buf, priv->wsm_caps.fw_label, sizeof(priv->wsm_caps.fw_label));
+	priv->wsm_caps.fw_label[sizeof(priv->wsm_caps.fw_label) - 1] = 0; /* Do not trust FW too much... */
+
+	if (WARN_ON(priv->wsm_caps.status))
+		return -EINVAL;
+
+	if (WARN_ON(priv->wsm_caps.fw_type > 4))
+		return -EINVAL;
+
+	pr_info("CW1200 WSM init done.\n"
+		"   Input buffers: %d x %d bytes\n"
+		"   Hardware: %d.%d\n"
+		"   %s firmware [%s], ver: %d, build: %d,"
+		"   api: %d, cap: 0x%.4X\n",
+		priv->wsm_caps.input_buffers,
+		priv->wsm_caps.input_buffer_size,
+		priv->wsm_caps.hw_id, priv->wsm_caps.hw_subid,
+		cw1200_fw_types[priv->wsm_caps.fw_type],
+		priv->wsm_caps.fw_label, priv->wsm_caps.fw_ver,
+		priv->wsm_caps.fw_build,
+		priv->wsm_caps.fw_api, priv->wsm_caps.fw_cap);
+
+	/* Disable unsupported frequency bands */
+	if (!(priv->wsm_caps.fw_cap & 0x1))
+		priv->hw->wiphy->bands[IEEE80211_BAND_2GHZ] = NULL;
+	if (!(priv->wsm_caps.fw_cap & 0x2))
+		priv->hw->wiphy->bands[IEEE80211_BAND_5GHZ] = NULL;
+
+	priv->firmware_ready = 1;
+	wake_up(&priv->wsm_startup_done);
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+static int wsm_receive_indication(struct cw1200_common *priv,
+				  int link_id,
+				  struct wsm_buf *buf,
+				  struct sk_buff **skb_p)
+{
+	struct wsm_rx rx;
+	struct ieee80211_hdr *hdr;
+	size_t hdr_len;
+	__le16 fctl;
+
+	rx.status = WSM_GET32(buf);
+	rx.channel_number = WSM_GET16(buf);
+	rx.rx_rate = WSM_GET8(buf);
+	rx.rcpi_rssi = WSM_GET8(buf);
+	rx.flags = WSM_GET32(buf);
+
+	/* FW Workaround: Drop probe resp or
+	   beacon when RSSI is 0
+	*/
+	hdr = (struct ieee80211_hdr *)(*skb_p)->data;
+
+	if (!rx.rcpi_rssi &&
+	    (ieee80211_is_probe_resp(hdr->frame_control) ||
+	     ieee80211_is_beacon(hdr->frame_control)))
+		return 0;
+
+	/* If no RSSI subscription has been made,
+	 * convert RCPI to RSSI here
+	 */
+	if (!priv->cqm_use_rssi)
+		rx.rcpi_rssi = rx.rcpi_rssi / 2 - 110;
+
+	fctl = *(__le16 *)buf->data;
+	hdr_len = buf->data - buf->begin;
+	skb_pull(*skb_p, hdr_len);
+	if (!rx.status && ieee80211_is_deauth(fctl)) {
+		if (priv->join_status == CW1200_JOIN_STATUS_STA) {
+			/* Shedule unjoin work */
+			pr_debug("[WSM] Issue unjoin command (RX).\n");
+			wsm_lock_tx_async(priv);
+			if (queue_work(priv->workqueue,
+				       &priv->unjoin_work) <= 0)
+				wsm_unlock_tx(priv);
+		}
+	}
+	cw1200_rx_cb(priv, &rx, link_id, skb_p);
+	if (*skb_p)
+		skb_push(*skb_p, hdr_len);
+
+	return 0;
+
+underflow:
+	return -EINVAL;
+}
+
+static int wsm_event_indication(struct cw1200_common *priv, struct wsm_buf *buf)
+{
+	int first;
+	struct cw1200_wsm_event *event;
+
+	if (priv->mode == NL80211_IFTYPE_UNSPECIFIED) {
+		/* STA is stopped. */
+		return 0;
+	}
+
+	event = kzalloc(sizeof(struct cw1200_wsm_event), GFP_KERNEL);
+	if (!event)
+		return -ENOMEM;
+
+	event->evt.id = WSM_GET32(buf);
+	event->evt.data = WSM_GET32(buf);
+
+	pr_debug("[WSM] Event: %d(%d)\n",
+		 event->evt.id, event->evt.data);
+
+	spin_lock(&priv->event_queue_lock);
+	first = list_empty(&priv->event_queue);
+	list_add_tail(&event->link, &priv->event_queue);
+	spin_unlock(&priv->event_queue_lock);
+
+	if (first)
+		queue_work(priv->workqueue, &priv->event_handler);
+
+	return 0;
+
+underflow:
+	kfree(event);
+	return -EINVAL;
+}
+
+static int wsm_channel_switch_indication(struct cw1200_common *priv,
+					 struct wsm_buf *buf)
+{
+	WARN_ON(WSM_GET32(buf));
+
+	priv->channel_switch_in_progress = 0;
+	wake_up(&priv->channel_switch_done);
+
+	wsm_unlock_tx(priv);
+
+	return 0;
+
+underflow:
+	return -EINVAL;
+}
+
+static int wsm_set_pm_indication(struct cw1200_common *priv,
+				 struct wsm_buf *buf)
+{
+	/* TODO:  Check buf (struct wsm_set_pm_complete) for validity */
+	if (priv->ps_mode_switch_in_progress) {
+		priv->ps_mode_switch_in_progress = 0;
+		wake_up(&priv->ps_mode_switch_done);
+	}
+	return 0;
+}
+
+static int wsm_scan_started(struct cw1200_common *priv, void *arg,
+			    struct wsm_buf *buf)
+{
+	u32 status = WSM_GET32(buf);
+	if (status != WSM_STATUS_SUCCESS) {
+		cw1200_scan_failed_cb(priv);
+		return -EINVAL;
+	}
+	return 0;
+
+underflow:
+	WARN_ON(1);
+	return -EINVAL;
+}
+
+static int wsm_scan_complete_indication(struct cw1200_common *priv,
+					struct wsm_buf *buf)
+{
+	struct wsm_scan_complete arg;
+	arg.status = WSM_GET32(buf);
+	arg.psm = WSM_GET8(buf);
+	arg.num_channels = WSM_GET8(buf);
+	cw1200_scan_complete_cb(priv, &arg);
+
+	return 0;
+
+underflow:
+	return -EINVAL;
+}
+
+static int wsm_join_complete_indication(struct cw1200_common *priv,
+					struct wsm_buf *buf)
+{
+	struct wsm_join_complete arg;
+	arg.status = WSM_GET32(buf);
+	pr_debug("[WSM] Join complete indication, status: %d\n", arg.status);
+	cw1200_join_complete_cb(priv, &arg);
+
+	return 0;
+
+underflow:
+	return -EINVAL;
+}
+
+static int wsm_find_complete_indication(struct cw1200_common *priv,
+					struct wsm_buf *buf)
+{
+	pr_warn("Implement find_complete_indication\n");
+	return 0;
+}
+
+static int wsm_ba_timeout_indication(struct cw1200_common *priv,
+				     struct wsm_buf *buf)
+{
+	u32 dummy;
+	u8 tid;
+	u8 dummy2;
+	u8 addr[ETH_ALEN];
+
+	dummy = WSM_GET32(buf);
+	tid = WSM_GET8(buf);
+	dummy2 = WSM_GET8(buf);
+	WSM_GET(buf, addr, ETH_ALEN);
+
+	pr_info("BlockACK timeout, tid %d, addr %pM\n",
+		tid, addr);
+
+	return 0;
+
+underflow:
+	return -EINVAL;
+}
+
+static int wsm_suspend_resume_indication(struct cw1200_common *priv,
+					 int link_id, struct wsm_buf *buf)
+{
+	u32 flags;
+	struct wsm_suspend_resume arg;
+
+	flags = WSM_GET32(buf);
+	arg.link_id = link_id;
+	arg.stop = !(flags & 1);
+	arg.multicast = !!(flags & 8);
+	arg.queue = (flags >> 1) & 3;
+
+	cw1200_suspend_resume(priv, &arg);
+
+	return 0;
+
+underflow:
+	return -EINVAL;
+}
+
+
+/* ******************************************************************** */
+/* WSM TX								*/
+
+static int wsm_cmd_send(struct cw1200_common *priv,
+			struct wsm_buf *buf,
+			void *arg, u16 cmd, long tmo)
+{
+	size_t buf_len = buf->data - buf->begin;
+	int ret;
+
+	/* Don't bother if we're dead. */
+	if (priv->bh_error) {
+		ret = 0;
+		goto done;
+	}
+
+	/* Block until the cmd buffer is completed.  Tortuous. */
+	spin_lock(&priv->wsm_cmd.lock);
+	while (!priv->wsm_cmd.done) {
+		spin_unlock(&priv->wsm_cmd.lock);
+		spin_lock(&priv->wsm_cmd.lock);
+	}
+	priv->wsm_cmd.done = 0;
+	spin_unlock(&priv->wsm_cmd.lock);
+
+	if (cmd == WSM_WRITE_MIB_REQ_ID ||
+	    cmd == WSM_READ_MIB_REQ_ID)
+		pr_debug("[WSM] >>> 0x%.4X [MIB: 0x%.4X] (%zu)\n",
+			 cmd, __le16_to_cpu(((__le16 *)buf->begin)[2]),
+			 buf_len);
+	else
+		pr_debug("[WSM] >>> 0x%.4X (%zu)\n", cmd, buf_len);
+
+	/* Due to buggy SPI on CW1200, we need to
+	 * pad the message by a few bytes to ensure
+	 * that it's completely received.
+	 */
+	buf_len += 4;
+
+	/* Fill HI message header */
+	/* BH will add sequence number */
+	((__le16 *)buf->begin)[0] = __cpu_to_le16(buf_len);
+	((__le16 *)buf->begin)[1] = __cpu_to_le16(cmd);
+
+	spin_lock(&priv->wsm_cmd.lock);
+	BUG_ON(priv->wsm_cmd.ptr);
+	priv->wsm_cmd.ptr = buf->begin;
+	priv->wsm_cmd.len = buf_len;
+	priv->wsm_cmd.arg = arg;
+	priv->wsm_cmd.cmd = cmd;
+	spin_unlock(&priv->wsm_cmd.lock);
+
+	cw1200_bh_wakeup(priv);
+
+	/* Wait for command completion */
+	ret = wait_event_timeout(priv->wsm_cmd_wq,
+				 priv->wsm_cmd.done, tmo);
+
+	if (!ret && !priv->wsm_cmd.done) {
+		spin_lock(&priv->wsm_cmd.lock);
+		priv->wsm_cmd.done = 1;
+		priv->wsm_cmd.ptr = NULL;
+		spin_unlock(&priv->wsm_cmd.lock);
+		if (priv->bh_error) {
+			/* Return ok to help system cleanup */
+			ret = 0;
+		} else {
+			pr_err("CMD req (0x%04x) stuck in firmware, killing BH\n", priv->wsm_cmd.cmd);
+			print_hex_dump_bytes("REQDUMP: ", DUMP_PREFIX_NONE,
+					     buf->begin, buf_len);
+			pr_err("Outstanding outgoing frames:  %d\n", priv->hw_bufs_used);
+
+			/* Kill BH thread to report the error to the top layer. */
+			atomic_add(1, &priv->bh_term);
+			wake_up(&priv->bh_wq);
+			ret = -ETIMEDOUT;
+		}
+	} else {
+		spin_lock(&priv->wsm_cmd.lock);
+		BUG_ON(!priv->wsm_cmd.done);
+		ret = priv->wsm_cmd.ret;
+		spin_unlock(&priv->wsm_cmd.lock);
+	}
+done:
+	wsm_buf_reset(buf);
+	return ret;
+}
+
+/* ******************************************************************** */
+/* WSM TX port control							*/
+
+void wsm_lock_tx(struct cw1200_common *priv)
+{
+	wsm_cmd_lock(priv);
+	if (atomic_add_return(1, &priv->tx_lock) == 1) {
+		if (wsm_flush_tx(priv))
+			pr_debug("[WSM] TX is locked.\n");
+	}
+	wsm_cmd_unlock(priv);
+}
+
+void wsm_lock_tx_async(struct cw1200_common *priv)
+{
+	if (atomic_add_return(1, &priv->tx_lock) == 1)
+		pr_debug("[WSM] TX is locked (async).\n");
+}
+
+bool wsm_flush_tx(struct cw1200_common *priv)
+{
+	unsigned long timestamp = jiffies;
+	bool pending = false;
+	long timeout;
+	int i;
+
+	/* Flush must be called with TX lock held. */
+	BUG_ON(!atomic_read(&priv->tx_lock));
+
+	/* First check if we really need to do something.
+	 * It is safe to use unprotected access, as hw_bufs_used
+	 * can only decrements.
+	 */
+	if (!priv->hw_bufs_used)
+		return true;
+
+	if (priv->bh_error) {
+		/* In case of failure do not wait for magic. */
+		pr_err("[WSM] Fatal error occured, will not flush TX.\n");
+		return false;
+	} else {
+		/* Get a timestamp of "oldest" frame */
+		for (i = 0; i < 4; ++i)
+			pending |= cw1200_queue_get_xmit_timestamp(
+					&priv->tx_queue[i],
+					&timestamp, 0xffffffff);
+		/* If there's nothing pending, we're good */
+		if (!pending)
+			return true;
+
+		timeout = timestamp + WSM_CMD_LAST_CHANCE_TIMEOUT - jiffies;
+		if (timeout < 0 || wait_event_timeout(priv->bh_evt_wq,
+						      !priv->hw_bufs_used,
+						      timeout) <= 0) {
+			/* Hmmm... Not good. Frame had stuck in firmware. */
+			priv->bh_error = 1;
+			wiphy_err(priv->hw->wiphy, "[WSM] TX Frames (%d) stuck in firmware, killing BH\n", priv->hw_bufs_used);
+			wake_up(&priv->bh_wq);
+			return false;
+		}
+
+		/* Ok, everything is flushed. */
+		return true;
+	}
+}
+
+void wsm_unlock_tx(struct cw1200_common *priv)
+{
+	int tx_lock;
+	tx_lock = atomic_sub_return(1, &priv->tx_lock);
+	BUG_ON(tx_lock < 0);
+
+	if (tx_lock == 0) {
+		if (!priv->bh_error)
+			cw1200_bh_wakeup(priv);
+		pr_debug("[WSM] TX is unlocked.\n");
+	}
+}
+
+/* ******************************************************************** */
+/* WSM RX								*/
+
+int wsm_handle_exception(struct cw1200_common *priv, u8 *data, size_t len)
+{
+	struct wsm_buf buf;
+	u32 reason;
+	u32 reg[18];
+	char fname[48];
+	unsigned int i;
+
+	static const char * const reason_str[] = {
+		"undefined instruction",
+		"prefetch abort",
+		"data abort",
+		"unknown error",
+	};
+
+	buf.begin = buf.data = data;
+	buf.end = &buf.begin[len];
+
+	reason = WSM_GET32(&buf);
+	for (i = 0; i < ARRAY_SIZE(reg); ++i)
+		reg[i] = WSM_GET32(&buf);
+	WSM_GET(&buf, fname, sizeof(fname));
+
+	if (reason < 4)
+		wiphy_err(priv->hw->wiphy,
+			  "Firmware exception: %s.\n",
+			  reason_str[reason]);
+	else
+		wiphy_err(priv->hw->wiphy,
+			  "Firmware assert at %.*s, line %d\n",
+			  (int) sizeof(fname), fname, reg[1]);
+
+	for (i = 0; i < 12; i += 4)
+		wiphy_err(priv->hw->wiphy,
+			  "R%d: 0x%.8X, R%d: 0x%.8X, R%d: 0x%.8X, R%d: 0x%.8X,\n",
+			  i + 0, reg[i + 0], i + 1, reg[i + 1],
+			  i + 2, reg[i + 2], i + 3, reg[i + 3]);
+	wiphy_err(priv->hw->wiphy,
+		  "R12: 0x%.8X, SP: 0x%.8X, LR: 0x%.8X, PC: 0x%.8X,\n",
+		  reg[i + 0], reg[i + 1], reg[i + 2], reg[i + 3]);
+	i += 4;
+	wiphy_err(priv->hw->wiphy,
+		  "CPSR: 0x%.8X, SPSR: 0x%.8X\n",
+		  reg[i + 0], reg[i + 1]);
+
+	print_hex_dump_bytes("R1: ", DUMP_PREFIX_NONE,
+			     fname, sizeof(fname));
+	return 0;
+
+underflow:
+	wiphy_err(priv->hw->wiphy, "Firmware exception.\n");
+	print_hex_dump_bytes("Exception: ", DUMP_PREFIX_NONE,
+			     data, len);
+	return -EINVAL;
+}
+
+int wsm_handle_rx(struct cw1200_common *priv, u16 id,
+		  struct wsm_hdr *wsm, struct sk_buff **skb_p)
+{
+	int ret = 0;
+	struct wsm_buf wsm_buf;
+	int link_id = (id >> 6) & 0x0F;
+
+	/* Strip link id. */
+	id &= ~WSM_TX_LINK_ID(WSM_TX_LINK_ID_MAX);
+
+	wsm_buf.begin = (u8 *)&wsm[0];
+	wsm_buf.data = (u8 *)&wsm[1];
+	wsm_buf.end = &wsm_buf.begin[__le16_to_cpu(wsm->len)];
+
+	pr_debug("[WSM] <<< 0x%.4X (%td)\n", id,
+		 wsm_buf.end - wsm_buf.begin);
+
+	if (id == WSM_TX_CONFIRM_IND_ID) {
+		ret = wsm_tx_confirm(priv, &wsm_buf, link_id);
+	} else if (id == WSM_MULTI_TX_CONFIRM_ID) {
+		ret = wsm_multi_tx_confirm(priv, &wsm_buf, link_id);
+	} else if (id & 0x0400) {
+		void *wsm_arg;
+		u16 wsm_cmd;
+
+		/* Do not trust FW too much. Protection against repeated
+		 * response and race condition removal (see above).
+		 */
+		spin_lock(&priv->wsm_cmd.lock);
+		wsm_arg = priv->wsm_cmd.arg;
+		wsm_cmd = priv->wsm_cmd.cmd &
+				~WSM_TX_LINK_ID(WSM_TX_LINK_ID_MAX);
+		priv->wsm_cmd.cmd = 0xFFFF;
+		spin_unlock(&priv->wsm_cmd.lock);
+
+		if (WARN_ON((id & ~0x0400) != wsm_cmd)) {
+			/* Note that any non-zero is a fatal retcode. */
+			ret = -EINVAL;
+			goto out;
+		}
+
+		/* Note that wsm_arg can be NULL in case of timeout in
+		 * wsm_cmd_send().
+		 */
+
+		switch (id) {
+		case WSM_READ_MIB_RESP_ID:
+			if (wsm_arg)
+				ret = wsm_read_mib_confirm(priv, wsm_arg,
+								&wsm_buf);
+			break;
+		case WSM_WRITE_MIB_RESP_ID:
+			if (wsm_arg)
+				ret = wsm_write_mib_confirm(priv, wsm_arg,
+							    &wsm_buf);
+			break;
+		case WSM_START_SCAN_RESP_ID:
+			if (wsm_arg)
+				ret = wsm_scan_started(priv, wsm_arg, &wsm_buf);
+			break;
+		case WSM_CONFIGURATION_RESP_ID:
+			if (wsm_arg)
+				ret = wsm_configuration_confirm(priv, wsm_arg,
+								&wsm_buf);
+			break;
+		case WSM_JOIN_RESP_ID:
+			if (wsm_arg)
+				ret = wsm_join_confirm(priv, wsm_arg, &wsm_buf);
+			break;
+		case WSM_STOP_SCAN_RESP_ID:
+		case WSM_RESET_RESP_ID:
+		case WSM_ADD_KEY_RESP_ID:
+		case WSM_REMOVE_KEY_RESP_ID:
+		case WSM_SET_PM_RESP_ID:
+		case WSM_SET_BSS_PARAMS_RESP_ID:
+		case 0x0412: /* set_tx_queue_params */
+		case WSM_EDCA_PARAMS_RESP_ID:
+		case WSM_SWITCH_CHANNEL_RESP_ID:
+		case WSM_START_RESP_ID:
+		case WSM_BEACON_TRANSMIT_RESP_ID:
+		case 0x0419: /* start_find */
+		case 0x041A: /* stop_find */
+		case 0x041B: /* update_ie */
+		case 0x041C: /* map_link */
+			WARN_ON(wsm_arg != NULL);
+			ret = wsm_generic_confirm(priv, wsm_arg, &wsm_buf);
+			if (ret) {
+				wiphy_warn(priv->hw->wiphy,
+					   "wsm_generic_confirm failed for request 0x%04x.\n",
+					   id & ~0x0400);
+
+				/* often 0x407 and 0x410 occur, this means we're dead.. */
+				if (priv->join_status >= CW1200_JOIN_STATUS_JOINING) {
+					wsm_lock_tx(priv);
+					if (queue_work(priv->workqueue, &priv->unjoin_work) <= 0)
+						wsm_unlock_tx(priv);
+				}
+			}
+			break;
+		default:
+			wiphy_warn(priv->hw->wiphy,
+				   "Unrecognized confirmation 0x%04x\n",
+				   id & ~0x0400);
+		}
+
+		spin_lock(&priv->wsm_cmd.lock);
+		priv->wsm_cmd.ret = ret;
+		priv->wsm_cmd.done = 1;
+		spin_unlock(&priv->wsm_cmd.lock);
+
+		ret = 0; /* Error response from device should ne stop BH. */
+
+		wake_up(&priv->wsm_cmd_wq);
+	} else if (id & 0x0800) {
+		switch (id) {
+		case WSM_STARTUP_IND_ID:
+			ret = wsm_startup_indication(priv, &wsm_buf);
+			break;
+		case WSM_RECEIVE_IND_ID:
+			ret = wsm_receive_indication(priv, link_id,
+						     &wsm_buf, skb_p);
+			break;
+		case 0x0805:
+			ret = wsm_event_indication(priv, &wsm_buf);
+			break;
+		case WSM_SCAN_COMPLETE_IND_ID:
+			ret = wsm_scan_complete_indication(priv, &wsm_buf);
+			break;
+		case 0x0808:
+			ret = wsm_ba_timeout_indication(priv, &wsm_buf);
+			break;
+		case 0x0809:
+			ret = wsm_set_pm_indication(priv, &wsm_buf);
+			break;
+		case 0x080A:
+			ret = wsm_channel_switch_indication(priv, &wsm_buf);
+			break;
+		case 0x080B:
+			ret = wsm_find_complete_indication(priv, &wsm_buf);
+			break;
+		case 0x080C:
+			ret = wsm_suspend_resume_indication(priv,
+					link_id, &wsm_buf);
+			break;
+		case 0x080F:
+			ret = wsm_join_complete_indication(priv, &wsm_buf);
+			break;
+		default:
+			pr_warn("Unrecognised WSM ID %04x\n", id);
+		}
+	} else {
+		WARN_ON(1);
+		ret = -EINVAL;
+	}
+out:
+	return ret;
+}
+
+static bool wsm_handle_tx_data(struct cw1200_common *priv,
+			       struct wsm_tx *wsm,
+			       const struct ieee80211_tx_info *tx_info,
+			       const struct cw1200_txpriv *txpriv,
+			       struct cw1200_queue *queue)
+{
+	bool handled = false;
+	const struct ieee80211_hdr *frame =
+		(struct ieee80211_hdr *)&((u8 *)wsm)[txpriv->offset];
+	__le16 fctl = frame->frame_control;
+	enum {
+		do_probe,
+		do_drop,
+		do_wep,
+		do_tx,
+	} action = do_tx;
+
+	switch (priv->mode) {
+	case NL80211_IFTYPE_STATION:
+		if (priv->join_status == CW1200_JOIN_STATUS_MONITOR)
+			action = do_tx;
+		else if (priv->join_status < CW1200_JOIN_STATUS_PRE_STA)
+			action = do_drop;
+		break;
+	case NL80211_IFTYPE_AP:
+		if (!priv->join_status) {
+			action = do_drop;
+		} else if (!(BIT(txpriv->raw_link_id) &
+			     (BIT(0) | priv->link_id_map))) {
+			wiphy_warn(priv->hw->wiphy,
+				   "A frame with expired link id is dropped.\n");
+			action = do_drop;
+		}
+		if (cw1200_queue_get_generation(wsm->packet_id) >
+				CW1200_MAX_REQUEUE_ATTEMPTS) {
+			/* HACK!!! WSM324 firmware has tendency to requeue
+			 * multicast frames in a loop, causing performance
+			 * drop and high power consumption of the driver.
+			 * In this situation it is better just to drop
+			 * the problematic frame.
+			 */
+			wiphy_warn(priv->hw->wiphy,
+				   "Too many attempts to requeue a frame; dropped.\n");
+			action = do_drop;
+		}
+		break;
+	case NL80211_IFTYPE_ADHOC:
+		if (priv->join_status != CW1200_JOIN_STATUS_IBSS)
+			action = do_drop;
+		break;
+	case NL80211_IFTYPE_MESH_POINT:
+		action = do_tx; /* TODO:  Test me! */
+		break;
+	case NL80211_IFTYPE_MONITOR:
+	default:
+		action = do_drop;
+		break;
+	}
+
+	if (action == do_tx) {
+		if (ieee80211_is_nullfunc(fctl)) {
+			spin_lock(&priv->bss_loss_lock);
+			if (priv->bss_loss_state) {
+				priv->bss_loss_confirm_id = wsm->packet_id;
+				wsm->queue_id = WSM_QUEUE_VOICE;
+			}
+			spin_unlock(&priv->bss_loss_lock);
+		} else if (ieee80211_is_probe_req(fctl)) {
+			action = do_probe;
+		} else if (ieee80211_is_deauth(fctl) &&
+			   priv->mode != NL80211_IFTYPE_AP) {
+			pr_debug("[WSM] Issue unjoin command due to tx deauth.\n");
+			wsm_lock_tx_async(priv);
+			if (queue_work(priv->workqueue,
+				       &priv->unjoin_work) <= 0)
+				wsm_unlock_tx(priv);
+		} else if (ieee80211_has_protected(fctl) &&
+			   tx_info->control.hw_key &&
+			   tx_info->control.hw_key->keyidx != priv->wep_default_key_id &&
+			   (tx_info->control.hw_key->cipher == WLAN_CIPHER_SUITE_WEP40 ||
+			    tx_info->control.hw_key->cipher == WLAN_CIPHER_SUITE_WEP104)) {
+			action = do_wep;
+		}
+	}
+
+	switch (action) {
+	case do_probe:
+		/* An interesting FW "feature". Device filters probe responses.
+		 * The easiest way to get it back is to convert
+		 * probe request into WSM start_scan command.
+		 */
+		pr_debug("[WSM] Convert probe request to scan.\n");
+		wsm_lock_tx_async(priv);
+		priv->pending_frame_id = wsm->packet_id;
+		if (queue_delayed_work(priv->workqueue,
+				       &priv->scan.probe_work, 0) <= 0)
+			wsm_unlock_tx(priv);
+		handled = true;
+		break;
+	case do_drop:
+		pr_debug("[WSM] Drop frame (0x%.4X).\n", fctl);
+		BUG_ON(cw1200_queue_remove(queue, wsm->packet_id));
+		handled = true;
+		break;
+	case do_wep:
+		pr_debug("[WSM] Issue set_default_wep_key.\n");
+		wsm_lock_tx_async(priv);
+		priv->wep_default_key_id = tx_info->control.hw_key->keyidx;
+		priv->pending_frame_id = wsm->packet_id;
+		if (queue_work(priv->workqueue, &priv->wep_key_work) <= 0)
+			wsm_unlock_tx(priv);
+		handled = true;
+		break;
+	case do_tx:
+		pr_debug("[WSM] Transmit frame.\n");
+		break;
+	default:
+		/* Do nothing */
+		break;
+	}
+	return handled;
+}
+
+static int cw1200_get_prio_queue(struct cw1200_common *priv,
+				 u32 link_id_map, int *total)
+{
+	static const int urgent = BIT(CW1200_LINK_ID_AFTER_DTIM) |
+		BIT(CW1200_LINK_ID_UAPSD);
+	struct wsm_edca_queue_params *edca;
+	unsigned score, best = -1;
+	int winner = -1;
+	int queued;
+	int i;
+
+	/* search for a winner using edca params */
+	for (i = 0; i < 4; ++i) {
+		queued = cw1200_queue_get_num_queued(&priv->tx_queue[i],
+				link_id_map);
+		if (!queued)
+			continue;
+		*total += queued;
+		edca = &priv->edca.params[i];
+		score = ((edca->aifns + edca->cwmin) << 16) +
+			((edca->cwmax - edca->cwmin) *
+			 (get_random_int() & 0xFFFF));
+		if (score < best && (winner < 0 || i != 3)) {
+			best = score;
+			winner = i;
+		}
+	}
+
+	/* override winner if bursting */
+	if (winner >= 0 && priv->tx_burst_idx >= 0 &&
+	    winner != priv->tx_burst_idx &&
+	    !cw1200_queue_get_num_queued(
+		    &priv->tx_queue[winner],
+		    link_id_map & urgent) &&
+	    cw1200_queue_get_num_queued(
+		    &priv->tx_queue[priv->tx_burst_idx],
+		    link_id_map))
+		winner = priv->tx_burst_idx;
+
+	return winner;
+}
+
+static int wsm_get_tx_queue_and_mask(struct cw1200_common *priv,
+				     struct cw1200_queue **queue_p,
+				     u32 *tx_allowed_mask_p,
+				     bool *more)
+{
+	int idx;
+	u32 tx_allowed_mask;
+	int total = 0;
+
+	/* Search for a queue with multicast frames buffered */
+	if (priv->tx_multicast) {
+		tx_allowed_mask = BIT(CW1200_LINK_ID_AFTER_DTIM);
+		idx = cw1200_get_prio_queue(priv,
+				tx_allowed_mask, &total);
+		if (idx >= 0) {
+			*more = total > 1;
+			goto found;
+		}
+	}
+
+	/* Search for unicast traffic */
+	tx_allowed_mask = ~priv->sta_asleep_mask;
+	tx_allowed_mask |= BIT(CW1200_LINK_ID_UAPSD);
+	if (priv->sta_asleep_mask) {
+		tx_allowed_mask |= priv->pspoll_mask;
+		tx_allowed_mask &= ~BIT(CW1200_LINK_ID_AFTER_DTIM);
+	} else {
+		tx_allowed_mask |= BIT(CW1200_LINK_ID_AFTER_DTIM);
+	}
+	idx = cw1200_get_prio_queue(priv,
+			tx_allowed_mask, &total);
+	if (idx < 0)
+		return -ENOENT;
+
+found:
+	*queue_p = &priv->tx_queue[idx];
+	*tx_allowed_mask_p = tx_allowed_mask;
+	return 0;
+}
+
+int wsm_get_tx(struct cw1200_common *priv, u8 **data,
+	       size_t *tx_len, int *burst)
+{
+	struct wsm_tx *wsm = NULL;
+	struct ieee80211_tx_info *tx_info;
+	struct cw1200_queue *queue = NULL;
+	int queue_num;
+	u32 tx_allowed_mask = 0;
+	const struct cw1200_txpriv *txpriv = NULL;
+	int count = 0;
+
+	/* More is used only for broadcasts. */
+	bool more = false;
+
+	if (priv->wsm_cmd.ptr) { /* CMD request */
+		++count;
+		spin_lock(&priv->wsm_cmd.lock);
+		BUG_ON(!priv->wsm_cmd.ptr);
+		*data = priv->wsm_cmd.ptr;
+		*tx_len = priv->wsm_cmd.len;
+		*burst = 1;
+		spin_unlock(&priv->wsm_cmd.lock);
+	} else {
+		for (;;) {
+			int ret;
+
+			if (atomic_add_return(0, &priv->tx_lock))
+				break;
+
+			spin_lock_bh(&priv->ps_state_lock);
+
+			ret = wsm_get_tx_queue_and_mask(priv, &queue,
+							&tx_allowed_mask, &more);
+			queue_num = queue - priv->tx_queue;
+
+			if (priv->buffered_multicasts &&
+			    (ret || !more) &&
+			    (priv->tx_multicast || !priv->sta_asleep_mask)) {
+				priv->buffered_multicasts = false;
+				if (priv->tx_multicast) {
+					priv->tx_multicast = false;
+					queue_work(priv->workqueue,
+						   &priv->multicast_stop_work);
+				}
+			}
+
+			spin_unlock_bh(&priv->ps_state_lock);
+
+			if (ret)
+				break;
+
+			if (cw1200_queue_get(queue,
+					     tx_allowed_mask,
+					     &wsm, &tx_info, &txpriv))
+				continue;
+
+			if (wsm_handle_tx_data(priv, wsm,
+					       tx_info, txpriv, queue))
+				continue;  /* Handled by WSM */
+
+			wsm->hdr.id &= __cpu_to_le16(
+				~WSM_TX_LINK_ID(WSM_TX_LINK_ID_MAX));
+			wsm->hdr.id |= cpu_to_le16(
+				WSM_TX_LINK_ID(txpriv->raw_link_id));
+			priv->pspoll_mask &= ~BIT(txpriv->raw_link_id);
+
+			*data = (u8 *)wsm;
+			*tx_len = __le16_to_cpu(wsm->hdr.len);
+
+			/* allow bursting if txop is set */
+			if (priv->edca.params[queue_num].txop_limit)
+				*burst = min(*burst,
+					     (int)cw1200_queue_get_num_queued(queue, tx_allowed_mask) + 1);
+			else
+				*burst = 1;
+
+			/* store index of bursting queue */
+			if (*burst > 1)
+				priv->tx_burst_idx = queue_num;
+			else
+				priv->tx_burst_idx = -1;
+
+			if (more) {
+				struct ieee80211_hdr *hdr =
+					(struct ieee80211_hdr *)
+					&((u8 *)wsm)[txpriv->offset];
+				/* more buffered multicast/broadcast frames
+				 *  ==> set MoreData flag in IEEE 802.11 header
+				 *  to inform PS STAs
+				 */
+				hdr->frame_control |=
+					cpu_to_le16(IEEE80211_FCTL_MOREDATA);
+			}
+
+			pr_debug("[WSM] >>> 0x%.4X (%zu) %p %c\n",
+				 0x0004, *tx_len, *data,
+				 wsm->more ? 'M' : ' ');
+			++count;
+			break;
+		}
+	}
+
+	return count;
+}
+
+void wsm_txed(struct cw1200_common *priv, u8 *data)
+{
+	if (data == priv->wsm_cmd.ptr) {
+		spin_lock(&priv->wsm_cmd.lock);
+		priv->wsm_cmd.ptr = NULL;
+		spin_unlock(&priv->wsm_cmd.lock);
+	}
+}
+
+/* ******************************************************************** */
+/* WSM buffer								*/
+
+void wsm_buf_init(struct wsm_buf *buf)
+{
+	BUG_ON(buf->begin);
+	buf->begin = kmalloc(FWLOAD_BLOCK_SIZE, GFP_KERNEL | GFP_DMA);
+	buf->end = buf->begin ? &buf->begin[FWLOAD_BLOCK_SIZE] : buf->begin;
+	wsm_buf_reset(buf);
+}
+
+void wsm_buf_deinit(struct wsm_buf *buf)
+{
+	kfree(buf->begin);
+	buf->begin = buf->data = buf->end = NULL;
+}
+
+static void wsm_buf_reset(struct wsm_buf *buf)
+{
+	if (buf->begin) {
+		buf->data = &buf->begin[4];
+		*(u32 *)buf->begin = 0;
+	} else {
+		buf->data = buf->begin;
+	}
+}
+
+static int wsm_buf_reserve(struct wsm_buf *buf, size_t extra_size)
+{
+	size_t pos = buf->data - buf->begin;
+	size_t size = pos + extra_size;
+
+	size = round_up(size, FWLOAD_BLOCK_SIZE);
+
+	buf->begin = krealloc(buf->begin, size, GFP_KERNEL | GFP_DMA);
+	if (buf->begin) {
+		buf->data = &buf->begin[pos];
+		buf->end = &buf->begin[size];
+		return 0;
+	} else {
+		buf->end = buf->data = buf->begin;
+		return -ENOMEM;
+	}
+}
diff --git a/drivers/net/wireless/cw1200/wsm.h b/drivers/net/wireless/cw1200/wsm.h
new file mode 100644
index 0000000..7afc613
--- /dev/null
+++ b/drivers/net/wireless/cw1200/wsm.h
@@ -0,0 +1,1870 @@
+/*
+ * WSM host interface (HI) interface for ST-Ericsson CW1200 mac80211 drivers
+ *
+ * Copyright (c) 2010, ST-Ericsson
+ * Author: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no>
+ *
+ * Based on CW1200 UMAC WSM API, which is
+ * Copyright (C) ST-Ericsson SA 2010
+ * Author: Stewart Mathers <stewart.mathers@stericsson.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef CW1200_WSM_H_INCLUDED
+#define CW1200_WSM_H_INCLUDED
+
+#include <linux/spinlock.h>
+
+struct cw1200_common;
+
+/* Bands */
+/* Radio band 2.412 -2.484 GHz. */
+#define WSM_PHY_BAND_2_4G		(0)
+
+/* Radio band 4.9375-5.8250 GHz. */
+#define WSM_PHY_BAND_5G			(1)
+
+/* Transmit rates */
+/* 1   Mbps            ERP-DSSS */
+#define WSM_TRANSMIT_RATE_1		(0)
+
+/* 2   Mbps            ERP-DSSS */
+#define WSM_TRANSMIT_RATE_2		(1)
+
+/* 5.5 Mbps            ERP-CCK */
+#define WSM_TRANSMIT_RATE_5		(2)
+
+/* 11  Mbps            ERP-CCK */
+#define WSM_TRANSMIT_RATE_11		(3)
+
+/* 22  Mbps            ERP-PBCC (Not supported) */
+/* #define WSM_TRANSMIT_RATE_22		(4) */
+
+/* 33  Mbps            ERP-PBCC (Not supported) */
+/* #define WSM_TRANSMIT_RATE_33		(5) */
+
+/* 6   Mbps   (3 Mbps) ERP-OFDM, BPSK coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_6		(6)
+
+/* 9   Mbps (4.5 Mbps) ERP-OFDM, BPSK coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_9		(7)
+
+/* 12  Mbps  (6 Mbps)  ERP-OFDM, QPSK coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_12		(8)
+
+/* 18  Mbps  (9 Mbps)  ERP-OFDM, QPSK coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_18		(9)
+
+/* 24  Mbps (12 Mbps)  ERP-OFDM, 16QAM coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_24		(10)
+
+/* 36  Mbps (18 Mbps)  ERP-OFDM, 16QAM coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_36		(11)
+
+/* 48  Mbps (24 Mbps)  ERP-OFDM, 64QAM coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_48		(12)
+
+/* 54  Mbps (27 Mbps)  ERP-OFDM, 64QAM coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_54		(13)
+
+/* 6.5 Mbps            HT-OFDM, BPSK coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_HT_6		(14)
+
+/* 13  Mbps            HT-OFDM, QPSK coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_HT_13		(15)
+
+/* 19.5 Mbps           HT-OFDM, QPSK coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_HT_19		(16)
+
+/* 26  Mbps            HT-OFDM, 16QAM coding rate 1/2 */
+#define WSM_TRANSMIT_RATE_HT_26		(17)
+
+/* 39  Mbps            HT-OFDM, 16QAM coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_HT_39		(18)
+
+/* 52  Mbps            HT-OFDM, 64QAM coding rate 2/3 */
+#define WSM_TRANSMIT_RATE_HT_52		(19)
+
+/* 58.5 Mbps           HT-OFDM, 64QAM coding rate 3/4 */
+#define WSM_TRANSMIT_RATE_HT_58		(20)
+
+/* 65  Mbps            HT-OFDM, 64QAM coding rate 5/6 */
+#define WSM_TRANSMIT_RATE_HT_65		(21)
+
+/* Scan types */
+/* Foreground scan */
+#define WSM_SCAN_TYPE_FOREGROUND	(0)
+
+/* Background scan */
+#define WSM_SCAN_TYPE_BACKGROUND	(1)
+
+/* Auto scan */
+#define WSM_SCAN_TYPE_AUTO		(2)
+
+/* Scan flags */
+/* Forced background scan means if the station cannot */
+/* enter the power-save mode, it shall force to perform a */
+/* background scan. Only valid when ScanType is */
+/* background scan. */
+#define WSM_SCAN_FLAG_FORCE_BACKGROUND	(BIT(0))
+
+/* The WLAN device scans one channel at a time so */
+/* that disturbance to the data traffic is minimized. */
+#define WSM_SCAN_FLAG_SPLIT_METHOD	(BIT(1))
+
+/* Preamble Type. Long if not set. */
+#define WSM_SCAN_FLAG_SHORT_PREAMBLE	(BIT(2))
+
+/* 11n Tx Mode. Mixed if not set. */
+#define WSM_SCAN_FLAG_11N_GREENFIELD	(BIT(3))
+
+/* Scan constraints */
+/* Maximum number of channels to be scanned. */
+#define WSM_SCAN_MAX_NUM_OF_CHANNELS	(48)
+
+/* The maximum number of SSIDs that the device can scan for. */
+#define WSM_SCAN_MAX_NUM_OF_SSIDS	(2)
+
+/* Power management modes */
+/* 802.11 Active mode */
+#define WSM_PSM_ACTIVE			(0)
+
+/* 802.11 PS mode */
+#define WSM_PSM_PS			BIT(0)
+
+/* Fast Power Save bit */
+#define WSM_PSM_FAST_PS_FLAG		BIT(7)
+
+/* Dynamic aka Fast power save */
+#define WSM_PSM_FAST_PS			(BIT(0) | BIT(7))
+
+/* Undetermined */
+/* Note : Undetermined status is reported when the */
+/* NULL data frame used to advertise the PM mode to */
+/* the AP at Pre or Post Background Scan is not Acknowledged */
+#define WSM_PSM_UNKNOWN			BIT(1)
+
+/* Queue IDs */
+/* best effort/legacy */
+#define WSM_QUEUE_BEST_EFFORT		(0)
+
+/* background */
+#define WSM_QUEUE_BACKGROUND		(1)
+
+/* video */
+#define WSM_QUEUE_VIDEO			(2)
+
+/* voice */
+#define WSM_QUEUE_VOICE			(3)
+
+/* HT TX parameters */
+/* Non-HT */
+#define WSM_HT_TX_NON_HT		(0)
+
+/* Mixed format */
+#define WSM_HT_TX_MIXED			(1)
+
+/* Greenfield format */
+#define WSM_HT_TX_GREENFIELD		(2)
+
+/* STBC allowed */
+#define WSM_HT_TX_STBC			(BIT(7))
+
+/* EPTA prioirty flags for BT Coex */
+/* default epta priority */
+#define WSM_EPTA_PRIORITY_DEFAULT	4
+/* use for normal data */
+#define WSM_EPTA_PRIORITY_DATA		4
+/* use for connect/disconnect/roaming*/
+#define WSM_EPTA_PRIORITY_MGT		5
+/* use for action frames */
+#define WSM_EPTA_PRIORITY_ACTION	5
+/* use for AC_VI data */
+#define WSM_EPTA_PRIORITY_VIDEO		5
+/* use for AC_VO data */
+#define WSM_EPTA_PRIORITY_VOICE		6
+/* use for EAPOL exchange */
+#define WSM_EPTA_PRIORITY_EAPOL		7
+
+/* TX status */
+/* Frame was sent aggregated */
+/* Only valid for WSM_SUCCESS status. */
+#define WSM_TX_STATUS_AGGREGATION	(BIT(0))
+
+/* Host should requeue this frame later. */
+/* Valid only when status is WSM_REQUEUE. */
+#define WSM_TX_STATUS_REQUEUE		(BIT(1))
+
+/* Normal Ack */
+#define WSM_TX_STATUS_NORMAL_ACK	(0<<2)
+
+/* No Ack */
+#define WSM_TX_STATUS_NO_ACK		(1<<2)
+
+/* No explicit acknowledgement */
+#define WSM_TX_STATUS_NO_EXPLICIT_ACK	(2<<2)
+
+/* Block Ack */
+/* Only valid for WSM_SUCCESS status. */
+#define WSM_TX_STATUS_BLOCK_ACK		(3<<2)
+
+/* RX status */
+/* Unencrypted */
+#define WSM_RX_STATUS_UNENCRYPTED	(0<<0)
+
+/* WEP */
+#define WSM_RX_STATUS_WEP		(1<<0)
+
+/* TKIP */
+#define WSM_RX_STATUS_TKIP		(2<<0)
+
+/* AES */
+#define WSM_RX_STATUS_AES		(3<<0)
+
+/* WAPI */
+#define WSM_RX_STATUS_WAPI		(4<<0)
+
+/* Macro to fetch encryption subfield. */
+#define WSM_RX_STATUS_ENCRYPTION(status) ((status) & 0x07)
+
+/* Frame was part of an aggregation */
+#define WSM_RX_STATUS_AGGREGATE		(BIT(3))
+
+/* Frame was first in the aggregation */
+#define WSM_RX_STATUS_AGGREGATE_FIRST	(BIT(4))
+
+/* Frame was last in the aggregation */
+#define WSM_RX_STATUS_AGGREGATE_LAST	(BIT(5))
+
+/* Indicates a defragmented frame */
+#define WSM_RX_STATUS_DEFRAGMENTED	(BIT(6))
+
+/* Indicates a Beacon frame */
+#define WSM_RX_STATUS_BEACON		(BIT(7))
+
+/* Indicates STA bit beacon TIM field */
+#define WSM_RX_STATUS_TIM		(BIT(8))
+
+/* Indicates Beacon frame's virtual bitmap contains multicast bit */
+#define WSM_RX_STATUS_MULTICAST		(BIT(9))
+
+/* Indicates frame contains a matching SSID */
+#define WSM_RX_STATUS_MATCHING_SSID	(BIT(10))
+
+/* Indicates frame contains a matching BSSI */
+#define WSM_RX_STATUS_MATCHING_BSSI	(BIT(11))
+
+/* Indicates More bit set in Framectl field */
+#define WSM_RX_STATUS_MORE_DATA		(BIT(12))
+
+/* Indicates frame received during a measurement process */
+#define WSM_RX_STATUS_MEASUREMENT	(BIT(13))
+
+/* Indicates frame received as an HT packet */
+#define WSM_RX_STATUS_HT		(BIT(14))
+
+/* Indicates frame received with STBC */
+#define WSM_RX_STATUS_STBC		(BIT(15))
+
+/* Indicates Address 1 field matches dot11StationId */
+#define WSM_RX_STATUS_ADDRESS1		(BIT(16))
+
+/* Indicates Group address present in the Address 1 field */
+#define WSM_RX_STATUS_GROUP		(BIT(17))
+
+/* Indicates Broadcast address present in the Address 1 field */
+#define WSM_RX_STATUS_BROADCAST		(BIT(18))
+
+/* Indicates group key used with encrypted frames */
+#define WSM_RX_STATUS_GROUP_KEY		(BIT(19))
+
+/* Macro to fetch encryption key index. */
+#define WSM_RX_STATUS_KEY_IDX(status)	(((status >> 20)) & 0x0F)
+
+/* Indicates TSF inclusion after 802.11 frame body */
+#define WSM_RX_STATUS_TSF_INCLUDED	(BIT(24))
+
+/* Frame Control field starts at Frame offset + 2 */
+#define WSM_TX_2BYTES_SHIFT		(BIT(7))
+
+/* Join mode */
+/* IBSS */
+#define WSM_JOIN_MODE_IBSS		(0)
+
+/* BSS */
+#define WSM_JOIN_MODE_BSS		(1)
+
+/* PLCP preamble type */
+/* For long preamble */
+#define WSM_JOIN_PREAMBLE_LONG		(0)
+
+/* For short preamble (Long for 1Mbps) */
+#define WSM_JOIN_PREAMBLE_SHORT		(1)
+
+/* For short preamble (Long for 1 and 2Mbps) */
+#define WSM_JOIN_PREAMBLE_SHORT_2	(2)
+
+/* Join flags */
+/* Unsynchronized */
+#define WSM_JOIN_FLAGS_UNSYNCRONIZED	BIT(0)
+/* The BSS owner is a P2P GO */
+#define WSM_JOIN_FLAGS_P2P_GO		BIT(1)
+/* Force to join BSS with the BSSID and the
+ * SSID specified without waiting for beacons. The
+ * ProbeForJoin parameter is ignored.
+ */
+#define WSM_JOIN_FLAGS_FORCE		BIT(2)
+/* Give probe request/response higher
+ * priority over the BT traffic
+ */
+#define WSM_JOIN_FLAGS_PRIO		BIT(3)
+/* Issue immediate join confirmation and use
+ * join complete to notify about completion
+ */
+#define WSM_JOIN_FLAGS_FORCE_WITH_COMPLETE_IND BIT(5)
+
+/* Key types */
+#define WSM_KEY_TYPE_WEP_DEFAULT	(0)
+#define WSM_KEY_TYPE_WEP_PAIRWISE	(1)
+#define WSM_KEY_TYPE_TKIP_GROUP		(2)
+#define WSM_KEY_TYPE_TKIP_PAIRWISE	(3)
+#define WSM_KEY_TYPE_AES_GROUP		(4)
+#define WSM_KEY_TYPE_AES_PAIRWISE	(5)
+#define WSM_KEY_TYPE_WAPI_GROUP		(6)
+#define WSM_KEY_TYPE_WAPI_PAIRWISE	(7)
+
+/* Key indexes */
+#define WSM_KEY_MAX_INDEX		(10)
+
+/* ACK policy */
+#define WSM_ACK_POLICY_NORMAL		(0)
+#define WSM_ACK_POLICY_NO_ACK		(1)
+
+/* Start modes */
+#define WSM_START_MODE_AP		(0)	/* Mini AP */
+#define WSM_START_MODE_P2P_GO		(1)	/* P2P GO */
+#define WSM_START_MODE_P2P_DEV		(2)	/* P2P device */
+
+/* SetAssociationMode MIB flags */
+#define WSM_ASSOCIATION_MODE_USE_PREAMBLE_TYPE		(BIT(0))
+#define WSM_ASSOCIATION_MODE_USE_HT_MODE		(BIT(1))
+#define WSM_ASSOCIATION_MODE_USE_BASIC_RATE_SET		(BIT(2))
+#define WSM_ASSOCIATION_MODE_USE_MPDU_START_SPACING	(BIT(3))
+#define WSM_ASSOCIATION_MODE_SNOOP_ASSOC_FRAMES		(BIT(4))
+
+/* RcpiRssiThreshold MIB flags */
+#define WSM_RCPI_RSSI_THRESHOLD_ENABLE	(BIT(0))
+#define WSM_RCPI_RSSI_USE_RSSI		(BIT(1))
+#define WSM_RCPI_RSSI_DONT_USE_UPPER	(BIT(2))
+#define WSM_RCPI_RSSI_DONT_USE_LOWER	(BIT(3))
+
+/* Update-ie constants */
+#define WSM_UPDATE_IE_BEACON		(BIT(0))
+#define WSM_UPDATE_IE_PROBE_RESP	(BIT(1))
+#define WSM_UPDATE_IE_PROBE_REQ		(BIT(2))
+
+/* WSM events */
+/* Error */
+#define WSM_EVENT_ERROR			(0)
+
+/* BSS lost */
+#define WSM_EVENT_BSS_LOST		(1)
+
+/* BSS regained */
+#define WSM_EVENT_BSS_REGAINED		(2)
+
+/* Radar detected */
+#define WSM_EVENT_RADAR_DETECTED	(3)
+
+/* RCPI or RSSI threshold triggered */
+#define WSM_EVENT_RCPI_RSSI		(4)
+
+/* BT inactive */
+#define WSM_EVENT_BT_INACTIVE		(5)
+
+/* BT active */
+#define WSM_EVENT_BT_ACTIVE		(6)
+
+/* MIB IDs */
+/* 4.1  dot11StationId */
+#define WSM_MIB_ID_DOT11_STATION_ID		0x0000
+
+/* 4.2  dot11MaxtransmitMsduLifeTime */
+#define WSM_MIB_ID_DOT11_MAX_TRANSMIT_LIFTIME	0x0001
+
+/* 4.3  dot11MaxReceiveLifeTime */
+#define WSM_MIB_ID_DOT11_MAX_RECEIVE_LIFETIME	0x0002
+
+/* 4.4  dot11SlotTime */
+#define WSM_MIB_ID_DOT11_SLOT_TIME		0x0003
+
+/* 4.5  dot11GroupAddressesTable */
+#define WSM_MIB_ID_DOT11_GROUP_ADDRESSES_TABLE	0x0004
+#define WSM_MAX_GRP_ADDRTABLE_ENTRIES		8
+
+/* 4.6  dot11WepDefaultKeyId */
+#define WSM_MIB_ID_DOT11_WEP_DEFAULT_KEY_ID	0x0005
+
+/* 4.7  dot11CurrentTxPowerLevel */
+#define WSM_MIB_ID_DOT11_CURRENT_TX_POWER_LEVEL	0x0006
+
+/* 4.8  dot11RTSThreshold */
+#define WSM_MIB_ID_DOT11_RTS_THRESHOLD		0x0007
+
+/* 4.9  NonErpProtection */
+#define WSM_MIB_ID_NON_ERP_PROTECTION		0x1000
+
+/* 4.10 ArpIpAddressesTable */
+#define WSM_MIB_ID_ARP_IP_ADDRESSES_TABLE	0x1001
+#define WSM_MAX_ARP_IP_ADDRTABLE_ENTRIES	1
+
+/* 4.11 TemplateFrame */
+#define WSM_MIB_ID_TEMPLATE_FRAME		0x1002
+
+/* 4.12 RxFilter */
+#define WSM_MIB_ID_RX_FILTER			0x1003
+
+/* 4.13 BeaconFilterTable */
+#define WSM_MIB_ID_BEACON_FILTER_TABLE		0x1004
+
+/* 4.14 BeaconFilterEnable */
+#define WSM_MIB_ID_BEACON_FILTER_ENABLE		0x1005
+
+/* 4.15 OperationalPowerMode */
+#define WSM_MIB_ID_OPERATIONAL_POWER_MODE	0x1006
+
+/* 4.16 BeaconWakeUpPeriod */
+#define WSM_MIB_ID_BEACON_WAKEUP_PERIOD		0x1007
+
+/* 4.17 RcpiRssiThreshold */
+#define WSM_MIB_ID_RCPI_RSSI_THRESHOLD		0x1009
+
+/* 4.18 StatisticsTable */
+#define WSM_MIB_ID_STATISTICS_TABLE		0x100A
+
+/* 4.19 IbssPsConfig */
+#define WSM_MIB_ID_IBSS_PS_CONFIG		0x100B
+
+/* 4.20 CountersTable */
+#define WSM_MIB_ID_COUNTERS_TABLE		0x100C
+
+/* 4.21 BlockAckPolicy */
+#define WSM_MIB_ID_BLOCK_ACK_POLICY		0x100E
+
+/* 4.22 OverrideInternalTxRate */
+#define WSM_MIB_ID_OVERRIDE_INTERNAL_TX_RATE	0x100F
+
+/* 4.23 SetAssociationMode */
+#define WSM_MIB_ID_SET_ASSOCIATION_MODE		0x1010
+
+/* 4.24 UpdateEptaConfigData */
+#define WSM_MIB_ID_UPDATE_EPTA_CONFIG_DATA	0x1011
+
+/* 4.25 SelectCcaMethod */
+#define WSM_MIB_ID_SELECT_CCA_METHOD		0x1012
+
+/* 4.26 SetUpasdInformation */
+#define WSM_MIB_ID_SET_UAPSD_INFORMATION	0x1013
+
+/* 4.27 SetAutoCalibrationMode  WBF00004073 */
+#define WSM_MIB_ID_SET_AUTO_CALIBRATION_MODE	0x1015
+
+/* 4.28 SetTxRateRetryPolicy */
+#define WSM_MIB_ID_SET_TX_RATE_RETRY_POLICY	0x1016
+
+/* 4.29 SetHostMessageTypeFilter */
+#define WSM_MIB_ID_SET_HOST_MSG_TYPE_FILTER	0x1017
+
+/* 4.30 P2PFindInfo */
+#define WSM_MIB_ID_P2P_FIND_INFO		0x1018
+
+/* 4.31 P2PPsModeInfo */
+#define WSM_MIB_ID_P2P_PS_MODE_INFO		0x1019
+
+/* 4.32 SetEtherTypeDataFrameFilter */
+#define WSM_MIB_ID_SET_ETHERTYPE_DATAFRAME_FILTER 0x101A
+
+/* 4.33 SetUDPPortDataFrameFilter */
+#define WSM_MIB_ID_SET_UDPPORT_DATAFRAME_FILTER	0x101B
+
+/* 4.34 SetMagicDataFrameFilter */
+#define WSM_MIB_ID_SET_MAGIC_DATAFRAME_FILTER	0x101C
+
+/* 4.35 P2PDeviceInfo */
+#define WSM_MIB_ID_P2P_DEVICE_INFO		0x101D
+
+/* 4.36 SetWCDMABand */
+#define WSM_MIB_ID_SET_WCDMA_BAND		0x101E
+
+/* 4.37 GroupTxSequenceCounter */
+#define WSM_MIB_ID_GRP_SEQ_COUNTER		0x101F
+
+/* 4.38 ProtectedMgmtPolicy */
+#define WSM_MIB_ID_PROTECTED_MGMT_POLICY	0x1020
+
+/* 4.39 SetHtProtection */
+#define WSM_MIB_ID_SET_HT_PROTECTION		0x1021
+
+/* 4.40 GPIO Command */
+#define WSM_MIB_ID_GPIO_COMMAND			0x1022
+
+/* 4.41 TSF Counter Value */
+#define WSM_MIB_ID_TSF_COUNTER			0x1023
+
+/* Test Purposes Only */
+#define WSM_MIB_ID_BLOCK_ACK_INFO		0x100D
+
+/* 4.42 UseMultiTxConfMessage */
+#define WSM_MIB_USE_MULTI_TX_CONF		0x1024
+
+/* 4.43 Keep-alive period */
+#define WSM_MIB_ID_KEEP_ALIVE_PERIOD		0x1025
+
+/* 4.44 Disable BSSID filter */
+#define WSM_MIB_ID_DISABLE_BSSID_FILTER		0x1026
+
+/* Frame template types */
+#define WSM_FRAME_TYPE_PROBE_REQUEST	(0)
+#define WSM_FRAME_TYPE_BEACON		(1)
+#define WSM_FRAME_TYPE_NULL		(2)
+#define WSM_FRAME_TYPE_QOS_NULL		(3)
+#define WSM_FRAME_TYPE_PS_POLL		(4)
+#define WSM_FRAME_TYPE_PROBE_RESPONSE	(5)
+
+#define WSM_FRAME_GREENFIELD		(0x80)	/* See 4.11 */
+
+/* Status */
+/* The WSM firmware has completed a request */
+/* successfully. */
+#define WSM_STATUS_SUCCESS              (0)
+
+/* This is a generic failure code if other error codes do */
+/* not apply. */
+#define WSM_STATUS_FAILURE              (1)
+
+/* A request contains one or more invalid parameters. */
+#define WSM_INVALID_PARAMETER           (2)
+
+/* The request cannot perform because the device is in */
+/* an inappropriate mode. */
+#define WSM_ACCESS_DENIED               (3)
+
+/* The frame received includes a decryption error. */
+#define WSM_STATUS_DECRYPTFAILURE       (4)
+
+/* A MIC failure is detected in the received packets. */
+#define WSM_STATUS_MICFAILURE           (5)
+
+/* The transmit request failed due to retry limit being */
+/* exceeded. */
+#define WSM_STATUS_RETRY_EXCEEDED       (6)
+
+/* The transmit request failed due to MSDU life time */
+/* being exceeded. */
+#define WSM_STATUS_TX_LIFETIME_EXCEEDED (7)
+
+/* The link to the AP is lost. */
+#define WSM_STATUS_LINK_LOST            (8)
+
+/* No key was found for the encrypted frame */
+#define WSM_STATUS_NO_KEY_FOUND         (9)
+
+/* Jammer was detected when transmitting this frame */
+#define WSM_STATUS_JAMMER_DETECTED      (10)
+
+/* The message should be requeued later. */
+/* This is applicable only to Transmit */
+#define WSM_REQUEUE                     (11)
+
+/* Advanced filtering options */
+#define WSM_MAX_FILTER_ELEMENTS		(4)
+
+#define WSM_FILTER_ACTION_IGNORE	(0)
+#define WSM_FILTER_ACTION_FILTER_IN	(1)
+#define WSM_FILTER_ACTION_FILTER_OUT	(2)
+
+#define WSM_FILTER_PORT_TYPE_DST	(0)
+#define WSM_FILTER_PORT_TYPE_SRC	(1)
+
+/* Actual header of WSM messages */
+struct wsm_hdr {
+	__le16 len;
+	__le16 id;
+};
+
+#define WSM_TX_SEQ_MAX			(7)
+#define WSM_TX_SEQ(seq)			\
+		((seq & WSM_TX_SEQ_MAX) << 13)
+#define WSM_TX_LINK_ID_MAX		(0x0F)
+#define WSM_TX_LINK_ID(link_id)		\
+		((link_id & WSM_TX_LINK_ID_MAX) << 6)
+
+#define MAX_BEACON_SKIP_TIME_MS 1000
+
+#define WSM_CMD_LAST_CHANCE_TIMEOUT (HZ * 3 / 2)
+
+/* ******************************************************************** */
+/* WSM capability							*/
+
+#define WSM_STARTUP_IND_ID 0x0801
+
+struct wsm_startup_ind {
+	u16 input_buffers;
+	u16 input_buffer_size;
+	u16 status;
+	u16 hw_id;
+	u16 hw_subid;
+	u16 fw_cap;
+	u16 fw_type;
+	u16 fw_api;
+	u16 fw_build;
+	u16 fw_ver;
+	char fw_label[128];
+	u32 config[4];
+};
+
+/* ******************************************************************** */
+/* WSM commands								*/
+
+/* 3.1 */
+#define WSM_CONFIGURATION_REQ_ID 0x0009
+#define WSM_CONFIGURATION_RESP_ID 0x0409
+
+struct wsm_tx_power_range {
+	int min_power_level;
+	int max_power_level;
+	u32 stepping;
+};
+
+struct wsm_configuration {
+	/* [in] */ u32 dot11MaxTransmitMsduLifeTime;
+	/* [in] */ u32 dot11MaxReceiveLifeTime;
+	/* [in] */ u32 dot11RtsThreshold;
+	/* [in, out] */ u8 *dot11StationId;
+	/* [in] */ const void *dpdData;
+	/* [in] */ size_t dpdData_size;
+	/* [out] */ u8 dot11FrequencyBandsSupported;
+	/* [out] */ u32 supportedRateMask;
+	/* [out] */ struct wsm_tx_power_range txPowerRange[2];
+};
+
+int wsm_configuration(struct cw1200_common *priv,
+		      struct wsm_configuration *arg);
+
+/* 3.3 */
+#define WSM_RESET_REQ_ID 0x000A
+#define WSM_RESET_RESP_ID 0x040A
+struct wsm_reset {
+	/* [in] */ int link_id;
+	/* [in] */ bool reset_statistics;
+};
+
+int wsm_reset(struct cw1200_common *priv, const struct wsm_reset *arg);
+
+/* 3.5 */
+#define WSM_READ_MIB_REQ_ID 0x0005
+#define WSM_READ_MIB_RESP_ID 0x0405
+int wsm_read_mib(struct cw1200_common *priv, u16 mib_id, void *buf,
+		 size_t buf_size);
+
+/* 3.7 */
+#define WSM_WRITE_MIB_REQ_ID 0x0006
+#define WSM_WRITE_MIB_RESP_ID 0x0406
+int wsm_write_mib(struct cw1200_common *priv, u16 mib_id, void *buf,
+		  size_t buf_size);
+
+/* 3.9 */
+#define WSM_START_SCAN_REQ_ID 0x0007
+#define WSM_START_SCAN_RESP_ID 0x0407
+
+struct wsm_ssid {
+	u8 ssid[32];
+	u32 length;
+};
+
+struct wsm_scan_ch {
+	u16 number;
+	u32 min_chan_time;
+	u32 max_chan_time;
+	u32 tx_power_level;
+};
+
+struct wsm_scan {
+	/* WSM_PHY_BAND_... */
+	u8 band;
+
+	/* WSM_SCAN_TYPE_... */
+	u8 type;
+
+	/* WSM_SCAN_FLAG_... */
+	u8 flags;
+
+	/* WSM_TRANSMIT_RATE_... */
+	u8 max_tx_rate;
+
+	/* Interval period in TUs that the device shall the re- */
+	/* execute the requested scan. Max value supported by the device */
+	/* is 256s. */
+	u32 auto_scan_interval;
+
+	/* Number of probe requests (per SSID) sent to one (1) */
+	/* channel. Zero (0) means that none is send, which */
+	/* means that a passive scan is to be done. Value */
+	/* greater than zero (0) means that an active scan is to */
+	/* be done. */
+	u32 num_probes;
+
+	/* Number of channels to be scanned. */
+	/* Maximum value is WSM_SCAN_MAX_NUM_OF_CHANNELS. */
+	u8 num_channels;
+
+	/* Number of SSID provided in the scan command (this */
+	/* is zero (0) in broadcast scan) */
+	/* The maximum number of SSIDs is WSM_SCAN_MAX_NUM_OF_SSIDS. */
+	u8 num_ssids;
+
+	/* The delay time (in microseconds) period */
+	/* before sending a probe-request. */
+	u8 probe_delay;
+
+	/* SSIDs to be scanned [numOfSSIDs]; */
+	struct wsm_ssid *ssids;
+
+	/* Channels to be scanned [numOfChannels]; */
+	struct wsm_scan_ch *ch;
+};
+
+int wsm_scan(struct cw1200_common *priv, const struct wsm_scan *arg);
+
+/* 3.11 */
+#define WSM_STOP_SCAN_REQ_ID 0x0008
+#define WSM_STOP_SCAN_RESP_ID 0x0408
+int wsm_stop_scan(struct cw1200_common *priv);
+
+/* 3.13 */
+#define WSM_SCAN_COMPLETE_IND_ID 0x0806
+struct wsm_scan_complete {
+	/* WSM_STATUS_... */
+	u32 status;
+
+	/* WSM_PSM_... */
+	u8 psm;
+
+	/* Number of channels that the scan operation completed. */
+	u8 num_channels;
+};
+
+/* 3.14 */
+#define WSM_TX_CONFIRM_IND_ID 0x0404
+#define WSM_MULTI_TX_CONFIRM_ID 0x041E
+
+struct wsm_tx_confirm {
+	/* Packet identifier used in wsm_tx. */
+	u32 packet_id;
+
+	/* WSM_STATUS_... */
+	u32 status;
+
+	/* WSM_TRANSMIT_RATE_... */
+	u8 tx_rate;
+
+	/* The number of times the frame was transmitted */
+	/* without receiving an acknowledgement. */
+	u8 ack_failures;
+
+	/* WSM_TX_STATUS_... */
+	u16 flags;
+
+	/* The total time in microseconds that the frame spent in */
+	/* the WLAN device before transmission as completed. */
+	u32 media_delay;
+
+	/* The total time in microseconds that the frame spent in */
+	/* the WLAN device before transmission was started. */
+	u32 tx_queue_delay;
+};
+
+/* 3.15 */
+typedef void (*wsm_tx_confirm_cb) (struct cw1200_common *priv,
+				   struct wsm_tx_confirm *arg);
+
+/* Note that ideology of wsm_tx struct is different against the rest of
+ * WSM API. wsm_hdr is /not/ a caller-adapted struct to be used as an input
+ * argument for WSM call, but a prepared bytestream to be sent to firmware.
+ * It is filled partly in cw1200_tx, partly in low-level WSM code.
+ * Please pay attention once again: ideology is different.
+ *
+ * Legend:
+ * - [in]: cw1200_tx must fill this field.
+ * - [wsm]: the field is filled by low-level WSM.
+ */
+struct wsm_tx {
+	/* common WSM header */
+	struct wsm_hdr hdr;
+
+	/* Packet identifier that meant to be used in completion. */
+	u32 packet_id;  /* Note this is actually a cookie */
+
+	/* WSM_TRANSMIT_RATE_... */
+	u8 max_tx_rate;
+
+	/* WSM_QUEUE_... */
+	u8 queue_id;
+
+	/* True: another packet is pending on the host for transmission. */
+	u8 more;
+
+	/* Bit 0 = 0 - Start expiry time from first Tx attempt (default) */
+	/* Bit 0 = 1 - Start expiry time from receipt of Tx Request */
+	/* Bits 3:1  - PTA Priority */
+	/* Bits 6:4  - Tx Rate Retry Policy */
+	/* Bit 7 - Reserved */
+	u8 flags;
+
+	/* Should be 0. */
+	u32 reserved;
+
+	/* The elapsed time in TUs, after the initial transmission */
+	/* of an MSDU, after which further attempts to transmit */
+	/* the MSDU shall be terminated. Overrides the global */
+	/* dot11MaxTransmitMsduLifeTime setting [optional] */
+	/* Device will set the default value if this is 0. */
+	u32 expire_time;
+
+	/* WSM_HT_TX_... */
+	__le32 ht_tx_parameters;
+} __packed;
+
+/* = sizeof(generic hi hdr) + sizeof(wsm hdr) + sizeof(alignment) */
+#define WSM_TX_EXTRA_HEADROOM (28)
+
+/* 3.16 */
+#define WSM_RECEIVE_IND_ID 0x0804
+
+struct wsm_rx {
+	/* WSM_STATUS_... */
+	u32 status;
+
+	/* Specifies the channel of the received packet. */
+	u16 channel_number;
+
+	/* WSM_TRANSMIT_RATE_... */
+	u8 rx_rate;
+
+	/* This value is expressed in signed Q8.0 format for */
+	/* RSSI and unsigned Q7.1 format for RCPI. */
+	u8 rcpi_rssi;
+
+	/* WSM_RX_STATUS_... */
+	u32 flags;
+};
+
+/* = sizeof(generic hi hdr) + sizeof(wsm hdr) */
+#define WSM_RX_EXTRA_HEADROOM (16)
+
+typedef void (*wsm_rx_cb) (struct cw1200_common *priv, struct wsm_rx *arg,
+			   struct sk_buff **skb_p);
+
+/* 3.17 */
+struct wsm_event {
+	/* WSM_STATUS_... */
+	/* [out] */ u32 id;
+
+	/* Indication parameters. */
+	/* For error indication, this shall be a 32-bit WSM status. */
+	/* For RCPI or RSSI indication, this should be an 8-bit */
+	/* RCPI or RSSI value. */
+	/* [out] */ u32 data;
+};
+
+struct cw1200_wsm_event {
+	struct list_head link;
+	struct wsm_event evt;
+};
+
+/* 3.18 - 3.22 */
+/* Measurement. Skipped for now. Irrelevent. */
+
+typedef void (*wsm_event_cb) (struct cw1200_common *priv,
+			      struct wsm_event *arg);
+
+/* 3.23 */
+#define WSM_JOIN_REQ_ID 0x000B
+#define WSM_JOIN_RESP_ID 0x040B
+
+struct wsm_join {
+	/* WSM_JOIN_MODE_... */
+	u8 mode;
+
+	/* WSM_PHY_BAND_... */
+	u8 band;
+
+	/* Specifies the channel number to join. The channel */
+	/* number will be mapped to an actual frequency */
+	/* according to the band */
+	u16 channel_number;
+
+	/* Specifies the BSSID of the BSS or IBSS to be joined */
+	/* or the IBSS to be started. */
+	u8 bssid[6];
+
+	/* ATIM window of IBSS */
+	/* When ATIM window is zero the initiated IBSS does */
+	/* not support power saving. */
+	u16 atim_window;
+
+	/* WSM_JOIN_PREAMBLE_... */
+	u8 preamble_type;
+
+	/* Specifies if a probe request should be send with the */
+	/* specified SSID when joining to the network. */
+	u8 probe_for_join;
+
+	/* DTIM Period (In multiples of beacon interval) */
+	u8 dtim_period;
+
+	/* WSM_JOIN_FLAGS_... */
+	u8 flags;
+
+	/* Length of the SSID */
+	u32 ssid_len;
+
+	/* Specifies the SSID of the IBSS to join or start */
+	u8 ssid[32];
+
+	/* Specifies the time between TBTTs in TUs */
+	u32 beacon_interval;
+
+	/* A bit mask that defines the BSS basic rate set. */
+	u32 basic_rate_set;
+};
+
+struct wsm_join_cnf {
+	u32 status;
+
+	/* Minimum transmission power level in units of 0.1dBm */
+	u32 min_power_level;
+
+	/* Maximum transmission power level in units of 0.1dBm */
+	u32 max_power_level;
+};
+
+int wsm_join(struct cw1200_common *priv, struct wsm_join *arg);
+
+/* 3.24 */
+struct wsm_join_complete {
+	/* WSM_STATUS_... */
+	u32 status;
+};
+
+/* 3.25 */
+#define WSM_SET_PM_REQ_ID 0x0010
+#define WSM_SET_PM_RESP_ID 0x0410
+struct wsm_set_pm {
+	/* WSM_PSM_... */
+	u8 mode;
+
+	/* in unit of 500us; 0 to use default */
+	u8 fast_psm_idle_period;
+
+	/* in unit of 500us; 0 to use default */
+	u8 ap_psm_change_period;
+
+	/* in unit of 500us; 0 to disable auto-pspoll */
+	u8 min_auto_pspoll_period;
+};
+
+int wsm_set_pm(struct cw1200_common *priv, const struct wsm_set_pm *arg);
+
+/* 3.27 */
+struct wsm_set_pm_complete {
+	u8 psm;			/* WSM_PSM_... */
+};
+
+/* 3.28 */
+#define WSM_SET_BSS_PARAMS_REQ_ID 0x0011
+#define WSM_SET_BSS_PARAMS_RESP_ID 0x0411
+struct wsm_set_bss_params {
+	/* This resets the beacon loss counters only */
+	u8 reset_beacon_loss;
+
+	/* The number of lost consecutive beacons after which */
+	/* the WLAN device should indicate the BSS-Lost event */
+	/* to the WLAN host driver. */
+	u8 beacon_lost_count;
+
+	/* The AID received during the association process. */
+	u16 aid;
+
+	/* The operational rate set mask */
+	u32 operational_rate_set;
+};
+
+int wsm_set_bss_params(struct cw1200_common *priv,
+		       const struct wsm_set_bss_params *arg);
+
+/* 3.30 */
+#define WSM_ADD_KEY_REQ_ID         0x000C
+#define WSM_ADD_KEY_RESP_ID        0x040C
+struct wsm_add_key {
+	u8 type;		/* WSM_KEY_TYPE_... */
+	u8 index;		/* Key entry index: 0 -- WSM_KEY_MAX_INDEX */
+	u16 reserved;
+	union {
+		struct {
+			u8 peer[6];	/* MAC address of the peer station */
+			u8 reserved;
+			u8 keylen;		/* Key length in bytes */
+			u8 keydata[16];		/* Key data */
+		} __packed wep_pairwise;
+		struct {
+			u8 keyid;	/* Unique per key identifier (0..3) */
+			u8 keylen;		/* Key length in bytes */
+			u16 reserved;
+			u8 keydata[16];		/* Key data */
+		} __packed wep_group;
+		struct {
+			u8 peer[6];	/* MAC address of the peer station */
+			u16 reserved;
+			u8 keydata[16];	/* TKIP key data */
+			u8 rx_mic_key[8];		/* Rx MIC key */
+			u8 tx_mic_key[8];		/* Tx MIC key */
+		} __packed tkip_pairwise;
+		struct {
+			u8 keydata[16];	/* TKIP key data */
+			u8 rx_mic_key[8];		/* Rx MIC key */
+			u8 keyid;		/* Key ID */
+			u8 reserved[3];
+			u8 rx_seqnum[8];	/* Receive Sequence Counter */
+		} __packed tkip_group;
+		struct {
+			u8 peer[6];	/* MAC address of the peer station */
+			u16 reserved;
+			u8 keydata[16];	/* AES key data */
+		} __packed aes_pairwise;
+		struct {
+			u8 keydata[16];	/* AES key data */
+			u8 keyid;		/* Key ID */
+			u8 reserved[3];
+			u8 rx_seqnum[8];	/* Receive Sequence Counter */
+		} __packed aes_group;
+		struct {
+			u8 peer[6];	/* MAC address of the peer station */
+			u8 keyid;		/* Key ID */
+			u8 reserved;
+			u8 keydata[16];	/* WAPI key data */
+			u8 mic_key[16];	/* MIC key data */
+		} __packed wapi_pairwise;
+		struct {
+			u8 keydata[16];	/* WAPI key data */
+			u8 mic_key[16];	/* MIC key data */
+			u8 keyid;		/* Key ID */
+			u8 reserved[3];
+		} __packed wapi_group;
+	} __packed;
+} __packed;
+
+int wsm_add_key(struct cw1200_common *priv, const struct wsm_add_key *arg);
+
+/* 3.32 */
+#define WSM_REMOVE_KEY_REQ_ID         0x000D
+#define WSM_REMOVE_KEY_RESP_ID        0x040D
+struct wsm_remove_key {
+	u8 index; /* Key entry index : 0-10 */
+};
+
+int wsm_remove_key(struct cw1200_common *priv,
+		   const struct wsm_remove_key *arg);
+
+/* 3.34 */
+struct wsm_set_tx_queue_params {
+	/* WSM_ACK_POLICY_... */
+	u8 ackPolicy;
+
+	/* Medium Time of TSPEC (in 32us units) allowed per */
+	/* One Second Averaging Period for this queue. */
+	u16 allowedMediumTime;
+
+	/* dot11MaxTransmitMsduLifetime to be used for the */
+	/* specified queue. */
+	u32 maxTransmitLifetime;
+};
+
+struct wsm_tx_queue_params {
+	/* NOTE: index is a linux queue id. */
+	struct wsm_set_tx_queue_params params[4];
+};
+
+
+#define WSM_TX_QUEUE_SET(queue_params, queue, ack_policy, allowed_time,\
+		max_life_time)	\
+do {							\
+	struct wsm_set_tx_queue_params *p = &(queue_params)->params[queue]; \
+	p->ackPolicy = (ack_policy);				\
+	p->allowedMediumTime = (allowed_time);				\
+	p->maxTransmitLifetime = (max_life_time);			\
+} while (0)
+
+int wsm_set_tx_queue_params(struct cw1200_common *priv,
+			    const struct wsm_set_tx_queue_params *arg, u8 id);
+
+/* 3.36 */
+#define WSM_EDCA_PARAMS_REQ_ID 0x0013
+#define WSM_EDCA_PARAMS_RESP_ID 0x0413
+struct wsm_edca_queue_params {
+	/* CWmin (in slots) for the access class. */
+	u16 cwmin;
+
+	/* CWmax (in slots) for the access class. */
+	u16 cwmax;
+
+	/* AIFS (in slots) for the access class. */
+	u16 aifns;
+
+	/* TX OP Limit (in microseconds) for the access class. */
+	u16 txop_limit;
+
+	/* dot11MaxReceiveLifetime to be used for the specified */
+	/* the access class. Overrides the global */
+	/* dot11MaxReceiveLifetime value */
+	u32 max_rx_lifetime;
+};
+
+struct wsm_edca_params {
+	/* NOTE: index is a linux queue id. */
+	struct wsm_edca_queue_params params[4];
+	bool uapsd_enable[4];
+};
+
+#define TXOP_UNIT 32
+#define WSM_EDCA_SET(__edca, __queue, __aifs, __cw_min, __cw_max, __txop, __lifetime,\
+		     __uapsd) \
+	do {							\
+		struct wsm_edca_queue_params *p = &(__edca)->params[__queue]; \
+		p->cwmin = __cw_min;					\
+		p->cwmax = __cw_max;					\
+		p->aifns = __aifs;					\
+		p->txop_limit = ((__txop) * TXOP_UNIT);			\
+		p->max_rx_lifetime = __lifetime;			\
+		(__edca)->uapsd_enable[__queue] = (__uapsd);		\
+	} while (0)
+
+int wsm_set_edca_params(struct cw1200_common *priv,
+			const struct wsm_edca_params *arg);
+
+int wsm_set_uapsd_param(struct cw1200_common *priv,
+			const struct wsm_edca_params *arg);
+
+/* 3.38 */
+/* Set-System info. Skipped for now. Irrelevent. */
+
+/* 3.40 */
+#define WSM_SWITCH_CHANNEL_REQ_ID 0x0016
+#define WSM_SWITCH_CHANNEL_RESP_ID 0x0416
+
+struct wsm_switch_channel {
+	/* 1 - means the STA shall not transmit any further */
+	/* frames until the channel switch has completed */
+	u8 mode;
+
+	/* Number of TBTTs until channel switch occurs. */
+	/* 0 - indicates switch shall occur at any time */
+	/* 1 - occurs immediately before the next TBTT */
+	u8 switch_count;
+
+	/* The new channel number to switch to. */
+	/* Note this is defined as per section 2.7. */
+	u16 channel_number;
+};
+
+int wsm_switch_channel(struct cw1200_common *priv,
+		       const struct wsm_switch_channel *arg);
+
+typedef void (*wsm_channel_switch_cb) (struct cw1200_common *priv);
+
+#define WSM_START_REQ_ID 0x0017
+#define WSM_START_RESP_ID 0x0417
+
+struct wsm_start {
+	/* WSM_START_MODE_... */
+	/* [in] */ u8 mode;
+
+	/* WSM_PHY_BAND_... */
+	/* [in] */ u8 band;
+
+	/* Channel number */
+	/* [in] */ u16 channel_number;
+
+	/* Client Traffic window in units of TU */
+	/* Valid only when mode == ..._P2P */
+	/* [in] */ u32 ct_window;
+
+	/* Interval between two consecutive */
+	/* beacon transmissions in TU. */
+	/* [in] */ u32 beacon_interval;
+
+	/* DTIM period in terms of beacon intervals */
+	/* [in] */ u8 dtim_period;
+
+	/* WSM_JOIN_PREAMBLE_... */
+	/* [in] */ u8 preamble;
+
+	/* The delay time (in microseconds) period */
+	/* before sending a probe-request. */
+	/* [in] */ u8 probe_delay;
+
+	/* Length of the SSID */
+	/* [in] */ u8 ssid_len;
+
+	/* SSID of the BSS or P2P_GO to be started now. */
+	/* [in] */ u8 ssid[32];
+
+	/* The basic supported rates for the MiniAP. */
+	/* [in] */ u32 basic_rate_set;
+};
+
+int wsm_start(struct cw1200_common *priv, const struct wsm_start *arg);
+
+#define WSM_BEACON_TRANSMIT_REQ_ID 0x0018
+#define WSM_BEACON_TRANSMIT_RESP_ID 0x0418
+
+struct wsm_beacon_transmit {
+	/* 1: enable; 0: disable */
+	/* [in] */ u8 enable_beaconing;
+};
+
+int wsm_beacon_transmit(struct cw1200_common *priv,
+			const struct wsm_beacon_transmit *arg);
+
+int wsm_start_find(struct cw1200_common *priv);
+
+int wsm_stop_find(struct cw1200_common *priv);
+
+typedef void (*wsm_find_complete_cb) (struct cw1200_common *priv, u32 status);
+
+struct wsm_suspend_resume {
+	/* See 3.52 */
+	/* Link ID */
+	/* [out] */ int link_id;
+	/* Stop sending further Tx requests down to device for this link */
+	/* [out] */ bool stop;
+	/* Transmit multicast Frames */
+	/* [out] */ bool multicast;
+	/* The AC on which Tx to be suspended /resumed. */
+	/* This is applicable only for U-APSD */
+	/* WSM_QUEUE_... */
+	/* [out] */ int queue;
+};
+
+typedef void (*wsm_suspend_resume_cb) (struct cw1200_common *priv,
+				       struct wsm_suspend_resume *arg);
+
+/* 3.54 Update-IE request. */
+struct wsm_update_ie {
+	/* WSM_UPDATE_IE_... */
+	/* [in] */ u16 what;
+	/* [in] */ u16 count;
+	/* [in] */ u8 *ies;
+	/* [in] */ size_t length;
+};
+
+int wsm_update_ie(struct cw1200_common *priv,
+		  const struct wsm_update_ie *arg);
+
+/* 3.56 */
+struct wsm_map_link {
+	/* MAC address of the remote device */
+	/* [in] */ u8 mac_addr[6];
+	/* [in] */ u8 link_id;
+};
+
+int wsm_map_link(struct cw1200_common *priv, const struct wsm_map_link *arg);
+
+/* ******************************************************************** */
+/* MIB shortcats							*/
+
+static inline int wsm_set_output_power(struct cw1200_common *priv,
+				       int power_level)
+{
+	__le32 val = __cpu_to_le32(power_level);
+	return wsm_write_mib(priv, WSM_MIB_ID_DOT11_CURRENT_TX_POWER_LEVEL,
+			     &val, sizeof(val));
+}
+
+static inline int wsm_set_beacon_wakeup_period(struct cw1200_common *priv,
+					       unsigned dtim_interval,
+					       unsigned listen_interval)
+{
+	struct {
+		u8 numBeaconPeriods;
+		u8 reserved;
+		__le16 listenInterval;
+	} val = {
+		dtim_interval, 0, __cpu_to_le16(listen_interval)
+	};
+
+	if (dtim_interval > 0xFF || listen_interval > 0xFFFF)
+		return -EINVAL;
+	else
+		return wsm_write_mib(priv, WSM_MIB_ID_BEACON_WAKEUP_PERIOD,
+				     &val, sizeof(val));
+}
+
+struct wsm_rcpi_rssi_threshold {
+	u8 rssiRcpiMode;	/* WSM_RCPI_RSSI_... */
+	u8 lowerThreshold;
+	u8 upperThreshold;
+	u8 rollingAverageCount;
+};
+
+static inline int wsm_set_rcpi_rssi_threshold(struct cw1200_common *priv,
+					struct wsm_rcpi_rssi_threshold *arg)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_RCPI_RSSI_THRESHOLD, arg,
+			     sizeof(*arg));
+}
+
+struct wsm_mib_counters_table {
+	__le32 plcp_errors;
+	__le32 fcs_errors;
+	__le32 tx_packets;
+	__le32 rx_packets;
+	__le32 rx_packet_errors;
+	__le32 rx_decryption_failures;
+	__le32 rx_mic_failures;
+	__le32 rx_no_key_failures;
+	__le32 tx_multicast_frames;
+	__le32 tx_frames_success;
+	__le32 tx_frame_failures;
+	__le32 tx_frames_retried;
+	__le32 tx_frames_multi_retried;
+	__le32 rx_frame_duplicates;
+	__le32 rts_success;
+	__le32 rts_failures;
+	__le32 ack_failures;
+	__le32 rx_multicast_frames;
+	__le32 rx_frames_success;
+	__le32 rx_cmac_icv_errors;
+	__le32 rx_cmac_replays;
+	__le32 rx_mgmt_ccmp_replays;
+} __packed;
+
+static inline int wsm_get_counters_table(struct cw1200_common *priv,
+					 struct wsm_mib_counters_table *arg)
+{
+	return wsm_read_mib(priv, WSM_MIB_ID_COUNTERS_TABLE,
+			    arg, sizeof(*arg));
+}
+
+static inline int wsm_get_station_id(struct cw1200_common *priv, u8 *mac)
+{
+	return wsm_read_mib(priv, WSM_MIB_ID_DOT11_STATION_ID, mac, ETH_ALEN);
+}
+
+struct wsm_rx_filter {
+	bool promiscuous;
+	bool bssid;
+	bool fcs;
+	bool probeResponder;
+};
+
+static inline int wsm_set_rx_filter(struct cw1200_common *priv,
+				    const struct wsm_rx_filter *arg)
+{
+	__le32 val = 0;
+	if (arg->promiscuous)
+		val |= __cpu_to_le32(BIT(0));
+	if (arg->bssid)
+		val |= __cpu_to_le32(BIT(1));
+	if (arg->fcs)
+		val |= __cpu_to_le32(BIT(2));
+	if (arg->probeResponder)
+		val |= __cpu_to_le32(BIT(3));
+	return wsm_write_mib(priv, WSM_MIB_ID_RX_FILTER, &val, sizeof(val));
+}
+
+int wsm_set_probe_responder(struct cw1200_common *priv, bool enable);
+
+#define WSM_BEACON_FILTER_IE_HAS_CHANGED	BIT(0)
+#define WSM_BEACON_FILTER_IE_NO_LONGER_PRESENT	BIT(1)
+#define WSM_BEACON_FILTER_IE_HAS_APPEARED	BIT(2)
+
+struct wsm_beacon_filter_table_entry {
+	u8	ie_id;
+	u8	flags;
+	u8	oui[3];
+	u8	match_data[3];
+} __packed;
+
+struct wsm_mib_beacon_filter_table {
+	__le32 num;
+	struct wsm_beacon_filter_table_entry entry[10];
+} __packed;
+
+static inline int wsm_set_beacon_filter_table(struct cw1200_common *priv,
+					      struct wsm_mib_beacon_filter_table *ft)
+{
+	size_t size = __le32_to_cpu(ft->num) *
+		     sizeof(struct wsm_beacon_filter_table_entry) +
+		     sizeof(__le32);
+
+	return wsm_write_mib(priv, WSM_MIB_ID_BEACON_FILTER_TABLE, ft, size);
+}
+
+#define WSM_BEACON_FILTER_ENABLE	BIT(0) /* Enable/disable beacon filtering */
+#define WSM_BEACON_FILTER_AUTO_ERP	BIT(1) /* If 1 FW will handle ERP IE changes internally */
+
+struct wsm_beacon_filter_control {
+	int enabled;
+	int bcn_count;
+};
+
+static inline int wsm_beacon_filter_control(struct cw1200_common *priv,
+					struct wsm_beacon_filter_control *arg)
+{
+	struct {
+		__le32 enabled;
+		__le32 bcn_count;
+	} val;
+	val.enabled = __cpu_to_le32(arg->enabled);
+	val.bcn_count = __cpu_to_le32(arg->bcn_count);
+	return wsm_write_mib(priv, WSM_MIB_ID_BEACON_FILTER_ENABLE, &val,
+			     sizeof(val));
+}
+
+enum wsm_power_mode {
+	wsm_power_mode_active = 0,
+	wsm_power_mode_doze = 1,
+	wsm_power_mode_quiescent = 2,
+};
+
+struct wsm_operational_mode {
+	enum wsm_power_mode power_mode;
+	int disable_more_flag_usage;
+	int perform_ant_diversity;
+};
+
+static inline int wsm_set_operational_mode(struct cw1200_common *priv,
+					const struct wsm_operational_mode *arg)
+{
+	u8 val = arg->power_mode;
+	if (arg->disable_more_flag_usage)
+		val |= BIT(4);
+	if (arg->perform_ant_diversity)
+		val |= BIT(5);
+	return wsm_write_mib(priv, WSM_MIB_ID_OPERATIONAL_POWER_MODE, &val,
+			     sizeof(val));
+}
+
+struct wsm_template_frame {
+	u8 frame_type;
+	u8 rate;
+	struct sk_buff *skb;
+};
+
+static inline int wsm_set_template_frame(struct cw1200_common *priv,
+					 struct wsm_template_frame *arg)
+{
+	int ret;
+	u8 *p = skb_push(arg->skb, 4);
+	p[0] = arg->frame_type;
+	p[1] = arg->rate;
+	((__le16 *)p)[1] = __cpu_to_le16(arg->skb->len - 4);
+	ret = wsm_write_mib(priv, WSM_MIB_ID_TEMPLATE_FRAME, p, arg->skb->len);
+	skb_pull(arg->skb, 4);
+	return ret;
+}
+
+
+struct wsm_protected_mgmt_policy {
+	bool protectedMgmtEnable;
+	bool unprotectedMgmtFramesAllowed;
+	bool encryptionForAuthFrame;
+};
+
+static inline int wsm_set_protected_mgmt_policy(struct cw1200_common *priv,
+		struct wsm_protected_mgmt_policy *arg)
+{
+	__le32 val = 0;
+	int ret;
+	if (arg->protectedMgmtEnable)
+		val |= __cpu_to_le32(BIT(0));
+	if (arg->unprotectedMgmtFramesAllowed)
+		val |= __cpu_to_le32(BIT(1));
+	if (arg->encryptionForAuthFrame)
+		val |= __cpu_to_le32(BIT(2));
+	ret = wsm_write_mib(priv, WSM_MIB_ID_PROTECTED_MGMT_POLICY,
+			&val, sizeof(val));
+	return ret;
+}
+
+struct wsm_mib_block_ack_policy {
+	u8 tx_tid;
+	u8 reserved1;
+	u8 rx_tid;
+	u8 reserved2;
+} __packed;
+
+static inline int wsm_set_block_ack_policy(struct cw1200_common *priv,
+					   u8 tx_tid_policy,
+					   u8 rx_tid_policy)
+{
+	struct wsm_mib_block_ack_policy val = {
+		.tx_tid = tx_tid_policy,
+		.rx_tid = rx_tid_policy,
+	};
+	return wsm_write_mib(priv, WSM_MIB_ID_BLOCK_ACK_POLICY, &val,
+			     sizeof(val));
+}
+
+struct wsm_mib_association_mode {
+	u8 flags;		/* WSM_ASSOCIATION_MODE_... */
+	u8 preamble;	/* WSM_JOIN_PREAMBLE_... */
+	u8 greenfield;	/* 1 for greenfield */
+	u8 mpdu_start_spacing;
+	__le32 basic_rate_set;
+} __packed;
+
+static inline int wsm_set_association_mode(struct cw1200_common *priv,
+					   struct wsm_mib_association_mode *arg)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_SET_ASSOCIATION_MODE, arg,
+			     sizeof(*arg));
+}
+
+#define WSM_TX_RATE_POLICY_FLAG_TERMINATE_WHEN_FINISHED BIT(2)
+#define WSM_TX_RATE_POLICY_FLAG_COUNT_INITIAL_TRANSMIT BIT(3)
+struct wsm_tx_rate_retry_policy {
+	u8 index;
+	u8 short_retries;
+	u8 long_retries;
+	/* BIT(2) - Terminate retries when Tx rate retry policy
+	 *          finishes.
+	 * BIT(3) - Count initial frame transmission as part of
+	 *          rate retry counting but not as a retry
+	 *          attempt
+	 */
+	u8 flags;
+	u8 rate_recoveries;
+	u8 reserved[3];
+	__le32 rate_count_indices[3];
+} __packed;
+
+struct wsm_set_tx_rate_retry_policy {
+	u8 num;
+	u8 reserved[3];
+	struct wsm_tx_rate_retry_policy tbl[8];
+} __packed;
+
+static inline int wsm_set_tx_rate_retry_policy(struct cw1200_common *priv,
+				struct wsm_set_tx_rate_retry_policy *arg)
+{
+	size_t size = 4 + arg->num * sizeof(struct wsm_tx_rate_retry_policy);
+	return wsm_write_mib(priv, WSM_MIB_ID_SET_TX_RATE_RETRY_POLICY, arg,
+			     size);
+}
+
+/* 4.32 SetEtherTypeDataFrameFilter */
+struct wsm_ether_type_filter_hdr {
+	u8 num;		/* Up to WSM_MAX_FILTER_ELEMENTS */
+	u8 reserved[3];
+} __packed;
+
+struct wsm_ether_type_filter {
+	u8 action;	/* WSM_FILTER_ACTION_XXX */
+	u8 reserved;
+	__le16 type;	/* Type of ethernet frame */
+} __packed;
+
+static inline int wsm_set_ether_type_filter(struct cw1200_common *priv,
+				struct wsm_ether_type_filter_hdr *arg)
+{
+	size_t size = sizeof(struct wsm_ether_type_filter_hdr) +
+		arg->num * sizeof(struct wsm_ether_type_filter);
+	return wsm_write_mib(priv, WSM_MIB_ID_SET_ETHERTYPE_DATAFRAME_FILTER,
+		arg, size);
+}
+
+/* 4.33 SetUDPPortDataFrameFilter */
+struct wsm_udp_port_filter_hdr {
+	u8 num;		/* Up to WSM_MAX_FILTER_ELEMENTS */
+	u8 reserved[3];
+} __packed;
+
+struct wsm_udp_port_filter {
+	u8 action;	/* WSM_FILTER_ACTION_XXX */
+	u8 type;		/* WSM_FILTER_PORT_TYPE_XXX */
+	__le16 port;		/* Port number */
+} __packed;
+
+static inline int wsm_set_udp_port_filter(struct cw1200_common *priv,
+				struct wsm_udp_port_filter_hdr *arg)
+{
+	size_t size = sizeof(struct wsm_udp_port_filter_hdr) +
+		arg->num * sizeof(struct wsm_udp_port_filter);
+	return wsm_write_mib(priv, WSM_MIB_ID_SET_UDPPORT_DATAFRAME_FILTER,
+		arg, size);
+}
+
+/* Undocumented MIBs: */
+/* 4.35 P2PDeviceInfo */
+#define D11_MAX_SSID_LEN		(32)
+
+struct wsm_p2p_device_type {
+	__le16 category_id;
+	u8 oui[4];
+	__le16 subcategory_id;
+} __packed;
+
+struct wsm_p2p_device_info {
+	struct wsm_p2p_device_type primaryDevice;
+	u8 reserved1[3];
+	u8 devname_size;
+	u8 local_devname[D11_MAX_SSID_LEN];
+	u8 reserved2[3];
+	u8 num_secdev_supported;
+	struct wsm_p2p_device_type secdevs[0];
+} __packed;
+
+/* 4.36 SetWCDMABand - WO */
+struct wsm_cdma_band {
+	u8 wcdma_band;
+	u8 reserved[3];
+} __packed;
+
+/* 4.37 GroupTxSequenceCounter - RO */
+struct wsm_group_tx_seq {
+	__le32 bits_47_16;
+	__le16 bits_15_00;
+	__le16 reserved;
+} __packed;
+
+/* 4.39 SetHtProtection - WO */
+#define WSM_DUAL_CTS_PROT_ENB		(1 << 0)
+#define WSM_NON_GREENFIELD_STA_PRESENT  (1 << 1)
+#define WSM_HT_PROT_MODE__NO_PROT	(0 << 2)
+#define WSM_HT_PROT_MODE__NON_MEMBER	(1 << 2)
+#define WSM_HT_PROT_MODE__20_MHZ	(2 << 2)
+#define WSM_HT_PROT_MODE__NON_HT_MIXED	(3 << 2)
+#define WSM_LSIG_TXOP_PROT_FULL		(1 << 4)
+#define WSM_LARGE_L_LENGTH_PROT		(1 << 5)
+
+struct wsm_ht_protection {
+	__le32 flags;
+} __packed;
+
+/* 4.40 GPIO Command - R/W */
+#define WSM_GPIO_COMMAND_SETUP	0
+#define WSM_GPIO_COMMAND_READ	1
+#define WSM_GPIO_COMMAND_WRITE	2
+#define WSM_GPIO_COMMAND_RESET	3
+#define WSM_GPIO_ALL_PINS	0xFF
+
+struct wsm_gpio_command {
+	u8 command;
+	u8 pin;
+	__le16 config;
+} __packed;
+
+/* 4.41 TSFCounter - RO */
+struct wsm_tsf_counter {
+	__le64 tsf_counter;
+} __packed;
+
+/* 4.43 Keep alive period */
+struct wsm_keep_alive_period {
+	__le16 period;
+	u8 reserved[2];
+} __packed;
+
+static inline int wsm_keep_alive_period(struct cw1200_common *priv,
+					int period)
+{
+	struct wsm_keep_alive_period arg = {
+		.period = __cpu_to_le16(period),
+	};
+	return wsm_write_mib(priv, WSM_MIB_ID_KEEP_ALIVE_PERIOD,
+			&arg, sizeof(arg));
+};
+
+/* BSSID filtering */
+struct wsm_set_bssid_filtering {
+	u8 filter;
+	u8 reserved[3];
+} __packed;
+
+static inline int wsm_set_bssid_filtering(struct cw1200_common *priv,
+					  bool enabled)
+{
+	struct wsm_set_bssid_filtering arg = {
+		.filter = !enabled,
+	};
+	return wsm_write_mib(priv, WSM_MIB_ID_DISABLE_BSSID_FILTER,
+			&arg, sizeof(arg));
+}
+
+/* Multicast filtering - 4.5 */
+struct wsm_mib_multicast_filter {
+	__le32 enable;
+	__le32 num_addrs;
+	u8 macaddrs[WSM_MAX_GRP_ADDRTABLE_ENTRIES][ETH_ALEN];
+} __packed;
+
+static inline int wsm_set_multicast_filter(struct cw1200_common *priv,
+					   struct wsm_mib_multicast_filter *fp)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_DOT11_GROUP_ADDRESSES_TABLE,
+			     fp, sizeof(*fp));
+}
+
+/* ARP IPv4 filtering - 4.10 */
+struct wsm_mib_arp_ipv4_filter {
+	__le32 enable;
+	__be32 ipv4addrs[WSM_MAX_ARP_IP_ADDRTABLE_ENTRIES];
+} __packed;
+
+static inline int wsm_set_arp_ipv4_filter(struct cw1200_common *priv,
+					  struct wsm_mib_arp_ipv4_filter *fp)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_ARP_IP_ADDRESSES_TABLE,
+			    fp, sizeof(*fp));
+}
+
+/* P2P Power Save Mode Info - 4.31 */
+struct wsm_p2p_ps_modeinfo {
+	u8	opp_ps_ct_window;
+	u8	count;
+	u8	reserved;
+	u8	dtim_count;
+	__le32	duration;
+	__le32	interval;
+	__le32	start_time;
+} __packed;
+
+static inline int wsm_set_p2p_ps_modeinfo(struct cw1200_common *priv,
+					  struct wsm_p2p_ps_modeinfo *mi)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_P2P_PS_MODE_INFO,
+			     mi, sizeof(*mi));
+}
+
+static inline int wsm_get_p2p_ps_modeinfo(struct cw1200_common *priv,
+					  struct wsm_p2p_ps_modeinfo *mi)
+{
+	return wsm_read_mib(priv, WSM_MIB_ID_P2P_PS_MODE_INFO,
+			    mi, sizeof(*mi));
+}
+
+/* UseMultiTxConfMessage */
+
+static inline int wsm_use_multi_tx_conf(struct cw1200_common *priv,
+					bool enabled)
+{
+	__le32 arg = enabled ? __cpu_to_le32(1) : 0;
+
+	return wsm_write_mib(priv, WSM_MIB_USE_MULTI_TX_CONF,
+			&arg, sizeof(arg));
+}
+
+
+/* 4.26 SetUpasdInformation */
+struct wsm_uapsd_info {
+	__le16 uapsd_flags;
+	__le16 min_auto_trigger_interval;
+	__le16 max_auto_trigger_interval;
+	__le16 auto_trigger_step;
+};
+
+static inline int wsm_set_uapsd_info(struct cw1200_common *priv,
+				     struct wsm_uapsd_info *arg)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_SET_UAPSD_INFORMATION,
+				arg, sizeof(*arg));
+}
+
+/* 4.22 OverrideInternalTxRate */
+struct wsm_override_internal_txrate {
+	u8 internalTxRate;
+	u8 nonErpInternalTxRate;
+	u8 reserved[2];
+} __packed;
+
+static inline int wsm_set_override_internal_txrate(struct cw1200_common *priv,
+				     struct wsm_override_internal_txrate *arg)
+{
+	return wsm_write_mib(priv, WSM_MIB_ID_OVERRIDE_INTERNAL_TX_RATE,
+				arg, sizeof(*arg));
+}
+
+/* ******************************************************************** */
+/* WSM TX port control							*/
+
+void wsm_lock_tx(struct cw1200_common *priv);
+void wsm_lock_tx_async(struct cw1200_common *priv);
+bool wsm_flush_tx(struct cw1200_common *priv);
+void wsm_unlock_tx(struct cw1200_common *priv);
+
+/* ******************************************************************** */
+/* WSM / BH API								*/
+
+int wsm_handle_exception(struct cw1200_common *priv, u8 *data, size_t len);
+int wsm_handle_rx(struct cw1200_common *priv, u16 id, struct wsm_hdr *wsm,
+		  struct sk_buff **skb_p);
+
+/* ******************************************************************** */
+/* wsm_buf API								*/
+
+struct wsm_buf {
+	u8 *begin;
+	u8 *data;
+	u8 *end;
+};
+
+void wsm_buf_init(struct wsm_buf *buf);
+void wsm_buf_deinit(struct wsm_buf *buf);
+
+/* ******************************************************************** */
+/* wsm_cmd API								*/
+
+struct wsm_cmd {
+	spinlock_t lock; /* Protect structure from multiple access */
+	int done;
+	u8 *ptr;
+	size_t len;
+	void *arg;
+	int ret;
+	u16 cmd;
+};
+
+/* ******************************************************************** */
+/* WSM TX buffer access							*/
+
+int wsm_get_tx(struct cw1200_common *priv, u8 **data,
+	       size_t *tx_len, int *burst);
+void wsm_txed(struct cw1200_common *priv, u8 *data);
+
+/* ******************************************************************** */
+/* Queue mapping: WSM <---> linux					*/
+/* Linux: VO VI BE BK							*/
+/* WSM:   BE BK VI VO							*/
+
+static inline u8 wsm_queue_id_to_linux(u8 queue_id)
+{
+	static const u8 queue_mapping[] = {
+		2, 3, 1, 0
+	};
+	return queue_mapping[queue_id];
+}
+
+static inline u8 wsm_queue_id_to_wsm(u8 queue_id)
+{
+	static const u8 queue_mapping[] = {
+		3, 2, 0, 1
+	};
+	return queue_mapping[queue_id];
+}
+
+#endif /* CW1200_HWIO_H_INCLUDED */
diff --git a/drivers/net/wireless/ipw2x00/ipw2100.c b/drivers/net/wireless/ipw2x00/ipw2100.c
index 15920aa..f8ab193 100644
--- a/drivers/net/wireless/ipw2x00/ipw2100.c
+++ b/drivers/net/wireless/ipw2x00/ipw2100.c
@@ -6242,8 +6242,6 @@ static int ipw2100_pci_init_one(struct pci_dev *pci_dev,
 	if ((val & 0x0000ff00) != 0)
 		pci_write_config_dword(pci_dev, 0x40, val & 0xffff00ff);
 
-	pci_set_power_state(pci_dev, PCI_D0);
-
 	if (!ipw2100_hw_is_adapter_in_system(dev)) {
 		printk(KERN_WARNING DRV_NAME
 		       "Device not found via register read.\n");
diff --git a/drivers/net/wireless/ipw2x00/ipw2200.c b/drivers/net/wireless/ipw2x00/ipw2200.c
index 4ed5e45..6b823a1 100644
--- a/drivers/net/wireless/ipw2x00/ipw2200.c
+++ b/drivers/net/wireless/ipw2x00/ipw2200.c
@@ -3548,6 +3548,7 @@ static int ipw_load(struct ipw_priv *priv)
 		ipw_rx_queue_reset(priv, priv->rxq);
 	if (!priv->rxq) {
 		IPW_ERROR("Unable to initialize Rx queue\n");
+		rc = -ENOMEM;
 		goto error;
 	}
 
diff --git a/drivers/net/wireless/ipw2x00/libipw_rx.c b/drivers/net/wireless/ipw2x00/libipw_rx.c
index 95a1ca1..9ffe659 100644
--- a/drivers/net/wireless/ipw2x00/libipw_rx.c
+++ b/drivers/net/wireless/ipw2x00/libipw_rx.c
@@ -1195,7 +1195,7 @@ static int libipw_parse_info_param(struct libipw_info_element
 #ifdef CONFIG_LIBIPW_DEBUG
 				p += snprintf(p, sizeof(rates_str) -
 					      (p - rates_str), "%02X ",
-					      network->rates[i]);
+					      network->rates_ex[i]);
 #endif
 				if (libipw_is_ofdm_rate
 				    (info_element->data[i])) {
diff --git a/drivers/net/wireless/iwlegacy/3945-mac.c b/drivers/net/wireless/iwlegacy/3945-mac.c
index b37a582..9581d07 100644
--- a/drivers/net/wireless/iwlegacy/3945-mac.c
+++ b/drivers/net/wireless/iwlegacy/3945-mac.c
@@ -3119,7 +3119,7 @@ il3945_store_debug_level(struct device *d, struct device_attribute *attr,
 	unsigned long val;
 	int ret;
 
-	ret = strict_strtoul(buf, 0, &val);
+	ret = kstrtoul(buf, 0, &val);
 	if (ret)
 		IL_INFO("%s is not in hex or decimal form.\n", buf);
 	else
@@ -3727,7 +3727,8 @@ il3945_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	 * 5. Setup HW Constants
 	 * ********************/
 	/* Device-specific setup */
-	if (il3945_hw_set_hw_params(il)) {
+	err = il3945_hw_set_hw_params(il);
+	if (err) {
 		IL_ERR("failed to set hw settings\n");
 		goto out_eeprom_free;
 	}
diff --git a/drivers/net/wireless/iwlegacy/3945.c b/drivers/net/wireless/iwlegacy/3945.c
index dc1e6da..c092033 100644
--- a/drivers/net/wireless/iwlegacy/3945.c
+++ b/drivers/net/wireless/iwlegacy/3945.c
@@ -331,6 +331,19 @@ il3945_hdl_tx(struct il_priv *il, struct il_rx_buf *rxb)
 		return;
 	}
 
+	/*
+	 * Firmware will not transmit frame on passive channel, if it not yet
+	 * received some valid frame on that channel. When this error happen
+	 * we have to wait until firmware will unblock itself i.e. when we
+	 * note received beacon or other frame. We unblock queues in
+	 * il3945_pass_packet_to_mac80211 or in il_mac_bss_info_changed.
+	 */
+	if (unlikely((status & TX_STATUS_MSK) == TX_STATUS_FAIL_PASSIVE_NO_RX) &&
+	    il->iw_mode == NL80211_IFTYPE_STATION) {
+		il_stop_queues_by_reason(il, IL_STOP_REASON_PASSIVE);
+		D_INFO("Stopped queues - RX waiting on passive channel\n");
+	}
+
 	txq->time_stamp = jiffies;
 	info = IEEE80211_SKB_CB(txq->skbs[txq->q.read_ptr]);
 	ieee80211_tx_info_clear_status(info);
@@ -488,6 +501,11 @@ il3945_pass_packet_to_mac80211(struct il_priv *il, struct il_rx_buf *rxb,
 		return;
 	}
 
+	if (unlikely(test_bit(IL_STOP_REASON_PASSIVE, &il->stop_reason))) {
+		il_wake_queues_by_reason(il, IL_STOP_REASON_PASSIVE);
+		D_INFO("Woke queues - frame received on passive channel\n");
+	}
+
 	skb = dev_alloc_skb(128);
 	if (!skb) {
 		IL_ERR("dev_alloc_skb failed\n");
diff --git a/drivers/net/wireless/iwlegacy/4965-mac.c b/drivers/net/wireless/iwlegacy/4965-mac.c
index 9a95045..b9b2bb5 100644
--- a/drivers/net/wireless/iwlegacy/4965-mac.c
+++ b/drivers/net/wireless/iwlegacy/4965-mac.c
@@ -588,6 +588,11 @@ il4965_pass_packet_to_mac80211(struct il_priv *il, struct ieee80211_hdr *hdr,
 		return;
 	}
 
+	if (unlikely(test_bit(IL_STOP_REASON_PASSIVE, &il->stop_reason))) {
+		il_wake_queues_by_reason(il, IL_STOP_REASON_PASSIVE);
+		D_INFO("Woke queues - frame received on passive channel\n");
+	}
+
 	/* In case of HW accelerated crypto and bad decryption, drop */
 	if (!il->cfg->mod_params->sw_crypto &&
 	    il_set_decrypted_flag(il, hdr, ampdu_status, stats))
@@ -2806,6 +2811,19 @@ il4965_hdl_tx(struct il_priv *il, struct il_rx_buf *rxb)
 		return;
 	}
 
+	/*
+	 * Firmware will not transmit frame on passive channel, if it not yet
+	 * received some valid frame on that channel. When this error happen
+	 * we have to wait until firmware will unblock itself i.e. when we
+	 * note received beacon or other frame. We unblock queues in
+	 * il4965_pass_packet_to_mac80211 or in il_mac_bss_info_changed.
+	 */
+	if (unlikely((status & TX_STATUS_MSK) == TX_STATUS_FAIL_PASSIVE_NO_RX) &&
+	    il->iw_mode == NL80211_IFTYPE_STATION) {
+		il_stop_queues_by_reason(il, IL_STOP_REASON_PASSIVE);
+		D_INFO("Stopped queues - RX waiting on passive channel\n");
+	}
+
 	spin_lock_irqsave(&il->sta_lock, flags);
 	if (txq->sched_retry) {
 		const u32 scd_ssn = il4965_get_scd_ssn(tx_resp);
@@ -4567,7 +4585,7 @@ il4965_store_debug_level(struct device *d, struct device_attribute *attr,
 	unsigned long val;
 	int ret;
 
-	ret = strict_strtoul(buf, 0, &val);
+	ret = kstrtoul(buf, 0, &val);
 	if (ret)
 		IL_ERR("%s is not in hex or decimal form.\n", buf);
 	else
@@ -4614,7 +4632,7 @@ il4965_store_tx_power(struct device *d, struct device_attribute *attr,
 	unsigned long val;
 	int ret;
 
-	ret = strict_strtoul(buf, 10, &val);
+	ret = kstrtoul(buf, 10, &val);
 	if (ret)
 		IL_INFO("%s is not in decimal form.\n", buf);
 	else {
@@ -5741,7 +5759,8 @@ il4965_mac_setup_register(struct il_priv *il, u32 max_probe_length)
 	hw->flags =
 	    IEEE80211_HW_SIGNAL_DBM | IEEE80211_HW_AMPDU_AGGREGATION |
 	    IEEE80211_HW_NEED_DTIM_BEFORE_ASSOC | IEEE80211_HW_SPECTRUM_MGMT |
-	    IEEE80211_HW_SUPPORTS_PS | IEEE80211_HW_SUPPORTS_DYNAMIC_PS;
+	    IEEE80211_HW_REPORTS_TX_ACK_STATUS | IEEE80211_HW_SUPPORTS_PS |
+	    IEEE80211_HW_SUPPORTS_DYNAMIC_PS;
 	if (il->cfg->sku & IL_SKU_N)
 		hw->flags |=
 		    IEEE80211_HW_SUPPORTS_DYNAMIC_SMPS |
diff --git a/drivers/net/wireless/iwlegacy/commands.h b/drivers/net/wireless/iwlegacy/commands.h
index 3b6c994..0484215 100644
--- a/drivers/net/wireless/iwlegacy/commands.h
+++ b/drivers/net/wireless/iwlegacy/commands.h
@@ -1348,14 +1348,6 @@ struct il_rx_mpdu_res_start {
 #define TX_CMD_SEC_KEY128	0x08
 
 /*
- * security overhead sizes
- */
-#define WEP_IV_LEN 4
-#define WEP_ICV_LEN 4
-#define CCMP_MIC_LEN 8
-#define TKIP_ICV_LEN 4
-
-/*
  * C_TX = 0x1c (command)
  */
 
diff --git a/drivers/net/wireless/iwlegacy/common.c b/drivers/net/wireless/iwlegacy/common.c
index e9a3cbc..3195aad 100644
--- a/drivers/net/wireless/iwlegacy/common.c
+++ b/drivers/net/wireless/iwlegacy/common.c
@@ -5307,6 +5307,17 @@ il_mac_bss_info_changed(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
 		D_MAC80211("BSSID %pM\n", bss_conf->bssid);
 
 		/*
+		 * On passive channel we wait with blocked queues to see if
+		 * there is traffic on that channel. If no frame will be
+		 * received (what is very unlikely since scan detects AP on
+		 * that channel, but theoretically possible), mac80211 associate
+		 * procedure will time out and mac80211 will call us with NULL
+		 * bssid. We have to unblock queues on such condition.
+		 */
+		if (is_zero_ether_addr(bss_conf->bssid))
+			il_wake_queues_by_reason(il, IL_STOP_REASON_PASSIVE);
+
+		/*
 		 * If there is currently a HW scan going on in the background,
 		 * then we need to cancel it, otherwise sometimes we are not
 		 * able to authenticate (FIXME: why ?)
diff --git a/drivers/net/wireless/iwlegacy/common.h b/drivers/net/wireless/iwlegacy/common.h
index 4caaf52..83f8ed8 100644
--- a/drivers/net/wireless/iwlegacy/common.h
+++ b/drivers/net/wireless/iwlegacy/common.h
@@ -1299,6 +1299,8 @@ struct il_priv {
 	/* queue refcounts */
 #define IL_MAX_HW_QUEUES	32
 	unsigned long queue_stopped[BITS_TO_LONGS(IL_MAX_HW_QUEUES)];
+#define IL_STOP_REASON_PASSIVE	0
+	unsigned long stop_reason;
 	/* for each AC */
 	atomic_t queue_stop_count[4];
 
@@ -2257,6 +2259,19 @@ il_set_swq_id(struct il_tx_queue *txq, u8 ac, u8 hwq)
 }
 
 static inline void
+_il_wake_queue(struct il_priv *il, u8 ac)
+{
+	if (atomic_dec_return(&il->queue_stop_count[ac]) <= 0)
+		ieee80211_wake_queue(il->hw, ac);
+}
+
+static inline void
+_il_stop_queue(struct il_priv *il, u8 ac)
+{
+	if (atomic_inc_return(&il->queue_stop_count[ac]) > 0)
+		ieee80211_stop_queue(il->hw, ac);
+}
+static inline void
 il_wake_queue(struct il_priv *il, struct il_tx_queue *txq)
 {
 	u8 queue = txq->swq_id;
@@ -2264,8 +2279,7 @@ il_wake_queue(struct il_priv *il, struct il_tx_queue *txq)
 	u8 hwq = (queue >> 2) & 0x1f;
 
 	if (test_and_clear_bit(hwq, il->queue_stopped))
-		if (atomic_dec_return(&il->queue_stop_count[ac]) <= 0)
-			ieee80211_wake_queue(il->hw, ac);
+		_il_wake_queue(il, ac);
 }
 
 static inline void
@@ -2276,8 +2290,27 @@ il_stop_queue(struct il_priv *il, struct il_tx_queue *txq)
 	u8 hwq = (queue >> 2) & 0x1f;
 
 	if (!test_and_set_bit(hwq, il->queue_stopped))
-		if (atomic_inc_return(&il->queue_stop_count[ac]) > 0)
-			ieee80211_stop_queue(il->hw, ac);
+		_il_stop_queue(il, ac);
+}
+
+static inline void
+il_wake_queues_by_reason(struct il_priv *il, int reason)
+{
+	u8 ac;
+
+	if (test_and_clear_bit(reason, &il->stop_reason))
+		for (ac = 0; ac < 4; ac++)
+			_il_wake_queue(il, ac);
+}
+
+static inline void
+il_stop_queues_by_reason(struct il_priv *il, int reason)
+{
+	u8 ac;
+
+	if (!test_and_set_bit(reason, &il->stop_reason))
+		for (ac = 0; ac < 4; ac++)
+			_il_stop_queue(il, ac);
 }
 
 #ifdef ieee80211_stop_queue
diff --git a/drivers/net/wireless/iwlwifi/Kconfig b/drivers/net/wireless/iwlwifi/Kconfig
index 56c2040..cbaa5c2 100644
--- a/drivers/net/wireless/iwlwifi/Kconfig
+++ b/drivers/net/wireless/iwlwifi/Kconfig
@@ -128,16 +128,6 @@ config IWLWIFI_DEVICE_TRACING
 	  occur.
 endmenu
 
-config IWLWIFI_DEVICE_TESTMODE
-	def_bool y
-	depends on IWLWIFI
-	depends on NL80211_TESTMODE
-	help
-	  This option enables the testmode support for iwlwifi device through
-	  NL80211_TESTMODE. This provide the capabilities of enable user space
-	  validation applications to interacts with the device through the
-	  generic netlink message via NL80211_TESTMODE channel.
-
 config IWLWIFI_P2P
 	def_bool y
 	bool "iwlwifi experimental P2P support"
diff --git a/drivers/net/wireless/iwlwifi/Makefile b/drivers/net/wireless/iwlwifi/Makefile
index 3b5613e..1fa6442 100644
--- a/drivers/net/wireless/iwlwifi/Makefile
+++ b/drivers/net/wireless/iwlwifi/Makefile
@@ -7,14 +7,15 @@ iwlwifi-objs		+= iwl-notif-wait.o
 iwlwifi-objs		+= iwl-eeprom-read.o iwl-eeprom-parse.o
 iwlwifi-objs		+= iwl-phy-db.o iwl-nvm-parse.o
 iwlwifi-objs		+= pcie/drv.o pcie/rx.o pcie/tx.o pcie/trans.o
-iwlwifi-objs		+= iwl-1000.o iwl-2000.o iwl-5000.o iwl-6000.o iwl-7000.o
+iwlwifi-$(CONFIG_IWLDVM) += iwl-1000.o iwl-2000.o iwl-5000.o iwl-6000.o
+iwlwifi-$(CONFIG_IWLMVM) += iwl-7000.o
+
+iwlwifi-objs += $(iwlwifi-m)
 
 iwlwifi-$(CONFIG_IWLWIFI_DEVICE_TRACING) += iwl-devtrace.o
-iwlwifi-$(CONFIG_IWLWIFI_DEVICE_TESTMODE) += iwl-test.o
 
 ccflags-y += -D__CHECK_ENDIAN__ -I$(src)
 
-
 obj-$(CONFIG_IWLDVM)	+= dvm/
 obj-$(CONFIG_IWLMVM)	+= mvm/
 
diff --git a/drivers/net/wireless/iwlwifi/dvm/Makefile b/drivers/net/wireless/iwlwifi/dvm/Makefile
index 5ff76b2..dce7ab2 100644
--- a/drivers/net/wireless/iwlwifi/dvm/Makefile
+++ b/drivers/net/wireless/iwlwifi/dvm/Makefile
@@ -8,6 +8,5 @@ iwldvm-objs		+= scan.o led.o
 iwldvm-objs		+= rxon.o devices.o
 
 iwldvm-$(CONFIG_IWLWIFI_DEBUGFS) += debugfs.o
-iwldvm-$(CONFIG_IWLWIFI_DEVICE_TESTMODE) += testmode.o
 
 ccflags-y += -D__CHECK_ENDIAN__ -I$(src)/../
diff --git a/drivers/net/wireless/iwlwifi/dvm/agn.h b/drivers/net/wireless/iwlwifi/dvm/agn.h
index 48545ab..1835511 100644
--- a/drivers/net/wireless/iwlwifi/dvm/agn.h
+++ b/drivers/net/wireless/iwlwifi/dvm/agn.h
@@ -76,13 +76,16 @@
 #define IWL_INVALID_STATION	255
 
 /* device operations */
-extern struct iwl_lib_ops iwl1000_lib;
-extern struct iwl_lib_ops iwl2000_lib;
-extern struct iwl_lib_ops iwl2030_lib;
-extern struct iwl_lib_ops iwl5000_lib;
-extern struct iwl_lib_ops iwl5150_lib;
-extern struct iwl_lib_ops iwl6000_lib;
-extern struct iwl_lib_ops iwl6030_lib;
+extern const struct iwl_dvm_cfg iwl_dvm_1000_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_2000_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_105_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_2030_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_5000_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_5150_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_6000_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_6005_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_6050_cfg;
+extern const struct iwl_dvm_cfg iwl_dvm_6030_cfg;
 
 
 #define TIME_UNIT		1024
@@ -291,8 +294,8 @@ void iwlagn_bt_adjust_rssi_monitor(struct iwl_priv *priv, bool rssi_ena);
 
 static inline bool iwl_advanced_bt_coexist(struct iwl_priv *priv)
 {
-	return priv->cfg->bt_params &&
-	       priv->cfg->bt_params->advanced_bt_coexist;
+	return priv->lib->bt_params &&
+	       priv->lib->bt_params->advanced_bt_coexist;
 }
 
 #ifdef CONFIG_IWLWIFI_DEBUG
@@ -402,43 +405,6 @@ static inline __le32 iwl_hw_set_rate_n_flags(u8 rate, u32 flags)
 
 extern int iwl_alive_start(struct iwl_priv *priv);
 
-/* testmode support */
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-
-extern int iwlagn_mac_testmode_cmd(struct ieee80211_hw *hw, void *data,
-				   int len);
-extern int iwlagn_mac_testmode_dump(struct ieee80211_hw *hw,
-				    struct sk_buff *skb,
-				    struct netlink_callback *cb,
-				    void *data, int len);
-extern void iwl_testmode_init(struct iwl_priv *priv);
-extern void iwl_testmode_free(struct iwl_priv *priv);
-
-#else
-
-static inline
-int iwlagn_mac_testmode_cmd(struct ieee80211_hw *hw, void *data, int len)
-{
-	return -ENOSYS;
-}
-
-static inline
-int iwlagn_mac_testmode_dump(struct ieee80211_hw *hw, struct sk_buff *skb,
-		      struct netlink_callback *cb,
-		      void *data, int len)
-{
-	return -ENOSYS;
-}
-
-static inline void iwl_testmode_init(struct iwl_priv *priv)
-{
-}
-
-static inline void iwl_testmode_free(struct iwl_priv *priv)
-{
-}
-#endif
-
 #ifdef CONFIG_IWLWIFI_DEBUG
 void iwl_print_rx_config_cmd(struct iwl_priv *priv,
 			     enum iwl_rxon_context_id ctxid);
diff --git a/drivers/net/wireless/iwlwifi/dvm/calib.c b/drivers/net/wireless/iwlwifi/dvm/calib.c
index d6c4cf2..1b0f0d5 100644
--- a/drivers/net/wireless/iwlwifi/dvm/calib.c
+++ b/drivers/net/wireless/iwlwifi/dvm/calib.c
@@ -521,7 +521,7 @@ static int iwl_enhance_sensitivity_write(struct iwl_priv *priv)
 
 	iwl_prepare_legacy_sensitivity_tbl(priv, data, &cmd.enhance_table[0]);
 
-	if (priv->cfg->base_params->hd_v2) {
+	if (priv->lib->hd_v2) {
 		cmd.enhance_table[HD_INA_NON_SQUARE_DET_OFDM_INDEX] =
 			HD_INA_NON_SQUARE_DET_OFDM_DATA_V2;
 		cmd.enhance_table[HD_INA_NON_SQUARE_DET_CCK_INDEX] =
@@ -895,7 +895,7 @@ static void iwlagn_gain_computation(struct iwl_priv *priv,
 			continue;
 		}
 
-		delta_g = (priv->cfg->base_params->chain_noise_scale *
+		delta_g = (priv->lib->chain_noise_scale *
 			((s32)average_noise[default_chain] -
 			(s32)average_noise[i])) / 1500;
 
@@ -1051,8 +1051,8 @@ void iwl_chain_noise_calibration(struct iwl_priv *priv)
 		return;
 
 	/* Analyze signal for disconnected antenna */
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist) {
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		/* Disable disconnected antenna algorithm for advanced
 		   bt coex, assuming valid antennas are connected */
 		data->active_chains = priv->nvm_data->valid_rx_ant;
diff --git a/drivers/net/wireless/iwlwifi/dvm/commands.h b/drivers/net/wireless/iwlwifi/dvm/commands.h
index 95ca026..ebdac90 100644
--- a/drivers/net/wireless/iwlwifi/dvm/commands.h
+++ b/drivers/net/wireless/iwlwifi/dvm/commands.h
@@ -838,10 +838,6 @@ struct iwl_qosparam_cmd {
 #define STA_MODIFY_DELBA_TID_MSK	0x10
 #define STA_MODIFY_SLEEP_TX_COUNT_MSK	0x20
 
-/* Receiver address (actually, Rx station's index into station table),
- * combined with Traffic ID (QOS priority), in format used by Tx Scheduler */
-#define BUILD_RAxTID(sta_id, tid)	(((sta_id) << 4) + (tid))
-
 /* agn */
 struct iwl_keyinfo {
 	__le16 key_flags;
@@ -1225,14 +1221,6 @@ struct iwl_rx_mpdu_res_start {
 #define TX_CMD_SEC_KEY128	0x08
 
 /*
- * security overhead sizes
- */
-#define WEP_IV_LEN 4
-#define WEP_ICV_LEN 4
-#define CCMP_MIC_LEN 8
-#define TKIP_ICV_LEN 4
-
-/*
  * REPLY_TX = 0x1c (command)
  */
 
diff --git a/drivers/net/wireless/iwlwifi/dvm/dev.h b/drivers/net/wireless/iwlwifi/dvm/dev.h
index 71ea775..60a4e0d 100644
--- a/drivers/net/wireless/iwlwifi/dvm/dev.h
+++ b/drivers/net/wireless/iwlwifi/dvm/dev.h
@@ -52,8 +52,6 @@
 #include "rs.h"
 #include "tt.h"
 
-#include "iwl-test.h"
-
 /* CT-KILL constants */
 #define CT_KILL_THRESHOLD_LEGACY   110 /* in Celsius */
 #define CT_KILL_THRESHOLD	   114 /* in Celsius */
@@ -568,16 +566,61 @@ struct iwl_hw_params {
 	const struct iwl_sensitivity_ranges *sens;
 };
 
-struct iwl_lib_ops {
-	/* set hw dependent parameters */
+/**
+ * struct iwl_dvm_bt_params - DVM specific BT (coex) parameters
+ * @advanced_bt_coexist: support advanced bt coexist
+ * @bt_init_traffic_load: specify initial bt traffic load
+ * @bt_prio_boost: default bt priority boost value
+ * @agg_time_limit: maximum number of uSec in aggregation
+ * @bt_sco_disable: uCode should not response to BT in SCO/ESCO mode
+ */
+struct iwl_dvm_bt_params {
+	bool advanced_bt_coexist;
+	u8 bt_init_traffic_load;
+	u32 bt_prio_boost;
+	u16 agg_time_limit;
+	bool bt_sco_disable;
+	bool bt_session_2;
+};
+
+/**
+ * struct iwl_dvm_cfg - DVM firmware specific device configuration
+ * @set_hw_params: set hardware parameters
+ * @set_channel_switch: send channel switch command
+ * @nic_config: apply device specific configuration
+ * @temperature: read temperature
+ * @adv_thermal_throttle: support advance thermal throttle
+ * @support_ct_kill_exit: support ct kill exit condition
+ * @plcp_delta_threshold: plcp error rate threshold used to trigger
+ *	radio tuning when there is a high receiving plcp error rate
+ * @chain_noise_scale: default chain noise scale used for gain computation
+ * @hd_v2: v2 of enhanced sensitivity value, used for 2000 series and up
+ * @no_idle_support: do not support idle mode
+ * @bt_params: pointer to BT parameters
+ * @need_temp_offset_calib: need to perform temperature offset calibration
+ * @no_xtal_calib: some devices do not need crystal calibration data,
+ *	don't send it to those
+ * @temp_offset_v2: support v2 of temperature offset calibration
+ * @adv_pm: advanced power management
+ */
+struct iwl_dvm_cfg {
 	void (*set_hw_params)(struct iwl_priv *priv);
 	int (*set_channel_switch)(struct iwl_priv *priv,
 				  struct ieee80211_channel_switch *ch_switch);
-	/* device specific configuration */
 	void (*nic_config)(struct iwl_priv *priv);
-
-	/* temperature */
 	void (*temperature)(struct iwl_priv *priv);
+
+	const struct iwl_dvm_bt_params *bt_params;
+	s32 chain_noise_scale;
+	u8 plcp_delta_threshold;
+	bool adv_thermal_throttle;
+	bool support_ct_kill_exit;
+	bool hd_v2;
+	bool no_idle_support;
+	bool need_temp_offset_calib;
+	bool no_xtal_calib;
+	bool temp_offset_v2;
+	bool adv_pm;
 };
 
 struct iwl_wipan_noa_data {
@@ -610,7 +653,7 @@ struct iwl_priv {
 	struct device *dev;		/* for debug prints only */
 	const struct iwl_cfg *cfg;
 	const struct iwl_fw *fw;
-	const struct iwl_lib_ops *lib;
+	const struct iwl_dvm_cfg *lib;
 	unsigned long status;
 
 	spinlock_t sta_lock;
@@ -646,10 +689,6 @@ struct iwl_priv {
 	struct iwl_spectrum_notification measure_report;
 	u8 measurement_status;
 
-#define IWL_OWNERSHIP_DRIVER	0
-#define IWL_OWNERSHIP_TM	1
-	u8 ucode_owner;
-
 	/* ucode beacon time */
 	u32 ucode_beacon_time;
 	int missed_beacon_threshold;
@@ -844,7 +883,7 @@ struct iwl_priv {
 #endif /* CONFIG_IWLWIFI_DEBUGFS */
 
 	struct iwl_nvm_data *nvm_data;
-	/* eeprom blob for debugfs/testmode */
+	/* eeprom blob for debugfs */
 	u8 *eeprom_blob;
 	size_t eeprom_blob_size;
 
@@ -860,16 +899,14 @@ struct iwl_priv {
 	unsigned long blink_on, blink_off;
 	bool led_registered;
 
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-	struct iwl_test tst;
-	u32 tm_fixed_rate;
-#endif
-
 	/* WoWLAN GTK rekey data */
 	u8 kck[NL80211_KCK_LEN], kek[NL80211_KEK_LEN];
 	__le64 replay_ctr;
 	__le16 last_seq_ctl;
 	bool have_rekey_data;
+#ifdef CONFIG_PM_SLEEP
+	struct wiphy_wowlan_support wowlan_support;
+#endif
 
 	/* device_pointers: pointers to ucode event tables */
 	struct {
diff --git a/drivers/net/wireless/iwlwifi/dvm/devices.c b/drivers/net/wireless/iwlwifi/dvm/devices.c
index c48907c..352c6cb 100644
--- a/drivers/net/wireless/iwlwifi/dvm/devices.c
+++ b/drivers/net/wireless/iwlwifi/dvm/devices.c
@@ -174,10 +174,13 @@ static void iwl1000_hw_set_hw_params(struct iwl_priv *priv)
 	priv->hw_params.sens = &iwl1000_sensitivity;
 }
 
-struct iwl_lib_ops iwl1000_lib = {
+const struct iwl_dvm_cfg iwl_dvm_1000_cfg = {
 	.set_hw_params = iwl1000_hw_set_hw_params,
 	.nic_config = iwl1000_nic_config,
 	.temperature = iwlagn_temperature,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_EXT_LONG_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
 };
 
 
@@ -232,16 +235,56 @@ static void iwl2000_hw_set_hw_params(struct iwl_priv *priv)
 	priv->hw_params.sens = &iwl2000_sensitivity;
 }
 
-struct iwl_lib_ops iwl2000_lib = {
+const struct iwl_dvm_cfg iwl_dvm_2000_cfg = {
 	.set_hw_params = iwl2000_hw_set_hw_params,
 	.nic_config = iwl2000_nic_config,
 	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.hd_v2 = true,
+	.need_temp_offset_calib = true,
+	.temp_offset_v2 = true,
 };
 
-struct iwl_lib_ops iwl2030_lib = {
+const struct iwl_dvm_cfg iwl_dvm_105_cfg = {
 	.set_hw_params = iwl2000_hw_set_hw_params,
 	.nic_config = iwl2000_nic_config,
 	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.hd_v2 = true,
+	.need_temp_offset_calib = true,
+	.temp_offset_v2 = true,
+	.adv_pm = true,
+};
+
+static const struct iwl_dvm_bt_params iwl2030_bt_params = {
+	/* Due to bluetooth, we transmit 2.4 GHz probes only on antenna A */
+	.advanced_bt_coexist = true,
+	.agg_time_limit = BT_AGG_THRESHOLD_DEF,
+	.bt_init_traffic_load = IWL_BT_COEX_TRAFFIC_LOAD_NONE,
+	.bt_prio_boost = IWLAGN_BT_PRIO_BOOST_DEFAULT32,
+	.bt_sco_disable = true,
+	.bt_session_2 = true,
+};
+
+const struct iwl_dvm_cfg iwl_dvm_2030_cfg = {
+	.set_hw_params = iwl2000_hw_set_hw_params,
+	.nic_config = iwl2000_nic_config,
+	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.hd_v2 = true,
+	.bt_params = &iwl2030_bt_params,
+	.need_temp_offset_calib = true,
+	.temp_offset_v2 = true,
+	.adv_pm = true,
 };
 
 /*
@@ -420,16 +463,23 @@ static int iwl5000_hw_channel_switch(struct iwl_priv *priv,
 	return iwl_dvm_send_cmd(priv, &hcmd);
 }
 
-struct iwl_lib_ops iwl5000_lib = {
+const struct iwl_dvm_cfg iwl_dvm_5000_cfg = {
 	.set_hw_params = iwl5000_hw_set_hw_params,
 	.set_channel_switch = iwl5000_hw_channel_switch,
 	.temperature = iwlagn_temperature,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_LONG_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.no_idle_support = true,
 };
 
-struct iwl_lib_ops iwl5150_lib = {
+const struct iwl_dvm_cfg iwl_dvm_5150_cfg = {
 	.set_hw_params = iwl5150_hw_set_hw_params,
 	.set_channel_switch = iwl5000_hw_channel_switch,
 	.temperature = iwl5150_temperature,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_LONG_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.no_idle_support = true,
+	.no_xtal_calib = true,
 };
 
 
@@ -584,16 +634,59 @@ static int iwl6000_hw_channel_switch(struct iwl_priv *priv,
 	return err;
 }
 
-struct iwl_lib_ops iwl6000_lib = {
+const struct iwl_dvm_cfg iwl_dvm_6000_cfg = {
 	.set_hw_params = iwl6000_hw_set_hw_params,
 	.set_channel_switch = iwl6000_hw_channel_switch,
 	.nic_config = iwl6000_nic_config,
 	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+};
+
+const struct iwl_dvm_cfg iwl_dvm_6005_cfg = {
+	.set_hw_params = iwl6000_hw_set_hw_params,
+	.set_channel_switch = iwl6000_hw_channel_switch,
+	.nic_config = iwl6000_nic_config,
+	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.need_temp_offset_calib = true,
+};
+
+const struct iwl_dvm_cfg iwl_dvm_6050_cfg = {
+	.set_hw_params = iwl6000_hw_set_hw_params,
+	.set_channel_switch = iwl6000_hw_channel_switch,
+	.nic_config = iwl6000_nic_config,
+	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1500,
+};
+
+static const struct iwl_dvm_bt_params iwl6000_bt_params = {
+	/* Due to bluetooth, we transmit 2.4 GHz probes only on antenna A */
+	.advanced_bt_coexist = true,
+	.agg_time_limit = BT_AGG_THRESHOLD_DEF,
+	.bt_init_traffic_load = IWL_BT_COEX_TRAFFIC_LOAD_NONE,
+	.bt_prio_boost = IWLAGN_BT_PRIO_BOOST_DEFAULT,
+	.bt_sco_disable = true,
 };
 
-struct iwl_lib_ops iwl6030_lib = {
+const struct iwl_dvm_cfg iwl_dvm_6030_cfg = {
 	.set_hw_params = iwl6000_hw_set_hw_params,
 	.set_channel_switch = iwl6000_hw_channel_switch,
 	.nic_config = iwl6000_nic_config,
 	.temperature = iwlagn_temperature,
+	.adv_thermal_throttle = true,
+	.support_ct_kill_exit = true,
+	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
+	.chain_noise_scale = 1000,
+	.bt_params = &iwl6000_bt_params,
+	.need_temp_offset_calib = true,
+	.adv_pm = true,
 };
diff --git a/drivers/net/wireless/iwlwifi/dvm/lib.c b/drivers/net/wireless/iwlwifi/dvm/lib.c
index 54f5533..3d5bdc4 100644
--- a/drivers/net/wireless/iwlwifi/dvm/lib.c
+++ b/drivers/net/wireless/iwlwifi/dvm/lib.c
@@ -254,23 +254,23 @@ void iwlagn_send_advance_bt_config(struct iwl_priv *priv)
 	BUILD_BUG_ON(sizeof(iwlagn_def_3w_lookup) !=
 			sizeof(basic.bt3_lookup_table));
 
-	if (priv->cfg->bt_params) {
+	if (priv->lib->bt_params) {
 		/*
 		 * newer generation of devices (2000 series and newer)
 		 * use the version 2 of the bt command
 		 * we need to make sure sending the host command
 		 * with correct data structure to avoid uCode assert
 		 */
-		if (priv->cfg->bt_params->bt_session_2) {
+		if (priv->lib->bt_params->bt_session_2) {
 			bt_cmd_v2.prio_boost = cpu_to_le32(
-				priv->cfg->bt_params->bt_prio_boost);
+				priv->lib->bt_params->bt_prio_boost);
 			bt_cmd_v2.tx_prio_boost = 0;
 			bt_cmd_v2.rx_prio_boost = 0;
 		} else {
 			/* older version only has 8 bits */
-			WARN_ON(priv->cfg->bt_params->bt_prio_boost & ~0xFF);
+			WARN_ON(priv->lib->bt_params->bt_prio_boost & ~0xFF);
 			bt_cmd_v1.prio_boost =
-				priv->cfg->bt_params->bt_prio_boost;
+				priv->lib->bt_params->bt_prio_boost;
 			bt_cmd_v1.tx_prio_boost = 0;
 			bt_cmd_v1.rx_prio_boost = 0;
 		}
@@ -330,7 +330,7 @@ void iwlagn_send_advance_bt_config(struct iwl_priv *priv)
 		       priv->bt_full_concurrent ?
 		       "full concurrency" : "3-wire");
 
-	if (priv->cfg->bt_params->bt_session_2) {
+	if (priv->lib->bt_params->bt_session_2) {
 		memcpy(&bt_cmd_v2.basic, &basic,
 			sizeof(basic));
 		ret = iwl_dvm_send_cmd_pdu(priv, REPLY_BT_CONFIG,
@@ -758,8 +758,8 @@ static bool is_single_rx_stream(struct iwl_priv *priv)
  */
 static int iwl_get_active_rx_chain_count(struct iwl_priv *priv)
 {
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist &&
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist &&
 	    (priv->bt_full_concurrent ||
 	     priv->bt_traffic_load >= IWL_BT_COEX_TRAFFIC_LOAD_HIGH)) {
 		/*
@@ -830,8 +830,8 @@ void iwlagn_set_rxon_chain(struct iwl_priv *priv, struct iwl_rxon_context *ctx)
 	else
 		active_chains = priv->nvm_data->valid_rx_ant;
 
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist &&
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist &&
 	    (priv->bt_full_concurrent ||
 	     priv->bt_traffic_load >= IWL_BT_COEX_TRAFFIC_LOAD_HIGH)) {
 		/*
@@ -1288,12 +1288,6 @@ int iwl_dvm_send_cmd(struct iwl_priv *priv, struct iwl_host_cmd *cmd)
 	if (!(cmd->flags & CMD_ASYNC))
 		lockdep_assert_held(&priv->mutex);
 
-	if (priv->ucode_owner == IWL_OWNERSHIP_TM &&
-	    !(cmd->flags & CMD_ON_DEMAND)) {
-		IWL_DEBUG_HC(priv, "tm own the uCode, no regular hcmd send\n");
-		return -EIO;
-	}
-
 	return iwl_trans_send_cmd(priv->trans, cmd);
 }
 
diff --git a/drivers/net/wireless/iwlwifi/dvm/mac80211.c b/drivers/net/wireless/iwlwifi/dvm/mac80211.c
index cab23af..822f1a0 100644
--- a/drivers/net/wireless/iwlwifi/dvm/mac80211.c
+++ b/drivers/net/wireless/iwlwifi/dvm/mac80211.c
@@ -208,20 +208,21 @@ int iwlagn_mac_setup_register(struct iwl_priv *priv,
 	    priv->trans->ops->d3_suspend &&
 	    priv->trans->ops->d3_resume &&
 	    device_can_wakeup(priv->trans->dev)) {
-		hw->wiphy->wowlan.flags = WIPHY_WOWLAN_MAGIC_PKT |
-					  WIPHY_WOWLAN_DISCONNECT |
-					  WIPHY_WOWLAN_EAP_IDENTITY_REQ |
-					  WIPHY_WOWLAN_RFKILL_RELEASE;
+		priv->wowlan_support.flags = WIPHY_WOWLAN_MAGIC_PKT |
+					     WIPHY_WOWLAN_DISCONNECT |
+					     WIPHY_WOWLAN_EAP_IDENTITY_REQ |
+					     WIPHY_WOWLAN_RFKILL_RELEASE;
 		if (!iwlwifi_mod_params.sw_crypto)
-			hw->wiphy->wowlan.flags |=
+			priv->wowlan_support.flags |=
 				WIPHY_WOWLAN_SUPPORTS_GTK_REKEY |
 				WIPHY_WOWLAN_GTK_REKEY_FAILURE;
 
-		hw->wiphy->wowlan.n_patterns = IWLAGN_WOWLAN_MAX_PATTERNS;
-		hw->wiphy->wowlan.pattern_min_len =
+		priv->wowlan_support.n_patterns = IWLAGN_WOWLAN_MAX_PATTERNS;
+		priv->wowlan_support.pattern_min_len =
 					IWLAGN_WOWLAN_MIN_PATTERN_LEN;
-		hw->wiphy->wowlan.pattern_max_len =
+		priv->wowlan_support.pattern_max_len =
 					IWLAGN_WOWLAN_MAX_PATTERN_LEN;
+		hw->wiphy->wowlan = &priv->wowlan_support;
 	}
 #endif
 
@@ -426,7 +427,11 @@ static int iwlagn_mac_suspend(struct ieee80211_hw *hw,
 	if (ret)
 		goto error;
 
-	iwl_trans_d3_suspend(priv->trans);
+	/* let the ucode operate on its own */
+	iwl_write32(priv->trans, CSR_UCODE_DRV_GP1_SET,
+		    CSR_UCODE_DRV_GP1_BIT_D3_CFG_COMPLETE);
+
+	iwl_trans_d3_suspend(priv->trans, false);
 
 	goto out;
 
@@ -500,7 +505,7 @@ static int iwlagn_mac_resume(struct ieee80211_hw *hw)
 	/* we'll clear ctx->vif during iwlagn_prepare_restart() */
 	vif = ctx->vif;
 
-	ret = iwl_trans_d3_resume(priv->trans, &d3_status);
+	ret = iwl_trans_d3_resume(priv->trans, &d3_status, false);
 	if (ret)
 		goto out_unlock;
 
@@ -509,6 +514,10 @@ static int iwlagn_mac_resume(struct ieee80211_hw *hw)
 		goto out_unlock;
 	}
 
+	/* uCode is no longer operating by itself */
+	iwl_write32(priv->trans, CSR_UCODE_DRV_GP1_CLR,
+		    CSR_UCODE_DRV_GP1_BIT_D3_CFG_COMPLETE);
+
 	base = priv->device_pointers.error_event_table;
 	if (!iwlagn_hw_valid_rtc_data_addr(base)) {
 		IWL_WARN(priv, "Invalid error table during resume!\n");
@@ -1276,8 +1285,8 @@ static void iwlagn_mac_rssi_callback(struct ieee80211_hw *hw,
 	IWL_DEBUG_MAC80211(priv, "enter\n");
 	mutex_lock(&priv->mutex);
 
-	if (priv->cfg->bt_params &&
-			priv->cfg->bt_params->advanced_bt_coexist) {
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		if (rssi_event == RSSI_EVENT_LOW)
 			priv->bt_enable_pspoll = true;
 		else if (rssi_event == RSSI_EVENT_HIGH)
@@ -1387,7 +1396,7 @@ static int iwl_setup_interface(struct iwl_priv *priv,
 		return err;
 	}
 
-	if (priv->cfg->bt_params && priv->cfg->bt_params->advanced_bt_coexist &&
+	if (priv->lib->bt_params && priv->lib->bt_params->advanced_bt_coexist &&
 	    vif->type == NL80211_IFTYPE_ADHOC) {
 		/*
 		 * pretend to have high BT traffic as long as we
@@ -1757,8 +1766,6 @@ struct ieee80211_ops iwlagn_hw_ops = {
 	.remain_on_channel = iwlagn_mac_remain_on_channel,
 	.cancel_remain_on_channel = iwlagn_mac_cancel_remain_on_channel,
 	.rssi_callback = iwlagn_mac_rssi_callback,
-	CFG80211_TESTMODE_CMD(iwlagn_mac_testmode_cmd)
-	CFG80211_TESTMODE_DUMP(iwlagn_mac_testmode_dump)
 	.set_tim = iwlagn_mac_set_tim,
 };
 
diff --git a/drivers/net/wireless/iwlwifi/dvm/main.c b/drivers/net/wireless/iwlwifi/dvm/main.c
index 74d7572..3952ddf 100644
--- a/drivers/net/wireless/iwlwifi/dvm/main.c
+++ b/drivers/net/wireless/iwlwifi/dvm/main.c
@@ -615,7 +615,7 @@ static void iwl_rf_kill_ct_config(struct iwl_priv *priv)
 
 	priv->thermal_throttle.ct_kill_toggle = false;
 
-	if (priv->cfg->base_params->support_ct_kill_exit) {
+	if (priv->lib->support_ct_kill_exit) {
 		adv_cmd.critical_temperature_enter =
 			cpu_to_le32(priv->hw_params.ct_kill_threshold);
 		adv_cmd.critical_temperature_exit =
@@ -732,10 +732,10 @@ int iwl_alive_start(struct iwl_priv *priv)
 	}
 
 	/* download priority table before any calibration request */
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist) {
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		/* Configure Bluetooth device coexistence support */
-		if (priv->cfg->bt_params->bt_sco_disable)
+		if (priv->lib->bt_params->bt_sco_disable)
 			priv->bt_enable_pspoll = false;
 		else
 			priv->bt_enable_pspoll = true;
@@ -873,9 +873,9 @@ void iwl_down(struct iwl_priv *priv)
 	priv->bt_status = 0;
 	priv->cur_rssi_ctx = NULL;
 	priv->bt_is_sco = 0;
-	if (priv->cfg->bt_params)
+	if (priv->lib->bt_params)
 		priv->bt_traffic_load =
-			 priv->cfg->bt_params->bt_init_traffic_load;
+			 priv->lib->bt_params->bt_init_traffic_load;
 	else
 		priv->bt_traffic_load = 0;
 	priv->bt_full_concurrent = false;
@@ -1058,7 +1058,7 @@ static void iwl_setup_deferred_work(struct iwl_priv *priv)
 
 	iwl_setup_scan_deferred_work(priv);
 
-	if (priv->cfg->bt_params)
+	if (priv->lib->bt_params)
 		iwlagn_bt_setup_deferred_work(priv);
 
 	init_timer(&priv->statistics_periodic);
@@ -1072,7 +1072,7 @@ static void iwl_setup_deferred_work(struct iwl_priv *priv)
 
 void iwl_cancel_deferred_work(struct iwl_priv *priv)
 {
-	if (priv->cfg->bt_params)
+	if (priv->lib->bt_params)
 		iwlagn_bt_cancel_deferred_work(priv);
 
 	cancel_work_sync(&priv->run_time_calib_work);
@@ -1098,16 +1098,13 @@ static int iwl_init_drv(struct iwl_priv *priv)
 
 	priv->band = IEEE80211_BAND_2GHZ;
 
-	priv->plcp_delta_threshold =
-		priv->cfg->base_params->plcp_delta_threshold;
+	priv->plcp_delta_threshold = priv->lib->plcp_delta_threshold;
 
 	priv->iw_mode = NL80211_IFTYPE_STATION;
 	priv->current_ht_config.smps = IEEE80211_SMPS_STATIC;
 	priv->missed_beacon_threshold = IWL_MISSED_BEACON_THRESHOLD_DEF;
 	priv->agg_tids_count = 0;
 
-	priv->ucode_owner = IWL_OWNERSHIP_DRIVER;
-
 	priv->rx_statistics_jiffies = jiffies;
 
 	/* Choose which receivers/antennas to use */
@@ -1116,8 +1113,8 @@ static int iwl_init_drv(struct iwl_priv *priv)
 	iwl_init_scan_params(priv);
 
 	/* init bt coex */
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist) {
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		priv->kill_ack_mask = IWLAGN_BT_KILL_ACK_MASK_DEFAULT;
 		priv->kill_cts_mask = IWLAGN_BT_KILL_CTS_MASK_DEFAULT;
 		priv->bt_valid = IWLAGN_BT_ALL_VALID_MSK;
@@ -1173,12 +1170,6 @@ static void iwl_option_config(struct iwl_priv *priv)
 	IWL_INFO(priv, "CONFIG_IWLWIFI_DEVICE_TRACING disabled\n");
 #endif
 
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-	IWL_INFO(priv, "CONFIG_IWLWIFI_DEVICE_TESTMODE enabled\n");
-#else
-	IWL_INFO(priv, "CONFIG_IWLWIFI_DEVICE_TESTMODE disabled\n");
-#endif
-
 #ifdef CONFIG_IWLWIFI_P2P
 	IWL_INFO(priv, "CONFIG_IWLWIFI_P2P enabled\n");
 #else
@@ -1264,31 +1255,37 @@ static struct iwl_op_mode *iwl_op_mode_dvm_start(struct iwl_trans *trans,
 	switch (priv->cfg->device_family) {
 	case IWL_DEVICE_FAMILY_1000:
 	case IWL_DEVICE_FAMILY_100:
-		priv->lib = &iwl1000_lib;
+		priv->lib = &iwl_dvm_1000_cfg;
 		break;
 	case IWL_DEVICE_FAMILY_2000:
+		priv->lib = &iwl_dvm_2000_cfg;
+		break;
 	case IWL_DEVICE_FAMILY_105:
-		priv->lib = &iwl2000_lib;
+		priv->lib = &iwl_dvm_105_cfg;
 		break;
 	case IWL_DEVICE_FAMILY_2030:
 	case IWL_DEVICE_FAMILY_135:
-		priv->lib = &iwl2030_lib;
+		priv->lib = &iwl_dvm_2030_cfg;
 		break;
 	case IWL_DEVICE_FAMILY_5000:
-		priv->lib = &iwl5000_lib;
+		priv->lib = &iwl_dvm_5000_cfg;
 		break;
 	case IWL_DEVICE_FAMILY_5150:
-		priv->lib = &iwl5150_lib;
+		priv->lib = &iwl_dvm_5150_cfg;
 		break;
 	case IWL_DEVICE_FAMILY_6000:
-	case IWL_DEVICE_FAMILY_6005:
 	case IWL_DEVICE_FAMILY_6000i:
+		priv->lib = &iwl_dvm_6000_cfg;
+		break;
+	case IWL_DEVICE_FAMILY_6005:
+		priv->lib = &iwl_dvm_6005_cfg;
+		break;
 	case IWL_DEVICE_FAMILY_6050:
 	case IWL_DEVICE_FAMILY_6150:
-		priv->lib = &iwl6000_lib;
+		priv->lib = &iwl_dvm_6050_cfg;
 		break;
 	case IWL_DEVICE_FAMILY_6030:
-		priv->lib = &iwl6030_lib;
+		priv->lib = &iwl_dvm_6030_cfg;
 		break;
 	default:
 		break;
@@ -1350,8 +1347,8 @@ static struct iwl_op_mode *iwl_op_mode_dvm_start(struct iwl_trans *trans,
 			IWL_BT_ANTENNA_COUPLING_THRESHOLD) ?
 			true : false;
 
-	/* enable/disable bt channel inhibition */
-	priv->bt_ch_announce = iwlwifi_mod_params.bt_ch_announce;
+	/* bt channel inhibition enabled*/
+	priv->bt_ch_announce = true;
 	IWL_DEBUG_INFO(priv, "BT channel inhibition is %s\n",
 		       (priv->bt_ch_announce) ? "On" : "Off");
 
@@ -1446,7 +1443,6 @@ static struct iwl_op_mode *iwl_op_mode_dvm_start(struct iwl_trans *trans,
 	 ********************/
 	iwl_setup_deferred_work(priv);
 	iwl_setup_rx_handlers(priv);
-	iwl_testmode_init(priv);
 
 	iwl_power_initialize(priv);
 	iwl_tt_initialize(priv);
@@ -1483,7 +1479,6 @@ out_mac80211_unregister:
 	iwlagn_mac_unregister(priv);
 out_destroy_workqueue:
 	iwl_tt_exit(priv);
-	iwl_testmode_free(priv);
 	iwl_cancel_deferred_work(priv);
 	destroy_workqueue(priv->workqueue);
 	priv->workqueue = NULL;
@@ -1505,7 +1500,6 @@ static void iwl_op_mode_dvm_stop(struct iwl_op_mode *op_mode)
 
 	IWL_DEBUG_INFO(priv, "*** UNLOAD DRIVER ***\n");
 
-	iwl_testmode_free(priv);
 	iwlagn_mac_unregister(priv);
 
 	iwl_tt_exit(priv);
@@ -1854,14 +1848,9 @@ int iwl_dump_nic_event_log(struct iwl_priv *priv, bool full_log,
 		return pos;
 	}
 
-#ifdef CONFIG_IWLWIFI_DEBUG
 	if (!(iwl_have_debug_level(IWL_DL_FW_ERRORS)) && !full_log)
 		size = (size > DEFAULT_DUMP_EVENT_LOG_ENTRIES)
 			? DEFAULT_DUMP_EVENT_LOG_ENTRIES : size;
-#else
-	size = (size > DEFAULT_DUMP_EVENT_LOG_ENTRIES)
-		? DEFAULT_DUMP_EVENT_LOG_ENTRIES : size;
-#endif
 	IWL_ERR(priv, "Start IWL Event Log Dump: display last %u entries\n",
 		size);
 
@@ -1905,10 +1894,8 @@ static void iwlagn_fw_error(struct iwl_priv *priv, bool ondemand)
 	unsigned int reload_msec;
 	unsigned long reload_jiffies;
 
-#ifdef CONFIG_IWLWIFI_DEBUG
 	if (iwl_have_debug_level(IWL_DL_FW_ERRORS))
 		iwl_print_rx_config_cmd(priv, IWL_RXON_CTX_BSS);
-#endif
 
 	/* uCode is no longer loaded. */
 	priv->ucode_loaded = false;
diff --git a/drivers/net/wireless/iwlwifi/dvm/power.c b/drivers/net/wireless/iwlwifi/dvm/power.c
index bd69018..77cb597 100644
--- a/drivers/net/wireless/iwlwifi/dvm/power.c
+++ b/drivers/net/wireless/iwlwifi/dvm/power.c
@@ -163,7 +163,7 @@ static void iwl_static_sleep_cmd(struct iwl_priv *priv,
 	u8 skip;
 	u32 slp_itrvl;
 
-	if (priv->cfg->adv_pm) {
+	if (priv->lib->adv_pm) {
 		table = apm_range_2;
 		if (period <= IWL_DTIM_RANGE_1_MAX)
 			table = apm_range_1;
@@ -217,7 +217,7 @@ static void iwl_static_sleep_cmd(struct iwl_priv *priv,
 		cmd->flags &= ~IWL_POWER_SHADOW_REG_ENA;
 
 	if (iwl_advanced_bt_coexist(priv)) {
-		if (!priv->cfg->bt_params->bt_sco_disable)
+		if (!priv->lib->bt_params->bt_sco_disable)
 			cmd->flags |= IWL_POWER_BT_SCO_ENA;
 		else
 			cmd->flags &= ~IWL_POWER_BT_SCO_ENA;
@@ -293,7 +293,7 @@ static void iwl_power_build_cmd(struct iwl_priv *priv,
 
 	if (priv->wowlan)
 		iwl_static_sleep_cmd(priv, cmd, IWL_POWER_INDEX_5, dtimper);
-	else if (!priv->cfg->base_params->no_idle_support &&
+	else if (!priv->lib->no_idle_support &&
 		 priv->hw->conf.flags & IEEE80211_CONF_IDLE)
 		iwl_static_sleep_cmd(priv, cmd, IWL_POWER_INDEX_5, 20);
 	else if (iwl_tt_is_low_power_state(priv)) {
diff --git a/drivers/net/wireless/iwlwifi/dvm/rs.c b/drivers/net/wireless/iwlwifi/dvm/rs.c
index 10fbb17..1b69394 100644
--- a/drivers/net/wireless/iwlwifi/dvm/rs.c
+++ b/drivers/net/wireless/iwlwifi/dvm/rs.c
@@ -351,12 +351,6 @@ static void rs_program_fix_rate(struct iwl_priv *priv,
 	lq_sta->active_mimo2_rate  = 0x1FD0;	/* 6 - 60 MBits, no 9, no CCK */
 	lq_sta->active_mimo3_rate  = 0x1FD0;	/* 6 - 60 MBits, no 9, no CCK */
 
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-	/* testmode has higher priority to overwirte the fixed rate */
-	if (priv->tm_fixed_rate)
-		lq_sta->dbg_fixed_rate = priv->tm_fixed_rate;
-#endif
-
 	IWL_DEBUG_RATE(priv, "sta_id %d rate 0x%X\n",
 		lq_sta->lq.sta_id, lq_sta->dbg_fixed_rate);
 
@@ -419,23 +413,18 @@ static int rs_tl_turn_on_agg_for_tid(struct iwl_priv *priv,
 
 	load = rs_tl_get_load(lq_data, tid);
 
-	if ((iwlwifi_mod_params.auto_agg) || (load > IWL_AGG_LOAD_THRESHOLD)) {
-		IWL_DEBUG_HT(priv, "Starting Tx agg: STA: %pM tid: %d\n",
-				sta->addr, tid);
-		ret = ieee80211_start_tx_ba_session(sta, tid, 5000);
-		if (ret == -EAGAIN) {
-			/*
-			 * driver and mac80211 is out of sync
-			 * this might be cause by reloading firmware
-			 * stop the tx ba session here
-			 */
-			IWL_ERR(priv, "Fail start Tx agg on tid: %d\n",
-				tid);
-			ieee80211_stop_tx_ba_session(sta, tid);
-		}
-	} else {
-		IWL_DEBUG_HT(priv, "Aggregation not enabled for tid %d "
-			"because load = %u\n", tid, load);
+	IWL_DEBUG_HT(priv, "Starting Tx agg: STA: %pM tid: %d\n",
+			sta->addr, tid);
+	ret = ieee80211_start_tx_ba_session(sta, tid, 5000);
+	if (ret == -EAGAIN) {
+		/*
+		 * driver and mac80211 is out of sync
+		 * this might be cause by reloading firmware
+		 * stop the tx ba session here
+		 */
+		IWL_ERR(priv, "Fail start Tx agg on tid: %d\n",
+			tid);
+		ieee80211_stop_tx_ba_session(sta, tid);
 	}
 	return ret;
 }
@@ -1083,12 +1072,7 @@ done:
 	if (sta && sta->supp_rates[sband->band])
 		rs_rate_scale_perform(priv, skb, sta, lq_sta);
 
-#if defined(CONFIG_MAC80211_DEBUGFS) && defined(CONFIG_IWLWIFI_DEVICE_TESTMODE)
-	if ((priv->tm_fixed_rate) &&
-	    (priv->tm_fixed_rate != lq_sta->dbg_fixed_rate))
-		rs_program_fix_rate(priv, lq_sta);
-#endif
-	if (priv->cfg->bt_params && priv->cfg->bt_params->advanced_bt_coexist)
+	if (priv->lib->bt_params && priv->lib->bt_params->advanced_bt_coexist)
 		rs_bt_update_lq(priv, ctx, lq_sta);
 }
 
@@ -2913,9 +2897,6 @@ void iwl_rs_rate_init(struct iwl_priv *priv, struct ieee80211_sta *sta, u8 sta_i
 	if (sband->band == IEEE80211_BAND_5GHZ)
 		lq_sta->last_txrate_idx += IWL_FIRST_OFDM_RATE;
 	lq_sta->is_agg = 0;
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-	priv->tm_fixed_rate = 0;
-#endif
 #ifdef CONFIG_MAC80211_DEBUGFS
 	lq_sta->dbg_fixed_rate = 0;
 #endif
@@ -3064,11 +3045,11 @@ static void rs_fill_link_cmd(struct iwl_priv *priv,
 	 * overwrite if needed, pass aggregation time limit
 	 * to uCode in uSec
 	 */
-	if (priv && priv->cfg->bt_params &&
-	    priv->cfg->bt_params->agg_time_limit &&
+	if (priv && priv->lib->bt_params &&
+	    priv->lib->bt_params->agg_time_limit &&
 	    priv->bt_traffic_load >= IWL_BT_COEX_TRAFFIC_LOAD_HIGH)
 		lq_cmd->agg_params.agg_time_limit =
-			cpu_to_le16(priv->cfg->bt_params->agg_time_limit);
+			cpu_to_le16(priv->lib->bt_params->agg_time_limit);
 }
 
 static void *rs_alloc(struct ieee80211_hw *hw, struct dentry *debugfsdir)
diff --git a/drivers/net/wireless/iwlwifi/dvm/rx.c b/drivers/net/wireless/iwlwifi/dvm/rx.c
index a4eed20..d71776d 100644
--- a/drivers/net/wireless/iwlwifi/dvm/rx.c
+++ b/drivers/net/wireless/iwlwifi/dvm/rx.c
@@ -335,8 +335,7 @@ static void iwlagn_recover_from_statistics(struct iwl_priv *priv,
 	if (msecs < 99)
 		return;
 
-	if (iwlwifi_mod_params.plcp_check &&
-	    !iwlagn_good_plcp_health(priv, cur_ofdm, cur_ofdm_ht, msecs))
+	if (!iwlagn_good_plcp_health(priv, cur_ofdm, cur_ofdm_ht, msecs))
 		iwl_force_rf_reset(priv, false);
 }
 
@@ -1102,7 +1101,7 @@ void iwl_setup_rx_handlers(struct iwl_priv *priv)
 	iwl_notification_wait_init(&priv->notif_wait);
 
 	/* Set up BT Rx handlers */
-	if (priv->cfg->bt_params)
+	if (priv->lib->bt_params)
 		iwlagn_bt_rx_handler_setup(priv);
 }
 
@@ -1120,32 +1119,17 @@ int iwl_rx_dispatch(struct iwl_op_mode *op_mode, struct iwl_rx_cmd_buffer *rxb,
 	 */
 	iwl_notification_wait_notify(&priv->notif_wait, pkt);
 
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-	/*
-	 * RX data may be forwarded to userspace in one
-	 * of two cases: the user owns the fw through testmode or when
-	 * the user requested to monitor the rx w/o affecting the regular flow.
-	 * In these cases the iwl_test object will handle forwarding the rx
-	 * data to user space.
-	 * Note that if the ownership flag != IWL_OWNERSHIP_TM the flow
-	 * continues.
-	 */
-	iwl_test_rx(&priv->tst, rxb);
-#endif
-
-	if (priv->ucode_owner != IWL_OWNERSHIP_TM) {
-		/* Based on type of command response or notification,
-		 *   handle those that need handling via function in
-		 *   rx_handlers table.  See iwl_setup_rx_handlers() */
-		if (priv->rx_handlers[pkt->hdr.cmd]) {
-			priv->rx_handlers_stats[pkt->hdr.cmd]++;
-			err = priv->rx_handlers[pkt->hdr.cmd] (priv, rxb, cmd);
-		} else {
-			/* No handling needed */
-			IWL_DEBUG_RX(priv, "No handler needed for %s, 0x%02x\n",
-				     iwl_dvm_get_cmd_string(pkt->hdr.cmd),
-				     pkt->hdr.cmd);
-		}
+	/* Based on type of command response or notification,
+	 *   handle those that need handling via function in
+	 *   rx_handlers table.  See iwl_setup_rx_handlers() */
+	if (priv->rx_handlers[pkt->hdr.cmd]) {
+		priv->rx_handlers_stats[pkt->hdr.cmd]++;
+		err = priv->rx_handlers[pkt->hdr.cmd] (priv, rxb, cmd);
+	} else {
+		/* No handling needed */
+		IWL_DEBUG_RX(priv, "No handler needed for %s, 0x%02x\n",
+			     iwl_dvm_get_cmd_string(pkt->hdr.cmd),
+			     pkt->hdr.cmd);
 	}
 	return err;
 }
diff --git a/drivers/net/wireless/iwlwifi/dvm/scan.c b/drivers/net/wireless/iwlwifi/dvm/scan.c
index d69b558..8c686a5 100644
--- a/drivers/net/wireless/iwlwifi/dvm/scan.c
+++ b/drivers/net/wireless/iwlwifi/dvm/scan.c
@@ -801,8 +801,8 @@ static int iwlagn_request_scan(struct iwl_priv *priv, struct ieee80211_vif *vif)
 		 * Internal scans are passive, so we can indiscriminately set
 		 * the BT ignore flag on 2.4 GHz since it applies to TX only.
 		 */
-		if (priv->cfg->bt_params &&
-		    priv->cfg->bt_params->advanced_bt_coexist)
+		if (priv->lib->bt_params &&
+		    priv->lib->bt_params->advanced_bt_coexist)
 			scan->tx_cmd.tx_flags |= TX_CMD_FLG_IGNORE_BT;
 		break;
 	case IEEE80211_BAND_5GHZ:
@@ -844,8 +844,8 @@ static int iwlagn_request_scan(struct iwl_priv *priv, struct ieee80211_vif *vif)
 	band = priv->scan_band;
 
 	if (band == IEEE80211_BAND_2GHZ &&
-	    priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist) {
+	    priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		/* transmit 2.4 GHz probes only on first antenna */
 		scan_tx_antennas = first_antenna(scan_tx_antennas);
 	}
@@ -873,8 +873,8 @@ static int iwlagn_request_scan(struct iwl_priv *priv, struct ieee80211_vif *vif)
 
 		rx_ant = first_antenna(active_chains);
 	}
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist &&
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist &&
 	    priv->bt_full_concurrent) {
 		/* operated as 1x1 in full concurrency mode */
 		rx_ant = first_antenna(rx_ant);
diff --git a/drivers/net/wireless/iwlwifi/dvm/testmode.c b/drivers/net/wireless/iwlwifi/dvm/testmode.c
deleted file mode 100644
index b89b9d9..0000000
--- a/drivers/net/wireless/iwlwifi/dvm/testmode.c
+++ /dev/null
@@ -1,471 +0,0 @@
-/******************************************************************************
- *
- * This file is provided under a dual BSD/GPLv2 license.  When using or
- * redistributing this file, you may do so under either license.
- *
- * GPL LICENSE SUMMARY
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110,
- * USA
- *
- * The full GNU General Public License is included in this distribution
- * in the file called COPYING.
- *
- * Contact Information:
- *  Intel Linux Wireless <ilw@linux.intel.com>
- * Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
- *
- * BSD LICENSE
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- *  * Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- *  * Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *  * Neither the name Intel Corporation nor the names of its
- *    contributors may be used to endorse or promote products derived
- *    from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- *****************************************************************************/
-
-#include <linux/init.h>
-#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/dma-mapping.h>
-#include <net/net_namespace.h>
-#include <linux/netdevice.h>
-#include <net/cfg80211.h>
-#include <net/mac80211.h>
-#include <net/netlink.h>
-
-#include "iwl-debug.h"
-#include "iwl-trans.h"
-#include "dev.h"
-#include "agn.h"
-#include "iwl-test.h"
-#include "iwl-testmode.h"
-
-static int iwl_testmode_send_cmd(struct iwl_op_mode *op_mode,
-				 struct iwl_host_cmd *cmd)
-{
-	struct iwl_priv *priv = IWL_OP_MODE_GET_DVM(op_mode);
-	return iwl_dvm_send_cmd(priv, cmd);
-}
-
-static bool iwl_testmode_valid_hw_addr(u32 addr)
-{
-	if (iwlagn_hw_valid_rtc_data_addr(addr))
-		return true;
-
-	if (IWLAGN_RTC_INST_LOWER_BOUND <= addr &&
-	    addr < IWLAGN_RTC_INST_UPPER_BOUND)
-		return true;
-
-	return false;
-}
-
-static u32 iwl_testmode_get_fw_ver(struct iwl_op_mode *op_mode)
-{
-	struct iwl_priv *priv = IWL_OP_MODE_GET_DVM(op_mode);
-	return priv->fw->ucode_ver;
-}
-
-static struct sk_buff*
-iwl_testmode_alloc_reply(struct iwl_op_mode *op_mode, int len)
-{
-	struct iwl_priv *priv = IWL_OP_MODE_GET_DVM(op_mode);
-	return cfg80211_testmode_alloc_reply_skb(priv->hw->wiphy, len);
-}
-
-static int iwl_testmode_reply(struct iwl_op_mode *op_mode, struct sk_buff *skb)
-{
-	return cfg80211_testmode_reply(skb);
-}
-
-static struct sk_buff *iwl_testmode_alloc_event(struct iwl_op_mode *op_mode,
-						int len)
-{
-	struct iwl_priv *priv = IWL_OP_MODE_GET_DVM(op_mode);
-	return cfg80211_testmode_alloc_event_skb(priv->hw->wiphy, len,
-						 GFP_ATOMIC);
-}
-
-static void iwl_testmode_event(struct iwl_op_mode *op_mode, struct sk_buff *skb)
-{
-	return cfg80211_testmode_event(skb, GFP_ATOMIC);
-}
-
-static struct iwl_test_ops tst_ops = {
-	.send_cmd = iwl_testmode_send_cmd,
-	.valid_hw_addr = iwl_testmode_valid_hw_addr,
-	.get_fw_ver = iwl_testmode_get_fw_ver,
-	.alloc_reply = iwl_testmode_alloc_reply,
-	.reply = iwl_testmode_reply,
-	.alloc_event = iwl_testmode_alloc_event,
-	.event = iwl_testmode_event,
-};
-
-void iwl_testmode_init(struct iwl_priv *priv)
-{
-	iwl_test_init(&priv->tst, priv->trans, &tst_ops);
-}
-
-void iwl_testmode_free(struct iwl_priv *priv)
-{
-	iwl_test_free(&priv->tst);
-}
-
-static int iwl_testmode_cfg_init_calib(struct iwl_priv *priv)
-{
-	struct iwl_notification_wait calib_wait;
-	static const u8 calib_complete[] = {
-		CALIBRATION_COMPLETE_NOTIFICATION
-	};
-	int ret;
-
-	iwl_init_notification_wait(&priv->notif_wait, &calib_wait,
-				   calib_complete, ARRAY_SIZE(calib_complete),
-				   NULL, NULL);
-	ret = iwl_init_alive_start(priv);
-	if (ret) {
-		IWL_ERR(priv, "Fail init calibration: %d\n", ret);
-		goto cfg_init_calib_error;
-	}
-
-	ret = iwl_wait_notification(&priv->notif_wait, &calib_wait, 2 * HZ);
-	if (ret)
-		IWL_ERR(priv, "Error detecting"
-			" CALIBRATION_COMPLETE_NOTIFICATION: %d\n", ret);
-	return ret;
-
-cfg_init_calib_error:
-	iwl_remove_notification(&priv->notif_wait, &calib_wait);
-	return ret;
-}
-
-/*
- * This function handles the user application commands for driver.
- *
- * It retrieves command ID carried with IWL_TM_ATTR_COMMAND and calls to the
- * handlers respectively.
- *
- * If it's an unknown commdn ID, -ENOSYS is replied; otherwise, the returned
- * value of the actual command execution is replied to the user application.
- *
- * If there's any message responding to the user space, IWL_TM_ATTR_SYNC_RSP
- * is used for carry the message while IWL_TM_ATTR_COMMAND must set to
- * IWL_TM_CMD_DEV2APP_SYNC_RSP.
- *
- * @hw: ieee80211_hw object that represents the device
- * @tb: gnl message fields from the user space
- */
-static int iwl_testmode_driver(struct ieee80211_hw *hw, struct nlattr **tb)
-{
-	struct iwl_priv *priv = IWL_MAC80211_GET_DVM(hw);
-	struct iwl_trans *trans = priv->trans;
-	struct sk_buff *skb;
-	unsigned char *rsp_data_ptr = NULL;
-	int status = 0, rsp_data_len = 0;
-	u32 inst_size = 0, data_size = 0;
-	const struct fw_img *img;
-
-	switch (nla_get_u32(tb[IWL_TM_ATTR_COMMAND])) {
-	case IWL_TM_CMD_APP2DEV_GET_DEVICENAME:
-		rsp_data_ptr = (unsigned char *)priv->cfg->name;
-		rsp_data_len = strlen(priv->cfg->name);
-		skb = cfg80211_testmode_alloc_reply_skb(hw->wiphy,
-							rsp_data_len + 20);
-		if (!skb) {
-			IWL_ERR(priv, "Memory allocation fail\n");
-			return -ENOMEM;
-		}
-		if (nla_put_u32(skb, IWL_TM_ATTR_COMMAND,
-				IWL_TM_CMD_DEV2APP_SYNC_RSP) ||
-		    nla_put(skb, IWL_TM_ATTR_SYNC_RSP,
-			    rsp_data_len, rsp_data_ptr))
-			goto nla_put_failure;
-		status = cfg80211_testmode_reply(skb);
-		if (status < 0)
-			IWL_ERR(priv, "Error sending msg : %d\n", status);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_LOAD_INIT_FW:
-		status = iwl_load_ucode_wait_alive(priv, IWL_UCODE_INIT);
-		if (status)
-			IWL_ERR(priv, "Error loading init ucode: %d\n", status);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_CFG_INIT_CALIB:
-		iwl_testmode_cfg_init_calib(priv);
-		priv->ucode_loaded = false;
-		iwl_trans_stop_device(trans);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_LOAD_RUNTIME_FW:
-		status = iwl_load_ucode_wait_alive(priv, IWL_UCODE_REGULAR);
-		if (status) {
-			IWL_ERR(priv,
-				"Error loading runtime ucode: %d\n", status);
-			break;
-		}
-		status = iwl_alive_start(priv);
-		if (status)
-			IWL_ERR(priv,
-				"Error starting the device: %d\n", status);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_LOAD_WOWLAN_FW:
-		iwl_scan_cancel_timeout(priv, 200);
-		priv->ucode_loaded = false;
-		iwl_trans_stop_device(trans);
-		status = iwl_load_ucode_wait_alive(priv, IWL_UCODE_WOWLAN);
-		if (status) {
-			IWL_ERR(priv,
-				"Error loading WOWLAN ucode: %d\n", status);
-			break;
-		}
-		status = iwl_alive_start(priv);
-		if (status)
-			IWL_ERR(priv,
-				"Error starting the device: %d\n", status);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_GET_EEPROM:
-		if (priv->eeprom_blob) {
-			skb = cfg80211_testmode_alloc_reply_skb(hw->wiphy,
-				priv->eeprom_blob_size + 20);
-			if (!skb) {
-				IWL_ERR(priv, "Memory allocation fail\n");
-				return -ENOMEM;
-			}
-			if (nla_put_u32(skb, IWL_TM_ATTR_COMMAND,
-					IWL_TM_CMD_DEV2APP_EEPROM_RSP) ||
-			    nla_put(skb, IWL_TM_ATTR_EEPROM,
-				    priv->eeprom_blob_size,
-				    priv->eeprom_blob))
-				goto nla_put_failure;
-			status = cfg80211_testmode_reply(skb);
-			if (status < 0)
-				IWL_ERR(priv, "Error sending msg : %d\n",
-					status);
-		} else
-			return -ENODATA;
-		break;
-
-	case IWL_TM_CMD_APP2DEV_FIXRATE_REQ:
-		if (!tb[IWL_TM_ATTR_FIXRATE]) {
-			IWL_ERR(priv, "Missing fixrate setting\n");
-			return -ENOMSG;
-		}
-		priv->tm_fixed_rate = nla_get_u32(tb[IWL_TM_ATTR_FIXRATE]);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_GET_FW_INFO:
-		skb = cfg80211_testmode_alloc_reply_skb(hw->wiphy, 20 + 8);
-		if (!skb) {
-			IWL_ERR(priv, "Memory allocation fail\n");
-			return -ENOMEM;
-		}
-		if (!priv->ucode_loaded) {
-			IWL_ERR(priv, "No uCode has not been loaded\n");
-			return -EINVAL;
-		} else {
-			img = &priv->fw->img[priv->cur_ucode];
-			inst_size = img->sec[IWL_UCODE_SECTION_INST].len;
-			data_size = img->sec[IWL_UCODE_SECTION_DATA].len;
-		}
-		if (nla_put_u32(skb, IWL_TM_ATTR_FW_TYPE, priv->cur_ucode) ||
-		    nla_put_u32(skb, IWL_TM_ATTR_FW_INST_SIZE, inst_size) ||
-		    nla_put_u32(skb, IWL_TM_ATTR_FW_DATA_SIZE, data_size))
-			goto nla_put_failure;
-		status = cfg80211_testmode_reply(skb);
-		if (status < 0)
-			IWL_ERR(priv, "Error sending msg : %d\n", status);
-		break;
-
-	default:
-		IWL_ERR(priv, "Unknown testmode driver command ID\n");
-		return -ENOSYS;
-	}
-	return status;
-
-nla_put_failure:
-	kfree_skb(skb);
-	return -EMSGSIZE;
-}
-
-/*
- * This function handles the user application switch ucode ownership.
- *
- * It retrieves the mandatory fields IWL_TM_ATTR_UCODE_OWNER and
- * decide who the current owner of the uCode
- *
- * If the current owner is OWNERSHIP_TM, then the only host command
- * can deliver to uCode is from testmode, all the other host commands
- * will dropped.
- *
- * default driver is the owner of uCode in normal operational mode
- *
- * @hw: ieee80211_hw object that represents the device
- * @tb: gnl message fields from the user space
- */
-static int iwl_testmode_ownership(struct ieee80211_hw *hw, struct nlattr **tb)
-{
-	struct iwl_priv *priv = IWL_MAC80211_GET_DVM(hw);
-	u8 owner;
-
-	if (!tb[IWL_TM_ATTR_UCODE_OWNER]) {
-		IWL_ERR(priv, "Missing ucode owner\n");
-		return -ENOMSG;
-	}
-
-	owner = nla_get_u8(tb[IWL_TM_ATTR_UCODE_OWNER]);
-	if (owner == IWL_OWNERSHIP_DRIVER) {
-		priv->ucode_owner = owner;
-		iwl_test_enable_notifications(&priv->tst, false);
-	} else if (owner == IWL_OWNERSHIP_TM) {
-		priv->ucode_owner = owner;
-		iwl_test_enable_notifications(&priv->tst, true);
-	} else {
-		IWL_ERR(priv, "Invalid owner\n");
-		return -EINVAL;
-	}
-	return 0;
-}
-
-/* The testmode gnl message handler that takes the gnl message from the
- * user space and parses it per the policy iwl_testmode_gnl_msg_policy, then
- * invoke the corresponding handlers.
- *
- * This function is invoked when there is user space application sending
- * gnl message through the testmode tunnel NL80211_CMD_TESTMODE regulated
- * by nl80211.
- *
- * It retrieves the mandatory field, IWL_TM_ATTR_COMMAND, before
- * dispatching it to the corresponding handler.
- *
- * If IWL_TM_ATTR_COMMAND is missing, -ENOMSG is replied to user application;
- * -ENOSYS is replied to the user application if the command is unknown;
- * Otherwise, the command is dispatched to the respective handler.
- *
- * @hw: ieee80211_hw object that represents the device
- * @data: pointer to user space message
- * @len: length in byte of @data
- */
-int iwlagn_mac_testmode_cmd(struct ieee80211_hw *hw, void *data, int len)
-{
-	struct nlattr *tb[IWL_TM_ATTR_MAX];
-	struct iwl_priv *priv = IWL_MAC80211_GET_DVM(hw);
-	int result;
-
-	result = iwl_test_parse(&priv->tst, tb, data, len);
-	if (result)
-		return result;
-
-	/* in case multiple accesses to the device happens */
-	mutex_lock(&priv->mutex);
-	switch (nla_get_u32(tb[IWL_TM_ATTR_COMMAND])) {
-	case IWL_TM_CMD_APP2DEV_UCODE:
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_READ32:
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE32:
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE8:
-	case IWL_TM_CMD_APP2DEV_BEGIN_TRACE:
-	case IWL_TM_CMD_APP2DEV_END_TRACE:
-	case IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_READ:
-	case IWL_TM_CMD_APP2DEV_NOTIFICATIONS:
-	case IWL_TM_CMD_APP2DEV_GET_FW_VERSION:
-	case IWL_TM_CMD_APP2DEV_GET_DEVICE_ID:
-	case IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_WRITE:
-		result = iwl_test_handle_cmd(&priv->tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_GET_DEVICENAME:
-	case IWL_TM_CMD_APP2DEV_LOAD_INIT_FW:
-	case IWL_TM_CMD_APP2DEV_CFG_INIT_CALIB:
-	case IWL_TM_CMD_APP2DEV_LOAD_RUNTIME_FW:
-	case IWL_TM_CMD_APP2DEV_GET_EEPROM:
-	case IWL_TM_CMD_APP2DEV_FIXRATE_REQ:
-	case IWL_TM_CMD_APP2DEV_LOAD_WOWLAN_FW:
-	case IWL_TM_CMD_APP2DEV_GET_FW_INFO:
-		IWL_DEBUG_INFO(priv, "testmode cmd to driver\n");
-		result = iwl_testmode_driver(hw, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_OWNERSHIP:
-		IWL_DEBUG_INFO(priv, "testmode change uCode ownership\n");
-		result = iwl_testmode_ownership(hw, tb);
-		break;
-
-	default:
-		IWL_ERR(priv, "Unknown testmode command\n");
-		result = -ENOSYS;
-		break;
-	}
-	mutex_unlock(&priv->mutex);
-
-	if (result)
-		IWL_ERR(priv, "Test cmd failed result=%d\n", result);
-	return result;
-}
-
-int iwlagn_mac_testmode_dump(struct ieee80211_hw *hw, struct sk_buff *skb,
-		      struct netlink_callback *cb,
-		      void *data, int len)
-{
-	struct iwl_priv *priv = IWL_MAC80211_GET_DVM(hw);
-	int result;
-	u32 cmd;
-
-	if (cb->args[3]) {
-		/* offset by 1 since commands start at 0 */
-		cmd = cb->args[3] - 1;
-	} else {
-		struct nlattr *tb[IWL_TM_ATTR_MAX];
-
-		result = iwl_test_parse(&priv->tst, tb, data, len);
-		if (result)
-			return result;
-
-		cmd = nla_get_u32(tb[IWL_TM_ATTR_COMMAND]);
-		cb->args[3] = cmd + 1;
-	}
-
-	/* in case multiple accesses to the device happens */
-	mutex_lock(&priv->mutex);
-	result = iwl_test_dump(&priv->tst, cmd, skb, cb);
-	mutex_unlock(&priv->mutex);
-	return result;
-}
diff --git a/drivers/net/wireless/iwlwifi/dvm/tt.c b/drivers/net/wireless/iwlwifi/dvm/tt.c
index 03f9bc0..fbeee08 100644
--- a/drivers/net/wireless/iwlwifi/dvm/tt.c
+++ b/drivers/net/wireless/iwlwifi/dvm/tt.c
@@ -627,7 +627,7 @@ void iwl_tt_initialize(struct iwl_priv *priv)
 	INIT_WORK(&priv->ct_enter, iwl_bg_ct_enter);
 	INIT_WORK(&priv->ct_exit, iwl_bg_ct_exit);
 
-	if (priv->cfg->base_params->adv_thermal_throttle) {
+	if (priv->lib->adv_thermal_throttle) {
 		IWL_DEBUG_TEMP(priv, "Advanced Thermal Throttling\n");
 		tt->restriction = kcalloc(IWL_TI_STATE_MAX,
 					  sizeof(struct iwl_tt_restriction),
diff --git a/drivers/net/wireless/iwlwifi/dvm/tx.c b/drivers/net/wireless/iwlwifi/dvm/tx.c
index a900aaf..5ee983f 100644
--- a/drivers/net/wireless/iwlwifi/dvm/tx.c
+++ b/drivers/net/wireless/iwlwifi/dvm/tx.c
@@ -83,8 +83,8 @@ static void iwlagn_tx_cmd_build_basic(struct iwl_priv *priv,
 	else if (ieee80211_is_back_req(fc))
 		tx_flags |= TX_CMD_FLG_ACK_MSK | TX_CMD_FLG_IMM_BA_RSP_MASK;
 	else if (info->band == IEEE80211_BAND_2GHZ &&
-		 priv->cfg->bt_params &&
-		 priv->cfg->bt_params->advanced_bt_coexist &&
+		 priv->lib->bt_params &&
+		 priv->lib->bt_params->advanced_bt_coexist &&
 		 (ieee80211_is_auth(fc) || ieee80211_is_assoc_req(fc) ||
 		 ieee80211_is_reassoc_req(fc) ||
 		 skb->protocol == cpu_to_be16(ETH_P_PAE)))
@@ -162,18 +162,6 @@ static void iwlagn_tx_cmd_build_rate(struct iwl_priv *priv,
 	if (ieee80211_is_data(fc)) {
 		tx_cmd->initial_rate_index = 0;
 		tx_cmd->tx_flags |= TX_CMD_FLG_STA_RATE_MSK;
-#ifdef CONFIG_IWLWIFI_DEVICE_TESTMODE
-		if (priv->tm_fixed_rate) {
-			/*
-			 * rate overwrite by testmode
-			 * we not only send lq command to change rate
-			 * we also re-enforce per data pkt base.
-			 */
-			tx_cmd->tx_flags &= ~TX_CMD_FLG_STA_RATE_MSK;
-			memcpy(&tx_cmd->rate_n_flags, &priv->tm_fixed_rate,
-			       sizeof(tx_cmd->rate_n_flags));
-		}
-#endif
 		return;
 	} else if (ieee80211_is_back_req(fc))
 		tx_cmd->tx_flags |= TX_CMD_FLG_STA_RATE_MSK;
@@ -202,8 +190,8 @@ static void iwlagn_tx_cmd_build_rate(struct iwl_priv *priv,
 		rate_flags |= RATE_MCS_CCK_MSK;
 
 	/* Set up antennas */
-	 if (priv->cfg->bt_params &&
-	     priv->cfg->bt_params->advanced_bt_coexist &&
+	 if (priv->lib->bt_params &&
+	     priv->lib->bt_params->advanced_bt_coexist &&
 	     priv->bt_full_concurrent) {
 		/* operated as 1x1 in full concurrency mode */
 		priv->mgmt_tx_ant = iwl_toggle_tx_ant(priv, priv->mgmt_tx_ant,
@@ -986,8 +974,8 @@ static void iwl_rx_reply_tx_agg(struct iwl_priv *priv,
 	 * notification again.
 	 */
 	if (tx_resp->bt_kill_count && tx_resp->frame_count == 1 &&
-	    priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist) {
+	    priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		IWL_DEBUG_COEX(priv, "receive reply tx w/ bt_kill\n");
 	}
 
diff --git a/drivers/net/wireless/iwlwifi/dvm/ucode.c b/drivers/net/wireless/iwlwifi/dvm/ucode.c
index 0a1cdc5..86270b6 100644
--- a/drivers/net/wireless/iwlwifi/dvm/ucode.c
+++ b/drivers/net/wireless/iwlwifi/dvm/ucode.c
@@ -132,8 +132,8 @@ int iwl_init_alive_start(struct iwl_priv *priv)
 {
 	int ret;
 
-	if (priv->cfg->bt_params &&
-	    priv->cfg->bt_params->advanced_bt_coexist) {
+	if (priv->lib->bt_params &&
+	    priv->lib->bt_params->advanced_bt_coexist) {
 		/*
 		 * Tell uCode we are ready to perform calibration
 		 * need to perform this before any calibration
@@ -155,8 +155,8 @@ int iwl_init_alive_start(struct iwl_priv *priv)
 	 * temperature offset calibration is only needed for runtime ucode,
 	 * so prepare the value now.
 	 */
-	if (priv->cfg->need_temp_offset_calib) {
-		if (priv->cfg->temp_offset_v2)
+	if (priv->lib->need_temp_offset_calib) {
+		if (priv->lib->temp_offset_v2)
 			return iwl_set_temperature_offset_calib_v2(priv);
 		else
 			return iwl_set_temperature_offset_calib(priv);
@@ -277,7 +277,7 @@ static int iwl_alive_notify(struct iwl_priv *priv)
 	if (ret)
 		return ret;
 
-	if (!priv->cfg->no_xtal_calib) {
+	if (!priv->lib->no_xtal_calib) {
 		ret = iwl_set_Xtal_calib(priv);
 		if (ret)
 			return ret;
diff --git a/drivers/net/wireless/iwlwifi/iwl-1000.c b/drivers/net/wireless/iwlwifi/iwl-1000.c
index c080ae3..0d2afe0 100644
--- a/drivers/net/wireless/iwlwifi/iwl-1000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-1000.c
@@ -60,9 +60,6 @@ static const struct iwl_base_params iwl1000_base_params = {
 	.max_ll_items = OTP_MAX_LL_ITEMS_1000,
 	.shadow_ram_support = false,
 	.led_compensation = 51,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_EXT_LONG_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_WATCHDOG_DISABLED,
 	.max_event_log_size = 128,
 };
diff --git a/drivers/net/wireless/iwlwifi/iwl-2000.c b/drivers/net/wireless/iwlwifi/iwl-2000.c
index a6ddd2f..c727ec7 100644
--- a/drivers/net/wireless/iwlwifi/iwl-2000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-2000.c
@@ -72,14 +72,9 @@ static const struct iwl_base_params iwl2000_base_params = {
 	.max_ll_items = OTP_MAX_LL_ITEMS_2x00,
 	.shadow_ram_support = true,
 	.led_compensation = 51,
-	.adv_thermal_throttle = true,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_DEF_WD_TIMEOUT,
 	.max_event_log_size = 512,
 	.shadow_reg_enable = false, /* TODO: fix bugs using this feature */
-	.hd_v2 = true,
 };
 
 
@@ -90,14 +85,9 @@ static const struct iwl_base_params iwl2030_base_params = {
 	.max_ll_items = OTP_MAX_LL_ITEMS_2x00,
 	.shadow_ram_support = true,
 	.led_compensation = 57,
-	.adv_thermal_throttle = true,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_LONG_WD_TIMEOUT,
 	.max_event_log_size = 512,
 	.shadow_reg_enable = false, /* TODO: fix bugs using this feature */
-	.hd_v2 = true,
 };
 
 static const struct iwl_ht_params iwl2000_ht_params = {
@@ -106,16 +96,6 @@ static const struct iwl_ht_params iwl2000_ht_params = {
 	.ht40_bands = BIT(IEEE80211_BAND_2GHZ),
 };
 
-static const struct iwl_bt_params iwl2030_bt_params = {
-	/* Due to bluetooth, we transmit 2.4 GHz probes only on antenna A */
-	.advanced_bt_coexist = true,
-	.agg_time_limit = BT_AGG_THRESHOLD_DEF,
-	.bt_init_traffic_load = IWL_BT_COEX_TRAFFIC_LOAD_NONE,
-	.bt_prio_boost = IWLAGN_BT_PRIO_BOOST_DEFAULT32,
-	.bt_sco_disable = true,
-	.bt_session_2 = true,
-};
-
 static const struct iwl_eeprom_params iwl20x0_eeprom_params = {
 	.regulatory_bands = {
 		EEPROM_REG_BAND_1_CHANNELS,
@@ -137,12 +117,10 @@ static const struct iwl_eeprom_params iwl20x0_eeprom_params = {
 	.device_family = IWL_DEVICE_FAMILY_2000,		\
 	.max_inst_size = IWL60_RTC_INST_SIZE,			\
 	.max_data_size = IWL60_RTC_DATA_SIZE,			\
-	.nvm_ver = EEPROM_2000_EEPROM_VERSION,		\
-	.nvm_calib_ver = EEPROM_2000_TX_POWER_VERSION,	\
+	.nvm_ver = EEPROM_2000_EEPROM_VERSION,			\
+	.nvm_calib_ver = EEPROM_2000_TX_POWER_VERSION,		\
 	.base_params = &iwl2000_base_params,			\
 	.eeprom_params = &iwl20x0_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
-	.temp_offset_v2 = true,					\
 	.led_mode = IWL_LED_RF_STATE
 
 const struct iwl_cfg iwl2000_2bgn_cfg = {
@@ -168,12 +146,8 @@ const struct iwl_cfg iwl2000_2bgn_d_cfg = {
 	.nvm_ver = EEPROM_2000_EEPROM_VERSION,		\
 	.nvm_calib_ver = EEPROM_2000_TX_POWER_VERSION,	\
 	.base_params = &iwl2030_base_params,			\
-	.bt_params = &iwl2030_bt_params,			\
 	.eeprom_params = &iwl20x0_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
-	.temp_offset_v2 = true,					\
-	.led_mode = IWL_LED_RF_STATE,				\
-	.adv_pm = true
+	.led_mode = IWL_LED_RF_STATE
 
 const struct iwl_cfg iwl2030_2bgn_cfg = {
 	.name = "Intel(R) Centrino(R) Wireless-N 2230 BGN",
@@ -193,10 +167,7 @@ const struct iwl_cfg iwl2030_2bgn_cfg = {
 	.nvm_calib_ver = EEPROM_2000_TX_POWER_VERSION,	\
 	.base_params = &iwl2000_base_params,			\
 	.eeprom_params = &iwl20x0_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
-	.temp_offset_v2 = true,					\
 	.led_mode = IWL_LED_RF_STATE,				\
-	.adv_pm = true,						\
 	.rx_with_siso_diversity = true
 
 const struct iwl_cfg iwl105_bgn_cfg = {
@@ -222,12 +193,8 @@ const struct iwl_cfg iwl105_bgn_d_cfg = {
 	.nvm_ver = EEPROM_2000_EEPROM_VERSION,		\
 	.nvm_calib_ver = EEPROM_2000_TX_POWER_VERSION,	\
 	.base_params = &iwl2030_base_params,			\
-	.bt_params = &iwl2030_bt_params,			\
 	.eeprom_params = &iwl20x0_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
-	.temp_offset_v2 = true,					\
 	.led_mode = IWL_LED_RF_STATE,				\
-	.adv_pm = true,						\
 	.rx_with_siso_diversity = true
 
 const struct iwl_cfg iwl135_bgn_cfg = {
diff --git a/drivers/net/wireless/iwlwifi/iwl-5000.c b/drivers/net/wireless/iwlwifi/iwl-5000.c
index 403f3f2..ecc01e1 100644
--- a/drivers/net/wireless/iwlwifi/iwl-5000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-5000.c
@@ -59,11 +59,8 @@ static const struct iwl_base_params iwl5000_base_params = {
 	.num_of_queues = IWLAGN_NUM_QUEUES,
 	.pll_cfg_val = CSR50_ANA_PLL_CFG_VAL,
 	.led_compensation = 51,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_LONG_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_WATCHDOG_DISABLED,
 	.max_event_log_size = 512,
-	.no_idle_support = true,
 };
 
 static const struct iwl_ht_params iwl5000_ht_params = {
@@ -159,7 +156,6 @@ const struct iwl_cfg iwl5350_agn_cfg = {
 	.nvm_calib_ver = EEPROM_5050_TX_POWER_VERSION,	\
 	.base_params = &iwl5000_base_params,			\
 	.eeprom_params = &iwl5000_eeprom_params,		\
-	.no_xtal_calib = true,					\
 	.led_mode = IWL_LED_BLINK,				\
 	.internal_wimax_coex = true
 
diff --git a/drivers/net/wireless/iwlwifi/iwl-6000.c b/drivers/net/wireless/iwlwifi/iwl-6000.c
index b5ab8d1..30d45e2 100644
--- a/drivers/net/wireless/iwlwifi/iwl-6000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-6000.c
@@ -82,10 +82,6 @@ static const struct iwl_base_params iwl6000_base_params = {
 	.max_ll_items = OTP_MAX_LL_ITEMS_6x00,
 	.shadow_ram_support = true,
 	.led_compensation = 51,
-	.adv_thermal_throttle = true,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_DEF_WD_TIMEOUT,
 	.max_event_log_size = 512,
 	.shadow_reg_enable = false, /* TODO: fix bugs using this feature */
@@ -98,10 +94,6 @@ static const struct iwl_base_params iwl6050_base_params = {
 	.max_ll_items = OTP_MAX_LL_ITEMS_6x50,
 	.shadow_ram_support = true,
 	.led_compensation = 51,
-	.adv_thermal_throttle = true,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
-	.chain_noise_scale = 1500,
 	.wd_timeout = IWL_DEF_WD_TIMEOUT,
 	.max_event_log_size = 1024,
 	.shadow_reg_enable = false, /* TODO: fix bugs using this feature */
@@ -114,10 +106,6 @@ static const struct iwl_base_params iwl6000_g2_base_params = {
 	.max_ll_items = OTP_MAX_LL_ITEMS_6x00,
 	.shadow_ram_support = true,
 	.led_compensation = 57,
-	.adv_thermal_throttle = true,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_LONG_WD_TIMEOUT,
 	.max_event_log_size = 512,
 	.shadow_reg_enable = false, /* TODO: fix bugs using this feature */
@@ -129,15 +117,6 @@ static const struct iwl_ht_params iwl6000_ht_params = {
 	.ht40_bands = BIT(IEEE80211_BAND_2GHZ) | BIT(IEEE80211_BAND_5GHZ),
 };
 
-static const struct iwl_bt_params iwl6000_bt_params = {
-	/* Due to bluetooth, we transmit 2.4 GHz probes only on antenna A */
-	.advanced_bt_coexist = true,
-	.agg_time_limit = BT_AGG_THRESHOLD_DEF,
-	.bt_init_traffic_load = IWL_BT_COEX_TRAFFIC_LOAD_NONE,
-	.bt_prio_boost = IWLAGN_BT_PRIO_BOOST_DEFAULT,
-	.bt_sco_disable = true,
-};
-
 static const struct iwl_eeprom_params iwl6000_eeprom_params = {
 	.regulatory_bands = {
 		EEPROM_REG_BAND_1_CHANNELS,
@@ -163,7 +142,6 @@ static const struct iwl_eeprom_params iwl6000_eeprom_params = {
 	.nvm_calib_ver = EEPROM_6005_TX_POWER_VERSION,	\
 	.base_params = &iwl6000_g2_base_params,			\
 	.eeprom_params = &iwl6000_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
 	.led_mode = IWL_LED_RF_STATE
 
 const struct iwl_cfg iwl6005_2agn_cfg = {
@@ -217,11 +195,8 @@ const struct iwl_cfg iwl6005_2agn_mow2_cfg = {
 	.nvm_ver = EEPROM_6030_EEPROM_VERSION,		\
 	.nvm_calib_ver = EEPROM_6030_TX_POWER_VERSION,	\
 	.base_params = &iwl6000_g2_base_params,			\
-	.bt_params = &iwl6000_bt_params,			\
 	.eeprom_params = &iwl6000_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
-	.led_mode = IWL_LED_RF_STATE,				\
-	.adv_pm = true						\
+	.led_mode = IWL_LED_RF_STATE
 
 const struct iwl_cfg iwl6030_2agn_cfg = {
 	.name = "Intel(R) Centrino(R) Advanced-N 6230 AGN",
@@ -256,11 +231,8 @@ const struct iwl_cfg iwl6030_2bg_cfg = {
 	.nvm_ver = EEPROM_6030_EEPROM_VERSION,		\
 	.nvm_calib_ver = EEPROM_6030_TX_POWER_VERSION,	\
 	.base_params = &iwl6000_g2_base_params,			\
-	.bt_params = &iwl6000_bt_params,			\
 	.eeprom_params = &iwl6000_eeprom_params,		\
-	.need_temp_offset_calib = true,				\
-	.led_mode = IWL_LED_RF_STATE,				\
-	.adv_pm = true
+	.led_mode = IWL_LED_RF_STATE
 
 const struct iwl_cfg iwl6035_2agn_cfg = {
 	.name = "Intel(R) Centrino(R) Advanced-N 6235 AGN",
diff --git a/drivers/net/wireless/iwlwifi/iwl-7000.c b/drivers/net/wireless/iwlwifi/iwl-7000.c
index 50263e8..22b7fa5 100644
--- a/drivers/net/wireless/iwlwifi/iwl-7000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-7000.c
@@ -67,16 +67,16 @@
 #include "iwl-agn-hw.h"
 
 /* Highest firmware API version supported */
-#define IWL7260_UCODE_API_MAX	6
-#define IWL3160_UCODE_API_MAX	6
+#define IWL7260_UCODE_API_MAX	7
+#define IWL3160_UCODE_API_MAX	7
 
 /* Oldest version we won't warn about */
-#define IWL7260_UCODE_API_OK	6
-#define IWL3160_UCODE_API_OK	6
+#define IWL7260_UCODE_API_OK	7
+#define IWL3160_UCODE_API_OK	7
 
 /* Lowest firmware API version supported */
-#define IWL7260_UCODE_API_MIN	6
-#define IWL3160_UCODE_API_MIN	6
+#define IWL7260_UCODE_API_MIN	7
+#define IWL3160_UCODE_API_MIN	7
 
 /* NVM versions */
 #define IWL7260_NVM_VERSION		0x0a1d
@@ -96,13 +96,9 @@ static const struct iwl_base_params iwl7000_base_params = {
 	.pll_cfg_val = 0,
 	.shadow_ram_support = true,
 	.led_compensation = 57,
-	.adv_thermal_throttle = true,
-	.support_ct_kill_exit = true,
-	.plcp_delta_threshold = IWL_MAX_PLCP_ERR_THRESHOLD_DEF,
-	.chain_noise_scale = 1000,
 	.wd_timeout = IWL_LONG_WD_TIMEOUT,
 	.max_event_log_size = 512,
-	.shadow_reg_enable = false, /* TODO: fix bugs using this feature */
+	.shadow_reg_enable = true,
 };
 
 static const struct iwl_ht_params iwl7000_ht_params = {
@@ -118,14 +114,11 @@ static const struct iwl_ht_params iwl7000_ht_params = {
 	.max_inst_size = IWL60_RTC_INST_SIZE,			\
 	.max_data_size = IWL60_RTC_DATA_SIZE,			\
 	.base_params = &iwl7000_base_params,			\
-	/* TODO: .bt_params? */					\
-	.need_temp_offset_calib = true,				\
-	.led_mode = IWL_LED_RF_STATE,				\
-	.adv_pm = true						\
+	.led_mode = IWL_LED_RF_STATE
 
 
 const struct iwl_cfg iwl7260_2ac_cfg = {
-	.name = "Intel(R) Dual Band Wireless AC7260",
+	.name = "Intel(R) Dual Band Wireless AC 7260",
 	.fw_name_pre = IWL7260_FW_PRE,
 	IWL_DEVICE_7000,
 	.ht_params = &iwl7000_ht_params,
@@ -133,8 +126,44 @@ const struct iwl_cfg iwl7260_2ac_cfg = {
 	.nvm_calib_ver = IWL7260_TX_POWER_VERSION,
 };
 
-const struct iwl_cfg iwl3160_ac_cfg = {
-	.name = "Intel(R) Dual Band Wireless AC3160",
+const struct iwl_cfg iwl7260_2n_cfg = {
+	.name = "Intel(R) Dual Band Wireless N 7260",
+	.fw_name_pre = IWL7260_FW_PRE,
+	IWL_DEVICE_7000,
+	.ht_params = &iwl7000_ht_params,
+	.nvm_ver = IWL7260_NVM_VERSION,
+	.nvm_calib_ver = IWL7260_TX_POWER_VERSION,
+};
+
+const struct iwl_cfg iwl7260_n_cfg = {
+	.name = "Intel(R) Wireless N 7260",
+	.fw_name_pre = IWL7260_FW_PRE,
+	IWL_DEVICE_7000,
+	.ht_params = &iwl7000_ht_params,
+	.nvm_ver = IWL7260_NVM_VERSION,
+	.nvm_calib_ver = IWL7260_TX_POWER_VERSION,
+};
+
+const struct iwl_cfg iwl3160_2ac_cfg = {
+	.name = "Intel(R) Dual Band Wireless AC 3160",
+	.fw_name_pre = IWL3160_FW_PRE,
+	IWL_DEVICE_7000,
+	.ht_params = &iwl7000_ht_params,
+	.nvm_ver = IWL3160_NVM_VERSION,
+	.nvm_calib_ver = IWL3160_TX_POWER_VERSION,
+};
+
+const struct iwl_cfg iwl3160_2n_cfg = {
+	.name = "Intel(R) Dual Band Wireless N 3160",
+	.fw_name_pre = IWL3160_FW_PRE,
+	IWL_DEVICE_7000,
+	.ht_params = &iwl7000_ht_params,
+	.nvm_ver = IWL3160_NVM_VERSION,
+	.nvm_calib_ver = IWL3160_TX_POWER_VERSION,
+};
+
+const struct iwl_cfg iwl3160_n_cfg = {
+	.name = "Intel(R) Wireless N 3160",
 	.fw_name_pre = IWL3160_FW_PRE,
 	IWL_DEVICE_7000,
 	.ht_params = &iwl7000_ht_params,
diff --git a/drivers/net/wireless/iwlwifi/iwl-config.h b/drivers/net/wireless/iwlwifi/iwl-config.h
index c38aa8f..83b9ff6 100644
--- a/drivers/net/wireless/iwlwifi/iwl-config.h
+++ b/drivers/net/wireless/iwlwifi/iwl-config.h
@@ -136,17 +136,9 @@ enum iwl_led_mode {
  * @led_compensation: compensate on the led on/off time per HW according
  *	to the deviation to achieve the desired led frequency.
  *	The detail algorithm is described in iwl-led.c
- * @chain_noise_num_beacons: number of beacons used to compute chain noise
- * @adv_thermal_throttle: support advance thermal throttle
- * @support_ct_kill_exit: support ct kill exit condition
- * @plcp_delta_threshold: plcp error rate threshold used to trigger
- *	radio tuning when there is a high receiving plcp error rate
- * @chain_noise_scale: default chain noise scale used for gain computation
  * @wd_timeout: TX queues watchdog timeout
  * @max_event_log_size: size of event log buffer size for ucode event logging
  * @shadow_reg_enable: HW shadow register support
- * @hd_v2: v2 of enhanced sensitivity value, used for 2000 series and up
- * @no_idle_support: do not support idle mode
  */
 struct iwl_base_params {
 	int eeprom_size;
@@ -157,31 +149,9 @@ struct iwl_base_params {
 	const u16 max_ll_items;
 	const bool shadow_ram_support;
 	u16 led_compensation;
-	bool adv_thermal_throttle;
-	bool support_ct_kill_exit;
-	u8 plcp_delta_threshold;
-	s32 chain_noise_scale;
 	unsigned int wd_timeout;
 	u32 max_event_log_size;
 	const bool shadow_reg_enable;
-	const bool hd_v2;
-	const bool no_idle_support;
-};
-
-/*
- * @advanced_bt_coexist: support advanced bt coexist
- * @bt_init_traffic_load: specify initial bt traffic load
- * @bt_prio_boost: default bt priority boost value
- * @agg_time_limit: maximum number of uSec in aggregation
- * @bt_sco_disable: uCode should not response to BT in SCO/ESCO mode
- */
-struct iwl_bt_params {
-	bool advanced_bt_coexist;
-	u8 bt_init_traffic_load;
-	u32 bt_prio_boost;
-	u16 agg_time_limit;
-	bool bt_sco_disable;
-	bool bt_session_2;
 };
 
 /*
@@ -231,16 +201,10 @@ struct iwl_eeprom_params {
  * @nvm_calib_ver: NVM calibration version
  * @lib: pointer to the lib ops
  * @base_params: pointer to basic parameters
- * @ht_params: point to ht patameters
- * @bt_params: pointer to bt parameters
- * @need_temp_offset_calib: need to perform temperature offset calibration
- * @no_xtal_calib: some devices do not need crystal calibration data,
- *	don't send it to those
+ * @ht_params: point to ht parameters
  * @led_mode: 0=blinking, 1=On(RF On)/Off(RF Off)
- * @adv_pm: advance power management
  * @rx_with_siso_diversity: 1x1 device with rx antenna diversity
  * @internal_wimax_coex: internal wifi/wimax combo device
- * @temp_offset_v2: support v2 of temperature offset calibration
  *
  * We enable the driver to be backward compatible wrt. hardware features.
  * API differences in uCode shouldn't be handled here but through TLVs
@@ -258,26 +222,23 @@ struct iwl_cfg {
 	const u32 max_inst_size;
 	u8   valid_tx_ant;
 	u8   valid_rx_ant;
+	bool bt_shared_single_ant;
 	u16  nvm_ver;
 	u16  nvm_calib_ver;
 	/* params not likely to change within a device family */
 	const struct iwl_base_params *base_params;
 	/* params likely to change within a device family */
 	const struct iwl_ht_params *ht_params;
-	const struct iwl_bt_params *bt_params;
 	const struct iwl_eeprom_params *eeprom_params;
-	const bool need_temp_offset_calib; /* if used set to true */
-	const bool no_xtal_calib;
 	enum iwl_led_mode led_mode;
-	const bool adv_pm;
 	const bool rx_with_siso_diversity;
 	const bool internal_wimax_coex;
-	const bool temp_offset_v2;
 };
 
 /*
  * This list declares the config structures for all devices.
  */
+#if IS_ENABLED(CONFIG_IWLDVM)
 extern const struct iwl_cfg iwl5300_agn_cfg;
 extern const struct iwl_cfg iwl5100_agn_cfg;
 extern const struct iwl_cfg iwl5350_agn_cfg;
@@ -319,7 +280,14 @@ extern const struct iwl_cfg iwl6035_2agn_cfg;
 extern const struct iwl_cfg iwl105_bgn_cfg;
 extern const struct iwl_cfg iwl105_bgn_d_cfg;
 extern const struct iwl_cfg iwl135_bgn_cfg;
+#endif /* CONFIG_IWLDVM */
+#if IS_ENABLED(CONFIG_IWLMVM)
 extern const struct iwl_cfg iwl7260_2ac_cfg;
-extern const struct iwl_cfg iwl3160_ac_cfg;
+extern const struct iwl_cfg iwl7260_2n_cfg;
+extern const struct iwl_cfg iwl7260_n_cfg;
+extern const struct iwl_cfg iwl3160_2ac_cfg;
+extern const struct iwl_cfg iwl3160_2n_cfg;
+extern const struct iwl_cfg iwl3160_n_cfg;
+#endif /* CONFIG_IWLMVM */
 
 #endif /* __IWL_CONFIG_H__ */
diff --git a/drivers/net/wireless/iwlwifi/iwl-csr.h b/drivers/net/wireless/iwlwifi/iwl-csr.h
index 20e845d..a276af4 100644
--- a/drivers/net/wireless/iwlwifi/iwl-csr.h
+++ b/drivers/net/wireless/iwlwifi/iwl-csr.h
@@ -472,4 +472,23 @@
 #define IWL_HOST_INT_CALIB_TIMEOUT_DEF	(0x10)
 #define IWL_HOST_INT_CALIB_TIMEOUT_MIN	(0x0)
 
+/*****************************************************************************
+ *                        7000/3000 series SHR DTS addresses                 *
+ *****************************************************************************/
+
+/* Diode Results Register Structure: */
+enum dtd_diode_reg {
+	DTS_DIODE_REG_DIG_VAL			= 0x000000FF, /* bits [7:0] */
+	DTS_DIODE_REG_VREF_LOW			= 0x0000FF00, /* bits [15:8] */
+	DTS_DIODE_REG_VREF_HIGH			= 0x00FF0000, /* bits [23:16] */
+	DTS_DIODE_REG_VREF_ID			= 0x03000000, /* bits [25:24] */
+	DTS_DIODE_REG_PASS_ONCE			= 0x80000000, /* bits [31:31] */
+	DTS_DIODE_REG_FLAGS_MSK			= 0xFF000000, /* bits [31:24] */
+/* Those are the masks INSIDE the flags bit-field: */
+	DTS_DIODE_REG_FLAGS_VREFS_ID_POS	= 0,
+	DTS_DIODE_REG_FLAGS_VREFS_ID		= 0x00000003, /* bits [1:0] */
+	DTS_DIODE_REG_FLAGS_PASS_ONCE_POS	= 7,
+	DTS_DIODE_REG_FLAGS_PASS_ONCE		= 0x00000080, /* bits [7:7] */
+};
+
 #endif /* !__iwl_csr_h__ */
diff --git a/drivers/net/wireless/iwlwifi/iwl-debug.h b/drivers/net/wireless/iwlwifi/iwl-debug.h
index 8cf5db7..7edb851 100644
--- a/drivers/net/wireless/iwlwifi/iwl-debug.h
+++ b/drivers/net/wireless/iwlwifi/iwl-debug.h
@@ -34,7 +34,11 @@
 
 static inline bool iwl_have_debug_level(u32 level)
 {
+#ifdef CONFIG_IWLWIFI_DEBUG
 	return iwlwifi_mod_params.debug_level & level;
+#else
+	return false;
+#endif
 }
 
 void __iwl_err(struct device *dev, bool rfkill_prefix, bool only_trace,
diff --git a/drivers/net/wireless/iwlwifi/iwl-drv.c b/drivers/net/wireless/iwlwifi/iwl-drv.c
index 40fed1f..d0162d4 100644
--- a/drivers/net/wireless/iwlwifi/iwl-drv.c
+++ b/drivers/net/wireless/iwlwifi/iwl-drv.c
@@ -1111,11 +1111,8 @@ void iwl_drv_stop(struct iwl_drv *drv)
 /* shared module parameters */
 struct iwl_mod_params iwlwifi_mod_params = {
 	.restart_fw = true,
-	.plcp_check = true,
 	.bt_coex_active = true,
 	.power_level = IWL_POWER_INDEX_1,
-	.bt_ch_announce = true,
-	.auto_agg = true,
 	.wd_disable = true,
 	/* the rest are 0 by default */
 };
@@ -1223,19 +1220,14 @@ module_param_named(antenna_coupling, iwlwifi_mod_params.ant_coupling,
 MODULE_PARM_DESC(antenna_coupling,
 		 "specify antenna coupling in dB (defualt: 0 dB)");
 
-module_param_named(bt_ch_inhibition, iwlwifi_mod_params.bt_ch_announce,
-		   bool, S_IRUGO);
-MODULE_PARM_DESC(bt_ch_inhibition,
-		 "Enable BT channel inhibition (default: enable)");
-
-module_param_named(plcp_check, iwlwifi_mod_params.plcp_check, bool, S_IRUGO);
-MODULE_PARM_DESC(plcp_check, "Check plcp health (default: 1 [enabled])");
-
 module_param_named(wd_disable, iwlwifi_mod_params.wd_disable, int, S_IRUGO);
 MODULE_PARM_DESC(wd_disable,
 		"Disable stuck queue watchdog timer 0=system default, "
 		"1=disable, 2=enable (default: 0)");
 
+module_param_named(nvm_file, iwlwifi_mod_params.nvm_file, charp, S_IRUGO);
+MODULE_PARM_DESC(nvm_file, "NVM file name");
+
 /*
  * set bt_coex_active to true, uCode will do kill/defer
  * every time the priority line is asserted (BT is sending signals on the
@@ -1269,8 +1261,3 @@ module_param_named(power_level, iwlwifi_mod_params.power_level,
 		int, S_IRUGO);
 MODULE_PARM_DESC(power_level,
 		 "default power save level (range from 1 - 5, default: 1)");
-
-module_param_named(auto_agg, iwlwifi_mod_params.auto_agg,
-		bool, S_IRUGO);
-MODULE_PARM_DESC(auto_agg,
-		 "enable agg w/o check traffic load (default: enable)");
diff --git a/drivers/net/wireless/iwlwifi/iwl-drv.h b/drivers/net/wireless/iwlwifi/iwl-drv.h
index 7d14509..429337a 100644
--- a/drivers/net/wireless/iwlwifi/iwl-drv.h
+++ b/drivers/net/wireless/iwlwifi/iwl-drv.h
@@ -62,8 +62,7 @@
 
 #ifndef __iwl_drv_h__
 #define __iwl_drv_h__
-
-#include <linux/module.h>
+#include <linux/export.h>
 
 /* for all modules */
 #define DRV_NAME        "iwlwifi"
diff --git a/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.c b/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.c
index 600c9fd..4c887f3 100644
--- a/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.c
+++ b/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.c
@@ -732,17 +732,16 @@ int iwl_init_sband_channels(struct iwl_nvm_data *data,
 void iwl_init_ht_hw_capab(const struct iwl_cfg *cfg,
 			  struct iwl_nvm_data *data,
 			  struct ieee80211_sta_ht_cap *ht_info,
-			  enum ieee80211_band band)
+			  enum ieee80211_band band,
+			  u8 tx_chains, u8 rx_chains)
 {
 	int max_bit_rate = 0;
-	u8 rx_chains;
-	u8 tx_chains;
 
-	tx_chains = hweight8(data->valid_tx_ant);
+	tx_chains = hweight8(tx_chains);
 	if (cfg->rx_with_siso_diversity)
 		rx_chains = 1;
 	else
-		rx_chains = hweight8(data->valid_rx_ant);
+		rx_chains = hweight8(rx_chains);
 
 	if (!(data->sku_cap_11n_enable) || !cfg->ht_params) {
 		ht_info->ht_supported = false;
@@ -806,7 +805,8 @@ static void iwl_init_sbands(struct device *dev, const struct iwl_cfg *cfg,
 	sband->n_bitrates = N_RATES_24;
 	n_used += iwl_init_sband_channels(data, sband, n_channels,
 					  IEEE80211_BAND_2GHZ);
-	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_2GHZ);
+	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_2GHZ,
+			     data->valid_tx_ant, data->valid_rx_ant);
 
 	sband = &data->bands[IEEE80211_BAND_5GHZ];
 	sband->band = IEEE80211_BAND_5GHZ;
@@ -814,7 +814,8 @@ static void iwl_init_sbands(struct device *dev, const struct iwl_cfg *cfg,
 	sband->n_bitrates = N_RATES_52;
 	n_used += iwl_init_sband_channels(data, sband, n_channels,
 					  IEEE80211_BAND_5GHZ);
-	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_5GHZ);
+	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_5GHZ,
+			     data->valid_tx_ant, data->valid_rx_ant);
 
 	if (n_channels != n_used)
 		IWL_ERR_DEV(dev, "EEPROM: used only %d of %d channels\n",
diff --git a/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.h b/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.h
index 37f1153..d73304a 100644
--- a/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.h
+++ b/drivers/net/wireless/iwlwifi/iwl-eeprom-parse.h
@@ -133,6 +133,7 @@ int iwl_init_sband_channels(struct iwl_nvm_data *data,
 void iwl_init_ht_hw_capab(const struct iwl_cfg *cfg,
 			  struct iwl_nvm_data *data,
 			  struct ieee80211_sta_ht_cap *ht_info,
-			  enum ieee80211_band band);
+			  enum ieee80211_band band,
+			  u8 tx_chains, u8 rx_chains);
 
 #endif /* __iwl_eeprom_parse_h__ */
diff --git a/drivers/net/wireless/iwlwifi/iwl-fw.h b/drivers/net/wireless/iwlwifi/iwl-fw.h
index c4c446d..f844d5c 100644
--- a/drivers/net/wireless/iwlwifi/iwl-fw.h
+++ b/drivers/net/wireless/iwlwifi/iwl-fw.h
@@ -106,11 +106,14 @@ enum iwl_ucode_type {
 
 /*
  * enumeration of ucode section.
- * This enumeration is used for legacy tlv style (before 16.0 uCode).
+ * This enumeration is used directly for older firmware (before 16.0).
+ * For new firmware, there can be up to 4 sections (see below) but the
+ * first one packaged into the firmware file is the DATA section and
+ * some debugging code accesses that.
  */
 enum iwl_ucode_sec {
-	IWL_UCODE_SECTION_INST,
 	IWL_UCODE_SECTION_DATA,
+	IWL_UCODE_SECTION_INST,
 };
 /*
  * For 16.0 uCode and above, there is no differentiation between sections,
diff --git a/drivers/net/wireless/iwlwifi/iwl-modparams.h b/drivers/net/wireless/iwlwifi/iwl-modparams.h
index d6f6c37..a1f580c 100644
--- a/drivers/net/wireless/iwlwifi/iwl-modparams.h
+++ b/drivers/net/wireless/iwlwifi/iwl-modparams.h
@@ -93,7 +93,6 @@ enum iwl_power_level {
  *	use IWL_DISABLE_HT_* constants
  * @amsdu_size_8K: enable 8K amsdu size, default = 0
  * @restart_fw: restart firmware, default = 1
- * @plcp_check: enable plcp health check, default = true
  * @wd_disable: enable stuck queue check, default = 0
  * @bt_coex_active: enable bt coex, default = true
  * @led_mode: system default, default = 0
@@ -101,24 +100,22 @@ enum iwl_power_level {
  * @power_level: power level, default = 1
  * @debug_level: levels are IWL_DL_*
  * @ant_coupling: antenna coupling in dB, default = 0
- * @bt_ch_announce: BT channel inhibition, default = enable
- * @auto_agg: enable agg. without check, default = true
  */
 struct iwl_mod_params {
 	int sw_crypto;
 	unsigned int disable_11n;
 	int amsdu_size_8K;
 	bool restart_fw;
-	bool plcp_check;
 	int  wd_disable;
 	bool bt_coex_active;
 	int led_mode;
 	bool power_save;
 	int power_level;
+#ifdef CONFIG_IWLWIFI_DEBUG
 	u32 debug_level;
+#endif
 	int ant_coupling;
-	bool bt_ch_announce;
-	bool auto_agg;
+	char *nvm_file;
 };
 
 #endif /* #__iwl_modparams_h__ */
diff --git a/drivers/net/wireless/iwlwifi/iwl-nvm-parse.c b/drivers/net/wireless/iwlwifi/iwl-nvm-parse.c
index 6199a0a..acd2665 100644
--- a/drivers/net/wireless/iwlwifi/iwl-nvm-parse.c
+++ b/drivers/net/wireless/iwlwifi/iwl-nvm-parse.c
@@ -89,6 +89,7 @@ enum nvm_sku_bits {
 	NVM_SKU_CAP_BAND_24GHZ	= BIT(0),
 	NVM_SKU_CAP_BAND_52GHZ	= BIT(1),
 	NVM_SKU_CAP_11N_ENABLE	= BIT(2),
+	NVM_SKU_CAP_11AC_ENABLE	= BIT(3),
 };
 
 /* radio config bits (actual values from NVM definition) */
@@ -258,8 +259,6 @@ static void iwl_init_vht_hw_capab(const struct iwl_cfg *cfg,
 				  struct iwl_nvm_data *data,
 				  struct ieee80211_sta_vht_cap *vht_cap)
 {
-	/* For now, assume new devices with NVM are VHT capable */
-
 	vht_cap->vht_supported = true;
 
 	vht_cap->cap = IEEE80211_VHT_CAP_SHORT_GI_80 |
@@ -292,7 +291,8 @@ static void iwl_init_vht_hw_capab(const struct iwl_cfg *cfg,
 }
 
 static void iwl_init_sbands(struct device *dev, const struct iwl_cfg *cfg,
-			    struct iwl_nvm_data *data, const __le16 *nvm_sw)
+			    struct iwl_nvm_data *data, const __le16 *nvm_sw,
+			    bool enable_vht, u8 tx_chains, u8 rx_chains)
 {
 	int n_channels = iwl_init_channel_map(dev, cfg, data,
 			&nvm_sw[NVM_CHANNELS]);
@@ -305,7 +305,8 @@ static void iwl_init_sbands(struct device *dev, const struct iwl_cfg *cfg,
 	sband->n_bitrates = N_RATES_24;
 	n_used += iwl_init_sband_channels(data, sband, n_channels,
 					  IEEE80211_BAND_2GHZ);
-	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_2GHZ);
+	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_2GHZ,
+			     tx_chains, rx_chains);
 
 	sband = &data->bands[IEEE80211_BAND_5GHZ];
 	sband->band = IEEE80211_BAND_5GHZ;
@@ -313,8 +314,10 @@ static void iwl_init_sbands(struct device *dev, const struct iwl_cfg *cfg,
 	sband->n_bitrates = N_RATES_52;
 	n_used += iwl_init_sband_channels(data, sband, n_channels,
 					  IEEE80211_BAND_5GHZ);
-	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_5GHZ);
-	iwl_init_vht_hw_capab(cfg, data, &sband->vht_cap);
+	iwl_init_ht_hw_capab(cfg, data, &sband->ht_cap, IEEE80211_BAND_5GHZ,
+			     tx_chains, rx_chains);
+	if (enable_vht)
+		iwl_init_vht_hw_capab(cfg, data, &sband->vht_cap);
 
 	if (n_channels != n_used)
 		IWL_ERR_DEV(dev, "NVM: used only %d of %d channels\n",
@@ -324,7 +327,7 @@ static void iwl_init_sbands(struct device *dev, const struct iwl_cfg *cfg,
 struct iwl_nvm_data *
 iwl_parse_nvm_data(struct device *dev, const struct iwl_cfg *cfg,
 		   const __le16 *nvm_hw, const __le16 *nvm_sw,
-		   const __le16 *nvm_calib)
+		   const __le16 *nvm_calib, u8 tx_chains, u8 rx_chains)
 {
 	struct iwl_nvm_data *data;
 	u8 hw_addr[ETH_ALEN];
@@ -380,7 +383,8 @@ iwl_parse_nvm_data(struct device *dev, const struct iwl_cfg *cfg,
 	data->hw_addr[4] = hw_addr[5];
 	data->hw_addr[5] = hw_addr[4];
 
-	iwl_init_sbands(dev, cfg, data, nvm_sw);
+	iwl_init_sbands(dev, cfg, data, nvm_sw, sku & NVM_SKU_CAP_11AC_ENABLE,
+			tx_chains, rx_chains);
 
 	data->calib_version = 255;   /* TODO:
 					this value will prevent some checks from
diff --git a/drivers/net/wireless/iwlwifi/iwl-nvm-parse.h b/drivers/net/wireless/iwlwifi/iwl-nvm-parse.h
index e57fb98..3325059 100644
--- a/drivers/net/wireless/iwlwifi/iwl-nvm-parse.h
+++ b/drivers/net/wireless/iwlwifi/iwl-nvm-parse.h
@@ -75,6 +75,6 @@
 struct iwl_nvm_data *
 iwl_parse_nvm_data(struct device *dev, const struct iwl_cfg *cfg,
 		   const __le16 *nvm_hw, const __le16 *nvm_sw,
-		   const __le16 *nvm_calib);
+		   const __le16 *nvm_calib, u8 tx_chains, u8 rx_chains);
 
 #endif /* __iwl_nvm_parse_h__ */
diff --git a/drivers/net/wireless/iwlwifi/iwl-phy-db.c b/drivers/net/wireless/iwlwifi/iwl-phy-db.c
index 25745da..1a405ae 100644
--- a/drivers/net/wireless/iwlwifi/iwl-phy-db.c
+++ b/drivers/net/wireless/iwlwifi/iwl-phy-db.c
@@ -92,20 +92,16 @@ struct iwl_phy_db_entry {
 struct iwl_phy_db {
 	struct iwl_phy_db_entry	cfg;
 	struct iwl_phy_db_entry	calib_nch;
-	struct iwl_phy_db_entry	calib_ch;
 	struct iwl_phy_db_entry	calib_ch_group_papd[IWL_NUM_PAPD_CH_GROUPS];
 	struct iwl_phy_db_entry	calib_ch_group_txp[IWL_NUM_TXP_CH_GROUPS];
 
-	u32 channel_num;
-	u32 channel_size;
-
 	struct iwl_trans *trans;
 };
 
 enum iwl_phy_db_section_type {
 	IWL_PHY_DB_CFG = 1,
 	IWL_PHY_DB_CALIB_NCH,
-	IWL_PHY_DB_CALIB_CH,
+	IWL_PHY_DB_UNUSED,
 	IWL_PHY_DB_CALIB_CHG_PAPD,
 	IWL_PHY_DB_CALIB_CHG_TXP,
 	IWL_PHY_DB_MAX
@@ -169,8 +165,6 @@ iwl_phy_db_get_section(struct iwl_phy_db *phy_db,
 		return &phy_db->cfg;
 	case IWL_PHY_DB_CALIB_NCH:
 		return &phy_db->calib_nch;
-	case IWL_PHY_DB_CALIB_CH:
-		return &phy_db->calib_ch;
 	case IWL_PHY_DB_CALIB_CHG_PAPD:
 		if (chg_id >= IWL_NUM_PAPD_CH_GROUPS)
 			return NULL;
@@ -208,7 +202,6 @@ void iwl_phy_db_free(struct iwl_phy_db *phy_db)
 
 	iwl_phy_db_free_section(phy_db, IWL_PHY_DB_CFG, 0);
 	iwl_phy_db_free_section(phy_db, IWL_PHY_DB_CALIB_NCH, 0);
-	iwl_phy_db_free_section(phy_db, IWL_PHY_DB_CALIB_CH, 0);
 	for (i = 0; i < IWL_NUM_PAPD_CH_GROUPS; i++)
 		iwl_phy_db_free_section(phy_db, IWL_PHY_DB_CALIB_CHG_PAPD, i);
 	for (i = 0; i < IWL_NUM_TXP_CH_GROUPS; i++)
@@ -248,13 +241,6 @@ int iwl_phy_db_set_section(struct iwl_phy_db *phy_db, struct iwl_rx_packet *pkt,
 
 	entry->size = size;
 
-	if (type == IWL_PHY_DB_CALIB_CH) {
-		phy_db->channel_num =
-			le32_to_cpup((__le32 *)phy_db_notif->data);
-		phy_db->channel_size =
-			(size - CHANNEL_NUM_SIZE) / phy_db->channel_num;
-	}
-
 	IWL_DEBUG_INFO(phy_db->trans,
 		       "%s(%d): [PHYDB]SET: Type %d , Size: %d\n",
 		       __func__, __LINE__, type, size);
@@ -328,10 +314,7 @@ int iwl_phy_db_get_section_data(struct iwl_phy_db *phy_db,
 				u32 type, u8 **data, u16 *size, u16 ch_id)
 {
 	struct iwl_phy_db_entry *entry;
-	u32 channel_num;
-	u32 channel_size;
 	u16 ch_group_id = 0;
-	u16 index;
 
 	if (!phy_db)
 		return -EINVAL;
@@ -346,21 +329,8 @@ int iwl_phy_db_get_section_data(struct iwl_phy_db *phy_db,
 	if (!entry)
 		return -EINVAL;
 
-	if (type == IWL_PHY_DB_CALIB_CH) {
-		index = ch_id_to_ch_index(ch_id);
-		channel_num = phy_db->channel_num;
-		channel_size = phy_db->channel_size;
-		if (index >= channel_num) {
-			IWL_ERR(phy_db->trans, "Wrong channel number %d\n",
-				ch_id);
-			return -EINVAL;
-		}
-		*data = entry->data + CHANNEL_NUM_SIZE + index * channel_size;
-		*size = channel_size;
-	} else {
-		*data = entry->data;
-		*size = entry->size;
-	}
+	*data = entry->data;
+	*size = entry->size;
 
 	IWL_DEBUG_INFO(phy_db->trans,
 		       "%s(%d): [PHYDB] GET: Type %d , Size: %d\n",
@@ -413,6 +383,9 @@ static int iwl_phy_db_send_all_channel_groups(
 		if (!entry)
 			return -EINVAL;
 
+		if (WARN_ON_ONCE(!entry->size))
+			continue;
+
 		/* Send the requested PHY DB section */
 		err = iwl_send_phy_db_cmd(phy_db,
 					  type,
diff --git a/drivers/net/wireless/iwlwifi/iwl-prph.h b/drivers/net/wireless/iwlwifi/iwl-prph.h
index 386f2a7..ff8cc75 100644
--- a/drivers/net/wireless/iwlwifi/iwl-prph.h
+++ b/drivers/net/wireless/iwlwifi/iwl-prph.h
@@ -100,6 +100,18 @@
 /* Device system time */
 #define DEVICE_SYSTEM_TIME_REG 0xA0206C
 
+/*****************************************************************************
+ *                        7000/3000 series SHR DTS addresses                 *
+ *****************************************************************************/
+
+#define SHR_MISC_WFM_DTS_EN	(0x00a10024)
+#define DTSC_CFG_MODE		(0x00a10604)
+#define DTSC_VREF_AVG		(0x00a10648)
+#define DTSC_VREF5_AVG		(0x00a1064c)
+#define DTSC_CFG_MODE_PERIODIC	(0x2)
+#define DTSC_PTAT_AVG		(0x00a10650)
+
+
 /**
  * Tx Scheduler
  *
diff --git a/drivers/net/wireless/iwlwifi/iwl-test.c b/drivers/net/wireless/iwlwifi/iwl-test.c
deleted file mode 100644
index 5cfd55b..0000000
--- a/drivers/net/wireless/iwlwifi/iwl-test.c
+++ /dev/null
@@ -1,852 +0,0 @@
-/******************************************************************************
- *
- * This file is provided under a dual BSD/GPLv2 license.  When using or
- * redistributing this file, you may do so under either license.
- *
- * GPL LICENSE SUMMARY
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110,
- * USA
- *
- * The full GNU General Public License is included in this distribution
- * in the file called COPYING.
- *
- * Contact Information:
- *  Intel Linux Wireless <ilw@linux.intel.com>
- * Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
- *
- * BSD LICENSE
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- *  * Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- *  * Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *  * Neither the name Intel Corporation nor the names of its
- *    contributors may be used to endorse or promote products derived
- *    from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- *****************************************************************************/
-
-#include <linux/export.h>
-#include <net/netlink.h>
-
-#include "iwl-drv.h"
-#include "iwl-io.h"
-#include "iwl-fh.h"
-#include "iwl-prph.h"
-#include "iwl-trans.h"
-#include "iwl-test.h"
-#include "iwl-csr.h"
-#include "iwl-testmode.h"
-
-/*
- * Periphery registers absolute lower bound. This is used in order to
- * differentiate registery access through HBUS_TARG_PRPH_* and
- * HBUS_TARG_MEM_* accesses.
- */
-#define IWL_ABS_PRPH_START (0xA00000)
-
-/*
- * The TLVs used in the gnl message policy between the kernel module and
- * user space application. iwl_testmode_gnl_msg_policy is to be carried
- * through the NL80211_CMD_TESTMODE channel regulated by nl80211.
- * See iwl-testmode.h
- */
-static
-struct nla_policy iwl_testmode_gnl_msg_policy[IWL_TM_ATTR_MAX] = {
-	[IWL_TM_ATTR_COMMAND] = { .type = NLA_U32, },
-
-	[IWL_TM_ATTR_UCODE_CMD_ID] = { .type = NLA_U8, },
-	[IWL_TM_ATTR_UCODE_CMD_DATA] = { .type = NLA_UNSPEC, },
-
-	[IWL_TM_ATTR_REG_OFFSET] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_REG_VALUE8] = { .type = NLA_U8, },
-	[IWL_TM_ATTR_REG_VALUE32] = { .type = NLA_U32, },
-
-	[IWL_TM_ATTR_SYNC_RSP] = { .type = NLA_UNSPEC, },
-	[IWL_TM_ATTR_UCODE_RX_PKT] = { .type = NLA_UNSPEC, },
-
-	[IWL_TM_ATTR_EEPROM] = { .type = NLA_UNSPEC, },
-
-	[IWL_TM_ATTR_TRACE_ADDR] = { .type = NLA_UNSPEC, },
-	[IWL_TM_ATTR_TRACE_DUMP] = { .type = NLA_UNSPEC, },
-	[IWL_TM_ATTR_TRACE_SIZE] = { .type = NLA_U32, },
-
-	[IWL_TM_ATTR_FIXRATE] = { .type = NLA_U32, },
-
-	[IWL_TM_ATTR_UCODE_OWNER] = { .type = NLA_U8, },
-
-	[IWL_TM_ATTR_MEM_ADDR] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_BUFFER_SIZE] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_BUFFER_DUMP] = { .type = NLA_UNSPEC, },
-
-	[IWL_TM_ATTR_FW_VERSION] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_DEVICE_ID] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_FW_TYPE] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_FW_INST_SIZE] = { .type = NLA_U32, },
-	[IWL_TM_ATTR_FW_DATA_SIZE] = { .type = NLA_U32, },
-
-	[IWL_TM_ATTR_ENABLE_NOTIFICATION] = {.type = NLA_FLAG, },
-};
-
-static inline void iwl_test_trace_clear(struct iwl_test *tst)
-{
-	memset(&tst->trace, 0, sizeof(struct iwl_test_trace));
-}
-
-static void iwl_test_trace_stop(struct iwl_test *tst)
-{
-	if (!tst->trace.enabled)
-		return;
-
-	if (tst->trace.cpu_addr && tst->trace.dma_addr)
-		dma_free_coherent(tst->trans->dev,
-				  tst->trace.tsize,
-				  tst->trace.cpu_addr,
-				  tst->trace.dma_addr);
-
-	iwl_test_trace_clear(tst);
-}
-
-static inline void iwl_test_mem_clear(struct iwl_test *tst)
-{
-	memset(&tst->mem, 0, sizeof(struct iwl_test_mem));
-}
-
-static inline void iwl_test_mem_stop(struct iwl_test *tst)
-{
-	if (!tst->mem.in_read)
-		return;
-
-	iwl_test_mem_clear(tst);
-}
-
-/*
- * Initializes the test object
- * During the lifetime of the test object it is assumed that the transport is
- * started. The test object should be stopped before the transport is stopped.
- */
-void iwl_test_init(struct iwl_test *tst, struct iwl_trans *trans,
-		   struct iwl_test_ops *ops)
-{
-	tst->trans = trans;
-	tst->ops = ops;
-
-	iwl_test_trace_clear(tst);
-	iwl_test_mem_clear(tst);
-}
-EXPORT_SYMBOL_GPL(iwl_test_init);
-
-/*
- * Stop the test object
- */
-void iwl_test_free(struct iwl_test *tst)
-{
-	iwl_test_mem_stop(tst);
-	iwl_test_trace_stop(tst);
-}
-EXPORT_SYMBOL_GPL(iwl_test_free);
-
-static inline int iwl_test_send_cmd(struct iwl_test *tst,
-				    struct iwl_host_cmd *cmd)
-{
-	return tst->ops->send_cmd(tst->trans->op_mode, cmd);
-}
-
-static inline bool iwl_test_valid_hw_addr(struct iwl_test *tst, u32 addr)
-{
-	return tst->ops->valid_hw_addr(addr);
-}
-
-static inline u32 iwl_test_fw_ver(struct iwl_test *tst)
-{
-	return tst->ops->get_fw_ver(tst->trans->op_mode);
-}
-
-static inline struct sk_buff*
-iwl_test_alloc_reply(struct iwl_test *tst, int len)
-{
-	return tst->ops->alloc_reply(tst->trans->op_mode, len);
-}
-
-static inline int iwl_test_reply(struct iwl_test *tst, struct sk_buff *skb)
-{
-	return tst->ops->reply(tst->trans->op_mode, skb);
-}
-
-static inline struct sk_buff*
-iwl_test_alloc_event(struct iwl_test *tst, int len)
-{
-	return tst->ops->alloc_event(tst->trans->op_mode, len);
-}
-
-static inline void
-iwl_test_event(struct iwl_test *tst, struct sk_buff *skb)
-{
-	return tst->ops->event(tst->trans->op_mode, skb);
-}
-
-/*
- * This function handles the user application commands to the fw. The fw
- * commands are sent in a synchronuous manner. In case that the user requested
- * to get commands response, it is send to the user.
- */
-static int iwl_test_fw_cmd(struct iwl_test *tst, struct nlattr **tb)
-{
-	struct iwl_host_cmd cmd;
-	struct iwl_rx_packet *pkt;
-	struct sk_buff *skb;
-	void *reply_buf;
-	u32 reply_len;
-	int ret;
-	bool cmd_want_skb;
-
-	memset(&cmd, 0, sizeof(struct iwl_host_cmd));
-
-	if (!tb[IWL_TM_ATTR_UCODE_CMD_ID] ||
-	    !tb[IWL_TM_ATTR_UCODE_CMD_DATA]) {
-		IWL_ERR(tst->trans, "Missing fw command mandatory fields\n");
-		return -ENOMSG;
-	}
-
-	cmd.flags = CMD_ON_DEMAND | CMD_SYNC;
-	cmd_want_skb = nla_get_flag(tb[IWL_TM_ATTR_UCODE_CMD_SKB]);
-	if (cmd_want_skb)
-		cmd.flags |= CMD_WANT_SKB;
-
-	cmd.id = nla_get_u8(tb[IWL_TM_ATTR_UCODE_CMD_ID]);
-	cmd.data[0] = nla_data(tb[IWL_TM_ATTR_UCODE_CMD_DATA]);
-	cmd.len[0] = nla_len(tb[IWL_TM_ATTR_UCODE_CMD_DATA]);
-	cmd.dataflags[0] = IWL_HCMD_DFL_NOCOPY;
-	IWL_DEBUG_INFO(tst->trans, "test fw cmd=0x%x, flags 0x%x, len %d\n",
-		       cmd.id, cmd.flags, cmd.len[0]);
-
-	ret = iwl_test_send_cmd(tst, &cmd);
-	if (ret) {
-		IWL_ERR(tst->trans, "Failed to send hcmd\n");
-		return ret;
-	}
-	if (!cmd_want_skb)
-		return ret;
-
-	/* Handling return of SKB to the user */
-	pkt = cmd.resp_pkt;
-	if (!pkt) {
-		IWL_ERR(tst->trans, "HCMD received a null response packet\n");
-		return ret;
-	}
-
-	reply_len = le32_to_cpu(pkt->len_n_flags) & FH_RSCSR_FRAME_SIZE_MSK;
-	skb = iwl_test_alloc_reply(tst, reply_len + 20);
-	reply_buf = kmemdup(&pkt->hdr, reply_len, GFP_KERNEL);
-	if (!skb || !reply_buf) {
-		kfree_skb(skb);
-		kfree(reply_buf);
-		return -ENOMEM;
-	}
-
-	/* The reply is in a page, that we cannot send to user space. */
-	iwl_free_resp(&cmd);
-
-	if (nla_put_u32(skb, IWL_TM_ATTR_COMMAND,
-			IWL_TM_CMD_DEV2APP_UCODE_RX_PKT) ||
-	    nla_put(skb, IWL_TM_ATTR_UCODE_RX_PKT, reply_len, reply_buf))
-		goto nla_put_failure;
-	return iwl_test_reply(tst, skb);
-
-nla_put_failure:
-	IWL_DEBUG_INFO(tst->trans, "Failed creating NL attributes\n");
-	kfree(reply_buf);
-	kfree_skb(skb);
-	return -ENOMSG;
-}
-
-/*
- * Handles the user application commands for register access.
- */
-static int iwl_test_reg(struct iwl_test *tst, struct nlattr **tb)
-{
-	u32 ofs, val32, cmd;
-	u8 val8;
-	struct sk_buff *skb;
-	int status = 0;
-	struct iwl_trans *trans = tst->trans;
-
-	if (!tb[IWL_TM_ATTR_REG_OFFSET]) {
-		IWL_ERR(trans, "Missing reg offset\n");
-		return -ENOMSG;
-	}
-
-	ofs = nla_get_u32(tb[IWL_TM_ATTR_REG_OFFSET]);
-	IWL_DEBUG_INFO(trans, "test reg access cmd offset=0x%x\n", ofs);
-
-	cmd = nla_get_u32(tb[IWL_TM_ATTR_COMMAND]);
-
-	/*
-	 * Allow access only to FH/CSR/HBUS in direct mode.
-	 * Since we don't have the upper bounds for the CSR and HBUS segments,
-	 * we will use only the upper bound of FH for sanity check.
-	 */
-	if (ofs >= FH_MEM_UPPER_BOUND) {
-		IWL_ERR(trans, "offset out of segment (0x0 - 0x%x)\n",
-			FH_MEM_UPPER_BOUND);
-		return -EINVAL;
-	}
-
-	switch (cmd) {
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_READ32:
-		val32 = iwl_read_direct32(tst->trans, ofs);
-		IWL_DEBUG_INFO(trans, "32 value to read 0x%x\n", val32);
-
-		skb = iwl_test_alloc_reply(tst, 20);
-		if (!skb) {
-			IWL_ERR(trans, "Memory allocation fail\n");
-			return -ENOMEM;
-		}
-		if (nla_put_u32(skb, IWL_TM_ATTR_REG_VALUE32, val32))
-			goto nla_put_failure;
-		status = iwl_test_reply(tst, skb);
-		if (status < 0)
-			IWL_ERR(trans, "Error sending msg : %d\n", status);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE32:
-		if (!tb[IWL_TM_ATTR_REG_VALUE32]) {
-			IWL_ERR(trans, "Missing value to write\n");
-			return -ENOMSG;
-		} else {
-			val32 = nla_get_u32(tb[IWL_TM_ATTR_REG_VALUE32]);
-			IWL_DEBUG_INFO(trans, "32b write val=0x%x\n", val32);
-			iwl_write_direct32(tst->trans, ofs, val32);
-		}
-		break;
-
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE8:
-		if (!tb[IWL_TM_ATTR_REG_VALUE8]) {
-			IWL_ERR(trans, "Missing value to write\n");
-			return -ENOMSG;
-		} else {
-			val8 = nla_get_u8(tb[IWL_TM_ATTR_REG_VALUE8]);
-			IWL_DEBUG_INFO(trans, "8b write val=0x%x\n", val8);
-			iwl_write8(tst->trans, ofs, val8);
-		}
-		break;
-
-	default:
-		IWL_ERR(trans, "Unknown test register cmd ID\n");
-		return -ENOMSG;
-	}
-
-	return status;
-
-nla_put_failure:
-	kfree_skb(skb);
-	return -EMSGSIZE;
-}
-
-/*
- * Handles the request to start FW tracing. Allocates of the trace buffer
- * and sends a reply to user space with the address of the allocated buffer.
- */
-static int iwl_test_trace_begin(struct iwl_test *tst, struct nlattr **tb)
-{
-	struct sk_buff *skb;
-	int status = 0;
-
-	if (tst->trace.enabled)
-		return -EBUSY;
-
-	if (!tb[IWL_TM_ATTR_TRACE_SIZE])
-		tst->trace.size = TRACE_BUFF_SIZE_DEF;
-	else
-		tst->trace.size =
-			nla_get_u32(tb[IWL_TM_ATTR_TRACE_SIZE]);
-
-	if (!tst->trace.size)
-		return -EINVAL;
-
-	if (tst->trace.size < TRACE_BUFF_SIZE_MIN ||
-	    tst->trace.size > TRACE_BUFF_SIZE_MAX)
-		return -EINVAL;
-
-	tst->trace.tsize = tst->trace.size + TRACE_BUFF_PADD;
-	tst->trace.cpu_addr = dma_alloc_coherent(tst->trans->dev,
-						 tst->trace.tsize,
-						 &tst->trace.dma_addr,
-						 GFP_KERNEL);
-	if (!tst->trace.cpu_addr)
-		return -ENOMEM;
-
-	tst->trace.enabled = true;
-	tst->trace.trace_addr = (u8 *)PTR_ALIGN(tst->trace.cpu_addr, 0x100);
-
-	memset(tst->trace.trace_addr, 0x03B, tst->trace.size);
-
-	skb = iwl_test_alloc_reply(tst, sizeof(tst->trace.dma_addr) + 20);
-	if (!skb) {
-		IWL_ERR(tst->trans, "Memory allocation fail\n");
-		iwl_test_trace_stop(tst);
-		return -ENOMEM;
-	}
-
-	if (nla_put(skb, IWL_TM_ATTR_TRACE_ADDR,
-		    sizeof(tst->trace.dma_addr),
-		    (u64 *)&tst->trace.dma_addr))
-		goto nla_put_failure;
-
-	status = iwl_test_reply(tst, skb);
-	if (status < 0)
-		IWL_ERR(tst->trans, "Error sending msg : %d\n", status);
-
-	tst->trace.nchunks = DIV_ROUND_UP(tst->trace.size,
-					  DUMP_CHUNK_SIZE);
-
-	return status;
-
-nla_put_failure:
-	kfree_skb(skb);
-	if (nla_get_u32(tb[IWL_TM_ATTR_COMMAND]) ==
-	    IWL_TM_CMD_APP2DEV_BEGIN_TRACE)
-		iwl_test_trace_stop(tst);
-	return -EMSGSIZE;
-}
-
-/*
- * Handles indirect read from the periphery or the SRAM. The read is performed
- * to a temporary buffer. The user space application should later issue a dump
- */
-static int iwl_test_indirect_read(struct iwl_test *tst, u32 addr, u32 size)
-{
-	struct iwl_trans *trans = tst->trans;
-	unsigned long flags;
-	int i;
-
-	if (size & 0x3)
-		return -EINVAL;
-
-	tst->mem.size = size;
-	tst->mem.addr = kmalloc(tst->mem.size, GFP_KERNEL);
-	if (tst->mem.addr == NULL)
-		return -ENOMEM;
-
-	/* Hard-coded periphery absolute address */
-	if (IWL_ABS_PRPH_START <= addr &&
-	    addr < IWL_ABS_PRPH_START + PRPH_END) {
-			if (!iwl_trans_grab_nic_access(trans, false, &flags)) {
-				return -EIO;
-			}
-			iwl_write32(trans, HBUS_TARG_PRPH_RADDR,
-				    addr | (3 << 24));
-			for (i = 0; i < size; i += 4)
-				*(u32 *)(tst->mem.addr + i) =
-					iwl_read32(trans, HBUS_TARG_PRPH_RDAT);
-			iwl_trans_release_nic_access(trans, &flags);
-	} else { /* target memory (SRAM) */
-		iwl_trans_read_mem(trans, addr, tst->mem.addr,
-				   tst->mem.size / 4);
-	}
-
-	tst->mem.nchunks =
-		DIV_ROUND_UP(tst->mem.size, DUMP_CHUNK_SIZE);
-	tst->mem.in_read = true;
-	return 0;
-
-}
-
-/*
- * Handles indirect write to the periphery or SRAM. The  is performed to a
- * temporary buffer.
- */
-static int iwl_test_indirect_write(struct iwl_test *tst, u32 addr,
-	u32 size, unsigned char *buf)
-{
-	struct iwl_trans *trans = tst->trans;
-	u32 val, i;
-	unsigned long flags;
-
-	if (IWL_ABS_PRPH_START <= addr &&
-	    addr < IWL_ABS_PRPH_START + PRPH_END) {
-		/* Periphery writes can be 1-3 bytes long, or DWORDs */
-		if (size < 4) {
-			memcpy(&val, buf, size);
-			if (!iwl_trans_grab_nic_access(trans, false, &flags))
-					return -EIO;
-			iwl_write32(trans, HBUS_TARG_PRPH_WADDR,
-				    (addr & 0x0000FFFF) |
-				    ((size - 1) << 24));
-			iwl_write32(trans, HBUS_TARG_PRPH_WDAT, val);
-			iwl_trans_release_nic_access(trans, &flags);
-		} else {
-			if (size % 4)
-				return -EINVAL;
-			for (i = 0; i < size; i += 4)
-				iwl_write_prph(trans, addr+i,
-					       *(u32 *)(buf+i));
-		}
-	} else if (iwl_test_valid_hw_addr(tst, addr)) {
-		iwl_trans_write_mem(trans, addr, buf, size / 4);
-	} else {
-		return -EINVAL;
-	}
-	return 0;
-}
-
-/*
- * Handles the user application commands for indirect read/write
- * to/from the periphery or the SRAM.
- */
-static int iwl_test_indirect_mem(struct iwl_test *tst, struct nlattr **tb)
-{
-	u32 addr, size, cmd;
-	unsigned char *buf;
-
-	/* Both read and write should be blocked, for atomicity */
-	if (tst->mem.in_read)
-		return -EBUSY;
-
-	cmd = nla_get_u32(tb[IWL_TM_ATTR_COMMAND]);
-	if (!tb[IWL_TM_ATTR_MEM_ADDR]) {
-		IWL_ERR(tst->trans, "Error finding memory offset address\n");
-		return -ENOMSG;
-	}
-	addr = nla_get_u32(tb[IWL_TM_ATTR_MEM_ADDR]);
-	if (!tb[IWL_TM_ATTR_BUFFER_SIZE]) {
-		IWL_ERR(tst->trans, "Error finding size for memory reading\n");
-		return -ENOMSG;
-	}
-	size = nla_get_u32(tb[IWL_TM_ATTR_BUFFER_SIZE]);
-
-	if (cmd == IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_READ) {
-		return iwl_test_indirect_read(tst, addr,  size);
-	} else {
-		if (!tb[IWL_TM_ATTR_BUFFER_DUMP])
-			return -EINVAL;
-		buf = (unsigned char *)nla_data(tb[IWL_TM_ATTR_BUFFER_DUMP]);
-		return iwl_test_indirect_write(tst, addr, size, buf);
-	}
-}
-
-/*
- * Enable notifications to user space
- */
-static int iwl_test_notifications(struct iwl_test *tst,
-				  struct nlattr **tb)
-{
-	tst->notify = nla_get_flag(tb[IWL_TM_ATTR_ENABLE_NOTIFICATION]);
-	return 0;
-}
-
-/*
- * Handles the request to get the device id
- */
-static int iwl_test_get_dev_id(struct iwl_test *tst, struct nlattr **tb)
-{
-	u32 devid = tst->trans->hw_id;
-	struct sk_buff *skb;
-	int status;
-
-	IWL_DEBUG_INFO(tst->trans, "hw version: 0x%x\n", devid);
-
-	skb = iwl_test_alloc_reply(tst, 20);
-	if (!skb) {
-		IWL_ERR(tst->trans, "Memory allocation fail\n");
-		return -ENOMEM;
-	}
-
-	if (nla_put_u32(skb, IWL_TM_ATTR_DEVICE_ID, devid))
-		goto nla_put_failure;
-	status = iwl_test_reply(tst, skb);
-	if (status < 0)
-		IWL_ERR(tst->trans, "Error sending msg : %d\n", status);
-
-	return 0;
-
-nla_put_failure:
-	kfree_skb(skb);
-	return -EMSGSIZE;
-}
-
-/*
- * Handles the request to get the FW version
- */
-static int iwl_test_get_fw_ver(struct iwl_test *tst, struct nlattr **tb)
-{
-	struct sk_buff *skb;
-	int status;
-	u32 ver = iwl_test_fw_ver(tst);
-
-	IWL_DEBUG_INFO(tst->trans, "uCode version raw: 0x%x\n", ver);
-
-	skb = iwl_test_alloc_reply(tst, 20);
-	if (!skb) {
-		IWL_ERR(tst->trans, "Memory allocation fail\n");
-		return -ENOMEM;
-	}
-
-	if (nla_put_u32(skb, IWL_TM_ATTR_FW_VERSION, ver))
-		goto nla_put_failure;
-
-	status = iwl_test_reply(tst, skb);
-	if (status < 0)
-		IWL_ERR(tst->trans, "Error sending msg : %d\n", status);
-
-	return 0;
-
-nla_put_failure:
-	kfree_skb(skb);
-	return -EMSGSIZE;
-}
-
-/*
- * Parse the netlink message and validate that the IWL_TM_ATTR_CMD exists
- */
-int iwl_test_parse(struct iwl_test *tst, struct nlattr **tb,
-		   void *data, int len)
-{
-	int result;
-
-	result = nla_parse(tb, IWL_TM_ATTR_MAX - 1, data, len,
-			iwl_testmode_gnl_msg_policy);
-	if (result) {
-		IWL_ERR(tst->trans, "Fail parse gnl msg: %d\n", result);
-		return result;
-	}
-
-	/* IWL_TM_ATTR_COMMAND is absolutely mandatory */
-	if (!tb[IWL_TM_ATTR_COMMAND]) {
-		IWL_ERR(tst->trans, "Missing testmode command type\n");
-		return -ENOMSG;
-	}
-	return 0;
-}
-IWL_EXPORT_SYMBOL(iwl_test_parse);
-
-/*
- * Handle test commands.
- * Returns 1 for unknown commands (not handled by the test object); negative
- * value in case of error.
- */
-int iwl_test_handle_cmd(struct iwl_test *tst, struct nlattr **tb)
-{
-	int result;
-
-	switch (nla_get_u32(tb[IWL_TM_ATTR_COMMAND])) {
-	case IWL_TM_CMD_APP2DEV_UCODE:
-		IWL_DEBUG_INFO(tst->trans, "test cmd to uCode\n");
-		result = iwl_test_fw_cmd(tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_READ32:
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE32:
-	case IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE8:
-		IWL_DEBUG_INFO(tst->trans, "test cmd to register\n");
-		result = iwl_test_reg(tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_BEGIN_TRACE:
-		IWL_DEBUG_INFO(tst->trans, "test uCode trace cmd to driver\n");
-		result = iwl_test_trace_begin(tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_END_TRACE:
-		iwl_test_trace_stop(tst);
-		result = 0;
-		break;
-
-	case IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_READ:
-	case IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_WRITE:
-		IWL_DEBUG_INFO(tst->trans, "test indirect memory cmd\n");
-		result = iwl_test_indirect_mem(tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_NOTIFICATIONS:
-		IWL_DEBUG_INFO(tst->trans, "test notifications cmd\n");
-		result = iwl_test_notifications(tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_GET_FW_VERSION:
-		IWL_DEBUG_INFO(tst->trans, "test get FW ver cmd\n");
-		result = iwl_test_get_fw_ver(tst, tb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_GET_DEVICE_ID:
-		IWL_DEBUG_INFO(tst->trans, "test Get device ID cmd\n");
-		result = iwl_test_get_dev_id(tst, tb);
-		break;
-
-	default:
-		IWL_DEBUG_INFO(tst->trans, "Unknown test command\n");
-		result = 1;
-		break;
-	}
-	return result;
-}
-IWL_EXPORT_SYMBOL(iwl_test_handle_cmd);
-
-static int iwl_test_trace_dump(struct iwl_test *tst, struct sk_buff *skb,
-			       struct netlink_callback *cb)
-{
-	int idx, length;
-
-	if (!tst->trace.enabled || !tst->trace.trace_addr)
-		return -EFAULT;
-
-	idx = cb->args[4];
-	if (idx >= tst->trace.nchunks)
-		return -ENOENT;
-
-	length = DUMP_CHUNK_SIZE;
-	if (((idx + 1) == tst->trace.nchunks) &&
-	    (tst->trace.size % DUMP_CHUNK_SIZE))
-		length = tst->trace.size %
-			DUMP_CHUNK_SIZE;
-
-	if (nla_put(skb, IWL_TM_ATTR_TRACE_DUMP, length,
-		    tst->trace.trace_addr + (DUMP_CHUNK_SIZE * idx)))
-		goto nla_put_failure;
-
-	cb->args[4] = ++idx;
-	return 0;
-
- nla_put_failure:
-	return -ENOBUFS;
-}
-
-static int iwl_test_buffer_dump(struct iwl_test *tst, struct sk_buff *skb,
-				struct netlink_callback *cb)
-{
-	int idx, length;
-
-	if (!tst->mem.in_read)
-		return -EFAULT;
-
-	idx = cb->args[4];
-	if (idx >= tst->mem.nchunks) {
-		iwl_test_mem_stop(tst);
-		return -ENOENT;
-	}
-
-	length = DUMP_CHUNK_SIZE;
-	if (((idx + 1) == tst->mem.nchunks) &&
-	    (tst->mem.size % DUMP_CHUNK_SIZE))
-		length = tst->mem.size % DUMP_CHUNK_SIZE;
-
-	if (nla_put(skb, IWL_TM_ATTR_BUFFER_DUMP, length,
-		    tst->mem.addr + (DUMP_CHUNK_SIZE * idx)))
-		goto nla_put_failure;
-
-	cb->args[4] = ++idx;
-	return 0;
-
- nla_put_failure:
-	return -ENOBUFS;
-}
-
-/*
- * Handle dump commands.
- * Returns 1 for unknown commands (not handled by the test object); negative
- * value in case of error.
- */
-int iwl_test_dump(struct iwl_test *tst, u32 cmd, struct sk_buff *skb,
-		  struct netlink_callback *cb)
-{
-	int result;
-
-	switch (cmd) {
-	case IWL_TM_CMD_APP2DEV_READ_TRACE:
-		IWL_DEBUG_INFO(tst->trans, "uCode trace cmd\n");
-		result = iwl_test_trace_dump(tst, skb, cb);
-		break;
-
-	case IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_DUMP:
-		IWL_DEBUG_INFO(tst->trans, "testmode sram dump cmd\n");
-		result = iwl_test_buffer_dump(tst, skb, cb);
-		break;
-
-	default:
-		result = 1;
-		break;
-	}
-	return result;
-}
-IWL_EXPORT_SYMBOL(iwl_test_dump);
-
-/*
- * Multicast a spontaneous messages from the device to the user space.
- */
-static void iwl_test_send_rx(struct iwl_test *tst,
-			     struct iwl_rx_cmd_buffer *rxb)
-{
-	struct sk_buff *skb;
-	struct iwl_rx_packet *data;
-	int length;
-
-	data = rxb_addr(rxb);
-	length = le32_to_cpu(data->len_n_flags) & FH_RSCSR_FRAME_SIZE_MSK;
-
-	/* the length doesn't include len_n_flags field, so add it manually */
-	length += sizeof(__le32);
-
-	skb = iwl_test_alloc_event(tst, length + 20);
-	if (skb == NULL) {
-		IWL_ERR(tst->trans, "Out of memory for message to user\n");
-		return;
-	}
-
-	if (nla_put_u32(skb, IWL_TM_ATTR_COMMAND,
-			IWL_TM_CMD_DEV2APP_UCODE_RX_PKT) ||
-	    nla_put(skb, IWL_TM_ATTR_UCODE_RX_PKT, length, data))
-		goto nla_put_failure;
-
-	iwl_test_event(tst, skb);
-	return;
-
-nla_put_failure:
-	kfree_skb(skb);
-	IWL_ERR(tst->trans, "Ouch, overran buffer, check allocation!\n");
-}
-
-/*
- * Called whenever a Rx frames is recevied from the device. If notifications to
- * the user space are requested, sends the frames to the user.
- */
-void iwl_test_rx(struct iwl_test *tst, struct iwl_rx_cmd_buffer *rxb)
-{
-	if (tst->notify)
-		iwl_test_send_rx(tst, rxb);
-}
-IWL_EXPORT_SYMBOL(iwl_test_rx);
diff --git a/drivers/net/wireless/iwlwifi/iwl-test.h b/drivers/net/wireless/iwlwifi/iwl-test.h
deleted file mode 100644
index 8fbd217..0000000
--- a/drivers/net/wireless/iwlwifi/iwl-test.h
+++ /dev/null
@@ -1,161 +0,0 @@
-/******************************************************************************
- *
- * This file is provided under a dual BSD/GPLv2 license.  When using or
- * redistributing this file, you may do so under either license.
- *
- * GPL LICENSE SUMMARY
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110,
- * USA
- *
- * The full GNU General Public License is included in this distribution
- * in the file called COPYING.
- *
- * Contact Information:
- *  Intel Linux Wireless <ilw@linux.intel.com>
- * Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
- *
- * BSD LICENSE
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- *  * Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- *  * Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *  * Neither the name Intel Corporation nor the names of its
- *    contributors may be used to endorse or promote products derived
- *    from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- *****************************************************************************/
-
-#ifndef __IWL_TEST_H__
-#define __IWL_TEST_H__
-
-#include <linux/types.h>
-#include "iwl-trans.h"
-
-struct iwl_test_trace {
-	u32 size;
-	u32 tsize;
-	u32 nchunks;
-	u8 *cpu_addr;
-	u8 *trace_addr;
-	dma_addr_t dma_addr;
-	bool enabled;
-};
-
-struct iwl_test_mem {
-	u32 size;
-	u32 nchunks;
-	u8 *addr;
-	bool in_read;
-};
-
-/*
- * struct iwl_test_ops: callback to the op mode
- *
- * The structure defines the callbacks that the op_mode should handle,
- * inorder to handle logic that is out of the scope of iwl_test. The
- * op_mode must set all the callbacks.
-
- * @send_cmd: handler that is used by the test object to request the
- *  op_mode to send a command to the fw.
- *
- * @valid_hw_addr: handler that is used by the test object to request the
- *  op_mode to check if the given address is a valid address.
- *
- * @get_fw_ver: handler used to get the FW version.
- *
- * @alloc_reply: handler used by the test object to request the op_mode
- *  to allocate an skb for sending a reply to the user, and initialize
- *  the skb. It is assumed that the test object only fills the required
- *  attributes.
- *
- * @reply: handler used by the test object to request the op_mode to reply
- *  to a request. The skb is an skb previously allocated by the the
- *  alloc_reply callback.
- I
- * @alloc_event: handler used by the test object to request the op_mode
- *  to allocate an skb for sending an event, and initialize
- *  the skb. It is assumed that the test object only fills the required
- *  attributes.
- *
- * @reply: handler used by the test object to request the op_mode to send
- *  an event. The skb is an skb previously allocated by the the
- *  alloc_event callback.
- */
-struct iwl_test_ops {
-	int (*send_cmd)(struct iwl_op_mode *op_modes,
-			struct iwl_host_cmd *cmd);
-	bool (*valid_hw_addr)(u32 addr);
-	u32 (*get_fw_ver)(struct iwl_op_mode *op_mode);
-
-	struct sk_buff *(*alloc_reply)(struct iwl_op_mode *op_mode, int len);
-	int (*reply)(struct iwl_op_mode *op_mode, struct sk_buff *skb);
-	struct sk_buff* (*alloc_event)(struct iwl_op_mode *op_mode, int len);
-	void (*event)(struct iwl_op_mode *op_mode, struct sk_buff *skb);
-};
-
-struct iwl_test {
-	struct iwl_trans *trans;
-	struct iwl_test_ops *ops;
-	struct iwl_test_trace trace;
-	struct iwl_test_mem mem;
-	bool notify;
-};
-
-void iwl_test_init(struct iwl_test *tst, struct iwl_trans *trans,
-		   struct iwl_test_ops *ops);
-
-void iwl_test_free(struct iwl_test *tst);
-
-int iwl_test_parse(struct iwl_test *tst, struct nlattr **tb,
-		   void *data, int len);
-
-int iwl_test_handle_cmd(struct iwl_test *tst, struct nlattr **tb);
-
-int iwl_test_dump(struct iwl_test *tst, u32 cmd, struct sk_buff *skb,
-		  struct netlink_callback *cb);
-
-void iwl_test_rx(struct iwl_test *tst, struct iwl_rx_cmd_buffer *rxb);
-
-static inline void iwl_test_enable_notifications(struct iwl_test *tst,
-						 bool enable)
-{
-	tst->notify = enable;
-}
-
-#endif
diff --git a/drivers/net/wireless/iwlwifi/iwl-testmode.h b/drivers/net/wireless/iwlwifi/iwl-testmode.h
deleted file mode 100644
index 98f48a9..0000000
--- a/drivers/net/wireless/iwlwifi/iwl-testmode.h
+++ /dev/null
@@ -1,309 +0,0 @@
-/******************************************************************************
- *
- * This file is provided under a dual BSD/GPLv2 license.  When using or
- * redistributing this file, you may do so under either license.
- *
- * GPL LICENSE SUMMARY
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110,
- * USA
- *
- * The full GNU General Public License is included in this distribution
- * in the file called COPYING.
- *
- * Contact Information:
- *  Intel Linux Wireless <ilw@linux.intel.com>
- * Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
- *
- * BSD LICENSE
- *
- * Copyright(c) 2010 - 2013 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- *  * Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- *  * Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in
- *    the documentation and/or other materials provided with the
- *    distribution.
- *  * Neither the name Intel Corporation nor the names of its
- *    contributors may be used to endorse or promote products derived
- *    from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
- *****************************************************************************/
-#ifndef __IWL_TESTMODE_H__
-#define __IWL_TESTMODE_H__
-
-#include <linux/types.h>
-
-
-/*
- * Commands from user space to kernel space(IWL_TM_CMD_ID_APP2DEV_XX) and
- * from and kernel space to user space(IWL_TM_CMD_ID_DEV2APP_XX).
- * The command ID is carried with IWL_TM_ATTR_COMMAND.
- *
- * @IWL_TM_CMD_APP2DEV_UCODE:
- *	commands from user application to the uCode,
- *	the actual uCode host command ID is carried with
- *	IWL_TM_ATTR_UCODE_CMD_ID
- *
- * @IWL_TM_CMD_APP2DEV_DIRECT_REG_READ32:
- * @IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE32:
- * @IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE8:
- *	commands from user applicaiton to access register
- *
- * @IWL_TM_CMD_APP2DEV_GET_DEVICENAME: retrieve device name
- * @IWL_TM_CMD_APP2DEV_LOAD_INIT_FW: load initial uCode image
- * @IWL_TM_CMD_APP2DEV_CFG_INIT_CALIB: perform calibration
- * @IWL_TM_CMD_APP2DEV_LOAD_RUNTIME_FW: load runtime uCode image
- * @IWL_TM_CMD_APP2DEV_GET_EEPROM: request EEPROM data
- * @IWL_TM_CMD_APP2DEV_FIXRATE_REQ: set fix MCS
- *	commands fom user space for pure driver level operations
- *
- * @IWL_TM_CMD_APP2DEV_BEGIN_TRACE:
- * @IWL_TM_CMD_APP2DEV_END_TRACE:
- * @IWL_TM_CMD_APP2DEV_READ_TRACE:
- *	commands fom user space for uCode trace operations
- *
- * @IWL_TM_CMD_DEV2APP_SYNC_RSP:
- *	commands from kernel space to carry the synchronous response
- *	to user application
- * @IWL_TM_CMD_DEV2APP_UCODE_RX_PKT:
- *	commands from kernel space to multicast the spontaneous messages
- *	to user application, or reply of host commands
- * @IWL_TM_CMD_DEV2APP_EEPROM_RSP:
- *	commands from kernel space to carry the eeprom response
- *	to user application
- *
- * @IWL_TM_CMD_APP2DEV_OWNERSHIP:
- *	commands from user application to own change the ownership of the uCode
- *	if application has the ownership, the only host command from
- *	testmode will deliver to uCode. Default owner is driver
- *
- * @IWL_TM_CMD_APP2DEV_LOAD_WOWLAN_FW: load Wake On Wireless LAN uCode image
- * @IWL_TM_CMD_APP2DEV_GET_FW_VERSION: retrieve uCode version
- * @IWL_TM_CMD_APP2DEV_GET_DEVICE_ID: retrieve ID information in device
- * @IWL_TM_CMD_APP2DEV_GET_FW_INFO:
- *	retrieve information of existing loaded uCode image
- *
- * @IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_READ:
- * @IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_DUMP:
- * @IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_WRITE:
- *	Commands to read/write data from periphery or SRAM memory ranges.
- *	Fore reading, a READ command is sent from the userspace and the data
- *	is returned when the user calls a DUMP command.
- *	For writing, only a WRITE command is used.
- * @IWL_TM_CMD_APP2DEV_NOTIFICATIONS:
- *	Command to enable/disable notifications (currently RX packets) from the
- *	driver to userspace.
- */
-enum iwl_tm_cmd_t {
-	IWL_TM_CMD_APP2DEV_UCODE		= 1,
-	IWL_TM_CMD_APP2DEV_DIRECT_REG_READ32	= 2,
-	IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE32	= 3,
-	IWL_TM_CMD_APP2DEV_DIRECT_REG_WRITE8	= 4,
-	IWL_TM_CMD_APP2DEV_GET_DEVICENAME	= 5,
-	IWL_TM_CMD_APP2DEV_LOAD_INIT_FW		= 6,
-	IWL_TM_CMD_APP2DEV_CFG_INIT_CALIB	= 7,
-	IWL_TM_CMD_APP2DEV_LOAD_RUNTIME_FW	= 8,
-	IWL_TM_CMD_APP2DEV_GET_EEPROM		= 9,
-	IWL_TM_CMD_APP2DEV_FIXRATE_REQ		= 10,
-	IWL_TM_CMD_APP2DEV_BEGIN_TRACE		= 11,
-	IWL_TM_CMD_APP2DEV_END_TRACE		= 12,
-	IWL_TM_CMD_APP2DEV_READ_TRACE		= 13,
-	IWL_TM_CMD_DEV2APP_SYNC_RSP		= 14,
-	IWL_TM_CMD_DEV2APP_UCODE_RX_PKT		= 15,
-	IWL_TM_CMD_DEV2APP_EEPROM_RSP		= 16,
-	IWL_TM_CMD_APP2DEV_OWNERSHIP		= 17,
-	RESERVED_18				= 18,
-	RESERVED_19				= 19,
-	RESERVED_20				= 20,
-	RESERVED_21				= 21,
-	IWL_TM_CMD_APP2DEV_LOAD_WOWLAN_FW	= 22,
-	IWL_TM_CMD_APP2DEV_GET_FW_VERSION	= 23,
-	IWL_TM_CMD_APP2DEV_GET_DEVICE_ID	= 24,
-	IWL_TM_CMD_APP2DEV_GET_FW_INFO		= 25,
-	IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_READ = 26,
-	IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_DUMP = 27,
-	IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_WRITE = 28,
-	IWL_TM_CMD_APP2DEV_NOTIFICATIONS	= 29,
-	IWL_TM_CMD_MAX				= 30,
-};
-
-/*
- * Atrribute filed in testmode command
- * See enum iwl_tm_cmd_t.
- *
- * @IWL_TM_ATTR_NOT_APPLICABLE:
- *	The attribute is not applicable or invalid
- * @IWL_TM_ATTR_COMMAND:
- *	From user space to kernel space:
- *	the command either destines to ucode, driver, or register;
- *	From kernel space to user space:
- *	the command either carries synchronous response,
- *	or the spontaneous message multicast from the device;
- *
- * @IWL_TM_ATTR_UCODE_CMD_ID:
- * @IWL_TM_ATTR_UCODE_CMD_DATA:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_UCODE,
- *	The mandatory fields are :
- *	IWL_TM_ATTR_UCODE_CMD_ID for recognizable command ID;
- *	IWL_TM_ATTR_UCODE_CMD_DATA for the actual command payload
- *	to the ucode
- *
- * @IWL_TM_ATTR_REG_OFFSET:
- * @IWL_TM_ATTR_REG_VALUE8:
- * @IWL_TM_ATTR_REG_VALUE32:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_REG_XXX,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_REG_OFFSET for the offset of the target register;
- *	IWL_TM_ATTR_REG_VALUE8 or IWL_TM_ATTR_REG_VALUE32 for value
- *
- * @IWL_TM_ATTR_SYNC_RSP:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_DEV2APP_SYNC_RSP,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_SYNC_RSP for the data content responding to the user
- *	application command
- *
- * @IWL_TM_ATTR_UCODE_RX_PKT:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_DEV2APP_UCODE_RX_PKT,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_UCODE_RX_PKT for the data content multicast to the user
- *	application
- *
- * @IWL_TM_ATTR_EEPROM:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_DEV2APP_EEPROM,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_EEPROM for the data content responging to the user
- *	application
- *
- * @IWL_TM_ATTR_TRACE_ADDR:
- * @IWL_TM_ATTR_TRACE_SIZE:
- * @IWL_TM_ATTR_TRACE_DUMP:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_XXX_TRACE,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_MEM_TRACE_ADDR for the trace address
- *	IWL_TM_ATTR_MEM_TRACE_SIZE for the trace buffer size
- *	IWL_TM_ATTR_MEM_TRACE_DUMP for the trace dump
- *
- * @IWL_TM_ATTR_FIXRATE:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_FIXRATE_REQ,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_FIXRATE for the fixed rate
- *
- * @IWL_TM_ATTR_UCODE_OWNER:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_OWNERSHIP,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_UCODE_OWNER for the new owner
- *
- * @IWL_TM_ATTR_MEM_ADDR:
- * @IWL_TM_ATTR_BUFFER_SIZE:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_READ
- *	or IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_WRITE.
- *	The mandatory fields are:
- *	IWL_TM_ATTR_MEM_ADDR for the address in SRAM/periphery to read/write
- *	IWL_TM_ATTR_BUFFER_SIZE for the buffer size of data to read/write.
- *
- * @IWL_TM_ATTR_BUFFER_DUMP:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_DUMP,
- *	IWL_TM_ATTR_BUFFER_DUMP is used for the data that was read.
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_INDIRECT_BUFFER_WRITE,
- *	this attribute contains the data to write.
- *
- * @IWL_TM_ATTR_FW_VERSION:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_GET_FW_VERSION,
- *	IWL_TM_ATTR_FW_VERSION for the uCode version
- *
- * @IWL_TM_ATTR_DEVICE_ID:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_GET_DEVICE_ID,
- *	IWL_TM_ATTR_DEVICE_ID for the device ID information
- *
- * @IWL_TM_ATTR_FW_TYPE:
- * @IWL_TM_ATTR_FW_INST_SIZE:
- * @IWL_TM_ATTR_FW_DATA_SIZE:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_GET_FW_INFO,
- *	The mandatory fields are:
- *	IWL_TM_ATTR_FW_TYPE for the uCode type (INIT/RUNTIME/...)
- *	IWL_TM_ATTR_FW_INST_SIZE for the size of instruction section
- *	IWL_TM_ATTR_FW_DATA_SIZE for the size of data section
- *
- * @IWL_TM_ATTR_UCODE_CMD_SKB:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_UCODE this flag
- *	indicates that the user wants to receive the response of the command
- *	in a reply SKB. If it's not present, the response is not returned.
- * @IWL_TM_ATTR_ENABLE_NOTIFICATIONS:
- *	When IWL_TM_ATTR_COMMAND is IWL_TM_CMD_APP2DEV_NOTIFICATIONS, this
- *	flag enables (if present) or disables (if not) the forwarding
- *	to userspace.
- */
-enum iwl_tm_attr_t {
-	IWL_TM_ATTR_NOT_APPLICABLE		= 0,
-	IWL_TM_ATTR_COMMAND			= 1,
-	IWL_TM_ATTR_UCODE_CMD_ID		= 2,
-	IWL_TM_ATTR_UCODE_CMD_DATA		= 3,
-	IWL_TM_ATTR_REG_OFFSET			= 4,
-	IWL_TM_ATTR_REG_VALUE8			= 5,
-	IWL_TM_ATTR_REG_VALUE32			= 6,
-	IWL_TM_ATTR_SYNC_RSP			= 7,
-	IWL_TM_ATTR_UCODE_RX_PKT		= 8,
-	IWL_TM_ATTR_EEPROM			= 9,
-	IWL_TM_ATTR_TRACE_ADDR			= 10,
-	IWL_TM_ATTR_TRACE_SIZE			= 11,
-	IWL_TM_ATTR_TRACE_DUMP			= 12,
-	IWL_TM_ATTR_FIXRATE			= 13,
-	IWL_TM_ATTR_UCODE_OWNER			= 14,
-	IWL_TM_ATTR_MEM_ADDR			= 15,
-	IWL_TM_ATTR_BUFFER_SIZE			= 16,
-	IWL_TM_ATTR_BUFFER_DUMP			= 17,
-	IWL_TM_ATTR_FW_VERSION			= 18,
-	IWL_TM_ATTR_DEVICE_ID			= 19,
-	IWL_TM_ATTR_FW_TYPE			= 20,
-	IWL_TM_ATTR_FW_INST_SIZE		= 21,
-	IWL_TM_ATTR_FW_DATA_SIZE		= 22,
-	IWL_TM_ATTR_UCODE_CMD_SKB		= 23,
-	IWL_TM_ATTR_ENABLE_NOTIFICATION		= 24,
-	IWL_TM_ATTR_MAX				= 25,
-};
-
-/* uCode trace buffer */
-#define TRACE_BUFF_SIZE_MAX	0x200000
-#define TRACE_BUFF_SIZE_MIN	0x20000
-#define TRACE_BUFF_SIZE_DEF	TRACE_BUFF_SIZE_MIN
-#define TRACE_BUFF_PADD		0x2000
-
-/* Maximum data size of each dump it packet */
-#define DUMP_CHUNK_SIZE		(PAGE_SIZE - 1024)
-
-/* Address offset of data segment in SRAM */
-#define SRAM_DATA_SEG_OFFSET   0x800000
-
-#endif
diff --git a/drivers/net/wireless/iwlwifi/iwl-trans.h b/drivers/net/wireless/iwlwifi/iwl-trans.h
index 7a13790..8d91422c 100644
--- a/drivers/net/wireless/iwlwifi/iwl-trans.h
+++ b/drivers/net/wireless/iwlwifi/iwl-trans.h
@@ -183,13 +183,12 @@ struct iwl_rx_packet {
  * @CMD_ASYNC: Return right away and don't want for the response
  * @CMD_WANT_SKB: valid only with CMD_SYNC. The caller needs the buffer of the
  *	response. The caller needs to call iwl_free_resp when done.
- * @CMD_ON_DEMAND: This command is sent by the test mode pipe.
  */
 enum CMD_MODE {
 	CMD_SYNC		= 0,
 	CMD_ASYNC		= BIT(0),
 	CMD_WANT_SKB		= BIT(1),
-	CMD_ON_DEMAND		= BIT(2),
+	CMD_SEND_IN_RFKILL	= BIT(2),
 };
 
 #define DEF_CMD_PAYLOAD_SIZE 320
@@ -427,8 +426,9 @@ struct iwl_trans_ops {
 	void (*fw_alive)(struct iwl_trans *trans, u32 scd_addr);
 	void (*stop_device)(struct iwl_trans *trans);
 
-	void (*d3_suspend)(struct iwl_trans *trans);
-	int (*d3_resume)(struct iwl_trans *trans, enum iwl_d3_status *status);
+	void (*d3_suspend)(struct iwl_trans *trans, bool test);
+	int (*d3_resume)(struct iwl_trans *trans, enum iwl_d3_status *status,
+			 bool test);
 
 	int (*send_cmd)(struct iwl_trans *trans, struct iwl_host_cmd *cmd);
 
@@ -455,7 +455,7 @@ struct iwl_trans_ops {
 	int (*read_mem)(struct iwl_trans *trans, u32 addr,
 			void *buf, int dwords);
 	int (*write_mem)(struct iwl_trans *trans, u32 addr,
-			 void *buf, int dwords);
+			 const void *buf, int dwords);
 	void (*configure)(struct iwl_trans *trans,
 			  const struct iwl_trans_config *trans_cfg);
 	void (*set_pmi)(struct iwl_trans *trans, bool state);
@@ -587,17 +587,18 @@ static inline void iwl_trans_stop_device(struct iwl_trans *trans)
 	trans->state = IWL_TRANS_NO_FW;
 }
 
-static inline void iwl_trans_d3_suspend(struct iwl_trans *trans)
+static inline void iwl_trans_d3_suspend(struct iwl_trans *trans, bool test)
 {
 	might_sleep();
-	trans->ops->d3_suspend(trans);
+	trans->ops->d3_suspend(trans, test);
 }
 
 static inline int iwl_trans_d3_resume(struct iwl_trans *trans,
-				      enum iwl_d3_status *status)
+				      enum iwl_d3_status *status,
+				      bool test)
 {
 	might_sleep();
-	return trans->ops->d3_resume(trans, status);
+	return trans->ops->d3_resume(trans, status, test);
 }
 
 static inline int iwl_trans_send_cmd(struct iwl_trans *trans,
@@ -761,7 +762,7 @@ static inline u32 iwl_trans_read_mem32(struct iwl_trans *trans, u32 addr)
 }
 
 static inline int iwl_trans_write_mem(struct iwl_trans *trans, u32 addr,
-				      void *buf, int dwords)
+				      const void *buf, int dwords)
 {
 	return trans->ops->write_mem(trans, addr, buf, dwords);
 }
diff --git a/drivers/net/wireless/iwlwifi/mvm/Makefile b/drivers/net/wireless/iwlwifi/mvm/Makefile
index 2acc44b..ff856e5 100644
--- a/drivers/net/wireless/iwlwifi/mvm/Makefile
+++ b/drivers/net/wireless/iwlwifi/mvm/Makefile
@@ -3,7 +3,7 @@ iwlmvm-y += fw.o mac80211.o nvm.o ops.o phy-ctxt.o mac-ctxt.o
 iwlmvm-y += utils.o rx.o tx.o binding.o quota.o sta.o
 iwlmvm-y += scan.o time-event.o rs.o
 iwlmvm-y += power.o bt-coex.o
-iwlmvm-y += led.o
+iwlmvm-y += led.o tt.o
 iwlmvm-$(CONFIG_IWLWIFI_DEBUGFS) += debugfs.o
 iwlmvm-$(CONFIG_PM_SLEEP) += d3.o
 
diff --git a/drivers/net/wireless/iwlwifi/mvm/bt-coex.c b/drivers/net/wireless/iwlwifi/mvm/bt-coex.c
index 810bfa5..dbd622a 100644
--- a/drivers/net/wireless/iwlwifi/mvm/bt-coex.c
+++ b/drivers/net/wireless/iwlwifi/mvm/bt-coex.c
@@ -174,7 +174,7 @@ static const __le32 iwl_tight_lookup[BT_COEX_LUT_SIZE] = {
 static const __le32 iwl_loose_lookup[BT_COEX_LUT_SIZE] = {
 	cpu_to_le32(0xaaaaaaaa),
 	cpu_to_le32(0xaaaaaaaa),
-	cpu_to_le32(0xaeaaaaaa),
+	cpu_to_le32(0xaaaaaaaa),
 	cpu_to_le32(0xaaaaaaaa),
 	cpu_to_le32(0xcc00ff28),
 	cpu_to_le32(0x0000aaaa),
@@ -202,6 +202,22 @@ static const __le32 iwl_concurrent_lookup[BT_COEX_LUT_SIZE] = {
 	cpu_to_le32(0x00000000),
 };
 
+/* single shared antenna */
+static const __le32 iwl_single_shared_ant_lookup[BT_COEX_LUT_SIZE] = {
+	cpu_to_le32(0x40000000),
+	cpu_to_le32(0x00000000),
+	cpu_to_le32(0x44000000),
+	cpu_to_le32(0x00000000),
+	cpu_to_le32(0x40000000),
+	cpu_to_le32(0x00000000),
+	cpu_to_le32(0x44000000),
+	cpu_to_le32(0x00000000),
+	cpu_to_le32(0xC0004000),
+	cpu_to_le32(0xF0005000),
+	cpu_to_le32(0xC0004000),
+	cpu_to_le32(0xF0005000),
+};
+
 int iwl_send_bt_init_conf(struct iwl_mvm *mvm)
 {
 	struct iwl_bt_coex_cmd cmd = {
@@ -225,7 +241,10 @@ int iwl_send_bt_init_conf(struct iwl_mvm *mvm)
 					BT_VALID_REDUCED_TX_POWER |
 					BT_VALID_LUT);
 
-	if (is_loose_coex())
+	if (mvm->cfg->bt_shared_single_ant)
+		memcpy(&cmd.decision_lut, iwl_single_shared_ant_lookup,
+		       sizeof(iwl_single_shared_ant_lookup));
+	else if (is_loose_coex())
 		memcpy(&cmd.decision_lut, iwl_loose_lookup,
 		       sizeof(iwl_tight_lookup));
 	else
@@ -351,6 +370,7 @@ static void iwl_mvm_bt_notif_iterator(void *_data, u8 *mac,
 	enum ieee80211_band band;
 	int ave_rssi;
 
+	lockdep_assert_held(&mvm->mutex);
 	if (vif->type != NL80211_IFTYPE_STATION)
 		return;
 
@@ -365,7 +385,8 @@ static void iwl_mvm_bt_notif_iterator(void *_data, u8 *mac,
 	smps_mode = IEEE80211_SMPS_AUTOMATIC;
 
 	if (band != IEEE80211_BAND_2GHZ) {
-		ieee80211_request_smps(vif, smps_mode);
+		iwl_mvm_update_smps(mvm, vif, IWL_MVM_SMPS_REQ_BT_COEX,
+				    smps_mode);
 		return;
 	}
 
@@ -380,7 +401,7 @@ static void iwl_mvm_bt_notif_iterator(void *_data, u8 *mac,
 		       mvmvif->id,  data->notif->bt_status,
 		       data->notif->bt_traffic_load, smps_mode);
 
-	ieee80211_request_smps(vif, smps_mode);
+	iwl_mvm_update_smps(mvm, vif, IWL_MVM_SMPS_REQ_BT_COEX, smps_mode);
 
 	/* don't reduce the Tx power if in loose scheme */
 	if (is_loose_coex())
diff --git a/drivers/net/wireless/iwlwifi/mvm/d3.c b/drivers/net/wireless/iwlwifi/mvm/d3.c
index 16bbdcc..7e5e5c2 100644
--- a/drivers/net/wireless/iwlwifi/mvm/d3.c
+++ b/drivers/net/wireless/iwlwifi/mvm/d3.c
@@ -63,6 +63,7 @@
 
 #include <linux/etherdevice.h>
 #include <linux/ip.h>
+#include <linux/fs.h>
 #include <net/cfg80211.h>
 #include <net/ipv6.h>
 #include <net/tcp.h>
@@ -419,8 +420,7 @@ static __le16 pseudo_hdr_check(int len, __be32 saddr, __be32 daddr)
 	return cpu_to_le16(be16_to_cpu((__force __be16)check));
 }
 
-static void iwl_mvm_build_tcp_packet(struct iwl_mvm *mvm,
-				     struct ieee80211_vif *vif,
+static void iwl_mvm_build_tcp_packet(struct ieee80211_vif *vif,
 				     struct cfg80211_wowlan_tcp *tcp,
 				     void *_pkt, u8 *mask,
 				     __le16 *pseudo_hdr_csum,
@@ -566,21 +566,21 @@ static int iwl_mvm_send_remote_wake_cfg(struct iwl_mvm *mvm,
 
 	/* SYN (TX) */
 	iwl_mvm_build_tcp_packet(
-		mvm, vif, tcp, cfg->syn_tx.data, NULL,
+		vif, tcp, cfg->syn_tx.data, NULL,
 		&cfg->syn_tx.info.tcp_pseudo_header_checksum,
 		MVM_TCP_TX_SYN);
 	cfg->syn_tx.info.tcp_payload_length = 0;
 
 	/* SYN/ACK (RX) */
 	iwl_mvm_build_tcp_packet(
-		mvm, vif, tcp, cfg->synack_rx.data, cfg->synack_rx.rx_mask,
+		vif, tcp, cfg->synack_rx.data, cfg->synack_rx.rx_mask,
 		&cfg->synack_rx.info.tcp_pseudo_header_checksum,
 		MVM_TCP_RX_SYNACK);
 	cfg->synack_rx.info.tcp_payload_length = 0;
 
 	/* KEEPALIVE/ACK (TX) */
 	iwl_mvm_build_tcp_packet(
-		mvm, vif, tcp, cfg->keepalive_tx.data, NULL,
+		vif, tcp, cfg->keepalive_tx.data, NULL,
 		&cfg->keepalive_tx.info.tcp_pseudo_header_checksum,
 		MVM_TCP_TX_DATA);
 	cfg->keepalive_tx.info.tcp_payload_length =
@@ -604,7 +604,7 @@ static int iwl_mvm_send_remote_wake_cfg(struct iwl_mvm *mvm,
 
 	/* ACK (RX) */
 	iwl_mvm_build_tcp_packet(
-		mvm, vif, tcp, cfg->keepalive_ack_rx.data,
+		vif, tcp, cfg->keepalive_ack_rx.data,
 		cfg->keepalive_ack_rx.rx_mask,
 		&cfg->keepalive_ack_rx.info.tcp_pseudo_header_checksum,
 		MVM_TCP_RX_ACK);
@@ -612,7 +612,7 @@ static int iwl_mvm_send_remote_wake_cfg(struct iwl_mvm *mvm,
 
 	/* WAKEUP (RX) */
 	iwl_mvm_build_tcp_packet(
-		mvm, vif, tcp, cfg->wake_rx.data, cfg->wake_rx.rx_mask,
+		vif, tcp, cfg->wake_rx.data, cfg->wake_rx.rx_mask,
 		&cfg->wake_rx.info.tcp_pseudo_header_checksum,
 		MVM_TCP_RX_WAKE);
 	cfg->wake_rx.info.tcp_payload_length =
@@ -620,7 +620,7 @@ static int iwl_mvm_send_remote_wake_cfg(struct iwl_mvm *mvm,
 
 	/* FIN */
 	iwl_mvm_build_tcp_packet(
-		mvm, vif, tcp, cfg->fin_tx.data, NULL,
+		vif, tcp, cfg->fin_tx.data, NULL,
 		&cfg->fin_tx.info.tcp_pseudo_header_checksum,
 		MVM_TCP_TX_FIN);
 	cfg->fin_tx.info.tcp_payload_length = 0;
@@ -756,7 +756,9 @@ static int iwl_mvm_d3_reprogram(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 	return 0;
 }
 
-int iwl_mvm_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
+static int __iwl_mvm_suspend(struct ieee80211_hw *hw,
+			     struct cfg80211_wowlan *wowlan,
+			     bool test)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
 	struct iwl_d3_iter_data suspend_iter_data = {
@@ -769,7 +771,7 @@ int iwl_mvm_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
 	struct iwl_wowlan_config_cmd wowlan_config_cmd = {};
 	struct iwl_wowlan_kek_kck_material_cmd kek_kck_cmd = {};
 	struct iwl_wowlan_tkip_params_cmd tkip_cmd = {};
-	struct iwl_d3_manager_config d3_cfg_cmd = {
+	struct iwl_d3_manager_config d3_cfg_cmd_data = {
 		/*
 		 * Program the minimum sleep time to 10 seconds, as many
 		 * platforms have issues processing a wakeup signal while
@@ -777,17 +779,30 @@ int iwl_mvm_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
 		 */
 		.min_sleep_time = cpu_to_le32(10 * 1000 * 1000),
 	};
+	struct iwl_host_cmd d3_cfg_cmd = {
+		.id = D3_CONFIG_CMD,
+		.flags = CMD_SYNC | CMD_WANT_SKB,
+		.data[0] = &d3_cfg_cmd_data,
+		.len[0] = sizeof(d3_cfg_cmd_data),
+	};
 	struct wowlan_key_data key_data = {
 		.use_rsc_tsc = false,
 		.tkip = &tkip_cmd,
 		.use_tkip = false,
 	};
 	int ret, i;
+	int len __maybe_unused;
 	u16 seq;
 	u8 old_aux_sta_id, old_ap_sta_id = IWL_MVM_STATION_COUNT;
 
-	if (WARN_ON(!wowlan))
+	if (!wowlan) {
+		/*
+		 * mac80211 shouldn't get here, but for D3 test
+		 * it doesn't warrant a warning
+		 */
+		WARN_ON(!test);
 		return -EINVAL;
+	}
 
 	key_data.rsc_tsc = kzalloc(sizeof(*key_data.rsc_tsc), GFP_KERNEL);
 	if (!key_data.rsc_tsc)
@@ -1007,15 +1022,37 @@ int iwl_mvm_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
 	if (ret)
 		goto out;
 
+	ret = iwl_mvm_power_update_mode(mvm, vif);
+	if (ret)
+		goto out;
+
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	if (mvm->d3_wake_sysassert)
+		d3_cfg_cmd_data.wakeup_flags |=
+			cpu_to_le32(IWL_WAKEUP_D3_CONFIG_FW_ERROR);
+#endif
+
 	/* must be last -- this switches firmware state */
-	ret = iwl_mvm_send_cmd_pdu(mvm, D3_CONFIG_CMD, CMD_SYNC,
-				   sizeof(d3_cfg_cmd), &d3_cfg_cmd);
+	ret = iwl_mvm_send_cmd(mvm, &d3_cfg_cmd);
 	if (ret)
 		goto out;
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	len = le32_to_cpu(d3_cfg_cmd.resp_pkt->len_n_flags) &
+		FH_RSCSR_FRAME_SIZE_MSK;
+	if (len >= sizeof(u32) * 2) {
+		mvm->d3_test_pme_ptr =
+			le32_to_cpup((__le32 *)d3_cfg_cmd.resp_pkt->data);
+	} else if (test) {
+		/* in test mode we require the pointer */
+		ret = -EIO;
+		goto out;
+	}
+#endif
+	iwl_free_resp(&d3_cfg_cmd);
 
 	clear_bit(IWL_MVM_STATUS_IN_HW_RESTART, &mvm->status);
 
-	iwl_trans_d3_suspend(mvm->trans);
+	iwl_trans_d3_suspend(mvm->trans, test);
  out:
 	mvm->aux_sta.sta_id = old_aux_sta_id;
 	mvm_ap_sta->sta_id = old_ap_sta_id;
@@ -1030,6 +1067,11 @@ int iwl_mvm_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
 	return ret;
 }
 
+int iwl_mvm_suspend(struct ieee80211_hw *hw, struct cfg80211_wowlan *wowlan)
+{
+	return __iwl_mvm_suspend(hw, wowlan, false);
+}
+
 static void iwl_mvm_query_wakeup_reasons(struct iwl_mvm *mvm,
 					 struct ieee80211_vif *vif)
 {
@@ -1214,9 +1256,28 @@ static void iwl_mvm_query_wakeup_reasons(struct iwl_mvm *mvm,
 	iwl_free_resp(&cmd);
 }
 
-int iwl_mvm_resume(struct ieee80211_hw *hw)
+static void iwl_mvm_read_d3_sram(struct iwl_mvm *mvm)
+{
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	const struct fw_img *img = &mvm->fw->img[IWL_UCODE_WOWLAN];
+	u32 len = img->sec[IWL_UCODE_SECTION_DATA].len;
+	u32 offs = img->sec[IWL_UCODE_SECTION_DATA].offset;
+
+	if (!mvm->store_d3_resume_sram)
+		return;
+
+	if (!mvm->d3_resume_sram) {
+		mvm->d3_resume_sram = kzalloc(len, GFP_KERNEL);
+		if (!mvm->d3_resume_sram)
+			return;
+	}
+
+	iwl_trans_read_mem_bytes(mvm->trans, offs, mvm->d3_resume_sram, len);
+#endif
+}
+
+static int __iwl_mvm_resume(struct iwl_mvm *mvm, bool test)
 {
-	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
 	struct iwl_d3_iter_data resume_iter_data = {
 		.mvm = mvm,
 	};
@@ -1236,7 +1297,7 @@ int iwl_mvm_resume(struct ieee80211_hw *hw)
 
 	vif = resume_iter_data.vif;
 
-	ret = iwl_trans_d3_resume(mvm->trans, &d3_status);
+	ret = iwl_trans_d3_resume(mvm->trans, &d3_status, test);
 	if (ret)
 		goto out_unlock;
 
@@ -1245,12 +1306,15 @@ int iwl_mvm_resume(struct ieee80211_hw *hw)
 		goto out_unlock;
 	}
 
+	/* query SRAM first in case we want event logging */
+	iwl_mvm_read_d3_sram(mvm);
+
 	iwl_mvm_query_wakeup_reasons(mvm, vif);
 
  out_unlock:
 	mutex_unlock(&mvm->mutex);
 
-	if (vif)
+	if (!test && vif)
 		ieee80211_resume_disconnect(vif);
 
 	/* return 1 to reconfigure the device */
@@ -1258,9 +1322,106 @@ int iwl_mvm_resume(struct ieee80211_hw *hw)
 	return 1;
 }
 
+int iwl_mvm_resume(struct ieee80211_hw *hw)
+{
+	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
+
+	return __iwl_mvm_resume(mvm, false);
+}
+
 void iwl_mvm_set_wakeup(struct ieee80211_hw *hw, bool enabled)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
 
 	device_set_wakeup_enable(mvm->trans->dev, enabled);
 }
+
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+static int iwl_mvm_d3_test_open(struct inode *inode, struct file *file)
+{
+	struct iwl_mvm *mvm = inode->i_private;
+	int err;
+
+	if (mvm->d3_test_active)
+		return -EBUSY;
+
+	file->private_data = inode->i_private;
+
+	ieee80211_stop_queues(mvm->hw);
+	synchronize_net();
+
+	/* start pseudo D3 */
+	rtnl_lock();
+	err = __iwl_mvm_suspend(mvm->hw, mvm->hw->wiphy->wowlan_config, true);
+	rtnl_unlock();
+	if (err > 0)
+		err = -EINVAL;
+	if (err) {
+		ieee80211_wake_queues(mvm->hw);
+		return err;
+	}
+	mvm->d3_test_active = true;
+	return 0;
+}
+
+static ssize_t iwl_mvm_d3_test_read(struct file *file, char __user *user_buf,
+				    size_t count, loff_t *ppos)
+{
+	struct iwl_mvm *mvm = file->private_data;
+	u32 pme_asserted;
+
+	while (true) {
+		pme_asserted = iwl_trans_read_mem32(mvm->trans,
+						    mvm->d3_test_pme_ptr);
+		if (pme_asserted)
+			break;
+		if (msleep_interruptible(100))
+			break;
+	}
+
+	return 0;
+}
+
+static void iwl_mvm_d3_test_disconn_work_iter(void *_data, u8 *mac,
+					      struct ieee80211_vif *vif)
+{
+	if (vif->type == NL80211_IFTYPE_STATION)
+		ieee80211_connection_loss(vif);
+}
+
+static int iwl_mvm_d3_test_release(struct inode *inode, struct file *file)
+{
+	struct iwl_mvm *mvm = inode->i_private;
+	int remaining_time = 10;
+
+	mvm->d3_test_active = false;
+	__iwl_mvm_resume(mvm, true);
+	iwl_abort_notification_waits(&mvm->notif_wait);
+	ieee80211_restart_hw(mvm->hw);
+
+	/* wait for restart and disconnect all interfaces */
+	while (test_bit(IWL_MVM_STATUS_IN_HW_RESTART, &mvm->status) &&
+	       remaining_time > 0) {
+		remaining_time--;
+		msleep(1000);
+	}
+
+	if (remaining_time == 0)
+		IWL_ERR(mvm, "Timed out waiting for HW restart to finish!\n");
+
+	ieee80211_iterate_active_interfaces_atomic(
+		mvm->hw, IEEE80211_IFACE_ITER_NORMAL,
+		iwl_mvm_d3_test_disconn_work_iter, NULL);
+
+	ieee80211_wake_queues(mvm->hw);
+
+	return 0;
+}
+
+const struct file_operations iwl_dbgfs_d3_test_ops = {
+	.llseek = no_llseek,
+	.open = iwl_mvm_d3_test_open,
+	.read = iwl_mvm_d3_test_read,
+	.release = iwl_mvm_d3_test_release,
+};
+#endif
diff --git a/drivers/net/wireless/iwlwifi/mvm/debugfs.c b/drivers/net/wireless/iwlwifi/mvm/debugfs.c
index 2053dcc..e56ed2a 100644
--- a/drivers/net/wireless/iwlwifi/mvm/debugfs.c
+++ b/drivers/net/wireless/iwlwifi/mvm/debugfs.c
@@ -145,15 +145,18 @@ static ssize_t iwl_dbgfs_sram_read(struct file *file, char __user *user_buf,
 	char *buf;
 	u8 *ptr;
 
+	if (!mvm->ucode_loaded)
+		return -EINVAL;
+
 	/* default is to dump the entire data segment */
 	if (!mvm->dbgfs_sram_offset && !mvm->dbgfs_sram_len) {
-		mvm->dbgfs_sram_offset = 0x800000;
-		if (!mvm->ucode_loaded)
-			return -EINVAL;
 		img = &mvm->fw->img[mvm->cur_ucode];
-		mvm->dbgfs_sram_len = img->sec[IWL_UCODE_SECTION_DATA].len;
+		ofs = img->sec[IWL_UCODE_SECTION_DATA].offset;
+		len = img->sec[IWL_UCODE_SECTION_DATA].len;
+	} else {
+		ofs = mvm->dbgfs_sram_offset;
+		len = mvm->dbgfs_sram_len;
 	}
-	len = mvm->dbgfs_sram_len;
 
 	bufsz = len * 4 + 256;
 	buf = kzalloc(bufsz, GFP_KERNEL);
@@ -167,12 +170,9 @@ static ssize_t iwl_dbgfs_sram_read(struct file *file, char __user *user_buf,
 	}
 
 	pos += scnprintf(buf + pos, bufsz - pos, "sram_len: 0x%x\n", len);
-	pos += scnprintf(buf + pos, bufsz - pos, "sram_offset: 0x%x\n",
-			 mvm->dbgfs_sram_offset);
+	pos += scnprintf(buf + pos, bufsz - pos, "sram_offset: 0x%x\n", ofs);
 
-	iwl_trans_read_mem_bytes(mvm->trans,
-				 mvm->dbgfs_sram_offset,
-				 ptr, len);
+	iwl_trans_read_mem_bytes(mvm->trans, ofs, ptr, len);
 	for (ofs = 0; ofs < len; ofs += 16) {
 		pos += scnprintf(buf + pos, bufsz - pos, "0x%.4x ", ofs);
 		hex_dump_to_buffer(ptr + ofs, 16, 16, 1, buf + pos,
@@ -300,6 +300,168 @@ static ssize_t iwl_dbgfs_power_down_d3_allow_write(struct file *file,
 	return count;
 }
 
+static void iwl_dbgfs_update_pm(struct iwl_mvm *mvm,
+				 struct ieee80211_vif *vif,
+				 enum iwl_dbgfs_pm_mask param, int val)
+{
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_dbgfs_pm *dbgfs_pm = &mvmvif->dbgfs_pm;
+
+	dbgfs_pm->mask |= param;
+
+	switch (param) {
+	case MVM_DEBUGFS_PM_KEEP_ALIVE: {
+		struct ieee80211_hw *hw = mvm->hw;
+		int dtimper = hw->conf.ps_dtim_period ?: 1;
+		int dtimper_msec = dtimper * vif->bss_conf.beacon_int;
+
+		IWL_DEBUG_POWER(mvm, "debugfs: set keep_alive= %d sec\n", val);
+		if (val * MSEC_PER_SEC < 3 * dtimper_msec) {
+			IWL_WARN(mvm,
+				 "debugfs: keep alive period (%ld msec) is less than minimum required (%d msec)\n",
+				 val * MSEC_PER_SEC, 3 * dtimper_msec);
+		}
+		dbgfs_pm->keep_alive_seconds = val;
+		break;
+	}
+	case MVM_DEBUGFS_PM_SKIP_OVER_DTIM:
+		IWL_DEBUG_POWER(mvm, "skip_over_dtim %s\n",
+				val ? "enabled" : "disabled");
+		dbgfs_pm->skip_over_dtim = val;
+		break;
+	case MVM_DEBUGFS_PM_SKIP_DTIM_PERIODS:
+		IWL_DEBUG_POWER(mvm, "skip_dtim_periods=%d\n", val);
+		dbgfs_pm->skip_dtim_periods = val;
+		break;
+	case MVM_DEBUGFS_PM_RX_DATA_TIMEOUT:
+		IWL_DEBUG_POWER(mvm, "rx_data_timeout=%d\n", val);
+		dbgfs_pm->rx_data_timeout = val;
+		break;
+	case MVM_DEBUGFS_PM_TX_DATA_TIMEOUT:
+		IWL_DEBUG_POWER(mvm, "tx_data_timeout=%d\n", val);
+		dbgfs_pm->tx_data_timeout = val;
+		break;
+	case MVM_DEBUGFS_PM_DISABLE_POWER_OFF:
+		IWL_DEBUG_POWER(mvm, "disable_power_off=%d\n", val);
+		dbgfs_pm->disable_power_off = val;
+	case MVM_DEBUGFS_PM_LPRX_ENA:
+		IWL_DEBUG_POWER(mvm, "lprx %s\n", val ? "enabled" : "disabled");
+		dbgfs_pm->lprx_ena = val;
+		break;
+	case MVM_DEBUGFS_PM_LPRX_RSSI_THRESHOLD:
+		IWL_DEBUG_POWER(mvm, "lprx_rssi_threshold=%d\n", val);
+		dbgfs_pm->lprx_rssi_threshold = val;
+		break;
+	}
+}
+
+static ssize_t iwl_dbgfs_pm_params_write(struct file *file,
+					 const char __user *user_buf,
+					 size_t count, loff_t *ppos)
+{
+	struct ieee80211_vif *vif = file->private_data;
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_mvm *mvm = mvmvif->dbgfs_data;
+	enum iwl_dbgfs_pm_mask param;
+	char buf[32] = {};
+	int val;
+	int ret;
+
+	if (copy_from_user(buf, user_buf, sizeof(buf)))
+		return -EFAULT;
+
+	if (!strncmp("keep_alive=", buf, 11)) {
+		if (sscanf(buf + 11, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_KEEP_ALIVE;
+	} else if (!strncmp("skip_over_dtim=", buf, 15)) {
+		if (sscanf(buf + 15, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_SKIP_OVER_DTIM;
+	} else if (!strncmp("skip_dtim_periods=", buf, 18)) {
+		if (sscanf(buf + 18, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_SKIP_DTIM_PERIODS;
+	} else if (!strncmp("rx_data_timeout=", buf, 16)) {
+		if (sscanf(buf + 16, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_RX_DATA_TIMEOUT;
+	} else if (!strncmp("tx_data_timeout=", buf, 16)) {
+		if (sscanf(buf + 16, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_TX_DATA_TIMEOUT;
+	} else if (!strncmp("disable_power_off=", buf, 18)) {
+		if (sscanf(buf + 18, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_DISABLE_POWER_OFF;
+	} else if (!strncmp("lprx=", buf, 5)) {
+		if (sscanf(buf + 5, "%d", &val) != 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_LPRX_ENA;
+	} else if (!strncmp("lprx_rssi_threshold=", buf, 20)) {
+		if (sscanf(buf + 20, "%d", &val) != 1)
+			return -EINVAL;
+		if (val > POWER_LPRX_RSSI_THRESHOLD_MAX || val <
+		    POWER_LPRX_RSSI_THRESHOLD_MIN)
+			return -EINVAL;
+		param = MVM_DEBUGFS_PM_LPRX_RSSI_THRESHOLD;
+	} else {
+		return -EINVAL;
+	}
+
+	mutex_lock(&mvm->mutex);
+	iwl_dbgfs_update_pm(mvm, vif, param, val);
+	ret = iwl_mvm_power_update_mode(mvm, vif);
+	mutex_unlock(&mvm->mutex);
+
+	return ret ?: count;
+}
+
+static ssize_t iwl_dbgfs_pm_params_read(struct file *file,
+					char __user *user_buf,
+					size_t count, loff_t *ppos)
+{
+	struct ieee80211_vif *vif = file->private_data;
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_mvm *mvm = mvmvif->dbgfs_data;
+	struct iwl_powertable_cmd cmd = {};
+	char buf[256];
+	int bufsz = sizeof(buf);
+	int pos = 0;
+
+	iwl_mvm_power_build_cmd(mvm, vif, &cmd);
+
+	pos += scnprintf(buf+pos, bufsz-pos, "disable_power_off = %d\n",
+			 (cmd.flags &
+			 cpu_to_le16(POWER_FLAGS_POWER_SAVE_ENA_MSK)) ?
+			 0 : 1);
+	pos += scnprintf(buf+pos, bufsz-pos, "skip_dtim_periods = %d\n",
+			 le32_to_cpu(cmd.skip_dtim_periods));
+	pos += scnprintf(buf+pos, bufsz-pos, "power_scheme = %d\n",
+			 iwlmvm_mod_params.power_scheme);
+	pos += scnprintf(buf+pos, bufsz-pos, "flags = 0x%x\n",
+			 le16_to_cpu(cmd.flags));
+	pos += scnprintf(buf+pos, bufsz-pos, "keep_alive = %d\n",
+			 cmd.keep_alive_seconds);
+
+	if (cmd.flags & cpu_to_le16(POWER_FLAGS_POWER_MANAGEMENT_ENA_MSK)) {
+		pos += scnprintf(buf+pos, bufsz-pos, "skip_over_dtim = %d\n",
+				 (cmd.flags &
+				 cpu_to_le16(POWER_FLAGS_SKIP_OVER_DTIM_MSK)) ?
+				 1 : 0);
+		pos += scnprintf(buf+pos, bufsz-pos, "rx_data_timeout = %d\n",
+				 le32_to_cpu(cmd.rx_data_timeout));
+		pos += scnprintf(buf+pos, bufsz-pos, "tx_data_timeout = %d\n",
+				 le32_to_cpu(cmd.tx_data_timeout));
+		if (cmd.flags & cpu_to_le16(POWER_FLAGS_LPRX_ENA_MSK))
+			pos += scnprintf(buf+pos, bufsz-pos,
+					 "lprx_rssi_threshold = %d\n",
+					 le32_to_cpu(cmd.lprx_rssi_threshold));
+	}
+
+	return simple_read_from_buffer(user_buf, count, ppos, buf, pos);
+}
+
 static ssize_t iwl_dbgfs_mac_params_read(struct file *file,
 					 char __user *user_buf,
 					 size_t count, loff_t *ppos)
@@ -481,6 +643,255 @@ static ssize_t iwl_dbgfs_fw_restart_write(struct file *file,
 	return count;
 }
 
+static void iwl_dbgfs_update_bf(struct ieee80211_vif *vif,
+				enum iwl_dbgfs_bf_mask param, int value)
+{
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_dbgfs_bf *dbgfs_bf = &mvmvif->dbgfs_bf;
+
+	dbgfs_bf->mask |= param;
+
+	switch (param) {
+	case MVM_DEBUGFS_BF_ENERGY_DELTA:
+		dbgfs_bf->bf_energy_delta = value;
+		break;
+	case MVM_DEBUGFS_BF_ROAMING_ENERGY_DELTA:
+		dbgfs_bf->bf_roaming_energy_delta = value;
+		break;
+	case MVM_DEBUGFS_BF_ROAMING_STATE:
+		dbgfs_bf->bf_roaming_state = value;
+		break;
+	case MVM_DEBUGFS_BF_TEMPERATURE_DELTA:
+		dbgfs_bf->bf_temperature_delta = value;
+		break;
+	case MVM_DEBUGFS_BF_ENABLE_BEACON_FILTER:
+		dbgfs_bf->bf_enable_beacon_filter = value;
+		break;
+	case MVM_DEBUGFS_BF_DEBUG_FLAG:
+		dbgfs_bf->bf_debug_flag = value;
+		break;
+	case MVM_DEBUGFS_BF_ESCAPE_TIMER:
+		dbgfs_bf->bf_escape_timer = value;
+		break;
+	case MVM_DEBUGFS_BA_ENABLE_BEACON_ABORT:
+		dbgfs_bf->ba_enable_beacon_abort = value;
+		break;
+	case MVM_DEBUGFS_BA_ESCAPE_TIMER:
+		dbgfs_bf->ba_escape_timer = value;
+		break;
+	}
+}
+
+static ssize_t iwl_dbgfs_bf_params_write(struct file *file,
+					 const char __user *user_buf,
+					 size_t count, loff_t *ppos)
+{
+	struct ieee80211_vif *vif = file->private_data;
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_mvm *mvm = mvmvif->dbgfs_data;
+	enum iwl_dbgfs_bf_mask param;
+	char buf[256];
+	int buf_size;
+	int value;
+	int ret = 0;
+
+	memset(buf, 0, sizeof(buf));
+	buf_size = min(count, sizeof(buf) - 1);
+	if (copy_from_user(buf, user_buf, buf_size))
+		return -EFAULT;
+
+	if (!strncmp("bf_energy_delta=", buf, 16)) {
+		if (sscanf(buf+16, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < IWL_BF_ENERGY_DELTA_MIN ||
+		    value > IWL_BF_ENERGY_DELTA_MAX)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_ENERGY_DELTA;
+	} else if (!strncmp("bf_roaming_energy_delta=", buf, 24)) {
+		if (sscanf(buf+24, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < IWL_BF_ROAMING_ENERGY_DELTA_MIN ||
+		    value > IWL_BF_ROAMING_ENERGY_DELTA_MAX)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_ROAMING_ENERGY_DELTA;
+	} else if (!strncmp("bf_roaming_state=", buf, 17)) {
+		if (sscanf(buf+17, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < IWL_BF_ROAMING_STATE_MIN ||
+		    value > IWL_BF_ROAMING_STATE_MAX)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_ROAMING_STATE;
+	} else if (!strncmp("bf_temperature_delta=", buf, 21)) {
+		if (sscanf(buf+21, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < IWL_BF_TEMPERATURE_DELTA_MIN ||
+		    value > IWL_BF_TEMPERATURE_DELTA_MAX)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_TEMPERATURE_DELTA;
+	} else if (!strncmp("bf_enable_beacon_filter=", buf, 24)) {
+		if (sscanf(buf+24, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < 0 || value > 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_ENABLE_BEACON_FILTER;
+	} else if (!strncmp("bf_debug_flag=", buf, 14)) {
+		if (sscanf(buf+14, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < 0 || value > 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_DEBUG_FLAG;
+	} else if (!strncmp("bf_escape_timer=", buf, 16)) {
+		if (sscanf(buf+16, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < IWL_BF_ESCAPE_TIMER_MIN ||
+		    value > IWL_BF_ESCAPE_TIMER_MAX)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BF_ESCAPE_TIMER;
+	} else if (!strncmp("ba_escape_timer=", buf, 16)) {
+		if (sscanf(buf+16, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < IWL_BA_ESCAPE_TIMER_MIN ||
+		    value > IWL_BA_ESCAPE_TIMER_MAX)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BA_ESCAPE_TIMER;
+	} else if (!strncmp("ba_enable_beacon_abort=", buf, 23)) {
+		if (sscanf(buf+23, "%d", &value) != 1)
+			return -EINVAL;
+		if (value < 0 || value > 1)
+			return -EINVAL;
+		param = MVM_DEBUGFS_BA_ENABLE_BEACON_ABORT;
+	} else {
+		return -EINVAL;
+	}
+
+	mutex_lock(&mvm->mutex);
+	iwl_dbgfs_update_bf(vif, param, value);
+	if (param == MVM_DEBUGFS_BF_ENABLE_BEACON_FILTER && !value) {
+		ret = iwl_mvm_disable_beacon_filter(mvm, vif);
+	} else {
+		if (mvmvif->bf_enabled)
+			ret = iwl_mvm_enable_beacon_filter(mvm, vif);
+		else
+			ret = iwl_mvm_disable_beacon_filter(mvm, vif);
+	}
+	mutex_unlock(&mvm->mutex);
+
+	return ret ?: count;
+}
+
+static ssize_t iwl_dbgfs_bf_params_read(struct file *file,
+					char __user *user_buf,
+					size_t count, loff_t *ppos)
+{
+	struct ieee80211_vif *vif = file->private_data;
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	char buf[256];
+	int pos = 0;
+	const size_t bufsz = sizeof(buf);
+	struct iwl_beacon_filter_cmd cmd = {
+		.bf_energy_delta = IWL_BF_ENERGY_DELTA_DEFAULT,
+		.bf_roaming_energy_delta = IWL_BF_ROAMING_ENERGY_DELTA_DEFAULT,
+		.bf_roaming_state = IWL_BF_ROAMING_STATE_DEFAULT,
+		.bf_temperature_delta = IWL_BF_TEMPERATURE_DELTA_DEFAULT,
+		.bf_enable_beacon_filter = IWL_BF_ENABLE_BEACON_FILTER_DEFAULT,
+		.bf_debug_flag = IWL_BF_DEBUG_FLAG_DEFAULT,
+		.bf_escape_timer = cpu_to_le32(IWL_BF_ESCAPE_TIMER_DEFAULT),
+		.ba_escape_timer = cpu_to_le32(IWL_BA_ESCAPE_TIMER_DEFAULT),
+		.ba_enable_beacon_abort = IWL_BA_ENABLE_BEACON_ABORT_DEFAULT,
+	};
+
+	iwl_mvm_beacon_filter_debugfs_parameters(vif, &cmd);
+	if (mvmvif->bf_enabled)
+		cmd.bf_enable_beacon_filter = 1;
+	else
+		cmd.bf_enable_beacon_filter = 0;
+
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_energy_delta = %d\n",
+			 cmd.bf_energy_delta);
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_roaming_energy_delta = %d\n",
+			 cmd.bf_roaming_energy_delta);
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_roaming_state = %d\n",
+			 cmd.bf_roaming_state);
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_temperature_delta = %d\n",
+			 cmd.bf_temperature_delta);
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_enable_beacon_filter = %d\n",
+			 cmd.bf_enable_beacon_filter);
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_debug_flag = %d\n",
+			 cmd.bf_debug_flag);
+	pos += scnprintf(buf+pos, bufsz-pos, "bf_escape_timer = %d\n",
+			 cmd.bf_escape_timer);
+	pos += scnprintf(buf+pos, bufsz-pos, "ba_escape_timer = %d\n",
+			 cmd.ba_escape_timer);
+	pos += scnprintf(buf+pos, bufsz-pos, "ba_enable_beacon_abort = %d\n",
+			 cmd.ba_enable_beacon_abort);
+
+	return simple_read_from_buffer(user_buf, count, ppos, buf, pos);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static ssize_t iwl_dbgfs_d3_sram_write(struct file *file,
+				       const char __user *user_buf,
+				       size_t count, loff_t *ppos)
+{
+	struct iwl_mvm *mvm = file->private_data;
+	char buf[8] = {};
+	int store;
+
+	if (copy_from_user(buf, user_buf, sizeof(buf)))
+		return -EFAULT;
+
+	if (sscanf(buf, "%d", &store) != 1)
+		return -EINVAL;
+
+	mvm->store_d3_resume_sram = store;
+
+	return count;
+}
+
+static ssize_t iwl_dbgfs_d3_sram_read(struct file *file, char __user *user_buf,
+				      size_t count, loff_t *ppos)
+{
+	struct iwl_mvm *mvm = file->private_data;
+	const struct fw_img *img;
+	int ofs, len, pos = 0;
+	size_t bufsz, ret;
+	char *buf;
+	u8 *ptr = mvm->d3_resume_sram;
+
+	img = &mvm->fw->img[IWL_UCODE_WOWLAN];
+	len = img->sec[IWL_UCODE_SECTION_DATA].len;
+
+	bufsz = len * 4 + 256;
+	buf = kzalloc(bufsz, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	pos += scnprintf(buf, bufsz, "D3 SRAM capture: %sabled\n",
+			 mvm->store_d3_resume_sram ? "en" : "dis");
+
+	if (ptr) {
+		for (ofs = 0; ofs < len; ofs += 16) {
+			pos += scnprintf(buf + pos, bufsz - pos,
+					 "0x%.4x ", ofs);
+			hex_dump_to_buffer(ptr + ofs, 16, 16, 1, buf + pos,
+					   bufsz - pos, false);
+			pos += strlen(buf + pos);
+			if (bufsz - pos > 0)
+				buf[pos++] = '\n';
+		}
+	} else {
+		pos += scnprintf(buf + pos, bufsz - pos,
+				 "(no data captured)\n");
+	}
+
+	ret = simple_read_from_buffer(user_buf, count, ppos, buf, pos);
+
+	kfree(buf);
+
+	return ret;
+}
+#endif
+
 #define MVM_DEBUGFS_READ_FILE_OPS(name)					\
 static const struct file_operations iwl_dbgfs_##name##_ops = {	\
 	.read = iwl_dbgfs_##name##_read,				\
@@ -524,9 +935,14 @@ MVM_DEBUGFS_READ_FILE_OPS(bt_notif);
 MVM_DEBUGFS_WRITE_FILE_OPS(power_down_allow);
 MVM_DEBUGFS_WRITE_FILE_OPS(power_down_d3_allow);
 MVM_DEBUGFS_WRITE_FILE_OPS(fw_restart);
+#ifdef CONFIG_PM_SLEEP
+MVM_DEBUGFS_READ_WRITE_FILE_OPS(d3_sram);
+#endif
 
 /* Interface specific debugfs entries */
 MVM_DEBUGFS_READ_FILE_OPS(mac_params);
+MVM_DEBUGFS_READ_WRITE_FILE_OPS(pm_params);
+MVM_DEBUGFS_READ_WRITE_FILE_OPS(bf_params);
 
 int iwl_mvm_dbgfs_register(struct iwl_mvm *mvm, struct dentry *dbgfs_dir)
 {
@@ -542,6 +958,13 @@ int iwl_mvm_dbgfs_register(struct iwl_mvm *mvm, struct dentry *dbgfs_dir)
 	MVM_DEBUGFS_ADD_FILE(power_down_allow, mvm->debugfs_dir, S_IWUSR);
 	MVM_DEBUGFS_ADD_FILE(power_down_d3_allow, mvm->debugfs_dir, S_IWUSR);
 	MVM_DEBUGFS_ADD_FILE(fw_restart, mvm->debugfs_dir, S_IWUSR);
+#ifdef CONFIG_PM_SLEEP
+	MVM_DEBUGFS_ADD_FILE(d3_sram, mvm->debugfs_dir, S_IRUSR | S_IWUSR);
+	MVM_DEBUGFS_ADD_FILE(d3_test, mvm->debugfs_dir, S_IRUSR);
+	if (!debugfs_create_bool("d3_wake_sysassert", S_IRUSR | S_IWUSR,
+				 mvm->debugfs_dir, &mvm->d3_wake_sysassert))
+		goto err;
+#endif
 
 	/*
 	 * Create a symlink with mac80211. It will be removed when mac80211
@@ -577,9 +1000,19 @@ void iwl_mvm_vif_dbgfs_register(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 		return;
 	}
 
+	if (iwlmvm_mod_params.power_scheme != IWL_POWER_SCHEME_CAM &&
+	    vif->type == NL80211_IFTYPE_STATION && !vif->p2p)
+		MVM_DEBUGFS_ADD_FILE_VIF(pm_params, mvmvif->dbgfs_dir, S_IWUSR |
+					 S_IRUSR);
+
 	MVM_DEBUGFS_ADD_FILE_VIF(mac_params, mvmvif->dbgfs_dir,
 				 S_IRUSR);
 
+	if (vif->type == NL80211_IFTYPE_STATION && !vif->p2p &&
+	    mvmvif == mvm->bf_allowed_vif)
+		MVM_DEBUGFS_ADD_FILE_VIF(bf_params, mvmvif->dbgfs_dir,
+					 S_IRUSR | S_IWUSR);
+
 	/*
 	 * Create symlink for convenience pointing to interface specific
 	 * debugfs entries for the driver. For example, under
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw-api-d3.h b/drivers/net/wireless/iwlwifi/mvm/fw-api-d3.h
index 51e015d..6f8b2c1 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw-api-d3.h
+++ b/drivers/net/wireless/iwlwifi/mvm/fw-api-d3.h
@@ -75,13 +75,15 @@ enum iwl_d3_wakeup_flags {
  * struct iwl_d3_manager_config - D3 manager configuration command
  * @min_sleep_time: minimum sleep time (in usec)
  * @wakeup_flags: wakeup flags, see &enum iwl_d3_wakeup_flags
+ * @wakeup_host_timer: force wakeup after this many seconds
  *
  * The structure is used for the D3_CONFIG_CMD command.
  */
 struct iwl_d3_manager_config {
 	__le32 min_sleep_time;
 	__le32 wakeup_flags;
-} __packed; /* D3_MANAGER_CONFIG_CMD_S_VER_3 */
+	__le32 wakeup_host_timer;
+} __packed; /* D3_MANAGER_CONFIG_CMD_S_VER_4 */
 
 
 /* TODO: OFFLOADS_QUERY_API_S_VER_1 */
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw-api-mac.h b/drivers/net/wireless/iwlwifi/mvm/fw-api-mac.h
index d68640e..98b1feb 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw-api-mac.h
+++ b/drivers/net/wireless/iwlwifi/mvm/fw-api-mac.h
@@ -71,7 +71,13 @@
 #define MAC_INDEX_MIN_DRIVER	0
 #define NUM_MAC_INDEX_DRIVER	MAC_INDEX_AUX
 
-#define AC_NUM	4 /* Number of access categories */
+enum iwl_ac {
+	AC_BK,
+	AC_BE,
+	AC_VI,
+	AC_VO,
+	AC_NUM,
+};
 
 /**
  * enum iwl_mac_protection_flags - MAC context flags
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw-api-power.h b/drivers/net/wireless/iwlwifi/mvm/fw-api-power.h
index 81fe45f..a6da359 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw-api-power.h
+++ b/drivers/net/wireless/iwlwifi/mvm/fw-api-power.h
@@ -66,6 +66,11 @@
 
 /* Power Management Commands, Responses, Notifications */
 
+/* Radio LP RX Energy Threshold measured in dBm */
+#define POWER_LPRX_RSSI_THRESHOLD	75
+#define POWER_LPRX_RSSI_THRESHOLD_MAX	94
+#define POWER_LPRX_RSSI_THRESHOLD_MIN	30
+
 /**
  * enum iwl_scan_flags - masks for power table command flags
  * @POWER_FLAGS_POWER_SAVE_ENA_MSK: '1' Allow to save power by turning off
@@ -101,20 +106,107 @@ enum iwl_power_flags {
  * @tx_data_timeout:    Minimum time (usec) from last Tx packet for AM to
  *			PSM transition - legacy PM
  * @sleep_interval:	not in use
- * @keep_alive_beacons:	not in use
+ * @skip_dtim_periods:	Number of DTIM periods to skip if Skip over DTIM flag
+ *			is set. For example, if it is required to skip over
+ *			one DTIM, this value need to be set to 2 (DTIM periods).
  * @lprx_rssi_threshold: Signal strength up to which LP RX can be enabled.
  *			Default: 80dbm
  */
 struct iwl_powertable_cmd {
-	/* PM_POWER_TABLE_CMD_API_S_VER_5 */
+	/* PM_POWER_TABLE_CMD_API_S_VER_6 */
 	__le16 flags;
 	u8 keep_alive_seconds;
 	u8 debug_flags;
 	__le32 rx_data_timeout;
 	__le32 tx_data_timeout;
 	__le32 sleep_interval[IWL_POWER_VEC_SIZE];
-	__le32 keep_alive_beacons;
+	__le32 skip_dtim_periods;
 	__le32 lprx_rssi_threshold;
 } __packed;
 
+/**
+ * struct iwl_beacon_filter_cmd
+ * REPLY_BEACON_FILTERING_CMD = 0xd2 (command)
+ * @id_and_color: MAC contex identifier
+ * @bf_energy_delta: Used for RSSI filtering, if in 'normal' state. Send beacon
+ *      to driver if delta in Energy values calculated for this and last
+ *      passed beacon is greater than this threshold. Zero value means that
+ *      the Energy change is ignored for beacon filtering, and beacon will
+ *      not be forced to be sent to driver regardless of this delta. Typical
+ *      energy delta 5dB.
+ * @bf_roaming_energy_delta: Used for RSSI filtering, if in 'roaming' state.
+ *      Send beacon to driver if delta in Energy values calculated for this
+ *      and last passed beacon is greater than this threshold. Zero value
+ *      means that the Energy change is ignored for beacon filtering while in
+ *      Roaming state, typical energy delta 1dB.
+ * @bf_roaming_state: Used for RSSI filtering. If absolute Energy values
+ *      calculated for current beacon is less than the threshold, use
+ *      Roaming Energy Delta Threshold, otherwise use normal Energy Delta
+ *      Threshold. Typical energy threshold is -72dBm.
+ * @bf_temperature_delta: Send Beacon to driver if delta in temperature values
+ *      calculated for this and the last passed beacon is greater than  this
+ *      threshold. Zero value means that the temperature changeis ignored for
+ *      beacon filtering; beacons will not be  forced to be sent to driver
+ *      regardless of whether its temerature has been changed.
+ * @bf_enable_beacon_filter: 1, beacon filtering is enabled; 0, disabled.
+ * @bf_filter_escape_timer: Send beacons to to driver if no beacons were passed
+ *      for a specific period of time. Units: Beacons.
+ * @ba_escape_timer: Fully receive and parse beacon if no beacons were passed
+ *      for a longer period of time then this escape-timeout. Units: Beacons.
+ * @ba_enable_beacon_abort: 1, beacon abort is enabled; 0, disabled.
+ */
+struct iwl_beacon_filter_cmd {
+	u8 bf_energy_delta;
+	u8 bf_roaming_energy_delta;
+	u8 bf_roaming_state;
+	u8 bf_temperature_delta;
+	u8 bf_enable_beacon_filter;
+	u8 bf_debug_flag;
+	__le16 reserved1;
+	__le32 bf_escape_timer;
+	__le32 ba_escape_timer;
+	u8 ba_enable_beacon_abort;
+	u8 reserved2[3];
+} __packed;
+
+/* Beacon filtering and beacon abort */
+#define IWL_BF_ENERGY_DELTA_DEFAULT 5
+#define IWL_BF_ENERGY_DELTA_MAX 255
+#define IWL_BF_ENERGY_DELTA_MIN 0
+
+#define IWL_BF_ROAMING_ENERGY_DELTA_DEFAULT 1
+#define IWL_BF_ROAMING_ENERGY_DELTA_MAX 255
+#define IWL_BF_ROAMING_ENERGY_DELTA_MIN 0
+
+#define IWL_BF_ROAMING_STATE_DEFAULT 72
+#define IWL_BF_ROAMING_STATE_MAX 255
+#define IWL_BF_ROAMING_STATE_MIN 0
+
+#define IWL_BF_TEMPERATURE_DELTA_DEFAULT 5
+#define IWL_BF_TEMPERATURE_DELTA_MAX 255
+#define IWL_BF_TEMPERATURE_DELTA_MIN 0
+
+#define IWL_BF_ENABLE_BEACON_FILTER_DEFAULT 1
+
+#define IWL_BF_DEBUG_FLAG_DEFAULT 0
+
+#define IWL_BF_ESCAPE_TIMER_DEFAULT 50
+#define IWL_BF_ESCAPE_TIMER_MAX 1024
+#define IWL_BF_ESCAPE_TIMER_MIN 0
+
+#define IWL_BA_ESCAPE_TIMER_DEFAULT 3
+#define IWL_BA_ESCAPE_TIMER_MAX 1024
+#define IWL_BA_ESCAPE_TIMER_MIN 0
+
+#define IWL_BA_ENABLE_BEACON_ABORT_DEFAULT 1
+
+#define IWL_BF_CMD_CONFIG_DEFAULTS					\
+	.bf_energy_delta = IWL_BF_ENERGY_DELTA_DEFAULT,			\
+	.bf_roaming_energy_delta = IWL_BF_ROAMING_ENERGY_DELTA_DEFAULT,	\
+	.bf_roaming_state = IWL_BF_ROAMING_STATE_DEFAULT,		\
+	.bf_temperature_delta = IWL_BF_TEMPERATURE_DELTA_DEFAULT,	\
+	.bf_debug_flag = IWL_BF_DEBUG_FLAG_DEFAULT,			\
+	.bf_escape_timer = cpu_to_le32(IWL_BF_ESCAPE_TIMER_DEFAULT),	\
+	.ba_escape_timer = cpu_to_le32(IWL_BA_ESCAPE_TIMER_DEFAULT)
+
 #endif
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw-api-tx.h b/drivers/net/wireless/iwlwifi/mvm/fw-api-tx.h
index 007a93b..700cce7 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw-api-tx.h
+++ b/drivers/net/wireless/iwlwifi/mvm/fw-api-tx.h
@@ -134,6 +134,7 @@ enum iwl_tx_flags {
 #define TX_CMD_SEC_WEP			0x01
 #define TX_CMD_SEC_CCM			0x02
 #define TX_CMD_SEC_TKIP			0x03
+#define TX_CMD_SEC_MSK			0x07
 #define TX_CMD_SEC_WEP_KEY_IDX_POS	6
 #define TX_CMD_SEC_WEP_KEY_IDX_MSK	0xc0
 #define TX_CMD_SEC_KEY128		0x08
@@ -227,10 +228,11 @@ struct iwl_tx_cmd {
 	__le16 len;
 	__le16 next_frame_len;
 	__le32 tx_flags;
-	/* DRAM_SCRATCH_API_U_VER_1 */
-	u8 try_cnt;
-	u8 btkill_cnt;
-	__le16 reserved;
+	struct {
+		u8 try_cnt;
+		u8 btkill_cnt;
+		__le16 reserved;
+	} scratch; /* DRAM_SCRATCH_API_U_VER_1 */
 	__le32 rate_n_flags;
 	u8 sta_id;
 	u8 sec_ctl;
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw-api.h b/drivers/net/wireless/iwlwifi/mvm/fw-api.h
index c638455..cbfb3be 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw-api.h
+++ b/drivers/net/wireless/iwlwifi/mvm/fw-api.h
@@ -139,6 +139,9 @@ enum {
 	/* Power */
 	POWER_TABLE_CMD = 0x77,
 
+	/* Thermal Throttling*/
+	REPLY_THERMAL_MNG_BACKOFF = 0x7e,
+
 	/* Scanning */
 	SCAN_REQUEST_CMD = 0x80,
 	SCAN_ABORT_CMD = 0x81,
@@ -161,6 +164,8 @@ enum {
 	CARD_STATE_CMD = 0xa0,
 	CARD_STATE_NOTIFICATION = 0xa1,
 
+	MISSED_BEACONS_NOTIFICATION = 0xa2,
+
 	REPLY_RX_PHY_CMD = 0xc0,
 	REPLY_RX_MPDU_CMD = 0xc1,
 	BA_NOTIF = 0xc5,
@@ -170,6 +175,8 @@ enum {
 	BT_COEX_PROT_ENV = 0xcd,
 	BT_PROFILE_NOTIFICATION = 0xce,
 
+	REPLY_BEACON_FILTERING_CMD = 0xd2,
+
 	REPLY_DEBUG_CMD = 0xf0,
 	DEBUG_LOG_MSG = 0xf7,
 
@@ -938,6 +945,24 @@ struct iwl_card_state_notif {
 } __packed; /* CARD_STATE_NTFY_API_S_VER_1 */
 
 /**
+ * struct iwl_missed_beacons_notif - information on missed beacons
+ * ( MISSED_BEACONS_NOTIFICATION = 0xa2 )
+ * @mac_id: interface ID
+ * @consec_missed_beacons_since_last_rx: number of consecutive missed
+ *	beacons since last RX.
+ * @consec_missed_beacons: number of consecutive missed beacons
+ * @num_expected_beacons:
+ * @num_recvd_beacons:
+ */
+struct iwl_missed_beacons_notif {
+	__le32 mac_id;
+	__le32 consec_missed_beacons_since_last_rx;
+	__le32 consec_missed_beacons;
+	__le32 num_expected_beacons;
+	__le32 num_recvd_beacons;
+} __packed; /* MISSED_BEACON_NTFY_API_S_VER_3 */
+
+/**
  * struct iwl_set_calib_default_cmd - set default value for calibration.
  * ( SET_CALIB_DEFAULT_CMD = 0x8e )
  * @calib_index: the calibration to set value for
@@ -975,4 +1000,212 @@ struct iwl_mcast_filter_cmd {
 	u8 addr_list[0];
 } __packed; /* MCAST_FILTERING_CMD_API_S_VER_1 */
 
+struct mvm_statistics_dbg {
+	__le32 burst_check;
+	__le32 burst_count;
+	__le32 wait_for_silence_timeout_cnt;
+	__le32 reserved[3];
+} __packed; /* STATISTICS_DEBUG_API_S_VER_2 */
+
+struct mvm_statistics_div {
+	__le32 tx_on_a;
+	__le32 tx_on_b;
+	__le32 exec_time;
+	__le32 probe_time;
+	__le32 rssi_ant;
+	__le32 reserved2;
+} __packed; /* STATISTICS_SLOW_DIV_API_S_VER_2 */
+
+struct mvm_statistics_general_common {
+	__le32 temperature;   /* radio temperature */
+	__le32 temperature_m; /* radio voltage */
+	struct mvm_statistics_dbg dbg;
+	__le32 sleep_time;
+	__le32 slots_out;
+	__le32 slots_idle;
+	__le32 ttl_timestamp;
+	struct mvm_statistics_div div;
+	__le32 rx_enable_counter;
+	/*
+	 * num_of_sos_states:
+	 *  count the number of times we have to re-tune
+	 *  in order to get out of bad PHY status
+	 */
+	__le32 num_of_sos_states;
+} __packed; /* STATISTICS_GENERAL_API_S_VER_5 */
+
+struct mvm_statistics_rx_non_phy {
+	__le32 bogus_cts;	/* CTS received when not expecting CTS */
+	__le32 bogus_ack;	/* ACK received when not expecting ACK */
+	__le32 non_bssid_frames;	/* number of frames with BSSID that
+					 * doesn't belong to the STA BSSID */
+	__le32 filtered_frames;	/* count frames that were dumped in the
+				 * filtering process */
+	__le32 non_channel_beacons;	/* beacons with our bss id but not on
+					 * our serving channel */
+	__le32 channel_beacons;	/* beacons with our bss id and in our
+				 * serving channel */
+	__le32 num_missed_bcon;	/* number of missed beacons */
+	__le32 adc_rx_saturation_time;	/* count in 0.8us units the time the
+					 * ADC was in saturation */
+	__le32 ina_detection_search_time;/* total time (in 0.8us) searched
+					  * for INA */
+	__le32 beacon_silence_rssi_a;	/* RSSI silence after beacon frame */
+	__le32 beacon_silence_rssi_b;	/* RSSI silence after beacon frame */
+	__le32 beacon_silence_rssi_c;	/* RSSI silence after beacon frame */
+	__le32 interference_data_flag;	/* flag for interference data
+					 * availability. 1 when data is
+					 * available. */
+	__le32 channel_load;		/* counts RX Enable time in uSec */
+	__le32 dsp_false_alarms;	/* DSP false alarm (both OFDM
+					 * and CCK) counter */
+	__le32 beacon_rssi_a;
+	__le32 beacon_rssi_b;
+	__le32 beacon_rssi_c;
+	__le32 beacon_energy_a;
+	__le32 beacon_energy_b;
+	__le32 beacon_energy_c;
+	__le32 num_bt_kills;
+	__le32 mac_id;
+	__le32 directed_data_mpdu;
+} __packed; /* STATISTICS_RX_NON_PHY_API_S_VER_3 */
+
+struct mvm_statistics_rx_phy {
+	__le32 ina_cnt;
+	__le32 fina_cnt;
+	__le32 plcp_err;
+	__le32 crc32_err;
+	__le32 overrun_err;
+	__le32 early_overrun_err;
+	__le32 crc32_good;
+	__le32 false_alarm_cnt;
+	__le32 fina_sync_err_cnt;
+	__le32 sfd_timeout;
+	__le32 fina_timeout;
+	__le32 unresponded_rts;
+	__le32 rxe_frame_limit_overrun;
+	__le32 sent_ack_cnt;
+	__le32 sent_cts_cnt;
+	__le32 sent_ba_rsp_cnt;
+	__le32 dsp_self_kill;
+	__le32 mh_format_err;
+	__le32 re_acq_main_rssi_sum;
+	__le32 reserved;
+} __packed; /* STATISTICS_RX_PHY_API_S_VER_2 */
+
+struct mvm_statistics_rx_ht_phy {
+	__le32 plcp_err;
+	__le32 overrun_err;
+	__le32 early_overrun_err;
+	__le32 crc32_good;
+	__le32 crc32_err;
+	__le32 mh_format_err;
+	__le32 agg_crc32_good;
+	__le32 agg_mpdu_cnt;
+	__le32 agg_cnt;
+	__le32 unsupport_mcs;
+} __packed;  /* STATISTICS_HT_RX_PHY_API_S_VER_1 */
+
+#define MAX_CHAINS 3
+
+struct mvm_statistics_tx_non_phy_agg {
+	__le32 ba_timeout;
+	__le32 ba_reschedule_frames;
+	__le32 scd_query_agg_frame_cnt;
+	__le32 scd_query_no_agg;
+	__le32 scd_query_agg;
+	__le32 scd_query_mismatch;
+	__le32 frame_not_ready;
+	__le32 underrun;
+	__le32 bt_prio_kill;
+	__le32 rx_ba_rsp_cnt;
+	__s8 txpower[MAX_CHAINS];
+	__s8 reserved;
+	__le32 reserved2;
+} __packed; /* STATISTICS_TX_NON_PHY_AGG_API_S_VER_1 */
+
+struct mvm_statistics_tx_channel_width {
+	__le32 ext_cca_narrow_ch20[1];
+	__le32 ext_cca_narrow_ch40[2];
+	__le32 ext_cca_narrow_ch80[3];
+	__le32 ext_cca_narrow_ch160[4];
+	__le32 last_tx_ch_width_indx;
+	__le32 rx_detected_per_ch_width[4];
+	__le32 success_per_ch_width[4];
+	__le32 fail_per_ch_width[4];
+}; /* STATISTICS_TX_CHANNEL_WIDTH_API_S_VER_1 */
+
+struct mvm_statistics_tx {
+	__le32 preamble_cnt;
+	__le32 rx_detected_cnt;
+	__le32 bt_prio_defer_cnt;
+	__le32 bt_prio_kill_cnt;
+	__le32 few_bytes_cnt;
+	__le32 cts_timeout;
+	__le32 ack_timeout;
+	__le32 expected_ack_cnt;
+	__le32 actual_ack_cnt;
+	__le32 dump_msdu_cnt;
+	__le32 burst_abort_next_frame_mismatch_cnt;
+	__le32 burst_abort_missing_next_frame_cnt;
+	__le32 cts_timeout_collision;
+	__le32 ack_or_ba_timeout_collision;
+	struct mvm_statistics_tx_non_phy_agg agg;
+	struct mvm_statistics_tx_channel_width channel_width;
+} __packed; /* STATISTICS_TX_API_S_VER_4 */
+
+
+struct mvm_statistics_bt_activity {
+	__le32 hi_priority_tx_req_cnt;
+	__le32 hi_priority_tx_denied_cnt;
+	__le32 lo_priority_tx_req_cnt;
+	__le32 lo_priority_tx_denied_cnt;
+	__le32 hi_priority_rx_req_cnt;
+	__le32 hi_priority_rx_denied_cnt;
+	__le32 lo_priority_rx_req_cnt;
+	__le32 lo_priority_rx_denied_cnt;
+} __packed;  /* STATISTICS_BT_ACTIVITY_API_S_VER_1 */
+
+struct mvm_statistics_general {
+	struct mvm_statistics_general_common common;
+	__le32 beacon_filtered;
+	__le32 missed_beacons;
+	__s8 beacon_filter_everage_energy;
+	__s8 beacon_filter_reason;
+	__s8 beacon_filter_current_energy;
+	__s8 beacon_filter_reserved;
+	__le32 beacon_filter_delta_time;
+	struct mvm_statistics_bt_activity bt_activity;
+} __packed; /* STATISTICS_GENERAL_API_S_VER_5 */
+
+struct mvm_statistics_rx {
+	struct mvm_statistics_rx_phy ofdm;
+	struct mvm_statistics_rx_phy cck;
+	struct mvm_statistics_rx_non_phy general;
+	struct mvm_statistics_rx_ht_phy ofdm_ht;
+} __packed; /* STATISTICS_RX_API_S_VER_3 */
+
+/*
+ * STATISTICS_NOTIFICATION = 0x9d (notification only, not a command)
+ *
+ * By default, uCode issues this notification after receiving a beacon
+ * while associated.  To disable this behavior, set DISABLE_NOTIF flag in the
+ * REPLY_STATISTICS_CMD 0x9c, above.
+ *
+ * Statistics counters continue to increment beacon after beacon, but are
+ * cleared when changing channels or when driver issues REPLY_STATISTICS_CMD
+ * 0x9c with CLEAR_STATS bit set (see above).
+ *
+ * uCode also issues this notification during scans.  uCode clears statistics
+ * appropriately so that each notification contains statistics for only the
+ * one channel that has just been scanned.
+ */
+
+struct iwl_notif_statistics { /* STATISTICS_NTFY_API_S_VER_8 */
+	__le32 flag;
+	struct mvm_statistics_rx rx;
+	struct mvm_statistics_tx tx;
+	struct mvm_statistics_general general;
+} __packed;
+
 #endif /* __fw_api_h__ */
diff --git a/drivers/net/wireless/iwlwifi/mvm/fw.c b/drivers/net/wireless/iwlwifi/mvm/fw.c
index e18c92d..cd7c003 100644
--- a/drivers/net/wireless/iwlwifi/mvm/fw.c
+++ b/drivers/net/wireless/iwlwifi/mvm/fw.c
@@ -326,6 +326,17 @@ int iwl_run_init_mvm_ucode(struct iwl_mvm *mvm, bool read_nvm)
 	ret = iwl_nvm_check_version(mvm->nvm_data, mvm->trans);
 	WARN_ON(ret);
 
+	/*
+	 * abort after reading the nvm in case RF Kill is on, we will complete
+	 * the init seq later when RF kill will switch to off
+	 */
+	if (iwl_mvm_is_radio_killed(mvm)) {
+		IWL_DEBUG_RF_KILL(mvm,
+				  "jump over all phy activities due to RF kill\n");
+		iwl_remove_notification(&mvm->notif_wait, &calib_wait);
+		return 1;
+	}
+
 	/* Send TX valid antennas before triggering calibrations */
 	ret = iwl_send_tx_ant_cfg(mvm, iwl_fw_valid_tx_ant(mvm->fw));
 	if (ret)
@@ -388,6 +399,8 @@ out:
 int iwl_mvm_up(struct iwl_mvm *mvm)
 {
 	int ret, i;
+	struct ieee80211_channel *chan;
+	struct cfg80211_chan_def chandef;
 
 	lockdep_assert_held(&mvm->mutex);
 
@@ -400,8 +413,16 @@ int iwl_mvm_up(struct iwl_mvm *mvm)
 		ret = iwl_run_init_mvm_ucode(mvm, false);
 		if (ret && !iwlmvm_mod_params.init_dbg) {
 			IWL_ERR(mvm, "Failed to run INIT ucode: %d\n", ret);
+			/* this can't happen */
+			if (WARN_ON(ret > 0))
+				ret = -ERFKILL;
 			goto error;
 		}
+		/* should stop & start HW since that INIT image just loaded */
+		iwl_trans_stop_hw(mvm->trans, false);
+		ret = iwl_trans_start_hw(mvm->trans);
+		if (ret)
+			return ret;
 	}
 
 	if (iwlmvm_mod_params.init_dbg)
@@ -443,8 +464,22 @@ int iwl_mvm_up(struct iwl_mvm *mvm)
 	if (ret)
 		goto error;
 
-	IWL_DEBUG_INFO(mvm, "RT uCode started.\n");
+	/* Add all the PHY contexts */
+	chan = &mvm->hw->wiphy->bands[IEEE80211_BAND_2GHZ]->channels[0];
+	cfg80211_chandef_create(&chandef, chan, NL80211_CHAN_NO_HT);
+	for (i = 0; i < NUM_PHY_CTX; i++) {
+		/*
+		 * The channel used here isn't relevant as it's
+		 * going to be overwritten in the other flows.
+		 * For now use the first channel we have.
+		 */
+		ret = iwl_mvm_phy_ctxt_add(mvm, &mvm->phy_ctxts[i],
+					   &chandef, 1, 1);
+		if (ret)
+			goto error;
+	}
 
+	IWL_DEBUG_INFO(mvm, "RT uCode started.\n");
 	return 0;
  error:
 	iwl_trans_stop_device(mvm->trans);
diff --git a/drivers/net/wireless/iwlwifi/mvm/mac-ctxt.c b/drivers/net/wireless/iwlwifi/mvm/mac-ctxt.c
index b2cc3d9..94aae9c 100644
--- a/drivers/net/wireless/iwlwifi/mvm/mac-ctxt.c
+++ b/drivers/net/wireless/iwlwifi/mvm/mac-ctxt.c
@@ -193,14 +193,11 @@ static void iwl_mvm_mac_iface_iterator(void *_data, u8 *mac,
 u32 iwl_mvm_mac_get_queues_mask(struct iwl_mvm *mvm,
 				struct ieee80211_vif *vif)
 {
-	u32 qmask, ac;
+	u32 qmask = 0, ac;
 
 	if (vif->type == NL80211_IFTYPE_P2P_DEVICE)
 		return BIT(IWL_MVM_OFFCHANNEL_QUEUE);
 
-	qmask = (vif->cab_queue != IEEE80211_INVAL_HW_QUEUE) ?
-		BIT(vif->cab_queue) : 0;
-
 	for (ac = 0; ac < IEEE80211_NUM_ACS; ac++)
 		if (vif->hw_queue[ac] != IEEE80211_INVAL_HW_QUEUE)
 			qmask |= BIT(vif->hw_queue[ac]);
@@ -227,7 +224,7 @@ static int iwl_mvm_mac_ctxt_allocate_resources(struct iwl_mvm *mvm,
 		.found_vif = false,
 	};
 	u32 ac;
-	int ret;
+	int ret, i;
 
 	/*
 	 * Allocate a MAC ID and a TSF for this MAC, along with the queues
@@ -335,6 +332,9 @@ static int iwl_mvm_mac_ctxt_allocate_resources(struct iwl_mvm *mvm,
 	mvmvif->bcast_sta.sta_id = IWL_MVM_STATION_COUNT;
 	mvmvif->ap_sta_id = IWL_MVM_STATION_COUNT;
 
+	for (i = 0; i < NUM_IWL_MVM_SMPS_REQ; i++)
+		mvmvif->smps_requests[i] = IEEE80211_SMPS_AUTOMATIC;
+
 	return 0;
 
 exit_fail:
@@ -362,7 +362,7 @@ int iwl_mvm_mac_ctxt_init(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 		break;
 	case NL80211_IFTYPE_AP:
 		iwl_trans_ac_txq_enable(mvm->trans, vif->cab_queue,
-					IWL_MVM_TX_FIFO_VO);
+					IWL_MVM_TX_FIFO_MCAST);
 		/* fall through */
 	default:
 		for (ac = 0; ac < IEEE80211_NUM_ACS; ac++)
@@ -550,6 +550,10 @@ static void iwl_mvm_mac_ctxt_cmd_common(struct iwl_mvm *mvm,
 		cmd->ac[i].fifos_mask = BIT(iwl_mvm_ac_to_tx_fifo[i]);
 	}
 
+	/* in AP mode, the MCAST FIFO takes the EDCA params from VO */
+	if (vif->type == NL80211_IFTYPE_AP)
+		cmd->ac[AC_VO].fifos_mask |= BIT(IWL_MVM_TX_FIFO_MCAST);
+
 	if (vif->bss_conf.qos)
 		cmd->qos_flags |= cpu_to_le32(MAC_QOS_FLG_UPDATE_EDCA);
 
@@ -861,6 +865,30 @@ int iwl_mvm_mac_ctxt_beacon_changed(struct iwl_mvm *mvm,
 	return ret;
 }
 
+struct iwl_mvm_mac_ap_iterator_data {
+	struct iwl_mvm *mvm;
+	struct ieee80211_vif *vif;
+	u32 beacon_device_ts;
+	u16 beacon_int;
+};
+
+/* Find the beacon_device_ts and beacon_int for a managed interface */
+static void iwl_mvm_mac_ap_iterator(void *_data, u8 *mac,
+				    struct ieee80211_vif *vif)
+{
+	struct iwl_mvm_mac_ap_iterator_data *data = _data;
+
+	if (vif->type != NL80211_IFTYPE_STATION || !vif->bss_conf.assoc)
+		return;
+
+	/* Station client has higher priority over P2P client*/
+	if (vif->p2p && data->beacon_device_ts)
+		return;
+
+	data->beacon_device_ts = vif->bss_conf.sync_device_ts;
+	data->beacon_int = vif->bss_conf.beacon_int;
+}
+
 /*
  * Fill the specific data for mac context of type AP of P2P GO
  */
@@ -870,6 +898,11 @@ static void iwl_mvm_mac_ctxt_cmd_fill_ap(struct iwl_mvm *mvm,
 					 bool add)
 {
 	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_mvm_mac_ap_iterator_data data = {
+		.mvm = mvm,
+		.vif = vif,
+		.beacon_device_ts = 0
+	};
 
 	ctxt_ap->bi = cpu_to_le32(vif->bss_conf.beacon_int);
 	ctxt_ap->bi_reciprocal =
@@ -883,16 +916,33 @@ static void iwl_mvm_mac_ctxt_cmd_fill_ap(struct iwl_mvm *mvm,
 	ctxt_ap->mcast_qid = cpu_to_le32(vif->cab_queue);
 
 	/*
-	 * Only read the system time when the MAC is being added, when we
+	 * Only set the beacon time when the MAC is being added, when we
 	 * just modify the MAC then we should keep the time -- the firmware
 	 * can otherwise have a "jumping" TBTT.
 	 */
-	if (add)
-		mvmvif->ap_beacon_time =
-			iwl_read_prph(mvm->trans, DEVICE_SYSTEM_TIME_REG);
+	if (add) {
+		/*
+		 * If there is a station/P2P client interface which is
+		 * associated, set the AP's TBTT far enough from the station's
+		 * TBTT. Otherwise, set it to the current system time
+		 */
+		ieee80211_iterate_active_interfaces_atomic(
+			mvm->hw, IEEE80211_IFACE_ITER_RESUME_ALL,
+			iwl_mvm_mac_ap_iterator, &data);
+
+		if (data.beacon_device_ts) {
+			u32 rand = (prandom_u32() % (80 - 20)) + 20;
+			mvmvif->ap_beacon_time = data.beacon_device_ts +
+				ieee80211_tu_to_usec(data.beacon_int * rand /
+						     100);
+		} else {
+			mvmvif->ap_beacon_time =
+				iwl_read_prph(mvm->trans,
+					      DEVICE_SYSTEM_TIME_REG);
+		}
+	}
 
 	ctxt_ap->beacon_time = cpu_to_le32(mvmvif->ap_beacon_time);
-
 	ctxt_ap->beacon_tsf = 0; /* unused */
 
 	/* TODO: Assume that the beacon id == mac context id */
@@ -1047,3 +1097,28 @@ int iwl_mvm_rx_beacon_notif(struct iwl_mvm *mvm,
 		     rate);
 	return 0;
 }
+
+static void iwl_mvm_beacon_loss_iterator(void *_data, u8 *mac,
+					 struct ieee80211_vif *vif)
+{
+	u16 *id = _data;
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+
+	if (mvmvif->id == *id)
+		ieee80211_beacon_loss(vif);
+}
+
+int iwl_mvm_rx_missed_beacons_notif(struct iwl_mvm *mvm,
+				    struct iwl_rx_cmd_buffer *rxb,
+				    struct iwl_device_cmd *cmd)
+{
+	struct iwl_rx_packet *pkt = rxb_addr(rxb);
+	struct iwl_missed_beacons_notif *missed_beacons = (void *)pkt->data;
+	u16 id = (u16)le32_to_cpu(missed_beacons->mac_id);
+
+	ieee80211_iterate_active_interfaces_atomic(mvm->hw,
+						   IEEE80211_IFACE_ITER_NORMAL,
+						   iwl_mvm_beacon_loss_iterator,
+						   &id);
+	return 0;
+}
diff --git a/drivers/net/wireless/iwlwifi/mvm/mac80211.c b/drivers/net/wireless/iwlwifi/mvm/mac80211.c
index a5eb8c8..e08683b 100644
--- a/drivers/net/wireless/iwlwifi/mvm/mac80211.c
+++ b/drivers/net/wireless/iwlwifi/mvm/mac80211.c
@@ -81,12 +81,12 @@
 static const struct ieee80211_iface_limit iwl_mvm_limits[] = {
 	{
 		.max = 1,
-		.types = BIT(NL80211_IFTYPE_STATION) |
-			BIT(NL80211_IFTYPE_AP),
+		.types = BIT(NL80211_IFTYPE_STATION),
 	},
 	{
 		.max = 1,
-		.types = BIT(NL80211_IFTYPE_P2P_CLIENT) |
+		.types = BIT(NL80211_IFTYPE_AP) |
+			BIT(NL80211_IFTYPE_P2P_CLIENT) |
 			BIT(NL80211_IFTYPE_P2P_GO),
 	},
 	{
@@ -127,6 +127,17 @@ static const struct wiphy_wowlan_tcp_support iwl_mvm_wowlan_tcp_support = {
 };
 #endif
 
+static void iwl_mvm_reset_phy_ctxts(struct iwl_mvm *mvm)
+{
+	int i;
+
+	memset(mvm->phy_ctxts, 0, sizeof(mvm->phy_ctxts));
+	for (i = 0; i < NUM_PHY_CTX; i++) {
+		mvm->phy_ctxts[i].id = i;
+		mvm->phy_ctxts[i].ref = 0;
+	}
+}
+
 int iwl_mvm_mac_setup_register(struct iwl_mvm *mvm)
 {
 	struct ieee80211_hw *hw = mvm->hw;
@@ -141,7 +152,8 @@ int iwl_mvm_mac_setup_register(struct iwl_mvm *mvm)
 		    IEEE80211_HW_SUPPORTS_PS |
 		    IEEE80211_HW_SUPPORTS_DYNAMIC_PS |
 		    IEEE80211_HW_AMPDU_AGGREGATION |
-		    IEEE80211_HW_TIMING_BEACON_ONLY;
+		    IEEE80211_HW_TIMING_BEACON_ONLY |
+		    IEEE80211_HW_CONNECTION_MONITOR;
 
 	hw->queues = IWL_MVM_FIRST_AGG_QUEUE;
 	hw->offchannel_tx_hw_queue = IWL_MVM_OFFCHANNEL_QUEUE;
@@ -158,7 +170,7 @@ int iwl_mvm_mac_setup_register(struct iwl_mvm *mvm)
 
 	hw->sta_data_size = sizeof(struct iwl_mvm_sta);
 	hw->vif_data_size = sizeof(struct iwl_mvm_vif);
-	hw->chanctx_data_size = sizeof(struct iwl_mvm_phy_ctxt);
+	hw->chanctx_data_size = sizeof(u16);
 
 	hw->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION) |
 		BIT(NL80211_IFTYPE_P2P_CLIENT) |
@@ -193,6 +205,8 @@ int iwl_mvm_mac_setup_register(struct iwl_mvm *mvm)
 		hw->wiphy->n_addresses++;
 	}
 
+	iwl_mvm_reset_phy_ctxts(mvm);
+
 	/* we create the 802.11 header and a max-length SSID element */
 	hw->wiphy->max_scan_ie_len =
 		mvm->fw->ucode_capa.max_probe_length - 24 - 34;
@@ -222,20 +236,20 @@ int iwl_mvm_mac_setup_register(struct iwl_mvm *mvm)
 	    mvm->trans->ops->d3_suspend &&
 	    mvm->trans->ops->d3_resume &&
 	    device_can_wakeup(mvm->trans->dev)) {
-		hw->wiphy->wowlan.flags = WIPHY_WOWLAN_MAGIC_PKT |
-					  WIPHY_WOWLAN_DISCONNECT |
-					  WIPHY_WOWLAN_EAP_IDENTITY_REQ |
-					  WIPHY_WOWLAN_RFKILL_RELEASE;
+		mvm->wowlan.flags = WIPHY_WOWLAN_MAGIC_PKT |
+				    WIPHY_WOWLAN_DISCONNECT |
+				    WIPHY_WOWLAN_EAP_IDENTITY_REQ |
+				    WIPHY_WOWLAN_RFKILL_RELEASE;
 		if (!iwlwifi_mod_params.sw_crypto)
-			hw->wiphy->wowlan.flags |=
-				WIPHY_WOWLAN_SUPPORTS_GTK_REKEY |
-				WIPHY_WOWLAN_GTK_REKEY_FAILURE |
-				WIPHY_WOWLAN_4WAY_HANDSHAKE;
-
-		hw->wiphy->wowlan.n_patterns = IWL_WOWLAN_MAX_PATTERNS;
-		hw->wiphy->wowlan.pattern_min_len = IWL_WOWLAN_MIN_PATTERN_LEN;
-		hw->wiphy->wowlan.pattern_max_len = IWL_WOWLAN_MAX_PATTERN_LEN;
-		hw->wiphy->wowlan.tcp = &iwl_mvm_wowlan_tcp_support;
+			mvm->wowlan.flags |= WIPHY_WOWLAN_SUPPORTS_GTK_REKEY |
+					     WIPHY_WOWLAN_GTK_REKEY_FAILURE |
+					     WIPHY_WOWLAN_4WAY_HANDSHAKE;
+
+		mvm->wowlan.n_patterns = IWL_WOWLAN_MAX_PATTERNS;
+		mvm->wowlan.pattern_min_len = IWL_WOWLAN_MIN_PATTERN_LEN;
+		mvm->wowlan.pattern_max_len = IWL_WOWLAN_MAX_PATTERN_LEN;
+		mvm->wowlan.tcp = &iwl_mvm_wowlan_tcp_support;
+		hw->wiphy->wowlan = &mvm->wowlan;
 	}
 #endif
 
@@ -252,8 +266,8 @@ static void iwl_mvm_mac_tx(struct ieee80211_hw *hw,
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
 
-	if (test_bit(IWL_MVM_STATUS_HW_RFKILL, &mvm->status)) {
-		IWL_DEBUG_DROP(mvm, "Dropping - RF KILL\n");
+	if (iwl_mvm_is_radio_killed(mvm)) {
+		IWL_DEBUG_DROP(mvm, "Dropping - RF/CT KILL\n");
 		goto drop;
 	}
 
@@ -345,8 +359,7 @@ static void iwl_mvm_cleanup_iterator(void *data, u8 *mac,
 	iwl_mvm_te_clear_data(mvm, &mvmvif->time_event_data);
 	spin_unlock_bh(&mvm->time_event_lock);
 
-	if (vif->type != NL80211_IFTYPE_P2P_DEVICE)
-		mvmvif->phy_ctxt = NULL;
+	mvmvif->phy_ctxt = NULL;
 }
 
 static void iwl_mvm_restart_cleanup(struct iwl_mvm *mvm)
@@ -363,6 +376,9 @@ static void iwl_mvm_restart_cleanup(struct iwl_mvm *mvm)
 		mvm->hw, IEEE80211_IFACE_ITER_RESUME_ALL,
 		iwl_mvm_cleanup_iterator, mvm);
 
+	mvm->p2p_device_vif = NULL;
+
+	iwl_mvm_reset_phy_ctxts(mvm);
 	memset(mvm->fw_key_table, 0, sizeof(mvm->fw_key_table));
 	memset(mvm->sta_drained, 0, sizeof(mvm->sta_drained));
 
@@ -456,6 +472,20 @@ static void iwl_mvm_power_update_iterator(void *data, u8 *mac,
 	iwl_mvm_power_update_mode(mvm, vif);
 }
 
+static struct iwl_mvm_phy_ctxt *iwl_mvm_get_free_phy_ctxt(struct iwl_mvm *mvm)
+{
+	u16 i;
+
+	lockdep_assert_held(&mvm->mutex);
+
+	for (i = 0; i < NUM_PHY_CTX; i++)
+		if (!mvm->phy_ctxts[i].ref)
+			return &mvm->phy_ctxts[i];
+
+	IWL_ERR(mvm, "No available PHY context\n");
+	return NULL;
+}
+
 static int iwl_mvm_mac_add_interface(struct ieee80211_hw *hw,
 				     struct ieee80211_vif *vif)
 {
@@ -530,32 +560,34 @@ static int iwl_mvm_mac_add_interface(struct ieee80211_hw *hw,
 	 */
 	iwl_mvm_power_update_mode(mvm, vif);
 
+	/* beacon filtering */
+	if (!mvm->bf_allowed_vif &&
+	    vif->type == NL80211_IFTYPE_STATION && !vif->p2p){
+		mvm->bf_allowed_vif = mvmvif;
+		vif->driver_flags |= IEEE80211_VIF_BEACON_FILTER;
+	}
+
+	ret = iwl_mvm_disable_beacon_filter(mvm, vif);
+	if (ret)
+		goto out_release;
+
 	/*
 	 * P2P_DEVICE interface does not have a channel context assigned to it,
 	 * so a dedicated PHY context is allocated to it and the corresponding
 	 * MAC context is bound to it at this stage.
 	 */
 	if (vif->type == NL80211_IFTYPE_P2P_DEVICE) {
-		struct ieee80211_channel *chan;
-		struct cfg80211_chan_def chandef;
 
-		mvmvif->phy_ctxt = &mvm->phy_ctxt_roc;
-
-		/*
-		 * The channel used here isn't relevant as it's
-		 * going to be overwritten as part of the ROC flow.
-		 * For now use the first channel we have.
-		 */
-		chan = &mvm->hw->wiphy->bands[IEEE80211_BAND_2GHZ]->channels[0];
-		cfg80211_chandef_create(&chandef, chan, NL80211_CHAN_NO_HT);
-		ret = iwl_mvm_phy_ctxt_add(mvm, mvmvif->phy_ctxt,
-					   &chandef, 1, 1);
-		if (ret)
+		mvmvif->phy_ctxt = iwl_mvm_get_free_phy_ctxt(mvm);
+		if (!mvmvif->phy_ctxt) {
+			ret = -ENOSPC;
 			goto out_remove_mac;
+		}
 
+		iwl_mvm_phy_ctxt_ref(mvm, mvmvif->phy_ctxt);
 		ret = iwl_mvm_binding_add_vif(mvm, vif);
 		if (ret)
-			goto out_remove_phy;
+			goto out_unref_phy;
 
 		ret = iwl_mvm_add_bcast_sta(mvm, vif, &mvmvif->bcast_sta);
 		if (ret)
@@ -571,27 +603,17 @@ static int iwl_mvm_mac_add_interface(struct ieee80211_hw *hw,
 
  out_unbind:
 	iwl_mvm_binding_remove_vif(mvm, vif);
- out_remove_phy:
-	iwl_mvm_phy_ctxt_remove(mvm, mvmvif->phy_ctxt);
+ out_unref_phy:
+	iwl_mvm_phy_ctxt_unref(mvm, mvmvif->phy_ctxt);
  out_remove_mac:
 	mvmvif->phy_ctxt = NULL;
 	iwl_mvm_mac_ctxt_remove(mvm, vif);
  out_release:
-	/*
-	 * TODO: remove this temporary code.
-	 * Currently MVM FW supports power management only on single MAC.
-	 * Check if only one additional interface remains after releasing
-	 * current one. Update power mode on the remaining interface.
-	 */
 	if (vif->type != NL80211_IFTYPE_P2P_DEVICE)
 		mvm->vif_count--;
-	IWL_DEBUG_MAC80211(mvm, "Currently %d interfaces active\n",
-			   mvm->vif_count);
-	if (mvm->vif_count == 1) {
-		ieee80211_iterate_active_interfaces(
-					mvm->hw, IEEE80211_IFACE_ITER_NORMAL,
-					iwl_mvm_power_update_iterator, mvm);
-	}
+	ieee80211_iterate_active_interfaces(
+		mvm->hw, IEEE80211_IFACE_ITER_NORMAL,
+		iwl_mvm_power_update_iterator, mvm);
 	iwl_mvm_mac_ctxt_release(mvm, vif);
  out_unlock:
 	mutex_unlock(&mvm->mutex);
@@ -629,8 +651,7 @@ static void iwl_mvm_prepare_mac_removal(struct iwl_mvm *mvm,
 		 * By now, all the AC queues are empty. The AGG queues are
 		 * empty too. We already got all the Tx responses for all the
 		 * packets in the queues. The drain work can have been
-		 * triggered. Flush it. This work item takes the mutex, so kill
-		 * it before we take it.
+		 * triggered. Flush it.
 		 */
 		flush_work(&mvm->sta_drained_wk);
 	}
@@ -646,6 +667,11 @@ static void iwl_mvm_mac_remove_interface(struct ieee80211_hw *hw,
 
 	mutex_lock(&mvm->mutex);
 
+	if (mvm->bf_allowed_vif == mvmvif) {
+		mvm->bf_allowed_vif = NULL;
+		vif->driver_flags &= ~IEEE80211_VIF_BEACON_FILTER;
+	}
+
 	iwl_mvm_vif_dbgfs_clean(mvm, vif);
 
 	/*
@@ -661,7 +687,7 @@ static void iwl_mvm_mac_remove_interface(struct ieee80211_hw *hw,
 		mvm->p2p_device_vif = NULL;
 		iwl_mvm_rm_bcast_sta(mvm, &mvmvif->bcast_sta);
 		iwl_mvm_binding_remove_vif(mvm, vif);
-		iwl_mvm_phy_ctxt_remove(mvm, mvmvif->phy_ctxt);
+		iwl_mvm_phy_ctxt_unref(mvm, mvmvif->phy_ctxt);
 		mvmvif->phy_ctxt = NULL;
 	}
 
@@ -748,7 +774,10 @@ static void iwl_mvm_bss_info_changed_station(struct iwl_mvm *mvm,
 			if (ret)
 				IWL_ERR(mvm, "failed to update quotas\n");
 		}
-	} else if (changes & BSS_CHANGED_DTIM_PERIOD) {
+		ret = iwl_mvm_power_update_mode(mvm, vif);
+		if (ret)
+			IWL_ERR(mvm, "failed to update power mode\n");
+	} else if (changes & BSS_CHANGED_BEACON_INFO) {
 		/*
 		 * We received a beacon _after_ association so
 		 * remove the session protection.
@@ -756,19 +785,9 @@ static void iwl_mvm_bss_info_changed_station(struct iwl_mvm *mvm,
 		iwl_mvm_remove_time_event(mvm, mvmvif,
 					  &mvmvif->time_event_data);
 	} else if (changes & BSS_CHANGED_PS) {
-		/*
-		 * TODO: remove this temporary code.
-		 * Currently MVM FW supports power management only on single
-		 * MAC. Avoid power mode update if more than one interface
-		 * is active.
-		 */
-		IWL_DEBUG_MAC80211(mvm, "Currently %d interfaces active\n",
-				   mvm->vif_count);
-		if (mvm->vif_count == 1) {
-			ret = iwl_mvm_power_update_mode(mvm, vif);
-			if (ret)
-				IWL_ERR(mvm, "failed to update power mode\n");
-		}
+		ret = iwl_mvm_power_update_mode(mvm, vif);
+		if (ret)
+			IWL_ERR(mvm, "failed to update power mode\n");
 	}
 }
 
@@ -999,9 +1018,13 @@ static int iwl_mvm_mac_sta_state(struct ieee80211_hw *hw,
 					     mvmvif->phy_ctxt->channel->band);
 	} else if (old_state == IEEE80211_STA_ASSOC &&
 		   new_state == IEEE80211_STA_AUTHORIZED) {
+		/* enable beacon filtering */
+		WARN_ON(iwl_mvm_enable_beacon_filter(mvm, vif));
 		ret = 0;
 	} else if (old_state == IEEE80211_STA_AUTHORIZED &&
 		   new_state == IEEE80211_STA_ASSOC) {
+		/* disable beacon filtering */
+		WARN_ON(iwl_mvm_disable_beacon_filter(mvm, vif));
 		ret = 0;
 	} else if (old_state == IEEE80211_STA_ASSOC &&
 		   new_state == IEEE80211_STA_AUTH) {
@@ -1167,29 +1190,107 @@ static int iwl_mvm_roc(struct ieee80211_hw *hw,
 		       enum ieee80211_roc_type type)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
 	struct cfg80211_chan_def chandef;
-	int ret;
+	struct iwl_mvm_phy_ctxt *phy_ctxt;
+	int ret, i;
+
+	IWL_DEBUG_MAC80211(mvm, "enter (%d, %d, %d)\n", channel->hw_value,
+			   duration, type);
 
 	if (vif->type != NL80211_IFTYPE_P2P_DEVICE) {
 		IWL_ERR(mvm, "vif isn't a P2P_DEVICE: %d\n", vif->type);
 		return -EINVAL;
 	}
 
-	IWL_DEBUG_MAC80211(mvm, "enter (%d, %d, %d)\n", channel->hw_value,
-			   duration, type);
-
 	mutex_lock(&mvm->mutex);
 
+	for (i = 0; i < NUM_PHY_CTX; i++) {
+		phy_ctxt = &mvm->phy_ctxts[i];
+		if (phy_ctxt->ref == 0 || mvmvif->phy_ctxt == phy_ctxt)
+			continue;
+
+		if (phy_ctxt->ref && channel == phy_ctxt->channel) {
+			/*
+			 * Unbind the P2P_DEVICE from the current PHY context,
+			 * and if the PHY context is not used remove it.
+			 */
+			ret = iwl_mvm_binding_remove_vif(mvm, vif);
+			if (WARN(ret, "Failed unbinding P2P_DEVICE\n"))
+				goto out_unlock;
+
+			iwl_mvm_phy_ctxt_unref(mvm, mvmvif->phy_ctxt);
+
+			/* Bind the P2P_DEVICE to the current PHY Context */
+			mvmvif->phy_ctxt = phy_ctxt;
+
+			ret = iwl_mvm_binding_add_vif(mvm, vif);
+			if (WARN(ret, "Failed binding P2P_DEVICE\n"))
+				goto out_unlock;
+
+			iwl_mvm_phy_ctxt_ref(mvm, mvmvif->phy_ctxt);
+			goto schedule_time_event;
+		}
+	}
+
+	/* Need to update the PHY context only if the ROC channel changed */
+	if (channel == mvmvif->phy_ctxt->channel)
+		goto schedule_time_event;
+
 	cfg80211_chandef_create(&chandef, channel, NL80211_CHAN_NO_HT);
-	ret = iwl_mvm_phy_ctxt_changed(mvm, &mvm->phy_ctxt_roc,
-				       &chandef, 1, 1);
 
+	/*
+	 * Change the PHY context configuration as it is currently referenced
+	 * only by the P2P Device MAC
+	 */
+	if (mvmvif->phy_ctxt->ref == 1) {
+		ret = iwl_mvm_phy_ctxt_changed(mvm, mvmvif->phy_ctxt,
+					       &chandef, 1, 1);
+		if (ret)
+			goto out_unlock;
+	} else {
+		/*
+		 * The PHY context is shared with other MACs. Need to remove the
+		 * P2P Device from the binding, allocate an new PHY context and
+		 * create a new binding
+		 */
+		phy_ctxt = iwl_mvm_get_free_phy_ctxt(mvm);
+		if (!phy_ctxt) {
+			ret = -ENOSPC;
+			goto out_unlock;
+		}
+
+		ret = iwl_mvm_phy_ctxt_changed(mvm, phy_ctxt, &chandef,
+					       1, 1);
+		if (ret) {
+			IWL_ERR(mvm, "Failed to change PHY context\n");
+			goto out_unlock;
+		}
+
+		/* Unbind the P2P_DEVICE from the current PHY context */
+		ret = iwl_mvm_binding_remove_vif(mvm, vif);
+		if (WARN(ret, "Failed unbinding P2P_DEVICE\n"))
+			goto out_unlock;
+
+		iwl_mvm_phy_ctxt_unref(mvm, mvmvif->phy_ctxt);
+
+		/* Bind the P2P_DEVICE to the new allocated PHY context */
+		mvmvif->phy_ctxt = phy_ctxt;
+
+		ret = iwl_mvm_binding_add_vif(mvm, vif);
+		if (WARN(ret, "Failed binding P2P_DEVICE\n"))
+			goto out_unlock;
+
+		iwl_mvm_phy_ctxt_ref(mvm, mvmvif->phy_ctxt);
+	}
+
+schedule_time_event:
 	/* Schedule the time events */
 	ret = iwl_mvm_start_p2p_roc(mvm, vif, duration, type);
 
+out_unlock:
 	mutex_unlock(&mvm->mutex);
 	IWL_DEBUG_MAC80211(mvm, "leave\n");
-
 	return ret;
 }
 
@@ -1211,15 +1312,30 @@ static int iwl_mvm_add_chanctx(struct ieee80211_hw *hw,
 			       struct ieee80211_chanctx_conf *ctx)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
-	struct iwl_mvm_phy_ctxt *phy_ctxt = (void *)ctx->drv_priv;
+	u16 *phy_ctxt_id = (u16 *)ctx->drv_priv;
+	struct iwl_mvm_phy_ctxt *phy_ctxt;
 	int ret;
 
+	IWL_DEBUG_MAC80211(mvm, "Add channel context\n");
+
 	mutex_lock(&mvm->mutex);
+	phy_ctxt = iwl_mvm_get_free_phy_ctxt(mvm);
+	if (!phy_ctxt) {
+		ret = -ENOSPC;
+		goto out;
+	}
 
-	IWL_DEBUG_MAC80211(mvm, "Add PHY context\n");
-	ret = iwl_mvm_phy_ctxt_add(mvm, phy_ctxt, &ctx->def,
-				   ctx->rx_chains_static,
-				   ctx->rx_chains_dynamic);
+	ret = iwl_mvm_phy_ctxt_changed(mvm, phy_ctxt, &ctx->def,
+				       ctx->rx_chains_static,
+				       ctx->rx_chains_dynamic);
+	if (ret) {
+		IWL_ERR(mvm, "Failed to add PHY context\n");
+		goto out;
+	}
+
+	iwl_mvm_phy_ctxt_ref(mvm, phy_ctxt);
+	*phy_ctxt_id = phy_ctxt->id;
+out:
 	mutex_unlock(&mvm->mutex);
 	return ret;
 }
@@ -1228,10 +1344,11 @@ static void iwl_mvm_remove_chanctx(struct ieee80211_hw *hw,
 				   struct ieee80211_chanctx_conf *ctx)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
-	struct iwl_mvm_phy_ctxt *phy_ctxt = (void *)ctx->drv_priv;
+	u16 *phy_ctxt_id = (u16 *)ctx->drv_priv;
+	struct iwl_mvm_phy_ctxt *phy_ctxt = &mvm->phy_ctxts[*phy_ctxt_id];
 
 	mutex_lock(&mvm->mutex);
-	iwl_mvm_phy_ctxt_remove(mvm, phy_ctxt);
+	iwl_mvm_phy_ctxt_unref(mvm, phy_ctxt);
 	mutex_unlock(&mvm->mutex);
 }
 
@@ -1240,7 +1357,16 @@ static void iwl_mvm_change_chanctx(struct ieee80211_hw *hw,
 				   u32 changed)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
-	struct iwl_mvm_phy_ctxt *phy_ctxt = (void *)ctx->drv_priv;
+	u16 *phy_ctxt_id = (u16 *)ctx->drv_priv;
+	struct iwl_mvm_phy_ctxt *phy_ctxt = &mvm->phy_ctxts[*phy_ctxt_id];
+
+	if (WARN_ONCE((phy_ctxt->ref > 1) &&
+		      (changed & ~(IEEE80211_CHANCTX_CHANGE_WIDTH |
+				   IEEE80211_CHANCTX_CHANGE_RX_CHAINS |
+				   IEEE80211_CHANCTX_CHANGE_RADAR)),
+		      "Cannot change PHY. Ref=%d, changed=0x%X\n",
+		      phy_ctxt->ref, changed))
+		return;
 
 	mutex_lock(&mvm->mutex);
 	iwl_mvm_phy_ctxt_changed(mvm, phy_ctxt, &ctx->def,
@@ -1254,13 +1380,14 @@ static int iwl_mvm_assign_vif_chanctx(struct ieee80211_hw *hw,
 				      struct ieee80211_chanctx_conf *ctx)
 {
 	struct iwl_mvm *mvm = IWL_MAC80211_GET_MVM(hw);
-	struct iwl_mvm_phy_ctxt *phyctx = (void *)ctx->drv_priv;
+	u16 *phy_ctxt_id = (u16 *)ctx->drv_priv;
+	struct iwl_mvm_phy_ctxt *phy_ctxt = &mvm->phy_ctxts[*phy_ctxt_id];
 	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
 	int ret;
 
 	mutex_lock(&mvm->mutex);
 
-	mvmvif->phy_ctxt = phyctx;
+	mvmvif->phy_ctxt = phy_ctxt;
 
 	switch (vif->type) {
 	case NL80211_IFTYPE_AP:
diff --git a/drivers/net/wireless/iwlwifi/mvm/mvm.h b/drivers/net/wireless/iwlwifi/mvm/mvm.h
index 9f46b23..d40d7db 100644
--- a/drivers/net/wireless/iwlwifi/mvm/mvm.h
+++ b/drivers/net/wireless/iwlwifi/mvm/mvm.h
@@ -73,7 +73,6 @@
 #include "iwl-trans.h"
 #include "iwl-notif-wait.h"
 #include "iwl-eeprom-parse.h"
-#include "iwl-test.h"
 #include "iwl-trans.h"
 #include "sta.h"
 #include "fw-api.h"
@@ -88,6 +87,7 @@ enum iwl_mvm_tx_fifo {
 	IWL_MVM_TX_FIFO_BE,
 	IWL_MVM_TX_FIFO_VI,
 	IWL_MVM_TX_FIFO_VO,
+	IWL_MVM_TX_FIFO_MCAST = 5,
 };
 
 extern struct ieee80211_ops iwl_mvm_hw_ops;
@@ -109,6 +109,7 @@ extern struct iwl_mvm_mod_params iwlmvm_mod_params;
 struct iwl_mvm_phy_ctxt {
 	u16 id;
 	u16 color;
+	u32 ref;
 
 	/*
 	 * TODO: This should probably be removed. Currently here only for rate
@@ -149,6 +150,64 @@ enum iwl_power_scheme {
 
 #define IWL_CONN_MAX_LISTEN_INTERVAL	70
 
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+enum iwl_dbgfs_pm_mask {
+	MVM_DEBUGFS_PM_KEEP_ALIVE = BIT(0),
+	MVM_DEBUGFS_PM_SKIP_OVER_DTIM = BIT(1),
+	MVM_DEBUGFS_PM_SKIP_DTIM_PERIODS = BIT(2),
+	MVM_DEBUGFS_PM_RX_DATA_TIMEOUT = BIT(3),
+	MVM_DEBUGFS_PM_TX_DATA_TIMEOUT = BIT(4),
+	MVM_DEBUGFS_PM_DISABLE_POWER_OFF = BIT(5),
+	MVM_DEBUGFS_PM_LPRX_ENA = BIT(6),
+	MVM_DEBUGFS_PM_LPRX_RSSI_THRESHOLD = BIT(7),
+};
+
+struct iwl_dbgfs_pm {
+	u8 keep_alive_seconds;
+	u32 rx_data_timeout;
+	u32 tx_data_timeout;
+	bool skip_over_dtim;
+	u8 skip_dtim_periods;
+	bool disable_power_off;
+	bool lprx_ena;
+	u32 lprx_rssi_threshold;
+	int mask;
+};
+
+/* beacon filtering */
+
+enum iwl_dbgfs_bf_mask {
+	MVM_DEBUGFS_BF_ENERGY_DELTA = BIT(0),
+	MVM_DEBUGFS_BF_ROAMING_ENERGY_DELTA = BIT(1),
+	MVM_DEBUGFS_BF_ROAMING_STATE = BIT(2),
+	MVM_DEBUGFS_BF_TEMPERATURE_DELTA = BIT(3),
+	MVM_DEBUGFS_BF_ENABLE_BEACON_FILTER = BIT(4),
+	MVM_DEBUGFS_BF_DEBUG_FLAG = BIT(5),
+	MVM_DEBUGFS_BF_ESCAPE_TIMER = BIT(6),
+	MVM_DEBUGFS_BA_ESCAPE_TIMER = BIT(7),
+	MVM_DEBUGFS_BA_ENABLE_BEACON_ABORT = BIT(8),
+};
+
+struct iwl_dbgfs_bf {
+	u8 bf_energy_delta;
+	u8 bf_roaming_energy_delta;
+	u8 bf_roaming_state;
+	u8 bf_temperature_delta;
+	u8 bf_enable_beacon_filter;
+	u8 bf_debug_flag;
+	u32 bf_escape_timer;
+	u32 ba_escape_timer;
+	u8 ba_enable_beacon_abort;
+	int mask;
+};
+#endif
+
+enum iwl_mvm_smps_type_request {
+	IWL_MVM_SMPS_REQ_BT_COEX,
+	IWL_MVM_SMPS_REQ_TT,
+	NUM_IWL_MVM_SMPS_REQ,
+};
+
 /**
  * struct iwl_mvm_vif - data per Virtual Interface, it is a MAC context
  * @id: between 0 and 3
@@ -163,6 +222,8 @@ enum iwl_power_scheme {
  * @bcast_sta: station used for broadcast packets. Used by the following
  *  vifs: P2P_DEVICE, GO and AP.
  * @beacon_skb: the skb used to hold the AP/GO beacon template
+ * @smps_requests: the requests of of differents parts of the driver, regard
+	the desired smps mode.
  */
 struct iwl_mvm_vif {
 	u16 id;
@@ -172,6 +233,8 @@ struct iwl_mvm_vif {
 	bool uploaded;
 	bool ap_active;
 	bool monitor_active;
+	/* indicate whether beacon filtering is enabled */
+	bool bf_enabled;
 
 	u32 ap_beacon_time;
 
@@ -214,7 +277,11 @@ struct iwl_mvm_vif {
 	struct dentry *dbgfs_dir;
 	struct dentry *dbgfs_slink;
 	void *dbgfs_data;
+	struct iwl_dbgfs_pm dbgfs_pm;
+	struct iwl_dbgfs_bf dbgfs_bf;
 #endif
+
+	enum ieee80211_smps_mode smps_requests[NUM_IWL_MVM_SMPS_REQ];
 };
 
 static inline struct iwl_mvm_vif *
@@ -223,12 +290,6 @@ iwl_mvm_vif_from_mac80211(struct ieee80211_vif *vif)
 	return (void *)vif->drv_priv;
 }
 
-enum iwl_mvm_status {
-	IWL_MVM_STATUS_HW_RFKILL,
-	IWL_MVM_STATUS_ROC_RUNNING,
-	IWL_MVM_STATUS_IN_HW_RESTART,
-};
-
 enum iwl_scan_status {
 	IWL_MVM_SCAN_NONE,
 	IWL_MVM_SCAN_OS,
@@ -246,6 +307,65 @@ struct iwl_nvm_section {
 	const u8 *data;
 };
 
+/*
+ * Tx-backoff threshold
+ * @temperature: The threshold in Celsius
+ * @backoff: The tx-backoff in uSec
+ */
+struct iwl_tt_tx_backoff {
+	s32 temperature;
+	u32 backoff;
+};
+
+#define TT_TX_BACKOFF_SIZE 6
+
+/**
+ * struct iwl_tt_params - thermal throttling parameters
+ * @ct_kill_entry: CT Kill entry threshold
+ * @ct_kill_exit: CT Kill exit threshold
+ * @ct_kill_duration: The time  intervals (in uSec) in which the driver needs
+ *	to checks whether to exit CT Kill.
+ * @dynamic_smps_entry: Dynamic SMPS entry threshold
+ * @dynamic_smps_exit: Dynamic SMPS exit threshold
+ * @tx_protection_entry: TX protection entry threshold
+ * @tx_protection_exit: TX protection exit threshold
+ * @tx_backoff: Array of thresholds for tx-backoff , in ascending order.
+ * @support_ct_kill: Support CT Kill?
+ * @support_dynamic_smps: Support dynamic SMPS?
+ * @support_tx_protection: Support tx protection?
+ * @support_tx_backoff: Support tx-backoff?
+ */
+struct iwl_tt_params {
+	s32 ct_kill_entry;
+	s32 ct_kill_exit;
+	u32 ct_kill_duration;
+	s32 dynamic_smps_entry;
+	s32 dynamic_smps_exit;
+	s32 tx_protection_entry;
+	s32 tx_protection_exit;
+	struct iwl_tt_tx_backoff tx_backoff[TT_TX_BACKOFF_SIZE];
+	bool support_ct_kill;
+	bool support_dynamic_smps;
+	bool support_tx_protection;
+	bool support_tx_backoff;
+};
+
+/**
+ * struct iwl_mvm_tt_mgnt - Thermal Throttling Management structure
+ * @ct_kill_exit: worker to exit thermal kill
+ * @dynamic_smps: Is thermal throttling enabled dynamic_smps?
+ * @tx_backoff: The current thremal throttling tx backoff in uSec.
+ * @params: Parameters to configure the thermal throttling algorithm.
+ * @throttle: Is thermal throttling is active?
+ */
+struct iwl_mvm_tt_mgmt {
+	struct delayed_work ct_kill_exit;
+	bool dynamic_smps;
+	u32 tx_backoff;
+	const struct iwl_tt_params *params;
+	bool throttle;
+};
+
 struct iwl_mvm {
 	/* for logger access */
 	struct device *dev;
@@ -266,6 +386,12 @@ struct iwl_mvm {
 
 	unsigned long status;
 
+	/*
+	 * for beacon filtering -
+	 * currently only one interface can be supported
+	 */
+	struct iwl_mvm_vif *bf_allowed_vif;
+
 	enum iwl_ucode_type cur_ucode;
 	bool ucode_loaded;
 	bool init_ucode_run;
@@ -313,7 +439,7 @@ struct iwl_mvm {
 	bool prevent_power_down_d3;
 #endif
 
-	struct iwl_mvm_phy_ctxt phy_ctxt_roc;
+	struct iwl_mvm_phy_ctxt phy_ctxts[NUM_PHY_CTX];
 
 	struct list_head time_event_list;
 	spinlock_t time_event_lock;
@@ -337,12 +463,24 @@ struct iwl_mvm {
 	struct ieee80211_vif *p2p_device_vif;
 
 #ifdef CONFIG_PM_SLEEP
+	struct wiphy_wowlan_support wowlan;
 	int gtk_ivlen, gtk_icvlen, ptk_ivlen, ptk_icvlen;
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	u32 d3_wake_sysassert; /* must be u32 for debugfs_create_bool */
+	bool d3_test_active;
+	bool store_d3_resume_sram;
+	void *d3_resume_sram;
+	u32 d3_test_pme_ptr;
+#endif
 #endif
 
 	/* BT-Coex */
 	u8 bt_kill_msk;
 	struct iwl_bt_coex_profile_notif last_bt_notif;
+
+	/* Thermal Throttling and CTkill */
+	struct iwl_mvm_tt_mgmt thermal_throttle;
+	s32 temperature;	/* Celsius */
 };
 
 /* Extract MVM priv from op_mode and _hw */
@@ -352,6 +490,19 @@ struct iwl_mvm {
 #define IWL_MAC80211_GET_MVM(_hw)			\
 	IWL_OP_MODE_GET_MVM((struct iwl_op_mode *)((_hw)->priv))
 
+enum iwl_mvm_status {
+	IWL_MVM_STATUS_HW_RFKILL,
+	IWL_MVM_STATUS_HW_CTKILL,
+	IWL_MVM_STATUS_ROC_RUNNING,
+	IWL_MVM_STATUS_IN_HW_RESTART,
+};
+
+static inline bool iwl_mvm_is_radio_killed(struct iwl_mvm *mvm)
+{
+	return test_bit(IWL_MVM_STATUS_HW_RFKILL, &mvm->status) ||
+	       test_bit(IWL_MVM_STATUS_HW_CTKILL, &mvm->status);
+}
+
 extern const u8 iwl_mvm_ac_to_tx_fifo[];
 
 struct iwl_rate_info {
@@ -443,8 +594,10 @@ int iwl_mvm_phy_ctxt_add(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt,
 int iwl_mvm_phy_ctxt_changed(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt,
 			     struct cfg80211_chan_def *chandef,
 			     u8 chains_static, u8 chains_dynamic);
-void iwl_mvm_phy_ctxt_remove(struct iwl_mvm *mvm,
-			     struct iwl_mvm_phy_ctxt *ctxt);
+void iwl_mvm_phy_ctxt_ref(struct iwl_mvm *mvm,
+			  struct iwl_mvm_phy_ctxt *ctxt);
+void iwl_mvm_phy_ctxt_unref(struct iwl_mvm *mvm,
+			    struct iwl_mvm_phy_ctxt *ctxt);
 
 /* MAC (virtual interface) programming */
 int iwl_mvm_mac_ctxt_init(struct iwl_mvm *mvm, struct ieee80211_vif *vif);
@@ -459,6 +612,9 @@ int iwl_mvm_mac_ctxt_beacon_changed(struct iwl_mvm *mvm,
 int iwl_mvm_rx_beacon_notif(struct iwl_mvm *mvm,
 			    struct iwl_rx_cmd_buffer *rxb,
 			    struct iwl_device_cmd *cmd);
+int iwl_mvm_rx_missed_beacons_notif(struct iwl_mvm *mvm,
+				    struct iwl_rx_cmd_buffer *rxb,
+				    struct iwl_device_cmd *cmd);
 
 /* Bindings */
 int iwl_mvm_binding_add_vif(struct iwl_mvm *mvm, struct ieee80211_vif *vif);
@@ -523,6 +679,7 @@ void iwl_mvm_ipv6_addr_change(struct ieee80211_hw *hw,
 			      struct inet6_dev *idev);
 void iwl_mvm_set_default_unicast_key(struct ieee80211_hw *hw,
 				     struct ieee80211_vif *vif, int idx);
+extern const struct file_operations iwl_dbgfs_d3_test_ops;
 
 /* BT Coex */
 int iwl_send_bt_prio_tbl(struct iwl_mvm *mvm);
@@ -534,4 +691,31 @@ void iwl_mvm_bt_rssi_event(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 			   enum ieee80211_rssi_event rssi_event);
 void iwl_mvm_bt_coex_vif_assoc(struct iwl_mvm *mvm, struct ieee80211_vif *vif);
 
+/* beacon filtering */
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+void
+iwl_mvm_beacon_filter_debugfs_parameters(struct ieee80211_vif *vif,
+					 struct iwl_beacon_filter_cmd *cmd);
+#else
+static inline void
+iwl_mvm_beacon_filter_debugfs_parameters(struct ieee80211_vif *vif,
+					 struct iwl_beacon_filter_cmd *cmd)
+{}
+#endif
+int iwl_mvm_enable_beacon_filter(struct iwl_mvm *mvm,
+				 struct ieee80211_vif *vif);
+int iwl_mvm_disable_beacon_filter(struct iwl_mvm *mvm,
+				  struct ieee80211_vif *vif);
+
+/* SMPS */
+void iwl_mvm_update_smps(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
+				enum iwl_mvm_smps_type_request req_type,
+				enum ieee80211_smps_mode smps_request);
+
+/* Thermal management and CT-kill */
+void iwl_mvm_tt_handler(struct iwl_mvm *mvm);
+void iwl_mvm_tt_initialize(struct iwl_mvm *mvm);
+void iwl_mvm_tt_exit(struct iwl_mvm *mvm);
+void iwl_mvm_set_hw_ctkill_state(struct iwl_mvm *mvm, bool state);
+
 #endif /* __IWL_MVM_H__ */
diff --git a/drivers/net/wireless/iwlwifi/mvm/nvm.c b/drivers/net/wireless/iwlwifi/mvm/nvm.c
index b8ec02f..edb94ea 100644
--- a/drivers/net/wireless/iwlwifi/mvm/nvm.c
+++ b/drivers/net/wireless/iwlwifi/mvm/nvm.c
@@ -60,6 +60,7 @@
  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  *
  *****************************************************************************/
+#include <linux/firmware.h>
 #include "iwl-trans.h"
 #include "mvm.h"
 #include "iwl-eeprom-parse.h"
@@ -75,31 +76,56 @@ static const int nvm_to_read[] = {
 };
 
 /* Default NVM size to read */
-#define IWL_NVM_DEFAULT_CHUNK_SIZE (2*1024);
+#define IWL_NVM_DEFAULT_CHUNK_SIZE (2*1024)
+#define IWL_MAX_NVM_SECTION_SIZE 6000
 
-static inline void iwl_nvm_fill_read(struct iwl_nvm_access_cmd *cmd,
-				     u16 offset, u16 length, u16 section)
+#define NVM_WRITE_OPCODE 1
+#define NVM_READ_OPCODE 0
+
+/*
+ * prepare the NVM host command w/ the pointers to the nvm buffer
+ * and send it to fw
+ */
+static int iwl_nvm_write_chunk(struct iwl_mvm *mvm, u16 section,
+			       u16 offset, u16 length, const u8 *data)
 {
-	cmd->offset = cpu_to_le16(offset);
-	cmd->length = cpu_to_le16(length);
-	cmd->type = cpu_to_le16(section);
+	struct iwl_nvm_access_cmd nvm_access_cmd = {
+		.offset = cpu_to_le16(offset),
+		.length = cpu_to_le16(length),
+		.type = cpu_to_le16(section),
+		.op_code = NVM_WRITE_OPCODE,
+	};
+	struct iwl_host_cmd cmd = {
+		.id = NVM_ACCESS_CMD,
+		.len = { sizeof(struct iwl_nvm_access_cmd), length },
+		.flags = CMD_SYNC | CMD_SEND_IN_RFKILL,
+		.data = { &nvm_access_cmd, data },
+		/* data may come from vmalloc, so use _DUP */
+		.dataflags = { 0, IWL_HCMD_DFL_DUP },
+	};
+
+	return iwl_mvm_send_cmd(mvm, &cmd);
 }
 
 static int iwl_nvm_read_chunk(struct iwl_mvm *mvm, u16 section,
 			      u16 offset, u16 length, u8 *data)
 {
-	struct iwl_nvm_access_cmd nvm_access_cmd = {};
+	struct iwl_nvm_access_cmd nvm_access_cmd = {
+		.offset = cpu_to_le16(offset),
+		.length = cpu_to_le16(length),
+		.type = cpu_to_le16(section),
+		.op_code = NVM_READ_OPCODE,
+	};
 	struct iwl_nvm_access_resp *nvm_resp;
 	struct iwl_rx_packet *pkt;
 	struct iwl_host_cmd cmd = {
 		.id = NVM_ACCESS_CMD,
-		.flags = CMD_SYNC | CMD_WANT_SKB,
+		.flags = CMD_SYNC | CMD_WANT_SKB | CMD_SEND_IN_RFKILL,
 		.data = { &nvm_access_cmd, },
 	};
 	int ret, bytes_read, offset_read;
 	u8 *resp_data;
 
-	iwl_nvm_fill_read(&nvm_access_cmd, offset, length, section);
 	cmd.len[0] = sizeof(struct iwl_nvm_access_cmd);
 
 	ret = iwl_mvm_send_cmd(mvm, &cmd);
@@ -144,6 +170,30 @@ exit:
 	return ret;
 }
 
+static int iwl_nvm_write_section(struct iwl_mvm *mvm, u16 section,
+				 const u8 *data, u16 length)
+{
+	int offset = 0;
+
+	/* copy data in chunks of 2k (and remainder if any) */
+
+	while (offset < length) {
+		int chunk_size, ret;
+
+		chunk_size = min(IWL_NVM_DEFAULT_CHUNK_SIZE,
+				 length - offset);
+
+		ret = iwl_nvm_write_chunk(mvm, section, offset,
+					  chunk_size, data + offset);
+		if (ret < 0)
+			return ret;
+
+		offset += chunk_size;
+	}
+
+	return 0;
+}
+
 /*
  * Reads an NVM section completely.
  * NICs prior to 7000 family doesn't have a real NVM, but just read
@@ -177,7 +227,8 @@ static int iwl_nvm_read_section(struct iwl_mvm *mvm, u16 section,
 		offset += ret;
 	}
 
-	IWL_INFO(mvm, "NVM section %d read completed\n", section);
+	IWL_DEBUG_EEPROM(mvm->trans->dev,
+			 "NVM section %d read completed\n", section);
 	return offset;
 }
 
@@ -200,7 +251,130 @@ iwl_parse_nvm_sections(struct iwl_mvm *mvm)
 	hw = (const __le16 *)sections[NVM_SECTION_TYPE_HW].data;
 	sw = (const __le16 *)sections[NVM_SECTION_TYPE_SW].data;
 	calib = (const __le16 *)sections[NVM_SECTION_TYPE_CALIBRATION].data;
-	return iwl_parse_nvm_data(mvm->trans->dev, mvm->cfg, hw, sw, calib);
+	return iwl_parse_nvm_data(mvm->trans->dev, mvm->cfg, hw, sw, calib,
+				  iwl_fw_valid_tx_ant(mvm->fw),
+				  iwl_fw_valid_rx_ant(mvm->fw));
+}
+
+#define MAX_NVM_FILE_LEN	16384
+
+/*
+ * HOW TO CREATE THE NVM FILE FORMAT:
+ * ------------------------------
+ * 1. create hex file, format:
+ *      3800 -> header
+ *      0000 -> header
+ *      5a40 -> data
+ *
+ *   rev - 6 bit (word1)
+ *   len - 10 bit (word1)
+ *   id - 4 bit (word2)
+ *   rsv - 12 bit (word2)
+ *
+ * 2. flip 8bits with 8 bits per line to get the right NVM file format
+ *
+ * 3. create binary file from the hex file
+ *
+ * 4. save as "iNVM_xxx.bin" under /lib/firmware
+ */
+static int iwl_mvm_load_external_nvm(struct iwl_mvm *mvm)
+{
+	int ret, section_id, section_size;
+	const struct firmware *fw_entry;
+	const struct {
+		__le16 word1;
+		__le16 word2;
+		u8 data[];
+	} *file_sec;
+	const u8 *eof;
+
+#define NVM_WORD1_LEN(x) (8 * (x & 0x03FF))
+#define NVM_WORD2_ID(x) (x >> 12)
+
+	/*
+	 * Obtain NVM image via request_firmware. Since we already used
+	 * request_firmware_nowait() for the firmware binary load and only
+	 * get here after that we assume the NVM request can be satisfied
+	 * synchronously.
+	 */
+	ret = request_firmware(&fw_entry, iwlwifi_mod_params.nvm_file,
+			       mvm->trans->dev);
+	if (ret) {
+		IWL_ERR(mvm, "ERROR: %s isn't available %d\n",
+			iwlwifi_mod_params.nvm_file, ret);
+		return ret;
+	}
+
+	IWL_INFO(mvm, "Loaded NVM file %s (%zu bytes)\n",
+		 iwlwifi_mod_params.nvm_file, fw_entry->size);
+
+	if (fw_entry->size < sizeof(*file_sec)) {
+		IWL_ERR(mvm, "NVM file too small\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (fw_entry->size > MAX_NVM_FILE_LEN) {
+		IWL_ERR(mvm, "NVM file too large\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
+	eof = fw_entry->data + fw_entry->size;
+
+	file_sec = (void *)fw_entry->data;
+
+	while (true) {
+		if (file_sec->data > eof) {
+			IWL_ERR(mvm,
+				"ERROR - NVM file too short for section header\n");
+			ret = -EINVAL;
+			break;
+		}
+
+		/* check for EOF marker */
+		if (!file_sec->word1 && !file_sec->word2) {
+			ret = 0;
+			break;
+		}
+
+		section_size = 2 * NVM_WORD1_LEN(le16_to_cpu(file_sec->word1));
+		section_id = NVM_WORD2_ID(le16_to_cpu(file_sec->word2));
+
+		if (section_size > IWL_MAX_NVM_SECTION_SIZE) {
+			IWL_ERR(mvm, "ERROR - section too large (%d)\n",
+				section_size);
+			ret = -EINVAL;
+			break;
+		}
+
+		if (!section_size) {
+			IWL_ERR(mvm, "ERROR - section empty\n");
+			ret = -EINVAL;
+			break;
+		}
+
+		if (file_sec->data + section_size > eof) {
+			IWL_ERR(mvm,
+				"ERROR - NVM file too short for section (%d bytes)\n",
+				section_size);
+			ret = -EINVAL;
+			break;
+		}
+
+		ret = iwl_nvm_write_section(mvm, section_id, file_sec->data,
+					    section_size);
+		if (ret < 0) {
+			IWL_ERR(mvm, "iwl_mvm_send_cmd failed: %d\n", ret);
+			break;
+		}
+
+		/* advance to the next section */
+		file_sec = (void *)(file_sec->data + section_size);
+	}
+out:
+	release_firmware(fw_entry);
+	return ret;
 }
 
 int iwl_nvm_init(struct iwl_mvm *mvm)
@@ -208,6 +382,17 @@ int iwl_nvm_init(struct iwl_mvm *mvm)
 	int ret, i, section;
 	u8 *nvm_buffer, *temp;
 
+	/* load external NVM if configured */
+	if (iwlwifi_mod_params.nvm_file) {
+		/* move to External NVM flow */
+		ret = iwl_mvm_load_external_nvm(mvm);
+		if (ret)
+			return ret;
+	}
+
+	/* Read From FW NVM */
+	IWL_DEBUG_EEPROM(mvm->trans->dev, "Read from NVM\n");
+
 	/* TODO: find correct NVM max size for a section */
 	nvm_buffer = kmalloc(mvm->cfg->base_params->eeprom_size,
 			     GFP_KERNEL);
@@ -231,8 +416,9 @@ int iwl_nvm_init(struct iwl_mvm *mvm)
 	if (ret < 0)
 		return ret;
 
-	ret = 0;
 	mvm->nvm_data = iwl_parse_nvm_sections(mvm);
+	if (!mvm->nvm_data)
+		return -ENODATA;
 
-	return ret;
+	return 0;
 }
diff --git a/drivers/net/wireless/iwlwifi/mvm/ops.c b/drivers/net/wireless/iwlwifi/mvm/ops.c
index b29c31a..af79a14 100644
--- a/drivers/net/wireless/iwlwifi/mvm/ops.c
+++ b/drivers/net/wireless/iwlwifi/mvm/ops.c
@@ -215,17 +215,22 @@ static const struct iwl_rx_handlers iwl_mvm_rx_handlers[] = {
 	RX_HANDLER(REPLY_RX_PHY_CMD, iwl_mvm_rx_rx_phy_cmd, false),
 	RX_HANDLER(TX_CMD, iwl_mvm_rx_tx_cmd, false),
 	RX_HANDLER(BA_NOTIF, iwl_mvm_rx_ba_notif, false),
+
+	RX_HANDLER(BT_PROFILE_NOTIFICATION, iwl_mvm_rx_bt_coex_notif, true),
+	RX_HANDLER(BEACON_NOTIFICATION, iwl_mvm_rx_beacon_notif, false),
+	RX_HANDLER(STATISTICS_NOTIFICATION, iwl_mvm_rx_statistics, true),
+
 	RX_HANDLER(TIME_EVENT_NOTIFICATION, iwl_mvm_rx_time_event_notif, false),
 
 	RX_HANDLER(SCAN_REQUEST_CMD, iwl_mvm_rx_scan_response, false),
 	RX_HANDLER(SCAN_COMPLETE_NOTIFICATION, iwl_mvm_rx_scan_complete, false),
 
-	RX_HANDLER(BT_PROFILE_NOTIFICATION, iwl_mvm_rx_bt_coex_notif, true),
-	RX_HANDLER(BEACON_NOTIFICATION, iwl_mvm_rx_beacon_notif, false),
-
 	RX_HANDLER(RADIO_VERSION_NOTIFICATION, iwl_mvm_rx_radio_ver, false),
 	RX_HANDLER(CARD_STATE_NOTIFICATION, iwl_mvm_rx_card_state_notif, false),
 
+	RX_HANDLER(MISSED_BEACONS_NOTIFICATION, iwl_mvm_rx_missed_beacons_notif,
+		   false),
+
 	RX_HANDLER(REPLY_ERROR, iwl_mvm_rx_fw_error, false),
 };
 #undef RX_HANDLER
@@ -288,11 +293,14 @@ static const char *iwl_mvm_cmd_strings[REPLY_MAX] = {
 	CMD(NET_DETECT_HOTSPOTS_CMD),
 	CMD(NET_DETECT_HOTSPOTS_QUERY_CMD),
 	CMD(CARD_STATE_NOTIFICATION),
+	CMD(MISSED_BEACONS_NOTIFICATION),
 	CMD(BT_COEX_PRIO_TABLE),
 	CMD(BT_COEX_PROT_ENV),
 	CMD(BT_PROFILE_NOTIFICATION),
 	CMD(BT_CONFIG),
 	CMD(MCAST_FILTER_CMD),
+	CMD(REPLY_BEACON_FILTERING_CMD),
+	CMD(REPLY_THERMAL_MNG_BACKOFF),
 };
 #undef CMD
 
@@ -393,10 +401,13 @@ iwl_op_mode_mvm_start(struct iwl_trans *trans, const struct iwl_cfg *cfg,
 	if (err)
 		goto out_free;
 
+	iwl_mvm_tt_initialize(mvm);
+
 	mutex_lock(&mvm->mutex);
 	err = iwl_run_init_mvm_ucode(mvm, true);
 	mutex_unlock(&mvm->mutex);
-	if (err && !iwlmvm_mod_params.init_dbg) {
+	/* returns 0 if successful, 1 if success but in rfkill */
+	if (err < 0 && !iwlmvm_mod_params.init_dbg) {
 		IWL_ERR(mvm, "Failed to run INIT ucode: %d\n", err);
 		goto out_free;
 	}
@@ -439,10 +450,16 @@ static void iwl_op_mode_mvm_stop(struct iwl_op_mode *op_mode)
 
 	iwl_mvm_leds_exit(mvm);
 
+	iwl_mvm_tt_exit(mvm);
+
 	ieee80211_unregister_hw(mvm->hw);
 
 	kfree(mvm->scan_cmd);
 
+#if defined(CONFIG_PM_SLEEP) && defined(CONFIG_IWLWIFI_DEBUGFS)
+	kfree(mvm->d3_resume_sram);
+#endif
+
 	iwl_trans_stop_hw(mvm->trans, true);
 
 	iwl_phy_db_free(mvm->phy_db);
@@ -589,6 +606,16 @@ static void iwl_mvm_wake_sw_queue(struct iwl_op_mode *op_mode, int queue)
 	ieee80211_wake_queue(mvm->hw, mq);
 }
 
+void iwl_mvm_set_hw_ctkill_state(struct iwl_mvm *mvm, bool state)
+{
+	if (state)
+		set_bit(IWL_MVM_STATUS_HW_CTKILL, &mvm->status);
+	else
+		clear_bit(IWL_MVM_STATUS_HW_CTKILL, &mvm->status);
+
+	wiphy_rfkill_set_hw_state(mvm->hw->wiphy, iwl_mvm_is_radio_killed(mvm));
+}
+
 static void iwl_mvm_set_hw_rfkill_state(struct iwl_op_mode *op_mode, bool state)
 {
 	struct iwl_mvm *mvm = IWL_OP_MODE_GET_MVM(op_mode);
@@ -598,7 +625,7 @@ static void iwl_mvm_set_hw_rfkill_state(struct iwl_op_mode *op_mode, bool state)
 	else
 		clear_bit(IWL_MVM_STATUS_HW_RFKILL, &mvm->status);
 
-	wiphy_rfkill_set_hw_state(mvm->hw->wiphy, state);
+	wiphy_rfkill_set_hw_state(mvm->hw->wiphy, iwl_mvm_is_radio_killed(mvm));
 }
 
 static void iwl_mvm_free_skb(struct iwl_op_mode *op_mode, struct sk_buff *skb)
diff --git a/drivers/net/wireless/iwlwifi/mvm/phy-ctxt.c b/drivers/net/wireless/iwlwifi/mvm/phy-ctxt.c
index a28a1d1..a8652dd 100644
--- a/drivers/net/wireless/iwlwifi/mvm/phy-ctxt.c
+++ b/drivers/net/wireless/iwlwifi/mvm/phy-ctxt.c
@@ -195,21 +195,6 @@ static int iwl_mvm_phy_ctxt_apply(struct iwl_mvm *mvm,
 	return ret;
 }
 
-
-struct phy_ctx_used_data {
-	unsigned long used[BITS_TO_LONGS(NUM_PHY_CTX)];
-};
-
-static void iwl_mvm_phy_ctx_used_iter(struct ieee80211_hw *hw,
-				      struct ieee80211_chanctx_conf *ctx,
-				      void *_data)
-{
-	struct phy_ctx_used_data *data = _data;
-	struct iwl_mvm_phy_ctxt *phy_ctxt = (void *)ctx->drv_priv;
-
-	__set_bit(phy_ctxt->id, data->used);
-}
-
 /*
  * Send a command to add a PHY context based on the current HW configuration.
  */
@@ -217,34 +202,28 @@ int iwl_mvm_phy_ctxt_add(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt,
 			 struct cfg80211_chan_def *chandef,
 			 u8 chains_static, u8 chains_dynamic)
 {
-	struct phy_ctx_used_data data = {
-		.used = { },
-	};
-
-	/*
-	 * If this is a regular PHY context (not the ROC one)
-	 * skip the ROC PHY context's ID.
-	 */
-	if (ctxt != &mvm->phy_ctxt_roc)
-		__set_bit(mvm->phy_ctxt_roc.id, data.used);
+	int ret;
 
+	WARN_ON(!test_bit(IWL_MVM_STATUS_IN_HW_RESTART, &mvm->status) &&
+		ctxt->ref);
 	lockdep_assert_held(&mvm->mutex);
-	ctxt->color++;
 
-	if (!test_bit(IWL_MVM_STATUS_IN_HW_RESTART, &mvm->status)) {
-		ieee80211_iter_chan_contexts_atomic(
-			mvm->hw, iwl_mvm_phy_ctx_used_iter, &data);
+	ctxt->channel = chandef->chan;
+	ret = iwl_mvm_phy_ctxt_apply(mvm, ctxt, chandef,
+				     chains_static, chains_dynamic,
+				     FW_CTXT_ACTION_ADD, 0);
 
-		ctxt->id = find_first_zero_bit(data.used, NUM_PHY_CTX);
-		if (WARN_ONCE(ctxt->id == NUM_PHY_CTX,
-			      "Failed to init PHY context - no free ID!\n"))
-			return -EIO;
-	}
+	return ret;
+}
 
-	ctxt->channel = chandef->chan;
-	return iwl_mvm_phy_ctxt_apply(mvm, ctxt, chandef,
-				      chains_static, chains_dynamic,
-				      FW_CTXT_ACTION_ADD, 0);
+/*
+ * Update the number of references to the given PHY context. This is valid only
+ * in case the PHY context was already created, i.e., its reference count > 0.
+ */
+void iwl_mvm_phy_ctxt_ref(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt)
+{
+	lockdep_assert_held(&mvm->mutex);
+	ctxt->ref++;
 }
 
 /*
@@ -264,23 +243,12 @@ int iwl_mvm_phy_ctxt_changed(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt,
 				      FW_CTXT_ACTION_MODIFY, 0);
 }
 
-/*
- * Send a command to the FW to remove the given phy context.
- * Once the command is sent, regardless of success or failure, the context is
- * marked as invalid
- */
-void iwl_mvm_phy_ctxt_remove(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt)
+void iwl_mvm_phy_ctxt_unref(struct iwl_mvm *mvm, struct iwl_mvm_phy_ctxt *ctxt)
 {
-	struct iwl_phy_context_cmd cmd;
-	int ret;
-
 	lockdep_assert_held(&mvm->mutex);
 
-	iwl_mvm_phy_ctxt_cmd_hdr(ctxt, &cmd, FW_CTXT_ACTION_REMOVE, 0);
-	ret = iwl_mvm_send_cmd_pdu(mvm, PHY_CONTEXT_CMD, CMD_SYNC,
-				   sizeof(struct iwl_phy_context_cmd),
-				   &cmd);
-	if (ret)
-		IWL_ERR(mvm, "Failed to send PHY remove: ctxt id=%d\n",
-			ctxt->id);
+	if (WARN_ON_ONCE(!ctxt))
+		return;
+
+	ctxt->ref--;
 }
diff --git a/drivers/net/wireless/iwlwifi/mvm/power.c b/drivers/net/wireless/iwlwifi/mvm/power.c
index ed77e43..e7ca965 100644
--- a/drivers/net/wireless/iwlwifi/mvm/power.c
+++ b/drivers/net/wireless/iwlwifi/mvm/power.c
@@ -75,6 +75,54 @@
 
 #define POWER_KEEP_ALIVE_PERIOD_SEC    25
 
+static int iwl_mvm_beacon_filter_send_cmd(struct iwl_mvm *mvm,
+					  struct iwl_beacon_filter_cmd *cmd)
+{
+	int ret;
+
+	ret = iwl_mvm_send_cmd_pdu(mvm, REPLY_BEACON_FILTERING_CMD, CMD_SYNC,
+				   sizeof(struct iwl_beacon_filter_cmd), cmd);
+
+	if (!ret) {
+		IWL_DEBUG_POWER(mvm, "ba_enable_beacon_abort is: %d\n",
+				cmd->ba_enable_beacon_abort);
+		IWL_DEBUG_POWER(mvm, "ba_escape_timer is: %d\n",
+				cmd->ba_escape_timer);
+		IWL_DEBUG_POWER(mvm, "bf_debug_flag is: %d\n",
+				cmd->bf_debug_flag);
+		IWL_DEBUG_POWER(mvm, "bf_enable_beacon_filter is: %d\n",
+				cmd->bf_enable_beacon_filter);
+		IWL_DEBUG_POWER(mvm, "bf_energy_delta is: %d\n",
+				cmd->bf_energy_delta);
+		IWL_DEBUG_POWER(mvm, "bf_escape_timer is: %d\n",
+				cmd->bf_escape_timer);
+		IWL_DEBUG_POWER(mvm, "bf_roaming_energy_delta is: %d\n",
+				cmd->bf_roaming_energy_delta);
+		IWL_DEBUG_POWER(mvm, "bf_roaming_state is: %d\n",
+				cmd->bf_roaming_state);
+		IWL_DEBUG_POWER(mvm, "bf_temperature_delta is: %d\n",
+				cmd->bf_temperature_delta);
+	}
+	return ret;
+}
+
+static int iwl_mvm_update_beacon_abort(struct iwl_mvm *mvm,
+				       struct ieee80211_vif *vif, bool enable)
+{
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_beacon_filter_cmd cmd = {
+		IWL_BF_CMD_CONFIG_DEFAULTS,
+		.bf_enable_beacon_filter = 1,
+		.ba_enable_beacon_abort = enable,
+	};
+
+	if (!mvmvif->bf_enabled)
+		return 0;
+
+	iwl_mvm_beacon_filter_debugfs_parameters(vif, &cmd);
+	return iwl_mvm_beacon_filter_send_cmd(mvm, &cmd);
+}
+
 static void iwl_mvm_power_log(struct iwl_mvm *mvm,
 			      struct iwl_powertable_cmd *cmd)
 {
@@ -89,8 +137,12 @@ static void iwl_mvm_power_log(struct iwl_mvm *mvm,
 				le32_to_cpu(cmd->rx_data_timeout));
 		IWL_DEBUG_POWER(mvm, "Tx timeout = %u usec\n",
 				le32_to_cpu(cmd->tx_data_timeout));
-		IWL_DEBUG_POWER(mvm, "LP RX RSSI threshold = %u\n",
-				cmd->lprx_rssi_threshold);
+		if (cmd->flags & cpu_to_le16(POWER_FLAGS_SKIP_OVER_DTIM_MSK))
+			IWL_DEBUG_POWER(mvm, "DTIM periods to skip = %u\n",
+					le32_to_cpu(cmd->skip_dtim_periods));
+		if (cmd->flags & cpu_to_le16(POWER_FLAGS_LPRX_ENA_MSK))
+			IWL_DEBUG_POWER(mvm, "LP RX RSSI threshold = %u\n",
+					le32_to_cpu(cmd->lprx_rssi_threshold));
 	}
 }
 
@@ -103,6 +155,8 @@ void iwl_mvm_power_build_cmd(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 	int dtimper, dtimper_msec;
 	int keep_alive;
 	bool radar_detect = false;
+	struct iwl_mvm_vif *mvmvif __maybe_unused =
+		iwl_mvm_vif_from_mac80211(vif);
 
 	/*
 	 * Regardless of power management state the driver must set
@@ -115,12 +169,27 @@ void iwl_mvm_power_build_cmd(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 		return;
 
 	cmd->flags |= cpu_to_le16(POWER_FLAGS_POWER_SAVE_ENA_MSK);
+	if (!vif->bss_conf.assoc)
+		cmd->flags |= cpu_to_le16(POWER_FLAGS_POWER_MANAGEMENT_ENA_MSK);
 
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_DISABLE_POWER_OFF &&
+	    mvmvif->dbgfs_pm.disable_power_off)
+		cmd->flags &= cpu_to_le16(~POWER_FLAGS_POWER_SAVE_ENA_MSK);
+#endif
 	if (!vif->bss_conf.ps)
 		return;
 
 	cmd->flags |= cpu_to_le16(POWER_FLAGS_POWER_MANAGEMENT_ENA_MSK);
 
+	if (vif->bss_conf.beacon_rate &&
+	    (vif->bss_conf.beacon_rate->bitrate == 10 ||
+	     vif->bss_conf.beacon_rate->bitrate == 60)) {
+		cmd->flags |= cpu_to_le16(POWER_FLAGS_LPRX_ENA_MSK);
+		cmd->lprx_rssi_threshold =
+			cpu_to_le32(POWER_LPRX_RSSI_THRESHOLD);
+	}
+
 	dtimper = hw->conf.ps_dtim_period ?: 1;
 
 	/* Check if radar detection is required on current channel */
@@ -135,8 +204,11 @@ void iwl_mvm_power_build_cmd(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 
 	/* Check skip over DTIM conditions */
 	if (!radar_detect && (dtimper <= 10) &&
-	    (iwlmvm_mod_params.power_scheme == IWL_POWER_SCHEME_LP))
+	    (iwlmvm_mod_params.power_scheme == IWL_POWER_SCHEME_LP ||
+	     mvm->cur_ucode == IWL_UCODE_WOWLAN)) {
 		cmd->flags |= cpu_to_le16(POWER_FLAGS_SKIP_OVER_DTIM_MSK);
+		cmd->skip_dtim_periods = cpu_to_le32(3);
+	}
 
 	/* Check that keep alive period is at least 3 * DTIM */
 	dtimper_msec = dtimper * vif->bss_conf.beacon_int;
@@ -145,27 +217,85 @@ void iwl_mvm_power_build_cmd(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 	keep_alive = DIV_ROUND_UP(keep_alive, MSEC_PER_SEC);
 	cmd->keep_alive_seconds = keep_alive;
 
-	cmd->rx_data_timeout = cpu_to_le32(100 * USEC_PER_MSEC);
-	cmd->tx_data_timeout = cpu_to_le32(100 * USEC_PER_MSEC);
+	if (mvm->cur_ucode != IWL_UCODE_WOWLAN) {
+		cmd->rx_data_timeout = cpu_to_le32(100 * USEC_PER_MSEC);
+		cmd->tx_data_timeout = cpu_to_le32(100 * USEC_PER_MSEC);
+	} else {
+		cmd->rx_data_timeout = cpu_to_le32(10 * USEC_PER_MSEC);
+		cmd->tx_data_timeout = cpu_to_le32(10 * USEC_PER_MSEC);
+	}
+
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_KEEP_ALIVE)
+		cmd->keep_alive_seconds = mvmvif->dbgfs_pm.keep_alive_seconds;
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_SKIP_OVER_DTIM) {
+		if (mvmvif->dbgfs_pm.skip_over_dtim)
+			cmd->flags |=
+				cpu_to_le16(POWER_FLAGS_SKIP_OVER_DTIM_MSK);
+		else
+			cmd->flags &=
+				cpu_to_le16(~POWER_FLAGS_SKIP_OVER_DTIM_MSK);
+	}
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_RX_DATA_TIMEOUT)
+		cmd->rx_data_timeout =
+			cpu_to_le32(mvmvif->dbgfs_pm.rx_data_timeout);
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_TX_DATA_TIMEOUT)
+		cmd->tx_data_timeout =
+			cpu_to_le32(mvmvif->dbgfs_pm.tx_data_timeout);
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_SKIP_DTIM_PERIODS)
+		cmd->skip_dtim_periods =
+			cpu_to_le32(mvmvif->dbgfs_pm.skip_dtim_periods);
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_LPRX_ENA) {
+		if (mvmvif->dbgfs_pm.lprx_ena)
+			cmd->flags |= cpu_to_le16(POWER_FLAGS_LPRX_ENA_MSK);
+		else
+			cmd->flags &= cpu_to_le16(~POWER_FLAGS_LPRX_ENA_MSK);
+	}
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_LPRX_RSSI_THRESHOLD)
+		cmd->lprx_rssi_threshold =
+			cpu_to_le32(mvmvif->dbgfs_pm.lprx_rssi_threshold);
+#endif /* CONFIG_IWLWIFI_DEBUGFS */
 }
 
 int iwl_mvm_power_update_mode(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 {
+	int ret;
+	bool ba_enable;
 	struct iwl_powertable_cmd cmd = {};
 
 	if (vif->type != NL80211_IFTYPE_STATION || vif->p2p)
 		return 0;
 
+	/*
+	 * TODO: The following vif_count verification is temporary condition.
+	 * Avoid power mode update if more than one interface is currently
+	 * active. Remove this condition when FW will support power management
+	 * on multiple MACs.
+	 */
+	IWL_DEBUG_POWER(mvm, "Currently %d interfaces active\n",
+			mvm->vif_count);
+	if (mvm->vif_count > 1)
+		return 0;
+
 	iwl_mvm_power_build_cmd(mvm, vif, &cmd);
 	iwl_mvm_power_log(mvm, &cmd);
 
-	return iwl_mvm_send_cmd_pdu(mvm, POWER_TABLE_CMD, CMD_SYNC,
-				    sizeof(cmd), &cmd);
+	ret = iwl_mvm_send_cmd_pdu(mvm, POWER_TABLE_CMD, CMD_SYNC,
+				   sizeof(cmd), &cmd);
+	if (ret)
+		return ret;
+
+	ba_enable = !!(cmd.flags &
+		       cpu_to_le16(POWER_FLAGS_POWER_MANAGEMENT_ENA_MSK));
+
+	return iwl_mvm_update_beacon_abort(mvm, vif, ba_enable);
 }
 
 int iwl_mvm_power_disable(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 {
 	struct iwl_powertable_cmd cmd = {};
+	struct iwl_mvm_vif *mvmvif __maybe_unused =
+		iwl_mvm_vif_from_mac80211(vif);
 
 	if (vif->type != NL80211_IFTYPE_STATION || vif->p2p)
 		return 0;
@@ -173,8 +303,82 @@ int iwl_mvm_power_disable(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 	if (iwlmvm_mod_params.power_scheme != IWL_POWER_SCHEME_CAM)
 		cmd.flags |= cpu_to_le16(POWER_FLAGS_POWER_SAVE_ENA_MSK);
 
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+	if (mvmvif->dbgfs_pm.mask & MVM_DEBUGFS_PM_DISABLE_POWER_OFF &&
+	    mvmvif->dbgfs_pm.disable_power_off)
+		cmd.flags &= cpu_to_le16(~POWER_FLAGS_POWER_SAVE_ENA_MSK);
+#endif
 	iwl_mvm_power_log(mvm, &cmd);
 
 	return iwl_mvm_send_cmd_pdu(mvm, POWER_TABLE_CMD, CMD_ASYNC,
 				    sizeof(cmd), &cmd);
 }
+
+#ifdef CONFIG_IWLWIFI_DEBUGFS
+void
+iwl_mvm_beacon_filter_debugfs_parameters(struct ieee80211_vif *vif,
+					 struct iwl_beacon_filter_cmd *cmd)
+{
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_dbgfs_bf *dbgfs_bf = &mvmvif->dbgfs_bf;
+
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BF_ENERGY_DELTA)
+		cmd->bf_energy_delta = dbgfs_bf->bf_energy_delta;
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BF_ROAMING_ENERGY_DELTA)
+		cmd->bf_roaming_energy_delta =
+				 dbgfs_bf->bf_roaming_energy_delta;
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BF_ROAMING_STATE)
+		cmd->bf_roaming_state = dbgfs_bf->bf_roaming_state;
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BF_TEMPERATURE_DELTA)
+		cmd->bf_temperature_delta = dbgfs_bf->bf_temperature_delta;
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BF_DEBUG_FLAG)
+		cmd->bf_debug_flag = dbgfs_bf->bf_debug_flag;
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BF_ESCAPE_TIMER)
+		cmd->bf_escape_timer = cpu_to_le32(dbgfs_bf->bf_escape_timer);
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BA_ESCAPE_TIMER)
+		cmd->ba_escape_timer = cpu_to_le32(dbgfs_bf->ba_escape_timer);
+	if (dbgfs_bf->mask & MVM_DEBUGFS_BA_ENABLE_BEACON_ABORT)
+		cmd->ba_enable_beacon_abort = dbgfs_bf->ba_enable_beacon_abort;
+}
+#endif
+
+int iwl_mvm_enable_beacon_filter(struct iwl_mvm *mvm,
+				 struct ieee80211_vif *vif)
+{
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	struct iwl_beacon_filter_cmd cmd = {
+		IWL_BF_CMD_CONFIG_DEFAULTS,
+		.bf_enable_beacon_filter = 1,
+	};
+	int ret;
+
+	if (mvmvif != mvm->bf_allowed_vif ||
+	    vif->type != NL80211_IFTYPE_STATION || vif->p2p)
+		return 0;
+
+	iwl_mvm_beacon_filter_debugfs_parameters(vif, &cmd);
+	ret = iwl_mvm_beacon_filter_send_cmd(mvm, &cmd);
+
+	if (!ret)
+		mvmvif->bf_enabled = true;
+
+	return ret;
+}
+
+int iwl_mvm_disable_beacon_filter(struct iwl_mvm *mvm,
+				  struct ieee80211_vif *vif)
+{
+	struct iwl_beacon_filter_cmd cmd = {};
+	struct iwl_mvm_vif *mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	int ret;
+
+	if (vif->type != NL80211_IFTYPE_STATION || vif->p2p)
+		return 0;
+
+	ret = iwl_mvm_beacon_filter_send_cmd(mvm, &cmd);
+
+	if (!ret)
+		mvmvif->bf_enabled = false;
+
+	return ret;
+}
diff --git a/drivers/net/wireless/iwlwifi/mvm/quota.c b/drivers/net/wireless/iwlwifi/mvm/quota.c
index a1e3e92..29d49cf 100644
--- a/drivers/net/wireless/iwlwifi/mvm/quota.c
+++ b/drivers/net/wireless/iwlwifi/mvm/quota.c
@@ -169,27 +169,34 @@ int iwl_mvm_update_quotas(struct iwl_mvm *mvm, struct ieee80211_vif *newvif)
 			num_active_bindings++;
 	}
 
-	if (!num_active_bindings)
-		goto send_cmd;
-
-	quota = IWL_MVM_MAX_QUOTA / num_active_bindings;
-	quota_rem = IWL_MVM_MAX_QUOTA % num_active_bindings;
+	quota = 0;
+	quota_rem = 0;
+	if (num_active_bindings) {
+		quota = IWL_MVM_MAX_QUOTA / num_active_bindings;
+		quota_rem = IWL_MVM_MAX_QUOTA % num_active_bindings;
+	}
 
 	for (idx = 0, i = 0; i < MAX_BINDINGS; i++) {
-		if (data.n_interfaces[i] <= 0)
+		if (data.colors[i] < 0)
 			continue;
 
 		cmd.quotas[idx].id_and_color =
 			cpu_to_le32(FW_CMD_ID_AND_COLOR(i, data.colors[i]));
-		cmd.quotas[idx].quota = cpu_to_le32(quota);
-		cmd.quotas[idx].max_duration = cpu_to_le32(IWL_MVM_MAX_QUOTA);
+
+		if (data.n_interfaces[i] <= 0) {
+			cmd.quotas[idx].quota = cpu_to_le32(0);
+			cmd.quotas[idx].max_duration = cpu_to_le32(0);
+		} else {
+			cmd.quotas[idx].quota = cpu_to_le32(quota);
+			cmd.quotas[idx].max_duration =
+				cpu_to_le32(IWL_MVM_MAX_QUOTA);
+		}
 		idx++;
 	}
 
 	/* Give the remainder of the session to the first binding */
 	le32_add_cpu(&cmd.quotas[0].quota, quota_rem);
 
-send_cmd:
 	ret = iwl_mvm_send_cmd_pdu(mvm, TIME_QUOTA_CMD, CMD_SYNC,
 				   sizeof(cmd), &cmd);
 	if (ret)
diff --git a/drivers/net/wireless/iwlwifi/mvm/rs.c b/drivers/net/wireless/iwlwifi/mvm/rs.c
index b99fe31..b328a98 100644
--- a/drivers/net/wireless/iwlwifi/mvm/rs.c
+++ b/drivers/net/wireless/iwlwifi/mvm/rs.c
@@ -401,24 +401,29 @@ static int rs_tl_turn_on_agg_for_tid(struct iwl_mvm *mvm,
 
 	load = rs_tl_get_load(lq_data, tid);
 
-	if ((iwlwifi_mod_params.auto_agg) || (load > IWL_AGG_LOAD_THRESHOLD)) {
-		IWL_DEBUG_HT(mvm, "Starting Tx agg: STA: %pM tid: %d\n",
-			     sta->addr, tid);
-		ret = ieee80211_start_tx_ba_session(sta, tid, 5000);
-		if (ret == -EAGAIN) {
-			/*
-			 * driver and mac80211 is out of sync
-			 * this might be cause by reloading firmware
-			 * stop the tx ba session here
-			 */
-			IWL_ERR(mvm, "Fail start Tx agg on tid: %d\n",
-				tid);
-			ieee80211_stop_tx_ba_session(sta, tid);
-		}
-	} else {
-		IWL_DEBUG_HT(mvm,
-			     "Aggregation not enabled for tid %d because load = %u\n",
-			     tid, load);
+	/*
+	 * Don't create TX aggregation sessions when in high
+	 * BT traffic, as they would just be disrupted by BT.
+	 */
+	if (BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD) >= 2) {
+		IWL_DEBUG_COEX(mvm, "BT traffic (%d), no aggregation allowed\n",
+			       BT_MBOX_MSG(&mvm->last_bt_notif,
+					   3, TRAFFIC_LOAD));
+		return ret;
+	}
+
+	IWL_DEBUG_HT(mvm, "Starting Tx agg: STA: %pM tid: %d\n",
+		     sta->addr, tid);
+	ret = ieee80211_start_tx_ba_session(sta, tid, 5000);
+	if (ret == -EAGAIN) {
+		/*
+		 * driver and mac80211 is out of sync
+		 * this might be cause by reloading firmware
+		 * stop the tx ba session here
+		 */
+		IWL_ERR(mvm, "Fail start Tx agg on tid: %d\n",
+			tid);
+		ieee80211_stop_tx_ba_session(sta, tid);
 	}
 	return ret;
 }
@@ -1519,6 +1524,29 @@ static int rs_move_siso_to_other(struct iwl_mvm *mvm,
 	u8 update_search_tbl_counter = 0;
 	int ret;
 
+	switch (BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD)) {
+	case IWL_BT_COEX_TRAFFIC_LOAD_NONE:
+		/* nothing */
+		break;
+	case IWL_BT_COEX_TRAFFIC_LOAD_LOW:
+		/* avoid antenna B unless MIMO */
+		if (tbl->action == IWL_SISO_SWITCH_ANTENNA2)
+			tbl->action = IWL_SISO_SWITCH_MIMO2_AB;
+		break;
+	case IWL_BT_COEX_TRAFFIC_LOAD_HIGH:
+	case IWL_BT_COEX_TRAFFIC_LOAD_CONTINUOUS:
+		/* avoid antenna B and MIMO */
+		valid_tx_ant =
+			first_antenna(iwl_fw_valid_tx_ant(mvm->fw));
+		if (tbl->action != IWL_SISO_SWITCH_ANTENNA1)
+			tbl->action = IWL_SISO_SWITCH_ANTENNA1;
+		break;
+	default:
+		IWL_ERR(mvm, "Invalid BT load %d",
+			BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD));
+		break;
+	}
+
 	start_action = tbl->action;
 	while (1) {
 		lq_sta->action_counter++;
@@ -1532,7 +1560,9 @@ static int rs_move_siso_to_other(struct iwl_mvm *mvm,
 			     tx_chains_num <= 2))
 				break;
 
-			if (window->success_ratio >= IWL_RS_GOOD_RATIO)
+			if (window->success_ratio >= IWL_RS_GOOD_RATIO &&
+			    BT_MBOX_MSG(&mvm->last_bt_notif, 3,
+					TRAFFIC_LOAD) == 0)
 				break;
 
 			memcpy(search_tbl, tbl, sz);
@@ -1654,6 +1684,28 @@ static int rs_move_mimo2_to_other(struct iwl_mvm *mvm,
 	u8 update_search_tbl_counter = 0;
 	int ret;
 
+	switch (BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD)) {
+	case IWL_BT_COEX_TRAFFIC_LOAD_NONE:
+		/* nothing */
+		break;
+	case IWL_BT_COEX_TRAFFIC_LOAD_HIGH:
+	case IWL_BT_COEX_TRAFFIC_LOAD_CONTINUOUS:
+		/* avoid antenna B and MIMO */
+		if (tbl->action != IWL_MIMO2_SWITCH_SISO_A)
+			tbl->action = IWL_MIMO2_SWITCH_SISO_A;
+		break;
+	case IWL_BT_COEX_TRAFFIC_LOAD_LOW:
+		/* avoid antenna B unless MIMO */
+		if (tbl->action == IWL_MIMO2_SWITCH_SISO_B ||
+		    tbl->action == IWL_MIMO2_SWITCH_SISO_C)
+			tbl->action = IWL_MIMO2_SWITCH_SISO_A;
+		break;
+	default:
+		IWL_ERR(mvm, "Invalid BT load %d",
+			BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD));
+		break;
+	}
+
 	start_action = tbl->action;
 	while (1) {
 		lq_sta->action_counter++;
@@ -1791,6 +1843,28 @@ static int rs_move_mimo3_to_other(struct iwl_mvm *mvm,
 	int ret;
 	u8 update_search_tbl_counter = 0;
 
+	switch (BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD)) {
+	case IWL_BT_COEX_TRAFFIC_LOAD_NONE:
+		/* nothing */
+		break;
+	case IWL_BT_COEX_TRAFFIC_LOAD_HIGH:
+	case IWL_BT_COEX_TRAFFIC_LOAD_CONTINUOUS:
+		/* avoid antenna B and MIMO */
+		if (tbl->action != IWL_MIMO3_SWITCH_SISO_A)
+			tbl->action = IWL_MIMO3_SWITCH_SISO_A;
+		break;
+	case IWL_BT_COEX_TRAFFIC_LOAD_LOW:
+		/* avoid antenna B unless MIMO */
+		if (tbl->action == IWL_MIMO3_SWITCH_SISO_B ||
+		    tbl->action == IWL_MIMO3_SWITCH_SISO_C)
+			tbl->action = IWL_MIMO3_SWITCH_SISO_A;
+		break;
+	default:
+		IWL_ERR(mvm, "Invalid BT load %d",
+			BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD));
+		break;
+	}
+
 	start_action = tbl->action;
 	while (1) {
 		lq_sta->action_counter++;
@@ -2302,6 +2376,32 @@ static void rs_rate_scale_perform(struct iwl_mvm *mvm,
 	     (current_tpt > (100 * tbl->expected_tpt[low]))))
 		scale_action = 0;
 
+	if ((BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD) >=
+	     IWL_BT_COEX_TRAFFIC_LOAD_HIGH) &&
+	     (is_mimo2(tbl->lq_type) || is_mimo3(tbl->lq_type))) {
+		if (lq_sta->last_bt_traffic >
+		    BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD)) {
+			/*
+			 * don't set scale_action, don't want to scale up if
+			 * the rate scale doesn't otherwise think that is a
+			 * good idea.
+			 */
+		} else if (lq_sta->last_bt_traffic <=
+			   BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD)) {
+			scale_action = -1;
+		}
+	}
+	lq_sta->last_bt_traffic =
+		BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD);
+
+	if ((BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD) >=
+	     IWL_BT_COEX_TRAFFIC_LOAD_HIGH) &&
+	     (is_mimo2(tbl->lq_type) || is_mimo3(tbl->lq_type))) {
+		/* search for a new modulation */
+		rs_stay_in_table(lq_sta, true);
+		goto lq_update;
+	}
+
 	switch (scale_action) {
 	case -1:
 		/* Decrease starting rate, update uCode's rate table */
@@ -2783,6 +2883,13 @@ static void rs_fill_link_cmd(struct iwl_mvm *mvm,
 
 	lq_cmd->agg_time_limit =
 		cpu_to_le16(LINK_QUAL_AGG_TIME_LIMIT_DEF);
+
+	/*
+	 * overwrite if needed, pass aggregation time limit
+	 * to uCode in uSec - This is racy - but heh, at least it helps...
+	 */
+	if (mvm && BT_MBOX_MSG(&mvm->last_bt_notif, 3, TRAFFIC_LOAD) >= 2)
+		lq_cmd->agg_time_limit = cpu_to_le16(1200);
 }
 
 static void *rs_alloc(struct ieee80211_hw *hw, struct dentry *debugfsdir)
@@ -3081,3 +3188,29 @@ void iwl_mvm_rate_control_unregister(void)
 {
 	ieee80211_rate_control_unregister(&rs_mvm_ops);
 }
+
+/**
+ * iwl_mvm_tx_protection - Gets LQ command, change it to enable/disable
+ * Tx protection, according to this rquest and previous requests,
+ * and send the LQ command.
+ * @lq: The LQ command
+ * @mvmsta: The station
+ * @enable: Enable Tx protection?
+ */
+int iwl_mvm_tx_protection(struct iwl_mvm *mvm, struct iwl_lq_cmd *lq,
+			  struct iwl_mvm_sta *mvmsta, bool enable)
+{
+	lockdep_assert_held(&mvm->mutex);
+
+	if (enable) {
+		if (mvmsta->tx_protection == 0)
+			lq->flags |= LQ_FLAG_SET_STA_TLC_RTS_MSK;
+		mvmsta->tx_protection++;
+	} else {
+		mvmsta->tx_protection--;
+		if (mvmsta->tx_protection == 0)
+			lq->flags &= ~LQ_FLAG_SET_STA_TLC_RTS_MSK;
+	}
+
+	return iwl_mvm_send_lq_cmd(mvm, lq, CMD_ASYNC, false);
+}
diff --git a/drivers/net/wireless/iwlwifi/mvm/rs.h b/drivers/net/wireless/iwlwifi/mvm/rs.h
index 219c685..cff4f6d 100644
--- a/drivers/net/wireless/iwlwifi/mvm/rs.h
+++ b/drivers/net/wireless/iwlwifi/mvm/rs.h
@@ -358,6 +358,18 @@ struct iwl_lq_sta {
 	u8 last_bt_traffic;
 };
 
+enum iwl_bt_coex_profile_traffic_load {
+	IWL_BT_COEX_TRAFFIC_LOAD_NONE		= 0,
+	IWL_BT_COEX_TRAFFIC_LOAD_LOW		= 1,
+	IWL_BT_COEX_TRAFFIC_LOAD_HIGH		= 2,
+	IWL_BT_COEX_TRAFFIC_LOAD_CONTINUOUS	= 3,
+/*
+ * There are no more even though below is a u8, the
+ * indication from the BT device only has two bits.
+ */
+};
+
+
 static inline u8 num_of_ant(u8 mask)
 {
 	return  !!((mask) & ANT_A) +
@@ -390,4 +402,9 @@ extern int iwl_mvm_rate_control_register(void);
  */
 extern void iwl_mvm_rate_control_unregister(void);
 
+struct iwl_mvm_sta;
+
+int iwl_mvm_tx_protection(struct iwl_mvm *mvm, struct iwl_lq_cmd *lq,
+			  struct iwl_mvm_sta *mvmsta, bool enable);
+
 #endif /* __rs__ */
diff --git a/drivers/net/wireless/iwlwifi/mvm/rx.c b/drivers/net/wireless/iwlwifi/mvm/rx.c
index 4dfc21a..e4930d5 100644
--- a/drivers/net/wireless/iwlwifi/mvm/rx.c
+++ b/drivers/net/wireless/iwlwifi/mvm/rx.c
@@ -363,3 +363,25 @@ int iwl_mvm_rx_rx_mpdu(struct iwl_mvm *mvm, struct iwl_rx_cmd_buffer *rxb,
 					rxb, &rx_status);
 	return 0;
 }
+
+/*
+ * iwl_mvm_rx_statistics - STATISTICS_NOTIFICATION handler
+ *
+ * TODO: This handler is implemented partially.
+ * It only gets the NIC's temperature.
+ */
+int iwl_mvm_rx_statistics(struct iwl_mvm *mvm,
+			  struct iwl_rx_cmd_buffer *rxb,
+			  struct iwl_device_cmd *cmd)
+{
+	struct iwl_rx_packet *pkt = rxb_addr(rxb);
+	struct iwl_notif_statistics *stats = (void *)&pkt->data;
+	struct mvm_statistics_general_common *common = &stats->general.common;
+
+	if (mvm->temperature != le32_to_cpu(common->temperature)) {
+		mvm->temperature = le32_to_cpu(common->temperature);
+		iwl_mvm_tt_handler(mvm);
+	}
+
+	return 0;
+}
diff --git a/drivers/net/wireless/iwlwifi/mvm/scan.c b/drivers/net/wireless/iwlwifi/mvm/scan.c
index 2476e43..2157b0f 100644
--- a/drivers/net/wireless/iwlwifi/mvm/scan.c
+++ b/drivers/net/wireless/iwlwifi/mvm/scan.c
@@ -298,12 +298,6 @@ int iwl_mvm_scan_request(struct iwl_mvm *mvm,
 	else
 		cmd->type = cpu_to_le32(SCAN_TYPE_FORCED);
 
-	/*
-	 * TODO: This is a WA due to a bug in the FW AUX framework that does not
-	 * properly handle time events that fail to be scheduled
-	 */
-	cmd->type = cpu_to_le32(SCAN_TYPE_FORCED);
-
 	cmd->repeats = cpu_to_le32(1);
 
 	/*
diff --git a/drivers/net/wireless/iwlwifi/mvm/sta.c b/drivers/net/wireless/iwlwifi/mvm/sta.c
index 5c664ed..62fe520 100644
--- a/drivers/net/wireless/iwlwifi/mvm/sta.c
+++ b/drivers/net/wireless/iwlwifi/mvm/sta.c
@@ -64,6 +64,7 @@
 
 #include "mvm.h"
 #include "sta.h"
+#include "rs.h"
 
 static int iwl_mvm_find_free_sta_id(struct iwl_mvm *mvm)
 {
@@ -217,6 +218,8 @@ int iwl_mvm_add_sta(struct iwl_mvm *mvm,
 						      mvmvif->color);
 	mvm_sta->vif = vif;
 	mvm_sta->max_agg_bufsize = LINK_QUAL_AGG_FRAME_LIMIT_DEF;
+	mvm_sta->tx_protection = 0;
+	mvm_sta->tt_tx_protection = false;
 
 	/* HW restart, don't assume the memory has been zeroed */
 	atomic_set(&mvm->pending_frames[sta_id], 0);
@@ -226,9 +229,6 @@ int iwl_mvm_add_sta(struct iwl_mvm *mvm,
 		if (vif->hw_queue[i] != IEEE80211_INVAL_HW_QUEUE)
 			mvm_sta->tfd_queue_msk |= BIT(vif->hw_queue[i]);
 
-	if (vif->cab_queue != IEEE80211_INVAL_HW_QUEUE)
-		mvm_sta->tfd_queue_msk |= BIT(vif->cab_queue);
-
 	/* for HW restart - need to reset the seq_number etc... */
 	memset(mvm_sta->tid_data, 0, sizeof(mvm_sta->tid_data));
 
@@ -798,21 +798,23 @@ int iwl_mvm_sta_tx_agg_oper(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
 		min(mvmsta->max_agg_bufsize, buf_size);
 	mvmsta->lq_sta.lq.agg_frame_cnt_limit = mvmsta->max_agg_bufsize;
 
+	IWL_DEBUG_HT(mvm, "Tx aggregation enabled on ra = %pM tid = %d\n",
+		     sta->addr, tid);
+
 	if (mvm->cfg->ht_params->use_rts_for_aggregation) {
 		/*
 		 * switch to RTS/CTS if it is the prefer protection
 		 * method for HT traffic
+		 * this function also sends the LQ command
 		 */
-		mvmsta->lq_sta.lq.flags |= LQ_FLAG_SET_STA_TLC_RTS_MSK;
+		return iwl_mvm_tx_protection(mvm, &mvmsta->lq_sta.lq,
+					     mvmsta, true);
 		/*
 		 * TODO: remove the TLC_RTS flag when we tear down the last
 		 * AGG session (agg_tids_count in DVM)
 		 */
 	}
 
-	IWL_DEBUG_HT(mvm, "Tx aggregation enabled on ra = %pM tid = %d\n",
-		     sta->addr, tid);
-
 	return iwl_mvm_send_lq_cmd(mvm, &mvmsta->lq_sta.lq, CMD_ASYNC, false);
 }
 
@@ -1287,17 +1289,11 @@ void iwl_mvm_sta_modify_ps_wake(struct iwl_mvm *mvm,
 	struct iwl_mvm_add_sta_cmd cmd = {
 		.add_modify = STA_MODE_MODIFY,
 		.sta_id = mvmsta->sta_id,
-		.modify_mask = STA_MODIFY_SLEEPING_STA_TX_COUNT,
-		.sleep_state_flags = cpu_to_le16(STA_SLEEP_STATE_AWAKE),
+		.station_flags_msk = cpu_to_le32(STA_FLG_PS),
 		.mac_id_n_color = cpu_to_le32(mvmsta->mac_id_n_color),
 	};
 	int ret;
 
-	/*
-	 * Same modify mask for sleep_tx_count and sleep_state_flags but this
-	 * should be fine since if we set the STA as "awake", then
-	 * sleep_tx_count is not relevant.
-	 */
 	ret = iwl_mvm_send_cmd_pdu(mvm, ADD_STA, CMD_ASYNC, sizeof(cmd), &cmd);
 	if (ret)
 		IWL_ERR(mvm, "Failed to send ADD_STA command (%d)\n", ret);
diff --git a/drivers/net/wireless/iwlwifi/mvm/sta.h b/drivers/net/wireless/iwlwifi/mvm/sta.h
index a4ddce7..94b265e 100644
--- a/drivers/net/wireless/iwlwifi/mvm/sta.h
+++ b/drivers/net/wireless/iwlwifi/mvm/sta.h
@@ -250,7 +250,6 @@ enum iwl_mvm_agg_state {
  *	the first packet to be sent in legacy HW queue in Tx AGG stop flow.
  *	Basically when next_reclaimed reaches ssn, we can tell mac80211 that
  *	we are ready to finish the Tx AGG stop / start flow.
- * @wait_for_ba: Expect block-ack before next Tx reply
  */
 struct iwl_mvm_tid_data {
 	u16 seq_number;
@@ -260,7 +259,6 @@ struct iwl_mvm_tid_data {
 	enum iwl_mvm_agg_state state;
 	u16 txq_id;
 	u16 ssn;
-	bool wait_for_ba;
 };
 
 /**
@@ -275,6 +273,8 @@ struct iwl_mvm_tid_data {
  * @lock: lock to protect the whole struct. Since %tid_data is access from Tx
  * and from Tx response flow, it needs a spinlock.
  * @tid_data: per tid data. Look at %iwl_mvm_tid_data.
+ * @tx_protection: reference counter for controlling the Tx protection.
+ * @tt_tx_protection: is thermal throttling enable Tx protection?
  *
  * When mac80211 creates a station it reserves some space (hw->sta_data_size)
  * in the structure for use by driver. This structure is placed in that
@@ -296,6 +296,10 @@ struct iwl_mvm_sta {
 #ifdef CONFIG_PM_SLEEP
 	u16 last_seq_ctl;
 #endif
+
+	/* Temporary, until the new TLC will control the Tx protection */
+	s8 tx_protection;
+	bool tt_tx_protection;
 };
 
 /**
diff --git a/drivers/net/wireless/iwlwifi/mvm/tt.c b/drivers/net/wireless/iwlwifi/mvm/tt.c
new file mode 100644
index 0000000..d6ae7f1
--- /dev/null
+++ b/drivers/net/wireless/iwlwifi/mvm/tt.c
@@ -0,0 +1,530 @@
+/******************************************************************************
+ *
+ * This file is provided under a dual BSD/GPLv2 license.  When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * Copyright(c) 2013 Intel Corporation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110,
+ * USA
+ *
+ * The full GNU General Public License is included in this distribution
+ * in the file called COPYING.
+ *
+ * Contact Information:
+ *  Intel Linux Wireless <ilw@linux.intel.com>
+ * Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+ *
+ * BSD LICENSE
+ *
+ * Copyright(c) 2012 - 2013 Intel Corporation. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ *  * Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ *  * Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *  * Neither the name Intel Corporation nor the names of its
+ *    contributors may be used to endorse or promote products derived
+ *    from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ *****************************************************************************/
+
+#include "mvm.h"
+#include "iwl-config.h"
+#include "iwl-io.h"
+#include "iwl-csr.h"
+#include "iwl-prph.h"
+
+#define OTP_DTS_DIODE_DEVIATION 96 /*in words*/
+/* VBG - Voltage Band Gap error data (temperature offset) */
+#define OTP_WP_DTS_VBG			(OTP_DTS_DIODE_DEVIATION + 2)
+#define MEAS_VBG_MIN_VAL		2300
+#define MEAS_VBG_MAX_VAL		3000
+#define MEAS_VBG_DEFAULT_VAL		2700
+#define DTS_DIODE_VALID(flags)		(flags & DTS_DIODE_REG_FLAGS_PASS_ONCE)
+#define MIN_TEMPERATURE			0
+#define MAX_TEMPERATURE			125
+#define TEMPERATURE_ERROR		(MAX_TEMPERATURE + 1)
+#define PTAT_DIGITAL_VALUE_MIN_VALUE	0
+#define PTAT_DIGITAL_VALUE_MAX_VALUE	0xFF
+#define DTS_VREFS_NUM			5
+static inline u32 DTS_DIODE_GET_VREFS_ID(u32 flags)
+{
+	return (flags & DTS_DIODE_REG_FLAGS_VREFS_ID) >>
+					DTS_DIODE_REG_FLAGS_VREFS_ID_POS;
+}
+
+#define CALC_VREFS_MIN_DIFF	43
+#define CALC_VREFS_MAX_DIFF	51
+#define CALC_LUT_SIZE		(1 + CALC_VREFS_MAX_DIFF - CALC_VREFS_MIN_DIFF)
+#define CALC_LUT_INDEX_OFFSET	CALC_VREFS_MIN_DIFF
+#define CALC_TEMPERATURE_RESULT_SHIFT_OFFSET	23
+
+/*
+ * @digital_value: The diode's digital-value sampled (temperature/voltage)
+ * @vref_low: The lower voltage-reference (the vref just below the diode's
+ *	sampled digital-value)
+ * @vref_high: The higher voltage-reference (the vref just above the diode's
+ *	sampled digital-value)
+ * @flags: bits[1:0]: The ID of the Vrefs pair (lowVref,highVref)
+ *	bits[6:2]: Reserved.
+ *	bits[7:7]: Indicates completion of at least 1 successful sample
+ *	since last DTS reset.
+ */
+struct iwl_mvm_dts_diode_bits {
+	u8 digital_value;
+	u8 vref_low;
+	u8 vref_high;
+	u8 flags;
+} __packed;
+
+union dts_diode_results {
+	u32 reg_value;
+	struct iwl_mvm_dts_diode_bits bits;
+} __packed;
+
+static s16 iwl_mvm_dts_get_volt_band_gap(struct iwl_mvm *mvm)
+{
+	struct iwl_nvm_section calib_sec;
+	const __le16 *calib;
+	u16 vbg;
+
+	/* TODO: move parsing to NVM code */
+	calib_sec = mvm->nvm_sections[NVM_SECTION_TYPE_CALIBRATION];
+	calib = (__le16 *)calib_sec.data;
+
+	vbg = le16_to_cpu(calib[OTP_WP_DTS_VBG]);
+
+	if (vbg < MEAS_VBG_MIN_VAL || vbg > MEAS_VBG_MAX_VAL)
+		vbg = MEAS_VBG_DEFAULT_VAL;
+
+	return vbg;
+}
+
+static u16 iwl_mvm_dts_get_ptat_deviation_offset(struct iwl_mvm *mvm)
+{
+	const u8 *calib;
+	u8 ptat, pa1, pa2, median;
+
+	/* TODO: move parsing to NVM code */
+	calib = mvm->nvm_sections[NVM_SECTION_TYPE_CALIBRATION].data;
+	ptat = calib[OTP_DTS_DIODE_DEVIATION];
+	pa1 = calib[OTP_DTS_DIODE_DEVIATION + 1];
+	pa2 = calib[OTP_DTS_DIODE_DEVIATION + 2];
+
+	/* get the median: */
+	if (ptat > pa1) {
+		if (ptat > pa2)
+			median = (pa1 > pa2) ? pa1 : pa2;
+		else
+			median = ptat;
+	} else {
+		if (pa1 > pa2)
+			median = (ptat > pa2) ? ptat : pa2;
+		else
+			median = pa1;
+	}
+
+	return ptat - median;
+}
+
+static u8 iwl_mvm_dts_calibrate_ptat_deviation(struct iwl_mvm *mvm, u8 value)
+{
+	/* Calibrate the PTAT digital value, based on PTAT deviation data: */
+	s16 new_val = value - iwl_mvm_dts_get_ptat_deviation_offset(mvm);
+
+	if (new_val > PTAT_DIGITAL_VALUE_MAX_VALUE)
+		new_val = PTAT_DIGITAL_VALUE_MAX_VALUE;
+	else if (new_val < PTAT_DIGITAL_VALUE_MIN_VALUE)
+		new_val = PTAT_DIGITAL_VALUE_MIN_VALUE;
+
+	return new_val;
+}
+
+static bool dts_get_adjacent_vrefs(struct iwl_mvm *mvm,
+				   union dts_diode_results *avg_ptat)
+{
+	u8 vrefs_results[DTS_VREFS_NUM];
+	u8 low_vref_index = 0, flags;
+	u32 reg;
+
+	reg = iwl_read_prph(mvm->trans, DTSC_VREF_AVG);
+	memcpy(vrefs_results, &reg, sizeof(reg));
+	reg = iwl_read_prph(mvm->trans, DTSC_VREF5_AVG);
+	vrefs_results[4] = reg & 0xff;
+
+	if (avg_ptat->bits.digital_value < vrefs_results[0] ||
+	    avg_ptat->bits.digital_value > vrefs_results[4])
+		return false;
+
+	if (avg_ptat->bits.digital_value > vrefs_results[3])
+		low_vref_index = 3;
+	else if (avg_ptat->bits.digital_value > vrefs_results[2])
+		low_vref_index = 2;
+	else if (avg_ptat->bits.digital_value > vrefs_results[1])
+		low_vref_index = 1;
+
+	avg_ptat->bits.vref_low  = vrefs_results[low_vref_index];
+	avg_ptat->bits.vref_high = vrefs_results[low_vref_index + 1];
+	flags = avg_ptat->bits.flags;
+	avg_ptat->bits.flags =
+		(flags & ~DTS_DIODE_REG_FLAGS_VREFS_ID) |
+		(low_vref_index & DTS_DIODE_REG_FLAGS_VREFS_ID);
+	return true;
+}
+
+/*
+ * return true it the results are valid, and false otherwise.
+ */
+static bool dts_read_ptat_avg_results(struct iwl_mvm *mvm,
+				      union dts_diode_results *avg_ptat)
+{
+	u32 reg;
+	u8 tmp;
+
+	/* fill the diode value and pass_once with avg-reg results */
+	reg = iwl_read_prph(mvm->trans, DTSC_PTAT_AVG);
+	reg &= DTS_DIODE_REG_DIG_VAL | DTS_DIODE_REG_PASS_ONCE;
+	avg_ptat->reg_value = reg;
+
+	/* calibrate the PTAT digital value */
+	tmp = avg_ptat->bits.digital_value;
+	tmp = iwl_mvm_dts_calibrate_ptat_deviation(mvm, tmp);
+	avg_ptat->bits.digital_value = tmp;
+
+	/*
+	 * fill vrefs fields, based on the avgVrefs results
+	 * and the diode value
+	 */
+	return dts_get_adjacent_vrefs(mvm, avg_ptat) &&
+		DTS_DIODE_VALID(avg_ptat->bits.flags);
+}
+
+static s32 calculate_nic_temperature(union dts_diode_results avg_ptat,
+				     u16 volt_band_gap)
+{
+	u32 tmp_result;
+	u8 vrefs_diff;
+	/*
+	 * For temperature calculation (at the end, shift right by 23)
+	 * LUT[(D2-D1)] = ROUND{ 2^23 / ((D2-D1)*9*10) }
+	 * (D2-D1) ==   43    44    45    46    47    48    49    50    51
+	 */
+	static const u16 calc_lut[CALC_LUT_SIZE] = {
+		2168, 2118, 2071, 2026, 1983, 1942, 1902, 1864, 1828,
+	};
+
+	/*
+	 * The diff between the high and low voltage-references is assumed
+	 * to be strictly be in range of [60,68]
+	 */
+	vrefs_diff = avg_ptat.bits.vref_high - avg_ptat.bits.vref_low;
+
+	if (vrefs_diff < CALC_VREFS_MIN_DIFF ||
+	    vrefs_diff > CALC_VREFS_MAX_DIFF)
+		return TEMPERATURE_ERROR;
+
+	/* calculate the result: */
+	tmp_result =
+		vrefs_diff * (DTS_DIODE_GET_VREFS_ID(avg_ptat.bits.flags) + 9);
+	tmp_result += avg_ptat.bits.digital_value;
+	tmp_result -= avg_ptat.bits.vref_high;
+
+	/* multiply by the LUT value (based on the diff) */
+	tmp_result *= calc_lut[vrefs_diff - CALC_LUT_INDEX_OFFSET];
+
+	/*
+	 * Get the BandGap (the voltage refereces source) error data
+	 * (temperature offset)
+	 */
+	tmp_result *= volt_band_gap;
+
+	/*
+	 * here, tmp_result value can be up to 32-bits. We want to right-shift
+	 * it *without* sign-extend.
+	 */
+	tmp_result = tmp_result >> CALC_TEMPERATURE_RESULT_SHIFT_OFFSET;
+
+	/*
+	 * at this point, tmp_result should be in the range:
+	 * 200 <= tmp_result <= 365
+	 */
+	return (s16)tmp_result - 240;
+}
+
+static s32 check_nic_temperature(struct iwl_mvm *mvm)
+{
+	u16 volt_band_gap;
+	union dts_diode_results avg_ptat;
+
+	volt_band_gap = iwl_mvm_dts_get_volt_band_gap(mvm);
+
+	/* disable DTS */
+	iwl_write_prph(mvm->trans, SHR_MISC_WFM_DTS_EN, 0);
+
+	/* SV initialization */
+	iwl_write_prph(mvm->trans, SHR_MISC_WFM_DTS_EN, 1);
+	iwl_write_prph(mvm->trans, DTSC_CFG_MODE,
+		       DTSC_CFG_MODE_PERIODIC);
+
+	/* wait for results */
+	msleep(100);
+	if (!dts_read_ptat_avg_results(mvm, &avg_ptat))
+		return TEMPERATURE_ERROR;
+
+	/* disable DTS */
+	iwl_write_prph(mvm->trans, SHR_MISC_WFM_DTS_EN, 0);
+
+	return calculate_nic_temperature(avg_ptat, volt_band_gap);
+}
+
+static void iwl_mvm_enter_ctkill(struct iwl_mvm *mvm)
+{
+	u32 duration = mvm->thermal_throttle.params->ct_kill_duration;
+
+	IWL_ERR(mvm, "Enter CT Kill\n");
+	iwl_mvm_set_hw_ctkill_state(mvm, true);
+	schedule_delayed_work(&mvm->thermal_throttle.ct_kill_exit,
+			      round_jiffies_relative(duration * HZ));
+}
+
+static void iwl_mvm_exit_ctkill(struct iwl_mvm *mvm)
+{
+	IWL_ERR(mvm, "Exit CT Kill\n");
+	iwl_mvm_set_hw_ctkill_state(mvm, false);
+}
+
+static void check_exit_ctkill(struct work_struct *work)
+{
+	struct iwl_mvm_tt_mgmt *tt;
+	struct iwl_mvm *mvm;
+	u32 duration;
+	s32 temp;
+
+	tt = container_of(work, struct iwl_mvm_tt_mgmt, ct_kill_exit.work);
+	mvm = container_of(tt, struct iwl_mvm, thermal_throttle);
+
+	duration = tt->params->ct_kill_duration;
+
+	iwl_trans_start_hw(mvm->trans);
+	temp = check_nic_temperature(mvm);
+	iwl_trans_stop_hw(mvm->trans, false);
+
+	if (temp < MIN_TEMPERATURE || temp > MAX_TEMPERATURE) {
+		IWL_DEBUG_TEMP(mvm, "Failed to measure NIC temperature\n");
+		goto reschedule;
+	}
+	IWL_DEBUG_TEMP(mvm, "NIC temperature: %d\n", temp);
+
+	if (temp <= tt->params->ct_kill_exit) {
+		iwl_mvm_exit_ctkill(mvm);
+		return;
+	}
+
+reschedule:
+	schedule_delayed_work(&mvm->thermal_throttle.ct_kill_exit,
+			      round_jiffies(duration * HZ));
+}
+
+static void iwl_mvm_tt_smps_iterator(void *_data, u8 *mac,
+				     struct ieee80211_vif *vif)
+{
+	struct iwl_mvm *mvm = _data;
+	enum ieee80211_smps_mode smps_mode;
+
+	lockdep_assert_held(&mvm->mutex);
+
+	if (mvm->thermal_throttle.dynamic_smps)
+		smps_mode = IEEE80211_SMPS_DYNAMIC;
+	else
+		smps_mode = IEEE80211_SMPS_AUTOMATIC;
+
+	if (vif->type != NL80211_IFTYPE_STATION)
+		return;
+
+	iwl_mvm_update_smps(mvm, vif, IWL_MVM_SMPS_REQ_TT, smps_mode);
+}
+
+static void iwl_mvm_tt_tx_protection(struct iwl_mvm *mvm, bool enable)
+{
+	struct ieee80211_sta *sta;
+	struct iwl_mvm_sta *mvmsta;
+	int i, err;
+
+	for (i = 0; i < IWL_MVM_STATION_COUNT; i++) {
+		sta = rcu_dereference_protected(mvm->fw_id_to_mac_id[i],
+						lockdep_is_held(&mvm->mutex));
+		if (IS_ERR_OR_NULL(sta))
+			continue;
+		mvmsta = (void *)sta->drv_priv;
+		if (enable == mvmsta->tt_tx_protection)
+			continue;
+		err = iwl_mvm_tx_protection(mvm, &mvmsta->lq_sta.lq,
+					    mvmsta, enable);
+		if (err) {
+			IWL_ERR(mvm, "Failed to %s Tx protection\n",
+				enable ? "enable" : "disable");
+		} else {
+			IWL_DEBUG_TEMP(mvm, "%s Tx protection\n",
+				       enable ? "Enable" : "Disable");
+			mvmsta->tt_tx_protection = enable;
+		}
+	}
+}
+
+static void iwl_mvm_tt_tx_backoff(struct iwl_mvm *mvm, u32 backoff)
+{
+	struct iwl_host_cmd cmd = {
+		.id = REPLY_THERMAL_MNG_BACKOFF,
+		.len = { sizeof(u32), },
+		.data = { &backoff, },
+		.flags = CMD_SYNC,
+	};
+
+	if (iwl_mvm_send_cmd(mvm, &cmd) == 0) {
+		IWL_DEBUG_TEMP(mvm, "Set Thermal Tx backoff to: %u\n",
+			       backoff);
+		mvm->thermal_throttle.tx_backoff = backoff;
+	} else {
+		IWL_ERR(mvm, "Failed to change Thermal Tx backoff\n");
+	}
+}
+
+void iwl_mvm_tt_handler(struct iwl_mvm *mvm)
+{
+	const struct iwl_tt_params *params = mvm->thermal_throttle.params;
+	struct iwl_mvm_tt_mgmt *tt = &mvm->thermal_throttle;
+	s32 temperature = mvm->temperature;
+	bool throttle_enable = false;
+	int i;
+	u32 tx_backoff;
+
+	IWL_DEBUG_TEMP(mvm, "NIC temperature: %d\n", mvm->temperature);
+
+	if (params->support_ct_kill && temperature >= params->ct_kill_entry) {
+		iwl_mvm_enter_ctkill(mvm);
+		return;
+	}
+
+	if (params->support_dynamic_smps) {
+		if (!tt->dynamic_smps &&
+		    temperature >= params->dynamic_smps_entry) {
+			IWL_DEBUG_TEMP(mvm, "Enable dynamic SMPS\n");
+			tt->dynamic_smps = true;
+			ieee80211_iterate_active_interfaces_atomic(
+					mvm->hw, IEEE80211_IFACE_ITER_NORMAL,
+					iwl_mvm_tt_smps_iterator, mvm);
+			throttle_enable = true;
+		} else if (tt->dynamic_smps &&
+			   temperature <= params->dynamic_smps_exit) {
+			IWL_DEBUG_TEMP(mvm, "Disable dynamic SMPS\n");
+			tt->dynamic_smps = false;
+			ieee80211_iterate_active_interfaces_atomic(
+					mvm->hw, IEEE80211_IFACE_ITER_NORMAL,
+					iwl_mvm_tt_smps_iterator, mvm);
+		}
+	}
+
+	if (params->support_tx_protection) {
+		if (temperature >= params->tx_protection_entry) {
+			iwl_mvm_tt_tx_protection(mvm, true);
+			throttle_enable = true;
+		} else if (temperature <= params->tx_protection_exit) {
+			iwl_mvm_tt_tx_protection(mvm, false);
+		}
+	}
+
+	if (params->support_tx_backoff) {
+		tx_backoff = 0;
+		for (i = 0; i < TT_TX_BACKOFF_SIZE; i++) {
+			if (temperature < params->tx_backoff[i].temperature)
+				break;
+			tx_backoff = params->tx_backoff[i].backoff;
+		}
+		if (tx_backoff != 0)
+			throttle_enable = true;
+		if (tt->tx_backoff != tx_backoff)
+			iwl_mvm_tt_tx_backoff(mvm, tx_backoff);
+	}
+
+	if (!tt->throttle && throttle_enable) {
+		IWL_WARN(mvm,
+			 "Due to high temperature thermal throttling initiated\n");
+		tt->throttle = true;
+	} else if (tt->throttle && !tt->dynamic_smps && tt->tx_backoff == 0 &&
+		   temperature <= params->tx_protection_exit) {
+		IWL_WARN(mvm,
+			 "Temperature is back to normal thermal throttling stopped\n");
+		tt->throttle = false;
+	}
+}
+
+static const struct iwl_tt_params iwl7000_tt_params = {
+	.ct_kill_entry = 118,
+	.ct_kill_exit = 96,
+	.ct_kill_duration = 5,
+	.dynamic_smps_entry = 114,
+	.dynamic_smps_exit = 110,
+	.tx_protection_entry = 114,
+	.tx_protection_exit = 108,
+	.tx_backoff = {
+		{.temperature = 112, .backoff = 200},
+		{.temperature = 113, .backoff = 600},
+		{.temperature = 114, .backoff = 1200},
+		{.temperature = 115, .backoff = 2000},
+		{.temperature = 116, .backoff = 4000},
+		{.temperature = 117, .backoff = 10000},
+	},
+	.support_ct_kill = true,
+	.support_dynamic_smps = true,
+	.support_tx_protection = true,
+	.support_tx_backoff = true,
+};
+
+void iwl_mvm_tt_initialize(struct iwl_mvm *mvm)
+{
+	struct iwl_mvm_tt_mgmt *tt = &mvm->thermal_throttle;
+
+	IWL_DEBUG_TEMP(mvm, "Initialize Thermal Throttling\n");
+	tt->params = &iwl7000_tt_params;
+	tt->throttle = false;
+	INIT_DELAYED_WORK(&tt->ct_kill_exit, check_exit_ctkill);
+}
+
+void iwl_mvm_tt_exit(struct iwl_mvm *mvm)
+{
+	cancel_delayed_work_sync(&mvm->thermal_throttle.ct_kill_exit);
+	IWL_DEBUG_TEMP(mvm, "Exit Thermal Throttling\n");
+}
diff --git a/drivers/net/wireless/iwlwifi/mvm/tx.c b/drivers/net/wireless/iwlwifi/mvm/tx.c
index 48c1891..f0e96a9 100644
--- a/drivers/net/wireless/iwlwifi/mvm/tx.c
+++ b/drivers/net/wireless/iwlwifi/mvm/tx.c
@@ -175,7 +175,7 @@ static void iwl_mvm_set_tx_cmd_rate(struct iwl_mvm *mvm,
 	 * table is controlled by LINK_QUALITY commands
 	 */
 
-	if (ieee80211_is_data(fc)) {
+	if (ieee80211_is_data(fc) && sta) {
 		tx_cmd->initial_rate_index = 0;
 		tx_cmd->tx_flags |= cpu_to_le32(TX_CMD_FLG_STA_RATE);
 		return;
@@ -408,7 +408,6 @@ int iwl_mvm_tx_skb(struct iwl_mvm *mvm, struct sk_buff *skb,
 	IWL_DEBUG_TX(mvm, "TX to [%d|%d] Q:%d - seq: 0x%x\n", mvmsta->sta_id,
 		     tid, txq_id, seq_number);
 
-	/* NOTE: aggregation will need changes here (for txq id) */
 	if (iwl_trans_tx(mvm->trans, skb, dev_cmd, txq_id))
 		goto drop_unlock_sta;
 
@@ -610,8 +609,8 @@ static void iwl_mvm_rx_tx_cmd_single(struct iwl_mvm *mvm,
 		    !(info->flags & IEEE80211_TX_STAT_ACK))
 			info->flags |= IEEE80211_TX_STAT_AMPDU_NO_BACK;
 
-		/* W/A FW bug: seq_ctl is wrong when the queue is flushed */
-		if (status == TX_STATUS_FAIL_FIFO_FLUSHED) {
+		/* W/A FW bug: seq_ctl is wrong when the status isn't success */
+		if (status != TX_STATUS_SUCCESS) {
 			struct ieee80211_hdr *hdr = (void *)skb->data;
 			seq_ctl = le16_to_cpu(hdr->seq_ctrl);
 		}
diff --git a/drivers/net/wireless/iwlwifi/mvm/utils.c b/drivers/net/wireless/iwlwifi/mvm/utils.c
index 687b34e..1e13328 100644
--- a/drivers/net/wireless/iwlwifi/mvm/utils.c
+++ b/drivers/net/wireless/iwlwifi/mvm/utils.c
@@ -76,6 +76,11 @@ int iwl_mvm_send_cmd(struct iwl_mvm *mvm, struct iwl_host_cmd *cmd)
 {
 	int ret;
 
+#if defined(CONFIG_IWLWIFI_DEBUGFS) && defined(CONFIG_PM_SLEEP)
+	if (WARN_ON(mvm->d3_test_active))
+		return -EIO;
+#endif
+
 	/*
 	 * Synchronous commands from this op-mode must hold
 	 * the mutex, this ensures we don't try to send two
@@ -125,6 +130,11 @@ int iwl_mvm_send_cmd_status(struct iwl_mvm *mvm, struct iwl_host_cmd *cmd,
 
 	lockdep_assert_held(&mvm->mutex);
 
+#if defined(CONFIG_IWLWIFI_DEBUGFS) && defined(CONFIG_PM_SLEEP)
+	if (WARN_ON(mvm->d3_test_active))
+		return -EIO;
+#endif
+
 	/*
 	 * Only synchronous commands can wait for status,
 	 * we use WANT_SKB so the caller can't.
@@ -471,3 +481,34 @@ int iwl_mvm_send_lq_cmd(struct iwl_mvm *mvm, struct iwl_lq_cmd *lq,
 
 	return iwl_mvm_send_cmd(mvm, &cmd);
 }
+
+/**
+ * iwl_mvm_update_smps - Get a requst to change the SMPS mode
+ * @req_type: The part of the driver who call for a change.
+ * @smps_requests: The request to change the SMPS mode.
+ *
+ * Get a requst to change the SMPS mode,
+ * and change it according to all other requests in the driver.
+ */
+void iwl_mvm_update_smps(struct iwl_mvm *mvm, struct ieee80211_vif *vif,
+			 enum iwl_mvm_smps_type_request req_type,
+			 enum ieee80211_smps_mode smps_request)
+{
+	struct iwl_mvm_vif *mvmvif;
+	enum ieee80211_smps_mode smps_mode = IEEE80211_SMPS_AUTOMATIC;
+	int i;
+
+	lockdep_assert_held(&mvm->mutex);
+	mvmvif = iwl_mvm_vif_from_mac80211(vif);
+	mvmvif->smps_requests[req_type] = smps_request;
+	for (i = 0; i < NUM_IWL_MVM_SMPS_REQ; i++) {
+		if (mvmvif->smps_requests[i] == IEEE80211_SMPS_STATIC) {
+			smps_mode = IEEE80211_SMPS_STATIC;
+			break;
+		}
+		if (mvmvif->smps_requests[i] == IEEE80211_SMPS_DYNAMIC)
+			smps_mode = IEEE80211_SMPS_DYNAMIC;
+	}
+
+	ieee80211_request_smps(vif, smps_mode);
+}
diff --git a/drivers/net/wireless/iwlwifi/pcie/drv.c b/drivers/net/wireless/iwlwifi/pcie/drv.c
index 8cb53ec..81f3ea5 100644
--- a/drivers/net/wireless/iwlwifi/pcie/drv.c
+++ b/drivers/net/wireless/iwlwifi/pcie/drv.c
@@ -78,6 +78,7 @@
 
 /* Hardware specific file defines the PCI IDs table for that hardware module */
 static DEFINE_PCI_DEVICE_TABLE(iwl_hw_card_ids) = {
+#if IS_ENABLED(CONFIG_IWLDVM)
 	{IWL_PCI_DEVICE(0x4232, 0x1201, iwl5100_agn_cfg)}, /* Mini Card */
 	{IWL_PCI_DEVICE(0x4232, 0x1301, iwl5100_agn_cfg)}, /* Half Mini Card */
 	{IWL_PCI_DEVICE(0x4232, 0x1204, iwl5100_agn_cfg)}, /* Mini Card */
@@ -253,13 +254,60 @@ static DEFINE_PCI_DEVICE_TABLE(iwl_hw_card_ids) = {
 	{IWL_PCI_DEVICE(0x0892, 0x0062, iwl135_bgn_cfg)},
 	{IWL_PCI_DEVICE(0x0893, 0x0262, iwl135_bgn_cfg)},
 	{IWL_PCI_DEVICE(0x0892, 0x0462, iwl135_bgn_cfg)},
+#endif /* CONFIG_IWLDVM */
 
+#if IS_ENABLED(CONFIG_IWLMVM)
 /* 7000 Series */
 	{IWL_PCI_DEVICE(0x08B1, 0x4070, iwl7260_2ac_cfg)},
-	{IWL_PCI_DEVICE(0x08B1, 0x4062, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4170, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4060, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4160, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4062, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4162, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0x4270, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0x4260, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0x4262, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4470, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4460, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4462, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4870, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x486E, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4A70, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4A6E, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4A6C, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4020, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0x4220, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0x4420, iwl7260_2n_cfg)},
 	{IWL_PCI_DEVICE(0x08B1, 0xC070, iwl7260_2ac_cfg)},
-	{IWL_PCI_DEVICE(0x08B3, 0x0070, iwl3160_ac_cfg)},
-	{IWL_PCI_DEVICE(0x08B3, 0x8070, iwl3160_ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC170, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC060, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC160, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC062, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC162, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0xC270, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0xC260, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0xC262, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC470, iwl7260_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC460, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC462, iwl7260_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC020, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B2, 0xC220, iwl7260_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B1, 0xC420, iwl7260_2n_cfg)},
+
+/* 3160 Series */
+	{IWL_PCI_DEVICE(0x08B3, 0x0070, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x0170, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x0060, iwl3160_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x0062, iwl3160_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B4, 0x0270, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x0470, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x8070, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x8170, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x8060, iwl3160_2n_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x8062, iwl3160_n_cfg)},
+	{IWL_PCI_DEVICE(0x08B4, 0x8270, iwl3160_2ac_cfg)},
+	{IWL_PCI_DEVICE(0x08B3, 0x8470, iwl3160_2ac_cfg)},
+#endif /* CONFIG_IWLMVM */
 
 	{0}
 };
diff --git a/drivers/net/wireless/iwlwifi/pcie/internal.h b/drivers/net/wireless/iwlwifi/pcie/internal.h
index 148843e..b654dcd 100644
--- a/drivers/net/wireless/iwlwifi/pcie/internal.h
+++ b/drivers/net/wireless/iwlwifi/pcie/internal.h
@@ -217,6 +217,7 @@ struct iwl_pcie_txq_scratch_buf {
  * @trans_pcie: pointer back to transport (for timer)
  * @need_update: indicates need to update read/write index
  * @active: stores if queue is active
+ * @ampdu: true if this queue is an ampdu queue for an specific RA/TID
  *
  * A Tx queue consists of circular buffer of BDs (a.k.a. TFDs, transmit frame
  * descriptors) and required locking structures.
@@ -232,6 +233,7 @@ struct iwl_txq {
 	struct iwl_trans_pcie *trans_pcie;
 	u8 need_update;
 	u8 active;
+	bool ampdu;
 };
 
 static inline dma_addr_t
diff --git a/drivers/net/wireless/iwlwifi/pcie/rx.c b/drivers/net/wireless/iwlwifi/pcie/rx.c
index 567e67a..fd848cd 100644
--- a/drivers/net/wireless/iwlwifi/pcie/rx.c
+++ b/drivers/net/wireless/iwlwifi/pcie/rx.c
@@ -110,9 +110,10 @@
 /*
  * iwl_rxq_space - Return number of free slots available in queue.
  */
-static int iwl_rxq_space(const struct iwl_rxq *q)
+static int iwl_rxq_space(const struct iwl_rxq *rxq)
 {
-	int s = q->read - q->write;
+	int s = rxq->read - rxq->write;
+
 	if (s <= 0)
 		s += RX_QUEUE_SIZE;
 	/* keep some buffer to not confuse full and empty queue */
@@ -143,21 +144,22 @@ int iwl_pcie_rx_stop(struct iwl_trans *trans)
 /*
  * iwl_pcie_rxq_inc_wr_ptr - Update the write pointer for the RX queue
  */
-static void iwl_pcie_rxq_inc_wr_ptr(struct iwl_trans *trans, struct iwl_rxq *q)
+static void iwl_pcie_rxq_inc_wr_ptr(struct iwl_trans *trans,
+				    struct iwl_rxq *rxq)
 {
 	unsigned long flags;
 	u32 reg;
 
-	spin_lock_irqsave(&q->lock, flags);
+	spin_lock_irqsave(&rxq->lock, flags);
 
-	if (q->need_update == 0)
+	if (rxq->need_update == 0)
 		goto exit_unlock;
 
 	if (trans->cfg->base_params->shadow_reg_enable) {
 		/* shadow register enabled */
 		/* Device expects a multiple of 8 */
-		q->write_actual = (q->write & ~0x7);
-		iwl_write32(trans, FH_RSCSR_CHNL0_WPTR, q->write_actual);
+		rxq->write_actual = (rxq->write & ~0x7);
+		iwl_write32(trans, FH_RSCSR_CHNL0_WPTR, rxq->write_actual);
 	} else {
 		struct iwl_trans_pcie *trans_pcie =
 			IWL_TRANS_GET_PCIE_TRANS(trans);
@@ -175,22 +177,22 @@ static void iwl_pcie_rxq_inc_wr_ptr(struct iwl_trans *trans, struct iwl_rxq *q)
 				goto exit_unlock;
 			}
 
-			q->write_actual = (q->write & ~0x7);
+			rxq->write_actual = (rxq->write & ~0x7);
 			iwl_write_direct32(trans, FH_RSCSR_CHNL0_WPTR,
-					q->write_actual);
+					   rxq->write_actual);
 
 		/* Else device is assumed to be awake */
 		} else {
 			/* Device expects a multiple of 8 */
-			q->write_actual = (q->write & ~0x7);
+			rxq->write_actual = (rxq->write & ~0x7);
 			iwl_write_direct32(trans, FH_RSCSR_CHNL0_WPTR,
-				q->write_actual);
+					   rxq->write_actual);
 		}
 	}
-	q->need_update = 0;
+	rxq->need_update = 0;
 
  exit_unlock:
-	spin_unlock_irqrestore(&q->lock, flags);
+	spin_unlock_irqrestore(&rxq->lock, flags);
 }
 
 /*
@@ -355,19 +357,16 @@ static void iwl_pcie_rxq_free_rbs(struct iwl_trans *trans)
 	struct iwl_rxq *rxq = &trans_pcie->rxq;
 	int i;
 
-	/* Fill the rx_used queue with _all_ of the Rx buffers */
+	lockdep_assert_held(&rxq->lock);
+
 	for (i = 0; i < RX_FREE_BUFFERS + RX_QUEUE_SIZE; i++) {
-		/* In the reset function, these buffers may have been allocated
-		 * to an SKB, so we need to unmap and free potential storage */
-		if (rxq->pool[i].page != NULL) {
-			dma_unmap_page(trans->dev, rxq->pool[i].page_dma,
-				       PAGE_SIZE << trans_pcie->rx_page_order,
-				       DMA_FROM_DEVICE);
-			__free_pages(rxq->pool[i].page,
-				     trans_pcie->rx_page_order);
-			rxq->pool[i].page = NULL;
-		}
-		list_add_tail(&rxq->pool[i].list, &rxq->rx_used);
+		if (!rxq->pool[i].page)
+			continue;
+		dma_unmap_page(trans->dev, rxq->pool[i].page_dma,
+			       PAGE_SIZE << trans_pcie->rx_page_order,
+			       DMA_FROM_DEVICE);
+		__free_pages(rxq->pool[i].page, trans_pcie->rx_page_order);
+		rxq->pool[i].page = NULL;
 	}
 }
 
@@ -491,6 +490,20 @@ static void iwl_pcie_rx_hw_init(struct iwl_trans *trans, struct iwl_rxq *rxq)
 	iwl_write8(trans, CSR_INT_COALESCING, IWL_HOST_INT_TIMEOUT_DEF);
 }
 
+static void iwl_pcie_rx_init_rxb_lists(struct iwl_rxq *rxq)
+{
+	int i;
+
+	lockdep_assert_held(&rxq->lock);
+
+	INIT_LIST_HEAD(&rxq->rx_free);
+	INIT_LIST_HEAD(&rxq->rx_used);
+	rxq->free_count = 0;
+
+	for (i = 0; i < RX_FREE_BUFFERS + RX_QUEUE_SIZE; i++)
+		list_add(&rxq->pool[i].list, &rxq->rx_used);
+}
+
 int iwl_pcie_rx_init(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
@@ -505,13 +518,12 @@ int iwl_pcie_rx_init(struct iwl_trans *trans)
 	}
 
 	spin_lock_irqsave(&rxq->lock, flags);
-	INIT_LIST_HEAD(&rxq->rx_free);
-	INIT_LIST_HEAD(&rxq->rx_used);
 
-	INIT_WORK(&trans_pcie->rx_replenish,
-		  iwl_pcie_rx_replenish_work);
+	INIT_WORK(&trans_pcie->rx_replenish, iwl_pcie_rx_replenish_work);
 
+	/* free all first - we might be reconfigured for a different size */
 	iwl_pcie_rxq_free_rbs(trans);
+	iwl_pcie_rx_init_rxb_lists(rxq);
 
 	for (i = 0; i < RX_QUEUE_SIZE; i++)
 		rxq->queue[i] = NULL;
@@ -520,7 +532,6 @@ int iwl_pcie_rx_init(struct iwl_trans *trans)
 	 * not restocked the Rx queue with fresh buffers */
 	rxq->read = rxq->write = 0;
 	rxq->write_actual = 0;
-	rxq->free_count = 0;
 	memset(rxq->rb_stts, 0, sizeof(*rxq->rb_stts));
 	spin_unlock_irqrestore(&rxq->lock, flags);
 
@@ -802,9 +813,6 @@ irqreturn_t iwl_pcie_irq_handler(int irq, void *dev_id)
 	u32 handled = 0;
 	unsigned long flags;
 	u32 i;
-#ifdef CONFIG_IWLWIFI_DEBUG
-	u32 inta_mask;
-#endif
 
 	lock_map_acquire(&trans->sync_cmd_lockdep_map);
 
@@ -826,14 +834,9 @@ irqreturn_t iwl_pcie_irq_handler(int irq, void *dev_id)
 
 	inta = trans_pcie->inta;
 
-#ifdef CONFIG_IWLWIFI_DEBUG
-	if (iwl_have_debug_level(IWL_DL_ISR)) {
-		/* just for debug */
-		inta_mask = iwl_read32(trans, CSR_INT_MASK);
+	if (iwl_have_debug_level(IWL_DL_ISR))
 		IWL_DEBUG_ISR(trans, "inta 0x%08x, enabled 0x%08x\n",
-			      inta, inta_mask);
-	}
-#endif
+			      inta, iwl_read32(trans, CSR_INT_MASK));
 
 	/* saved interrupt in inta variable now we can reset trans_pcie->inta */
 	trans_pcie->inta = 0;
@@ -855,12 +858,11 @@ irqreturn_t iwl_pcie_irq_handler(int irq, void *dev_id)
 		goto out;
 	}
 
-#ifdef CONFIG_IWLWIFI_DEBUG
 	if (iwl_have_debug_level(IWL_DL_ISR)) {
 		/* NIC fires this, but we don't use it, redundant with WAKEUP */
 		if (inta & CSR_INT_BIT_SCD) {
-			IWL_DEBUG_ISR(trans, "Scheduler finished to transmit "
-				      "the frame/frames.\n");
+			IWL_DEBUG_ISR(trans,
+				      "Scheduler finished to transmit the frame/frames.\n");
 			isr_stats->sch++;
 		}
 
@@ -870,7 +872,7 @@ irqreturn_t iwl_pcie_irq_handler(int irq, void *dev_id)
 			isr_stats->alive++;
 		}
 	}
-#endif
+
 	/* Safely ignore these bits for debug checks below */
 	inta &= ~(CSR_INT_BIT_SCD | CSR_INT_BIT_ALIVE);
 
@@ -1118,9 +1120,6 @@ static irqreturn_t iwl_pcie_isr(int irq, void *data)
 	struct iwl_trans *trans = data;
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 	u32 inta, inta_mask;
-#ifdef CONFIG_IWLWIFI_DEBUG
-	u32 inta_fh;
-#endif
 
 	lockdep_assert_held(&trans_pcie->irq_lock);
 
@@ -1159,13 +1158,11 @@ static irqreturn_t iwl_pcie_isr(int irq, void *data)
 		return IRQ_HANDLED;
 	}
 
-#ifdef CONFIG_IWLWIFI_DEBUG
-	if (iwl_have_debug_level(IWL_DL_ISR)) {
-		inta_fh = iwl_read32(trans, CSR_FH_INT_STATUS);
-		IWL_DEBUG_ISR(trans, "ISR inta 0x%08x, enabled 0x%08x, "
-			      "fh 0x%08x\n", inta, inta_mask, inta_fh);
-	}
-#endif
+	if (iwl_have_debug_level(IWL_DL_ISR))
+		IWL_DEBUG_ISR(trans,
+			      "ISR inta 0x%08x, enabled 0x%08x, fh 0x%08x\n",
+			      inta, inta_mask,
+			      iwl_read32(trans, CSR_FH_INT_STATUS));
 
 	trans_pcie->inta |= inta;
 	/* the thread will service interrupts and re-enable them */
@@ -1198,7 +1195,7 @@ irqreturn_t iwl_pcie_isr_ict(int irq, void *data)
 {
 	struct iwl_trans *trans = data;
 	struct iwl_trans_pcie *trans_pcie;
-	u32 inta, inta_mask;
+	u32 inta;
 	u32 val = 0;
 	u32 read;
 	unsigned long flags;
@@ -1226,7 +1223,6 @@ irqreturn_t iwl_pcie_isr_ict(int irq, void *data)
 	 * If we have something to service, the tasklet will re-enable ints.
 	 * If we *don't* have something, we'll re-enable before leaving here.
 	 */
-	inta_mask = iwl_read32(trans, CSR_INT_MASK);
 	iwl_write32(trans, CSR_INT_MASK, 0x00000000);
 
 	/* Ignore interrupt if there's nothing in NIC to service.
@@ -1271,8 +1267,11 @@ irqreturn_t iwl_pcie_isr_ict(int irq, void *data)
 		val |= 0x8000;
 
 	inta = (0xff & val) | ((0xff00 & val) << 16);
-	IWL_DEBUG_ISR(trans, "ISR inta 0x%08x, enabled 0x%08x ict 0x%08x\n",
-		      inta, inta_mask, val);
+	IWL_DEBUG_ISR(trans, "ISR inta 0x%08x, enabled(sw) 0x%08x ict 0x%08x\n",
+		      inta, trans_pcie->inta_mask, val);
+	if (iwl_have_debug_level(IWL_DL_ISR))
+		IWL_DEBUG_ISR(trans, "enabled(hw) 0x%08x\n",
+			      iwl_read32(trans, CSR_INT_MASK));
 
 	inta &= trans_pcie->inta_mask;
 	trans_pcie->inta |= inta;
diff --git a/drivers/net/wireless/iwlwifi/pcie/trans.c b/drivers/net/wireless/iwlwifi/pcie/trans.c
index 50ba0a4..826c156 100644
--- a/drivers/net/wireless/iwlwifi/pcie/trans.c
+++ b/drivers/net/wireless/iwlwifi/pcie/trans.c
@@ -405,20 +405,27 @@ static int iwl_pcie_load_section(struct iwl_trans *trans, u8 section_num,
 {
 	u8 *v_addr;
 	dma_addr_t p_addr;
-	u32 offset;
+	u32 offset, chunk_sz = section->len;
 	int ret = 0;
 
 	IWL_DEBUG_FW(trans, "[%d] uCode section being loaded...\n",
 		     section_num);
 
-	v_addr = dma_alloc_coherent(trans->dev, PAGE_SIZE, &p_addr, GFP_KERNEL);
-	if (!v_addr)
-		return -ENOMEM;
+	v_addr = dma_alloc_coherent(trans->dev, chunk_sz, &p_addr,
+				    GFP_KERNEL | __GFP_NOWARN);
+	if (!v_addr) {
+		IWL_DEBUG_INFO(trans, "Falling back to small chunks of DMA\n");
+		chunk_sz = PAGE_SIZE;
+		v_addr = dma_alloc_coherent(trans->dev, chunk_sz,
+					    &p_addr, GFP_KERNEL);
+		if (!v_addr)
+			return -ENOMEM;
+	}
 
-	for (offset = 0; offset < section->len; offset += PAGE_SIZE) {
+	for (offset = 0; offset < section->len; offset += chunk_sz) {
 		u32 copy_size;
 
-		copy_size = min_t(u32, PAGE_SIZE, section->len - offset);
+		copy_size = min_t(u32, chunk_sz, section->len - offset);
 
 		memcpy(v_addr, (u8 *)section->data + offset, copy_size);
 		ret = iwl_pcie_load_firmware_chunk(trans,
@@ -432,7 +439,7 @@ static int iwl_pcie_load_section(struct iwl_trans *trans, u8 section_num,
 		}
 	}
 
-	dma_free_coherent(trans->dev, PAGE_SIZE, v_addr, p_addr);
+	dma_free_coherent(trans->dev, chunk_sz, v_addr, p_addr);
 	return ret;
 }
 
@@ -571,13 +578,17 @@ static void iwl_trans_pcie_stop_device(struct iwl_trans *trans)
 	clear_bit(STATUS_RFKILL, &trans_pcie->status);
 }
 
-static void iwl_trans_pcie_d3_suspend(struct iwl_trans *trans)
+static void iwl_trans_pcie_d3_suspend(struct iwl_trans *trans, bool test)
 {
-	/* let the ucode operate on its own */
-	iwl_write32(trans, CSR_UCODE_DRV_GP1_SET,
-		    CSR_UCODE_DRV_GP1_BIT_D3_CFG_COMPLETE);
-
 	iwl_disable_interrupts(trans);
+
+	/*
+	 * in testing mode, the host stays awake and the
+	 * hardware won't be reset (not even partially)
+	 */
+	if (test)
+		return;
+
 	iwl_pcie_disable_ict(trans);
 
 	iwl_clear_bit(trans, CSR_GP_CNTRL,
@@ -596,11 +607,18 @@ static void iwl_trans_pcie_d3_suspend(struct iwl_trans *trans)
 }
 
 static int iwl_trans_pcie_d3_resume(struct iwl_trans *trans,
-				    enum iwl_d3_status *status)
+				    enum iwl_d3_status *status,
+				    bool test)
 {
 	u32 val;
 	int ret;
 
+	if (test) {
+		iwl_enable_interrupts(trans);
+		*status = IWL_D3_STATUS_ALIVE;
+		return 0;
+	}
+
 	iwl_pcie_set_pwr(trans, false);
 
 	val = iwl_read32(trans, CSR_RESET);
@@ -636,9 +654,6 @@ static int iwl_trans_pcie_d3_resume(struct iwl_trans *trans,
 		return ret;
 	}
 
-	iwl_write32(trans, CSR_UCODE_DRV_GP1_CLR,
-		    CSR_UCODE_DRV_GP1_BIT_D3_CFG_COMPLETE);
-
 	*status = IWL_D3_STATUS_ALIVE;
 	return 0;
 }
@@ -823,8 +838,9 @@ static bool iwl_trans_pcie_grab_nic_access(struct iwl_trans *trans, bool silent,
 						unsigned long *flags)
 {
 	int ret;
-	struct iwl_trans_pcie *pcie_trans = IWL_TRANS_GET_PCIE_TRANS(trans);
-	spin_lock_irqsave(&pcie_trans->reg_lock, *flags);
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+
+	spin_lock_irqsave(&trans_pcie->reg_lock, *flags);
 
 	/* this bit wakes up the NIC */
 	__iwl_trans_pcie_set_bit(trans, CSR_GP_CNTRL,
@@ -860,7 +876,7 @@ static bool iwl_trans_pcie_grab_nic_access(struct iwl_trans *trans, bool silent,
 			WARN_ONCE(1,
 				  "Timeout waiting for hardware access (CSR_GP_CNTRL 0x%08x)\n",
 				  val);
-			spin_unlock_irqrestore(&pcie_trans->reg_lock, *flags);
+			spin_unlock_irqrestore(&trans_pcie->reg_lock, *flags);
 			return false;
 		}
 	}
@@ -869,22 +885,22 @@ static bool iwl_trans_pcie_grab_nic_access(struct iwl_trans *trans, bool silent,
 	 * Fool sparse by faking we release the lock - sparse will
 	 * track nic_access anyway.
 	 */
-	__release(&pcie_trans->reg_lock);
+	__release(&trans_pcie->reg_lock);
 	return true;
 }
 
 static void iwl_trans_pcie_release_nic_access(struct iwl_trans *trans,
 					      unsigned long *flags)
 {
-	struct iwl_trans_pcie *pcie_trans = IWL_TRANS_GET_PCIE_TRANS(trans);
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 
-	lockdep_assert_held(&pcie_trans->reg_lock);
+	lockdep_assert_held(&trans_pcie->reg_lock);
 
 	/*
 	 * Fool sparse by faking we acquiring the lock - sparse will
 	 * track nic_access anyway.
 	 */
-	__acquire(&pcie_trans->reg_lock);
+	__acquire(&trans_pcie->reg_lock);
 
 	__iwl_trans_pcie_clear_bit(trans, CSR_GP_CNTRL,
 				   CSR_GP_CNTRL_REG_FLAG_MAC_ACCESS_REQ);
@@ -895,7 +911,7 @@ static void iwl_trans_pcie_release_nic_access(struct iwl_trans *trans,
 	 * scheduled on different CPUs (after we drop reg_lock).
 	 */
 	mmiowb();
-	spin_unlock_irqrestore(&pcie_trans->reg_lock, *flags);
+	spin_unlock_irqrestore(&trans_pcie->reg_lock, *flags);
 }
 
 static int iwl_trans_pcie_read_mem(struct iwl_trans *trans, u32 addr,
@@ -917,11 +933,11 @@ static int iwl_trans_pcie_read_mem(struct iwl_trans *trans, u32 addr,
 }
 
 static int iwl_trans_pcie_write_mem(struct iwl_trans *trans, u32 addr,
-				    void *buf, int dwords)
+				    const void *buf, int dwords)
 {
 	unsigned long flags;
 	int offs, ret = 0;
-	u32 *vals = buf;
+	const u32 *vals = buf;
 
 	if (iwl_trans_grab_nic_access(trans, false, &flags)) {
 		iwl_write32(trans, HBUS_TARG_MEM_WADDR, addr);
diff --git a/drivers/net/wireless/iwlwifi/pcie/tx.c b/drivers/net/wireless/iwlwifi/pcie/tx.c
index c5e3029..c47c921 100644
--- a/drivers/net/wireless/iwlwifi/pcie/tx.c
+++ b/drivers/net/wireless/iwlwifi/pcie/tx.c
@@ -224,13 +224,13 @@ static void iwl_pcie_txq_update_byte_cnt_tbl(struct iwl_trans *trans,
 
 	switch (sec_ctl & TX_CMD_SEC_MSK) {
 	case TX_CMD_SEC_CCM:
-		len += CCMP_MIC_LEN;
+		len += IEEE80211_CCMP_MIC_LEN;
 		break;
 	case TX_CMD_SEC_TKIP:
-		len += TKIP_ICV_LEN;
+		len += IEEE80211_TKIP_ICV_LEN;
 		break;
 	case TX_CMD_SEC_WEP:
-		len += WEP_IV_LEN + WEP_ICV_LEN;
+		len += IEEE80211_WEP_IV_LEN + IEEE80211_WEP_ICV_LEN;
 		break;
 	}
 
@@ -576,10 +576,16 @@ static void iwl_pcie_txq_unmap(struct iwl_trans *trans, int txq_id)
 
 	spin_lock_bh(&txq->lock);
 	while (q->write_ptr != q->read_ptr) {
+		IWL_DEBUG_TX_REPLY(trans, "Q %d Free %d\n",
+				   txq_id, q->read_ptr);
 		iwl_pcie_txq_free_tfd(trans, txq);
 		q->read_ptr = iwl_queue_inc_wrap(q->read_ptr, q->n_bd);
 	}
+	txq->active = false;
 	spin_unlock_bh(&txq->lock);
+
+	/* just in case - this queue may have been stopped */
+	iwl_wake_queue(trans, txq);
 }
 
 /*
@@ -927,6 +933,12 @@ void iwl_trans_pcie_reclaim(struct iwl_trans *trans, int txq_id, int ssn,
 
 	spin_lock_bh(&txq->lock);
 
+	if (!txq->active) {
+		IWL_DEBUG_TX_QUEUES(trans, "Q %d inactive - ignoring idx %d\n",
+				    txq_id, ssn);
+		goto out;
+	}
+
 	if (txq->q.read_ptr == tfd_num)
 		goto out;
 
@@ -1045,6 +1057,10 @@ static inline void iwl_pcie_txq_set_inactive(struct iwl_trans *trans,
 		(1 << SCD_QUEUE_STTS_REG_POS_SCD_ACT_EN));
 }
 
+/* Receiver address (actually, Rx station's index into station table),
+ * combined with Traffic ID (QOS priority), in format used by Tx Scheduler */
+#define BUILD_RAxTID(sta_id, tid)	(((sta_id) << 4) + (tid))
+
 void iwl_trans_pcie_txq_enable(struct iwl_trans *trans, int txq_id, int fifo,
 			       int sta_id, int tid, int frame_limit, u16 ssn)
 {
@@ -1069,6 +1085,7 @@ void iwl_trans_pcie_txq_enable(struct iwl_trans *trans, int txq_id, int fifo,
 
 		/* enable aggregations for the queue */
 		iwl_set_bits_prph(trans, SCD_AGGR_SEL, BIT(txq_id));
+		trans_pcie->txq[txq_id].ampdu = true;
 	} else {
 		/*
 		 * disable aggregations for the queue, this will also make the
@@ -1103,6 +1120,7 @@ void iwl_trans_pcie_txq_enable(struct iwl_trans *trans, int txq_id, int fifo,
 		       (fifo << SCD_QUEUE_STTS_REG_POS_TXF) |
 		       (1 << SCD_QUEUE_STTS_REG_POS_WSL) |
 		       SCD_QUEUE_STTS_REG_MSK);
+	trans_pcie->txq[txq_id].active = true;
 	IWL_DEBUG_TX_QUEUES(trans, "Activate queue %d on FIFO %d WrPtr: %d\n",
 			    txq_id, fifo, ssn & 0xff);
 }
@@ -1125,6 +1143,7 @@ void iwl_trans_pcie_txq_disable(struct iwl_trans *trans, int txq_id)
 			    ARRAY_SIZE(zero_val));
 
 	iwl_pcie_txq_unmap(trans, txq_id);
+	trans_pcie->txq[txq_id].ampdu = false;
 
 	IWL_DEBUG_TX_QUEUES(trans, "Deactivate queue %d\n", txq_id);
 }
@@ -1518,11 +1537,13 @@ static int iwl_pcie_send_hcmd_sync(struct iwl_trans *trans,
 	if (test_bit(STATUS_FW_ERROR, &trans_pcie->status)) {
 		IWL_ERR(trans, "FW error in SYNC CMD %s\n",
 			get_cmd_string(trans_pcie, cmd->id));
+		dump_stack();
 		ret = -EIO;
 		goto cancel;
 	}
 
-	if (test_bit(STATUS_RFKILL, &trans_pcie->status)) {
+	if (!(cmd->flags & CMD_SEND_IN_RFKILL) &&
+	    test_bit(STATUS_RFKILL, &trans_pcie->status)) {
 		IWL_DEBUG_RF_KILL(trans, "RFKILL in SYNC CMD... no rsp\n");
 		ret = -ERFKILL;
 		goto cancel;
@@ -1564,7 +1585,8 @@ int iwl_trans_pcie_send_hcmd(struct iwl_trans *trans, struct iwl_host_cmd *cmd)
 	if (test_bit(STATUS_FW_ERROR, &trans_pcie->status))
 		return -EIO;
 
-	if (test_bit(STATUS_RFKILL, &trans_pcie->status)) {
+	if (!(cmd->flags & CMD_SEND_IN_RFKILL) &&
+	    test_bit(STATUS_RFKILL, &trans_pcie->status)) {
 		IWL_DEBUG_RF_KILL(trans, "Dropping CMD 0x%x: RF KILL\n",
 				  cmd->id);
 		return -ERFKILL;
@@ -1592,7 +1614,7 @@ int iwl_trans_pcie_tx(struct iwl_trans *trans, struct sk_buff *skb,
 	u8 wait_write_ptr = 0;
 	__le16 fc = hdr->frame_control;
 	u8 hdr_len = ieee80211_hdrlen(fc);
-	u16 __maybe_unused wifi_seq;
+	u16 wifi_seq;
 
 	txq = &trans_pcie->txq[txq_id];
 	q = &txq->q;
@@ -1609,13 +1631,11 @@ int iwl_trans_pcie_tx(struct iwl_trans *trans, struct sk_buff *skb,
 	 * the BA.
 	 * Check here that the packets are in the right place on the ring.
 	 */
-#ifdef CONFIG_IWLWIFI_DEBUG
 	wifi_seq = IEEE80211_SEQ_TO_SN(le16_to_cpu(hdr->seq_ctrl));
-	WARN_ONCE((iwl_read_prph(trans, SCD_AGGR_SEL) & BIT(txq_id)) &&
-		  ((wifi_seq & 0xff) != q->write_ptr),
+	WARN_ONCE(trans_pcie->txq[txq_id].ampdu &&
+		  (wifi_seq & 0xff) != q->write_ptr,
 		  "Q: %d WiFi Seq %d tfdNum %d",
 		  txq_id, wifi_seq, q->write_ptr);
-#endif
 
 	/* Set up driver data for this TFD */
 	txq->entries[q->write_ptr].skb = skb;
diff --git a/drivers/net/wireless/libertas/mesh.c b/drivers/net/wireless/libertas/mesh.c
index 3e81264..efae07e 100644
--- a/drivers/net/wireless/libertas/mesh.c
+++ b/drivers/net/wireless/libertas/mesh.c
@@ -240,7 +240,7 @@ static ssize_t lbs_prb_rsp_limit_set(struct device *dev,
 	memset(&mesh_access, 0, sizeof(mesh_access));
 	mesh_access.data[0] = cpu_to_le32(CMD_ACT_SET);
 
-	if (!strict_strtoul(buf, 10, &retry_limit))
+	if (!kstrtoul(buf, 10, &retry_limit))
 		return -ENOTSUPP;
 	if (retry_limit > 15)
 		return -ENOTSUPP;
diff --git a/drivers/net/wireless/mwifiex/11h.c b/drivers/net/wireless/mwifiex/11h.c
new file mode 100644
index 0000000..8d68307
--- /dev/null
+++ b/drivers/net/wireless/mwifiex/11h.c
@@ -0,0 +1,101 @@
+/*
+ * Marvell Wireless LAN device driver: 802.11h
+ *
+ * Copyright (C) 2013, Marvell International Ltd.
+ *
+ * This software file (the "File") is distributed by Marvell International
+ * Ltd. under the terms of the GNU General Public License Version 2, June 1991
+ * (the "License").  You may use, redistribute and/or modify this File in
+ * accordance with the terms and conditions of the License, a copy of which
+ * is available by writing to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA or on the
+ * worldwide web at http://www.gnu.org/licenses/old-licenses/gpl-2.0.txt.
+ *
+ * THE FILE IS DISTRIBUTED AS-IS, WITHOUT WARRANTY OF ANY KIND, AND THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE
+ * ARE EXPRESSLY DISCLAIMED.  The License provides additional details about
+ * this warranty disclaimer.
+ */
+
+#include "main.h"
+#include "fw.h"
+
+
+/* This function appends 11h info to a buffer while joining an
+ * infrastructure BSS
+ */
+static void
+mwifiex_11h_process_infra_join(struct mwifiex_private *priv, u8 **buffer,
+			       struct mwifiex_bssdescriptor *bss_desc)
+{
+	struct mwifiex_ie_types_header *ie_header;
+	struct mwifiex_ie_types_pwr_capability *cap;
+	struct mwifiex_ie_types_local_pwr_constraint *constraint;
+	struct ieee80211_supported_band *sband;
+	u8 radio_type;
+	int i;
+
+	if (!buffer || !(*buffer))
+		return;
+
+	radio_type = mwifiex_band_to_radio_type((u8) bss_desc->bss_band);
+	sband = priv->wdev->wiphy->bands[radio_type];
+
+	cap = (struct mwifiex_ie_types_pwr_capability *)*buffer;
+	cap->header.type = cpu_to_le16(WLAN_EID_PWR_CAPABILITY);
+	cap->header.len = cpu_to_le16(2);
+	cap->min_pwr = 0;
+	cap->max_pwr = 0;
+	*buffer += sizeof(*cap);
+
+	constraint = (struct mwifiex_ie_types_local_pwr_constraint *)*buffer;
+	constraint->header.type = cpu_to_le16(WLAN_EID_PWR_CONSTRAINT);
+	constraint->header.len = cpu_to_le16(2);
+	constraint->chan = bss_desc->channel;
+	constraint->constraint = bss_desc->local_constraint;
+	*buffer += sizeof(*constraint);
+
+	ie_header = (struct mwifiex_ie_types_header *)*buffer;
+	ie_header->type = cpu_to_le16(TLV_TYPE_PASSTHROUGH);
+	ie_header->len  = cpu_to_le16(2 * sband->n_channels + 2);
+	*buffer += sizeof(*ie_header);
+	*(*buffer)++ = WLAN_EID_SUPPORTED_CHANNELS;
+	*(*buffer)++ = 2 * sband->n_channels;
+	for (i = 0; i < sband->n_channels; i++) {
+		*(*buffer)++ = ieee80211_frequency_to_channel(
+					sband->channels[i].center_freq);
+		*(*buffer)++ = 1; /* one channel in the subband */
+	}
+}
+
+/* Enable or disable the 11h extensions in the firmware */
+static int mwifiex_11h_activate(struct mwifiex_private *priv, bool flag)
+{
+	u32 enable = flag;
+
+	return mwifiex_send_cmd_sync(priv, HostCmd_CMD_802_11_SNMP_MIB,
+				     HostCmd_ACT_GEN_SET, DOT11H_I, &enable);
+}
+
+/* This functions processes TLV buffer for a pending BSS Join command.
+ *
+ * Activate 11h functionality in the firmware if the spectrum management
+ * capability bit is found in the network we are joining. Also, necessary
+ * TLVs are set based on requested network's 11h capability.
+ */
+void mwifiex_11h_process_join(struct mwifiex_private *priv, u8 **buffer,
+			      struct mwifiex_bssdescriptor *bss_desc)
+{
+	if (bss_desc->sensed_11h) {
+		/* Activate 11h functions in firmware, turns on capability
+		 * bit
+		 */
+		mwifiex_11h_activate(priv, true);
+		bss_desc->cap_info_bitmap |= WLAN_CAPABILITY_SPECTRUM_MGMT;
+		mwifiex_11h_process_infra_join(priv, buffer, bss_desc);
+	} else {
+		/* Deactivate 11h functions in the firmware */
+		mwifiex_11h_activate(priv, false);
+		bss_desc->cap_info_bitmap &= ~WLAN_CAPABILITY_SPECTRUM_MGMT;
+	}
+}
diff --git a/drivers/net/wireless/mwifiex/Kconfig b/drivers/net/wireless/mwifiex/Kconfig
index 4f614aa..f7ff472 100644
--- a/drivers/net/wireless/mwifiex/Kconfig
+++ b/drivers/net/wireless/mwifiex/Kconfig
@@ -3,13 +3,13 @@ config MWIFIEX
 	depends on CFG80211
 	---help---
 	  This adds support for wireless adapters based on Marvell
-	  802.11n chipsets.
+	  802.11n/ac chipsets.
 
 	  If you choose to build it as a module, it will be called
 	  mwifiex.
 
 config MWIFIEX_SDIO
-	tristate "Marvell WiFi-Ex Driver for SD8786/SD8787/SD8797"
+	tristate "Marvell WiFi-Ex Driver for SD8786/SD8787/SD8797/SD8897"
 	depends on MWIFIEX && MMC
 	select FW_LOADER
 	---help---
diff --git a/drivers/net/wireless/mwifiex/Makefile b/drivers/net/wireless/mwifiex/Makefile
index ecf2846..a42a506 100644
--- a/drivers/net/wireless/mwifiex/Makefile
+++ b/drivers/net/wireless/mwifiex/Makefile
@@ -40,6 +40,7 @@ mwifiex-y += sta_rx.o
 mwifiex-y += uap_txrx.o
 mwifiex-y += cfg80211.o
 mwifiex-y += ethtool.o
+mwifiex-y += 11h.o
 mwifiex-$(CONFIG_DEBUG_FS) += debugfs.o
 obj-$(CONFIG_MWIFIEX) += mwifiex.o
 
diff --git a/drivers/net/wireless/mwifiex/cfg80211.c b/drivers/net/wireless/mwifiex/cfg80211.c
index e42b266..ef5fa89 100644
--- a/drivers/net/wireless/mwifiex/cfg80211.c
+++ b/drivers/net/wireless/mwifiex/cfg80211.c
@@ -20,6 +20,9 @@
 #include "cfg80211.h"
 #include "main.h"
 
+static char *reg_alpha2;
+module_param(reg_alpha2, charp, 0);
+
 static const struct ieee80211_iface_limit mwifiex_ap_sta_limits[] = {
 	{
 		.max = 2, .types = BIT(NL80211_IFTYPE_STATION),
@@ -1231,6 +1234,51 @@ static int mwifiex_cfg80211_change_beacon(struct wiphy *wiphy,
 	return 0;
 }
 
+/* cfg80211 operation handler for del_station.
+ * Function deauthenticates station which value is provided in mac parameter.
+ * If mac is NULL/broadcast, all stations in associated station list are
+ * deauthenticated. If bss is not started or there are no stations in
+ * associated stations list, no action is taken.
+ */
+static int
+mwifiex_cfg80211_del_station(struct wiphy *wiphy, struct net_device *dev,
+			     u8 *mac)
+{
+	struct mwifiex_private *priv = mwifiex_netdev_get_priv(dev);
+	struct mwifiex_sta_node *sta_node;
+	unsigned long flags;
+
+	if (list_empty(&priv->sta_list) || !priv->bss_started)
+		return 0;
+
+	if (!mac || is_broadcast_ether_addr(mac)) {
+		wiphy_dbg(wiphy, "%s: NULL/broadcast mac address\n", __func__);
+		list_for_each_entry(sta_node, &priv->sta_list, list) {
+			if (mwifiex_send_cmd_sync(priv,
+						  HostCmd_CMD_UAP_STA_DEAUTH,
+						  HostCmd_ACT_GEN_SET, 0,
+						  sta_node->mac_addr))
+				return -1;
+			mwifiex_uap_del_sta_data(priv, sta_node);
+		}
+	} else {
+		wiphy_dbg(wiphy, "%s: mac address %pM\n", __func__, mac);
+		spin_lock_irqsave(&priv->sta_list_spinlock, flags);
+		sta_node = mwifiex_get_sta_entry(priv, mac);
+		spin_unlock_irqrestore(&priv->sta_list_spinlock, flags);
+		if (sta_node) {
+			if (mwifiex_send_cmd_sync(priv,
+						  HostCmd_CMD_UAP_STA_DEAUTH,
+						  HostCmd_ACT_GEN_SET, 0,
+						  sta_node->mac_addr))
+				return -1;
+			mwifiex_uap_del_sta_data(priv, sta_node);
+		}
+	}
+
+	return 0;
+}
+
 static int
 mwifiex_cfg80211_set_antenna(struct wiphy *wiphy, u32 tx_ant, u32 rx_ant)
 {
@@ -1859,6 +1907,7 @@ mwifiex_cfg80211_scan(struct wiphy *wiphy,
 	int i, offset, ret;
 	struct ieee80211_channel *chan;
 	struct ieee_types_header *ie;
+	struct mwifiex_user_scan_cfg *user_scan_cfg;
 
 	wiphy_dbg(wiphy, "info: received scan request on %s\n", dev->name);
 
@@ -1869,20 +1918,22 @@ mwifiex_cfg80211_scan(struct wiphy *wiphy,
 		return -EBUSY;
 	}
 
-	if (priv->user_scan_cfg) {
+	/* Block scan request if scan operation or scan cleanup when interface
+	 * is disabled is in process
+	 */
+	if (priv->scan_request || priv->scan_aborting) {
 		dev_err(priv->adapter->dev, "cmd: Scan already in process..\n");
 		return -EBUSY;
 	}
 
-	priv->user_scan_cfg = kzalloc(sizeof(struct mwifiex_user_scan_cfg),
-				      GFP_KERNEL);
-	if (!priv->user_scan_cfg)
+	user_scan_cfg = kzalloc(sizeof(*user_scan_cfg), GFP_KERNEL);
+	if (!user_scan_cfg)
 		return -ENOMEM;
 
 	priv->scan_request = request;
 
-	priv->user_scan_cfg->num_ssids = request->n_ssids;
-	priv->user_scan_cfg->ssid_list = request->ssids;
+	user_scan_cfg->num_ssids = request->n_ssids;
+	user_scan_cfg->ssid_list = request->ssids;
 
 	if (request->ie && request->ie_len) {
 		offset = 0;
@@ -1902,25 +1953,25 @@ mwifiex_cfg80211_scan(struct wiphy *wiphy,
 	for (i = 0; i < min_t(u32, request->n_channels,
 			      MWIFIEX_USER_SCAN_CHAN_MAX); i++) {
 		chan = request->channels[i];
-		priv->user_scan_cfg->chan_list[i].chan_number = chan->hw_value;
-		priv->user_scan_cfg->chan_list[i].radio_type = chan->band;
+		user_scan_cfg->chan_list[i].chan_number = chan->hw_value;
+		user_scan_cfg->chan_list[i].radio_type = chan->band;
 
 		if (chan->flags & IEEE80211_CHAN_PASSIVE_SCAN)
-			priv->user_scan_cfg->chan_list[i].scan_type =
+			user_scan_cfg->chan_list[i].scan_type =
 						MWIFIEX_SCAN_TYPE_PASSIVE;
 		else
-			priv->user_scan_cfg->chan_list[i].scan_type =
+			user_scan_cfg->chan_list[i].scan_type =
 						MWIFIEX_SCAN_TYPE_ACTIVE;
 
-		priv->user_scan_cfg->chan_list[i].scan_time = 0;
+		user_scan_cfg->chan_list[i].scan_time = 0;
 	}
 
-	ret = mwifiex_scan_networks(priv, priv->user_scan_cfg);
+	ret = mwifiex_scan_networks(priv, user_scan_cfg);
+	kfree(user_scan_cfg);
 	if (ret) {
 		dev_err(priv->adapter->dev, "scan failed: %d\n", ret);
+		priv->scan_aborting = false;
 		priv->scan_request = NULL;
-		kfree(priv->user_scan_cfg);
-		priv->user_scan_cfg = NULL;
 		return ret;
 	}
 
@@ -2419,6 +2470,7 @@ static struct cfg80211_ops mwifiex_cfg80211_ops = {
 	.change_beacon = mwifiex_cfg80211_change_beacon,
 	.set_cqm_rssi_config = mwifiex_cfg80211_set_cqm_rssi_config,
 	.set_antenna = mwifiex_cfg80211_set_antenna,
+	.del_station = mwifiex_cfg80211_del_station,
 #ifdef CONFIG_PM
 	.suspend = mwifiex_cfg80211_suspend,
 	.resume = mwifiex_cfg80211_resume,
@@ -2426,6 +2478,27 @@ static struct cfg80211_ops mwifiex_cfg80211_ops = {
 #endif
 };
 
+#ifdef CONFIG_PM
+static const struct wiphy_wowlan_support mwifiex_wowlan_support = {
+	.flags = WIPHY_WOWLAN_MAGIC_PKT,
+	.n_patterns = MWIFIEX_MAX_FILTERS,
+	.pattern_min_len = 1,
+	.pattern_max_len = MWIFIEX_MAX_PATTERN_LEN,
+	.max_pkt_offset = MWIFIEX_MAX_OFFSET_LEN,
+};
+#endif
+
+static bool mwifiex_is_valid_alpha2(const char *alpha2)
+{
+	if (!alpha2 || strlen(alpha2) != 2)
+		return false;
+
+	if (isalpha(alpha2[0]) && isalpha(alpha2[1]))
+		return true;
+
+	return false;
+}
+
 /*
  * This function registers the device with CFG802.11 subsystem.
  *
@@ -2478,16 +2551,13 @@ int mwifiex_register_cfg80211(struct mwifiex_adapter *adapter)
 			WIPHY_FLAG_AP_PROBE_RESP_OFFLOAD |
 			WIPHY_FLAG_AP_UAPSD |
 			WIPHY_FLAG_CUSTOM_REGULATORY |
+			WIPHY_FLAG_STRICT_REGULATORY |
 			WIPHY_FLAG_HAS_REMAIN_ON_CHANNEL;
 
 	wiphy_apply_custom_regulatory(wiphy, &mwifiex_world_regdom_custom);
 
 #ifdef CONFIG_PM
-	wiphy->wowlan.flags = WIPHY_WOWLAN_MAGIC_PKT;
-	wiphy->wowlan.n_patterns = MWIFIEX_MAX_FILTERS;
-	wiphy->wowlan.pattern_min_len = 1;
-	wiphy->wowlan.pattern_max_len = MWIFIEX_MAX_PATTERN_LEN;
-	wiphy->wowlan.max_pkt_offset = MWIFIEX_MAX_OFFSET_LEN;
+	wiphy->wowlan = &mwifiex_wowlan_support;
 #endif
 
 	wiphy->probe_resp_offload = NL80211_PROBE_RESP_OFFLOAD_SUPPORT_WPS |
@@ -2519,10 +2589,16 @@ int mwifiex_register_cfg80211(struct mwifiex_adapter *adapter)
 		wiphy_free(wiphy);
 		return ret;
 	}
-	country_code = mwifiex_11d_code_2_region(priv->adapter->region_code);
-	if (country_code)
-		dev_info(adapter->dev,
-			 "ignoring F/W country code %2.2s\n", country_code);
+
+	if (reg_alpha2 && mwifiex_is_valid_alpha2(reg_alpha2)) {
+		wiphy_info(wiphy, "driver hint alpha2: %2.2s\n", reg_alpha2);
+		regulatory_hint(wiphy, reg_alpha2);
+	} else {
+		country_code = mwifiex_11d_code_2_region(adapter->region_code);
+		if (country_code)
+			wiphy_info(wiphy, "ignoring F/W country code %2.2s\n",
+				   country_code);
+	}
 
 	adapter->wiphy = wiphy;
 	return ret;
diff --git a/drivers/net/wireless/mwifiex/cmdevt.c b/drivers/net/wireless/mwifiex/cmdevt.c
index 26755d9..2d76147 100644
--- a/drivers/net/wireless/mwifiex/cmdevt.c
+++ b/drivers/net/wireless/mwifiex/cmdevt.c
@@ -570,6 +570,7 @@ int mwifiex_send_cmd_async(struct mwifiex_private *priv, uint16_t cmd_no,
 		case HostCmd_CMD_UAP_SYS_CONFIG:
 		case HostCmd_CMD_UAP_BSS_START:
 		case HostCmd_CMD_UAP_BSS_STOP:
+		case HostCmd_CMD_UAP_STA_DEAUTH:
 			ret = mwifiex_uap_prepare_cmd(priv, cmd_no, cmd_action,
 						      cmd_oid, data_buf,
 						      cmd_ptr);
diff --git a/drivers/net/wireless/mwifiex/fw.h b/drivers/net/wireless/mwifiex/fw.h
index 1f7578d..1b45aa5 100644
--- a/drivers/net/wireless/mwifiex/fw.h
+++ b/drivers/net/wireless/mwifiex/fw.h
@@ -245,6 +245,8 @@ enum MWIFIEX_802_11_PRIVACY_FILTER {
 #define HT_BW_20    0
 #define HT_BW_40    1
 
+#define DFS_CHAN_MOVE_TIME      10000
+
 #define HostCmd_CMD_GET_HW_SPEC                       0x0003
 #define HostCmd_CMD_802_11_SCAN                       0x0006
 #define HostCmd_CMD_802_11_GET_LOG                    0x000b
@@ -271,6 +273,7 @@ enum MWIFIEX_802_11_PRIVACY_FILTER {
 #define HostCmd_CMD_802_11_SUBSCRIBE_EVENT            0x0075
 #define HostCmd_CMD_802_11_TX_RATE_QUERY              0x007f
 #define HostCmd_CMD_802_11_IBSS_COALESCING_STATUS     0x0083
+#define HostCmd_CMD_CFG_DATA                          0x008f
 #define HostCmd_CMD_VERSION_EXT                       0x0097
 #define HostCmd_CMD_MEF_CFG                           0x009a
 #define HostCmd_CMD_RSSI_INFO                         0x00a4
@@ -279,6 +282,7 @@ enum MWIFIEX_802_11_PRIVACY_FILTER {
 #define HostCmd_CMD_UAP_SYS_CONFIG                    0x00b0
 #define HostCmd_CMD_UAP_BSS_START                     0x00b1
 #define HostCmd_CMD_UAP_BSS_STOP                      0x00b2
+#define HostCmd_CMD_UAP_STA_DEAUTH                    0x00b5
 #define HostCmd_CMD_11N_CFG                           0x00cd
 #define HostCmd_CMD_11N_ADDBA_REQ                     0x00ce
 #define HostCmd_CMD_11N_ADDBA_RSP                     0x00cf
@@ -436,6 +440,7 @@ enum P2P_MODES {
 #define EVENT_BW_CHANGE                 0x00000048
 #define EVENT_UAP_MIC_COUNTERMEASURES   0x0000004c
 #define EVENT_HOSTWAKE_STAIE		0x0000004d
+#define EVENT_CHANNEL_SWITCH_ANN        0x00000050
 #define EVENT_REMAIN_ON_CHAN_EXPIRED    0x0000005f
 
 #define EVENT_ID_MASK                   0xffff
@@ -464,6 +469,8 @@ enum P2P_MODES {
 #define MWIFIEX_CRITERIA_UNICAST	BIT(1)
 #define MWIFIEX_CRITERIA_MULTICAST	BIT(3)
 
+#define CFG_DATA_TYPE_CAL		2
+
 struct mwifiex_ie_types_header {
 	__le16 type;
 	__le16 len;
@@ -971,6 +978,7 @@ enum SNMP_MIB_INDEX {
 	LONG_RETRY_LIM_I = 7,
 	FRAG_THRESH_I = 8,
 	DOT11D_I = 9,
+	DOT11H_I = 10,
 };
 
 #define MAX_SNMP_BUF_SIZE   128
@@ -1197,6 +1205,23 @@ struct host_cmd_ds_amsdu_aggr_ctrl {
 	__le16 curr_buf_size;
 } __packed;
 
+struct host_cmd_ds_sta_deauth {
+	u8 mac[ETH_ALEN];
+	__le16 reason;
+} __packed;
+
+struct mwifiex_ie_types_pwr_capability {
+	struct mwifiex_ie_types_header header;
+	s8 min_pwr;
+	s8 max_pwr;
+};
+
+struct mwifiex_ie_types_local_pwr_constraint {
+	struct mwifiex_ie_types_header header;
+	u8 chan;
+	u8 constraint;
+};
+
 struct mwifiex_ie_types_wmm_param_set {
 	struct mwifiex_ie_types_header header;
 	u8 wmm_ie[1];
@@ -1573,6 +1598,12 @@ struct mwifiex_ie_list {
 	struct mwifiex_ie ie_list[MAX_MGMT_IE_INDEX];
 } __packed;
 
+struct host_cmd_ds_802_11_cfg_data {
+	__le16 action;
+	__le16 type;
+	__le16 data_len;
+} __packed;
+
 struct host_cmd_ds_command {
 	__le16 command;
 	__le16 size;
@@ -1630,7 +1661,9 @@ struct host_cmd_ds_command {
 		struct host_cmd_ds_802_11_eeprom_access eeprom;
 		struct host_cmd_ds_802_11_subsc_evt subsc_evt;
 		struct host_cmd_ds_sys_config uap_sys_config;
+		struct host_cmd_ds_sta_deauth sta_deauth;
 		struct host_cmd_11ac_vht_cfg vht_cfg;
+		struct host_cmd_ds_802_11_cfg_data cfg_data;
 	} params;
 } __packed;
 
diff --git a/drivers/net/wireless/mwifiex/init.c b/drivers/net/wireless/mwifiex/init.c
index 9f44fda..caaf4bd 100644
--- a/drivers/net/wireless/mwifiex/init.c
+++ b/drivers/net/wireless/mwifiex/init.c
@@ -52,87 +52,6 @@ static int mwifiex_add_bss_prio_tbl(struct mwifiex_private *priv)
 	return 0;
 }
 
-static void scan_delay_timer_fn(unsigned long data)
-{
-	struct mwifiex_private *priv = (struct mwifiex_private *)data;
-	struct mwifiex_adapter *adapter = priv->adapter;
-	struct cmd_ctrl_node *cmd_node, *tmp_node;
-	unsigned long flags;
-
-	if (adapter->scan_delay_cnt == MWIFIEX_MAX_SCAN_DELAY_CNT) {
-		/*
-		 * Abort scan operation by cancelling all pending scan
-		 * commands
-		 */
-		spin_lock_irqsave(&adapter->scan_pending_q_lock, flags);
-		list_for_each_entry_safe(cmd_node, tmp_node,
-					 &adapter->scan_pending_q, list) {
-			list_del(&cmd_node->list);
-			mwifiex_insert_cmd_to_free_q(adapter, cmd_node);
-		}
-		spin_unlock_irqrestore(&adapter->scan_pending_q_lock, flags);
-
-		spin_lock_irqsave(&adapter->mwifiex_cmd_lock, flags);
-		adapter->scan_processing = false;
-		adapter->scan_delay_cnt = 0;
-		adapter->empty_tx_q_cnt = 0;
-		spin_unlock_irqrestore(&adapter->mwifiex_cmd_lock, flags);
-
-		if (priv->user_scan_cfg) {
-			if (priv->scan_request) {
-				dev_dbg(priv->adapter->dev,
-					"info: aborting scan\n");
-				cfg80211_scan_done(priv->scan_request, 1);
-				priv->scan_request = NULL;
-			} else {
-				dev_dbg(priv->adapter->dev,
-					"info: scan already aborted\n");
-			}
-
-			kfree(priv->user_scan_cfg);
-			priv->user_scan_cfg = NULL;
-		}
-		goto done;
-	}
-
-	if (!atomic_read(&priv->adapter->is_tx_received)) {
-		adapter->empty_tx_q_cnt++;
-		if (adapter->empty_tx_q_cnt == MWIFIEX_MAX_EMPTY_TX_Q_CNT) {
-			/*
-			 * No Tx traffic for 200msec. Get scan command from
-			 * scan pending queue and put to cmd pending queue to
-			 * resume scan operation
-			 */
-			adapter->scan_delay_cnt = 0;
-			adapter->empty_tx_q_cnt = 0;
-			spin_lock_irqsave(&adapter->scan_pending_q_lock, flags);
-			cmd_node = list_first_entry(&adapter->scan_pending_q,
-						    struct cmd_ctrl_node, list);
-			list_del(&cmd_node->list);
-			spin_unlock_irqrestore(&adapter->scan_pending_q_lock,
-					       flags);
-
-			mwifiex_insert_cmd_to_pending_q(adapter, cmd_node,
-							true);
-			queue_work(adapter->workqueue, &adapter->main_work);
-			goto done;
-		}
-	} else {
-		adapter->empty_tx_q_cnt = 0;
-	}
-
-	/* Delay scan operation further by 20msec */
-	mod_timer(&priv->scan_delay_timer, jiffies +
-		  msecs_to_jiffies(MWIFIEX_SCAN_DELAY_MSEC));
-	adapter->scan_delay_cnt++;
-
-done:
-	if (atomic_read(&priv->adapter->is_tx_received))
-		atomic_set(&priv->adapter->is_tx_received, false);
-
-	return;
-}
-
 /*
  * This function initializes the private structure and sets default
  * values to the members.
@@ -214,8 +133,8 @@ int mwifiex_init_priv(struct mwifiex_private *priv)
 
 	priv->scan_block = false;
 
-	setup_timer(&priv->scan_delay_timer, scan_delay_timer_fn,
-		    (unsigned long)priv);
+	priv->csa_chan = 0;
+	priv->csa_expire_time = 0;
 
 	return mwifiex_add_bss_prio_tbl(priv);
 }
@@ -447,23 +366,29 @@ static void mwifiex_free_lock_list(struct mwifiex_adapter *adapter)
 }
 
 /*
- * This function frees the adapter structure.
+ * This function performs cleanup for adapter structure.
  *
- * The freeing operation is done recursively, by canceling all
- * pending commands, freeing the member buffers previously
- * allocated (command buffers, scan table buffer, sleep confirm
- * command buffer), stopping the timers and calling the cleanup
- * routines for every interface, before the actual adapter
- * structure is freed.
+ * The cleanup is done recursively, by canceling all pending
+ * commands, freeing the member buffers previously allocated
+ * (command buffers, scan table buffer, sleep confirm command
+ * buffer), stopping the timers and calling the cleanup routines
+ * for every interface.
  */
 static void
-mwifiex_free_adapter(struct mwifiex_adapter *adapter)
+mwifiex_adapter_cleanup(struct mwifiex_adapter *adapter)
 {
+	int i;
+
 	if (!adapter) {
 		pr_err("%s: adapter is NULL\n", __func__);
 		return;
 	}
 
+	for (i = 0; i < adapter->priv_num; i++) {
+		if (adapter->priv[i])
+			del_timer_sync(&adapter->priv[i]->scan_delay_timer);
+	}
+
 	mwifiex_cancel_all_pending_cmd(adapter);
 
 	/* Free lock variables */
@@ -684,7 +609,6 @@ mwifiex_shutdown_drv(struct mwifiex_adapter *adapter)
 	int ret = -EINPROGRESS;
 	struct mwifiex_private *priv;
 	s32 i;
-	unsigned long flags;
 	struct sk_buff *skb;
 
 	/* mwifiex already shutdown */
@@ -719,7 +643,7 @@ mwifiex_shutdown_drv(struct mwifiex_adapter *adapter)
 		}
 	}
 
-	spin_lock_irqsave(&adapter->mwifiex_lock, flags);
+	spin_lock(&adapter->mwifiex_lock);
 
 	if (adapter->if_ops.data_complete) {
 		while ((skb = skb_dequeue(&adapter->usb_rx_data_q))) {
@@ -733,10 +657,9 @@ mwifiex_shutdown_drv(struct mwifiex_adapter *adapter)
 		}
 	}
 
-	/* Free adapter structure */
-	mwifiex_free_adapter(adapter);
+	mwifiex_adapter_cleanup(adapter);
 
-	spin_unlock_irqrestore(&adapter->mwifiex_lock, flags);
+	spin_unlock(&adapter->mwifiex_lock);
 
 	/* Notify completion */
 	ret = mwifiex_shutdown_fw_complete(adapter);
diff --git a/drivers/net/wireless/mwifiex/join.c b/drivers/net/wireless/mwifiex/join.c
index 6bcb66e..1c8a771 100644
--- a/drivers/net/wireless/mwifiex/join.c
+++ b/drivers/net/wireless/mwifiex/join.c
@@ -534,6 +534,8 @@ int mwifiex_cmd_802_11_associate(struct mwifiex_private *priv,
 
 	mwifiex_cmd_append_tsf_tlv(priv, &pos, bss_desc);
 
+	mwifiex_11h_process_join(priv, &pos, bss_desc);
+
 	cmd->size = cpu_to_le16((u16) (pos - (u8 *) assoc) + S_DS_GEN);
 
 	/* Set the Capability info at last */
@@ -919,9 +921,8 @@ mwifiex_cmd_802_11_ad_hoc_start(struct mwifiex_private *priv,
 	memcpy(&priv->curr_bss_params.data_rates,
 	       &adhoc_start->data_rate, priv->curr_bss_params.num_of_rates);
 
-	dev_dbg(adapter->dev, "info: ADHOC_S_CMD: rates=%02x %02x %02x %02x\n",
-		adhoc_start->data_rate[0], adhoc_start->data_rate[1],
-		adhoc_start->data_rate[2], adhoc_start->data_rate[3]);
+	dev_dbg(adapter->dev, "info: ADHOC_S_CMD: rates=%4ph\n",
+		adhoc_start->data_rate);
 
 	dev_dbg(adapter->dev, "info: ADHOC_S_CMD: AD-HOC Start command is ready\n");
 
diff --git a/drivers/net/wireless/mwifiex/main.c b/drivers/net/wireless/mwifiex/main.c
index 2eb88ea..e15ab72 100644
--- a/drivers/net/wireless/mwifiex/main.c
+++ b/drivers/net/wireless/mwifiex/main.c
@@ -25,6 +25,86 @@
 #define VERSION	"1.0"
 
 const char driver_version[] = "mwifiex " VERSION " (%s) ";
+static char *cal_data_cfg;
+module_param(cal_data_cfg, charp, 0);
+
+static void scan_delay_timer_fn(unsigned long data)
+{
+	struct mwifiex_private *priv = (struct mwifiex_private *)data;
+	struct mwifiex_adapter *adapter = priv->adapter;
+	struct cmd_ctrl_node *cmd_node, *tmp_node;
+	unsigned long flags;
+
+	if (adapter->surprise_removed)
+		return;
+
+	if (adapter->scan_delay_cnt == MWIFIEX_MAX_SCAN_DELAY_CNT) {
+		/*
+		 * Abort scan operation by cancelling all pending scan
+		 * commands
+		 */
+		spin_lock_irqsave(&adapter->scan_pending_q_lock, flags);
+		list_for_each_entry_safe(cmd_node, tmp_node,
+					 &adapter->scan_pending_q, list) {
+			list_del(&cmd_node->list);
+			mwifiex_insert_cmd_to_free_q(adapter, cmd_node);
+		}
+		spin_unlock_irqrestore(&adapter->scan_pending_q_lock, flags);
+
+		spin_lock_irqsave(&adapter->mwifiex_cmd_lock, flags);
+		adapter->scan_processing = false;
+		adapter->scan_delay_cnt = 0;
+		adapter->empty_tx_q_cnt = 0;
+		spin_unlock_irqrestore(&adapter->mwifiex_cmd_lock, flags);
+
+		if (priv->scan_request) {
+			dev_dbg(adapter->dev, "info: aborting scan\n");
+			cfg80211_scan_done(priv->scan_request, 1);
+			priv->scan_request = NULL;
+		} else {
+			priv->scan_aborting = false;
+			dev_dbg(adapter->dev, "info: scan already aborted\n");
+		}
+		goto done;
+	}
+
+	if (!atomic_read(&priv->adapter->is_tx_received)) {
+		adapter->empty_tx_q_cnt++;
+		if (adapter->empty_tx_q_cnt == MWIFIEX_MAX_EMPTY_TX_Q_CNT) {
+			/*
+			 * No Tx traffic for 200msec. Get scan command from
+			 * scan pending queue and put to cmd pending queue to
+			 * resume scan operation
+			 */
+			adapter->scan_delay_cnt = 0;
+			adapter->empty_tx_q_cnt = 0;
+			spin_lock_irqsave(&adapter->scan_pending_q_lock, flags);
+			cmd_node = list_first_entry(&adapter->scan_pending_q,
+						    struct cmd_ctrl_node, list);
+			list_del(&cmd_node->list);
+			spin_unlock_irqrestore(&adapter->scan_pending_q_lock,
+					       flags);
+
+			mwifiex_insert_cmd_to_pending_q(adapter, cmd_node,
+							true);
+			queue_work(adapter->workqueue, &adapter->main_work);
+			goto done;
+		}
+	} else {
+		adapter->empty_tx_q_cnt = 0;
+	}
+
+	/* Delay scan operation further by 20msec */
+	mod_timer(&priv->scan_delay_timer, jiffies +
+		  msecs_to_jiffies(MWIFIEX_SCAN_DELAY_MSEC));
+	adapter->scan_delay_cnt++;
+
+done:
+	if (atomic_read(&priv->adapter->is_tx_received))
+		atomic_set(&priv->adapter->is_tx_received, false);
+
+	return;
+}
 
 /*
  * This function registers the device and performs all the necessary
@@ -73,6 +153,10 @@ static int mwifiex_register(void *card, struct mwifiex_if_ops *if_ops,
 
 		adapter->priv[i]->adapter = adapter;
 		adapter->priv_num++;
+
+		setup_timer(&adapter->priv[i]->scan_delay_timer,
+			    scan_delay_timer_fn,
+			    (unsigned long)adapter->priv[i]);
 	}
 	mwifiex_init_lock_list(adapter);
 
@@ -336,6 +420,13 @@ static void mwifiex_fw_dpc(const struct firmware *firmware, void *context)
 
 	dev_notice(adapter->dev, "WLAN FW is active\n");
 
+	if (cal_data_cfg) {
+		if ((request_firmware(&adapter->cal_data, cal_data_cfg,
+				      adapter->dev)) < 0)
+			dev_err(adapter->dev,
+				"Cal data request_firmware() failed\n");
+	}
+
 	adapter->init_wait_q_woken = false;
 	ret = mwifiex_init_fw(adapter);
 	if (ret == -1) {
@@ -390,6 +481,10 @@ err_init_fw:
 	pr_debug("info: %s: unregister device\n", __func__);
 	adapter->if_ops.unregister_dev(adapter);
 done:
+	if (adapter->cal_data) {
+		release_firmware(adapter->cal_data);
+		adapter->cal_data = NULL;
+	}
 	release_firmware(adapter->firmware);
 	complete(&adapter->fw_load);
 	return;
@@ -436,6 +531,7 @@ mwifiex_close(struct net_device *dev)
 		dev_dbg(priv->adapter->dev, "aborting scan on ndo_stop\n");
 		cfg80211_scan_done(priv->scan_request, 1);
 		priv->scan_request = NULL;
+		priv->scan_aborting = true;
 	}
 
 	return 0;
@@ -573,9 +669,8 @@ static void mwifiex_set_multicast_list(struct net_device *dev)
 		mcast_list.mode = MWIFIEX_ALL_MULTI_MODE;
 	} else {
 		mcast_list.mode = MWIFIEX_MULTICAST_MODE;
-		if (netdev_mc_count(dev))
-			mcast_list.num_multicast_addr =
-				mwifiex_copy_mcast_addr(&mcast_list, dev);
+		mcast_list.num_multicast_addr =
+			mwifiex_copy_mcast_addr(&mcast_list, dev);
 	}
 	mwifiex_request_set_multicast_list(priv, &mcast_list);
 }
diff --git a/drivers/net/wireless/mwifiex/main.h b/drivers/net/wireless/mwifiex/main.h
index 4ef67fc..3da73d3 100644
--- a/drivers/net/wireless/mwifiex/main.h
+++ b/drivers/net/wireless/mwifiex/main.h
@@ -309,6 +309,9 @@ struct mwifiex_bssdescriptor {
 	u16 wapi_offset;
 	u8 *beacon_buf;
 	u32 beacon_buf_size;
+	u8 sensed_11h;
+	u8 local_constraint;
+	u8 chan_sw_ie_present;
 };
 
 struct mwifiex_current_bss_params {
@@ -492,7 +495,6 @@ struct mwifiex_private {
 	struct semaphore async_sem;
 	u8 report_scan_result;
 	struct cfg80211_scan_request *scan_request;
-	struct mwifiex_user_scan_cfg *user_scan_cfg;
 	u8 cfg_bssid[6];
 	struct wps wps;
 	u8 scan_block;
@@ -510,6 +512,9 @@ struct mwifiex_private {
 	u8 ap_11ac_enabled;
 	u32 mgmt_frame_mask;
 	struct mwifiex_roc_cfg roc_cfg;
+	bool scan_aborting;
+	u8 csa_chan;
+	unsigned long csa_expire_time;
 };
 
 enum mwifiex_ba_status {
@@ -730,6 +735,7 @@ struct mwifiex_adapter {
 	u16 max_mgmt_ie_index;
 	u8 scan_delay_cnt;
 	u8 empty_tx_q_cnt;
+	const struct firmware *cal_data;
 
 	/* 11AC */
 	u32 is_hw_11ac_capable;
@@ -1017,6 +1023,24 @@ static inline bool mwifiex_is_skb_mgmt_frame(struct sk_buff *skb)
 	return (*(u32 *)skb->data == PKT_TYPE_MGMT);
 }
 
+/* This function retrieves channel closed for operation by Channel
+ * Switch Announcement.
+ */
+static inline u8
+mwifiex_11h_get_csa_closed_channel(struct mwifiex_private *priv)
+{
+	if (!priv->csa_chan)
+		return 0;
+
+	/* Clear csa channel, if DFS channel move time has passed */
+	if (jiffies > priv->csa_expire_time) {
+		priv->csa_chan = 0;
+		priv->csa_expire_time = 0;
+	}
+
+	return priv->csa_chan;
+}
+
 int mwifiex_init_shutdown_fw(struct mwifiex_private *priv,
 			     u32 func_init_shutdown);
 int mwifiex_add_card(void *, struct semaphore *, struct mwifiex_if_ops *, u8);
@@ -1115,6 +1139,12 @@ int mwifiex_set_mgmt_ies(struct mwifiex_private *priv,
 			 struct cfg80211_beacon_data *data);
 int mwifiex_del_mgmt_ies(struct mwifiex_private *priv);
 u8 *mwifiex_11d_code_2_region(u8 code);
+void mwifiex_uap_del_sta_data(struct mwifiex_private *priv,
+			      struct mwifiex_sta_node *node);
+
+void mwifiex_11h_process_join(struct mwifiex_private *priv, u8 **buffer,
+			      struct mwifiex_bssdescriptor *bss_desc);
+int mwifiex_11h_handle_event_chanswann(struct mwifiex_private *priv);
 
 extern const struct ethtool_ops mwifiex_ethtool_ops;
 
diff --git a/drivers/net/wireless/mwifiex/scan.c b/drivers/net/wireless/mwifiex/scan.c
index 9cf5d8f..c447d9b 100644
--- a/drivers/net/wireless/mwifiex/scan.c
+++ b/drivers/net/wireless/mwifiex/scan.c
@@ -391,6 +391,12 @@ mwifiex_is_network_compatible(struct mwifiex_private *priv,
 		return 0;
 	}
 
+	if (bss_desc->chan_sw_ie_present) {
+		dev_err(adapter->dev,
+			"Don't connect to AP with WLAN_EID_CHANNEL_SWITCH\n");
+		return -1;
+	}
+
 	if (mwifiex_is_bss_wapi(priv, bss_desc)) {
 		dev_dbg(adapter->dev, "info: return success for WAPI AP\n");
 		return 0;
@@ -569,6 +575,9 @@ mwifiex_scan_channel_list(struct mwifiex_private *priv,
 		return -1;
 	}
 
+	/* Check csa channel expiry before preparing scan list */
+	mwifiex_11h_get_csa_closed_channel(priv);
+
 	chan_tlv_out->header.type = cpu_to_le16(TLV_TYPE_CHANLIST);
 
 	/* Set the temp channel struct pointer to the start of the desired
@@ -598,6 +607,11 @@ mwifiex_scan_channel_list(struct mwifiex_private *priv,
 		while (tlv_idx < max_chan_per_scan &&
 		       tmp_chan_list->chan_number && !done_early) {
 
+			if (tmp_chan_list->chan_number == priv->csa_chan) {
+				tmp_chan_list++;
+				continue;
+			}
+
 			dev_dbg(priv->adapter->dev,
 				"info: Scan: Chan(%3d), Radio(%d),"
 				" Mode(%d, %d), Dur(%d)\n",
@@ -1169,6 +1183,19 @@ int mwifiex_update_bss_desc_with_ie(struct mwifiex_adapter *adapter,
 			bss_entry->erp_flags = *(current_ptr + 2);
 			break;
 
+		case WLAN_EID_PWR_CONSTRAINT:
+			bss_entry->local_constraint = *(current_ptr + 2);
+			bss_entry->sensed_11h = true;
+			break;
+
+		case WLAN_EID_CHANNEL_SWITCH:
+			bss_entry->chan_sw_ie_present = true;
+		case WLAN_EID_PWR_CAPABILITY:
+		case WLAN_EID_TPC_REPORT:
+		case WLAN_EID_QUIET:
+			bss_entry->sensed_11h = true;
+		    break;
+
 		case WLAN_EID_EXT_SUPP_RATES:
 			/*
 			 * Only process extended supported rate
@@ -1575,6 +1602,9 @@ int mwifiex_ret_802_11_scan(struct mwifiex_private *priv,
 		goto check_next_scan;
 	}
 
+	/* Check csa channel expiry before parsing scan response */
+	mwifiex_11h_get_csa_closed_channel(priv);
+
 	bytes_left = le16_to_cpu(scan_rsp->bss_descript_size);
 	dev_dbg(adapter->dev, "info: SCAN_RESP: bss_descript_size %d\n",
 		bytes_left);
@@ -1727,6 +1757,13 @@ int mwifiex_ret_802_11_scan(struct mwifiex_private *priv,
 			struct ieee80211_channel *chan;
 			u8 band;
 
+			/* Skip entry if on csa closed channel */
+			if (channel == priv->csa_chan) {
+				dev_dbg(adapter->dev,
+					"Dropping entry on csa closed channel\n");
+				continue;
+			}
+
 			band = BAND_G;
 			if (chan_band_tlv) {
 				chan_band =
@@ -1784,22 +1821,17 @@ check_next_scan:
 		if (priv->report_scan_result)
 			priv->report_scan_result = false;
 
-		if (priv->user_scan_cfg) {
-			if (priv->scan_request) {
-				dev_dbg(priv->adapter->dev,
-					"info: notifying scan done\n");
-				cfg80211_scan_done(priv->scan_request, 0);
-				priv->scan_request = NULL;
-			} else {
-				dev_dbg(priv->adapter->dev,
-					"info: scan already aborted\n");
-			}
-
-			kfree(priv->user_scan_cfg);
-			priv->user_scan_cfg = NULL;
+		if (priv->scan_request) {
+			dev_dbg(adapter->dev, "info: notifying scan done\n");
+			cfg80211_scan_done(priv->scan_request, 0);
+			priv->scan_request = NULL;
+		} else {
+			priv->scan_aborting = false;
+			dev_dbg(adapter->dev, "info: scan already aborted\n");
 		}
 	} else {
-		if (priv->user_scan_cfg && !priv->scan_request) {
+		if ((priv->scan_aborting && !priv->scan_request) ||
+		    priv->scan_block) {
 			spin_unlock_irqrestore(&adapter->scan_pending_q_lock,
 					       flags);
 			adapter->scan_delay_cnt = MWIFIEX_MAX_SCAN_DELAY_CNT;
diff --git a/drivers/net/wireless/mwifiex/sdio.c b/drivers/net/wireless/mwifiex/sdio.c
index 363ba31..5ee5ed0 100644
--- a/drivers/net/wireless/mwifiex/sdio.c
+++ b/drivers/net/wireless/mwifiex/sdio.c
@@ -77,6 +77,17 @@ mwifiex_sdio_probe(struct sdio_func *func, const struct sdio_device_id *id)
 
 	func->card->quirks |= MMC_QUIRK_BLKSZ_FOR_BYTE_MODE;
 
+	if (id->driver_data) {
+		struct mwifiex_sdio_device *data = (void *)id->driver_data;
+
+		card->firmware = data->firmware;
+		card->reg = data->reg;
+		card->max_ports = data->max_ports;
+		card->mp_agg_pkt_limit = data->mp_agg_pkt_limit;
+		card->supports_sdio_new_mode = data->supports_sdio_new_mode;
+		card->has_control_mask = data->has_control_mask;
+	}
+
 	sdio_claim_host(func);
 	ret = sdio_enable_func(func);
 	sdio_release_host(func);
@@ -251,12 +262,19 @@ static int mwifiex_sdio_resume(struct device *dev)
 #define SDIO_DEVICE_ID_MARVELL_8787   (0x9119)
 /* Device ID for SD8797 */
 #define SDIO_DEVICE_ID_MARVELL_8797   (0x9129)
+/* Device ID for SD8897 */
+#define SDIO_DEVICE_ID_MARVELL_8897   (0x912d)
 
 /* WLAN IDs */
 static const struct sdio_device_id mwifiex_ids[] = {
-	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8786)},
-	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8787)},
-	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8797)},
+	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8786),
+		.driver_data = (unsigned long) &mwifiex_sdio_sd8786},
+	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8787),
+		.driver_data = (unsigned long) &mwifiex_sdio_sd8787},
+	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8797),
+		.driver_data = (unsigned long) &mwifiex_sdio_sd8797},
+	{SDIO_DEVICE(SDIO_VENDOR_ID_MARVELL, SDIO_DEVICE_ID_MARVELL_8897),
+		.driver_data = (unsigned long) &mwifiex_sdio_sd8897},
 	{},
 };
 
@@ -282,13 +300,13 @@ static struct sdio_driver mwifiex_sdio = {
  * This function writes data into SDIO card register.
  */
 static int
-mwifiex_write_reg(struct mwifiex_adapter *adapter, u32 reg, u32 data)
+mwifiex_write_reg(struct mwifiex_adapter *adapter, u32 reg, u8 data)
 {
 	struct sdio_mmc_card *card = adapter->card;
 	int ret = -1;
 
 	sdio_claim_host(card->func);
-	sdio_writeb(card->func, (u8) data, reg, &ret);
+	sdio_writeb(card->func, data, reg, &ret);
 	sdio_release_host(card->func);
 
 	return ret;
@@ -298,7 +316,7 @@ mwifiex_write_reg(struct mwifiex_adapter *adapter, u32 reg, u32 data)
  * This function reads data from SDIO card register.
  */
 static int
-mwifiex_read_reg(struct mwifiex_adapter *adapter, u32 reg, u32 *data)
+mwifiex_read_reg(struct mwifiex_adapter *adapter, u32 reg, u8 *data)
 {
 	struct sdio_mmc_card *card = adapter->card;
 	int ret = -1;
@@ -400,7 +418,40 @@ static int mwifiex_pm_wakeup_card_complete(struct mwifiex_adapter *adapter)
 }
 
 /*
- * This function initializes the IO ports.
+ * This function is used to initialize IO ports for the
+ * chipsets supporting SDIO new mode eg SD8897.
+ */
+static int mwifiex_init_sdio_new_mode(struct mwifiex_adapter *adapter)
+{
+	u8 reg;
+
+	adapter->ioport = MEM_PORT;
+
+	/* enable sdio new mode */
+	if (mwifiex_read_reg(adapter, CARD_CONFIG_2_1_REG, &reg))
+		return -1;
+	if (mwifiex_write_reg(adapter, CARD_CONFIG_2_1_REG,
+			      reg | CMD53_NEW_MODE))
+		return -1;
+
+	/* Configure cmd port and enable reading rx length from the register */
+	if (mwifiex_read_reg(adapter, CMD_CONFIG_0, &reg))
+		return -1;
+	if (mwifiex_write_reg(adapter, CMD_CONFIG_0, reg | CMD_PORT_RD_LEN_EN))
+		return -1;
+
+	/* Enable Dnld/Upld ready auto reset for cmd port after cmd53 is
+	 * completed
+	 */
+	if (mwifiex_read_reg(adapter, CMD_CONFIG_1, &reg))
+		return -1;
+	if (mwifiex_write_reg(adapter, CMD_CONFIG_1, reg | CMD_PORT_AUTO_EN))
+		return -1;
+
+	return 0;
+}
+
+/* This function initializes the IO ports.
  *
  * The following operations are performed -
  *      - Read the IO ports (0, 1 and 2)
@@ -409,10 +460,17 @@ static int mwifiex_pm_wakeup_card_complete(struct mwifiex_adapter *adapter)
  */
 static int mwifiex_init_sdio_ioport(struct mwifiex_adapter *adapter)
 {
-	u32 reg;
+	u8 reg;
+	struct sdio_mmc_card *card = adapter->card;
 
 	adapter->ioport = 0;
 
+	if (card->supports_sdio_new_mode) {
+		if (mwifiex_init_sdio_new_mode(adapter))
+			return -1;
+		goto cont;
+	}
+
 	/* Read the IO port */
 	if (!mwifiex_read_reg(adapter, IO_PORT_0_REG, &reg))
 		adapter->ioport |= (reg & 0xff);
@@ -428,19 +486,19 @@ static int mwifiex_init_sdio_ioport(struct mwifiex_adapter *adapter)
 		adapter->ioport |= ((reg & 0xff) << 16);
 	else
 		return -1;
-
+cont:
 	pr_debug("info: SDIO FUNC1 IO port: %#x\n", adapter->ioport);
 
 	/* Set Host interrupt reset to read to clear */
 	if (!mwifiex_read_reg(adapter, HOST_INT_RSR_REG, &reg))
 		mwifiex_write_reg(adapter, HOST_INT_RSR_REG,
-				  reg | SDIO_INT_MASK);
+				  reg | card->reg->sdio_int_mask);
 	else
 		return -1;
 
 	/* Dnld/Upld ready set to auto reset */
-	if (!mwifiex_read_reg(adapter, CARD_MISC_CFG_REG, &reg))
-		mwifiex_write_reg(adapter, CARD_MISC_CFG_REG,
+	if (!mwifiex_read_reg(adapter, card->reg->card_misc_cfg_reg, &reg))
+		mwifiex_write_reg(adapter, card->reg->card_misc_cfg_reg,
 				  reg | AUTO_RE_ENABLE_INT);
 	else
 		return -1;
@@ -486,34 +544,42 @@ static int mwifiex_write_data_to_card(struct mwifiex_adapter *adapter,
 static int mwifiex_get_rd_port(struct mwifiex_adapter *adapter, u8 *port)
 {
 	struct sdio_mmc_card *card = adapter->card;
-	u16 rd_bitmap = card->mp_rd_bitmap;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
+	u32 rd_bitmap = card->mp_rd_bitmap;
 
-	dev_dbg(adapter->dev, "data: mp_rd_bitmap=0x%04x\n", rd_bitmap);
+	dev_dbg(adapter->dev, "data: mp_rd_bitmap=0x%08x\n", rd_bitmap);
 
-	if (!(rd_bitmap & (CTRL_PORT_MASK | DATA_PORT_MASK)))
-		return -1;
+	if (card->supports_sdio_new_mode) {
+		if (!(rd_bitmap & reg->data_port_mask))
+			return -1;
+	} else {
+		if (!(rd_bitmap & (CTRL_PORT_MASK | reg->data_port_mask)))
+			return -1;
+	}
 
-	if (card->mp_rd_bitmap & CTRL_PORT_MASK) {
-		card->mp_rd_bitmap &= (u16) (~CTRL_PORT_MASK);
+	if ((card->has_control_mask) &&
+	    (card->mp_rd_bitmap & CTRL_PORT_MASK)) {
+		card->mp_rd_bitmap &= (u32) (~CTRL_PORT_MASK);
 		*port = CTRL_PORT;
-		dev_dbg(adapter->dev, "data: port=%d mp_rd_bitmap=0x%04x\n",
+		dev_dbg(adapter->dev, "data: port=%d mp_rd_bitmap=0x%08x\n",
 			*port, card->mp_rd_bitmap);
-	} else {
-		if (card->mp_rd_bitmap & (1 << card->curr_rd_port)) {
-			card->mp_rd_bitmap &= (u16)
-						(~(1 << card->curr_rd_port));
-			*port = card->curr_rd_port;
+		return 0;
+	}
 
-			if (++card->curr_rd_port == MAX_PORT)
-				card->curr_rd_port = 1;
-		} else {
-			return -1;
-		}
+	if (!(card->mp_rd_bitmap & (1 << card->curr_rd_port)))
+		return -1;
+
+	/* We are now handling the SDIO data ports */
+	card->mp_rd_bitmap &= (u32)(~(1 << card->curr_rd_port));
+	*port = card->curr_rd_port;
+
+	if (++card->curr_rd_port == card->max_ports)
+		card->curr_rd_port = reg->start_rd_port;
+
+	dev_dbg(adapter->dev,
+		"data: port=%d mp_rd_bitmap=0x%08x -> 0x%08x\n",
+		*port, rd_bitmap, card->mp_rd_bitmap);
 
-		dev_dbg(adapter->dev,
-			"data: port=%d mp_rd_bitmap=0x%04x -> 0x%04x\n",
-			*port, rd_bitmap, card->mp_rd_bitmap);
-	}
 	return 0;
 }
 
@@ -524,35 +590,45 @@ static int mwifiex_get_rd_port(struct mwifiex_adapter *adapter, u8 *port)
  * increased (provided it does not reach the maximum limit, in which
  * case it is reset to 1)
  */
-static int mwifiex_get_wr_port_data(struct mwifiex_adapter *adapter, u8 *port)
+static int mwifiex_get_wr_port_data(struct mwifiex_adapter *adapter, u32 *port)
 {
 	struct sdio_mmc_card *card = adapter->card;
-	u16 wr_bitmap = card->mp_wr_bitmap;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
+	u32 wr_bitmap = card->mp_wr_bitmap;
 
-	dev_dbg(adapter->dev, "data: mp_wr_bitmap=0x%04x\n", wr_bitmap);
+	dev_dbg(adapter->dev, "data: mp_wr_bitmap=0x%08x\n", wr_bitmap);
 
-	if (!(wr_bitmap & card->mp_data_port_mask))
+	if (card->supports_sdio_new_mode &&
+	    !(wr_bitmap & reg->data_port_mask)) {
+		adapter->data_sent = true;
+		return -EBUSY;
+	} else if (!card->supports_sdio_new_mode &&
+		   !(wr_bitmap & card->mp_data_port_mask)) {
 		return -1;
+	}
 
 	if (card->mp_wr_bitmap & (1 << card->curr_wr_port)) {
-		card->mp_wr_bitmap &= (u16) (~(1 << card->curr_wr_port));
+		card->mp_wr_bitmap &= (u32) (~(1 << card->curr_wr_port));
 		*port = card->curr_wr_port;
-		if (++card->curr_wr_port == card->mp_end_port)
-			card->curr_wr_port = 1;
+		if (((card->supports_sdio_new_mode) &&
+		     (++card->curr_wr_port == card->max_ports)) ||
+		    ((!card->supports_sdio_new_mode) &&
+		     (++card->curr_wr_port == card->mp_end_port)))
+			card->curr_wr_port = reg->start_wr_port;
 	} else {
 		adapter->data_sent = true;
 		return -EBUSY;
 	}
 
-	if (*port == CTRL_PORT) {
-		dev_err(adapter->dev, "invalid data port=%d cur port=%d"
-			" mp_wr_bitmap=0x%04x -> 0x%04x\n",
+	if ((card->has_control_mask) && (*port == CTRL_PORT)) {
+		dev_err(adapter->dev,
+			"invalid data port=%d cur port=%d mp_wr_bitmap=0x%08x -> 0x%08x\n",
 			*port, card->curr_wr_port, wr_bitmap,
 			card->mp_wr_bitmap);
 		return -1;
 	}
 
-	dev_dbg(adapter->dev, "data: port=%d mp_wr_bitmap=0x%04x -> 0x%04x\n",
+	dev_dbg(adapter->dev, "data: port=%d mp_wr_bitmap=0x%08x -> 0x%08x\n",
 		*port, wr_bitmap, card->mp_wr_bitmap);
 
 	return 0;
@@ -564,11 +640,12 @@ static int mwifiex_get_wr_port_data(struct mwifiex_adapter *adapter, u8 *port)
 static int
 mwifiex_sdio_poll_card_status(struct mwifiex_adapter *adapter, u8 bits)
 {
+	struct sdio_mmc_card *card = adapter->card;
 	u32 tries;
-	u32 cs;
+	u8 cs;
 
 	for (tries = 0; tries < MAX_POLL_TRIES; tries++) {
-		if (mwifiex_read_reg(adapter, CARD_STATUS_REG, &cs))
+		if (mwifiex_read_reg(adapter, card->reg->poll_reg, &cs))
 			break;
 		else if ((cs & bits) == bits)
 			return 0;
@@ -587,12 +664,14 @@ mwifiex_sdio_poll_card_status(struct mwifiex_adapter *adapter, u8 bits)
 static int
 mwifiex_sdio_read_fw_status(struct mwifiex_adapter *adapter, u16 *dat)
 {
-	u32 fws0, fws1;
+	struct sdio_mmc_card *card = adapter->card;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
+	u8 fws0, fws1;
 
-	if (mwifiex_read_reg(adapter, CARD_FW_STATUS0_REG, &fws0))
+	if (mwifiex_read_reg(adapter, reg->status_reg_0, &fws0))
 		return -1;
 
-	if (mwifiex_read_reg(adapter, CARD_FW_STATUS1_REG, &fws1))
+	if (mwifiex_read_reg(adapter, reg->status_reg_1, &fws1))
 		return -1;
 
 	*dat = (u16) ((fws1 << 8) | fws0);
@@ -608,14 +687,14 @@ mwifiex_sdio_read_fw_status(struct mwifiex_adapter *adapter, u16 *dat)
  */
 static int mwifiex_sdio_disable_host_int(struct mwifiex_adapter *adapter)
 {
-	u32 host_int_mask;
+	u8 host_int_mask, host_int_disable = HOST_INT_DISABLE;
 
 	/* Read back the host_int_mask register */
 	if (mwifiex_read_reg(adapter, HOST_INT_MASK_REG, &host_int_mask))
 		return -1;
 
 	/* Update with the mask and write back to the register */
-	host_int_mask &= ~HOST_INT_DISABLE;
+	host_int_mask &= ~host_int_disable;
 
 	if (mwifiex_write_reg(adapter, HOST_INT_MASK_REG, host_int_mask)) {
 		dev_err(adapter->dev, "disable host interrupt failed\n");
@@ -633,8 +712,11 @@ static int mwifiex_sdio_disable_host_int(struct mwifiex_adapter *adapter)
  */
 static int mwifiex_sdio_enable_host_int(struct mwifiex_adapter *adapter)
 {
+	struct sdio_mmc_card *card = adapter->card;
+
 	/* Simply write the mask to the register */
-	if (mwifiex_write_reg(adapter, HOST_INT_MASK_REG, HOST_INT_ENABLE)) {
+	if (mwifiex_write_reg(adapter, HOST_INT_MASK_REG,
+			      card->reg->host_int_enable)) {
 		dev_err(adapter->dev, "enable host interrupt failed\n");
 		return -1;
 	}
@@ -686,11 +768,13 @@ static int mwifiex_sdio_card_to_host(struct mwifiex_adapter *adapter,
 static int mwifiex_prog_fw_w_helper(struct mwifiex_adapter *adapter,
 				    struct mwifiex_fw_image *fw)
 {
+	struct sdio_mmc_card *card = adapter->card;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
 	int ret;
 	u8 *firmware = fw->fw_buf;
 	u32 firmware_len = fw->fw_len;
 	u32 offset = 0;
-	u32 base0, base1;
+	u8 base0, base1;
 	u8 *fwbuf;
 	u16 len = 0;
 	u32 txlen, tx_blocks = 0, tries;
@@ -727,7 +811,7 @@ static int mwifiex_prog_fw_w_helper(struct mwifiex_adapter *adapter,
 			break;
 
 		for (tries = 0; tries < MAX_POLL_TRIES; tries++) {
-			ret = mwifiex_read_reg(adapter, HOST_F1_RD_BASE_0,
+			ret = mwifiex_read_reg(adapter, reg->base_0_reg,
 					       &base0);
 			if (ret) {
 				dev_err(adapter->dev,
@@ -736,7 +820,7 @@ static int mwifiex_prog_fw_w_helper(struct mwifiex_adapter *adapter,
 					base0, base0);
 				goto done;
 			}
-			ret = mwifiex_read_reg(adapter, HOST_F1_RD_BASE_1,
+			ret = mwifiex_read_reg(adapter, reg->base_1_reg,
 					       &base1);
 			if (ret) {
 				dev_err(adapter->dev,
@@ -828,10 +912,11 @@ done:
 static int mwifiex_check_fw_status(struct mwifiex_adapter *adapter,
 				   u32 poll_num)
 {
+	struct sdio_mmc_card *card = adapter->card;
 	int ret = 0;
 	u16 firmware_stat;
 	u32 tries;
-	u32 winner_status;
+	u8 winner_status;
 
 	/* Wait for firmware initialization event */
 	for (tries = 0; tries < poll_num; tries++) {
@@ -849,7 +934,7 @@ static int mwifiex_check_fw_status(struct mwifiex_adapter *adapter,
 
 	if (ret) {
 		if (mwifiex_read_reg
-		    (adapter, CARD_FW_STATUS0_REG, &winner_status))
+		    (adapter, card->reg->status_reg_0, &winner_status))
 			winner_status = 0;
 
 		if (winner_status)
@@ -866,12 +951,12 @@ static int mwifiex_check_fw_status(struct mwifiex_adapter *adapter,
 static void mwifiex_interrupt_status(struct mwifiex_adapter *adapter)
 {
 	struct sdio_mmc_card *card = adapter->card;
-	u32 sdio_ireg;
+	u8 sdio_ireg;
 	unsigned long flags;
 
-	if (mwifiex_read_data_sync(adapter, card->mp_regs, MAX_MP_REGS,
-				   REG_PORT | MWIFIEX_SDIO_BYTE_MODE_MASK,
-				   0)) {
+	if (mwifiex_read_data_sync(adapter, card->mp_regs,
+				   card->reg->max_mp_regs,
+				   REG_PORT | MWIFIEX_SDIO_BYTE_MODE_MASK, 0)) {
 		dev_err(adapter->dev, "read mp_regs failed\n");
 		return;
 	}
@@ -880,6 +965,9 @@ static void mwifiex_interrupt_status(struct mwifiex_adapter *adapter)
 	if (sdio_ireg) {
 		/*
 		 * DN_LD_HOST_INT_STATUS and/or UP_LD_HOST_INT_STATUS
+		 * For SDIO new mode CMD port interrupts
+		 *	DN_LD_CMD_PORT_HOST_INT_STATUS and/or
+		 *	UP_LD_CMD_PORT_HOST_INT_STATUS
 		 * Clear the interrupt status register
 		 */
 		dev_dbg(adapter->dev, "int: sdio_ireg = %#x\n", sdio_ireg);
@@ -1003,11 +1091,11 @@ static int mwifiex_sdio_card_to_host_mp_aggr(struct mwifiex_adapter *adapter,
 	s32 f_aggr_cur = 0;
 	struct sk_buff *skb_deaggr;
 	u32 pind;
-	u32 pkt_len, pkt_type = 0;
+	u32 pkt_len, pkt_type, mport;
 	u8 *curr_ptr;
 	u32 rx_len = skb->len;
 
-	if (port == CTRL_PORT) {
+	if ((card->has_control_mask) && (port == CTRL_PORT)) {
 		/* Read the command Resp without aggr */
 		dev_dbg(adapter->dev, "info: %s: no aggregation for cmd "
 			"response\n", __func__);
@@ -1024,7 +1112,10 @@ static int mwifiex_sdio_card_to_host_mp_aggr(struct mwifiex_adapter *adapter,
 		goto rx_curr_single;
 	}
 
-	if (card->mp_rd_bitmap & (~((u16) CTRL_PORT_MASK))) {
+	if ((!card->has_control_mask && (card->mp_rd_bitmap &
+					 card->reg->data_port_mask)) ||
+	    (card->has_control_mask && (card->mp_rd_bitmap &
+					(~((u32) CTRL_PORT_MASK))))) {
 		/* Some more data RX pending */
 		dev_dbg(adapter->dev, "info: %s: not last packet\n", __func__);
 
@@ -1060,10 +1151,10 @@ static int mwifiex_sdio_card_to_host_mp_aggr(struct mwifiex_adapter *adapter,
 	if (f_aggr_cur) {
 		dev_dbg(adapter->dev, "info: current packet aggregation\n");
 		/* Curr pkt can be aggregated */
-		MP_RX_AGGR_SETUP(card, skb, port);
+		mp_rx_aggr_setup(card, skb, port);
 
 		if (MP_RX_AGGR_PKT_LIMIT_REACHED(card) ||
-		    MP_RX_AGGR_PORT_LIMIT_REACHED(card)) {
+		    mp_rx_aggr_port_limit_reached(card)) {
 			dev_dbg(adapter->dev, "info: %s: aggregated packet "
 				"limit reached\n", __func__);
 			/* No more pkts allowed in Aggr buf, rx it */
@@ -1076,11 +1167,28 @@ static int mwifiex_sdio_card_to_host_mp_aggr(struct mwifiex_adapter *adapter,
 		dev_dbg(adapter->dev, "info: do_rx_aggr: num of packets: %d\n",
 			card->mpa_rx.pkt_cnt);
 
+		if (card->supports_sdio_new_mode) {
+			int i;
+			u32 port_count;
+
+			for (i = 0, port_count = 0; i < card->max_ports; i++)
+				if (card->mpa_rx.ports & BIT(i))
+					port_count++;
+
+			/* Reading data from "start_port + 0" to "start_port +
+			 * port_count -1", so decrease the count by 1
+			 */
+			port_count--;
+			mport = (adapter->ioport | SDIO_MPA_ADDR_BASE |
+				 (port_count << 8)) + card->mpa_rx.start_port;
+		} else {
+			mport = (adapter->ioport | SDIO_MPA_ADDR_BASE |
+				 (card->mpa_rx.ports << 4)) +
+				 card->mpa_rx.start_port;
+		}
+
 		if (mwifiex_read_data_sync(adapter, card->mpa_rx.buf,
-					   card->mpa_rx.buf_len,
-					   (adapter->ioport | 0x1000 |
-					    (card->mpa_rx.ports << 4)) +
-					   card->mpa_rx.start_port, 1))
+					   card->mpa_rx.buf_len, mport, 1))
 			goto error;
 
 		curr_ptr = card->mpa_rx.buf;
@@ -1167,6 +1275,7 @@ error:
 static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 {
 	struct sdio_mmc_card *card = adapter->card;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
 	int ret = 0;
 	u8 sdio_ireg;
 	struct sk_buff *skb;
@@ -1175,6 +1284,8 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 	u32 rx_blocks;
 	u16 rx_len;
 	unsigned long flags;
+	u32 bitmap;
+	u8 cr;
 
 	spin_lock_irqsave(&adapter->int_lock, flags);
 	sdio_ireg = adapter->int_status;
@@ -1184,10 +1295,60 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 	if (!sdio_ireg)
 		return ret;
 
+	/* Following interrupt is only for SDIO new mode */
+	if (sdio_ireg & DN_LD_CMD_PORT_HOST_INT_STATUS && adapter->cmd_sent)
+		adapter->cmd_sent = false;
+
+	/* Following interrupt is only for SDIO new mode */
+	if (sdio_ireg & UP_LD_CMD_PORT_HOST_INT_STATUS) {
+		u32 pkt_type;
+
+		/* read the len of control packet */
+		rx_len = card->mp_regs[CMD_RD_LEN_1] << 8;
+		rx_len |= (u16) card->mp_regs[CMD_RD_LEN_0];
+		rx_blocks = DIV_ROUND_UP(rx_len, MWIFIEX_SDIO_BLOCK_SIZE);
+		if (rx_len <= INTF_HEADER_LEN ||
+		    (rx_blocks * MWIFIEX_SDIO_BLOCK_SIZE) >
+		     MWIFIEX_RX_DATA_BUF_SIZE)
+			return -1;
+		rx_len = (u16) (rx_blocks * MWIFIEX_SDIO_BLOCK_SIZE);
+
+		skb = dev_alloc_skb(rx_len);
+		if (!skb)
+			return -1;
+
+		skb_put(skb, rx_len);
+
+		if (mwifiex_sdio_card_to_host(adapter, &pkt_type, skb->data,
+					      skb->len, adapter->ioport |
+							CMD_PORT_SLCT)) {
+			dev_err(adapter->dev,
+				"%s: failed to card_to_host", __func__);
+			dev_kfree_skb_any(skb);
+			goto term_cmd;
+		}
+
+		if ((pkt_type != MWIFIEX_TYPE_CMD) &&
+		    (pkt_type != MWIFIEX_TYPE_EVENT))
+			dev_err(adapter->dev,
+				"%s:Received wrong packet on cmd port",
+				__func__);
+
+		mwifiex_decode_rx_packet(adapter, skb, pkt_type);
+	}
+
 	if (sdio_ireg & DN_LD_HOST_INT_STATUS) {
-		card->mp_wr_bitmap = ((u16) card->mp_regs[WR_BITMAP_U]) << 8;
-		card->mp_wr_bitmap |= (u16) card->mp_regs[WR_BITMAP_L];
-		dev_dbg(adapter->dev, "int: DNLD: wr_bitmap=0x%04x\n",
+		bitmap = (u32) card->mp_regs[reg->wr_bitmap_l];
+		bitmap |= ((u32) card->mp_regs[reg->wr_bitmap_u]) << 8;
+		if (card->supports_sdio_new_mode) {
+			bitmap |=
+				((u32) card->mp_regs[reg->wr_bitmap_1l]) << 16;
+			bitmap |=
+				((u32) card->mp_regs[reg->wr_bitmap_1u]) << 24;
+		}
+		card->mp_wr_bitmap = bitmap;
+
+		dev_dbg(adapter->dev, "int: DNLD: wr_bitmap=0x%x\n",
 			card->mp_wr_bitmap);
 		if (adapter->data_sent &&
 		    (card->mp_wr_bitmap & card->mp_data_port_mask)) {
@@ -1200,11 +1361,11 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 	/* As firmware will not generate download ready interrupt if the port
 	   updated is command port only, cmd_sent should be done for any SDIO
 	   interrupt. */
-	if (adapter->cmd_sent) {
+	if (card->has_control_mask && adapter->cmd_sent) {
 		/* Check if firmware has attach buffer at command port and
 		   update just that in wr_bit_map. */
 		card->mp_wr_bitmap |=
-			(u16) card->mp_regs[WR_BITMAP_L] & CTRL_PORT_MASK;
+			(u32) card->mp_regs[reg->wr_bitmap_l] & CTRL_PORT_MASK;
 		if (card->mp_wr_bitmap & CTRL_PORT_MASK)
 			adapter->cmd_sent = false;
 	}
@@ -1212,9 +1373,16 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 	dev_dbg(adapter->dev, "info: cmd_sent=%d data_sent=%d\n",
 		adapter->cmd_sent, adapter->data_sent);
 	if (sdio_ireg & UP_LD_HOST_INT_STATUS) {
-		card->mp_rd_bitmap = ((u16) card->mp_regs[RD_BITMAP_U]) << 8;
-		card->mp_rd_bitmap |= (u16) card->mp_regs[RD_BITMAP_L];
-		dev_dbg(adapter->dev, "int: UPLD: rd_bitmap=0x%04x\n",
+		bitmap = (u32) card->mp_regs[reg->rd_bitmap_l];
+		bitmap |= ((u32) card->mp_regs[reg->rd_bitmap_u]) << 8;
+		if (card->supports_sdio_new_mode) {
+			bitmap |=
+				((u32) card->mp_regs[reg->rd_bitmap_1l]) << 16;
+			bitmap |=
+				((u32) card->mp_regs[reg->rd_bitmap_1u]) << 24;
+		}
+		card->mp_rd_bitmap = bitmap;
+		dev_dbg(adapter->dev, "int: UPLD: rd_bitmap=0x%x\n",
 			card->mp_rd_bitmap);
 
 		while (true) {
@@ -1224,8 +1392,8 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 					"info: no more rd_port available\n");
 				break;
 			}
-			len_reg_l = RD_LEN_P0_L + (port << 1);
-			len_reg_u = RD_LEN_P0_U + (port << 1);
+			len_reg_l = reg->rd_len_p0_l + (port << 1);
+			len_reg_u = reg->rd_len_p0_u + (port << 1);
 			rx_len = ((u16) card->mp_regs[len_reg_u]) << 8;
 			rx_len |= (u16) card->mp_regs[len_reg_l];
 			dev_dbg(adapter->dev, "info: RX: port=%d rx_len=%u\n",
@@ -1257,37 +1425,33 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
 
 			if (mwifiex_sdio_card_to_host_mp_aggr(adapter, skb,
 							      port)) {
-				u32 cr = 0;
-
 				dev_err(adapter->dev, "card_to_host_mpa failed:"
 					" int status=%#x\n", sdio_ireg);
-				if (mwifiex_read_reg(adapter,
-						     CONFIGURATION_REG, &cr))
-					dev_err(adapter->dev,
-						"read CFG reg failed\n");
-
-				dev_dbg(adapter->dev,
-					"info: CFG reg val = %d\n", cr);
-				if (mwifiex_write_reg(adapter,
-						      CONFIGURATION_REG,
-						      (cr | 0x04)))
-					dev_err(adapter->dev,
-						"write CFG reg failed\n");
-
-				dev_dbg(adapter->dev, "info: write success\n");
-				if (mwifiex_read_reg(adapter,
-						     CONFIGURATION_REG, &cr))
-					dev_err(adapter->dev,
-						"read CFG reg failed\n");
-
-				dev_dbg(adapter->dev,
-					"info: CFG reg val =%x\n", cr);
-				return -1;
+				goto term_cmd;
 			}
 		}
 	}
 
 	return 0;
+
+term_cmd:
+	/* terminate cmd */
+	if (mwifiex_read_reg(adapter, CONFIGURATION_REG, &cr))
+		dev_err(adapter->dev, "read CFG reg failed\n");
+	else
+		dev_dbg(adapter->dev, "info: CFG reg val = %d\n", cr);
+
+	if (mwifiex_write_reg(adapter, CONFIGURATION_REG, (cr | 0x04)))
+		dev_err(adapter->dev, "write CFG reg failed\n");
+	else
+		dev_dbg(adapter->dev, "info: write success\n");
+
+	if (mwifiex_read_reg(adapter, CONFIGURATION_REG, &cr))
+		dev_err(adapter->dev, "read CFG reg failed\n");
+	else
+		dev_dbg(adapter->dev, "info: CFG reg val =%x\n", cr);
+
+	return -1;
 }
 
 /*
@@ -1305,7 +1469,7 @@ static int mwifiex_process_int_status(struct mwifiex_adapter *adapter)
  * and return.
  */
 static int mwifiex_host_to_card_mp_aggr(struct mwifiex_adapter *adapter,
-					u8 *payload, u32 pkt_len, u8 port,
+					u8 *payload, u32 pkt_len, u32 port,
 					u32 next_pkt_len)
 {
 	struct sdio_mmc_card *card = adapter->card;
@@ -1314,8 +1478,11 @@ static int mwifiex_host_to_card_mp_aggr(struct mwifiex_adapter *adapter,
 	s32 f_send_cur_buf = 0;
 	s32 f_precopy_cur_buf = 0;
 	s32 f_postcopy_cur_buf = 0;
+	u32 mport;
 
-	if ((!card->mpa_tx.enabled) || (port == CTRL_PORT)) {
+	if (!card->mpa_tx.enabled ||
+	    (card->has_control_mask && (port == CTRL_PORT)) ||
+	    (card->supports_sdio_new_mode && (port == CMD_PORT_SLCT))) {
 		dev_dbg(adapter->dev, "info: %s: tx aggregation disabled\n",
 			__func__);
 
@@ -1329,7 +1496,7 @@ static int mwifiex_host_to_card_mp_aggr(struct mwifiex_adapter *adapter,
 			__func__);
 
 		if (MP_TX_AGGR_IN_PROGRESS(card)) {
-			if (!MP_TX_AGGR_PORT_LIMIT_REACHED(card) &&
+			if (!mp_tx_aggr_port_limit_reached(card) &&
 			    MP_TX_AGGR_BUF_HAS_ROOM(card, pkt_len)) {
 				f_precopy_cur_buf = 1;
 
@@ -1342,7 +1509,7 @@ static int mwifiex_host_to_card_mp_aggr(struct mwifiex_adapter *adapter,
 				/* No room in Aggr buf, send it */
 				f_send_aggr_buf = 1;
 
-				if (MP_TX_AGGR_PORT_LIMIT_REACHED(card) ||
+				if (mp_tx_aggr_port_limit_reached(card) ||
 				    !(card->mp_wr_bitmap &
 				      (1 << card->curr_wr_port)))
 					f_send_cur_buf = 1;
@@ -1381,7 +1548,7 @@ static int mwifiex_host_to_card_mp_aggr(struct mwifiex_adapter *adapter,
 		MP_TX_AGGR_BUF_PUT(card, payload, pkt_len, port);
 
 		if (MP_TX_AGGR_PKT_LIMIT_REACHED(card) ||
-		    MP_TX_AGGR_PORT_LIMIT_REACHED(card))
+		    mp_tx_aggr_port_limit_reached(card))
 			/* No more pkts allowed in Aggr buf, send it */
 			f_send_aggr_buf = 1;
 	}
@@ -1390,11 +1557,28 @@ static int mwifiex_host_to_card_mp_aggr(struct mwifiex_adapter *adapter,
 		dev_dbg(adapter->dev, "data: %s: send aggr buffer: %d %d\n",
 			__func__,
 				card->mpa_tx.start_port, card->mpa_tx.ports);
+		if (card->supports_sdio_new_mode) {
+			u32 port_count;
+			int i;
+
+			for (i = 0, port_count = 0; i < card->max_ports; i++)
+				if (card->mpa_tx.ports & BIT(i))
+					port_count++;
+
+			/* Writing data from "start_port + 0" to "start_port +
+			 * port_count -1", so decrease the count by 1
+			 */
+			port_count--;
+			mport = (adapter->ioport | SDIO_MPA_ADDR_BASE |
+				 (port_count << 8)) + card->mpa_tx.start_port;
+		} else {
+			mport = (adapter->ioport | SDIO_MPA_ADDR_BASE |
+				 (card->mpa_tx.ports << 4)) +
+				 card->mpa_tx.start_port;
+		}
+
 		ret = mwifiex_write_data_to_card(adapter, card->mpa_tx.buf,
-						 card->mpa_tx.buf_len,
-						 (adapter->ioport | 0x1000 |
-						 (card->mpa_tx.ports << 4)) +
-						  card->mpa_tx.start_port);
+						 card->mpa_tx.buf_len, mport);
 
 		MP_TX_AGGR_BUF_RESET(card);
 	}
@@ -1434,7 +1618,7 @@ static int mwifiex_sdio_host_to_card(struct mwifiex_adapter *adapter,
 	int ret;
 	u32 buf_block_len;
 	u32 blk_size;
-	u8 port = CTRL_PORT;
+	u32 port = CTRL_PORT;
 	u8 *payload = (u8 *)skb->data;
 	u32 pkt_len = skb->len;
 
@@ -1465,6 +1649,9 @@ static int mwifiex_sdio_host_to_card(struct mwifiex_adapter *adapter,
 		    pkt_len > MWIFIEX_UPLD_SIZE)
 			dev_err(adapter->dev, "%s: payload=%p, nb=%d\n",
 				__func__, payload, pkt_len);
+
+		if (card->supports_sdio_new_mode)
+			port = CMD_PORT_SLCT;
 	}
 
 	/* Transfer data to card */
@@ -1586,18 +1773,7 @@ static int mwifiex_register_dev(struct mwifiex_adapter *adapter)
 
 	adapter->dev = &func->dev;
 
-	switch (func->device) {
-	case SDIO_DEVICE_ID_MARVELL_8786:
-		strcpy(adapter->fw_name, SD8786_DEFAULT_FW_NAME);
-		break;
-	case SDIO_DEVICE_ID_MARVELL_8797:
-		strcpy(adapter->fw_name, SD8797_DEFAULT_FW_NAME);
-		break;
-	case SDIO_DEVICE_ID_MARVELL_8787:
-	default:
-		strcpy(adapter->fw_name, SD8787_DEFAULT_FW_NAME);
-		break;
-	}
+	strcpy(adapter->fw_name, card->firmware);
 
 	return 0;
 
@@ -1626,8 +1802,9 @@ disable_func:
 static int mwifiex_init_sdio(struct mwifiex_adapter *adapter)
 {
 	struct sdio_mmc_card *card = adapter->card;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
 	int ret;
-	u32 sdio_ireg;
+	u8 sdio_ireg;
 
 	/*
 	 * Read the HOST_INT_STATUS_REG for ACK the first interrupt got
@@ -1645,30 +1822,35 @@ static int mwifiex_init_sdio(struct mwifiex_adapter *adapter)
 	/* Initialize SDIO variables in card */
 	card->mp_rd_bitmap = 0;
 	card->mp_wr_bitmap = 0;
-	card->curr_rd_port = 1;
-	card->curr_wr_port = 1;
+	card->curr_rd_port = reg->start_rd_port;
+	card->curr_wr_port = reg->start_wr_port;
 
-	card->mp_data_port_mask = DATA_PORT_MASK;
+	card->mp_data_port_mask = reg->data_port_mask;
 
 	card->mpa_tx.buf_len = 0;
 	card->mpa_tx.pkt_cnt = 0;
 	card->mpa_tx.start_port = 0;
 
 	card->mpa_tx.enabled = 1;
-	card->mpa_tx.pkt_aggr_limit = SDIO_MP_AGGR_DEF_PKT_LIMIT;
+	card->mpa_tx.pkt_aggr_limit = card->mp_agg_pkt_limit;
 
 	card->mpa_rx.buf_len = 0;
 	card->mpa_rx.pkt_cnt = 0;
 	card->mpa_rx.start_port = 0;
 
 	card->mpa_rx.enabled = 1;
-	card->mpa_rx.pkt_aggr_limit = SDIO_MP_AGGR_DEF_PKT_LIMIT;
+	card->mpa_rx.pkt_aggr_limit = card->mp_agg_pkt_limit;
 
 	/* Allocate buffers for SDIO MP-A */
-	card->mp_regs = kzalloc(MAX_MP_REGS, GFP_KERNEL);
+	card->mp_regs = kzalloc(reg->max_mp_regs, GFP_KERNEL);
 	if (!card->mp_regs)
 		return -ENOMEM;
 
+	/* Allocate skb pointer buffers */
+	card->mpa_rx.skb_arr = kzalloc((sizeof(void *)) *
+				       card->mp_agg_pkt_limit, GFP_KERNEL);
+	card->mpa_rx.len_arr = kzalloc(sizeof(*card->mpa_rx.len_arr) *
+				       card->mp_agg_pkt_limit, GFP_KERNEL);
 	ret = mwifiex_alloc_sdio_mpa_buffers(adapter,
 					     SDIO_MP_TX_AGGR_DEF_BUF_SIZE,
 					     SDIO_MP_RX_AGGR_DEF_BUF_SIZE);
@@ -1705,6 +1887,8 @@ static void mwifiex_cleanup_sdio(struct mwifiex_adapter *adapter)
 	struct sdio_mmc_card *card = adapter->card;
 
 	kfree(card->mp_regs);
+	kfree(card->mpa_rx.skb_arr);
+	kfree(card->mpa_rx.len_arr);
 	kfree(card->mpa_tx.buf);
 	kfree(card->mpa_rx.buf);
 }
@@ -1716,16 +1900,20 @@ static void
 mwifiex_update_mp_end_port(struct mwifiex_adapter *adapter, u16 port)
 {
 	struct sdio_mmc_card *card = adapter->card;
+	const struct mwifiex_sdio_card_reg *reg = card->reg;
 	int i;
 
 	card->mp_end_port = port;
 
-	card->mp_data_port_mask = DATA_PORT_MASK;
+	card->mp_data_port_mask = reg->data_port_mask;
 
-	for (i = 1; i <= MAX_PORT - card->mp_end_port; i++)
-		card->mp_data_port_mask &= ~(1 << (MAX_PORT - i));
+	if (reg->start_wr_port) {
+		for (i = 1; i <= card->max_ports - card->mp_end_port; i++)
+			card->mp_data_port_mask &=
+					~(1 << (card->max_ports - i));
+	}
 
-	card->curr_wr_port = 1;
+	card->curr_wr_port = reg->start_wr_port;
 
 	dev_dbg(adapter->dev, "cmd: mp_end_port %d, data port mask 0x%x\n",
 		port, card->mp_data_port_mask);
@@ -1831,3 +2019,4 @@ MODULE_LICENSE("GPL v2");
 MODULE_FIRMWARE(SD8786_DEFAULT_FW_NAME);
 MODULE_FIRMWARE(SD8787_DEFAULT_FW_NAME);
 MODULE_FIRMWARE(SD8797_DEFAULT_FW_NAME);
+MODULE_FIRMWARE(SD8897_DEFAULT_FW_NAME);
diff --git a/drivers/net/wireless/mwifiex/sdio.h b/drivers/net/wireless/mwifiex/sdio.h
index 8cc5468..6d51dfd 100644
--- a/drivers/net/wireless/mwifiex/sdio.h
+++ b/drivers/net/wireless/mwifiex/sdio.h
@@ -32,30 +32,37 @@
 #define SD8786_DEFAULT_FW_NAME "mrvl/sd8786_uapsta.bin"
 #define SD8787_DEFAULT_FW_NAME "mrvl/sd8787_uapsta.bin"
 #define SD8797_DEFAULT_FW_NAME "mrvl/sd8797_uapsta.bin"
+#define SD8897_DEFAULT_FW_NAME "mrvl/sd8897_uapsta.bin"
 
 #define BLOCK_MODE	1
 #define BYTE_MODE	0
 
 #define REG_PORT			0
-#define RD_BITMAP_L			0x04
-#define RD_BITMAP_U			0x05
-#define WR_BITMAP_L			0x06
-#define WR_BITMAP_U			0x07
-#define RD_LEN_P0_L			0x08
-#define RD_LEN_P0_U			0x09
 
 #define MWIFIEX_SDIO_IO_PORT_MASK		0xfffff
 
 #define MWIFIEX_SDIO_BYTE_MODE_MASK	0x80000000
 
+#define SDIO_MPA_ADDR_BASE		0x1000
 #define CTRL_PORT			0
 #define CTRL_PORT_MASK			0x0001
-#define DATA_PORT_MASK			0xfffe
 
-#define MAX_MP_REGS			64
-#define MAX_PORT			16
-
-#define SDIO_MP_AGGR_DEF_PKT_LIMIT	8
+#define CMD_PORT_UPLD_INT_MASK		(0x1U<<6)
+#define CMD_PORT_DNLD_INT_MASK		(0x1U<<7)
+#define HOST_TERM_CMD53			(0x1U << 2)
+#define REG_PORT			0
+#define MEM_PORT			0x10000
+#define CMD_RD_LEN_0			0xB4
+#define CMD_RD_LEN_1			0xB5
+#define CARD_CONFIG_2_1_REG             0xCD
+#define CMD53_NEW_MODE			(0x1U << 0)
+#define CMD_CONFIG_0			0xB8
+#define CMD_PORT_RD_LEN_EN		(0x1U << 2)
+#define CMD_CONFIG_1			0xB9
+#define CMD_PORT_AUTO_EN		(0x1U << 0)
+#define CMD_PORT_SLCT			0x8000
+#define UP_LD_CMD_PORT_HOST_INT_STATUS	(0x40U)
+#define DN_LD_CMD_PORT_HOST_INT_STATUS	(0x80U)
 
 #define SDIO_MP_TX_AGGR_DEF_BUF_SIZE        (8192)	/* 8K */
 
@@ -75,14 +82,8 @@
 
 /* Host Control Registers : Configuration */
 #define CONFIGURATION_REG		0x00
-/* Host Control Registers : Host without Command 53 finish host*/
-#define HOST_TO_CARD_EVENT       (0x1U << 3)
-/* Host Control Registers : Host without Command 53 finish host */
-#define HOST_WO_CMD53_FINISH_HOST	(0x1U << 2)
 /* Host Control Registers : Host power up */
 #define HOST_POWER_UP			(0x1U << 1)
-/* Host Control Registers : Host power down */
-#define HOST_POWER_DOWN			(0x1U << 0)
 
 /* Host Control Registers : Host interrupt mask */
 #define HOST_INT_MASK_REG		0x02
@@ -90,8 +91,7 @@
 #define UP_LD_HOST_INT_MASK		(0x1U)
 /* Host Control Registers : Download host interrupt mask */
 #define DN_LD_HOST_INT_MASK		(0x2U)
-/* Enable Host interrupt mask */
-#define HOST_INT_ENABLE	(UP_LD_HOST_INT_MASK | DN_LD_HOST_INT_MASK)
+
 /* Disable Host interrupt mask */
 #define	HOST_INT_DISABLE		0xff
 
@@ -104,74 +104,15 @@
 
 /* Host Control Registers : Host interrupt RSR */
 #define HOST_INT_RSR_REG		0x01
-/* Host Control Registers : Upload host interrupt RSR */
-#define UP_LD_HOST_INT_RSR		(0x1U)
-#define SDIO_INT_MASK			0x3F
 
 /* Host Control Registers : Host interrupt status */
 #define HOST_INT_STATUS_REG		0x28
-/* Host Control Registers : Upload CRC error */
-#define UP_LD_CRC_ERR			(0x1U << 2)
-/* Host Control Registers : Upload restart */
-#define UP_LD_RESTART                   (0x1U << 1)
-/* Host Control Registers : Download restart */
-#define DN_LD_RESTART                   (0x1U << 0)
-
-/* Card Control Registers : Card status register */
-#define CARD_STATUS_REG                 0x30
+
 /* Card Control Registers : Card I/O ready */
 #define CARD_IO_READY                   (0x1U << 3)
-/* Card Control Registers : CIS card ready */
-#define CIS_CARD_RDY                    (0x1U << 2)
-/* Card Control Registers : Upload card ready */
-#define UP_LD_CARD_RDY                  (0x1U << 1)
 /* Card Control Registers : Download card ready */
 #define DN_LD_CARD_RDY                  (0x1U << 0)
 
-/* Card Control Registers : Host interrupt mask register */
-#define HOST_INTERRUPT_MASK_REG         0x34
-/* Card Control Registers : Host power interrupt mask */
-#define HOST_POWER_INT_MASK             (0x1U << 3)
-/* Card Control Registers : Abort card interrupt mask */
-#define ABORT_CARD_INT_MASK             (0x1U << 2)
-/* Card Control Registers : Upload card interrupt mask */
-#define UP_LD_CARD_INT_MASK             (0x1U << 1)
-/* Card Control Registers : Download card interrupt mask */
-#define DN_LD_CARD_INT_MASK             (0x1U << 0)
-
-/* Card Control Registers : Card interrupt status register */
-#define CARD_INTERRUPT_STATUS_REG       0x38
-/* Card Control Registers : Power up interrupt */
-#define POWER_UP_INT                    (0x1U << 4)
-/* Card Control Registers : Power down interrupt */
-#define POWER_DOWN_INT                  (0x1U << 3)
-
-/* Card Control Registers : Card interrupt RSR register */
-#define CARD_INTERRUPT_RSR_REG          0x3c
-/* Card Control Registers : Power up RSR */
-#define POWER_UP_RSR                    (0x1U << 4)
-/* Card Control Registers : Power down RSR */
-#define POWER_DOWN_RSR                  (0x1U << 3)
-
-/* Card Control Registers : Miscellaneous Configuration Register */
-#define CARD_MISC_CFG_REG               0x6C
-
-/* Host F1 read base 0 */
-#define HOST_F1_RD_BASE_0		0x0040
-/* Host F1 read base 1 */
-#define HOST_F1_RD_BASE_1		0x0041
-/* Host F1 card ready */
-#define HOST_F1_CARD_RDY		0x0020
-
-/* Firmware status 0 register */
-#define CARD_FW_STATUS0_REG		0x60
-/* Firmware status 1 register */
-#define CARD_FW_STATUS1_REG		0x61
-/* Rx length register */
-#define CARD_RX_LEN_REG			0x62
-/* Rx unit register */
-#define CARD_RX_UNIT_REG		0x63
-
 /* Max retry number of CMD53 write */
 #define MAX_WRITE_IOMEM_RETRY		2
 
@@ -192,7 +133,8 @@
 	if (a->mpa_tx.start_port <= port)				\
 		a->mpa_tx.ports |= (1<<(a->mpa_tx.pkt_cnt));		\
 	else								\
-		a->mpa_tx.ports |= (1<<(a->mpa_tx.pkt_cnt+1+(MAX_PORT -	\
+		a->mpa_tx.ports |= (1<<(a->mpa_tx.pkt_cnt+1+		\
+						(a->max_ports -	\
 						a->mp_end_port)));	\
 	a->mpa_tx.pkt_cnt++;						\
 } while (0)
@@ -201,12 +143,6 @@
 #define MP_TX_AGGR_PKT_LIMIT_REACHED(a)					\
 			(a->mpa_tx.pkt_cnt == a->mpa_tx.pkt_aggr_limit)
 
-/* SDIO Tx aggregation port limit ? */
-#define MP_TX_AGGR_PORT_LIMIT_REACHED(a) ((a->curr_wr_port <		\
-			a->mpa_tx.start_port) && (((MAX_PORT -		\
-			a->mpa_tx.start_port) + a->curr_wr_port) >=	\
-				SDIO_MP_AGGR_DEF_PKT_LIMIT))
-
 /* Reset SDIO Tx aggregation buffer parameters */
 #define MP_TX_AGGR_BUF_RESET(a) do {					\
 	a->mpa_tx.pkt_cnt = 0;						\
@@ -219,12 +155,6 @@
 #define MP_RX_AGGR_PKT_LIMIT_REACHED(a)					\
 			(a->mpa_rx.pkt_cnt == a->mpa_rx.pkt_aggr_limit)
 
-/* SDIO Tx aggregation port limit ? */
-#define MP_RX_AGGR_PORT_LIMIT_REACHED(a) ((a->curr_rd_port <		\
-			a->mpa_rx.start_port) && (((MAX_PORT -		\
-			a->mpa_rx.start_port) + a->curr_rd_port) >=	\
-			SDIO_MP_AGGR_DEF_PKT_LIMIT))
-
 /* SDIO Rx aggregation in progress ? */
 #define MP_RX_AGGR_IN_PROGRESS(a) (a->mpa_rx.pkt_cnt > 0)
 
@@ -232,20 +162,6 @@
 #define MP_RX_AGGR_BUF_HAS_ROOM(a, rx_len)				\
 			((a->mpa_rx.buf_len+rx_len) <= a->mpa_rx.buf_size)
 
-/* Prepare to copy current packet from card to SDIO Rx aggregation buffer */
-#define MP_RX_AGGR_SETUP(a, skb, port) do {				\
-	a->mpa_rx.buf_len += skb->len;					\
-	if (!a->mpa_rx.pkt_cnt)						\
-		a->mpa_rx.start_port = port;				\
-	if (a->mpa_rx.start_port <= port)				\
-		a->mpa_rx.ports |= (1<<(a->mpa_rx.pkt_cnt));		\
-	else								\
-		a->mpa_rx.ports |= (1<<(a->mpa_rx.pkt_cnt+1));		\
-	a->mpa_rx.skb_arr[a->mpa_rx.pkt_cnt] = skb;			\
-	a->mpa_rx.len_arr[a->mpa_rx.pkt_cnt] = skb->len;		\
-	a->mpa_rx.pkt_cnt++;						\
-} while (0)
-
 /* Reset SDIO Rx aggregation buffer parameters */
 #define MP_RX_AGGR_BUF_RESET(a) do {					\
 	a->mpa_rx.pkt_cnt = 0;						\
@@ -254,14 +170,13 @@
 	a->mpa_rx.start_port = 0;					\
 } while (0)
 
-
 /* data structure for SDIO MPA TX */
 struct mwifiex_sdio_mpa_tx {
 	/* multiport tx aggregation buffer pointer */
 	u8 *buf;
 	u32 buf_len;
 	u32 pkt_cnt;
-	u16 ports;
+	u32 ports;
 	u16 start_port;
 	u8 enabled;
 	u32 buf_size;
@@ -272,11 +187,11 @@ struct mwifiex_sdio_mpa_rx {
 	u8 *buf;
 	u32 buf_len;
 	u32 pkt_cnt;
-	u16 ports;
+	u32 ports;
 	u16 start_port;
 
-	struct sk_buff *skb_arr[SDIO_MP_AGGR_DEF_PKT_LIMIT];
-	u32 len_arr[SDIO_MP_AGGR_DEF_PKT_LIMIT];
+	struct sk_buff **skb_arr;
+	u32 *len_arr;
 
 	u8 enabled;
 	u32 buf_size;
@@ -286,15 +201,47 @@ struct mwifiex_sdio_mpa_rx {
 int mwifiex_bus_register(void);
 void mwifiex_bus_unregister(void);
 
+struct mwifiex_sdio_card_reg {
+	u8 start_rd_port;
+	u8 start_wr_port;
+	u8 base_0_reg;
+	u8 base_1_reg;
+	u8 poll_reg;
+	u8 host_int_enable;
+	u8 status_reg_0;
+	u8 status_reg_1;
+	u8 sdio_int_mask;
+	u32 data_port_mask;
+	u8 max_mp_regs;
+	u8 rd_bitmap_l;
+	u8 rd_bitmap_u;
+	u8 rd_bitmap_1l;
+	u8 rd_bitmap_1u;
+	u8 wr_bitmap_l;
+	u8 wr_bitmap_u;
+	u8 wr_bitmap_1l;
+	u8 wr_bitmap_1u;
+	u8 rd_len_p0_l;
+	u8 rd_len_p0_u;
+	u8 card_misc_cfg_reg;
+};
+
 struct sdio_mmc_card {
 	struct sdio_func *func;
 	struct mwifiex_adapter *adapter;
 
-	u16 mp_rd_bitmap;
-	u16 mp_wr_bitmap;
+	const char *firmware;
+	const struct mwifiex_sdio_card_reg *reg;
+	u8 max_ports;
+	u8 mp_agg_pkt_limit;
+	bool supports_sdio_new_mode;
+	bool has_control_mask;
+
+	u32 mp_rd_bitmap;
+	u32 mp_wr_bitmap;
 
 	u16 mp_end_port;
-	u16 mp_data_port_mask;
+	u32 mp_data_port_mask;
 
 	u8 curr_rd_port;
 	u8 curr_wr_port;
@@ -305,6 +252,98 @@ struct sdio_mmc_card {
 	struct mwifiex_sdio_mpa_rx mpa_rx;
 };
 
+struct mwifiex_sdio_device {
+	const char *firmware;
+	const struct mwifiex_sdio_card_reg *reg;
+	u8 max_ports;
+	u8 mp_agg_pkt_limit;
+	bool supports_sdio_new_mode;
+	bool has_control_mask;
+};
+
+static const struct mwifiex_sdio_card_reg mwifiex_reg_sd87xx = {
+	.start_rd_port = 1,
+	.start_wr_port = 1,
+	.base_0_reg = 0x0040,
+	.base_1_reg = 0x0041,
+	.poll_reg = 0x30,
+	.host_int_enable = UP_LD_HOST_INT_MASK | DN_LD_HOST_INT_MASK,
+	.status_reg_0 = 0x60,
+	.status_reg_1 = 0x61,
+	.sdio_int_mask = 0x3f,
+	.data_port_mask = 0x0000fffe,
+	.max_mp_regs = 64,
+	.rd_bitmap_l = 0x04,
+	.rd_bitmap_u = 0x05,
+	.wr_bitmap_l = 0x06,
+	.wr_bitmap_u = 0x07,
+	.rd_len_p0_l = 0x08,
+	.rd_len_p0_u = 0x09,
+	.card_misc_cfg_reg = 0x6c,
+};
+
+static const struct mwifiex_sdio_card_reg mwifiex_reg_sd8897 = {
+	.start_rd_port = 0,
+	.start_wr_port = 0,
+	.base_0_reg = 0x60,
+	.base_1_reg = 0x61,
+	.poll_reg = 0x50,
+	.host_int_enable = UP_LD_HOST_INT_MASK | DN_LD_HOST_INT_MASK |
+			CMD_PORT_UPLD_INT_MASK | CMD_PORT_DNLD_INT_MASK,
+	.status_reg_0 = 0xc0,
+	.status_reg_1 = 0xc1,
+	.sdio_int_mask = 0xff,
+	.data_port_mask = 0xffffffff,
+	.max_mp_regs = 184,
+	.rd_bitmap_l = 0x04,
+	.rd_bitmap_u = 0x05,
+	.rd_bitmap_1l = 0x06,
+	.rd_bitmap_1u = 0x07,
+	.wr_bitmap_l = 0x08,
+	.wr_bitmap_u = 0x09,
+	.wr_bitmap_1l = 0x0a,
+	.wr_bitmap_1u = 0x0b,
+	.rd_len_p0_l = 0x0c,
+	.rd_len_p0_u = 0x0d,
+	.card_misc_cfg_reg = 0xcc,
+};
+
+static const struct mwifiex_sdio_device mwifiex_sdio_sd8786 = {
+	.firmware = SD8786_DEFAULT_FW_NAME,
+	.reg = &mwifiex_reg_sd87xx,
+	.max_ports = 16,
+	.mp_agg_pkt_limit = 8,
+	.supports_sdio_new_mode = false,
+	.has_control_mask = true,
+};
+
+static const struct mwifiex_sdio_device mwifiex_sdio_sd8787 = {
+	.firmware = SD8787_DEFAULT_FW_NAME,
+	.reg = &mwifiex_reg_sd87xx,
+	.max_ports = 16,
+	.mp_agg_pkt_limit = 8,
+	.supports_sdio_new_mode = false,
+	.has_control_mask = true,
+};
+
+static const struct mwifiex_sdio_device mwifiex_sdio_sd8797 = {
+	.firmware = SD8797_DEFAULT_FW_NAME,
+	.reg = &mwifiex_reg_sd87xx,
+	.max_ports = 16,
+	.mp_agg_pkt_limit = 8,
+	.supports_sdio_new_mode = false,
+	.has_control_mask = true,
+};
+
+static const struct mwifiex_sdio_device mwifiex_sdio_sd8897 = {
+	.firmware = SD8897_DEFAULT_FW_NAME,
+	.reg = &mwifiex_reg_sd8897,
+	.max_ports = 32,
+	.mp_agg_pkt_limit = 16,
+	.supports_sdio_new_mode = true,
+	.has_control_mask = false,
+};
+
 /*
  * .cmdrsp_complete handler
  */
@@ -325,4 +364,77 @@ static inline int mwifiex_sdio_event_complete(struct mwifiex_adapter *adapter,
 	return 0;
 }
 
+static inline bool
+mp_rx_aggr_port_limit_reached(struct sdio_mmc_card *card)
+{
+	u8 tmp;
+
+	if (card->curr_rd_port < card->mpa_rx.start_port) {
+		if (card->supports_sdio_new_mode)
+			tmp = card->mp_end_port >> 1;
+		else
+			tmp = card->mp_agg_pkt_limit;
+
+		if (((card->max_ports - card->mpa_rx.start_port) +
+		    card->curr_rd_port) >= tmp)
+			return true;
+	}
+
+	if (!card->supports_sdio_new_mode)
+		return false;
+
+	if ((card->curr_rd_port - card->mpa_rx.start_port) >=
+	    (card->mp_end_port >> 1))
+		return true;
+
+	return false;
+}
+
+static inline bool
+mp_tx_aggr_port_limit_reached(struct sdio_mmc_card *card)
+{
+	u16 tmp;
+
+	if (card->curr_wr_port < card->mpa_tx.start_port) {
+		if (card->supports_sdio_new_mode)
+			tmp = card->mp_end_port >> 1;
+		else
+			tmp = card->mp_agg_pkt_limit;
+
+		if (((card->max_ports - card->mpa_tx.start_port) +
+		    card->curr_wr_port) >= tmp)
+			return true;
+	}
+
+	if (!card->supports_sdio_new_mode)
+		return false;
+
+	if ((card->curr_wr_port - card->mpa_tx.start_port) >=
+	    (card->mp_end_port >> 1))
+		return true;
+
+	return false;
+}
+
+/* Prepare to copy current packet from card to SDIO Rx aggregation buffer */
+static inline void mp_rx_aggr_setup(struct sdio_mmc_card *card,
+				    struct sk_buff *skb, u8 port)
+{
+	card->mpa_rx.buf_len += skb->len;
+
+	if (!card->mpa_rx.pkt_cnt)
+		card->mpa_rx.start_port = port;
+
+	if (card->supports_sdio_new_mode) {
+		card->mpa_rx.ports |= (1 << port);
+	} else {
+		if (card->mpa_rx.start_port <= port)
+			card->mpa_rx.ports |= 1 << (card->mpa_rx.pkt_cnt);
+		else
+			card->mpa_rx.ports |= 1 << (card->mpa_rx.pkt_cnt + 1);
+	}
+	card->mpa_rx.skb_arr[card->mpa_rx.pkt_cnt] = skb;
+	card->mpa_rx.len_arr[card->mpa_rx.pkt_cnt] = skb->len;
+	card->mpa_rx.pkt_cnt++;
+}
 #endif /* _MWIFIEX_SDIO_H */
diff --git a/drivers/net/wireless/mwifiex/sta_cmd.c b/drivers/net/wireless/mwifiex/sta_cmd.c
index b193e25..8ece485 100644
--- a/drivers/net/wireless/mwifiex/sta_cmd.c
+++ b/drivers/net/wireless/mwifiex/sta_cmd.c
@@ -1134,6 +1134,55 @@ mwifiex_cmd_mef_cfg(struct mwifiex_private *priv,
 	return 0;
 }
 
+/* This function parse cal data from ASCII to hex */
+static u32 mwifiex_parse_cal_cfg(u8 *src, size_t len, u8 *dst)
+{
+	u8 *s = src, *d = dst;
+
+	while (s - src < len) {
+		if (*s && (isspace(*s) || *s == '\t')) {
+			s++;
+			continue;
+		}
+		if (isxdigit(*s)) {
+			*d++ = simple_strtol(s, NULL, 16);
+			s += 2;
+		} else {
+			s++;
+		}
+	}
+
+	return d - dst;
+}
+
+/* This function prepares command of set_cfg_data. */
+static int mwifiex_cmd_cfg_data(struct mwifiex_private *priv,
+				struct host_cmd_ds_command *cmd,
+				u16 cmd_action)
+{
+	struct host_cmd_ds_802_11_cfg_data *cfg_data = &cmd->params.cfg_data;
+	struct mwifiex_adapter *adapter = priv->adapter;
+	u32 len, cal_data_offset;
+	u8 *tmp_cmd = (u8 *)cmd;
+
+	cal_data_offset = S_DS_GEN + sizeof(*cfg_data);
+	if ((adapter->cal_data->data) && (adapter->cal_data->size > 0))
+		len = mwifiex_parse_cal_cfg((u8 *)adapter->cal_data->data,
+					    adapter->cal_data->size,
+					    (u8 *)(tmp_cmd + cal_data_offset));
+	else
+		return -1;
+
+	cfg_data->action = cpu_to_le16(cmd_action);
+	cfg_data->type = cpu_to_le16(CFG_DATA_TYPE_CAL);
+	cfg_data->data_len = cpu_to_le16(len);
+
+	cmd->command = cpu_to_le16(HostCmd_CMD_CFG_DATA);
+	cmd->size = cpu_to_le16(S_DS_GEN + sizeof(*cfg_data) + len);
+
+	return 0;
+}
+
 /*
  * This function prepares the commands before sending them to the firmware.
  *
@@ -1152,6 +1201,9 @@ int mwifiex_sta_prepare_cmd(struct mwifiex_private *priv, uint16_t cmd_no,
 	case HostCmd_CMD_GET_HW_SPEC:
 		ret = mwifiex_cmd_get_hw_spec(priv, cmd_ptr);
 		break;
+	case HostCmd_CMD_CFG_DATA:
+		ret = mwifiex_cmd_cfg_data(priv, cmd_ptr, cmd_action);
+		break;
 	case HostCmd_CMD_MAC_CONTROL:
 		ret = mwifiex_cmd_mac_control(priv, cmd_ptr, cmd_action,
 					      data_buf);
@@ -1384,6 +1436,7 @@ int mwifiex_sta_prepare_cmd(struct mwifiex_private *priv, uint16_t cmd_no,
  */
 int mwifiex_sta_init_cmd(struct mwifiex_private *priv, u8 first_sta)
 {
+	struct mwifiex_adapter *adapter = priv->adapter;
 	int ret;
 	u16 enable = true;
 	struct mwifiex_ds_11n_amsdu_aggr_ctrl amsdu_aggr_ctrl;
@@ -1404,6 +1457,15 @@ int mwifiex_sta_init_cmd(struct mwifiex_private *priv, u8 first_sta)
 					    HostCmd_ACT_GEN_SET, 0, NULL);
 		if (ret)
 			return -1;
+
+		/* Download calibration data to firmware */
+		if (adapter->cal_data) {
+			ret = mwifiex_send_cmd_sync(priv, HostCmd_CMD_CFG_DATA,
+						HostCmd_ACT_GEN_SET, 0, NULL);
+			if (ret)
+				return -1;
+		}
+
 		/* Read MAC address from HW */
 		ret = mwifiex_send_cmd_sync(priv, HostCmd_CMD_GET_HW_SPEC,
 					    HostCmd_ACT_GEN_GET, 0, NULL);
diff --git a/drivers/net/wireless/mwifiex/sta_cmdresp.c b/drivers/net/wireless/mwifiex/sta_cmdresp.c
index 9f990e1..d85df15 100644
--- a/drivers/net/wireless/mwifiex/sta_cmdresp.c
+++ b/drivers/net/wireless/mwifiex/sta_cmdresp.c
@@ -818,6 +818,18 @@ static int mwifiex_ret_subsc_evt(struct mwifiex_private *priv,
 	return 0;
 }
 
+/* This function handles the command response of set_cfg_data */
+static int mwifiex_ret_cfg_data(struct mwifiex_private *priv,
+				struct host_cmd_ds_command *resp)
+{
+	if (resp->result != HostCmd_RESULT_OK) {
+		dev_err(priv->adapter->dev, "Cal data cmd resp failed\n");
+		return -1;
+	}
+
+	return 0;
+}
+
 /*
  * This function handles the command responses.
  *
@@ -841,6 +853,9 @@ int mwifiex_process_sta_cmdresp(struct mwifiex_private *priv, u16 cmdresp_no,
 	case HostCmd_CMD_GET_HW_SPEC:
 		ret = mwifiex_ret_get_hw_spec(priv, resp);
 		break;
+	case HostCmd_CMD_CFG_DATA:
+		ret = mwifiex_ret_cfg_data(priv, resp);
+		break;
 	case HostCmd_CMD_MAC_CONTROL:
 		break;
 	case HostCmd_CMD_802_11_MAC_ADDRESS:
@@ -978,6 +993,8 @@ int mwifiex_process_sta_cmdresp(struct mwifiex_private *priv, u16 cmdresp_no,
 	case HostCmd_CMD_UAP_BSS_STOP:
 		priv->bss_started = 0;
 		break;
+	case HostCmd_CMD_UAP_STA_DEAUTH:
+		break;
 	case HostCmd_CMD_MEF_CFG:
 		break;
 	default:
diff --git a/drivers/net/wireless/mwifiex/sta_event.c b/drivers/net/wireless/mwifiex/sta_event.c
index 41aafc7..ea265ec 100644
--- a/drivers/net/wireless/mwifiex/sta_event.c
+++ b/drivers/net/wireless/mwifiex/sta_event.c
@@ -427,6 +427,17 @@ int mwifiex_process_sta_event(struct mwifiex_private *priv)
 
 		break;
 
+	case EVENT_CHANNEL_SWITCH_ANN:
+		dev_dbg(adapter->dev, "event: Channel Switch Announcement\n");
+		priv->csa_expire_time =
+				jiffies + msecs_to_jiffies(DFS_CHAN_MOVE_TIME);
+		priv->csa_chan = priv->curr_bss_params.bss_descriptor.channel;
+		ret = mwifiex_send_cmd_async(priv,
+			HostCmd_CMD_802_11_DEAUTHENTICATE,
+			HostCmd_ACT_GEN_SET, 0,
+			priv->curr_bss_params.bss_descriptor.mac_address);
+		break;
+
 	default:
 		dev_dbg(adapter->dev, "event: unknown event id: %#x\n",
 			eventcause);
diff --git a/drivers/net/wireless/mwifiex/sta_ioctl.c b/drivers/net/wireless/mwifiex/sta_ioctl.c
index 1a8a19d..206c3e0 100644
--- a/drivers/net/wireless/mwifiex/sta_ioctl.c
+++ b/drivers/net/wireless/mwifiex/sta_ioctl.c
@@ -104,16 +104,14 @@ int mwifiex_request_set_multicast_list(struct mwifiex_private *priv,
 		} else {
 			priv->curr_pkt_filter &=
 				~HostCmd_ACT_MAC_ALL_MULTICAST_ENABLE;
-			if (mcast_list->num_multicast_addr) {
-				dev_dbg(priv->adapter->dev,
-					"info: Set multicast list=%d\n",
-				       mcast_list->num_multicast_addr);
-				/* Send multicast addresses to firmware */
-				ret = mwifiex_send_cmd_async(priv,
-					HostCmd_CMD_MAC_MULTICAST_ADR,
-					HostCmd_ACT_GEN_SET, 0,
-					mcast_list);
-			}
+			dev_dbg(priv->adapter->dev,
+				"info: Set multicast list=%d\n",
+				mcast_list->num_multicast_addr);
+			/* Send multicast addresses to firmware */
+			ret = mwifiex_send_cmd_async(priv,
+				HostCmd_CMD_MAC_MULTICAST_ADR,
+				HostCmd_ACT_GEN_SET, 0,
+				mcast_list);
 		}
 	}
 	dev_dbg(priv->adapter->dev,
@@ -180,6 +178,9 @@ int mwifiex_fill_new_bss_desc(struct mwifiex_private *priv,
 	 */
 	bss_desc->disable_11ac = true;
 
+	if (bss_desc->cap_info_bitmap & WLAN_CAPABILITY_SPECTRUM_MGMT)
+		bss_desc->sensed_11h = true;
+
 	return mwifiex_update_bss_desc_with_ie(priv->adapter, bss_desc);
 }
 
@@ -257,30 +258,37 @@ int mwifiex_bss_start(struct mwifiex_private *priv, struct cfg80211_bss *bss,
 	}
 
 	if (priv->bss_mode == NL80211_IFTYPE_STATION) {
+		u8 config_bands;
+
 		/* Infra mode */
 		ret = mwifiex_deauthenticate(priv, NULL);
 		if (ret)
 			goto done;
 
-		if (bss_desc) {
-			u8 config_bands = 0;
+		if (!bss_desc)
+			return -1;
 
-			if (mwifiex_band_to_radio_type((u8) bss_desc->bss_band)
-			    == HostCmd_SCAN_RADIO_TYPE_BG)
-				config_bands = BAND_B | BAND_G | BAND_GN |
-					       BAND_GAC;
-			else
-				config_bands = BAND_A | BAND_AN | BAND_AAC;
+		if (mwifiex_band_to_radio_type(bss_desc->bss_band) ==
+						HostCmd_SCAN_RADIO_TYPE_BG)
+			config_bands = BAND_B | BAND_G | BAND_GN | BAND_GAC;
+		else
+			config_bands = BAND_A | BAND_AN | BAND_AAC;
 
-			if (!((config_bands | adapter->fw_bands) &
-			      ~adapter->fw_bands))
-				adapter->config_bands = config_bands;
-		}
+		if (!((config_bands | adapter->fw_bands) & ~adapter->fw_bands))
+			adapter->config_bands = config_bands;
 
 		ret = mwifiex_check_network_compatibility(priv, bss_desc);
 		if (ret)
 			goto done;
 
+		if (mwifiex_11h_get_csa_closed_channel(priv) ==
+							(u8)bss_desc->channel) {
+			dev_err(adapter->dev,
+				"Attempt to reconnect on csa closed chan(%d)\n",
+				bss_desc->channel);
+			goto done;
+		}
+
 		dev_dbg(adapter->dev, "info: SSID found in scan list ... "
 				      "associating...\n");
 
diff --git a/drivers/net/wireless/mwifiex/uap_cmd.c b/drivers/net/wireless/mwifiex/uap_cmd.c
index b04b1db..2de882d 100644
--- a/drivers/net/wireless/mwifiex/uap_cmd.c
+++ b/drivers/net/wireless/mwifiex/uap_cmd.c
@@ -689,6 +689,23 @@ mwifiex_cmd_uap_sys_config(struct host_cmd_ds_command *cmd, u16 cmd_action,
 	return 0;
 }
 
+/* This function prepares AP specific deauth command with mac supplied in
+ * function parameter.
+ */
+static int mwifiex_cmd_uap_sta_deauth(struct mwifiex_private *priv,
+				      struct host_cmd_ds_command *cmd, u8 *mac)
+{
+	struct host_cmd_ds_sta_deauth *sta_deauth = &cmd->params.sta_deauth;
+
+	cmd->command = cpu_to_le16(HostCmd_CMD_UAP_STA_DEAUTH);
+	memcpy(sta_deauth->mac, mac, ETH_ALEN);
+	sta_deauth->reason = cpu_to_le16(WLAN_REASON_DEAUTH_LEAVING);
+
+	cmd->size = cpu_to_le16(sizeof(struct host_cmd_ds_sta_deauth) +
+				S_DS_GEN);
+	return 0;
+}
+
 /* This function prepares the AP specific commands before sending them
  * to the firmware.
  * This is a generic function which calls specific command preparation
@@ -710,6 +727,10 @@ int mwifiex_uap_prepare_cmd(struct mwifiex_private *priv, u16 cmd_no,
 		cmd->command = cpu_to_le16(cmd_no);
 		cmd->size = cpu_to_le16(S_DS_GEN);
 		break;
+	case HostCmd_CMD_UAP_STA_DEAUTH:
+		if (mwifiex_cmd_uap_sta_deauth(priv, cmd, data_buf))
+			return -1;
+		break;
 	default:
 		dev_err(priv->adapter->dev,
 			"PREP_CMD: unknown cmd %#x\n", cmd_no);
diff --git a/drivers/net/wireless/mwifiex/uap_event.c b/drivers/net/wireless/mwifiex/uap_event.c
index 21c640d..7180665 100644
--- a/drivers/net/wireless/mwifiex/uap_event.c
+++ b/drivers/net/wireless/mwifiex/uap_event.c
@@ -107,18 +107,15 @@ mwifiex_set_sta_ht_cap(struct mwifiex_private *priv, const u8 *ies,
  */
 static void mwifiex_del_sta_entry(struct mwifiex_private *priv, u8 *mac)
 {
-	struct mwifiex_sta_node *node, *tmp;
+	struct mwifiex_sta_node *node;
 	unsigned long flags;
 
 	spin_lock_irqsave(&priv->sta_list_spinlock, flags);
 
 	node = mwifiex_get_sta_entry(priv, mac);
 	if (node) {
-		list_for_each_entry_safe(node, tmp, &priv->sta_list,
-					 list) {
-			list_del(&node->list);
-			kfree(node);
-		}
+		list_del(&node->list);
+		kfree(node);
 	}
 
 	spin_unlock_irqrestore(&priv->sta_list_spinlock, flags);
@@ -295,3 +292,19 @@ int mwifiex_process_uap_event(struct mwifiex_private *priv)
 
 	return 0;
 }
+
+/* This function deletes station entry from associated station list.
+ * Also if both AP and STA are 11n enabled, RxReorder tables and TxBA stream
+ * tables created for this station are deleted.
+ */
+void mwifiex_uap_del_sta_data(struct mwifiex_private *priv,
+			      struct mwifiex_sta_node *node)
+{
+	if (priv->ap_11n_enabled && node->is_11n_enabled) {
+		mwifiex_11n_del_rx_reorder_tbl_by_ta(priv, node->mac_addr);
+		mwifiex_del_tx_ba_stream_tbl_by_ra(priv, node->mac_addr);
+	}
+	mwifiex_del_sta_entry(priv, node->mac_addr);
+
+	return;
+}
diff --git a/drivers/net/wireless/mwifiex/wmm.c b/drivers/net/wireless/mwifiex/wmm.c
index 4be3d33..944e884 100644
--- a/drivers/net/wireless/mwifiex/wmm.c
+++ b/drivers/net/wireless/mwifiex/wmm.c
@@ -37,6 +37,9 @@
 /* Offset for TOS field in the IP header */
 #define IPTOS_OFFSET 5
 
+static bool enable_tx_amsdu;
+module_param(enable_tx_amsdu, bool, 0644);
+
 /* WMM information IE */
 static const u8 wmm_info_ie[] = { WLAN_EID_VENDOR_SPECIFIC, 0x07,
 	0x00, 0x50, 0xf2, 0x02,
@@ -1233,7 +1236,7 @@ mwifiex_dequeue_tx_packet(struct mwifiex_adapter *adapter)
 				mwifiex_send_delba(priv, tid_del, ra, 1);
 			}
 		}
-		if (mwifiex_is_amsdu_allowed(priv, tid) &&
+		if (enable_tx_amsdu && mwifiex_is_amsdu_allowed(priv, tid) &&
 		    mwifiex_is_11n_aggragation_possible(priv, ptr,
 							adapter->tx_buf_size))
 			mwifiex_11n_aggregate_pkt(priv, ptr, INTF_HEADER_LEN,
diff --git a/drivers/net/wireless/mwl8k.c b/drivers/net/wireless/mwl8k.c
index 6820fce..a3707fd 100644
--- a/drivers/net/wireless/mwl8k.c
+++ b/drivers/net/wireless/mwl8k.c
@@ -1548,7 +1548,7 @@ static int mwl8k_tx_wait_empty(struct ieee80211_hw *hw)
 	if (!priv->pending_tx_pkts)
 		return 0;
 
-	retry = 0;
+	retry = 1;
 	rc = 0;
 
 	spin_lock_bh(&priv->tx_lock);
@@ -1572,13 +1572,19 @@ static int mwl8k_tx_wait_empty(struct ieee80211_hw *hw)
 
 		spin_lock_bh(&priv->tx_lock);
 
-		if (timeout) {
+		if (timeout || !priv->pending_tx_pkts) {
 			WARN_ON(priv->pending_tx_pkts);
 			if (retry)
 				wiphy_notice(hw->wiphy, "tx rings drained\n");
 			break;
 		}
 
+		if (retry) {
+			mwl8k_tx_start(priv);
+			retry = 0;
+			continue;
+		}
+
 		if (priv->pending_tx_pkts < oldcount) {
 			wiphy_notice(hw->wiphy,
 				     "waiting for tx rings to drain (%d -> %d pkts)\n",
@@ -2055,6 +2061,7 @@ mwl8k_txq_xmit(struct ieee80211_hw *hw,
 				mwl8k_remove_stream(hw, stream);
 				spin_unlock(&priv->stream_lock);
 			}
+			mwl8k_tx_start(priv);
 			spin_unlock_bh(&priv->tx_lock);
 			pci_unmap_single(priv->pdev, dma, skb->len,
 					 PCI_DMA_TODEVICE);
diff --git a/drivers/net/wireless/orinoco/orinoco_pci.h b/drivers/net/wireless/orinoco/orinoco_pci.h
index ea7231a..43f5b9f 100644
--- a/drivers/net/wireless/orinoco/orinoco_pci.h
+++ b/drivers/net/wireless/orinoco/orinoco_pci.h
@@ -38,7 +38,7 @@ static int orinoco_pci_resume(struct pci_dev *pdev)
 	struct net_device *dev = priv->ndev;
 	int err;
 
-	pci_set_power_state(pdev, 0);
+	pci_set_power_state(pdev, PCI_D0);
 	err = pci_enable_device(pdev);
 	if (err) {
 		printk(KERN_ERR "%s: pci_enable_device failed on resume\n",
diff --git a/drivers/net/wireless/orinoco/orinoco_usb.c b/drivers/net/wireless/orinoco/orinoco_usb.c
index 1f9cb55..bdfe637 100644
--- a/drivers/net/wireless/orinoco/orinoco_usb.c
+++ b/drivers/net/wireless/orinoco/orinoco_usb.c
@@ -881,7 +881,8 @@ static int ezusb_access_ltv(struct ezusb_priv *upriv,
 
 	if (!upriv->udev) {
 		dbg("Device disconnected");
-		return -ENODEV;
+		retval = -ENODEV;
+		goto exit;
 	}
 
 	if (upriv->read_urb->status != -EINPROGRESS)
diff --git a/drivers/net/wireless/p54/p54spi.c b/drivers/net/wireless/p54/p54spi.c
index 978e7eb..7fc46f2 100644
--- a/drivers/net/wireless/p54/p54spi.c
+++ b/drivers/net/wireless/p54/p54spi.c
@@ -42,8 +42,7 @@
 
 MODULE_FIRMWARE("3826.arm");
 
-/*
- * gpios should be handled in board files and provided via platform data,
+/* gpios should be handled in board files and provided via platform data,
  * but because it's currently impossible for p54spi to have a header file
  * in include/linux, let's use module paramaters for now
  */
@@ -191,8 +190,7 @@ static int p54spi_request_eeprom(struct ieee80211_hw *dev)
 	const struct firmware *eeprom;
 	int ret;
 
-	/*
-	 * allow users to customize their eeprom.
+	/* allow users to customize their eeprom.
 	 */
 
 	ret = request_firmware(&eeprom, "3826.eeprom", &priv->spi->dev);
@@ -285,8 +283,7 @@ static void p54spi_power_on(struct p54s_priv *priv)
 	gpio_set_value(p54spi_gpio_power, 1);
 	enable_irq(gpio_to_irq(p54spi_gpio_irq));
 
-	/*
-	 * need to wait a while before device can be accessed, the length
+	/* need to wait a while before device can be accessed, the length
 	 * is just a guess
 	 */
 	msleep(10);
@@ -365,7 +362,8 @@ static int p54spi_rx(struct p54s_priv *priv)
 	/* Firmware may insert up to 4 padding bytes after the lmac header,
 	 * but it does not amend the size of SPI data transfer.
 	 * Such packets has correct data size in header, thus referencing
-	 * past the end of allocated skb. Reserve extra 4 bytes for this case */
+	 * past the end of allocated skb. Reserve extra 4 bytes for this case
+	 */
 	skb = dev_alloc_skb(len + 4);
 	if (!skb) {
 		p54spi_sleep(priv);
@@ -383,7 +381,8 @@ static int p54spi_rx(struct p54s_priv *priv)
 	}
 	p54spi_sleep(priv);
 	/* Put additional bytes to compensate for the possible
-	 * alignment-caused truncation */
+	 * alignment-caused truncation
+	 */
 	skb_put(skb, 4);
 
 	if (p54_rx(priv->hw, skb) == 0)
@@ -713,27 +712,7 @@ static struct spi_driver p54spi_driver = {
 	.remove		= p54spi_remove,
 };
 
-static int __init p54spi_init(void)
-{
-	int ret;
-
-	ret = spi_register_driver(&p54spi_driver);
-	if (ret < 0) {
-		printk(KERN_ERR "failed to register SPI driver: %d", ret);
-		goto out;
-	}
-
-out:
-	return ret;
-}
-
-static void __exit p54spi_exit(void)
-{
-	spi_unregister_driver(&p54spi_driver);
-}
-
-module_init(p54spi_init);
-module_exit(p54spi_exit);
+module_spi_driver(p54spi_driver);
 
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Christian Lamparter <chunkeey@web.de>");
diff --git a/drivers/net/wireless/rt2x00/rt2400pci.c b/drivers/net/wireless/rt2x00/rt2400pci.c
index f714373..3d53a09 100644
--- a/drivers/net/wireless/rt2x00/rt2400pci.c
+++ b/drivers/net/wireless/rt2x00/rt2400pci.c
@@ -1767,33 +1767,45 @@ static const struct rt2x00lib_ops rt2400pci_rt2x00_ops = {
 	.config			= rt2400pci_config,
 };
 
-static const struct data_queue_desc rt2400pci_queue_rx = {
-	.entry_num		= 24,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= RXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+static void rt2400pci_queue_init(struct data_queue *queue)
+{
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 24;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = RXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2400pci_queue_tx = {
-	.entry_num		= 24,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 24;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2400pci_queue_bcn = {
-	.entry_num		= 1,
-	.data_size		= MGMT_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_BEACON:
+		queue->limit = 1;
+		queue->data_size = MGMT_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2400pci_queue_atim = {
-	.entry_num		= 8,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_ATIM:
+		queue->limit = 8;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
+
+	default:
+		BUG();
+		break;
+	}
+}
 
 static const struct rt2x00_ops rt2400pci_ops = {
 	.name			= KBUILD_MODNAME,
@@ -1801,11 +1813,7 @@ static const struct rt2x00_ops rt2400pci_ops = {
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= 0,
-	.rx			= &rt2400pci_queue_rx,
-	.tx			= &rt2400pci_queue_tx,
-	.bcn			= &rt2400pci_queue_bcn,
-	.atim			= &rt2400pci_queue_atim,
+	.queue_init		= rt2400pci_queue_init,
 	.lib			= &rt2400pci_rt2x00_ops,
 	.hw			= &rt2400pci_mac80211_ops,
 #ifdef CONFIG_RT2X00_LIB_DEBUGFS
diff --git a/drivers/net/wireless/rt2x00/rt2500pci.c b/drivers/net/wireless/rt2x00/rt2500pci.c
index 77e45b2..0ac5c58 100644
--- a/drivers/net/wireless/rt2x00/rt2500pci.c
+++ b/drivers/net/wireless/rt2x00/rt2500pci.c
@@ -2056,33 +2056,45 @@ static const struct rt2x00lib_ops rt2500pci_rt2x00_ops = {
 	.config			= rt2500pci_config,
 };
 
-static const struct data_queue_desc rt2500pci_queue_rx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= RXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+static void rt2500pci_queue_init(struct data_queue *queue)
+{
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = RXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2500pci_queue_tx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2500pci_queue_bcn = {
-	.entry_num		= 1,
-	.data_size		= MGMT_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_BEACON:
+		queue->limit = 1;
+		queue->data_size = MGMT_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2500pci_queue_atim = {
-	.entry_num		= 8,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_ATIM:
+		queue->limit = 8;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
+
+	default:
+		BUG();
+		break;
+	}
+}
 
 static const struct rt2x00_ops rt2500pci_ops = {
 	.name			= KBUILD_MODNAME,
@@ -2090,11 +2102,7 @@ static const struct rt2x00_ops rt2500pci_ops = {
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= 0,
-	.rx			= &rt2500pci_queue_rx,
-	.tx			= &rt2500pci_queue_tx,
-	.bcn			= &rt2500pci_queue_bcn,
-	.atim			= &rt2500pci_queue_atim,
+	.queue_init		= rt2500pci_queue_init,
 	.lib			= &rt2500pci_rt2x00_ops,
 	.hw			= &rt2500pci_mac80211_ops,
 #ifdef CONFIG_RT2X00_LIB_DEBUGFS
diff --git a/drivers/net/wireless/rt2x00/rt2500usb.c b/drivers/net/wireless/rt2x00/rt2500usb.c
index a7f7b36..85acc79 100644
--- a/drivers/net/wireless/rt2x00/rt2500usb.c
+++ b/drivers/net/wireless/rt2x00/rt2500usb.c
@@ -1867,33 +1867,45 @@ static const struct rt2x00lib_ops rt2500usb_rt2x00_ops = {
 	.config			= rt2500usb_config,
 };
 
-static const struct data_queue_desc rt2500usb_queue_rx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= RXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+static void rt2500usb_queue_init(struct data_queue *queue)
+{
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = RXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
-static const struct data_queue_desc rt2500usb_queue_tx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
-static const struct data_queue_desc rt2500usb_queue_bcn = {
-	.entry_num		= 1,
-	.data_size		= MGMT_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb_bcn),
-};
+	case QID_BEACON:
+		queue->limit = 1;
+		queue->data_size = MGMT_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb_bcn);
+		break;
 
-static const struct data_queue_desc rt2500usb_queue_atim = {
-	.entry_num		= 8,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	case QID_ATIM:
+		queue->limit = 8;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
+
+	default:
+		BUG();
+		break;
+	}
+}
 
 static const struct rt2x00_ops rt2500usb_ops = {
 	.name			= KBUILD_MODNAME,
@@ -1901,11 +1913,7 @@ static const struct rt2x00_ops rt2500usb_ops = {
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= TXD_DESC_SIZE,
-	.rx			= &rt2500usb_queue_rx,
-	.tx			= &rt2500usb_queue_tx,
-	.bcn			= &rt2500usb_queue_bcn,
-	.atim			= &rt2500usb_queue_atim,
+	.queue_init		= rt2500usb_queue_init,
 	.lib			= &rt2500usb_rt2x00_ops,
 	.hw			= &rt2500usb_mac80211_ops,
 #ifdef CONFIG_RT2X00_LIB_DEBUGFS
diff --git a/drivers/net/wireless/rt2x00/rt2800.h b/drivers/net/wireless/rt2x00/rt2800.h
index a7630d5..d78c495 100644
--- a/drivers/net/wireless/rt2x00/rt2800.h
+++ b/drivers/net/wireless/rt2x00/rt2800.h
@@ -100,7 +100,7 @@
 #define CSR_REG_BASE			0x1000
 #define CSR_REG_SIZE			0x0800
 #define EEPROM_BASE			0x0000
-#define EEPROM_SIZE			0x0110
+#define EEPROM_SIZE			0x0200
 #define BBP_BASE			0x0000
 #define BBP_SIZE			0x00ff
 #define RF_BASE				0x0004
@@ -2625,11 +2625,13 @@ struct mac_iveiv_entry {
 /*
  * DMA descriptor defines.
  */
-#define TXWI_DESC_SIZE			(4 * sizeof(__le32))
-#define RXWI_DESC_SIZE			(4 * sizeof(__le32))
 
-#define TXWI_DESC_SIZE_5592		(5 * sizeof(__le32))
-#define RXWI_DESC_SIZE_5592		(6 * sizeof(__le32))
+#define TXWI_DESC_SIZE_4WORDS		(4 * sizeof(__le32))
+#define TXWI_DESC_SIZE_5WORDS		(5 * sizeof(__le32))
+
+#define RXWI_DESC_SIZE_4WORDS		(4 * sizeof(__le32))
+#define RXWI_DESC_SIZE_6WORDS		(6 * sizeof(__le32))
+
 /*
  * TX WI structure
  */
diff --git a/drivers/net/wireless/rt2x00/rt2800lib.c b/drivers/net/wireless/rt2x00/rt2800lib.c
index 72f32e5..1f80ea5 100644
--- a/drivers/net/wireless/rt2x00/rt2800lib.c
+++ b/drivers/net/wireless/rt2x00/rt2800lib.c
@@ -840,7 +840,7 @@ static inline void rt2800_clear_beacon_register(struct rt2x00_dev *rt2x00dev,
 						unsigned int beacon_base)
 {
 	int i;
-	const int txwi_desc_size = rt2x00dev->ops->bcn->winfo_size;
+	const int txwi_desc_size = rt2x00dev->bcn->winfo_size;
 
 	/*
 	 * For the Beacon base registers we only need to clear
@@ -2392,7 +2392,7 @@ static void rt2800_config_channel_rf55xx(struct rt2x00_dev *rt2x00dev,
 	rt2800_rfcsr_write(rt2x00dev, 49, rfcsr);
 
 	rt2800_rfcsr_read(rt2x00dev, 50, &rfcsr);
-	if (info->default_power1 > power_bound)
+	if (info->default_power2 > power_bound)
 		rt2x00_set_field8(&rfcsr, RFCSR50_TX, power_bound);
 	else
 		rt2x00_set_field8(&rfcsr, RFCSR50_TX, info->default_power2);
@@ -2678,30 +2678,53 @@ static void rt2800_config_channel(struct rt2x00_dev *rt2x00dev,
 
 	tx_pin = 0;
 
-	/* Turn on unused PA or LNA when not using 1T or 1R */
-	if (rt2x00dev->default_ant.tx_chain_num == 2) {
+	switch (rt2x00dev->default_ant.tx_chain_num) {
+	case 3:
+		/* Turn on tertiary PAs */
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_A2_EN,
+				   rf->channel > 14);
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_G2_EN,
+				   rf->channel <= 14);
+		/* fall-through */
+	case 2:
+		/* Turn on secondary PAs */
 		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_A1_EN,
 				   rf->channel > 14);
 		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_G1_EN,
 				   rf->channel <= 14);
+		/* fall-through */
+	case 1:
+		/* Turn on primary PAs */
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_A0_EN,
+				   rf->channel > 14);
+		if (test_bit(CAPABILITY_BT_COEXIST, &rt2x00dev->cap_flags))
+			rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_G0_EN, 1);
+		else
+			rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_G0_EN,
+					   rf->channel <= 14);
+		break;
 	}
 
-	/* Turn on unused PA or LNA when not using 1T or 1R */
-	if (rt2x00dev->default_ant.rx_chain_num == 2) {
+	switch (rt2x00dev->default_ant.rx_chain_num) {
+	case 3:
+		/* Turn on tertiary LNAs */
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_A2_EN, 1);
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_G2_EN, 1);
+		/* fall-through */
+	case 2:
+		/* Turn on secondary LNAs */
 		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_A1_EN, 1);
 		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_G1_EN, 1);
+		/* fall-through */
+	case 1:
+		/* Turn on primary LNAs */
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_A0_EN, 1);
+		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_G0_EN, 1);
+		break;
 	}
 
-	rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_A0_EN, 1);
-	rt2x00_set_field32(&tx_pin, TX_PIN_CFG_LNA_PE_G0_EN, 1);
 	rt2x00_set_field32(&tx_pin, TX_PIN_CFG_RFTR_EN, 1);
 	rt2x00_set_field32(&tx_pin, TX_PIN_CFG_TRSW_EN, 1);
-	if (test_bit(CAPABILITY_BT_COEXIST, &rt2x00dev->cap_flags))
-		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_G0_EN, 1);
-	else
-		rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_G0_EN,
-				   rf->channel <= 14);
-	rt2x00_set_field32(&tx_pin, TX_PIN_CFG_PA_PE_A0_EN, rf->channel > 14);
 
 	rt2800_register_write(rt2x00dev, TX_PIN_CFG, tx_pin);
 
@@ -3960,379 +3983,577 @@ static void rt2800_init_bbp_early(struct rt2x00_dev *rt2x00dev)
 	rt2800_bbp_write(rt2x00dev, 106, 0x35);
 }
 
-static void rt2800_init_bbp_5592(struct rt2x00_dev *rt2x00dev)
+static void rt2800_disable_unused_dac_adc(struct rt2x00_dev *rt2x00dev)
 {
-	int ant, div_mode;
 	u16 eeprom;
 	u8 value;
 
-	rt2800_init_bbp_early(rt2x00dev);
+	rt2800_bbp_read(rt2x00dev, 138, &value);
+	rt2x00_eeprom_read(rt2x00dev, EEPROM_NIC_CONF0, &eeprom);
+	if (rt2x00_get_field16(eeprom, EEPROM_NIC_CONF0_TXPATH) == 1)
+		value |= 0x20;
+	if (rt2x00_get_field16(eeprom, EEPROM_NIC_CONF0_RXPATH) == 1)
+		value &= ~0x02;
+	rt2800_bbp_write(rt2x00dev, 138, value);
+}
 
-	rt2800_bbp_read(rt2x00dev, 105, &value);
-	rt2x00_set_field8(&value, BBP105_MLD,
-			  rt2x00dev->default_ant.rx_chain_num == 2);
-	rt2800_bbp_write(rt2x00dev, 105, value);
+static void rt2800_init_bbp_305x_soc(struct rt2x00_dev *rt2x00dev)
+{
+	rt2800_bbp_write(rt2x00dev, 31, 0x08);
+
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
+
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
+	rt2800_bbp_write(rt2x00dev, 73, 0x10);
+
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
+
+	rt2800_bbp_write(rt2x00dev, 78, 0x0e);
+	rt2800_bbp_write(rt2x00dev, 80, 0x08);
+
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
+
+	rt2800_bbp_write(rt2x00dev, 83, 0x6a);
+
+	rt2800_bbp_write(rt2x00dev, 84, 0x99);
+
+	rt2800_bbp_write(rt2x00dev, 86, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
+	rt2800_bbp_write(rt2x00dev, 92, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+
+	rt2800_bbp_write(rt2x00dev, 105, 0x01);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x35);
+}
+
+static void rt2800_init_bbp_28xx(struct rt2x00_dev *rt2x00dev)
+{
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
+
+	if (rt2x00_rt_rev(rt2x00dev, RT2860, REV_RT2860C)) {
+		rt2800_bbp_write(rt2x00dev, 69, 0x16);
+		rt2800_bbp_write(rt2x00dev, 73, 0x12);
+	} else {
+		rt2800_bbp_write(rt2x00dev, 69, 0x12);
+		rt2800_bbp_write(rt2x00dev, 73, 0x10);
+	}
+
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
+
+	rt2800_bbp_write(rt2x00dev, 81, 0x37);
+
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
+
+	rt2800_bbp_write(rt2x00dev, 83, 0x6a);
+
+	if (rt2x00_rt_rev(rt2x00dev, RT2860, REV_RT2860D))
+		rt2800_bbp_write(rt2x00dev, 84, 0x19);
+	else
+		rt2800_bbp_write(rt2x00dev, 84, 0x99);
+
+	rt2800_bbp_write(rt2x00dev, 86, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
+	rt2800_bbp_write(rt2x00dev, 92, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 103, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 105, 0x05);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x35);
+}
+
+static void rt2800_init_bbp_30xx(struct rt2x00_dev *rt2x00dev)
+{
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
+
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
+	rt2800_bbp_write(rt2x00dev, 73, 0x10);
+
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
+
+	rt2800_bbp_write(rt2x00dev, 79, 0x13);
+	rt2800_bbp_write(rt2x00dev, 80, 0x05);
+	rt2800_bbp_write(rt2x00dev, 81, 0x33);
+
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
+
+	rt2800_bbp_write(rt2x00dev, 83, 0x6a);
+
+	rt2800_bbp_write(rt2x00dev, 84, 0x99);
+
+	rt2800_bbp_write(rt2x00dev, 86, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
+	rt2800_bbp_write(rt2x00dev, 92, 0x00);
+
+	if (rt2x00_rt_rev_gte(rt2x00dev, RT3070, REV_RT3070F) ||
+	    rt2x00_rt_rev_gte(rt2x00dev, RT3071, REV_RT3071E) ||
+	    rt2x00_rt_rev_gte(rt2x00dev, RT3090, REV_RT3090E))
+		rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+	else
+		rt2800_bbp_write(rt2x00dev, 103, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 105, 0x05);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x35);
+
+	if (rt2x00_rt(rt2x00dev, RT3071) ||
+	    rt2x00_rt(rt2x00dev, RT3090))
+		rt2800_disable_unused_dac_adc(rt2x00dev);
+}
+
+static void rt2800_init_bbp_3290(struct rt2x00_dev *rt2x00dev)
+{
+	u8 value;
 
 	rt2800_bbp4_mac_if_ctrl(rt2x00dev);
 
-	rt2800_bbp_write(rt2x00dev, 20, 0x06);
 	rt2800_bbp_write(rt2x00dev, 31, 0x08);
-	rt2800_bbp_write(rt2x00dev, 65, 0x2C);
-	rt2800_bbp_write(rt2x00dev, 68, 0xDD);
-	rt2800_bbp_write(rt2x00dev, 69, 0x1A);
-	rt2800_bbp_write(rt2x00dev, 70, 0x05);
+
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
+
+	rt2800_bbp_write(rt2x00dev, 68, 0x0b);
+
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
 	rt2800_bbp_write(rt2x00dev, 73, 0x13);
-	rt2800_bbp_write(rt2x00dev, 74, 0x0F);
-	rt2800_bbp_write(rt2x00dev, 75, 0x4F);
+	rt2800_bbp_write(rt2x00dev, 75, 0x46);
 	rt2800_bbp_write(rt2x00dev, 76, 0x28);
+
+	rt2800_bbp_write(rt2x00dev, 77, 0x58);
+
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
+
+	rt2800_bbp_write(rt2x00dev, 74, 0x0b);
+	rt2800_bbp_write(rt2x00dev, 79, 0x18);
+	rt2800_bbp_write(rt2x00dev, 80, 0x09);
+	rt2800_bbp_write(rt2x00dev, 81, 0x33);
+
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
+
+	rt2800_bbp_write(rt2x00dev, 83, 0x7a);
+
+	rt2800_bbp_write(rt2x00dev, 84, 0x9a);
+
+	rt2800_bbp_write(rt2x00dev, 86, 0x38);
+
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
+	rt2800_bbp_write(rt2x00dev, 92, 0x02);
+
+	rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+
+	rt2800_bbp_write(rt2x00dev, 104, 0x92);
+
+	rt2800_bbp_write(rt2x00dev, 105, 0x1c);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x03);
+
+	rt2800_bbp_write(rt2x00dev, 128, 0x12);
+
+	rt2800_bbp_write(rt2x00dev, 67, 0x24);
+	rt2800_bbp_write(rt2x00dev, 143, 0x04);
+	rt2800_bbp_write(rt2x00dev, 142, 0x99);
+	rt2800_bbp_write(rt2x00dev, 150, 0x30);
+	rt2800_bbp_write(rt2x00dev, 151, 0x2e);
+	rt2800_bbp_write(rt2x00dev, 152, 0x20);
+	rt2800_bbp_write(rt2x00dev, 153, 0x34);
+	rt2800_bbp_write(rt2x00dev, 154, 0x40);
+	rt2800_bbp_write(rt2x00dev, 155, 0x3b);
+	rt2800_bbp_write(rt2x00dev, 253, 0x04);
+
+	rt2800_bbp_read(rt2x00dev, 47, &value);
+	rt2x00_set_field8(&value, BBP47_TSSI_ADC6, 1);
+	rt2800_bbp_write(rt2x00dev, 47, value);
+
+	/* Use 5-bit ADC for Acquisition and 8-bit ADC for data */
+	rt2800_bbp_read(rt2x00dev, 3, &value);
+	rt2x00_set_field8(&value, BBP3_ADC_MODE_SWITCH, 1);
+	rt2x00_set_field8(&value, BBP3_ADC_INIT_MODE, 1);
+	rt2800_bbp_write(rt2x00dev, 3, value);
+}
+
+static void rt2800_init_bbp_3352(struct rt2x00_dev *rt2x00dev)
+{
+	rt2800_bbp_write(rt2x00dev, 3, 0x00);
+	rt2800_bbp_write(rt2x00dev, 4, 0x50);
+
+	rt2800_bbp_write(rt2x00dev, 31, 0x08);
+
+	rt2800_bbp_write(rt2x00dev, 47, 0x48);
+
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
+
+	rt2800_bbp_write(rt2x00dev, 68, 0x0b);
+
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
+	rt2800_bbp_write(rt2x00dev, 73, 0x13);
+	rt2800_bbp_write(rt2x00dev, 75, 0x46);
+	rt2800_bbp_write(rt2x00dev, 76, 0x28);
+
 	rt2800_bbp_write(rt2x00dev, 77, 0x59);
-	rt2800_bbp_write(rt2x00dev, 84, 0x9A);
+
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
+
+	rt2800_bbp_write(rt2x00dev, 78, 0x0e);
+	rt2800_bbp_write(rt2x00dev, 80, 0x08);
+	rt2800_bbp_write(rt2x00dev, 81, 0x37);
+
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
+
+	rt2800_bbp_write(rt2x00dev, 83, 0x6a);
+
+	rt2800_bbp_write(rt2x00dev, 84, 0x99);
+
 	rt2800_bbp_write(rt2x00dev, 86, 0x38);
+
 	rt2800_bbp_write(rt2x00dev, 88, 0x90);
+
 	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
 	rt2800_bbp_write(rt2x00dev, 92, 0x02);
-	rt2800_bbp_write(rt2x00dev, 95, 0x9a);
-	rt2800_bbp_write(rt2x00dev, 98, 0x12);
-	rt2800_bbp_write(rt2x00dev, 103, 0xC0);
+
+	rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+
 	rt2800_bbp_write(rt2x00dev, 104, 0x92);
-	/* FIXME BBP105 owerwrite */
-	rt2800_bbp_write(rt2x00dev, 105, 0x3C);
-	rt2800_bbp_write(rt2x00dev, 106, 0x35);
-	rt2800_bbp_write(rt2x00dev, 128, 0x12);
-	rt2800_bbp_write(rt2x00dev, 134, 0xD0);
-	rt2800_bbp_write(rt2x00dev, 135, 0xF6);
-	rt2800_bbp_write(rt2x00dev, 137, 0x0F);
 
-	/* Initialize GLRT (Generalized Likehood Radio Test) */
-	rt2800_init_bbp_5592_glrt(rt2x00dev);
+	rt2800_bbp_write(rt2x00dev, 105, 0x34);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x05);
+
+	rt2800_bbp_write(rt2x00dev, 120, 0x50);
+
+	rt2800_bbp_write(rt2x00dev, 137, 0x0f);
+
+	rt2800_bbp_write(rt2x00dev, 163, 0xbd);
+	/* Set ITxBF timeout to 0x9c40=1000msec */
+	rt2800_bbp_write(rt2x00dev, 179, 0x02);
+	rt2800_bbp_write(rt2x00dev, 180, 0x00);
+	rt2800_bbp_write(rt2x00dev, 182, 0x40);
+	rt2800_bbp_write(rt2x00dev, 180, 0x01);
+	rt2800_bbp_write(rt2x00dev, 182, 0x9c);
+	rt2800_bbp_write(rt2x00dev, 179, 0x00);
+	/* Reprogram the inband interface to put right values in RXWI */
+	rt2800_bbp_write(rt2x00dev, 142, 0x04);
+	rt2800_bbp_write(rt2x00dev, 143, 0x3b);
+	rt2800_bbp_write(rt2x00dev, 142, 0x06);
+	rt2800_bbp_write(rt2x00dev, 143, 0xa0);
+	rt2800_bbp_write(rt2x00dev, 142, 0x07);
+	rt2800_bbp_write(rt2x00dev, 143, 0xa1);
+	rt2800_bbp_write(rt2x00dev, 142, 0x08);
+	rt2800_bbp_write(rt2x00dev, 143, 0xa2);
+
+	rt2800_bbp_write(rt2x00dev, 148, 0xc8);
+}
 
-	rt2800_bbp4_mac_if_ctrl(rt2x00dev);
+static void rt2800_init_bbp_3390(struct rt2x00_dev *rt2x00dev)
+{
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
 
-	rt2x00_eeprom_read(rt2x00dev, EEPROM_NIC_CONF1, &eeprom);
-	div_mode = rt2x00_get_field16(eeprom, EEPROM_NIC_CONF1_ANT_DIVERSITY);
-	ant = (div_mode == 3) ? 1 : 0;
-	rt2800_bbp_read(rt2x00dev, 152, &value);
-	if (ant == 0) {
-		/* Main antenna */
-		rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 1);
-	} else {
-		/* Auxiliary antenna */
-		rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 0);
-	}
-	rt2800_bbp_write(rt2x00dev, 152, value);
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
+	rt2800_bbp_write(rt2x00dev, 73, 0x10);
 
-	if (rt2x00_rt_rev_gte(rt2x00dev, RT5592, REV_RT5592C)) {
-		rt2800_bbp_read(rt2x00dev, 254, &value);
-		rt2x00_set_field8(&value, BBP254_BIT7, 1);
-		rt2800_bbp_write(rt2x00dev, 254, value);
-	}
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
 
-	rt2800_init_freq_calibration(rt2x00dev);
+	rt2800_bbp_write(rt2x00dev, 79, 0x13);
+	rt2800_bbp_write(rt2x00dev, 80, 0x05);
+	rt2800_bbp_write(rt2x00dev, 81, 0x33);
 
-	rt2800_bbp_write(rt2x00dev, 84, 0x19);
-	if (rt2x00_rt_rev_gte(rt2x00dev, RT5592, REV_RT5592C))
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
+
+	rt2800_bbp_write(rt2x00dev, 83, 0x6a);
+
+	rt2800_bbp_write(rt2x00dev, 84, 0x99);
+
+	rt2800_bbp_write(rt2x00dev, 86, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
+	rt2800_bbp_write(rt2x00dev, 92, 0x00);
+
+	if (rt2x00_rt_rev_gte(rt2x00dev, RT3390, REV_RT3390E))
 		rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+	else
+		rt2800_bbp_write(rt2x00dev, 103, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 105, 0x05);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x35);
+
+	rt2800_disable_unused_dac_adc(rt2x00dev);
 }
 
-static int rt2800_init_bbp(struct rt2x00_dev *rt2x00dev)
+static void rt2800_init_bbp_3572(struct rt2x00_dev *rt2x00dev)
 {
-	unsigned int i;
-	u16 eeprom;
-	u8 reg_id;
-	u8 value;
+	rt2800_bbp_write(rt2x00dev, 31, 0x08);
 
-	if (unlikely(rt2800_wait_bbp_rf_ready(rt2x00dev) ||
-		     rt2800_wait_bbp_ready(rt2x00dev)))
-		return -EACCES;
+	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
+	rt2800_bbp_write(rt2x00dev, 66, 0x38);
 
-	if (rt2x00_rt(rt2x00dev, RT5592)) {
-		rt2800_init_bbp_5592(rt2x00dev);
-		return 0;
-	}
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
+	rt2800_bbp_write(rt2x00dev, 73, 0x10);
 
-	if (rt2x00_rt(rt2x00dev, RT3352)) {
-		rt2800_bbp_write(rt2x00dev, 3, 0x00);
-		rt2800_bbp_write(rt2x00dev, 4, 0x50);
-	}
+	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp4_mac_if_ctrl(rt2x00dev);
+	rt2800_bbp_write(rt2x00dev, 79, 0x13);
+	rt2800_bbp_write(rt2x00dev, 80, 0x05);
+	rt2800_bbp_write(rt2x00dev, 81, 0x33);
 
-	if (rt2800_is_305x_soc(rt2x00dev) ||
-	    rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT3572) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 31, 0x08);
+	rt2800_bbp_write(rt2x00dev, 82, 0x62);
 
-	if (rt2x00_rt(rt2x00dev, RT3352))
-		rt2800_bbp_write(rt2x00dev, 47, 0x48);
+	rt2800_bbp_write(rt2x00dev, 83, 0x6a);
+
+	rt2800_bbp_write(rt2x00dev, 84, 0x99);
+
+	rt2800_bbp_write(rt2x00dev, 86, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+
+	rt2800_bbp_write(rt2x00dev, 92, 0x00);
+
+	rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+
+	rt2800_bbp_write(rt2x00dev, 105, 0x05);
+
+	rt2800_bbp_write(rt2x00dev, 106, 0x35);
+
+	rt2800_disable_unused_dac_adc(rt2x00dev);
+}
+
+static void rt2800_init_bbp_53xx(struct rt2x00_dev *rt2x00dev)
+{
+	int ant, div_mode;
+	u16 eeprom;
+	u8 value;
+
+	rt2800_bbp4_mac_if_ctrl(rt2x00dev);
+
+	rt2800_bbp_write(rt2x00dev, 31, 0x08);
 
 	rt2800_bbp_write(rt2x00dev, 65, 0x2c);
 	rt2800_bbp_write(rt2x00dev, 66, 0x38);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 68, 0x0b);
+	rt2800_bbp_write(rt2x00dev, 68, 0x0b);
 
-	if (rt2x00_rt_rev(rt2x00dev, RT2860, REV_RT2860C)) {
-		rt2800_bbp_write(rt2x00dev, 69, 0x16);
-		rt2800_bbp_write(rt2x00dev, 73, 0x12);
-	} else if (rt2x00_rt(rt2x00dev, RT3290) ||
-		   rt2x00_rt(rt2x00dev, RT3352) ||
-		   rt2x00_rt(rt2x00dev, RT5390) ||
-		   rt2x00_rt(rt2x00dev, RT5392)) {
-		rt2800_bbp_write(rt2x00dev, 69, 0x12);
-		rt2800_bbp_write(rt2x00dev, 73, 0x13);
-		rt2800_bbp_write(rt2x00dev, 75, 0x46);
-		rt2800_bbp_write(rt2x00dev, 76, 0x28);
+	rt2800_bbp_write(rt2x00dev, 69, 0x12);
+	rt2800_bbp_write(rt2x00dev, 73, 0x13);
+	rt2800_bbp_write(rt2x00dev, 75, 0x46);
+	rt2800_bbp_write(rt2x00dev, 76, 0x28);
 
-		if (rt2x00_rt(rt2x00dev, RT3290))
-			rt2800_bbp_write(rt2x00dev, 77, 0x58);
-		else
-			rt2800_bbp_write(rt2x00dev, 77, 0x59);
-	} else {
-		rt2800_bbp_write(rt2x00dev, 69, 0x12);
-		rt2800_bbp_write(rt2x00dev, 73, 0x10);
-	}
+	rt2800_bbp_write(rt2x00dev, 77, 0x59);
 
 	rt2800_bbp_write(rt2x00dev, 70, 0x0a);
 
-	if (rt2x00_rt(rt2x00dev, RT3070) ||
-	    rt2x00_rt(rt2x00dev, RT3071) ||
-	    rt2x00_rt(rt2x00dev, RT3090) ||
-	    rt2x00_rt(rt2x00dev, RT3390) ||
-	    rt2x00_rt(rt2x00dev, RT3572) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392)) {
-		rt2800_bbp_write(rt2x00dev, 79, 0x13);
-		rt2800_bbp_write(rt2x00dev, 80, 0x05);
-		rt2800_bbp_write(rt2x00dev, 81, 0x33);
-	} else if (rt2800_is_305x_soc(rt2x00dev)) {
-		rt2800_bbp_write(rt2x00dev, 78, 0x0e);
-		rt2800_bbp_write(rt2x00dev, 80, 0x08);
-	} else if (rt2x00_rt(rt2x00dev, RT3290)) {
-		rt2800_bbp_write(rt2x00dev, 74, 0x0b);
-		rt2800_bbp_write(rt2x00dev, 79, 0x18);
-		rt2800_bbp_write(rt2x00dev, 80, 0x09);
-		rt2800_bbp_write(rt2x00dev, 81, 0x33);
-	} else if (rt2x00_rt(rt2x00dev, RT3352)) {
-		rt2800_bbp_write(rt2x00dev, 78, 0x0e);
-		rt2800_bbp_write(rt2x00dev, 80, 0x08);
-		rt2800_bbp_write(rt2x00dev, 81, 0x37);
-	} else {
-		rt2800_bbp_write(rt2x00dev, 81, 0x37);
-	}
+	rt2800_bbp_write(rt2x00dev, 79, 0x13);
+	rt2800_bbp_write(rt2x00dev, 80, 0x05);
+	rt2800_bbp_write(rt2x00dev, 81, 0x33);
 
 	rt2800_bbp_write(rt2x00dev, 82, 0x62);
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 83, 0x7a);
-	else
-		rt2800_bbp_write(rt2x00dev, 83, 0x6a);
 
-	if (rt2x00_rt_rev(rt2x00dev, RT2860, REV_RT2860D))
-		rt2800_bbp_write(rt2x00dev, 84, 0x19);
-	else if (rt2x00_rt(rt2x00dev, RT3290) ||
-		 rt2x00_rt(rt2x00dev, RT5390) ||
-		 rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 84, 0x9a);
-	else
-		rt2800_bbp_write(rt2x00dev, 84, 0x99);
+	rt2800_bbp_write(rt2x00dev, 83, 0x7a);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 86, 0x38);
-	else
-		rt2800_bbp_write(rt2x00dev, 86, 0x00);
+	rt2800_bbp_write(rt2x00dev, 84, 0x9a);
 
-	if (rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
+	rt2800_bbp_write(rt2x00dev, 86, 0x38);
+
+	if (rt2x00_rt(rt2x00dev, RT5392))
 		rt2800_bbp_write(rt2x00dev, 88, 0x90);
 
 	rt2800_bbp_write(rt2x00dev, 91, 0x04);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 92, 0x02);
-	else
-		rt2800_bbp_write(rt2x00dev, 92, 0x00);
+	rt2800_bbp_write(rt2x00dev, 92, 0x02);
 
 	if (rt2x00_rt(rt2x00dev, RT5392)) {
 		rt2800_bbp_write(rt2x00dev, 95, 0x9a);
 		rt2800_bbp_write(rt2x00dev, 98, 0x12);
 	}
 
-	if (rt2x00_rt_rev_gte(rt2x00dev, RT3070, REV_RT3070F) ||
-	    rt2x00_rt_rev_gte(rt2x00dev, RT3071, REV_RT3071E) ||
-	    rt2x00_rt_rev_gte(rt2x00dev, RT3090, REV_RT3090E) ||
-	    rt2x00_rt_rev_gte(rt2x00dev, RT3390, REV_RT3390E) ||
-	    rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT3572) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392) ||
-	    rt2800_is_305x_soc(rt2x00dev))
-		rt2800_bbp_write(rt2x00dev, 103, 0xc0);
-	else
-		rt2800_bbp_write(rt2x00dev, 103, 0x00);
+	rt2800_bbp_write(rt2x00dev, 103, 0xc0);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT3352) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 104, 0x92);
+	rt2800_bbp_write(rt2x00dev, 104, 0x92);
 
-	if (rt2800_is_305x_soc(rt2x00dev))
-		rt2800_bbp_write(rt2x00dev, 105, 0x01);
-	else if (rt2x00_rt(rt2x00dev, RT3290))
-		rt2800_bbp_write(rt2x00dev, 105, 0x1c);
-	else if (rt2x00_rt(rt2x00dev, RT3352))
-		rt2800_bbp_write(rt2x00dev, 105, 0x34);
-	else if (rt2x00_rt(rt2x00dev, RT5390) ||
-		 rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 105, 0x3c);
-	else
-		rt2800_bbp_write(rt2x00dev, 105, 0x05);
+	rt2800_bbp_write(rt2x00dev, 105, 0x3c);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT5390))
+	if (rt2x00_rt(rt2x00dev, RT5390))
 		rt2800_bbp_write(rt2x00dev, 106, 0x03);
-	else if (rt2x00_rt(rt2x00dev, RT3352))
-		rt2800_bbp_write(rt2x00dev, 106, 0x05);
 	else if (rt2x00_rt(rt2x00dev, RT5392))
 		rt2800_bbp_write(rt2x00dev, 106, 0x12);
 	else
-		rt2800_bbp_write(rt2x00dev, 106, 0x35);
-
-	if (rt2x00_rt(rt2x00dev, RT3352))
-		rt2800_bbp_write(rt2x00dev, 120, 0x50);
+		WARN_ON(1);
 
-	if (rt2x00_rt(rt2x00dev, RT3290) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392))
-		rt2800_bbp_write(rt2x00dev, 128, 0x12);
+	rt2800_bbp_write(rt2x00dev, 128, 0x12);
 
 	if (rt2x00_rt(rt2x00dev, RT5392)) {
 		rt2800_bbp_write(rt2x00dev, 134, 0xd0);
 		rt2800_bbp_write(rt2x00dev, 135, 0xf6);
 	}
 
-	if (rt2x00_rt(rt2x00dev, RT3352))
-		rt2800_bbp_write(rt2x00dev, 137, 0x0f);
+	rt2800_disable_unused_dac_adc(rt2x00dev);
 
-	if (rt2x00_rt(rt2x00dev, RT3071) ||
-	    rt2x00_rt(rt2x00dev, RT3090) ||
-	    rt2x00_rt(rt2x00dev, RT3390) ||
-	    rt2x00_rt(rt2x00dev, RT3572) ||
-	    rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392)) {
-		rt2800_bbp_read(rt2x00dev, 138, &value);
+	rt2x00_eeprom_read(rt2x00dev, EEPROM_NIC_CONF1, &eeprom);
+	div_mode = rt2x00_get_field16(eeprom,
+				      EEPROM_NIC_CONF1_ANT_DIVERSITY);
+	ant = (div_mode == 3) ? 1 : 0;
 
-		rt2x00_eeprom_read(rt2x00dev, EEPROM_NIC_CONF0, &eeprom);
-		if (rt2x00_get_field16(eeprom, EEPROM_NIC_CONF0_TXPATH) == 1)
-			value |= 0x20;
-		if (rt2x00_get_field16(eeprom, EEPROM_NIC_CONF0_RXPATH) == 1)
-			value &= ~0x02;
+	/* check if this is a Bluetooth combo card */
+	if (test_bit(CAPABILITY_BT_COEXIST, &rt2x00dev->cap_flags)) {
+		u32 reg;
 
-		rt2800_bbp_write(rt2x00dev, 138, value);
+		rt2800_register_read(rt2x00dev, GPIO_CTRL, &reg);
+		rt2x00_set_field32(&reg, GPIO_CTRL_DIR3, 0);
+		rt2x00_set_field32(&reg, GPIO_CTRL_DIR6, 0);
+		rt2x00_set_field32(&reg, GPIO_CTRL_VAL3, 0);
+		rt2x00_set_field32(&reg, GPIO_CTRL_VAL6, 0);
+		if (ant == 0)
+			rt2x00_set_field32(&reg, GPIO_CTRL_VAL3, 1);
+		else if (ant == 1)
+			rt2x00_set_field32(&reg, GPIO_CTRL_VAL6, 1);
+		rt2800_register_write(rt2x00dev, GPIO_CTRL, reg);
 	}
 
-	if (rt2x00_rt(rt2x00dev, RT3290)) {
-		rt2800_bbp_write(rt2x00dev, 67, 0x24);
-		rt2800_bbp_write(rt2x00dev, 143, 0x04);
-		rt2800_bbp_write(rt2x00dev, 142, 0x99);
-		rt2800_bbp_write(rt2x00dev, 150, 0x30);
-		rt2800_bbp_write(rt2x00dev, 151, 0x2e);
-		rt2800_bbp_write(rt2x00dev, 152, 0x20);
-		rt2800_bbp_write(rt2x00dev, 153, 0x34);
-		rt2800_bbp_write(rt2x00dev, 154, 0x40);
-		rt2800_bbp_write(rt2x00dev, 155, 0x3b);
-		rt2800_bbp_write(rt2x00dev, 253, 0x04);
-
-		rt2800_bbp_read(rt2x00dev, 47, &value);
-		rt2x00_set_field8(&value, BBP47_TSSI_ADC6, 1);
-		rt2800_bbp_write(rt2x00dev, 47, value);
-
-		/* Use 5-bit ADC for Acquisition and 8-bit ADC for data */
-		rt2800_bbp_read(rt2x00dev, 3, &value);
-		rt2x00_set_field8(&value, BBP3_ADC_MODE_SWITCH, 1);
-		rt2x00_set_field8(&value, BBP3_ADC_INIT_MODE, 1);
-		rt2800_bbp_write(rt2x00dev, 3, value);
+	/* This chip has hardware antenna diversity*/
+	if (rt2x00_rt_rev_gte(rt2x00dev, RT5390, REV_RT5390R)) {
+		rt2800_bbp_write(rt2x00dev, 150, 0); /* Disable Antenna Software OFDM */
+		rt2800_bbp_write(rt2x00dev, 151, 0); /* Disable Antenna Software CCK */
+		rt2800_bbp_write(rt2x00dev, 154, 0); /* Clear previously selected antenna */
 	}
 
-	if (rt2x00_rt(rt2x00dev, RT3352)) {
-		rt2800_bbp_write(rt2x00dev, 163, 0xbd);
-		/* Set ITxBF timeout to 0x9c40=1000msec */
-		rt2800_bbp_write(rt2x00dev, 179, 0x02);
-		rt2800_bbp_write(rt2x00dev, 180, 0x00);
-		rt2800_bbp_write(rt2x00dev, 182, 0x40);
-		rt2800_bbp_write(rt2x00dev, 180, 0x01);
-		rt2800_bbp_write(rt2x00dev, 182, 0x9c);
-		rt2800_bbp_write(rt2x00dev, 179, 0x00);
-		/* Reprogram the inband interface to put right values in RXWI */
-		rt2800_bbp_write(rt2x00dev, 142, 0x04);
-		rt2800_bbp_write(rt2x00dev, 143, 0x3b);
-		rt2800_bbp_write(rt2x00dev, 142, 0x06);
-		rt2800_bbp_write(rt2x00dev, 143, 0xa0);
-		rt2800_bbp_write(rt2x00dev, 142, 0x07);
-		rt2800_bbp_write(rt2x00dev, 143, 0xa1);
-		rt2800_bbp_write(rt2x00dev, 142, 0x08);
-		rt2800_bbp_write(rt2x00dev, 143, 0xa2);
-
-		rt2800_bbp_write(rt2x00dev, 148, 0xc8);
+	rt2800_bbp_read(rt2x00dev, 152, &value);
+	if (ant == 0)
+		rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 1);
+	else
+		rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 0);
+	rt2800_bbp_write(rt2x00dev, 152, value);
+
+	rt2800_init_freq_calibration(rt2x00dev);
+}
+
+static void rt2800_init_bbp_5592(struct rt2x00_dev *rt2x00dev)
+{
+	int ant, div_mode;
+	u16 eeprom;
+	u8 value;
+
+	rt2800_init_bbp_early(rt2x00dev);
+
+	rt2800_bbp_read(rt2x00dev, 105, &value);
+	rt2x00_set_field8(&value, BBP105_MLD,
+			  rt2x00dev->default_ant.rx_chain_num == 2);
+	rt2800_bbp_write(rt2x00dev, 105, value);
+
+	rt2800_bbp4_mac_if_ctrl(rt2x00dev);
+
+	rt2800_bbp_write(rt2x00dev, 20, 0x06);
+	rt2800_bbp_write(rt2x00dev, 31, 0x08);
+	rt2800_bbp_write(rt2x00dev, 65, 0x2C);
+	rt2800_bbp_write(rt2x00dev, 68, 0xDD);
+	rt2800_bbp_write(rt2x00dev, 69, 0x1A);
+	rt2800_bbp_write(rt2x00dev, 70, 0x05);
+	rt2800_bbp_write(rt2x00dev, 73, 0x13);
+	rt2800_bbp_write(rt2x00dev, 74, 0x0F);
+	rt2800_bbp_write(rt2x00dev, 75, 0x4F);
+	rt2800_bbp_write(rt2x00dev, 76, 0x28);
+	rt2800_bbp_write(rt2x00dev, 77, 0x59);
+	rt2800_bbp_write(rt2x00dev, 84, 0x9A);
+	rt2800_bbp_write(rt2x00dev, 86, 0x38);
+	rt2800_bbp_write(rt2x00dev, 88, 0x90);
+	rt2800_bbp_write(rt2x00dev, 91, 0x04);
+	rt2800_bbp_write(rt2x00dev, 92, 0x02);
+	rt2800_bbp_write(rt2x00dev, 95, 0x9a);
+	rt2800_bbp_write(rt2x00dev, 98, 0x12);
+	rt2800_bbp_write(rt2x00dev, 103, 0xC0);
+	rt2800_bbp_write(rt2x00dev, 104, 0x92);
+	/* FIXME BBP105 owerwrite */
+	rt2800_bbp_write(rt2x00dev, 105, 0x3C);
+	rt2800_bbp_write(rt2x00dev, 106, 0x35);
+	rt2800_bbp_write(rt2x00dev, 128, 0x12);
+	rt2800_bbp_write(rt2x00dev, 134, 0xD0);
+	rt2800_bbp_write(rt2x00dev, 135, 0xF6);
+	rt2800_bbp_write(rt2x00dev, 137, 0x0F);
+
+	/* Initialize GLRT (Generalized Likehood Radio Test) */
+	rt2800_init_bbp_5592_glrt(rt2x00dev);
+
+	rt2800_bbp4_mac_if_ctrl(rt2x00dev);
+
+	rt2x00_eeprom_read(rt2x00dev, EEPROM_NIC_CONF1, &eeprom);
+	div_mode = rt2x00_get_field16(eeprom, EEPROM_NIC_CONF1_ANT_DIVERSITY);
+	ant = (div_mode == 3) ? 1 : 0;
+	rt2800_bbp_read(rt2x00dev, 152, &value);
+	if (ant == 0) {
+		/* Main antenna */
+		rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 1);
+	} else {
+		/* Auxiliary antenna */
+		rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 0);
 	}
+	rt2800_bbp_write(rt2x00dev, 152, value);
 
-	if (rt2x00_rt(rt2x00dev, RT5390) ||
-	    rt2x00_rt(rt2x00dev, RT5392)) {
-		int ant, div_mode;
+	if (rt2x00_rt_rev_gte(rt2x00dev, RT5592, REV_RT5592C)) {
+		rt2800_bbp_read(rt2x00dev, 254, &value);
+		rt2x00_set_field8(&value, BBP254_BIT7, 1);
+		rt2800_bbp_write(rt2x00dev, 254, value);
+	}
 
-		rt2x00_eeprom_read(rt2x00dev, EEPROM_NIC_CONF1, &eeprom);
-		div_mode = rt2x00_get_field16(eeprom,
-					      EEPROM_NIC_CONF1_ANT_DIVERSITY);
-		ant = (div_mode == 3) ? 1 : 0;
+	rt2800_init_freq_calibration(rt2x00dev);
 
-		/* check if this is a Bluetooth combo card */
-		if (test_bit(CAPABILITY_BT_COEXIST, &rt2x00dev->cap_flags)) {
-			u32 reg;
-
-			rt2800_register_read(rt2x00dev, GPIO_CTRL, &reg);
-			rt2x00_set_field32(&reg, GPIO_CTRL_DIR3, 0);
-			rt2x00_set_field32(&reg, GPIO_CTRL_DIR6, 0);
-			rt2x00_set_field32(&reg, GPIO_CTRL_VAL3, 0);
-			rt2x00_set_field32(&reg, GPIO_CTRL_VAL6, 0);
-			if (ant == 0)
-				rt2x00_set_field32(&reg, GPIO_CTRL_VAL3, 1);
-			else if (ant == 1)
-				rt2x00_set_field32(&reg, GPIO_CTRL_VAL6, 1);
-			rt2800_register_write(rt2x00dev, GPIO_CTRL, reg);
-		}
+	rt2800_bbp_write(rt2x00dev, 84, 0x19);
+	if (rt2x00_rt_rev_gte(rt2x00dev, RT5592, REV_RT5592C))
+		rt2800_bbp_write(rt2x00dev, 103, 0xc0);
+}
 
-		/* This chip has hardware antenna diversity*/
-		if (rt2x00_rt_rev_gte(rt2x00dev, RT5390, REV_RT5390R)) {
-			rt2800_bbp_write(rt2x00dev, 150, 0); /* Disable Antenna Software OFDM */
-			rt2800_bbp_write(rt2x00dev, 151, 0); /* Disable Antenna Software CCK */
-			rt2800_bbp_write(rt2x00dev, 154, 0); /* Clear previously selected antenna */
-		}
+static void rt2800_init_bbp(struct rt2x00_dev *rt2x00dev)
+{
+	unsigned int i;
+	u16 eeprom;
+	u8 reg_id;
+	u8 value;
 
-		rt2800_bbp_read(rt2x00dev, 152, &value);
-		if (ant == 0)
-			rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 1);
-		else
-			rt2x00_set_field8(&value, BBP152_RX_DEFAULT_ANT, 0);
-		rt2800_bbp_write(rt2x00dev, 152, value);
+	if (rt2800_is_305x_soc(rt2x00dev))
+		rt2800_init_bbp_305x_soc(rt2x00dev);
 
-		rt2800_init_freq_calibration(rt2x00dev);
+	switch (rt2x00dev->chip.rt) {
+	case RT2860:
+	case RT2872:
+	case RT2883:
+		rt2800_init_bbp_28xx(rt2x00dev);
+		break;
+	case RT3070:
+	case RT3071:
+	case RT3090:
+		rt2800_init_bbp_30xx(rt2x00dev);
+		break;
+	case RT3290:
+		rt2800_init_bbp_3290(rt2x00dev);
+		break;
+	case RT3352:
+		rt2800_init_bbp_3352(rt2x00dev);
+		break;
+	case RT3390:
+		rt2800_init_bbp_3390(rt2x00dev);
+		break;
+	case RT3572:
+		rt2800_init_bbp_3572(rt2x00dev);
+		break;
+	case RT5390:
+	case RT5392:
+		rt2800_init_bbp_53xx(rt2x00dev);
+		break;
+	case RT5592:
+		rt2800_init_bbp_5592(rt2x00dev);
+		return;
 	}
 
 	for (i = 0; i < EEPROM_BBP_SIZE; i++) {
@@ -4344,8 +4565,6 @@ static int rt2800_init_bbp(struct rt2x00_dev *rt2x00dev)
 			rt2800_bbp_write(rt2x00dev, reg_id, value);
 		}
 	}
-
-	return 0;
 }
 
 static void rt2800_led_open_drain_enable(struct rt2x00_dev *rt2x00dev)
@@ -5196,9 +5415,11 @@ int rt2800_enable_radio(struct rt2x00_dev *rt2x00dev)
 	}
 	msleep(1);
 
-	if (unlikely(rt2800_init_bbp(rt2x00dev)))
+	if (unlikely(rt2800_wait_bbp_rf_ready(rt2x00dev) ||
+		     rt2800_wait_bbp_ready(rt2x00dev)))
 		return -EIO;
 
+	rt2800_init_bbp(rt2x00dev);
 	rt2800_init_rfcsr(rt2x00dev);
 
 	if (rt2x00_is_usb(rt2x00dev) &&
@@ -6056,8 +6277,8 @@ static int rt2800_probe_hw_mode(struct rt2x00_dev *rt2x00dev)
 		default_power2 = rt2x00_eeprom_addr(rt2x00dev, EEPROM_TXPOWER_A2);
 
 		for (i = 14; i < spec->num_channels; i++) {
-			info[i].default_power1 = default_power1[i];
-			info[i].default_power2 = default_power2[i];
+			info[i].default_power1 = default_power1[i - 14];
+			info[i].default_power2 = default_power2[i - 14];
 		}
 	}
 
diff --git a/drivers/net/wireless/rt2x00/rt2800pci.c b/drivers/net/wireless/rt2x00/rt2800pci.c
index 6f4a861..0005562 100644
--- a/drivers/net/wireless/rt2x00/rt2800pci.c
+++ b/drivers/net/wireless/rt2x00/rt2800pci.c
@@ -637,6 +637,7 @@ static void rt2800pci_write_tx_desc(struct queue_entry *entry,
 	struct queue_entry_priv_mmio *entry_priv = entry->priv_data;
 	__le32 *txd = entry_priv->desc;
 	u32 word;
+	const unsigned int txwi_size = entry->queue->winfo_size;
 
 	/*
 	 * The buffers pointed by SD_PTR0/SD_LEN0 and SD_PTR1/SD_LEN1
@@ -659,14 +660,14 @@ static void rt2800pci_write_tx_desc(struct queue_entry *entry,
 			   !test_bit(ENTRY_TXD_MORE_FRAG, &txdesc->flags));
 	rt2x00_set_field32(&word, TXD_W1_BURST,
 			   test_bit(ENTRY_TXD_BURST, &txdesc->flags));
-	rt2x00_set_field32(&word, TXD_W1_SD_LEN0, TXWI_DESC_SIZE);
+	rt2x00_set_field32(&word, TXD_W1_SD_LEN0, txwi_size);
 	rt2x00_set_field32(&word, TXD_W1_LAST_SEC0, 0);
 	rt2x00_set_field32(&word, TXD_W1_DMA_DONE, 0);
 	rt2x00_desc_write(txd, 1, word);
 
 	word = 0;
 	rt2x00_set_field32(&word, TXD_W2_SD_PTR1,
-			   skbdesc->skb_dma + TXWI_DESC_SIZE);
+			   skbdesc->skb_dma + txwi_size);
 	rt2x00_desc_write(txd, 2, word);
 
 	word = 0;
@@ -1014,7 +1015,7 @@ static void rt2800pci_txstatus_interrupt(struct rt2x00_dev *rt2x00dev)
 	 * Since we have only one producer and one consumer we don't
 	 * need to lock the kfifo.
 	 */
-	for (i = 0; i < rt2x00dev->ops->tx->entry_num; i++) {
+	for (i = 0; i < rt2x00dev->tx->limit; i++) {
 		rt2x00mmio_register_read(rt2x00dev, TX_STA_FIFO, &status);
 
 		if (!rt2x00_get_field32(status, TX_STA_FIFO_VALID))
@@ -1186,29 +1187,43 @@ static const struct rt2x00lib_ops rt2800pci_rt2x00_ops = {
 	.sta_remove		= rt2800_sta_remove,
 };
 
-static const struct data_queue_desc rt2800pci_queue_rx = {
-	.entry_num		= 128,
-	.data_size		= AGGREGATION_SIZE,
-	.desc_size		= RXD_DESC_SIZE,
-	.winfo_size		= RXWI_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+static void rt2800pci_queue_init(struct data_queue *queue)
+{
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 128;
+		queue->data_size = AGGREGATION_SIZE;
+		queue->desc_size = RXD_DESC_SIZE;
+		queue->winfo_size = RXWI_DESC_SIZE_4WORDS;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2800pci_queue_tx = {
-	.entry_num		= 64,
-	.data_size		= AGGREGATION_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.winfo_size		= TXWI_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 64;
+		queue->data_size = AGGREGATION_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->winfo_size = TXWI_DESC_SIZE_4WORDS;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt2800pci_queue_bcn = {
-	.entry_num		= 8,
-	.data_size		= 0, /* No DMA required for beacons */
-	.desc_size		= TXD_DESC_SIZE,
-	.winfo_size		= TXWI_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_BEACON:
+		queue->limit = 8;
+		queue->data_size = 0; /* No DMA required for beacons */
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->winfo_size = TXWI_DESC_SIZE_4WORDS;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
+
+	case QID_ATIM:
+		/* fallthrough */
+	default:
+		BUG();
+		break;
+	}
+}
 
 static const struct rt2x00_ops rt2800pci_ops = {
 	.name			= KBUILD_MODNAME,
@@ -1217,10 +1232,7 @@ static const struct rt2x00_ops rt2800pci_ops = {
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= TXWI_DESC_SIZE,
-	.rx			= &rt2800pci_queue_rx,
-	.tx			= &rt2800pci_queue_tx,
-	.bcn			= &rt2800pci_queue_bcn,
+	.queue_init		= rt2800pci_queue_init,
 	.lib			= &rt2800pci_rt2x00_ops,
 	.drv			= &rt2800pci_rt2800_ops,
 	.hw			= &rt2800pci_mac80211_ops,
diff --git a/drivers/net/wireless/rt2x00/rt2800usb.c b/drivers/net/wireless/rt2x00/rt2800usb.c
index ac854d7..840833b 100644
--- a/drivers/net/wireless/rt2x00/rt2800usb.c
+++ b/drivers/net/wireless/rt2x00/rt2800usb.c
@@ -327,7 +327,7 @@ static int rt2800usb_enable_radio(struct rt2x00_dev *rt2x00dev)
 	 * this limit so reduce the number to prevent errors.
 	 */
 	rt2x00_set_field32(&reg, USB_DMA_CFG_RX_BULK_AGG_LIMIT,
-			   ((rt2x00dev->ops->rx->entry_num * DATA_FRAME_SIZE)
+			   ((rt2x00dev->rx->limit * DATA_FRAME_SIZE)
 			    / 1024) - 3);
 	rt2x00_set_field32(&reg, USB_DMA_CFG_RX_BULK_EN, 1);
 	rt2x00_set_field32(&reg, USB_DMA_CFG_TX_BULK_EN, 1);
@@ -849,85 +849,63 @@ static const struct rt2x00lib_ops rt2800usb_rt2x00_ops = {
 	.sta_remove		= rt2800_sta_remove,
 };
 
-static const struct data_queue_desc rt2800usb_queue_rx = {
-	.entry_num		= 128,
-	.data_size		= AGGREGATION_SIZE,
-	.desc_size		= RXINFO_DESC_SIZE,
-	.winfo_size		= RXWI_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
-
-static const struct data_queue_desc rt2800usb_queue_tx = {
-	.entry_num		= 16,
-	.data_size		= AGGREGATION_SIZE,
-	.desc_size		= TXINFO_DESC_SIZE,
-	.winfo_size		= TXWI_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
-
-static const struct data_queue_desc rt2800usb_queue_bcn = {
-	.entry_num		= 8,
-	.data_size		= MGMT_FRAME_SIZE,
-	.desc_size		= TXINFO_DESC_SIZE,
-	.winfo_size		= TXWI_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+static void rt2800usb_queue_init(struct data_queue *queue)
+{
+	struct rt2x00_dev *rt2x00dev = queue->rt2x00dev;
+	unsigned short txwi_size, rxwi_size;
 
-static const struct rt2x00_ops rt2800usb_ops = {
-	.name			= KBUILD_MODNAME,
-	.drv_data_size		= sizeof(struct rt2800_drv_data),
-	.max_ap_intf		= 8,
-	.eeprom_size		= EEPROM_SIZE,
-	.rf_size		= RF_SIZE,
-	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= TXINFO_DESC_SIZE + TXWI_DESC_SIZE,
-	.rx			= &rt2800usb_queue_rx,
-	.tx			= &rt2800usb_queue_tx,
-	.bcn			= &rt2800usb_queue_bcn,
-	.lib			= &rt2800usb_rt2x00_ops,
-	.drv			= &rt2800usb_rt2800_ops,
-	.hw			= &rt2800usb_mac80211_ops,
-#ifdef CONFIG_RT2X00_LIB_DEBUGFS
-	.debugfs		= &rt2800_rt2x00debug,
-#endif /* CONFIG_RT2X00_LIB_DEBUGFS */
-};
+	if (rt2x00_rt(rt2x00dev, RT5592)) {
+		txwi_size = TXWI_DESC_SIZE_5WORDS;
+		rxwi_size = RXWI_DESC_SIZE_6WORDS;
+	} else {
+		txwi_size = TXWI_DESC_SIZE_4WORDS;
+		rxwi_size = RXWI_DESC_SIZE_4WORDS;
+	}
 
-static const struct data_queue_desc rt2800usb_queue_rx_5592 = {
-	.entry_num		= 128,
-	.data_size		= AGGREGATION_SIZE,
-	.desc_size		= RXINFO_DESC_SIZE,
-	.winfo_size		= RXWI_DESC_SIZE_5592,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 128;
+		queue->data_size = AGGREGATION_SIZE;
+		queue->desc_size = RXINFO_DESC_SIZE;
+		queue->winfo_size = rxwi_size;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
-static const struct data_queue_desc rt2800usb_queue_tx_5592 = {
-	.entry_num		= 16,
-	.data_size		= AGGREGATION_SIZE,
-	.desc_size		= TXINFO_DESC_SIZE,
-	.winfo_size		= TXWI_DESC_SIZE_5592,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 16;
+		queue->data_size = AGGREGATION_SIZE;
+		queue->desc_size = TXINFO_DESC_SIZE;
+		queue->winfo_size = txwi_size;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
-static const struct data_queue_desc rt2800usb_queue_bcn_5592 = {
-	.entry_num		= 8,
-	.data_size		= MGMT_FRAME_SIZE,
-	.desc_size		= TXINFO_DESC_SIZE,
-	.winfo_size		= TXWI_DESC_SIZE_5592,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	case QID_BEACON:
+		queue->limit = 8;
+		queue->data_size = MGMT_FRAME_SIZE;
+		queue->desc_size = TXINFO_DESC_SIZE;
+		queue->winfo_size = txwi_size;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
+	case QID_ATIM:
+		/* fallthrough */
+	default:
+		BUG();
+		break;
+	}
+}
 
-static const struct rt2x00_ops rt2800usb_ops_5592 = {
+static const struct rt2x00_ops rt2800usb_ops = {
 	.name			= KBUILD_MODNAME,
 	.drv_data_size		= sizeof(struct rt2800_drv_data),
 	.max_ap_intf		= 8,
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= TXINFO_DESC_SIZE + TXWI_DESC_SIZE_5592,
-	.rx			= &rt2800usb_queue_rx_5592,
-	.tx			= &rt2800usb_queue_tx_5592,
-	.bcn			= &rt2800usb_queue_bcn_5592,
+	.queue_init		= rt2800usb_queue_init,
 	.lib			= &rt2800usb_rt2x00_ops,
 	.drv			= &rt2800usb_rt2800_ops,
 	.hw			= &rt2800usb_mac80211_ops,
@@ -1248,15 +1226,15 @@ static struct usb_device_id rt2800usb_device_table[] = {
 #endif
 #ifdef CONFIG_RT2800USB_RT55XX
 	/* Arcadyan */
-	{ USB_DEVICE(0x043e, 0x7a32), .driver_info = 5592 },
+	{ USB_DEVICE(0x043e, 0x7a32) },
 	/* AVM GmbH */
-	{ USB_DEVICE(0x057c, 0x8501), .driver_info = 5592 },
+	{ USB_DEVICE(0x057c, 0x8501) },
 	/* D-Link DWA-160-B2 */
-	{ USB_DEVICE(0x2001, 0x3c1a), .driver_info = 5592 },
+	{ USB_DEVICE(0x2001, 0x3c1a) },
 	/* Proware */
-	{ USB_DEVICE(0x043e, 0x7a13), .driver_info = 5592 },
+	{ USB_DEVICE(0x043e, 0x7a13) },
 	/* Ralink */
-	{ USB_DEVICE(0x148f, 0x5572), .driver_info = 5592 },
+	{ USB_DEVICE(0x148f, 0x5572) },
 #endif
 #ifdef CONFIG_RT2800USB_UNKNOWN
 	/*
@@ -1361,9 +1339,6 @@ MODULE_LICENSE("GPL");
 static int rt2800usb_probe(struct usb_interface *usb_intf,
 			   const struct usb_device_id *id)
 {
-	if (id->driver_info == 5592)
-		return rt2x00usb_probe(usb_intf, &rt2800usb_ops_5592);
-
 	return rt2x00usb_probe(usb_intf, &rt2800usb_ops);
 }
 
diff --git a/drivers/net/wireless/rt2x00/rt2x00.h b/drivers/net/wireless/rt2x00/rt2x00.h
index 7510723..ee3fc57 100644
--- a/drivers/net/wireless/rt2x00/rt2x00.h
+++ b/drivers/net/wireless/rt2x00/rt2x00.h
@@ -648,11 +648,7 @@ struct rt2x00_ops {
 	const unsigned int eeprom_size;
 	const unsigned int rf_size;
 	const unsigned int tx_queues;
-	const unsigned int extra_tx_headroom;
-	const struct data_queue_desc *rx;
-	const struct data_queue_desc *tx;
-	const struct data_queue_desc *bcn;
-	const struct data_queue_desc *atim;
+	void (*queue_init)(struct data_queue *queue);
 	const struct rt2x00lib_ops *lib;
 	const void *drv;
 	const struct ieee80211_ops *hw;
@@ -1010,6 +1006,9 @@ struct rt2x00_dev {
 	 */
 	struct list_head bar_list;
 	spinlock_t bar_list_lock;
+
+	/* Extra TX headroom required for alignment purposes. */
+	unsigned int extra_tx_headroom;
 };
 
 struct rt2x00_bar_list_entry {
diff --git a/drivers/net/wireless/rt2x00/rt2x00dev.c b/drivers/net/wireless/rt2x00/rt2x00dev.c
index c8b9ef0..b16521e 100644
--- a/drivers/net/wireless/rt2x00/rt2x00dev.c
+++ b/drivers/net/wireless/rt2x00/rt2x00dev.c
@@ -334,7 +334,7 @@ void rt2x00lib_txdone(struct queue_entry *entry,
 	/*
 	 * Remove the extra tx headroom from the skb.
 	 */
-	skb_pull(entry->skb, rt2x00dev->ops->extra_tx_headroom);
+	skb_pull(entry->skb, rt2x00dev->extra_tx_headroom);
 
 	/*
 	 * Signal that the TX descriptor is no longer in the skb.
@@ -1049,7 +1049,7 @@ static int rt2x00lib_probe_hw(struct rt2x00_dev *rt2x00dev)
 	 */
 	rt2x00dev->hw->extra_tx_headroom =
 		max_t(unsigned int, IEEE80211_TX_STATUS_HEADROOM,
-		      rt2x00dev->ops->extra_tx_headroom);
+		      rt2x00dev->extra_tx_headroom);
 
 	/*
 	 * Take TX headroom required for alignment into account.
@@ -1077,7 +1077,7 @@ static int rt2x00lib_probe_hw(struct rt2x00_dev *rt2x00dev)
 		 */
 		int kfifo_size =
 			roundup_pow_of_two(rt2x00dev->ops->tx_queues *
-					   rt2x00dev->ops->tx->entry_num *
+					   rt2x00dev->tx->limit *
 					   sizeof(u32));
 
 		status = kfifo_alloc(&rt2x00dev->txstatus_fifo, kfifo_size,
@@ -1256,6 +1256,17 @@ static inline void rt2x00lib_set_if_combinations(struct rt2x00_dev *rt2x00dev)
 	rt2x00dev->hw->wiphy->n_iface_combinations = 1;
 }
 
+static unsigned int rt2x00dev_extra_tx_headroom(struct rt2x00_dev *rt2x00dev)
+{
+	if (WARN_ON(!rt2x00dev->tx))
+		return 0;
+
+	if (rt2x00_is_usb(rt2x00dev))
+		return rt2x00dev->tx[0].winfo_size + rt2x00dev->tx[0].desc_size;
+
+	return rt2x00dev->tx[0].winfo_size;
+}
+
 /*
  * driver allocation handlers.
  */
@@ -1301,23 +1312,6 @@ int rt2x00lib_probe_dev(struct rt2x00_dev *rt2x00dev)
 		(rt2x00dev->ops->max_ap_intf - 1);
 
 	/*
-	 * Determine which operating modes are supported, all modes
-	 * which require beaconing, depend on the availability of
-	 * beacon entries.
-	 */
-	rt2x00dev->hw->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION);
-	if (rt2x00dev->ops->bcn->entry_num > 0)
-		rt2x00dev->hw->wiphy->interface_modes |=
-		    BIT(NL80211_IFTYPE_ADHOC) |
-		    BIT(NL80211_IFTYPE_AP) |
-#ifdef CONFIG_MAC80211_MESH
-		    BIT(NL80211_IFTYPE_MESH_POINT) |
-#endif
-		    BIT(NL80211_IFTYPE_WDS);
-
-	rt2x00dev->hw->wiphy->flags |= WIPHY_FLAG_IBSS_RSN;
-
-	/*
 	 * Initialize work.
 	 */
 	rt2x00dev->workqueue =
@@ -1347,6 +1341,26 @@ int rt2x00lib_probe_dev(struct rt2x00_dev *rt2x00dev)
 	if (retval)
 		goto exit;
 
+	/* Cache TX headroom value */
+	rt2x00dev->extra_tx_headroom = rt2x00dev_extra_tx_headroom(rt2x00dev);
+
+	/*
+	 * Determine which operating modes are supported, all modes
+	 * which require beaconing, depend on the availability of
+	 * beacon entries.
+	 */
+	rt2x00dev->hw->wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION);
+	if (rt2x00dev->bcn->limit > 0)
+		rt2x00dev->hw->wiphy->interface_modes |=
+		    BIT(NL80211_IFTYPE_ADHOC) |
+		    BIT(NL80211_IFTYPE_AP) |
+#ifdef CONFIG_MAC80211_MESH
+		    BIT(NL80211_IFTYPE_MESH_POINT) |
+#endif
+		    BIT(NL80211_IFTYPE_WDS);
+
+	rt2x00dev->hw->wiphy->flags |= WIPHY_FLAG_IBSS_RSN;
+
 	/*
 	 * Initialize ieee80211 structure.
 	 */
diff --git a/drivers/net/wireless/rt2x00/rt2x00pci.c b/drivers/net/wireless/rt2x00/rt2x00pci.c
index dc49e52..76d95de 100644
--- a/drivers/net/wireless/rt2x00/rt2x00pci.c
+++ b/drivers/net/wireless/rt2x00/rt2x00pci.c
@@ -105,11 +105,13 @@ int rt2x00pci_probe(struct pci_dev *pci_dev, const struct rt2x00_ops *ops)
 		goto exit_release_regions;
 	}
 
+	pci_enable_msi(pci_dev);
+
 	hw = ieee80211_alloc_hw(sizeof(struct rt2x00_dev), ops->hw);
 	if (!hw) {
 		rt2x00_probe_err("Failed to allocate hardware\n");
 		retval = -ENOMEM;
-		goto exit_release_regions;
+		goto exit_disable_msi;
 	}
 
 	pci_set_drvdata(pci_dev, hw);
@@ -150,6 +152,9 @@ exit_free_reg:
 exit_free_device:
 	ieee80211_free_hw(hw);
 
+exit_disable_msi:
+	pci_disable_msi(pci_dev);
+
 exit_release_regions:
 	pci_release_regions(pci_dev);
 
@@ -174,6 +179,8 @@ void rt2x00pci_remove(struct pci_dev *pci_dev)
 	rt2x00pci_free_reg(rt2x00dev);
 	ieee80211_free_hw(hw);
 
+	pci_disable_msi(pci_dev);
+
 	/*
 	 * Free the PCI device data.
 	 */
diff --git a/drivers/net/wireless/rt2x00/rt2x00queue.c b/drivers/net/wireless/rt2x00/rt2x00queue.c
index 2c12311..6c0a91f 100644
--- a/drivers/net/wireless/rt2x00/rt2x00queue.c
+++ b/drivers/net/wireless/rt2x00/rt2x00queue.c
@@ -542,8 +542,8 @@ static int rt2x00queue_write_tx_data(struct queue_entry *entry,
 	/*
 	 * Add the requested extra tx headroom in front of the skb.
 	 */
-	skb_push(entry->skb, rt2x00dev->ops->extra_tx_headroom);
-	memset(entry->skb->data, 0, rt2x00dev->ops->extra_tx_headroom);
+	skb_push(entry->skb, rt2x00dev->extra_tx_headroom);
+	memset(entry->skb->data, 0, rt2x00dev->extra_tx_headroom);
 
 	/*
 	 * Call the driver's write_tx_data function, if it exists.
@@ -596,7 +596,7 @@ static void rt2x00queue_bar_check(struct queue_entry *entry)
 {
 	struct rt2x00_dev *rt2x00dev = entry->queue->rt2x00dev;
 	struct ieee80211_bar *bar = (void *) (entry->skb->data +
-				    rt2x00dev->ops->extra_tx_headroom);
+				    rt2x00dev->extra_tx_headroom);
 	struct rt2x00_bar_list_entry *bar_entry;
 
 	if (likely(!ieee80211_is_back_req(bar->frame_control)))
@@ -1161,8 +1161,7 @@ void rt2x00queue_init_queues(struct rt2x00_dev *rt2x00dev)
 	}
 }
 
-static int rt2x00queue_alloc_entries(struct data_queue *queue,
-				     const struct data_queue_desc *qdesc)
+static int rt2x00queue_alloc_entries(struct data_queue *queue)
 {
 	struct queue_entry *entries;
 	unsigned int entry_size;
@@ -1170,16 +1169,10 @@ static int rt2x00queue_alloc_entries(struct data_queue *queue,
 
 	rt2x00queue_reset(queue);
 
-	queue->limit = qdesc->entry_num;
-	queue->threshold = DIV_ROUND_UP(qdesc->entry_num, 10);
-	queue->data_size = qdesc->data_size;
-	queue->desc_size = qdesc->desc_size;
-	queue->winfo_size = qdesc->winfo_size;
-
 	/*
 	 * Allocate all queue entries.
 	 */
-	entry_size = sizeof(*entries) + qdesc->priv_size;
+	entry_size = sizeof(*entries) + queue->priv_size;
 	entries = kcalloc(queue->limit, entry_size, GFP_KERNEL);
 	if (!entries)
 		return -ENOMEM;
@@ -1195,7 +1188,7 @@ static int rt2x00queue_alloc_entries(struct data_queue *queue,
 		entries[i].entry_idx = i;
 		entries[i].priv_data =
 		    QUEUE_ENTRY_PRIV_OFFSET(entries, i, queue->limit,
-					    sizeof(*entries), qdesc->priv_size);
+					    sizeof(*entries), queue->priv_size);
 	}
 
 #undef QUEUE_ENTRY_PRIV_OFFSET
@@ -1237,23 +1230,22 @@ int rt2x00queue_initialize(struct rt2x00_dev *rt2x00dev)
 	struct data_queue *queue;
 	int status;
 
-	status = rt2x00queue_alloc_entries(rt2x00dev->rx, rt2x00dev->ops->rx);
+	status = rt2x00queue_alloc_entries(rt2x00dev->rx);
 	if (status)
 		goto exit;
 
 	tx_queue_for_each(rt2x00dev, queue) {
-		status = rt2x00queue_alloc_entries(queue, rt2x00dev->ops->tx);
+		status = rt2x00queue_alloc_entries(queue);
 		if (status)
 			goto exit;
 	}
 
-	status = rt2x00queue_alloc_entries(rt2x00dev->bcn, rt2x00dev->ops->bcn);
+	status = rt2x00queue_alloc_entries(rt2x00dev->bcn);
 	if (status)
 		goto exit;
 
 	if (test_bit(REQUIRE_ATIM_QUEUE, &rt2x00dev->cap_flags)) {
-		status = rt2x00queue_alloc_entries(rt2x00dev->atim,
-						   rt2x00dev->ops->atim);
+		status = rt2x00queue_alloc_entries(rt2x00dev->atim);
 		if (status)
 			goto exit;
 	}
@@ -1297,6 +1289,10 @@ static void rt2x00queue_init(struct rt2x00_dev *rt2x00dev,
 	queue->aifs = 2;
 	queue->cw_min = 5;
 	queue->cw_max = 10;
+
+	rt2x00dev->ops->queue_init(queue);
+
+	queue->threshold = DIV_ROUND_UP(queue->limit, 10);
 }
 
 int rt2x00queue_allocate(struct rt2x00_dev *rt2x00dev)
diff --git a/drivers/net/wireless/rt2x00/rt2x00queue.h b/drivers/net/wireless/rt2x00/rt2x00queue.h
index 4a7b34e..ebe1172 100644
--- a/drivers/net/wireless/rt2x00/rt2x00queue.h
+++ b/drivers/net/wireless/rt2x00/rt2x00queue.h
@@ -453,6 +453,7 @@ enum data_queue_flags {
  * @cw_max: The cw max value for outgoing frames (field ignored in RX queue).
  * @data_size: Maximum data size for the frames in this queue.
  * @desc_size: Hardware descriptor size for the data in this queue.
+ * @priv_size: Size of per-queue_entry private data.
  * @usb_endpoint: Device endpoint used for communication (USB only)
  * @usb_maxpacket: Max packet size for given endpoint (USB only)
  */
@@ -481,31 +482,13 @@ struct data_queue {
 	unsigned short data_size;
 	unsigned char  desc_size;
 	unsigned char  winfo_size;
+	unsigned short priv_size;
 
 	unsigned short usb_endpoint;
 	unsigned short usb_maxpacket;
 };
 
 /**
- * struct data_queue_desc: Data queue description
- *
- * The information in this structure is used by drivers
- * to inform rt2x00lib about the creation of the data queue.
- *
- * @entry_num: Maximum number of entries for a queue.
- * @data_size: Maximum data size for the frames in this queue.
- * @desc_size: Hardware descriptor size for the data in this queue.
- * @priv_size: Size of per-queue_entry private data.
- */
-struct data_queue_desc {
-	unsigned short entry_num;
-	unsigned short data_size;
-	unsigned char  desc_size;
-	unsigned char  winfo_size;
-	unsigned short priv_size;
-};
-
-/**
  * queue_end - Return pointer to the last queue (HELPER MACRO).
  * @__dev: Pointer to &struct rt2x00_dev
  *
diff --git a/drivers/net/wireless/rt2x00/rt61pci.c b/drivers/net/wireless/rt2x00/rt61pci.c
index 0dc8180..54d3ddf 100644
--- a/drivers/net/wireless/rt2x00/rt61pci.c
+++ b/drivers/net/wireless/rt2x00/rt61pci.c
@@ -2175,7 +2175,7 @@ static void rt61pci_txdone(struct rt2x00_dev *rt2x00dev)
 	 * that the TX_STA_FIFO stack has a size of 16. We stick to our
 	 * tx ring size for now.
 	 */
-	for (i = 0; i < rt2x00dev->ops->tx->entry_num; i++) {
+	for (i = 0; i < rt2x00dev->tx->limit; i++) {
 		rt2x00mmio_register_read(rt2x00dev, STA_CSR4, &reg);
 		if (!rt2x00_get_field32(reg, STA_CSR4_VALID))
 			break;
@@ -2825,7 +2825,8 @@ static int rt61pci_probe_hw_mode(struct rt2x00_dev *rt2x00dev)
 		tx_power = rt2x00_eeprom_addr(rt2x00dev, EEPROM_TXPOWER_A_START);
 		for (i = 14; i < spec->num_channels; i++) {
 			info[i].max_power = MAX_TXPOWER;
-			info[i].default_power1 = TXPOWER_FROM_DEV(tx_power[i]);
+			info[i].default_power1 =
+					TXPOWER_FROM_DEV(tx_power[i - 14]);
 		}
 	}
 
@@ -3025,26 +3026,40 @@ static const struct rt2x00lib_ops rt61pci_rt2x00_ops = {
 	.config			= rt61pci_config,
 };
 
-static const struct data_queue_desc rt61pci_queue_rx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= RXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+static void rt61pci_queue_init(struct data_queue *queue)
+{
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = RXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt61pci_queue_tx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
 
-static const struct data_queue_desc rt61pci_queue_bcn = {
-	.entry_num		= 4,
-	.data_size		= 0, /* No DMA required for beacons */
-	.desc_size		= TXINFO_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_mmio),
-};
+	case QID_BEACON:
+		queue->limit = 4;
+		queue->data_size = 0; /* No DMA required for beacons */
+		queue->desc_size = TXINFO_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_mmio);
+		break;
+
+	case QID_ATIM:
+		/* fallthrough */
+	default:
+		BUG();
+		break;
+	}
+}
 
 static const struct rt2x00_ops rt61pci_ops = {
 	.name			= KBUILD_MODNAME,
@@ -3052,10 +3067,7 @@ static const struct rt2x00_ops rt61pci_ops = {
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= 0,
-	.rx			= &rt61pci_queue_rx,
-	.tx			= &rt61pci_queue_tx,
-	.bcn			= &rt61pci_queue_bcn,
+	.queue_init		= rt61pci_queue_init,
 	.lib			= &rt61pci_rt2x00_ops,
 	.hw			= &rt61pci_mac80211_ops,
 #ifdef CONFIG_RT2X00_LIB_DEBUGFS
diff --git a/drivers/net/wireless/rt2x00/rt73usb.c b/drivers/net/wireless/rt2x00/rt73usb.c
index 377e09b..1d3880e 100644
--- a/drivers/net/wireless/rt2x00/rt73usb.c
+++ b/drivers/net/wireless/rt2x00/rt73usb.c
@@ -2167,7 +2167,8 @@ static int rt73usb_probe_hw_mode(struct rt2x00_dev *rt2x00dev)
 		tx_power = rt2x00_eeprom_addr(rt2x00dev, EEPROM_TXPOWER_A_START);
 		for (i = 14; i < spec->num_channels; i++) {
 			info[i].max_power = MAX_TXPOWER;
-			info[i].default_power1 = TXPOWER_FROM_DEV(tx_power[i]);
+			info[i].default_power1 =
+					TXPOWER_FROM_DEV(tx_power[i - 14]);
 		}
 	}
 
@@ -2359,26 +2360,40 @@ static const struct rt2x00lib_ops rt73usb_rt2x00_ops = {
 	.config			= rt73usb_config,
 };
 
-static const struct data_queue_desc rt73usb_queue_rx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= RXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+static void rt73usb_queue_init(struct data_queue *queue)
+{
+	switch (queue->qid) {
+	case QID_RX:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = RXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
-static const struct data_queue_desc rt73usb_queue_tx = {
-	.entry_num		= 32,
-	.data_size		= DATA_FRAME_SIZE,
-	.desc_size		= TXD_DESC_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	case QID_AC_VO:
+	case QID_AC_VI:
+	case QID_AC_BE:
+	case QID_AC_BK:
+		queue->limit = 32;
+		queue->data_size = DATA_FRAME_SIZE;
+		queue->desc_size = TXD_DESC_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
 
-static const struct data_queue_desc rt73usb_queue_bcn = {
-	.entry_num		= 4,
-	.data_size		= MGMT_FRAME_SIZE,
-	.desc_size		= TXINFO_SIZE,
-	.priv_size		= sizeof(struct queue_entry_priv_usb),
-};
+	case QID_BEACON:
+		queue->limit = 4;
+		queue->data_size = MGMT_FRAME_SIZE;
+		queue->desc_size = TXINFO_SIZE;
+		queue->priv_size = sizeof(struct queue_entry_priv_usb);
+		break;
+
+	case QID_ATIM:
+		/* fallthrough */
+	default:
+		BUG();
+		break;
+	}
+}
 
 static const struct rt2x00_ops rt73usb_ops = {
 	.name			= KBUILD_MODNAME,
@@ -2386,10 +2401,7 @@ static const struct rt2x00_ops rt73usb_ops = {
 	.eeprom_size		= EEPROM_SIZE,
 	.rf_size		= RF_SIZE,
 	.tx_queues		= NUM_TX_QUEUES,
-	.extra_tx_headroom	= TXD_DESC_SIZE,
-	.rx			= &rt73usb_queue_rx,
-	.tx			= &rt73usb_queue_tx,
-	.bcn			= &rt73usb_queue_bcn,
+	.queue_init		= rt73usb_queue_init,
 	.lib			= &rt73usb_rt2x00_ops,
 	.hw			= &rt73usb_mac80211_ops,
 #ifdef CONFIG_RT2X00_LIB_DEBUGFS
diff --git a/drivers/net/wireless/rtlwifi/base.c b/drivers/net/wireless/rtlwifi/base.c
index a5f2231..9d558ac 100644
--- a/drivers/net/wireless/rtlwifi/base.c
+++ b/drivers/net/wireless/rtlwifi/base.c
@@ -1817,7 +1817,7 @@ static ssize_t rtl_store_debug_level(struct device *d,
 	unsigned long val;
 	int ret;
 
-	ret = strict_strtoul(buf, 0, &val);
+	ret = kstrtoul(buf, 0, &val);
 	if (ret) {
 		printk(KERN_DEBUG "%s is not in hex or decimal form.\n", buf);
 	} else {
diff --git a/drivers/net/wireless/rtlwifi/rtl8192cu/rf.c b/drivers/net/wireless/rtlwifi/rtl8192cu/rf.c
index 953f1a0..2119313 100644
--- a/drivers/net/wireless/rtlwifi/rtl8192cu/rf.c
+++ b/drivers/net/wireless/rtlwifi/rtl8192cu/rf.c
@@ -104,7 +104,7 @@ void rtl92cu_phy_rf6052_set_cck_txpower(struct ieee80211_hw *hw,
 			tx_agc[RF90_PATH_A] = 0x10101010;
 			tx_agc[RF90_PATH_B] = 0x10101010;
 		} else if (rtlpriv->dm.dynamic_txhighpower_lvl ==
-			   TXHIGHPWRLEVEL_LEVEL1) {
+			   TXHIGHPWRLEVEL_LEVEL2) {
 			tx_agc[RF90_PATH_A] = 0x00000000;
 			tx_agc[RF90_PATH_B] = 0x00000000;
 		} else{
diff --git a/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c b/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
index 826f085..2bd5985 100644
--- a/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
+++ b/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
@@ -359,6 +359,7 @@ static struct usb_device_id rtl8192c_usb_ids[] = {
 	{RTL_USB_DEVICE(0x2001, 0x330a, rtl92cu_hal_cfg)}, /*D-Link-Alpha*/
 	{RTL_USB_DEVICE(0x2019, 0xab2b, rtl92cu_hal_cfg)}, /*Planex -Abocom*/
 	{RTL_USB_DEVICE(0x20f4, 0x624d, rtl92cu_hal_cfg)}, /*TRENDNet*/
+	{RTL_USB_DEVICE(0x2357, 0x0100, rtl92cu_hal_cfg)}, /*TP-Link WN8200ND*/
 	{RTL_USB_DEVICE(0x7392, 0x7822, rtl92cu_hal_cfg)}, /*Edimax -Edimax*/
 	{}
 };
diff --git a/drivers/net/wireless/rtlwifi/rtl8192de/dm.c b/drivers/net/wireless/rtlwifi/rtl8192de/dm.c
index 19a7655..47875ba 100644
--- a/drivers/net/wireless/rtlwifi/rtl8192de/dm.c
+++ b/drivers/net/wireless/rtlwifi/rtl8192de/dm.c
@@ -842,7 +842,7 @@ static void rtl92d_dm_txpower_tracking_callback_thermalmeter(
 	long val_y, ele_c = 0;
 	u8 ofdm_index[2];
 	s8 cck_index = 0;
-	u8 ofdm_index_old[2];
+	u8 ofdm_index_old[2] = {0, 0};
 	s8 cck_index_old = 0;
 	u8 index;
 	int i;
diff --git a/drivers/net/wireless/rtlwifi/rtl8723ae/sw.c b/drivers/net/wireless/rtlwifi/rtl8723ae/sw.c
index e4c4cdc..d9ee2ef 100644
--- a/drivers/net/wireless/rtlwifi/rtl8723ae/sw.c
+++ b/drivers/net/wireless/rtlwifi/rtl8723ae/sw.c
@@ -251,7 +251,7 @@ static struct rtl_hal_cfg rtl8723ae_hal_cfg = {
 	.bar_id = 2,
 	.write_readback = true,
 	.name = "rtl8723ae_pci",
-	.fw_name = "rtlwifi/rtl8723aefw.bin",
+	.fw_name = "rtlwifi/rtl8723fw.bin",
 	.ops = &rtl8723ae_hal_ops,
 	.mod_params = &rtl8723ae_mod_params,
 	.maps[SYS_ISO_CTRL] = REG_SYS_ISO_CTRL,
@@ -353,8 +353,8 @@ MODULE_AUTHOR("Realtek WlanFAE	<wlanfae@realtek.com>");
 MODULE_AUTHOR("Larry Finger	<Larry.Finger@lwfinger.net>");
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("Realtek 8723E 802.11n PCI wireless");
-MODULE_FIRMWARE("rtlwifi/rtl8723aefw.bin");
-MODULE_FIRMWARE("rtlwifi/rtl8723aefw_B.bin");
+MODULE_FIRMWARE("rtlwifi/rtl8723fw.bin");
+MODULE_FIRMWARE("rtlwifi/rtl8723fw_B.bin");
 
 module_param_named(swenc, rtl8723ae_mod_params.sw_crypto, bool, 0444);
 module_param_named(debug, rtl8723ae_mod_params.debug, int, 0444);
diff --git a/drivers/net/wireless/ti/wl1251/spi.c b/drivers/net/wireless/ti/wl1251/spi.c
index 4c67c2f..c7dc6fe 100644
--- a/drivers/net/wireless/ti/wl1251/spi.c
+++ b/drivers/net/wireless/ti/wl1251/spi.c
@@ -93,8 +93,7 @@ static void wl1251_spi_wake(struct wl1251 *wl)
 	memset(&t, 0, sizeof(t));
 	spi_message_init(&m);
 
-	/*
-	 * Set WSPI_INIT_COMMAND
+	/* Set WSPI_INIT_COMMAND
 	 * the data is being send from the MSB to LSB
 	 */
 	cmd[2] = 0xff;
@@ -262,7 +261,8 @@ static int wl1251_spi_probe(struct spi_device *spi)
 	wl->if_ops = &wl1251_spi_ops;
 
 	/* This is the only SPI value that we need to set here, the rest
-	 * comes from the board-peripherals file */
+	 * comes from the board-peripherals file
+	 */
 	spi->bits_per_word = 32;
 
 	ret = spi_setup(spi);
@@ -329,29 +329,7 @@ static struct spi_driver wl1251_spi_driver = {
 	.remove		= wl1251_spi_remove,
 };
 
-static int __init wl1251_spi_init(void)
-{
-	int ret;
-
-	ret = spi_register_driver(&wl1251_spi_driver);
-	if (ret < 0) {
-		wl1251_error("failed to register spi driver: %d", ret);
-		goto out;
-	}
-
-out:
-	return ret;
-}
-
-static void __exit wl1251_spi_exit(void)
-{
-	spi_unregister_driver(&wl1251_spi_driver);
-
-	wl1251_notice("unloaded");
-}
-
-module_init(wl1251_spi_init);
-module_exit(wl1251_spi_exit);
+module_spi_driver(wl1251_spi_driver);
 
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Kalle Valo <kvalo@adurom.com>");
diff --git a/drivers/net/wireless/ti/wl18xx/main.c b/drivers/net/wireless/ti/wl18xx/main.c
index 9fa692d..7aa0eb8 100644
--- a/drivers/net/wireless/ti/wl18xx/main.c
+++ b/drivers/net/wireless/ti/wl18xx/main.c
@@ -23,6 +23,7 @@
 #include <linux/platform_device.h>
 #include <linux/ip.h>
 #include <linux/firmware.h>
+#include <linux/etherdevice.h>
 
 #include "../wlcore/wlcore.h"
 #include "../wlcore/debug.h"
@@ -594,8 +595,8 @@ static const struct wlcore_partition_set wl18xx_ptable[PART_TABLE_LEN] = {
 		.mem3 = { .start = 0x00000000, .size  = 0x00000000 },
 	},
 	[PART_PHY_INIT] = {
-		.mem  = { .start = 0x80926000,
-			  .size = sizeof(struct wl18xx_mac_and_phy_params) },
+		.mem  = { .start = WL18XX_PHY_INIT_MEM_ADDR,
+			  .size  = WL18XX_PHY_INIT_MEM_SIZE },
 		.reg  = { .start = 0x00000000, .size = 0x00000000 },
 		.mem2 = { .start = 0x00000000, .size = 0x00000000 },
 		.mem3 = { .start = 0x00000000, .size = 0x00000000 },
@@ -799,6 +800,9 @@ static int wl18xx_pre_upload(struct wl1271 *wl)
 	u32 tmp;
 	int ret;
 
+	BUILD_BUG_ON(sizeof(struct wl18xx_mac_and_phy_params) >
+		WL18XX_PHY_INIT_MEM_SIZE);
+
 	ret = wlcore_set_partition(wl, &wl->ptable[PART_BOOT]);
 	if (ret < 0)
 		goto out;
@@ -815,6 +819,35 @@ static int wl18xx_pre_upload(struct wl1271 *wl)
 	wl1271_debug(DEBUG_BOOT, "chip id 0x%x", tmp);
 
 	ret = wlcore_read32(wl, WL18XX_SCR_PAD2, &tmp);
+	if (ret < 0)
+		goto out;
+
+	/*
+	 * Workaround for FDSP code RAM corruption (needed for PG2.1
+	 * and newer; for older chips it's a NOP).  Change FDSP clock
+	 * settings so that it's muxed to the ATGP clock instead of
+	 * its own clock.
+	 */
+
+	ret = wlcore_set_partition(wl, &wl->ptable[PART_PHY_INIT]);
+	if (ret < 0)
+		goto out;
+
+	/* disable FDSP clock */
+	ret = wlcore_write32(wl, WL18XX_PHY_FPGA_SPARE_1,
+			     MEM_FDSP_CLK_120_DISABLE);
+	if (ret < 0)
+		goto out;
+
+	/* set ATPG clock toward FDSP Code RAM rather than its own clock */
+	ret = wlcore_write32(wl, WL18XX_PHY_FPGA_SPARE_1,
+			     MEM_FDSP_CODERAM_FUNC_CLK_SEL);
+	if (ret < 0)
+		goto out;
+
+	/* re-enable FDSP clock */
+	ret = wlcore_write32(wl, WL18XX_PHY_FPGA_SPARE_1,
+			     MEM_FDSP_CLK_120_ENABLE);
 
 out:
 	return ret;
@@ -1286,6 +1319,16 @@ static int wl18xx_get_mac(struct wl1271 *wl)
 		((mac1 & 0xff000000) >> 24);
 	wl->fuse_nic_addr = (mac1 & 0xffffff);
 
+	if (!wl->fuse_oui_addr && !wl->fuse_nic_addr) {
+		u8 mac[ETH_ALEN];
+
+		eth_random_addr(mac);
+
+		wl->fuse_oui_addr = (mac[0] << 16) + (mac[1] << 8) + mac[2];
+		wl->fuse_nic_addr = (mac[3] << 16) + (mac[4] << 8) + mac[5];
+		wl1271_warning("MAC address from fuse not available, using random locally administered addresses.");
+	}
+
 	ret = wlcore_set_partition(wl, &wl->ptable[PART_DOWN]);
 
 out:
diff --git a/drivers/net/wireless/ti/wl18xx/reg.h b/drivers/net/wireless/ti/wl18xx/reg.h
index 6306e04..05dd8ba 100644
--- a/drivers/net/wireless/ti/wl18xx/reg.h
+++ b/drivers/net/wireless/ti/wl18xx/reg.h
@@ -38,6 +38,9 @@
 #define WL18XX_REG_BOOT_PART_SIZE  0x00014578
 
 #define WL18XX_PHY_INIT_MEM_ADDR   0x80926000
+#define WL18XX_PHY_END_MEM_ADDR	   0x8093CA44
+#define WL18XX_PHY_INIT_MEM_SIZE \
+	(WL18XX_PHY_END_MEM_ADDR - WL18XX_PHY_INIT_MEM_ADDR)
 
 #define WL18XX_SDIO_WSPI_BASE		(WL18XX_REGISTERS_BASE)
 #define WL18XX_REG_CONFIG_BASE		(WL18XX_REGISTERS_BASE + 0x02000)
@@ -217,4 +220,16 @@ static const char * const rdl_names[] = {
 	[RDL_4_SP]	= "1897 MIMO",
 };
 
+/* FPGA_SPARE_1 register - used to change the PHY ATPG clock at boot time */
+#define WL18XX_PHY_FPGA_SPARE_1		0x8093CA40
+
+/* command to disable FDSP clock */
+#define MEM_FDSP_CLK_120_DISABLE        0x80000000
+
+/* command to set ATPG clock toward FDSP Code RAM rather than its own clock */
+#define MEM_FDSP_CODERAM_FUNC_CLK_SEL	0xC0000000
+
+/* command to re-enable FDSP clock */
+#define MEM_FDSP_CLK_120_ENABLE		0x40000000
+
 #endif /* __REG_H__ */
diff --git a/drivers/net/wireless/ti/wlcore/Makefile b/drivers/net/wireless/ti/wlcore/Makefile
index b21398f..4f23931 100644
--- a/drivers/net/wireless/ti/wlcore/Makefile
+++ b/drivers/net/wireless/ti/wlcore/Makefile
@@ -1,5 +1,5 @@
 wlcore-objs		= main.o cmd.o io.o event.o tx.o rx.o ps.o acx.o \
-			  boot.o init.o debugfs.o scan.o
+			  boot.o init.o debugfs.o scan.o sysfs.o
 
 wlcore_spi-objs 	= spi.o
 wlcore_sdio-objs	= sdio.o
diff --git a/drivers/net/wireless/ti/wlcore/main.c b/drivers/net/wireless/ti/wlcore/main.c
index 953111a..b8db55c 100644
--- a/drivers/net/wireless/ti/wlcore/main.c
+++ b/drivers/net/wireless/ti/wlcore/main.c
@@ -1,10 +1,9 @@
 
 /*
- * This file is part of wl1271
+ * This file is part of wlcore
  *
  * Copyright (C) 2008-2010 Nokia Corporation
- *
- * Contact: Luciano Coelho <luciano.coelho@nokia.com>
+ * Copyright (C) 2011-2013 Texas Instruments Inc.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License
@@ -24,34 +23,23 @@
 
 #include <linux/module.h>
 #include <linux/firmware.h>
-#include <linux/delay.h>
-#include <linux/spi/spi.h>
-#include <linux/crc32.h>
 #include <linux/etherdevice.h>
 #include <linux/vmalloc.h>
-#include <linux/platform_device.h>
-#include <linux/slab.h>
 #include <linux/wl12xx.h>
-#include <linux/sched.h>
 #include <linux/interrupt.h>
 
 #include "wlcore.h"
 #include "debug.h"
 #include "wl12xx_80211.h"
 #include "io.h"
-#include "event.h"
 #include "tx.h"
-#include "rx.h"
 #include "ps.h"
 #include "init.h"
 #include "debugfs.h"
-#include "cmd.h"
-#include "boot.h"
 #include "testmode.h"
 #include "scan.h"
 #include "hw_ops.h"
-
-#define WL1271_BOOT_RETRIES 3
+#include "sysfs.h"
 
 #define WL1271_BOOT_RETRIES 3
 
@@ -65,8 +53,7 @@ static void __wl1271_op_remove_interface(struct wl1271 *wl,
 static void wlcore_op_stop_locked(struct wl1271 *wl);
 static void wl1271_free_ap_keys(struct wl1271 *wl, struct wl12xx_vif *wlvif);
 
-static int wl12xx_set_authorized(struct wl1271 *wl,
-				 struct wl12xx_vif *wlvif)
+static int wl12xx_set_authorized(struct wl1271 *wl, struct wl12xx_vif *wlvif)
 {
 	int ret;
 
@@ -983,7 +970,7 @@ static int wlcore_fw_wakeup(struct wl1271 *wl)
 
 static int wl1271_setup(struct wl1271 *wl)
 {
-	wl->fw_status_1 = kmalloc(WLCORE_FW_STATUS_1_LEN(wl->num_rx_desc) +
+	wl->fw_status_1 = kzalloc(WLCORE_FW_STATUS_1_LEN(wl->num_rx_desc) +
 				  sizeof(*wl->fw_status_2) +
 				  wl->fw_status_priv_len, GFP_KERNEL);
 	if (!wl->fw_status_1)
@@ -993,7 +980,7 @@ static int wl1271_setup(struct wl1271 *wl)
 				(((u8 *) wl->fw_status_1) +
 				WLCORE_FW_STATUS_1_LEN(wl->num_rx_desc));
 
-	wl->tx_res_if = kmalloc(sizeof(*wl->tx_res_if), GFP_KERNEL);
+	wl->tx_res_if = kzalloc(sizeof(*wl->tx_res_if), GFP_KERNEL);
 	if (!wl->tx_res_if) {
 		kfree(wl->fw_status_1);
 		return -ENOMEM;
@@ -1668,8 +1655,7 @@ static int wl1271_configure_suspend(struct wl1271 *wl,
 	return 0;
 }
 
-static void wl1271_configure_resume(struct wl1271 *wl,
-				    struct wl12xx_vif *wlvif)
+static void wl1271_configure_resume(struct wl1271 *wl, struct wl12xx_vif *wlvif)
 {
 	int ret = 0;
 	bool is_ap = wlvif->bss_type == BSS_TYPE_AP_BSS;
@@ -2603,6 +2589,7 @@ unlock:
 	cancel_work_sync(&wlvif->rx_streaming_enable_work);
 	cancel_work_sync(&wlvif->rx_streaming_disable_work);
 	cancel_delayed_work_sync(&wlvif->connection_loss_work);
+	cancel_delayed_work_sync(&wlvif->channel_switch_work);
 
 	mutex_lock(&wl->mutex);
 }
@@ -3210,14 +3197,6 @@ static int wl1271_set_key(struct wl1271 *wl, struct wl12xx_vif *wlvif,
 		if (ret < 0)
 			return ret;
 
-		/* the default WEP key needs to be configured at least once */
-		if (key_type == KEY_WEP) {
-			ret = wl12xx_cmd_set_default_wep_key(wl,
-							wlvif->default_key,
-							wlvif->sta.hlid);
-			if (ret < 0)
-				return ret;
-		}
 	}
 
 	return 0;
@@ -3374,6 +3353,46 @@ int wlcore_set_key(struct wl1271 *wl, enum set_key_cmd cmd,
 }
 EXPORT_SYMBOL_GPL(wlcore_set_key);
 
+static void wl1271_op_set_default_key_idx(struct ieee80211_hw *hw,
+					  struct ieee80211_vif *vif,
+					  int key_idx)
+{
+	struct wl1271 *wl = hw->priv;
+	struct wl12xx_vif *wlvif = wl12xx_vif_to_data(vif);
+	int ret;
+
+	wl1271_debug(DEBUG_MAC80211, "mac80211 set default key idx %d",
+		     key_idx);
+
+	mutex_lock(&wl->mutex);
+
+	if (unlikely(wl->state != WLCORE_STATE_ON)) {
+		ret = -EAGAIN;
+		goto out_unlock;
+	}
+
+	ret = wl1271_ps_elp_wakeup(wl);
+	if (ret < 0)
+		goto out_unlock;
+
+	wlvif->default_key = key_idx;
+
+	/* the default WEP key needs to be configured at least once */
+	if (wlvif->encryption_type == KEY_WEP) {
+		ret = wl12xx_cmd_set_default_wep_key(wl,
+				key_idx,
+				wlvif->sta.hlid);
+		if (ret < 0)
+			goto out_sleep;
+	}
+
+out_sleep:
+	wl1271_ps_elp_sleep(wl);
+
+out_unlock:
+	mutex_unlock(&wl->mutex);
+}
+
 void wlcore_regdomain_config(struct wl1271 *wl)
 {
 	int ret;
@@ -3782,8 +3801,7 @@ static int wlcore_set_beacon_template(struct wl1271 *wl,
 	struct ieee80211_hdr *hdr;
 	u32 min_rate;
 	int ret;
-	int ieoffset = offsetof(struct ieee80211_mgmt,
-				u.beacon.variable);
+	int ieoffset = offsetof(struct ieee80211_mgmt, u.beacon.variable);
 	struct sk_buff *beacon = ieee80211_beacon_get(wl->hw, vif);
 	u16 tmpl_id;
 
@@ -4230,8 +4248,7 @@ static void wl1271_bss_info_changed_sta(struct wl1271 *wl,
 	}
 
 	/* Handle new association with HT. Do this after join. */
-	if (sta_exists &&
-	    (changed & BSS_CHANGED_HT)) {
+	if (sta_exists) {
 		bool enabled =
 			bss_conf->chandef.width != NL80211_CHAN_WIDTH_20_NOHT;
 
@@ -5368,6 +5385,7 @@ static const struct ieee80211_ops wl1271_ops = {
 	.ampdu_action = wl1271_op_ampdu_action,
 	.tx_frames_pending = wl1271_tx_frames_pending,
 	.set_bitrate_mask = wl12xx_set_bitrate_mask,
+	.set_default_unicast_key = wl1271_op_set_default_key_idx,
 	.channel_switch = wl12xx_op_channel_switch,
 	.flush = wlcore_op_flush,
 	.remain_on_channel = wlcore_op_remain_on_channel,
@@ -5403,151 +5421,6 @@ u8 wlcore_rate_to_idx(struct wl1271 *wl, u8 rate, enum ieee80211_band band)
 	return idx;
 }
 
-static ssize_t wl1271_sysfs_show_bt_coex_state(struct device *dev,
-					       struct device_attribute *attr,
-					       char *buf)
-{
-	struct wl1271 *wl = dev_get_drvdata(dev);
-	ssize_t len;
-
-	len = PAGE_SIZE;
-
-	mutex_lock(&wl->mutex);
-	len = snprintf(buf, len, "%d\n\n0 - off\n1 - on\n",
-		       wl->sg_enabled);
-	mutex_unlock(&wl->mutex);
-
-	return len;
-
-}
-
-static ssize_t wl1271_sysfs_store_bt_coex_state(struct device *dev,
-						struct device_attribute *attr,
-						const char *buf, size_t count)
-{
-	struct wl1271 *wl = dev_get_drvdata(dev);
-	unsigned long res;
-	int ret;
-
-	ret = kstrtoul(buf, 10, &res);
-	if (ret < 0) {
-		wl1271_warning("incorrect value written to bt_coex_mode");
-		return count;
-	}
-
-	mutex_lock(&wl->mutex);
-
-	res = !!res;
-
-	if (res == wl->sg_enabled)
-		goto out;
-
-	wl->sg_enabled = res;
-
-	if (unlikely(wl->state != WLCORE_STATE_ON))
-		goto out;
-
-	ret = wl1271_ps_elp_wakeup(wl);
-	if (ret < 0)
-		goto out;
-
-	wl1271_acx_sg_enable(wl, wl->sg_enabled);
-	wl1271_ps_elp_sleep(wl);
-
- out:
-	mutex_unlock(&wl->mutex);
-	return count;
-}
-
-static DEVICE_ATTR(bt_coex_state, S_IRUGO | S_IWUSR,
-		   wl1271_sysfs_show_bt_coex_state,
-		   wl1271_sysfs_store_bt_coex_state);
-
-static ssize_t wl1271_sysfs_show_hw_pg_ver(struct device *dev,
-					   struct device_attribute *attr,
-					   char *buf)
-{
-	struct wl1271 *wl = dev_get_drvdata(dev);
-	ssize_t len;
-
-	len = PAGE_SIZE;
-
-	mutex_lock(&wl->mutex);
-	if (wl->hw_pg_ver >= 0)
-		len = snprintf(buf, len, "%d\n", wl->hw_pg_ver);
-	else
-		len = snprintf(buf, len, "n/a\n");
-	mutex_unlock(&wl->mutex);
-
-	return len;
-}
-
-static DEVICE_ATTR(hw_pg_ver, S_IRUGO,
-		   wl1271_sysfs_show_hw_pg_ver, NULL);
-
-static ssize_t wl1271_sysfs_read_fwlog(struct file *filp, struct kobject *kobj,
-				       struct bin_attribute *bin_attr,
-				       char *buffer, loff_t pos, size_t count)
-{
-	struct device *dev = container_of(kobj, struct device, kobj);
-	struct wl1271 *wl = dev_get_drvdata(dev);
-	ssize_t len;
-	int ret;
-
-	ret = mutex_lock_interruptible(&wl->mutex);
-	if (ret < 0)
-		return -ERESTARTSYS;
-
-	/* Let only one thread read the log at a time, blocking others */
-	while (wl->fwlog_size == 0) {
-		DEFINE_WAIT(wait);
-
-		prepare_to_wait_exclusive(&wl->fwlog_waitq,
-					  &wait,
-					  TASK_INTERRUPTIBLE);
-
-		if (wl->fwlog_size != 0) {
-			finish_wait(&wl->fwlog_waitq, &wait);
-			break;
-		}
-
-		mutex_unlock(&wl->mutex);
-
-		schedule();
-		finish_wait(&wl->fwlog_waitq, &wait);
-
-		if (signal_pending(current))
-			return -ERESTARTSYS;
-
-		ret = mutex_lock_interruptible(&wl->mutex);
-		if (ret < 0)
-			return -ERESTARTSYS;
-	}
-
-	/* Check if the fwlog is still valid */
-	if (wl->fwlog_size < 0) {
-		mutex_unlock(&wl->mutex);
-		return 0;
-	}
-
-	/* Seeking is not supported - old logs are not kept. Disregard pos. */
-	len = min(count, (size_t)wl->fwlog_size);
-	wl->fwlog_size -= len;
-	memcpy(buffer, wl->fwlog, len);
-
-	/* Make room for new messages */
-	memmove(wl->fwlog, wl->fwlog + len, wl->fwlog_size);
-
-	mutex_unlock(&wl->mutex);
-
-	return len;
-}
-
-static struct bin_attribute fwlog_attr = {
-	.attr = {.name = "fwlog", .mode = S_IRUSR},
-	.read = wl1271_sysfs_read_fwlog,
-};
-
 static void wl12xx_derive_mac_addresses(struct wl1271 *wl, u32 oui, u32 nic)
 {
 	int i;
@@ -5827,8 +5700,6 @@ static int wl1271_init_ieee80211(struct wl1271 *wl)
 	return 0;
 }
 
-#define WL1271_DEFAULT_CHANNEL 0
-
 struct ieee80211_hw *wlcore_alloc_hw(size_t priv_size, u32 aggr_buf_size,
 				     u32 mbox_size)
 {
@@ -5881,7 +5752,7 @@ struct ieee80211_hw *wlcore_alloc_hw(size_t priv_size, u32 aggr_buf_size,
 		goto err_hw;
 	}
 
-	wl->channel = WL1271_DEFAULT_CHANNEL;
+	wl->channel = 0;
 	wl->rx_counter = 0;
 	wl->power_level = WL1271_DEFAULT_POWER_LEVEL;
 	wl->band = IEEE80211_BAND_2GHZ;
@@ -5988,11 +5859,8 @@ int wlcore_free_hw(struct wl1271 *wl)
 	wake_up_interruptible_all(&wl->fwlog_waitq);
 	mutex_unlock(&wl->mutex);
 
-	device_remove_bin_file(wl->dev, &fwlog_attr);
-
-	device_remove_file(wl->dev, &dev_attr_hw_pg_ver);
+	wlcore_sysfs_free(wl);
 
-	device_remove_file(wl->dev, &dev_attr_bt_coex_state);
 	kfree(wl->buffer_32);
 	kfree(wl->mbox);
 	free_page((unsigned long)wl->fwlog);
@@ -6018,6 +5886,15 @@ int wlcore_free_hw(struct wl1271 *wl)
 }
 EXPORT_SYMBOL_GPL(wlcore_free_hw);
 
+#ifdef CONFIG_PM
+static const struct wiphy_wowlan_support wlcore_wowlan_support = {
+	.flags = WIPHY_WOWLAN_ANY,
+	.n_patterns = WL1271_MAX_RX_FILTERS,
+	.pattern_min_len = 1,
+	.pattern_max_len = WL1271_RX_FILTER_MAX_PATTERN_SIZE,
+};
+#endif
+
 static void wlcore_nvs_cb(const struct firmware *fw, void *context)
 {
 	struct wl1271 *wl = context;
@@ -6071,14 +5948,8 @@ static void wlcore_nvs_cb(const struct firmware *fw, void *context)
 	if (!ret) {
 		wl->irq_wake_enabled = true;
 		device_init_wakeup(wl->dev, 1);
-		if (pdata->pwr_in_suspend) {
-			wl->hw->wiphy->wowlan.flags = WIPHY_WOWLAN_ANY;
-			wl->hw->wiphy->wowlan.n_patterns =
-				WL1271_MAX_RX_FILTERS;
-			wl->hw->wiphy->wowlan.pattern_min_len = 1;
-			wl->hw->wiphy->wowlan.pattern_max_len =
-				WL1271_RX_FILTER_MAX_PATTERN_SIZE;
-		}
+		if (pdata->pwr_in_suspend)
+			wl->hw->wiphy->wowlan = &wlcore_wowlan_support;
 	}
 #endif
 	disable_irq(wl->irq);
@@ -6101,36 +5972,13 @@ static void wlcore_nvs_cb(const struct firmware *fw, void *context)
 	if (ret)
 		goto out_irq;
 
-	/* Create sysfs file to control bt coex state */
-	ret = device_create_file(wl->dev, &dev_attr_bt_coex_state);
-	if (ret < 0) {
-		wl1271_error("failed to create sysfs file bt_coex_state");
+	ret = wlcore_sysfs_init(wl);
+	if (ret)
 		goto out_unreg;
-	}
-
-	/* Create sysfs file to get HW PG version */
-	ret = device_create_file(wl->dev, &dev_attr_hw_pg_ver);
-	if (ret < 0) {
-		wl1271_error("failed to create sysfs file hw_pg_ver");
-		goto out_bt_coex_state;
-	}
-
-	/* Create sysfs file for the FW log */
-	ret = device_create_bin_file(wl->dev, &fwlog_attr);
-	if (ret < 0) {
-		wl1271_error("failed to create sysfs file fwlog");
-		goto out_hw_pg_ver;
-	}
 
 	wl->initialized = true;
 	goto out;
 
-out_hw_pg_ver:
-	device_remove_file(wl->dev, &dev_attr_hw_pg_ver);
-
-out_bt_coex_state:
-	device_remove_file(wl->dev, &dev_attr_bt_coex_state);
-
 out_unreg:
 	wl1271_unregister_hw(wl);
 
diff --git a/drivers/net/wireless/ti/wlcore/ps.c b/drivers/net/wireless/ti/wlcore/ps.c
index 9654577..98066d4 100644
--- a/drivers/net/wireless/ti/wlcore/ps.c
+++ b/drivers/net/wireless/ti/wlcore/ps.c
@@ -110,7 +110,7 @@ int wl1271_ps_elp_wakeup(struct wl1271 *wl)
 	DECLARE_COMPLETION_ONSTACK(compl);
 	unsigned long flags;
 	int ret;
-	u32 start_time = jiffies;
+	unsigned long start_time = jiffies;
 	bool pending = false;
 
 	/*
diff --git a/drivers/net/wireless/ti/wlcore/spi.c b/drivers/net/wireless/ti/wlcore/spi.c
index e264478..1b0cd98 100644
--- a/drivers/net/wireless/ti/wlcore/spi.c
+++ b/drivers/net/wireless/ti/wlcore/spi.c
@@ -434,19 +434,7 @@ static struct spi_driver wl1271_spi_driver = {
 	.remove		= wl1271_remove,
 };
 
-static int __init wl1271_init(void)
-{
-	return spi_register_driver(&wl1271_spi_driver);
-}
-
-static void __exit wl1271_exit(void)
-{
-	spi_unregister_driver(&wl1271_spi_driver);
-}
-
-module_init(wl1271_init);
-module_exit(wl1271_exit);
-
+module_spi_driver(wl1271_spi_driver);
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Luciano Coelho <coelho@ti.com>");
 MODULE_AUTHOR("Juuso Oikarinen <juuso.oikarinen@nokia.com>");
diff --git a/drivers/net/wireless/ti/wlcore/sysfs.c b/drivers/net/wireless/ti/wlcore/sysfs.c
new file mode 100644
index 0000000..8e58349
--- /dev/null
+++ b/drivers/net/wireless/ti/wlcore/sysfs.c
@@ -0,0 +1,216 @@
+/*
+ * This file is part of wlcore
+ *
+ * Copyright (C) 2013 Texas Instruments Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
+ * 02110-1301 USA
+ *
+ */
+
+#include "wlcore.h"
+#include "debug.h"
+#include "ps.h"
+#include "sysfs.h"
+
+static ssize_t wl1271_sysfs_show_bt_coex_state(struct device *dev,
+					       struct device_attribute *attr,
+					       char *buf)
+{
+	struct wl1271 *wl = dev_get_drvdata(dev);
+	ssize_t len;
+
+	len = PAGE_SIZE;
+
+	mutex_lock(&wl->mutex);
+	len = snprintf(buf, len, "%d\n\n0 - off\n1 - on\n",
+		       wl->sg_enabled);
+	mutex_unlock(&wl->mutex);
+
+	return len;
+
+}
+
+static ssize_t wl1271_sysfs_store_bt_coex_state(struct device *dev,
+						struct device_attribute *attr,
+						const char *buf, size_t count)
+{
+	struct wl1271 *wl = dev_get_drvdata(dev);
+	unsigned long res;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &res);
+	if (ret < 0) {
+		wl1271_warning("incorrect value written to bt_coex_mode");
+		return count;
+	}
+
+	mutex_lock(&wl->mutex);
+
+	res = !!res;
+
+	if (res == wl->sg_enabled)
+		goto out;
+
+	wl->sg_enabled = res;
+
+	if (unlikely(wl->state != WLCORE_STATE_ON))
+		goto out;
+
+	ret = wl1271_ps_elp_wakeup(wl);
+	if (ret < 0)
+		goto out;
+
+	wl1271_acx_sg_enable(wl, wl->sg_enabled);
+	wl1271_ps_elp_sleep(wl);
+
+ out:
+	mutex_unlock(&wl->mutex);
+	return count;
+}
+
+static DEVICE_ATTR(bt_coex_state, S_IRUGO | S_IWUSR,
+		   wl1271_sysfs_show_bt_coex_state,
+		   wl1271_sysfs_store_bt_coex_state);
+
+static ssize_t wl1271_sysfs_show_hw_pg_ver(struct device *dev,
+					   struct device_attribute *attr,
+					   char *buf)
+{
+	struct wl1271 *wl = dev_get_drvdata(dev);
+	ssize_t len;
+
+	len = PAGE_SIZE;
+
+	mutex_lock(&wl->mutex);
+	if (wl->hw_pg_ver >= 0)
+		len = snprintf(buf, len, "%d\n", wl->hw_pg_ver);
+	else
+		len = snprintf(buf, len, "n/a\n");
+	mutex_unlock(&wl->mutex);
+
+	return len;
+}
+
+static DEVICE_ATTR(hw_pg_ver, S_IRUGO,
+		   wl1271_sysfs_show_hw_pg_ver, NULL);
+
+static ssize_t wl1271_sysfs_read_fwlog(struct file *filp, struct kobject *kobj,
+				       struct bin_attribute *bin_attr,
+				       char *buffer, loff_t pos, size_t count)
+{
+	struct device *dev = container_of(kobj, struct device, kobj);
+	struct wl1271 *wl = dev_get_drvdata(dev);
+	ssize_t len;
+	int ret;
+
+	ret = mutex_lock_interruptible(&wl->mutex);
+	if (ret < 0)
+		return -ERESTARTSYS;
+
+	/* Let only one thread read the log at a time, blocking others */
+	while (wl->fwlog_size == 0) {
+		DEFINE_WAIT(wait);
+
+		prepare_to_wait_exclusive(&wl->fwlog_waitq,
+					  &wait,
+					  TASK_INTERRUPTIBLE);
+
+		if (wl->fwlog_size != 0) {
+			finish_wait(&wl->fwlog_waitq, &wait);
+			break;
+		}
+
+		mutex_unlock(&wl->mutex);
+
+		schedule();
+		finish_wait(&wl->fwlog_waitq, &wait);
+
+		if (signal_pending(current))
+			return -ERESTARTSYS;
+
+		ret = mutex_lock_interruptible(&wl->mutex);
+		if (ret < 0)
+			return -ERESTARTSYS;
+	}
+
+	/* Check if the fwlog is still valid */
+	if (wl->fwlog_size < 0) {
+		mutex_unlock(&wl->mutex);
+		return 0;
+	}
+
+	/* Seeking is not supported - old logs are not kept. Disregard pos. */
+	len = min(count, (size_t)wl->fwlog_size);
+	wl->fwlog_size -= len;
+	memcpy(buffer, wl->fwlog, len);
+
+	/* Make room for new messages */
+	memmove(wl->fwlog, wl->fwlog + len, wl->fwlog_size);
+
+	mutex_unlock(&wl->mutex);
+
+	return len;
+}
+
+static struct bin_attribute fwlog_attr = {
+	.attr = {.name = "fwlog", .mode = S_IRUSR},
+	.read = wl1271_sysfs_read_fwlog,
+};
+
+int wlcore_sysfs_init(struct wl1271 *wl)
+{
+	int ret;
+
+	/* Create sysfs file to control bt coex state */
+	ret = device_create_file(wl->dev, &dev_attr_bt_coex_state);
+	if (ret < 0) {
+		wl1271_error("failed to create sysfs file bt_coex_state");
+		goto out;
+	}
+
+	/* Create sysfs file to get HW PG version */
+	ret = device_create_file(wl->dev, &dev_attr_hw_pg_ver);
+	if (ret < 0) {
+		wl1271_error("failed to create sysfs file hw_pg_ver");
+		goto out_bt_coex_state;
+	}
+
+	/* Create sysfs file for the FW log */
+	ret = device_create_bin_file(wl->dev, &fwlog_attr);
+	if (ret < 0) {
+		wl1271_error("failed to create sysfs file fwlog");
+		goto out_hw_pg_ver;
+	}
+
+	goto out;
+
+out_hw_pg_ver:
+	device_remove_file(wl->dev, &dev_attr_hw_pg_ver);
+
+out_bt_coex_state:
+	device_remove_file(wl->dev, &dev_attr_bt_coex_state);
+
+out:
+	return ret;
+}
+
+void wlcore_sysfs_free(struct wl1271 *wl)
+{
+	device_remove_bin_file(wl->dev, &fwlog_attr);
+
+	device_remove_file(wl->dev, &dev_attr_hw_pg_ver);
+
+	device_remove_file(wl->dev, &dev_attr_bt_coex_state);
+}
diff --git a/drivers/net/wireless/ti/wlcore/sysfs.h b/drivers/net/wireless/ti/wlcore/sysfs.h
new file mode 100644
index 0000000..c148892
--- /dev/null
+++ b/drivers/net/wireless/ti/wlcore/sysfs.h
@@ -0,0 +1,28 @@
+/*
+ * This file is part of wlcore
+ *
+ * Copyright (C) 2013 Texas Instruments Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
+ * 02110-1301 USA
+ *
+ */
+
+#ifndef __SYSFS_H__
+#define __SYSFS_H__
+
+int wlcore_sysfs_init(struct wl1271 *wl);
+void wlcore_sysfs_free(struct wl1271 *wl);
+
+#endif
diff --git a/drivers/net/wireless/ti/wlcore/tx.c b/drivers/net/wireless/ti/wlcore/tx.c
index 004d02e..7e93fe6 100644
--- a/drivers/net/wireless/ti/wlcore/tx.c
+++ b/drivers/net/wireless/ti/wlcore/tx.c
@@ -386,7 +386,7 @@ static int wl1271_prepare_tx_frame(struct wl1271 *wl, struct wl12xx_vif *wlvif,
 		is_wep = (cipher == WLAN_CIPHER_SUITE_WEP40) ||
 			 (cipher == WLAN_CIPHER_SUITE_WEP104);
 
-		if (unlikely(is_wep && wlvif->default_key != idx)) {
+		if (WARN_ON(is_wep && wlvif->default_key != idx)) {
 			ret = wl1271_set_default_wep_key(wl, wlvif, idx);
 			if (ret < 0)
 				return ret;
diff --git a/drivers/net/xen-netback/common.h b/drivers/net/xen-netback/common.h
index 9d7f172..8a4d77e 100644
--- a/drivers/net/xen-netback/common.h
+++ b/drivers/net/xen-netback/common.h
@@ -57,8 +57,12 @@ struct xenvif {
 
 	u8               fe_dev_addr[6];
 
-	/* Physical parameters of the comms window. */
-	unsigned int     irq;
+	/* When feature-split-event-channels = 0, tx_irq = rx_irq. */
+	unsigned int tx_irq;
+	unsigned int rx_irq;
+	/* Only used when feature-split-event-channels = 1 */
+	char tx_irq_name[IFNAMSIZ+4]; /* DEVNAME-tx */
+	char rx_irq_name[IFNAMSIZ+4]; /* DEVNAME-rx */
 
 	/* List of frontends to notify after a batch of frames sent. */
 	struct list_head notify_list;
@@ -113,13 +117,15 @@ struct xenvif *xenvif_alloc(struct device *parent,
 			    unsigned int handle);
 
 int xenvif_connect(struct xenvif *vif, unsigned long tx_ring_ref,
-		   unsigned long rx_ring_ref, unsigned int evtchn);
+		   unsigned long rx_ring_ref, unsigned int tx_evtchn,
+		   unsigned int rx_evtchn);
 void xenvif_disconnect(struct xenvif *vif);
 
 void xenvif_get(struct xenvif *vif);
 void xenvif_put(struct xenvif *vif);
 
 int xenvif_xenbus_init(void);
+void xenvif_xenbus_fini(void);
 
 int xenvif_schedulable(struct xenvif *vif);
 
@@ -157,4 +163,6 @@ void xenvif_carrier_off(struct xenvif *vif);
 /* Returns number of ring slots required to send an skb to the frontend */
 unsigned int xen_netbk_count_skb_slots(struct xenvif *vif, struct sk_buff *skb);
 
+extern bool separate_tx_rx_irq;
+
 #endif /* __XEN_NETBACK__COMMON_H__ */
diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c
index d984141..087d2db 100644
--- a/drivers/net/xen-netback/interface.c
+++ b/drivers/net/xen-netback/interface.c
@@ -60,21 +60,39 @@ static int xenvif_rx_schedulable(struct xenvif *vif)
 	return xenvif_schedulable(vif) && !xen_netbk_rx_ring_full(vif);
 }
 
-static irqreturn_t xenvif_interrupt(int irq, void *dev_id)
+static irqreturn_t xenvif_tx_interrupt(int irq, void *dev_id)
 {
 	struct xenvif *vif = dev_id;
 
 	if (vif->netbk == NULL)
-		return IRQ_NONE;
+		return IRQ_HANDLED;
 
 	xen_netbk_schedule_xenvif(vif);
 
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t xenvif_rx_interrupt(int irq, void *dev_id)
+{
+	struct xenvif *vif = dev_id;
+
+	if (vif->netbk == NULL)
+		return IRQ_HANDLED;
+
 	if (xenvif_rx_schedulable(vif))
 		netif_wake_queue(vif->dev);
 
 	return IRQ_HANDLED;
 }
 
+static irqreturn_t xenvif_interrupt(int irq, void *dev_id)
+{
+	xenvif_tx_interrupt(irq, dev_id);
+	xenvif_rx_interrupt(irq, dev_id);
+
+	return IRQ_HANDLED;
+}
+
 static int xenvif_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct xenvif *vif = netdev_priv(dev);
@@ -125,13 +143,17 @@ static struct net_device_stats *xenvif_get_stats(struct net_device *dev)
 static void xenvif_up(struct xenvif *vif)
 {
 	xen_netbk_add_xenvif(vif);
-	enable_irq(vif->irq);
+	enable_irq(vif->tx_irq);
+	if (vif->tx_irq != vif->rx_irq)
+		enable_irq(vif->rx_irq);
 	xen_netbk_check_rx_xenvif(vif);
 }
 
 static void xenvif_down(struct xenvif *vif)
 {
-	disable_irq(vif->irq);
+	disable_irq(vif->tx_irq);
+	if (vif->tx_irq != vif->rx_irq)
+		disable_irq(vif->rx_irq);
 	del_timer_sync(&vif->credit_timeout);
 	xen_netbk_deschedule_xenvif(vif);
 	xen_netbk_remove_xenvif(vif);
@@ -308,25 +330,52 @@ struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
 }
 
 int xenvif_connect(struct xenvif *vif, unsigned long tx_ring_ref,
-		   unsigned long rx_ring_ref, unsigned int evtchn)
+		   unsigned long rx_ring_ref, unsigned int tx_evtchn,
+		   unsigned int rx_evtchn)
 {
 	int err = -ENOMEM;
 
 	/* Already connected through? */
-	if (vif->irq)
+	if (vif->tx_irq)
 		return 0;
 
+	__module_get(THIS_MODULE);
+
 	err = xen_netbk_map_frontend_rings(vif, tx_ring_ref, rx_ring_ref);
 	if (err < 0)
 		goto err;
 
-	err = bind_interdomain_evtchn_to_irqhandler(
-		vif->domid, evtchn, xenvif_interrupt, 0,
-		vif->dev->name, vif);
-	if (err < 0)
-		goto err_unmap;
-	vif->irq = err;
-	disable_irq(vif->irq);
+	if (tx_evtchn == rx_evtchn) {
+		/* feature-split-event-channels == 0 */
+		err = bind_interdomain_evtchn_to_irqhandler(
+			vif->domid, tx_evtchn, xenvif_interrupt, 0,
+			vif->dev->name, vif);
+		if (err < 0)
+			goto err_unmap;
+		vif->tx_irq = vif->rx_irq = err;
+		disable_irq(vif->tx_irq);
+	} else {
+		/* feature-split-event-channels == 1 */
+		snprintf(vif->tx_irq_name, sizeof(vif->tx_irq_name),
+			 "%s-tx", vif->dev->name);
+		err = bind_interdomain_evtchn_to_irqhandler(
+			vif->domid, tx_evtchn, xenvif_tx_interrupt, 0,
+			vif->tx_irq_name, vif);
+		if (err < 0)
+			goto err_unmap;
+		vif->tx_irq = err;
+		disable_irq(vif->tx_irq);
+
+		snprintf(vif->rx_irq_name, sizeof(vif->rx_irq_name),
+			 "%s-rx", vif->dev->name);
+		err = bind_interdomain_evtchn_to_irqhandler(
+			vif->domid, rx_evtchn, xenvif_rx_interrupt, 0,
+			vif->rx_irq_name, vif);
+		if (err < 0)
+			goto err_tx_unbind;
+		vif->rx_irq = err;
+		disable_irq(vif->rx_irq);
+	}
 
 	xenvif_get(vif);
 
@@ -340,9 +389,13 @@ int xenvif_connect(struct xenvif *vif, unsigned long tx_ring_ref,
 	rtnl_unlock();
 
 	return 0;
+err_tx_unbind:
+	unbind_from_irqhandler(vif->tx_irq, vif);
+	vif->tx_irq = 0;
 err_unmap:
 	xen_netbk_unmap_frontend_rings(vif);
 err:
+	module_put(THIS_MODULE);
 	return err;
 }
 
@@ -360,18 +413,37 @@ void xenvif_carrier_off(struct xenvif *vif)
 
 void xenvif_disconnect(struct xenvif *vif)
 {
+	/* Disconnect funtion might get called by generic framework
+	 * even before vif connects, so we need to check if we really
+	 * need to do a module_put.
+	 */
+	int need_module_put = 0;
+
 	if (netif_carrier_ok(vif->dev))
 		xenvif_carrier_off(vif);
 
 	atomic_dec(&vif->refcnt);
 	wait_event(vif->waiting_to_free, atomic_read(&vif->refcnt) == 0);
 
-	if (vif->irq)
-		unbind_from_irqhandler(vif->irq, vif);
+	if (vif->tx_irq) {
+		if (vif->tx_irq == vif->rx_irq)
+			unbind_from_irqhandler(vif->tx_irq, vif);
+		else {
+			unbind_from_irqhandler(vif->tx_irq, vif);
+			unbind_from_irqhandler(vif->rx_irq, vif);
+		}
+		/* vif->irq is valid, we had a module_get in
+		 * xenvif_connect.
+		 */
+		need_module_put = 1;
+	}
 
 	unregister_netdev(vif->dev);
 
 	xen_netbk_unmap_frontend_rings(vif);
 
 	free_netdev(vif->dev);
+
+	if (need_module_put)
+		module_put(THIS_MODULE);
 }
diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c
index 8c20935..64828de 100644
--- a/drivers/net/xen-netback/netback.c
+++ b/drivers/net/xen-netback/netback.c
@@ -47,6 +47,13 @@
 #include <asm/xen/hypercall.h>
 #include <asm/xen/page.h>
 
+/* Provide an option to disable split event channels at load time as
+ * event channels are limited resource. Split event channels are
+ * enabled by default.
+ */
+bool separate_tx_rx_irq = 1;
+module_param(separate_tx_rx_irq, bool, 0644);
+
 /*
  * This is the maximum slots a skb can have. If a guest sends a skb
  * which exceeds this limit it is considered malicious.
@@ -783,7 +790,7 @@ static void xen_netbk_rx_action(struct xen_netbk *netbk)
 	}
 
 	list_for_each_entry_safe(vif, tmp, &notify, notify_list) {
-		notify_remote_via_irq(vif->irq);
+		notify_remote_via_irq(vif->rx_irq);
 		list_del_init(&vif->notify_list);
 		xenvif_put(vif);
 	}
@@ -1763,7 +1770,7 @@ static void make_tx_response(struct xenvif *vif,
 	vif->tx.rsp_prod_pvt = ++i;
 	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&vif->tx, notify);
 	if (notify)
-		notify_remote_via_irq(vif->irq);
+		notify_remote_via_irq(vif->tx_irq);
 }
 
 static struct xen_netif_rx_response *make_rx_response(struct xenvif *vif,
@@ -1883,9 +1890,8 @@ static int __init netback_init(void)
 		return -ENODEV;
 
 	if (fatal_skb_slots < XEN_NETBK_LEGACY_SLOTS_MAX) {
-		printk(KERN_INFO
-		       "xen-netback: fatal_skb_slots too small (%d), bump it to XEN_NETBK_LEGACY_SLOTS_MAX (%d)\n",
-		       fatal_skb_slots, XEN_NETBK_LEGACY_SLOTS_MAX);
+		pr_info("fatal_skb_slots too small (%d), bump it to XEN_NETBK_LEGACY_SLOTS_MAX (%d)\n",
+			fatal_skb_slots, XEN_NETBK_LEGACY_SLOTS_MAX);
 		fatal_skb_slots = XEN_NETBK_LEGACY_SLOTS_MAX;
 	}
 
@@ -1914,7 +1920,7 @@ static int __init netback_init(void)
 					     "netback/%u", group);
 
 		if (IS_ERR(netbk->task)) {
-			printk(KERN_ALERT "kthread_create() fails at netback\n");
+			pr_alert("kthread_create() fails at netback\n");
 			del_timer(&netbk->net_timer);
 			rc = PTR_ERR(netbk->task);
 			goto failed_init;
@@ -1940,10 +1946,6 @@ static int __init netback_init(void)
 failed_init:
 	while (--group >= 0) {
 		struct xen_netbk *netbk = &xen_netbk[group];
-		for (i = 0; i < MAX_PENDING_REQS; i++) {
-			if (netbk->mmap_pages[i])
-				__free_page(netbk->mmap_pages[i]);
-		}
 		del_timer(&netbk->net_timer);
 		kthread_stop(netbk->task);
 	}
@@ -1954,5 +1956,25 @@ failed_init:
 
 module_init(netback_init);
 
+static void __exit netback_fini(void)
+{
+	int i, j;
+
+	xenvif_xenbus_fini();
+
+	for (i = 0; i < xen_netbk_group_nr; i++) {
+		struct xen_netbk *netbk = &xen_netbk[i];
+		del_timer_sync(&netbk->net_timer);
+		kthread_stop(netbk->task);
+		for (j = 0; j < MAX_PENDING_REQS; j++) {
+			if (netbk->mmap_pages[j])
+				__free_page(netbk->mmap_pages[j]);
+		}
+	}
+
+	vfree(xen_netbk);
+}
+module_exit(netback_fini);
+
 MODULE_LICENSE("Dual BSD/GPL");
 MODULE_ALIAS("xen-backend:vif");
diff --git a/drivers/net/xen-netback/xenbus.c b/drivers/net/xen-netback/xenbus.c
index 410018c..1fe48fe3 100644
--- a/drivers/net/xen-netback/xenbus.c
+++ b/drivers/net/xen-netback/xenbus.c
@@ -122,6 +122,16 @@ static int netback_probe(struct xenbus_device *dev,
 		goto fail;
 	}
 
+	/*
+	 * Split event channels support, this is optional so it is not
+	 * put inside the above loop.
+	 */
+	err = xenbus_printf(XBT_NIL, dev->nodename,
+			    "feature-split-event-channels",
+			    "%u", separate_tx_rx_irq);
+	if (err)
+		pr_debug("Error writing feature-split-event-channels\n");
+
 	err = xenbus_switch_state(dev, XenbusStateInitWait);
 	if (err)
 		goto fail;
@@ -135,7 +145,7 @@ abort_transaction:
 	xenbus_transaction_end(xbt, 1);
 	xenbus_dev_fatal(dev, err, "%s", message);
 fail:
-	pr_debug("failed");
+	pr_debug("failed\n");
 	netback_remove(dev);
 	return err;
 }
@@ -218,15 +228,14 @@ static void frontend_changed(struct xenbus_device *dev,
 {
 	struct backend_info *be = dev_get_drvdata(&dev->dev);
 
-	pr_debug("frontend state %s", xenbus_strstate(frontend_state));
+	pr_debug("frontend state %s\n", xenbus_strstate(frontend_state));
 
 	be->frontend_state = frontend_state;
 
 	switch (frontend_state) {
 	case XenbusStateInitialising:
 		if (dev->state == XenbusStateClosed) {
-			printk(KERN_INFO "%s: %s: prepare for reconnect\n",
-			       __func__, dev->nodename);
+			pr_info("%s: prepare for reconnect\n", dev->nodename);
 			xenbus_switch_state(dev, XenbusStateInitWait);
 		}
 		break;
@@ -393,21 +402,36 @@ static int connect_rings(struct backend_info *be)
 	struct xenvif *vif = be->vif;
 	struct xenbus_device *dev = be->dev;
 	unsigned long tx_ring_ref, rx_ring_ref;
-	unsigned int evtchn, rx_copy;
+	unsigned int tx_evtchn, rx_evtchn, rx_copy;
 	int err;
 	int val;
 
 	err = xenbus_gather(XBT_NIL, dev->otherend,
 			    "tx-ring-ref", "%lu", &tx_ring_ref,
-			    "rx-ring-ref", "%lu", &rx_ring_ref,
-			    "event-channel", "%u", &evtchn, NULL);
+			    "rx-ring-ref", "%lu", &rx_ring_ref, NULL);
 	if (err) {
 		xenbus_dev_fatal(dev, err,
-				 "reading %s/ring-ref and event-channel",
+				 "reading %s/ring-ref",
 				 dev->otherend);
 		return err;
 	}
 
+	/* Try split event channels first, then single event channel. */
+	err = xenbus_gather(XBT_NIL, dev->otherend,
+			    "event-channel-tx", "%u", &tx_evtchn,
+			    "event-channel-rx", "%u", &rx_evtchn, NULL);
+	if (err < 0) {
+		err = xenbus_scanf(XBT_NIL, dev->otherend,
+				   "event-channel", "%u", &tx_evtchn);
+		if (err < 0) {
+			xenbus_dev_fatal(dev, err,
+					 "reading %s/event-channel(-tx/rx)",
+					 dev->otherend);
+			return err;
+		}
+		rx_evtchn = tx_evtchn;
+	}
+
 	err = xenbus_scanf(XBT_NIL, dev->otherend, "request-rx-copy", "%u",
 			   &rx_copy);
 	if (err == -ENOENT) {
@@ -454,11 +478,13 @@ static int connect_rings(struct backend_info *be)
 	vif->csum = !val;
 
 	/* Map the shared frame, irq etc. */
-	err = xenvif_connect(vif, tx_ring_ref, rx_ring_ref, evtchn);
+	err = xenvif_connect(vif, tx_ring_ref, rx_ring_ref,
+			     tx_evtchn, rx_evtchn);
 	if (err) {
 		xenbus_dev_fatal(dev, err,
-				 "mapping shared-frames %lu/%lu port %u",
-				 tx_ring_ref, rx_ring_ref, evtchn);
+				 "mapping shared-frames %lu/%lu port tx %u rx %u",
+				 tx_ring_ref, rx_ring_ref,
+				 tx_evtchn, rx_evtchn);
 		return err;
 	}
 	return 0;
@@ -485,3 +511,8 @@ int xenvif_xenbus_init(void)
 {
 	return xenbus_register_backend(&netback_driver);
 }
+
+void xenvif_xenbus_fini(void)
+{
+	return xenbus_unregister_driver(&netback_driver);
+}
diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c
index 1db10141..ff7f111 100644
--- a/drivers/net/xen-netfront.c
+++ b/drivers/net/xen-netfront.c
@@ -29,6 +29,8 @@
  * IN THE SOFTWARE.
  */
 
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/netdevice.h>
@@ -85,7 +87,15 @@ struct netfront_info {
 
 	struct napi_struct napi;
 
-	unsigned int evtchn;
+	/* Split event channels support, tx_* == rx_* when using
+	 * single event channel.
+	 */
+	unsigned int tx_evtchn, rx_evtchn;
+	unsigned int tx_irq, rx_irq;
+	/* Only used when split event channels support is enabled */
+	char tx_irq_name[IFNAMSIZ+4]; /* DEVNAME-tx */
+	char rx_irq_name[IFNAMSIZ+4]; /* DEVNAME-rx */
+
 	struct xenbus_device *xbdev;
 
 	spinlock_t   tx_lock;
@@ -330,7 +340,7 @@ no_skb:
  push:
 	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->rx, notify);
 	if (notify)
-		notify_remote_via_irq(np->netdev->irq);
+		notify_remote_via_irq(np->rx_irq);
 }
 
 static int xennet_open(struct net_device *dev)
@@ -377,9 +387,8 @@ static void xennet_tx_buf_gc(struct net_device *dev)
 			skb = np->tx_skbs[id].skb;
 			if (unlikely(gnttab_query_foreign_access(
 				np->grant_tx_ref[id]) != 0)) {
-				printk(KERN_ALERT "xennet_tx_buf_gc: warning "
-				       "-- grant still in use by backend "
-				       "domain.\n");
+				pr_alert("%s: warning -- grant still in use by backend domain\n",
+					 __func__);
 				BUG();
 			}
 			gnttab_end_foreign_access_ref(
@@ -623,7 +632,7 @@ static int xennet_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->tx, notify);
 	if (notify)
-		notify_remote_via_irq(np->netdev->irq);
+		notify_remote_via_irq(np->tx_irq);
 
 	u64_stats_update_begin(&stats->syncp);
 	stats->tx_bytes += skb->len;
@@ -796,14 +805,14 @@ static int xennet_set_skb_gso(struct sk_buff *skb,
 {
 	if (!gso->u.gso.size) {
 		if (net_ratelimit())
-			printk(KERN_WARNING "GSO size must not be zero.\n");
+			pr_warn("GSO size must not be zero\n");
 		return -EINVAL;
 	}
 
 	/* Currently only TCPv4 S.O. is supported. */
 	if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4) {
 		if (net_ratelimit())
-			printk(KERN_WARNING "Bad GSO type %d.\n", gso->u.gso.type);
+			pr_warn("Bad GSO type %d\n", gso->u.gso.type);
 		return -EINVAL;
 	}
 
@@ -850,7 +859,6 @@ static RING_IDX xennet_fill_frags(struct netfront_info *np,
 static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
 {
 	struct iphdr *iph;
-	unsigned char *th;
 	int err = -EPROTO;
 	int recalculate_partial_csum = 0;
 
@@ -875,27 +883,27 @@ static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
 		goto out;
 
 	iph = (void *)skb->data;
-	th = skb->data + 4 * iph->ihl;
-	if (th >= skb_tail_pointer(skb))
-		goto out;
 
-	skb->csum_start = th - skb->head;
 	switch (iph->protocol) {
 	case IPPROTO_TCP:
-		skb->csum_offset = offsetof(struct tcphdr, check);
+		if (!skb_partial_csum_set(skb, 4 * iph->ihl,
+					  offsetof(struct tcphdr, check)))
+			goto out;
 
 		if (recalculate_partial_csum) {
-			struct tcphdr *tcph = (struct tcphdr *)th;
+			struct tcphdr *tcph = tcp_hdr(skb);
 			tcph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
 							 skb->len - iph->ihl*4,
 							 IPPROTO_TCP, 0);
 		}
 		break;
 	case IPPROTO_UDP:
-		skb->csum_offset = offsetof(struct udphdr, check);
+		if (!skb_partial_csum_set(skb, 4 * iph->ihl,
+					  offsetof(struct udphdr, check)))
+			goto out;
 
 		if (recalculate_partial_csum) {
-			struct udphdr *udph = (struct udphdr *)th;
+			struct udphdr *udph = udp_hdr(skb);
 			udph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr,
 							 skb->len - iph->ihl*4,
 							 IPPROTO_UDP, 0);
@@ -903,15 +911,11 @@ static int checksum_setup(struct net_device *dev, struct sk_buff *skb)
 		break;
 	default:
 		if (net_ratelimit())
-			printk(KERN_ERR "Attempting to checksum a non-"
-			       "TCP/UDP packet, dropping a protocol"
-			       " %d packet", iph->protocol);
+			pr_err("Attempting to checksum a non-TCP/UDP packet, dropping a protocol %d packet\n",
+			       iph->protocol);
 		goto out;
 	}
 
-	if ((th + skb->csum_offset + 2) > skb_tail_pointer(skb))
-		goto out;
-
 	err = 0;
 
 out:
@@ -1254,23 +1258,35 @@ static int xennet_set_features(struct net_device *dev,
 	return 0;
 }
 
-static irqreturn_t xennet_interrupt(int irq, void *dev_id)
+static irqreturn_t xennet_tx_interrupt(int irq, void *dev_id)
 {
-	struct net_device *dev = dev_id;
-	struct netfront_info *np = netdev_priv(dev);
+	struct netfront_info *np = dev_id;
+	struct net_device *dev = np->netdev;
 	unsigned long flags;
 
 	spin_lock_irqsave(&np->tx_lock, flags);
+	xennet_tx_buf_gc(dev);
+	spin_unlock_irqrestore(&np->tx_lock, flags);
 
-	if (likely(netif_carrier_ok(dev))) {
-		xennet_tx_buf_gc(dev);
-		/* Under tx_lock: protects access to rx shared-ring indexes. */
-		if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx))
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t xennet_rx_interrupt(int irq, void *dev_id)
+{
+	struct netfront_info *np = dev_id;
+	struct net_device *dev = np->netdev;
+
+	if (likely(netif_carrier_ok(dev) &&
+		   RING_HAS_UNCONSUMED_RESPONSES(&np->rx)))
 			napi_schedule(&np->napi);
-	}
 
-	spin_unlock_irqrestore(&np->tx_lock, flags);
+	return IRQ_HANDLED;
+}
 
+static irqreturn_t xennet_interrupt(int irq, void *dev_id)
+{
+	xennet_tx_interrupt(irq, dev_id);
+	xennet_rx_interrupt(irq, dev_id);
 	return IRQ_HANDLED;
 }
 
@@ -1343,14 +1359,14 @@ static struct net_device *xennet_create_dev(struct xenbus_device *dev)
 	/* A grant for every tx ring slot */
 	if (gnttab_alloc_grant_references(TX_MAX_TARGET,
 					  &np->gref_tx_head) < 0) {
-		printk(KERN_ALERT "#### netfront can't alloc tx grant refs\n");
+		pr_alert("can't alloc tx grant refs\n");
 		err = -ENOMEM;
 		goto exit_free_stats;
 	}
 	/* A grant for every rx ring slot */
 	if (gnttab_alloc_grant_references(RX_MAX_TARGET,
 					  &np->gref_rx_head) < 0) {
-		printk(KERN_ALERT "#### netfront can't alloc rx grant refs\n");
+		pr_alert("can't alloc rx grant refs\n");
 		err = -ENOMEM;
 		goto exit_free_tx;
 	}
@@ -1414,16 +1430,14 @@ static int netfront_probe(struct xenbus_device *dev,
 
 	err = register_netdev(info->netdev);
 	if (err) {
-		printk(KERN_WARNING "%s: register_netdev err=%d\n",
-		       __func__, err);
+		pr_warn("%s: register_netdev err=%d\n", __func__, err);
 		goto fail;
 	}
 
 	err = xennet_sysfs_addif(info->netdev);
 	if (err) {
 		unregister_netdev(info->netdev);
-		printk(KERN_WARNING "%s: add sysfs failed err=%d\n",
-		       __func__, err);
+		pr_warn("%s: add sysfs failed err=%d\n", __func__, err);
 		goto fail;
 	}
 
@@ -1451,9 +1465,14 @@ static void xennet_disconnect_backend(struct netfront_info *info)
 	spin_unlock_irq(&info->tx_lock);
 	spin_unlock_bh(&info->rx_lock);
 
-	if (info->netdev->irq)
-		unbind_from_irqhandler(info->netdev->irq, info->netdev);
-	info->evtchn = info->netdev->irq = 0;
+	if (info->tx_irq && (info->tx_irq == info->rx_irq))
+		unbind_from_irqhandler(info->tx_irq, info);
+	if (info->tx_irq && (info->tx_irq != info->rx_irq)) {
+		unbind_from_irqhandler(info->tx_irq, info);
+		unbind_from_irqhandler(info->rx_irq, info);
+	}
+	info->tx_evtchn = info->rx_evtchn = 0;
+	info->tx_irq = info->rx_irq = 0;
 
 	/* End access and free the pages */
 	xennet_end_access(info->tx_ring_ref, info->tx.sring);
@@ -1503,12 +1522,82 @@ static int xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
 	return 0;
 }
 
+static int setup_netfront_single(struct netfront_info *info)
+{
+	int err;
+
+	err = xenbus_alloc_evtchn(info->xbdev, &info->tx_evtchn);
+	if (err < 0)
+		goto fail;
+
+	err = bind_evtchn_to_irqhandler(info->tx_evtchn,
+					xennet_interrupt,
+					0, info->netdev->name, info);
+	if (err < 0)
+		goto bind_fail;
+	info->rx_evtchn = info->tx_evtchn;
+	info->rx_irq = info->tx_irq = err;
+
+	return 0;
+
+bind_fail:
+	xenbus_free_evtchn(info->xbdev, info->tx_evtchn);
+	info->tx_evtchn = 0;
+fail:
+	return err;
+}
+
+static int setup_netfront_split(struct netfront_info *info)
+{
+	int err;
+
+	err = xenbus_alloc_evtchn(info->xbdev, &info->tx_evtchn);
+	if (err < 0)
+		goto fail;
+	err = xenbus_alloc_evtchn(info->xbdev, &info->rx_evtchn);
+	if (err < 0)
+		goto alloc_rx_evtchn_fail;
+
+	snprintf(info->tx_irq_name, sizeof(info->tx_irq_name),
+		 "%s-tx", info->netdev->name);
+	err = bind_evtchn_to_irqhandler(info->tx_evtchn,
+					xennet_tx_interrupt,
+					0, info->tx_irq_name, info);
+	if (err < 0)
+		goto bind_tx_fail;
+	info->tx_irq = err;
+
+	snprintf(info->rx_irq_name, sizeof(info->rx_irq_name),
+		 "%s-rx", info->netdev->name);
+	err = bind_evtchn_to_irqhandler(info->rx_evtchn,
+					xennet_rx_interrupt,
+					0, info->rx_irq_name, info);
+	if (err < 0)
+		goto bind_rx_fail;
+	info->rx_irq = err;
+
+	return 0;
+
+bind_rx_fail:
+	unbind_from_irqhandler(info->tx_irq, info);
+	info->tx_irq = 0;
+bind_tx_fail:
+	xenbus_free_evtchn(info->xbdev, info->rx_evtchn);
+	info->rx_evtchn = 0;
+alloc_rx_evtchn_fail:
+	xenbus_free_evtchn(info->xbdev, info->tx_evtchn);
+	info->tx_evtchn = 0;
+fail:
+	return err;
+}
+
 static int setup_netfront(struct xenbus_device *dev, struct netfront_info *info)
 {
 	struct xen_netif_tx_sring *txs;
 	struct xen_netif_rx_sring *rxs;
 	int err;
 	struct net_device *netdev = info->netdev;
+	unsigned int feature_split_evtchn;
 
 	info->tx_ring_ref = GRANT_INVALID_REF;
 	info->rx_ring_ref = GRANT_INVALID_REF;
@@ -1516,6 +1605,12 @@ static int setup_netfront(struct xenbus_device *dev, struct netfront_info *info)
 	info->tx.sring = NULL;
 	netdev->irq = 0;
 
+	err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+			   "feature-split-event-channels", "%u",
+			   &feature_split_evtchn);
+	if (err < 0)
+		feature_split_evtchn = 0;
+
 	err = xen_net_read_mac(dev, netdev->dev_addr);
 	if (err) {
 		xenbus_dev_fatal(dev, err, "parsing %s/mac", dev->nodename);
@@ -1532,40 +1627,50 @@ static int setup_netfront(struct xenbus_device *dev, struct netfront_info *info)
 	FRONT_RING_INIT(&info->tx, txs, PAGE_SIZE);
 
 	err = xenbus_grant_ring(dev, virt_to_mfn(txs));
-	if (err < 0) {
-		free_page((unsigned long)txs);
-		goto fail;
-	}
+	if (err < 0)
+		goto grant_tx_ring_fail;
 
 	info->tx_ring_ref = err;
 	rxs = (struct xen_netif_rx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
 	if (!rxs) {
 		err = -ENOMEM;
 		xenbus_dev_fatal(dev, err, "allocating rx ring page");
-		goto fail;
+		goto alloc_rx_ring_fail;
 	}
 	SHARED_RING_INIT(rxs);
 	FRONT_RING_INIT(&info->rx, rxs, PAGE_SIZE);
 
 	err = xenbus_grant_ring(dev, virt_to_mfn(rxs));
-	if (err < 0) {
-		free_page((unsigned long)rxs);
-		goto fail;
-	}
+	if (err < 0)
+		goto grant_rx_ring_fail;
 	info->rx_ring_ref = err;
 
-	err = xenbus_alloc_evtchn(dev, &info->evtchn);
+	if (feature_split_evtchn)
+		err = setup_netfront_split(info);
+	/* setup single event channel if
+	 *  a) feature-split-event-channels == 0
+	 *  b) feature-split-event-channels == 1 but failed to setup
+	 */
+	if (!feature_split_evtchn || (feature_split_evtchn && err))
+		err = setup_netfront_single(info);
+
 	if (err)
-		goto fail;
+		goto alloc_evtchn_fail;
 
-	err = bind_evtchn_to_irqhandler(info->evtchn, xennet_interrupt,
-					0, netdev->name, netdev);
-	if (err < 0)
-		goto fail;
-	netdev->irq = err;
 	return 0;
 
- fail:
+	/* If we fail to setup netfront, it is safe to just revoke access to
+	 * granted pages because backend is not accessing it at this point.
+	 */
+alloc_evtchn_fail:
+	gnttab_end_foreign_access_ref(info->rx_ring_ref, 0);
+grant_rx_ring_fail:
+	free_page((unsigned long)rxs);
+alloc_rx_ring_fail:
+	gnttab_end_foreign_access_ref(info->tx_ring_ref, 0);
+grant_tx_ring_fail:
+	free_page((unsigned long)txs);
+fail:
 	return err;
 }
 
@@ -1601,11 +1706,27 @@ again:
 		message = "writing rx ring-ref";
 		goto abort_transaction;
 	}
-	err = xenbus_printf(xbt, dev->nodename,
-			    "event-channel", "%u", info->evtchn);
-	if (err) {
-		message = "writing event-channel";
-		goto abort_transaction;
+
+	if (info->tx_evtchn == info->rx_evtchn) {
+		err = xenbus_printf(xbt, dev->nodename,
+				    "event-channel", "%u", info->tx_evtchn);
+		if (err) {
+			message = "writing event-channel";
+			goto abort_transaction;
+		}
+	} else {
+		err = xenbus_printf(xbt, dev->nodename,
+				    "event-channel-tx", "%u", info->tx_evtchn);
+		if (err) {
+			message = "writing event-channel-tx";
+			goto abort_transaction;
+		}
+		err = xenbus_printf(xbt, dev->nodename,
+				    "event-channel-rx", "%u", info->rx_evtchn);
+		if (err) {
+			message = "writing event-channel-rx";
+			goto abort_transaction;
+		}
 	}
 
 	err = xenbus_printf(xbt, dev->nodename, "request-rx-copy", "%u",
@@ -1718,7 +1839,9 @@ static int xennet_connect(struct net_device *dev)
 	 * packets.
 	 */
 	netif_carrier_on(np->netdev);
-	notify_remote_via_irq(np->netdev->irq);
+	notify_remote_via_irq(np->tx_irq);
+	if (np->tx_irq != np->rx_irq)
+		notify_remote_via_irq(np->rx_irq);
 	xennet_tx_buf_gc(dev);
 	xennet_alloc_rx_buffers(dev);
 
@@ -1991,7 +2114,7 @@ static int __init netif_init(void)
 	if (xen_hvm_domain() && !xen_platform_pci_unplug)
 		return -ENODEV;
 
-	printk(KERN_INFO "Initialising Xen virtual ethernet driver.\n");
+	pr_info("Initialising Xen virtual ethernet driver\n");
 
 	return xenbus_register_frontend(&netfront_driver);
 }
diff --git a/drivers/nfc/Kconfig b/drivers/nfc/Kconfig
index 74a852e..b0b64cc 100644
--- a/drivers/nfc/Kconfig
+++ b/drivers/nfc/Kconfig
@@ -36,6 +36,16 @@ config NFC_MEI_PHY
 
 	  If unsure, say N.
 
+config NFC_SIM
+	tristate "NFC hardware simulator driver"
+	help
+	  This driver declares two virtual NFC devices supporting NFC-DEP
+	  protocol. An LLCP connection can be established between them and
+	  all packets sent from one device is sent back to the other, acting as
+	  loopback devices.
+
+	  If unsure, say N.
+
 source "drivers/nfc/pn544/Kconfig"
 source "drivers/nfc/microread/Kconfig"
 
diff --git a/drivers/nfc/Makefile b/drivers/nfc/Makefile
index aa6bd65..be7636a 100644
--- a/drivers/nfc/Makefile
+++ b/drivers/nfc/Makefile
@@ -7,5 +7,6 @@ obj-$(CONFIG_NFC_MICROREAD)	+= microread/
 obj-$(CONFIG_NFC_PN533)		+= pn533.o
 obj-$(CONFIG_NFC_WILINK)	+= nfcwilink.o
 obj-$(CONFIG_NFC_MEI_PHY)	+= mei_phy.o
+obj-$(CONFIG_NFC_SIM)		+= nfcsim.o
 
 ccflags-$(CONFIG_NFC_DEBUG) := -DDEBUG
diff --git a/drivers/nfc/mei_phy.c b/drivers/nfc/mei_phy.c
index 1201bdb..606bf55 100644
--- a/drivers/nfc/mei_phy.c
+++ b/drivers/nfc/mei_phy.c
@@ -30,7 +30,7 @@ struct mei_nfc_hdr {
 	u16 req_id;
 	u32 reserved;
 	u16 data_size;
-} __attribute__((packed));
+} __packed;
 
 #define MEI_NFC_MAX_READ (MEI_NFC_HEADER_SIZE + MEI_NFC_MAX_HCI_PAYLOAD)
 
@@ -60,8 +60,8 @@ int nfc_mei_phy_enable(void *phy_id)
 
 	r = mei_cl_enable_device(phy->device);
 	if (r < 0) {
-                pr_err("MEI_PHY: Could not enable device\n");
-                return r;
+		pr_err("MEI_PHY: Could not enable device\n");
+		return r;
 	}
 
 	r = mei_cl_register_event_cb(phy->device, nfc_mei_event_cb, phy);
diff --git a/drivers/nfc/microread/microread.c b/drivers/nfc/microread/microread.c
index 3420d83..cdb9f6d 100644
--- a/drivers/nfc/microread/microread.c
+++ b/drivers/nfc/microread/microread.c
@@ -650,7 +650,7 @@ int microread_probe(void *phy_id, struct nfc_phy_ops *phy_ops, char *llc_name,
 {
 	struct microread_info *info;
 	unsigned long quirks = 0;
-	u32 protocols, se;
+	u32 protocols;
 	struct nfc_hci_init_data init_data;
 	int r;
 
@@ -678,10 +678,8 @@ int microread_probe(void *phy_id, struct nfc_phy_ops *phy_ops, char *llc_name,
 		    NFC_PROTO_ISO14443_B_MASK |
 		    NFC_PROTO_NFC_DEP_MASK;
 
-	se = NFC_SE_UICC | NFC_SE_EMBEDDED;
-
 	info->hdev = nfc_hci_allocate_device(&microread_hci_ops, &init_data,
-					     quirks, protocols, se, llc_name,
+					     quirks, protocols, llc_name,
 					     phy_headroom +
 					     MICROREAD_CMDS_HEADROOM,
 					     phy_tailroom +
diff --git a/drivers/nfc/nfcsim.c b/drivers/nfc/nfcsim.c
new file mode 100644
index 0000000..c5c30fb
--- /dev/null
+++ b/drivers/nfc/nfcsim.c
@@ -0,0 +1,541 @@
+/*
+ * NFC hardware simulation driver
+ * Copyright (c) 2013, Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ */
+
+#include <linux/device.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/nfc.h>
+#include <net/nfc/nfc.h>
+
+#define DEV_ERR(_dev, fmt, args...) nfc_dev_err(&_dev->nfc_dev->dev, \
+						"%s: " fmt, __func__, ## args)
+
+#define DEV_DBG(_dev, fmt, args...) nfc_dev_dbg(&_dev->nfc_dev->dev, \
+						"%s: " fmt, __func__, ## args)
+
+#define NFCSIM_VERSION "0.1"
+
+#define NFCSIM_POLL_NONE	0
+#define NFCSIM_POLL_INITIATOR	1
+#define NFCSIM_POLL_TARGET	2
+#define NFCSIM_POLL_DUAL	(NFCSIM_POLL_INITIATOR | NFCSIM_POLL_TARGET)
+
+struct nfcsim {
+	struct nfc_dev *nfc_dev;
+
+	struct mutex lock;
+
+	struct delayed_work recv_work;
+
+	struct sk_buff *clone_skb;
+
+	struct delayed_work poll_work;
+	u8 polling_mode;
+	u8 curr_polling_mode;
+
+	u8 shutting_down;
+
+	u8 up;
+
+	u8 initiator;
+
+	data_exchange_cb_t cb;
+	void *cb_context;
+
+	struct nfcsim *peer_dev;
+};
+
+static struct nfcsim *dev0;
+static struct nfcsim *dev1;
+
+struct workqueue_struct *wq;
+
+static void nfcsim_cleanup_dev(struct nfcsim *dev, u8 shutdown)
+{
+	DEV_DBG(dev, "shutdown=%d", shutdown);
+
+	mutex_lock(&dev->lock);
+
+	dev->polling_mode = NFCSIM_POLL_NONE;
+	dev->shutting_down = shutdown;
+	dev->cb = NULL;
+	dev_kfree_skb(dev->clone_skb);
+	dev->clone_skb = NULL;
+
+	mutex_unlock(&dev->lock);
+
+	cancel_delayed_work_sync(&dev->poll_work);
+	cancel_delayed_work_sync(&dev->recv_work);
+}
+
+static int nfcsim_target_found(struct nfcsim *dev)
+{
+	struct nfc_target nfc_tgt;
+
+	DEV_DBG(dev, "");
+
+	memset(&nfc_tgt, 0, sizeof(struct nfc_target));
+
+	nfc_tgt.supported_protocols = NFC_PROTO_NFC_DEP_MASK;
+	nfc_targets_found(dev->nfc_dev, &nfc_tgt, 1);
+
+	return 0;
+}
+
+static int nfcsim_dev_up(struct nfc_dev *nfc_dev)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+
+	DEV_DBG(dev, "");
+
+	mutex_lock(&dev->lock);
+
+	dev->up = 1;
+
+	mutex_unlock(&dev->lock);
+
+	return 0;
+}
+
+static int nfcsim_dev_down(struct nfc_dev *nfc_dev)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+
+	DEV_DBG(dev, "");
+
+	mutex_lock(&dev->lock);
+
+	dev->up = 0;
+
+	mutex_unlock(&dev->lock);
+
+	return 0;
+}
+
+static int nfcsim_dep_link_up(struct nfc_dev *nfc_dev,
+			      struct nfc_target *target,
+			      u8 comm_mode, u8 *gb, size_t gb_len)
+{
+	int rc;
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+	struct nfcsim *peer = dev->peer_dev;
+	u8 *remote_gb;
+	size_t remote_gb_len;
+
+	DEV_DBG(dev, "target_idx: %d, comm_mode: %d\n", target->idx, comm_mode);
+
+	mutex_lock(&peer->lock);
+
+	nfc_tm_activated(peer->nfc_dev, NFC_PROTO_NFC_DEP_MASK,
+			 NFC_COMM_ACTIVE, gb, gb_len);
+
+	remote_gb = nfc_get_local_general_bytes(peer->nfc_dev, &remote_gb_len);
+	if (!remote_gb) {
+		DEV_ERR(peer, "Can't get remote general bytes");
+
+		mutex_unlock(&peer->lock);
+		return -EINVAL;
+	}
+
+	mutex_unlock(&peer->lock);
+
+	mutex_lock(&dev->lock);
+
+	rc = nfc_set_remote_general_bytes(nfc_dev, remote_gb, remote_gb_len);
+	if (rc) {
+		DEV_ERR(dev, "Can't set remote general bytes");
+		mutex_unlock(&dev->lock);
+		return rc;
+	}
+
+	rc = nfc_dep_link_is_up(nfc_dev, target->idx, NFC_COMM_ACTIVE,
+				NFC_RF_INITIATOR);
+
+	mutex_unlock(&dev->lock);
+
+	return rc;
+}
+
+static int nfcsim_dep_link_down(struct nfc_dev *nfc_dev)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+
+	DEV_DBG(dev, "");
+
+	nfcsim_cleanup_dev(dev, 0);
+
+	return 0;
+}
+
+static int nfcsim_start_poll(struct nfc_dev *nfc_dev,
+			     u32 im_protocols, u32 tm_protocols)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+	int rc;
+
+	mutex_lock(&dev->lock);
+
+	if (dev->polling_mode != NFCSIM_POLL_NONE) {
+		DEV_ERR(dev, "Already in polling mode");
+		rc = -EBUSY;
+		goto exit;
+	}
+
+	if (im_protocols & NFC_PROTO_NFC_DEP_MASK)
+		dev->polling_mode |= NFCSIM_POLL_INITIATOR;
+
+	if (tm_protocols & NFC_PROTO_NFC_DEP_MASK)
+		dev->polling_mode |= NFCSIM_POLL_TARGET;
+
+	if (dev->polling_mode == NFCSIM_POLL_NONE) {
+		DEV_ERR(dev, "Unsupported polling mode");
+		rc = -EINVAL;
+		goto exit;
+	}
+
+	dev->initiator = 0;
+	dev->curr_polling_mode = NFCSIM_POLL_NONE;
+
+	queue_delayed_work(wq, &dev->poll_work, 0);
+
+	DEV_DBG(dev, "Start polling: im: 0x%X, tm: 0x%X", im_protocols,
+		tm_protocols);
+
+	rc = 0;
+exit:
+	mutex_unlock(&dev->lock);
+
+	return rc;
+}
+
+static void nfcsim_stop_poll(struct nfc_dev *nfc_dev)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+
+	DEV_DBG(dev, "Stop poll");
+
+	mutex_lock(&dev->lock);
+
+	dev->polling_mode = NFCSIM_POLL_NONE;
+
+	mutex_unlock(&dev->lock);
+
+	cancel_delayed_work_sync(&dev->poll_work);
+}
+
+static int nfcsim_activate_target(struct nfc_dev *nfc_dev,
+				  struct nfc_target *target, u32 protocol)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+
+	DEV_DBG(dev, "");
+
+	return -ENOTSUPP;
+}
+
+static void nfcsim_deactivate_target(struct nfc_dev *nfc_dev,
+				     struct nfc_target *target)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+
+	DEV_DBG(dev, "");
+}
+
+static void nfcsim_wq_recv(struct work_struct *work)
+{
+	struct nfcsim *dev = container_of(work, struct nfcsim,
+					  recv_work.work);
+
+	mutex_lock(&dev->lock);
+
+	if (dev->shutting_down || !dev->up || !dev->clone_skb) {
+		dev_kfree_skb(dev->clone_skb);
+		goto exit;
+	}
+
+	if (dev->initiator) {
+		if (!dev->cb) {
+			DEV_ERR(dev, "Null recv callback");
+			dev_kfree_skb(dev->clone_skb);
+			goto exit;
+		}
+
+		dev->cb(dev->cb_context, dev->clone_skb, 0);
+		dev->cb = NULL;
+	} else {
+		nfc_tm_data_received(dev->nfc_dev, dev->clone_skb);
+	}
+
+exit:
+	dev->clone_skb = NULL;
+
+	mutex_unlock(&dev->lock);
+}
+
+static int nfcsim_tx(struct nfc_dev *nfc_dev, struct nfc_target *target,
+		     struct sk_buff *skb, data_exchange_cb_t cb,
+		     void *cb_context)
+{
+	struct nfcsim *dev = nfc_get_drvdata(nfc_dev);
+	struct nfcsim *peer = dev->peer_dev;
+	int err;
+
+	mutex_lock(&dev->lock);
+
+	if (dev->shutting_down || !dev->up) {
+		mutex_unlock(&dev->lock);
+		err = -ENODEV;
+		goto exit;
+	}
+
+	dev->cb = cb;
+	dev->cb_context = cb_context;
+
+	mutex_unlock(&dev->lock);
+
+	mutex_lock(&peer->lock);
+
+	peer->clone_skb = skb_clone(skb, GFP_KERNEL);
+
+	if (!peer->clone_skb) {
+		DEV_ERR(dev, "skb_clone failed");
+		mutex_unlock(&peer->lock);
+		err = -ENOMEM;
+		goto exit;
+	}
+
+	/* This simulates an arbitrary transmission delay between the 2 devices.
+	 * If packet transmission occurs immediately between them, we have a
+	 * non-stop flow of several tens of thousands SYMM packets per second
+	 * and a burning cpu.
+	 *
+	 * TODO: Add support for a sysfs entry to control this delay.
+	 */
+	queue_delayed_work(wq, &peer->recv_work, msecs_to_jiffies(5));
+
+	mutex_unlock(&peer->lock);
+
+	err = 0;
+exit:
+	dev_kfree_skb(skb);
+
+	return err;
+}
+
+static int nfcsim_im_transceive(struct nfc_dev *nfc_dev,
+				struct nfc_target *target, struct sk_buff *skb,
+				data_exchange_cb_t cb, void *cb_context)
+{
+	return nfcsim_tx(nfc_dev, target, skb, cb, cb_context);
+}
+
+static int nfcsim_tm_send(struct nfc_dev *nfc_dev, struct sk_buff *skb)
+{
+	return nfcsim_tx(nfc_dev, NULL, skb, NULL, NULL);
+}
+
+static struct nfc_ops nfcsim_nfc_ops = {
+	.dev_up = nfcsim_dev_up,
+	.dev_down = nfcsim_dev_down,
+	.dep_link_up = nfcsim_dep_link_up,
+	.dep_link_down = nfcsim_dep_link_down,
+	.start_poll = nfcsim_start_poll,
+	.stop_poll = nfcsim_stop_poll,
+	.activate_target = nfcsim_activate_target,
+	.deactivate_target = nfcsim_deactivate_target,
+	.im_transceive = nfcsim_im_transceive,
+	.tm_send = nfcsim_tm_send,
+};
+
+static void nfcsim_set_polling_mode(struct nfcsim *dev)
+{
+	if (dev->polling_mode == NFCSIM_POLL_NONE) {
+		dev->curr_polling_mode = NFCSIM_POLL_NONE;
+		return;
+	}
+
+	if (dev->curr_polling_mode == NFCSIM_POLL_NONE) {
+		if (dev->polling_mode & NFCSIM_POLL_INITIATOR)
+			dev->curr_polling_mode = NFCSIM_POLL_INITIATOR;
+		else
+			dev->curr_polling_mode = NFCSIM_POLL_TARGET;
+
+		return;
+	}
+
+	if (dev->polling_mode == NFCSIM_POLL_DUAL) {
+		if (dev->curr_polling_mode == NFCSIM_POLL_TARGET)
+			dev->curr_polling_mode = NFCSIM_POLL_INITIATOR;
+		else
+			dev->curr_polling_mode = NFCSIM_POLL_TARGET;
+	}
+}
+
+static void nfcsim_wq_poll(struct work_struct *work)
+{
+	struct nfcsim *dev = container_of(work, struct nfcsim, poll_work.work);
+	struct nfcsim *peer = dev->peer_dev;
+
+	/* These work items run on an ordered workqueue and are therefore
+	 * serialized. So we can take both mutexes without being dead locked.
+	 */
+	mutex_lock(&dev->lock);
+	mutex_lock(&peer->lock);
+
+	nfcsim_set_polling_mode(dev);
+
+	if (dev->curr_polling_mode == NFCSIM_POLL_NONE) {
+		DEV_DBG(dev, "Not polling");
+		goto unlock;
+	}
+
+	DEV_DBG(dev, "Polling as %s",
+		dev->curr_polling_mode == NFCSIM_POLL_INITIATOR ?
+		"initiator" : "target");
+
+	if (dev->curr_polling_mode == NFCSIM_POLL_TARGET)
+		goto sched_work;
+
+	if (peer->curr_polling_mode == NFCSIM_POLL_TARGET) {
+		peer->polling_mode = NFCSIM_POLL_NONE;
+		dev->polling_mode = NFCSIM_POLL_NONE;
+
+		dev->initiator = 1;
+
+		nfcsim_target_found(dev);
+
+		goto unlock;
+	}
+
+sched_work:
+	/* This defines the delay for an initiator to check if the other device
+	 * is polling in target mode.
+	 * If the device starts in dual mode polling, it switches between
+	 * initiator and target at every round.
+	 * Because the wq is ordered and only 1 work item is executed at a time,
+	 * we'll always have one device polling as initiator and the other as
+	 * target at some point, even if both are started in dual mode.
+	 */
+	queue_delayed_work(wq, &dev->poll_work, msecs_to_jiffies(200));
+
+unlock:
+	mutex_unlock(&peer->lock);
+	mutex_unlock(&dev->lock);
+}
+
+static struct nfcsim *nfcsim_init_dev(void)
+{
+	struct nfcsim *dev;
+	int rc = -ENOMEM;
+
+	dev = kzalloc(sizeof(*dev), GFP_KERNEL);
+	if (dev == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	mutex_init(&dev->lock);
+
+	INIT_DELAYED_WORK(&dev->recv_work, nfcsim_wq_recv);
+	INIT_DELAYED_WORK(&dev->poll_work, nfcsim_wq_poll);
+
+	dev->nfc_dev = nfc_allocate_device(&nfcsim_nfc_ops,
+					   NFC_PROTO_NFC_DEP_MASK,
+					   0, 0);
+	if (!dev->nfc_dev)
+		goto error;
+
+	nfc_set_drvdata(dev->nfc_dev, dev);
+
+	rc = nfc_register_device(dev->nfc_dev);
+	if (rc)
+		goto free_nfc_dev;
+
+	return dev;
+
+free_nfc_dev:
+	nfc_free_device(dev->nfc_dev);
+
+error:
+	kfree(dev);
+
+	return ERR_PTR(rc);
+}
+
+static void nfcsim_free_device(struct nfcsim *dev)
+{
+	nfc_unregister_device(dev->nfc_dev);
+
+	nfc_free_device(dev->nfc_dev);
+
+	kfree(dev);
+}
+
+int __init nfcsim_init(void)
+{
+	int rc;
+
+	/* We need an ordered wq to ensure that poll_work items are executed
+	 * one at a time.
+	 */
+	wq = alloc_ordered_workqueue("nfcsim", 0);
+	if (!wq) {
+		rc = -ENOMEM;
+		goto exit;
+	}
+
+	dev0 = nfcsim_init_dev();
+	if (IS_ERR(dev0)) {
+		rc = PTR_ERR(dev0);
+		goto exit;
+	}
+
+	dev1 = nfcsim_init_dev();
+	if (IS_ERR(dev1)) {
+		kfree(dev0);
+
+		rc = PTR_ERR(dev1);
+		goto exit;
+	}
+
+	dev0->peer_dev = dev1;
+	dev1->peer_dev = dev0;
+
+	pr_debug("NFCsim " NFCSIM_VERSION " initialized\n");
+
+	rc = 0;
+exit:
+	if (rc)
+		pr_err("Failed to initialize nfcsim driver (%d)\n",
+		       rc);
+
+	return rc;
+}
+
+void __exit nfcsim_exit(void)
+{
+	nfcsim_cleanup_dev(dev0, 1);
+	nfcsim_cleanup_dev(dev1, 1);
+
+	nfcsim_free_device(dev0);
+	nfcsim_free_device(dev1);
+
+	destroy_workqueue(wq);
+}
+
+module_init(nfcsim_init);
+module_exit(nfcsim_exit);
+
+MODULE_DESCRIPTION("NFCSim driver ver " NFCSIM_VERSION);
+MODULE_VERSION(NFCSIM_VERSION);
+MODULE_LICENSE("GPL");
diff --git a/drivers/nfc/nfcwilink.c b/drivers/nfc/nfcwilink.c
index 3b731ac..59f95d8 100644
--- a/drivers/nfc/nfcwilink.c
+++ b/drivers/nfc/nfcwilink.c
@@ -109,7 +109,7 @@ enum {
 	NFCWILINK_FW_DOWNLOAD,
 };
 
-static int nfcwilink_send(struct sk_buff *skb);
+static int nfcwilink_send(struct nci_dev *ndev, struct sk_buff *skb);
 
 static inline struct sk_buff *nfcwilink_skb_alloc(unsigned int len, gfp_t how)
 {
@@ -156,8 +156,6 @@ static int nfcwilink_get_bts_file_name(struct nfcwilink *drv, char *file_name)
 		return -ENOMEM;
 	}
 
-	skb->dev = (void *)drv->ndev;
-
 	cmd = (struct nci_vs_nfcc_info_cmd *)
 			skb_put(skb, sizeof(struct nci_vs_nfcc_info_cmd));
 	cmd->gid = NCI_VS_NFCC_INFO_CMD_GID;
@@ -166,7 +164,7 @@ static int nfcwilink_get_bts_file_name(struct nfcwilink *drv, char *file_name)
 
 	drv->nfcc_info.plen = 0;
 
-	rc = nfcwilink_send(skb);
+	rc = nfcwilink_send(drv->ndev, skb);
 	if (rc)
 		return rc;
 
@@ -232,11 +230,9 @@ static int nfcwilink_send_bts_cmd(struct nfcwilink *drv, __u8 *data, int len)
 		return -ENOMEM;
 	}
 
-	skb->dev = (void *)drv->ndev;
-
 	memcpy(skb_put(skb, len), data, len);
 
-	rc = nfcwilink_send(skb);
+	rc = nfcwilink_send(drv->ndev, skb);
 	if (rc)
 		return rc;
 
@@ -371,10 +367,8 @@ static long nfcwilink_receive(void *priv_data, struct sk_buff *skb)
 		return 0;
 	}
 
-	skb->dev = (void *) drv->ndev;
-
 	/* Forward skb to NCI core layer */
-	rc = nci_recv_frame(skb);
+	rc = nci_recv_frame(drv->ndev, skb);
 	if (rc < 0) {
 		nfc_dev_err(&drv->pdev->dev, "nci_recv_frame failed %d", rc);
 		return rc;
@@ -480,9 +474,8 @@ static int nfcwilink_close(struct nci_dev *ndev)
 	return rc;
 }
 
-static int nfcwilink_send(struct sk_buff *skb)
+static int nfcwilink_send(struct nci_dev *ndev, struct sk_buff *skb)
 {
-	struct nci_dev *ndev = (struct nci_dev *)skb->dev;
 	struct nfcwilink *drv = nci_get_drvdata(ndev);
 	struct nfcwilink_hdr hdr = {NFCWILINK_CHNL, NFCWILINK_OPCODE, 0x0000};
 	long len;
@@ -542,7 +535,6 @@ static int nfcwilink_probe(struct platform_device *pdev)
 
 	drv->ndev = nci_allocate_device(&nfcwilink_ops,
 					protocols,
-					NFC_SE_NONE,
 					NFCWILINK_HDR_LEN,
 					0);
 	if (!drv->ndev) {
diff --git a/drivers/nfc/pn533.c b/drivers/nfc/pn533.c
index ec269e6..daf92ac 100644
--- a/drivers/nfc/pn533.c
+++ b/drivers/nfc/pn533.c
@@ -258,7 +258,7 @@ static const struct pn533_poll_modulations poll_mod[] = {
 				.opcode = PN533_FELICA_OPC_SENSF_REQ,
 				.sc = PN533_FELICA_SENSF_SC_ALL,
 				.rc = PN533_FELICA_SENSF_RC_NO_SYSTEM_CODE,
-				.tsn = 0,
+				.tsn = 0x03,
 			},
 		},
 		.len = 7,
@@ -271,7 +271,7 @@ static const struct pn533_poll_modulations poll_mod[] = {
 				.opcode = PN533_FELICA_OPC_SENSF_REQ,
 				.sc = PN533_FELICA_SENSF_SC_ALL,
 				.rc = PN533_FELICA_SENSF_RC_NO_SYSTEM_CODE,
-				.tsn = 0,
+				.tsn = 0x03,
 			},
 		 },
 		.len = 7,
@@ -1235,7 +1235,7 @@ static int pn533_target_found_type_a(struct nfc_target *nfc_tgt, u8 *tgt_data,
 struct pn533_target_felica {
 	u8 pol_res;
 	u8 opcode;
-	u8 nfcid2[8];
+	u8 nfcid2[NFC_NFCID2_MAXSIZE];
 	u8 pad[8];
 	/* optional */
 	u8 syst_code[];
@@ -1275,6 +1275,9 @@ static int pn533_target_found_felica(struct nfc_target *nfc_tgt, u8 *tgt_data,
 	memcpy(nfc_tgt->sensf_res, &tgt_felica->opcode, 9);
 	nfc_tgt->sensf_res_len = 9;
 
+	memcpy(nfc_tgt->nfcid2, tgt_felica->nfcid2, NFC_NFCID2_MAXSIZE);
+	nfc_tgt->nfcid2_len = NFC_NFCID2_MAXSIZE;
+
 	return 0;
 }
 
@@ -2084,6 +2087,9 @@ static int pn533_dep_link_up(struct nfc_dev *nfc_dev, struct nfc_target *target,
 	if (comm_mode == NFC_COMM_PASSIVE)
 		skb_len += PASSIVE_DATA_LEN;
 
+	if (target && target->nfcid2_len)
+		skb_len += NFC_NFCID3_MAXSIZE;
+
 	skb = pn533_alloc_skb(dev, skb_len);
 	if (!skb)
 		return -ENOMEM;
@@ -2100,6 +2106,12 @@ static int pn533_dep_link_up(struct nfc_dev *nfc_dev, struct nfc_target *target,
 		*next |= 1;
 	}
 
+	if (target && target->nfcid2_len) {
+		memcpy(skb_put(skb, NFC_NFCID3_MAXSIZE), target->nfcid2,
+		       target->nfcid2_len);
+		*next |= 2;
+	}
+
 	if (gb != NULL && gb_len > 0) {
 		memcpy(skb_put(skb, gb_len), gb, gb_len);
 		*next |= 4; /* We have some Gi */
@@ -2489,7 +2501,7 @@ static void pn533_acr122_poweron_rdr_resp(struct urb *urb)
 
 	nfc_dev_dbg(&urb->dev->dev, "%s", __func__);
 
-	print_hex_dump(KERN_ERR, "ACR122 RX: ", DUMP_PREFIX_NONE, 16, 1,
+	print_hex_dump_debug("ACR122 RX: ", DUMP_PREFIX_NONE, 16, 1,
 		       urb->transfer_buffer, urb->transfer_buffer_length,
 		       false);
 
@@ -2520,7 +2532,7 @@ static int pn533_acr122_poweron_rdr(struct pn533 *dev)
 	dev->out_urb->transfer_buffer = cmd;
 	dev->out_urb->transfer_buffer_length = sizeof(cmd);
 
-	print_hex_dump(KERN_ERR, "ACR122 TX: ", DUMP_PREFIX_NONE, 16, 1,
+	print_hex_dump_debug("ACR122 TX: ", DUMP_PREFIX_NONE, 16, 1,
 		       cmd, sizeof(cmd), false);
 
 	rc = usb_submit_urb(dev->out_urb, GFP_KERNEL);
@@ -2774,17 +2786,18 @@ static int pn533_probe(struct usb_interface *interface,
 		goto destroy_wq;
 
 	nfc_dev_info(&dev->interface->dev,
-		     "NXP PN533 firmware ver %d.%d now attached",
-		     fw_ver.ver, fw_ver.rev);
+		     "NXP PN5%02X firmware ver %d.%d now attached",
+		     fw_ver.ic, fw_ver.ver, fw_ver.rev);
 
 
 	dev->nfc_dev = nfc_allocate_device(&pn533_nfc_ops, protocols,
-					   NFC_SE_NONE,
 					   dev->ops->tx_header_len +
 					   PN533_CMD_DATAEXCH_HEAD_LEN,
 					   dev->ops->tx_tail_len);
-	if (!dev->nfc_dev)
+	if (!dev->nfc_dev) {
+		rc = -ENOMEM;
 		goto destroy_wq;
+	}
 
 	nfc_set_parent_dev(dev->nfc_dev, &interface->dev);
 	nfc_set_drvdata(dev->nfc_dev, dev);
diff --git a/drivers/nfc/pn544/pn544.c b/drivers/nfc/pn544/pn544.c
index 9c5f16e..0d17da7 100644
--- a/drivers/nfc/pn544/pn544.c
+++ b/drivers/nfc/pn544/pn544.c
@@ -551,20 +551,25 @@ static int pn544_hci_complete_target_discovered(struct nfc_hci_dev *hdev,
 			return -EPROTO;
 		}
 
-		r = nfc_hci_send_cmd(hdev, PN544_RF_READER_F_GATE,
-				     PN544_RF_READER_CMD_ACTIVATE_NEXT,
-				     uid_skb->data, uid_skb->len, NULL);
-		kfree_skb(uid_skb);
-
-		r = nfc_hci_send_cmd(hdev,
+		/* Type F NFC-DEP IDm has prefix 0x01FE */
+		if ((uid_skb->data[0] == 0x01) && (uid_skb->data[1] == 0xfe)) {
+			kfree_skb(uid_skb);
+			r = nfc_hci_send_cmd(hdev,
 					PN544_RF_READER_NFCIP1_INITIATOR_GATE,
 					PN544_HCI_CMD_CONTINUE_ACTIVATION,
 					NULL, 0, NULL);
-		if (r < 0)
-			return r;
+			if (r < 0)
+				return r;
 
-		target->hci_reader_gate = PN544_RF_READER_NFCIP1_INITIATOR_GATE;
-		target->supported_protocols = NFC_PROTO_NFC_DEP_MASK;
+			target->supported_protocols = NFC_PROTO_NFC_DEP_MASK;
+			target->hci_reader_gate =
+				PN544_RF_READER_NFCIP1_INITIATOR_GATE;
+		} else {
+			r = nfc_hci_send_cmd(hdev, PN544_RF_READER_F_GATE,
+					     PN544_RF_READER_CMD_ACTIVATE_NEXT,
+					     uid_skb->data, uid_skb->len, NULL);
+			kfree_skb(uid_skb);
+		}
 	} else if (target->supported_protocols & NFC_PROTO_ISO14443_MASK) {
 		/*
 		 * TODO: maybe other ISO 14443 require some kind of continue
@@ -706,12 +711,9 @@ static int pn544_hci_check_presence(struct nfc_hci_dev *hdev,
 		 return nfc_hci_send_cmd(hdev, NFC_HCI_RF_READER_A_GATE,
 				     PN544_RF_READER_CMD_ACTIVATE_NEXT,
 				     target->nfcid1, target->nfcid1_len, NULL);
-	} else if (target->supported_protocols & NFC_PROTO_JEWEL_MASK) {
-		return nfc_hci_send_cmd(hdev, target->hci_reader_gate,
-					PN544_JEWEL_RAW_CMD, NULL, 0, NULL);
-	} else if (target->supported_protocols & NFC_PROTO_FELICA_MASK) {
-		return nfc_hci_send_cmd(hdev, PN544_RF_READER_F_GATE,
-					PN544_FELICA_RAW, NULL, 0, NULL);
+	} else if (target->supported_protocols & (NFC_PROTO_JEWEL_MASK |
+						NFC_PROTO_FELICA_MASK)) {
+		return -EOPNOTSUPP;
 	} else if (target->supported_protocols & NFC_PROTO_NFC_DEP_MASK) {
 		return nfc_hci_send_cmd(hdev, target->hci_reader_gate,
 					PN544_HCI_CMD_ATTREQUEST,
@@ -801,7 +803,7 @@ int pn544_hci_probe(void *phy_id, struct nfc_phy_ops *phy_ops, char *llc_name,
 		    struct nfc_hci_dev **hdev)
 {
 	struct pn544_hci_info *info;
-	u32 protocols, se;
+	u32 protocols;
 	struct nfc_hci_init_data init_data;
 	int r;
 
@@ -834,10 +836,8 @@ int pn544_hci_probe(void *phy_id, struct nfc_phy_ops *phy_ops, char *llc_name,
 		    NFC_PROTO_ISO14443_B_MASK |
 		    NFC_PROTO_NFC_DEP_MASK;
 
-	se = NFC_SE_UICC | NFC_SE_EMBEDDED;
-
 	info->hdev = nfc_hci_allocate_device(&pn544_hci_ops, &init_data, 0,
-					     protocols, se, llc_name,
+					     protocols, llc_name,
 					     phy_headroom + PN544_CMDS_HEADROOM,
 					     phy_tailroom, phy_payload);
 	if (!info->hdev) {
diff --git a/drivers/of/of_net.c b/drivers/of/of_net.c
index ffab033..ea174c8 100644
--- a/drivers/of/of_net.c
+++ b/drivers/of/of_net.c
@@ -22,6 +22,7 @@ static const char *phy_modes[] = {
 	[PHY_INTERFACE_MODE_GMII]	= "gmii",
 	[PHY_INTERFACE_MODE_SGMII]	= "sgmii",
 	[PHY_INTERFACE_MODE_TBI]	= "tbi",
+	[PHY_INTERFACE_MODE_REVMII]	= "rev-mii",
 	[PHY_INTERFACE_MODE_RMII]	= "rmii",
 	[PHY_INTERFACE_MODE_RGMII]	= "rgmii",
 	[PHY_INTERFACE_MODE_RGMII_ID]	= "rgmii-id",
diff --git a/drivers/power/reset/restart-poweroff.c b/drivers/power/reset/restart-poweroff.c
index 059cd15..5758033 100644
--- a/drivers/power/reset/restart-poweroff.c
+++ b/drivers/power/reset/restart-poweroff.c
@@ -15,11 +15,12 @@
 #include <linux/platform_device.h>
 #include <linux/of_platform.h>
 #include <linux/module.h>
+#include <linux/reboot.h>
 #include <asm/system_misc.h>
 
 static void restart_poweroff_do_poweroff(void)
 {
-	arm_pm_restart('h', NULL);
+	arm_pm_restart(REBOOT_HARD, NULL);
 }
 
 static int restart_poweroff_probe(struct platform_device *pdev)
diff --git a/drivers/power/reset/vexpress-poweroff.c b/drivers/power/reset/vexpress-poweroff.c
index 469e696..476aa49 100644
--- a/drivers/power/reset/vexpress-poweroff.c
+++ b/drivers/power/reset/vexpress-poweroff.c
@@ -48,7 +48,7 @@ static void vexpress_power_off(void)
 
 static struct device *vexpress_restart_device;
 
-static void vexpress_restart(char str, const char *cmd)
+static void vexpress_restart(enum reboot_mode reboot_mode, const char *cmd)
 {
 	vexpress_reset_do(vexpress_restart_device, "restart");
 }
diff --git a/drivers/rtc/rtc-stmp3xxx.c b/drivers/rtc/rtc-stmp3xxx.c
index 90a3e86..767fee2 100644
--- a/drivers/rtc/rtc-stmp3xxx.c
+++ b/drivers/rtc/rtc-stmp3xxx.c
@@ -261,7 +261,12 @@ static int stmp3xxx_rtc_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, rtc_data);
 
-	stmp_reset_block(rtc_data->io);
+	err = stmp_reset_block(rtc_data->io);
+	if (err) {
+		dev_err(&pdev->dev, "stmp_reset_block failed: %d\n", err);
+		return err;
+	}
+
 	writel(STMP3XXX_RTC_PERSISTENT0_ALARM_EN |
 			STMP3XXX_RTC_PERSISTENT0_ALARM_WAKE_EN |
 			STMP3XXX_RTC_PERSISTENT0_ALARM_WAKE,
diff --git a/drivers/s390/net/netiucv.c b/drivers/s390/net/netiucv.c
index 9ca3996..279ad50 100644
--- a/drivers/s390/net/netiucv.c
+++ b/drivers/s390/net/netiucv.c
@@ -130,26 +130,6 @@ static inline int iucv_dbf_passes(debug_info_t *dbf_grp, int level)
 /**
  * some more debug stuff
  */
-#define IUCV_HEXDUMP16(importance,header,ptr) \
-PRINT_##importance(header "%02x %02x %02x %02x  %02x %02x %02x %02x  " \
-		   "%02x %02x %02x %02x  %02x %02x %02x %02x\n", \
-		   *(((char*)ptr)),*(((char*)ptr)+1),*(((char*)ptr)+2), \
-		   *(((char*)ptr)+3),*(((char*)ptr)+4),*(((char*)ptr)+5), \
-		   *(((char*)ptr)+6),*(((char*)ptr)+7),*(((char*)ptr)+8), \
-		   *(((char*)ptr)+9),*(((char*)ptr)+10),*(((char*)ptr)+11), \
-		   *(((char*)ptr)+12),*(((char*)ptr)+13), \
-		   *(((char*)ptr)+14),*(((char*)ptr)+15)); \
-PRINT_##importance(header "%02x %02x %02x %02x  %02x %02x %02x %02x  " \
-		   "%02x %02x %02x %02x  %02x %02x %02x %02x\n", \
-		   *(((char*)ptr)+16),*(((char*)ptr)+17), \
-		   *(((char*)ptr)+18),*(((char*)ptr)+19), \
-		   *(((char*)ptr)+20),*(((char*)ptr)+21), \
-		   *(((char*)ptr)+22),*(((char*)ptr)+23), \
-		   *(((char*)ptr)+24),*(((char*)ptr)+25), \
-		   *(((char*)ptr)+26),*(((char*)ptr)+27), \
-		   *(((char*)ptr)+28),*(((char*)ptr)+29), \
-		   *(((char*)ptr)+30),*(((char*)ptr)+31));
-
 #define PRINTK_HEADER " iucv: "       /* for debugging */
 
 /* dummy device to make sure netiucv_pm functions are called */
diff --git a/drivers/s390/net/qeth_core.h b/drivers/s390/net/qeth_core.h
index c4f392d..41ef943 100644
--- a/drivers/s390/net/qeth_core.h
+++ b/drivers/s390/net/qeth_core.h
@@ -738,7 +738,7 @@ struct qeth_rx {
 	int qdio_err;
 };
 
-#define QETH_NAPI_WEIGHT 128
+#define QETH_NAPI_WEIGHT NAPI_POLL_WEIGHT
 
 struct qeth_card {
 	struct list_head list;
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index 70ce6b6..0a328d0 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -1282,8 +1282,10 @@ static void qeth_free_qdio_buffers(struct qeth_card *card)
 
 	qeth_free_cq(card);
 	cancel_delayed_work_sync(&card->buffer_reclaim_work);
-	for (j = 0; j < QDIO_MAX_BUFFERS_PER_Q; ++j)
-		dev_kfree_skb_any(card->qdio.in_q->bufs[j].rx_skb);
+	for (j = 0; j < QDIO_MAX_BUFFERS_PER_Q; ++j) {
+		if (card->qdio.in_q->bufs[j].rx_skb)
+			dev_kfree_skb_any(card->qdio.in_q->bufs[j].rx_skb);
+	}
 	kfree(card->qdio.in_q);
 	card->qdio.in_q = NULL;
 	/* inbound buffer pool */
@@ -1729,14 +1731,14 @@ static void qeth_configure_blkt_default(struct qeth_card *card, char *prcd)
 	QETH_DBF_TEXT(SETUP, 2, "cfgblkt");
 
 	if (prcd[74] == 0xF0 && prcd[75] == 0xF0 &&
-	    (prcd[76] == 0xF5 || prcd[76] == 0xF6)) {
-		card->info.blkt.time_total = 250;
-		card->info.blkt.inter_packet = 5;
-		card->info.blkt.inter_packet_jumbo = 15;
-	} else {
+	    prcd[76] >= 0xF1 && prcd[76] <= 0xF4) {
 		card->info.blkt.time_total = 0;
 		card->info.blkt.inter_packet = 0;
 		card->info.blkt.inter_packet_jumbo = 0;
+	} else {
+		card->info.blkt.time_total = 250;
+		card->info.blkt.inter_packet = 5;
+		card->info.blkt.inter_packet_jumbo = 15;
 	}
 }
 
@@ -2198,11 +2200,11 @@ static inline int qeth_get_initial_mtu_for_card(struct qeth_card *card)
 		case QETH_LINK_TYPE_LANE_TR:
 			return 2000;
 		default:
-			return 1492;
+			return card->options.layer2 ? 1500 : 1492;
 		}
 	case QETH_CARD_TYPE_OSM:
 	case QETH_CARD_TYPE_OSX:
-		return 1492;
+		return card->options.layer2 ? 1500 : 1492;
 	default:
 		return 1500;
 	}
@@ -2275,9 +2277,10 @@ static int qeth_ulp_enable_cb(struct qeth_card *card, struct qeth_reply *reply,
 		card->info.max_mtu = mtu;
 		card->qdio.in_buf_size = mtu + 2 * PAGE_SIZE;
 	} else {
-		card->info.initial_mtu = qeth_get_initial_mtu_for_card(card);
 		card->info.max_mtu = *(__u16 *)QETH_ULP_ENABLE_RESP_MAX_MTU(
 			iob->data);
+		card->info.initial_mtu = min(card->info.max_mtu,
+					qeth_get_initial_mtu_for_card(card));
 		card->qdio.in_buf_size = QETH_IN_BUF_SIZE_DEFAULT;
 	}
 
diff --git a/drivers/scsi/cxgbi/cxgb4i/cxgb4i.c b/drivers/scsi/cxgbi/cxgb4i/cxgb4i.c
index e659feb..5a9f842 100644
--- a/drivers/scsi/cxgbi/cxgb4i/cxgb4i.c
+++ b/drivers/scsi/cxgbi/cxgb4i/cxgb4i.c
@@ -433,7 +433,7 @@ static inline unsigned int calc_tx_flits_ofld(const struct sk_buff *skb)
 		return DIV_ROUND_UP(skb->len, 8);
 	flits = skb_transport_offset(skb) / 8;
 	cnt = skb_shinfo(skb)->nr_frags;
-	if (skb->tail != skb->transport_header)
+	if (skb_tail_pointer(skb) != skb_transport_header(skb))
 		cnt++;
 	return flits + sgl_len(cnt);
 }
diff --git a/drivers/scsi/fcoe/fcoe.c b/drivers/scsi/fcoe/fcoe.c
index 32ae6c6..4a05d04 100644
--- a/drivers/scsi/fcoe/fcoe.c
+++ b/drivers/scsi/fcoe/fcoe.c
@@ -1978,7 +1978,7 @@ static int fcoe_device_notification(struct notifier_block *notifier,
 {
 	struct fcoe_ctlr_device *cdev;
 	struct fc_lport *lport = NULL;
-	struct net_device *netdev = ptr;
+	struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
 	struct fcoe_ctlr *ctlr;
 	struct fcoe_interface *fcoe;
 	struct fcoe_port *port;
diff --git a/drivers/scsi/fcoe/fcoe_transport.c b/drivers/scsi/fcoe/fcoe_transport.c
index f3a5a53..01adbe0 100644
--- a/drivers/scsi/fcoe/fcoe_transport.c
+++ b/drivers/scsi/fcoe/fcoe_transport.c
@@ -704,7 +704,7 @@ static struct net_device *fcoe_if_to_netdev(const char *buffer)
 static int libfcoe_device_notification(struct notifier_block *notifier,
 				    ulong event, void *ptr)
 {
-	struct net_device *netdev = ptr;
+	struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
 
 	switch (event) {
 	case NETDEV_UNREGISTER:
diff --git a/drivers/scsi/libiscsi_tcp.c b/drivers/scsi/libiscsi_tcp.c
index 92deec5..1d58d53 100644
--- a/drivers/scsi/libiscsi_tcp.c
+++ b/drivers/scsi/libiscsi_tcp.c
@@ -906,7 +906,6 @@ int iscsi_tcp_recv_skb(struct iscsi_conn *conn, struct sk_buff *skb,
 			ISCSI_DBG_TCP(conn, "no more data avail. Consumed %d\n",
 				      consumed);
 			*status = ISCSI_TCP_SKB_DONE;
-			skb_abort_seq_read(&seq);
 			goto skb_done;
 		}
 		BUG_ON(segment->copied >= segment->size);
diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 0a537a0..d055450 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -439,10 +439,7 @@ static int fill_from_dev_buffer(struct scsi_cmnd *scp, unsigned char *arr,
 
 	act_len = sg_copy_from_buffer(sdb->table.sgl, sdb->table.nents,
 				      arr, arr_len);
-	if (sdb->resid)
-		sdb->resid -= act_len;
-	else
-		sdb->resid = scsi_bufflen(scp) - act_len;
+	sdb->resid = scsi_bufflen(scp) - act_len;
 
 	return 0;
 }
@@ -1693,24 +1690,48 @@ static int check_device_access_params(struct sdebug_dev_info *devi,
 	return 0;
 }
 
+/* Returns number of bytes copied or -1 if error. */
 static int do_device_access(struct scsi_cmnd *scmd,
 			    struct sdebug_dev_info *devi,
 			    unsigned long long lba, unsigned int num, int write)
 {
 	int ret;
 	unsigned long long block, rest = 0;
-	int (*func)(struct scsi_cmnd *, unsigned char *, int);
+	struct scsi_data_buffer *sdb;
+	enum dma_data_direction dir;
+	size_t (*func)(struct scatterlist *, unsigned int, void *, size_t,
+		       off_t);
+
+	if (write) {
+		sdb = scsi_out(scmd);
+		dir = DMA_TO_DEVICE;
+		func = sg_pcopy_to_buffer;
+	} else {
+		sdb = scsi_in(scmd);
+		dir = DMA_FROM_DEVICE;
+		func = sg_pcopy_from_buffer;
+	}
 
-	func = write ? fetch_to_dev_buffer : fill_from_dev_buffer;
+	if (!sdb->length)
+		return 0;
+	if (!(scsi_bidi_cmnd(scmd) || scmd->sc_data_direction == dir))
+		return -1;
 
 	block = do_div(lba, sdebug_store_sectors);
 	if (block + num > sdebug_store_sectors)
 		rest = block + num - sdebug_store_sectors;
 
-	ret = func(scmd, fake_storep + (block * scsi_debug_sector_size),
-		   (num - rest) * scsi_debug_sector_size);
-	if (!ret && rest)
-		ret = func(scmd, fake_storep, rest * scsi_debug_sector_size);
+	ret = func(sdb->table.sgl, sdb->table.nents,
+		   fake_storep + (block * scsi_debug_sector_size),
+		   (num - rest) * scsi_debug_sector_size, 0);
+	if (ret != (num - rest) * scsi_debug_sector_size)
+		return ret;
+
+	if (rest) {
+		ret += func(sdb->table.sgl, sdb->table.nents,
+			    fake_storep, rest * scsi_debug_sector_size,
+			    (num - rest) * scsi_debug_sector_size);
+	}
 
 	return ret;
 }
@@ -1849,7 +1870,12 @@ static int resp_read(struct scsi_cmnd *SCpnt, unsigned long long lba,
 	read_lock_irqsave(&atomic_rw, iflags);
 	ret = do_device_access(SCpnt, devip, lba, num, 0);
 	read_unlock_irqrestore(&atomic_rw, iflags);
-	return ret;
+	if (ret == -1)
+		return DID_ERROR << 16;
+
+	scsi_in(SCpnt)->resid = scsi_bufflen(SCpnt) - ret;
+
+	return 0;
 }
 
 void dump_sector(unsigned char *buf, int len)
diff --git a/drivers/ssb/driver_chipcommon_sflash.c b/drivers/ssb/driver_chipcommon_sflash.c
index 720665c..e84cf04 100644
--- a/drivers/ssb/driver_chipcommon_sflash.c
+++ b/drivers/ssb/driver_chipcommon_sflash.c
@@ -9,6 +9,19 @@
 
 #include "ssb_private.h"
 
+static struct resource ssb_sflash_resource = {
+	.name	= "ssb_sflash",
+	.start	= SSB_FLASH2,
+	.end	= 0,
+	.flags  = IORESOURCE_MEM | IORESOURCE_READONLY,
+};
+
+struct platform_device ssb_sflash_dev = {
+	.name		= "ssb_sflash",
+	.resource	= &ssb_sflash_resource,
+	.num_resources	= 1,
+};
+
 struct ssb_sflash_tbl_e {
 	char *name;
 	u32 id;
@@ -16,7 +29,7 @@ struct ssb_sflash_tbl_e {
 	u16 numblocks;
 };
 
-static struct ssb_sflash_tbl_e ssb_sflash_st_tbl[] = {
+static const struct ssb_sflash_tbl_e ssb_sflash_st_tbl[] = {
 	{ "M25P20", 0x11, 0x10000, 4, },
 	{ "M25P40", 0x12, 0x10000, 8, },
 
@@ -27,7 +40,7 @@ static struct ssb_sflash_tbl_e ssb_sflash_st_tbl[] = {
 	{ 0 },
 };
 
-static struct ssb_sflash_tbl_e ssb_sflash_sst_tbl[] = {
+static const struct ssb_sflash_tbl_e ssb_sflash_sst_tbl[] = {
 	{ "SST25WF512", 1, 0x1000, 16, },
 	{ "SST25VF512", 0x48, 0x1000, 16, },
 	{ "SST25WF010", 2, 0x1000, 32, },
@@ -45,7 +58,7 @@ static struct ssb_sflash_tbl_e ssb_sflash_sst_tbl[] = {
 	{ 0 },
 };
 
-static struct ssb_sflash_tbl_e ssb_sflash_at_tbl[] = {
+static const struct ssb_sflash_tbl_e ssb_sflash_at_tbl[] = {
 	{ "AT45DB011", 0xc, 256, 512, },
 	{ "AT45DB021", 0x14, 256, 1024, },
 	{ "AT45DB041", 0x1c, 256, 2048, },
@@ -73,7 +86,8 @@ static void ssb_sflash_cmd(struct ssb_chipcommon *cc, u32 opcode)
 /* Initialize serial flash access */
 int ssb_sflash_init(struct ssb_chipcommon *cc)
 {
-	struct ssb_sflash_tbl_e *e;
+	struct ssb_sflash *sflash = &cc->dev->bus->mipscore.sflash;
+	const struct ssb_sflash_tbl_e *e;
 	u32 id, id2;
 
 	switch (cc->capabilities & SSB_CHIPCO_CAP_FLASHT) {
@@ -131,9 +145,21 @@ int ssb_sflash_init(struct ssb_chipcommon *cc)
 		return -ENOTSUPP;
 	}
 
+	sflash->window = SSB_FLASH2;
+	sflash->blocksize = e->blocksize;
+	sflash->numblocks = e->numblocks;
+	sflash->size = sflash->blocksize * sflash->numblocks;
+	sflash->present = true;
+
 	pr_info("Found %s serial flash (blocksize: 0x%X, blocks: %d)\n",
 		e->name, e->blocksize, e->numblocks);
 
+	/* Prepare platform device, but don't register it yet. It's too early,
+	 * malloc (required by device_private_init) is not available yet. */
+	ssb_sflash_dev.resource[0].end = ssb_sflash_dev.resource[0].start +
+					 sflash->size;
+	ssb_sflash_dev.dev.platform_data = sflash;
+
 	pr_err("Serial flash support is not implemented yet!\n");
 
 	return -ENOTSUPP;
diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
index 812775a..e55ddf7 100644
--- a/drivers/ssb/main.c
+++ b/drivers/ssb/main.c
@@ -553,6 +553,14 @@ static int ssb_devices_register(struct ssb_bus *bus)
 	}
 #endif
 
+#ifdef CONFIG_SSB_SFLASH
+	if (bus->mipscore.sflash.present) {
+		err = platform_device_register(&ssb_sflash_dev);
+		if (err)
+			pr_err("Error registering serial flash\n");
+	}
+#endif
+
 	return 0;
 error:
 	/* Unwind the already registered devices. */
diff --git a/drivers/ssb/pcihost_wrapper.c b/drivers/ssb/pcihost_wrapper.c
index 32ed1fa..69161bb 100644
--- a/drivers/ssb/pcihost_wrapper.c
+++ b/drivers/ssb/pcihost_wrapper.c
@@ -38,7 +38,7 @@ static int ssb_pcihost_resume(struct pci_dev *dev)
 	struct ssb_bus *ssb = pci_get_drvdata(dev);
 	int err;
 
-	pci_set_power_state(dev, 0);
+	pci_set_power_state(dev, PCI_D0);
 	err = pci_enable_device(dev);
 	if (err)
 		return err;
diff --git a/drivers/ssb/sprom.c b/drivers/ssb/sprom.c
index a3b2364..e753fbe 100644
--- a/drivers/ssb/sprom.c
+++ b/drivers/ssb/sprom.c
@@ -54,7 +54,7 @@ static int hex2sprom(u16 *sprom, const char *dump, size_t len,
 	while (cnt < sprom_size_words) {
 		memcpy(tmp, dump, 4);
 		dump += 4;
-		err = strict_strtoul(tmp, 16, &parsed);
+		err = kstrtoul(tmp, 16, &parsed);
 		if (err)
 			return err;
 		sprom[cnt++] = swab16((u16)parsed);
diff --git a/drivers/ssb/ssb_private.h b/drivers/ssb/ssb_private.h
index 4671f17..eb507a5 100644
--- a/drivers/ssb/ssb_private.h
+++ b/drivers/ssb/ssb_private.h
@@ -243,6 +243,10 @@ static inline int ssb_sflash_init(struct ssb_chipcommon *cc)
 extern struct platform_device ssb_pflash_dev;
 #endif
 
+#ifdef CONFIG_SSB_SFLASH
+extern struct platform_device ssb_sflash_dev;
+#endif
+
 #ifdef CONFIG_SSB_DRIVER_EXTIF
 extern u32 ssb_extif_watchdog_timer_set_wdt(struct bcm47xx_wdt *wdt, u32 ticks);
 extern u32 ssb_extif_watchdog_timer_set_ms(struct bcm47xx_wdt *wdt, u32 ms);
diff --git a/drivers/staging/csr/netdev.c b/drivers/staging/csr/netdev.c
index 5ead2d4..9c716c1 100644
--- a/drivers/staging/csr/netdev.c
+++ b/drivers/staging/csr/netdev.c
@@ -2891,7 +2891,7 @@ void uf_net_get_name(struct net_device *dev, char *name, int len)
  */
 static int
 uf_netdev_event(struct notifier_block *notif, unsigned long event, void* ptr) {
-    struct net_device *netdev = ptr;
+    struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
     netInterface_priv_t *interfacePriv = (netInterface_priv_t *)netdev_priv(netdev);
     unifi_priv_t *priv = NULL;
     static const CsrWifiMacAddress broadcast_address = {{0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}};
diff --git a/drivers/staging/ft1000/ft1000-pcmcia/ft1000_proc.c b/drivers/staging/ft1000/ft1000-pcmcia/ft1000_proc.c
index 94e426e..b2330f1 100644
--- a/drivers/staging/ft1000/ft1000-pcmcia/ft1000_proc.c
+++ b/drivers/staging/ft1000/ft1000-pcmcia/ft1000_proc.c
@@ -164,7 +164,7 @@ static const struct file_operations ft1000_proc_fops = {
 static int ft1000NotifyProc(struct notifier_block *this, unsigned long event,
 				void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct ft1000_info *info;
 
 	info = netdev_priv(dev);
diff --git a/drivers/staging/ft1000/ft1000-usb/ft1000_proc.c b/drivers/staging/ft1000/ft1000-usb/ft1000_proc.c
index eca6f02..5ead942 100644
--- a/drivers/staging/ft1000/ft1000-usb/ft1000_proc.c
+++ b/drivers/staging/ft1000/ft1000-usb/ft1000_proc.c
@@ -166,7 +166,7 @@ static const struct file_operations ft1000_proc_fops = {
 static int
 ft1000NotifyProc(struct notifier_block *this, unsigned long event, void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct ft1000_info *info;
 	struct proc_dir_entry *ft1000_proc_file;
 
diff --git a/drivers/staging/imx-drm/ipuv3-crtc.c b/drivers/staging/imx-drm/ipuv3-crtc.c
index 4a7eedf..9176a81 100644
--- a/drivers/staging/imx-drm/ipuv3-crtc.c
+++ b/drivers/staging/imx-drm/ipuv3-crtc.c
@@ -359,17 +359,12 @@ static void ipu_crtc_commit(struct drm_crtc *crtc)
 	ipu_fb_enable(ipu_crtc);
 }
 
-static void ipu_crtc_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static struct drm_crtc_helper_funcs ipu_helper_funcs = {
 	.dpms = ipu_crtc_dpms,
 	.mode_fixup = ipu_crtc_mode_fixup,
 	.mode_set = ipu_crtc_mode_set,
 	.prepare = ipu_crtc_prepare,
 	.commit = ipu_crtc_commit,
-	.load_lut = ipu_crtc_load_lut,
 };
 
 static int ipu_enable_vblank(struct drm_crtc *crtc)
diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h b/drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h
index f050808..a8e9c0c 100644
--- a/drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h
+++ b/drivers/staging/lustre/lustre/include/linux/lustre_patchless_compat.h
@@ -60,8 +60,6 @@ truncate_complete_page(struct address_space *mapping, struct page *page)
 	ll_delete_from_page_cache(page);
 }
 
-#  define d_refcount(d)		 ((d)->d_count)
-
 #ifdef ATTR_OPEN
 # define ATTR_FROM_OPEN ATTR_OPEN
 #else
diff --git a/drivers/staging/lustre/lustre/include/linux/lvfs.h b/drivers/staging/lustre/lustre/include/linux/lvfs.h
index b4db6cb..eb59ac7 100644
--- a/drivers/staging/lustre/lustre/include/linux/lvfs.h
+++ b/drivers/staging/lustre/lustre/include/linux/lvfs.h
@@ -99,7 +99,7 @@ static inline void l_dput(struct dentry *de)
 	if (!de || IS_ERR(de))
 		return;
 	//shrink_dcache_parent(de);
-	LASSERT(d_refcount(de) > 0);
+	LASSERT(d_count(de) > 0);
 	dput(de);
 }
 
diff --git a/drivers/staging/lustre/lustre/include/lprocfs_status.h b/drivers/staging/lustre/lustre/include/lprocfs_status.h
index e770d02..55f1822 100644
--- a/drivers/staging/lustre/lustre/include/lprocfs_status.h
+++ b/drivers/staging/lustre/lustre/include/lprocfs_status.h
@@ -53,7 +53,7 @@ struct lprocfs_vars {
 	/**
 	 * /proc file mode.
 	 */
-	mode_t			proc_mode;
+	umode_t			proc_mode;
 };
 
 struct lprocfs_static_vars {
@@ -600,11 +600,11 @@ extern int lprocfs_obd_setup(struct obd_device *obd, struct lprocfs_vars *list);
 extern int lprocfs_obd_cleanup(struct obd_device *obd);
 
 extern int lprocfs_seq_create(proc_dir_entry_t *parent, const char *name,
-			      mode_t mode,
+			      umode_t mode,
 			      const struct file_operations *seq_fops,
 			      void *data);
 extern int lprocfs_obd_seq_create(struct obd_device *dev, const char *name,
-				  mode_t mode,
+				  umode_t mode,
 				  const struct file_operations *seq_fops,
 				  void *data);
 
diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c
index 7d6abff..ff0d085 100644
--- a/drivers/staging/lustre/lustre/llite/dcache.c
+++ b/drivers/staging/lustre/lustre/llite/dcache.c
@@ -98,7 +98,7 @@ int ll_dcompare(const struct dentry *parent, const struct inode *pinode,
 
 	CDEBUG(D_DENTRY, "found name %.*s(%p) flags %#x refc %d\n",
 	       name->len, name->name, dentry, dentry->d_flags,
-	       d_refcount(dentry));
+	       d_count(dentry));
 
 	/* mountpoint is always valid */
 	if (d_mountpoint((struct dentry *)dentry))
@@ -165,7 +165,7 @@ static int ll_ddelete(const struct dentry *de)
 	       list_empty(&de->d_subdirs) ? "" : "subdirs");
 
 	/* kernel >= 2.6.38 last refcount is decreased after this function. */
-	LASSERT(d_refcount(de) == 1);
+	LASSERT(d_count(de) == 1);
 
 	/* Disable this piece of code temproarily because this is called
 	 * inside dcache_lock so it's not appropriate to do lots of work
@@ -190,7 +190,7 @@ static int ll_set_dd(struct dentry *de)
 
 	CDEBUG(D_DENTRY, "ldd on dentry %.*s (%p) parent %p inode %p refc %d\n",
 		de->d_name.len, de->d_name.name, de, de->d_parent, de->d_inode,
-		d_refcount(de));
+		d_count(de));
 
 	if (de->d_fsdata == NULL) {
 		struct ll_dentry_data *lld;
@@ -540,7 +540,7 @@ out:
 		CDEBUG(D_DENTRY, "revalidated dentry %.*s (%p) parent %p "
 		       "inode %p refc %d\n", de->d_name.len,
 		       de->d_name.name, de, de->d_parent, de->d_inode,
-		       d_refcount(de));
+		       d_count(de));
 
 		ll_set_lock_data(exp, de->d_inode, it, &bits);
 
diff --git a/drivers/staging/lustre/lustre/llite/llite_internal.h b/drivers/staging/lustre/lustre/llite/llite_internal.h
index 992cd20..5227c5c 100644
--- a/drivers/staging/lustre/lustre/llite/llite_internal.h
+++ b/drivers/staging/lustre/lustre/llite/llite_internal.h
@@ -1529,12 +1529,12 @@ static inline void d_lustre_invalidate(struct dentry *dentry, int nested)
 {
 	CDEBUG(D_DENTRY, "invalidate dentry %.*s (%p) parent %p inode %p "
 	       "refc %d\n", dentry->d_name.len, dentry->d_name.name, dentry,
-	       dentry->d_parent, dentry->d_inode, d_refcount(dentry));
+	       dentry->d_parent, dentry->d_inode, d_count(dentry));
 
 	spin_lock_nested(&dentry->d_lock,
 			 nested ? DENTRY_D_LOCK_NESTED : DENTRY_D_LOCK_NORMAL);
 	__d_lustre_invalidate(dentry);
-	if (d_refcount(dentry) == 0)
+	if (d_count(dentry) == 0)
 		__d_drop(dentry);
 	spin_unlock(&dentry->d_lock);
 }
diff --git a/drivers/staging/lustre/lustre/llite/llite_lib.c b/drivers/staging/lustre/lustre/llite/llite_lib.c
index 2311b20..afae801 100644
--- a/drivers/staging/lustre/lustre/llite/llite_lib.c
+++ b/drivers/staging/lustre/lustre/llite/llite_lib.c
@@ -659,7 +659,7 @@ void lustre_dump_dentry(struct dentry *dentry, int recur)
 	       " flags=0x%x, fsdata=%p, %d subdirs\n", dentry,
 	       dentry->d_name.len, dentry->d_name.name,
 	       dentry->d_parent->d_name.len, dentry->d_parent->d_name.name,
-	       dentry->d_parent, dentry->d_inode, d_refcount(dentry),
+	       dentry->d_parent, dentry->d_inode, d_count(dentry),
 	       dentry->d_flags, dentry->d_fsdata, subdirs);
 	if (dentry->d_inode != NULL)
 		ll_dump_inode(dentry->d_inode);
diff --git a/drivers/staging/lustre/lustre/llite/namei.c b/drivers/staging/lustre/lustre/llite/namei.c
index 58d59aa..ff8f63d 100644
--- a/drivers/staging/lustre/lustre/llite/namei.c
+++ b/drivers/staging/lustre/lustre/llite/namei.c
@@ -409,7 +409,7 @@ struct dentry *ll_splice_alias(struct inode *inode, struct dentry *de)
 			iput(inode);
 			CDEBUG(D_DENTRY,
 			       "Reuse dentry %p inode %p refc %d flags %#x\n",
-			      new, new->d_inode, d_refcount(new), new->d_flags);
+			      new, new->d_inode, d_count(new), new->d_flags);
 			return new;
 		}
 	}
@@ -417,7 +417,7 @@ struct dentry *ll_splice_alias(struct inode *inode, struct dentry *de)
 	__d_lustre_invalidate(de);
 	d_add(de, inode);
 	CDEBUG(D_DENTRY, "Add dentry %p inode %p refc %d flags %#x\n",
-	       de, de->d_inode, d_refcount(de), de->d_flags);
+	       de, de->d_inode, d_count(de), de->d_flags);
 	return de;
 }
 
diff --git a/drivers/staging/lustre/lustre/lvfs/lvfs_linux.c b/drivers/staging/lustre/lustre/lvfs/lvfs_linux.c
index 1e6f32c..e70d8fe 100644
--- a/drivers/staging/lustre/lustre/lvfs/lvfs_linux.c
+++ b/drivers/staging/lustre/lustre/lvfs/lvfs_linux.c
@@ -121,8 +121,8 @@ void push_ctxt(struct lvfs_run_ctxt *save, struct lvfs_run_ctxt *new_ctx,
 	OBD_SET_CTXT_MAGIC(save);
 
 	save->fs = get_fs();
-	LASSERT(d_refcount(cfs_fs_pwd(current->fs)));
-	LASSERT(d_refcount(new_ctx->pwd));
+	LASSERT(d_count(cfs_fs_pwd(current->fs)));
+	LASSERT(d_count(new_ctx->pwd));
 	save->pwd = dget(cfs_fs_pwd(current->fs));
 	save->pwdmnt = mntget(cfs_fs_mnt(current->fs));
 	save->luc.luc_umask = current_umask();
diff --git a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
index 3b157f8..f7af3d6 100644
--- a/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
+++ b/drivers/staging/lustre/lustre/obdclass/lprocfs_status.c
@@ -73,7 +73,7 @@ proc_dir_entry_t *lprocfs_add_simple(struct proc_dir_entry *root,
 				     struct file_operations *fops)
 {
 	proc_dir_entry_t *proc;
-	mode_t mode = 0;
+	umode_t mode = 0;
 
 	if (root == NULL || name == NULL || fops == NULL)
 		return ERR_PTR(-EINVAL);
@@ -140,7 +140,7 @@ int lprocfs_add_vars(struct proc_dir_entry *root, struct lprocfs_vars *list,
 
 	while (list->name != NULL) {
 		struct proc_dir_entry *proc;
-		mode_t mode = 0;
+		umode_t mode = 0;
 
 		if (list->proc_mode != 0000) {
 			mode = list->proc_mode;
@@ -1899,7 +1899,7 @@ EXPORT_SYMBOL(lprocfs_find_named_value);
 
 int lprocfs_seq_create(proc_dir_entry_t *parent,
 		       const char *name,
-		       mode_t mode,
+		       umode_t mode,
 		       const struct file_operations *seq_fops,
 		       void *data)
 {
@@ -1919,7 +1919,7 @@ EXPORT_SYMBOL(lprocfs_seq_create);
 
 int lprocfs_obd_seq_create(struct obd_device *dev,
 			   const char *name,
-			   mode_t mode,
+			   umode_t mode,
 			   const struct file_operations *seq_fops,
 			   void *data)
 {
diff --git a/drivers/staging/rtl8192u/r8192U_core.c b/drivers/staging/rtl8192u/r8192U_core.c
index c880adc..14c14c2 100644
--- a/drivers/staging/rtl8192u/r8192U_core.c
+++ b/drivers/staging/rtl8192u/r8192U_core.c
@@ -1144,8 +1144,8 @@ struct sk_buff *DrvAggr_Aggregation(struct net_device *dev, struct ieee80211_drv
 		/* Subframe drv Tx descriptor and firmware info setting */
 		skb = pSendList->tx_agg_frames[i];
 		tcb_desc = (cb_desc *)(skb->cb + MAX_DEV_ADDR_SIZE);
-		tx_agg_desc = (tx_desc_819x_usb_aggr_subframe *)agg_skb->tail;
-		tx_fwinfo = (tx_fwinfo_819x_usb *)(agg_skb->tail + sizeof(tx_desc_819x_usb_aggr_subframe));
+		tx_agg_desc = (tx_desc_819x_usb_aggr_subframe *)skb_tail_pointer(agg_skb);
+		tx_fwinfo = (tx_fwinfo_819x_usb *)(skb_tail_pointer(agg_skb) + sizeof(tx_desc_819x_usb_aggr_subframe));
 
 		memset(tx_fwinfo, 0, sizeof(tx_fwinfo_819x_usb));
 		/* DWORD 0 */
diff --git a/drivers/staging/silicom/Kconfig b/drivers/staging/silicom/Kconfig
index eda2e7d..6651bd8 100644
--- a/drivers/staging/silicom/Kconfig
+++ b/drivers/staging/silicom/Kconfig
@@ -5,7 +5,7 @@
 config NET_VENDOR_SILICOM
 	bool "Silicom devices"
 	default y
-	depends on PCI
+	depends on PCI && NETDEVICES
 	---help---
 	  If you have a network card (Ethernet) belonging to this class,
 	  say Y.
@@ -19,7 +19,7 @@ if NET_VENDOR_SILICOM
 
 config SBYPASS
 	tristate "Silicom BypassCTL library support"
-	depends on PCI && NET
+	depends on PCI
 	depends on m
 	---help---
 	  If you have a network (Ethernet) controller of this type, say Y
@@ -29,10 +29,9 @@ config SBYPASS
 
 config BPCTL
 	tristate "Silicom BypassCTL net support"
-	depends on PCI && NET
+	depends on PCI
 	depends on m
 	select SBYPASS
-	select NET_CORE
 	select MII
 	---help---
 	  If you have a network (Ethernet) controller of this type, say Y
diff --git a/drivers/staging/silicom/bpctl_mod.c b/drivers/staging/silicom/bpctl_mod.c
index 4b3a1ae..48b9fb1 100644
--- a/drivers/staging/silicom/bpctl_mod.c
+++ b/drivers/staging/silicom/bpctl_mod.c
@@ -128,7 +128,7 @@ static unsigned long str_to_hex(char *p);
 static int bp_device_event(struct notifier_block *unused,
 			   unsigned long event, void *ptr)
 {
-	struct net_device *dev = ptr;
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	static bpctl_dev_t *pbpctl_dev, *pbpctl_dev_m;
 	int dev_num = 0, ret = 0, ret_d = 0, time_left = 0;
 	/* printk("BP_PROC_SUPPORT event =%d %s %d\n", event,dev->name, dev->ifindex ); */
diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index f80d3dd..8ca5ac7 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c
@@ -150,6 +150,11 @@ static void vhost_net_ubuf_put_and_wait(struct vhost_net_ubuf_ref *ubufs)
 {
 	kref_put(&ubufs->kref, vhost_net_zerocopy_done_signal);
 	wait_event(ubufs->wait, !atomic_read(&ubufs->kref.refcount));
+}
+
+static void vhost_net_ubuf_put_wait_and_free(struct vhost_net_ubuf_ref *ubufs)
+{
+	vhost_net_ubuf_put_and_wait(ubufs);
 	kfree(ubufs);
 }
 
@@ -948,7 +953,7 @@ static long vhost_net_set_backend(struct vhost_net *n, unsigned index, int fd)
 	mutex_unlock(&vq->mutex);
 
 	if (oldubufs) {
-		vhost_net_ubuf_put_and_wait(oldubufs);
+		vhost_net_ubuf_put_wait_and_free(oldubufs);
 		mutex_lock(&vq->mutex);
 		vhost_zerocopy_signal_used(n, vq);
 		mutex_unlock(&vq->mutex);
@@ -966,7 +971,7 @@ err_used:
 	rcu_assign_pointer(vq->private_data, oldsock);
 	vhost_net_enable_vq(n, vq);
 	if (ubufs)
-		vhost_net_ubuf_put_and_wait(ubufs);
+		vhost_net_ubuf_put_wait_and_free(ubufs);
 err_ubufs:
 	fput(sock->file);
 err_vq:
diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
index 2e937bd..4cf1e1d 100644
--- a/drivers/video/Kconfig
+++ b/drivers/video/Kconfig
@@ -367,6 +367,8 @@ config FB_IMX
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
 	select FB_CFB_IMAGEBLIT
+	select FB_MODE_HELPERS
+	select VIDEOMODE_HELPERS
 
 config FB_CYBER2000
 	tristate "CyberPro 2000/2010/5000 support"
@@ -2188,7 +2190,7 @@ config FB_PS3_DEFAULT_SIZE_M
 
 config FB_XILINX
 	tristate "Xilinx frame buffer support"
-	depends on FB && (XILINX_VIRTEX || MICROBLAZE)
+	depends on FB && (XILINX_VIRTEX || MICROBLAZE || ARCH_ZYNQ)
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
 	select FB_CFB_IMAGEBLIT
diff --git a/drivers/video/aty/aty128fb.c b/drivers/video/aty/aty128fb.c
index 8c55011..a4dfe8c 100644
--- a/drivers/video/aty/aty128fb.c
+++ b/drivers/video/aty/aty128fb.c
@@ -2016,7 +2016,7 @@ static int aty128_init(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	aty128_init_engine(par);
 
-	par->pm_reg = pci_find_capability(pdev, PCI_CAP_ID_PM);
+	par->pm_reg = pdev->pm_cap;
 	par->pdev = pdev;
 	par->asleep = 0;
 	par->lock_blank = 0;
diff --git a/drivers/video/aty/atyfb_base.c b/drivers/video/aty/atyfb_base.c
index 4f27fdc..a89c15d 100644
--- a/drivers/video/aty/atyfb_base.c
+++ b/drivers/video/aty/atyfb_base.c
@@ -58,6 +58,7 @@
 #include <linux/slab.h>
 #include <linux/vmalloc.h>
 #include <linux/delay.h>
+#include <linux/compiler.h>
 #include <linux/console.h>
 #include <linux/fb.h>
 #include <linux/init.h>
@@ -434,8 +435,8 @@ static int correct_chipset(struct atyfb_par *par)
 	const char *name;
 	int i;
 
-	for (i = ARRAY_SIZE(aty_chips) - 1; i >= 0; i--)
-		if (par->pci_id == aty_chips[i].pci_id)
+	for (i = ARRAY_SIZE(aty_chips); i > 0; i--)
+		if (par->pci_id == aty_chips[i - 1].pci_id)
 			break;
 
 	if (i < 0)
@@ -531,8 +532,8 @@ static int correct_chipset(struct atyfb_par *par)
 	return 0;
 }
 
-static char ram_dram[] = "DRAM";
-static char ram_resv[] = "RESV";
+static char ram_dram[] __maybe_unused = "DRAM";
+static char ram_resv[] __maybe_unused = "RESV";
 #ifdef CONFIG_FB_ATY_GX
 static char ram_vram[] = "VRAM";
 #endif /* CONFIG_FB_ATY_GX */
diff --git a/drivers/video/aty/radeon_pm.c b/drivers/video/aty/radeon_pm.c
index 92bda58..f7091ec 100644
--- a/drivers/video/aty/radeon_pm.c
+++ b/drivers/video/aty/radeon_pm.c
@@ -2805,7 +2805,7 @@ static void radeonfb_early_resume(void *data)
 void radeonfb_pm_init(struct radeonfb_info *rinfo, int dynclk, int ignore_devlist, int force_sleep)
 {
 	/* Find PM registers in config space if any*/
-	rinfo->pm_reg = pci_find_capability(rinfo->pdev, PCI_CAP_ID_PM);
+	rinfo->pm_reg = rinfo->pdev->pm_cap;
 
 	/* Enable/Disable dynamic clocks: TODO add sysfs access */
 	if (rinfo->family == CHIP_FAMILY_RS480)
diff --git a/drivers/video/au1100fb.c b/drivers/video/au1100fb.c
index ebeb971..a54ccdc 100644
--- a/drivers/video/au1100fb.c
+++ b/drivers/video/au1100fb.c
@@ -577,7 +577,6 @@ failed:
 	if (fbdev->info.cmap.len != 0) {
 		fb_dealloc_cmap(&fbdev->info.cmap);
 	}
-	platform_set_drvdata(dev, NULL);
 
 	return -ENODEV;
 }
diff --git a/drivers/video/bf54x-lq043fb.c b/drivers/video/bf54x-lq043fb.c
index 2726a5b..87f288b 100644
--- a/drivers/video/bf54x-lq043fb.c
+++ b/drivers/video/bf54x-lq043fb.c
@@ -681,7 +681,6 @@ out3:
 out2:
 	free_dma(CH_EPPI0);
 out1:
-	platform_set_drvdata(pdev, NULL);
 
 	return ret;
 }
diff --git a/drivers/video/bfin-lq035q1-fb.c b/drivers/video/bfin-lq035q1-fb.c
index 29d8c04..b594a58 100644
--- a/drivers/video/bfin-lq035q1-fb.c
+++ b/drivers/video/bfin-lq035q1-fb.c
@@ -170,16 +170,19 @@ static int lq035q1_spidev_remove(struct spi_device *spi)
 	return lq035q1_control(spi, LQ035_SHUT_CTL, LQ035_SHUT);
 }
 
-#ifdef CONFIG_PM
-static int lq035q1_spidev_suspend(struct spi_device *spi, pm_message_t state)
+#ifdef CONFIG_PM_SLEEP
+static int lq035q1_spidev_suspend(struct device *dev)
 {
+	struct spi_device *spi = to_spi_device(dev);
+
 	return lq035q1_control(spi, LQ035_SHUT_CTL, LQ035_SHUT);
 }
 
-static int lq035q1_spidev_resume(struct spi_device *spi)
+static int lq035q1_spidev_resume(struct device *dev)
 {
-	int ret;
+	struct spi_device *spi = to_spi_device(dev);
 	struct spi_control *ctl = spi_get_drvdata(spi);
+	int ret;
 
 	ret = lq035q1_control(spi, LQ035_DRIVER_OUTPUT_CTL, ctl->mode);
 	if (ret)
@@ -187,9 +190,13 @@ static int lq035q1_spidev_resume(struct spi_device *spi)
 
 	return lq035q1_control(spi, LQ035_SHUT_CTL, LQ035_ON);
 }
+
+static SIMPLE_DEV_PM_OPS(lq035q1_spidev_pm_ops, lq035q1_spidev_suspend,
+	lq035q1_spidev_resume);
+#define LQ035Q1_SPIDEV_PM_OPS (&lq035q1_spidev_pm_ops)
+
 #else
-# define lq035q1_spidev_suspend NULL
-# define lq035q1_spidev_resume  NULL
+#define LQ035Q1_SPIDEV_PM_OPS NULL
 #endif
 
 /* Power down all displays on reboot, poweroff or halt */
@@ -708,8 +715,7 @@ static int bfin_lq035q1_probe(struct platform_device *pdev)
 	info->spidrv.probe    = lq035q1_spidev_probe;
 	info->spidrv.remove   = lq035q1_spidev_remove;
 	info->spidrv.shutdown = lq035q1_spidev_shutdown;
-	info->spidrv.suspend  = lq035q1_spidev_suspend;
-	info->spidrv.resume   = lq035q1_spidev_resume;
+	info->spidrv.driver.pm = LQ035Q1_SPIDEV_PM_OPS;
 
 	ret = spi_register_driver(&info->spidrv);
 	if (ret < 0) {
@@ -759,7 +765,6 @@ static int bfin_lq035q1_probe(struct platform_device *pdev)
  out2:
 	free_dma(CH_PPI);
  out1:
-	platform_set_drvdata(pdev, NULL);
 
 	return ret;
 }
@@ -788,7 +793,6 @@ static int bfin_lq035q1_remove(struct platform_device *pdev)
 	bfin_lq035q1_free_ports(info->disp_info->ppi_mode ==
 				USE_RGB565_16_BIT_PPI);
 
-	platform_set_drvdata(pdev, NULL);
 	framebuffer_release(fbinfo);
 
 	dev_info(&pdev->dev, "unregistered LCD driver\n");
diff --git a/drivers/video/bfin-t350mcqb-fb.c b/drivers/video/bfin-t350mcqb-fb.c
index d46da01..48c0c4e 100644
--- a/drivers/video/bfin-t350mcqb-fb.c
+++ b/drivers/video/bfin-t350mcqb-fb.c
@@ -578,7 +578,6 @@ out3:
 out2:
 	free_dma(CH_PPI);
 out1:
-	platform_set_drvdata(pdev, NULL);
 
 	return ret;
 }
@@ -608,7 +607,6 @@ static int bfin_t350mcqb_remove(struct platform_device *pdev)
 
 	bfin_t350mcqb_request_ports(0);
 
-	platform_set_drvdata(pdev, NULL);
 	framebuffer_release(fbinfo);
 
 	printk(KERN_INFO DRIVER_NAME ": Unregister LCD driver.\n");
diff --git a/drivers/video/console/vgacon.c b/drivers/video/console/vgacon.c
index 5855d17..9d8feac 100644
--- a/drivers/video/console/vgacon.c
+++ b/drivers/video/console/vgacon.c
@@ -42,6 +42,7 @@
 #include <linux/kd.h>
 #include <linux/slab.h>
 #include <linux/vt_kern.h>
+#include <linux/sched.h>
 #include <linux/selection.h>
 #include <linux/spinlock.h>
 #include <linux/ioport.h>
@@ -1124,11 +1125,15 @@ static int vgacon_do_font_op(struct vgastate *state,char *arg,int set,int ch512)
 
 	if (arg) {
 		if (set)
-			for (i = 0; i < cmapsz; i++)
+			for (i = 0; i < cmapsz; i++) {
 				vga_writeb(arg[i], charmap + i);
+				cond_resched();
+			}
 		else
-			for (i = 0; i < cmapsz; i++)
+			for (i = 0; i < cmapsz; i++) {
 				arg[i] = vga_readb(charmap + i);
+				cond_resched();
+			}
 
 		/*
 		 * In 512-character mode, the character map is not contiguous if
@@ -1139,11 +1144,15 @@ static int vgacon_do_font_op(struct vgastate *state,char *arg,int set,int ch512)
 			charmap += 2 * cmapsz;
 			arg += cmapsz;
 			if (set)
-				for (i = 0; i < cmapsz; i++)
+				for (i = 0; i < cmapsz; i++) {
 					vga_writeb(arg[i], charmap + i);
+					cond_resched();
+				}
 			else
-				for (i = 0; i < cmapsz; i++)
+				for (i = 0; i < cmapsz; i++) {
 					arg[i] = vga_readb(charmap + i);
+					cond_resched();
+				}
 		}
 	}
 
diff --git a/drivers/video/ep93xx-fb.c b/drivers/video/ep93xx-fb.c
index ee1ee54..28a837d 100644
--- a/drivers/video/ep93xx-fb.c
+++ b/drivers/video/ep93xx-fb.c
@@ -595,7 +595,6 @@ failed_videomem:
 	fb_dealloc_cmap(&info->cmap);
 failed_cmap:
 	kfree(info);
-	platform_set_drvdata(pdev, NULL);
 
 	return err;
 }
@@ -614,7 +613,6 @@ static int ep93xxfb_remove(struct platform_device *pdev)
 		fbi->mach_info->teardown(pdev);
 
 	kfree(info);
-	platform_set_drvdata(pdev, NULL);
 
 	return 0;
 }
diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
index 098bfc6..36e1fe2 100644
--- a/drivers/video/fbmem.c
+++ b/drivers/video/fbmem.c
@@ -1305,7 +1305,9 @@ static int do_fscreeninfo_to_user(struct fb_fix_screeninfo *fix,
 	err |= copy_to_user(fix32->reserved, fix->reserved,
 			    sizeof(fix->reserved));
 
-	return err;
+	if (err)
+		return -EFAULT;
+	return 0;
 }
 
 static int fb_get_fscreeninfo(struct fb_info *info, unsigned int cmd,
@@ -1881,7 +1883,7 @@ static int ofonly __read_mostly;
  *
  * NOTE: Needed to maintain backwards compatibility
  */
-int fb_get_options(char *name, char **option)
+int fb_get_options(const char *name, char **option)
 {
 	char *opt, *options = NULL;
 	int retval = 0;
diff --git a/drivers/video/fsl-diu-fb.c b/drivers/video/fsl-diu-fb.c
index 6c27805..6dd7225 100644
--- a/drivers/video/fsl-diu-fb.c
+++ b/drivers/video/fsl-diu-fb.c
@@ -469,7 +469,7 @@ static enum fsl_diu_monitor_port fsl_diu_name_to_port(const char *s)
 	unsigned long val;
 
 	if (s) {
-		if (!strict_strtoul(s, 10, &val) && (val <= 2))
+		if (!kstrtoul(s, 10, &val) && (val <= 2))
 			port = (enum fsl_diu_monitor_port) val;
 		else if (strncmp(s, "lvds", 4) == 0)
 			port = FSL_DIU_PORT_LVDS;
@@ -1853,7 +1853,7 @@ static int __init fsl_diu_setup(char *options)
 		if (!strncmp(opt, "monitor=", 8)) {
 			monitor_port = fsl_diu_name_to_port(opt + 8);
 		} else if (!strncmp(opt, "bpp=", 4)) {
-			if (!strict_strtoul(opt + 4, 10, &val))
+			if (!kstrtoul(opt + 4, 10, &val))
 				default_bpp = val;
 		} else
 			fb_mode = opt;
diff --git a/drivers/video/i740fb.c b/drivers/video/i740fb.c
index cfd0c52..6c48388 100644
--- a/drivers/video/i740fb.c
+++ b/drivers/video/i740fb.c
@@ -1302,7 +1302,7 @@ static int  __init i740fb_setup(char *options)
 }
 #endif
 
-int __init i740fb_init(void)
+static int __init i740fb_init(void)
 {
 #ifndef MODULE
 	char *option = NULL;
diff --git a/drivers/video/imxfb.c b/drivers/video/imxfb.c
index 0abf2bf..38733ac 100644
--- a/drivers/video/imxfb.c
+++ b/drivers/video/imxfb.c
@@ -31,6 +31,12 @@
 #include <linux/dma-mapping.h>
 #include <linux/io.h>
 #include <linux/math64.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+
+#include <video/of_display_timing.h>
+#include <video/of_videomode.h>
+#include <video/videomode.h>
 
 #include <linux/platform_data/video-imxfb.h>
 
@@ -112,10 +118,11 @@
 #define LCDISR_EOF	(1<<1)
 #define LCDISR_BOF	(1<<0)
 
+#define IMXFB_LSCR1_DEFAULT 0x00120300
+
 /* Used fb-mode. Can be set on kernel command line, therefore file-static. */
 static const char *fb_mode;
 
-
 /*
  * These are the bitfields for each
  * display depth that we support.
@@ -187,6 +194,19 @@ static struct platform_device_id imxfb_devtype[] = {
 };
 MODULE_DEVICE_TABLE(platform, imxfb_devtype);
 
+static struct of_device_id imxfb_of_dev_id[] = {
+	{
+		.compatible = "fsl,imx1-fb",
+		.data = &imxfb_devtype[IMX1_FB],
+	}, {
+		.compatible = "fsl,imx21-fb",
+		.data = &imxfb_devtype[IMX21_FB],
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(of, imxfb_of_dev_id);
+
 static inline int is_imx1_fb(struct imxfb_info *fbi)
 {
 	return fbi->devtype == IMX1_FB;
@@ -319,6 +339,9 @@ static const struct imx_fb_videomode *imxfb_find_mode(struct imxfb_info *fbi)
 	struct imx_fb_videomode *m;
 	int i;
 
+	if (!fb_mode)
+		return &fbi->mode[0];
+
 	for (i = 0, m = &fbi->mode[0]; i < fbi->num_modes; i++, m++) {
 		if (!strcmp(m->mode.name, fb_mode))
 			return m;
@@ -474,6 +497,9 @@ static int imxfb_bl_update_status(struct backlight_device *bl)
 	struct imxfb_info *fbi = bl_get_data(bl);
 	int brightness = bl->props.brightness;
 
+	if (!fbi->pwmr)
+		return 0;
+
 	if (bl->props.power != FB_BLANK_UNBLANK)
 		brightness = 0;
 	if (bl->props.fb_blank != FB_BLANK_UNBLANK)
@@ -684,10 +710,14 @@ static int imxfb_activate_var(struct fb_var_screeninfo *var, struct fb_info *inf
 
 	writel(fbi->pcr, fbi->regs + LCDC_PCR);
 #ifndef PWMR_BACKLIGHT_AVAILABLE
-	writel(fbi->pwmr, fbi->regs + LCDC_PWMR);
+	if (fbi->pwmr)
+		writel(fbi->pwmr, fbi->regs + LCDC_PWMR);
 #endif
 	writel(fbi->lscr1, fbi->regs + LCDC_LSCR1);
-	writel(fbi->dmacr, fbi->regs + LCDC_DMACR);
+
+	/* dmacr = 0 is no valid value, as we need DMA control marks. */
+	if (fbi->dmacr)
+		writel(fbi->dmacr, fbi->regs + LCDC_DMACR);
 
 	return 0;
 }
@@ -723,13 +753,12 @@ static int imxfb_resume(struct platform_device *dev)
 #define imxfb_resume	NULL
 #endif
 
-static int __init imxfb_init_fbinfo(struct platform_device *pdev)
+static int imxfb_init_fbinfo(struct platform_device *pdev)
 {
 	struct imx_fb_platform_data *pdata = pdev->dev.platform_data;
 	struct fb_info *info = dev_get_drvdata(&pdev->dev);
 	struct imxfb_info *fbi = info->par;
-	struct imx_fb_videomode *m;
-	int i;
+	struct device_node *np;
 
 	pr_debug("%s\n",__func__);
 
@@ -760,41 +789,95 @@ static int __init imxfb_init_fbinfo(struct platform_device *pdev)
 	info->fbops			= &imxfb_ops;
 	info->flags			= FBINFO_FLAG_DEFAULT |
 					  FBINFO_READS_FAST;
-	info->var.grayscale		= pdata->cmap_greyscale;
-	fbi->cmap_inverse		= pdata->cmap_inverse;
-	fbi->cmap_static		= pdata->cmap_static;
-	fbi->lscr1			= pdata->lscr1;
-	fbi->dmacr			= pdata->dmacr;
-	fbi->pwmr			= pdata->pwmr;
-	fbi->lcd_power			= pdata->lcd_power;
-	fbi->backlight_power		= pdata->backlight_power;
-
-	for (i = 0, m = &pdata->mode[0]; i < pdata->num_modes; i++, m++)
-		info->fix.smem_len = max_t(size_t, info->fix.smem_len,
-				m->mode.xres * m->mode.yres * m->bpp / 8);
+	if (pdata) {
+		info->var.grayscale		= pdata->cmap_greyscale;
+		fbi->cmap_inverse		= pdata->cmap_inverse;
+		fbi->cmap_static		= pdata->cmap_static;
+		fbi->lscr1			= pdata->lscr1;
+		fbi->dmacr			= pdata->dmacr;
+		fbi->pwmr			= pdata->pwmr;
+		fbi->lcd_power			= pdata->lcd_power;
+		fbi->backlight_power		= pdata->backlight_power;
+	} else {
+		np = pdev->dev.of_node;
+		info->var.grayscale = of_property_read_bool(np,
+						"cmap-greyscale");
+		fbi->cmap_inverse = of_property_read_bool(np, "cmap-inverse");
+		fbi->cmap_static = of_property_read_bool(np, "cmap-static");
+
+		fbi->lscr1 = IMXFB_LSCR1_DEFAULT;
+		of_property_read_u32(np, "fsl,lscr1", &fbi->lscr1);
+
+		of_property_read_u32(np, "fsl,dmacr", &fbi->dmacr);
+
+		/* These two function pointers could be used by some specific
+		 * platforms. */
+		fbi->lcd_power = NULL;
+		fbi->backlight_power = NULL;
+	}
 
 	return 0;
 }
 
-static int __init imxfb_probe(struct platform_device *pdev)
+static int imxfb_of_read_mode(struct device *dev, struct device_node *np,
+		struct imx_fb_videomode *imxfb_mode)
+{
+	int ret;
+	struct fb_videomode *of_mode = &imxfb_mode->mode;
+	u32 bpp;
+	u32 pcr;
+
+	ret = of_property_read_string(np, "model", &of_mode->name);
+	if (ret)
+		of_mode->name = NULL;
+
+	ret = of_get_fb_videomode(np, of_mode, OF_USE_NATIVE_MODE);
+	if (ret) {
+		dev_err(dev, "Failed to get videomode from DT\n");
+		return ret;
+	}
+
+	ret = of_property_read_u32(np, "bits-per-pixel", &bpp);
+	ret |= of_property_read_u32(np, "fsl,pcr", &pcr);
+
+	if (ret) {
+		dev_err(dev, "Failed to read bpp and pcr from DT\n");
+		return -EINVAL;
+	}
+
+	if (bpp < 1 || bpp > 255) {
+		dev_err(dev, "Bits per pixel have to be between 1 and 255\n");
+		return -EINVAL;
+	}
+
+	imxfb_mode->bpp = bpp;
+	imxfb_mode->pcr = pcr;
+
+	return 0;
+}
+
+static int imxfb_probe(struct platform_device *pdev)
 {
 	struct imxfb_info *fbi;
 	struct fb_info *info;
 	struct imx_fb_platform_data *pdata;
 	struct resource *res;
+	struct imx_fb_videomode *m;
+	const struct of_device_id *of_id;
 	int ret, i;
+	int bytes_per_pixel;
 
 	dev_info(&pdev->dev, "i.MX Framebuffer driver\n");
 
+	of_id = of_match_device(imxfb_of_dev_id, &pdev->dev);
+	if (of_id)
+		pdev->id_entry = of_id->data;
+
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!res)
 		return -ENODEV;
 
 	pdata = pdev->dev.platform_data;
-	if (!pdata) {
-		dev_err(&pdev->dev,"No platform_data available\n");
-		return -ENOMEM;
-	}
 
 	info = framebuffer_alloc(sizeof(struct imxfb_info), &pdev->dev);
 	if (!info)
@@ -802,15 +885,55 @@ static int __init imxfb_probe(struct platform_device *pdev)
 
 	fbi = info->par;
 
-	if (!fb_mode)
-		fb_mode = pdata->mode[0].mode.name;
-
 	platform_set_drvdata(pdev, info);
 
 	ret = imxfb_init_fbinfo(pdev);
 	if (ret < 0)
 		goto failed_init;
 
+	if (pdata) {
+		if (!fb_mode)
+			fb_mode = pdata->mode[0].mode.name;
+
+		fbi->mode = pdata->mode;
+		fbi->num_modes = pdata->num_modes;
+	} else {
+		struct device_node *display_np;
+		fb_mode = NULL;
+
+		display_np = of_parse_phandle(pdev->dev.of_node, "display", 0);
+		if (!display_np) {
+			dev_err(&pdev->dev, "No display defined in devicetree\n");
+			ret = -EINVAL;
+			goto failed_of_parse;
+		}
+
+		/*
+		 * imxfb does not support more modes, we choose only the native
+		 * mode.
+		 */
+		fbi->num_modes = 1;
+
+		fbi->mode = devm_kzalloc(&pdev->dev,
+				sizeof(struct imx_fb_videomode), GFP_KERNEL);
+		if (!fbi->mode) {
+			ret = -ENOMEM;
+			goto failed_of_parse;
+		}
+
+		ret = imxfb_of_read_mode(&pdev->dev, display_np, fbi->mode);
+		if (ret)
+			goto failed_of_parse;
+	}
+
+	/* Calculate maximum bytes used per pixel. In most cases this should
+	 * be the same as m->bpp/8 */
+	m = &fbi->mode[0];
+	bytes_per_pixel = (m->bpp + 7) / 8;
+	for (i = 0; i < fbi->num_modes; i++, m++)
+		info->fix.smem_len = max_t(size_t, info->fix.smem_len,
+				m->mode.xres * m->mode.yres * bytes_per_pixel);
+
 	res = request_mem_region(res->start, resource_size(res),
 				DRIVER_NAME);
 	if (!res) {
@@ -843,7 +966,8 @@ static int __init imxfb_probe(struct platform_device *pdev)
 		goto failed_ioremap;
 	}
 
-	if (!pdata->fixed_screen_cpu) {
+	/* Seems not being used by anyone, so no support for oftree */
+	if (!pdata || !pdata->fixed_screen_cpu) {
 		fbi->map_size = PAGE_ALIGN(info->fix.smem_len);
 		fbi->map_cpu = dma_alloc_writecombine(&pdev->dev,
 				fbi->map_size, &fbi->map_dma, GFP_KERNEL);
@@ -868,18 +992,16 @@ static int __init imxfb_probe(struct platform_device *pdev)
 		info->fix.smem_start = fbi->screen_dma;
 	}
 
-	if (pdata->init) {
+	if (pdata && pdata->init) {
 		ret = pdata->init(fbi->pdev);
 		if (ret)
 			goto failed_platform_init;
 	}
 
-	fbi->mode = pdata->mode;
-	fbi->num_modes = pdata->num_modes;
 
 	INIT_LIST_HEAD(&info->modelist);
-	for (i = 0; i < pdata->num_modes; i++)
-		fb_add_videomode(&pdata->mode[i].mode, &info->modelist);
+	for (i = 0; i < fbi->num_modes; i++)
+		fb_add_videomode(&fbi->mode[i].mode, &info->modelist);
 
 	/*
 	 * This makes sure that our colour bitfield
@@ -909,10 +1031,10 @@ static int __init imxfb_probe(struct platform_device *pdev)
 failed_register:
 	fb_dealloc_cmap(&info->cmap);
 failed_cmap:
-	if (pdata->exit)
+	if (pdata && pdata->exit)
 		pdata->exit(fbi->pdev);
 failed_platform_init:
-	if (!pdata->fixed_screen_cpu)
+	if (pdata && !pdata->fixed_screen_cpu)
 		dma_free_writecombine(&pdev->dev,fbi->map_size,fbi->map_cpu,
 			fbi->map_dma);
 failed_map:
@@ -921,9 +1043,9 @@ failed_ioremap:
 failed_getclock:
 	release_mem_region(res->start, resource_size(res));
 failed_req:
+failed_of_parse:
 	kfree(info->pseudo_palette);
 failed_init:
-	platform_set_drvdata(pdev, NULL);
 	framebuffer_release(info);
 	return ret;
 }
@@ -945,7 +1067,7 @@ static int imxfb_remove(struct platform_device *pdev)
 	unregister_framebuffer(info);
 
 	pdata = pdev->dev.platform_data;
-	if (pdata->exit)
+	if (pdata && pdata->exit)
 		pdata->exit(fbi->pdev);
 
 	fb_dealloc_cmap(&info->cmap);
@@ -955,12 +1077,10 @@ static int imxfb_remove(struct platform_device *pdev)
 	iounmap(fbi->regs);
 	release_mem_region(res->start, resource_size(res));
 
-	platform_set_drvdata(pdev, NULL);
-
 	return 0;
 }
 
-void  imxfb_shutdown(struct platform_device * dev)
+static void imxfb_shutdown(struct platform_device *dev)
 {
 	struct fb_info *info = platform_get_drvdata(dev);
 	struct imxfb_info *fbi = info->par;
@@ -974,6 +1094,7 @@ static struct platform_driver imxfb_driver = {
 	.shutdown	= imxfb_shutdown,
 	.driver		= {
 		.name	= DRIVER_NAME,
+		.of_match_table = imxfb_of_dev_id,
 	},
 	.id_table	= imxfb_devtype,
 };
@@ -999,7 +1120,7 @@ static int imxfb_setup(void)
 	return 0;
 }
 
-int __init imxfb_init(void)
+static int __init imxfb_init(void)
 {
 	int ret = imxfb_setup();
 
diff --git a/drivers/video/jz4740_fb.c b/drivers/video/jz4740_fb.c
index 36979b4..2c49112 100644
--- a/drivers/video/jz4740_fb.c
+++ b/drivers/video/jz4740_fb.c
@@ -737,8 +737,6 @@ static int jzfb_remove(struct platform_device *pdev)
 	fb_dealloc_cmap(&jzfb->fb->cmap);
 	jzfb_free_devmem(jzfb);
 
-	platform_set_drvdata(pdev, NULL);
-
 	framebuffer_release(jzfb->fb);
 
 	return 0;
diff --git a/drivers/video/mmp/fb/mmpfb.c b/drivers/video/mmp/fb/mmpfb.c
index 6d1fa96..4ab95b8 100644
--- a/drivers/video/mmp/fb/mmpfb.c
+++ b/drivers/video/mmp/fb/mmpfb.c
@@ -659,7 +659,6 @@ failed_destroy_mutex:
 	mutex_destroy(&fbi->access_ok);
 failed:
 	dev_err(fbi->dev, "mmp-fb: frame buffer device init failed\n");
-	platform_set_drvdata(pdev, NULL);
 
 	framebuffer_release(info);
 
diff --git a/drivers/video/mmp/hw/mmp_ctrl.c b/drivers/video/mmp/hw/mmp_ctrl.c
index 4bd31b2..75dca19 100644
--- a/drivers/video/mmp/hw/mmp_ctrl.c
+++ b/drivers/video/mmp/hw/mmp_ctrl.c
@@ -165,9 +165,9 @@ static void overlay_set_win(struct mmp_overlay *overlay, struct mmp_win *win)
 
 static void dmafetch_onoff(struct mmp_overlay *overlay, int on)
 {
-	u32 mask = overlay_is_vid(overlay) ? CFG_GRA_ENA_MASK :
-		   CFG_DMA_ENA_MASK;
-	u32 enable = overlay_is_vid(overlay) ? CFG_GRA_ENA(1) : CFG_DMA_ENA(1);
+	u32 mask = overlay_is_vid(overlay) ? CFG_DMA_ENA_MASK :
+		   CFG_GRA_ENA_MASK;
+	u32 enable = overlay_is_vid(overlay) ? CFG_DMA_ENA(1) : CFG_GRA_ENA(1);
 	u32 tmp;
 	struct mmp_path *path = overlay->path;
 
@@ -238,7 +238,7 @@ static int overlay_set_addr(struct mmp_overlay *overlay, struct mmp_addr *addr)
 	struct lcd_regs *regs = path_regs(overlay->path);
 
 	/* FIXME: assert addr supported */
-	memcpy(&overlay->addr, addr, sizeof(struct mmp_win));
+	memcpy(&overlay->addr, addr, sizeof(struct mmp_addr));
 	writel(addr->phys[0], &regs->g_0);
 
 	return overlay->addr.phys[0];
@@ -566,7 +566,6 @@ failed:
 		devm_kfree(ctrl->dev, ctrl);
 	}
 
-	platform_set_drvdata(pdev, NULL);
 	dev_err(&pdev->dev, "device init failed\n");
 
 	return ret;
diff --git a/drivers/video/mxsfb.c b/drivers/video/mxsfb.c
index 21223d4..3ba3771 100644
--- a/drivers/video/mxsfb.c
+++ b/drivers/video/mxsfb.c
@@ -899,7 +899,6 @@ static int mxsfb_probe(struct platform_device *pdev)
 
 	host->base = devm_ioremap_resource(&pdev->dev, res);
 	if (IS_ERR(host->base)) {
-		dev_err(&pdev->dev, "ioremap failed\n");
 		ret = PTR_ERR(host->base);
 		goto fb_release;
 	}
@@ -986,8 +985,6 @@ static int mxsfb_remove(struct platform_device *pdev)
 
 	framebuffer_release(fb_info);
 
-	platform_set_drvdata(pdev, NULL);
-
 	return 0;
 }
 
diff --git a/drivers/video/nuc900fb.c b/drivers/video/nuc900fb.c
index 32581c7..8c527e5 100644
--- a/drivers/video/nuc900fb.c
+++ b/drivers/video/nuc900fb.c
@@ -707,7 +707,6 @@ static int nuc900fb_remove(struct platform_device *pdev)
 	release_resource(fbi->mem);
 	kfree(fbi->mem);
 
-	platform_set_drvdata(pdev, NULL);
 	framebuffer_release(fbinfo);
 
 	return 0;
diff --git a/drivers/video/of_display_timing.c b/drivers/video/of_display_timing.c
index 56009bc..171821d 100644
--- a/drivers/video/of_display_timing.c
+++ b/drivers/video/of_display_timing.c
@@ -23,7 +23,7 @@
  * Every display_timing can be specified with either just the typical value or
  * a range consisting of min/typ/max. This function helps handling this
  **/
-static int parse_timing_property(struct device_node *np, const char *name,
+static int parse_timing_property(const struct device_node *np, const char *name,
 			  struct timing_entry *result)
 {
 	struct property *prop;
@@ -53,21 +53,16 @@ static int parse_timing_property(struct device_node *np, const char *name,
 }
 
 /**
- * of_get_display_timing - parse display_timing entry from device_node
+ * of_parse_display_timing - parse display_timing entry from device_node
  * @np: device_node with the properties
  **/
-static struct display_timing *of_get_display_timing(struct device_node *np)
+static int of_parse_display_timing(const struct device_node *np,
+		struct display_timing *dt)
 {
-	struct display_timing *dt;
 	u32 val = 0;
 	int ret = 0;
 
-	dt = kzalloc(sizeof(*dt), GFP_KERNEL);
-	if (!dt) {
-		pr_err("%s: could not allocate display_timing struct\n",
-			of_node_full_name(np));
-		return NULL;
-	}
+	memset(dt, 0, sizeof(*dt));
 
 	ret |= parse_timing_property(np, "hback-porch", &dt->hback_porch);
 	ret |= parse_timing_property(np, "hfront-porch", &dt->hfront_porch);
@@ -97,16 +92,44 @@ static struct display_timing *of_get_display_timing(struct device_node *np)
 		dt->flags |= DISPLAY_FLAGS_INTERLACED;
 	if (of_property_read_bool(np, "doublescan"))
 		dt->flags |= DISPLAY_FLAGS_DOUBLESCAN;
+	if (of_property_read_bool(np, "doubleclk"))
+		dt->flags |= DISPLAY_FLAGS_DOUBLECLK;
 
 	if (ret) {
 		pr_err("%s: error reading timing properties\n",
 			of_node_full_name(np));
-		kfree(dt);
-		return NULL;
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * of_get_display_timing - parse a display_timing entry
+ * @np: device_node with the timing subnode
+ * @name: name of the timing node
+ * @dt: display_timing struct to fill
+ **/
+int of_get_display_timing(struct device_node *np, const char *name,
+		struct display_timing *dt)
+{
+	struct device_node *timing_np;
+
+	if (!np) {
+		pr_err("%s: no devicenode given\n", of_node_full_name(np));
+		return -EINVAL;
 	}
 
-	return dt;
+	timing_np = of_find_node_by_name(np, name);
+	if (!timing_np) {
+		pr_err("%s: could not find node '%s'\n",
+			of_node_full_name(np), name);
+		return -ENOENT;
+	}
+
+	return of_parse_display_timing(timing_np, dt);
 }
+EXPORT_SYMBOL_GPL(of_get_display_timing);
 
 /**
  * of_get_display_timings - parse all display_timing entries from a device_node
@@ -174,9 +197,17 @@ struct display_timings *of_get_display_timings(struct device_node *np)
 
 	for_each_child_of_node(timings_np, entry) {
 		struct display_timing *dt;
+		int r;
 
-		dt = of_get_display_timing(entry);
+		dt = kzalloc(sizeof(*dt), GFP_KERNEL);
 		if (!dt) {
+			pr_err("%s: could not allocate display_timing struct\n",
+					of_node_full_name(np));
+			goto timingfail;
+		}
+
+		r = of_parse_display_timing(entry, dt);
+		if (r) {
 			/*
 			 * to not encourage wrong devicetrees, fail in case of
 			 * an error
diff --git a/drivers/video/omap2/Kconfig b/drivers/video/omap2/Kconfig
index b07b2b0..56cad0f 100644
--- a/drivers/video/omap2/Kconfig
+++ b/drivers/video/omap2/Kconfig
@@ -6,5 +6,6 @@ if ARCH_OMAP2PLUS
 source "drivers/video/omap2/dss/Kconfig"
 source "drivers/video/omap2/omapfb/Kconfig"
 source "drivers/video/omap2/displays/Kconfig"
+source "drivers/video/omap2/displays-new/Kconfig"
 
 endif
diff --git a/drivers/video/omap2/Makefile b/drivers/video/omap2/Makefile
index 296e5c5..86873c2 100644
--- a/drivers/video/omap2/Makefile
+++ b/drivers/video/omap2/Makefile
@@ -2,4 +2,5 @@ obj-$(CONFIG_OMAP2_VRFB) += vrfb.o
 
 obj-$(CONFIG_OMAP2_DSS) += dss/
 obj-y += displays/
+obj-y += displays-new/
 obj-$(CONFIG_FB_OMAP2) += omapfb/
diff --git a/drivers/video/omap2/displays-new/Kconfig b/drivers/video/omap2/displays-new/Kconfig
new file mode 100644
index 0000000..6c90885
--- /dev/null
+++ b/drivers/video/omap2/displays-new/Kconfig
@@ -0,0 +1,73 @@
+menu "OMAP Display Device Drivers (new device model)"
+        depends on OMAP2_DSS
+
+config DISPLAY_ENCODER_TFP410
+        tristate "TFP410 DPI to DVI Encoder"
+	help
+	  Driver for TFP410 DPI to DVI encoder.
+
+config DISPLAY_ENCODER_TPD12S015
+        tristate "TPD12S015 HDMI ESD protection and level shifter"
+	help
+	  Driver for TPD12S015, which offers HDMI ESD protection and level
+	  shifting.
+
+config DISPLAY_CONNECTOR_DVI
+        tristate "DVI Connector"
+	depends on I2C
+	help
+	  Driver for a generic DVI connector.
+
+config DISPLAY_CONNECTOR_HDMI
+        tristate "HDMI Connector"
+	help
+	  Driver for a generic HDMI connector.
+
+config DISPLAY_CONNECTOR_ANALOG_TV
+        tristate "Analog TV Connector"
+	help
+	  Driver for a generic analog TV connector.
+
+config DISPLAY_PANEL_DPI
+	tristate "Generic DPI panel"
+	help
+	  Driver for generic DPI panels.
+
+config DISPLAY_PANEL_DSI_CM
+	tristate "Generic DSI Command Mode Panel"
+	help
+	  Driver for generic DSI command mode panels.
+
+config DISPLAY_PANEL_SONY_ACX565AKM
+	tristate "ACX565AKM Panel"
+	depends on SPI && BACKLIGHT_CLASS_DEVICE
+	help
+	  This is the LCD panel used on Nokia N900
+
+config DISPLAY_PANEL_LGPHILIPS_LB035Q02
+	tristate "LG.Philips LB035Q02 LCD Panel"
+	depends on SPI
+	help
+	  LCD Panel used on the Gumstix Overo Palo35
+
+config DISPLAY_PANEL_SHARP_LS037V7DW01
+        tristate "Sharp LS037V7DW01 LCD Panel"
+        depends on BACKLIGHT_CLASS_DEVICE
+        help
+          LCD Panel used in TI's SDP3430 and EVM boards
+
+config DISPLAY_PANEL_TPO_TD043MTEA1
+        tristate "TPO TD043MTEA1 LCD Panel"
+        depends on SPI
+        help
+          LCD Panel used in OMAP3 Pandora
+
+config DISPLAY_PANEL_NEC_NL8048HL11
+	tristate "NEC NL8048HL11 Panel"
+	depends on SPI
+	depends on BACKLIGHT_CLASS_DEVICE
+	help
+		This NEC NL8048HL11 panel is TFT LCD used in the
+		Zoom2/3/3630 sdp boards.
+
+endmenu
diff --git a/drivers/video/omap2/displays-new/Makefile b/drivers/video/omap2/displays-new/Makefile
new file mode 100644
index 0000000..5aeb11b
--- /dev/null
+++ b/drivers/video/omap2/displays-new/Makefile
@@ -0,0 +1,12 @@
+obj-$(CONFIG_DISPLAY_ENCODER_TFP410) += encoder-tfp410.o
+obj-$(CONFIG_DISPLAY_ENCODER_TPD12S015) += encoder-tpd12s015.o
+obj-$(CONFIG_DISPLAY_CONNECTOR_DVI) += connector-dvi.o
+obj-$(CONFIG_DISPLAY_CONNECTOR_HDMI) += connector-hdmi.o
+obj-$(CONFIG_DISPLAY_CONNECTOR_ANALOG_TV) += connector-analog-tv.o
+obj-$(CONFIG_DISPLAY_PANEL_DPI) += panel-dpi.o
+obj-$(CONFIG_DISPLAY_PANEL_DSI_CM) += panel-dsi-cm.o
+obj-$(CONFIG_DISPLAY_PANEL_SONY_ACX565AKM) += panel-sony-acx565akm.o
+obj-$(CONFIG_DISPLAY_PANEL_LGPHILIPS_LB035Q02) += panel-lgphilips-lb035q02.o
+obj-$(CONFIG_DISPLAY_PANEL_SHARP_LS037V7DW01) += panel-sharp-ls037v7dw01.o
+obj-$(CONFIG_DISPLAY_PANEL_TPO_TD043MTEA1) += panel-tpo-td043mtea1.o
+obj-$(CONFIG_DISPLAY_PANEL_NEC_NL8048HL11) += panel-nec-nl8048hl11.o
diff --git a/drivers/video/omap2/displays-new/connector-analog-tv.c b/drivers/video/omap2/displays-new/connector-analog-tv.c
new file mode 100644
index 0000000..5338f36
--- /dev/null
+++ b/drivers/video/omap2/displays-new/connector-analog-tv.c
@@ -0,0 +1,265 @@
+/*
+ * Analog TV Connector driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	struct device *dev;
+
+	struct omap_video_timings timings;
+
+	enum omap_dss_venc_type connector_type;
+	bool invert_polarity;
+};
+
+#define to_panel_data(x) container_of(x, struct panel_drv_data, dssdev)
+
+static int tvc_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(ddata->dev, "connect\n");
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.atv->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void tvc_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(ddata->dev, "disconnect\n");
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.atv->disconnect(in, dssdev);
+}
+
+static int tvc_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(ddata->dev, "enable\n");
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.atv->set_timings(in, &ddata->timings);
+
+	in->ops.atv->set_type(in, ddata->connector_type);
+	in->ops.atv->invert_vid_out_polarity(in, ddata->invert_polarity);
+
+	r = in->ops.atv->enable(in);
+	if (r)
+		return r;
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return r;
+}
+
+static void tvc_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(ddata->dev, "disable\n");
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	in->ops.atv->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void tvc_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->timings = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.atv->set_timings(in, timings);
+}
+
+static void tvc_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->timings;
+}
+
+static int tvc_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.atv->check_timings(in, timings);
+}
+
+static u32 tvc_get_wss(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.atv->get_wss(in);
+}
+
+static int tvc_set_wss(struct omap_dss_device *dssdev, u32 wss)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.atv->set_wss(in, wss);
+}
+
+static struct omap_dss_driver tvc_driver = {
+	.connect		= tvc_connect,
+	.disconnect		= tvc_disconnect,
+
+	.enable			= tvc_enable,
+	.disable		= tvc_disable,
+
+	.set_timings		= tvc_set_timings,
+	.get_timings		= tvc_get_timings,
+	.check_timings		= tvc_check_timings,
+
+	.get_resolution		= omapdss_default_get_resolution,
+
+	.get_wss		= tvc_get_wss,
+	.set_wss		= tvc_set_wss,
+};
+
+static int tvc_probe_pdata(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct connector_atv_platform_data *pdata;
+	struct omap_dss_device *in, *dssdev;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "Failed to find video source\n");
+		return -ENODEV;
+	}
+
+	ddata->in = in;
+
+	ddata->connector_type = pdata->connector_type;
+	ddata->invert_polarity = ddata->invert_polarity;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int tvc_probe(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (!ddata)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+	ddata->dev = &pdev->dev;
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = tvc_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	ddata->timings = omap_dss_pal_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->driver = &tvc_driver;
+	dssdev->dev = &pdev->dev;
+	dssdev->type = OMAP_DISPLAY_TYPE_VENC;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = omap_dss_pal_timings;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+err_reg:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit tvc_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_display(&ddata->dssdev);
+
+	tvc_disable(dssdev);
+	tvc_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct platform_driver tvc_connector_driver = {
+	.probe	= tvc_probe,
+	.remove	= __exit_p(tvc_remove),
+	.driver	= {
+		.name	= "connector-analog-tv",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_platform_driver(tvc_connector_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("Analog TV Connector driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/connector-dvi.c b/drivers/video/omap2/displays-new/connector-dvi.c
new file mode 100644
index 0000000..bc5f8ce
--- /dev/null
+++ b/drivers/video/omap2/displays-new/connector-dvi.c
@@ -0,0 +1,351 @@
+/*
+ * Generic DVI Connector driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/i2c.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include <drm/drm_edid.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+static const struct omap_video_timings dvic_default_timings = {
+	.x_res		= 640,
+	.y_res		= 480,
+
+	.pixel_clock	= 23500,
+
+	.hfp		= 48,
+	.hsw		= 32,
+	.hbp		= 80,
+
+	.vfp		= 3,
+	.vsw		= 4,
+	.vbp		= 7,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.data_pclk_edge	= OMAPDSS_DRIVE_SIG_RISING_EDGE,
+	.de_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.sync_pclk_edge	= OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES,
+};
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	struct omap_video_timings timings;
+
+	struct i2c_adapter *i2c_adapter;
+};
+
+#define to_panel_data(x) container_of(x, struct panel_drv_data, dssdev)
+
+static int dvic_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dvi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void dvic_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dvi->disconnect(in, dssdev);
+}
+
+static int dvic_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dvi->set_timings(in, &ddata->timings);
+
+	r = in->ops.dvi->enable(in);
+	if (r)
+		return r;
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void dvic_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	in->ops.dvi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void dvic_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->timings = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dvi->set_timings(in, timings);
+}
+
+static void dvic_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->timings;
+}
+
+static int dvic_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dvi->check_timings(in, timings);
+}
+
+static int dvic_ddc_read(struct i2c_adapter *adapter,
+		unsigned char *buf, u16 count, u8 offset)
+{
+	int r, retries;
+
+	for (retries = 3; retries > 0; retries--) {
+		struct i2c_msg msgs[] = {
+			{
+				.addr   = DDC_ADDR,
+				.flags  = 0,
+				.len    = 1,
+				.buf    = &offset,
+			}, {
+				.addr   = DDC_ADDR,
+				.flags  = I2C_M_RD,
+				.len    = count,
+				.buf    = buf,
+			}
+		};
+
+		r = i2c_transfer(adapter, msgs, 2);
+		if (r == 2)
+			return 0;
+
+		if (r != -EAGAIN)
+			break;
+	}
+
+	return r < 0 ? r : -EIO;
+}
+
+static int dvic_read_edid(struct omap_dss_device *dssdev,
+		u8 *edid, int len)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	int r, l, bytes_read;
+
+	if (!ddata->i2c_adapter)
+		return -ENODEV;
+
+	l = min(EDID_LENGTH, len);
+	r = dvic_ddc_read(ddata->i2c_adapter, edid, l, 0);
+	if (r)
+		return r;
+
+	bytes_read = l;
+
+	/* if there are extensions, read second block */
+	if (len > EDID_LENGTH && edid[0x7e] > 0) {
+		l = min(EDID_LENGTH, len - EDID_LENGTH);
+
+		r = dvic_ddc_read(ddata->i2c_adapter, edid + EDID_LENGTH,
+				l, EDID_LENGTH);
+		if (r)
+			return r;
+
+		bytes_read += l;
+	}
+
+	return bytes_read;
+}
+
+static bool dvic_detect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	unsigned char out;
+	int r;
+
+	if (!ddata->i2c_adapter)
+		return true;
+
+	r = dvic_ddc_read(ddata->i2c_adapter, &out, 1, 0);
+
+	return r == 0;
+}
+
+static struct omap_dss_driver dvic_driver = {
+	.connect	= dvic_connect,
+	.disconnect	= dvic_disconnect,
+
+	.enable		= dvic_enable,
+	.disable	= dvic_disable,
+
+	.set_timings	= dvic_set_timings,
+	.get_timings	= dvic_get_timings,
+	.check_timings	= dvic_check_timings,
+
+	.get_resolution	= omapdss_default_get_resolution,
+
+	.read_edid	= dvic_read_edid,
+	.detect		= dvic_detect,
+};
+
+static int dvic_probe_pdata(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct connector_dvi_platform_data *pdata;
+	struct omap_dss_device *in, *dssdev;
+	int i2c_bus_num;
+
+	pdata = dev_get_platdata(&pdev->dev);
+	i2c_bus_num = pdata->i2c_bus_num;
+
+	if (i2c_bus_num != -1) {
+		struct i2c_adapter *adapter;
+
+		adapter = i2c_get_adapter(i2c_bus_num);
+		if (!adapter) {
+			dev_err(&pdev->dev,
+					"Failed to get I2C adapter, bus %d\n",
+					i2c_bus_num);
+			return -EPROBE_DEFER;
+		}
+
+		ddata->i2c_adapter = adapter;
+	}
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "Failed to find video source\n");
+		return -ENODEV;
+	}
+
+	ddata->in = in;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int dvic_probe(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (!ddata)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = dvic_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	ddata->timings = dvic_default_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->driver = &dvic_driver;
+	dssdev->dev = &pdev->dev;
+	dssdev->type = OMAP_DISPLAY_TYPE_DVI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = dvic_default_timings;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit dvic_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_display(&ddata->dssdev);
+
+	dvic_disable(dssdev);
+	dvic_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	if (ddata->i2c_adapter)
+		i2c_put_adapter(ddata->i2c_adapter);
+
+	return 0;
+}
+
+static struct platform_driver dvi_connector_driver = {
+	.probe	= dvic_probe,
+	.remove	= __exit_p(dvic_remove),
+	.driver	= {
+		.name	= "connector-dvi",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_platform_driver(dvi_connector_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("Generic DVI Connector driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/connector-hdmi.c b/drivers/video/omap2/displays-new/connector-hdmi.c
new file mode 100644
index 0000000..c582671
--- /dev/null
+++ b/drivers/video/omap2/displays-new/connector-hdmi.c
@@ -0,0 +1,375 @@
+/*
+ * HDMI Connector driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+
+#include <drm/drm_edid.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+static const struct omap_video_timings hdmic_default_timings = {
+	.x_res		= 640,
+	.y_res		= 480,
+	.pixel_clock	= 25175,
+	.hsw		= 96,
+	.hfp		= 16,
+	.hbp		= 48,
+	.vsw		= 2,
+	.vfp		= 11,
+	.vbp		= 31,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+
+	.interlace	= false,
+};
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	struct device *dev;
+
+	struct omap_video_timings timings;
+};
+
+#define to_panel_data(x) container_of(x, struct panel_drv_data, dssdev)
+
+static int hdmic_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(ddata->dev, "connect\n");
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.hdmi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void hdmic_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(ddata->dev, "disconnect\n");
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.hdmi->disconnect(in, dssdev);
+}
+
+static int hdmic_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(ddata->dev, "enable\n");
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.hdmi->set_timings(in, &ddata->timings);
+
+	r = in->ops.hdmi->enable(in);
+	if (r)
+		return r;
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return r;
+}
+
+static void hdmic_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(ddata->dev, "disable\n");
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	in->ops.hdmi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void hdmic_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->timings = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.hdmi->set_timings(in, timings);
+}
+
+static void hdmic_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->timings;
+}
+
+static int hdmic_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->check_timings(in, timings);
+}
+
+static int hdmic_read_edid(struct omap_dss_device *dssdev,
+		u8 *edid, int len)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->read_edid(in, edid, len);
+}
+
+static bool hdmic_detect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->detect(in);
+}
+
+static int hdmic_audio_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	/* enable audio only if the display is active */
+	if (!omapdss_device_is_enabled(dssdev))
+		return -EPERM;
+
+	r = in->ops.hdmi->audio_enable(in);
+	if (r)
+		return r;
+
+	dssdev->audio_state = OMAP_DSS_AUDIO_ENABLED;
+
+	return 0;
+}
+
+static void hdmic_audio_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	in->ops.hdmi->audio_disable(in);
+
+	dssdev->audio_state = OMAP_DSS_AUDIO_DISABLED;
+}
+
+static int hdmic_audio_start(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	/*
+	 * No need to check the panel state. It was checked when trasitioning
+	 * to AUDIO_ENABLED.
+	 */
+	if (dssdev->audio_state != OMAP_DSS_AUDIO_ENABLED)
+		return -EPERM;
+
+	r = in->ops.hdmi->audio_start(in);
+	if (r)
+		return r;
+
+	dssdev->audio_state = OMAP_DSS_AUDIO_PLAYING;
+
+	return 0;
+}
+
+static void hdmic_audio_stop(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	in->ops.hdmi->audio_stop(in);
+
+	dssdev->audio_state = OMAP_DSS_AUDIO_ENABLED;
+}
+
+static bool hdmic_audio_supported(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return false;
+
+	return in->ops.hdmi->audio_supported(in);
+}
+
+static int hdmic_audio_config(struct omap_dss_device *dssdev,
+		struct omap_dss_audio *audio)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	/* config audio only if the display is active */
+	if (!omapdss_device_is_enabled(dssdev))
+		return -EPERM;
+
+	r = in->ops.hdmi->audio_config(in, audio);
+	if (r)
+		return r;
+
+	dssdev->audio_state = OMAP_DSS_AUDIO_CONFIGURED;
+
+	return 0;
+}
+
+static struct omap_dss_driver hdmic_driver = {
+	.connect		= hdmic_connect,
+	.disconnect		= hdmic_disconnect,
+
+	.enable			= hdmic_enable,
+	.disable		= hdmic_disable,
+
+	.set_timings		= hdmic_set_timings,
+	.get_timings		= hdmic_get_timings,
+	.check_timings		= hdmic_check_timings,
+
+	.get_resolution		= omapdss_default_get_resolution,
+
+	.read_edid		= hdmic_read_edid,
+	.detect			= hdmic_detect,
+
+	.audio_enable		= hdmic_audio_enable,
+	.audio_disable		= hdmic_audio_disable,
+	.audio_start		= hdmic_audio_start,
+	.audio_stop		= hdmic_audio_stop,
+	.audio_supported	= hdmic_audio_supported,
+	.audio_config		= hdmic_audio_config,
+};
+
+static int hdmic_probe_pdata(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct connector_hdmi_platform_data *pdata;
+	struct omap_dss_device *in, *dssdev;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "Failed to find video source\n");
+		return -ENODEV;
+	}
+
+	ddata->in = in;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int hdmic_probe(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (!ddata)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+	ddata->dev = &pdev->dev;
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = hdmic_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	ddata->timings = hdmic_default_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->driver = &hdmic_driver;
+	dssdev->dev = &pdev->dev;
+	dssdev->type = OMAP_DISPLAY_TYPE_HDMI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = hdmic_default_timings;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+err_reg:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit hdmic_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_display(&ddata->dssdev);
+
+	hdmic_disable(dssdev);
+	hdmic_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct platform_driver hdmi_connector_driver = {
+	.probe	= hdmic_probe,
+	.remove	= __exit_p(hdmic_remove),
+	.driver	= {
+		.name	= "connector-hdmi",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_platform_driver(hdmi_connector_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("HDMI Connector driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/encoder-tfp410.c b/drivers/video/omap2/displays-new/encoder-tfp410.c
new file mode 100644
index 0000000..a04f658
--- /dev/null
+++ b/drivers/video/omap2/displays-new/encoder-tfp410.c
@@ -0,0 +1,267 @@
+/*
+ * TFP410 DPI-to-DVI encoder driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/gpio.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	int pd_gpio;
+	int data_lines;
+
+	struct omap_video_timings timings;
+};
+
+#define to_panel_data(x) container_of(x, struct panel_drv_data, dssdev)
+
+static int tfp410_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return -EBUSY;
+
+	r = in->ops.dpi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	dst->output = dssdev;
+	dssdev->device = dst;
+
+	return 0;
+}
+
+static void tfp410_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	WARN_ON(!omapdss_device_is_connected(dssdev));
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	WARN_ON(dst != dssdev->device);
+	if (dst != dssdev->device)
+		return;
+
+	dst->output = NULL;
+	dssdev->device = NULL;
+
+	in->ops.dpi->disconnect(in, &ddata->dssdev);
+}
+
+static int tfp410_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dpi->set_timings(in, &ddata->timings);
+	in->ops.dpi->set_data_lines(in, ddata->data_lines);
+
+	r = in->ops.dpi->enable(in);
+	if (r)
+		return r;
+
+	if (gpio_is_valid(ddata->pd_gpio))
+		gpio_set_value_cansleep(ddata->pd_gpio, 1);
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void tfp410_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	if (gpio_is_valid(ddata->pd_gpio))
+		gpio_set_value_cansleep(ddata->pd_gpio, 0);
+
+	in->ops.dpi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void tfp410_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->timings = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dpi->set_timings(in, timings);
+}
+
+static void tfp410_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->timings;
+}
+
+static int tfp410_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dpi->check_timings(in, timings);
+}
+
+static const struct omapdss_dvi_ops tfp410_dvi_ops = {
+	.connect	= tfp410_connect,
+	.disconnect	= tfp410_disconnect,
+
+	.enable		= tfp410_enable,
+	.disable	= tfp410_disable,
+
+	.check_timings	= tfp410_check_timings,
+	.set_timings	= tfp410_set_timings,
+	.get_timings	= tfp410_get_timings,
+};
+
+static int tfp410_probe_pdata(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct encoder_tfp410_platform_data *pdata;
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	ddata->pd_gpio = pdata->power_down_gpio;
+
+	ddata->data_lines = pdata->data_lines;
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "Failed to find video source\n");
+		return -ENODEV;
+	}
+
+	ddata->in = in;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int tfp410_probe(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (!ddata)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = tfp410_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	if (gpio_is_valid(ddata->pd_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->pd_gpio,
+				GPIOF_OUT_INIT_LOW, "tfp410 PD");
+		if (r) {
+			dev_err(&pdev->dev, "Failed to request PD GPIO %d\n",
+					ddata->pd_gpio);
+			goto err_gpio;
+		}
+	}
+
+	dssdev = &ddata->dssdev;
+	dssdev->ops.dvi = &tfp410_dvi_ops;
+	dssdev->dev = &pdev->dev;
+	dssdev->type = OMAP_DISPLAY_TYPE_DPI;
+	dssdev->output_type = OMAP_DISPLAY_TYPE_DVI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->phy.dpi.data_lines = ddata->data_lines;
+
+	r = omapdss_register_output(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register output\n");
+		goto err_reg;
+	}
+
+	return 0;
+err_reg:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit tfp410_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_output(&ddata->dssdev);
+
+	WARN_ON(omapdss_device_is_enabled(dssdev));
+	if (omapdss_device_is_enabled(dssdev))
+		tfp410_disable(dssdev);
+
+	WARN_ON(omapdss_device_is_connected(dssdev));
+	if (omapdss_device_is_connected(dssdev))
+		tfp410_disconnect(dssdev, dssdev->device);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct platform_driver tfp410_driver = {
+	.probe	= tfp410_probe,
+	.remove	= __exit_p(tfp410_remove),
+	.driver	= {
+		.name	= "tfp410",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_platform_driver(tfp410_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("TFP410 DPI to DVI encoder driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/encoder-tpd12s015.c b/drivers/video/omap2/displays-new/encoder-tpd12s015.c
new file mode 100644
index 0000000..ce0e010
--- /dev/null
+++ b/drivers/video/omap2/displays-new/encoder-tpd12s015.c
@@ -0,0 +1,395 @@
+/*
+ * TPD12S015 HDMI ESD protection & level shifter chip driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/completion.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/gpio.h>
+#include <linux/platform_device.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	int ct_cp_hpd_gpio;
+	int ls_oe_gpio;
+	int hpd_gpio;
+
+	struct omap_video_timings timings;
+
+	struct completion hpd_completion;
+};
+
+#define to_panel_data(x) container_of(x, struct panel_drv_data, dssdev)
+
+static irqreturn_t tpd_hpd_irq_handler(int irq, void *data)
+{
+	struct panel_drv_data *ddata = data;
+	bool hpd;
+
+	hpd = gpio_get_value_cansleep(ddata->hpd_gpio);
+
+	dev_dbg(ddata->dssdev.dev, "hpd %d\n", hpd);
+
+	if (gpio_is_valid(ddata->ls_oe_gpio)) {
+		if (hpd)
+			gpio_set_value_cansleep(ddata->ls_oe_gpio, 1);
+		else
+			gpio_set_value_cansleep(ddata->ls_oe_gpio, 0);
+	}
+
+	complete_all(&ddata->hpd_completion);
+
+	return IRQ_HANDLED;
+}
+
+static int tpd_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	r = in->ops.hdmi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	dst->output = dssdev;
+	dssdev->device = dst;
+
+	INIT_COMPLETION(ddata->hpd_completion);
+
+	gpio_set_value_cansleep(ddata->ct_cp_hpd_gpio, 1);
+	/* DC-DC converter needs at max 300us to get to 90% of 5V */
+	udelay(300);
+
+	/*
+	 * If there's a cable connected, wait for the hpd irq to trigger,
+	 * which turns on the level shifters.
+	 */
+	if (gpio_get_value_cansleep(ddata->hpd_gpio)) {
+		unsigned long to;
+		to = wait_for_completion_timeout(&ddata->hpd_completion,
+				msecs_to_jiffies(250));
+		WARN_ON_ONCE(to == 0);
+	}
+
+	return 0;
+}
+
+static void tpd_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	WARN_ON(dst != dssdev->device);
+
+	if (dst != dssdev->device)
+		return;
+
+	gpio_set_value_cansleep(ddata->ct_cp_hpd_gpio, 0);
+
+	dst->output = NULL;
+	dssdev->device = NULL;
+
+	in->ops.hdmi->disconnect(in, &ddata->dssdev);
+}
+
+static int tpd_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE)
+		return 0;
+
+	in->ops.hdmi->set_timings(in, &ddata->timings);
+
+	r = in->ops.hdmi->enable(in);
+	if (r)
+		return r;
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return r;
+}
+
+static void tpd_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (dssdev->state != OMAP_DSS_DISPLAY_ACTIVE)
+		return;
+
+	in->ops.hdmi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void tpd_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->timings = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.hdmi->set_timings(in, timings);
+}
+
+static void tpd_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->timings;
+}
+
+static int tpd_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	r = in->ops.hdmi->check_timings(in, timings);
+
+	return r;
+}
+
+static int tpd_read_edid(struct omap_dss_device *dssdev,
+		u8 *edid, int len)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!gpio_get_value_cansleep(ddata->hpd_gpio))
+		return -ENODEV;
+
+	return in->ops.hdmi->read_edid(in, edid, len);
+}
+
+static bool tpd_detect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	return gpio_get_value_cansleep(ddata->hpd_gpio);
+}
+
+static int tpd_audio_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->audio_enable(in);
+}
+
+static void tpd_audio_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	in->ops.hdmi->audio_disable(in);
+}
+
+static int tpd_audio_start(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->audio_start(in);
+}
+
+static void tpd_audio_stop(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	in->ops.hdmi->audio_stop(in);
+}
+
+static bool tpd_audio_supported(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->audio_supported(in);
+}
+
+static int tpd_audio_config(struct omap_dss_device *dssdev,
+		struct omap_dss_audio *audio)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.hdmi->audio_config(in, audio);
+}
+
+static const struct omapdss_hdmi_ops tpd_hdmi_ops = {
+	.connect		= tpd_connect,
+	.disconnect		= tpd_disconnect,
+
+	.enable			= tpd_enable,
+	.disable		= tpd_disable,
+
+	.check_timings		= tpd_check_timings,
+	.set_timings		= tpd_set_timings,
+	.get_timings		= tpd_get_timings,
+
+	.read_edid		= tpd_read_edid,
+	.detect			= tpd_detect,
+
+	.audio_enable		= tpd_audio_enable,
+	.audio_disable		= tpd_audio_disable,
+	.audio_start		= tpd_audio_start,
+	.audio_stop		= tpd_audio_stop,
+	.audio_supported	= tpd_audio_supported,
+	.audio_config		= tpd_audio_config,
+};
+
+static int tpd_probe_pdata(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct encoder_tpd12s015_platform_data *pdata;
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	ddata->ct_cp_hpd_gpio = pdata->ct_cp_hpd_gpio;
+	ddata->ls_oe_gpio = pdata->ls_oe_gpio;
+	ddata->hpd_gpio = pdata->hpd_gpio;
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "Failed to find video source\n");
+		return -ENODEV;
+	}
+
+	ddata->in = in;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int tpd_probe(struct platform_device *pdev)
+{
+	struct omap_dss_device *in, *dssdev;
+	struct panel_drv_data *ddata;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (!ddata)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+
+	init_completion(&ddata->hpd_completion);
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = tpd_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	r = devm_gpio_request_one(&pdev->dev, ddata->ct_cp_hpd_gpio,
+			GPIOF_OUT_INIT_LOW, "hdmi_ct_cp_hpd");
+	if (r)
+		goto err_gpio;
+
+	if (gpio_is_valid(ddata->ls_oe_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->ls_oe_gpio,
+				GPIOF_OUT_INIT_LOW, "hdmi_ls_oe");
+		if (r)
+			goto err_gpio;
+	}
+
+	r = devm_gpio_request_one(&pdev->dev, ddata->hpd_gpio,
+			GPIOF_DIR_IN, "hdmi_hpd");
+	if (r)
+		goto err_gpio;
+
+	r = devm_request_threaded_irq(&pdev->dev, gpio_to_irq(ddata->hpd_gpio),
+				 NULL, tpd_hpd_irq_handler,
+				 IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING |
+				 IRQF_ONESHOT, "hpd", ddata);
+	if (r)
+		goto err_irq;
+
+	dssdev = &ddata->dssdev;
+	dssdev->ops.hdmi = &tpd_hdmi_ops;
+	dssdev->dev = &pdev->dev;
+	dssdev->type = OMAP_DISPLAY_TYPE_HDMI;
+	dssdev->output_type = OMAP_DISPLAY_TYPE_HDMI;
+	dssdev->owner = THIS_MODULE;
+
+	in = ddata->in;
+
+	r = omapdss_register_output(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register output\n");
+		goto err_reg;
+	}
+
+	return 0;
+err_reg:
+err_irq:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit tpd_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_output(&ddata->dssdev);
+
+	WARN_ON(omapdss_device_is_enabled(dssdev));
+	if (omapdss_device_is_enabled(dssdev))
+		tpd_disable(dssdev);
+
+	WARN_ON(omapdss_device_is_connected(dssdev));
+	if (omapdss_device_is_connected(dssdev))
+		tpd_disconnect(dssdev, dssdev->device);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct platform_driver tpd_driver = {
+	.probe	= tpd_probe,
+	.remove	= __exit_p(tpd_remove),
+	.driver	= {
+		.name	= "tpd12s015",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_platform_driver(tpd_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("TPD12S015 driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-dpi.c b/drivers/video/omap2/displays-new/panel-dpi.c
new file mode 100644
index 0000000..5f8f7e7
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-dpi.c
@@ -0,0 +1,270 @@
+/*
+ * Generic MIPI DPI Panel Driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/gpio.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	int data_lines;
+
+	struct omap_video_timings videomode;
+
+	int backlight_gpio;
+	int enable_gpio;
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static int panel_dpi_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dpi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void panel_dpi_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dpi->disconnect(in, dssdev);
+}
+
+static int panel_dpi_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dpi->set_data_lines(in, ddata->data_lines);
+	in->ops.dpi->set_timings(in, &ddata->videomode);
+
+	r = in->ops.dpi->enable(in);
+	if (r)
+		return r;
+
+	if (gpio_is_valid(ddata->enable_gpio))
+		gpio_set_value_cansleep(ddata->enable_gpio, 1);
+
+	if (gpio_is_valid(ddata->backlight_gpio))
+		gpio_set_value_cansleep(ddata->backlight_gpio, 1);
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void panel_dpi_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	if (gpio_is_valid(ddata->enable_gpio))
+		gpio_set_value_cansleep(ddata->enable_gpio, 0);
+
+	if (gpio_is_valid(ddata->backlight_gpio))
+		gpio_set_value_cansleep(ddata->backlight_gpio, 0);
+
+	in->ops.dpi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void panel_dpi_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->videomode = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dpi->set_timings(in, timings);
+}
+
+static void panel_dpi_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->videomode;
+}
+
+static int panel_dpi_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dpi->check_timings(in, timings);
+}
+
+static struct omap_dss_driver panel_dpi_ops = {
+	.connect	= panel_dpi_connect,
+	.disconnect	= panel_dpi_disconnect,
+
+	.enable		= panel_dpi_enable,
+	.disable	= panel_dpi_disable,
+
+	.set_timings	= panel_dpi_set_timings,
+	.get_timings	= panel_dpi_get_timings,
+	.check_timings	= panel_dpi_check_timings,
+
+	.get_resolution	= omapdss_default_get_resolution,
+};
+
+static int panel_dpi_probe_pdata(struct platform_device *pdev)
+{
+	const struct panel_dpi_platform_data *pdata;
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev, *in;
+	struct videomode vm;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "failed to find video source '%s'\n",
+				pdata->source);
+		return -EPROBE_DEFER;
+	}
+
+	ddata->in = in;
+
+	ddata->data_lines = pdata->data_lines;
+
+	videomode_from_timing(pdata->display_timing, &vm);
+	videomode_to_omap_video_timings(&vm, &ddata->videomode);
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	ddata->enable_gpio = pdata->enable_gpio;
+	ddata->backlight_gpio = pdata->backlight_gpio;
+
+	return 0;
+}
+
+static int panel_dpi_probe(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (ddata == NULL)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = panel_dpi_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	if (gpio_is_valid(ddata->enable_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->enable_gpio,
+				GPIOF_OUT_INIT_LOW, "panel enable");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->backlight_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->backlight_gpio,
+				GPIOF_OUT_INIT_LOW, "panel backlight");
+		if (r)
+			goto err_gpio;
+	}
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = &pdev->dev;
+	dssdev->driver = &panel_dpi_ops;
+	dssdev->type = OMAP_DISPLAY_TYPE_DPI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = ddata->videomode;
+	dssdev->phy.dpi.data_lines = ddata->data_lines;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit panel_dpi_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_display(dssdev);
+
+	panel_dpi_disable(dssdev);
+	panel_dpi_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct platform_driver panel_dpi_driver = {
+	.probe = panel_dpi_probe,
+	.remove = __exit_p(panel_dpi_remove),
+	.driver = {
+		.name = "panel-dpi",
+		.owner = THIS_MODULE,
+	},
+};
+
+module_platform_driver(panel_dpi_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("Generic MIPI DPI Panel Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-dsi-cm.c b/drivers/video/omap2/displays-new/panel-dsi-cm.c
new file mode 100644
index 0000000..aaaea64
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-dsi-cm.c
@@ -0,0 +1,1336 @@
+/*
+ * Generic DSI Command Mode panel driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+/* #define DEBUG */
+
+#include <linux/backlight.h>
+#include <linux/delay.h>
+#include <linux/fb.h>
+#include <linux/gpio.h>
+#include <linux/interrupt.h>
+#include <linux/jiffies.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+#include <video/mipi_display.h>
+
+/* DSI Virtual channel. Hardcoded for now. */
+#define TCH 0
+
+#define DCS_READ_NUM_ERRORS	0x05
+#define DCS_BRIGHTNESS		0x51
+#define DCS_CTRL_DISPLAY	0x53
+#define DCS_GET_ID1		0xda
+#define DCS_GET_ID2		0xdb
+#define DCS_GET_ID3		0xdc
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	struct omap_video_timings timings;
+
+	struct platform_device *pdev;
+
+	struct mutex lock;
+
+	struct backlight_device *bldev;
+
+	unsigned long	hw_guard_end;	/* next value of jiffies when we can
+					 * issue the next sleep in/out command
+					 */
+	unsigned long	hw_guard_wait;	/* max guard time in jiffies */
+
+	/* panel HW configuration from DT or platform data */
+	int reset_gpio;
+	int ext_te_gpio;
+
+	bool use_dsi_backlight;
+
+	struct omap_dsi_pin_config pin_config;
+
+	/* runtime variables */
+	bool enabled;
+
+	bool te_enabled;
+
+	atomic_t do_update;
+	int channel;
+
+	struct delayed_work te_timeout_work;
+
+	bool intro_printed;
+
+	struct workqueue_struct *workqueue;
+
+	bool ulps_enabled;
+	unsigned ulps_timeout;
+	struct delayed_work ulps_work;
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static irqreturn_t dsicm_te_isr(int irq, void *data);
+static void dsicm_te_timeout_work_callback(struct work_struct *work);
+static int _dsicm_enable_te(struct panel_drv_data *ddata, bool enable);
+
+static int dsicm_panel_reset(struct panel_drv_data *ddata);
+
+static void dsicm_ulps_work(struct work_struct *work);
+
+static void hw_guard_start(struct panel_drv_data *ddata, int guard_msec)
+{
+	ddata->hw_guard_wait = msecs_to_jiffies(guard_msec);
+	ddata->hw_guard_end = jiffies + ddata->hw_guard_wait;
+}
+
+static void hw_guard_wait(struct panel_drv_data *ddata)
+{
+	unsigned long wait = ddata->hw_guard_end - jiffies;
+
+	if ((long)wait > 0 && wait <= ddata->hw_guard_wait) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		schedule_timeout(wait);
+	}
+}
+
+static int dsicm_dcs_read_1(struct panel_drv_data *ddata, u8 dcs_cmd, u8 *data)
+{
+	struct omap_dss_device *in = ddata->in;
+	int r;
+	u8 buf[1];
+
+	r = in->ops.dsi->dcs_read(in, ddata->channel, dcs_cmd, buf, 1);
+
+	if (r < 0)
+		return r;
+
+	*data = buf[0];
+
+	return 0;
+}
+
+static int dsicm_dcs_write_0(struct panel_drv_data *ddata, u8 dcs_cmd)
+{
+	struct omap_dss_device *in = ddata->in;
+	return in->ops.dsi->dcs_write(in, ddata->channel, &dcs_cmd, 1);
+}
+
+static int dsicm_dcs_write_1(struct panel_drv_data *ddata, u8 dcs_cmd, u8 param)
+{
+	struct omap_dss_device *in = ddata->in;
+	u8 buf[2] = { dcs_cmd, param };
+
+	return in->ops.dsi->dcs_write(in, ddata->channel, buf, 2);
+}
+
+static int dsicm_sleep_in(struct panel_drv_data *ddata)
+
+{
+	struct omap_dss_device *in = ddata->in;
+	u8 cmd;
+	int r;
+
+	hw_guard_wait(ddata);
+
+	cmd = MIPI_DCS_ENTER_SLEEP_MODE;
+	r = in->ops.dsi->dcs_write_nosync(in, ddata->channel, &cmd, 1);
+	if (r)
+		return r;
+
+	hw_guard_start(ddata, 120);
+
+	usleep_range(5000, 10000);
+
+	return 0;
+}
+
+static int dsicm_sleep_out(struct panel_drv_data *ddata)
+{
+	int r;
+
+	hw_guard_wait(ddata);
+
+	r = dsicm_dcs_write_0(ddata, MIPI_DCS_EXIT_SLEEP_MODE);
+	if (r)
+		return r;
+
+	hw_guard_start(ddata, 120);
+
+	usleep_range(5000, 10000);
+
+	return 0;
+}
+
+static int dsicm_get_id(struct panel_drv_data *ddata, u8 *id1, u8 *id2, u8 *id3)
+{
+	int r;
+
+	r = dsicm_dcs_read_1(ddata, DCS_GET_ID1, id1);
+	if (r)
+		return r;
+	r = dsicm_dcs_read_1(ddata, DCS_GET_ID2, id2);
+	if (r)
+		return r;
+	r = dsicm_dcs_read_1(ddata, DCS_GET_ID3, id3);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static int dsicm_set_update_window(struct panel_drv_data *ddata,
+		u16 x, u16 y, u16 w, u16 h)
+{
+	struct omap_dss_device *in = ddata->in;
+	int r;
+	u16 x1 = x;
+	u16 x2 = x + w - 1;
+	u16 y1 = y;
+	u16 y2 = y + h - 1;
+
+	u8 buf[5];
+	buf[0] = MIPI_DCS_SET_COLUMN_ADDRESS;
+	buf[1] = (x1 >> 8) & 0xff;
+	buf[2] = (x1 >> 0) & 0xff;
+	buf[3] = (x2 >> 8) & 0xff;
+	buf[4] = (x2 >> 0) & 0xff;
+
+	r = in->ops.dsi->dcs_write_nosync(in, ddata->channel, buf, sizeof(buf));
+	if (r)
+		return r;
+
+	buf[0] = MIPI_DCS_SET_PAGE_ADDRESS;
+	buf[1] = (y1 >> 8) & 0xff;
+	buf[2] = (y1 >> 0) & 0xff;
+	buf[3] = (y2 >> 8) & 0xff;
+	buf[4] = (y2 >> 0) & 0xff;
+
+	r = in->ops.dsi->dcs_write_nosync(in, ddata->channel, buf, sizeof(buf));
+	if (r)
+		return r;
+
+	in->ops.dsi->bta_sync(in, ddata->channel);
+
+	return r;
+}
+
+static void dsicm_queue_ulps_work(struct panel_drv_data *ddata)
+{
+	if (ddata->ulps_timeout > 0)
+		queue_delayed_work(ddata->workqueue, &ddata->ulps_work,
+				msecs_to_jiffies(ddata->ulps_timeout));
+}
+
+static void dsicm_cancel_ulps_work(struct panel_drv_data *ddata)
+{
+	cancel_delayed_work(&ddata->ulps_work);
+}
+
+static int dsicm_enter_ulps(struct panel_drv_data *ddata)
+{
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (ddata->ulps_enabled)
+		return 0;
+
+	dsicm_cancel_ulps_work(ddata);
+
+	r = _dsicm_enable_te(ddata, false);
+	if (r)
+		goto err;
+
+	if (gpio_is_valid(ddata->ext_te_gpio))
+		disable_irq(gpio_to_irq(ddata->ext_te_gpio));
+
+	in->ops.dsi->disable(in, false, true);
+
+	ddata->ulps_enabled = true;
+
+	return 0;
+
+err:
+	dev_err(&ddata->pdev->dev, "enter ULPS failed");
+	dsicm_panel_reset(ddata);
+
+	ddata->ulps_enabled = false;
+
+	dsicm_queue_ulps_work(ddata);
+
+	return r;
+}
+
+static int dsicm_exit_ulps(struct panel_drv_data *ddata)
+{
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!ddata->ulps_enabled)
+		return 0;
+
+	r = in->ops.dsi->enable(in);
+	if (r) {
+		dev_err(&ddata->pdev->dev, "failed to enable DSI\n");
+		goto err1;
+	}
+
+	in->ops.dsi->enable_hs(in, ddata->channel, true);
+
+	r = _dsicm_enable_te(ddata, true);
+	if (r) {
+		dev_err(&ddata->pdev->dev, "failed to re-enable TE");
+		goto err2;
+	}
+
+	if (gpio_is_valid(ddata->ext_te_gpio))
+		enable_irq(gpio_to_irq(ddata->ext_te_gpio));
+
+	dsicm_queue_ulps_work(ddata);
+
+	ddata->ulps_enabled = false;
+
+	return 0;
+
+err2:
+	dev_err(&ddata->pdev->dev, "failed to exit ULPS");
+
+	r = dsicm_panel_reset(ddata);
+	if (!r) {
+		if (gpio_is_valid(ddata->ext_te_gpio))
+			enable_irq(gpio_to_irq(ddata->ext_te_gpio));
+		ddata->ulps_enabled = false;
+	}
+err1:
+	dsicm_queue_ulps_work(ddata);
+
+	return r;
+}
+
+static int dsicm_wake_up(struct panel_drv_data *ddata)
+{
+	if (ddata->ulps_enabled)
+		return dsicm_exit_ulps(ddata);
+
+	dsicm_cancel_ulps_work(ddata);
+	dsicm_queue_ulps_work(ddata);
+	return 0;
+}
+
+static int dsicm_bl_update_status(struct backlight_device *dev)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&dev->dev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+	int level;
+
+	if (dev->props.fb_blank == FB_BLANK_UNBLANK &&
+			dev->props.power == FB_BLANK_UNBLANK)
+		level = dev->props.brightness;
+	else
+		level = 0;
+
+	dev_dbg(&ddata->pdev->dev, "update brightness to %d\n", level);
+
+	mutex_lock(&ddata->lock);
+
+	if (ddata->enabled) {
+		in->ops.dsi->bus_lock(in);
+
+		r = dsicm_wake_up(ddata);
+		if (!r)
+			r = dsicm_dcs_write_1(ddata, DCS_BRIGHTNESS, level);
+
+		in->ops.dsi->bus_unlock(in);
+	} else {
+		r = 0;
+	}
+
+	mutex_unlock(&ddata->lock);
+
+	return r;
+}
+
+static int dsicm_bl_get_intensity(struct backlight_device *dev)
+{
+	if (dev->props.fb_blank == FB_BLANK_UNBLANK &&
+			dev->props.power == FB_BLANK_UNBLANK)
+		return dev->props.brightness;
+
+	return 0;
+}
+
+static const struct backlight_ops dsicm_bl_ops = {
+	.get_brightness = dsicm_bl_get_intensity,
+	.update_status  = dsicm_bl_update_status,
+};
+
+static void dsicm_get_resolution(struct omap_dss_device *dssdev,
+		u16 *xres, u16 *yres)
+{
+	*xres = dssdev->panel.timings.x_res;
+	*yres = dssdev->panel.timings.y_res;
+}
+
+static ssize_t dsicm_num_errors_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *in = ddata->in;
+	u8 errors = 0;
+	int r;
+
+	mutex_lock(&ddata->lock);
+
+	if (ddata->enabled) {
+		in->ops.dsi->bus_lock(in);
+
+		r = dsicm_wake_up(ddata);
+		if (!r)
+			r = dsicm_dcs_read_1(ddata, DCS_READ_NUM_ERRORS,
+					&errors);
+
+		in->ops.dsi->bus_unlock(in);
+	} else {
+		r = -ENODEV;
+	}
+
+	mutex_unlock(&ddata->lock);
+
+	if (r)
+		return r;
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", errors);
+}
+
+static ssize_t dsicm_hw_revision_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *in = ddata->in;
+	u8 id1, id2, id3;
+	int r;
+
+	mutex_lock(&ddata->lock);
+
+	if (ddata->enabled) {
+		in->ops.dsi->bus_lock(in);
+
+		r = dsicm_wake_up(ddata);
+		if (!r)
+			r = dsicm_get_id(ddata, &id1, &id2, &id3);
+
+		in->ops.dsi->bus_unlock(in);
+	} else {
+		r = -ENODEV;
+	}
+
+	mutex_unlock(&ddata->lock);
+
+	if (r)
+		return r;
+
+	return snprintf(buf, PAGE_SIZE, "%02x.%02x.%02x\n", id1, id2, id3);
+}
+
+static ssize_t dsicm_store_ulps(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf, size_t count)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *in = ddata->in;
+	unsigned long t;
+	int r;
+
+	r = kstrtoul(buf, 0, &t);
+	if (r)
+		return r;
+
+	mutex_lock(&ddata->lock);
+
+	if (ddata->enabled) {
+		in->ops.dsi->bus_lock(in);
+
+		if (t)
+			r = dsicm_enter_ulps(ddata);
+		else
+			r = dsicm_wake_up(ddata);
+
+		in->ops.dsi->bus_unlock(in);
+	}
+
+	mutex_unlock(&ddata->lock);
+
+	if (r)
+		return r;
+
+	return count;
+}
+
+static ssize_t dsicm_show_ulps(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	unsigned t;
+
+	mutex_lock(&ddata->lock);
+	t = ddata->ulps_enabled;
+	mutex_unlock(&ddata->lock);
+
+	return snprintf(buf, PAGE_SIZE, "%u\n", t);
+}
+
+static ssize_t dsicm_store_ulps_timeout(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf, size_t count)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *in = ddata->in;
+	unsigned long t;
+	int r;
+
+	r = kstrtoul(buf, 0, &t);
+	if (r)
+		return r;
+
+	mutex_lock(&ddata->lock);
+	ddata->ulps_timeout = t;
+
+	if (ddata->enabled) {
+		/* dsicm_wake_up will restart the timer */
+		in->ops.dsi->bus_lock(in);
+		r = dsicm_wake_up(ddata);
+		in->ops.dsi->bus_unlock(in);
+	}
+
+	mutex_unlock(&ddata->lock);
+
+	if (r)
+		return r;
+
+	return count;
+}
+
+static ssize_t dsicm_show_ulps_timeout(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	unsigned t;
+
+	mutex_lock(&ddata->lock);
+	t = ddata->ulps_timeout;
+	mutex_unlock(&ddata->lock);
+
+	return snprintf(buf, PAGE_SIZE, "%u\n", t);
+}
+
+static DEVICE_ATTR(num_dsi_errors, S_IRUGO, dsicm_num_errors_show, NULL);
+static DEVICE_ATTR(hw_revision, S_IRUGO, dsicm_hw_revision_show, NULL);
+static DEVICE_ATTR(ulps, S_IRUGO | S_IWUSR,
+		dsicm_show_ulps, dsicm_store_ulps);
+static DEVICE_ATTR(ulps_timeout, S_IRUGO | S_IWUSR,
+		dsicm_show_ulps_timeout, dsicm_store_ulps_timeout);
+
+static struct attribute *dsicm_attrs[] = {
+	&dev_attr_num_dsi_errors.attr,
+	&dev_attr_hw_revision.attr,
+	&dev_attr_ulps.attr,
+	&dev_attr_ulps_timeout.attr,
+	NULL,
+};
+
+static struct attribute_group dsicm_attr_group = {
+	.attrs = dsicm_attrs,
+};
+
+static void dsicm_hw_reset(struct panel_drv_data *ddata)
+{
+	if (!gpio_is_valid(ddata->reset_gpio))
+		return;
+
+	gpio_set_value(ddata->reset_gpio, 1);
+	udelay(10);
+	/* reset the panel */
+	gpio_set_value(ddata->reset_gpio, 0);
+	/* assert reset */
+	udelay(10);
+	gpio_set_value(ddata->reset_gpio, 1);
+	/* wait after releasing reset */
+	usleep_range(5000, 10000);
+}
+
+static int dsicm_power_on(struct panel_drv_data *ddata)
+{
+	struct omap_dss_device *in = ddata->in;
+	u8 id1, id2, id3;
+	int r;
+	struct omap_dss_dsi_config dsi_config = {
+		.mode = OMAP_DSS_DSI_CMD_MODE,
+		.pixel_format = OMAP_DSS_DSI_FMT_RGB888,
+		.timings = &ddata->timings,
+		.hs_clk_min = 150000000,
+		.hs_clk_max = 300000000,
+		.lp_clk_min = 7000000,
+		.lp_clk_max = 10000000,
+	};
+
+	r = in->ops.dsi->configure_pins(in, &ddata->pin_config);
+	if (r) {
+		dev_err(&ddata->pdev->dev, "failed to configure DSI pins\n");
+		goto err0;
+	};
+
+	r = in->ops.dsi->set_config(in, &dsi_config);
+	if (r) {
+		dev_err(&ddata->pdev->dev, "failed to configure DSI\n");
+		goto err0;
+	}
+
+	r = in->ops.dsi->enable(in);
+	if (r) {
+		dev_err(&ddata->pdev->dev, "failed to enable DSI\n");
+		goto err0;
+	}
+
+	dsicm_hw_reset(ddata);
+
+	in->ops.dsi->enable_hs(in, ddata->channel, false);
+
+	r = dsicm_sleep_out(ddata);
+	if (r)
+		goto err;
+
+	r = dsicm_get_id(ddata, &id1, &id2, &id3);
+	if (r)
+		goto err;
+
+	r = dsicm_dcs_write_1(ddata, DCS_BRIGHTNESS, 0xff);
+	if (r)
+		goto err;
+
+	r = dsicm_dcs_write_1(ddata, DCS_CTRL_DISPLAY,
+			(1<<2) | (1<<5));	/* BL | BCTRL */
+	if (r)
+		goto err;
+
+	r = dsicm_dcs_write_1(ddata, MIPI_DCS_SET_PIXEL_FORMAT,
+		MIPI_DCS_PIXEL_FMT_24BIT);
+	if (r)
+		goto err;
+
+	r = dsicm_dcs_write_0(ddata, MIPI_DCS_SET_DISPLAY_ON);
+	if (r)
+		goto err;
+
+	r = _dsicm_enable_te(ddata, ddata->te_enabled);
+	if (r)
+		goto err;
+
+	r = in->ops.dsi->enable_video_output(in, ddata->channel);
+	if (r)
+		goto err;
+
+	ddata->enabled = 1;
+
+	if (!ddata->intro_printed) {
+		dev_info(&ddata->pdev->dev, "panel revision %02x.%02x.%02x\n",
+			id1, id2, id3);
+		ddata->intro_printed = true;
+	}
+
+	in->ops.dsi->enable_hs(in, ddata->channel, true);
+
+	return 0;
+err:
+	dev_err(&ddata->pdev->dev, "error while enabling panel, issuing HW reset\n");
+
+	dsicm_hw_reset(ddata);
+
+	in->ops.dsi->disable(in, true, false);
+err0:
+	return r;
+}
+
+static void dsicm_power_off(struct panel_drv_data *ddata)
+{
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	in->ops.dsi->disable_video_output(in, ddata->channel);
+
+	r = dsicm_dcs_write_0(ddata, MIPI_DCS_SET_DISPLAY_OFF);
+	if (!r)
+		r = dsicm_sleep_in(ddata);
+
+	if (r) {
+		dev_err(&ddata->pdev->dev,
+				"error disabling panel, issuing HW reset\n");
+		dsicm_hw_reset(ddata);
+	}
+
+	in->ops.dsi->disable(in, true, false);
+
+	ddata->enabled = 0;
+}
+
+static int dsicm_panel_reset(struct panel_drv_data *ddata)
+{
+	dev_err(&ddata->pdev->dev, "performing LCD reset\n");
+
+	dsicm_power_off(ddata);
+	dsicm_hw_reset(ddata);
+	return dsicm_power_on(ddata);
+}
+
+static int dsicm_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	struct device *dev = &ddata->pdev->dev;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dsi->connect(in, dssdev);
+	if (r) {
+		dev_err(dev, "Failed to connect to video source\n");
+		return r;
+	}
+
+	r = in->ops.dsi->request_vc(ddata->in, &ddata->channel);
+	if (r) {
+		dev_err(dev, "failed to get virtual channel\n");
+		goto err_req_vc;
+	}
+
+	r = in->ops.dsi->set_vc_id(ddata->in, ddata->channel, TCH);
+	if (r) {
+		dev_err(dev, "failed to set VC_ID\n");
+		goto err_vc_id;
+	}
+
+	return 0;
+
+err_vc_id:
+	in->ops.dsi->release_vc(ddata->in, ddata->channel);
+err_req_vc:
+	in->ops.dsi->disconnect(in, dssdev);
+	return r;
+}
+
+static void dsicm_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dsi->release_vc(in, ddata->channel);
+	in->ops.dsi->disconnect(in, dssdev);
+}
+
+static int dsicm_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(&ddata->pdev->dev, "enable\n");
+
+	mutex_lock(&ddata->lock);
+
+	if (!omapdss_device_is_connected(dssdev)) {
+		r = -ENODEV;
+		goto err;
+	}
+
+	if (omapdss_device_is_enabled(dssdev)) {
+		r = 0;
+		goto err;
+	}
+
+	in->ops.dsi->bus_lock(in);
+
+	r = dsicm_power_on(ddata);
+
+	in->ops.dsi->bus_unlock(in);
+
+	if (r)
+		goto err;
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	mutex_unlock(&ddata->lock);
+
+	return 0;
+err:
+	dev_dbg(&ddata->pdev->dev, "enable failed\n");
+	mutex_unlock(&ddata->lock);
+	return r;
+}
+
+static void dsicm_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(&ddata->pdev->dev, "disable\n");
+
+	mutex_lock(&ddata->lock);
+
+	dsicm_cancel_ulps_work(ddata);
+
+	in->ops.dsi->bus_lock(in);
+
+	if (omapdss_device_is_enabled(dssdev)) {
+		r = dsicm_wake_up(ddata);
+		if (!r)
+			dsicm_power_off(ddata);
+	}
+
+	in->ops.dsi->bus_unlock(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+
+	mutex_unlock(&ddata->lock);
+}
+
+static void dsicm_framedone_cb(int err, void *data)
+{
+	struct panel_drv_data *ddata = data;
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(&ddata->pdev->dev, "framedone, err %d\n", err);
+	in->ops.dsi->bus_unlock(ddata->in);
+}
+
+static irqreturn_t dsicm_te_isr(int irq, void *data)
+{
+	struct panel_drv_data *ddata = data;
+	struct omap_dss_device *in = ddata->in;
+	int old;
+	int r;
+
+	old = atomic_cmpxchg(&ddata->do_update, 1, 0);
+
+	if (old) {
+		cancel_delayed_work(&ddata->te_timeout_work);
+
+		r = in->ops.dsi->update(in, ddata->channel, dsicm_framedone_cb,
+				ddata);
+		if (r)
+			goto err;
+	}
+
+	return IRQ_HANDLED;
+err:
+	dev_err(&ddata->pdev->dev, "start update failed\n");
+	in->ops.dsi->bus_unlock(in);
+	return IRQ_HANDLED;
+}
+
+static void dsicm_te_timeout_work_callback(struct work_struct *work)
+{
+	struct panel_drv_data *ddata = container_of(work, struct panel_drv_data,
+					te_timeout_work.work);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_err(&ddata->pdev->dev, "TE not received for 250ms!\n");
+
+	atomic_set(&ddata->do_update, 0);
+	in->ops.dsi->bus_unlock(in);
+}
+
+static int dsicm_update(struct omap_dss_device *dssdev,
+				    u16 x, u16 y, u16 w, u16 h)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(&ddata->pdev->dev, "update %d, %d, %d x %d\n", x, y, w, h);
+
+	mutex_lock(&ddata->lock);
+	in->ops.dsi->bus_lock(in);
+
+	r = dsicm_wake_up(ddata);
+	if (r)
+		goto err;
+
+	if (!ddata->enabled) {
+		r = 0;
+		goto err;
+	}
+
+	/* XXX no need to send this every frame, but dsi break if not done */
+	r = dsicm_set_update_window(ddata, 0, 0,
+			dssdev->panel.timings.x_res,
+			dssdev->panel.timings.y_res);
+	if (r)
+		goto err;
+
+	if (ddata->te_enabled && gpio_is_valid(ddata->ext_te_gpio)) {
+		schedule_delayed_work(&ddata->te_timeout_work,
+				msecs_to_jiffies(250));
+		atomic_set(&ddata->do_update, 1);
+	} else {
+		r = in->ops.dsi->update(in, ddata->channel, dsicm_framedone_cb,
+				ddata);
+		if (r)
+			goto err;
+	}
+
+	/* note: no bus_unlock here. unlock is in framedone_cb */
+	mutex_unlock(&ddata->lock);
+	return 0;
+err:
+	in->ops.dsi->bus_unlock(in);
+	mutex_unlock(&ddata->lock);
+	return r;
+}
+
+static int dsicm_sync(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(&ddata->pdev->dev, "sync\n");
+
+	mutex_lock(&ddata->lock);
+	in->ops.dsi->bus_lock(in);
+	in->ops.dsi->bus_unlock(in);
+	mutex_unlock(&ddata->lock);
+
+	dev_dbg(&ddata->pdev->dev, "sync done\n");
+
+	return 0;
+}
+
+static int _dsicm_enable_te(struct panel_drv_data *ddata, bool enable)
+{
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (enable)
+		r = dsicm_dcs_write_1(ddata, MIPI_DCS_SET_TEAR_ON, 0);
+	else
+		r = dsicm_dcs_write_0(ddata, MIPI_DCS_SET_TEAR_OFF);
+
+	if (!gpio_is_valid(ddata->ext_te_gpio))
+		in->ops.dsi->enable_te(in, enable);
+
+	/* possible panel bug */
+	msleep(100);
+
+	return r;
+}
+
+static int dsicm_enable_te(struct omap_dss_device *dssdev, bool enable)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	mutex_lock(&ddata->lock);
+
+	if (ddata->te_enabled == enable)
+		goto end;
+
+	in->ops.dsi->bus_lock(in);
+
+	if (ddata->enabled) {
+		r = dsicm_wake_up(ddata);
+		if (r)
+			goto err;
+
+		r = _dsicm_enable_te(ddata, enable);
+		if (r)
+			goto err;
+	}
+
+	ddata->te_enabled = enable;
+
+	in->ops.dsi->bus_unlock(in);
+end:
+	mutex_unlock(&ddata->lock);
+
+	return 0;
+err:
+	in->ops.dsi->bus_unlock(in);
+	mutex_unlock(&ddata->lock);
+
+	return r;
+}
+
+static int dsicm_get_te(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	int r;
+
+	mutex_lock(&ddata->lock);
+	r = ddata->te_enabled;
+	mutex_unlock(&ddata->lock);
+
+	return r;
+}
+
+static int dsicm_memory_read(struct omap_dss_device *dssdev,
+		void *buf, size_t size,
+		u16 x, u16 y, u16 w, u16 h)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+	int first = 1;
+	int plen;
+	unsigned buf_used = 0;
+
+	if (size < w * h * 3)
+		return -ENOMEM;
+
+	mutex_lock(&ddata->lock);
+
+	if (!ddata->enabled) {
+		r = -ENODEV;
+		goto err1;
+	}
+
+	size = min(w * h * 3,
+			dssdev->panel.timings.x_res *
+			dssdev->panel.timings.y_res * 3);
+
+	in->ops.dsi->bus_lock(in);
+
+	r = dsicm_wake_up(ddata);
+	if (r)
+		goto err2;
+
+	/* plen 1 or 2 goes into short packet. until checksum error is fixed,
+	 * use short packets. plen 32 works, but bigger packets seem to cause
+	 * an error. */
+	if (size % 2)
+		plen = 1;
+	else
+		plen = 2;
+
+	dsicm_set_update_window(ddata, x, y, w, h);
+
+	r = in->ops.dsi->set_max_rx_packet_size(in, ddata->channel, plen);
+	if (r)
+		goto err2;
+
+	while (buf_used < size) {
+		u8 dcs_cmd = first ? 0x2e : 0x3e;
+		first = 0;
+
+		r = in->ops.dsi->dcs_read(in, ddata->channel, dcs_cmd,
+				buf + buf_used, size - buf_used);
+
+		if (r < 0) {
+			dev_err(dssdev->dev, "read error\n");
+			goto err3;
+		}
+
+		buf_used += r;
+
+		if (r < plen) {
+			dev_err(&ddata->pdev->dev, "short read\n");
+			break;
+		}
+
+		if (signal_pending(current)) {
+			dev_err(&ddata->pdev->dev, "signal pending, "
+					"aborting memory read\n");
+			r = -ERESTARTSYS;
+			goto err3;
+		}
+	}
+
+	r = buf_used;
+
+err3:
+	in->ops.dsi->set_max_rx_packet_size(in, ddata->channel, 1);
+err2:
+	in->ops.dsi->bus_unlock(in);
+err1:
+	mutex_unlock(&ddata->lock);
+	return r;
+}
+
+static void dsicm_ulps_work(struct work_struct *work)
+{
+	struct panel_drv_data *ddata = container_of(work, struct panel_drv_data,
+			ulps_work.work);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	mutex_lock(&ddata->lock);
+
+	if (dssdev->state != OMAP_DSS_DISPLAY_ACTIVE || !ddata->enabled) {
+		mutex_unlock(&ddata->lock);
+		return;
+	}
+
+	in->ops.dsi->bus_lock(in);
+
+	dsicm_enter_ulps(ddata);
+
+	in->ops.dsi->bus_unlock(in);
+	mutex_unlock(&ddata->lock);
+}
+
+static struct omap_dss_driver dsicm_ops = {
+	.connect	= dsicm_connect,
+	.disconnect	= dsicm_disconnect,
+
+	.enable		= dsicm_enable,
+	.disable	= dsicm_disable,
+
+	.update		= dsicm_update,
+	.sync		= dsicm_sync,
+
+	.get_resolution	= dsicm_get_resolution,
+	.get_recommended_bpp = omapdss_default_get_recommended_bpp,
+
+	.enable_te	= dsicm_enable_te,
+	.get_te		= dsicm_get_te,
+
+	.memory_read	= dsicm_memory_read,
+};
+
+static int dsicm_probe_pdata(struct platform_device *pdev)
+{
+	const struct panel_dsicm_platform_data *pdata;
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "failed to find video source\n");
+		return -EPROBE_DEFER;
+	}
+	ddata->in = in;
+
+	ddata->reset_gpio = pdata->reset_gpio;
+
+	if (pdata->use_ext_te)
+		ddata->ext_te_gpio = pdata->ext_te_gpio;
+	else
+		ddata->ext_te_gpio = -1;
+
+	ddata->ulps_timeout = pdata->ulps_timeout;
+
+	ddata->use_dsi_backlight = pdata->use_dsi_backlight;
+
+	ddata->pin_config = pdata->pin_config;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int dsicm_probe(struct platform_device *pdev)
+{
+	struct backlight_properties props;
+	struct panel_drv_data *ddata;
+	struct backlight_device *bldev = NULL;
+	struct device *dev = &pdev->dev;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	dev_dbg(dev, "probe\n");
+
+	ddata = devm_kzalloc(dev, sizeof(*ddata), GFP_KERNEL);
+	if (!ddata)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+	ddata->pdev = pdev;
+
+	if (dev_get_platdata(dev)) {
+		r = dsicm_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	ddata->timings.x_res = 864;
+	ddata->timings.y_res = 480;
+	ddata->timings.pixel_clock = DIV_ROUND_UP(864 * 480 * 60, 1000);
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = dev;
+	dssdev->driver = &dsicm_ops;
+	dssdev->panel.timings = ddata->timings;
+	dssdev->type = OMAP_DISPLAY_TYPE_DSI;
+	dssdev->owner = THIS_MODULE;
+
+	dssdev->panel.dsi_pix_fmt = OMAP_DSS_DSI_FMT_RGB888;
+	dssdev->caps = OMAP_DSS_DISPLAY_CAP_MANUAL_UPDATE |
+		OMAP_DSS_DISPLAY_CAP_TEAR_ELIM;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	mutex_init(&ddata->lock);
+
+	atomic_set(&ddata->do_update, 0);
+
+	if (gpio_is_valid(ddata->reset_gpio)) {
+		r = devm_gpio_request_one(dev, ddata->reset_gpio,
+				GPIOF_OUT_INIT_LOW, "taal rst");
+		if (r) {
+			dev_err(dev, "failed to request reset gpio\n");
+			return r;
+		}
+	}
+
+	if (gpio_is_valid(ddata->ext_te_gpio)) {
+		r = devm_gpio_request_one(dev, ddata->ext_te_gpio,
+				GPIOF_IN, "taal irq");
+		if (r) {
+			dev_err(dev, "GPIO request failed\n");
+			return r;
+		}
+
+		r = devm_request_irq(dev, gpio_to_irq(ddata->ext_te_gpio),
+				dsicm_te_isr,
+				IRQF_TRIGGER_RISING,
+				"taal vsync", ddata);
+
+		if (r) {
+			dev_err(dev, "IRQ request failed\n");
+			return r;
+		}
+
+		INIT_DEFERRABLE_WORK(&ddata->te_timeout_work,
+					dsicm_te_timeout_work_callback);
+
+		dev_dbg(dev, "Using GPIO TE\n");
+	}
+
+	ddata->workqueue = create_singlethread_workqueue("dsicm_wq");
+	if (ddata->workqueue == NULL) {
+		dev_err(dev, "can't create workqueue\n");
+		return -ENOMEM;
+	}
+	INIT_DELAYED_WORK(&ddata->ulps_work, dsicm_ulps_work);
+
+	dsicm_hw_reset(ddata);
+
+	if (ddata->use_dsi_backlight) {
+		memset(&props, 0, sizeof(struct backlight_properties));
+		props.max_brightness = 255;
+
+		props.type = BACKLIGHT_RAW;
+		bldev = backlight_device_register(dev_name(dev),
+				dev, ddata, &dsicm_bl_ops, &props);
+		if (IS_ERR(bldev)) {
+			r = PTR_ERR(bldev);
+			goto err_bl;
+		}
+
+		ddata->bldev = bldev;
+
+		bldev->props.fb_blank = FB_BLANK_UNBLANK;
+		bldev->props.power = FB_BLANK_UNBLANK;
+		bldev->props.brightness = 255;
+
+		dsicm_bl_update_status(bldev);
+	}
+
+	r = sysfs_create_group(&dev->kobj, &dsicm_attr_group);
+	if (r) {
+		dev_err(dev, "failed to create sysfs files\n");
+		goto err_sysfs_create;
+	}
+
+	return 0;
+
+err_sysfs_create:
+	if (bldev != NULL)
+		backlight_device_unregister(bldev);
+err_bl:
+	destroy_workqueue(ddata->workqueue);
+err_reg:
+	return r;
+}
+
+static int __exit dsicm_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct backlight_device *bldev;
+
+	dev_dbg(&pdev->dev, "remove\n");
+
+	omapdss_unregister_display(dssdev);
+
+	dsicm_disable(dssdev);
+	dsicm_disconnect(dssdev);
+
+	sysfs_remove_group(&pdev->dev.kobj, &dsicm_attr_group);
+
+	bldev = ddata->bldev;
+	if (bldev != NULL) {
+		bldev->props.power = FB_BLANK_POWERDOWN;
+		dsicm_bl_update_status(bldev);
+		backlight_device_unregister(bldev);
+	}
+
+	omap_dss_put_device(ddata->in);
+
+	dsicm_cancel_ulps_work(ddata);
+	destroy_workqueue(ddata->workqueue);
+
+	/* reset, to be sure that the panel is in a valid state */
+	dsicm_hw_reset(ddata);
+
+	return 0;
+}
+
+static struct platform_driver dsicm_driver = {
+	.probe = dsicm_probe,
+	.remove = __exit_p(dsicm_remove),
+	.driver = {
+		.name = "panel-dsi-cm",
+		.owner = THIS_MODULE,
+	},
+};
+
+module_platform_driver(dsicm_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("Generic DSI Command Mode Panel Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-lgphilips-lb035q02.c b/drivers/video/omap2/displays-new/panel-lgphilips-lb035q02.c
new file mode 100644
index 0000000..6e8977b
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-lgphilips-lb035q02.c
@@ -0,0 +1,358 @@
+/*
+ * LG.Philips LB035Q02 LCD Panel driver
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ * Based on a driver by: Steve Sakoman <steve@sakoman.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/spi/spi.h>
+#include <linux/mutex.h>
+#include <linux/gpio.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+static struct omap_video_timings lb035q02_timings = {
+	.x_res = 320,
+	.y_res = 240,
+
+	.pixel_clock	= 6500,
+
+	.hsw		= 2,
+	.hfp		= 20,
+	.hbp		= 68,
+
+	.vsw		= 2,
+	.vfp		= 4,
+	.vbp		= 18,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.data_pclk_edge	= OMAPDSS_DRIVE_SIG_RISING_EDGE,
+	.de_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.sync_pclk_edge	= OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES,
+};
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	struct spi_device *spi;
+
+	int data_lines;
+
+	struct omap_video_timings videomode;
+
+	int reset_gpio;
+	int backlight_gpio;
+	int enable_gpio;
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static int lb035q02_write_reg(struct spi_device *spi, u8 reg, u16 val)
+{
+	struct spi_message msg;
+	struct spi_transfer index_xfer = {
+		.len		= 3,
+		.cs_change	= 1,
+	};
+	struct spi_transfer value_xfer = {
+		.len		= 3,
+	};
+	u8	buffer[16];
+
+	spi_message_init(&msg);
+
+	/* register index */
+	buffer[0] = 0x70;
+	buffer[1] = 0x00;
+	buffer[2] = reg & 0x7f;
+	index_xfer.tx_buf = buffer;
+	spi_message_add_tail(&index_xfer, &msg);
+
+	/* register value */
+	buffer[4] = 0x72;
+	buffer[5] = val >> 8;
+	buffer[6] = val;
+	value_xfer.tx_buf = buffer + 4;
+	spi_message_add_tail(&value_xfer, &msg);
+
+	return spi_sync(spi, &msg);
+}
+
+static void init_lb035q02_panel(struct spi_device *spi)
+{
+	/* Init sequence from page 28 of the lb035q02 spec */
+	lb035q02_write_reg(spi, 0x01, 0x6300);
+	lb035q02_write_reg(spi, 0x02, 0x0200);
+	lb035q02_write_reg(spi, 0x03, 0x0177);
+	lb035q02_write_reg(spi, 0x04, 0x04c7);
+	lb035q02_write_reg(spi, 0x05, 0xffc0);
+	lb035q02_write_reg(spi, 0x06, 0xe806);
+	lb035q02_write_reg(spi, 0x0a, 0x4008);
+	lb035q02_write_reg(spi, 0x0b, 0x0000);
+	lb035q02_write_reg(spi, 0x0d, 0x0030);
+	lb035q02_write_reg(spi, 0x0e, 0x2800);
+	lb035q02_write_reg(spi, 0x0f, 0x0000);
+	lb035q02_write_reg(spi, 0x16, 0x9f80);
+	lb035q02_write_reg(spi, 0x17, 0x0a0f);
+	lb035q02_write_reg(spi, 0x1e, 0x00c1);
+	lb035q02_write_reg(spi, 0x30, 0x0300);
+	lb035q02_write_reg(spi, 0x31, 0x0007);
+	lb035q02_write_reg(spi, 0x32, 0x0000);
+	lb035q02_write_reg(spi, 0x33, 0x0000);
+	lb035q02_write_reg(spi, 0x34, 0x0707);
+	lb035q02_write_reg(spi, 0x35, 0x0004);
+	lb035q02_write_reg(spi, 0x36, 0x0302);
+	lb035q02_write_reg(spi, 0x37, 0x0202);
+	lb035q02_write_reg(spi, 0x3a, 0x0a0d);
+	lb035q02_write_reg(spi, 0x3b, 0x0806);
+}
+
+static int lb035q02_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dpi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	init_lb035q02_panel(ddata->spi);
+
+	return 0;
+}
+
+static void lb035q02_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dpi->disconnect(in, dssdev);
+}
+
+static int lb035q02_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dpi->set_data_lines(in, ddata->data_lines);
+	in->ops.dpi->set_timings(in, &ddata->videomode);
+
+	r = in->ops.dpi->enable(in);
+	if (r)
+		return r;
+
+	if (gpio_is_valid(ddata->enable_gpio))
+		gpio_set_value_cansleep(ddata->enable_gpio, 1);
+
+	if (gpio_is_valid(ddata->backlight_gpio))
+		gpio_set_value_cansleep(ddata->backlight_gpio, 1);
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void lb035q02_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	if (gpio_is_valid(ddata->enable_gpio))
+		gpio_set_value_cansleep(ddata->enable_gpio, 0);
+
+	if (gpio_is_valid(ddata->backlight_gpio))
+		gpio_set_value_cansleep(ddata->backlight_gpio, 0);
+
+	in->ops.dpi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void lb035q02_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->videomode = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dpi->set_timings(in, timings);
+}
+
+static void lb035q02_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->videomode;
+}
+
+static int lb035q02_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dpi->check_timings(in, timings);
+}
+
+static struct omap_dss_driver lb035q02_ops = {
+	.connect	= lb035q02_connect,
+	.disconnect	= lb035q02_disconnect,
+
+	.enable		= lb035q02_enable,
+	.disable	= lb035q02_disable,
+
+	.set_timings	= lb035q02_set_timings,
+	.get_timings	= lb035q02_get_timings,
+	.check_timings	= lb035q02_check_timings,
+
+	.get_resolution	= omapdss_default_get_resolution,
+};
+
+static int lb035q02_probe_pdata(struct spi_device *spi)
+{
+	const struct panel_lb035q02_platform_data *pdata;
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&spi->dev);
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&spi->dev, "failed to find video source '%s'\n",
+				pdata->source);
+		return -EPROBE_DEFER;
+	}
+
+	ddata->in = in;
+
+	ddata->data_lines = pdata->data_lines;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	ddata->enable_gpio = pdata->enable_gpio;
+	ddata->backlight_gpio = pdata->backlight_gpio;
+
+	return 0;
+}
+
+static int lb035q02_panel_spi_probe(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&spi->dev, sizeof(*ddata), GFP_KERNEL);
+	if (ddata == NULL)
+		return -ENOMEM;
+
+	dev_set_drvdata(&spi->dev, ddata);
+
+	ddata->spi = spi;
+
+	if (dev_get_platdata(&spi->dev)) {
+		r = lb035q02_probe_pdata(spi);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	if (gpio_is_valid(ddata->enable_gpio)) {
+		r = devm_gpio_request_one(&spi->dev, ddata->enable_gpio,
+				GPIOF_OUT_INIT_LOW, "panel enable");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->backlight_gpio)) {
+		r = devm_gpio_request_one(&spi->dev, ddata->backlight_gpio,
+				GPIOF_OUT_INIT_LOW, "panel backlight");
+		if (r)
+			goto err_gpio;
+	}
+
+	ddata->videomode = lb035q02_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = &spi->dev;
+	dssdev->driver = &lb035q02_ops;
+	dssdev->type = OMAP_DISPLAY_TYPE_DPI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = ddata->videomode;
+	dssdev->phy.dpi.data_lines = ddata->data_lines;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&spi->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int lb035q02_panel_spi_remove(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_display(dssdev);
+
+	lb035q02_disable(dssdev);
+	lb035q02_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct spi_driver lb035q02_spi_driver = {
+	.probe		= lb035q02_panel_spi_probe,
+	.remove		= lb035q02_panel_spi_remove,
+	.driver		= {
+		.name	= "panel_lgphilips_lb035q02",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_spi_driver(lb035q02_spi_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("LG.Philips LB035Q02 LCD Panel driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-nec-nl8048hl11.c b/drivers/video/omap2/displays-new/panel-nec-nl8048hl11.c
new file mode 100644
index 0000000..bb217da
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-nec-nl8048hl11.c
@@ -0,0 +1,394 @@
+/*
+ * NEC NL8048HL11 Panel driver
+ *
+ * Copyright (C) 2010 Texas Instruments Inc.
+ * Author: Erik Gilling <konkers@android.com>
+ * Converted to new DSS device model: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/spi/spi.h>
+#include <linux/fb.h>
+#include <linux/gpio.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+struct panel_drv_data {
+	struct omap_dss_device	dssdev;
+	struct omap_dss_device *in;
+
+	struct omap_video_timings videomode;
+
+	int data_lines;
+
+	int res_gpio;
+	int qvga_gpio;
+
+	struct spi_device *spi;
+};
+
+#define LCD_XRES		800
+#define LCD_YRES		480
+/*
+ * NEC PIX Clock Ratings
+ * MIN:21.8MHz TYP:23.8MHz MAX:25.7MHz
+ */
+#define LCD_PIXEL_CLOCK		23800
+
+static const struct {
+	unsigned char addr;
+	unsigned char dat;
+} nec_8048_init_seq[] = {
+	{ 3, 0x01 }, { 0, 0x00 }, { 1, 0x01 }, { 4, 0x00 }, { 5, 0x14 },
+	{ 6, 0x24 }, { 16, 0xD7 }, { 17, 0x00 }, { 18, 0x00 }, { 19, 0x55 },
+	{ 20, 0x01 }, { 21, 0x70 }, { 22, 0x1E }, { 23, 0x25 },	{ 24, 0x25 },
+	{ 25, 0x02 }, { 26, 0x02 }, { 27, 0xA0 }, { 32, 0x2F }, { 33, 0x0F },
+	{ 34, 0x0F }, { 35, 0x0F }, { 36, 0x0F }, { 37, 0x0F },	{ 38, 0x0F },
+	{ 39, 0x00 }, { 40, 0x02 }, { 41, 0x02 }, { 42, 0x02 },	{ 43, 0x0F },
+	{ 44, 0x0F }, { 45, 0x0F }, { 46, 0x0F }, { 47, 0x0F },	{ 48, 0x0F },
+	{ 49, 0x0F }, { 50, 0x00 }, { 51, 0x02 }, { 52, 0x02 }, { 53, 0x02 },
+	{ 80, 0x0C }, { 83, 0x42 }, { 84, 0x42 }, { 85, 0x41 },	{ 86, 0x14 },
+	{ 89, 0x88 }, { 90, 0x01 }, { 91, 0x00 }, { 92, 0x02 },	{ 93, 0x0C },
+	{ 94, 0x1C }, { 95, 0x27 }, { 98, 0x49 }, { 99, 0x27 }, { 102, 0x76 },
+	{ 103, 0x27 }, { 112, 0x01 }, { 113, 0x0E }, { 114, 0x02 },
+	{ 115, 0x0C }, { 118, 0x0C }, { 121, 0x30 }, { 130, 0x00 },
+	{ 131, 0x00 }, { 132, 0xFC }, { 134, 0x00 }, { 136, 0x00 },
+	{ 138, 0x00 }, { 139, 0x00 }, { 140, 0x00 }, { 141, 0xFC },
+	{ 143, 0x00 }, { 145, 0x00 }, { 147, 0x00 }, { 148, 0x00 },
+	{ 149, 0x00 }, { 150, 0xFC }, { 152, 0x00 }, { 154, 0x00 },
+	{ 156, 0x00 }, { 157, 0x00 }, { 2, 0x00 },
+};
+
+static const struct omap_video_timings nec_8048_panel_timings = {
+	.x_res		= LCD_XRES,
+	.y_res		= LCD_YRES,
+	.pixel_clock	= LCD_PIXEL_CLOCK,
+	.hfp		= 6,
+	.hsw		= 1,
+	.hbp		= 4,
+	.vfp		= 3,
+	.vsw		= 1,
+	.vbp		= 4,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.data_pclk_edge	= OMAPDSS_DRIVE_SIG_RISING_EDGE,
+	.de_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.sync_pclk_edge	= OMAPDSS_DRIVE_SIG_RISING_EDGE,
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static int nec_8048_spi_send(struct spi_device *spi, unsigned char reg_addr,
+			unsigned char reg_data)
+{
+	int ret = 0;
+	unsigned int cmd = 0, data = 0;
+
+	cmd = 0x0000 | reg_addr; /* register address write */
+	data = 0x0100 | reg_data; /* register data write */
+	data = (cmd << 16) | data;
+
+	ret = spi_write(spi, (unsigned char *)&data, 4);
+	if (ret)
+		pr_err("error in spi_write %x\n", data);
+
+	return ret;
+}
+
+static int init_nec_8048_wvga_lcd(struct spi_device *spi)
+{
+	unsigned int i;
+	/* Initialization Sequence */
+	/* nec_8048_spi_send(spi, REG, VAL) */
+	for (i = 0; i < (ARRAY_SIZE(nec_8048_init_seq) - 1); i++)
+		nec_8048_spi_send(spi, nec_8048_init_seq[i].addr,
+				nec_8048_init_seq[i].dat);
+	udelay(20);
+	nec_8048_spi_send(spi, nec_8048_init_seq[i].addr,
+				nec_8048_init_seq[i].dat);
+	return 0;
+}
+
+static int nec_8048_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dpi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void nec_8048_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dpi->disconnect(in, dssdev);
+}
+
+static int nec_8048_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dpi->set_data_lines(in, ddata->data_lines);
+	in->ops.dpi->set_timings(in, &ddata->videomode);
+
+	r = in->ops.dpi->enable(in);
+	if (r)
+		return r;
+
+	if (gpio_is_valid(ddata->res_gpio))
+		gpio_set_value_cansleep(ddata->res_gpio, 1);
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void nec_8048_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	if (gpio_is_valid(ddata->res_gpio))
+		gpio_set_value_cansleep(ddata->res_gpio, 0);
+
+	in->ops.dpi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void nec_8048_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->videomode = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dpi->set_timings(in, timings);
+}
+
+static void nec_8048_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->videomode;
+}
+
+static int nec_8048_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dpi->check_timings(in, timings);
+}
+
+static struct omap_dss_driver nec_8048_ops = {
+	.connect	= nec_8048_connect,
+	.disconnect	= nec_8048_disconnect,
+
+	.enable		= nec_8048_enable,
+	.disable	= nec_8048_disable,
+
+	.set_timings	= nec_8048_set_timings,
+	.get_timings	= nec_8048_get_timings,
+	.check_timings	= nec_8048_check_timings,
+
+	.get_resolution	= omapdss_default_get_resolution,
+};
+
+
+static int nec_8048_probe_pdata(struct spi_device *spi)
+{
+	const struct panel_nec_nl8048hl11_platform_data *pdata;
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&spi->dev);
+
+	ddata->qvga_gpio = pdata->qvga_gpio;
+	ddata->res_gpio = pdata->res_gpio;
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&spi->dev, "failed to find video source '%s'\n",
+				pdata->source);
+		return -EPROBE_DEFER;
+	}
+	ddata->in = in;
+
+	ddata->data_lines = pdata->data_lines;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int nec_8048_probe(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	dev_dbg(&spi->dev, "%s\n", __func__);
+
+	spi->mode = SPI_MODE_0;
+	spi->bits_per_word = 32;
+
+	r = spi_setup(spi);
+	if (r < 0) {
+		dev_err(&spi->dev, "spi_setup failed: %d\n", r);
+		return r;
+	}
+
+	init_nec_8048_wvga_lcd(spi);
+
+	ddata = devm_kzalloc(&spi->dev, sizeof(*ddata), GFP_KERNEL);
+	if (ddata == NULL)
+		return -ENOMEM;
+
+	dev_set_drvdata(&spi->dev, ddata);
+
+	ddata->spi = spi;
+
+	if (dev_get_platdata(&spi->dev)) {
+		r = nec_8048_probe_pdata(spi);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	if (gpio_is_valid(ddata->qvga_gpio)) {
+		r = devm_gpio_request_one(&spi->dev, ddata->qvga_gpio,
+				GPIOF_OUT_INIT_HIGH, "lcd QVGA");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->res_gpio)) {
+		r = devm_gpio_request_one(&spi->dev, ddata->res_gpio,
+				GPIOF_OUT_INIT_LOW, "lcd RES");
+		if (r)
+			goto err_gpio;
+	}
+
+	ddata->videomode = nec_8048_panel_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = &spi->dev;
+	dssdev->driver = &nec_8048_ops;
+	dssdev->type = OMAP_DISPLAY_TYPE_DPI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = ddata->videomode;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&spi->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int nec_8048_remove(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(&ddata->spi->dev, "%s\n", __func__);
+
+	omapdss_unregister_display(dssdev);
+
+	nec_8048_disable(dssdev);
+	nec_8048_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int nec_8048_suspend(struct device *dev)
+{
+	struct spi_device *spi = to_spi_device(dev);
+
+	nec_8048_spi_send(spi, 2, 0x01);
+	mdelay(40);
+
+	return 0;
+}
+
+static int nec_8048_resume(struct device *dev)
+{
+	struct spi_device *spi = to_spi_device(dev);
+
+	/* reinitialize the panel */
+	spi_setup(spi);
+	nec_8048_spi_send(spi, 2, 0x00);
+	init_nec_8048_wvga_lcd(spi);
+
+	return 0;
+}
+static SIMPLE_DEV_PM_OPS(nec_8048_pm_ops, nec_8048_suspend,
+		nec_8048_resume);
+#define NEC_8048_PM_OPS (&nec_8048_pm_ops)
+#else
+#define NEC_8048_PM_OPS NULL
+#endif
+
+static struct spi_driver nec_8048_driver = {
+	.driver = {
+		.name	= "panel-nec-nl8048hl11",
+		.owner	= THIS_MODULE,
+		.pm	= NEC_8048_PM_OPS,
+	},
+	.probe	= nec_8048_probe,
+	.remove	= nec_8048_remove,
+};
+
+module_spi_driver(nec_8048_driver);
+
+MODULE_AUTHOR("Erik Gilling <konkers@android.com>");
+MODULE_DESCRIPTION("NEC-NL8048HL11 Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-sharp-ls037v7dw01.c b/drivers/video/omap2/displays-new/panel-sharp-ls037v7dw01.c
new file mode 100644
index 0000000..72a4fb5
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-sharp-ls037v7dw01.c
@@ -0,0 +1,324 @@
+/*
+ * LCD panel driver for Sharp LS037V7DW01
+ *
+ * Copyright (C) 2013 Texas Instruments
+ * Author: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/delay.h>
+#include <linux/gpio.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+struct panel_drv_data {
+	struct omap_dss_device dssdev;
+	struct omap_dss_device *in;
+
+	int data_lines;
+
+	struct omap_video_timings videomode;
+
+	int resb_gpio;
+	int ini_gpio;
+	int mo_gpio;
+	int lr_gpio;
+	int ud_gpio;
+};
+
+static const struct omap_video_timings sharp_ls_timings = {
+	.x_res = 480,
+	.y_res = 640,
+
+	.pixel_clock	= 19200,
+
+	.hsw		= 2,
+	.hfp		= 1,
+	.hbp		= 28,
+
+	.vsw		= 1,
+	.vfp		= 1,
+	.vbp		= 1,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.data_pclk_edge	= OMAPDSS_DRIVE_SIG_RISING_EDGE,
+	.de_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.sync_pclk_edge	= OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES,
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static int sharp_ls_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dpi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void sharp_ls_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dpi->disconnect(in, dssdev);
+}
+
+static int sharp_ls_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dpi->set_data_lines(in, ddata->data_lines);
+	in->ops.dpi->set_timings(in, &ddata->videomode);
+
+	r = in->ops.dpi->enable(in);
+	if (r)
+		return r;
+
+	/* wait couple of vsyncs until enabling the LCD */
+	msleep(50);
+
+	if (gpio_is_valid(ddata->resb_gpio))
+		gpio_set_value_cansleep(ddata->resb_gpio, 1);
+
+	if (gpio_is_valid(ddata->ini_gpio))
+		gpio_set_value_cansleep(ddata->ini_gpio, 1);
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void sharp_ls_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	if (gpio_is_valid(ddata->ini_gpio))
+		gpio_set_value_cansleep(ddata->ini_gpio, 0);
+
+	if (gpio_is_valid(ddata->resb_gpio))
+		gpio_set_value_cansleep(ddata->resb_gpio, 0);
+
+	/* wait at least 5 vsyncs after disabling the LCD */
+
+	msleep(100);
+
+	in->ops.dpi->disable(in);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void sharp_ls_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->videomode = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dpi->set_timings(in, timings);
+}
+
+static void sharp_ls_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->videomode;
+}
+
+static int sharp_ls_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dpi->check_timings(in, timings);
+}
+
+static struct omap_dss_driver sharp_ls_ops = {
+	.connect	= sharp_ls_connect,
+	.disconnect	= sharp_ls_disconnect,
+
+	.enable		= sharp_ls_enable,
+	.disable	= sharp_ls_disable,
+
+	.set_timings	= sharp_ls_set_timings,
+	.get_timings	= sharp_ls_get_timings,
+	.check_timings	= sharp_ls_check_timings,
+
+	.get_resolution	= omapdss_default_get_resolution,
+};
+
+static int sharp_ls_probe_pdata(struct platform_device *pdev)
+{
+	const struct panel_sharp_ls037v7dw01_platform_data *pdata;
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&pdev->dev);
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&pdev->dev, "failed to find video source '%s'\n",
+				pdata->source);
+		return -EPROBE_DEFER;
+	}
+
+	ddata->in = in;
+
+	ddata->data_lines = pdata->data_lines;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	ddata->resb_gpio = pdata->resb_gpio;
+	ddata->ini_gpio = pdata->ini_gpio;
+	ddata->mo_gpio = pdata->mo_gpio;
+	ddata->lr_gpio = pdata->lr_gpio;
+	ddata->ud_gpio = pdata->ud_gpio;
+
+	return 0;
+}
+
+static int sharp_ls_probe(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	ddata = devm_kzalloc(&pdev->dev, sizeof(*ddata), GFP_KERNEL);
+	if (ddata == NULL)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, ddata);
+
+	if (dev_get_platdata(&pdev->dev)) {
+		r = sharp_ls_probe_pdata(pdev);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	if (gpio_is_valid(ddata->mo_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->mo_gpio,
+				GPIOF_OUT_INIT_LOW, "lcd MO");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->lr_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->lr_gpio,
+				GPIOF_OUT_INIT_HIGH, "lcd LR");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->ud_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->ud_gpio,
+				GPIOF_OUT_INIT_HIGH, "lcd UD");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->resb_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->resb_gpio,
+				GPIOF_OUT_INIT_LOW, "lcd RESB");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->ini_gpio)) {
+		r = devm_gpio_request_one(&pdev->dev, ddata->ini_gpio,
+				GPIOF_OUT_INIT_LOW, "lcd INI");
+		if (r)
+			goto err_gpio;
+	}
+
+	ddata->videomode = sharp_ls_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = &pdev->dev;
+	dssdev->driver = &sharp_ls_ops;
+	dssdev->type = OMAP_DISPLAY_TYPE_DPI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = ddata->videomode;
+	dssdev->phy.dpi.data_lines = ddata->data_lines;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&pdev->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int __exit sharp_ls_remove(struct platform_device *pdev)
+{
+	struct panel_drv_data *ddata = platform_get_drvdata(pdev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	omapdss_unregister_display(dssdev);
+
+	sharp_ls_disable(dssdev);
+	sharp_ls_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct platform_driver sharp_ls_driver = {
+	.probe = sharp_ls_probe,
+	.remove = __exit_p(sharp_ls_remove),
+	.driver = {
+		.name = "panel-sharp-ls037v7dw01",
+		.owner = THIS_MODULE,
+	},
+};
+
+module_platform_driver(sharp_ls_driver);
+
+MODULE_AUTHOR("Tomi Valkeinen <tomi.valkeinen@ti.com>");
+MODULE_DESCRIPTION("Sharp LS037V7DW01 Panel Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-sony-acx565akm.c b/drivers/video/omap2/displays-new/panel-sony-acx565akm.c
new file mode 100644
index 0000000..e6d56f7
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-sony-acx565akm.c
@@ -0,0 +1,865 @@
+/*
+ * Sony ACX565AKM LCD Panel driver
+ *
+ * Copyright (C) 2010 Nokia Corporation
+ *
+ * Original Driver Author: Imre Deak <imre.deak@nokia.com>
+ * Based on panel-generic.c by Tomi Valkeinen <tomi.valkeinen@nokia.com>
+ * Adapted to new DSS2 framework: Roger Quadros <roger.quadros@nokia.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published by
+ * the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/spi/spi.h>
+#include <linux/jiffies.h>
+#include <linux/sched.h>
+#include <linux/backlight.h>
+#include <linux/fb.h>
+#include <linux/gpio.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+#define MIPID_CMD_READ_DISP_ID		0x04
+#define MIPID_CMD_READ_RED		0x06
+#define MIPID_CMD_READ_GREEN		0x07
+#define MIPID_CMD_READ_BLUE		0x08
+#define MIPID_CMD_READ_DISP_STATUS	0x09
+#define MIPID_CMD_RDDSDR		0x0F
+#define MIPID_CMD_SLEEP_IN		0x10
+#define MIPID_CMD_SLEEP_OUT		0x11
+#define MIPID_CMD_DISP_OFF		0x28
+#define MIPID_CMD_DISP_ON		0x29
+#define MIPID_CMD_WRITE_DISP_BRIGHTNESS	0x51
+#define MIPID_CMD_READ_DISP_BRIGHTNESS	0x52
+#define MIPID_CMD_WRITE_CTRL_DISP	0x53
+
+#define CTRL_DISP_BRIGHTNESS_CTRL_ON	(1 << 5)
+#define CTRL_DISP_AMBIENT_LIGHT_CTRL_ON	(1 << 4)
+#define CTRL_DISP_BACKLIGHT_ON		(1 << 2)
+#define CTRL_DISP_AUTO_BRIGHTNESS_ON	(1 << 1)
+
+#define MIPID_CMD_READ_CTRL_DISP	0x54
+#define MIPID_CMD_WRITE_CABC		0x55
+#define MIPID_CMD_READ_CABC		0x56
+
+#define MIPID_VER_LPH8923		3
+#define MIPID_VER_LS041Y3		4
+#define MIPID_VER_L4F00311		8
+#define MIPID_VER_ACX565AKM		9
+
+struct panel_drv_data {
+	struct omap_dss_device	dssdev;
+	struct omap_dss_device *in;
+
+	int reset_gpio;
+	int datapairs;
+
+	struct omap_video_timings videomode;
+
+	char		*name;
+	int		enabled;
+	int		model;
+	int		revision;
+	u8		display_id[3];
+	unsigned	has_bc:1;
+	unsigned	has_cabc:1;
+	unsigned	cabc_mode;
+	unsigned long	hw_guard_end;		/* next value of jiffies
+						   when we can issue the
+						   next sleep in/out command */
+	unsigned long	hw_guard_wait;		/* max guard time in jiffies */
+
+	struct spi_device	*spi;
+	struct mutex		mutex;
+
+	struct backlight_device *bl_dev;
+};
+
+static const struct omap_video_timings acx565akm_panel_timings = {
+	.x_res		= 800,
+	.y_res		= 480,
+	.pixel_clock	= 24000,
+	.hfp		= 28,
+	.hsw		= 4,
+	.hbp		= 24,
+	.vfp		= 3,
+	.vsw		= 3,
+	.vbp		= 4,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+
+	.data_pclk_edge	= OMAPDSS_DRIVE_SIG_RISING_EDGE,
+	.de_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.sync_pclk_edge	= OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES,
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static void acx565akm_transfer(struct panel_drv_data *ddata, int cmd,
+			      const u8 *wbuf, int wlen, u8 *rbuf, int rlen)
+{
+	struct spi_message	m;
+	struct spi_transfer	*x, xfer[5];
+	int			r;
+
+	BUG_ON(ddata->spi == NULL);
+
+	spi_message_init(&m);
+
+	memset(xfer, 0, sizeof(xfer));
+	x = &xfer[0];
+
+	cmd &=  0xff;
+	x->tx_buf = &cmd;
+	x->bits_per_word = 9;
+	x->len = 2;
+
+	if (rlen > 1 && wlen == 0) {
+		/*
+		 * Between the command and the response data there is a
+		 * dummy clock cycle. Add an extra bit after the command
+		 * word to account for this.
+		 */
+		x->bits_per_word = 10;
+		cmd <<= 1;
+	}
+	spi_message_add_tail(x, &m);
+
+	if (wlen) {
+		x++;
+		x->tx_buf = wbuf;
+		x->len = wlen;
+		x->bits_per_word = 9;
+		spi_message_add_tail(x, &m);
+	}
+
+	if (rlen) {
+		x++;
+		x->rx_buf	= rbuf;
+		x->len		= rlen;
+		spi_message_add_tail(x, &m);
+	}
+
+	r = spi_sync(ddata->spi, &m);
+	if (r < 0)
+		dev_dbg(&ddata->spi->dev, "spi_sync %d\n", r);
+}
+
+static inline void acx565akm_cmd(struct panel_drv_data *ddata, int cmd)
+{
+	acx565akm_transfer(ddata, cmd, NULL, 0, NULL, 0);
+}
+
+static inline void acx565akm_write(struct panel_drv_data *ddata,
+			       int reg, const u8 *buf, int len)
+{
+	acx565akm_transfer(ddata, reg, buf, len, NULL, 0);
+}
+
+static inline void acx565akm_read(struct panel_drv_data *ddata,
+			      int reg, u8 *buf, int len)
+{
+	acx565akm_transfer(ddata, reg, NULL, 0, buf, len);
+}
+
+static void hw_guard_start(struct panel_drv_data *ddata, int guard_msec)
+{
+	ddata->hw_guard_wait = msecs_to_jiffies(guard_msec);
+	ddata->hw_guard_end = jiffies + ddata->hw_guard_wait;
+}
+
+static void hw_guard_wait(struct panel_drv_data *ddata)
+{
+	unsigned long wait = ddata->hw_guard_end - jiffies;
+
+	if ((long)wait > 0 && wait <= ddata->hw_guard_wait) {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		schedule_timeout(wait);
+	}
+}
+
+static void set_sleep_mode(struct panel_drv_data *ddata, int on)
+{
+	int cmd;
+
+	if (on)
+		cmd = MIPID_CMD_SLEEP_IN;
+	else
+		cmd = MIPID_CMD_SLEEP_OUT;
+	/*
+	 * We have to keep 120msec between sleep in/out commands.
+	 * (8.2.15, 8.2.16).
+	 */
+	hw_guard_wait(ddata);
+	acx565akm_cmd(ddata, cmd);
+	hw_guard_start(ddata, 120);
+}
+
+static void set_display_state(struct panel_drv_data *ddata, int enabled)
+{
+	int cmd = enabled ? MIPID_CMD_DISP_ON : MIPID_CMD_DISP_OFF;
+
+	acx565akm_cmd(ddata, cmd);
+}
+
+static int panel_enabled(struct panel_drv_data *ddata)
+{
+	u32 disp_status;
+	int enabled;
+
+	acx565akm_read(ddata, MIPID_CMD_READ_DISP_STATUS,
+			(u8 *)&disp_status, 4);
+	disp_status = __be32_to_cpu(disp_status);
+	enabled = (disp_status & (1 << 17)) && (disp_status & (1 << 10));
+	dev_dbg(&ddata->spi->dev,
+		"LCD panel %senabled by bootloader (status 0x%04x)\n",
+		enabled ? "" : "not ", disp_status);
+	return enabled;
+}
+
+static int panel_detect(struct panel_drv_data *ddata)
+{
+	acx565akm_read(ddata, MIPID_CMD_READ_DISP_ID, ddata->display_id, 3);
+	dev_dbg(&ddata->spi->dev, "MIPI display ID: %02x%02x%02x\n",
+		ddata->display_id[0],
+		ddata->display_id[1],
+		ddata->display_id[2]);
+
+	switch (ddata->display_id[0]) {
+	case 0x10:
+		ddata->model = MIPID_VER_ACX565AKM;
+		ddata->name = "acx565akm";
+		ddata->has_bc = 1;
+		ddata->has_cabc = 1;
+		break;
+	case 0x29:
+		ddata->model = MIPID_VER_L4F00311;
+		ddata->name = "l4f00311";
+		break;
+	case 0x45:
+		ddata->model = MIPID_VER_LPH8923;
+		ddata->name = "lph8923";
+		break;
+	case 0x83:
+		ddata->model = MIPID_VER_LS041Y3;
+		ddata->name = "ls041y3";
+		break;
+	default:
+		ddata->name = "unknown";
+		dev_err(&ddata->spi->dev, "invalid display ID\n");
+		return -ENODEV;
+	}
+
+	ddata->revision = ddata->display_id[1];
+
+	dev_info(&ddata->spi->dev, "omapfb: %s rev %02x LCD detected\n",
+			ddata->name, ddata->revision);
+
+	return 0;
+}
+
+/*----------------------Backlight Control-------------------------*/
+
+static void enable_backlight_ctrl(struct panel_drv_data *ddata, int enable)
+{
+	u16 ctrl;
+
+	acx565akm_read(ddata, MIPID_CMD_READ_CTRL_DISP, (u8 *)&ctrl, 1);
+	if (enable) {
+		ctrl |= CTRL_DISP_BRIGHTNESS_CTRL_ON |
+			CTRL_DISP_BACKLIGHT_ON;
+	} else {
+		ctrl &= ~(CTRL_DISP_BRIGHTNESS_CTRL_ON |
+			  CTRL_DISP_BACKLIGHT_ON);
+	}
+
+	ctrl |= 1 << 8;
+	acx565akm_write(ddata, MIPID_CMD_WRITE_CTRL_DISP, (u8 *)&ctrl, 2);
+}
+
+static void set_cabc_mode(struct panel_drv_data *ddata, unsigned mode)
+{
+	u16 cabc_ctrl;
+
+	ddata->cabc_mode = mode;
+	if (!ddata->enabled)
+		return;
+	cabc_ctrl = 0;
+	acx565akm_read(ddata, MIPID_CMD_READ_CABC, (u8 *)&cabc_ctrl, 1);
+	cabc_ctrl &= ~3;
+	cabc_ctrl |= (1 << 8) | (mode & 3);
+	acx565akm_write(ddata, MIPID_CMD_WRITE_CABC, (u8 *)&cabc_ctrl, 2);
+}
+
+static unsigned get_cabc_mode(struct panel_drv_data *ddata)
+{
+	return ddata->cabc_mode;
+}
+
+static unsigned get_hw_cabc_mode(struct panel_drv_data *ddata)
+{
+	u8 cabc_ctrl;
+
+	acx565akm_read(ddata, MIPID_CMD_READ_CABC, &cabc_ctrl, 1);
+	return cabc_ctrl & 3;
+}
+
+static void acx565akm_set_brightness(struct panel_drv_data *ddata, int level)
+{
+	int bv;
+
+	bv = level | (1 << 8);
+	acx565akm_write(ddata, MIPID_CMD_WRITE_DISP_BRIGHTNESS, (u8 *)&bv, 2);
+
+	if (level)
+		enable_backlight_ctrl(ddata, 1);
+	else
+		enable_backlight_ctrl(ddata, 0);
+}
+
+static int acx565akm_get_actual_brightness(struct panel_drv_data *ddata)
+{
+	u8 bv;
+
+	acx565akm_read(ddata, MIPID_CMD_READ_DISP_BRIGHTNESS, &bv, 1);
+
+	return bv;
+}
+
+
+static int acx565akm_bl_update_status(struct backlight_device *dev)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&dev->dev);
+	int r;
+	int level;
+
+	dev_dbg(&ddata->spi->dev, "%s\n", __func__);
+
+	mutex_lock(&ddata->mutex);
+
+	if (dev->props.fb_blank == FB_BLANK_UNBLANK &&
+			dev->props.power == FB_BLANK_UNBLANK)
+		level = dev->props.brightness;
+	else
+		level = 0;
+
+	r = 0;
+	if (ddata->has_bc)
+		acx565akm_set_brightness(ddata, level);
+	else
+		r = -ENODEV;
+
+	mutex_unlock(&ddata->mutex);
+
+	return r;
+}
+
+static int acx565akm_bl_get_intensity(struct backlight_device *dev)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&dev->dev);
+
+	dev_dbg(&dev->dev, "%s\n", __func__);
+
+	if (!ddata->has_bc)
+		return -ENODEV;
+
+	if (dev->props.fb_blank == FB_BLANK_UNBLANK &&
+			dev->props.power == FB_BLANK_UNBLANK) {
+		if (ddata->has_bc)
+			return acx565akm_get_actual_brightness(ddata);
+		else
+			return dev->props.brightness;
+	}
+
+	return 0;
+}
+
+static const struct backlight_ops acx565akm_bl_ops = {
+	.get_brightness = acx565akm_bl_get_intensity,
+	.update_status  = acx565akm_bl_update_status,
+};
+
+/*--------------------Auto Brightness control via Sysfs---------------------*/
+
+static const char * const cabc_modes[] = {
+	"off",		/* always used when CABC is not supported */
+	"ui",
+	"still-image",
+	"moving-image",
+};
+
+static ssize_t show_cabc_mode(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	const char *mode_str;
+	int mode;
+	int len;
+
+	if (!ddata->has_cabc)
+		mode = 0;
+	else
+		mode = get_cabc_mode(ddata);
+	mode_str = "unknown";
+	if (mode >= 0 && mode < ARRAY_SIZE(cabc_modes))
+		mode_str = cabc_modes[mode];
+	len = snprintf(buf, PAGE_SIZE, "%s\n", mode_str);
+
+	return len < PAGE_SIZE - 1 ? len : PAGE_SIZE - 1;
+}
+
+static ssize_t store_cabc_mode(struct device *dev,
+		struct device_attribute *attr,
+		const char *buf, size_t count)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(cabc_modes); i++) {
+		const char *mode_str = cabc_modes[i];
+		int cmp_len = strlen(mode_str);
+
+		if (count > 0 && buf[count - 1] == '\n')
+			count--;
+		if (count != cmp_len)
+			continue;
+
+		if (strncmp(buf, mode_str, cmp_len) == 0)
+			break;
+	}
+
+	if (i == ARRAY_SIZE(cabc_modes))
+		return -EINVAL;
+
+	if (!ddata->has_cabc && i != 0)
+		return -EINVAL;
+
+	mutex_lock(&ddata->mutex);
+	set_cabc_mode(ddata, i);
+	mutex_unlock(&ddata->mutex);
+
+	return count;
+}
+
+static ssize_t show_cabc_available_modes(struct device *dev,
+		struct device_attribute *attr,
+		char *buf)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	int len;
+	int i;
+
+	if (!ddata->has_cabc)
+		return snprintf(buf, PAGE_SIZE, "%s\n", cabc_modes[0]);
+
+	for (i = 0, len = 0;
+	     len < PAGE_SIZE && i < ARRAY_SIZE(cabc_modes); i++)
+		len += snprintf(&buf[len], PAGE_SIZE - len, "%s%s%s",
+			i ? " " : "", cabc_modes[i],
+			i == ARRAY_SIZE(cabc_modes) - 1 ? "\n" : "");
+
+	return len < PAGE_SIZE ? len : PAGE_SIZE - 1;
+}
+
+static DEVICE_ATTR(cabc_mode, S_IRUGO | S_IWUSR,
+		show_cabc_mode, store_cabc_mode);
+static DEVICE_ATTR(cabc_available_modes, S_IRUGO,
+		show_cabc_available_modes, NULL);
+
+static struct attribute *bldev_attrs[] = {
+	&dev_attr_cabc_mode.attr,
+	&dev_attr_cabc_available_modes.attr,
+	NULL,
+};
+
+static struct attribute_group bldev_attr_group = {
+	.attrs = bldev_attrs,
+};
+
+static int acx565akm_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.sdi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void acx565akm_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.sdi->disconnect(in, dssdev);
+}
+
+static int acx565akm_panel_power_on(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	dev_dbg(&ddata->spi->dev, "%s\n", __func__);
+
+	in->ops.sdi->set_timings(in, &ddata->videomode);
+	in->ops.sdi->set_datapairs(in, ddata->datapairs);
+
+	r = in->ops.sdi->enable(in);
+	if (r) {
+		pr_err("%s sdi enable failed\n", __func__);
+		return r;
+	}
+
+	/*FIXME tweak me */
+	msleep(50);
+
+	if (gpio_is_valid(ddata->reset_gpio))
+		gpio_set_value(ddata->reset_gpio, 1);
+
+	if (ddata->enabled) {
+		dev_dbg(&ddata->spi->dev, "panel already enabled\n");
+		return 0;
+	}
+
+	/*
+	 * We have to meet all the following delay requirements:
+	 * 1. tRW: reset pulse width 10usec (7.12.1)
+	 * 2. tRT: reset cancel time 5msec (7.12.1)
+	 * 3. Providing PCLK,HS,VS signals for 2 frames = ~50msec worst
+	 *    case (7.6.2)
+	 * 4. 120msec before the sleep out command (7.12.1)
+	 */
+	msleep(120);
+
+	set_sleep_mode(ddata, 0);
+	ddata->enabled = 1;
+
+	/* 5msec between sleep out and the next command. (8.2.16) */
+	usleep_range(5000, 10000);
+	set_display_state(ddata, 1);
+	set_cabc_mode(ddata, ddata->cabc_mode);
+
+	mutex_unlock(&ddata->mutex);
+
+	return acx565akm_bl_update_status(ddata->bl_dev);
+}
+
+static void acx565akm_panel_power_off(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(dssdev->dev, "%s\n", __func__);
+
+	if (!ddata->enabled)
+		return;
+
+	set_display_state(ddata, 0);
+	set_sleep_mode(ddata, 1);
+	ddata->enabled = 0;
+	/*
+	 * We have to provide PCLK,HS,VS signals for 2 frames (worst case
+	 * ~50msec) after sending the sleep in command and asserting the
+	 * reset signal. We probably could assert the reset w/o the delay
+	 * but we still delay to avoid possible artifacts. (7.6.1)
+	 */
+	msleep(50);
+
+	if (gpio_is_valid(ddata->reset_gpio))
+		gpio_set_value(ddata->reset_gpio, 0);
+
+	/* FIXME need to tweak this delay */
+	msleep(100);
+
+	in->ops.sdi->disable(in);
+}
+
+static int acx565akm_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	int r;
+
+	dev_dbg(dssdev->dev, "%s\n", __func__);
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	mutex_lock(&ddata->mutex);
+	r = acx565akm_panel_power_on(dssdev);
+	mutex_unlock(&ddata->mutex);
+
+	if (r)
+		return r;
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void acx565akm_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	dev_dbg(dssdev->dev, "%s\n", __func__);
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	mutex_lock(&ddata->mutex);
+	acx565akm_panel_power_off(dssdev);
+	mutex_unlock(&ddata->mutex);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void acx565akm_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->videomode = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.sdi->set_timings(in, timings);
+}
+
+static void acx565akm_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->videomode;
+}
+
+static int acx565akm_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.sdi->check_timings(in, timings);
+}
+
+static struct omap_dss_driver acx565akm_ops = {
+	.connect	= acx565akm_connect,
+	.disconnect	= acx565akm_disconnect,
+
+	.enable		= acx565akm_enable,
+	.disable	= acx565akm_disable,
+
+	.set_timings	= acx565akm_set_timings,
+	.get_timings	= acx565akm_get_timings,
+	.check_timings	= acx565akm_check_timings,
+
+	.get_resolution	= omapdss_default_get_resolution,
+};
+
+static int acx565akm_probe_pdata(struct spi_device *spi)
+{
+	const struct panel_acx565akm_platform_data *pdata;
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&spi->dev);
+
+	ddata->reset_gpio = pdata->reset_gpio;
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&spi->dev, "failed to find video source '%s'\n",
+				pdata->source);
+		return -EPROBE_DEFER;
+	}
+	ddata->in = in;
+
+	ddata->datapairs = pdata->datapairs;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int acx565akm_probe(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	struct backlight_device *bldev;
+	int max_brightness, brightness;
+	struct backlight_properties props;
+	int r;
+
+	dev_dbg(&spi->dev, "%s\n", __func__);
+
+	spi->mode = SPI_MODE_3;
+
+	ddata = devm_kzalloc(&spi->dev, sizeof(*ddata), GFP_KERNEL);
+	if (ddata == NULL)
+		return -ENOMEM;
+
+	dev_set_drvdata(&spi->dev, ddata);
+
+	ddata->spi = spi;
+
+	mutex_init(&ddata->mutex);
+
+	if (dev_get_platdata(&spi->dev)) {
+		r = acx565akm_probe_pdata(spi);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	if (gpio_is_valid(ddata->reset_gpio)) {
+		r = devm_gpio_request_one(&spi->dev, ddata->reset_gpio,
+				GPIOF_OUT_INIT_LOW, "lcd reset");
+		if (r)
+			goto err_gpio;
+	}
+
+	if (gpio_is_valid(ddata->reset_gpio))
+		gpio_set_value(ddata->reset_gpio, 1);
+
+	/*
+	 * After reset we have to wait 5 msec before the first
+	 * command can be sent.
+	 */
+	usleep_range(5000, 10000);
+
+	ddata->enabled = panel_enabled(ddata);
+
+	r = panel_detect(ddata);
+
+	if (!ddata->enabled && gpio_is_valid(ddata->reset_gpio))
+		gpio_set_value(ddata->reset_gpio, 0);
+
+	if (r) {
+		dev_err(&spi->dev, "%s panel detect error\n", __func__);
+		goto err_detect;
+	}
+
+	memset(&props, 0, sizeof(props));
+	props.fb_blank = FB_BLANK_UNBLANK;
+	props.power = FB_BLANK_UNBLANK;
+	props.type = BACKLIGHT_RAW;
+
+	bldev = backlight_device_register("acx565akm", &ddata->spi->dev,
+			ddata, &acx565akm_bl_ops, &props);
+	ddata->bl_dev = bldev;
+	if (ddata->has_cabc) {
+		r = sysfs_create_group(&bldev->dev.kobj, &bldev_attr_group);
+		if (r) {
+			dev_err(&bldev->dev,
+				"%s failed to create sysfs files\n", __func__);
+			goto err_sysfs;
+		}
+		ddata->cabc_mode = get_hw_cabc_mode(ddata);
+	}
+
+	max_brightness = 255;
+
+	if (ddata->has_bc)
+		brightness = acx565akm_get_actual_brightness(ddata);
+	else
+		brightness = 0;
+
+	bldev->props.max_brightness = max_brightness;
+	bldev->props.brightness = brightness;
+
+	acx565akm_bl_update_status(bldev);
+
+
+	ddata->videomode = acx565akm_panel_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = &spi->dev;
+	dssdev->driver = &acx565akm_ops;
+	dssdev->type = OMAP_DISPLAY_TYPE_SDI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = ddata->videomode;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&spi->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+	sysfs_remove_group(&bldev->dev.kobj, &bldev_attr_group);
+err_sysfs:
+	backlight_device_unregister(bldev);
+err_detect:
+err_gpio:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int acx565akm_remove(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(&ddata->spi->dev, "%s\n", __func__);
+
+	sysfs_remove_group(&ddata->bl_dev->dev.kobj, &bldev_attr_group);
+	backlight_device_unregister(ddata->bl_dev);
+
+	omapdss_unregister_display(dssdev);
+
+	acx565akm_disable(dssdev);
+	acx565akm_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	return 0;
+}
+
+static struct spi_driver acx565akm_driver = {
+	.driver = {
+		.name	= "acx565akm",
+		.owner	= THIS_MODULE,
+	},
+	.probe	= acx565akm_probe,
+	.remove	= acx565akm_remove,
+};
+
+module_spi_driver(acx565akm_driver);
+
+MODULE_AUTHOR("Nokia Corporation");
+MODULE_DESCRIPTION("acx565akm LCD Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays-new/panel-tpo-td043mtea1.c b/drivers/video/omap2/displays-new/panel-tpo-td043mtea1.c
new file mode 100644
index 0000000..eadc652
--- /dev/null
+++ b/drivers/video/omap2/displays-new/panel-tpo-td043mtea1.c
@@ -0,0 +1,646 @@
+/*
+ * TPO TD043MTEA1 Panel driver
+ *
+ * Author: Gražvydas Ignotas <notasas@gmail.com>
+ * Converted to new DSS device model: Tomi Valkeinen <tomi.valkeinen@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/spi/spi.h>
+#include <linux/regulator/consumer.h>
+#include <linux/gpio.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+
+#include <video/omapdss.h>
+#include <video/omap-panel-data.h>
+
+#define TPO_R02_MODE(x)		((x) & 7)
+#define TPO_R02_MODE_800x480	7
+#define TPO_R02_NCLK_RISING	BIT(3)
+#define TPO_R02_HSYNC_HIGH	BIT(4)
+#define TPO_R02_VSYNC_HIGH	BIT(5)
+
+#define TPO_R03_NSTANDBY	BIT(0)
+#define TPO_R03_EN_CP_CLK	BIT(1)
+#define TPO_R03_EN_VGL_PUMP	BIT(2)
+#define TPO_R03_EN_PWM		BIT(3)
+#define TPO_R03_DRIVING_CAP_100	BIT(4)
+#define TPO_R03_EN_PRE_CHARGE	BIT(6)
+#define TPO_R03_SOFTWARE_CTL	BIT(7)
+
+#define TPO_R04_NFLIP_H		BIT(0)
+#define TPO_R04_NFLIP_V		BIT(1)
+#define TPO_R04_CP_CLK_FREQ_1H	BIT(2)
+#define TPO_R04_VGL_FREQ_1H	BIT(4)
+
+#define TPO_R03_VAL_NORMAL (TPO_R03_NSTANDBY | TPO_R03_EN_CP_CLK | \
+			TPO_R03_EN_VGL_PUMP |  TPO_R03_EN_PWM | \
+			TPO_R03_DRIVING_CAP_100 | TPO_R03_EN_PRE_CHARGE | \
+			TPO_R03_SOFTWARE_CTL)
+
+#define TPO_R03_VAL_STANDBY (TPO_R03_DRIVING_CAP_100 | \
+			TPO_R03_EN_PRE_CHARGE | TPO_R03_SOFTWARE_CTL)
+
+static const u16 tpo_td043_def_gamma[12] = {
+	105, 315, 381, 431, 490, 537, 579, 686, 780, 837, 880, 1023
+};
+
+struct panel_drv_data {
+	struct omap_dss_device	dssdev;
+	struct omap_dss_device *in;
+
+	struct omap_video_timings videomode;
+
+	int data_lines;
+
+	struct spi_device *spi;
+	struct regulator *vcc_reg;
+	int nreset_gpio;
+	u16 gamma[12];
+	u32 mode;
+	u32 hmirror:1;
+	u32 vmirror:1;
+	u32 powered_on:1;
+	u32 spi_suspended:1;
+	u32 power_on_resume:1;
+};
+
+static const struct omap_video_timings tpo_td043_timings = {
+	.x_res		= 800,
+	.y_res		= 480,
+
+	.pixel_clock	= 36000,
+
+	.hsw		= 1,
+	.hfp		= 68,
+	.hbp		= 214,
+
+	.vsw		= 1,
+	.vfp		= 39,
+	.vbp		= 34,
+
+	.vsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.hsync_level	= OMAPDSS_SIG_ACTIVE_LOW,
+	.data_pclk_edge	= OMAPDSS_DRIVE_SIG_FALLING_EDGE,
+	.de_level	= OMAPDSS_SIG_ACTIVE_HIGH,
+	.sync_pclk_edge	= OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES,
+};
+
+#define to_panel_data(p) container_of(p, struct panel_drv_data, dssdev)
+
+static int tpo_td043_write(struct spi_device *spi, u8 addr, u8 data)
+{
+	struct spi_message	m;
+	struct spi_transfer	xfer;
+	u16			w;
+	int			r;
+
+	spi_message_init(&m);
+
+	memset(&xfer, 0, sizeof(xfer));
+
+	w = ((u16)addr << 10) | (1 << 8) | data;
+	xfer.tx_buf = &w;
+	xfer.bits_per_word = 16;
+	xfer.len = 2;
+	spi_message_add_tail(&xfer, &m);
+
+	r = spi_sync(spi, &m);
+	if (r < 0)
+		dev_warn(&spi->dev, "failed to write to LCD reg (%d)\n", r);
+	return r;
+}
+
+static void tpo_td043_write_gamma(struct spi_device *spi, u16 gamma[12])
+{
+	u8 i, val;
+
+	/* gamma bits [9:8] */
+	for (val = i = 0; i < 4; i++)
+		val |= (gamma[i] & 0x300) >> ((i + 1) * 2);
+	tpo_td043_write(spi, 0x11, val);
+
+	for (val = i = 0; i < 4; i++)
+		val |= (gamma[i+4] & 0x300) >> ((i + 1) * 2);
+	tpo_td043_write(spi, 0x12, val);
+
+	for (val = i = 0; i < 4; i++)
+		val |= (gamma[i+8] & 0x300) >> ((i + 1) * 2);
+	tpo_td043_write(spi, 0x13, val);
+
+	/* gamma bits [7:0] */
+	for (val = i = 0; i < 12; i++)
+		tpo_td043_write(spi, 0x14 + i, gamma[i] & 0xff);
+}
+
+static int tpo_td043_write_mirror(struct spi_device *spi, bool h, bool v)
+{
+	u8 reg4 = TPO_R04_NFLIP_H | TPO_R04_NFLIP_V |
+		TPO_R04_CP_CLK_FREQ_1H | TPO_R04_VGL_FREQ_1H;
+	if (h)
+		reg4 &= ~TPO_R04_NFLIP_H;
+	if (v)
+		reg4 &= ~TPO_R04_NFLIP_V;
+
+	return tpo_td043_write(spi, 4, reg4);
+}
+
+static int tpo_td043_set_hmirror(struct omap_dss_device *dssdev, bool enable)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
+
+	ddata->hmirror = enable;
+	return tpo_td043_write_mirror(ddata->spi, ddata->hmirror,
+			ddata->vmirror);
+}
+
+static bool tpo_td043_get_hmirror(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
+
+	return ddata->hmirror;
+}
+
+static ssize_t tpo_td043_vmirror_show(struct device *dev,
+	struct device_attribute *attr, char *buf)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", ddata->vmirror);
+}
+
+static ssize_t tpo_td043_vmirror_store(struct device *dev,
+	struct device_attribute *attr, const char *buf, size_t count)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	int val;
+	int ret;
+
+	ret = kstrtoint(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+
+	val = !!val;
+
+	ret = tpo_td043_write_mirror(ddata->spi, ddata->hmirror, val);
+	if (ret < 0)
+		return ret;
+
+	ddata->vmirror = val;
+
+	return count;
+}
+
+static ssize_t tpo_td043_mode_show(struct device *dev,
+	struct device_attribute *attr, char *buf)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+
+	return snprintf(buf, PAGE_SIZE, "%d\n", ddata->mode);
+}
+
+static ssize_t tpo_td043_mode_store(struct device *dev,
+	struct device_attribute *attr, const char *buf, size_t count)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	long val;
+	int ret;
+
+	ret = kstrtol(buf, 0, &val);
+	if (ret != 0 || val & ~7)
+		return -EINVAL;
+
+	ddata->mode = val;
+
+	val |= TPO_R02_NCLK_RISING;
+	tpo_td043_write(ddata->spi, 2, val);
+
+	return count;
+}
+
+static ssize_t tpo_td043_gamma_show(struct device *dev,
+	struct device_attribute *attr, char *buf)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	ssize_t len = 0;
+	int ret;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ddata->gamma); i++) {
+		ret = snprintf(buf + len, PAGE_SIZE - len, "%u ",
+				ddata->gamma[i]);
+		if (ret < 0)
+			return ret;
+		len += ret;
+	}
+	buf[len - 1] = '\n';
+
+	return len;
+}
+
+static ssize_t tpo_td043_gamma_store(struct device *dev,
+	struct device_attribute *attr, const char *buf, size_t count)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	unsigned int g[12];
+	int ret;
+	int i;
+
+	ret = sscanf(buf, "%u %u %u %u %u %u %u %u %u %u %u %u",
+			&g[0], &g[1], &g[2], &g[3], &g[4], &g[5],
+			&g[6], &g[7], &g[8], &g[9], &g[10], &g[11]);
+
+	if (ret != 12)
+		return -EINVAL;
+
+	for (i = 0; i < 12; i++)
+		ddata->gamma[i] = g[i];
+
+	tpo_td043_write_gamma(ddata->spi, ddata->gamma);
+
+	return count;
+}
+
+static DEVICE_ATTR(vmirror, S_IRUGO | S_IWUSR,
+		tpo_td043_vmirror_show, tpo_td043_vmirror_store);
+static DEVICE_ATTR(mode, S_IRUGO | S_IWUSR,
+		tpo_td043_mode_show, tpo_td043_mode_store);
+static DEVICE_ATTR(gamma, S_IRUGO | S_IWUSR,
+		tpo_td043_gamma_show, tpo_td043_gamma_store);
+
+static struct attribute *tpo_td043_attrs[] = {
+	&dev_attr_vmirror.attr,
+	&dev_attr_mode.attr,
+	&dev_attr_gamma.attr,
+	NULL,
+};
+
+static struct attribute_group tpo_td043_attr_group = {
+	.attrs = tpo_td043_attrs,
+};
+
+static int tpo_td043_power_on(struct panel_drv_data *ddata)
+{
+	int r;
+
+	if (ddata->powered_on)
+		return 0;
+
+	r = regulator_enable(ddata->vcc_reg);
+	if (r != 0)
+		return r;
+
+	/* wait for panel to stabilize */
+	msleep(160);
+
+	if (gpio_is_valid(ddata->nreset_gpio))
+		gpio_set_value(ddata->nreset_gpio, 1);
+
+	tpo_td043_write(ddata->spi, 2,
+			TPO_R02_MODE(ddata->mode) | TPO_R02_NCLK_RISING);
+	tpo_td043_write(ddata->spi, 3, TPO_R03_VAL_NORMAL);
+	tpo_td043_write(ddata->spi, 0x20, 0xf0);
+	tpo_td043_write(ddata->spi, 0x21, 0xf0);
+	tpo_td043_write_mirror(ddata->spi, ddata->hmirror,
+			ddata->vmirror);
+	tpo_td043_write_gamma(ddata->spi, ddata->gamma);
+
+	ddata->powered_on = 1;
+	return 0;
+}
+
+static void tpo_td043_power_off(struct panel_drv_data *ddata)
+{
+	if (!ddata->powered_on)
+		return;
+
+	tpo_td043_write(ddata->spi, 3,
+			TPO_R03_VAL_STANDBY | TPO_R03_EN_PWM);
+
+	if (gpio_is_valid(ddata->nreset_gpio))
+		gpio_set_value(ddata->nreset_gpio, 0);
+
+	/* wait for at least 2 vsyncs before cutting off power */
+	msleep(50);
+
+	tpo_td043_write(ddata->spi, 3, TPO_R03_VAL_STANDBY);
+
+	regulator_disable(ddata->vcc_reg);
+
+	ddata->powered_on = 0;
+}
+
+static int tpo_td043_connect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (omapdss_device_is_connected(dssdev))
+		return 0;
+
+	r = in->ops.dpi->connect(in, dssdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void tpo_td043_disconnect(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return;
+
+	in->ops.dpi->disconnect(in, dssdev);
+}
+
+static int tpo_td043_enable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+	int r;
+
+	if (!omapdss_device_is_connected(dssdev))
+		return -ENODEV;
+
+	if (omapdss_device_is_enabled(dssdev))
+		return 0;
+
+	in->ops.dpi->set_data_lines(in, ddata->data_lines);
+	in->ops.dpi->set_timings(in, &ddata->videomode);
+
+	r = in->ops.dpi->enable(in);
+	if (r)
+		return r;
+
+	/*
+	 * If we are resuming from system suspend, SPI clocks might not be
+	 * enabled yet, so we'll program the LCD from SPI PM resume callback.
+	 */
+	if (!ddata->spi_suspended) {
+		r = tpo_td043_power_on(ddata);
+		if (r) {
+			in->ops.dpi->disable(in);
+			return r;
+		}
+	}
+
+	dssdev->state = OMAP_DSS_DISPLAY_ACTIVE;
+
+	return 0;
+}
+
+static void tpo_td043_disable(struct omap_dss_device *dssdev)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	if (!omapdss_device_is_enabled(dssdev))
+		return;
+
+	in->ops.dpi->disable(in);
+
+	if (!ddata->spi_suspended)
+		tpo_td043_power_off(ddata);
+
+	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
+}
+
+static void tpo_td043_set_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	ddata->videomode = *timings;
+	dssdev->panel.timings = *timings;
+
+	in->ops.dpi->set_timings(in, timings);
+}
+
+static void tpo_td043_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+
+	*timings = ddata->videomode;
+}
+
+static int tpo_td043_check_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	struct panel_drv_data *ddata = to_panel_data(dssdev);
+	struct omap_dss_device *in = ddata->in;
+
+	return in->ops.dpi->check_timings(in, timings);
+}
+
+static struct omap_dss_driver tpo_td043_ops = {
+	.connect	= tpo_td043_connect,
+	.disconnect	= tpo_td043_disconnect,
+
+	.enable		= tpo_td043_enable,
+	.disable	= tpo_td043_disable,
+
+	.set_timings	= tpo_td043_set_timings,
+	.get_timings	= tpo_td043_get_timings,
+	.check_timings	= tpo_td043_check_timings,
+
+	.set_mirror	= tpo_td043_set_hmirror,
+	.get_mirror	= tpo_td043_get_hmirror,
+
+	.get_resolution	= omapdss_default_get_resolution,
+};
+
+
+static int tpo_td043_probe_pdata(struct spi_device *spi)
+{
+	const struct panel_tpo_td043mtea1_platform_data *pdata;
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev, *in;
+
+	pdata = dev_get_platdata(&spi->dev);
+
+	ddata->nreset_gpio = pdata->nreset_gpio;
+
+	in = omap_dss_find_output(pdata->source);
+	if (in == NULL) {
+		dev_err(&spi->dev, "failed to find video source '%s'\n",
+				pdata->source);
+		return -EPROBE_DEFER;
+	}
+	ddata->in = in;
+
+	ddata->data_lines = pdata->data_lines;
+
+	dssdev = &ddata->dssdev;
+	dssdev->name = pdata->name;
+
+	return 0;
+}
+
+static int tpo_td043_probe(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata;
+	struct omap_dss_device *dssdev;
+	int r;
+
+	dev_dbg(&spi->dev, "%s\n", __func__);
+
+	spi->bits_per_word = 16;
+	spi->mode = SPI_MODE_0;
+
+	r = spi_setup(spi);
+	if (r < 0) {
+		dev_err(&spi->dev, "spi_setup failed: %d\n", r);
+		return r;
+	}
+
+	ddata = devm_kzalloc(&spi->dev, sizeof(*ddata), GFP_KERNEL);
+	if (ddata == NULL)
+		return -ENOMEM;
+
+	dev_set_drvdata(&spi->dev, ddata);
+
+	ddata->spi = spi;
+
+	if (dev_get_platdata(&spi->dev)) {
+		r = tpo_td043_probe_pdata(spi);
+		if (r)
+			return r;
+	} else {
+		return -ENODEV;
+	}
+
+	ddata->mode = TPO_R02_MODE_800x480;
+	memcpy(ddata->gamma, tpo_td043_def_gamma, sizeof(ddata->gamma));
+
+	ddata->vcc_reg = devm_regulator_get(&spi->dev, "vcc");
+	if (IS_ERR(ddata->vcc_reg)) {
+		dev_err(&spi->dev, "failed to get LCD VCC regulator\n");
+		r = PTR_ERR(ddata->vcc_reg);
+		goto err_regulator;
+	}
+
+	if (gpio_is_valid(ddata->nreset_gpio)) {
+		r = devm_gpio_request_one(&spi->dev,
+				ddata->nreset_gpio, GPIOF_OUT_INIT_LOW,
+				"lcd reset");
+		if (r < 0) {
+			dev_err(&spi->dev, "couldn't request reset GPIO\n");
+			goto err_gpio_req;
+		}
+	}
+
+	r = sysfs_create_group(&spi->dev.kobj, &tpo_td043_attr_group);
+	if (r) {
+		dev_err(&spi->dev, "failed to create sysfs files\n");
+		goto err_sysfs;
+	}
+
+	ddata->videomode = tpo_td043_timings;
+
+	dssdev = &ddata->dssdev;
+	dssdev->dev = &spi->dev;
+	dssdev->driver = &tpo_td043_ops;
+	dssdev->type = OMAP_DISPLAY_TYPE_DPI;
+	dssdev->owner = THIS_MODULE;
+	dssdev->panel.timings = ddata->videomode;
+
+	r = omapdss_register_display(dssdev);
+	if (r) {
+		dev_err(&spi->dev, "Failed to register panel\n");
+		goto err_reg;
+	}
+
+	return 0;
+
+err_reg:
+	sysfs_remove_group(&spi->dev.kobj, &tpo_td043_attr_group);
+err_sysfs:
+err_gpio_req:
+err_regulator:
+	omap_dss_put_device(ddata->in);
+	return r;
+}
+
+static int tpo_td043_remove(struct spi_device *spi)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(&spi->dev);
+	struct omap_dss_device *dssdev = &ddata->dssdev;
+	struct omap_dss_device *in = ddata->in;
+
+	dev_dbg(&ddata->spi->dev, "%s\n", __func__);
+
+	omapdss_unregister_display(dssdev);
+
+	tpo_td043_disable(dssdev);
+	tpo_td043_disconnect(dssdev);
+
+	omap_dss_put_device(in);
+
+	sysfs_remove_group(&spi->dev.kobj, &tpo_td043_attr_group);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int tpo_td043_spi_suspend(struct device *dev)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+
+	dev_dbg(dev, "tpo_td043_spi_suspend, tpo %p\n", ddata);
+
+	ddata->power_on_resume = ddata->powered_on;
+	tpo_td043_power_off(ddata);
+	ddata->spi_suspended = 1;
+
+	return 0;
+}
+
+static int tpo_td043_spi_resume(struct device *dev)
+{
+	struct panel_drv_data *ddata = dev_get_drvdata(dev);
+	int ret;
+
+	dev_dbg(dev, "tpo_td043_spi_resume\n");
+
+	if (ddata->power_on_resume) {
+		ret = tpo_td043_power_on(ddata);
+		if (ret)
+			return ret;
+	}
+	ddata->spi_suspended = 0;
+
+	return 0;
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(tpo_td043_spi_pm,
+	tpo_td043_spi_suspend, tpo_td043_spi_resume);
+
+static struct spi_driver tpo_td043_spi_driver = {
+	.driver = {
+		.name	= "panel-tpo-td043mtea1",
+		.owner	= THIS_MODULE,
+		.pm	= &tpo_td043_spi_pm,
+	},
+	.probe	= tpo_td043_probe,
+	.remove	= tpo_td043_remove,
+};
+
+module_spi_driver(tpo_td043_spi_driver);
+
+MODULE_AUTHOR("Gražvydas Ignotas <notasas@gmail.com>");
+MODULE_DESCRIPTION("TPO TD043MTEA1 LCD Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/video/omap2/displays/Kconfig b/drivers/video/omap2/displays/Kconfig
index c3853c9..e80ac1c 100644
--- a/drivers/video/omap2/displays/Kconfig
+++ b/drivers/video/omap2/displays/Kconfig
@@ -1,4 +1,4 @@
-menu "OMAP2/3 Display Device Drivers"
+menu "OMAP2/3 Display Device Drivers (old device model)"
         depends on OMAP2_DSS
 
 config PANEL_GENERIC_DPI
diff --git a/drivers/video/omap2/displays/panel-acx565akm.c b/drivers/video/omap2/displays/panel-acx565akm.c
index d7f69c0..3fd100f 100644
--- a/drivers/video/omap2/displays/panel-acx565akm.c
+++ b/drivers/video/omap2/displays/panel-acx565akm.c
@@ -510,7 +510,7 @@ static int acx_panel_probe(struct omap_dss_device *dssdev)
 	int max_brightness, brightness;
 	struct backlight_properties props;
 
-	dev_dbg(&dssdev->dev, "%s\n", __func__);
+	dev_dbg(dssdev->dev, "%s\n", __func__);
 
 	if (!panel_data)
 		return -EINVAL;
@@ -519,7 +519,7 @@ static int acx_panel_probe(struct omap_dss_device *dssdev)
 	dssdev->panel.timings = acx_panel_timings;
 
 	if (gpio_is_valid(panel_data->reset_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, panel_data->reset_gpio,
+		r = devm_gpio_request_one(dssdev->dev, panel_data->reset_gpio,
 				GPIOF_OUT_INIT_LOW, "lcd reset");
 		if (r)
 			return r;
@@ -538,7 +538,7 @@ static int acx_panel_probe(struct omap_dss_device *dssdev)
 
 	r = panel_detect(md);
 	if (r) {
-		dev_err(&dssdev->dev, "%s panel detect error\n", __func__);
+		dev_err(dssdev->dev, "%s panel detect error\n", __func__);
 		if (!md->enabled && gpio_is_valid(panel_data->reset_gpio))
 			gpio_set_value(panel_data->reset_gpio, 0);
 
@@ -593,7 +593,7 @@ static void acx_panel_remove(struct omap_dss_device *dssdev)
 {
 	struct acx565akm_device *md = &acx_dev;
 
-	dev_dbg(&dssdev->dev, "%s\n", __func__);
+	dev_dbg(dssdev->dev, "%s\n", __func__);
 	sysfs_remove_group(&md->bl_dev->dev.kobj, &bldev_attr_group);
 	backlight_device_unregister(md->bl_dev);
 	mutex_lock(&acx_dev.mutex);
@@ -607,7 +607,7 @@ static int acx_panel_power_on(struct omap_dss_device *dssdev)
 	struct panel_acx565akm_data *panel_data = get_panel_data(dssdev);
 	int r;
 
-	dev_dbg(&dssdev->dev, "%s\n", __func__);
+	dev_dbg(dssdev->dev, "%s\n", __func__);
 
 	if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE)
 		return 0;
@@ -667,7 +667,7 @@ static void acx_panel_power_off(struct omap_dss_device *dssdev)
 	struct acx565akm_device *md = &acx_dev;
 	struct panel_acx565akm_data *panel_data = get_panel_data(dssdev);
 
-	dev_dbg(&dssdev->dev, "%s\n", __func__);
+	dev_dbg(dssdev->dev, "%s\n", __func__);
 
 	if (dssdev->state != OMAP_DSS_DISPLAY_ACTIVE)
 		return;
@@ -704,7 +704,7 @@ static int acx_panel_enable(struct omap_dss_device *dssdev)
 {
 	int r;
 
-	dev_dbg(&dssdev->dev, "%s\n", __func__);
+	dev_dbg(dssdev->dev, "%s\n", __func__);
 	r = acx_panel_power_on(dssdev);
 
 	if (r)
@@ -716,7 +716,7 @@ static int acx_panel_enable(struct omap_dss_device *dssdev)
 
 static void acx_panel_disable(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "%s\n", __func__);
+	dev_dbg(dssdev->dev, "%s\n", __func__);
 	acx_panel_power_off(dssdev);
 	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
 }
diff --git a/drivers/video/omap2/displays/panel-generic-dpi.c b/drivers/video/omap2/displays/panel-generic-dpi.c
index 97363f7..bebebd4 100644
--- a/drivers/video/omap2/displays/panel-generic-dpi.c
+++ b/drivers/video/omap2/displays/panel-generic-dpi.c
@@ -536,7 +536,7 @@ static int generic_dpi_panel_power_on(struct omap_dss_device *dssdev)
 {
 	int r, i;
 	struct panel_generic_dpi_data *panel_data = get_panel_data(dssdev);
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 	struct panel_config *panel_config = drv_data->panel_config;
 
 	if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE)
@@ -567,7 +567,7 @@ err0:
 static void generic_dpi_panel_power_off(struct omap_dss_device *dssdev)
 {
 	struct panel_generic_dpi_data *panel_data = get_panel_data(dssdev);
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 	struct panel_config *panel_config = drv_data->panel_config;
 	int i;
 
@@ -593,7 +593,7 @@ static int generic_dpi_panel_probe(struct omap_dss_device *dssdev)
 	struct panel_drv_data *drv_data = NULL;
 	int i, r;
 
-	dev_dbg(&dssdev->dev, "probe\n");
+	dev_dbg(dssdev->dev, "probe\n");
 
 	if (!panel_data || !panel_data->name)
 		return -EINVAL;
@@ -609,7 +609,7 @@ static int generic_dpi_panel_probe(struct omap_dss_device *dssdev)
 		return -EINVAL;
 
 	for (i = 0; i < panel_data->num_gpios; ++i) {
-		r = devm_gpio_request_one(&dssdev->dev, panel_data->gpios[i],
+		r = devm_gpio_request_one(dssdev->dev, panel_data->gpios[i],
 				panel_data->gpio_invert[i] ?
 				GPIOF_OUT_INIT_HIGH : GPIOF_OUT_INIT_LOW,
 				"panel gpio");
@@ -619,7 +619,7 @@ static int generic_dpi_panel_probe(struct omap_dss_device *dssdev)
 
 	dssdev->panel.timings = panel_config->timings;
 
-	drv_data = devm_kzalloc(&dssdev->dev, sizeof(*drv_data), GFP_KERNEL);
+	drv_data = devm_kzalloc(dssdev->dev, sizeof(*drv_data), GFP_KERNEL);
 	if (!drv_data)
 		return -ENOMEM;
 
@@ -628,21 +628,21 @@ static int generic_dpi_panel_probe(struct omap_dss_device *dssdev)
 
 	mutex_init(&drv_data->lock);
 
-	dev_set_drvdata(&dssdev->dev, drv_data);
+	dev_set_drvdata(dssdev->dev, drv_data);
 
 	return 0;
 }
 
 static void __exit generic_dpi_panel_remove(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "remove\n");
+	dev_dbg(dssdev->dev, "remove\n");
 
-	dev_set_drvdata(&dssdev->dev, NULL);
+	dev_set_drvdata(dssdev->dev, NULL);
 }
 
 static int generic_dpi_panel_enable(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&drv_data->lock);
@@ -660,7 +660,7 @@ err:
 
 static void generic_dpi_panel_disable(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&drv_data->lock);
 
@@ -674,7 +674,7 @@ static void generic_dpi_panel_disable(struct omap_dss_device *dssdev)
 static void generic_dpi_panel_set_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&drv_data->lock);
 
@@ -688,7 +688,7 @@ static void generic_dpi_panel_set_timings(struct omap_dss_device *dssdev,
 static void generic_dpi_panel_get_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&drv_data->lock);
 
@@ -700,7 +700,7 @@ static void generic_dpi_panel_get_timings(struct omap_dss_device *dssdev,
 static int generic_dpi_panel_check_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	struct panel_drv_data *drv_data = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *drv_data = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&drv_data->lock);
diff --git a/drivers/video/omap2/displays/panel-lgphilips-lb035q02.c b/drivers/video/omap2/displays/panel-lgphilips-lb035q02.c
index 4ea6548..6c51430 100644
--- a/drivers/video/omap2/displays/panel-lgphilips-lb035q02.c
+++ b/drivers/video/omap2/displays/panel-lgphilips-lb035q02.c
@@ -109,12 +109,12 @@ static int lb035q02_panel_probe(struct omap_dss_device *dssdev)
 
 	dssdev->panel.timings = lb035q02_timings;
 
-	ld = devm_kzalloc(&dssdev->dev, sizeof(*ld), GFP_KERNEL);
+	ld = devm_kzalloc(dssdev->dev, sizeof(*ld), GFP_KERNEL);
 	if (!ld)
 		return -ENOMEM;
 
 	for (i = 0; i < panel_data->num_gpios; ++i) {
-		r = devm_gpio_request_one(&dssdev->dev, panel_data->gpios[i],
+		r = devm_gpio_request_one(dssdev->dev, panel_data->gpios[i],
 				panel_data->gpio_invert[i] ?
 				GPIOF_OUT_INIT_HIGH : GPIOF_OUT_INIT_LOW,
 				"panel gpio");
@@ -123,7 +123,7 @@ static int lb035q02_panel_probe(struct omap_dss_device *dssdev)
 	}
 
 	mutex_init(&ld->lock);
-	dev_set_drvdata(&dssdev->dev, ld);
+	dev_set_drvdata(dssdev->dev, ld);
 
 	return 0;
 }
@@ -134,7 +134,7 @@ static void lb035q02_panel_remove(struct omap_dss_device *dssdev)
 
 static int lb035q02_panel_enable(struct omap_dss_device *dssdev)
 {
-	struct lb035q02_data *ld = dev_get_drvdata(&dssdev->dev);
+	struct lb035q02_data *ld = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&ld->lock);
@@ -153,7 +153,7 @@ err:
 
 static void lb035q02_panel_disable(struct omap_dss_device *dssdev)
 {
-	struct lb035q02_data *ld = dev_get_drvdata(&dssdev->dev);
+	struct lb035q02_data *ld = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&ld->lock);
 
diff --git a/drivers/video/omap2/displays/panel-n8x0.c b/drivers/video/omap2/displays/panel-n8x0.c
index 860b180..1d525fc 100644
--- a/drivers/video/omap2/displays/panel-n8x0.c
+++ b/drivers/video/omap2/displays/panel-n8x0.c
@@ -311,16 +311,16 @@ static int n8x0_panel_power_on(struct omap_dss_device *dssdev)
 	switch (rev & 0xfc) {
 	case 0x9c:
 		ddata->blizzard_ver = BLIZZARD_VERSION_S1D13744;
-		dev_info(&dssdev->dev, "s1d13744 LCD controller rev %d "
+		dev_info(dssdev->dev, "s1d13744 LCD controller rev %d "
 			"initialized (CNF pins %x)\n", rev & 0x03, conf & 0x07);
 		break;
 	case 0xa4:
 		ddata->blizzard_ver = BLIZZARD_VERSION_S1D13745;
-		dev_info(&dssdev->dev, "s1d13745 LCD controller rev %d "
+		dev_info(dssdev->dev, "s1d13745 LCD controller rev %d "
 			"initialized (CNF pins %x)\n", rev & 0x03, conf & 0x07);
 		break;
 	default:
-		dev_err(&dssdev->dev, "invalid s1d1374x revision %02x\n", rev);
+		dev_err(dssdev->dev, "invalid s1d1374x revision %02x\n", rev);
 		r = -ENODEV;
 		goto err_inv_chip;
 	}
@@ -341,13 +341,13 @@ static int n8x0_panel_power_on(struct omap_dss_device *dssdev)
 		panel_name = "ls041y3";
 		break;
 	default:
-		dev_err(&dssdev->dev, "invalid display ID 0x%x\n",
+		dev_err(dssdev->dev, "invalid display ID 0x%x\n",
 				display_id[0]);
 		r = -ENODEV;
 		goto err_inv_panel;
 	}
 
-	dev_info(&dssdev->dev, "%s rev %02x LCD detected\n",
+	dev_info(dssdev->dev, "%s rev %02x LCD detected\n",
 			panel_name, display_id[1]);
 
 	send_sleep_out(spi);
@@ -416,7 +416,7 @@ static int n8x0_panel_probe(struct omap_dss_device *dssdev)
 	struct panel_drv_data *ddata;
 	int r;
 
-	dev_dbg(&dssdev->dev, "probe\n");
+	dev_dbg(dssdev->dev, "probe\n");
 
 	if (!bdata)
 		return -EINVAL;
@@ -434,14 +434,14 @@ static int n8x0_panel_probe(struct omap_dss_device *dssdev)
 	dssdev->caps = OMAP_DSS_DISPLAY_CAP_MANUAL_UPDATE;
 
 	if (gpio_is_valid(bdata->panel_reset)) {
-		r = devm_gpio_request_one(&dssdev->dev, bdata->panel_reset,
+		r = devm_gpio_request_one(dssdev->dev, bdata->panel_reset,
 				GPIOF_OUT_INIT_LOW, "PANEL RESET");
 		if (r)
 			return r;
 	}
 
 	if (gpio_is_valid(bdata->ctrl_pwrdown)) {
-		r = devm_gpio_request_one(&dssdev->dev, bdata->ctrl_pwrdown,
+		r = devm_gpio_request_one(dssdev->dev, bdata->ctrl_pwrdown,
 				GPIOF_OUT_INIT_LOW, "PANEL PWRDOWN");
 		if (r)
 			return r;
@@ -452,9 +452,9 @@ static int n8x0_panel_probe(struct omap_dss_device *dssdev)
 
 static void n8x0_panel_remove(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "remove\n");
+	dev_dbg(dssdev->dev, "remove\n");
 
-	dev_set_drvdata(&dssdev->dev, NULL);
+	dev_set_drvdata(dssdev->dev, NULL);
 }
 
 static int n8x0_panel_enable(struct omap_dss_device *dssdev)
@@ -462,7 +462,7 @@ static int n8x0_panel_enable(struct omap_dss_device *dssdev)
 	struct panel_drv_data *ddata = get_drv_data(dssdev);
 	int r;
 
-	dev_dbg(&dssdev->dev, "enable\n");
+	dev_dbg(dssdev->dev, "enable\n");
 
 	mutex_lock(&ddata->lock);
 
@@ -488,7 +488,7 @@ static void n8x0_panel_disable(struct omap_dss_device *dssdev)
 {
 	struct panel_drv_data *ddata = get_drv_data(dssdev);
 
-	dev_dbg(&dssdev->dev, "disable\n");
+	dev_dbg(dssdev->dev, "disable\n");
 
 	mutex_lock(&ddata->lock);
 
@@ -521,13 +521,13 @@ static int n8x0_panel_update(struct omap_dss_device *dssdev,
 	struct panel_drv_data *ddata = get_drv_data(dssdev);
 	u16 dw, dh;
 
-	dev_dbg(&dssdev->dev, "update\n");
+	dev_dbg(dssdev->dev, "update\n");
 
 	dw = dssdev->panel.timings.x_res;
 	dh = dssdev->panel.timings.y_res;
 
 	if (x != 0 || y != 0 || w != dw || h != dh) {
-		dev_err(&dssdev->dev, "invalid update region %d, %d, %d, %d\n",
+		dev_err(dssdev->dev, "invalid update region %d, %d, %d, %d\n",
 			x, y, w, h);
 		return -EINVAL;
 	}
@@ -548,7 +548,7 @@ static int n8x0_panel_sync(struct omap_dss_device *dssdev)
 {
 	struct panel_drv_data *ddata = get_drv_data(dssdev);
 
-	dev_dbg(&dssdev->dev, "sync\n");
+	dev_dbg(dssdev->dev, "sync\n");
 
 	mutex_lock(&ddata->lock);
 	rfbi_bus_lock();
diff --git a/drivers/video/omap2/displays/panel-nec-nl8048hl11-01b.c b/drivers/video/omap2/displays/panel-nec-nl8048hl11-01b.c
index 20c3cd9..6b9f792 100644
--- a/drivers/video/omap2/displays/panel-nec-nl8048hl11-01b.c
+++ b/drivers/video/omap2/displays/panel-nec-nl8048hl11-01b.c
@@ -98,14 +98,14 @@ static int nec_8048_panel_probe(struct omap_dss_device *dssdev)
 	dssdev->panel.timings = nec_8048_panel_timings;
 
 	if (gpio_is_valid(pd->qvga_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->qvga_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->qvga_gpio,
 				GPIOF_OUT_INIT_HIGH, "lcd QVGA");
 		if (r)
 			return r;
 	}
 
 	if (gpio_is_valid(pd->res_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->res_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->res_gpio,
 				GPIOF_OUT_INIT_LOW, "lcd RES");
 		if (r)
 			return r;
diff --git a/drivers/video/omap2/displays/panel-picodlp.c b/drivers/video/omap2/displays/panel-picodlp.c
index 62f2db0..153e9be 100644
--- a/drivers/video/omap2/displays/panel-picodlp.c
+++ b/drivers/video/omap2/displays/panel-picodlp.c
@@ -351,7 +351,7 @@ static struct i2c_driver picodlp_i2c_driver = {
 static int picodlp_panel_power_on(struct omap_dss_device *dssdev)
 {
 	int r, trial = 100;
-	struct picodlp_data *picod = dev_get_drvdata(&dssdev->dev);
+	struct picodlp_data *picod = dev_get_drvdata(dssdev->dev);
 	struct picodlp_panel_data *picodlp_pdata = get_panel_data(dssdev);
 
 	gpio_set_value(picodlp_pdata->pwrgood_gpio, 0);
@@ -360,7 +360,7 @@ static int picodlp_panel_power_on(struct omap_dss_device *dssdev)
 
 	while (!gpio_get_value(picodlp_pdata->emu_done_gpio)) {
 		if (!trial--) {
-			dev_err(&dssdev->dev, "emu_done signal not"
+			dev_err(dssdev->dev, "emu_done signal not"
 						" going high\n");
 			return -ETIMEDOUT;
 		}
@@ -378,7 +378,7 @@ static int picodlp_panel_power_on(struct omap_dss_device *dssdev)
 
 	r = omapdss_dpi_display_enable(dssdev);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to enable DPI\n");
+		dev_err(dssdev->dev, "failed to enable DPI\n");
 		goto err1;
 	}
 
@@ -418,7 +418,7 @@ static int picodlp_panel_probe(struct omap_dss_device *dssdev)
 	if (!picodlp_pdata)
 		return -EINVAL;
 
-	picod = devm_kzalloc(&dssdev->dev, sizeof(*picod), GFP_KERNEL);
+	picod = devm_kzalloc(dssdev->dev, sizeof(*picod), GFP_KERNEL);
 	if (!picod)
 		return -ENOMEM;
 
@@ -428,23 +428,23 @@ static int picodlp_panel_probe(struct omap_dss_device *dssdev)
 
 	adapter = i2c_get_adapter(picodlp_adapter_id);
 	if (!adapter) {
-		dev_err(&dssdev->dev, "can't get i2c adapter\n");
+		dev_err(dssdev->dev, "can't get i2c adapter\n");
 		return -ENODEV;
 	}
 
 	picodlp_i2c_client = i2c_new_device(adapter, &picodlp_i2c_board_info);
 	if (!picodlp_i2c_client) {
-		dev_err(&dssdev->dev, "can't add i2c device::"
+		dev_err(dssdev->dev, "can't add i2c device::"
 					 " picodlp_i2c_client is NULL\n");
 		return -ENODEV;
 	}
 
 	picod->picodlp_i2c_client = picodlp_i2c_client;
 
-	dev_set_drvdata(&dssdev->dev, picod);
+	dev_set_drvdata(dssdev->dev, picod);
 
 	if (gpio_is_valid(picodlp_pdata->emu_done_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev,
+		r = devm_gpio_request_one(dssdev->dev,
 				picodlp_pdata->emu_done_gpio,
 				GPIOF_IN, "DLP EMU DONE");
 		if (r)
@@ -452,7 +452,7 @@ static int picodlp_panel_probe(struct omap_dss_device *dssdev)
 	}
 
 	if (gpio_is_valid(picodlp_pdata->pwrgood_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev,
+		r = devm_gpio_request_one(dssdev->dev,
 				picodlp_pdata->pwrgood_gpio,
 				GPIOF_OUT_INIT_LOW, "DLP PWRGOOD");
 		if (r)
@@ -464,21 +464,19 @@ static int picodlp_panel_probe(struct omap_dss_device *dssdev)
 
 static void picodlp_panel_remove(struct omap_dss_device *dssdev)
 {
-	struct picodlp_data *picod = dev_get_drvdata(&dssdev->dev);
+	struct picodlp_data *picod = dev_get_drvdata(dssdev->dev);
 
 	i2c_unregister_device(picod->picodlp_i2c_client);
-	dev_set_drvdata(&dssdev->dev, NULL);
-	dev_dbg(&dssdev->dev, "removing picodlp panel\n");
-
-	kfree(picod);
+	dev_set_drvdata(dssdev->dev, NULL);
+	dev_dbg(dssdev->dev, "removing picodlp panel\n");
 }
 
 static int picodlp_panel_enable(struct omap_dss_device *dssdev)
 {
-	struct picodlp_data *picod = dev_get_drvdata(&dssdev->dev);
+	struct picodlp_data *picod = dev_get_drvdata(dssdev->dev);
 	int r;
 
-	dev_dbg(&dssdev->dev, "enabling picodlp panel\n");
+	dev_dbg(dssdev->dev, "enabling picodlp panel\n");
 
 	mutex_lock(&picod->lock);
 	if (dssdev->state != OMAP_DSS_DISPLAY_DISABLED) {
@@ -494,7 +492,7 @@ static int picodlp_panel_enable(struct omap_dss_device *dssdev)
 
 static void picodlp_panel_disable(struct omap_dss_device *dssdev)
 {
-	struct picodlp_data *picod = dev_get_drvdata(&dssdev->dev);
+	struct picodlp_data *picod = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&picod->lock);
 	/* Turn off DLP Power */
@@ -504,7 +502,7 @@ static void picodlp_panel_disable(struct omap_dss_device *dssdev)
 	dssdev->state = OMAP_DSS_DISPLAY_DISABLED;
 	mutex_unlock(&picod->lock);
 
-	dev_dbg(&dssdev->dev, "disabling picodlp panel\n");
+	dev_dbg(dssdev->dev, "disabling picodlp panel\n");
 }
 
 static void picodlp_get_resolution(struct omap_dss_device *dssdev,
diff --git a/drivers/video/omap2/displays/panel-sharp-ls037v7dw01.c b/drivers/video/omap2/displays/panel-sharp-ls037v7dw01.c
index 74cb0eb..78f0a67 100644
--- a/drivers/video/omap2/displays/panel-sharp-ls037v7dw01.c
+++ b/drivers/video/omap2/displays/panel-sharp-ls037v7dw01.c
@@ -66,35 +66,35 @@ static int sharp_ls_panel_probe(struct omap_dss_device *dssdev)
 	dssdev->panel.timings = sharp_ls_timings;
 
 	if (gpio_is_valid(pd->mo_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->mo_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->mo_gpio,
 				GPIOF_OUT_INIT_LOW, "lcd MO");
 		if (r)
 			return r;
 	}
 
 	if (gpio_is_valid(pd->lr_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->lr_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->lr_gpio,
 				GPIOF_OUT_INIT_HIGH, "lcd LR");
 		if (r)
 			return r;
 	}
 
 	if (gpio_is_valid(pd->ud_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->ud_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->ud_gpio,
 				GPIOF_OUT_INIT_HIGH, "lcd UD");
 		if (r)
 			return r;
 	}
 
 	if (gpio_is_valid(pd->resb_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->resb_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->resb_gpio,
 				GPIOF_OUT_INIT_LOW, "lcd RESB");
 		if (r)
 			return r;
 	}
 
 	if (gpio_is_valid(pd->ini_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, pd->ini_gpio,
+		r = devm_gpio_request_one(dssdev->dev, pd->ini_gpio,
 				GPIOF_OUT_INIT_LOW, "lcd INI");
 		if (r)
 			return r;
diff --git a/drivers/video/omap2/displays/panel-taal.c b/drivers/video/omap2/displays/panel-taal.c
index c4f78bd..54a07da 100644
--- a/drivers/video/omap2/displays/panel-taal.c
+++ b/drivers/video/omap2/displays/panel-taal.c
@@ -237,7 +237,7 @@ static int taal_set_update_window(struct taal_data *td,
 
 static void taal_queue_esd_work(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	if (td->esd_interval > 0)
 		queue_delayed_work(td->workqueue, &td->esd_work,
@@ -246,14 +246,14 @@ static void taal_queue_esd_work(struct omap_dss_device *dssdev)
 
 static void taal_cancel_esd_work(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	cancel_delayed_work(&td->esd_work);
 }
 
 static void taal_queue_ulps_work(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	if (td->ulps_timeout > 0)
 		queue_delayed_work(td->workqueue, &td->ulps_work,
@@ -262,14 +262,14 @@ static void taal_queue_ulps_work(struct omap_dss_device *dssdev)
 
 static void taal_cancel_ulps_work(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	cancel_delayed_work(&td->ulps_work);
 }
 
 static int taal_enter_ulps(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	if (td->ulps_enabled)
@@ -291,7 +291,7 @@ static int taal_enter_ulps(struct omap_dss_device *dssdev)
 	return 0;
 
 err:
-	dev_err(&dssdev->dev, "enter ULPS failed");
+	dev_err(dssdev->dev, "enter ULPS failed");
 	taal_panel_reset(dssdev);
 
 	td->ulps_enabled = false;
@@ -303,7 +303,7 @@ err:
 
 static int taal_exit_ulps(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	if (!td->ulps_enabled)
@@ -311,7 +311,7 @@ static int taal_exit_ulps(struct omap_dss_device *dssdev)
 
 	r = omapdss_dsi_display_enable(dssdev);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to enable DSI\n");
+		dev_err(dssdev->dev, "failed to enable DSI\n");
 		goto err1;
 	}
 
@@ -319,7 +319,7 @@ static int taal_exit_ulps(struct omap_dss_device *dssdev)
 
 	r = _taal_enable_te(dssdev, true);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to re-enable TE");
+		dev_err(dssdev->dev, "failed to re-enable TE");
 		goto err2;
 	}
 
@@ -333,7 +333,7 @@ static int taal_exit_ulps(struct omap_dss_device *dssdev)
 	return 0;
 
 err2:
-	dev_err(&dssdev->dev, "failed to exit ULPS");
+	dev_err(dssdev->dev, "failed to exit ULPS");
 
 	r = taal_panel_reset(dssdev);
 	if (!r) {
@@ -349,7 +349,7 @@ err1:
 
 static int taal_wake_up(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	if (td->ulps_enabled)
 		return taal_exit_ulps(dssdev);
@@ -362,7 +362,7 @@ static int taal_wake_up(struct omap_dss_device *dssdev)
 static int taal_bl_update_status(struct backlight_device *dev)
 {
 	struct omap_dss_device *dssdev = dev_get_drvdata(&dev->dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 	int level;
 
@@ -372,7 +372,7 @@ static int taal_bl_update_status(struct backlight_device *dev)
 	else
 		level = 0;
 
-	dev_dbg(&dssdev->dev, "update brightness to %d\n", level);
+	dev_dbg(dssdev->dev, "update brightness to %d\n", level);
 
 	mutex_lock(&td->lock);
 
@@ -418,7 +418,7 @@ static ssize_t taal_num_errors_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	u8 errors = 0;
 	int r;
 
@@ -448,7 +448,7 @@ static ssize_t taal_hw_revision_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	u8 id1, id2, id3;
 	int r;
 
@@ -486,7 +486,7 @@ static ssize_t show_cabc_mode(struct device *dev,
 		char *buf)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	const char *mode_str;
 	int mode;
 	int len;
@@ -506,7 +506,7 @@ static ssize_t store_cabc_mode(struct device *dev,
 		const char *buf, size_t count)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int i;
 	int r;
 
@@ -568,12 +568,12 @@ static ssize_t taal_store_esd_interval(struct device *dev,
 		const char *buf, size_t count)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	unsigned long t;
 	int r;
 
-	r = strict_strtoul(buf, 10, &t);
+	r = kstrtoul(buf, 10, &t);
 	if (r)
 		return r;
 
@@ -592,7 +592,7 @@ static ssize_t taal_show_esd_interval(struct device *dev,
 		char *buf)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	unsigned t;
 
 	mutex_lock(&td->lock);
@@ -607,11 +607,11 @@ static ssize_t taal_store_ulps(struct device *dev,
 		const char *buf, size_t count)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	unsigned long t;
 	int r;
 
-	r = strict_strtoul(buf, 10, &t);
+	r = kstrtoul(buf, 10, &t);
 	if (r)
 		return r;
 
@@ -641,7 +641,7 @@ static ssize_t taal_show_ulps(struct device *dev,
 		char *buf)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	unsigned t;
 
 	mutex_lock(&td->lock);
@@ -656,11 +656,11 @@ static ssize_t taal_store_ulps_timeout(struct device *dev,
 		const char *buf, size_t count)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	unsigned long t;
 	int r;
 
-	r = strict_strtoul(buf, 10, &t);
+	r = kstrtoul(buf, 10, &t);
 	if (r)
 		return r;
 
@@ -687,7 +687,7 @@ static ssize_t taal_show_ulps_timeout(struct device *dev,
 		char *buf)
 {
 	struct omap_dss_device *dssdev = to_dss_device(dev);
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	unsigned t;
 
 	mutex_lock(&td->lock);
@@ -727,7 +727,7 @@ static struct attribute_group taal_attr_group = {
 
 static void taal_hw_reset(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	if (!gpio_is_valid(td->reset_gpio))
 		return;
@@ -768,13 +768,13 @@ static int taal_probe(struct omap_dss_device *dssdev)
 	struct backlight_device *bldev = NULL;
 	int r;
 
-	dev_dbg(&dssdev->dev, "probe\n");
+	dev_dbg(dssdev->dev, "probe\n");
 
-	td = devm_kzalloc(&dssdev->dev, sizeof(*td), GFP_KERNEL);
+	td = devm_kzalloc(dssdev->dev, sizeof(*td), GFP_KERNEL);
 	if (!td)
 		return -ENOMEM;
 
-	dev_set_drvdata(&dssdev->dev, td);
+	dev_set_drvdata(dssdev->dev, td);
 	td->dssdev = dssdev;
 
 	if (dssdev->data) {
@@ -797,41 +797,41 @@ static int taal_probe(struct omap_dss_device *dssdev)
 	atomic_set(&td->do_update, 0);
 
 	if (gpio_is_valid(td->reset_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, td->reset_gpio,
+		r = devm_gpio_request_one(dssdev->dev, td->reset_gpio,
 				GPIOF_OUT_INIT_LOW, "taal rst");
 		if (r) {
-			dev_err(&dssdev->dev, "failed to request reset gpio\n");
+			dev_err(dssdev->dev, "failed to request reset gpio\n");
 			return r;
 		}
 	}
 
 	if (gpio_is_valid(td->ext_te_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, td->ext_te_gpio,
+		r = devm_gpio_request_one(dssdev->dev, td->ext_te_gpio,
 				GPIOF_IN, "taal irq");
 		if (r) {
-			dev_err(&dssdev->dev, "GPIO request failed\n");
+			dev_err(dssdev->dev, "GPIO request failed\n");
 			return r;
 		}
 
-		r = devm_request_irq(&dssdev->dev, gpio_to_irq(td->ext_te_gpio),
+		r = devm_request_irq(dssdev->dev, gpio_to_irq(td->ext_te_gpio),
 				taal_te_isr,
 				IRQF_TRIGGER_RISING,
 				"taal vsync", dssdev);
 
 		if (r) {
-			dev_err(&dssdev->dev, "IRQ request failed\n");
+			dev_err(dssdev->dev, "IRQ request failed\n");
 			return r;
 		}
 
 		INIT_DEFERRABLE_WORK(&td->te_timeout_work,
 					taal_te_timeout_work_callback);
 
-		dev_dbg(&dssdev->dev, "Using GPIO TE\n");
+		dev_dbg(dssdev->dev, "Using GPIO TE\n");
 	}
 
 	td->workqueue = create_singlethread_workqueue("taal_esd");
 	if (td->workqueue == NULL) {
-		dev_err(&dssdev->dev, "can't create ESD workqueue\n");
+		dev_err(dssdev->dev, "can't create ESD workqueue\n");
 		return -ENOMEM;
 	}
 	INIT_DEFERRABLE_WORK(&td->esd_work, taal_esd_work);
@@ -844,8 +844,8 @@ static int taal_probe(struct omap_dss_device *dssdev)
 		props.max_brightness = 255;
 
 		props.type = BACKLIGHT_RAW;
-		bldev = backlight_device_register(dev_name(&dssdev->dev),
-				&dssdev->dev, dssdev, &taal_bl_ops, &props);
+		bldev = backlight_device_register(dev_name(dssdev->dev),
+				dssdev->dev, dssdev, &taal_bl_ops, &props);
 		if (IS_ERR(bldev)) {
 			r = PTR_ERR(bldev);
 			goto err_bl;
@@ -862,19 +862,19 @@ static int taal_probe(struct omap_dss_device *dssdev)
 
 	r = omap_dsi_request_vc(dssdev, &td->channel);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to get virtual channel\n");
+		dev_err(dssdev->dev, "failed to get virtual channel\n");
 		goto err_req_vc;
 	}
 
 	r = omap_dsi_set_vc_id(dssdev, td->channel, TCH);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to set VC_ID\n");
+		dev_err(dssdev->dev, "failed to set VC_ID\n");
 		goto err_vc_id;
 	}
 
-	r = sysfs_create_group(&dssdev->dev.kobj, &taal_attr_group);
+	r = sysfs_create_group(&dssdev->dev->kobj, &taal_attr_group);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to create sysfs files\n");
+		dev_err(dssdev->dev, "failed to create sysfs files\n");
 		goto err_vc_id;
 	}
 
@@ -892,12 +892,12 @@ err_bl:
 
 static void __exit taal_remove(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	struct backlight_device *bldev;
 
-	dev_dbg(&dssdev->dev, "remove\n");
+	dev_dbg(dssdev->dev, "remove\n");
 
-	sysfs_remove_group(&dssdev->dev.kobj, &taal_attr_group);
+	sysfs_remove_group(&dssdev->dev->kobj, &taal_attr_group);
 	omap_dsi_release_vc(dssdev, td->channel);
 
 	bldev = td->bldev;
@@ -917,7 +917,7 @@ static void __exit taal_remove(struct omap_dss_device *dssdev)
 
 static int taal_power_on(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	u8 id1, id2, id3;
 	int r;
 	struct omap_dss_dsi_config dsi_config = {
@@ -932,19 +932,19 @@ static int taal_power_on(struct omap_dss_device *dssdev)
 
 	r = omapdss_dsi_configure_pins(dssdev, &td->pin_config);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to configure DSI pins\n");
+		dev_err(dssdev->dev, "failed to configure DSI pins\n");
 		goto err0;
 	};
 
 	r = omapdss_dsi_set_config(dssdev, &dsi_config);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to configure DSI\n");
+		dev_err(dssdev->dev, "failed to configure DSI\n");
 		goto err0;
 	}
 
 	r = omapdss_dsi_display_enable(dssdev);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to enable DSI\n");
+		dev_err(dssdev->dev, "failed to enable DSI\n");
 		goto err0;
 	}
 
@@ -999,10 +999,10 @@ static int taal_power_on(struct omap_dss_device *dssdev)
 	td->enabled = 1;
 
 	if (!td->intro_printed) {
-		dev_info(&dssdev->dev, "panel revision %02x.%02x.%02x\n",
+		dev_info(dssdev->dev, "panel revision %02x.%02x.%02x\n",
 			id1, id2, id3);
 		if (td->cabc_broken)
-			dev_info(&dssdev->dev,
+			dev_info(dssdev->dev,
 					"old Taal version, CABC disabled\n");
 		td->intro_printed = true;
 	}
@@ -1011,7 +1011,7 @@ static int taal_power_on(struct omap_dss_device *dssdev)
 
 	return 0;
 err:
-	dev_err(&dssdev->dev, "error while enabling panel, issuing HW reset\n");
+	dev_err(dssdev->dev, "error while enabling panel, issuing HW reset\n");
 
 	taal_hw_reset(dssdev);
 
@@ -1022,7 +1022,7 @@ err0:
 
 static void taal_power_off(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	dsi_disable_video_output(dssdev, td->channel);
@@ -1032,7 +1032,7 @@ static void taal_power_off(struct omap_dss_device *dssdev)
 		r = taal_sleep_in(td);
 
 	if (r) {
-		dev_err(&dssdev->dev,
+		dev_err(dssdev->dev,
 				"error disabling panel, issuing HW reset\n");
 		taal_hw_reset(dssdev);
 	}
@@ -1044,7 +1044,7 @@ static void taal_power_off(struct omap_dss_device *dssdev)
 
 static int taal_panel_reset(struct omap_dss_device *dssdev)
 {
-	dev_err(&dssdev->dev, "performing LCD reset\n");
+	dev_err(dssdev->dev, "performing LCD reset\n");
 
 	taal_power_off(dssdev);
 	taal_hw_reset(dssdev);
@@ -1053,10 +1053,10 @@ static int taal_panel_reset(struct omap_dss_device *dssdev)
 
 static int taal_enable(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
-	dev_dbg(&dssdev->dev, "enable\n");
+	dev_dbg(dssdev->dev, "enable\n");
 
 	mutex_lock(&td->lock);
 
@@ -1082,16 +1082,16 @@ static int taal_enable(struct omap_dss_device *dssdev)
 
 	return 0;
 err:
-	dev_dbg(&dssdev->dev, "enable failed\n");
+	dev_dbg(dssdev->dev, "enable failed\n");
 	mutex_unlock(&td->lock);
 	return r;
 }
 
 static void taal_disable(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
-	dev_dbg(&dssdev->dev, "disable\n");
+	dev_dbg(dssdev->dev, "disable\n");
 
 	mutex_lock(&td->lock);
 
@@ -1118,14 +1118,14 @@ static void taal_disable(struct omap_dss_device *dssdev)
 static void taal_framedone_cb(int err, void *data)
 {
 	struct omap_dss_device *dssdev = data;
-	dev_dbg(&dssdev->dev, "framedone, err %d\n", err);
+	dev_dbg(dssdev->dev, "framedone, err %d\n", err);
 	dsi_bus_unlock(dssdev);
 }
 
 static irqreturn_t taal_te_isr(int irq, void *data)
 {
 	struct omap_dss_device *dssdev = data;
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int old;
 	int r;
 
@@ -1142,7 +1142,7 @@ static irqreturn_t taal_te_isr(int irq, void *data)
 
 	return IRQ_HANDLED;
 err:
-	dev_err(&dssdev->dev, "start update failed\n");
+	dev_err(dssdev->dev, "start update failed\n");
 	dsi_bus_unlock(dssdev);
 	return IRQ_HANDLED;
 }
@@ -1153,7 +1153,7 @@ static void taal_te_timeout_work_callback(struct work_struct *work)
 					te_timeout_work.work);
 	struct omap_dss_device *dssdev = td->dssdev;
 
-	dev_err(&dssdev->dev, "TE not received for 250ms!\n");
+	dev_err(dssdev->dev, "TE not received for 250ms!\n");
 
 	atomic_set(&td->do_update, 0);
 	dsi_bus_unlock(dssdev);
@@ -1162,10 +1162,10 @@ static void taal_te_timeout_work_callback(struct work_struct *work)
 static int taal_update(struct omap_dss_device *dssdev,
 				    u16 x, u16 y, u16 w, u16 h)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
-	dev_dbg(&dssdev->dev, "update %d, %d, %d x %d\n", x, y, w, h);
+	dev_dbg(dssdev->dev, "update %d, %d, %d x %d\n", x, y, w, h);
 
 	mutex_lock(&td->lock);
 	dsi_bus_lock(dssdev);
@@ -1208,23 +1208,23 @@ err:
 
 static int taal_sync(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
-	dev_dbg(&dssdev->dev, "sync\n");
+	dev_dbg(dssdev->dev, "sync\n");
 
 	mutex_lock(&td->lock);
 	dsi_bus_lock(dssdev);
 	dsi_bus_unlock(dssdev);
 	mutex_unlock(&td->lock);
 
-	dev_dbg(&dssdev->dev, "sync done\n");
+	dev_dbg(dssdev->dev, "sync done\n");
 
 	return 0;
 }
 
 static int _taal_enable_te(struct omap_dss_device *dssdev, bool enable)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	if (enable)
@@ -1243,7 +1243,7 @@ static int _taal_enable_te(struct omap_dss_device *dssdev, bool enable)
 
 static int taal_enable_te(struct omap_dss_device *dssdev, bool enable)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&td->lock);
@@ -1279,7 +1279,7 @@ err:
 
 static int taal_get_te(struct omap_dss_device *dssdev)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&td->lock);
@@ -1291,7 +1291,7 @@ static int taal_get_te(struct omap_dss_device *dssdev)
 
 static int taal_run_test(struct omap_dss_device *dssdev, int test_num)
 {
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 	u8 id1, id2, id3;
 	int r;
 
@@ -1336,7 +1336,7 @@ static int taal_memory_read(struct omap_dss_device *dssdev,
 	int first = 1;
 	int plen;
 	unsigned buf_used = 0;
-	struct taal_data *td = dev_get_drvdata(&dssdev->dev);
+	struct taal_data *td = dev_get_drvdata(dssdev->dev);
 
 	if (size < w * h * 3)
 		return -ENOMEM;
@@ -1380,19 +1380,19 @@ static int taal_memory_read(struct omap_dss_device *dssdev,
 				buf + buf_used, size - buf_used);
 
 		if (r < 0) {
-			dev_err(&dssdev->dev, "read error\n");
+			dev_err(dssdev->dev, "read error\n");
 			goto err3;
 		}
 
 		buf_used += r;
 
 		if (r < plen) {
-			dev_err(&dssdev->dev, "short read\n");
+			dev_err(dssdev->dev, "short read\n");
 			break;
 		}
 
 		if (signal_pending(current)) {
-			dev_err(&dssdev->dev, "signal pending, "
+			dev_err(dssdev->dev, "signal pending, "
 					"aborting memory read\n");
 			r = -ERESTARTSYS;
 			goto err3;
@@ -1450,26 +1450,26 @@ static void taal_esd_work(struct work_struct *work)
 
 	r = taal_wake_up(dssdev);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to exit ULPS\n");
+		dev_err(dssdev->dev, "failed to exit ULPS\n");
 		goto err;
 	}
 
 	r = taal_dcs_read_1(td, MIPI_DCS_GET_DIAGNOSTIC_RESULT, &state1);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to read Taal status\n");
+		dev_err(dssdev->dev, "failed to read Taal status\n");
 		goto err;
 	}
 
 	/* Run self diagnostics */
 	r = taal_sleep_out(td);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to run Taal self-diagnostics\n");
+		dev_err(dssdev->dev, "failed to run Taal self-diagnostics\n");
 		goto err;
 	}
 
 	r = taal_dcs_read_1(td, MIPI_DCS_GET_DIAGNOSTIC_RESULT, &state2);
 	if (r) {
-		dev_err(&dssdev->dev, "failed to read Taal status\n");
+		dev_err(dssdev->dev, "failed to read Taal status\n");
 		goto err;
 	}
 
@@ -1477,7 +1477,7 @@ static void taal_esd_work(struct work_struct *work)
 	 * Bit6 if the test passes.
 	 */
 	if (!((state1 ^ state2) & (1 << 6))) {
-		dev_err(&dssdev->dev, "LCD self diagnostics failed\n");
+		dev_err(dssdev->dev, "LCD self diagnostics failed\n");
 		goto err;
 	}
 	/* Self-diagnostics result is also shown on TE GPIO line. We need
@@ -1495,7 +1495,7 @@ static void taal_esd_work(struct work_struct *work)
 	mutex_unlock(&td->lock);
 	return;
 err:
-	dev_err(&dssdev->dev, "performing LCD reset\n");
+	dev_err(dssdev->dev, "performing LCD reset\n");
 
 	taal_panel_reset(dssdev);
 
diff --git a/drivers/video/omap2/displays/panel-tfp410.c b/drivers/video/omap2/displays/panel-tfp410.c
index 46039c4..1fdfb15 100644
--- a/drivers/video/omap2/displays/panel-tfp410.c
+++ b/drivers/video/omap2/displays/panel-tfp410.c
@@ -59,7 +59,7 @@ struct panel_drv_data {
 
 static int tfp410_power_on(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE)
@@ -82,7 +82,7 @@ err0:
 
 static void tfp410_power_off(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 
 	if (dssdev->state != OMAP_DSS_DISPLAY_ACTIVE)
 		return;
@@ -99,7 +99,7 @@ static int tfp410_probe(struct omap_dss_device *dssdev)
 	int r;
 	int i2c_bus_num;
 
-	ddata = devm_kzalloc(&dssdev->dev, sizeof(*ddata), GFP_KERNEL);
+	ddata = devm_kzalloc(dssdev->dev, sizeof(*ddata), GFP_KERNEL);
 	if (!ddata)
 		return -ENOMEM;
 
@@ -119,10 +119,10 @@ static int tfp410_probe(struct omap_dss_device *dssdev)
 	}
 
 	if (gpio_is_valid(ddata->pd_gpio)) {
-		r = devm_gpio_request_one(&dssdev->dev, ddata->pd_gpio,
+		r = devm_gpio_request_one(dssdev->dev, ddata->pd_gpio,
 				GPIOF_OUT_INIT_LOW, "tfp410 pd");
 		if (r) {
-			dev_err(&dssdev->dev, "Failed to request PD GPIO %d\n",
+			dev_err(dssdev->dev, "Failed to request PD GPIO %d\n",
 					ddata->pd_gpio);
 			return r;
 		}
@@ -133,7 +133,7 @@ static int tfp410_probe(struct omap_dss_device *dssdev)
 
 		adapter = i2c_get_adapter(i2c_bus_num);
 		if (!adapter) {
-			dev_err(&dssdev->dev, "Failed to get I2C adapter, bus %d\n",
+			dev_err(dssdev->dev, "Failed to get I2C adapter, bus %d\n",
 					i2c_bus_num);
 			return -EPROBE_DEFER;
 		}
@@ -141,28 +141,28 @@ static int tfp410_probe(struct omap_dss_device *dssdev)
 		ddata->i2c_adapter = adapter;
 	}
 
-	dev_set_drvdata(&dssdev->dev, ddata);
+	dev_set_drvdata(dssdev->dev, ddata);
 
 	return 0;
 }
 
 static void __exit tfp410_remove(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&ddata->lock);
 
 	if (ddata->i2c_adapter)
 		i2c_put_adapter(ddata->i2c_adapter);
 
-	dev_set_drvdata(&dssdev->dev, NULL);
+	dev_set_drvdata(dssdev->dev, NULL);
 
 	mutex_unlock(&ddata->lock);
 }
 
 static int tfp410_enable(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&ddata->lock);
@@ -178,7 +178,7 @@ static int tfp410_enable(struct omap_dss_device *dssdev)
 
 static void tfp410_disable(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&ddata->lock);
 
@@ -192,7 +192,7 @@ static void tfp410_disable(struct omap_dss_device *dssdev)
 static void tfp410_set_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&ddata->lock);
 	omapdss_dpi_set_timings(dssdev, timings);
@@ -203,7 +203,7 @@ static void tfp410_set_timings(struct omap_dss_device *dssdev,
 static void tfp410_get_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 
 	mutex_lock(&ddata->lock);
 	*timings = dssdev->panel.timings;
@@ -213,7 +213,7 @@ static void tfp410_get_timings(struct omap_dss_device *dssdev,
 static int tfp410_check_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	mutex_lock(&ddata->lock);
@@ -258,7 +258,7 @@ static int tfp410_ddc_read(struct i2c_adapter *adapter,
 static int tfp410_read_edid(struct omap_dss_device *dssdev,
 		u8 *edid, int len)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 	int r, l, bytes_read;
 
 	mutex_lock(&ddata->lock);
@@ -298,7 +298,7 @@ err:
 
 static bool tfp410_detect(struct omap_dss_device *dssdev)
 {
-	struct panel_drv_data *ddata = dev_get_drvdata(&dssdev->dev);
+	struct panel_drv_data *ddata = dev_get_drvdata(dssdev->dev);
 	unsigned char out;
 	int r;
 
diff --git a/drivers/video/omap2/displays/panel-tpo-td043mtea1.c b/drivers/video/omap2/displays/panel-tpo-td043mtea1.c
index abf2bc4..7729b6f 100644
--- a/drivers/video/omap2/displays/panel-tpo-td043mtea1.c
+++ b/drivers/video/omap2/displays/panel-tpo-td043mtea1.c
@@ -126,7 +126,7 @@ static int tpo_td043_write_mirror(struct spi_device *spi, bool h, bool v)
 
 static int tpo_td043_set_hmirror(struct omap_dss_device *dssdev, bool enable)
 {
-	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(&dssdev->dev);
+	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(dssdev->dev);
 
 	tpo_td043->hmirror = enable;
 	return tpo_td043_write_mirror(tpo_td043->spi, tpo_td043->hmirror,
@@ -135,7 +135,7 @@ static int tpo_td043_set_hmirror(struct omap_dss_device *dssdev, bool enable)
 
 static bool tpo_td043_get_hmirror(struct omap_dss_device *dssdev)
 {
-	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(&dssdev->dev);
+	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(dssdev->dev);
 
 	return tpo_td043->hmirror;
 }
@@ -338,7 +338,7 @@ static void tpo_td043_power_off(struct tpo_td043_device *tpo_td043)
 
 static int tpo_td043_enable_dss(struct omap_dss_device *dssdev)
 {
-	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(&dssdev->dev);
+	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(dssdev->dev);
 	int r;
 
 	if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE)
@@ -372,7 +372,7 @@ err0:
 
 static void tpo_td043_disable_dss(struct omap_dss_device *dssdev)
 {
-	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(&dssdev->dev);
+	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(dssdev->dev);
 
 	if (dssdev->state != OMAP_DSS_DISPLAY_ACTIVE)
 		return;
@@ -385,14 +385,14 @@ static void tpo_td043_disable_dss(struct omap_dss_device *dssdev)
 
 static int tpo_td043_enable(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "enable\n");
+	dev_dbg(dssdev->dev, "enable\n");
 
 	return tpo_td043_enable_dss(dssdev);
 }
 
 static void tpo_td043_disable(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "disable\n");
+	dev_dbg(dssdev->dev, "disable\n");
 
 	tpo_td043_disable_dss(dssdev);
 
@@ -405,10 +405,10 @@ static int tpo_td043_probe(struct omap_dss_device *dssdev)
 	struct panel_tpo_td043_data *pdata = get_panel_data(dssdev);
 	int ret = 0;
 
-	dev_dbg(&dssdev->dev, "probe\n");
+	dev_dbg(dssdev->dev, "probe\n");
 
 	if (tpo_td043 == NULL) {
-		dev_err(&dssdev->dev, "missing tpo_td043_device\n");
+		dev_err(dssdev->dev, "missing tpo_td043_device\n");
 		return -ENODEV;
 	}
 
@@ -423,28 +423,28 @@ static int tpo_td043_probe(struct omap_dss_device *dssdev)
 	tpo_td043->mode = TPO_R02_MODE_800x480;
 	memcpy(tpo_td043->gamma, tpo_td043_def_gamma, sizeof(tpo_td043->gamma));
 
-	tpo_td043->vcc_reg = regulator_get(&dssdev->dev, "vcc");
+	tpo_td043->vcc_reg = regulator_get(dssdev->dev, "vcc");
 	if (IS_ERR(tpo_td043->vcc_reg)) {
-		dev_err(&dssdev->dev, "failed to get LCD VCC regulator\n");
+		dev_err(dssdev->dev, "failed to get LCD VCC regulator\n");
 		ret = PTR_ERR(tpo_td043->vcc_reg);
 		goto fail_regulator;
 	}
 
 	if (gpio_is_valid(tpo_td043->nreset_gpio)) {
-		ret = devm_gpio_request_one(&dssdev->dev,
+		ret = devm_gpio_request_one(dssdev->dev,
 				tpo_td043->nreset_gpio, GPIOF_OUT_INIT_LOW,
 				"lcd reset");
 		if (ret < 0) {
-			dev_err(&dssdev->dev, "couldn't request reset GPIO\n");
+			dev_err(dssdev->dev, "couldn't request reset GPIO\n");
 			goto fail_gpio_req;
 		}
 	}
 
-	ret = sysfs_create_group(&dssdev->dev.kobj, &tpo_td043_attr_group);
+	ret = sysfs_create_group(&dssdev->dev->kobj, &tpo_td043_attr_group);
 	if (ret)
-		dev_warn(&dssdev->dev, "failed to create sysfs files\n");
+		dev_warn(dssdev->dev, "failed to create sysfs files\n");
 
-	dev_set_drvdata(&dssdev->dev, tpo_td043);
+	dev_set_drvdata(dssdev->dev, tpo_td043);
 
 	return 0;
 
@@ -457,11 +457,11 @@ fail_regulator:
 
 static void tpo_td043_remove(struct omap_dss_device *dssdev)
 {
-	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(&dssdev->dev);
+	struct tpo_td043_device *tpo_td043 = dev_get_drvdata(dssdev->dev);
 
-	dev_dbg(&dssdev->dev, "remove\n");
+	dev_dbg(dssdev->dev, "remove\n");
 
-	sysfs_remove_group(&dssdev->dev.kobj, &tpo_td043_attr_group);
+	sysfs_remove_group(&dssdev->dev->kobj, &tpo_td043_attr_group);
 	regulator_put(tpo_td043->vcc_reg);
 }
 
diff --git a/drivers/video/omap2/dss/Kconfig b/drivers/video/omap2/dss/Kconfig
index cb0f145..8f70a83 100644
--- a/drivers/video/omap2/dss/Kconfig
+++ b/drivers/video/omap2/dss/Kconfig
@@ -1,5 +1,6 @@
 menuconfig OMAP2_DSS
         tristate "OMAP2+ Display Subsystem support"
+	select VIDEOMODE_HELPERS
         help
 	  OMAP2+ Display Subsystem support.
 
diff --git a/drivers/video/omap2/dss/apply.c b/drivers/video/omap2/dss/apply.c
index a4b356a..d6212d63 100644
--- a/drivers/video/omap2/dss/apply.c
+++ b/drivers/video/omap2/dss/apply.c
@@ -420,16 +420,26 @@ static void wait_pending_extra_info_updates(void)
 		DSSWARN("timeout in wait_pending_extra_info_updates\n");
 }
 
-static inline struct omap_dss_device *dss_ovl_get_device(struct omap_overlay *ovl)
+static struct omap_dss_device *dss_mgr_get_device(struct omap_overlay_manager *mgr)
 {
-	return ovl->manager ?
-		(ovl->manager->output ? ovl->manager->output->device : NULL) :
-		NULL;
+	struct omap_dss_device *dssdev;
+
+	dssdev = mgr->output;
+	if (dssdev == NULL)
+		return NULL;
+
+	while (dssdev->device)
+		dssdev = dssdev->device;
+
+	if (dssdev->driver)
+		return dssdev;
+	else
+		return NULL;
 }
 
-static inline struct omap_dss_device *dss_mgr_get_device(struct omap_overlay_manager *mgr)
+static struct omap_dss_device *dss_ovl_get_device(struct omap_overlay *ovl)
 {
-	return mgr->output ? mgr->output->device : NULL;
+	return ovl->manager ? dss_mgr_get_device(ovl->manager) : NULL;
 }
 
 static int dss_mgr_wait_for_vsync(struct omap_overlay_manager *mgr)
@@ -792,6 +802,18 @@ static void mgr_clear_shadow_dirty(struct omap_overlay_manager *mgr)
 	}
 }
 
+static int dss_mgr_connect_compat(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	return mgr->set_output(mgr, dst);
+}
+
+static void dss_mgr_disconnect_compat(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	mgr->unset_output(mgr);
+}
+
 static void dss_mgr_start_update_compat(struct omap_overlay_manager *mgr)
 {
 	struct mgr_priv_data *mp = get_mgr_priv(mgr);
@@ -1156,7 +1178,7 @@ static void dss_mgr_get_info(struct omap_overlay_manager *mgr,
 }
 
 static int dss_mgr_set_output(struct omap_overlay_manager *mgr,
-		struct omap_dss_output *output)
+		struct omap_dss_device *output)
 {
 	int r;
 
@@ -1554,6 +1576,8 @@ static void dss_mgr_unregister_framedone_handler_compat(struct omap_overlay_mana
 }
 
 static const struct dss_mgr_ops apply_mgr_ops = {
+	.connect = dss_mgr_connect_compat,
+	.disconnect = dss_mgr_disconnect_compat,
 	.start_update = dss_mgr_start_update_compat,
 	.enable = dss_mgr_enable_compat,
 	.disable = dss_mgr_disable_compat,
@@ -1569,7 +1593,6 @@ static DEFINE_MUTEX(compat_init_lock);
 int omapdss_compat_init(void)
 {
 	struct platform_device *pdev = dss_get_core_pdev();
-	struct omap_dss_device *dssdev = NULL;
 	int i, r;
 
 	mutex_lock(&compat_init_lock);
@@ -1579,7 +1602,7 @@ int omapdss_compat_init(void)
 
 	apply_init_priv();
 
-	dss_init_overlay_managers(pdev);
+	dss_init_overlay_managers_sysfs(pdev);
 	dss_init_overlays(pdev);
 
 	for (i = 0; i < omap_dss_get_num_overlay_managers(); i++) {
@@ -1615,12 +1638,9 @@ int omapdss_compat_init(void)
 	if (r)
 		goto err_mgr_ops;
 
-	for_each_dss_dev(dssdev) {
-		r = display_init_sysfs(pdev, dssdev);
-		/* XXX uninit sysfs files on error */
-		if (r)
-			goto err_disp_sysfs;
-	}
+	r = display_init_sysfs(pdev);
+	if (r)
+		goto err_disp_sysfs;
 
 	dispc_runtime_get();
 
@@ -1637,12 +1657,13 @@ out:
 
 err_init_irq:
 	dispc_runtime_put();
+	display_uninit_sysfs(pdev);
 
 err_disp_sysfs:
 	dss_uninstall_mgr_ops();
 
 err_mgr_ops:
-	dss_uninit_overlay_managers(pdev);
+	dss_uninit_overlay_managers_sysfs(pdev);
 	dss_uninit_overlays(pdev);
 
 	compat_refcnt--;
@@ -1656,7 +1677,6 @@ EXPORT_SYMBOL(omapdss_compat_init);
 void omapdss_compat_uninit(void)
 {
 	struct platform_device *pdev = dss_get_core_pdev();
-	struct omap_dss_device *dssdev = NULL;
 
 	mutex_lock(&compat_init_lock);
 
@@ -1665,12 +1685,11 @@ void omapdss_compat_uninit(void)
 
 	dss_dispc_uninitialize_irq();
 
-	for_each_dss_dev(dssdev)
-		display_uninit_sysfs(pdev, dssdev);
+	display_uninit_sysfs(pdev);
 
 	dss_uninstall_mgr_ops();
 
-	dss_uninit_overlay_managers(pdev);
+	dss_uninit_overlay_managers_sysfs(pdev);
 	dss_uninit_overlays(pdev);
 out:
 	mutex_unlock(&compat_init_lock);
diff --git a/drivers/video/omap2/dss/core.c b/drivers/video/omap2/dss/core.c
index c9c2252..1aeb274 100644
--- a/drivers/video/omap2/dss/core.c
+++ b/drivers/video/omap2/dss/core.c
@@ -88,7 +88,7 @@ struct regulator *dss_get_vdds_dsi(void)
 	if (core.vdds_dsi_reg != NULL)
 		return core.vdds_dsi_reg;
 
-	reg = regulator_get(&core.pdev->dev, "vdds_dsi");
+	reg = devm_regulator_get(&core.pdev->dev, "vdds_dsi");
 	if (!IS_ERR(reg))
 		core.vdds_dsi_reg = reg;
 
@@ -102,7 +102,7 @@ struct regulator *dss_get_vdds_sdi(void)
 	if (core.vdds_sdi_reg != NULL)
 		return core.vdds_sdi_reg;
 
-	reg = regulator_get(&core.pdev->dev, "vdds_sdi");
+	reg = devm_regulator_get(&core.pdev->dev, "vdds_sdi");
 	if (!IS_ERR(reg))
 		core.vdds_sdi_reg = reg;
 
@@ -243,6 +243,8 @@ static int __init omap_dss_probe(struct platform_device *pdev)
 
 	if (def_disp_name)
 		core.default_display_name = def_disp_name;
+	else if (pdata->default_display_name)
+		core.default_display_name = pdata->default_display_name;
 	else if (pdata->default_device)
 		core.default_display_name = pdata->default_device->name;
 
@@ -290,37 +292,9 @@ static int dss_bus_match(struct device *dev, struct device_driver *driver)
 	return strcmp(dssdev->driver_name, driver->name) == 0;
 }
 
-static ssize_t device_name_show(struct device *dev,
-		struct device_attribute *attr, char *buf)
-{
-	struct omap_dss_device *dssdev = to_dss_device(dev);
-	return snprintf(buf, PAGE_SIZE, "%s\n",
-			dssdev->name ?
-			dssdev->name : "");
-}
-
-static struct device_attribute default_dev_attrs[] = {
-	__ATTR(name, S_IRUGO, device_name_show, NULL),
-	__ATTR_NULL,
-};
-
-static ssize_t driver_name_show(struct device_driver *drv, char *buf)
-{
-	struct omap_dss_driver *dssdrv = to_dss_driver(drv);
-	return snprintf(buf, PAGE_SIZE, "%s\n",
-			dssdrv->driver.name ?
-			dssdrv->driver.name : "");
-}
-static struct driver_attribute default_drv_attrs[] = {
-	__ATTR(name, S_IRUGO, driver_name_show, NULL),
-	__ATTR_NULL,
-};
-
 static struct bus_type dss_bus_type = {
 	.name = "omapdss",
 	.match = dss_bus_match,
-	.dev_attrs = default_dev_attrs,
-	.drv_attrs = default_drv_attrs,
 };
 
 static void dss_bus_release(struct device *dev)
@@ -377,6 +351,46 @@ static int dss_driver_remove(struct device *dev)
 	return 0;
 }
 
+static int omapdss_default_connect(struct omap_dss_device *dssdev)
+{
+	struct omap_dss_device *out;
+	struct omap_overlay_manager *mgr;
+	int r;
+
+	out = dssdev->output;
+
+	if (out == NULL)
+		return -ENODEV;
+
+	mgr = omap_dss_get_overlay_manager(out->dispc_channel);
+	if (!mgr)
+		return -ENODEV;
+
+	r = dss_mgr_connect(mgr, out);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void omapdss_default_disconnect(struct omap_dss_device *dssdev)
+{
+	struct omap_dss_device *out;
+	struct omap_overlay_manager *mgr;
+
+	out = dssdev->output;
+
+	if (out == NULL)
+		return;
+
+	mgr = out->manager;
+
+	if (mgr == NULL)
+		return;
+
+	dss_mgr_disconnect(mgr, out);
+}
+
 int omap_dss_register_driver(struct omap_dss_driver *dssdriver)
 {
 	dssdriver->driver.bus = &dss_bus_type;
@@ -390,6 +404,10 @@ int omap_dss_register_driver(struct omap_dss_driver *dssdriver)
 			omapdss_default_get_recommended_bpp;
 	if (dssdriver->get_timings == NULL)
 		dssdriver->get_timings = omapdss_default_get_timings;
+	if (dssdriver->connect == NULL)
+		dssdriver->connect = omapdss_default_connect;
+	if (dssdriver->disconnect == NULL)
+		dssdriver->disconnect = omapdss_default_disconnect;
 
 	return driver_register(&dssdriver->driver);
 }
@@ -419,29 +437,33 @@ struct omap_dss_device *dss_alloc_and_init_device(struct device *parent)
 	if (!dssdev)
 		return NULL;
 
-	dssdev->dev.bus = &dss_bus_type;
-	dssdev->dev.parent = parent;
-	dssdev->dev.release = omap_dss_dev_release;
-	dev_set_name(&dssdev->dev, "display%d", disp_num_counter++);
+	dssdev->old_dev.bus = &dss_bus_type;
+	dssdev->old_dev.parent = parent;
+	dssdev->old_dev.release = omap_dss_dev_release;
+	dev_set_name(&dssdev->old_dev, "display%d", disp_num_counter++);
 
-	device_initialize(&dssdev->dev);
+	device_initialize(&dssdev->old_dev);
 
 	return dssdev;
 }
 
 int dss_add_device(struct omap_dss_device *dssdev)
 {
-	return device_add(&dssdev->dev);
+	dssdev->dev = &dssdev->old_dev;
+
+	omapdss_register_display(dssdev);
+	return device_add(&dssdev->old_dev);
 }
 
 void dss_put_device(struct omap_dss_device *dssdev)
 {
-	put_device(&dssdev->dev);
+	put_device(&dssdev->old_dev);
 }
 
 void dss_unregister_device(struct omap_dss_device *dssdev)
 {
-	device_unregister(&dssdev->dev);
+	device_unregister(&dssdev->old_dev);
+	omapdss_unregister_display(dssdev);
 }
 
 static int dss_unregister_dss_dev(struct device *dev, void *data)
@@ -618,16 +640,6 @@ static int __init omap_dss_init(void)
 
 static void __exit omap_dss_exit(void)
 {
-	if (core.vdds_dsi_reg != NULL) {
-		regulator_put(core.vdds_dsi_reg);
-		core.vdds_dsi_reg = NULL;
-	}
-
-	if (core.vdds_sdi_reg != NULL) {
-		regulator_put(core.vdds_sdi_reg);
-		core.vdds_sdi_reg = NULL;
-	}
-
 	omap_dss_unregister_drivers();
 
 	omap_dss_bus_unregister();
diff --git a/drivers/video/omap2/dss/dispc-compat.c b/drivers/video/omap2/dss/dispc-compat.c
index 928884c..83779c2 100644
--- a/drivers/video/omap2/dss/dispc-compat.c
+++ b/drivers/video/omap2/dss/dispc-compat.c
@@ -360,8 +360,7 @@ static void dispc_error_worker(struct work_struct *work)
 		if (bit & errors) {
 			DSSERR("FIFO UNDERFLOW on %s, disabling the overlay\n",
 					ovl->name);
-			dispc_ovl_enable(ovl->id, false);
-			dispc_mgr_go(ovl->manager->id);
+			ovl->disable(ovl);
 			msleep(50);
 		}
 	}
diff --git a/drivers/video/omap2/dss/dispc.c b/drivers/video/omap2/dss/dispc.c
index b33b016..02a7340 100644
--- a/drivers/video/omap2/dss/dispc.c
+++ b/drivers/video/omap2/dss/dispc.c
@@ -103,6 +103,7 @@ static struct {
 	int irq;
 
 	unsigned long core_clk_rate;
+	unsigned long tv_pclk_rate;
 
 	u32 fifo_size[DISPC_MAX_NR_FIFOS];
 	/* maps which plane is using a fifo. fifo-id -> plane-id */
@@ -3071,22 +3072,15 @@ unsigned long dispc_mgr_pclk_rate(enum omap_channel channel)
 
 		return r / pcd;
 	} else {
-		enum dss_hdmi_venc_clk_source_select source;
-
-		source = dss_get_hdmi_venc_clk_source();
-
-		switch (source) {
-		case DSS_VENC_TV_CLK:
-			return venc_get_pixel_clock();
-		case DSS_HDMI_M_PCLK:
-			return hdmi_get_pixel_clock();
-		default:
-			BUG();
-			return 0;
-		}
+		return dispc.tv_pclk_rate;
 	}
 }
 
+void dispc_set_tv_pclk(unsigned long pclk)
+{
+	dispc.tv_pclk_rate = pclk;
+}
+
 unsigned long dispc_core_clk_rate(void)
 {
 	return dispc.core_clk_rate;
@@ -3710,6 +3704,8 @@ static int __init omap_dispchw_probe(struct platform_device *pdev)
 
 	dispc_runtime_put();
 
+	dss_init_overlay_managers();
+
 	dss_debugfs_create_file("dispc", dispc_dump_regs);
 
 	return 0;
@@ -3723,6 +3719,8 @@ static int __exit omap_dispchw_remove(struct platform_device *pdev)
 {
 	pm_runtime_disable(&pdev->dev);
 
+	dss_uninit_overlay_managers();
+
 	return 0;
 }
 
diff --git a/drivers/video/omap2/dss/display-sysfs.c b/drivers/video/omap2/dss/display-sysfs.c
index 18211a9..21d7f77 100644
--- a/drivers/video/omap2/dss/display-sysfs.c
+++ b/drivers/video/omap2/dss/display-sysfs.c
@@ -22,42 +22,69 @@
 
 #include <linux/kernel.h>
 #include <linux/module.h>
-#include <linux/jiffies.h>
 #include <linux/platform_device.h>
+#include <linux/sysfs.h>
 
 #include <video/omapdss.h>
 #include "dss.h"
-#include "dss_features.h"
+
+static struct omap_dss_device *to_dss_device_sysfs(struct device *dev)
+{
+	struct omap_dss_device *dssdev = NULL;
+
+	for_each_dss_dev(dssdev) {
+		if (dssdev->dev == dev) {
+			omap_dss_put_device(dssdev);
+			return dssdev;
+		}
+	}
+
+	return NULL;
+}
+
+static ssize_t display_name_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
+
+	return snprintf(buf, PAGE_SIZE, "%s\n",
+			dssdev->name ?
+			dssdev->name : "");
+}
 
 static ssize_t display_enabled_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
-	bool enabled = dssdev->state != OMAP_DSS_DISPLAY_DISABLED;
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 
-	return snprintf(buf, PAGE_SIZE, "%d\n", enabled);
+	return snprintf(buf, PAGE_SIZE, "%d\n",
+			omapdss_device_is_enabled(dssdev));
 }
 
 static ssize_t display_enabled_store(struct device *dev,
 		struct device_attribute *attr,
 		const char *buf, size_t size)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	int r;
-	bool enabled;
+	bool enable;
 
-	r = strtobool(buf, &enabled);
+	r = strtobool(buf, &enable);
 	if (r)
 		return r;
 
-	if (enabled != (dssdev->state != OMAP_DSS_DISPLAY_DISABLED)) {
-		if (enabled) {
-			r = dssdev->driver->enable(dssdev);
-			if (r)
-				return r;
-		} else {
-			dssdev->driver->disable(dssdev);
-		}
+	if (enable == omapdss_device_is_enabled(dssdev))
+		return size;
+
+	if (omapdss_device_is_connected(dssdev) == false)
+		return -ENODEV;
+
+	if (enable) {
+		r = dssdev->driver->enable(dssdev);
+		if (r)
+			return r;
+	} else {
+		dssdev->driver->disable(dssdev);
 	}
 
 	return size;
@@ -66,7 +93,7 @@ static ssize_t display_enabled_store(struct device *dev,
 static ssize_t display_tear_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	return snprintf(buf, PAGE_SIZE, "%d\n",
 			dssdev->driver->get_te ?
 			dssdev->driver->get_te(dssdev) : 0);
@@ -75,7 +102,7 @@ static ssize_t display_tear_show(struct device *dev,
 static ssize_t display_tear_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t size)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	int r;
 	bool te;
 
@@ -96,7 +123,7 @@ static ssize_t display_tear_store(struct device *dev,
 static ssize_t display_timings_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	struct omap_video_timings t;
 
 	if (!dssdev->driver->get_timings)
@@ -113,7 +140,7 @@ static ssize_t display_timings_show(struct device *dev,
 static ssize_t display_timings_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t size)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	struct omap_video_timings t = dssdev->panel.timings;
 	int r, found;
 
@@ -152,7 +179,7 @@ static ssize_t display_timings_store(struct device *dev,
 static ssize_t display_rotate_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	int rotate;
 	if (!dssdev->driver->get_rotate)
 		return -ENOENT;
@@ -163,7 +190,7 @@ static ssize_t display_rotate_show(struct device *dev,
 static ssize_t display_rotate_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t size)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	int rot, r;
 
 	if (!dssdev->driver->set_rotate || !dssdev->driver->get_rotate)
@@ -183,7 +210,7 @@ static ssize_t display_rotate_store(struct device *dev,
 static ssize_t display_mirror_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	int mirror;
 	if (!dssdev->driver->get_mirror)
 		return -ENOENT;
@@ -194,7 +221,7 @@ static ssize_t display_mirror_show(struct device *dev,
 static ssize_t display_mirror_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t size)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	int r;
 	bool mirror;
 
@@ -215,7 +242,7 @@ static ssize_t display_mirror_store(struct device *dev,
 static ssize_t display_wss_show(struct device *dev,
 		struct device_attribute *attr, char *buf)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	unsigned int wss;
 
 	if (!dssdev->driver->get_wss)
@@ -229,7 +256,7 @@ static ssize_t display_wss_show(struct device *dev,
 static ssize_t display_wss_store(struct device *dev,
 		struct device_attribute *attr, const char *buf, size_t size)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = to_dss_device_sysfs(dev);
 	u32 wss;
 	int r;
 
@@ -250,6 +277,7 @@ static ssize_t display_wss_store(struct device *dev,
 	return size;
 }
 
+static DEVICE_ATTR(name, S_IRUGO, display_name_show, NULL);
 static DEVICE_ATTR(enabled, S_IRUGO|S_IWUSR,
 		display_enabled_show, display_enabled_store);
 static DEVICE_ATTR(tear_elim, S_IRUGO|S_IWUSR,
@@ -263,59 +291,55 @@ static DEVICE_ATTR(mirror, S_IRUGO|S_IWUSR,
 static DEVICE_ATTR(wss, S_IRUGO|S_IWUSR,
 		display_wss_show, display_wss_store);
 
-static struct device_attribute *display_sysfs_attrs[] = {
-	&dev_attr_enabled,
-	&dev_attr_tear_elim,
-	&dev_attr_timings,
-	&dev_attr_rotate,
-	&dev_attr_mirror,
-	&dev_attr_wss,
+static const struct attribute *display_sysfs_attrs[] = {
+	&dev_attr_name.attr,
+	&dev_attr_enabled.attr,
+	&dev_attr_tear_elim.attr,
+	&dev_attr_timings.attr,
+	&dev_attr_rotate.attr,
+	&dev_attr_mirror.attr,
+	&dev_attr_wss.attr,
 	NULL
 };
 
-int display_init_sysfs(struct platform_device *pdev,
-		struct omap_dss_device *dssdev)
+int display_init_sysfs(struct platform_device *pdev)
 {
-	struct device_attribute *attr;
-	int i, r;
+	struct omap_dss_device *dssdev = NULL;
+	int r;
 
-	/* create device sysfs files */
-	i = 0;
-	while ((attr = display_sysfs_attrs[i++]) != NULL) {
-		r = device_create_file(&dssdev->dev, attr);
-		if (r) {
-			for (i = i - 2; i >= 0; i--) {
-				attr = display_sysfs_attrs[i];
-				device_remove_file(&dssdev->dev, attr);
-			}
+	for_each_dss_dev(dssdev) {
+		struct kobject *kobj = &dssdev->dev->kobj;
 
-			DSSERR("failed to create sysfs file\n");
-			return r;
+		r = sysfs_create_files(kobj, display_sysfs_attrs);
+		if (r) {
+			DSSERR("failed to create sysfs files\n");
+			goto err;
 		}
-	}
 
-	/* create display? sysfs links */
-	r = sysfs_create_link(&pdev->dev.kobj, &dssdev->dev.kobj,
-			dev_name(&dssdev->dev));
-	if (r) {
-		while ((attr = display_sysfs_attrs[i++]) != NULL)
-			device_remove_file(&dssdev->dev, attr);
+		r = sysfs_create_link(&pdev->dev.kobj, kobj, dssdev->alias);
+		if (r) {
+			sysfs_remove_files(kobj, display_sysfs_attrs);
 
-		DSSERR("failed to create sysfs display link\n");
-		return r;
+			DSSERR("failed to create sysfs display link\n");
+			goto err;
+		}
 	}
 
 	return 0;
+
+err:
+	display_uninit_sysfs(pdev);
+
+	return r;
 }
 
-void display_uninit_sysfs(struct platform_device *pdev,
-		struct omap_dss_device *dssdev)
+void display_uninit_sysfs(struct platform_device *pdev)
 {
-	struct device_attribute *attr;
-	int i = 0;
-
-	sysfs_remove_link(&pdev->dev.kobj, dev_name(&dssdev->dev));
+	struct omap_dss_device *dssdev = NULL;
 
-	while ((attr = display_sysfs_attrs[i++]) != NULL)
-		device_remove_file(&dssdev->dev, attr);
+	for_each_dss_dev(dssdev) {
+		sysfs_remove_link(&pdev->dev.kobj, dssdev->alias);
+		sysfs_remove_files(&dssdev->dev->kobj,
+				display_sysfs_attrs);
+	}
 }
diff --git a/drivers/video/omap2/dss/display.c b/drivers/video/omap2/dss/display.c
index 0aa8ad8..fafe7c9 100644
--- a/drivers/video/omap2/dss/display.c
+++ b/drivers/video/omap2/dss/display.c
@@ -61,6 +61,7 @@ int omapdss_default_get_recommended_bpp(struct omap_dss_device *dssdev)
 	case OMAP_DISPLAY_TYPE_VENC:
 	case OMAP_DISPLAY_TYPE_SDI:
 	case OMAP_DISPLAY_TYPE_HDMI:
+	case OMAP_DISPLAY_TYPE_DVI:
 		return 24;
 	default:
 		BUG();
@@ -76,110 +77,154 @@ void omapdss_default_get_timings(struct omap_dss_device *dssdev,
 }
 EXPORT_SYMBOL(omapdss_default_get_timings);
 
-static int dss_suspend_device(struct device *dev, void *data)
+int dss_suspend_all_devices(void)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
-
-	if (dssdev->state != OMAP_DSS_DISPLAY_ACTIVE) {
-		dssdev->activate_after_resume = false;
-		return 0;
-	}
-
-	dssdev->driver->disable(dssdev);
-
-	dssdev->activate_after_resume = true;
+	struct omap_dss_device *dssdev = NULL;
 
-	return 0;
-}
+	for_each_dss_dev(dssdev) {
+		if (!dssdev->driver)
+			continue;
 
-int dss_suspend_all_devices(void)
-{
-	int r;
-	struct bus_type *bus = dss_get_bus();
-
-	r = bus_for_each_dev(bus, NULL, NULL, dss_suspend_device);
-	if (r) {
-		/* resume all displays that were suspended */
-		dss_resume_all_devices();
-		return r;
+		if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE) {
+			dssdev->driver->disable(dssdev);
+			dssdev->activate_after_resume = true;
+		} else {
+			dssdev->activate_after_resume = false;
+		}
 	}
 
 	return 0;
 }
 
-static int dss_resume_device(struct device *dev, void *data)
+int dss_resume_all_devices(void)
 {
-	int r;
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_device *dssdev = NULL;
 
-	if (dssdev->activate_after_resume) {
-		r = dssdev->driver->enable(dssdev);
-		if (r)
-			return r;
-	}
+	for_each_dss_dev(dssdev) {
+		if (!dssdev->driver)
+			continue;
 
-	dssdev->activate_after_resume = false;
+		if (dssdev->activate_after_resume) {
+			dssdev->driver->enable(dssdev);
+			dssdev->activate_after_resume = false;
+		}
+	}
 
 	return 0;
 }
 
-int dss_resume_all_devices(void)
+void dss_disable_all_devices(void)
 {
-	struct bus_type *bus = dss_get_bus();
+	struct omap_dss_device *dssdev = NULL;
+
+	for_each_dss_dev(dssdev) {
+		if (!dssdev->driver)
+			continue;
 
-	return bus_for_each_dev(bus, NULL, NULL, dss_resume_device);
+		if (dssdev->state == OMAP_DSS_DISPLAY_ACTIVE)
+			dssdev->driver->disable(dssdev);
+	}
 }
 
-static int dss_disable_device(struct device *dev, void *data)
+static LIST_HEAD(panel_list);
+static DEFINE_MUTEX(panel_list_mutex);
+static int disp_num_counter;
+
+int omapdss_register_display(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_device *dssdev = to_dss_device(dev);
+	struct omap_dss_driver *drv = dssdev->driver;
 
-	if (dssdev->state != OMAP_DSS_DISPLAY_DISABLED)
-		dssdev->driver->disable(dssdev);
+	snprintf(dssdev->alias, sizeof(dssdev->alias),
+			"display%d", disp_num_counter++);
 
+	if (drv && drv->get_resolution == NULL)
+		drv->get_resolution = omapdss_default_get_resolution;
+	if (drv && drv->get_recommended_bpp == NULL)
+		drv->get_recommended_bpp = omapdss_default_get_recommended_bpp;
+	if (drv && drv->get_timings == NULL)
+		drv->get_timings = omapdss_default_get_timings;
+
+	mutex_lock(&panel_list_mutex);
+	list_add_tail(&dssdev->panel_list, &panel_list);
+	mutex_unlock(&panel_list_mutex);
 	return 0;
 }
+EXPORT_SYMBOL(omapdss_register_display);
 
-void dss_disable_all_devices(void)
+void omapdss_unregister_display(struct omap_dss_device *dssdev)
 {
-	struct bus_type *bus = dss_get_bus();
-	bus_for_each_dev(bus, NULL, NULL, dss_disable_device);
+	mutex_lock(&panel_list_mutex);
+	list_del(&dssdev->panel_list);
+	mutex_unlock(&panel_list_mutex);
 }
+EXPORT_SYMBOL(omapdss_unregister_display);
 
-
-void omap_dss_get_device(struct omap_dss_device *dssdev)
+struct omap_dss_device *omap_dss_get_device(struct omap_dss_device *dssdev)
 {
-	get_device(&dssdev->dev);
+	if (!try_module_get(dssdev->owner))
+		return NULL;
+
+	if (get_device(dssdev->dev) == NULL) {
+		module_put(dssdev->owner);
+		return NULL;
+	}
+
+	return dssdev;
 }
 EXPORT_SYMBOL(omap_dss_get_device);
 
 void omap_dss_put_device(struct omap_dss_device *dssdev)
 {
-	put_device(&dssdev->dev);
+	put_device(dssdev->dev);
+	module_put(dssdev->owner);
 }
 EXPORT_SYMBOL(omap_dss_put_device);
 
-/* ref count of the found device is incremented. ref count
- * of from-device is decremented. */
+/*
+ * ref count of the found device is incremented.
+ * ref count of from-device is decremented.
+ */
 struct omap_dss_device *omap_dss_get_next_device(struct omap_dss_device *from)
 {
-	struct device *dev;
-	struct device *dev_start = NULL;
-	struct omap_dss_device *dssdev = NULL;
+	struct list_head *l;
+	struct omap_dss_device *dssdev;
+
+	mutex_lock(&panel_list_mutex);
 
-	int match(struct device *dev, void *data)
-	{
-		return 1;
+	if (list_empty(&panel_list)) {
+		dssdev = NULL;
+		goto out;
 	}
 
-	if (from)
-		dev_start = &from->dev;
-	dev = bus_find_device(dss_get_bus(), dev_start, NULL, match);
-	if (dev)
-		dssdev = to_dss_device(dev);
-	if (from)
-		put_device(&from->dev);
+	if (from == NULL) {
+		dssdev = list_first_entry(&panel_list, struct omap_dss_device,
+				panel_list);
+		omap_dss_get_device(dssdev);
+		goto out;
+	}
+
+	omap_dss_put_device(from);
+
+	list_for_each(l, &panel_list) {
+		dssdev = list_entry(l, struct omap_dss_device, panel_list);
+		if (dssdev == from) {
+			if (list_is_last(l, &panel_list)) {
+				dssdev = NULL;
+				goto out;
+			}
+
+			dssdev = list_entry(l->next, struct omap_dss_device,
+					panel_list);
+			omap_dss_get_device(dssdev);
+			goto out;
+		}
+	}
 
+	WARN(1, "'from' dssdev not found\n");
+
+	dssdev = NULL;
+out:
+	mutex_unlock(&panel_list_mutex);
 	return dssdev;
 }
 EXPORT_SYMBOL(omap_dss_get_next_device);
@@ -198,24 +243,72 @@ struct omap_dss_device *omap_dss_find_device(void *data,
 }
 EXPORT_SYMBOL(omap_dss_find_device);
 
-int omap_dss_start_device(struct omap_dss_device *dssdev)
+void videomode_to_omap_video_timings(const struct videomode *vm,
+		struct omap_video_timings *ovt)
 {
-	if (!dssdev->driver) {
-		DSSDBG("no driver\n");
-		return -ENODEV;
-	}
-
-	if (!try_module_get(dssdev->dev.driver->owner)) {
-		return -ENODEV;
-	}
-
-	return 0;
+	memset(ovt, 0, sizeof(*ovt));
+
+	ovt->pixel_clock = vm->pixelclock / 1000;
+	ovt->x_res = vm->hactive;
+	ovt->hbp = vm->hback_porch;
+	ovt->hfp = vm->hfront_porch;
+	ovt->hsw = vm->hsync_len;
+	ovt->y_res = vm->vactive;
+	ovt->vbp = vm->vback_porch;
+	ovt->vfp = vm->vfront_porch;
+	ovt->vsw = vm->vsync_len;
+
+	ovt->vsync_level = vm->flags & DISPLAY_FLAGS_VSYNC_HIGH ?
+		OMAPDSS_SIG_ACTIVE_HIGH :
+		OMAPDSS_SIG_ACTIVE_LOW;
+	ovt->hsync_level = vm->flags & DISPLAY_FLAGS_HSYNC_HIGH ?
+		OMAPDSS_SIG_ACTIVE_HIGH :
+		OMAPDSS_SIG_ACTIVE_LOW;
+	ovt->de_level = vm->flags & DISPLAY_FLAGS_DE_HIGH ?
+		OMAPDSS_SIG_ACTIVE_HIGH :
+		OMAPDSS_SIG_ACTIVE_HIGH;
+	ovt->data_pclk_edge = vm->flags & DISPLAY_FLAGS_PIXDATA_POSEDGE ?
+		OMAPDSS_DRIVE_SIG_RISING_EDGE :
+		OMAPDSS_DRIVE_SIG_FALLING_EDGE;
+
+	ovt->sync_pclk_edge = OMAPDSS_DRIVE_SIG_OPPOSITE_EDGES;
 }
-EXPORT_SYMBOL(omap_dss_start_device);
+EXPORT_SYMBOL(videomode_to_omap_video_timings);
 
-void omap_dss_stop_device(struct omap_dss_device *dssdev)
+void omap_video_timings_to_videomode(const struct omap_video_timings *ovt,
+		struct videomode *vm)
 {
-	module_put(dssdev->dev.driver->owner);
+	memset(vm, 0, sizeof(*vm));
+
+	vm->pixelclock = ovt->pixel_clock * 1000;
+
+	vm->hactive = ovt->x_res;
+	vm->hback_porch = ovt->hbp;
+	vm->hfront_porch = ovt->hfp;
+	vm->hsync_len = ovt->hsw;
+	vm->vactive = ovt->y_res;
+	vm->vback_porch = ovt->vbp;
+	vm->vfront_porch = ovt->vfp;
+	vm->vsync_len = ovt->vsw;
+
+	if (ovt->hsync_level == OMAPDSS_SIG_ACTIVE_HIGH)
+		vm->flags |= DISPLAY_FLAGS_HSYNC_HIGH;
+	else
+		vm->flags |= DISPLAY_FLAGS_HSYNC_LOW;
+
+	if (ovt->vsync_level == OMAPDSS_SIG_ACTIVE_HIGH)
+		vm->flags |= DISPLAY_FLAGS_VSYNC_HIGH;
+	else
+		vm->flags |= DISPLAY_FLAGS_VSYNC_LOW;
+
+	if (ovt->de_level == OMAPDSS_SIG_ACTIVE_HIGH)
+		vm->flags |= DISPLAY_FLAGS_DE_HIGH;
+	else
+		vm->flags |= DISPLAY_FLAGS_DE_LOW;
+
+	if (ovt->data_pclk_edge == OMAPDSS_DRIVE_SIG_RISING_EDGE)
+		vm->flags |= DISPLAY_FLAGS_PIXDATA_POSEDGE;
+	else
+		vm->flags |= DISPLAY_FLAGS_PIXDATA_NEGEDGE;
 }
-EXPORT_SYMBOL(omap_dss_stop_device);
-
+EXPORT_SYMBOL(omap_video_timings_to_videomode);
diff --git a/drivers/video/omap2/dss/dpi.c b/drivers/video/omap2/dss/dpi.c
index 757b57f..a6b331e 100644
--- a/drivers/video/omap2/dss/dpi.c
+++ b/drivers/video/omap2/dss/dpi.c
@@ -37,6 +37,8 @@
 #include "dss_features.h"
 
 static struct {
+	struct platform_device *pdev;
+
 	struct regulator *vdds_dsi_reg;
 	struct platform_device *dsidev;
 
@@ -46,7 +48,7 @@ static struct {
 	struct dss_lcd_mgr_config mgr_config;
 	int data_lines;
 
-	struct omap_dss_output output;
+	struct omap_dss_device output;
 } dpi;
 
 static struct platform_device *dpi_get_dsidev(enum omap_channel channel)
@@ -129,7 +131,7 @@ static bool dpi_calc_dispc_cb(int lckd, int pckd, unsigned long lck,
 	 * shifted. So skip all odd dividers when the pixel clock is on the
 	 * higher side.
 	 */
-	if (ctx->pck_min >= 1000000) {
+	if (ctx->pck_min >= 100000000) {
 		if (lckd > 1 && lckd % 2 != 0)
 			return false;
 
@@ -156,7 +158,7 @@ static bool dpi_calc_hsdiv_cb(int regm_dispc, unsigned long dispc,
 	 * shifted. So skip all odd dividers when the pixel clock is on the
 	 * higher side.
 	 */
-	if (regm_dispc > 1 && regm_dispc % 2 != 0 && ctx->pck_min >= 1000000)
+	if (regm_dispc > 1 && regm_dispc % 2 != 0 && ctx->pck_min >= 100000000)
 		return false;
 
 	ctx->dsi_cinfo.regm_dispc = regm_dispc;
@@ -345,7 +347,7 @@ static void dpi_config_lcd_manager(struct omap_overlay_manager *mgr)
 
 int omapdss_dpi_display_enable(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_output *out = &dpi.output;
+	struct omap_dss_device *out = &dpi.output;
 	int r;
 
 	mutex_lock(&dpi.lock);
@@ -362,12 +364,6 @@ int omapdss_dpi_display_enable(struct omap_dss_device *dssdev)
 		goto err_no_out_mgr;
 	}
 
-	r = omap_dss_start_device(dssdev);
-	if (r) {
-		DSSERR("failed to start device\n");
-		goto err_start_dev;
-	}
-
 	if (dss_has_feature(FEAT_DPI_USES_VDDS_DSI)) {
 		r = regulator_enable(dpi.vdds_dsi_reg);
 		if (r)
@@ -422,8 +418,6 @@ err_get_dispc:
 	if (dss_has_feature(FEAT_DPI_USES_VDDS_DSI))
 		regulator_disable(dpi.vdds_dsi_reg);
 err_reg_enable:
-	omap_dss_stop_device(dssdev);
-err_start_dev:
 err_no_out_mgr:
 err_no_reg:
 	mutex_unlock(&dpi.lock);
@@ -450,8 +444,6 @@ void omapdss_dpi_display_disable(struct omap_dss_device *dssdev)
 	if (dss_has_feature(FEAT_DPI_USES_VDDS_DSI))
 		regulator_disable(dpi.vdds_dsi_reg);
 
-	omap_dss_stop_device(dssdev);
-
 	mutex_unlock(&dpi.lock);
 }
 EXPORT_SYMBOL(omapdss_dpi_display_disable);
@@ -469,6 +461,16 @@ void omapdss_dpi_set_timings(struct omap_dss_device *dssdev,
 }
 EXPORT_SYMBOL(omapdss_dpi_set_timings);
 
+static void dpi_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	mutex_lock(&dpi.lock);
+
+	*timings = dpi.timings;
+
+	mutex_unlock(&dpi.lock);
+}
+
 int dpi_check_timings(struct omap_dss_device *dssdev,
 			struct omap_video_timings *timings)
 {
@@ -542,6 +544,50 @@ static int dpi_verify_dsi_pll(struct platform_device *dsidev)
 	return 0;
 }
 
+static int dpi_init_regulator(void)
+{
+	struct regulator *vdds_dsi;
+
+	if (!dss_has_feature(FEAT_DPI_USES_VDDS_DSI))
+		return 0;
+
+	if (dpi.vdds_dsi_reg)
+		return 0;
+
+	vdds_dsi = dss_get_vdds_dsi();
+
+	if (IS_ERR(vdds_dsi)) {
+		vdds_dsi = devm_regulator_get(&dpi.pdev->dev, "vdds_dsi");
+		if (IS_ERR(vdds_dsi)) {
+			DSSERR("can't get VDDS_DSI regulator\n");
+			return PTR_ERR(vdds_dsi);
+		}
+	}
+
+	dpi.vdds_dsi_reg = vdds_dsi;
+
+	return 0;
+}
+
+static void dpi_init_pll(void)
+{
+	struct platform_device *dsidev;
+
+	if (dpi.dsidev)
+		return;
+
+	dsidev = dpi_get_dsidev(dpi.output.dispc_channel);
+	if (!dsidev)
+		return;
+
+	if (dpi_verify_dsi_pll(dsidev)) {
+		DSSWARN("DSI PLL not operational\n");
+		return;
+	}
+
+	dpi.dsidev = dsidev;
+}
+
 /*
  * Return a hardcoded channel for the DPI output. This should work for
  * current use cases, but this can be later expanded to either resolve
@@ -572,41 +618,6 @@ static enum omap_channel dpi_get_channel(void)
 	}
 }
 
-static int dpi_init_display(struct omap_dss_device *dssdev)
-{
-	struct platform_device *dsidev;
-
-	DSSDBG("init_display\n");
-
-	if (dss_has_feature(FEAT_DPI_USES_VDDS_DSI) &&
-					dpi.vdds_dsi_reg == NULL) {
-		struct regulator *vdds_dsi;
-
-		vdds_dsi = dss_get_vdds_dsi();
-
-		if (IS_ERR(vdds_dsi)) {
-			DSSERR("can't get VDDS_DSI regulator\n");
-			return PTR_ERR(vdds_dsi);
-		}
-
-		dpi.vdds_dsi_reg = vdds_dsi;
-	}
-
-	dsidev = dpi_get_dsidev(dpi.output.dispc_channel);
-
-	if (dsidev && dpi_verify_dsi_pll(dsidev)) {
-		dsidev = NULL;
-		DSSWARN("DSI PLL not operational\n");
-	}
-
-	if (dsidev)
-		DSSDBG("using DSI PLL for DPI clock\n");
-
-	dpi.dsidev = dsidev;
-
-	return 0;
-}
-
 static struct omap_dss_device *dpi_find_dssdev(struct platform_device *pdev)
 {
 	struct omap_dss_board_info *pdata = pdev->dev.platform_data;
@@ -646,19 +657,18 @@ static int dpi_probe_pdata(struct platform_device *dpidev)
 	if (!plat_dssdev)
 		return 0;
 
+	r = dpi_init_regulator();
+	if (r)
+		return r;
+
+	dpi_init_pll();
+
 	dssdev = dss_alloc_and_init_device(&dpidev->dev);
 	if (!dssdev)
 		return -ENOMEM;
 
 	dss_copy_device_pdata(dssdev, plat_dssdev);
 
-	r = dpi_init_display(dssdev);
-	if (r) {
-		DSSERR("device %s init failed: %d\n", dssdev->name, r);
-		dss_put_device(dssdev);
-		return r;
-	}
-
 	r = omapdss_output_set_device(&dpi.output, dssdev);
 	if (r) {
 		DSSERR("failed to connect output to new device: %s\n",
@@ -678,41 +688,108 @@ static int dpi_probe_pdata(struct platform_device *dpidev)
 	return 0;
 }
 
+static int dpi_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct omap_overlay_manager *mgr;
+	int r;
+
+	r = dpi_init_regulator();
+	if (r)
+		return r;
+
+	dpi_init_pll();
+
+	mgr = omap_dss_get_overlay_manager(dssdev->dispc_channel);
+	if (!mgr)
+		return -ENODEV;
+
+	r = dss_mgr_connect(mgr, dssdev);
+	if (r)
+		return r;
+
+	r = omapdss_output_set_device(dssdev, dst);
+	if (r) {
+		DSSERR("failed to connect output to new device: %s\n",
+				dst->name);
+		dss_mgr_disconnect(mgr, dssdev);
+		return r;
+	}
+
+	return 0;
+}
+
+static void dpi_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	WARN_ON(dst != dssdev->device);
+
+	if (dst != dssdev->device)
+		return;
+
+	omapdss_output_unset_device(dssdev);
+
+	if (dssdev->manager)
+		dss_mgr_disconnect(dssdev->manager, dssdev);
+}
+
+static const struct omapdss_dpi_ops dpi_ops = {
+	.connect = dpi_connect,
+	.disconnect = dpi_disconnect,
+
+	.enable = omapdss_dpi_display_enable,
+	.disable = omapdss_dpi_display_disable,
+
+	.check_timings = dpi_check_timings,
+	.set_timings = omapdss_dpi_set_timings,
+	.get_timings = dpi_get_timings,
+
+	.set_data_lines = omapdss_dpi_set_data_lines,
+};
+
 static void dpi_init_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &dpi.output;
+	struct omap_dss_device *out = &dpi.output;
 
-	out->pdev = pdev;
+	out->dev = &pdev->dev;
 	out->id = OMAP_DSS_OUTPUT_DPI;
-	out->type = OMAP_DISPLAY_TYPE_DPI;
+	out->output_type = OMAP_DISPLAY_TYPE_DPI;
 	out->name = "dpi.0";
 	out->dispc_channel = dpi_get_channel();
+	out->ops.dpi = &dpi_ops;
+	out->owner = THIS_MODULE;
 
-	dss_register_output(out);
+	omapdss_register_output(out);
 }
 
 static void __exit dpi_uninit_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &dpi.output;
+	struct omap_dss_device *out = &dpi.output;
 
-	dss_unregister_output(out);
+	omapdss_unregister_output(out);
 }
 
 static int omap_dpi_probe(struct platform_device *pdev)
 {
 	int r;
 
+	dpi.pdev = pdev;
+
 	mutex_init(&dpi.lock);
 
 	dpi_init_output(pdev);
 
-	r = dpi_probe_pdata(pdev);
-	if (r) {
-		dpi_uninit_output(pdev);
-		return r;
+	if (pdev->dev.platform_data) {
+		r = dpi_probe_pdata(pdev);
+		if (r)
+			goto err_probe;
 	}
 
 	return 0;
+
+err_probe:
+	dpi_uninit_output(pdev);
+	return r;
 }
 
 static int __exit omap_dpi_remove(struct platform_device *pdev)
diff --git a/drivers/video/omap2/dss/dsi.c b/drivers/video/omap2/dss/dsi.c
index a73dedc..99a043b 100644
--- a/drivers/video/omap2/dss/dsi.c
+++ b/drivers/video/omap2/dss/dsi.c
@@ -363,7 +363,7 @@ struct dsi_data {
 	enum omap_dss_dsi_mode mode;
 	struct omap_dss_dsi_videomode_timings vm_timings;
 
-	struct omap_dss_output output;
+	struct omap_dss_device output;
 };
 
 struct dsi_packet_sent_handler_data {
@@ -383,12 +383,21 @@ static inline struct dsi_data *dsi_get_dsidrv_data(struct platform_device *dside
 
 static inline struct platform_device *dsi_get_dsidev_from_dssdev(struct omap_dss_device *dssdev)
 {
-	return dssdev->output->pdev;
+	/* HACK: dssdev can be either the panel device, when using old API, or
+	 * the dsi device itself, when using the new API. So we solve this for
+	 * now by checking the dssdev->id. This will be removed when the old API
+	 * is removed.
+	 */
+	if (dssdev->id == OMAP_DSS_OUTPUT_DSI1 ||
+			dssdev->id == OMAP_DSS_OUTPUT_DSI2)
+		return to_platform_device(dssdev->dev);
+
+	return to_platform_device(dssdev->output->dev);
 }
 
 struct platform_device *dsi_get_dsidev_from_id(int module)
 {
-	struct omap_dss_output *out;
+	struct omap_dss_device *out;
 	enum omap_dss_output_id	id;
 
 	switch (module) {
@@ -404,7 +413,7 @@ struct platform_device *dsi_get_dsidev_from_id(int module)
 
 	out = omap_dss_get_output(id);
 
-	return out ? out->pdev : NULL;
+	return out ? to_platform_device(out->dev) : NULL;
 }
 
 static inline void dsi_write_reg(struct platform_device *dsidev,
@@ -1114,6 +1123,30 @@ void dsi_runtime_put(struct platform_device *dsidev)
 	WARN_ON(r < 0 && r != -ENOSYS);
 }
 
+static int dsi_regulator_init(struct platform_device *dsidev)
+{
+	struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev);
+	struct regulator *vdds_dsi;
+
+	if (dsi->vdds_dsi_reg != NULL)
+		return 0;
+
+	vdds_dsi = devm_regulator_get(&dsi->pdev->dev, "vdds_dsi");
+
+	/* DT HACK: try VCXIO to make omapdss work for o4 sdp/panda */
+	if (IS_ERR(vdds_dsi))
+		vdds_dsi = devm_regulator_get(&dsi->pdev->dev, "VCXIO");
+
+	if (IS_ERR(vdds_dsi)) {
+		DSSERR("can't get VDDS_DSI regulator\n");
+		return PTR_ERR(vdds_dsi);
+	}
+
+	dsi->vdds_dsi_reg = vdds_dsi;
+
+	return 0;
+}
+
 /* source clock for DSI PLL. this could also be PCLKFREE */
 static inline void dsi_enable_pll_clock(struct platform_device *dsidev,
 		bool enable)
@@ -1592,22 +1625,9 @@ int dsi_pll_init(struct platform_device *dsidev, bool enable_hsclk,
 	 */
 	enable_hsclk = enable_hsdiv = true;
 
-	if (dsi->vdds_dsi_reg == NULL) {
-		struct regulator *vdds_dsi;
-
-		vdds_dsi = regulator_get(&dsi->pdev->dev, "vdds_dsi");
-
-		/* DT HACK: try VCXIO to make omapdss work for o4 sdp/panda */
-		if (IS_ERR(vdds_dsi))
-			vdds_dsi = regulator_get(&dsi->pdev->dev, "VCXIO");
-
-		if (IS_ERR(vdds_dsi)) {
-			DSSERR("can't get VDDS_DSI regulator\n");
-			return PTR_ERR(vdds_dsi);
-		}
-
-		dsi->vdds_dsi_reg = vdds_dsi;
-	}
+	r = dsi_regulator_init(dsidev);
+	if (r)
+		return r;
 
 	dsi_enable_pll_clock(dsidev, 1);
 	/*
@@ -4122,7 +4142,7 @@ int dsi_enable_video_output(struct omap_dss_device *dssdev, int channel)
 	struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev);
 	struct omap_overlay_manager *mgr = dsi->output.manager;
 	int bpp = dsi_get_pixel_size(dsi->pix_fmt);
-	struct omap_dss_output *out = &dsi->output;
+	struct omap_dss_device *out = &dsi->output;
 	u8 data_type;
 	u16 word_count;
 	int r;
@@ -4581,12 +4601,6 @@ int omapdss_dsi_display_enable(struct omap_dss_device *dssdev)
 
 	mutex_lock(&dsi->lock);
 
-	r = omap_dss_start_device(dssdev);
-	if (r) {
-		DSSERR("failed to start device\n");
-		goto err_start_dev;
-	}
-
 	r = dsi_runtime_get(dsidev);
 	if (r)
 		goto err_get_dsi;
@@ -4607,8 +4621,6 @@ err_init_dsi:
 	dsi_enable_pll_clock(dsidev, 0);
 	dsi_runtime_put(dsidev);
 err_get_dsi:
-	omap_dss_stop_device(dssdev);
-err_start_dev:
 	mutex_unlock(&dsi->lock);
 	DSSDBG("dsi_display_enable FAILED\n");
 	return r;
@@ -4637,8 +4649,6 @@ void omapdss_dsi_display_disable(struct omap_dss_device *dssdev,
 	dsi_runtime_put(dsidev);
 	dsi_enable_pll_clock(dsidev, 0);
 
-	omap_dss_stop_device(dssdev);
-
 	mutex_unlock(&dsi->lock);
 }
 EXPORT_SYMBOL(omapdss_dsi_display_disable);
@@ -5225,34 +5235,6 @@ static enum omap_channel dsi_get_channel(int module_id)
 	}
 }
 
-static int dsi_init_display(struct omap_dss_device *dssdev)
-{
-	struct platform_device *dsidev =
-			dsi_get_dsidev_from_id(dssdev->phy.dsi.module);
-	struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev);
-
-	DSSDBG("DSI init\n");
-
-	if (dsi->vdds_dsi_reg == NULL) {
-		struct regulator *vdds_dsi;
-
-		vdds_dsi = regulator_get(&dsi->pdev->dev, "vdds_dsi");
-
-		/* DT HACK: try VCXIO to make omapdss work for o4 sdp/panda */
-		if (IS_ERR(vdds_dsi))
-			vdds_dsi = regulator_get(&dsi->pdev->dev, "VCXIO");
-
-		if (IS_ERR(vdds_dsi)) {
-			DSSERR("can't get VDDS_DSI regulator\n");
-			return PTR_ERR(vdds_dsi);
-		}
-
-		dsi->vdds_dsi_reg = vdds_dsi;
-	}
-
-	return 0;
-}
-
 int omap_dsi_request_vc(struct omap_dss_device *dssdev, int *channel)
 {
 	struct platform_device *dsidev = dsi_get_dsidev_from_dssdev(dssdev);
@@ -5410,19 +5392,16 @@ static int dsi_probe_pdata(struct platform_device *dsidev)
 	if (!plat_dssdev)
 		return 0;
 
+	r = dsi_regulator_init(dsidev);
+	if (r)
+		return r;
+
 	dssdev = dss_alloc_and_init_device(&dsidev->dev);
 	if (!dssdev)
 		return -ENOMEM;
 
 	dss_copy_device_pdata(dssdev, plat_dssdev);
 
-	r = dsi_init_display(dssdev);
-	if (r) {
-		DSSERR("device %s init failed: %d\n", dssdev->name, r);
-		dss_put_device(dssdev);
-		return r;
-	}
-
 	r = omapdss_output_set_device(&dsi->output, dssdev);
 	if (r) {
 		DSSERR("failed to connect output to new device: %s\n",
@@ -5442,28 +5421,113 @@ static int dsi_probe_pdata(struct platform_device *dsidev)
 	return 0;
 }
 
+static int dsi_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct platform_device *dsidev = dsi_get_dsidev_from_dssdev(dssdev);
+	struct omap_overlay_manager *mgr;
+	int r;
+
+	r = dsi_regulator_init(dsidev);
+	if (r)
+		return r;
+
+	mgr = omap_dss_get_overlay_manager(dssdev->dispc_channel);
+	if (!mgr)
+		return -ENODEV;
+
+	r = dss_mgr_connect(mgr, dssdev);
+	if (r)
+		return r;
+
+	r = omapdss_output_set_device(dssdev, dst);
+	if (r) {
+		DSSERR("failed to connect output to new device: %s\n",
+				dssdev->name);
+		dss_mgr_disconnect(mgr, dssdev);
+		return r;
+	}
+
+	return 0;
+}
+
+static void dsi_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	WARN_ON(dst != dssdev->device);
+
+	if (dst != dssdev->device)
+		return;
+
+	omapdss_output_unset_device(dssdev);
+
+	if (dssdev->manager)
+		dss_mgr_disconnect(dssdev->manager, dssdev);
+}
+
+static const struct omapdss_dsi_ops dsi_ops = {
+	.connect = dsi_connect,
+	.disconnect = dsi_disconnect,
+
+	.bus_lock = dsi_bus_lock,
+	.bus_unlock = dsi_bus_unlock,
+
+	.enable = omapdss_dsi_display_enable,
+	.disable = omapdss_dsi_display_disable,
+
+	.enable_hs = omapdss_dsi_vc_enable_hs,
+
+	.configure_pins = omapdss_dsi_configure_pins,
+	.set_config = omapdss_dsi_set_config,
+
+	.enable_video_output = dsi_enable_video_output,
+	.disable_video_output = dsi_disable_video_output,
+
+	.update = omap_dsi_update,
+
+	.enable_te = omapdss_dsi_enable_te,
+
+	.request_vc = omap_dsi_request_vc,
+	.set_vc_id = omap_dsi_set_vc_id,
+	.release_vc = omap_dsi_release_vc,
+
+	.dcs_write = dsi_vc_dcs_write,
+	.dcs_write_nosync = dsi_vc_dcs_write_nosync,
+	.dcs_read = dsi_vc_dcs_read,
+
+	.gen_write = dsi_vc_generic_write,
+	.gen_write_nosync = dsi_vc_generic_write_nosync,
+	.gen_read = dsi_vc_generic_read,
+
+	.bta_sync = dsi_vc_send_bta_sync,
+
+	.set_max_rx_packet_size = dsi_vc_set_max_rx_packet_size,
+};
+
 static void dsi_init_output(struct platform_device *dsidev)
 {
 	struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev);
-	struct omap_dss_output *out = &dsi->output;
+	struct omap_dss_device *out = &dsi->output;
 
-	out->pdev = dsidev;
+	out->dev = &dsidev->dev;
 	out->id = dsi->module_id == 0 ?
 			OMAP_DSS_OUTPUT_DSI1 : OMAP_DSS_OUTPUT_DSI2;
 
-	out->type = OMAP_DISPLAY_TYPE_DSI;
+	out->output_type = OMAP_DISPLAY_TYPE_DSI;
 	out->name = dsi->module_id == 0 ? "dsi.0" : "dsi.1";
 	out->dispc_channel = dsi_get_channel(dsi->module_id);
+	out->ops.dsi = &dsi_ops;
+	out->owner = THIS_MODULE;
 
-	dss_register_output(out);
+	omapdss_register_output(out);
 }
 
 static void dsi_uninit_output(struct platform_device *dsidev)
 {
 	struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev);
-	struct omap_dss_output *out = &dsi->output;
+	struct omap_dss_device *out = &dsi->output;
 
-	dss_unregister_output(out);
+	omapdss_unregister_output(out);
 }
 
 /* DSI1 HW IP initialisation */
@@ -5563,12 +5627,10 @@ static int omap_dsihw_probe(struct platform_device *dsidev)
 
 	dsi_init_output(dsidev);
 
-	r = dsi_probe_pdata(dsidev);
-	if (r) {
-		dsi_runtime_put(dsidev);
-		dsi_uninit_output(dsidev);
-		pm_runtime_disable(&dsidev->dev);
-		return r;
+	if (dsidev->dev.platform_data) {
+		r = dsi_probe_pdata(dsidev);
+		if (r)
+			goto err_probe;
 	}
 
 	dsi_runtime_put(dsidev);
@@ -5586,6 +5648,9 @@ static int omap_dsihw_probe(struct platform_device *dsidev)
 #endif
 	return 0;
 
+err_probe:
+	dsi_runtime_put(dsidev);
+	dsi_uninit_output(dsidev);
 err_runtime_get:
 	pm_runtime_disable(&dsidev->dev);
 	return r;
@@ -5603,14 +5668,9 @@ static int __exit omap_dsihw_remove(struct platform_device *dsidev)
 
 	pm_runtime_disable(&dsidev->dev);
 
-	if (dsi->vdds_dsi_reg != NULL) {
-		if (dsi->vdds_dsi_enabled) {
-			regulator_disable(dsi->vdds_dsi_reg);
-			dsi->vdds_dsi_enabled = false;
-		}
-
-		regulator_put(dsi->vdds_dsi_reg);
-		dsi->vdds_dsi_reg = NULL;
+	if (dsi->vdds_dsi_reg != NULL && dsi->vdds_dsi_enabled) {
+		regulator_disable(dsi->vdds_dsi_reg);
+		dsi->vdds_dsi_enabled = false;
 	}
 
 	return 0;
diff --git a/drivers/video/omap2/dss/dss.c b/drivers/video/omap2/dss/dss.c
index 94f66f9..bd01608 100644
--- a/drivers/video/omap2/dss/dss.c
+++ b/drivers/video/omap2/dss/dss.c
@@ -157,7 +157,8 @@ static void dss_restore_context(void)
 
 int dss_get_ctx_loss_count(void)
 {
-	struct omap_dss_board_info *board_data = dss.pdev->dev.platform_data;
+	struct platform_device *core_pdev = dss_get_core_pdev();
+	struct omap_dss_board_info *board_data = core_pdev->dev.platform_data;
 	int cnt;
 
 	if (!board_data->get_context_loss_count)
diff --git a/drivers/video/omap2/dss/dss.h b/drivers/video/omap2/dss/dss.h
index 8475893..50a2362 100644
--- a/drivers/video/omap2/dss/dss.h
+++ b/drivers/video/omap2/dss/dss.h
@@ -179,23 +179,19 @@ void dss_put_device(struct omap_dss_device *dssdev);
 void dss_copy_device_pdata(struct omap_dss_device *dst,
 		const struct omap_dss_device *src);
 
-/* output */
-void dss_register_output(struct omap_dss_output *out);
-void dss_unregister_output(struct omap_dss_output *out);
-
 /* display */
 int dss_suspend_all_devices(void);
 int dss_resume_all_devices(void);
 void dss_disable_all_devices(void);
 
-int display_init_sysfs(struct platform_device *pdev,
-		struct omap_dss_device *dssdev);
-void display_uninit_sysfs(struct platform_device *pdev,
-		struct omap_dss_device *dssdev);
+int display_init_sysfs(struct platform_device *pdev);
+void display_uninit_sysfs(struct platform_device *pdev);
 
 /* manager */
-int dss_init_overlay_managers(struct platform_device *pdev);
-void dss_uninit_overlay_managers(struct platform_device *pdev);
+int dss_init_overlay_managers(void);
+void dss_uninit_overlay_managers(void);
+int dss_init_overlay_managers_sysfs(struct platform_device *pdev);
+void dss_uninit_overlay_managers_sysfs(struct platform_device *pdev);
 int dss_mgr_simple_check(struct omap_overlay_manager *mgr,
 		const struct omap_overlay_manager_info *info);
 int dss_mgr_check_timings(struct omap_overlay_manager *mgr,
@@ -426,6 +422,7 @@ void dispc_mgr_set_clock_div(enum omap_channel channel,
 		const struct dispc_clock_info *cinfo);
 int dispc_mgr_get_clock_div(enum omap_channel channel,
 		struct dispc_clock_info *cinfo);
+void dispc_set_tv_pclk(unsigned long pclk);
 
 u32 dispc_wb_get_framedone_irq(void);
 bool dispc_wb_go_busy(void);
@@ -437,17 +434,8 @@ int dispc_wb_setup(const struct omap_dss_writeback_info *wi,
 		bool mem_to_mem, const struct omap_video_timings *timings);
 
 /* VENC */
-#ifdef CONFIG_OMAP2_DSS_VENC
 int venc_init_platform_driver(void) __init;
 void venc_uninit_platform_driver(void) __exit;
-unsigned long venc_get_pixel_clock(void);
-#else
-static inline unsigned long venc_get_pixel_clock(void)
-{
-	WARN("%s: VENC not compiled in, returning pclk as 0\n", __func__);
-	return 0;
-}
-#endif
 int omapdss_venc_display_enable(struct omap_dss_device *dssdev);
 void omapdss_venc_display_disable(struct omap_dss_device *dssdev);
 void omapdss_venc_set_timings(struct omap_dss_device *dssdev,
@@ -464,17 +452,8 @@ int venc_panel_init(void);
 void venc_panel_exit(void);
 
 /* HDMI */
-#ifdef CONFIG_OMAP4_DSS_HDMI
 int hdmi_init_platform_driver(void) __init;
 void hdmi_uninit_platform_driver(void) __exit;
-unsigned long hdmi_get_pixel_clock(void);
-#else
-static inline unsigned long hdmi_get_pixel_clock(void)
-{
-	WARN("%s: HDMI not compiled in, returning pclk as 0\n", __func__);
-	return 0;
-}
-#endif
 int omapdss_hdmi_display_enable(struct omap_dss_device *dssdev);
 void omapdss_hdmi_display_disable(struct omap_dss_device *dssdev);
 int omapdss_hdmi_core_enable(struct omap_dss_device *dssdev);
diff --git a/drivers/video/omap2/dss/dss_features.c b/drivers/video/omap2/dss/dss_features.c
index 77dbe0c..b9cfebb 100644
--- a/drivers/video/omap2/dss/dss_features.c
+++ b/drivers/video/omap2/dss/dss_features.c
@@ -797,7 +797,6 @@ static const struct ti_hdmi_ip_ops omap4_hdmi_functions = {
 	.phy_enable		=	ti_hdmi_4xxx_phy_enable,
 	.phy_disable		=	ti_hdmi_4xxx_phy_disable,
 	.read_edid		=	ti_hdmi_4xxx_read_edid,
-	.detect			=	ti_hdmi_4xxx_detect,
 	.pll_enable		=	ti_hdmi_4xxx_pll_enable,
 	.pll_disable		=	ti_hdmi_4xxx_pll_disable,
 	.video_enable		=	ti_hdmi_4xxx_wp_video_start,
diff --git a/drivers/video/omap2/dss/hdmi.c b/drivers/video/omap2/dss/hdmi.c
index a109934..44a885b 100644
--- a/drivers/video/omap2/dss/hdmi.c
+++ b/drivers/video/omap2/dss/hdmi.c
@@ -70,7 +70,9 @@ static struct {
 	int ls_oe_gpio;
 	int hpd_gpio;
 
-	struct omap_dss_output output;
+	bool core_enabled;
+
+	struct omap_dss_device output;
 } hdmi;
 
 /*
@@ -328,6 +330,29 @@ static void hdmi_runtime_put(void)
 	WARN_ON(r < 0 && r != -ENOSYS);
 }
 
+static int hdmi_init_regulator(void)
+{
+	struct regulator *reg;
+
+	if (hdmi.vdda_hdmi_dac_reg != NULL)
+		return 0;
+
+	reg = devm_regulator_get(&hdmi.pdev->dev, "vdda_hdmi_dac");
+
+	/* DT HACK: try VDAC to make omapdss work for o4 sdp/panda */
+	if (IS_ERR(reg))
+		reg = devm_regulator_get(&hdmi.pdev->dev, "VDAC");
+
+	if (IS_ERR(reg)) {
+		DSSERR("can't get VDDA_HDMI_DAC regulator\n");
+		return PTR_ERR(reg);
+	}
+
+	hdmi.vdda_hdmi_dac_reg = reg;
+
+	return 0;
+}
+
 static int hdmi_init_display(struct omap_dss_device *dssdev)
 {
 	int r;
@@ -342,22 +367,9 @@ static int hdmi_init_display(struct omap_dss_device *dssdev)
 
 	dss_init_hdmi_ip_ops(&hdmi.ip_data, omapdss_get_version());
 
-	if (hdmi.vdda_hdmi_dac_reg == NULL) {
-		struct regulator *reg;
-
-		reg = devm_regulator_get(&hdmi.pdev->dev, "vdda_hdmi_dac");
-
-		/* DT HACK: try VDAC to make omapdss work for o4 sdp/panda */
-		if (IS_ERR(reg))
-			reg = devm_regulator_get(&hdmi.pdev->dev, "VDAC");
-
-		if (IS_ERR(reg)) {
-			DSSERR("can't get VDDA_HDMI_DAC regulator\n");
-			return PTR_ERR(reg);
-		}
-
-		hdmi.vdda_hdmi_dac_reg = reg;
-	}
+	r = hdmi_init_regulator();
+	if (r)
+		return r;
 
 	r = gpio_request_array(gpios, ARRAY_SIZE(gpios));
 	if (r)
@@ -455,12 +467,6 @@ end:	return cm;
 
 }
 
-unsigned long hdmi_get_pixel_clock(void)
-{
-	/* HDMI Pixel Clock in Mhz */
-	return hdmi.ip_data.cfg.timings.pixel_clock * 1000;
-}
-
 static void hdmi_compute_pll(struct omap_dss_device *dssdev, int phy,
 		struct hdmi_pll_info *pi)
 {
@@ -511,8 +517,10 @@ static int hdmi_power_on_core(struct omap_dss_device *dssdev)
 {
 	int r;
 
-	gpio_set_value(hdmi.ct_cp_hpd_gpio, 1);
-	gpio_set_value(hdmi.ls_oe_gpio, 1);
+	if (gpio_is_valid(hdmi.ct_cp_hpd_gpio))
+		gpio_set_value(hdmi.ct_cp_hpd_gpio, 1);
+	if (gpio_is_valid(hdmi.ls_oe_gpio))
+		gpio_set_value(hdmi.ls_oe_gpio, 1);
 
 	/* wait 300us after CT_CP_HPD for the 5V power output to reach 90% */
 	udelay(300);
@@ -528,29 +536,37 @@ static int hdmi_power_on_core(struct omap_dss_device *dssdev)
 	/* Make selection of HDMI in DSS */
 	dss_select_hdmi_venc_clk_source(DSS_HDMI_M_PCLK);
 
+	hdmi.core_enabled = true;
+
 	return 0;
 
 err_runtime_get:
 	regulator_disable(hdmi.vdda_hdmi_dac_reg);
 err_vdac_enable:
-	gpio_set_value(hdmi.ct_cp_hpd_gpio, 0);
-	gpio_set_value(hdmi.ls_oe_gpio, 0);
+	if (gpio_is_valid(hdmi.ct_cp_hpd_gpio))
+		gpio_set_value(hdmi.ct_cp_hpd_gpio, 0);
+	if (gpio_is_valid(hdmi.ls_oe_gpio))
+		gpio_set_value(hdmi.ls_oe_gpio, 0);
 	return r;
 }
 
 static void hdmi_power_off_core(struct omap_dss_device *dssdev)
 {
+	hdmi.core_enabled = false;
+
 	hdmi_runtime_put();
 	regulator_disable(hdmi.vdda_hdmi_dac_reg);
-	gpio_set_value(hdmi.ct_cp_hpd_gpio, 0);
-	gpio_set_value(hdmi.ls_oe_gpio, 0);
+	if (gpio_is_valid(hdmi.ct_cp_hpd_gpio))
+		gpio_set_value(hdmi.ct_cp_hpd_gpio, 0);
+	if (gpio_is_valid(hdmi.ls_oe_gpio))
+		gpio_set_value(hdmi.ls_oe_gpio, 0);
 }
 
 static int hdmi_power_on_full(struct omap_dss_device *dssdev)
 {
 	int r;
 	struct omap_video_timings *p;
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = hdmi.output.manager;
 	unsigned long phy;
 
 	r = hdmi_power_on_core(dssdev);
@@ -613,7 +629,7 @@ err_pll_enable:
 
 static void hdmi_power_off_full(struct omap_dss_device *dssdev)
 {
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = hdmi.output.manager;
 
 	dss_mgr_disable(mgr);
 
@@ -653,9 +669,23 @@ void omapdss_hdmi_display_set_timing(struct omap_dss_device *dssdev,
 	if (t != NULL)
 		hdmi.ip_data.cfg = *t;
 
+	dispc_set_tv_pclk(t->timings.pixel_clock * 1000);
+
 	mutex_unlock(&hdmi.lock);
 }
 
+static void omapdss_hdmi_display_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	const struct hdmi_config *cfg;
+
+	cfg = hdmi_get_timings();
+	if (cfg == NULL)
+		cfg = &vesa_timings[0];
+
+	memcpy(timings, &cfg->timings, sizeof(cfg->timings));
+}
+
 static void hdmi_dump_regs(struct seq_file *s)
 {
 	mutex_lock(&hdmi.lock);
@@ -700,7 +730,7 @@ bool omapdss_hdmi_detect(void)
 	r = hdmi_runtime_get();
 	BUG_ON(r);
 
-	r = hdmi.ip_data.ops->detect(&hdmi.ip_data);
+	r = gpio_get_value(hdmi.hpd_gpio);
 
 	hdmi_runtime_put();
 	mutex_unlock(&hdmi.lock);
@@ -710,7 +740,7 @@ bool omapdss_hdmi_detect(void)
 
 int omapdss_hdmi_display_enable(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_output *out = dssdev->output;
+	struct omap_dss_device *out = &hdmi.output;
 	int r = 0;
 
 	DSSDBG("ENTER hdmi_display_enable\n");
@@ -723,25 +753,15 @@ int omapdss_hdmi_display_enable(struct omap_dss_device *dssdev)
 		goto err0;
 	}
 
-	hdmi.ip_data.hpd_gpio = hdmi.hpd_gpio;
-
-	r = omap_dss_start_device(dssdev);
-	if (r) {
-		DSSERR("failed to start device\n");
-		goto err0;
-	}
-
 	r = hdmi_power_on_full(dssdev);
 	if (r) {
 		DSSERR("failed to power on device\n");
-		goto err1;
+		goto err0;
 	}
 
 	mutex_unlock(&hdmi.lock);
 	return 0;
 
-err1:
-	omap_dss_stop_device(dssdev);
 err0:
 	mutex_unlock(&hdmi.lock);
 	return r;
@@ -755,8 +775,6 @@ void omapdss_hdmi_display_disable(struct omap_dss_device *dssdev)
 
 	hdmi_power_off_full(dssdev);
 
-	omap_dss_stop_device(dssdev);
-
 	mutex_unlock(&hdmi.lock);
 }
 
@@ -768,8 +786,6 @@ int omapdss_hdmi_core_enable(struct omap_dss_device *dssdev)
 
 	mutex_lock(&hdmi.lock);
 
-	hdmi.ip_data.hpd_gpio = hdmi.hpd_gpio;
-
 	r = hdmi_power_on_core(dssdev);
 	if (r) {
 		DSSERR("failed to power on device\n");
@@ -1033,24 +1049,219 @@ static int hdmi_probe_pdata(struct platform_device *pdev)
 	return 0;
 }
 
+static int hdmi_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct omap_overlay_manager *mgr;
+	int r;
+
+	dss_init_hdmi_ip_ops(&hdmi.ip_data, omapdss_get_version());
+
+	r = hdmi_init_regulator();
+	if (r)
+		return r;
+
+	mgr = omap_dss_get_overlay_manager(dssdev->dispc_channel);
+	if (!mgr)
+		return -ENODEV;
+
+	r = dss_mgr_connect(mgr, dssdev);
+	if (r)
+		return r;
+
+	r = omapdss_output_set_device(dssdev, dst);
+	if (r) {
+		DSSERR("failed to connect output to new device: %s\n",
+				dst->name);
+		dss_mgr_disconnect(mgr, dssdev);
+		return r;
+	}
+
+	return 0;
+}
+
+static void hdmi_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	WARN_ON(dst != dssdev->device);
+
+	if (dst != dssdev->device)
+		return;
+
+	omapdss_output_unset_device(dssdev);
+
+	if (dssdev->manager)
+		dss_mgr_disconnect(dssdev->manager, dssdev);
+}
+
+static int hdmi_read_edid(struct omap_dss_device *dssdev,
+		u8 *edid, int len)
+{
+	bool need_enable;
+	int r;
+
+	need_enable = hdmi.core_enabled == false;
+
+	if (need_enable) {
+		r = omapdss_hdmi_core_enable(dssdev);
+		if (r)
+			return r;
+	}
+
+	r = omapdss_hdmi_read_edid(edid, len);
+
+	if (need_enable)
+		omapdss_hdmi_core_disable(dssdev);
+
+	return r;
+}
+
+#if defined(CONFIG_OMAP4_DSS_HDMI_AUDIO)
+static int omapdss_hdmi_audio_enable(struct omap_dss_device *dssdev)
+{
+	int r;
+
+	mutex_lock(&hdmi.lock);
+
+	if (!hdmi_mode_has_audio()) {
+		r = -EPERM;
+		goto err;
+	}
+
+	r = hdmi_audio_enable();
+	if (r)
+		goto err;
+
+	mutex_unlock(&hdmi.lock);
+	return 0;
+
+err:
+	mutex_unlock(&hdmi.lock);
+	return r;
+}
+
+static void omapdss_hdmi_audio_disable(struct omap_dss_device *dssdev)
+{
+	hdmi_audio_disable();
+}
+
+static int omapdss_hdmi_audio_start(struct omap_dss_device *dssdev)
+{
+	return hdmi_audio_start();
+}
+
+static void omapdss_hdmi_audio_stop(struct omap_dss_device *dssdev)
+{
+	hdmi_audio_stop();
+}
+
+static bool omapdss_hdmi_audio_supported(struct omap_dss_device *dssdev)
+{
+	bool r;
+
+	mutex_lock(&hdmi.lock);
+
+	r = hdmi_mode_has_audio();
+
+	mutex_unlock(&hdmi.lock);
+	return r;
+}
+
+static int omapdss_hdmi_audio_config(struct omap_dss_device *dssdev,
+		struct omap_dss_audio *audio)
+{
+	int r;
+
+	mutex_lock(&hdmi.lock);
+
+	if (!hdmi_mode_has_audio()) {
+		r = -EPERM;
+		goto err;
+	}
+
+	r = hdmi_audio_config(audio);
+	if (r)
+		goto err;
+
+	mutex_unlock(&hdmi.lock);
+	return 0;
+
+err:
+	mutex_unlock(&hdmi.lock);
+	return r;
+}
+#else
+static int omapdss_hdmi_audio_enable(struct omap_dss_device *dssdev)
+{
+	return -EPERM;
+}
+
+static void omapdss_hdmi_audio_disable(struct omap_dss_device *dssdev)
+{
+}
+
+static int omapdss_hdmi_audio_start(struct omap_dss_device *dssdev)
+{
+	return -EPERM;
+}
+
+static void omapdss_hdmi_audio_stop(struct omap_dss_device *dssdev)
+{
+}
+
+static bool omapdss_hdmi_audio_supported(struct omap_dss_device *dssdev)
+{
+	return false;
+}
+
+static int omapdss_hdmi_audio_config(struct omap_dss_device *dssdev,
+		struct omap_dss_audio *audio)
+{
+	return -EPERM;
+}
+#endif
+
+static const struct omapdss_hdmi_ops hdmi_ops = {
+	.connect		= hdmi_connect,
+	.disconnect		= hdmi_disconnect,
+
+	.enable			= omapdss_hdmi_display_enable,
+	.disable		= omapdss_hdmi_display_disable,
+
+	.check_timings		= omapdss_hdmi_display_check_timing,
+	.set_timings		= omapdss_hdmi_display_set_timing,
+	.get_timings		= omapdss_hdmi_display_get_timings,
+
+	.read_edid		= hdmi_read_edid,
+
+	.audio_enable		= omapdss_hdmi_audio_enable,
+	.audio_disable		= omapdss_hdmi_audio_disable,
+	.audio_start		= omapdss_hdmi_audio_start,
+	.audio_stop		= omapdss_hdmi_audio_stop,
+	.audio_supported	= omapdss_hdmi_audio_supported,
+	.audio_config		= omapdss_hdmi_audio_config,
+};
+
 static void hdmi_init_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &hdmi.output;
+	struct omap_dss_device *out = &hdmi.output;
 
-	out->pdev = pdev;
+	out->dev = &pdev->dev;
 	out->id = OMAP_DSS_OUTPUT_HDMI;
-	out->type = OMAP_DISPLAY_TYPE_HDMI;
+	out->output_type = OMAP_DISPLAY_TYPE_HDMI;
 	out->name = "hdmi.0";
 	out->dispc_channel = OMAP_DSS_CHANNEL_DIGIT;
+	out->ops.hdmi = &hdmi_ops;
+	out->owner = THIS_MODULE;
 
-	dss_register_output(out);
+	omapdss_register_output(out);
 }
 
 static void __exit hdmi_uninit_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &hdmi.output;
+	struct omap_dss_device *out = &hdmi.output;
 
-	dss_unregister_output(out);
+	omapdss_unregister_output(out);
 }
 
 /* HDMI HW IP initialisation */
@@ -1071,6 +1282,12 @@ static int omapdss_hdmihw_probe(struct platform_device *pdev)
 	if (IS_ERR(hdmi.ip_data.base_wp))
 		return PTR_ERR(hdmi.ip_data.base_wp);
 
+	hdmi.ip_data.irq = platform_get_irq(pdev, 0);
+	if (hdmi.ip_data.irq < 0) {
+		DSSERR("platform_get_irq failed\n");
+		return -ENODEV;
+	}
+
 	r = hdmi_get_clocks(pdev);
 	if (r) {
 		DSSERR("can't get clocks\n");
@@ -1084,6 +1301,10 @@ static int omapdss_hdmihw_probe(struct platform_device *pdev)
 	hdmi.ip_data.pll_offset = HDMI_PLLCTRL;
 	hdmi.ip_data.phy_offset = HDMI_PHY;
 
+	hdmi.ct_cp_hpd_gpio = -1;
+	hdmi.ls_oe_gpio = -1;
+	hdmi.hpd_gpio = -1;
+
 	hdmi_init_output(pdev);
 
 	r = hdmi_panel_init();
@@ -1094,15 +1315,19 @@ static int omapdss_hdmihw_probe(struct platform_device *pdev)
 
 	dss_debugfs_create_file("hdmi", hdmi_dump_regs);
 
-	r = hdmi_probe_pdata(pdev);
-	if (r) {
-		hdmi_panel_exit();
-		hdmi_uninit_output(pdev);
-		pm_runtime_disable(&pdev->dev);
-		return r;
+	if (pdev->dev.platform_data) {
+		r = hdmi_probe_pdata(pdev);
+		if (r)
+			goto err_probe;
 	}
 
 	return 0;
+
+err_probe:
+	hdmi_panel_exit();
+	hdmi_uninit_output(pdev);
+	pm_runtime_disable(&pdev->dev);
+	return r;
 }
 
 static int __exit hdmi_remove_child(struct device *dev, void *data)
diff --git a/drivers/video/omap2/dss/manager-sysfs.c b/drivers/video/omap2/dss/manager-sysfs.c
index 9a2fb59..de7e7b5 100644
--- a/drivers/video/omap2/dss/manager-sysfs.c
+++ b/drivers/video/omap2/dss/manager-sysfs.c
@@ -50,6 +50,7 @@ static ssize_t manager_display_store(struct omap_overlay_manager *mgr,
 	int r = 0;
 	size_t len = size;
 	struct omap_dss_device *dssdev = NULL;
+	struct omap_dss_device *old_dssdev;
 
 	int match(struct omap_dss_device *dssdev, void *data)
 	{
@@ -66,32 +67,44 @@ static ssize_t manager_display_store(struct omap_overlay_manager *mgr,
 	if (len > 0 && dssdev == NULL)
 		return -EINVAL;
 
-	if (dssdev)
+	if (dssdev) {
 		DSSDBG("display %s found\n", dssdev->name);
 
-	if (mgr->output) {
-		r = mgr->unset_output(mgr);
-		if (r) {
-			DSSERR("failed to unset current output\n");
+		if (omapdss_device_is_connected(dssdev)) {
+			DSSERR("new display is already connected\n");
+			r = -EINVAL;
+			goto put_device;
+		}
+
+		if (omapdss_device_is_enabled(dssdev)) {
+			DSSERR("new display is not disabled\n");
+			r = -EINVAL;
 			goto put_device;
 		}
 	}
 
-	if (dssdev) {
-		struct omap_dss_output *out = dssdev->output;
-
-		/*
-		 * a registered device should have an output connected to it
-		 * already
-		 */
-		if (!out) {
-			DSSERR("device has no output connected to it\n");
+	old_dssdev = mgr->get_device(mgr);
+	if (old_dssdev) {
+		if (omapdss_device_is_enabled(old_dssdev)) {
+			DSSERR("old display is not disabled\n");
+			r = -EINVAL;
 			goto put_device;
 		}
 
-		r = mgr->set_output(mgr, out);
+		old_dssdev->driver->disconnect(old_dssdev);
+	}
+
+	if (dssdev) {
+		r = dssdev->driver->connect(dssdev);
 		if (r) {
-			DSSERR("failed to set manager output\n");
+			DSSERR("failed to connect new device\n");
+			goto put_device;
+		}
+
+		old_dssdev = mgr->get_device(mgr);
+		if (old_dssdev != dssdev) {
+			DSSERR("failed to connect device to this manager\n");
+			dssdev->driver->disconnect(dssdev);
 			goto put_device;
 		}
 
@@ -509,4 +522,6 @@ void dss_manager_kobj_uninit(struct omap_overlay_manager *mgr)
 {
 	kobject_del(&mgr->kobj);
 	kobject_put(&mgr->kobj);
+
+	memset(&mgr->kobj, 0, sizeof(mgr->kobj));
 }
diff --git a/drivers/video/omap2/dss/manager.c b/drivers/video/omap2/dss/manager.c
index 2551eaa..1aac9b4 100644
--- a/drivers/video/omap2/dss/manager.c
+++ b/drivers/video/omap2/dss/manager.c
@@ -36,9 +36,9 @@
 static int num_managers;
 static struct omap_overlay_manager *managers;
 
-int dss_init_overlay_managers(struct platform_device *pdev)
+int dss_init_overlay_managers(void)
 {
-	int i, r;
+	int i;
 
 	num_managers = dss_feat_get_num_mgrs();
 
@@ -76,6 +76,17 @@ int dss_init_overlay_managers(struct platform_device *pdev)
 			dss_feat_get_supported_outputs(mgr->id);
 
 		INIT_LIST_HEAD(&mgr->overlays);
+	}
+
+	return 0;
+}
+
+int dss_init_overlay_managers_sysfs(struct platform_device *pdev)
+{
+	int i, r;
+
+	for (i = 0; i < num_managers; ++i) {
+		struct omap_overlay_manager *mgr = &managers[i];
 
 		r = dss_manager_kobj_init(mgr, pdev);
 		if (r)
@@ -85,18 +96,22 @@ int dss_init_overlay_managers(struct platform_device *pdev)
 	return 0;
 }
 
-void dss_uninit_overlay_managers(struct platform_device *pdev)
+void dss_uninit_overlay_managers(void)
+{
+	kfree(managers);
+	managers = NULL;
+	num_managers = 0;
+}
+
+void dss_uninit_overlay_managers_sysfs(struct platform_device *pdev)
 {
 	int i;
 
 	for (i = 0; i < num_managers; ++i) {
 		struct omap_overlay_manager *mgr = &managers[i];
+
 		dss_manager_kobj_uninit(mgr);
 	}
-
-	kfree(managers);
-	managers = NULL;
-	num_managers = 0;
 }
 
 int omap_dss_get_num_overlay_managers(void)
diff --git a/drivers/video/omap2/dss/output.c b/drivers/video/omap2/dss/output.c
index 5214df6..3f5c0a7 100644
--- a/drivers/video/omap2/dss/output.c
+++ b/drivers/video/omap2/dss/output.c
@@ -27,7 +27,7 @@
 static LIST_HEAD(output_list);
 static DEFINE_MUTEX(output_lock);
 
-int omapdss_output_set_device(struct omap_dss_output *out,
+int omapdss_output_set_device(struct omap_dss_device *out,
 		struct omap_dss_device *dssdev)
 {
 	int r;
@@ -41,7 +41,7 @@ int omapdss_output_set_device(struct omap_dss_output *out,
 		goto err;
 	}
 
-	if (out->type != dssdev->type) {
+	if (out->output_type != dssdev->type) {
 		DSSERR("output type and display type don't match\n");
 		r = -EINVAL;
 		goto err;
@@ -60,7 +60,7 @@ err:
 }
 EXPORT_SYMBOL(omapdss_output_set_device);
 
-int omapdss_output_unset_device(struct omap_dss_output *out)
+int omapdss_output_unset_device(struct omap_dss_device *out)
 {
 	int r;
 
@@ -92,19 +92,22 @@ err:
 }
 EXPORT_SYMBOL(omapdss_output_unset_device);
 
-void dss_register_output(struct omap_dss_output *out)
+int omapdss_register_output(struct omap_dss_device *out)
 {
 	list_add_tail(&out->list, &output_list);
+	return 0;
 }
+EXPORT_SYMBOL(omapdss_register_output);
 
-void dss_unregister_output(struct omap_dss_output *out)
+void omapdss_unregister_output(struct omap_dss_device *out)
 {
 	list_del(&out->list);
 }
+EXPORT_SYMBOL(omapdss_unregister_output);
 
-struct omap_dss_output *omap_dss_get_output(enum omap_dss_output_id id)
+struct omap_dss_device *omap_dss_get_output(enum omap_dss_output_id id)
 {
-	struct omap_dss_output *out;
+	struct omap_dss_device *out;
 
 	list_for_each_entry(out, &output_list, list) {
 		if (out->id == id)
@@ -115,6 +118,62 @@ struct omap_dss_output *omap_dss_get_output(enum omap_dss_output_id id)
 }
 EXPORT_SYMBOL(omap_dss_get_output);
 
+struct omap_dss_device *omap_dss_find_output(const char *name)
+{
+	struct omap_dss_device *out;
+
+	list_for_each_entry(out, &output_list, list) {
+		if (strcmp(out->name, name) == 0)
+			return omap_dss_get_device(out);
+	}
+
+	return NULL;
+}
+EXPORT_SYMBOL(omap_dss_find_output);
+
+struct omap_dss_device *omap_dss_find_output_by_node(struct device_node *node)
+{
+	struct omap_dss_device *out;
+
+	list_for_each_entry(out, &output_list, list) {
+		if (out->dev->of_node == node)
+			return omap_dss_get_device(out);
+	}
+
+	return NULL;
+}
+EXPORT_SYMBOL(omap_dss_find_output_by_node);
+
+struct omap_dss_device *omapdss_find_output_from_display(struct omap_dss_device *dssdev)
+{
+	while (dssdev->output)
+		dssdev = dssdev->output;
+
+	if (dssdev->id != 0)
+		return omap_dss_get_device(dssdev);
+
+	return NULL;
+}
+EXPORT_SYMBOL(omapdss_find_output_from_display);
+
+struct omap_overlay_manager *omapdss_find_mgr_from_display(struct omap_dss_device *dssdev)
+{
+	struct omap_dss_device *out;
+	struct omap_overlay_manager *mgr;
+
+	out = omapdss_find_output_from_display(dssdev);
+
+	if (out == NULL)
+		return NULL;
+
+	mgr = out->manager;
+
+	omap_dss_put_device(out);
+
+	return mgr;
+}
+EXPORT_SYMBOL(omapdss_find_mgr_from_display);
+
 static const struct dss_mgr_ops *dss_mgr_ops;
 
 int dss_install_mgr_ops(const struct dss_mgr_ops *mgr_ops)
@@ -134,6 +193,20 @@ void dss_uninstall_mgr_ops(void)
 }
 EXPORT_SYMBOL(dss_uninstall_mgr_ops);
 
+int dss_mgr_connect(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	return dss_mgr_ops->connect(mgr, dst);
+}
+EXPORT_SYMBOL(dss_mgr_connect);
+
+void dss_mgr_disconnect(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	dss_mgr_ops->disconnect(mgr, dst);
+}
+EXPORT_SYMBOL(dss_mgr_disconnect);
+
 void dss_mgr_set_timings(struct omap_overlay_manager *mgr,
 		const struct omap_video_timings *timings)
 {
diff --git a/drivers/video/omap2/dss/rfbi.c b/drivers/video/omap2/dss/rfbi.c
index 1a17dd1..fdfe6e6 100644
--- a/drivers/video/omap2/dss/rfbi.c
+++ b/drivers/video/omap2/dss/rfbi.c
@@ -117,7 +117,7 @@ static struct {
 	int data_lines;
 	struct rfbi_timings intf_timings;
 
-	struct omap_dss_output output;
+	struct omap_dss_device output;
 } rfbi;
 
 static inline void rfbi_write_reg(const struct rfbi_reg idx, u32 val)
@@ -312,7 +312,7 @@ static int rfbi_transfer_area(struct omap_dss_device *dssdev,
 {
 	u32 l;
 	int r;
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = rfbi.output.manager;
 	u16 width = rfbi.timings.x_res;
 	u16 height = rfbi.timings.y_res;
 
@@ -852,7 +852,7 @@ static void rfbi_dump_regs(struct seq_file *s)
 
 static void rfbi_config_lcd_manager(struct omap_dss_device *dssdev)
 {
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = rfbi.output.manager;
 	struct dss_lcd_mgr_config mgr_config;
 
 	mgr_config.io_pad_mode = DSS_IO_PAD_MODE_RFBI;
@@ -890,7 +890,7 @@ static void rfbi_config_lcd_manager(struct omap_dss_device *dssdev)
 
 int omapdss_rfbi_display_enable(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_output *out = dssdev->output;
+	struct omap_dss_device *out = &rfbi.output;
 	int r;
 
 	if (out == NULL || out->manager == NULL) {
@@ -902,12 +902,6 @@ int omapdss_rfbi_display_enable(struct omap_dss_device *dssdev)
 	if (r)
 		return r;
 
-	r = omap_dss_start_device(dssdev);
-	if (r) {
-		DSSERR("failed to start device\n");
-		goto err0;
-	}
-
 	r = dss_mgr_register_framedone_handler(out->manager,
 			framedone_callback, NULL);
 	if (r) {
@@ -924,8 +918,6 @@ int omapdss_rfbi_display_enable(struct omap_dss_device *dssdev)
 
 	return 0;
 err1:
-	omap_dss_stop_device(dssdev);
-err0:
 	rfbi_runtime_put();
 	return r;
 }
@@ -933,11 +925,10 @@ EXPORT_SYMBOL(omapdss_rfbi_display_enable);
 
 void omapdss_rfbi_display_disable(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_output *out = dssdev->output;
+	struct omap_dss_device *out = &rfbi.output;
 
 	dss_mgr_unregister_framedone_handler(out->manager,
 			framedone_callback, NULL);
-	omap_dss_stop_device(dssdev);
 
 	rfbi_runtime_put();
 }
@@ -1022,22 +1013,23 @@ static int rfbi_probe_pdata(struct platform_device *rfbidev)
 
 static void rfbi_init_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &rfbi.output;
+	struct omap_dss_device *out = &rfbi.output;
 
-	out->pdev = pdev;
+	out->dev = &pdev->dev;
 	out->id = OMAP_DSS_OUTPUT_DBI;
-	out->type = OMAP_DISPLAY_TYPE_DBI;
+	out->output_type = OMAP_DISPLAY_TYPE_DBI;
 	out->name = "rfbi.0";
 	out->dispc_channel = OMAP_DSS_CHANNEL_LCD;
+	out->owner = THIS_MODULE;
 
-	dss_register_output(out);
+	omapdss_register_output(out);
 }
 
 static void __exit rfbi_uninit_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &rfbi.output;
+	struct omap_dss_device *out = &rfbi.output;
 
-	dss_unregister_output(out);
+	omapdss_unregister_output(out);
 }
 
 /* RFBI HW IP initialisation */
@@ -1093,15 +1085,16 @@ static int omap_rfbihw_probe(struct platform_device *pdev)
 
 	rfbi_init_output(pdev);
 
-	r = rfbi_probe_pdata(pdev);
-	if (r) {
-		rfbi_uninit_output(pdev);
-		pm_runtime_disable(&pdev->dev);
-		return r;
+	if (pdev->dev.platform_data) {
+		r = rfbi_probe_pdata(pdev);
+		if (r)
+			goto err_probe;
 	}
 
 	return 0;
 
+err_probe:
+	rfbi_uninit_output(pdev);
 err_runtime_get:
 	pm_runtime_disable(&pdev->dev);
 	return r;
diff --git a/drivers/video/omap2/dss/sdi.c b/drivers/video/omap2/dss/sdi.c
index 0bcd302..856af2e 100644
--- a/drivers/video/omap2/dss/sdi.c
+++ b/drivers/video/omap2/dss/sdi.c
@@ -31,6 +31,8 @@
 #include "dss.h"
 
 static struct {
+	struct platform_device *pdev;
+
 	bool update_enabled;
 	struct regulator *vdds_sdi_reg;
 
@@ -38,7 +40,7 @@ static struct {
 	struct omap_video_timings timings;
 	int datapairs;
 
-	struct omap_dss_output output;
+	struct omap_dss_device output;
 } sdi;
 
 struct sdi_clk_calc_ctx {
@@ -109,7 +111,7 @@ static int sdi_calc_clock_div(unsigned long pclk,
 
 static void sdi_config_lcd_manager(struct omap_dss_device *dssdev)
 {
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = sdi.output.manager;
 
 	sdi.mgr_config.io_pad_mode = DSS_IO_PAD_MODE_BYPASS;
 
@@ -124,7 +126,7 @@ static void sdi_config_lcd_manager(struct omap_dss_device *dssdev)
 
 int omapdss_sdi_display_enable(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_output *out = dssdev->output;
+	struct omap_dss_device *out = &sdi.output;
 	struct omap_video_timings *t = &sdi.timings;
 	struct dss_clock_info dss_cinfo;
 	struct dispc_clock_info dispc_cinfo;
@@ -136,12 +138,6 @@ int omapdss_sdi_display_enable(struct omap_dss_device *dssdev)
 		return -ENODEV;
 	}
 
-	r = omap_dss_start_device(dssdev);
-	if (r) {
-		DSSERR("failed to start device\n");
-		goto err_start_dev;
-	}
-
 	r = regulator_enable(sdi.vdds_sdi_reg);
 	if (r)
 		goto err_reg_enable;
@@ -213,15 +209,13 @@ err_calc_clock_div:
 err_get_dispc:
 	regulator_disable(sdi.vdds_sdi_reg);
 err_reg_enable:
-	omap_dss_stop_device(dssdev);
-err_start_dev:
 	return r;
 }
 EXPORT_SYMBOL(omapdss_sdi_display_enable);
 
 void omapdss_sdi_display_disable(struct omap_dss_device *dssdev)
 {
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = sdi.output.manager;
 
 	dss_mgr_disable(mgr);
 
@@ -230,8 +224,6 @@ void omapdss_sdi_display_disable(struct omap_dss_device *dssdev)
 	dispc_runtime_put();
 
 	regulator_disable(sdi.vdds_sdi_reg);
-
-	omap_dss_stop_device(dssdev);
 }
 EXPORT_SYMBOL(omapdss_sdi_display_disable);
 
@@ -242,29 +234,51 @@ void omapdss_sdi_set_timings(struct omap_dss_device *dssdev,
 }
 EXPORT_SYMBOL(omapdss_sdi_set_timings);
 
+static void sdi_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	*timings = sdi.timings;
+}
+
+static int sdi_check_timings(struct omap_dss_device *dssdev,
+			struct omap_video_timings *timings)
+{
+	struct omap_overlay_manager *mgr = sdi.output.manager;
+
+	if (mgr && !dispc_mgr_timings_ok(mgr->id, timings))
+		return -EINVAL;
+
+	if (timings->pixel_clock == 0)
+		return -EINVAL;
+
+	return 0;
+}
+
 void omapdss_sdi_set_datapairs(struct omap_dss_device *dssdev, int datapairs)
 {
 	sdi.datapairs = datapairs;
 }
 EXPORT_SYMBOL(omapdss_sdi_set_datapairs);
 
-static int sdi_init_display(struct omap_dss_device *dssdev)
+static int sdi_init_regulator(void)
 {
-	DSSDBG("SDI init\n");
+	struct regulator *vdds_sdi;
 
-	if (sdi.vdds_sdi_reg == NULL) {
-		struct regulator *vdds_sdi;
+	if (sdi.vdds_sdi_reg)
+		return 0;
 
-		vdds_sdi = dss_get_vdds_sdi();
+	vdds_sdi = dss_get_vdds_sdi();
 
+	if (IS_ERR(vdds_sdi)) {
+		vdds_sdi = devm_regulator_get(&sdi.pdev->dev, "vdds_sdi");
 		if (IS_ERR(vdds_sdi)) {
 			DSSERR("can't get VDDS_SDI regulator\n");
 			return PTR_ERR(vdds_sdi);
 		}
-
-		sdi.vdds_sdi_reg = vdds_sdi;
 	}
 
+	sdi.vdds_sdi_reg = vdds_sdi;
+
 	return 0;
 }
 
@@ -313,7 +327,7 @@ static int sdi_probe_pdata(struct platform_device *sdidev)
 
 	dss_copy_device_pdata(dssdev, plat_dssdev);
 
-	r = sdi_init_display(dssdev);
+	r = sdi_init_regulator();
 	if (r) {
 		DSSERR("device %s init failed: %d\n", dssdev->name, r);
 		dss_put_device(dssdev);
@@ -339,39 +353,104 @@ static int sdi_probe_pdata(struct platform_device *sdidev)
 	return 0;
 }
 
+static int sdi_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct omap_overlay_manager *mgr;
+	int r;
+
+	r = sdi_init_regulator();
+	if (r)
+		return r;
+
+	mgr = omap_dss_get_overlay_manager(dssdev->dispc_channel);
+	if (!mgr)
+		return -ENODEV;
+
+	r = dss_mgr_connect(mgr, dssdev);
+	if (r)
+		return r;
+
+	r = omapdss_output_set_device(dssdev, dst);
+	if (r) {
+		DSSERR("failed to connect output to new device: %s\n",
+				dst->name);
+		dss_mgr_disconnect(mgr, dssdev);
+		return r;
+	}
+
+	return 0;
+}
+
+static void sdi_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	WARN_ON(dst != dssdev->device);
+
+	if (dst != dssdev->device)
+		return;
+
+	omapdss_output_unset_device(dssdev);
+
+	if (dssdev->manager)
+		dss_mgr_disconnect(dssdev->manager, dssdev);
+}
+
+static const struct omapdss_sdi_ops sdi_ops = {
+	.connect = sdi_connect,
+	.disconnect = sdi_disconnect,
+
+	.enable = omapdss_sdi_display_enable,
+	.disable = omapdss_sdi_display_disable,
+
+	.check_timings = sdi_check_timings,
+	.set_timings = omapdss_sdi_set_timings,
+	.get_timings = sdi_get_timings,
+
+	.set_datapairs = omapdss_sdi_set_datapairs,
+};
+
 static void sdi_init_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &sdi.output;
+	struct omap_dss_device *out = &sdi.output;
 
-	out->pdev = pdev;
+	out->dev = &pdev->dev;
 	out->id = OMAP_DSS_OUTPUT_SDI;
-	out->type = OMAP_DISPLAY_TYPE_SDI;
+	out->output_type = OMAP_DISPLAY_TYPE_SDI;
 	out->name = "sdi.0";
 	out->dispc_channel = OMAP_DSS_CHANNEL_LCD;
+	out->ops.sdi = &sdi_ops;
+	out->owner = THIS_MODULE;
 
-	dss_register_output(out);
+	omapdss_register_output(out);
 }
 
 static void __exit sdi_uninit_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &sdi.output;
+	struct omap_dss_device *out = &sdi.output;
 
-	dss_unregister_output(out);
+	omapdss_unregister_output(out);
 }
 
 static int omap_sdi_probe(struct platform_device *pdev)
 {
 	int r;
 
+	sdi.pdev = pdev;
+
 	sdi_init_output(pdev);
 
-	r = sdi_probe_pdata(pdev);
-	if (r) {
-		sdi_uninit_output(pdev);
-		return r;
+	if (pdev->dev.platform_data) {
+		r = sdi_probe_pdata(pdev);
+		if (r)
+			goto err_probe;
 	}
 
 	return 0;
+
+err_probe:
+	sdi_uninit_output(pdev);
+	return r;
 }
 
 static int __exit omap_sdi_remove(struct platform_device *pdev)
diff --git a/drivers/video/omap2/dss/ti_hdmi.h b/drivers/video/omap2/dss/ti_hdmi.h
index 216aa70..45215f4 100644
--- a/drivers/video/omap2/dss/ti_hdmi.h
+++ b/drivers/video/omap2/dss/ti_hdmi.h
@@ -73,8 +73,6 @@ struct ti_hdmi_ip_ops {
 
 	int (*read_edid)(struct hdmi_ip_data *ip_data, u8 *edid, int len);
 
-	bool (*detect)(struct hdmi_ip_data *ip_data);
-
 	int (*pll_enable)(struct hdmi_ip_data *ip_data);
 
 	void (*pll_disable)(struct hdmi_ip_data *ip_data);
@@ -155,19 +153,18 @@ struct hdmi_ip_data {
 	unsigned long	core_av_offset;
 	unsigned long	pll_offset;
 	unsigned long	phy_offset;
+	int		irq;
 	const struct ti_hdmi_ip_ops *ops;
 	struct hdmi_config cfg;
 	struct hdmi_pll_info pll_data;
 	struct hdmi_core_infoframe_avi avi_cfg;
 
 	/* ti_hdmi_4xxx_ip private data. These should be in a separate struct */
-	int hpd_gpio;
 	struct mutex lock;
 };
 int ti_hdmi_4xxx_phy_enable(struct hdmi_ip_data *ip_data);
 void ti_hdmi_4xxx_phy_disable(struct hdmi_ip_data *ip_data);
 int ti_hdmi_4xxx_read_edid(struct hdmi_ip_data *ip_data, u8 *edid, int len);
-bool ti_hdmi_4xxx_detect(struct hdmi_ip_data *ip_data);
 int ti_hdmi_4xxx_wp_video_start(struct hdmi_ip_data *ip_data);
 void ti_hdmi_4xxx_wp_video_stop(struct hdmi_ip_data *ip_data);
 int ti_hdmi_4xxx_pll_enable(struct hdmi_ip_data *ip_data);
diff --git a/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.c b/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.c
index e18b222..e242ed8 100644
--- a/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.c
+++ b/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.c
@@ -28,7 +28,6 @@
 #include <linux/delay.h>
 #include <linux/string.h>
 #include <linux/seq_file.h>
-#include <linux/gpio.h>
 #if defined(CONFIG_OMAP4_DSS_HDMI_AUDIO)
 #include <sound/asound.h>
 #include <sound/asoundef.h>
@@ -38,6 +37,9 @@
 #include "dss.h"
 #include "dss_features.h"
 
+#define HDMI_IRQ_LINK_CONNECT		(1 << 25)
+#define HDMI_IRQ_LINK_DISCONNECT	(1 << 26)
+
 static inline void hdmi_write_reg(void __iomem *base_addr,
 				const u16 idx, u32 val)
 {
@@ -233,37 +235,39 @@ void ti_hdmi_4xxx_pll_disable(struct hdmi_ip_data *ip_data)
 	hdmi_set_pll_pwr(ip_data, HDMI_PLLPWRCMD_ALLOFF);
 }
 
-static int hdmi_check_hpd_state(struct hdmi_ip_data *ip_data)
+static irqreturn_t hdmi_irq_handler(int irq, void *data)
 {
-	bool hpd;
-	int r;
-
-	mutex_lock(&ip_data->lock);
-
-	hpd = gpio_get_value(ip_data->hpd_gpio);
+	struct hdmi_ip_data *ip_data = data;
+	void __iomem *wp_base = hdmi_wp_base(ip_data);
+	u32 irqstatus;
+
+	irqstatus = hdmi_read_reg(wp_base, HDMI_WP_IRQSTATUS);
+	hdmi_write_reg(wp_base, HDMI_WP_IRQSTATUS, irqstatus);
+	/* flush posted write */
+	hdmi_read_reg(wp_base, HDMI_WP_IRQSTATUS);
+
+	if ((irqstatus & HDMI_IRQ_LINK_CONNECT) &&
+			irqstatus & HDMI_IRQ_LINK_DISCONNECT) {
+		/*
+		 * If we get both connect and disconnect interrupts at the same
+		 * time, turn off the PHY, clear interrupts, and restart, which
+		 * raises connect interrupt if a cable is connected, or nothing
+		 * if cable is not connected.
+		 */
+		hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_OFF);
 
-	if (hpd)
-		r = hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_TXON);
-	else
-		r = hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_LDOON);
+		hdmi_write_reg(wp_base, HDMI_WP_IRQSTATUS,
+			HDMI_IRQ_LINK_CONNECT | HDMI_IRQ_LINK_DISCONNECT);
+		/* flush posted write */
+		hdmi_read_reg(wp_base, HDMI_WP_IRQSTATUS);
 
-	if (r) {
-		DSSERR("Failed to %s PHY TX power\n",
-				hpd ? "enable" : "disable");
-		goto err;
+		hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_LDOON);
+	} else if (irqstatus & HDMI_IRQ_LINK_CONNECT) {
+		hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_TXON);
+	} else if (irqstatus & HDMI_IRQ_LINK_DISCONNECT) {
+		hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_LDOON);
 	}
 
-err:
-	mutex_unlock(&ip_data->lock);
-	return r;
-}
-
-static irqreturn_t hpd_irq_handler(int irq, void *data)
-{
-	struct hdmi_ip_data *ip_data = data;
-
-	hdmi_check_hpd_state(ip_data);
-
 	return IRQ_HANDLED;
 }
 
@@ -272,6 +276,12 @@ int ti_hdmi_4xxx_phy_enable(struct hdmi_ip_data *ip_data)
 	u16 r = 0;
 	void __iomem *phy_base = hdmi_phy_base(ip_data);
 
+	hdmi_write_reg(hdmi_wp_base(ip_data), HDMI_WP_IRQENABLE_CLR,
+			0xffffffff);
+
+	hdmi_write_reg(hdmi_wp_base(ip_data), HDMI_WP_IRQSTATUS,
+			HDMI_IRQ_LINK_CONNECT | HDMI_IRQ_LINK_DISCONNECT);
+
 	r = hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_LDOON);
 	if (r)
 		return r;
@@ -297,29 +307,23 @@ int ti_hdmi_4xxx_phy_enable(struct hdmi_ip_data *ip_data)
 	/* Write to phy address 3 to change the polarity control */
 	REG_FLD_MOD(phy_base, HDMI_TXPHY_PAD_CFG_CTRL, 0x1, 27, 27);
 
-	r = request_threaded_irq(gpio_to_irq(ip_data->hpd_gpio),
-				 NULL, hpd_irq_handler,
-				 IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING |
-				 IRQF_ONESHOT, "hpd", ip_data);
+	r = request_threaded_irq(ip_data->irq, NULL, hdmi_irq_handler,
+				 IRQF_ONESHOT, "OMAP HDMI", ip_data);
 	if (r) {
-		DSSERR("HPD IRQ request failed\n");
+		DSSERR("HDMI IRQ request failed\n");
 		hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_OFF);
 		return r;
 	}
 
-	r = hdmi_check_hpd_state(ip_data);
-	if (r) {
-		free_irq(gpio_to_irq(ip_data->hpd_gpio), ip_data);
-		hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_OFF);
-		return r;
-	}
+	hdmi_write_reg(hdmi_wp_base(ip_data), HDMI_WP_IRQENABLE_SET,
+			HDMI_IRQ_LINK_CONNECT | HDMI_IRQ_LINK_DISCONNECT);
 
 	return 0;
 }
 
 void ti_hdmi_4xxx_phy_disable(struct hdmi_ip_data *ip_data)
 {
-	free_irq(gpio_to_irq(ip_data->hpd_gpio), ip_data);
+	free_irq(ip_data->irq, ip_data);
 
 	hdmi_set_phy_pwr(ip_data, HDMI_PHYPWRCMD_OFF);
 }
@@ -476,11 +480,6 @@ int ti_hdmi_4xxx_read_edid(struct hdmi_ip_data *ip_data,
 	return l;
 }
 
-bool ti_hdmi_4xxx_detect(struct hdmi_ip_data *ip_data)
-{
-	return gpio_get_value(ip_data->hpd_gpio);
-}
-
 static void hdmi_core_init(struct hdmi_core_video_config *video_cfg,
 			struct hdmi_core_infoframe_avi *avi_cfg,
 			struct hdmi_core_packet_enable_repeat *repeat_cfg)
diff --git a/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.h b/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.h
index 8366ae1..6ef2f92 100644
--- a/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.h
+++ b/drivers/video/omap2/dss/ti_hdmi_4xxx_ip.h
@@ -33,6 +33,7 @@
 #define HDMI_WP_IRQSTATUS			0x28
 #define HDMI_WP_PWR_CTRL			0x40
 #define HDMI_WP_IRQENABLE_SET			0x2C
+#define HDMI_WP_IRQENABLE_CLR			0x30
 #define HDMI_WP_VIDEO_CFG			0x50
 #define HDMI_WP_VIDEO_SIZE			0x60
 #define HDMI_WP_VIDEO_TIMING_H			0x68
diff --git a/drivers/video/omap2/dss/venc.c b/drivers/video/omap2/dss/venc.c
index 74fdb3e..496a106 100644
--- a/drivers/video/omap2/dss/venc.c
+++ b/drivers/video/omap2/dss/venc.c
@@ -304,7 +304,7 @@ static struct {
 	enum omap_dss_venc_type type;
 	bool invert_polarity;
 
-	struct omap_dss_output output;
+	struct omap_dss_device output;
 } venc;
 
 static inline void venc_write_reg(int idx, u32 val)
@@ -429,7 +429,7 @@ static const struct venc_config *venc_timings_to_config(
 
 static int venc_power_on(struct omap_dss_device *dssdev)
 {
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = venc.output.manager;
 	u32 l;
 	int r;
 
@@ -480,7 +480,7 @@ err0:
 
 static void venc_power_off(struct omap_dss_device *dssdev)
 {
-	struct omap_overlay_manager *mgr = dssdev->output->manager;
+	struct omap_overlay_manager *mgr = venc.output.manager;
 
 	venc_write_reg(VENC_OUTPUT_CONTROL, 0);
 	dss_set_dac_pwrdn_bgz(0);
@@ -492,15 +492,9 @@ static void venc_power_off(struct omap_dss_device *dssdev)
 	venc_runtime_put();
 }
 
-unsigned long venc_get_pixel_clock(void)
-{
-	/* VENC Pixel Clock in Mhz */
-	return 13500000;
-}
-
 int omapdss_venc_display_enable(struct omap_dss_device *dssdev)
 {
-	struct omap_dss_output *out = dssdev->output;
+	struct omap_dss_device *out = &venc.output;
 	int r;
 
 	DSSDBG("venc_display_enable\n");
@@ -513,23 +507,15 @@ int omapdss_venc_display_enable(struct omap_dss_device *dssdev)
 		goto err0;
 	}
 
-	r = omap_dss_start_device(dssdev);
-	if (r) {
-		DSSERR("failed to start device\n");
-		goto err0;
-	}
-
 	r = venc_power_on(dssdev);
 	if (r)
-		goto err1;
+		goto err0;
 
 	venc.wss_data = 0;
 
 	mutex_unlock(&venc.venc_lock);
 
 	return 0;
-err1:
-	omap_dss_stop_device(dssdev);
 err0:
 	mutex_unlock(&venc.venc_lock);
 	return r;
@@ -543,8 +529,6 @@ void omapdss_venc_display_disable(struct omap_dss_device *dssdev)
 
 	venc_power_off(dssdev);
 
-	omap_dss_stop_device(dssdev);
-
 	mutex_unlock(&venc.venc_lock);
 }
 
@@ -561,6 +545,8 @@ void omapdss_venc_set_timings(struct omap_dss_device *dssdev,
 
 	venc.timings = *timings;
 
+	dispc_set_tv_pclk(13500000);
+
 	mutex_unlock(&venc.venc_lock);
 }
 
@@ -578,6 +564,16 @@ int omapdss_venc_check_timings(struct omap_dss_device *dssdev,
 	return -EINVAL;
 }
 
+static void venc_get_timings(struct omap_dss_device *dssdev,
+		struct omap_video_timings *timings)
+{
+	mutex_lock(&venc.venc_lock);
+
+	*timings = venc.timings;
+
+	mutex_unlock(&venc.venc_lock);
+}
+
 u32 omapdss_venc_get_wss(struct omap_dss_device *dssdev)
 {
 	/* Invert due to VENC_L21_WC_CTL:INV=1 */
@@ -633,23 +629,22 @@ void omapdss_venc_invert_vid_out_polarity(struct omap_dss_device *dssdev,
 	mutex_unlock(&venc.venc_lock);
 }
 
-static int venc_init_display(struct omap_dss_device *dssdev)
+static int venc_init_regulator(void)
 {
-	DSSDBG("init_display\n");
-
-	if (venc.vdda_dac_reg == NULL) {
-		struct regulator *vdda_dac;
+	struct regulator *vdda_dac;
 
-		vdda_dac = regulator_get(&venc.pdev->dev, "vdda_dac");
+	if (venc.vdda_dac_reg != NULL)
+		return 0;
 
-		if (IS_ERR(vdda_dac)) {
-			DSSERR("can't get VDDA_DAC regulator\n");
-			return PTR_ERR(vdda_dac);
-		}
+	vdda_dac = devm_regulator_get(&venc.pdev->dev, "vdda_dac");
 
-		venc.vdda_dac_reg = vdda_dac;
+	if (IS_ERR(vdda_dac)) {
+		DSSERR("can't get VDDA_DAC regulator\n");
+		return PTR_ERR(vdda_dac);
 	}
 
+	venc.vdda_dac_reg = vdda_dac;
+
 	return 0;
 }
 
@@ -765,19 +760,16 @@ static int venc_probe_pdata(struct platform_device *vencdev)
 	if (!plat_dssdev)
 		return 0;
 
+	r = venc_init_regulator();
+	if (r)
+		return r;
+
 	dssdev = dss_alloc_and_init_device(&vencdev->dev);
 	if (!dssdev)
 		return -ENOMEM;
 
 	dss_copy_device_pdata(dssdev, plat_dssdev);
 
-	r = venc_init_display(dssdev);
-	if (r) {
-		DSSERR("device %s init failed: %d\n", dssdev->name, r);
-		dss_put_device(dssdev);
-		return r;
-	}
-
 	r = omapdss_output_set_device(&venc.output, dssdev);
 	if (r) {
 		DSSERR("failed to connect output to new device: %s\n",
@@ -797,24 +789,87 @@ static int venc_probe_pdata(struct platform_device *vencdev)
 	return 0;
 }
 
+static int venc_connect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	struct omap_overlay_manager *mgr;
+	int r;
+
+	r = venc_init_regulator();
+	if (r)
+		return r;
+
+	mgr = omap_dss_get_overlay_manager(dssdev->dispc_channel);
+	if (!mgr)
+		return -ENODEV;
+
+	r = dss_mgr_connect(mgr, dssdev);
+	if (r)
+		return r;
+
+	r = omapdss_output_set_device(dssdev, dst);
+	if (r) {
+		DSSERR("failed to connect output to new device: %s\n",
+				dst->name);
+		dss_mgr_disconnect(mgr, dssdev);
+		return r;
+	}
+
+	return 0;
+}
+
+static void venc_disconnect(struct omap_dss_device *dssdev,
+		struct omap_dss_device *dst)
+{
+	WARN_ON(dst != dssdev->device);
+
+	if (dst != dssdev->device)
+		return;
+
+	omapdss_output_unset_device(dssdev);
+
+	if (dssdev->manager)
+		dss_mgr_disconnect(dssdev->manager, dssdev);
+}
+
+static const struct omapdss_atv_ops venc_ops = {
+	.connect = venc_connect,
+	.disconnect = venc_disconnect,
+
+	.enable = omapdss_venc_display_enable,
+	.disable = omapdss_venc_display_disable,
+
+	.check_timings = omapdss_venc_check_timings,
+	.set_timings = omapdss_venc_set_timings,
+	.get_timings = venc_get_timings,
+
+	.set_type = omapdss_venc_set_type,
+	.invert_vid_out_polarity = omapdss_venc_invert_vid_out_polarity,
+
+	.set_wss = omapdss_venc_set_wss,
+	.get_wss = omapdss_venc_get_wss,
+};
+
 static void venc_init_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &venc.output;
+	struct omap_dss_device *out = &venc.output;
 
-	out->pdev = pdev;
+	out->dev = &pdev->dev;
 	out->id = OMAP_DSS_OUTPUT_VENC;
-	out->type = OMAP_DISPLAY_TYPE_VENC;
+	out->output_type = OMAP_DISPLAY_TYPE_VENC;
 	out->name = "venc.0";
 	out->dispc_channel = OMAP_DSS_CHANNEL_DIGIT;
+	out->ops.atv = &venc_ops;
+	out->owner = THIS_MODULE;
 
-	dss_register_output(out);
+	omapdss_register_output(out);
 }
 
 static void __exit venc_uninit_output(struct platform_device *pdev)
 {
-	struct omap_dss_output *out = &venc.output;
+	struct omap_dss_device *out = &venc.output;
 
-	dss_unregister_output(out);
+	omapdss_unregister_output(out);
 }
 
 /* VENC HW IP initialisation */
@@ -866,16 +921,17 @@ static int omap_venchw_probe(struct platform_device *pdev)
 
 	venc_init_output(pdev);
 
-	r = venc_probe_pdata(pdev);
-	if (r) {
-		venc_panel_exit();
-		venc_uninit_output(pdev);
-		pm_runtime_disable(&pdev->dev);
-		return r;
+	if (pdev->dev.platform_data) {
+		r = venc_probe_pdata(pdev);
+		if (r)
+			goto err_probe;
 	}
 
 	return 0;
 
+err_probe:
+	venc_panel_exit();
+	venc_uninit_output(pdev);
 err_panel_init:
 err_runtime_get:
 	pm_runtime_disable(&pdev->dev);
@@ -886,11 +942,6 @@ static int __exit omap_venchw_remove(struct platform_device *pdev)
 {
 	dss_unregister_child_devices(&pdev->dev);
 
-	if (venc.vdda_dac_reg != NULL) {
-		regulator_put(venc.vdda_dac_reg);
-		venc.vdda_dac_reg = NULL;
-	}
-
 	venc_panel_exit();
 
 	venc_uninit_output(pdev);
diff --git a/drivers/video/omap2/dss/venc_panel.c b/drivers/video/omap2/dss/venc_panel.c
index 0d2b1a0..f7d92c5 100644
--- a/drivers/video/omap2/dss/venc_panel.c
+++ b/drivers/video/omap2/dss/venc_panel.c
@@ -107,19 +107,19 @@ static int venc_panel_probe(struct omap_dss_device *dssdev)
 
 	dssdev->panel.timings = default_timings;
 
-	return device_create_file(&dssdev->dev, &dev_attr_output_type);
+	return device_create_file(dssdev->dev, &dev_attr_output_type);
 }
 
 static void venc_panel_remove(struct omap_dss_device *dssdev)
 {
-	device_remove_file(&dssdev->dev, &dev_attr_output_type);
+	device_remove_file(dssdev->dev, &dev_attr_output_type);
 }
 
 static int venc_panel_enable(struct omap_dss_device *dssdev)
 {
 	int r;
 
-	dev_dbg(&dssdev->dev, "venc_panel_enable\n");
+	dev_dbg(dssdev->dev, "venc_panel_enable\n");
 
 	mutex_lock(&venc_panel.lock);
 
@@ -150,7 +150,7 @@ err:
 
 static void venc_panel_disable(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "venc_panel_disable\n");
+	dev_dbg(dssdev->dev, "venc_panel_disable\n");
 
 	mutex_lock(&venc_panel.lock);
 
@@ -167,7 +167,7 @@ end:
 static void venc_panel_set_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	dev_dbg(&dssdev->dev, "venc_panel_set_timings\n");
+	dev_dbg(dssdev->dev, "venc_panel_set_timings\n");
 
 	mutex_lock(&venc_panel.lock);
 
@@ -180,21 +180,21 @@ static void venc_panel_set_timings(struct omap_dss_device *dssdev,
 static int venc_panel_check_timings(struct omap_dss_device *dssdev,
 		struct omap_video_timings *timings)
 {
-	dev_dbg(&dssdev->dev, "venc_panel_check_timings\n");
+	dev_dbg(dssdev->dev, "venc_panel_check_timings\n");
 
 	return omapdss_venc_check_timings(dssdev, timings);
 }
 
 static u32 venc_panel_get_wss(struct omap_dss_device *dssdev)
 {
-	dev_dbg(&dssdev->dev, "venc_panel_get_wss\n");
+	dev_dbg(dssdev->dev, "venc_panel_get_wss\n");
 
 	return omapdss_venc_get_wss(dssdev);
 }
 
 static int venc_panel_set_wss(struct omap_dss_device *dssdev, u32 wss)
 {
-	dev_dbg(&dssdev->dev, "venc_panel_set_wss\n");
+	dev_dbg(dssdev->dev, "venc_panel_set_wss\n");
 
 	return omapdss_venc_set_wss(dssdev, wss);
 }
diff --git a/drivers/video/omap2/omapfb/omapfb-ioctl.c b/drivers/video/omap2/omapfb/omapfb-ioctl.c
index d30b45d..146b6f5 100644
--- a/drivers/video/omap2/omapfb/omapfb-ioctl.c
+++ b/drivers/video/omap2/omapfb/omapfb-ioctl.c
@@ -770,12 +770,17 @@ int omapfb_ioctl(struct fb_info *fbi, unsigned int cmd, unsigned long arg)
 
 	case OMAPFB_WAITFORVSYNC:
 		DBG("ioctl WAITFORVSYNC\n");
-		if (!display || !display->output || !display->output->manager) {
+
+		if (!display) {
 			r = -EINVAL;
 			break;
 		}
 
-		mgr = display->output->manager;
+		mgr = omapdss_find_mgr_from_display(display);
+		if (!mgr) {
+			r = -EINVAL;
+			break;
+		}
 
 		r = mgr->wait_for_vsync(mgr);
 		break;
diff --git a/drivers/video/omap2/omapfb/omapfb-main.c b/drivers/video/omap2/omapfb/omapfb-main.c
index 856917b..27d6905 100644
--- a/drivers/video/omap2/omapfb/omapfb-main.c
+++ b/drivers/video/omap2/omapfb/omapfb-main.c
@@ -1853,6 +1853,8 @@ static void omapfb_free_resources(struct omapfb2_device *fbdev)
 		if (dssdev->state != OMAP_DSS_DISPLAY_DISABLED)
 			dssdev->driver->disable(dssdev);
 
+		dssdev->driver->disconnect(dssdev);
+
 		omap_dss_put_device(dssdev);
 	}
 
@@ -2363,27 +2365,26 @@ static int omapfb_init_connections(struct omapfb2_device *fbdev,
 	int i, r;
 	struct omap_overlay_manager *mgr;
 
-	if (!def_dssdev->output) {
-		dev_err(fbdev->dev, "no output for the default display\n");
-		return -EINVAL;
+	r = def_dssdev->driver->connect(def_dssdev);
+	if (r) {
+		dev_err(fbdev->dev, "failed to connect default display\n");
+		return r;
 	}
 
 	for (i = 0; i < fbdev->num_displays; ++i) {
 		struct omap_dss_device *dssdev = fbdev->displays[i].dssdev;
-		struct omap_dss_output *out = dssdev->output;
-
-		mgr = omap_dss_get_overlay_manager(out->dispc_channel);
 
-		if (!mgr || !out)
+		if (dssdev == def_dssdev)
 			continue;
 
-		if (mgr->output)
-			mgr->unset_output(mgr);
-
-		mgr->set_output(mgr, out);
+		/*
+		 * We don't care if the connect succeeds or not. We just want to
+		 * connect as many displays as possible.
+		 */
+		dssdev->driver->connect(dssdev);
 	}
 
-	mgr = def_dssdev->output->manager;
+	mgr = omapdss_find_mgr_from_display(def_dssdev);
 
 	if (!mgr) {
 		dev_err(fbdev->dev, "no ovl manager for the default display\n");
@@ -2502,7 +2503,7 @@ static int omapfb_probe(struct platform_device *pdev)
 
 	if (def_display == NULL) {
 		dev_err(fbdev->dev, "failed to find default display\n");
-		r = -EINVAL;
+		r = -EPROBE_DEFER;
 		goto cleanup;
 	}
 
diff --git a/drivers/video/pxa3xx-gcu.c b/drivers/video/pxa3xx-gcu.c
index 7cf0b13..ad382b3 100644
--- a/drivers/video/pxa3xx-gcu.c
+++ b/drivers/video/pxa3xx-gcu.c
@@ -710,7 +710,6 @@ err_misc_deregister:
 	misc_deregister(&priv->misc_dev);
 
 err_free_priv:
-	platform_set_drvdata(dev, NULL);
 	free_buffers(dev, priv);
 	kfree(priv);
 	return ret;
@@ -728,7 +727,6 @@ static int pxa3xx_gcu_remove(struct platform_device *dev)
 			priv->shared, priv->shared_phys);
 	iounmap(priv->mmio_base);
 	release_mem_region(r->start, resource_size(r));
-	platform_set_drvdata(dev, NULL);
 	clk_disable(priv->clk);
 	free_buffers(dev, priv);
 	kfree(priv);
diff --git a/drivers/video/pxafb.c b/drivers/video/pxafb.c
index 580f80c..eca2de4 100644
--- a/drivers/video/pxafb.c
+++ b/drivers/video/pxafb.c
@@ -2256,7 +2256,6 @@ failed_free_res:
 	release_mem_region(r->start, resource_size(r));
 failed_fbi:
 	clk_put(fbi->clk);
-	platform_set_drvdata(dev, NULL);
 	kfree(fbi);
 failed:
 	return ret;
diff --git a/drivers/video/s3c2410fb.c b/drivers/video/s3c2410fb.c
index 76a0e7f..21a32ad 100644
--- a/drivers/video/s3c2410fb.c
+++ b/drivers/video/s3c2410fb.c
@@ -1005,7 +1005,6 @@ release_regs:
 release_mem:
 	release_mem_region(res->start, size);
 dealloc_fb:
-	platform_set_drvdata(pdev, NULL);
 	framebuffer_release(fbinfo);
 	return ret;
 }
@@ -1051,7 +1050,6 @@ static int s3c2410fb_remove(struct platform_device *pdev)
 
 	release_mem_region(info->mem->start, resource_size(info->mem));
 
-	platform_set_drvdata(pdev, NULL);
 	framebuffer_release(fbinfo);
 
 	return 0;
diff --git a/drivers/video/sa1100fb.c b/drivers/video/sa1100fb.c
index f34c858..de76da0 100644
--- a/drivers/video/sa1100fb.c
+++ b/drivers/video/sa1100fb.c
@@ -1271,7 +1271,6 @@ static int sa1100fb_probe(struct platform_device *pdev)
  failed:
 	if (fbi)
 		iounmap(fbi->base);
-	platform_set_drvdata(pdev, NULL);
 	kfree(fbi);
 	release_mem_region(res->start, resource_size(res));
 	return ret;
diff --git a/drivers/video/sh7760fb.c b/drivers/video/sh7760fb.c
index 5fbb0c7..a8c6c43 100644
--- a/drivers/video/sh7760fb.c
+++ b/drivers/video/sh7760fb.c
@@ -571,7 +571,6 @@ static int sh7760fb_remove(struct platform_device *dev)
 	iounmap(par->base);
 	release_mem_region(par->ioarea->start, resource_size(par->ioarea));
 	framebuffer_release(info);
-	platform_set_drvdata(dev, NULL);
 
 	return 0;
 }
diff --git a/drivers/video/sh_mipi_dsi.c b/drivers/video/sh_mipi_dsi.c
index 6cad530..8f6e8ff 100644
--- a/drivers/video/sh_mipi_dsi.c
+++ b/drivers/video/sh_mipi_dsi.c
@@ -567,7 +567,6 @@ static int sh_mipi_remove(struct platform_device *pdev)
 	iounmap(mipi->base);
 	if (res)
 		release_mem_region(res->start, resource_size(res));
-	platform_set_drvdata(pdev, NULL);
 	kfree(mipi);
 
 	return 0;
diff --git a/drivers/video/smscufx.c b/drivers/video/smscufx.c
index b2b33fc..e188ada 100644
--- a/drivers/video/smscufx.c
+++ b/drivers/video/smscufx.c
@@ -1622,7 +1622,7 @@ static int ufx_usb_probe(struct usb_interface *interface,
 {
 	struct usb_device *usbdev;
 	struct ufx_data *dev;
-	struct fb_info *info = 0;
+	struct fb_info *info = NULL;
 	int retval = -ENOMEM;
 	u32 id_rev, fpga_rev;
 
diff --git a/drivers/video/ssd1307fb.c b/drivers/video/ssd1307fb.c
index 9ef05d3..44967c8 100644
--- a/drivers/video/ssd1307fb.c
+++ b/drivers/video/ssd1307fb.c
@@ -16,24 +16,50 @@
 #include <linux/pwm.h>
 #include <linux/delay.h>
 
-#define SSD1307FB_WIDTH			96
-#define SSD1307FB_HEIGHT		16
-
 #define SSD1307FB_DATA			0x40
 #define SSD1307FB_COMMAND		0x80
 
+#define SSD1307FB_SET_ADDRESS_MODE	0x20
+#define SSD1307FB_SET_ADDRESS_MODE_HORIZONTAL	(0x00)
+#define SSD1307FB_SET_ADDRESS_MODE_VERTICAL	(0x01)
+#define SSD1307FB_SET_ADDRESS_MODE_PAGE		(0x02)
+#define SSD1307FB_SET_COL_RANGE		0x21
+#define SSD1307FB_SET_PAGE_RANGE	0x22
 #define SSD1307FB_CONTRAST		0x81
+#define	SSD1307FB_CHARGE_PUMP		0x8d
 #define SSD1307FB_SEG_REMAP_ON		0xa1
 #define SSD1307FB_DISPLAY_OFF		0xae
+#define SSD1307FB_SET_MULTIPLEX_RATIO	0xa8
 #define SSD1307FB_DISPLAY_ON		0xaf
 #define SSD1307FB_START_PAGE_ADDRESS	0xb0
+#define SSD1307FB_SET_DISPLAY_OFFSET	0xd3
+#define	SSD1307FB_SET_CLOCK_FREQ	0xd5
+#define	SSD1307FB_SET_PRECHARGE_PERIOD	0xd9
+#define	SSD1307FB_SET_COM_PINS_CONFIG	0xda
+#define	SSD1307FB_SET_VCOMH		0xdb
+
+struct ssd1307fb_par;
+
+struct ssd1307fb_ops {
+	int (*init)(struct ssd1307fb_par *);
+	int (*remove)(struct ssd1307fb_par *);
+};
 
 struct ssd1307fb_par {
 	struct i2c_client *client;
+	u32 height;
 	struct fb_info *info;
+	struct ssd1307fb_ops *ops;
+	u32 page_offset;
 	struct pwm_device *pwm;
 	u32 pwm_period;
 	int reset;
+	u32 width;
+};
+
+struct ssd1307fb_array {
+	u8	type;
+	u8	data[0];
 };
 
 static struct fb_fix_screeninfo ssd1307fb_fix = {
@@ -43,68 +69,87 @@ static struct fb_fix_screeninfo ssd1307fb_fix = {
 	.xpanstep	= 0,
 	.ypanstep	= 0,
 	.ywrapstep	= 0,
-	.line_length	= SSD1307FB_WIDTH / 8,
 	.accel		= FB_ACCEL_NONE,
 };
 
 static struct fb_var_screeninfo ssd1307fb_var = {
-	.xres		= SSD1307FB_WIDTH,
-	.yres		= SSD1307FB_HEIGHT,
-	.xres_virtual	= SSD1307FB_WIDTH,
-	.yres_virtual	= SSD1307FB_HEIGHT,
 	.bits_per_pixel	= 1,
 };
 
-static int ssd1307fb_write_array(struct i2c_client *client, u8 type, u8 *cmd, u32 len)
+static struct ssd1307fb_array *ssd1307fb_alloc_array(u32 len, u8 type)
 {
-	u8 *buf;
-	int ret = 0;
-
-	buf = kzalloc(len + 1, GFP_KERNEL);
-	if (!buf) {
-		dev_err(&client->dev, "Couldn't allocate sending buffer.\n");
-		return -ENOMEM;
-	}
+	struct ssd1307fb_array *array;
 
-	buf[0] = type;
-	memcpy(buf + 1, cmd, len);
+	array = kzalloc(sizeof(struct ssd1307fb_array) + len, GFP_KERNEL);
+	if (!array)
+		return NULL;
 
-	ret = i2c_master_send(client, buf, len + 1);
-	if (ret != len + 1) {
-		dev_err(&client->dev, "Couldn't send I2C command.\n");
-		goto error;
-	}
+	array->type = type;
 
-error:
-	kfree(buf);
-	return ret;
+	return array;
 }
 
-static inline int ssd1307fb_write_cmd_array(struct i2c_client *client, u8 *cmd, u32 len)
+static int ssd1307fb_write_array(struct i2c_client *client,
+				 struct ssd1307fb_array *array, u32 len)
 {
-	return ssd1307fb_write_array(client, SSD1307FB_COMMAND, cmd, len);
+	int ret;
+
+	len += sizeof(struct ssd1307fb_array);
+
+	ret = i2c_master_send(client, (u8 *)array, len);
+	if (ret != len) {
+		dev_err(&client->dev, "Couldn't send I2C command.\n");
+		return ret;
+	}
+
+	return 0;
 }
 
 static inline int ssd1307fb_write_cmd(struct i2c_client *client, u8 cmd)
 {
-	return ssd1307fb_write_cmd_array(client, &cmd, 1);
-}
+	struct ssd1307fb_array *array;
+	int ret;
 
-static inline int ssd1307fb_write_data_array(struct i2c_client *client, u8 *cmd, u32 len)
-{
-	return ssd1307fb_write_array(client, SSD1307FB_DATA, cmd, len);
+	array = ssd1307fb_alloc_array(1, SSD1307FB_COMMAND);
+	if (!array)
+		return -ENOMEM;
+
+	array->data[0] = cmd;
+
+	ret = ssd1307fb_write_array(client, array, 1);
+	kfree(array);
+
+	return ret;
 }
 
 static inline int ssd1307fb_write_data(struct i2c_client *client, u8 data)
 {
-	return ssd1307fb_write_data_array(client, &data, 1);
+	struct ssd1307fb_array *array;
+	int ret;
+
+	array = ssd1307fb_alloc_array(1, SSD1307FB_DATA);
+	if (!array)
+		return -ENOMEM;
+
+	array->data[0] = data;
+
+	ret = ssd1307fb_write_array(client, array, 1);
+	kfree(array);
+
+	return ret;
 }
 
 static void ssd1307fb_update_display(struct ssd1307fb_par *par)
 {
+	struct ssd1307fb_array *array;
 	u8 *vmem = par->info->screen_base;
 	int i, j, k;
 
+	array = ssd1307fb_alloc_array(par->width * par->height / 8,
+				      SSD1307FB_DATA);
+	if (!array)
+		return;
+
 	/*
 	 * The screen is divided in pages, each having a height of 8
 	 * pixels, and the width of the screen. When sending a byte of
@@ -134,24 +179,23 @@ static void ssd1307fb_update_display(struct ssd1307fb_par *par)
 	 *  (5) A4 B4 C4 D4 E4 F4 G4 H4
 	 */
 
-	for (i = 0; i < (SSD1307FB_HEIGHT / 8); i++) {
-		ssd1307fb_write_cmd(par->client, SSD1307FB_START_PAGE_ADDRESS + (i + 1));
-		ssd1307fb_write_cmd(par->client, 0x00);
-		ssd1307fb_write_cmd(par->client, 0x10);
-
-		for (j = 0; j < SSD1307FB_WIDTH; j++) {
-			u8 buf = 0;
+	for (i = 0; i < (par->height / 8); i++) {
+		for (j = 0; j < par->width; j++) {
+			u32 array_idx = i * par->width + j;
+			array->data[array_idx] = 0;
 			for (k = 0; k < 8; k++) {
-				u32 page_length = SSD1307FB_WIDTH * i;
-				u32 index = page_length + (SSD1307FB_WIDTH * k + j) / 8;
+				u32 page_length = par->width * i;
+				u32 index = page_length + (par->width * k + j) / 8;
 				u8 byte = *(vmem + index);
 				u8 bit = byte & (1 << (j % 8));
 				bit = bit >> (j % 8);
-				buf |= bit << k;
+				array->data[array_idx] |= bit << k;
 			}
-			ssd1307fb_write_data(par->client, buf);
 		}
 	}
+
+	ssd1307fb_write_array(par->client, array, par->width * par->height / 8);
+	kfree(array);
 }
 
 
@@ -227,16 +271,167 @@ static struct fb_deferred_io ssd1307fb_defio = {
 	.deferred_io	= ssd1307fb_deferred_io,
 };
 
+static int ssd1307fb_ssd1307_init(struct ssd1307fb_par *par)
+{
+	int ret;
+
+	par->pwm = pwm_get(&par->client->dev, NULL);
+	if (IS_ERR(par->pwm)) {
+		dev_err(&par->client->dev, "Could not get PWM from device tree!\n");
+		return PTR_ERR(par->pwm);
+	}
+
+	par->pwm_period = pwm_get_period(par->pwm);
+	/* Enable the PWM */
+	pwm_config(par->pwm, par->pwm_period / 2, par->pwm_period);
+	pwm_enable(par->pwm);
+
+	dev_dbg(&par->client->dev, "Using PWM%d with a %dns period.\n",
+		par->pwm->pwm, par->pwm_period);
+
+	/* Map column 127 of the OLED to segment 0 */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SEG_REMAP_ON);
+	if (ret < 0)
+		return ret;
+
+	/* Turn on the display */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_DISPLAY_ON);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static int ssd1307fb_ssd1307_remove(struct ssd1307fb_par *par)
+{
+	pwm_disable(par->pwm);
+	pwm_put(par->pwm);
+	return 0;
+}
+
+static struct ssd1307fb_ops ssd1307fb_ssd1307_ops = {
+	.init	= ssd1307fb_ssd1307_init,
+	.remove	= ssd1307fb_ssd1307_remove,
+};
+
+static int ssd1307fb_ssd1306_init(struct ssd1307fb_par *par)
+{
+	int ret;
+
+	/* Set initial contrast */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_CONTRAST);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x7f);
+	if (ret < 0)
+		return ret;
+
+	/* Set COM direction */
+	ret = ssd1307fb_write_cmd(par->client, 0xc8);
+	if (ret < 0)
+		return ret;
+
+	/* Set segment re-map */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SEG_REMAP_ON);
+	if (ret < 0)
+		return ret;
+
+	/* Set multiplex ratio value */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_MULTIPLEX_RATIO);
+	ret = ret & ssd1307fb_write_cmd(par->client, par->height - 1);
+	if (ret < 0)
+		return ret;
+
+	/* set display offset value */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_DISPLAY_OFFSET);
+	ret = ssd1307fb_write_cmd(par->client, 0x20);
+	if (ret < 0)
+		return ret;
+
+	/* Set clock frequency */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_CLOCK_FREQ);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0xf0);
+	if (ret < 0)
+		return ret;
+
+	/* Set precharge period in number of ticks from the internal clock */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_PRECHARGE_PERIOD);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x22);
+	if (ret < 0)
+		return ret;
+
+	/* Set COM pins configuration */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_COM_PINS_CONFIG);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x22);
+	if (ret < 0)
+		return ret;
+
+	/* Set VCOMH */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_VCOMH);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x49);
+	if (ret < 0)
+		return ret;
+
+	/* Turn on the DC-DC Charge Pump */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_CHARGE_PUMP);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x14);
+	if (ret < 0)
+		return ret;
+
+	/* Switch to horizontal addressing mode */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_ADDRESS_MODE);
+	ret = ret & ssd1307fb_write_cmd(par->client,
+					SSD1307FB_SET_ADDRESS_MODE_HORIZONTAL);
+	if (ret < 0)
+		return ret;
+
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_COL_RANGE);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x0);
+	ret = ret & ssd1307fb_write_cmd(par->client, par->width - 1);
+	if (ret < 0)
+		return ret;
+
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_SET_PAGE_RANGE);
+	ret = ret & ssd1307fb_write_cmd(par->client, 0x0);
+	ret = ret & ssd1307fb_write_cmd(par->client,
+					par->page_offset + (par->height / 8) - 1);
+	if (ret < 0)
+		return ret;
+
+	/* Turn on the display */
+	ret = ssd1307fb_write_cmd(par->client, SSD1307FB_DISPLAY_ON);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+static struct ssd1307fb_ops ssd1307fb_ssd1306_ops = {
+	.init	= ssd1307fb_ssd1306_init,
+};
+
+static const struct of_device_id ssd1307fb_of_match[] = {
+	{
+		.compatible = "solomon,ssd1306fb-i2c",
+		.data = (void *)&ssd1307fb_ssd1306_ops,
+	},
+	{
+		.compatible = "solomon,ssd1307fb-i2c",
+		.data = (void *)&ssd1307fb_ssd1307_ops,
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, ssd1307fb_of_match);
+
 static int ssd1307fb_probe(struct i2c_client *client,
 			   const struct i2c_device_id *id)
 {
 	struct fb_info *info;
-	u32 vmem_size = SSD1307FB_WIDTH * SSD1307FB_HEIGHT / 8;
+	struct device_node *node = client->dev.of_node;
+	u32 vmem_size;
 	struct ssd1307fb_par *par;
 	u8 *vmem;
 	int ret;
 
-	if (!client->dev.of_node) {
+	if (!node) {
 		dev_err(&client->dev, "No device tree data found!\n");
 		return -EINVAL;
 	}
@@ -247,6 +442,31 @@ static int ssd1307fb_probe(struct i2c_client *client,
 		return -ENOMEM;
 	}
 
+	par = info->par;
+	par->info = info;
+	par->client = client;
+
+	par->ops = (struct ssd1307fb_ops *)of_match_device(ssd1307fb_of_match,
+							   &client->dev)->data;
+
+	par->reset = of_get_named_gpio(client->dev.of_node,
+					 "reset-gpios", 0);
+	if (!gpio_is_valid(par->reset)) {
+		ret = -EINVAL;
+		goto fb_alloc_error;
+	}
+
+	if (of_property_read_u32(node, "solomon,width", &par->width))
+		par->width = 96;
+
+	if (of_property_read_u32(node, "solomon,height", &par->height))
+		par->width = 16;
+
+	if (of_property_read_u32(node, "solomon,page-offset", &par->page_offset))
+		par->page_offset = 1;
+
+	vmem_size = par->width * par->height / 8;
+
 	vmem = devm_kzalloc(&client->dev, vmem_size, GFP_KERNEL);
 	if (!vmem) {
 		dev_err(&client->dev, "Couldn't allocate graphical memory.\n");
@@ -256,9 +476,15 @@ static int ssd1307fb_probe(struct i2c_client *client,
 
 	info->fbops = &ssd1307fb_ops;
 	info->fix = ssd1307fb_fix;
+	info->fix.line_length = par->width / 8;
 	info->fbdefio = &ssd1307fb_defio;
 
 	info->var = ssd1307fb_var;
+	info->var.xres = par->width;
+	info->var.xres_virtual = par->width;
+	info->var.yres = par->height;
+	info->var.yres_virtual = par->height;
+
 	info->var.red.length = 1;
 	info->var.red.offset = 0;
 	info->var.green.length = 1;
@@ -272,17 +498,6 @@ static int ssd1307fb_probe(struct i2c_client *client,
 
 	fb_deferred_io_init(info);
 
-	par = info->par;
-	par->info = info;
-	par->client = client;
-
-	par->reset = of_get_named_gpio(client->dev.of_node,
-					 "reset-gpios", 0);
-	if (!gpio_is_valid(par->reset)) {
-		ret = -EINVAL;
-		goto reset_oled_error;
-	}
-
 	ret = devm_gpio_request_one(&client->dev, par->reset,
 				    GPIOF_OUT_INIT_HIGH,
 				    "oled-reset");
@@ -293,23 +508,6 @@ static int ssd1307fb_probe(struct i2c_client *client,
 		goto reset_oled_error;
 	}
 
-	par->pwm = pwm_get(&client->dev, NULL);
-	if (IS_ERR(par->pwm)) {
-		dev_err(&client->dev, "Could not get PWM from device tree!\n");
-		ret = PTR_ERR(par->pwm);
-		goto pwm_error;
-	}
-
-	par->pwm_period = pwm_get_period(par->pwm);
-
-	dev_dbg(&client->dev, "Using PWM%d with a %dns period.\n", par->pwm->pwm, par->pwm_period);
-
-	ret = register_framebuffer(info);
-	if (ret) {
-		dev_err(&client->dev, "Couldn't register the framebuffer\n");
-		goto fbreg_error;
-	}
-
 	i2c_set_clientdata(client, info);
 
 	/* Reset the screen */
@@ -318,34 +516,25 @@ static int ssd1307fb_probe(struct i2c_client *client,
 	gpio_set_value(par->reset, 1);
 	udelay(4);
 
-	/* Enable the PWM */
-	pwm_config(par->pwm, par->pwm_period / 2, par->pwm_period);
-	pwm_enable(par->pwm);
-
-	/* Map column 127 of the OLED to segment 0 */
-	ret = ssd1307fb_write_cmd(client, SSD1307FB_SEG_REMAP_ON);
-	if (ret < 0) {
-		dev_err(&client->dev, "Couldn't remap the screen.\n");
-		goto remap_error;
+	if (par->ops->init) {
+		ret = par->ops->init(par);
+		if (ret)
+			goto reset_oled_error;
 	}
 
-	/* Turn on the display */
-	ret = ssd1307fb_write_cmd(client, SSD1307FB_DISPLAY_ON);
-	if (ret < 0) {
-		dev_err(&client->dev, "Couldn't turn the display on.\n");
-		goto remap_error;
+	ret = register_framebuffer(info);
+	if (ret) {
+		dev_err(&client->dev, "Couldn't register the framebuffer\n");
+		goto panel_init_error;
 	}
 
 	dev_info(&client->dev, "fb%d: %s framebuffer device registered, using %d bytes of video memory\n", info->node, info->fix.id, vmem_size);
 
 	return 0;
 
-remap_error:
-	unregister_framebuffer(info);
-	pwm_disable(par->pwm);
-fbreg_error:
-	pwm_put(par->pwm);
-pwm_error:
+panel_init_error:
+	if (par->ops->remove)
+		par->ops->remove(par);
 reset_oled_error:
 	fb_deferred_io_cleanup(info);
 fb_alloc_error:
@@ -359,8 +548,8 @@ static int ssd1307fb_remove(struct i2c_client *client)
 	struct ssd1307fb_par *par = info->par;
 
 	unregister_framebuffer(info);
-	pwm_disable(par->pwm);
-	pwm_put(par->pwm);
+	if (par->ops->remove)
+		par->ops->remove(par);
 	fb_deferred_io_cleanup(info);
 	framebuffer_release(info);
 
@@ -368,17 +557,12 @@ static int ssd1307fb_remove(struct i2c_client *client)
 }
 
 static const struct i2c_device_id ssd1307fb_i2c_id[] = {
+	{ "ssd1306fb", 0 },
 	{ "ssd1307fb", 0 },
 	{ }
 };
 MODULE_DEVICE_TABLE(i2c, ssd1307fb_i2c_id);
 
-static const struct of_device_id ssd1307fb_of_match[] = {
-	{ .compatible = "solomon,ssd1307fb-i2c" },
-	{},
-};
-MODULE_DEVICE_TABLE(of, ssd1307fb_of_match);
-
 static struct i2c_driver ssd1307fb_driver = {
 	.probe = ssd1307fb_probe,
 	.remove = ssd1307fb_remove,
diff --git a/drivers/video/tmiofb.c b/drivers/video/tmiofb.c
index dc4fb86..deb8733 100644
--- a/drivers/video/tmiofb.c
+++ b/drivers/video/tmiofb.c
@@ -794,7 +794,6 @@ err_hw_init:
 		cell->disable(dev);
 err_enable:
 err_find_mode:
-	platform_set_drvdata(dev, NULL);
 	free_irq(irq, info);
 err_request_irq:
 	iounmap(info->screen_base);
@@ -823,8 +822,6 @@ static int tmiofb_remove(struct platform_device *dev)
 		if (cell->disable)
 			cell->disable(dev);
 
-		platform_set_drvdata(dev, NULL);
-
 		free_irq(irq, info);
 
 		iounmap(info->screen_base);
diff --git a/drivers/video/udlfb.c b/drivers/video/udlfb.c
index ec03e72..d2e5bc3 100644
--- a/drivers/video/udlfb.c
+++ b/drivers/video/udlfb.c
@@ -434,10 +434,10 @@ static void dlfb_compress_hline(
 
 	while ((pixel_end > pixel) &&
 	       (cmd_buffer_end - MIN_RLX_CMD_BYTES > cmd)) {
-		uint8_t *raw_pixels_count_byte = 0;
-		uint8_t *cmd_pixels_count_byte = 0;
-		const uint16_t *raw_pixel_start = 0;
-		const uint16_t *cmd_pixel_start, *cmd_pixel_end = 0;
+		uint8_t *raw_pixels_count_byte = NULL;
+		uint8_t *cmd_pixels_count_byte = NULL;
+		const uint16_t *raw_pixel_start = NULL;
+		const uint16_t *cmd_pixel_start, *cmd_pixel_end = NULL;
 
 		prefetchw((void *) cmd); /* pull in one cache line at least */
 
@@ -573,7 +573,7 @@ static int dlfb_render_hline(struct dlfb_data *dev, struct urb **urb_ptr,
 	return 0;
 }
 
-int dlfb_handle_damage(struct dlfb_data *dev, int x, int y,
+static int dlfb_handle_damage(struct dlfb_data *dev, int x, int y,
 	       int width, int height, char *data)
 {
 	int i, ret;
@@ -1588,7 +1588,7 @@ static int dlfb_usb_probe(struct usb_interface *interface,
 			const struct usb_device_id *id)
 {
 	struct usb_device *usbdev;
-	struct dlfb_data *dev = 0;
+	struct dlfb_data *dev = NULL;
 	int retval = -ENOMEM;
 
 	/* usb initialization */
diff --git a/drivers/video/uvesafb.c b/drivers/video/uvesafb.c
index e328a61..b963ea1 100644
--- a/drivers/video/uvesafb.c
+++ b/drivers/video/uvesafb.c
@@ -24,9 +24,6 @@
 #ifdef CONFIG_X86
 #include <video/vga.h>
 #endif
-#ifdef CONFIG_MTRR
-#include <asm/mtrr.h>
-#endif
 #include "edid.h"
 
 static struct cb_id uvesafb_cn_id = {
@@ -819,8 +816,8 @@ static int uvesafb_vbe_init(struct fb_info *info)
 	if (par->pmi_setpal || par->ypan) {
 		if (__supported_pte_mask & _PAGE_NX) {
 			par->pmi_setpal = par->ypan = 0;
-			printk(KERN_WARNING "uvesafb: NX protection is actively."
-				"We have better not to use the PMI.\n");
+			printk(KERN_WARNING "uvesafb: NX protection is active, "
+					    "better not use the PMI.\n");
 		} else {
 			uvesafb_vbe_getpmi(task, par);
 		}
@@ -1540,67 +1537,30 @@ static void uvesafb_init_info(struct fb_info *info, struct vbe_mode_ib *mode)
 
 static void uvesafb_init_mtrr(struct fb_info *info)
 {
-#ifdef CONFIG_MTRR
+	struct uvesafb_par *par = info->par;
+
 	if (mtrr && !(info->fix.smem_start & (PAGE_SIZE - 1))) {
 		int temp_size = info->fix.smem_len;
-		unsigned int type = 0;
 
-		switch (mtrr) {
-		case 1:
-			type = MTRR_TYPE_UNCACHABLE;
-			break;
-		case 2:
-			type = MTRR_TYPE_WRBACK;
-			break;
-		case 3:
-			type = MTRR_TYPE_WRCOMB;
-			break;
-		case 4:
-			type = MTRR_TYPE_WRTHROUGH;
-			break;
-		default:
-			type = 0;
-			break;
-		}
+		int rc;
 
-		if (type) {
-			int rc;
+		/* Find the largest power-of-two */
+		temp_size = roundup_pow_of_two(temp_size);
 
-			/* Find the largest power-of-two */
-			temp_size = roundup_pow_of_two(temp_size);
+		/* Try and find a power of two to add */
+		do {
+			rc = arch_phys_wc_add(info->fix.smem_start, temp_size);
+			temp_size >>= 1;
+		} while (temp_size >= PAGE_SIZE && rc == -EINVAL);
 
-			/* Try and find a power of two to add */
-			do {
-				rc = mtrr_add(info->fix.smem_start,
-					      temp_size, type, 1);
-				temp_size >>= 1;
-			} while (temp_size >= PAGE_SIZE && rc == -EINVAL);
-		}
+		if (rc >= 0)
+			par->mtrr_handle = rc;
 	}
-#endif /* CONFIG_MTRR */
 }
 
 static void uvesafb_ioremap(struct fb_info *info)
 {
-#ifdef CONFIG_X86
-	switch (mtrr) {
-	case 1: /* uncachable */
-		info->screen_base = ioremap_nocache(info->fix.smem_start, info->fix.smem_len);
-		break;
-	case 2: /* write-back */
-		info->screen_base = ioremap_cache(info->fix.smem_start, info->fix.smem_len);
-		break;
-	case 3: /* write-combining */
-		info->screen_base = ioremap_wc(info->fix.smem_start, info->fix.smem_len);
-		break;
-	case 4: /* write-through */
-	default:
-		info->screen_base = ioremap(info->fix.smem_start, info->fix.smem_len);
-		break;
-	}
-#else
-	info->screen_base = ioremap(info->fix.smem_start, info->fix.smem_len);
-#endif /* CONFIG_X86 */
+	info->screen_base = ioremap_wc(info->fix.smem_start, info->fix.smem_len);
 }
 
 static ssize_t uvesafb_show_vbe_ver(struct device *dev,
@@ -1851,6 +1811,7 @@ static int uvesafb_remove(struct platform_device *dev)
 		unregister_framebuffer(info);
 		release_region(0x3c0, 32);
 		iounmap(info->screen_base);
+		arch_phys_wc_del(par->mtrr_handle);
 		release_mem_region(info->fix.smem_start, info->fix.smem_len);
 		fb_destroy_modedb(info->monspecs.modedb);
 		fb_dealloc_cmap(&info->cmap);
@@ -1930,6 +1891,9 @@ static int uvesafb_setup(char *options)
 		}
 	}
 
+	if (mtrr != 3 && mtrr != 1)
+		pr_warn("uvesafb: mtrr should be set to 0 or 3; %d is unsupported", mtrr);
+
 	return 0;
 }
 #endif /* !MODULE */
diff --git a/drivers/video/vga16fb.c b/drivers/video/vga16fb.c
index 545faec..830ded4 100644
--- a/drivers/video/vga16fb.c
+++ b/drivers/video/vga16fb.c
@@ -1269,7 +1269,6 @@ static void vga16fb_destroy(struct fb_info *info)
 	iounmap(info->screen_base);
 	fb_dealloc_cmap(&info->cmap);
 	/* XXX unshare VGA regions */
-	platform_set_drvdata(dev, NULL);
 	framebuffer_release(info);
 }
 
diff --git a/drivers/video/vt8500lcdfb.c b/drivers/video/vt8500lcdfb.c
index 9547e18..897484903 100644
--- a/drivers/video/vt8500lcdfb.c
+++ b/drivers/video/vt8500lcdfb.c
@@ -448,7 +448,6 @@ failed_free_io:
 failed_free_res:
 	release_mem_region(res->start, resource_size(res));
 failed_fbi:
-	platform_set_drvdata(pdev, NULL);
 	kfree(fbi);
 failed:
 	return ret;
diff --git a/drivers/video/wm8505fb.c b/drivers/video/wm8505fb.c
index 01f9ace..3072f30 100644
--- a/drivers/video/wm8505fb.c
+++ b/drivers/video/wm8505fb.c
@@ -173,7 +173,7 @@ static ssize_t contrast_store(struct device *dev,
 	struct wm8505fb_info *fbi = to_wm8505fb_info(info);
 	unsigned long tmp;
 
-	if (strict_strtoul(buf, 10, &tmp) || (tmp > 0xff))
+	if (kstrtoul(buf, 10, &tmp) || (tmp > 0xff))
 		return -EINVAL;
 	fbi->contrast = tmp;
 
diff --git a/drivers/video/xilinxfb.c b/drivers/video/xilinxfb.c
index af0b4fd..f3d4a69 100644
--- a/drivers/video/xilinxfb.c
+++ b/drivers/video/xilinxfb.c
@@ -44,7 +44,7 @@
 
 
 /*
- * Xilinx calls it "PLB TFT LCD Controller" though it can also be used for
+ * Xilinx calls it "TFT LCD Controller" though it can also be used for
  * the VGA port on the Xilinx ML40x board. This is a hardware display
  * controller for a 640x480 resolution TFT or VGA screen.
  *
@@ -54,11 +54,11 @@
  * don't start thinking about scrolling).  The second allows the LCD to
  * be turned on or off as well as rotated 180 degrees.
  *
- * In case of direct PLB access the second control register will be at
+ * In case of direct BUS access the second control register will be at
  * an offset of 4 as compared to the DCR access where the offset is 1
  * i.e. REG_CTRL. So this is taken care in the function
- * xilinx_fb_out_be32 where it left shifts the offset 2 times in case of
- * direct PLB access.
+ * xilinx_fb_out32 where it left shifts the offset 2 times in case of
+ * direct BUS access.
  */
 #define NUM_REGS	2
 #define REG_FB_ADDR	0
@@ -116,7 +116,8 @@ static struct fb_var_screeninfo xilinx_fb_var = {
 };
 
 
-#define PLB_ACCESS_FLAG	0x1		/* 1 = PLB, 0 = DCR */
+#define BUS_ACCESS_FLAG		0x1 /* 1 = BUS, 0 = DCR */
+#define LITTLE_ENDIAN_ACCESS	0x2 /* LITTLE ENDIAN IO functions */
 
 struct xilinxfb_drvdata {
 
@@ -146,21 +147,40 @@ struct xilinxfb_drvdata {
 	container_of(_info, struct xilinxfb_drvdata, info)
 
 /*
- * The XPS TFT Controller can be accessed through PLB or DCR interface.
+ * The XPS TFT Controller can be accessed through BUS or DCR interface.
  * To perform the read/write on the registers we need to check on
  * which bus its connected and call the appropriate write API.
  */
-static void xilinx_fb_out_be32(struct xilinxfb_drvdata *drvdata, u32 offset,
+static void xilinx_fb_out32(struct xilinxfb_drvdata *drvdata, u32 offset,
 				u32 val)
 {
-	if (drvdata->flags & PLB_ACCESS_FLAG)
-		out_be32(drvdata->regs + (offset << 2), val);
+	if (drvdata->flags & BUS_ACCESS_FLAG) {
+		if (drvdata->flags & LITTLE_ENDIAN_ACCESS)
+			iowrite32(val, drvdata->regs + (offset << 2));
+		else
+			iowrite32be(val, drvdata->regs + (offset << 2));
+	}
 #ifdef CONFIG_PPC_DCR
 	else
 		dcr_write(drvdata->dcr_host, offset, val);
 #endif
 }
 
+static u32 xilinx_fb_in32(struct xilinxfb_drvdata *drvdata, u32 offset)
+{
+	if (drvdata->flags & BUS_ACCESS_FLAG) {
+		if (drvdata->flags & LITTLE_ENDIAN_ACCESS)
+			return ioread32(drvdata->regs + (offset << 2));
+		else
+			return ioread32be(drvdata->regs + (offset << 2));
+	}
+#ifdef CONFIG_PPC_DCR
+	else
+		return dcr_read(drvdata->dcr_host, offset);
+#endif
+	return 0;
+}
+
 static int
 xilinx_fb_setcolreg(unsigned regno, unsigned red, unsigned green, unsigned blue,
 	unsigned transp, struct fb_info *fbi)
@@ -197,7 +217,7 @@ xilinx_fb_blank(int blank_mode, struct fb_info *fbi)
 	switch (blank_mode) {
 	case FB_BLANK_UNBLANK:
 		/* turn on panel */
-		xilinx_fb_out_be32(drvdata, REG_CTRL, drvdata->reg_ctrl_default);
+		xilinx_fb_out32(drvdata, REG_CTRL, drvdata->reg_ctrl_default);
 		break;
 
 	case FB_BLANK_NORMAL:
@@ -205,7 +225,7 @@ xilinx_fb_blank(int blank_mode, struct fb_info *fbi)
 	case FB_BLANK_HSYNC_SUSPEND:
 	case FB_BLANK_POWERDOWN:
 		/* turn off panel */
-		xilinx_fb_out_be32(drvdata, REG_CTRL, 0);
+		xilinx_fb_out32(drvdata, REG_CTRL, 0);
 	default:
 		break;
 
@@ -227,33 +247,23 @@ static struct fb_ops xilinxfb_ops =
  * Bus independent setup/teardown
  */
 
-static int xilinxfb_assign(struct device *dev,
+static int xilinxfb_assign(struct platform_device *pdev,
 			   struct xilinxfb_drvdata *drvdata,
-			   unsigned long physaddr,
 			   struct xilinxfb_platform_data *pdata)
 {
 	int rc;
+	struct device *dev = &pdev->dev;
 	int fbsize = pdata->xvirt * pdata->yvirt * BYTES_PER_PIXEL;
 
-	if (drvdata->flags & PLB_ACCESS_FLAG) {
-		/*
-		 * Map the control registers in if the controller
-		 * is on direct PLB interface.
-		 */
-		if (!request_mem_region(physaddr, 8, DRIVER_NAME)) {
-			dev_err(dev, "Couldn't lock memory region at 0x%08lX\n",
-				physaddr);
-			rc = -ENODEV;
-			goto err_region;
-		}
+	if (drvdata->flags & BUS_ACCESS_FLAG) {
+		struct resource *res;
 
-		drvdata->regs_phys = physaddr;
-		drvdata->regs = ioremap(physaddr, 8);
+		res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+		drvdata->regs_phys = res->start;
+		drvdata->regs = devm_request_and_ioremap(&pdev->dev, res);
 		if (!drvdata->regs) {
-			dev_err(dev, "Couldn't lock memory region at 0x%08lX\n",
-				physaddr);
-			rc = -ENODEV;
-			goto err_map;
+			rc = -EADDRNOTAVAIL;
+			goto err_region;
 		}
 	}
 
@@ -270,7 +280,7 @@ static int xilinxfb_assign(struct device *dev,
 	if (!drvdata->fb_virt) {
 		dev_err(dev, "Could not allocate frame buffer memory\n");
 		rc = -ENOMEM;
-		if (drvdata->flags & PLB_ACCESS_FLAG)
+		if (drvdata->flags & BUS_ACCESS_FLAG)
 			goto err_fbmem;
 		else
 			goto err_region;
@@ -280,13 +290,19 @@ static int xilinxfb_assign(struct device *dev,
 	memset_io((void __iomem *)drvdata->fb_virt, 0, fbsize);
 
 	/* Tell the hardware where the frame buffer is */
-	xilinx_fb_out_be32(drvdata, REG_FB_ADDR, drvdata->fb_phys);
+	xilinx_fb_out32(drvdata, REG_FB_ADDR, drvdata->fb_phys);
+	rc = xilinx_fb_in32(drvdata, REG_FB_ADDR);
+	/* Endianess detection */
+	if (rc != drvdata->fb_phys) {
+		drvdata->flags |= LITTLE_ENDIAN_ACCESS;
+		xilinx_fb_out32(drvdata, REG_FB_ADDR, drvdata->fb_phys);
+	}
 
 	/* Turn on the display */
 	drvdata->reg_ctrl_default = REG_CTRL_ENABLE;
 	if (pdata->rotate_screen)
 		drvdata->reg_ctrl_default |= REG_CTRL_ROTATE;
-	xilinx_fb_out_be32(drvdata, REG_CTRL,
+	xilinx_fb_out32(drvdata, REG_CTRL,
 					drvdata->reg_ctrl_default);
 
 	/* Fill struct fb_info */
@@ -323,9 +339,9 @@ static int xilinxfb_assign(struct device *dev,
 		goto err_regfb;
 	}
 
-	if (drvdata->flags & PLB_ACCESS_FLAG) {
+	if (drvdata->flags & BUS_ACCESS_FLAG) {
 		/* Put a banner in the log (for DEBUG) */
-		dev_dbg(dev, "regs: phys=%lx, virt=%p\n", physaddr,
+		dev_dbg(dev, "regs: phys=%x, virt=%p\n", drvdata->regs_phys,
 					drvdata->regs);
 	}
 	/* Put a banner in the log (for DEBUG) */
@@ -345,15 +361,11 @@ err_cmap:
 		iounmap(drvdata->fb_virt);
 
 	/* Turn off the display */
-	xilinx_fb_out_be32(drvdata, REG_CTRL, 0);
+	xilinx_fb_out32(drvdata, REG_CTRL, 0);
 
 err_fbmem:
-	if (drvdata->flags & PLB_ACCESS_FLAG)
-		iounmap(drvdata->regs);
-
-err_map:
-	if (drvdata->flags & PLB_ACCESS_FLAG)
-		release_mem_region(physaddr, 8);
+	if (drvdata->flags & BUS_ACCESS_FLAG)
+		devm_iounmap(dev, drvdata->regs);
 
 err_region:
 	kfree(drvdata);
@@ -381,13 +393,11 @@ static int xilinxfb_release(struct device *dev)
 		iounmap(drvdata->fb_virt);
 
 	/* Turn off the display */
-	xilinx_fb_out_be32(drvdata, REG_CTRL, 0);
+	xilinx_fb_out32(drvdata, REG_CTRL, 0);
 
 	/* Release the resources, as allocated based on interface */
-	if (drvdata->flags & PLB_ACCESS_FLAG) {
-		iounmap(drvdata->regs);
-		release_mem_region(drvdata->regs_phys, 8);
-	}
+	if (drvdata->flags & BUS_ACCESS_FLAG)
+		devm_iounmap(dev, drvdata->regs);
 #ifdef CONFIG_PPC_DCR
 	else
 		dcr_unmap(drvdata->dcr_host, drvdata->dcr_len);
@@ -406,11 +416,9 @@ static int xilinxfb_release(struct device *dev)
 static int xilinxfb_of_probe(struct platform_device *op)
 {
 	const u32 *prop;
-	u32 *p;
-	u32 tft_access;
+	u32 tft_access = 0;
 	struct xilinxfb_platform_data pdata;
-	struct resource res;
-	int size, rc;
+	int size;
 	struct xilinxfb_drvdata *drvdata;
 
 	/* Copy with the default pdata (not a ptr reference!) */
@@ -424,34 +432,29 @@ static int xilinxfb_of_probe(struct platform_device *op)
 	}
 
 	/*
-	 * To check whether the core is connected directly to DCR or PLB
+	 * To check whether the core is connected directly to DCR or BUS
 	 * interface and initialize the tft_access accordingly.
 	 */
-	p = (u32 *)of_get_property(op->dev.of_node, "xlnx,dcr-splb-slave-if", NULL);
-	tft_access = p ? *p : 0;
+	of_property_read_u32(op->dev.of_node, "xlnx,dcr-splb-slave-if",
+			     &tft_access);
 
 	/*
-	 * Fill the resource structure if its direct PLB interface
+	 * Fill the resource structure if its direct BUS interface
 	 * otherwise fill the dcr_host structure.
 	 */
 	if (tft_access) {
-		drvdata->flags |= PLB_ACCESS_FLAG;
-		rc = of_address_to_resource(op->dev.of_node, 0, &res);
-		if (rc) {
-			dev_err(&op->dev, "invalid address\n");
-			goto err;
-		}
+		drvdata->flags |= BUS_ACCESS_FLAG;
 	}
 #ifdef CONFIG_PPC_DCR
 	else {
 		int start;
-		res.start = 0;
 		start = dcr_resource_start(op->dev.of_node, 0);
 		drvdata->dcr_len = dcr_resource_len(op->dev.of_node, 0);
 		drvdata->dcr_host = dcr_map(op->dev.of_node, start, drvdata->dcr_len);
 		if (!DCR_MAP_OK(drvdata->dcr_host)) {
 			dev_err(&op->dev, "invalid DCR address\n");
-			goto err;
+			kfree(drvdata);
+			return -ENODEV;
 		}
 	}
 #endif
@@ -478,11 +481,7 @@ static int xilinxfb_of_probe(struct platform_device *op)
 		pdata.rotate_screen = 1;
 
 	dev_set_drvdata(&op->dev, drvdata);
-	return xilinxfb_assign(&op->dev, drvdata, res.start, &pdata);
-
- err:
-	kfree(drvdata);
-	return -ENODEV;
+	return xilinxfb_assign(op, drvdata, &pdata);
 }
 
 static int xilinxfb_of_remove(struct platform_device *op)
diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index c702074..6b4a4db 100644
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -576,19 +576,21 @@ void virtqueue_disable_cb(struct virtqueue *_vq)
 EXPORT_SYMBOL_GPL(virtqueue_disable_cb);
 
 /**
- * virtqueue_enable_cb - restart callbacks after disable_cb.
+ * virtqueue_enable_cb_prepare - restart callbacks after disable_cb
  * @vq: the struct virtqueue we're talking about.
  *
- * This re-enables callbacks; it returns "false" if there are pending
- * buffers in the queue, to detect a possible race between the driver
- * checking for more work, and enabling callbacks.
+ * This re-enables callbacks; it returns current queue state
+ * in an opaque unsigned value. This value should be later tested by
+ * virtqueue_poll, to detect a possible race between the driver checking for
+ * more work, and enabling callbacks.
  *
  * Caller must ensure we don't call this with other virtqueue
  * operations at the same time (except where noted).
  */
-bool virtqueue_enable_cb(struct virtqueue *_vq)
+unsigned virtqueue_enable_cb_prepare(struct virtqueue *_vq)
 {
 	struct vring_virtqueue *vq = to_vvq(_vq);
+	u16 last_used_idx;
 
 	START_USE(vq);
 
@@ -598,15 +600,45 @@ bool virtqueue_enable_cb(struct virtqueue *_vq)
 	 * either clear the flags bit or point the event index at the next
 	 * entry. Always do both to keep code simple. */
 	vq->vring.avail->flags &= ~VRING_AVAIL_F_NO_INTERRUPT;
-	vring_used_event(&vq->vring) = vq->last_used_idx;
+	vring_used_event(&vq->vring) = last_used_idx = vq->last_used_idx;
+	END_USE(vq);
+	return last_used_idx;
+}
+EXPORT_SYMBOL_GPL(virtqueue_enable_cb_prepare);
+
+/**
+ * virtqueue_poll - query pending used buffers
+ * @vq: the struct virtqueue we're talking about.
+ * @last_used_idx: virtqueue state (from call to virtqueue_enable_cb_prepare).
+ *
+ * Returns "true" if there are pending used buffers in the queue.
+ *
+ * This does not need to be serialized.
+ */
+bool virtqueue_poll(struct virtqueue *_vq, unsigned last_used_idx)
+{
+	struct vring_virtqueue *vq = to_vvq(_vq);
+
 	virtio_mb(vq->weak_barriers);
-	if (unlikely(more_used(vq))) {
-		END_USE(vq);
-		return false;
-	}
+	return (u16)last_used_idx != vq->vring.used->idx;
+}
+EXPORT_SYMBOL_GPL(virtqueue_poll);
 
-	END_USE(vq);
-	return true;
+/**
+ * virtqueue_enable_cb - restart callbacks after disable_cb.
+ * @vq: the struct virtqueue we're talking about.
+ *
+ * This re-enables callbacks; it returns "false" if there are pending
+ * buffers in the queue, to detect a possible race between the driver
+ * checking for more work, and enabling callbacks.
+ *
+ * Caller must ensure we don't call this with other virtqueue
+ * operations at the same time (except where noted).
+ */
+bool virtqueue_enable_cb(struct virtqueue *_vq)
+{
+	unsigned last_used_idx = virtqueue_enable_cb_prepare(_vq);
+	return !virtqueue_poll(_vq, last_used_idx);
 }
 EXPORT_SYMBOL_GPL(virtqueue_enable_cb);